Page MenuHomeFreeBSD
Files F83601667

No OneTemporary
Actions

View Options

This file is larger than 256 KB, so syntax highlighting was skipped.
diff --git a/cddl/lib/libzpool/Makefile b/cddl/lib/libzpool/Makefile
index d05babdc26e0..c801cf571e88 100644
--- a/cddl/lib/libzpool/Makefile
+++ b/cddl/lib/libzpool/Makefile
@@ -1,336 +1,337 @@
ZFSTOP= ${SRCTOP}/sys/contrib/openzfs
# ZFS_COMMON_SRCS
.PATH: ${ZFSTOP}/module/zfs
.PATH: ${ZFSTOP}/module/zcommon
.PATH: ${ZFSTOP}/module/unicode
# LUA_SRCS
.PATH: ${ZFSTOP}/module/lua
# ZSTD_SRCS
.PATH: ${ZFSTOP}/module/zstd
.PATH: ${ZFSTOP}/module/zstd/lib/common
.PATH: ${ZFSTOP}/module/zstd/lib/compress
.PATH: ${ZFSTOP}/module/zstd/lib/decompress
.PATH: ${ZFSTOP}/module/os/linux/zfs
.PATH: ${ZFSTOP}/lib/libzpool
.if exists(${SRCTOP}/sys/cddl/contrib/opensolaris/common/atomic/${MACHINE_ARCH}/opensolaris_atomic.S)
.PATH: ${SRCTOP}/sys/cddl/contrib/opensolaris/common/atomic/${MACHINE_ARCH}
ATOMIC_SRCS= opensolaris_atomic.S
ACFLAGS+= -Wa,--noexecstack
.else
.PATH: ${SRCTOP}/sys/cddl/compat/opensolaris/kern
ATOMIC_SRCS= opensolaris_atomic.c
.endif
.if ${MACHINE_ARCH} == "powerpc" || ${MACHINE_ARCH} == "powerpcspe"
# Don't waste GOT entries on small data.
PICFLAG= -fPIC
.endif
PACKAGE= zfs
LIB= zpool
USER_C = \
kernel.c \
taskq.c \
util.c
KERNEL_C = \
zfeature_common.c \
zfs_comutil.c \
zfs_deleg.c \
zfs_fletcher.c \
zfs_fletcher_superscalar.c \
zfs_fletcher_superscalar4.c \
zfs_namecheck.c \
zfs_prop.c \
zfs_zstd.c \
zpool_prop.c \
zprop_common.c \
abd.c \
abd_os.c \
aggsum.c \
arc.c \
arc_os.c \
blake3_zfs.c \
blkptr.c \
bplist.c \
bpobj.c \
bptree.c \
bqueue.c \
btree.c \
brt.c \
cityhash.c \
dbuf.c \
dbuf_stats.c \
ddt.c \
ddt_zap.c \
dmu.c \
dmu_diff.c \
dmu_object.c \
dmu_objset.c \
dmu_recv.c \
dmu_redact.c \
dmu_send.c \
dmu_traverse.c \
dmu_tx.c \
dmu_zfetch.c \
dnode.c \
dnode_sync.c \
dsl_bookmark.c \
dsl_dataset.c \
dsl_deadlist.c \
dsl_deleg.c \
dsl_dir.c \
dsl_crypt.c \
dsl_pool.c \
dsl_prop.c \
dsl_scan.c \
dsl_synctask.c \
dsl_destroy.c \
dsl_userhold.c \
edonr_zfs.c \
entropy_common.c \
error_private.c \
fm.c \
fse_compress.c \
fse_decompress.c \
gzip.c \
hist.c \
hkdf.c \
huf_compress.c \
huf_decompress.c \
lzjb.c \
lz4.c \
lz4_zfs.c \
metaslab.c \
mmp.c \
multilist.c \
objlist.c \
pathname.c \
pool.c \
range_tree.c \
refcount.c \
rrwlock.c \
sa.c \
sha2_zfs.c \
skein_zfs.c \
spa.c \
spa_checkpoint.c \
spa_config.c \
spa_errlog.c \
spa_history.c \
spa_log_spacemap.c \
spa_misc.c \
spa_stats.c \
space_map.c \
space_reftree.c \
txg.c \
trace.c \
uberblock.c \
unique.c \
vdev.c \
vdev_draid.c \
vdev_draid_rand.c \
vdev_file.c \
vdev_indirect_births.c \
vdev_indirect.c \
vdev_indirect_mapping.c \
vdev_initialize.c \
vdev_label.c \
+ vdev_label_os.c \
vdev_mirror.c \
vdev_missing.c \
vdev_queue.c \
vdev_raidz.c \
vdev_raidz_math_aarch64_neon.c \
vdev_raidz_math_aarch64_neonx2.c \
vdev_raidz_math_avx2.c \
vdev_raidz_math_avx512bw.c \
vdev_raidz_math_avx512f.c \
vdev_raidz_math.c \
vdev_raidz_math_scalar.c \
vdev_rebuild.c \
vdev_removal.c \
vdev_root.c \
vdev_trim.c \
xxhash.c \
zap.c \
zap_leaf.c \
zap_micro.c \
zcp.c \
zcp_get.c \
zcp_global.c \
zcp_iter.c \
zcp_set.c \
zcp_synctask.c \
zfeature.c \
zfs_byteswap.c \
zfs_chksum.c \
zfs_debug.c \
zfs_fm.c \
zfs_fuid.c \
zfs_sa.c \
zfs_znode.c \
zfs_racct.c \
zfs_ratelimit.c \
zfs_rlock.c \
zil.c \
zio.c \
zio_checksum.c \
zio_compress.c \
zio_crypt.c \
zio_inject.c \
zle.c \
zrlock.c \
zstd_common.c \
zstd_compress.c \
zstd_compress_literals.c \
zstd_compress_sequences.c \
zstd_compress_superblock.c \
zstd_ddict.c \
zstd_decompress.c \
zstd_decompress_block.c \
zstd_double_fast.c \
zstd_fast.c \
zstd_lazy.c \
zstd_ldm.c \
zstd_opt.c \
zthr.c
ARCH_C =
.if ${MACHINE_ARCH} == "amd64" || ${MACHINE_ARCH} == "i386"
ARCH_C += vdev_raidz_math_sse2.c \
vdev_raidz_math_ssse3.c \
zfs_fletcher_intel.c \
zfs_fletcher_sse.c
CFLAGS += -DHAVE_SSE2 -DHAVE_SSE3
.endif
.if ${MACHINE_ARCH} == "amd64"
ARCH_C += zfs_fletcher_avx512.c
CFLAGS+= -DHAVE_AVX2 -DHAVE_AVX -D__x86_64 -DHAVE_AVX512F \
-DHAVE_AVX512BW
.endif
.if ${MACHINE_CPUARCH} == "aarch64"
ARCH_C += zfs_fletcher_aarch64_neon.c
.endif
LUA_C = \
lapi.c \
lauxlib.c \
lbaselib.c \
lcode.c \
lcompat.c \
lcorolib.c \
lctype.c \
ldebug.c \
ldo.c \
lfunc.c \
lgc.c \
llex.c \
lmem.c \
lobject.c \
lopcodes.c \
lparser.c \
lstate.c \
lstring.c \
lstrlib.c \
ltable.c \
ltablib.c \
ltm.c \
lvm.c \
lzio.c
UNICODE_C = u8_textprep.c uconv.c
SRCS= ${USER_C} ${KERNEL_C} ${LUA_C} ${UNICODE_C} ${ARCH_C}
WARNS?= 2
CFLAGS+= \
-DIN_BASE \
-I${ZFSTOP}/include \
-I${ZFSTOP}/lib/libspl/include \
-I${ZFSTOP}/lib/libspl/include/os/freebsd \
-I${SRCTOP}/sys \
-I${ZFSTOP}/include/os/freebsd/zfs \
-I${SRCTOP}/cddl/compat/opensolaris/include \
-I${ZFSTOP}/module/icp/include \
-include ${ZFSTOP}/include/os/freebsd/spl/sys/ccompile.h \
-DHAVE_ISSETUGID \
-include ${SRCTOP}/sys/modules/zfs/zfs_config.h \
-I${SRCTOP}/sys/modules/zfs \
-I${ZFSTOP}/include/os/freebsd/zfs \
-DLIB_ZPOOL_BUILD -DZFS_DEBUG \
# XXX: pthread doesn't have mutex_owned() equivalent, so we need to look
# into libthr private structures. That's sooo evil, but it's only for
# ZFS debugging tools needs.
CFLAGS+= -DWANTS_MUTEX_OWNED
CFLAGS+= -I${SRCTOP}/lib/libpthread/thread
CFLAGS+= -I${SRCTOP}/lib/libpthread/sys
CFLAGS+= -I${SRCTOP}/lib/libthr/arch/${MACHINE_CPUARCH}/include
CFLAGS.gcc+= -fms-extensions
LIBADD= md pthread z spl icp nvpair avl umem
# atomic.S doesn't like profiling.
MK_PROFILE= no
CSTD= c99
# Since there are many asserts in this library, it makes no sense to compile
# it without debugging.
CFLAGS+= -g -DDEBUG=1
# Pointer values are used as debugging "tags" to mark reference count
# ownerships and in some cases the tag reference is dropped after an
# object is freed.
CFLAGS.dbuf.c= ${NO_WUSE_AFTER_FREE}
CFLAGS.entropy_common.c= -fno-tree-vectorize
CFLAGS.entropy_common.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.error_private.c= -fno-tree-vectorize
CFLAGS.error_private.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.fse_compress.c= -fno-tree-vectorize
CFLAGS.fse_compress.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.fse_decompress.c= -fno-tree-vectorize
CFLAGS.fse_decompress.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.hist.c= -fno-tree-vectorize
CFLAGS.hist.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.huf_compress.c= -fno-tree-vectorize
CFLAGS.huf_compress.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.huf_decompress.c= -fno-tree-vectorize
CFLAGS.huf_decompress.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.pool.c= -fno-tree-vectorize
CFLAGS.pool.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.xxhash.c= -fno-tree-vectorize
CFLAGS.xxhash.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_compress.c= -fno-tree-vectorize
CFLAGS.zstd_compress.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_compress_literals.c= -fno-tree-vectorize
CFLAGS.zstd_compress_literals.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_compress_sequences.c= -fno-tree-vectorize
CFLAGS.zstd_compress_sequences.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_compress_superblock.c= -fno-tree-vectorize
CFLAGS.zstd_compress_superblock.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_double_fast.c= -fno-tree-vectorize
CFLAGS.zstd_double_fast.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_fast.c= -fno-tree-vectorize
CFLAGS.zstd_fast.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_lazy.c= -fno-tree-vectorize
CFLAGS.zstd_lazy.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_ldm.c= -fno-tree-vectorize
CFLAGS.zstd_ldm.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_opt.c= -fno-tree-vectorize
CFLAGS.zstd_opt.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_ddict.c= -fno-tree-vectorize
CFLAGS.zstd_ddict.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_decompress.c= -fno-tree-vectorize
CFLAGS.zstd_decompress.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_decompress_block.c= -fno-tree-vectorize
CFLAGS.zstd_decompress_block.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
.include <bsd.lib.mk>
diff --git a/sys/contrib/openzfs/cmd/raidz_test/raidz_bench.c b/sys/contrib/openzfs/cmd/raidz_test/raidz_bench.c
index 8be08558b36d..730e6e1a040b 100644
--- a/sys/contrib/openzfs/cmd/raidz_test/raidz_bench.c
+++ b/sys/contrib/openzfs/cmd/raidz_test/raidz_bench.c
@@ -1,242 +1,242 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (C) 2016 Gvozden Nešković. All rights reserved.
*/
#include <sys/zfs_context.h>
#include <sys/time.h>
#include <sys/wait.h>
#include <sys/zio.h>
#include <sys/vdev_raidz.h>
#include <sys/vdev_raidz_impl.h>
#include <stdio.h>
#include "raidz_test.h"
#define GEN_BENCH_MEMORY (((uint64_t)1ULL)<<32)
#define REC_BENCH_MEMORY (((uint64_t)1ULL)<<29)
#define BENCH_ASHIFT 12
#define MIN_CS_SHIFT BENCH_ASHIFT
#define MAX_CS_SHIFT SPA_MAXBLOCKSHIFT
static zio_t zio_bench;
static raidz_map_t *rm_bench;
static size_t max_data_size = SPA_MAXBLOCKSIZE;
static void
bench_init_raidz_map(void)
{
zio_bench.io_offset = 0;
zio_bench.io_size = max_data_size;
/*
* To permit larger column sizes these have to be done
* allocated using aligned alloc instead of zio_abd_buf_alloc
*/
zio_bench.io_abd = raidz_alloc(max_data_size);
init_zio_abd(&zio_bench);
}
static void
bench_fini_raidz_maps(void)
{
/* tear down golden zio */
raidz_free(zio_bench.io_abd, max_data_size);
memset(&zio_bench, 0, sizeof (zio_t));
}
static inline void
run_gen_bench_impl(const char *impl)
{
int fn, ncols;
uint64_t ds, iter_cnt, iter, disksize;
hrtime_t start;
double elapsed, d_bw;
/* Benchmark generate functions */
for (fn = 0; fn < RAIDZ_GEN_NUM; fn++) {
for (ds = MIN_CS_SHIFT; ds <= MAX_CS_SHIFT; ds++) {
/* create suitable raidz_map */
ncols = rto_opts.rto_dcols + fn + 1;
zio_bench.io_size = 1ULL << ds;
if (rto_opts.rto_expand) {
rm_bench = vdev_raidz_map_alloc_expanded(
- zio_bench.io_abd,
- zio_bench.io_size, zio_bench.io_offset,
+ &zio_bench,
rto_opts.rto_ashift, ncols+1, ncols,
- fn+1, rto_opts.rto_expand_offset);
+ fn+1, rto_opts.rto_expand_offset,
+ 0, B_FALSE);
} else {
rm_bench = vdev_raidz_map_alloc(&zio_bench,
BENCH_ASHIFT, ncols, fn+1);
}
/* estimate iteration count */
iter_cnt = GEN_BENCH_MEMORY;
iter_cnt /= zio_bench.io_size;
start = gethrtime();
for (iter = 0; iter < iter_cnt; iter++)
vdev_raidz_generate_parity(rm_bench);
elapsed = NSEC2SEC((double)(gethrtime() - start));
disksize = (1ULL << ds) / rto_opts.rto_dcols;
d_bw = (double)iter_cnt * (double)disksize;
d_bw /= (1024.0 * 1024.0 * elapsed);
LOG(D_ALL, "%10s, %8s, %zu, %10llu, %lf, %lf, %u\n",
impl,
raidz_gen_name[fn],
rto_opts.rto_dcols,
(1ULL<<ds),
d_bw,
d_bw * (double)(ncols),
(unsigned)iter_cnt);
vdev_raidz_map_free(rm_bench);
}
}
}
static void
run_gen_bench(void)
{
char **impl_name;
LOG(D_INFO, DBLSEP "\nBenchmarking parity generation...\n\n");
LOG(D_ALL, "impl, math, dcols, iosize, disk_bw, total_bw, iter\n");
for (impl_name = (char **)raidz_impl_names; *impl_name != NULL;
impl_name++) {
if (vdev_raidz_impl_set(*impl_name) != 0)
continue;
run_gen_bench_impl(*impl_name);
}
}
static void
run_rec_bench_impl(const char *impl)
{
int fn, ncols, nbad;
uint64_t ds, iter_cnt, iter, disksize;
hrtime_t start;
double elapsed, d_bw;
static const int tgt[7][3] = {
{1, 2, 3}, /* rec_p: bad QR & D[0] */
{0, 2, 3}, /* rec_q: bad PR & D[0] */
{0, 1, 3}, /* rec_r: bad PQ & D[0] */
{2, 3, 4}, /* rec_pq: bad R & D[0][1] */
{1, 3, 4}, /* rec_pr: bad Q & D[0][1] */
{0, 3, 4}, /* rec_qr: bad P & D[0][1] */
{3, 4, 5} /* rec_pqr: bad & D[0][1][2] */
};
for (fn = 0; fn < RAIDZ_REC_NUM; fn++) {
for (ds = MIN_CS_SHIFT; ds <= MAX_CS_SHIFT; ds++) {
/* create suitable raidz_map */
ncols = rto_opts.rto_dcols + PARITY_PQR;
zio_bench.io_size = 1ULL << ds;
/*
* raidz block is too short to test
* the requested method
*/
if (zio_bench.io_size / rto_opts.rto_dcols <
(1ULL << BENCH_ASHIFT))
continue;
if (rto_opts.rto_expand) {
rm_bench = vdev_raidz_map_alloc_expanded(
- zio_bench.io_abd,
- zio_bench.io_size, zio_bench.io_offset,
+ &zio_bench,
BENCH_ASHIFT, ncols+1, ncols,
- PARITY_PQR, rto_opts.rto_expand_offset);
+ PARITY_PQR,
+ rto_opts.rto_expand_offset, 0, B_FALSE);
} else {
rm_bench = vdev_raidz_map_alloc(&zio_bench,
BENCH_ASHIFT, ncols, PARITY_PQR);
}
/* estimate iteration count */
iter_cnt = (REC_BENCH_MEMORY);
iter_cnt /= zio_bench.io_size;
/* calculate how many bad columns there are */
nbad = MIN(3, raidz_ncols(rm_bench) -
raidz_parity(rm_bench));
start = gethrtime();
for (iter = 0; iter < iter_cnt; iter++)
vdev_raidz_reconstruct(rm_bench, tgt[fn], nbad);
elapsed = NSEC2SEC((double)(gethrtime() - start));
disksize = (1ULL << ds) / rto_opts.rto_dcols;
d_bw = (double)iter_cnt * (double)(disksize);
d_bw /= (1024.0 * 1024.0 * elapsed);
LOG(D_ALL, "%10s, %8s, %zu, %10llu, %lf, %lf, %u\n",
impl,
raidz_rec_name[fn],
rto_opts.rto_dcols,
(1ULL<<ds),
d_bw,
d_bw * (double)ncols,
(unsigned)iter_cnt);
vdev_raidz_map_free(rm_bench);
}
}
}
static void
run_rec_bench(void)
{
char **impl_name;
LOG(D_INFO, DBLSEP "\nBenchmarking data reconstruction...\n\n");
LOG(D_ALL, "impl, math, dcols, iosize, disk_bw, total_bw, iter\n");
for (impl_name = (char **)raidz_impl_names; *impl_name != NULL;
impl_name++) {
if (vdev_raidz_impl_set(*impl_name) != 0)
continue;
run_rec_bench_impl(*impl_name);
}
}
void
run_raidz_benchmark(void)
{
bench_init_raidz_map();
run_gen_bench();
run_rec_bench();
bench_fini_raidz_maps();
}
diff --git a/sys/contrib/openzfs/cmd/raidz_test/raidz_test.c b/sys/contrib/openzfs/cmd/raidz_test/raidz_test.c
index 195026d3a7ab..6a018ecf0737 100644
--- a/sys/contrib/openzfs/cmd/raidz_test/raidz_test.c
+++ b/sys/contrib/openzfs/cmd/raidz_test/raidz_test.c
@@ -1,1024 +1,840 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (C) 2016 Gvozden Nešković. All rights reserved.
*/
#include <sys/zfs_context.h>
#include <sys/time.h>
#include <sys/wait.h>
#include <sys/zio.h>
#include <umem.h>
#include <sys/vdev_raidz.h>
#include <sys/vdev_raidz_impl.h>
#include <assert.h>
#include <stdio.h>
#include "raidz_test.h"
static int *rand_data;
raidz_test_opts_t rto_opts;
static char pid_s[16];
static void sig_handler(int signo)
{
int old_errno = errno;
struct sigaction action;
/*
* Restore default action and re-raise signal so SIGSEGV and
* SIGABRT can trigger a core dump.
*/
action.sa_handler = SIG_DFL;
sigemptyset(&action.sa_mask);
action.sa_flags = 0;
(void) sigaction(signo, &action, NULL);
if (rto_opts.rto_gdb) {
pid_t pid = fork();
if (pid == 0) {
execlp("gdb", "gdb", "-ex", "set pagination 0",
"-p", pid_s, NULL);
_exit(-1);
} else if (pid > 0)
while (waitpid(pid, NULL, 0) == -1 && errno == EINTR)
;
}
raise(signo);
errno = old_errno;
}
static void print_opts(raidz_test_opts_t *opts, boolean_t force)
{
const char *verbose;
switch (opts->rto_v) {
case D_ALL:
verbose = "no";
break;
case D_INFO:
verbose = "info";
break;
case D_DEBUG:
default:
verbose = "debug";
break;
}
if (force || opts->rto_v >= D_INFO) {
(void) fprintf(stdout, DBLSEP "Running with options:\n"
" (-a) zio ashift : %zu\n"
" (-o) zio offset : 1 << %zu\n"
" (-e) expanded map : %s\n"
" (-r) reflow offset : %llx\n"
" (-d) number of raidz data columns : %zu\n"
" (-s) size of DATA : 1 << %zu\n"
" (-S) sweep parameters : %s \n"
" (-v) verbose : %s \n\n",
opts->rto_ashift, /* -a */
ilog2(opts->rto_offset), /* -o */
opts->rto_expand ? "yes" : "no", /* -e */
(u_longlong_t)opts->rto_expand_offset, /* -r */
opts->rto_dcols, /* -d */
ilog2(opts->rto_dsize), /* -s */
opts->rto_sweep ? "yes" : "no", /* -S */
verbose); /* -v */
}
}
static void usage(boolean_t requested)
{
const raidz_test_opts_t *o = &rto_opts_defaults;
FILE *fp = requested ? stdout : stderr;
(void) fprintf(fp, "Usage:\n"
"\t[-a zio ashift (default: %zu)]\n"
"\t[-o zio offset, exponent radix 2 (default: %zu)]\n"
"\t[-d number of raidz data columns (default: %zu)]\n"
"\t[-s zio size, exponent radix 2 (default: %zu)]\n"
"\t[-S parameter sweep (default: %s)]\n"
"\t[-t timeout for parameter sweep test]\n"
"\t[-B benchmark all raidz implementations]\n"
"\t[-e use expanded raidz map (default: %s)]\n"
"\t[-r expanded raidz map reflow offset (default: %llx)]\n"
"\t[-v increase verbosity (default: %d)]\n"
"\t[-h (print help)]\n"
"\t[-T test the test, see if failure would be detected]\n"
"\t[-D debug (attach gdb on SIGSEGV)]\n"
"",
o->rto_ashift, /* -a */
ilog2(o->rto_offset), /* -o */
o->rto_dcols, /* -d */
ilog2(o->rto_dsize), /* -s */
rto_opts.rto_sweep ? "yes" : "no", /* -S */
rto_opts.rto_expand ? "yes" : "no", /* -e */
(u_longlong_t)o->rto_expand_offset, /* -r */
o->rto_v); /* -v */
exit(requested ? 0 : 1);
}
static void process_options(int argc, char **argv)
{
size_t value;
int opt;
raidz_test_opts_t *o = &rto_opts;
memcpy(o, &rto_opts_defaults, sizeof (*o));
while ((opt = getopt(argc, argv, "TDBSvha:er:o:d:s:t:")) != -1) {
switch (opt) {
case 'a':
value = strtoull(optarg, NULL, 0);
o->rto_ashift = MIN(13, MAX(9, value));
break;
case 'e':
o->rto_expand = 1;
break;
case 'r':
o->rto_expand_offset = strtoull(optarg, NULL, 0);
break;
case 'o':
value = strtoull(optarg, NULL, 0);
o->rto_offset = ((1ULL << MIN(12, value)) >> 9) << 9;
break;
case 'd':
value = strtoull(optarg, NULL, 0);
o->rto_dcols = MIN(255, MAX(1, value));
break;
case 's':
value = strtoull(optarg, NULL, 0);
o->rto_dsize = 1ULL << MIN(SPA_MAXBLOCKSHIFT,
MAX(SPA_MINBLOCKSHIFT, value));
break;
case 't':
value = strtoull(optarg, NULL, 0);
o->rto_sweep_timeout = value;
break;
case 'v':
o->rto_v++;
break;
case 'S':
o->rto_sweep = 1;
break;
case 'B':
o->rto_benchmark = 1;
break;
case 'D':
o->rto_gdb = 1;
break;
case 'T':
o->rto_sanity = 1;
break;
case 'h':
usage(B_TRUE);
break;
case '?':
default:
usage(B_FALSE);
break;
}
}
}
#define DATA_COL(rr, i) ((rr)->rr_col[rr->rr_firstdatacol + (i)].rc_abd)
#define DATA_COL_SIZE(rr, i) ((rr)->rr_col[rr->rr_firstdatacol + (i)].rc_size)
#define CODE_COL(rr, i) ((rr)->rr_col[(i)].rc_abd)
#define CODE_COL_SIZE(rr, i) ((rr)->rr_col[(i)].rc_size)
static int
cmp_code(raidz_test_opts_t *opts, const raidz_map_t *rm, const int parity)
{
int r, i, ret = 0;
VERIFY(parity >= 1 && parity <= 3);
for (r = 0; r < rm->rm_nrows; r++) {
raidz_row_t * const rr = rm->rm_row[r];
raidz_row_t * const rrg = opts->rm_golden->rm_row[r];
for (i = 0; i < parity; i++) {
if (CODE_COL_SIZE(rrg, i) == 0) {
VERIFY0(CODE_COL_SIZE(rr, i));
continue;
}
if (abd_cmp(CODE_COL(rr, i),
CODE_COL(rrg, i)) != 0) {
ret++;
LOG_OPT(D_DEBUG, opts,
"\nParity block [%d] different!\n", i);
}
}
}
return (ret);
}
static int
cmp_data(raidz_test_opts_t *opts, raidz_map_t *rm)
{
int r, i, dcols, ret = 0;
for (r = 0; r < rm->rm_nrows; r++) {
raidz_row_t *rr = rm->rm_row[r];
raidz_row_t *rrg = opts->rm_golden->rm_row[r];
dcols = opts->rm_golden->rm_row[0]->rr_cols -
raidz_parity(opts->rm_golden);
for (i = 0; i < dcols; i++) {
if (DATA_COL_SIZE(rrg, i) == 0) {
VERIFY0(DATA_COL_SIZE(rr, i));
continue;
}
if (abd_cmp(DATA_COL(rrg, i),
DATA_COL(rr, i)) != 0) {
ret++;
LOG_OPT(D_DEBUG, opts,
"\nData block [%d] different!\n", i);
}
}
}
return (ret);
}
static int
init_rand(void *data, size_t size, void *private)
{
(void) private;
memcpy(data, rand_data, size);
return (0);
}
static void
corrupt_colums(raidz_map_t *rm, const int *tgts, const int cnt)
{
for (int r = 0; r < rm->rm_nrows; r++) {
raidz_row_t *rr = rm->rm_row[r];
for (int i = 0; i < cnt; i++) {
raidz_col_t *col = &rr->rr_col[tgts[i]];
abd_iterate_func(col->rc_abd, 0, col->rc_size,
init_rand, NULL);
}
}
}
void
init_zio_abd(zio_t *zio)
{
abd_iterate_func(zio->io_abd, 0, zio->io_size, init_rand, NULL);
}
static void
fini_raidz_map(zio_t **zio, raidz_map_t **rm)
{
vdev_raidz_map_free(*rm);
raidz_free((*zio)->io_abd, (*zio)->io_size);
umem_free(*zio, sizeof (zio_t));
*zio = NULL;
*rm = NULL;
}
static int
init_raidz_golden_map(raidz_test_opts_t *opts, const int parity)
{
int err = 0;
zio_t *zio_test;
raidz_map_t *rm_test;
const size_t total_ncols = opts->rto_dcols + parity;
if (opts->rm_golden) {
fini_raidz_map(&opts->zio_golden, &opts->rm_golden);
}
opts->zio_golden = umem_zalloc(sizeof (zio_t), UMEM_NOFAIL);
zio_test = umem_zalloc(sizeof (zio_t), UMEM_NOFAIL);
opts->zio_golden->io_offset = zio_test->io_offset = opts->rto_offset;
opts->zio_golden->io_size = zio_test->io_size = opts->rto_dsize;
opts->zio_golden->io_abd = raidz_alloc(opts->rto_dsize);
zio_test->io_abd = raidz_alloc(opts->rto_dsize);
init_zio_abd(opts->zio_golden);
init_zio_abd(zio_test);
VERIFY0(vdev_raidz_impl_set("original"));
if (opts->rto_expand) {
opts->rm_golden =
- vdev_raidz_map_alloc_expanded(opts->zio_golden->io_abd,
- opts->zio_golden->io_size, opts->zio_golden->io_offset,
+ vdev_raidz_map_alloc_expanded(opts->zio_golden,
opts->rto_ashift, total_ncols+1, total_ncols,
- parity, opts->rto_expand_offset);
- rm_test = vdev_raidz_map_alloc_expanded(zio_test->io_abd,
- zio_test->io_size, zio_test->io_offset,
+ parity, opts->rto_expand_offset, 0, B_FALSE);
+ rm_test = vdev_raidz_map_alloc_expanded(zio_test,
opts->rto_ashift, total_ncols+1, total_ncols,
- parity, opts->rto_expand_offset);
+ parity, opts->rto_expand_offset, 0, B_FALSE);
} else {
opts->rm_golden = vdev_raidz_map_alloc(opts->zio_golden,
opts->rto_ashift, total_ncols, parity);
rm_test = vdev_raidz_map_alloc(zio_test,
opts->rto_ashift, total_ncols, parity);
}
VERIFY(opts->zio_golden);
VERIFY(opts->rm_golden);
vdev_raidz_generate_parity(opts->rm_golden);
vdev_raidz_generate_parity(rm_test);
/* sanity check */
err |= cmp_data(opts, rm_test);
err |= cmp_code(opts, rm_test, parity);
if (err)
ERR("initializing the golden copy ... [FAIL]!\n");
/* tear down raidz_map of test zio */
fini_raidz_map(&zio_test, &rm_test);
return (err);
}
-/*
- * If reflow is not in progress, reflow_offset should be UINT64_MAX.
- * For each row, if the row is entirely before reflow_offset, it will
- * come from the new location. Otherwise this row will come from the
- * old location. Therefore, rows that straddle the reflow_offset will
- * come from the old location.
- *
- * NOTE: Until raidz expansion is implemented this function is only
- * needed by raidz_test.c to the multi-row raid_map_t functionality.
- */
-raidz_map_t *
-vdev_raidz_map_alloc_expanded(abd_t *abd, uint64_t size, uint64_t offset,
- uint64_t ashift, uint64_t physical_cols, uint64_t logical_cols,
- uint64_t nparity, uint64_t reflow_offset)
-{
- /* The zio's size in units of the vdev's minimum sector size. */
- uint64_t s = size >> ashift;
- uint64_t q, r, bc, devidx, asize = 0, tot;
-
- /*
- * "Quotient": The number of data sectors for this stripe on all but
- * the "big column" child vdevs that also contain "remainder" data.
- * AKA "full rows"
- */
- q = s / (logical_cols - nparity);
-
- /*
- * "Remainder": The number of partial stripe data sectors in this I/O.
- * This will add a sector to some, but not all, child vdevs.
- */
- r = s - q * (logical_cols - nparity);
-
- /* The number of "big columns" - those which contain remainder data. */
- bc = (r == 0 ? 0 : r + nparity);
-
- /*
- * The total number of data and parity sectors associated with
- * this I/O.
- */
- tot = s + nparity * (q + (r == 0 ? 0 : 1));
-
- /* How many rows contain data (not skip) */
- uint64_t rows = howmany(tot, logical_cols);
- int cols = MIN(tot, logical_cols);
-
- raidz_map_t *rm = kmem_zalloc(offsetof(raidz_map_t, rm_row[rows]),
- KM_SLEEP);
- rm->rm_nrows = rows;
-
- for (uint64_t row = 0; row < rows; row++) {
- raidz_row_t *rr = kmem_alloc(offsetof(raidz_row_t,
- rr_col[cols]), KM_SLEEP);
- rm->rm_row[row] = rr;
-
- /* The starting RAIDZ (parent) vdev sector of the row. */
- uint64_t b = (offset >> ashift) + row * logical_cols;
-
- /*
- * If we are in the middle of a reflow, and any part of this
- * row has not been copied, then use the old location of
- * this row.
- */
- int row_phys_cols = physical_cols;
- if (b + (logical_cols - nparity) > reflow_offset >> ashift)
- row_phys_cols--;
-
- /* starting child of this row */
- uint64_t child_id = b % row_phys_cols;
- /* The starting byte offset on each child vdev. */
- uint64_t child_offset = (b / row_phys_cols) << ashift;
-
- /*
- * We set cols to the entire width of the block, even
- * if this row is shorter. This is needed because parity
- * generation (for Q and R) needs to know the entire width,
- * because it treats the short row as though it was
- * full-width (and the "phantom" sectors were zero-filled).
- *
- * Another approach to this would be to set cols shorter
- * (to just the number of columns that we might do i/o to)
- * and have another mechanism to tell the parity generation
- * about the "entire width". Reconstruction (at least
- * vdev_raidz_reconstruct_general()) would also need to
- * know about the "entire width".
- */
- rr->rr_cols = cols;
- rr->rr_bigcols = bc;
- rr->rr_missingdata = 0;
- rr->rr_missingparity = 0;
- rr->rr_firstdatacol = nparity;
- rr->rr_abd_empty = NULL;
- rr->rr_nempty = 0;
-
- for (int c = 0; c < rr->rr_cols; c++, child_id++) {
- if (child_id >= row_phys_cols) {
- child_id -= row_phys_cols;
- child_offset += 1ULL << ashift;
- }
- rr->rr_col[c].rc_devidx = child_id;
- rr->rr_col[c].rc_offset = child_offset;
- rr->rr_col[c].rc_orig_data = NULL;
- rr->rr_col[c].rc_error = 0;
- rr->rr_col[c].rc_tried = 0;
- rr->rr_col[c].rc_skipped = 0;
- rr->rr_col[c].rc_need_orig_restore = B_FALSE;
-
- uint64_t dc = c - rr->rr_firstdatacol;
- if (c < rr->rr_firstdatacol) {
- rr->rr_col[c].rc_size = 1ULL << ashift;
- rr->rr_col[c].rc_abd =
- abd_alloc_linear(rr->rr_col[c].rc_size,
- B_TRUE);
- } else if (row == rows - 1 && bc != 0 && c >= bc) {
- /*
- * Past the end, this for parity generation.
- */
- rr->rr_col[c].rc_size = 0;
- rr->rr_col[c].rc_abd = NULL;
- } else {
- /*
- * "data column" (col excluding parity)
- * Add an ASCII art diagram here
- */
- uint64_t off;
-
- if (c < bc || r == 0) {
- off = dc * rows + row;
- } else {
- off = r * rows +
- (dc - r) * (rows - 1) + row;
- }
- rr->rr_col[c].rc_size = 1ULL << ashift;
- rr->rr_col[c].rc_abd = abd_get_offset_struct(
- &rr->rr_col[c].rc_abdstruct,
- abd, off << ashift, 1 << ashift);
- }
-
- asize += rr->rr_col[c].rc_size;
- }
- /*
- * If all data stored spans all columns, there's a danger that
- * parity will always be on the same device and, since parity
- * isn't read during normal operation, that that device's I/O
- * bandwidth won't be used effectively. We therefore switch
- * the parity every 1MB.
- *
- * ...at least that was, ostensibly, the theory. As a practical
- * matter unless we juggle the parity between all devices
- * evenly, we won't see any benefit. Further, occasional writes
- * that aren't a multiple of the LCM of the number of children
- * and the minimum stripe width are sufficient to avoid pessimal
- * behavior. Unfortunately, this decision created an implicit
- * on-disk format requirement that we need to support for all
- * eternity, but only for single-parity RAID-Z.
- *
- * If we intend to skip a sector in the zeroth column for
- * padding we must make sure to note this swap. We will never
- * intend to skip the first column since at least one data and
- * one parity column must appear in each row.
- */
- if (rr->rr_firstdatacol == 1 && rr->rr_cols > 1 &&
- (offset & (1ULL << 20))) {
- ASSERT(rr->rr_cols >= 2);
- ASSERT(rr->rr_col[0].rc_size == rr->rr_col[1].rc_size);
- devidx = rr->rr_col[0].rc_devidx;
- uint64_t o = rr->rr_col[0].rc_offset;
- rr->rr_col[0].rc_devidx = rr->rr_col[1].rc_devidx;
- rr->rr_col[0].rc_offset = rr->rr_col[1].rc_offset;
- rr->rr_col[1].rc_devidx = devidx;
- rr->rr_col[1].rc_offset = o;
- }
-
- }
- ASSERT3U(asize, ==, tot << ashift);
-
- /* init RAIDZ parity ops */
- rm->rm_ops = vdev_raidz_math_get_ops();
-
- return (rm);
-}
-
static raidz_map_t *
init_raidz_map(raidz_test_opts_t *opts, zio_t **zio, const int parity)
{
raidz_map_t *rm = NULL;
const size_t alloc_dsize = opts->rto_dsize;
const size_t total_ncols = opts->rto_dcols + parity;
const int ccols[] = { 0, 1, 2 };
VERIFY(zio);
VERIFY(parity <= 3 && parity >= 1);
*zio = umem_zalloc(sizeof (zio_t), UMEM_NOFAIL);
(*zio)->io_offset = 0;
(*zio)->io_size = alloc_dsize;
(*zio)->io_abd = raidz_alloc(alloc_dsize);
init_zio_abd(*zio);
if (opts->rto_expand) {
- rm = vdev_raidz_map_alloc_expanded((*zio)->io_abd,
- (*zio)->io_size, (*zio)->io_offset,
+ rm = vdev_raidz_map_alloc_expanded(*zio,
opts->rto_ashift, total_ncols+1, total_ncols,
- parity, opts->rto_expand_offset);
+ parity, opts->rto_expand_offset, 0, B_FALSE);
} else {
rm = vdev_raidz_map_alloc(*zio, opts->rto_ashift,
total_ncols, parity);
}
VERIFY(rm);
/* Make sure code columns are destroyed */
corrupt_colums(rm, ccols, parity);
return (rm);
}
static int
run_gen_check(raidz_test_opts_t *opts)
{
char **impl_name;
int fn, err = 0;
zio_t *zio_test;
raidz_map_t *rm_test;
err = init_raidz_golden_map(opts, PARITY_PQR);
if (0 != err)
return (err);
LOG(D_INFO, DBLSEP);
LOG(D_INFO, "Testing parity generation...\n");
for (impl_name = (char **)raidz_impl_names+1; *impl_name != NULL;
impl_name++) {
LOG(D_INFO, SEP);
LOG(D_INFO, "\tTesting [%s] implementation...", *impl_name);
if (0 != vdev_raidz_impl_set(*impl_name)) {
LOG(D_INFO, "[SKIP]\n");
continue;
} else {
LOG(D_INFO, "[SUPPORTED]\n");
}
for (fn = 0; fn < RAIDZ_GEN_NUM; fn++) {
/* Check if should stop */
if (rto_opts.rto_should_stop)
return (err);
/* create suitable raidz_map */
rm_test = init_raidz_map(opts, &zio_test, fn+1);
VERIFY(rm_test);
LOG(D_INFO, "\t\tTesting method [%s] ...",
raidz_gen_name[fn]);
if (!opts->rto_sanity)
vdev_raidz_generate_parity(rm_test);
if (cmp_code(opts, rm_test, fn+1) != 0) {
LOG(D_INFO, "[FAIL]\n");
err++;
} else
LOG(D_INFO, "[PASS]\n");
fini_raidz_map(&zio_test, &rm_test);
}
}
fini_raidz_map(&opts->zio_golden, &opts->rm_golden);
return (err);
}
static int
run_rec_check_impl(raidz_test_opts_t *opts, raidz_map_t *rm, const int fn)
{
int x0, x1, x2;
int tgtidx[3];
int err = 0;
static const int rec_tgts[7][3] = {
{1, 2, 3}, /* rec_p: bad QR & D[0] */
{0, 2, 3}, /* rec_q: bad PR & D[0] */
{0, 1, 3}, /* rec_r: bad PQ & D[0] */
{2, 3, 4}, /* rec_pq: bad R & D[0][1] */
{1, 3, 4}, /* rec_pr: bad Q & D[0][1] */
{0, 3, 4}, /* rec_qr: bad P & D[0][1] */
{3, 4, 5} /* rec_pqr: bad & D[0][1][2] */
};
memcpy(tgtidx, rec_tgts[fn], sizeof (tgtidx));
if (fn < RAIDZ_REC_PQ) {
/* can reconstruct 1 failed data disk */
for (x0 = 0; x0 < opts->rto_dcols; x0++) {
if (x0 >= rm->rm_row[0]->rr_cols - raidz_parity(rm))
continue;
/* Check if should stop */
if (rto_opts.rto_should_stop)
return (err);
LOG(D_DEBUG, "[%d] ", x0);
tgtidx[2] = x0 + raidz_parity(rm);
corrupt_colums(rm, tgtidx+2, 1);
if (!opts->rto_sanity)
vdev_raidz_reconstruct(rm, tgtidx, 3);
if (cmp_data(opts, rm) != 0) {
err++;
LOG(D_DEBUG, "\nREC D[%d]... [FAIL]\n", x0);
}
}
} else if (fn < RAIDZ_REC_PQR) {
/* can reconstruct 2 failed data disk */
for (x0 = 0; x0 < opts->rto_dcols; x0++) {
if (x0 >= rm->rm_row[0]->rr_cols - raidz_parity(rm))
continue;
for (x1 = x0 + 1; x1 < opts->rto_dcols; x1++) {
if (x1 >= rm->rm_row[0]->rr_cols -
raidz_parity(rm))
continue;
/* Check if should stop */
if (rto_opts.rto_should_stop)
return (err);
LOG(D_DEBUG, "[%d %d] ", x0, x1);
tgtidx[1] = x0 + raidz_parity(rm);
tgtidx[2] = x1 + raidz_parity(rm);
corrupt_colums(rm, tgtidx+1, 2);
if (!opts->rto_sanity)
vdev_raidz_reconstruct(rm, tgtidx, 3);
if (cmp_data(opts, rm) != 0) {
err++;
LOG(D_DEBUG, "\nREC D[%d %d]... "
"[FAIL]\n", x0, x1);
}
}
}
} else {
/* can reconstruct 3 failed data disk */
for (x0 = 0; x0 < opts->rto_dcols; x0++) {
if (x0 >= rm->rm_row[0]->rr_cols - raidz_parity(rm))
continue;
for (x1 = x0 + 1; x1 < opts->rto_dcols; x1++) {
if (x1 >= rm->rm_row[0]->rr_cols -
raidz_parity(rm))
continue;
for (x2 = x1 + 1; x2 < opts->rto_dcols; x2++) {
if (x2 >= rm->rm_row[0]->rr_cols -
raidz_parity(rm))
continue;
/* Check if should stop */
if (rto_opts.rto_should_stop)
return (err);
LOG(D_DEBUG, "[%d %d %d]", x0, x1, x2);
tgtidx[0] = x0 + raidz_parity(rm);
tgtidx[1] = x1 + raidz_parity(rm);
tgtidx[2] = x2 + raidz_parity(rm);
corrupt_colums(rm, tgtidx, 3);
if (!opts->rto_sanity)
vdev_raidz_reconstruct(rm,
tgtidx, 3);
if (cmp_data(opts, rm) != 0) {
err++;
LOG(D_DEBUG,
"\nREC D[%d %d %d]... "
"[FAIL]\n", x0, x1, x2);
}
}
}
}
}
return (err);
}
static int
run_rec_check(raidz_test_opts_t *opts)
{
char **impl_name;
unsigned fn, err = 0;
zio_t *zio_test;
raidz_map_t *rm_test;
err = init_raidz_golden_map(opts, PARITY_PQR);
if (0 != err)
return (err);
LOG(D_INFO, DBLSEP);
LOG(D_INFO, "Testing data reconstruction...\n");
for (impl_name = (char **)raidz_impl_names+1; *impl_name != NULL;
impl_name++) {
LOG(D_INFO, SEP);
LOG(D_INFO, "\tTesting [%s] implementation...", *impl_name);
if (vdev_raidz_impl_set(*impl_name) != 0) {
LOG(D_INFO, "[SKIP]\n");
continue;
} else
LOG(D_INFO, "[SUPPORTED]\n");
/* create suitable raidz_map */
rm_test = init_raidz_map(opts, &zio_test, PARITY_PQR);
/* generate parity */
vdev_raidz_generate_parity(rm_test);
for (fn = 0; fn < RAIDZ_REC_NUM; fn++) {
LOG(D_INFO, "\t\tTesting method [%s] ...",
raidz_rec_name[fn]);
if (run_rec_check_impl(opts, rm_test, fn) != 0) {
LOG(D_INFO, "[FAIL]\n");
err++;
} else
LOG(D_INFO, "[PASS]\n");
}
/* tear down test raidz_map */
fini_raidz_map(&zio_test, &rm_test);
}
fini_raidz_map(&opts->zio_golden, &opts->rm_golden);
return (err);
}
static int
run_test(raidz_test_opts_t *opts)
{
int err = 0;
if (opts == NULL)
opts = &rto_opts;
print_opts(opts, B_FALSE);
err |= run_gen_check(opts);
err |= run_rec_check(opts);
return (err);
}
#define SWEEP_RUNNING 0
#define SWEEP_FINISHED 1
#define SWEEP_ERROR 2
#define SWEEP_TIMEOUT 3
static int sweep_state = 0;
static raidz_test_opts_t failed_opts;
static kmutex_t sem_mtx;
static kcondvar_t sem_cv;
static int max_free_slots;
static int free_slots;
static __attribute__((noreturn)) void
sweep_thread(void *arg)
{
int err = 0;
raidz_test_opts_t *opts = (raidz_test_opts_t *)arg;
VERIFY(opts != NULL);
err = run_test(opts);
if (rto_opts.rto_sanity) {
/* 25% chance that a sweep test fails */
if (rand() < (RAND_MAX/4))
err = 1;
}
if (0 != err) {
mutex_enter(&sem_mtx);
memcpy(&failed_opts, opts, sizeof (raidz_test_opts_t));
sweep_state = SWEEP_ERROR;
mutex_exit(&sem_mtx);
}
umem_free(opts, sizeof (raidz_test_opts_t));
/* signal the next thread */
mutex_enter(&sem_mtx);
free_slots++;
cv_signal(&sem_cv);
mutex_exit(&sem_mtx);
thread_exit();
}
static int
run_sweep(void)
{
static const size_t dcols_v[] = { 1, 2, 3, 4, 5, 6, 7, 8, 12, 15, 16 };
static const size_t ashift_v[] = { 9, 12, 14 };
static const size_t size_v[] = { 1 << 9, 21 * (1 << 9), 13 * (1 << 12),
1 << 17, (1 << 20) - (1 << 12), SPA_MAXBLOCKSIZE };
(void) setvbuf(stdout, NULL, _IONBF, 0);
ulong_t total_comb = ARRAY_SIZE(size_v) * ARRAY_SIZE(ashift_v) *
ARRAY_SIZE(dcols_v);
ulong_t tried_comb = 0;
hrtime_t time_diff, start_time = gethrtime();
raidz_test_opts_t *opts;
int a, d, s;
max_free_slots = free_slots = MAX(2, boot_ncpus);
mutex_init(&sem_mtx, NULL, MUTEX_DEFAULT, NULL);
cv_init(&sem_cv, NULL, CV_DEFAULT, NULL);
for (s = 0; s < ARRAY_SIZE(size_v); s++)
for (a = 0; a < ARRAY_SIZE(ashift_v); a++)
for (d = 0; d < ARRAY_SIZE(dcols_v); d++) {
if (size_v[s] < (1 << ashift_v[a])) {
total_comb--;
continue;
}
if (++tried_comb % 20 == 0)
LOG(D_ALL, "%lu/%lu... ", tried_comb, total_comb);
/* wait for signal to start new thread */
mutex_enter(&sem_mtx);
while (cv_timedwait_sig(&sem_cv, &sem_mtx,
ddi_get_lbolt() + hz)) {
/* check if should stop the test (timeout) */
time_diff = (gethrtime() - start_time) / NANOSEC;
if (rto_opts.rto_sweep_timeout > 0 &&
time_diff >= rto_opts.rto_sweep_timeout) {
sweep_state = SWEEP_TIMEOUT;
rto_opts.rto_should_stop = B_TRUE;
mutex_exit(&sem_mtx);
goto exit;
}
/* check if should stop the test (error) */
if (sweep_state != SWEEP_RUNNING) {
mutex_exit(&sem_mtx);
goto exit;
}
/* exit loop if a slot is available */
if (free_slots > 0) {
break;
}
}
free_slots--;
mutex_exit(&sem_mtx);
opts = umem_zalloc(sizeof (raidz_test_opts_t), UMEM_NOFAIL);
opts->rto_ashift = ashift_v[a];
opts->rto_dcols = dcols_v[d];
opts->rto_offset = (1ULL << ashift_v[a]) * rand();
opts->rto_dsize = size_v[s];
opts->rto_expand = rto_opts.rto_expand;
opts->rto_expand_offset = rto_opts.rto_expand_offset;
opts->rto_v = 0; /* be quiet */
VERIFY3P(thread_create(NULL, 0, sweep_thread, (void *) opts,
0, NULL, TS_RUN, defclsyspri), !=, NULL);
}
exit:
LOG(D_ALL, "\nWaiting for test threads to finish...\n");
mutex_enter(&sem_mtx);
VERIFY(free_slots <= max_free_slots);
while (free_slots < max_free_slots) {
(void) cv_wait(&sem_cv, &sem_mtx);
}
mutex_exit(&sem_mtx);
if (sweep_state == SWEEP_ERROR) {
ERR("Sweep test failed! Failed option: \n");
print_opts(&failed_opts, B_TRUE);
} else {
if (sweep_state == SWEEP_TIMEOUT)
LOG(D_ALL, "Test timeout (%lus). Stopping...\n",
(ulong_t)rto_opts.rto_sweep_timeout);
LOG(D_ALL, "Sweep test succeeded on %lu raidz maps!\n",
(ulong_t)tried_comb);
}
mutex_destroy(&sem_mtx);
return (sweep_state == SWEEP_ERROR ? SWEEP_ERROR : 0);
}
int
main(int argc, char **argv)
{
size_t i;
struct sigaction action;
int err = 0;
/* init gdb pid string early */
(void) sprintf(pid_s, "%d", getpid());
action.sa_handler = sig_handler;
sigemptyset(&action.sa_mask);
action.sa_flags = 0;
if (sigaction(SIGSEGV, &action, NULL) < 0) {
ERR("raidz_test: cannot catch SIGSEGV: %s.\n", strerror(errno));
exit(EXIT_FAILURE);
}
(void) setvbuf(stdout, NULL, _IOLBF, 0);
dprintf_setup(&argc, argv);
process_options(argc, argv);
kernel_init(SPA_MODE_READ);
/* setup random data because rand() is not reentrant */
rand_data = (int *)umem_alloc(SPA_MAXBLOCKSIZE, UMEM_NOFAIL);
srand((unsigned)time(NULL) * getpid());
for (i = 0; i < SPA_MAXBLOCKSIZE / sizeof (int); i++)
rand_data[i] = rand();
mprotect(rand_data, SPA_MAXBLOCKSIZE, PROT_READ);
if (rto_opts.rto_benchmark) {
run_raidz_benchmark();
} else if (rto_opts.rto_sweep) {
err = run_sweep();
} else {
err = run_test(NULL);
}
umem_free(rand_data, SPA_MAXBLOCKSIZE);
kernel_fini();
return (err);
}
diff --git a/sys/contrib/openzfs/cmd/raidz_test/raidz_test.h b/sys/contrib/openzfs/cmd/raidz_test/raidz_test.h
index 163929defc73..f912e281f6f3 100644
--- a/sys/contrib/openzfs/cmd/raidz_test/raidz_test.h
+++ b/sys/contrib/openzfs/cmd/raidz_test/raidz_test.h
@@ -1,125 +1,122 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (C) 2016 Gvozden Nešković. All rights reserved.
*/
#ifndef RAIDZ_TEST_H
#define RAIDZ_TEST_H
#include <sys/spa.h>
static const char *const raidz_impl_names[] = {
"original",
"scalar",
"sse2",
"ssse3",
"avx2",
"avx512f",
"avx512bw",
"aarch64_neon",
"aarch64_neonx2",
"powerpc_altivec",
NULL
};
enum raidz_verbosity {
D_ALL,
D_INFO,
D_DEBUG,
};
typedef struct raidz_test_opts {
size_t rto_ashift;
uint64_t rto_offset;
size_t rto_dcols;
size_t rto_dsize;
enum raidz_verbosity rto_v;
size_t rto_sweep;
size_t rto_sweep_timeout;
size_t rto_benchmark;
size_t rto_expand;
uint64_t rto_expand_offset;
size_t rto_sanity;
size_t rto_gdb;
/* non-user options */
boolean_t rto_should_stop;
zio_t *zio_golden;
raidz_map_t *rm_golden;
} raidz_test_opts_t;
static const raidz_test_opts_t rto_opts_defaults = {
.rto_ashift = 9,
.rto_offset = 1ULL << 0,
.rto_dcols = 8,
.rto_dsize = 1<<19,
.rto_v = D_ALL,
.rto_sweep = 0,
.rto_benchmark = 0,
.rto_expand = 0,
.rto_expand_offset = -1ULL,
.rto_sanity = 0,
.rto_gdb = 0,
.rto_should_stop = B_FALSE
};
extern raidz_test_opts_t rto_opts;
static inline size_t ilog2(size_t a)
{
return (a > 1 ? 1 + ilog2(a >> 1) : 0);
}
#define LOG(lvl, ...) \
{ \
if (rto_opts.rto_v >= lvl) \
(void) fprintf(stdout, __VA_ARGS__); \
} \
#define LOG_OPT(lvl, opt, ...) \
{ \
if (opt->rto_v >= lvl) \
(void) fprintf(stdout, __VA_ARGS__); \
} \
#define ERR(...) (void) fprintf(stderr, __VA_ARGS__)
#define DBLSEP "================\n"
#define SEP "----------------\n"
#define raidz_alloc(size) abd_alloc(size, B_FALSE)
#define raidz_free(p, size) abd_free(p)
void init_zio_abd(zio_t *zio);
void run_raidz_benchmark(void);
-struct raidz_map *vdev_raidz_map_alloc_expanded(abd_t *, uint64_t, uint64_t,
- uint64_t, uint64_t, uint64_t, uint64_t, uint64_t);
-
#endif /* RAIDZ_TEST_H */
diff --git a/sys/contrib/openzfs/cmd/zdb/zdb.c b/sys/contrib/openzfs/cmd/zdb/zdb.c
index 3c282f3fc975..18221c4b92d2 100644
--- a/sys/contrib/openzfs/cmd/zdb/zdb.c
+++ b/sys/contrib/openzfs/cmd/zdb/zdb.c
@@ -1,9423 +1,9428 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2019 by Delphix. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright 2016 Nexenta Systems, Inc.
* Copyright (c) 2017, 2018 Lawrence Livermore National Security, LLC.
* Copyright (c) 2015, 2017, Intel Corporation.
* Copyright (c) 2020 Datto Inc.
* Copyright (c) 2020, The FreeBSD Foundation [1]
*
* [1] Portions of this software were developed by Allan Jude
* under sponsorship from the FreeBSD Foundation.
* Copyright (c) 2021 Allan Jude
* Copyright (c) 2021 Toomas Soome <tsoome@me.com>
* Copyright (c) 2023, Klara Inc.
*/
#include <stdio.h>
#include <unistd.h>
#include <stdlib.h>
#include <ctype.h>
#include <getopt.h>
#include <openssl/evp.h>
#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/dmu.h>
#include <sys/zap.h>
#include <sys/fs/zfs.h>
#include <sys/zfs_znode.h>
#include <sys/zfs_sa.h>
#include <sys/sa.h>
#include <sys/sa_impl.h>
#include <sys/vdev.h>
#include <sys/vdev_impl.h>
#include <sys/metaslab_impl.h>
#include <sys/dmu_objset.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_pool.h>
#include <sys/dsl_bookmark.h>
#include <sys/dbuf.h>
#include <sys/zil.h>
#include <sys/zil_impl.h>
#include <sys/stat.h>
#include <sys/resource.h>
#include <sys/dmu_send.h>
#include <sys/dmu_traverse.h>
#include <sys/zio_checksum.h>
#include <sys/zio_compress.h>
#include <sys/zfs_fuid.h>
#include <sys/arc.h>
#include <sys/arc_impl.h>
#include <sys/ddt.h>
#include <sys/zfeature.h>
#include <sys/abd.h>
#include <sys/blkptr.h>
#include <sys/dsl_crypt.h>
#include <sys/dsl_scan.h>
#include <sys/btree.h>
#include <sys/brt.h>
#include <zfs_comutil.h>
#include <sys/zstd/zstd.h>
#include <libnvpair.h>
#include <libzutil.h>
#include "zdb.h"
#define ZDB_COMPRESS_NAME(idx) ((idx) < ZIO_COMPRESS_FUNCTIONS ? \
zio_compress_table[(idx)].ci_name : "UNKNOWN")
#define ZDB_CHECKSUM_NAME(idx) ((idx) < ZIO_CHECKSUM_FUNCTIONS ? \
zio_checksum_table[(idx)].ci_name : "UNKNOWN")
#define ZDB_OT_TYPE(idx) ((idx) < DMU_OT_NUMTYPES ? (idx) : \
(idx) == DMU_OTN_ZAP_DATA || (idx) == DMU_OTN_ZAP_METADATA ? \
DMU_OT_ZAP_OTHER : \
(idx) == DMU_OTN_UINT64_DATA || (idx) == DMU_OTN_UINT64_METADATA ? \
DMU_OT_UINT64_OTHER : DMU_OT_NUMTYPES)
/* Some platforms require part of inode IDs to be remapped */
#ifdef __APPLE__
#define ZDB_MAP_OBJECT_ID(obj) INO_XNUTOZFS(obj, 2)
#else
#define ZDB_MAP_OBJECT_ID(obj) (obj)
#endif
static const char *
zdb_ot_name(dmu_object_type_t type)
{
if (type < DMU_OT_NUMTYPES)
return (dmu_ot[type].ot_name);
else if ((type & DMU_OT_NEWTYPE) &&
((type & DMU_OT_BYTESWAP_MASK) < DMU_BSWAP_NUMFUNCS))
return (dmu_ot_byteswap[type & DMU_OT_BYTESWAP_MASK].ob_name);
else
return ("UNKNOWN");
}
extern int reference_tracking_enable;
extern int zfs_recover;
extern uint_t zfs_vdev_async_read_max_active;
extern boolean_t spa_load_verify_dryrun;
extern boolean_t spa_mode_readable_spacemaps;
extern uint_t zfs_reconstruct_indirect_combinations_max;
extern uint_t zfs_btree_verify_intensity;
static const char cmdname[] = "zdb";
uint8_t dump_opt[256];
typedef void object_viewer_t(objset_t *, uint64_t, void *data, size_t size);
static uint64_t *zopt_metaslab = NULL;
static unsigned zopt_metaslab_args = 0;
typedef struct zopt_object_range {
uint64_t zor_obj_start;
uint64_t zor_obj_end;
uint64_t zor_flags;
} zopt_object_range_t;
static zopt_object_range_t *zopt_object_ranges = NULL;
static unsigned zopt_object_args = 0;
static int flagbits[256];
#define ZOR_FLAG_PLAIN_FILE 0x0001
#define ZOR_FLAG_DIRECTORY 0x0002
#define ZOR_FLAG_SPACE_MAP 0x0004
#define ZOR_FLAG_ZAP 0x0008
#define ZOR_FLAG_ALL_TYPES -1
#define ZOR_SUPPORTED_FLAGS (ZOR_FLAG_PLAIN_FILE | \
ZOR_FLAG_DIRECTORY | \
ZOR_FLAG_SPACE_MAP | \
ZOR_FLAG_ZAP)
#define ZDB_FLAG_CHECKSUM 0x0001
#define ZDB_FLAG_DECOMPRESS 0x0002
#define ZDB_FLAG_BSWAP 0x0004
#define ZDB_FLAG_GBH 0x0008
#define ZDB_FLAG_INDIRECT 0x0010
#define ZDB_FLAG_RAW 0x0020
#define ZDB_FLAG_PRINT_BLKPTR 0x0040
#define ZDB_FLAG_VERBOSE 0x0080
static uint64_t max_inflight_bytes = 256 * 1024 * 1024; /* 256MB */
static int leaked_objects = 0;
static range_tree_t *mos_refd_objs;
static void snprintf_blkptr_compact(char *, size_t, const blkptr_t *,
boolean_t);
static void mos_obj_refd(uint64_t);
static void mos_obj_refd_multiple(uint64_t);
static int dump_bpobj_cb(void *arg, const blkptr_t *bp, boolean_t free,
dmu_tx_t *tx);
typedef struct sublivelist_verify {
/* FREE's that haven't yet matched to an ALLOC, in one sub-livelist */
zfs_btree_t sv_pair;
/* ALLOC's without a matching FREE, accumulates across sub-livelists */
zfs_btree_t sv_leftover;
} sublivelist_verify_t;
static int
livelist_compare(const void *larg, const void *rarg)
{
const blkptr_t *l = larg;
const blkptr_t *r = rarg;
/* Sort them according to dva[0] */
uint64_t l_dva0_vdev, r_dva0_vdev;
l_dva0_vdev = DVA_GET_VDEV(&l->blk_dva[0]);
r_dva0_vdev = DVA_GET_VDEV(&r->blk_dva[0]);
if (l_dva0_vdev < r_dva0_vdev)
return (-1);
else if (l_dva0_vdev > r_dva0_vdev)
return (+1);
/* if vdevs are equal, sort by offsets. */
uint64_t l_dva0_offset;
uint64_t r_dva0_offset;
l_dva0_offset = DVA_GET_OFFSET(&l->blk_dva[0]);
r_dva0_offset = DVA_GET_OFFSET(&r->blk_dva[0]);
if (l_dva0_offset < r_dva0_offset) {
return (-1);
} else if (l_dva0_offset > r_dva0_offset) {
return (+1);
}
/*
* Since we're storing blkptrs without cancelling FREE/ALLOC pairs,
* it's possible the offsets are equal. In that case, sort by txg
*/
if (l->blk_birth < r->blk_birth) {
return (-1);
} else if (l->blk_birth > r->blk_birth) {
return (+1);
}
return (0);
}
typedef struct sublivelist_verify_block {
dva_t svb_dva;
/*
* We need this to check if the block marked as allocated
* in the livelist was freed (and potentially reallocated)
* in the metaslab spacemaps at a later TXG.
*/
uint64_t svb_allocated_txg;
} sublivelist_verify_block_t;
static void zdb_print_blkptr(const blkptr_t *bp, int flags);
typedef struct sublivelist_verify_block_refcnt {
/* block pointer entry in livelist being verified */
blkptr_t svbr_blk;
/*
* Refcount gets incremented to 1 when we encounter the first
* FREE entry for the svfbr block pointer and a node for it
* is created in our ZDB verification/tracking metadata.
*
* As we encounter more FREE entries we increment this counter
* and similarly decrement it whenever we find the respective
* ALLOC entries for this block.
*
* When the refcount gets to 0 it means that all the FREE and
* ALLOC entries of this block have paired up and we no longer
* need to track it in our verification logic (e.g. the node
* containing this struct in our verification data structure
* should be freed).
*
* [refer to sublivelist_verify_blkptr() for the actual code]
*/
uint32_t svbr_refcnt;
} sublivelist_verify_block_refcnt_t;
static int
sublivelist_block_refcnt_compare(const void *larg, const void *rarg)
{
const sublivelist_verify_block_refcnt_t *l = larg;
const sublivelist_verify_block_refcnt_t *r = rarg;
return (livelist_compare(&l->svbr_blk, &r->svbr_blk));
}
static int
sublivelist_verify_blkptr(void *arg, const blkptr_t *bp, boolean_t free,
dmu_tx_t *tx)
{
ASSERT3P(tx, ==, NULL);
struct sublivelist_verify *sv = arg;
sublivelist_verify_block_refcnt_t current = {
.svbr_blk = *bp,
/*
* Start with 1 in case this is the first free entry.
* This field is not used for our B-Tree comparisons
* anyway.
*/
.svbr_refcnt = 1,
};
zfs_btree_index_t where;
sublivelist_verify_block_refcnt_t *pair =
zfs_btree_find(&sv->sv_pair, &current, &where);
if (free) {
if (pair == NULL) {
/* first free entry for this block pointer */
zfs_btree_add(&sv->sv_pair, &current);
} else {
pair->svbr_refcnt++;
}
} else {
if (pair == NULL) {
/* block that is currently marked as allocated */
for (int i = 0; i < SPA_DVAS_PER_BP; i++) {
if (DVA_IS_EMPTY(&bp->blk_dva[i]))
break;
sublivelist_verify_block_t svb = {
.svb_dva = bp->blk_dva[i],
.svb_allocated_txg = bp->blk_birth
};
if (zfs_btree_find(&sv->sv_leftover, &svb,
&where) == NULL) {
zfs_btree_add_idx(&sv->sv_leftover,
&svb, &where);
}
}
} else {
/* alloc matches a free entry */
pair->svbr_refcnt--;
if (pair->svbr_refcnt == 0) {
/* all allocs and frees have been matched */
zfs_btree_remove_idx(&sv->sv_pair, &where);
}
}
}
return (0);
}
static int
sublivelist_verify_func(void *args, dsl_deadlist_entry_t *dle)
{
int err;
struct sublivelist_verify *sv = args;
zfs_btree_create(&sv->sv_pair, sublivelist_block_refcnt_compare, NULL,
sizeof (sublivelist_verify_block_refcnt_t));
err = bpobj_iterate_nofree(&dle->dle_bpobj, sublivelist_verify_blkptr,
sv, NULL);
sublivelist_verify_block_refcnt_t *e;
zfs_btree_index_t *cookie = NULL;
while ((e = zfs_btree_destroy_nodes(&sv->sv_pair, &cookie)) != NULL) {
char blkbuf[BP_SPRINTF_LEN];
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf),
&e->svbr_blk, B_TRUE);
(void) printf("\tERROR: %d unmatched FREE(s): %s\n",
e->svbr_refcnt, blkbuf);
}
zfs_btree_destroy(&sv->sv_pair);
return (err);
}
static int
livelist_block_compare(const void *larg, const void *rarg)
{
const sublivelist_verify_block_t *l = larg;
const sublivelist_verify_block_t *r = rarg;
if (DVA_GET_VDEV(&l->svb_dva) < DVA_GET_VDEV(&r->svb_dva))
return (-1);
else if (DVA_GET_VDEV(&l->svb_dva) > DVA_GET_VDEV(&r->svb_dva))
return (+1);
if (DVA_GET_OFFSET(&l->svb_dva) < DVA_GET_OFFSET(&r->svb_dva))
return (-1);
else if (DVA_GET_OFFSET(&l->svb_dva) > DVA_GET_OFFSET(&r->svb_dva))
return (+1);
if (DVA_GET_ASIZE(&l->svb_dva) < DVA_GET_ASIZE(&r->svb_dva))
return (-1);
else if (DVA_GET_ASIZE(&l->svb_dva) > DVA_GET_ASIZE(&r->svb_dva))
return (+1);
return (0);
}
/*
* Check for errors in a livelist while tracking all unfreed ALLOCs in the
* sublivelist_verify_t: sv->sv_leftover
*/
static void
livelist_verify(dsl_deadlist_t *dl, void *arg)
{
sublivelist_verify_t *sv = arg;
dsl_deadlist_iterate(dl, sublivelist_verify_func, sv);
}
/*
* Check for errors in the livelist entry and discard the intermediary
* data structures
*/
static int
sublivelist_verify_lightweight(void *args, dsl_deadlist_entry_t *dle)
{
(void) args;
sublivelist_verify_t sv;
zfs_btree_create(&sv.sv_leftover, livelist_block_compare, NULL,
sizeof (sublivelist_verify_block_t));
int err = sublivelist_verify_func(&sv, dle);
zfs_btree_clear(&sv.sv_leftover);
zfs_btree_destroy(&sv.sv_leftover);
return (err);
}
typedef struct metaslab_verify {
/*
* Tree containing all the leftover ALLOCs from the livelists
* that are part of this metaslab.
*/
zfs_btree_t mv_livelist_allocs;
/*
* Metaslab information.
*/
uint64_t mv_vdid;
uint64_t mv_msid;
uint64_t mv_start;
uint64_t mv_end;
/*
* What's currently allocated for this metaslab.
*/
range_tree_t *mv_allocated;
} metaslab_verify_t;
typedef void ll_iter_t(dsl_deadlist_t *ll, void *arg);
typedef int (*zdb_log_sm_cb_t)(spa_t *spa, space_map_entry_t *sme, uint64_t txg,
void *arg);
typedef struct unflushed_iter_cb_arg {
spa_t *uic_spa;
uint64_t uic_txg;
void *uic_arg;
zdb_log_sm_cb_t uic_cb;
} unflushed_iter_cb_arg_t;
static int
iterate_through_spacemap_logs_cb(space_map_entry_t *sme, void *arg)
{
unflushed_iter_cb_arg_t *uic = arg;
return (uic->uic_cb(uic->uic_spa, sme, uic->uic_txg, uic->uic_arg));
}
static void
iterate_through_spacemap_logs(spa_t *spa, zdb_log_sm_cb_t cb, void *arg)
{
if (!spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP))
return;
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
for (spa_log_sm_t *sls = avl_first(&spa->spa_sm_logs_by_txg);
sls; sls = AVL_NEXT(&spa->spa_sm_logs_by_txg, sls)) {
space_map_t *sm = NULL;
VERIFY0(space_map_open(&sm, spa_meta_objset(spa),
sls->sls_sm_obj, 0, UINT64_MAX, SPA_MINBLOCKSHIFT));
unflushed_iter_cb_arg_t uic = {
.uic_spa = spa,
.uic_txg = sls->sls_txg,
.uic_arg = arg,
.uic_cb = cb
};
VERIFY0(space_map_iterate(sm, space_map_length(sm),
iterate_through_spacemap_logs_cb, &uic));
space_map_close(sm);
}
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
static void
verify_livelist_allocs(metaslab_verify_t *mv, uint64_t txg,
uint64_t offset, uint64_t size)
{
sublivelist_verify_block_t svb = {{{0}}};
DVA_SET_VDEV(&svb.svb_dva, mv->mv_vdid);
DVA_SET_OFFSET(&svb.svb_dva, offset);
DVA_SET_ASIZE(&svb.svb_dva, size);
zfs_btree_index_t where;
uint64_t end_offset = offset + size;
/*
* Look for an exact match for spacemap entry in the livelist entries.
* Then, look for other livelist entries that fall within the range
* of the spacemap entry as it may have been condensed
*/
sublivelist_verify_block_t *found =
zfs_btree_find(&mv->mv_livelist_allocs, &svb, &where);
if (found == NULL) {
found = zfs_btree_next(&mv->mv_livelist_allocs, &where, &where);
}
for (; found != NULL && DVA_GET_VDEV(&found->svb_dva) == mv->mv_vdid &&
DVA_GET_OFFSET(&found->svb_dva) < end_offset;
found = zfs_btree_next(&mv->mv_livelist_allocs, &where, &where)) {
if (found->svb_allocated_txg <= txg) {
(void) printf("ERROR: Livelist ALLOC [%llx:%llx] "
"from TXG %llx FREED at TXG %llx\n",
(u_longlong_t)DVA_GET_OFFSET(&found->svb_dva),
(u_longlong_t)DVA_GET_ASIZE(&found->svb_dva),
(u_longlong_t)found->svb_allocated_txg,
(u_longlong_t)txg);
}
}
}
static int
metaslab_spacemap_validation_cb(space_map_entry_t *sme, void *arg)
{
metaslab_verify_t *mv = arg;
uint64_t offset = sme->sme_offset;
uint64_t size = sme->sme_run;
uint64_t txg = sme->sme_txg;
if (sme->sme_type == SM_ALLOC) {
if (range_tree_contains(mv->mv_allocated,
offset, size)) {
(void) printf("ERROR: DOUBLE ALLOC: "
"%llu [%llx:%llx] "
"%llu:%llu LOG_SM\n",
(u_longlong_t)txg, (u_longlong_t)offset,
(u_longlong_t)size, (u_longlong_t)mv->mv_vdid,
(u_longlong_t)mv->mv_msid);
} else {
range_tree_add(mv->mv_allocated,
offset, size);
}
} else {
if (!range_tree_contains(mv->mv_allocated,
offset, size)) {
(void) printf("ERROR: DOUBLE FREE: "
"%llu [%llx:%llx] "
"%llu:%llu LOG_SM\n",
(u_longlong_t)txg, (u_longlong_t)offset,
(u_longlong_t)size, (u_longlong_t)mv->mv_vdid,
(u_longlong_t)mv->mv_msid);
} else {
range_tree_remove(mv->mv_allocated,
offset, size);
}
}
if (sme->sme_type != SM_ALLOC) {
/*
* If something is freed in the spacemap, verify that
* it is not listed as allocated in the livelist.
*/
verify_livelist_allocs(mv, txg, offset, size);
}
return (0);
}
static int
spacemap_check_sm_log_cb(spa_t *spa, space_map_entry_t *sme,
uint64_t txg, void *arg)
{
metaslab_verify_t *mv = arg;
uint64_t offset = sme->sme_offset;
uint64_t vdev_id = sme->sme_vdev;
vdev_t *vd = vdev_lookup_top(spa, vdev_id);
/* skip indirect vdevs */
if (!vdev_is_concrete(vd))
return (0);
if (vdev_id != mv->mv_vdid)
return (0);
metaslab_t *ms = vd->vdev_ms[offset >> vd->vdev_ms_shift];
if (ms->ms_id != mv->mv_msid)
return (0);
if (txg < metaslab_unflushed_txg(ms))
return (0);
ASSERT3U(txg, ==, sme->sme_txg);
return (metaslab_spacemap_validation_cb(sme, mv));
}
static void
spacemap_check_sm_log(spa_t *spa, metaslab_verify_t *mv)
{
iterate_through_spacemap_logs(spa, spacemap_check_sm_log_cb, mv);
}
static void
spacemap_check_ms_sm(space_map_t *sm, metaslab_verify_t *mv)
{
if (sm == NULL)
return;
VERIFY0(space_map_iterate(sm, space_map_length(sm),
metaslab_spacemap_validation_cb, mv));
}
static void iterate_deleted_livelists(spa_t *spa, ll_iter_t func, void *arg);
/*
* Transfer blocks from sv_leftover tree to the mv_livelist_allocs if
* they are part of that metaslab (mv_msid).
*/
static void
mv_populate_livelist_allocs(metaslab_verify_t *mv, sublivelist_verify_t *sv)
{
zfs_btree_index_t where;
sublivelist_verify_block_t *svb;
ASSERT3U(zfs_btree_numnodes(&mv->mv_livelist_allocs), ==, 0);
for (svb = zfs_btree_first(&sv->sv_leftover, &where);
svb != NULL;
svb = zfs_btree_next(&sv->sv_leftover, &where, &where)) {
if (DVA_GET_VDEV(&svb->svb_dva) != mv->mv_vdid)
continue;
if (DVA_GET_OFFSET(&svb->svb_dva) < mv->mv_start &&
(DVA_GET_OFFSET(&svb->svb_dva) +
DVA_GET_ASIZE(&svb->svb_dva)) > mv->mv_start) {
(void) printf("ERROR: Found block that crosses "
"metaslab boundary: <%llu:%llx:%llx>\n",
(u_longlong_t)DVA_GET_VDEV(&svb->svb_dva),
(u_longlong_t)DVA_GET_OFFSET(&svb->svb_dva),
(u_longlong_t)DVA_GET_ASIZE(&svb->svb_dva));
continue;
}
if (DVA_GET_OFFSET(&svb->svb_dva) < mv->mv_start)
continue;
if (DVA_GET_OFFSET(&svb->svb_dva) >= mv->mv_end)
continue;
if ((DVA_GET_OFFSET(&svb->svb_dva) +
DVA_GET_ASIZE(&svb->svb_dva)) > mv->mv_end) {
(void) printf("ERROR: Found block that crosses "
"metaslab boundary: <%llu:%llx:%llx>\n",
(u_longlong_t)DVA_GET_VDEV(&svb->svb_dva),
(u_longlong_t)DVA_GET_OFFSET(&svb->svb_dva),
(u_longlong_t)DVA_GET_ASIZE(&svb->svb_dva));
continue;
}
zfs_btree_add(&mv->mv_livelist_allocs, svb);
}
for (svb = zfs_btree_first(&mv->mv_livelist_allocs, &where);
svb != NULL;
svb = zfs_btree_next(&mv->mv_livelist_allocs, &where, &where)) {
zfs_btree_remove(&sv->sv_leftover, svb);
}
}
/*
* [Livelist Check]
* Iterate through all the sublivelists and:
* - report leftover frees (**)
* - record leftover ALLOCs together with their TXG [see Cross Check]
*
* (**) Note: Double ALLOCs are valid in datasets that have dedup
* enabled. Similarly double FREEs are allowed as well but
* only if they pair up with a corresponding ALLOC entry once
* we our done with our sublivelist iteration.
*
* [Spacemap Check]
* for each metaslab:
* - iterate over spacemap and then the metaslab's entries in the
* spacemap log, then report any double FREEs and ALLOCs (do not
* blow up).
*
* [Cross Check]
* After finishing the Livelist Check phase and while being in the
* Spacemap Check phase, we find all the recorded leftover ALLOCs
* of the livelist check that are part of the metaslab that we are
* currently looking at in the Spacemap Check. We report any entries
* that are marked as ALLOCs in the livelists but have been actually
* freed (and potentially allocated again) after their TXG stamp in
* the spacemaps. Also report any ALLOCs from the livelists that
* belong to indirect vdevs (e.g. their vdev completed removal).
*
* Note that this will miss Log Spacemap entries that cancelled each other
* out before being flushed to the metaslab, so we are not guaranteed
* to match all erroneous ALLOCs.
*/
static void
livelist_metaslab_validate(spa_t *spa)
{
(void) printf("Verifying deleted livelist entries\n");
sublivelist_verify_t sv;
zfs_btree_create(&sv.sv_leftover, livelist_block_compare, NULL,
sizeof (sublivelist_verify_block_t));
iterate_deleted_livelists(spa, livelist_verify, &sv);
(void) printf("Verifying metaslab entries\n");
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
if (!vdev_is_concrete(vd))
continue;
for (uint64_t mid = 0; mid < vd->vdev_ms_count; mid++) {
metaslab_t *m = vd->vdev_ms[mid];
(void) fprintf(stderr,
"\rverifying concrete vdev %llu, "
"metaslab %llu of %llu ...",
(longlong_t)vd->vdev_id,
(longlong_t)mid,
(longlong_t)vd->vdev_ms_count);
uint64_t shift, start;
range_seg_type_t type =
metaslab_calculate_range_tree_type(vd, m,
&start, &shift);
metaslab_verify_t mv;
mv.mv_allocated = range_tree_create(NULL,
type, NULL, start, shift);
mv.mv_vdid = vd->vdev_id;
mv.mv_msid = m->ms_id;
mv.mv_start = m->ms_start;
mv.mv_end = m->ms_start + m->ms_size;
zfs_btree_create(&mv.mv_livelist_allocs,
livelist_block_compare, NULL,
sizeof (sublivelist_verify_block_t));
mv_populate_livelist_allocs(&mv, &sv);
spacemap_check_ms_sm(m->ms_sm, &mv);
spacemap_check_sm_log(spa, &mv);
range_tree_vacate(mv.mv_allocated, NULL, NULL);
range_tree_destroy(mv.mv_allocated);
zfs_btree_clear(&mv.mv_livelist_allocs);
zfs_btree_destroy(&mv.mv_livelist_allocs);
}
}
(void) fprintf(stderr, "\n");
/*
* If there are any segments in the leftover tree after we walked
* through all the metaslabs in the concrete vdevs then this means
* that we have segments in the livelists that belong to indirect
* vdevs and are marked as allocated.
*/
if (zfs_btree_numnodes(&sv.sv_leftover) == 0) {
zfs_btree_destroy(&sv.sv_leftover);
return;
}
(void) printf("ERROR: Found livelist blocks marked as allocated "
"for indirect vdevs:\n");
zfs_btree_index_t *where = NULL;
sublivelist_verify_block_t *svb;
while ((svb = zfs_btree_destroy_nodes(&sv.sv_leftover, &where)) !=
NULL) {
int vdev_id = DVA_GET_VDEV(&svb->svb_dva);
ASSERT3U(vdev_id, <, rvd->vdev_children);
vdev_t *vd = rvd->vdev_child[vdev_id];
ASSERT(!vdev_is_concrete(vd));
(void) printf("<%d:%llx:%llx> TXG %llx\n",
vdev_id, (u_longlong_t)DVA_GET_OFFSET(&svb->svb_dva),
(u_longlong_t)DVA_GET_ASIZE(&svb->svb_dva),
(u_longlong_t)svb->svb_allocated_txg);
}
(void) printf("\n");
zfs_btree_destroy(&sv.sv_leftover);
}
/*
* These libumem hooks provide a reasonable set of defaults for the allocator's
* debugging facilities.
*/
const char *
_umem_debug_init(void)
{
return ("default,verbose"); /* $UMEM_DEBUG setting */
}
const char *
_umem_logging_init(void)
{
return ("fail,contents"); /* $UMEM_LOGGING setting */
}
static void
usage(void)
{
(void) fprintf(stderr,
"Usage:\t%s [-AbcdDFGhikLMPsvXy] [-e [-V] [-p <path> ...]] "
"[-I <inflight I/Os>]\n"
"\t\t[-o <var>=<value>]... [-t <txg>] [-U <cache>] [-x <dumpdir>]\n"
"\t\t[-K <key>]\n"
"\t\t[<poolname>[/<dataset | objset id>] [<object | range> ...]]\n"
"\t%s [-AdiPv] [-e [-V] [-p <path> ...]] [-U <cache>] [-K <key>]\n"
"\t\t[<poolname>[/<dataset | objset id>] [<object | range> ...]\n"
"\t%s -B [-e [-V] [-p <path> ...]] [-I <inflight I/Os>]\n"
"\t\t[-o <var>=<value>]... [-t <txg>] [-U <cache>] [-x <dumpdir>]\n"
"\t\t[-K <key>] <poolname>/<objset id> [<backupflags>]\n"
"\t%s [-v] <bookmark>\n"
"\t%s -C [-A] [-U <cache>] [<poolname>]\n"
"\t%s -l [-Aqu] <device>\n"
"\t%s -m [-AFLPX] [-e [-V] [-p <path> ...]] [-t <txg>] "
"[-U <cache>]\n\t\t<poolname> [<vdev> [<metaslab> ...]]\n"
"\t%s -O [-K <key>] <dataset> <path>\n"
"\t%s -r [-K <key>] <dataset> <path> <destination>\n"
"\t%s -R [-A] [-e [-V] [-p <path> ...]] [-U <cache>]\n"
"\t\t<poolname> <vdev>:<offset>:<size>[:<flags>]\n"
"\t%s -E [-A] word0:word1:...:word15\n"
"\t%s -S [-AP] [-e [-V] [-p <path> ...]] [-U <cache>] "
"<poolname>\n\n",
cmdname, cmdname, cmdname, cmdname, cmdname, cmdname, cmdname,
cmdname, cmdname, cmdname, cmdname, cmdname);
(void) fprintf(stderr, " Dataset name must include at least one "
"separator character '/' or '@'\n");
(void) fprintf(stderr, " If dataset name is specified, only that "
"dataset is dumped\n");
(void) fprintf(stderr, " If object numbers or object number "
"ranges are specified, only those\n"
" objects or ranges are dumped.\n\n");
(void) fprintf(stderr,
" Object ranges take the form <start>:<end>[:<flags>]\n"
" start Starting object number\n"
" end Ending object number, or -1 for no upper bound\n"
" flags Optional flags to select object types:\n"
" A All objects (this is the default)\n"
" d ZFS directories\n"
" f ZFS files \n"
" m SPA space maps\n"
" z ZAPs\n"
" - Negate effect of next flag\n\n");
(void) fprintf(stderr, " Options to control amount of output:\n");
(void) fprintf(stderr, " -b --block-stats "
"block statistics\n");
(void) fprintf(stderr, " -B --backup "
"backup stream\n");
(void) fprintf(stderr, " -c --checksum "
"checksum all metadata (twice for all data) blocks\n");
(void) fprintf(stderr, " -C --config "
"config (or cachefile if alone)\n");
(void) fprintf(stderr, " -d --datasets "
"dataset(s)\n");
(void) fprintf(stderr, " -D --dedup-stats "
"dedup statistics\n");
(void) fprintf(stderr, " -E --embedded-block-pointer=INTEGER\n"
" decode and display block "
"from an embedded block pointer\n");
(void) fprintf(stderr, " -h --history "
"pool history\n");
(void) fprintf(stderr, " -i --intent-logs "
"intent logs\n");
(void) fprintf(stderr, " -l --label "
"read label contents\n");
(void) fprintf(stderr, " -k --checkpointed-state "
"examine the checkpointed state of the pool\n");
(void) fprintf(stderr, " -L --disable-leak-tracking "
"disable leak tracking (do not load spacemaps)\n");
(void) fprintf(stderr, " -m --metaslabs "
"metaslabs\n");
(void) fprintf(stderr, " -M --metaslab-groups "
"metaslab groups\n");
(void) fprintf(stderr, " -O --object-lookups "
"perform object lookups by path\n");
(void) fprintf(stderr, " -r --copy-object "
"copy an object by path to file\n");
(void) fprintf(stderr, " -R --read-block "
"read and display block from a device\n");
(void) fprintf(stderr, " -s --io-stats "
"report stats on zdb's I/O\n");
(void) fprintf(stderr, " -S --simulate-dedup "
"simulate dedup to measure effect\n");
(void) fprintf(stderr, " -v --verbose "
"verbose (applies to all others)\n");
(void) fprintf(stderr, " -y --livelist "
"perform livelist and metaslab validation on any livelists being "
"deleted\n\n");
(void) fprintf(stderr, " Below options are intended for use "
"with other options:\n");
(void) fprintf(stderr, " -A --ignore-assertions "
"ignore assertions (-A), enable panic recovery (-AA) or both "
"(-AAA)\n");
(void) fprintf(stderr, " -e --exported "
"pool is exported/destroyed/has altroot/not in a cachefile\n");
(void) fprintf(stderr, " -F --automatic-rewind "
"attempt automatic rewind within safe range of transaction "
"groups\n");
(void) fprintf(stderr, " -G --dump-debug-msg "
"dump zfs_dbgmsg buffer before exiting\n");
(void) fprintf(stderr, " -I --inflight=INTEGER "
"specify the maximum number of checksumming I/Os "
"[default is 200]\n");
(void) fprintf(stderr, " -K --key=KEY "
"decryption key for encrypted dataset\n");
(void) fprintf(stderr, " -o --option=\"OPTION=INTEGER\" "
"set global variable to an unsigned 32-bit integer\n");
(void) fprintf(stderr, " -p --path==PATH "
"use one or more with -e to specify path to vdev dir\n");
(void) fprintf(stderr, " -P --parseable "
"print numbers in parseable form\n");
(void) fprintf(stderr, " -q --skip-label "
"don't print label contents\n");
(void) fprintf(stderr, " -t --txg=INTEGER "
"highest txg to use when searching for uberblocks\n");
(void) fprintf(stderr, " -u --uberblock "
"uberblock\n");
(void) fprintf(stderr, " -U --cachefile=PATH "
"use alternate cachefile\n");
(void) fprintf(stderr, " -V --verbatim "
"do verbatim import\n");
(void) fprintf(stderr, " -x --dump-blocks=PATH "
"dump all read blocks into specified directory\n");
(void) fprintf(stderr, " -X --extreme-rewind "
"attempt extreme rewind (does not work with dataset)\n");
(void) fprintf(stderr, " -Y --all-reconstruction "
"attempt all reconstruction combinations for split blocks\n");
(void) fprintf(stderr, " -Z --zstd-headers "
"show ZSTD headers \n");
(void) fprintf(stderr, "Specify an option more than once (e.g. -bb) "
"to make only that option verbose\n");
(void) fprintf(stderr, "Default is to dump everything non-verbosely\n");
exit(1);
}
static void
dump_debug_buffer(void)
{
if (dump_opt['G']) {
(void) printf("\n");
(void) fflush(stdout);
zfs_dbgmsg_print("zdb");
}
}
/*
* Called for usage errors that are discovered after a call to spa_open(),
* dmu_bonus_hold(), or pool_match(). abort() is called for other errors.
*/
static void
fatal(const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
(void) fprintf(stderr, "%s: ", cmdname);
(void) vfprintf(stderr, fmt, ap);
va_end(ap);
(void) fprintf(stderr, "\n");
dump_debug_buffer();
exit(1);
}
static void
dump_packed_nvlist(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) size;
nvlist_t *nv;
size_t nvsize = *(uint64_t *)data;
char *packed = umem_alloc(nvsize, UMEM_NOFAIL);
VERIFY(0 == dmu_read(os, object, 0, nvsize, packed, DMU_READ_PREFETCH));
VERIFY(nvlist_unpack(packed, nvsize, &nv, 0) == 0);
umem_free(packed, nvsize);
dump_nvlist(nv, 8);
nvlist_free(nv);
}
static void
dump_history_offsets(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) size;
spa_history_phys_t *shp = data;
if (shp == NULL)
return;
(void) printf("\t\tpool_create_len = %llu\n",
(u_longlong_t)shp->sh_pool_create_len);
(void) printf("\t\tphys_max_off = %llu\n",
(u_longlong_t)shp->sh_phys_max_off);
(void) printf("\t\tbof = %llu\n",
(u_longlong_t)shp->sh_bof);
(void) printf("\t\teof = %llu\n",
(u_longlong_t)shp->sh_eof);
(void) printf("\t\trecords_lost = %llu\n",
(u_longlong_t)shp->sh_records_lost);
}
static void
zdb_nicenum(uint64_t num, char *buf, size_t buflen)
{
if (dump_opt['P'])
(void) snprintf(buf, buflen, "%llu", (longlong_t)num);
else
nicenum(num, buf, buflen);
}
static const char histo_stars[] = "****************************************";
static const uint64_t histo_width = sizeof (histo_stars) - 1;
static void
dump_histogram(const uint64_t *histo, int size, int offset)
{
int i;
int minidx = size - 1;
int maxidx = 0;
uint64_t max = 0;
for (i = 0; i < size; i++) {
if (histo[i] == 0)
continue;
if (histo[i] > max)
max = histo[i];
if (i > maxidx)
maxidx = i;
if (i < minidx)
minidx = i;
}
if (max < histo_width)
max = histo_width;
for (i = minidx; i <= maxidx; i++) {
(void) printf("\t\t\t%3u: %6llu %s\n",
i + offset, (u_longlong_t)histo[i],
&histo_stars[(max - histo[i]) * histo_width / max]);
}
}
static void
dump_zap_stats(objset_t *os, uint64_t object)
{
int error;
zap_stats_t zs;
error = zap_get_stats(os, object, &zs);
if (error)
return;
if (zs.zs_ptrtbl_len == 0) {
ASSERT(zs.zs_num_blocks == 1);
(void) printf("\tmicrozap: %llu bytes, %llu entries\n",
(u_longlong_t)zs.zs_blocksize,
(u_longlong_t)zs.zs_num_entries);
return;
}
(void) printf("\tFat ZAP stats:\n");
(void) printf("\t\tPointer table:\n");
(void) printf("\t\t\t%llu elements\n",
(u_longlong_t)zs.zs_ptrtbl_len);
(void) printf("\t\t\tzt_blk: %llu\n",
(u_longlong_t)zs.zs_ptrtbl_zt_blk);
(void) printf("\t\t\tzt_numblks: %llu\n",
(u_longlong_t)zs.zs_ptrtbl_zt_numblks);
(void) printf("\t\t\tzt_shift: %llu\n",
(u_longlong_t)zs.zs_ptrtbl_zt_shift);
(void) printf("\t\t\tzt_blks_copied: %llu\n",
(u_longlong_t)zs.zs_ptrtbl_blks_copied);
(void) printf("\t\t\tzt_nextblk: %llu\n",
(u_longlong_t)zs.zs_ptrtbl_nextblk);
(void) printf("\t\tZAP entries: %llu\n",
(u_longlong_t)zs.zs_num_entries);
(void) printf("\t\tLeaf blocks: %llu\n",
(u_longlong_t)zs.zs_num_leafs);
(void) printf("\t\tTotal blocks: %llu\n",
(u_longlong_t)zs.zs_num_blocks);
(void) printf("\t\tzap_block_type: 0x%llx\n",
(u_longlong_t)zs.zs_block_type);
(void) printf("\t\tzap_magic: 0x%llx\n",
(u_longlong_t)zs.zs_magic);
(void) printf("\t\tzap_salt: 0x%llx\n",
(u_longlong_t)zs.zs_salt);
(void) printf("\t\tLeafs with 2^n pointers:\n");
dump_histogram(zs.zs_leafs_with_2n_pointers, ZAP_HISTOGRAM_SIZE, 0);
(void) printf("\t\tBlocks with n*5 entries:\n");
dump_histogram(zs.zs_blocks_with_n5_entries, ZAP_HISTOGRAM_SIZE, 0);
(void) printf("\t\tBlocks n/10 full:\n");
dump_histogram(zs.zs_blocks_n_tenths_full, ZAP_HISTOGRAM_SIZE, 0);
(void) printf("\t\tEntries with n chunks:\n");
dump_histogram(zs.zs_entries_using_n_chunks, ZAP_HISTOGRAM_SIZE, 0);
(void) printf("\t\tBuckets with n entries:\n");
dump_histogram(zs.zs_buckets_with_n_entries, ZAP_HISTOGRAM_SIZE, 0);
}
static void
dump_none(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
}
static void
dump_unknown(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
(void) printf("\tUNKNOWN OBJECT TYPE\n");
}
static void
dump_uint8(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
}
static void
dump_uint64(objset_t *os, uint64_t object, void *data, size_t size)
{
uint64_t *arr;
uint64_t oursize;
if (dump_opt['d'] < 6)
return;
if (data == NULL) {
dmu_object_info_t doi;
VERIFY0(dmu_object_info(os, object, &doi));
size = doi.doi_max_offset;
/*
* We cap the size at 1 mebibyte here to prevent
* allocation failures and nigh-infinite printing if the
* object is extremely large.
*/
oursize = MIN(size, 1 << 20);
arr = kmem_alloc(oursize, KM_SLEEP);
int err = dmu_read(os, object, 0, oursize, arr, 0);
if (err != 0) {
(void) printf("got error %u from dmu_read\n", err);
kmem_free(arr, oursize);
return;
}
} else {
/*
* Even though the allocation is already done in this code path,
* we still cap the size to prevent excessive printing.
*/
oursize = MIN(size, 1 << 20);
arr = data;
}
if (size == 0) {
if (data == NULL)
kmem_free(arr, oursize);
(void) printf("\t\t[]\n");
return;
}
(void) printf("\t\t[%0llx", (u_longlong_t)arr[0]);
for (size_t i = 1; i * sizeof (uint64_t) < oursize; i++) {
if (i % 4 != 0)
(void) printf(", %0llx", (u_longlong_t)arr[i]);
else
(void) printf(",\n\t\t%0llx", (u_longlong_t)arr[i]);
}
if (oursize != size)
(void) printf(", ... ");
(void) printf("]\n");
if (data == NULL)
kmem_free(arr, oursize);
}
static void
dump_zap(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
zap_cursor_t zc;
zap_attribute_t attr;
void *prop;
unsigned i;
dump_zap_stats(os, object);
(void) printf("\n");
for (zap_cursor_init(&zc, os, object);
zap_cursor_retrieve(&zc, &attr) == 0;
zap_cursor_advance(&zc)) {
(void) printf("\t\t%s = ", attr.za_name);
if (attr.za_num_integers == 0) {
(void) printf("\n");
continue;
}
prop = umem_zalloc(attr.za_num_integers *
attr.za_integer_length, UMEM_NOFAIL);
(void) zap_lookup(os, object, attr.za_name,
attr.za_integer_length, attr.za_num_integers, prop);
if (attr.za_integer_length == 1) {
if (strcmp(attr.za_name,
DSL_CRYPTO_KEY_MASTER_KEY) == 0 ||
strcmp(attr.za_name,
DSL_CRYPTO_KEY_HMAC_KEY) == 0 ||
strcmp(attr.za_name, DSL_CRYPTO_KEY_IV) == 0 ||
strcmp(attr.za_name, DSL_CRYPTO_KEY_MAC) == 0 ||
strcmp(attr.za_name, DMU_POOL_CHECKSUM_SALT) == 0) {
uint8_t *u8 = prop;
for (i = 0; i < attr.za_num_integers; i++) {
(void) printf("%02x", u8[i]);
}
} else {
(void) printf("%s", (char *)prop);
}
} else {
for (i = 0; i < attr.za_num_integers; i++) {
switch (attr.za_integer_length) {
case 2:
(void) printf("%u ",
((uint16_t *)prop)[i]);
break;
case 4:
(void) printf("%u ",
((uint32_t *)prop)[i]);
break;
case 8:
(void) printf("%lld ",
(u_longlong_t)((int64_t *)prop)[i]);
break;
}
}
}
(void) printf("\n");
umem_free(prop, attr.za_num_integers * attr.za_integer_length);
}
zap_cursor_fini(&zc);
}
static void
dump_bpobj(objset_t *os, uint64_t object, void *data, size_t size)
{
bpobj_phys_t *bpop = data;
uint64_t i;
char bytes[32], comp[32], uncomp[32];
/* make sure the output won't get truncated */
_Static_assert(sizeof (bytes) >= NN_NUMBUF_SZ, "bytes truncated");
_Static_assert(sizeof (comp) >= NN_NUMBUF_SZ, "comp truncated");
_Static_assert(sizeof (uncomp) >= NN_NUMBUF_SZ, "uncomp truncated");
if (bpop == NULL)
return;
zdb_nicenum(bpop->bpo_bytes, bytes, sizeof (bytes));
zdb_nicenum(bpop->bpo_comp, comp, sizeof (comp));
zdb_nicenum(bpop->bpo_uncomp, uncomp, sizeof (uncomp));
(void) printf("\t\tnum_blkptrs = %llu\n",
(u_longlong_t)bpop->bpo_num_blkptrs);
(void) printf("\t\tbytes = %s\n", bytes);
if (size >= BPOBJ_SIZE_V1) {
(void) printf("\t\tcomp = %s\n", comp);
(void) printf("\t\tuncomp = %s\n", uncomp);
}
if (size >= BPOBJ_SIZE_V2) {
(void) printf("\t\tsubobjs = %llu\n",
(u_longlong_t)bpop->bpo_subobjs);
(void) printf("\t\tnum_subobjs = %llu\n",
(u_longlong_t)bpop->bpo_num_subobjs);
}
if (size >= sizeof (*bpop)) {
(void) printf("\t\tnum_freed = %llu\n",
(u_longlong_t)bpop->bpo_num_freed);
}
if (dump_opt['d'] < 5)
return;
for (i = 0; i < bpop->bpo_num_blkptrs; i++) {
char blkbuf[BP_SPRINTF_LEN];
blkptr_t bp;
int err = dmu_read(os, object,
i * sizeof (bp), sizeof (bp), &bp, 0);
if (err != 0) {
(void) printf("got error %u from dmu_read\n", err);
break;
}
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf), &bp,
BP_GET_FREE(&bp));
(void) printf("\t%s\n", blkbuf);
}
}
static void
dump_bpobj_subobjs(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
dmu_object_info_t doi;
int64_t i;
VERIFY0(dmu_object_info(os, object, &doi));
uint64_t *subobjs = kmem_alloc(doi.doi_max_offset, KM_SLEEP);
int err = dmu_read(os, object, 0, doi.doi_max_offset, subobjs, 0);
if (err != 0) {
(void) printf("got error %u from dmu_read\n", err);
kmem_free(subobjs, doi.doi_max_offset);
return;
}
int64_t last_nonzero = -1;
for (i = 0; i < doi.doi_max_offset / 8; i++) {
if (subobjs[i] != 0)
last_nonzero = i;
}
for (i = 0; i <= last_nonzero; i++) {
(void) printf("\t%llu\n", (u_longlong_t)subobjs[i]);
}
kmem_free(subobjs, doi.doi_max_offset);
}
static void
dump_ddt_zap(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
dump_zap_stats(os, object);
/* contents are printed elsewhere, properly decoded */
}
static void
dump_sa_attrs(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
zap_cursor_t zc;
zap_attribute_t attr;
dump_zap_stats(os, object);
(void) printf("\n");
for (zap_cursor_init(&zc, os, object);
zap_cursor_retrieve(&zc, &attr) == 0;
zap_cursor_advance(&zc)) {
(void) printf("\t\t%s = ", attr.za_name);
if (attr.za_num_integers == 0) {
(void) printf("\n");
continue;
}
(void) printf(" %llx : [%d:%d:%d]\n",
(u_longlong_t)attr.za_first_integer,
(int)ATTR_LENGTH(attr.za_first_integer),
(int)ATTR_BSWAP(attr.za_first_integer),
(int)ATTR_NUM(attr.za_first_integer));
}
zap_cursor_fini(&zc);
}
static void
dump_sa_layouts(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
zap_cursor_t zc;
zap_attribute_t attr;
uint16_t *layout_attrs;
unsigned i;
dump_zap_stats(os, object);
(void) printf("\n");
for (zap_cursor_init(&zc, os, object);
zap_cursor_retrieve(&zc, &attr) == 0;
zap_cursor_advance(&zc)) {
(void) printf("\t\t%s = [", attr.za_name);
if (attr.za_num_integers == 0) {
(void) printf("\n");
continue;
}
VERIFY(attr.za_integer_length == 2);
layout_attrs = umem_zalloc(attr.za_num_integers *
attr.za_integer_length, UMEM_NOFAIL);
VERIFY(zap_lookup(os, object, attr.za_name,
attr.za_integer_length,
attr.za_num_integers, layout_attrs) == 0);
for (i = 0; i != attr.za_num_integers; i++)
(void) printf(" %d ", (int)layout_attrs[i]);
(void) printf("]\n");
umem_free(layout_attrs,
attr.za_num_integers * attr.za_integer_length);
}
zap_cursor_fini(&zc);
}
static void
dump_zpldir(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
zap_cursor_t zc;
zap_attribute_t attr;
const char *typenames[] = {
/* 0 */ "not specified",
/* 1 */ "FIFO",
/* 2 */ "Character Device",
/* 3 */ "3 (invalid)",
/* 4 */ "Directory",
/* 5 */ "5 (invalid)",
/* 6 */ "Block Device",
/* 7 */ "7 (invalid)",
/* 8 */ "Regular File",
/* 9 */ "9 (invalid)",
/* 10 */ "Symbolic Link",
/* 11 */ "11 (invalid)",
/* 12 */ "Socket",
/* 13 */ "Door",
/* 14 */ "Event Port",
/* 15 */ "15 (invalid)",
};
dump_zap_stats(os, object);
(void) printf("\n");
for (zap_cursor_init(&zc, os, object);
zap_cursor_retrieve(&zc, &attr) == 0;
zap_cursor_advance(&zc)) {
(void) printf("\t\t%s = %lld (type: %s)\n",
attr.za_name, ZFS_DIRENT_OBJ(attr.za_first_integer),
typenames[ZFS_DIRENT_TYPE(attr.za_first_integer)]);
}
zap_cursor_fini(&zc);
}
static int
get_dtl_refcount(vdev_t *vd)
{
int refcount = 0;
if (vd->vdev_ops->vdev_op_leaf) {
space_map_t *sm = vd->vdev_dtl_sm;
if (sm != NULL &&
sm->sm_dbuf->db_size == sizeof (space_map_phys_t))
return (1);
return (0);
}
for (unsigned c = 0; c < vd->vdev_children; c++)
refcount += get_dtl_refcount(vd->vdev_child[c]);
return (refcount);
}
static int
get_metaslab_refcount(vdev_t *vd)
{
int refcount = 0;
if (vd->vdev_top == vd) {
for (uint64_t m = 0; m < vd->vdev_ms_count; m++) {
space_map_t *sm = vd->vdev_ms[m]->ms_sm;
if (sm != NULL &&
sm->sm_dbuf->db_size == sizeof (space_map_phys_t))
refcount++;
}
}
for (unsigned c = 0; c < vd->vdev_children; c++)
refcount += get_metaslab_refcount(vd->vdev_child[c]);
return (refcount);
}
static int
get_obsolete_refcount(vdev_t *vd)
{
uint64_t obsolete_sm_object;
int refcount = 0;
VERIFY0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
if (vd->vdev_top == vd && obsolete_sm_object != 0) {
dmu_object_info_t doi;
VERIFY0(dmu_object_info(vd->vdev_spa->spa_meta_objset,
obsolete_sm_object, &doi));
if (doi.doi_bonus_size == sizeof (space_map_phys_t)) {
refcount++;
}
} else {
ASSERT3P(vd->vdev_obsolete_sm, ==, NULL);
ASSERT3U(obsolete_sm_object, ==, 0);
}
for (unsigned c = 0; c < vd->vdev_children; c++) {
refcount += get_obsolete_refcount(vd->vdev_child[c]);
}
return (refcount);
}
static int
get_prev_obsolete_spacemap_refcount(spa_t *spa)
{
uint64_t prev_obj =
spa->spa_condensing_indirect_phys.scip_prev_obsolete_sm_object;
if (prev_obj != 0) {
dmu_object_info_t doi;
VERIFY0(dmu_object_info(spa->spa_meta_objset, prev_obj, &doi));
if (doi.doi_bonus_size == sizeof (space_map_phys_t)) {
return (1);
}
}
return (0);
}
static int
get_checkpoint_refcount(vdev_t *vd)
{
int refcount = 0;
if (vd->vdev_top == vd && vd->vdev_top_zap != 0 &&
zap_contains(spa_meta_objset(vd->vdev_spa),
vd->vdev_top_zap, VDEV_TOP_ZAP_POOL_CHECKPOINT_SM) == 0)
refcount++;
for (uint64_t c = 0; c < vd->vdev_children; c++)
refcount += get_checkpoint_refcount(vd->vdev_child[c]);
return (refcount);
}
static int
get_log_spacemap_refcount(spa_t *spa)
{
return (avl_numnodes(&spa->spa_sm_logs_by_txg));
}
static int
verify_spacemap_refcounts(spa_t *spa)
{
uint64_t expected_refcount = 0;
uint64_t actual_refcount;
(void) feature_get_refcount(spa,
&spa_feature_table[SPA_FEATURE_SPACEMAP_HISTOGRAM],
&expected_refcount);
actual_refcount = get_dtl_refcount(spa->spa_root_vdev);
actual_refcount += get_metaslab_refcount(spa->spa_root_vdev);
actual_refcount += get_obsolete_refcount(spa->spa_root_vdev);
actual_refcount += get_prev_obsolete_spacemap_refcount(spa);
actual_refcount += get_checkpoint_refcount(spa->spa_root_vdev);
actual_refcount += get_log_spacemap_refcount(spa);
if (expected_refcount != actual_refcount) {
(void) printf("space map refcount mismatch: expected %lld != "
"actual %lld\n",
(longlong_t)expected_refcount,
(longlong_t)actual_refcount);
return (2);
}
return (0);
}
static void
dump_spacemap(objset_t *os, space_map_t *sm)
{
const char *ddata[] = { "ALLOC", "FREE", "CONDENSE", "INVALID",
"INVALID", "INVALID", "INVALID", "INVALID" };
if (sm == NULL)
return;
(void) printf("space map object %llu:\n",
(longlong_t)sm->sm_object);
(void) printf(" smp_length = 0x%llx\n",
(longlong_t)sm->sm_phys->smp_length);
(void) printf(" smp_alloc = 0x%llx\n",
(longlong_t)sm->sm_phys->smp_alloc);
if (dump_opt['d'] < 6 && dump_opt['m'] < 4)
return;
/*
* Print out the freelist entries in both encoded and decoded form.
*/
uint8_t mapshift = sm->sm_shift;
int64_t alloc = 0;
uint64_t word, entry_id = 0;
for (uint64_t offset = 0; offset < space_map_length(sm);
offset += sizeof (word)) {
VERIFY0(dmu_read(os, space_map_object(sm), offset,
sizeof (word), &word, DMU_READ_PREFETCH));
if (sm_entry_is_debug(word)) {
uint64_t de_txg = SM_DEBUG_TXG_DECODE(word);
uint64_t de_sync_pass = SM_DEBUG_SYNCPASS_DECODE(word);
if (de_txg == 0) {
(void) printf(
"\t [%6llu] PADDING\n",
(u_longlong_t)entry_id);
} else {
(void) printf(
"\t [%6llu] %s: txg %llu pass %llu\n",
(u_longlong_t)entry_id,
ddata[SM_DEBUG_ACTION_DECODE(word)],
(u_longlong_t)de_txg,
(u_longlong_t)de_sync_pass);
}
entry_id++;
continue;
}
uint8_t words;
char entry_type;
uint64_t entry_off, entry_run, entry_vdev = SM_NO_VDEVID;
if (sm_entry_is_single_word(word)) {
entry_type = (SM_TYPE_DECODE(word) == SM_ALLOC) ?
'A' : 'F';
entry_off = (SM_OFFSET_DECODE(word) << mapshift) +
sm->sm_start;
entry_run = SM_RUN_DECODE(word) << mapshift;
words = 1;
} else {
/* it is a two-word entry so we read another word */
ASSERT(sm_entry_is_double_word(word));
uint64_t extra_word;
offset += sizeof (extra_word);
VERIFY0(dmu_read(os, space_map_object(sm), offset,
sizeof (extra_word), &extra_word,
DMU_READ_PREFETCH));
ASSERT3U(offset, <=, space_map_length(sm));
entry_run = SM2_RUN_DECODE(word) << mapshift;
entry_vdev = SM2_VDEV_DECODE(word);
entry_type = (SM2_TYPE_DECODE(extra_word) == SM_ALLOC) ?
'A' : 'F';
entry_off = (SM2_OFFSET_DECODE(extra_word) <<
mapshift) + sm->sm_start;
words = 2;
}
(void) printf("\t [%6llu] %c range:"
" %010llx-%010llx size: %06llx vdev: %06llu words: %u\n",
(u_longlong_t)entry_id,
entry_type, (u_longlong_t)entry_off,
(u_longlong_t)(entry_off + entry_run),
(u_longlong_t)entry_run,
(u_longlong_t)entry_vdev, words);
if (entry_type == 'A')
alloc += entry_run;
else
alloc -= entry_run;
entry_id++;
}
if (alloc != space_map_allocated(sm)) {
(void) printf("space_map_object alloc (%lld) INCONSISTENT "
"with space map summary (%lld)\n",
(longlong_t)space_map_allocated(sm), (longlong_t)alloc);
}
}
static void
dump_metaslab_stats(metaslab_t *msp)
{
char maxbuf[32];
range_tree_t *rt = msp->ms_allocatable;
zfs_btree_t *t = &msp->ms_allocatable_by_size;
int free_pct = range_tree_space(rt) * 100 / msp->ms_size;
/* max sure nicenum has enough space */
_Static_assert(sizeof (maxbuf) >= NN_NUMBUF_SZ, "maxbuf truncated");
zdb_nicenum(metaslab_largest_allocatable(msp), maxbuf, sizeof (maxbuf));
(void) printf("\t %25s %10lu %7s %6s %4s %4d%%\n",
"segments", zfs_btree_numnodes(t), "maxsize", maxbuf,
"freepct", free_pct);
(void) printf("\tIn-memory histogram:\n");
dump_histogram(rt->rt_histogram, RANGE_TREE_HISTOGRAM_SIZE, 0);
}
static void
dump_metaslab(metaslab_t *msp)
{
vdev_t *vd = msp->ms_group->mg_vd;
spa_t *spa = vd->vdev_spa;
space_map_t *sm = msp->ms_sm;
char freebuf[32];
zdb_nicenum(msp->ms_size - space_map_allocated(sm), freebuf,
sizeof (freebuf));
(void) printf(
"\tmetaslab %6llu offset %12llx spacemap %6llu free %5s\n",
(u_longlong_t)msp->ms_id, (u_longlong_t)msp->ms_start,
(u_longlong_t)space_map_object(sm), freebuf);
if (dump_opt['m'] > 2 && !dump_opt['L']) {
mutex_enter(&msp->ms_lock);
VERIFY0(metaslab_load(msp));
range_tree_stat_verify(msp->ms_allocatable);
dump_metaslab_stats(msp);
metaslab_unload(msp);
mutex_exit(&msp->ms_lock);
}
if (dump_opt['m'] > 1 && sm != NULL &&
spa_feature_is_active(spa, SPA_FEATURE_SPACEMAP_HISTOGRAM)) {
/*
* The space map histogram represents free space in chunks
* of sm_shift (i.e. bucket 0 refers to 2^sm_shift).
*/
(void) printf("\tOn-disk histogram:\t\tfragmentation %llu\n",
(u_longlong_t)msp->ms_fragmentation);
dump_histogram(sm->sm_phys->smp_histogram,
SPACE_MAP_HISTOGRAM_SIZE, sm->sm_shift);
}
if (vd->vdev_ops == &vdev_draid_ops)
ASSERT3U(msp->ms_size, <=, 1ULL << vd->vdev_ms_shift);
else
ASSERT3U(msp->ms_size, ==, 1ULL << vd->vdev_ms_shift);
dump_spacemap(spa->spa_meta_objset, msp->ms_sm);
if (spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP)) {
(void) printf("\tFlush data:\n\tunflushed txg=%llu\n\n",
(u_longlong_t)metaslab_unflushed_txg(msp));
}
}
static void
print_vdev_metaslab_header(vdev_t *vd)
{
vdev_alloc_bias_t alloc_bias = vd->vdev_alloc_bias;
const char *bias_str = "";
if (alloc_bias == VDEV_BIAS_LOG || vd->vdev_islog) {
bias_str = VDEV_ALLOC_BIAS_LOG;
} else if (alloc_bias == VDEV_BIAS_SPECIAL) {
bias_str = VDEV_ALLOC_BIAS_SPECIAL;
} else if (alloc_bias == VDEV_BIAS_DEDUP) {
bias_str = VDEV_ALLOC_BIAS_DEDUP;
}
uint64_t ms_flush_data_obj = 0;
if (vd->vdev_top_zap != 0) {
int error = zap_lookup(spa_meta_objset(vd->vdev_spa),
vd->vdev_top_zap, VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS,
sizeof (uint64_t), 1, &ms_flush_data_obj);
if (error != ENOENT) {
ASSERT0(error);
}
}
(void) printf("\tvdev %10llu %s",
(u_longlong_t)vd->vdev_id, bias_str);
if (ms_flush_data_obj != 0) {
(void) printf(" ms_unflushed_phys object %llu",
(u_longlong_t)ms_flush_data_obj);
}
(void) printf("\n\t%-10s%5llu %-19s %-15s %-12s\n",
"metaslabs", (u_longlong_t)vd->vdev_ms_count,
"offset", "spacemap", "free");
(void) printf("\t%15s %19s %15s %12s\n",
"---------------", "-------------------",
"---------------", "------------");
}
static void
dump_metaslab_groups(spa_t *spa, boolean_t show_special)
{
vdev_t *rvd = spa->spa_root_vdev;
metaslab_class_t *mc = spa_normal_class(spa);
metaslab_class_t *smc = spa_special_class(spa);
uint64_t fragmentation;
metaslab_class_histogram_verify(mc);
for (unsigned c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
metaslab_group_t *mg = tvd->vdev_mg;
if (mg == NULL || (mg->mg_class != mc &&
(!show_special || mg->mg_class != smc)))
continue;
metaslab_group_histogram_verify(mg);
mg->mg_fragmentation = metaslab_group_fragmentation(mg);
(void) printf("\tvdev %10llu\t\tmetaslabs%5llu\t\t"
"fragmentation",
(u_longlong_t)tvd->vdev_id,
(u_longlong_t)tvd->vdev_ms_count);
if (mg->mg_fragmentation == ZFS_FRAG_INVALID) {
(void) printf("%3s\n", "-");
} else {
(void) printf("%3llu%%\n",
(u_longlong_t)mg->mg_fragmentation);
}
dump_histogram(mg->mg_histogram, RANGE_TREE_HISTOGRAM_SIZE, 0);
}
(void) printf("\tpool %s\tfragmentation", spa_name(spa));
fragmentation = metaslab_class_fragmentation(mc);
if (fragmentation == ZFS_FRAG_INVALID)
(void) printf("\t%3s\n", "-");
else
(void) printf("\t%3llu%%\n", (u_longlong_t)fragmentation);
dump_histogram(mc->mc_histogram, RANGE_TREE_HISTOGRAM_SIZE, 0);
}
static void
print_vdev_indirect(vdev_t *vd)
{
vdev_indirect_config_t *vic = &vd->vdev_indirect_config;
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
vdev_indirect_births_t *vib = vd->vdev_indirect_births;
if (vim == NULL) {
ASSERT3P(vib, ==, NULL);
return;
}
ASSERT3U(vdev_indirect_mapping_object(vim), ==,
vic->vic_mapping_object);
ASSERT3U(vdev_indirect_births_object(vib), ==,
vic->vic_births_object);
(void) printf("indirect births obj %llu:\n",
(longlong_t)vic->vic_births_object);
(void) printf(" vib_count = %llu\n",
(longlong_t)vdev_indirect_births_count(vib));
for (uint64_t i = 0; i < vdev_indirect_births_count(vib); i++) {
vdev_indirect_birth_entry_phys_t *cur_vibe =
&vib->vib_entries[i];
(void) printf("\toffset %llx -> txg %llu\n",
(longlong_t)cur_vibe->vibe_offset,
(longlong_t)cur_vibe->vibe_phys_birth_txg);
}
(void) printf("\n");
(void) printf("indirect mapping obj %llu:\n",
(longlong_t)vic->vic_mapping_object);
(void) printf(" vim_max_offset = 0x%llx\n",
(longlong_t)vdev_indirect_mapping_max_offset(vim));
(void) printf(" vim_bytes_mapped = 0x%llx\n",
(longlong_t)vdev_indirect_mapping_bytes_mapped(vim));
(void) printf(" vim_count = %llu\n",
(longlong_t)vdev_indirect_mapping_num_entries(vim));
if (dump_opt['d'] <= 5 && dump_opt['m'] <= 3)
return;
uint32_t *counts = vdev_indirect_mapping_load_obsolete_counts(vim);
for (uint64_t i = 0; i < vdev_indirect_mapping_num_entries(vim); i++) {
vdev_indirect_mapping_entry_phys_t *vimep =
&vim->vim_entries[i];
(void) printf("\t<%llx:%llx:%llx> -> "
"<%llx:%llx:%llx> (%x obsolete)\n",
(longlong_t)vd->vdev_id,
(longlong_t)DVA_MAPPING_GET_SRC_OFFSET(vimep),
(longlong_t)DVA_GET_ASIZE(&vimep->vimep_dst),
(longlong_t)DVA_GET_VDEV(&vimep->vimep_dst),
(longlong_t)DVA_GET_OFFSET(&vimep->vimep_dst),
(longlong_t)DVA_GET_ASIZE(&vimep->vimep_dst),
counts[i]);
}
(void) printf("\n");
uint64_t obsolete_sm_object;
VERIFY0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
if (obsolete_sm_object != 0) {
objset_t *mos = vd->vdev_spa->spa_meta_objset;
(void) printf("obsolete space map object %llu:\n",
(u_longlong_t)obsolete_sm_object);
ASSERT(vd->vdev_obsolete_sm != NULL);
ASSERT3U(space_map_object(vd->vdev_obsolete_sm), ==,
obsolete_sm_object);
dump_spacemap(mos, vd->vdev_obsolete_sm);
(void) printf("\n");
}
}
static void
dump_metaslabs(spa_t *spa)
{
vdev_t *vd, *rvd = spa->spa_root_vdev;
uint64_t m, c = 0, children = rvd->vdev_children;
(void) printf("\nMetaslabs:\n");
if (!dump_opt['d'] && zopt_metaslab_args > 0) {
c = zopt_metaslab[0];
if (c >= children)
(void) fatal("bad vdev id: %llu", (u_longlong_t)c);
if (zopt_metaslab_args > 1) {
vd = rvd->vdev_child[c];
print_vdev_metaslab_header(vd);
for (m = 1; m < zopt_metaslab_args; m++) {
if (zopt_metaslab[m] < vd->vdev_ms_count)
dump_metaslab(
vd->vdev_ms[zopt_metaslab[m]]);
else
(void) fprintf(stderr, "bad metaslab "
"number %llu\n",
(u_longlong_t)zopt_metaslab[m]);
}
(void) printf("\n");
return;
}
children = c + 1;
}
for (; c < children; c++) {
vd = rvd->vdev_child[c];
print_vdev_metaslab_header(vd);
print_vdev_indirect(vd);
for (m = 0; m < vd->vdev_ms_count; m++)
dump_metaslab(vd->vdev_ms[m]);
(void) printf("\n");
}
}
static void
dump_log_spacemaps(spa_t *spa)
{
if (!spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP))
return;
(void) printf("\nLog Space Maps in Pool:\n");
for (spa_log_sm_t *sls = avl_first(&spa->spa_sm_logs_by_txg);
sls; sls = AVL_NEXT(&spa->spa_sm_logs_by_txg, sls)) {
space_map_t *sm = NULL;
VERIFY0(space_map_open(&sm, spa_meta_objset(spa),
sls->sls_sm_obj, 0, UINT64_MAX, SPA_MINBLOCKSHIFT));
(void) printf("Log Spacemap object %llu txg %llu\n",
(u_longlong_t)sls->sls_sm_obj, (u_longlong_t)sls->sls_txg);
dump_spacemap(spa->spa_meta_objset, sm);
space_map_close(sm);
}
(void) printf("\n");
}
static void
dump_dde(const ddt_t *ddt, const ddt_entry_t *dde, uint64_t index)
{
const ddt_phys_t *ddp = dde->dde_phys;
const ddt_key_t *ddk = &dde->dde_key;
const char *types[4] = { "ditto", "single", "double", "triple" };
char blkbuf[BP_SPRINTF_LEN];
blkptr_t blk;
int p;
for (p = 0; p < DDT_PHYS_TYPES; p++, ddp++) {
if (ddp->ddp_phys_birth == 0)
continue;
ddt_bp_create(ddt->ddt_checksum, ddk, ddp, &blk);
snprintf_blkptr(blkbuf, sizeof (blkbuf), &blk);
(void) printf("index %llx refcnt %llu %s %s\n",
(u_longlong_t)index, (u_longlong_t)ddp->ddp_refcnt,
types[p], blkbuf);
}
}
static void
dump_dedup_ratio(const ddt_stat_t *dds)
{
double rL, rP, rD, D, dedup, compress, copies;
if (dds->dds_blocks == 0)
return;
rL = (double)dds->dds_ref_lsize;
rP = (double)dds->dds_ref_psize;
rD = (double)dds->dds_ref_dsize;
D = (double)dds->dds_dsize;
dedup = rD / D;
compress = rL / rP;
copies = rD / rP;
(void) printf("dedup = %.2f, compress = %.2f, copies = %.2f, "
"dedup * compress / copies = %.2f\n\n",
dedup, compress, copies, dedup * compress / copies);
}
static void
dump_ddt(ddt_t *ddt, enum ddt_type type, enum ddt_class class)
{
char name[DDT_NAMELEN];
ddt_entry_t dde;
uint64_t walk = 0;
dmu_object_info_t doi;
uint64_t count, dspace, mspace;
int error;
error = ddt_object_info(ddt, type, class, &doi);
if (error == ENOENT)
return;
ASSERT(error == 0);
error = ddt_object_count(ddt, type, class, &count);
ASSERT(error == 0);
if (count == 0)
return;
dspace = doi.doi_physical_blocks_512 << 9;
mspace = doi.doi_fill_count * doi.doi_data_block_size;
ddt_object_name(ddt, type, class, name);
(void) printf("%s: %llu entries, size %llu on disk, %llu in core\n",
name,
(u_longlong_t)count,
(u_longlong_t)(dspace / count),
(u_longlong_t)(mspace / count));
if (dump_opt['D'] < 3)
return;
zpool_dump_ddt(NULL, &ddt->ddt_histogram[type][class]);
if (dump_opt['D'] < 4)
return;
if (dump_opt['D'] < 5 && class == DDT_CLASS_UNIQUE)
return;
(void) printf("%s contents:\n\n", name);
while ((error = ddt_object_walk(ddt, type, class, &walk, &dde)) == 0)
dump_dde(ddt, &dde, walk);
ASSERT3U(error, ==, ENOENT);
(void) printf("\n");
}
static void
dump_all_ddts(spa_t *spa)
{
ddt_histogram_t ddh_total = {{{0}}};
ddt_stat_t dds_total = {0};
for (enum zio_checksum c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
ddt_t *ddt = spa->spa_ddt[c];
for (enum ddt_type type = 0; type < DDT_TYPES; type++) {
for (enum ddt_class class = 0; class < DDT_CLASSES;
class++) {
dump_ddt(ddt, type, class);
}
}
}
ddt_get_dedup_stats(spa, &dds_total);
if (dds_total.dds_blocks == 0) {
(void) printf("All DDTs are empty\n");
return;
}
(void) printf("\n");
if (dump_opt['D'] > 1) {
(void) printf("DDT histogram (aggregated over all DDTs):\n");
ddt_get_dedup_histogram(spa, &ddh_total);
zpool_dump_ddt(&dds_total, &ddh_total);
}
dump_dedup_ratio(&dds_total);
}
static void
dump_dtl_seg(void *arg, uint64_t start, uint64_t size)
{
char *prefix = arg;
(void) printf("%s [%llu,%llu) length %llu\n",
prefix,
(u_longlong_t)start,
(u_longlong_t)(start + size),
(u_longlong_t)(size));
}
static void
dump_dtl(vdev_t *vd, int indent)
{
spa_t *spa = vd->vdev_spa;
boolean_t required;
const char *name[DTL_TYPES] = { "missing", "partial", "scrub",
"outage" };
char prefix[256];
spa_vdev_state_enter(spa, SCL_NONE);
required = vdev_dtl_required(vd);
(void) spa_vdev_state_exit(spa, NULL, 0);
if (indent == 0)
(void) printf("\nDirty time logs:\n\n");
(void) printf("\t%*s%s [%s]\n", indent, "",
vd->vdev_path ? vd->vdev_path :
vd->vdev_parent ? vd->vdev_ops->vdev_op_type : spa_name(spa),
required ? "DTL-required" : "DTL-expendable");
for (int t = 0; t < DTL_TYPES; t++) {
range_tree_t *rt = vd->vdev_dtl[t];
if (range_tree_space(rt) == 0)
continue;
(void) snprintf(prefix, sizeof (prefix), "\t%*s%s",
indent + 2, "", name[t]);
range_tree_walk(rt, dump_dtl_seg, prefix);
if (dump_opt['d'] > 5 && vd->vdev_children == 0)
dump_spacemap(spa->spa_meta_objset,
vd->vdev_dtl_sm);
}
for (unsigned c = 0; c < vd->vdev_children; c++)
dump_dtl(vd->vdev_child[c], indent + 4);
}
static void
dump_history(spa_t *spa)
{
nvlist_t **events = NULL;
char *buf;
uint64_t resid, len, off = 0;
uint_t num = 0;
int error;
char tbuf[30];
if ((buf = malloc(SPA_OLD_MAXBLOCKSIZE)) == NULL) {
(void) fprintf(stderr, "%s: unable to allocate I/O buffer\n",
__func__);
return;
}
do {
len = SPA_OLD_MAXBLOCKSIZE;
if ((error = spa_history_get(spa, &off, &len, buf)) != 0) {
(void) fprintf(stderr, "Unable to read history: "
"error %d\n", error);
free(buf);
return;
}
if (zpool_history_unpack(buf, len, &resid, &events, &num) != 0)
break;
off -= resid;
} while (len != 0);
(void) printf("\nHistory:\n");
for (unsigned i = 0; i < num; i++) {
boolean_t printed = B_FALSE;
if (nvlist_exists(events[i], ZPOOL_HIST_TIME)) {
time_t tsec;
struct tm t;
tsec = fnvlist_lookup_uint64(events[i],
ZPOOL_HIST_TIME);
(void) localtime_r(&tsec, &t);
(void) strftime(tbuf, sizeof (tbuf), "%F.%T", &t);
} else {
tbuf[0] = '\0';
}
if (nvlist_exists(events[i], ZPOOL_HIST_CMD)) {
(void) printf("%s %s\n", tbuf,
fnvlist_lookup_string(events[i], ZPOOL_HIST_CMD));
} else if (nvlist_exists(events[i], ZPOOL_HIST_INT_EVENT)) {
uint64_t ievent;
ievent = fnvlist_lookup_uint64(events[i],
ZPOOL_HIST_INT_EVENT);
if (ievent >= ZFS_NUM_LEGACY_HISTORY_EVENTS)
goto next;
(void) printf(" %s [internal %s txg:%ju] %s\n",
tbuf,
zfs_history_event_names[ievent],
fnvlist_lookup_uint64(events[i],
ZPOOL_HIST_TXG),
fnvlist_lookup_string(events[i],
ZPOOL_HIST_INT_STR));
} else if (nvlist_exists(events[i], ZPOOL_HIST_INT_NAME)) {
(void) printf("%s [txg:%ju] %s", tbuf,
fnvlist_lookup_uint64(events[i],
ZPOOL_HIST_TXG),
fnvlist_lookup_string(events[i],
ZPOOL_HIST_INT_NAME));
if (nvlist_exists(events[i], ZPOOL_HIST_DSNAME)) {
(void) printf(" %s (%llu)",
fnvlist_lookup_string(events[i],
ZPOOL_HIST_DSNAME),
(u_longlong_t)fnvlist_lookup_uint64(
events[i],
ZPOOL_HIST_DSID));
}
(void) printf(" %s\n", fnvlist_lookup_string(events[i],
ZPOOL_HIST_INT_STR));
} else if (nvlist_exists(events[i], ZPOOL_HIST_IOCTL)) {
(void) printf("%s ioctl %s\n", tbuf,
fnvlist_lookup_string(events[i],
ZPOOL_HIST_IOCTL));
if (nvlist_exists(events[i], ZPOOL_HIST_INPUT_NVL)) {
(void) printf(" input:\n");
dump_nvlist(fnvlist_lookup_nvlist(events[i],
ZPOOL_HIST_INPUT_NVL), 8);
}
if (nvlist_exists(events[i], ZPOOL_HIST_OUTPUT_NVL)) {
(void) printf(" output:\n");
dump_nvlist(fnvlist_lookup_nvlist(events[i],
ZPOOL_HIST_OUTPUT_NVL), 8);
}
if (nvlist_exists(events[i], ZPOOL_HIST_ERRNO)) {
(void) printf(" errno: %lld\n",
(longlong_t)fnvlist_lookup_int64(events[i],
ZPOOL_HIST_ERRNO));
}
} else {
goto next;
}
printed = B_TRUE;
next:
if (dump_opt['h'] > 1) {
if (!printed)
(void) printf("unrecognized record:\n");
dump_nvlist(events[i], 2);
}
}
free(buf);
}
static void
dump_dnode(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
}
static uint64_t
blkid2offset(const dnode_phys_t *dnp, const blkptr_t *bp,
const zbookmark_phys_t *zb)
{
if (dnp == NULL) {
ASSERT(zb->zb_level < 0);
if (zb->zb_object == 0)
return (zb->zb_blkid);
return (zb->zb_blkid * BP_GET_LSIZE(bp));
}
ASSERT(zb->zb_level >= 0);
return ((zb->zb_blkid <<
(zb->zb_level * (dnp->dn_indblkshift - SPA_BLKPTRSHIFT))) *
dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT);
}
static void
snprintf_zstd_header(spa_t *spa, char *blkbuf, size_t buflen,
const blkptr_t *bp)
{
abd_t *pabd;
void *buf;
zio_t *zio;
zfs_zstdhdr_t zstd_hdr;
int error;
if (BP_GET_COMPRESS(bp) != ZIO_COMPRESS_ZSTD)
return;
if (BP_IS_HOLE(bp))
return;
if (BP_IS_EMBEDDED(bp)) {
buf = malloc(SPA_MAXBLOCKSIZE);
if (buf == NULL) {
(void) fprintf(stderr, "out of memory\n");
exit(1);
}
decode_embedded_bp_compressed(bp, buf);
memcpy(&zstd_hdr, buf, sizeof (zstd_hdr));
free(buf);
zstd_hdr.c_len = BE_32(zstd_hdr.c_len);
zstd_hdr.raw_version_level = BE_32(zstd_hdr.raw_version_level);
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf),
" ZSTD:size=%u:version=%u:level=%u:EMBEDDED",
zstd_hdr.c_len, zfs_get_hdrversion(&zstd_hdr),
zfs_get_hdrlevel(&zstd_hdr));
return;
}
pabd = abd_alloc_for_io(SPA_MAXBLOCKSIZE, B_FALSE);
zio = zio_root(spa, NULL, NULL, 0);
/* Decrypt but don't decompress so we can read the compression header */
zio_nowait(zio_read(zio, spa, bp, pabd, BP_GET_PSIZE(bp), NULL, NULL,
ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW_COMPRESS,
NULL));
error = zio_wait(zio);
if (error) {
(void) fprintf(stderr, "read failed: %d\n", error);
return;
}
buf = abd_borrow_buf_copy(pabd, BP_GET_LSIZE(bp));
memcpy(&zstd_hdr, buf, sizeof (zstd_hdr));
zstd_hdr.c_len = BE_32(zstd_hdr.c_len);
zstd_hdr.raw_version_level = BE_32(zstd_hdr.raw_version_level);
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf),
" ZSTD:size=%u:version=%u:level=%u:NORMAL",
zstd_hdr.c_len, zfs_get_hdrversion(&zstd_hdr),
zfs_get_hdrlevel(&zstd_hdr));
abd_return_buf_copy(pabd, buf, BP_GET_LSIZE(bp));
}
static void
snprintf_blkptr_compact(char *blkbuf, size_t buflen, const blkptr_t *bp,
boolean_t bp_freed)
{
const dva_t *dva = bp->blk_dva;
int ndvas = dump_opt['d'] > 5 ? BP_GET_NDVAS(bp) : 1;
int i;
if (dump_opt['b'] >= 6) {
snprintf_blkptr(blkbuf, buflen, bp);
if (bp_freed) {
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf), " %s", "FREE");
}
return;
}
if (BP_IS_EMBEDDED(bp)) {
(void) sprintf(blkbuf,
"EMBEDDED et=%u %llxL/%llxP B=%llu",
(int)BPE_GET_ETYPE(bp),
(u_longlong_t)BPE_GET_LSIZE(bp),
(u_longlong_t)BPE_GET_PSIZE(bp),
(u_longlong_t)bp->blk_birth);
return;
}
blkbuf[0] = '\0';
for (i = 0; i < ndvas; i++)
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf), "%llu:%llx:%llx ",
(u_longlong_t)DVA_GET_VDEV(&dva[i]),
(u_longlong_t)DVA_GET_OFFSET(&dva[i]),
(u_longlong_t)DVA_GET_ASIZE(&dva[i]));
if (BP_IS_HOLE(bp)) {
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf),
"%llxL B=%llu",
(u_longlong_t)BP_GET_LSIZE(bp),
(u_longlong_t)bp->blk_birth);
} else {
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf),
"%llxL/%llxP F=%llu B=%llu/%llu",
(u_longlong_t)BP_GET_LSIZE(bp),
(u_longlong_t)BP_GET_PSIZE(bp),
(u_longlong_t)BP_GET_FILL(bp),
(u_longlong_t)bp->blk_birth,
(u_longlong_t)BP_PHYSICAL_BIRTH(bp));
if (bp_freed)
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf), " %s", "FREE");
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf),
" cksum=%016llx:%016llx:%016llx:%016llx",
(u_longlong_t)bp->blk_cksum.zc_word[0],
(u_longlong_t)bp->blk_cksum.zc_word[1],
(u_longlong_t)bp->blk_cksum.zc_word[2],
(u_longlong_t)bp->blk_cksum.zc_word[3]);
}
}
static void
print_indirect(spa_t *spa, blkptr_t *bp, const zbookmark_phys_t *zb,
const dnode_phys_t *dnp)
{
char blkbuf[BP_SPRINTF_LEN];
int l;
if (!BP_IS_EMBEDDED(bp)) {
ASSERT3U(BP_GET_TYPE(bp), ==, dnp->dn_type);
ASSERT3U(BP_GET_LEVEL(bp), ==, zb->zb_level);
}
(void) printf("%16llx ", (u_longlong_t)blkid2offset(dnp, bp, zb));
ASSERT(zb->zb_level >= 0);
for (l = dnp->dn_nlevels - 1; l >= -1; l--) {
if (l == zb->zb_level) {
(void) printf("L%llx", (u_longlong_t)zb->zb_level);
} else {
(void) printf(" ");
}
}
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf), bp, B_FALSE);
if (dump_opt['Z'] && BP_GET_COMPRESS(bp) == ZIO_COMPRESS_ZSTD)
snprintf_zstd_header(spa, blkbuf, sizeof (blkbuf), bp);
(void) printf("%s\n", blkbuf);
}
static int
visit_indirect(spa_t *spa, const dnode_phys_t *dnp,
blkptr_t *bp, const zbookmark_phys_t *zb)
{
int err = 0;
if (bp->blk_birth == 0)
return (0);
print_indirect(spa, bp, zb, dnp);
if (BP_GET_LEVEL(bp) > 0 && !BP_IS_HOLE(bp)) {
arc_flags_t flags = ARC_FLAG_WAIT;
int i;
blkptr_t *cbp;
int epb = BP_GET_LSIZE(bp) >> SPA_BLKPTRSHIFT;
arc_buf_t *buf;
uint64_t fill = 0;
ASSERT(!BP_IS_REDACTED(bp));
err = arc_read(NULL, spa, bp, arc_getbuf_func, &buf,
ZIO_PRIORITY_ASYNC_READ, ZIO_FLAG_CANFAIL, &flags, zb);
if (err)
return (err);
ASSERT(buf->b_data);
/* recursively visit blocks below this */
cbp = buf->b_data;
for (i = 0; i < epb; i++, cbp++) {
zbookmark_phys_t czb;
SET_BOOKMARK(&czb, zb->zb_objset, zb->zb_object,
zb->zb_level - 1,
zb->zb_blkid * epb + i);
err = visit_indirect(spa, dnp, cbp, &czb);
if (err)
break;
fill += BP_GET_FILL(cbp);
}
if (!err)
ASSERT3U(fill, ==, BP_GET_FILL(bp));
arc_buf_destroy(buf, &buf);
}
return (err);
}
static void
dump_indirect(dnode_t *dn)
{
dnode_phys_t *dnp = dn->dn_phys;
zbookmark_phys_t czb;
(void) printf("Indirect blocks:\n");
SET_BOOKMARK(&czb, dmu_objset_id(dn->dn_objset),
dn->dn_object, dnp->dn_nlevels - 1, 0);
for (int j = 0; j < dnp->dn_nblkptr; j++) {
czb.zb_blkid = j;
(void) visit_indirect(dmu_objset_spa(dn->dn_objset), dnp,
&dnp->dn_blkptr[j], &czb);
}
(void) printf("\n");
}
static void
dump_dsl_dir(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object;
dsl_dir_phys_t *dd = data;
time_t crtime;
char nice[32];
/* make sure nicenum has enough space */
_Static_assert(sizeof (nice) >= NN_NUMBUF_SZ, "nice truncated");
if (dd == NULL)
return;
ASSERT3U(size, >=, sizeof (dsl_dir_phys_t));
crtime = dd->dd_creation_time;
(void) printf("\t\tcreation_time = %s", ctime(&crtime));
(void) printf("\t\thead_dataset_obj = %llu\n",
(u_longlong_t)dd->dd_head_dataset_obj);
(void) printf("\t\tparent_dir_obj = %llu\n",
(u_longlong_t)dd->dd_parent_obj);
(void) printf("\t\torigin_obj = %llu\n",
(u_longlong_t)dd->dd_origin_obj);
(void) printf("\t\tchild_dir_zapobj = %llu\n",
(u_longlong_t)dd->dd_child_dir_zapobj);
zdb_nicenum(dd->dd_used_bytes, nice, sizeof (nice));
(void) printf("\t\tused_bytes = %s\n", nice);
zdb_nicenum(dd->dd_compressed_bytes, nice, sizeof (nice));
(void) printf("\t\tcompressed_bytes = %s\n", nice);
zdb_nicenum(dd->dd_uncompressed_bytes, nice, sizeof (nice));
(void) printf("\t\tuncompressed_bytes = %s\n", nice);
zdb_nicenum(dd->dd_quota, nice, sizeof (nice));
(void) printf("\t\tquota = %s\n", nice);
zdb_nicenum(dd->dd_reserved, nice, sizeof (nice));
(void) printf("\t\treserved = %s\n", nice);
(void) printf("\t\tprops_zapobj = %llu\n",
(u_longlong_t)dd->dd_props_zapobj);
(void) printf("\t\tdeleg_zapobj = %llu\n",
(u_longlong_t)dd->dd_deleg_zapobj);
(void) printf("\t\tflags = %llx\n",
(u_longlong_t)dd->dd_flags);
#define DO(which) \
zdb_nicenum(dd->dd_used_breakdown[DD_USED_ ## which], nice, \
sizeof (nice)); \
(void) printf("\t\tused_breakdown[" #which "] = %s\n", nice)
DO(HEAD);
DO(SNAP);
DO(CHILD);
DO(CHILD_RSRV);
DO(REFRSRV);
#undef DO
(void) printf("\t\tclones = %llu\n",
(u_longlong_t)dd->dd_clones);
}
static void
dump_dsl_dataset(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object;
dsl_dataset_phys_t *ds = data;
time_t crtime;
char used[32], compressed[32], uncompressed[32], unique[32];
char blkbuf[BP_SPRINTF_LEN];
/* make sure nicenum has enough space */
_Static_assert(sizeof (used) >= NN_NUMBUF_SZ, "used truncated");
_Static_assert(sizeof (compressed) >= NN_NUMBUF_SZ,
"compressed truncated");
_Static_assert(sizeof (uncompressed) >= NN_NUMBUF_SZ,
"uncompressed truncated");
_Static_assert(sizeof (unique) >= NN_NUMBUF_SZ, "unique truncated");
if (ds == NULL)
return;
ASSERT(size == sizeof (*ds));
crtime = ds->ds_creation_time;
zdb_nicenum(ds->ds_referenced_bytes, used, sizeof (used));
zdb_nicenum(ds->ds_compressed_bytes, compressed, sizeof (compressed));
zdb_nicenum(ds->ds_uncompressed_bytes, uncompressed,
sizeof (uncompressed));
zdb_nicenum(ds->ds_unique_bytes, unique, sizeof (unique));
snprintf_blkptr(blkbuf, sizeof (blkbuf), &ds->ds_bp);
(void) printf("\t\tdir_obj = %llu\n",
(u_longlong_t)ds->ds_dir_obj);
(void) printf("\t\tprev_snap_obj = %llu\n",
(u_longlong_t)ds->ds_prev_snap_obj);
(void) printf("\t\tprev_snap_txg = %llu\n",
(u_longlong_t)ds->ds_prev_snap_txg);
(void) printf("\t\tnext_snap_obj = %llu\n",
(u_longlong_t)ds->ds_next_snap_obj);
(void) printf("\t\tsnapnames_zapobj = %llu\n",
(u_longlong_t)ds->ds_snapnames_zapobj);
(void) printf("\t\tnum_children = %llu\n",
(u_longlong_t)ds->ds_num_children);
(void) printf("\t\tuserrefs_obj = %llu\n",
(u_longlong_t)ds->ds_userrefs_obj);
(void) printf("\t\tcreation_time = %s", ctime(&crtime));
(void) printf("\t\tcreation_txg = %llu\n",
(u_longlong_t)ds->ds_creation_txg);
(void) printf("\t\tdeadlist_obj = %llu\n",
(u_longlong_t)ds->ds_deadlist_obj);
(void) printf("\t\tused_bytes = %s\n", used);
(void) printf("\t\tcompressed_bytes = %s\n", compressed);
(void) printf("\t\tuncompressed_bytes = %s\n", uncompressed);
(void) printf("\t\tunique = %s\n", unique);
(void) printf("\t\tfsid_guid = %llu\n",
(u_longlong_t)ds->ds_fsid_guid);
(void) printf("\t\tguid = %llu\n",
(u_longlong_t)ds->ds_guid);
(void) printf("\t\tflags = %llx\n",
(u_longlong_t)ds->ds_flags);
(void) printf("\t\tnext_clones_obj = %llu\n",
(u_longlong_t)ds->ds_next_clones_obj);
(void) printf("\t\tprops_obj = %llu\n",
(u_longlong_t)ds->ds_props_obj);
(void) printf("\t\tbp = %s\n", blkbuf);
}
static int
dump_bptree_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
(void) arg, (void) tx;
char blkbuf[BP_SPRINTF_LEN];
if (bp->blk_birth != 0) {
snprintf_blkptr(blkbuf, sizeof (blkbuf), bp);
(void) printf("\t%s\n", blkbuf);
}
return (0);
}
static void
dump_bptree(objset_t *os, uint64_t obj, const char *name)
{
char bytes[32];
bptree_phys_t *bt;
dmu_buf_t *db;
/* make sure nicenum has enough space */
_Static_assert(sizeof (bytes) >= NN_NUMBUF_SZ, "bytes truncated");
if (dump_opt['d'] < 3)
return;
VERIFY3U(0, ==, dmu_bonus_hold(os, obj, FTAG, &db));
bt = db->db_data;
zdb_nicenum(bt->bt_bytes, bytes, sizeof (bytes));
(void) printf("\n %s: %llu datasets, %s\n",
name, (unsigned long long)(bt->bt_end - bt->bt_begin), bytes);
dmu_buf_rele(db, FTAG);
if (dump_opt['d'] < 5)
return;
(void) printf("\n");
(void) bptree_iterate(os, obj, B_FALSE, dump_bptree_cb, NULL, NULL);
}
static int
dump_bpobj_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed, dmu_tx_t *tx)
{
(void) arg, (void) tx;
char blkbuf[BP_SPRINTF_LEN];
ASSERT(bp->blk_birth != 0);
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf), bp, bp_freed);
(void) printf("\t%s\n", blkbuf);
return (0);
}
static void
dump_full_bpobj(bpobj_t *bpo, const char *name, int indent)
{
char bytes[32];
char comp[32];
char uncomp[32];
uint64_t i;
/* make sure nicenum has enough space */
_Static_assert(sizeof (bytes) >= NN_NUMBUF_SZ, "bytes truncated");
_Static_assert(sizeof (comp) >= NN_NUMBUF_SZ, "comp truncated");
_Static_assert(sizeof (uncomp) >= NN_NUMBUF_SZ, "uncomp truncated");
if (dump_opt['d'] < 3)
return;
zdb_nicenum(bpo->bpo_phys->bpo_bytes, bytes, sizeof (bytes));
if (bpo->bpo_havesubobj && bpo->bpo_phys->bpo_subobjs != 0) {
zdb_nicenum(bpo->bpo_phys->bpo_comp, comp, sizeof (comp));
zdb_nicenum(bpo->bpo_phys->bpo_uncomp, uncomp, sizeof (uncomp));
if (bpo->bpo_havefreed) {
(void) printf(" %*s: object %llu, %llu local "
"blkptrs, %llu freed, %llu subobjs in object %llu, "
"%s (%s/%s comp)\n",
indent * 8, name,
(u_longlong_t)bpo->bpo_object,
(u_longlong_t)bpo->bpo_phys->bpo_num_blkptrs,
(u_longlong_t)bpo->bpo_phys->bpo_num_freed,
(u_longlong_t)bpo->bpo_phys->bpo_num_subobjs,
(u_longlong_t)bpo->bpo_phys->bpo_subobjs,
bytes, comp, uncomp);
} else {
(void) printf(" %*s: object %llu, %llu local "
"blkptrs, %llu subobjs in object %llu, "
"%s (%s/%s comp)\n",
indent * 8, name,
(u_longlong_t)bpo->bpo_object,
(u_longlong_t)bpo->bpo_phys->bpo_num_blkptrs,
(u_longlong_t)bpo->bpo_phys->bpo_num_subobjs,
(u_longlong_t)bpo->bpo_phys->bpo_subobjs,
bytes, comp, uncomp);
}
for (i = 0; i < bpo->bpo_phys->bpo_num_subobjs; i++) {
uint64_t subobj;
bpobj_t subbpo;
int error;
VERIFY0(dmu_read(bpo->bpo_os,
bpo->bpo_phys->bpo_subobjs,
i * sizeof (subobj), sizeof (subobj), &subobj, 0));
error = bpobj_open(&subbpo, bpo->bpo_os, subobj);
if (error != 0) {
(void) printf("ERROR %u while trying to open "
"subobj id %llu\n",
error, (u_longlong_t)subobj);
continue;
}
dump_full_bpobj(&subbpo, "subobj", indent + 1);
bpobj_close(&subbpo);
}
} else {
if (bpo->bpo_havefreed) {
(void) printf(" %*s: object %llu, %llu blkptrs, "
"%llu freed, %s\n",
indent * 8, name,
(u_longlong_t)bpo->bpo_object,
(u_longlong_t)bpo->bpo_phys->bpo_num_blkptrs,
(u_longlong_t)bpo->bpo_phys->bpo_num_freed,
bytes);
} else {
(void) printf(" %*s: object %llu, %llu blkptrs, "
"%s\n",
indent * 8, name,
(u_longlong_t)bpo->bpo_object,
(u_longlong_t)bpo->bpo_phys->bpo_num_blkptrs,
bytes);
}
}
if (dump_opt['d'] < 5)
return;
if (indent == 0) {
(void) bpobj_iterate_nofree(bpo, dump_bpobj_cb, NULL, NULL);
(void) printf("\n");
}
}
static int
dump_bookmark(dsl_pool_t *dp, char *name, boolean_t print_redact,
boolean_t print_list)
{
int err = 0;
zfs_bookmark_phys_t prop;
objset_t *mos = dp->dp_spa->spa_meta_objset;
err = dsl_bookmark_lookup(dp, name, NULL, &prop);
if (err != 0) {
return (err);
}
(void) printf("\t#%s: ", strchr(name, '#') + 1);
(void) printf("{guid: %llx creation_txg: %llu creation_time: "
"%llu redaction_obj: %llu}\n", (u_longlong_t)prop.zbm_guid,
(u_longlong_t)prop.zbm_creation_txg,
(u_longlong_t)prop.zbm_creation_time,
(u_longlong_t)prop.zbm_redaction_obj);
IMPLY(print_list, print_redact);
if (!print_redact || prop.zbm_redaction_obj == 0)
return (0);
redaction_list_t *rl;
VERIFY0(dsl_redaction_list_hold_obj(dp,
prop.zbm_redaction_obj, FTAG, &rl));
redaction_list_phys_t *rlp = rl->rl_phys;
(void) printf("\tRedacted:\n\t\tProgress: ");
if (rlp->rlp_last_object != UINT64_MAX ||
rlp->rlp_last_blkid != UINT64_MAX) {
(void) printf("%llu %llu (incomplete)\n",
(u_longlong_t)rlp->rlp_last_object,
(u_longlong_t)rlp->rlp_last_blkid);
} else {
(void) printf("complete\n");
}
(void) printf("\t\tSnapshots: [");
for (unsigned int i = 0; i < rlp->rlp_num_snaps; i++) {
if (i > 0)
(void) printf(", ");
(void) printf("%0llu",
(u_longlong_t)rlp->rlp_snaps[i]);
}
(void) printf("]\n\t\tLength: %llu\n",
(u_longlong_t)rlp->rlp_num_entries);
if (!print_list) {
dsl_redaction_list_rele(rl, FTAG);
return (0);
}
if (rlp->rlp_num_entries == 0) {
dsl_redaction_list_rele(rl, FTAG);
(void) printf("\t\tRedaction List: []\n\n");
return (0);
}
redact_block_phys_t *rbp_buf;
uint64_t size;
dmu_object_info_t doi;
VERIFY0(dmu_object_info(mos, prop.zbm_redaction_obj, &doi));
size = doi.doi_max_offset;
rbp_buf = kmem_alloc(size, KM_SLEEP);
err = dmu_read(mos, prop.zbm_redaction_obj, 0, size,
rbp_buf, 0);
if (err != 0) {
dsl_redaction_list_rele(rl, FTAG);
kmem_free(rbp_buf, size);
return (err);
}
(void) printf("\t\tRedaction List: [{object: %llx, offset: "
"%llx, blksz: %x, count: %llx}",
(u_longlong_t)rbp_buf[0].rbp_object,
(u_longlong_t)rbp_buf[0].rbp_blkid,
(uint_t)(redact_block_get_size(&rbp_buf[0])),
(u_longlong_t)redact_block_get_count(&rbp_buf[0]));
for (size_t i = 1; i < rlp->rlp_num_entries; i++) {
(void) printf(",\n\t\t{object: %llx, offset: %llx, "
"blksz: %x, count: %llx}",
(u_longlong_t)rbp_buf[i].rbp_object,
(u_longlong_t)rbp_buf[i].rbp_blkid,
(uint_t)(redact_block_get_size(&rbp_buf[i])),
(u_longlong_t)redact_block_get_count(&rbp_buf[i]));
}
dsl_redaction_list_rele(rl, FTAG);
kmem_free(rbp_buf, size);
(void) printf("]\n\n");
return (0);
}
static void
dump_bookmarks(objset_t *os, int verbosity)
{
zap_cursor_t zc;
zap_attribute_t attr;
dsl_dataset_t *ds = dmu_objset_ds(os);
dsl_pool_t *dp = spa_get_dsl(os->os_spa);
objset_t *mos = os->os_spa->spa_meta_objset;
if (verbosity < 4)
return;
dsl_pool_config_enter(dp, FTAG);
for (zap_cursor_init(&zc, mos, ds->ds_bookmarks_obj);
zap_cursor_retrieve(&zc, &attr) == 0;
zap_cursor_advance(&zc)) {
char osname[ZFS_MAX_DATASET_NAME_LEN];
char buf[ZFS_MAX_DATASET_NAME_LEN];
int len;
dmu_objset_name(os, osname);
len = snprintf(buf, sizeof (buf), "%s#%s", osname,
attr.za_name);
VERIFY3S(len, <, ZFS_MAX_DATASET_NAME_LEN);
(void) dump_bookmark(dp, buf, verbosity >= 5, verbosity >= 6);
}
zap_cursor_fini(&zc);
dsl_pool_config_exit(dp, FTAG);
}
static void
bpobj_count_refd(bpobj_t *bpo)
{
mos_obj_refd(bpo->bpo_object);
if (bpo->bpo_havesubobj && bpo->bpo_phys->bpo_subobjs != 0) {
mos_obj_refd(bpo->bpo_phys->bpo_subobjs);
for (uint64_t i = 0; i < bpo->bpo_phys->bpo_num_subobjs; i++) {
uint64_t subobj;
bpobj_t subbpo;
int error;
VERIFY0(dmu_read(bpo->bpo_os,
bpo->bpo_phys->bpo_subobjs,
i * sizeof (subobj), sizeof (subobj), &subobj, 0));
error = bpobj_open(&subbpo, bpo->bpo_os, subobj);
if (error != 0) {
(void) printf("ERROR %u while trying to open "
"subobj id %llu\n",
error, (u_longlong_t)subobj);
continue;
}
bpobj_count_refd(&subbpo);
bpobj_close(&subbpo);
}
}
}
static int
dsl_deadlist_entry_count_refd(void *arg, dsl_deadlist_entry_t *dle)
{
spa_t *spa = arg;
uint64_t empty_bpobj = spa->spa_dsl_pool->dp_empty_bpobj;
if (dle->dle_bpobj.bpo_object != empty_bpobj)
bpobj_count_refd(&dle->dle_bpobj);
return (0);
}
static int
dsl_deadlist_entry_dump(void *arg, dsl_deadlist_entry_t *dle)
{
ASSERT(arg == NULL);
if (dump_opt['d'] >= 5) {
char buf[128];
(void) snprintf(buf, sizeof (buf),
"mintxg %llu -> obj %llu",
(longlong_t)dle->dle_mintxg,
(longlong_t)dle->dle_bpobj.bpo_object);
dump_full_bpobj(&dle->dle_bpobj, buf, 0);
} else {
(void) printf("mintxg %llu -> obj %llu\n",
(longlong_t)dle->dle_mintxg,
(longlong_t)dle->dle_bpobj.bpo_object);
}
return (0);
}
static void
dump_blkptr_list(dsl_deadlist_t *dl, const char *name)
{
char bytes[32];
char comp[32];
char uncomp[32];
char entries[32];
spa_t *spa = dmu_objset_spa(dl->dl_os);
uint64_t empty_bpobj = spa->spa_dsl_pool->dp_empty_bpobj;
if (dl->dl_oldfmt) {
if (dl->dl_bpobj.bpo_object != empty_bpobj)
bpobj_count_refd(&dl->dl_bpobj);
} else {
mos_obj_refd(dl->dl_object);
dsl_deadlist_iterate(dl, dsl_deadlist_entry_count_refd, spa);
}
/* make sure nicenum has enough space */
_Static_assert(sizeof (bytes) >= NN_NUMBUF_SZ, "bytes truncated");
_Static_assert(sizeof (comp) >= NN_NUMBUF_SZ, "comp truncated");
_Static_assert(sizeof (uncomp) >= NN_NUMBUF_SZ, "uncomp truncated");
_Static_assert(sizeof (entries) >= NN_NUMBUF_SZ, "entries truncated");
if (dump_opt['d'] < 3)
return;
if (dl->dl_oldfmt) {
dump_full_bpobj(&dl->dl_bpobj, "old-format deadlist", 0);
return;
}
zdb_nicenum(dl->dl_phys->dl_used, bytes, sizeof (bytes));
zdb_nicenum(dl->dl_phys->dl_comp, comp, sizeof (comp));
zdb_nicenum(dl->dl_phys->dl_uncomp, uncomp, sizeof (uncomp));
zdb_nicenum(avl_numnodes(&dl->dl_tree), entries, sizeof (entries));
(void) printf("\n %s: %s (%s/%s comp), %s entries\n",
name, bytes, comp, uncomp, entries);
if (dump_opt['d'] < 4)
return;
(void) putchar('\n');
dsl_deadlist_iterate(dl, dsl_deadlist_entry_dump, NULL);
}
static int
verify_dd_livelist(objset_t *os)
{
uint64_t ll_used, used, ll_comp, comp, ll_uncomp, uncomp;
dsl_pool_t *dp = spa_get_dsl(os->os_spa);
dsl_dir_t *dd = os->os_dsl_dataset->ds_dir;
ASSERT(!dmu_objset_is_snapshot(os));
if (!dsl_deadlist_is_open(&dd->dd_livelist))
return (0);
/* Iterate through the livelist to check for duplicates */
dsl_deadlist_iterate(&dd->dd_livelist, sublivelist_verify_lightweight,
NULL);
dsl_pool_config_enter(dp, FTAG);
dsl_deadlist_space(&dd->dd_livelist, &ll_used,
&ll_comp, &ll_uncomp);
dsl_dataset_t *origin_ds;
ASSERT(dsl_pool_config_held(dp));
VERIFY0(dsl_dataset_hold_obj(dp,
dsl_dir_phys(dd)->dd_origin_obj, FTAG, &origin_ds));
VERIFY0(dsl_dataset_space_written(origin_ds, os->os_dsl_dataset,
&used, &comp, &uncomp));
dsl_dataset_rele(origin_ds, FTAG);
dsl_pool_config_exit(dp, FTAG);
/*
* It's possible that the dataset's uncomp space is larger than the
* livelist's because livelists do not track embedded block pointers
*/
if (used != ll_used || comp != ll_comp || uncomp < ll_uncomp) {
char nice_used[32], nice_comp[32], nice_uncomp[32];
(void) printf("Discrepancy in space accounting:\n");
zdb_nicenum(used, nice_used, sizeof (nice_used));
zdb_nicenum(comp, nice_comp, sizeof (nice_comp));
zdb_nicenum(uncomp, nice_uncomp, sizeof (nice_uncomp));
(void) printf("dir: used %s, comp %s, uncomp %s\n",
nice_used, nice_comp, nice_uncomp);
zdb_nicenum(ll_used, nice_used, sizeof (nice_used));
zdb_nicenum(ll_comp, nice_comp, sizeof (nice_comp));
zdb_nicenum(ll_uncomp, nice_uncomp, sizeof (nice_uncomp));
(void) printf("livelist: used %s, comp %s, uncomp %s\n",
nice_used, nice_comp, nice_uncomp);
return (1);
}
return (0);
}
static char *key_material = NULL;
static boolean_t
zdb_derive_key(dsl_dir_t *dd, uint8_t *key_out)
{
uint64_t keyformat, salt, iters;
int i;
unsigned char c;
VERIFY0(zap_lookup(dd->dd_pool->dp_meta_objset, dd->dd_crypto_obj,
zfs_prop_to_name(ZFS_PROP_KEYFORMAT), sizeof (uint64_t),
1, &keyformat));
switch (keyformat) {
case ZFS_KEYFORMAT_HEX:
for (i = 0; i < WRAPPING_KEY_LEN * 2; i += 2) {
if (!isxdigit(key_material[i]) ||
!isxdigit(key_material[i+1]))
return (B_FALSE);
if (sscanf(&key_material[i], "%02hhx", &c) != 1)
return (B_FALSE);
key_out[i / 2] = c;
}
break;
case ZFS_KEYFORMAT_PASSPHRASE:
VERIFY0(zap_lookup(dd->dd_pool->dp_meta_objset,
dd->dd_crypto_obj, zfs_prop_to_name(ZFS_PROP_PBKDF2_SALT),
sizeof (uint64_t), 1, &salt));
VERIFY0(zap_lookup(dd->dd_pool->dp_meta_objset,
dd->dd_crypto_obj, zfs_prop_to_name(ZFS_PROP_PBKDF2_ITERS),
sizeof (uint64_t), 1, &iters));
if (PKCS5_PBKDF2_HMAC_SHA1(key_material, strlen(key_material),
((uint8_t *)&salt), sizeof (uint64_t), iters,
WRAPPING_KEY_LEN, key_out) != 1)
return (B_FALSE);
break;
default:
fatal("no support for key format %u\n",
(unsigned int) keyformat);
}
return (B_TRUE);
}
static char encroot[ZFS_MAX_DATASET_NAME_LEN];
static boolean_t key_loaded = B_FALSE;
static void
zdb_load_key(objset_t *os)
{
dsl_pool_t *dp;
dsl_dir_t *dd, *rdd;
uint8_t key[WRAPPING_KEY_LEN];
uint64_t rddobj;
int err;
dp = spa_get_dsl(os->os_spa);
dd = os->os_dsl_dataset->ds_dir;
dsl_pool_config_enter(dp, FTAG);
VERIFY0(zap_lookup(dd->dd_pool->dp_meta_objset, dd->dd_crypto_obj,
DSL_CRYPTO_KEY_ROOT_DDOBJ, sizeof (uint64_t), 1, &rddobj));
VERIFY0(dsl_dir_hold_obj(dd->dd_pool, rddobj, NULL, FTAG, &rdd));
dsl_dir_name(rdd, encroot);
dsl_dir_rele(rdd, FTAG);
if (!zdb_derive_key(dd, key))
fatal("couldn't derive encryption key");
dsl_pool_config_exit(dp, FTAG);
ASSERT3U(dsl_dataset_get_keystatus(dd), ==, ZFS_KEYSTATUS_UNAVAILABLE);
dsl_crypto_params_t *dcp;
nvlist_t *crypto_args;
crypto_args = fnvlist_alloc();
fnvlist_add_uint8_array(crypto_args, "wkeydata",
(uint8_t *)key, WRAPPING_KEY_LEN);
VERIFY0(dsl_crypto_params_create_nvlist(DCP_CMD_NONE,
NULL, crypto_args, &dcp));
err = spa_keystore_load_wkey(encroot, dcp, B_FALSE);
dsl_crypto_params_free(dcp, (err != 0));
fnvlist_free(crypto_args);
if (err != 0)
fatal(
"couldn't load encryption key for %s: %s",
encroot, err == ZFS_ERR_CRYPTO_NOTSUP ?
"crypto params not supported" : strerror(err));
ASSERT3U(dsl_dataset_get_keystatus(dd), ==, ZFS_KEYSTATUS_AVAILABLE);
printf("Unlocked encryption root: %s\n", encroot);
key_loaded = B_TRUE;
}
static void
zdb_unload_key(void)
{
if (!key_loaded)
return;
VERIFY0(spa_keystore_unload_wkey(encroot));
key_loaded = B_FALSE;
}
static avl_tree_t idx_tree;
static avl_tree_t domain_tree;
static boolean_t fuid_table_loaded;
static objset_t *sa_os = NULL;
static sa_attr_type_t *sa_attr_table = NULL;
static int
open_objset(const char *path, const void *tag, objset_t **osp)
{
int err;
uint64_t sa_attrs = 0;
uint64_t version = 0;
VERIFY3P(sa_os, ==, NULL);
/*
* We can't own an objset if it's redacted. Therefore, we do this
* dance: hold the objset, then acquire a long hold on its dataset, then
* release the pool (which is held as part of holding the objset).
*/
if (dump_opt['K']) {
/* decryption requested, try to load keys */
err = dmu_objset_hold(path, tag, osp);
if (err != 0) {
(void) fprintf(stderr, "failed to hold dataset "
"'%s': %s\n",
path, strerror(err));
return (err);
}
dsl_dataset_long_hold(dmu_objset_ds(*osp), tag);
dsl_pool_rele(dmu_objset_pool(*osp), tag);
/* succeeds or dies */
zdb_load_key(*osp);
/* release it all */
dsl_dataset_long_rele(dmu_objset_ds(*osp), tag);
dsl_dataset_rele(dmu_objset_ds(*osp), tag);
}
int ds_hold_flags = key_loaded ? DS_HOLD_FLAG_DECRYPT : 0;
err = dmu_objset_hold_flags(path, ds_hold_flags, tag, osp);
if (err != 0) {
(void) fprintf(stderr, "failed to hold dataset '%s': %s\n",
path, strerror(err));
return (err);
}
dsl_dataset_long_hold(dmu_objset_ds(*osp), tag);
dsl_pool_rele(dmu_objset_pool(*osp), tag);
if (dmu_objset_type(*osp) == DMU_OST_ZFS &&
(key_loaded || !(*osp)->os_encrypted)) {
(void) zap_lookup(*osp, MASTER_NODE_OBJ, ZPL_VERSION_STR,
8, 1, &version);
if (version >= ZPL_VERSION_SA) {
(void) zap_lookup(*osp, MASTER_NODE_OBJ, ZFS_SA_ATTRS,
8, 1, &sa_attrs);
}
err = sa_setup(*osp, sa_attrs, zfs_attr_table, ZPL_END,
&sa_attr_table);
if (err != 0) {
(void) fprintf(stderr, "sa_setup failed: %s\n",
strerror(err));
dsl_dataset_long_rele(dmu_objset_ds(*osp), tag);
dsl_dataset_rele_flags(dmu_objset_ds(*osp),
ds_hold_flags, tag);
*osp = NULL;
}
}
sa_os = *osp;
return (err);
}
static void
close_objset(objset_t *os, const void *tag)
{
VERIFY3P(os, ==, sa_os);
if (os->os_sa != NULL)
sa_tear_down(os);
dsl_dataset_long_rele(dmu_objset_ds(os), tag);
dsl_dataset_rele_flags(dmu_objset_ds(os),
key_loaded ? DS_HOLD_FLAG_DECRYPT : 0, tag);
sa_attr_table = NULL;
sa_os = NULL;
zdb_unload_key();
}
static void
fuid_table_destroy(void)
{
if (fuid_table_loaded) {
zfs_fuid_table_destroy(&idx_tree, &domain_tree);
fuid_table_loaded = B_FALSE;
}
}
/*
* print uid or gid information.
* For normal POSIX id just the id is printed in decimal format.
* For CIFS files with FUID the fuid is printed in hex followed by
* the domain-rid string.
*/
static void
print_idstr(uint64_t id, const char *id_type)
{
if (FUID_INDEX(id)) {
const char *domain =
zfs_fuid_idx_domain(&idx_tree, FUID_INDEX(id));
(void) printf("\t%s %llx [%s-%d]\n", id_type,
(u_longlong_t)id, domain, (int)FUID_RID(id));
} else {
(void) printf("\t%s %llu\n", id_type, (u_longlong_t)id);
}
}
static void
dump_uidgid(objset_t *os, uint64_t uid, uint64_t gid)
{
uint32_t uid_idx, gid_idx;
uid_idx = FUID_INDEX(uid);
gid_idx = FUID_INDEX(gid);
/* Load domain table, if not already loaded */
if (!fuid_table_loaded && (uid_idx || gid_idx)) {
uint64_t fuid_obj;
/* first find the fuid object. It lives in the master node */
VERIFY(zap_lookup(os, MASTER_NODE_OBJ, ZFS_FUID_TABLES,
8, 1, &fuid_obj) == 0);
zfs_fuid_avl_tree_create(&idx_tree, &domain_tree);
(void) zfs_fuid_table_load(os, fuid_obj,
&idx_tree, &domain_tree);
fuid_table_loaded = B_TRUE;
}
print_idstr(uid, "uid");
print_idstr(gid, "gid");
}
static void
dump_znode_sa_xattr(sa_handle_t *hdl)
{
nvlist_t *sa_xattr;
nvpair_t *elem = NULL;
int sa_xattr_size = 0;
int sa_xattr_entries = 0;
int error;
char *sa_xattr_packed;
error = sa_size(hdl, sa_attr_table[ZPL_DXATTR], &sa_xattr_size);
if (error || sa_xattr_size == 0)
return;
sa_xattr_packed = malloc(sa_xattr_size);
if (sa_xattr_packed == NULL)
return;
error = sa_lookup(hdl, sa_attr_table[ZPL_DXATTR],
sa_xattr_packed, sa_xattr_size);
if (error) {
free(sa_xattr_packed);
return;
}
error = nvlist_unpack(sa_xattr_packed, sa_xattr_size, &sa_xattr, 0);
if (error) {
free(sa_xattr_packed);
return;
}
while ((elem = nvlist_next_nvpair(sa_xattr, elem)) != NULL)
sa_xattr_entries++;
(void) printf("\tSA xattrs: %d bytes, %d entries\n\n",
sa_xattr_size, sa_xattr_entries);
while ((elem = nvlist_next_nvpair(sa_xattr, elem)) != NULL) {
boolean_t can_print = !dump_opt['P'];
uchar_t *value;
uint_t cnt, idx;
(void) printf("\t\t%s = ", nvpair_name(elem));
nvpair_value_byte_array(elem, &value, &cnt);
for (idx = 0; idx < cnt; ++idx) {
if (!isprint(value[idx])) {
can_print = B_FALSE;
break;
}
}
for (idx = 0; idx < cnt; ++idx) {
if (can_print)
(void) putchar(value[idx]);
else
(void) printf("\\%3.3o", value[idx]);
}
(void) putchar('\n');
}
nvlist_free(sa_xattr);
free(sa_xattr_packed);
}
static void
dump_znode_symlink(sa_handle_t *hdl)
{
int sa_symlink_size = 0;
char linktarget[MAXPATHLEN];
int error;
error = sa_size(hdl, sa_attr_table[ZPL_SYMLINK], &sa_symlink_size);
if (error || sa_symlink_size == 0) {
return;
}
if (sa_symlink_size >= sizeof (linktarget)) {
(void) printf("symlink size %d is too large\n",
sa_symlink_size);
return;
}
linktarget[sa_symlink_size] = '\0';
if (sa_lookup(hdl, sa_attr_table[ZPL_SYMLINK],
&linktarget, sa_symlink_size) == 0)
(void) printf("\ttarget %s\n", linktarget);
}
static void
dump_znode(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
char path[MAXPATHLEN * 2]; /* allow for xattr and failure prefix */
sa_handle_t *hdl;
uint64_t xattr, rdev, gen;
uint64_t uid, gid, mode, fsize, parent, links;
uint64_t pflags;
uint64_t acctm[2], modtm[2], chgtm[2], crtm[2];
time_t z_crtime, z_atime, z_mtime, z_ctime;
sa_bulk_attr_t bulk[12];
int idx = 0;
int error;
VERIFY3P(os, ==, sa_os);
if (sa_handle_get(os, object, NULL, SA_HDL_PRIVATE, &hdl)) {
(void) printf("Failed to get handle for SA znode\n");
return;
}
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_UID], NULL, &uid, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_GID], NULL, &gid, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_LINKS], NULL,
&links, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_GEN], NULL, &gen, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_MODE], NULL,
&mode, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_PARENT],
NULL, &parent, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_SIZE], NULL,
&fsize, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_ATIME], NULL,
acctm, 16);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_MTIME], NULL,
modtm, 16);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_CRTIME], NULL,
crtm, 16);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_CTIME], NULL,
chgtm, 16);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_FLAGS], NULL,
&pflags, 8);
if (sa_bulk_lookup(hdl, bulk, idx)) {
(void) sa_handle_destroy(hdl);
return;
}
z_crtime = (time_t)crtm[0];
z_atime = (time_t)acctm[0];
z_mtime = (time_t)modtm[0];
z_ctime = (time_t)chgtm[0];
if (dump_opt['d'] > 4) {
error = zfs_obj_to_path(os, object, path, sizeof (path));
if (error == ESTALE) {
(void) snprintf(path, sizeof (path), "on delete queue");
} else if (error != 0) {
leaked_objects++;
(void) snprintf(path, sizeof (path),
"path not found, possibly leaked");
}
(void) printf("\tpath %s\n", path);
}
if (S_ISLNK(mode))
dump_znode_symlink(hdl);
dump_uidgid(os, uid, gid);
(void) printf("\tatime %s", ctime(&z_atime));
(void) printf("\tmtime %s", ctime(&z_mtime));
(void) printf("\tctime %s", ctime(&z_ctime));
(void) printf("\tcrtime %s", ctime(&z_crtime));
(void) printf("\tgen %llu\n", (u_longlong_t)gen);
(void) printf("\tmode %llo\n", (u_longlong_t)mode);
(void) printf("\tsize %llu\n", (u_longlong_t)fsize);
(void) printf("\tparent %llu\n", (u_longlong_t)parent);
(void) printf("\tlinks %llu\n", (u_longlong_t)links);
(void) printf("\tpflags %llx\n", (u_longlong_t)pflags);
if (dmu_objset_projectquota_enabled(os) && (pflags & ZFS_PROJID)) {
uint64_t projid;
if (sa_lookup(hdl, sa_attr_table[ZPL_PROJID], &projid,
sizeof (uint64_t)) == 0)
(void) printf("\tprojid %llu\n", (u_longlong_t)projid);
}
if (sa_lookup(hdl, sa_attr_table[ZPL_XATTR], &xattr,
sizeof (uint64_t)) == 0)
(void) printf("\txattr %llu\n", (u_longlong_t)xattr);
if (sa_lookup(hdl, sa_attr_table[ZPL_RDEV], &rdev,
sizeof (uint64_t)) == 0)
(void) printf("\trdev 0x%016llx\n", (u_longlong_t)rdev);
dump_znode_sa_xattr(hdl);
sa_handle_destroy(hdl);
}
static void
dump_acl(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
}
static void
dump_dmu_objset(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
}
static object_viewer_t *object_viewer[DMU_OT_NUMTYPES + 1] = {
dump_none, /* unallocated */
dump_zap, /* object directory */
dump_uint64, /* object array */
dump_none, /* packed nvlist */
dump_packed_nvlist, /* packed nvlist size */
dump_none, /* bpobj */
dump_bpobj, /* bpobj header */
dump_none, /* SPA space map header */
dump_none, /* SPA space map */
dump_none, /* ZIL intent log */
dump_dnode, /* DMU dnode */
dump_dmu_objset, /* DMU objset */
dump_dsl_dir, /* DSL directory */
dump_zap, /* DSL directory child map */
dump_zap, /* DSL dataset snap map */
dump_zap, /* DSL props */
dump_dsl_dataset, /* DSL dataset */
dump_znode, /* ZFS znode */
dump_acl, /* ZFS V0 ACL */
dump_uint8, /* ZFS plain file */
dump_zpldir, /* ZFS directory */
dump_zap, /* ZFS master node */
dump_zap, /* ZFS delete queue */
dump_uint8, /* zvol object */
dump_zap, /* zvol prop */
dump_uint8, /* other uint8[] */
dump_uint64, /* other uint64[] */
dump_zap, /* other ZAP */
dump_zap, /* persistent error log */
dump_uint8, /* SPA history */
dump_history_offsets, /* SPA history offsets */
dump_zap, /* Pool properties */
dump_zap, /* DSL permissions */
dump_acl, /* ZFS ACL */
dump_uint8, /* ZFS SYSACL */
dump_none, /* FUID nvlist */
dump_packed_nvlist, /* FUID nvlist size */
dump_zap, /* DSL dataset next clones */
dump_zap, /* DSL scrub queue */
dump_zap, /* ZFS user/group/project used */
dump_zap, /* ZFS user/group/project quota */
dump_zap, /* snapshot refcount tags */
dump_ddt_zap, /* DDT ZAP object */
dump_zap, /* DDT statistics */
dump_znode, /* SA object */
dump_zap, /* SA Master Node */
dump_sa_attrs, /* SA attribute registration */
dump_sa_layouts, /* SA attribute layouts */
dump_zap, /* DSL scrub translations */
dump_none, /* fake dedup BP */
dump_zap, /* deadlist */
dump_none, /* deadlist hdr */
dump_zap, /* dsl clones */
dump_bpobj_subobjs, /* bpobj subobjs */
dump_unknown, /* Unknown type, must be last */
};
static boolean_t
match_object_type(dmu_object_type_t obj_type, uint64_t flags)
{
boolean_t match = B_TRUE;
switch (obj_type) {
case DMU_OT_DIRECTORY_CONTENTS:
if (!(flags & ZOR_FLAG_DIRECTORY))
match = B_FALSE;
break;
case DMU_OT_PLAIN_FILE_CONTENTS:
if (!(flags & ZOR_FLAG_PLAIN_FILE))
match = B_FALSE;
break;
case DMU_OT_SPACE_MAP:
if (!(flags & ZOR_FLAG_SPACE_MAP))
match = B_FALSE;
break;
default:
if (strcmp(zdb_ot_name(obj_type), "zap") == 0) {
if (!(flags & ZOR_FLAG_ZAP))
match = B_FALSE;
break;
}
/*
* If all bits except some of the supported flags are
* set, the user combined the all-types flag (A) with
* a negated flag to exclude some types (e.g. A-f to
* show all object types except plain files).
*/
if ((flags | ZOR_SUPPORTED_FLAGS) != ZOR_FLAG_ALL_TYPES)
match = B_FALSE;
break;
}
return (match);
}
static void
dump_object(objset_t *os, uint64_t object, int verbosity,
boolean_t *print_header, uint64_t *dnode_slots_used, uint64_t flags)
{
dmu_buf_t *db = NULL;
dmu_object_info_t doi;
dnode_t *dn;
boolean_t dnode_held = B_FALSE;
void *bonus = NULL;
size_t bsize = 0;
char iblk[32], dblk[32], lsize[32], asize[32], fill[32], dnsize[32];
char bonus_size[32];
char aux[50];
int error;
/* make sure nicenum has enough space */
_Static_assert(sizeof (iblk) >= NN_NUMBUF_SZ, "iblk truncated");
_Static_assert(sizeof (dblk) >= NN_NUMBUF_SZ, "dblk truncated");
_Static_assert(sizeof (lsize) >= NN_NUMBUF_SZ, "lsize truncated");
_Static_assert(sizeof (asize) >= NN_NUMBUF_SZ, "asize truncated");
_Static_assert(sizeof (bonus_size) >= NN_NUMBUF_SZ,
"bonus_size truncated");
if (*print_header) {
(void) printf("\n%10s %3s %5s %5s %5s %6s %5s %6s %s\n",
"Object", "lvl", "iblk", "dblk", "dsize", "dnsize",
"lsize", "%full", "type");
*print_header = 0;
}
if (object == 0) {
dn = DMU_META_DNODE(os);
dmu_object_info_from_dnode(dn, &doi);
} else {
/*
* Encrypted datasets will have sensitive bonus buffers
* encrypted. Therefore we cannot hold the bonus buffer and
* must hold the dnode itself instead.
*/
error = dmu_object_info(os, object, &doi);
if (error)
fatal("dmu_object_info() failed, errno %u", error);
if (!key_loaded && os->os_encrypted &&
DMU_OT_IS_ENCRYPTED(doi.doi_bonus_type)) {
error = dnode_hold(os, object, FTAG, &dn);
if (error)
fatal("dnode_hold() failed, errno %u", error);
dnode_held = B_TRUE;
} else {
error = dmu_bonus_hold(os, object, FTAG, &db);
if (error)
fatal("dmu_bonus_hold(%llu) failed, errno %u",
object, error);
bonus = db->db_data;
bsize = db->db_size;
dn = DB_DNODE((dmu_buf_impl_t *)db);
}
}
/*
* Default to showing all object types if no flags were specified.
*/
if (flags != 0 && flags != ZOR_FLAG_ALL_TYPES &&
!match_object_type(doi.doi_type, flags))
goto out;
if (dnode_slots_used)
*dnode_slots_used = doi.doi_dnodesize / DNODE_MIN_SIZE;
zdb_nicenum(doi.doi_metadata_block_size, iblk, sizeof (iblk));
zdb_nicenum(doi.doi_data_block_size, dblk, sizeof (dblk));
zdb_nicenum(doi.doi_max_offset, lsize, sizeof (lsize));
zdb_nicenum(doi.doi_physical_blocks_512 << 9, asize, sizeof (asize));
zdb_nicenum(doi.doi_bonus_size, bonus_size, sizeof (bonus_size));
zdb_nicenum(doi.doi_dnodesize, dnsize, sizeof (dnsize));
(void) snprintf(fill, sizeof (fill), "%6.2f", 100.0 *
doi.doi_fill_count * doi.doi_data_block_size / (object == 0 ?
DNODES_PER_BLOCK : 1) / doi.doi_max_offset);
aux[0] = '\0';
if (doi.doi_checksum != ZIO_CHECKSUM_INHERIT || verbosity >= 6) {
(void) snprintf(aux + strlen(aux), sizeof (aux) - strlen(aux),
" (K=%s)", ZDB_CHECKSUM_NAME(doi.doi_checksum));
}
if (doi.doi_compress == ZIO_COMPRESS_INHERIT &&
ZIO_COMPRESS_HASLEVEL(os->os_compress) && verbosity >= 6) {
const char *compname = NULL;
if (zfs_prop_index_to_string(ZFS_PROP_COMPRESSION,
ZIO_COMPRESS_RAW(os->os_compress, os->os_complevel),
&compname) == 0) {
(void) snprintf(aux + strlen(aux),
sizeof (aux) - strlen(aux), " (Z=inherit=%s)",
compname);
} else {
(void) snprintf(aux + strlen(aux),
sizeof (aux) - strlen(aux),
" (Z=inherit=%s-unknown)",
ZDB_COMPRESS_NAME(os->os_compress));
}
} else if (doi.doi_compress == ZIO_COMPRESS_INHERIT && verbosity >= 6) {
(void) snprintf(aux + strlen(aux), sizeof (aux) - strlen(aux),
" (Z=inherit=%s)", ZDB_COMPRESS_NAME(os->os_compress));
} else if (doi.doi_compress != ZIO_COMPRESS_INHERIT || verbosity >= 6) {
(void) snprintf(aux + strlen(aux), sizeof (aux) - strlen(aux),
" (Z=%s)", ZDB_COMPRESS_NAME(doi.doi_compress));
}
(void) printf("%10lld %3u %5s %5s %5s %6s %5s %6s %s%s\n",
(u_longlong_t)object, doi.doi_indirection, iblk, dblk,
asize, dnsize, lsize, fill, zdb_ot_name(doi.doi_type), aux);
if (doi.doi_bonus_type != DMU_OT_NONE && verbosity > 3) {
(void) printf("%10s %3s %5s %5s %5s %5s %5s %6s %s\n",
"", "", "", "", "", "", bonus_size, "bonus",
zdb_ot_name(doi.doi_bonus_type));
}
if (verbosity >= 4) {
(void) printf("\tdnode flags: %s%s%s%s\n",
(dn->dn_phys->dn_flags & DNODE_FLAG_USED_BYTES) ?
"USED_BYTES " : "",
(dn->dn_phys->dn_flags & DNODE_FLAG_USERUSED_ACCOUNTED) ?
"USERUSED_ACCOUNTED " : "",
(dn->dn_phys->dn_flags & DNODE_FLAG_USEROBJUSED_ACCOUNTED) ?
"USEROBJUSED_ACCOUNTED " : "",
(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR) ?
"SPILL_BLKPTR" : "");
(void) printf("\tdnode maxblkid: %llu\n",
(longlong_t)dn->dn_phys->dn_maxblkid);
if (!dnode_held) {
object_viewer[ZDB_OT_TYPE(doi.doi_bonus_type)](os,
object, bonus, bsize);
} else {
(void) printf("\t\t(bonus encrypted)\n");
}
if (key_loaded ||
(!os->os_encrypted || !DMU_OT_IS_ENCRYPTED(doi.doi_type))) {
object_viewer[ZDB_OT_TYPE(doi.doi_type)](os, object,
NULL, 0);
} else {
(void) printf("\t\t(object encrypted)\n");
}
*print_header = B_TRUE;
}
if (verbosity >= 5) {
if (dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR) {
char blkbuf[BP_SPRINTF_LEN];
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf),
DN_SPILL_BLKPTR(dn->dn_phys), B_FALSE);
(void) printf("\nSpill block: %s\n", blkbuf);
}
dump_indirect(dn);
}
if (verbosity >= 5) {
/*
* Report the list of segments that comprise the object.
*/
uint64_t start = 0;
uint64_t end;
uint64_t blkfill = 1;
int minlvl = 1;
if (dn->dn_type == DMU_OT_DNODE) {
minlvl = 0;
blkfill = DNODES_PER_BLOCK;
}
for (;;) {
char segsize[32];
/* make sure nicenum has enough space */
_Static_assert(sizeof (segsize) >= NN_NUMBUF_SZ,
"segsize truncated");
error = dnode_next_offset(dn,
0, &start, minlvl, blkfill, 0);
if (error)
break;
end = start;
error = dnode_next_offset(dn,
DNODE_FIND_HOLE, &end, minlvl, blkfill, 0);
zdb_nicenum(end - start, segsize, sizeof (segsize));
(void) printf("\t\tsegment [%016llx, %016llx)"
" size %5s\n", (u_longlong_t)start,
(u_longlong_t)end, segsize);
if (error)
break;
start = end;
}
}
out:
if (db != NULL)
dmu_buf_rele(db, FTAG);
if (dnode_held)
dnode_rele(dn, FTAG);
}
static void
count_dir_mos_objects(dsl_dir_t *dd)
{
mos_obj_refd(dd->dd_object);
mos_obj_refd(dsl_dir_phys(dd)->dd_child_dir_zapobj);
mos_obj_refd(dsl_dir_phys(dd)->dd_deleg_zapobj);
mos_obj_refd(dsl_dir_phys(dd)->dd_props_zapobj);
mos_obj_refd(dsl_dir_phys(dd)->dd_clones);
/*
* The dd_crypto_obj can be referenced by multiple dsl_dir's.
* Ignore the references after the first one.
*/
mos_obj_refd_multiple(dd->dd_crypto_obj);
}
static void
count_ds_mos_objects(dsl_dataset_t *ds)
{
mos_obj_refd(ds->ds_object);
mos_obj_refd(dsl_dataset_phys(ds)->ds_next_clones_obj);
mos_obj_refd(dsl_dataset_phys(ds)->ds_props_obj);
mos_obj_refd(dsl_dataset_phys(ds)->ds_userrefs_obj);
mos_obj_refd(dsl_dataset_phys(ds)->ds_snapnames_zapobj);
mos_obj_refd(ds->ds_bookmarks_obj);
if (!dsl_dataset_is_snapshot(ds)) {
count_dir_mos_objects(ds->ds_dir);
}
}
static const char *const objset_types[DMU_OST_NUMTYPES] = {
"NONE", "META", "ZPL", "ZVOL", "OTHER", "ANY" };
/*
* Parse a string denoting a range of object IDs of the form
* <start>[:<end>[:flags]], and store the results in zor.
* Return 0 on success. On error, return 1 and update the msg
* pointer to point to a descriptive error message.
*/
static int
parse_object_range(char *range, zopt_object_range_t *zor, const char **msg)
{
uint64_t flags = 0;
char *p, *s, *dup, *flagstr, *tmp = NULL;
size_t len;
int i;
int rc = 0;
if (strchr(range, ':') == NULL) {
zor->zor_obj_start = strtoull(range, &p, 0);
if (*p != '\0') {
*msg = "Invalid characters in object ID";
rc = 1;
}
zor->zor_obj_start = ZDB_MAP_OBJECT_ID(zor->zor_obj_start);
zor->zor_obj_end = zor->zor_obj_start;
return (rc);
}
if (strchr(range, ':') == range) {
*msg = "Invalid leading colon";
rc = 1;
return (rc);
}
len = strlen(range);
if (range[len - 1] == ':') {
*msg = "Invalid trailing colon";
rc = 1;
return (rc);
}
dup = strdup(range);
s = strtok_r(dup, ":", &tmp);
zor->zor_obj_start = strtoull(s, &p, 0);
if (*p != '\0') {
*msg = "Invalid characters in start object ID";
rc = 1;
goto out;
}
s = strtok_r(NULL, ":", &tmp);
zor->zor_obj_end = strtoull(s, &p, 0);
if (*p != '\0') {
*msg = "Invalid characters in end object ID";
rc = 1;
goto out;
}
if (zor->zor_obj_start > zor->zor_obj_end) {
*msg = "Start object ID may not exceed end object ID";
rc = 1;
goto out;
}
s = strtok_r(NULL, ":", &tmp);
if (s == NULL) {
zor->zor_flags = ZOR_FLAG_ALL_TYPES;
goto out;
} else if (strtok_r(NULL, ":", &tmp) != NULL) {
*msg = "Invalid colon-delimited field after flags";
rc = 1;
goto out;
}
flagstr = s;
for (i = 0; flagstr[i]; i++) {
int bit;
boolean_t negation = (flagstr[i] == '-');
if (negation) {
i++;
if (flagstr[i] == '\0') {
*msg = "Invalid trailing negation operator";
rc = 1;
goto out;
}
}
bit = flagbits[(uchar_t)flagstr[i]];
if (bit == 0) {
*msg = "Invalid flag";
rc = 1;
goto out;
}
if (negation)
flags &= ~bit;
else
flags |= bit;
}
zor->zor_flags = flags;
zor->zor_obj_start = ZDB_MAP_OBJECT_ID(zor->zor_obj_start);
zor->zor_obj_end = ZDB_MAP_OBJECT_ID(zor->zor_obj_end);
out:
free(dup);
return (rc);
}
static void
dump_objset(objset_t *os)
{
dmu_objset_stats_t dds = { 0 };
uint64_t object, object_count;
uint64_t refdbytes, usedobjs, scratch;
char numbuf[32];
char blkbuf[BP_SPRINTF_LEN + 20];
char osname[ZFS_MAX_DATASET_NAME_LEN];
const char *type = "UNKNOWN";
int verbosity = dump_opt['d'];
boolean_t print_header;
unsigned i;
int error;
uint64_t total_slots_used = 0;
uint64_t max_slot_used = 0;
uint64_t dnode_slots;
uint64_t obj_start;
uint64_t obj_end;
uint64_t flags;
/* make sure nicenum has enough space */
_Static_assert(sizeof (numbuf) >= NN_NUMBUF_SZ, "numbuf truncated");
dsl_pool_config_enter(dmu_objset_pool(os), FTAG);
dmu_objset_fast_stat(os, &dds);
dsl_pool_config_exit(dmu_objset_pool(os), FTAG);
print_header = B_TRUE;
if (dds.dds_type < DMU_OST_NUMTYPES)
type = objset_types[dds.dds_type];
if (dds.dds_type == DMU_OST_META) {
dds.dds_creation_txg = TXG_INITIAL;
usedobjs = BP_GET_FILL(os->os_rootbp);
refdbytes = dsl_dir_phys(os->os_spa->spa_dsl_pool->dp_mos_dir)->
dd_used_bytes;
} else {
dmu_objset_space(os, &refdbytes, &scratch, &usedobjs, &scratch);
}
ASSERT3U(usedobjs, ==, BP_GET_FILL(os->os_rootbp));
zdb_nicenum(refdbytes, numbuf, sizeof (numbuf));
if (verbosity >= 4) {
(void) snprintf(blkbuf, sizeof (blkbuf), ", rootbp ");
(void) snprintf_blkptr(blkbuf + strlen(blkbuf),
sizeof (blkbuf) - strlen(blkbuf), os->os_rootbp);
} else {
blkbuf[0] = '\0';
}
dmu_objset_name(os, osname);
(void) printf("Dataset %s [%s], ID %llu, cr_txg %llu, "
"%s, %llu objects%s%s\n",
osname, type, (u_longlong_t)dmu_objset_id(os),
(u_longlong_t)dds.dds_creation_txg,
numbuf, (u_longlong_t)usedobjs, blkbuf,
(dds.dds_inconsistent) ? " (inconsistent)" : "");
for (i = 0; i < zopt_object_args; i++) {
obj_start = zopt_object_ranges[i].zor_obj_start;
obj_end = zopt_object_ranges[i].zor_obj_end;
flags = zopt_object_ranges[i].zor_flags;
object = obj_start;
if (object == 0 || obj_start == obj_end)
dump_object(os, object, verbosity, &print_header, NULL,
flags);
else
object--;
while ((dmu_object_next(os, &object, B_FALSE, 0) == 0) &&
object <= obj_end) {
dump_object(os, object, verbosity, &print_header, NULL,
flags);
}
}
if (zopt_object_args > 0) {
(void) printf("\n");
return;
}
if (dump_opt['i'] != 0 || verbosity >= 2)
dump_intent_log(dmu_objset_zil(os));
if (dmu_objset_ds(os) != NULL) {
dsl_dataset_t *ds = dmu_objset_ds(os);
dump_blkptr_list(&ds->ds_deadlist, "Deadlist");
if (dsl_deadlist_is_open(&ds->ds_dir->dd_livelist) &&
!dmu_objset_is_snapshot(os)) {
dump_blkptr_list(&ds->ds_dir->dd_livelist, "Livelist");
if (verify_dd_livelist(os) != 0)
fatal("livelist is incorrect");
}
if (dsl_dataset_remap_deadlist_exists(ds)) {
(void) printf("ds_remap_deadlist:\n");
dump_blkptr_list(&ds->ds_remap_deadlist, "Deadlist");
}
count_ds_mos_objects(ds);
}
if (dmu_objset_ds(os) != NULL)
dump_bookmarks(os, verbosity);
if (verbosity < 2)
return;
if (BP_IS_HOLE(os->os_rootbp))
return;
dump_object(os, 0, verbosity, &print_header, NULL, 0);
object_count = 0;
if (DMU_USERUSED_DNODE(os) != NULL &&
DMU_USERUSED_DNODE(os)->dn_type != 0) {
dump_object(os, DMU_USERUSED_OBJECT, verbosity, &print_header,
NULL, 0);
dump_object(os, DMU_GROUPUSED_OBJECT, verbosity, &print_header,
NULL, 0);
}
if (DMU_PROJECTUSED_DNODE(os) != NULL &&
DMU_PROJECTUSED_DNODE(os)->dn_type != 0)
dump_object(os, DMU_PROJECTUSED_OBJECT, verbosity,
&print_header, NULL, 0);
object = 0;
while ((error = dmu_object_next(os, &object, B_FALSE, 0)) == 0) {
dump_object(os, object, verbosity, &print_header, &dnode_slots,
0);
object_count++;
total_slots_used += dnode_slots;
max_slot_used = object + dnode_slots - 1;
}
(void) printf("\n");
(void) printf(" Dnode slots:\n");
(void) printf("\tTotal used: %10llu\n",
(u_longlong_t)total_slots_used);
(void) printf("\tMax used: %10llu\n",
(u_longlong_t)max_slot_used);
(void) printf("\tPercent empty: %10lf\n",
(double)(max_slot_used - total_slots_used)*100 /
(double)max_slot_used);
(void) printf("\n");
if (error != ESRCH) {
(void) fprintf(stderr, "dmu_object_next() = %d\n", error);
abort();
}
ASSERT3U(object_count, ==, usedobjs);
if (leaked_objects != 0) {
(void) printf("%d potentially leaked objects detected\n",
leaked_objects);
leaked_objects = 0;
}
}
static void
dump_uberblock(uberblock_t *ub, const char *header, const char *footer)
{
time_t timestamp = ub->ub_timestamp;
(void) printf("%s", header ? header : "");
(void) printf("\tmagic = %016llx\n", (u_longlong_t)ub->ub_magic);
(void) printf("\tversion = %llu\n", (u_longlong_t)ub->ub_version);
(void) printf("\ttxg = %llu\n", (u_longlong_t)ub->ub_txg);
(void) printf("\tguid_sum = %llu\n", (u_longlong_t)ub->ub_guid_sum);
(void) printf("\ttimestamp = %llu UTC = %s",
(u_longlong_t)ub->ub_timestamp, ctime(&timestamp));
(void) printf("\tmmp_magic = %016llx\n",
(u_longlong_t)ub->ub_mmp_magic);
if (MMP_VALID(ub)) {
(void) printf("\tmmp_delay = %0llu\n",
(u_longlong_t)ub->ub_mmp_delay);
if (MMP_SEQ_VALID(ub))
(void) printf("\tmmp_seq = %u\n",
(unsigned int) MMP_SEQ(ub));
if (MMP_FAIL_INT_VALID(ub))
(void) printf("\tmmp_fail = %u\n",
(unsigned int) MMP_FAIL_INT(ub));
if (MMP_INTERVAL_VALID(ub))
(void) printf("\tmmp_write = %u\n",
(unsigned int) MMP_INTERVAL(ub));
/* After MMP_* to make summarize_uberblock_mmp cleaner */
(void) printf("\tmmp_valid = %x\n",
(unsigned int) ub->ub_mmp_config & 0xFF);
}
if (dump_opt['u'] >= 4) {
char blkbuf[BP_SPRINTF_LEN];
snprintf_blkptr(blkbuf, sizeof (blkbuf), &ub->ub_rootbp);
(void) printf("\trootbp = %s\n", blkbuf);
}
(void) printf("\tcheckpoint_txg = %llu\n",
(u_longlong_t)ub->ub_checkpoint_txg);
+
+ (void) printf("\traidz_reflow state=%u off=%llu\n",
+ (int)RRSS_GET_STATE(ub),
+ (u_longlong_t)RRSS_GET_OFFSET(ub));
+
(void) printf("%s", footer ? footer : "");
}
static void
dump_config(spa_t *spa)
{
dmu_buf_t *db;
size_t nvsize = 0;
int error = 0;
error = dmu_bonus_hold(spa->spa_meta_objset,
spa->spa_config_object, FTAG, &db);
if (error == 0) {
nvsize = *(uint64_t *)db->db_data;
dmu_buf_rele(db, FTAG);
(void) printf("\nMOS Configuration:\n");
dump_packed_nvlist(spa->spa_meta_objset,
spa->spa_config_object, (void *)&nvsize, 1);
} else {
(void) fprintf(stderr, "dmu_bonus_hold(%llu) failed, errno %d",
(u_longlong_t)spa->spa_config_object, error);
}
}
static void
dump_cachefile(const char *cachefile)
{
int fd;
struct stat64 statbuf;
char *buf;
nvlist_t *config;
if ((fd = open64(cachefile, O_RDONLY)) < 0) {
(void) printf("cannot open '%s': %s\n", cachefile,
strerror(errno));
exit(1);
}
if (fstat64(fd, &statbuf) != 0) {
(void) printf("failed to stat '%s': %s\n", cachefile,
strerror(errno));
exit(1);
}
if ((buf = malloc(statbuf.st_size)) == NULL) {
(void) fprintf(stderr, "failed to allocate %llu bytes\n",
(u_longlong_t)statbuf.st_size);
exit(1);
}
if (read(fd, buf, statbuf.st_size) != statbuf.st_size) {
(void) fprintf(stderr, "failed to read %llu bytes\n",
(u_longlong_t)statbuf.st_size);
exit(1);
}
(void) close(fd);
if (nvlist_unpack(buf, statbuf.st_size, &config, 0) != 0) {
(void) fprintf(stderr, "failed to unpack nvlist\n");
exit(1);
}
free(buf);
dump_nvlist(config, 0);
nvlist_free(config);
}
/*
* ZFS label nvlist stats
*/
typedef struct zdb_nvl_stats {
int zns_list_count;
int zns_leaf_count;
size_t zns_leaf_largest;
size_t zns_leaf_total;
nvlist_t *zns_string;
nvlist_t *zns_uint64;
nvlist_t *zns_boolean;
} zdb_nvl_stats_t;
static void
collect_nvlist_stats(nvlist_t *nvl, zdb_nvl_stats_t *stats)
{
nvlist_t *list, **array;
nvpair_t *nvp = NULL;
const char *name;
uint_t i, items;
stats->zns_list_count++;
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
name = nvpair_name(nvp);
switch (nvpair_type(nvp)) {
case DATA_TYPE_STRING:
fnvlist_add_string(stats->zns_string, name,
fnvpair_value_string(nvp));
break;
case DATA_TYPE_UINT64:
fnvlist_add_uint64(stats->zns_uint64, name,
fnvpair_value_uint64(nvp));
break;
case DATA_TYPE_BOOLEAN:
fnvlist_add_boolean(stats->zns_boolean, name);
break;
case DATA_TYPE_NVLIST:
if (nvpair_value_nvlist(nvp, &list) == 0)
collect_nvlist_stats(list, stats);
break;
case DATA_TYPE_NVLIST_ARRAY:
if (nvpair_value_nvlist_array(nvp, &array, &items) != 0)
break;
for (i = 0; i < items; i++) {
collect_nvlist_stats(array[i], stats);
/* collect stats on leaf vdev */
if (strcmp(name, "children") == 0) {
size_t size;
(void) nvlist_size(array[i], &size,
NV_ENCODE_XDR);
stats->zns_leaf_total += size;
if (size > stats->zns_leaf_largest)
stats->zns_leaf_largest = size;
stats->zns_leaf_count++;
}
}
break;
default:
(void) printf("skip type %d!\n", (int)nvpair_type(nvp));
}
}
}
static void
dump_nvlist_stats(nvlist_t *nvl, size_t cap)
{
zdb_nvl_stats_t stats = { 0 };
size_t size, sum = 0, total;
size_t noise;
/* requires nvlist with non-unique names for stat collection */
VERIFY0(nvlist_alloc(&stats.zns_string, 0, 0));
VERIFY0(nvlist_alloc(&stats.zns_uint64, 0, 0));
VERIFY0(nvlist_alloc(&stats.zns_boolean, 0, 0));
VERIFY0(nvlist_size(stats.zns_boolean, &noise, NV_ENCODE_XDR));
(void) printf("\n\nZFS Label NVList Config Stats:\n");
VERIFY0(nvlist_size(nvl, &total, NV_ENCODE_XDR));
(void) printf(" %d bytes used, %d bytes free (using %4.1f%%)\n\n",
(int)total, (int)(cap - total), 100.0 * total / cap);
collect_nvlist_stats(nvl, &stats);
VERIFY0(nvlist_size(stats.zns_uint64, &size, NV_ENCODE_XDR));
size -= noise;
sum += size;
(void) printf("%12s %4d %6d bytes (%5.2f%%)\n", "integers:",
(int)fnvlist_num_pairs(stats.zns_uint64),
(int)size, 100.0 * size / total);
VERIFY0(nvlist_size(stats.zns_string, &size, NV_ENCODE_XDR));
size -= noise;
sum += size;
(void) printf("%12s %4d %6d bytes (%5.2f%%)\n", "strings:",
(int)fnvlist_num_pairs(stats.zns_string),
(int)size, 100.0 * size / total);
VERIFY0(nvlist_size(stats.zns_boolean, &size, NV_ENCODE_XDR));
size -= noise;
sum += size;
(void) printf("%12s %4d %6d bytes (%5.2f%%)\n", "booleans:",
(int)fnvlist_num_pairs(stats.zns_boolean),
(int)size, 100.0 * size / total);
size = total - sum; /* treat remainder as nvlist overhead */
(void) printf("%12s %4d %6d bytes (%5.2f%%)\n\n", "nvlists:",
stats.zns_list_count, (int)size, 100.0 * size / total);
if (stats.zns_leaf_count > 0) {
size_t average = stats.zns_leaf_total / stats.zns_leaf_count;
(void) printf("%12s %4d %6d bytes average\n", "leaf vdevs:",
stats.zns_leaf_count, (int)average);
(void) printf("%24d bytes largest\n",
(int)stats.zns_leaf_largest);
if (dump_opt['l'] >= 3 && average > 0)
(void) printf(" space for %d additional leaf vdevs\n",
(int)((cap - total) / average));
}
(void) printf("\n");
nvlist_free(stats.zns_string);
nvlist_free(stats.zns_uint64);
nvlist_free(stats.zns_boolean);
}
typedef struct cksum_record {
zio_cksum_t cksum;
boolean_t labels[VDEV_LABELS];
avl_node_t link;
} cksum_record_t;
static int
cksum_record_compare(const void *x1, const void *x2)
{
const cksum_record_t *l = (cksum_record_t *)x1;
const cksum_record_t *r = (cksum_record_t *)x2;
int arraysize = ARRAY_SIZE(l->cksum.zc_word);
int difference = 0;
for (int i = 0; i < arraysize; i++) {
difference = TREE_CMP(l->cksum.zc_word[i], r->cksum.zc_word[i]);
if (difference)
break;
}
return (difference);
}
static cksum_record_t *
cksum_record_alloc(zio_cksum_t *cksum, int l)
{
cksum_record_t *rec;
rec = umem_zalloc(sizeof (*rec), UMEM_NOFAIL);
rec->cksum = *cksum;
rec->labels[l] = B_TRUE;
return (rec);
}
static cksum_record_t *
cksum_record_lookup(avl_tree_t *tree, zio_cksum_t *cksum)
{
cksum_record_t lookup = { .cksum = *cksum };
avl_index_t where;
return (avl_find(tree, &lookup, &where));
}
static cksum_record_t *
cksum_record_insert(avl_tree_t *tree, zio_cksum_t *cksum, int l)
{
cksum_record_t *rec;
rec = cksum_record_lookup(tree, cksum);
if (rec) {
rec->labels[l] = B_TRUE;
} else {
rec = cksum_record_alloc(cksum, l);
avl_add(tree, rec);
}
return (rec);
}
static int
first_label(cksum_record_t *rec)
{
for (int i = 0; i < VDEV_LABELS; i++)
if (rec->labels[i])
return (i);
return (-1);
}
static void
print_label_numbers(const char *prefix, const cksum_record_t *rec)
{
fputs(prefix, stdout);
for (int i = 0; i < VDEV_LABELS; i++)
if (rec->labels[i] == B_TRUE)
printf("%d ", i);
putchar('\n');
}
#define MAX_UBERBLOCK_COUNT (VDEV_UBERBLOCK_RING >> UBERBLOCK_SHIFT)
typedef struct zdb_label {
vdev_label_t label;
uint64_t label_offset;
nvlist_t *config_nv;
cksum_record_t *config;
cksum_record_t *uberblocks[MAX_UBERBLOCK_COUNT];
boolean_t header_printed;
boolean_t read_failed;
boolean_t cksum_valid;
} zdb_label_t;
static void
print_label_header(zdb_label_t *label, int l)
{
if (dump_opt['q'])
return;
if (label->header_printed == B_TRUE)
return;
(void) printf("------------------------------------\n");
(void) printf("LABEL %d %s\n", l,
label->cksum_valid ? "" : "(Bad label cksum)");
(void) printf("------------------------------------\n");
label->header_printed = B_TRUE;
}
static void
print_l2arc_header(void)
{
(void) printf("------------------------------------\n");
(void) printf("L2ARC device header\n");
(void) printf("------------------------------------\n");
}
static void
print_l2arc_log_blocks(void)
{
(void) printf("------------------------------------\n");
(void) printf("L2ARC device log blocks\n");
(void) printf("------------------------------------\n");
}
static void
dump_l2arc_log_entries(uint64_t log_entries,
l2arc_log_ent_phys_t *le, uint64_t i)
{
for (int j = 0; j < log_entries; j++) {
dva_t dva = le[j].le_dva;
(void) printf("lb[%4llu]\tle[%4d]\tDVA asize: %llu, "
"vdev: %llu, offset: %llu\n",
(u_longlong_t)i, j + 1,
(u_longlong_t)DVA_GET_ASIZE(&dva),
(u_longlong_t)DVA_GET_VDEV(&dva),
(u_longlong_t)DVA_GET_OFFSET(&dva));
(void) printf("|\t\t\t\tbirth: %llu\n",
(u_longlong_t)le[j].le_birth);
(void) printf("|\t\t\t\tlsize: %llu\n",
(u_longlong_t)L2BLK_GET_LSIZE((&le[j])->le_prop));
(void) printf("|\t\t\t\tpsize: %llu\n",
(u_longlong_t)L2BLK_GET_PSIZE((&le[j])->le_prop));
(void) printf("|\t\t\t\tcompr: %llu\n",
(u_longlong_t)L2BLK_GET_COMPRESS((&le[j])->le_prop));
(void) printf("|\t\t\t\tcomplevel: %llu\n",
(u_longlong_t)(&le[j])->le_complevel);
(void) printf("|\t\t\t\ttype: %llu\n",
(u_longlong_t)L2BLK_GET_TYPE((&le[j])->le_prop));
(void) printf("|\t\t\t\tprotected: %llu\n",
(u_longlong_t)L2BLK_GET_PROTECTED((&le[j])->le_prop));
(void) printf("|\t\t\t\tprefetch: %llu\n",
(u_longlong_t)L2BLK_GET_PREFETCH((&le[j])->le_prop));
(void) printf("|\t\t\t\taddress: %llu\n",
(u_longlong_t)le[j].le_daddr);
(void) printf("|\t\t\t\tARC state: %llu\n",
(u_longlong_t)L2BLK_GET_STATE((&le[j])->le_prop));
(void) printf("|\n");
}
(void) printf("\n");
}
static void
dump_l2arc_log_blkptr(const l2arc_log_blkptr_t *lbps)
{
(void) printf("|\t\tdaddr: %llu\n", (u_longlong_t)lbps->lbp_daddr);
(void) printf("|\t\tpayload_asize: %llu\n",
(u_longlong_t)lbps->lbp_payload_asize);
(void) printf("|\t\tpayload_start: %llu\n",
(u_longlong_t)lbps->lbp_payload_start);
(void) printf("|\t\tlsize: %llu\n",
(u_longlong_t)L2BLK_GET_LSIZE(lbps->lbp_prop));
(void) printf("|\t\tasize: %llu\n",
(u_longlong_t)L2BLK_GET_PSIZE(lbps->lbp_prop));
(void) printf("|\t\tcompralgo: %llu\n",
(u_longlong_t)L2BLK_GET_COMPRESS(lbps->lbp_prop));
(void) printf("|\t\tcksumalgo: %llu\n",
(u_longlong_t)L2BLK_GET_CHECKSUM(lbps->lbp_prop));
(void) printf("|\n\n");
}
static void
dump_l2arc_log_blocks(int fd, const l2arc_dev_hdr_phys_t *l2dhdr,
l2arc_dev_hdr_phys_t *rebuild)
{
l2arc_log_blk_phys_t this_lb;
uint64_t asize;
l2arc_log_blkptr_t lbps[2];
abd_t *abd;
zio_cksum_t cksum;
int failed = 0;
l2arc_dev_t dev;
if (!dump_opt['q'])
print_l2arc_log_blocks();
memcpy(lbps, l2dhdr->dh_start_lbps, sizeof (lbps));
dev.l2ad_evict = l2dhdr->dh_evict;
dev.l2ad_start = l2dhdr->dh_start;
dev.l2ad_end = l2dhdr->dh_end;
if (l2dhdr->dh_start_lbps[0].lbp_daddr == 0) {
/* no log blocks to read */
if (!dump_opt['q']) {
(void) printf("No log blocks to read\n");
(void) printf("\n");
}
return;
} else {
dev.l2ad_hand = lbps[0].lbp_daddr +
L2BLK_GET_PSIZE((&lbps[0])->lbp_prop);
}
dev.l2ad_first = !!(l2dhdr->dh_flags & L2ARC_DEV_HDR_EVICT_FIRST);
for (;;) {
if (!l2arc_log_blkptr_valid(&dev, &lbps[0]))
break;
/* L2BLK_GET_PSIZE returns aligned size for log blocks */
asize = L2BLK_GET_PSIZE((&lbps[0])->lbp_prop);
if (pread64(fd, &this_lb, asize, lbps[0].lbp_daddr) != asize) {
if (!dump_opt['q']) {
(void) printf("Error while reading next log "
"block\n\n");
}
break;
}
fletcher_4_native_varsize(&this_lb, asize, &cksum);
if (!ZIO_CHECKSUM_EQUAL(cksum, lbps[0].lbp_cksum)) {
failed++;
if (!dump_opt['q']) {
(void) printf("Invalid cksum\n");
dump_l2arc_log_blkptr(&lbps[0]);
}
break;
}
switch (L2BLK_GET_COMPRESS((&lbps[0])->lbp_prop)) {
case ZIO_COMPRESS_OFF:
break;
default:
abd = abd_alloc_for_io(asize, B_TRUE);
abd_copy_from_buf_off(abd, &this_lb, 0, asize);
if (zio_decompress_data(L2BLK_GET_COMPRESS(
(&lbps[0])->lbp_prop), abd, &this_lb,
asize, sizeof (this_lb), NULL) != 0) {
(void) printf("L2ARC block decompression "
"failed\n");
abd_free(abd);
goto out;
}
abd_free(abd);
break;
}
if (this_lb.lb_magic == BSWAP_64(L2ARC_LOG_BLK_MAGIC))
byteswap_uint64_array(&this_lb, sizeof (this_lb));
if (this_lb.lb_magic != L2ARC_LOG_BLK_MAGIC) {
if (!dump_opt['q'])
(void) printf("Invalid log block magic\n\n");
break;
}
rebuild->dh_lb_count++;
rebuild->dh_lb_asize += asize;
if (dump_opt['l'] > 1 && !dump_opt['q']) {
(void) printf("lb[%4llu]\tmagic: %llu\n",
(u_longlong_t)rebuild->dh_lb_count,
(u_longlong_t)this_lb.lb_magic);
dump_l2arc_log_blkptr(&lbps[0]);
}
if (dump_opt['l'] > 2 && !dump_opt['q'])
dump_l2arc_log_entries(l2dhdr->dh_log_entries,
this_lb.lb_entries,
rebuild->dh_lb_count);
if (l2arc_range_check_overlap(lbps[1].lbp_payload_start,
lbps[0].lbp_payload_start, dev.l2ad_evict) &&
!dev.l2ad_first)
break;
lbps[0] = lbps[1];
lbps[1] = this_lb.lb_prev_lbp;
}
out:
if (!dump_opt['q']) {
(void) printf("log_blk_count:\t %llu with valid cksum\n",
(u_longlong_t)rebuild->dh_lb_count);
(void) printf("\t\t %d with invalid cksum\n", failed);
(void) printf("log_blk_asize:\t %llu\n\n",
(u_longlong_t)rebuild->dh_lb_asize);
}
}
static int
dump_l2arc_header(int fd)
{
l2arc_dev_hdr_phys_t l2dhdr = {0}, rebuild = {0};
int error = B_FALSE;
if (pread64(fd, &l2dhdr, sizeof (l2dhdr),
VDEV_LABEL_START_SIZE) != sizeof (l2dhdr)) {
error = B_TRUE;
} else {
if (l2dhdr.dh_magic == BSWAP_64(L2ARC_DEV_HDR_MAGIC))
byteswap_uint64_array(&l2dhdr, sizeof (l2dhdr));
if (l2dhdr.dh_magic != L2ARC_DEV_HDR_MAGIC)
error = B_TRUE;
}
if (error) {
(void) printf("L2ARC device header not found\n\n");
/* Do not return an error here for backward compatibility */
return (0);
} else if (!dump_opt['q']) {
print_l2arc_header();
(void) printf(" magic: %llu\n",
(u_longlong_t)l2dhdr.dh_magic);
(void) printf(" version: %llu\n",
(u_longlong_t)l2dhdr.dh_version);
(void) printf(" pool_guid: %llu\n",
(u_longlong_t)l2dhdr.dh_spa_guid);
(void) printf(" flags: %llu\n",
(u_longlong_t)l2dhdr.dh_flags);
(void) printf(" start_lbps[0]: %llu\n",
(u_longlong_t)
l2dhdr.dh_start_lbps[0].lbp_daddr);
(void) printf(" start_lbps[1]: %llu\n",
(u_longlong_t)
l2dhdr.dh_start_lbps[1].lbp_daddr);
(void) printf(" log_blk_ent: %llu\n",
(u_longlong_t)l2dhdr.dh_log_entries);
(void) printf(" start: %llu\n",
(u_longlong_t)l2dhdr.dh_start);
(void) printf(" end: %llu\n",
(u_longlong_t)l2dhdr.dh_end);
(void) printf(" evict: %llu\n",
(u_longlong_t)l2dhdr.dh_evict);
(void) printf(" lb_asize_refcount: %llu\n",
(u_longlong_t)l2dhdr.dh_lb_asize);
(void) printf(" lb_count_refcount: %llu\n",
(u_longlong_t)l2dhdr.dh_lb_count);
(void) printf(" trim_action_time: %llu\n",
(u_longlong_t)l2dhdr.dh_trim_action_time);
(void) printf(" trim_state: %llu\n\n",
(u_longlong_t)l2dhdr.dh_trim_state);
}
dump_l2arc_log_blocks(fd, &l2dhdr, &rebuild);
/*
* The total aligned size of log blocks and the number of log blocks
* reported in the header of the device may be less than what zdb
* reports by dump_l2arc_log_blocks() which emulates l2arc_rebuild().
* This happens because dump_l2arc_log_blocks() lacks the memory
* pressure valve that l2arc_rebuild() has. Thus, if we are on a system
* with low memory, l2arc_rebuild will exit prematurely and dh_lb_asize
* and dh_lb_count will be lower to begin with than what exists on the
* device. This is normal and zdb should not exit with an error. The
* opposite case should never happen though, the values reported in the
* header should never be higher than what dump_l2arc_log_blocks() and
* l2arc_rebuild() report. If this happens there is a leak in the
* accounting of log blocks.
*/
if (l2dhdr.dh_lb_asize > rebuild.dh_lb_asize ||
l2dhdr.dh_lb_count > rebuild.dh_lb_count)
return (1);
return (0);
}
static void
dump_config_from_label(zdb_label_t *label, size_t buflen, int l)
{
if (dump_opt['q'])
return;
if ((dump_opt['l'] < 3) && (first_label(label->config) != l))
return;
print_label_header(label, l);
dump_nvlist(label->config_nv, 4);
print_label_numbers(" labels = ", label->config);
if (dump_opt['l'] >= 2)
dump_nvlist_stats(label->config_nv, buflen);
}
#define ZDB_MAX_UB_HEADER_SIZE 32
static void
dump_label_uberblocks(zdb_label_t *label, uint64_t ashift, int label_num)
{
vdev_t vd;
char header[ZDB_MAX_UB_HEADER_SIZE];
vd.vdev_ashift = ashift;
vd.vdev_top = &vd;
for (int i = 0; i < VDEV_UBERBLOCK_COUNT(&vd); i++) {
uint64_t uoff = VDEV_UBERBLOCK_OFFSET(&vd, i);
uberblock_t *ub = (void *)((char *)&label->label + uoff);
cksum_record_t *rec = label->uberblocks[i];
if (rec == NULL) {
if (dump_opt['u'] >= 2) {
print_label_header(label, label_num);
(void) printf(" Uberblock[%d] invalid\n", i);
}
continue;
}
if ((dump_opt['u'] < 3) && (first_label(rec) != label_num))
continue;
if ((dump_opt['u'] < 4) &&
(ub->ub_mmp_magic == MMP_MAGIC) && ub->ub_mmp_delay &&
(i >= VDEV_UBERBLOCK_COUNT(&vd) - MMP_BLOCKS_PER_LABEL))
continue;
print_label_header(label, label_num);
(void) snprintf(header, ZDB_MAX_UB_HEADER_SIZE,
" Uberblock[%d]\n", i);
dump_uberblock(ub, header, "");
print_label_numbers(" labels = ", rec);
}
}
static char curpath[PATH_MAX];
/*
* Iterate through the path components, recursively passing
* current one's obj and remaining path until we find the obj
* for the last one.
*/
static int
dump_path_impl(objset_t *os, uint64_t obj, char *name, uint64_t *retobj)
{
int err;
boolean_t header = B_TRUE;
uint64_t child_obj;
char *s;
dmu_buf_t *db;
dmu_object_info_t doi;
if ((s = strchr(name, '/')) != NULL)
*s = '\0';
err = zap_lookup(os, obj, name, 8, 1, &child_obj);
(void) strlcat(curpath, name, sizeof (curpath));
if (err != 0) {
(void) fprintf(stderr, "failed to lookup %s: %s\n",
curpath, strerror(err));
return (err);
}
child_obj = ZFS_DIRENT_OBJ(child_obj);
err = sa_buf_hold(os, child_obj, FTAG, &db);
if (err != 0) {
(void) fprintf(stderr,
"failed to get SA dbuf for obj %llu: %s\n",
(u_longlong_t)child_obj, strerror(err));
return (EINVAL);
}
dmu_object_info_from_db(db, &doi);
sa_buf_rele(db, FTAG);
if (doi.doi_bonus_type != DMU_OT_SA &&
doi.doi_bonus_type != DMU_OT_ZNODE) {
(void) fprintf(stderr, "invalid bonus type %d for obj %llu\n",
doi.doi_bonus_type, (u_longlong_t)child_obj);
return (EINVAL);
}
if (dump_opt['v'] > 6) {
(void) printf("obj=%llu %s type=%d bonustype=%d\n",
(u_longlong_t)child_obj, curpath, doi.doi_type,
doi.doi_bonus_type);
}
(void) strlcat(curpath, "/", sizeof (curpath));
switch (doi.doi_type) {
case DMU_OT_DIRECTORY_CONTENTS:
if (s != NULL && *(s + 1) != '\0')
return (dump_path_impl(os, child_obj, s + 1, retobj));
zfs_fallthrough;
case DMU_OT_PLAIN_FILE_CONTENTS:
if (retobj != NULL) {
*retobj = child_obj;
} else {
dump_object(os, child_obj, dump_opt['v'], &header,
NULL, 0);
}
return (0);
default:
(void) fprintf(stderr, "object %llu has non-file/directory "
"type %d\n", (u_longlong_t)obj, doi.doi_type);
break;
}
return (EINVAL);
}
/*
* Dump the blocks for the object specified by path inside the dataset.
*/
static int
dump_path(char *ds, char *path, uint64_t *retobj)
{
int err;
objset_t *os;
uint64_t root_obj;
err = open_objset(ds, FTAG, &os);
if (err != 0)
return (err);
err = zap_lookup(os, MASTER_NODE_OBJ, ZFS_ROOT_OBJ, 8, 1, &root_obj);
if (err != 0) {
(void) fprintf(stderr, "can't lookup root znode: %s\n",
strerror(err));
close_objset(os, FTAG);
return (EINVAL);
}
(void) snprintf(curpath, sizeof (curpath), "dataset=%s path=/", ds);
err = dump_path_impl(os, root_obj, path, retobj);
close_objset(os, FTAG);
return (err);
}
static int
dump_backup_bytes(objset_t *os, void *buf, int len, void *arg)
{
const char *p = (const char *)buf;
ssize_t nwritten;
(void) os;
(void) arg;
/* Write the data out, handling short writes and signals. */
while ((nwritten = write(STDOUT_FILENO, p, len)) < len) {
if (nwritten < 0) {
if (errno == EINTR)
continue;
return (errno);
}
p += nwritten;
len -= nwritten;
}
return (0);
}
static void
dump_backup(const char *pool, uint64_t objset_id, const char *flagstr)
{
boolean_t embed = B_FALSE;
boolean_t large_block = B_FALSE;
boolean_t compress = B_FALSE;
boolean_t raw = B_FALSE;
const char *c;
for (c = flagstr; c != NULL && *c != '\0'; c++) {
switch (*c) {
case 'e':
embed = B_TRUE;
break;
case 'L':
large_block = B_TRUE;
break;
case 'c':
compress = B_TRUE;
break;
case 'w':
raw = B_TRUE;
break;
default:
fprintf(stderr, "dump_backup: invalid flag "
"'%c'\n", *c);
return;
}
}
if (isatty(STDOUT_FILENO)) {
fprintf(stderr, "dump_backup: stream cannot be written "
"to a terminal\n");
return;
}
offset_t off = 0;
dmu_send_outparams_t out = {
.dso_outfunc = dump_backup_bytes,
.dso_dryrun = B_FALSE,
};
int err = dmu_send_obj(pool, objset_id, /* fromsnap */0, embed,
large_block, compress, raw, /* saved */ B_FALSE, STDOUT_FILENO,
&off, &out);
if (err != 0) {
fprintf(stderr, "dump_backup: dmu_send_obj: %s\n",
strerror(err));
return;
}
}
static int
zdb_copy_object(objset_t *os, uint64_t srcobj, char *destfile)
{
int err = 0;
uint64_t size, readsize, oursize, offset;
ssize_t writesize;
sa_handle_t *hdl;
(void) printf("Copying object %" PRIu64 " to file %s\n", srcobj,
destfile);
VERIFY3P(os, ==, sa_os);
if ((err = sa_handle_get(os, srcobj, NULL, SA_HDL_PRIVATE, &hdl))) {
(void) printf("Failed to get handle for SA znode\n");
return (err);
}
if ((err = sa_lookup(hdl, sa_attr_table[ZPL_SIZE], &size, 8))) {
(void) sa_handle_destroy(hdl);
return (err);
}
(void) sa_handle_destroy(hdl);
(void) printf("Object %" PRIu64 " is %" PRIu64 " bytes\n", srcobj,
size);
if (size == 0) {
return (EINVAL);
}
int fd = open(destfile, O_WRONLY | O_CREAT | O_TRUNC, 0644);
if (fd == -1)
return (errno);
/*
* We cap the size at 1 mebibyte here to prevent
* allocation failures and nigh-infinite printing if the
* object is extremely large.
*/
oursize = MIN(size, 1 << 20);
offset = 0;
char *buf = kmem_alloc(oursize, KM_NOSLEEP);
if (buf == NULL) {
(void) close(fd);
return (ENOMEM);
}
while (offset < size) {
readsize = MIN(size - offset, 1 << 20);
err = dmu_read(os, srcobj, offset, readsize, buf, 0);
if (err != 0) {
(void) printf("got error %u from dmu_read\n", err);
kmem_free(buf, oursize);
(void) close(fd);
return (err);
}
if (dump_opt['v'] > 3) {
(void) printf("Read offset=%" PRIu64 " size=%" PRIu64
" error=%d\n", offset, readsize, err);
}
writesize = write(fd, buf, readsize);
if (writesize < 0) {
err = errno;
break;
} else if (writesize != readsize) {
/* Incomplete write */
(void) fprintf(stderr, "Short write, only wrote %llu of"
" %" PRIu64 " bytes, exiting...\n",
(u_longlong_t)writesize, readsize);
break;
}
offset += readsize;
}
(void) close(fd);
if (buf != NULL)
kmem_free(buf, oursize);
return (err);
}
static boolean_t
label_cksum_valid(vdev_label_t *label, uint64_t offset)
{
zio_checksum_info_t *ci = &zio_checksum_table[ZIO_CHECKSUM_LABEL];
zio_cksum_t expected_cksum;
zio_cksum_t actual_cksum;
zio_cksum_t verifier;
zio_eck_t *eck;
int byteswap;
void *data = (char *)label + offsetof(vdev_label_t, vl_vdev_phys);
eck = (zio_eck_t *)((char *)(data) + VDEV_PHYS_SIZE) - 1;
offset += offsetof(vdev_label_t, vl_vdev_phys);
ZIO_SET_CHECKSUM(&verifier, offset, 0, 0, 0);
byteswap = (eck->zec_magic == BSWAP_64(ZEC_MAGIC));
if (byteswap)
byteswap_uint64_array(&verifier, sizeof (zio_cksum_t));
expected_cksum = eck->zec_cksum;
eck->zec_cksum = verifier;
abd_t *abd = abd_get_from_buf(data, VDEV_PHYS_SIZE);
ci->ci_func[byteswap](abd, VDEV_PHYS_SIZE, NULL, &actual_cksum);
abd_free(abd);
if (byteswap)
byteswap_uint64_array(&expected_cksum, sizeof (zio_cksum_t));
if (ZIO_CHECKSUM_EQUAL(actual_cksum, expected_cksum))
return (B_TRUE);
return (B_FALSE);
}
static int
dump_label(const char *dev)
{
char path[MAXPATHLEN];
zdb_label_t labels[VDEV_LABELS] = {{{{0}}}};
uint64_t psize, ashift, l2cache;
struct stat64 statbuf;
boolean_t config_found = B_FALSE;
boolean_t error = B_FALSE;
boolean_t read_l2arc_header = B_FALSE;
avl_tree_t config_tree;
avl_tree_t uberblock_tree;
void *node, *cookie;
int fd;
/*
* Check if we were given absolute path and use it as is.
* Otherwise if the provided vdev name doesn't point to a file,
* try prepending expected disk paths and partition numbers.
*/
(void) strlcpy(path, dev, sizeof (path));
if (dev[0] != '/' && stat64(path, &statbuf) != 0) {
int error;
error = zfs_resolve_shortname(dev, path, MAXPATHLEN);
if (error == 0 && zfs_dev_is_whole_disk(path)) {
if (zfs_append_partition(path, MAXPATHLEN) == -1)
error = ENOENT;
}
if (error || (stat64(path, &statbuf) != 0)) {
(void) printf("failed to find device %s, try "
"specifying absolute path instead\n", dev);
return (1);
}
}
if ((fd = open64(path, O_RDONLY)) < 0) {
(void) printf("cannot open '%s': %s\n", path, strerror(errno));
exit(1);
}
if (fstat64_blk(fd, &statbuf) != 0) {
(void) printf("failed to stat '%s': %s\n", path,
strerror(errno));
(void) close(fd);
exit(1);
}
if (S_ISBLK(statbuf.st_mode) && zfs_dev_flush(fd) != 0)
(void) printf("failed to invalidate cache '%s' : %s\n", path,
strerror(errno));
avl_create(&config_tree, cksum_record_compare,
sizeof (cksum_record_t), offsetof(cksum_record_t, link));
avl_create(&uberblock_tree, cksum_record_compare,
sizeof (cksum_record_t), offsetof(cksum_record_t, link));
psize = statbuf.st_size;
psize = P2ALIGN(psize, (uint64_t)sizeof (vdev_label_t));
ashift = SPA_MINBLOCKSHIFT;
/*
* 1. Read the label from disk
* 2. Verify label cksum
* 3. Unpack the configuration and insert in config tree.
* 4. Traverse all uberblocks and insert in uberblock tree.
*/
for (int l = 0; l < VDEV_LABELS; l++) {
zdb_label_t *label = &labels[l];
char *buf = label->label.vl_vdev_phys.vp_nvlist;
size_t buflen = sizeof (label->label.vl_vdev_phys.vp_nvlist);
nvlist_t *config;
cksum_record_t *rec;
zio_cksum_t cksum;
vdev_t vd;
label->label_offset = vdev_label_offset(psize, l, 0);
if (pread64(fd, &label->label, sizeof (label->label),
label->label_offset) != sizeof (label->label)) {
if (!dump_opt['q'])
(void) printf("failed to read label %d\n", l);
label->read_failed = B_TRUE;
error = B_TRUE;
continue;
}
label->read_failed = B_FALSE;
label->cksum_valid = label_cksum_valid(&label->label,
label->label_offset);
if (nvlist_unpack(buf, buflen, &config, 0) == 0) {
nvlist_t *vdev_tree = NULL;
size_t size;
if ((nvlist_lookup_nvlist(config,
ZPOOL_CONFIG_VDEV_TREE, &vdev_tree) != 0) ||
(nvlist_lookup_uint64(vdev_tree,
ZPOOL_CONFIG_ASHIFT, &ashift) != 0))
ashift = SPA_MINBLOCKSHIFT;
if (nvlist_size(config, &size, NV_ENCODE_XDR) != 0)
size = buflen;
/* If the device is a cache device read the header. */
if (!read_l2arc_header) {
if (nvlist_lookup_uint64(config,
ZPOOL_CONFIG_POOL_STATE, &l2cache) == 0 &&
l2cache == POOL_STATE_L2CACHE) {
read_l2arc_header = B_TRUE;
}
}
fletcher_4_native_varsize(buf, size, &cksum);
rec = cksum_record_insert(&config_tree, &cksum, l);
label->config = rec;
label->config_nv = config;
config_found = B_TRUE;
} else {
error = B_TRUE;
}
vd.vdev_ashift = ashift;
vd.vdev_top = &vd;
for (int i = 0; i < VDEV_UBERBLOCK_COUNT(&vd); i++) {
uint64_t uoff = VDEV_UBERBLOCK_OFFSET(&vd, i);
uberblock_t *ub = (void *)((char *)label + uoff);
if (uberblock_verify(ub))
continue;
fletcher_4_native_varsize(ub, sizeof (*ub), &cksum);
rec = cksum_record_insert(&uberblock_tree, &cksum, l);
label->uberblocks[i] = rec;
}
}
/*
* Dump the label and uberblocks.
*/
for (int l = 0; l < VDEV_LABELS; l++) {
zdb_label_t *label = &labels[l];
size_t buflen = sizeof (label->label.vl_vdev_phys.vp_nvlist);
if (label->read_failed == B_TRUE)
continue;
if (label->config_nv) {
dump_config_from_label(label, buflen, l);
} else {
if (!dump_opt['q'])
(void) printf("failed to unpack label %d\n", l);
}
if (dump_opt['u'])
dump_label_uberblocks(label, ashift, l);
nvlist_free(label->config_nv);
}
/*
* Dump the L2ARC header, if existent.
*/
if (read_l2arc_header)
error |= dump_l2arc_header(fd);
cookie = NULL;
while ((node = avl_destroy_nodes(&config_tree, &cookie)) != NULL)
umem_free(node, sizeof (cksum_record_t));
cookie = NULL;
while ((node = avl_destroy_nodes(&uberblock_tree, &cookie)) != NULL)
umem_free(node, sizeof (cksum_record_t));
avl_destroy(&config_tree);
avl_destroy(&uberblock_tree);
(void) close(fd);
return (config_found == B_FALSE ? 2 :
(error == B_TRUE ? 1 : 0));
}
static uint64_t dataset_feature_count[SPA_FEATURES];
static uint64_t global_feature_count[SPA_FEATURES];
static uint64_t remap_deadlist_count = 0;
static int
dump_one_objset(const char *dsname, void *arg)
{
(void) arg;
int error;
objset_t *os;
spa_feature_t f;
error = open_objset(dsname, FTAG, &os);
if (error != 0)
return (0);
for (f = 0; f < SPA_FEATURES; f++) {
if (!dsl_dataset_feature_is_active(dmu_objset_ds(os), f))
continue;
ASSERT(spa_feature_table[f].fi_flags &
ZFEATURE_FLAG_PER_DATASET);
dataset_feature_count[f]++;
}
if (dsl_dataset_remap_deadlist_exists(dmu_objset_ds(os))) {
remap_deadlist_count++;
}
for (dsl_bookmark_node_t *dbn =
avl_first(&dmu_objset_ds(os)->ds_bookmarks); dbn != NULL;
dbn = AVL_NEXT(&dmu_objset_ds(os)->ds_bookmarks, dbn)) {
mos_obj_refd(dbn->dbn_phys.zbm_redaction_obj);
if (dbn->dbn_phys.zbm_redaction_obj != 0) {
global_feature_count[
SPA_FEATURE_REDACTION_BOOKMARKS]++;
objset_t *mos = os->os_spa->spa_meta_objset;
dnode_t *rl;
VERIFY0(dnode_hold(mos,
dbn->dbn_phys.zbm_redaction_obj, FTAG, &rl));
if (rl->dn_have_spill) {
global_feature_count[
SPA_FEATURE_REDACTION_LIST_SPILL]++;
}
}
if (dbn->dbn_phys.zbm_flags & ZBM_FLAG_HAS_FBN)
global_feature_count[SPA_FEATURE_BOOKMARK_WRITTEN]++;
}
if (dsl_deadlist_is_open(&dmu_objset_ds(os)->ds_dir->dd_livelist) &&
!dmu_objset_is_snapshot(os)) {
global_feature_count[SPA_FEATURE_LIVELIST]++;
}
dump_objset(os);
close_objset(os, FTAG);
fuid_table_destroy();
return (0);
}
/*
* Block statistics.
*/
#define PSIZE_HISTO_SIZE (SPA_OLD_MAXBLOCKSIZE / SPA_MINBLOCKSIZE + 2)
typedef struct zdb_blkstats {
uint64_t zb_asize;
uint64_t zb_lsize;
uint64_t zb_psize;
uint64_t zb_count;
uint64_t zb_gangs;
uint64_t zb_ditto_samevdev;
uint64_t zb_ditto_same_ms;
uint64_t zb_psize_histogram[PSIZE_HISTO_SIZE];
} zdb_blkstats_t;
/*
* Extended object types to report deferred frees and dedup auto-ditto blocks.
*/
#define ZDB_OT_DEFERRED (DMU_OT_NUMTYPES + 0)
#define ZDB_OT_DITTO (DMU_OT_NUMTYPES + 1)
#define ZDB_OT_OTHER (DMU_OT_NUMTYPES + 2)
#define ZDB_OT_TOTAL (DMU_OT_NUMTYPES + 3)
static const char *zdb_ot_extname[] = {
"deferred free",
"dedup ditto",
"other",
"Total",
};
#define ZB_TOTAL DN_MAX_LEVELS
#define SPA_MAX_FOR_16M (SPA_MAXBLOCKSHIFT+1)
typedef struct zdb_brt_entry {
dva_t zbre_dva;
uint64_t zbre_refcount;
avl_node_t zbre_node;
} zdb_brt_entry_t;
typedef struct zdb_cb {
zdb_blkstats_t zcb_type[ZB_TOTAL + 1][ZDB_OT_TOTAL + 1];
uint64_t zcb_removing_size;
uint64_t zcb_checkpoint_size;
uint64_t zcb_dedup_asize;
uint64_t zcb_dedup_blocks;
uint64_t zcb_clone_asize;
uint64_t zcb_clone_blocks;
uint64_t zcb_psize_count[SPA_MAX_FOR_16M];
uint64_t zcb_lsize_count[SPA_MAX_FOR_16M];
uint64_t zcb_asize_count[SPA_MAX_FOR_16M];
uint64_t zcb_psize_len[SPA_MAX_FOR_16M];
uint64_t zcb_lsize_len[SPA_MAX_FOR_16M];
uint64_t zcb_asize_len[SPA_MAX_FOR_16M];
uint64_t zcb_psize_total;
uint64_t zcb_lsize_total;
uint64_t zcb_asize_total;
uint64_t zcb_embedded_blocks[NUM_BP_EMBEDDED_TYPES];
uint64_t zcb_embedded_histogram[NUM_BP_EMBEDDED_TYPES]
[BPE_PAYLOAD_SIZE + 1];
uint64_t zcb_start;
hrtime_t zcb_lastprint;
uint64_t zcb_totalasize;
uint64_t zcb_errors[256];
int zcb_readfails;
int zcb_haderrors;
spa_t *zcb_spa;
uint32_t **zcb_vd_obsolete_counts;
avl_tree_t zcb_brt;
boolean_t zcb_brt_is_active;
} zdb_cb_t;
/* test if two DVA offsets from same vdev are within the same metaslab */
static boolean_t
same_metaslab(spa_t *spa, uint64_t vdev, uint64_t off1, uint64_t off2)
{
vdev_t *vd = vdev_lookup_top(spa, vdev);
uint64_t ms_shift = vd->vdev_ms_shift;
return ((off1 >> ms_shift) == (off2 >> ms_shift));
}
/*
* Used to simplify reporting of the histogram data.
*/
typedef struct one_histo {
const char *name;
uint64_t *count;
uint64_t *len;
uint64_t cumulative;
} one_histo_t;
/*
* The number of separate histograms processed for psize, lsize and asize.
*/
#define NUM_HISTO 3
/*
* This routine will create a fixed column size output of three different
* histograms showing by blocksize of 512 - 2^ SPA_MAX_FOR_16M
* the count, length and cumulative length of the psize, lsize and
* asize blocks.
*
* All three types of blocks are listed on a single line
*
* By default the table is printed in nicenumber format (e.g. 123K) but
* if the '-P' parameter is specified then the full raw number (parseable)
* is printed out.
*/
static void
dump_size_histograms(zdb_cb_t *zcb)
{
/*
* A temporary buffer that allows us to convert a number into
* a string using zdb_nicenumber to allow either raw or human
* readable numbers to be output.
*/
char numbuf[32];
/*
* Define titles which are used in the headers of the tables
* printed by this routine.
*/
const char blocksize_title1[] = "block";
const char blocksize_title2[] = "size";
const char count_title[] = "Count";
const char length_title[] = "Size";
const char cumulative_title[] = "Cum.";
/*
* Setup the histogram arrays (psize, lsize, and asize).
*/
one_histo_t parm_histo[NUM_HISTO];
parm_histo[0].name = "psize";
parm_histo[0].count = zcb->zcb_psize_count;
parm_histo[0].len = zcb->zcb_psize_len;
parm_histo[0].cumulative = 0;
parm_histo[1].name = "lsize";
parm_histo[1].count = zcb->zcb_lsize_count;
parm_histo[1].len = zcb->zcb_lsize_len;
parm_histo[1].cumulative = 0;
parm_histo[2].name = "asize";
parm_histo[2].count = zcb->zcb_asize_count;
parm_histo[2].len = zcb->zcb_asize_len;
parm_histo[2].cumulative = 0;
(void) printf("\nBlock Size Histogram\n");
/*
* Print the first line titles
*/
if (dump_opt['P'])
(void) printf("\n%s\t", blocksize_title1);
else
(void) printf("\n%7s ", blocksize_title1);
for (int j = 0; j < NUM_HISTO; j++) {
if (dump_opt['P']) {
if (j < NUM_HISTO - 1) {
(void) printf("%s\t\t\t", parm_histo[j].name);
} else {
/* Don't print trailing spaces */
(void) printf(" %s", parm_histo[j].name);
}
} else {
if (j < NUM_HISTO - 1) {
/* Left aligned strings in the output */
(void) printf("%-7s ",
parm_histo[j].name);
} else {
/* Don't print trailing spaces */
(void) printf("%s", parm_histo[j].name);
}
}
}
(void) printf("\n");
/*
* Print the second line titles
*/
if (dump_opt['P']) {
(void) printf("%s\t", blocksize_title2);
} else {
(void) printf("%7s ", blocksize_title2);
}
for (int i = 0; i < NUM_HISTO; i++) {
if (dump_opt['P']) {
(void) printf("%s\t%s\t%s\t",
count_title, length_title, cumulative_title);
} else {
(void) printf("%7s%7s%7s",
count_title, length_title, cumulative_title);
}
}
(void) printf("\n");
/*
* Print the rows
*/
for (int i = SPA_MINBLOCKSHIFT; i < SPA_MAX_FOR_16M; i++) {
/*
* Print the first column showing the blocksize
*/
zdb_nicenum((1ULL << i), numbuf, sizeof (numbuf));
if (dump_opt['P']) {
printf("%s", numbuf);
} else {
printf("%7s:", numbuf);
}
/*
* Print the remaining set of 3 columns per size:
* for psize, lsize and asize
*/
for (int j = 0; j < NUM_HISTO; j++) {
parm_histo[j].cumulative += parm_histo[j].len[i];
zdb_nicenum(parm_histo[j].count[i],
numbuf, sizeof (numbuf));
if (dump_opt['P'])
(void) printf("\t%s", numbuf);
else
(void) printf("%7s", numbuf);
zdb_nicenum(parm_histo[j].len[i],
numbuf, sizeof (numbuf));
if (dump_opt['P'])
(void) printf("\t%s", numbuf);
else
(void) printf("%7s", numbuf);
zdb_nicenum(parm_histo[j].cumulative,
numbuf, sizeof (numbuf));
if (dump_opt['P'])
(void) printf("\t%s", numbuf);
else
(void) printf("%7s", numbuf);
}
(void) printf("\n");
}
}
static void
zdb_count_block(zdb_cb_t *zcb, zilog_t *zilog, const blkptr_t *bp,
dmu_object_type_t type)
{
uint64_t refcnt = 0;
int i;
ASSERT(type < ZDB_OT_TOTAL);
if (zilog && zil_bp_tree_add(zilog, bp) != 0)
return;
spa_config_enter(zcb->zcb_spa, SCL_CONFIG, FTAG, RW_READER);
for (i = 0; i < 4; i++) {
int l = (i < 2) ? BP_GET_LEVEL(bp) : ZB_TOTAL;
int t = (i & 1) ? type : ZDB_OT_TOTAL;
int equal;
zdb_blkstats_t *zb = &zcb->zcb_type[l][t];
zb->zb_asize += BP_GET_ASIZE(bp);
zb->zb_lsize += BP_GET_LSIZE(bp);
zb->zb_psize += BP_GET_PSIZE(bp);
zb->zb_count++;
/*
* The histogram is only big enough to record blocks up to
* SPA_OLD_MAXBLOCKSIZE; larger blocks go into the last,
* "other", bucket.
*/
unsigned idx = BP_GET_PSIZE(bp) >> SPA_MINBLOCKSHIFT;
idx = MIN(idx, SPA_OLD_MAXBLOCKSIZE / SPA_MINBLOCKSIZE + 1);
zb->zb_psize_histogram[idx]++;
zb->zb_gangs += BP_COUNT_GANG(bp);
switch (BP_GET_NDVAS(bp)) {
case 2:
if (DVA_GET_VDEV(&bp->blk_dva[0]) ==
DVA_GET_VDEV(&bp->blk_dva[1])) {
zb->zb_ditto_samevdev++;
if (same_metaslab(zcb->zcb_spa,
DVA_GET_VDEV(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[1])))
zb->zb_ditto_same_ms++;
}
break;
case 3:
equal = (DVA_GET_VDEV(&bp->blk_dva[0]) ==
DVA_GET_VDEV(&bp->blk_dva[1])) +
(DVA_GET_VDEV(&bp->blk_dva[0]) ==
DVA_GET_VDEV(&bp->blk_dva[2])) +
(DVA_GET_VDEV(&bp->blk_dva[1]) ==
DVA_GET_VDEV(&bp->blk_dva[2]));
if (equal != 0) {
zb->zb_ditto_samevdev++;
if (DVA_GET_VDEV(&bp->blk_dva[0]) ==
DVA_GET_VDEV(&bp->blk_dva[1]) &&
same_metaslab(zcb->zcb_spa,
DVA_GET_VDEV(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[1])))
zb->zb_ditto_same_ms++;
else if (DVA_GET_VDEV(&bp->blk_dva[0]) ==
DVA_GET_VDEV(&bp->blk_dva[2]) &&
same_metaslab(zcb->zcb_spa,
DVA_GET_VDEV(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[2])))
zb->zb_ditto_same_ms++;
else if (DVA_GET_VDEV(&bp->blk_dva[1]) ==
DVA_GET_VDEV(&bp->blk_dva[2]) &&
same_metaslab(zcb->zcb_spa,
DVA_GET_VDEV(&bp->blk_dva[1]),
DVA_GET_OFFSET(&bp->blk_dva[1]),
DVA_GET_OFFSET(&bp->blk_dva[2])))
zb->zb_ditto_same_ms++;
}
break;
}
}
spa_config_exit(zcb->zcb_spa, SCL_CONFIG, FTAG);
if (BP_IS_EMBEDDED(bp)) {
zcb->zcb_embedded_blocks[BPE_GET_ETYPE(bp)]++;
zcb->zcb_embedded_histogram[BPE_GET_ETYPE(bp)]
[BPE_GET_PSIZE(bp)]++;
return;
}
/*
* The binning histogram bins by powers of two up to
* SPA_MAXBLOCKSIZE rather than creating bins for
* every possible blocksize found in the pool.
*/
int bin = highbit64(BP_GET_PSIZE(bp)) - 1;
zcb->zcb_psize_count[bin]++;
zcb->zcb_psize_len[bin] += BP_GET_PSIZE(bp);
zcb->zcb_psize_total += BP_GET_PSIZE(bp);
bin = highbit64(BP_GET_LSIZE(bp)) - 1;
zcb->zcb_lsize_count[bin]++;
zcb->zcb_lsize_len[bin] += BP_GET_LSIZE(bp);
zcb->zcb_lsize_total += BP_GET_LSIZE(bp);
bin = highbit64(BP_GET_ASIZE(bp)) - 1;
zcb->zcb_asize_count[bin]++;
zcb->zcb_asize_len[bin] += BP_GET_ASIZE(bp);
zcb->zcb_asize_total += BP_GET_ASIZE(bp);
if (zcb->zcb_brt_is_active && brt_maybe_exists(zcb->zcb_spa, bp)) {
/*
* Cloned blocks are special. We need to count them, so we can
* later uncount them when reporting leaked space, and we must
* only claim them them once.
*
* To do this, we keep our own in-memory BRT. For each block
* we haven't seen before, we look it up in the real BRT and
* if its there, we note it and its refcount then proceed as
* normal. If we see the block again, we count it as a clone
* and then give it no further consideration.
*/
zdb_brt_entry_t zbre_search, *zbre;
avl_index_t where;
zbre_search.zbre_dva = bp->blk_dva[0];
zbre = avl_find(&zcb->zcb_brt, &zbre_search, &where);
if (zbre != NULL) {
zcb->zcb_clone_asize += BP_GET_ASIZE(bp);
zcb->zcb_clone_blocks++;
zbre->zbre_refcount--;
if (zbre->zbre_refcount == 0) {
avl_remove(&zcb->zcb_brt, zbre);
umem_free(zbre, sizeof (zdb_brt_entry_t));
}
return;
}
uint64_t crefcnt = brt_entry_get_refcount(zcb->zcb_spa, bp);
if (crefcnt > 0) {
zbre = umem_zalloc(sizeof (zdb_brt_entry_t),
UMEM_NOFAIL);
zbre->zbre_dva = bp->blk_dva[0];
zbre->zbre_refcount = crefcnt;
avl_insert(&zcb->zcb_brt, zbre, where);
}
}
if (dump_opt['L'])
return;
if (BP_GET_DEDUP(bp)) {
ddt_t *ddt;
ddt_entry_t *dde;
ddt = ddt_select(zcb->zcb_spa, bp);
ddt_enter(ddt);
dde = ddt_lookup(ddt, bp, B_FALSE);
if (dde == NULL) {
refcnt = 0;
} else {
ddt_phys_t *ddp = ddt_phys_select(dde, bp);
ddt_phys_decref(ddp);
refcnt = ddp->ddp_refcnt;
if (ddt_phys_total_refcnt(dde) == 0)
ddt_remove(ddt, dde);
}
ddt_exit(ddt);
}
VERIFY3U(zio_wait(zio_claim(NULL, zcb->zcb_spa,
refcnt ? 0 : spa_min_claim_txg(zcb->zcb_spa),
bp, NULL, NULL, ZIO_FLAG_CANFAIL)), ==, 0);
}
static void
zdb_blkptr_done(zio_t *zio)
{
spa_t *spa = zio->io_spa;
blkptr_t *bp = zio->io_bp;
int ioerr = zio->io_error;
zdb_cb_t *zcb = zio->io_private;
zbookmark_phys_t *zb = &zio->io_bookmark;
mutex_enter(&spa->spa_scrub_lock);
spa->spa_load_verify_bytes -= BP_GET_PSIZE(bp);
cv_broadcast(&spa->spa_scrub_io_cv);
if (ioerr && !(zio->io_flags & ZIO_FLAG_SPECULATIVE)) {
char blkbuf[BP_SPRINTF_LEN];
zcb->zcb_haderrors = 1;
zcb->zcb_errors[ioerr]++;
if (dump_opt['b'] >= 2)
snprintf_blkptr(blkbuf, sizeof (blkbuf), bp);
else
blkbuf[0] = '\0';
(void) printf("zdb_blkptr_cb: "
"Got error %d reading "
"<%llu, %llu, %lld, %llx> %s -- skipping\n",
ioerr,
(u_longlong_t)zb->zb_objset,
(u_longlong_t)zb->zb_object,
(u_longlong_t)zb->zb_level,
(u_longlong_t)zb->zb_blkid,
blkbuf);
}
mutex_exit(&spa->spa_scrub_lock);
abd_free(zio->io_abd);
}
static int
zdb_blkptr_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
const zbookmark_phys_t *zb, const dnode_phys_t *dnp, void *arg)
{
zdb_cb_t *zcb = arg;
dmu_object_type_t type;
boolean_t is_metadata;
if (zb->zb_level == ZB_DNODE_LEVEL)
return (0);
if (dump_opt['b'] >= 5 && bp->blk_birth > 0) {
char blkbuf[BP_SPRINTF_LEN];
snprintf_blkptr(blkbuf, sizeof (blkbuf), bp);
(void) printf("objset %llu object %llu "
"level %lld offset 0x%llx %s\n",
(u_longlong_t)zb->zb_objset,
(u_longlong_t)zb->zb_object,
(longlong_t)zb->zb_level,
(u_longlong_t)blkid2offset(dnp, bp, zb),
blkbuf);
}
if (BP_IS_HOLE(bp) || BP_IS_REDACTED(bp))
return (0);
type = BP_GET_TYPE(bp);
zdb_count_block(zcb, zilog, bp,
(type & DMU_OT_NEWTYPE) ? ZDB_OT_OTHER : type);
is_metadata = (BP_GET_LEVEL(bp) != 0 || DMU_OT_IS_METADATA(type));
if (!BP_IS_EMBEDDED(bp) &&
(dump_opt['c'] > 1 || (dump_opt['c'] && is_metadata))) {
size_t size = BP_GET_PSIZE(bp);
abd_t *abd = abd_alloc(size, B_FALSE);
int flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCRUB | ZIO_FLAG_RAW;
/* If it's an intent log block, failure is expected. */
if (zb->zb_level == ZB_ZIL_LEVEL)
flags |= ZIO_FLAG_SPECULATIVE;
mutex_enter(&spa->spa_scrub_lock);
while (spa->spa_load_verify_bytes > max_inflight_bytes)
cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock);
spa->spa_load_verify_bytes += size;
mutex_exit(&spa->spa_scrub_lock);
zio_nowait(zio_read(NULL, spa, bp, abd, size,
zdb_blkptr_done, zcb, ZIO_PRIORITY_ASYNC_READ, flags, zb));
}
zcb->zcb_readfails = 0;
/* only call gethrtime() every 100 blocks */
static int iters;
if (++iters > 100)
iters = 0;
else
return (0);
if (dump_opt['b'] < 5 && gethrtime() > zcb->zcb_lastprint + NANOSEC) {
uint64_t now = gethrtime();
char buf[10];
uint64_t bytes = zcb->zcb_type[ZB_TOTAL][ZDB_OT_TOTAL].zb_asize;
uint64_t kb_per_sec =
1 + bytes / (1 + ((now - zcb->zcb_start) / 1000 / 1000));
uint64_t sec_remaining =
(zcb->zcb_totalasize - bytes) / 1024 / kb_per_sec;
/* make sure nicenum has enough space */
_Static_assert(sizeof (buf) >= NN_NUMBUF_SZ, "buf truncated");
zfs_nicebytes(bytes, buf, sizeof (buf));
(void) fprintf(stderr,
"\r%5s completed (%4"PRIu64"MB/s) "
"estimated time remaining: "
"%"PRIu64"hr %02"PRIu64"min %02"PRIu64"sec ",
buf, kb_per_sec / 1024,
sec_remaining / 60 / 60,
sec_remaining / 60 % 60,
sec_remaining % 60);
zcb->zcb_lastprint = now;
}
return (0);
}
static void
zdb_leak(void *arg, uint64_t start, uint64_t size)
{
vdev_t *vd = arg;
(void) printf("leaked space: vdev %llu, offset 0x%llx, size %llu\n",
(u_longlong_t)vd->vdev_id, (u_longlong_t)start, (u_longlong_t)size);
}
static metaslab_ops_t zdb_metaslab_ops = {
NULL /* alloc */
};
static int
load_unflushed_svr_segs_cb(spa_t *spa, space_map_entry_t *sme,
uint64_t txg, void *arg)
{
spa_vdev_removal_t *svr = arg;
uint64_t offset = sme->sme_offset;
uint64_t size = sme->sme_run;
/* skip vdevs we don't care about */
if (sme->sme_vdev != svr->svr_vdev_id)
return (0);
vdev_t *vd = vdev_lookup_top(spa, sme->sme_vdev);
metaslab_t *ms = vd->vdev_ms[offset >> vd->vdev_ms_shift];
ASSERT(sme->sme_type == SM_ALLOC || sme->sme_type == SM_FREE);
if (txg < metaslab_unflushed_txg(ms))
return (0);
if (sme->sme_type == SM_ALLOC)
range_tree_add(svr->svr_allocd_segs, offset, size);
else
range_tree_remove(svr->svr_allocd_segs, offset, size);
return (0);
}
static void
claim_segment_impl_cb(uint64_t inner_offset, vdev_t *vd, uint64_t offset,
uint64_t size, void *arg)
{
(void) inner_offset, (void) arg;
/*
* This callback was called through a remap from
* a device being removed. Therefore, the vdev that
* this callback is applied to is a concrete
* vdev.
*/
ASSERT(vdev_is_concrete(vd));
VERIFY0(metaslab_claim_impl(vd, offset, size,
spa_min_claim_txg(vd->vdev_spa)));
}
static void
claim_segment_cb(void *arg, uint64_t offset, uint64_t size)
{
vdev_t *vd = arg;
vdev_indirect_ops.vdev_op_remap(vd, offset, size,
claim_segment_impl_cb, NULL);
}
/*
* After accounting for all allocated blocks that are directly referenced,
* we might have missed a reference to a block from a partially complete
* (and thus unused) indirect mapping object. We perform a secondary pass
* through the metaslabs we have already mapped and claim the destination
* blocks.
*/
static void
zdb_claim_removing(spa_t *spa, zdb_cb_t *zcb)
{
if (dump_opt['L'])
return;
if (spa->spa_vdev_removal == NULL)
return;
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
spa_vdev_removal_t *svr = spa->spa_vdev_removal;
vdev_t *vd = vdev_lookup_top(spa, svr->svr_vdev_id);
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
ASSERT0(range_tree_space(svr->svr_allocd_segs));
range_tree_t *allocs = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0);
for (uint64_t msi = 0; msi < vd->vdev_ms_count; msi++) {
metaslab_t *msp = vd->vdev_ms[msi];
ASSERT0(range_tree_space(allocs));
if (msp->ms_sm != NULL)
VERIFY0(space_map_load(msp->ms_sm, allocs, SM_ALLOC));
range_tree_vacate(allocs, range_tree_add, svr->svr_allocd_segs);
}
range_tree_destroy(allocs);
iterate_through_spacemap_logs(spa, load_unflushed_svr_segs_cb, svr);
/*
* Clear everything past what has been synced,
* because we have not allocated mappings for
* it yet.
*/
range_tree_clear(svr->svr_allocd_segs,
vdev_indirect_mapping_max_offset(vim),
vd->vdev_asize - vdev_indirect_mapping_max_offset(vim));
zcb->zcb_removing_size += range_tree_space(svr->svr_allocd_segs);
range_tree_vacate(svr->svr_allocd_segs, claim_segment_cb, vd);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
static int
increment_indirect_mapping_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
(void) tx;
zdb_cb_t *zcb = arg;
spa_t *spa = zcb->zcb_spa;
vdev_t *vd;
const dva_t *dva = &bp->blk_dva[0];
ASSERT(!bp_freed);
ASSERT(!dump_opt['L']);
ASSERT3U(BP_GET_NDVAS(bp), ==, 1);
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
vd = vdev_lookup_top(zcb->zcb_spa, DVA_GET_VDEV(dva));
ASSERT3P(vd, !=, NULL);
spa_config_exit(spa, SCL_VDEV, FTAG);
ASSERT(vd->vdev_indirect_config.vic_mapping_object != 0);
ASSERT3P(zcb->zcb_vd_obsolete_counts[vd->vdev_id], !=, NULL);
vdev_indirect_mapping_increment_obsolete_count(
vd->vdev_indirect_mapping,
DVA_GET_OFFSET(dva), DVA_GET_ASIZE(dva),
zcb->zcb_vd_obsolete_counts[vd->vdev_id]);
return (0);
}
static uint32_t *
zdb_load_obsolete_counts(vdev_t *vd)
{
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
spa_t *spa = vd->vdev_spa;
spa_condensing_indirect_phys_t *scip =
&spa->spa_condensing_indirect_phys;
uint64_t obsolete_sm_object;
uint32_t *counts;
VERIFY0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
EQUIV(obsolete_sm_object != 0, vd->vdev_obsolete_sm != NULL);
counts = vdev_indirect_mapping_load_obsolete_counts(vim);
if (vd->vdev_obsolete_sm != NULL) {
vdev_indirect_mapping_load_obsolete_spacemap(vim, counts,
vd->vdev_obsolete_sm);
}
if (scip->scip_vdev == vd->vdev_id &&
scip->scip_prev_obsolete_sm_object != 0) {
space_map_t *prev_obsolete_sm = NULL;
VERIFY0(space_map_open(&prev_obsolete_sm, spa->spa_meta_objset,
scip->scip_prev_obsolete_sm_object, 0, vd->vdev_asize, 0));
vdev_indirect_mapping_load_obsolete_spacemap(vim, counts,
prev_obsolete_sm);
space_map_close(prev_obsolete_sm);
}
return (counts);
}
static void
zdb_ddt_leak_init(spa_t *spa, zdb_cb_t *zcb)
{
ddt_bookmark_t ddb = {0};
ddt_entry_t dde;
int error;
int p;
ASSERT(!dump_opt['L']);
while ((error = ddt_walk(spa, &ddb, &dde)) == 0) {
blkptr_t blk;
ddt_phys_t *ddp = dde.dde_phys;
if (ddb.ddb_class == DDT_CLASS_UNIQUE)
return;
ASSERT(ddt_phys_total_refcnt(&dde) > 1);
for (p = 0; p < DDT_PHYS_TYPES; p++, ddp++) {
if (ddp->ddp_phys_birth == 0)
continue;
ddt_bp_create(ddb.ddb_checksum,
&dde.dde_key, ddp, &blk);
if (p == DDT_PHYS_DITTO) {
zdb_count_block(zcb, NULL, &blk, ZDB_OT_DITTO);
} else {
zcb->zcb_dedup_asize +=
BP_GET_ASIZE(&blk) * (ddp->ddp_refcnt - 1);
zcb->zcb_dedup_blocks++;
}
}
ddt_t *ddt = spa->spa_ddt[ddb.ddb_checksum];
ddt_enter(ddt);
VERIFY(ddt_lookup(ddt, &blk, B_TRUE) != NULL);
ddt_exit(ddt);
}
ASSERT(error == ENOENT);
}
typedef struct checkpoint_sm_exclude_entry_arg {
vdev_t *cseea_vd;
uint64_t cseea_checkpoint_size;
} checkpoint_sm_exclude_entry_arg_t;
static int
checkpoint_sm_exclude_entry_cb(space_map_entry_t *sme, void *arg)
{
checkpoint_sm_exclude_entry_arg_t *cseea = arg;
vdev_t *vd = cseea->cseea_vd;
metaslab_t *ms = vd->vdev_ms[sme->sme_offset >> vd->vdev_ms_shift];
uint64_t end = sme->sme_offset + sme->sme_run;
ASSERT(sme->sme_type == SM_FREE);
/*
* Since the vdev_checkpoint_sm exists in the vdev level
* and the ms_sm space maps exist in the metaslab level,
* an entry in the checkpoint space map could theoretically
* cross the boundaries of the metaslab that it belongs.
*
* In reality, because of the way that we populate and
* manipulate the checkpoint's space maps currently,
* there shouldn't be any entries that cross metaslabs.
* Hence the assertion below.
*
* That said, there is no fundamental requirement that
* the checkpoint's space map entries should not cross
* metaslab boundaries. So if needed we could add code
* that handles metaslab-crossing segments in the future.
*/
VERIFY3U(sme->sme_offset, >=, ms->ms_start);
VERIFY3U(end, <=, ms->ms_start + ms->ms_size);
/*
* By removing the entry from the allocated segments we
* also verify that the entry is there to begin with.
*/
mutex_enter(&ms->ms_lock);
range_tree_remove(ms->ms_allocatable, sme->sme_offset, sme->sme_run);
mutex_exit(&ms->ms_lock);
cseea->cseea_checkpoint_size += sme->sme_run;
return (0);
}
static void
zdb_leak_init_vdev_exclude_checkpoint(vdev_t *vd, zdb_cb_t *zcb)
{
spa_t *spa = vd->vdev_spa;
space_map_t *checkpoint_sm = NULL;
uint64_t checkpoint_sm_obj;
/*
* If there is no vdev_top_zap, we are in a pool whose
* version predates the pool checkpoint feature.
*/
if (vd->vdev_top_zap == 0)
return;
/*
* If there is no reference of the vdev_checkpoint_sm in
* the vdev_top_zap, then one of the following scenarios
* is true:
*
* 1] There is no checkpoint
* 2] There is a checkpoint, but no checkpointed blocks
* have been freed yet
* 3] The current vdev is indirect
*
* In these cases we return immediately.
*/
if (zap_contains(spa_meta_objset(spa), vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM) != 0)
return;
VERIFY0(zap_lookup(spa_meta_objset(spa), vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM, sizeof (uint64_t), 1,
&checkpoint_sm_obj));
checkpoint_sm_exclude_entry_arg_t cseea;
cseea.cseea_vd = vd;
cseea.cseea_checkpoint_size = 0;
VERIFY0(space_map_open(&checkpoint_sm, spa_meta_objset(spa),
checkpoint_sm_obj, 0, vd->vdev_asize, vd->vdev_ashift));
VERIFY0(space_map_iterate(checkpoint_sm,
space_map_length(checkpoint_sm),
checkpoint_sm_exclude_entry_cb, &cseea));
space_map_close(checkpoint_sm);
zcb->zcb_checkpoint_size += cseea.cseea_checkpoint_size;
}
static void
zdb_leak_init_exclude_checkpoint(spa_t *spa, zdb_cb_t *zcb)
{
ASSERT(!dump_opt['L']);
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
ASSERT3U(c, ==, rvd->vdev_child[c]->vdev_id);
zdb_leak_init_vdev_exclude_checkpoint(rvd->vdev_child[c], zcb);
}
}
static int
count_unflushed_space_cb(spa_t *spa, space_map_entry_t *sme,
uint64_t txg, void *arg)
{
int64_t *ualloc_space = arg;
uint64_t offset = sme->sme_offset;
uint64_t vdev_id = sme->sme_vdev;
vdev_t *vd = vdev_lookup_top(spa, vdev_id);
if (!vdev_is_concrete(vd))
return (0);
metaslab_t *ms = vd->vdev_ms[offset >> vd->vdev_ms_shift];
ASSERT(sme->sme_type == SM_ALLOC || sme->sme_type == SM_FREE);
if (txg < metaslab_unflushed_txg(ms))
return (0);
if (sme->sme_type == SM_ALLOC)
*ualloc_space += sme->sme_run;
else
*ualloc_space -= sme->sme_run;
return (0);
}
static int64_t
get_unflushed_alloc_space(spa_t *spa)
{
if (dump_opt['L'])
return (0);
int64_t ualloc_space = 0;
iterate_through_spacemap_logs(spa, count_unflushed_space_cb,
&ualloc_space);
return (ualloc_space);
}
static int
load_unflushed_cb(spa_t *spa, space_map_entry_t *sme, uint64_t txg, void *arg)
{
maptype_t *uic_maptype = arg;
uint64_t offset = sme->sme_offset;
uint64_t size = sme->sme_run;
uint64_t vdev_id = sme->sme_vdev;
vdev_t *vd = vdev_lookup_top(spa, vdev_id);
/* skip indirect vdevs */
if (!vdev_is_concrete(vd))
return (0);
metaslab_t *ms = vd->vdev_ms[offset >> vd->vdev_ms_shift];
ASSERT(sme->sme_type == SM_ALLOC || sme->sme_type == SM_FREE);
ASSERT(*uic_maptype == SM_ALLOC || *uic_maptype == SM_FREE);
if (txg < metaslab_unflushed_txg(ms))
return (0);
if (*uic_maptype == sme->sme_type)
range_tree_add(ms->ms_allocatable, offset, size);
else
range_tree_remove(ms->ms_allocatable, offset, size);
return (0);
}
static void
load_unflushed_to_ms_allocatables(spa_t *spa, maptype_t maptype)
{
iterate_through_spacemap_logs(spa, load_unflushed_cb, &maptype);
}
static void
load_concrete_ms_allocatable_trees(spa_t *spa, maptype_t maptype)
{
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t i = 0; i < rvd->vdev_children; i++) {
vdev_t *vd = rvd->vdev_child[i];
ASSERT3U(i, ==, vd->vdev_id);
if (vd->vdev_ops == &vdev_indirect_ops)
continue;
for (uint64_t m = 0; m < vd->vdev_ms_count; m++) {
metaslab_t *msp = vd->vdev_ms[m];
(void) fprintf(stderr,
"\rloading concrete vdev %llu, "
"metaslab %llu of %llu ...",
(longlong_t)vd->vdev_id,
(longlong_t)msp->ms_id,
(longlong_t)vd->vdev_ms_count);
mutex_enter(&msp->ms_lock);
range_tree_vacate(msp->ms_allocatable, NULL, NULL);
/*
* We don't want to spend the CPU manipulating the
* size-ordered tree, so clear the range_tree ops.
*/
msp->ms_allocatable->rt_ops = NULL;
if (msp->ms_sm != NULL) {
VERIFY0(space_map_load(msp->ms_sm,
msp->ms_allocatable, maptype));
}
if (!msp->ms_loaded)
msp->ms_loaded = B_TRUE;
mutex_exit(&msp->ms_lock);
}
}
load_unflushed_to_ms_allocatables(spa, maptype);
}
/*
* vm_idxp is an in-out parameter which (for indirect vdevs) is the
* index in vim_entries that has the first entry in this metaslab.
* On return, it will be set to the first entry after this metaslab.
*/
static void
load_indirect_ms_allocatable_tree(vdev_t *vd, metaslab_t *msp,
uint64_t *vim_idxp)
{
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
mutex_enter(&msp->ms_lock);
range_tree_vacate(msp->ms_allocatable, NULL, NULL);
/*
* We don't want to spend the CPU manipulating the
* size-ordered tree, so clear the range_tree ops.
*/
msp->ms_allocatable->rt_ops = NULL;
for (; *vim_idxp < vdev_indirect_mapping_num_entries(vim);
(*vim_idxp)++) {
vdev_indirect_mapping_entry_phys_t *vimep =
&vim->vim_entries[*vim_idxp];
uint64_t ent_offset = DVA_MAPPING_GET_SRC_OFFSET(vimep);
uint64_t ent_len = DVA_GET_ASIZE(&vimep->vimep_dst);
ASSERT3U(ent_offset, >=, msp->ms_start);
if (ent_offset >= msp->ms_start + msp->ms_size)
break;
/*
* Mappings do not cross metaslab boundaries,
* because we create them by walking the metaslabs.
*/
ASSERT3U(ent_offset + ent_len, <=,
msp->ms_start + msp->ms_size);
range_tree_add(msp->ms_allocatable, ent_offset, ent_len);
}
if (!msp->ms_loaded)
msp->ms_loaded = B_TRUE;
mutex_exit(&msp->ms_lock);
}
static void
zdb_leak_init_prepare_indirect_vdevs(spa_t *spa, zdb_cb_t *zcb)
{
ASSERT(!dump_opt['L']);
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
ASSERT3U(c, ==, vd->vdev_id);
if (vd->vdev_ops != &vdev_indirect_ops)
continue;
/*
* Note: we don't check for mapping leaks on
* removing vdevs because their ms_allocatable's
* are used to look for leaks in allocated space.
*/
zcb->zcb_vd_obsolete_counts[c] = zdb_load_obsolete_counts(vd);
/*
* Normally, indirect vdevs don't have any
* metaslabs. We want to set them up for
* zio_claim().
*/
vdev_metaslab_group_create(vd);
VERIFY0(vdev_metaslab_init(vd, 0));
vdev_indirect_mapping_t *vim __maybe_unused =
vd->vdev_indirect_mapping;
uint64_t vim_idx = 0;
for (uint64_t m = 0; m < vd->vdev_ms_count; m++) {
(void) fprintf(stderr,
"\rloading indirect vdev %llu, "
"metaslab %llu of %llu ...",
(longlong_t)vd->vdev_id,
(longlong_t)vd->vdev_ms[m]->ms_id,
(longlong_t)vd->vdev_ms_count);
load_indirect_ms_allocatable_tree(vd, vd->vdev_ms[m],
&vim_idx);
}
ASSERT3U(vim_idx, ==, vdev_indirect_mapping_num_entries(vim));
}
}
static void
zdb_leak_init(spa_t *spa, zdb_cb_t *zcb)
{
zcb->zcb_spa = spa;
if (dump_opt['L'])
return;
dsl_pool_t *dp = spa->spa_dsl_pool;
vdev_t *rvd = spa->spa_root_vdev;
/*
* We are going to be changing the meaning of the metaslab's
* ms_allocatable. Ensure that the allocator doesn't try to
* use the tree.
*/
spa->spa_normal_class->mc_ops = &zdb_metaslab_ops;
spa->spa_log_class->mc_ops = &zdb_metaslab_ops;
spa->spa_embedded_log_class->mc_ops = &zdb_metaslab_ops;
zcb->zcb_vd_obsolete_counts =
umem_zalloc(rvd->vdev_children * sizeof (uint32_t *),
UMEM_NOFAIL);
/*
* For leak detection, we overload the ms_allocatable trees
* to contain allocated segments instead of free segments.
* As a result, we can't use the normal metaslab_load/unload
* interfaces.
*/
zdb_leak_init_prepare_indirect_vdevs(spa, zcb);
load_concrete_ms_allocatable_trees(spa, SM_ALLOC);
/*
* On load_concrete_ms_allocatable_trees() we loaded all the
* allocated entries from the ms_sm to the ms_allocatable for
* each metaslab. If the pool has a checkpoint or is in the
* middle of discarding a checkpoint, some of these blocks
* may have been freed but their ms_sm may not have been
* updated because they are referenced by the checkpoint. In
* order to avoid false-positives during leak-detection, we
* go through the vdev's checkpoint space map and exclude all
* its entries from their relevant ms_allocatable.
*
* We also aggregate the space held by the checkpoint and add
* it to zcb_checkpoint_size.
*
* Note that at this point we are also verifying that all the
* entries on the checkpoint_sm are marked as allocated in
* the ms_sm of their relevant metaslab.
* [see comment in checkpoint_sm_exclude_entry_cb()]
*/
zdb_leak_init_exclude_checkpoint(spa, zcb);
ASSERT3U(zcb->zcb_checkpoint_size, ==, spa_get_checkpoint_space(spa));
/* for cleaner progress output */
(void) fprintf(stderr, "\n");
if (bpobj_is_open(&dp->dp_obsolete_bpobj)) {
ASSERT(spa_feature_is_enabled(spa,
SPA_FEATURE_DEVICE_REMOVAL));
(void) bpobj_iterate_nofree(&dp->dp_obsolete_bpobj,
increment_indirect_mapping_cb, zcb, NULL);
}
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
zdb_ddt_leak_init(spa, zcb);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
static boolean_t
zdb_check_for_obsolete_leaks(vdev_t *vd, zdb_cb_t *zcb)
{
boolean_t leaks = B_FALSE;
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
uint64_t total_leaked = 0;
boolean_t are_precise = B_FALSE;
ASSERT(vim != NULL);
for (uint64_t i = 0; i < vdev_indirect_mapping_num_entries(vim); i++) {
vdev_indirect_mapping_entry_phys_t *vimep =
&vim->vim_entries[i];
uint64_t obsolete_bytes = 0;
uint64_t offset = DVA_MAPPING_GET_SRC_OFFSET(vimep);
metaslab_t *msp = vd->vdev_ms[offset >> vd->vdev_ms_shift];
/*
* This is not very efficient but it's easy to
* verify correctness.
*/
for (uint64_t inner_offset = 0;
inner_offset < DVA_GET_ASIZE(&vimep->vimep_dst);
inner_offset += 1ULL << vd->vdev_ashift) {
if (range_tree_contains(msp->ms_allocatable,
offset + inner_offset, 1ULL << vd->vdev_ashift)) {
obsolete_bytes += 1ULL << vd->vdev_ashift;
}
}
int64_t bytes_leaked = obsolete_bytes -
zcb->zcb_vd_obsolete_counts[vd->vdev_id][i];
ASSERT3U(DVA_GET_ASIZE(&vimep->vimep_dst), >=,
zcb->zcb_vd_obsolete_counts[vd->vdev_id][i]);
VERIFY0(vdev_obsolete_counts_are_precise(vd, &are_precise));
if (bytes_leaked != 0 && (are_precise || dump_opt['d'] >= 5)) {
(void) printf("obsolete indirect mapping count "
"mismatch on %llu:%llx:%llx : %llx bytes leaked\n",
(u_longlong_t)vd->vdev_id,
(u_longlong_t)DVA_MAPPING_GET_SRC_OFFSET(vimep),
(u_longlong_t)DVA_GET_ASIZE(&vimep->vimep_dst),
(u_longlong_t)bytes_leaked);
}
total_leaked += ABS(bytes_leaked);
}
VERIFY0(vdev_obsolete_counts_are_precise(vd, &are_precise));
if (!are_precise && total_leaked > 0) {
int pct_leaked = total_leaked * 100 /
vdev_indirect_mapping_bytes_mapped(vim);
(void) printf("cannot verify obsolete indirect mapping "
"counts of vdev %llu because precise feature was not "
"enabled when it was removed: %d%% (%llx bytes) of mapping"
"unreferenced\n",
(u_longlong_t)vd->vdev_id, pct_leaked,
(u_longlong_t)total_leaked);
} else if (total_leaked > 0) {
(void) printf("obsolete indirect mapping count mismatch "
"for vdev %llu -- %llx total bytes mismatched\n",
(u_longlong_t)vd->vdev_id,
(u_longlong_t)total_leaked);
leaks |= B_TRUE;
}
vdev_indirect_mapping_free_obsolete_counts(vim,
zcb->zcb_vd_obsolete_counts[vd->vdev_id]);
zcb->zcb_vd_obsolete_counts[vd->vdev_id] = NULL;
return (leaks);
}
static boolean_t
zdb_leak_fini(spa_t *spa, zdb_cb_t *zcb)
{
if (dump_opt['L'])
return (B_FALSE);
boolean_t leaks = B_FALSE;
vdev_t *rvd = spa->spa_root_vdev;
for (unsigned c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
if (zcb->zcb_vd_obsolete_counts[c] != NULL) {
leaks |= zdb_check_for_obsolete_leaks(vd, zcb);
}
for (uint64_t m = 0; m < vd->vdev_ms_count; m++) {
metaslab_t *msp = vd->vdev_ms[m];
ASSERT3P(msp->ms_group, ==, (msp->ms_group->mg_class ==
spa_embedded_log_class(spa)) ?
vd->vdev_log_mg : vd->vdev_mg);
/*
* ms_allocatable has been overloaded
* to contain allocated segments. Now that
* we finished traversing all blocks, any
* block that remains in the ms_allocatable
* represents an allocated block that we
* did not claim during the traversal.
* Claimed blocks would have been removed
* from the ms_allocatable. For indirect
* vdevs, space remaining in the tree
* represents parts of the mapping that are
* not referenced, which is not a bug.
*/
if (vd->vdev_ops == &vdev_indirect_ops) {
range_tree_vacate(msp->ms_allocatable,
NULL, NULL);
} else {
range_tree_vacate(msp->ms_allocatable,
zdb_leak, vd);
}
if (msp->ms_loaded) {
msp->ms_loaded = B_FALSE;
}
}
}
umem_free(zcb->zcb_vd_obsolete_counts,
rvd->vdev_children * sizeof (uint32_t *));
zcb->zcb_vd_obsolete_counts = NULL;
return (leaks);
}
static int
count_block_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
(void) tx;
zdb_cb_t *zcb = arg;
if (dump_opt['b'] >= 5) {
char blkbuf[BP_SPRINTF_LEN];
snprintf_blkptr(blkbuf, sizeof (blkbuf), bp);
(void) printf("[%s] %s\n",
"deferred free", blkbuf);
}
zdb_count_block(zcb, NULL, bp, ZDB_OT_DEFERRED);
return (0);
}
/*
* Iterate over livelists which have been destroyed by the user but
* are still present in the MOS, waiting to be freed
*/
static void
iterate_deleted_livelists(spa_t *spa, ll_iter_t func, void *arg)
{
objset_t *mos = spa->spa_meta_objset;
uint64_t zap_obj;
int err = zap_lookup(mos, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_DELETED_CLONES, sizeof (uint64_t), 1, &zap_obj);
if (err == ENOENT)
return;
ASSERT0(err);
zap_cursor_t zc;
zap_attribute_t attr;
dsl_deadlist_t ll;
/* NULL out os prior to dsl_deadlist_open in case it's garbage */
ll.dl_os = NULL;
for (zap_cursor_init(&zc, mos, zap_obj);
zap_cursor_retrieve(&zc, &attr) == 0;
(void) zap_cursor_advance(&zc)) {
dsl_deadlist_open(&ll, mos, attr.za_first_integer);
func(&ll, arg);
dsl_deadlist_close(&ll);
}
zap_cursor_fini(&zc);
}
static int
bpobj_count_block_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
ASSERT(!bp_freed);
return (count_block_cb(arg, bp, tx));
}
static int
livelist_entry_count_blocks_cb(void *args, dsl_deadlist_entry_t *dle)
{
zdb_cb_t *zbc = args;
bplist_t blks;
bplist_create(&blks);
/* determine which blocks have been alloc'd but not freed */
VERIFY0(dsl_process_sub_livelist(&dle->dle_bpobj, &blks, NULL, NULL));
/* count those blocks */
(void) bplist_iterate(&blks, count_block_cb, zbc, NULL);
bplist_destroy(&blks);
return (0);
}
static void
livelist_count_blocks(dsl_deadlist_t *ll, void *arg)
{
dsl_deadlist_iterate(ll, livelist_entry_count_blocks_cb, arg);
}
/*
* Count the blocks in the livelists that have been destroyed by the user
* but haven't yet been freed.
*/
static void
deleted_livelists_count_blocks(spa_t *spa, zdb_cb_t *zbc)
{
iterate_deleted_livelists(spa, livelist_count_blocks, zbc);
}
static void
dump_livelist_cb(dsl_deadlist_t *ll, void *arg)
{
ASSERT3P(arg, ==, NULL);
global_feature_count[SPA_FEATURE_LIVELIST]++;
dump_blkptr_list(ll, "Deleted Livelist");
dsl_deadlist_iterate(ll, sublivelist_verify_lightweight, NULL);
}
/*
* Print out, register object references to, and increment feature counts for
* livelists that have been destroyed by the user but haven't yet been freed.
*/
static void
deleted_livelists_dump_mos(spa_t *spa)
{
uint64_t zap_obj;
objset_t *mos = spa->spa_meta_objset;
int err = zap_lookup(mos, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_DELETED_CLONES, sizeof (uint64_t), 1, &zap_obj);
if (err == ENOENT)
return;
mos_obj_refd(zap_obj);
iterate_deleted_livelists(spa, dump_livelist_cb, NULL);
}
static int
zdb_brt_entry_compare(const void *zcn1, const void *zcn2)
{
const dva_t *dva1 = &((const zdb_brt_entry_t *)zcn1)->zbre_dva;
const dva_t *dva2 = &((const zdb_brt_entry_t *)zcn2)->zbre_dva;
int cmp;
cmp = TREE_CMP(DVA_GET_VDEV(dva1), DVA_GET_VDEV(dva2));
if (cmp == 0)
cmp = TREE_CMP(DVA_GET_OFFSET(dva1), DVA_GET_OFFSET(dva2));
return (cmp);
}
static int
dump_block_stats(spa_t *spa)
{
zdb_cb_t *zcb;
zdb_blkstats_t *zb, *tzb;
uint64_t norm_alloc, norm_space, total_alloc, total_found;
int flags = TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA |
TRAVERSE_NO_DECRYPT | TRAVERSE_HARD;
boolean_t leaks = B_FALSE;
int e, c, err;
bp_embedded_type_t i;
zcb = umem_zalloc(sizeof (zdb_cb_t), UMEM_NOFAIL);
if (spa_feature_is_active(spa, SPA_FEATURE_BLOCK_CLONING)) {
avl_create(&zcb->zcb_brt, zdb_brt_entry_compare,
sizeof (zdb_brt_entry_t),
offsetof(zdb_brt_entry_t, zbre_node));
zcb->zcb_brt_is_active = B_TRUE;
}
(void) printf("\nTraversing all blocks %s%s%s%s%s...\n\n",
(dump_opt['c'] || !dump_opt['L']) ? "to verify " : "",
(dump_opt['c'] == 1) ? "metadata " : "",
dump_opt['c'] ? "checksums " : "",
(dump_opt['c'] && !dump_opt['L']) ? "and verify " : "",
!dump_opt['L'] ? "nothing leaked " : "");
/*
* When leak detection is enabled we load all space maps as SM_ALLOC
* maps, then traverse the pool claiming each block we discover. If
* the pool is perfectly consistent, the segment trees will be empty
* when we're done. Anything left over is a leak; any block we can't
* claim (because it's not part of any space map) is a double
* allocation, reference to a freed block, or an unclaimed log block.
*
* When leak detection is disabled (-L option) we still traverse the
* pool claiming each block we discover, but we skip opening any space
* maps.
*/
zdb_leak_init(spa, zcb);
/*
* If there's a deferred-free bplist, process that first.
*/
(void) bpobj_iterate_nofree(&spa->spa_deferred_bpobj,
bpobj_count_block_cb, zcb, NULL);
if (spa_version(spa) >= SPA_VERSION_DEADLISTS) {
(void) bpobj_iterate_nofree(&spa->spa_dsl_pool->dp_free_bpobj,
bpobj_count_block_cb, zcb, NULL);
}
zdb_claim_removing(spa, zcb);
if (spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY)) {
VERIFY3U(0, ==, bptree_iterate(spa->spa_meta_objset,
spa->spa_dsl_pool->dp_bptree_obj, B_FALSE, count_block_cb,
zcb, NULL));
}
deleted_livelists_count_blocks(spa, zcb);
if (dump_opt['c'] > 1)
flags |= TRAVERSE_PREFETCH_DATA;
zcb->zcb_totalasize = metaslab_class_get_alloc(spa_normal_class(spa));
zcb->zcb_totalasize += metaslab_class_get_alloc(spa_special_class(spa));
zcb->zcb_totalasize += metaslab_class_get_alloc(spa_dedup_class(spa));
zcb->zcb_totalasize +=
metaslab_class_get_alloc(spa_embedded_log_class(spa));
zcb->zcb_start = zcb->zcb_lastprint = gethrtime();
err = traverse_pool(spa, 0, flags, zdb_blkptr_cb, zcb);
/*
* If we've traversed the data blocks then we need to wait for those
* I/Os to complete. We leverage "The Godfather" zio to wait on
* all async I/Os to complete.
*/
if (dump_opt['c']) {
for (c = 0; c < max_ncpus; c++) {
(void) zio_wait(spa->spa_async_zio_root[c]);
spa->spa_async_zio_root[c] = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE |
ZIO_FLAG_GODFATHER);
}
}
ASSERT0(spa->spa_load_verify_bytes);
/*
* Done after zio_wait() since zcb_haderrors is modified in
* zdb_blkptr_done()
*/
zcb->zcb_haderrors |= err;
if (zcb->zcb_haderrors) {
(void) printf("\nError counts:\n\n");
(void) printf("\t%5s %s\n", "errno", "count");
for (e = 0; e < 256; e++) {
if (zcb->zcb_errors[e] != 0) {
(void) printf("\t%5d %llu\n",
e, (u_longlong_t)zcb->zcb_errors[e]);
}
}
}
/*
* Report any leaked segments.
*/
leaks |= zdb_leak_fini(spa, zcb);
tzb = &zcb->zcb_type[ZB_TOTAL][ZDB_OT_TOTAL];
norm_alloc = metaslab_class_get_alloc(spa_normal_class(spa));
norm_space = metaslab_class_get_space(spa_normal_class(spa));
total_alloc = norm_alloc +
metaslab_class_get_alloc(spa_log_class(spa)) +
metaslab_class_get_alloc(spa_embedded_log_class(spa)) +
metaslab_class_get_alloc(spa_special_class(spa)) +
metaslab_class_get_alloc(spa_dedup_class(spa)) +
get_unflushed_alloc_space(spa);
total_found =
tzb->zb_asize - zcb->zcb_dedup_asize - zcb->zcb_clone_asize +
zcb->zcb_removing_size + zcb->zcb_checkpoint_size;
if (total_found == total_alloc && !dump_opt['L']) {
(void) printf("\n\tNo leaks (block sum matches space"
" maps exactly)\n");
} else if (!dump_opt['L']) {
(void) printf("block traversal size %llu != alloc %llu "
"(%s %lld)\n",
(u_longlong_t)total_found,
(u_longlong_t)total_alloc,
(dump_opt['L']) ? "unreachable" : "leaked",
(longlong_t)(total_alloc - total_found));
leaks = B_TRUE;
}
if (tzb->zb_count == 0) {
umem_free(zcb, sizeof (zdb_cb_t));
return (2);
}
(void) printf("\n");
(void) printf("\t%-16s %14llu\n", "bp count:",
(u_longlong_t)tzb->zb_count);
(void) printf("\t%-16s %14llu\n", "ganged count:",
(longlong_t)tzb->zb_gangs);
(void) printf("\t%-16s %14llu avg: %6llu\n", "bp logical:",
(u_longlong_t)tzb->zb_lsize,
(u_longlong_t)(tzb->zb_lsize / tzb->zb_count));
(void) printf("\t%-16s %14llu avg: %6llu compression: %6.2f\n",
"bp physical:", (u_longlong_t)tzb->zb_psize,
(u_longlong_t)(tzb->zb_psize / tzb->zb_count),
(double)tzb->zb_lsize / tzb->zb_psize);
(void) printf("\t%-16s %14llu avg: %6llu compression: %6.2f\n",
"bp allocated:", (u_longlong_t)tzb->zb_asize,
(u_longlong_t)(tzb->zb_asize / tzb->zb_count),
(double)tzb->zb_lsize / tzb->zb_asize);
(void) printf("\t%-16s %14llu ref>1: %6llu deduplication: %6.2f\n",
"bp deduped:", (u_longlong_t)zcb->zcb_dedup_asize,
(u_longlong_t)zcb->zcb_dedup_blocks,
(double)zcb->zcb_dedup_asize / tzb->zb_asize + 1.0);
(void) printf("\t%-16s %14llu count: %6llu\n",
"bp cloned:", (u_longlong_t)zcb->zcb_clone_asize,
(u_longlong_t)zcb->zcb_clone_blocks);
(void) printf("\t%-16s %14llu used: %5.2f%%\n", "Normal class:",
(u_longlong_t)norm_alloc, 100.0 * norm_alloc / norm_space);
if (spa_special_class(spa)->mc_allocator[0].mca_rotor != NULL) {
uint64_t alloc = metaslab_class_get_alloc(
spa_special_class(spa));
uint64_t space = metaslab_class_get_space(
spa_special_class(spa));
(void) printf("\t%-16s %14llu used: %5.2f%%\n",
"Special class", (u_longlong_t)alloc,
100.0 * alloc / space);
}
if (spa_dedup_class(spa)->mc_allocator[0].mca_rotor != NULL) {
uint64_t alloc = metaslab_class_get_alloc(
spa_dedup_class(spa));
uint64_t space = metaslab_class_get_space(
spa_dedup_class(spa));
(void) printf("\t%-16s %14llu used: %5.2f%%\n",
"Dedup class", (u_longlong_t)alloc,
100.0 * alloc / space);
}
if (spa_embedded_log_class(spa)->mc_allocator[0].mca_rotor != NULL) {
uint64_t alloc = metaslab_class_get_alloc(
spa_embedded_log_class(spa));
uint64_t space = metaslab_class_get_space(
spa_embedded_log_class(spa));
(void) printf("\t%-16s %14llu used: %5.2f%%\n",
"Embedded log class", (u_longlong_t)alloc,
100.0 * alloc / space);
}
for (i = 0; i < NUM_BP_EMBEDDED_TYPES; i++) {
if (zcb->zcb_embedded_blocks[i] == 0)
continue;
(void) printf("\n");
(void) printf("\tadditional, non-pointer bps of type %u: "
"%10llu\n",
i, (u_longlong_t)zcb->zcb_embedded_blocks[i]);
if (dump_opt['b'] >= 3) {
(void) printf("\t number of (compressed) bytes: "
"number of bps\n");
dump_histogram(zcb->zcb_embedded_histogram[i],
sizeof (zcb->zcb_embedded_histogram[i]) /
sizeof (zcb->zcb_embedded_histogram[i][0]), 0);
}
}
if (tzb->zb_ditto_samevdev != 0) {
(void) printf("\tDittoed blocks on same vdev: %llu\n",
(longlong_t)tzb->zb_ditto_samevdev);
}
if (tzb->zb_ditto_same_ms != 0) {
(void) printf("\tDittoed blocks in same metaslab: %llu\n",
(longlong_t)tzb->zb_ditto_same_ms);
}
for (uint64_t v = 0; v < spa->spa_root_vdev->vdev_children; v++) {
vdev_t *vd = spa->spa_root_vdev->vdev_child[v];
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
if (vim == NULL) {
continue;
}
char mem[32];
zdb_nicenum(vdev_indirect_mapping_num_entries(vim),
mem, vdev_indirect_mapping_size(vim));
(void) printf("\tindirect vdev id %llu has %llu segments "
"(%s in memory)\n",
(longlong_t)vd->vdev_id,
(longlong_t)vdev_indirect_mapping_num_entries(vim), mem);
}
if (dump_opt['b'] >= 2) {
int l, t, level;
char csize[32], lsize[32], psize[32], asize[32];
char avg[32], gang[32];
(void) printf("\nBlocks\tLSIZE\tPSIZE\tASIZE"
"\t avg\t comp\t%%Total\tType\n");
zfs_blkstat_t *mdstats = umem_zalloc(sizeof (zfs_blkstat_t),
UMEM_NOFAIL);
for (t = 0; t <= ZDB_OT_TOTAL; t++) {
const char *typename;
/* make sure nicenum has enough space */
_Static_assert(sizeof (csize) >= NN_NUMBUF_SZ,
"csize truncated");
_Static_assert(sizeof (lsize) >= NN_NUMBUF_SZ,
"lsize truncated");
_Static_assert(sizeof (psize) >= NN_NUMBUF_SZ,
"psize truncated");
_Static_assert(sizeof (asize) >= NN_NUMBUF_SZ,
"asize truncated");
_Static_assert(sizeof (avg) >= NN_NUMBUF_SZ,
"avg truncated");
_Static_assert(sizeof (gang) >= NN_NUMBUF_SZ,
"gang truncated");
if (t < DMU_OT_NUMTYPES)
typename = dmu_ot[t].ot_name;
else
typename = zdb_ot_extname[t - DMU_OT_NUMTYPES];
if (zcb->zcb_type[ZB_TOTAL][t].zb_asize == 0) {
(void) printf("%6s\t%5s\t%5s\t%5s"
"\t%5s\t%5s\t%6s\t%s\n",
"-",
"-",
"-",
"-",
"-",
"-",
"-",
typename);
continue;
}
for (l = ZB_TOTAL - 1; l >= -1; l--) {
level = (l == -1 ? ZB_TOTAL : l);
zb = &zcb->zcb_type[level][t];
if (zb->zb_asize == 0)
continue;
if (level != ZB_TOTAL && t < DMU_OT_NUMTYPES &&
(level > 0 || DMU_OT_IS_METADATA(t))) {
mdstats->zb_count += zb->zb_count;
mdstats->zb_lsize += zb->zb_lsize;
mdstats->zb_psize += zb->zb_psize;
mdstats->zb_asize += zb->zb_asize;
mdstats->zb_gangs += zb->zb_gangs;
}
if (dump_opt['b'] < 3 && level != ZB_TOTAL)
continue;
if (level == 0 && zb->zb_asize ==
zcb->zcb_type[ZB_TOTAL][t].zb_asize)
continue;
zdb_nicenum(zb->zb_count, csize,
sizeof (csize));
zdb_nicenum(zb->zb_lsize, lsize,
sizeof (lsize));
zdb_nicenum(zb->zb_psize, psize,
sizeof (psize));
zdb_nicenum(zb->zb_asize, asize,
sizeof (asize));
zdb_nicenum(zb->zb_asize / zb->zb_count, avg,
sizeof (avg));
zdb_nicenum(zb->zb_gangs, gang, sizeof (gang));
(void) printf("%6s\t%5s\t%5s\t%5s\t%5s"
"\t%5.2f\t%6.2f\t",
csize, lsize, psize, asize, avg,
(double)zb->zb_lsize / zb->zb_psize,
100.0 * zb->zb_asize / tzb->zb_asize);
if (level == ZB_TOTAL)
(void) printf("%s\n", typename);
else
(void) printf(" L%d %s\n",
level, typename);
if (dump_opt['b'] >= 3 && zb->zb_gangs > 0) {
(void) printf("\t number of ganged "
"blocks: %s\n", gang);
}
if (dump_opt['b'] >= 4) {
(void) printf("psize "
"(in 512-byte sectors): "
"number of blocks\n");
dump_histogram(zb->zb_psize_histogram,
PSIZE_HISTO_SIZE, 0);
}
}
}
zdb_nicenum(mdstats->zb_count, csize,
sizeof (csize));
zdb_nicenum(mdstats->zb_lsize, lsize,
sizeof (lsize));
zdb_nicenum(mdstats->zb_psize, psize,
sizeof (psize));
zdb_nicenum(mdstats->zb_asize, asize,
sizeof (asize));
zdb_nicenum(mdstats->zb_asize / mdstats->zb_count, avg,
sizeof (avg));
zdb_nicenum(mdstats->zb_gangs, gang, sizeof (gang));
(void) printf("%6s\t%5s\t%5s\t%5s\t%5s"
"\t%5.2f\t%6.2f\t",
csize, lsize, psize, asize, avg,
(double)mdstats->zb_lsize / mdstats->zb_psize,
100.0 * mdstats->zb_asize / tzb->zb_asize);
(void) printf("%s\n", "Metadata Total");
/* Output a table summarizing block sizes in the pool */
if (dump_opt['b'] >= 2) {
dump_size_histograms(zcb);
}
umem_free(mdstats, sizeof (zfs_blkstat_t));
}
(void) printf("\n");
if (leaks) {
umem_free(zcb, sizeof (zdb_cb_t));
return (2);
}
if (zcb->zcb_haderrors) {
umem_free(zcb, sizeof (zdb_cb_t));
return (3);
}
umem_free(zcb, sizeof (zdb_cb_t));
return (0);
}
typedef struct zdb_ddt_entry {
ddt_key_t zdde_key;
uint64_t zdde_ref_blocks;
uint64_t zdde_ref_lsize;
uint64_t zdde_ref_psize;
uint64_t zdde_ref_dsize;
avl_node_t zdde_node;
} zdb_ddt_entry_t;
static int
zdb_ddt_add_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
const zbookmark_phys_t *zb, const dnode_phys_t *dnp, void *arg)
{
(void) zilog, (void) dnp;
avl_tree_t *t = arg;
avl_index_t where;
zdb_ddt_entry_t *zdde, zdde_search;
if (zb->zb_level == ZB_DNODE_LEVEL || BP_IS_HOLE(bp) ||
BP_IS_EMBEDDED(bp))
return (0);
if (dump_opt['S'] > 1 && zb->zb_level == ZB_ROOT_LEVEL) {
(void) printf("traversing objset %llu, %llu objects, "
"%lu blocks so far\n",
(u_longlong_t)zb->zb_objset,
(u_longlong_t)BP_GET_FILL(bp),
avl_numnodes(t));
}
if (BP_IS_HOLE(bp) || BP_GET_CHECKSUM(bp) == ZIO_CHECKSUM_OFF ||
BP_GET_LEVEL(bp) > 0 || DMU_OT_IS_METADATA(BP_GET_TYPE(bp)))
return (0);
ddt_key_fill(&zdde_search.zdde_key, bp);
zdde = avl_find(t, &zdde_search, &where);
if (zdde == NULL) {
zdde = umem_zalloc(sizeof (*zdde), UMEM_NOFAIL);
zdde->zdde_key = zdde_search.zdde_key;
avl_insert(t, zdde, where);
}
zdde->zdde_ref_blocks += 1;
zdde->zdde_ref_lsize += BP_GET_LSIZE(bp);
zdde->zdde_ref_psize += BP_GET_PSIZE(bp);
zdde->zdde_ref_dsize += bp_get_dsize_sync(spa, bp);
return (0);
}
static void
dump_simulated_ddt(spa_t *spa)
{
avl_tree_t t;
void *cookie = NULL;
zdb_ddt_entry_t *zdde;
ddt_histogram_t ddh_total = {{{0}}};
ddt_stat_t dds_total = {0};
avl_create(&t, ddt_entry_compare,
sizeof (zdb_ddt_entry_t), offsetof(zdb_ddt_entry_t, zdde_node));
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
(void) traverse_pool(spa, 0, TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA |
TRAVERSE_NO_DECRYPT, zdb_ddt_add_cb, &t);
spa_config_exit(spa, SCL_CONFIG, FTAG);
while ((zdde = avl_destroy_nodes(&t, &cookie)) != NULL) {
ddt_stat_t dds;
uint64_t refcnt = zdde->zdde_ref_blocks;
ASSERT(refcnt != 0);
dds.dds_blocks = zdde->zdde_ref_blocks / refcnt;
dds.dds_lsize = zdde->zdde_ref_lsize / refcnt;
dds.dds_psize = zdde->zdde_ref_psize / refcnt;
dds.dds_dsize = zdde->zdde_ref_dsize / refcnt;
dds.dds_ref_blocks = zdde->zdde_ref_blocks;
dds.dds_ref_lsize = zdde->zdde_ref_lsize;
dds.dds_ref_psize = zdde->zdde_ref_psize;
dds.dds_ref_dsize = zdde->zdde_ref_dsize;
ddt_stat_add(&ddh_total.ddh_stat[highbit64(refcnt) - 1],
&dds, 0);
umem_free(zdde, sizeof (*zdde));
}
avl_destroy(&t);
ddt_histogram_stat(&dds_total, &ddh_total);
(void) printf("Simulated DDT histogram:\n");
zpool_dump_ddt(&dds_total, &ddh_total);
dump_dedup_ratio(&dds_total);
}
static int
verify_device_removal_feature_counts(spa_t *spa)
{
uint64_t dr_feature_refcount = 0;
uint64_t oc_feature_refcount = 0;
uint64_t indirect_vdev_count = 0;
uint64_t precise_vdev_count = 0;
uint64_t obsolete_counts_object_count = 0;
uint64_t obsolete_sm_count = 0;
uint64_t obsolete_counts_count = 0;
uint64_t scip_count = 0;
uint64_t obsolete_bpobj_count = 0;
int ret = 0;
spa_condensing_indirect_phys_t *scip =
&spa->spa_condensing_indirect_phys;
if (scip->scip_next_mapping_object != 0) {
vdev_t *vd = spa->spa_root_vdev->vdev_child[scip->scip_vdev];
ASSERT(scip->scip_prev_obsolete_sm_object != 0);
ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops);
(void) printf("Condensing indirect vdev %llu: new mapping "
"object %llu, prev obsolete sm %llu\n",
(u_longlong_t)scip->scip_vdev,
(u_longlong_t)scip->scip_next_mapping_object,
(u_longlong_t)scip->scip_prev_obsolete_sm_object);
if (scip->scip_prev_obsolete_sm_object != 0) {
space_map_t *prev_obsolete_sm = NULL;
VERIFY0(space_map_open(&prev_obsolete_sm,
spa->spa_meta_objset,
scip->scip_prev_obsolete_sm_object,
0, vd->vdev_asize, 0));
dump_spacemap(spa->spa_meta_objset, prev_obsolete_sm);
(void) printf("\n");
space_map_close(prev_obsolete_sm);
}
scip_count += 2;
}
for (uint64_t i = 0; i < spa->spa_root_vdev->vdev_children; i++) {
vdev_t *vd = spa->spa_root_vdev->vdev_child[i];
vdev_indirect_config_t *vic = &vd->vdev_indirect_config;
if (vic->vic_mapping_object != 0) {
ASSERT(vd->vdev_ops == &vdev_indirect_ops ||
vd->vdev_removing);
indirect_vdev_count++;
if (vd->vdev_indirect_mapping->vim_havecounts) {
obsolete_counts_count++;
}
}
boolean_t are_precise;
VERIFY0(vdev_obsolete_counts_are_precise(vd, &are_precise));
if (are_precise) {
ASSERT(vic->vic_mapping_object != 0);
precise_vdev_count++;
}
uint64_t obsolete_sm_object;
VERIFY0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
if (obsolete_sm_object != 0) {
ASSERT(vic->vic_mapping_object != 0);
obsolete_sm_count++;
}
}
(void) feature_get_refcount(spa,
&spa_feature_table[SPA_FEATURE_DEVICE_REMOVAL],
&dr_feature_refcount);
(void) feature_get_refcount(spa,
&spa_feature_table[SPA_FEATURE_OBSOLETE_COUNTS],
&oc_feature_refcount);
if (dr_feature_refcount != indirect_vdev_count) {
ret = 1;
(void) printf("Number of indirect vdevs (%llu) " \
"does not match feature count (%llu)\n",
(u_longlong_t)indirect_vdev_count,
(u_longlong_t)dr_feature_refcount);
} else {
(void) printf("Verified device_removal feature refcount " \
"of %llu is correct\n",
(u_longlong_t)dr_feature_refcount);
}
if (zap_contains(spa_meta_objset(spa), DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_OBSOLETE_BPOBJ) == 0) {
obsolete_bpobj_count++;
}
obsolete_counts_object_count = precise_vdev_count;
obsolete_counts_object_count += obsolete_sm_count;
obsolete_counts_object_count += obsolete_counts_count;
obsolete_counts_object_count += scip_count;
obsolete_counts_object_count += obsolete_bpobj_count;
obsolete_counts_object_count += remap_deadlist_count;
if (oc_feature_refcount != obsolete_counts_object_count) {
ret = 1;
(void) printf("Number of obsolete counts objects (%llu) " \
"does not match feature count (%llu)\n",
(u_longlong_t)obsolete_counts_object_count,
(u_longlong_t)oc_feature_refcount);
(void) printf("pv:%llu os:%llu oc:%llu sc:%llu "
"ob:%llu rd:%llu\n",
(u_longlong_t)precise_vdev_count,
(u_longlong_t)obsolete_sm_count,
(u_longlong_t)obsolete_counts_count,
(u_longlong_t)scip_count,
(u_longlong_t)obsolete_bpobj_count,
(u_longlong_t)remap_deadlist_count);
} else {
(void) printf("Verified indirect_refcount feature refcount " \
"of %llu is correct\n",
(u_longlong_t)oc_feature_refcount);
}
return (ret);
}
static void
zdb_set_skip_mmp(char *target)
{
spa_t *spa;
/*
* Disable the activity check to allow examination of
* active pools.
*/
mutex_enter(&spa_namespace_lock);
if ((spa = spa_lookup(target)) != NULL) {
spa->spa_import_flags |= ZFS_IMPORT_SKIP_MMP;
}
mutex_exit(&spa_namespace_lock);
}
#define BOGUS_SUFFIX "_CHECKPOINTED_UNIVERSE"
/*
* Import the checkpointed state of the pool specified by the target
* parameter as readonly. The function also accepts a pool config
* as an optional parameter, else it attempts to infer the config by
* the name of the target pool.
*
* Note that the checkpointed state's pool name will be the name of
* the original pool with the above suffix appended to it. In addition,
* if the target is not a pool name (e.g. a path to a dataset) then
* the new_path parameter is populated with the updated path to
* reflect the fact that we are looking into the checkpointed state.
*
* The function returns a newly-allocated copy of the name of the
* pool containing the checkpointed state. When this copy is no
* longer needed it should be freed with free(3C). Same thing
* applies to the new_path parameter if allocated.
*/
static char *
import_checkpointed_state(char *target, nvlist_t *cfg, char **new_path)
{
int error = 0;
char *poolname, *bogus_name = NULL;
boolean_t freecfg = B_FALSE;
/* If the target is not a pool, the extract the pool name */
char *path_start = strchr(target, '/');
if (path_start != NULL) {
size_t poolname_len = path_start - target;
poolname = strndup(target, poolname_len);
} else {
poolname = target;
}
if (cfg == NULL) {
zdb_set_skip_mmp(poolname);
error = spa_get_stats(poolname, &cfg, NULL, 0);
if (error != 0) {
fatal("Tried to read config of pool \"%s\" but "
"spa_get_stats() failed with error %d\n",
poolname, error);
}
freecfg = B_TRUE;
}
if (asprintf(&bogus_name, "%s%s", poolname, BOGUS_SUFFIX) == -1) {
if (target != poolname)
free(poolname);
return (NULL);
}
fnvlist_add_string(cfg, ZPOOL_CONFIG_POOL_NAME, bogus_name);
error = spa_import(bogus_name, cfg, NULL,
ZFS_IMPORT_MISSING_LOG | ZFS_IMPORT_CHECKPOINT |
ZFS_IMPORT_SKIP_MMP);
if (freecfg)
nvlist_free(cfg);
if (error != 0) {
fatal("Tried to import pool \"%s\" but spa_import() failed "
"with error %d\n", bogus_name, error);
}
if (new_path != NULL && path_start != NULL) {
if (asprintf(new_path, "%s%s", bogus_name, path_start) == -1) {
free(bogus_name);
if (path_start != NULL)
free(poolname);
return (NULL);
}
}
if (target != poolname)
free(poolname);
return (bogus_name);
}
typedef struct verify_checkpoint_sm_entry_cb_arg {
vdev_t *vcsec_vd;
/* the following fields are only used for printing progress */
uint64_t vcsec_entryid;
uint64_t vcsec_num_entries;
} verify_checkpoint_sm_entry_cb_arg_t;
#define ENTRIES_PER_PROGRESS_UPDATE 10000
static int
verify_checkpoint_sm_entry_cb(space_map_entry_t *sme, void *arg)
{
verify_checkpoint_sm_entry_cb_arg_t *vcsec = arg;
vdev_t *vd = vcsec->vcsec_vd;
metaslab_t *ms = vd->vdev_ms[sme->sme_offset >> vd->vdev_ms_shift];
uint64_t end = sme->sme_offset + sme->sme_run;
ASSERT(sme->sme_type == SM_FREE);
if ((vcsec->vcsec_entryid % ENTRIES_PER_PROGRESS_UPDATE) == 0) {
(void) fprintf(stderr,
"\rverifying vdev %llu, space map entry %llu of %llu ...",
(longlong_t)vd->vdev_id,
(longlong_t)vcsec->vcsec_entryid,
(longlong_t)vcsec->vcsec_num_entries);
}
vcsec->vcsec_entryid++;
/*
* See comment in checkpoint_sm_exclude_entry_cb()
*/
VERIFY3U(sme->sme_offset, >=, ms->ms_start);
VERIFY3U(end, <=, ms->ms_start + ms->ms_size);
/*
* The entries in the vdev_checkpoint_sm should be marked as
* allocated in the checkpointed state of the pool, therefore
* their respective ms_allocateable trees should not contain them.
*/
mutex_enter(&ms->ms_lock);
range_tree_verify_not_present(ms->ms_allocatable,
sme->sme_offset, sme->sme_run);
mutex_exit(&ms->ms_lock);
return (0);
}
/*
* Verify that all segments in the vdev_checkpoint_sm are allocated
* according to the checkpoint's ms_sm (i.e. are not in the checkpoint's
* ms_allocatable).
*
* Do so by comparing the checkpoint space maps (vdev_checkpoint_sm) of
* each vdev in the current state of the pool to the metaslab space maps
* (ms_sm) of the checkpointed state of the pool.
*
* Note that the function changes the state of the ms_allocatable
* trees of the current spa_t. The entries of these ms_allocatable
* trees are cleared out and then repopulated from with the free
* entries of their respective ms_sm space maps.
*/
static void
verify_checkpoint_vdev_spacemaps(spa_t *checkpoint, spa_t *current)
{
vdev_t *ckpoint_rvd = checkpoint->spa_root_vdev;
vdev_t *current_rvd = current->spa_root_vdev;
load_concrete_ms_allocatable_trees(checkpoint, SM_FREE);
for (uint64_t c = 0; c < ckpoint_rvd->vdev_children; c++) {
vdev_t *ckpoint_vd = ckpoint_rvd->vdev_child[c];
vdev_t *current_vd = current_rvd->vdev_child[c];
space_map_t *checkpoint_sm = NULL;
uint64_t checkpoint_sm_obj;
if (ckpoint_vd->vdev_ops == &vdev_indirect_ops) {
/*
* Since we don't allow device removal in a pool
* that has a checkpoint, we expect that all removed
* vdevs were removed from the pool before the
* checkpoint.
*/
ASSERT3P(current_vd->vdev_ops, ==, &vdev_indirect_ops);
continue;
}
/*
* If the checkpoint space map doesn't exist, then nothing
* here is checkpointed so there's nothing to verify.
*/
if (current_vd->vdev_top_zap == 0 ||
zap_contains(spa_meta_objset(current),
current_vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM) != 0)
continue;
VERIFY0(zap_lookup(spa_meta_objset(current),
current_vd->vdev_top_zap, VDEV_TOP_ZAP_POOL_CHECKPOINT_SM,
sizeof (uint64_t), 1, &checkpoint_sm_obj));
VERIFY0(space_map_open(&checkpoint_sm, spa_meta_objset(current),
checkpoint_sm_obj, 0, current_vd->vdev_asize,
current_vd->vdev_ashift));
verify_checkpoint_sm_entry_cb_arg_t vcsec;
vcsec.vcsec_vd = ckpoint_vd;
vcsec.vcsec_entryid = 0;
vcsec.vcsec_num_entries =
space_map_length(checkpoint_sm) / sizeof (uint64_t);
VERIFY0(space_map_iterate(checkpoint_sm,
space_map_length(checkpoint_sm),
verify_checkpoint_sm_entry_cb, &vcsec));
if (dump_opt['m'] > 3)
dump_spacemap(current->spa_meta_objset, checkpoint_sm);
space_map_close(checkpoint_sm);
}
/*
* If we've added vdevs since we took the checkpoint, ensure
* that their checkpoint space maps are empty.
*/
if (ckpoint_rvd->vdev_children < current_rvd->vdev_children) {
for (uint64_t c = ckpoint_rvd->vdev_children;
c < current_rvd->vdev_children; c++) {
vdev_t *current_vd = current_rvd->vdev_child[c];
VERIFY3P(current_vd->vdev_checkpoint_sm, ==, NULL);
}
}
/* for cleaner progress output */
(void) fprintf(stderr, "\n");
}
/*
* Verifies that all space that's allocated in the checkpoint is
* still allocated in the current version, by checking that everything
* in checkpoint's ms_allocatable (which is actually allocated, not
* allocatable/free) is not present in current's ms_allocatable.
*
* Note that the function changes the state of the ms_allocatable
* trees of both spas when called. The entries of all ms_allocatable
* trees are cleared out and then repopulated from their respective
* ms_sm space maps. In the checkpointed state we load the allocated
* entries, and in the current state we load the free entries.
*/
static void
verify_checkpoint_ms_spacemaps(spa_t *checkpoint, spa_t *current)
{
vdev_t *ckpoint_rvd = checkpoint->spa_root_vdev;
vdev_t *current_rvd = current->spa_root_vdev;
load_concrete_ms_allocatable_trees(checkpoint, SM_ALLOC);
load_concrete_ms_allocatable_trees(current, SM_FREE);
for (uint64_t i = 0; i < ckpoint_rvd->vdev_children; i++) {
vdev_t *ckpoint_vd = ckpoint_rvd->vdev_child[i];
vdev_t *current_vd = current_rvd->vdev_child[i];
if (ckpoint_vd->vdev_ops == &vdev_indirect_ops) {
/*
* See comment in verify_checkpoint_vdev_spacemaps()
*/
ASSERT3P(current_vd->vdev_ops, ==, &vdev_indirect_ops);
continue;
}
for (uint64_t m = 0; m < ckpoint_vd->vdev_ms_count; m++) {
metaslab_t *ckpoint_msp = ckpoint_vd->vdev_ms[m];
metaslab_t *current_msp = current_vd->vdev_ms[m];
(void) fprintf(stderr,
"\rverifying vdev %llu of %llu, "
"metaslab %llu of %llu ...",
(longlong_t)current_vd->vdev_id,
(longlong_t)current_rvd->vdev_children,
(longlong_t)current_vd->vdev_ms[m]->ms_id,
(longlong_t)current_vd->vdev_ms_count);
/*
* We walk through the ms_allocatable trees that
* are loaded with the allocated blocks from the
* ms_sm spacemaps of the checkpoint. For each
* one of these ranges we ensure that none of them
* exists in the ms_allocatable trees of the
* current state which are loaded with the ranges
* that are currently free.
*
* This way we ensure that none of the blocks that
* are part of the checkpoint were freed by mistake.
*/
range_tree_walk(ckpoint_msp->ms_allocatable,
(range_tree_func_t *)range_tree_verify_not_present,
current_msp->ms_allocatable);
}
}
/* for cleaner progress output */
(void) fprintf(stderr, "\n");
}
static void
verify_checkpoint_blocks(spa_t *spa)
{
ASSERT(!dump_opt['L']);
spa_t *checkpoint_spa;
char *checkpoint_pool;
int error = 0;
/*
* We import the checkpointed state of the pool (under a different
* name) so we can do verification on it against the current state
* of the pool.
*/
checkpoint_pool = import_checkpointed_state(spa->spa_name, NULL,
NULL);
ASSERT(strcmp(spa->spa_name, checkpoint_pool) != 0);
error = spa_open(checkpoint_pool, &checkpoint_spa, FTAG);
if (error != 0) {
fatal("Tried to open pool \"%s\" but spa_open() failed with "
"error %d\n", checkpoint_pool, error);
}
/*
* Ensure that ranges in the checkpoint space maps of each vdev
* are allocated according to the checkpointed state's metaslab
* space maps.
*/
verify_checkpoint_vdev_spacemaps(checkpoint_spa, spa);
/*
* Ensure that allocated ranges in the checkpoint's metaslab
* space maps remain allocated in the metaslab space maps of
* the current state.
*/
verify_checkpoint_ms_spacemaps(checkpoint_spa, spa);
/*
* Once we are done, we get rid of the checkpointed state.
*/
spa_close(checkpoint_spa, FTAG);
free(checkpoint_pool);
}
static void
dump_leftover_checkpoint_blocks(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t i = 0; i < rvd->vdev_children; i++) {
vdev_t *vd = rvd->vdev_child[i];
space_map_t *checkpoint_sm = NULL;
uint64_t checkpoint_sm_obj;
if (vd->vdev_top_zap == 0)
continue;
if (zap_contains(spa_meta_objset(spa), vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM) != 0)
continue;
VERIFY0(zap_lookup(spa_meta_objset(spa), vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM,
sizeof (uint64_t), 1, &checkpoint_sm_obj));
VERIFY0(space_map_open(&checkpoint_sm, spa_meta_objset(spa),
checkpoint_sm_obj, 0, vd->vdev_asize, vd->vdev_ashift));
dump_spacemap(spa->spa_meta_objset, checkpoint_sm);
space_map_close(checkpoint_sm);
}
}
static int
verify_checkpoint(spa_t *spa)
{
uberblock_t checkpoint;
int error;
if (!spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT))
return (0);
error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ZPOOL_CHECKPOINT, sizeof (uint64_t),
sizeof (uberblock_t) / sizeof (uint64_t), &checkpoint);
if (error == ENOENT && !dump_opt['L']) {
/*
* If the feature is active but the uberblock is missing
* then we must be in the middle of discarding the
* checkpoint.
*/
(void) printf("\nPartially discarded checkpoint "
"state found:\n");
if (dump_opt['m'] > 3)
dump_leftover_checkpoint_blocks(spa);
return (0);
} else if (error != 0) {
(void) printf("lookup error %d when looking for "
"checkpointed uberblock in MOS\n", error);
return (error);
}
dump_uberblock(&checkpoint, "\nCheckpointed uberblock found:\n", "\n");
if (checkpoint.ub_checkpoint_txg == 0) {
(void) printf("\nub_checkpoint_txg not set in checkpointed "
"uberblock\n");
error = 3;
}
if (error == 0 && !dump_opt['L'])
verify_checkpoint_blocks(spa);
return (error);
}
static void
mos_leaks_cb(void *arg, uint64_t start, uint64_t size)
{
(void) arg;
for (uint64_t i = start; i < size; i++) {
(void) printf("MOS object %llu referenced but not allocated\n",
(u_longlong_t)i);
}
}
static void
mos_obj_refd(uint64_t obj)
{
if (obj != 0 && mos_refd_objs != NULL)
range_tree_add(mos_refd_objs, obj, 1);
}
/*
* Call on a MOS object that may already have been referenced.
*/
static void
mos_obj_refd_multiple(uint64_t obj)
{
if (obj != 0 && mos_refd_objs != NULL &&
!range_tree_contains(mos_refd_objs, obj, 1))
range_tree_add(mos_refd_objs, obj, 1);
}
static void
mos_leak_vdev_top_zap(vdev_t *vd)
{
uint64_t ms_flush_data_obj;
int error = zap_lookup(spa_meta_objset(vd->vdev_spa),
vd->vdev_top_zap, VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS,
sizeof (ms_flush_data_obj), 1, &ms_flush_data_obj);
if (error == ENOENT)
return;
ASSERT0(error);
mos_obj_refd(ms_flush_data_obj);
}
static void
mos_leak_vdev(vdev_t *vd)
{
mos_obj_refd(vd->vdev_dtl_object);
mos_obj_refd(vd->vdev_ms_array);
mos_obj_refd(vd->vdev_indirect_config.vic_births_object);
mos_obj_refd(vd->vdev_indirect_config.vic_mapping_object);
mos_obj_refd(vd->vdev_leaf_zap);
if (vd->vdev_checkpoint_sm != NULL)
mos_obj_refd(vd->vdev_checkpoint_sm->sm_object);
if (vd->vdev_indirect_mapping != NULL) {
mos_obj_refd(vd->vdev_indirect_mapping->
vim_phys->vimp_counts_object);
}
if (vd->vdev_obsolete_sm != NULL)
mos_obj_refd(vd->vdev_obsolete_sm->sm_object);
for (uint64_t m = 0; m < vd->vdev_ms_count; m++) {
metaslab_t *ms = vd->vdev_ms[m];
mos_obj_refd(space_map_object(ms->ms_sm));
}
if (vd->vdev_root_zap != 0)
mos_obj_refd(vd->vdev_root_zap);
if (vd->vdev_top_zap != 0) {
mos_obj_refd(vd->vdev_top_zap);
mos_leak_vdev_top_zap(vd);
}
for (uint64_t c = 0; c < vd->vdev_children; c++) {
mos_leak_vdev(vd->vdev_child[c]);
}
}
static void
mos_leak_log_spacemaps(spa_t *spa)
{
uint64_t spacemap_zap;
int error = zap_lookup(spa_meta_objset(spa),
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_LOG_SPACEMAP_ZAP,
sizeof (spacemap_zap), 1, &spacemap_zap);
if (error == ENOENT)
return;
ASSERT0(error);
mos_obj_refd(spacemap_zap);
for (spa_log_sm_t *sls = avl_first(&spa->spa_sm_logs_by_txg);
sls; sls = AVL_NEXT(&spa->spa_sm_logs_by_txg, sls))
mos_obj_refd(sls->sls_sm_obj);
}
static void
errorlog_count_refd(objset_t *mos, uint64_t errlog)
{
zap_cursor_t zc;
zap_attribute_t za;
for (zap_cursor_init(&zc, mos, errlog);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
mos_obj_refd(za.za_first_integer);
}
zap_cursor_fini(&zc);
}
static int
dump_mos_leaks(spa_t *spa)
{
int rv = 0;
objset_t *mos = spa->spa_meta_objset;
dsl_pool_t *dp = spa->spa_dsl_pool;
/* Visit and mark all referenced objects in the MOS */
mos_obj_refd(DMU_POOL_DIRECTORY_OBJECT);
mos_obj_refd(spa->spa_pool_props_object);
mos_obj_refd(spa->spa_config_object);
mos_obj_refd(spa->spa_ddt_stat_object);
mos_obj_refd(spa->spa_feat_desc_obj);
mos_obj_refd(spa->spa_feat_enabled_txg_obj);
mos_obj_refd(spa->spa_feat_for_read_obj);
mos_obj_refd(spa->spa_feat_for_write_obj);
mos_obj_refd(spa->spa_history);
mos_obj_refd(spa->spa_errlog_last);
mos_obj_refd(spa->spa_errlog_scrub);
if (spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) {
errorlog_count_refd(mos, spa->spa_errlog_last);
errorlog_count_refd(mos, spa->spa_errlog_scrub);
}
mos_obj_refd(spa->spa_all_vdev_zaps);
mos_obj_refd(spa->spa_dsl_pool->dp_bptree_obj);
mos_obj_refd(spa->spa_dsl_pool->dp_tmp_userrefs_obj);
mos_obj_refd(spa->spa_dsl_pool->dp_scan->scn_phys.scn_queue_obj);
bpobj_count_refd(&spa->spa_deferred_bpobj);
mos_obj_refd(dp->dp_empty_bpobj);
bpobj_count_refd(&dp->dp_obsolete_bpobj);
bpobj_count_refd(&dp->dp_free_bpobj);
mos_obj_refd(spa->spa_l2cache.sav_object);
mos_obj_refd(spa->spa_spares.sav_object);
if (spa->spa_syncing_log_sm != NULL)
mos_obj_refd(spa->spa_syncing_log_sm->sm_object);
mos_leak_log_spacemaps(spa);
mos_obj_refd(spa->spa_condensing_indirect_phys.
scip_next_mapping_object);
mos_obj_refd(spa->spa_condensing_indirect_phys.
scip_prev_obsolete_sm_object);
if (spa->spa_condensing_indirect_phys.scip_next_mapping_object != 0) {
vdev_indirect_mapping_t *vim =
vdev_indirect_mapping_open(mos,
spa->spa_condensing_indirect_phys.scip_next_mapping_object);
mos_obj_refd(vim->vim_phys->vimp_counts_object);
vdev_indirect_mapping_close(vim);
}
deleted_livelists_dump_mos(spa);
if (dp->dp_origin_snap != NULL) {
dsl_dataset_t *ds;
dsl_pool_config_enter(dp, FTAG);
VERIFY0(dsl_dataset_hold_obj(dp,
dsl_dataset_phys(dp->dp_origin_snap)->ds_next_snap_obj,
FTAG, &ds));
count_ds_mos_objects(ds);
dump_blkptr_list(&ds->ds_deadlist, "Deadlist");
dsl_dataset_rele(ds, FTAG);
dsl_pool_config_exit(dp, FTAG);
count_ds_mos_objects(dp->dp_origin_snap);
dump_blkptr_list(&dp->dp_origin_snap->ds_deadlist, "Deadlist");
}
count_dir_mos_objects(dp->dp_mos_dir);
if (dp->dp_free_dir != NULL)
count_dir_mos_objects(dp->dp_free_dir);
if (dp->dp_leak_dir != NULL)
count_dir_mos_objects(dp->dp_leak_dir);
mos_leak_vdev(spa->spa_root_vdev);
for (uint64_t class = 0; class < DDT_CLASSES; class++) {
for (uint64_t type = 0; type < DDT_TYPES; type++) {
for (uint64_t cksum = 0;
cksum < ZIO_CHECKSUM_FUNCTIONS; cksum++) {
ddt_t *ddt = spa->spa_ddt[cksum];
mos_obj_refd(ddt->ddt_object[type][class]);
}
}
}
/*
* Visit all allocated objects and make sure they are referenced.
*/
uint64_t object = 0;
while (dmu_object_next(mos, &object, B_FALSE, 0) == 0) {
if (range_tree_contains(mos_refd_objs, object, 1)) {
range_tree_remove(mos_refd_objs, object, 1);
} else {
dmu_object_info_t doi;
const char *name;
VERIFY0(dmu_object_info(mos, object, &doi));
if (doi.doi_type & DMU_OT_NEWTYPE) {
dmu_object_byteswap_t bswap =
DMU_OT_BYTESWAP(doi.doi_type);
name = dmu_ot_byteswap[bswap].ob_name;
} else {
name = dmu_ot[doi.doi_type].ot_name;
}
(void) printf("MOS object %llu (%s) leaked\n",
(u_longlong_t)object, name);
rv = 2;
}
}
(void) range_tree_walk(mos_refd_objs, mos_leaks_cb, NULL);
if (!range_tree_is_empty(mos_refd_objs))
rv = 2;
range_tree_vacate(mos_refd_objs, NULL, NULL);
range_tree_destroy(mos_refd_objs);
return (rv);
}
typedef struct log_sm_obsolete_stats_arg {
uint64_t lsos_current_txg;
uint64_t lsos_total_entries;
uint64_t lsos_valid_entries;
uint64_t lsos_sm_entries;
uint64_t lsos_valid_sm_entries;
} log_sm_obsolete_stats_arg_t;
static int
log_spacemap_obsolete_stats_cb(spa_t *spa, space_map_entry_t *sme,
uint64_t txg, void *arg)
{
log_sm_obsolete_stats_arg_t *lsos = arg;
uint64_t offset = sme->sme_offset;
uint64_t vdev_id = sme->sme_vdev;
if (lsos->lsos_current_txg == 0) {
/* this is the first log */
lsos->lsos_current_txg = txg;
} else if (lsos->lsos_current_txg < txg) {
/* we just changed log - print stats and reset */
(void) printf("%-8llu valid entries out of %-8llu - txg %llu\n",
(u_longlong_t)lsos->lsos_valid_sm_entries,
(u_longlong_t)lsos->lsos_sm_entries,
(u_longlong_t)lsos->lsos_current_txg);
lsos->lsos_valid_sm_entries = 0;
lsos->lsos_sm_entries = 0;
lsos->lsos_current_txg = txg;
}
ASSERT3U(lsos->lsos_current_txg, ==, txg);
lsos->lsos_sm_entries++;
lsos->lsos_total_entries++;
vdev_t *vd = vdev_lookup_top(spa, vdev_id);
if (!vdev_is_concrete(vd))
return (0);
metaslab_t *ms = vd->vdev_ms[offset >> vd->vdev_ms_shift];
ASSERT(sme->sme_type == SM_ALLOC || sme->sme_type == SM_FREE);
if (txg < metaslab_unflushed_txg(ms))
return (0);
lsos->lsos_valid_sm_entries++;
lsos->lsos_valid_entries++;
return (0);
}
static void
dump_log_spacemap_obsolete_stats(spa_t *spa)
{
if (!spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP))
return;
log_sm_obsolete_stats_arg_t lsos = {0};
(void) printf("Log Space Map Obsolete Entry Statistics:\n");
iterate_through_spacemap_logs(spa,
log_spacemap_obsolete_stats_cb, &lsos);
/* print stats for latest log */
(void) printf("%-8llu valid entries out of %-8llu - txg %llu\n",
(u_longlong_t)lsos.lsos_valid_sm_entries,
(u_longlong_t)lsos.lsos_sm_entries,
(u_longlong_t)lsos.lsos_current_txg);
(void) printf("%-8llu valid entries out of %-8llu - total\n\n",
(u_longlong_t)lsos.lsos_valid_entries,
(u_longlong_t)lsos.lsos_total_entries);
}
static void
dump_zpool(spa_t *spa)
{
dsl_pool_t *dp = spa_get_dsl(spa);
int rc = 0;
if (dump_opt['y']) {
livelist_metaslab_validate(spa);
}
if (dump_opt['S']) {
dump_simulated_ddt(spa);
return;
}
if (!dump_opt['e'] && dump_opt['C'] > 1) {
(void) printf("\nCached configuration:\n");
dump_nvlist(spa->spa_config, 8);
}
if (dump_opt['C'])
dump_config(spa);
if (dump_opt['u'])
dump_uberblock(&spa->spa_uberblock, "\nUberblock:\n", "\n");
if (dump_opt['D'])
dump_all_ddts(spa);
if (dump_opt['d'] > 2 || dump_opt['m'])
dump_metaslabs(spa);
if (dump_opt['M'])
dump_metaslab_groups(spa, dump_opt['M'] > 1);
if (dump_opt['d'] > 2 || dump_opt['m']) {
dump_log_spacemaps(spa);
dump_log_spacemap_obsolete_stats(spa);
}
if (dump_opt['d'] || dump_opt['i']) {
spa_feature_t f;
mos_refd_objs = range_tree_create(NULL, RANGE_SEG64, NULL, 0,
0);
dump_objset(dp->dp_meta_objset);
if (dump_opt['d'] >= 3) {
dsl_pool_t *dp = spa->spa_dsl_pool;
dump_full_bpobj(&spa->spa_deferred_bpobj,
"Deferred frees", 0);
if (spa_version(spa) >= SPA_VERSION_DEADLISTS) {
dump_full_bpobj(&dp->dp_free_bpobj,
"Pool snapshot frees", 0);
}
if (bpobj_is_open(&dp->dp_obsolete_bpobj)) {
ASSERT(spa_feature_is_enabled(spa,
SPA_FEATURE_DEVICE_REMOVAL));
dump_full_bpobj(&dp->dp_obsolete_bpobj,
"Pool obsolete blocks", 0);
}
if (spa_feature_is_active(spa,
SPA_FEATURE_ASYNC_DESTROY)) {
dump_bptree(spa->spa_meta_objset,
dp->dp_bptree_obj,
"Pool dataset frees");
}
dump_dtl(spa->spa_root_vdev, 0);
}
for (spa_feature_t f = 0; f < SPA_FEATURES; f++)
global_feature_count[f] = UINT64_MAX;
global_feature_count[SPA_FEATURE_REDACTION_BOOKMARKS] = 0;
global_feature_count[SPA_FEATURE_REDACTION_LIST_SPILL] = 0;
global_feature_count[SPA_FEATURE_BOOKMARK_WRITTEN] = 0;
global_feature_count[SPA_FEATURE_LIVELIST] = 0;
(void) dmu_objset_find(spa_name(spa), dump_one_objset,
NULL, DS_FIND_SNAPSHOTS | DS_FIND_CHILDREN);
if (rc == 0 && !dump_opt['L'])
rc = dump_mos_leaks(spa);
for (f = 0; f < SPA_FEATURES; f++) {
uint64_t refcount;
uint64_t *arr;
if (!(spa_feature_table[f].fi_flags &
ZFEATURE_FLAG_PER_DATASET)) {
if (global_feature_count[f] == UINT64_MAX)
continue;
if (!spa_feature_is_enabled(spa, f)) {
ASSERT0(global_feature_count[f]);
continue;
}
arr = global_feature_count;
} else {
if (!spa_feature_is_enabled(spa, f)) {
ASSERT0(dataset_feature_count[f]);
continue;
}
arr = dataset_feature_count;
}
if (feature_get_refcount(spa, &spa_feature_table[f],
&refcount) == ENOTSUP)
continue;
if (arr[f] != refcount) {
(void) printf("%s feature refcount mismatch: "
"%lld consumers != %lld refcount\n",
spa_feature_table[f].fi_uname,
(longlong_t)arr[f], (longlong_t)refcount);
rc = 2;
} else {
(void) printf("Verified %s feature refcount "
"of %llu is correct\n",
spa_feature_table[f].fi_uname,
(longlong_t)refcount);
}
}
if (rc == 0)
rc = verify_device_removal_feature_counts(spa);
}
if (rc == 0 && (dump_opt['b'] || dump_opt['c']))
rc = dump_block_stats(spa);
if (rc == 0)
rc = verify_spacemap_refcounts(spa);
if (dump_opt['s'])
show_pool_stats(spa);
if (dump_opt['h'])
dump_history(spa);
if (rc == 0)
rc = verify_checkpoint(spa);
if (rc != 0) {
dump_debug_buffer();
exit(rc);
}
}
#define ZDB_FLAG_CHECKSUM 0x0001
#define ZDB_FLAG_DECOMPRESS 0x0002
#define ZDB_FLAG_BSWAP 0x0004
#define ZDB_FLAG_GBH 0x0008
#define ZDB_FLAG_INDIRECT 0x0010
#define ZDB_FLAG_RAW 0x0020
#define ZDB_FLAG_PRINT_BLKPTR 0x0040
#define ZDB_FLAG_VERBOSE 0x0080
static int flagbits[256];
static char flagbitstr[16];
static void
zdb_print_blkptr(const blkptr_t *bp, int flags)
{
char blkbuf[BP_SPRINTF_LEN];
if (flags & ZDB_FLAG_BSWAP)
byteswap_uint64_array((void *)bp, sizeof (blkptr_t));
snprintf_blkptr(blkbuf, sizeof (blkbuf), bp);
(void) printf("%s\n", blkbuf);
}
static void
zdb_dump_indirect(blkptr_t *bp, int nbps, int flags)
{
int i;
for (i = 0; i < nbps; i++)
zdb_print_blkptr(&bp[i], flags);
}
static void
zdb_dump_gbh(void *buf, int flags)
{
zdb_dump_indirect((blkptr_t *)buf, SPA_GBH_NBLKPTRS, flags);
}
static void
zdb_dump_block_raw(void *buf, uint64_t size, int flags)
{
if (flags & ZDB_FLAG_BSWAP)
byteswap_uint64_array(buf, size);
VERIFY(write(fileno(stdout), buf, size) == size);
}
static void
zdb_dump_block(char *label, void *buf, uint64_t size, int flags)
{
uint64_t *d = (uint64_t *)buf;
unsigned nwords = size / sizeof (uint64_t);
int do_bswap = !!(flags & ZDB_FLAG_BSWAP);
unsigned i, j;
const char *hdr;
char *c;
if (do_bswap)
hdr = " 7 6 5 4 3 2 1 0 f e d c b a 9 8";
else
hdr = " 0 1 2 3 4 5 6 7 8 9 a b c d e f";
(void) printf("\n%s\n%6s %s 0123456789abcdef\n", label, "", hdr);
#ifdef _LITTLE_ENDIAN
/* correct the endianness */
do_bswap = !do_bswap;
#endif
for (i = 0; i < nwords; i += 2) {
(void) printf("%06llx: %016llx %016llx ",
(u_longlong_t)(i * sizeof (uint64_t)),
(u_longlong_t)(do_bswap ? BSWAP_64(d[i]) : d[i]),
(u_longlong_t)(do_bswap ? BSWAP_64(d[i + 1]) : d[i + 1]));
c = (char *)&d[i];
for (j = 0; j < 2 * sizeof (uint64_t); j++)
(void) printf("%c", isprint(c[j]) ? c[j] : '.');
(void) printf("\n");
}
}
/*
* There are two acceptable formats:
* leaf_name - For example: c1t0d0 or /tmp/ztest.0a
* child[.child]* - For example: 0.1.1
*
* The second form can be used to specify arbitrary vdevs anywhere
* in the hierarchy. For example, in a pool with a mirror of
* RAID-Zs, you can specify either RAID-Z vdev with 0.0 or 0.1 .
*/
static vdev_t *
zdb_vdev_lookup(vdev_t *vdev, const char *path)
{
char *s, *p, *q;
unsigned i;
if (vdev == NULL)
return (NULL);
/* First, assume the x.x.x.x format */
i = strtoul(path, &s, 10);
if (s == path || (s && *s != '.' && *s != '\0'))
goto name;
if (i >= vdev->vdev_children)
return (NULL);
vdev = vdev->vdev_child[i];
if (s && *s == '\0')
return (vdev);
return (zdb_vdev_lookup(vdev, s+1));
name:
for (i = 0; i < vdev->vdev_children; i++) {
vdev_t *vc = vdev->vdev_child[i];
if (vc->vdev_path == NULL) {
vc = zdb_vdev_lookup(vc, path);
if (vc == NULL)
continue;
else
return (vc);
}
p = strrchr(vc->vdev_path, '/');
p = p ? p + 1 : vc->vdev_path;
q = &vc->vdev_path[strlen(vc->vdev_path) - 2];
if (strcmp(vc->vdev_path, path) == 0)
return (vc);
if (strcmp(p, path) == 0)
return (vc);
if (strcmp(q, "s0") == 0 && strncmp(p, path, q - p) == 0)
return (vc);
}
return (NULL);
}
static int
name_from_objset_id(spa_t *spa, uint64_t objset_id, char *outstr)
{
dsl_dataset_t *ds;
dsl_pool_config_enter(spa->spa_dsl_pool, FTAG);
int error = dsl_dataset_hold_obj(spa->spa_dsl_pool, objset_id,
NULL, &ds);
if (error != 0) {
(void) fprintf(stderr, "failed to hold objset %llu: %s\n",
(u_longlong_t)objset_id, strerror(error));
dsl_pool_config_exit(spa->spa_dsl_pool, FTAG);
return (error);
}
dsl_dataset_name(ds, outstr);
dsl_dataset_rele(ds, NULL);
dsl_pool_config_exit(spa->spa_dsl_pool, FTAG);
return (0);
}
static boolean_t
zdb_parse_block_sizes(char *sizes, uint64_t *lsize, uint64_t *psize)
{
char *s0, *s1, *tmp = NULL;
if (sizes == NULL)
return (B_FALSE);
s0 = strtok_r(sizes, "/", &tmp);
if (s0 == NULL)
return (B_FALSE);
s1 = strtok_r(NULL, "/", &tmp);
*lsize = strtoull(s0, NULL, 16);
*psize = s1 ? strtoull(s1, NULL, 16) : *lsize;
return (*lsize >= *psize && *psize > 0);
}
#define ZIO_COMPRESS_MASK(alg) (1ULL << (ZIO_COMPRESS_##alg))
static boolean_t
zdb_decompress_block(abd_t *pabd, void *buf, void *lbuf, uint64_t lsize,
uint64_t psize, int flags)
{
(void) buf;
boolean_t exceeded = B_FALSE;
/*
* We don't know how the data was compressed, so just try
* every decompress function at every inflated blocksize.
*/
void *lbuf2 = umem_alloc(SPA_MAXBLOCKSIZE, UMEM_NOFAIL);
int cfuncs[ZIO_COMPRESS_FUNCTIONS] = { 0 };
int *cfuncp = cfuncs;
uint64_t maxlsize = SPA_MAXBLOCKSIZE;
uint64_t mask = ZIO_COMPRESS_MASK(ON) | ZIO_COMPRESS_MASK(OFF) |
ZIO_COMPRESS_MASK(INHERIT) | ZIO_COMPRESS_MASK(EMPTY) |
(getenv("ZDB_NO_ZLE") ? ZIO_COMPRESS_MASK(ZLE) : 0);
*cfuncp++ = ZIO_COMPRESS_LZ4;
*cfuncp++ = ZIO_COMPRESS_LZJB;
mask |= ZIO_COMPRESS_MASK(LZ4) | ZIO_COMPRESS_MASK(LZJB);
for (int c = 0; c < ZIO_COMPRESS_FUNCTIONS; c++)
if (((1ULL << c) & mask) == 0)
*cfuncp++ = c;
/*
* On the one hand, with SPA_MAXBLOCKSIZE at 16MB, this
* could take a while and we should let the user know
* we are not stuck. On the other hand, printing progress
* info gets old after a while. User can specify 'v' flag
* to see the progression.
*/
if (lsize == psize)
lsize += SPA_MINBLOCKSIZE;
else
maxlsize = lsize;
for (; lsize <= maxlsize; lsize += SPA_MINBLOCKSIZE) {
for (cfuncp = cfuncs; *cfuncp; cfuncp++) {
if (flags & ZDB_FLAG_VERBOSE) {
(void) fprintf(stderr,
"Trying %05llx -> %05llx (%s)\n",
(u_longlong_t)psize,
(u_longlong_t)lsize,
zio_compress_table[*cfuncp].\
ci_name);
}
/*
* We randomize lbuf2, and decompress to both
* lbuf and lbuf2. This way, we will know if
* decompression fill exactly to lsize.
*/
VERIFY0(random_get_pseudo_bytes(lbuf2, lsize));
if (zio_decompress_data(*cfuncp, pabd,
lbuf, psize, lsize, NULL) == 0 &&
zio_decompress_data(*cfuncp, pabd,
lbuf2, psize, lsize, NULL) == 0 &&
memcmp(lbuf, lbuf2, lsize) == 0)
break;
}
if (*cfuncp != 0)
break;
}
umem_free(lbuf2, SPA_MAXBLOCKSIZE);
if (lsize > maxlsize) {
exceeded = B_TRUE;
}
if (*cfuncp == ZIO_COMPRESS_ZLE) {
printf("\nZLE decompression was selected. If you "
"suspect the results are wrong,\ntry avoiding ZLE "
"by setting and exporting ZDB_NO_ZLE=\"true\"\n");
}
return (exceeded);
}
/*
* Read a block from a pool and print it out. The syntax of the
* block descriptor is:
*
* pool:vdev_specifier:offset:[lsize/]psize[:flags]
*
* pool - The name of the pool you wish to read from
* vdev_specifier - Which vdev (see comment for zdb_vdev_lookup)
* offset - offset, in hex, in bytes
* size - Amount of data to read, in hex, in bytes
* flags - A string of characters specifying options
* b: Decode a blkptr at given offset within block
* c: Calculate and display checksums
* d: Decompress data before dumping
* e: Byteswap data before dumping
* g: Display data as a gang block header
* i: Display as an indirect block
* r: Dump raw data to stdout
* v: Verbose
*
*/
static void
zdb_read_block(char *thing, spa_t *spa)
{
blkptr_t blk, *bp = &blk;
dva_t *dva = bp->blk_dva;
int flags = 0;
uint64_t offset = 0, psize = 0, lsize = 0, blkptr_offset = 0;
zio_t *zio;
vdev_t *vd;
abd_t *pabd;
void *lbuf, *buf;
char *s, *p, *dup, *flagstr, *sizes, *tmp = NULL;
const char *vdev, *errmsg = NULL;
int i, error;
boolean_t borrowed = B_FALSE, found = B_FALSE;
dup = strdup(thing);
s = strtok_r(dup, ":", &tmp);
vdev = s ?: "";
s = strtok_r(NULL, ":", &tmp);
offset = strtoull(s ? s : "", NULL, 16);
sizes = strtok_r(NULL, ":", &tmp);
s = strtok_r(NULL, ":", &tmp);
flagstr = strdup(s ?: "");
if (!zdb_parse_block_sizes(sizes, &lsize, &psize))
errmsg = "invalid size(s)";
if (!IS_P2ALIGNED(psize, DEV_BSIZE) || !IS_P2ALIGNED(lsize, DEV_BSIZE))
errmsg = "size must be a multiple of sector size";
if (!IS_P2ALIGNED(offset, DEV_BSIZE))
errmsg = "offset must be a multiple of sector size";
if (errmsg) {
(void) printf("Invalid block specifier: %s - %s\n",
thing, errmsg);
goto done;
}
tmp = NULL;
for (s = strtok_r(flagstr, ":", &tmp);
s != NULL;
s = strtok_r(NULL, ":", &tmp)) {
for (i = 0; i < strlen(flagstr); i++) {
int bit = flagbits[(uchar_t)flagstr[i]];
if (bit == 0) {
(void) printf("***Ignoring flag: %c\n",
(uchar_t)flagstr[i]);
continue;
}
found = B_TRUE;
flags |= bit;
p = &flagstr[i + 1];
if (*p != ':' && *p != '\0') {
int j = 0, nextbit = flagbits[(uchar_t)*p];
char *end, offstr[8] = { 0 };
if ((bit == ZDB_FLAG_PRINT_BLKPTR) &&
(nextbit == 0)) {
/* look ahead to isolate the offset */
while (nextbit == 0 &&
strchr(flagbitstr, *p) == NULL) {
offstr[j] = *p;
j++;
if (i + j > strlen(flagstr))
break;
p++;
nextbit = flagbits[(uchar_t)*p];
}
blkptr_offset = strtoull(offstr, &end,
16);
i += j;
} else if (nextbit == 0) {
(void) printf("***Ignoring flag arg:"
" '%c'\n", (uchar_t)*p);
}
}
}
}
if (blkptr_offset % sizeof (blkptr_t)) {
printf("Block pointer offset 0x%llx "
"must be divisible by 0x%x\n",
(longlong_t)blkptr_offset, (int)sizeof (blkptr_t));
goto done;
}
if (found == B_FALSE && strlen(flagstr) > 0) {
printf("Invalid flag arg: '%s'\n", flagstr);
goto done;
}
vd = zdb_vdev_lookup(spa->spa_root_vdev, vdev);
if (vd == NULL) {
(void) printf("***Invalid vdev: %s\n", vdev);
goto done;
} else {
if (vd->vdev_path)
(void) fprintf(stderr, "Found vdev: %s\n",
vd->vdev_path);
else
(void) fprintf(stderr, "Found vdev type: %s\n",
vd->vdev_ops->vdev_op_type);
}
pabd = abd_alloc_for_io(SPA_MAXBLOCKSIZE, B_FALSE);
lbuf = umem_alloc(SPA_MAXBLOCKSIZE, UMEM_NOFAIL);
BP_ZERO(bp);
DVA_SET_VDEV(&dva[0], vd->vdev_id);
DVA_SET_OFFSET(&dva[0], offset);
DVA_SET_GANG(&dva[0], !!(flags & ZDB_FLAG_GBH));
DVA_SET_ASIZE(&dva[0], vdev_psize_to_asize(vd, psize));
BP_SET_BIRTH(bp, TXG_INITIAL, TXG_INITIAL);
BP_SET_LSIZE(bp, lsize);
BP_SET_PSIZE(bp, psize);
BP_SET_COMPRESS(bp, ZIO_COMPRESS_OFF);
BP_SET_CHECKSUM(bp, ZIO_CHECKSUM_OFF);
BP_SET_TYPE(bp, DMU_OT_NONE);
BP_SET_LEVEL(bp, 0);
BP_SET_DEDUP(bp, 0);
BP_SET_BYTEORDER(bp, ZFS_HOST_BYTEORDER);
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
zio = zio_root(spa, NULL, NULL, 0);
if (vd == vd->vdev_top) {
/*
* Treat this as a normal block read.
*/
zio_nowait(zio_read(zio, spa, bp, pabd, psize, NULL, NULL,
ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW, NULL));
} else {
/*
* Treat this as a vdev child I/O.
*/
zio_nowait(zio_vdev_child_io(zio, bp, vd, offset, pabd,
psize, ZIO_TYPE_READ, ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY |
ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW | ZIO_FLAG_OPTIONAL,
NULL, NULL));
}
error = zio_wait(zio);
spa_config_exit(spa, SCL_STATE, FTAG);
if (error) {
(void) printf("Read of %s failed, error: %d\n", thing, error);
goto out;
}
uint64_t orig_lsize = lsize;
buf = lbuf;
if (flags & ZDB_FLAG_DECOMPRESS) {
boolean_t failed = zdb_decompress_block(pabd, buf, lbuf,
lsize, psize, flags);
if (failed) {
(void) printf("Decompress of %s failed\n", thing);
goto out;
}
} else {
buf = abd_borrow_buf_copy(pabd, lsize);
borrowed = B_TRUE;
}
/*
* Try to detect invalid block pointer. If invalid, try
* decompressing.
*/
if ((flags & ZDB_FLAG_PRINT_BLKPTR || flags & ZDB_FLAG_INDIRECT) &&
!(flags & ZDB_FLAG_DECOMPRESS)) {
const blkptr_t *b = (const blkptr_t *)(void *)
((uintptr_t)buf + (uintptr_t)blkptr_offset);
if (zfs_blkptr_verify(spa, b,
BLK_CONFIG_NEEDED, BLK_VERIFY_ONLY) == B_FALSE) {
abd_return_buf_copy(pabd, buf, lsize);
borrowed = B_FALSE;
buf = lbuf;
boolean_t failed = zdb_decompress_block(pabd, buf,
lbuf, lsize, psize, flags);
b = (const blkptr_t *)(void *)
((uintptr_t)buf + (uintptr_t)blkptr_offset);
if (failed || zfs_blkptr_verify(spa, b,
BLK_CONFIG_NEEDED, BLK_VERIFY_LOG) == B_FALSE) {
printf("invalid block pointer at this DVA\n");
goto out;
}
}
}
if (flags & ZDB_FLAG_PRINT_BLKPTR)
zdb_print_blkptr((blkptr_t *)(void *)
((uintptr_t)buf + (uintptr_t)blkptr_offset), flags);
else if (flags & ZDB_FLAG_RAW)
zdb_dump_block_raw(buf, lsize, flags);
else if (flags & ZDB_FLAG_INDIRECT)
zdb_dump_indirect((blkptr_t *)buf,
orig_lsize / sizeof (blkptr_t), flags);
else if (flags & ZDB_FLAG_GBH)
zdb_dump_gbh(buf, flags);
else
zdb_dump_block(thing, buf, lsize, flags);
/*
* If :c was specified, iterate through the checksum table to
* calculate and display each checksum for our specified
* DVA and length.
*/
if ((flags & ZDB_FLAG_CHECKSUM) && !(flags & ZDB_FLAG_RAW) &&
!(flags & ZDB_FLAG_GBH)) {
zio_t *czio;
(void) printf("\n");
for (enum zio_checksum ck = ZIO_CHECKSUM_LABEL;
ck < ZIO_CHECKSUM_FUNCTIONS; ck++) {
if ((zio_checksum_table[ck].ci_flags &
ZCHECKSUM_FLAG_EMBEDDED) ||
ck == ZIO_CHECKSUM_NOPARITY) {
continue;
}
BP_SET_CHECKSUM(bp, ck);
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
czio = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL);
if (vd == vd->vdev_top) {
zio_nowait(zio_read(czio, spa, bp, pabd, psize,
NULL, NULL,
ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW |
ZIO_FLAG_DONT_RETRY, NULL));
} else {
zio_nowait(zio_vdev_child_io(czio, bp, vd,
offset, pabd, psize, ZIO_TYPE_READ,
ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_DONT_PROPAGATE |
ZIO_FLAG_DONT_RETRY |
ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW |
ZIO_FLAG_SPECULATIVE |
ZIO_FLAG_OPTIONAL, NULL, NULL));
}
error = zio_wait(czio);
if (error == 0 || error == ECKSUM) {
zio_t *ck_zio = zio_null(NULL, spa, NULL,
NULL, NULL, 0);
ck_zio->io_offset =
DVA_GET_OFFSET(&bp->blk_dva[0]);
ck_zio->io_bp = bp;
zio_checksum_compute(ck_zio, ck, pabd, lsize);
printf(
"%12s\t"
"cksum=%016llx:%016llx:%016llx:%016llx\n",
zio_checksum_table[ck].ci_name,
(u_longlong_t)bp->blk_cksum.zc_word[0],
(u_longlong_t)bp->blk_cksum.zc_word[1],
(u_longlong_t)bp->blk_cksum.zc_word[2],
(u_longlong_t)bp->blk_cksum.zc_word[3]);
zio_wait(ck_zio);
} else {
printf("error %d reading block\n", error);
}
spa_config_exit(spa, SCL_STATE, FTAG);
}
}
if (borrowed)
abd_return_buf_copy(pabd, buf, lsize);
out:
abd_free(pabd);
umem_free(lbuf, SPA_MAXBLOCKSIZE);
done:
free(flagstr);
free(dup);
}
static void
zdb_embedded_block(char *thing)
{
blkptr_t bp = {{{{0}}}};
unsigned long long *words = (void *)&bp;
char *buf;
int err;
err = sscanf(thing, "%llx:%llx:%llx:%llx:%llx:%llx:%llx:%llx:"
"%llx:%llx:%llx:%llx:%llx:%llx:%llx:%llx",
words + 0, words + 1, words + 2, words + 3,
words + 4, words + 5, words + 6, words + 7,
words + 8, words + 9, words + 10, words + 11,
words + 12, words + 13, words + 14, words + 15);
if (err != 16) {
(void) fprintf(stderr, "invalid input format\n");
exit(1);
}
ASSERT3U(BPE_GET_LSIZE(&bp), <=, SPA_MAXBLOCKSIZE);
buf = malloc(SPA_MAXBLOCKSIZE);
if (buf == NULL) {
(void) fprintf(stderr, "out of memory\n");
exit(1);
}
err = decode_embedded_bp(&bp, buf, BPE_GET_LSIZE(&bp));
if (err != 0) {
(void) fprintf(stderr, "decode failed: %u\n", err);
exit(1);
}
zdb_dump_block_raw(buf, BPE_GET_LSIZE(&bp), 0);
free(buf);
}
/* check for valid hex or decimal numeric string */
static boolean_t
zdb_numeric(char *str)
{
int i = 0;
if (strlen(str) == 0)
return (B_FALSE);
if (strncmp(str, "0x", 2) == 0 || strncmp(str, "0X", 2) == 0)
i = 2;
for (; i < strlen(str); i++) {
if (!isxdigit(str[i]))
return (B_FALSE);
}
return (B_TRUE);
}
int
main(int argc, char **argv)
{
int c;
spa_t *spa = NULL;
objset_t *os = NULL;
int dump_all = 1;
int verbose = 0;
int error = 0;
char **searchdirs = NULL;
int nsearch = 0;
char *target, *target_pool, dsname[ZFS_MAX_DATASET_NAME_LEN];
nvlist_t *policy = NULL;
uint64_t max_txg = UINT64_MAX;
int64_t objset_id = -1;
uint64_t object;
int flags = ZFS_IMPORT_MISSING_LOG;
int rewind = ZPOOL_NEVER_REWIND;
char *spa_config_path_env, *objset_str;
boolean_t target_is_spa = B_TRUE, dataset_lookup = B_FALSE;
nvlist_t *cfg = NULL;
dprintf_setup(&argc, argv);
/*
* If there is an environment variable SPA_CONFIG_PATH it overrides
* default spa_config_path setting. If -U flag is specified it will
* override this environment variable settings once again.
*/
spa_config_path_env = getenv("SPA_CONFIG_PATH");
if (spa_config_path_env != NULL)
spa_config_path = spa_config_path_env;
/*
* For performance reasons, we set this tunable down. We do so before
* the arg parsing section so that the user can override this value if
* they choose.
*/
zfs_btree_verify_intensity = 3;
struct option long_options[] = {
{"ignore-assertions", no_argument, NULL, 'A'},
{"block-stats", no_argument, NULL, 'b'},
{"backup", no_argument, NULL, 'B'},
{"checksum", no_argument, NULL, 'c'},
{"config", no_argument, NULL, 'C'},
{"datasets", no_argument, NULL, 'd'},
{"dedup-stats", no_argument, NULL, 'D'},
{"exported", no_argument, NULL, 'e'},
{"embedded-block-pointer", no_argument, NULL, 'E'},
{"automatic-rewind", no_argument, NULL, 'F'},
{"dump-debug-msg", no_argument, NULL, 'G'},
{"history", no_argument, NULL, 'h'},
{"intent-logs", no_argument, NULL, 'i'},
{"inflight", required_argument, NULL, 'I'},
{"checkpointed-state", no_argument, NULL, 'k'},
{"key", required_argument, NULL, 'K'},
{"label", no_argument, NULL, 'l'},
{"disable-leak-tracking", no_argument, NULL, 'L'},
{"metaslabs", no_argument, NULL, 'm'},
{"metaslab-groups", no_argument, NULL, 'M'},
{"numeric", no_argument, NULL, 'N'},
{"option", required_argument, NULL, 'o'},
{"object-lookups", no_argument, NULL, 'O'},
{"path", required_argument, NULL, 'p'},
{"parseable", no_argument, NULL, 'P'},
{"skip-label", no_argument, NULL, 'q'},
{"copy-object", no_argument, NULL, 'r'},
{"read-block", no_argument, NULL, 'R'},
{"io-stats", no_argument, NULL, 's'},
{"simulate-dedup", no_argument, NULL, 'S'},
{"txg", required_argument, NULL, 't'},
{"uberblock", no_argument, NULL, 'u'},
{"cachefile", required_argument, NULL, 'U'},
{"verbose", no_argument, NULL, 'v'},
{"verbatim", no_argument, NULL, 'V'},
{"dump-blocks", required_argument, NULL, 'x'},
{"extreme-rewind", no_argument, NULL, 'X'},
{"all-reconstruction", no_argument, NULL, 'Y'},
{"livelist", no_argument, NULL, 'y'},
{"zstd-headers", no_argument, NULL, 'Z'},
{0, 0, 0, 0}
};
while ((c = getopt_long(argc, argv,
"AbBcCdDeEFGhiI:kK:lLmMNo:Op:PqrRsSt:uU:vVx:XYyZ",
long_options, NULL)) != -1) {
switch (c) {
case 'b':
case 'B':
case 'c':
case 'C':
case 'd':
case 'D':
case 'E':
case 'G':
case 'h':
case 'i':
case 'l':
case 'm':
case 'M':
case 'N':
case 'O':
case 'r':
case 'R':
case 's':
case 'S':
case 'u':
case 'y':
case 'Z':
dump_opt[c]++;
dump_all = 0;
break;
case 'A':
case 'e':
case 'F':
case 'k':
case 'L':
case 'P':
case 'q':
case 'X':
dump_opt[c]++;
break;
case 'Y':
zfs_reconstruct_indirect_combinations_max = INT_MAX;
zfs_deadman_enabled = 0;
break;
/* NB: Sort single match options below. */
case 'I':
max_inflight_bytes = strtoull(optarg, NULL, 0);
if (max_inflight_bytes == 0) {
(void) fprintf(stderr, "maximum number "
"of inflight bytes must be greater "
"than 0\n");
usage();
}
break;
case 'K':
dump_opt[c]++;
key_material = strdup(optarg);
/* redact key material in process table */
while (*optarg != '\0') { *optarg++ = '*'; }
break;
case 'o':
error = set_global_var(optarg);
if (error != 0)
usage();
break;
case 'p':
if (searchdirs == NULL) {
searchdirs = umem_alloc(sizeof (char *),
UMEM_NOFAIL);
} else {
char **tmp = umem_alloc((nsearch + 1) *
sizeof (char *), UMEM_NOFAIL);
memcpy(tmp, searchdirs, nsearch *
sizeof (char *));
umem_free(searchdirs,
nsearch * sizeof (char *));
searchdirs = tmp;
}
searchdirs[nsearch++] = optarg;
break;
case 't':
max_txg = strtoull(optarg, NULL, 0);
if (max_txg < TXG_INITIAL) {
(void) fprintf(stderr, "incorrect txg "
"specified: %s\n", optarg);
usage();
}
break;
case 'U':
spa_config_path = optarg;
if (spa_config_path[0] != '/') {
(void) fprintf(stderr,
"cachefile must be an absolute path "
"(i.e. start with a slash)\n");
usage();
}
break;
case 'v':
verbose++;
break;
case 'V':
flags = ZFS_IMPORT_VERBATIM;
break;
case 'x':
vn_dumpdir = optarg;
break;
default:
usage();
break;
}
}
if (!dump_opt['e'] && searchdirs != NULL) {
(void) fprintf(stderr, "-p option requires use of -e\n");
usage();
}
#if defined(_LP64)
/*
* ZDB does not typically re-read blocks; therefore limit the ARC
* to 256 MB, which can be used entirely for metadata.
*/
zfs_arc_min = 2ULL << SPA_MAXBLOCKSHIFT;
zfs_arc_max = 256 * 1024 * 1024;
#endif
/*
* "zdb -c" uses checksum-verifying scrub i/os which are async reads.
* "zdb -b" uses traversal prefetch which uses async reads.
* For good performance, let several of them be active at once.
*/
zfs_vdev_async_read_max_active = 10;
/*
* Disable reference tracking for better performance.
*/
reference_tracking_enable = B_FALSE;
/*
* Do not fail spa_load when spa_load_verify fails. This is needed
* to load non-idle pools.
*/
spa_load_verify_dryrun = B_TRUE;
/*
* ZDB should have ability to read spacemaps.
*/
spa_mode_readable_spacemaps = B_TRUE;
kernel_init(SPA_MODE_READ);
if (dump_all)
verbose = MAX(verbose, 1);
for (c = 0; c < 256; c++) {
if (dump_all && strchr("ABeEFkKlLNOPrRSXy", c) == NULL)
dump_opt[c] = 1;
if (dump_opt[c])
dump_opt[c] += verbose;
}
libspl_set_assert_ok((dump_opt['A'] == 1) || (dump_opt['A'] > 2));
zfs_recover = (dump_opt['A'] > 1);
argc -= optind;
argv += optind;
if (argc < 2 && dump_opt['R'])
usage();
if (dump_opt['E']) {
if (argc != 1)
usage();
zdb_embedded_block(argv[0]);
return (0);
}
if (argc < 1) {
if (!dump_opt['e'] && dump_opt['C']) {
dump_cachefile(spa_config_path);
return (0);
}
usage();
}
if (dump_opt['l'])
return (dump_label(argv[0]));
if (dump_opt['O']) {
if (argc != 2)
usage();
dump_opt['v'] = verbose + 3;
return (dump_path(argv[0], argv[1], NULL));
}
if (dump_opt['r']) {
target_is_spa = B_FALSE;
if (argc != 3)
usage();
dump_opt['v'] = verbose;
error = dump_path(argv[0], argv[1], &object);
if (error != 0)
fatal("internal error: %s", strerror(error));
}
if (dump_opt['X'] || dump_opt['F'])
rewind = ZPOOL_DO_REWIND |
(dump_opt['X'] ? ZPOOL_EXTREME_REWIND : 0);
/* -N implies -d */
if (dump_opt['N'] && dump_opt['d'] == 0)
dump_opt['d'] = dump_opt['N'];
if (nvlist_alloc(&policy, NV_UNIQUE_NAME_TYPE, 0) != 0 ||
nvlist_add_uint64(policy, ZPOOL_LOAD_REQUEST_TXG, max_txg) != 0 ||
nvlist_add_uint32(policy, ZPOOL_LOAD_REWIND_POLICY, rewind) != 0)
fatal("internal error: %s", strerror(ENOMEM));
error = 0;
target = argv[0];
if (strpbrk(target, "/@") != NULL) {
size_t targetlen;
target_pool = strdup(target);
*strpbrk(target_pool, "/@") = '\0';
target_is_spa = B_FALSE;
targetlen = strlen(target);
if (targetlen && target[targetlen - 1] == '/')
target[targetlen - 1] = '\0';
/*
* See if an objset ID was supplied (-d <pool>/<objset ID>).
* To disambiguate tank/100, consider the 100 as objsetID
* if -N was given, otherwise 100 is an objsetID iff
* tank/100 as a named dataset fails on lookup.
*/
objset_str = strchr(target, '/');
if (objset_str && strlen(objset_str) > 1 &&
zdb_numeric(objset_str + 1)) {
char *endptr;
errno = 0;
objset_str++;
objset_id = strtoull(objset_str, &endptr, 0);
/* dataset 0 is the same as opening the pool */
if (errno == 0 && endptr != objset_str &&
objset_id != 0) {
if (dump_opt['N'])
dataset_lookup = B_TRUE;
}
/* normal dataset name not an objset ID */
if (endptr == objset_str) {
objset_id = -1;
}
} else if (objset_str && !zdb_numeric(objset_str + 1) &&
dump_opt['N']) {
printf("Supply a numeric objset ID with -N\n");
exit(1);
}
} else {
target_pool = target;
}
if (dump_opt['e']) {
importargs_t args = { 0 };
args.paths = nsearch;
args.path = searchdirs;
args.can_be_active = B_TRUE;
libpc_handle_t lpch = {
.lpc_lib_handle = NULL,
.lpc_ops = &libzpool_config_ops,
.lpc_printerr = B_TRUE
};
error = zpool_find_config(&lpch, target_pool, &cfg, &args);
if (error == 0) {
if (nvlist_add_nvlist(cfg,
ZPOOL_LOAD_POLICY, policy) != 0) {
fatal("can't open '%s': %s",
target, strerror(ENOMEM));
}
if (dump_opt['C'] > 1) {
(void) printf("\nConfiguration for import:\n");
dump_nvlist(cfg, 8);
}
/*
* Disable the activity check to allow examination of
* active pools.
*/
error = spa_import(target_pool, cfg, NULL,
flags | ZFS_IMPORT_SKIP_MMP);
}
}
if (searchdirs != NULL) {
umem_free(searchdirs, nsearch * sizeof (char *));
searchdirs = NULL;
}
/*
* import_checkpointed_state makes the assumption that the
* target pool that we pass it is already part of the spa
* namespace. Because of that we need to make sure to call
* it always after the -e option has been processed, which
* imports the pool to the namespace if it's not in the
* cachefile.
*/
char *checkpoint_pool = NULL;
char *checkpoint_target = NULL;
if (dump_opt['k']) {
checkpoint_pool = import_checkpointed_state(target, cfg,
&checkpoint_target);
if (checkpoint_target != NULL)
target = checkpoint_target;
}
if (cfg != NULL) {
nvlist_free(cfg);
cfg = NULL;
}
if (target_pool != target)
free(target_pool);
if (error == 0) {
if (dump_opt['k'] && (target_is_spa || dump_opt['R'])) {
ASSERT(checkpoint_pool != NULL);
ASSERT(checkpoint_target == NULL);
error = spa_open(checkpoint_pool, &spa, FTAG);
if (error != 0) {
fatal("Tried to open pool \"%s\" but "
"spa_open() failed with error %d\n",
checkpoint_pool, error);
}
} else if (target_is_spa || dump_opt['R'] || dump_opt['B'] ||
objset_id == 0) {
zdb_set_skip_mmp(target);
error = spa_open_rewind(target, &spa, FTAG, policy,
NULL);
if (error) {
/*
* If we're missing the log device then
* try opening the pool after clearing the
* log state.
*/
mutex_enter(&spa_namespace_lock);
if ((spa = spa_lookup(target)) != NULL &&
spa->spa_log_state == SPA_LOG_MISSING) {
spa->spa_log_state = SPA_LOG_CLEAR;
error = 0;
}
mutex_exit(&spa_namespace_lock);
if (!error) {
error = spa_open_rewind(target, &spa,
FTAG, policy, NULL);
}
}
} else if (strpbrk(target, "#") != NULL) {
dsl_pool_t *dp;
error = dsl_pool_hold(target, FTAG, &dp);
if (error != 0) {
fatal("can't dump '%s': %s", target,
strerror(error));
}
error = dump_bookmark(dp, target, B_TRUE, verbose > 1);
dsl_pool_rele(dp, FTAG);
if (error != 0) {
fatal("can't dump '%s': %s", target,
strerror(error));
}
return (error);
} else {
target_pool = strdup(target);
if (strpbrk(target, "/@") != NULL)
*strpbrk(target_pool, "/@") = '\0';
zdb_set_skip_mmp(target);
/*
* If -N was supplied, the user has indicated that
* zdb -d <pool>/<objsetID> is in effect. Otherwise
* we first assume that the dataset string is the
* dataset name. If dmu_objset_hold fails with the
* dataset string, and we have an objset_id, retry the
* lookup with the objsetID.
*/
boolean_t retry = B_TRUE;
retry_lookup:
if (dataset_lookup == B_TRUE) {
/*
* Use the supplied id to get the name
* for open_objset.
*/
error = spa_open(target_pool, &spa, FTAG);
if (error == 0) {
error = name_from_objset_id(spa,
objset_id, dsname);
spa_close(spa, FTAG);
if (error == 0)
target = dsname;
}
}
if (error == 0) {
if (objset_id > 0 && retry) {
int err = dmu_objset_hold(target, FTAG,
&os);
if (err) {
dataset_lookup = B_TRUE;
retry = B_FALSE;
goto retry_lookup;
} else {
dmu_objset_rele(os, FTAG);
}
}
error = open_objset(target, FTAG, &os);
}
if (error == 0)
spa = dmu_objset_spa(os);
free(target_pool);
}
}
nvlist_free(policy);
if (error)
fatal("can't open '%s': %s", target, strerror(error));
/*
* Set the pool failure mode to panic in order to prevent the pool
* from suspending. A suspended I/O will have no way to resume and
* can prevent the zdb(8) command from terminating as expected.
*/
if (spa != NULL)
spa->spa_failmode = ZIO_FAILURE_MODE_PANIC;
argv++;
argc--;
if (dump_opt['r']) {
error = zdb_copy_object(os, object, argv[1]);
} else if (!dump_opt['R']) {
flagbits['d'] = ZOR_FLAG_DIRECTORY;
flagbits['f'] = ZOR_FLAG_PLAIN_FILE;
flagbits['m'] = ZOR_FLAG_SPACE_MAP;
flagbits['z'] = ZOR_FLAG_ZAP;
flagbits['A'] = ZOR_FLAG_ALL_TYPES;
if (argc > 0 && dump_opt['d']) {
zopt_object_args = argc;
zopt_object_ranges = calloc(zopt_object_args,
sizeof (zopt_object_range_t));
for (unsigned i = 0; i < zopt_object_args; i++) {
int err;
const char *msg = NULL;
err = parse_object_range(argv[i],
&zopt_object_ranges[i], &msg);
if (err != 0)
fatal("Bad object or range: '%s': %s\n",
argv[i], msg ?: "");
}
} else if (argc > 0 && dump_opt['m']) {
zopt_metaslab_args = argc;
zopt_metaslab = calloc(zopt_metaslab_args,
sizeof (uint64_t));
for (unsigned i = 0; i < zopt_metaslab_args; i++) {
errno = 0;
zopt_metaslab[i] = strtoull(argv[i], NULL, 0);
if (zopt_metaslab[i] == 0 && errno != 0)
fatal("bad number %s: %s", argv[i],
strerror(errno));
}
}
if (dump_opt['B']) {
dump_backup(target, objset_id,
argc > 0 ? argv[0] : NULL);
} else if (os != NULL) {
dump_objset(os);
} else if (zopt_object_args > 0 && !dump_opt['m']) {
dump_objset(spa->spa_meta_objset);
} else {
dump_zpool(spa);
}
} else {
flagbits['b'] = ZDB_FLAG_PRINT_BLKPTR;
flagbits['c'] = ZDB_FLAG_CHECKSUM;
flagbits['d'] = ZDB_FLAG_DECOMPRESS;
flagbits['e'] = ZDB_FLAG_BSWAP;
flagbits['g'] = ZDB_FLAG_GBH;
flagbits['i'] = ZDB_FLAG_INDIRECT;
flagbits['r'] = ZDB_FLAG_RAW;
flagbits['v'] = ZDB_FLAG_VERBOSE;
for (int i = 0; i < argc; i++)
zdb_read_block(argv[i], spa);
}
if (dump_opt['k']) {
free(checkpoint_pool);
if (!target_is_spa)
free(checkpoint_target);
}
if (os != NULL) {
close_objset(os, FTAG);
} else {
spa_close(spa, FTAG);
}
fuid_table_destroy();
dump_debug_buffer();
kernel_fini();
return (error);
}
diff --git a/sys/contrib/openzfs/cmd/zpool/zpool_main.c b/sys/contrib/openzfs/cmd/zpool/zpool_main.c
index 5507f9d3fd67..9dd1d2109004 100644
--- a/sys/contrib/openzfs/cmd/zpool/zpool_main.c
+++ b/sys/contrib/openzfs/cmd/zpool/zpool_main.c
@@ -1,11237 +1,11355 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright 2011 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright (c) 2012 by Frederik Wessels. All rights reserved.
* Copyright (c) 2012 by Cyril Plisko. All rights reserved.
* Copyright (c) 2013 by Prasad Joshi (sTec). All rights reserved.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>.
* Copyright (c) 2017 Datto Inc.
* Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019, loli10K <ezomori.nozomu@gmail.com>
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
* Copyright (c) 2021, Klara Inc.
* Copyright [2021] Hewlett Packard Enterprise Development LP
*/
#include <assert.h>
#include <ctype.h>
#include <dirent.h>
#include <errno.h>
#include <fcntl.h>
#include <getopt.h>
#include <libgen.h>
#include <libintl.h>
#include <libuutil.h>
#include <locale.h>
#include <pthread.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <time.h>
#include <unistd.h>
#include <pwd.h>
#include <zone.h>
#include <sys/wait.h>
#include <zfs_prop.h>
#include <sys/fs/zfs.h>
#include <sys/stat.h>
#include <sys/systeminfo.h>
#include <sys/fm/fs/zfs.h>
#include <sys/fm/util.h>
#include <sys/fm/protocol.h>
#include <sys/zfs_ioctl.h>
#include <sys/mount.h>
#include <sys/sysmacros.h>
#include <math.h>
#include <libzfs.h>
#include <libzutil.h>
#include "zpool_util.h"
#include "zfs_comutil.h"
#include "zfeature_common.h"
#include "statcommon.h"
libzfs_handle_t *g_zfs;
static int zpool_do_create(int, char **);
static int zpool_do_destroy(int, char **);
static int zpool_do_add(int, char **);
static int zpool_do_remove(int, char **);
static int zpool_do_labelclear(int, char **);
static int zpool_do_checkpoint(int, char **);
static int zpool_do_list(int, char **);
static int zpool_do_iostat(int, char **);
static int zpool_do_status(int, char **);
static int zpool_do_online(int, char **);
static int zpool_do_offline(int, char **);
static int zpool_do_clear(int, char **);
static int zpool_do_reopen(int, char **);
static int zpool_do_reguid(int, char **);
static int zpool_do_attach(int, char **);
static int zpool_do_detach(int, char **);
static int zpool_do_replace(int, char **);
static int zpool_do_split(int, char **);
static int zpool_do_initialize(int, char **);
static int zpool_do_scrub(int, char **);
static int zpool_do_resilver(int, char **);
static int zpool_do_trim(int, char **);
static int zpool_do_import(int, char **);
static int zpool_do_export(int, char **);
static int zpool_do_upgrade(int, char **);
static int zpool_do_history(int, char **);
static int zpool_do_events(int, char **);
static int zpool_do_get(int, char **);
static int zpool_do_set(int, char **);
static int zpool_do_sync(int, char **);
static int zpool_do_version(int, char **);
static int zpool_do_wait(int, char **);
static int zpool_do_help(int argc, char **argv);
static zpool_compat_status_t zpool_do_load_compat(
const char *, boolean_t *);
/*
* These libumem hooks provide a reasonable set of defaults for the allocator's
* debugging facilities.
*/
#ifdef DEBUG
const char *
_umem_debug_init(void)
{
return ("default,verbose"); /* $UMEM_DEBUG setting */
}
const char *
_umem_logging_init(void)
{
return ("fail,contents"); /* $UMEM_LOGGING setting */
}
#endif
typedef enum {
HELP_ADD,
HELP_ATTACH,
HELP_CLEAR,
HELP_CREATE,
HELP_CHECKPOINT,
HELP_DESTROY,
HELP_DETACH,
HELP_EXPORT,
HELP_HISTORY,
HELP_IMPORT,
HELP_IOSTAT,
HELP_LABELCLEAR,
HELP_LIST,
HELP_OFFLINE,
HELP_ONLINE,
HELP_REPLACE,
HELP_REMOVE,
HELP_INITIALIZE,
HELP_SCRUB,
HELP_RESILVER,
HELP_TRIM,
HELP_STATUS,
HELP_UPGRADE,
HELP_EVENTS,
HELP_GET,
HELP_SET,
HELP_SPLIT,
HELP_SYNC,
HELP_REGUID,
HELP_REOPEN,
HELP_VERSION,
HELP_WAIT
} zpool_help_t;
/*
* Flags for stats to display with "zpool iostats"
*/
enum iostat_type {
IOS_DEFAULT = 0,
IOS_LATENCY = 1,
IOS_QUEUES = 2,
IOS_L_HISTO = 3,
IOS_RQ_HISTO = 4,
IOS_COUNT, /* always last element */
};
/* iostat_type entries as bitmasks */
#define IOS_DEFAULT_M (1ULL << IOS_DEFAULT)
#define IOS_LATENCY_M (1ULL << IOS_LATENCY)
#define IOS_QUEUES_M (1ULL << IOS_QUEUES)
#define IOS_L_HISTO_M (1ULL << IOS_L_HISTO)
#define IOS_RQ_HISTO_M (1ULL << IOS_RQ_HISTO)
/* Mask of all the histo bits */
#define IOS_ANYHISTO_M (IOS_L_HISTO_M | IOS_RQ_HISTO_M)
/*
* Lookup table for iostat flags to nvlist names. Basically a list
* of all the nvlists a flag requires. Also specifies the order in
* which data gets printed in zpool iostat.
*/
static const char *vsx_type_to_nvlist[IOS_COUNT][15] = {
[IOS_L_HISTO] = {
ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO,
ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO,
NULL},
[IOS_LATENCY] = {
ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO,
ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO,
NULL},
[IOS_QUEUES] = {
ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE,
NULL},
[IOS_RQ_HISTO] = {
ZPOOL_CONFIG_VDEV_SYNC_IND_R_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_AGG_R_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_IND_W_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_AGG_W_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_IND_R_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_AGG_R_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_IND_W_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_AGG_W_HISTO,
ZPOOL_CONFIG_VDEV_IND_SCRUB_HISTO,
ZPOOL_CONFIG_VDEV_AGG_SCRUB_HISTO,
ZPOOL_CONFIG_VDEV_IND_TRIM_HISTO,
ZPOOL_CONFIG_VDEV_AGG_TRIM_HISTO,
ZPOOL_CONFIG_VDEV_IND_REBUILD_HISTO,
ZPOOL_CONFIG_VDEV_AGG_REBUILD_HISTO,
NULL},
};
/*
* Given a cb->cb_flags with a histogram bit set, return the iostat_type.
* Right now, only one histo bit is ever set at one time, so we can
* just do a highbit64(a)
*/
#define IOS_HISTO_IDX(a) (highbit64(a & IOS_ANYHISTO_M) - 1)
typedef struct zpool_command {
const char *name;
int (*func)(int, char **);
zpool_help_t usage;
} zpool_command_t;
/*
* Master command table. Each ZFS command has a name, associated function, and
* usage message. The usage messages need to be internationalized, so we have
* to have a function to return the usage message based on a command index.
*
* These commands are organized according to how they are displayed in the usage
* message. An empty command (one with a NULL name) indicates an empty line in
* the generic usage message.
*/
static zpool_command_t command_table[] = {
{ "version", zpool_do_version, HELP_VERSION },
{ NULL },
{ "create", zpool_do_create, HELP_CREATE },
{ "destroy", zpool_do_destroy, HELP_DESTROY },
{ NULL },
{ "add", zpool_do_add, HELP_ADD },
{ "remove", zpool_do_remove, HELP_REMOVE },
{ NULL },
{ "labelclear", zpool_do_labelclear, HELP_LABELCLEAR },
{ NULL },
{ "checkpoint", zpool_do_checkpoint, HELP_CHECKPOINT },
{ NULL },
{ "list", zpool_do_list, HELP_LIST },
{ "iostat", zpool_do_iostat, HELP_IOSTAT },
{ "status", zpool_do_status, HELP_STATUS },
{ NULL },
{ "online", zpool_do_online, HELP_ONLINE },
{ "offline", zpool_do_offline, HELP_OFFLINE },
{ "clear", zpool_do_clear, HELP_CLEAR },
{ "reopen", zpool_do_reopen, HELP_REOPEN },
{ NULL },
{ "attach", zpool_do_attach, HELP_ATTACH },
{ "detach", zpool_do_detach, HELP_DETACH },
{ "replace", zpool_do_replace, HELP_REPLACE },
{ "split", zpool_do_split, HELP_SPLIT },
{ NULL },
{ "initialize", zpool_do_initialize, HELP_INITIALIZE },
{ "resilver", zpool_do_resilver, HELP_RESILVER },
{ "scrub", zpool_do_scrub, HELP_SCRUB },
{ "trim", zpool_do_trim, HELP_TRIM },
{ NULL },
{ "import", zpool_do_import, HELP_IMPORT },
{ "export", zpool_do_export, HELP_EXPORT },
{ "upgrade", zpool_do_upgrade, HELP_UPGRADE },
{ "reguid", zpool_do_reguid, HELP_REGUID },
{ NULL },
{ "history", zpool_do_history, HELP_HISTORY },
{ "events", zpool_do_events, HELP_EVENTS },
{ NULL },
{ "get", zpool_do_get, HELP_GET },
{ "set", zpool_do_set, HELP_SET },
{ "sync", zpool_do_sync, HELP_SYNC },
{ NULL },
{ "wait", zpool_do_wait, HELP_WAIT },
};
#define NCOMMAND (ARRAY_SIZE(command_table))
#define VDEV_ALLOC_CLASS_LOGS "logs"
static zpool_command_t *current_command;
static zfs_type_t current_prop_type = (ZFS_TYPE_POOL | ZFS_TYPE_VDEV);
static char history_str[HIS_MAX_RECORD_LEN];
static boolean_t log_history = B_TRUE;
static uint_t timestamp_fmt = NODATE;
static const char *
get_usage(zpool_help_t idx)
{
switch (idx) {
case HELP_ADD:
return (gettext("\tadd [-fgLnP] [-o property=value] "
"<pool> <vdev> ...\n"));
case HELP_ATTACH:
return (gettext("\tattach [-fsw] [-o property=value] "
"<pool> <device> <new-device>\n"));
case HELP_CLEAR:
return (gettext("\tclear [-nF] <pool> [device]\n"));
case HELP_CREATE:
return (gettext("\tcreate [-fnd] [-o property=value] ... \n"
"\t [-O file-system-property=value] ... \n"
"\t [-m mountpoint] [-R root] <pool> <vdev> ...\n"));
case HELP_CHECKPOINT:
return (gettext("\tcheckpoint [-d [-w]] <pool> ...\n"));
case HELP_DESTROY:
return (gettext("\tdestroy [-f] <pool>\n"));
case HELP_DETACH:
return (gettext("\tdetach <pool> <device>\n"));
case HELP_EXPORT:
return (gettext("\texport [-af] <pool> ...\n"));
case HELP_HISTORY:
return (gettext("\thistory [-il] [<pool>] ...\n"));
case HELP_IMPORT:
return (gettext("\timport [-d dir] [-D]\n"
"\timport [-o mntopts] [-o property=value] ... \n"
"\t [-d dir | -c cachefile] [-D] [-l] [-f] [-m] [-N] "
"[-R root] [-F [-n]] -a\n"
"\timport [-o mntopts] [-o property=value] ... \n"
"\t [-d dir | -c cachefile] [-D] [-l] [-f] [-m] [-N] "
"[-R root] [-F [-n]]\n"
"\t [--rewind-to-checkpoint] <pool | id> [newpool]\n"));
case HELP_IOSTAT:
return (gettext("\tiostat [[[-c [script1,script2,...]"
"[-lq]]|[-rw]] [-T d | u] [-ghHLpPvy]\n"
"\t [[pool ...]|[pool vdev ...]|[vdev ...]]"
" [[-n] interval [count]]\n"));
case HELP_LABELCLEAR:
return (gettext("\tlabelclear [-f] <vdev>\n"));
case HELP_LIST:
return (gettext("\tlist [-gHLpPv] [-o property[,...]] "
"[-T d|u] [pool] ... \n"
"\t [interval [count]]\n"));
case HELP_OFFLINE:
return (gettext("\toffline [-f] [-t] <pool> <device> ...\n"));
case HELP_ONLINE:
return (gettext("\tonline [-e] <pool> <device> ...\n"));
case HELP_REPLACE:
return (gettext("\treplace [-fsw] [-o property=value] "
"<pool> <device> [new-device]\n"));
case HELP_REMOVE:
return (gettext("\tremove [-npsw] <pool> <device> ...\n"));
case HELP_REOPEN:
return (gettext("\treopen [-n] <pool>\n"));
case HELP_INITIALIZE:
return (gettext("\tinitialize [-c | -s | -u] [-w] <pool> "
"[<device> ...]\n"));
case HELP_SCRUB:
return (gettext("\tscrub [-s | -p] [-w] [-e] <pool> ...\n"));
case HELP_RESILVER:
return (gettext("\tresilver <pool> ...\n"));
case HELP_TRIM:
return (gettext("\ttrim [-dw] [-r <rate>] [-c | -s] <pool> "
"[<device> ...]\n"));
case HELP_STATUS:
return (gettext("\tstatus [-c [script1,script2,...]] "
"[-igLpPstvxD] [-T d|u] [pool] ... \n"
"\t [interval [count]]\n"));
case HELP_UPGRADE:
return (gettext("\tupgrade\n"
"\tupgrade -v\n"
"\tupgrade [-V version] <-a | pool ...>\n"));
case HELP_EVENTS:
return (gettext("\tevents [-vHf [pool] | -c]\n"));
case HELP_GET:
return (gettext("\tget [-Hp] [-o \"all\" | field[,...]] "
"<\"all\" | property[,...]> <pool> ...\n"));
case HELP_SET:
return (gettext("\tset <property=value> <pool>\n"
"\tset <vdev_property=value> <pool> <vdev>\n"));
case HELP_SPLIT:
return (gettext("\tsplit [-gLnPl] [-R altroot] [-o mntopts]\n"
"\t [-o property=value] <pool> <newpool> "
"[<device> ...]\n"));
case HELP_REGUID:
return (gettext("\treguid <pool>\n"));
case HELP_SYNC:
return (gettext("\tsync [pool] ...\n"));
case HELP_VERSION:
return (gettext("\tversion\n"));
case HELP_WAIT:
return (gettext("\twait [-Hp] [-T d|u] [-t <activity>[,...]] "
"<pool> [interval]\n"));
default:
__builtin_unreachable();
}
}
static void
zpool_collect_leaves(zpool_handle_t *zhp, nvlist_t *nvroot, nvlist_t *res)
{
uint_t children = 0;
nvlist_t **child;
uint_t i;
(void) nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children);
if (children == 0) {
char *path = zpool_vdev_name(g_zfs, zhp, nvroot,
VDEV_NAME_PATH);
if (strcmp(path, VDEV_TYPE_INDIRECT) != 0 &&
strcmp(path, VDEV_TYPE_HOLE) != 0)
fnvlist_add_boolean(res, path);
free(path);
return;
}
for (i = 0; i < children; i++) {
zpool_collect_leaves(zhp, child[i], res);
}
}
/*
* Callback routine that will print out a pool property value.
*/
static int
print_pool_prop_cb(int prop, void *cb)
{
FILE *fp = cb;
(void) fprintf(fp, "\t%-19s ", zpool_prop_to_name(prop));
if (zpool_prop_readonly(prop))
(void) fprintf(fp, " NO ");
else
(void) fprintf(fp, " YES ");
if (zpool_prop_values(prop) == NULL)
(void) fprintf(fp, "-\n");
else
(void) fprintf(fp, "%s\n", zpool_prop_values(prop));
return (ZPROP_CONT);
}
/*
* Callback routine that will print out a vdev property value.
*/
static int
print_vdev_prop_cb(int prop, void *cb)
{
FILE *fp = cb;
(void) fprintf(fp, "\t%-19s ", vdev_prop_to_name(prop));
if (vdev_prop_readonly(prop))
(void) fprintf(fp, " NO ");
else
(void) fprintf(fp, " YES ");
if (vdev_prop_values(prop) == NULL)
(void) fprintf(fp, "-\n");
else
(void) fprintf(fp, "%s\n", vdev_prop_values(prop));
return (ZPROP_CONT);
}
/*
* Display usage message. If we're inside a command, display only the usage for
* that command. Otherwise, iterate over the entire command table and display
* a complete usage message.
*/
static __attribute__((noreturn)) void
usage(boolean_t requested)
{
FILE *fp = requested ? stdout : stderr;
if (current_command == NULL) {
int i;
(void) fprintf(fp, gettext("usage: zpool command args ...\n"));
(void) fprintf(fp,
gettext("where 'command' is one of the following:\n\n"));
for (i = 0; i < NCOMMAND; i++) {
if (command_table[i].name == NULL)
(void) fprintf(fp, "\n");
else
(void) fprintf(fp, "%s",
get_usage(command_table[i].usage));
}
(void) fprintf(fp,
gettext("\nFor further help on a command or topic, "
"run: %s\n"), "zpool help [<topic>]");
} else {
(void) fprintf(fp, gettext("usage:\n"));
(void) fprintf(fp, "%s", get_usage(current_command->usage));
}
if (current_command != NULL &&
current_prop_type != (ZFS_TYPE_POOL | ZFS_TYPE_VDEV) &&
((strcmp(current_command->name, "set") == 0) ||
(strcmp(current_command->name, "get") == 0) ||
(strcmp(current_command->name, "list") == 0))) {
(void) fprintf(fp, "%s",
gettext("\nthe following properties are supported:\n"));
(void) fprintf(fp, "\n\t%-19s %s %s\n\n",
"PROPERTY", "EDIT", "VALUES");
/* Iterate over all properties */
if (current_prop_type == ZFS_TYPE_POOL) {
(void) zprop_iter(print_pool_prop_cb, fp, B_FALSE,
B_TRUE, current_prop_type);
(void) fprintf(fp, "\t%-19s ", "feature@...");
(void) fprintf(fp, "YES "
"disabled | enabled | active\n");
(void) fprintf(fp, gettext("\nThe feature@ properties "
"must be appended with a feature name.\n"
"See zpool-features(7).\n"));
} else if (current_prop_type == ZFS_TYPE_VDEV) {
(void) zprop_iter(print_vdev_prop_cb, fp, B_FALSE,
B_TRUE, current_prop_type);
}
}
/*
* See comments at end of main().
*/
if (getenv("ZFS_ABORT") != NULL) {
(void) printf("dumping core by request\n");
abort();
}
exit(requested ? 0 : 2);
}
/*
* zpool initialize [-c | -s | -u] [-w] <pool> [<vdev> ...]
* Initialize all unused blocks in the specified vdevs, or all vdevs in the pool
* if none specified.
*
* -c Cancel. Ends active initializing.
* -s Suspend. Initializing can then be restarted with no flags.
* -u Uninitialize. Clears initialization state.
* -w Wait. Blocks until initializing has completed.
*/
int
zpool_do_initialize(int argc, char **argv)
{
int c;
char *poolname;
zpool_handle_t *zhp;
nvlist_t *vdevs;
int err = 0;
boolean_t wait = B_FALSE;
struct option long_options[] = {
{"cancel", no_argument, NULL, 'c'},
{"suspend", no_argument, NULL, 's'},
{"uninit", no_argument, NULL, 'u'},
{"wait", no_argument, NULL, 'w'},
{0, 0, 0, 0}
};
pool_initialize_func_t cmd_type = POOL_INITIALIZE_START;
while ((c = getopt_long(argc, argv, "csuw", long_options,
NULL)) != -1) {
switch (c) {
case 'c':
if (cmd_type != POOL_INITIALIZE_START &&
cmd_type != POOL_INITIALIZE_CANCEL) {
(void) fprintf(stderr, gettext("-c cannot be "
"combined with other options\n"));
usage(B_FALSE);
}
cmd_type = POOL_INITIALIZE_CANCEL;
break;
case 's':
if (cmd_type != POOL_INITIALIZE_START &&
cmd_type != POOL_INITIALIZE_SUSPEND) {
(void) fprintf(stderr, gettext("-s cannot be "
"combined with other options\n"));
usage(B_FALSE);
}
cmd_type = POOL_INITIALIZE_SUSPEND;
break;
case 'u':
if (cmd_type != POOL_INITIALIZE_START &&
cmd_type != POOL_INITIALIZE_UNINIT) {
(void) fprintf(stderr, gettext("-u cannot be "
"combined with other options\n"));
usage(B_FALSE);
}
cmd_type = POOL_INITIALIZE_UNINIT;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
if (optopt != 0) {
(void) fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
} else {
(void) fprintf(stderr,
gettext("invalid option '%s'\n"),
argv[optind - 1]);
}
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
return (-1);
}
if (wait && (cmd_type != POOL_INITIALIZE_START)) {
(void) fprintf(stderr, gettext("-w cannot be used with -c, -s"
"or -u\n"));
usage(B_FALSE);
}
poolname = argv[0];
zhp = zpool_open(g_zfs, poolname);
if (zhp == NULL)
return (-1);
vdevs = fnvlist_alloc();
if (argc == 1) {
/* no individual leaf vdevs specified, so add them all */
nvlist_t *config = zpool_get_config(zhp, NULL);
nvlist_t *nvroot = fnvlist_lookup_nvlist(config,
ZPOOL_CONFIG_VDEV_TREE);
zpool_collect_leaves(zhp, nvroot, vdevs);
} else {
for (int i = 1; i < argc; i++) {
fnvlist_add_boolean(vdevs, argv[i]);
}
}
if (wait)
err = zpool_initialize_wait(zhp, cmd_type, vdevs);
else
err = zpool_initialize(zhp, cmd_type, vdevs);
fnvlist_free(vdevs);
zpool_close(zhp);
return (err);
}
/*
* print a pool vdev config for dry runs
*/
static void
print_vdev_tree(zpool_handle_t *zhp, const char *name, nvlist_t *nv, int indent,
const char *match, int name_flags)
{
nvlist_t **child;
uint_t c, children;
char *vname;
boolean_t printed = B_FALSE;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0) {
if (name != NULL)
(void) printf("\t%*s%s\n", indent, "", name);
return;
}
for (c = 0; c < children; c++) {
uint64_t is_log = B_FALSE, is_hole = B_FALSE;
const char *class = "";
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE,
&is_hole);
if (is_hole == B_TRUE) {
continue;
}
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&is_log);
if (is_log)
class = VDEV_ALLOC_BIAS_LOG;
(void) nvlist_lookup_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &class);
if (strcmp(match, class) != 0)
continue;
if (!printed && name != NULL) {
(void) printf("\t%*s%s\n", indent, "", name);
printed = B_TRUE;
}
vname = zpool_vdev_name(g_zfs, zhp, child[c], name_flags);
print_vdev_tree(zhp, vname, child[c], indent + 2, "",
name_flags);
free(vname);
}
}
/*
* Print the list of l2cache devices for dry runs.
*/
static void
print_cache_list(nvlist_t *nv, int indent)
{
nvlist_t **child;
uint_t c, children;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE,
&child, &children) == 0 && children > 0) {
(void) printf("\t%*s%s\n", indent, "", "cache");
} else {
return;
}
for (c = 0; c < children; c++) {
char *vname;
vname = zpool_vdev_name(g_zfs, NULL, child[c], 0);
(void) printf("\t%*s%s\n", indent + 2, "", vname);
free(vname);
}
}
/*
* Print the list of spares for dry runs.
*/
static void
print_spare_list(nvlist_t *nv, int indent)
{
nvlist_t **child;
uint_t c, children;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES,
&child, &children) == 0 && children > 0) {
(void) printf("\t%*s%s\n", indent, "", "spares");
} else {
return;
}
for (c = 0; c < children; c++) {
char *vname;
vname = zpool_vdev_name(g_zfs, NULL, child[c], 0);
(void) printf("\t%*s%s\n", indent + 2, "", vname);
free(vname);
}
}
static boolean_t
prop_list_contains_feature(nvlist_t *proplist)
{
nvpair_t *nvp;
for (nvp = nvlist_next_nvpair(proplist, NULL); NULL != nvp;
nvp = nvlist_next_nvpair(proplist, nvp)) {
if (zpool_prop_feature(nvpair_name(nvp)))
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Add a property pair (name, string-value) into a property nvlist.
*/
static int
add_prop_list(const char *propname, const char *propval, nvlist_t **props,
boolean_t poolprop)
{
zpool_prop_t prop = ZPOOL_PROP_INVAL;
nvlist_t *proplist;
const char *normnm;
const char *strval;
if (*props == NULL &&
nvlist_alloc(props, NV_UNIQUE_NAME, 0) != 0) {
(void) fprintf(stderr,
gettext("internal error: out of memory\n"));
return (1);
}
proplist = *props;
if (poolprop) {
const char *vname = zpool_prop_to_name(ZPOOL_PROP_VERSION);
const char *cname =
zpool_prop_to_name(ZPOOL_PROP_COMPATIBILITY);
if ((prop = zpool_name_to_prop(propname)) == ZPOOL_PROP_INVAL &&
(!zpool_prop_feature(propname) &&
!zpool_prop_vdev(propname))) {
(void) fprintf(stderr, gettext("property '%s' is "
"not a valid pool or vdev property\n"), propname);
return (2);
}
/*
* feature@ properties and version should not be specified
* at the same time.
*/
if ((prop == ZPOOL_PROP_INVAL && zpool_prop_feature(propname) &&
nvlist_exists(proplist, vname)) ||
(prop == ZPOOL_PROP_VERSION &&
prop_list_contains_feature(proplist))) {
(void) fprintf(stderr, gettext("'feature@' and "
"'version' properties cannot be specified "
"together\n"));
return (2);
}
/*
* if version is specified, only "legacy" compatibility
* may be requested
*/
if ((prop == ZPOOL_PROP_COMPATIBILITY &&
strcmp(propval, ZPOOL_COMPAT_LEGACY) != 0 &&
nvlist_exists(proplist, vname)) ||
(prop == ZPOOL_PROP_VERSION &&
nvlist_exists(proplist, cname) &&
strcmp(fnvlist_lookup_string(proplist, cname),
ZPOOL_COMPAT_LEGACY) != 0)) {
(void) fprintf(stderr, gettext("when 'version' is "
"specified, the 'compatibility' feature may only "
"be set to '" ZPOOL_COMPAT_LEGACY "'\n"));
return (2);
}
if (zpool_prop_feature(propname) || zpool_prop_vdev(propname))
normnm = propname;
else
normnm = zpool_prop_to_name(prop);
} else {
zfs_prop_t fsprop = zfs_name_to_prop(propname);
if (zfs_prop_valid_for_type(fsprop, ZFS_TYPE_FILESYSTEM,
B_FALSE)) {
normnm = zfs_prop_to_name(fsprop);
} else if (zfs_prop_user(propname) ||
zfs_prop_userquota(propname)) {
normnm = propname;
} else {
(void) fprintf(stderr, gettext("property '%s' is "
"not a valid filesystem property\n"), propname);
return (2);
}
}
if (nvlist_lookup_string(proplist, normnm, &strval) == 0 &&
prop != ZPOOL_PROP_CACHEFILE) {
(void) fprintf(stderr, gettext("property '%s' "
"specified multiple times\n"), propname);
return (2);
}
if (nvlist_add_string(proplist, normnm, propval) != 0) {
(void) fprintf(stderr, gettext("internal "
"error: out of memory\n"));
return (1);
}
return (0);
}
/*
* Set a default property pair (name, string-value) in a property nvlist
*/
static int
add_prop_list_default(const char *propname, const char *propval,
nvlist_t **props)
{
const char *pval;
if (nvlist_lookup_string(*props, propname, &pval) == 0)
return (0);
return (add_prop_list(propname, propval, props, B_TRUE));
}
/*
* zpool add [-fgLnP] [-o property=value] <pool> <vdev> ...
*
* -f Force addition of devices, even if they appear in use
* -g Display guid for individual vdev name.
* -L Follow links when resolving vdev path name.
* -n Do not add the devices, but display the resulting layout if
* they were to be added.
* -o Set property=value.
* -P Display full path for vdev name.
*
* Adds the given vdevs to 'pool'. As with create, the bulk of this work is
* handled by make_root_vdev(), which constructs the nvlist needed to pass to
* libzfs.
*/
int
zpool_do_add(int argc, char **argv)
{
boolean_t force = B_FALSE;
boolean_t dryrun = B_FALSE;
int name_flags = 0;
int c;
nvlist_t *nvroot;
char *poolname;
int ret;
zpool_handle_t *zhp;
nvlist_t *config;
nvlist_t *props = NULL;
char *propval;
/* check options */
while ((c = getopt(argc, argv, "fgLno:P")) != -1) {
switch (c) {
case 'f':
force = B_TRUE;
break;
case 'g':
name_flags |= VDEV_NAME_GUID;
break;
case 'L':
name_flags |= VDEV_NAME_FOLLOW_LINKS;
break;
case 'n':
dryrun = B_TRUE;
break;
case 'o':
if ((propval = strchr(optarg, '=')) == NULL) {
(void) fprintf(stderr, gettext("missing "
"'=' for -o option\n"));
usage(B_FALSE);
}
*propval = '\0';
propval++;
if ((strcmp(optarg, ZPOOL_CONFIG_ASHIFT) != 0) ||
(add_prop_list(optarg, propval, &props, B_TRUE)))
usage(B_FALSE);
break;
case 'P':
name_flags |= VDEV_NAME_PATH;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing vdev specification\n"));
usage(B_FALSE);
}
poolname = argv[0];
argc--;
argv++;
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
if ((config = zpool_get_config(zhp, NULL)) == NULL) {
(void) fprintf(stderr, gettext("pool '%s' is unavailable\n"),
poolname);
zpool_close(zhp);
return (1);
}
/* unless manually specified use "ashift" pool property (if set) */
if (!nvlist_exists(props, ZPOOL_CONFIG_ASHIFT)) {
int intval;
zprop_source_t src;
char strval[ZPOOL_MAXPROPLEN];
intval = zpool_get_prop_int(zhp, ZPOOL_PROP_ASHIFT, &src);
if (src != ZPROP_SRC_DEFAULT) {
(void) sprintf(strval, "%" PRId32, intval);
verify(add_prop_list(ZPOOL_CONFIG_ASHIFT, strval,
&props, B_TRUE) == 0);
}
}
/* pass off to make_root_vdev for processing */
nvroot = make_root_vdev(zhp, props, force, !force, B_FALSE, dryrun,
argc, argv);
if (nvroot == NULL) {
zpool_close(zhp);
return (1);
}
if (dryrun) {
nvlist_t *poolnvroot;
nvlist_t **l2child, **sparechild;
uint_t l2children, sparechildren, c;
char *vname;
boolean_t hadcache = B_FALSE, hadspare = B_FALSE;
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&poolnvroot) == 0);
(void) printf(gettext("would update '%s' to the following "
"configuration:\n\n"), zpool_get_name(zhp));
/* print original main pool and new tree */
print_vdev_tree(zhp, poolname, poolnvroot, 0, "",
name_flags | VDEV_NAME_TYPE_ID);
print_vdev_tree(zhp, NULL, nvroot, 0, "", name_flags);
/* print other classes: 'dedup', 'special', and 'log' */
if (zfs_special_devs(poolnvroot, VDEV_ALLOC_BIAS_DEDUP)) {
print_vdev_tree(zhp, "dedup", poolnvroot, 0,
VDEV_ALLOC_BIAS_DEDUP, name_flags);
print_vdev_tree(zhp, NULL, nvroot, 0,
VDEV_ALLOC_BIAS_DEDUP, name_flags);
} else if (zfs_special_devs(nvroot, VDEV_ALLOC_BIAS_DEDUP)) {
print_vdev_tree(zhp, "dedup", nvroot, 0,
VDEV_ALLOC_BIAS_DEDUP, name_flags);
}
if (zfs_special_devs(poolnvroot, VDEV_ALLOC_BIAS_SPECIAL)) {
print_vdev_tree(zhp, "special", poolnvroot, 0,
VDEV_ALLOC_BIAS_SPECIAL, name_flags);
print_vdev_tree(zhp, NULL, nvroot, 0,
VDEV_ALLOC_BIAS_SPECIAL, name_flags);
} else if (zfs_special_devs(nvroot, VDEV_ALLOC_BIAS_SPECIAL)) {
print_vdev_tree(zhp, "special", nvroot, 0,
VDEV_ALLOC_BIAS_SPECIAL, name_flags);
}
if (num_logs(poolnvroot) > 0) {
print_vdev_tree(zhp, "logs", poolnvroot, 0,
VDEV_ALLOC_BIAS_LOG, name_flags);
print_vdev_tree(zhp, NULL, nvroot, 0,
VDEV_ALLOC_BIAS_LOG, name_flags);
} else if (num_logs(nvroot) > 0) {
print_vdev_tree(zhp, "logs", nvroot, 0,
VDEV_ALLOC_BIAS_LOG, name_flags);
}
/* Do the same for the caches */
if (nvlist_lookup_nvlist_array(poolnvroot, ZPOOL_CONFIG_L2CACHE,
&l2child, &l2children) == 0 && l2children) {
hadcache = B_TRUE;
(void) printf(gettext("\tcache\n"));
for (c = 0; c < l2children; c++) {
vname = zpool_vdev_name(g_zfs, NULL,
l2child[c], name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2child, &l2children) == 0 && l2children) {
if (!hadcache)
(void) printf(gettext("\tcache\n"));
for (c = 0; c < l2children; c++) {
vname = zpool_vdev_name(g_zfs, NULL,
l2child[c], name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
/* And finally the spares */
if (nvlist_lookup_nvlist_array(poolnvroot, ZPOOL_CONFIG_SPARES,
&sparechild, &sparechildren) == 0 && sparechildren > 0) {
hadspare = B_TRUE;
(void) printf(gettext("\tspares\n"));
for (c = 0; c < sparechildren; c++) {
vname = zpool_vdev_name(g_zfs, NULL,
sparechild[c], name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&sparechild, &sparechildren) == 0 && sparechildren > 0) {
if (!hadspare)
(void) printf(gettext("\tspares\n"));
for (c = 0; c < sparechildren; c++) {
vname = zpool_vdev_name(g_zfs, NULL,
sparechild[c], name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
ret = 0;
} else {
ret = (zpool_add(zhp, nvroot) != 0);
}
nvlist_free(props);
nvlist_free(nvroot);
zpool_close(zhp);
return (ret);
}
/*
* zpool remove [-npsw] <pool> <vdev> ...
*
* Removes the given vdev from the pool.
*/
int
zpool_do_remove(int argc, char **argv)
{
char *poolname;
int i, ret = 0;
zpool_handle_t *zhp = NULL;
boolean_t stop = B_FALSE;
int c;
boolean_t noop = B_FALSE;
boolean_t parsable = B_FALSE;
boolean_t wait = B_FALSE;
/* check options */
while ((c = getopt(argc, argv, "npsw")) != -1) {
switch (c) {
case 'n':
noop = B_TRUE;
break;
case 'p':
parsable = B_TRUE;
break;
case 's':
stop = B_TRUE;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
poolname = argv[0];
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
if (stop && noop) {
zpool_close(zhp);
(void) fprintf(stderr, gettext("stop request ignored\n"));
return (0);
}
if (stop) {
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if (zpool_vdev_remove_cancel(zhp) != 0)
ret = 1;
if (wait) {
(void) fprintf(stderr, gettext("invalid option "
"combination: -w cannot be used with -s\n"));
usage(B_FALSE);
}
} else {
if (argc < 2) {
(void) fprintf(stderr, gettext("missing device\n"));
usage(B_FALSE);
}
for (i = 1; i < argc; i++) {
if (noop) {
uint64_t size;
if (zpool_vdev_indirect_size(zhp, argv[i],
&size) != 0) {
ret = 1;
break;
}
if (parsable) {
(void) printf("%s %llu\n",
argv[i], (unsigned long long)size);
} else {
char valstr[32];
zfs_nicenum(size, valstr,
sizeof (valstr));
(void) printf("Memory that will be "
"used after removing %s: %s\n",
argv[i], valstr);
}
} else {
if (zpool_vdev_remove(zhp, argv[i]) != 0)
ret = 1;
}
}
if (ret == 0 && wait)
ret = zpool_wait(zhp, ZPOOL_WAIT_REMOVE);
}
zpool_close(zhp);
return (ret);
}
/*
* Return 1 if a vdev is active (being used in a pool)
* Return 0 if a vdev is inactive (offlined or faulted, or not in active pool)
*
* This is useful for checking if a disk in an active pool is offlined or
* faulted.
*/
static int
vdev_is_active(char *vdev_path)
{
int fd;
fd = open(vdev_path, O_EXCL);
if (fd < 0) {
return (1); /* cant open O_EXCL - disk is active */
}
close(fd);
return (0); /* disk is inactive in the pool */
}
/*
* zpool labelclear [-f] <vdev>
*
* -f Force clearing the label for the vdevs which are members of
* the exported or foreign pools.
*
* Verifies that the vdev is not active and zeros out the label information
* on the device.
*/
int
zpool_do_labelclear(int argc, char **argv)
{
char vdev[MAXPATHLEN];
char *name = NULL;
int c, fd = -1, ret = 0;
nvlist_t *config;
pool_state_t state;
boolean_t inuse = B_FALSE;
boolean_t force = B_FALSE;
/* check options */
while ((c = getopt(argc, argv, "f")) != -1) {
switch (c) {
case 'f':
force = B_TRUE;
break;
default:
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get vdev name */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing vdev name\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
(void) strlcpy(vdev, argv[0], sizeof (vdev));
/*
* If we cannot open an absolute path, we quit.
* Otherwise if the provided vdev name doesn't point to a file,
* try prepending expected disk paths and partition numbers.
*/
if ((fd = open(vdev, O_RDWR)) < 0) {
int error;
if (vdev[0] == '/') {
(void) fprintf(stderr, gettext("failed to open "
"%s: %s\n"), vdev, strerror(errno));
return (1);
}
error = zfs_resolve_shortname(argv[0], vdev, MAXPATHLEN);
if (error == 0 && zfs_dev_is_whole_disk(vdev)) {
if (zfs_append_partition(vdev, MAXPATHLEN) == -1)
error = ENOENT;
}
if (error || ((fd = open(vdev, O_RDWR)) < 0)) {
if (errno == ENOENT) {
(void) fprintf(stderr, gettext(
"failed to find device %s, try "
"specifying absolute path instead\n"),
argv[0]);
return (1);
}
(void) fprintf(stderr, gettext("failed to open %s:"
" %s\n"), vdev, strerror(errno));
return (1);
}
}
/*
* Flush all dirty pages for the block device. This should not be
* fatal when the device does not support BLKFLSBUF as would be the
* case for a file vdev.
*/
if ((zfs_dev_flush(fd) != 0) && (errno != ENOTTY))
(void) fprintf(stderr, gettext("failed to invalidate "
"cache for %s: %s\n"), vdev, strerror(errno));
if (zpool_read_label(fd, &config, NULL) != 0) {
(void) fprintf(stderr,
gettext("failed to read label from %s\n"), vdev);
ret = 1;
goto errout;
}
nvlist_free(config);
ret = zpool_in_use(g_zfs, fd, &state, &name, &inuse);
if (ret != 0) {
(void) fprintf(stderr,
gettext("failed to check state for %s\n"), vdev);
ret = 1;
goto errout;
}
if (!inuse)
goto wipe_label;
switch (state) {
default:
case POOL_STATE_ACTIVE:
case POOL_STATE_SPARE:
case POOL_STATE_L2CACHE:
/*
* We allow the user to call 'zpool offline -f'
* on an offlined disk in an active pool. We can check if
* the disk is online by calling vdev_is_active().
*/
if (force && !vdev_is_active(vdev))
break;
(void) fprintf(stderr, gettext(
"%s is a member (%s) of pool \"%s\""),
vdev, zpool_pool_state_to_name(state), name);
if (force) {
(void) fprintf(stderr, gettext(
". Offline the disk first to clear its label."));
}
printf("\n");
ret = 1;
goto errout;
case POOL_STATE_EXPORTED:
if (force)
break;
(void) fprintf(stderr, gettext(
"use '-f' to override the following error:\n"
"%s is a member of exported pool \"%s\"\n"),
vdev, name);
ret = 1;
goto errout;
case POOL_STATE_POTENTIALLY_ACTIVE:
if (force)
break;
(void) fprintf(stderr, gettext(
"use '-f' to override the following error:\n"
"%s is a member of potentially active pool \"%s\"\n"),
vdev, name);
ret = 1;
goto errout;
case POOL_STATE_DESTROYED:
/* inuse should never be set for a destroyed pool */
assert(0);
break;
}
wipe_label:
ret = zpool_clear_label(fd);
if (ret != 0) {
(void) fprintf(stderr,
gettext("failed to clear label for %s\n"), vdev);
}
errout:
free(name);
(void) close(fd);
return (ret);
}
/*
* zpool create [-fnd] [-o property=value] ...
* [-O file-system-property=value] ...
* [-R root] [-m mountpoint] <pool> <dev> ...
*
* -f Force creation, even if devices appear in use
* -n Do not create the pool, but display the resulting layout if it
* were to be created.
* -R Create a pool under an alternate root
* -m Set default mountpoint for the root dataset. By default it's
* '/<pool>'
* -o Set property=value.
* -o Set feature@feature=enabled|disabled.
* -d Don't automatically enable all supported pool features
* (individual features can be enabled with -o).
* -O Set fsproperty=value in the pool's root file system
*
* Creates the named pool according to the given vdev specification. The
* bulk of the vdev processing is done in make_root_vdev() in zpool_vdev.c.
* Once we get the nvlist back from make_root_vdev(), we either print out the
* contents (if '-n' was specified), or pass it to libzfs to do the creation.
*/
int
zpool_do_create(int argc, char **argv)
{
boolean_t force = B_FALSE;
boolean_t dryrun = B_FALSE;
boolean_t enable_pool_features = B_TRUE;
int c;
nvlist_t *nvroot = NULL;
char *poolname;
char *tname = NULL;
int ret = 1;
char *altroot = NULL;
char *compat = NULL;
char *mountpoint = NULL;
nvlist_t *fsprops = NULL;
nvlist_t *props = NULL;
char *propval;
/* check options */
while ((c = getopt(argc, argv, ":fndR:m:o:O:t:")) != -1) {
switch (c) {
case 'f':
force = B_TRUE;
break;
case 'n':
dryrun = B_TRUE;
break;
case 'd':
enable_pool_features = B_FALSE;
break;
case 'R':
altroot = optarg;
if (add_prop_list(zpool_prop_to_name(
ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE))
goto errout;
if (add_prop_list_default(zpool_prop_to_name(
ZPOOL_PROP_CACHEFILE), "none", &props))
goto errout;
break;
case 'm':
/* Equivalent to -O mountpoint=optarg */
mountpoint = optarg;
break;
case 'o':
if ((propval = strchr(optarg, '=')) == NULL) {
(void) fprintf(stderr, gettext("missing "
"'=' for -o option\n"));
goto errout;
}
*propval = '\0';
propval++;
if (add_prop_list(optarg, propval, &props, B_TRUE))
goto errout;
/*
* If the user is creating a pool that doesn't support
* feature flags, don't enable any features.
*/
if (zpool_name_to_prop(optarg) == ZPOOL_PROP_VERSION) {
char *end;
u_longlong_t ver;
ver = strtoull(propval, &end, 10);
if (*end == '\0' &&
ver < SPA_VERSION_FEATURES) {
enable_pool_features = B_FALSE;
}
}
if (zpool_name_to_prop(optarg) == ZPOOL_PROP_ALTROOT)
altroot = propval;
if (zpool_name_to_prop(optarg) ==
ZPOOL_PROP_COMPATIBILITY)
compat = propval;
break;
case 'O':
if ((propval = strchr(optarg, '=')) == NULL) {
(void) fprintf(stderr, gettext("missing "
"'=' for -O option\n"));
goto errout;
}
*propval = '\0';
propval++;
/*
* Mountpoints are checked and then added later.
* Uniquely among properties, they can be specified
* more than once, to avoid conflict with -m.
*/
if (0 == strcmp(optarg,
zfs_prop_to_name(ZFS_PROP_MOUNTPOINT))) {
mountpoint = propval;
} else if (add_prop_list(optarg, propval, &fsprops,
B_FALSE)) {
goto errout;
}
break;
case 't':
/*
* Sanity check temporary pool name.
*/
if (strchr(optarg, '/') != NULL) {
(void) fprintf(stderr, gettext("cannot create "
"'%s': invalid character '/' in temporary "
"name\n"), optarg);
(void) fprintf(stderr, gettext("use 'zfs "
"create' to create a dataset\n"));
goto errout;
}
if (add_prop_list(zpool_prop_to_name(
ZPOOL_PROP_TNAME), optarg, &props, B_TRUE))
goto errout;
if (add_prop_list_default(zpool_prop_to_name(
ZPOOL_PROP_CACHEFILE), "none", &props))
goto errout;
tname = optarg;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
goto badusage;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
goto badusage;
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
goto badusage;
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing vdev specification\n"));
goto badusage;
}
poolname = argv[0];
/*
* As a special case, check for use of '/' in the name, and direct the
* user to use 'zfs create' instead.
*/
if (strchr(poolname, '/') != NULL) {
(void) fprintf(stderr, gettext("cannot create '%s': invalid "
"character '/' in pool name\n"), poolname);
(void) fprintf(stderr, gettext("use 'zfs create' to "
"create a dataset\n"));
goto errout;
}
/* pass off to make_root_vdev for bulk processing */
nvroot = make_root_vdev(NULL, props, force, !force, B_FALSE, dryrun,
argc - 1, argv + 1);
if (nvroot == NULL)
goto errout;
/* make_root_vdev() allows 0 toplevel children if there are spares */
if (!zfs_allocatable_devs(nvroot)) {
(void) fprintf(stderr, gettext("invalid vdev "
"specification: at least one toplevel vdev must be "
"specified\n"));
goto errout;
}
if (altroot != NULL && altroot[0] != '/') {
(void) fprintf(stderr, gettext("invalid alternate root '%s': "
"must be an absolute path\n"), altroot);
goto errout;
}
/*
* Check the validity of the mountpoint and direct the user to use the
* '-m' mountpoint option if it looks like its in use.
*/
if (mountpoint == NULL ||
(strcmp(mountpoint, ZFS_MOUNTPOINT_LEGACY) != 0 &&
strcmp(mountpoint, ZFS_MOUNTPOINT_NONE) != 0)) {
char buf[MAXPATHLEN];
DIR *dirp;
if (mountpoint && mountpoint[0] != '/') {
(void) fprintf(stderr, gettext("invalid mountpoint "
"'%s': must be an absolute path, 'legacy', or "
"'none'\n"), mountpoint);
goto errout;
}
if (mountpoint == NULL) {
if (altroot != NULL)
(void) snprintf(buf, sizeof (buf), "%s/%s",
altroot, poolname);
else
(void) snprintf(buf, sizeof (buf), "/%s",
poolname);
} else {
if (altroot != NULL)
(void) snprintf(buf, sizeof (buf), "%s%s",
altroot, mountpoint);
else
(void) snprintf(buf, sizeof (buf), "%s",
mountpoint);
}
if ((dirp = opendir(buf)) == NULL && errno != ENOENT) {
(void) fprintf(stderr, gettext("mountpoint '%s' : "
"%s\n"), buf, strerror(errno));
(void) fprintf(stderr, gettext("use '-m' "
"option to provide a different default\n"));
goto errout;
} else if (dirp) {
int count = 0;
while (count < 3 && readdir(dirp) != NULL)
count++;
(void) closedir(dirp);
if (count > 2) {
(void) fprintf(stderr, gettext("mountpoint "
"'%s' exists and is not empty\n"), buf);
(void) fprintf(stderr, gettext("use '-m' "
"option to provide a "
"different default\n"));
goto errout;
}
}
}
/*
* Now that the mountpoint's validity has been checked, ensure that
* the property is set appropriately prior to creating the pool.
*/
if (mountpoint != NULL) {
ret = add_prop_list(zfs_prop_to_name(ZFS_PROP_MOUNTPOINT),
mountpoint, &fsprops, B_FALSE);
if (ret != 0)
goto errout;
}
ret = 1;
if (dryrun) {
/*
* For a dry run invocation, print out a basic message and run
* through all the vdevs in the list and print out in an
* appropriate hierarchy.
*/
(void) printf(gettext("would create '%s' with the "
"following layout:\n\n"), poolname);
print_vdev_tree(NULL, poolname, nvroot, 0, "", 0);
print_vdev_tree(NULL, "dedup", nvroot, 0,
VDEV_ALLOC_BIAS_DEDUP, 0);
print_vdev_tree(NULL, "special", nvroot, 0,
VDEV_ALLOC_BIAS_SPECIAL, 0);
print_vdev_tree(NULL, "logs", nvroot, 0,
VDEV_ALLOC_BIAS_LOG, 0);
print_cache_list(nvroot, 0);
print_spare_list(nvroot, 0);
ret = 0;
} else {
/*
* Load in feature set.
* Note: if compatibility property not given, we'll have
* NULL, which means 'all features'.
*/
boolean_t requested_features[SPA_FEATURES];
if (zpool_do_load_compat(compat, requested_features) !=
ZPOOL_COMPATIBILITY_OK)
goto errout;
/*
* props contains list of features to enable.
* For each feature:
* - remove it if feature@name=disabled
* - leave it there if feature@name=enabled
* - add it if:
* - enable_pool_features (ie: no '-d' or '-o version')
* - it's supported by the kernel module
* - it's in the requested feature set
* - warn if it's enabled but not in compat
*/
for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
char propname[MAXPATHLEN];
const char *propval;
zfeature_info_t *feat = &spa_feature_table[i];
(void) snprintf(propname, sizeof (propname),
"feature@%s", feat->fi_uname);
if (!nvlist_lookup_string(props, propname, &propval)) {
if (strcmp(propval,
ZFS_FEATURE_DISABLED) == 0) {
(void) nvlist_remove_all(props,
propname);
} else if (strcmp(propval,
ZFS_FEATURE_ENABLED) == 0 &&
!requested_features[i]) {
(void) fprintf(stderr, gettext(
"Warning: feature \"%s\" enabled "
"but is not in specified "
"'compatibility' feature set.\n"),
feat->fi_uname);
}
} else if (
enable_pool_features &&
feat->fi_zfs_mod_supported &&
requested_features[i]) {
ret = add_prop_list(propname,
ZFS_FEATURE_ENABLED, &props, B_TRUE);
if (ret != 0)
goto errout;
}
}
ret = 1;
if (zpool_create(g_zfs, poolname,
nvroot, props, fsprops) == 0) {
zfs_handle_t *pool = zfs_open(g_zfs,
tname ? tname : poolname, ZFS_TYPE_FILESYSTEM);
if (pool != NULL) {
if (zfs_mount(pool, NULL, 0) == 0) {
ret = zfs_share(pool, NULL);
zfs_commit_shares(NULL);
}
zfs_close(pool);
}
} else if (libzfs_errno(g_zfs) == EZFS_INVALIDNAME) {
(void) fprintf(stderr, gettext("pool name may have "
"been omitted\n"));
}
}
errout:
nvlist_free(nvroot);
nvlist_free(fsprops);
nvlist_free(props);
return (ret);
badusage:
nvlist_free(fsprops);
nvlist_free(props);
usage(B_FALSE);
return (2);
}
/*
* zpool destroy <pool>
*
* -f Forcefully unmount any datasets
*
* Destroy the given pool. Automatically unmounts any datasets in the pool.
*/
int
zpool_do_destroy(int argc, char **argv)
{
boolean_t force = B_FALSE;
int c;
char *pool;
zpool_handle_t *zhp;
int ret;
/* check options */
while ((c = getopt(argc, argv, "f")) != -1) {
switch (c) {
case 'f':
force = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* check arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
pool = argv[0];
if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL) {
/*
* As a special case, check for use of '/' in the name, and
* direct the user to use 'zfs destroy' instead.
*/
if (strchr(pool, '/') != NULL)
(void) fprintf(stderr, gettext("use 'zfs destroy' to "
"destroy a dataset\n"));
return (1);
}
if (zpool_disable_datasets(zhp, force) != 0) {
(void) fprintf(stderr, gettext("could not destroy '%s': "
"could not unmount datasets\n"), zpool_get_name(zhp));
zpool_close(zhp);
return (1);
}
/* The history must be logged as part of the export */
log_history = B_FALSE;
ret = (zpool_destroy(zhp, history_str) != 0);
zpool_close(zhp);
return (ret);
}
typedef struct export_cbdata {
boolean_t force;
boolean_t hardforce;
} export_cbdata_t;
/*
* Export one pool
*/
static int
zpool_export_one(zpool_handle_t *zhp, void *data)
{
export_cbdata_t *cb = data;
if (zpool_disable_datasets(zhp, cb->force) != 0)
return (1);
/* The history must be logged as part of the export */
log_history = B_FALSE;
if (cb->hardforce) {
if (zpool_export_force(zhp, history_str) != 0)
return (1);
} else if (zpool_export(zhp, cb->force, history_str) != 0) {
return (1);
}
return (0);
}
/*
* zpool export [-f] <pool> ...
*
* -a Export all pools
* -f Forcefully unmount datasets
*
* Export the given pools. By default, the command will attempt to cleanly
* unmount any active datasets within the pool. If the '-f' flag is specified,
* then the datasets will be forcefully unmounted.
*/
int
zpool_do_export(int argc, char **argv)
{
export_cbdata_t cb;
boolean_t do_all = B_FALSE;
boolean_t force = B_FALSE;
boolean_t hardforce = B_FALSE;
int c, ret;
/* check options */
while ((c = getopt(argc, argv, "afF")) != -1) {
switch (c) {
case 'a':
do_all = B_TRUE;
break;
case 'f':
force = B_TRUE;
break;
case 'F':
hardforce = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
cb.force = force;
cb.hardforce = hardforce;
argc -= optind;
argv += optind;
if (do_all) {
if (argc != 0) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
return (for_each_pool(argc, argv, B_TRUE, NULL,
ZFS_TYPE_POOL, B_FALSE, zpool_export_one, &cb));
}
/* check arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool argument\n"));
usage(B_FALSE);
}
ret = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, zpool_export_one, &cb);
return (ret);
}
/*
* Given a vdev configuration, determine the maximum width needed for the device
* name column.
*/
static int
max_width(zpool_handle_t *zhp, nvlist_t *nv, int depth, int max,
int name_flags)
{
static const char *const subtypes[] =
{ZPOOL_CONFIG_SPARES, ZPOOL_CONFIG_L2CACHE, ZPOOL_CONFIG_CHILDREN};
char *name = zpool_vdev_name(g_zfs, zhp, nv, name_flags);
max = MAX(strlen(name) + depth, max);
free(name);
nvlist_t **child;
uint_t children;
for (size_t i = 0; i < ARRAY_SIZE(subtypes); ++i)
if (nvlist_lookup_nvlist_array(nv, subtypes[i],
&child, &children) == 0)
for (uint_t c = 0; c < children; ++c)
max = MAX(max_width(zhp, child[c], depth + 2,
max, name_flags), max);
return (max);
}
typedef struct spare_cbdata {
uint64_t cb_guid;
zpool_handle_t *cb_zhp;
} spare_cbdata_t;
static boolean_t
find_vdev(nvlist_t *nv, uint64_t search)
{
uint64_t guid;
nvlist_t **child;
uint_t c, children;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0 &&
search == guid)
return (B_TRUE);
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) == 0) {
for (c = 0; c < children; c++)
if (find_vdev(child[c], search))
return (B_TRUE);
}
return (B_FALSE);
}
static int
find_spare(zpool_handle_t *zhp, void *data)
{
spare_cbdata_t *cbp = data;
nvlist_t *config, *nvroot;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
if (find_vdev(nvroot, cbp->cb_guid)) {
cbp->cb_zhp = zhp;
return (1);
}
zpool_close(zhp);
return (0);
}
typedef struct status_cbdata {
int cb_count;
int cb_name_flags;
int cb_namewidth;
boolean_t cb_allpools;
boolean_t cb_verbose;
boolean_t cb_literal;
boolean_t cb_explain;
boolean_t cb_first;
boolean_t cb_dedup_stats;
boolean_t cb_print_status;
boolean_t cb_print_slow_ios;
boolean_t cb_print_vdev_init;
boolean_t cb_print_vdev_trim;
vdev_cmd_data_list_t *vcdl;
} status_cbdata_t;
/* Return 1 if string is NULL, empty, or whitespace; return 0 otherwise. */
static boolean_t
is_blank_str(const char *str)
{
for (; str != NULL && *str != '\0'; ++str)
if (!isblank(*str))
return (B_FALSE);
return (B_TRUE);
}
/* Print command output lines for specific vdev in a specific pool */
static void
zpool_print_cmd(vdev_cmd_data_list_t *vcdl, const char *pool, const char *path)
{
vdev_cmd_data_t *data;
int i, j;
const char *val;
for (i = 0; i < vcdl->count; i++) {
if ((strcmp(vcdl->data[i].path, path) != 0) ||
(strcmp(vcdl->data[i].pool, pool) != 0)) {
/* Not the vdev we're looking for */
continue;
}
data = &vcdl->data[i];
/* Print out all the output values for this vdev */
for (j = 0; j < vcdl->uniq_cols_cnt; j++) {
val = NULL;
/* Does this vdev have values for this column? */
for (int k = 0; k < data->cols_cnt; k++) {
if (strcmp(data->cols[k],
vcdl->uniq_cols[j]) == 0) {
/* yes it does, record the value */
val = data->lines[k];
break;
}
}
/*
* Mark empty values with dashes to make output
* awk-able.
*/
if (val == NULL || is_blank_str(val))
val = "-";
printf("%*s", vcdl->uniq_cols_width[j], val);
if (j < vcdl->uniq_cols_cnt - 1)
fputs(" ", stdout);
}
/* Print out any values that aren't in a column at the end */
for (j = data->cols_cnt; j < data->lines_cnt; j++) {
/* Did we have any columns? If so print a spacer. */
if (vcdl->uniq_cols_cnt > 0)
fputs(" ", stdout);
val = data->lines[j];
fputs(val ?: "", stdout);
}
break;
}
}
/*
* Print vdev initialization status for leaves
*/
static void
print_status_initialize(vdev_stat_t *vs, boolean_t verbose)
{
if (verbose) {
if ((vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE ||
vs->vs_initialize_state == VDEV_INITIALIZE_SUSPENDED ||
vs->vs_initialize_state == VDEV_INITIALIZE_COMPLETE) &&
!vs->vs_scan_removing) {
char zbuf[1024];
char tbuf[256];
struct tm zaction_ts;
time_t t = vs->vs_initialize_action_time;
int initialize_pct = 100;
if (vs->vs_initialize_state !=
VDEV_INITIALIZE_COMPLETE) {
initialize_pct = (vs->vs_initialize_bytes_done *
100 / (vs->vs_initialize_bytes_est + 1));
}
(void) localtime_r(&t, &zaction_ts);
(void) strftime(tbuf, sizeof (tbuf), "%c", &zaction_ts);
switch (vs->vs_initialize_state) {
case VDEV_INITIALIZE_SUSPENDED:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("suspended, started at"), tbuf);
break;
case VDEV_INITIALIZE_ACTIVE:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("started at"), tbuf);
break;
case VDEV_INITIALIZE_COMPLETE:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("completed at"), tbuf);
break;
}
(void) printf(gettext(" (%d%% initialized%s)"),
initialize_pct, zbuf);
} else {
(void) printf(gettext(" (uninitialized)"));
}
} else if (vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE) {
(void) printf(gettext(" (initializing)"));
}
}
/*
* Print vdev TRIM status for leaves
*/
static void
print_status_trim(vdev_stat_t *vs, boolean_t verbose)
{
if (verbose) {
if ((vs->vs_trim_state == VDEV_TRIM_ACTIVE ||
vs->vs_trim_state == VDEV_TRIM_SUSPENDED ||
vs->vs_trim_state == VDEV_TRIM_COMPLETE) &&
!vs->vs_scan_removing) {
char zbuf[1024];
char tbuf[256];
struct tm zaction_ts;
time_t t = vs->vs_trim_action_time;
int trim_pct = 100;
if (vs->vs_trim_state != VDEV_TRIM_COMPLETE) {
trim_pct = (vs->vs_trim_bytes_done *
100 / (vs->vs_trim_bytes_est + 1));
}
(void) localtime_r(&t, &zaction_ts);
(void) strftime(tbuf, sizeof (tbuf), "%c", &zaction_ts);
switch (vs->vs_trim_state) {
case VDEV_TRIM_SUSPENDED:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("suspended, started at"), tbuf);
break;
case VDEV_TRIM_ACTIVE:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("started at"), tbuf);
break;
case VDEV_TRIM_COMPLETE:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("completed at"), tbuf);
break;
}
(void) printf(gettext(" (%d%% trimmed%s)"),
trim_pct, zbuf);
} else if (vs->vs_trim_notsup) {
(void) printf(gettext(" (trim unsupported)"));
} else {
(void) printf(gettext(" (untrimmed)"));
}
} else if (vs->vs_trim_state == VDEV_TRIM_ACTIVE) {
(void) printf(gettext(" (trimming)"));
}
}
/*
* Return the color associated with a health string. This includes returning
* NULL for no color change.
*/
static const char *
health_str_to_color(const char *health)
{
if (strcmp(health, gettext("FAULTED")) == 0 ||
strcmp(health, gettext("SUSPENDED")) == 0 ||
strcmp(health, gettext("UNAVAIL")) == 0) {
return (ANSI_RED);
}
if (strcmp(health, gettext("OFFLINE")) == 0 ||
strcmp(health, gettext("DEGRADED")) == 0 ||
strcmp(health, gettext("REMOVED")) == 0) {
return (ANSI_YELLOW);
}
return (NULL);
}
/*
* Print out configuration state as requested by status_callback.
*/
static void
print_status_config(zpool_handle_t *zhp, status_cbdata_t *cb, const char *name,
nvlist_t *nv, int depth, boolean_t isspare, vdev_rebuild_stat_t *vrs)
{
nvlist_t **child, *root;
uint_t c, i, vsc, children;
pool_scan_stat_t *ps = NULL;
vdev_stat_t *vs;
char rbuf[6], wbuf[6], cbuf[6];
char *vname;
uint64_t notpresent;
spare_cbdata_t spare_cb;
const char *state;
const char *type;
const char *path = NULL;
const char *rcolor = NULL, *wcolor = NULL, *ccolor = NULL;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &vsc) == 0);
verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0);
if (strcmp(type, VDEV_TYPE_INDIRECT) == 0)
return;
state = zpool_state_to_name(vs->vs_state, vs->vs_aux);
if (isspare) {
/*
* For hot spares, we use the terms 'INUSE' and 'AVAILABLE' for
* online drives.
*/
if (vs->vs_aux == VDEV_AUX_SPARED)
state = gettext("INUSE");
else if (vs->vs_state == VDEV_STATE_HEALTHY)
state = gettext("AVAIL");
}
printf_color(health_str_to_color(state),
"\t%*s%-*s %-8s", depth, "", cb->cb_namewidth - depth,
name, state);
if (!isspare) {
if (vs->vs_read_errors)
rcolor = ANSI_RED;
if (vs->vs_write_errors)
wcolor = ANSI_RED;
if (vs->vs_checksum_errors)
ccolor = ANSI_RED;
if (cb->cb_literal) {
fputc(' ', stdout);
printf_color(rcolor, "%5llu",
(u_longlong_t)vs->vs_read_errors);
fputc(' ', stdout);
printf_color(wcolor, "%5llu",
(u_longlong_t)vs->vs_write_errors);
fputc(' ', stdout);
printf_color(ccolor, "%5llu",
(u_longlong_t)vs->vs_checksum_errors);
} else {
zfs_nicenum(vs->vs_read_errors, rbuf, sizeof (rbuf));
zfs_nicenum(vs->vs_write_errors, wbuf, sizeof (wbuf));
zfs_nicenum(vs->vs_checksum_errors, cbuf,
sizeof (cbuf));
fputc(' ', stdout);
printf_color(rcolor, "%5s", rbuf);
fputc(' ', stdout);
printf_color(wcolor, "%5s", wbuf);
fputc(' ', stdout);
printf_color(ccolor, "%5s", cbuf);
}
if (cb->cb_print_slow_ios) {
if (children == 0) {
/* Only leafs vdevs have slow IOs */
zfs_nicenum(vs->vs_slow_ios, rbuf,
sizeof (rbuf));
} else {
snprintf(rbuf, sizeof (rbuf), "-");
}
if (cb->cb_literal)
printf(" %5llu", (u_longlong_t)vs->vs_slow_ios);
else
printf(" %5s", rbuf);
}
}
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT,
&notpresent) == 0) {
verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0);
(void) printf(" %s %s", gettext("was"), path);
} else if (vs->vs_aux != 0) {
(void) printf(" ");
color_start(ANSI_RED);
switch (vs->vs_aux) {
case VDEV_AUX_OPEN_FAILED:
(void) printf(gettext("cannot open"));
break;
case VDEV_AUX_BAD_GUID_SUM:
(void) printf(gettext("missing device"));
break;
case VDEV_AUX_NO_REPLICAS:
(void) printf(gettext("insufficient replicas"));
break;
case VDEV_AUX_VERSION_NEWER:
(void) printf(gettext("newer version"));
break;
case VDEV_AUX_UNSUP_FEAT:
(void) printf(gettext("unsupported feature(s)"));
break;
case VDEV_AUX_ASHIFT_TOO_BIG:
(void) printf(gettext("unsupported minimum blocksize"));
break;
case VDEV_AUX_SPARED:
verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID,
&spare_cb.cb_guid) == 0);
if (zpool_iter(g_zfs, find_spare, &spare_cb) == 1) {
if (strcmp(zpool_get_name(spare_cb.cb_zhp),
zpool_get_name(zhp)) == 0)
(void) printf(gettext("currently in "
"use"));
else
(void) printf(gettext("in use by "
"pool '%s'"),
zpool_get_name(spare_cb.cb_zhp));
zpool_close(spare_cb.cb_zhp);
} else {
(void) printf(gettext("currently in use"));
}
break;
case VDEV_AUX_ERR_EXCEEDED:
(void) printf(gettext("too many errors"));
break;
case VDEV_AUX_IO_FAILURE:
(void) printf(gettext("experienced I/O failures"));
break;
case VDEV_AUX_BAD_LOG:
(void) printf(gettext("bad intent log"));
break;
case VDEV_AUX_EXTERNAL:
(void) printf(gettext("external device fault"));
break;
case VDEV_AUX_SPLIT_POOL:
(void) printf(gettext("split into new pool"));
break;
case VDEV_AUX_ACTIVE:
(void) printf(gettext("currently in use"));
break;
case VDEV_AUX_CHILDREN_OFFLINE:
(void) printf(gettext("all children offline"));
break;
case VDEV_AUX_BAD_LABEL:
(void) printf(gettext("invalid label"));
break;
default:
(void) printf(gettext("corrupted data"));
break;
}
color_end();
} else if (children == 0 && !isspare &&
getenv("ZPOOL_STATUS_NON_NATIVE_ASHIFT_IGNORE") == NULL &&
VDEV_STAT_VALID(vs_physical_ashift, vsc) &&
vs->vs_configured_ashift < vs->vs_physical_ashift) {
(void) printf(
gettext(" block size: %dB configured, %dB native"),
1 << vs->vs_configured_ashift, 1 << vs->vs_physical_ashift);
}
if (vs->vs_scan_removing != 0) {
(void) printf(gettext(" (removing)"));
} else if (VDEV_STAT_VALID(vs_noalloc, vsc) && vs->vs_noalloc != 0) {
(void) printf(gettext(" (non-allocating)"));
}
/* The root vdev has the scrub/resilver stats */
root = fnvlist_lookup_nvlist(zpool_get_config(zhp, NULL),
ZPOOL_CONFIG_VDEV_TREE);
(void) nvlist_lookup_uint64_array(root, ZPOOL_CONFIG_SCAN_STATS,
(uint64_t **)&ps, &c);
/*
* If you force fault a drive that's resilvering, its scan stats can
* get frozen in time, giving the false impression that it's
* being resilvered. That's why we check the state to see if the vdev
* is healthy before reporting "resilvering" or "repairing".
*/
if (ps != NULL && ps->pss_state == DSS_SCANNING && children == 0 &&
vs->vs_state == VDEV_STATE_HEALTHY) {
if (vs->vs_scan_processed != 0) {
(void) printf(gettext(" (%s)"),
(ps->pss_func == POOL_SCAN_RESILVER) ?
"resilvering" : "repairing");
} else if (vs->vs_resilver_deferred) {
(void) printf(gettext(" (awaiting resilver)"));
}
}
/* The top-level vdevs have the rebuild stats */
if (vrs != NULL && vrs->vrs_state == VDEV_REBUILD_ACTIVE &&
children == 0 && vs->vs_state == VDEV_STATE_HEALTHY) {
if (vs->vs_rebuild_processed != 0) {
(void) printf(gettext(" (resilvering)"));
}
}
if (cb->vcdl != NULL) {
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0) {
printf(" ");
zpool_print_cmd(cb->vcdl, zpool_get_name(zhp), path);
}
}
/* Display vdev initialization and trim status for leaves. */
if (children == 0) {
print_status_initialize(vs, cb->cb_print_vdev_init);
print_status_trim(vs, cb->cb_print_vdev_trim);
}
(void) printf("\n");
for (c = 0; c < children; c++) {
uint64_t islog = B_FALSE, ishole = B_FALSE;
/* Don't print logs or holes here */
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&islog);
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE,
&ishole);
if (islog || ishole)
continue;
/* Only print normal classes here */
if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS))
continue;
/* Provide vdev_rebuild_stats to children if available */
if (vrs == NULL) {
(void) nvlist_lookup_uint64_array(nv,
ZPOOL_CONFIG_REBUILD_STATS,
(uint64_t **)&vrs, &i);
}
vname = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags | VDEV_NAME_TYPE_ID);
print_status_config(zhp, cb, vname, child[c], depth + 2,
isspare, vrs);
free(vname);
}
}
/*
* Print the configuration of an exported pool. Iterate over all vdevs in the
* pool, printing out the name and status for each one.
*/
static void
print_import_config(status_cbdata_t *cb, const char *name, nvlist_t *nv,
int depth)
{
nvlist_t **child;
uint_t c, children;
vdev_stat_t *vs;
const char *type;
char *vname;
verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0);
if (strcmp(type, VDEV_TYPE_MISSING) == 0 ||
strcmp(type, VDEV_TYPE_HOLE) == 0)
return;
verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &c) == 0);
(void) printf("\t%*s%-*s", depth, "", cb->cb_namewidth - depth, name);
(void) printf(" %s", zpool_state_to_name(vs->vs_state, vs->vs_aux));
if (vs->vs_aux != 0) {
(void) printf(" ");
switch (vs->vs_aux) {
case VDEV_AUX_OPEN_FAILED:
(void) printf(gettext("cannot open"));
break;
case VDEV_AUX_BAD_GUID_SUM:
(void) printf(gettext("missing device"));
break;
case VDEV_AUX_NO_REPLICAS:
(void) printf(gettext("insufficient replicas"));
break;
case VDEV_AUX_VERSION_NEWER:
(void) printf(gettext("newer version"));
break;
case VDEV_AUX_UNSUP_FEAT:
(void) printf(gettext("unsupported feature(s)"));
break;
case VDEV_AUX_ERR_EXCEEDED:
(void) printf(gettext("too many errors"));
break;
case VDEV_AUX_ACTIVE:
(void) printf(gettext("currently in use"));
break;
case VDEV_AUX_CHILDREN_OFFLINE:
(void) printf(gettext("all children offline"));
break;
case VDEV_AUX_BAD_LABEL:
(void) printf(gettext("invalid label"));
break;
default:
(void) printf(gettext("corrupted data"));
break;
}
}
(void) printf("\n");
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
return;
for (c = 0; c < children; c++) {
uint64_t is_log = B_FALSE;
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&is_log);
if (is_log)
continue;
if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS))
continue;
vname = zpool_vdev_name(g_zfs, NULL, child[c],
cb->cb_name_flags | VDEV_NAME_TYPE_ID);
print_import_config(cb, vname, child[c], depth + 2);
free(vname);
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE,
&child, &children) == 0) {
(void) printf(gettext("\tcache\n"));
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(g_zfs, NULL, child[c],
cb->cb_name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES,
&child, &children) == 0) {
(void) printf(gettext("\tspares\n"));
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(g_zfs, NULL, child[c],
cb->cb_name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
}
/*
* Print specialized class vdevs.
*
* These are recorded as top level vdevs in the main pool child array
* but with "is_log" set to 1 or an "alloc_bias" string. We use either
* print_status_config() or print_import_config() to print the top level
* class vdevs then any of their children (eg mirrored slogs) are printed
* recursively - which works because only the top level vdev is marked.
*/
static void
print_class_vdevs(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t *nv,
const char *class)
{
uint_t c, children;
nvlist_t **child;
boolean_t printed = B_FALSE;
assert(zhp != NULL || !cb->cb_verbose);
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child,
&children) != 0)
return;
for (c = 0; c < children; c++) {
uint64_t is_log = B_FALSE;
const char *bias = NULL;
const char *type = NULL;
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&is_log);
if (is_log) {
bias = (char *)VDEV_ALLOC_CLASS_LOGS;
} else {
(void) nvlist_lookup_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &bias);
(void) nvlist_lookup_string(child[c],
ZPOOL_CONFIG_TYPE, &type);
}
if (bias == NULL || strcmp(bias, class) != 0)
continue;
if (!is_log && strcmp(type, VDEV_TYPE_INDIRECT) == 0)
continue;
if (!printed) {
(void) printf("\t%s\t\n", gettext(class));
printed = B_TRUE;
}
char *name = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags | VDEV_NAME_TYPE_ID);
if (cb->cb_print_status)
print_status_config(zhp, cb, name, child[c], 2,
B_FALSE, NULL);
else
print_import_config(cb, name, child[c], 2);
free(name);
}
}
/*
* Display the status for the given pool.
*/
static int
show_import(nvlist_t *config, boolean_t report_error)
{
uint64_t pool_state;
vdev_stat_t *vs;
const char *name;
uint64_t guid;
uint64_t hostid = 0;
const char *msgid;
const char *hostname = "unknown";
nvlist_t *nvroot, *nvinfo;
zpool_status_t reason;
zpool_errata_t errata;
const char *health;
uint_t vsc;
const char *comment;
status_cbdata_t cb = { 0 };
verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
&name) == 0);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
&guid) == 0);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
&pool_state) == 0);
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &vsc) == 0);
health = zpool_state_to_name(vs->vs_state, vs->vs_aux);
reason = zpool_import_status(config, &msgid, &errata);
/*
* If we're importing using a cachefile, then we won't report any
* errors unless we are in the scan phase of the import.
*/
if (reason != ZPOOL_STATUS_OK && !report_error)
return (reason);
(void) printf(gettext(" pool: %s\n"), name);
(void) printf(gettext(" id: %llu\n"), (u_longlong_t)guid);
(void) printf(gettext(" state: %s"), health);
if (pool_state == POOL_STATE_DESTROYED)
(void) printf(gettext(" (DESTROYED)"));
(void) printf("\n");
switch (reason) {
case ZPOOL_STATUS_MISSING_DEV_R:
case ZPOOL_STATUS_MISSING_DEV_NR:
case ZPOOL_STATUS_BAD_GUID_SUM:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"missing from the system.\n"));
break;
case ZPOOL_STATUS_CORRUPT_LABEL_R:
case ZPOOL_STATUS_CORRUPT_LABEL_NR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices contains"
" corrupted data.\n"));
break;
case ZPOOL_STATUS_CORRUPT_DATA:
(void) printf(
gettext(" status: The pool data is corrupted.\n"));
break;
case ZPOOL_STATUS_OFFLINE_DEV:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices "
"are offlined.\n"));
break;
case ZPOOL_STATUS_CORRUPT_POOL:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool metadata is "
"corrupted.\n"));
break;
case ZPOOL_STATUS_VERSION_OLDER:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool is formatted using "
"a legacy on-disk version.\n"));
break;
case ZPOOL_STATUS_VERSION_NEWER:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool is formatted using "
"an incompatible version.\n"));
break;
case ZPOOL_STATUS_FEAT_DISABLED:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Some supported "
"features are not enabled on the pool.\n\t"
"(Note that they may be intentionally disabled "
"if the\n\t'compatibility' property is set.)\n"));
break;
case ZPOOL_STATUS_COMPATIBILITY_ERR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Error reading or parsing "
"the file(s) indicated by the 'compatibility'\n"
"property.\n"));
break;
case ZPOOL_STATUS_INCOMPATIBLE_FEAT:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more features "
"are enabled on the pool despite not being\n"
"requested by the 'compatibility' property.\n"));
break;
case ZPOOL_STATUS_UNSUP_FEAT_READ:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool uses the following "
"feature(s) not supported on this system:\n"));
color_start(ANSI_YELLOW);
zpool_print_unsup_feat(config);
color_end();
break;
case ZPOOL_STATUS_UNSUP_FEAT_WRITE:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool can only be "
"accessed in read-only mode on this system. It\n\tcannot be"
" accessed in read-write mode because it uses the "
"following\n\tfeature(s) not supported on this system:\n"));
color_start(ANSI_YELLOW);
zpool_print_unsup_feat(config);
color_end();
break;
case ZPOOL_STATUS_HOSTID_ACTIVE:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool is currently "
"imported by another system.\n"));
break;
case ZPOOL_STATUS_HOSTID_REQUIRED:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool has the "
"multihost property on. It cannot\n\tbe safely imported "
"when the system hostid is not set.\n"));
break;
case ZPOOL_STATUS_HOSTID_MISMATCH:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool was last accessed "
"by another system.\n"));
break;
case ZPOOL_STATUS_FAULTED_DEV_R:
case ZPOOL_STATUS_FAULTED_DEV_NR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"faulted.\n"));
break;
case ZPOOL_STATUS_BAD_LOG:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("An intent log record cannot "
"be read.\n"));
break;
case ZPOOL_STATUS_RESILVERING:
case ZPOOL_STATUS_REBUILDING:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices were "
"being resilvered.\n"));
break;
case ZPOOL_STATUS_ERRATA:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Errata #%d detected.\n"),
errata);
break;
case ZPOOL_STATUS_NON_NATIVE_ASHIFT:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"configured to use a non-native block size.\n"
"\tExpect reduced performance.\n"));
break;
default:
/*
* No other status can be seen when importing pools.
*/
assert(reason == ZPOOL_STATUS_OK);
}
/*
* Print out an action according to the overall state of the pool.
*/
if (vs->vs_state == VDEV_STATE_HEALTHY) {
if (reason == ZPOOL_STATUS_VERSION_OLDER ||
reason == ZPOOL_STATUS_FEAT_DISABLED) {
(void) printf(gettext(" action: The pool can be "
"imported using its name or numeric identifier, "
"though\n\tsome features will not be available "
"without an explicit 'zpool upgrade'.\n"));
} else if (reason == ZPOOL_STATUS_COMPATIBILITY_ERR) {
(void) printf(gettext(" action: The pool can be "
"imported using its name or numeric\n\tidentifier, "
"though the file(s) indicated by its "
"'compatibility'\n\tproperty cannot be parsed at "
"this time.\n"));
} else if (reason == ZPOOL_STATUS_HOSTID_MISMATCH) {
(void) printf(gettext(" action: The pool can be "
"imported using its name or numeric "
"identifier and\n\tthe '-f' flag.\n"));
} else if (reason == ZPOOL_STATUS_ERRATA) {
switch (errata) {
case ZPOOL_ERRATA_NONE:
break;
case ZPOOL_ERRATA_ZOL_2094_SCRUB:
(void) printf(gettext(" action: The pool can "
"be imported using its name or numeric "
"identifier,\n\thowever there is a compat"
"ibility issue which should be corrected"
"\n\tby running 'zpool scrub'\n"));
break;
case ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY:
(void) printf(gettext(" action: The pool can"
"not be imported with this version of ZFS "
"due to\n\tan active asynchronous destroy. "
"Revert to an earlier version\n\tand "
"allow the destroy to complete before "
"updating.\n"));
break;
case ZPOOL_ERRATA_ZOL_6845_ENCRYPTION:
(void) printf(gettext(" action: Existing "
"encrypted datasets contain an on-disk "
"incompatibility, which\n\tneeds to be "
"corrected. Backup these datasets to new "
"encrypted datasets\n\tand destroy the "
"old ones.\n"));
break;
case ZPOOL_ERRATA_ZOL_8308_ENCRYPTION:
(void) printf(gettext(" action: Existing "
"encrypted snapshots and bookmarks contain "
"an on-disk\n\tincompatibility. This may "
"cause on-disk corruption if they are used"
"\n\twith 'zfs recv'. To correct the "
"issue, enable the bookmark_v2 feature.\n\t"
"No additional action is needed if there "
"are no encrypted snapshots or\n\t"
"bookmarks. If preserving the encrypted "
"snapshots and bookmarks is\n\trequired, "
"use a non-raw send to backup and restore "
"them. Alternately,\n\tthey may be removed"
" to resolve the incompatibility.\n"));
break;
default:
/*
* All errata must contain an action message.
*/
assert(0);
}
} else {
(void) printf(gettext(" action: The pool can be "
"imported using its name or numeric "
"identifier.\n"));
}
} else if (vs->vs_state == VDEV_STATE_DEGRADED) {
(void) printf(gettext(" action: The pool can be imported "
"despite missing or damaged devices. The\n\tfault "
"tolerance of the pool may be compromised if imported.\n"));
} else {
switch (reason) {
case ZPOOL_STATUS_VERSION_NEWER:
(void) printf(gettext(" action: The pool cannot be "
"imported. Access the pool on a system running "
"newer\n\tsoftware, or recreate the pool from "
"backup.\n"));
break;
case ZPOOL_STATUS_UNSUP_FEAT_READ:
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("The pool cannot be "
"imported. Access the pool on a system that "
"supports\n\tthe required feature(s), or recreate "
"the pool from backup.\n"));
break;
case ZPOOL_STATUS_UNSUP_FEAT_WRITE:
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("The pool cannot be "
"imported in read-write mode. Import the pool "
"with\n"
"\t\"-o readonly=on\", access the pool on a system "
"that supports the\n\trequired feature(s), or "
"recreate the pool from backup.\n"));
break;
case ZPOOL_STATUS_MISSING_DEV_R:
case ZPOOL_STATUS_MISSING_DEV_NR:
case ZPOOL_STATUS_BAD_GUID_SUM:
(void) printf(gettext(" action: The pool cannot be "
"imported. Attach the missing\n\tdevices and try "
"again.\n"));
break;
case ZPOOL_STATUS_HOSTID_ACTIVE:
VERIFY0(nvlist_lookup_nvlist(config,
ZPOOL_CONFIG_LOAD_INFO, &nvinfo));
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME))
hostname = fnvlist_lookup_string(nvinfo,
ZPOOL_CONFIG_MMP_HOSTNAME);
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTID))
hostid = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_HOSTID);
(void) printf(gettext(" action: The pool must be "
"exported from %s (hostid=%"PRIx64")\n\tbefore it "
"can be safely imported.\n"), hostname, hostid);
break;
case ZPOOL_STATUS_HOSTID_REQUIRED:
(void) printf(gettext(" action: Set a unique system "
"hostid with the zgenhostid(8) command.\n"));
break;
default:
(void) printf(gettext(" action: The pool cannot be "
"imported due to damaged devices or data.\n"));
}
}
/* Print the comment attached to the pool. */
if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMMENT, &comment) == 0)
(void) printf(gettext("comment: %s\n"), comment);
/*
* If the state is "closed" or "can't open", and the aux state
* is "corrupt data":
*/
if (((vs->vs_state == VDEV_STATE_CLOSED) ||
(vs->vs_state == VDEV_STATE_CANT_OPEN)) &&
(vs->vs_aux == VDEV_AUX_CORRUPT_DATA)) {
if (pool_state == POOL_STATE_DESTROYED)
(void) printf(gettext("\tThe pool was destroyed, "
"but can be imported using the '-Df' flags.\n"));
else if (pool_state != POOL_STATE_EXPORTED)
(void) printf(gettext("\tThe pool may be active on "
"another system, but can be imported using\n\t"
"the '-f' flag.\n"));
}
if (msgid != NULL) {
(void) printf(gettext(
" see: https://openzfs.github.io/openzfs-docs/msg/%s\n"),
msgid);
}
(void) printf(gettext(" config:\n\n"));
cb.cb_namewidth = max_width(NULL, nvroot, 0, strlen(name),
VDEV_NAME_TYPE_ID);
if (cb.cb_namewidth < 10)
cb.cb_namewidth = 10;
print_import_config(&cb, name, nvroot, 0);
print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_BIAS_DEDUP);
print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_BIAS_SPECIAL);
print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_CLASS_LOGS);
if (reason == ZPOOL_STATUS_BAD_GUID_SUM) {
(void) printf(gettext("\n\tAdditional devices are known to "
"be part of this pool, though their\n\texact "
"configuration cannot be determined.\n"));
}
return (0);
}
static boolean_t
zfs_force_import_required(nvlist_t *config)
{
uint64_t state;
uint64_t hostid = 0;
nvlist_t *nvinfo;
state = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE);
nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
/*
* The hostid on LOAD_INFO comes from the MOS label via
* spa_tryimport(). If its not there then we're likely talking to an
* older kernel, so use the top one, which will be from the label
* discovered in zpool_find_import(), or if a cachefile is in use, the
* local hostid.
*/
if (nvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_HOSTID, &hostid) != 0)
nvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID, &hostid);
if (state != POOL_STATE_EXPORTED && hostid != get_system_hostid())
return (B_TRUE);
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_STATE)) {
mmp_state_t mmp_state = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_STATE);
if (mmp_state != MMP_STATE_INACTIVE)
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Perform the import for the given configuration. This passes the heavy
* lifting off to zpool_import_props(), and then mounts the datasets contained
* within the pool.
*/
static int
do_import(nvlist_t *config, const char *newname, const char *mntopts,
nvlist_t *props, int flags)
{
int ret = 0;
int ms_status = 0;
zpool_handle_t *zhp;
const char *name;
uint64_t version;
name = fnvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME);
version = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION);
if (!SPA_VERSION_IS_SUPPORTED(version)) {
(void) fprintf(stderr, gettext("cannot import '%s': pool "
"is formatted using an unsupported ZFS version\n"), name);
return (1);
} else if (zfs_force_import_required(config) &&
!(flags & ZFS_IMPORT_ANY_HOST)) {
mmp_state_t mmp_state = MMP_STATE_INACTIVE;
nvlist_t *nvinfo;
nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_STATE))
mmp_state = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_STATE);
if (mmp_state == MMP_STATE_ACTIVE) {
const char *hostname = "<unknown>";
uint64_t hostid = 0;
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME))
hostname = fnvlist_lookup_string(nvinfo,
ZPOOL_CONFIG_MMP_HOSTNAME);
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTID))
hostid = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_HOSTID);
(void) fprintf(stderr, gettext("cannot import '%s': "
"pool is imported on %s (hostid: "
"0x%"PRIx64")\nExport the pool on the other "
"system, then run 'zpool import'.\n"),
name, hostname, hostid);
} else if (mmp_state == MMP_STATE_NO_HOSTID) {
(void) fprintf(stderr, gettext("Cannot import '%s': "
"pool has the multihost property on and the\n"
"system's hostid is not set. Set a unique hostid "
"with the zgenhostid(8) command.\n"), name);
} else {
const char *hostname = "<unknown>";
time_t timestamp = 0;
uint64_t hostid = 0;
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_HOSTNAME))
hostname = fnvlist_lookup_string(nvinfo,
ZPOOL_CONFIG_HOSTNAME);
else 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(nvinfo, ZPOOL_CONFIG_HOSTID))
hostid = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_HOSTID);
else if (nvlist_exists(config, ZPOOL_CONFIG_HOSTID))
hostid = fnvlist_lookup_uint64(config,
ZPOOL_CONFIG_HOSTID);
(void) fprintf(stderr, gettext("cannot import '%s': "
"pool was previously in use from another system.\n"
"Last accessed by %s (hostid=%"PRIx64") at %s"
"The pool can be imported, use 'zpool import -f' "
"to import the pool.\n"), name, hostname,
hostid, ctime(&timestamp));
}
return (1);
}
if (zpool_import_props(g_zfs, config, newname, props, flags) != 0)
return (1);
if (newname != NULL)
name = newname;
if ((zhp = zpool_open_canfail(g_zfs, name)) == NULL)
return (1);
/*
* Loading keys is best effort. We don't want to return immediately
* if it fails but we do want to give the error to the caller.
*/
if (flags & ZFS_IMPORT_LOAD_KEYS &&
zfs_crypto_attempt_load_keys(g_zfs, name) != 0)
ret = 1;
if (zpool_get_state(zhp) != POOL_STATE_UNAVAIL &&
!(flags & ZFS_IMPORT_ONLY)) {
ms_status = zpool_enable_datasets(zhp, mntopts, 0);
if (ms_status == EZFS_SHAREFAILED) {
(void) fprintf(stderr, gettext("Import was "
"successful, but unable to share some datasets"));
} else if (ms_status == EZFS_MOUNTFAILED) {
(void) fprintf(stderr, gettext("Import was "
"successful, but unable to mount some datasets"));
}
}
zpool_close(zhp);
return (ret);
}
static int
import_pools(nvlist_t *pools, nvlist_t *props, char *mntopts, int flags,
char *orig_name, char *new_name,
boolean_t do_destroyed, boolean_t pool_specified, boolean_t do_all,
importargs_t *import)
{
nvlist_t *config = NULL;
nvlist_t *found_config = NULL;
uint64_t pool_state;
/*
* At this point we have a list of import candidate configs. Even if
* we were searching by pool name or guid, we still need to
* post-process the list to deal with pool state and possible
* duplicate names.
*/
int err = 0;
nvpair_t *elem = NULL;
boolean_t first = B_TRUE;
while ((elem = nvlist_next_nvpair(pools, elem)) != NULL) {
verify(nvpair_value_nvlist(elem, &config) == 0);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
&pool_state) == 0);
if (!do_destroyed && pool_state == POOL_STATE_DESTROYED)
continue;
if (do_destroyed && pool_state != POOL_STATE_DESTROYED)
continue;
verify(nvlist_add_nvlist(config, ZPOOL_LOAD_POLICY,
import->policy) == 0);
if (!pool_specified) {
if (first)
first = B_FALSE;
else if (!do_all)
(void) fputc('\n', stdout);
if (do_all) {
err |= do_import(config, NULL, mntopts,
props, flags);
} else {
/*
* If we're importing from cachefile, then
* we don't want to report errors until we
* are in the scan phase of the import. If
* we get an error, then we return that error
* to invoke the scan phase.
*/
if (import->cachefile && !import->scan)
err = show_import(config, B_FALSE);
else
(void) show_import(config, B_TRUE);
}
} else if (import->poolname != NULL) {
const char *name;
/*
* We are searching for a pool based on name.
*/
verify(nvlist_lookup_string(config,
ZPOOL_CONFIG_POOL_NAME, &name) == 0);
if (strcmp(name, import->poolname) == 0) {
if (found_config != NULL) {
(void) fprintf(stderr, gettext(
"cannot import '%s': more than "
"one matching pool\n"),
import->poolname);
(void) fprintf(stderr, gettext(
"import by numeric ID instead\n"));
err = B_TRUE;
}
found_config = config;
}
} else {
uint64_t guid;
/*
* Search for a pool by guid.
*/
verify(nvlist_lookup_uint64(config,
ZPOOL_CONFIG_POOL_GUID, &guid) == 0);
if (guid == import->guid)
found_config = config;
}
}
/*
* If we were searching for a specific pool, verify that we found a
* pool, and then do the import.
*/
if (pool_specified && err == 0) {
if (found_config == NULL) {
(void) fprintf(stderr, gettext("cannot import '%s': "
"no such pool available\n"), orig_name);
err = B_TRUE;
} else {
err |= do_import(found_config, new_name,
mntopts, props, flags);
}
}
/*
* If we were just looking for pools, report an error if none were
* found.
*/
if (!pool_specified && first)
(void) fprintf(stderr,
gettext("no pools available to import\n"));
return (err);
}
typedef struct target_exists_args {
const char *poolname;
uint64_t poolguid;
} target_exists_args_t;
static int
name_or_guid_exists(zpool_handle_t *zhp, void *data)
{
target_exists_args_t *args = data;
nvlist_t *config = zpool_get_config(zhp, NULL);
int found = 0;
if (config == NULL)
return (0);
if (args->poolname != NULL) {
const char *pool_name;
verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
&pool_name) == 0);
if (strcmp(pool_name, args->poolname) == 0)
found = 1;
} else {
uint64_t pool_guid;
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
&pool_guid) == 0);
if (pool_guid == args->poolguid)
found = 1;
}
zpool_close(zhp);
return (found);
}
/*
* zpool checkpoint <pool>
* checkpoint --discard <pool>
*
* -d Discard the checkpoint from a checkpointed
* --discard pool.
*
* -w Wait for discarding a checkpoint to complete.
* --wait
*
* Checkpoints the specified pool, by taking a "snapshot" of its
* current state. A pool can only have one checkpoint at a time.
*/
int
zpool_do_checkpoint(int argc, char **argv)
{
boolean_t discard, wait;
char *pool;
zpool_handle_t *zhp;
int c, err;
struct option long_options[] = {
{"discard", no_argument, NULL, 'd'},
{"wait", no_argument, NULL, 'w'},
{0, 0, 0, 0}
};
discard = B_FALSE;
wait = B_FALSE;
while ((c = getopt_long(argc, argv, ":dw", long_options, NULL)) != -1) {
switch (c) {
case 'd':
discard = B_TRUE;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
if (wait && !discard) {
(void) fprintf(stderr, gettext("--wait only valid when "
"--discard also specified\n"));
usage(B_FALSE);
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
pool = argv[0];
if ((zhp = zpool_open(g_zfs, pool)) == NULL) {
/* As a special case, check for use of '/' in the name */
if (strchr(pool, '/') != NULL)
(void) fprintf(stderr, gettext("'zpool checkpoint' "
"doesn't work on datasets. To save the state "
"of a dataset from a specific point in time "
"please use 'zfs snapshot'\n"));
return (1);
}
if (discard) {
err = (zpool_discard_checkpoint(zhp) != 0);
if (err == 0 && wait)
err = zpool_wait(zhp, ZPOOL_WAIT_CKPT_DISCARD);
} else {
err = (zpool_checkpoint(zhp) != 0);
}
zpool_close(zhp);
return (err);
}
#define CHECKPOINT_OPT 1024
/*
* zpool import [-d dir] [-D]
* import [-o mntopts] [-o prop=value] ... [-R root] [-D] [-l]
* [-d dir | -c cachefile | -s] [-f] -a
* import [-o mntopts] [-o prop=value] ... [-R root] [-D] [-l]
* [-d dir | -c cachefile | -s] [-f] [-n] [-F] <pool | id>
* [newpool]
*
* -c Read pool information from a cachefile instead of searching
* devices. If importing from a cachefile config fails, then
* fallback to searching for devices only in the directories that
* exist in the cachefile.
*
* -d Scan in a specific directory, other than /dev/. More than
* one directory can be specified using multiple '-d' options.
*
* -D Scan for previously destroyed pools or import all or only
* specified destroyed pools.
*
* -R Temporarily import the pool, with all mountpoints relative to
* the given root. The pool will remain exported when the machine
* is rebooted.
*
* -V Import even in the presence of faulted vdevs. This is an
* intentionally undocumented option for testing purposes, and
* treats the pool configuration as complete, leaving any bad
* vdevs in the FAULTED state. In other words, it does verbatim
* import.
*
* -f Force import, even if it appears that the pool is active.
*
* -F Attempt rewind if necessary.
*
* -n See if rewind would work, but don't actually rewind.
*
* -N Import the pool but don't mount datasets.
*
* -T Specify a starting txg to use for import. This option is
* intentionally undocumented option for testing purposes.
*
* -a Import all pools found.
*
* -l Load encryption keys while importing.
*
* -o Set property=value and/or temporary mount options (without '=').
*
* -s Scan using the default search path, the libblkid cache will
* not be consulted.
*
* --rewind-to-checkpoint
* Import the pool and revert back to the checkpoint.
*
* The import command scans for pools to import, and import pools based on pool
* name and GUID. The pool can also be renamed as part of the import process.
*/
int
zpool_do_import(int argc, char **argv)
{
char **searchdirs = NULL;
char *env, *envdup = NULL;
int nsearch = 0;
int c;
int err = 0;
nvlist_t *pools = NULL;
boolean_t do_all = B_FALSE;
boolean_t do_destroyed = B_FALSE;
char *mntopts = NULL;
uint64_t searchguid = 0;
char *searchname = NULL;
char *propval;
nvlist_t *policy = NULL;
nvlist_t *props = NULL;
int flags = ZFS_IMPORT_NORMAL;
uint32_t rewind_policy = ZPOOL_NO_REWIND;
boolean_t dryrun = B_FALSE;
boolean_t do_rewind = B_FALSE;
boolean_t xtreme_rewind = B_FALSE;
boolean_t do_scan = B_FALSE;
boolean_t pool_exists = B_FALSE;
boolean_t pool_specified = B_FALSE;
uint64_t txg = -1ULL;
char *cachefile = NULL;
importargs_t idata = { 0 };
char *endptr;
struct option long_options[] = {
{"rewind-to-checkpoint", no_argument, NULL, CHECKPOINT_OPT},
{0, 0, 0, 0}
};
/* check options */
while ((c = getopt_long(argc, argv, ":aCc:d:DEfFlmnNo:R:stT:VX",
long_options, NULL)) != -1) {
switch (c) {
case 'a':
do_all = B_TRUE;
break;
case 'c':
cachefile = optarg;
break;
case 'd':
searchdirs = safe_realloc(searchdirs,
(nsearch + 1) * sizeof (char *));
searchdirs[nsearch++] = optarg;
break;
case 'D':
do_destroyed = B_TRUE;
break;
case 'f':
flags |= ZFS_IMPORT_ANY_HOST;
break;
case 'F':
do_rewind = B_TRUE;
break;
case 'l':
flags |= ZFS_IMPORT_LOAD_KEYS;
break;
case 'm':
flags |= ZFS_IMPORT_MISSING_LOG;
break;
case 'n':
dryrun = B_TRUE;
break;
case 'N':
flags |= ZFS_IMPORT_ONLY;
break;
case 'o':
if ((propval = strchr(optarg, '=')) != NULL) {
*propval = '\0';
propval++;
if (add_prop_list(optarg, propval,
&props, B_TRUE))
goto error;
} else {
mntopts = optarg;
}
break;
case 'R':
if (add_prop_list(zpool_prop_to_name(
ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE))
goto error;
if (add_prop_list_default(zpool_prop_to_name(
ZPOOL_PROP_CACHEFILE), "none", &props))
goto error;
break;
case 's':
do_scan = B_TRUE;
break;
case 't':
flags |= ZFS_IMPORT_TEMP_NAME;
if (add_prop_list_default(zpool_prop_to_name(
ZPOOL_PROP_CACHEFILE), "none", &props))
goto error;
break;
case 'T':
errno = 0;
txg = strtoull(optarg, &endptr, 0);
if (errno != 0 || *endptr != '\0') {
(void) fprintf(stderr,
gettext("invalid txg value\n"));
usage(B_FALSE);
}
rewind_policy = ZPOOL_DO_REWIND | ZPOOL_EXTREME_REWIND;
break;
case 'V':
flags |= ZFS_IMPORT_VERBATIM;
break;
case 'X':
xtreme_rewind = B_TRUE;
break;
case CHECKPOINT_OPT:
flags |= ZFS_IMPORT_CHECKPOINT;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (cachefile && nsearch != 0) {
(void) fprintf(stderr, gettext("-c is incompatible with -d\n"));
usage(B_FALSE);
}
if (cachefile && do_scan) {
(void) fprintf(stderr, gettext("-c is incompatible with -s\n"));
usage(B_FALSE);
}
if ((flags & ZFS_IMPORT_LOAD_KEYS) && (flags & ZFS_IMPORT_ONLY)) {
(void) fprintf(stderr, gettext("-l is incompatible with -N\n"));
usage(B_FALSE);
}
if ((flags & ZFS_IMPORT_LOAD_KEYS) && !do_all && argc == 0) {
(void) fprintf(stderr, gettext("-l is only meaningful during "
"an import\n"));
usage(B_FALSE);
}
if ((dryrun || xtreme_rewind) && !do_rewind) {
(void) fprintf(stderr,
gettext("-n or -X only meaningful with -F\n"));
usage(B_FALSE);
}
if (dryrun)
rewind_policy = ZPOOL_TRY_REWIND;
else if (do_rewind)
rewind_policy = ZPOOL_DO_REWIND;
if (xtreme_rewind)
rewind_policy |= ZPOOL_EXTREME_REWIND;
/* In the future, we can capture further policy and include it here */
if (nvlist_alloc(&policy, NV_UNIQUE_NAME, 0) != 0 ||
nvlist_add_uint64(policy, ZPOOL_LOAD_REQUEST_TXG, txg) != 0 ||
nvlist_add_uint32(policy, ZPOOL_LOAD_REWIND_POLICY,
rewind_policy) != 0)
goto error;
/* check argument count */
if (do_all) {
if (argc != 0) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
} else {
if (argc > 2) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
}
/*
* Check for the effective uid. We do this explicitly here because
* otherwise any attempt to discover pools will silently fail.
*/
if (argc == 0 && geteuid() != 0) {
(void) fprintf(stderr, gettext("cannot "
"discover pools: permission denied\n"));
free(searchdirs);
nvlist_free(props);
nvlist_free(policy);
return (1);
}
/*
* Depending on the arguments given, we do one of the following:
*
* <none> Iterate through all pools and display information about
* each one.
*
* -a Iterate through all pools and try to import each one.
*
* <id> Find the pool that corresponds to the given GUID/pool
* name and import that one.
*
* -D Above options applies only to destroyed pools.
*/
if (argc != 0) {
char *endptr;
errno = 0;
searchguid = strtoull(argv[0], &endptr, 10);
if (errno != 0 || *endptr != '\0') {
searchname = argv[0];
searchguid = 0;
}
pool_specified = B_TRUE;
/*
* User specified a name or guid. Ensure it's unique.
*/
target_exists_args_t search = {searchname, searchguid};
pool_exists = zpool_iter(g_zfs, name_or_guid_exists, &search);
}
/*
* Check the environment for the preferred search path.
*/
if ((searchdirs == NULL) && (env = getenv("ZPOOL_IMPORT_PATH"))) {
char *dir, *tmp = NULL;
envdup = strdup(env);
for (dir = strtok_r(envdup, ":", &tmp);
dir != NULL;
dir = strtok_r(NULL, ":", &tmp)) {
searchdirs = safe_realloc(searchdirs,
(nsearch + 1) * sizeof (char *));
searchdirs[nsearch++] = dir;
}
}
idata.path = searchdirs;
idata.paths = nsearch;
idata.poolname = searchname;
idata.guid = searchguid;
idata.cachefile = cachefile;
idata.scan = do_scan;
idata.policy = policy;
libpc_handle_t lpch = {
.lpc_lib_handle = g_zfs,
.lpc_ops = &libzfs_config_ops,
.lpc_printerr = B_TRUE
};
pools = zpool_search_import(&lpch, &idata);
if (pools != NULL && pool_exists &&
(argc == 1 || strcmp(argv[0], argv[1]) == 0)) {
(void) fprintf(stderr, gettext("cannot import '%s': "
"a pool with that name already exists\n"),
argv[0]);
(void) fprintf(stderr, gettext("use the form '%s "
"<pool | id> <newpool>' to give it a new name\n"),
"zpool import");
err = 1;
} else if (pools == NULL && pool_exists) {
(void) fprintf(stderr, gettext("cannot import '%s': "
"a pool with that name is already created/imported,\n"),
argv[0]);
(void) fprintf(stderr, gettext("and no additional pools "
"with that name were found\n"));
err = 1;
} else if (pools == NULL) {
if (argc != 0) {
(void) fprintf(stderr, gettext("cannot import '%s': "
"no such pool available\n"), argv[0]);
}
err = 1;
}
if (err == 1) {
free(searchdirs);
free(envdup);
nvlist_free(policy);
nvlist_free(pools);
nvlist_free(props);
return (1);
}
err = import_pools(pools, props, mntopts, flags,
argc >= 1 ? argv[0] : NULL,
argc >= 2 ? argv[1] : NULL,
do_destroyed, pool_specified, do_all, &idata);
/*
* If we're using the cachefile and we failed to import, then
* fallback to scanning the directory for pools that match
* those in the cachefile.
*/
if (err != 0 && cachefile != NULL) {
(void) printf(gettext("cachefile import failed, retrying\n"));
/*
* We use the scan flag to gather the directories that exist
* in the cachefile. If we need to fallback to searching for
* the pool config, we will only search devices in these
* directories.
*/
idata.scan = B_TRUE;
nvlist_free(pools);
pools = zpool_search_import(&lpch, &idata);
err = import_pools(pools, props, mntopts, flags,
argc >= 1 ? argv[0] : NULL,
argc >= 2 ? argv[1] : NULL,
do_destroyed, pool_specified, do_all, &idata);
}
error:
nvlist_free(props);
nvlist_free(pools);
nvlist_free(policy);
free(searchdirs);
free(envdup);
return (err ? 1 : 0);
}
/*
* zpool sync [-f] [pool] ...
*
* -f (undocumented) force uberblock (and config including zpool cache file)
* update.
*
* Sync the specified pool(s).
* Without arguments "zpool sync" will sync all pools.
* This command initiates TXG sync(s) and will return after the TXG(s) commit.
*
*/
static int
zpool_do_sync(int argc, char **argv)
{
int ret;
boolean_t force = B_FALSE;
/* check options */
while ((ret = getopt(argc, argv, "f")) != -1) {
switch (ret) {
case 'f':
force = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* if argc == 0 we will execute zpool_sync_one on all pools */
ret = for_each_pool(argc, argv, B_FALSE, NULL, ZFS_TYPE_POOL,
B_FALSE, zpool_sync_one, &force);
return (ret);
}
typedef struct iostat_cbdata {
uint64_t cb_flags;
int cb_namewidth;
int cb_iteration;
boolean_t cb_verbose;
boolean_t cb_literal;
boolean_t cb_scripted;
zpool_list_t *cb_list;
vdev_cmd_data_list_t *vcdl;
vdev_cbdata_t cb_vdevs;
} iostat_cbdata_t;
/* iostat labels */
typedef struct name_and_columns {
const char *name; /* Column name */
unsigned int columns; /* Center name to this number of columns */
} name_and_columns_t;
#define IOSTAT_MAX_LABELS 15 /* Max number of labels on one line */
static const name_and_columns_t iostat_top_labels[][IOSTAT_MAX_LABELS] =
{
[IOS_DEFAULT] = {{"capacity", 2}, {"operations", 2}, {"bandwidth", 2},
{NULL}},
[IOS_LATENCY] = {{"total_wait", 2}, {"disk_wait", 2}, {"syncq_wait", 2},
{"asyncq_wait", 2}, {"scrub", 1}, {"trim", 1}, {"rebuild", 1},
{NULL}},
[IOS_QUEUES] = {{"syncq_read", 2}, {"syncq_write", 2},
{"asyncq_read", 2}, {"asyncq_write", 2}, {"scrubq_read", 2},
{"trimq_write", 2}, {"rebuildq_write", 2}, {NULL}},
[IOS_L_HISTO] = {{"total_wait", 2}, {"disk_wait", 2}, {"syncq_wait", 2},
{"asyncq_wait", 2}, {NULL}},
[IOS_RQ_HISTO] = {{"sync_read", 2}, {"sync_write", 2},
{"async_read", 2}, {"async_write", 2}, {"scrub", 2},
{"trim", 2}, {"rebuild", 2}, {NULL}},
};
/* Shorthand - if "columns" field not set, default to 1 column */
static const name_and_columns_t iostat_bottom_labels[][IOSTAT_MAX_LABELS] =
{
[IOS_DEFAULT] = {{"alloc"}, {"free"}, {"read"}, {"write"}, {"read"},
{"write"}, {NULL}},
[IOS_LATENCY] = {{"read"}, {"write"}, {"read"}, {"write"}, {"read"},
{"write"}, {"read"}, {"write"}, {"wait"}, {"wait"}, {"wait"},
{NULL}},
[IOS_QUEUES] = {{"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"},
{"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"},
{"pend"}, {"activ"}, {"pend"}, {"activ"}, {NULL}},
[IOS_L_HISTO] = {{"read"}, {"write"}, {"read"}, {"write"}, {"read"},
{"write"}, {"read"}, {"write"}, {"scrub"}, {"trim"}, {"rebuild"},
{NULL}},
[IOS_RQ_HISTO] = {{"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"},
{"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"},
{"ind"}, {"agg"}, {NULL}},
};
static const char *histo_to_title[] = {
[IOS_L_HISTO] = "latency",
[IOS_RQ_HISTO] = "req_size",
};
/*
* Return the number of labels in a null-terminated name_and_columns_t
* array.
*
*/
static unsigned int
label_array_len(const name_and_columns_t *labels)
{
int i = 0;
while (labels[i].name)
i++;
return (i);
}
/*
* Return the number of strings in a null-terminated string array.
* For example:
*
* const char foo[] = {"bar", "baz", NULL}
*
* returns 2
*/
static uint64_t
str_array_len(const char *array[])
{
uint64_t i = 0;
while (array[i])
i++;
return (i);
}
/*
* Return a default column width for default/latency/queue columns. This does
* not include histograms, which have their columns autosized.
*/
static unsigned int
default_column_width(iostat_cbdata_t *cb, enum iostat_type type)
{
unsigned long column_width = 5; /* Normal niceprint */
static unsigned long widths[] = {
/*
* Choose some sane default column sizes for printing the
* raw numbers.
*/
[IOS_DEFAULT] = 15, /* 1PB capacity */
[IOS_LATENCY] = 10, /* 1B ns = 10sec */
[IOS_QUEUES] = 6, /* 1M queue entries */
[IOS_L_HISTO] = 10, /* 1B ns = 10sec */
[IOS_RQ_HISTO] = 6, /* 1M queue entries */
};
if (cb->cb_literal)
column_width = widths[type];
return (column_width);
}
/*
* Print the column labels, i.e:
*
* capacity operations bandwidth
* alloc free read write read write ...
*
* If force_column_width is set, use it for the column width. If not set, use
* the default column width.
*/
static void
print_iostat_labels(iostat_cbdata_t *cb, unsigned int force_column_width,
const name_and_columns_t labels[][IOSTAT_MAX_LABELS])
{
int i, idx, s;
int text_start, rw_column_width, spaces_to_end;
uint64_t flags = cb->cb_flags;
uint64_t f;
unsigned int column_width = force_column_width;
/* For each bit set in flags */
for (f = flags; f; f &= ~(1ULL << idx)) {
idx = lowbit64(f) - 1;
if (!force_column_width)
column_width = default_column_width(cb, idx);
/* Print our top labels centered over "read write" label. */
for (i = 0; i < label_array_len(labels[idx]); i++) {
const char *name = labels[idx][i].name;
/*
* We treat labels[][].columns == 0 as shorthand
* for one column. It makes writing out the label
* tables more concise.
*/
unsigned int columns = MAX(1, labels[idx][i].columns);
unsigned int slen = strlen(name);
rw_column_width = (column_width * columns) +
(2 * (columns - 1));
text_start = (int)((rw_column_width) / columns -
slen / columns);
if (text_start < 0)
text_start = 0;
printf(" "); /* Two spaces between columns */
/* Space from beginning of column to label */
for (s = 0; s < text_start; s++)
printf(" ");
printf("%s", name);
/* Print space after label to end of column */
spaces_to_end = rw_column_width - text_start - slen;
if (spaces_to_end < 0)
spaces_to_end = 0;
for (s = 0; s < spaces_to_end; s++)
printf(" ");
}
}
}
/*
* print_cmd_columns - Print custom column titles from -c
*
* If the user specified the "zpool status|iostat -c" then print their custom
* column titles in the header. For example, print_cmd_columns() would print
* the " col1 col2" part of this:
*
* $ zpool iostat -vc 'echo col1=val1; echo col2=val2'
* ...
* capacity operations bandwidth
* pool alloc free read write read write col1 col2
* ---------- ----- ----- ----- ----- ----- ----- ---- ----
* mypool 269K 1008M 0 0 107 946
* mirror 269K 1008M 0 0 107 946
* sdb - - 0 0 102 473 val1 val2
* sdc - - 0 0 5 473 val1 val2
* ---------- ----- ----- ----- ----- ----- ----- ---- ----
*/
static void
print_cmd_columns(vdev_cmd_data_list_t *vcdl, int use_dashes)
{
int i, j;
vdev_cmd_data_t *data = &vcdl->data[0];
if (vcdl->count == 0 || data == NULL)
return;
/*
* Each vdev cmd should have the same column names unless the user did
* something weird with their cmd. Just take the column names from the
* first vdev and assume it works for all of them.
*/
for (i = 0; i < vcdl->uniq_cols_cnt; i++) {
printf(" ");
if (use_dashes) {
for (j = 0; j < vcdl->uniq_cols_width[i]; j++)
printf("-");
} else {
printf_color(ANSI_BOLD, "%*s", vcdl->uniq_cols_width[i],
vcdl->uniq_cols[i]);
}
}
}
/*
* Utility function to print out a line of dashes like:
*
* -------------------------------- ----- ----- ----- ----- -----
*
* ...or a dashed named-row line like:
*
* logs - - - - -
*
* @cb: iostat data
*
* @force_column_width If non-zero, use the value as the column width.
* Otherwise use the default column widths.
*
* @name: Print a dashed named-row line starting
* with @name. Otherwise, print a regular
* dashed line.
*/
static void
print_iostat_dashes(iostat_cbdata_t *cb, unsigned int force_column_width,
const char *name)
{
int i;
unsigned int namewidth;
uint64_t flags = cb->cb_flags;
uint64_t f;
int idx;
const name_and_columns_t *labels;
const char *title;
if (cb->cb_flags & IOS_ANYHISTO_M) {
title = histo_to_title[IOS_HISTO_IDX(cb->cb_flags)];
} else if (cb->cb_vdevs.cb_names_count) {
title = "vdev";
} else {
title = "pool";
}
namewidth = MAX(MAX(strlen(title), cb->cb_namewidth),
name ? strlen(name) : 0);
if (name) {
printf("%-*s", namewidth, name);
} else {
for (i = 0; i < namewidth; i++)
(void) printf("-");
}
/* For each bit in flags */
for (f = flags; f; f &= ~(1ULL << idx)) {
unsigned int column_width;
idx = lowbit64(f) - 1;
if (force_column_width)
column_width = force_column_width;
else
column_width = default_column_width(cb, idx);
labels = iostat_bottom_labels[idx];
for (i = 0; i < label_array_len(labels); i++) {
if (name)
printf(" %*s-", column_width - 1, " ");
else
printf(" %.*s", column_width,
"--------------------");
}
}
}
static void
print_iostat_separator_impl(iostat_cbdata_t *cb,
unsigned int force_column_width)
{
print_iostat_dashes(cb, force_column_width, NULL);
}
static void
print_iostat_separator(iostat_cbdata_t *cb)
{
print_iostat_separator_impl(cb, 0);
}
static void
print_iostat_header_impl(iostat_cbdata_t *cb, unsigned int force_column_width,
const char *histo_vdev_name)
{
unsigned int namewidth;
const char *title;
color_start(ANSI_BOLD);
if (cb->cb_flags & IOS_ANYHISTO_M) {
title = histo_to_title[IOS_HISTO_IDX(cb->cb_flags)];
} else if (cb->cb_vdevs.cb_names_count) {
title = "vdev";
} else {
title = "pool";
}
namewidth = MAX(MAX(strlen(title), cb->cb_namewidth),
histo_vdev_name ? strlen(histo_vdev_name) : 0);
if (histo_vdev_name)
printf("%-*s", namewidth, histo_vdev_name);
else
printf("%*s", namewidth, "");
print_iostat_labels(cb, force_column_width, iostat_top_labels);
printf("\n");
printf("%-*s", namewidth, title);
print_iostat_labels(cb, force_column_width, iostat_bottom_labels);
if (cb->vcdl != NULL)
print_cmd_columns(cb->vcdl, 0);
printf("\n");
print_iostat_separator_impl(cb, force_column_width);
if (cb->vcdl != NULL)
print_cmd_columns(cb->vcdl, 1);
color_end();
printf("\n");
}
static void
print_iostat_header(iostat_cbdata_t *cb)
{
print_iostat_header_impl(cb, 0, NULL);
}
/*
* Prints a size string (i.e. 120M) with the suffix ("M") colored
* by order of magnitude. Uses column_size to add padding.
*/
static void
print_stat_color(const char *statbuf, unsigned int column_size)
{
fputs(" ", stdout);
size_t len = strlen(statbuf);
while (len < column_size) {
fputc(' ', stdout);
column_size--;
}
if (*statbuf == '0') {
color_start(ANSI_GRAY);
fputc('0', stdout);
} else {
for (; *statbuf; statbuf++) {
if (*statbuf == 'K') color_start(ANSI_GREEN);
else if (*statbuf == 'M') color_start(ANSI_YELLOW);
else if (*statbuf == 'G') color_start(ANSI_RED);
else if (*statbuf == 'T') color_start(ANSI_BOLD_BLUE);
else if (*statbuf == 'P') color_start(ANSI_MAGENTA);
else if (*statbuf == 'E') color_start(ANSI_CYAN);
fputc(*statbuf, stdout);
if (--column_size <= 0)
break;
}
}
color_end();
}
/*
* Display a single statistic.
*/
static void
print_one_stat(uint64_t value, enum zfs_nicenum_format format,
unsigned int column_size, boolean_t scripted)
{
char buf[64];
zfs_nicenum_format(value, buf, sizeof (buf), format);
if (scripted)
printf("\t%s", buf);
else
print_stat_color(buf, column_size);
}
/*
* Calculate the default vdev stats
*
* Subtract oldvs from newvs, apply a scaling factor, and save the resulting
* stats into calcvs.
*/
static void
calc_default_iostats(vdev_stat_t *oldvs, vdev_stat_t *newvs,
vdev_stat_t *calcvs)
{
int i;
memcpy(calcvs, newvs, sizeof (*calcvs));
for (i = 0; i < ARRAY_SIZE(calcvs->vs_ops); i++)
calcvs->vs_ops[i] = (newvs->vs_ops[i] - oldvs->vs_ops[i]);
for (i = 0; i < ARRAY_SIZE(calcvs->vs_bytes); i++)
calcvs->vs_bytes[i] = (newvs->vs_bytes[i] - oldvs->vs_bytes[i]);
}
/*
* Internal representation of the extended iostats data.
*
* The extended iostat stats are exported in nvlists as either uint64_t arrays
* or single uint64_t's. We make both look like arrays to make them easier
* to process. In order to make single uint64_t's look like arrays, we set
* __data to the stat data, and then set *data = &__data with count = 1. Then,
* we can just use *data and count.
*/
struct stat_array {
uint64_t *data;
uint_t count; /* Number of entries in data[] */
uint64_t __data; /* Only used when data is a single uint64_t */
};
static uint64_t
stat_histo_max(struct stat_array *nva, unsigned int len)
{
uint64_t max = 0;
int i;
for (i = 0; i < len; i++)
max = MAX(max, array64_max(nva[i].data, nva[i].count));
return (max);
}
/*
* Helper function to lookup a uint64_t array or uint64_t value and store its
* data as a stat_array. If the nvpair is a single uint64_t value, then we make
* it look like a one element array to make it easier to process.
*/
static int
nvpair64_to_stat_array(nvlist_t *nvl, const char *name,
struct stat_array *nva)
{
nvpair_t *tmp;
int ret;
verify(nvlist_lookup_nvpair(nvl, name, &tmp) == 0);
switch (nvpair_type(tmp)) {
case DATA_TYPE_UINT64_ARRAY:
ret = nvpair_value_uint64_array(tmp, &nva->data, &nva->count);
break;
case DATA_TYPE_UINT64:
ret = nvpair_value_uint64(tmp, &nva->__data);
nva->data = &nva->__data;
nva->count = 1;
break;
default:
/* Not a uint64_t */
ret = EINVAL;
break;
}
return (ret);
}
/*
* Given a list of nvlist names, look up the extended stats in newnv and oldnv,
* subtract them, and return the results in a newly allocated stat_array.
* You must free the returned array after you are done with it with
* free_calc_stats().
*
* Additionally, you can set "oldnv" to NULL if you simply want the newnv
* values.
*/
static struct stat_array *
calc_and_alloc_stats_ex(const char **names, unsigned int len, nvlist_t *oldnv,
nvlist_t *newnv)
{
nvlist_t *oldnvx = NULL, *newnvx;
struct stat_array *oldnva, *newnva, *calcnva;
int i, j;
unsigned int alloc_size = (sizeof (struct stat_array)) * len;
/* Extract our extended stats nvlist from the main list */
verify(nvlist_lookup_nvlist(newnv, ZPOOL_CONFIG_VDEV_STATS_EX,
&newnvx) == 0);
if (oldnv) {
verify(nvlist_lookup_nvlist(oldnv, ZPOOL_CONFIG_VDEV_STATS_EX,
&oldnvx) == 0);
}
newnva = safe_malloc(alloc_size);
oldnva = safe_malloc(alloc_size);
calcnva = safe_malloc(alloc_size);
for (j = 0; j < len; j++) {
verify(nvpair64_to_stat_array(newnvx, names[j],
&newnva[j]) == 0);
calcnva[j].count = newnva[j].count;
alloc_size = calcnva[j].count * sizeof (calcnva[j].data[0]);
calcnva[j].data = safe_malloc(alloc_size);
memcpy(calcnva[j].data, newnva[j].data, alloc_size);
if (oldnvx) {
verify(nvpair64_to_stat_array(oldnvx, names[j],
&oldnva[j]) == 0);
for (i = 0; i < oldnva[j].count; i++)
calcnva[j].data[i] -= oldnva[j].data[i];
}
}
free(newnva);
free(oldnva);
return (calcnva);
}
static void
free_calc_stats(struct stat_array *nva, unsigned int len)
{
int i;
for (i = 0; i < len; i++)
free(nva[i].data);
free(nva);
}
static void
print_iostat_histo(struct stat_array *nva, unsigned int len,
iostat_cbdata_t *cb, unsigned int column_width, unsigned int namewidth,
double scale)
{
int i, j;
char buf[6];
uint64_t val;
enum zfs_nicenum_format format;
unsigned int buckets;
unsigned int start_bucket;
if (cb->cb_literal)
format = ZFS_NICENUM_RAW;
else
format = ZFS_NICENUM_1024;
/* All these histos are the same size, so just use nva[0].count */
buckets = nva[0].count;
if (cb->cb_flags & IOS_RQ_HISTO_M) {
/* Start at 512 - req size should never be lower than this */
start_bucket = 9;
} else {
start_bucket = 0;
}
for (j = start_bucket; j < buckets; j++) {
/* Print histogram bucket label */
if (cb->cb_flags & IOS_L_HISTO_M) {
/* Ending range of this bucket */
val = (1UL << (j + 1)) - 1;
zfs_nicetime(val, buf, sizeof (buf));
} else {
/* Request size (starting range of bucket) */
val = (1UL << j);
zfs_nicenum(val, buf, sizeof (buf));
}
if (cb->cb_scripted)
printf("%llu", (u_longlong_t)val);
else
printf("%-*s", namewidth, buf);
/* Print the values on the line */
for (i = 0; i < len; i++) {
print_one_stat(nva[i].data[j] * scale, format,
column_width, cb->cb_scripted);
}
printf("\n");
}
}
static void
print_solid_separator(unsigned int length)
{
while (length--)
printf("-");
printf("\n");
}
static void
print_iostat_histos(iostat_cbdata_t *cb, nvlist_t *oldnv,
nvlist_t *newnv, double scale, const char *name)
{
unsigned int column_width;
unsigned int namewidth;
unsigned int entire_width;
enum iostat_type type;
struct stat_array *nva;
const char **names;
unsigned int names_len;
/* What type of histo are we? */
type = IOS_HISTO_IDX(cb->cb_flags);
/* Get NULL-terminated array of nvlist names for our histo */
names = vsx_type_to_nvlist[type];
names_len = str_array_len(names); /* num of names */
nva = calc_and_alloc_stats_ex(names, names_len, oldnv, newnv);
if (cb->cb_literal) {
column_width = MAX(5,
(unsigned int) log10(stat_histo_max(nva, names_len)) + 1);
} else {
column_width = 5;
}
namewidth = MAX(cb->cb_namewidth,
strlen(histo_to_title[IOS_HISTO_IDX(cb->cb_flags)]));
/*
* Calculate the entire line width of what we're printing. The
* +2 is for the two spaces between columns:
*/
/* read write */
/* ----- ----- */
/* |___| <---------- column_width */
/* */
/* |__________| <--- entire_width */
/* */
entire_width = namewidth + (column_width + 2) *
label_array_len(iostat_bottom_labels[type]);
if (cb->cb_scripted)
printf("%s\n", name);
else
print_iostat_header_impl(cb, column_width, name);
print_iostat_histo(nva, names_len, cb, column_width,
namewidth, scale);
free_calc_stats(nva, names_len);
if (!cb->cb_scripted)
print_solid_separator(entire_width);
}
/*
* Calculate the average latency of a power-of-two latency histogram
*/
static uint64_t
single_histo_average(uint64_t *histo, unsigned int buckets)
{
int i;
uint64_t count = 0, total = 0;
for (i = 0; i < buckets; i++) {
/*
* Our buckets are power-of-two latency ranges. Use the
* midpoint latency of each bucket to calculate the average.
* For example:
*
* Bucket Midpoint
* 8ns-15ns: 12ns
* 16ns-31ns: 24ns
* ...
*/
if (histo[i] != 0) {
total += histo[i] * (((1UL << i) + ((1UL << i)/2)));
count += histo[i];
}
}
/* Prevent divide by zero */
return (count == 0 ? 0 : total / count);
}
static void
print_iostat_queues(iostat_cbdata_t *cb, nvlist_t *newnv)
{
const char *names[] = {
ZPOOL_CONFIG_VDEV_SYNC_R_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_SYNC_W_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_R_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_W_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_SCRUB_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_TRIM_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_REBUILD_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE,
};
struct stat_array *nva;
unsigned int column_width = default_column_width(cb, IOS_QUEUES);
enum zfs_nicenum_format format;
nva = calc_and_alloc_stats_ex(names, ARRAY_SIZE(names), NULL, newnv);
if (cb->cb_literal)
format = ZFS_NICENUM_RAW;
else
format = ZFS_NICENUM_1024;
for (int i = 0; i < ARRAY_SIZE(names); i++) {
uint64_t val = nva[i].data[0];
print_one_stat(val, format, column_width, cb->cb_scripted);
}
free_calc_stats(nva, ARRAY_SIZE(names));
}
static void
print_iostat_latency(iostat_cbdata_t *cb, nvlist_t *oldnv,
nvlist_t *newnv)
{
int i;
uint64_t val;
const char *names[] = {
ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO,
ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO,
};
struct stat_array *nva;
unsigned int column_width = default_column_width(cb, IOS_LATENCY);
enum zfs_nicenum_format format;
nva = calc_and_alloc_stats_ex(names, ARRAY_SIZE(names), oldnv, newnv);
if (cb->cb_literal)
format = ZFS_NICENUM_RAWTIME;
else
format = ZFS_NICENUM_TIME;
/* Print our avg latencies on the line */
for (i = 0; i < ARRAY_SIZE(names); i++) {
/* Compute average latency for a latency histo */
val = single_histo_average(nva[i].data, nva[i].count);
print_one_stat(val, format, column_width, cb->cb_scripted);
}
free_calc_stats(nva, ARRAY_SIZE(names));
}
/*
* Print default statistics (capacity/operations/bandwidth)
*/
static void
print_iostat_default(vdev_stat_t *vs, iostat_cbdata_t *cb, double scale)
{
unsigned int column_width = default_column_width(cb, IOS_DEFAULT);
enum zfs_nicenum_format format;
char na; /* char to print for "not applicable" values */
if (cb->cb_literal) {
format = ZFS_NICENUM_RAW;
na = '0';
} else {
format = ZFS_NICENUM_1024;
na = '-';
}
/* only toplevel vdevs have capacity stats */
if (vs->vs_space == 0) {
if (cb->cb_scripted)
printf("\t%c\t%c", na, na);
else
printf(" %*c %*c", column_width, na, column_width,
na);
} else {
print_one_stat(vs->vs_alloc, format, column_width,
cb->cb_scripted);
print_one_stat(vs->vs_space - vs->vs_alloc, format,
column_width, cb->cb_scripted);
}
print_one_stat((uint64_t)(vs->vs_ops[ZIO_TYPE_READ] * scale),
format, column_width, cb->cb_scripted);
print_one_stat((uint64_t)(vs->vs_ops[ZIO_TYPE_WRITE] * scale),
format, column_width, cb->cb_scripted);
print_one_stat((uint64_t)(vs->vs_bytes[ZIO_TYPE_READ] * scale),
format, column_width, cb->cb_scripted);
print_one_stat((uint64_t)(vs->vs_bytes[ZIO_TYPE_WRITE] * scale),
format, column_width, cb->cb_scripted);
}
static const char *const class_name[] = {
VDEV_ALLOC_BIAS_DEDUP,
VDEV_ALLOC_BIAS_SPECIAL,
VDEV_ALLOC_CLASS_LOGS
};
/*
* Print out all the statistics for the given vdev. This can either be the
* toplevel configuration, or called recursively. If 'name' is NULL, then this
* is a verbose output, and we don't want to display the toplevel pool stats.
*
* Returns the number of stat lines printed.
*/
static unsigned int
print_vdev_stats(zpool_handle_t *zhp, const char *name, nvlist_t *oldnv,
nvlist_t *newnv, iostat_cbdata_t *cb, int depth)
{
nvlist_t **oldchild, **newchild;
uint_t c, children, oldchildren;
vdev_stat_t *oldvs, *newvs, *calcvs;
vdev_stat_t zerovs = { 0 };
char *vname;
int i;
int ret = 0;
uint64_t tdelta;
double scale;
if (strcmp(name, VDEV_TYPE_INDIRECT) == 0)
return (ret);
calcvs = safe_malloc(sizeof (*calcvs));
if (oldnv != NULL) {
verify(nvlist_lookup_uint64_array(oldnv,
ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&oldvs, &c) == 0);
} else {
oldvs = &zerovs;
}
/* Do we only want to see a specific vdev? */
for (i = 0; i < cb->cb_vdevs.cb_names_count; i++) {
/* Yes we do. Is this the vdev? */
if (strcmp(name, cb->cb_vdevs.cb_names[i]) == 0) {
/*
* This is our vdev. Since it is the only vdev we
* will be displaying, make depth = 0 so that it
* doesn't get indented.
*/
depth = 0;
break;
}
}
if (cb->cb_vdevs.cb_names_count && (i == cb->cb_vdevs.cb_names_count)) {
/* Couldn't match the name */
goto children;
}
verify(nvlist_lookup_uint64_array(newnv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&newvs, &c) == 0);
/*
* Print the vdev name unless it's is a histogram. Histograms
* display the vdev name in the header itself.
*/
if (!(cb->cb_flags & IOS_ANYHISTO_M)) {
if (cb->cb_scripted) {
printf("%s", name);
} else {
if (strlen(name) + depth > cb->cb_namewidth)
(void) printf("%*s%s", depth, "", name);
else
(void) printf("%*s%s%*s", depth, "", name,
(int)(cb->cb_namewidth - strlen(name) -
depth), "");
}
}
/* Calculate our scaling factor */
tdelta = newvs->vs_timestamp - oldvs->vs_timestamp;
if ((oldvs->vs_timestamp == 0) && (cb->cb_flags & IOS_ANYHISTO_M)) {
/*
* If we specify printing histograms with no time interval, then
* print the histogram numbers over the entire lifetime of the
* vdev.
*/
scale = 1;
} else {
if (tdelta == 0)
scale = 1.0;
else
scale = (double)NANOSEC / tdelta;
}
if (cb->cb_flags & IOS_DEFAULT_M) {
calc_default_iostats(oldvs, newvs, calcvs);
print_iostat_default(calcvs, cb, scale);
}
if (cb->cb_flags & IOS_LATENCY_M)
print_iostat_latency(cb, oldnv, newnv);
if (cb->cb_flags & IOS_QUEUES_M)
print_iostat_queues(cb, newnv);
if (cb->cb_flags & IOS_ANYHISTO_M) {
printf("\n");
print_iostat_histos(cb, oldnv, newnv, scale, name);
}
if (cb->vcdl != NULL) {
const char *path;
if (nvlist_lookup_string(newnv, ZPOOL_CONFIG_PATH,
&path) == 0) {
printf(" ");
zpool_print_cmd(cb->vcdl, zpool_get_name(zhp), path);
}
}
if (!(cb->cb_flags & IOS_ANYHISTO_M))
printf("\n");
ret++;
children:
free(calcvs);
if (!cb->cb_verbose)
return (ret);
if (nvlist_lookup_nvlist_array(newnv, ZPOOL_CONFIG_CHILDREN,
&newchild, &children) != 0)
return (ret);
if (oldnv) {
if (nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_CHILDREN,
&oldchild, &oldchildren) != 0)
return (ret);
children = MIN(oldchildren, children);
}
/*
* print normal top-level devices
*/
for (c = 0; c < children; c++) {
uint64_t ishole = B_FALSE, islog = B_FALSE;
(void) nvlist_lookup_uint64(newchild[c], ZPOOL_CONFIG_IS_HOLE,
&ishole);
(void) nvlist_lookup_uint64(newchild[c], ZPOOL_CONFIG_IS_LOG,
&islog);
if (ishole || islog)
continue;
if (nvlist_exists(newchild[c], ZPOOL_CONFIG_ALLOCATION_BIAS))
continue;
vname = zpool_vdev_name(g_zfs, zhp, newchild[c],
cb->cb_vdevs.cb_name_flags | VDEV_NAME_TYPE_ID);
ret += print_vdev_stats(zhp, vname, oldnv ? oldchild[c] : NULL,
newchild[c], cb, depth + 2);
free(vname);
}
/*
* print all other top-level devices
*/
for (uint_t n = 0; n < ARRAY_SIZE(class_name); n++) {
boolean_t printed = B_FALSE;
for (c = 0; c < children; c++) {
uint64_t islog = B_FALSE;
const char *bias = NULL;
const char *type = NULL;
(void) nvlist_lookup_uint64(newchild[c],
ZPOOL_CONFIG_IS_LOG, &islog);
if (islog) {
bias = VDEV_ALLOC_CLASS_LOGS;
} else {
(void) nvlist_lookup_string(newchild[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &bias);
(void) nvlist_lookup_string(newchild[c],
ZPOOL_CONFIG_TYPE, &type);
}
if (bias == NULL || strcmp(bias, class_name[n]) != 0)
continue;
if (!islog && strcmp(type, VDEV_TYPE_INDIRECT) == 0)
continue;
if (!printed) {
if ((!(cb->cb_flags & IOS_ANYHISTO_M)) &&
!cb->cb_scripted &&
!cb->cb_vdevs.cb_names) {
print_iostat_dashes(cb, 0,
class_name[n]);
}
printf("\n");
printed = B_TRUE;
}
vname = zpool_vdev_name(g_zfs, zhp, newchild[c],
cb->cb_vdevs.cb_name_flags | VDEV_NAME_TYPE_ID);
ret += print_vdev_stats(zhp, vname, oldnv ?
oldchild[c] : NULL, newchild[c], cb, depth + 2);
free(vname);
}
}
/*
* Include level 2 ARC devices in iostat output
*/
if (nvlist_lookup_nvlist_array(newnv, ZPOOL_CONFIG_L2CACHE,
&newchild, &children) != 0)
return (ret);
if (oldnv) {
if (nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_L2CACHE,
&oldchild, &oldchildren) != 0)
return (ret);
children = MIN(oldchildren, children);
}
if (children > 0) {
if ((!(cb->cb_flags & IOS_ANYHISTO_M)) && !cb->cb_scripted &&
!cb->cb_vdevs.cb_names) {
print_iostat_dashes(cb, 0, "cache");
}
printf("\n");
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(g_zfs, zhp, newchild[c],
cb->cb_vdevs.cb_name_flags);
ret += print_vdev_stats(zhp, vname, oldnv ? oldchild[c]
: NULL, newchild[c], cb, depth + 2);
free(vname);
}
}
return (ret);
}
static int
refresh_iostat(zpool_handle_t *zhp, void *data)
{
iostat_cbdata_t *cb = data;
boolean_t missing;
/*
* If the pool has disappeared, remove it from the list and continue.
*/
if (zpool_refresh_stats(zhp, &missing) != 0)
return (-1);
if (missing)
pool_list_remove(cb->cb_list, zhp);
return (0);
}
/*
* Callback to print out the iostats for the given pool.
*/
static int
print_iostat(zpool_handle_t *zhp, void *data)
{
iostat_cbdata_t *cb = data;
nvlist_t *oldconfig, *newconfig;
nvlist_t *oldnvroot, *newnvroot;
int ret;
newconfig = zpool_get_config(zhp, &oldconfig);
if (cb->cb_iteration == 1)
oldconfig = NULL;
verify(nvlist_lookup_nvlist(newconfig, ZPOOL_CONFIG_VDEV_TREE,
&newnvroot) == 0);
if (oldconfig == NULL)
oldnvroot = NULL;
else
verify(nvlist_lookup_nvlist(oldconfig, ZPOOL_CONFIG_VDEV_TREE,
&oldnvroot) == 0);
ret = print_vdev_stats(zhp, zpool_get_name(zhp), oldnvroot, newnvroot,
cb, 0);
if ((ret != 0) && !(cb->cb_flags & IOS_ANYHISTO_M) &&
!cb->cb_scripted && cb->cb_verbose &&
!cb->cb_vdevs.cb_names_count) {
print_iostat_separator(cb);
if (cb->vcdl != NULL) {
print_cmd_columns(cb->vcdl, 1);
}
printf("\n");
}
return (ret);
}
static int
get_columns(void)
{
struct winsize ws;
int columns = 80;
int error;
if (isatty(STDOUT_FILENO)) {
error = ioctl(STDOUT_FILENO, TIOCGWINSZ, &ws);
if (error == 0)
columns = ws.ws_col;
} else {
columns = 999;
}
return (columns);
}
/*
* Return the required length of the pool/vdev name column. The minimum
* allowed width and output formatting flags must be provided.
*/
static int
get_namewidth(zpool_handle_t *zhp, int min_width, int flags, boolean_t verbose)
{
nvlist_t *config, *nvroot;
int width = min_width;
if ((config = zpool_get_config(zhp, NULL)) != NULL) {
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
size_t poolname_len = strlen(zpool_get_name(zhp));
if (verbose == B_FALSE) {
width = MAX(poolname_len, min_width);
} else {
width = MAX(poolname_len,
max_width(zhp, nvroot, 0, min_width, flags));
}
}
return (width);
}
/*
* Parse the input string, get the 'interval' and 'count' value if there is one.
*/
static void
get_interval_count(int *argcp, char **argv, float *iv,
unsigned long *cnt)
{
float interval = 0;
unsigned long count = 0;
int argc = *argcp;
/*
* Determine if the last argument is an integer or a pool name
*/
if (argc > 0 && zfs_isnumber(argv[argc - 1])) {
char *end;
errno = 0;
interval = strtof(argv[argc - 1], &end);
if (*end == '\0' && errno == 0) {
if (interval == 0) {
(void) fprintf(stderr, gettext(
"interval cannot be zero\n"));
usage(B_FALSE);
}
/*
* Ignore the last parameter
*/
argc--;
} else {
/*
* If this is not a valid number, just plow on. The
* user will get a more informative error message later
* on.
*/
interval = 0;
}
}
/*
* If the last argument is also an integer, then we have both a count
* and an interval.
*/
if (argc > 0 && zfs_isnumber(argv[argc - 1])) {
char *end;
errno = 0;
count = interval;
interval = strtof(argv[argc - 1], &end);
if (*end == '\0' && errno == 0) {
if (interval == 0) {
(void) fprintf(stderr, gettext(
"interval cannot be zero\n"));
usage(B_FALSE);
}
/*
* Ignore the last parameter
*/
argc--;
} else {
interval = 0;
}
}
*iv = interval;
*cnt = count;
*argcp = argc;
}
static void
get_timestamp_arg(char c)
{
if (c == 'u')
timestamp_fmt = UDATE;
else if (c == 'd')
timestamp_fmt = DDATE;
else
usage(B_FALSE);
}
/*
* Return stat flags that are supported by all pools by both the module and
* zpool iostat. "*data" should be initialized to all 0xFFs before running.
* It will get ANDed down until only the flags that are supported on all pools
* remain.
*/
static int
get_stat_flags_cb(zpool_handle_t *zhp, void *data)
{
uint64_t *mask = data;
nvlist_t *config, *nvroot, *nvx;
uint64_t flags = 0;
int i, j;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
/* Default stats are always supported, but for completeness.. */
if (nvlist_exists(nvroot, ZPOOL_CONFIG_VDEV_STATS))
flags |= IOS_DEFAULT_M;
/* Get our extended stats nvlist from the main list */
if (nvlist_lookup_nvlist(nvroot, ZPOOL_CONFIG_VDEV_STATS_EX,
&nvx) != 0) {
/*
* No extended stats; they're probably running an older
* module. No big deal, we support that too.
*/
goto end;
}
/* For each extended stat, make sure all its nvpairs are supported */
for (j = 0; j < ARRAY_SIZE(vsx_type_to_nvlist); j++) {
if (!vsx_type_to_nvlist[j][0])
continue;
/* Start off by assuming the flag is supported, then check */
flags |= (1ULL << j);
for (i = 0; vsx_type_to_nvlist[j][i]; i++) {
if (!nvlist_exists(nvx, vsx_type_to_nvlist[j][i])) {
/* flag isn't supported */
flags = flags & ~(1ULL << j);
break;
}
}
}
end:
*mask = *mask & flags;
return (0);
}
/*
* Return a bitmask of stats that are supported on all pools by both the module
* and zpool iostat.
*/
static uint64_t
get_stat_flags(zpool_list_t *list)
{
uint64_t mask = -1;
/*
* get_stat_flags_cb() will lop off bits from "mask" until only the
* flags that are supported on all pools remain.
*/
pool_list_iter(list, B_FALSE, get_stat_flags_cb, &mask);
return (mask);
}
/*
* Return 1 if cb_data->cb_names[0] is this vdev's name, 0 otherwise.
*/
static int
is_vdev_cb(void *zhp_data, nvlist_t *nv, void *cb_data)
{
uint64_t guid;
vdev_cbdata_t *cb = cb_data;
zpool_handle_t *zhp = zhp_data;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0)
return (0);
return (guid == zpool_vdev_path_to_guid(zhp, cb->cb_names[0]));
}
/*
* Returns 1 if cb_data->cb_names[0] is a vdev name, 0 otherwise.
*/
static int
is_vdev(zpool_handle_t *zhp, void *cb_data)
{
return (for_each_vdev(zhp, is_vdev_cb, cb_data));
}
/*
* Check if vdevs are in a pool
*
* Return 1 if all argv[] strings are vdev names in pool "pool_name". Otherwise
* return 0. If pool_name is NULL, then search all pools.
*/
static int
are_vdevs_in_pool(int argc, char **argv, char *pool_name,
vdev_cbdata_t *cb)
{
char **tmp_name;
int ret = 0;
int i;
int pool_count = 0;
if ((argc == 0) || !*argv)
return (0);
if (pool_name)
pool_count = 1;
/* Temporarily hijack cb_names for a second... */
tmp_name = cb->cb_names;
/* Go though our list of prospective vdev names */
for (i = 0; i < argc; i++) {
cb->cb_names = argv + i;
/* Is this name a vdev in our pools? */
ret = for_each_pool(pool_count, &pool_name, B_TRUE, NULL,
ZFS_TYPE_POOL, B_FALSE, is_vdev, cb);
if (!ret) {
/* No match */
break;
}
}
cb->cb_names = tmp_name;
return (ret);
}
static int
is_pool_cb(zpool_handle_t *zhp, void *data)
{
char *name = data;
if (strcmp(name, zpool_get_name(zhp)) == 0)
return (1);
return (0);
}
/*
* Do we have a pool named *name? If so, return 1, otherwise 0.
*/
static int
is_pool(char *name)
{
return (for_each_pool(0, NULL, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE,
is_pool_cb, name));
}
/* Are all our argv[] strings pool names? If so return 1, 0 otherwise. */
static int
are_all_pools(int argc, char **argv)
{
if ((argc == 0) || !*argv)
return (0);
while (--argc >= 0)
if (!is_pool(argv[argc]))
return (0);
return (1);
}
/*
* Helper function to print out vdev/pool names we can't resolve. Used for an
* error message.
*/
static void
error_list_unresolved_vdevs(int argc, char **argv, char *pool_name,
vdev_cbdata_t *cb)
{
int i;
char *name;
char *str;
for (i = 0; i < argc; i++) {
name = argv[i];
if (is_pool(name))
str = gettext("pool");
else if (are_vdevs_in_pool(1, &name, pool_name, cb))
str = gettext("vdev in this pool");
else if (are_vdevs_in_pool(1, &name, NULL, cb))
str = gettext("vdev in another pool");
else
str = gettext("unknown");
fprintf(stderr, "\t%s (%s)\n", name, str);
}
}
/*
* Same as get_interval_count(), but with additional checks to not misinterpret
* guids as interval/count values. Assumes VDEV_NAME_GUID is set in
* cb.cb_vdevs.cb_name_flags.
*/
static void
get_interval_count_filter_guids(int *argc, char **argv, float *interval,
unsigned long *count, iostat_cbdata_t *cb)
{
char **tmpargv = argv;
int argc_for_interval = 0;
/* Is the last arg an interval value? Or a guid? */
if (*argc >= 1 && !are_vdevs_in_pool(1, &argv[*argc - 1], NULL,
&cb->cb_vdevs)) {
/*
* The last arg is not a guid, so it's probably an
* interval value.
*/
argc_for_interval++;
if (*argc >= 2 &&
!are_vdevs_in_pool(1, &argv[*argc - 2], NULL,
&cb->cb_vdevs)) {
/*
* The 2nd to last arg is not a guid, so it's probably
* an interval value.
*/
argc_for_interval++;
}
}
/* Point to our list of possible intervals */
tmpargv = &argv[*argc - argc_for_interval];
*argc = *argc - argc_for_interval;
get_interval_count(&argc_for_interval, tmpargv,
interval, count);
}
/*
* Floating point sleep(). Allows you to pass in a floating point value for
* seconds.
*/
static void
fsleep(float sec)
{
struct timespec req;
req.tv_sec = floor(sec);
req.tv_nsec = (sec - (float)req.tv_sec) * NANOSEC;
nanosleep(&req, NULL);
}
/*
* Terminal height, in rows. Returns -1 if stdout is not connected to a TTY or
* if we were unable to determine its size.
*/
static int
terminal_height(void)
{
struct winsize win;
if (isatty(STDOUT_FILENO) == 0)
return (-1);
if (ioctl(STDOUT_FILENO, TIOCGWINSZ, &win) != -1 && win.ws_row > 0)
return (win.ws_row);
return (-1);
}
/*
* Run one of the zpool status/iostat -c scripts with the help (-h) option and
* print the result.
*
* name: Short name of the script ('iostat').
* path: Full path to the script ('/usr/local/etc/zfs/zpool.d/iostat');
*/
static void
print_zpool_script_help(char *name, char *path)
{
char *argv[] = {path, (char *)"-h", NULL};
char **lines = NULL;
int lines_cnt = 0;
int rc;
rc = libzfs_run_process_get_stdout_nopath(path, argv, NULL, &lines,
&lines_cnt);
if (rc != 0 || lines == NULL || lines_cnt <= 0) {
if (lines != NULL)
libzfs_free_str_array(lines, lines_cnt);
return;
}
for (int i = 0; i < lines_cnt; i++)
if (!is_blank_str(lines[i]))
printf(" %-14s %s\n", name, lines[i]);
libzfs_free_str_array(lines, lines_cnt);
}
/*
* Go though the zpool status/iostat -c scripts in the user's path, run their
* help option (-h), and print out the results.
*/
static void
print_zpool_dir_scripts(char *dirpath)
{
DIR *dir;
struct dirent *ent;
char fullpath[MAXPATHLEN];
struct stat dir_stat;
if ((dir = opendir(dirpath)) != NULL) {
/* print all the files and directories within directory */
while ((ent = readdir(dir)) != NULL) {
if (snprintf(fullpath, sizeof (fullpath), "%s/%s",
dirpath, ent->d_name) >= sizeof (fullpath)) {
(void) fprintf(stderr,
gettext("internal error: "
"ZPOOL_SCRIPTS_PATH too large.\n"));
exit(1);
}
/* Print the scripts */
if (stat(fullpath, &dir_stat) == 0)
if (dir_stat.st_mode & S_IXUSR &&
S_ISREG(dir_stat.st_mode))
print_zpool_script_help(ent->d_name,
fullpath);
}
closedir(dir);
}
}
/*
* Print out help text for all zpool status/iostat -c scripts.
*/
static void
print_zpool_script_list(const char *subcommand)
{
char *dir, *sp, *tmp;
printf(gettext("Available 'zpool %s -c' commands:\n"), subcommand);
sp = zpool_get_cmd_search_path();
if (sp == NULL)
return;
for (dir = strtok_r(sp, ":", &tmp);
dir != NULL;
dir = strtok_r(NULL, ":", &tmp))
print_zpool_dir_scripts(dir);
free(sp);
}
/*
* Set the minimum pool/vdev name column width. The width must be at least 10,
* but may be as large as the column width - 42 so it still fits on one line.
* NOTE: 42 is the width of the default capacity/operations/bandwidth output
*/
static int
get_namewidth_iostat(zpool_handle_t *zhp, void *data)
{
iostat_cbdata_t *cb = data;
int width, available_width;
/*
* get_namewidth() returns the maximum width of any name in that column
* for any pool/vdev/device line that will be output.
*/
width = get_namewidth(zhp, cb->cb_namewidth,
cb->cb_vdevs.cb_name_flags | VDEV_NAME_TYPE_ID, cb->cb_verbose);
/*
* The width we are calculating is the width of the header and also the
* padding width for names that are less than maximum width. The stats
* take up 42 characters, so the width available for names is:
*/
available_width = get_columns() - 42;
/*
* If the maximum width fits on a screen, then great! Make everything
* line up by justifying all lines to the same width. If that max
* width is larger than what's available, the name plus stats won't fit
* on one line, and justifying to that width would cause every line to
* wrap on the screen. We only want lines with long names to wrap.
* Limit the padding to what won't wrap.
*/
if (width > available_width)
width = available_width;
/*
* And regardless of whatever the screen width is (get_columns can
* return 0 if the width is not known or less than 42 for a narrow
* terminal) have the width be a minimum of 10.
*/
if (width < 10)
width = 10;
/* Save the calculated width */
cb->cb_namewidth = width;
return (0);
}
/*
* zpool iostat [[-c [script1,script2,...]] [-lq]|[-rw]] [-ghHLpPvy] [-n name]
* [-T d|u] [[ pool ...]|[pool vdev ...]|[vdev ...]]
* [interval [count]]
*
* -c CMD For each vdev, run command CMD
* -g Display guid for individual vdev name.
* -L Follow links when resolving vdev path name.
* -P Display full path for vdev name.
* -v Display statistics for individual vdevs
* -h Display help
* -p Display values in parsable (exact) format.
* -H Scripted mode. Don't display headers, and separate properties
* by a single tab.
* -l Display average latency
* -q Display queue depths
* -w Display latency histograms
* -r Display request size histogram
* -T Display a timestamp in date(1) or Unix format
* -n Only print headers once
*
* This command can be tricky because we want to be able to deal with pool
* creation/destruction as well as vdev configuration changes. The bulk of this
* processing is handled by the pool_list_* routines in zpool_iter.c. We rely
* on pool_list_update() to detect the addition of new pools. Configuration
* changes are all handled within libzfs.
*/
int
zpool_do_iostat(int argc, char **argv)
{
int c;
int ret;
int npools;
float interval = 0;
unsigned long count = 0;
int winheight = 24;
zpool_list_t *list;
boolean_t verbose = B_FALSE;
boolean_t latency = B_FALSE, l_histo = B_FALSE, rq_histo = B_FALSE;
boolean_t queues = B_FALSE, parsable = B_FALSE, scripted = B_FALSE;
boolean_t omit_since_boot = B_FALSE;
boolean_t guid = B_FALSE;
boolean_t follow_links = B_FALSE;
boolean_t full_name = B_FALSE;
boolean_t headers_once = B_FALSE;
iostat_cbdata_t cb = { 0 };
char *cmd = NULL;
/* Used for printing error message */
const char flag_to_arg[] = {[IOS_LATENCY] = 'l', [IOS_QUEUES] = 'q',
[IOS_L_HISTO] = 'w', [IOS_RQ_HISTO] = 'r'};
uint64_t unsupported_flags;
/* check options */
while ((c = getopt(argc, argv, "c:gLPT:vyhplqrwnH")) != -1) {
switch (c) {
case 'c':
if (cmd != NULL) {
fprintf(stderr,
gettext("Can't set -c flag twice\n"));
exit(1);
}
if (getenv("ZPOOL_SCRIPTS_ENABLED") != NULL &&
!libzfs_envvar_is_set("ZPOOL_SCRIPTS_ENABLED")) {
fprintf(stderr, gettext(
"Can't run -c, disabled by "
"ZPOOL_SCRIPTS_ENABLED.\n"));
exit(1);
}
if ((getuid() <= 0 || geteuid() <= 0) &&
!libzfs_envvar_is_set("ZPOOL_SCRIPTS_AS_ROOT")) {
fprintf(stderr, gettext(
"Can't run -c with root privileges "
"unless ZPOOL_SCRIPTS_AS_ROOT is set.\n"));
exit(1);
}
cmd = optarg;
verbose = B_TRUE;
break;
case 'g':
guid = B_TRUE;
break;
case 'L':
follow_links = B_TRUE;
break;
case 'P':
full_name = B_TRUE;
break;
case 'T':
get_timestamp_arg(*optarg);
break;
case 'v':
verbose = B_TRUE;
break;
case 'p':
parsable = B_TRUE;
break;
case 'l':
latency = B_TRUE;
break;
case 'q':
queues = B_TRUE;
break;
case 'H':
scripted = B_TRUE;
break;
case 'w':
l_histo = B_TRUE;
break;
case 'r':
rq_histo = B_TRUE;
break;
case 'y':
omit_since_boot = B_TRUE;
break;
case 'n':
headers_once = B_TRUE;
break;
case 'h':
usage(B_FALSE);
break;
case '?':
if (optopt == 'c') {
print_zpool_script_list("iostat");
exit(0);
} else {
fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
}
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
cb.cb_literal = parsable;
cb.cb_scripted = scripted;
if (guid)
cb.cb_vdevs.cb_name_flags |= VDEV_NAME_GUID;
if (follow_links)
cb.cb_vdevs.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS;
if (full_name)
cb.cb_vdevs.cb_name_flags |= VDEV_NAME_PATH;
cb.cb_iteration = 0;
cb.cb_namewidth = 0;
cb.cb_verbose = verbose;
/* Get our interval and count values (if any) */
if (guid) {
get_interval_count_filter_guids(&argc, argv, &interval,
&count, &cb);
} else {
get_interval_count(&argc, argv, &interval, &count);
}
if (argc == 0) {
/* No args, so just print the defaults. */
} else if (are_all_pools(argc, argv)) {
/* All the args are pool names */
} else if (are_vdevs_in_pool(argc, argv, NULL, &cb.cb_vdevs)) {
/* All the args are vdevs */
cb.cb_vdevs.cb_names = argv;
cb.cb_vdevs.cb_names_count = argc;
argc = 0; /* No pools to process */
} else if (are_all_pools(1, argv)) {
/* The first arg is a pool name */
if (are_vdevs_in_pool(argc - 1, argv + 1, argv[0],
&cb.cb_vdevs)) {
/* ...and the rest are vdev names */
cb.cb_vdevs.cb_names = argv + 1;
cb.cb_vdevs.cb_names_count = argc - 1;
argc = 1; /* One pool to process */
} else {
fprintf(stderr, gettext("Expected either a list of "));
fprintf(stderr, gettext("pools, or list of vdevs in"));
fprintf(stderr, " \"%s\", ", argv[0]);
fprintf(stderr, gettext("but got:\n"));
error_list_unresolved_vdevs(argc - 1, argv + 1,
argv[0], &cb.cb_vdevs);
fprintf(stderr, "\n");
usage(B_FALSE);
return (1);
}
} else {
/*
* The args don't make sense. The first arg isn't a pool name,
* nor are all the args vdevs.
*/
fprintf(stderr, gettext("Unable to parse pools/vdevs list.\n"));
fprintf(stderr, "\n");
return (1);
}
if (cb.cb_vdevs.cb_names_count != 0) {
/*
* If user specified vdevs, it implies verbose.
*/
cb.cb_verbose = B_TRUE;
}
/*
* Construct the list of all interesting pools.
*/
ret = 0;
if ((list = pool_list_get(argc, argv, NULL, ZFS_TYPE_POOL, parsable,
&ret)) == NULL)
return (1);
if (pool_list_count(list) == 0 && argc != 0) {
pool_list_free(list);
return (1);
}
if (pool_list_count(list) == 0 && interval == 0) {
pool_list_free(list);
(void) fprintf(stderr, gettext("no pools available\n"));
return (1);
}
if ((l_histo || rq_histo) && (cmd != NULL || latency || queues)) {
pool_list_free(list);
(void) fprintf(stderr,
gettext("[-r|-w] isn't allowed with [-c|-l|-q]\n"));
usage(B_FALSE);
return (1);
}
if (l_histo && rq_histo) {
pool_list_free(list);
(void) fprintf(stderr,
gettext("Only one of [-r|-w] can be passed at a time\n"));
usage(B_FALSE);
return (1);
}
/*
* Enter the main iostat loop.
*/
cb.cb_list = list;
if (l_histo) {
/*
* Histograms tables look out of place when you try to display
* them with the other stats, so make a rule that you can only
* print histograms by themselves.
*/
cb.cb_flags = IOS_L_HISTO_M;
} else if (rq_histo) {
cb.cb_flags = IOS_RQ_HISTO_M;
} else {
cb.cb_flags = IOS_DEFAULT_M;
if (latency)
cb.cb_flags |= IOS_LATENCY_M;
if (queues)
cb.cb_flags |= IOS_QUEUES_M;
}
/*
* See if the module supports all the stats we want to display.
*/
unsupported_flags = cb.cb_flags & ~get_stat_flags(list);
if (unsupported_flags) {
uint64_t f;
int idx;
fprintf(stderr,
gettext("The loaded zfs module doesn't support:"));
/* for each bit set in unsupported_flags */
for (f = unsupported_flags; f; f &= ~(1ULL << idx)) {
idx = lowbit64(f) - 1;
fprintf(stderr, " -%c", flag_to_arg[idx]);
}
fprintf(stderr, ". Try running a newer module.\n");
pool_list_free(list);
return (1);
}
for (;;) {
if ((npools = pool_list_count(list)) == 0)
(void) fprintf(stderr, gettext("no pools available\n"));
else {
/*
* If this is the first iteration and -y was supplied
* we skip any printing.
*/
boolean_t skip = (omit_since_boot &&
cb.cb_iteration == 0);
/*
* Refresh all statistics. This is done as an
* explicit step before calculating the maximum name
* width, so that any * configuration changes are
* properly accounted for.
*/
(void) pool_list_iter(list, B_FALSE, refresh_iostat,
&cb);
/*
* Iterate over all pools to determine the maximum width
* for the pool / device name column across all pools.
*/
cb.cb_namewidth = 0;
(void) pool_list_iter(list, B_FALSE,
get_namewidth_iostat, &cb);
if (timestamp_fmt != NODATE)
print_timestamp(timestamp_fmt);
if (cmd != NULL && cb.cb_verbose &&
!(cb.cb_flags & IOS_ANYHISTO_M)) {
cb.vcdl = all_pools_for_each_vdev_run(argc,
argv, cmd, g_zfs, cb.cb_vdevs.cb_names,
cb.cb_vdevs.cb_names_count,
cb.cb_vdevs.cb_name_flags);
} else {
cb.vcdl = NULL;
}
/*
* Check terminal size so we can print headers
* even when terminal window has its height
* changed.
*/
winheight = terminal_height();
/*
* Are we connected to TTY? If not, headers_once
* should be true, to avoid breaking scripts.
*/
if (winheight < 0)
headers_once = B_TRUE;
/*
* If it's the first time and we're not skipping it,
* or either skip or verbose mode, print the header.
*
* The histogram code explicitly prints its header on
* every vdev, so skip this for histograms.
*/
if (((++cb.cb_iteration == 1 && !skip) ||
(skip != verbose) ||
(!headers_once &&
(cb.cb_iteration % winheight) == 0)) &&
(!(cb.cb_flags & IOS_ANYHISTO_M)) &&
!cb.cb_scripted)
print_iostat_header(&cb);
if (skip) {
(void) fsleep(interval);
continue;
}
pool_list_iter(list, B_FALSE, print_iostat, &cb);
/*
* If there's more than one pool, and we're not in
* verbose mode (which prints a separator for us),
* then print a separator.
*
* In addition, if we're printing specific vdevs then
* we also want an ending separator.
*/
if (((npools > 1 && !verbose &&
!(cb.cb_flags & IOS_ANYHISTO_M)) ||
(!(cb.cb_flags & IOS_ANYHISTO_M) &&
cb.cb_vdevs.cb_names_count)) &&
!cb.cb_scripted) {
print_iostat_separator(&cb);
if (cb.vcdl != NULL)
print_cmd_columns(cb.vcdl, 1);
printf("\n");
}
if (cb.vcdl != NULL)
free_vdev_cmd_data_list(cb.vcdl);
}
/*
* Flush the output so that redirection to a file isn't buffered
* indefinitely.
*/
(void) fflush(stdout);
if (interval == 0)
break;
if (count != 0 && --count == 0)
break;
(void) fsleep(interval);
}
pool_list_free(list);
return (ret);
}
typedef struct list_cbdata {
boolean_t cb_verbose;
int cb_name_flags;
int cb_namewidth;
boolean_t cb_scripted;
zprop_list_t *cb_proplist;
boolean_t cb_literal;
} list_cbdata_t;
/*
* Given a list of columns to display, output appropriate headers for each one.
*/
static void
print_header(list_cbdata_t *cb)
{
zprop_list_t *pl = cb->cb_proplist;
char headerbuf[ZPOOL_MAXPROPLEN];
const char *header;
boolean_t first = B_TRUE;
boolean_t right_justify;
size_t width = 0;
for (; pl != NULL; pl = pl->pl_next) {
width = pl->pl_width;
if (first && cb->cb_verbose) {
/*
* Reset the width to accommodate the verbose listing
* of devices.
*/
width = cb->cb_namewidth;
}
if (!first)
(void) fputs(" ", stdout);
else
first = B_FALSE;
right_justify = B_FALSE;
if (pl->pl_prop != ZPROP_USERPROP) {
header = zpool_prop_column_name(pl->pl_prop);
right_justify = zpool_prop_align_right(pl->pl_prop);
} else {
int i;
for (i = 0; pl->pl_user_prop[i] != '\0'; i++)
headerbuf[i] = toupper(pl->pl_user_prop[i]);
headerbuf[i] = '\0';
header = headerbuf;
}
if (pl->pl_next == NULL && !right_justify)
(void) fputs(header, stdout);
else if (right_justify)
(void) printf("%*s", (int)width, header);
else
(void) printf("%-*s", (int)width, header);
}
(void) fputc('\n', stdout);
}
/*
* Given a pool and a list of properties, print out all the properties according
* to the described layout. Used by zpool_do_list().
*/
static void
print_pool(zpool_handle_t *zhp, list_cbdata_t *cb)
{
zprop_list_t *pl = cb->cb_proplist;
boolean_t first = B_TRUE;
char property[ZPOOL_MAXPROPLEN];
const char *propstr;
boolean_t right_justify;
size_t width;
for (; pl != NULL; pl = pl->pl_next) {
width = pl->pl_width;
if (first && cb->cb_verbose) {
/*
* Reset the width to accommodate the verbose listing
* of devices.
*/
width = cb->cb_namewidth;
}
if (!first) {
if (cb->cb_scripted)
(void) fputc('\t', stdout);
else
(void) fputs(" ", stdout);
} else {
first = B_FALSE;
}
right_justify = B_FALSE;
if (pl->pl_prop != ZPROP_USERPROP) {
if (zpool_get_prop(zhp, pl->pl_prop, property,
sizeof (property), NULL, cb->cb_literal) != 0)
propstr = "-";
else
propstr = property;
right_justify = zpool_prop_align_right(pl->pl_prop);
} else if ((zpool_prop_feature(pl->pl_user_prop) ||
zpool_prop_unsupported(pl->pl_user_prop)) &&
zpool_prop_get_feature(zhp, pl->pl_user_prop, property,
sizeof (property)) == 0) {
propstr = property;
} else if (zfs_prop_user(pl->pl_user_prop) &&
zpool_get_userprop(zhp, pl->pl_user_prop, property,
sizeof (property), NULL) == 0) {
propstr = property;
} else {
propstr = "-";
}
/*
* If this is being called in scripted mode, or if this is the
* last column and it is left-justified, don't include a width
* format specifier.
*/
if (cb->cb_scripted || (pl->pl_next == NULL && !right_justify))
(void) fputs(propstr, stdout);
else if (right_justify)
(void) printf("%*s", (int)width, propstr);
else
(void) printf("%-*s", (int)width, propstr);
}
(void) fputc('\n', stdout);
}
static void
print_one_column(zpool_prop_t prop, uint64_t value, const char *str,
boolean_t scripted, boolean_t valid, enum zfs_nicenum_format format)
{
char propval[64];
boolean_t fixed;
size_t width = zprop_width(prop, &fixed, ZFS_TYPE_POOL);
switch (prop) {
case ZPOOL_PROP_SIZE:
case ZPOOL_PROP_EXPANDSZ:
case ZPOOL_PROP_CHECKPOINT:
case ZPOOL_PROP_DEDUPRATIO:
if (value == 0)
(void) strlcpy(propval, "-", sizeof (propval));
else
zfs_nicenum_format(value, propval, sizeof (propval),
format);
break;
case ZPOOL_PROP_FRAGMENTATION:
if (value == ZFS_FRAG_INVALID) {
(void) strlcpy(propval, "-", sizeof (propval));
} else if (format == ZFS_NICENUM_RAW) {
(void) snprintf(propval, sizeof (propval), "%llu",
(unsigned long long)value);
} else {
(void) snprintf(propval, sizeof (propval), "%llu%%",
(unsigned long long)value);
}
break;
case ZPOOL_PROP_CAPACITY:
/* capacity value is in parts-per-10,000 (aka permyriad) */
if (format == ZFS_NICENUM_RAW)
(void) snprintf(propval, sizeof (propval), "%llu",
(unsigned long long)value / 100);
else
(void) snprintf(propval, sizeof (propval),
value < 1000 ? "%1.2f%%" : value < 10000 ?
"%2.1f%%" : "%3.0f%%", value / 100.0);
break;
case ZPOOL_PROP_HEALTH:
width = 8;
(void) strlcpy(propval, str, sizeof (propval));
break;
default:
zfs_nicenum_format(value, propval, sizeof (propval), format);
}
if (!valid)
(void) strlcpy(propval, "-", sizeof (propval));
if (scripted)
(void) printf("\t%s", propval);
else
(void) printf(" %*s", (int)width, propval);
}
/*
* print static default line per vdev
* not compatible with '-o' <proplist> option
*/
static void
print_list_stats(zpool_handle_t *zhp, const char *name, nvlist_t *nv,
list_cbdata_t *cb, int depth, boolean_t isspare)
{
nvlist_t **child;
vdev_stat_t *vs;
uint_t c, children;
char *vname;
boolean_t scripted = cb->cb_scripted;
uint64_t islog = B_FALSE;
const char *dashes = "%-*s - - - - "
"- - - - -\n";
verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &c) == 0);
if (name != NULL) {
boolean_t toplevel = (vs->vs_space != 0);
uint64_t cap;
enum zfs_nicenum_format format;
const char *state;
if (cb->cb_literal)
format = ZFS_NICENUM_RAW;
else
format = ZFS_NICENUM_1024;
if (strcmp(name, VDEV_TYPE_INDIRECT) == 0)
return;
if (scripted)
(void) printf("\t%s", name);
else if (strlen(name) + depth > cb->cb_namewidth)
(void) printf("%*s%s", depth, "", name);
else
(void) printf("%*s%s%*s", depth, "", name,
(int)(cb->cb_namewidth - strlen(name) - depth), "");
/*
* Print the properties for the individual vdevs. Some
* properties are only applicable to toplevel vdevs. The
* 'toplevel' boolean value is passed to the print_one_column()
* to indicate that the value is valid.
*/
if (VDEV_STAT_VALID(vs_pspace, c) && vs->vs_pspace)
print_one_column(ZPOOL_PROP_SIZE, vs->vs_pspace, NULL,
scripted, B_TRUE, format);
else
print_one_column(ZPOOL_PROP_SIZE, vs->vs_space, NULL,
scripted, toplevel, format);
print_one_column(ZPOOL_PROP_ALLOCATED, vs->vs_alloc, NULL,
scripted, toplevel, format);
print_one_column(ZPOOL_PROP_FREE, vs->vs_space - vs->vs_alloc,
NULL, scripted, toplevel, format);
print_one_column(ZPOOL_PROP_CHECKPOINT,
vs->vs_checkpoint_space, NULL, scripted, toplevel, format);
print_one_column(ZPOOL_PROP_EXPANDSZ, vs->vs_esize, NULL,
scripted, B_TRUE, format);
print_one_column(ZPOOL_PROP_FRAGMENTATION,
vs->vs_fragmentation, NULL, scripted,
(vs->vs_fragmentation != ZFS_FRAG_INVALID && toplevel),
format);
cap = (vs->vs_space == 0) ? 0 :
(vs->vs_alloc * 10000 / vs->vs_space);
print_one_column(ZPOOL_PROP_CAPACITY, cap, NULL,
scripted, toplevel, format);
print_one_column(ZPOOL_PROP_DEDUPRATIO, 0, NULL,
scripted, toplevel, format);
state = zpool_state_to_name(vs->vs_state, vs->vs_aux);
if (isspare) {
if (vs->vs_aux == VDEV_AUX_SPARED)
state = "INUSE";
else if (vs->vs_state == VDEV_STATE_HEALTHY)
state = "AVAIL";
}
print_one_column(ZPOOL_PROP_HEALTH, 0, state, scripted,
B_TRUE, format);
(void) fputc('\n', stdout);
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
return;
/* list the normal vdevs first */
for (c = 0; c < children; c++) {
uint64_t ishole = B_FALSE;
if (nvlist_lookup_uint64(child[c],
ZPOOL_CONFIG_IS_HOLE, &ishole) == 0 && ishole)
continue;
if (nvlist_lookup_uint64(child[c],
ZPOOL_CONFIG_IS_LOG, &islog) == 0 && islog)
continue;
if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS))
continue;
vname = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags | VDEV_NAME_TYPE_ID);
print_list_stats(zhp, vname, child[c], cb, depth + 2, B_FALSE);
free(vname);
}
/* list the classes: 'logs', 'dedup', and 'special' */
for (uint_t n = 0; n < ARRAY_SIZE(class_name); n++) {
boolean_t printed = B_FALSE;
for (c = 0; c < children; c++) {
const char *bias = NULL;
const char *type = NULL;
if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&islog) == 0 && islog) {
bias = VDEV_ALLOC_CLASS_LOGS;
} else {
(void) nvlist_lookup_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &bias);
(void) nvlist_lookup_string(child[c],
ZPOOL_CONFIG_TYPE, &type);
}
if (bias == NULL || strcmp(bias, class_name[n]) != 0)
continue;
if (!islog && strcmp(type, VDEV_TYPE_INDIRECT) == 0)
continue;
if (!printed) {
/* LINTED E_SEC_PRINTF_VAR_FMT */
(void) printf(dashes, cb->cb_namewidth,
class_name[n]);
printed = B_TRUE;
}
vname = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags | VDEV_NAME_TYPE_ID);
print_list_stats(zhp, vname, child[c], cb, depth + 2,
B_FALSE);
free(vname);
}
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE,
&child, &children) == 0 && children > 0) {
/* LINTED E_SEC_PRINTF_VAR_FMT */
(void) printf(dashes, cb->cb_namewidth, "cache");
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags);
print_list_stats(zhp, vname, child[c], cb, depth + 2,
B_FALSE);
free(vname);
}
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child,
&children) == 0 && children > 0) {
/* LINTED E_SEC_PRINTF_VAR_FMT */
(void) printf(dashes, cb->cb_namewidth, "spare");
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags);
print_list_stats(zhp, vname, child[c], cb, depth + 2,
B_TRUE);
free(vname);
}
}
}
/*
* Generic callback function to list a pool.
*/
static int
list_callback(zpool_handle_t *zhp, void *data)
{
list_cbdata_t *cbp = data;
print_pool(zhp, cbp);
if (cbp->cb_verbose) {
nvlist_t *config, *nvroot;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
print_list_stats(zhp, NULL, nvroot, cbp, 0, B_FALSE);
}
return (0);
}
/*
* Set the minimum pool/vdev name column width. The width must be at least 9,
* but may be as large as needed.
*/
static int
get_namewidth_list(zpool_handle_t *zhp, void *data)
{
list_cbdata_t *cb = data;
int width;
width = get_namewidth(zhp, cb->cb_namewidth,
cb->cb_name_flags | VDEV_NAME_TYPE_ID, cb->cb_verbose);
if (width < 9)
width = 9;
cb->cb_namewidth = width;
return (0);
}
/*
* zpool list [-gHLpP] [-o prop[,prop]*] [-T d|u] [pool] ... [interval [count]]
*
* -g Display guid for individual vdev name.
* -H Scripted mode. Don't display headers, and separate properties
* by a single tab.
* -L Follow links when resolving vdev path name.
* -o List of properties to display. Defaults to
* "name,size,allocated,free,expandsize,fragmentation,capacity,"
* "dedupratio,health,altroot"
* -p Display values in parsable (exact) format.
* -P Display full path for vdev name.
* -T Display a timestamp in date(1) or Unix format
*
* List all pools in the system, whether or not they're healthy. Output space
* statistics for each one, as well as health status summary.
*/
int
zpool_do_list(int argc, char **argv)
{
int c;
int ret = 0;
list_cbdata_t cb = { 0 };
static char default_props[] =
"name,size,allocated,free,checkpoint,expandsize,fragmentation,"
"capacity,dedupratio,health,altroot";
char *props = default_props;
float interval = 0;
unsigned long count = 0;
zpool_list_t *list;
boolean_t first = B_TRUE;
current_prop_type = ZFS_TYPE_POOL;
/* check options */
while ((c = getopt(argc, argv, ":gHLo:pPT:v")) != -1) {
switch (c) {
case 'g':
cb.cb_name_flags |= VDEV_NAME_GUID;
break;
case 'H':
cb.cb_scripted = B_TRUE;
break;
case 'L':
cb.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS;
break;
case 'o':
props = optarg;
break;
case 'P':
cb.cb_name_flags |= VDEV_NAME_PATH;
break;
case 'p':
cb.cb_literal = B_TRUE;
break;
case 'T':
get_timestamp_arg(*optarg);
break;
case 'v':
cb.cb_verbose = B_TRUE;
cb.cb_namewidth = 8; /* 8 until precalc is avail */
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
get_interval_count(&argc, argv, &interval, &count);
if (zprop_get_list(g_zfs, props, &cb.cb_proplist, ZFS_TYPE_POOL) != 0)
usage(B_FALSE);
for (;;) {
if ((list = pool_list_get(argc, argv, &cb.cb_proplist,
ZFS_TYPE_POOL, cb.cb_literal, &ret)) == NULL)
return (1);
if (pool_list_count(list) == 0)
break;
cb.cb_namewidth = 0;
(void) pool_list_iter(list, B_FALSE, get_namewidth_list, &cb);
if (timestamp_fmt != NODATE)
print_timestamp(timestamp_fmt);
if (!cb.cb_scripted && (first || cb.cb_verbose)) {
print_header(&cb);
first = B_FALSE;
}
ret = pool_list_iter(list, B_TRUE, list_callback, &cb);
if (interval == 0)
break;
if (count != 0 && --count == 0)
break;
pool_list_free(list);
(void) fsleep(interval);
}
if (argc == 0 && !cb.cb_scripted && pool_list_count(list) == 0) {
(void) printf(gettext("no pools available\n"));
ret = 0;
}
pool_list_free(list);
zprop_free_list(cb.cb_proplist);
return (ret);
}
static int
zpool_do_attach_or_replace(int argc, char **argv, int replacing)
{
boolean_t force = B_FALSE;
boolean_t rebuild = B_FALSE;
boolean_t wait = B_FALSE;
int c;
nvlist_t *nvroot;
char *poolname, *old_disk, *new_disk;
zpool_handle_t *zhp;
nvlist_t *props = NULL;
char *propval;
int ret;
/* check options */
while ((c = getopt(argc, argv, "fo:sw")) != -1) {
switch (c) {
case 'f':
force = B_TRUE;
break;
case 'o':
if ((propval = strchr(optarg, '=')) == NULL) {
(void) fprintf(stderr, gettext("missing "
"'=' for -o option\n"));
usage(B_FALSE);
}
*propval = '\0';
propval++;
if ((strcmp(optarg, ZPOOL_CONFIG_ASHIFT) != 0) ||
(add_prop_list(optarg, propval, &props, B_TRUE)))
usage(B_FALSE);
break;
case 's':
rebuild = B_TRUE;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
poolname = argv[0];
if (argc < 2) {
(void) fprintf(stderr,
gettext("missing <device> specification\n"));
usage(B_FALSE);
}
old_disk = argv[1];
if (argc < 3) {
if (!replacing) {
(void) fprintf(stderr,
gettext("missing <new_device> specification\n"));
usage(B_FALSE);
}
new_disk = old_disk;
argc -= 1;
argv += 1;
} else {
new_disk = argv[2];
argc -= 2;
argv += 2;
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if ((zhp = zpool_open(g_zfs, poolname)) == NULL) {
nvlist_free(props);
return (1);
}
if (zpool_get_config(zhp, NULL) == NULL) {
(void) fprintf(stderr, gettext("pool '%s' is unavailable\n"),
poolname);
zpool_close(zhp);
nvlist_free(props);
return (1);
}
/* unless manually specified use "ashift" pool property (if set) */
if (!nvlist_exists(props, ZPOOL_CONFIG_ASHIFT)) {
int intval;
zprop_source_t src;
char strval[ZPOOL_MAXPROPLEN];
intval = zpool_get_prop_int(zhp, ZPOOL_PROP_ASHIFT, &src);
if (src != ZPROP_SRC_DEFAULT) {
(void) sprintf(strval, "%" PRId32, intval);
verify(add_prop_list(ZPOOL_CONFIG_ASHIFT, strval,
&props, B_TRUE) == 0);
}
}
nvroot = make_root_vdev(zhp, props, force, B_FALSE, replacing, B_FALSE,
argc, argv);
if (nvroot == NULL) {
zpool_close(zhp);
nvlist_free(props);
return (1);
}
ret = zpool_vdev_attach(zhp, old_disk, new_disk, nvroot, replacing,
rebuild);
- if (ret == 0 && wait)
- ret = zpool_wait(zhp,
- replacing ? ZPOOL_WAIT_REPLACE : ZPOOL_WAIT_RESILVER);
+ if (ret == 0 && wait) {
+ zpool_wait_activity_t activity = ZPOOL_WAIT_RESILVER;
+ char raidz_prefix[] = "raidz";
+ if (replacing) {
+ activity = ZPOOL_WAIT_REPLACE;
+ } else if (strncmp(old_disk,
+ raidz_prefix, strlen(raidz_prefix)) == 0) {
+ activity = ZPOOL_WAIT_RAIDZ_EXPAND;
+ }
+ ret = zpool_wait(zhp, activity);
+ }
nvlist_free(props);
nvlist_free(nvroot);
zpool_close(zhp);
return (ret);
}
/*
* zpool replace [-fsw] [-o property=value] <pool> <device> <new_device>
*
* -f Force attach, even if <new_device> appears to be in use.
* -s Use sequential instead of healing reconstruction for resilver.
* -o Set property=value.
* -w Wait for replacing to complete before returning
*
* Replace <device> with <new_device>.
*/
int
zpool_do_replace(int argc, char **argv)
{
return (zpool_do_attach_or_replace(argc, argv, B_TRUE));
}
/*
- * zpool attach [-fsw] [-o property=value] <pool> <device> <new_device>
+ * zpool attach [-fsw] [-o property=value] <pool> <device>|<vdev> <new_device>
*
* -f Force attach, even if <new_device> appears to be in use.
* -s Use sequential instead of healing reconstruction for resilver.
* -o Set property=value.
- * -w Wait for resilvering to complete before returning
+ * -w Wait for resilvering (mirror) or expansion (raidz) to complete
+ * before returning.
*
- * Attach <new_device> to the mirror containing <device>. If <device> is not
- * part of a mirror, then <device> will be transformed into a mirror of
- * <device> and <new_device>. In either case, <new_device> will begin life
- * with a DTL of [0, now], and will immediately begin to resilver itself.
+ * Attach <new_device> to a <device> or <vdev>, where the vdev can be of type
+ * mirror or raidz. If <device> is not part of a mirror, then <device> will
+ * be transformed into a mirror of <device> and <new_device>. When a mirror
+ * is involved, <new_device> will begin life with a DTL of [0, now], and will
+ * immediately begin to resilver itself. For the raidz case, a expansion will
+ * commence and reflow the raidz data across all the disks including the
+ * <new_device>.
*/
int
zpool_do_attach(int argc, char **argv)
{
return (zpool_do_attach_or_replace(argc, argv, B_FALSE));
}
/*
* zpool detach [-f] <pool> <device>
*
* -f Force detach of <device>, even if DTLs argue against it
* (not supported yet)
*
* Detach a device from a mirror. The operation will be refused if <device>
* is the last device in the mirror, or if the DTLs indicate that this device
* has the only valid copy of some data.
*/
int
zpool_do_detach(int argc, char **argv)
{
int c;
char *poolname, *path;
zpool_handle_t *zhp;
int ret;
/* check options */
while ((c = getopt(argc, argv, "")) != -1) {
switch (c) {
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr,
gettext("missing <device> specification\n"));
usage(B_FALSE);
}
poolname = argv[0];
path = argv[1];
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
ret = zpool_vdev_detach(zhp, path);
zpool_close(zhp);
return (ret);
}
/*
* zpool split [-gLnP] [-o prop=val] ...
* [-o mntopt] ...
* [-R altroot] <pool> <newpool> [<device> ...]
*
* -g Display guid for individual vdev name.
* -L Follow links when resolving vdev path name.
* -n Do not split the pool, but display the resulting layout if
* it were to be split.
* -o Set property=value, or set mount options.
* -P Display full path for vdev name.
* -R Mount the split-off pool under an alternate root.
* -l Load encryption keys while importing.
*
* Splits the named pool and gives it the new pool name. Devices to be split
* off may be listed, provided that no more than one device is specified
* per top-level vdev mirror. The newly split pool is left in an exported
* state unless -R is specified.
*
* Restrictions: the top-level of the pool pool must only be made up of
* mirrors; all devices in the pool must be healthy; no device may be
* undergoing a resilvering operation.
*/
int
zpool_do_split(int argc, char **argv)
{
char *srcpool, *newpool, *propval;
char *mntopts = NULL;
splitflags_t flags;
int c, ret = 0;
int ms_status = 0;
boolean_t loadkeys = B_FALSE;
zpool_handle_t *zhp;
nvlist_t *config, *props = NULL;
flags.dryrun = B_FALSE;
flags.import = B_FALSE;
flags.name_flags = 0;
/* check options */
while ((c = getopt(argc, argv, ":gLR:lno:P")) != -1) {
switch (c) {
case 'g':
flags.name_flags |= VDEV_NAME_GUID;
break;
case 'L':
flags.name_flags |= VDEV_NAME_FOLLOW_LINKS;
break;
case 'R':
flags.import = B_TRUE;
if (add_prop_list(
zpool_prop_to_name(ZPOOL_PROP_ALTROOT), optarg,
&props, B_TRUE) != 0) {
nvlist_free(props);
usage(B_FALSE);
}
break;
case 'l':
loadkeys = B_TRUE;
break;
case 'n':
flags.dryrun = B_TRUE;
break;
case 'o':
if ((propval = strchr(optarg, '=')) != NULL) {
*propval = '\0';
propval++;
if (add_prop_list(optarg, propval,
&props, B_TRUE) != 0) {
nvlist_free(props);
usage(B_FALSE);
}
} else {
mntopts = optarg;
}
break;
case 'P':
flags.name_flags |= VDEV_NAME_PATH;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
break;
}
}
if (!flags.import && mntopts != NULL) {
(void) fprintf(stderr, gettext("setting mntopts is only "
"valid when importing the pool\n"));
usage(B_FALSE);
}
if (!flags.import && loadkeys) {
(void) fprintf(stderr, gettext("loading keys is only "
"valid when importing the pool\n"));
usage(B_FALSE);
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("Missing pool name\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("Missing new pool name\n"));
usage(B_FALSE);
}
srcpool = argv[0];
newpool = argv[1];
argc -= 2;
argv += 2;
if ((zhp = zpool_open(g_zfs, srcpool)) == NULL) {
nvlist_free(props);
return (1);
}
config = split_mirror_vdev(zhp, newpool, props, flags, argc, argv);
if (config == NULL) {
ret = 1;
} else {
if (flags.dryrun) {
(void) printf(gettext("would create '%s' with the "
"following layout:\n\n"), newpool);
print_vdev_tree(NULL, newpool, config, 0, "",
flags.name_flags);
print_vdev_tree(NULL, "dedup", config, 0,
VDEV_ALLOC_BIAS_DEDUP, 0);
print_vdev_tree(NULL, "special", config, 0,
VDEV_ALLOC_BIAS_SPECIAL, 0);
}
}
zpool_close(zhp);
if (ret != 0 || flags.dryrun || !flags.import) {
nvlist_free(config);
nvlist_free(props);
return (ret);
}
/*
* The split was successful. Now we need to open the new
* pool and import it.
*/
if ((zhp = zpool_open_canfail(g_zfs, newpool)) == NULL) {
nvlist_free(config);
nvlist_free(props);
return (1);
}
if (loadkeys) {
ret = zfs_crypto_attempt_load_keys(g_zfs, newpool);
if (ret != 0)
ret = 1;
}
if (zpool_get_state(zhp) != POOL_STATE_UNAVAIL) {
ms_status = zpool_enable_datasets(zhp, mntopts, 0);
if (ms_status == EZFS_SHAREFAILED) {
(void) fprintf(stderr, gettext("Split was successful, "
"datasets are mounted but sharing of some datasets "
"has failed\n"));
} else if (ms_status == EZFS_MOUNTFAILED) {
(void) fprintf(stderr, gettext("Split was successful"
", but some datasets could not be mounted\n"));
(void) fprintf(stderr, gettext("Try doing '%s' with a "
"different altroot\n"), "zpool import");
}
}
zpool_close(zhp);
nvlist_free(config);
nvlist_free(props);
return (ret);
}
/*
* zpool online <pool> <device> ...
*/
int
zpool_do_online(int argc, char **argv)
{
int c, i;
char *poolname;
zpool_handle_t *zhp;
int ret = 0;
vdev_state_t newstate;
int flags = 0;
/* check options */
while ((c = getopt(argc, argv, "e")) != -1) {
switch (c) {
case 'e':
flags |= ZFS_ONLINE_EXPAND;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing device name\n"));
usage(B_FALSE);
}
poolname = argv[0];
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
for (i = 1; i < argc; i++) {
vdev_state_t oldstate;
boolean_t avail_spare, l2cache;
nvlist_t *tgt = zpool_find_vdev(zhp, argv[i], &avail_spare,
&l2cache, NULL);
if (tgt == NULL) {
ret = 1;
continue;
}
uint_t vsc;
oldstate = ((vdev_stat_t *)fnvlist_lookup_uint64_array(tgt,
ZPOOL_CONFIG_VDEV_STATS, &vsc))->vs_state;
if (zpool_vdev_online(zhp, argv[i], flags, &newstate) == 0) {
if (newstate != VDEV_STATE_HEALTHY) {
(void) printf(gettext("warning: device '%s' "
"onlined, but remains in faulted state\n"),
argv[i]);
if (newstate == VDEV_STATE_FAULTED)
(void) printf(gettext("use 'zpool "
"clear' to restore a faulted "
"device\n"));
else
(void) printf(gettext("use 'zpool "
"replace' to replace devices "
"that are no longer present\n"));
if ((flags & ZFS_ONLINE_EXPAND)) {
(void) printf(gettext("%s: failed "
"to expand usable space on "
"unhealthy device '%s'\n"),
(oldstate >= VDEV_STATE_DEGRADED ?
"error" : "warning"), argv[i]);
if (oldstate >= VDEV_STATE_DEGRADED) {
ret = 1;
break;
}
}
}
} else {
ret = 1;
}
}
zpool_close(zhp);
return (ret);
}
/*
* zpool offline [-ft] <pool> <device> ...
*
* -f Force the device into a faulted state.
*
* -t Only take the device off-line temporarily. The offline/faulted
* state will not be persistent across reboots.
*/
int
zpool_do_offline(int argc, char **argv)
{
int c, i;
char *poolname;
zpool_handle_t *zhp;
int ret = 0;
boolean_t istmp = B_FALSE;
boolean_t fault = B_FALSE;
/* check options */
while ((c = getopt(argc, argv, "ft")) != -1) {
switch (c) {
case 'f':
fault = B_TRUE;
break;
case 't':
istmp = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing device name\n"));
usage(B_FALSE);
}
poolname = argv[0];
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
for (i = 1; i < argc; i++) {
if (fault) {
uint64_t guid = zpool_vdev_path_to_guid(zhp, argv[i]);
vdev_aux_t aux;
if (istmp == B_FALSE) {
/* Force the fault to persist across imports */
aux = VDEV_AUX_EXTERNAL_PERSIST;
} else {
aux = VDEV_AUX_EXTERNAL;
}
if (guid == 0 || zpool_vdev_fault(zhp, guid, aux) != 0)
ret = 1;
} else {
if (zpool_vdev_offline(zhp, argv[i], istmp) != 0)
ret = 1;
}
}
zpool_close(zhp);
return (ret);
}
/*
* zpool clear <pool> [device]
*
* Clear all errors associated with a pool or a particular device.
*/
int
zpool_do_clear(int argc, char **argv)
{
int c;
int ret = 0;
boolean_t dryrun = B_FALSE;
boolean_t do_rewind = B_FALSE;
boolean_t xtreme_rewind = B_FALSE;
uint32_t rewind_policy = ZPOOL_NO_REWIND;
nvlist_t *policy = NULL;
zpool_handle_t *zhp;
char *pool, *device;
/* check options */
while ((c = getopt(argc, argv, "FnX")) != -1) {
switch (c) {
case 'F':
do_rewind = B_TRUE;
break;
case 'n':
dryrun = B_TRUE;
break;
case 'X':
xtreme_rewind = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name\n"));
usage(B_FALSE);
}
if (argc > 2) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if ((dryrun || xtreme_rewind) && !do_rewind) {
(void) fprintf(stderr,
gettext("-n or -X only meaningful with -F\n"));
usage(B_FALSE);
}
if (dryrun)
rewind_policy = ZPOOL_TRY_REWIND;
else if (do_rewind)
rewind_policy = ZPOOL_DO_REWIND;
if (xtreme_rewind)
rewind_policy |= ZPOOL_EXTREME_REWIND;
/* In future, further rewind policy choices can be passed along here */
if (nvlist_alloc(&policy, NV_UNIQUE_NAME, 0) != 0 ||
nvlist_add_uint32(policy, ZPOOL_LOAD_REWIND_POLICY,
rewind_policy) != 0) {
return (1);
}
pool = argv[0];
device = argc == 2 ? argv[1] : NULL;
if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL) {
nvlist_free(policy);
return (1);
}
if (zpool_clear(zhp, device, policy) != 0)
ret = 1;
zpool_close(zhp);
nvlist_free(policy);
return (ret);
}
/*
* zpool reguid <pool>
*/
int
zpool_do_reguid(int argc, char **argv)
{
int c;
char *poolname;
zpool_handle_t *zhp;
int ret = 0;
/* check options */
while ((c = getopt(argc, argv, "")) != -1) {
switch (c) {
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
poolname = argv[0];
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
ret = zpool_reguid(zhp);
zpool_close(zhp);
return (ret);
}
/*
* zpool reopen <pool>
*
* Reopen the pool so that the kernel can update the sizes of all vdevs.
*/
int
zpool_do_reopen(int argc, char **argv)
{
int c;
int ret = 0;
boolean_t scrub_restart = B_TRUE;
/* check options */
while ((c = getopt(argc, argv, "n")) != -1) {
switch (c) {
case 'n':
scrub_restart = B_FALSE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* if argc == 0 we will execute zpool_reopen_one on all pools */
ret = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, zpool_reopen_one, &scrub_restart);
return (ret);
}
typedef struct scrub_cbdata {
int cb_type;
pool_scrub_cmd_t cb_scrub_cmd;
} scrub_cbdata_t;
static boolean_t
zpool_has_checkpoint(zpool_handle_t *zhp)
{
nvlist_t *config, *nvroot;
config = zpool_get_config(zhp, NULL);
if (config != NULL) {
pool_checkpoint_stat_t *pcs = NULL;
uint_t c;
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c);
if (pcs == NULL || pcs->pcs_state == CS_NONE)
return (B_FALSE);
assert(pcs->pcs_state == CS_CHECKPOINT_EXISTS ||
pcs->pcs_state == CS_CHECKPOINT_DISCARDING);
return (B_TRUE);
}
return (B_FALSE);
}
static int
scrub_callback(zpool_handle_t *zhp, void *data)
{
scrub_cbdata_t *cb = data;
int err;
/*
* Ignore faulted pools.
*/
if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) {
(void) fprintf(stderr, gettext("cannot scan '%s': pool is "
"currently unavailable\n"), zpool_get_name(zhp));
return (1);
}
err = zpool_scan(zhp, cb->cb_type, cb->cb_scrub_cmd);
if (err == 0 && zpool_has_checkpoint(zhp) &&
cb->cb_type == POOL_SCAN_SCRUB) {
(void) printf(gettext("warning: will not scrub state that "
"belongs to the checkpoint of pool '%s'\n"),
zpool_get_name(zhp));
}
return (err != 0);
}
static int
wait_callback(zpool_handle_t *zhp, void *data)
{
zpool_wait_activity_t *act = data;
return (zpool_wait(zhp, *act));
}
/*
* zpool scrub [-s | -p] [-w] [-e] <pool> ...
*
* -e Only scrub blocks in the error log.
* -s Stop. Stops any in-progress scrub.
* -p Pause. Pause in-progress scrub.
* -w Wait. Blocks until scrub has completed.
*/
int
zpool_do_scrub(int argc, char **argv)
{
int c;
scrub_cbdata_t cb;
boolean_t wait = B_FALSE;
int error;
cb.cb_type = POOL_SCAN_SCRUB;
cb.cb_scrub_cmd = POOL_SCRUB_NORMAL;
boolean_t is_error_scrub = B_FALSE;
boolean_t is_pause = B_FALSE;
boolean_t is_stop = B_FALSE;
/* check options */
while ((c = getopt(argc, argv, "spwe")) != -1) {
switch (c) {
case 'e':
is_error_scrub = B_TRUE;
break;
case 's':
is_stop = B_TRUE;
break;
case 'p':
is_pause = B_TRUE;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
if (is_pause && is_stop) {
(void) fprintf(stderr, gettext("invalid option "
"combination :-s and -p are mutually exclusive\n"));
usage(B_FALSE);
} else {
if (is_error_scrub)
cb.cb_type = POOL_SCAN_ERRORSCRUB;
if (is_pause) {
cb.cb_scrub_cmd = POOL_SCRUB_PAUSE;
} else if (is_stop) {
cb.cb_type = POOL_SCAN_NONE;
} else {
cb.cb_scrub_cmd = POOL_SCRUB_NORMAL;
}
}
if (wait && (cb.cb_type == POOL_SCAN_NONE ||
cb.cb_scrub_cmd == POOL_SCRUB_PAUSE)) {
(void) fprintf(stderr, gettext("invalid option combination: "
"-w cannot be used with -p or -s\n"));
usage(B_FALSE);
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
error = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, scrub_callback, &cb);
if (wait && !error) {
zpool_wait_activity_t act = ZPOOL_WAIT_SCRUB;
error = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, wait_callback, &act);
}
return (error);
}
/*
* zpool resilver <pool> ...
*
* Restarts any in-progress resilver
*/
int
zpool_do_resilver(int argc, char **argv)
{
int c;
scrub_cbdata_t cb;
cb.cb_type = POOL_SCAN_RESILVER;
cb.cb_scrub_cmd = POOL_SCRUB_NORMAL;
/* check options */
while ((c = getopt(argc, argv, "")) != -1) {
switch (c) {
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
return (for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, scrub_callback, &cb));
}
/*
* zpool trim [-d] [-r <rate>] [-c | -s] <pool> [<device> ...]
*
* -c Cancel. Ends any in-progress trim.
* -d Secure trim. Requires kernel and device support.
* -r <rate> Sets the TRIM rate in bytes (per second). Supports
* adding a multiplier suffix such as 'k' or 'm'.
* -s Suspend. TRIM can then be restarted with no flags.
* -w Wait. Blocks until trimming has completed.
*/
int
zpool_do_trim(int argc, char **argv)
{
struct option long_options[] = {
{"cancel", no_argument, NULL, 'c'},
{"secure", no_argument, NULL, 'd'},
{"rate", required_argument, NULL, 'r'},
{"suspend", no_argument, NULL, 's'},
{"wait", no_argument, NULL, 'w'},
{0, 0, 0, 0}
};
pool_trim_func_t cmd_type = POOL_TRIM_START;
uint64_t rate = 0;
boolean_t secure = B_FALSE;
boolean_t wait = B_FALSE;
int c;
while ((c = getopt_long(argc, argv, "cdr:sw", long_options, NULL))
!= -1) {
switch (c) {
case 'c':
if (cmd_type != POOL_TRIM_START &&
cmd_type != POOL_TRIM_CANCEL) {
(void) fprintf(stderr, gettext("-c cannot be "
"combined with other options\n"));
usage(B_FALSE);
}
cmd_type = POOL_TRIM_CANCEL;
break;
case 'd':
if (cmd_type != POOL_TRIM_START) {
(void) fprintf(stderr, gettext("-d cannot be "
"combined with the -c or -s options\n"));
usage(B_FALSE);
}
secure = B_TRUE;
break;
case 'r':
if (cmd_type != POOL_TRIM_START) {
(void) fprintf(stderr, gettext("-r cannot be "
"combined with the -c or -s options\n"));
usage(B_FALSE);
}
if (zfs_nicestrtonum(g_zfs, optarg, &rate) == -1) {
(void) fprintf(stderr, "%s: %s\n",
gettext("invalid value for rate"),
libzfs_error_description(g_zfs));
usage(B_FALSE);
}
break;
case 's':
if (cmd_type != POOL_TRIM_START &&
cmd_type != POOL_TRIM_SUSPEND) {
(void) fprintf(stderr, gettext("-s cannot be "
"combined with other options\n"));
usage(B_FALSE);
}
cmd_type = POOL_TRIM_SUSPEND;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
if (optopt != 0) {
(void) fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
} else {
(void) fprintf(stderr,
gettext("invalid option '%s'\n"),
argv[optind - 1]);
}
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
return (-1);
}
if (wait && (cmd_type != POOL_TRIM_START)) {
(void) fprintf(stderr, gettext("-w cannot be used with -c or "
"-s\n"));
usage(B_FALSE);
}
char *poolname = argv[0];
zpool_handle_t *zhp = zpool_open(g_zfs, poolname);
if (zhp == NULL)
return (-1);
trimflags_t trim_flags = {
.secure = secure,
.rate = rate,
.wait = wait,
};
nvlist_t *vdevs = fnvlist_alloc();
if (argc == 1) {
/* no individual leaf vdevs specified, so add them all */
nvlist_t *config = zpool_get_config(zhp, NULL);
nvlist_t *nvroot = fnvlist_lookup_nvlist(config,
ZPOOL_CONFIG_VDEV_TREE);
zpool_collect_leaves(zhp, nvroot, vdevs);
trim_flags.fullpool = B_TRUE;
} else {
trim_flags.fullpool = B_FALSE;
for (int i = 1; i < argc; i++) {
fnvlist_add_boolean(vdevs, argv[i]);
}
}
int error = zpool_trim(zhp, cmd_type, vdevs, &trim_flags);
fnvlist_free(vdevs);
zpool_close(zhp);
return (error);
}
/*
* Converts a total number of seconds to a human readable string broken
* down in to days/hours/minutes/seconds.
*/
static void
secs_to_dhms(uint64_t total, char *buf)
{
uint64_t days = total / 60 / 60 / 24;
uint64_t hours = (total / 60 / 60) % 24;
uint64_t mins = (total / 60) % 60;
uint64_t secs = (total % 60);
if (days > 0) {
(void) sprintf(buf, "%llu days %02llu:%02llu:%02llu",
(u_longlong_t)days, (u_longlong_t)hours,
(u_longlong_t)mins, (u_longlong_t)secs);
} else {
(void) sprintf(buf, "%02llu:%02llu:%02llu",
(u_longlong_t)hours, (u_longlong_t)mins,
(u_longlong_t)secs);
}
}
/*
* Print out detailed error scrub status.
*/
static void
print_err_scrub_status(pool_scan_stat_t *ps)
{
time_t start, end, pause;
uint64_t total_secs_left;
uint64_t secs_left, mins_left, hours_left, days_left;
uint64_t examined, to_be_examined;
if (ps == NULL || ps->pss_error_scrub_func != POOL_SCAN_ERRORSCRUB) {
return;
}
(void) printf(gettext(" scrub: "));
start = ps->pss_error_scrub_start;
end = ps->pss_error_scrub_end;
pause = ps->pss_pass_error_scrub_pause;
examined = ps->pss_error_scrub_examined;
to_be_examined = ps->pss_error_scrub_to_be_examined;
assert(ps->pss_error_scrub_func == POOL_SCAN_ERRORSCRUB);
if (ps->pss_error_scrub_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);
(void) printf(gettext("scrubbed %llu error blocks in %llu days "
"%02llu:%02llu:%02llu on %s"), (u_longlong_t)examined,
(u_longlong_t)days_left, (u_longlong_t)hours_left,
(u_longlong_t)mins_left, (u_longlong_t)secs_left,
ctime(&end));
return;
} else if (ps->pss_error_scrub_state == DSS_CANCELED) {
(void) printf(gettext("error scrub canceled on %s"),
ctime(&end));
return;
}
assert(ps->pss_error_scrub_state == DSS_ERRORSCRUBBING);
/* Error scrub is in progress. */
if (pause == 0) {
(void) printf(gettext("error scrub in progress since %s"),
ctime(&start));
} else {
(void) printf(gettext("error scrub paused since %s"),
ctime(&pause));
(void) printf(gettext("\terror scrub started on %s"),
ctime(&start));
}
double fraction_done = (double)examined / (to_be_examined + examined);
(void) printf(gettext("\t%.2f%% done, issued I/O for %llu error"
" blocks"), 100 * fraction_done, (u_longlong_t)examined);
(void) printf("\n");
}
/*
* Print out detailed scrub status.
*/
static void
print_scan_scrub_resilver_status(pool_scan_stat_t *ps)
{
time_t start, end, pause;
uint64_t pass_scanned, scanned, pass_issued, issued, total_s, total_i;
uint64_t elapsed, scan_rate, issue_rate;
double fraction_done;
char processed_buf[7], scanned_buf[7], issued_buf[7], total_s_buf[7];
char total_i_buf[7], srate_buf[7], irate_buf[7], time_buf[32];
printf(" ");
printf_color(ANSI_BOLD, gettext("scan:"));
printf(" ");
/* If there's never been a scan, there's not much to say. */
if (ps == NULL || ps->pss_func == POOL_SCAN_NONE ||
ps->pss_func >= POOL_SCAN_FUNCS) {
(void) printf(gettext("none requested\n"));
return;
}
start = ps->pss_start_time;
end = ps->pss_end_time;
pause = ps->pss_pass_scrub_pause;
zfs_nicebytes(ps->pss_processed, processed_buf, sizeof (processed_buf));
int is_resilver = ps->pss_func == POOL_SCAN_RESILVER;
int is_scrub = ps->pss_func == POOL_SCAN_SCRUB;
assert(is_resilver || is_scrub);
/* Scan is finished or canceled. */
if (ps->pss_state == DSS_FINISHED) {
secs_to_dhms(end - start, time_buf);
if (is_scrub) {
(void) printf(gettext("scrub repaired %s "
"in %s with %llu errors on %s"), processed_buf,
time_buf, (u_longlong_t)ps->pss_errors,
ctime(&end));
} else if (is_resilver) {
(void) printf(gettext("resilvered %s "
"in %s with %llu errors on %s"), processed_buf,
time_buf, (u_longlong_t)ps->pss_errors,
ctime(&end));
}
return;
} else if (ps->pss_state == DSS_CANCELED) {
if (is_scrub) {
(void) printf(gettext("scrub canceled on %s"),
ctime(&end));
} else if (is_resilver) {
(void) printf(gettext("resilver canceled on %s"),
ctime(&end));
}
return;
}
assert(ps->pss_state == DSS_SCANNING);
/* Scan is in progress. Resilvers can't be paused. */
if (is_scrub) {
if (pause == 0) {
(void) printf(gettext("scrub in progress since %s"),
ctime(&start));
} else {
(void) printf(gettext("scrub paused since %s"),
ctime(&pause));
(void) printf(gettext("\tscrub started on %s"),
ctime(&start));
}
} else if (is_resilver) {
(void) printf(gettext("resilver in progress since %s"),
ctime(&start));
}
scanned = ps->pss_examined;
pass_scanned = ps->pss_pass_exam;
issued = ps->pss_issued;
pass_issued = ps->pss_pass_issued;
total_s = ps->pss_to_examine;
total_i = ps->pss_to_examine - ps->pss_skipped;
/* we are only done with a block once we have issued the IO for it */
fraction_done = (double)issued / total_i;
/* 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;
/* format all of the numbers we will be reporting */
zfs_nicebytes(scanned, scanned_buf, sizeof (scanned_buf));
zfs_nicebytes(issued, issued_buf, sizeof (issued_buf));
zfs_nicebytes(total_s, total_s_buf, sizeof (total_s_buf));
zfs_nicebytes(total_i, total_i_buf, sizeof (total_i_buf));
/* do not print estimated time if we have a paused scrub */
(void) printf(gettext("\t%s / %s scanned"), scanned_buf, total_s_buf);
if (pause == 0 && scan_rate > 0) {
zfs_nicebytes(scan_rate, srate_buf, sizeof (srate_buf));
(void) printf(gettext(" at %s/s"), srate_buf);
}
(void) printf(gettext(", %s / %s issued"), issued_buf, total_i_buf);
if (pause == 0 && issue_rate > 0) {
zfs_nicebytes(issue_rate, irate_buf, sizeof (irate_buf));
(void) printf(gettext(" at %s/s"), irate_buf);
}
(void) printf(gettext("\n"));
if (is_resilver) {
(void) printf(gettext("\t%s resilvered, %.2f%% done"),
processed_buf, 100 * fraction_done);
} else if (is_scrub) {
(void) printf(gettext("\t%s repaired, %.2f%% done"),
processed_buf, 100 * fraction_done);
}
if (pause == 0) {
/*
* Only provide an estimate iff:
* 1) we haven't yet issued all we expected, and
* 2) the issue rate exceeds 10 MB/s, and
* 3) it's either:
* a) a resilver which has started repairs, or
* b) a scrub which has entered the issue phase.
*/
if (total_i >= issued && issue_rate >= 10 * 1024 * 1024 &&
((is_resilver && ps->pss_processed > 0) ||
(is_scrub && issued > 0))) {
secs_to_dhms((total_i - issued) / issue_rate, time_buf);
(void) printf(gettext(", %s to go\n"), time_buf);
} else {
(void) printf(gettext(", no estimated "
"completion time\n"));
}
} else {
(void) printf(gettext("\n"));
}
}
static void
print_rebuild_status_impl(vdev_rebuild_stat_t *vrs, uint_t c, char *vdev_name)
{
if (vrs == NULL || vrs->vrs_state == VDEV_REBUILD_NONE)
return;
printf(" ");
printf_color(ANSI_BOLD, gettext("scan:"));
printf(" ");
uint64_t bytes_scanned = vrs->vrs_bytes_scanned;
uint64_t bytes_issued = vrs->vrs_bytes_issued;
uint64_t bytes_rebuilt = vrs->vrs_bytes_rebuilt;
uint64_t bytes_est_s = vrs->vrs_bytes_est;
uint64_t bytes_est_i = vrs->vrs_bytes_est;
if (c > offsetof(vdev_rebuild_stat_t, vrs_pass_bytes_skipped) / 8)
bytes_est_i -= vrs->vrs_pass_bytes_skipped;
uint64_t scan_rate = (vrs->vrs_pass_bytes_scanned /
(vrs->vrs_pass_time_ms + 1)) * 1000;
uint64_t issue_rate = (vrs->vrs_pass_bytes_issued /
(vrs->vrs_pass_time_ms + 1)) * 1000;
double scan_pct = MIN((double)bytes_scanned * 100 /
(bytes_est_s + 1), 100);
/* Format all of the numbers we will be reporting */
char bytes_scanned_buf[7], bytes_issued_buf[7];
char bytes_rebuilt_buf[7], bytes_est_s_buf[7], bytes_est_i_buf[7];
char scan_rate_buf[7], issue_rate_buf[7], time_buf[32];
zfs_nicebytes(bytes_scanned, bytes_scanned_buf,
sizeof (bytes_scanned_buf));
zfs_nicebytes(bytes_issued, bytes_issued_buf,
sizeof (bytes_issued_buf));
zfs_nicebytes(bytes_rebuilt, bytes_rebuilt_buf,
sizeof (bytes_rebuilt_buf));
zfs_nicebytes(bytes_est_s, bytes_est_s_buf, sizeof (bytes_est_s_buf));
zfs_nicebytes(bytes_est_i, bytes_est_i_buf, sizeof (bytes_est_i_buf));
time_t start = vrs->vrs_start_time;
time_t end = vrs->vrs_end_time;
/* Rebuild is finished or canceled. */
if (vrs->vrs_state == VDEV_REBUILD_COMPLETE) {
secs_to_dhms(vrs->vrs_scan_time_ms / 1000, time_buf);
(void) printf(gettext("resilvered (%s) %s in %s "
"with %llu errors on %s"), vdev_name, bytes_rebuilt_buf,
time_buf, (u_longlong_t)vrs->vrs_errors, ctime(&end));
return;
} else if (vrs->vrs_state == VDEV_REBUILD_CANCELED) {
(void) printf(gettext("resilver (%s) canceled on %s"),
vdev_name, ctime(&end));
return;
} else if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) {
(void) printf(gettext("resilver (%s) in progress since %s"),
vdev_name, ctime(&start));
}
assert(vrs->vrs_state == VDEV_REBUILD_ACTIVE);
(void) printf(gettext("\t%s / %s scanned"), bytes_scanned_buf,
bytes_est_s_buf);
if (scan_rate > 0) {
zfs_nicebytes(scan_rate, scan_rate_buf, sizeof (scan_rate_buf));
(void) printf(gettext(" at %s/s"), scan_rate_buf);
}
(void) printf(gettext(", %s / %s issued"), bytes_issued_buf,
bytes_est_i_buf);
if (issue_rate > 0) {
zfs_nicebytes(issue_rate, issue_rate_buf,
sizeof (issue_rate_buf));
(void) printf(gettext(" at %s/s"), issue_rate_buf);
}
(void) printf(gettext("\n"));
(void) printf(gettext("\t%s resilvered, %.2f%% done"),
bytes_rebuilt_buf, scan_pct);
if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) {
if (bytes_est_s >= bytes_scanned &&
scan_rate >= 10 * 1024 * 1024) {
secs_to_dhms((bytes_est_s - bytes_scanned) / scan_rate,
time_buf);
(void) printf(gettext(", %s to go\n"), time_buf);
} else {
(void) printf(gettext(", no estimated "
"completion time\n"));
}
} else {
(void) printf(gettext("\n"));
}
}
/*
* Print rebuild status for top-level vdevs.
*/
static void
print_rebuild_status(zpool_handle_t *zhp, nvlist_t *nvroot)
{
nvlist_t **child;
uint_t children;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
for (uint_t c = 0; c < children; c++) {
vdev_rebuild_stat_t *vrs;
uint_t i;
if (nvlist_lookup_uint64_array(child[c],
ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i) == 0) {
char *name = zpool_vdev_name(g_zfs, zhp,
child[c], VDEV_NAME_TYPE_ID);
print_rebuild_status_impl(vrs, i, name);
free(name);
}
}
}
/*
* As we don't scrub checkpointed blocks, we want to warn the user that we
* skipped scanning some blocks if a checkpoint exists or existed at any
* time during the scan. If a sequential instead of healing reconstruction
* was performed then the blocks were reconstructed. However, their checksums
* have not been verified so we still print the warning.
*/
static void
print_checkpoint_scan_warning(pool_scan_stat_t *ps, pool_checkpoint_stat_t *pcs)
{
if (ps == NULL || pcs == NULL)
return;
if (pcs->pcs_state == CS_NONE ||
pcs->pcs_state == CS_CHECKPOINT_DISCARDING)
return;
assert(pcs->pcs_state == CS_CHECKPOINT_EXISTS);
if (ps->pss_state == DSS_NONE)
return;
if ((ps->pss_state == DSS_FINISHED || ps->pss_state == DSS_CANCELED) &&
ps->pss_end_time < pcs->pcs_start_time)
return;
if (ps->pss_state == DSS_FINISHED || ps->pss_state == DSS_CANCELED) {
(void) printf(gettext(" scan warning: skipped blocks "
"that are only referenced by the checkpoint.\n"));
} else {
assert(ps->pss_state == DSS_SCANNING);
(void) printf(gettext(" scan warning: skipping blocks "
"that are only referenced by the checkpoint.\n"));
}
}
/*
* Returns B_TRUE if there is an active rebuild in progress. Otherwise,
* B_FALSE is returned and 'rebuild_end_time' is set to the end time for
* the last completed (or cancelled) rebuild.
*/
static boolean_t
check_rebuilding(nvlist_t *nvroot, uint64_t *rebuild_end_time)
{
nvlist_t **child;
uint_t children;
boolean_t rebuilding = B_FALSE;
uint64_t end_time = 0;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
for (uint_t c = 0; c < children; c++) {
vdev_rebuild_stat_t *vrs;
uint_t i;
if (nvlist_lookup_uint64_array(child[c],
ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i) == 0) {
if (vrs->vrs_end_time > end_time)
end_time = vrs->vrs_end_time;
if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) {
rebuilding = B_TRUE;
end_time = 0;
break;
}
}
}
if (rebuild_end_time != NULL)
*rebuild_end_time = end_time;
return (rebuilding);
}
/*
* Print the scan status.
*/
static void
print_scan_status(zpool_handle_t *zhp, nvlist_t *nvroot)
{
uint64_t rebuild_end_time = 0, resilver_end_time = 0;
boolean_t have_resilver = B_FALSE, have_scrub = B_FALSE;
boolean_t have_errorscrub = B_FALSE;
boolean_t active_resilver = B_FALSE;
pool_checkpoint_stat_t *pcs = NULL;
pool_scan_stat_t *ps = NULL;
uint_t c;
time_t scrub_start = 0, errorscrub_start = 0;
if (nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS,
(uint64_t **)&ps, &c) == 0) {
if (ps->pss_func == POOL_SCAN_RESILVER) {
resilver_end_time = ps->pss_end_time;
active_resilver = (ps->pss_state == DSS_SCANNING);
}
have_resilver = (ps->pss_func == POOL_SCAN_RESILVER);
have_scrub = (ps->pss_func == POOL_SCAN_SCRUB);
scrub_start = ps->pss_start_time;
if (c > offsetof(pool_scan_stat_t,
pss_pass_error_scrub_pause) / 8) {
have_errorscrub = (ps->pss_error_scrub_func ==
POOL_SCAN_ERRORSCRUB);
errorscrub_start = ps->pss_error_scrub_start;
}
}
boolean_t active_rebuild = check_rebuilding(nvroot, &rebuild_end_time);
boolean_t have_rebuild = (active_rebuild || (rebuild_end_time > 0));
/* Always print the scrub status when available. */
if (have_scrub && scrub_start > errorscrub_start)
print_scan_scrub_resilver_status(ps);
else if (have_errorscrub && errorscrub_start >= scrub_start)
print_err_scrub_status(ps);
/*
* When there is an active resilver or rebuild print its status.
* Otherwise print the status of the last resilver or rebuild.
*/
if (active_resilver || (!active_rebuild && have_resilver &&
resilver_end_time && resilver_end_time > rebuild_end_time)) {
print_scan_scrub_resilver_status(ps);
} else if (active_rebuild || (!active_resilver && have_rebuild &&
rebuild_end_time && rebuild_end_time > resilver_end_time)) {
print_rebuild_status(zhp, nvroot);
}
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c);
print_checkpoint_scan_warning(ps, pcs);
}
/*
* Print out detailed removal status.
*/
static void
print_removal_status(zpool_handle_t *zhp, pool_removal_stat_t *prs)
{
char copied_buf[7], examined_buf[7], total_buf[7], rate_buf[7];
time_t start, end;
nvlist_t *config, *nvroot;
nvlist_t **child;
uint_t children;
char *vdev_name;
if (prs == NULL || prs->prs_state == DSS_NONE)
return;
/*
* Determine name of vdev.
*/
config = zpool_get_config(zhp, NULL);
nvroot = fnvlist_lookup_nvlist(config,
ZPOOL_CONFIG_VDEV_TREE);
verify(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) == 0);
assert(prs->prs_removing_vdev < children);
vdev_name = zpool_vdev_name(g_zfs, zhp,
child[prs->prs_removing_vdev], B_TRUE);
printf_color(ANSI_BOLD, gettext("remove: "));
start = prs->prs_start_time;
end = prs->prs_end_time;
zfs_nicenum(prs->prs_copied, copied_buf, sizeof (copied_buf));
/*
* Removal is finished or canceled.
*/
if (prs->prs_state == DSS_FINISHED) {
uint64_t minutes_taken = (end - start) / 60;
(void) printf(gettext("Removal of vdev %llu copied %s "
"in %lluh%um, completed on %s"),
(longlong_t)prs->prs_removing_vdev,
copied_buf,
(u_longlong_t)(minutes_taken / 60),
(uint_t)(minutes_taken % 60),
ctime((time_t *)&end));
} else if (prs->prs_state == DSS_CANCELED) {
(void) printf(gettext("Removal of %s canceled on %s"),
vdev_name, ctime(&end));
} else {
uint64_t copied, total, elapsed, mins_left, hours_left;
double fraction_done;
uint_t rate;
assert(prs->prs_state == DSS_SCANNING);
/*
* Removal is in progress.
*/
(void) printf(gettext(
"Evacuation of %s in progress since %s"),
vdev_name, ctime(&start));
copied = prs->prs_copied > 0 ? prs->prs_copied : 1;
total = prs->prs_to_copy;
fraction_done = (double)copied / total;
/* elapsed time for this pass */
elapsed = time(NULL) - prs->prs_start_time;
elapsed = elapsed > 0 ? elapsed : 1;
rate = copied / elapsed;
rate = rate > 0 ? rate : 1;
mins_left = ((total - copied) / rate) / 60;
hours_left = mins_left / 60;
zfs_nicenum(copied, examined_buf, sizeof (examined_buf));
zfs_nicenum(total, total_buf, sizeof (total_buf));
zfs_nicenum(rate, rate_buf, sizeof (rate_buf));
/*
* do not print estimated time if hours_left is more than
* 30 days
*/
(void) printf(gettext(
"\t%s copied out of %s at %s/s, %.2f%% done"),
examined_buf, total_buf, rate_buf, 100 * fraction_done);
if (hours_left < (30 * 24)) {
(void) printf(gettext(", %lluh%um to go\n"),
(u_longlong_t)hours_left, (uint_t)(mins_left % 60));
} else {
(void) printf(gettext(
", (copy is slow, no estimated time)\n"));
}
}
free(vdev_name);
if (prs->prs_mapping_memory > 0) {
char mem_buf[7];
zfs_nicenum(prs->prs_mapping_memory, mem_buf, sizeof (mem_buf));
(void) printf(gettext(
"\t%s memory used for removed device mappings\n"),
mem_buf);
}
}
+/*
+ * Print out detailed raidz expansion status.
+ */
+static void
+print_raidz_expand_status(zpool_handle_t *zhp, pool_raidz_expand_stat_t *pres)
+{
+ char copied_buf[7];
+
+ if (pres == NULL || pres->pres_state == DSS_NONE)
+ return;
+
+ /*
+ * Determine name of vdev.
+ */
+ nvlist_t *config = zpool_get_config(zhp, NULL);
+ nvlist_t *nvroot = fnvlist_lookup_nvlist(config,
+ ZPOOL_CONFIG_VDEV_TREE);
+ nvlist_t **child;
+ uint_t children;
+ verify(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
+ &child, &children) == 0);
+ assert(pres->pres_expanding_vdev < children);
+
+ printf_color(ANSI_BOLD, gettext("expand: "));
+
+ time_t start = pres->pres_start_time;
+ time_t end = pres->pres_end_time;
+ char *vname =
+ zpool_vdev_name(g_zfs, zhp, child[pres->pres_expanding_vdev], 0);
+ zfs_nicenum(pres->pres_reflowed, copied_buf, sizeof (copied_buf));
+
+ /*
+ * Expansion is finished or canceled.
+ */
+ if (pres->pres_state == DSS_FINISHED) {
+ char time_buf[32];
+ secs_to_dhms(end - start, time_buf);
+
+ (void) printf(gettext("expanded %s-%u copied %s in %s, "
+ "on %s"), vname, (int)pres->pres_expanding_vdev,
+ copied_buf, time_buf, ctime((time_t *)&end));
+ } else {
+ char examined_buf[7], total_buf[7], rate_buf[7];
+ uint64_t copied, total, elapsed, secs_left;
+ double fraction_done;
+ uint_t rate;
+
+ assert(pres->pres_state == DSS_SCANNING);
+
+ /*
+ * Expansion is in progress.
+ */
+ (void) printf(gettext(
+ "expansion of %s-%u in progress since %s"),
+ vname, (int)pres->pres_expanding_vdev, ctime(&start));
+
+ copied = pres->pres_reflowed > 0 ? pres->pres_reflowed : 1;
+ total = pres->pres_to_reflow;
+ fraction_done = (double)copied / total;
+
+ /* elapsed time for this pass */
+ elapsed = time(NULL) - pres->pres_start_time;
+ elapsed = elapsed > 0 ? elapsed : 1;
+ rate = copied / elapsed;
+ rate = rate > 0 ? rate : 1;
+ secs_left = (total - copied) / rate;
+
+ zfs_nicenum(copied, examined_buf, sizeof (examined_buf));
+ zfs_nicenum(total, total_buf, sizeof (total_buf));
+ zfs_nicenum(rate, rate_buf, sizeof (rate_buf));
+
+ /*
+ * do not print estimated time if hours_left is more than
+ * 30 days
+ */
+ (void) printf(gettext("\t%s / %s copied at %s/s, %.2f%% done"),
+ examined_buf, total_buf, rate_buf, 100 * fraction_done);
+ if (pres->pres_waiting_for_resilver) {
+ (void) printf(gettext(", paused for resilver or "
+ "clear\n"));
+ } else if (secs_left < (30 * 24 * 3600)) {
+ char time_buf[32];
+ secs_to_dhms(secs_left, time_buf);
+ (void) printf(gettext(", %s to go\n"), time_buf);
+ } else {
+ (void) printf(gettext(
+ ", (copy is slow, no estimated time)\n"));
+ }
+ }
+ free(vname);
+}
static void
print_checkpoint_status(pool_checkpoint_stat_t *pcs)
{
time_t start;
char space_buf[7];
if (pcs == NULL || pcs->pcs_state == CS_NONE)
return;
(void) printf(gettext("checkpoint: "));
start = pcs->pcs_start_time;
zfs_nicenum(pcs->pcs_space, space_buf, sizeof (space_buf));
if (pcs->pcs_state == CS_CHECKPOINT_EXISTS) {
char *date = ctime(&start);
/*
* ctime() adds a newline at the end of the generated
* string, thus the weird format specifier and the
* strlen() call used to chop it off from the output.
*/
(void) printf(gettext("created %.*s, consumes %s\n"),
(int)(strlen(date) - 1), date, space_buf);
return;
}
assert(pcs->pcs_state == CS_CHECKPOINT_DISCARDING);
(void) printf(gettext("discarding, %s remaining.\n"),
space_buf);
}
static void
print_error_log(zpool_handle_t *zhp)
{
nvlist_t *nverrlist = NULL;
nvpair_t *elem;
char *pathname;
size_t len = MAXPATHLEN * 2;
if (zpool_get_errlog(zhp, &nverrlist) != 0)
return;
(void) printf("errors: Permanent errors have been "
"detected in the following files:\n\n");
pathname = safe_malloc(len);
elem = NULL;
while ((elem = nvlist_next_nvpair(nverrlist, elem)) != NULL) {
nvlist_t *nv;
uint64_t dsobj, obj;
verify(nvpair_value_nvlist(elem, &nv) == 0);
verify(nvlist_lookup_uint64(nv, ZPOOL_ERR_DATASET,
&dsobj) == 0);
verify(nvlist_lookup_uint64(nv, ZPOOL_ERR_OBJECT,
&obj) == 0);
zpool_obj_to_path(zhp, dsobj, obj, pathname, len);
(void) printf("%7s %s\n", "", pathname);
}
free(pathname);
nvlist_free(nverrlist);
}
static void
print_spares(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t **spares,
uint_t nspares)
{
uint_t i;
char *name;
if (nspares == 0)
return;
(void) printf(gettext("\tspares\n"));
for (i = 0; i < nspares; i++) {
name = zpool_vdev_name(g_zfs, zhp, spares[i],
cb->cb_name_flags);
print_status_config(zhp, cb, name, spares[i], 2, B_TRUE, NULL);
free(name);
}
}
static void
print_l2cache(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t **l2cache,
uint_t nl2cache)
{
uint_t i;
char *name;
if (nl2cache == 0)
return;
(void) printf(gettext("\tcache\n"));
for (i = 0; i < nl2cache; i++) {
name = zpool_vdev_name(g_zfs, zhp, l2cache[i],
cb->cb_name_flags);
print_status_config(zhp, cb, name, l2cache[i], 2,
B_FALSE, NULL);
free(name);
}
}
static void
print_dedup_stats(nvlist_t *config)
{
ddt_histogram_t *ddh;
ddt_stat_t *dds;
ddt_object_t *ddo;
uint_t c;
char dspace[6], mspace[6];
/*
* If the pool was faulted then we may not have been able to
* obtain the config. Otherwise, if we have anything in the dedup
* table continue processing the stats.
*/
if (nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_OBJ_STATS,
(uint64_t **)&ddo, &c) != 0)
return;
(void) printf("\n");
(void) printf(gettext(" dedup: "));
if (ddo->ddo_count == 0) {
(void) printf(gettext("no DDT entries\n"));
return;
}
zfs_nicebytes(ddo->ddo_dspace, dspace, sizeof (dspace));
zfs_nicebytes(ddo->ddo_mspace, mspace, sizeof (mspace));
(void) printf("DDT entries %llu, size %s on disk, %s in core\n",
(u_longlong_t)ddo->ddo_count,
dspace,
mspace);
verify(nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_STATS,
(uint64_t **)&dds, &c) == 0);
verify(nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_HISTOGRAM,
(uint64_t **)&ddh, &c) == 0);
zpool_dump_ddt(dds, ddh);
}
/*
* Display a summary of pool status. Displays a summary such as:
*
* pool: tank
* status: DEGRADED
* reason: One or more devices ...
* see: https://openzfs.github.io/openzfs-docs/msg/ZFS-xxxx-01
* config:
* mirror DEGRADED
* c1t0d0 OK
* c2t0d0 UNAVAIL
*
* When given the '-v' option, we print out the complete config. If the '-e'
* option is specified, then we print out error rate information as well.
*/
static int
status_callback(zpool_handle_t *zhp, void *data)
{
status_cbdata_t *cbp = data;
nvlist_t *config, *nvroot;
const char *msgid;
zpool_status_t reason;
zpool_errata_t errata;
const char *health;
uint_t c;
vdev_stat_t *vs;
config = zpool_get_config(zhp, NULL);
reason = zpool_get_status(zhp, &msgid, &errata);
cbp->cb_count++;
/*
* If we were given 'zpool status -x', only report those pools with
* problems.
*/
if (cbp->cb_explain &&
(reason == ZPOOL_STATUS_OK ||
reason == ZPOOL_STATUS_VERSION_OLDER ||
reason == ZPOOL_STATUS_FEAT_DISABLED ||
reason == ZPOOL_STATUS_COMPATIBILITY_ERR ||
reason == ZPOOL_STATUS_INCOMPATIBLE_FEAT)) {
if (!cbp->cb_allpools) {
(void) printf(gettext("pool '%s' is healthy\n"),
zpool_get_name(zhp));
if (cbp->cb_first)
cbp->cb_first = B_FALSE;
}
return (0);
}
if (cbp->cb_first)
cbp->cb_first = B_FALSE;
else
(void) printf("\n");
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &c) == 0);
health = zpool_get_state_str(zhp);
printf(" ");
printf_color(ANSI_BOLD, gettext("pool:"));
printf(" %s\n", zpool_get_name(zhp));
fputc(' ', stdout);
printf_color(ANSI_BOLD, gettext("state: "));
printf_color(health_str_to_color(health), "%s", health);
fputc('\n', stdout);
switch (reason) {
case ZPOOL_STATUS_MISSING_DEV_R:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices could "
"not be opened. Sufficient replicas exist for\n\tthe pool "
"to continue functioning in a degraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Attach the missing device "
"and online it using 'zpool online'.\n"));
break;
case ZPOOL_STATUS_MISSING_DEV_NR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices could "
"not be opened. There are insufficient\n\treplicas for the"
" pool to continue functioning.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Attach the missing device "
"and online it using 'zpool online'.\n"));
break;
case ZPOOL_STATUS_CORRUPT_LABEL_R:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices could "
"not be used because the label is missing or\n\tinvalid. "
"Sufficient replicas exist for the pool to continue\n\t"
"functioning in a degraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Replace the device using "
"'zpool replace'.\n"));
break;
case ZPOOL_STATUS_CORRUPT_LABEL_NR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices could "
"not be used because the label is missing \n\tor invalid. "
"There are insufficient replicas for the pool to "
"continue\n\tfunctioning.\n"));
zpool_explain_recover(zpool_get_handle(zhp),
zpool_get_name(zhp), reason, config);
break;
case ZPOOL_STATUS_FAILING_DEV:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices has "
"experienced an unrecoverable error. An\n\tattempt was "
"made to correct the error. Applications are "
"unaffected.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Determine if the "
"device needs to be replaced, and clear the errors\n\tusing"
" 'zpool clear' or replace the device with 'zpool "
"replace'.\n"));
break;
case ZPOOL_STATUS_OFFLINE_DEV:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices has "
"been taken offline by the administrator.\n\tSufficient "
"replicas exist for the pool to continue functioning in "
"a\n\tdegraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Online the device "
"using 'zpool online' or replace the device with\n\t'zpool "
"replace'.\n"));
break;
case ZPOOL_STATUS_REMOVED_DEV:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices has "
"been removed by the administrator.\n\tSufficient "
"replicas exist for the pool to continue functioning in "
"a\n\tdegraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Online the device "
"using zpool online' or replace the device with\n\t'zpool "
"replace'.\n"));
break;
case ZPOOL_STATUS_RESILVERING:
case ZPOOL_STATUS_REBUILDING:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices is "
"currently being resilvered. The pool will\n\tcontinue "
"to function, possibly in a degraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Wait for the resilver to "
"complete.\n"));
break;
case ZPOOL_STATUS_REBUILD_SCRUB:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices have "
"been sequentially resilvered, scrubbing\n\tthe pool "
"is recommended.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Use 'zpool scrub' to "
"verify all data checksums.\n"));
break;
case ZPOOL_STATUS_CORRUPT_DATA:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices has "
"experienced an error resulting in data\n\tcorruption. "
"Applications may be affected.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Restore the file in question"
" if possible. Otherwise restore the\n\tentire pool from "
"backup.\n"));
break;
case ZPOOL_STATUS_CORRUPT_POOL:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool metadata is "
"corrupted and the pool cannot be opened.\n"));
zpool_explain_recover(zpool_get_handle(zhp),
zpool_get_name(zhp), reason, config);
break;
case ZPOOL_STATUS_VERSION_OLDER:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool is formatted using "
"a legacy on-disk format. The pool can\n\tstill be used, "
"but some features are unavailable.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Upgrade the pool using "
"'zpool upgrade'. Once this is done, the\n\tpool will no "
"longer be accessible on software that does not support\n\t"
"feature flags.\n"));
break;
case ZPOOL_STATUS_VERSION_NEWER:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool has been upgraded "
"to a newer, incompatible on-disk version.\n\tThe pool "
"cannot be accessed on this system.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Access the pool from a "
"system running more recent software, or\n\trestore the "
"pool from backup.\n"));
break;
case ZPOOL_STATUS_FEAT_DISABLED:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Some supported and "
"requested features are not enabled on the pool.\n\t"
"The pool can still be used, but some features are "
"unavailable.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Enable all features using "
"'zpool upgrade'. Once this is done,\n\tthe pool may no "
"longer be accessible by software that does not support\n\t"
"the features. See zpool-features(7) for details.\n"));
break;
case ZPOOL_STATUS_COMPATIBILITY_ERR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("This pool has a "
"compatibility list specified, but it could not be\n\t"
"read/parsed at this time. The pool can still be used, "
"but this\n\tshould be investigated.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Check the value of the "
"'compatibility' property against the\n\t"
"appropriate file in " ZPOOL_SYSCONF_COMPAT_D " or "
ZPOOL_DATA_COMPAT_D ".\n"));
break;
case ZPOOL_STATUS_INCOMPATIBLE_FEAT:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more features "
"are enabled on the pool despite not being\n\t"
"requested by the 'compatibility' property.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Consider setting "
"'compatibility' to an appropriate value, or\n\t"
"adding needed features to the relevant file in\n\t"
ZPOOL_SYSCONF_COMPAT_D " or " ZPOOL_DATA_COMPAT_D ".\n"));
break;
case ZPOOL_STATUS_UNSUP_FEAT_READ:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool cannot be accessed "
"on this system because it uses the\n\tfollowing feature(s)"
" not supported on this system:\n"));
zpool_print_unsup_feat(config);
(void) printf("\n");
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Access the pool from a "
"system that supports the required feature(s),\n\tor "
"restore the pool from backup.\n"));
break;
case ZPOOL_STATUS_UNSUP_FEAT_WRITE:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool can only be "
"accessed in read-only mode on this system. It\n\tcannot be"
" accessed in read-write mode because it uses the "
"following\n\tfeature(s) not supported on this system:\n"));
zpool_print_unsup_feat(config);
(void) printf("\n");
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("The pool cannot be accessed "
"in read-write mode. Import the pool with\n"
"\t\"-o readonly=on\", access the pool from a system that "
"supports the\n\trequired feature(s), or restore the "
"pool from backup.\n"));
break;
case ZPOOL_STATUS_FAULTED_DEV_R:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"faulted in response to persistent errors.\n\tSufficient "
"replicas exist for the pool to continue functioning "
"in a\n\tdegraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Replace the faulted device, "
"or use 'zpool clear' to mark the device\n\trepaired.\n"));
break;
case ZPOOL_STATUS_FAULTED_DEV_NR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"faulted in response to persistent errors. There are "
"insufficient replicas for the pool to\n\tcontinue "
"functioning.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Destroy and re-create the "
"pool from a backup source. Manually marking the device\n"
"\trepaired using 'zpool clear' may allow some data "
"to be recovered.\n"));
break;
case ZPOOL_STATUS_IO_FAILURE_MMP:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool is suspended "
"because multihost writes failed or were delayed;\n\t"
"another system could import the pool undetected.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Make sure the pool's devices"
" are connected, then reboot your system and\n\timport the "
"pool.\n"));
break;
case ZPOOL_STATUS_IO_FAILURE_WAIT:
case ZPOOL_STATUS_IO_FAILURE_CONTINUE:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"faulted in response to IO failures.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Make sure the affected "
"devices are connected, then run 'zpool clear'.\n"));
break;
case ZPOOL_STATUS_BAD_LOG:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("An intent log record "
"could not be read.\n"
"\tWaiting for administrator intervention to fix the "
"faulted pool.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Either restore the affected "
"device(s) and run 'zpool online',\n"
"\tor ignore the intent log records by running "
"'zpool clear'.\n"));
break;
case ZPOOL_STATUS_NON_NATIVE_ASHIFT:
(void) printf(gettext("status: One or more devices are "
"configured to use a non-native block size.\n"
"\tExpect reduced performance.\n"));
(void) printf(gettext("action: Replace affected devices with "
"devices that support the\n\tconfigured block size, or "
"migrate data to a properly configured\n\tpool.\n"));
break;
case ZPOOL_STATUS_HOSTID_MISMATCH:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Mismatch between pool hostid"
" and system hostid on imported pool.\n\tThis pool was "
"previously imported into a system with a different "
"hostid,\n\tand then was verbatim imported into this "
"system.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Export this pool on all "
"systems on which it is imported.\n"
"\tThen import it to correct the mismatch.\n"));
break;
case ZPOOL_STATUS_ERRATA:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Errata #%d detected.\n"),
errata);
switch (errata) {
case ZPOOL_ERRATA_NONE:
break;
case ZPOOL_ERRATA_ZOL_2094_SCRUB:
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("To correct the issue"
" run 'zpool scrub'.\n"));
break;
case ZPOOL_ERRATA_ZOL_6845_ENCRYPTION:
(void) printf(gettext("\tExisting encrypted datasets "
"contain an on-disk incompatibility\n\twhich "
"needs to be corrected.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("To correct the issue"
" backup existing encrypted datasets to new\n\t"
"encrypted datasets and destroy the old ones. "
"'zfs mount -o ro' can\n\tbe used to temporarily "
"mount existing encrypted datasets readonly.\n"));
break;
case ZPOOL_ERRATA_ZOL_8308_ENCRYPTION:
(void) printf(gettext("\tExisting encrypted snapshots "
"and bookmarks contain an on-disk\n\tincompat"
"ibility. This may cause on-disk corruption if "
"they are used\n\twith 'zfs recv'.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("To correct the"
"issue, enable the bookmark_v2 feature. No "
"additional\n\taction is needed if there are no "
"encrypted snapshots or bookmarks.\n\tIf preserving"
"the encrypted snapshots and bookmarks is required,"
" use\n\ta non-raw send to backup and restore them."
" Alternately, they may be\n\tremoved to resolve "
"the incompatibility.\n"));
break;
default:
/*
* All errata which allow the pool to be imported
* must contain an action message.
*/
assert(0);
}
break;
default:
/*
* The remaining errors can't actually be generated, yet.
*/
assert(reason == ZPOOL_STATUS_OK);
}
if (msgid != NULL) {
printf(" ");
printf_color(ANSI_BOLD, gettext("see:"));
printf(gettext(
" https://openzfs.github.io/openzfs-docs/msg/%s\n"),
msgid);
}
if (config != NULL) {
uint64_t nerr;
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
- pool_checkpoint_stat_t *pcs = NULL;
- pool_removal_stat_t *prs = NULL;
print_scan_status(zhp, nvroot);
+ pool_removal_stat_t *prs = NULL;
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t **)&prs, &c);
print_removal_status(zhp, prs);
+ pool_checkpoint_stat_t *pcs = NULL;
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c);
print_checkpoint_status(pcs);
+ pool_raidz_expand_stat_t *pres = NULL;
+ (void) nvlist_lookup_uint64_array(nvroot,
+ ZPOOL_CONFIG_RAIDZ_EXPAND_STATS, (uint64_t **)&pres, &c);
+ print_raidz_expand_status(zhp, pres);
+
cbp->cb_namewidth = max_width(zhp, nvroot, 0, 0,
cbp->cb_name_flags | VDEV_NAME_TYPE_ID);
if (cbp->cb_namewidth < 10)
cbp->cb_namewidth = 10;
color_start(ANSI_BOLD);
(void) printf(gettext("config:\n\n"));
(void) printf(gettext("\t%-*s %-8s %5s %5s %5s"),
cbp->cb_namewidth, "NAME", "STATE", "READ", "WRITE",
"CKSUM");
color_end();
if (cbp->cb_print_slow_ios) {
printf_color(ANSI_BOLD, " %5s", gettext("SLOW"));
}
if (cbp->vcdl != NULL)
print_cmd_columns(cbp->vcdl, 0);
printf("\n");
print_status_config(zhp, cbp, zpool_get_name(zhp), nvroot, 0,
B_FALSE, NULL);
print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_BIAS_DEDUP);
print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_BIAS_SPECIAL);
print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_CLASS_LOGS);
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache) == 0)
print_l2cache(zhp, cbp, l2cache, nl2cache);
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares) == 0)
print_spares(zhp, cbp, spares, nspares);
if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT,
&nerr) == 0) {
(void) printf("\n");
if (nerr == 0) {
(void) printf(gettext(
"errors: No known data errors\n"));
} else if (!cbp->cb_verbose) {
(void) printf(gettext("errors: %llu data "
"errors, use '-v' for a list\n"),
(u_longlong_t)nerr);
} else {
print_error_log(zhp);
}
}
if (cbp->cb_dedup_stats)
print_dedup_stats(config);
} else {
(void) printf(gettext("config: The configuration cannot be "
"determined.\n"));
}
return (0);
}
/*
* zpool status [-c [script1,script2,...]] [-igLpPstvx] [-T d|u] [pool] ...
* [interval [count]]
*
* -c CMD For each vdev, run command CMD
* -i Display vdev initialization status.
* -g Display guid for individual vdev name.
* -L Follow links when resolving vdev path name.
* -p Display values in parsable (exact) format.
* -P Display full path for vdev name.
* -s Display slow IOs column.
* -v Display complete error logs
* -x Display only pools with potential problems
* -D Display dedup status (undocumented)
* -t Display vdev TRIM status.
* -T Display a timestamp in date(1) or Unix format
*
* Describes the health status of all pools or some subset.
*/
int
zpool_do_status(int argc, char **argv)
{
int c;
int ret;
float interval = 0;
unsigned long count = 0;
status_cbdata_t cb = { 0 };
char *cmd = NULL;
/* check options */
while ((c = getopt(argc, argv, "c:igLpPsvxDtT:")) != -1) {
switch (c) {
case 'c':
if (cmd != NULL) {
fprintf(stderr,
gettext("Can't set -c flag twice\n"));
exit(1);
}
if (getenv("ZPOOL_SCRIPTS_ENABLED") != NULL &&
!libzfs_envvar_is_set("ZPOOL_SCRIPTS_ENABLED")) {
fprintf(stderr, gettext(
"Can't run -c, disabled by "
"ZPOOL_SCRIPTS_ENABLED.\n"));
exit(1);
}
if ((getuid() <= 0 || geteuid() <= 0) &&
!libzfs_envvar_is_set("ZPOOL_SCRIPTS_AS_ROOT")) {
fprintf(stderr, gettext(
"Can't run -c with root privileges "
"unless ZPOOL_SCRIPTS_AS_ROOT is set.\n"));
exit(1);
}
cmd = optarg;
break;
case 'i':
cb.cb_print_vdev_init = B_TRUE;
break;
case 'g':
cb.cb_name_flags |= VDEV_NAME_GUID;
break;
case 'L':
cb.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS;
break;
case 'p':
cb.cb_literal = B_TRUE;
break;
case 'P':
cb.cb_name_flags |= VDEV_NAME_PATH;
break;
case 's':
cb.cb_print_slow_ios = B_TRUE;
break;
case 'v':
cb.cb_verbose = B_TRUE;
break;
case 'x':
cb.cb_explain = B_TRUE;
break;
case 'D':
cb.cb_dedup_stats = B_TRUE;
break;
case 't':
cb.cb_print_vdev_trim = B_TRUE;
break;
case 'T':
get_timestamp_arg(*optarg);
break;
case '?':
if (optopt == 'c') {
print_zpool_script_list("status");
exit(0);
} else {
fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
}
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
get_interval_count(&argc, argv, &interval, &count);
if (argc == 0)
cb.cb_allpools = B_TRUE;
cb.cb_first = B_TRUE;
cb.cb_print_status = B_TRUE;
for (;;) {
if (timestamp_fmt != NODATE)
print_timestamp(timestamp_fmt);
if (cmd != NULL)
cb.vcdl = all_pools_for_each_vdev_run(argc, argv, cmd,
NULL, NULL, 0, 0);
ret = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
cb.cb_literal, status_callback, &cb);
if (cb.vcdl != NULL)
free_vdev_cmd_data_list(cb.vcdl);
if (argc == 0 && cb.cb_count == 0)
(void) fprintf(stderr, gettext("no pools available\n"));
else if (cb.cb_explain && cb.cb_first && cb.cb_allpools)
(void) printf(gettext("all pools are healthy\n"));
if (ret != 0)
return (ret);
if (interval == 0)
break;
if (count != 0 && --count == 0)
break;
(void) fsleep(interval);
}
return (0);
}
typedef struct upgrade_cbdata {
int cb_first;
int cb_argc;
uint64_t cb_version;
char **cb_argv;
} upgrade_cbdata_t;
static int
check_unsupp_fs(zfs_handle_t *zhp, void *unsupp_fs)
{
int zfs_version = (int)zfs_prop_get_int(zhp, ZFS_PROP_VERSION);
int *count = (int *)unsupp_fs;
if (zfs_version > ZPL_VERSION) {
(void) printf(gettext("%s (v%d) is not supported by this "
"implementation of ZFS.\n"),
zfs_get_name(zhp), zfs_version);
(*count)++;
}
zfs_iter_filesystems_v2(zhp, 0, check_unsupp_fs, unsupp_fs);
zfs_close(zhp);
return (0);
}
static int
upgrade_version(zpool_handle_t *zhp, uint64_t version)
{
int ret;
nvlist_t *config;
uint64_t oldversion;
int unsupp_fs = 0;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
&oldversion) == 0);
char compat[ZFS_MAXPROPLEN];
if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compat,
ZFS_MAXPROPLEN, NULL, B_FALSE) != 0)
compat[0] = '\0';
assert(SPA_VERSION_IS_SUPPORTED(oldversion));
assert(oldversion < version);
ret = zfs_iter_root(zpool_get_handle(zhp), check_unsupp_fs, &unsupp_fs);
if (ret != 0)
return (ret);
if (unsupp_fs) {
(void) fprintf(stderr, gettext("Upgrade not performed due "
"to %d unsupported filesystems (max v%d).\n"),
unsupp_fs, (int)ZPL_VERSION);
return (1);
}
if (strcmp(compat, ZPOOL_COMPAT_LEGACY) == 0) {
(void) fprintf(stderr, gettext("Upgrade not performed because "
"'compatibility' property set to '"
ZPOOL_COMPAT_LEGACY "'.\n"));
return (1);
}
ret = zpool_upgrade(zhp, version);
if (ret != 0)
return (ret);
if (version >= SPA_VERSION_FEATURES) {
(void) printf(gettext("Successfully upgraded "
"'%s' from version %llu to feature flags.\n"),
zpool_get_name(zhp), (u_longlong_t)oldversion);
} else {
(void) printf(gettext("Successfully upgraded "
"'%s' from version %llu to version %llu.\n"),
zpool_get_name(zhp), (u_longlong_t)oldversion,
(u_longlong_t)version);
}
return (0);
}
static int
upgrade_enable_all(zpool_handle_t *zhp, int *countp)
{
int i, ret, count;
boolean_t firstff = B_TRUE;
nvlist_t *enabled = zpool_get_features(zhp);
char compat[ZFS_MAXPROPLEN];
if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compat,
ZFS_MAXPROPLEN, NULL, B_FALSE) != 0)
compat[0] = '\0';
boolean_t requested_features[SPA_FEATURES];
if (zpool_do_load_compat(compat, requested_features) !=
ZPOOL_COMPATIBILITY_OK)
return (-1);
count = 0;
for (i = 0; i < SPA_FEATURES; i++) {
const char *fname = spa_feature_table[i].fi_uname;
const char *fguid = spa_feature_table[i].fi_guid;
if (!spa_feature_table[i].fi_zfs_mod_supported)
continue;
if (!nvlist_exists(enabled, fguid) && requested_features[i]) {
char *propname;
verify(-1 != asprintf(&propname, "feature@%s", fname));
ret = zpool_set_prop(zhp, propname,
ZFS_FEATURE_ENABLED);
if (ret != 0) {
free(propname);
return (ret);
}
count++;
if (firstff) {
(void) printf(gettext("Enabled the "
"following features on '%s':\n"),
zpool_get_name(zhp));
firstff = B_FALSE;
}
(void) printf(gettext(" %s\n"), fname);
free(propname);
}
}
if (countp != NULL)
*countp = count;
return (0);
}
static int
upgrade_cb(zpool_handle_t *zhp, void *arg)
{
upgrade_cbdata_t *cbp = arg;
nvlist_t *config;
uint64_t version;
boolean_t modified_pool = B_FALSE;
int ret;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
&version) == 0);
assert(SPA_VERSION_IS_SUPPORTED(version));
if (version < cbp->cb_version) {
cbp->cb_first = B_FALSE;
ret = upgrade_version(zhp, cbp->cb_version);
if (ret != 0)
return (ret);
modified_pool = B_TRUE;
/*
* If they did "zpool upgrade -a", then we could
* be doing ioctls to different pools. We need
* to log this history once to each pool, and bypass
* the normal history logging that happens in main().
*/
(void) zpool_log_history(g_zfs, history_str);
log_history = B_FALSE;
}
if (cbp->cb_version >= SPA_VERSION_FEATURES) {
int count;
ret = upgrade_enable_all(zhp, &count);
if (ret != 0)
return (ret);
if (count > 0) {
cbp->cb_first = B_FALSE;
modified_pool = B_TRUE;
}
}
if (modified_pool) {
(void) printf("\n");
(void) after_zpool_upgrade(zhp);
}
return (0);
}
static int
upgrade_list_older_cb(zpool_handle_t *zhp, void *arg)
{
upgrade_cbdata_t *cbp = arg;
nvlist_t *config;
uint64_t version;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
&version) == 0);
assert(SPA_VERSION_IS_SUPPORTED(version));
if (version < SPA_VERSION_FEATURES) {
if (cbp->cb_first) {
(void) printf(gettext("The following pools are "
"formatted with legacy version numbers and can\n"
"be upgraded to use feature flags. After "
"being upgraded, these pools\nwill no "
"longer be accessible by software that does not "
"support feature\nflags.\n\n"
"Note that setting a pool's 'compatibility' "
"feature to '" ZPOOL_COMPAT_LEGACY "' will\n"
"inhibit upgrades.\n\n"));
(void) printf(gettext("VER POOL\n"));
(void) printf(gettext("--- ------------\n"));
cbp->cb_first = B_FALSE;
}
(void) printf("%2llu %s\n", (u_longlong_t)version,
zpool_get_name(zhp));
}
return (0);
}
static int
upgrade_list_disabled_cb(zpool_handle_t *zhp, void *arg)
{
upgrade_cbdata_t *cbp = arg;
nvlist_t *config;
uint64_t version;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
&version) == 0);
if (version >= SPA_VERSION_FEATURES) {
int i;
boolean_t poolfirst = B_TRUE;
nvlist_t *enabled = zpool_get_features(zhp);
for (i = 0; i < SPA_FEATURES; i++) {
const char *fguid = spa_feature_table[i].fi_guid;
const char *fname = spa_feature_table[i].fi_uname;
if (!spa_feature_table[i].fi_zfs_mod_supported)
continue;
if (!nvlist_exists(enabled, fguid)) {
if (cbp->cb_first) {
(void) printf(gettext("\nSome "
"supported features are not "
"enabled on the following pools. "
"Once a\nfeature is enabled the "
"pool may become incompatible with "
"software\nthat does not support "
"the feature. See "
"zpool-features(7) for "
"details.\n\n"
"Note that the pool "
"'compatibility' feature can be "
"used to inhibit\nfeature "
"upgrades.\n\n"));
(void) printf(gettext("POOL "
"FEATURE\n"));
(void) printf(gettext("------"
"---------\n"));
cbp->cb_first = B_FALSE;
}
if (poolfirst) {
(void) printf(gettext("%s\n"),
zpool_get_name(zhp));
poolfirst = B_FALSE;
}
(void) printf(gettext(" %s\n"), fname);
}
/*
* If they did "zpool upgrade -a", then we could
* be doing ioctls to different pools. We need
* to log this history once to each pool, and bypass
* the normal history logging that happens in main().
*/
(void) zpool_log_history(g_zfs, history_str);
log_history = B_FALSE;
}
}
return (0);
}
static int
upgrade_one(zpool_handle_t *zhp, void *data)
{
boolean_t modified_pool = B_FALSE;
upgrade_cbdata_t *cbp = data;
uint64_t cur_version;
int ret;
if (strcmp("log", zpool_get_name(zhp)) == 0) {
(void) fprintf(stderr, gettext("'log' is now a reserved word\n"
"Pool 'log' must be renamed using export and import"
" to upgrade.\n"));
return (1);
}
cur_version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL);
if (cur_version > cbp->cb_version) {
(void) printf(gettext("Pool '%s' is already formatted "
"using more current version '%llu'.\n\n"),
zpool_get_name(zhp), (u_longlong_t)cur_version);
return (0);
}
if (cbp->cb_version != SPA_VERSION && cur_version == cbp->cb_version) {
(void) printf(gettext("Pool '%s' is already formatted "
"using version %llu.\n\n"), zpool_get_name(zhp),
(u_longlong_t)cbp->cb_version);
return (0);
}
if (cur_version != cbp->cb_version) {
modified_pool = B_TRUE;
ret = upgrade_version(zhp, cbp->cb_version);
if (ret != 0)
return (ret);
}
if (cbp->cb_version >= SPA_VERSION_FEATURES) {
int count = 0;
ret = upgrade_enable_all(zhp, &count);
if (ret != 0)
return (ret);
if (count != 0) {
modified_pool = B_TRUE;
} else if (cur_version == SPA_VERSION) {
(void) printf(gettext("Pool '%s' already has all "
"supported and requested features enabled.\n"),
zpool_get_name(zhp));
}
}
if (modified_pool) {
(void) printf("\n");
(void) after_zpool_upgrade(zhp);
}
return (0);
}
/*
* zpool upgrade
* zpool upgrade -v
* zpool upgrade [-V version] <-a | pool ...>
*
* With no arguments, display downrev'd ZFS pool available for upgrade.
* Individual pools can be upgraded by specifying the pool, and '-a' will
* upgrade all pools.
*/
int
zpool_do_upgrade(int argc, char **argv)
{
int c;
upgrade_cbdata_t cb = { 0 };
int ret = 0;
boolean_t showversions = B_FALSE;
boolean_t upgradeall = B_FALSE;
char *end;
/* check options */
while ((c = getopt(argc, argv, ":avV:")) != -1) {
switch (c) {
case 'a':
upgradeall = B_TRUE;
break;
case 'v':
showversions = B_TRUE;
break;
case 'V':
cb.cb_version = strtoll(optarg, &end, 10);
if (*end != '\0' ||
!SPA_VERSION_IS_SUPPORTED(cb.cb_version)) {
(void) fprintf(stderr,
gettext("invalid version '%s'\n"), optarg);
usage(B_FALSE);
}
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
cb.cb_argc = argc;
cb.cb_argv = argv;
argc -= optind;
argv += optind;
if (cb.cb_version == 0) {
cb.cb_version = SPA_VERSION;
} else if (!upgradeall && argc == 0) {
(void) fprintf(stderr, gettext("-V option is "
"incompatible with other arguments\n"));
usage(B_FALSE);
}
if (showversions) {
if (upgradeall || argc != 0) {
(void) fprintf(stderr, gettext("-v option is "
"incompatible with other arguments\n"));
usage(B_FALSE);
}
} else if (upgradeall) {
if (argc != 0) {
(void) fprintf(stderr, gettext("-a option should not "
"be used along with a pool name\n"));
usage(B_FALSE);
}
}
(void) printf("%s", gettext("This system supports ZFS pool feature "
"flags.\n\n"));
if (showversions) {
int i;
(void) printf(gettext("The following features are "
"supported:\n\n"));
(void) printf(gettext("FEAT DESCRIPTION\n"));
(void) printf("----------------------------------------------"
"---------------\n");
for (i = 0; i < SPA_FEATURES; i++) {
zfeature_info_t *fi = &spa_feature_table[i];
if (!fi->fi_zfs_mod_supported)
continue;
const char *ro =
(fi->fi_flags & ZFEATURE_FLAG_READONLY_COMPAT) ?
" (read-only compatible)" : "";
(void) printf("%-37s%s\n", fi->fi_uname, ro);
(void) printf(" %s\n", fi->fi_desc);
}
(void) printf("\n");
(void) printf(gettext("The following legacy versions are also "
"supported:\n\n"));
(void) printf(gettext("VER DESCRIPTION\n"));
(void) printf("--- -----------------------------------------"
"---------------\n");
(void) printf(gettext(" 1 Initial ZFS version\n"));
(void) printf(gettext(" 2 Ditto blocks "
"(replicated metadata)\n"));
(void) printf(gettext(" 3 Hot spares and double parity "
"RAID-Z\n"));
(void) printf(gettext(" 4 zpool history\n"));
(void) printf(gettext(" 5 Compression using the gzip "
"algorithm\n"));
(void) printf(gettext(" 6 bootfs pool property\n"));
(void) printf(gettext(" 7 Separate intent log devices\n"));
(void) printf(gettext(" 8 Delegated administration\n"));
(void) printf(gettext(" 9 refquota and refreservation "
"properties\n"));
(void) printf(gettext(" 10 Cache devices\n"));
(void) printf(gettext(" 11 Improved scrub performance\n"));
(void) printf(gettext(" 12 Snapshot properties\n"));
(void) printf(gettext(" 13 snapused property\n"));
(void) printf(gettext(" 14 passthrough-x aclinherit\n"));
(void) printf(gettext(" 15 user/group space accounting\n"));
(void) printf(gettext(" 16 stmf property support\n"));
(void) printf(gettext(" 17 Triple-parity RAID-Z\n"));
(void) printf(gettext(" 18 Snapshot user holds\n"));
(void) printf(gettext(" 19 Log device removal\n"));
(void) printf(gettext(" 20 Compression using zle "
"(zero-length encoding)\n"));
(void) printf(gettext(" 21 Deduplication\n"));
(void) printf(gettext(" 22 Received properties\n"));
(void) printf(gettext(" 23 Slim ZIL\n"));
(void) printf(gettext(" 24 System attributes\n"));
(void) printf(gettext(" 25 Improved scrub stats\n"));
(void) printf(gettext(" 26 Improved snapshot deletion "
"performance\n"));
(void) printf(gettext(" 27 Improved snapshot creation "
"performance\n"));
(void) printf(gettext(" 28 Multiple vdev replacements\n"));
(void) printf(gettext("\nFor more information on a particular "
"version, including supported releases,\n"));
(void) printf(gettext("see the ZFS Administration Guide.\n\n"));
} else if (argc == 0 && upgradeall) {
cb.cb_first = B_TRUE;
ret = zpool_iter(g_zfs, upgrade_cb, &cb);
if (ret == 0 && cb.cb_first) {
if (cb.cb_version == SPA_VERSION) {
(void) printf(gettext("All pools are already "
"formatted using feature flags.\n\n"));
(void) printf(gettext("Every feature flags "
"pool already has all supported and "
"requested features enabled.\n"));
} else {
(void) printf(gettext("All pools are already "
"formatted with version %llu or higher.\n"),
(u_longlong_t)cb.cb_version);
}
}
} else if (argc == 0) {
cb.cb_first = B_TRUE;
ret = zpool_iter(g_zfs, upgrade_list_older_cb, &cb);
assert(ret == 0);
if (cb.cb_first) {
(void) printf(gettext("All pools are formatted "
"using feature flags.\n\n"));
} else {
(void) printf(gettext("\nUse 'zpool upgrade -v' "
"for a list of available legacy versions.\n"));
}
cb.cb_first = B_TRUE;
ret = zpool_iter(g_zfs, upgrade_list_disabled_cb, &cb);
assert(ret == 0);
if (cb.cb_first) {
(void) printf(gettext("Every feature flags pool has "
"all supported and requested features enabled.\n"));
} else {
(void) printf(gettext("\n"));
}
} else {
ret = for_each_pool(argc, argv, B_FALSE, NULL, ZFS_TYPE_POOL,
B_FALSE, upgrade_one, &cb);
}
return (ret);
}
typedef struct hist_cbdata {
boolean_t first;
boolean_t longfmt;
boolean_t internal;
} hist_cbdata_t;
static void
print_history_records(nvlist_t *nvhis, hist_cbdata_t *cb)
{
nvlist_t **records;
uint_t numrecords;
int i;
verify(nvlist_lookup_nvlist_array(nvhis, ZPOOL_HIST_RECORD,
&records, &numrecords) == 0);
for (i = 0; i < numrecords; i++) {
nvlist_t *rec = records[i];
char tbuf[64] = "";
if (nvlist_exists(rec, ZPOOL_HIST_TIME)) {
time_t tsec;
struct tm t;
tsec = fnvlist_lookup_uint64(records[i],
ZPOOL_HIST_TIME);
(void) localtime_r(&tsec, &t);
(void) strftime(tbuf, sizeof (tbuf), "%F.%T", &t);
}
if (nvlist_exists(rec, ZPOOL_HIST_ELAPSED_NS)) {
uint64_t elapsed_ns = fnvlist_lookup_int64(records[i],
ZPOOL_HIST_ELAPSED_NS);
(void) snprintf(tbuf + strlen(tbuf),
sizeof (tbuf) - strlen(tbuf),
" (%lldms)", (long long)elapsed_ns / 1000 / 1000);
}
if (nvlist_exists(rec, ZPOOL_HIST_CMD)) {
(void) printf("%s %s", tbuf,
fnvlist_lookup_string(rec, ZPOOL_HIST_CMD));
} else if (nvlist_exists(rec, ZPOOL_HIST_INT_EVENT)) {
int ievent =
fnvlist_lookup_uint64(rec, ZPOOL_HIST_INT_EVENT);
if (!cb->internal)
continue;
if (ievent >= ZFS_NUM_LEGACY_HISTORY_EVENTS) {
(void) printf("%s unrecognized record:\n",
tbuf);
dump_nvlist(rec, 4);
continue;
}
(void) printf("%s [internal %s txg:%lld] %s", tbuf,
zfs_history_event_names[ievent],
(longlong_t)fnvlist_lookup_uint64(
rec, ZPOOL_HIST_TXG),
fnvlist_lookup_string(rec, ZPOOL_HIST_INT_STR));
} else if (nvlist_exists(rec, ZPOOL_HIST_INT_NAME)) {
if (!cb->internal)
continue;
(void) printf("%s [txg:%lld] %s", tbuf,
(longlong_t)fnvlist_lookup_uint64(
rec, ZPOOL_HIST_TXG),
fnvlist_lookup_string(rec, ZPOOL_HIST_INT_NAME));
if (nvlist_exists(rec, ZPOOL_HIST_DSNAME)) {
(void) printf(" %s (%llu)",
fnvlist_lookup_string(rec,
ZPOOL_HIST_DSNAME),
(u_longlong_t)fnvlist_lookup_uint64(rec,
ZPOOL_HIST_DSID));
}
(void) printf(" %s", fnvlist_lookup_string(rec,
ZPOOL_HIST_INT_STR));
} else if (nvlist_exists(rec, ZPOOL_HIST_IOCTL)) {
if (!cb->internal)
continue;
(void) printf("%s ioctl %s\n", tbuf,
fnvlist_lookup_string(rec, ZPOOL_HIST_IOCTL));
if (nvlist_exists(rec, ZPOOL_HIST_INPUT_NVL)) {
(void) printf(" input:\n");
dump_nvlist(fnvlist_lookup_nvlist(rec,
ZPOOL_HIST_INPUT_NVL), 8);
}
if (nvlist_exists(rec, ZPOOL_HIST_OUTPUT_NVL)) {
(void) printf(" output:\n");
dump_nvlist(fnvlist_lookup_nvlist(rec,
ZPOOL_HIST_OUTPUT_NVL), 8);
}
if (nvlist_exists(rec, ZPOOL_HIST_OUTPUT_SIZE)) {
(void) printf(" output nvlist omitted; "
"original size: %lldKB\n",
(longlong_t)fnvlist_lookup_int64(rec,
ZPOOL_HIST_OUTPUT_SIZE) / 1024);
}
if (nvlist_exists(rec, ZPOOL_HIST_ERRNO)) {
(void) printf(" errno: %lld\n",
(longlong_t)fnvlist_lookup_int64(rec,
ZPOOL_HIST_ERRNO));
}
} else {
if (!cb->internal)
continue;
(void) printf("%s unrecognized record:\n", tbuf);
dump_nvlist(rec, 4);
}
if (!cb->longfmt) {
(void) printf("\n");
continue;
}
(void) printf(" [");
if (nvlist_exists(rec, ZPOOL_HIST_WHO)) {
uid_t who = fnvlist_lookup_uint64(rec, ZPOOL_HIST_WHO);
struct passwd *pwd = getpwuid(who);
(void) printf("user %d ", (int)who);
if (pwd != NULL)
(void) printf("(%s) ", pwd->pw_name);
}
if (nvlist_exists(rec, ZPOOL_HIST_HOST)) {
(void) printf("on %s",
fnvlist_lookup_string(rec, ZPOOL_HIST_HOST));
}
if (nvlist_exists(rec, ZPOOL_HIST_ZONE)) {
(void) printf(":%s",
fnvlist_lookup_string(rec, ZPOOL_HIST_ZONE));
}
(void) printf("]");
(void) printf("\n");
}
}
/*
* Print out the command history for a specific pool.
*/
static int
get_history_one(zpool_handle_t *zhp, void *data)
{
nvlist_t *nvhis;
int ret;
hist_cbdata_t *cb = (hist_cbdata_t *)data;
uint64_t off = 0;
boolean_t eof = B_FALSE;
cb->first = B_FALSE;
(void) printf(gettext("History for '%s':\n"), zpool_get_name(zhp));
while (!eof) {
if ((ret = zpool_get_history(zhp, &nvhis, &off, &eof)) != 0)
return (ret);
print_history_records(nvhis, cb);
nvlist_free(nvhis);
}
(void) printf("\n");
return (ret);
}
/*
* zpool history <pool>
*
* Displays the history of commands that modified pools.
*/
int
zpool_do_history(int argc, char **argv)
{
hist_cbdata_t cbdata = { 0 };
int ret;
int c;
cbdata.first = B_TRUE;
/* check options */
while ((c = getopt(argc, argv, "li")) != -1) {
switch (c) {
case 'l':
cbdata.longfmt = B_TRUE;
break;
case 'i':
cbdata.internal = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
ret = for_each_pool(argc, argv, B_FALSE, NULL, ZFS_TYPE_POOL,
B_FALSE, get_history_one, &cbdata);
if (argc == 0 && cbdata.first == B_TRUE) {
(void) fprintf(stderr, gettext("no pools available\n"));
return (0);
}
return (ret);
}
typedef struct ev_opts {
int verbose;
int scripted;
int follow;
int clear;
char poolname[ZFS_MAX_DATASET_NAME_LEN];
} ev_opts_t;
static void
zpool_do_events_short(nvlist_t *nvl, ev_opts_t *opts)
{
char ctime_str[26], str[32];
const char *ptr;
int64_t *tv;
uint_t n;
verify(nvlist_lookup_int64_array(nvl, FM_EREPORT_TIME, &tv, &n) == 0);
memset(str, ' ', 32);
(void) ctime_r((const time_t *)&tv[0], ctime_str);
(void) memcpy(str, ctime_str+4, 6); /* 'Jun 30' */
(void) memcpy(str+7, ctime_str+20, 4); /* '1993' */
(void) memcpy(str+12, ctime_str+11, 8); /* '21:49:08' */
(void) sprintf(str+20, ".%09lld", (longlong_t)tv[1]); /* '.123456789' */
if (opts->scripted)
(void) printf(gettext("%s\t"), str);
else
(void) printf(gettext("%s "), str);
verify(nvlist_lookup_string(nvl, FM_CLASS, &ptr) == 0);
(void) printf(gettext("%s\n"), ptr);
}
static void
zpool_do_events_nvprint(nvlist_t *nvl, int depth)
{
nvpair_t *nvp;
for (nvp = nvlist_next_nvpair(nvl, NULL);
nvp != NULL; nvp = nvlist_next_nvpair(nvl, nvp)) {
data_type_t type = nvpair_type(nvp);
const char *name = nvpair_name(nvp);
boolean_t b;
uint8_t i8;
uint16_t i16;
uint32_t i32;
uint64_t i64;
const char *str;
nvlist_t *cnv;
printf(gettext("%*s%s = "), depth, "", name);
switch (type) {
case DATA_TYPE_BOOLEAN:
printf(gettext("%s"), "1");
break;
case DATA_TYPE_BOOLEAN_VALUE:
(void) nvpair_value_boolean_value(nvp, &b);
printf(gettext("%s"), b ? "1" : "0");
break;
case DATA_TYPE_BYTE:
(void) nvpair_value_byte(nvp, &i8);
printf(gettext("0x%x"), i8);
break;
case DATA_TYPE_INT8:
(void) nvpair_value_int8(nvp, (void *)&i8);
printf(gettext("0x%x"), i8);
break;
case DATA_TYPE_UINT8:
(void) nvpair_value_uint8(nvp, &i8);
printf(gettext("0x%x"), i8);
break;
case DATA_TYPE_INT16:
(void) nvpair_value_int16(nvp, (void *)&i16);
printf(gettext("0x%x"), i16);
break;
case DATA_TYPE_UINT16:
(void) nvpair_value_uint16(nvp, &i16);
printf(gettext("0x%x"), i16);
break;
case DATA_TYPE_INT32:
(void) nvpair_value_int32(nvp, (void *)&i32);
printf(gettext("0x%x"), i32);
break;
case DATA_TYPE_UINT32:
(void) nvpair_value_uint32(nvp, &i32);
printf(gettext("0x%x"), i32);
break;
case DATA_TYPE_INT64:
(void) nvpair_value_int64(nvp, (void *)&i64);
printf(gettext("0x%llx"), (u_longlong_t)i64);
break;
case DATA_TYPE_UINT64:
(void) nvpair_value_uint64(nvp, &i64);
/*
* translate vdev state values to readable
* strings to aide zpool events consumers
*/
if (strcmp(name,
FM_EREPORT_PAYLOAD_ZFS_VDEV_STATE) == 0 ||
strcmp(name,
FM_EREPORT_PAYLOAD_ZFS_VDEV_LASTSTATE) == 0) {
printf(gettext("\"%s\" (0x%llx)"),
zpool_state_to_name(i64, VDEV_AUX_NONE),
(u_longlong_t)i64);
} else {
printf(gettext("0x%llx"), (u_longlong_t)i64);
}
break;
case DATA_TYPE_HRTIME:
(void) nvpair_value_hrtime(nvp, (void *)&i64);
printf(gettext("0x%llx"), (u_longlong_t)i64);
break;
case DATA_TYPE_STRING:
(void) nvpair_value_string(nvp, &str);
printf(gettext("\"%s\""), str ? str : "<NULL>");
break;
case DATA_TYPE_NVLIST:
printf(gettext("(embedded nvlist)\n"));
(void) nvpair_value_nvlist(nvp, &cnv);
zpool_do_events_nvprint(cnv, depth + 8);
printf(gettext("%*s(end %s)"), depth, "", name);
break;
case DATA_TYPE_NVLIST_ARRAY: {
nvlist_t **val;
uint_t i, nelem;
(void) nvpair_value_nvlist_array(nvp, &val, &nelem);
printf(gettext("(%d embedded nvlists)\n"), nelem);
for (i = 0; i < nelem; i++) {
printf(gettext("%*s%s[%d] = %s\n"),
depth, "", name, i, "(embedded nvlist)");
zpool_do_events_nvprint(val[i], depth + 8);
printf(gettext("%*s(end %s[%i])\n"),
depth, "", name, i);
}
printf(gettext("%*s(end %s)\n"), depth, "", name);
}
break;
case DATA_TYPE_INT8_ARRAY: {
int8_t *val;
uint_t i, nelem;
(void) nvpair_value_int8_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_UINT8_ARRAY: {
uint8_t *val;
uint_t i, nelem;
(void) nvpair_value_uint8_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_INT16_ARRAY: {
int16_t *val;
uint_t i, nelem;
(void) nvpair_value_int16_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_UINT16_ARRAY: {
uint16_t *val;
uint_t i, nelem;
(void) nvpair_value_uint16_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_INT32_ARRAY: {
int32_t *val;
uint_t i, nelem;
(void) nvpair_value_int32_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_UINT32_ARRAY: {
uint32_t *val;
uint_t i, nelem;
(void) nvpair_value_uint32_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_INT64_ARRAY: {
int64_t *val;
uint_t i, nelem;
(void) nvpair_value_int64_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%llx "),
(u_longlong_t)val[i]);
break;
}
case DATA_TYPE_UINT64_ARRAY: {
uint64_t *val;
uint_t i, nelem;
(void) nvpair_value_uint64_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%llx "),
(u_longlong_t)val[i]);
break;
}
case DATA_TYPE_STRING_ARRAY: {
const char **str;
uint_t i, nelem;
(void) nvpair_value_string_array(nvp, &str, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("\"%s\" "),
str[i] ? str[i] : "<NULL>");
break;
}
case DATA_TYPE_BOOLEAN_ARRAY:
case DATA_TYPE_BYTE_ARRAY:
case DATA_TYPE_DOUBLE:
case DATA_TYPE_DONTCARE:
case DATA_TYPE_UNKNOWN:
printf(gettext("<unknown>"));
break;
}
printf(gettext("\n"));
}
}
static int
zpool_do_events_next(ev_opts_t *opts)
{
nvlist_t *nvl;
int zevent_fd, ret, dropped;
const char *pool;
zevent_fd = open(ZFS_DEV, O_RDWR);
VERIFY(zevent_fd >= 0);
if (!opts->scripted)
(void) printf(gettext("%-30s %s\n"), "TIME", "CLASS");
while (1) {
ret = zpool_events_next(g_zfs, &nvl, &dropped,
(opts->follow ? ZEVENT_NONE : ZEVENT_NONBLOCK), zevent_fd);
if (ret || nvl == NULL)
break;
if (dropped > 0)
(void) printf(gettext("dropped %d events\n"), dropped);
if (strlen(opts->poolname) > 0 &&
nvlist_lookup_string(nvl, FM_FMRI_ZFS_POOL, &pool) == 0 &&
strcmp(opts->poolname, pool) != 0)
continue;
zpool_do_events_short(nvl, opts);
if (opts->verbose) {
zpool_do_events_nvprint(nvl, 8);
printf(gettext("\n"));
}
(void) fflush(stdout);
nvlist_free(nvl);
}
VERIFY(0 == close(zevent_fd));
return (ret);
}
static int
zpool_do_events_clear(void)
{
int count, ret;
ret = zpool_events_clear(g_zfs, &count);
if (!ret)
(void) printf(gettext("cleared %d events\n"), count);
return (ret);
}
/*
* zpool events [-vHf [pool] | -c]
*
* Displays events logs by ZFS.
*/
int
zpool_do_events(int argc, char **argv)
{
ev_opts_t opts = { 0 };
int ret;
int c;
/* check options */
while ((c = getopt(argc, argv, "vHfc")) != -1) {
switch (c) {
case 'v':
opts.verbose = 1;
break;
case 'H':
opts.scripted = 1;
break;
case 'f':
opts.follow = 1;
break;
case 'c':
opts.clear = 1;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
} else if (argc == 1) {
(void) strlcpy(opts.poolname, argv[0], sizeof (opts.poolname));
if (!zfs_name_valid(opts.poolname, ZFS_TYPE_POOL)) {
(void) fprintf(stderr,
gettext("invalid pool name '%s'\n"), opts.poolname);
usage(B_FALSE);
}
}
if ((argc == 1 || opts.verbose || opts.scripted || opts.follow) &&
opts.clear) {
(void) fprintf(stderr,
gettext("invalid options combined with -c\n"));
usage(B_FALSE);
}
if (opts.clear)
ret = zpool_do_events_clear();
else
ret = zpool_do_events_next(&opts);
return (ret);
}
static int
get_callback_vdev(zpool_handle_t *zhp, char *vdevname, void *data)
{
zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data;
char value[ZFS_MAXPROPLEN];
zprop_source_t srctype;
for (zprop_list_t *pl = cbp->cb_proplist; pl != NULL;
pl = pl->pl_next) {
char *prop_name;
/*
* If the first property is pool name, it is a special
* placeholder that we can skip. This will also skip
* over the name property when 'all' is specified.
*/
if (pl->pl_prop == ZPOOL_PROP_NAME &&
pl == cbp->cb_proplist)
continue;
if (pl->pl_prop == ZPROP_INVAL) {
prop_name = pl->pl_user_prop;
} else {
prop_name = (char *)vdev_prop_to_name(pl->pl_prop);
}
if (zpool_get_vdev_prop(zhp, vdevname, pl->pl_prop,
prop_name, value, sizeof (value), &srctype,
cbp->cb_literal) == 0) {
zprop_print_one_property(vdevname, cbp, prop_name,
value, srctype, NULL, NULL);
}
}
return (0);
}
static int
get_callback_vdev_cb(void *zhp_data, nvlist_t *nv, void *data)
{
zpool_handle_t *zhp = zhp_data;
zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data;
char *vdevname;
const char *type;
int ret;
/*
* zpool_vdev_name() transforms the root vdev name (i.e., root-0) to the
* pool name for display purposes, which is not desired. Fallback to
* zpool_vdev_name() when not dealing with the root vdev.
*/
type = fnvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE);
if (zhp != NULL && strcmp(type, "root") == 0)
vdevname = strdup("root-0");
else
vdevname = zpool_vdev_name(g_zfs, zhp, nv,
cbp->cb_vdevs.cb_name_flags);
(void) vdev_expand_proplist(zhp, vdevname, &cbp->cb_proplist);
ret = get_callback_vdev(zhp, vdevname, data);
free(vdevname);
return (ret);
}
static int
get_callback(zpool_handle_t *zhp, void *data)
{
zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data;
char value[ZFS_MAXPROPLEN];
zprop_source_t srctype;
zprop_list_t *pl;
int vid;
if (cbp->cb_type == ZFS_TYPE_VDEV) {
if (strcmp(cbp->cb_vdevs.cb_names[0], "all-vdevs") == 0) {
for_each_vdev(zhp, get_callback_vdev_cb, data);
} else {
/* Adjust column widths for vdev properties */
for (vid = 0; vid < cbp->cb_vdevs.cb_names_count;
vid++) {
vdev_expand_proplist(zhp,
cbp->cb_vdevs.cb_names[vid],
&cbp->cb_proplist);
}
/* Display the properties */
for (vid = 0; vid < cbp->cb_vdevs.cb_names_count;
vid++) {
get_callback_vdev(zhp,
cbp->cb_vdevs.cb_names[vid], data);
}
}
} else {
assert(cbp->cb_type == ZFS_TYPE_POOL);
for (pl = cbp->cb_proplist; pl != NULL; pl = pl->pl_next) {
/*
* Skip the special fake placeholder. This will also
* skip over the name property when 'all' is specified.
*/
if (pl->pl_prop == ZPOOL_PROP_NAME &&
pl == cbp->cb_proplist)
continue;
if (pl->pl_prop == ZPROP_INVAL &&
zfs_prop_user(pl->pl_user_prop)) {
srctype = ZPROP_SRC_LOCAL;
if (zpool_get_userprop(zhp, pl->pl_user_prop,
value, sizeof (value), &srctype) != 0)
continue;
zprop_print_one_property(zpool_get_name(zhp),
cbp, pl->pl_user_prop, value, srctype,
NULL, NULL);
} else if (pl->pl_prop == ZPROP_INVAL &&
(zpool_prop_feature(pl->pl_user_prop) ||
zpool_prop_unsupported(pl->pl_user_prop))) {
srctype = ZPROP_SRC_LOCAL;
if (zpool_prop_get_feature(zhp,
pl->pl_user_prop, value,
sizeof (value)) == 0) {
zprop_print_one_property(
zpool_get_name(zhp), cbp,
pl->pl_user_prop, value, srctype,
NULL, NULL);
}
} else {
if (zpool_get_prop(zhp, pl->pl_prop, value,
sizeof (value), &srctype,
cbp->cb_literal) != 0)
continue;
zprop_print_one_property(zpool_get_name(zhp),
cbp, zpool_prop_to_name(pl->pl_prop),
value, srctype, NULL, NULL);
}
}
}
return (0);
}
/*
* zpool get [-Hp] [-o "all" | field[,...]] <"all" | property[,...]> <pool> ...
*
* -H Scripted mode. Don't display headers, and separate properties
* by a single tab.
* -o List of columns to display. Defaults to
* "name,property,value,source".
* -p Display values in parsable (exact) format.
*
* Get properties of pools in the system. Output space statistics
* for each one as well as other attributes.
*/
int
zpool_do_get(int argc, char **argv)
{
zprop_get_cbdata_t cb = { 0 };
zprop_list_t fake_name = { 0 };
int ret;
int c, i;
char *propstr = NULL;
char *vdev = NULL;
cb.cb_first = B_TRUE;
/*
* Set up default columns and sources.
*/
cb.cb_sources = ZPROP_SRC_ALL;
cb.cb_columns[0] = GET_COL_NAME;
cb.cb_columns[1] = GET_COL_PROPERTY;
cb.cb_columns[2] = GET_COL_VALUE;
cb.cb_columns[3] = GET_COL_SOURCE;
cb.cb_type = ZFS_TYPE_POOL;
cb.cb_vdevs.cb_name_flags |= VDEV_NAME_TYPE_ID;
current_prop_type = cb.cb_type;
/* check options */
while ((c = getopt(argc, argv, ":Hpo:")) != -1) {
switch (c) {
case 'p':
cb.cb_literal = B_TRUE;
break;
case 'H':
cb.cb_scripted = B_TRUE;
break;
case 'o':
memset(&cb.cb_columns, 0, sizeof (cb.cb_columns));
i = 0;
for (char *tok; (tok = strsep(&optarg, ",")); ) {
static const char *const col_opts[] =
{ "name", "property", "value", "source",
"all" };
static const zfs_get_column_t col_cols[] =
{ GET_COL_NAME, GET_COL_PROPERTY, GET_COL_VALUE,
GET_COL_SOURCE };
if (i == ZFS_GET_NCOLS - 1) {
(void) fprintf(stderr, gettext("too "
"many fields given to -o "
"option\n"));
usage(B_FALSE);
}
for (c = 0; c < ARRAY_SIZE(col_opts); ++c)
if (strcmp(tok, col_opts[c]) == 0)
goto found;
(void) fprintf(stderr,
gettext("invalid column name '%s'\n"), tok);
usage(B_FALSE);
found:
if (c >= 4) {
if (i > 0) {
(void) fprintf(stderr,
gettext("\"all\" conflicts "
"with specific fields "
"given to -o option\n"));
usage(B_FALSE);
}
memcpy(cb.cb_columns, col_cols,
sizeof (col_cols));
i = ZFS_GET_NCOLS - 1;
} else
cb.cb_columns[i++] = col_cols[c];
}
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing property "
"argument\n"));
usage(B_FALSE);
}
/* Properties list is needed later by zprop_get_list() */
propstr = argv[0];
argc--;
argv++;
if (argc == 0) {
/* No args, so just print the defaults. */
} else if (are_all_pools(argc, argv)) {
/* All the args are pool names */
} else if (are_all_pools(1, argv)) {
/* The first arg is a pool name */
if ((argc == 2 && strcmp(argv[1], "all-vdevs") == 0) ||
(argc == 2 && strcmp(argv[1], "root") == 0) ||
are_vdevs_in_pool(argc - 1, argv + 1, argv[0],
&cb.cb_vdevs)) {
if (strcmp(argv[1], "root") == 0)
vdev = strdup("root-0");
else
vdev = strdup(argv[1]);
/* ... and the rest are vdev names */
cb.cb_vdevs.cb_names = &vdev;
cb.cb_vdevs.cb_names_count = argc - 1;
cb.cb_type = ZFS_TYPE_VDEV;
argc = 1; /* One pool to process */
} else {
fprintf(stderr, gettext("Expected a list of vdevs in"
" \"%s\", but got:\n"), argv[0]);
error_list_unresolved_vdevs(argc - 1, argv + 1,
argv[0], &cb.cb_vdevs);
fprintf(stderr, "\n");
usage(B_FALSE);
return (1);
}
} else {
/*
* The first arg isn't a pool name,
*/
fprintf(stderr, gettext("missing pool name.\n"));
fprintf(stderr, "\n");
usage(B_FALSE);
return (1);
}
if (zprop_get_list(g_zfs, propstr, &cb.cb_proplist,
cb.cb_type) != 0) {
/* Use correct list of valid properties (pool or vdev) */
current_prop_type = cb.cb_type;
usage(B_FALSE);
}
if (cb.cb_proplist != NULL) {
fake_name.pl_prop = ZPOOL_PROP_NAME;
fake_name.pl_width = strlen(gettext("NAME"));
fake_name.pl_next = cb.cb_proplist;
cb.cb_proplist = &fake_name;
}
ret = for_each_pool(argc, argv, B_TRUE, &cb.cb_proplist, cb.cb_type,
cb.cb_literal, get_callback, &cb);
if (cb.cb_proplist == &fake_name)
zprop_free_list(fake_name.pl_next);
else
zprop_free_list(cb.cb_proplist);
if (vdev != NULL)
free(vdev);
return (ret);
}
typedef struct set_cbdata {
char *cb_propname;
char *cb_value;
zfs_type_t cb_type;
vdev_cbdata_t cb_vdevs;
boolean_t cb_any_successful;
} set_cbdata_t;
static int
set_pool_callback(zpool_handle_t *zhp, set_cbdata_t *cb)
{
int error;
/* Check if we have out-of-bounds features */
if (strcmp(cb->cb_propname, ZPOOL_CONFIG_COMPATIBILITY) == 0) {
boolean_t features[SPA_FEATURES];
if (zpool_do_load_compat(cb->cb_value, features) !=
ZPOOL_COMPATIBILITY_OK)
return (-1);
nvlist_t *enabled = zpool_get_features(zhp);
spa_feature_t i;
for (i = 0; i < SPA_FEATURES; i++) {
const char *fguid = spa_feature_table[i].fi_guid;
if (nvlist_exists(enabled, fguid) && !features[i])
break;
}
if (i < SPA_FEATURES)
(void) fprintf(stderr, gettext("Warning: one or "
"more features already enabled on pool '%s'\n"
"are not present in this compatibility set.\n"),
zpool_get_name(zhp));
}
/* if we're setting a feature, check it's in compatibility set */
if (zpool_prop_feature(cb->cb_propname) &&
strcmp(cb->cb_value, ZFS_FEATURE_ENABLED) == 0) {
char *fname = strchr(cb->cb_propname, '@') + 1;
spa_feature_t f;
if (zfeature_lookup_name(fname, &f) == 0) {
char compat[ZFS_MAXPROPLEN];
if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY,
compat, ZFS_MAXPROPLEN, NULL, B_FALSE) != 0)
compat[0] = '\0';
boolean_t features[SPA_FEATURES];
if (zpool_do_load_compat(compat, features) !=
ZPOOL_COMPATIBILITY_OK) {
(void) fprintf(stderr, gettext("Error: "
"cannot enable feature '%s' on pool '%s'\n"
"because the pool's 'compatibility' "
"property cannot be parsed.\n"),
fname, zpool_get_name(zhp));
return (-1);
}
if (!features[f]) {
(void) fprintf(stderr, gettext("Error: "
"cannot enable feature '%s' on pool '%s'\n"
"as it is not specified in this pool's "
"current compatibility set.\n"
"Consider setting 'compatibility' to a "
"less restrictive set, or to 'off'.\n"),
fname, zpool_get_name(zhp));
return (-1);
}
}
}
error = zpool_set_prop(zhp, cb->cb_propname, cb->cb_value);
return (error);
}
static int
set_callback(zpool_handle_t *zhp, void *data)
{
int error;
set_cbdata_t *cb = (set_cbdata_t *)data;
if (cb->cb_type == ZFS_TYPE_VDEV) {
error = zpool_set_vdev_prop(zhp, *cb->cb_vdevs.cb_names,
cb->cb_propname, cb->cb_value);
} else {
assert(cb->cb_type == ZFS_TYPE_POOL);
error = set_pool_callback(zhp, cb);
}
cb->cb_any_successful = !error;
return (error);
}
int
zpool_do_set(int argc, char **argv)
{
set_cbdata_t cb = { 0 };
int error;
char *vdev = NULL;
current_prop_type = ZFS_TYPE_POOL;
if (argc > 1 && argv[1][0] == '-') {
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
argv[1][1]);
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing property=value "
"argument\n"));
usage(B_FALSE);
}
if (argc < 3) {
(void) fprintf(stderr, gettext("missing pool name\n"));
usage(B_FALSE);
}
if (argc > 4) {
(void) fprintf(stderr, gettext("too many pool names\n"));
usage(B_FALSE);
}
cb.cb_propname = argv[1];
cb.cb_type = ZFS_TYPE_POOL;
cb.cb_vdevs.cb_name_flags |= VDEV_NAME_TYPE_ID;
cb.cb_value = strchr(cb.cb_propname, '=');
if (cb.cb_value == NULL) {
(void) fprintf(stderr, gettext("missing value in "
"property=value argument\n"));
usage(B_FALSE);
}
*(cb.cb_value) = '\0';
cb.cb_value++;
argc -= 2;
argv += 2;
/* argv[0] is pool name */
if (!is_pool(argv[0])) {
(void) fprintf(stderr,
gettext("cannot open '%s': is not a pool\n"), argv[0]);
return (EINVAL);
}
/* argv[1], when supplied, is vdev name */
if (argc == 2) {
if (strcmp(argv[1], "root") == 0)
vdev = strdup("root-0");
else
vdev = strdup(argv[1]);
if (!are_vdevs_in_pool(1, &vdev, argv[0], &cb.cb_vdevs)) {
(void) fprintf(stderr, gettext(
"cannot find '%s' in '%s': device not in pool\n"),
vdev, argv[0]);
free(vdev);
return (EINVAL);
}
cb.cb_vdevs.cb_names = &vdev;
cb.cb_vdevs.cb_names_count = 1;
cb.cb_type = ZFS_TYPE_VDEV;
}
error = for_each_pool(1, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, set_callback, &cb);
if (vdev != NULL)
free(vdev);
return (error);
}
/* Add up the total number of bytes left to initialize/trim across all vdevs */
static uint64_t
vdev_activity_remaining(nvlist_t *nv, zpool_wait_activity_t activity)
{
uint64_t bytes_remaining;
nvlist_t **child;
uint_t c, children;
vdev_stat_t *vs;
assert(activity == ZPOOL_WAIT_INITIALIZE ||
activity == ZPOOL_WAIT_TRIM);
verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &c) == 0);
if (activity == ZPOOL_WAIT_INITIALIZE &&
vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE)
bytes_remaining = vs->vs_initialize_bytes_est -
vs->vs_initialize_bytes_done;
else if (activity == ZPOOL_WAIT_TRIM &&
vs->vs_trim_state == VDEV_TRIM_ACTIVE)
bytes_remaining = vs->vs_trim_bytes_est -
vs->vs_trim_bytes_done;
else
bytes_remaining = 0;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
for (c = 0; c < children; c++)
bytes_remaining += vdev_activity_remaining(child[c], activity);
return (bytes_remaining);
}
/* Add up the total number of bytes left to rebuild across top-level vdevs */
static uint64_t
vdev_activity_top_remaining(nvlist_t *nv)
{
uint64_t bytes_remaining = 0;
nvlist_t **child;
uint_t children;
int error;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
for (uint_t c = 0; c < children; c++) {
vdev_rebuild_stat_t *vrs;
uint_t i;
error = nvlist_lookup_uint64_array(child[c],
ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i);
if (error == 0) {
if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) {
bytes_remaining += (vrs->vrs_bytes_est -
vrs->vrs_bytes_rebuilt);
}
}
}
return (bytes_remaining);
}
/* Whether any vdevs are 'spare' or 'replacing' vdevs */
static boolean_t
vdev_any_spare_replacing(nvlist_t *nv)
{
nvlist_t **child;
uint_t c, children;
const char *vdev_type;
(void) nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &vdev_type);
if (strcmp(vdev_type, VDEV_TYPE_REPLACING) == 0 ||
strcmp(vdev_type, VDEV_TYPE_SPARE) == 0 ||
strcmp(vdev_type, VDEV_TYPE_DRAID_SPARE) == 0) {
return (B_TRUE);
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
for (c = 0; c < children; c++) {
if (vdev_any_spare_replacing(child[c]))
return (B_TRUE);
}
return (B_FALSE);
}
typedef struct wait_data {
char *wd_poolname;
boolean_t wd_scripted;
boolean_t wd_exact;
boolean_t wd_headers_once;
boolean_t wd_should_exit;
/* Which activities to wait for */
boolean_t wd_enabled[ZPOOL_WAIT_NUM_ACTIVITIES];
float wd_interval;
pthread_cond_t wd_cv;
pthread_mutex_t wd_mutex;
} wait_data_t;
/*
* Print to stdout a single line, containing one column for each activity that
* we are waiting for specifying how many bytes of work are left for that
* activity.
*/
static void
print_wait_status_row(wait_data_t *wd, zpool_handle_t *zhp, int row)
{
nvlist_t *config, *nvroot;
uint_t c;
int i;
pool_checkpoint_stat_t *pcs = NULL;
pool_scan_stat_t *pss = NULL;
pool_removal_stat_t *prs = NULL;
+ pool_raidz_expand_stat_t *pres = NULL;
const char *const headers[] = {"DISCARD", "FREE", "INITIALIZE",
- "REPLACE", "REMOVE", "RESILVER", "SCRUB", "TRIM"};
+ "REPLACE", "REMOVE", "RESILVER", "SCRUB", "TRIM", "RAIDZ_EXPAND"};
int col_widths[ZPOOL_WAIT_NUM_ACTIVITIES];
/* Calculate the width of each column */
for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) {
/*
* Make sure we have enough space in the col for pretty-printed
* numbers and for the column header, and then leave a couple
* spaces between cols for readability.
*/
col_widths[i] = MAX(strlen(headers[i]), 6) + 2;
}
/* Print header if appropriate */
int term_height = terminal_height();
boolean_t reprint_header = (!wd->wd_headers_once && term_height > 0 &&
row % (term_height-1) == 0);
if (!wd->wd_scripted && (row == 0 || reprint_header)) {
for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) {
if (wd->wd_enabled[i])
(void) printf("%*s", col_widths[i], headers[i]);
}
(void) fputc('\n', stdout);
}
/* Bytes of work remaining in each activity */
int64_t bytes_rem[ZPOOL_WAIT_NUM_ACTIVITIES] = {0};
bytes_rem[ZPOOL_WAIT_FREE] =
zpool_get_prop_int(zhp, ZPOOL_PROP_FREEING, NULL);
config = zpool_get_config(zhp, NULL);
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c);
if (pcs != NULL && pcs->pcs_state == CS_CHECKPOINT_DISCARDING)
bytes_rem[ZPOOL_WAIT_CKPT_DISCARD] = pcs->pcs_space;
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t **)&prs, &c);
if (prs != NULL && prs->prs_state == DSS_SCANNING)
bytes_rem[ZPOOL_WAIT_REMOVE] = prs->prs_to_copy -
prs->prs_copied;
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&pss, &c);
if (pss != NULL && pss->pss_state == DSS_SCANNING &&
pss->pss_pass_scrub_pause == 0) {
int64_t rem = pss->pss_to_examine - pss->pss_issued;
if (pss->pss_func == POOL_SCAN_SCRUB)
bytes_rem[ZPOOL_WAIT_SCRUB] = rem;
else
bytes_rem[ZPOOL_WAIT_RESILVER] = rem;
} else if (check_rebuilding(nvroot, NULL)) {
bytes_rem[ZPOOL_WAIT_RESILVER] =
vdev_activity_top_remaining(nvroot);
}
+ (void) nvlist_lookup_uint64_array(nvroot,
+ ZPOOL_CONFIG_RAIDZ_EXPAND_STATS, (uint64_t **)&pres, &c);
+ if (pres != NULL && pres->pres_state == DSS_SCANNING) {
+ int64_t rem = pres->pres_to_reflow - pres->pres_reflowed;
+ bytes_rem[ZPOOL_WAIT_RAIDZ_EXPAND] = rem;
+ }
+
bytes_rem[ZPOOL_WAIT_INITIALIZE] =
vdev_activity_remaining(nvroot, ZPOOL_WAIT_INITIALIZE);
bytes_rem[ZPOOL_WAIT_TRIM] =
vdev_activity_remaining(nvroot, ZPOOL_WAIT_TRIM);
/*
* A replace finishes after resilvering finishes, so the amount of work
* left for a replace is the same as for resilvering.
*
* It isn't quite correct to say that if we have any 'spare' or
* 'replacing' vdevs and a resilver is happening, then a replace is in
* progress, like we do here. When a hot spare is used, the faulted vdev
* is not removed after the hot spare is resilvered, so parent 'spare'
* vdev is not removed either. So we could have a 'spare' vdev, but be
* resilvering for a different reason. However, we use it as a heuristic
* because we don't have access to the DTLs, which could tell us whether
* or not we have really finished resilvering a hot spare.
*/
if (vdev_any_spare_replacing(nvroot))
bytes_rem[ZPOOL_WAIT_REPLACE] = bytes_rem[ZPOOL_WAIT_RESILVER];
if (timestamp_fmt != NODATE)
print_timestamp(timestamp_fmt);
for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) {
char buf[64];
if (!wd->wd_enabled[i])
continue;
- if (wd->wd_exact)
+ if (wd->wd_exact) {
(void) snprintf(buf, sizeof (buf), "%" PRIi64,
bytes_rem[i]);
- else
+ } else {
zfs_nicenum(bytes_rem[i], buf, sizeof (buf));
+ }
if (wd->wd_scripted)
(void) printf(i == 0 ? "%s" : "\t%s", buf);
else
(void) printf(" %*s", col_widths[i] - 1, buf);
}
(void) printf("\n");
(void) fflush(stdout);
}
static void *
wait_status_thread(void *arg)
{
wait_data_t *wd = (wait_data_t *)arg;
zpool_handle_t *zhp;
if ((zhp = zpool_open(g_zfs, wd->wd_poolname)) == NULL)
return (void *)(1);
for (int row = 0; ; row++) {
boolean_t missing;
struct timespec timeout;
int ret = 0;
(void) clock_gettime(CLOCK_REALTIME, &timeout);
if (zpool_refresh_stats(zhp, &missing) != 0 || missing ||
zpool_props_refresh(zhp) != 0) {
zpool_close(zhp);
return (void *)(uintptr_t)(missing ? 0 : 1);
}
print_wait_status_row(wd, zhp, row);
timeout.tv_sec += floor(wd->wd_interval);
long nanos = timeout.tv_nsec +
(wd->wd_interval - floor(wd->wd_interval)) * NANOSEC;
if (nanos >= NANOSEC) {
timeout.tv_sec++;
timeout.tv_nsec = nanos - NANOSEC;
} else {
timeout.tv_nsec = nanos;
}
pthread_mutex_lock(&wd->wd_mutex);
if (!wd->wd_should_exit)
ret = pthread_cond_timedwait(&wd->wd_cv, &wd->wd_mutex,
&timeout);
pthread_mutex_unlock(&wd->wd_mutex);
if (ret == 0) {
break; /* signaled by main thread */
} else if (ret != ETIMEDOUT) {
(void) fprintf(stderr, gettext("pthread_cond_timedwait "
"failed: %s\n"), strerror(ret));
zpool_close(zhp);
return (void *)(uintptr_t)(1);
}
}
zpool_close(zhp);
return (void *)(0);
}
int
zpool_do_wait(int argc, char **argv)
{
boolean_t verbose = B_FALSE;
int c, i;
unsigned long count;
pthread_t status_thr;
int error = 0;
zpool_handle_t *zhp;
wait_data_t wd;
wd.wd_scripted = B_FALSE;
wd.wd_exact = B_FALSE;
wd.wd_headers_once = B_FALSE;
wd.wd_should_exit = B_FALSE;
pthread_mutex_init(&wd.wd_mutex, NULL);
pthread_cond_init(&wd.wd_cv, NULL);
/* By default, wait for all types of activity. */
for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++)
wd.wd_enabled[i] = B_TRUE;
while ((c = getopt(argc, argv, "HpT:t:")) != -1) {
switch (c) {
case 'H':
wd.wd_scripted = B_TRUE;
break;
case 'n':
wd.wd_headers_once = B_TRUE;
break;
case 'p':
wd.wd_exact = B_TRUE;
break;
case 'T':
get_timestamp_arg(*optarg);
break;
case 't':
/* Reset activities array */
memset(&wd.wd_enabled, 0, sizeof (wd.wd_enabled));
for (char *tok; (tok = strsep(&optarg, ",")); ) {
static const char *const col_opts[] = {
"discard", "free", "initialize", "replace",
- "remove", "resilver", "scrub", "trim" };
+ "remove", "resilver", "scrub", "trim",
+ "raidz_expand" };
for (i = 0; i < ARRAY_SIZE(col_opts); ++i)
if (strcmp(tok, col_opts[i]) == 0) {
wd.wd_enabled[i] = B_TRUE;
goto found;
}
(void) fprintf(stderr,
gettext("invalid activity '%s'\n"), tok);
usage(B_FALSE);
found:;
}
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
get_interval_count(&argc, argv, &wd.wd_interval, &count);
if (count != 0) {
/* This subcmd only accepts an interval, not a count */
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if (wd.wd_interval != 0)
verbose = B_TRUE;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing 'pool' argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
wd.wd_poolname = argv[0];
if ((zhp = zpool_open(g_zfs, wd.wd_poolname)) == NULL)
return (1);
if (verbose) {
/*
* We use a separate thread for printing status updates because
* the main thread will call lzc_wait(), which blocks as long
* as an activity is in progress, which can be a long time.
*/
if (pthread_create(&status_thr, NULL, wait_status_thread, &wd)
!= 0) {
(void) fprintf(stderr, gettext("failed to create status"
"thread: %s\n"), strerror(errno));
zpool_close(zhp);
return (1);
}
}
/*
* Loop over all activities that we are supposed to wait for until none
* of them are in progress. Note that this means we can end up waiting
* for more activities to complete than just those that were in progress
* when we began waiting; if an activity we are interested in begins
* while we are waiting for another activity, we will wait for both to
* complete before exiting.
*/
for (;;) {
boolean_t missing = B_FALSE;
boolean_t any_waited = B_FALSE;
for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) {
boolean_t waited;
if (!wd.wd_enabled[i])
continue;
error = zpool_wait_status(zhp, i, &missing, &waited);
if (error != 0 || missing)
break;
any_waited = (any_waited || waited);
}
if (error != 0 || missing || !any_waited)
break;
}
zpool_close(zhp);
if (verbose) {
uintptr_t status;
pthread_mutex_lock(&wd.wd_mutex);
wd.wd_should_exit = B_TRUE;
pthread_cond_signal(&wd.wd_cv);
pthread_mutex_unlock(&wd.wd_mutex);
(void) pthread_join(status_thr, (void *)&status);
if (status != 0)
error = status;
}
pthread_mutex_destroy(&wd.wd_mutex);
pthread_cond_destroy(&wd.wd_cv);
return (error);
}
static int
find_command_idx(const char *command, int *idx)
{
for (int i = 0; i < NCOMMAND; ++i) {
if (command_table[i].name == NULL)
continue;
if (strcmp(command, command_table[i].name) == 0) {
*idx = i;
return (0);
}
}
return (1);
}
/*
* Display version message
*/
static int
zpool_do_version(int argc, char **argv)
{
(void) argc, (void) argv;
return (zfs_version_print() != 0);
}
/* Display documentation */
static int
zpool_do_help(int argc, char **argv)
{
char page[MAXNAMELEN];
if (argc < 3 || strcmp(argv[2], "zpool") == 0)
strcpy(page, "zpool");
else if (strcmp(argv[2], "concepts") == 0 ||
strcmp(argv[2], "props") == 0)
snprintf(page, sizeof (page), "zpool%s", argv[2]);
else
snprintf(page, sizeof (page), "zpool-%s", argv[2]);
execlp("man", "man", page, NULL);
fprintf(stderr, "couldn't run man program: %s", strerror(errno));
return (-1);
}
/*
* Do zpool_load_compat() and print error message on failure
*/
static zpool_compat_status_t
zpool_do_load_compat(const char *compat, boolean_t *list)
{
char report[1024];
zpool_compat_status_t ret;
ret = zpool_load_compat(compat, list, report, 1024);
switch (ret) {
case ZPOOL_COMPATIBILITY_OK:
break;
case ZPOOL_COMPATIBILITY_NOFILES:
case ZPOOL_COMPATIBILITY_BADFILE:
case ZPOOL_COMPATIBILITY_BADTOKEN:
(void) fprintf(stderr, "Error: %s\n", report);
break;
case ZPOOL_COMPATIBILITY_WARNTOKEN:
(void) fprintf(stderr, "Warning: %s\n", report);
ret = ZPOOL_COMPATIBILITY_OK;
break;
}
return (ret);
}
int
main(int argc, char **argv)
{
int ret = 0;
int i = 0;
char *cmdname;
char **newargv;
(void) setlocale(LC_ALL, "");
(void) setlocale(LC_NUMERIC, "C");
(void) textdomain(TEXT_DOMAIN);
srand(time(NULL));
opterr = 0;
/*
* Make sure the user has specified some command.
*/
if (argc < 2) {
(void) fprintf(stderr, gettext("missing command\n"));
usage(B_FALSE);
}
cmdname = argv[1];
/*
* Special case '-?'
*/
if ((strcmp(cmdname, "-?") == 0) || strcmp(cmdname, "--help") == 0)
usage(B_TRUE);
/*
* Special case '-V|--version'
*/
if ((strcmp(cmdname, "-V") == 0) || (strcmp(cmdname, "--version") == 0))
return (zpool_do_version(argc, argv));
/*
* Special case 'help'
*/
if (strcmp(cmdname, "help") == 0)
return (zpool_do_help(argc, argv));
if ((g_zfs = libzfs_init()) == NULL) {
(void) fprintf(stderr, "%s\n", libzfs_error_init(errno));
return (1);
}
libzfs_print_on_error(g_zfs, B_TRUE);
zfs_save_arguments(argc, argv, history_str, sizeof (history_str));
/*
* Many commands modify input strings for string parsing reasons.
* We create a copy to protect the original argv.
*/
newargv = safe_malloc((argc + 1) * sizeof (newargv[0]));
for (i = 0; i < argc; i++)
newargv[i] = strdup(argv[i]);
newargv[argc] = NULL;
/*
* Run the appropriate command.
*/
if (find_command_idx(cmdname, &i) == 0) {
current_command = &command_table[i];
ret = command_table[i].func(argc - 1, newargv + 1);
} else if (strchr(cmdname, '=')) {
verify(find_command_idx("set", &i) == 0);
current_command = &command_table[i];
ret = command_table[i].func(argc, newargv);
} else if (strcmp(cmdname, "freeze") == 0 && argc == 3) {
/*
* 'freeze' is a vile debugging abomination, so we treat
* it as such.
*/
zfs_cmd_t zc = {"\0"};
(void) strlcpy(zc.zc_name, argv[2], sizeof (zc.zc_name));
ret = zfs_ioctl(g_zfs, ZFS_IOC_POOL_FREEZE, &zc);
if (ret != 0) {
(void) fprintf(stderr,
gettext("failed to freeze pool: %d\n"), errno);
ret = 1;
}
log_history = 0;
} else {
(void) fprintf(stderr, gettext("unrecognized "
"command '%s'\n"), cmdname);
usage(B_FALSE);
ret = 1;
}
for (i = 0; i < argc; i++)
free(newargv[i]);
free(newargv);
if (ret == 0 && log_history)
(void) zpool_log_history(g_zfs, history_str);
libzfs_fini(g_zfs);
/*
* The 'ZFS_ABORT' environment variable causes us to dump core on exit
* for the purposes of running ::findleaks.
*/
if (getenv("ZFS_ABORT") != NULL) {
(void) printf("dumping core by request\n");
abort();
}
return (ret);
}
diff --git a/sys/contrib/openzfs/cmd/ztest.c b/sys/contrib/openzfs/cmd/ztest.c
index 8cfbdfe1c2e2..1d414a9f6fd5 100644
--- a/sys/contrib/openzfs/cmd/ztest.c
+++ b/sys/contrib/openzfs/cmd/ztest.c
@@ -1,8326 +1,9040 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2018 by Delphix. All rights reserved.
* Copyright 2011 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2013 Steven Hartland. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright 2017 Joyent, Inc.
* Copyright (c) 2017, Intel Corporation.
*/
/*
* The objective of this program is to provide a DMU/ZAP/SPA stress test
* that runs entirely in userland, is easy to use, and easy to extend.
*
* The overall design of the ztest program is as follows:
*
* (1) For each major functional area (e.g. adding vdevs to a pool,
* creating and destroying datasets, reading and writing objects, etc)
* we have a simple routine to test that functionality. These
* individual routines do not have to do anything "stressful".
*
* (2) We turn these simple functionality tests into a stress test by
* running them all in parallel, with as many threads as desired,
* and spread across as many datasets, objects, and vdevs as desired.
*
* (3) While all this is happening, we inject faults into the pool to
* verify that self-healing data really works.
*
* (4) Every time we open a dataset, we change its checksum and compression
* functions. Thus even individual objects vary from block to block
* in which checksum they use and whether they're compressed.
*
* (5) To verify that we never lose on-disk consistency after a crash,
* we run the entire test in a child of the main process.
* At random times, the child self-immolates with a SIGKILL.
* This is the software equivalent of pulling the power cord.
* The parent then runs the test again, using the existing
* storage pool, as many times as desired. If backwards compatibility
* testing is enabled ztest will sometimes run the "older" version
* of ztest after a SIGKILL.
*
* (6) To verify that we don't have future leaks or temporal incursions,
* many of the functional tests record the transaction group number
* as part of their data. When reading old data, they verify that
* the transaction group number is less than the current, open txg.
* If you add a new test, please do this if applicable.
*
* (7) Threads are created with a reduced stack size, for sanity checking.
* Therefore, it's important not to allocate huge buffers on the stack.
*
* When run with no arguments, ztest runs for about five minutes and
* produces no output if successful. To get a little bit of information,
* specify -V. To get more information, specify -VV, and so on.
*
* To turn this into an overnight stress test, use -T to specify run time.
*
* You can ask more vdevs [-v], datasets [-d], or threads [-t]
* to increase the pool capacity, fanout, and overall stress level.
*
* Use the -k option to set the desired frequency of kills.
*
* When ztest invokes itself it passes all relevant information through a
* temporary file which is mmap-ed in the child process. This allows shared
* memory to survive the exec syscall. The ztest_shared_hdr_t struct is always
* stored at offset 0 of this file and contains information on the size and
* number of shared structures in the file. The information stored in this file
* must remain backwards compatible with older versions of ztest so that
* ztest can invoke them during backwards compatibility testing (-B).
*/
#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/dmu.h>
#include <sys/txg.h>
#include <sys/dbuf.h>
#include <sys/zap.h>
#include <sys/dmu_objset.h>
#include <sys/poll.h>
#include <sys/stat.h>
#include <sys/time.h>
#include <sys/wait.h>
#include <sys/mman.h>
#include <sys/resource.h>
#include <sys/zio.h>
#include <sys/zil.h>
#include <sys/zil_impl.h>
#include <sys/vdev_draid.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_file.h>
#include <sys/vdev_initialize.h>
#include <sys/vdev_raidz.h>
#include <sys/vdev_trim.h>
#include <sys/spa_impl.h>
#include <sys/metaslab_impl.h>
#include <sys/dsl_prop.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_destroy.h>
#include <sys/dsl_scan.h>
#include <sys/zio_checksum.h>
#include <sys/zfs_refcount.h>
#include <sys/zfeature.h>
#include <sys/dsl_userhold.h>
#include <sys/abd.h>
#include <sys/blake3.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <getopt.h>
#include <signal.h>
#include <umem.h>
#include <ctype.h>
#include <math.h>
#include <sys/fs/zfs.h>
#include <zfs_fletcher.h>
#include <libnvpair.h>
#include <libzutil.h>
#include <sys/crypto/icp.h>
#include <sys/zfs_impl.h>
#if (__GLIBC__ && !__UCLIBC__)
#include <execinfo.h> /* for backtrace() */
#endif
static int ztest_fd_data = -1;
static int ztest_fd_rand = -1;
typedef struct ztest_shared_hdr {
uint64_t zh_hdr_size;
uint64_t zh_opts_size;
uint64_t zh_size;
uint64_t zh_stats_size;
uint64_t zh_stats_count;
uint64_t zh_ds_size;
uint64_t zh_ds_count;
+ uint64_t zh_scratch_state_size;
} ztest_shared_hdr_t;
static ztest_shared_hdr_t *ztest_shared_hdr;
enum ztest_class_state {
ZTEST_VDEV_CLASS_OFF,
ZTEST_VDEV_CLASS_ON,
ZTEST_VDEV_CLASS_RND
};
+/* Dedicated RAIDZ Expansion test states */
+typedef enum {
+ RAIDZ_EXPAND_NONE, /* Default is none, must opt-in */
+ RAIDZ_EXPAND_REQUESTED, /* The '-X' option was used */
+ RAIDZ_EXPAND_STARTED, /* Testing has commenced */
+ RAIDZ_EXPAND_KILLED, /* Reached the proccess kill */
+ RAIDZ_EXPAND_CHECKED, /* Pool scrub verification done */
+} raidz_expand_test_state_t;
+
+
#define ZO_GVARS_MAX_ARGLEN ((size_t)64)
#define ZO_GVARS_MAX_COUNT ((size_t)10)
typedef struct ztest_shared_opts {
char zo_pool[ZFS_MAX_DATASET_NAME_LEN];
char zo_dir[ZFS_MAX_DATASET_NAME_LEN];
char zo_alt_ztest[MAXNAMELEN];
char zo_alt_libpath[MAXNAMELEN];
uint64_t zo_vdevs;
uint64_t zo_vdevtime;
size_t zo_vdev_size;
int zo_ashift;
int zo_mirrors;
+ int zo_raid_do_expand;
int zo_raid_children;
int zo_raid_parity;
char zo_raid_type[8];
int zo_draid_data;
int zo_draid_spares;
int zo_datasets;
int zo_threads;
uint64_t zo_passtime;
uint64_t zo_killrate;
int zo_verbose;
int zo_init;
uint64_t zo_time;
uint64_t zo_maxloops;
uint64_t zo_metaslab_force_ganging;
+ raidz_expand_test_state_t zo_raidz_expand_test;
int zo_mmp_test;
int zo_special_vdevs;
int zo_dump_dbgmsg;
int zo_gvars_count;
char zo_gvars[ZO_GVARS_MAX_COUNT][ZO_GVARS_MAX_ARGLEN];
} ztest_shared_opts_t;
/* Default values for command line options. */
#define DEFAULT_POOL "ztest"
#define DEFAULT_VDEV_DIR "/tmp"
#define DEFAULT_VDEV_COUNT 5
#define DEFAULT_VDEV_SIZE (SPA_MINDEVSIZE * 4) /* 256m default size */
#define DEFAULT_VDEV_SIZE_STR "256M"
#define DEFAULT_ASHIFT SPA_MINBLOCKSHIFT
#define DEFAULT_MIRRORS 2
#define DEFAULT_RAID_CHILDREN 4
#define DEFAULT_RAID_PARITY 1
#define DEFAULT_DRAID_DATA 4
#define DEFAULT_DRAID_SPARES 1
#define DEFAULT_DATASETS_COUNT 7
#define DEFAULT_THREADS 23
#define DEFAULT_RUN_TIME 300 /* 300 seconds */
#define DEFAULT_RUN_TIME_STR "300 sec"
#define DEFAULT_PASS_TIME 60 /* 60 seconds */
#define DEFAULT_PASS_TIME_STR "60 sec"
#define DEFAULT_KILL_RATE 70 /* 70% kill rate */
#define DEFAULT_KILLRATE_STR "70%"
#define DEFAULT_INITS 1
#define DEFAULT_MAX_LOOPS 50 /* 5 minutes */
#define DEFAULT_FORCE_GANGING (64 << 10)
#define DEFAULT_FORCE_GANGING_STR "64K"
/* Simplifying assumption: -1 is not a valid default. */
#define NO_DEFAULT -1
static const ztest_shared_opts_t ztest_opts_defaults = {
.zo_pool = DEFAULT_POOL,
.zo_dir = DEFAULT_VDEV_DIR,
.zo_alt_ztest = { '\0' },
.zo_alt_libpath = { '\0' },
.zo_vdevs = DEFAULT_VDEV_COUNT,
.zo_ashift = DEFAULT_ASHIFT,
.zo_mirrors = DEFAULT_MIRRORS,
.zo_raid_children = DEFAULT_RAID_CHILDREN,
.zo_raid_parity = DEFAULT_RAID_PARITY,
.zo_raid_type = VDEV_TYPE_RAIDZ,
.zo_vdev_size = DEFAULT_VDEV_SIZE,
.zo_draid_data = DEFAULT_DRAID_DATA, /* data drives */
.zo_draid_spares = DEFAULT_DRAID_SPARES, /* distributed spares */
.zo_datasets = DEFAULT_DATASETS_COUNT,
.zo_threads = DEFAULT_THREADS,
.zo_passtime = DEFAULT_PASS_TIME,
.zo_killrate = DEFAULT_KILL_RATE,
.zo_verbose = 0,
.zo_mmp_test = 0,
.zo_init = DEFAULT_INITS,
.zo_time = DEFAULT_RUN_TIME,
.zo_maxloops = DEFAULT_MAX_LOOPS, /* max loops during spa_freeze() */
.zo_metaslab_force_ganging = DEFAULT_FORCE_GANGING,
.zo_special_vdevs = ZTEST_VDEV_CLASS_RND,
.zo_gvars_count = 0,
+ .zo_raidz_expand_test = RAIDZ_EXPAND_NONE,
};
extern uint64_t metaslab_force_ganging;
extern uint64_t metaslab_df_alloc_threshold;
extern uint64_t zfs_deadman_synctime_ms;
extern uint_t metaslab_preload_limit;
extern int zfs_compressed_arc_enabled;
extern int zfs_abd_scatter_enabled;
extern uint_t dmu_object_alloc_chunk_shift;
extern boolean_t zfs_force_some_double_word_sm_entries;
extern unsigned long zio_decompress_fail_fraction;
extern unsigned long zfs_reconstruct_indirect_damage_fraction;
+extern uint64_t raidz_expand_max_reflow_bytes;
+extern uint_t raidz_expand_pause_point;
static ztest_shared_opts_t *ztest_shared_opts;
static ztest_shared_opts_t ztest_opts;
static const char *const ztest_wkeydata = "abcdefghijklmnopqrstuvwxyz012345";
typedef struct ztest_shared_ds {
uint64_t zd_seq;
} ztest_shared_ds_t;
static ztest_shared_ds_t *ztest_shared_ds;
#define ZTEST_GET_SHARED_DS(d) (&ztest_shared_ds[d])
+typedef struct ztest_scratch_state {
+ uint64_t zs_raidz_scratch_verify_pause;
+} ztest_shared_scratch_state_t;
+
+static ztest_shared_scratch_state_t *ztest_scratch_state;
+
#define BT_MAGIC 0x123456789abcdefULL
#define MAXFAULTS(zs) \
(MAX((zs)->zs_mirrors, 1) * (ztest_opts.zo_raid_parity + 1) - 1)
enum ztest_io_type {
ZTEST_IO_WRITE_TAG,
ZTEST_IO_WRITE_PATTERN,
ZTEST_IO_WRITE_ZEROES,
ZTEST_IO_TRUNCATE,
ZTEST_IO_SETATTR,
ZTEST_IO_REWRITE,
ZTEST_IO_TYPES
};
typedef struct ztest_block_tag {
uint64_t bt_magic;
uint64_t bt_objset;
uint64_t bt_object;
uint64_t bt_dnodesize;
uint64_t bt_offset;
uint64_t bt_gen;
uint64_t bt_txg;
uint64_t bt_crtxg;
} ztest_block_tag_t;
typedef struct bufwad {
uint64_t bw_index;
uint64_t bw_txg;
uint64_t bw_data;
} bufwad_t;
/*
* It would be better to use a rangelock_t per object. Unfortunately
* the rangelock_t is not a drop-in replacement for rl_t, because we
* still need to map from object ID to rangelock_t.
*/
typedef enum {
- RL_READER,
- RL_WRITER,
- RL_APPEND
+ ZTRL_READER,
+ ZTRL_WRITER,
+ ZTRL_APPEND
} rl_type_t;
typedef struct rll {
void *rll_writer;
int rll_readers;
kmutex_t rll_lock;
kcondvar_t rll_cv;
} rll_t;
typedef struct rl {
uint64_t rl_object;
uint64_t rl_offset;
uint64_t rl_size;
rll_t *rl_lock;
} rl_t;
#define ZTEST_RANGE_LOCKS 64
#define ZTEST_OBJECT_LOCKS 64
/*
* Object descriptor. Used as a template for object lookup/create/remove.
*/
typedef struct ztest_od {
uint64_t od_dir;
uint64_t od_object;
dmu_object_type_t od_type;
dmu_object_type_t od_crtype;
uint64_t od_blocksize;
uint64_t od_crblocksize;
uint64_t od_crdnodesize;
uint64_t od_gen;
uint64_t od_crgen;
char od_name[ZFS_MAX_DATASET_NAME_LEN];
} ztest_od_t;
/*
* Per-dataset state.
*/
typedef struct ztest_ds {
ztest_shared_ds_t *zd_shared;
objset_t *zd_os;
pthread_rwlock_t zd_zilog_lock;
zilog_t *zd_zilog;
ztest_od_t *zd_od; /* debugging aid */
char zd_name[ZFS_MAX_DATASET_NAME_LEN];
kmutex_t zd_dirobj_lock;
rll_t zd_object_lock[ZTEST_OBJECT_LOCKS];
rll_t zd_range_lock[ZTEST_RANGE_LOCKS];
} ztest_ds_t;
/*
* Per-iteration state.
*/
typedef void ztest_func_t(ztest_ds_t *zd, uint64_t id);
typedef struct ztest_info {
ztest_func_t *zi_func; /* test function */
uint64_t zi_iters; /* iterations per execution */
uint64_t *zi_interval; /* execute every <interval> seconds */
const char *zi_funcname; /* name of test function */
} ztest_info_t;
typedef struct ztest_shared_callstate {
uint64_t zc_count; /* per-pass count */
uint64_t zc_time; /* per-pass time */
uint64_t zc_next; /* next time to call this function */
} ztest_shared_callstate_t;
static ztest_shared_callstate_t *ztest_shared_callstate;
#define ZTEST_GET_SHARED_CALLSTATE(c) (&ztest_shared_callstate[c])
ztest_func_t ztest_dmu_read_write;
ztest_func_t ztest_dmu_write_parallel;
ztest_func_t ztest_dmu_object_alloc_free;
ztest_func_t ztest_dmu_object_next_chunk;
ztest_func_t ztest_dmu_commit_callbacks;
ztest_func_t ztest_zap;
ztest_func_t ztest_zap_parallel;
ztest_func_t ztest_zil_commit;
ztest_func_t ztest_zil_remount;
ztest_func_t ztest_dmu_read_write_zcopy;
ztest_func_t ztest_dmu_objset_create_destroy;
ztest_func_t ztest_dmu_prealloc;
ztest_func_t ztest_fzap;
ztest_func_t ztest_dmu_snapshot_create_destroy;
ztest_func_t ztest_dsl_prop_get_set;
ztest_func_t ztest_spa_prop_get_set;
ztest_func_t ztest_spa_create_destroy;
ztest_func_t ztest_fault_inject;
ztest_func_t ztest_dmu_snapshot_hold;
ztest_func_t ztest_mmp_enable_disable;
ztest_func_t ztest_scrub;
ztest_func_t ztest_dsl_dataset_promote_busy;
ztest_func_t ztest_vdev_attach_detach;
+ztest_func_t ztest_vdev_raidz_attach;
ztest_func_t ztest_vdev_LUN_growth;
ztest_func_t ztest_vdev_add_remove;
ztest_func_t ztest_vdev_class_add;
ztest_func_t ztest_vdev_aux_add_remove;
ztest_func_t ztest_split_pool;
ztest_func_t ztest_reguid;
ztest_func_t ztest_spa_upgrade;
ztest_func_t ztest_device_removal;
ztest_func_t ztest_spa_checkpoint_create_discard;
ztest_func_t ztest_initialize;
ztest_func_t ztest_trim;
ztest_func_t ztest_blake3;
ztest_func_t ztest_fletcher;
ztest_func_t ztest_fletcher_incr;
ztest_func_t ztest_verify_dnode_bt;
static uint64_t zopt_always = 0ULL * NANOSEC; /* all the time */
static uint64_t zopt_incessant = 1ULL * NANOSEC / 10; /* every 1/10 second */
static uint64_t zopt_often = 1ULL * NANOSEC; /* every second */
static uint64_t zopt_sometimes = 10ULL * NANOSEC; /* every 10 seconds */
static uint64_t zopt_rarely = 60ULL * NANOSEC; /* every 60 seconds */
#define ZTI_INIT(func, iters, interval) \
{ .zi_func = (func), \
.zi_iters = (iters), \
.zi_interval = (interval), \
.zi_funcname = # func }
static ztest_info_t ztest_info[] = {
ZTI_INIT(ztest_dmu_read_write, 1, &zopt_always),
ZTI_INIT(ztest_dmu_write_parallel, 10, &zopt_always),
ZTI_INIT(ztest_dmu_object_alloc_free, 1, &zopt_always),
ZTI_INIT(ztest_dmu_object_next_chunk, 1, &zopt_sometimes),
ZTI_INIT(ztest_dmu_commit_callbacks, 1, &zopt_always),
ZTI_INIT(ztest_zap, 30, &zopt_always),
ZTI_INIT(ztest_zap_parallel, 100, &zopt_always),
ZTI_INIT(ztest_split_pool, 1, &zopt_sometimes),
ZTI_INIT(ztest_zil_commit, 1, &zopt_incessant),
ZTI_INIT(ztest_zil_remount, 1, &zopt_sometimes),
ZTI_INIT(ztest_dmu_read_write_zcopy, 1, &zopt_often),
ZTI_INIT(ztest_dmu_objset_create_destroy, 1, &zopt_often),
ZTI_INIT(ztest_dsl_prop_get_set, 1, &zopt_often),
ZTI_INIT(ztest_spa_prop_get_set, 1, &zopt_sometimes),
#if 0
ZTI_INIT(ztest_dmu_prealloc, 1, &zopt_sometimes),
#endif
ZTI_INIT(ztest_fzap, 1, &zopt_sometimes),
ZTI_INIT(ztest_dmu_snapshot_create_destroy, 1, &zopt_sometimes),
ZTI_INIT(ztest_spa_create_destroy, 1, &zopt_sometimes),
ZTI_INIT(ztest_fault_inject, 1, &zopt_sometimes),
ZTI_INIT(ztest_dmu_snapshot_hold, 1, &zopt_sometimes),
ZTI_INIT(ztest_mmp_enable_disable, 1, &zopt_sometimes),
ZTI_INIT(ztest_reguid, 1, &zopt_rarely),
ZTI_INIT(ztest_scrub, 1, &zopt_rarely),
ZTI_INIT(ztest_spa_upgrade, 1, &zopt_rarely),
ZTI_INIT(ztest_dsl_dataset_promote_busy, 1, &zopt_rarely),
ZTI_INIT(ztest_vdev_attach_detach, 1, &zopt_sometimes),
+ ZTI_INIT(ztest_vdev_raidz_attach, 1, &zopt_sometimes),
ZTI_INIT(ztest_vdev_LUN_growth, 1, &zopt_rarely),
ZTI_INIT(ztest_vdev_add_remove, 1, &ztest_opts.zo_vdevtime),
ZTI_INIT(ztest_vdev_class_add, 1, &ztest_opts.zo_vdevtime),
ZTI_INIT(ztest_vdev_aux_add_remove, 1, &ztest_opts.zo_vdevtime),
ZTI_INIT(ztest_device_removal, 1, &zopt_sometimes),
ZTI_INIT(ztest_spa_checkpoint_create_discard, 1, &zopt_rarely),
ZTI_INIT(ztest_initialize, 1, &zopt_sometimes),
ZTI_INIT(ztest_trim, 1, &zopt_sometimes),
ZTI_INIT(ztest_blake3, 1, &zopt_rarely),
ZTI_INIT(ztest_fletcher, 1, &zopt_rarely),
ZTI_INIT(ztest_fletcher_incr, 1, &zopt_rarely),
ZTI_INIT(ztest_verify_dnode_bt, 1, &zopt_sometimes),
};
#define ZTEST_FUNCS (sizeof (ztest_info) / sizeof (ztest_info_t))
/*
* The following struct is used to hold a list of uncalled commit callbacks.
* The callbacks are ordered by txg number.
*/
typedef struct ztest_cb_list {
kmutex_t zcl_callbacks_lock;
list_t zcl_callbacks;
} ztest_cb_list_t;
/*
* Stuff we need to share writably between parent and child.
*/
typedef struct ztest_shared {
boolean_t zs_do_init;
hrtime_t zs_proc_start;
hrtime_t zs_proc_stop;
hrtime_t zs_thread_start;
hrtime_t zs_thread_stop;
hrtime_t zs_thread_kill;
uint64_t zs_enospc_count;
uint64_t zs_vdev_next_leaf;
uint64_t zs_vdev_aux;
uint64_t zs_alloc;
uint64_t zs_space;
uint64_t zs_splits;
uint64_t zs_mirrors;
uint64_t zs_metaslab_sz;
uint64_t zs_metaslab_df_alloc_threshold;
uint64_t zs_guid;
} ztest_shared_t;
#define ID_PARALLEL -1ULL
static char ztest_dev_template[] = "%s/%s.%llua";
static char ztest_aux_template[] = "%s/%s.%s.%llu";
static ztest_shared_t *ztest_shared;
static spa_t *ztest_spa = NULL;
static ztest_ds_t *ztest_ds;
static kmutex_t ztest_vdev_lock;
static boolean_t ztest_device_removal_active = B_FALSE;
static boolean_t ztest_pool_scrubbed = B_FALSE;
static kmutex_t ztest_checkpoint_lock;
/*
* The ztest_name_lock protects the pool and dataset namespace used by
* the individual tests. To modify the namespace, consumers must grab
* this lock as writer. Grabbing the lock as reader will ensure that the
* namespace does not change while the lock is held.
*/
static pthread_rwlock_t ztest_name_lock;
static boolean_t ztest_dump_core = B_TRUE;
static boolean_t ztest_exiting;
/* Global commit callback list */
static ztest_cb_list_t zcl;
/* Commit cb delay */
static uint64_t zc_min_txg_delay = UINT64_MAX;
static int zc_cb_counter = 0;
/*
* Minimum number of commit callbacks that need to be registered for us to check
* whether the minimum txg delay is acceptable.
*/
#define ZTEST_COMMIT_CB_MIN_REG 100
/*
* If a number of txgs equal to this threshold have been created after a commit
* callback has been registered but not called, then we assume there is an
* implementation bug.
*/
#define ZTEST_COMMIT_CB_THRESH (TXG_CONCURRENT_STATES + 1000)
enum ztest_object {
ZTEST_META_DNODE = 0,
ZTEST_DIROBJ,
ZTEST_OBJECTS
};
static __attribute__((noreturn)) void usage(boolean_t requested);
static int ztest_scrub_impl(spa_t *spa);
/*
* These libumem hooks provide a reasonable set of defaults for the allocator's
* debugging facilities.
*/
const char *
_umem_debug_init(void)
{
return ("default,verbose"); /* $UMEM_DEBUG setting */
}
const char *
_umem_logging_init(void)
{
return ("fail,contents"); /* $UMEM_LOGGING setting */
}
static void
dump_debug_buffer(void)
{
ssize_t ret __attribute__((unused));
if (!ztest_opts.zo_dump_dbgmsg)
return;
/*
* We use write() instead of printf() so that this function
* is safe to call from a signal handler.
*/
ret = write(STDOUT_FILENO, "\n", 1);
zfs_dbgmsg_print("ztest");
}
#define BACKTRACE_SZ 100
static void sig_handler(int signo)
{
struct sigaction action;
#if (__GLIBC__ && !__UCLIBC__) /* backtrace() is a GNU extension */
int nptrs;
void *buffer[BACKTRACE_SZ];
nptrs = backtrace(buffer, BACKTRACE_SZ);
backtrace_symbols_fd(buffer, nptrs, STDERR_FILENO);
#endif
dump_debug_buffer();
/*
* Restore default action and re-raise signal so SIGSEGV and
* SIGABRT can trigger a core dump.
*/
action.sa_handler = SIG_DFL;
sigemptyset(&action.sa_mask);
action.sa_flags = 0;
(void) sigaction(signo, &action, NULL);
raise(signo);
}
#define FATAL_MSG_SZ 1024
static const char *fatal_msg;
static __attribute__((format(printf, 2, 3))) __attribute__((noreturn)) void
fatal(int do_perror, const char *message, ...)
{
va_list args;
int save_errno = errno;
char *buf;
(void) fflush(stdout);
buf = umem_alloc(FATAL_MSG_SZ, UMEM_NOFAIL);
if (buf == NULL)
goto out;
va_start(args, message);
(void) sprintf(buf, "ztest: ");
/* LINTED */
(void) vsprintf(buf + strlen(buf), message, args);
va_end(args);
if (do_perror) {
(void) snprintf(buf + strlen(buf), FATAL_MSG_SZ - strlen(buf),
": %s", strerror(save_errno));
}
(void) fprintf(stderr, "%s\n", buf);
fatal_msg = buf; /* to ease debugging */
out:
if (ztest_dump_core)
abort();
else
dump_debug_buffer();
exit(3);
}
static int
str2shift(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)) {
(void) fprintf(stderr, "ztest: invalid bytes suffix: %s\n",
buf);
usage(B_FALSE);
}
if (buf[1] == '\0' || (toupper(buf[1]) == 'B' && buf[2] == '\0')) {
return (10*i);
}
(void) fprintf(stderr, "ztest: invalid bytes suffix: %s\n", buf);
usage(B_FALSE);
}
static uint64_t
nicenumtoull(const char *buf)
{
char *end;
uint64_t val;
val = strtoull(buf, &end, 0);
if (end == buf) {
(void) fprintf(stderr, "ztest: bad numeric value: %s\n", buf);
usage(B_FALSE);
} else if (end[0] == '.') {
double fval = strtod(buf, &end);
fval *= pow(2, str2shift(end));
/*
* UINT64_MAX is not exactly representable as a double.
* The closest representation is UINT64_MAX + 1, so we
* use a >= comparison instead of > for the bounds check.
*/
if (fval >= (double)UINT64_MAX) {
(void) fprintf(stderr, "ztest: value too large: %s\n",
buf);
usage(B_FALSE);
}
val = (uint64_t)fval;
} else {
int shift = str2shift(end);
if (shift >= 64 || (val << shift) >> shift != val) {
(void) fprintf(stderr, "ztest: value too large: %s\n",
buf);
usage(B_FALSE);
}
val <<= shift;
}
return (val);
}
typedef struct ztest_option {
const char short_opt;
const char *long_opt;
const char *long_opt_param;
const char *comment;
unsigned int default_int;
const char *default_str;
} ztest_option_t;
/*
* The following option_table is used for generating the usage info as well as
* the long and short option information for calling getopt_long().
*/
static ztest_option_t option_table[] = {
{ 'v', "vdevs", "INTEGER", "Number of vdevs", DEFAULT_VDEV_COUNT,
NULL},
{ 's', "vdev-size", "INTEGER", "Size of each vdev",
NO_DEFAULT, DEFAULT_VDEV_SIZE_STR},
{ 'a', "alignment-shift", "INTEGER",
"Alignment shift; use 0 for random", DEFAULT_ASHIFT, NULL},
{ 'm', "mirror-copies", "INTEGER", "Number of mirror copies",
DEFAULT_MIRRORS, NULL},
{ 'r', "raid-disks", "INTEGER", "Number of raidz/draid disks",
DEFAULT_RAID_CHILDREN, NULL},
{ 'R', "raid-parity", "INTEGER", "Raid parity",
DEFAULT_RAID_PARITY, NULL},
- { 'K', "raid-kind", "raidz|draid|random", "Raid kind",
+ { 'K', "raid-kind", "raidz|eraidz|draid|random", "Raid kind",
NO_DEFAULT, "random"},
{ 'D', "draid-data", "INTEGER", "Number of draid data drives",
DEFAULT_DRAID_DATA, NULL},
{ 'S', "draid-spares", "INTEGER", "Number of draid spares",
DEFAULT_DRAID_SPARES, NULL},
{ 'd', "datasets", "INTEGER", "Number of datasets",
DEFAULT_DATASETS_COUNT, NULL},
{ 't', "threads", "INTEGER", "Number of ztest threads",
DEFAULT_THREADS, NULL},
{ 'g', "gang-block-threshold", "INTEGER",
"Metaslab gang block threshold",
NO_DEFAULT, DEFAULT_FORCE_GANGING_STR},
{ 'i', "init-count", "INTEGER", "Number of times to initialize pool",
DEFAULT_INITS, NULL},
{ 'k', "kill-percentage", "INTEGER", "Kill percentage",
NO_DEFAULT, DEFAULT_KILLRATE_STR},
{ 'p', "pool-name", "STRING", "Pool name",
NO_DEFAULT, DEFAULT_POOL},
{ 'f', "vdev-file-directory", "PATH", "File directory for vdev files",
NO_DEFAULT, DEFAULT_VDEV_DIR},
{ 'M', "multi-host", NULL,
"Multi-host; simulate pool imported on remote host",
NO_DEFAULT, NULL},
{ 'E', "use-existing-pool", NULL,
"Use existing pool instead of creating new one", NO_DEFAULT, NULL},
{ 'T', "run-time", "INTEGER", "Total run time",
NO_DEFAULT, DEFAULT_RUN_TIME_STR},
{ 'P', "pass-time", "INTEGER", "Time per pass",
NO_DEFAULT, DEFAULT_PASS_TIME_STR},
{ 'F', "freeze-loops", "INTEGER", "Max loops in spa_freeze()",
DEFAULT_MAX_LOOPS, NULL},
{ 'B', "alt-ztest", "PATH", "Alternate ztest path",
NO_DEFAULT, NULL},
{ 'C', "vdev-class-state", "on|off|random", "vdev class state",
NO_DEFAULT, "random"},
+ { 'X', "raidz-expansion", NULL,
+ "Perform a dedicated raidz expansion test",
+ NO_DEFAULT, NULL},
{ 'o', "option", "\"OPTION=INTEGER\"",
"Set global variable to an unsigned 32-bit integer value",
NO_DEFAULT, NULL},
{ 'G', "dump-debug-msg", NULL,
"Dump zfs_dbgmsg buffer before exiting due to an error",
NO_DEFAULT, NULL},
{ 'V', "verbose", NULL,
"Verbose (use multiple times for ever more verbosity)",
NO_DEFAULT, NULL},
{ 'h', "help", NULL, "Show this help",
NO_DEFAULT, NULL},
{0, 0, 0, 0, 0, 0}
};
static struct option *long_opts = NULL;
static char *short_opts = NULL;
static void
init_options(void)
{
ASSERT3P(long_opts, ==, NULL);
ASSERT3P(short_opts, ==, NULL);
int count = sizeof (option_table) / sizeof (option_table[0]);
long_opts = umem_alloc(sizeof (struct option) * count, UMEM_NOFAIL);
short_opts = umem_alloc(sizeof (char) * 2 * count, UMEM_NOFAIL);
int short_opt_index = 0;
for (int i = 0; i < count; i++) {
long_opts[i].val = option_table[i].short_opt;
long_opts[i].name = option_table[i].long_opt;
long_opts[i].has_arg = option_table[i].long_opt_param != NULL
? required_argument : no_argument;
long_opts[i].flag = NULL;
short_opts[short_opt_index++] = option_table[i].short_opt;
if (option_table[i].long_opt_param != NULL) {
short_opts[short_opt_index++] = ':';
}
}
}
static void
fini_options(void)
{
int count = sizeof (option_table) / sizeof (option_table[0]);
umem_free(long_opts, sizeof (struct option) * count);
umem_free(short_opts, sizeof (char) * 2 * count);
long_opts = NULL;
short_opts = NULL;
}
static __attribute__((noreturn)) void
usage(boolean_t requested)
{
char option[80];
FILE *fp = requested ? stdout : stderr;
(void) fprintf(fp, "Usage: %s [OPTIONS...]\n", DEFAULT_POOL);
for (int i = 0; option_table[i].short_opt != 0; i++) {
if (option_table[i].long_opt_param != NULL) {
(void) sprintf(option, " -%c --%s=%s",
option_table[i].short_opt,
option_table[i].long_opt,
option_table[i].long_opt_param);
} else {
(void) sprintf(option, " -%c --%s",
option_table[i].short_opt,
option_table[i].long_opt);
}
- (void) fprintf(fp, " %-40s%s", option,
+ (void) fprintf(fp, " %-43s%s", option,
option_table[i].comment);
if (option_table[i].long_opt_param != NULL) {
if (option_table[i].default_str != NULL) {
(void) fprintf(fp, " (default: %s)",
option_table[i].default_str);
} else if (option_table[i].default_int != NO_DEFAULT) {
(void) fprintf(fp, " (default: %u)",
option_table[i].default_int);
}
}
(void) fprintf(fp, "\n");
}
exit(requested ? 0 : 1);
}
static uint64_t
ztest_random(uint64_t range)
{
uint64_t r;
ASSERT3S(ztest_fd_rand, >=, 0);
if (range == 0)
return (0);
if (read(ztest_fd_rand, &r, sizeof (r)) != sizeof (r))
fatal(B_TRUE, "short read from /dev/urandom");
return (r % range);
}
static void
ztest_parse_name_value(const char *input, ztest_shared_opts_t *zo)
{
char name[32];
char *value;
int state = ZTEST_VDEV_CLASS_RND;
(void) strlcpy(name, input, sizeof (name));
value = strchr(name, '=');
if (value == NULL) {
(void) fprintf(stderr, "missing value in property=value "
"'-C' argument (%s)\n", input);
usage(B_FALSE);
}
*(value) = '\0';
value++;
if (strcmp(value, "on") == 0) {
state = ZTEST_VDEV_CLASS_ON;
} else if (strcmp(value, "off") == 0) {
state = ZTEST_VDEV_CLASS_OFF;
} else if (strcmp(value, "random") == 0) {
state = ZTEST_VDEV_CLASS_RND;
} else {
(void) fprintf(stderr, "invalid property value '%s'\n", value);
usage(B_FALSE);
}
if (strcmp(name, "special") == 0) {
zo->zo_special_vdevs = state;
} else {
(void) fprintf(stderr, "invalid property name '%s'\n", name);
usage(B_FALSE);
}
if (zo->zo_verbose >= 3)
(void) printf("%s vdev state is '%s'\n", name, value);
}
static void
process_options(int argc, char **argv)
{
char *path;
ztest_shared_opts_t *zo = &ztest_opts;
int opt;
uint64_t value;
const char *raid_kind = "random";
memcpy(zo, &ztest_opts_defaults, sizeof (*zo));
init_options();
while ((opt = getopt_long(argc, argv, short_opts, long_opts,
NULL)) != EOF) {
value = 0;
switch (opt) {
case 'v':
case 's':
case 'a':
case 'm':
case 'r':
case 'R':
case 'D':
case 'S':
case 'd':
case 't':
case 'g':
case 'i':
case 'k':
case 'T':
case 'P':
case 'F':
value = nicenumtoull(optarg);
}
switch (opt) {
case 'v':
zo->zo_vdevs = value;
break;
case 's':
zo->zo_vdev_size = MAX(SPA_MINDEVSIZE, value);
break;
case 'a':
zo->zo_ashift = value;
break;
case 'm':
zo->zo_mirrors = value;
break;
case 'r':
zo->zo_raid_children = MAX(1, value);
break;
case 'R':
zo->zo_raid_parity = MIN(MAX(value, 1), 3);
break;
case 'K':
raid_kind = optarg;
break;
case 'D':
zo->zo_draid_data = MAX(1, value);
break;
case 'S':
zo->zo_draid_spares = MAX(1, value);
break;
case 'd':
zo->zo_datasets = MAX(1, value);
break;
case 't':
zo->zo_threads = MAX(1, value);
break;
case 'g':
zo->zo_metaslab_force_ganging =
MAX(SPA_MINBLOCKSIZE << 1, value);
break;
case 'i':
zo->zo_init = value;
break;
case 'k':
zo->zo_killrate = value;
break;
case 'p':
(void) strlcpy(zo->zo_pool, optarg,
sizeof (zo->zo_pool));
break;
case 'f':
path = realpath(optarg, NULL);
if (path == NULL) {
(void) fprintf(stderr, "error: %s: %s\n",
optarg, strerror(errno));
usage(B_FALSE);
} else {
(void) strlcpy(zo->zo_dir, path,
sizeof (zo->zo_dir));
free(path);
}
break;
case 'M':
zo->zo_mmp_test = 1;
break;
case 'V':
zo->zo_verbose++;
break;
+ case 'X':
+ zo->zo_raidz_expand_test = RAIDZ_EXPAND_REQUESTED;
+ break;
case 'E':
zo->zo_init = 0;
break;
case 'T':
zo->zo_time = value;
break;
case 'P':
zo->zo_passtime = MAX(1, value);
break;
case 'F':
zo->zo_maxloops = MAX(1, value);
break;
case 'B':
(void) strlcpy(zo->zo_alt_ztest, optarg,
sizeof (zo->zo_alt_ztest));
break;
case 'C':
ztest_parse_name_value(optarg, zo);
break;
case 'o':
if (zo->zo_gvars_count >= ZO_GVARS_MAX_COUNT) {
(void) fprintf(stderr,
"max global var count (%zu) exceeded\n",
ZO_GVARS_MAX_COUNT);
usage(B_FALSE);
}
char *v = zo->zo_gvars[zo->zo_gvars_count];
if (strlcpy(v, optarg, ZO_GVARS_MAX_ARGLEN) >=
ZO_GVARS_MAX_ARGLEN) {
(void) fprintf(stderr,
"global var option '%s' is too long\n",
optarg);
usage(B_FALSE);
}
zo->zo_gvars_count++;
break;
case 'G':
zo->zo_dump_dbgmsg = 1;
break;
case 'h':
usage(B_TRUE);
break;
case '?':
default:
usage(B_FALSE);
break;
}
}
fini_options();
- /* When raid choice is 'random' add a draid pool 50% of the time */
+ /* Force compatible options for raidz expansion run */
+ if (zo->zo_raidz_expand_test == RAIDZ_EXPAND_REQUESTED) {
+ zo->zo_mmp_test = 0;
+ zo->zo_mirrors = 0;
+ zo->zo_vdevs = 1;
+ zo->zo_vdev_size = DEFAULT_VDEV_SIZE * 2;
+ zo->zo_raid_do_expand = B_FALSE;
+ raid_kind = "raidz";
+ }
+
if (strcmp(raid_kind, "random") == 0) {
- raid_kind = (ztest_random(2) == 0) ? "draid" : "raidz";
+ switch (ztest_random(3)) {
+ case 0:
+ raid_kind = "raidz";
+ break;
+ case 1:
+ raid_kind = "eraidz";
+ break;
+ case 2:
+ raid_kind = "draid";
+ break;
+ }
if (ztest_opts.zo_verbose >= 3)
(void) printf("choosing RAID type '%s'\n", raid_kind);
}
if (strcmp(raid_kind, "draid") == 0) {
uint64_t min_devsize;
/* With fewer disk use 256M, otherwise 128M is OK */
min_devsize = (ztest_opts.zo_raid_children < 16) ?
(256ULL << 20) : (128ULL << 20);
/* No top-level mirrors with dRAID for now */
zo->zo_mirrors = 0;
/* Use more appropriate defaults for dRAID */
if (zo->zo_vdevs == ztest_opts_defaults.zo_vdevs)
zo->zo_vdevs = 1;
if (zo->zo_raid_children ==
ztest_opts_defaults.zo_raid_children)
zo->zo_raid_children = 16;
if (zo->zo_ashift < 12)
zo->zo_ashift = 12;
if (zo->zo_vdev_size < min_devsize)
zo->zo_vdev_size = min_devsize;
if (zo->zo_draid_data + zo->zo_raid_parity >
zo->zo_raid_children - zo->zo_draid_spares) {
(void) fprintf(stderr, "error: too few draid "
"children (%d) for stripe width (%d)\n",
zo->zo_raid_children,
zo->zo_draid_data + zo->zo_raid_parity);
usage(B_FALSE);
}
(void) strlcpy(zo->zo_raid_type, VDEV_TYPE_DRAID,
sizeof (zo->zo_raid_type));
+ } else if (strcmp(raid_kind, "eraidz") == 0) {
+ /* using eraidz (expandable raidz) */
+ zo->zo_raid_do_expand = B_TRUE;
+
+ /* tests expect top-level to be raidz */
+ zo->zo_mirrors = 0;
+ zo->zo_vdevs = 1;
+
+ /* Make sure parity is less than data columns */
+ zo->zo_raid_parity = MIN(zo->zo_raid_parity,
+ zo->zo_raid_children - 1);
+
} else /* using raidz */ {
ASSERT0(strcmp(raid_kind, "raidz"));
zo->zo_raid_parity = MIN(zo->zo_raid_parity,
zo->zo_raid_children - 1);
}
zo->zo_vdevtime =
(zo->zo_vdevs > 0 ? zo->zo_time * NANOSEC / zo->zo_vdevs :
UINT64_MAX >> 2);
if (*zo->zo_alt_ztest) {
const char *invalid_what = "ztest";
char *val = zo->zo_alt_ztest;
if (0 != access(val, X_OK) ||
(strrchr(val, '/') == NULL && (errno == EINVAL)))
goto invalid;
int dirlen = strrchr(val, '/') - val;
strlcpy(zo->zo_alt_libpath, val,
MIN(sizeof (zo->zo_alt_libpath), dirlen + 1));
invalid_what = "library path", val = zo->zo_alt_libpath;
if (strrchr(val, '/') == NULL && (errno == EINVAL))
goto invalid;
*strrchr(val, '/') = '\0';
strlcat(val, "/lib", sizeof (zo->zo_alt_libpath));
if (0 != access(zo->zo_alt_libpath, X_OK))
goto invalid;
return;
invalid:
ztest_dump_core = B_FALSE;
fatal(B_TRUE, "invalid alternate %s %s", invalid_what, val);
}
}
static void
ztest_kill(ztest_shared_t *zs)
{
zs->zs_alloc = metaslab_class_get_alloc(spa_normal_class(ztest_spa));
zs->zs_space = metaslab_class_get_space(spa_normal_class(ztest_spa));
/*
* Before we kill ourselves, make sure that the config is updated.
* See comment above spa_write_cachefile().
*/
- mutex_enter(&spa_namespace_lock);
- spa_write_cachefile(ztest_spa, B_FALSE, B_FALSE, B_FALSE);
- mutex_exit(&spa_namespace_lock);
+ if (raidz_expand_pause_point != RAIDZ_EXPAND_PAUSE_NONE) {
+ if (mutex_tryenter(&spa_namespace_lock)) {
+ spa_write_cachefile(ztest_spa, B_FALSE, B_FALSE,
+ B_FALSE);
+ mutex_exit(&spa_namespace_lock);
+
+ ztest_scratch_state->zs_raidz_scratch_verify_pause =
+ raidz_expand_pause_point;
+ } else {
+ /*
+ * Do not verify scratch object in case if
+ * spa_namespace_lock cannot be acquired,
+ * it can cause deadlock in spa_config_update().
+ */
+ raidz_expand_pause_point = RAIDZ_EXPAND_PAUSE_NONE;
+
+ return;
+ }
+ } else {
+ mutex_enter(&spa_namespace_lock);
+ spa_write_cachefile(ztest_spa, B_FALSE, B_FALSE, B_FALSE);
+ mutex_exit(&spa_namespace_lock);
+ }
(void) raise(SIGKILL);
}
static void
ztest_record_enospc(const char *s)
{
(void) s;
ztest_shared->zs_enospc_count++;
}
static uint64_t
ztest_get_ashift(void)
{
if (ztest_opts.zo_ashift == 0)
return (SPA_MINBLOCKSHIFT + ztest_random(5));
return (ztest_opts.zo_ashift);
}
static boolean_t
ztest_is_draid_spare(const char *name)
{
uint64_t spare_id = 0, parity = 0, vdev_id = 0;
if (sscanf(name, VDEV_TYPE_DRAID "%"PRIu64"-%"PRIu64"-%"PRIu64"",
&parity, &vdev_id, &spare_id) == 3) {
return (B_TRUE);
}
return (B_FALSE);
}
static nvlist_t *
make_vdev_file(const char *path, const char *aux, const char *pool,
size_t size, uint64_t ashift)
{
char *pathbuf = NULL;
uint64_t vdev;
nvlist_t *file;
boolean_t draid_spare = B_FALSE;
if (ashift == 0)
ashift = ztest_get_ashift();
if (path == NULL) {
pathbuf = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
path = pathbuf;
if (aux != NULL) {
vdev = ztest_shared->zs_vdev_aux;
(void) snprintf(pathbuf, MAXPATHLEN,
ztest_aux_template, ztest_opts.zo_dir,
pool == NULL ? ztest_opts.zo_pool : pool,
aux, vdev);
} else {
vdev = ztest_shared->zs_vdev_next_leaf++;
(void) snprintf(pathbuf, MAXPATHLEN,
ztest_dev_template, ztest_opts.zo_dir,
pool == NULL ? ztest_opts.zo_pool : pool, vdev);
}
} else {
draid_spare = ztest_is_draid_spare(path);
}
if (size != 0 && !draid_spare) {
int fd = open(path, O_RDWR | O_CREAT | O_TRUNC, 0666);
if (fd == -1)
fatal(B_TRUE, "can't open %s", path);
if (ftruncate(fd, size) != 0)
fatal(B_TRUE, "can't ftruncate %s", path);
(void) close(fd);
}
file = fnvlist_alloc();
fnvlist_add_string(file, ZPOOL_CONFIG_TYPE,
draid_spare ? VDEV_TYPE_DRAID_SPARE : VDEV_TYPE_FILE);
fnvlist_add_string(file, ZPOOL_CONFIG_PATH, path);
fnvlist_add_uint64(file, ZPOOL_CONFIG_ASHIFT, ashift);
umem_free(pathbuf, MAXPATHLEN);
return (file);
}
static nvlist_t *
make_vdev_raid(const char *path, const char *aux, const char *pool, size_t size,
uint64_t ashift, int r)
{
nvlist_t *raid, **child;
int c;
if (r < 2)
return (make_vdev_file(path, aux, pool, size, ashift));
child = umem_alloc(r * sizeof (nvlist_t *), UMEM_NOFAIL);
for (c = 0; c < r; c++)
child[c] = make_vdev_file(path, aux, pool, size, ashift);
raid = fnvlist_alloc();
fnvlist_add_string(raid, ZPOOL_CONFIG_TYPE,
ztest_opts.zo_raid_type);
fnvlist_add_uint64(raid, ZPOOL_CONFIG_NPARITY,
ztest_opts.zo_raid_parity);
fnvlist_add_nvlist_array(raid, ZPOOL_CONFIG_CHILDREN,
(const nvlist_t **)child, r);
if (strcmp(ztest_opts.zo_raid_type, VDEV_TYPE_DRAID) == 0) {
uint64_t ndata = ztest_opts.zo_draid_data;
uint64_t nparity = ztest_opts.zo_raid_parity;
uint64_t nspares = ztest_opts.zo_draid_spares;
uint64_t children = ztest_opts.zo_raid_children;
uint64_t ngroups = 1;
/*
* Calculate the minimum number of groups required to fill a
* slice. This is the LCM of the stripe width (data + parity)
* and the number of data drives (children - spares).
*/
while (ngroups * (ndata + nparity) % (children - nspares) != 0)
ngroups++;
/* Store the basic dRAID configuration. */
fnvlist_add_uint64(raid, ZPOOL_CONFIG_DRAID_NDATA, ndata);
fnvlist_add_uint64(raid, ZPOOL_CONFIG_DRAID_NSPARES, nspares);
fnvlist_add_uint64(raid, ZPOOL_CONFIG_DRAID_NGROUPS, ngroups);
}
for (c = 0; c < r; c++)
fnvlist_free(child[c]);
umem_free(child, r * sizeof (nvlist_t *));
return (raid);
}
static nvlist_t *
make_vdev_mirror(const char *path, const char *aux, const char *pool,
size_t size, uint64_t ashift, int r, int m)
{
nvlist_t *mirror, **child;
int c;
if (m < 1)
return (make_vdev_raid(path, aux, pool, size, ashift, r));
child = umem_alloc(m * sizeof (nvlist_t *), UMEM_NOFAIL);
for (c = 0; c < m; c++)
child[c] = make_vdev_raid(path, aux, pool, size, ashift, r);
mirror = fnvlist_alloc();
fnvlist_add_string(mirror, ZPOOL_CONFIG_TYPE, VDEV_TYPE_MIRROR);
fnvlist_add_nvlist_array(mirror, ZPOOL_CONFIG_CHILDREN,
(const nvlist_t **)child, m);
for (c = 0; c < m; c++)
fnvlist_free(child[c]);
umem_free(child, m * sizeof (nvlist_t *));
return (mirror);
}
static nvlist_t *
make_vdev_root(const char *path, const char *aux, const char *pool, size_t size,
uint64_t ashift, const char *class, int r, int m, int t)
{
nvlist_t *root, **child;
int c;
boolean_t log;
ASSERT3S(t, >, 0);
log = (class != NULL && strcmp(class, "log") == 0);
child = umem_alloc(t * sizeof (nvlist_t *), UMEM_NOFAIL);
for (c = 0; c < t; c++) {
child[c] = make_vdev_mirror(path, aux, pool, size, ashift,
r, m);
fnvlist_add_uint64(child[c], ZPOOL_CONFIG_IS_LOG, log);
if (class != NULL && class[0] != '\0') {
ASSERT(m > 1 || log); /* expecting a mirror */
fnvlist_add_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, class);
}
}
root = fnvlist_alloc();
fnvlist_add_string(root, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT);
fnvlist_add_nvlist_array(root, aux ? aux : ZPOOL_CONFIG_CHILDREN,
(const nvlist_t **)child, t);
for (c = 0; c < t; c++)
fnvlist_free(child[c]);
umem_free(child, t * sizeof (nvlist_t *));
return (root);
}
/*
* Find a random spa version. Returns back a random spa version in the
* range [initial_version, SPA_VERSION_FEATURES].
*/
static uint64_t
ztest_random_spa_version(uint64_t initial_version)
{
uint64_t version = initial_version;
if (version <= SPA_VERSION_BEFORE_FEATURES) {
version = version +
ztest_random(SPA_VERSION_BEFORE_FEATURES - version + 1);
}
if (version > SPA_VERSION_BEFORE_FEATURES)
version = SPA_VERSION_FEATURES;
ASSERT(SPA_VERSION_IS_SUPPORTED(version));
return (version);
}
static int
ztest_random_blocksize(void)
{
ASSERT3U(ztest_spa->spa_max_ashift, !=, 0);
/*
* Choose a block size >= the ashift.
* If the SPA supports new MAXBLOCKSIZE, test up to 1MB blocks.
*/
int maxbs = SPA_OLD_MAXBLOCKSHIFT;
if (spa_maxblocksize(ztest_spa) == SPA_MAXBLOCKSIZE)
maxbs = 20;
uint64_t block_shift =
ztest_random(maxbs - ztest_spa->spa_max_ashift + 1);
return (1 << (SPA_MINBLOCKSHIFT + block_shift));
}
static int
ztest_random_dnodesize(void)
{
int slots;
int max_slots = spa_maxdnodesize(ztest_spa) >> DNODE_SHIFT;
if (max_slots == DNODE_MIN_SLOTS)
return (DNODE_MIN_SIZE);
/*
* Weight the random distribution more heavily toward smaller
* dnode sizes since that is more likely to reflect real-world
* usage.
*/
ASSERT3U(max_slots, >, 4);
switch (ztest_random(10)) {
case 0:
slots = 5 + ztest_random(max_slots - 4);
break;
case 1 ... 4:
slots = 2 + ztest_random(3);
break;
default:
slots = 1;
break;
}
return (slots << DNODE_SHIFT);
}
static int
ztest_random_ibshift(void)
{
return (DN_MIN_INDBLKSHIFT +
ztest_random(DN_MAX_INDBLKSHIFT - DN_MIN_INDBLKSHIFT + 1));
}
static uint64_t
ztest_random_vdev_top(spa_t *spa, boolean_t log_ok)
{
uint64_t top;
vdev_t *rvd = spa->spa_root_vdev;
vdev_t *tvd;
ASSERT3U(spa_config_held(spa, SCL_ALL, RW_READER), !=, 0);
do {
top = ztest_random(rvd->vdev_children);
tvd = rvd->vdev_child[top];
} while (!vdev_is_concrete(tvd) || (tvd->vdev_islog && !log_ok) ||
tvd->vdev_mg == NULL || tvd->vdev_mg->mg_class == NULL);
return (top);
}
static uint64_t
ztest_random_dsl_prop(zfs_prop_t prop)
{
uint64_t value;
do {
value = zfs_prop_random_value(prop, ztest_random(-1ULL));
} while (prop == ZFS_PROP_CHECKSUM && value == ZIO_CHECKSUM_OFF);
return (value);
}
static int
ztest_dsl_prop_set_uint64(char *osname, zfs_prop_t prop, uint64_t value,
boolean_t inherit)
{
const char *propname = zfs_prop_to_name(prop);
const char *valname;
char *setpoint;
uint64_t curval;
int error;
error = dsl_prop_set_int(osname, propname,
(inherit ? ZPROP_SRC_NONE : ZPROP_SRC_LOCAL), value);
if (error == ENOSPC) {
ztest_record_enospc(FTAG);
return (error);
}
ASSERT0(error);
setpoint = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
VERIFY0(dsl_prop_get_integer(osname, propname, &curval, setpoint));
if (ztest_opts.zo_verbose >= 6) {
int err;
err = zfs_prop_index_to_string(prop, curval, &valname);
if (err)
(void) printf("%s %s = %llu at '%s'\n", osname,
propname, (unsigned long long)curval, setpoint);
else
(void) printf("%s %s = %s at '%s'\n",
osname, propname, valname, setpoint);
}
umem_free(setpoint, MAXPATHLEN);
return (error);
}
static int
ztest_spa_prop_set_uint64(zpool_prop_t prop, uint64_t value)
{
spa_t *spa = ztest_spa;
nvlist_t *props = NULL;
int error;
props = fnvlist_alloc();
fnvlist_add_uint64(props, zpool_prop_to_name(prop), value);
error = spa_prop_set(spa, props);
fnvlist_free(props);
if (error == ENOSPC) {
ztest_record_enospc(FTAG);
return (error);
}
ASSERT0(error);
return (error);
}
static int
ztest_dmu_objset_own(const char *name, dmu_objset_type_t type,
boolean_t readonly, boolean_t decrypt, const void *tag, objset_t **osp)
{
int err;
char *cp = NULL;
char ddname[ZFS_MAX_DATASET_NAME_LEN];
strlcpy(ddname, name, sizeof (ddname));
cp = strchr(ddname, '@');
if (cp != NULL)
*cp = '\0';
err = dmu_objset_own(name, type, readonly, decrypt, tag, osp);
while (decrypt && err == EACCES) {
dsl_crypto_params_t *dcp;
nvlist_t *crypto_args = fnvlist_alloc();
fnvlist_add_uint8_array(crypto_args, "wkeydata",
(uint8_t *)ztest_wkeydata, WRAPPING_KEY_LEN);
VERIFY0(dsl_crypto_params_create_nvlist(DCP_CMD_NONE, NULL,
crypto_args, &dcp));
err = spa_keystore_load_wkey(ddname, dcp, B_FALSE);
/*
* Note: if there was an error loading, the wkey was not
* consumed, and needs to be freed.
*/
dsl_crypto_params_free(dcp, (err != 0));
fnvlist_free(crypto_args);
if (err == EINVAL) {
/*
* We couldn't load a key for this dataset so try
* the parent. This loop will eventually hit the
* encryption root since ztest only makes clones
* as children of their origin datasets.
*/
cp = strrchr(ddname, '/');
if (cp == NULL)
return (err);
*cp = '\0';
err = EACCES;
continue;
} else if (err != 0) {
break;
}
err = dmu_objset_own(name, type, readonly, decrypt, tag, osp);
break;
}
return (err);
}
static void
ztest_rll_init(rll_t *rll)
{
rll->rll_writer = NULL;
rll->rll_readers = 0;
mutex_init(&rll->rll_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&rll->rll_cv, NULL, CV_DEFAULT, NULL);
}
static void
ztest_rll_destroy(rll_t *rll)
{
ASSERT3P(rll->rll_writer, ==, NULL);
ASSERT0(rll->rll_readers);
mutex_destroy(&rll->rll_lock);
cv_destroy(&rll->rll_cv);
}
static void
ztest_rll_lock(rll_t *rll, rl_type_t type)
{
mutex_enter(&rll->rll_lock);
- if (type == RL_READER) {
+ if (type == ZTRL_READER) {
while (rll->rll_writer != NULL)
(void) cv_wait(&rll->rll_cv, &rll->rll_lock);
rll->rll_readers++;
} else {
while (rll->rll_writer != NULL || rll->rll_readers)
(void) cv_wait(&rll->rll_cv, &rll->rll_lock);
rll->rll_writer = curthread;
}
mutex_exit(&rll->rll_lock);
}
static void
ztest_rll_unlock(rll_t *rll)
{
mutex_enter(&rll->rll_lock);
if (rll->rll_writer) {
ASSERT0(rll->rll_readers);
rll->rll_writer = NULL;
} else {
ASSERT3S(rll->rll_readers, >, 0);
ASSERT3P(rll->rll_writer, ==, NULL);
rll->rll_readers--;
}
if (rll->rll_writer == NULL && rll->rll_readers == 0)
cv_broadcast(&rll->rll_cv);
mutex_exit(&rll->rll_lock);
}
static void
ztest_object_lock(ztest_ds_t *zd, uint64_t object, rl_type_t type)
{
rll_t *rll = &zd->zd_object_lock[object & (ZTEST_OBJECT_LOCKS - 1)];
ztest_rll_lock(rll, type);
}
static void
ztest_object_unlock(ztest_ds_t *zd, uint64_t object)
{
rll_t *rll = &zd->zd_object_lock[object & (ZTEST_OBJECT_LOCKS - 1)];
ztest_rll_unlock(rll);
}
static rl_t *
ztest_range_lock(ztest_ds_t *zd, uint64_t object, uint64_t offset,
uint64_t size, rl_type_t type)
{
uint64_t hash = object ^ (offset % (ZTEST_RANGE_LOCKS + 1));
rll_t *rll = &zd->zd_range_lock[hash & (ZTEST_RANGE_LOCKS - 1)];
rl_t *rl;
rl = umem_alloc(sizeof (*rl), UMEM_NOFAIL);
rl->rl_object = object;
rl->rl_offset = offset;
rl->rl_size = size;
rl->rl_lock = rll;
ztest_rll_lock(rll, type);
return (rl);
}
static void
ztest_range_unlock(rl_t *rl)
{
rll_t *rll = rl->rl_lock;
ztest_rll_unlock(rll);
umem_free(rl, sizeof (*rl));
}
static void
ztest_zd_init(ztest_ds_t *zd, ztest_shared_ds_t *szd, objset_t *os)
{
zd->zd_os = os;
zd->zd_zilog = dmu_objset_zil(os);
zd->zd_shared = szd;
dmu_objset_name(os, zd->zd_name);
int l;
if (zd->zd_shared != NULL)
zd->zd_shared->zd_seq = 0;
VERIFY0(pthread_rwlock_init(&zd->zd_zilog_lock, NULL));
mutex_init(&zd->zd_dirobj_lock, NULL, MUTEX_DEFAULT, NULL);
for (l = 0; l < ZTEST_OBJECT_LOCKS; l++)
ztest_rll_init(&zd->zd_object_lock[l]);
for (l = 0; l < ZTEST_RANGE_LOCKS; l++)
ztest_rll_init(&zd->zd_range_lock[l]);
}
static void
ztest_zd_fini(ztest_ds_t *zd)
{
int l;
mutex_destroy(&zd->zd_dirobj_lock);
(void) pthread_rwlock_destroy(&zd->zd_zilog_lock);
for (l = 0; l < ZTEST_OBJECT_LOCKS; l++)
ztest_rll_destroy(&zd->zd_object_lock[l]);
for (l = 0; l < ZTEST_RANGE_LOCKS; l++)
ztest_rll_destroy(&zd->zd_range_lock[l]);
}
#define TXG_MIGHTWAIT (ztest_random(10) == 0 ? TXG_NOWAIT : TXG_WAIT)
static uint64_t
ztest_tx_assign(dmu_tx_t *tx, uint64_t txg_how, const char *tag)
{
uint64_t txg;
int error;
/*
* Attempt to assign tx to some transaction group.
*/
error = dmu_tx_assign(tx, txg_how);
if (error) {
if (error == ERESTART) {
ASSERT3U(txg_how, ==, TXG_NOWAIT);
dmu_tx_wait(tx);
} else {
ASSERT3U(error, ==, ENOSPC);
ztest_record_enospc(tag);
}
dmu_tx_abort(tx);
return (0);
}
txg = dmu_tx_get_txg(tx);
ASSERT3U(txg, !=, 0);
return (txg);
}
static void
ztest_bt_generate(ztest_block_tag_t *bt, objset_t *os, uint64_t object,
uint64_t dnodesize, uint64_t offset, uint64_t gen, uint64_t txg,
uint64_t crtxg)
{
bt->bt_magic = BT_MAGIC;
bt->bt_objset = dmu_objset_id(os);
bt->bt_object = object;
bt->bt_dnodesize = dnodesize;
bt->bt_offset = offset;
bt->bt_gen = gen;
bt->bt_txg = txg;
bt->bt_crtxg = crtxg;
}
static void
ztest_bt_verify(ztest_block_tag_t *bt, objset_t *os, uint64_t object,
uint64_t dnodesize, uint64_t offset, uint64_t gen, uint64_t txg,
uint64_t crtxg)
{
ASSERT3U(bt->bt_magic, ==, BT_MAGIC);
ASSERT3U(bt->bt_objset, ==, dmu_objset_id(os));
ASSERT3U(bt->bt_object, ==, object);
ASSERT3U(bt->bt_dnodesize, ==, dnodesize);
ASSERT3U(bt->bt_offset, ==, offset);
ASSERT3U(bt->bt_gen, <=, gen);
ASSERT3U(bt->bt_txg, <=, txg);
ASSERT3U(bt->bt_crtxg, ==, crtxg);
}
static ztest_block_tag_t *
ztest_bt_bonus(dmu_buf_t *db)
{
dmu_object_info_t doi;
ztest_block_tag_t *bt;
dmu_object_info_from_db(db, &doi);
ASSERT3U(doi.doi_bonus_size, <=, db->db_size);
ASSERT3U(doi.doi_bonus_size, >=, sizeof (*bt));
bt = (void *)((char *)db->db_data + doi.doi_bonus_size - sizeof (*bt));
return (bt);
}
/*
* Generate a token to fill up unused bonus buffer space. Try to make
* it unique to the object, generation, and offset to verify that data
* is not getting overwritten by data from other dnodes.
*/
#define ZTEST_BONUS_FILL_TOKEN(obj, ds, gen, offset) \
(((ds) << 48) | ((gen) << 32) | ((obj) << 8) | (offset))
/*
* Fill up the unused bonus buffer region before the block tag with a
* verifiable pattern. Filling the whole bonus area with non-zero data
* helps ensure that all dnode traversal code properly skips the
* interior regions of large dnodes.
*/
static void
ztest_fill_unused_bonus(dmu_buf_t *db, void *end, uint64_t obj,
objset_t *os, uint64_t gen)
{
uint64_t *bonusp;
ASSERT(IS_P2ALIGNED((char *)end - (char *)db->db_data, 8));
for (bonusp = db->db_data; bonusp < (uint64_t *)end; bonusp++) {
uint64_t token = ZTEST_BONUS_FILL_TOKEN(obj, dmu_objset_id(os),
gen, bonusp - (uint64_t *)db->db_data);
*bonusp = token;
}
}
/*
* Verify that the unused area of a bonus buffer is filled with the
* expected tokens.
*/
static void
ztest_verify_unused_bonus(dmu_buf_t *db, void *end, uint64_t obj,
objset_t *os, uint64_t gen)
{
uint64_t *bonusp;
for (bonusp = db->db_data; bonusp < (uint64_t *)end; bonusp++) {
uint64_t token = ZTEST_BONUS_FILL_TOKEN(obj, dmu_objset_id(os),
gen, bonusp - (uint64_t *)db->db_data);
VERIFY3U(*bonusp, ==, token);
}
}
/*
* ZIL logging ops
*/
#define lrz_type lr_mode
#define lrz_blocksize lr_uid
#define lrz_ibshift lr_gid
#define lrz_bonustype lr_rdev
#define lrz_dnodesize lr_crtime[1]
static void
ztest_log_create(ztest_ds_t *zd, dmu_tx_t *tx, lr_create_t *lr)
{
char *name = (void *)(lr + 1); /* name follows lr */
size_t namesize = strlen(name) + 1;
itx_t *itx;
if (zil_replaying(zd->zd_zilog, tx))
return;
itx = zil_itx_create(TX_CREATE, sizeof (*lr) + namesize);
memcpy(&itx->itx_lr + 1, &lr->lr_common + 1,
sizeof (*lr) + namesize - sizeof (lr_t));
zil_itx_assign(zd->zd_zilog, itx, tx);
}
static void
ztest_log_remove(ztest_ds_t *zd, dmu_tx_t *tx, lr_remove_t *lr, uint64_t object)
{
char *name = (void *)(lr + 1); /* name follows lr */
size_t namesize = strlen(name) + 1;
itx_t *itx;
if (zil_replaying(zd->zd_zilog, tx))
return;
itx = zil_itx_create(TX_REMOVE, sizeof (*lr) + namesize);
memcpy(&itx->itx_lr + 1, &lr->lr_common + 1,
sizeof (*lr) + namesize - sizeof (lr_t));
itx->itx_oid = object;
zil_itx_assign(zd->zd_zilog, itx, tx);
}
static void
ztest_log_write(ztest_ds_t *zd, dmu_tx_t *tx, lr_write_t *lr)
{
itx_t *itx;
itx_wr_state_t write_state = ztest_random(WR_NUM_STATES);
if (zil_replaying(zd->zd_zilog, tx))
return;
if (lr->lr_length > zil_max_log_data(zd->zd_zilog, sizeof (lr_write_t)))
write_state = WR_INDIRECT;
itx = zil_itx_create(TX_WRITE,
sizeof (*lr) + (write_state == WR_COPIED ? lr->lr_length : 0));
if (write_state == WR_COPIED &&
dmu_read(zd->zd_os, lr->lr_foid, lr->lr_offset, lr->lr_length,
((lr_write_t *)&itx->itx_lr) + 1, DMU_READ_NO_PREFETCH) != 0) {
zil_itx_destroy(itx);
itx = zil_itx_create(TX_WRITE, sizeof (*lr));
write_state = WR_NEED_COPY;
}
itx->itx_private = zd;
itx->itx_wr_state = write_state;
itx->itx_sync = (ztest_random(8) == 0);
memcpy(&itx->itx_lr + 1, &lr->lr_common + 1,
sizeof (*lr) - sizeof (lr_t));
zil_itx_assign(zd->zd_zilog, itx, tx);
}
static void
ztest_log_truncate(ztest_ds_t *zd, dmu_tx_t *tx, lr_truncate_t *lr)
{
itx_t *itx;
if (zil_replaying(zd->zd_zilog, tx))
return;
itx = zil_itx_create(TX_TRUNCATE, sizeof (*lr));
memcpy(&itx->itx_lr + 1, &lr->lr_common + 1,
sizeof (*lr) - sizeof (lr_t));
itx->itx_sync = B_FALSE;
zil_itx_assign(zd->zd_zilog, itx, tx);
}
static void
ztest_log_setattr(ztest_ds_t *zd, dmu_tx_t *tx, lr_setattr_t *lr)
{
itx_t *itx;
if (zil_replaying(zd->zd_zilog, tx))
return;
itx = zil_itx_create(TX_SETATTR, sizeof (*lr));
memcpy(&itx->itx_lr + 1, &lr->lr_common + 1,
sizeof (*lr) - sizeof (lr_t));
itx->itx_sync = B_FALSE;
zil_itx_assign(zd->zd_zilog, itx, tx);
}
/*
* ZIL replay ops
*/
static int
ztest_replay_create(void *arg1, void *arg2, boolean_t byteswap)
{
ztest_ds_t *zd = arg1;
lr_create_t *lr = arg2;
char *name = (void *)(lr + 1); /* name follows lr */
objset_t *os = zd->zd_os;
ztest_block_tag_t *bbt;
dmu_buf_t *db;
dmu_tx_t *tx;
uint64_t txg;
int error = 0;
int bonuslen;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
ASSERT3U(lr->lr_doid, ==, ZTEST_DIROBJ);
ASSERT3S(name[0], !=, '\0');
tx = dmu_tx_create(os);
dmu_tx_hold_zap(tx, lr->lr_doid, B_TRUE, name);
if (lr->lrz_type == DMU_OT_ZAP_OTHER) {
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, B_TRUE, NULL);
} else {
dmu_tx_hold_bonus(tx, DMU_NEW_OBJECT);
}
txg = ztest_tx_assign(tx, TXG_WAIT, FTAG);
if (txg == 0)
return (ENOSPC);
ASSERT3U(dmu_objset_zil(os)->zl_replay, ==, !!lr->lr_foid);
bonuslen = DN_BONUS_SIZE(lr->lrz_dnodesize);
if (lr->lrz_type == DMU_OT_ZAP_OTHER) {
if (lr->lr_foid == 0) {
lr->lr_foid = zap_create_dnsize(os,
lr->lrz_type, lr->lrz_bonustype,
bonuslen, lr->lrz_dnodesize, tx);
} else {
error = zap_create_claim_dnsize(os, lr->lr_foid,
lr->lrz_type, lr->lrz_bonustype,
bonuslen, lr->lrz_dnodesize, tx);
}
} else {
if (lr->lr_foid == 0) {
lr->lr_foid = dmu_object_alloc_dnsize(os,
lr->lrz_type, 0, lr->lrz_bonustype,
bonuslen, lr->lrz_dnodesize, tx);
} else {
error = dmu_object_claim_dnsize(os, lr->lr_foid,
lr->lrz_type, 0, lr->lrz_bonustype,
bonuslen, lr->lrz_dnodesize, tx);
}
}
if (error) {
ASSERT3U(error, ==, EEXIST);
ASSERT(zd->zd_zilog->zl_replay);
dmu_tx_commit(tx);
return (error);
}
ASSERT3U(lr->lr_foid, !=, 0);
if (lr->lrz_type != DMU_OT_ZAP_OTHER)
VERIFY0(dmu_object_set_blocksize(os, lr->lr_foid,
lr->lrz_blocksize, lr->lrz_ibshift, tx));
VERIFY0(dmu_bonus_hold(os, lr->lr_foid, FTAG, &db));
bbt = ztest_bt_bonus(db);
dmu_buf_will_dirty(db, tx);
ztest_bt_generate(bbt, os, lr->lr_foid, lr->lrz_dnodesize, -1ULL,
lr->lr_gen, txg, txg);
ztest_fill_unused_bonus(db, bbt, lr->lr_foid, os, lr->lr_gen);
dmu_buf_rele(db, FTAG);
VERIFY0(zap_add(os, lr->lr_doid, name, sizeof (uint64_t), 1,
&lr->lr_foid, tx));
(void) ztest_log_create(zd, tx, lr);
dmu_tx_commit(tx);
return (0);
}
static int
ztest_replay_remove(void *arg1, void *arg2, boolean_t byteswap)
{
ztest_ds_t *zd = arg1;
lr_remove_t *lr = arg2;
char *name = (void *)(lr + 1); /* name follows lr */
objset_t *os = zd->zd_os;
dmu_object_info_t doi;
dmu_tx_t *tx;
uint64_t object, txg;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
ASSERT3U(lr->lr_doid, ==, ZTEST_DIROBJ);
ASSERT3S(name[0], !=, '\0');
VERIFY0(
zap_lookup(os, lr->lr_doid, name, sizeof (object), 1, &object));
ASSERT3U(object, !=, 0);
- ztest_object_lock(zd, object, RL_WRITER);
+ ztest_object_lock(zd, object, ZTRL_WRITER);
VERIFY0(dmu_object_info(os, object, &doi));
tx = dmu_tx_create(os);
dmu_tx_hold_zap(tx, lr->lr_doid, B_FALSE, name);
dmu_tx_hold_free(tx, object, 0, DMU_OBJECT_END);
txg = ztest_tx_assign(tx, TXG_WAIT, FTAG);
if (txg == 0) {
ztest_object_unlock(zd, object);
return (ENOSPC);
}
if (doi.doi_type == DMU_OT_ZAP_OTHER) {
VERIFY0(zap_destroy(os, object, tx));
} else {
VERIFY0(dmu_object_free(os, object, tx));
}
VERIFY0(zap_remove(os, lr->lr_doid, name, tx));
(void) ztest_log_remove(zd, tx, lr, object);
dmu_tx_commit(tx);
ztest_object_unlock(zd, object);
return (0);
}
static int
ztest_replay_write(void *arg1, void *arg2, boolean_t byteswap)
{
ztest_ds_t *zd = arg1;
lr_write_t *lr = arg2;
objset_t *os = zd->zd_os;
void *data = lr + 1; /* data follows lr */
uint64_t offset, length;
ztest_block_tag_t *bt = data;
ztest_block_tag_t *bbt;
uint64_t gen, txg, lrtxg, crtxg;
dmu_object_info_t doi;
dmu_tx_t *tx;
dmu_buf_t *db;
arc_buf_t *abuf = NULL;
rl_t *rl;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
offset = lr->lr_offset;
length = lr->lr_length;
/* If it's a dmu_sync() block, write the whole block */
if (lr->lr_common.lrc_reclen == sizeof (lr_write_t)) {
uint64_t blocksize = BP_GET_LSIZE(&lr->lr_blkptr);
if (length < blocksize) {
offset -= offset % blocksize;
length = blocksize;
}
}
if (bt->bt_magic == BSWAP_64(BT_MAGIC))
byteswap_uint64_array(bt, sizeof (*bt));
if (bt->bt_magic != BT_MAGIC)
bt = NULL;
- ztest_object_lock(zd, lr->lr_foid, RL_READER);
- rl = ztest_range_lock(zd, lr->lr_foid, offset, length, RL_WRITER);
+ ztest_object_lock(zd, lr->lr_foid, ZTRL_READER);
+ rl = ztest_range_lock(zd, lr->lr_foid, offset, length, ZTRL_WRITER);
VERIFY0(dmu_bonus_hold(os, lr->lr_foid, FTAG, &db));
dmu_object_info_from_db(db, &doi);
bbt = ztest_bt_bonus(db);
ASSERT3U(bbt->bt_magic, ==, BT_MAGIC);
gen = bbt->bt_gen;
crtxg = bbt->bt_crtxg;
lrtxg = lr->lr_common.lrc_txg;
tx = dmu_tx_create(os);
dmu_tx_hold_write(tx, lr->lr_foid, offset, length);
if (ztest_random(8) == 0 && length == doi.doi_data_block_size &&
P2PHASE(offset, length) == 0)
abuf = dmu_request_arcbuf(db, length);
txg = ztest_tx_assign(tx, TXG_WAIT, FTAG);
if (txg == 0) {
if (abuf != NULL)
dmu_return_arcbuf(abuf);
dmu_buf_rele(db, FTAG);
ztest_range_unlock(rl);
ztest_object_unlock(zd, lr->lr_foid);
return (ENOSPC);
}
if (bt != NULL) {
/*
* Usually, verify the old data before writing new data --
* but not always, because we also want to verify correct
* behavior when the data was not recently read into cache.
*/
ASSERT(doi.doi_data_block_size);
ASSERT0(offset % doi.doi_data_block_size);
if (ztest_random(4) != 0) {
int prefetch = ztest_random(2) ?
DMU_READ_PREFETCH : DMU_READ_NO_PREFETCH;
ztest_block_tag_t rbt;
VERIFY(dmu_read(os, lr->lr_foid, offset,
sizeof (rbt), &rbt, prefetch) == 0);
if (rbt.bt_magic == BT_MAGIC) {
ztest_bt_verify(&rbt, os, lr->lr_foid, 0,
offset, gen, txg, crtxg);
}
}
/*
* Writes can appear to be newer than the bonus buffer because
* the ztest_get_data() callback does a dmu_read() of the
* open-context data, which may be different than the data
* as it was when the write was generated.
*/
if (zd->zd_zilog->zl_replay) {
ztest_bt_verify(bt, os, lr->lr_foid, 0, offset,
MAX(gen, bt->bt_gen), MAX(txg, lrtxg),
bt->bt_crtxg);
}
/*
* Set the bt's gen/txg to the bonus buffer's gen/txg
* so that all of the usual ASSERTs will work.
*/
ztest_bt_generate(bt, os, lr->lr_foid, 0, offset, gen, txg,
crtxg);
}
if (abuf == NULL) {
dmu_write(os, lr->lr_foid, offset, length, data, tx);
} else {
memcpy(abuf->b_data, data, length);
VERIFY0(dmu_assign_arcbuf_by_dbuf(db, offset, abuf, tx));
}
(void) ztest_log_write(zd, tx, lr);
dmu_buf_rele(db, FTAG);
dmu_tx_commit(tx);
ztest_range_unlock(rl);
ztest_object_unlock(zd, lr->lr_foid);
return (0);
}
static int
ztest_replay_truncate(void *arg1, void *arg2, boolean_t byteswap)
{
ztest_ds_t *zd = arg1;
lr_truncate_t *lr = arg2;
objset_t *os = zd->zd_os;
dmu_tx_t *tx;
uint64_t txg;
rl_t *rl;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
- ztest_object_lock(zd, lr->lr_foid, RL_READER);
+ ztest_object_lock(zd, lr->lr_foid, ZTRL_READER);
rl = ztest_range_lock(zd, lr->lr_foid, lr->lr_offset, lr->lr_length,
- RL_WRITER);
+ ZTRL_WRITER);
tx = dmu_tx_create(os);
dmu_tx_hold_free(tx, lr->lr_foid, lr->lr_offset, lr->lr_length);
txg = ztest_tx_assign(tx, TXG_WAIT, FTAG);
if (txg == 0) {
ztest_range_unlock(rl);
ztest_object_unlock(zd, lr->lr_foid);
return (ENOSPC);
}
VERIFY0(dmu_free_range(os, lr->lr_foid, lr->lr_offset,
lr->lr_length, tx));
(void) ztest_log_truncate(zd, tx, lr);
dmu_tx_commit(tx);
ztest_range_unlock(rl);
ztest_object_unlock(zd, lr->lr_foid);
return (0);
}
static int
ztest_replay_setattr(void *arg1, void *arg2, boolean_t byteswap)
{
ztest_ds_t *zd = arg1;
lr_setattr_t *lr = arg2;
objset_t *os = zd->zd_os;
dmu_tx_t *tx;
dmu_buf_t *db;
ztest_block_tag_t *bbt;
uint64_t txg, lrtxg, crtxg, dnodesize;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
- ztest_object_lock(zd, lr->lr_foid, RL_WRITER);
+ ztest_object_lock(zd, lr->lr_foid, ZTRL_WRITER);
VERIFY0(dmu_bonus_hold(os, lr->lr_foid, FTAG, &db));
tx = dmu_tx_create(os);
dmu_tx_hold_bonus(tx, lr->lr_foid);
txg = ztest_tx_assign(tx, TXG_WAIT, FTAG);
if (txg == 0) {
dmu_buf_rele(db, FTAG);
ztest_object_unlock(zd, lr->lr_foid);
return (ENOSPC);
}
bbt = ztest_bt_bonus(db);
ASSERT3U(bbt->bt_magic, ==, BT_MAGIC);
crtxg = bbt->bt_crtxg;
lrtxg = lr->lr_common.lrc_txg;
dnodesize = bbt->bt_dnodesize;
if (zd->zd_zilog->zl_replay) {
ASSERT3U(lr->lr_size, !=, 0);
ASSERT3U(lr->lr_mode, !=, 0);
ASSERT3U(lrtxg, !=, 0);
} else {
/*
* Randomly change the size and increment the generation.
*/
lr->lr_size = (ztest_random(db->db_size / sizeof (*bbt)) + 1) *
sizeof (*bbt);
lr->lr_mode = bbt->bt_gen + 1;
ASSERT0(lrtxg);
}
/*
* Verify that the current bonus buffer is not newer than our txg.
*/
ztest_bt_verify(bbt, os, lr->lr_foid, dnodesize, -1ULL, lr->lr_mode,
MAX(txg, lrtxg), crtxg);
dmu_buf_will_dirty(db, tx);
ASSERT3U(lr->lr_size, >=, sizeof (*bbt));
ASSERT3U(lr->lr_size, <=, db->db_size);
VERIFY0(dmu_set_bonus(db, lr->lr_size, tx));
bbt = ztest_bt_bonus(db);
ztest_bt_generate(bbt, os, lr->lr_foid, dnodesize, -1ULL, lr->lr_mode,
txg, crtxg);
ztest_fill_unused_bonus(db, bbt, lr->lr_foid, os, bbt->bt_gen);
dmu_buf_rele(db, FTAG);
(void) ztest_log_setattr(zd, tx, lr);
dmu_tx_commit(tx);
ztest_object_unlock(zd, lr->lr_foid);
return (0);
}
static zil_replay_func_t *ztest_replay_vector[TX_MAX_TYPE] = {
NULL, /* 0 no such transaction type */
ztest_replay_create, /* TX_CREATE */
NULL, /* TX_MKDIR */
NULL, /* TX_MKXATTR */
NULL, /* TX_SYMLINK */
ztest_replay_remove, /* TX_REMOVE */
NULL, /* TX_RMDIR */
NULL, /* TX_LINK */
NULL, /* TX_RENAME */
ztest_replay_write, /* TX_WRITE */
ztest_replay_truncate, /* TX_TRUNCATE */
ztest_replay_setattr, /* TX_SETATTR */
NULL, /* TX_ACL */
NULL, /* TX_CREATE_ACL */
NULL, /* TX_CREATE_ATTR */
NULL, /* TX_CREATE_ACL_ATTR */
NULL, /* TX_MKDIR_ACL */
NULL, /* TX_MKDIR_ATTR */
NULL, /* TX_MKDIR_ACL_ATTR */
NULL, /* TX_WRITE2 */
NULL, /* TX_SETSAXATTR */
NULL, /* TX_RENAME_EXCHANGE */
NULL, /* TX_RENAME_WHITEOUT */
};
/*
* ZIL get_data callbacks
*/
static void
ztest_get_done(zgd_t *zgd, int error)
{
(void) error;
ztest_ds_t *zd = zgd->zgd_private;
uint64_t object = ((rl_t *)zgd->zgd_lr)->rl_object;
if (zgd->zgd_db)
dmu_buf_rele(zgd->zgd_db, zgd);
ztest_range_unlock((rl_t *)zgd->zgd_lr);
ztest_object_unlock(zd, object);
umem_free(zgd, sizeof (*zgd));
}
static int
ztest_get_data(void *arg, uint64_t arg2, lr_write_t *lr, char *buf,
struct lwb *lwb, zio_t *zio)
{
(void) arg2;
ztest_ds_t *zd = arg;
objset_t *os = zd->zd_os;
uint64_t object = lr->lr_foid;
uint64_t offset = lr->lr_offset;
uint64_t size = lr->lr_length;
uint64_t txg = lr->lr_common.lrc_txg;
uint64_t crtxg;
dmu_object_info_t doi;
dmu_buf_t *db;
zgd_t *zgd;
int error;
ASSERT3P(lwb, !=, NULL);
ASSERT3U(size, !=, 0);
- ztest_object_lock(zd, object, RL_READER);
+ ztest_object_lock(zd, object, ZTRL_READER);
error = dmu_bonus_hold(os, object, FTAG, &db);
if (error) {
ztest_object_unlock(zd, object);
return (error);
}
crtxg = ztest_bt_bonus(db)->bt_crtxg;
if (crtxg == 0 || crtxg > txg) {
dmu_buf_rele(db, FTAG);
ztest_object_unlock(zd, object);
return (ENOENT);
}
dmu_object_info_from_db(db, &doi);
dmu_buf_rele(db, FTAG);
db = NULL;
zgd = umem_zalloc(sizeof (*zgd), UMEM_NOFAIL);
zgd->zgd_lwb = lwb;
zgd->zgd_private = zd;
if (buf != NULL) { /* immediate write */
zgd->zgd_lr = (struct zfs_locked_range *)ztest_range_lock(zd,
- object, offset, size, RL_READER);
+ object, offset, size, ZTRL_READER);
error = dmu_read(os, object, offset, size, buf,
DMU_READ_NO_PREFETCH);
ASSERT0(error);
} else {
ASSERT3P(zio, !=, NULL);
size = doi.doi_data_block_size;
if (ISP2(size)) {
offset = P2ALIGN(offset, size);
} else {
ASSERT3U(offset, <, size);
offset = 0;
}
zgd->zgd_lr = (struct zfs_locked_range *)ztest_range_lock(zd,
- object, offset, size, RL_READER);
+ object, offset, size, ZTRL_READER);
error = dmu_buf_hold_noread(os, object, offset, zgd, &db);
if (error == 0) {
blkptr_t *bp = &lr->lr_blkptr;
zgd->zgd_db = db;
zgd->zgd_bp = bp;
ASSERT3U(db->db_offset, ==, offset);
ASSERT3U(db->db_size, ==, size);
error = dmu_sync(zio, lr->lr_common.lrc_txg,
ztest_get_done, zgd);
if (error == 0)
return (0);
}
}
ztest_get_done(zgd, error);
return (error);
}
static void *
ztest_lr_alloc(size_t lrsize, char *name)
{
char *lr;
size_t namesize = name ? strlen(name) + 1 : 0;
lr = umem_zalloc(lrsize + namesize, UMEM_NOFAIL);
if (name)
memcpy(lr + lrsize, name, namesize);
return (lr);
}
static void
ztest_lr_free(void *lr, size_t lrsize, char *name)
{
size_t namesize = name ? strlen(name) + 1 : 0;
umem_free(lr, lrsize + namesize);
}
/*
* Lookup a bunch of objects. Returns the number of objects not found.
*/
static int
ztest_lookup(ztest_ds_t *zd, ztest_od_t *od, int count)
{
int missing = 0;
int error;
int i;
ASSERT(MUTEX_HELD(&zd->zd_dirobj_lock));
for (i = 0; i < count; i++, od++) {
od->od_object = 0;
error = zap_lookup(zd->zd_os, od->od_dir, od->od_name,
sizeof (uint64_t), 1, &od->od_object);
if (error) {
ASSERT3S(error, ==, ENOENT);
ASSERT0(od->od_object);
missing++;
} else {
dmu_buf_t *db;
ztest_block_tag_t *bbt;
dmu_object_info_t doi;
ASSERT3U(od->od_object, !=, 0);
ASSERT0(missing); /* there should be no gaps */
- ztest_object_lock(zd, od->od_object, RL_READER);
+ ztest_object_lock(zd, od->od_object, ZTRL_READER);
VERIFY0(dmu_bonus_hold(zd->zd_os, od->od_object,
FTAG, &db));
dmu_object_info_from_db(db, &doi);
bbt = ztest_bt_bonus(db);
ASSERT3U(bbt->bt_magic, ==, BT_MAGIC);
od->od_type = doi.doi_type;
od->od_blocksize = doi.doi_data_block_size;
od->od_gen = bbt->bt_gen;
dmu_buf_rele(db, FTAG);
ztest_object_unlock(zd, od->od_object);
}
}
return (missing);
}
static int
ztest_create(ztest_ds_t *zd, ztest_od_t *od, int count)
{
int missing = 0;
int i;
ASSERT(MUTEX_HELD(&zd->zd_dirobj_lock));
for (i = 0; i < count; i++, od++) {
if (missing) {
od->od_object = 0;
missing++;
continue;
}
lr_create_t *lr = ztest_lr_alloc(sizeof (*lr), od->od_name);
lr->lr_doid = od->od_dir;
lr->lr_foid = 0; /* 0 to allocate, > 0 to claim */
lr->lrz_type = od->od_crtype;
lr->lrz_blocksize = od->od_crblocksize;
lr->lrz_ibshift = ztest_random_ibshift();
lr->lrz_bonustype = DMU_OT_UINT64_OTHER;
lr->lrz_dnodesize = od->od_crdnodesize;
lr->lr_gen = od->od_crgen;
lr->lr_crtime[0] = time(NULL);
if (ztest_replay_create(zd, lr, B_FALSE) != 0) {
ASSERT0(missing);
od->od_object = 0;
missing++;
} else {
od->od_object = lr->lr_foid;
od->od_type = od->od_crtype;
od->od_blocksize = od->od_crblocksize;
od->od_gen = od->od_crgen;
ASSERT3U(od->od_object, !=, 0);
}
ztest_lr_free(lr, sizeof (*lr), od->od_name);
}
return (missing);
}
static int
ztest_remove(ztest_ds_t *zd, ztest_od_t *od, int count)
{
int missing = 0;
int error;
int i;
ASSERT(MUTEX_HELD(&zd->zd_dirobj_lock));
od += count - 1;
for (i = count - 1; i >= 0; i--, od--) {
if (missing) {
missing++;
continue;
}
/*
* No object was found.
*/
if (od->od_object == 0)
continue;
lr_remove_t *lr = ztest_lr_alloc(sizeof (*lr), od->od_name);
lr->lr_doid = od->od_dir;
if ((error = ztest_replay_remove(zd, lr, B_FALSE)) != 0) {
ASSERT3U(error, ==, ENOSPC);
missing++;
} else {
od->od_object = 0;
}
ztest_lr_free(lr, sizeof (*lr), od->od_name);
}
return (missing);
}
static int
ztest_write(ztest_ds_t *zd, uint64_t object, uint64_t offset, uint64_t size,
- void *data)
+ const void *data)
{
lr_write_t *lr;
int error;
lr = ztest_lr_alloc(sizeof (*lr) + size, NULL);
lr->lr_foid = object;
lr->lr_offset = offset;
lr->lr_length = size;
lr->lr_blkoff = 0;
BP_ZERO(&lr->lr_blkptr);
memcpy(lr + 1, data, size);
error = ztest_replay_write(zd, lr, B_FALSE);
ztest_lr_free(lr, sizeof (*lr) + size, NULL);
return (error);
}
static int
ztest_truncate(ztest_ds_t *zd, uint64_t object, uint64_t offset, uint64_t size)
{
lr_truncate_t *lr;
int error;
lr = ztest_lr_alloc(sizeof (*lr), NULL);
lr->lr_foid = object;
lr->lr_offset = offset;
lr->lr_length = size;
error = ztest_replay_truncate(zd, lr, B_FALSE);
ztest_lr_free(lr, sizeof (*lr), NULL);
return (error);
}
static int
ztest_setattr(ztest_ds_t *zd, uint64_t object)
{
lr_setattr_t *lr;
int error;
lr = ztest_lr_alloc(sizeof (*lr), NULL);
lr->lr_foid = object;
lr->lr_size = 0;
lr->lr_mode = 0;
error = ztest_replay_setattr(zd, lr, B_FALSE);
ztest_lr_free(lr, sizeof (*lr), NULL);
return (error);
}
static void
ztest_prealloc(ztest_ds_t *zd, uint64_t object, uint64_t offset, uint64_t size)
{
objset_t *os = zd->zd_os;
dmu_tx_t *tx;
uint64_t txg;
rl_t *rl;
txg_wait_synced(dmu_objset_pool(os), 0);
- ztest_object_lock(zd, object, RL_READER);
- rl = ztest_range_lock(zd, object, offset, size, RL_WRITER);
+ ztest_object_lock(zd, object, ZTRL_READER);
+ rl = ztest_range_lock(zd, object, offset, size, ZTRL_WRITER);
tx = dmu_tx_create(os);
dmu_tx_hold_write(tx, object, offset, size);
txg = ztest_tx_assign(tx, TXG_WAIT, FTAG);
if (txg != 0) {
dmu_prealloc(os, object, offset, size, tx);
dmu_tx_commit(tx);
txg_wait_synced(dmu_objset_pool(os), txg);
} else {
(void) dmu_free_long_range(os, object, offset, size);
}
ztest_range_unlock(rl);
ztest_object_unlock(zd, object);
}
static void
ztest_io(ztest_ds_t *zd, uint64_t object, uint64_t offset)
{
int err;
ztest_block_tag_t wbt;
dmu_object_info_t doi;
enum ztest_io_type io_type;
uint64_t blocksize;
void *data;
VERIFY0(dmu_object_info(zd->zd_os, object, &doi));
blocksize = doi.doi_data_block_size;
data = umem_alloc(blocksize, UMEM_NOFAIL);
/*
* Pick an i/o type at random, biased toward writing block tags.
*/
io_type = ztest_random(ZTEST_IO_TYPES);
if (ztest_random(2) == 0)
io_type = ZTEST_IO_WRITE_TAG;
(void) pthread_rwlock_rdlock(&zd->zd_zilog_lock);
switch (io_type) {
case ZTEST_IO_WRITE_TAG:
ztest_bt_generate(&wbt, zd->zd_os, object, doi.doi_dnodesize,
offset, 0, 0, 0);
(void) ztest_write(zd, object, offset, sizeof (wbt), &wbt);
break;
case ZTEST_IO_WRITE_PATTERN:
(void) memset(data, 'a' + (object + offset) % 5, blocksize);
if (ztest_random(2) == 0) {
/*
* Induce fletcher2 collisions to ensure that
* zio_ddt_collision() detects and resolves them
* when using fletcher2-verify for deduplication.
*/
((uint64_t *)data)[0] ^= 1ULL << 63;
((uint64_t *)data)[4] ^= 1ULL << 63;
}
(void) ztest_write(zd, object, offset, blocksize, data);
break;
case ZTEST_IO_WRITE_ZEROES:
memset(data, 0, blocksize);
(void) ztest_write(zd, object, offset, blocksize, data);
break;
case ZTEST_IO_TRUNCATE:
(void) ztest_truncate(zd, object, offset, blocksize);
break;
case ZTEST_IO_SETATTR:
(void) ztest_setattr(zd, object);
break;
default:
break;
case ZTEST_IO_REWRITE:
(void) pthread_rwlock_rdlock(&ztest_name_lock);
err = ztest_dsl_prop_set_uint64(zd->zd_name,
ZFS_PROP_CHECKSUM, spa_dedup_checksum(ztest_spa),
B_FALSE);
ASSERT(err == 0 || err == ENOSPC);
err = ztest_dsl_prop_set_uint64(zd->zd_name,
ZFS_PROP_COMPRESSION,
ztest_random_dsl_prop(ZFS_PROP_COMPRESSION),
B_FALSE);
ASSERT(err == 0 || err == ENOSPC);
(void) pthread_rwlock_unlock(&ztest_name_lock);
VERIFY0(dmu_read(zd->zd_os, object, offset, blocksize, data,
DMU_READ_NO_PREFETCH));
(void) ztest_write(zd, object, offset, blocksize, data);
break;
}
(void) pthread_rwlock_unlock(&zd->zd_zilog_lock);
umem_free(data, blocksize);
}
/*
* Initialize an object description template.
*/
static void
ztest_od_init(ztest_od_t *od, uint64_t id, const char *tag, uint64_t index,
dmu_object_type_t type, uint64_t blocksize, uint64_t dnodesize,
uint64_t gen)
{
od->od_dir = ZTEST_DIROBJ;
od->od_object = 0;
od->od_crtype = type;
od->od_crblocksize = blocksize ? blocksize : ztest_random_blocksize();
od->od_crdnodesize = dnodesize ? dnodesize : ztest_random_dnodesize();
od->od_crgen = gen;
od->od_type = DMU_OT_NONE;
od->od_blocksize = 0;
od->od_gen = 0;
(void) snprintf(od->od_name, sizeof (od->od_name),
"%s(%"PRId64")[%"PRIu64"]",
tag, id, index);
}
/*
* Lookup or create the objects for a test using the od template.
* If the objects do not all exist, or if 'remove' is specified,
* remove any existing objects and create new ones. Otherwise,
* use the existing objects.
*/
static int
ztest_object_init(ztest_ds_t *zd, ztest_od_t *od, size_t size, boolean_t remove)
{
int count = size / sizeof (*od);
int rv = 0;
mutex_enter(&zd->zd_dirobj_lock);
if ((ztest_lookup(zd, od, count) != 0 || remove) &&
(ztest_remove(zd, od, count) != 0 ||
ztest_create(zd, od, count) != 0))
rv = -1;
zd->zd_od = od;
mutex_exit(&zd->zd_dirobj_lock);
return (rv);
}
void
ztest_zil_commit(ztest_ds_t *zd, uint64_t id)
{
(void) id;
zilog_t *zilog = zd->zd_zilog;
(void) pthread_rwlock_rdlock(&zd->zd_zilog_lock);
zil_commit(zilog, ztest_random(ZTEST_OBJECTS));
/*
* Remember the committed values in zd, which is in parent/child
* shared memory. If we die, the next iteration of ztest_run()
* will verify that the log really does contain this record.
*/
mutex_enter(&zilog->zl_lock);
ASSERT3P(zd->zd_shared, !=, NULL);
ASSERT3U(zd->zd_shared->zd_seq, <=, zilog->zl_commit_lr_seq);
zd->zd_shared->zd_seq = zilog->zl_commit_lr_seq;
mutex_exit(&zilog->zl_lock);
(void) pthread_rwlock_unlock(&zd->zd_zilog_lock);
}
/*
* This function is designed to simulate the operations that occur during a
* mount/unmount operation. We hold the dataset across these operations in an
* attempt to expose any implicit assumptions about ZIL management.
*/
void
ztest_zil_remount(ztest_ds_t *zd, uint64_t id)
{
(void) id;
objset_t *os = zd->zd_os;
/*
* We hold the ztest_vdev_lock so we don't cause problems with
* other threads that wish to remove a log device, such as
* ztest_device_removal().
*/
mutex_enter(&ztest_vdev_lock);
/*
* We grab the zd_dirobj_lock to ensure that no other thread is
* updating the zil (i.e. adding in-memory log records) and the
* zd_zilog_lock to block any I/O.
*/
mutex_enter(&zd->zd_dirobj_lock);
(void) pthread_rwlock_wrlock(&zd->zd_zilog_lock);
/* zfsvfs_teardown() */
zil_close(zd->zd_zilog);
/* zfsvfs_setup() */
VERIFY3P(zil_open(os, ztest_get_data, NULL), ==, zd->zd_zilog);
zil_replay(os, zd, ztest_replay_vector);
(void) pthread_rwlock_unlock(&zd->zd_zilog_lock);
mutex_exit(&zd->zd_dirobj_lock);
mutex_exit(&ztest_vdev_lock);
}
/*
* Verify that we can't destroy an active pool, create an existing pool,
* or create a pool with a bad vdev spec.
*/
void
ztest_spa_create_destroy(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
ztest_shared_opts_t *zo = &ztest_opts;
spa_t *spa;
nvlist_t *nvroot;
if (zo->zo_mmp_test)
return;
/*
* Attempt to create using a bad file.
*/
nvroot = make_vdev_root("/dev/bogus", NULL, NULL, 0, 0, NULL, 0, 0, 1);
VERIFY3U(ENOENT, ==,
spa_create("ztest_bad_file", nvroot, NULL, NULL, NULL));
fnvlist_free(nvroot);
/*
* Attempt to create using a bad mirror.
*/
nvroot = make_vdev_root("/dev/bogus", NULL, NULL, 0, 0, NULL, 0, 2, 1);
VERIFY3U(ENOENT, ==,
spa_create("ztest_bad_mirror", nvroot, NULL, NULL, NULL));
fnvlist_free(nvroot);
/*
* Attempt to create an existing pool. It shouldn't matter
* what's in the nvroot; we should fail with EEXIST.
*/
(void) pthread_rwlock_rdlock(&ztest_name_lock);
nvroot = make_vdev_root("/dev/bogus", NULL, NULL, 0, 0, NULL, 0, 0, 1);
VERIFY3U(EEXIST, ==,
spa_create(zo->zo_pool, nvroot, NULL, NULL, NULL));
fnvlist_free(nvroot);
/*
* We open a reference to the spa and then we try to export it
* expecting one of the following errors:
*
* EBUSY
* Because of the reference we just opened.
*
* ZFS_ERR_EXPORT_IN_PROGRESS
* For the case that there is another ztest thread doing
* an export concurrently.
*/
VERIFY0(spa_open(zo->zo_pool, &spa, FTAG));
int error = spa_destroy(zo->zo_pool);
if (error != EBUSY && error != ZFS_ERR_EXPORT_IN_PROGRESS) {
fatal(B_FALSE, "spa_destroy(%s) returned unexpected value %d",
spa->spa_name, error);
}
spa_close(spa, FTAG);
(void) pthread_rwlock_unlock(&ztest_name_lock);
}
/*
* Start and then stop the MMP threads to ensure the startup and shutdown code
* works properly. Actual protection and property-related code tested via ZTS.
*/
void
ztest_mmp_enable_disable(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
ztest_shared_opts_t *zo = &ztest_opts;
spa_t *spa = ztest_spa;
if (zo->zo_mmp_test)
return;
/*
* Since enabling MMP involves setting a property, it could not be done
* while the pool is suspended.
*/
if (spa_suspended(spa))
return;
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
mutex_enter(&spa->spa_props_lock);
zfs_multihost_fail_intervals = 0;
if (!spa_multihost(spa)) {
spa->spa_multihost = B_TRUE;
mmp_thread_start(spa);
}
mutex_exit(&spa->spa_props_lock);
spa_config_exit(spa, SCL_CONFIG, FTAG);
txg_wait_synced(spa_get_dsl(spa), 0);
mmp_signal_all_threads();
txg_wait_synced(spa_get_dsl(spa), 0);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
mutex_enter(&spa->spa_props_lock);
if (spa_multihost(spa)) {
mmp_thread_stop(spa);
spa->spa_multihost = B_FALSE;
}
mutex_exit(&spa->spa_props_lock);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
+static int
+ztest_get_raidz_children(spa_t *spa)
+{
+ (void) spa;
+ vdev_t *raidvd;
+
+ ASSERT(MUTEX_HELD(&ztest_vdev_lock));
+
+ if (ztest_opts.zo_raid_do_expand) {
+ raidvd = ztest_spa->spa_root_vdev->vdev_child[0];
+
+ ASSERT(raidvd->vdev_ops == &vdev_raidz_ops);
+
+ return (raidvd->vdev_children);
+ }
+
+ return (ztest_opts.zo_raid_children);
+}
+
void
ztest_spa_upgrade(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
spa_t *spa;
uint64_t initial_version = SPA_VERSION_INITIAL;
- uint64_t version, newversion;
+ uint64_t raidz_children, version, newversion;
nvlist_t *nvroot, *props;
char *name;
if (ztest_opts.zo_mmp_test)
return;
/* dRAID added after feature flags, skip upgrade test. */
if (strcmp(ztest_opts.zo_raid_type, VDEV_TYPE_DRAID) == 0)
return;
mutex_enter(&ztest_vdev_lock);
name = kmem_asprintf("%s_upgrade", ztest_opts.zo_pool);
/*
* Clean up from previous runs.
*/
(void) spa_destroy(name);
+ raidz_children = ztest_get_raidz_children(ztest_spa);
+
nvroot = make_vdev_root(NULL, NULL, name, ztest_opts.zo_vdev_size, 0,
- NULL, ztest_opts.zo_raid_children, ztest_opts.zo_mirrors, 1);
+ NULL, raidz_children, ztest_opts.zo_mirrors, 1);
/*
* If we're configuring a RAIDZ device then make sure that the
* initial version is capable of supporting that feature.
*/
switch (ztest_opts.zo_raid_parity) {
case 0:
case 1:
initial_version = SPA_VERSION_INITIAL;
break;
case 2:
initial_version = SPA_VERSION_RAIDZ2;
break;
case 3:
initial_version = SPA_VERSION_RAIDZ3;
break;
}
/*
* Create a pool with a spa version that can be upgraded. Pick
* a value between initial_version and SPA_VERSION_BEFORE_FEATURES.
*/
do {
version = ztest_random_spa_version(initial_version);
} while (version > SPA_VERSION_BEFORE_FEATURES);
props = fnvlist_alloc();
fnvlist_add_uint64(props,
zpool_prop_to_name(ZPOOL_PROP_VERSION), version);
VERIFY0(spa_create(name, nvroot, props, NULL, NULL));
fnvlist_free(nvroot);
fnvlist_free(props);
VERIFY0(spa_open(name, &spa, FTAG));
VERIFY3U(spa_version(spa), ==, version);
newversion = ztest_random_spa_version(version + 1);
if (ztest_opts.zo_verbose >= 4) {
(void) printf("upgrading spa version from "
"%"PRIu64" to %"PRIu64"\n",
version, newversion);
}
spa_upgrade(spa, newversion);
VERIFY3U(spa_version(spa), >, version);
VERIFY3U(spa_version(spa), ==, fnvlist_lookup_uint64(spa->spa_config,
zpool_prop_to_name(ZPOOL_PROP_VERSION)));
spa_close(spa, FTAG);
kmem_strfree(name);
mutex_exit(&ztest_vdev_lock);
}
static void
ztest_spa_checkpoint(spa_t *spa)
{
ASSERT(MUTEX_HELD(&ztest_checkpoint_lock));
int error = spa_checkpoint(spa->spa_name);
switch (error) {
case 0:
case ZFS_ERR_DEVRM_IN_PROGRESS:
case ZFS_ERR_DISCARDING_CHECKPOINT:
case ZFS_ERR_CHECKPOINT_EXISTS:
+ case ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS:
break;
case ENOSPC:
ztest_record_enospc(FTAG);
break;
default:
fatal(B_FALSE, "spa_checkpoint(%s) = %d", spa->spa_name, error);
}
}
static void
ztest_spa_discard_checkpoint(spa_t *spa)
{
ASSERT(MUTEX_HELD(&ztest_checkpoint_lock));
int error = spa_checkpoint_discard(spa->spa_name);
switch (error) {
case 0:
case ZFS_ERR_DISCARDING_CHECKPOINT:
case ZFS_ERR_NO_CHECKPOINT:
break;
default:
fatal(B_FALSE, "spa_discard_checkpoint(%s) = %d",
spa->spa_name, error);
}
}
void
ztest_spa_checkpoint_create_discard(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
spa_t *spa = ztest_spa;
mutex_enter(&ztest_checkpoint_lock);
if (ztest_random(2) == 0) {
ztest_spa_checkpoint(spa);
} else {
ztest_spa_discard_checkpoint(spa);
}
mutex_exit(&ztest_checkpoint_lock);
}
static vdev_t *
vdev_lookup_by_path(vdev_t *vd, const char *path)
{
vdev_t *mvd;
int c;
if (vd->vdev_path != NULL && strcmp(path, vd->vdev_path) == 0)
return (vd);
for (c = 0; c < vd->vdev_children; c++)
if ((mvd = vdev_lookup_by_path(vd->vdev_child[c], path)) !=
NULL)
return (mvd);
return (NULL);
}
static int
spa_num_top_vdevs(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
ASSERT3U(spa_config_held(spa, SCL_VDEV, RW_READER), ==, SCL_VDEV);
return (rvd->vdev_children);
}
/*
* Verify that vdev_add() works as expected.
*/
void
ztest_vdev_add_remove(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
ztest_shared_t *zs = ztest_shared;
spa_t *spa = ztest_spa;
uint64_t leaves;
uint64_t guid;
+ uint64_t raidz_children;
+
nvlist_t *nvroot;
int error;
if (ztest_opts.zo_mmp_test)
return;
mutex_enter(&ztest_vdev_lock);
- leaves = MAX(zs->zs_mirrors + zs->zs_splits, 1) *
- ztest_opts.zo_raid_children;
+ raidz_children = ztest_get_raidz_children(spa);
+ leaves = MAX(zs->zs_mirrors + zs->zs_splits, 1) * raidz_children;
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
ztest_shared->zs_vdev_next_leaf = spa_num_top_vdevs(spa) * leaves;
/*
* If we have slogs then remove them 1/4 of the time.
*/
if (spa_has_slogs(spa) && ztest_random(4) == 0) {
metaslab_group_t *mg;
/*
* find the first real slog in log allocation class
*/
mg = spa_log_class(spa)->mc_allocator[0].mca_rotor;
while (!mg->mg_vd->vdev_islog)
mg = mg->mg_next;
guid = mg->mg_vd->vdev_guid;
spa_config_exit(spa, SCL_VDEV, FTAG);
/*
* We have to grab the zs_name_lock as writer to
* prevent a race between removing a slog (dmu_objset_find)
* and destroying a dataset. Removing the slog will
* grab a reference on the dataset which may cause
* dsl_destroy_head() to fail with EBUSY thus
* leaving the dataset in an inconsistent state.
*/
pthread_rwlock_wrlock(&ztest_name_lock);
error = spa_vdev_remove(spa, guid, B_FALSE);
pthread_rwlock_unlock(&ztest_name_lock);
switch (error) {
case 0:
case EEXIST: /* Generic zil_reset() error */
case EBUSY: /* Replay required */
case EACCES: /* Crypto key not loaded */
case ZFS_ERR_CHECKPOINT_EXISTS:
case ZFS_ERR_DISCARDING_CHECKPOINT:
break;
default:
fatal(B_FALSE, "spa_vdev_remove() = %d", error);
}
} else {
spa_config_exit(spa, SCL_VDEV, FTAG);
/*
* Make 1/4 of the devices be log devices
*/
nvroot = make_vdev_root(NULL, NULL, NULL,
ztest_opts.zo_vdev_size, 0, (ztest_random(4) == 0) ?
- "log" : NULL, ztest_opts.zo_raid_children, zs->zs_mirrors,
+ "log" : NULL, raidz_children, zs->zs_mirrors,
1);
error = spa_vdev_add(spa, nvroot);
fnvlist_free(nvroot);
switch (error) {
case 0:
break;
case ENOSPC:
ztest_record_enospc("spa_vdev_add");
break;
default:
fatal(B_FALSE, "spa_vdev_add() = %d", error);
}
}
mutex_exit(&ztest_vdev_lock);
}
void
ztest_vdev_class_add(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
ztest_shared_t *zs = ztest_shared;
spa_t *spa = ztest_spa;
uint64_t leaves;
nvlist_t *nvroot;
+ uint64_t raidz_children;
const char *class = (ztest_random(2) == 0) ?
VDEV_ALLOC_BIAS_SPECIAL : VDEV_ALLOC_BIAS_DEDUP;
int error;
/*
* By default add a special vdev 50% of the time
*/
if ((ztest_opts.zo_special_vdevs == ZTEST_VDEV_CLASS_OFF) ||
(ztest_opts.zo_special_vdevs == ZTEST_VDEV_CLASS_RND &&
ztest_random(2) == 0)) {
return;
}
mutex_enter(&ztest_vdev_lock);
/* Only test with mirrors */
if (zs->zs_mirrors < 2) {
mutex_exit(&ztest_vdev_lock);
return;
}
/* requires feature@allocation_classes */
if (!spa_feature_is_enabled(spa, SPA_FEATURE_ALLOCATION_CLASSES)) {
mutex_exit(&ztest_vdev_lock);
return;
}
- leaves = MAX(zs->zs_mirrors + zs->zs_splits, 1) *
- ztest_opts.zo_raid_children;
+ raidz_children = ztest_get_raidz_children(spa);
+ leaves = MAX(zs->zs_mirrors + zs->zs_splits, 1) * raidz_children;
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
ztest_shared->zs_vdev_next_leaf = spa_num_top_vdevs(spa) * leaves;
spa_config_exit(spa, SCL_VDEV, FTAG);
nvroot = make_vdev_root(NULL, NULL, NULL, ztest_opts.zo_vdev_size, 0,
- class, ztest_opts.zo_raid_children, zs->zs_mirrors, 1);
+ class, raidz_children, zs->zs_mirrors, 1);
error = spa_vdev_add(spa, nvroot);
fnvlist_free(nvroot);
if (error == ENOSPC)
ztest_record_enospc("spa_vdev_add");
else if (error != 0)
fatal(B_FALSE, "spa_vdev_add() = %d", error);
/*
* 50% of the time allow small blocks in the special class
*/
if (error == 0 &&
spa_special_class(spa)->mc_groups == 1 && ztest_random(2) == 0) {
if (ztest_opts.zo_verbose >= 3)
(void) printf("Enabling special VDEV small blocks\n");
error = ztest_dsl_prop_set_uint64(zd->zd_name,
ZFS_PROP_SPECIAL_SMALL_BLOCKS, 32768, B_FALSE);
ASSERT(error == 0 || error == ENOSPC);
}
mutex_exit(&ztest_vdev_lock);
if (ztest_opts.zo_verbose >= 3) {
metaslab_class_t *mc;
if (strcmp(class, VDEV_ALLOC_BIAS_SPECIAL) == 0)
mc = spa_special_class(spa);
else
mc = spa_dedup_class(spa);
(void) printf("Added a %s mirrored vdev (of %d)\n",
class, (int)mc->mc_groups);
}
}
/*
* Verify that adding/removing aux devices (l2arc, hot spare) works as expected.
*/
void
ztest_vdev_aux_add_remove(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
ztest_shared_t *zs = ztest_shared;
spa_t *spa = ztest_spa;
vdev_t *rvd = spa->spa_root_vdev;
spa_aux_vdev_t *sav;
const char *aux;
char *path;
uint64_t guid = 0;
int error, ignore_err = 0;
if (ztest_opts.zo_mmp_test)
return;
path = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
if (ztest_random(2) == 0) {
sav = &spa->spa_spares;
aux = ZPOOL_CONFIG_SPARES;
} else {
sav = &spa->spa_l2cache;
aux = ZPOOL_CONFIG_L2CACHE;
}
mutex_enter(&ztest_vdev_lock);
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
if (sav->sav_count != 0 && ztest_random(4) == 0) {
/*
* Pick a random device to remove.
*/
vdev_t *svd = sav->sav_vdevs[ztest_random(sav->sav_count)];
/* dRAID spares cannot be removed; try anyways to see ENOTSUP */
if (strstr(svd->vdev_path, VDEV_TYPE_DRAID) != NULL)
ignore_err = ENOTSUP;
guid = svd->vdev_guid;
} else {
/*
* Find an unused device we can add.
*/
zs->zs_vdev_aux = 0;
for (;;) {
int c;
(void) snprintf(path, MAXPATHLEN, ztest_aux_template,
ztest_opts.zo_dir, ztest_opts.zo_pool, aux,
zs->zs_vdev_aux);
for (c = 0; c < sav->sav_count; c++)
if (strcmp(sav->sav_vdevs[c]->vdev_path,
path) == 0)
break;
if (c == sav->sav_count &&
vdev_lookup_by_path(rvd, path) == NULL)
break;
zs->zs_vdev_aux++;
}
}
spa_config_exit(spa, SCL_VDEV, FTAG);
if (guid == 0) {
/*
* Add a new device.
*/
nvlist_t *nvroot = make_vdev_root(NULL, aux, NULL,
(ztest_opts.zo_vdev_size * 5) / 4, 0, NULL, 0, 0, 1);
error = spa_vdev_add(spa, nvroot);
switch (error) {
case 0:
break;
default:
fatal(B_FALSE, "spa_vdev_add(%p) = %d", nvroot, error);
}
fnvlist_free(nvroot);
} else {
/*
* Remove an existing device. Sometimes, dirty its
* vdev state first to make sure we handle removal
* of devices that have pending state changes.
*/
if (ztest_random(2) == 0)
(void) vdev_online(spa, guid, 0, NULL);
error = spa_vdev_remove(spa, guid, B_FALSE);
switch (error) {
case 0:
case EBUSY:
case ZFS_ERR_CHECKPOINT_EXISTS:
case ZFS_ERR_DISCARDING_CHECKPOINT:
break;
default:
if (error != ignore_err)
fatal(B_FALSE,
"spa_vdev_remove(%"PRIu64") = %d",
guid, error);
}
}
mutex_exit(&ztest_vdev_lock);
umem_free(path, MAXPATHLEN);
}
/*
* split a pool if it has mirror tlvdevs
*/
void
ztest_split_pool(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
ztest_shared_t *zs = ztest_shared;
spa_t *spa = ztest_spa;
vdev_t *rvd = spa->spa_root_vdev;
nvlist_t *tree, **child, *config, *split, **schild;
uint_t c, children, schildren = 0, lastlogid = 0;
int error = 0;
if (ztest_opts.zo_mmp_test)
return;
mutex_enter(&ztest_vdev_lock);
/* ensure we have a usable config; mirrors of raidz aren't supported */
if (zs->zs_mirrors < 3 || ztest_opts.zo_raid_children > 1) {
mutex_exit(&ztest_vdev_lock);
return;
}
/* clean up the old pool, if any */
(void) spa_destroy("splitp");
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
/* generate a config from the existing config */
mutex_enter(&spa->spa_props_lock);
tree = fnvlist_lookup_nvlist(spa->spa_config, ZPOOL_CONFIG_VDEV_TREE);
mutex_exit(&spa->spa_props_lock);
VERIFY0(nvlist_lookup_nvlist_array(tree, ZPOOL_CONFIG_CHILDREN,
&child, &children));
schild = umem_alloc(rvd->vdev_children * sizeof (nvlist_t *),
UMEM_NOFAIL);
for (c = 0; c < children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
nvlist_t **mchild;
uint_t mchildren;
if (tvd->vdev_islog || tvd->vdev_ops == &vdev_hole_ops) {
schild[schildren] = fnvlist_alloc();
fnvlist_add_string(schild[schildren],
ZPOOL_CONFIG_TYPE, VDEV_TYPE_HOLE);
fnvlist_add_uint64(schild[schildren],
ZPOOL_CONFIG_IS_HOLE, 1);
if (lastlogid == 0)
lastlogid = schildren;
++schildren;
continue;
}
lastlogid = 0;
VERIFY0(nvlist_lookup_nvlist_array(child[c],
ZPOOL_CONFIG_CHILDREN, &mchild, &mchildren));
schild[schildren++] = fnvlist_dup(mchild[0]);
}
/* OK, create a config that can be used to split */
split = fnvlist_alloc();
fnvlist_add_string(split, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT);
fnvlist_add_nvlist_array(split, ZPOOL_CONFIG_CHILDREN,
(const nvlist_t **)schild, lastlogid != 0 ? lastlogid : schildren);
config = fnvlist_alloc();
fnvlist_add_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, split);
for (c = 0; c < schildren; c++)
fnvlist_free(schild[c]);
umem_free(schild, rvd->vdev_children * sizeof (nvlist_t *));
fnvlist_free(split);
spa_config_exit(spa, SCL_VDEV, FTAG);
(void) pthread_rwlock_wrlock(&ztest_name_lock);
error = spa_vdev_split_mirror(spa, "splitp", config, NULL, B_FALSE);
(void) pthread_rwlock_unlock(&ztest_name_lock);
fnvlist_free(config);
if (error == 0) {
(void) printf("successful split - results:\n");
mutex_enter(&spa_namespace_lock);
show_pool_stats(spa);
show_pool_stats(spa_lookup("splitp"));
mutex_exit(&spa_namespace_lock);
++zs->zs_splits;
--zs->zs_mirrors;
}
mutex_exit(&ztest_vdev_lock);
}
/*
* Verify that we can attach and detach devices.
*/
void
ztest_vdev_attach_detach(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
ztest_shared_t *zs = ztest_shared;
spa_t *spa = ztest_spa;
spa_aux_vdev_t *sav = &spa->spa_spares;
vdev_t *rvd = spa->spa_root_vdev;
vdev_t *oldvd, *newvd, *pvd;
nvlist_t *root;
uint64_t leaves;
uint64_t leaf, top;
uint64_t ashift = ztest_get_ashift();
uint64_t oldguid, pguid;
uint64_t oldsize, newsize;
+ uint64_t raidz_children;
char *oldpath, *newpath;
int replacing;
int oldvd_has_siblings = B_FALSE;
int newvd_is_spare = B_FALSE;
int newvd_is_dspare = B_FALSE;
int oldvd_is_log;
int oldvd_is_special;
int error, expected_error;
if (ztest_opts.zo_mmp_test)
return;
oldpath = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
newpath = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
mutex_enter(&ztest_vdev_lock);
- leaves = MAX(zs->zs_mirrors, 1) * ztest_opts.zo_raid_children;
+ raidz_children = ztest_get_raidz_children(spa);
+ leaves = MAX(zs->zs_mirrors, 1) * raidz_children;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
/*
* If a vdev is in the process of being removed, its removal may
* finish while we are in progress, leading to an unexpected error
* value. Don't bother trying to attach while we are in the middle
* of removal.
*/
if (ztest_device_removal_active) {
spa_config_exit(spa, SCL_ALL, FTAG);
goto out;
}
+ /*
+ * RAIDZ leaf VDEV mirrors are not currently supported while a
+ * RAIDZ expansion is in progress.
+ */
+ if (ztest_opts.zo_raid_do_expand) {
+ spa_config_exit(spa, SCL_ALL, FTAG);
+ goto out;
+ }
+
/*
* Decide whether to do an attach or a replace.
*/
replacing = ztest_random(2);
/*
* Pick a random top-level vdev.
*/
top = ztest_random_vdev_top(spa, B_TRUE);
/*
* Pick a random leaf within it.
*/
leaf = ztest_random(leaves);
/*
* Locate this vdev.
*/
oldvd = rvd->vdev_child[top];
/* pick a child from the mirror */
if (zs->zs_mirrors >= 1) {
ASSERT3P(oldvd->vdev_ops, ==, &vdev_mirror_ops);
ASSERT3U(oldvd->vdev_children, >=, zs->zs_mirrors);
- oldvd = oldvd->vdev_child[leaf / ztest_opts.zo_raid_children];
+ oldvd = oldvd->vdev_child[leaf / raidz_children];
}
/* pick a child out of the raidz group */
if (ztest_opts.zo_raid_children > 1) {
if (strcmp(oldvd->vdev_ops->vdev_op_type, "raidz") == 0)
ASSERT3P(oldvd->vdev_ops, ==, &vdev_raidz_ops);
else
ASSERT3P(oldvd->vdev_ops, ==, &vdev_draid_ops);
- ASSERT3U(oldvd->vdev_children, ==, ztest_opts.zo_raid_children);
- oldvd = oldvd->vdev_child[leaf % ztest_opts.zo_raid_children];
+ oldvd = oldvd->vdev_child[leaf % raidz_children];
}
/*
* If we're already doing an attach or replace, oldvd may be a
* mirror vdev -- in which case, pick a random child.
*/
while (oldvd->vdev_children != 0) {
oldvd_has_siblings = B_TRUE;
ASSERT3U(oldvd->vdev_children, >=, 2);
oldvd = oldvd->vdev_child[ztest_random(oldvd->vdev_children)];
}
oldguid = oldvd->vdev_guid;
oldsize = vdev_get_min_asize(oldvd);
oldvd_is_log = oldvd->vdev_top->vdev_islog;
oldvd_is_special =
oldvd->vdev_top->vdev_alloc_bias == VDEV_BIAS_SPECIAL ||
oldvd->vdev_top->vdev_alloc_bias == VDEV_BIAS_DEDUP;
(void) strlcpy(oldpath, oldvd->vdev_path, MAXPATHLEN);
pvd = oldvd->vdev_parent;
pguid = pvd->vdev_guid;
/*
* If oldvd has siblings, then half of the time, detach it. Prior
* to the detach the pool is scrubbed in order to prevent creating
* unrepairable blocks as a result of the data corruption injection.
*/
if (oldvd_has_siblings && ztest_random(2) == 0) {
spa_config_exit(spa, SCL_ALL, FTAG);
error = ztest_scrub_impl(spa);
if (error)
goto out;
error = spa_vdev_detach(spa, oldguid, pguid, B_FALSE);
if (error != 0 && error != ENODEV && error != EBUSY &&
error != ENOTSUP && error != ZFS_ERR_CHECKPOINT_EXISTS &&
error != ZFS_ERR_DISCARDING_CHECKPOINT)
fatal(B_FALSE, "detach (%s) returned %d",
oldpath, error);
goto out;
}
/*
* For the new vdev, choose with equal probability between the two
* standard paths (ending in either 'a' or 'b') or a random hot spare.
*/
if (sav->sav_count != 0 && ztest_random(3) == 0) {
newvd = sav->sav_vdevs[ztest_random(sav->sav_count)];
newvd_is_spare = B_TRUE;
if (newvd->vdev_ops == &vdev_draid_spare_ops)
newvd_is_dspare = B_TRUE;
(void) strlcpy(newpath, newvd->vdev_path, MAXPATHLEN);
} else {
(void) snprintf(newpath, MAXPATHLEN, ztest_dev_template,
ztest_opts.zo_dir, ztest_opts.zo_pool,
top * leaves + leaf);
if (ztest_random(2) == 0)
newpath[strlen(newpath) - 1] = 'b';
newvd = vdev_lookup_by_path(rvd, newpath);
}
if (newvd) {
/*
* Reopen to ensure the vdev's asize field isn't stale.
*/
vdev_reopen(newvd);
newsize = vdev_get_min_asize(newvd);
} else {
/*
* Make newsize a little bigger or smaller than oldsize.
* If it's smaller, the attach should fail.
* If it's larger, and we're doing a replace,
* we should get dynamic LUN growth when we're done.
*/
newsize = 10 * oldsize / (9 + ztest_random(3));
}
/*
* If pvd is not a mirror or root, the attach should fail with ENOTSUP,
* unless it's a replace; in that case any non-replacing parent is OK.
*
* If newvd is already part of the pool, it should fail with EBUSY.
*
* If newvd is too small, it should fail with EOVERFLOW.
*
* If newvd is a distributed spare and it's being attached to a
* dRAID which is not its parent it should fail with EINVAL.
*/
if (pvd->vdev_ops != &vdev_mirror_ops &&
pvd->vdev_ops != &vdev_root_ops && (!replacing ||
pvd->vdev_ops == &vdev_replacing_ops ||
pvd->vdev_ops == &vdev_spare_ops))
expected_error = ENOTSUP;
else if (newvd_is_spare &&
(!replacing || oldvd_is_log || oldvd_is_special))
expected_error = ENOTSUP;
else if (newvd == oldvd)
expected_error = replacing ? 0 : EBUSY;
else if (vdev_lookup_by_path(rvd, newpath) != NULL)
expected_error = EBUSY;
else if (!newvd_is_dspare && newsize < oldsize)
expected_error = EOVERFLOW;
else if (ashift > oldvd->vdev_top->vdev_ashift)
expected_error = EDOM;
else if (newvd_is_dspare && pvd != vdev_draid_spare_get_parent(newvd))
expected_error = EINVAL;
else
expected_error = 0;
spa_config_exit(spa, SCL_ALL, FTAG);
/*
* Build the nvlist describing newpath.
*/
root = make_vdev_root(newpath, NULL, NULL, newvd == NULL ? newsize : 0,
ashift, NULL, 0, 0, 1);
/*
* When supported select either a healing or sequential resilver.
*/
boolean_t rebuilding = B_FALSE;
if (pvd->vdev_ops == &vdev_mirror_ops ||
pvd->vdev_ops == &vdev_root_ops) {
rebuilding = !!ztest_random(2);
}
error = spa_vdev_attach(spa, oldguid, root, replacing, rebuilding);
fnvlist_free(root);
/*
* If our parent was the replacing vdev, but the replace completed,
* then instead of failing with ENOTSUP we may either succeed,
* fail with ENODEV, or fail with EOVERFLOW.
*/
if (expected_error == ENOTSUP &&
(error == 0 || error == ENODEV || error == EOVERFLOW))
expected_error = error;
/*
* If someone grew the LUN, the replacement may be too small.
*/
if (error == EOVERFLOW || error == EBUSY)
expected_error = error;
if (error == ZFS_ERR_CHECKPOINT_EXISTS ||
error == ZFS_ERR_DISCARDING_CHECKPOINT ||
error == ZFS_ERR_RESILVER_IN_PROGRESS ||
error == ZFS_ERR_REBUILD_IN_PROGRESS)
expected_error = error;
if (error != expected_error && expected_error != EBUSY) {
fatal(B_FALSE, "attach (%s %"PRIu64", %s %"PRIu64", %d) "
"returned %d, expected %d",
oldpath, oldsize, newpath,
newsize, replacing, error, expected_error);
}
out:
mutex_exit(&ztest_vdev_lock);
umem_free(oldpath, MAXPATHLEN);
umem_free(newpath, MAXPATHLEN);
}
+static void
+raidz_scratch_verify(void)
+{
+ spa_t *spa;
+ uint64_t write_size, logical_size, offset;
+ raidz_reflow_scratch_state_t state;
+ vdev_raidz_expand_t *vre;
+ vdev_t *raidvd;
+
+ ASSERT(raidz_expand_pause_point == RAIDZ_EXPAND_PAUSE_NONE);
+
+ if (ztest_scratch_state->zs_raidz_scratch_verify_pause == 0)
+ return;
+
+ kernel_init(SPA_MODE_READ);
+
+ mutex_enter(&spa_namespace_lock);
+ spa = spa_lookup(ztest_opts.zo_pool);
+ ASSERT(spa);
+ spa->spa_import_flags |= ZFS_IMPORT_SKIP_MMP;
+ mutex_exit(&spa_namespace_lock);
+
+ VERIFY0(spa_open(ztest_opts.zo_pool, &spa, FTAG));
+
+ ASSERT3U(RRSS_GET_OFFSET(&spa->spa_uberblock), !=, UINT64_MAX);
+
+ mutex_enter(&ztest_vdev_lock);
+
+ spa_config_enter(spa, SCL_ALL, FTAG, RW_READER);
+
+ vre = spa->spa_raidz_expand;
+ if (vre == NULL)
+ goto out;
+
+ raidvd = vdev_lookup_top(spa, vre->vre_vdev_id);
+ offset = RRSS_GET_OFFSET(&spa->spa_uberblock);
+ state = RRSS_GET_STATE(&spa->spa_uberblock);
+ write_size = P2ALIGN(VDEV_BOOT_SIZE, 1 << raidvd->vdev_ashift);
+ logical_size = write_size * raidvd->vdev_children;
+
+ switch (state) {
+ /*
+ * Initial state of reflow process. RAIDZ expansion was
+ * requested by user, but scratch object was not created.
+ */
+ case RRSS_SCRATCH_NOT_IN_USE:
+ ASSERT3U(offset, ==, 0);
+ break;
+
+ /*
+ * Scratch object was synced and stored in boot area.
+ */
+ case RRSS_SCRATCH_VALID:
+
+ /*
+ * Scratch object was synced back to raidz start offset,
+ * raidz is ready for sector by sector reflow process.
+ */
+ case RRSS_SCRATCH_INVALID_SYNCED:
+
+ /*
+ * Scratch object was synced back to raidz start offset
+ * on zpool importing, raidz is ready for sector by sector
+ * reflow process.
+ */
+ case RRSS_SCRATCH_INVALID_SYNCED_ON_IMPORT:
+ ASSERT3U(offset, ==, logical_size);
+ break;
+
+ /*
+ * Sector by sector reflow process started.
+ */
+ case RRSS_SCRATCH_INVALID_SYNCED_REFLOW:
+ ASSERT3U(offset, >=, logical_size);
+ break;
+ }
+
+out:
+ spa_config_exit(spa, SCL_ALL, FTAG);
+
+ mutex_exit(&ztest_vdev_lock);
+
+ ztest_scratch_state->zs_raidz_scratch_verify_pause = 0;
+
+ spa_close(spa, FTAG);
+ kernel_fini();
+}
+
+static void
+ztest_scratch_thread(void *arg)
+{
+ (void) arg;
+
+ /* wait up to 10 seconds */
+ for (int t = 100; t > 0; t -= 1) {
+ if (raidz_expand_pause_point == RAIDZ_EXPAND_PAUSE_NONE)
+ thread_exit();
+
+ (void) poll(NULL, 0, 100);
+ }
+
+ /* killed when the scratch area progress reached a certain point */
+ ztest_kill(ztest_shared);
+}
+
+/*
+ * Verify that we can attach raidz device.
+ */
+void
+ztest_vdev_raidz_attach(ztest_ds_t *zd, uint64_t id)
+{
+ (void) zd, (void) id;
+ ztest_shared_t *zs = ztest_shared;
+ spa_t *spa = ztest_spa;
+ uint64_t leaves, raidz_children, newsize, ashift = ztest_get_ashift();
+ kthread_t *scratch_thread = NULL;
+ vdev_t *newvd, *pvd;
+ nvlist_t *root;
+ char *newpath = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
+ int error, expected_error = 0;
+
+ mutex_enter(&ztest_vdev_lock);
+
+ spa_config_enter(spa, SCL_ALL, FTAG, RW_READER);
+
+ /* Only allow attach when raid-kind = 'eraidz' */
+ if (!ztest_opts.zo_raid_do_expand) {
+ spa_config_exit(spa, SCL_ALL, FTAG);
+ goto out;
+ }
+
+ if (ztest_opts.zo_mmp_test) {
+ spa_config_exit(spa, SCL_ALL, FTAG);
+ goto out;
+ }
+
+ if (ztest_device_removal_active) {
+ spa_config_exit(spa, SCL_ALL, FTAG);
+ goto out;
+ }
+
+ pvd = vdev_lookup_top(spa, 0);
+
+ ASSERT(pvd->vdev_ops == &vdev_raidz_ops);
+
+ /*
+ * Get size of a child of the raidz group,
+ * make sure device is a bit bigger
+ */
+ newvd = pvd->vdev_child[ztest_random(pvd->vdev_children)];
+ newsize = 10 * vdev_get_min_asize(newvd) / (9 + ztest_random(2));
+
+ /*
+ * Get next attached leaf id
+ */
+ raidz_children = ztest_get_raidz_children(spa);
+ leaves = MAX(zs->zs_mirrors + zs->zs_splits, 1) * raidz_children;
+ zs->zs_vdev_next_leaf = spa_num_top_vdevs(spa) * leaves;
+
+ if (spa->spa_raidz_expand)
+ expected_error = ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS;
+
+ spa_config_exit(spa, SCL_ALL, FTAG);
+
+ /*
+ * Path to vdev to be attached
+ */
+ (void) snprintf(newpath, MAXPATHLEN, ztest_dev_template,
+ ztest_opts.zo_dir, ztest_opts.zo_pool, zs->zs_vdev_next_leaf);
+
+ /*
+ * Build the nvlist describing newpath.
+ */
+ root = make_vdev_root(newpath, NULL, NULL, newsize, ashift, NULL,
+ 0, 0, 1);
+
+ /*
+ * 50% of the time, set raidz_expand_pause_point to cause
+ * raidz_reflow_scratch_sync() to pause at a certain point and
+ * then kill the test after 10 seconds so raidz_scratch_verify()
+ * can confirm consistency when the pool is imported.
+ */
+ if (ztest_random(2) == 0 && expected_error == 0) {
+ raidz_expand_pause_point =
+ ztest_random(RAIDZ_EXPAND_PAUSE_SCRATCH_POST_REFLOW_2) + 1;
+ scratch_thread = thread_create(NULL, 0, ztest_scratch_thread,
+ ztest_shared, 0, NULL, TS_RUN | TS_JOINABLE, defclsyspri);
+ }
+
+ error = spa_vdev_attach(spa, pvd->vdev_guid, root, B_FALSE, B_FALSE);
+
+ nvlist_free(root);
+
+ if (error == EOVERFLOW || error == ENXIO ||
+ error == ZFS_ERR_CHECKPOINT_EXISTS ||
+ error == ZFS_ERR_DISCARDING_CHECKPOINT)
+ expected_error = error;
+
+ if (error != 0 && error != expected_error) {
+ fatal(0, "raidz attach (%s %"PRIu64") returned %d, expected %d",
+ newpath, newsize, error, expected_error);
+ }
+
+ if (raidz_expand_pause_point) {
+ if (error != 0) {
+ /*
+ * Do not verify scratch object in case of error
+ * returned by vdev attaching.
+ */
+ raidz_expand_pause_point = RAIDZ_EXPAND_PAUSE_NONE;
+ }
+
+ VERIFY0(thread_join(scratch_thread));
+ }
+out:
+ mutex_exit(&ztest_vdev_lock);
+
+ umem_free(newpath, MAXPATHLEN);
+}
+
void
ztest_device_removal(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
spa_t *spa = ztest_spa;
vdev_t *vd;
uint64_t guid;
int error;
mutex_enter(&ztest_vdev_lock);
if (ztest_device_removal_active) {
mutex_exit(&ztest_vdev_lock);
return;
}
/*
* Remove a random top-level vdev and wait for removal to finish.
*/
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
vd = vdev_lookup_top(spa, ztest_random_vdev_top(spa, B_FALSE));
guid = vd->vdev_guid;
spa_config_exit(spa, SCL_VDEV, FTAG);
error = spa_vdev_remove(spa, guid, B_FALSE);
if (error == 0) {
ztest_device_removal_active = B_TRUE;
mutex_exit(&ztest_vdev_lock);
/*
* spa->spa_vdev_removal is created in a sync task that
* is initiated via dsl_sync_task_nowait(). Since the
* task may not run before spa_vdev_remove() returns, we
* must wait at least 1 txg to ensure that the removal
* struct has been created.
*/
txg_wait_synced(spa_get_dsl(spa), 0);
while (spa->spa_removing_phys.sr_state == DSS_SCANNING)
txg_wait_synced(spa_get_dsl(spa), 0);
} else {
mutex_exit(&ztest_vdev_lock);
return;
}
/*
* The pool needs to be scrubbed after completing device removal.
* Failure to do so may result in checksum errors due to the
* strategy employed by ztest_fault_inject() when selecting which
* offset are redundant and can be damaged.
*/
error = spa_scan(spa, POOL_SCAN_SCRUB);
if (error == 0) {
while (dsl_scan_scrubbing(spa_get_dsl(spa)))
txg_wait_synced(spa_get_dsl(spa), 0);
}
mutex_enter(&ztest_vdev_lock);
ztest_device_removal_active = B_FALSE;
mutex_exit(&ztest_vdev_lock);
}
/*
* Callback function which expands the physical size of the vdev.
*/
static vdev_t *
grow_vdev(vdev_t *vd, void *arg)
{
spa_t *spa __maybe_unused = vd->vdev_spa;
size_t *newsize = arg;
size_t fsize;
int fd;
ASSERT3S(spa_config_held(spa, SCL_STATE, RW_READER), ==, SCL_STATE);
ASSERT(vd->vdev_ops->vdev_op_leaf);
if ((fd = open(vd->vdev_path, O_RDWR)) == -1)
return (vd);
fsize = lseek(fd, 0, SEEK_END);
VERIFY0(ftruncate(fd, *newsize));
if (ztest_opts.zo_verbose >= 6) {
(void) printf("%s grew from %lu to %lu bytes\n",
vd->vdev_path, (ulong_t)fsize, (ulong_t)*newsize);
}
(void) close(fd);
return (NULL);
}
/*
* Callback function which expands a given vdev by calling vdev_online().
*/
static vdev_t *
online_vdev(vdev_t *vd, void *arg)
{
(void) arg;
spa_t *spa = vd->vdev_spa;
vdev_t *tvd = vd->vdev_top;
uint64_t guid = vd->vdev_guid;
uint64_t generation = spa->spa_config_generation + 1;
vdev_state_t newstate = VDEV_STATE_UNKNOWN;
int error;
ASSERT3S(spa_config_held(spa, SCL_STATE, RW_READER), ==, SCL_STATE);
ASSERT(vd->vdev_ops->vdev_op_leaf);
/* Calling vdev_online will initialize the new metaslabs */
spa_config_exit(spa, SCL_STATE, spa);
error = vdev_online(spa, guid, ZFS_ONLINE_EXPAND, &newstate);
spa_config_enter(spa, SCL_STATE, spa, RW_READER);
/*
* If vdev_online returned an error or the underlying vdev_open
* failed then we abort the expand. The only way to know that
* vdev_open fails is by checking the returned newstate.
*/
if (error || newstate != VDEV_STATE_HEALTHY) {
if (ztest_opts.zo_verbose >= 5) {
(void) printf("Unable to expand vdev, state %u, "
"error %d\n", newstate, error);
}
return (vd);
}
ASSERT3U(newstate, ==, VDEV_STATE_HEALTHY);
/*
* Since we dropped the lock we need to ensure that we're
* still talking to the original vdev. It's possible this
* vdev may have been detached/replaced while we were
* trying to online it.
*/
if (generation != spa->spa_config_generation) {
if (ztest_opts.zo_verbose >= 5) {
(void) printf("vdev configuration has changed, "
"guid %"PRIu64", state %"PRIu64", "
"expected gen %"PRIu64", got gen %"PRIu64"\n",
guid,
tvd->vdev_state,
generation,
spa->spa_config_generation);
}
return (vd);
}
return (NULL);
}
/*
* Traverse the vdev tree calling the supplied function.
* We continue to walk the tree until we either have walked all
* children or we receive a non-NULL return from the callback.
* If a NULL callback is passed, then we just return back the first
* leaf vdev we encounter.
*/
static vdev_t *
vdev_walk_tree(vdev_t *vd, vdev_t *(*func)(vdev_t *, void *), void *arg)
{
uint_t c;
if (vd->vdev_ops->vdev_op_leaf) {
if (func == NULL)
return (vd);
else
return (func(vd, arg));
}
for (c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if ((cvd = vdev_walk_tree(cvd, func, arg)) != NULL)
return (cvd);
}
return (NULL);
}
/*
* Verify that dynamic LUN growth works as expected.
*/
void
ztest_vdev_LUN_growth(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
spa_t *spa = ztest_spa;
vdev_t *vd, *tvd;
metaslab_class_t *mc;
metaslab_group_t *mg;
size_t psize, newsize;
uint64_t top;
uint64_t old_class_space, new_class_space, old_ms_count, new_ms_count;
mutex_enter(&ztest_checkpoint_lock);
mutex_enter(&ztest_vdev_lock);
spa_config_enter(spa, SCL_STATE, spa, RW_READER);
/*
* If there is a vdev removal in progress, it could complete while
* we are running, in which case we would not be able to verify
* that the metaslab_class space increased (because it decreases
* when the device removal completes).
*/
if (ztest_device_removal_active) {
spa_config_exit(spa, SCL_STATE, spa);
mutex_exit(&ztest_vdev_lock);
mutex_exit(&ztest_checkpoint_lock);
return;
}
+ /*
+ * If we are under raidz expansion, the test can failed because the
+ * metaslabs count will not increase immediately after the vdev is
+ * expanded. It will happen only after raidz expansion completion.
+ */
+ if (spa->spa_raidz_expand) {
+ spa_config_exit(spa, SCL_STATE, spa);
+ mutex_exit(&ztest_vdev_lock);
+ mutex_exit(&ztest_checkpoint_lock);
+ return;
+ }
+
top = ztest_random_vdev_top(spa, B_TRUE);
tvd = spa->spa_root_vdev->vdev_child[top];
mg = tvd->vdev_mg;
mc = mg->mg_class;
old_ms_count = tvd->vdev_ms_count;
old_class_space = metaslab_class_get_space(mc);
/*
* Determine the size of the first leaf vdev associated with
* our top-level device.
*/
vd = vdev_walk_tree(tvd, NULL, NULL);
ASSERT3P(vd, !=, NULL);
ASSERT(vd->vdev_ops->vdev_op_leaf);
psize = vd->vdev_psize;
/*
* We only try to expand the vdev if it's healthy, less than 4x its
* original size, and it has a valid psize.
*/
if (tvd->vdev_state != VDEV_STATE_HEALTHY ||
psize == 0 || psize >= 4 * ztest_opts.zo_vdev_size) {
spa_config_exit(spa, SCL_STATE, spa);
mutex_exit(&ztest_vdev_lock);
mutex_exit(&ztest_checkpoint_lock);
return;
}
ASSERT3U(psize, >, 0);
newsize = psize + MAX(psize / 8, SPA_MAXBLOCKSIZE);
ASSERT3U(newsize, >, psize);
if (ztest_opts.zo_verbose >= 6) {
(void) printf("Expanding LUN %s from %lu to %lu\n",
vd->vdev_path, (ulong_t)psize, (ulong_t)newsize);
}
/*
* Growing the vdev is a two step process:
* 1). expand the physical size (i.e. relabel)
* 2). online the vdev to create the new metaslabs
*/
if (vdev_walk_tree(tvd, grow_vdev, &newsize) != NULL ||
vdev_walk_tree(tvd, online_vdev, NULL) != NULL ||
tvd->vdev_state != VDEV_STATE_HEALTHY) {
if (ztest_opts.zo_verbose >= 5) {
(void) printf("Could not expand LUN because "
"the vdev configuration changed.\n");
}
spa_config_exit(spa, SCL_STATE, spa);
mutex_exit(&ztest_vdev_lock);
mutex_exit(&ztest_checkpoint_lock);
return;
}
spa_config_exit(spa, SCL_STATE, spa);
/*
* Expanding the LUN will update the config asynchronously,
* thus we must wait for the async thread to complete any
* pending tasks before proceeding.
*/
for (;;) {
boolean_t done;
mutex_enter(&spa->spa_async_lock);
done = (spa->spa_async_thread == NULL && !spa->spa_async_tasks);
mutex_exit(&spa->spa_async_lock);
if (done)
break;
txg_wait_synced(spa_get_dsl(spa), 0);
(void) poll(NULL, 0, 100);
}
spa_config_enter(spa, SCL_STATE, spa, RW_READER);
tvd = spa->spa_root_vdev->vdev_child[top];
new_ms_count = tvd->vdev_ms_count;
new_class_space = metaslab_class_get_space(mc);
if (tvd->vdev_mg != mg || mg->mg_class != mc) {
if (ztest_opts.zo_verbose >= 5) {
(void) printf("Could not verify LUN expansion due to "
"intervening vdev offline or remove.\n");
}
spa_config_exit(spa, SCL_STATE, spa);
mutex_exit(&ztest_vdev_lock);
mutex_exit(&ztest_checkpoint_lock);
return;
}
/*
* Make sure we were able to grow the vdev.
*/
if (new_ms_count <= old_ms_count) {
fatal(B_FALSE,
"LUN expansion failed: ms_count %"PRIu64" < %"PRIu64"\n",
old_ms_count, new_ms_count);
}
/*
* Make sure we were able to grow the pool.
*/
if (new_class_space <= old_class_space) {
fatal(B_FALSE,
"LUN expansion failed: class_space %"PRIu64" < %"PRIu64"\n",
old_class_space, new_class_space);
}
if (ztest_opts.zo_verbose >= 5) {
char oldnumbuf[NN_NUMBUF_SZ], newnumbuf[NN_NUMBUF_SZ];
nicenum(old_class_space, oldnumbuf, sizeof (oldnumbuf));
nicenum(new_class_space, newnumbuf, sizeof (newnumbuf));
(void) printf("%s grew from %s to %s\n",
spa->spa_name, oldnumbuf, newnumbuf);
}
spa_config_exit(spa, SCL_STATE, spa);
mutex_exit(&ztest_vdev_lock);
mutex_exit(&ztest_checkpoint_lock);
}
/*
* Verify that dmu_objset_{create,destroy,open,close} work as expected.
*/
static void
ztest_objset_create_cb(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx)
{
(void) arg, (void) cr;
/*
* Create the objects common to all ztest datasets.
*/
VERIFY0(zap_create_claim(os, ZTEST_DIROBJ,
DMU_OT_ZAP_OTHER, DMU_OT_NONE, 0, tx));
}
static int
ztest_dataset_create(char *dsname)
{
int err;
uint64_t rand;
dsl_crypto_params_t *dcp = NULL;
/*
* 50% of the time, we create encrypted datasets
* using a random cipher suite and a hard-coded
* wrapping key.
*/
rand = ztest_random(2);
if (rand != 0) {
nvlist_t *crypto_args = fnvlist_alloc();
nvlist_t *props = fnvlist_alloc();
/* slight bias towards the default cipher suite */
rand = ztest_random(ZIO_CRYPT_FUNCTIONS);
if (rand < ZIO_CRYPT_AES_128_CCM)
rand = ZIO_CRYPT_ON;
fnvlist_add_uint64(props,
zfs_prop_to_name(ZFS_PROP_ENCRYPTION), rand);
fnvlist_add_uint8_array(crypto_args, "wkeydata",
(uint8_t *)ztest_wkeydata, WRAPPING_KEY_LEN);
/*
* These parameters aren't really used by the kernel. They
* are simply stored so that userspace knows how to load
* the wrapping key.
*/
fnvlist_add_uint64(props,
zfs_prop_to_name(ZFS_PROP_KEYFORMAT), ZFS_KEYFORMAT_RAW);
fnvlist_add_string(props,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION), "prompt");
fnvlist_add_uint64(props,
zfs_prop_to_name(ZFS_PROP_PBKDF2_SALT), 0ULL);
fnvlist_add_uint64(props,
zfs_prop_to_name(ZFS_PROP_PBKDF2_ITERS), 0ULL);
VERIFY0(dsl_crypto_params_create_nvlist(DCP_CMD_NONE, props,
crypto_args, &dcp));
/*
* Cycle through all available encryption implementations
* to verify interoperability.
*/
VERIFY0(gcm_impl_set("cycle"));
VERIFY0(aes_impl_set("cycle"));
fnvlist_free(crypto_args);
fnvlist_free(props);
}
err = dmu_objset_create(dsname, DMU_OST_OTHER, 0, dcp,
ztest_objset_create_cb, NULL);
dsl_crypto_params_free(dcp, !!err);
rand = ztest_random(100);
if (err || rand < 80)
return (err);
if (ztest_opts.zo_verbose >= 5)
(void) printf("Setting dataset %s to sync always\n", dsname);
return (ztest_dsl_prop_set_uint64(dsname, ZFS_PROP_SYNC,
ZFS_SYNC_ALWAYS, B_FALSE));
}
static int
ztest_objset_destroy_cb(const char *name, void *arg)
{
(void) arg;
objset_t *os;
dmu_object_info_t doi;
int error;
/*
* Verify that the dataset contains a directory object.
*/
VERIFY0(ztest_dmu_objset_own(name, DMU_OST_OTHER, B_TRUE,
B_TRUE, FTAG, &os));
error = dmu_object_info(os, ZTEST_DIROBJ, &doi);
if (error != ENOENT) {
/* We could have crashed in the middle of destroying it */
ASSERT0(error);
ASSERT3U(doi.doi_type, ==, DMU_OT_ZAP_OTHER);
ASSERT3S(doi.doi_physical_blocks_512, >=, 0);
}
dmu_objset_disown(os, B_TRUE, FTAG);
/*
* Destroy the dataset.
*/
if (strchr(name, '@') != NULL) {
error = dsl_destroy_snapshot(name, B_TRUE);
if (error != ECHRNG) {
/*
* The program was executed, but encountered a runtime
* error, such as insufficient slop, or a hold on the
* dataset.
*/
ASSERT0(error);
}
} else {
error = dsl_destroy_head(name);
if (error == ENOSPC) {
/* There could be checkpoint or insufficient slop */
ztest_record_enospc(FTAG);
} else if (error != EBUSY) {
/* There could be a hold on this dataset */
ASSERT0(error);
}
}
return (0);
}
static boolean_t
ztest_snapshot_create(char *osname, uint64_t id)
{
char snapname[ZFS_MAX_DATASET_NAME_LEN];
int error;
(void) snprintf(snapname, sizeof (snapname), "%"PRIu64"", id);
error = dmu_objset_snapshot_one(osname, snapname);
if (error == ENOSPC) {
ztest_record_enospc(FTAG);
return (B_FALSE);
}
if (error != 0 && error != EEXIST && error != ECHRNG) {
fatal(B_FALSE, "ztest_snapshot_create(%s@%s) = %d", osname,
snapname, error);
}
return (B_TRUE);
}
static boolean_t
ztest_snapshot_destroy(char *osname, uint64_t id)
{
char snapname[ZFS_MAX_DATASET_NAME_LEN];
int error;
(void) snprintf(snapname, sizeof (snapname), "%s@%"PRIu64"",
osname, id);
error = dsl_destroy_snapshot(snapname, B_FALSE);
if (error != 0 && error != ENOENT && error != ECHRNG)
fatal(B_FALSE, "ztest_snapshot_destroy(%s) = %d",
snapname, error);
return (B_TRUE);
}
void
ztest_dmu_objset_create_destroy(ztest_ds_t *zd, uint64_t id)
{
(void) zd;
ztest_ds_t *zdtmp;
int iters;
int error;
objset_t *os, *os2;
char name[ZFS_MAX_DATASET_NAME_LEN];
zilog_t *zilog;
int i;
zdtmp = umem_alloc(sizeof (ztest_ds_t), UMEM_NOFAIL);
(void) pthread_rwlock_rdlock(&ztest_name_lock);
(void) snprintf(name, sizeof (name), "%s/temp_%"PRIu64"",
ztest_opts.zo_pool, id);
/*
* If this dataset exists from a previous run, process its replay log
* half of the time. If we don't replay it, then dsl_destroy_head()
* (invoked from ztest_objset_destroy_cb()) should just throw it away.
*/
if (ztest_random(2) == 0 &&
ztest_dmu_objset_own(name, DMU_OST_OTHER, B_FALSE,
B_TRUE, FTAG, &os) == 0) {
ztest_zd_init(zdtmp, NULL, os);
zil_replay(os, zdtmp, ztest_replay_vector);
ztest_zd_fini(zdtmp);
dmu_objset_disown(os, B_TRUE, FTAG);
}
/*
* There may be an old instance of the dataset we're about to
* create lying around from a previous run. If so, destroy it
* and all of its snapshots.
*/
(void) dmu_objset_find(name, ztest_objset_destroy_cb, NULL,
DS_FIND_CHILDREN | DS_FIND_SNAPSHOTS);
/*
* Verify that the destroyed dataset is no longer in the namespace.
* It may still be present if the destroy above fails with ENOSPC.
*/
error = ztest_dmu_objset_own(name, DMU_OST_OTHER, B_TRUE, B_TRUE,
FTAG, &os);
if (error == 0) {
dmu_objset_disown(os, B_TRUE, FTAG);
ztest_record_enospc(FTAG);
goto out;
}
VERIFY3U(ENOENT, ==, error);
/*
* Verify that we can create a new dataset.
*/
error = ztest_dataset_create(name);
if (error) {
if (error == ENOSPC) {
ztest_record_enospc(FTAG);
goto out;
}
fatal(B_FALSE, "dmu_objset_create(%s) = %d", name, error);
}
VERIFY0(ztest_dmu_objset_own(name, DMU_OST_OTHER, B_FALSE, B_TRUE,
FTAG, &os));
ztest_zd_init(zdtmp, NULL, os);
/*
* Open the intent log for it.
*/
zilog = zil_open(os, ztest_get_data, NULL);
/*
* Put some objects in there, do a little I/O to them,
* and randomly take a couple of snapshots along the way.
*/
iters = ztest_random(5);
for (i = 0; i < iters; i++) {
ztest_dmu_object_alloc_free(zdtmp, id);
if (ztest_random(iters) == 0)
(void) ztest_snapshot_create(name, i);
}
/*
* Verify that we cannot create an existing dataset.
*/
VERIFY3U(EEXIST, ==,
dmu_objset_create(name, DMU_OST_OTHER, 0, NULL, NULL, NULL));
/*
* Verify that we can hold an objset that is also owned.
*/
VERIFY0(dmu_objset_hold(name, FTAG, &os2));
dmu_objset_rele(os2, FTAG);
/*
* Verify that we cannot own an objset that is already owned.
*/
VERIFY3U(EBUSY, ==, ztest_dmu_objset_own(name, DMU_OST_OTHER,
B_FALSE, B_TRUE, FTAG, &os2));
zil_close(zilog);
dmu_objset_disown(os, B_TRUE, FTAG);
ztest_zd_fini(zdtmp);
out:
(void) pthread_rwlock_unlock(&ztest_name_lock);
umem_free(zdtmp, sizeof (ztest_ds_t));
}
/*
* Verify that dmu_snapshot_{create,destroy,open,close} work as expected.
*/
void
ztest_dmu_snapshot_create_destroy(ztest_ds_t *zd, uint64_t id)
{
(void) pthread_rwlock_rdlock(&ztest_name_lock);
(void) ztest_snapshot_destroy(zd->zd_name, id);
(void) ztest_snapshot_create(zd->zd_name, id);
(void) pthread_rwlock_unlock(&ztest_name_lock);
}
/*
* Cleanup non-standard snapshots and clones.
*/
static void
ztest_dsl_dataset_cleanup(char *osname, uint64_t id)
{
char *snap1name;
char *clone1name;
char *snap2name;
char *clone2name;
char *snap3name;
int error;
snap1name = umem_alloc(ZFS_MAX_DATASET_NAME_LEN, UMEM_NOFAIL);
clone1name = umem_alloc(ZFS_MAX_DATASET_NAME_LEN, UMEM_NOFAIL);
snap2name = umem_alloc(ZFS_MAX_DATASET_NAME_LEN, UMEM_NOFAIL);
clone2name = umem_alloc(ZFS_MAX_DATASET_NAME_LEN, UMEM_NOFAIL);
snap3name = umem_alloc(ZFS_MAX_DATASET_NAME_LEN, UMEM_NOFAIL);
(void) snprintf(snap1name, ZFS_MAX_DATASET_NAME_LEN, "%s@s1_%"PRIu64"",
osname, id);
(void) snprintf(clone1name, ZFS_MAX_DATASET_NAME_LEN, "%s/c1_%"PRIu64"",
osname, id);
(void) snprintf(snap2name, ZFS_MAX_DATASET_NAME_LEN, "%s@s2_%"PRIu64"",
clone1name, id);
(void) snprintf(clone2name, ZFS_MAX_DATASET_NAME_LEN, "%s/c2_%"PRIu64"",
osname, id);
(void) snprintf(snap3name, ZFS_MAX_DATASET_NAME_LEN, "%s@s3_%"PRIu64"",
clone1name, id);
error = dsl_destroy_head(clone2name);
if (error && error != ENOENT)
fatal(B_FALSE, "dsl_destroy_head(%s) = %d", clone2name, error);
error = dsl_destroy_snapshot(snap3name, B_FALSE);
if (error && error != ENOENT)
fatal(B_FALSE, "dsl_destroy_snapshot(%s) = %d",
snap3name, error);
error = dsl_destroy_snapshot(snap2name, B_FALSE);
if (error && error != ENOENT)
fatal(B_FALSE, "dsl_destroy_snapshot(%s) = %d",
snap2name, error);
error = dsl_destroy_head(clone1name);
if (error && error != ENOENT)
fatal(B_FALSE, "dsl_destroy_head(%s) = %d", clone1name, error);
error = dsl_destroy_snapshot(snap1name, B_FALSE);
if (error && error != ENOENT)
fatal(B_FALSE, "dsl_destroy_snapshot(%s) = %d",
snap1name, error);
umem_free(snap1name, ZFS_MAX_DATASET_NAME_LEN);
umem_free(clone1name, ZFS_MAX_DATASET_NAME_LEN);
umem_free(snap2name, ZFS_MAX_DATASET_NAME_LEN);
umem_free(clone2name, ZFS_MAX_DATASET_NAME_LEN);
umem_free(snap3name, ZFS_MAX_DATASET_NAME_LEN);
}
/*
* Verify dsl_dataset_promote handles EBUSY
*/
void
ztest_dsl_dataset_promote_busy(ztest_ds_t *zd, uint64_t id)
{
objset_t *os;
char *snap1name;
char *clone1name;
char *snap2name;
char *clone2name;
char *snap3name;
char *osname = zd->zd_name;
int error;
snap1name = umem_alloc(ZFS_MAX_DATASET_NAME_LEN, UMEM_NOFAIL);
clone1name = umem_alloc(ZFS_MAX_DATASET_NAME_LEN, UMEM_NOFAIL);
snap2name = umem_alloc(ZFS_MAX_DATASET_NAME_LEN, UMEM_NOFAIL);
clone2name = umem_alloc(ZFS_MAX_DATASET_NAME_LEN, UMEM_NOFAIL);
snap3name = umem_alloc(ZFS_MAX_DATASET_NAME_LEN, UMEM_NOFAIL);
(void) pthread_rwlock_rdlock(&ztest_name_lock);
ztest_dsl_dataset_cleanup(osname, id);
(void) snprintf(snap1name, ZFS_MAX_DATASET_NAME_LEN, "%s@s1_%"PRIu64"",
osname, id);
(void) snprintf(clone1name, ZFS_MAX_DATASET_NAME_LEN, "%s/c1_%"PRIu64"",
osname, id);
(void) snprintf(snap2name, ZFS_MAX_DATASET_NAME_LEN, "%s@s2_%"PRIu64"",
clone1name, id);
(void) snprintf(clone2name, ZFS_MAX_DATASET_NAME_LEN, "%s/c2_%"PRIu64"",
osname, id);
(void) snprintf(snap3name, ZFS_MAX_DATASET_NAME_LEN, "%s@s3_%"PRIu64"",
clone1name, id);
error = dmu_objset_snapshot_one(osname, strchr(snap1name, '@') + 1);
if (error && error != EEXIST) {
if (error == ENOSPC) {
ztest_record_enospc(FTAG);
goto out;
}
fatal(B_FALSE, "dmu_take_snapshot(%s) = %d", snap1name, error);
}
error = dmu_objset_clone(clone1name, snap1name);
if (error) {
if (error == ENOSPC) {
ztest_record_enospc(FTAG);
goto out;
}
fatal(B_FALSE, "dmu_objset_create(%s) = %d", clone1name, error);
}
error = dmu_objset_snapshot_one(clone1name, strchr(snap2name, '@') + 1);
if (error && error != EEXIST) {
if (error == ENOSPC) {
ztest_record_enospc(FTAG);
goto out;
}
fatal(B_FALSE, "dmu_open_snapshot(%s) = %d", snap2name, error);
}
error = dmu_objset_snapshot_one(clone1name, strchr(snap3name, '@') + 1);
if (error && error != EEXIST) {
if (error == ENOSPC) {
ztest_record_enospc(FTAG);
goto out;
}
fatal(B_FALSE, "dmu_open_snapshot(%s) = %d", snap3name, error);
}
error = dmu_objset_clone(clone2name, snap3name);
if (error) {
if (error == ENOSPC) {
ztest_record_enospc(FTAG);
goto out;
}
fatal(B_FALSE, "dmu_objset_create(%s) = %d", clone2name, error);
}
error = ztest_dmu_objset_own(snap2name, DMU_OST_ANY, B_TRUE, B_TRUE,
FTAG, &os);
if (error)
fatal(B_FALSE, "dmu_objset_own(%s) = %d", snap2name, error);
error = dsl_dataset_promote(clone2name, NULL);
if (error == ENOSPC) {
dmu_objset_disown(os, B_TRUE, FTAG);
ztest_record_enospc(FTAG);
goto out;
}
if (error != EBUSY)
fatal(B_FALSE, "dsl_dataset_promote(%s), %d, not EBUSY",
clone2name, error);
dmu_objset_disown(os, B_TRUE, FTAG);
out:
ztest_dsl_dataset_cleanup(osname, id);
(void) pthread_rwlock_unlock(&ztest_name_lock);
umem_free(snap1name, ZFS_MAX_DATASET_NAME_LEN);
umem_free(clone1name, ZFS_MAX_DATASET_NAME_LEN);
umem_free(snap2name, ZFS_MAX_DATASET_NAME_LEN);
umem_free(clone2name, ZFS_MAX_DATASET_NAME_LEN);
umem_free(snap3name, ZFS_MAX_DATASET_NAME_LEN);
}
#undef OD_ARRAY_SIZE
#define OD_ARRAY_SIZE 4
/*
* Verify that dmu_object_{alloc,free} work as expected.
*/
void
ztest_dmu_object_alloc_free(ztest_ds_t *zd, uint64_t id)
{
ztest_od_t *od;
int batchsize;
int size;
int b;
size = sizeof (ztest_od_t) * OD_ARRAY_SIZE;
od = umem_alloc(size, UMEM_NOFAIL);
batchsize = OD_ARRAY_SIZE;
for (b = 0; b < batchsize; b++)
ztest_od_init(od + b, id, FTAG, b, DMU_OT_UINT64_OTHER,
0, 0, 0);
/*
* Destroy the previous batch of objects, create a new batch,
* and do some I/O on the new objects.
*/
if (ztest_object_init(zd, od, size, B_TRUE) != 0) {
zd->zd_od = NULL;
umem_free(od, size);
return;
}
while (ztest_random(4 * batchsize) != 0)
ztest_io(zd, od[ztest_random(batchsize)].od_object,
ztest_random(ZTEST_RANGE_LOCKS) << SPA_MAXBLOCKSHIFT);
umem_free(od, size);
}
/*
* Rewind the global allocator to verify object allocation backfilling.
*/
void
ztest_dmu_object_next_chunk(ztest_ds_t *zd, uint64_t id)
{
(void) id;
objset_t *os = zd->zd_os;
uint_t dnodes_per_chunk = 1 << dmu_object_alloc_chunk_shift;
uint64_t object;
/*
* Rewind the global allocator randomly back to a lower object number
* to force backfilling and reclamation of recently freed dnodes.
*/
mutex_enter(&os->os_obj_lock);
object = ztest_random(os->os_obj_next_chunk);
os->os_obj_next_chunk = P2ALIGN(object, dnodes_per_chunk);
mutex_exit(&os->os_obj_lock);
}
#undef OD_ARRAY_SIZE
#define OD_ARRAY_SIZE 2
/*
* Verify that dmu_{read,write} work as expected.
*/
void
ztest_dmu_read_write(ztest_ds_t *zd, uint64_t id)
{
int size;
ztest_od_t *od;
objset_t *os = zd->zd_os;
size = sizeof (ztest_od_t) * OD_ARRAY_SIZE;
od = umem_alloc(size, UMEM_NOFAIL);
dmu_tx_t *tx;
int freeit, error;
uint64_t i, n, s, txg;
bufwad_t *packbuf, *bigbuf, *pack, *bigH, *bigT;
uint64_t packobj, packoff, packsize, bigobj, bigoff, bigsize;
uint64_t chunksize = (1000 + ztest_random(1000)) * sizeof (uint64_t);
uint64_t regions = 997;
uint64_t stride = 123456789ULL;
uint64_t width = 40;
int free_percent = 5;
/*
* This test uses two objects, packobj and bigobj, that are always
* updated together (i.e. in the same tx) so that their contents are
* in sync and can be compared. Their contents relate to each other
* in a simple way: packobj is a dense array of 'bufwad' structures,
* while bigobj is a sparse array of the same bufwads. Specifically,
* for any index n, there are three bufwads that should be identical:
*
* packobj, at offset n * sizeof (bufwad_t)
* bigobj, at the head of the nth chunk
* bigobj, at the tail of the nth chunk
*
* The chunk size is arbitrary. It doesn't have to be a power of two,
* and it doesn't have any relation to the object blocksize.
* The only requirement is that it can hold at least two bufwads.
*
* Normally, we write the bufwad to each of these locations.
* However, free_percent of the time we instead write zeroes to
* packobj and perform a dmu_free_range() on bigobj. By comparing
* bigobj to packobj, we can verify that the DMU is correctly
* tracking which parts of an object are allocated and free,
* and that the contents of the allocated blocks are correct.
*/
/*
* Read the directory info. If it's the first time, set things up.
*/
ztest_od_init(od, id, FTAG, 0, DMU_OT_UINT64_OTHER, 0, 0, chunksize);
ztest_od_init(od + 1, id, FTAG, 1, DMU_OT_UINT64_OTHER, 0, 0,
chunksize);
if (ztest_object_init(zd, od, size, B_FALSE) != 0) {
umem_free(od, size);
return;
}
bigobj = od[0].od_object;
packobj = od[1].od_object;
chunksize = od[0].od_gen;
ASSERT3U(chunksize, ==, od[1].od_gen);
/*
* Prefetch a random chunk of the big object.
* Our aim here is to get some async reads in flight
* for blocks that we may free below; the DMU should
* handle this race correctly.
*/
n = ztest_random(regions) * stride + ztest_random(width);
s = 1 + ztest_random(2 * width - 1);
dmu_prefetch(os, bigobj, 0, n * chunksize, s * chunksize,
ZIO_PRIORITY_SYNC_READ);
/*
* Pick a random index and compute the offsets into packobj and bigobj.
*/
n = ztest_random(regions) * stride + ztest_random(width);
s = 1 + ztest_random(width - 1);
packoff = n * sizeof (bufwad_t);
packsize = s * sizeof (bufwad_t);
bigoff = n * chunksize;
bigsize = s * chunksize;
packbuf = umem_alloc(packsize, UMEM_NOFAIL);
bigbuf = umem_alloc(bigsize, UMEM_NOFAIL);
/*
* free_percent of the time, free a range of bigobj rather than
* overwriting it.
*/
freeit = (ztest_random(100) < free_percent);
/*
* Read the current contents of our objects.
*/
error = dmu_read(os, packobj, packoff, packsize, packbuf,
DMU_READ_PREFETCH);
ASSERT0(error);
error = dmu_read(os, bigobj, bigoff, bigsize, bigbuf,
DMU_READ_PREFETCH);
ASSERT0(error);
/*
* Get a tx for the mods to both packobj and bigobj.
*/
tx = dmu_tx_create(os);
dmu_tx_hold_write(tx, packobj, packoff, packsize);
if (freeit)
dmu_tx_hold_free(tx, bigobj, bigoff, bigsize);
else
dmu_tx_hold_write(tx, bigobj, bigoff, bigsize);
/* This accounts for setting the checksum/compression. */
dmu_tx_hold_bonus(tx, bigobj);
txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG);
if (txg == 0) {
umem_free(packbuf, packsize);
umem_free(bigbuf, bigsize);
umem_free(od, size);
return;
}
enum zio_checksum cksum;
do {
cksum = (enum zio_checksum)
ztest_random_dsl_prop(ZFS_PROP_CHECKSUM);
} while (cksum >= ZIO_CHECKSUM_LEGACY_FUNCTIONS);
dmu_object_set_checksum(os, bigobj, cksum, tx);
enum zio_compress comp;
do {
comp = (enum zio_compress)
ztest_random_dsl_prop(ZFS_PROP_COMPRESSION);
} while (comp >= ZIO_COMPRESS_LEGACY_FUNCTIONS);
dmu_object_set_compress(os, bigobj, comp, tx);
/*
* For each index from n to n + s, verify that the existing bufwad
* in packobj matches the bufwads at the head and tail of the
* corresponding chunk in bigobj. Then update all three bufwads
* with the new values we want to write out.
*/
for (i = 0; i < s; i++) {
/* LINTED */
pack = (bufwad_t *)((char *)packbuf + i * sizeof (bufwad_t));
/* LINTED */
bigH = (bufwad_t *)((char *)bigbuf + i * chunksize);
/* LINTED */
bigT = (bufwad_t *)((char *)bigH + chunksize) - 1;
ASSERT3U((uintptr_t)bigH - (uintptr_t)bigbuf, <, bigsize);
ASSERT3U((uintptr_t)bigT - (uintptr_t)bigbuf, <, bigsize);
if (pack->bw_txg > txg)
fatal(B_FALSE,
"future leak: got %"PRIx64", open txg is %"PRIx64"",
pack->bw_txg, txg);
if (pack->bw_data != 0 && pack->bw_index != n + i)
fatal(B_FALSE, "wrong index: "
"got %"PRIx64", wanted %"PRIx64"+%"PRIx64"",
pack->bw_index, n, i);
if (memcmp(pack, bigH, sizeof (bufwad_t)) != 0)
fatal(B_FALSE, "pack/bigH mismatch in %p/%p",
pack, bigH);
if (memcmp(pack, bigT, sizeof (bufwad_t)) != 0)
fatal(B_FALSE, "pack/bigT mismatch in %p/%p",
pack, bigT);
if (freeit) {
memset(pack, 0, sizeof (bufwad_t));
} else {
pack->bw_index = n + i;
pack->bw_txg = txg;
pack->bw_data = 1 + ztest_random(-2ULL);
}
*bigH = *pack;
*bigT = *pack;
}
/*
* We've verified all the old bufwads, and made new ones.
* Now write them out.
*/
dmu_write(os, packobj, packoff, packsize, packbuf, tx);
if (freeit) {
if (ztest_opts.zo_verbose >= 7) {
(void) printf("freeing offset %"PRIx64" size %"PRIx64""
" txg %"PRIx64"\n",
bigoff, bigsize, txg);
}
VERIFY0(dmu_free_range(os, bigobj, bigoff, bigsize, tx));
} else {
if (ztest_opts.zo_verbose >= 7) {
(void) printf("writing offset %"PRIx64" size %"PRIx64""
" txg %"PRIx64"\n",
bigoff, bigsize, txg);
}
dmu_write(os, bigobj, bigoff, bigsize, bigbuf, tx);
}
dmu_tx_commit(tx);
/*
* Sanity check the stuff we just wrote.
*/
{
void *packcheck = umem_alloc(packsize, UMEM_NOFAIL);
void *bigcheck = umem_alloc(bigsize, UMEM_NOFAIL);
VERIFY0(dmu_read(os, packobj, packoff,
packsize, packcheck, DMU_READ_PREFETCH));
VERIFY0(dmu_read(os, bigobj, bigoff,
bigsize, bigcheck, DMU_READ_PREFETCH));
ASSERT0(memcmp(packbuf, packcheck, packsize));
ASSERT0(memcmp(bigbuf, bigcheck, bigsize));
umem_free(packcheck, packsize);
umem_free(bigcheck, bigsize);
}
umem_free(packbuf, packsize);
umem_free(bigbuf, bigsize);
umem_free(od, size);
}
static void
compare_and_update_pbbufs(uint64_t s, bufwad_t *packbuf, bufwad_t *bigbuf,
uint64_t bigsize, uint64_t n, uint64_t chunksize, uint64_t txg)
{
uint64_t i;
bufwad_t *pack;
bufwad_t *bigH;
bufwad_t *bigT;
/*
* For each index from n to n + s, verify that the existing bufwad
* in packobj matches the bufwads at the head and tail of the
* corresponding chunk in bigobj. Then update all three bufwads
* with the new values we want to write out.
*/
for (i = 0; i < s; i++) {
/* LINTED */
pack = (bufwad_t *)((char *)packbuf + i * sizeof (bufwad_t));
/* LINTED */
bigH = (bufwad_t *)((char *)bigbuf + i * chunksize);
/* LINTED */
bigT = (bufwad_t *)((char *)bigH + chunksize) - 1;
ASSERT3U((uintptr_t)bigH - (uintptr_t)bigbuf, <, bigsize);
ASSERT3U((uintptr_t)bigT - (uintptr_t)bigbuf, <, bigsize);
if (pack->bw_txg > txg)
fatal(B_FALSE,
"future leak: got %"PRIx64", open txg is %"PRIx64"",
pack->bw_txg, txg);
if (pack->bw_data != 0 && pack->bw_index != n + i)
fatal(B_FALSE, "wrong index: "
"got %"PRIx64", wanted %"PRIx64"+%"PRIx64"",
pack->bw_index, n, i);
if (memcmp(pack, bigH, sizeof (bufwad_t)) != 0)
fatal(B_FALSE, "pack/bigH mismatch in %p/%p",
pack, bigH);
if (memcmp(pack, bigT, sizeof (bufwad_t)) != 0)
fatal(B_FALSE, "pack/bigT mismatch in %p/%p",
pack, bigT);
pack->bw_index = n + i;
pack->bw_txg = txg;
pack->bw_data = 1 + ztest_random(-2ULL);
*bigH = *pack;
*bigT = *pack;
}
}
#undef OD_ARRAY_SIZE
#define OD_ARRAY_SIZE 2
void
ztest_dmu_read_write_zcopy(ztest_ds_t *zd, uint64_t id)
{
objset_t *os = zd->zd_os;
ztest_od_t *od;
dmu_tx_t *tx;
uint64_t i;
int error;
int size;
uint64_t n, s, txg;
bufwad_t *packbuf, *bigbuf;
uint64_t packobj, packoff, packsize, bigobj, bigoff, bigsize;
uint64_t blocksize = ztest_random_blocksize();
uint64_t chunksize = blocksize;
uint64_t regions = 997;
uint64_t stride = 123456789ULL;
uint64_t width = 9;
dmu_buf_t *bonus_db;
arc_buf_t **bigbuf_arcbufs;
dmu_object_info_t doi;
size = sizeof (ztest_od_t) * OD_ARRAY_SIZE;
od = umem_alloc(size, UMEM_NOFAIL);
/*
* This test uses two objects, packobj and bigobj, that are always
* updated together (i.e. in the same tx) so that their contents are
* in sync and can be compared. Their contents relate to each other
* in a simple way: packobj is a dense array of 'bufwad' structures,
* while bigobj is a sparse array of the same bufwads. Specifically,
* for any index n, there are three bufwads that should be identical:
*
* packobj, at offset n * sizeof (bufwad_t)
* bigobj, at the head of the nth chunk
* bigobj, at the tail of the nth chunk
*
* The chunk size is set equal to bigobj block size so that
* dmu_assign_arcbuf_by_dbuf() can be tested for object updates.
*/
/*
* Read the directory info. If it's the first time, set things up.
*/
ztest_od_init(od, id, FTAG, 0, DMU_OT_UINT64_OTHER, blocksize, 0, 0);
ztest_od_init(od + 1, id, FTAG, 1, DMU_OT_UINT64_OTHER, 0, 0,
chunksize);
if (ztest_object_init(zd, od, size, B_FALSE) != 0) {
umem_free(od, size);
return;
}
bigobj = od[0].od_object;
packobj = od[1].od_object;
blocksize = od[0].od_blocksize;
chunksize = blocksize;
ASSERT3U(chunksize, ==, od[1].od_gen);
VERIFY0(dmu_object_info(os, bigobj, &doi));
VERIFY(ISP2(doi.doi_data_block_size));
VERIFY3U(chunksize, ==, doi.doi_data_block_size);
VERIFY3U(chunksize, >=, 2 * sizeof (bufwad_t));
/*
* Pick a random index and compute the offsets into packobj and bigobj.
*/
n = ztest_random(regions) * stride + ztest_random(width);
s = 1 + ztest_random(width - 1);
packoff = n * sizeof (bufwad_t);
packsize = s * sizeof (bufwad_t);
bigoff = n * chunksize;
bigsize = s * chunksize;
packbuf = umem_zalloc(packsize, UMEM_NOFAIL);
bigbuf = umem_zalloc(bigsize, UMEM_NOFAIL);
VERIFY0(dmu_bonus_hold(os, bigobj, FTAG, &bonus_db));
bigbuf_arcbufs = umem_zalloc(2 * s * sizeof (arc_buf_t *), UMEM_NOFAIL);
/*
* Iteration 0 test zcopy for DB_UNCACHED dbufs.
* Iteration 1 test zcopy to already referenced dbufs.
* Iteration 2 test zcopy to dirty dbuf in the same txg.
* Iteration 3 test zcopy to dbuf dirty in previous txg.
* Iteration 4 test zcopy when dbuf is no longer dirty.
* Iteration 5 test zcopy when it can't be done.
* Iteration 6 one more zcopy write.
*/
for (i = 0; i < 7; i++) {
uint64_t j;
uint64_t off;
/*
* In iteration 5 (i == 5) use arcbufs
* that don't match bigobj blksz to test
* dmu_assign_arcbuf_by_dbuf() when it can't directly
* assign an arcbuf to a dbuf.
*/
for (j = 0; j < s; j++) {
if (i != 5 || chunksize < (SPA_MINBLOCKSIZE * 2)) {
bigbuf_arcbufs[j] =
dmu_request_arcbuf(bonus_db, chunksize);
} else {
bigbuf_arcbufs[2 * j] =
dmu_request_arcbuf(bonus_db, chunksize / 2);
bigbuf_arcbufs[2 * j + 1] =
dmu_request_arcbuf(bonus_db, chunksize / 2);
}
}
/*
* Get a tx for the mods to both packobj and bigobj.
*/
tx = dmu_tx_create(os);
dmu_tx_hold_write(tx, packobj, packoff, packsize);
dmu_tx_hold_write(tx, bigobj, bigoff, bigsize);
txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG);
if (txg == 0) {
umem_free(packbuf, packsize);
umem_free(bigbuf, bigsize);
for (j = 0; j < s; j++) {
if (i != 5 ||
chunksize < (SPA_MINBLOCKSIZE * 2)) {
dmu_return_arcbuf(bigbuf_arcbufs[j]);
} else {
dmu_return_arcbuf(
bigbuf_arcbufs[2 * j]);
dmu_return_arcbuf(
bigbuf_arcbufs[2 * j + 1]);
}
}
umem_free(bigbuf_arcbufs, 2 * s * sizeof (arc_buf_t *));
umem_free(od, size);
dmu_buf_rele(bonus_db, FTAG);
return;
}
/*
* 50% of the time don't read objects in the 1st iteration to
* test dmu_assign_arcbuf_by_dbuf() for the case when there are
* no existing dbufs for the specified offsets.
*/
if (i != 0 || ztest_random(2) != 0) {
error = dmu_read(os, packobj, packoff,
packsize, packbuf, DMU_READ_PREFETCH);
ASSERT0(error);
error = dmu_read(os, bigobj, bigoff, bigsize,
bigbuf, DMU_READ_PREFETCH);
ASSERT0(error);
}
compare_and_update_pbbufs(s, packbuf, bigbuf, bigsize,
n, chunksize, txg);
/*
* We've verified all the old bufwads, and made new ones.
* Now write them out.
*/
dmu_write(os, packobj, packoff, packsize, packbuf, tx);
if (ztest_opts.zo_verbose >= 7) {
(void) printf("writing offset %"PRIx64" size %"PRIx64""
" txg %"PRIx64"\n",
bigoff, bigsize, txg);
}
for (off = bigoff, j = 0; j < s; j++, off += chunksize) {
dmu_buf_t *dbt;
if (i != 5 || chunksize < (SPA_MINBLOCKSIZE * 2)) {
memcpy(bigbuf_arcbufs[j]->b_data,
(caddr_t)bigbuf + (off - bigoff),
chunksize);
} else {
memcpy(bigbuf_arcbufs[2 * j]->b_data,
(caddr_t)bigbuf + (off - bigoff),
chunksize / 2);
memcpy(bigbuf_arcbufs[2 * j + 1]->b_data,
(caddr_t)bigbuf + (off - bigoff) +
chunksize / 2,
chunksize / 2);
}
if (i == 1) {
VERIFY(dmu_buf_hold(os, bigobj, off,
FTAG, &dbt, DMU_READ_NO_PREFETCH) == 0);
}
if (i != 5 || chunksize < (SPA_MINBLOCKSIZE * 2)) {
VERIFY0(dmu_assign_arcbuf_by_dbuf(bonus_db,
off, bigbuf_arcbufs[j], tx));
} else {
VERIFY0(dmu_assign_arcbuf_by_dbuf(bonus_db,
off, bigbuf_arcbufs[2 * j], tx));
VERIFY0(dmu_assign_arcbuf_by_dbuf(bonus_db,
off + chunksize / 2,
bigbuf_arcbufs[2 * j + 1], tx));
}
if (i == 1) {
dmu_buf_rele(dbt, FTAG);
}
}
dmu_tx_commit(tx);
/*
* Sanity check the stuff we just wrote.
*/
{
void *packcheck = umem_alloc(packsize, UMEM_NOFAIL);
void *bigcheck = umem_alloc(bigsize, UMEM_NOFAIL);
VERIFY0(dmu_read(os, packobj, packoff,
packsize, packcheck, DMU_READ_PREFETCH));
VERIFY0(dmu_read(os, bigobj, bigoff,
bigsize, bigcheck, DMU_READ_PREFETCH));
ASSERT0(memcmp(packbuf, packcheck, packsize));
ASSERT0(memcmp(bigbuf, bigcheck, bigsize));
umem_free(packcheck, packsize);
umem_free(bigcheck, bigsize);
}
if (i == 2) {
txg_wait_open(dmu_objset_pool(os), 0, B_TRUE);
} else if (i == 3) {
txg_wait_synced(dmu_objset_pool(os), 0);
}
}
dmu_buf_rele(bonus_db, FTAG);
umem_free(packbuf, packsize);
umem_free(bigbuf, bigsize);
umem_free(bigbuf_arcbufs, 2 * s * sizeof (arc_buf_t *));
umem_free(od, size);
}
void
ztest_dmu_write_parallel(ztest_ds_t *zd, uint64_t id)
{
(void) id;
ztest_od_t *od;
od = umem_alloc(sizeof (ztest_od_t), UMEM_NOFAIL);
uint64_t offset = (1ULL << (ztest_random(20) + 43)) +
(ztest_random(ZTEST_RANGE_LOCKS) << SPA_MAXBLOCKSHIFT);
/*
* Have multiple threads write to large offsets in an object
* to verify that parallel writes to an object -- even to the
* same blocks within the object -- doesn't cause any trouble.
*/
ztest_od_init(od, ID_PARALLEL, FTAG, 0, DMU_OT_UINT64_OTHER, 0, 0, 0);
if (ztest_object_init(zd, od, sizeof (ztest_od_t), B_FALSE) != 0)
return;
while (ztest_random(10) != 0)
ztest_io(zd, od->od_object, offset);
umem_free(od, sizeof (ztest_od_t));
}
void
ztest_dmu_prealloc(ztest_ds_t *zd, uint64_t id)
{
ztest_od_t *od;
uint64_t offset = (1ULL << (ztest_random(4) + SPA_MAXBLOCKSHIFT)) +
(ztest_random(ZTEST_RANGE_LOCKS) << SPA_MAXBLOCKSHIFT);
uint64_t count = ztest_random(20) + 1;
uint64_t blocksize = ztest_random_blocksize();
void *data;
od = umem_alloc(sizeof (ztest_od_t), UMEM_NOFAIL);
ztest_od_init(od, id, FTAG, 0, DMU_OT_UINT64_OTHER, blocksize, 0, 0);
if (ztest_object_init(zd, od, sizeof (ztest_od_t),
!ztest_random(2)) != 0) {
umem_free(od, sizeof (ztest_od_t));
return;
}
if (ztest_truncate(zd, od->od_object, offset, count * blocksize) != 0) {
umem_free(od, sizeof (ztest_od_t));
return;
}
ztest_prealloc(zd, od->od_object, offset, count * blocksize);
data = umem_zalloc(blocksize, UMEM_NOFAIL);
while (ztest_random(count) != 0) {
uint64_t randoff = offset + (ztest_random(count) * blocksize);
if (ztest_write(zd, od->od_object, randoff, blocksize,
data) != 0)
break;
while (ztest_random(4) != 0)
ztest_io(zd, od->od_object, randoff);
}
umem_free(data, blocksize);
umem_free(od, sizeof (ztest_od_t));
}
/*
* Verify that zap_{create,destroy,add,remove,update} work as expected.
*/
#define ZTEST_ZAP_MIN_INTS 1
#define ZTEST_ZAP_MAX_INTS 4
#define ZTEST_ZAP_MAX_PROPS 1000
void
ztest_zap(ztest_ds_t *zd, uint64_t id)
{
objset_t *os = zd->zd_os;
ztest_od_t *od;
uint64_t object;
uint64_t txg, last_txg;
uint64_t value[ZTEST_ZAP_MAX_INTS];
uint64_t zl_ints, zl_intsize, prop;
int i, ints;
dmu_tx_t *tx;
char propname[100], txgname[100];
int error;
const char *const hc[2] = { "s.acl.h", ".s.open.h.hyLZlg" };
od = umem_alloc(sizeof (ztest_od_t), UMEM_NOFAIL);
ztest_od_init(od, id, FTAG, 0, DMU_OT_ZAP_OTHER, 0, 0, 0);
if (ztest_object_init(zd, od, sizeof (ztest_od_t),
!ztest_random(2)) != 0)
goto out;
object = od->od_object;
/*
* Generate a known hash collision, and verify that
* we can lookup and remove both entries.
*/
tx = dmu_tx_create(os);
dmu_tx_hold_zap(tx, object, B_TRUE, NULL);
txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG);
if (txg == 0)
goto out;
for (i = 0; i < 2; i++) {
value[i] = i;
VERIFY0(zap_add(os, object, hc[i], sizeof (uint64_t),
1, &value[i], tx));
}
for (i = 0; i < 2; i++) {
VERIFY3U(EEXIST, ==, zap_add(os, object, hc[i],
sizeof (uint64_t), 1, &value[i], tx));
VERIFY0(
zap_length(os, object, hc[i], &zl_intsize, &zl_ints));
ASSERT3U(zl_intsize, ==, sizeof (uint64_t));
ASSERT3U(zl_ints, ==, 1);
}
for (i = 0; i < 2; i++) {
VERIFY0(zap_remove(os, object, hc[i], tx));
}
dmu_tx_commit(tx);
/*
* Generate a bunch of random entries.
*/
ints = MAX(ZTEST_ZAP_MIN_INTS, object % ZTEST_ZAP_MAX_INTS);
prop = ztest_random(ZTEST_ZAP_MAX_PROPS);
(void) sprintf(propname, "prop_%"PRIu64"", prop);
(void) sprintf(txgname, "txg_%"PRIu64"", prop);
memset(value, 0, sizeof (value));
last_txg = 0;
/*
* If these zap entries already exist, validate their contents.
*/
error = zap_length(os, object, txgname, &zl_intsize, &zl_ints);
if (error == 0) {
ASSERT3U(zl_intsize, ==, sizeof (uint64_t));
ASSERT3U(zl_ints, ==, 1);
VERIFY0(zap_lookup(os, object, txgname, zl_intsize,
zl_ints, &last_txg));
VERIFY0(zap_length(os, object, propname, &zl_intsize,
&zl_ints));
ASSERT3U(zl_intsize, ==, sizeof (uint64_t));
ASSERT3U(zl_ints, ==, ints);
VERIFY0(zap_lookup(os, object, propname, zl_intsize,
zl_ints, value));
for (i = 0; i < ints; i++) {
ASSERT3U(value[i], ==, last_txg + object + i);
}
} else {
ASSERT3U(error, ==, ENOENT);
}
/*
* Atomically update two entries in our zap object.
* The first is named txg_%llu, and contains the txg
* in which the property was last updated. The second
* is named prop_%llu, and the nth element of its value
* should be txg + object + n.
*/
tx = dmu_tx_create(os);
dmu_tx_hold_zap(tx, object, B_TRUE, NULL);
txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG);
if (txg == 0)
goto out;
if (last_txg > txg)
fatal(B_FALSE, "zap future leak: old %"PRIu64" new %"PRIu64"",
last_txg, txg);
for (i = 0; i < ints; i++)
value[i] = txg + object + i;
VERIFY0(zap_update(os, object, txgname, sizeof (uint64_t),
1, &txg, tx));
VERIFY0(zap_update(os, object, propname, sizeof (uint64_t),
ints, value, tx));
dmu_tx_commit(tx);
/*
* Remove a random pair of entries.
*/
prop = ztest_random(ZTEST_ZAP_MAX_PROPS);
(void) sprintf(propname, "prop_%"PRIu64"", prop);
(void) sprintf(txgname, "txg_%"PRIu64"", prop);
error = zap_length(os, object, txgname, &zl_intsize, &zl_ints);
if (error == ENOENT)
goto out;
ASSERT0(error);
tx = dmu_tx_create(os);
dmu_tx_hold_zap(tx, object, B_TRUE, NULL);
txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG);
if (txg == 0)
goto out;
VERIFY0(zap_remove(os, object, txgname, tx));
VERIFY0(zap_remove(os, object, propname, tx));
dmu_tx_commit(tx);
out:
umem_free(od, sizeof (ztest_od_t));
}
/*
* Test case to test the upgrading of a microzap to fatzap.
*/
void
ztest_fzap(ztest_ds_t *zd, uint64_t id)
{
objset_t *os = zd->zd_os;
ztest_od_t *od;
uint64_t object, txg, value;
od = umem_alloc(sizeof (ztest_od_t), UMEM_NOFAIL);
ztest_od_init(od, id, FTAG, 0, DMU_OT_ZAP_OTHER, 0, 0, 0);
if (ztest_object_init(zd, od, sizeof (ztest_od_t),
!ztest_random(2)) != 0)
goto out;
object = od->od_object;
/*
* Add entries to this ZAP and make sure it spills over
* and gets upgraded to a fatzap. Also, since we are adding
* 2050 entries we should see ptrtbl growth and leaf-block split.
*/
for (value = 0; value < 2050; value++) {
char name[ZFS_MAX_DATASET_NAME_LEN];
dmu_tx_t *tx;
int error;
(void) snprintf(name, sizeof (name), "fzap-%"PRIu64"-%"PRIu64"",
id, value);
tx = dmu_tx_create(os);
dmu_tx_hold_zap(tx, object, B_TRUE, name);
txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG);
if (txg == 0)
goto out;
error = zap_add(os, object, name, sizeof (uint64_t), 1,
&value, tx);
ASSERT(error == 0 || error == EEXIST);
dmu_tx_commit(tx);
}
out:
umem_free(od, sizeof (ztest_od_t));
}
void
ztest_zap_parallel(ztest_ds_t *zd, uint64_t id)
{
(void) id;
objset_t *os = zd->zd_os;
ztest_od_t *od;
uint64_t txg, object, count, wsize, wc, zl_wsize, zl_wc;
dmu_tx_t *tx;
int i, namelen, error;
int micro = ztest_random(2);
char name[20], string_value[20];
void *data;
od = umem_alloc(sizeof (ztest_od_t), UMEM_NOFAIL);
ztest_od_init(od, ID_PARALLEL, FTAG, micro, DMU_OT_ZAP_OTHER, 0, 0, 0);
if (ztest_object_init(zd, od, sizeof (ztest_od_t), B_FALSE) != 0) {
umem_free(od, sizeof (ztest_od_t));
return;
}
object = od->od_object;
/*
* Generate a random name of the form 'xxx.....' where each
* x is a random printable character and the dots are dots.
* There are 94 such characters, and the name length goes from
* 6 to 20, so there are 94^3 * 15 = 12,458,760 possible names.
*/
namelen = ztest_random(sizeof (name) - 5) + 5 + 1;
for (i = 0; i < 3; i++)
name[i] = '!' + ztest_random('~' - '!' + 1);
for (; i < namelen - 1; i++)
name[i] = '.';
name[i] = '\0';
if ((namelen & 1) || micro) {
wsize = sizeof (txg);
wc = 1;
data = &txg;
} else {
wsize = 1;
wc = namelen;
data = string_value;
}
count = -1ULL;
VERIFY0(zap_count(os, object, &count));
ASSERT3S(count, !=, -1ULL);
/*
* Select an operation: length, lookup, add, update, remove.
*/
i = ztest_random(5);
if (i >= 2) {
tx = dmu_tx_create(os);
dmu_tx_hold_zap(tx, object, B_TRUE, NULL);
txg = ztest_tx_assign(tx, TXG_MIGHTWAIT, FTAG);
if (txg == 0) {
umem_free(od, sizeof (ztest_od_t));
return;
}
memcpy(string_value, name, namelen);
} else {
tx = NULL;
txg = 0;
memset(string_value, 0, namelen);
}
switch (i) {
case 0:
error = zap_length(os, object, name, &zl_wsize, &zl_wc);
if (error == 0) {
ASSERT3U(wsize, ==, zl_wsize);
ASSERT3U(wc, ==, zl_wc);
} else {
ASSERT3U(error, ==, ENOENT);
}
break;
case 1:
error = zap_lookup(os, object, name, wsize, wc, data);
if (error == 0) {
if (data == string_value &&
memcmp(name, data, namelen) != 0)
fatal(B_FALSE, "name '%s' != val '%s' len %d",
name, (char *)data, namelen);
} else {
ASSERT3U(error, ==, ENOENT);
}
break;
case 2:
error = zap_add(os, object, name, wsize, wc, data, tx);
ASSERT(error == 0 || error == EEXIST);
break;
case 3:
VERIFY0(zap_update(os, object, name, wsize, wc, data, tx));
break;
case 4:
error = zap_remove(os, object, name, tx);
ASSERT(error == 0 || error == ENOENT);
break;
}
if (tx != NULL)
dmu_tx_commit(tx);
umem_free(od, sizeof (ztest_od_t));
}
/*
* Commit callback data.
*/
typedef struct ztest_cb_data {
list_node_t zcd_node;
uint64_t zcd_txg;
int zcd_expected_err;
boolean_t zcd_added;
boolean_t zcd_called;
spa_t *zcd_spa;
} ztest_cb_data_t;
/* This is the actual commit callback function */
static void
ztest_commit_callback(void *arg, int error)
{
ztest_cb_data_t *data = arg;
uint64_t synced_txg;
VERIFY3P(data, !=, NULL);
VERIFY3S(data->zcd_expected_err, ==, error);
VERIFY(!data->zcd_called);
synced_txg = spa_last_synced_txg(data->zcd_spa);
if (data->zcd_txg > synced_txg)
fatal(B_FALSE,
"commit callback of txg %"PRIu64" called prematurely, "
"last synced txg = %"PRIu64"\n",
data->zcd_txg, synced_txg);
data->zcd_called = B_TRUE;
if (error == ECANCELED) {
ASSERT0(data->zcd_txg);
ASSERT(!data->zcd_added);
/*
* The private callback data should be destroyed here, but
* since we are going to check the zcd_called field after
* dmu_tx_abort(), we will destroy it there.
*/
return;
}
ASSERT(data->zcd_added);
ASSERT3U(data->zcd_txg, !=, 0);
(void) mutex_enter(&zcl.zcl_callbacks_lock);
/* See if this cb was called more quickly */
if ((synced_txg - data->zcd_txg) < zc_min_txg_delay)
zc_min_txg_delay = synced_txg - data->zcd_txg;
/* Remove our callback from the list */
list_remove(&zcl.zcl_callbacks, data);
(void) mutex_exit(&zcl.zcl_callbacks_lock);
umem_free(data, sizeof (ztest_cb_data_t));
}
/* Allocate and initialize callback data structure */
static ztest_cb_data_t *
ztest_create_cb_data(objset_t *os, uint64_t txg)
{
ztest_cb_data_t *cb_data;
cb_data = umem_zalloc(sizeof (ztest_cb_data_t), UMEM_NOFAIL);
cb_data->zcd_txg = txg;
cb_data->zcd_spa = dmu_objset_spa(os);
list_link_init(&cb_data->zcd_node);
return (cb_data);
}
/*
* Commit callback test.
*/
void
ztest_dmu_commit_callbacks(ztest_ds_t *zd, uint64_t id)
{
objset_t *os = zd->zd_os;
ztest_od_t *od;
dmu_tx_t *tx;
ztest_cb_data_t *cb_data[3], *tmp_cb;
uint64_t old_txg, txg;
int i, error = 0;
od = umem_alloc(sizeof (ztest_od_t), UMEM_NOFAIL);
ztest_od_init(od, id, FTAG, 0, DMU_OT_UINT64_OTHER, 0, 0, 0);
if (ztest_object_init(zd, od, sizeof (ztest_od_t), B_FALSE) != 0) {
umem_free(od, sizeof (ztest_od_t));
return;
}
tx = dmu_tx_create(os);
cb_data[0] = ztest_create_cb_data(os, 0);
dmu_tx_callback_register(tx, ztest_commit_callback, cb_data[0]);
dmu_tx_hold_write(tx, od->od_object, 0, sizeof (uint64_t));
/* Every once in a while, abort the transaction on purpose */
if (ztest_random(100) == 0)
error = -1;
if (!error)
error = dmu_tx_assign(tx, TXG_NOWAIT);
txg = error ? 0 : dmu_tx_get_txg(tx);
cb_data[0]->zcd_txg = txg;
cb_data[1] = ztest_create_cb_data(os, txg);
dmu_tx_callback_register(tx, ztest_commit_callback, cb_data[1]);
if (error) {
/*
* It's not a strict requirement to call the registered
* callbacks from inside dmu_tx_abort(), but that's what
* it's supposed to happen in the current implementation
* so we will check for that.
*/
for (i = 0; i < 2; i++) {
cb_data[i]->zcd_expected_err = ECANCELED;
VERIFY(!cb_data[i]->zcd_called);
}
dmu_tx_abort(tx);
for (i = 0; i < 2; i++) {
VERIFY(cb_data[i]->zcd_called);
umem_free(cb_data[i], sizeof (ztest_cb_data_t));
}
umem_free(od, sizeof (ztest_od_t));
return;
}
cb_data[2] = ztest_create_cb_data(os, txg);
dmu_tx_callback_register(tx, ztest_commit_callback, cb_data[2]);
/*
* Read existing data to make sure there isn't a future leak.
*/
VERIFY0(dmu_read(os, od->od_object, 0, sizeof (uint64_t),
&old_txg, DMU_READ_PREFETCH));
if (old_txg > txg)
fatal(B_FALSE,
"future leak: got %"PRIu64", open txg is %"PRIu64"",
old_txg, txg);
dmu_write(os, od->od_object, 0, sizeof (uint64_t), &txg, tx);
(void) mutex_enter(&zcl.zcl_callbacks_lock);
/*
* Since commit callbacks don't have any ordering requirement and since
* it is theoretically possible for a commit callback to be called
* after an arbitrary amount of time has elapsed since its txg has been
* synced, it is difficult to reliably determine whether a commit
* callback hasn't been called due to high load or due to a flawed
* implementation.
*
* In practice, we will assume that if after a certain number of txgs a
* commit callback hasn't been called, then most likely there's an
* implementation bug..
*/
tmp_cb = list_head(&zcl.zcl_callbacks);
if (tmp_cb != NULL &&
tmp_cb->zcd_txg + ZTEST_COMMIT_CB_THRESH < txg) {
fatal(B_FALSE,
"Commit callback threshold exceeded, "
"oldest txg: %"PRIu64", open txg: %"PRIu64"\n",
tmp_cb->zcd_txg, txg);
}
/*
* Let's find the place to insert our callbacks.
*
* Even though the list is ordered by txg, it is possible for the
* insertion point to not be the end because our txg may already be
* quiescing at this point and other callbacks in the open txg
* (from other objsets) may have sneaked in.
*/
tmp_cb = list_tail(&zcl.zcl_callbacks);
while (tmp_cb != NULL && tmp_cb->zcd_txg > txg)
tmp_cb = list_prev(&zcl.zcl_callbacks, tmp_cb);
/* Add the 3 callbacks to the list */
for (i = 0; i < 3; i++) {
if (tmp_cb == NULL)
list_insert_head(&zcl.zcl_callbacks, cb_data[i]);
else
list_insert_after(&zcl.zcl_callbacks, tmp_cb,
cb_data[i]);
cb_data[i]->zcd_added = B_TRUE;
VERIFY(!cb_data[i]->zcd_called);
tmp_cb = cb_data[i];
}
zc_cb_counter += 3;
(void) mutex_exit(&zcl.zcl_callbacks_lock);
dmu_tx_commit(tx);
umem_free(od, sizeof (ztest_od_t));
}
/*
* Visit each object in the dataset. Verify that its properties
* are consistent what was stored in the block tag when it was created,
* and that its unused bonus buffer space has not been overwritten.
*/
void
ztest_verify_dnode_bt(ztest_ds_t *zd, uint64_t id)
{
(void) id;
objset_t *os = zd->zd_os;
uint64_t obj;
int err = 0;
for (obj = 0; err == 0; err = dmu_object_next(os, &obj, FALSE, 0)) {
ztest_block_tag_t *bt = NULL;
dmu_object_info_t doi;
dmu_buf_t *db;
- ztest_object_lock(zd, obj, RL_READER);
+ ztest_object_lock(zd, obj, ZTRL_READER);
if (dmu_bonus_hold(os, obj, FTAG, &db) != 0) {
ztest_object_unlock(zd, obj);
continue;
}
dmu_object_info_from_db(db, &doi);
if (doi.doi_bonus_size >= sizeof (*bt))
bt = ztest_bt_bonus(db);
if (bt && bt->bt_magic == BT_MAGIC) {
ztest_bt_verify(bt, os, obj, doi.doi_dnodesize,
bt->bt_offset, bt->bt_gen, bt->bt_txg,
bt->bt_crtxg);
ztest_verify_unused_bonus(db, bt, obj, os, bt->bt_gen);
}
dmu_buf_rele(db, FTAG);
ztest_object_unlock(zd, obj);
}
}
void
ztest_dsl_prop_get_set(ztest_ds_t *zd, uint64_t id)
{
(void) id;
zfs_prop_t proplist[] = {
ZFS_PROP_CHECKSUM,
ZFS_PROP_COMPRESSION,
ZFS_PROP_COPIES,
ZFS_PROP_DEDUP
};
(void) pthread_rwlock_rdlock(&ztest_name_lock);
for (int p = 0; p < sizeof (proplist) / sizeof (proplist[0]); p++) {
int error = ztest_dsl_prop_set_uint64(zd->zd_name, proplist[p],
ztest_random_dsl_prop(proplist[p]), (int)ztest_random(2));
ASSERT(error == 0 || error == ENOSPC);
}
int error = ztest_dsl_prop_set_uint64(zd->zd_name, ZFS_PROP_RECORDSIZE,
ztest_random_blocksize(), (int)ztest_random(2));
ASSERT(error == 0 || error == ENOSPC);
(void) pthread_rwlock_unlock(&ztest_name_lock);
}
void
ztest_spa_prop_get_set(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
nvlist_t *props = NULL;
(void) pthread_rwlock_rdlock(&ztest_name_lock);
(void) ztest_spa_prop_set_uint64(ZPOOL_PROP_AUTOTRIM, ztest_random(2));
VERIFY0(spa_prop_get(ztest_spa, &props));
if (ztest_opts.zo_verbose >= 6)
dump_nvlist(props, 4);
fnvlist_free(props);
(void) pthread_rwlock_unlock(&ztest_name_lock);
}
static int
user_release_one(const char *snapname, const char *holdname)
{
nvlist_t *snaps, *holds;
int error;
snaps = fnvlist_alloc();
holds = fnvlist_alloc();
fnvlist_add_boolean(holds, holdname);
fnvlist_add_nvlist(snaps, snapname, holds);
fnvlist_free(holds);
error = dsl_dataset_user_release(snaps, NULL);
fnvlist_free(snaps);
return (error);
}
/*
* Test snapshot hold/release and deferred destroy.
*/
void
ztest_dmu_snapshot_hold(ztest_ds_t *zd, uint64_t id)
{
int error;
objset_t *os = zd->zd_os;
objset_t *origin;
char snapname[100];
char fullname[100];
char clonename[100];
char tag[100];
char osname[ZFS_MAX_DATASET_NAME_LEN];
nvlist_t *holds;
(void) pthread_rwlock_rdlock(&ztest_name_lock);
dmu_objset_name(os, osname);
(void) snprintf(snapname, sizeof (snapname), "sh1_%"PRIu64"", id);
(void) snprintf(fullname, sizeof (fullname), "%s@%s", osname, snapname);
(void) snprintf(clonename, sizeof (clonename), "%s/ch1_%"PRIu64"",
osname, id);
(void) snprintf(tag, sizeof (tag), "tag_%"PRIu64"", id);
/*
* Clean up from any previous run.
*/
error = dsl_destroy_head(clonename);
if (error != ENOENT)
ASSERT0(error);
error = user_release_one(fullname, tag);
if (error != ESRCH && error != ENOENT)
ASSERT0(error);
error = dsl_destroy_snapshot(fullname, B_FALSE);
if (error != ENOENT)
ASSERT0(error);
/*
* Create snapshot, clone it, mark snap for deferred destroy,
* destroy clone, verify snap was also destroyed.
*/
error = dmu_objset_snapshot_one(osname, snapname);
if (error) {
if (error == ENOSPC) {
ztest_record_enospc("dmu_objset_snapshot");
goto out;
}
fatal(B_FALSE, "dmu_objset_snapshot(%s) = %d", fullname, error);
}
error = dmu_objset_clone(clonename, fullname);
if (error) {
if (error == ENOSPC) {
ztest_record_enospc("dmu_objset_clone");
goto out;
}
fatal(B_FALSE, "dmu_objset_clone(%s) = %d", clonename, error);
}
error = dsl_destroy_snapshot(fullname, B_TRUE);
if (error) {
fatal(B_FALSE, "dsl_destroy_snapshot(%s, B_TRUE) = %d",
fullname, error);
}
error = dsl_destroy_head(clonename);
if (error)
fatal(B_FALSE, "dsl_destroy_head(%s) = %d", clonename, error);
error = dmu_objset_hold(fullname, FTAG, &origin);
if (error != ENOENT)
fatal(B_FALSE, "dmu_objset_hold(%s) = %d", fullname, error);
/*
* Create snapshot, add temporary hold, verify that we can't
* destroy a held snapshot, mark for deferred destroy,
* release hold, verify snapshot was destroyed.
*/
error = dmu_objset_snapshot_one(osname, snapname);
if (error) {
if (error == ENOSPC) {
ztest_record_enospc("dmu_objset_snapshot");
goto out;
}
fatal(B_FALSE, "dmu_objset_snapshot(%s) = %d", fullname, error);
}
holds = fnvlist_alloc();
fnvlist_add_string(holds, fullname, tag);
error = dsl_dataset_user_hold(holds, 0, NULL);
fnvlist_free(holds);
if (error == ENOSPC) {
ztest_record_enospc("dsl_dataset_user_hold");
goto out;
} else if (error) {
fatal(B_FALSE, "dsl_dataset_user_hold(%s, %s) = %u",
fullname, tag, error);
}
error = dsl_destroy_snapshot(fullname, B_FALSE);
if (error != EBUSY) {
fatal(B_FALSE, "dsl_destroy_snapshot(%s, B_FALSE) = %d",
fullname, error);
}
error = dsl_destroy_snapshot(fullname, B_TRUE);
if (error) {
fatal(B_FALSE, "dsl_destroy_snapshot(%s, B_TRUE) = %d",
fullname, error);
}
error = user_release_one(fullname, tag);
if (error)
fatal(B_FALSE, "user_release_one(%s, %s) = %d",
fullname, tag, error);
VERIFY3U(dmu_objset_hold(fullname, FTAG, &origin), ==, ENOENT);
out:
(void) pthread_rwlock_unlock(&ztest_name_lock);
}
/*
* Inject random faults into the on-disk data.
*/
void
ztest_fault_inject(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
ztest_shared_t *zs = ztest_shared;
spa_t *spa = ztest_spa;
int fd;
uint64_t offset;
uint64_t leaves;
uint64_t bad = 0x1990c0ffeedecadeull;
uint64_t top, leaf;
+ uint64_t raidz_children;
char *path0;
char *pathrand;
size_t fsize;
int bshift = SPA_MAXBLOCKSHIFT + 2;
int iters = 1000;
int maxfaults;
int mirror_save;
vdev_t *vd0 = NULL;
uint64_t guid0 = 0;
boolean_t islog = B_FALSE;
+ boolean_t injected = B_FALSE;
path0 = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
pathrand = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
mutex_enter(&ztest_vdev_lock);
/*
* Device removal is in progress, fault injection must be disabled
* until it completes and the pool is scrubbed. The fault injection
* strategy for damaging blocks does not take in to account evacuated
* blocks which may have already been damaged.
*/
- if (ztest_device_removal_active) {
- mutex_exit(&ztest_vdev_lock);
+ if (ztest_device_removal_active)
+ goto out;
+
+ /*
+ * The fault injection strategy for damaging blocks cannot be used
+ * if raidz expansion is in progress. The leaves value
+ * (attached raidz children) is variable and strategy for damaging
+ * blocks will corrupt same data blocks on different child vdevs
+ * because of the reflow process.
+ */
+ if (spa->spa_raidz_expand != NULL)
goto out;
- }
maxfaults = MAXFAULTS(zs);
- leaves = MAX(zs->zs_mirrors, 1) * ztest_opts.zo_raid_children;
+ raidz_children = ztest_get_raidz_children(spa);
+ leaves = MAX(zs->zs_mirrors, 1) * raidz_children;
mirror_save = zs->zs_mirrors;
- mutex_exit(&ztest_vdev_lock);
ASSERT3U(leaves, >=, 1);
/*
* While ztest is running the number of leaves will not change. This
* is critical for the fault injection logic as it determines where
* errors can be safely injected such that they are always repairable.
*
* When restarting ztest a different number of leaves may be requested
* which will shift the regions to be damaged. This is fine as long
* as the pool has been scrubbed prior to using the new mapping.
* Failure to do can result in non-repairable damage being injected.
*/
if (ztest_pool_scrubbed == B_FALSE)
goto out;
/*
* Grab the name lock as reader. There are some operations
* which don't like to have their vdevs changed while
* they are in progress (i.e. spa_change_guid). Those
* operations will have grabbed the name lock as writer.
*/
(void) pthread_rwlock_rdlock(&ztest_name_lock);
/*
* We need SCL_STATE here because we're going to look at vd0->vdev_tsd.
*/
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
if (ztest_random(2) == 0) {
/*
* Inject errors on a normal data device or slog device.
*/
top = ztest_random_vdev_top(spa, B_TRUE);
leaf = ztest_random(leaves) + zs->zs_splits;
/*
* Generate paths to the first leaf in this top-level vdev,
* and to the random leaf we selected. We'll induce transient
* write failures and random online/offline activity on leaf 0,
* and we'll write random garbage to the randomly chosen leaf.
*/
(void) snprintf(path0, MAXPATHLEN, ztest_dev_template,
ztest_opts.zo_dir, ztest_opts.zo_pool,
top * leaves + zs->zs_splits);
(void) snprintf(pathrand, MAXPATHLEN, ztest_dev_template,
ztest_opts.zo_dir, ztest_opts.zo_pool,
top * leaves + leaf);
vd0 = vdev_lookup_by_path(spa->spa_root_vdev, path0);
if (vd0 != NULL && vd0->vdev_top->vdev_islog)
islog = B_TRUE;
/*
* If the top-level vdev needs to be resilvered
* then we only allow faults on the device that is
* resilvering.
*/
if (vd0 != NULL && maxfaults != 1 &&
(!vdev_resilver_needed(vd0->vdev_top, NULL, NULL) ||
vd0->vdev_resilver_txg != 0)) {
/*
* Make vd0 explicitly claim to be unreadable,
* or unwritable, or reach behind its back
* and close the underlying fd. We can do this if
* maxfaults == 0 because we'll fail and reexecute,
* and we can do it if maxfaults >= 2 because we'll
* have enough redundancy. If maxfaults == 1, the
* combination of this with injection of random data
* corruption below exceeds the pool's fault tolerance.
*/
vdev_file_t *vf = vd0->vdev_tsd;
zfs_dbgmsg("injecting fault to vdev %llu; maxfaults=%d",
(long long)vd0->vdev_id, (int)maxfaults);
if (vf != NULL && ztest_random(3) == 0) {
(void) close(vf->vf_file->f_fd);
vf->vf_file->f_fd = -1;
} else if (ztest_random(2) == 0) {
vd0->vdev_cant_read = B_TRUE;
} else {
vd0->vdev_cant_write = B_TRUE;
}
guid0 = vd0->vdev_guid;
}
} else {
/*
* Inject errors on an l2cache device.
*/
spa_aux_vdev_t *sav = &spa->spa_l2cache;
if (sav->sav_count == 0) {
spa_config_exit(spa, SCL_STATE, FTAG);
(void) pthread_rwlock_unlock(&ztest_name_lock);
goto out;
}
vd0 = sav->sav_vdevs[ztest_random(sav->sav_count)];
guid0 = vd0->vdev_guid;
(void) strlcpy(path0, vd0->vdev_path, MAXPATHLEN);
(void) strlcpy(pathrand, vd0->vdev_path, MAXPATHLEN);
leaf = 0;
leaves = 1;
maxfaults = INT_MAX; /* no limit on cache devices */
}
spa_config_exit(spa, SCL_STATE, FTAG);
(void) pthread_rwlock_unlock(&ztest_name_lock);
/*
* If we can tolerate two or more faults, or we're dealing
* with a slog, randomly online/offline vd0.
*/
if ((maxfaults >= 2 || islog) && guid0 != 0) {
if (ztest_random(10) < 6) {
int flags = (ztest_random(2) == 0 ?
ZFS_OFFLINE_TEMPORARY : 0);
/*
* We have to grab the zs_name_lock as writer to
* prevent a race between offlining a slog and
* destroying a dataset. Offlining the slog will
* grab a reference on the dataset which may cause
* dsl_destroy_head() to fail with EBUSY thus
* leaving the dataset in an inconsistent state.
*/
if (islog)
(void) pthread_rwlock_wrlock(&ztest_name_lock);
VERIFY3U(vdev_offline(spa, guid0, flags), !=, EBUSY);
if (islog)
(void) pthread_rwlock_unlock(&ztest_name_lock);
} else {
/*
* Ideally we would like to be able to randomly
* call vdev_[on|off]line without holding locks
* to force unpredictable failures but the side
* effects of vdev_[on|off]line prevent us from
- * doing so. We grab the ztest_vdev_lock here to
- * prevent a race between injection testing and
- * aux_vdev removal.
+ * doing so.
*/
- mutex_enter(&ztest_vdev_lock);
(void) vdev_online(spa, guid0, 0, NULL);
- mutex_exit(&ztest_vdev_lock);
}
}
if (maxfaults == 0)
goto out;
/*
* We have at least single-fault tolerance, so inject data corruption.
*/
fd = open(pathrand, O_RDWR);
if (fd == -1) /* we hit a gap in the device namespace */
goto out;
fsize = lseek(fd, 0, SEEK_END);
while (--iters != 0) {
/*
* The offset must be chosen carefully to ensure that
* we do not inject a given logical block with errors
* on two different leaf devices, because ZFS can not
* tolerate that (if maxfaults==1).
*
* To achieve this we divide each leaf device into
* chunks of size (# leaves * SPA_MAXBLOCKSIZE * 4).
* Each chunk is further divided into error-injection
* ranges (can accept errors) and clear ranges (we do
* not inject errors in those). Each error-injection
* range can accept errors only for a single leaf vdev.
* Error-injection ranges are separated by clear ranges.
*
* For example, with 3 leaves, each chunk looks like:
* 0 to 32M: injection range for leaf 0
* 32M to 64M: clear range - no injection allowed
* 64M to 96M: injection range for leaf 1
* 96M to 128M: clear range - no injection allowed
* 128M to 160M: injection range for leaf 2
* 160M to 192M: clear range - no injection allowed
*
* Each clear range must be large enough such that a
* single block cannot straddle it. This way a block
* can't be a target in two different injection ranges
* (on different leaf vdevs).
*/
offset = ztest_random(fsize / (leaves << bshift)) *
(leaves << bshift) + (leaf << bshift) +
(ztest_random(1ULL << (bshift - 1)) & -8ULL);
/*
* Only allow damage to the labels at one end of the vdev.
*
* If all labels are damaged, the device will be totally
* inaccessible, which will result in loss of data,
* because we also damage (parts of) the other side of
* the mirror/raidz.
*
* Additionally, we will always have both an even and an
* odd label, so that we can handle crashes in the
* middle of vdev_config_sync().
*/
if ((leaf & 1) == 0 && offset < VDEV_LABEL_START_SIZE)
continue;
/*
* The two end labels are stored at the "end" of the disk, but
* the end of the disk (vdev_psize) is aligned to
* sizeof (vdev_label_t).
*/
uint64_t psize = P2ALIGN(fsize, sizeof (vdev_label_t));
if ((leaf & 1) == 1 &&
offset + sizeof (bad) > psize - VDEV_LABEL_END_SIZE)
continue;
- mutex_enter(&ztest_vdev_lock);
if (mirror_save != zs->zs_mirrors) {
- mutex_exit(&ztest_vdev_lock);
(void) close(fd);
goto out;
}
if (pwrite(fd, &bad, sizeof (bad), offset) != sizeof (bad))
fatal(B_TRUE,
"can't inject bad word at 0x%"PRIx64" in %s",
offset, pathrand);
- mutex_exit(&ztest_vdev_lock);
-
if (ztest_opts.zo_verbose >= 7)
(void) printf("injected bad word into %s,"
" offset 0x%"PRIx64"\n", pathrand, offset);
+
+ injected = B_TRUE;
}
(void) close(fd);
out:
+ mutex_exit(&ztest_vdev_lock);
+
+ if (injected && ztest_opts.zo_raid_do_expand) {
+ int error = spa_scan(spa, POOL_SCAN_SCRUB);
+ if (error == 0) {
+ while (dsl_scan_scrubbing(spa_get_dsl(spa)))
+ txg_wait_synced(spa_get_dsl(spa), 0);
+ }
+ }
+
umem_free(path0, MAXPATHLEN);
umem_free(pathrand, MAXPATHLEN);
}
/*
* By design ztest will never inject uncorrectable damage in to the pool.
* Issue a scrub, wait for it to complete, and verify there is never any
* persistent damage.
*
* Only after a full scrub has been completed is it safe to start injecting
* data corruption. See the comment in zfs_fault_inject().
*/
static int
ztest_scrub_impl(spa_t *spa)
{
int error = spa_scan(spa, POOL_SCAN_SCRUB);
if (error)
return (error);
while (dsl_scan_scrubbing(spa_get_dsl(spa)))
txg_wait_synced(spa_get_dsl(spa), 0);
if (spa_approx_errlog_size(spa) > 0)
return (ECKSUM);
ztest_pool_scrubbed = B_TRUE;
return (0);
}
/*
* Scrub the pool.
*/
void
ztest_scrub(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
spa_t *spa = ztest_spa;
int error;
/*
* Scrub in progress by device removal.
*/
if (ztest_device_removal_active)
return;
/*
* Start a scrub, wait a moment, then force a restart.
*/
(void) spa_scan(spa, POOL_SCAN_SCRUB);
(void) poll(NULL, 0, 100);
error = ztest_scrub_impl(spa);
if (error == EBUSY)
error = 0;
ASSERT0(error);
}
/*
* Change the guid for the pool.
*/
void
ztest_reguid(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
spa_t *spa = ztest_spa;
uint64_t orig, load;
int error;
ztest_shared_t *zs = ztest_shared;
if (ztest_opts.zo_mmp_test)
return;
orig = spa_guid(spa);
load = spa_load_guid(spa);
(void) pthread_rwlock_wrlock(&ztest_name_lock);
error = spa_change_guid(spa);
zs->zs_guid = spa_guid(spa);
(void) pthread_rwlock_unlock(&ztest_name_lock);
if (error != 0)
return;
if (ztest_opts.zo_verbose >= 4) {
(void) printf("Changed guid old %"PRIu64" -> %"PRIu64"\n",
orig, spa_guid(spa));
}
VERIFY3U(orig, !=, spa_guid(spa));
VERIFY3U(load, ==, spa_load_guid(spa));
}
void
ztest_blake3(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
hrtime_t end = gethrtime() + NANOSEC;
zio_cksum_salt_t salt;
void *salt_ptr = &salt.zcs_bytes;
struct abd *abd_data, *abd_meta;
void *buf, *templ;
int i, *ptr;
uint32_t size;
BLAKE3_CTX ctx;
const zfs_impl_t *blake3 = zfs_impl_get_ops("blake3");
size = ztest_random_blocksize();
buf = umem_alloc(size, UMEM_NOFAIL);
abd_data = abd_alloc(size, B_FALSE);
abd_meta = abd_alloc(size, B_TRUE);
for (i = 0, ptr = buf; i < size / sizeof (*ptr); i++, ptr++)
*ptr = ztest_random(UINT_MAX);
memset(salt_ptr, 'A', 32);
abd_copy_from_buf_off(abd_data, buf, 0, size);
abd_copy_from_buf_off(abd_meta, buf, 0, size);
while (gethrtime() <= end) {
int run_count = 100;
zio_cksum_t zc_ref1, zc_ref2;
zio_cksum_t zc_res1, zc_res2;
void *ref1 = &zc_ref1;
void *ref2 = &zc_ref2;
void *res1 = &zc_res1;
void *res2 = &zc_res2;
/* BLAKE3_KEY_LEN = 32 */
VERIFY0(blake3->setname("generic"));
templ = abd_checksum_blake3_tmpl_init(&salt);
Blake3_InitKeyed(&ctx, salt_ptr);
Blake3_Update(&ctx, buf, size);
Blake3_Final(&ctx, ref1);
zc_ref2 = zc_ref1;
ZIO_CHECKSUM_BSWAP(&zc_ref2);
abd_checksum_blake3_tmpl_free(templ);
VERIFY0(blake3->setname("cycle"));
while (run_count-- > 0) {
/* Test current implementation */
Blake3_InitKeyed(&ctx, salt_ptr);
Blake3_Update(&ctx, buf, size);
Blake3_Final(&ctx, res1);
zc_res2 = zc_res1;
ZIO_CHECKSUM_BSWAP(&zc_res2);
VERIFY0(memcmp(ref1, res1, 32));
VERIFY0(memcmp(ref2, res2, 32));
/* Test ABD - data */
templ = abd_checksum_blake3_tmpl_init(&salt);
abd_checksum_blake3_native(abd_data, size,
templ, &zc_res1);
abd_checksum_blake3_byteswap(abd_data, size,
templ, &zc_res2);
VERIFY0(memcmp(ref1, res1, 32));
VERIFY0(memcmp(ref2, res2, 32));
/* Test ABD - metadata */
abd_checksum_blake3_native(abd_meta, size,
templ, &zc_res1);
abd_checksum_blake3_byteswap(abd_meta, size,
templ, &zc_res2);
abd_checksum_blake3_tmpl_free(templ);
VERIFY0(memcmp(ref1, res1, 32));
VERIFY0(memcmp(ref2, res2, 32));
}
}
abd_free(abd_data);
abd_free(abd_meta);
umem_free(buf, size);
}
void
ztest_fletcher(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
hrtime_t end = gethrtime() + NANOSEC;
while (gethrtime() <= end) {
int run_count = 100;
void *buf;
struct abd *abd_data, *abd_meta;
uint32_t size;
int *ptr;
int i;
zio_cksum_t zc_ref;
zio_cksum_t zc_ref_byteswap;
size = ztest_random_blocksize();
buf = umem_alloc(size, UMEM_NOFAIL);
abd_data = abd_alloc(size, B_FALSE);
abd_meta = abd_alloc(size, B_TRUE);
for (i = 0, ptr = buf; i < size / sizeof (*ptr); i++, ptr++)
*ptr = ztest_random(UINT_MAX);
abd_copy_from_buf_off(abd_data, buf, 0, size);
abd_copy_from_buf_off(abd_meta, buf, 0, size);
VERIFY0(fletcher_4_impl_set("scalar"));
fletcher_4_native(buf, size, NULL, &zc_ref);
fletcher_4_byteswap(buf, size, NULL, &zc_ref_byteswap);
VERIFY0(fletcher_4_impl_set("cycle"));
while (run_count-- > 0) {
zio_cksum_t zc;
zio_cksum_t zc_byteswap;
fletcher_4_byteswap(buf, size, NULL, &zc_byteswap);
fletcher_4_native(buf, size, NULL, &zc);
VERIFY0(memcmp(&zc, &zc_ref, sizeof (zc)));
VERIFY0(memcmp(&zc_byteswap, &zc_ref_byteswap,
sizeof (zc_byteswap)));
/* Test ABD - data */
abd_fletcher_4_byteswap(abd_data, size, NULL,
&zc_byteswap);
abd_fletcher_4_native(abd_data, size, NULL, &zc);
VERIFY0(memcmp(&zc, &zc_ref, sizeof (zc)));
VERIFY0(memcmp(&zc_byteswap, &zc_ref_byteswap,
sizeof (zc_byteswap)));
/* Test ABD - metadata */
abd_fletcher_4_byteswap(abd_meta, size, NULL,
&zc_byteswap);
abd_fletcher_4_native(abd_meta, size, NULL, &zc);
VERIFY0(memcmp(&zc, &zc_ref, sizeof (zc)));
VERIFY0(memcmp(&zc_byteswap, &zc_ref_byteswap,
sizeof (zc_byteswap)));
}
umem_free(buf, size);
abd_free(abd_data);
abd_free(abd_meta);
}
}
void
ztest_fletcher_incr(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
void *buf;
size_t size;
int *ptr;
int i;
zio_cksum_t zc_ref;
zio_cksum_t zc_ref_bswap;
hrtime_t end = gethrtime() + NANOSEC;
while (gethrtime() <= end) {
int run_count = 100;
size = ztest_random_blocksize();
buf = umem_alloc(size, UMEM_NOFAIL);
for (i = 0, ptr = buf; i < size / sizeof (*ptr); i++, ptr++)
*ptr = ztest_random(UINT_MAX);
VERIFY0(fletcher_4_impl_set("scalar"));
fletcher_4_native(buf, size, NULL, &zc_ref);
fletcher_4_byteswap(buf, size, NULL, &zc_ref_bswap);
VERIFY0(fletcher_4_impl_set("cycle"));
while (run_count-- > 0) {
zio_cksum_t zc;
zio_cksum_t zc_bswap;
size_t pos = 0;
ZIO_SET_CHECKSUM(&zc, 0, 0, 0, 0);
ZIO_SET_CHECKSUM(&zc_bswap, 0, 0, 0, 0);
while (pos < size) {
size_t inc = 64 * ztest_random(size / 67);
/* sometimes add few bytes to test non-simd */
if (ztest_random(100) < 10)
inc += P2ALIGN(ztest_random(64),
sizeof (uint32_t));
if (inc > (size - pos))
inc = size - pos;
fletcher_4_incremental_native(buf + pos, inc,
&zc);
fletcher_4_incremental_byteswap(buf + pos, inc,
&zc_bswap);
pos += inc;
}
VERIFY3U(pos, ==, size);
VERIFY(ZIO_CHECKSUM_EQUAL(zc, zc_ref));
VERIFY(ZIO_CHECKSUM_EQUAL(zc_bswap, zc_ref_bswap));
/*
* verify if incremental on the whole buffer is
* equivalent to non-incremental version
*/
ZIO_SET_CHECKSUM(&zc, 0, 0, 0, 0);
ZIO_SET_CHECKSUM(&zc_bswap, 0, 0, 0, 0);
fletcher_4_incremental_native(buf, size, &zc);
fletcher_4_incremental_byteswap(buf, size, &zc_bswap);
VERIFY(ZIO_CHECKSUM_EQUAL(zc, zc_ref));
VERIFY(ZIO_CHECKSUM_EQUAL(zc_bswap, zc_ref_bswap));
}
umem_free(buf, size);
}
}
static int
ztest_set_global_vars(void)
{
for (size_t i = 0; i < ztest_opts.zo_gvars_count; i++) {
char *kv = ztest_opts.zo_gvars[i];
VERIFY3U(strlen(kv), <=, ZO_GVARS_MAX_ARGLEN);
VERIFY3U(strlen(kv), >, 0);
int err = set_global_var(kv);
if (ztest_opts.zo_verbose > 0) {
(void) printf("setting global var %s ... %s\n", kv,
err ? "failed" : "ok");
}
if (err != 0) {
(void) fprintf(stderr,
"failed to set global var '%s'\n", kv);
return (err);
}
}
return (0);
}
static char **
ztest_global_vars_to_zdb_args(void)
{
char **args = calloc(2*ztest_opts.zo_gvars_count + 1, sizeof (char *));
char **cur = args;
if (args == NULL)
return (NULL);
for (size_t i = 0; i < ztest_opts.zo_gvars_count; i++) {
*cur++ = (char *)"-o";
*cur++ = ztest_opts.zo_gvars[i];
}
ASSERT3P(cur, ==, &args[2*ztest_opts.zo_gvars_count]);
*cur = NULL;
return (args);
}
/* The end of strings is indicated by a NULL element */
static char *
join_strings(char **strings, const char *sep)
{
size_t totallen = 0;
for (char **sp = strings; *sp != NULL; sp++) {
totallen += strlen(*sp);
totallen += strlen(sep);
}
if (totallen > 0) {
ASSERT(totallen >= strlen(sep));
totallen -= strlen(sep);
}
size_t buflen = totallen + 1;
char *o = umem_alloc(buflen, UMEM_NOFAIL); /* trailing 0 byte */
o[0] = '\0';
for (char **sp = strings; *sp != NULL; sp++) {
size_t would;
would = strlcat(o, *sp, buflen);
VERIFY3U(would, <, buflen);
if (*(sp+1) == NULL) {
break;
}
would = strlcat(o, sep, buflen);
VERIFY3U(would, <, buflen);
}
ASSERT3S(strlen(o), ==, totallen);
return (o);
}
static int
ztest_check_path(char *path)
{
struct stat s;
/* return true on success */
return (!stat(path, &s));
}
static void
ztest_get_zdb_bin(char *bin, int len)
{
char *zdb_path;
/*
* Try to use $ZDB and in-tree zdb path. If not successful, just
* let popen to search through PATH.
*/
if ((zdb_path = getenv("ZDB"))) {
strlcpy(bin, zdb_path, len); /* In env */
if (!ztest_check_path(bin)) {
ztest_dump_core = 0;
fatal(B_TRUE, "invalid ZDB '%s'", bin);
}
return;
}
VERIFY3P(realpath(getexecname(), bin), !=, NULL);
if (strstr(bin, ".libs/ztest")) {
strstr(bin, ".libs/ztest")[0] = '\0'; /* In-tree */
strcat(bin, "zdb");
if (ztest_check_path(bin))
return;
}
strcpy(bin, "zdb");
}
static vdev_t *
ztest_random_concrete_vdev_leaf(vdev_t *vd)
{
if (vd == NULL)
return (NULL);
if (vd->vdev_children == 0)
return (vd);
vdev_t *eligible[vd->vdev_children];
int eligible_idx = 0, i;
for (i = 0; i < vd->vdev_children; i++) {
vdev_t *cvd = vd->vdev_child[i];
if (cvd->vdev_top->vdev_removing)
continue;
if (cvd->vdev_children > 0 ||
(vdev_is_concrete(cvd) && !cvd->vdev_detached)) {
eligible[eligible_idx++] = cvd;
}
}
VERIFY3S(eligible_idx, >, 0);
uint64_t child_no = ztest_random(eligible_idx);
return (ztest_random_concrete_vdev_leaf(eligible[child_no]));
}
void
ztest_initialize(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
spa_t *spa = ztest_spa;
int error = 0;
mutex_enter(&ztest_vdev_lock);
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
/* Random leaf vdev */
vdev_t *rand_vd = ztest_random_concrete_vdev_leaf(spa->spa_root_vdev);
if (rand_vd == NULL) {
spa_config_exit(spa, SCL_VDEV, FTAG);
mutex_exit(&ztest_vdev_lock);
return;
}
/*
* The random vdev we've selected may change as soon as we
* drop the spa_config_lock. We create local copies of things
* we're interested in.
*/
uint64_t guid = rand_vd->vdev_guid;
char *path = strdup(rand_vd->vdev_path);
boolean_t active = rand_vd->vdev_initialize_thread != NULL;
zfs_dbgmsg("vd %px, guid %llu", rand_vd, (u_longlong_t)guid);
spa_config_exit(spa, SCL_VDEV, FTAG);
uint64_t cmd = ztest_random(POOL_INITIALIZE_FUNCS);
nvlist_t *vdev_guids = fnvlist_alloc();
nvlist_t *vdev_errlist = fnvlist_alloc();
fnvlist_add_uint64(vdev_guids, path, guid);
error = spa_vdev_initialize(spa, vdev_guids, cmd, vdev_errlist);
fnvlist_free(vdev_guids);
fnvlist_free(vdev_errlist);
switch (cmd) {
case POOL_INITIALIZE_CANCEL:
if (ztest_opts.zo_verbose >= 4) {
(void) printf("Cancel initialize %s", path);
if (!active)
(void) printf(" failed (no initialize active)");
(void) printf("\n");
}
break;
case POOL_INITIALIZE_START:
if (ztest_opts.zo_verbose >= 4) {
(void) printf("Start initialize %s", path);
if (active && error == 0)
(void) printf(" failed (already active)");
else if (error != 0)
(void) printf(" failed (error %d)", error);
(void) printf("\n");
}
break;
case POOL_INITIALIZE_SUSPEND:
if (ztest_opts.zo_verbose >= 4) {
(void) printf("Suspend initialize %s", path);
if (!active)
(void) printf(" failed (no initialize active)");
(void) printf("\n");
}
break;
}
free(path);
mutex_exit(&ztest_vdev_lock);
}
void
ztest_trim(ztest_ds_t *zd, uint64_t id)
{
(void) zd, (void) id;
spa_t *spa = ztest_spa;
int error = 0;
mutex_enter(&ztest_vdev_lock);
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
/* Random leaf vdev */
vdev_t *rand_vd = ztest_random_concrete_vdev_leaf(spa->spa_root_vdev);
if (rand_vd == NULL) {
spa_config_exit(spa, SCL_VDEV, FTAG);
mutex_exit(&ztest_vdev_lock);
return;
}
/*
* The random vdev we've selected may change as soon as we
* drop the spa_config_lock. We create local copies of things
* we're interested in.
*/
uint64_t guid = rand_vd->vdev_guid;
char *path = strdup(rand_vd->vdev_path);
boolean_t active = rand_vd->vdev_trim_thread != NULL;
zfs_dbgmsg("vd %p, guid %llu", rand_vd, (u_longlong_t)guid);
spa_config_exit(spa, SCL_VDEV, FTAG);
uint64_t cmd = ztest_random(POOL_TRIM_FUNCS);
uint64_t rate = 1 << ztest_random(30);
boolean_t partial = (ztest_random(5) > 0);
boolean_t secure = (ztest_random(5) > 0);
nvlist_t *vdev_guids = fnvlist_alloc();
nvlist_t *vdev_errlist = fnvlist_alloc();
fnvlist_add_uint64(vdev_guids, path, guid);
error = spa_vdev_trim(spa, vdev_guids, cmd, rate, partial,
secure, vdev_errlist);
fnvlist_free(vdev_guids);
fnvlist_free(vdev_errlist);
switch (cmd) {
case POOL_TRIM_CANCEL:
if (ztest_opts.zo_verbose >= 4) {
(void) printf("Cancel TRIM %s", path);
if (!active)
(void) printf(" failed (no TRIM active)");
(void) printf("\n");
}
break;
case POOL_TRIM_START:
if (ztest_opts.zo_verbose >= 4) {
(void) printf("Start TRIM %s", path);
if (active && error == 0)
(void) printf(" failed (already active)");
else if (error != 0)
(void) printf(" failed (error %d)", error);
(void) printf("\n");
}
break;
case POOL_TRIM_SUSPEND:
if (ztest_opts.zo_verbose >= 4) {
(void) printf("Suspend TRIM %s", path);
if (!active)
(void) printf(" failed (no TRIM active)");
(void) printf("\n");
}
break;
}
free(path);
mutex_exit(&ztest_vdev_lock);
}
/*
* Verify pool integrity by running zdb.
*/
static void
ztest_run_zdb(uint64_t guid)
{
int status;
char *bin;
char *zdb;
char *zbuf;
const int len = MAXPATHLEN + MAXNAMELEN + 20;
FILE *fp;
bin = umem_alloc(len, UMEM_NOFAIL);
zdb = umem_alloc(len, UMEM_NOFAIL);
zbuf = umem_alloc(1024, UMEM_NOFAIL);
ztest_get_zdb_bin(bin, len);
char **set_gvars_args = ztest_global_vars_to_zdb_args();
if (set_gvars_args == NULL) {
fatal(B_FALSE, "Failed to allocate memory in "
"ztest_global_vars_to_zdb_args(). Cannot run zdb.\n");
}
char *set_gvars_args_joined = join_strings(set_gvars_args, " ");
free(set_gvars_args);
size_t would = snprintf(zdb, len,
"%s -bcc%s%s -G -d -Y -e -y %s -p %s %"PRIu64,
bin,
ztest_opts.zo_verbose >= 3 ? "s" : "",
ztest_opts.zo_verbose >= 4 ? "v" : "",
set_gvars_args_joined,
ztest_opts.zo_dir,
guid);
ASSERT3U(would, <, len);
umem_free(set_gvars_args_joined, strlen(set_gvars_args_joined) + 1);
if (ztest_opts.zo_verbose >= 5)
(void) printf("Executing %s\n", zdb);
fp = popen(zdb, "r");
while (fgets(zbuf, 1024, fp) != NULL)
if (ztest_opts.zo_verbose >= 3)
(void) printf("%s", zbuf);
status = pclose(fp);
if (status == 0)
goto out;
ztest_dump_core = 0;
if (WIFEXITED(status))
fatal(B_FALSE, "'%s' exit code %d", zdb, WEXITSTATUS(status));
else
fatal(B_FALSE, "'%s' died with signal %d",
zdb, WTERMSIG(status));
out:
umem_free(bin, len);
umem_free(zdb, len);
umem_free(zbuf, 1024);
}
static void
ztest_walk_pool_directory(const char *header)
{
spa_t *spa = NULL;
if (ztest_opts.zo_verbose >= 6)
(void) puts(header);
mutex_enter(&spa_namespace_lock);
while ((spa = spa_next(spa)) != NULL)
if (ztest_opts.zo_verbose >= 6)
(void) printf("\t%s\n", spa_name(spa));
mutex_exit(&spa_namespace_lock);
}
static void
ztest_spa_import_export(char *oldname, char *newname)
{
nvlist_t *config, *newconfig;
uint64_t pool_guid;
spa_t *spa;
int error;
if (ztest_opts.zo_verbose >= 4) {
(void) printf("import/export: old = %s, new = %s\n",
oldname, newname);
}
/*
* Clean up from previous runs.
*/
(void) spa_destroy(newname);
/*
* Get the pool's configuration and guid.
*/
VERIFY0(spa_open(oldname, &spa, FTAG));
/*
* Kick off a scrub to tickle scrub/export races.
*/
if (ztest_random(2) == 0)
(void) spa_scan(spa, POOL_SCAN_SCRUB);
pool_guid = spa_guid(spa);
spa_close(spa, FTAG);
ztest_walk_pool_directory("pools before export");
/*
* Export it.
*/
VERIFY0(spa_export(oldname, &config, B_FALSE, B_FALSE));
ztest_walk_pool_directory("pools after export");
/*
* Try to import it.
*/
newconfig = spa_tryimport(config);
ASSERT3P(newconfig, !=, NULL);
fnvlist_free(newconfig);
/*
* Import it under the new name.
*/
error = spa_import(newname, config, NULL, 0);
if (error != 0) {
dump_nvlist(config, 0);
fatal(B_FALSE, "couldn't import pool %s as %s: error %u",
oldname, newname, error);
}
ztest_walk_pool_directory("pools after import");
/*
* Try to import it again -- should fail with EEXIST.
*/
VERIFY3U(EEXIST, ==, spa_import(newname, config, NULL, 0));
/*
* Try to import it under a different name -- should fail with EEXIST.
*/
VERIFY3U(EEXIST, ==, spa_import(oldname, config, NULL, 0));
/*
* Verify that the pool is no longer visible under the old name.
*/
VERIFY3U(ENOENT, ==, spa_open(oldname, &spa, FTAG));
/*
* Verify that we can open and close the pool using the new name.
*/
VERIFY0(spa_open(newname, &spa, FTAG));
ASSERT3U(pool_guid, ==, spa_guid(spa));
spa_close(spa, FTAG);
fnvlist_free(config);
}
static void
ztest_resume(spa_t *spa)
{
if (spa_suspended(spa) && ztest_opts.zo_verbose >= 6)
(void) printf("resuming from suspended state\n");
spa_vdev_state_enter(spa, SCL_NONE);
vdev_clear(spa, NULL);
(void) spa_vdev_state_exit(spa, NULL, 0);
(void) zio_resume(spa);
}
static __attribute__((noreturn)) void
ztest_resume_thread(void *arg)
{
spa_t *spa = arg;
while (!ztest_exiting) {
if (spa_suspended(spa))
ztest_resume(spa);
(void) poll(NULL, 0, 100);
/*
* Periodically change the zfs_compressed_arc_enabled setting.
*/
if (ztest_random(10) == 0)
zfs_compressed_arc_enabled = ztest_random(2);
/*
* Periodically change the zfs_abd_scatter_enabled setting.
*/
if (ztest_random(10) == 0)
zfs_abd_scatter_enabled = ztest_random(2);
}
thread_exit();
}
static __attribute__((noreturn)) void
ztest_deadman_thread(void *arg)
{
ztest_shared_t *zs = arg;
spa_t *spa = ztest_spa;
hrtime_t delay, overdue, last_run = gethrtime();
delay = (zs->zs_thread_stop - zs->zs_thread_start) +
MSEC2NSEC(zfs_deadman_synctime_ms);
while (!ztest_exiting) {
/*
* Wait for the delay timer while checking occasionally
* if we should stop.
*/
if (gethrtime() < last_run + delay) {
(void) poll(NULL, 0, 1000);
continue;
}
/*
* If the pool is suspended then fail immediately. Otherwise,
* check to see if the pool is making any progress. If
* vdev_deadman() discovers that there hasn't been any recent
* I/Os then it will end up aborting the tests.
*/
if (spa_suspended(spa) || spa->spa_root_vdev == NULL) {
fatal(B_FALSE,
"aborting test after %llu seconds because "
"pool has transitioned to a suspended state.",
(u_longlong_t)zfs_deadman_synctime_ms / 1000);
}
vdev_deadman(spa->spa_root_vdev, FTAG);
/*
* If the process doesn't complete within a grace period of
* zfs_deadman_synctime_ms over the expected finish time,
* then it may be hung and is terminated.
*/
overdue = zs->zs_proc_stop + MSEC2NSEC(zfs_deadman_synctime_ms);
if (gethrtime() > overdue) {
fatal(B_FALSE,
"aborting test after %llu seconds because "
"the process is overdue for termination.",
(gethrtime() - zs->zs_proc_start) / NANOSEC);
}
(void) printf("ztest has been running for %lld seconds\n",
(gethrtime() - zs->zs_proc_start) / NANOSEC);
last_run = gethrtime();
delay = MSEC2NSEC(zfs_deadman_checktime_ms);
}
thread_exit();
}
static void
ztest_execute(int test, ztest_info_t *zi, uint64_t id)
{
ztest_ds_t *zd = &ztest_ds[id % ztest_opts.zo_datasets];
ztest_shared_callstate_t *zc = ZTEST_GET_SHARED_CALLSTATE(test);
hrtime_t functime = gethrtime();
int i;
for (i = 0; i < zi->zi_iters; i++)
zi->zi_func(zd, id);
functime = gethrtime() - functime;
atomic_add_64(&zc->zc_count, 1);
atomic_add_64(&zc->zc_time, functime);
if (ztest_opts.zo_verbose >= 4)
(void) printf("%6.2f sec in %s\n",
(double)functime / NANOSEC, zi->zi_funcname);
}
+typedef struct ztest_raidz_expand_io {
+ uint64_t rzx_id;
+ uint64_t rzx_amount;
+ uint64_t rzx_bufsize;
+ const void *rzx_buffer;
+ uint64_t rzx_alloc_max;
+ spa_t *rzx_spa;
+} ztest_expand_io_t;
+
+#undef OD_ARRAY_SIZE
+#define OD_ARRAY_SIZE 10
+
+/*
+ * Write a request amount of data to some dataset objects.
+ * There will be ztest_opts.zo_threads count of these running in parallel.
+ */
+static __attribute__((noreturn)) void
+ztest_rzx_thread(void *arg)
+{
+ ztest_expand_io_t *info = (ztest_expand_io_t *)arg;
+ ztest_od_t *od;
+ int batchsize;
+ int od_size;
+ ztest_ds_t *zd = &ztest_ds[info->rzx_id % ztest_opts.zo_datasets];
+ spa_t *spa = info->rzx_spa;
+
+ od_size = sizeof (ztest_od_t) * OD_ARRAY_SIZE;
+ od = umem_alloc(od_size, UMEM_NOFAIL);
+ batchsize = OD_ARRAY_SIZE;
+
+ /* Create objects to write to */
+ for (int b = 0; b < batchsize; b++) {
+ ztest_od_init(od + b, info->rzx_id, FTAG, b,
+ DMU_OT_UINT64_OTHER, 0, 0, 0);
+ }
+ if (ztest_object_init(zd, od, od_size, B_FALSE) != 0) {
+ umem_free(od, od_size);
+ thread_exit();
+ }
+
+ for (uint64_t offset = 0, written = 0; written < info->rzx_amount;
+ offset += info->rzx_bufsize) {
+ /* write to 10 objects */
+ for (int i = 0; i < batchsize && written < info->rzx_amount;
+ i++) {
+ (void) pthread_rwlock_rdlock(&zd->zd_zilog_lock);
+ ztest_write(zd, od[i].od_object, offset,
+ info->rzx_bufsize, info->rzx_buffer);
+ (void) pthread_rwlock_unlock(&zd->zd_zilog_lock);
+ written += info->rzx_bufsize;
+ }
+ txg_wait_synced(spa_get_dsl(spa), 0);
+ /* due to inflation, we'll typically bail here */
+ if (metaslab_class_get_alloc(spa_normal_class(spa)) >
+ info->rzx_alloc_max) {
+ break;
+ }
+ }
+
+ /* Remove a few objects to leave some holes in allocation space */
+ mutex_enter(&zd->zd_dirobj_lock);
+ (void) ztest_remove(zd, od, 2);
+ mutex_exit(&zd->zd_dirobj_lock);
+
+ umem_free(od, od_size);
+
+ thread_exit();
+}
+
static __attribute__((noreturn)) void
ztest_thread(void *arg)
{
int rand;
uint64_t id = (uintptr_t)arg;
ztest_shared_t *zs = ztest_shared;
uint64_t call_next;
hrtime_t now;
ztest_info_t *zi;
ztest_shared_callstate_t *zc;
while ((now = gethrtime()) < zs->zs_thread_stop) {
/*
* See if it's time to force a crash.
*/
- if (now > zs->zs_thread_kill)
+ if (now > zs->zs_thread_kill &&
+ raidz_expand_pause_point == RAIDZ_EXPAND_PAUSE_NONE) {
ztest_kill(zs);
+ }
/*
* If we're getting ENOSPC with some regularity, stop.
*/
if (zs->zs_enospc_count > 10)
break;
/*
* Pick a random function to execute.
*/
rand = ztest_random(ZTEST_FUNCS);
zi = &ztest_info[rand];
zc = ZTEST_GET_SHARED_CALLSTATE(rand);
call_next = zc->zc_next;
if (now >= call_next &&
atomic_cas_64(&zc->zc_next, call_next, call_next +
ztest_random(2 * zi->zi_interval[0] + 1)) == call_next) {
ztest_execute(rand, zi, id);
}
}
thread_exit();
}
static void
ztest_dataset_name(char *dsname, const char *pool, int d)
{
(void) snprintf(dsname, ZFS_MAX_DATASET_NAME_LEN, "%s/ds_%d", pool, d);
}
static void
ztest_dataset_destroy(int d)
{
char name[ZFS_MAX_DATASET_NAME_LEN];
int t;
ztest_dataset_name(name, ztest_opts.zo_pool, d);
if (ztest_opts.zo_verbose >= 3)
(void) printf("Destroying %s to free up space\n", name);
/*
* Cleanup any non-standard clones and snapshots. In general,
* ztest thread t operates on dataset (t % zopt_datasets),
* so there may be more than one thing to clean up.
*/
for (t = d; t < ztest_opts.zo_threads;
t += ztest_opts.zo_datasets)
ztest_dsl_dataset_cleanup(name, t);
(void) dmu_objset_find(name, ztest_objset_destroy_cb, NULL,
DS_FIND_SNAPSHOTS | DS_FIND_CHILDREN);
}
static void
ztest_dataset_dirobj_verify(ztest_ds_t *zd)
{
uint64_t usedobjs, dirobjs, scratch;
/*
* ZTEST_DIROBJ is the object directory for the entire dataset.
* Therefore, the number of objects in use should equal the
* number of ZTEST_DIROBJ entries, +1 for ZTEST_DIROBJ itself.
* If not, we have an object leak.
*
* Note that we can only check this in ztest_dataset_open(),
* when the open-context and syncing-context values agree.
* That's because zap_count() returns the open-context value,
* while dmu_objset_space() returns the rootbp fill count.
*/
VERIFY0(zap_count(zd->zd_os, ZTEST_DIROBJ, &dirobjs));
dmu_objset_space(zd->zd_os, &scratch, &scratch, &usedobjs, &scratch);
ASSERT3U(dirobjs + 1, ==, usedobjs);
}
static int
ztest_dataset_open(int d)
{
ztest_ds_t *zd = &ztest_ds[d];
uint64_t committed_seq = ZTEST_GET_SHARED_DS(d)->zd_seq;
objset_t *os;
zilog_t *zilog;
char name[ZFS_MAX_DATASET_NAME_LEN];
int error;
ztest_dataset_name(name, ztest_opts.zo_pool, d);
(void) pthread_rwlock_rdlock(&ztest_name_lock);
error = ztest_dataset_create(name);
if (error == ENOSPC) {
(void) pthread_rwlock_unlock(&ztest_name_lock);
ztest_record_enospc(FTAG);
return (error);
}
ASSERT(error == 0 || error == EEXIST);
VERIFY0(ztest_dmu_objset_own(name, DMU_OST_OTHER, B_FALSE,
B_TRUE, zd, &os));
(void) pthread_rwlock_unlock(&ztest_name_lock);
ztest_zd_init(zd, ZTEST_GET_SHARED_DS(d), os);
zilog = zd->zd_zilog;
if (zilog->zl_header->zh_claim_lr_seq != 0 &&
zilog->zl_header->zh_claim_lr_seq < committed_seq)
fatal(B_FALSE, "missing log records: "
"claimed %"PRIu64" < committed %"PRIu64"",
zilog->zl_header->zh_claim_lr_seq, committed_seq);
ztest_dataset_dirobj_verify(zd);
zil_replay(os, zd, ztest_replay_vector);
ztest_dataset_dirobj_verify(zd);
if (ztest_opts.zo_verbose >= 6)
(void) printf("%s replay %"PRIu64" blocks, "
"%"PRIu64" records, seq %"PRIu64"\n",
zd->zd_name,
zilog->zl_parse_blk_count,
zilog->zl_parse_lr_count,
zilog->zl_replaying_seq);
zilog = zil_open(os, ztest_get_data, NULL);
if (zilog->zl_replaying_seq != 0 &&
zilog->zl_replaying_seq < committed_seq)
fatal(B_FALSE, "missing log records: "
"replayed %"PRIu64" < committed %"PRIu64"",
zilog->zl_replaying_seq, committed_seq);
return (0);
}
static void
ztest_dataset_close(int d)
{
ztest_ds_t *zd = &ztest_ds[d];
zil_close(zd->zd_zilog);
dmu_objset_disown(zd->zd_os, B_TRUE, zd);
ztest_zd_fini(zd);
}
static int
ztest_replay_zil_cb(const char *name, void *arg)
{
(void) arg;
objset_t *os;
ztest_ds_t *zdtmp;
VERIFY0(ztest_dmu_objset_own(name, DMU_OST_ANY, B_TRUE,
B_TRUE, FTAG, &os));
zdtmp = umem_alloc(sizeof (ztest_ds_t), UMEM_NOFAIL);
ztest_zd_init(zdtmp, NULL, os);
zil_replay(os, zdtmp, ztest_replay_vector);
ztest_zd_fini(zdtmp);
if (dmu_objset_zil(os)->zl_parse_lr_count != 0 &&
ztest_opts.zo_verbose >= 6) {
zilog_t *zilog = dmu_objset_zil(os);
(void) printf("%s replay %"PRIu64" blocks, "
"%"PRIu64" records, seq %"PRIu64"\n",
name,
zilog->zl_parse_blk_count,
zilog->zl_parse_lr_count,
zilog->zl_replaying_seq);
}
umem_free(zdtmp, sizeof (ztest_ds_t));
dmu_objset_disown(os, B_TRUE, FTAG);
return (0);
}
static void
ztest_freeze(void)
{
ztest_ds_t *zd = &ztest_ds[0];
spa_t *spa;
int numloops = 0;
+ /* freeze not supported during RAIDZ expansion */
+ if (ztest_opts.zo_raid_do_expand)
+ return;
+
if (ztest_opts.zo_verbose >= 3)
(void) printf("testing spa_freeze()...\n");
+ raidz_scratch_verify();
kernel_init(SPA_MODE_READ | SPA_MODE_WRITE);
VERIFY0(spa_open(ztest_opts.zo_pool, &spa, FTAG));
VERIFY0(ztest_dataset_open(0));
ztest_spa = spa;
/*
* Force the first log block to be transactionally allocated.
* We have to do this before we freeze the pool -- otherwise
* the log chain won't be anchored.
*/
while (BP_IS_HOLE(&zd->zd_zilog->zl_header->zh_log)) {
ztest_dmu_object_alloc_free(zd, 0);
zil_commit(zd->zd_zilog, 0);
}
txg_wait_synced(spa_get_dsl(spa), 0);
/*
* Freeze the pool. This stops spa_sync() from doing anything,
* so that the only way to record changes from now on is the ZIL.
*/
spa_freeze(spa);
/*
* Because it is hard to predict how much space a write will actually
* require beforehand, we leave ourselves some fudge space to write over
* capacity.
*/
uint64_t capacity = metaslab_class_get_space(spa_normal_class(spa)) / 2;
/*
* Run tests that generate log records but don't alter the pool config
* or depend on DSL sync tasks (snapshots, objset create/destroy, etc).
* We do a txg_wait_synced() after each iteration to force the txg
* to increase well beyond the last synced value in the uberblock.
* The ZIL should be OK with that.
*
* Run a random number of times less than zo_maxloops and ensure we do
* not run out of space on the pool.
*/
while (ztest_random(10) != 0 &&
numloops++ < ztest_opts.zo_maxloops &&
metaslab_class_get_alloc(spa_normal_class(spa)) < capacity) {
ztest_od_t od;
ztest_od_init(&od, 0, FTAG, 0, DMU_OT_UINT64_OTHER, 0, 0, 0);
VERIFY0(ztest_object_init(zd, &od, sizeof (od), B_FALSE));
ztest_io(zd, od.od_object,
ztest_random(ZTEST_RANGE_LOCKS) << SPA_MAXBLOCKSHIFT);
txg_wait_synced(spa_get_dsl(spa), 0);
}
/*
* Commit all of the changes we just generated.
*/
zil_commit(zd->zd_zilog, 0);
txg_wait_synced(spa_get_dsl(spa), 0);
/*
* Close our dataset and close the pool.
*/
ztest_dataset_close(0);
spa_close(spa, FTAG);
kernel_fini();
/*
* Open and close the pool and dataset to induce log replay.
*/
+ raidz_scratch_verify();
kernel_init(SPA_MODE_READ | SPA_MODE_WRITE);
VERIFY0(spa_open(ztest_opts.zo_pool, &spa, FTAG));
ASSERT3U(spa_freeze_txg(spa), ==, UINT64_MAX);
VERIFY0(ztest_dataset_open(0));
ztest_spa = spa;
txg_wait_synced(spa_get_dsl(spa), 0);
ztest_dataset_close(0);
ztest_reguid(NULL, 0);
spa_close(spa, FTAG);
kernel_fini();
}
static void
ztest_import_impl(void)
{
importargs_t args = { 0 };
nvlist_t *cfg = NULL;
int nsearch = 1;
char *searchdirs[nsearch];
int flags = ZFS_IMPORT_MISSING_LOG;
searchdirs[0] = ztest_opts.zo_dir;
args.paths = nsearch;
args.path = searchdirs;
args.can_be_active = B_FALSE;
libpc_handle_t lpch = {
.lpc_lib_handle = NULL,
.lpc_ops = &libzpool_config_ops,
.lpc_printerr = B_TRUE
};
VERIFY0(zpool_find_config(&lpch, ztest_opts.zo_pool, &cfg, &args));
VERIFY0(spa_import(ztest_opts.zo_pool, cfg, NULL, flags));
fnvlist_free(cfg);
}
/*
* Import a storage pool with the given name.
*/
static void
ztest_import(ztest_shared_t *zs)
{
spa_t *spa;
mutex_init(&ztest_vdev_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&ztest_checkpoint_lock, NULL, MUTEX_DEFAULT, NULL);
VERIFY0(pthread_rwlock_init(&ztest_name_lock, NULL));
+ raidz_scratch_verify();
kernel_init(SPA_MODE_READ | SPA_MODE_WRITE);
ztest_import_impl();
VERIFY0(spa_open(ztest_opts.zo_pool, &spa, FTAG));
zs->zs_metaslab_sz =
1ULL << spa->spa_root_vdev->vdev_child[0]->vdev_ms_shift;
zs->zs_guid = spa_guid(spa);
spa_close(spa, FTAG);
kernel_fini();
if (!ztest_opts.zo_mmp_test) {
ztest_run_zdb(zs->zs_guid);
ztest_freeze();
ztest_run_zdb(zs->zs_guid);
}
(void) pthread_rwlock_destroy(&ztest_name_lock);
mutex_destroy(&ztest_vdev_lock);
mutex_destroy(&ztest_checkpoint_lock);
}
/*
- * Kick off threads to run tests on all datasets in parallel.
+ * After the expansion was killed, check that the pool is healthy
+ */
+static void
+ztest_raidz_expand_check(spa_t *spa)
+{
+ ASSERT3U(ztest_opts.zo_raidz_expand_test, ==, RAIDZ_EXPAND_KILLED);
+ /*
+ * Set pool check done flag, main program will run a zdb check
+ * of the pool when we exit.
+ */
+ ztest_shared_opts->zo_raidz_expand_test = RAIDZ_EXPAND_CHECKED;
+
+ /* Wait for reflow to finish */
+ if (ztest_opts.zo_verbose >= 1) {
+ (void) printf("\nwaiting for reflow to finish ...\n");
+ }
+ pool_raidz_expand_stat_t rzx_stats;
+ pool_raidz_expand_stat_t *pres = &rzx_stats;
+ do {
+ txg_wait_synced(spa_get_dsl(spa), 0);
+ (void) poll(NULL, 0, 500); /* wait 1/2 second */
+
+ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
+ (void) spa_raidz_expand_get_stats(spa, pres);
+ spa_config_exit(spa, SCL_CONFIG, FTAG);
+ } while (pres->pres_state != DSS_FINISHED &&
+ pres->pres_reflowed < pres->pres_to_reflow);
+
+ if (ztest_opts.zo_verbose >= 1) {
+ (void) printf("verifying an interrupted raidz "
+ "expansion using a pool scrub ...\n");
+ }
+ /* Will fail here if there is non-recoverable corruption detected */
+ VERIFY0(ztest_scrub_impl(spa));
+ if (ztest_opts.zo_verbose >= 1) {
+ (void) printf("raidz expansion scrub check complete\n");
+ }
+}
+
+/*
+ * Start a raidz expansion test. We run some I/O on the pool for a while
+ * to get some data in the pool. Then we grow the raidz and
+ * kill the test at the requested offset into the reflow, verifying that
+ * doing such does not lead to pool corruption.
+ */
+static void
+ztest_raidz_expand_run(ztest_shared_t *zs, spa_t *spa)
+{
+ nvlist_t *root;
+ pool_raidz_expand_stat_t rzx_stats;
+ pool_raidz_expand_stat_t *pres = &rzx_stats;
+ kthread_t **run_threads;
+ vdev_t *cvd, *rzvd = spa->spa_root_vdev->vdev_child[0];
+ int total_disks = rzvd->vdev_children;
+ int data_disks = total_disks - vdev_get_nparity(rzvd);
+ uint64_t alloc_goal;
+ uint64_t csize;
+ int error, t;
+ int threads = ztest_opts.zo_threads;
+ ztest_expand_io_t *thread_args;
+
+ ASSERT3U(ztest_opts.zo_raidz_expand_test, !=, RAIDZ_EXPAND_NONE);
+ ASSERT3U(rzvd->vdev_ops, ==, &vdev_raidz_ops);
+ ztest_opts.zo_raidz_expand_test = RAIDZ_EXPAND_STARTED;
+
+ /* Setup a 1 MiB buffer of random data */
+ uint64_t bufsize = 1024 * 1024;
+ void *buffer = umem_alloc(bufsize, UMEM_NOFAIL);
+
+ if (read(ztest_fd_rand, buffer, bufsize) != bufsize) {
+ fatal(B_TRUE, "short read from /dev/urandom");
+ }
+ /*
+ * Put some data in the pool and then attach a vdev to initiate
+ * reflow.
+ */
+ run_threads = umem_zalloc(threads * sizeof (kthread_t *), UMEM_NOFAIL);
+ thread_args = umem_zalloc(threads * sizeof (ztest_expand_io_t),
+ UMEM_NOFAIL);
+ /* Aim for roughly 25% of allocatable space up to 1GB */
+ alloc_goal = (vdev_get_min_asize(rzvd) * data_disks) / total_disks;
+ alloc_goal = MIN(alloc_goal >> 2, 1024*1024*1024);
+ if (ztest_opts.zo_verbose >= 1) {
+ (void) printf("adding data to pool '%s', goal %llu bytes\n",
+ ztest_opts.zo_pool, (u_longlong_t)alloc_goal);
+ }
+
+ /*
+ * Kick off all the I/O generators that run in parallel.
+ */
+ for (t = 0; t < threads; t++) {
+ if (t < ztest_opts.zo_datasets && ztest_dataset_open(t) != 0) {
+ umem_free(run_threads, threads * sizeof (kthread_t *));
+ umem_free(buffer, bufsize);
+ return;
+ }
+ thread_args[t].rzx_id = t;
+ thread_args[t].rzx_amount = alloc_goal / threads;
+ thread_args[t].rzx_bufsize = bufsize;
+ thread_args[t].rzx_buffer = buffer;
+ thread_args[t].rzx_alloc_max = alloc_goal;
+ thread_args[t].rzx_spa = spa;
+ run_threads[t] = thread_create(NULL, 0, ztest_rzx_thread,
+ &thread_args[t], 0, NULL, TS_RUN | TS_JOINABLE,
+ defclsyspri);
+ }
+
+ /*
+ * Wait for all of the writers to complete.
+ */
+ for (t = 0; t < threads; t++)
+ VERIFY0(thread_join(run_threads[t]));
+
+ /*
+ * Close all datasets. This must be done after all the threads
+ * are joined so we can be sure none of the datasets are in-use
+ * by any of the threads.
+ */
+ for (t = 0; t < ztest_opts.zo_threads; t++) {
+ if (t < ztest_opts.zo_datasets)
+ ztest_dataset_close(t);
+ }
+
+ txg_wait_synced(spa_get_dsl(spa), 0);
+
+ zs->zs_alloc = metaslab_class_get_alloc(spa_normal_class(spa));
+ zs->zs_space = metaslab_class_get_space(spa_normal_class(spa));
+
+ umem_free(buffer, bufsize);
+ umem_free(run_threads, threads * sizeof (kthread_t *));
+ umem_free(thread_args, threads * sizeof (ztest_expand_io_t));
+
+ /* Set our reflow target to 25%, 50% or 75% of allocated size */
+ uint_t multiple = ztest_random(3) + 1;
+ uint64_t reflow_max = (rzvd->vdev_stat.vs_alloc * multiple) / 4;
+ raidz_expand_max_reflow_bytes = reflow_max;
+
+ if (ztest_opts.zo_verbose >= 1) {
+ (void) printf("running raidz expansion test, killing when "
+ "reflow reaches %llu bytes (%u/4 of allocated space)\n",
+ (u_longlong_t)reflow_max, multiple);
+ }
+
+ /* XXX - do we want some I/O load during the reflow? */
+
+ /*
+ * Use a disk size that is larger than existing ones
+ */
+ cvd = rzvd->vdev_child[0];
+ csize = vdev_get_min_asize(cvd);
+ csize += csize / 10;
+ /*
+ * Path to vdev to be attached
+ */
+ char *newpath = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
+ (void) snprintf(newpath, MAXPATHLEN, ztest_dev_template,
+ ztest_opts.zo_dir, ztest_opts.zo_pool, rzvd->vdev_children);
+ /*
+ * Build the nvlist describing newpath.
+ */
+ root = make_vdev_root(newpath, NULL, NULL, csize, ztest_get_ashift(),
+ NULL, 0, 0, 1);
+ /*
+ * Expand the raidz vdev by attaching the new disk
+ */
+ if (ztest_opts.zo_verbose >= 1) {
+ (void) printf("expanding raidz: %d wide to %d wide with '%s'\n",
+ (int)rzvd->vdev_children, (int)rzvd->vdev_children + 1,
+ newpath);
+ }
+ error = spa_vdev_attach(spa, rzvd->vdev_guid, root, B_FALSE, B_FALSE);
+ nvlist_free(root);
+ if (error != 0) {
+ fatal(0, "raidz expand: attach (%s %llu) returned %d",
+ newpath, (long long)csize, error);
+ }
+
+ /*
+ * Wait for reflow to begin
+ */
+ while (spa->spa_raidz_expand == NULL) {
+ txg_wait_synced(spa_get_dsl(spa), 0);
+ (void) poll(NULL, 0, 100); /* wait 1/10 second */
+ }
+ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
+ (void) spa_raidz_expand_get_stats(spa, pres);
+ spa_config_exit(spa, SCL_CONFIG, FTAG);
+ while (pres->pres_state != DSS_SCANNING) {
+ txg_wait_synced(spa_get_dsl(spa), 0);
+ (void) poll(NULL, 0, 100); /* wait 1/10 second */
+ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
+ (void) spa_raidz_expand_get_stats(spa, pres);
+ spa_config_exit(spa, SCL_CONFIG, FTAG);
+ }
+
+ ASSERT3U(pres->pres_state, ==, DSS_SCANNING);
+ ASSERT3U(pres->pres_to_reflow, !=, 0);
+ /*
+ * Set so when we are killed we go to raidz checking rather than
+ * restarting test.
+ */
+ ztest_shared_opts->zo_raidz_expand_test = RAIDZ_EXPAND_KILLED;
+ if (ztest_opts.zo_verbose >= 1) {
+ (void) printf("raidz expansion reflow started, waiting for "
+ "%llu bytes to be copied\n", (u_longlong_t)reflow_max);
+ }
+
+ /*
+ * Wait for reflow maximum to be reached and then kill the test
+ */
+ while (pres->pres_reflowed < reflow_max) {
+ txg_wait_synced(spa_get_dsl(spa), 0);
+ (void) poll(NULL, 0, 100); /* wait 1/10 second */
+ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
+ (void) spa_raidz_expand_get_stats(spa, pres);
+ spa_config_exit(spa, SCL_CONFIG, FTAG);
+ }
+
+ /* Reset the reflow pause before killing */
+ raidz_expand_max_reflow_bytes = 0;
+
+ if (ztest_opts.zo_verbose >= 1) {
+ (void) printf("killing raidz expansion test after reflow "
+ "reached %llu bytes\n", (u_longlong_t)pres->pres_reflowed);
+ }
+
+ /*
+ * Kill ourself to simulate a panic during a reflow. Our parent will
+ * restart the test and the changed flag value will drive the test
+ * through the scrub/check code to verify the pool is not corrupted.
+ */
+ ztest_kill(zs);
+}
+
+static void
+ztest_generic_run(ztest_shared_t *zs, spa_t *spa)
+{
+ kthread_t **run_threads;
+ int t;
+
+ run_threads = umem_zalloc(ztest_opts.zo_threads * sizeof (kthread_t *),
+ UMEM_NOFAIL);
+
+ /*
+ * Kick off all the tests that run in parallel.
+ */
+ for (t = 0; t < ztest_opts.zo_threads; t++) {
+ if (t < ztest_opts.zo_datasets && ztest_dataset_open(t) != 0) {
+ umem_free(run_threads, ztest_opts.zo_threads *
+ sizeof (kthread_t *));
+ return;
+ }
+
+ run_threads[t] = thread_create(NULL, 0, ztest_thread,
+ (void *)(uintptr_t)t, 0, NULL, TS_RUN | TS_JOINABLE,
+ defclsyspri);
+ }
+
+ /*
+ * Wait for all of the tests to complete.
+ */
+ for (t = 0; t < ztest_opts.zo_threads; t++)
+ VERIFY0(thread_join(run_threads[t]));
+
+ /*
+ * Close all datasets. This must be done after all the threads
+ * are joined so we can be sure none of the datasets are in-use
+ * by any of the threads.
+ */
+ for (t = 0; t < ztest_opts.zo_threads; t++) {
+ if (t < ztest_opts.zo_datasets)
+ ztest_dataset_close(t);
+ }
+
+ txg_wait_synced(spa_get_dsl(spa), 0);
+
+ zs->zs_alloc = metaslab_class_get_alloc(spa_normal_class(spa));
+ zs->zs_space = metaslab_class_get_space(spa_normal_class(spa));
+
+ umem_free(run_threads, ztest_opts.zo_threads * sizeof (kthread_t *));
+}
+
+/*
+ * Setup our test context and kick off threads to run tests on all datasets
+ * in parallel.
*/
static void
ztest_run(ztest_shared_t *zs)
{
spa_t *spa;
objset_t *os;
kthread_t *resume_thread, *deadman_thread;
- kthread_t **run_threads;
uint64_t object;
int error;
int t, d;
ztest_exiting = B_FALSE;
/*
* Initialize parent/child shared state.
*/
mutex_init(&ztest_vdev_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&ztest_checkpoint_lock, NULL, MUTEX_DEFAULT, NULL);
VERIFY0(pthread_rwlock_init(&ztest_name_lock, NULL));
zs->zs_thread_start = gethrtime();
zs->zs_thread_stop =
zs->zs_thread_start + ztest_opts.zo_passtime * NANOSEC;
zs->zs_thread_stop = MIN(zs->zs_thread_stop, zs->zs_proc_stop);
zs->zs_thread_kill = zs->zs_thread_stop;
if (ztest_random(100) < ztest_opts.zo_killrate) {
zs->zs_thread_kill -=
ztest_random(ztest_opts.zo_passtime * NANOSEC);
}
mutex_init(&zcl.zcl_callbacks_lock, NULL, MUTEX_DEFAULT, NULL);
list_create(&zcl.zcl_callbacks, sizeof (ztest_cb_data_t),
offsetof(ztest_cb_data_t, zcd_node));
/*
* Open our pool. It may need to be imported first depending on
* what tests were running when the previous pass was terminated.
*/
+ raidz_scratch_verify();
kernel_init(SPA_MODE_READ | SPA_MODE_WRITE);
error = spa_open(ztest_opts.zo_pool, &spa, FTAG);
if (error) {
VERIFY3S(error, ==, ENOENT);
ztest_import_impl();
VERIFY0(spa_open(ztest_opts.zo_pool, &spa, FTAG));
zs->zs_metaslab_sz =
1ULL << spa->spa_root_vdev->vdev_child[0]->vdev_ms_shift;
}
metaslab_preload_limit = ztest_random(20) + 1;
ztest_spa = spa;
- VERIFY0(vdev_raidz_impl_set("cycle"));
+ /*
+ * XXX - BUGBUG raidz expansion do not run this for generic for now
+ */
+ if (ztest_opts.zo_raidz_expand_test != RAIDZ_EXPAND_NONE)
+ VERIFY0(vdev_raidz_impl_set("cycle"));
dmu_objset_stats_t dds;
VERIFY0(ztest_dmu_objset_own(ztest_opts.zo_pool,
DMU_OST_ANY, B_TRUE, B_TRUE, FTAG, &os));
dsl_pool_config_enter(dmu_objset_pool(os), FTAG);
dmu_objset_fast_stat(os, &dds);
dsl_pool_config_exit(dmu_objset_pool(os), FTAG);
dmu_objset_disown(os, B_TRUE, FTAG);
+ /* Give the dedicated raidz expansion test more grace time */
+ if (ztest_opts.zo_raidz_expand_test != RAIDZ_EXPAND_NONE)
+ zfs_deadman_synctime_ms *= 2;
+
/*
* Create a thread to periodically resume suspended I/O.
*/
resume_thread = thread_create(NULL, 0, ztest_resume_thread,
spa, 0, NULL, TS_RUN | TS_JOINABLE, defclsyspri);
/*
* Create a deadman thread and set to panic if we hang.
*/
deadman_thread = thread_create(NULL, 0, ztest_deadman_thread,
zs, 0, NULL, TS_RUN | TS_JOINABLE, defclsyspri);
spa->spa_deadman_failmode = ZIO_FAILURE_MODE_PANIC;
/*
* Verify that we can safely inquire about any object,
* whether it's allocated or not. To make it interesting,
* we probe a 5-wide window around each power of two.
* This hits all edge cases, including zero and the max.
*/
for (t = 0; t < 64; t++) {
for (d = -5; d <= 5; d++) {
error = dmu_object_info(spa->spa_meta_objset,
(1ULL << t) + d, NULL);
ASSERT(error == 0 || error == ENOENT ||
error == EINVAL);
}
}
/*
* If we got any ENOSPC errors on the previous run, destroy something.
*/
if (zs->zs_enospc_count != 0) {
+ /* Not expecting ENOSPC errors during raidz expansion tests */
+ ASSERT3U(ztest_opts.zo_raidz_expand_test, ==,
+ RAIDZ_EXPAND_NONE);
+
int d = ztest_random(ztest_opts.zo_datasets);
ztest_dataset_destroy(d);
}
zs->zs_enospc_count = 0;
/*
* If we were in the middle of ztest_device_removal() and were killed
* we need to ensure the removal and scrub complete before running
* any tests that check ztest_device_removal_active. The removal will
* be restarted automatically when the spa is opened, but we need to
* initiate the scrub manually if it is not already in progress. Note
* that we always run the scrub whenever an indirect vdev exists
* because we have no way of knowing for sure if ztest_device_removal()
* fully completed its scrub before the pool was reimported.
+ *
+ * Does not apply for the RAIDZ expansion specific test runs
*/
- if (spa->spa_removing_phys.sr_state == DSS_SCANNING ||
- spa->spa_removing_phys.sr_prev_indirect_vdev != -1) {
+ if (ztest_opts.zo_raidz_expand_test == RAIDZ_EXPAND_NONE &&
+ (spa->spa_removing_phys.sr_state == DSS_SCANNING ||
+ spa->spa_removing_phys.sr_prev_indirect_vdev != -1)) {
while (spa->spa_removing_phys.sr_state == DSS_SCANNING)
txg_wait_synced(spa_get_dsl(spa), 0);
error = ztest_scrub_impl(spa);
if (error == EBUSY)
error = 0;
ASSERT0(error);
}
- run_threads = umem_zalloc(ztest_opts.zo_threads * sizeof (kthread_t *),
- UMEM_NOFAIL);
-
if (ztest_opts.zo_verbose >= 4)
(void) printf("starting main threads...\n");
/*
* Replay all logs of all datasets in the pool. This is primarily for
* temporary datasets which wouldn't otherwise get replayed, which
* can trigger failures when attempting to offline a SLOG in
* ztest_fault_inject().
*/
(void) dmu_objset_find(ztest_opts.zo_pool, ztest_replay_zil_cb,
NULL, DS_FIND_CHILDREN);
- /*
- * Kick off all the tests that run in parallel.
- */
- for (t = 0; t < ztest_opts.zo_threads; t++) {
- if (t < ztest_opts.zo_datasets && ztest_dataset_open(t) != 0) {
- umem_free(run_threads, ztest_opts.zo_threads *
- sizeof (kthread_t *));
- return;
- }
-
- run_threads[t] = thread_create(NULL, 0, ztest_thread,
- (void *)(uintptr_t)t, 0, NULL, TS_RUN | TS_JOINABLE,
- defclsyspri);
- }
-
- /*
- * Wait for all of the tests to complete.
- */
- for (t = 0; t < ztest_opts.zo_threads; t++)
- VERIFY0(thread_join(run_threads[t]));
-
- /*
- * Close all datasets. This must be done after all the threads
- * are joined so we can be sure none of the datasets are in-use
- * by any of the threads.
- */
- for (t = 0; t < ztest_opts.zo_threads; t++) {
- if (t < ztest_opts.zo_datasets)
- ztest_dataset_close(t);
- }
-
- txg_wait_synced(spa_get_dsl(spa), 0);
-
- zs->zs_alloc = metaslab_class_get_alloc(spa_normal_class(spa));
- zs->zs_space = metaslab_class_get_space(spa_normal_class(spa));
-
- umem_free(run_threads, ztest_opts.zo_threads * sizeof (kthread_t *));
+ if (ztest_opts.zo_raidz_expand_test == RAIDZ_EXPAND_REQUESTED)
+ ztest_raidz_expand_run(zs, spa);
+ else if (ztest_opts.zo_raidz_expand_test == RAIDZ_EXPAND_KILLED)
+ ztest_raidz_expand_check(spa);
+ else
+ ztest_generic_run(zs, spa);
/* Kill the resume and deadman threads */
ztest_exiting = B_TRUE;
VERIFY0(thread_join(resume_thread));
VERIFY0(thread_join(deadman_thread));
ztest_resume(spa);
/*
* Right before closing the pool, kick off a bunch of async I/O;
* spa_close() should wait for it to complete.
*/
for (object = 1; object < 50; object++) {
dmu_prefetch(spa->spa_meta_objset, object, 0, 0, 1ULL << 20,
ZIO_PRIORITY_SYNC_READ);
}
/* Verify that at least one commit cb was called in a timely fashion */
if (zc_cb_counter >= ZTEST_COMMIT_CB_MIN_REG)
VERIFY0(zc_min_txg_delay);
spa_close(spa, FTAG);
/*
* Verify that we can loop over all pools.
*/
mutex_enter(&spa_namespace_lock);
for (spa = spa_next(NULL); spa != NULL; spa = spa_next(spa))
if (ztest_opts.zo_verbose > 3)
(void) printf("spa_next: found %s\n", spa_name(spa));
mutex_exit(&spa_namespace_lock);
/*
* Verify that we can export the pool and reimport it under a
* different name.
*/
if ((ztest_random(2) == 0) && !ztest_opts.zo_mmp_test) {
char name[ZFS_MAX_DATASET_NAME_LEN];
(void) snprintf(name, sizeof (name), "%s_import",
ztest_opts.zo_pool);
ztest_spa_import_export(ztest_opts.zo_pool, name);
ztest_spa_import_export(name, ztest_opts.zo_pool);
}
kernel_fini();
list_destroy(&zcl.zcl_callbacks);
mutex_destroy(&zcl.zcl_callbacks_lock);
(void) pthread_rwlock_destroy(&ztest_name_lock);
mutex_destroy(&ztest_vdev_lock);
mutex_destroy(&ztest_checkpoint_lock);
}
static void
print_time(hrtime_t t, char *timebuf)
{
hrtime_t s = t / NANOSEC;
hrtime_t m = s / 60;
hrtime_t h = m / 60;
hrtime_t d = h / 24;
s -= m * 60;
m -= h * 60;
h -= d * 24;
timebuf[0] = '\0';
if (d)
(void) sprintf(timebuf,
"%llud%02lluh%02llum%02llus", d, h, m, s);
else if (h)
(void) sprintf(timebuf, "%lluh%02llum%02llus", h, m, s);
else if (m)
(void) sprintf(timebuf, "%llum%02llus", m, s);
else
(void) sprintf(timebuf, "%llus", s);
}
static nvlist_t *
make_random_props(void)
{
nvlist_t *props;
props = fnvlist_alloc();
if (ztest_random(2) == 0)
return (props);
fnvlist_add_uint64(props,
zpool_prop_to_name(ZPOOL_PROP_AUTOREPLACE), 1);
return (props);
}
/*
* Create a storage pool with the given name and initial vdev size.
* Then test spa_freeze() functionality.
*/
static void
ztest_init(ztest_shared_t *zs)
{
spa_t *spa;
nvlist_t *nvroot, *props;
int i;
mutex_init(&ztest_vdev_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&ztest_checkpoint_lock, NULL, MUTEX_DEFAULT, NULL);
VERIFY0(pthread_rwlock_init(&ztest_name_lock, NULL));
+ raidz_scratch_verify();
kernel_init(SPA_MODE_READ | SPA_MODE_WRITE);
/*
* Create the storage pool.
*/
(void) spa_destroy(ztest_opts.zo_pool);
ztest_shared->zs_vdev_next_leaf = 0;
zs->zs_splits = 0;
zs->zs_mirrors = ztest_opts.zo_mirrors;
nvroot = make_vdev_root(NULL, NULL, NULL, ztest_opts.zo_vdev_size, 0,
NULL, ztest_opts.zo_raid_children, zs->zs_mirrors, 1);
props = make_random_props();
/*
* We don't expect the pool to suspend unless maxfaults == 0,
* in which case ztest_fault_inject() temporarily takes away
* the only valid replica.
*/
fnvlist_add_uint64(props,
zpool_prop_to_name(ZPOOL_PROP_FAILUREMODE),
MAXFAULTS(zs) ? ZIO_FAILURE_MODE_PANIC : ZIO_FAILURE_MODE_WAIT);
for (i = 0; i < SPA_FEATURES; i++) {
char *buf;
if (!spa_feature_table[i].fi_zfs_mod_supported)
continue;
/*
* 75% chance of using the log space map feature. We want ztest
* to exercise both the code paths that use the log space map
* feature and the ones that don't.
*/
if (i == SPA_FEATURE_LOG_SPACEMAP && ztest_random(4) == 0)
continue;
VERIFY3S(-1, !=, asprintf(&buf, "feature@%s",
spa_feature_table[i].fi_uname));
fnvlist_add_uint64(props, buf, 0);
free(buf);
}
VERIFY0(spa_create(ztest_opts.zo_pool, nvroot, props, NULL, NULL));
fnvlist_free(nvroot);
fnvlist_free(props);
VERIFY0(spa_open(ztest_opts.zo_pool, &spa, FTAG));
zs->zs_metaslab_sz =
1ULL << spa->spa_root_vdev->vdev_child[0]->vdev_ms_shift;
zs->zs_guid = spa_guid(spa);
spa_close(spa, FTAG);
kernel_fini();
if (!ztest_opts.zo_mmp_test) {
ztest_run_zdb(zs->zs_guid);
ztest_freeze();
ztest_run_zdb(zs->zs_guid);
}
(void) pthread_rwlock_destroy(&ztest_name_lock);
mutex_destroy(&ztest_vdev_lock);
mutex_destroy(&ztest_checkpoint_lock);
}
static void
setup_data_fd(void)
{
static char ztest_name_data[] = "/tmp/ztest.data.XXXXXX";
ztest_fd_data = mkstemp(ztest_name_data);
ASSERT3S(ztest_fd_data, >=, 0);
(void) unlink(ztest_name_data);
}
static int
shared_data_size(ztest_shared_hdr_t *hdr)
{
int size;
size = hdr->zh_hdr_size;
size += hdr->zh_opts_size;
size += hdr->zh_size;
size += hdr->zh_stats_size * hdr->zh_stats_count;
size += hdr->zh_ds_size * hdr->zh_ds_count;
+ size += hdr->zh_scratch_state_size;
return (size);
}
static void
setup_hdr(void)
{
int size;
ztest_shared_hdr_t *hdr;
hdr = (void *)mmap(0, P2ROUNDUP(sizeof (*hdr), getpagesize()),
PROT_READ | PROT_WRITE, MAP_SHARED, ztest_fd_data, 0);
ASSERT3P(hdr, !=, MAP_FAILED);
VERIFY0(ftruncate(ztest_fd_data, sizeof (ztest_shared_hdr_t)));
hdr->zh_hdr_size = sizeof (ztest_shared_hdr_t);
hdr->zh_opts_size = sizeof (ztest_shared_opts_t);
hdr->zh_size = sizeof (ztest_shared_t);
hdr->zh_stats_size = sizeof (ztest_shared_callstate_t);
hdr->zh_stats_count = ZTEST_FUNCS;
hdr->zh_ds_size = sizeof (ztest_shared_ds_t);
hdr->zh_ds_count = ztest_opts.zo_datasets;
+ hdr->zh_scratch_state_size = sizeof (ztest_shared_scratch_state_t);
size = shared_data_size(hdr);
VERIFY0(ftruncate(ztest_fd_data, size));
(void) munmap((caddr_t)hdr, P2ROUNDUP(sizeof (*hdr), getpagesize()));
}
static void
setup_data(void)
{
int size, offset;
ztest_shared_hdr_t *hdr;
uint8_t *buf;
hdr = (void *)mmap(0, P2ROUNDUP(sizeof (*hdr), getpagesize()),
PROT_READ, MAP_SHARED, ztest_fd_data, 0);
ASSERT3P(hdr, !=, MAP_FAILED);
size = shared_data_size(hdr);
(void) munmap((caddr_t)hdr, P2ROUNDUP(sizeof (*hdr), getpagesize()));
hdr = ztest_shared_hdr = (void *)mmap(0, P2ROUNDUP(size, getpagesize()),
PROT_READ | PROT_WRITE, MAP_SHARED, ztest_fd_data, 0);
ASSERT3P(hdr, !=, MAP_FAILED);
buf = (uint8_t *)hdr;
offset = hdr->zh_hdr_size;
ztest_shared_opts = (void *)&buf[offset];
offset += hdr->zh_opts_size;
ztest_shared = (void *)&buf[offset];
offset += hdr->zh_size;
ztest_shared_callstate = (void *)&buf[offset];
offset += hdr->zh_stats_size * hdr->zh_stats_count;
ztest_shared_ds = (void *)&buf[offset];
+ offset += hdr->zh_ds_size * hdr->zh_ds_count;
+ ztest_scratch_state = (void *)&buf[offset];
}
static boolean_t
exec_child(char *cmd, char *libpath, boolean_t ignorekill, int *statusp)
{
pid_t pid;
int status;
char *cmdbuf = NULL;
pid = fork();
if (cmd == NULL) {
cmdbuf = umem_alloc(MAXPATHLEN, UMEM_NOFAIL);
(void) strlcpy(cmdbuf, getexecname(), MAXPATHLEN);
cmd = cmdbuf;
}
if (pid == -1)
fatal(B_TRUE, "fork failed");
if (pid == 0) { /* child */
char fd_data_str[12];
VERIFY3S(11, >=,
snprintf(fd_data_str, 12, "%d", ztest_fd_data));
VERIFY0(setenv("ZTEST_FD_DATA", fd_data_str, 1));
if (libpath != NULL) {
const char *curlp = getenv("LD_LIBRARY_PATH");
if (curlp == NULL)
VERIFY0(setenv("LD_LIBRARY_PATH", libpath, 1));
else {
char *newlp = NULL;
VERIFY3S(-1, !=,
asprintf(&newlp, "%s:%s", libpath, curlp));
VERIFY0(setenv("LD_LIBRARY_PATH", newlp, 1));
free(newlp);
}
}
(void) execl(cmd, cmd, (char *)NULL);
ztest_dump_core = B_FALSE;
fatal(B_TRUE, "exec failed: %s", cmd);
}
if (cmdbuf != NULL) {
umem_free(cmdbuf, MAXPATHLEN);
cmd = NULL;
}
while (waitpid(pid, &status, 0) != pid)
continue;
if (statusp != NULL)
*statusp = status;
if (WIFEXITED(status)) {
if (WEXITSTATUS(status) != 0) {
(void) fprintf(stderr, "child exited with code %d\n",
WEXITSTATUS(status));
exit(2);
}
return (B_FALSE);
} else if (WIFSIGNALED(status)) {
if (!ignorekill || WTERMSIG(status) != SIGKILL) {
(void) fprintf(stderr, "child died with signal %d\n",
WTERMSIG(status));
exit(3);
}
return (B_TRUE);
} else {
(void) fprintf(stderr, "something strange happened to child\n");
exit(4);
}
}
static void
ztest_run_init(void)
{
int i;
ztest_shared_t *zs = ztest_shared;
/*
* Blow away any existing copy of zpool.cache
*/
(void) remove(spa_config_path);
if (ztest_opts.zo_init == 0) {
if (ztest_opts.zo_verbose >= 1)
(void) printf("Importing pool %s\n",
ztest_opts.zo_pool);
ztest_import(zs);
return;
}
/*
* Create and initialize our storage pool.
*/
for (i = 1; i <= ztest_opts.zo_init; i++) {
memset(zs, 0, sizeof (*zs));
if (ztest_opts.zo_verbose >= 3 &&
ztest_opts.zo_init != 1) {
(void) printf("ztest_init(), pass %d\n", i);
}
ztest_init(zs);
}
}
int
main(int argc, char **argv)
{
int kills = 0;
int iters = 0;
int older = 0;
int newer = 0;
ztest_shared_t *zs;
ztest_info_t *zi;
ztest_shared_callstate_t *zc;
char timebuf[100];
char numbuf[NN_NUMBUF_SZ];
char *cmd;
boolean_t hasalt;
int f, err;
char *fd_data_str = getenv("ZTEST_FD_DATA");
struct sigaction action;
(void) setvbuf(stdout, NULL, _IOLBF, 0);
dprintf_setup(&argc, argv);
zfs_deadman_synctime_ms = 300000;
zfs_deadman_checktime_ms = 30000;
/*
* As two-word space map entries may not come up often (especially
* if pool and vdev sizes are small) we want to force at least some
* of them so the feature get tested.
*/
zfs_force_some_double_word_sm_entries = B_TRUE;
/*
* Verify that even extensively damaged split blocks with many
* segments can be reconstructed in a reasonable amount of time
* when reconstruction is known to be possible.
*
* Note: the lower this value is, the more damage we inflict, and
* the more time ztest spends in recovering that damage. We chose
* to induce damage 1/100th of the time so recovery is tested but
* not so frequently that ztest doesn't get to test other code paths.
*/
zfs_reconstruct_indirect_damage_fraction = 100;
action.sa_handler = sig_handler;
sigemptyset(&action.sa_mask);
action.sa_flags = 0;
if (sigaction(SIGSEGV, &action, NULL) < 0) {
(void) fprintf(stderr, "ztest: cannot catch SIGSEGV: %s.\n",
strerror(errno));
exit(EXIT_FAILURE);
}
if (sigaction(SIGABRT, &action, NULL) < 0) {
(void) fprintf(stderr, "ztest: cannot catch SIGABRT: %s.\n",
strerror(errno));
exit(EXIT_FAILURE);
}
/*
* Force random_get_bytes() to use /dev/urandom in order to prevent
* ztest from needlessly depleting the system entropy pool.
*/
random_path = "/dev/urandom";
ztest_fd_rand = open(random_path, O_RDONLY | O_CLOEXEC);
ASSERT3S(ztest_fd_rand, >=, 0);
if (!fd_data_str) {
process_options(argc, argv);
setup_data_fd();
setup_hdr();
setup_data();
memcpy(ztest_shared_opts, &ztest_opts,
sizeof (*ztest_shared_opts));
} else {
ztest_fd_data = atoi(fd_data_str);
setup_data();
memcpy(&ztest_opts, ztest_shared_opts, sizeof (ztest_opts));
}
ASSERT3U(ztest_opts.zo_datasets, ==, ztest_shared_hdr->zh_ds_count);
err = ztest_set_global_vars();
if (err != 0 && !fd_data_str) {
/* error message done by ztest_set_global_vars */
exit(EXIT_FAILURE);
} else {
/* children should not be spawned if setting gvars fails */
VERIFY3S(err, ==, 0);
}
/* Override location of zpool.cache */
VERIFY3S(asprintf((char **)&spa_config_path, "%s/zpool.cache",
ztest_opts.zo_dir), !=, -1);
ztest_ds = umem_alloc(ztest_opts.zo_datasets * sizeof (ztest_ds_t),
UMEM_NOFAIL);
zs = ztest_shared;
if (fd_data_str) {
metaslab_force_ganging = ztest_opts.zo_metaslab_force_ganging;
metaslab_df_alloc_threshold =
zs->zs_metaslab_df_alloc_threshold;
if (zs->zs_do_init)
ztest_run_init();
else
ztest_run(zs);
exit(0);
}
hasalt = (strlen(ztest_opts.zo_alt_ztest) != 0);
if (ztest_opts.zo_verbose >= 1) {
- (void) printf("%"PRIu64" vdevs, %d datasets, %d threads,"
- "%d %s disks, %"PRIu64" seconds...\n\n",
+ (void) printf("%"PRIu64" vdevs, %d datasets, %d threads, "
+ "%d %s disks, parity %d, %"PRIu64" seconds...\n\n",
ztest_opts.zo_vdevs,
ztest_opts.zo_datasets,
ztest_opts.zo_threads,
ztest_opts.zo_raid_children,
ztest_opts.zo_raid_type,
+ ztest_opts.zo_raid_parity,
ztest_opts.zo_time);
}
cmd = umem_alloc(MAXNAMELEN, UMEM_NOFAIL);
(void) strlcpy(cmd, getexecname(), MAXNAMELEN);
zs->zs_do_init = B_TRUE;
if (strlen(ztest_opts.zo_alt_ztest) != 0) {
if (ztest_opts.zo_verbose >= 1) {
(void) printf("Executing older ztest for "
"initialization: %s\n", ztest_opts.zo_alt_ztest);
}
VERIFY(!exec_child(ztest_opts.zo_alt_ztest,
ztest_opts.zo_alt_libpath, B_FALSE, NULL));
} else {
VERIFY(!exec_child(NULL, NULL, B_FALSE, NULL));
}
zs->zs_do_init = B_FALSE;
zs->zs_proc_start = gethrtime();
zs->zs_proc_stop = zs->zs_proc_start + ztest_opts.zo_time * NANOSEC;
for (f = 0; f < ZTEST_FUNCS; f++) {
zi = &ztest_info[f];
zc = ZTEST_GET_SHARED_CALLSTATE(f);
if (zs->zs_proc_start + zi->zi_interval[0] > zs->zs_proc_stop)
zc->zc_next = UINT64_MAX;
else
zc->zc_next = zs->zs_proc_start +
ztest_random(2 * zi->zi_interval[0] + 1);
}
/*
* Run the tests in a loop. These tests include fault injection
* to verify that self-healing data works, and forced crashes
* to verify that we never lose on-disk consistency.
*/
while (gethrtime() < zs->zs_proc_stop) {
int status;
boolean_t killed;
/*
* Initialize the workload counters for each function.
*/
for (f = 0; f < ZTEST_FUNCS; f++) {
zc = ZTEST_GET_SHARED_CALLSTATE(f);
zc->zc_count = 0;
zc->zc_time = 0;
}
/* Set the allocation switch size */
zs->zs_metaslab_df_alloc_threshold =
ztest_random(zs->zs_metaslab_sz / 4) + 1;
if (!hasalt || ztest_random(2) == 0) {
if (hasalt && ztest_opts.zo_verbose >= 1) {
(void) printf("Executing newer ztest: %s\n",
cmd);
}
newer++;
killed = exec_child(cmd, NULL, B_TRUE, &status);
} else {
if (hasalt && ztest_opts.zo_verbose >= 1) {
(void) printf("Executing older ztest: %s\n",
ztest_opts.zo_alt_ztest);
}
older++;
killed = exec_child(ztest_opts.zo_alt_ztest,
ztest_opts.zo_alt_libpath, B_TRUE, &status);
}
if (killed)
kills++;
iters++;
if (ztest_opts.zo_verbose >= 1) {
hrtime_t now = gethrtime();
now = MIN(now, zs->zs_proc_stop);
print_time(zs->zs_proc_stop - now, timebuf);
nicenum(zs->zs_space, numbuf, sizeof (numbuf));
(void) printf("Pass %3d, %8s, %3"PRIu64" ENOSPC, "
"%4.1f%% of %5s used, %3.0f%% done, %8s to go\n",
iters,
WIFEXITED(status) ? "Complete" : "SIGKILL",
zs->zs_enospc_count,
100.0 * zs->zs_alloc / zs->zs_space,
numbuf,
100.0 * (now - zs->zs_proc_start) /
(ztest_opts.zo_time * NANOSEC), timebuf);
}
if (ztest_opts.zo_verbose >= 2) {
(void) printf("\nWorkload summary:\n\n");
(void) printf("%7s %9s %s\n",
"Calls", "Time", "Function");
(void) printf("%7s %9s %s\n",
"-----", "----", "--------");
for (f = 0; f < ZTEST_FUNCS; f++) {
zi = &ztest_info[f];
zc = ZTEST_GET_SHARED_CALLSTATE(f);
print_time(zc->zc_time, timebuf);
(void) printf("%7"PRIu64" %9s %s\n",
zc->zc_count, timebuf,
zi->zi_funcname);
}
(void) printf("\n");
}
if (!ztest_opts.zo_mmp_test)
ztest_run_zdb(zs->zs_guid);
+ if (ztest_shared_opts->zo_raidz_expand_test ==
+ RAIDZ_EXPAND_CHECKED)
+ break; /* raidz expand test complete */
}
if (ztest_opts.zo_verbose >= 1) {
if (hasalt) {
(void) printf("%d runs of older ztest: %s\n", older,
ztest_opts.zo_alt_ztest);
(void) printf("%d runs of newer ztest: %s\n", newer,
cmd);
}
(void) printf("%d killed, %d completed, %.0f%% kill rate\n",
kills, iters - kills, (100.0 * kills) / MAX(1, iters));
}
umem_free(cmd, MAXNAMELEN);
return (0);
}
diff --git a/sys/contrib/openzfs/contrib/debian/control b/sys/contrib/openzfs/contrib/debian/control
index f4e97fe16145..98beb900d0fa 100644
--- a/sys/contrib/openzfs/contrib/debian/control
+++ b/sys/contrib/openzfs/contrib/debian/control
@@ -1,327 +1,326 @@
Source: openzfs-linux
Section: contrib/kernel
Priority: optional
Maintainer: ZFS on Linux specific mailing list <zfs-discuss@list.zfsonlinux.org>
Build-Depends: debhelper-compat (= 12),
dh-python,
dh-sequence-dkms | dkms (>> 2.1.1.2-5),
libaio-dev,
libblkid-dev,
libcurl4-openssl-dev,
libelf-dev,
libpam0g-dev,
libssl-dev | libssl1.0-dev,
libtool,
libudev-dev,
lsb-release,
po-debconf,
python3-all-dev,
python3-cffi,
python3-setuptools,
python3-sphinx,
uuid-dev,
zlib1g-dev
Standards-Version: 4.5.1
Homepage: https://openzfs.org/
Vcs-Git: https://github.com/openzfs/zfs.git
Vcs-Browser: https://github.com/openzfs/zfs
Rules-Requires-Root: no
XS-Autobuild: yes
Package: openzfs-libnvpair3
Section: contrib/libs
Architecture: linux-any
Depends: ${misc:Depends}, ${shlibs:Depends}
Breaks: libnvpair1, libnvpair3
Replaces: libnvpair1, libnvpair3, libnvpair3linux
Conflicts: libnvpair3linux
Description: Solaris name-value library for Linux
This library provides routines for packing and unpacking nv pairs for
transporting data across process boundaries, transporting between
kernel and userland, and possibly saving onto disk files.
Package: openzfs-libpam-zfs
Section: contrib/admin
Architecture: linux-any
Depends: libpam-runtime, ${misc:Depends}, ${shlibs:Depends}
Replaces: libpam-zfs
Conflicts: libpam-zfs
Description: PAM module for managing encryption keys for ZFS
OpenZFS is a storage platform that encompasses the functionality of
traditional filesystems and volume managers. It supports data checksums,
compression, encryption, snapshots, and more.
.
This provides a Pluggable Authentication Module (PAM) that automatically
unlocks encrypted ZFS datasets upon login.
Package: openzfs-libuutil3
Section: contrib/libs
Architecture: linux-any
Depends: ${misc:Depends}, ${shlibs:Depends}
Breaks: libuutil1, libuutil3
Replaces: libuutil1, libuutil3, libuutil3linux
Conflicts: libuutil3linux
Description: Solaris userland utility library for Linux
This library provides a variety of glue functions for ZFS on Linux:
* libspl: The Solaris Porting Layer userland library, which provides APIs
that make it possible to run Solaris user code in a Linux environment
with relatively minimal modification.
* libavl: The Adelson-Velskii Landis balanced binary tree manipulation
library.
* libefi: The Extensible Firmware Interface library for GUID disk
partitioning.
* libshare: NFS, SMB, and iSCSI service integration for ZFS.
Package: openzfs-libzfs-dev
Section: contrib/libdevel
Architecture: linux-any
Depends: libssl-dev | libssl1.0-dev,
openzfs-libnvpair3 (= ${binary:Version}),
openzfs-libuutil3 (= ${binary:Version}),
openzfs-libzfs4 (= ${binary:Version}),
openzfs-libzfsbootenv1 (= ${binary:Version}),
openzfs-libzpool5 (= ${binary:Version}),
${misc:Depends}
Replaces: libzfslinux-dev
Conflicts: libzfslinux-dev
Provides: libnvpair-dev, libuutil-dev
Description: OpenZFS filesystem development files for Linux
Header files and static libraries for compiling software against
libraries of OpenZFS filesystem.
.
This package includes the development files of libnvpair3, libuutil3,
libzpool5 and libzfs4.
Package: openzfs-libzfs4
Section: contrib/libs
Architecture: linux-any
Depends: ${misc:Depends}, ${shlibs:Depends}
# The libcurl4 is loaded through dlopen("libcurl.so.4").
# https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=988521
Recommends: libcurl4
Breaks: libzfs2, libzfs4
Replaces: libzfs2, libzfs4, libzfs4linux
Conflicts: libzfs4linux
Description: OpenZFS filesystem library for Linux - general support
OpenZFS is a storage platform that encompasses the functionality of
traditional filesystems and volume managers. It supports data checksums,
compression, encryption, snapshots, and more.
.
The OpenZFS library provides support for managing OpenZFS filesystems.
Package: openzfs-libzfsbootenv1
Section: contrib/libs
Architecture: linux-any
Depends: ${misc:Depends}, ${shlibs:Depends}
Breaks: libzfs2, libzfs4
Replaces: libzfs2, libzfs4, libzfsbootenv1linux
Conflicts: libzfsbootenv1linux
Description: OpenZFS filesystem library for Linux - label info support
OpenZFS is a storage platform that encompasses the functionality of
traditional filesystems and volume managers. It supports data checksums,
compression, encryption, snapshots, and more.
.
The zfsbootenv library provides support for modifying ZFS label information.
Package: openzfs-libzpool5
Section: contrib/libs
Architecture: linux-any
Depends: ${misc:Depends}, ${shlibs:Depends}
Breaks: libzpool2, libzpool5
Replaces: libzpool2, libzpool5, libzpool5linux
Conflicts: libzpool5linux
Description: OpenZFS pool library for Linux
OpenZFS is a storage platform that encompasses the functionality of
traditional filesystems and volume managers. It supports data checksums,
compression, encryption, snapshots, and more.
.
This zpool library provides support for managing zpools.
Package: openzfs-python3-pyzfs
Section: contrib/python
Architecture: linux-any
Depends: python3-cffi,
openzfs-zfsutils (= ${binary:Version}),
${misc:Depends},
${python3:Depends}
Replaces: python3-pyzfs
Conflicts: python3-pyzfs
Description: wrapper for libzfs_core C library
libzfs_core is intended to be a stable interface for programmatic
administration of ZFS. This wrapper provides one-to-one wrappers for
libzfs_core API functions, but the signatures and types are more natural to
Python.
.
nvlists are wrapped as dictionaries or lists depending on their usage.
Some parameters have default values depending on typical use for
increased convenience. Enumerations and bit flags become strings and lists
of strings in Python. Errors are reported as exceptions rather than integer
errno-style error codes. The wrapper takes care to provide one-to-many
mapping of the error codes to the exceptions by interpreting a context
in which the error code is produced.
Package: openzfs-pyzfs-doc
Section: contrib/doc
Architecture: all
Depends: ${misc:Depends}, ${sphinxdoc:Depends}
Recommends: openzfs-python3-pyzfs
Replaces: pyzfs-doc
Conflicts: pyzfs-doc
Description: wrapper for libzfs_core C library (documentation)
libzfs_core is intended to be a stable interface for programmatic
administration of ZFS. This wrapper provides one-to-one wrappers for
libzfs_core API functions, but the signatures and types are more natural to
Python.
.
nvlists are wrapped as dictionaries or lists depending on their usage.
Some parameters have default values depending on typical use for
increased convenience. Enumerations and bit flags become strings and lists
of strings in Python. Errors are reported as exceptions rather than integer
errno-style error codes. The wrapper takes care to provide one-to-many
mapping of the error codes to the exceptions by interpreting a context
in which the error code is produced.
.
This package contains the documentation.
Package: openzfs-zfs-dkms
Architecture: all
Depends: dkms (>> 2.1.1.2-5),
file,
libc6-dev | libc-dev,
lsb-release,
python3-distutils | libpython3-stdlib (<< 3.6.4),
${misc:Depends},
${perl:Depends}
Recommends: openzfs-zfs-zed, openzfs-zfsutils (>= ${source:Version}), ${linux:Recommends}
# suggests debhelper because e.g. `dkms mkdeb -m zfs -v 0.8.2` needs dh_testdir (#909183)
Suggests: debhelper
Breaks: spl-dkms (<< 0.8.0~rc1)
Replaces: spl-dkms, zfs-dkms
-Conflicts: zfs-dkms
Provides: openzfs-zfs-modules
Description: OpenZFS filesystem kernel modules for Linux
OpenZFS is a storage platform that encompasses the functionality of
traditional filesystems and volume managers. It supports data checksums,
compression, encryption, snapshots, and more.
.
This DKMS package includes the SPA, DMU, ZVOL, and ZPL components of
OpenZFS.
Package: openzfs-zfs-initramfs
Architecture: all
Depends: busybox-initramfs | busybox-static | busybox,
initramfs-tools,
openzfs-zfs-modules | openzfs-zfs-dkms,
openzfs-zfsutils (>= ${source:Version}),
${misc:Depends}
Breaks: zfsutils-linux (<= 0.7.11-2)
Replaces: zfsutils-linux (<= 0.7.11-2), zfs-initramfs
Conflicts: zfs-initramfs
Description: OpenZFS root filesystem capabilities for Linux - initramfs
OpenZFS is a storage platform that encompasses the functionality of
traditional filesystems and volume managers. It supports data checksums,
compression, encryption, snapshots, and more.
.
This package adds OpenZFS to the system initramfs with a hook
for the initramfs-tools infrastructure.
Package: openzfs-zfs-dracut
Architecture: all
Depends: dracut,
openzfs-zfs-modules | openzfs-zfs-dkms,
openzfs-zfsutils (>= ${source:Version}),
${misc:Depends}
Conflicts: zfs-dracut
Replaces: zfs-dracut
Description: OpenZFS root filesystem capabilities for Linux - dracut
OpenZFS is a storage platform that encompasses the functionality of
traditional filesystems and volume managers. It supports data checksums,
compression, encryption, snapshots, and more.
.
This package adds OpenZFS to the system initramfs with a hook
for the dracut infrastructure.
Package: openzfs-zfsutils
Section: contrib/admin
Architecture: linux-any
Pre-Depends: ${misc:Pre-Depends}
Depends: openzfs-libnvpair3 (= ${binary:Version}),
openzfs-libuutil3 (= ${binary:Version}),
openzfs-libzfs4 (= ${binary:Version}),
openzfs-libzpool5 (= ${binary:Version}),
python3,
${misc:Depends},
${shlibs:Depends}
Recommends: lsb-base, openzfs-zfs-modules | openzfs-zfs-dkms, openzfs-zfs-zed
Breaks: openrc,
spl (<< 0.7.9-2),
spl-dkms (<< 0.8.0~rc1),
openzfs-zfs-dkms (<< ${source:Version}),
openzfs-zfs-dkms (>> ${source:Version}...)
Replaces: spl (<< 0.7.9-2), spl-dkms, zfsutils-linux
Conflicts: zfs, zfs-fuse, zfsutils-linux
Suggests: nfs-kernel-server,
samba-common-bin (>= 3.0.23),
openzfs-zfs-initramfs | openzfs-zfs-dracut
Provides: openzfsutils
Description: command-line tools to manage OpenZFS filesystems
OpenZFS is a storage platform that encompasses the functionality of
traditional filesystems and volume managers. It supports data checksums,
compression, encryption, snapshots, and more.
.
This package provides the zfs and zpool commands to create and administer
OpenZFS filesystems.
Package: openzfs-zfs-zed
Section: contrib/admin
Architecture: linux-any
Pre-Depends: ${misc:Pre-Depends}
Depends: openzfs-zfs-modules | openzfs-zfs-dkms,
openzfs-zfsutils (>= ${binary:Version}),
${misc:Depends},
${shlibs:Depends}
Conflicts: zfs, zfs-zed
Replaces: zfs-zed
Description: OpenZFS Event Daemon
OpenZFS is a storage platform that encompasses the functionality of
traditional filesystems and volume managers. It supports data checksums,
compression, encryption, snapshots, and more.
.
ZED (ZFS Event Daemon) monitors events generated by the ZFS kernel
module. When a zevent (ZFS Event) is posted, ZED will run any ZEDLETs
(ZFS Event Daemon Linkage for Executable Tasks) that have been enabled
for the corresponding zevent class.
.
This package provides the OpenZFS Event Daemon (zed).
Package: openzfs-zfs-test
Section: contrib/admin
Architecture: linux-any
Depends: acl,
attr,
bc,
fio,
ksh,
lsscsi,
mdadm,
parted,
python3,
openzfs-python3-pyzfs,
sudo,
sysstat,
openzfs-zfs-modules | openzfs-zfs-dkms,
openzfs-zfsutils (>=${binary:Version}),
${misc:Depends},
${shlibs:Depends}
Recommends: nfs-kernel-server
Breaks: zfsutils-linux (<= 0.7.9-2)
Replaces: zfsutils-linux (<= 0.7.9-2), zfs-test
Conflicts: zutils, zfs-test
Description: OpenZFS test infrastructure and support scripts
OpenZFS is a storage platform that encompasses the functionality of
traditional filesystems and volume managers. It supports data checksums,
compression, encryption, snapshots, and more.
.
This package provides the OpenZFS test infrastructure for destructively
testing and validating a system using OpenZFS. It is entirely optional
and should only be installed and used in test environments.
diff --git a/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/_constants.py b/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/_constants.py
index 5ee422dfa803..9c40ece1a7df 100644
--- a/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/_constants.py
+++ b/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/_constants.py
@@ -1,119 +1,121 @@
#
# Copyright 2015 ClusterHQ
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
"""
Important `libzfs_core` constants.
"""
from __future__ import absolute_import, division, print_function
import errno
import sys
# Compat for platform-specific errnos
if sys.platform.startswith('freebsd'):
ECHRNG = errno.ENXIO
ECKSUM = 97 # EINTEGRITY
ETIME = errno.ETIMEDOUT
else:
ECHRNG = errno.ECHRNG
ECKSUM = errno.EBADE
ETIME = errno.ETIME
# https://stackoverflow.com/a/1695250
def enum_with_offset(offset, sequential, named):
enums = dict(((b, a + offset) for a, b in enumerate(sequential)), **named)
return type('Enum', (), enums)
def enum(*sequential, **named):
return enum_with_offset(0, sequential, named)
#: Maximum length of any ZFS name.
MAXNAMELEN = 255
#: Default channel program limits
ZCP_DEFAULT_INSTRLIMIT = 10 * 1000 * 1000
ZCP_DEFAULT_MEMLIMIT = 10 * 1024 * 1024
#: Encryption wrapping key length
WRAPPING_KEY_LEN = 32
#: Encryption key location enum
zfs_key_location = enum(
'ZFS_KEYLOCATION_NONE',
'ZFS_KEYLOCATION_PROMPT',
'ZFS_KEYLOCATION_URI'
)
#: Encryption key format enum
zfs_keyformat = enum(
'ZFS_KEYFORMAT_NONE',
'ZFS_KEYFORMAT_RAW',
'ZFS_KEYFORMAT_HEX',
'ZFS_KEYFORMAT_PASSPHRASE'
)
# Encryption algorithms enum
zio_encrypt = enum(
'ZIO_CRYPT_INHERIT',
'ZIO_CRYPT_ON',
'ZIO_CRYPT_OFF',
'ZIO_CRYPT_AES_128_CCM',
'ZIO_CRYPT_AES_192_CCM',
'ZIO_CRYPT_AES_256_CCM',
'ZIO_CRYPT_AES_128_GCM',
'ZIO_CRYPT_AES_192_GCM',
'ZIO_CRYPT_AES_256_GCM'
)
# ZFS-specific error codes
zfs_errno = enum_with_offset(1024, [
'ZFS_ERR_CHECKPOINT_EXISTS',
'ZFS_ERR_DISCARDING_CHECKPOINT',
'ZFS_ERR_NO_CHECKPOINT',
'ZFS_ERR_DEVRM_IN_PROGRESS',
'ZFS_ERR_VDEV_TOO_BIG',
'ZFS_ERR_IOC_CMD_UNAVAIL',
'ZFS_ERR_IOC_ARG_UNAVAIL',
'ZFS_ERR_IOC_ARG_REQUIRED',
'ZFS_ERR_IOC_ARG_BADTYPE',
'ZFS_ERR_WRONG_PARENT',
'ZFS_ERR_FROM_IVSET_GUID_MISSING',
'ZFS_ERR_FROM_IVSET_GUID_MISMATCH',
'ZFS_ERR_SPILL_BLOCK_FLAG_MISSING',
'ZFS_ERR_UNKNOWN_SEND_STREAM_FEATURE',
'ZFS_ERR_EXPORT_IN_PROGRESS',
'ZFS_ERR_BOOKMARK_SOURCE_NOT_ANCESTOR',
'ZFS_ERR_STREAM_TRUNCATED',
'ZFS_ERR_STREAM_LARGE_BLOCK_MISMATCH',
'ZFS_ERR_RESILVER_IN_PROGRESS',
'ZFS_ERR_REBUILD_IN_PROGRESS',
'ZFS_ERR_BADPROP',
'ZFS_ERR_VDEV_NOTSUP',
'ZFS_ERR_NOT_USER_NAMESPACE',
'ZFS_ERR_RESUME_EXISTS',
'ZFS_ERR_CRYPTO_NOTSUP',
+ 'ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS',
],
{}
)
# compat before we used the enum helper for these values
ZFS_ERR_CHECKPOINT_EXISTS = zfs_errno.ZFS_ERR_CHECKPOINT_EXISTS
assert (ZFS_ERR_CHECKPOINT_EXISTS == 1024)
ZFS_ERR_DISCARDING_CHECKPOINT = zfs_errno.ZFS_ERR_DISCARDING_CHECKPOINT
ZFS_ERR_NO_CHECKPOINT = zfs_errno.ZFS_ERR_NO_CHECKPOINT
ZFS_ERR_DEVRM_IN_PROGRESS = zfs_errno.ZFS_ERR_DEVRM_IN_PROGRESS
ZFS_ERR_VDEV_TOO_BIG = zfs_errno.ZFS_ERR_VDEV_TOO_BIG
ZFS_ERR_WRONG_PARENT = zfs_errno.ZFS_ERR_WRONG_PARENT
ZFS_ERR_VDEV_NOTSUP = zfs_errno.ZFS_ERR_VDEV_NOTSUP
+ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS = zfs_errno.ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS
# vim: softtabstop=4 tabstop=4 expandtab shiftwidth=4
diff --git a/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/_error_translation.py b/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/_error_translation.py
index 26676db398c5..3d1a2d573e39 100644
--- a/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/_error_translation.py
+++ b/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/_error_translation.py
@@ -1,842 +1,845 @@
#
# Copyright 2015 ClusterHQ
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
"""
Helper routines for converting ``errno`` style error codes from C functions
to Python exceptions defined by `libzfs_core` API.
The conversion heavily depends on the context of the error: the attempted
operation and the input parameters. For this reason, there is a conversion
routine for each `libzfs_core` interface function. The conversion routines
have the return code as a parameter as well as all the parameters of the
corresponding interface functions.
The parameters and exceptions are documented in the `libzfs_core` interfaces.
"""
from __future__ import absolute_import, division, print_function
import errno
import re
import string
from . import exceptions as lzc_exc
from ._constants import (
ECHRNG,
ECKSUM,
ETIME,
MAXNAMELEN,
ZFS_ERR_CHECKPOINT_EXISTS,
ZFS_ERR_DISCARDING_CHECKPOINT,
ZFS_ERR_NO_CHECKPOINT,
ZFS_ERR_DEVRM_IN_PROGRESS,
ZFS_ERR_VDEV_TOO_BIG,
ZFS_ERR_WRONG_PARENT,
+ ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS,
zfs_errno
)
def lzc_create_translate_error(ret, name, ds_type, props):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(name)
raise lzc_exc.PropertyInvalid(name)
if ret == errno.EEXIST:
raise lzc_exc.FilesystemExists(name)
if ret == errno.ENOENT:
raise lzc_exc.ParentNotFound(name)
if ret == ZFS_ERR_WRONG_PARENT:
raise lzc_exc.WrongParent(_fs_name(name))
if ret == zfs_errno.ZFS_ERR_BADPROP:
raise lzc_exc.PropertyInvalid(name)
raise _generic_exception(ret, name, "Failed to create filesystem")
def lzc_clone_translate_error(ret, name, origin, props):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(name)
_validate_snap_name(origin)
raise lzc_exc.PropertyInvalid(name)
if ret == errno.EXDEV:
raise lzc_exc.PoolsDiffer(name)
if ret == errno.EEXIST:
raise lzc_exc.FilesystemExists(name)
if ret == errno.ENOENT:
if not _is_valid_snap_name(origin):
raise lzc_exc.SnapshotNameInvalid(origin)
raise lzc_exc.DatasetNotFound(name)
raise _generic_exception(ret, name, "Failed to create clone")
def lzc_rollback_translate_error(ret, name):
if ret == 0:
return
if ret == errno.ESRCH:
raise lzc_exc.SnapshotNotFound(name)
if ret == errno.EINVAL:
_validate_fs_name(name)
raise lzc_exc.NameInvalid(name)
if ret == errno.ENOENT:
if not _is_valid_fs_name(name):
raise lzc_exc.NameInvalid(name)
else:
raise lzc_exc.FilesystemNotFound(name)
raise _generic_exception(ret, name, "Failed to rollback")
def lzc_rollback_to_translate_error(ret, name, snap):
if ret == errno.EEXIST:
raise lzc_exc.SnapshotNotLatest(snap)
else:
lzc_rollback_translate_error(ret, name)
def lzc_snapshot_translate_errors(ret, errlist, snaps, props):
if ret == 0:
return
def _map(ret, name):
if ret == errno.EXDEV:
pool_names = iter(map(_pool_name, snaps))
pool_name = next(pool_names, None)
same_pool = all(x == pool_name for x in pool_names)
if same_pool:
return lzc_exc.DuplicateSnapshots(name)
else:
return lzc_exc.PoolsDiffer(name)
elif ret == errno.EINVAL:
if any(not _is_valid_snap_name(s) for s in snaps):
return lzc_exc.NameInvalid(name)
elif any(len(s) > MAXNAMELEN for s in snaps):
return lzc_exc.NameTooLong(name)
else:
return lzc_exc.PropertyInvalid(name)
if ret == errno.EEXIST:
return lzc_exc.SnapshotExists(name)
if ret == errno.ENOENT:
return lzc_exc.FilesystemNotFound(name)
return _generic_exception(ret, name, "Failed to create snapshot")
_handle_err_list(ret, errlist, snaps, lzc_exc.SnapshotFailure, _map)
def lzc_destroy_snaps_translate_errors(ret, errlist, snaps, defer):
if ret == 0:
return
def _map(ret, name):
if ret == errno.EEXIST:
return lzc_exc.SnapshotIsCloned(name)
if ret == errno.ENOENT:
return lzc_exc.PoolNotFound(name)
if ret == errno.EBUSY:
return lzc_exc.SnapshotIsHeld(name)
return _generic_exception(ret, name, "Failed to destroy snapshot")
_handle_err_list(
ret, errlist, snaps, lzc_exc.SnapshotDestructionFailure, _map)
def lzc_bookmark_translate_errors(ret, errlist, bookmarks):
if ret == 0:
return
def _map(ret, name):
source = bookmarks[name]
if ret == errno.EINVAL:
if name:
pool_names = map(_pool_name, bookmarks.keys())
# use _validate* functions for MAXNAMELEN check
try:
_validate_bmark_name(name)
except lzc_exc.ZFSError as e:
return e
try:
_validate_snap_name(source)
source_is_snap = True
except lzc_exc.ZFSError:
source_is_snap = False
try:
_validate_bmark_name(source)
source_is_bmark = True
except lzc_exc.ZFSError:
source_is_bmark = False
if not source_is_snap and not source_is_bmark:
return lzc_exc.BookmarkSourceInvalid(source)
if any(x != _pool_name(name) for x in pool_names):
return lzc_exc.PoolsDiffer(name)
else:
invalid_names = [
b for b in bookmarks.keys() if not _is_valid_bmark_name(b)]
if invalid_names:
return lzc_exc.BookmarkNameInvalid(invalid_names[0])
if ret == errno.EEXIST:
return lzc_exc.BookmarkExists(name)
if ret == errno.ENOENT:
return lzc_exc.SnapshotNotFound(name)
if ret == errno.ENOTSUP:
return lzc_exc.BookmarkNotSupported(name)
if ret == zfs_errno.ZFS_ERR_BOOKMARK_SOURCE_NOT_ANCESTOR:
return lzc_exc.BookmarkMismatch(source)
return _generic_exception(ret, name, "Failed to create bookmark")
_handle_err_list(
ret, errlist, bookmarks.keys(), lzc_exc.BookmarkFailure, _map)
def lzc_get_bookmarks_translate_error(ret, fsname, props):
if ret == 0:
return
if ret == errno.ENOENT:
raise lzc_exc.FilesystemNotFound(fsname)
raise _generic_exception(ret, fsname, "Failed to list bookmarks")
def lzc_destroy_bookmarks_translate_errors(ret, errlist, bookmarks):
if ret == 0:
return
def _map(ret, name):
if ret == errno.EINVAL:
return lzc_exc.NameInvalid(name)
return _generic_exception(ret, name, "Failed to destroy bookmark")
_handle_err_list(
ret, errlist, bookmarks, lzc_exc.BookmarkDestructionFailure, _map)
def lzc_snaprange_space_translate_error(ret, firstsnap, lastsnap):
if ret == 0:
return
if ret == errno.EXDEV and firstsnap is not None:
if _pool_name(firstsnap) != _pool_name(lastsnap):
raise lzc_exc.PoolsDiffer(lastsnap)
else:
raise lzc_exc.SnapshotMismatch(lastsnap)
if ret == errno.EINVAL:
if not _is_valid_snap_name(firstsnap):
raise lzc_exc.NameInvalid(firstsnap)
elif not _is_valid_snap_name(lastsnap):
raise lzc_exc.NameInvalid(lastsnap)
elif len(firstsnap) > MAXNAMELEN:
raise lzc_exc.NameTooLong(firstsnap)
elif len(lastsnap) > MAXNAMELEN:
raise lzc_exc.NameTooLong(lastsnap)
elif _pool_name(firstsnap) != _pool_name(lastsnap):
raise lzc_exc.PoolsDiffer(lastsnap)
else:
raise lzc_exc.SnapshotMismatch(lastsnap)
if ret == errno.ENOENT:
raise lzc_exc.SnapshotNotFound(lastsnap)
raise _generic_exception(
ret, lastsnap, "Failed to calculate space used by range of snapshots")
def lzc_hold_translate_errors(ret, errlist, holds, fd):
if ret == 0:
return
def _map(ret, name):
if ret == errno.EXDEV:
return lzc_exc.PoolsDiffer(name)
elif ret == errno.EINVAL:
if name:
pool_names = map(_pool_name, holds.keys())
if not _is_valid_snap_name(name):
return lzc_exc.NameInvalid(name)
elif len(name) > MAXNAMELEN:
return lzc_exc.NameTooLong(name)
elif any(x != _pool_name(name) for x in pool_names):
return lzc_exc.PoolsDiffer(name)
else:
invalid_names = [
b for b in holds.keys() if not _is_valid_snap_name(b)]
if invalid_names:
return lzc_exc.NameInvalid(invalid_names[0])
fs_name = None
hold_name = None
pool_name = None
if name is not None:
fs_name = _fs_name(name)
pool_name = _pool_name(name)
hold_name = holds[name]
if ret == errno.ENOENT:
return lzc_exc.FilesystemNotFound(fs_name)
if ret == errno.EEXIST:
return lzc_exc.HoldExists(name)
if ret == errno.E2BIG:
return lzc_exc.NameTooLong(hold_name)
if ret == errno.ENOTSUP:
return lzc_exc.FeatureNotSupported(pool_name)
return _generic_exception(ret, name, "Failed to hold snapshot")
if ret == errno.EBADF:
raise lzc_exc.BadHoldCleanupFD()
_handle_err_list(ret, errlist, holds.keys(), lzc_exc.HoldFailure, _map)
def lzc_release_translate_errors(ret, errlist, holds):
if ret == 0:
return
for snap in holds:
hold_list = holds[snap]
if not isinstance(hold_list, list):
raise lzc_exc.TypeError('holds must be in a list')
def _map(ret, name):
if ret == errno.EXDEV:
return lzc_exc.PoolsDiffer(name)
elif ret == errno.EINVAL:
if name:
pool_names = map(_pool_name, holds.keys())
if not _is_valid_snap_name(name):
return lzc_exc.NameInvalid(name)
elif len(name) > MAXNAMELEN:
return lzc_exc.NameTooLong(name)
elif any(x != _pool_name(name) for x in pool_names):
return lzc_exc.PoolsDiffer(name)
else:
invalid_names = [
b for b in holds.keys() if not _is_valid_snap_name(b)]
if invalid_names:
return lzc_exc.NameInvalid(invalid_names[0])
elif ret == errno.ENOENT:
return lzc_exc.HoldNotFound(name)
elif ret == errno.E2BIG:
tag_list = holds[name]
too_long_tags = [t for t in tag_list if len(t) > MAXNAMELEN]
return lzc_exc.NameTooLong(too_long_tags[0])
elif ret == errno.ENOTSUP:
pool_name = None
if name is not None:
pool_name = _pool_name(name)
return lzc_exc.FeatureNotSupported(pool_name)
else:
return _generic_exception(
ret, name, "Failed to release snapshot hold")
_handle_err_list(
ret, errlist, holds.keys(), lzc_exc.HoldReleaseFailure, _map)
def lzc_get_holds_translate_error(ret, snapname):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_snap_name(snapname)
if ret == errno.ENOENT:
raise lzc_exc.SnapshotNotFound(snapname)
if ret == errno.ENOTSUP:
raise lzc_exc.FeatureNotSupported(_pool_name(snapname))
raise _generic_exception(ret, snapname, "Failed to get holds on snapshot")
def lzc_send_translate_error(ret, snapname, fromsnap, fd, flags):
if ret == 0:
return
if ret == errno.EXDEV and fromsnap is not None:
if _pool_name(fromsnap) != _pool_name(snapname):
raise lzc_exc.PoolsDiffer(snapname)
else:
raise lzc_exc.SnapshotMismatch(snapname)
elif ret == errno.EINVAL:
if (fromsnap is not None and not _is_valid_snap_name(fromsnap) and
not _is_valid_bmark_name(fromsnap)):
raise lzc_exc.NameInvalid(fromsnap)
elif (not _is_valid_snap_name(snapname) and
not _is_valid_fs_name(snapname)):
raise lzc_exc.NameInvalid(snapname)
elif fromsnap is not None and len(fromsnap) > MAXNAMELEN:
raise lzc_exc.NameTooLong(fromsnap)
elif len(snapname) > MAXNAMELEN:
raise lzc_exc.NameTooLong(snapname)
elif (fromsnap is not None and
_pool_name(fromsnap) != _pool_name(snapname)):
raise lzc_exc.PoolsDiffer(snapname)
elif ret == errno.ENOENT:
if (fromsnap is not None and not _is_valid_snap_name(fromsnap) and
not _is_valid_bmark_name(fromsnap)):
raise lzc_exc.NameInvalid(fromsnap)
raise lzc_exc.SnapshotNotFound(snapname)
elif ret == errno.ENAMETOOLONG:
if fromsnap is not None and len(fromsnap) > MAXNAMELEN:
raise lzc_exc.NameTooLong(fromsnap)
else:
raise lzc_exc.NameTooLong(snapname)
raise lzc_exc.StreamIOError(ret)
def lzc_send_space_translate_error(ret, snapname, fromsnap):
if ret == 0:
return
if ret == errno.EXDEV and fromsnap is not None:
if _pool_name(fromsnap) != _pool_name(snapname):
raise lzc_exc.PoolsDiffer(snapname)
else:
raise lzc_exc.SnapshotMismatch(snapname)
elif ret == errno.EINVAL:
if fromsnap is not None and not _is_valid_snap_name(fromsnap):
raise lzc_exc.NameInvalid(fromsnap)
elif not _is_valid_snap_name(snapname):
raise lzc_exc.NameInvalid(snapname)
elif fromsnap is not None and len(fromsnap) > MAXNAMELEN:
raise lzc_exc.NameTooLong(fromsnap)
elif len(snapname) > MAXNAMELEN:
raise lzc_exc.NameTooLong(snapname)
elif (fromsnap is not None and
_pool_name(fromsnap) != _pool_name(snapname)):
raise lzc_exc.PoolsDiffer(snapname)
elif ret == errno.ENOENT and fromsnap is not None:
if not _is_valid_snap_name(fromsnap):
raise lzc_exc.NameInvalid(fromsnap)
if ret == errno.ENOENT:
raise lzc_exc.SnapshotNotFound(snapname)
raise _generic_exception(
ret, snapname, "Failed to estimate backup stream size")
def lzc_receive_translate_errors(
ret, snapname, fd, force, raw, resumable, embedded, origin, properrs
):
if ret == 0:
if properrs is not None and len(properrs) > 0:
def _map(ret, name):
if ret == errno.EINVAL:
return lzc_exc.PropertyInvalid(name)
if ret == zfs_errno.ZFS_ERR_BADPROP:
return lzc_exc.PropertyInvalid(name)
return _generic_exception(ret, name, "Failed to set property")
_handle_err_list(
errno.EINVAL, properrs, [snapname],
lzc_exc.ReceivePropertyFailure, _map)
else:
return
if ret == errno.EINVAL:
if (not _is_valid_snap_name(snapname) and
not _is_valid_fs_name(snapname)):
raise lzc_exc.NameInvalid(snapname)
elif len(snapname) > MAXNAMELEN:
raise lzc_exc.NameTooLong(snapname)
elif origin is not None and not _is_valid_snap_name(origin):
raise lzc_exc.NameInvalid(origin)
elif resumable:
raise lzc_exc.StreamFeatureInvalid()
elif embedded and not raw:
raise lzc_exc.StreamFeatureIncompatible()
else:
raise lzc_exc.BadStream()
if ret == errno.ENOENT:
if not _is_valid_snap_name(snapname):
raise lzc_exc.NameInvalid(snapname)
else:
raise lzc_exc.DatasetNotFound(snapname)
if ret == errno.EEXIST:
raise lzc_exc.DatasetExists(snapname)
if ret == errno.ENOTSUP:
raise lzc_exc.StreamFeatureNotSupported()
if ret == errno.ENODEV:
raise lzc_exc.StreamMismatch(_fs_name(snapname))
if ret == errno.ETXTBSY:
raise lzc_exc.DestinationModified(_fs_name(snapname))
if ret == errno.EBUSY:
raise lzc_exc.DatasetBusy(_fs_name(snapname))
if ret == errno.ENOSPC:
raise lzc_exc.NoSpace(_fs_name(snapname))
if ret == errno.EDQUOT:
raise lzc_exc.QuotaExceeded(_fs_name(snapname))
if ret == errno.ENAMETOOLONG:
raise lzc_exc.NameTooLong(snapname)
if ret == errno.EROFS:
raise lzc_exc.ReadOnlyPool(_pool_name(snapname))
if ret == errno.EAGAIN:
raise lzc_exc.SuspendedPool(_pool_name(snapname))
if ret == errno.EACCES:
raise lzc_exc.EncryptionKeyNotLoaded()
if ret == ECKSUM:
raise lzc_exc.BadStream()
if ret == ZFS_ERR_WRONG_PARENT:
raise lzc_exc.WrongParent(_fs_name(snapname))
if ret == zfs_errno.ZFS_ERR_STREAM_TRUNCATED:
raise lzc_exc.StreamTruncated()
if ret == zfs_errno.ZFS_ERR_BADPROP:
raise lzc_exc.PropertyInvalid(snapname)
raise lzc_exc.StreamIOError(ret)
def lzc_promote_translate_error(ret, name):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(name)
raise lzc_exc.NotClone(name)
if ret == errno.ENOTSOCK:
raise lzc_exc.NotClone(name)
if ret == errno.ENOENT:
raise lzc_exc.FilesystemNotFound(name)
if ret == errno.EEXIST:
raise lzc_exc.SnapshotExists(name)
raise _generic_exception(ret, name, "Failed to promote dataset")
def lzc_change_key_translate_error(ret, name):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(name)
raise lzc_exc.PropertyInvalid(name)
if ret == errno.ENOENT:
raise lzc_exc.FilesystemNotFound(name)
if ret == errno.EACCES:
raise lzc_exc.EncryptionKeyNotLoaded()
raise _generic_exception(ret, name, "Failed to change encryption key")
def lzc_load_key_translate_error(ret, name, noop):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(name)
raise lzc_exc.PropertyInvalid(name)
if ret == errno.ENOENT:
raise lzc_exc.FilesystemNotFound(name)
if ret == errno.EACCES:
raise lzc_exc.EncryptionKeyInvalid()
if ret == errno.EEXIST:
raise lzc_exc.EncryptionKeyAlreadyLoaded()
if noop:
raise _generic_exception(ret, name, "Failed to load encryption key")
else:
raise _generic_exception(ret, name, "Failed to verify encryption key")
def lzc_unload_key_translate_error(ret, name):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(name)
raise lzc_exc.PropertyInvalid(name)
if ret == errno.ENOENT:
raise lzc_exc.FilesystemNotFound(name)
if ret == errno.EACCES:
raise lzc_exc.EncryptionKeyNotLoaded()
raise _generic_exception(ret, name, "Failed to unload encryption key")
def lzc_sync_translate_error(ret, name):
if ret == 0:
return
if ret == errno.ENOENT:
raise lzc_exc.PoolNotFound(name)
raise _generic_exception(ret, name, "Failed to sync pool")
def lzc_reopen_translate_error(ret, name):
if ret == 0:
return
if ret == errno.ENOENT:
raise lzc_exc.PoolNotFound(name)
raise _generic_exception(ret, name, "Failed to reopen pool")
def lzc_channel_program_translate_error(ret, name, error):
if ret == 0:
return
if ret == errno.ENOENT:
raise lzc_exc.PoolNotFound(name)
if ret == ETIME:
raise lzc_exc.ZCPTimeout()
if ret == errno.ENOMEM:
raise lzc_exc.ZCPMemoryError()
if ret == errno.ENOSPC:
raise lzc_exc.ZCPSpaceError()
if ret == errno.EPERM:
raise lzc_exc.ZCPPermissionError()
if ret == ECHRNG:
raise lzc_exc.ZCPRuntimeError(error)
if ret == errno.EINVAL:
if error is None:
raise lzc_exc.ZCPLimitInvalid()
else:
raise lzc_exc.ZCPSyntaxError(error)
raise _generic_exception(ret, name, "Failed to execute channel program")
def lzc_pool_checkpoint_translate_error(ret, name, discard=False):
if ret == 0:
return
if ret == errno.ENOENT:
raise lzc_exc.PoolNotFound(name)
if ret == ZFS_ERR_CHECKPOINT_EXISTS:
raise lzc_exc.CheckpointExists()
if ret == ZFS_ERR_NO_CHECKPOINT:
raise lzc_exc.CheckpointNotFound()
if ret == ZFS_ERR_DISCARDING_CHECKPOINT:
raise lzc_exc.CheckpointDiscarding()
if ret == ZFS_ERR_DEVRM_IN_PROGRESS:
raise lzc_exc.DeviceRemovalRunning()
if ret == ZFS_ERR_VDEV_TOO_BIG:
raise lzc_exc.DeviceTooBig()
+ if ret == ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS:
+ raise lzc_exc.RaidzExpansionRunning()
if discard:
raise _generic_exception(
ret, name, "Failed to discard pool checkpoint")
else:
raise _generic_exception(ret, name, "Failed to create pool checkpoint")
def lzc_pool_checkpoint_discard_translate_error(ret, name):
lzc_pool_checkpoint_translate_error(ret, name, discard=True)
def lzc_rename_translate_error(ret, source, target):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(source)
_validate_fs_name(target)
if _pool_name(source) != _pool_name(target):
raise lzc_exc.PoolsDiffer(source)
if ret == errno.EEXIST:
raise lzc_exc.FilesystemExists(target)
if ret == errno.ENOENT:
raise lzc_exc.FilesystemNotFound(source)
if ret == ZFS_ERR_WRONG_PARENT:
raise lzc_exc.WrongParent(target)
raise _generic_exception(ret, source, "Failed to rename dataset")
def lzc_destroy_translate_error(ret, name):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(name)
if ret == errno.ENOENT:
raise lzc_exc.FilesystemNotFound(name)
raise _generic_exception(ret, name, "Failed to destroy dataset")
def lzc_inherit_prop_translate_error(ret, name, prop):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(name)
raise lzc_exc.PropertyInvalid(prop)
if ret == errno.ENOENT:
raise lzc_exc.DatasetNotFound(name)
raise _generic_exception(ret, name, "Failed to inherit a property")
def lzc_set_prop_translate_error(ret, name, prop, val):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_or_snap_name(name)
raise lzc_exc.PropertyInvalid(prop)
if ret == errno.ENOENT:
raise lzc_exc.DatasetNotFound(name)
raise _generic_exception(ret, name, "Failed to set a property")
def lzc_get_props_translate_error(ret, name):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_or_snap_name(name)
if ret == errno.ENOENT:
raise lzc_exc.DatasetNotFound(name)
raise _generic_exception(ret, name, "Failed to get properties")
def lzc_list_children_translate_error(ret, name):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(name)
raise _generic_exception(ret, name, "Error while iterating children")
def lzc_list_snaps_translate_error(ret, name):
if ret == 0:
return
if ret == errno.EINVAL:
_validate_fs_name(name)
raise _generic_exception(ret, name, "Error while iterating snapshots")
def lzc_list_translate_error(ret, name, opts):
if ret == 0:
return
if ret == errno.ENOENT:
raise lzc_exc.DatasetNotFound(name)
if ret == errno.EINVAL:
_validate_fs_or_snap_name(name)
raise _generic_exception(ret, name, "Error obtaining a list")
def _handle_err_list(ret, errlist, names, exception, mapper):
'''
Convert one or more errors from an operation into the requested exception.
:param int ret: the overall return code.
:param errlist: the dictionary that maps entity names to their specific
error codes.
:type errlist: dict of bytes:int
:param names: the list of all names of the entities on which the operation
was attempted.
:param type exception: the type of the exception to raise if an error
occurred. The exception should be a subclass of
``MultipleOperationsFailure``.
:param function mapper: the function that maps an error code and a name to
a Python exception.
Unless ``ret`` is zero this function will raise the ``exception``.
If the ``errlist`` is not empty, then the compound exception will contain
a list of exceptions corresponding to each individual error code in the
``errlist``.
Otherwise, the ``exception`` will contain a list with a single exception
corresponding to the ``ret`` value. If the ``names`` list contains only one
element, that is, the operation was attempted on a single entity, then the
name of that entity is passed to the ``mapper``.
If the operation was attempted on multiple entities, but the ``errlist``
is empty, then we can not know which entity caused the error and, thus,
``None`` is used as a name to signify that fact.
.. note::
Note that the ``errlist`` can contain a special element with a key of
"N_MORE_ERRORS".
That element means that there were too many errors to place on the
``errlist``.
Those errors are suppressed and only their count is provided as a
value of the special ``N_MORE_ERRORS`` element.
'''
if ret == 0:
return
if len(errlist) == 0:
suppressed_count = 0
names = list(zip(names, range(2)))
if len(names) == 1:
name, _ = names[0]
else:
name = None
errors = [mapper(ret, name)]
else:
errors = []
suppressed_count = errlist.pop('N_MORE_ERRORS', 0)
for name in errlist:
err = errlist[name]
errors.append(mapper(err, name))
raise exception(errors, suppressed_count)
def _pool_name(name):
'''
Extract a pool name from the given dataset or bookmark name.
'/' separates dataset name components.
'@' separates a snapshot name from the rest of the dataset name.
'#' separates a bookmark name from the rest of the dataset name.
'''
return re.split(b'[/@#]', name, 1)[0]
def _fs_name(name):
'''
Extract a dataset name from the given snapshot or bookmark name.
'@' separates a snapshot name from the rest of the dataset name.
'#' separates a bookmark name from the rest of the dataset name.
'''
return re.split(b'[@#]', name, 1)[0]
def _is_valid_name_component(component):
allowed = string.ascii_letters + string.digits + u'-_.: '
return component and all(x in allowed.encode() for x in component)
def _is_valid_fs_name(name):
return name and all(_is_valid_name_component(c) for c in name.split(b'/'))
def _is_valid_snap_name(name):
parts = name.split(b'@')
return (len(parts) == 2 and _is_valid_fs_name(parts[0]) and
_is_valid_name_component(parts[1]))
def _is_valid_bmark_name(name):
parts = name.split(b'#')
return (len(parts) == 2 and _is_valid_fs_name(parts[0]) and
_is_valid_name_component(parts[1]))
def _validate_fs_name(name):
if not _is_valid_fs_name(name):
raise lzc_exc.FilesystemNameInvalid(name)
elif len(name) > MAXNAMELEN:
raise lzc_exc.NameTooLong(name)
def _validate_snap_name(name):
if not _is_valid_snap_name(name):
raise lzc_exc.SnapshotNameInvalid(name)
elif len(name) > MAXNAMELEN:
raise lzc_exc.NameTooLong(name)
def _validate_bmark_name(name):
if not _is_valid_bmark_name(name):
raise lzc_exc.BookmarkNameInvalid(name)
elif len(name) > MAXNAMELEN:
raise lzc_exc.NameTooLong(name)
def _validate_fs_or_snap_name(name):
if not _is_valid_fs_name(name) and not _is_valid_snap_name(name):
raise lzc_exc.NameInvalid(name)
elif len(name) > MAXNAMELEN:
raise lzc_exc.NameTooLong(name)
def _generic_exception(err, name, message):
if err in _error_to_exception:
return _error_to_exception[err](name)
else:
return lzc_exc.ZFSGenericError(err, message, name)
_error_to_exception = {e.errno: e for e in [
lzc_exc.ZIOError,
lzc_exc.NoSpace,
lzc_exc.QuotaExceeded,
lzc_exc.DatasetBusy,
lzc_exc.NameTooLong,
lzc_exc.ReadOnlyPool,
lzc_exc.SuspendedPool,
lzc_exc.PoolsDiffer,
lzc_exc.PropertyNotSupported,
]}
# vim: softtabstop=4 tabstop=4 expandtab shiftwidth=4
diff --git a/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/exceptions.py b/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/exceptions.py
index e484b07b6450..ba8f7e49093c 100644
--- a/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/exceptions.py
+++ b/sys/contrib/openzfs/contrib/pyzfs/libzfs_core/exceptions.py
@@ -1,601 +1,607 @@
#
# Copyright 2015 ClusterHQ
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
"""
Exceptions that can be raised by libzfs_core operations.
"""
from __future__ import absolute_import, division, print_function
import errno
from ._constants import (
ECHRNG,
ECKSUM,
ETIME,
ZFS_ERR_CHECKPOINT_EXISTS,
ZFS_ERR_DISCARDING_CHECKPOINT,
ZFS_ERR_NO_CHECKPOINT,
ZFS_ERR_DEVRM_IN_PROGRESS,
ZFS_ERR_VDEV_TOO_BIG,
ZFS_ERR_WRONG_PARENT,
+ ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS,
zfs_errno
)
class ZFSError(Exception):
errno = None
message = None
name = None
def __str__(self):
if self.name is not None:
return "[Errno %d] %s: '%s'" % (
self.errno, self.message, self.name)
else:
return "[Errno %d] %s" % (self.errno, self.message)
def __repr__(self):
return "%s(%r, %r)" % (
self.__class__.__name__, self.errno, self.message)
class ZFSGenericError(ZFSError):
def __init__(self, errno, name, message):
self.errno = errno
self.message = message
self.name = name
class ZFSInitializationFailed(ZFSError):
message = "Failed to initialize libzfs_core"
def __init__(self, errno):
self.errno = errno
class MultipleOperationsFailure(ZFSError):
def __init__(self, errors, suppressed_count):
# Use first of the individual error codes
# as an overall error code. This is more consistent.
self.errno = errors[0].errno
self.errors = errors
# this many errors were encountered but not placed on the `errors` list
self.suppressed_count = suppressed_count
def __str__(self):
return "%s, %d errors included, %d suppressed" % (
ZFSError.__str__(self), len(self.errors), self.suppressed_count)
def __repr__(self):
return "%s(%r, %r, errors=%r, suppressed=%r)" % (
self.__class__.__name__, self.errno, self.message, self.errors,
self.suppressed_count)
class DatasetNotFound(ZFSError):
"""
This exception is raised when an operation failure can be caused by a
missing snapshot or a missing filesystem and it is impossible to
distinguish between the causes.
"""
errno = errno.ENOENT
message = "Dataset not found"
def __init__(self, name):
self.name = name
class DatasetExists(ZFSError):
"""
This exception is raised when an operation failure can be caused by an
existing snapshot or filesystem and it is impossible to distinguish between
the causes.
"""
errno = errno.EEXIST
message = "Dataset already exists"
def __init__(self, name):
self.name = name
class NotClone(ZFSError):
errno = errno.EINVAL
message = "Filesystem is not a clone, can not promote"
def __init__(self, name):
self.name = name
class FilesystemExists(DatasetExists):
message = "Filesystem already exists"
def __init__(self, name):
self.name = name
class FilesystemNotFound(DatasetNotFound):
message = "Filesystem not found"
def __init__(self, name):
self.name = name
class ParentNotFound(ZFSError):
errno = errno.ENOENT
message = "Parent not found"
def __init__(self, name):
self.name = name
class WrongParent(ZFSError):
errno = ZFS_ERR_WRONG_PARENT
message = "Parent dataset is not a filesystem"
def __init__(self, name):
self.name = name
class SnapshotExists(DatasetExists):
message = "Snapshot already exists"
def __init__(self, name):
self.name = name
class SnapshotNotFound(DatasetNotFound):
message = "Snapshot not found"
def __init__(self, name):
self.name = name
class SnapshotNotLatest(ZFSError):
errno = errno.EEXIST
message = "Snapshot is not the latest"
def __init__(self, name):
self.name = name
class SnapshotIsCloned(ZFSError):
errno = errno.EEXIST
message = "Snapshot is cloned"
def __init__(self, name):
self.name = name
class SnapshotIsHeld(ZFSError):
errno = errno.EBUSY
message = "Snapshot is held"
def __init__(self, name):
self.name = name
class DuplicateSnapshots(ZFSError):
errno = errno.EXDEV
message = "Requested multiple snapshots of the same filesystem"
def __init__(self, name):
self.name = name
class SnapshotFailure(MultipleOperationsFailure):
message = "Creation of snapshot(s) failed for one or more reasons"
def __init__(self, errors, suppressed_count):
super(SnapshotFailure, self).__init__(errors, suppressed_count)
class SnapshotDestructionFailure(MultipleOperationsFailure):
message = "Destruction of snapshot(s) failed for one or more reasons"
def __init__(self, errors, suppressed_count):
super(SnapshotDestructionFailure, self).__init__(
errors, suppressed_count)
class BookmarkExists(ZFSError):
errno = errno.EEXIST
message = "Bookmark already exists"
def __init__(self, name):
self.name = name
class BookmarkNotFound(ZFSError):
errno = errno.ENOENT
message = "Bookmark not found"
def __init__(self, name):
self.name = name
class BookmarkMismatch(ZFSError):
errno = errno.EINVAL
message = "source is not an ancestor of the new bookmark's dataset"
def __init__(self, name):
self.name = name
class BookmarkSourceInvalid(ZFSError):
errno = errno.EINVAL
message = "Bookmark source is not a valid snapshot or existing bookmark"
def __init__(self, name):
self.name = name
class BookmarkNotSupported(ZFSError):
errno = errno.ENOTSUP
message = "Bookmark feature is not supported"
def __init__(self, name):
self.name = name
class BookmarkFailure(MultipleOperationsFailure):
message = "Creation of bookmark(s) failed for one or more reasons"
def __init__(self, errors, suppressed_count):
super(BookmarkFailure, self).__init__(errors, suppressed_count)
class BookmarkDestructionFailure(MultipleOperationsFailure):
message = "Destruction of bookmark(s) failed for one or more reasons"
def __init__(self, errors, suppressed_count):
super(BookmarkDestructionFailure, self).__init__(
errors, suppressed_count)
class BadHoldCleanupFD(ZFSError):
errno = errno.EBADF
message = "Bad file descriptor as cleanup file descriptor"
class HoldExists(ZFSError):
errno = errno.EEXIST
message = "Hold with a given tag already exists on snapshot"
def __init__(self, name):
self.name = name
class HoldNotFound(ZFSError):
errno = errno.ENOENT
message = "Hold with a given tag does not exist on snapshot"
def __init__(self, name):
self.name = name
class HoldFailure(MultipleOperationsFailure):
message = "Placement of hold(s) failed for one or more reasons"
def __init__(self, errors, suppressed_count):
super(HoldFailure, self).__init__(errors, suppressed_count)
class HoldReleaseFailure(MultipleOperationsFailure):
message = "Release of hold(s) failed for one or more reasons"
def __init__(self, errors, suppressed_count):
super(HoldReleaseFailure, self).__init__(errors, suppressed_count)
class SnapshotMismatch(ZFSError):
errno = errno.ENODEV
message = "Snapshot is not descendant of source snapshot"
def __init__(self, name):
self.name = name
class StreamMismatch(ZFSError):
errno = errno.ENODEV
message = "Stream is not applicable to destination dataset"
def __init__(self, name):
self.name = name
class DestinationModified(ZFSError):
errno = errno.ETXTBSY
message = "Destination dataset has modifications that can not be undone"
def __init__(self, name):
self.name = name
class BadStream(ZFSError):
errno = ECKSUM
message = "Bad backup stream"
class StreamFeatureNotSupported(ZFSError):
errno = errno.ENOTSUP
message = "Stream contains unsupported feature"
class UnknownStreamFeature(ZFSError):
errno = errno.ENOTSUP
message = "Unknown feature requested for stream"
class StreamFeatureInvalid(ZFSError):
errno = errno.EINVAL
message = "Kernel modules must be upgraded to receive this stream"
class StreamFeatureIncompatible(ZFSError):
errno = errno.EINVAL
message = "Incompatible embedded feature with encrypted receive"
class StreamTruncated(ZFSError):
errno = zfs_errno.ZFS_ERR_STREAM_TRUNCATED
message = "incomplete stream"
class ReceivePropertyFailure(MultipleOperationsFailure):
message = "Receiving of properties failed for one or more reasons"
def __init__(self, errors, suppressed_count):
super(ReceivePropertyFailure, self).__init__(errors, suppressed_count)
class StreamIOError(ZFSError):
message = "I/O error while writing or reading stream"
def __init__(self, errno):
self.errno = errno
class ZIOError(ZFSError):
errno = errno.EIO
message = "I/O error"
def __init__(self, name):
self.name = name
class NoSpace(ZFSError):
errno = errno.ENOSPC
message = "No space left"
def __init__(self, name):
self.name = name
class QuotaExceeded(ZFSError):
errno = errno.EDQUOT
message = "Quota exceeded"
def __init__(self, name):
self.name = name
class DatasetBusy(ZFSError):
errno = errno.EBUSY
message = "Dataset is busy"
def __init__(self, name):
self.name = name
class NameTooLong(ZFSError):
errno = errno.ENAMETOOLONG
message = "Dataset name is too long"
def __init__(self, name):
self.name = name
class NameInvalid(ZFSError):
errno = errno.EINVAL
message = "Invalid name"
def __init__(self, name):
self.name = name
class SnapshotNameInvalid(NameInvalid):
message = "Invalid name for snapshot"
def __init__(self, name):
self.name = name
class FilesystemNameInvalid(NameInvalid):
message = "Invalid name for filesystem or volume"
def __init__(self, name):
self.name = name
class BookmarkNameInvalid(NameInvalid):
message = "Invalid name for bookmark"
def __init__(self, name):
self.name = name
class ReadOnlyPool(ZFSError):
errno = errno.EROFS
message = "Pool is read-only"
def __init__(self, name):
self.name = name
class SuspendedPool(ZFSError):
errno = errno.EAGAIN
message = "Pool is suspended"
def __init__(self, name):
self.name = name
class PoolNotFound(ZFSError):
errno = errno.EXDEV
message = "No such pool"
def __init__(self, name):
self.name = name
class PoolsDiffer(ZFSError):
errno = errno.EXDEV
message = "Source and target belong to different pools"
def __init__(self, name):
self.name = name
class FeatureNotSupported(ZFSError):
errno = errno.ENOTSUP
message = "Feature is not supported in this version"
def __init__(self, name):
self.name = name
class PropertyNotSupported(ZFSError):
errno = errno.ENOTSUP
message = "Property is not supported in this version"
def __init__(self, name):
self.name = name
class PropertyInvalid(ZFSError):
errno = errno.EINVAL
message = "Invalid property or property value"
def __init__(self, name):
self.name = name
class DatasetTypeInvalid(ZFSError):
errno = errno.EINVAL
message = "Specified dataset type is unknown"
def __init__(self, name):
self.name = name
class UnknownCryptCommand(ZFSError):
errno = errno.EINVAL
message = "Specified crypt command is invalid"
def __init__(self, name):
self.name = name
class EncryptionKeyNotLoaded(ZFSError):
errno = errno.EACCES
message = "Encryption key is not currently loaded"
class EncryptionKeyAlreadyLoaded(ZFSError):
errno = errno.EEXIST
message = "Encryption key is already loaded"
class EncryptionKeyInvalid(ZFSError):
errno = errno.EACCES
message = "Incorrect encryption key provided"
class ZCPError(ZFSError):
errno = None
message = None
class ZCPSyntaxError(ZCPError):
errno = errno.EINVAL
message = "Channel program contains syntax errors"
def __init__(self, details):
self.details = details
class ZCPRuntimeError(ZCPError):
errno = ECHRNG
message = "Channel programs encountered a runtime error"
def __init__(self, details):
self.details = details
class ZCPLimitInvalid(ZCPError):
errno = errno.EINVAL
message = "Channel program called with invalid limits"
class ZCPTimeout(ZCPError):
errno = ETIME
message = "Channel program timed out"
class ZCPSpaceError(ZCPError):
errno = errno.ENOSPC
message = "Channel program exhausted the memory limit"
class ZCPMemoryError(ZCPError):
errno = errno.ENOMEM
message = "Channel program return value too large"
class ZCPPermissionError(ZCPError):
errno = errno.EPERM
message = "Channel programs must be run as root"
class CheckpointExists(ZFSError):
errno = ZFS_ERR_CHECKPOINT_EXISTS
message = "Pool already has a checkpoint"
class CheckpointNotFound(ZFSError):
errno = ZFS_ERR_NO_CHECKPOINT
message = "Pool does not have a checkpoint"
class CheckpointDiscarding(ZFSError):
errno = ZFS_ERR_DISCARDING_CHECKPOINT
message = "Pool checkpoint is being discarded"
class DeviceRemovalRunning(ZFSError):
errno = ZFS_ERR_DEVRM_IN_PROGRESS
message = "A vdev is currently being removed"
class DeviceTooBig(ZFSError):
errno = ZFS_ERR_VDEV_TOO_BIG
message = "One or more top-level vdevs exceed the maximum vdev size"
+class RaidzExpansionRunning(ZFSError):
+ errno = ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS
+ message = "A raidz device is currently expanding"
+
+
# vim: softtabstop=4 tabstop=4 expandtab shiftwidth=4
diff --git a/sys/contrib/openzfs/include/libzfs.h b/sys/contrib/openzfs/include/libzfs.h
index 4adfa38e87be..dbb6340b0a43 100644
--- a/sys/contrib/openzfs/include/libzfs.h
+++ b/sys/contrib/openzfs/include/libzfs.h
@@ -1,1049 +1,1050 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2022 by Delphix. All rights reserved.
* Copyright Joyent, Inc.
* Copyright (c) 2013 Steven Hartland. All rights reserved.
* Copyright (c) 2016, Intel Corporation.
* Copyright 2016 Nexenta Systems, Inc.
* Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
* Copyright (c) 2019 Datto Inc.
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
*/
#ifndef _LIBZFS_H
#define _LIBZFS_H extern __attribute__((visibility("default")))
#include <assert.h>
#include <libshare.h>
#include <libnvpair.h>
#include <sys/mnttab.h>
#include <sys/param.h>
#include <sys/types.h>
#include <sys/fs/zfs.h>
#include <sys/avl.h>
#include <libzfs_core.h>
#ifdef __cplusplus
extern "C" {
#endif
/*
* Miscellaneous ZFS constants
*/
#define ZFS_MAXPROPLEN MAXPATHLEN
#define ZPOOL_MAXPROPLEN MAXPATHLEN
/*
* libzfs errors
*/
typedef enum zfs_error {
EZFS_SUCCESS = 0, /* no error -- success */
EZFS_NOMEM = 2000, /* out of memory */
EZFS_BADPROP, /* invalid property value */
EZFS_PROPREADONLY, /* cannot set readonly property */
EZFS_PROPTYPE, /* property does not apply to dataset type */
EZFS_PROPNONINHERIT, /* property is not inheritable */
EZFS_PROPSPACE, /* bad quota or reservation */
EZFS_BADTYPE, /* dataset is not of appropriate type */
EZFS_BUSY, /* pool or dataset is busy */
EZFS_EXISTS, /* pool or dataset already exists */
EZFS_NOENT, /* no such pool or dataset */
EZFS_BADSTREAM, /* bad backup stream */
EZFS_DSREADONLY, /* dataset is readonly */
EZFS_VOLTOOBIG, /* volume is too large for 32-bit system */
EZFS_INVALIDNAME, /* invalid dataset name */
EZFS_BADRESTORE, /* unable to restore to destination */
EZFS_BADBACKUP, /* backup failed */
EZFS_BADTARGET, /* bad attach/detach/replace target */
EZFS_NODEVICE, /* no such device in pool */
EZFS_BADDEV, /* invalid device to add */
EZFS_NOREPLICAS, /* no valid replicas */
EZFS_RESILVERING, /* resilvering (healing reconstruction) */
EZFS_BADVERSION, /* unsupported version */
EZFS_POOLUNAVAIL, /* pool is currently unavailable */
EZFS_DEVOVERFLOW, /* too many devices in one vdev */
EZFS_BADPATH, /* must be an absolute path */
EZFS_CROSSTARGET, /* rename or clone across pool or dataset */
EZFS_ZONED, /* used improperly in local zone */
EZFS_MOUNTFAILED, /* failed to mount dataset */
EZFS_UMOUNTFAILED, /* failed to unmount dataset */
EZFS_UNSHARENFSFAILED, /* failed to unshare over nfs */
EZFS_SHARENFSFAILED, /* failed to share over nfs */
EZFS_PERM, /* permission denied */
EZFS_NOSPC, /* out of space */
EZFS_FAULT, /* bad address */
EZFS_IO, /* I/O error */
EZFS_INTR, /* signal received */
EZFS_ISSPARE, /* device is a hot spare */
EZFS_INVALCONFIG, /* invalid vdev configuration */
EZFS_RECURSIVE, /* recursive dependency */
EZFS_NOHISTORY, /* no history object */
EZFS_POOLPROPS, /* couldn't retrieve pool props */
EZFS_POOL_NOTSUP, /* ops not supported for this type of pool */
EZFS_POOL_INVALARG, /* invalid argument for this pool operation */
EZFS_NAMETOOLONG, /* dataset name is too long */
EZFS_OPENFAILED, /* open of device failed */
EZFS_NOCAP, /* couldn't get capacity */
EZFS_LABELFAILED, /* write of label failed */
EZFS_BADWHO, /* invalid permission who */
EZFS_BADPERM, /* invalid permission */
EZFS_BADPERMSET, /* invalid permission set name */
EZFS_NODELEGATION, /* delegated administration is disabled */
EZFS_UNSHARESMBFAILED, /* failed to unshare over smb */
EZFS_SHARESMBFAILED, /* failed to share over smb */
EZFS_BADCACHE, /* bad cache file */
EZFS_ISL2CACHE, /* device is for the level 2 ARC */
EZFS_VDEVNOTSUP, /* unsupported vdev type */
EZFS_NOTSUP, /* ops not supported on this dataset */
EZFS_ACTIVE_SPARE, /* pool has active shared spare devices */
EZFS_UNPLAYED_LOGS, /* log device has unplayed logs */
EZFS_REFTAG_RELE, /* snapshot release: tag not found */
EZFS_REFTAG_HOLD, /* snapshot hold: tag already exists */
EZFS_TAGTOOLONG, /* snapshot hold/rele: tag too long */
EZFS_PIPEFAILED, /* pipe create failed */
EZFS_THREADCREATEFAILED, /* thread create failed */
EZFS_POSTSPLIT_ONLINE, /* onlining a disk after splitting it */
EZFS_SCRUBBING, /* currently scrubbing */
EZFS_ERRORSCRUBBING, /* currently error scrubbing */
EZFS_ERRORSCRUB_PAUSED, /* error scrub currently paused */
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_SCRUB_PAUSED_TO_CANCEL, /* scrub currently paused */
EZFS_ACTIVE_POOL, /* pool is imported on a different system */
EZFS_CRYPTOFAILED, /* failed to setup encryption */
EZFS_NO_PENDING, /* cannot cancel, no operation is pending */
EZFS_CHECKPOINT_EXISTS, /* checkpoint exists */
EZFS_DISCARDING_CHECKPOINT, /* currently discarding a checkpoint */
EZFS_NO_CHECKPOINT, /* pool has no checkpoint */
EZFS_DEVRM_IN_PROGRESS, /* a device is currently being removed */
EZFS_VDEV_TOO_BIG, /* a device is too big to be used */
EZFS_IOC_NOTSUPPORTED, /* operation not supported by zfs module */
EZFS_TOOMANY, /* argument list too long */
EZFS_INITIALIZING, /* currently initializing */
EZFS_NO_INITIALIZE, /* no active initialize */
EZFS_WRONG_PARENT, /* invalid parent dataset (e.g ZVOL) */
EZFS_TRIMMING, /* currently trimming */
EZFS_NO_TRIM, /* no active trim */
EZFS_TRIM_NOTSUP, /* device does not support trim */
EZFS_NO_RESILVER_DEFER, /* pool doesn't support resilver_defer */
EZFS_EXPORT_IN_PROGRESS, /* currently exporting the pool */
EZFS_REBUILDING, /* resilvering (sequential reconstrution) */
EZFS_VDEV_NOTSUP, /* ops not supported for this type of vdev */
EZFS_NOT_USER_NAMESPACE, /* a file is not a user namespace */
EZFS_CKSUM, /* insufficient replicas */
EZFS_RESUME_EXISTS, /* Resume on existing dataset without force */
EZFS_SHAREFAILED, /* filesystem share failed */
+ EZFS_RAIDZ_EXPAND_IN_PROGRESS, /* a raidz is currently expanding */
EZFS_UNKNOWN
} zfs_error_t;
/*
* The following data structures are all part
* of the zfs_allow_t data structure which is
* used for printing 'allow' permissions.
* It is a linked list of zfs_allow_t's which
* then contain avl tree's for user/group/sets/...
* and each one of the entries in those trees have
* avl tree's for the permissions they belong to and
* whether they are local,descendent or local+descendent
* permissions. The AVL trees are used primarily for
* sorting purposes, but also so that we can quickly find
* a given user and or permission.
*/
typedef struct zfs_perm_node {
avl_node_t z_node;
char z_pname[MAXPATHLEN];
} zfs_perm_node_t;
typedef struct zfs_allow_node {
avl_node_t z_node;
char z_key[MAXPATHLEN]; /* name, such as joe */
avl_tree_t z_localdescend; /* local+descendent perms */
avl_tree_t z_local; /* local permissions */
avl_tree_t z_descend; /* descendent permissions */
} zfs_allow_node_t;
typedef struct zfs_allow {
struct zfs_allow *z_next;
char z_setpoint[MAXPATHLEN];
avl_tree_t z_sets;
avl_tree_t z_crperms;
avl_tree_t z_user;
avl_tree_t z_group;
avl_tree_t z_everyone;
} zfs_allow_t;
/*
* Basic handle types
*/
typedef struct zfs_handle zfs_handle_t;
typedef struct zpool_handle zpool_handle_t;
typedef struct libzfs_handle libzfs_handle_t;
_LIBZFS_H int zpool_wait(zpool_handle_t *, zpool_wait_activity_t);
_LIBZFS_H int zpool_wait_status(zpool_handle_t *, zpool_wait_activity_t,
boolean_t *, boolean_t *);
/*
* Library initialization
*/
_LIBZFS_H libzfs_handle_t *libzfs_init(void);
_LIBZFS_H void libzfs_fini(libzfs_handle_t *);
_LIBZFS_H libzfs_handle_t *zpool_get_handle(zpool_handle_t *);
_LIBZFS_H libzfs_handle_t *zfs_get_handle(zfs_handle_t *);
_LIBZFS_H void libzfs_print_on_error(libzfs_handle_t *, boolean_t);
_LIBZFS_H void zfs_save_arguments(int argc, char **, char *, int);
_LIBZFS_H int zpool_log_history(libzfs_handle_t *, const char *);
_LIBZFS_H int libzfs_errno(libzfs_handle_t *);
_LIBZFS_H const char *libzfs_error_init(int);
_LIBZFS_H const char *libzfs_error_action(libzfs_handle_t *);
_LIBZFS_H const char *libzfs_error_description(libzfs_handle_t *);
_LIBZFS_H int zfs_standard_error(libzfs_handle_t *, int, const char *);
_LIBZFS_H void libzfs_mnttab_init(libzfs_handle_t *);
_LIBZFS_H void libzfs_mnttab_fini(libzfs_handle_t *);
_LIBZFS_H void libzfs_mnttab_cache(libzfs_handle_t *, boolean_t);
_LIBZFS_H int libzfs_mnttab_find(libzfs_handle_t *, const char *,
struct mnttab *);
_LIBZFS_H void libzfs_mnttab_add(libzfs_handle_t *, const char *,
const char *, const char *);
_LIBZFS_H void libzfs_mnttab_remove(libzfs_handle_t *, const char *);
/*
* Basic handle functions
*/
_LIBZFS_H zpool_handle_t *zpool_open(libzfs_handle_t *, const char *);
_LIBZFS_H zpool_handle_t *zpool_open_canfail(libzfs_handle_t *, const char *);
_LIBZFS_H void zpool_close(zpool_handle_t *);
_LIBZFS_H const char *zpool_get_name(zpool_handle_t *);
_LIBZFS_H int zpool_get_state(zpool_handle_t *);
_LIBZFS_H const char *zpool_state_to_name(vdev_state_t, vdev_aux_t);
_LIBZFS_H const char *zpool_pool_state_to_name(pool_state_t);
_LIBZFS_H void zpool_free_handles(libzfs_handle_t *);
/*
* Iterate over all active pools in the system.
*/
typedef int (*zpool_iter_f)(zpool_handle_t *, void *);
_LIBZFS_H int zpool_iter(libzfs_handle_t *, zpool_iter_f, void *);
_LIBZFS_H boolean_t zpool_skip_pool(const char *);
/*
* Functions to create and destroy pools
*/
_LIBZFS_H int zpool_create(libzfs_handle_t *, const char *, nvlist_t *,
nvlist_t *, nvlist_t *);
_LIBZFS_H int zpool_destroy(zpool_handle_t *, const char *);
_LIBZFS_H int zpool_add(zpool_handle_t *, nvlist_t *);
typedef struct splitflags {
/* do not split, but return the config that would be split off */
unsigned int dryrun : 1;
/* after splitting, import the pool */
unsigned int import : 1;
int name_flags;
} splitflags_t;
typedef struct trimflags {
/* requested vdevs are for the entire pool */
boolean_t fullpool;
/* request a secure trim, requires support from device */
boolean_t secure;
/* after starting trim, block until trim completes */
boolean_t wait;
/* trim at the requested rate in bytes/second */
uint64_t rate;
} trimflags_t;
/*
* Functions to manipulate pool and vdev state
*/
_LIBZFS_H int zpool_scan(zpool_handle_t *, pool_scan_func_t, pool_scrub_cmd_t);
_LIBZFS_H int zpool_initialize(zpool_handle_t *, pool_initialize_func_t,
nvlist_t *);
_LIBZFS_H int zpool_initialize_wait(zpool_handle_t *, pool_initialize_func_t,
nvlist_t *);
_LIBZFS_H int zpool_trim(zpool_handle_t *, pool_trim_func_t, nvlist_t *,
trimflags_t *);
_LIBZFS_H int zpool_clear(zpool_handle_t *, const char *, nvlist_t *);
_LIBZFS_H int zpool_reguid(zpool_handle_t *);
_LIBZFS_H int zpool_reopen_one(zpool_handle_t *, void *);
_LIBZFS_H int zpool_sync_one(zpool_handle_t *, void *);
_LIBZFS_H int zpool_vdev_online(zpool_handle_t *, const char *, int,
vdev_state_t *);
_LIBZFS_H int zpool_vdev_offline(zpool_handle_t *, const char *, boolean_t);
_LIBZFS_H int zpool_vdev_attach(zpool_handle_t *, const char *,
const char *, nvlist_t *, int, boolean_t);
_LIBZFS_H int zpool_vdev_detach(zpool_handle_t *, const char *);
_LIBZFS_H int zpool_vdev_remove(zpool_handle_t *, const char *);
_LIBZFS_H int zpool_vdev_remove_cancel(zpool_handle_t *);
_LIBZFS_H int zpool_vdev_indirect_size(zpool_handle_t *, const char *,
uint64_t *);
_LIBZFS_H int zpool_vdev_split(zpool_handle_t *, char *, nvlist_t **,
nvlist_t *, splitflags_t);
_LIBZFS_H int zpool_vdev_remove_wanted(zpool_handle_t *, const char *);
_LIBZFS_H int zpool_vdev_fault(zpool_handle_t *, uint64_t, vdev_aux_t);
_LIBZFS_H int zpool_vdev_degrade(zpool_handle_t *, uint64_t, vdev_aux_t);
_LIBZFS_H int zpool_vdev_clear(zpool_handle_t *, uint64_t);
_LIBZFS_H nvlist_t *zpool_find_vdev(zpool_handle_t *, const char *, boolean_t *,
boolean_t *, boolean_t *);
_LIBZFS_H nvlist_t *zpool_find_vdev_by_physpath(zpool_handle_t *, const char *,
boolean_t *, boolean_t *, boolean_t *);
_LIBZFS_H int zpool_label_disk(libzfs_handle_t *, zpool_handle_t *,
const char *);
_LIBZFS_H int zpool_prepare_disk(zpool_handle_t *zhp, nvlist_t *vdev_nv,
const char *prepare_str, char **lines[], int *lines_cnt);
_LIBZFS_H int zpool_prepare_and_label_disk(libzfs_handle_t *hdl,
zpool_handle_t *, const char *, nvlist_t *vdev_nv, const char *prepare_str,
char **lines[], int *lines_cnt);
_LIBZFS_H char ** zpool_vdev_script_alloc_env(const char *pool_name,
const char *vdev_path, const char *vdev_upath,
const char *vdev_enc_sysfs_path, const char *opt_key, const char *opt_val);
_LIBZFS_H void zpool_vdev_script_free_env(char **env);
_LIBZFS_H uint64_t zpool_vdev_path_to_guid(zpool_handle_t *zhp,
const char *path);
_LIBZFS_H const char *zpool_get_state_str(zpool_handle_t *);
/*
* Functions to manage pool properties
*/
_LIBZFS_H int zpool_set_prop(zpool_handle_t *, const char *, const char *);
_LIBZFS_H int zpool_get_prop(zpool_handle_t *, zpool_prop_t, char *,
size_t proplen, zprop_source_t *, boolean_t literal);
_LIBZFS_H int zpool_get_userprop(zpool_handle_t *, const char *, char *,
size_t proplen, zprop_source_t *);
_LIBZFS_H uint64_t zpool_get_prop_int(zpool_handle_t *, zpool_prop_t,
zprop_source_t *);
_LIBZFS_H int zpool_props_refresh(zpool_handle_t *);
_LIBZFS_H const char *zpool_prop_to_name(zpool_prop_t);
_LIBZFS_H const char *zpool_prop_values(zpool_prop_t);
/*
* Functions to manage vdev properties
*/
_LIBZFS_H int zpool_get_vdev_prop_value(nvlist_t *, vdev_prop_t, char *, char *,
size_t, zprop_source_t *, boolean_t);
_LIBZFS_H int zpool_get_vdev_prop(zpool_handle_t *, const char *, vdev_prop_t,
char *, char *, size_t, zprop_source_t *, boolean_t);
_LIBZFS_H int zpool_get_all_vdev_props(zpool_handle_t *, const char *,
nvlist_t **);
_LIBZFS_H int zpool_set_vdev_prop(zpool_handle_t *, const char *, const char *,
const char *);
_LIBZFS_H const char *vdev_prop_to_name(vdev_prop_t);
_LIBZFS_H const char *vdev_prop_values(vdev_prop_t);
_LIBZFS_H boolean_t vdev_prop_user(const char *name);
_LIBZFS_H const char *vdev_prop_column_name(vdev_prop_t);
_LIBZFS_H boolean_t vdev_prop_align_right(vdev_prop_t);
/*
* Pool health statistics.
*/
typedef enum {
/*
* The following correspond to faults as defined in the (fault.fs.zfs.*)
* event namespace. Each is associated with a corresponding message ID.
* This must be kept in sync with the zfs_msgid_table in
* lib/libzfs/libzfs_status.c.
*/
ZPOOL_STATUS_CORRUPT_CACHE, /* corrupt /kernel/drv/zpool.cache */
ZPOOL_STATUS_MISSING_DEV_R, /* missing device with replicas */
ZPOOL_STATUS_MISSING_DEV_NR, /* missing device with no replicas */
ZPOOL_STATUS_CORRUPT_LABEL_R, /* bad device label with replicas */
ZPOOL_STATUS_CORRUPT_LABEL_NR, /* bad device label with no replicas */
ZPOOL_STATUS_BAD_GUID_SUM, /* sum of device guids didn't match */
ZPOOL_STATUS_CORRUPT_POOL, /* pool metadata is corrupted */
ZPOOL_STATUS_CORRUPT_DATA, /* data errors in user (meta)data */
ZPOOL_STATUS_FAILING_DEV, /* device experiencing errors */
ZPOOL_STATUS_VERSION_NEWER, /* newer on-disk version */
ZPOOL_STATUS_HOSTID_MISMATCH, /* last accessed by another system */
ZPOOL_STATUS_HOSTID_ACTIVE, /* currently active on another system */
ZPOOL_STATUS_HOSTID_REQUIRED, /* multihost=on and hostid=0 */
ZPOOL_STATUS_IO_FAILURE_WAIT, /* failed I/O, failmode 'wait' */
ZPOOL_STATUS_IO_FAILURE_CONTINUE, /* failed I/O, failmode 'continue' */
ZPOOL_STATUS_IO_FAILURE_MMP, /* failed MMP, failmode not 'panic' */
ZPOOL_STATUS_BAD_LOG, /* cannot read log chain(s) */
ZPOOL_STATUS_ERRATA, /* informational errata available */
/*
* If the pool has unsupported features but can still be opened in
* read-only mode, its status is ZPOOL_STATUS_UNSUP_FEAT_WRITE. If the
* pool has unsupported features but cannot be opened at all, its
* status is ZPOOL_STATUS_UNSUP_FEAT_READ.
*/
ZPOOL_STATUS_UNSUP_FEAT_READ, /* unsupported features for read */
ZPOOL_STATUS_UNSUP_FEAT_WRITE, /* unsupported features for write */
/*
* These faults have no corresponding message ID. At the time we are
* checking the status, the original reason for the FMA fault (I/O or
* checksum errors) has been lost.
*/
ZPOOL_STATUS_FAULTED_DEV_R, /* faulted device with replicas */
ZPOOL_STATUS_FAULTED_DEV_NR, /* faulted device with no replicas */
/*
* The following are not faults per se, but still an error possibly
* requiring administrative attention. There is no corresponding
* message ID.
*/
ZPOOL_STATUS_VERSION_OLDER, /* older legacy on-disk version */
ZPOOL_STATUS_FEAT_DISABLED, /* supported features are disabled */
ZPOOL_STATUS_RESILVERING, /* device being resilvered */
ZPOOL_STATUS_OFFLINE_DEV, /* device offline */
ZPOOL_STATUS_REMOVED_DEV, /* removed device */
ZPOOL_STATUS_REBUILDING, /* device being rebuilt */
ZPOOL_STATUS_REBUILD_SCRUB, /* recommend scrubbing the pool */
ZPOOL_STATUS_NON_NATIVE_ASHIFT, /* (e.g. 512e dev with ashift of 9) */
ZPOOL_STATUS_COMPATIBILITY_ERR, /* bad 'compatibility' property */
ZPOOL_STATUS_INCOMPATIBLE_FEAT, /* feature set outside compatibility */
/*
* Finally, the following indicates a healthy pool.
*/
ZPOOL_STATUS_OK
} zpool_status_t;
_LIBZFS_H zpool_status_t zpool_get_status(zpool_handle_t *, const char **,
zpool_errata_t *);
_LIBZFS_H zpool_status_t zpool_import_status(nvlist_t *, const char **,
zpool_errata_t *);
/*
* Statistics and configuration functions.
*/
_LIBZFS_H nvlist_t *zpool_get_config(zpool_handle_t *, nvlist_t **);
_LIBZFS_H nvlist_t *zpool_get_features(zpool_handle_t *);
_LIBZFS_H int zpool_refresh_stats(zpool_handle_t *, boolean_t *);
_LIBZFS_H int zpool_get_errlog(zpool_handle_t *, nvlist_t **);
/*
* Import and export functions
*/
_LIBZFS_H int zpool_export(zpool_handle_t *, boolean_t, const char *);
_LIBZFS_H int zpool_export_force(zpool_handle_t *, const char *);
_LIBZFS_H int zpool_import(libzfs_handle_t *, nvlist_t *, const char *,
char *altroot);
_LIBZFS_H int zpool_import_props(libzfs_handle_t *, nvlist_t *, const char *,
nvlist_t *, int);
_LIBZFS_H void zpool_print_unsup_feat(nvlist_t *config);
/*
* Miscellaneous pool functions
*/
struct zfs_cmd;
_LIBZFS_H const char *const zfs_history_event_names[];
typedef enum {
VDEV_NAME_PATH = 1 << 0,
VDEV_NAME_GUID = 1 << 1,
VDEV_NAME_FOLLOW_LINKS = 1 << 2,
VDEV_NAME_TYPE_ID = 1 << 3,
} vdev_name_t;
_LIBZFS_H char *zpool_vdev_name(libzfs_handle_t *, zpool_handle_t *, nvlist_t *,
int name_flags);
_LIBZFS_H int zpool_upgrade(zpool_handle_t *, uint64_t);
_LIBZFS_H int zpool_get_history(zpool_handle_t *, nvlist_t **, uint64_t *,
boolean_t *);
_LIBZFS_H int zpool_events_next(libzfs_handle_t *, nvlist_t **, int *, unsigned,
int);
_LIBZFS_H int zpool_events_clear(libzfs_handle_t *, int *);
_LIBZFS_H int zpool_events_seek(libzfs_handle_t *, uint64_t, int);
_LIBZFS_H void zpool_obj_to_path_ds(zpool_handle_t *, uint64_t, uint64_t,
char *, size_t);
_LIBZFS_H void zpool_obj_to_path(zpool_handle_t *, uint64_t, uint64_t, char *,
size_t);
_LIBZFS_H int zfs_ioctl(libzfs_handle_t *, int, struct zfs_cmd *);
_LIBZFS_H void zpool_explain_recover(libzfs_handle_t *, const char *, int,
nvlist_t *);
_LIBZFS_H int zpool_checkpoint(zpool_handle_t *);
_LIBZFS_H int zpool_discard_checkpoint(zpool_handle_t *);
_LIBZFS_H boolean_t zpool_is_draid_spare(const char *);
/*
* Basic handle manipulations. These functions do not create or destroy the
* underlying datasets, only the references to them.
*/
_LIBZFS_H zfs_handle_t *zfs_open(libzfs_handle_t *, const char *, int);
_LIBZFS_H zfs_handle_t *zfs_handle_dup(zfs_handle_t *);
_LIBZFS_H void zfs_close(zfs_handle_t *);
_LIBZFS_H zfs_type_t zfs_get_type(const zfs_handle_t *);
_LIBZFS_H zfs_type_t zfs_get_underlying_type(const zfs_handle_t *);
_LIBZFS_H const char *zfs_get_name(const zfs_handle_t *);
_LIBZFS_H zpool_handle_t *zfs_get_pool_handle(const zfs_handle_t *);
_LIBZFS_H const char *zfs_get_pool_name(const zfs_handle_t *);
/*
* Property management functions. Some functions are shared with the kernel,
* and are found in sys/fs/zfs.h.
*/
/*
* zfs dataset property management
*/
_LIBZFS_H const char *zfs_prop_default_string(zfs_prop_t);
_LIBZFS_H uint64_t zfs_prop_default_numeric(zfs_prop_t);
_LIBZFS_H const char *zfs_prop_column_name(zfs_prop_t);
_LIBZFS_H boolean_t zfs_prop_align_right(zfs_prop_t);
_LIBZFS_H nvlist_t *zfs_valid_proplist(libzfs_handle_t *, zfs_type_t,
nvlist_t *, uint64_t, zfs_handle_t *, zpool_handle_t *, boolean_t,
const char *);
_LIBZFS_H const char *zfs_prop_to_name(zfs_prop_t);
_LIBZFS_H int zfs_prop_set(zfs_handle_t *, const char *, const char *);
_LIBZFS_H int zfs_prop_set_list(zfs_handle_t *, nvlist_t *);
_LIBZFS_H int zfs_prop_set_list_flags(zfs_handle_t *, nvlist_t *, int);
_LIBZFS_H int zfs_prop_get(zfs_handle_t *, zfs_prop_t, char *, size_t,
zprop_source_t *, char *, size_t, boolean_t);
_LIBZFS_H int zfs_prop_get_recvd(zfs_handle_t *, const char *, char *, size_t,
boolean_t);
_LIBZFS_H int zfs_prop_get_numeric(zfs_handle_t *, zfs_prop_t, uint64_t *,
zprop_source_t *, char *, size_t);
_LIBZFS_H int zfs_prop_get_userquota_int(zfs_handle_t *zhp,
const char *propname, uint64_t *propvalue);
_LIBZFS_H int zfs_prop_get_userquota(zfs_handle_t *zhp, const char *propname,
char *propbuf, int proplen, boolean_t literal);
_LIBZFS_H int zfs_prop_get_written_int(zfs_handle_t *zhp, const char *propname,
uint64_t *propvalue);
_LIBZFS_H int zfs_prop_get_written(zfs_handle_t *zhp, const char *propname,
char *propbuf, int proplen, boolean_t literal);
_LIBZFS_H int zfs_prop_get_feature(zfs_handle_t *zhp, const char *propname,
char *buf, size_t len);
_LIBZFS_H uint64_t getprop_uint64(zfs_handle_t *, zfs_prop_t, const char **);
_LIBZFS_H uint64_t zfs_prop_get_int(zfs_handle_t *, zfs_prop_t);
_LIBZFS_H int zfs_prop_inherit(zfs_handle_t *, const char *, boolean_t);
_LIBZFS_H const char *zfs_prop_values(zfs_prop_t);
_LIBZFS_H int zfs_prop_is_string(zfs_prop_t prop);
_LIBZFS_H nvlist_t *zfs_get_all_props(zfs_handle_t *);
_LIBZFS_H nvlist_t *zfs_get_user_props(zfs_handle_t *);
_LIBZFS_H nvlist_t *zfs_get_recvd_props(zfs_handle_t *);
_LIBZFS_H nvlist_t *zfs_get_clones_nvl(zfs_handle_t *);
_LIBZFS_H int zfs_wait_status(zfs_handle_t *, zfs_wait_activity_t,
boolean_t *, boolean_t *);
/*
* zfs encryption management
*/
_LIBZFS_H int zfs_crypto_get_encryption_root(zfs_handle_t *, boolean_t *,
char *);
_LIBZFS_H int zfs_crypto_create(libzfs_handle_t *, char *, nvlist_t *,
nvlist_t *, boolean_t stdin_available, uint8_t **, uint_t *);
_LIBZFS_H int zfs_crypto_clone_check(libzfs_handle_t *, zfs_handle_t *, char *,
nvlist_t *);
_LIBZFS_H int zfs_crypto_attempt_load_keys(libzfs_handle_t *, const char *);
_LIBZFS_H int zfs_crypto_load_key(zfs_handle_t *, boolean_t, const char *);
_LIBZFS_H int zfs_crypto_unload_key(zfs_handle_t *);
_LIBZFS_H int zfs_crypto_rewrap(zfs_handle_t *, nvlist_t *, boolean_t);
typedef struct zprop_list {
int pl_prop;
char *pl_user_prop;
struct zprop_list *pl_next;
boolean_t pl_all;
size_t pl_width;
size_t pl_recvd_width;
boolean_t pl_fixed;
} zprop_list_t;
_LIBZFS_H int zfs_expand_proplist(zfs_handle_t *, zprop_list_t **, boolean_t,
boolean_t);
_LIBZFS_H void zfs_prune_proplist(zfs_handle_t *, uint8_t *);
_LIBZFS_H int vdev_expand_proplist(zpool_handle_t *, const char *,
zprop_list_t **);
#define ZFS_MOUNTPOINT_NONE "none"
#define ZFS_MOUNTPOINT_LEGACY "legacy"
#define ZFS_FEATURE_DISABLED "disabled"
#define ZFS_FEATURE_ENABLED "enabled"
#define ZFS_FEATURE_ACTIVE "active"
#define ZFS_UNSUPPORTED_INACTIVE "inactive"
#define ZFS_UNSUPPORTED_READONLY "readonly"
/*
* zpool property management
*/
_LIBZFS_H int zpool_expand_proplist(zpool_handle_t *, zprop_list_t **,
zfs_type_t, boolean_t);
_LIBZFS_H int zpool_prop_get_feature(zpool_handle_t *, const char *, char *,
size_t);
_LIBZFS_H const char *zpool_prop_default_string(zpool_prop_t);
_LIBZFS_H uint64_t zpool_prop_default_numeric(zpool_prop_t);
_LIBZFS_H const char *zpool_prop_column_name(zpool_prop_t);
_LIBZFS_H boolean_t zpool_prop_align_right(zpool_prop_t);
/*
* Functions shared by zfs and zpool property management.
*/
_LIBZFS_H int zprop_iter(zprop_func func, void *cb, boolean_t show_all,
boolean_t ordered, zfs_type_t type);
_LIBZFS_H int zprop_get_list(libzfs_handle_t *, char *, zprop_list_t **,
zfs_type_t);
_LIBZFS_H void zprop_free_list(zprop_list_t *);
#define ZFS_GET_NCOLS 5
typedef enum {
GET_COL_NONE,
GET_COL_NAME,
GET_COL_PROPERTY,
GET_COL_VALUE,
GET_COL_RECVD,
GET_COL_SOURCE
} zfs_get_column_t;
/*
* Functions for printing zfs or zpool properties
*/
typedef struct vdev_cbdata {
int cb_name_flags;
char **cb_names;
unsigned int cb_names_count;
} vdev_cbdata_t;
typedef struct zprop_get_cbdata {
int cb_sources;
zfs_get_column_t cb_columns[ZFS_GET_NCOLS];
int cb_colwidths[ZFS_GET_NCOLS + 1];
boolean_t cb_scripted;
boolean_t cb_literal;
boolean_t cb_first;
zprop_list_t *cb_proplist;
zfs_type_t cb_type;
vdev_cbdata_t cb_vdevs;
} zprop_get_cbdata_t;
#define ZFS_SET_NOMOUNT 1
typedef struct zprop_set_cbdata {
int cb_flags;
nvlist_t *cb_proplist;
} zprop_set_cbdata_t;
_LIBZFS_H void zprop_print_one_property(const char *, zprop_get_cbdata_t *,
const char *, const char *, zprop_source_t, const char *,
const char *);
/*
* Iterator functions.
*/
#define ZFS_ITER_RECURSE (1 << 0)
#define ZFS_ITER_ARGS_CAN_BE_PATHS (1 << 1)
#define ZFS_ITER_PROP_LISTSNAPS (1 << 2)
#define ZFS_ITER_DEPTH_LIMIT (1 << 3)
#define ZFS_ITER_RECVD_PROPS (1 << 4)
#define ZFS_ITER_LITERAL_PROPS (1 << 5)
#define ZFS_ITER_SIMPLE (1 << 6)
typedef int (*zfs_iter_f)(zfs_handle_t *, void *);
_LIBZFS_H int zfs_iter_root(libzfs_handle_t *, zfs_iter_f, void *);
_LIBZFS_H int zfs_iter_children(zfs_handle_t *, zfs_iter_f, void *);
_LIBZFS_H int zfs_iter_dependents(zfs_handle_t *, boolean_t, zfs_iter_f,
void *);
_LIBZFS_H int zfs_iter_filesystems(zfs_handle_t *, zfs_iter_f, void *);
_LIBZFS_H int zfs_iter_snapshots(zfs_handle_t *, boolean_t, zfs_iter_f, void *,
uint64_t, uint64_t);
_LIBZFS_H int zfs_iter_snapshots_sorted(zfs_handle_t *, zfs_iter_f, void *,
uint64_t, uint64_t);
_LIBZFS_H int zfs_iter_snapspec(zfs_handle_t *, const char *, zfs_iter_f,
void *);
_LIBZFS_H int zfs_iter_bookmarks(zfs_handle_t *, zfs_iter_f, void *);
_LIBZFS_H int zfs_iter_children_v2(zfs_handle_t *, int, zfs_iter_f, void *);
_LIBZFS_H int zfs_iter_dependents_v2(zfs_handle_t *, int, boolean_t, zfs_iter_f,
void *);
_LIBZFS_H int zfs_iter_filesystems_v2(zfs_handle_t *, int, zfs_iter_f, void *);
_LIBZFS_H int zfs_iter_snapshots_v2(zfs_handle_t *, int, zfs_iter_f, void *,
uint64_t, uint64_t);
_LIBZFS_H int zfs_iter_snapshots_sorted_v2(zfs_handle_t *, int, zfs_iter_f,
void *, uint64_t, uint64_t);
_LIBZFS_H int zfs_iter_snapspec_v2(zfs_handle_t *, int, const char *,
zfs_iter_f, void *);
_LIBZFS_H int zfs_iter_bookmarks_v2(zfs_handle_t *, int, zfs_iter_f, void *);
_LIBZFS_H int zfs_iter_mounted(zfs_handle_t *, zfs_iter_f, void *);
typedef struct get_all_cb {
zfs_handle_t **cb_handles;
size_t cb_alloc;
size_t cb_used;
} get_all_cb_t;
_LIBZFS_H void zfs_foreach_mountpoint(libzfs_handle_t *, zfs_handle_t **,
size_t, zfs_iter_f, void *, boolean_t);
_LIBZFS_H void libzfs_add_handle(get_all_cb_t *, zfs_handle_t *);
/*
* Functions to create and destroy datasets.
*/
_LIBZFS_H int zfs_create(libzfs_handle_t *, const char *, zfs_type_t,
nvlist_t *);
_LIBZFS_H int zfs_create_ancestors(libzfs_handle_t *, const char *);
_LIBZFS_H int zfs_destroy(zfs_handle_t *, boolean_t);
_LIBZFS_H int zfs_destroy_snaps(zfs_handle_t *, char *, boolean_t);
_LIBZFS_H int zfs_destroy_snaps_nvl(libzfs_handle_t *, nvlist_t *, boolean_t);
_LIBZFS_H int zfs_destroy_snaps_nvl_os(libzfs_handle_t *, nvlist_t *);
_LIBZFS_H int zfs_clone(zfs_handle_t *, const char *, nvlist_t *);
_LIBZFS_H int zfs_snapshot(libzfs_handle_t *, const char *, boolean_t,
nvlist_t *);
_LIBZFS_H int zfs_snapshot_nvl(libzfs_handle_t *hdl, nvlist_t *snaps,
nvlist_t *props);
_LIBZFS_H int zfs_rollback(zfs_handle_t *, zfs_handle_t *, boolean_t);
typedef struct renameflags {
/* recursive rename */
unsigned int recursive : 1;
/* don't unmount file systems */
unsigned int nounmount : 1;
/* force unmount file systems */
unsigned int forceunmount : 1;
} renameflags_t;
_LIBZFS_H int zfs_rename(zfs_handle_t *, const char *, renameflags_t);
typedef struct sendflags {
/* Amount of extra information to print. */
int verbosity;
/* recursive send (ie, -R) */
boolean_t replicate;
/* for recursive send, skip sending missing snapshots */
boolean_t skipmissing;
/* for incrementals, do all intermediate snapshots */
boolean_t doall;
/* if dataset is a clone, do incremental from its origin */
boolean_t fromorigin;
/* field no longer used, maintained for backwards compatibility */
boolean_t pad;
/* send properties (ie, -p) */
boolean_t props;
/* do not send (no-op, ie. -n) */
boolean_t dryrun;
/* parsable verbose output (ie. -P) */
boolean_t parsable;
/* show progress (ie. -v) */
boolean_t progress;
/* show progress as process title (ie. -V) */
boolean_t progressastitle;
/* large blocks (>128K) are permitted */
boolean_t largeblock;
/* WRITE_EMBEDDED records of type DATA are permitted */
boolean_t embed_data;
/* compressed WRITE records are permitted */
boolean_t compress;
/* raw encrypted records are permitted */
boolean_t raw;
/* only send received properties (ie. -b) */
boolean_t backup;
/* include snapshot holds in send stream */
boolean_t holds;
/* stream represents a partially received dataset */
boolean_t saved;
} sendflags_t;
typedef boolean_t (snapfilter_cb_t)(zfs_handle_t *, void *);
_LIBZFS_H int zfs_send(zfs_handle_t *, const char *, const char *,
sendflags_t *, int, snapfilter_cb_t, void *, nvlist_t **);
_LIBZFS_H int zfs_send_one(zfs_handle_t *, const char *, int, sendflags_t *,
const char *);
_LIBZFS_H int zfs_send_progress(zfs_handle_t *, int, uint64_t *, uint64_t *);
_LIBZFS_H int zfs_send_resume(libzfs_handle_t *, sendflags_t *, int outfd,
const char *);
_LIBZFS_H int zfs_send_saved(zfs_handle_t *, sendflags_t *, int, const char *);
_LIBZFS_H nvlist_t *zfs_send_resume_token_to_nvlist(libzfs_handle_t *hdl,
const char *token);
_LIBZFS_H int zfs_promote(zfs_handle_t *);
_LIBZFS_H int zfs_hold(zfs_handle_t *, const char *, const char *,
boolean_t, int);
_LIBZFS_H int zfs_hold_nvl(zfs_handle_t *, int, nvlist_t *);
_LIBZFS_H int zfs_release(zfs_handle_t *, const char *, const char *,
boolean_t);
_LIBZFS_H int zfs_get_holds(zfs_handle_t *, nvlist_t **);
_LIBZFS_H uint64_t zvol_volsize_to_reservation(zpool_handle_t *, uint64_t,
nvlist_t *);
typedef int (*zfs_userspace_cb_t)(void *arg, const char *domain,
uid_t rid, uint64_t space);
_LIBZFS_H int zfs_userspace(zfs_handle_t *, zfs_userquota_prop_t,
zfs_userspace_cb_t, void *);
_LIBZFS_H int zfs_get_fsacl(zfs_handle_t *, nvlist_t **);
_LIBZFS_H int zfs_set_fsacl(zfs_handle_t *, boolean_t, nvlist_t *);
typedef struct recvflags {
/* print informational messages (ie, -v was specified) */
boolean_t verbose;
/* the destination is a prefix, not the exact fs (ie, -d) */
boolean_t isprefix;
/*
* Only the tail of the sent snapshot path is appended to the
* destination to determine the received snapshot name (ie, -e).
*/
boolean_t istail;
/* do not actually do the recv, just check if it would work (ie, -n) */
boolean_t dryrun;
/* rollback/destroy filesystems as necessary (eg, -F) */
boolean_t force;
/* set "canmount=off" on all modified filesystems */
boolean_t canmountoff;
/*
* Mark the file systems as "resumable" and do not destroy them if the
* receive is interrupted
*/
boolean_t resumable;
/* byteswap flag is used internally; callers need not specify */
boolean_t byteswap;
/* do not mount file systems as they are extracted (private) */
boolean_t nomount;
/* Was holds flag set in the compound header? */
boolean_t holds;
/* skip receive of snapshot holds */
boolean_t skipholds;
/* mount the filesystem unless nomount is specified */
boolean_t domount;
/* force unmount while recv snapshot (private) */
boolean_t forceunmount;
/* use this recv to check (and heal if needed) an existing snapshot */
boolean_t heal;
} recvflags_t;
_LIBZFS_H int zfs_receive(libzfs_handle_t *, const char *, nvlist_t *,
recvflags_t *, int, avl_tree_t *);
typedef enum diff_flags {
ZFS_DIFF_PARSEABLE = 1 << 0,
ZFS_DIFF_TIMESTAMP = 1 << 1,
ZFS_DIFF_CLASSIFY = 1 << 2,
ZFS_DIFF_NO_MANGLE = 1 << 3
} diff_flags_t;
_LIBZFS_H int zfs_show_diffs(zfs_handle_t *, int, const char *, const char *,
int);
/*
* Miscellaneous functions.
*/
_LIBZFS_H const char *zfs_type_to_name(zfs_type_t);
_LIBZFS_H void zfs_refresh_properties(zfs_handle_t *);
_LIBZFS_H int zfs_name_valid(const char *, zfs_type_t);
_LIBZFS_H zfs_handle_t *zfs_path_to_zhandle(libzfs_handle_t *, const char *,
zfs_type_t);
_LIBZFS_H int zfs_parent_name(zfs_handle_t *, char *, size_t);
_LIBZFS_H boolean_t zfs_dataset_exists(libzfs_handle_t *, const char *,
zfs_type_t);
_LIBZFS_H int zfs_spa_version(zfs_handle_t *, int *);
_LIBZFS_H boolean_t zfs_bookmark_exists(const char *path);
/*
* Mount support functions.
*/
_LIBZFS_H boolean_t is_mounted(libzfs_handle_t *, const char *special, char **);
_LIBZFS_H boolean_t zfs_is_mounted(zfs_handle_t *, char **);
_LIBZFS_H int zfs_mount(zfs_handle_t *, const char *, int);
_LIBZFS_H int zfs_mount_at(zfs_handle_t *, const char *, int, const char *);
_LIBZFS_H int zfs_unmount(zfs_handle_t *, const char *, int);
_LIBZFS_H int zfs_unmountall(zfs_handle_t *, int);
_LIBZFS_H int zfs_mount_delegation_check(void);
#if defined(__linux__) || defined(__APPLE__)
_LIBZFS_H int zfs_parse_mount_options(const char *mntopts,
unsigned long *mntflags, unsigned long *zfsflags, int sloppy, char *badopt,
char *mtabopt);
_LIBZFS_H void zfs_adjust_mount_options(zfs_handle_t *zhp, const char *mntpoint,
char *mntopts, char *mtabopt);
#endif
/*
* Share support functions.
*
* enum sa_protocol * lists are terminated with SA_NO_PROTOCOL,
* NULL means "all/any known to this libzfs".
*/
#define SA_NO_PROTOCOL -1
_LIBZFS_H boolean_t zfs_is_shared(zfs_handle_t *zhp, char **where,
const enum sa_protocol *proto);
_LIBZFS_H int zfs_share(zfs_handle_t *zhp, const enum sa_protocol *proto);
_LIBZFS_H int zfs_unshare(zfs_handle_t *zhp, const char *mountpoint,
const enum sa_protocol *proto);
_LIBZFS_H int zfs_unshareall(zfs_handle_t *zhp,
const enum sa_protocol *proto);
_LIBZFS_H void zfs_commit_shares(const enum sa_protocol *proto);
_LIBZFS_H void zfs_truncate_shares(const enum sa_protocol *proto);
_LIBZFS_H int zfs_nicestrtonum(libzfs_handle_t *, const char *, uint64_t *);
/*
* Utility functions to run an external process.
*/
#define STDOUT_VERBOSE 0x01
#define STDERR_VERBOSE 0x02
#define NO_DEFAULT_PATH 0x04 /* Don't use $PATH to lookup the command */
_LIBZFS_H int libzfs_run_process(const char *, char **, int);
_LIBZFS_H int libzfs_run_process_get_stdout(const char *, char *[], char *[],
char **[], int *);
_LIBZFS_H int libzfs_run_process_get_stdout_nopath(const char *, char *[],
char *[], char **[], int *);
_LIBZFS_H void libzfs_free_str_array(char **, int);
_LIBZFS_H boolean_t libzfs_envvar_is_set(const char *);
/*
* Utility functions for zfs version
*/
_LIBZFS_H const char *zfs_version_userland(void);
_LIBZFS_H char *zfs_version_kernel(void);
_LIBZFS_H int zfs_version_print(void);
/*
* Given a device or file, determine if it is part of a pool.
*/
_LIBZFS_H int zpool_in_use(libzfs_handle_t *, int, pool_state_t *, char **,
boolean_t *);
/*
* Label manipulation.
*/
_LIBZFS_H int zpool_clear_label(int);
_LIBZFS_H int zpool_set_bootenv(zpool_handle_t *, const nvlist_t *);
_LIBZFS_H int zpool_get_bootenv(zpool_handle_t *, nvlist_t **);
/*
* Management interfaces for SMB ACL files
*/
_LIBZFS_H int zfs_smb_acl_add(libzfs_handle_t *, char *, char *, char *);
_LIBZFS_H int zfs_smb_acl_remove(libzfs_handle_t *, char *, char *, char *);
_LIBZFS_H int zfs_smb_acl_purge(libzfs_handle_t *, char *, char *);
_LIBZFS_H int zfs_smb_acl_rename(libzfs_handle_t *, char *, char *, char *,
char *);
/*
* Enable and disable datasets within a pool by mounting/unmounting and
* sharing/unsharing them.
*/
_LIBZFS_H int zpool_enable_datasets(zpool_handle_t *, const char *, int);
_LIBZFS_H int zpool_disable_datasets(zpool_handle_t *, boolean_t);
_LIBZFS_H void zpool_disable_datasets_os(zpool_handle_t *, boolean_t);
_LIBZFS_H void zpool_disable_volume_os(const char *);
/*
* Parse a features file for -o compatibility
*/
typedef enum {
ZPOOL_COMPATIBILITY_OK,
ZPOOL_COMPATIBILITY_WARNTOKEN,
ZPOOL_COMPATIBILITY_BADTOKEN,
ZPOOL_COMPATIBILITY_BADFILE,
ZPOOL_COMPATIBILITY_NOFILES
} zpool_compat_status_t;
_LIBZFS_H zpool_compat_status_t zpool_load_compat(const char *,
boolean_t *, char *, size_t);
#ifdef __FreeBSD__
/*
* Attach/detach the given filesystem to/from the given jail.
*/
_LIBZFS_H int zfs_jail(zfs_handle_t *zhp, int jailid, int attach);
/*
* Set loader options for next boot.
*/
_LIBZFS_H int zpool_nextboot(libzfs_handle_t *, uint64_t, uint64_t,
const char *);
#endif /* __FreeBSD__ */
#ifdef __linux__
/*
* Add or delete the given filesystem to/from the given user namespace.
*/
_LIBZFS_H int zfs_userns(zfs_handle_t *zhp, const char *nspath, int attach);
#endif
#ifdef __cplusplus
}
#endif
#endif /* _LIBZFS_H */
diff --git a/sys/contrib/openzfs/include/sys/fs/zfs.h b/sys/contrib/openzfs/include/sys/fs/zfs.h
index a978a9db5075..c6f7dcca78b3 100644
--- a/sys/contrib/openzfs/include/sys/fs/zfs.h
+++ b/sys/contrib/openzfs/include/sys/fs/zfs.h
@@ -1,1868 +1,1893 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright 2011 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2013, 2017 Joyent, Inc. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019 Datto Inc.
* Portions Copyright 2010 Robert Milkowski
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
* Copyright (c) 2022 Hewlett Packard Enterprise Development LP.
*/
#ifndef _SYS_FS_ZFS_H
#define _SYS_FS_ZFS_H extern __attribute__((visibility("default")))
#include <sys/time.h>
#include <sys/zio_priority.h>
#ifdef __cplusplus
extern "C" {
#endif
/*
* Types and constants shared between userland and the kernel.
*/
/*
* Each dataset can be one of the following types. These constants can be
* combined into masks that can be passed to various functions.
*/
typedef enum {
ZFS_TYPE_INVALID = 0,
ZFS_TYPE_FILESYSTEM = (1 << 0),
ZFS_TYPE_SNAPSHOT = (1 << 1),
ZFS_TYPE_VOLUME = (1 << 2),
ZFS_TYPE_POOL = (1 << 3),
ZFS_TYPE_BOOKMARK = (1 << 4),
ZFS_TYPE_VDEV = (1 << 5),
} zfs_type_t;
/*
* NB: lzc_dataset_type should be updated whenever a new objset type is added,
* if it represents a real type of a dataset that can be created from userland.
*/
typedef enum dmu_objset_type {
DMU_OST_NONE,
DMU_OST_META,
DMU_OST_ZFS,
DMU_OST_ZVOL,
DMU_OST_OTHER, /* For testing only! */
DMU_OST_ANY, /* Be careful! */
DMU_OST_NUMTYPES
} dmu_objset_type_t;
#define ZFS_TYPE_DATASET \
(ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME | ZFS_TYPE_SNAPSHOT)
/*
* All of these include the terminating NUL byte.
*/
#define ZAP_MAXNAMELEN 256
#define ZAP_MAXVALUELEN (1024 * 8)
#define ZAP_OLDMAXVALUELEN 1024
#define ZFS_MAX_DATASET_NAME_LEN 256
/*
* Dataset properties are identified by these constants and must be added to
* the end of this list to ensure that external consumers are not affected
* by the change. If you make any changes to this list, be sure to update
* the property table in module/zcommon/zfs_prop.c.
*/
typedef enum {
ZPROP_CONT = -2,
ZPROP_INVAL = -1,
ZPROP_USERPROP = ZPROP_INVAL,
ZFS_PROP_TYPE = 0,
ZFS_PROP_CREATION,
ZFS_PROP_USED,
ZFS_PROP_AVAILABLE,
ZFS_PROP_REFERENCED,
ZFS_PROP_COMPRESSRATIO,
ZFS_PROP_MOUNTED,
ZFS_PROP_ORIGIN,
ZFS_PROP_QUOTA,
ZFS_PROP_RESERVATION,
ZFS_PROP_VOLSIZE,
ZFS_PROP_VOLBLOCKSIZE,
ZFS_PROP_RECORDSIZE,
ZFS_PROP_MOUNTPOINT,
ZFS_PROP_SHARENFS,
ZFS_PROP_CHECKSUM,
ZFS_PROP_COMPRESSION,
ZFS_PROP_ATIME,
ZFS_PROP_DEVICES,
ZFS_PROP_EXEC,
ZFS_PROP_SETUID,
ZFS_PROP_READONLY,
ZFS_PROP_ZONED,
ZFS_PROP_SNAPDIR,
ZFS_PROP_ACLMODE,
ZFS_PROP_ACLINHERIT,
ZFS_PROP_CREATETXG,
ZFS_PROP_NAME, /* not exposed to the user */
ZFS_PROP_CANMOUNT,
ZFS_PROP_ISCSIOPTIONS, /* not exposed to the user */
ZFS_PROP_XATTR,
ZFS_PROP_NUMCLONES, /* not exposed to the user */
ZFS_PROP_COPIES,
ZFS_PROP_VERSION,
ZFS_PROP_UTF8ONLY,
ZFS_PROP_NORMALIZE,
ZFS_PROP_CASE,
ZFS_PROP_VSCAN,
ZFS_PROP_NBMAND,
ZFS_PROP_SHARESMB,
ZFS_PROP_REFQUOTA,
ZFS_PROP_REFRESERVATION,
ZFS_PROP_GUID,
ZFS_PROP_PRIMARYCACHE,
ZFS_PROP_SECONDARYCACHE,
ZFS_PROP_USEDSNAP,
ZFS_PROP_USEDDS,
ZFS_PROP_USEDCHILD,
ZFS_PROP_USEDREFRESERV,
ZFS_PROP_USERACCOUNTING, /* not exposed to the user */
ZFS_PROP_STMF_SHAREINFO, /* not exposed to the user */
ZFS_PROP_DEFER_DESTROY,
ZFS_PROP_USERREFS,
ZFS_PROP_LOGBIAS,
ZFS_PROP_UNIQUE, /* not exposed to the user */
ZFS_PROP_OBJSETID,
ZFS_PROP_DEDUP,
ZFS_PROP_MLSLABEL,
ZFS_PROP_SYNC,
ZFS_PROP_DNODESIZE,
ZFS_PROP_REFRATIO,
ZFS_PROP_WRITTEN,
ZFS_PROP_CLONES,
ZFS_PROP_LOGICALUSED,
ZFS_PROP_LOGICALREFERENCED,
ZFS_PROP_INCONSISTENT, /* not exposed to the user */
ZFS_PROP_VOLMODE,
ZFS_PROP_FILESYSTEM_LIMIT,
ZFS_PROP_SNAPSHOT_LIMIT,
ZFS_PROP_FILESYSTEM_COUNT,
ZFS_PROP_SNAPSHOT_COUNT,
ZFS_PROP_SNAPDEV,
ZFS_PROP_ACLTYPE,
ZFS_PROP_SELINUX_CONTEXT,
ZFS_PROP_SELINUX_FSCONTEXT,
ZFS_PROP_SELINUX_DEFCONTEXT,
ZFS_PROP_SELINUX_ROOTCONTEXT,
ZFS_PROP_RELATIME,
ZFS_PROP_REDUNDANT_METADATA,
ZFS_PROP_OVERLAY,
ZFS_PROP_PREV_SNAP,
ZFS_PROP_RECEIVE_RESUME_TOKEN,
ZFS_PROP_ENCRYPTION,
ZFS_PROP_KEYLOCATION,
ZFS_PROP_KEYFORMAT,
ZFS_PROP_PBKDF2_SALT,
ZFS_PROP_PBKDF2_ITERS,
ZFS_PROP_ENCRYPTION_ROOT,
ZFS_PROP_KEY_GUID,
ZFS_PROP_KEYSTATUS,
ZFS_PROP_REMAPTXG, /* obsolete - no longer used */
ZFS_PROP_SPECIAL_SMALL_BLOCKS,
ZFS_PROP_IVSET_GUID, /* not exposed to the user */
ZFS_PROP_REDACTED,
ZFS_PROP_REDACT_SNAPS,
ZFS_PROP_SNAPSHOTS_CHANGED,
ZFS_PROP_PREFETCH,
ZFS_PROP_VOLTHREADING,
ZFS_NUM_PROPS
} zfs_prop_t;
typedef enum {
ZFS_PROP_USERUSED,
ZFS_PROP_USERQUOTA,
ZFS_PROP_GROUPUSED,
ZFS_PROP_GROUPQUOTA,
ZFS_PROP_USEROBJUSED,
ZFS_PROP_USEROBJQUOTA,
ZFS_PROP_GROUPOBJUSED,
ZFS_PROP_GROUPOBJQUOTA,
ZFS_PROP_PROJECTUSED,
ZFS_PROP_PROJECTQUOTA,
ZFS_PROP_PROJECTOBJUSED,
ZFS_PROP_PROJECTOBJQUOTA,
ZFS_NUM_USERQUOTA_PROPS
} zfs_userquota_prop_t;
_SYS_FS_ZFS_H const char *const zfs_userquota_prop_prefixes[
ZFS_NUM_USERQUOTA_PROPS];
/*
* Pool properties are identified by these constants and must be added to the
* end of this list to ensure that external consumers are not affected
* by the change. Properties must be registered in zfs_prop_init().
*/
typedef enum {
ZPOOL_PROP_INVAL = -1,
ZPOOL_PROP_NAME,
ZPOOL_PROP_SIZE,
ZPOOL_PROP_CAPACITY,
ZPOOL_PROP_ALTROOT,
ZPOOL_PROP_HEALTH,
ZPOOL_PROP_GUID,
ZPOOL_PROP_VERSION,
ZPOOL_PROP_BOOTFS,
ZPOOL_PROP_DELEGATION,
ZPOOL_PROP_AUTOREPLACE,
ZPOOL_PROP_CACHEFILE,
ZPOOL_PROP_FAILUREMODE,
ZPOOL_PROP_LISTSNAPS,
ZPOOL_PROP_AUTOEXPAND,
ZPOOL_PROP_DEDUPDITTO,
ZPOOL_PROP_DEDUPRATIO,
ZPOOL_PROP_FREE,
ZPOOL_PROP_ALLOCATED,
ZPOOL_PROP_READONLY,
ZPOOL_PROP_ASHIFT,
ZPOOL_PROP_COMMENT,
ZPOOL_PROP_EXPANDSZ,
ZPOOL_PROP_FREEING,
ZPOOL_PROP_FRAGMENTATION,
ZPOOL_PROP_LEAKED,
ZPOOL_PROP_MAXBLOCKSIZE,
ZPOOL_PROP_TNAME,
ZPOOL_PROP_MAXDNODESIZE,
ZPOOL_PROP_MULTIHOST,
ZPOOL_PROP_CHECKPOINT,
ZPOOL_PROP_LOAD_GUID,
ZPOOL_PROP_AUTOTRIM,
ZPOOL_PROP_COMPATIBILITY,
ZPOOL_PROP_BCLONEUSED,
ZPOOL_PROP_BCLONESAVED,
ZPOOL_PROP_BCLONERATIO,
ZPOOL_NUM_PROPS
} zpool_prop_t;
/* Small enough to not hog a whole line of printout in zpool(8). */
#define ZPROP_MAX_COMMENT 32
#define ZPROP_BOOLEAN_NA 2
#define ZPROP_VALUE "value"
#define ZPROP_SOURCE "source"
typedef enum {
ZPROP_SRC_NONE = 0x1,
ZPROP_SRC_DEFAULT = 0x2,
ZPROP_SRC_TEMPORARY = 0x4,
ZPROP_SRC_LOCAL = 0x8,
ZPROP_SRC_INHERITED = 0x10,
ZPROP_SRC_RECEIVED = 0x20
} zprop_source_t;
#define ZPROP_SRC_ALL 0x3f
#define ZPROP_SOURCE_VAL_RECVD "$recvd"
#define ZPROP_N_MORE_ERRORS "N_MORE_ERRORS"
/*
* Dataset flag implemented as a special entry in the props zap object
* indicating that the dataset has received properties on or after
* SPA_VERSION_RECVD_PROPS. The first such receive blows away local properties
* just as it did in earlier versions, and thereafter, local properties are
* preserved.
*/
#define ZPROP_HAS_RECVD "$hasrecvd"
typedef enum {
ZPROP_ERR_NOCLEAR = 0x1, /* failure to clear existing props */
ZPROP_ERR_NORESTORE = 0x2 /* failure to restore props on error */
} zprop_errflags_t;
typedef int (*zprop_func)(int, void *);
/*
* Properties to be set on the root file system of a new pool
* are stuffed into their own nvlist, which is then included in
* the properties nvlist with the pool properties.
*/
#define ZPOOL_ROOTFS_PROPS "root-props-nvl"
/*
* Length of 'written@' and 'written#'
*/
#define ZFS_WRITTEN_PROP_PREFIX_LEN 8
/*
* VDEV properties are identified by these constants and must be added to the
* end of this list to ensure that external consumers are not affected
* by the change. If you make any changes to this list, be sure to update
* the property table in usr/src/common/zfs/zpool_prop.c.
*/
typedef enum {
VDEV_PROP_INVAL = -1,
VDEV_PROP_USERPROP = VDEV_PROP_INVAL,
VDEV_PROP_NAME,
VDEV_PROP_CAPACITY,
VDEV_PROP_STATE,
VDEV_PROP_GUID,
VDEV_PROP_ASIZE,
VDEV_PROP_PSIZE,
VDEV_PROP_ASHIFT,
VDEV_PROP_SIZE,
VDEV_PROP_FREE,
VDEV_PROP_ALLOCATED,
VDEV_PROP_COMMENT,
VDEV_PROP_EXPANDSZ,
VDEV_PROP_FRAGMENTATION,
VDEV_PROP_BOOTSIZE,
VDEV_PROP_PARITY,
VDEV_PROP_PATH,
VDEV_PROP_DEVID,
VDEV_PROP_PHYS_PATH,
VDEV_PROP_ENC_PATH,
VDEV_PROP_FRU,
VDEV_PROP_PARENT,
VDEV_PROP_CHILDREN,
VDEV_PROP_NUMCHILDREN,
VDEV_PROP_READ_ERRORS,
VDEV_PROP_WRITE_ERRORS,
VDEV_PROP_CHECKSUM_ERRORS,
VDEV_PROP_INITIALIZE_ERRORS,
VDEV_PROP_OPS_NULL,
VDEV_PROP_OPS_READ,
VDEV_PROP_OPS_WRITE,
VDEV_PROP_OPS_FREE,
VDEV_PROP_OPS_CLAIM,
VDEV_PROP_OPS_TRIM,
VDEV_PROP_BYTES_NULL,
VDEV_PROP_BYTES_READ,
VDEV_PROP_BYTES_WRITE,
VDEV_PROP_BYTES_FREE,
VDEV_PROP_BYTES_CLAIM,
VDEV_PROP_BYTES_TRIM,
VDEV_PROP_REMOVING,
VDEV_PROP_ALLOCATING,
VDEV_PROP_FAILFAST,
VDEV_PROP_CHECKSUM_N,
VDEV_PROP_CHECKSUM_T,
VDEV_PROP_IO_N,
VDEV_PROP_IO_T,
+ VDEV_PROP_RAIDZ_EXPANDING,
VDEV_NUM_PROPS
} vdev_prop_t;
/*
* Dataset property functions shared between libzfs and kernel.
*/
_SYS_FS_ZFS_H const char *zfs_prop_default_string(zfs_prop_t);
_SYS_FS_ZFS_H uint64_t zfs_prop_default_numeric(zfs_prop_t);
_SYS_FS_ZFS_H boolean_t zfs_prop_readonly(zfs_prop_t);
_SYS_FS_ZFS_H boolean_t zfs_prop_visible(zfs_prop_t prop);
_SYS_FS_ZFS_H boolean_t zfs_prop_inheritable(zfs_prop_t);
_SYS_FS_ZFS_H boolean_t zfs_prop_setonce(zfs_prop_t);
_SYS_FS_ZFS_H boolean_t zfs_prop_encryption_key_param(zfs_prop_t);
_SYS_FS_ZFS_H boolean_t zfs_prop_valid_keylocation(const char *, boolean_t);
_SYS_FS_ZFS_H const char *zfs_prop_to_name(zfs_prop_t);
_SYS_FS_ZFS_H zfs_prop_t zfs_name_to_prop(const char *);
_SYS_FS_ZFS_H boolean_t zfs_prop_user(const char *);
_SYS_FS_ZFS_H boolean_t zfs_prop_userquota(const char *);
_SYS_FS_ZFS_H boolean_t zfs_prop_written(const char *);
_SYS_FS_ZFS_H int zfs_prop_index_to_string(zfs_prop_t, uint64_t, const char **);
_SYS_FS_ZFS_H int zfs_prop_string_to_index(zfs_prop_t, const char *,
uint64_t *);
_SYS_FS_ZFS_H uint64_t zfs_prop_random_value(zfs_prop_t, uint64_t seed);
_SYS_FS_ZFS_H boolean_t zfs_prop_valid_for_type(int, zfs_type_t, boolean_t);
/*
* Pool property functions shared between libzfs and kernel.
*/
_SYS_FS_ZFS_H zpool_prop_t zpool_name_to_prop(const char *);
_SYS_FS_ZFS_H const char *zpool_prop_to_name(zpool_prop_t);
_SYS_FS_ZFS_H const char *zpool_prop_default_string(zpool_prop_t);
_SYS_FS_ZFS_H uint64_t zpool_prop_default_numeric(zpool_prop_t);
_SYS_FS_ZFS_H boolean_t zpool_prop_readonly(zpool_prop_t);
_SYS_FS_ZFS_H boolean_t zpool_prop_setonce(zpool_prop_t);
_SYS_FS_ZFS_H boolean_t zpool_prop_feature(const char *);
_SYS_FS_ZFS_H boolean_t zpool_prop_unsupported(const char *);
_SYS_FS_ZFS_H int zpool_prop_index_to_string(zpool_prop_t, uint64_t,
const char **);
_SYS_FS_ZFS_H int zpool_prop_string_to_index(zpool_prop_t, const char *,
uint64_t *);
_SYS_FS_ZFS_H uint64_t zpool_prop_random_value(zpool_prop_t, uint64_t seed);
/*
* VDEV property functions shared between libzfs and kernel.
*/
_SYS_FS_ZFS_H vdev_prop_t vdev_name_to_prop(const char *);
_SYS_FS_ZFS_H boolean_t vdev_prop_user(const char *name);
_SYS_FS_ZFS_H const char *vdev_prop_to_name(vdev_prop_t);
_SYS_FS_ZFS_H const char *vdev_prop_default_string(vdev_prop_t);
_SYS_FS_ZFS_H uint64_t vdev_prop_default_numeric(vdev_prop_t);
_SYS_FS_ZFS_H boolean_t vdev_prop_readonly(vdev_prop_t prop);
_SYS_FS_ZFS_H int vdev_prop_index_to_string(vdev_prop_t, uint64_t,
const char **);
_SYS_FS_ZFS_H int vdev_prop_string_to_index(vdev_prop_t, const char *,
uint64_t *);
_SYS_FS_ZFS_H boolean_t zpool_prop_vdev(const char *name);
_SYS_FS_ZFS_H uint64_t vdev_prop_random_value(vdev_prop_t prop, uint64_t seed);
/*
* Definitions for the Delegation.
*/
typedef enum {
ZFS_DELEG_WHO_UNKNOWN = 0,
ZFS_DELEG_USER = 'u',
ZFS_DELEG_USER_SETS = 'U',
ZFS_DELEG_GROUP = 'g',
ZFS_DELEG_GROUP_SETS = 'G',
ZFS_DELEG_EVERYONE = 'e',
ZFS_DELEG_EVERYONE_SETS = 'E',
ZFS_DELEG_CREATE = 'c',
ZFS_DELEG_CREATE_SETS = 'C',
ZFS_DELEG_NAMED_SET = 's',
ZFS_DELEG_NAMED_SET_SETS = 'S'
} zfs_deleg_who_type_t;
typedef enum {
ZFS_DELEG_NONE = 0,
ZFS_DELEG_PERM_LOCAL = 1,
ZFS_DELEG_PERM_DESCENDENT = 2,
ZFS_DELEG_PERM_LOCALDESCENDENT = 3,
ZFS_DELEG_PERM_CREATE = 4
} zfs_deleg_inherit_t;
#define ZFS_DELEG_PERM_UID "uid"
#define ZFS_DELEG_PERM_GID "gid"
#define ZFS_DELEG_PERM_GROUPS "groups"
#define ZFS_MLSLABEL_DEFAULT "none"
#define ZFS_SMB_ACL_SRC "src"
#define ZFS_SMB_ACL_TARGET "target"
typedef enum {
ZFS_CANMOUNT_OFF = 0,
ZFS_CANMOUNT_ON = 1,
ZFS_CANMOUNT_NOAUTO = 2
} zfs_canmount_type_t;
typedef enum {
ZFS_LOGBIAS_LATENCY = 0,
ZFS_LOGBIAS_THROUGHPUT = 1
} zfs_logbias_op_t;
typedef enum zfs_share_op {
ZFS_SHARE_NFS = 0,
ZFS_UNSHARE_NFS = 1,
ZFS_SHARE_SMB = 2,
ZFS_UNSHARE_SMB = 3
} zfs_share_op_t;
typedef enum zfs_smb_acl_op {
ZFS_SMB_ACL_ADD,
ZFS_SMB_ACL_REMOVE,
ZFS_SMB_ACL_RENAME,
ZFS_SMB_ACL_PURGE
} zfs_smb_acl_op_t;
typedef enum zfs_cache_type {
ZFS_CACHE_NONE = 0,
ZFS_CACHE_METADATA = 1,
ZFS_CACHE_ALL = 2
} zfs_cache_type_t;
typedef enum {
ZFS_SYNC_STANDARD = 0,
ZFS_SYNC_ALWAYS = 1,
ZFS_SYNC_DISABLED = 2
} zfs_sync_type_t;
typedef enum {
ZFS_XATTR_OFF = 0,
ZFS_XATTR_DIR = 1,
ZFS_XATTR_SA = 2
} zfs_xattr_type_t;
typedef enum {
ZFS_DNSIZE_LEGACY = 0,
ZFS_DNSIZE_AUTO = 1,
ZFS_DNSIZE_1K = 1024,
ZFS_DNSIZE_2K = 2048,
ZFS_DNSIZE_4K = 4096,
ZFS_DNSIZE_8K = 8192,
ZFS_DNSIZE_16K = 16384
} zfs_dnsize_type_t;
typedef enum {
ZFS_REDUNDANT_METADATA_ALL,
ZFS_REDUNDANT_METADATA_MOST,
ZFS_REDUNDANT_METADATA_SOME,
ZFS_REDUNDANT_METADATA_NONE
} zfs_redundant_metadata_type_t;
typedef enum {
ZFS_VOLMODE_DEFAULT = 0,
ZFS_VOLMODE_GEOM = 1,
ZFS_VOLMODE_DEV = 2,
ZFS_VOLMODE_NONE = 3
} zfs_volmode_t;
typedef enum zfs_keystatus {
ZFS_KEYSTATUS_NONE = 0,
ZFS_KEYSTATUS_UNAVAILABLE,
ZFS_KEYSTATUS_AVAILABLE,
} zfs_keystatus_t;
typedef enum zfs_keyformat {
ZFS_KEYFORMAT_NONE = 0,
ZFS_KEYFORMAT_RAW,
ZFS_KEYFORMAT_HEX,
ZFS_KEYFORMAT_PASSPHRASE,
ZFS_KEYFORMAT_FORMATS
} zfs_keyformat_t;
typedef enum zfs_key_location {
ZFS_KEYLOCATION_NONE = 0,
ZFS_KEYLOCATION_PROMPT,
ZFS_KEYLOCATION_URI,
ZFS_KEYLOCATION_LOCATIONS
} zfs_keylocation_t;
typedef enum {
ZFS_PREFETCH_NONE = 0,
ZFS_PREFETCH_METADATA = 1,
ZFS_PREFETCH_ALL = 2
} zfs_prefetch_type_t;
#define DEFAULT_PBKDF2_ITERATIONS 350000
#define MIN_PBKDF2_ITERATIONS 100000
/*
* On-disk version number.
*/
#define SPA_VERSION_1 1ULL
#define SPA_VERSION_2 2ULL
#define SPA_VERSION_3 3ULL
#define SPA_VERSION_4 4ULL
#define SPA_VERSION_5 5ULL
#define SPA_VERSION_6 6ULL
#define SPA_VERSION_7 7ULL
#define SPA_VERSION_8 8ULL
#define SPA_VERSION_9 9ULL
#define SPA_VERSION_10 10ULL
#define SPA_VERSION_11 11ULL
#define SPA_VERSION_12 12ULL
#define SPA_VERSION_13 13ULL
#define SPA_VERSION_14 14ULL
#define SPA_VERSION_15 15ULL
#define SPA_VERSION_16 16ULL
#define SPA_VERSION_17 17ULL
#define SPA_VERSION_18 18ULL
#define SPA_VERSION_19 19ULL
#define SPA_VERSION_20 20ULL
#define SPA_VERSION_21 21ULL
#define SPA_VERSION_22 22ULL
#define SPA_VERSION_23 23ULL
#define SPA_VERSION_24 24ULL
#define SPA_VERSION_25 25ULL
#define SPA_VERSION_26 26ULL
#define SPA_VERSION_27 27ULL
#define SPA_VERSION_28 28ULL
#define SPA_VERSION_5000 5000ULL
/*
* The incrementing pool version number has been replaced by pool feature
* flags. For more details, see zfeature.c.
*/
#define SPA_VERSION SPA_VERSION_5000
#define SPA_VERSION_STRING "5000"
/*
* Symbolic names for the changes that caused a SPA_VERSION switch.
* Used in the code when checking for presence or absence of a feature.
* Feel free to define multiple symbolic names for each version if there
* were multiple changes to on-disk structures during that version.
*
* NOTE: When checking the current SPA_VERSION in your code, be sure
* to use spa_version() since it reports the version of the
* last synced uberblock. Checking the in-flight version can
* be dangerous in some cases.
*/
#define SPA_VERSION_INITIAL SPA_VERSION_1
#define SPA_VERSION_DITTO_BLOCKS SPA_VERSION_2
#define SPA_VERSION_SPARES SPA_VERSION_3
#define SPA_VERSION_RAIDZ2 SPA_VERSION_3
#define SPA_VERSION_BPOBJ_ACCOUNT SPA_VERSION_3
#define SPA_VERSION_RAIDZ_DEFLATE SPA_VERSION_3
#define SPA_VERSION_DNODE_BYTES SPA_VERSION_3
#define SPA_VERSION_ZPOOL_HISTORY SPA_VERSION_4
#define SPA_VERSION_GZIP_COMPRESSION SPA_VERSION_5
#define SPA_VERSION_BOOTFS SPA_VERSION_6
#define SPA_VERSION_SLOGS SPA_VERSION_7
#define SPA_VERSION_DELEGATED_PERMS SPA_VERSION_8
#define SPA_VERSION_FUID SPA_VERSION_9
#define SPA_VERSION_REFRESERVATION SPA_VERSION_9
#define SPA_VERSION_REFQUOTA SPA_VERSION_9
#define SPA_VERSION_UNIQUE_ACCURATE SPA_VERSION_9
#define SPA_VERSION_L2CACHE SPA_VERSION_10
#define SPA_VERSION_NEXT_CLONES SPA_VERSION_11
#define SPA_VERSION_ORIGIN SPA_VERSION_11
#define SPA_VERSION_DSL_SCRUB SPA_VERSION_11
#define SPA_VERSION_SNAP_PROPS SPA_VERSION_12
#define SPA_VERSION_USED_BREAKDOWN SPA_VERSION_13
#define SPA_VERSION_PASSTHROUGH_X SPA_VERSION_14
#define SPA_VERSION_USERSPACE SPA_VERSION_15
#define SPA_VERSION_STMF_PROP SPA_VERSION_16
#define SPA_VERSION_RAIDZ3 SPA_VERSION_17
#define SPA_VERSION_USERREFS SPA_VERSION_18
#define SPA_VERSION_HOLES SPA_VERSION_19
#define SPA_VERSION_ZLE_COMPRESSION SPA_VERSION_20
#define SPA_VERSION_DEDUP SPA_VERSION_21
#define SPA_VERSION_RECVD_PROPS SPA_VERSION_22
#define SPA_VERSION_SLIM_ZIL SPA_VERSION_23
#define SPA_VERSION_SA SPA_VERSION_24
#define SPA_VERSION_SCAN SPA_VERSION_25
#define SPA_VERSION_DIR_CLONES SPA_VERSION_26
#define SPA_VERSION_DEADLISTS SPA_VERSION_26
#define SPA_VERSION_FAST_SNAP SPA_VERSION_27
#define SPA_VERSION_MULTI_REPLACE SPA_VERSION_28
#define SPA_VERSION_BEFORE_FEATURES SPA_VERSION_28
#define SPA_VERSION_FEATURES SPA_VERSION_5000
#define SPA_VERSION_IS_SUPPORTED(v) \
(((v) >= SPA_VERSION_INITIAL && (v) <= SPA_VERSION_BEFORE_FEATURES) || \
((v) >= SPA_VERSION_FEATURES && (v) <= SPA_VERSION))
/*
* ZPL version - rev'd whenever an incompatible on-disk format change
* occurs. This is independent of SPA/DMU/ZAP versioning. You must
* also update the version_table[] and help message in zfs_prop.c.
*/
#define ZPL_VERSION_1 1ULL
#define ZPL_VERSION_2 2ULL
#define ZPL_VERSION_3 3ULL
#define ZPL_VERSION_4 4ULL
#define ZPL_VERSION_5 5ULL
#define ZPL_VERSION ZPL_VERSION_5
#define ZPL_VERSION_STRING "5"
#define ZPL_VERSION_INITIAL ZPL_VERSION_1
#define ZPL_VERSION_DIRENT_TYPE ZPL_VERSION_2
#define ZPL_VERSION_FUID ZPL_VERSION_3
#define ZPL_VERSION_NORMALIZATION ZPL_VERSION_3
#define ZPL_VERSION_SYSATTR ZPL_VERSION_3
#define ZPL_VERSION_USERSPACE ZPL_VERSION_4
#define ZPL_VERSION_SA ZPL_VERSION_5
/* Persistent L2ARC version */
#define L2ARC_PERSISTENT_VERSION_1 1ULL
#define L2ARC_PERSISTENT_VERSION L2ARC_PERSISTENT_VERSION_1
#define L2ARC_PERSISTENT_VERSION_STRING "1"
/* Rewind policy information */
#define ZPOOL_NO_REWIND 1 /* No policy - default behavior */
#define ZPOOL_NEVER_REWIND 2 /* Do not search for best txg or rewind */
#define ZPOOL_TRY_REWIND 4 /* Search for best txg, but do not rewind */
#define ZPOOL_DO_REWIND 8 /* Rewind to best txg w/in deferred frees */
#define ZPOOL_EXTREME_REWIND 16 /* Allow extreme measures to find best txg */
#define ZPOOL_REWIND_MASK 28 /* All the possible rewind bits */
#define ZPOOL_REWIND_POLICIES 31 /* All the possible policy bits */
typedef struct zpool_load_policy {
uint32_t zlp_rewind; /* rewind policy requested */
uint64_t zlp_maxmeta; /* max acceptable meta-data errors */
uint64_t zlp_maxdata; /* max acceptable data errors */
uint64_t zlp_txg; /* specific txg to load */
} zpool_load_policy_t;
/*
* The following are configuration names used in the nvlist describing a pool's
* configuration. New on-disk names should be prefixed with "<reversed-DNS>:"
* (e.g. "org.openzfs:") to avoid conflicting names being developed
* independently.
*/
#define ZPOOL_CONFIG_VERSION "version"
#define ZPOOL_CONFIG_POOL_NAME "name"
#define ZPOOL_CONFIG_POOL_STATE "state"
#define ZPOOL_CONFIG_POOL_TXG "txg"
#define ZPOOL_CONFIG_POOL_GUID "pool_guid"
#define ZPOOL_CONFIG_CREATE_TXG "create_txg"
#define ZPOOL_CONFIG_TOP_GUID "top_guid"
#define ZPOOL_CONFIG_VDEV_TREE "vdev_tree"
#define ZPOOL_CONFIG_TYPE "type"
#define ZPOOL_CONFIG_CHILDREN "children"
#define ZPOOL_CONFIG_ID "id"
#define ZPOOL_CONFIG_GUID "guid"
#define ZPOOL_CONFIG_INDIRECT_OBJECT "com.delphix:indirect_object"
#define ZPOOL_CONFIG_INDIRECT_BIRTHS "com.delphix:indirect_births"
#define ZPOOL_CONFIG_PREV_INDIRECT_VDEV "com.delphix:prev_indirect_vdev"
#define ZPOOL_CONFIG_PATH "path"
#define ZPOOL_CONFIG_DEVID "devid"
#define ZPOOL_CONFIG_SPARE_ID "spareid"
#define ZPOOL_CONFIG_METASLAB_ARRAY "metaslab_array"
#define ZPOOL_CONFIG_METASLAB_SHIFT "metaslab_shift"
#define ZPOOL_CONFIG_ASHIFT "ashift"
#define ZPOOL_CONFIG_ASIZE "asize"
#define ZPOOL_CONFIG_DTL "DTL"
#define ZPOOL_CONFIG_SCAN_STATS "scan_stats" /* not stored on disk */
#define ZPOOL_CONFIG_REMOVAL_STATS "removal_stats" /* not stored on disk */
#define ZPOOL_CONFIG_CHECKPOINT_STATS "checkpoint_stats" /* not on disk */
+#define ZPOOL_CONFIG_RAIDZ_EXPAND_STATS "raidz_expand_stats" /* not on disk */
#define ZPOOL_CONFIG_VDEV_STATS "vdev_stats" /* not stored on disk */
#define ZPOOL_CONFIG_INDIRECT_SIZE "indirect_size" /* not stored on disk */
/* container nvlist of extended stats */
#define ZPOOL_CONFIG_VDEV_STATS_EX "vdev_stats_ex"
/* Active queue read/write stats */
#define ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE "vdev_sync_r_active_queue"
#define ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE "vdev_sync_w_active_queue"
#define ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE "vdev_async_r_active_queue"
#define ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE "vdev_async_w_active_queue"
#define ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE "vdev_async_scrub_active_queue"
#define ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE "vdev_async_trim_active_queue"
#define ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE "vdev_rebuild_active_queue"
/* Queue sizes */
#define ZPOOL_CONFIG_VDEV_SYNC_R_PEND_QUEUE "vdev_sync_r_pend_queue"
#define ZPOOL_CONFIG_VDEV_SYNC_W_PEND_QUEUE "vdev_sync_w_pend_queue"
#define ZPOOL_CONFIG_VDEV_ASYNC_R_PEND_QUEUE "vdev_async_r_pend_queue"
#define ZPOOL_CONFIG_VDEV_ASYNC_W_PEND_QUEUE "vdev_async_w_pend_queue"
#define ZPOOL_CONFIG_VDEV_SCRUB_PEND_QUEUE "vdev_async_scrub_pend_queue"
#define ZPOOL_CONFIG_VDEV_TRIM_PEND_QUEUE "vdev_async_trim_pend_queue"
#define ZPOOL_CONFIG_VDEV_REBUILD_PEND_QUEUE "vdev_rebuild_pend_queue"
/* Latency read/write histogram stats */
#define ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO "vdev_tot_r_lat_histo"
#define ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO "vdev_tot_w_lat_histo"
#define ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO "vdev_disk_r_lat_histo"
#define ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO "vdev_disk_w_lat_histo"
#define ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO "vdev_sync_r_lat_histo"
#define ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO "vdev_sync_w_lat_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO "vdev_async_r_lat_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO "vdev_async_w_lat_histo"
#define ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO "vdev_scrub_histo"
#define ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO "vdev_trim_histo"
#define ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO "vdev_rebuild_histo"
/* Request size histograms */
#define ZPOOL_CONFIG_VDEV_SYNC_IND_R_HISTO "vdev_sync_ind_r_histo"
#define ZPOOL_CONFIG_VDEV_SYNC_IND_W_HISTO "vdev_sync_ind_w_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_IND_R_HISTO "vdev_async_ind_r_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_IND_W_HISTO "vdev_async_ind_w_histo"
#define ZPOOL_CONFIG_VDEV_IND_SCRUB_HISTO "vdev_ind_scrub_histo"
#define ZPOOL_CONFIG_VDEV_IND_TRIM_HISTO "vdev_ind_trim_histo"
#define ZPOOL_CONFIG_VDEV_IND_REBUILD_HISTO "vdev_ind_rebuild_histo"
#define ZPOOL_CONFIG_VDEV_SYNC_AGG_R_HISTO "vdev_sync_agg_r_histo"
#define ZPOOL_CONFIG_VDEV_SYNC_AGG_W_HISTO "vdev_sync_agg_w_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_AGG_R_HISTO "vdev_async_agg_r_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_AGG_W_HISTO "vdev_async_agg_w_histo"
#define ZPOOL_CONFIG_VDEV_AGG_SCRUB_HISTO "vdev_agg_scrub_histo"
#define ZPOOL_CONFIG_VDEV_AGG_TRIM_HISTO "vdev_agg_trim_histo"
#define ZPOOL_CONFIG_VDEV_AGG_REBUILD_HISTO "vdev_agg_rebuild_histo"
/* Number of slow IOs */
#define ZPOOL_CONFIG_VDEV_SLOW_IOS "vdev_slow_ios"
/* vdev enclosure sysfs path */
#define ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH "vdev_enc_sysfs_path"
#define ZPOOL_CONFIG_WHOLE_DISK "whole_disk"
#define ZPOOL_CONFIG_ERRCOUNT "error_count"
#define ZPOOL_CONFIG_NOT_PRESENT "not_present"
#define ZPOOL_CONFIG_SPARES "spares"
#define ZPOOL_CONFIG_IS_SPARE "is_spare"
#define ZPOOL_CONFIG_NPARITY "nparity"
+#define ZPOOL_CONFIG_RAIDZ_EXPANDING "raidz_expanding"
+#define ZPOOL_CONFIG_RAIDZ_EXPAND_TXGS "raidz_expand_txgs"
#define ZPOOL_CONFIG_HOSTID "hostid"
#define ZPOOL_CONFIG_HOSTNAME "hostname"
#define ZPOOL_CONFIG_LOADED_TIME "initial_load_time"
#define ZPOOL_CONFIG_UNSPARE "unspare"
#define ZPOOL_CONFIG_PHYS_PATH "phys_path"
#define ZPOOL_CONFIG_IS_LOG "is_log"
#define ZPOOL_CONFIG_L2CACHE "l2cache"
#define ZPOOL_CONFIG_HOLE_ARRAY "hole_array"
#define ZPOOL_CONFIG_VDEV_CHILDREN "vdev_children"
#define ZPOOL_CONFIG_IS_HOLE "is_hole"
#define ZPOOL_CONFIG_DDT_HISTOGRAM "ddt_histogram"
#define ZPOOL_CONFIG_DDT_OBJ_STATS "ddt_object_stats"
#define ZPOOL_CONFIG_DDT_STATS "ddt_stats"
#define ZPOOL_CONFIG_SPLIT "splitcfg"
#define ZPOOL_CONFIG_ORIG_GUID "orig_guid"
#define ZPOOL_CONFIG_SPLIT_GUID "split_guid"
#define ZPOOL_CONFIG_SPLIT_LIST "guid_list"
#define ZPOOL_CONFIG_NONALLOCATING "non_allocating"
#define ZPOOL_CONFIG_REMOVING "removing"
#define ZPOOL_CONFIG_RESILVER_TXG "resilver_txg"
#define ZPOOL_CONFIG_REBUILD_TXG "rebuild_txg"
#define ZPOOL_CONFIG_COMMENT "comment"
#define ZPOOL_CONFIG_SUSPENDED "suspended" /* not stored on disk */
#define ZPOOL_CONFIG_SUSPENDED_REASON "suspended_reason" /* not stored */
#define ZPOOL_CONFIG_TIMESTAMP "timestamp" /* not stored on disk */
#define ZPOOL_CONFIG_BOOTFS "bootfs" /* not stored on disk */
#define ZPOOL_CONFIG_MISSING_DEVICES "missing_vdevs" /* not stored on disk */
#define ZPOOL_CONFIG_LOAD_INFO "load_info" /* not stored on disk */
#define ZPOOL_CONFIG_REWIND_INFO "rewind_info" /* not stored on disk */
#define ZPOOL_CONFIG_UNSUP_FEAT "unsup_feat" /* not stored on disk */
#define ZPOOL_CONFIG_ENABLED_FEAT "enabled_feat" /* not stored on disk */
#define ZPOOL_CONFIG_CAN_RDONLY "can_rdonly" /* not stored on disk */
#define ZPOOL_CONFIG_FEATURES_FOR_READ "features_for_read"
#define ZPOOL_CONFIG_FEATURE_STATS "feature_stats" /* not stored on disk */
#define ZPOOL_CONFIG_ERRATA "errata" /* not stored on disk */
#define ZPOOL_CONFIG_VDEV_ROOT_ZAP "com.klarasystems:vdev_zap_root"
#define ZPOOL_CONFIG_VDEV_TOP_ZAP "com.delphix:vdev_zap_top"
#define ZPOOL_CONFIG_VDEV_LEAF_ZAP "com.delphix:vdev_zap_leaf"
#define ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS "com.delphix:has_per_vdev_zaps"
#define ZPOOL_CONFIG_RESILVER_DEFER "com.datto:resilver_defer"
#define ZPOOL_CONFIG_CACHEFILE "cachefile" /* not stored on disk */
#define ZPOOL_CONFIG_MMP_STATE "mmp_state" /* not stored on disk */
#define ZPOOL_CONFIG_MMP_TXG "mmp_txg" /* not stored on disk */
#define ZPOOL_CONFIG_MMP_SEQ "mmp_seq" /* not stored on disk */
#define ZPOOL_CONFIG_MMP_HOSTNAME "mmp_hostname" /* not stored on disk */
#define ZPOOL_CONFIG_MMP_HOSTID "mmp_hostid" /* not stored on disk */
#define ZPOOL_CONFIG_ALLOCATION_BIAS "alloc_bias" /* not stored on disk */
#define ZPOOL_CONFIG_EXPANSION_TIME "expansion_time" /* not stored */
#define ZPOOL_CONFIG_REBUILD_STATS "org.openzfs:rebuild_stats"
#define ZPOOL_CONFIG_COMPATIBILITY "compatibility"
/*
* The persistent vdev state is stored as separate values rather than a single
* 'vdev_state' entry. This is because a device can be in multiple states, such
* as offline and degraded.
*/
#define ZPOOL_CONFIG_OFFLINE "offline"
#define ZPOOL_CONFIG_FAULTED "faulted"
#define ZPOOL_CONFIG_DEGRADED "degraded"
#define ZPOOL_CONFIG_REMOVED "removed"
#define ZPOOL_CONFIG_FRU "fru"
#define ZPOOL_CONFIG_AUX_STATE "aux_state"
/* Pool load policy parameters */
#define ZPOOL_LOAD_POLICY "load-policy"
#define ZPOOL_LOAD_REWIND_POLICY "load-rewind-policy"
#define ZPOOL_LOAD_REQUEST_TXG "load-request-txg"
#define ZPOOL_LOAD_META_THRESH "load-meta-thresh"
#define ZPOOL_LOAD_DATA_THRESH "load-data-thresh"
/* Rewind data discovered */
#define ZPOOL_CONFIG_LOAD_TIME "rewind_txg_ts"
#define ZPOOL_CONFIG_LOAD_META_ERRORS "verify_meta_errors"
#define ZPOOL_CONFIG_LOAD_DATA_ERRORS "verify_data_errors"
#define ZPOOL_CONFIG_REWIND_TIME "seconds_of_rewind"
/* dRAID configuration */
#define ZPOOL_CONFIG_DRAID_NDATA "draid_ndata"
#define ZPOOL_CONFIG_DRAID_NSPARES "draid_nspares"
#define ZPOOL_CONFIG_DRAID_NGROUPS "draid_ngroups"
#define VDEV_TYPE_ROOT "root"
#define VDEV_TYPE_MIRROR "mirror"
#define VDEV_TYPE_REPLACING "replacing"
#define VDEV_TYPE_RAIDZ "raidz"
#define VDEV_TYPE_DRAID "draid"
#define VDEV_TYPE_DRAID_SPARE "dspare"
#define VDEV_TYPE_DISK "disk"
#define VDEV_TYPE_FILE "file"
#define VDEV_TYPE_MISSING "missing"
#define VDEV_TYPE_HOLE "hole"
#define VDEV_TYPE_SPARE "spare"
#define VDEV_TYPE_LOG "log"
#define VDEV_TYPE_L2CACHE "l2cache"
#define VDEV_TYPE_INDIRECT "indirect"
#define VDEV_RAIDZ_MAXPARITY 3
#define VDEV_DRAID_MAXPARITY 3
#define VDEV_DRAID_MIN_CHILDREN 2
#define VDEV_DRAID_MAX_CHILDREN UINT8_MAX
/* VDEV_TOP_ZAP_* are used in top-level vdev ZAP objects. */
#define VDEV_TOP_ZAP_INDIRECT_OBSOLETE_SM \
"com.delphix:indirect_obsolete_sm"
#define VDEV_TOP_ZAP_OBSOLETE_COUNTS_ARE_PRECISE \
"com.delphix:obsolete_counts_are_precise"
#define VDEV_TOP_ZAP_POOL_CHECKPOINT_SM \
"com.delphix:pool_checkpoint_sm"
#define VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS \
"com.delphix:ms_unflushed_phys_txgs"
#define VDEV_TOP_ZAP_VDEV_REBUILD_PHYS \
"org.openzfs:vdev_rebuild"
#define VDEV_TOP_ZAP_ALLOCATION_BIAS \
"org.zfsonlinux:allocation_bias"
+#define VDEV_TOP_ZAP_RAIDZ_EXPAND_STATE \
+ "org.openzfs:raidz_expand_state"
+#define VDEV_TOP_ZAP_RAIDZ_EXPAND_START_TIME \
+ "org.openzfs:raidz_expand_start_time"
+#define VDEV_TOP_ZAP_RAIDZ_EXPAND_END_TIME \
+ "org.openzfs:raidz_expand_end_time"
+#define VDEV_TOP_ZAP_RAIDZ_EXPAND_BYTES_COPIED \
+ "org.openzfs:raidz_expand_bytes_copied"
+
/* vdev metaslab allocation bias */
#define VDEV_ALLOC_BIAS_LOG "log"
#define VDEV_ALLOC_BIAS_SPECIAL "special"
#define VDEV_ALLOC_BIAS_DEDUP "dedup"
/* vdev initialize state */
#define VDEV_LEAF_ZAP_INITIALIZE_LAST_OFFSET \
"com.delphix:next_offset_to_initialize"
#define VDEV_LEAF_ZAP_INITIALIZE_STATE \
"com.delphix:vdev_initialize_state"
#define VDEV_LEAF_ZAP_INITIALIZE_ACTION_TIME \
"com.delphix:vdev_initialize_action_time"
/* vdev TRIM state */
#define VDEV_LEAF_ZAP_TRIM_LAST_OFFSET \
"org.zfsonlinux:next_offset_to_trim"
#define VDEV_LEAF_ZAP_TRIM_STATE \
"org.zfsonlinux:vdev_trim_state"
#define VDEV_LEAF_ZAP_TRIM_ACTION_TIME \
"org.zfsonlinux:vdev_trim_action_time"
#define VDEV_LEAF_ZAP_TRIM_RATE \
"org.zfsonlinux:vdev_trim_rate"
#define VDEV_LEAF_ZAP_TRIM_PARTIAL \
"org.zfsonlinux:vdev_trim_partial"
#define VDEV_LEAF_ZAP_TRIM_SECURE \
"org.zfsonlinux:vdev_trim_secure"
/*
* This is needed in userland to report the minimum necessary device size.
*/
#define SPA_MINDEVSIZE (64ULL << 20)
/*
* Set if the fragmentation has not yet been calculated. This can happen
* because the space maps have not been upgraded or the histogram feature
* is not enabled.
*/
#define ZFS_FRAG_INVALID UINT64_MAX
/*
* The location of the pool configuration repository, shared between kernel and
* userland.
*/
#define ZPOOL_CACHE_BOOT "/boot/zfs/zpool.cache"
#define ZPOOL_CACHE "/etc/zfs/zpool.cache"
/*
* Settings for zpool compatibility features files
*/
#define ZPOOL_SYSCONF_COMPAT_D SYSCONFDIR "/zfs/compatibility.d"
#define ZPOOL_DATA_COMPAT_D PKGDATADIR "/compatibility.d"
#define ZPOOL_COMPAT_MAXSIZE 16384
/*
* Hard-wired compatibility settings
*/
#define ZPOOL_COMPAT_LEGACY "legacy"
#define ZPOOL_COMPAT_OFF "off"
/*
* vdev states are ordered from least to most healthy.
* A vdev that's CANT_OPEN or below is considered unusable.
*/
typedef enum vdev_state {
VDEV_STATE_UNKNOWN = 0, /* Uninitialized vdev */
VDEV_STATE_CLOSED, /* Not currently open */
VDEV_STATE_OFFLINE, /* Not allowed to open */
VDEV_STATE_REMOVED, /* Explicitly removed from system */
VDEV_STATE_CANT_OPEN, /* Tried to open, but failed */
VDEV_STATE_FAULTED, /* External request to fault device */
VDEV_STATE_DEGRADED, /* Replicated vdev with unhealthy kids */
VDEV_STATE_HEALTHY /* Presumed good */
} vdev_state_t;
#define VDEV_STATE_ONLINE VDEV_STATE_HEALTHY
/*
* vdev aux states. When a vdev is in the CANT_OPEN state, the aux field
* of the vdev stats structure uses these constants to distinguish why.
*/
typedef enum vdev_aux {
VDEV_AUX_NONE, /* no error */
VDEV_AUX_OPEN_FAILED, /* ldi_open_*() or vn_open() failed */
VDEV_AUX_CORRUPT_DATA, /* bad label or disk contents */
VDEV_AUX_NO_REPLICAS, /* insufficient number of replicas */
VDEV_AUX_BAD_GUID_SUM, /* vdev guid sum doesn't match */
VDEV_AUX_TOO_SMALL, /* vdev size is too small */
VDEV_AUX_BAD_LABEL, /* the label is OK but invalid */
VDEV_AUX_VERSION_NEWER, /* on-disk version is too new */
VDEV_AUX_VERSION_OLDER, /* on-disk version is too old */
VDEV_AUX_UNSUP_FEAT, /* unsupported features */
VDEV_AUX_SPARED, /* hot spare used in another pool */
VDEV_AUX_ERR_EXCEEDED, /* too many errors */
VDEV_AUX_IO_FAILURE, /* experienced I/O failure */
VDEV_AUX_BAD_LOG, /* cannot read log chain(s) */
VDEV_AUX_EXTERNAL, /* external diagnosis or forced fault */
VDEV_AUX_SPLIT_POOL, /* vdev was split off into another pool */
VDEV_AUX_BAD_ASHIFT, /* vdev ashift is invalid */
VDEV_AUX_EXTERNAL_PERSIST, /* persistent forced fault */
VDEV_AUX_ACTIVE, /* vdev active on a different host */
VDEV_AUX_CHILDREN_OFFLINE, /* all children are offline */
VDEV_AUX_ASHIFT_TOO_BIG, /* vdev's min block size is too large */
} vdev_aux_t;
/*
* pool state. The following states are written to disk as part of the normal
* SPA lifecycle: ACTIVE, EXPORTED, DESTROYED, SPARE, L2CACHE. The remaining
* states are software abstractions used at various levels to communicate
* pool state.
*/
typedef enum pool_state {
POOL_STATE_ACTIVE = 0, /* In active use */
POOL_STATE_EXPORTED, /* Explicitly exported */
POOL_STATE_DESTROYED, /* Explicitly destroyed */
POOL_STATE_SPARE, /* Reserved for hot spare use */
POOL_STATE_L2CACHE, /* Level 2 ARC device */
POOL_STATE_UNINITIALIZED, /* Internal spa_t state */
POOL_STATE_UNAVAIL, /* Internal libzfs state */
POOL_STATE_POTENTIALLY_ACTIVE /* Internal libzfs state */
} pool_state_t;
/*
* mmp state. The following states provide additional detail describing
* why a pool couldn't be safely imported.
*/
typedef enum mmp_state {
MMP_STATE_ACTIVE = 0, /* In active use */
MMP_STATE_INACTIVE, /* Inactive and safe to import */
MMP_STATE_NO_HOSTID /* System hostid is not set */
} mmp_state_t;
/*
* Scan Functions.
*/
typedef enum pool_scan_func {
POOL_SCAN_NONE,
POOL_SCAN_SCRUB,
POOL_SCAN_RESILVER,
POOL_SCAN_ERRORSCRUB,
POOL_SCAN_FUNCS
} pool_scan_func_t;
/*
* Used to control scrub pause and resume.
*/
typedef enum pool_scrub_cmd {
POOL_SCRUB_NORMAL = 0,
POOL_SCRUB_PAUSE,
POOL_SCRUB_FLAGS_END
} pool_scrub_cmd_t;
typedef enum {
CS_NONE,
CS_CHECKPOINT_EXISTS,
CS_CHECKPOINT_DISCARDING,
CS_NUM_STATES
} checkpoint_state_t;
typedef struct pool_checkpoint_stat {
uint64_t pcs_state; /* checkpoint_state_t */
uint64_t pcs_start_time; /* time checkpoint/discard started */
uint64_t pcs_space; /* checkpointed space */
} pool_checkpoint_stat_t;
/*
* ZIO types. Needed to interpret vdev statistics below.
*/
typedef enum zio_type {
ZIO_TYPE_NULL = 0,
ZIO_TYPE_READ,
ZIO_TYPE_WRITE,
ZIO_TYPE_FREE,
ZIO_TYPE_CLAIM,
ZIO_TYPE_IOCTL,
ZIO_TYPE_TRIM,
ZIO_TYPES
} zio_type_t;
/*
* Pool statistics. Note: all fields should be 64-bit because this
* is passed between kernel and userland as an nvlist uint64 array.
*/
typedef struct pool_scan_stat {
/* values stored on disk */
uint64_t pss_func; /* pool_scan_func_t */
uint64_t pss_state; /* dsl_scan_state_t */
uint64_t pss_start_time; /* scan start time */
uint64_t pss_end_time; /* scan end time */
uint64_t pss_to_examine; /* total bytes to scan */
uint64_t pss_examined; /* total bytes located by scanner */
uint64_t pss_skipped; /* total bytes skipped by scanner */
uint64_t pss_processed; /* total processed bytes */
uint64_t pss_errors; /* scan errors */
/* values not stored on disk */
uint64_t pss_pass_exam; /* examined bytes per scan pass */
uint64_t pss_pass_start; /* start time of a scan pass */
uint64_t pss_pass_scrub_pause; /* pause time of a scrub pass */
/* cumulative time scrub spent paused, needed for rate calculation */
uint64_t pss_pass_scrub_spent_paused;
uint64_t pss_pass_issued; /* issued bytes per scan pass */
uint64_t pss_issued; /* total bytes checked by scanner */
/* error scrub values stored on disk */
uint64_t pss_error_scrub_func; /* pool_scan_func_t */
uint64_t pss_error_scrub_state; /* dsl_scan_state_t */
uint64_t pss_error_scrub_start; /* error scrub start time */
uint64_t pss_error_scrub_end; /* error scrub end time */
uint64_t pss_error_scrub_examined; /* error blocks issued I/O */
/* error blocks to be issued I/O */
uint64_t pss_error_scrub_to_be_examined;
/* error scrub values not stored on disk */
/* error scrub pause time in milliseconds */
uint64_t pss_pass_error_scrub_pause;
} 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 struct pool_raidz_expand_stat {
+ uint64_t pres_state; /* dsl_scan_state_t */
+ uint64_t pres_expanding_vdev;
+ uint64_t pres_start_time;
+ uint64_t pres_end_time;
+ uint64_t pres_to_reflow; /* bytes that need to be moved */
+ uint64_t pres_reflowed; /* bytes moved so far */
+ uint64_t pres_waiting_for_resilver;
+} pool_raidz_expand_stat_t;
+
typedef enum dsl_scan_state {
DSS_NONE,
DSS_SCANNING,
DSS_FINISHED,
DSS_CANCELED,
DSS_ERRORSCRUBBING,
DSS_NUM_STATES
} dsl_scan_state_t;
typedef struct vdev_rebuild_stat {
uint64_t vrs_state; /* vdev_rebuild_state_t */
uint64_t vrs_start_time; /* time_t */
uint64_t vrs_end_time; /* time_t */
uint64_t vrs_scan_time_ms; /* total run time (millisecs) */
uint64_t vrs_bytes_scanned; /* allocated bytes scanned */
uint64_t vrs_bytes_issued; /* read bytes issued */
uint64_t vrs_bytes_rebuilt; /* rebuilt bytes */
uint64_t vrs_bytes_est; /* total bytes to scan */
uint64_t vrs_errors; /* scanning errors */
uint64_t vrs_pass_time_ms; /* pass run time (millisecs) */
uint64_t vrs_pass_bytes_scanned; /* bytes scanned since start/resume */
uint64_t vrs_pass_bytes_issued; /* bytes rebuilt since start/resume */
uint64_t vrs_pass_bytes_skipped; /* bytes skipped since start/resume */
} vdev_rebuild_stat_t;
/*
* Errata described by https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-ER.
* The ordering of this enum must be maintained to ensure the errata identifiers
* map to the correct documentation. New errata may only be appended to the
* list and must contain corresponding documentation at the above link.
*/
typedef enum zpool_errata {
ZPOOL_ERRATA_NONE,
ZPOOL_ERRATA_ZOL_2094_SCRUB,
ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY,
ZPOOL_ERRATA_ZOL_6845_ENCRYPTION,
ZPOOL_ERRATA_ZOL_8308_ENCRYPTION,
} zpool_errata_t;
/*
* Vdev statistics. Note: all fields should be 64-bit because this
* is passed between kernel and user land as an nvlist uint64 array.
*
* The vs_ops[] and vs_bytes[] arrays must always be an array size of 6 in
* order to keep subsequent members at their known fixed offsets. When
* adding a new field it must be added to the end the structure.
*/
#define VS_ZIO_TYPES 6
typedef struct vdev_stat {
hrtime_t vs_timestamp; /* time since vdev load */
uint64_t vs_state; /* vdev state */
uint64_t vs_aux; /* see vdev_aux_t */
uint64_t vs_alloc; /* space allocated */
uint64_t vs_space; /* total capacity */
uint64_t vs_dspace; /* deflated capacity */
uint64_t vs_rsize; /* replaceable dev size */
uint64_t vs_esize; /* expandable dev size */
uint64_t vs_ops[VS_ZIO_TYPES]; /* operation count */
uint64_t vs_bytes[VS_ZIO_TYPES]; /* bytes read/written */
uint64_t vs_read_errors; /* read errors */
uint64_t vs_write_errors; /* write errors */
uint64_t vs_checksum_errors; /* checksum errors */
uint64_t vs_initialize_errors; /* initializing errors */
uint64_t vs_self_healed; /* self-healed bytes */
uint64_t vs_scan_removing; /* removing? */
uint64_t vs_scan_processed; /* scan processed bytes */
uint64_t vs_fragmentation; /* device fragmentation */
uint64_t vs_initialize_bytes_done; /* bytes initialized */
uint64_t vs_initialize_bytes_est; /* total bytes to initialize */
uint64_t vs_initialize_state; /* vdev_initializing_state_t */
uint64_t vs_initialize_action_time; /* time_t */
uint64_t vs_checkpoint_space; /* checkpoint-consumed space */
uint64_t vs_resilver_deferred; /* resilver deferred */
uint64_t vs_slow_ios; /* slow IOs */
uint64_t vs_trim_errors; /* trimming errors */
uint64_t vs_trim_notsup; /* supported by device */
uint64_t vs_trim_bytes_done; /* bytes trimmed */
uint64_t vs_trim_bytes_est; /* total bytes to trim */
uint64_t vs_trim_state; /* vdev_trim_state_t */
uint64_t vs_trim_action_time; /* time_t */
uint64_t vs_rebuild_processed; /* bytes rebuilt */
uint64_t vs_configured_ashift; /* TLV vdev_ashift */
uint64_t vs_logical_ashift; /* vdev_logical_ashift */
uint64_t vs_physical_ashift; /* vdev_physical_ashift */
uint64_t vs_noalloc; /* allocations halted? */
uint64_t vs_pspace; /* physical capacity */
} vdev_stat_t;
#define VDEV_STAT_VALID(field, uint64_t_field_count) \
((uint64_t_field_count * sizeof (uint64_t)) >= \
(offsetof(vdev_stat_t, field) + sizeof (((vdev_stat_t *)NULL)->field)))
/*
* Extended stats
*
* These are stats which aren't included in the original iostat output. For
* convenience, they are grouped together in vdev_stat_ex, although each stat
* is individually exported as an nvlist.
*/
typedef struct vdev_stat_ex {
/* Number of ZIOs issued to disk and waiting to finish */
uint64_t vsx_active_queue[ZIO_PRIORITY_NUM_QUEUEABLE];
/* Number of ZIOs pending to be issued to disk */
uint64_t vsx_pend_queue[ZIO_PRIORITY_NUM_QUEUEABLE];
/*
* Below are the histograms for various latencies. Buckets are in
* units of nanoseconds.
*/
/*
* 2^37 nanoseconds = 134s. Timeouts will probably start kicking in
* before this.
*/
#define VDEV_L_HISTO_BUCKETS 37 /* Latency histo buckets */
#define VDEV_RQ_HISTO_BUCKETS 25 /* Request size histo buckets */
/* Amount of time in ZIO queue (ns) */
uint64_t vsx_queue_histo[ZIO_PRIORITY_NUM_QUEUEABLE]
[VDEV_L_HISTO_BUCKETS];
/* Total ZIO latency (ns). Includes queuing and disk access time */
uint64_t vsx_total_histo[ZIO_TYPES][VDEV_L_HISTO_BUCKETS];
/* Amount of time to read/write the disk (ns) */
uint64_t vsx_disk_histo[ZIO_TYPES][VDEV_L_HISTO_BUCKETS];
/* "lookup the bucket for a value" histogram macros */
#define HISTO(val, buckets) (val != 0 ? MIN(highbit64(val) - 1, \
buckets - 1) : 0)
#define L_HISTO(a) HISTO(a, VDEV_L_HISTO_BUCKETS)
#define RQ_HISTO(a) HISTO(a, VDEV_RQ_HISTO_BUCKETS)
/* Physical IO histogram */
uint64_t vsx_ind_histo[ZIO_PRIORITY_NUM_QUEUEABLE]
[VDEV_RQ_HISTO_BUCKETS];
/* Delegated (aggregated) physical IO histogram */
uint64_t vsx_agg_histo[ZIO_PRIORITY_NUM_QUEUEABLE]
[VDEV_RQ_HISTO_BUCKETS];
} vdev_stat_ex_t;
/*
* Initialize functions.
*/
typedef enum pool_initialize_func {
POOL_INITIALIZE_START,
POOL_INITIALIZE_CANCEL,
POOL_INITIALIZE_SUSPEND,
POOL_INITIALIZE_UNINIT,
POOL_INITIALIZE_FUNCS
} pool_initialize_func_t;
/*
* TRIM functions.
*/
typedef enum pool_trim_func {
POOL_TRIM_START,
POOL_TRIM_CANCEL,
POOL_TRIM_SUSPEND,
POOL_TRIM_FUNCS
} pool_trim_func_t;
/*
* DDT statistics. Note: all fields should be 64-bit because this
* is passed between kernel and userland as an nvlist uint64 array.
*/
typedef struct ddt_object {
uint64_t ddo_count; /* number of elements in ddt */
uint64_t ddo_dspace; /* size of ddt on disk */
uint64_t ddo_mspace; /* size of ddt in-core */
} ddt_object_t;
typedef struct ddt_stat {
uint64_t dds_blocks; /* blocks */
uint64_t dds_lsize; /* logical size */
uint64_t dds_psize; /* physical size */
uint64_t dds_dsize; /* deflated allocated size */
uint64_t dds_ref_blocks; /* referenced blocks */
uint64_t dds_ref_lsize; /* referenced lsize * refcnt */
uint64_t dds_ref_psize; /* referenced psize * refcnt */
uint64_t dds_ref_dsize; /* referenced dsize * refcnt */
} ddt_stat_t;
typedef struct ddt_histogram {
ddt_stat_t ddh_stat[64]; /* power-of-two histogram buckets */
} ddt_histogram_t;
#define ZVOL_DRIVER "zvol"
#define ZFS_DRIVER "zfs"
#define ZFS_DEV "/dev/zfs"
#define ZFS_DEVDIR "/dev"
#define ZFS_SUPER_MAGIC 0x2fc12fc1
/* general zvol path */
#define ZVOL_DIR "/dev/zvol/"
#define ZVOL_MAJOR 230
#define ZVOL_MINOR_BITS 4
#define ZVOL_MINOR_MASK ((1U << ZVOL_MINOR_BITS) - 1)
#define ZVOL_MINORS (1 << 4)
#define ZVOL_DEV_NAME "zd"
#define ZVOL_PROP_NAME "name"
#define ZVOL_DEFAULT_BLOCKSIZE 16384
typedef enum {
VDEV_INITIALIZE_NONE,
VDEV_INITIALIZE_ACTIVE,
VDEV_INITIALIZE_CANCELED,
VDEV_INITIALIZE_SUSPENDED,
VDEV_INITIALIZE_COMPLETE
} vdev_initializing_state_t;
typedef enum {
VDEV_TRIM_NONE,
VDEV_TRIM_ACTIVE,
VDEV_TRIM_CANCELED,
VDEV_TRIM_SUSPENDED,
VDEV_TRIM_COMPLETE,
} vdev_trim_state_t;
typedef enum {
VDEV_REBUILD_NONE,
VDEV_REBUILD_ACTIVE,
VDEV_REBUILD_CANCELED,
VDEV_REBUILD_COMPLETE,
} vdev_rebuild_state_t;
/*
* nvlist name constants. Facilitate restricting snapshot iteration range for
* the "list next snapshot" ioctl
*/
#define SNAP_ITER_MIN_TXG "snap_iter_min_txg"
#define SNAP_ITER_MAX_TXG "snap_iter_max_txg"
/*
* /dev/zfs ioctl numbers.
*
* These numbers cannot change over time. New ioctl numbers must be appended.
*/
typedef enum zfs_ioc {
/*
* Core features - 88/128 numbers reserved.
*/
#ifdef __FreeBSD__
ZFS_IOC_FIRST = 0,
#else
ZFS_IOC_FIRST = ('Z' << 8),
#endif
ZFS_IOC = ZFS_IOC_FIRST,
ZFS_IOC_POOL_CREATE = ZFS_IOC_FIRST, /* 0x5a00 */
ZFS_IOC_POOL_DESTROY, /* 0x5a01 */
ZFS_IOC_POOL_IMPORT, /* 0x5a02 */
ZFS_IOC_POOL_EXPORT, /* 0x5a03 */
ZFS_IOC_POOL_CONFIGS, /* 0x5a04 */
ZFS_IOC_POOL_STATS, /* 0x5a05 */
ZFS_IOC_POOL_TRYIMPORT, /* 0x5a06 */
ZFS_IOC_POOL_SCAN, /* 0x5a07 */
ZFS_IOC_POOL_FREEZE, /* 0x5a08 */
ZFS_IOC_POOL_UPGRADE, /* 0x5a09 */
ZFS_IOC_POOL_GET_HISTORY, /* 0x5a0a */
ZFS_IOC_VDEV_ADD, /* 0x5a0b */
ZFS_IOC_VDEV_REMOVE, /* 0x5a0c */
ZFS_IOC_VDEV_SET_STATE, /* 0x5a0d */
ZFS_IOC_VDEV_ATTACH, /* 0x5a0e */
ZFS_IOC_VDEV_DETACH, /* 0x5a0f */
ZFS_IOC_VDEV_SETPATH, /* 0x5a10 */
ZFS_IOC_VDEV_SETFRU, /* 0x5a11 */
ZFS_IOC_OBJSET_STATS, /* 0x5a12 */
ZFS_IOC_OBJSET_ZPLPROPS, /* 0x5a13 */
ZFS_IOC_DATASET_LIST_NEXT, /* 0x5a14 */
ZFS_IOC_SNAPSHOT_LIST_NEXT, /* 0x5a15 */
ZFS_IOC_SET_PROP, /* 0x5a16 */
ZFS_IOC_CREATE, /* 0x5a17 */
ZFS_IOC_DESTROY, /* 0x5a18 */
ZFS_IOC_ROLLBACK, /* 0x5a19 */
ZFS_IOC_RENAME, /* 0x5a1a */
ZFS_IOC_RECV, /* 0x5a1b */
ZFS_IOC_SEND, /* 0x5a1c */
ZFS_IOC_INJECT_FAULT, /* 0x5a1d */
ZFS_IOC_CLEAR_FAULT, /* 0x5a1e */
ZFS_IOC_INJECT_LIST_NEXT, /* 0x5a1f */
ZFS_IOC_ERROR_LOG, /* 0x5a20 */
ZFS_IOC_CLEAR, /* 0x5a21 */
ZFS_IOC_PROMOTE, /* 0x5a22 */
ZFS_IOC_SNAPSHOT, /* 0x5a23 */
ZFS_IOC_DSOBJ_TO_DSNAME, /* 0x5a24 */
ZFS_IOC_OBJ_TO_PATH, /* 0x5a25 */
ZFS_IOC_POOL_SET_PROPS, /* 0x5a26 */
ZFS_IOC_POOL_GET_PROPS, /* 0x5a27 */
ZFS_IOC_SET_FSACL, /* 0x5a28 */
ZFS_IOC_GET_FSACL, /* 0x5a29 */
ZFS_IOC_SHARE, /* 0x5a2a */
ZFS_IOC_INHERIT_PROP, /* 0x5a2b */
ZFS_IOC_SMB_ACL, /* 0x5a2c */
ZFS_IOC_USERSPACE_ONE, /* 0x5a2d */
ZFS_IOC_USERSPACE_MANY, /* 0x5a2e */
ZFS_IOC_USERSPACE_UPGRADE, /* 0x5a2f */
ZFS_IOC_HOLD, /* 0x5a30 */
ZFS_IOC_RELEASE, /* 0x5a31 */
ZFS_IOC_GET_HOLDS, /* 0x5a32 */
ZFS_IOC_OBJSET_RECVD_PROPS, /* 0x5a33 */
ZFS_IOC_VDEV_SPLIT, /* 0x5a34 */
ZFS_IOC_NEXT_OBJ, /* 0x5a35 */
ZFS_IOC_DIFF, /* 0x5a36 */
ZFS_IOC_TMP_SNAPSHOT, /* 0x5a37 */
ZFS_IOC_OBJ_TO_STATS, /* 0x5a38 */
ZFS_IOC_SPACE_WRITTEN, /* 0x5a39 */
ZFS_IOC_SPACE_SNAPS, /* 0x5a3a */
ZFS_IOC_DESTROY_SNAPS, /* 0x5a3b */
ZFS_IOC_POOL_REGUID, /* 0x5a3c */
ZFS_IOC_POOL_REOPEN, /* 0x5a3d */
ZFS_IOC_SEND_PROGRESS, /* 0x5a3e */
ZFS_IOC_LOG_HISTORY, /* 0x5a3f */
ZFS_IOC_SEND_NEW, /* 0x5a40 */
ZFS_IOC_SEND_SPACE, /* 0x5a41 */
ZFS_IOC_CLONE, /* 0x5a42 */
ZFS_IOC_BOOKMARK, /* 0x5a43 */
ZFS_IOC_GET_BOOKMARKS, /* 0x5a44 */
ZFS_IOC_DESTROY_BOOKMARKS, /* 0x5a45 */
ZFS_IOC_RECV_NEW, /* 0x5a46 */
ZFS_IOC_POOL_SYNC, /* 0x5a47 */
ZFS_IOC_CHANNEL_PROGRAM, /* 0x5a48 */
ZFS_IOC_LOAD_KEY, /* 0x5a49 */
ZFS_IOC_UNLOAD_KEY, /* 0x5a4a */
ZFS_IOC_CHANGE_KEY, /* 0x5a4b */
ZFS_IOC_REMAP, /* 0x5a4c */
ZFS_IOC_POOL_CHECKPOINT, /* 0x5a4d */
ZFS_IOC_POOL_DISCARD_CHECKPOINT, /* 0x5a4e */
ZFS_IOC_POOL_INITIALIZE, /* 0x5a4f */
ZFS_IOC_POOL_TRIM, /* 0x5a50 */
ZFS_IOC_REDACT, /* 0x5a51 */
ZFS_IOC_GET_BOOKMARK_PROPS, /* 0x5a52 */
ZFS_IOC_WAIT, /* 0x5a53 */
ZFS_IOC_WAIT_FS, /* 0x5a54 */
ZFS_IOC_VDEV_GET_PROPS, /* 0x5a55 */
ZFS_IOC_VDEV_SET_PROPS, /* 0x5a56 */
ZFS_IOC_POOL_SCRUB, /* 0x5a57 */
/*
* Per-platform (Optional) - 8/128 numbers reserved.
*/
ZFS_IOC_PLATFORM = ZFS_IOC_FIRST + 0x80,
ZFS_IOC_EVENTS_NEXT, /* 0x81 (Linux) */
ZFS_IOC_EVENTS_CLEAR, /* 0x82 (Linux) */
ZFS_IOC_EVENTS_SEEK, /* 0x83 (Linux) */
ZFS_IOC_NEXTBOOT, /* 0x84 (FreeBSD) */
ZFS_IOC_JAIL, /* 0x85 (FreeBSD) */
ZFS_IOC_USERNS_ATTACH = ZFS_IOC_JAIL, /* 0x85 (Linux) */
ZFS_IOC_UNJAIL, /* 0x86 (FreeBSD) */
ZFS_IOC_USERNS_DETACH = ZFS_IOC_UNJAIL, /* 0x86 (Linux) */
ZFS_IOC_SET_BOOTENV, /* 0x87 */
ZFS_IOC_GET_BOOTENV, /* 0x88 */
ZFS_IOC_LAST
} zfs_ioc_t;
/*
* zvol ioctl to get dataset name
*/
#define BLKZNAME _IOR(0x12, 125, char[ZFS_MAX_DATASET_NAME_LEN])
#ifdef __linux__
/*
* IOCTLs to update and retrieve additional file level attributes on
* Linux.
*/
#define ZFS_IOC_GETDOSFLAGS _IOR(0x83, 1, uint64_t)
#define ZFS_IOC_SETDOSFLAGS _IOW(0x83, 2, uint64_t)
/*
* Additional file level attributes, that are stored
* in the upper half of z_pflags
*/
#define ZFS_READONLY 0x0000000100000000ull
#define ZFS_HIDDEN 0x0000000200000000ull
#define ZFS_SYSTEM 0x0000000400000000ull
#define ZFS_ARCHIVE 0x0000000800000000ull
#define ZFS_IMMUTABLE 0x0000001000000000ull
#define ZFS_NOUNLINK 0x0000002000000000ull
#define ZFS_APPENDONLY 0x0000004000000000ull
#define ZFS_NODUMP 0x0000008000000000ull
#define ZFS_OPAQUE 0x0000010000000000ull
#define ZFS_AV_QUARANTINED 0x0000020000000000ull
#define ZFS_AV_MODIFIED 0x0000040000000000ull
#define ZFS_REPARSE 0x0000080000000000ull
#define ZFS_OFFLINE 0x0000100000000000ull
#define ZFS_SPARSE 0x0000200000000000ull
#define ZFS_DOS_FL_USER_VISIBLE (ZFS_IMMUTABLE | ZFS_APPENDONLY | \
ZFS_NOUNLINK | ZFS_ARCHIVE | ZFS_NODUMP | ZFS_SYSTEM | \
ZFS_HIDDEN | ZFS_READONLY | ZFS_REPARSE | ZFS_OFFLINE | \
ZFS_SPARSE)
#endif
/*
* ZFS-specific error codes used for returning descriptive errors
* to the userland through zfs ioctls.
*
* The enum implicitly includes all the error codes from errno.h.
* New code should use and extend this enum for errors that are
* not described precisely by generic errno codes.
*
* These numbers should not change over time. New entries should be appended.
*
* (Keep in sync with contrib/pyzfs/libzfs_core/_constants.py)
*/
typedef enum {
ZFS_ERR_CHECKPOINT_EXISTS = 1024,
ZFS_ERR_DISCARDING_CHECKPOINT,
ZFS_ERR_NO_CHECKPOINT,
ZFS_ERR_DEVRM_IN_PROGRESS,
ZFS_ERR_VDEV_TOO_BIG,
ZFS_ERR_IOC_CMD_UNAVAIL,
ZFS_ERR_IOC_ARG_UNAVAIL,
ZFS_ERR_IOC_ARG_REQUIRED,
ZFS_ERR_IOC_ARG_BADTYPE,
ZFS_ERR_WRONG_PARENT,
ZFS_ERR_FROM_IVSET_GUID_MISSING,
ZFS_ERR_FROM_IVSET_GUID_MISMATCH,
ZFS_ERR_SPILL_BLOCK_FLAG_MISSING,
ZFS_ERR_UNKNOWN_SEND_STREAM_FEATURE,
ZFS_ERR_EXPORT_IN_PROGRESS,
ZFS_ERR_BOOKMARK_SOURCE_NOT_ANCESTOR,
ZFS_ERR_STREAM_TRUNCATED,
ZFS_ERR_STREAM_LARGE_BLOCK_MISMATCH,
ZFS_ERR_RESILVER_IN_PROGRESS,
ZFS_ERR_REBUILD_IN_PROGRESS,
ZFS_ERR_BADPROP,
ZFS_ERR_VDEV_NOTSUP,
ZFS_ERR_NOT_USER_NAMESPACE,
ZFS_ERR_RESUME_EXISTS,
ZFS_ERR_CRYPTO_NOTSUP,
+ ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS,
} zfs_errno_t;
/*
* Internal SPA load state. Used by FMA diagnosis engine.
*/
typedef enum {
SPA_LOAD_NONE, /* no load in progress */
SPA_LOAD_OPEN, /* normal open */
SPA_LOAD_IMPORT, /* import in progress */
SPA_LOAD_TRYIMPORT, /* tryimport in progress */
SPA_LOAD_RECOVER, /* recovery requested */
SPA_LOAD_ERROR, /* load failed */
SPA_LOAD_CREATE /* creation in progress */
} spa_load_state_t;
typedef enum {
ZPOOL_WAIT_CKPT_DISCARD,
ZPOOL_WAIT_FREE,
ZPOOL_WAIT_INITIALIZE,
ZPOOL_WAIT_REPLACE,
ZPOOL_WAIT_REMOVE,
ZPOOL_WAIT_RESILVER,
ZPOOL_WAIT_SCRUB,
ZPOOL_WAIT_TRIM,
+ ZPOOL_WAIT_RAIDZ_EXPAND,
ZPOOL_WAIT_NUM_ACTIVITIES
} zpool_wait_activity_t;
typedef enum {
ZFS_WAIT_DELETEQ,
ZFS_WAIT_NUM_ACTIVITIES
} zfs_wait_activity_t;
/*
* Bookmark name values.
*/
#define ZPOOL_ERR_LIST "error list"
#define ZPOOL_ERR_DATASET "dataset"
#define ZPOOL_ERR_OBJECT "object"
#define HIS_MAX_RECORD_LEN (MAXPATHLEN + MAXPATHLEN + 1)
/*
* The following are names used in the nvlist describing
* the pool's history log.
*/
#define ZPOOL_HIST_RECORD "history record"
#define ZPOOL_HIST_TIME "history time"
#define ZPOOL_HIST_CMD "history command"
#define ZPOOL_HIST_WHO "history who"
#define ZPOOL_HIST_ZONE "history zone"
#define ZPOOL_HIST_HOST "history hostname"
#define ZPOOL_HIST_TXG "history txg"
#define ZPOOL_HIST_INT_EVENT "history internal event"
#define ZPOOL_HIST_INT_STR "history internal str"
#define ZPOOL_HIST_INT_NAME "internal_name"
#define ZPOOL_HIST_IOCTL "ioctl"
#define ZPOOL_HIST_INPUT_NVL "in_nvl"
#define ZPOOL_HIST_OUTPUT_NVL "out_nvl"
#define ZPOOL_HIST_OUTPUT_SIZE "out_size"
#define ZPOOL_HIST_DSNAME "dsname"
#define ZPOOL_HIST_DSID "dsid"
#define ZPOOL_HIST_ERRNO "errno"
#define ZPOOL_HIST_ELAPSED_NS "elapsed_ns"
/*
* Special nvlist name that will not have its args recorded in the pool's
* history log.
*/
#define ZPOOL_HIDDEN_ARGS "hidden_args"
/*
* The following are names used when invoking ZFS_IOC_POOL_INITIALIZE.
*/
#define ZPOOL_INITIALIZE_COMMAND "initialize_command"
#define ZPOOL_INITIALIZE_VDEVS "initialize_vdevs"
/*
* The following are names used when invoking ZFS_IOC_POOL_TRIM.
*/
#define ZPOOL_TRIM_COMMAND "trim_command"
#define ZPOOL_TRIM_VDEVS "trim_vdevs"
#define ZPOOL_TRIM_RATE "trim_rate"
#define ZPOOL_TRIM_SECURE "trim_secure"
/*
* The following are names used when invoking ZFS_IOC_POOL_WAIT.
*/
#define ZPOOL_WAIT_ACTIVITY "wait_activity"
#define ZPOOL_WAIT_TAG "wait_tag"
#define ZPOOL_WAIT_WAITED "wait_waited"
/*
* The following are names used when invoking ZFS_IOC_VDEV_GET_PROP.
*/
#define ZPOOL_VDEV_PROPS_GET_VDEV "vdevprops_get_vdev"
#define ZPOOL_VDEV_PROPS_GET_PROPS "vdevprops_get_props"
/*
* The following are names used when invoking ZFS_IOC_VDEV_SET_PROP.
*/
#define ZPOOL_VDEV_PROPS_SET_VDEV "vdevprops_set_vdev"
#define ZPOOL_VDEV_PROPS_SET_PROPS "vdevprops_set_props"
/*
* The following are names used when invoking ZFS_IOC_WAIT_FS.
*/
#define ZFS_WAIT_ACTIVITY "wait_activity"
#define ZFS_WAIT_WAITED "wait_waited"
/*
* Flags for ZFS_IOC_VDEV_SET_STATE
*/
#define ZFS_ONLINE_CHECKREMOVE 0x1
#define ZFS_ONLINE_UNSPARE 0x2
#define ZFS_ONLINE_FORCEFAULT 0x4
#define ZFS_ONLINE_EXPAND 0x8
#define ZFS_ONLINE_SPARE 0x10
#define ZFS_OFFLINE_TEMPORARY 0x1
/*
* Flags for ZFS_IOC_POOL_IMPORT
*/
#define ZFS_IMPORT_NORMAL 0x0
#define ZFS_IMPORT_VERBATIM 0x1
#define ZFS_IMPORT_ANY_HOST 0x2
#define ZFS_IMPORT_MISSING_LOG 0x4
#define ZFS_IMPORT_ONLY 0x8
#define ZFS_IMPORT_TEMP_NAME 0x10
#define ZFS_IMPORT_SKIP_MMP 0x20
#define ZFS_IMPORT_LOAD_KEYS 0x40
#define ZFS_IMPORT_CHECKPOINT 0x80
/*
* Channel program argument/return nvlist keys and defaults.
*/
#define ZCP_ARG_PROGRAM "program"
#define ZCP_ARG_ARGLIST "arg"
#define ZCP_ARG_SYNC "sync"
#define ZCP_ARG_INSTRLIMIT "instrlimit"
#define ZCP_ARG_MEMLIMIT "memlimit"
#define ZCP_ARG_CLIARGV "argv"
#define ZCP_RET_ERROR "error"
#define ZCP_RET_RETURN "return"
#define ZCP_DEFAULT_INSTRLIMIT (10 * 1000 * 1000)
#define ZCP_MAX_INSTRLIMIT (10 * ZCP_DEFAULT_INSTRLIMIT)
#define ZCP_DEFAULT_MEMLIMIT (10 * 1024 * 1024)
#define ZCP_MAX_MEMLIMIT (10 * ZCP_DEFAULT_MEMLIMIT)
/*
* Sysevent payload members. ZFS will generate the following sysevents with the
* given payloads:
*
* ESC_ZFS_RESILVER_START
* ESC_ZFS_RESILVER_FINISH
*
* ZFS_EV_POOL_NAME DATA_TYPE_STRING
* ZFS_EV_POOL_GUID DATA_TYPE_UINT64
* ZFS_EV_RESILVER_TYPE DATA_TYPE_STRING
*
* ESC_ZFS_POOL_DESTROY
* ESC_ZFS_POOL_REGUID
*
* ZFS_EV_POOL_NAME DATA_TYPE_STRING
* ZFS_EV_POOL_GUID DATA_TYPE_UINT64
*
* ESC_ZFS_VDEV_REMOVE
* ESC_ZFS_VDEV_CLEAR
* ESC_ZFS_VDEV_CHECK
*
* ZFS_EV_POOL_NAME DATA_TYPE_STRING
* ZFS_EV_POOL_GUID DATA_TYPE_UINT64
* ZFS_EV_VDEV_PATH DATA_TYPE_STRING (optional)
* ZFS_EV_VDEV_GUID DATA_TYPE_UINT64
*
* ESC_ZFS_HISTORY_EVENT
*
* ZFS_EV_POOL_NAME DATA_TYPE_STRING
* ZFS_EV_POOL_GUID DATA_TYPE_UINT64
* ZFS_EV_HIST_TIME DATA_TYPE_UINT64 (optional)
* ZFS_EV_HIST_CMD DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_WHO DATA_TYPE_UINT64 (optional)
* ZFS_EV_HIST_ZONE DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_HOST DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_TXG DATA_TYPE_UINT64 (optional)
* ZFS_EV_HIST_INT_EVENT DATA_TYPE_UINT64 (optional)
* ZFS_EV_HIST_INT_STR DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_INT_NAME DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_IOCTL DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_DSNAME DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_DSID DATA_TYPE_UINT64 (optional)
*
* The ZFS_EV_HIST_* members will correspond to the ZPOOL_HIST_* members in the
* history log nvlist. The keynames will be free of any spaces or other
* characters that could be potentially unexpected to consumers of the
* sysevents.
*/
#define ZFS_EV_POOL_NAME "pool_name"
#define ZFS_EV_POOL_GUID "pool_guid"
#define ZFS_EV_VDEV_PATH "vdev_path"
#define ZFS_EV_VDEV_GUID "vdev_guid"
#define ZFS_EV_HIST_TIME "history_time"
#define ZFS_EV_HIST_CMD "history_command"
#define ZFS_EV_HIST_WHO "history_who"
#define ZFS_EV_HIST_ZONE "history_zone"
#define ZFS_EV_HIST_HOST "history_hostname"
#define ZFS_EV_HIST_TXG "history_txg"
#define ZFS_EV_HIST_INT_EVENT "history_internal_event"
#define ZFS_EV_HIST_INT_STR "history_internal_str"
#define ZFS_EV_HIST_INT_NAME "history_internal_name"
#define ZFS_EV_HIST_IOCTL "history_ioctl"
#define ZFS_EV_HIST_DSNAME "history_dsname"
#define ZFS_EV_HIST_DSID "history_dsid"
#define ZFS_EV_RESILVER_TYPE "resilver_type"
/*
* We currently support block sizes from 512 bytes to 16MB.
* The benefits of larger blocks, and thus larger IO, need to be weighed
* against the cost of COWing a giant block to modify one byte, and the
* large latency of reading or writing a large block.
*
* The recordsize property can not be set larger than zfs_max_recordsize
* (default 16MB on 64-bit and 1MB on 32-bit). See the comment near
* zfs_max_recordsize in dsl_dataset.c for details.
*
* Note that although the LSIZE field of the blkptr_t can store sizes up
* to 32MB, the dnode's dn_datablkszsec can only store sizes up to
* 32MB - 512 bytes. Therefore, we limit SPA_MAXBLOCKSIZE to 16MB.
*/
#define SPA_MINBLOCKSHIFT 9
#define SPA_OLD_MAXBLOCKSHIFT 17
#define SPA_MAXBLOCKSHIFT 24
#define SPA_MINBLOCKSIZE (1ULL << SPA_MINBLOCKSHIFT)
#define SPA_OLD_MAXBLOCKSIZE (1ULL << SPA_OLD_MAXBLOCKSHIFT)
#define SPA_MAXBLOCKSIZE (1ULL << SPA_MAXBLOCKSHIFT)
/* supported encryption algorithms */
enum zio_encrypt {
ZIO_CRYPT_INHERIT = 0,
ZIO_CRYPT_ON,
ZIO_CRYPT_OFF,
ZIO_CRYPT_AES_128_CCM,
ZIO_CRYPT_AES_192_CCM,
ZIO_CRYPT_AES_256_CCM,
ZIO_CRYPT_AES_128_GCM,
ZIO_CRYPT_AES_192_GCM,
ZIO_CRYPT_AES_256_GCM,
ZIO_CRYPT_FUNCTIONS
};
#define ZIO_CRYPT_ON_VALUE ZIO_CRYPT_AES_256_GCM
#define ZIO_CRYPT_DEFAULT ZIO_CRYPT_OFF
/*
* xattr namespace prefixes. These are forbidden in xattr names.
*
* For cross-platform compatibility, xattrs in the user namespace should not be
* prefixed with the namespace name, but for backwards compatibility with older
* ZFS on Linux versions we do prefix the namespace.
*/
#define ZFS_XA_NS_FREEBSD_PREFIX "freebsd:"
#define ZFS_XA_NS_FREEBSD_PREFIX_LEN strlen("freebsd:")
#define ZFS_XA_NS_LINUX_SECURITY_PREFIX "security."
#define ZFS_XA_NS_LINUX_SECURITY_PREFIX_LEN strlen("security.")
#define ZFS_XA_NS_LINUX_SYSTEM_PREFIX "system."
#define ZFS_XA_NS_LINUX_SYSTEM_PREFIX_LEN strlen("system.")
#define ZFS_XA_NS_LINUX_TRUSTED_PREFIX "trusted."
#define ZFS_XA_NS_LINUX_TRUSTED_PREFIX_LEN strlen("trusted.")
#define ZFS_XA_NS_LINUX_USER_PREFIX "user."
#define ZFS_XA_NS_LINUX_USER_PREFIX_LEN strlen("user.")
#define ZFS_XA_NS_PREFIX_MATCH(ns, name) \
(strncmp(name, ZFS_XA_NS_##ns##_PREFIX, \
ZFS_XA_NS_##ns##_PREFIX_LEN) == 0)
#define ZFS_XA_NS_PREFIX_FORBIDDEN(name) \
(ZFS_XA_NS_PREFIX_MATCH(FREEBSD, name) || \
ZFS_XA_NS_PREFIX_MATCH(LINUX_SECURITY, name) || \
ZFS_XA_NS_PREFIX_MATCH(LINUX_SYSTEM, name) || \
ZFS_XA_NS_PREFIX_MATCH(LINUX_TRUSTED, name) || \
ZFS_XA_NS_PREFIX_MATCH(LINUX_USER, name))
#ifdef __cplusplus
}
#endif
#endif /* _SYS_FS_ZFS_H */
diff --git a/sys/contrib/openzfs/include/sys/spa_impl.h b/sys/contrib/openzfs/include/sys/spa_impl.h
index b1eb06f94fcc..ee91816ac48e 100644
--- a/sys/contrib/openzfs/include/sys/spa_impl.h
+++ b/sys/contrib/openzfs/include/sys/spa_impl.h
@@ -1,488 +1,492 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2019 by Delphix. All rights reserved.
* Copyright 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) 2016 Actifio, Inc. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019 Datto Inc.
*/
#ifndef _SYS_SPA_IMPL_H
#define _SYS_SPA_IMPL_H
#include <sys/spa.h>
#include <sys/spa_checkpoint.h>
#include <sys/spa_log_spacemap.h>
#include <sys/vdev.h>
#include <sys/vdev_rebuild.h>
#include <sys/vdev_removal.h>
+#include <sys/vdev_raidz.h>
#include <sys/metaslab.h>
#include <sys/dmu.h>
#include <sys/dsl_pool.h>
#include <sys/uberblock_impl.h>
#include <sys/zfs_context.h>
#include <sys/avl.h>
#include <sys/zfs_refcount.h>
#include <sys/bplist.h>
#include <sys/bpobj.h>
#include <sys/dsl_crypt.h>
#include <sys/zfeature.h>
#include <sys/zthr.h>
#include <sys/dsl_deadlist.h>
#include <zfeature_common.h>
#ifdef __cplusplus
extern "C" {
#endif
typedef struct spa_alloc {
kmutex_t spaa_lock;
avl_tree_t spaa_tree;
} ____cacheline_aligned spa_alloc_t;
typedef struct spa_error_entry {
zbookmark_phys_t se_bookmark;
char *se_name;
avl_node_t se_avl;
zbookmark_err_phys_t se_zep; /* not accounted in avl_find */
} 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;
int scl_count;
kcondvar_t scl_cv;
} ____cacheline_aligned 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;
/* one for each thread in the spa sync taskq */
typedef struct spa_syncthread_info {
kthread_t *sti_thread;
taskq_t *sti_wr_iss_tq; /* assigned wr_iss taskq */
} spa_syncthread_info_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 */
uint32_t 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? */
boolean_t spa_is_splitting; /* in the middle of a split? */
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 */
boolean_t spa_is_exporting; /* true while exporting pool */
metaslab_class_t *spa_normal_class; /* normal data class */
metaslab_class_t *spa_log_class; /* intent log data class */
metaslab_class_t *spa_embedded_log_class; /* log on normal vdevs */
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 */
inode_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 */
uint64_t spa_min_ashift; /* of vdevs in normal class */
uint64_t spa_max_ashift; /* of vdevs in normal class */
uint64_t spa_min_alloc; /* of vdevs in normal class */
uint64_t spa_gcd_alloc; /* 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_allocs is an array, whose lengths is stored in spa_alloc_count.
* There is one tree and one lock for each allocator, to help improve
* allocation performance in write-heavy workloads.
*/
spa_alloc_t *spa_allocs;
int spa_alloc_count;
int spa_active_allocator; /* selectable allocator */
/* per-allocator sync thread taskqs */
taskq_t *spa_sync_tq;
spa_syncthread_info_t *spa_syncthreads;
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 */
/* in-flight verification bytes */
uint64_t spa_load_verify_bytes;
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 */
/* error scrub pause time in milliseconds */
uint64_t spa_scan_pass_errorscrub_pause;
/* total error scrub paused time in milliseconds */
uint64_t spa_scan_pass_errorscrub_spent_paused;
/*
* 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? */
uint64_t spa_nonallocating_dspace;
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. */
+ vdev_raidz_expand_t *spa_raidz_expand;
+ zthr_t *spa_raidz_expand_zthr;
+
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;
space_map_t *spa_syncing_log_sm; /* current log space map */
avl_tree_t spa_sm_logs_by_txg;
kmutex_t spa_flushed_ms_lock; /* for metaslabs_by_flushed */
avl_tree_t spa_metaslabs_by_flushed;
spa_unflushed_stats_t spa_unflushed_stats;
list_t spa_log_summary;
uint64_t spa_log_flushall_txg;
zthr_t *spa_livelist_delete_zthr; /* deleting livelists */
zthr_t *spa_livelist_condense_zthr; /* condensing livelists */
uint64_t spa_livelists_to_delete; /* set of livelists to free */
livelist_condense_entry_t spa_to_condense; /* next to condense */
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 */
avl_tree_t spa_errlist_healed; /* list of healed blocks */
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_deadman_failmode; /* failure mode for deadman */
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 */
spa_mode_t spa_mode; /* SPA_MODE_{READ|WRITE} */
boolean_t spa_read_spacemaps; /* spacemaps available if ro */
spa_log_state_t spa_log_state; /* log state */
uint64_t spa_autoexpand; /* lun expansion on/off */
ddt_t *spa_ddt[ZIO_CHECKSUM_FUNCTIONS]; /* in-core DDTs */
uint64_t spa_ddt_stat_object; /* DDT statistics */
uint64_t spa_dedup_dspace; /* Cache get_dedup_dspace() */
uint64_t spa_dedup_checksum; /* default dedup checksum */
uint64_t spa_dspace; /* dspace in normal class */
struct brt *spa_brt; /* in-core BRT */
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 */
proc_t *spa_proc; /* "zpool-poolname" process */
uintptr_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 */
kmutex_t spa_feat_stats_lock; /* protects spa_feat_stats */
nvlist_t *spa_feat_stats; /* Cache of enabled features */
/* cache feature refcounts */
uint64_t spa_feat_refcount_cache[SPA_FEATURES];
taskqid_t spa_deadman_tqid; /* Task id */
uint64_t spa_deadman_calls; /* number of deadman calls */
hrtime_t spa_sync_starttime; /* starting time of spa_sync */
uint64_t spa_deadman_synctime; /* deadman sync expiration */
uint64_t spa_deadman_ziotime; /* deadman zio expiration */
uint64_t spa_all_vdev_zaps; /* ZAP of per-vd ZAP obj #s */
spa_avz_action_t spa_avz_action; /* destroy/rebuild AVZ? */
uint64_t spa_autotrim; /* automatic background trim? */
uint64_t spa_errata; /* errata issues detected */
spa_stats_t spa_stats; /* assorted spa statistics */
spa_keystore_t spa_keystore; /* loaded crypto keys */
/* 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 */
taskq_t *spa_zvol_taskq; /* Taskq for minor management */
taskq_t *spa_metaslab_taskq; /* Taskq for metaslab preload */
taskq_t *spa_prefetch_taskq; /* Taskq for prefetch threads */
taskq_t *spa_upgrade_taskq; /* Taskq for upgrade jobs */
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 */
uint32_t spa_hostid; /* cached system hostid */
/* synchronization for threads in spa_wait */
kmutex_t spa_activities_lock;
kcondvar_t spa_activities_cv;
kcondvar_t spa_waiters_cv;
int spa_waiters; /* number of waiting threads */
boolean_t spa_waiters_cancel; /* waiters should return */
char *spa_compatibility; /* compatibility file(s) */
/*
* spa_refcount & spa_config_lock must be the last elements
* 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 char *spa_config_path;
extern const char *zfs_deadman_failmode;
extern uint_t spa_slop_shift;
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, zio_t *zio);
extern void spa_taskq_dispatch_sync(spa_t *, zio_type_t t, zio_taskq_type_t q,
task_func_t *func, void *arg, uint_t flags);
extern void spa_load_spares(spa_t *spa);
extern void spa_load_l2cache(spa_t *spa);
extern sysevent_t *spa_event_create(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl,
const char *name);
extern void spa_event_post(sysevent_t *ev);
extern int param_set_deadman_failmode_common(const char *val);
extern void spa_set_deadman_synctime(hrtime_t ns);
extern void spa_set_deadman_ziotime(hrtime_t ns);
extern const char *spa_history_zone(void);
extern const char *zfs_active_allocator;
extern int param_set_active_allocator_common(const char *val);
#ifdef __cplusplus
}
#endif
#endif /* _SYS_SPA_IMPL_H */
diff --git a/sys/contrib/openzfs/include/sys/uberblock_impl.h b/sys/contrib/openzfs/include/sys/uberblock_impl.h
index 03bcfa8f4dd1..d3a71cc8f84b 100644
--- a/sys/contrib/openzfs/include/sys/uberblock_impl.h
+++ b/sys/contrib/openzfs/include/sys/uberblock_impl.h
@@ -1,145 +1,180 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2016, 2017 by Delphix. All rights reserved.
*/
#ifndef _SYS_UBERBLOCK_IMPL_H
#define _SYS_UBERBLOCK_IMPL_H
#include <sys/uberblock.h>
#ifdef __cplusplus
extern "C" {
#endif
/*
* The uberblock version is incremented whenever an incompatible on-disk
* format change is made to the SPA, DMU, or ZAP.
*
* Note: the first two fields should never be moved. When a storage pool
* is opened, the uberblock must be read off the disk before the version
* can be checked. If the ub_version field is moved, we may not detect
* version mismatch. If the ub_magic field is moved, applications that
* expect the magic number in the first word won't work.
*/
#define UBERBLOCK_MAGIC 0x00bab10c /* oo-ba-bloc! */
#define UBERBLOCK_SHIFT 10 /* up to 1K */
#define MMP_MAGIC 0xa11cea11 /* all-see-all */
#define MMP_INTERVAL_VALID_BIT 0x01
#define MMP_SEQ_VALID_BIT 0x02
#define MMP_FAIL_INT_VALID_BIT 0x04
#define MMP_VALID(ubp) (ubp->ub_magic == UBERBLOCK_MAGIC && \
ubp->ub_mmp_magic == MMP_MAGIC)
#define MMP_INTERVAL_VALID(ubp) (MMP_VALID(ubp) && (ubp->ub_mmp_config & \
MMP_INTERVAL_VALID_BIT))
#define MMP_SEQ_VALID(ubp) (MMP_VALID(ubp) && (ubp->ub_mmp_config & \
MMP_SEQ_VALID_BIT))
#define MMP_FAIL_INT_VALID(ubp) (MMP_VALID(ubp) && (ubp->ub_mmp_config & \
MMP_FAIL_INT_VALID_BIT))
#define MMP_INTERVAL(ubp) ((ubp->ub_mmp_config & 0x00000000FFFFFF00) \
>> 8)
#define MMP_SEQ(ubp) ((ubp->ub_mmp_config & 0x0000FFFF00000000) \
>> 32)
#define MMP_FAIL_INT(ubp) ((ubp->ub_mmp_config & 0xFFFF000000000000) \
>> 48)
#define MMP_INTERVAL_SET(write) \
(((uint64_t)(write & 0xFFFFFF) << 8) | MMP_INTERVAL_VALID_BIT)
#define MMP_SEQ_SET(seq) \
(((uint64_t)(seq & 0xFFFF) << 32) | MMP_SEQ_VALID_BIT)
#define MMP_FAIL_INT_SET(fail) \
(((uint64_t)(fail & 0xFFFF) << 48) | MMP_FAIL_INT_VALID_BIT)
+/*
+ * RAIDZ expansion reflow information.
+ *
+ * 64 56 48 40 32 24 16 8 0
+ * +-------+-------+-------+-------+-------+-------+-------+-------+
+ * |Scratch | Reflow |
+ * | State | Offset |
+ * +-------+-------+-------+-------+-------+-------+-------+-------+
+ */
+typedef enum raidz_reflow_scratch_state {
+ RRSS_SCRATCH_NOT_IN_USE = 0,
+ RRSS_SCRATCH_VALID,
+ RRSS_SCRATCH_INVALID_SYNCED,
+ RRSS_SCRATCH_INVALID_SYNCED_ON_IMPORT,
+ RRSS_SCRATCH_INVALID_SYNCED_REFLOW
+} raidz_reflow_scratch_state_t;
+
+#define RRSS_GET_OFFSET(ub) \
+ BF64_GET_SB((ub)->ub_raidz_reflow_info, 0, 55, SPA_MINBLOCKSHIFT, 0)
+#define RRSS_SET_OFFSET(ub, x) \
+ BF64_SET_SB((ub)->ub_raidz_reflow_info, 0, 55, SPA_MINBLOCKSHIFT, 0, x)
+
+#define RRSS_GET_STATE(ub) \
+ BF64_GET((ub)->ub_raidz_reflow_info, 55, 9)
+#define RRSS_SET_STATE(ub, x) \
+ BF64_SET((ub)->ub_raidz_reflow_info, 55, 9, x)
+
+#define RAIDZ_REFLOW_SET(ub, state, offset) do { \
+ (ub)->ub_raidz_reflow_info = 0; \
+ RRSS_SET_OFFSET(ub, offset); \
+ RRSS_SET_STATE(ub, state); \
+} while (0)
+
struct uberblock {
uint64_t ub_magic; /* UBERBLOCK_MAGIC */
uint64_t ub_version; /* SPA_VERSION */
uint64_t ub_txg; /* txg of last sync */
uint64_t ub_guid_sum; /* sum of all vdev guids */
uint64_t ub_timestamp; /* UTC time of last sync */
blkptr_t ub_rootbp; /* MOS objset_phys_t */
/* highest SPA_VERSION supported by software that wrote this txg */
uint64_t ub_software_version;
/* Maybe missing in uberblocks we read, but always written */
uint64_t ub_mmp_magic; /* MMP_MAGIC */
/*
* If ub_mmp_delay == 0 and ub_mmp_magic is valid, MMP is off.
* Otherwise, nanosec since last MMP write.
*/
uint64_t ub_mmp_delay;
/*
* The ub_mmp_config contains the multihost write interval, multihost
* fail intervals, sequence number for sub-second granularity, and
* valid bit mask. This layout is as follows:
*
* 64 56 48 40 32 24 16 8 0
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 0 | Fail Intervals| Seq | Write Interval (ms) | VALID |
* +-------+-------+-------+-------+-------+-------+-------+-------+
*
* This allows a write_interval of (2^24/1000)s, over 4.5 hours
*
* VALID Bits:
* - 0x01 - Write Interval (ms)
* - 0x02 - Sequence number exists
* - 0x04 - Fail Intervals
* - 0xf8 - Reserved
*/
uint64_t ub_mmp_config;
/*
* ub_checkpoint_txg indicates two things about the current uberblock:
*
* 1] If it is not zero then this uberblock is a checkpoint. If it is
* zero, then this uberblock is not a checkpoint.
*
* 2] On checkpointed uberblocks, the value of ub_checkpoint_txg is
* the ub_txg that the uberblock had at the time we moved it to
* the MOS config.
*
* The field is set when we checkpoint the uberblock and continues to
* hold that value even after we've rewound (unlike the ub_txg that
* is reset to a higher value).
*
* Besides checks used to determine whether we are reopening the
* pool from a checkpointed uberblock [see spa_ld_select_uberblock()],
* the value of the field is used to determine which ZIL blocks have
* been allocated according to the ms_sm when we are rewinding to a
* checkpoint. Specifically, if blk_birth > ub_checkpoint_txg, then
* the ZIL block is not allocated [see uses of spa_min_claim_txg()].
*/
uint64_t ub_checkpoint_txg;
+
+ uint64_t ub_raidz_reflow_info;
};
#ifdef __cplusplus
}
#endif
#endif /* _SYS_UBERBLOCK_IMPL_H */
diff --git a/sys/contrib/openzfs/include/sys/vdev.h b/sys/contrib/openzfs/include/sys/vdev.h
index 03e1f438aaf9..38f62b07dc59 100644
--- a/sys/contrib/openzfs/include/sys/vdev.h
+++ b/sys/contrib/openzfs/include/sys/vdev.h
@@ -1,226 +1,232 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019, Datto Inc. All rights reserved.
*/
#ifndef _SYS_VDEV_H
#define _SYS_VDEV_H
#include <sys/spa.h>
#include <sys/zio.h>
#include <sys/dmu.h>
#include <sys/space_map.h>
#include <sys/metaslab.h>
#include <sys/fs/zfs.h>
#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 int zfs_nocacheflush;
typedef boolean_t vdev_open_children_func_t(vdev_t *vd);
extern void vdev_dbgmsg(vdev_t *vd, const char *fmt, ...)
__attribute__((format(printf, 2, 3)));
extern void vdev_dbgmsg_print_tree(vdev_t *, int);
extern int vdev_open(vdev_t *);
extern void vdev_open_children(vdev_t *);
extern void vdev_open_children_subset(vdev_t *, vdev_open_children_func_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_default_need_resilver(vdev_t *vd, const dva_t *dva,
size_t psize, uint64_t phys_birth);
extern boolean_t vdev_dtl_need_resilver(vdev_t *vd, const dva_t *dva,
size_t psize, uint64_t phys_birth);
extern void vdev_dtl_reassess(vdev_t *vd, uint64_t txg, uint64_t scrub_txg,
boolean_t scrub_done, boolean_t rebuild_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 boolean_t vdev_replace_in_progress(vdev_t *vdev);
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, const char *tag);
typedef void vdev_xlate_func_t(void *arg, range_seg64_t *physical_rs);
extern boolean_t vdev_xlate_is_empty(range_seg64_t *rs);
extern void vdev_xlate(vdev_t *vd, const range_seg64_t *logical_rs,
range_seg64_t *physical_rs, range_seg64_t *remain_rs);
extern void vdev_xlate_walk(vdev_t *vd, const range_seg64_t *logical_rs,
vdev_xlate_func_t *func, void *arg);
extern void vdev_get_stats_ex(vdev_t *vd, vdev_stat_t *vs, vdev_stat_ex_t *vsx);
extern metaslab_group_t *vdev_get_mg(vdev_t *vd, metaslab_class_t *mc);
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_txg(vdev_t *vd, uint64_t psize,
+ uint64_t txg);
extern uint64_t vdev_psize_to_asize(vdev_t *vd, uint64_t psize);
/*
- * Return the amount of space allocated for a gang block header.
+ * Return the amount of space allocated for a gang block header. Note that
+ * since the physical birth txg is not provided, this must be constant for
+ * a given vdev. (e.g. raidz expansion can't change this)
*/
static inline uint64_t
vdev_gang_header_asize(vdev_t *vd)
{
- return (vdev_psize_to_asize(vd, SPA_GANGBLOCKSIZE));
+ return (vdev_psize_to_asize_txg(vd, SPA_GANGBLOCKSIZE, 0));
}
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 int vdev_remove_wanted(spa_t *spa, uint64_t guid);
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_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 uint32_t vdev_queue_length(vdev_t *vd);
extern uint64_t vdev_queue_last_offset(vdev_t *vd);
extern uint64_t vdev_queue_class_length(vdev_t *vq, zio_priority_t p);
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_defer_resilver(vdev_t *vd);
extern boolean_t vdev_clear_resilver_deferred(vdev_t *vd, dmu_tx_t *tx);
typedef enum vdev_config_flag {
VDEV_CONFIG_SPARE = 1 << 0,
VDEV_CONFIG_L2CACHE = 1 << 1,
VDEV_CONFIG_MOS = 1 << 2,
VDEV_CONFIG_MISSING = 1 << 3
} vdev_config_flag_t;
extern void vdev_post_kobj_evt(vdev_t *vd);
extern void vdev_clear_kobj_evt(vdev_t *vd);
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_config_generate_stats(vdev_t *vd, nvlist_t *nv);
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 *priv, int flags);
extern int vdev_label_read_bootenv(vdev_t *, nvlist_t *);
extern int vdev_label_write_bootenv(vdev_t *, nvlist_t *);
+extern int vdev_uberblock_sync_list(vdev_t **, int, struct uberblock *, int);
+extern int vdev_check_boot_reserve(spa_t *, vdev_t *);
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);
extern int vdev_prop_set(vdev_t *vd, nvlist_t *innvl, nvlist_t *outnvl);
extern int vdev_prop_get(vdev_t *vd, nvlist_t *nvprops, nvlist_t *outnvl);
#ifdef __cplusplus
}
#endif
#endif /* _SYS_VDEV_H */
diff --git a/sys/contrib/openzfs/include/sys/vdev_impl.h b/sys/contrib/openzfs/include/sys/vdev_impl.h
index 3f2312c23438..dafab66c70f9 100644
--- a/sys/contrib/openzfs/include/sys/vdev_impl.h
+++ b/sys/contrib/openzfs/include/sys/vdev_impl.h
@@ -1,652 +1,654 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
*/
#ifndef _SYS_VDEV_IMPL_H
#define _SYS_VDEV_IMPL_H
#include <sys/avl.h>
#include <sys/bpobj.h>
#include <sys/dmu.h>
#include <sys/metaslab.h>
#include <sys/nvpair.h>
#include <sys/space_map.h>
#include <sys/vdev.h>
#include <sys/dkio.h>
#include <sys/uberblock_impl.h>
#include <sys/vdev_indirect_mapping.h>
#include <sys/vdev_indirect_births.h>
#include <sys/vdev_rebuild.h>
#include <sys/vdev_removal.h>
#include <sys/zfs_ratelimit.h>
#ifdef __cplusplus
extern "C" {
#endif
/*
* Virtual device descriptors.
*
* All storage pool operations go through the virtual device framework,
* which provides data replication and I/O scheduling.
*/
/*
* Forward declarations that lots of things need.
*/
typedef struct vdev_queue vdev_queue_t;
struct abd;
extern uint_t zfs_vdev_queue_depth_pct;
extern uint_t zfs_vdev_def_queue_depth;
extern uint_t zfs_vdev_async_write_max_active;
/*
* Virtual device operations
*/
typedef int vdev_init_func_t(spa_t *spa, nvlist_t *nv, void **tsd);
typedef void vdev_kobj_post_evt_func_t(vdev_t *vd);
typedef void vdev_fini_func_t(vdev_t *vd);
typedef int vdev_open_func_t(vdev_t *vd, uint64_t *size, uint64_t *max_size,
uint64_t *ashift, uint64_t *pshift);
typedef void vdev_close_func_t(vdev_t *vd);
-typedef uint64_t vdev_asize_func_t(vdev_t *vd, uint64_t psize);
+typedef uint64_t vdev_asize_func_t(vdev_t *vd, uint64_t psize, uint64_t txg);
typedef uint64_t vdev_min_asize_func_t(vdev_t *vd);
typedef uint64_t vdev_min_alloc_func_t(vdev_t *vd);
typedef void vdev_io_start_func_t(zio_t *zio);
typedef void vdev_io_done_func_t(zio_t *zio);
typedef void vdev_state_change_func_t(vdev_t *vd, int, int);
typedef boolean_t vdev_need_resilver_func_t(vdev_t *vd, const dva_t *dva,
size_t psize, uint64_t phys_birth);
typedef void vdev_hold_func_t(vdev_t *vd);
typedef void vdev_rele_func_t(vdev_t *vd);
typedef void vdev_remap_cb_t(uint64_t inner_offset, vdev_t *vd,
uint64_t offset, uint64_t size, void *arg);
typedef void vdev_remap_func_t(vdev_t *vd, uint64_t offset, uint64_t size,
vdev_remap_cb_t callback, void *arg);
/*
* Given a target vdev, translates the logical range "in" to the physical
* range "res"
*/
typedef void vdev_xlation_func_t(vdev_t *cvd, const range_seg64_t *logical,
range_seg64_t *physical, range_seg64_t *remain);
typedef uint64_t vdev_rebuild_asize_func_t(vdev_t *vd, uint64_t start,
uint64_t size, uint64_t max_segment);
typedef void vdev_metaslab_init_func_t(vdev_t *vd, uint64_t *startp,
uint64_t *sizep);
typedef void vdev_config_generate_func_t(vdev_t *vd, nvlist_t *nv);
typedef uint64_t vdev_nparity_func_t(vdev_t *vd);
typedef uint64_t vdev_ndisks_func_t(vdev_t *vd);
typedef const struct vdev_ops {
vdev_init_func_t *vdev_op_init;
vdev_fini_func_t *vdev_op_fini;
vdev_open_func_t *vdev_op_open;
vdev_close_func_t *vdev_op_close;
vdev_asize_func_t *vdev_op_asize;
vdev_min_asize_func_t *vdev_op_min_asize;
vdev_min_alloc_func_t *vdev_op_min_alloc;
vdev_io_start_func_t *vdev_op_io_start;
vdev_io_done_func_t *vdev_op_io_done;
vdev_state_change_func_t *vdev_op_state_change;
vdev_need_resilver_func_t *vdev_op_need_resilver;
vdev_hold_func_t *vdev_op_hold;
vdev_rele_func_t *vdev_op_rele;
vdev_remap_func_t *vdev_op_remap;
vdev_xlation_func_t *vdev_op_xlate;
vdev_rebuild_asize_func_t *vdev_op_rebuild_asize;
vdev_metaslab_init_func_t *vdev_op_metaslab_init;
vdev_config_generate_func_t *vdev_op_config_generate;
vdev_nparity_func_t *vdev_op_nparity;
vdev_ndisks_func_t *vdev_op_ndisks;
vdev_kobj_post_evt_func_t *vdev_op_kobj_evt_post;
char vdev_op_type[16];
boolean_t vdev_op_leaf;
} vdev_ops_t;
/*
* Virtual device properties
*/
typedef union vdev_queue_class {
struct {
ulong_t vqc_list_numnodes;
list_t vqc_list;
};
avl_tree_t vqc_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_read_offset_tree;
avl_tree_t vq_write_offset_tree;
uint64_t vq_last_offset;
zio_priority_t vq_last_prio; /* Last sent I/O priority. */
uint32_t vq_cqueued; /* Classes with queued I/Os. */
uint32_t vq_cactive[ZIO_PRIORITY_NUM_QUEUEABLE];
uint32_t vq_active; /* Number of active I/Os. */
uint32_t vq_ia_active; /* Active interactive I/Os. */
uint32_t vq_nia_credit; /* Non-interactive I/Os credit. */
list_t vq_active_list; /* List of active I/Os. */
hrtime_t vq_io_complete_ts; /* time last i/o completed */
hrtime_t vq_io_delta_ts;
zio_t vq_io_search; /* used as local for stack reduction */
kmutex_t vq_lock;
};
typedef enum vdev_alloc_bias {
VDEV_BIAS_NONE,
VDEV_BIAS_LOG, /* dedicated to ZIL data (SLOG) */
VDEV_BIAS_SPECIAL, /* dedicated to ddt, metadata, and small blks */
VDEV_BIAS_DEDUP /* dedicated to dedup metadata */
} vdev_alloc_bias_t;
/*
* On-disk indirect vdev state.
*
* An indirect vdev is described exclusively in the MOS config of a pool.
* The config for an indirect vdev includes several fields, which are
* accessed in memory by a vdev_indirect_config_t.
*/
typedef struct vdev_indirect_config {
/*
* Object (in MOS) which contains the indirect mapping. This object
* contains an array of vdev_indirect_mapping_entry_phys_t ordered by
* vimep_src. The bonus buffer for this object is a
* vdev_indirect_mapping_phys_t. This object is allocated when a vdev
* removal is initiated.
*
* Note that this object can be empty if none of the data on the vdev
* has been copied yet.
*/
uint64_t vic_mapping_object;
/*
* Object (in MOS) which contains the birth times for the mapping
* entries. This object contains an array of
* vdev_indirect_birth_entry_phys_t sorted by vibe_offset. The bonus
* buffer for this object is a vdev_indirect_birth_phys_t. This object
* is allocated when a vdev removal is initiated.
*
* Note that this object can be empty if none of the vdev has yet been
* copied.
*/
uint64_t vic_births_object;
/*
* This is the vdev ID which was removed previous to this vdev, or
* UINT64_MAX if there are no previously removed vdevs.
*/
uint64_t vic_prev_indirect_vdev;
} vdev_indirect_config_t;
/*
* Virtual device descriptor
*/
struct vdev {
/*
* Common to all vdev types.
*/
uint64_t vdev_id; /* child number in vdev parent */
uint64_t vdev_guid; /* unique ID for this vdev */
uint64_t vdev_guid_sum; /* self guid + all child guids */
uint64_t vdev_orig_guid; /* orig. guid prior to remove */
uint64_t vdev_asize; /* allocatable device capacity */
uint64_t vdev_min_asize; /* min acceptable asize */
uint64_t vdev_max_asize; /* max acceptable asize */
uint64_t vdev_ashift; /* block alignment shift */
/*
* Logical block alignment shift
*
* The smallest sized/aligned I/O supported by the device.
*/
uint64_t vdev_logical_ashift;
/*
* Physical block alignment shift
*
* The device supports logical I/Os with vdev_logical_ashift
* size/alignment, but optimum performance will be achieved by
* aligning/sizing requests to vdev_physical_ashift. Smaller
* requests may be inflated or incur device level read-modify-write
* operations.
*
* May be 0 to indicate no preference (i.e. use vdev_logical_ashift).
*/
uint64_t vdev_physical_ashift;
uint64_t vdev_state; /* see VDEV_STATE_* #defines */
uint64_t vdev_prevstate; /* used when reopening a vdev */
vdev_ops_t *vdev_ops; /* vdev operations */
spa_t *vdev_spa; /* spa for this vdev */
void *vdev_tsd; /* type-specific data */
vdev_t *vdev_top; /* top-level vdev */
vdev_t *vdev_parent; /* parent vdev */
vdev_t **vdev_child; /* array of children */
uint64_t vdev_children; /* number of children */
vdev_stat_t vdev_stat; /* virtual device statistics */
vdev_stat_ex_t vdev_stat_ex; /* extended statistics */
boolean_t vdev_expanding; /* expand the vdev? */
boolean_t vdev_reopening; /* reopen in progress? */
boolean_t vdev_nonrot; /* true if solid state */
int vdev_load_error; /* error on last load */
int vdev_open_error; /* error on last open */
int vdev_validate_error; /* error on last validate */
kthread_t *vdev_open_thread; /* thread opening children */
kthread_t *vdev_validate_thread; /* thread validating children */
uint64_t vdev_crtxg; /* txg when top-level was added */
uint64_t vdev_root_zap;
/*
* Top-level vdev state.
*/
uint64_t vdev_ms_array; /* metaslab array object */
uint64_t vdev_ms_shift; /* metaslab size shift */
uint64_t vdev_ms_count; /* number of metaslabs */
metaslab_group_t *vdev_mg; /* metaslab group */
metaslab_group_t *vdev_log_mg; /* embedded slog metaslab group */
metaslab_t **vdev_ms; /* metaslab array */
txg_list_t vdev_ms_list; /* per-txg dirty metaslab lists */
txg_list_t vdev_dtl_list; /* per-txg dirty DTL lists */
txg_node_t vdev_txg_node; /* per-txg dirty vdev linkage */
boolean_t vdev_remove_wanted; /* async remove wanted? */
boolean_t vdev_probe_wanted; /* async probe wanted? */
list_node_t vdev_config_dirty_node; /* config dirty list */
list_node_t vdev_state_dirty_node; /* state dirty list */
uint64_t vdev_deflate_ratio; /* deflation ratio (x512) */
uint64_t vdev_islog; /* is an intent log device */
uint64_t vdev_noalloc; /* device is passivated? */
uint64_t vdev_removing; /* device is being removed? */
uint64_t vdev_failfast; /* device failfast setting */
+ boolean_t vdev_rz_expanding; /* raidz is being expanded? */
boolean_t vdev_ishole; /* is a hole in the namespace */
uint64_t vdev_top_zap;
vdev_alloc_bias_t vdev_alloc_bias; /* metaslab allocation bias */
/* pool checkpoint related */
space_map_t *vdev_checkpoint_sm; /* contains reserved blocks */
/* Initialize related */
boolean_t vdev_initialize_exit_wanted;
vdev_initializing_state_t vdev_initialize_state;
list_node_t vdev_initialize_node;
kthread_t *vdev_initialize_thread;
/* Protects vdev_initialize_thread and vdev_initialize_state. */
kmutex_t vdev_initialize_lock;
kcondvar_t vdev_initialize_cv;
uint64_t vdev_initialize_offset[TXG_SIZE];
uint64_t vdev_initialize_last_offset;
range_tree_t *vdev_initialize_tree; /* valid while initializing */
uint64_t vdev_initialize_bytes_est;
uint64_t vdev_initialize_bytes_done;
uint64_t vdev_initialize_action_time; /* start and end time */
/* TRIM related */
boolean_t vdev_trim_exit_wanted;
boolean_t vdev_autotrim_exit_wanted;
vdev_trim_state_t vdev_trim_state;
list_node_t vdev_trim_node;
kmutex_t vdev_autotrim_lock;
kcondvar_t vdev_autotrim_cv;
kcondvar_t vdev_autotrim_kick_cv;
kthread_t *vdev_autotrim_thread;
/* Protects vdev_trim_thread and vdev_trim_state. */
kmutex_t vdev_trim_lock;
kcondvar_t vdev_trim_cv;
kthread_t *vdev_trim_thread;
uint64_t vdev_trim_offset[TXG_SIZE];
uint64_t vdev_trim_last_offset;
uint64_t vdev_trim_bytes_est;
uint64_t vdev_trim_bytes_done;
uint64_t vdev_trim_rate; /* requested rate (bytes/sec) */
uint64_t vdev_trim_partial; /* requested partial TRIM */
uint64_t vdev_trim_secure; /* requested secure TRIM */
uint64_t vdev_trim_action_time; /* start and end time */
/* Rebuild related */
boolean_t vdev_rebuilding;
boolean_t vdev_rebuild_exit_wanted;
boolean_t vdev_rebuild_cancel_wanted;
boolean_t vdev_rebuild_reset_wanted;
kmutex_t vdev_rebuild_lock;
kcondvar_t vdev_rebuild_cv;
kthread_t *vdev_rebuild_thread;
vdev_rebuild_t vdev_rebuild_config;
/* For limiting outstanding I/Os (initialize, TRIM) */
kmutex_t vdev_initialize_io_lock;
kcondvar_t vdev_initialize_io_cv;
uint64_t vdev_initialize_inflight;
kmutex_t vdev_trim_io_lock;
kcondvar_t vdev_trim_io_cv;
uint64_t vdev_trim_inflight[3];
/*
* Values stored in the config for an indirect or removing vdev.
*/
vdev_indirect_config_t vdev_indirect_config;
/*
* The vdev_indirect_rwlock protects the vdev_indirect_mapping
* pointer from changing on indirect vdevs (when it is condensed).
* Note that removing (not yet indirect) vdevs have different
* access patterns (the mapping is not accessed from open context,
* e.g. from zio_read) and locking strategy (e.g. svr_lock).
*/
krwlock_t vdev_indirect_rwlock;
vdev_indirect_mapping_t *vdev_indirect_mapping;
vdev_indirect_births_t *vdev_indirect_births;
/*
* In memory data structures used to manage the obsolete sm, for
* indirect or removing vdevs.
*
* The vdev_obsolete_segments is the in-core record of the segments
* that are no longer referenced anywhere in the pool (due to
* being freed or remapped and not referenced by any snapshots).
* During a sync, segments are added to vdev_obsolete_segments
* via vdev_indirect_mark_obsolete(); at the end of each sync
* pass, this is appended to vdev_obsolete_sm via
* vdev_indirect_sync_obsolete(). The vdev_obsolete_lock
* protects against concurrent modifications of vdev_obsolete_segments
* from multiple zio threads.
*/
kmutex_t vdev_obsolete_lock;
range_tree_t *vdev_obsolete_segments;
space_map_t *vdev_obsolete_sm;
/*
* Protects the vdev_scan_io_queue field itself as well as the
* structure's contents (when present).
*/
kmutex_t vdev_scan_io_queue_lock;
struct dsl_scan_io_queue *vdev_scan_io_queue;
/*
* Leaf vdev state.
*/
range_tree_t *vdev_dtl[DTL_TYPES]; /* dirty time logs */
space_map_t *vdev_dtl_sm; /* dirty time log space map */
txg_node_t vdev_dtl_node; /* per-txg dirty DTL linkage */
uint64_t vdev_dtl_object; /* DTL object */
uint64_t vdev_psize; /* physical device capacity */
uint64_t vdev_wholedisk; /* true if this is a whole disk */
uint64_t vdev_offline; /* persistent offline state */
uint64_t vdev_faulted; /* persistent faulted state */
uint64_t vdev_degraded; /* persistent degraded state */
uint64_t vdev_removed; /* persistent removed state */
uint64_t vdev_resilver_txg; /* persistent resilvering state */
uint64_t vdev_rebuild_txg; /* persistent rebuilding state */
char *vdev_path; /* vdev path (if any) */
char *vdev_devid; /* vdev devid (if any) */
char *vdev_physpath; /* vdev device path (if any) */
char *vdev_enc_sysfs_path; /* enclosure sysfs path */
char *vdev_fru; /* physical FRU location */
uint64_t vdev_not_present; /* not present during import */
uint64_t vdev_unspare; /* unspare when resilvering done */
boolean_t vdev_nowritecache; /* true if flushwritecache failed */
boolean_t vdev_has_trim; /* TRIM is supported */
boolean_t vdev_has_securetrim; /* secure TRIM is supported */
boolean_t vdev_checkremove; /* temporary online test */
boolean_t vdev_forcefault; /* force online fault */
boolean_t vdev_splitting; /* split or repair in progress */
boolean_t vdev_delayed_close; /* delayed device close? */
boolean_t vdev_tmpoffline; /* device taken offline temporarily? */
boolean_t vdev_detached; /* device detached? */
boolean_t vdev_cant_read; /* vdev is failing all reads */
boolean_t vdev_cant_write; /* vdev is failing all writes */
boolean_t vdev_isspare; /* was a hot spare */
boolean_t vdev_isl2cache; /* was a l2cache device */
boolean_t vdev_copy_uberblocks; /* post expand copy uberblocks */
boolean_t vdev_resilver_deferred; /* resilver deferred */
boolean_t vdev_kobj_flag; /* kobj event record */
boolean_t vdev_attaching; /* vdev attach ashift handling */
vdev_queue_t vdev_queue; /* I/O deadline schedule queue */
spa_aux_vdev_t *vdev_aux; /* for l2cache and spares vdevs */
zio_t *vdev_probe_zio; /* root of current probe */
vdev_aux_t vdev_label_aux; /* on-disk aux state */
uint64_t vdev_leaf_zap;
hrtime_t vdev_mmp_pending; /* 0 if write finished */
uint64_t vdev_mmp_kstat_id; /* to find kstat entry */
uint64_t vdev_expansion_time; /* vdev's last expansion time */
list_node_t vdev_leaf_node; /* leaf vdev list */
/*
* For DTrace to work in userland (libzpool) context, these fields must
* remain at the end of the structure. DTrace will use the kernel's
* CTF definition for 'struct vdev', and since the size of a kmutex_t is
* larger in userland, the offsets for the rest of the fields would be
* incorrect.
*/
kmutex_t vdev_dtl_lock; /* vdev_dtl_{map,resilver} */
kmutex_t vdev_stat_lock; /* vdev_stat */
kmutex_t vdev_probe_lock; /* protects vdev_probe_zio */
/*
* We rate limit ZIO delay, deadman, and checksum events, since they
* can flood ZED with tons of events when a drive is acting up.
*/
zfs_ratelimit_t vdev_delay_rl;
zfs_ratelimit_t vdev_deadman_rl;
zfs_ratelimit_t vdev_checksum_rl;
/*
* Checksum and IO thresholds for tuning ZED
*/
uint64_t vdev_checksum_n;
uint64_t vdev_checksum_t;
uint64_t vdev_io_n;
uint64_t vdev_io_t;
};
#define VDEV_PAD_SIZE (8 << 10)
/* 2 padding areas (vl_pad1 and vl_be) to skip */
#define VDEV_SKIP_SIZE VDEV_PAD_SIZE * 2
#define VDEV_PHYS_SIZE (112 << 10)
#define VDEV_UBERBLOCK_RING (128 << 10)
/*
* MMP blocks occupy the last MMP_BLOCKS_PER_LABEL slots in the uberblock
* ring when MMP is enabled.
*/
#define MMP_BLOCKS_PER_LABEL 1
/* The largest uberblock we support is 8k. */
#define MAX_UBERBLOCK_SHIFT (13)
#define VDEV_UBERBLOCK_SHIFT(vd) \
MIN(MAX((vd)->vdev_top->vdev_ashift, UBERBLOCK_SHIFT), \
MAX_UBERBLOCK_SHIFT)
#define VDEV_UBERBLOCK_COUNT(vd) \
(VDEV_UBERBLOCK_RING >> VDEV_UBERBLOCK_SHIFT(vd))
#define VDEV_UBERBLOCK_OFFSET(vd, n) \
offsetof(vdev_label_t, vl_uberblock[(n) << VDEV_UBERBLOCK_SHIFT(vd)])
#define VDEV_UBERBLOCK_SIZE(vd) (1ULL << VDEV_UBERBLOCK_SHIFT(vd))
typedef struct vdev_phys {
char vp_nvlist[VDEV_PHYS_SIZE - sizeof (zio_eck_t)];
zio_eck_t vp_zbt;
} vdev_phys_t;
typedef enum vbe_vers {
/*
* The bootenv file is stored as ascii text in the envblock.
* It is used by the GRUB bootloader used on Linux to store the
* contents of the grubenv file. The file is stored as raw ASCII,
* and is protected by an embedded checksum. By default, GRUB will
* check if the boot filesystem supports storing the environment data
* in a special location, and if so, will invoke filesystem specific
* logic to retrieve it. This can be overridden by a variable, should
* the user so desire.
*/
VB_RAW = 0,
/*
* The bootenv file is converted to an nvlist and then packed into the
* envblock.
*/
VB_NVLIST = 1
} vbe_vers_t;
typedef struct vdev_boot_envblock {
uint64_t vbe_version;
char vbe_bootenv[VDEV_PAD_SIZE - sizeof (uint64_t) -
sizeof (zio_eck_t)];
zio_eck_t vbe_zbt;
} vdev_boot_envblock_t;
_Static_assert(sizeof (vdev_boot_envblock_t) == VDEV_PAD_SIZE,
"vdev_boot_envblock_t wrong size");
typedef struct vdev_label {
char vl_pad1[VDEV_PAD_SIZE]; /* 8K */
vdev_boot_envblock_t vl_be; /* 8K */
vdev_phys_t vl_vdev_phys; /* 112K */
char vl_uberblock[VDEV_UBERBLOCK_RING]; /* 128K */
} vdev_label_t; /* 256K total */
/*
* vdev_dirty() flags
*/
#define VDD_METASLAB 0x01
#define VDD_DTL 0x02
/* Offset of embedded boot loader region on each label */
#define VDEV_BOOT_OFFSET (2 * sizeof (vdev_label_t))
/*
* Size of embedded boot loader region on each label.
* The total size of the first two labels plus the boot area is 4MB.
+ * On RAIDZ, this space is overwritten during RAIDZ expansion.
*/
#define VDEV_BOOT_SIZE (7ULL << 19) /* 3.5M */
/*
* Size of label regions at the start and end of each leaf device.
*/
#define VDEV_LABEL_START_SIZE (2 * sizeof (vdev_label_t) + VDEV_BOOT_SIZE)
#define VDEV_LABEL_END_SIZE (2 * sizeof (vdev_label_t))
#define VDEV_LABELS 4
#define VDEV_BEST_LABEL VDEV_LABELS
#define VDEV_OFFSET_IS_LABEL(vd, off) \
(((off) < VDEV_LABEL_START_SIZE) || \
((off) >= ((vd)->vdev_psize - VDEV_LABEL_END_SIZE)))
#define VDEV_ALLOC_LOAD 0
#define VDEV_ALLOC_ADD 1
#define VDEV_ALLOC_SPARE 2
#define VDEV_ALLOC_L2CACHE 3
#define VDEV_ALLOC_ROOTPOOL 4
#define VDEV_ALLOC_SPLIT 5
#define VDEV_ALLOC_ATTACH 6
/*
* Allocate or free a vdev
*/
extern vdev_t *vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid,
vdev_ops_t *ops);
extern int vdev_alloc(spa_t *spa, vdev_t **vdp, nvlist_t *config,
vdev_t *parent, uint_t id, int alloctype);
extern void vdev_free(vdev_t *vd);
/*
* Add or remove children and parents
*/
extern void vdev_add_child(vdev_t *pvd, vdev_t *cvd);
extern void vdev_remove_child(vdev_t *pvd, vdev_t *cvd);
extern void vdev_compact_children(vdev_t *pvd);
extern vdev_t *vdev_add_parent(vdev_t *cvd, vdev_ops_t *ops);
extern void vdev_remove_parent(vdev_t *cvd);
/*
* vdev sync load and sync
*/
extern boolean_t vdev_log_state_valid(vdev_t *vd);
extern int vdev_load(vdev_t *vd);
extern int vdev_dtl_load(vdev_t *vd);
extern void vdev_sync(vdev_t *vd, uint64_t txg);
extern void vdev_sync_done(vdev_t *vd, uint64_t txg);
extern void vdev_dirty(vdev_t *vd, int flags, void *arg, uint64_t txg);
extern void vdev_dirty_leaves(vdev_t *vd, int flags, uint64_t txg);
/*
* Available vdev types.
*/
extern vdev_ops_t vdev_root_ops;
extern vdev_ops_t vdev_mirror_ops;
extern vdev_ops_t vdev_replacing_ops;
extern vdev_ops_t vdev_raidz_ops;
extern vdev_ops_t vdev_draid_ops;
extern vdev_ops_t vdev_draid_spare_ops;
extern vdev_ops_t vdev_disk_ops;
extern vdev_ops_t vdev_file_ops;
extern vdev_ops_t vdev_missing_ops;
extern vdev_ops_t vdev_hole_ops;
extern vdev_ops_t vdev_spare_ops;
extern vdev_ops_t vdev_indirect_ops;
/*
* Common size functions
*/
extern void vdev_default_xlate(vdev_t *vd, const range_seg64_t *logical_rs,
range_seg64_t *physical_rs, range_seg64_t *remain_rs);
-extern uint64_t vdev_default_asize(vdev_t *vd, uint64_t psize);
+extern uint64_t vdev_default_asize(vdev_t *vd, uint64_t psize, uint64_t txg);
extern uint64_t vdev_default_min_asize(vdev_t *vd);
extern uint64_t vdev_get_min_asize(vdev_t *vd);
extern void vdev_set_min_asize(vdev_t *vd);
extern uint64_t vdev_get_min_alloc(vdev_t *vd);
extern uint64_t vdev_get_nparity(vdev_t *vd);
extern uint64_t vdev_get_ndisks(vdev_t *vd);
/*
* Global variables
*/
extern int zfs_vdev_standard_sm_blksz;
/*
* Functions from vdev_indirect.c
*/
extern void vdev_indirect_sync_obsolete(vdev_t *vd, dmu_tx_t *tx);
extern boolean_t vdev_indirect_should_condense(vdev_t *vd);
extern void spa_condense_indirect_start_sync(vdev_t *vd, dmu_tx_t *tx);
extern int vdev_obsolete_sm_object(vdev_t *vd, uint64_t *sm_obj);
extern int vdev_obsolete_counts_are_precise(vdev_t *vd, boolean_t *are_precise);
/*
* Other miscellaneous functions
*/
int vdev_checkpoint_sm_object(vdev_t *vd, uint64_t *sm_obj);
void vdev_metaslab_group_create(vdev_t *vd);
uint64_t vdev_best_ashift(uint64_t logical, uint64_t a, uint64_t b);
/*
* Vdev ashift optimization tunables
*/
extern uint_t zfs_vdev_min_auto_ashift;
extern uint_t zfs_vdev_max_auto_ashift;
int param_set_min_auto_ashift(ZFS_MODULE_PARAM_ARGS);
int param_set_max_auto_ashift(ZFS_MODULE_PARAM_ARGS);
#ifdef __cplusplus
}
#endif
#endif /* _SYS_VDEV_IMPL_H */
diff --git a/sys/contrib/openzfs/include/sys/vdev_raidz.h b/sys/contrib/openzfs/include/sys/vdev_raidz.h
index e34b6e4b158e..a34bc00ca4df 100644
--- a/sys/contrib/openzfs/include/sys/vdev_raidz.h
+++ b/sys/contrib/openzfs/include/sys/vdev_raidz.h
@@ -1,77 +1,176 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (C) 2016 Gvozden Neskovic <neskovic@compeng.uni-frankfurt.de>.
*/
#ifndef _SYS_VDEV_RAIDZ_H
#define _SYS_VDEV_RAIDZ_H
#include <sys/types.h>
+#include <sys/zfs_rlock.h>
#ifdef __cplusplus
extern "C" {
#endif
struct zio;
struct raidz_col;
struct raidz_row;
struct raidz_map;
+struct vdev_raidz;
+struct uberblock;
#if !defined(_KERNEL)
struct kernel_param {};
#endif
/*
* vdev_raidz interface
*/
struct raidz_map *vdev_raidz_map_alloc(struct zio *, uint64_t, uint64_t,
uint64_t);
+struct raidz_map *vdev_raidz_map_alloc_expanded(struct zio *,
+ uint64_t, uint64_t, uint64_t, uint64_t, uint64_t, uint64_t, boolean_t);
void vdev_raidz_map_free(struct raidz_map *);
+void vdev_raidz_free(struct vdev_raidz *);
void vdev_raidz_generate_parity_row(struct raidz_map *, struct raidz_row *);
void vdev_raidz_generate_parity(struct raidz_map *);
void vdev_raidz_reconstruct(struct raidz_map *, const int *, int);
void vdev_raidz_child_done(zio_t *);
void vdev_raidz_io_done(zio_t *);
void vdev_raidz_checksum_error(zio_t *, struct raidz_col *, abd_t *);
+struct raidz_row *vdev_raidz_row_alloc(int);
+void vdev_raidz_reflow_copy_scratch(spa_t *);
+void raidz_dtl_reassessed(vdev_t *);
extern const zio_vsd_ops_t vdev_raidz_vsd_ops;
/*
* vdev_raidz_math interface
*/
void vdev_raidz_math_init(void);
void vdev_raidz_math_fini(void);
const struct raidz_impl_ops *vdev_raidz_math_get_ops(void);
int vdev_raidz_math_generate(struct raidz_map *, struct raidz_row *);
int vdev_raidz_math_reconstruct(struct raidz_map *, struct raidz_row *,
const int *, const int *, const int);
int vdev_raidz_impl_set(const char *);
+typedef struct vdev_raidz_expand {
+ uint64_t vre_vdev_id;
+
+ kmutex_t vre_lock;
+ kcondvar_t vre_cv;
+
+ /*
+ * How much i/o is outstanding (issued and not completed).
+ */
+ uint64_t vre_outstanding_bytes;
+
+ /*
+ * Next offset to issue i/o for.
+ */
+ uint64_t vre_offset;
+
+ /*
+ * Lowest offset of a failed expansion i/o. The expansion will retry
+ * from here. Once the expansion thread notices the failure and exits,
+ * vre_failed_offset is reset back to UINT64_MAX, and
+ * vre_waiting_for_resilver will be set.
+ */
+ uint64_t vre_failed_offset;
+ boolean_t vre_waiting_for_resilver;
+
+ /*
+ * Offset that is completing each txg
+ */
+ uint64_t vre_offset_pertxg[TXG_SIZE];
+
+ /*
+ * Bytes copied in each txg.
+ */
+ uint64_t vre_bytes_copied_pertxg[TXG_SIZE];
+
+ /*
+ * The rangelock prevents normal read/write zio's from happening while
+ * there are expansion (reflow) i/os in progress to the same offsets.
+ */
+ zfs_rangelock_t vre_rangelock;
+
+ /*
+ * These fields are stored on-disk in the vdev_top_zap:
+ */
+ dsl_scan_state_t vre_state;
+ uint64_t vre_start_time;
+ uint64_t vre_end_time;
+ uint64_t vre_bytes_copied;
+} vdev_raidz_expand_t;
+
typedef struct vdev_raidz {
- int vd_logical_width;
+ /*
+ * Number of child vdevs when this raidz vdev was created (i.e. before
+ * any raidz expansions).
+ */
+ int vd_original_width;
+
+ /*
+ * The current number of child vdevs, which may be more than the
+ * original width if an expansion is in progress or has completed.
+ */
+ int vd_physical_width;
+
int vd_nparity;
+
+ /*
+ * Tree of reflow_node_t's. The lock protects the avl tree only.
+ * The reflow_node_t's describe completed expansions, and are used
+ * to determine the logical width given a block's birth time.
+ */
+ avl_tree_t vd_expand_txgs;
+ kmutex_t vd_expand_lock;
+
+ /*
+ * If this vdev is being expanded, spa_raidz_expand is set to this
+ */
+ vdev_raidz_expand_t vn_vre;
} vdev_raidz_t;
+extern int vdev_raidz_attach_check(vdev_t *);
+extern void vdev_raidz_attach_sync(void *, dmu_tx_t *);
+extern void spa_start_raidz_expansion_thread(spa_t *);
+extern int spa_raidz_expand_get_stats(spa_t *, pool_raidz_expand_stat_t *);
+extern int vdev_raidz_load(vdev_t *);
+
+/* RAIDZ scratch area pause points (for testing) */
+#define RAIDZ_EXPAND_PAUSE_NONE 0
+#define RAIDZ_EXPAND_PAUSE_PRE_SCRATCH_1 1
+#define RAIDZ_EXPAND_PAUSE_PRE_SCRATCH_2 2
+#define RAIDZ_EXPAND_PAUSE_PRE_SCRATCH_3 3
+#define RAIDZ_EXPAND_PAUSE_SCRATCH_VALID 4
+#define RAIDZ_EXPAND_PAUSE_SCRATCH_REFLOWED 5
+#define RAIDZ_EXPAND_PAUSE_SCRATCH_POST_REFLOW_1 6
+#define RAIDZ_EXPAND_PAUSE_SCRATCH_POST_REFLOW_2 7
+
#ifdef __cplusplus
}
#endif
#endif /* _SYS_VDEV_RAIDZ_H */
diff --git a/sys/contrib/openzfs/include/sys/vdev_raidz_impl.h b/sys/contrib/openzfs/include/sys/vdev_raidz_impl.h
index c1037fa12e30..fae03f8f53da 100644
--- a/sys/contrib/openzfs/include/sys/vdev_raidz_impl.h
+++ b/sys/contrib/openzfs/include/sys/vdev_raidz_impl.h
@@ -1,389 +1,410 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (C) 2016 Gvozden Nešković. All rights reserved.
*/
#ifndef _VDEV_RAIDZ_H
#define _VDEV_RAIDZ_H
#include <sys/types.h>
#include <sys/debug.h>
#include <sys/kstat.h>
#include <sys/abd.h>
#include <sys/vdev_impl.h>
+#include <sys/abd_impl.h>
+#include <sys/zfs_rlock.h>
#ifdef __cplusplus
extern "C" {
#endif
#define CODE_P (0U)
#define CODE_Q (1U)
#define CODE_R (2U)
#define PARITY_P (1U)
#define PARITY_PQ (2U)
#define PARITY_PQR (3U)
#define TARGET_X (0U)
#define TARGET_Y (1U)
#define TARGET_Z (2U)
/*
* Parity generation methods indexes
*/
enum raidz_math_gen_op {
RAIDZ_GEN_P = 0,
RAIDZ_GEN_PQ,
RAIDZ_GEN_PQR,
RAIDZ_GEN_NUM = 3
};
/*
* Data reconstruction methods indexes
*/
enum raidz_rec_op {
RAIDZ_REC_P = 0,
RAIDZ_REC_Q,
RAIDZ_REC_R,
RAIDZ_REC_PQ,
RAIDZ_REC_PR,
RAIDZ_REC_QR,
RAIDZ_REC_PQR,
RAIDZ_REC_NUM = 7
};
extern const char *const raidz_gen_name[RAIDZ_GEN_NUM];
extern const char *const raidz_rec_name[RAIDZ_REC_NUM];
/*
* Methods used to define raidz implementation
*
* @raidz_gen_f Parity generation function
* @par1 pointer to raidz_map
* @raidz_rec_f Data reconstruction function
* @par1 pointer to raidz_map
* @par2 array of reconstruction targets
* @will_work_f Function returns TRUE if impl. is supported on the system
* @init_impl_f Function is called once on init
* @fini_impl_f Function is called once on fini
*/
typedef void (*raidz_gen_f)(void *);
typedef int (*raidz_rec_f)(void *, const int *);
typedef boolean_t (*will_work_f)(void);
typedef void (*init_impl_f)(void);
typedef void (*fini_impl_f)(void);
#define RAIDZ_IMPL_NAME_MAX (20)
typedef struct raidz_impl_ops {
init_impl_f init;
fini_impl_f fini;
raidz_gen_f gen[RAIDZ_GEN_NUM]; /* Parity generate functions */
raidz_rec_f rec[RAIDZ_REC_NUM]; /* Data reconstruction functions */
will_work_f is_supported; /* Support check function */
char name[RAIDZ_IMPL_NAME_MAX]; /* Name of the implementation */
} raidz_impl_ops_t;
+
typedef struct raidz_col {
- uint64_t rc_devidx; /* child device index for I/O */
+ int rc_devidx; /* child device index for I/O */
+ uint32_t rc_size; /* I/O size */
uint64_t rc_offset; /* device offset */
- uint64_t rc_size; /* I/O size */
abd_t rc_abdstruct; /* rc_abd probably points here */
abd_t *rc_abd; /* I/O data */
abd_t *rc_orig_data; /* pre-reconstruction */
int rc_error; /* I/O error for this device */
- uint8_t rc_tried; /* Did we attempt this I/O column? */
- uint8_t rc_skipped; /* Did we skip this I/O column? */
- uint8_t rc_need_orig_restore; /* need to restore from orig_data? */
- uint8_t rc_force_repair; /* Write good data to this column */
- uint8_t rc_allow_repair; /* Allow repair I/O to this column */
+ uint8_t rc_tried:1; /* Did we attempt this I/O column? */
+ uint8_t rc_skipped:1; /* Did we skip this I/O column? */
+ uint8_t rc_need_orig_restore:1; /* need to restore from orig_data? */
+ uint8_t rc_force_repair:1; /* Write good data to this column */
+ uint8_t rc_allow_repair:1; /* Allow repair I/O to this column */
+ int rc_shadow_devidx; /* for double write during expansion */
+ int rc_shadow_error; /* for double write during expansion */
+ uint64_t rc_shadow_offset; /* for double write during expansion */
} raidz_col_t;
typedef struct raidz_row {
- uint64_t rr_cols; /* Regular column count */
- uint64_t rr_scols; /* Count including skipped columns */
- uint64_t rr_bigcols; /* Remainder data column count */
- uint64_t rr_missingdata; /* Count of missing data devices */
- uint64_t rr_missingparity; /* Count of missing parity devices */
- uint64_t rr_firstdatacol; /* First data column/parity count */
+ int rr_cols; /* Regular column count */
+ int rr_scols; /* Count including skipped columns */
+ int rr_bigcols; /* Remainder data column count */
+ int rr_missingdata; /* Count of missing data devices */
+ int rr_missingparity; /* Count of missing parity devices */
+ int rr_firstdatacol; /* First data column/parity count */
abd_t *rr_abd_empty; /* dRAID empty sector buffer */
int rr_nempty; /* empty sectors included in parity */
#ifdef ZFS_DEBUG
uint64_t rr_offset; /* Logical offset for *_io_verify() */
uint64_t rr_size; /* Physical size for *_io_verify() */
#endif
raidz_col_t rr_col[0]; /* Flexible array of I/O columns */
} raidz_row_t;
typedef struct raidz_map {
boolean_t rm_ecksuminjected; /* checksum error was injected */
int rm_nrows; /* Regular row count */
int rm_nskip; /* RAIDZ sectors skipped for padding */
int rm_skipstart; /* Column index of padding start */
+ int rm_original_width; /* pre-expansion width of raidz vdev */
+ int rm_nphys_cols; /* num entries in rm_phys_col[] */
+ zfs_locked_range_t *rm_lr;
const raidz_impl_ops_t *rm_ops; /* RAIDZ math operations */
+ raidz_col_t *rm_phys_col; /* if non-NULL, read i/o aggregation */
raidz_row_t *rm_row[0]; /* flexible array of rows */
} raidz_map_t;
+/*
+ * Nodes in vdev_raidz_t:vd_expand_txgs.
+ * Blocks with physical birth time of re_txg or later have the specified
+ * logical width (until the next node).
+ */
+typedef struct reflow_node {
+ uint64_t re_txg;
+ uint64_t re_logical_width;
+ avl_node_t re_link;
+} reflow_node_t;
+
#define RAIDZ_ORIGINAL_IMPL (INT_MAX)
extern const raidz_impl_ops_t vdev_raidz_scalar_impl;
extern boolean_t raidz_will_scalar_work(void);
#if defined(__x86_64) && defined(HAVE_SSE2) /* only x86_64 for now */
extern const raidz_impl_ops_t vdev_raidz_sse2_impl;
#endif
#if defined(__x86_64) && defined(HAVE_SSSE3) /* only x86_64 for now */
extern const raidz_impl_ops_t vdev_raidz_ssse3_impl;
#endif
#if defined(__x86_64) && defined(HAVE_AVX2) /* only x86_64 for now */
extern const raidz_impl_ops_t vdev_raidz_avx2_impl;
#endif
#if defined(__x86_64) && defined(HAVE_AVX512F) /* only x86_64 for now */
extern const raidz_impl_ops_t vdev_raidz_avx512f_impl;
#endif
#if defined(__x86_64) && defined(HAVE_AVX512BW) /* only x86_64 for now */
extern const raidz_impl_ops_t vdev_raidz_avx512bw_impl;
#endif
#if defined(__aarch64__)
extern const raidz_impl_ops_t vdev_raidz_aarch64_neon_impl;
extern const raidz_impl_ops_t vdev_raidz_aarch64_neonx2_impl;
#endif
#if defined(__powerpc__)
extern const raidz_impl_ops_t vdev_raidz_powerpc_altivec_impl;
#endif
/*
* Commonly used raidz_map helpers
*
* raidz_parity Returns parity of the RAIDZ block
* raidz_ncols Returns number of columns the block spans
* Note, all rows have the same number of columns.
* raidz_nbigcols Returns number of big columns
* raidz_col_p Returns pointer to a column
* raidz_col_size Returns size of a column
* raidz_big_size Returns size of big columns
* raidz_short_size Returns size of short columns
*/
#define raidz_parity(rm) ((rm)->rm_row[0]->rr_firstdatacol)
#define raidz_ncols(rm) ((rm)->rm_row[0]->rr_cols)
#define raidz_nbigcols(rm) ((rm)->rm_bigcols)
#define raidz_col_p(rm, c) ((rm)->rm_col + (c))
#define raidz_col_size(rm, c) ((rm)->rm_col[c].rc_size)
#define raidz_big_size(rm) (raidz_col_size(rm, CODE_P))
#define raidz_short_size(rm) (raidz_col_size(rm, raidz_ncols(rm)-1))
/*
* Macro defines an RAIDZ parity generation method
*
* @code parity the function produce
* @impl name of the implementation
*/
#define _RAIDZ_GEN_WRAP(code, impl) \
static void \
impl ## _gen_ ## code(void *rrp) \
{ \
raidz_row_t *rr = (raidz_row_t *)rrp; \
raidz_generate_## code ## _impl(rr); \
}
/*
* Macro defines an RAIDZ data reconstruction method
*
* @code parity the function produce
* @impl name of the implementation
*/
#define _RAIDZ_REC_WRAP(code, impl) \
static int \
impl ## _rec_ ## code(void *rrp, const int *tgtidx) \
{ \
raidz_row_t *rr = (raidz_row_t *)rrp; \
return (raidz_reconstruct_## code ## _impl(rr, tgtidx)); \
}
/*
* Define all gen methods for an implementation
*
* @impl name of the implementation
*/
#define DEFINE_GEN_METHODS(impl) \
_RAIDZ_GEN_WRAP(p, impl); \
_RAIDZ_GEN_WRAP(pq, impl); \
_RAIDZ_GEN_WRAP(pqr, impl)
/*
* Define all rec functions for an implementation
*
* @impl name of the implementation
*/
#define DEFINE_REC_METHODS(impl) \
_RAIDZ_REC_WRAP(p, impl); \
_RAIDZ_REC_WRAP(q, impl); \
_RAIDZ_REC_WRAP(r, impl); \
_RAIDZ_REC_WRAP(pq, impl); \
_RAIDZ_REC_WRAP(pr, impl); \
_RAIDZ_REC_WRAP(qr, impl); \
_RAIDZ_REC_WRAP(pqr, impl)
#define RAIDZ_GEN_METHODS(impl) \
{ \
[RAIDZ_GEN_P] = & impl ## _gen_p, \
[RAIDZ_GEN_PQ] = & impl ## _gen_pq, \
[RAIDZ_GEN_PQR] = & impl ## _gen_pqr \
}
#define RAIDZ_REC_METHODS(impl) \
{ \
[RAIDZ_REC_P] = & impl ## _rec_p, \
[RAIDZ_REC_Q] = & impl ## _rec_q, \
[RAIDZ_REC_R] = & impl ## _rec_r, \
[RAIDZ_REC_PQ] = & impl ## _rec_pq, \
[RAIDZ_REC_PR] = & impl ## _rec_pr, \
[RAIDZ_REC_QR] = & impl ## _rec_qr, \
[RAIDZ_REC_PQR] = & impl ## _rec_pqr \
}
typedef struct raidz_impl_kstat {
uint64_t gen[RAIDZ_GEN_NUM]; /* gen method speed B/s */
uint64_t rec[RAIDZ_REC_NUM]; /* rec method speed B/s */
} raidz_impl_kstat_t;
/*
* Enumerate various multiplication constants
* used in reconstruction methods
*/
typedef enum raidz_mul_info {
/* Reconstruct Q */
MUL_Q_X = 0,
/* Reconstruct R */
MUL_R_X = 0,
/* Reconstruct PQ */
MUL_PQ_X = 0,
MUL_PQ_Y = 1,
/* Reconstruct PR */
MUL_PR_X = 0,
MUL_PR_Y = 1,
/* Reconstruct QR */
MUL_QR_XQ = 0,
MUL_QR_X = 1,
MUL_QR_YQ = 2,
MUL_QR_Y = 3,
/* Reconstruct PQR */
MUL_PQR_XP = 0,
MUL_PQR_XQ = 1,
MUL_PQR_XR = 2,
MUL_PQR_YU = 3,
MUL_PQR_YP = 4,
MUL_PQR_YQ = 5,
MUL_CNT = 6
} raidz_mul_info_t;
/*
* Powers of 2 in the Galois field.
*/
extern const uint8_t vdev_raidz_pow2[256] __attribute__((aligned(256)));
/* Logs of 2 in the Galois field defined above. */
extern const uint8_t vdev_raidz_log2[256] __attribute__((aligned(256)));
/*
* Multiply a given number by 2 raised to the given power.
*/
static inline uint8_t
vdev_raidz_exp2(const uint8_t a, const unsigned exp)
{
if (a == 0)
return (0);
return (vdev_raidz_pow2[(exp + (unsigned)vdev_raidz_log2[a]) % 255]);
}
/*
* Galois Field operations.
*
* gf_exp2 - computes 2 raised to the given power
* gf_exp4 - computes 4 raised to the given power
* gf_mul - multiplication
* gf_div - division
* gf_inv - multiplicative inverse
*/
typedef unsigned gf_t;
typedef unsigned gf_log_t;
static inline gf_t
gf_mul(const gf_t a, const gf_t b)
{
gf_log_t logsum;
if (a == 0 || b == 0)
return (0);
logsum = (gf_log_t)vdev_raidz_log2[a] + (gf_log_t)vdev_raidz_log2[b];
return ((gf_t)vdev_raidz_pow2[logsum % 255]);
}
static inline gf_t
gf_div(const gf_t a, const gf_t b)
{
gf_log_t logsum;
ASSERT3U(b, >, 0);
if (a == 0)
return (0);
logsum = (gf_log_t)255 + (gf_log_t)vdev_raidz_log2[a] -
(gf_log_t)vdev_raidz_log2[b];
return ((gf_t)vdev_raidz_pow2[logsum % 255]);
}
static inline gf_t
gf_inv(const gf_t a)
{
gf_log_t logsum;
ASSERT3U(a, >, 0);
logsum = (gf_log_t)255 - (gf_log_t)vdev_raidz_log2[a];
return ((gf_t)vdev_raidz_pow2[logsum]);
}
static inline gf_t
gf_exp2(gf_log_t exp)
{
return (vdev_raidz_pow2[exp % 255]);
}
static inline gf_t
gf_exp4(gf_log_t exp)
{
ASSERT3U(exp, <=, 255);
return ((gf_t)vdev_raidz_pow2[(2 * exp) % 255]);
}
#ifdef __cplusplus
}
#endif
#endif /* _VDEV_RAIDZ_H */
diff --git a/sys/contrib/openzfs/include/sys/zfs_debug.h b/sys/contrib/openzfs/include/sys/zfs_debug.h
index a1dfef1d89ff..8d94557a5882 100644
--- a/sys/contrib/openzfs/include/sys/zfs_debug.h
+++ b/sys/contrib/openzfs/include/sys/zfs_debug.h
@@ -1,113 +1,114 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2019 by Delphix. All rights reserved.
*/
#ifndef _SYS_ZFS_DEBUG_H
#define _SYS_ZFS_DEBUG_H
#ifdef __cplusplus
extern "C" {
#endif
#ifndef TRUE
#define TRUE 1
#endif
#ifndef FALSE
#define FALSE 0
#endif
extern int zfs_flags;
extern int zfs_recover;
extern int zfs_free_leak_on_eio;
extern int zfs_dbgmsg_enable;
#define ZFS_DEBUG_DPRINTF (1 << 0)
#define ZFS_DEBUG_DBUF_VERIFY (1 << 1)
#define ZFS_DEBUG_DNODE_VERIFY (1 << 2)
#define ZFS_DEBUG_SNAPNAMES (1 << 3)
#define ZFS_DEBUG_MODIFY (1 << 4)
/* 1<<5 was previously used, try not to reuse */
#define ZFS_DEBUG_ZIO_FREE (1 << 6)
#define ZFS_DEBUG_HISTOGRAM_VERIFY (1 << 7)
#define ZFS_DEBUG_METASLAB_VERIFY (1 << 8)
#define ZFS_DEBUG_SET_ERROR (1 << 9)
#define ZFS_DEBUG_INDIRECT_REMAP (1 << 10)
#define ZFS_DEBUG_TRIM (1 << 11)
#define ZFS_DEBUG_LOG_SPACEMAP (1 << 12)
#define ZFS_DEBUG_METASLAB_ALLOC (1 << 13)
#define ZFS_DEBUG_BRT (1 << 14)
+#define ZFS_DEBUG_RAIDZ_RECONSTRUCT (1 << 15)
extern void __set_error(const char *file, const char *func, int line, int err);
extern void __zfs_dbgmsg(char *buf);
extern void __dprintf(boolean_t dprint, const char *file, const char *func,
int line, const char *fmt, ...) __attribute__((format(printf, 5, 6)));
/*
* Some general principles for using zfs_dbgmsg():
* 1. We don't want to pollute the log with typically-irrelevant messages,
* so don't print too many messages in the "normal" code path - O(1)
* per txg.
* 2. We want to know for sure what happened, so make the message specific
* (e.g. *which* thing am I operating on).
* 3. Do print a message when something unusual or unexpected happens
* (e.g. error cases).
* 4. Print a message when making user-initiated on-disk changes.
*
* Note that besides principle 1, another reason that we don't want to
* use zfs_dbgmsg in high-frequency routines is the potential impact
* that it can have on performance.
*/
#define zfs_dbgmsg(...) \
if (zfs_dbgmsg_enable) \
__dprintf(B_FALSE, __FILE__, __func__, __LINE__, __VA_ARGS__)
#ifdef ZFS_DEBUG
/*
* To enable this:
*
* $ echo 1 >/sys/module/zfs/parameters/zfs_flags
*/
#define dprintf(...) \
if (zfs_flags & ZFS_DEBUG_DPRINTF) \
__dprintf(B_TRUE, __FILE__, __func__, __LINE__, __VA_ARGS__)
#else
#define dprintf(...) ((void)0)
#endif /* ZFS_DEBUG */
extern void zfs_panic_recover(const char *fmt, ...);
extern void zfs_dbgmsg_init(void);
extern void zfs_dbgmsg_fini(void);
#ifndef _KERNEL
extern int dprintf_find_string(const char *string);
extern void zfs_dbgmsg_print(const char *tag);
#endif
#ifdef __cplusplus
}
#endif
#endif /* _SYS_ZFS_DEBUG_H */
diff --git a/sys/contrib/openzfs/include/zfeature_common.h b/sys/contrib/openzfs/include/zfeature_common.h
index 1025c44738ba..2515ba321759 100644
--- a/sys/contrib/openzfs/include/zfeature_common.h
+++ b/sys/contrib/openzfs/include/zfeature_common.h
@@ -1,141 +1,142 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2011, 2018 by Delphix. All rights reserved.
* Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
* Copyright (c) 2013, Joyent, Inc. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
*/
#ifndef _ZFEATURE_COMMON_H
#define _ZFEATURE_COMMON_H extern __attribute__((visibility("default")))
#include <sys/fs/zfs.h>
#include <sys/inttypes.h>
#include <sys/types.h>
#ifdef __cplusplus
extern "C" {
#endif
struct zfeature_info;
typedef enum spa_feature {
SPA_FEATURE_NONE = -1,
SPA_FEATURE_ASYNC_DESTROY,
SPA_FEATURE_EMPTY_BPOBJ,
SPA_FEATURE_LZ4_COMPRESS,
SPA_FEATURE_MULTI_VDEV_CRASH_DUMP,
SPA_FEATURE_SPACEMAP_HISTOGRAM,
SPA_FEATURE_ENABLED_TXG,
SPA_FEATURE_HOLE_BIRTH,
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_EMBEDDED_DATA,
SPA_FEATURE_BOOKMARKS,
SPA_FEATURE_FS_SS_LIMIT,
SPA_FEATURE_LARGE_BLOCKS,
SPA_FEATURE_LARGE_DNODE,
SPA_FEATURE_SHA512,
SPA_FEATURE_SKEIN,
SPA_FEATURE_EDONR,
SPA_FEATURE_USEROBJ_ACCOUNTING,
SPA_FEATURE_ENCRYPTION,
SPA_FEATURE_PROJECT_QUOTA,
SPA_FEATURE_DEVICE_REMOVAL,
SPA_FEATURE_OBSOLETE_COUNTS,
SPA_FEATURE_POOL_CHECKPOINT,
SPA_FEATURE_SPACEMAP_V2,
SPA_FEATURE_ALLOCATION_CLASSES,
SPA_FEATURE_RESILVER_DEFER,
SPA_FEATURE_BOOKMARK_V2,
SPA_FEATURE_REDACTION_BOOKMARKS,
SPA_FEATURE_REDACTED_DATASETS,
SPA_FEATURE_BOOKMARK_WRITTEN,
SPA_FEATURE_LOG_SPACEMAP,
SPA_FEATURE_LIVELIST,
SPA_FEATURE_DEVICE_REBUILD,
SPA_FEATURE_ZSTD_COMPRESS,
SPA_FEATURE_DRAID,
SPA_FEATURE_ZILSAXATTR,
SPA_FEATURE_HEAD_ERRLOG,
SPA_FEATURE_BLAKE3,
SPA_FEATURE_BLOCK_CLONING,
SPA_FEATURE_AVZ_V2,
SPA_FEATURE_REDACTION_LIST_SPILL,
+ SPA_FEATURE_RAIDZ_EXPANSION,
SPA_FEATURES
} spa_feature_t;
#define SPA_FEATURE_DISABLED (-1ULL)
typedef enum zfeature_flags {
/* Can open pool readonly even if this feature is not supported. */
ZFEATURE_FLAG_READONLY_COMPAT = (1 << 0),
/*
* Is this feature necessary to load the pool? i.e. do we need this
* feature to read the full feature list out of the MOS?
*/
ZFEATURE_FLAG_MOS = (1 << 1),
/* Activate this feature at the same time it is enabled. */
ZFEATURE_FLAG_ACTIVATE_ON_ENABLE = (1 << 2),
/* Each dataset has a field set if it has ever used this feature. */
ZFEATURE_FLAG_PER_DATASET = (1 << 3)
} zfeature_flags_t;
typedef enum zfeature_type {
ZFEATURE_TYPE_BOOLEAN,
ZFEATURE_TYPE_UINT64_ARRAY,
ZFEATURE_NUM_TYPES
} zfeature_type_t;
typedef struct zfeature_info {
spa_feature_t fi_feature;
const char *fi_uname; /* User-facing feature name */
const char *fi_guid; /* On-disk feature identifier */
const char *fi_desc; /* Feature description */
zfeature_flags_t fi_flags;
boolean_t fi_zfs_mod_supported; /* supported by running zfs module */
zfeature_type_t fi_type; /* Only relevant for PER_DATASET features */
/* array of dependencies, terminated by SPA_FEATURE_NONE */
const spa_feature_t *fi_depends;
} zfeature_info_t;
typedef int (zfeature_func_t)(zfeature_info_t *, void *);
#define ZFS_FEATURE_DEBUG
_ZFEATURE_COMMON_H zfeature_info_t spa_feature_table[SPA_FEATURES];
_ZFEATURE_COMMON_H boolean_t zfeature_checks_disable;
_ZFEATURE_COMMON_H boolean_t zfeature_is_valid_guid(const char *);
_ZFEATURE_COMMON_H boolean_t zfeature_is_supported(const char *);
_ZFEATURE_COMMON_H int zfeature_lookup_guid(const char *, spa_feature_t *);
_ZFEATURE_COMMON_H int zfeature_lookup_name(const char *, spa_feature_t *);
_ZFEATURE_COMMON_H boolean_t zfeature_depends_on(spa_feature_t, spa_feature_t);
_ZFEATURE_COMMON_H void zpool_feature_init(void);
#ifdef __cplusplus
}
#endif
#endif /* _ZFEATURE_COMMON_H */
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs.abi b/sys/contrib/openzfs/lib/libzfs/libzfs.abi
index bcdcff97666e..51b0368f8cf7 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs.abi
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs.abi
@@ -1,9437 +1,9584 @@
<abi-corpus version='2.0' architecture='elf-amd-x86_64' soname='libzfs.so.4'>
<elf-needed>
<dependency name='libzfs_core.so.3'/>
<dependency name='libnvpair.so.3'/>
<dependency name='libuuid.so.1'/>
<dependency name='libblkid.so.1'/>
<dependency name='libudev.so.1'/>
<dependency name='libuutil.so.3'/>
<dependency name='libm.so.6'/>
<dependency name='libcrypto.so.3'/>
<dependency name='libz.so.1'/>
<dependency name='libc.so.6'/>
<dependency name='ld-linux-x86-64.so.2'/>
</elf-needed>
<elf-function-symbols>
<elf-symbol name='_sol_getmntent' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_char' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_char_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_int_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_long' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_long_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_ptr_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_short' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_short_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_clear_long_excl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_set_long_excl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_char' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_char_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_int_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_long' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_long_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_ptr_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_short' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_short_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_add' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_destroy_nodes' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_find' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_first' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_insert' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_insert_here' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_is_empty' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_last' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_nearest' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_numnodes' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_swap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_update' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_update_gt' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_update_lt' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_walk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='bookmark_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='cityhash4' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='color_end' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='color_start' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='dataset_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='dataset_nestcheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_alloc_and_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_alloc_and_read' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_err_check' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_free' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_rescan' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_use_whole_disk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_write' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='entity_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_2_byteswap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_2_incremental_byteswap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_2_incremental_native' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_2_native' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_byteswap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_fini' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_impl_set' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_incremental_byteswap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_incremental_native' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_native' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_native_varsize' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='get_dataset_depth' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='get_system_hostid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getexecname' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getextmntent' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getmntany' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getprop_uint64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getzoneid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='is_mounted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='is_mpath_whole_disk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libpc_error_description' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libspl_assertf' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libspl_set_assert_ok' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_add_handle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_envvar_is_set' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_errno' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_error_action' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_error_description' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_error_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_fini' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_free_str_array' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_add' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_cache' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_find' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_fini' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_print_on_error' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_run_process' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_run_process_get_stdout' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_run_process_get_stdout_nopath' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_head' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_after' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_before' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_head' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_is_empty' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_link_active' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_link_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_link_replace' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_move_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_next' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_prev' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_remove_head' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_remove_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_consumer' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_enter' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_exit' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_producer' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_sync' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='mkdirp' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='mountpoint_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='permset_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='pool_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='print_timestamp' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='printf_color' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_commit_shares' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_disable_share' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_enable_share' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_errorstr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_is_shared' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_truncate_shares' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_validate_shareopts' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='snapshot_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='spl_pagesize' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='strlcat' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='strlcpy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_abandon' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_dispatch' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_member' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_resume' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_suspend' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_suspended' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_wait' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='update_vdev_config_dev_strs' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='update_vdev_config_dev_sysfs_path' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='use_color' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_expand_proplist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_name_to_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_align_right' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_column_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_default_numeric' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_default_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_get_table' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_get_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_index_to_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_random_value' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_readonly' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_string_to_index' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_user' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_values' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_depends_on' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_is_supported' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_is_valid_guid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_lookup_guid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_lookup_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_adjust_mount_options' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_allocatable_devs' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_append_partition' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_basename' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_bookmark_exists' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_clone' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_close' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_commit_shares' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_component_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_create_ancestors' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_attempt_load_keys' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_clone_check' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_get_encryption_root' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_load_key' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_rewrap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_unload_key' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dataset_exists' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dataset_name_hidden' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_deleg_canonicalize_perm' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_deleg_verify_nvlist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_deleg_whokey' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_destroy_snaps' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_destroy_snaps_nvl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_destroy_snaps_nvl_os' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dev_flush' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dev_is_dm' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dev_is_whole_disk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_device_get_devid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_device_get_physical' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dirnamelen' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_expand_proplist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_foreach_mountpoint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_all_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_clones_nvl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_enclosure_sysfs_path' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_fsacl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_handle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_holds' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_pool_handle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_pool_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_recvd_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_underlying_path' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_underlying_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_user_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_handle_dup' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_hold' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_hold_nvl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_ioctl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_is_mounted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_is_shared' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_isnumber' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_bookmarks' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_bookmarks_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_children' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_children_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_dependents' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_dependents_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_filesystems' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_filesystems_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_mounted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_root' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_snapshots' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_snapshots_sorted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_snapshots_sorted_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_snapshots_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_snapspec' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_snapspec_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_mod_supported' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_mount' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_mount_at' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_mount_delegation_check' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_name_to_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_name_valid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_nicebytes' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_nicenum' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_nicenum_format' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_niceraw' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_nicestrtonum' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_nicetime' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_open' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_parent_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_parse_mount_options' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_path_to_zhandle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_promote' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_align_right' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_column_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_default_numeric' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_default_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_delegatable' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_encryption_key_param' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_numeric' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_recvd' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_table' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_userquota' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_userquota_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_written' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_written_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_index_to_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_inherit' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_inheritable' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_is_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_random_value' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_readonly' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_set' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_set_list' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_set_list_flags' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_setonce' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_string_to_index' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_user' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_userquota' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_valid_for_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_valid_keylocation' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_values' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_visible' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_written' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prune_proplist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_receive' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_refresh_properties' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_release' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_rename' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_resolve_shortname' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_rollback' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_save_arguments' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send_one' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send_progress' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send_resume' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send_resume_token_to_nvlist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send_saved' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_set_fsacl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_setproctitle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_setproctitle_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_share' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_show_diffs' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_smb_acl_add' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_smb_acl_purge' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_smb_acl_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_smb_acl_rename' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_snapshot' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_snapshot_nvl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_spa_version' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_spa_version_map' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_special_devs' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_standard_error' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_strcmp_pathname' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_strip_partition' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_strip_path' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_truncate_shares' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_type_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_unmount' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_unmountall' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_unshare' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_unshareall' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_userns' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_userspace' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_valid_proplist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_version_kernel' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_version_print' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_version_userland' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_wait_status' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_zpl_version_map' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_add' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_checkpoint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_clear' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_clear_label' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_close' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_default_search_paths' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_disable_datasets' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_disable_datasets_os' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_disable_volume_os' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_discard_checkpoint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_dump_ddt' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_enable_datasets' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_events_clear' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_events_next' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_events_seek' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_expand_proplist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_explain_recover' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_export' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_export_force' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_feature_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_find_config' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_find_vdev' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_find_vdev_by_physpath' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_free_handles' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_all_vdev_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_bootenv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_config' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_errlog' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_features' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_handle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_history' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_load_policy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_prop_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_state' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_state_str' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_status' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_userprop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_vdev_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_vdev_prop_value' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_history_unpack' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_import' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_import_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_import_status' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_in_use' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_initialize' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_initialize_wait' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_is_draid_spare' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_iter' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_label_disk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_label_disk_wait' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_load_compat' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_log_history' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_name_to_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_obj_to_path' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_obj_to_path_ds' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_open' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_open_canfail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_pool_state_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prepare_and_label_disk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prepare_disk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_print_unsup_feat' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_align_right' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_column_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_default_numeric' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_default_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_feature' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_get_feature' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_get_table' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_get_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_index_to_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_random_value' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_readonly' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_setonce' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_string_to_index' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_unsupported' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_values' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_vdev' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_props_refresh' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_read_label' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_refresh_stats' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_reguid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_reopen_one' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_scan' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_search_import' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_set_bootenv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_set_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_set_vdev_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_skip_pool' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_state_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_sync_one' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_trim' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_upgrade' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_attach' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_clear' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_degrade' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_detach' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_fault' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_indirect_size' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_offline' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_online' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_path_to_guid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_remove_cancel' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_remove_wanted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_script_alloc_env' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_script_free_env' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_split' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_wait' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_wait_status' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_free_list' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_get_list' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_index_to_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_iter' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_iter_common' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_name_to_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_print_one_property' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_random_value' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_register_hidden' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_register_impl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_register_index' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_register_number' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_register_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_string_to_index' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_valid_char' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_valid_for_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_values' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_width' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zvol_volsize_to_reservation' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
</elf-function-symbols>
<elf-variable-symbols>
<elf-symbol name='efi_debug' size='4' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_abd_ops' size='24' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_avx2_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_avx512bw_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_avx512f_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_sse2_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_ssse3_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_superscalar4_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_superscalar_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_config_ops' size='16' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_protocol_names' size='16' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
- <elf-symbol name='spa_feature_table' size='2240' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+ <elf-symbol name='spa_feature_table' size='2296' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_checks_disable' size='4' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_deleg_perm_tab' size='512' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_history_event_names' size='328' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_max_dataset_nesting' size='4' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_userquota_prop_prefixes' size='96' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
</elf-variable-symbols>
<abi-instr address-size='64' path='lib/libefi/rdwr_efi.c' language='LANG_C99'>
<typedef-decl name='uInt' type-id='f0981eeb' id='09110a74'/>
<var-decl name='efi_debug' type-id='95e97e5e' mangled-name='efi_debug' visibility='default' elf-symbol-id='efi_debug'/>
<function-decl name='uuid_generate' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='cf536864'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uuid_is_null' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='354f7eb9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='crc32' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5bbcce85'/>
<parameter type-id='e8cb3e0e'/>
<parameter type-id='09110a74'/>
<return type-id='5bbcce85'/>
</function-decl>
<function-decl name='efi_err_check' mangled-name='efi_err_check' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_err_check'>
<parameter type-id='0d8119a8' name='vtoc'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libshare/libshare.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='b99c00c9' size-in-bits='128' id='2d6895a3'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
<var-decl name='sa_protocol_names' type-id='2d6895a3' mangled-name='sa_protocol_names' visibility='default' elf-symbol-id='sa_protocol_names'/>
<type-decl name='unsigned long int' size-in-bits='64' id='7359adad'/>
</abi-instr>
<abi-instr address-size='64' path='lib/libshare/nfs.c' language='LANG_C99'>
<function-decl name='rename' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='memchr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='flock' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fchmod' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='e1c52942'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='mkdir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='e1c52942'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libshare/os/linux/nfs.c' language='LANG_C99'>
<class-decl name='sa_share_impl' size-in-bits='192' is-struct='yes' visibility='default' id='72b09bf8'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='sa_zfsname' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='sa_mountpoint' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='sa_shareopts' type-id='80f4b756' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='sa_share_impl_t' type-id='946a2c6b' id='a48b47d0'/>
<class-decl name='sa_fstype_t' size-in-bits='384' is-struct='yes' naming-typedef-id='639af739' visibility='default' id='944afa86'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='enable_share' type-id='2f78a9c1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='disable_share' type-id='2f78a9c1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='is_shared' type-id='81020bc2' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='validate_shareopts' type-id='f194a8fb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='commit_shares' type-id='797ee7da' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='truncate_shares' type-id='5d51038b' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='sa_fstype_t' type-id='944afa86' id='639af739'/>
<qualified-type-def type-id='639af739' const='yes' id='d19dbca9'/>
<qualified-type-def type-id='72b09bf8' const='yes' id='484950e3'/>
<pointer-type-def type-id='484950e3' size-in-bits='64' id='946a2c6b'/>
<pointer-type-def type-id='276427e1' size-in-bits='64' id='1db260e5'/>
<qualified-type-def type-id='1db260e5' const='yes' id='797ee7da'/>
<pointer-type-def type-id='5113b296' size-in-bits='64' id='70487b28'/>
<qualified-type-def type-id='70487b28' const='yes' id='f194a8fb'/>
<pointer-type-def type-id='c13578bc' size-in-bits='64' id='fa1f29ce'/>
<qualified-type-def type-id='fa1f29ce' const='yes' id='2f78a9c1'/>
<pointer-type-def type-id='723e6cf2' size-in-bits='64' id='1d99e49c'/>
<pointer-type-def type-id='86373eb1' size-in-bits='64' id='f337456d'/>
<qualified-type-def type-id='f337456d' const='yes' id='81020bc2'/>
<qualified-type-def type-id='953b12f8' const='yes' id='5d51038b'/>
<var-decl name='libshare_nfs_type' type-id='d19dbca9' visibility='default'/>
<function-decl name='nfs_escape_mountpoint' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='9b23c9ad'/>
<parameter type-id='37e3bd22'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nfs_is_shared_impl' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='a48b47d0'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='nfs_toggle_share' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='a48b47d0'/>
<parameter type-id='1d99e49c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nfs_reset_shares' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-type size-in-bits='64' id='276427e1'>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='5113b296'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='c13578bc'>
<parameter type-id='a48b47d0'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='723e6cf2'>
<parameter type-id='a48b47d0'/>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='86373eb1'>
<parameter type-id='a48b47d0'/>
<return type-id='c19b74c3'/>
</function-type>
<function-type size-in-bits='64' id='ee076206'>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libshare/os/linux/smb.c' language='LANG_C99'>
<var-decl name='libshare_smb_type' type-id='d19dbca9' visibility='default'/>
<function-decl name='__fgets_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='266fe297'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='e75a27e9'/>
<return type-id='26a90f95'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/assert.c' language='LANG_C99'>
<function-decl name='libspl_set_assert_ok' mangled-name='libspl_set_assert_ok' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libspl_set_assert_ok'>
<parameter type-id='c19b74c3' name='val'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/atomic.c' language='LANG_C99'>
<typedef-decl name='int8_t' type-id='2171a512' id='ee31ee44'/>
<typedef-decl name='__int8_t' type-id='28577a57' id='2171a512'/>
<qualified-type-def type-id='149c6638' volatile='yes' id='5120c5f7'/>
<pointer-type-def type-id='5120c5f7' size-in-bits='64' id='93977ae7'/>
<qualified-type-def type-id='b96825af' volatile='yes' id='84ff7d66'/>
<pointer-type-def type-id='84ff7d66' size-in-bits='64' id='aa323ea4'/>
<qualified-type-def type-id='ee1f298e' volatile='yes' id='6f7e09cb'/>
<pointer-type-def type-id='6f7e09cb' size-in-bits='64' id='64698d33'/>
<function-decl name='atomic_inc_8' mangled-name='atomic_inc_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_8'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_16' mangled-name='atomic_inc_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_16'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_32' mangled-name='atomic_inc_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_32'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_ulong' mangled-name='atomic_inc_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_ulong'>
<parameter type-id='64698d33' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_8' mangled-name='atomic_dec_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_8'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_16' mangled-name='atomic_dec_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_16'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_32' mangled-name='atomic_dec_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_32'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_ulong' mangled-name='atomic_dec_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_ulong'>
<parameter type-id='64698d33' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_ptr' mangled-name='atomic_add_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_8' mangled-name='atomic_add_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_16' mangled-name='atomic_add_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_32' mangled-name='atomic_add_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_ptr' mangled-name='atomic_sub_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_8' mangled-name='atomic_sub_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_16' mangled-name='atomic_sub_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_32' mangled-name='atomic_sub_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_8' mangled-name='atomic_or_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_16' mangled-name='atomic_or_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_32' mangled-name='atomic_or_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_ulong' mangled-name='atomic_or_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_8' mangled-name='atomic_and_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_16' mangled-name='atomic_and_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_32' mangled-name='atomic_and_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_ulong' mangled-name='atomic_and_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_8_nv' mangled-name='atomic_inc_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_inc_16_nv' mangled-name='atomic_inc_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_inc_32_nv' mangled-name='atomic_inc_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_inc_ulong_nv' mangled-name='atomic_inc_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_dec_8_nv' mangled-name='atomic_dec_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_dec_16_nv' mangled-name='atomic_dec_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_dec_32_nv' mangled-name='atomic_dec_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_dec_ulong_nv' mangled-name='atomic_dec_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_add_ptr_nv' mangled-name='atomic_add_ptr_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_ptr_nv'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_add_8_nv' mangled-name='atomic_add_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_add_16_nv' mangled-name='atomic_add_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_add_32_nv' mangled-name='atomic_add_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_add_long_nv' mangled-name='atomic_add_long_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_long_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='bd54fe1a' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_sub_ptr_nv' mangled-name='atomic_sub_ptr_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_ptr_nv'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_sub_8_nv' mangled-name='atomic_sub_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_sub_16_nv' mangled-name='atomic_sub_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_sub_32_nv' mangled-name='atomic_sub_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_sub_long_nv' mangled-name='atomic_sub_long_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_long_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='bd54fe1a' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_or_8_nv' mangled-name='atomic_or_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_or_16_nv' mangled-name='atomic_or_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_or_32_nv' mangled-name='atomic_or_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_or_ulong_nv' mangled-name='atomic_or_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_and_8_nv' mangled-name='atomic_and_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_and_16_nv' mangled-name='atomic_and_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_and_32_nv' mangled-name='atomic_and_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_and_ulong_nv' mangled-name='atomic_and_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_cas_ptr' mangled-name='atomic_cas_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='eaa32e2f' name='exp'/>
<parameter type-id='eaa32e2f' name='des'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_cas_8' mangled-name='atomic_cas_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='exp'/>
<parameter type-id='b96825af' name='des'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_cas_16' mangled-name='atomic_cas_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='exp'/>
<parameter type-id='149c6638' name='des'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_cas_32' mangled-name='atomic_cas_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='exp'/>
<parameter type-id='8f92235e' name='des'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_cas_ulong' mangled-name='atomic_cas_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='exp'/>
<parameter type-id='ee1f298e' name='des'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_swap_8' mangled-name='atomic_swap_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_swap_16' mangled-name='atomic_swap_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_swap_ulong' mangled-name='atomic_swap_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_swap_ptr' mangled-name='atomic_swap_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='eaa32e2f' name='bits'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_set_long_excl' mangled-name='atomic_set_long_excl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_set_long_excl'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='3502e3ff' name='value'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='atomic_clear_long_excl' mangled-name='atomic_clear_long_excl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_clear_long_excl'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='3502e3ff' name='value'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='membar_enter' mangled-name='membar_enter' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='membar_enter'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='membar_consumer' mangled-name='membar_consumer' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='membar_consumer'>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/getexecname.c' language='LANG_C99'>
<function-decl name='getexecname' mangled-name='getexecname' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getexecname'>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='getexecname_impl' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='26a90f95'/>
<return type-id='79a0948f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/list.c' language='LANG_C99'>
<typedef-decl name='list_node_t' type-id='b0b5e45e' id='b21843b2'/>
<typedef-decl name='list_t' type-id='e824dae9' id='0899125f'/>
<class-decl name='list_node' size-in-bits='128' is-struct='yes' visibility='default' id='b0b5e45e'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='next' type-id='b03eadb4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='prev' type-id='b03eadb4' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='list' size-in-bits='256' is-struct='yes' visibility='default' id='e824dae9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='list_size' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='list_offset' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='list_head' type-id='b0b5e45e' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='b0b5e45e' size-in-bits='64' id='b03eadb4'/>
<pointer-type-def type-id='b21843b2' size-in-bits='64' id='ccc38265'/>
<pointer-type-def type-id='0899125f' size-in-bits='64' id='352ec160'/>
<function-decl name='list_create' mangled-name='list_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_create'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='b59d7dce' name='size'/>
<parameter type-id='b59d7dce' name='offset'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_destroy' mangled-name='list_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_destroy'>
<parameter type-id='352ec160' name='list'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_after' mangled-name='list_insert_after' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_after'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<parameter type-id='eaa32e2f' name='nobject'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_before' mangled-name='list_insert_before' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_before'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<parameter type-id='eaa32e2f' name='nobject'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_head' mangled-name='list_insert_head' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_head'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_tail' mangled-name='list_insert_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_tail'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_remove' mangled-name='list_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_remove'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_remove_head' mangled-name='list_remove_head' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_remove_head'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_remove_tail' mangled-name='list_remove_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_remove_tail'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_head' mangled-name='list_head' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_head'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_tail' mangled-name='list_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_tail'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_next' mangled-name='list_next' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_next'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_prev' mangled-name='list_prev' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_prev'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_move_tail' mangled-name='list_move_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_move_tail'>
<parameter type-id='352ec160' name='dst'/>
<parameter type-id='352ec160' name='src'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_link_replace' mangled-name='list_link_replace' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_link_replace'>
<parameter type-id='ccc38265' name='lold'/>
<parameter type-id='ccc38265' name='lnew'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_link_init' mangled-name='list_link_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_link_init'>
<parameter type-id='ccc38265' name='ln'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_link_active' mangled-name='list_link_active' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_link_active'>
<parameter type-id='ccc38265' name='ln'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='list_is_empty' mangled-name='list_is_empty' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_is_empty'>
<parameter type-id='352ec160' name='list'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/mkdirp.c' language='LANG_C99'>
<typedef-decl name='wchar_t' type-id='95e97e5e' id='928221d2'/>
<qualified-type-def type-id='928221d2' const='yes' id='effb3702'/>
<pointer-type-def type-id='effb3702' size-in-bits='64' id='f077d3f8'/>
<qualified-type-def type-id='f077d3f8' restrict='yes' id='598aab80'/>
<pointer-type-def type-id='928221d2' size-in-bits='64' id='323d93c1'/>
<qualified-type-def type-id='323d93c1' restrict='yes' id='f1358bc3'/>
<function-decl name='__mbstowcs_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f1358bc3'/>
<parameter type-id='9d26089a'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='__wcstombs_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='266fe297'/>
<parameter type-id='598aab80'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/os/linux/getmntany.c' language='LANG_C99'>
<pointer-type-def type-id='56fe4a37' size-in-bits='64' id='b6b61d2f'/>
<qualified-type-def type-id='b6b61d2f' restrict='yes' id='3cad23cd'/>
<function-decl name='getmntent_r' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e75a27e9'/>
<parameter type-id='3cad23cd'/>
<parameter type-id='266fe297'/>
<parameter type-id='95e97e5e'/>
<return type-id='b6b61d2f'/>
</function-decl>
<function-decl name='feof' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/timestamp.c' language='LANG_C99'>
<typedef-decl name='nl_item' type-id='95e97e5e' id='03b79a94'/>
<function-decl name='nl_langinfo' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='03b79a94'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='print_timestamp' mangled-name='print_timestamp' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='print_timestamp'>
<parameter type-id='3502e3ff' name='timestamp_fmt'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libtpool/thread_pool.c' language='LANG_C99'>
- <array-type-def dimensions='1' type-id='8901473c' size-in-bits='576' id='f5da478b'>
- <subrange length='1' type-id='7359adad' id='52f813b4'/>
- </array-type-def>
<array-type-def dimensions='1' type-id='49ef3ffd' size-in-bits='1024' id='a14403f5'>
<subrange length='16' type-id='7359adad' id='848d0938'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='384' id='36d7f119'>
<subrange length='48' type-id='7359adad' id='8f6d2a81'/>
</array-type-def>
- <array-type-def dimensions='1' type-id='bd54fe1a' size-in-bits='512' id='5d4efd44'>
- <subrange length='8' type-id='7359adad' id='56e0c0b1'/>
- </array-type-def>
<array-type-def dimensions='1' type-id='f0981eeb' size-in-bits='64' id='0d532ec1'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
- <array-type-def dimensions='1' type-id='eaa32e2f' size-in-bits='256' id='209ef23f'>
- <subrange length='4' type-id='7359adad' id='16fe7105'/>
- </array-type-def>
- <class-decl name='__cancel_jmp_buf_tag' size-in-bits='576' is-struct='yes' visibility='default' id='8901473c'>
- <data-member access='public' layout-offset-in-bits='0'>
- <var-decl name='__cancel_jmp_buf' type-id='379a1ab7' visibility='default'/>
- </data-member>
- <data-member access='public' layout-offset-in-bits='512'>
- <var-decl name='__mask_was_saved' type-id='95e97e5e' visibility='default'/>
- </data-member>
- </class-decl>
- <class-decl name='__pthread_unwind_buf_t' size-in-bits='832' is-struct='yes' naming-typedef-id='4423cf7f' visibility='default' id='a0abc656'>
- <data-member access='public' layout-offset-in-bits='0'>
- <var-decl name='__cancel_jmp_buf' type-id='f5da478b' visibility='default'/>
- </data-member>
- <data-member access='public' layout-offset-in-bits='576'>
- <var-decl name='__pad' type-id='209ef23f' visibility='default'/>
- </data-member>
- </class-decl>
- <typedef-decl name='__pthread_unwind_buf_t' type-id='a0abc656' id='4423cf7f'/>
<union-decl name='__atomic_wide_counter' size-in-bits='64' naming-typedef-id='f3b40860' visibility='default' id='613ce450'>
<data-member access='public'>
<var-decl name='__value64' type-id='3a47d82b' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__value32' type-id='e7f43f72' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='__anonymous_struct__' size-in-bits='64' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f72'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__low' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='__high' type-id='f0981eeb' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__atomic_wide_counter' type-id='613ce450' id='f3b40860'/>
<typedef-decl name='__cpu_mask' type-id='7359adad' id='49ef3ffd'/>
<class-decl name='cpu_set_t' size-in-bits='1024' is-struct='yes' naming-typedef-id='8037c762' visibility='default' id='1f20d231'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__bits' type-id='a14403f5' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='cpu_set_t' type-id='1f20d231' id='8037c762'/>
<union-decl name='pthread_condattr_t' size-in-bits='32' naming-typedef-id='836265dd' visibility='default' id='33dd3aad'>
<data-member access='public'>
<var-decl name='__size' type-id='8e0573fd' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='95e97e5e' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_condattr_t' type-id='33dd3aad' id='836265dd'/>
<union-decl name='pthread_cond_t' size-in-bits='384' naming-typedef-id='62fab762' visibility='default' id='cbb12c12'>
<data-member access='public'>
<var-decl name='__data' type-id='c987b47c' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__size' type-id='36d7f119' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='1eb56b1e' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_cond_t' type-id='cbb12c12' id='62fab762'/>
- <typedef-decl name='__jmp_buf' type-id='5d4efd44' id='379a1ab7'/>
<class-decl name='__pthread_cond_s' size-in-bits='384' is-struct='yes' visibility='default' id='c987b47c'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__wseq' type-id='f3b40860' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='__g1_start' type-id='f3b40860' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='__g_refs' type-id='0d532ec1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='__g_size' type-id='0d532ec1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='__g1_orig_size' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='__wrefs' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='__g_signals' type-id='0d532ec1' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='sched_param' size-in-bits='32' is-struct='yes' visibility='default' id='0897719a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='sched_priority' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='tpool_job_t' type-id='3b8579e5' id='66a0afc9'/>
<class-decl name='tpool_job' size-in-bits='192' is-struct='yes' visibility='default' id='3b8579e5'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tpj_next' type-id='f32b30e4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tpj_func' type-id='b7f9d8e6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='tpj_arg' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='tpool_active_t' type-id='c8d086f4' id='6fcda10e'/>
<class-decl name='tpool_active' size-in-bits='128' is-struct='yes' visibility='default' id='c8d086f4'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tpa_next' type-id='ad33e5e7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tpa_tid' type-id='4051f5e7' visibility='default'/>
</data-member>
</class-decl>
- <pointer-type-def type-id='8901473c' size-in-bits='64' id='eb91b7ea'/>
- <pointer-type-def type-id='4423cf7f' size-in-bits='64' id='ba7c727c'/>
- <pointer-type-def type-id='b9c97942' size-in-bits='64' id='bbf06c47'/>
- <qualified-type-def type-id='bbf06c47' restrict='yes' id='65e6ec45'/>
- <qualified-type-def type-id='b9c97942' const='yes' id='191f6b72'/>
- <pointer-type-def type-id='191f6b72' size-in-bits='64' id='e475fb88'/>
- <qualified-type-def type-id='e475fb88' restrict='yes' id='5a8729d0'/>
<qualified-type-def type-id='8037c762' const='yes' id='f50ea9b2'/>
<pointer-type-def type-id='f50ea9b2' size-in-bits='64' id='5e14fa48'/>
<qualified-type-def type-id='836265dd' const='yes' id='7d24c58d'/>
<pointer-type-def type-id='7d24c58d' size-in-bits='64' id='a7e325e5'/>
<qualified-type-def type-id='a7e325e5' restrict='yes' id='4c428e67'/>
<qualified-type-def type-id='0897719a' const='yes' id='c4a7b189'/>
<pointer-type-def type-id='c4a7b189' size-in-bits='64' id='36fca399'/>
<qualified-type-def type-id='36fca399' restrict='yes' id='37e4897b'/>
<qualified-type-def type-id='a9c79a1f' const='yes' id='cd087e36'/>
<pointer-type-def type-id='cd087e36' size-in-bits='64' id='e05e8614'/>
<qualified-type-def type-id='e05e8614' restrict='yes' id='0be2e71c'/>
<pointer-type-def type-id='8037c762' size-in-bits='64' id='d74a6869'/>
<qualified-type-def type-id='7292109c' restrict='yes' id='6942f6a4'/>
<qualified-type-def type-id='7347a39e' restrict='yes' id='578ba182'/>
<pointer-type-def type-id='62fab762' size-in-bits='64' id='db285b03'/>
<qualified-type-def type-id='db285b03' restrict='yes' id='2a468b41'/>
<qualified-type-def type-id='18c91f9e' restrict='yes' id='6e745582'/>
<pointer-type-def type-id='0897719a' size-in-bits='64' id='23cbcb08'/>
<qualified-type-def type-id='23cbcb08' restrict='yes' id='b09b2050'/>
<pointer-type-def type-id='6fcda10e' size-in-bits='64' id='ad33e5e7'/>
<pointer-type-def type-id='66a0afc9' size-in-bits='64' id='f32b30e4'/>
<qualified-type-def type-id='63e171df' restrict='yes' id='9e7a3a7d'/>
<function-decl name='pthread_self' visibility='default' binding='global' size-in-bits='64'>
<return type-id='4051f5e7'/>
</function-decl>
<function-decl name='pthread_attr_init' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_destroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getdetachstate' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='540db505'/>
<parameter type-id='7292109c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setdetachstate' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getguardsize' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='540db505'/>
<parameter type-id='78c01427'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setguardsize' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getschedparam' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e1815e87'/>
<parameter type-id='b09b2050'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setschedparam' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='578ba182'/>
<parameter type-id='37e4897b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getschedpolicy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e1815e87'/>
<parameter type-id='6942f6a4'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setschedpolicy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getinheritsched' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e1815e87'/>
<parameter type-id='6942f6a4'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setinheritsched' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getscope' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e1815e87'/>
<parameter type-id='6942f6a4'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setscope' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getstack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e1815e87'/>
<parameter type-id='9e7a3a7d'/>
<parameter type-id='d19b2c25'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setstack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setaffinity_np' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='5e14fa48'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getaffinity_np' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='540db505'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='d74a6869'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_setcancelstate' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='7292109c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_setcanceltype' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='7292109c'/>
<return type-id='95e97e5e'/>
</function-decl>
- <function-decl name='__pthread_register_cancel' visibility='default' binding='global' size-in-bits='64'>
- <parameter type-id='ba7c727c'/>
- <return type-id='48b5725f'/>
- </function-decl>
<function-decl name='__pthread_unregister_cancel' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='ba7c727c'/>
<return type-id='48b5725f'/>
</function-decl>
- <function-decl name='__pthread_unwind_next' visibility='default' binding='global' size-in-bits='64'>
- <parameter type-id='ba7c727c'/>
- <return type-id='48b5725f'/>
- </function-decl>
<function-decl name='pthread_cond_init' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='2a468b41'/>
<parameter type-id='4c428e67'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_cond_signal' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='db285b03'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_cond_broadcast' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='db285b03'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_cond_wait' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='2a468b41'/>
<parameter type-id='6e745582'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_cond_timedwait' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='2a468b41'/>
<parameter type-id='6e745582'/>
<parameter type-id='0be2e71c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__sysconf' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='bd54fe1a'/>
</function-decl>
- <function-decl name='pthread_sigmask' visibility='default' binding='global' size-in-bits='64'>
- <parameter type-id='95e97e5e'/>
- <parameter type-id='5a8729d0'/>
- <parameter type-id='65e6ec45'/>
- <return type-id='95e97e5e'/>
- </function-decl>
<function-decl name='tpool_abandon' mangled-name='tpool_abandon' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_abandon'>
<parameter type-id='9cf59a50' name='tpool'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='tpool_suspend' mangled-name='tpool_suspend' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_suspend'>
<parameter type-id='9cf59a50' name='tpool'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='tpool_suspended' mangled-name='tpool_suspended' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_suspended'>
<parameter type-id='9cf59a50' name='tpool'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='tpool_resume' mangled-name='tpool_resume' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_resume'>
<parameter type-id='9cf59a50' name='tpool'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='tpool_member' mangled-name='tpool_member' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_member'>
<parameter type-id='9cf59a50' name='tpool'/>
<return type-id='95e97e5e'/>
</function-decl>
<pointer-type-def type-id='b1bbf10d' size-in-bits='64' id='9cf59a50'/>
<typedef-decl name='tpool_t' type-id='88d1b7f9' id='b1bbf10d'/>
<class-decl name='tpool' size-in-bits='2496' is-struct='yes' visibility='default' id='88d1b7f9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tp_forw' type-id='9cf59a50' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tp_back' type-id='9cf59a50' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='tp_mutex' type-id='7a6844eb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='tp_busycv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='tp_workcv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1216'>
<var-decl name='tp_waitcv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1600'>
<var-decl name='tp_active' type-id='ad33e5e7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1664'>
<var-decl name='tp_head' type-id='f32b30e4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1728'>
<var-decl name='tp_tail' type-id='f32b30e4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1792'>
<var-decl name='tp_attr' type-id='7d8569fd' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2240'>
<var-decl name='tp_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2272'>
<var-decl name='tp_linger' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2304'>
<var-decl name='tp_njobs' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2336'>
<var-decl name='tp_minimum' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2368'>
<var-decl name='tp_maximum' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2400'>
<var-decl name='tp_current' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2432'>
<var-decl name='tp_idle' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<function-type size-in-bits='64' id='c5c76c9c'>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_changelist.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='bf311473' size-in-bits='128' id='f0f65199'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
<type-decl name='char' size-in-bits='8' id='a84c031d'/>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='8192' id='b54ce520'>
<subrange length='1024' type-id='7359adad' id='c60446f8'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='2048' id='d1617432'>
<subrange length='256' type-id='7359adad' id='36e5b9fa'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='320' id='36c46961'>
<subrange length='40' type-id='7359adad' id='8f80b239'/>
</array-type-def>
<class-decl name='re_dfa_t' is-struct='yes' visibility='default' is-declaration-only='yes' id='b48d2441'/>
<class-decl name='uu_avl' is-struct='yes' visibility='default' is-declaration-only='yes' id='4af029d1'/>
<class-decl name='uu_avl_pool' is-struct='yes' visibility='default' is-declaration-only='yes' id='12a530a8'/>
<class-decl name='uu_avl_walk' is-struct='yes' visibility='default' is-declaration-only='yes' id='e70a39e3'/>
<type-decl name='int' size-in-bits='32' id='95e97e5e'/>
<type-decl name='long int' size-in-bits='64' id='bd54fe1a'/>
<type-decl name='long long int' size-in-bits='64' id='1eb56b1e'/>
<type-decl name='short int' size-in-bits='16' id='a2185560'/>
<array-type-def dimensions='1' type-id='e475ab95' size-in-bits='192' id='0ce65a8b'>
<subrange length='3' type-id='7359adad' id='56f209d2'/>
</array-type-def>
<type-decl name='unnamed-enum-underlying-type-32' is-anonymous='yes' size-in-bits='32' alignment-in-bits='32' id='9cac1fee'/>
<type-decl name='unsigned char' size-in-bits='8' id='002ac4a6'/>
<type-decl name='unsigned int' size-in-bits='32' id='f0981eeb'/>
<type-decl name='unsigned long int' size-in-bits='64' id='7359adad'/>
<type-decl name='void' id='48b5725f'/>
<typedef-decl name='uu_compare_fn_t' type-id='add6e811' id='40f93560'/>
<typedef-decl name='uu_avl_pool_t' type-id='12a530a8' id='7f84e390'/>
<typedef-decl name='uu_avl_t' type-id='4af029d1' id='bb7f0973'/>
<class-decl name='uu_avl_node' size-in-bits='192' is-struct='yes' visibility='default' id='f65f4326'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='uan_opaque' type-id='0ce65a8b' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='uu_avl_node_t' type-id='f65f4326' id='73a65116'/>
<typedef-decl name='uu_avl_walk_t' type-id='e70a39e3' id='edd8457b'/>
<typedef-decl name='uu_avl_index_t' type-id='e475ab95' id='5d7f5fc8'/>
<typedef-decl name='zfs_handle_t' type-id='f6ee4445' id='775509eb'/>
<typedef-decl name='zpool_handle_t' type-id='67002a8a' id='b1efc708'/>
<typedef-decl name='libzfs_handle_t' type-id='c8a9d9d8' id='95942d0c'/>
<typedef-decl name='zfs_iter_f' type-id='5571cde4' id='d8e49ab9'/>
<typedef-decl name='avl_tree_t' type-id='b351119f' id='f20fbd51'/>
<class-decl name='avl_node' size-in-bits='192' is-struct='yes' visibility='default' id='428b67b3'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='avl_child' type-id='f0f65199' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='avl_pcb' type-id='e475ab95' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='avl_tree' size-in-bits='320' is-struct='yes' visibility='default' id='b351119f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='avl_root' type-id='bf311473' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='avl_compar' type-id='585e1de9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='avl_offset' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='avl_numnodes' type-id='ee1f298e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='avl_pad' type-id='b59d7dce' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='dmu_objset_stats' size-in-bits='2304' is-struct='yes' visibility='default' id='098f0221'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='dds_num_clones' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='dds_creation_txg' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='dds_guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='dds_type' type-id='230f1e16' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='dds_is_snapshot' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='232'>
<var-decl name='dds_inconsistent' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='240'>
<var-decl name='dds_redacted' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='248'>
<var-decl name='dds_origin' type-id='d1617432' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='dmu_objset_stats_t' type-id='098f0221' id='b2c14f17'/>
<enum-decl name='zfs_type_t' naming-typedef-id='2e45de5d' id='5d8f7321'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_TYPE_INVALID' value='0'/>
<enumerator name='ZFS_TYPE_FILESYSTEM' value='1'/>
<enumerator name='ZFS_TYPE_SNAPSHOT' value='2'/>
<enumerator name='ZFS_TYPE_VOLUME' value='4'/>
<enumerator name='ZFS_TYPE_POOL' value='8'/>
<enumerator name='ZFS_TYPE_BOOKMARK' value='16'/>
<enumerator name='ZFS_TYPE_VDEV' value='32'/>
</enum-decl>
<typedef-decl name='zfs_type_t' type-id='5d8f7321' id='2e45de5d'/>
<enum-decl name='dmu_objset_type' id='6b1b19f9'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DMU_OST_NONE' value='0'/>
<enumerator name='DMU_OST_META' value='1'/>
<enumerator name='DMU_OST_ZFS' value='2'/>
<enumerator name='DMU_OST_ZVOL' value='3'/>
<enumerator name='DMU_OST_OTHER' value='4'/>
<enumerator name='DMU_OST_ANY' value='5'/>
<enumerator name='DMU_OST_NUMTYPES' value='6'/>
</enum-decl>
<typedef-decl name='dmu_objset_type_t' type-id='6b1b19f9' id='230f1e16'/>
<enum-decl name='zfs_prop_t' naming-typedef-id='58603c44' id='4b000d60'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPROP_CONT' value='-2'/>
<enumerator name='ZPROP_INVAL' value='-1'/>
<enumerator name='ZPROP_USERPROP' value='-1'/>
<enumerator name='ZFS_PROP_TYPE' value='0'/>
<enumerator name='ZFS_PROP_CREATION' value='1'/>
<enumerator name='ZFS_PROP_USED' value='2'/>
<enumerator name='ZFS_PROP_AVAILABLE' value='3'/>
<enumerator name='ZFS_PROP_REFERENCED' value='4'/>
<enumerator name='ZFS_PROP_COMPRESSRATIO' value='5'/>
<enumerator name='ZFS_PROP_MOUNTED' value='6'/>
<enumerator name='ZFS_PROP_ORIGIN' value='7'/>
<enumerator name='ZFS_PROP_QUOTA' value='8'/>
<enumerator name='ZFS_PROP_RESERVATION' value='9'/>
<enumerator name='ZFS_PROP_VOLSIZE' value='10'/>
<enumerator name='ZFS_PROP_VOLBLOCKSIZE' value='11'/>
<enumerator name='ZFS_PROP_RECORDSIZE' value='12'/>
<enumerator name='ZFS_PROP_MOUNTPOINT' value='13'/>
<enumerator name='ZFS_PROP_SHARENFS' value='14'/>
<enumerator name='ZFS_PROP_CHECKSUM' value='15'/>
<enumerator name='ZFS_PROP_COMPRESSION' value='16'/>
<enumerator name='ZFS_PROP_ATIME' value='17'/>
<enumerator name='ZFS_PROP_DEVICES' value='18'/>
<enumerator name='ZFS_PROP_EXEC' value='19'/>
<enumerator name='ZFS_PROP_SETUID' value='20'/>
<enumerator name='ZFS_PROP_READONLY' value='21'/>
<enumerator name='ZFS_PROP_ZONED' value='22'/>
<enumerator name='ZFS_PROP_SNAPDIR' value='23'/>
<enumerator name='ZFS_PROP_ACLMODE' value='24'/>
<enumerator name='ZFS_PROP_ACLINHERIT' value='25'/>
<enumerator name='ZFS_PROP_CREATETXG' value='26'/>
<enumerator name='ZFS_PROP_NAME' value='27'/>
<enumerator name='ZFS_PROP_CANMOUNT' value='28'/>
<enumerator name='ZFS_PROP_ISCSIOPTIONS' value='29'/>
<enumerator name='ZFS_PROP_XATTR' value='30'/>
<enumerator name='ZFS_PROP_NUMCLONES' value='31'/>
<enumerator name='ZFS_PROP_COPIES' value='32'/>
<enumerator name='ZFS_PROP_VERSION' value='33'/>
<enumerator name='ZFS_PROP_UTF8ONLY' value='34'/>
<enumerator name='ZFS_PROP_NORMALIZE' value='35'/>
<enumerator name='ZFS_PROP_CASE' value='36'/>
<enumerator name='ZFS_PROP_VSCAN' value='37'/>
<enumerator name='ZFS_PROP_NBMAND' value='38'/>
<enumerator name='ZFS_PROP_SHARESMB' value='39'/>
<enumerator name='ZFS_PROP_REFQUOTA' value='40'/>
<enumerator name='ZFS_PROP_REFRESERVATION' value='41'/>
<enumerator name='ZFS_PROP_GUID' value='42'/>
<enumerator name='ZFS_PROP_PRIMARYCACHE' value='43'/>
<enumerator name='ZFS_PROP_SECONDARYCACHE' value='44'/>
<enumerator name='ZFS_PROP_USEDSNAP' value='45'/>
<enumerator name='ZFS_PROP_USEDDS' value='46'/>
<enumerator name='ZFS_PROP_USEDCHILD' value='47'/>
<enumerator name='ZFS_PROP_USEDREFRESERV' value='48'/>
<enumerator name='ZFS_PROP_USERACCOUNTING' value='49'/>
<enumerator name='ZFS_PROP_STMF_SHAREINFO' value='50'/>
<enumerator name='ZFS_PROP_DEFER_DESTROY' value='51'/>
<enumerator name='ZFS_PROP_USERREFS' value='52'/>
<enumerator name='ZFS_PROP_LOGBIAS' value='53'/>
<enumerator name='ZFS_PROP_UNIQUE' value='54'/>
<enumerator name='ZFS_PROP_OBJSETID' value='55'/>
<enumerator name='ZFS_PROP_DEDUP' value='56'/>
<enumerator name='ZFS_PROP_MLSLABEL' value='57'/>
<enumerator name='ZFS_PROP_SYNC' value='58'/>
<enumerator name='ZFS_PROP_DNODESIZE' value='59'/>
<enumerator name='ZFS_PROP_REFRATIO' value='60'/>
<enumerator name='ZFS_PROP_WRITTEN' value='61'/>
<enumerator name='ZFS_PROP_CLONES' value='62'/>
<enumerator name='ZFS_PROP_LOGICALUSED' value='63'/>
<enumerator name='ZFS_PROP_LOGICALREFERENCED' value='64'/>
<enumerator name='ZFS_PROP_INCONSISTENT' value='65'/>
<enumerator name='ZFS_PROP_VOLMODE' value='66'/>
<enumerator name='ZFS_PROP_FILESYSTEM_LIMIT' value='67'/>
<enumerator name='ZFS_PROP_SNAPSHOT_LIMIT' value='68'/>
<enumerator name='ZFS_PROP_FILESYSTEM_COUNT' value='69'/>
<enumerator name='ZFS_PROP_SNAPSHOT_COUNT' value='70'/>
<enumerator name='ZFS_PROP_SNAPDEV' value='71'/>
<enumerator name='ZFS_PROP_ACLTYPE' value='72'/>
<enumerator name='ZFS_PROP_SELINUX_CONTEXT' value='73'/>
<enumerator name='ZFS_PROP_SELINUX_FSCONTEXT' value='74'/>
<enumerator name='ZFS_PROP_SELINUX_DEFCONTEXT' value='75'/>
<enumerator name='ZFS_PROP_SELINUX_ROOTCONTEXT' value='76'/>
<enumerator name='ZFS_PROP_RELATIME' value='77'/>
<enumerator name='ZFS_PROP_REDUNDANT_METADATA' value='78'/>
<enumerator name='ZFS_PROP_OVERLAY' value='79'/>
<enumerator name='ZFS_PROP_PREV_SNAP' value='80'/>
<enumerator name='ZFS_PROP_RECEIVE_RESUME_TOKEN' value='81'/>
<enumerator name='ZFS_PROP_ENCRYPTION' value='82'/>
<enumerator name='ZFS_PROP_KEYLOCATION' value='83'/>
<enumerator name='ZFS_PROP_KEYFORMAT' value='84'/>
<enumerator name='ZFS_PROP_PBKDF2_SALT' value='85'/>
<enumerator name='ZFS_PROP_PBKDF2_ITERS' value='86'/>
<enumerator name='ZFS_PROP_ENCRYPTION_ROOT' value='87'/>
<enumerator name='ZFS_PROP_KEY_GUID' value='88'/>
<enumerator name='ZFS_PROP_KEYSTATUS' value='89'/>
<enumerator name='ZFS_PROP_REMAPTXG' value='90'/>
<enumerator name='ZFS_PROP_SPECIAL_SMALL_BLOCKS' value='91'/>
<enumerator name='ZFS_PROP_IVSET_GUID' value='92'/>
<enumerator name='ZFS_PROP_REDACTED' value='93'/>
<enumerator name='ZFS_PROP_REDACT_SNAPS' value='94'/>
<enumerator name='ZFS_PROP_SNAPSHOTS_CHANGED' value='95'/>
<enumerator name='ZFS_PROP_PREFETCH' value='96'/>
<enumerator name='ZFS_PROP_VOLTHREADING' value='97'/>
<enumerator name='ZFS_NUM_PROPS' value='98'/>
</enum-decl>
<typedef-decl name='zfs_prop_t' type-id='4b000d60' id='58603c44'/>
<enum-decl name='zprop_source_t' naming-typedef-id='a2256d42' id='5903f80e'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPROP_SRC_NONE' value='1'/>
<enumerator name='ZPROP_SRC_DEFAULT' value='2'/>
<enumerator name='ZPROP_SRC_TEMPORARY' value='4'/>
<enumerator name='ZPROP_SRC_LOCAL' value='8'/>
<enumerator name='ZPROP_SRC_INHERITED' value='16'/>
<enumerator name='ZPROP_SRC_RECEIVED' value='32'/>
</enum-decl>
<typedef-decl name='zprop_source_t' type-id='5903f80e' id='a2256d42'/>
<class-decl name='nvlist' size-in-bits='192' is-struct='yes' visibility='default' id='ac266fd9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='nvl_version' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='nvl_nvflag' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='nvl_priv' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='nvl_flag' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='nvl_pad' type-id='3ff5601b' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='nvlist_t' type-id='ac266fd9' id='8e8d4be3'/>
<enum-decl name='sa_protocol' id='9155d4b5'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='SA_PROTOCOL_NFS' value='0'/>
<enumerator name='SA_PROTOCOL_SMB' value='1'/>
<enumerator name='SA_PROTOCOL_COUNT' value='2'/>
</enum-decl>
<enum-decl name='boolean_t' naming-typedef-id='c19b74c3' id='f58c8277'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='B_FALSE' value='0'/>
<enumerator name='B_TRUE' value='1'/>
</enum-decl>
<typedef-decl name='boolean_t' type-id='f58c8277' id='c19b74c3'/>
<typedef-decl name='uint_t' type-id='f0981eeb' id='3502e3ff'/>
<typedef-decl name='ulong_t' type-id='7359adad' id='ee1f298e'/>
<typedef-decl name='longlong_t' type-id='1eb56b1e' id='9b3ff54f'/>
<typedef-decl name='diskaddr_t' type-id='9b3ff54f' id='804dc465'/>
<typedef-decl name='zoneid_t' type-id='3502e3ff' id='4da03624'/>
<typedef-decl name='__re_long_size_t' type-id='7359adad' id='ba516949'/>
<typedef-decl name='reg_syntax_t' type-id='7359adad' id='1b72c3b3'/>
<class-decl name='re_pattern_buffer' size-in-bits='512' is-struct='yes' visibility='default' id='19fc9a8c'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='buffer' type-id='33976309' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='allocated' type-id='ba516949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='used' type-id='ba516949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='syntax' type-id='1b72c3b3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='fastmap' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='translate' type-id='cf536864' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='re_nsub' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='can_be_null' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='449'>
<var-decl name='regs_allocated' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='451'>
<var-decl name='fastmap_accurate' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='452'>
<var-decl name='no_sub' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='453'>
<var-decl name='not_bol' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='454'>
<var-decl name='not_eol' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='455'>
<var-decl name='newline_anchor' type-id='f0981eeb' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='regex_t' type-id='19fc9a8c' id='aca3bac8'/>
<typedef-decl name='uintptr_t' type-id='7359adad' id='e475ab95'/>
<union-decl name='pthread_mutex_t' size-in-bits='320' naming-typedef-id='7a6844eb' visibility='default' id='70681f9b'>
<data-member access='public'>
<var-decl name='__data' type-id='4c734837' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__size' type-id='36c46961' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='bd54fe1a' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_mutex_t' type-id='70681f9b' id='7a6844eb'/>
<typedef-decl name='int32_t' type-id='33f57a65' id='3ff5601b'/>
<typedef-decl name='uint8_t' type-id='c51d6389' id='b96825af'/>
<typedef-decl name='uint32_t' type-id='62f1140c' id='8f92235e'/>
<typedef-decl name='uint64_t' type-id='8910171f' id='9c313c2d'/>
<class-decl name='__pthread_mutex_s' size-in-bits='320' is-struct='yes' visibility='default' id='4c734837'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__lock' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='__count' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='__owner' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='__nusers' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='__kind' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='__spins' type-id='a2185560' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='176'>
<var-decl name='__elision' type-id='a2185560' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='__list' type-id='518fb49c' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__pthread_internal_list' size-in-bits='128' is-struct='yes' visibility='default' id='0e01899c'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__prev' type-id='4d98cd5a' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='__next' type-id='4d98cd5a' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__pthread_list_t' type-id='0e01899c' id='518fb49c'/>
<typedef-decl name='__uint8_t' type-id='002ac4a6' id='c51d6389'/>
<typedef-decl name='__int32_t' type-id='95e97e5e' id='33f57a65'/>
<typedef-decl name='__uint32_t' type-id='f0981eeb' id='62f1140c'/>
<typedef-decl name='__uint64_t' type-id='7359adad' id='8910171f'/>
<typedef-decl name='size_t' type-id='7359adad' id='b59d7dce'/>
<class-decl name='libzfs_handle' size-in-bits='18240' is-struct='yes' visibility='default' id='c8a9d9d8'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='libzfs_error' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='libzfs_fd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='libzfs_pool_handles' type-id='4c81de99' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='libzfs_ns_avlpool' type-id='de82c773' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='libzfs_ns_avl' type-id='a5c21a38' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='libzfs_ns_gen' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='libzfs_desc_active' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='libzfs_action' type-id='b54ce520' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8544'>
<var-decl name='libzfs_desc' type-id='b54ce520' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='16736'>
<var-decl name='libzfs_printerr' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='16768'>
<var-decl name='libzfs_mnttab_enable' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='16832'>
<var-decl name='libzfs_mnttab_cache_lock' type-id='7a6844eb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='17152'>
<var-decl name='libzfs_mnttab_cache' type-id='f20fbd51' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='17472'>
<var-decl name='libzfs_pool_iter' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='17504'>
<var-decl name='libzfs_prop_debug' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='17536'>
<var-decl name='libzfs_urire' type-id='aca3bac8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='18048'>
<var-decl name='libzfs_max_nvlist' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='18112'>
<var-decl name='libfetch' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='18176'>
<var-decl name='libfetch_load_error' type-id='26a90f95' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='zfs_handle' size-in-bits='4928' is-struct='yes' visibility='default' id='f6ee4445'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zfs_hdl' type-id='b0382bb3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zpool_hdl' type-id='4c81de99' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zfs_name' type-id='d1617432' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2176'>
<var-decl name='zfs_type' type-id='2e45de5d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2208'>
<var-decl name='zfs_head_type' type-id='2e45de5d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2240'>
<var-decl name='zfs_dmustats' type-id='b2c14f17' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4544'>
<var-decl name='zfs_props' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4608'>
<var-decl name='zfs_user_props' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4672'>
<var-decl name='zfs_recvd_props' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4736'>
<var-decl name='zfs_mntcheck' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4800'>
<var-decl name='zfs_mntopts' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4864'>
<var-decl name='zfs_props_table' type-id='ae3e8ca6' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='zpool_handle' size-in-bits='2560' is-struct='yes' visibility='default' id='67002a8a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zpool_hdl' type-id='b0382bb3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zpool_next' type-id='4c81de99' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zpool_name' type-id='d1617432' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2176'>
<var-decl name='zpool_state' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2240'>
<var-decl name='zpool_config_size' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2304'>
<var-decl name='zpool_config' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2368'>
<var-decl name='zpool_old_config' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2432'>
<var-decl name='zpool_props' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2496'>
<var-decl name='zpool_start_block' type-id='804dc465' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='0e01899c' size-in-bits='64' id='4d98cd5a'/>
<pointer-type-def type-id='428b67b3' size-in-bits='64' id='bf311473'/>
<pointer-type-def type-id='a84c031d' size-in-bits='64' id='26a90f95'/>
<pointer-type-def type-id='26a90f95' size-in-bits='64' id='9b23c9ad'/>
<qualified-type-def type-id='a84c031d' const='yes' id='9b45d938'/>
<pointer-type-def type-id='9b45d938' size-in-bits='64' id='80f4b756'/>
<qualified-type-def type-id='9155d4b5' const='yes' id='9f2c1699'/>
<pointer-type-def type-id='9f2c1699' size-in-bits='64' id='4567bbc9'/>
<qualified-type-def type-id='775509eb' const='yes' id='5eadf2db'/>
<pointer-type-def type-id='5eadf2db' size-in-bits='64' id='fcd57163'/>
<pointer-type-def type-id='96ee24a5' size-in-bits='64' id='585e1de9'/>
<pointer-type-def type-id='cb9628fa' size-in-bits='64' id='5571cde4'/>
<pointer-type-def type-id='95942d0c' size-in-bits='64' id='b0382bb3'/>
<pointer-type-def type-id='8e8d4be3' size-in-bits='64' id='5ce45b60'/>
<pointer-type-def type-id='b48d2441' size-in-bits='64' id='33976309'/>
<pointer-type-def type-id='b96825af' size-in-bits='64' id='ae3e8ca6'/>
<pointer-type-def type-id='002ac4a6' size-in-bits='64' id='cf536864'/>
<pointer-type-def type-id='5d7f5fc8' size-in-bits='64' id='813a2225'/>
<pointer-type-def type-id='73a65116' size-in-bits='64' id='2dc35b9d'/>
<pointer-type-def type-id='7f84e390' size-in-bits='64' id='de82c773'/>
<pointer-type-def type-id='bb7f0973' size-in-bits='64' id='a5c21a38'/>
<pointer-type-def type-id='edd8457b' size-in-bits='64' id='5842d146'/>
<pointer-type-def type-id='40f93560' size-in-bits='64' id='d502b39f'/>
<pointer-type-def type-id='48b5725f' size-in-bits='64' id='eaa32e2f'/>
<pointer-type-def type-id='775509eb' size-in-bits='64' id='9200a744'/>
<pointer-type-def type-id='b1efc708' size-in-bits='64' id='4c81de99'/>
<pointer-type-def type-id='a2256d42' size-in-bits='64' id='debc6aa3'/>
<class-decl name='re_dfa_t' is-struct='yes' visibility='default' is-declaration-only='yes' id='b48d2441'/>
<class-decl name='uu_avl' is-struct='yes' visibility='default' is-declaration-only='yes' id='4af029d1'/>
<class-decl name='uu_avl_pool' is-struct='yes' visibility='default' is-declaration-only='yes' id='12a530a8'/>
<class-decl name='uu_avl_walk' is-struct='yes' visibility='default' is-declaration-only='yes' id='e70a39e3'/>
<function-decl name='uu_avl_pool_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='d502b39f'/>
<parameter type-id='8f92235e'/>
<return type-id='de82c773'/>
</function-decl>
<function-decl name='uu_avl_pool_destroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='de82c773'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uu_avl_node_init' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='2dc35b9d'/>
<parameter type-id='de82c773'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uu_avl_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='de82c773'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='8f92235e'/>
<return type-id='a5c21a38'/>
</function-decl>
<function-decl name='uu_avl_destroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uu_avl_last' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='uu_avl_walk_start' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='8f92235e'/>
<return type-id='5842d146'/>
</function-decl>
<function-decl name='uu_avl_walk_next' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5842d146'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='uu_avl_walk_end' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5842d146'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uu_avl_find' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='813a2225'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='uu_avl_insert' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='5d7f5fc8'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uu_avl_remove' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_get_handle' mangled-name='zfs_get_handle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_handle'>
<parameter type-id='9200a744'/>
<return type-id='b0382bb3'/>
</function-decl>
<function-decl name='zfs_open' mangled-name='zfs_open' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_open'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='zfs_close' mangled-name='zfs_close' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_close'>
<parameter type-id='9200a744'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_get_name' mangled-name='zfs_get_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_name'>
<parameter type-id='fcd57163'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_prop_get' mangled-name='zfs_prop_get' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get'>
<parameter type-id='9200a744'/>
<parameter type-id='58603c44'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='debc6aa3'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='c19b74c3'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_int' mangled-name='zfs_prop_get_int' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_int'>
<parameter type-id='9200a744'/>
<parameter type-id='58603c44'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zfs_iter_children_v2' mangled-name='zfs_iter_children_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_children_v2'>
<parameter type-id='9200a744'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_dependents_v2' mangled-name='zfs_iter_dependents_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_dependents_v2'>
<parameter type-id='9200a744'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_mounted' mangled-name='zfs_iter_mounted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_mounted'>
<parameter type-id='9200a744'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_refresh_properties' mangled-name='zfs_refresh_properties' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_refresh_properties'>
<parameter type-id='9200a744'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_is_mounted' mangled-name='zfs_is_mounted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_is_mounted'>
<parameter type-id='9200a744'/>
<parameter type-id='9b23c9ad'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_mount' mangled-name='zfs_mount' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_mount'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_unmount' mangled-name='zfs_unmount' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_unmount'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_is_shared' mangled-name='zfs_is_shared' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_is_shared'>
<parameter type-id='9200a744'/>
<parameter type-id='9b23c9ad'/>
<parameter type-id='4567bbc9'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_share' mangled-name='zfs_share' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_share'>
<parameter type-id='9200a744'/>
<parameter type-id='4567bbc9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_unshare' mangled-name='zfs_unshare' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_unshare'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='4567bbc9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_commit_shares' mangled-name='zfs_commit_shares' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_commit_shares'>
<parameter type-id='4567bbc9'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='getzoneid' mangled-name='getzoneid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getzoneid'>
<return type-id='4da03624'/>
</function-decl>
<function-decl name='sa_commit_shares' mangled-name='sa_commit_shares' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_commit_shares'>
<parameter type-id='9155d4b5'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strlcat' mangled-name='strlcat' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='strlcat'>
<parameter type-id='26a90f95'/>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='strlcpy' mangled-name='strlcpy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='strlcpy'>
<parameter type-id='26a90f95'/>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strcmp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='strncmp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='strlen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='zfs_error' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_alloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='remove_mountpoint' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<return type-id='48b5725f'/>
</function-decl>
<function-type size-in-bits='64' id='96ee24a5'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='add6e811'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='cb9628fa'>
<parameter type-id='9200a744'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_config.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='32768' id='d16c6df4'>
<subrange length='4096' type-id='7359adad' id='bc1b5ddc'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='65536' id='163f6aa5'>
<subrange length='8192' type-id='7359adad' id='c88f397d'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='infinite' id='e84913bd'>
<subrange length='infinite' type-id='7359adad' id='031f2035'/>
</array-type-def>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='128' id='c1c22e6c'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='24' id='d3490169'>
<subrange length='3' type-id='7359adad' id='56f209d2'/>
</array-type-def>
<type-decl name='variadic parameter type' id='2c1145c5'/>
<typedef-decl name='zpool_iter_f' type-id='3aebb66f' id='fa476e62'/>
<enum-decl name='data_type_t' naming-typedef-id='8d0687d2' id='aeeae136'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DATA_TYPE_DONTCARE' value='-1'/>
<enumerator name='DATA_TYPE_UNKNOWN' value='0'/>
<enumerator name='DATA_TYPE_BOOLEAN' value='1'/>
<enumerator name='DATA_TYPE_BYTE' value='2'/>
<enumerator name='DATA_TYPE_INT16' value='3'/>
<enumerator name='DATA_TYPE_UINT16' value='4'/>
<enumerator name='DATA_TYPE_INT32' value='5'/>
<enumerator name='DATA_TYPE_UINT32' value='6'/>
<enumerator name='DATA_TYPE_INT64' value='7'/>
<enumerator name='DATA_TYPE_UINT64' value='8'/>
<enumerator name='DATA_TYPE_STRING' value='9'/>
<enumerator name='DATA_TYPE_BYTE_ARRAY' value='10'/>
<enumerator name='DATA_TYPE_INT16_ARRAY' value='11'/>
<enumerator name='DATA_TYPE_UINT16_ARRAY' value='12'/>
<enumerator name='DATA_TYPE_INT32_ARRAY' value='13'/>
<enumerator name='DATA_TYPE_UINT32_ARRAY' value='14'/>
<enumerator name='DATA_TYPE_INT64_ARRAY' value='15'/>
<enumerator name='DATA_TYPE_UINT64_ARRAY' value='16'/>
<enumerator name='DATA_TYPE_STRING_ARRAY' value='17'/>
<enumerator name='DATA_TYPE_HRTIME' value='18'/>
<enumerator name='DATA_TYPE_NVLIST' value='19'/>
<enumerator name='DATA_TYPE_NVLIST_ARRAY' value='20'/>
<enumerator name='DATA_TYPE_BOOLEAN_VALUE' value='21'/>
<enumerator name='DATA_TYPE_INT8' value='22'/>
<enumerator name='DATA_TYPE_UINT8' value='23'/>
<enumerator name='DATA_TYPE_BOOLEAN_ARRAY' value='24'/>
<enumerator name='DATA_TYPE_INT8_ARRAY' value='25'/>
<enumerator name='DATA_TYPE_UINT8_ARRAY' value='26'/>
<enumerator name='DATA_TYPE_DOUBLE' value='27'/>
</enum-decl>
<typedef-decl name='data_type_t' type-id='aeeae136' id='8d0687d2'/>
<class-decl name='nvpair' size-in-bits='128' is-struct='yes' visibility='default' id='1c34e459'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='nvp_size' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='nvp_name_sz' type-id='23bd8cb5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='48'>
<var-decl name='nvp_reserve' type-id='23bd8cb5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='nvp_value_elem' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='nvp_type' type-id='8d0687d2' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='nvp_name' type-id='e84913bd' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='nvpair_t' type-id='1c34e459' id='57928edf'/>
<class-decl name='drr_begin' size-in-bits='2432' is-struct='yes' visibility='default' id='09fcdc01'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_magic' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_versioninfo' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_creation_time' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_type' type-id='230f1e16' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='drr_flags' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_fromguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_toname' type-id='d1617432' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='zinject_record' size-in-bits='2816' is-struct='yes' visibility='default' id='3216f820'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zi_objset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zi_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zi_start' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='zi_end' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='zi_guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='zi_level' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='zi_error' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='zi_type' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='zi_freq' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='zi_failfast' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='zi_func' type-id='d1617432' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2560'>
<var-decl name='zi_iotype' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2592'>
<var-decl name='zi_duration' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2624'>
<var-decl name='zi_timer' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2688'>
<var-decl name='zi_nlanes' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2752'>
<var-decl name='zi_cmd' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2784'>
<var-decl name='zi_dvas' type-id='8f92235e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zinject_record_t' type-id='3216f820' id='a4301ca6'/>
<class-decl name='zfs_share' size-in-bits='256' is-struct='yes' visibility='default' id='feb6f2da'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='z_exportdata' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='z_sharedata' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='z_sharetype' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='z_sharemax' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_share_t' type-id='feb6f2da' id='ee5cec36'/>
<class-decl name='zfs_cmd' size-in-bits='109952' is-struct='yes' visibility='default' id='3522cd69'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zc_name' type-id='d16c6df4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32768'>
<var-decl name='zc_nvlist_src' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32832'>
<var-decl name='zc_nvlist_src_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32896'>
<var-decl name='zc_nvlist_dst' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32960'>
<var-decl name='zc_nvlist_dst_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33024'>
<var-decl name='zc_nvlist_dst_filled' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33056'>
<var-decl name='zc_pad2' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33088'>
<var-decl name='zc_history' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33152'>
<var-decl name='zc_value' type-id='163f6aa5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='98688'>
<var-decl name='zc_string' type-id='d1617432' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100736'>
<var-decl name='zc_guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100800'>
<var-decl name='zc_nvlist_conf' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100864'>
<var-decl name='zc_nvlist_conf_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100928'>
<var-decl name='zc_cookie' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100992'>
<var-decl name='zc_objset_type' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101056'>
<var-decl name='zc_perm_action' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101120'>
<var-decl name='zc_history_len' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101184'>
<var-decl name='zc_history_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101248'>
<var-decl name='zc_obj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101312'>
<var-decl name='zc_iflags' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101376'>
<var-decl name='zc_share' type-id='ee5cec36' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101632'>
<var-decl name='zc_objset_stats' type-id='b2c14f17' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='103936'>
<var-decl name='zc_begin_record' type-id='09fcdc01' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='106368'>
<var-decl name='zc_inject_record' type-id='a4301ca6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109184'>
<var-decl name='zc_defer_destroy' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109216'>
<var-decl name='zc_flags' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109248'>
<var-decl name='zc_action_handle' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109312'>
<var-decl name='zc_cleanup_fd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109344'>
<var-decl name='zc_simple' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109352'>
<var-decl name='zc_pad' type-id='d3490169' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109376'>
<var-decl name='zc_sendobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109440'>
<var-decl name='zc_fromobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109504'>
<var-decl name='zc_createtxg' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109568'>
<var-decl name='zc_stat' type-id='0371a9c7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109888'>
<var-decl name='zc_zoneid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_cmd_t' type-id='3522cd69' id='a5559cdd'/>
<class-decl name='zfs_stat' size-in-bits='320' is-struct='yes' visibility='default' id='6417f0b9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zs_gen' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zs_mode' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zs_links' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='zs_ctime' type-id='c1c22e6c' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_stat_t' type-id='6417f0b9' id='0371a9c7'/>
<typedef-decl name='int16_t' type-id='03896e23' id='23bd8cb5'/>
<typedef-decl name='__int16_t' type-id='a2185560' id='03896e23'/>
<pointer-type-def type-id='c19b74c3' size-in-bits='64' id='37e3bd22'/>
<qualified-type-def type-id='8e8d4be3' const='yes' id='693c3853'/>
<pointer-type-def type-id='693c3853' size-in-bits='64' id='22cce67b'/>
<qualified-type-def type-id='57928edf' const='yes' id='642ee20f'/>
<pointer-type-def type-id='642ee20f' size-in-bits='64' id='dace003f'/>
<pointer-type-def type-id='2bce87e3' size-in-bits='64' id='3aebb66f'/>
<pointer-type-def type-id='95e97e5e' size-in-bits='64' id='7292109c'/>
<pointer-type-def type-id='5ce45b60' size-in-bits='64' id='857bb57e'/>
<pointer-type-def type-id='57928edf' size-in-bits='64' id='3fa542f0'/>
<pointer-type-def type-id='eaa32e2f' size-in-bits='64' id='63e171df'/>
<pointer-type-def type-id='3522cd69' size-in-bits='64' id='b65f7fd1'/>
<pointer-type-def type-id='a5559cdd' size-in-bits='64' id='e4ec4540'/>
<pointer-type-def type-id='4c81de99' size-in-bits='64' id='237193c9'/>
<function-decl name='uu_avl_first' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='uu_avl_next' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='eaa32e2f'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='uu_avl_teardown' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='63e171df'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='zfs_standard_error' mangled-name='zfs_standard_error' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_standard_error'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_ioctl' mangled-name='zfs_ioctl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_ioctl'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='b65f7fd1'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='nvlist_dup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='857bb57e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_exists' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='nvlist_next_nvpair' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='dace003f'/>
<return type-id='3fa542f0'/>
</function-decl>
<function-decl name='nvpair_name' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='fnvpair_value_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='3fa542f0'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='libspl_assertf' mangled-name='libspl_assertf' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libspl_assertf'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__errno_location' visibility='default' binding='global' size-in-bits='64'>
<return type-id='7292109c'/>
</function-decl>
<function-decl name='dcgettext' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='getenv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strchr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zpool_get_config' mangled-name='zpool_get_config' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_config'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='857bb57e' name='oldconfig'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zpool_get_features' mangled-name='zpool_get_features' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_features'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zpool_refresh_stats' mangled-name='zpool_refresh_stats' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_refresh_stats'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='37e3bd22' name='missing'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_skip_pool' mangled-name='zpool_skip_pool' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_skip_pool'>
<parameter type-id='80f4b756' name='poolname'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_iter' mangled-name='zpool_iter' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_iter'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='fa476e62' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_root' mangled-name='zfs_iter_root' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_root'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_strdup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='no_memory' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zcmd_alloc_dst_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<parameter type-id='b59d7dce'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zcmd_expand_dst_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zcmd_read_dst_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zcmd_free_nvlists' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e4ec4540'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='make_dataset_handle' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='zpool_open_silent' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='237193c9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-type size-in-bits='64' id='2bce87e3'>
<parameter type-id='4c81de99'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_crypto.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='fb7c6451' size-in-bits='256' id='64177143'>
<subrange length='32' type-id='7359adad' id='ae5bde82'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='8' id='89feb1ec'>
<subrange length='1' type-id='7359adad' id='52f813b4'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='160' id='664ac0b7'>
<subrange length='20' type-id='7359adad' id='fdca39cf'/>
</array-type-def>
<class-decl name='_IO_codecvt' is-struct='yes' visibility='default' is-declaration-only='yes' id='a4036571'/>
<class-decl name='_IO_marker' is-struct='yes' visibility='default' is-declaration-only='yes' id='010ae0b9'/>
<class-decl name='_IO_wide_data' is-struct='yes' visibility='default' is-declaration-only='yes' id='79bd3751'/>
<array-type-def dimensions='1' type-id='95e97e5e' size-in-bits='896' id='47394ee0'>
<subrange length='28' type-id='7359adad' id='3db583d7'/>
</array-type-def>
<type-decl name='signed char' size-in-bits='8' id='28577a57'/>
<array-type-def dimensions='1' type-id='7359adad' size-in-bits='1024' id='d2baa450'>
<subrange length='16' type-id='7359adad' id='848d0938'/>
</array-type-def>
<type-decl name='unsigned short int' size-in-bits='16' id='8efea9e5'/>
<enum-decl name='zpool_prop_t' naming-typedef-id='5d0c23fb' id='af1ba157'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_PROP_INVAL' value='-1'/>
<enumerator name='ZPOOL_PROP_NAME' value='0'/>
<enumerator name='ZPOOL_PROP_SIZE' value='1'/>
<enumerator name='ZPOOL_PROP_CAPACITY' value='2'/>
<enumerator name='ZPOOL_PROP_ALTROOT' value='3'/>
<enumerator name='ZPOOL_PROP_HEALTH' value='4'/>
<enumerator name='ZPOOL_PROP_GUID' value='5'/>
<enumerator name='ZPOOL_PROP_VERSION' value='6'/>
<enumerator name='ZPOOL_PROP_BOOTFS' value='7'/>
<enumerator name='ZPOOL_PROP_DELEGATION' value='8'/>
<enumerator name='ZPOOL_PROP_AUTOREPLACE' value='9'/>
<enumerator name='ZPOOL_PROP_CACHEFILE' value='10'/>
<enumerator name='ZPOOL_PROP_FAILUREMODE' value='11'/>
<enumerator name='ZPOOL_PROP_LISTSNAPS' value='12'/>
<enumerator name='ZPOOL_PROP_AUTOEXPAND' value='13'/>
<enumerator name='ZPOOL_PROP_DEDUPDITTO' value='14'/>
<enumerator name='ZPOOL_PROP_DEDUPRATIO' value='15'/>
<enumerator name='ZPOOL_PROP_FREE' value='16'/>
<enumerator name='ZPOOL_PROP_ALLOCATED' value='17'/>
<enumerator name='ZPOOL_PROP_READONLY' value='18'/>
<enumerator name='ZPOOL_PROP_ASHIFT' value='19'/>
<enumerator name='ZPOOL_PROP_COMMENT' value='20'/>
<enumerator name='ZPOOL_PROP_EXPANDSZ' value='21'/>
<enumerator name='ZPOOL_PROP_FREEING' value='22'/>
<enumerator name='ZPOOL_PROP_FRAGMENTATION' value='23'/>
<enumerator name='ZPOOL_PROP_LEAKED' value='24'/>
<enumerator name='ZPOOL_PROP_MAXBLOCKSIZE' value='25'/>
<enumerator name='ZPOOL_PROP_TNAME' value='26'/>
<enumerator name='ZPOOL_PROP_MAXDNODESIZE' value='27'/>
<enumerator name='ZPOOL_PROP_MULTIHOST' value='28'/>
<enumerator name='ZPOOL_PROP_CHECKPOINT' value='29'/>
<enumerator name='ZPOOL_PROP_LOAD_GUID' value='30'/>
<enumerator name='ZPOOL_PROP_AUTOTRIM' value='31'/>
<enumerator name='ZPOOL_PROP_COMPATIBILITY' value='32'/>
<enumerator name='ZPOOL_PROP_BCLONEUSED' value='33'/>
<enumerator name='ZPOOL_PROP_BCLONESAVED' value='34'/>
<enumerator name='ZPOOL_PROP_BCLONERATIO' value='35'/>
<enumerator name='ZPOOL_NUM_PROPS' value='36'/>
</enum-decl>
<typedef-decl name='zpool_prop_t' type-id='af1ba157' id='5d0c23fb'/>
<typedef-decl name='regoff_t' type-id='95e97e5e' id='54a2a2a8'/>
<class-decl name='regmatch_t' size-in-bits='64' is-struct='yes' naming-typedef-id='1b941664' visibility='default' id='4f932615'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='rm_so' type-id='54a2a2a8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='rm_eo' type-id='54a2a2a8' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='regmatch_t' type-id='4f932615' id='1b941664'/>
<typedef-decl name='__sighandler_t' type-id='03347643' id='8cdd9566'/>
<typedef-decl name='ssize_t' type-id='41060289' id='79a0948f'/>
<class-decl name='sigaction' size-in-bits='1216' is-struct='yes' visibility='default' id='fe391c48'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__sigaction_handler' type-id='ac5ab595' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='sa_mask' type-id='b9c97942' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1088'>
<var-decl name='sa_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1152'>
<var-decl name='sa_restorer' type-id='953b12f8' visibility='default'/>
</data-member>
</class-decl>
<union-decl name='__anonymous_union__' size-in-bits='64' is-anonymous='yes' visibility='default' id='ac5ab595'>
<data-member access='public'>
<var-decl name='sa_handler' type-id='8cdd9566' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='sa_sigaction' type-id='6e756877' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='termios' size-in-bits='480' is-struct='yes' visibility='default' id='ad55d2bc'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='c_iflag' type-id='241ce6f8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='c_oflag' type-id='241ce6f8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='c_cflag' type-id='241ce6f8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='c_lflag' type-id='241ce6f8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='c_line' type-id='fb7c6451' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='136'>
<var-decl name='c_cc' type-id='64177143' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='416'>
<var-decl name='c_ispeed' type-id='6a8e8a14' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='c_ospeed' type-id='6a8e8a14' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='cc_t' type-id='002ac4a6' id='fb7c6451'/>
<typedef-decl name='speed_t' type-id='f0981eeb' id='6a8e8a14'/>
<typedef-decl name='tcflag_t' type-id='f0981eeb' id='241ce6f8'/>
<typedef-decl name='__uid_t' type-id='f0981eeb' id='cc5fcceb'/>
<typedef-decl name='__off_t' type-id='bd54fe1a' id='79989e9c'/>
<typedef-decl name='__off64_t' type-id='bd54fe1a' id='724e4de6'/>
<typedef-decl name='__pid_t' type-id='95e97e5e' id='3629bad8'/>
<typedef-decl name='__clock_t' type-id='bd54fe1a' id='4d66c6d7'/>
<typedef-decl name='__ssize_t' type-id='bd54fe1a' id='41060289'/>
<typedef-decl name='FILE' type-id='ec1ed955' id='aa12d1ba'/>
<class-decl name='__sigset_t' size-in-bits='1024' is-struct='yes' naming-typedef-id='b9c97942' visibility='default' id='2616147f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__val' type-id='d2baa450' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__sigset_t' type-id='2616147f' id='b9c97942'/>
<union-decl name='sigval' size-in-bits='64' visibility='default' id='a094b870'>
<data-member access='public'>
<var-decl name='sival_int' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='sival_ptr' type-id='eaa32e2f' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='__sigval_t' type-id='a094b870' id='eabacd01'/>
<class-decl name='siginfo_t' size-in-bits='1024' is-struct='yes' naming-typedef-id='cb681f62' visibility='default' id='d8149419'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_signo' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='si_errno' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_code' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='__pad0' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='_sifields' type-id='ac5ab596' visibility='default'/>
</data-member>
</class-decl>
<union-decl name='__anonymous_union__1' size-in-bits='896' is-anonymous='yes' visibility='default' id='ac5ab596'>
<data-member access='public'>
<var-decl name='_pad' type-id='47394ee0' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_kill' type-id='e7f43f73' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_timer' type-id='e7f43f74' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_rt' type-id='e7f43f75' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_sigchld' type-id='e7f43f76' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_sigfault' type-id='e7f43f77' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_sigpoll' type-id='e7f43f78' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_sigsys' type-id='e7f43f79' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='__anonymous_struct__1' size-in-bits='64' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f73'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_pid' type-id='3629bad8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='si_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__2' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f74'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_tid' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='si_overrun' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_sigval' type-id='eabacd01' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__3' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f75'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_pid' type-id='3629bad8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='si_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_sigval' type-id='eabacd01' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__4' size-in-bits='256' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f76'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_pid' type-id='3629bad8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='si_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_status' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='si_utime' type-id='4d66c6d7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='si_stime' type-id='4d66c6d7' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__5' size-in-bits='256' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f77'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_addr' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_addr_lsb' type-id='a2185560' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='_bounds' type-id='ac5ab597' visibility='default'/>
</data-member>
</class-decl>
<union-decl name='__anonymous_union__2' size-in-bits='128' is-anonymous='yes' visibility='default' id='ac5ab597'>
<data-member access='public'>
<var-decl name='_addr_bnd' type-id='e7f43f7a' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_pkey' type-id='62f1140c' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='__anonymous_struct__6' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f7a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='_lower' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='_upper' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__7' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f78'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_band' type-id='bd54fe1a' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_fd' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__8' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f79'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='_call_addr' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='_syscall' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='_arch' type-id='f0981eeb' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='siginfo_t' type-id='d8149419' id='cb681f62'/>
<typedef-decl name='sigset_t' type-id='b9c97942' id='daf33c64'/>
<typedef-decl name='_IO_lock_t' type-id='48b5725f' id='bb4788fa'/>
<class-decl name='_IO_FILE' size-in-bits='1728' is-struct='yes' visibility='default' id='ec1ed955'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='_IO_read_ptr' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='_IO_read_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='_IO_read_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='_IO_write_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='_IO_write_ptr' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='_IO_write_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='_IO_buf_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='_IO_buf_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='_IO_save_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='640'>
<var-decl name='_IO_backup_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='_IO_save_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='768'>
<var-decl name='_markers' type-id='e4c6fa61' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='_chain' type-id='dca988a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='896'>
<var-decl name='_fileno' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='928'>
<var-decl name='_flags2' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='_old_offset' type-id='79989e9c' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1024'>
<var-decl name='_cur_column' type-id='8efea9e5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1040'>
<var-decl name='_vtable_offset' type-id='28577a57' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1048'>
<var-decl name='_shortbuf' type-id='89feb1ec' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1088'>
<var-decl name='_lock' type-id='cecf4ea7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1152'>
<var-decl name='_offset' type-id='724e4de6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1216'>
<var-decl name='_codecvt' type-id='570f8c59' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1280'>
<var-decl name='_wide_data' type-id='c65a1f29' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1344'>
<var-decl name='_freeres_list' type-id='dca988a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1408'>
<var-decl name='_freeres_buf' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1472'>
<var-decl name='__pad5' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1536'>
<var-decl name='_mode' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1568'>
<var-decl name='_unused2' type-id='664ac0b7' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='aa12d1ba' size-in-bits='64' id='822cd80b'/>
<qualified-type-def type-id='822cd80b' restrict='yes' id='e75a27e9'/>
<pointer-type-def type-id='ec1ed955' size-in-bits='64' id='dca988a5'/>
<pointer-type-def type-id='a4036571' size-in-bits='64' id='570f8c59'/>
<pointer-type-def type-id='bb4788fa' size-in-bits='64' id='cecf4ea7'/>
<pointer-type-def type-id='010ae0b9' size-in-bits='64' id='e4c6fa61'/>
<pointer-type-def type-id='79bd3751' size-in-bits='64' id='c65a1f29'/>
<qualified-type-def type-id='9b23c9ad' restrict='yes' id='8c85230f'/>
<qualified-type-def type-id='80f4b756' restrict='yes' id='9d26089a'/>
<pointer-type-def type-id='80f4b756' size-in-bits='64' id='7d3cd834'/>
<qualified-type-def type-id='aca3bac8' const='yes' id='2498fd78'/>
<pointer-type-def type-id='2498fd78' size-in-bits='64' id='eed6c816'/>
<qualified-type-def type-id='eed6c816' restrict='yes' id='a431a9da'/>
<qualified-type-def type-id='fe391c48' const='yes' id='14a93b33'/>
<pointer-type-def type-id='14a93b33' size-in-bits='64' id='9f68085b'/>
<qualified-type-def type-id='9f68085b' restrict='yes' id='e2a5e6f9'/>
<qualified-type-def type-id='ad55d2bc' const='yes' id='a46bf13f'/>
<pointer-type-def type-id='a46bf13f' size-in-bits='64' id='eaec840f'/>
<qualified-type-def type-id='002ac4a6' const='yes' id='ea86de29'/>
<pointer-type-def type-id='ea86de29' size-in-bits='64' id='354f7eb9'/>
<qualified-type-def type-id='8efea9e5' const='yes' id='3beb2af4'/>
<pointer-type-def type-id='3beb2af4' size-in-bits='64' id='31347b7a'/>
<pointer-type-def type-id='31347b7a' size-in-bits='64' id='c59e1ef0'/>
<pointer-type-def type-id='1b941664' size-in-bits='64' id='7e2979d5'/>
<qualified-type-def type-id='7e2979d5' restrict='yes' id='fc212857'/>
<pointer-type-def type-id='fe391c48' size-in-bits='64' id='568dd84e'/>
<qualified-type-def type-id='568dd84e' restrict='yes' id='3d8ee6f2'/>
<pointer-type-def type-id='cb681f62' size-in-bits='64' id='185869c1'/>
<pointer-type-def type-id='daf33c64' size-in-bits='64' id='9e80f729'/>
<pointer-type-def type-id='b59d7dce' size-in-bits='64' id='78c01427'/>
<qualified-type-def type-id='78c01427' restrict='yes' id='d19b2c25'/>
<pointer-type-def type-id='ad55d2bc' size-in-bits='64' id='665a4eda'/>
<pointer-type-def type-id='9c313c2d' size-in-bits='64' id='5d6479ae'/>
<pointer-type-def type-id='ae3e8ca6' size-in-bits='64' id='d8774064'/>
<pointer-type-def type-id='3502e3ff' size-in-bits='64' id='4dd26a40'/>
<pointer-type-def type-id='ee076206' size-in-bits='64' id='953b12f8'/>
<pointer-type-def type-id='f712e2b7' size-in-bits='64' id='03347643'/>
<pointer-type-def type-id='ef70d893' size-in-bits='64' id='6e756877'/>
<qualified-type-def type-id='eaa32e2f' restrict='yes' id='1b7446cd'/>
<class-decl name='_IO_codecvt' is-struct='yes' visibility='default' is-declaration-only='yes' id='a4036571'/>
<class-decl name='_IO_marker' is-struct='yes' visibility='default' is-declaration-only='yes' id='010ae0b9'/>
<class-decl name='_IO_wide_data' is-struct='yes' visibility='default' is-declaration-only='yes' id='79bd3751'/>
<function-decl name='zpool_get_prop_int' mangled-name='zpool_get_prop_int' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_prop_int'>
<parameter type-id='4c81de99'/>
<parameter type-id='5d0c23fb'/>
<parameter type-id='debc6aa3'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zfs_handle_dup' mangled-name='zfs_handle_dup' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_handle_dup'>
<parameter type-id='9200a744'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='zfs_valid_proplist' mangled-name='zfs_valid_proplist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_valid_proplist'>
<parameter type-id='b0382bb3'/>
<parameter type-id='2e45de5d'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9200a744'/>
<parameter type-id='4c81de99'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='80f4b756'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_prop_to_name' mangled-name='zfs_prop_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_to_name'>
<parameter type-id='58603c44'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_iter_filesystems_v2' mangled-name='zfs_iter_filesystems_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_filesystems_v2'>
<parameter type-id='9200a744'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_parent_name' mangled-name='zfs_parent_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_parent_name'>
<parameter type-id='9200a744'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_load_key' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='ae3e8ca6'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_unload_key' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_change_key' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='ae3e8ca6'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_name_to_prop' mangled-name='zfs_name_to_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_name_to_prop'>
<parameter type-id='80f4b756'/>
<return type-id='58603c44'/>
</function-decl>
<function-decl name='nvlist_add_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_add_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fnvlist_alloc' visibility='default' binding='global' size-in-bits='64'>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='__ctype_b_loc' visibility='default' binding='global' size-in-bits='64'>
<return type-id='c59e1ef0'/>
</function-decl>
<function-decl name='dlopen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='dlsym' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='1b7446cd'/>
<parameter type-id='9d26089a'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='dlerror' visibility='default' binding='global' size-in-bits='64'>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='PKCS5_PBKDF2_HMAC_SHA1' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='354f7eb9'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='cf536864'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='regexec' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a431a9da'/>
<parameter type-id='9d26089a'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='fc212857'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='kill' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='3629bad8'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sigemptyset' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9e80f729'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sigaction' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='e2a5e6f9'/>
<parameter type-id='3d8ee6f2'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fclose' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fflush' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fdopen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<return type-id='822cd80b'/>
</function-decl>
<function-decl name='fputc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__getdelim' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='8c85230f'/>
<parameter type-id='d19b2c25'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='e75a27e9'/>
<return type-id='41060289'/>
</function-decl>
<function-decl name='rewind' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='ferror' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fileno' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='malloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='calloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='strdup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strerror' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='tcgetattr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='665a4eda'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='tcsetattr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaec840f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='close' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getpid' visibility='default' binding='global' size-in-bits='64'>
<return type-id='3629bad8'/>
</function-decl>
<function-decl name='isatty' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='unlink' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__open_too_many_args' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__open_missing_mode' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__printf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__asprintf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='9d26089a'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__fread_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='1b7446cd'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='e75a27e9'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='__read_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='79a0948f'/>
</function-decl>
<function-decl name='zfs_crypto_get_encryption_root' mangled-name='zfs_crypto_get_encryption_root' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_get_encryption_root'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='37e3bd22' name='is_encroot'/>
<parameter type-id='26a90f95' name='buf'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_create' mangled-name='zfs_crypto_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_create'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='parent_name'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='5ce45b60' name='pool_props'/>
<parameter type-id='c19b74c3' name='stdin_available'/>
<parameter type-id='d8774064' name='wkeydata_out'/>
<parameter type-id='4dd26a40' name='wkeylen_out'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_clone_check' mangled-name='zfs_crypto_clone_check' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_clone_check'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='9200a744' name='origin_zhp'/>
<parameter type-id='26a90f95' name='parent_name'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_attempt_load_keys' mangled-name='zfs_crypto_attempt_load_keys' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_attempt_load_keys'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='fsname'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_load_key' mangled-name='zfs_crypto_load_key' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_load_key'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='c19b74c3' name='noop'/>
<parameter type-id='80f4b756' name='alt_keylocation'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_unload_key' mangled-name='zfs_crypto_unload_key' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_unload_key'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_rewrap' mangled-name='zfs_crypto_rewrap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_rewrap'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='5ce45b60' name='raw_props'/>
<parameter type-id='c19b74c3' name='inheritkey'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_error_aux' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='48b5725f'/>
</function-decl>
<function-type size-in-bits='64' id='f712e2b7'>
<parameter type-id='95e97e5e'/>
<return type-id='48b5725f'/>
</function-type>
<function-type size-in-bits='64' id='ef70d893'>
<parameter type-id='95e97e5e'/>
<parameter type-id='185869c1'/>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_dataset.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='32' id='8e0573fd'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<class-decl name='prop_changelist' is-struct='yes' visibility='default' is-declaration-only='yes' id='d86edc51'/>
<class-decl name='zprop_list' size-in-bits='448' is-struct='yes' visibility='default' id='bd9b4291'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='pl_prop' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='pl_user_prop' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='pl_next' type-id='9f1a1109' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='pl_all' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='pl_width' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='pl_recvd_width' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='pl_fixed' type-id='c19b74c3' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zprop_list_t' type-id='bd9b4291' id='bdb8ac4f'/>
<class-decl name='renameflags' size-in-bits='32' is-struct='yes' visibility='default' id='7aee5792'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='recursive' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1'>
<var-decl name='nounmount' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2'>
<var-decl name='forceunmount' type-id='f0981eeb' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='renameflags_t' type-id='7aee5792' id='067170c2'/>
<typedef-decl name='zfs_userspace_cb_t' type-id='ca64ff60' id='16c5f410'/>
<enum-decl name='lzc_dataset_type' id='bc9887f1'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='LZC_DATSET_TYPE_ZFS' value='2'/>
<enumerator name='LZC_DATSET_TYPE_ZVOL' value='3'/>
</enum-decl>
<typedef-decl name='avl_index_t' type-id='e475ab95' id='fba6cb51'/>
<enum-decl name='zfs_userquota_prop_t' naming-typedef-id='279fde6a' id='5258d2f6'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_PROP_USERUSED' value='0'/>
<enumerator name='ZFS_PROP_USERQUOTA' value='1'/>
<enumerator name='ZFS_PROP_GROUPUSED' value='2'/>
<enumerator name='ZFS_PROP_GROUPQUOTA' value='3'/>
<enumerator name='ZFS_PROP_USEROBJUSED' value='4'/>
<enumerator name='ZFS_PROP_USEROBJQUOTA' value='5'/>
<enumerator name='ZFS_PROP_GROUPOBJUSED' value='6'/>
<enumerator name='ZFS_PROP_GROUPOBJQUOTA' value='7'/>
<enumerator name='ZFS_PROP_PROJECTUSED' value='8'/>
<enumerator name='ZFS_PROP_PROJECTQUOTA' value='9'/>
<enumerator name='ZFS_PROP_PROJECTOBJUSED' value='10'/>
<enumerator name='ZFS_PROP_PROJECTOBJQUOTA' value='11'/>
<enumerator name='ZFS_NUM_USERQUOTA_PROPS' value='12'/>
</enum-decl>
<typedef-decl name='zfs_userquota_prop_t' type-id='5258d2f6' id='279fde6a'/>
<enum-decl name='zfs_wait_activity_t' naming-typedef-id='3024501a' id='527d5dc6'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_WAIT_DELETEQ' value='0'/>
<enumerator name='ZFS_WAIT_NUM_ACTIVITIES' value='1'/>
</enum-decl>
<typedef-decl name='zfs_wait_activity_t' type-id='527d5dc6' id='3024501a'/>
<enum-decl name='namecheck_err_t' naming-typedef-id='8e0af06e' id='f43bbcda'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='NAME_ERR_LEADING_SLASH' value='0'/>
<enumerator name='NAME_ERR_EMPTY_COMPONENT' value='1'/>
<enumerator name='NAME_ERR_TRAILING_SLASH' value='2'/>
<enumerator name='NAME_ERR_INVALCHAR' value='3'/>
<enumerator name='NAME_ERR_MULTIPLE_DELIMITERS' value='4'/>
<enumerator name='NAME_ERR_NOLETTER' value='5'/>
<enumerator name='NAME_ERR_RESERVED' value='6'/>
<enumerator name='NAME_ERR_DISKLIKE' value='7'/>
<enumerator name='NAME_ERR_TOOLONG' value='8'/>
<enumerator name='NAME_ERR_SELF_REF' value='9'/>
<enumerator name='NAME_ERR_PARENT_REF' value='10'/>
<enumerator name='NAME_ERR_NO_AT' value='11'/>
<enumerator name='NAME_ERR_NO_POUND' value='12'/>
</enum-decl>
<typedef-decl name='namecheck_err_t' type-id='f43bbcda' id='8e0af06e'/>
<enum-decl name='zprop_type_t' naming-typedef-id='31429eff' id='87676253'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='PROP_TYPE_NUMBER' value='0'/>
<enumerator name='PROP_TYPE_STRING' value='1'/>
<enumerator name='PROP_TYPE_INDEX' value='2'/>
</enum-decl>
<typedef-decl name='zprop_type_t' type-id='87676253' id='31429eff'/>
<class-decl name='mnttab' size-in-bits='256' is-struct='yes' visibility='default' id='1b055409'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='mnt_special' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='mnt_mountp' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='mnt_fstype' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='mnt_mntopts' type-id='26a90f95' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='group' size-in-bits='256' is-struct='yes' visibility='default' id='01a1b934'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='gr_name' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='gr_passwd' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='gr_gid' type-id='d94ec6d9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='gr_mem' type-id='9b23c9ad' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='mntent' size-in-bits='320' is-struct='yes' visibility='default' id='56fe4a37'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='mnt_fsname' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='mnt_dir' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='mnt_type' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='mnt_opts' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='mnt_freq' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='mnt_passno' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='passwd' size-in-bits='384' is-struct='yes' visibility='default' id='a63d15a3'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='pw_name' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='pw_passwd' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='pw_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='pw_gid' type-id='d94ec6d9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='pw_gecos' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='pw_dir' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='pw_shell' type-id='26a90f95' visibility='default'/>
</data-member>
</class-decl>
<union-decl name='pthread_mutexattr_t' size-in-bits='32' naming-typedef-id='8afd6070' visibility='default' id='7300eb00'>
<data-member access='public'>
<var-decl name='__size' type-id='8e0573fd' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='95e97e5e' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_mutexattr_t' type-id='7300eb00' id='8afd6070'/>
<typedef-decl name='int64_t' type-id='0c9942d2' id='9da381c4'/>
<typedef-decl name='__int64_t' type-id='bd54fe1a' id='0c9942d2'/>
<typedef-decl name='__gid_t' type-id='f0981eeb' id='d94ec6d9'/>
<typedef-decl name='__time_t' type-id='bd54fe1a' id='65eda9c0'/>
<class-decl name='tm' size-in-bits='448' is-struct='yes' visibility='default' id='dddf6ca2'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tm_sec' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='tm_min' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tm_hour' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='tm_mday' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='tm_mon' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='tm_year' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='tm_wday' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='tm_yday' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='tm_isdst' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='tm_gmtoff' type-id='bd54fe1a' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='tm_zone' type-id='80f4b756' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='time_t' type-id='65eda9c0' id='c9d12d66'/>
<typedef-decl name='uid_t' type-id='cc5fcceb' id='354978ed'/>
<typedef-decl name='prop_changelist_t' type-id='d86edc51' id='eae6431d'/>
<pointer-type-def type-id='fba6cb51' size-in-bits='64' id='32adbf30'/>
<pointer-type-def type-id='f20fbd51' size-in-bits='64' id='a3681dea'/>
<qualified-type-def type-id='26a90f95' restrict='yes' id='266fe297'/>
<qualified-type-def type-id='56fe4a37' const='yes' id='a75125ce'/>
<pointer-type-def type-id='a75125ce' size-in-bits='64' id='48bea5ec'/>
<qualified-type-def type-id='8afd6070' const='yes' id='1d853360'/>
<pointer-type-def type-id='1d853360' size-in-bits='64' id='c2afbd7e'/>
<qualified-type-def type-id='c9d12d66' const='yes' id='588b3216'/>
<pointer-type-def type-id='588b3216' size-in-bits='64' id='9f201474'/>
<qualified-type-def type-id='9f201474' restrict='yes' id='d6e2847c'/>
<qualified-type-def type-id='dddf6ca2' const='yes' id='e824a34f'/>
<pointer-type-def type-id='e824a34f' size-in-bits='64' id='d6ad37ff'/>
<qualified-type-def type-id='d6ad37ff' restrict='yes' id='f8c6051d'/>
<qualified-type-def type-id='9c313c2d' const='yes' id='c3b7ba7d'/>
<pointer-type-def type-id='c3b7ba7d' size-in-bits='64' id='713a56f5'/>
<pointer-type-def type-id='01a1b934' size-in-bits='64' id='566b3f52'/>
<pointer-type-def type-id='7e291ce6' size-in-bits='64' id='ca64ff60'/>
<pointer-type-def type-id='9da381c4' size-in-bits='64' id='cb785ebf'/>
<pointer-type-def type-id='1b055409' size-in-bits='64' id='9d424d31'/>
<pointer-type-def type-id='8e0af06e' size-in-bits='64' id='053457bd'/>
<pointer-type-def type-id='857bb57e' size-in-bits='64' id='75be733c'/>
<pointer-type-def type-id='a63d15a3' size-in-bits='64' id='a195f4a3'/>
<pointer-type-def type-id='eae6431d' size-in-bits='64' id='0d41d328'/>
<pointer-type-def type-id='7a6844eb' size-in-bits='64' id='18c91f9e'/>
<pointer-type-def type-id='dddf6ca2' size-in-bits='64' id='d915a820'/>
<qualified-type-def type-id='d915a820' restrict='yes' id='f099ad08'/>
<pointer-type-def type-id='5d6479ae' size-in-bits='64' id='892b4acc'/>
<pointer-type-def type-id='bd9b4291' size-in-bits='64' id='9f1a1109'/>
<pointer-type-def type-id='bdb8ac4f' size-in-bits='64' id='3a9b2288'/>
<pointer-type-def type-id='3a9b2288' size-in-bits='64' id='e4378506'/>
<class-decl name='prop_changelist' is-struct='yes' visibility='default' is-declaration-only='yes' id='d86edc51'/>
<function-decl name='zpool_open' mangled-name='zpool_open' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_open'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<return type-id='4c81de99'/>
</function-decl>
<function-decl name='zpool_open_canfail' mangled-name='zpool_open_canfail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_open_canfail'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<return type-id='4c81de99'/>
</function-decl>
<function-decl name='zpool_close' mangled-name='zpool_close' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_close'>
<parameter type-id='4c81de99'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_get_name' mangled-name='zpool_get_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_name'>
<parameter type-id='4c81de99'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_get_prop' mangled-name='zpool_get_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_prop'>
<parameter type-id='4c81de99'/>
<parameter type-id='5d0c23fb'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='debc6aa3'/>
<parameter type-id='c19b74c3'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_default_string' mangled-name='zfs_prop_default_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_default_string'>
<parameter type-id='58603c44'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_prop_default_numeric' mangled-name='zfs_prop_default_numeric' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_default_numeric'>
<parameter type-id='58603c44'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zpool_prop_get_feature' mangled-name='zpool_prop_get_feature' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_get_feature'>
<parameter type-id='4c81de99'/>
<parameter type-id='80f4b756'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_snapshots_v2' mangled-name='zfs_iter_snapshots_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapshots_v2'>
<parameter type-id='9200a744'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9c313c2d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_bookmarks_v2' mangled-name='zfs_iter_bookmarks_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_bookmarks_v2'>
<parameter type-id='9200a744'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_destroy_snaps_nvl_os' mangled-name='zfs_destroy_snaps_nvl_os' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_destroy_snaps_nvl_os'>
<parameter type-id='b0382bb3'/>
<parameter type-id='5ce45b60'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_nicestrtonum' mangled-name='zfs_nicestrtonum' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_nicestrtonum'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_snapshot' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='bc9887f1'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='ae3e8ca6'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_clone' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_promote' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='26a90f95'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_destroy_snaps' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_bookmarks' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_destroy_bookmarks' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_hold' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_release' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_holds' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_exists' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='lzc_rollback_to' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_destroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_channel_program_nosync' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_wait_fs' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='3024501a'/>
<parameter type-id='37e3bd22'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_nicebytes' mangled-name='zfs_nicebytes' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_nicebytes'>
<parameter type-id='9c313c2d'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_nicenum' mangled-name='zfs_nicenum' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_nicenum'>
<parameter type-id='9c313c2d'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_create' mangled-name='avl_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_create'>
<parameter type-id='a3681dea'/>
<parameter type-id='585e1de9'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_find' mangled-name='avl_find' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_find'>
<parameter type-id='a3681dea'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='32adbf30'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='avl_add' mangled-name='avl_add' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_add'>
<parameter type-id='a3681dea'/>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_remove' mangled-name='avl_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_remove'>
<parameter type-id='a3681dea'/>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_numnodes' mangled-name='avl_numnodes' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_numnodes'>
<parameter type-id='a3681dea'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='avl_destroy_nodes' mangled-name='avl_destroy_nodes' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_destroy_nodes'>
<parameter type-id='a3681dea'/>
<parameter type-id='63e171df'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='avl_destroy' mangled-name='avl_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_destroy'>
<parameter type-id='a3681dea'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_prop_readonly' mangled-name='zfs_prop_readonly' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_readonly'>
<parameter type-id='58603c44'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_inheritable' mangled-name='zfs_prop_inheritable' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_inheritable'>
<parameter type-id='58603c44'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_setonce' mangled-name='zfs_prop_setonce' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_setonce'>
<parameter type-id='58603c44'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_encryption_key_param' mangled-name='zfs_prop_encryption_key_param' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_encryption_key_param'>
<parameter type-id='58603c44'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_valid_keylocation' mangled-name='zfs_prop_valid_keylocation' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_valid_keylocation'>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_user' mangled-name='zfs_prop_user' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_user'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_userquota' mangled-name='zfs_prop_userquota' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_userquota'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_written' mangled-name='zfs_prop_written' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_written'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_index_to_string' mangled-name='zfs_prop_index_to_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_index_to_string'>
<parameter type-id='58603c44'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_valid_for_type' mangled-name='zfs_prop_valid_for_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_valid_for_type'>
<parameter type-id='95e97e5e'/>
<parameter type-id='2e45de5d'/>
<parameter type-id='c19b74c3'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='nvlist_alloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='857bb57e'/>
<parameter type-id='3502e3ff'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_size' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='78c01427'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_pack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='9b23c9ad'/>
<parameter type-id='78c01427'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_unpack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='857bb57e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_add_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='22cce67b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_add_uint64_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='713a56f5'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_remove' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='8d0687d2'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_remove_all' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_int64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<parameter type-id='cb785ebf'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_uint64_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='892b4acc'/>
<parameter type-id='4dd26a40'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_nvlist_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='75be733c'/>
<parameter type-id='4dd26a40'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_empty' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='nvpair_type' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='8d0687d2'/>
</function-decl>
<function-decl name='nvpair_value_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvpair_value_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fnvlist_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_boolean' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_lookup_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='fnvlist_lookup_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='fnvlist_lookup_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='fnvpair_value_int32' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='3ff5601b'/>
</function-decl>
<function-decl name='fnvpair_value_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='entity_namecheck' mangled-name='entity_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='entity_namecheck'>
<parameter type-id='80f4b756'/>
<parameter type-id='053457bd'/>
<parameter type-id='26a90f95'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='dataset_nestcheck' mangled-name='dataset_nestcheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='dataset_nestcheck'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='mountpoint_namecheck' mangled-name='mountpoint_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='mountpoint_namecheck'>
<parameter type-id='80f4b756'/>
<parameter type-id='053457bd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_type' mangled-name='zfs_prop_get_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_type'>
<parameter type-id='58603c44'/>
<return type-id='31429eff'/>
</function-decl>
<function-decl name='sa_validate_shareopts' mangled-name='sa_validate_shareopts' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_validate_shareopts'>
<parameter type-id='80f4b756'/>
<parameter type-id='9155d4b5'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getmntany' mangled-name='getmntany' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getmntany'>
<parameter type-id='822cd80b'/>
<parameter type-id='9d424d31'/>
<parameter type-id='9d424d31'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='_sol_getmntent' mangled-name='_sol_getmntent' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='_sol_getmntent'>
<parameter type-id='822cd80b'/>
<parameter type-id='9d424d31'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getgrnam' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='566b3f52'/>
</function-decl>
<function-decl name='hasmntopt' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='48bea5ec'/>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='pthread_mutex_init' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='18c91f9e'/>
<parameter type-id='c2afbd7e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_mutex_destroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='18c91f9e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_mutex_lock' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='18c91f9e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_mutex_unlock' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='18c91f9e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getpwnam' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='a195f4a3'/>
</function-decl>
<function-decl name='strtol' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<return type-id='bd54fe1a'/>
</function-decl>
<function-decl name='strtoul' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<return type-id='7359adad'/>
</function-decl>
<function-decl name='abort' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strrchr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strcspn' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='strstr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strsep' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='8c85230f'/>
<parameter type-id='9d26089a'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strftime' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='266fe297'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='9d26089a'/>
<parameter type-id='f8c6051d'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='localtime_r' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='d6e2847c'/>
<parameter type-id='f099ad08'/>
<return type-id='d915a820'/>
</function-decl>
<function-decl name='__fprintf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e75a27e9'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='9d26089a'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='ioctl' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='7359adad'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_type_to_name' mangled-name='zfs_type_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_type_to_name'>
<parameter type-id='2e45de5d' name='type'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_name_valid' mangled-name='zfs_name_valid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_name_valid'>
<parameter type-id='80f4b756' name='name'/>
<parameter type-id='2e45de5d' name='type'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_free_handles' mangled-name='zpool_free_handles' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_free_handles'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_bookmark_exists' mangled-name='zfs_bookmark_exists' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_bookmark_exists'>
<parameter type-id='80f4b756' name='path'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='libzfs_mnttab_init' mangled-name='libzfs_mnttab_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_init'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_mnttab_fini' mangled-name='libzfs_mnttab_fini' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_fini'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_mnttab_cache' mangled-name='libzfs_mnttab_cache' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_cache'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='c19b74c3' name='enable'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_mnttab_find' mangled-name='libzfs_mnttab_find' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_find'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='fsname'/>
<parameter type-id='9d424d31' name='entry'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_mnttab_add' mangled-name='libzfs_mnttab_add' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_add'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='special'/>
<parameter type-id='80f4b756' name='mountp'/>
<parameter type-id='80f4b756' name='mntopts'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_mnttab_remove' mangled-name='libzfs_mnttab_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_remove'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='fsname'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_spa_version' mangled-name='zfs_spa_version' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_spa_version'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='7292109c' name='spa_version'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_set' mangled-name='zfs_prop_set' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_set'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='80f4b756' name='propval'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_set_list' mangled-name='zfs_prop_set_list' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_set_list'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_set_list_flags' mangled-name='zfs_prop_set_list_flags' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_set_list_flags'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_inherit' mangled-name='zfs_prop_inherit' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_inherit'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='c19b74c3' name='received'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getprop_uint64' mangled-name='getprop_uint64' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getprop_uint64'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='58603c44' name='prop'/>
<parameter type-id='7d3cd834' name='source'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zfs_prop_get_recvd' mangled-name='zfs_prop_get_recvd' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_recvd'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='26a90f95' name='propbuf'/>
<parameter type-id='b59d7dce' name='proplen'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_get_clones_nvl' mangled-name='zfs_get_clones_nvl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_clones_nvl'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_prop_get_numeric' mangled-name='zfs_prop_get_numeric' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_numeric'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='58603c44' name='prop'/>
<parameter type-id='5d6479ae' name='value'/>
<parameter type-id='debc6aa3' name='src'/>
<parameter type-id='26a90f95' name='statbuf'/>
<parameter type-id='b59d7dce' name='statlen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_userquota_int' mangled-name='zfs_prop_get_userquota_int' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_userquota_int'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='5d6479ae' name='propvalue'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_userquota' mangled-name='zfs_prop_get_userquota' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_userquota'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='26a90f95' name='propbuf'/>
<parameter type-id='95e97e5e' name='proplen'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_written_int' mangled-name='zfs_prop_get_written_int' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_written_int'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='5d6479ae' name='propvalue'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_written' mangled-name='zfs_prop_get_written' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_written'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='26a90f95' name='propbuf'/>
<parameter type-id='95e97e5e' name='proplen'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_get_pool_name' mangled-name='zfs_get_pool_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_pool_name'>
<parameter type-id='fcd57163' name='zhp'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_get_type' mangled-name='zfs_get_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_type'>
<parameter type-id='fcd57163' name='zhp'/>
<return type-id='2e45de5d'/>
</function-decl>
<function-decl name='zfs_get_underlying_type' mangled-name='zfs_get_underlying_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_underlying_type'>
<parameter type-id='fcd57163' name='zhp'/>
<return type-id='2e45de5d'/>
</function-decl>
<function-decl name='zfs_dataset_exists' mangled-name='zfs_dataset_exists' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dataset_exists'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='2e45de5d' name='types'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_create_ancestors' mangled-name='zfs_create_ancestors' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_create_ancestors'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_create' mangled-name='zfs_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_create'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='2e45de5d' name='type'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_destroy' mangled-name='zfs_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_destroy'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='c19b74c3' name='defer'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_destroy_snaps' mangled-name='zfs_destroy_snaps' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_destroy_snaps'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='26a90f95' name='snapname'/>
<parameter type-id='c19b74c3' name='defer'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_destroy_snaps_nvl' mangled-name='zfs_destroy_snaps_nvl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_destroy_snaps_nvl'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='5ce45b60' name='snaps'/>
<parameter type-id='c19b74c3' name='defer'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_clone' mangled-name='zfs_clone' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_clone'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='target'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_promote' mangled-name='zfs_promote' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_promote'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_snapshot_nvl' mangled-name='zfs_snapshot_nvl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_snapshot_nvl'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='5ce45b60' name='snaps'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_snapshot' mangled-name='zfs_snapshot' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_snapshot'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='c19b74c3' name='recursive'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_rollback' mangled-name='zfs_rollback' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_rollback'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='9200a744' name='snap'/>
<parameter type-id='c19b74c3' name='force'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_rename' mangled-name='zfs_rename' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_rename'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='target'/>
<parameter type-id='067170c2' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_get_all_props' mangled-name='zfs_get_all_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_all_props'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_get_recvd_props' mangled-name='zfs_get_recvd_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_recvd_props'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_get_user_props' mangled-name='zfs_get_user_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_user_props'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_expand_proplist' mangled-name='zfs_expand_proplist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_expand_proplist'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='e4378506' name='plp'/>
<parameter type-id='c19b74c3' name='received'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prune_proplist' mangled-name='zfs_prune_proplist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prune_proplist'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='ae3e8ca6' name='props'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_smb_acl_add' mangled-name='zfs_smb_acl_add' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_smb_acl_add'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='dataset'/>
<parameter type-id='26a90f95' name='path'/>
<parameter type-id='26a90f95' name='resource'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_smb_acl_remove' mangled-name='zfs_smb_acl_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_smb_acl_remove'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='dataset'/>
<parameter type-id='26a90f95' name='path'/>
<parameter type-id='26a90f95' name='resource'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_smb_acl_purge' mangled-name='zfs_smb_acl_purge' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_smb_acl_purge'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='dataset'/>
<parameter type-id='26a90f95' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_smb_acl_rename' mangled-name='zfs_smb_acl_rename' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_smb_acl_rename'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='dataset'/>
<parameter type-id='26a90f95' name='path'/>
<parameter type-id='26a90f95' name='oldname'/>
<parameter type-id='26a90f95' name='newname'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_userspace' mangled-name='zfs_userspace' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_userspace'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='279fde6a' name='type'/>
<parameter type-id='16c5f410' name='func'/>
<parameter type-id='eaa32e2f' name='arg'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_hold' mangled-name='zfs_hold' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_hold'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='80f4b756' name='tag'/>
<parameter type-id='c19b74c3' name='recursive'/>
<parameter type-id='95e97e5e' name='cleanup_fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_hold_nvl' mangled-name='zfs_hold_nvl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_hold_nvl'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='95e97e5e' name='cleanup_fd'/>
<parameter type-id='5ce45b60' name='holds'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_release' mangled-name='zfs_release' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_release'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='80f4b756' name='tag'/>
<parameter type-id='c19b74c3' name='recursive'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_get_fsacl' mangled-name='zfs_get_fsacl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_fsacl'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='857bb57e' name='nvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_set_fsacl' mangled-name='zfs_set_fsacl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_set_fsacl'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='c19b74c3' name='un'/>
<parameter type-id='5ce45b60' name='nvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_get_holds' mangled-name='zfs_get_holds' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_holds'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='857bb57e' name='nvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zvol_volsize_to_reservation' mangled-name='zvol_volsize_to_reservation' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zvol_volsize_to_reservation'>
<parameter type-id='4c81de99' name='zph'/>
<parameter type-id='9c313c2d' name='volsize'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zfs_wait_status' mangled-name='zfs_wait_status' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_wait_status'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='3024501a' name='activity'/>
<parameter type-id='37e3bd22' name='missing'/>
<parameter type-id='37e3bd22' name='waited'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_error_fmt' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_standard_error_fmt' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_setprop_error' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='58603c44'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='26a90f95'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_parse_value' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='3fa542f0'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='2e45de5d'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='7d3cd834'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_expand_list' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4378506'/>
<parameter type-id='2e45de5d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zcmd_write_src_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='changelist_prefix' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='changelist_postfix' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='changelist_rename' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='changelist_remove' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='changelist_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='changelist_gather' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<parameter type-id='58603c44'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<return type-id='0d41d328'/>
</function-decl>
<function-decl name='changelist_haszonedchild' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_name_valid' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-type size-in-bits='64' id='7e291ce6'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='80f4b756'/>
<parameter type-id='354978ed'/>
<parameter type-id='9c313c2d'/>
<return type-id='95e97e5e'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_diff.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='448' id='6093ff7c'>
<subrange length='56' type-id='7359adad' id='f8137894'/>
</array-type-def>
<typedef-decl name='pthread_t' type-id='7359adad' id='4051f5e7'/>
<union-decl name='pthread_attr_t' size-in-bits='448' visibility='default' id='b63afacd'>
<data-member access='public'>
<var-decl name='__size' type-id='6093ff7c' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='bd54fe1a' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_attr_t' type-id='b63afacd' id='7d8569fd'/>
<class-decl name='differ_info' size-in-bits='9088' is-struct='yes' visibility='default' id='d41965ee'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zhp' type-id='9200a744' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='fromsnap' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='frommnt' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='tosnap' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='tomnt' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='ds' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='dsmnt' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='tmpsnap' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='errbuf' type-id='b54ce520' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8704'>
<var-decl name='isclone' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8736'>
<var-decl name='scripted' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8768'>
<var-decl name='classify' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8800'>
<var-decl name='timestamped' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8832'>
<var-decl name='no_mangle' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8896'>
<var-decl name='shares' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8960'>
<var-decl name='zerr' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8992'>
<var-decl name='cleanupfd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='9024'>
<var-decl name='outputfd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='9056'>
<var-decl name='datafd' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='differ_info_t' type-id='d41965ee' id='e8525f0e'/>
<qualified-type-def type-id='7d8569fd' const='yes' id='e06dee2d'/>
<pointer-type-def type-id='e06dee2d' size-in-bits='64' id='540db505'/>
<qualified-type-def type-id='540db505' restrict='yes' id='e1815e87'/>
<pointer-type-def type-id='e8525f0e' size-in-bits='64' id='ee78f675'/>
<pointer-type-def type-id='4051f5e7' size-in-bits='64' id='e01b5462'/>
<qualified-type-def type-id='e01b5462' restrict='yes' id='cc338b26'/>
<pointer-type-def type-id='cd5d79f4' size-in-bits='64' id='5ad9edb6'/>
<function-decl name='is_mounted' mangled-name='is_mounted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='is_mounted'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9b23c9ad'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='color_start' mangled-name='color_start' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='color_start'>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='color_end' mangled-name='color_end' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='color_end'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='pthread_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='cc338b26'/>
<parameter type-id='e1815e87'/>
<parameter type-id='5ad9edb6'/>
<parameter type-id='1b7446cd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_join' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='4051f5e7'/>
<parameter type-id='63e171df'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_cancel' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='4051f5e7'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fputs' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='e75a27e9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pipe2' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7292109c'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_show_diffs' mangled-name='zfs_show_diffs' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_show_diffs'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='95e97e5e' name='outfd'/>
<parameter type-id='80f4b756' name='fromsnap'/>
<parameter type-id='80f4b756' name='tosnap'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_asprintf' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zfs_validate_name' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='c19b74c3'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='find_shares_object' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='ee78f675'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-type size-in-bits='64' id='cd5d79f4'>
<parameter type-id='eaa32e2f'/>
<return type-id='eaa32e2f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_import.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='03085adc' size-in-bits='192' id='083f8d58'>
<subrange length='3' type-id='7359adad' id='56f209d2'/>
</array-type-def>
<typedef-decl name='refresh_config_func_t' type-id='29f040d2' id='b7c58eaa'/>
<typedef-decl name='pool_active_func_t' type-id='baa42fef' id='de5d1d8f'/>
<class-decl name='pool_config_ops' size-in-bits='128' is-struct='yes' visibility='default' id='8b092c69'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='pco_refresh_config' type-id='e7c00489' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='pco_pool_active' type-id='9eadf5e0' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='pool_config_ops_t' type-id='1a21babe' id='b1e62775'/>
<enum-decl name='pool_state' id='4871ac24'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_STATE_ACTIVE' value='0'/>
<enumerator name='POOL_STATE_EXPORTED' value='1'/>
<enumerator name='POOL_STATE_DESTROYED' value='2'/>
<enumerator name='POOL_STATE_SPARE' value='3'/>
<enumerator name='POOL_STATE_L2CACHE' value='4'/>
<enumerator name='POOL_STATE_UNINITIALIZED' value='5'/>
<enumerator name='POOL_STATE_UNAVAIL' value='6'/>
<enumerator name='POOL_STATE_POTENTIALLY_ACTIVE' value='7'/>
</enum-decl>
<typedef-decl name='pool_state_t' type-id='4871ac24' id='084a08a3'/>
<class-decl name='stat64' size-in-bits='1152' is-struct='yes' visibility='default' id='0bbec9cd'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='st_dev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='st_ino' type-id='71288a47' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='st_nlink' type-id='80f0b9df' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='st_mode' type-id='e1c52942' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='st_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='st_gid' type-id='d94ec6d9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='__pad0' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='st_rdev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='st_size' type-id='79989e9c' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='st_blksize' type-id='d3f10a7f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='st_blocks' type-id='4e711bf1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='st_atim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='st_mtim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='st_ctim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='__glibc_reserved' type-id='083f8d58' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__dev_t' type-id='7359adad' id='35ed8932'/>
<typedef-decl name='__ino64_t' type-id='7359adad' id='71288a47'/>
<typedef-decl name='__mode_t' type-id='f0981eeb' id='e1c52942'/>
<typedef-decl name='__nlink_t' type-id='7359adad' id='80f0b9df'/>
<typedef-decl name='__blksize_t' type-id='bd54fe1a' id='d3f10a7f'/>
<typedef-decl name='__blkcnt64_t' type-id='bd54fe1a' id='4e711bf1'/>
<typedef-decl name='__syscall_slong_t' type-id='bd54fe1a' id='03085adc'/>
<class-decl name='timespec' size-in-bits='128' is-struct='yes' visibility='default' id='a9c79a1f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tv_sec' type-id='65eda9c0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tv_nsec' type-id='03085adc' visibility='default'/>
</data-member>
</class-decl>
<qualified-type-def type-id='8b092c69' const='yes' id='1a21babe'/>
<pointer-type-def type-id='de5d1d8f' size-in-bits='64' id='9eadf5e0'/>
<pointer-type-def type-id='084a08a3' size-in-bits='64' id='b9ea57b8'/>
<pointer-type-def type-id='b7c58eaa' size-in-bits='64' id='e7c00489'/>
<pointer-type-def type-id='0bbec9cd' size-in-bits='64' id='62f7a03d'/>
<var-decl name='libzfs_config_ops' type-id='b1e62775' mangled-name='libzfs_config_ops' visibility='default' elf-symbol-id='libzfs_config_ops'/>
<function-decl name='zpool_read_label' mangled-name='zpool_read_label' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_read_label'>
<parameter type-id='95e97e5e'/>
<parameter type-id='857bb57e'/>
<parameter type-id='7292109c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pwrite64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='724e4de6'/>
<return type-id='79a0948f'/>
</function-decl>
<function-decl name='__pread64_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='724e4de6'/>
<parameter type-id='b59d7dce'/>
<return type-id='79a0948f'/>
</function-decl>
<function-decl name='fstat64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='62f7a03d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zcmd_write_conf_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_clear_label' mangled-name='zpool_clear_label' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_clear_label'>
<parameter type-id='95e97e5e' name='fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_in_use' mangled-name='zpool_in_use' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_in_use'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='b9ea57b8' name='state'/>
<parameter type-id='9b23c9ad' name='namestr'/>
<parameter type-id='37e3bd22' name='inuse'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-type size-in-bits='64' id='baa42fef'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='37e3bd22'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='29f040d2'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='5ce45b60'/>
<return type-id='5ce45b60'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_iter.c' language='LANG_C99'>
<pointer-type-def type-id='b351119f' size-in-bits='64' id='716943c7'/>
<function-decl name='avl_first' mangled-name='avl_first' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_first'>
<parameter type-id='a3681dea'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='avl_walk' mangled-name='avl_walk' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_walk'>
<parameter type-id='716943c7'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='95e97e5e'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='make_dataset_handle_zc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='make_dataset_simple_handle_zc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<parameter type-id='e4ec4540'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='make_bookmark_handle' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='zfs_iter_filesystems' mangled-name='zfs_iter_filesystems' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_filesystems'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_snapshots' mangled-name='zfs_iter_snapshots' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapshots'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='c19b74c3' name='simple'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<parameter type-id='9c313c2d' name='min_txg'/>
<parameter type-id='9c313c2d' name='max_txg'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_bookmarks' mangled-name='zfs_iter_bookmarks' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_bookmarks'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_snapshots_sorted' mangled-name='zfs_iter_snapshots_sorted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapshots_sorted'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='d8e49ab9' name='callback'/>
<parameter type-id='eaa32e2f' name='data'/>
<parameter type-id='9c313c2d' name='min_txg'/>
<parameter type-id='9c313c2d' name='max_txg'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_snapshots_sorted_v2' mangled-name='zfs_iter_snapshots_sorted_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapshots_sorted_v2'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='95e97e5e' name='flags'/>
<parameter type-id='d8e49ab9' name='callback'/>
<parameter type-id='eaa32e2f' name='data'/>
<parameter type-id='9c313c2d' name='min_txg'/>
<parameter type-id='9c313c2d' name='max_txg'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_snapspec' mangled-name='zfs_iter_snapspec' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapspec'>
<parameter type-id='9200a744' name='fs_zhp'/>
<parameter type-id='80f4b756' name='spec_orig'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='arg'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_snapspec_v2' mangled-name='zfs_iter_snapspec_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapspec_v2'>
<parameter type-id='9200a744' name='fs_zhp'/>
<parameter type-id='95e97e5e' name='flags'/>
<parameter type-id='80f4b756' name='spec_orig'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='arg'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_children' mangled-name='zfs_iter_children' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_children'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_dependents' mangled-name='zfs_iter_dependents' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_dependents'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='c19b74c3' name='allowrecursion'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_mount.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='6028cbfe' size-in-bits='256' id='b39b9aa7'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<class-decl name='__dirstream' is-struct='yes' visibility='default' is-declaration-only='yes' id='20cd73f2'/>
<class-decl name='tpool' size-in-bits='2496' is-struct='yes' visibility='default' id='88d1b7f9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tp_forw' type-id='9cf59a50' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tp_back' type-id='9cf59a50' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='tp_mutex' type-id='7a6844eb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='tp_busycv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='tp_workcv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1216'>
<var-decl name='tp_waitcv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1600'>
<var-decl name='tp_active' type-id='ad33e5e7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1664'>
<var-decl name='tp_head' type-id='f32b30e4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1728'>
<var-decl name='tp_tail' type-id='f32b30e4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1792'>
<var-decl name='tp_attr' type-id='7d8569fd' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2240'>
<var-decl name='tp_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2272'>
<var-decl name='tp_linger' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2304'>
<var-decl name='tp_njobs' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2336'>
<var-decl name='tp_minimum' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2368'>
<var-decl name='tp_maximum' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2400'>
<var-decl name='tp_current' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2432'>
<var-decl name='tp_idle' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<array-type-def dimensions='1' type-id='95e97e5e' size-in-bits='64' id='e4266c7e'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
<class-decl name='get_all_cb' size-in-bits='192' is-struct='yes' visibility='default' id='803dac95'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='cb_handles' type-id='4507922a' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='cb_alloc' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='cb_used' type-id='b59d7dce' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='get_all_cb_t' type-id='803dac95' id='9b293607'/>
<typedef-decl name='tpool_t' type-id='88d1b7f9' id='b1bbf10d'/>
<typedef-decl name='DIR' type-id='20cd73f2' id='54a5d683'/>
<typedef-decl name='mode_t' type-id='e1c52942' id='d50d396c'/>
<typedef-decl name='__compar_fn_t' type-id='585e1de9' id='aba7edd8'/>
<class-decl name='dirent64' size-in-bits='2240' is-struct='yes' visibility='default' id='5725d813'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='d_ino' type-id='71288a47' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='d_off' type-id='724e4de6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='d_reclen' type-id='8efea9e5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='144'>
<var-decl name='d_type' type-id='002ac4a6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='152'>
<var-decl name='d_name' type-id='d1617432' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='statfs64' size-in-bits='960' is-struct='yes' visibility='default' id='a2a6be1a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='f_type' type-id='6028cbfe' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='f_bsize' type-id='6028cbfe' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='f_blocks' type-id='95fe1a02' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='f_bfree' type-id='95fe1a02' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='f_bavail' type-id='95fe1a02' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='f_files' type-id='0c3a4dde' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='f_ffree' type-id='0c3a4dde' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='f_fsid' type-id='0f35d263' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='f_namelen' type-id='6028cbfe' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='f_frsize' type-id='6028cbfe' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='640'>
<var-decl name='f_flags' type-id='6028cbfe' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='f_spare' type-id='b39b9aa7' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='stat' size-in-bits='1152' is-struct='yes' visibility='default' id='aafc373f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='st_dev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='st_ino' type-id='e43e523d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='st_nlink' type-id='80f0b9df' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='st_mode' type-id='e1c52942' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='st_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='st_gid' type-id='d94ec6d9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='__pad0' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='st_rdev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='st_size' type-id='79989e9c' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='st_blksize' type-id='d3f10a7f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='st_blocks' type-id='dbc43803' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='st_atim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='st_mtim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='st_ctim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='__glibc_reserved' type-id='083f8d58' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__ino_t' type-id='7359adad' id='e43e523d'/>
<class-decl name='__fsid_t' size-in-bits='64' is-struct='yes' naming-typedef-id='0f35d263' visibility='default' id='ea35c84a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__val' type-id='e4266c7e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__fsid_t' type-id='ea35c84a' id='0f35d263'/>
<typedef-decl name='__blkcnt_t' type-id='bd54fe1a' id='dbc43803'/>
<typedef-decl name='__fsblkcnt64_t' type-id='7359adad' id='95fe1a02'/>
<typedef-decl name='__fsfilcnt64_t' type-id='7359adad' id='0c3a4dde'/>
<typedef-decl name='__fsword_t' type-id='bd54fe1a' id='6028cbfe'/>
<pointer-type-def type-id='54a5d683' size-in-bits='64' id='f09217ba'/>
<pointer-type-def type-id='5725d813' size-in-bits='64' id='07b96073'/>
<pointer-type-def type-id='9b293607' size-in-bits='64' id='77bf1784'/>
<pointer-type-def type-id='7d8569fd' size-in-bits='64' id='7347a39e'/>
<pointer-type-def type-id='aafc373f' size-in-bits='64' id='4330df87'/>
<qualified-type-def type-id='4330df87' restrict='yes' id='73665405'/>
<pointer-type-def type-id='a2a6be1a' size-in-bits='64' id='7fd094c8'/>
<pointer-type-def type-id='b1bbf10d' size-in-bits='64' id='9cf59a50'/>
<pointer-type-def type-id='c5c76c9c' size-in-bits='64' id='b7f9d8e6'/>
<pointer-type-def type-id='9200a744' size-in-bits='64' id='4507922a'/>
<class-decl name='__dirstream' is-struct='yes' visibility='default' is-declaration-only='yes' id='20cd73f2'/>
<function-decl name='zpool_disable_datasets_os' mangled-name='zpool_disable_datasets_os' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_disable_datasets_os'>
<parameter type-id='4c81de99'/>
<parameter type-id='c19b74c3'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_disable_volume_os' mangled-name='zpool_disable_volume_os' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_disable_volume_os'>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='tpool_create' mangled-name='tpool_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_create'>
<parameter type-id='3502e3ff'/>
<parameter type-id='3502e3ff'/>
<parameter type-id='3502e3ff'/>
<parameter type-id='7347a39e'/>
<return type-id='9cf59a50'/>
</function-decl>
<function-decl name='tpool_dispatch' mangled-name='tpool_dispatch' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_dispatch'>
<parameter type-id='9cf59a50'/>
<parameter type-id='b7f9d8e6'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='tpool_destroy' mangled-name='tpool_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_destroy'>
<parameter type-id='9cf59a50'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='tpool_wait' mangled-name='tpool_wait' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_wait'>
<parameter type-id='9cf59a50'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='mkdirp' mangled-name='mkdirp' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='mkdirp'>
<parameter type-id='80f4b756'/>
<parameter type-id='d50d396c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sa_errorstr' mangled-name='sa_errorstr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_errorstr'>
<parameter type-id='95e97e5e'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='sa_enable_share' mangled-name='sa_enable_share' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_enable_share'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9155d4b5'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sa_disable_share' mangled-name='sa_disable_share' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_disable_share'>
<parameter type-id='80f4b756'/>
<parameter type-id='9155d4b5'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sa_is_shared' mangled-name='sa_is_shared' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_is_shared'>
<parameter type-id='80f4b756'/>
<parameter type-id='9155d4b5'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='sa_truncate_shares' mangled-name='sa_truncate_shares' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_truncate_shares'>
<parameter type-id='9155d4b5'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fdopendir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='f09217ba'/>
</function-decl>
<function-decl name='closedir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f09217ba'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='readdir64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f09217ba'/>
<return type-id='07b96073'/>
</function-decl>
<function-decl name='qsort' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='aba7edd8'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='rmdir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__openat_too_many_args' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__openat_missing_mode' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='statfs64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='7fd094c8'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_realloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='changelist_unshare' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<parameter type-id='4567bbc9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='do_mount' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='do_unmount' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_mount_at' mangled-name='zfs_mount_at' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_mount_at'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='options'/>
<parameter type-id='95e97e5e' name='flags'/>
<parameter type-id='80f4b756' name='mountpoint'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_unmountall' mangled-name='zfs_unmountall' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_unmountall'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_truncate_shares' mangled-name='zfs_truncate_shares' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_truncate_shares'>
<parameter type-id='4567bbc9' name='proto'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_unshareall' mangled-name='zfs_unshareall' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_unshareall'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='4567bbc9' name='proto'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_add_handle' mangled-name='libzfs_add_handle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_add_handle'>
<parameter type-id='77bf1784' name='cbp'/>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_foreach_mountpoint' mangled-name='zfs_foreach_mountpoint' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_foreach_mountpoint'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='4507922a' name='handles'/>
<parameter type-id='b59d7dce' name='num_handles'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<parameter type-id='c19b74c3' name='parallel'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_enable_datasets' mangled-name='zpool_enable_datasets' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_enable_datasets'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='mntopts'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_disable_datasets' mangled-name='zpool_disable_datasets' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_disable_datasets'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='c19b74c3' name='force'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_pool.c' language='LANG_C99'>
<type-decl name='long long unsigned int' size-in-bits='64' id='3a47d82b'/>
<class-decl name='splitflags' size-in-bits='64' is-struct='yes' visibility='default' id='dc01bf52'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='dryrun' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1'>
<var-decl name='import' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='name_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='splitflags_t' type-id='dc01bf52' id='325c1e34'/>
<class-decl name='trimflags' size-in-bits='192' is-struct='yes' visibility='default' id='8ef58008'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='fullpool' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='secure' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='wait' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='rate' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='trimflags_t' type-id='8ef58008' id='a093cbb8'/>
<enum-decl name='zpool_status_t' naming-typedef-id='d3dd6294' id='5e770b40'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_STATUS_CORRUPT_CACHE' value='0'/>
<enumerator name='ZPOOL_STATUS_MISSING_DEV_R' value='1'/>
<enumerator name='ZPOOL_STATUS_MISSING_DEV_NR' value='2'/>
<enumerator name='ZPOOL_STATUS_CORRUPT_LABEL_R' value='3'/>
<enumerator name='ZPOOL_STATUS_CORRUPT_LABEL_NR' value='4'/>
<enumerator name='ZPOOL_STATUS_BAD_GUID_SUM' value='5'/>
<enumerator name='ZPOOL_STATUS_CORRUPT_POOL' value='6'/>
<enumerator name='ZPOOL_STATUS_CORRUPT_DATA' value='7'/>
<enumerator name='ZPOOL_STATUS_FAILING_DEV' value='8'/>
<enumerator name='ZPOOL_STATUS_VERSION_NEWER' value='9'/>
<enumerator name='ZPOOL_STATUS_HOSTID_MISMATCH' value='10'/>
<enumerator name='ZPOOL_STATUS_HOSTID_ACTIVE' value='11'/>
<enumerator name='ZPOOL_STATUS_HOSTID_REQUIRED' value='12'/>
<enumerator name='ZPOOL_STATUS_IO_FAILURE_WAIT' value='13'/>
<enumerator name='ZPOOL_STATUS_IO_FAILURE_CONTINUE' value='14'/>
<enumerator name='ZPOOL_STATUS_IO_FAILURE_MMP' value='15'/>
<enumerator name='ZPOOL_STATUS_BAD_LOG' value='16'/>
<enumerator name='ZPOOL_STATUS_ERRATA' value='17'/>
<enumerator name='ZPOOL_STATUS_UNSUP_FEAT_READ' value='18'/>
<enumerator name='ZPOOL_STATUS_UNSUP_FEAT_WRITE' value='19'/>
<enumerator name='ZPOOL_STATUS_FAULTED_DEV_R' value='20'/>
<enumerator name='ZPOOL_STATUS_FAULTED_DEV_NR' value='21'/>
<enumerator name='ZPOOL_STATUS_VERSION_OLDER' value='22'/>
<enumerator name='ZPOOL_STATUS_FEAT_DISABLED' value='23'/>
<enumerator name='ZPOOL_STATUS_RESILVERING' value='24'/>
<enumerator name='ZPOOL_STATUS_OFFLINE_DEV' value='25'/>
<enumerator name='ZPOOL_STATUS_REMOVED_DEV' value='26'/>
<enumerator name='ZPOOL_STATUS_REBUILDING' value='27'/>
<enumerator name='ZPOOL_STATUS_REBUILD_SCRUB' value='28'/>
<enumerator name='ZPOOL_STATUS_NON_NATIVE_ASHIFT' value='29'/>
<enumerator name='ZPOOL_STATUS_COMPATIBILITY_ERR' value='30'/>
<enumerator name='ZPOOL_STATUS_INCOMPATIBLE_FEAT' value='31'/>
<enumerator name='ZPOOL_STATUS_OK' value='32'/>
</enum-decl>
<typedef-decl name='zpool_status_t' type-id='5e770b40' id='d3dd6294'/>
<enum-decl name='zpool_compat_status_t' naming-typedef-id='901b78d1' id='20676925'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_COMPATIBILITY_OK' value='0'/>
<enumerator name='ZPOOL_COMPATIBILITY_WARNTOKEN' value='1'/>
<enumerator name='ZPOOL_COMPATIBILITY_BADTOKEN' value='2'/>
<enumerator name='ZPOOL_COMPATIBILITY_BADFILE' value='3'/>
<enumerator name='ZPOOL_COMPATIBILITY_NOFILES' value='4'/>
</enum-decl>
<typedef-decl name='zpool_compat_status_t' type-id='20676925' id='901b78d1'/>
<enum-decl name='vdev_prop_t' naming-typedef-id='5aa5c90c' id='1573bec8'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='VDEV_PROP_INVAL' value='-1'/>
<enumerator name='VDEV_PROP_USERPROP' value='-1'/>
<enumerator name='VDEV_PROP_NAME' value='0'/>
<enumerator name='VDEV_PROP_CAPACITY' value='1'/>
<enumerator name='VDEV_PROP_STATE' value='2'/>
<enumerator name='VDEV_PROP_GUID' value='3'/>
<enumerator name='VDEV_PROP_ASIZE' value='4'/>
<enumerator name='VDEV_PROP_PSIZE' value='5'/>
<enumerator name='VDEV_PROP_ASHIFT' value='6'/>
<enumerator name='VDEV_PROP_SIZE' value='7'/>
<enumerator name='VDEV_PROP_FREE' value='8'/>
<enumerator name='VDEV_PROP_ALLOCATED' value='9'/>
<enumerator name='VDEV_PROP_COMMENT' value='10'/>
<enumerator name='VDEV_PROP_EXPANDSZ' value='11'/>
<enumerator name='VDEV_PROP_FRAGMENTATION' value='12'/>
<enumerator name='VDEV_PROP_BOOTSIZE' value='13'/>
<enumerator name='VDEV_PROP_PARITY' value='14'/>
<enumerator name='VDEV_PROP_PATH' value='15'/>
<enumerator name='VDEV_PROP_DEVID' value='16'/>
<enumerator name='VDEV_PROP_PHYS_PATH' value='17'/>
<enumerator name='VDEV_PROP_ENC_PATH' value='18'/>
<enumerator name='VDEV_PROP_FRU' value='19'/>
<enumerator name='VDEV_PROP_PARENT' value='20'/>
<enumerator name='VDEV_PROP_CHILDREN' value='21'/>
<enumerator name='VDEV_PROP_NUMCHILDREN' value='22'/>
<enumerator name='VDEV_PROP_READ_ERRORS' value='23'/>
<enumerator name='VDEV_PROP_WRITE_ERRORS' value='24'/>
<enumerator name='VDEV_PROP_CHECKSUM_ERRORS' value='25'/>
<enumerator name='VDEV_PROP_INITIALIZE_ERRORS' value='26'/>
<enumerator name='VDEV_PROP_OPS_NULL' value='27'/>
<enumerator name='VDEV_PROP_OPS_READ' value='28'/>
<enumerator name='VDEV_PROP_OPS_WRITE' value='29'/>
<enumerator name='VDEV_PROP_OPS_FREE' value='30'/>
<enumerator name='VDEV_PROP_OPS_CLAIM' value='31'/>
<enumerator name='VDEV_PROP_OPS_TRIM' value='32'/>
<enumerator name='VDEV_PROP_BYTES_NULL' value='33'/>
<enumerator name='VDEV_PROP_BYTES_READ' value='34'/>
<enumerator name='VDEV_PROP_BYTES_WRITE' value='35'/>
<enumerator name='VDEV_PROP_BYTES_FREE' value='36'/>
<enumerator name='VDEV_PROP_BYTES_CLAIM' value='37'/>
<enumerator name='VDEV_PROP_BYTES_TRIM' value='38'/>
<enumerator name='VDEV_PROP_REMOVING' value='39'/>
<enumerator name='VDEV_PROP_ALLOCATING' value='40'/>
<enumerator name='VDEV_PROP_FAILFAST' value='41'/>
<enumerator name='VDEV_PROP_CHECKSUM_N' value='42'/>
<enumerator name='VDEV_PROP_CHECKSUM_T' value='43'/>
<enumerator name='VDEV_PROP_IO_N' value='44'/>
<enumerator name='VDEV_PROP_IO_T' value='45'/>
- <enumerator name='VDEV_NUM_PROPS' value='46'/>
+ <enumerator name='VDEV_PROP_RAIDZ_EXPANDING' value='46'/>
+ <enumerator name='VDEV_NUM_PROPS' value='47'/>
</enum-decl>
<typedef-decl name='vdev_prop_t' type-id='1573bec8' id='5aa5c90c'/>
<class-decl name='zpool_load_policy' size-in-bits='256' is-struct='yes' visibility='default' id='2f65b36f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zlp_rewind' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zlp_maxmeta' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zlp_maxdata' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='zlp_txg' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zpool_load_policy_t' type-id='2f65b36f' id='d11b7617'/>
<enum-decl name='vdev_state' id='21566197'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='VDEV_STATE_UNKNOWN' value='0'/>
<enumerator name='VDEV_STATE_CLOSED' value='1'/>
<enumerator name='VDEV_STATE_OFFLINE' value='2'/>
<enumerator name='VDEV_STATE_REMOVED' value='3'/>
<enumerator name='VDEV_STATE_CANT_OPEN' value='4'/>
<enumerator name='VDEV_STATE_FAULTED' value='5'/>
<enumerator name='VDEV_STATE_DEGRADED' value='6'/>
<enumerator name='VDEV_STATE_HEALTHY' value='7'/>
</enum-decl>
<typedef-decl name='vdev_state_t' type-id='21566197' id='35acf840'/>
<enum-decl name='vdev_aux' id='7f5bcca4'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='VDEV_AUX_NONE' value='0'/>
<enumerator name='VDEV_AUX_OPEN_FAILED' value='1'/>
<enumerator name='VDEV_AUX_CORRUPT_DATA' value='2'/>
<enumerator name='VDEV_AUX_NO_REPLICAS' value='3'/>
<enumerator name='VDEV_AUX_BAD_GUID_SUM' value='4'/>
<enumerator name='VDEV_AUX_TOO_SMALL' value='5'/>
<enumerator name='VDEV_AUX_BAD_LABEL' value='6'/>
<enumerator name='VDEV_AUX_VERSION_NEWER' value='7'/>
<enumerator name='VDEV_AUX_VERSION_OLDER' value='8'/>
<enumerator name='VDEV_AUX_UNSUP_FEAT' value='9'/>
<enumerator name='VDEV_AUX_SPARED' value='10'/>
<enumerator name='VDEV_AUX_ERR_EXCEEDED' value='11'/>
<enumerator name='VDEV_AUX_IO_FAILURE' value='12'/>
<enumerator name='VDEV_AUX_BAD_LOG' value='13'/>
<enumerator name='VDEV_AUX_EXTERNAL' value='14'/>
<enumerator name='VDEV_AUX_SPLIT_POOL' value='15'/>
<enumerator name='VDEV_AUX_BAD_ASHIFT' value='16'/>
<enumerator name='VDEV_AUX_EXTERNAL_PERSIST' value='17'/>
<enumerator name='VDEV_AUX_ACTIVE' value='18'/>
<enumerator name='VDEV_AUX_CHILDREN_OFFLINE' value='19'/>
<enumerator name='VDEV_AUX_ASHIFT_TOO_BIG' value='20'/>
</enum-decl>
<typedef-decl name='vdev_aux_t' type-id='7f5bcca4' id='9d774e0b'/>
<enum-decl name='pool_scan_func' id='1b092565'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_SCAN_NONE' value='0'/>
<enumerator name='POOL_SCAN_SCRUB' value='1'/>
<enumerator name='POOL_SCAN_RESILVER' value='2'/>
<enumerator name='POOL_SCAN_ERRORSCRUB' value='3'/>
<enumerator name='POOL_SCAN_FUNCS' value='4'/>
</enum-decl>
<typedef-decl name='pool_scan_func_t' type-id='1b092565' id='7313fbe2'/>
<enum-decl name='pool_scrub_cmd' id='a1474cbd'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_SCRUB_NORMAL' value='0'/>
<enumerator name='POOL_SCRUB_PAUSE' value='1'/>
<enumerator name='POOL_SCRUB_FLAGS_END' value='2'/>
</enum-decl>
<typedef-decl name='pool_scrub_cmd_t' type-id='a1474cbd' id='b51cf3c2'/>
<enum-decl name='zpool_errata' id='d9abbf54'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_ERRATA_NONE' value='0'/>
<enumerator name='ZPOOL_ERRATA_ZOL_2094_SCRUB' value='1'/>
<enumerator name='ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY' value='2'/>
<enumerator name='ZPOOL_ERRATA_ZOL_6845_ENCRYPTION' value='3'/>
<enumerator name='ZPOOL_ERRATA_ZOL_8308_ENCRYPTION' value='4'/>
</enum-decl>
<typedef-decl name='zpool_errata_t' type-id='d9abbf54' id='688c495b'/>
<enum-decl name='pool_initialize_func' id='5c246ad4'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_INITIALIZE_START' value='0'/>
<enumerator name='POOL_INITIALIZE_CANCEL' value='1'/>
<enumerator name='POOL_INITIALIZE_SUSPEND' value='2'/>
<enumerator name='POOL_INITIALIZE_UNINIT' value='3'/>
<enumerator name='POOL_INITIALIZE_FUNCS' value='4'/>
</enum-decl>
<typedef-decl name='pool_initialize_func_t' type-id='5c246ad4' id='7063e1ab'/>
<enum-decl name='pool_trim_func' id='54ed608a'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_TRIM_START' value='0'/>
<enumerator name='POOL_TRIM_CANCEL' value='1'/>
<enumerator name='POOL_TRIM_SUSPEND' value='2'/>
<enumerator name='POOL_TRIM_FUNCS' value='3'/>
</enum-decl>
<typedef-decl name='pool_trim_func_t' type-id='54ed608a' id='b1146b8d'/>
+ <enum-decl name='zfs_ioc' id='12033f13'>
+ <underlying-type type-id='9cac1fee'/>
+ <enumerator name='ZFS_IOC_FIRST' value='23040'/>
+ <enumerator name='ZFS_IOC' value='23040'/>
+ <enumerator name='ZFS_IOC_POOL_CREATE' value='23040'/>
+ <enumerator name='ZFS_IOC_POOL_DESTROY' value='23041'/>
+ <enumerator name='ZFS_IOC_POOL_IMPORT' value='23042'/>
+ <enumerator name='ZFS_IOC_POOL_EXPORT' value='23043'/>
+ <enumerator name='ZFS_IOC_POOL_CONFIGS' value='23044'/>
+ <enumerator name='ZFS_IOC_POOL_STATS' value='23045'/>
+ <enumerator name='ZFS_IOC_POOL_TRYIMPORT' value='23046'/>
+ <enumerator name='ZFS_IOC_POOL_SCAN' value='23047'/>
+ <enumerator name='ZFS_IOC_POOL_FREEZE' value='23048'/>
+ <enumerator name='ZFS_IOC_POOL_UPGRADE' value='23049'/>
+ <enumerator name='ZFS_IOC_POOL_GET_HISTORY' value='23050'/>
+ <enumerator name='ZFS_IOC_VDEV_ADD' value='23051'/>
+ <enumerator name='ZFS_IOC_VDEV_REMOVE' value='23052'/>
+ <enumerator name='ZFS_IOC_VDEV_SET_STATE' value='23053'/>
+ <enumerator name='ZFS_IOC_VDEV_ATTACH' value='23054'/>
+ <enumerator name='ZFS_IOC_VDEV_DETACH' value='23055'/>
+ <enumerator name='ZFS_IOC_VDEV_SETPATH' value='23056'/>
+ <enumerator name='ZFS_IOC_VDEV_SETFRU' value='23057'/>
+ <enumerator name='ZFS_IOC_OBJSET_STATS' value='23058'/>
+ <enumerator name='ZFS_IOC_OBJSET_ZPLPROPS' value='23059'/>
+ <enumerator name='ZFS_IOC_DATASET_LIST_NEXT' value='23060'/>
+ <enumerator name='ZFS_IOC_SNAPSHOT_LIST_NEXT' value='23061'/>
+ <enumerator name='ZFS_IOC_SET_PROP' value='23062'/>
+ <enumerator name='ZFS_IOC_CREATE' value='23063'/>
+ <enumerator name='ZFS_IOC_DESTROY' value='23064'/>
+ <enumerator name='ZFS_IOC_ROLLBACK' value='23065'/>
+ <enumerator name='ZFS_IOC_RENAME' value='23066'/>
+ <enumerator name='ZFS_IOC_RECV' value='23067'/>
+ <enumerator name='ZFS_IOC_SEND' value='23068'/>
+ <enumerator name='ZFS_IOC_INJECT_FAULT' value='23069'/>
+ <enumerator name='ZFS_IOC_CLEAR_FAULT' value='23070'/>
+ <enumerator name='ZFS_IOC_INJECT_LIST_NEXT' value='23071'/>
+ <enumerator name='ZFS_IOC_ERROR_LOG' value='23072'/>
+ <enumerator name='ZFS_IOC_CLEAR' value='23073'/>
+ <enumerator name='ZFS_IOC_PROMOTE' value='23074'/>
+ <enumerator name='ZFS_IOC_SNAPSHOT' value='23075'/>
+ <enumerator name='ZFS_IOC_DSOBJ_TO_DSNAME' value='23076'/>
+ <enumerator name='ZFS_IOC_OBJ_TO_PATH' value='23077'/>
+ <enumerator name='ZFS_IOC_POOL_SET_PROPS' value='23078'/>
+ <enumerator name='ZFS_IOC_POOL_GET_PROPS' value='23079'/>
+ <enumerator name='ZFS_IOC_SET_FSACL' value='23080'/>
+ <enumerator name='ZFS_IOC_GET_FSACL' value='23081'/>
+ <enumerator name='ZFS_IOC_SHARE' value='23082'/>
+ <enumerator name='ZFS_IOC_INHERIT_PROP' value='23083'/>
+ <enumerator name='ZFS_IOC_SMB_ACL' value='23084'/>
+ <enumerator name='ZFS_IOC_USERSPACE_ONE' value='23085'/>
+ <enumerator name='ZFS_IOC_USERSPACE_MANY' value='23086'/>
+ <enumerator name='ZFS_IOC_USERSPACE_UPGRADE' value='23087'/>
+ <enumerator name='ZFS_IOC_HOLD' value='23088'/>
+ <enumerator name='ZFS_IOC_RELEASE' value='23089'/>
+ <enumerator name='ZFS_IOC_GET_HOLDS' value='23090'/>
+ <enumerator name='ZFS_IOC_OBJSET_RECVD_PROPS' value='23091'/>
+ <enumerator name='ZFS_IOC_VDEV_SPLIT' value='23092'/>
+ <enumerator name='ZFS_IOC_NEXT_OBJ' value='23093'/>
+ <enumerator name='ZFS_IOC_DIFF' value='23094'/>
+ <enumerator name='ZFS_IOC_TMP_SNAPSHOT' value='23095'/>
+ <enumerator name='ZFS_IOC_OBJ_TO_STATS' value='23096'/>
+ <enumerator name='ZFS_IOC_SPACE_WRITTEN' value='23097'/>
+ <enumerator name='ZFS_IOC_SPACE_SNAPS' value='23098'/>
+ <enumerator name='ZFS_IOC_DESTROY_SNAPS' value='23099'/>
+ <enumerator name='ZFS_IOC_POOL_REGUID' value='23100'/>
+ <enumerator name='ZFS_IOC_POOL_REOPEN' value='23101'/>
+ <enumerator name='ZFS_IOC_SEND_PROGRESS' value='23102'/>
+ <enumerator name='ZFS_IOC_LOG_HISTORY' value='23103'/>
+ <enumerator name='ZFS_IOC_SEND_NEW' value='23104'/>
+ <enumerator name='ZFS_IOC_SEND_SPACE' value='23105'/>
+ <enumerator name='ZFS_IOC_CLONE' value='23106'/>
+ <enumerator name='ZFS_IOC_BOOKMARK' value='23107'/>
+ <enumerator name='ZFS_IOC_GET_BOOKMARKS' value='23108'/>
+ <enumerator name='ZFS_IOC_DESTROY_BOOKMARKS' value='23109'/>
+ <enumerator name='ZFS_IOC_RECV_NEW' value='23110'/>
+ <enumerator name='ZFS_IOC_POOL_SYNC' value='23111'/>
+ <enumerator name='ZFS_IOC_CHANNEL_PROGRAM' value='23112'/>
+ <enumerator name='ZFS_IOC_LOAD_KEY' value='23113'/>
+ <enumerator name='ZFS_IOC_UNLOAD_KEY' value='23114'/>
+ <enumerator name='ZFS_IOC_CHANGE_KEY' value='23115'/>
+ <enumerator name='ZFS_IOC_REMAP' value='23116'/>
+ <enumerator name='ZFS_IOC_POOL_CHECKPOINT' value='23117'/>
+ <enumerator name='ZFS_IOC_POOL_DISCARD_CHECKPOINT' value='23118'/>
+ <enumerator name='ZFS_IOC_POOL_INITIALIZE' value='23119'/>
+ <enumerator name='ZFS_IOC_POOL_TRIM' value='23120'/>
+ <enumerator name='ZFS_IOC_REDACT' value='23121'/>
+ <enumerator name='ZFS_IOC_GET_BOOKMARK_PROPS' value='23122'/>
+ <enumerator name='ZFS_IOC_WAIT' value='23123'/>
+ <enumerator name='ZFS_IOC_WAIT_FS' value='23124'/>
+ <enumerator name='ZFS_IOC_VDEV_GET_PROPS' value='23125'/>
+ <enumerator name='ZFS_IOC_VDEV_SET_PROPS' value='23126'/>
+ <enumerator name='ZFS_IOC_POOL_SCRUB' value='23127'/>
+ <enumerator name='ZFS_IOC_PLATFORM' value='23168'/>
+ <enumerator name='ZFS_IOC_EVENTS_NEXT' value='23169'/>
+ <enumerator name='ZFS_IOC_EVENTS_CLEAR' value='23170'/>
+ <enumerator name='ZFS_IOC_EVENTS_SEEK' value='23171'/>
+ <enumerator name='ZFS_IOC_NEXTBOOT' value='23172'/>
+ <enumerator name='ZFS_IOC_JAIL' value='23173'/>
+ <enumerator name='ZFS_IOC_USERNS_ATTACH' value='23173'/>
+ <enumerator name='ZFS_IOC_UNJAIL' value='23174'/>
+ <enumerator name='ZFS_IOC_USERNS_DETACH' value='23174'/>
+ <enumerator name='ZFS_IOC_SET_BOOTENV' value='23175'/>
+ <enumerator name='ZFS_IOC_GET_BOOTENV' value='23176'/>
+ <enumerator name='ZFS_IOC_LAST' value='23177'/>
+ </enum-decl>
+ <typedef-decl name='zfs_ioc_t' type-id='12033f13' id='5b35941c'/>
<enum-decl name='zpool_wait_activity_t' naming-typedef-id='73446457' id='849338e3'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_WAIT_CKPT_DISCARD' value='0'/>
<enumerator name='ZPOOL_WAIT_FREE' value='1'/>
<enumerator name='ZPOOL_WAIT_INITIALIZE' value='2'/>
<enumerator name='ZPOOL_WAIT_REPLACE' value='3'/>
<enumerator name='ZPOOL_WAIT_REMOVE' value='4'/>
<enumerator name='ZPOOL_WAIT_RESILVER' value='5'/>
<enumerator name='ZPOOL_WAIT_SCRUB' value='6'/>
<enumerator name='ZPOOL_WAIT_TRIM' value='7'/>
- <enumerator name='ZPOOL_WAIT_NUM_ACTIVITIES' value='8'/>
+ <enumerator name='ZPOOL_WAIT_RAIDZ_EXPAND' value='8'/>
+ <enumerator name='ZPOOL_WAIT_NUM_ACTIVITIES' value='9'/>
</enum-decl>
<typedef-decl name='zpool_wait_activity_t' type-id='849338e3' id='73446457'/>
<enum-decl name='spa_feature' id='33ecb627'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='SPA_FEATURE_NONE' value='-1'/>
<enumerator name='SPA_FEATURE_ASYNC_DESTROY' value='0'/>
<enumerator name='SPA_FEATURE_EMPTY_BPOBJ' value='1'/>
<enumerator name='SPA_FEATURE_LZ4_COMPRESS' value='2'/>
<enumerator name='SPA_FEATURE_MULTI_VDEV_CRASH_DUMP' value='3'/>
<enumerator name='SPA_FEATURE_SPACEMAP_HISTOGRAM' value='4'/>
<enumerator name='SPA_FEATURE_ENABLED_TXG' value='5'/>
<enumerator name='SPA_FEATURE_HOLE_BIRTH' value='6'/>
<enumerator name='SPA_FEATURE_EXTENSIBLE_DATASET' value='7'/>
<enumerator name='SPA_FEATURE_EMBEDDED_DATA' value='8'/>
<enumerator name='SPA_FEATURE_BOOKMARKS' value='9'/>
<enumerator name='SPA_FEATURE_FS_SS_LIMIT' value='10'/>
<enumerator name='SPA_FEATURE_LARGE_BLOCKS' value='11'/>
<enumerator name='SPA_FEATURE_LARGE_DNODE' value='12'/>
<enumerator name='SPA_FEATURE_SHA512' value='13'/>
<enumerator name='SPA_FEATURE_SKEIN' value='14'/>
<enumerator name='SPA_FEATURE_EDONR' value='15'/>
<enumerator name='SPA_FEATURE_USEROBJ_ACCOUNTING' value='16'/>
<enumerator name='SPA_FEATURE_ENCRYPTION' value='17'/>
<enumerator name='SPA_FEATURE_PROJECT_QUOTA' value='18'/>
<enumerator name='SPA_FEATURE_DEVICE_REMOVAL' value='19'/>
<enumerator name='SPA_FEATURE_OBSOLETE_COUNTS' value='20'/>
<enumerator name='SPA_FEATURE_POOL_CHECKPOINT' value='21'/>
<enumerator name='SPA_FEATURE_SPACEMAP_V2' value='22'/>
<enumerator name='SPA_FEATURE_ALLOCATION_CLASSES' value='23'/>
<enumerator name='SPA_FEATURE_RESILVER_DEFER' value='24'/>
<enumerator name='SPA_FEATURE_BOOKMARK_V2' value='25'/>
<enumerator name='SPA_FEATURE_REDACTION_BOOKMARKS' value='26'/>
<enumerator name='SPA_FEATURE_REDACTED_DATASETS' value='27'/>
<enumerator name='SPA_FEATURE_BOOKMARK_WRITTEN' value='28'/>
<enumerator name='SPA_FEATURE_LOG_SPACEMAP' value='29'/>
<enumerator name='SPA_FEATURE_LIVELIST' value='30'/>
<enumerator name='SPA_FEATURE_DEVICE_REBUILD' value='31'/>
<enumerator name='SPA_FEATURE_ZSTD_COMPRESS' value='32'/>
<enumerator name='SPA_FEATURE_DRAID' value='33'/>
<enumerator name='SPA_FEATURE_ZILSAXATTR' value='34'/>
<enumerator name='SPA_FEATURE_HEAD_ERRLOG' value='35'/>
<enumerator name='SPA_FEATURE_BLAKE3' value='36'/>
<enumerator name='SPA_FEATURE_BLOCK_CLONING' value='37'/>
<enumerator name='SPA_FEATURE_AVZ_V2' value='38'/>
<enumerator name='SPA_FEATURE_REDACTION_LIST_SPILL' value='39'/>
- <enumerator name='SPA_FEATURES' value='40'/>
+ <enumerator name='SPA_FEATURE_RAIDZ_EXPANSION' value='40'/>
+ <enumerator name='SPA_FEATURES' value='41'/>
</enum-decl>
<typedef-decl name='spa_feature_t' type-id='33ecb627' id='d6618c78'/>
<qualified-type-def type-id='22cce67b' const='yes' id='d2816df0'/>
<pointer-type-def type-id='d2816df0' size-in-bits='64' id='3bbfee2e'/>
<qualified-type-def type-id='b96825af' const='yes' id='2b61797f'/>
<pointer-type-def type-id='2b61797f' size-in-bits='64' id='9f7200cf'/>
<pointer-type-def type-id='d6618c78' size-in-bits='64' id='a8425263'/>
<qualified-type-def type-id='62f7a03d' restrict='yes' id='f1cadedf'/>
<pointer-type-def type-id='a093cbb8' size-in-bits='64' id='b13f38c3'/>
<pointer-type-def type-id='35acf840' size-in-bits='64' id='17f3480d'/>
<pointer-type-def type-id='688c495b' size-in-bits='64' id='cec6f2e4'/>
<pointer-type-def type-id='d11b7617' size-in-bits='64' id='23432aaa'/>
<function-decl name='zpool_get_handle' mangled-name='zpool_get_handle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_handle'>
<parameter type-id='4c81de99'/>
<return type-id='b0382bb3'/>
</function-decl>
<function-decl name='zpool_prop_to_name' mangled-name='zpool_prop_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_to_name'>
<parameter type-id='5d0c23fb'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='vdev_prop_to_name' mangled-name='vdev_prop_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_to_name'>
<parameter type-id='5aa5c90c'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='vdev_prop_user' mangled-name='vdev_prop_user' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_user'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_get_status' mangled-name='zpool_get_status' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_status'>
<parameter type-id='4c81de99'/>
<parameter type-id='7d3cd834'/>
<parameter type-id='cec6f2e4'/>
<return type-id='d3dd6294'/>
</function-decl>
<function-decl name='zpool_prop_default_string' mangled-name='zpool_prop_default_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_default_string'>
<parameter type-id='5d0c23fb'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_prop_default_numeric' mangled-name='zpool_prop_default_numeric' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_default_numeric'>
<parameter type-id='5d0c23fb'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='libzfs_envvar_is_set' mangled-name='libzfs_envvar_is_set' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_envvar_is_set'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='lzc_initialize' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='7063e1ab'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_trim' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='b1146b8d'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_sync' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_reopen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_pool_checkpoint' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_pool_checkpoint_discard' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_wait' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='73446457'/>
<parameter type-id='37e3bd22'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_wait_tag' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='73446457'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='37e3bd22'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_set_bootenv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='22cce67b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_bootenv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_vdev_prop' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_set_vdev_prop' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
+ <function-decl name='lzc_scrub' visibility='default' binding='global' size-in-bits='64'>
+ <parameter type-id='5b35941c'/>
+ <parameter type-id='80f4b756'/>
+ <parameter type-id='5ce45b60'/>
+ <parameter type-id='857bb57e'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
<function-decl name='zfs_resolve_shortname' mangled-name='zfs_resolve_shortname' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_resolve_shortname'>
<parameter type-id='80f4b756'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_strip_partition' mangled-name='zfs_strip_partition' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_strip_partition'>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zfs_strip_path' mangled-name='zfs_strip_path' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_strip_path'>
<parameter type-id='80f4b756'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_strcmp_pathname' mangled-name='zfs_strcmp_pathname' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_strcmp_pathname'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_history_unpack' mangled-name='zpool_history_unpack' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_history_unpack'>
<parameter type-id='26a90f95'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='75be733c'/>
<parameter type-id='4dd26a40'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_basename' mangled-name='zfs_basename' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_basename'>
<parameter type-id='80f4b756'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_name_to_prop' mangled-name='zpool_name_to_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_name_to_prop'>
<parameter type-id='80f4b756'/>
<return type-id='5d0c23fb'/>
</function-decl>
<function-decl name='zpool_prop_readonly' mangled-name='zpool_prop_readonly' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_readonly'>
<parameter type-id='5d0c23fb'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_prop_setonce' mangled-name='zpool_prop_setonce' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_setonce'>
<parameter type-id='5d0c23fb'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_prop_feature' mangled-name='zpool_prop_feature' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_feature'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_prop_index_to_string' mangled-name='zpool_prop_index_to_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_index_to_string'>
<parameter type-id='5d0c23fb'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='vdev_name_to_prop' mangled-name='vdev_name_to_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_name_to_prop'>
<parameter type-id='80f4b756'/>
<return type-id='5aa5c90c'/>
</function-decl>
<function-decl name='vdev_prop_default_string' mangled-name='vdev_prop_default_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_default_string'>
<parameter type-id='5aa5c90c'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='vdev_prop_default_numeric' mangled-name='vdev_prop_default_numeric' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_default_numeric'>
<parameter type-id='5aa5c90c'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='vdev_prop_readonly' mangled-name='vdev_prop_readonly' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_readonly'>
<parameter type-id='5aa5c90c'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='vdev_prop_index_to_string' mangled-name='vdev_prop_index_to_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_index_to_string'>
<parameter type-id='5aa5c90c'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_prop_vdev' mangled-name='zpool_prop_vdev' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_vdev'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='nvlist_add_nvpair' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='3fa542f0'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_add_uint8_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9f7200cf'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_add_nvlist_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='3bbfee2e'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvpair_value_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='3fa542f0'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fnvlist_add_boolean_value' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_int64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9da381c4'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_nvlist_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='3bbfee2e'/>
<parameter type-id='3502e3ff'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_lookup_uint64_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='4dd26a40'/>
<return type-id='5d6479ae'/>
</function-decl>
<function-decl name='fnvpair_value_int64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='9da381c4'/>
</function-decl>
<function-decl name='fnvpair_value_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfeature_is_supported' mangled-name='zfeature_is_supported' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfeature_is_supported'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfeature_lookup_guid' mangled-name='zfeature_lookup_guid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfeature_lookup_guid'>
<parameter type-id='80f4b756'/>
<parameter type-id='a8425263'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfeature_lookup_name' mangled-name='zfeature_lookup_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfeature_lookup_name'>
<parameter type-id='80f4b756'/>
<parameter type-id='a8425263'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_load_policy' mangled-name='zpool_get_load_policy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_load_policy'>
<parameter type-id='5ce45b60'/>
<parameter type-id='23432aaa'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='pool_namecheck' mangled-name='pool_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='pool_namecheck'>
<parameter type-id='80f4b756'/>
<parameter type-id='053457bd'/>
<parameter type-id='26a90f95'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_prop_get_type' mangled-name='zpool_prop_get_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_get_type'>
<parameter type-id='5d0c23fb'/>
<return type-id='31429eff'/>
</function-decl>
<function-decl name='vdev_prop_get_type' mangled-name='vdev_prop_get_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_get_type'>
<parameter type-id='5aa5c90c'/>
<return type-id='31429eff'/>
</function-decl>
<function-decl name='get_system_hostid' mangled-name='get_system_hostid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='get_system_hostid'>
<return type-id='7359adad'/>
</function-decl>
<function-decl name='strtoull' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<return type-id='3a47d82b'/>
</function-decl>
<function-decl name='memcmp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='strtok_r' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='266fe297'/>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='__realpath_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='266fe297'/>
<parameter type-id='b59d7dce'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='munmap' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='stat64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='f1cadedf'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_standard_error' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_standard_error_fmt' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_relabel_disk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_props_refresh' mangled-name='zpool_props_refresh' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_props_refresh'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_state_to_name' mangled-name='zpool_state_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_state_to_name'>
<parameter type-id='35acf840' name='state'/>
<parameter type-id='9d774e0b' name='aux'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_pool_state_to_name' mangled-name='zpool_pool_state_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_pool_state_to_name'>
<parameter type-id='084a08a3' name='state'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_get_state_str' mangled-name='zpool_get_state_str' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_state_str'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_get_userprop' mangled-name='zpool_get_userprop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_userprop'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='len'/>
<parameter type-id='debc6aa3' name='srctype'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_set_prop' mangled-name='zpool_set_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_set_prop'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='80f4b756' name='propval'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_expand_proplist' mangled-name='zpool_expand_proplist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_expand_proplist'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='e4378506' name='plp'/>
<parameter type-id='2e45de5d' name='type'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='vdev_expand_proplist' mangled-name='vdev_expand_proplist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_expand_proplist'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='vdevname'/>
<parameter type-id='e4378506' name='plp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_state' mangled-name='zpool_get_state' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_state'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_is_draid_spare' mangled-name='zpool_is_draid_spare' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_is_draid_spare'>
<parameter type-id='80f4b756' name='name'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_create' mangled-name='zpool_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_create'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='pool'/>
<parameter type-id='5ce45b60' name='nvroot'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='5ce45b60' name='fsprops'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_destroy' mangled-name='zpool_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_destroy'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='log_str'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_checkpoint' mangled-name='zpool_checkpoint' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_checkpoint'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_discard_checkpoint' mangled-name='zpool_discard_checkpoint' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_discard_checkpoint'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_add' mangled-name='zpool_add' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_add'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='5ce45b60' name='nvroot'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_export' mangled-name='zpool_export' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_export'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='c19b74c3' name='force'/>
<parameter type-id='80f4b756' name='log_str'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_export_force' mangled-name='zpool_export_force' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_export_force'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='log_str'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_explain_recover' mangled-name='zpool_explain_recover' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_explain_recover'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='name'/>
<parameter type-id='95e97e5e' name='reason'/>
<parameter type-id='5ce45b60' name='config'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_import' mangled-name='zpool_import' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_import'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='5ce45b60' name='config'/>
<parameter type-id='80f4b756' name='newname'/>
<parameter type-id='26a90f95' name='altroot'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_print_unsup_feat' mangled-name='zpool_print_unsup_feat' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_print_unsup_feat'>
<parameter type-id='5ce45b60' name='config'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_import_props' mangled-name='zpool_import_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_import_props'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='5ce45b60' name='config'/>
<parameter type-id='80f4b756' name='newname'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_initialize' mangled-name='zpool_initialize' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_initialize'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='7063e1ab' name='cmd_type'/>
<parameter type-id='5ce45b60' name='vds'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_initialize_wait' mangled-name='zpool_initialize_wait' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_initialize_wait'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='7063e1ab' name='cmd_type'/>
<parameter type-id='5ce45b60' name='vds'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_trim' mangled-name='zpool_trim' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_trim'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='b1146b8d' name='cmd_type'/>
<parameter type-id='5ce45b60' name='vds'/>
<parameter type-id='b13f38c3' name='trim_flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_scan' mangled-name='zpool_scan' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_scan'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='7313fbe2' name='func'/>
<parameter type-id='b51cf3c2' name='cmd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_find_vdev_by_physpath' mangled-name='zpool_find_vdev_by_physpath' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_find_vdev_by_physpath'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='ppath'/>
<parameter type-id='37e3bd22' name='avail_spare'/>
<parameter type-id='37e3bd22' name='l2cache'/>
<parameter type-id='37e3bd22' name='log'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zpool_find_vdev' mangled-name='zpool_find_vdev' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_find_vdev'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='37e3bd22' name='avail_spare'/>
<parameter type-id='37e3bd22' name='l2cache'/>
<parameter type-id='37e3bd22' name='log'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zpool_vdev_path_to_guid' mangled-name='zpool_vdev_path_to_guid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_path_to_guid'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zpool_vdev_online' mangled-name='zpool_vdev_online' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_online'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='95e97e5e' name='flags'/>
<parameter type-id='17f3480d' name='newstate'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_offline' mangled-name='zpool_vdev_offline' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_offline'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='c19b74c3' name='istmp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_remove_wanted' mangled-name='zpool_vdev_remove_wanted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_remove_wanted'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_fault' mangled-name='zpool_vdev_fault' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_fault'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='guid'/>
<parameter type-id='9d774e0b' name='aux'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_degrade' mangled-name='zpool_vdev_degrade' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_degrade'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='guid'/>
<parameter type-id='9d774e0b' name='aux'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_attach' mangled-name='zpool_vdev_attach' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_attach'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='old_disk'/>
<parameter type-id='80f4b756' name='new_disk'/>
<parameter type-id='5ce45b60' name='nvroot'/>
<parameter type-id='95e97e5e' name='replacing'/>
<parameter type-id='c19b74c3' name='rebuild'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_detach' mangled-name='zpool_vdev_detach' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_detach'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_split' mangled-name='zpool_vdev_split' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_split'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='26a90f95' name='newname'/>
<parameter type-id='857bb57e' name='newroot'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='325c1e34' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_remove' mangled-name='zpool_vdev_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_remove'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_remove_cancel' mangled-name='zpool_vdev_remove_cancel' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_remove_cancel'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_indirect_size' mangled-name='zpool_vdev_indirect_size' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_indirect_size'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='5d6479ae' name='sizep'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_clear' mangled-name='zpool_clear' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_clear'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='5ce45b60' name='rewindnvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_clear' mangled-name='zpool_vdev_clear' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_clear'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='guid'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_reguid' mangled-name='zpool_reguid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_reguid'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_reopen_one' mangled-name='zpool_reopen_one' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_reopen_one'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_sync_one' mangled-name='zpool_sync_one' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_sync_one'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_name' mangled-name='zpool_vdev_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_name'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='5ce45b60' name='nv'/>
<parameter type-id='95e97e5e' name='name_flags'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zpool_get_errlog' mangled-name='zpool_get_errlog' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_errlog'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='857bb57e' name='nverrlistp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_upgrade' mangled-name='zpool_upgrade' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_upgrade'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='new_version'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_save_arguments' mangled-name='zfs_save_arguments' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_save_arguments'>
<parameter type-id='95e97e5e' name='argc'/>
<parameter type-id='9b23c9ad' name='argv'/>
<parameter type-id='26a90f95' name='string'/>
<parameter type-id='95e97e5e' name='len'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_log_history' mangled-name='zpool_log_history' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_log_history'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='message'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_history' mangled-name='zpool_get_history' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_history'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='857bb57e' name='nvhisp'/>
<parameter type-id='5d6479ae' name='off'/>
<parameter type-id='37e3bd22' name='eof'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_events_next' mangled-name='zpool_events_next' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_events_next'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='857bb57e' name='nvp'/>
<parameter type-id='7292109c' name='dropped'/>
<parameter type-id='f0981eeb' name='flags'/>
<parameter type-id='95e97e5e' name='zevent_fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_events_clear' mangled-name='zpool_events_clear' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_events_clear'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='7292109c' name='count'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_events_seek' mangled-name='zpool_events_seek' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_events_seek'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='9c313c2d' name='eid'/>
<parameter type-id='95e97e5e' name='zevent_fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_obj_to_path' mangled-name='zpool_obj_to_path' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_obj_to_path'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='dsobj'/>
<parameter type-id='9c313c2d' name='obj'/>
<parameter type-id='26a90f95' name='pathname'/>
<parameter type-id='b59d7dce' name='len'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_obj_to_path_ds' mangled-name='zpool_obj_to_path_ds' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_obj_to_path_ds'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='dsobj'/>
<parameter type-id='9c313c2d' name='obj'/>
<parameter type-id='26a90f95' name='pathname'/>
<parameter type-id='b59d7dce' name='len'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_wait' mangled-name='zpool_wait' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_wait'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='73446457' name='activity'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_wait_status' mangled-name='zpool_wait_status' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_wait_status'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='73446457' name='activity'/>
<parameter type-id='37e3bd22' name='missing'/>
<parameter type-id='37e3bd22' name='waited'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_set_bootenv' mangled-name='zpool_set_bootenv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_set_bootenv'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='22cce67b' name='envmap'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_bootenv' mangled-name='zpool_get_bootenv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_bootenv'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='857bb57e' name='nvlp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_load_compat' mangled-name='zpool_load_compat' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_load_compat'>
<parameter type-id='80f4b756' name='compat'/>
<parameter type-id='37e3bd22' name='features'/>
<parameter type-id='26a90f95' name='report'/>
<parameter type-id='b59d7dce' name='rlen'/>
<return type-id='901b78d1'/>
</function-decl>
<function-decl name='zpool_get_vdev_prop_value' mangled-name='zpool_get_vdev_prop_value' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_vdev_prop_value'>
<parameter type-id='5ce45b60' name='nvprop'/>
<parameter type-id='5aa5c90c' name='prop'/>
<parameter type-id='26a90f95' name='prop_name'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='len'/>
<parameter type-id='debc6aa3' name='srctype'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_vdev_prop' mangled-name='zpool_get_vdev_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_vdev_prop'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='vdevname'/>
<parameter type-id='5aa5c90c' name='prop'/>
<parameter type-id='26a90f95' name='prop_name'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='len'/>
<parameter type-id='debc6aa3' name='srctype'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_all_vdev_props' mangled-name='zpool_get_all_vdev_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_all_vdev_props'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='vdevname'/>
<parameter type-id='857bb57e' name='outnvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_set_vdev_prop' mangled-name='zpool_set_vdev_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_set_vdev_prop'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='vdevname'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='80f4b756' name='propval'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_sendrecv.c' language='LANG_C99'>
+ <array-type-def dimensions='1' type-id='8901473c' size-in-bits='576' id='f5da478b'>
+ <subrange length='1' type-id='7359adad' id='52f813b4'/>
+ </array-type-def>
+ <array-type-def dimensions='1' type-id='95e97e5e' size-in-bits='384' id='73b82f0f'>
+ <subrange length='12' type-id='7359adad' id='84827bdc'/>
+ </array-type-def>
+ <array-type-def dimensions='1' type-id='bd54fe1a' size-in-bits='512' id='5d4efd44'>
+ <subrange length='8' type-id='7359adad' id='56e0c0b1'/>
+ </array-type-def>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='2176' id='8c2bcad1'>
<subrange length='34' type-id='7359adad' id='6a6a7e00'/>
</array-type-def>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='256' id='85c64d26'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='96' id='fa8ef949'>
<subrange length='12' type-id='7359adad' id='84827bdc'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='128' id='fa9986a5'>
<subrange length='16' type-id='7359adad' id='848d0938'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='40' id='0f4ddd0b'>
<subrange length='5' type-id='7359adad' id='53010e10'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='48' id='0f562bd0'>
<subrange length='6' type-id='7359adad' id='52fa524b'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='64' id='13339fda'>
<subrange length='8' type-id='7359adad' id='56e0c0b1'/>
</array-type-def>
+ <array-type-def dimensions='1' type-id='eaa32e2f' size-in-bits='256' id='209ef23f'>
+ <subrange length='4' type-id='7359adad' id='16fe7105'/>
+ </array-type-def>
<class-decl name='sendflags' size-in-bits='576' is-struct='yes' visibility='default' id='f6aa15be'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='verbosity' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='replicate' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='skipmissing' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='doall' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='fromorigin' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='pad' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='props' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='dryrun' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='parsable' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='progress' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='progressastitle' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='largeblock' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='embed_data' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='416'>
<var-decl name='compress' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='raw' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='backup' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='holds' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='544'>
<var-decl name='saved' type-id='c19b74c3' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='sendflags_t' type-id='f6aa15be' id='945467e6'/>
<typedef-decl name='snapfilter_cb_t' type-id='d2a5e211' id='3d3ffb69'/>
<class-decl name='recvflags' size-in-bits='448' is-struct='yes' visibility='default' id='34a384dc'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='verbose' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='isprefix' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='istail' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='dryrun' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='force' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='canmountoff' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='resumable' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='byteswap' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='nomount' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='holds' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='skipholds' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='domount' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='forceunmount' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='416'>
<var-decl name='heal' type-id='c19b74c3' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='recvflags_t' type-id='34a384dc' id='9e59d1d4'/>
<enum-decl name='lzc_send_flags' id='bfbd3c8e'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='LZC_SEND_FLAG_EMBED_DATA' value='1'/>
<enumerator name='LZC_SEND_FLAG_LARGE_BLOCK' value='2'/>
<enumerator name='LZC_SEND_FLAG_COMPRESS' value='4'/>
<enumerator name='LZC_SEND_FLAG_RAW' value='8'/>
<enumerator name='LZC_SEND_FLAG_SAVED' value='16'/>
</enum-decl>
<class-decl name='ddt_key' size-in-bits='320' is-struct='yes' visibility='default' id='e0a4a1cb'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='ddk_cksum' type-id='39730d0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='ddk_prop' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='ddt_key_t' type-id='e0a4a1cb' id='67f6d2cf'/>
<enum-decl name='dmu_object_type' id='04b3b0b9'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DMU_OT_NONE' value='0'/>
<enumerator name='DMU_OT_OBJECT_DIRECTORY' value='1'/>
<enumerator name='DMU_OT_OBJECT_ARRAY' value='2'/>
<enumerator name='DMU_OT_PACKED_NVLIST' value='3'/>
<enumerator name='DMU_OT_PACKED_NVLIST_SIZE' value='4'/>
<enumerator name='DMU_OT_BPOBJ' value='5'/>
<enumerator name='DMU_OT_BPOBJ_HDR' value='6'/>
<enumerator name='DMU_OT_SPACE_MAP_HEADER' value='7'/>
<enumerator name='DMU_OT_SPACE_MAP' value='8'/>
<enumerator name='DMU_OT_INTENT_LOG' value='9'/>
<enumerator name='DMU_OT_DNODE' value='10'/>
<enumerator name='DMU_OT_OBJSET' value='11'/>
<enumerator name='DMU_OT_DSL_DIR' value='12'/>
<enumerator name='DMU_OT_DSL_DIR_CHILD_MAP' value='13'/>
<enumerator name='DMU_OT_DSL_DS_SNAP_MAP' value='14'/>
<enumerator name='DMU_OT_DSL_PROPS' value='15'/>
<enumerator name='DMU_OT_DSL_DATASET' value='16'/>
<enumerator name='DMU_OT_ZNODE' value='17'/>
<enumerator name='DMU_OT_OLDACL' value='18'/>
<enumerator name='DMU_OT_PLAIN_FILE_CONTENTS' value='19'/>
<enumerator name='DMU_OT_DIRECTORY_CONTENTS' value='20'/>
<enumerator name='DMU_OT_MASTER_NODE' value='21'/>
<enumerator name='DMU_OT_UNLINKED_SET' value='22'/>
<enumerator name='DMU_OT_ZVOL' value='23'/>
<enumerator name='DMU_OT_ZVOL_PROP' value='24'/>
<enumerator name='DMU_OT_PLAIN_OTHER' value='25'/>
<enumerator name='DMU_OT_UINT64_OTHER' value='26'/>
<enumerator name='DMU_OT_ZAP_OTHER' value='27'/>
<enumerator name='DMU_OT_ERROR_LOG' value='28'/>
<enumerator name='DMU_OT_SPA_HISTORY' value='29'/>
<enumerator name='DMU_OT_SPA_HISTORY_OFFSETS' value='30'/>
<enumerator name='DMU_OT_POOL_PROPS' value='31'/>
<enumerator name='DMU_OT_DSL_PERMS' value='32'/>
<enumerator name='DMU_OT_ACL' value='33'/>
<enumerator name='DMU_OT_SYSACL' value='34'/>
<enumerator name='DMU_OT_FUID' value='35'/>
<enumerator name='DMU_OT_FUID_SIZE' value='36'/>
<enumerator name='DMU_OT_NEXT_CLONES' value='37'/>
<enumerator name='DMU_OT_SCAN_QUEUE' value='38'/>
<enumerator name='DMU_OT_USERGROUP_USED' value='39'/>
<enumerator name='DMU_OT_USERGROUP_QUOTA' value='40'/>
<enumerator name='DMU_OT_USERREFS' value='41'/>
<enumerator name='DMU_OT_DDT_ZAP' value='42'/>
<enumerator name='DMU_OT_DDT_STATS' value='43'/>
<enumerator name='DMU_OT_SA' value='44'/>
<enumerator name='DMU_OT_SA_MASTER_NODE' value='45'/>
<enumerator name='DMU_OT_SA_ATTR_REGISTRATION' value='46'/>
<enumerator name='DMU_OT_SA_ATTR_LAYOUTS' value='47'/>
<enumerator name='DMU_OT_SCAN_XLATE' value='48'/>
<enumerator name='DMU_OT_DEDUP' value='49'/>
<enumerator name='DMU_OT_DEADLIST' value='50'/>
<enumerator name='DMU_OT_DEADLIST_HDR' value='51'/>
<enumerator name='DMU_OT_DSL_CLONES' value='52'/>
<enumerator name='DMU_OT_BPOBJ_SUBOBJ' value='53'/>
<enumerator name='DMU_OT_NUMTYPES' value='54'/>
<enumerator name='DMU_OTN_UINT8_DATA' value='128'/>
<enumerator name='DMU_OTN_UINT8_METADATA' value='192'/>
<enumerator name='DMU_OTN_UINT16_DATA' value='129'/>
<enumerator name='DMU_OTN_UINT16_METADATA' value='193'/>
<enumerator name='DMU_OTN_UINT32_DATA' value='130'/>
<enumerator name='DMU_OTN_UINT32_METADATA' value='194'/>
<enumerator name='DMU_OTN_UINT64_DATA' value='131'/>
<enumerator name='DMU_OTN_UINT64_METADATA' value='195'/>
<enumerator name='DMU_OTN_ZAP_DATA' value='132'/>
<enumerator name='DMU_OTN_ZAP_METADATA' value='196'/>
<enumerator name='DMU_OTN_UINT8_ENC_DATA' value='160'/>
<enumerator name='DMU_OTN_UINT8_ENC_METADATA' value='224'/>
<enumerator name='DMU_OTN_UINT16_ENC_DATA' value='161'/>
<enumerator name='DMU_OTN_UINT16_ENC_METADATA' value='225'/>
<enumerator name='DMU_OTN_UINT32_ENC_DATA' value='162'/>
<enumerator name='DMU_OTN_UINT32_ENC_METADATA' value='226'/>
<enumerator name='DMU_OTN_UINT64_ENC_DATA' value='163'/>
<enumerator name='DMU_OTN_UINT64_ENC_METADATA' value='227'/>
<enumerator name='DMU_OTN_ZAP_ENC_DATA' value='164'/>
<enumerator name='DMU_OTN_ZAP_ENC_METADATA' value='228'/>
</enum-decl>
<typedef-decl name='dmu_object_type_t' type-id='04b3b0b9' id='5c9d8906'/>
<class-decl name='zio_cksum' size-in-bits='256' is-struct='yes' visibility='default' id='1d53e28b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zc_word' type-id='85c64d26' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zio_cksum_t' type-id='1d53e28b' id='39730d0b'/>
<class-decl name='dmu_replay_record' size-in-bits='2496' is-struct='yes' visibility='default' id='781a52d7'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_type' type-id='08f5ca17' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='drr_payloadlen' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_u' type-id='ac5ab598' visibility='default'/>
</data-member>
</class-decl>
<enum-decl name='__anonymous_enum__' is-anonymous='yes' id='08f5ca17'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DRR_BEGIN' value='0'/>
<enumerator name='DRR_OBJECT' value='1'/>
<enumerator name='DRR_FREEOBJECTS' value='2'/>
<enumerator name='DRR_WRITE' value='3'/>
<enumerator name='DRR_FREE' value='4'/>
<enumerator name='DRR_END' value='5'/>
<enumerator name='DRR_WRITE_BYREF' value='6'/>
<enumerator name='DRR_SPILL' value='7'/>
<enumerator name='DRR_WRITE_EMBEDDED' value='8'/>
<enumerator name='DRR_OBJECT_RANGE' value='9'/>
<enumerator name='DRR_REDACT' value='10'/>
<enumerator name='DRR_NUMTYPES' value='11'/>
</enum-decl>
<union-decl name='__anonymous_union__' size-in-bits='2432' is-anonymous='yes' visibility='default' id='ac5ab598'>
<data-member access='public'>
<var-decl name='drr_begin' type-id='09fcdc01' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_end' type-id='6ee25631' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_object' type-id='f9ad530b' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_freeobjects' type-id='a27d958e' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_write' type-id='4cc69e4b' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_free' type-id='c836cfd2' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_write_byref' type-id='e511cdce' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_spill' type-id='1e69a80a' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_write_embedded' type-id='98b1345e' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_object_range' type-id='aba1f9e1' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_redact' type-id='50389039' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_checksum' type-id='a5fe3647' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='drr_end' size-in-bits='320' is-struct='yes' visibility='default' id='6ee25631'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_checksum' type-id='39730d0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_object' size-in-bits='448' is-struct='yes' visibility='default' id='f9ad530b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_type' type-id='5c9d8906' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='drr_bonustype' type-id='5c9d8906' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_blksz' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='drr_bonuslen' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_checksumtype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='200'>
<var-decl name='drr_compress' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='208'>
<var-decl name='drr_dn_slots' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='216'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='drr_raw_bonuslen' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_indblkshift' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='328'>
<var-decl name='drr_nlevels' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='336'>
<var-decl name='drr_nblkptr' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='344'>
<var-decl name='drr_pad' type-id='0f4ddd0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_maxblkid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_freeobjects' size-in-bits='192' is-struct='yes' visibility='default' id='a27d958e'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_firstobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_numobjs' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_write' size-in-bits='1088' is-struct='yes' visibility='default' id='4cc69e4b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_type' type-id='5c9d8906' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='drr_pad' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_logical_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_checksumtype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='328'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='336'>
<var-decl name='drr_compressiontype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='344'>
<var-decl name='drr_pad2' type-id='0f4ddd0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_key' type-id='67f6d2cf' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='drr_compressed_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='768'>
<var-decl name='drr_salt' type-id='13339fda' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='drr_iv' type-id='fa8ef949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='928'>
<var-decl name='drr_mac' type-id='fa9986a5' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_free' size-in-bits='256' is-struct='yes' visibility='default' id='c836cfd2'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_write_byref' size-in-bits='832' is-struct='yes' visibility='default' id='e511cdce'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_refguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_refobject' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_refoffset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='drr_checksumtype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='456'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='464'>
<var-decl name='drr_pad2' type-id='0f562bd0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='drr_key' type-id='67f6d2cf' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_spill' size-in-bits='640' is-struct='yes' visibility='default' id='1e69a80a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='200'>
<var-decl name='drr_compressiontype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='208'>
<var-decl name='drr_pad' type-id='0f562bd0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_compressed_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_salt' type-id='13339fda' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_iv' type-id='fa8ef949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='drr_mac' type-id='fa9986a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='608'>
<var-decl name='drr_type' type-id='5c9d8906' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_write_embedded' size-in-bits='384' is-struct='yes' visibility='default' id='98b1345e'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_compression' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='264'>
<var-decl name='drr_etype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='272'>
<var-decl name='drr_pad' type-id='0f562bd0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_lsize' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='drr_psize' type-id='8f92235e' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_object_range' size-in-bits='512' is-struct='yes' visibility='default' id='aba1f9e1'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_firstobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_numslots' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_salt' type-id='13339fda' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_iv' type-id='fa8ef949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='drr_mac' type-id='fa9986a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='488'>
<var-decl name='drr_pad' type-id='d3490169' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_redact' size-in-bits='256' is-struct='yes' visibility='default' id='50389039'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_checksum' size-in-bits='2432' is-struct='yes' visibility='default' id='a5fe3647'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_pad' type-id='8c2bcad1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2176'>
<var-decl name='drr_checksum' type-id='39730d0b' visibility='default'/>
</data-member>
</class-decl>
+ <class-decl name='__cancel_jmp_buf_tag' size-in-bits='576' is-struct='yes' visibility='default' id='8901473c'>
+ <data-member access='public' layout-offset-in-bits='0'>
+ <var-decl name='__cancel_jmp_buf' type-id='379a1ab7' visibility='default'/>
+ </data-member>
+ <data-member access='public' layout-offset-in-bits='512'>
+ <var-decl name='__mask_was_saved' type-id='95e97e5e' visibility='default'/>
+ </data-member>
+ </class-decl>
+ <class-decl name='__pthread_unwind_buf_t' size-in-bits='832' is-struct='yes' naming-typedef-id='4423cf7f' visibility='default' id='a0abc656'>
+ <data-member access='public' layout-offset-in-bits='0'>
+ <var-decl name='__cancel_jmp_buf' type-id='f5da478b' visibility='default'/>
+ </data-member>
+ <data-member access='public' layout-offset-in-bits='576'>
+ <var-decl name='__pad' type-id='209ef23f' visibility='default'/>
+ </data-member>
+ </class-decl>
+ <typedef-decl name='__pthread_unwind_buf_t' type-id='a0abc656' id='4423cf7f'/>
+ <typedef-decl name='__jmp_buf' type-id='5d4efd44' id='379a1ab7'/>
<typedef-decl name='__clockid_t' type-id='95e97e5e' id='08f9a87a'/>
+ <typedef-decl name='__timer_t' type-id='eaa32e2f' id='df209b60'/>
<typedef-decl name='clockid_t' type-id='08f9a87a' id='a1c3b834'/>
+ <class-decl name='sigevent' size-in-bits='512' is-struct='yes' visibility='default' id='519bc206'>
+ <data-member access='public' layout-offset-in-bits='0'>
+ <var-decl name='sigev_value' type-id='eabacd01' visibility='default'/>
+ </data-member>
+ <data-member access='public' layout-offset-in-bits='64'>
+ <var-decl name='sigev_signo' type-id='95e97e5e' visibility='default'/>
+ </data-member>
+ <data-member access='public' layout-offset-in-bits='96'>
+ <var-decl name='sigev_notify' type-id='95e97e5e' visibility='default'/>
+ </data-member>
+ <data-member access='public' layout-offset-in-bits='128'>
+ <var-decl name='_sigev_un' type-id='ac5ab599' visibility='default'/>
+ </data-member>
+ </class-decl>
+ <union-decl name='__anonymous_union__1' size-in-bits='384' is-anonymous='yes' visibility='default' id='ac5ab599'>
+ <data-member access='public'>
+ <var-decl name='_pad' type-id='73b82f0f' visibility='default'/>
+ </data-member>
+ <data-member access='public'>
+ <var-decl name='_tid' type-id='3629bad8' visibility='default'/>
+ </data-member>
+ <data-member access='public'>
+ <var-decl name='_sigev_thread' type-id='e7f43f7b' visibility='default'/>
+ </data-member>
+ </union-decl>
+ <class-decl name='__anonymous_struct__' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f7b'>
+ <data-member access='public' layout-offset-in-bits='0'>
+ <var-decl name='_function' type-id='5f147c28' visibility='default'/>
+ </data-member>
+ <data-member access='public' layout-offset-in-bits='64'>
+ <var-decl name='_attribute' type-id='7347a39e' visibility='default'/>
+ </data-member>
+ </class-decl>
+ <class-decl name='itimerspec' size-in-bits='256' is-struct='yes' visibility='default' id='acbdbcc6'>
+ <data-member access='public' layout-offset-in-bits='0'>
+ <var-decl name='it_interval' type-id='a9c79a1f' visibility='default'/>
+ </data-member>
+ <data-member access='public' layout-offset-in-bits='128'>
+ <var-decl name='it_value' type-id='a9c79a1f' visibility='default'/>
+ </data-member>
+ </class-decl>
+ <typedef-decl name='timer_t' type-id='df209b60' id='b07ae406'/>
<typedef-decl name='Byte' type-id='002ac4a6' id='efb9ba06'/>
<typedef-decl name='uLong' type-id='7359adad' id='5bbcce85'/>
<typedef-decl name='Bytef' type-id='efb9ba06' id='c1606520'/>
<typedef-decl name='uLongf' type-id='5bbcce85' id='4d39af59'/>
<pointer-type-def type-id='c1606520' size-in-bits='64' id='4c667223'/>
+ <pointer-type-def type-id='8901473c' size-in-bits='64' id='eb91b7ea'/>
+ <pointer-type-def type-id='4423cf7f' size-in-bits='64' id='ba7c727c'/>
+ <pointer-type-def type-id='b9c97942' size-in-bits='64' id='bbf06c47'/>
+ <qualified-type-def type-id='bbf06c47' restrict='yes' id='65e6ec45'/>
<qualified-type-def type-id='c1606520' const='yes' id='a6124a50'/>
<pointer-type-def type-id='a6124a50' size-in-bits='64' id='e8cb3e0e'/>
+ <qualified-type-def type-id='b9c97942' const='yes' id='191f6b72'/>
+ <pointer-type-def type-id='191f6b72' size-in-bits='64' id='e475fb88'/>
+ <qualified-type-def type-id='e475fb88' restrict='yes' id='5a8729d0'/>
<qualified-type-def type-id='781a52d7' const='yes' id='413ab2b8'/>
<pointer-type-def type-id='413ab2b8' size-in-bits='64' id='41671bd6'/>
+ <qualified-type-def type-id='acbdbcc6' const='yes' id='4ba62af7'/>
+ <pointer-type-def type-id='4ba62af7' size-in-bits='64' id='f39579e7'/>
+ <qualified-type-def type-id='f39579e7' restrict='yes' id='9b23e165'/>
<pointer-type-def type-id='c70fa2e8' size-in-bits='64' id='2e711a2a'/>
<pointer-type-def type-id='3ff5601b' size-in-bits='64' id='4aafb922'/>
+ <pointer-type-def type-id='acbdbcc6' size-in-bits='64' id='116842ac'/>
+ <qualified-type-def type-id='116842ac' restrict='yes' id='3d3c4cf4'/>
<pointer-type-def type-id='9e59d1d4' size-in-bits='64' id='4ea84b4f'/>
<pointer-type-def type-id='945467e6' size-in-bits='64' id='8def7735'/>
+ <pointer-type-def type-id='519bc206' size-in-bits='64' id='ef2f159c'/>
+ <qualified-type-def type-id='ef2f159c' restrict='yes' id='de0eb5a4'/>
<pointer-type-def type-id='3d3ffb69' size-in-bits='64' id='72a26210'/>
<pointer-type-def type-id='c9d12d66' size-in-bits='64' id='b2eb2c3f'/>
+ <pointer-type-def type-id='b07ae406' size-in-bits='64' id='36e89359'/>
+ <qualified-type-def type-id='36e89359' restrict='yes' id='de98c2bb'/>
<pointer-type-def type-id='a9c79a1f' size-in-bits='64' id='3d83ba87'/>
<pointer-type-def type-id='4d39af59' size-in-bits='64' id='60db3356'/>
+ <pointer-type-def type-id='f1abb096' size-in-bits='64' id='5f147c28'/>
<pointer-type-def type-id='39730d0b' size-in-bits='64' id='c24fc2ee'/>
<function-decl name='nvlist_print' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_get_pool_handle' mangled-name='zfs_get_pool_handle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_pool_handle'>
<parameter type-id='fcd57163'/>
<return type-id='4c81de99'/>
</function-decl>
<function-decl name='lzc_send_wrapper' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='2e711a2a'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_redacted' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='bfbd3c8e'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_resume_redacted' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='bfbd3c8e'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_receive_with_cmdprops' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='ae3e8ca6'/>
<parameter type-id='3502e3ff'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='41671bd6'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_receive_with_heal' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='ae3e8ca6'/>
<parameter type-id='3502e3ff'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='41671bd6'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_space' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='bfbd3c8e'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_space_resume_redacted' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='bfbd3c8e'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_rename' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_setproctitle' mangled-name='zfs_setproctitle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_setproctitle'>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_insert' mangled-name='avl_insert' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_insert'>
<parameter type-id='a3681dea'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='fba6cb51'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='nvlist_add_boolean' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_boolean' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvpair_value_int32' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<parameter type-id='4aafb922'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fnvlist_size' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='fnvlist_merge' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_remove' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_lookup_boolean_value' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='fletcher_4_native_varsize' mangled-name='fletcher_4_native_varsize' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_native_varsize'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='c24fc2ee'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_4_incremental_native' mangled-name='fletcher_4_incremental_native' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_incremental_native'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fletcher_4_incremental_byteswap' mangled-name='fletcher_4_incremental_byteswap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_incremental_byteswap'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
+ <function-decl name='pthread_exit' visibility='default' binding='global' size-in-bits='64'>
+ <parameter type-id='eaa32e2f'/>
+ <return type-id='48b5725f'/>
+ </function-decl>
+ <function-decl name='__pthread_register_cancel' visibility='default' binding='global' size-in-bits='64'>
+ <parameter type-id='ba7c727c'/>
+ <return type-id='48b5725f'/>
+ </function-decl>
+ <function-decl name='__pthread_unwind_next' visibility='default' binding='global' size-in-bits='64'>
+ <parameter type-id='ba7c727c'/>
+ <return type-id='48b5725f'/>
+ </function-decl>
+ <function-decl name='sigaddset' visibility='default' binding='global' size-in-bits='64'>
+ <parameter type-id='9e80f729'/>
+ <parameter type-id='95e97e5e'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
<function-decl name='perror' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strndup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='time' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b2eb2c3f'/>
<return type-id='c9d12d66'/>
</function-decl>
<function-decl name='clock_gettime' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a1c3b834'/>
<parameter type-id='3d83ba87'/>
<return type-id='95e97e5e'/>
</function-decl>
+ <function-decl name='timer_create' visibility='default' binding='global' size-in-bits='64'>
+ <parameter type-id='a1c3b834'/>
+ <parameter type-id='de0eb5a4'/>
+ <parameter type-id='de98c2bb'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
+ <function-decl name='timer_delete' visibility='default' binding='global' size-in-bits='64'>
+ <parameter type-id='b07ae406'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
+ <function-decl name='timer_settime' visibility='default' binding='global' size-in-bits='64'>
+ <parameter type-id='b07ae406'/>
+ <parameter type-id='95e97e5e'/>
+ <parameter type-id='9b23e165'/>
+ <parameter type-id='3d3c4cf4'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
<function-decl name='write' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<return type-id='79a0948f'/>
</function-decl>
- <function-decl name='sleep' visibility='default' binding='global' size-in-bits='64'>
- <parameter type-id='f0981eeb'/>
- <return type-id='f0981eeb'/>
+ <function-decl name='pause' visibility='default' binding='global' size-in-bits='64'>
+ <return type-id='95e97e5e'/>
+ </function-decl>
+ <function-decl name='pthread_sigmask' visibility='default' binding='global' size-in-bits='64'>
+ <parameter type-id='95e97e5e'/>
+ <parameter type-id='5a8729d0'/>
+ <parameter type-id='65e6ec45'/>
+ <return type-id='95e97e5e'/>
</function-decl>
<function-decl name='uncompress' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='4c667223'/>
<parameter type-id='60db3356'/>
<parameter type-id='e8cb3e0e'/>
<parameter type-id='5bbcce85'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='create_parents' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='26a90f95'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_send_progress' mangled-name='zfs_send_progress' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send_progress'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='5d6479ae' name='bytes_written'/>
<parameter type-id='5d6479ae' name='blocks_visited'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_send_resume_token_to_nvlist' mangled-name='zfs_send_resume_token_to_nvlist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send_resume_token_to_nvlist'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='token'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_send_resume' mangled-name='zfs_send_resume' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send_resume'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='8def7735' name='flags'/>
<parameter type-id='95e97e5e' name='outfd'/>
<parameter type-id='80f4b756' name='resume_token'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_send_saved' mangled-name='zfs_send_saved' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send_saved'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='8def7735' name='flags'/>
<parameter type-id='95e97e5e' name='outfd'/>
<parameter type-id='80f4b756' name='resume_token'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_send' mangled-name='zfs_send' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='fromsnap'/>
<parameter type-id='80f4b756' name='tosnap'/>
<parameter type-id='8def7735' name='flags'/>
<parameter type-id='95e97e5e' name='outfd'/>
<parameter type-id='72a26210' name='filter_func'/>
<parameter type-id='eaa32e2f' name='cb_arg'/>
<parameter type-id='857bb57e' name='debugnvp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_send_one' mangled-name='zfs_send_one' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send_one'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='from'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='8def7735' name='flags'/>
<parameter type-id='80f4b756' name='redactbook'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_receive' mangled-name='zfs_receive' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_receive'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='tosnap'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='4ea84b4f' name='flags'/>
<parameter type-id='95e97e5e' name='infd'/>
<parameter type-id='a3681dea' name='stream_avl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-type size-in-bits='64' id='c70fa2e8'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='d2a5e211'>
<parameter type-id='9200a744'/>
<parameter type-id='eaa32e2f'/>
<return type-id='c19b74c3'/>
</function-type>
+ <function-type size-in-bits='64' id='f1abb096'>
+ <parameter type-id='eabacd01'/>
+ <return type-id='48b5725f'/>
+ </function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_status.c' language='LANG_C99'>
<function-decl name='zpool_import_status' mangled-name='zpool_import_status' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_import_status'>
<parameter type-id='5ce45b60' name='config'/>
<parameter type-id='7d3cd834' name='msgid'/>
<parameter type-id='cec6f2e4' name='errata'/>
<return type-id='d3dd6294'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_util.c' language='LANG_C99'>
<class-decl name='__va_list_tag' size-in-bits='192' is-struct='yes' visibility='default' id='d5027220'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='gp_offset' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='fp_offset' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='overflow_arg_area' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='reg_save_area' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<type-decl name='double' size-in-bits='64' id='a0eb0f08'/>
<array-type-def dimensions='1' type-id='95e97e5e' size-in-bits='192' id='e41bdf22'>
<subrange length='6' type-id='7359adad' id='52fa524b'/>
</array-type-def>
<array-type-def dimensions='1' type-id='19cefcee' size-in-bits='160' alignment-in-bits='32' id='3fcf57d2'>
<subrange length='5' type-id='7359adad' id='53010e10'/>
</array-type-def>
<enum-decl name='zfs_get_column_t' naming-typedef-id='19cefcee' id='223bdcaa'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='GET_COL_NONE' value='0'/>
<enumerator name='GET_COL_NAME' value='1'/>
<enumerator name='GET_COL_PROPERTY' value='2'/>
<enumerator name='GET_COL_VALUE' value='3'/>
<enumerator name='GET_COL_RECVD' value='4'/>
<enumerator name='GET_COL_SOURCE' value='5'/>
</enum-decl>
<typedef-decl name='zfs_get_column_t' type-id='223bdcaa' id='19cefcee'/>
<class-decl name='vdev_cbdata' size-in-bits='192' is-struct='yes' visibility='default' id='b8006be8'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='cb_name_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='cb_names' type-id='9b23c9ad' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='cb_names_count' type-id='f0981eeb' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='vdev_cbdata_t' type-id='b8006be8' id='a9679c94'/>
<class-decl name='zprop_get_cbdata' size-in-bits='832' is-struct='yes' visibility='default' id='f3d3c319'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='cb_sources' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='cb_columns' type-id='3fcf57d2' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='cb_colwidths' type-id='e41bdf22' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='cb_scripted' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='416'>
<var-decl name='cb_literal' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='cb_first' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='cb_proplist' type-id='3a9b2288' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='cb_type' type-id='2e45de5d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='640'>
<var-decl name='cb_vdevs' type-id='a9679c94' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zprop_get_cbdata_t' type-id='f3d3c319' id='f3d87113'/>
<typedef-decl name='zprop_func' type-id='2e711a2a' id='1ec3747a'/>
<enum-decl name='zprop_attr_t' naming-typedef-id='999701cc' id='77d05200'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='PROP_DEFAULT' value='0'/>
<enumerator name='PROP_READONLY' value='1'/>
<enumerator name='PROP_INHERIT' value='2'/>
<enumerator name='PROP_ONETIME' value='3'/>
<enumerator name='PROP_ONETIME_DEFAULT' value='4'/>
</enum-decl>
<typedef-decl name='zprop_attr_t' type-id='77d05200' id='999701cc'/>
<class-decl name='zfs_index' size-in-bits='128' is-struct='yes' visibility='default' id='87957af9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='pi_name' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='pi_value' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zprop_index_t' type-id='87957af9' id='64636ce3'/>
<class-decl name='zprop_desc_t' size-in-bits='640' is-struct='yes' naming-typedef-id='ffa52b96' visibility='default' id='bbff5e4b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='pd_name' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='pd_propnum' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='pd_proptype' type-id='31429eff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='pd_strdefault' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='pd_numdefault' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='pd_attr' type-id='999701cc' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='pd_types' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='pd_values' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='pd_colname' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='pd_rightalign' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='449'>
<var-decl name='pd_visible' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='450'>
<var-decl name='pd_zfs_mod_supported' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='451'>
<var-decl name='pd_always_flex' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='pd_table' type-id='c8bc397b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='pd_table_size' type-id='b59d7dce' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zprop_desc_t' type-id='bbff5e4b' id='ffa52b96'/>
<class-decl name='extmnttab' size-in-bits='320' is-struct='yes' visibility='default' id='0c544dc0'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='mnt_special' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='mnt_mountp' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='mnt_fstype' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='mnt_mntopts' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='mnt_major' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='mnt_minor' type-id='3502e3ff' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='d5027220' size-in-bits='64' id='b7f2d5e6'/>
<qualified-type-def type-id='26a90f95' const='yes' id='57de658a'/>
<pointer-type-def type-id='57de658a' size-in-bits='64' id='f319fae0'/>
<pointer-type-def type-id='9b23c9ad' size-in-bits='64' id='c0563f85'/>
<qualified-type-def type-id='33f57a65' const='yes' id='21fd6035'/>
<pointer-type-def type-id='21fd6035' size-in-bits='64' id='a0de50cd'/>
<pointer-type-def type-id='a0de50cd' size-in-bits='64' id='24f95ba5'/>
<qualified-type-def type-id='64636ce3' const='yes' id='072f7953'/>
<pointer-type-def type-id='072f7953' size-in-bits='64' id='c8bc397b'/>
<pointer-type-def type-id='0c544dc0' size-in-bits='64' id='394fc496'/>
<pointer-type-def type-id='aca3bac8' size-in-bits='64' id='d33f11cb'/>
<qualified-type-def type-id='d33f11cb' restrict='yes' id='5c53ba29'/>
<pointer-type-def type-id='ffa52b96' size-in-bits='64' id='76c8174b'/>
<pointer-type-def type-id='f3d87113' size-in-bits='64' id='0d2a0670'/>
<function-decl name='zfs_version_kernel' mangled-name='zfs_version_kernel' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_version_kernel'>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='libzfs_core_init' visibility='default' binding='global' size-in-bits='64'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_core_fini' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_prop_unsupported' mangled-name='zpool_prop_unsupported' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_unsupported'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_feature_init' mangled-name='zpool_feature_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_feature_init'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_4_init' mangled-name='fletcher_4_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_init'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_4_fini' mangled-name='fletcher_4_fini' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_fini'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_prop_init' mangled-name='zfs_prop_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_init'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_prop_get_table' mangled-name='zfs_prop_get_table' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_table'>
<return type-id='76c8174b'/>
</function-decl>
<function-decl name='zpool_prop_init' mangled-name='zpool_prop_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_init'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_prop_get_table' mangled-name='zpool_prop_get_table' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_get_table'>
<return type-id='76c8174b'/>
</function-decl>
<function-decl name='vdev_prop_init' mangled-name='vdev_prop_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_init'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_iter_common' mangled-name='zprop_iter_common' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_iter_common'>
<parameter type-id='1ec3747a'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='2e45de5d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_name_to_prop' mangled-name='zprop_name_to_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_name_to_prop'>
<parameter type-id='80f4b756'/>
<parameter type-id='2e45de5d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_string_to_index' mangled-name='zprop_string_to_index' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_string_to_index'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='2e45de5d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_values' mangled-name='zprop_values' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_values'>
<parameter type-id='95e97e5e'/>
<parameter type-id='2e45de5d'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zprop_width' mangled-name='zprop_width' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_width'>
<parameter type-id='95e97e5e'/>
<parameter type-id='37e3bd22'/>
<parameter type-id='2e45de5d'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='zprop_valid_for_type' mangled-name='zprop_valid_for_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_valid_for_type'>
<parameter type-id='95e97e5e'/>
<parameter type-id='2e45de5d'/>
<parameter type-id='c19b74c3'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='getextmntent' mangled-name='getextmntent' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getextmntent'>
<parameter type-id='80f4b756'/>
<parameter type-id='394fc496'/>
<parameter type-id='62f7a03d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__ctype_toupper_loc' visibility='default' binding='global' size-in-bits='64'>
<return type-id='24f95ba5'/>
</function-decl>
<function-decl name='dlclose' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='regcomp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5c53ba29'/>
<parameter type-id='9d26089a'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='regfree' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='d33f11cb'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='puts' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='strtod' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<return type-id='a0eb0f08'/>
</function-decl>
<function-decl name='realloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='exit' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strnlen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='strncasecmp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='dup2' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='execve' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='f319fae0'/>
<parameter type-id='f319fae0'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='execv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='f319fae0'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='execvp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='f319fae0'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='execvpe' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='f319fae0'/>
<parameter type-id='f319fae0'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='_exit' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fork' visibility='default' binding='global' size-in-bits='64'>
<return type-id='3629bad8'/>
</function-decl>
<function-decl name='pow' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a0eb0f08'/>
<parameter type-id='a0eb0f08'/>
<return type-id='a0eb0f08'/>
</function-decl>
<function-decl name='__vfprintf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e75a27e9'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='9d26089a'/>
<parameter type-id='b7f2d5e6'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__vasprintf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='9d26089a'/>
<parameter type-id='b7f2d5e6'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='waitpid' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='3629bad8'/>
<parameter type-id='7292109c'/>
<parameter type-id='95e97e5e'/>
<return type-id='3629bad8'/>
</function-decl>
<function-decl name='namespace_clear' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_load_module' visibility='default' binding='global' size-in-bits='64'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_errno' mangled-name='libzfs_errno' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_errno'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_error_action' mangled-name='libzfs_error_action' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_error_action'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='libzfs_error_description' mangled-name='libzfs_error_description' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_error_description'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='libzfs_print_on_error' mangled-name='libzfs_print_on_error' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_print_on_error'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='c19b74c3' name='printerr'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_run_process' mangled-name='libzfs_run_process' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_run_process'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='9b23c9ad' name='argv'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_run_process_get_stdout' mangled-name='libzfs_run_process_get_stdout' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_run_process_get_stdout'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='9b23c9ad' name='argv'/>
<parameter type-id='9b23c9ad' name='env'/>
<parameter type-id='c0563f85' name='lines'/>
<parameter type-id='7292109c' name='lines_cnt'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_run_process_get_stdout_nopath' mangled-name='libzfs_run_process_get_stdout_nopath' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_run_process_get_stdout_nopath'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='9b23c9ad' name='argv'/>
<parameter type-id='9b23c9ad' name='env'/>
<parameter type-id='c0563f85' name='lines'/>
<parameter type-id='7292109c' name='lines_cnt'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_free_str_array' mangled-name='libzfs_free_str_array' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_free_str_array'>
<parameter type-id='9b23c9ad' name='strs'/>
<parameter type-id='95e97e5e' name='count'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_init' mangled-name='libzfs_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_init'>
<return type-id='b0382bb3'/>
</function-decl>
<function-decl name='libzfs_fini' mangled-name='libzfs_fini' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_fini'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_path_to_zhandle' mangled-name='zfs_path_to_zhandle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_path_to_zhandle'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='2e45de5d' name='argtype'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='zprop_print_one_property' mangled-name='zprop_print_one_property' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_print_one_property'>
<parameter type-id='80f4b756' name='name'/>
<parameter type-id='0d2a0670' name='cbp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='80f4b756' name='value'/>
<parameter type-id='a2256d42' name='sourcetype'/>
<parameter type-id='80f4b756' name='source'/>
<parameter type-id='80f4b756' name='recvd_value'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_get_list' mangled-name='zprop_get_list' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_get_list'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='props'/>
<parameter type-id='e4378506' name='listp'/>
<parameter type-id='2e45de5d' name='type'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_free_list' mangled-name='zprop_free_list' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_free_list'>
<parameter type-id='3a9b2288' name='pl'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_iter' mangled-name='zprop_iter' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_iter'>
<parameter type-id='1ec3747a' name='func'/>
<parameter type-id='eaa32e2f' name='cb'/>
<parameter type-id='c19b74c3' name='show_all'/>
<parameter type-id='c19b74c3' name='ordered'/>
<parameter type-id='2e45de5d' name='type'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_version_userland' mangled-name='zfs_version_userland' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_version_userland'>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_version_print' mangled-name='zfs_version_print' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_version_print'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='use_color' mangled-name='use_color' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='use_color'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='printf_color' mangled-name='printf_color' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='printf_color'>
<parameter type-id='80f4b756' name='color'/>
<parameter type-id='80f4b756' name='format'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/os/linux/libzfs_mount_os.c' language='LANG_C99'>
<pointer-type-def type-id='7359adad' size-in-bits='64' id='1d2c2b85'/>
<function-decl name='geteuid' visibility='default' binding='global' size-in-bits='64'>
<return type-id='cc5fcceb'/>
</function-decl>
<function-decl name='mount' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='7359adad'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='umount2' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_parse_mount_options' mangled-name='zfs_parse_mount_options' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_parse_mount_options'>
<parameter type-id='80f4b756' name='mntopts'/>
<parameter type-id='1d2c2b85' name='mntflags'/>
<parameter type-id='1d2c2b85' name='zfsflags'/>
<parameter type-id='95e97e5e' name='sloppy'/>
<parameter type-id='26a90f95' name='badopt'/>
<parameter type-id='26a90f95' name='mtabopt'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_adjust_mount_options' mangled-name='zfs_adjust_mount_options' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_adjust_mount_options'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='mntpoint'/>
<parameter type-id='26a90f95' name='mntopts'/>
<parameter type-id='26a90f95' name='mtabopt'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_mount_delegation_check' mangled-name='zfs_mount_delegation_check' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_mount_delegation_check'>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/os/linux/libzfs_pool_os.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='288' id='16e6f2c6'>
<subrange length='36' type-id='7359adad' id='ae666bde'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a65ae39c' size-in-bits='960' id='fa198beb'>
<subrange length='1' type-id='7359adad' id='52f813b4'/>
</array-type-def>
<array-type-def dimensions='1' type-id='3502e3ff' size-in-bits='384' id='dba89ba3'>
<subrange length='12' type-id='7359adad' id='84827bdc'/>
</array-type-def>
<array-type-def dimensions='1' type-id='3502e3ff' size-in-bits='256' id='01d84ed4'>
<subrange length='8' type-id='7359adad' id='56e0c0b1'/>
</array-type-def>
<class-decl name='dk_part' size-in-bits='960' is-struct='yes' visibility='default' id='a65ae39c'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='p_start' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='p_size' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='p_guid' type-id='214f32ea' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='p_tag' type-id='d908a348' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='272'>
<var-decl name='p_flag' type-id='d908a348' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='p_name' type-id='16e6f2c6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='p_uguid' type-id='214f32ea' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='p_resv' type-id='01d84ed4' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='dk_gpt' size-in-bits='1920' is-struct='yes' visibility='default' id='dd4a2e5a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='efi_version' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='efi_nparts' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='efi_part_size' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='efi_lbasize' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='efi_last_lba' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='efi_first_u_lba' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='efi_last_u_lba' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='efi_disk_uguid' type-id='214f32ea' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='efi_flags' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='efi_reserved1' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='efi_altern_lba' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='efi_reserved' type-id='dba89ba3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='efi_parts' type-id='fa198beb' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='uuid' size-in-bits='128' is-struct='yes' visibility='default' id='214f32ea'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='time_low' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='time_mid' type-id='149c6638' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='48'>
<var-decl name='time_hi_and_version' type-id='149c6638' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='clock_seq_hi_and_reserved' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='72'>
<var-decl name='clock_seq_low' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='80'>
<var-decl name='node_addr' type-id='0f562bd0' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='ushort_t' type-id='8efea9e5' id='d908a348'/>
<typedef-decl name='uint16_t' type-id='253c2d2a' id='149c6638'/>
<typedef-decl name='__uint16_t' type-id='8efea9e5' id='253c2d2a'/>
<pointer-type-def type-id='dd4a2e5a' size-in-bits='64' id='0d8119a8'/>
<pointer-type-def type-id='0d8119a8' size-in-bits='64' id='c43b27a6'/>
<function-decl name='zpool_label_disk_wait' mangled-name='zpool_label_disk_wait' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_label_disk_wait'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_append_partition' mangled-name='zfs_append_partition' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_append_partition'>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='efi_alloc_and_init' mangled-name='efi_alloc_and_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_alloc_and_init'>
<parameter type-id='95e97e5e'/>
<parameter type-id='8f92235e'/>
<parameter type-id='c43b27a6'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='efi_alloc_and_read' mangled-name='efi_alloc_and_read' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_alloc_and_read'>
<parameter type-id='95e97e5e'/>
<parameter type-id='c43b27a6'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='efi_write' mangled-name='efi_write' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_write'>
<parameter type-id='95e97e5e'/>
<parameter type-id='0d8119a8'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='efi_rescan' mangled-name='efi_rescan' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_rescan'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='efi_free' mangled-name='efi_free' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_free'>
<parameter type-id='0d8119a8'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='efi_use_whole_disk' mangled-name='efi_use_whole_disk' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_use_whole_disk'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='rand' visibility='default' binding='global' size-in-bits='64'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fsync' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_label_disk' mangled-name='zpool_label_disk' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_label_disk'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='name'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/os/linux/libzfs_util_os.c' language='LANG_C99'>
- <class-decl name='itimerspec' size-in-bits='256' is-struct='yes' visibility='default' id='acbdbcc6'>
- <data-member access='public' layout-offset-in-bits='0'>
- <var-decl name='it_interval' type-id='a9c79a1f' visibility='default'/>
- </data-member>
- <data-member access='public' layout-offset-in-bits='128'>
- <var-decl name='it_value' type-id='a9c79a1f' visibility='default'/>
- </data-member>
- </class-decl>
<typedef-decl name='nfds_t' type-id='7359adad' id='555eef66'/>
<class-decl name='pollfd' size-in-bits='64' is-struct='yes' visibility='default' id='b440e872'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='fd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='events' type-id='a2185560' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='48'>
<var-decl name='revents' type-id='a2185560' visibility='default'/>
</data-member>
</class-decl>
- <qualified-type-def type-id='acbdbcc6' const='yes' id='4ba62af7'/>
- <pointer-type-def type-id='4ba62af7' size-in-bits='64' id='f39579e7'/>
- <pointer-type-def type-id='acbdbcc6' size-in-bits='64' id='116842ac'/>
<pointer-type-def type-id='b440e872' size-in-bits='64' id='3ac36db0'/>
<function-decl name='access' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__poll_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='3ac36db0'/>
<parameter type-id='555eef66'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='7359adad'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='inotify_init1' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='inotify_add_watch' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='8f92235e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='timerfd_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='08f9a87a'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='timerfd_settime' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='f39579e7'/>
<parameter type-id='116842ac'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_error_init' mangled-name='libzfs_error_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_error_init'>
<parameter type-id='95e97e5e' name='error'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_userns' mangled-name='zfs_userns' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_userns'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='nspath'/>
<parameter type-id='95e97e5e' name='attach'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/os/linux/zutil_device_path_os.c' language='LANG_C99'>
<class-decl name='udev' is-struct='yes' visibility='default' is-declaration-only='yes' id='e4a7fb7f'/>
<class-decl name='udev_device' is-struct='yes' visibility='default' is-declaration-only='yes' id='640b33ca'/>
<pointer-type-def type-id='e4a7fb7f' size-in-bits='64' id='025eefe7'/>
<pointer-type-def type-id='640b33ca' size-in-bits='64' id='b32bae08'/>
<class-decl name='udev' is-struct='yes' visibility='default' is-declaration-only='yes' id='e4a7fb7f'/>
<class-decl name='udev_device' is-struct='yes' visibility='default' is-declaration-only='yes' id='640b33ca'/>
<function-decl name='udev_new' visibility='default' binding='global' size-in-bits='64'>
<return type-id='025eefe7'/>
</function-decl>
<function-decl name='udev_device_unref' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b32bae08'/>
<return type-id='b32bae08'/>
</function-decl>
<function-decl name='udev_device_new_from_subsystem_sysname' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='025eefe7'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='b32bae08'/>
</function-decl>
<function-decl name='udev_device_get_property_value' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b32bae08'/>
<parameter type-id='80f4b756'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='__readlink_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='266fe297'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='79a0948f'/>
</function-decl>
<function-decl name='zfs_get_enclosure_sysfs_path' mangled-name='zfs_get_enclosure_sysfs_path' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_enclosure_sysfs_path'>
<parameter type-id='80f4b756' name='dev_name'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zfs_dev_is_dm' mangled-name='zfs_dev_is_dm' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dev_is_dm'>
<parameter type-id='80f4b756' name='dev_name'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_dev_is_whole_disk' mangled-name='zfs_dev_is_whole_disk' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dev_is_whole_disk'>
<parameter type-id='80f4b756' name='dev_name'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_get_underlying_path' mangled-name='zfs_get_underlying_path' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_underlying_path'>
<parameter type-id='80f4b756' name='dev_name'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='is_mpath_whole_disk' mangled-name='is_mpath_whole_disk' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='is_mpath_whole_disk'>
<parameter type-id='80f4b756' name='path'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/os/linux/zutil_import_os.c' language='LANG_C99'>
<class-decl name='blkid_struct_cache' is-struct='yes' visibility='default' is-declaration-only='yes' id='09286066'/>
<class-decl name='blkid_struct_dev' is-struct='yes' visibility='default' is-declaration-only='yes' id='86223623'/>
<class-decl name='blkid_struct_dev_iterate' is-struct='yes' visibility='default' is-declaration-only='yes' id='d88420d6'/>
<class-decl name='udev_list_entry' is-struct='yes' visibility='default' is-declaration-only='yes' id='e7dbdca3'/>
<typedef-decl name='pool_vdev_iter_f' type-id='6c16a6c8' id='dff793e0'/>
<typedef-decl name='blkid_dev' type-id='8433f053' id='f47b023a'/>
<typedef-decl name='blkid_cache' type-id='940e3afc' id='0882dfdf'/>
<typedef-decl name='blkid_dev_iterate' type-id='b8fa2efc' id='f4760fa7'/>
<typedef-decl name='__useconds_t' type-id='f0981eeb' id='4e80d4b1'/>
<pointer-type-def type-id='0882dfdf' size-in-bits='64' id='2e3e7caa'/>
<pointer-type-def type-id='f47b023a' size-in-bits='64' id='d87f9b75'/>
<pointer-type-def type-id='09286066' size-in-bits='64' id='940e3afc'/>
<pointer-type-def type-id='86223623' size-in-bits='64' id='8433f053'/>
<pointer-type-def type-id='d88420d6' size-in-bits='64' id='b8fa2efc'/>
<pointer-type-def type-id='2ec2411e' size-in-bits='64' id='6c16a6c8'/>
<pointer-type-def type-id='e7dbdca3' size-in-bits='64' id='deabd0d3'/>
<class-decl name='blkid_struct_cache' is-struct='yes' visibility='default' is-declaration-only='yes' id='09286066'/>
<class-decl name='blkid_struct_dev' is-struct='yes' visibility='default' is-declaration-only='yes' id='86223623'/>
<class-decl name='blkid_struct_dev_iterate' is-struct='yes' visibility='default' is-declaration-only='yes' id='d88420d6'/>
<class-decl name='udev_list_entry' is-struct='yes' visibility='default' is-declaration-only='yes' id='e7dbdca3'/>
<function-decl name='for_each_vdev_in_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='dff793e0'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='label_paths' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5507783b'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='7d3cd834'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zutil_alloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5507783b'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='zutil_strdup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5507783b'/>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='slice_cache_compare' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='blkid_put_cache' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0882dfdf'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='blkid_get_cache' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='2e3e7caa'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='blkid_dev_devname' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f47b023a'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='blkid_dev_iterate_begin' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0882dfdf'/>
<return type-id='f4760fa7'/>
</function-decl>
<function-decl name='blkid_dev_set_search' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f4760fa7'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='blkid_dev_next' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f4760fa7'/>
<parameter type-id='d87f9b75'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='blkid_dev_iterate_end' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f4760fa7'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='blkid_probe_all_new' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0882dfdf'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='udev_unref' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='025eefe7'/>
<return type-id='025eefe7'/>
</function-decl>
<function-decl name='udev_list_entry_get_next' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='deabd0d3'/>
<return type-id='deabd0d3'/>
</function-decl>
<function-decl name='udev_list_entry_get_name' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='deabd0d3'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='udev_device_get_parent_with_subsystem_devtype' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b32bae08'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='b32bae08'/>
</function-decl>
<function-decl name='udev_device_get_devlinks_list_entry' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b32bae08'/>
<return type-id='deabd0d3'/>
</function-decl>
<function-decl name='sched_yield' visibility='default' binding='global' size-in-bits='64'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='usleep' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='4e80d4b1'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_dev_flush' mangled-name='zfs_dev_flush' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dev_flush'>
<parameter type-id='95e97e5e' name='fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_device_get_devid' mangled-name='zfs_device_get_devid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_device_get_devid'>
<parameter type-id='b32bae08' name='dev'/>
<parameter type-id='26a90f95' name='bufptr'/>
<parameter type-id='b59d7dce' name='buflen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_device_get_physical' mangled-name='zfs_device_get_physical' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_device_get_physical'>
<parameter type-id='b32bae08' name='dev'/>
<parameter type-id='26a90f95' name='bufptr'/>
<parameter type-id='b59d7dce' name='buflen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='update_vdev_config_dev_sysfs_path' mangled-name='update_vdev_config_dev_sysfs_path' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='update_vdev_config_dev_sysfs_path'>
<parameter type-id='5ce45b60' name='nv'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='80f4b756' name='key'/>
<return type-id='48b5725f'/>
</function-decl>
<function-type size-in-bits='64' id='2ec2411e'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/os/linux/zutil_setproctitle.c' language='LANG_C99'>
<function-decl name='warnx' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='setenv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_setproctitle_init' mangled-name='zfs_setproctitle_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_setproctitle_init'>
<parameter type-id='95e97e5e' name='argc'/>
<parameter type-id='9b23c9ad' name='argv'/>
<parameter type-id='9b23c9ad' name='envp'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/zutil_device_path.c' language='LANG_C99'>
<pointer-type-def type-id='b99c00c9' size-in-bits='64' id='13956559'/>
<function-decl name='zpool_default_search_paths' mangled-name='zpool_default_search_paths' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_default_search_paths'>
<parameter type-id='78c01427'/>
<return type-id='13956559'/>
</function-decl>
<function-decl name='strspn' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='zfs_dirnamelen' mangled-name='zfs_dirnamelen' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dirnamelen'>
<parameter type-id='80f4b756' name='path'/>
<return type-id='79a0948f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/zutil_import.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='256' id='16dc656a'>
<subrange length='32' type-id='7359adad' id='ae5bde82'/>
</array-type-def>
- <array-type-def dimensions='1' type-id='95e97e5e' size-in-bits='384' id='73b82f0f'>
- <subrange length='12' type-id='7359adad' id='84827bdc'/>
- </array-type-def>
<class-decl name='importargs' size-in-bits='448' is-struct='yes' visibility='default' id='7ac83801'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='path' type-id='9b23c9ad' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='paths' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='poolname' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='cachefile' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='can_be_active' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='scan' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='policy' type-id='5ce45b60' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='importargs_t' type-id='7ac83801' id='7a842a6b'/>
<class-decl name='libpc_handle' size-in-bits='8448' is-struct='yes' visibility='default' id='7c8737f0'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='lpc_error' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='lpc_printerr' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='lpc_open_access_error' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='lpc_desc_active' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='lpc_desc' type-id='b54ce520' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8320'>
<var-decl name='lpc_ops' type-id='f095e320' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8384'>
<var-decl name='lpc_lib_handle' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='libpc_handle_t' type-id='7c8737f0' id='8a70a786'/>
<class-decl name='aiocb' size-in-bits='1344' is-struct='yes' visibility='default' id='e4957c49'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='aio_fildes' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='aio_lio_opcode' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='aio_reqprio' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='aio_buf' type-id='fe09dd29' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='aio_nbytes' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='aio_sigevent' type-id='519bc206' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='768'>
<var-decl name='__next_prio' type-id='924bbc81' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='__abs_prio' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='864'>
<var-decl name='__policy' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='896'>
<var-decl name='__error_code' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='__return_value' type-id='41060289' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1024'>
<var-decl name='aio_offset' type-id='724e4de6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1088'>
<var-decl name='__glibc_reserved' type-id='16dc656a' visibility='default'/>
</data-member>
</class-decl>
- <class-decl name='sigevent' size-in-bits='512' is-struct='yes' visibility='default' id='519bc206'>
- <data-member access='public' layout-offset-in-bits='0'>
- <var-decl name='sigev_value' type-id='eabacd01' visibility='default'/>
- </data-member>
- <data-member access='public' layout-offset-in-bits='64'>
- <var-decl name='sigev_signo' type-id='95e97e5e' visibility='default'/>
- </data-member>
- <data-member access='public' layout-offset-in-bits='96'>
- <var-decl name='sigev_notify' type-id='95e97e5e' visibility='default'/>
- </data-member>
- <data-member access='public' layout-offset-in-bits='128'>
- <var-decl name='_sigev_un' type-id='ac5ab599' visibility='default'/>
- </data-member>
- </class-decl>
- <union-decl name='__anonymous_union__' size-in-bits='384' is-anonymous='yes' visibility='default' id='ac5ab599'>
- <data-member access='public'>
- <var-decl name='_pad' type-id='73b82f0f' visibility='default'/>
- </data-member>
- <data-member access='public'>
- <var-decl name='_tid' type-id='3629bad8' visibility='default'/>
- </data-member>
- <data-member access='public'>
- <var-decl name='_sigev_thread' type-id='e7f43f7b' visibility='default'/>
- </data-member>
- </union-decl>
- <class-decl name='__anonymous_struct__' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f7b'>
- <data-member access='public' layout-offset-in-bits='0'>
- <var-decl name='_function' type-id='5f147c28' visibility='default'/>
- </data-member>
- <data-member access='public' layout-offset-in-bits='64'>
- <var-decl name='_attribute' type-id='7347a39e' visibility='default'/>
- </data-member>
- </class-decl>
<pointer-type-def type-id='e4957c49' size-in-bits='64' id='924bbc81'/>
<qualified-type-def type-id='924bbc81' const='yes' id='5499dcde'/>
<pointer-type-def type-id='5499dcde' size-in-bits='64' id='2236d41c'/>
<qualified-type-def type-id='2236d41c' restrict='yes' id='31488924'/>
<pointer-type-def type-id='a3681dea' size-in-bits='64' id='fce6d540'/>
<qualified-type-def type-id='e4957c49' const='yes' id='fced9da2'/>
<pointer-type-def type-id='fced9da2' size-in-bits='64' id='b20efd18'/>
<pointer-type-def type-id='7a842a6b' size-in-bits='64' id='07ee4a58'/>
<pointer-type-def type-id='8a70a786' size-in-bits='64' id='5507783b'/>
<pointer-type-def type-id='b1e62775' size-in-bits='64' id='f095e320'/>
- <pointer-type-def type-id='519bc206' size-in-bits='64' id='ef2f159c'/>
- <qualified-type-def type-id='ef2f159c' restrict='yes' id='de0eb5a4'/>
- <pointer-type-def type-id='f1abb096' size-in-bits='64' id='5f147c28'/>
<qualified-type-def type-id='48b5725f' volatile='yes' id='b0b3cbf9'/>
<pointer-type-def type-id='b0b3cbf9' size-in-bits='64' id='fe09dd29'/>
<function-decl name='update_vdev_config_dev_strs' mangled-name='update_vdev_config_dev_strs' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='update_vdev_config_dev_strs'>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='update_vdevs_config_dev_sysfs_path' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_dup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='spl_pagesize' mangled-name='spl_pagesize' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='spl_pagesize'>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='posix_memalign' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='63e171df'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sysconf' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='bd54fe1a'/>
</function-decl>
<function-decl name='libpc_error_description' mangled-name='libpc_error_description' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libpc_error_description'>
<parameter type-id='5507783b' name='hdl'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_search_import' mangled-name='zpool_search_import' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_search_import'>
<parameter type-id='5507783b' name='hdl'/>
<parameter type-id='07ee4a58' name='import'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zpool_find_config' mangled-name='zpool_find_config' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_find_config'>
<parameter type-id='5507783b' name='hdl'/>
<parameter type-id='80f4b756' name='target'/>
<parameter type-id='857bb57e' name='configp'/>
<parameter type-id='07ee4a58' name='args'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_find_import_blkid' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5507783b'/>
<parameter type-id='18c91f9e'/>
<parameter type-id='fce6d540'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_open_func' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
- <function-type size-in-bits='64' id='f1abb096'>
- <parameter type-id='eabacd01'/>
- <return type-id='48b5725f'/>
- </function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/zutil_nicenum.c' language='LANG_C99'>
<type-decl name='long double' size-in-bits='128' id='e095c704'/>
<enum-decl name='zfs_nicenum_format' id='29cf1969'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_NICENUM_1024' value='0'/>
<enumerator name='ZFS_NICENUM_BYTES' value='1'/>
<enumerator name='ZFS_NICENUM_TIME' value='2'/>
<enumerator name='ZFS_NICENUM_RAW' value='3'/>
<enumerator name='ZFS_NICENUM_RAWTIME' value='4'/>
</enum-decl>
<function-decl name='powl' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e095c704'/>
<parameter type-id='e095c704'/>
<return type-id='e095c704'/>
</function-decl>
<function-decl name='floor' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a0eb0f08'/>
<return type-id='a0eb0f08'/>
</function-decl>
<function-decl name='zfs_isnumber' mangled-name='zfs_isnumber' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_isnumber'>
<parameter type-id='80f4b756' name='str'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_nicenum_format' mangled-name='zfs_nicenum_format' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_nicenum_format'>
<parameter type-id='9c313c2d' name='num'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='buflen'/>
<parameter type-id='29cf1969' name='format'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_nicetime' mangled-name='zfs_nicetime' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_nicetime'>
<parameter type-id='9c313c2d' name='num'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='buflen'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_niceraw' mangled-name='zfs_niceraw' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_niceraw'>
<parameter type-id='9c313c2d' name='num'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='buflen'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/zutil_pool.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='853fd5dc' size-in-bits='32768' id='b505fc2f'>
<subrange length='64' type-id='7359adad' id='b10be967'/>
</array-type-def>
<class-decl name='ddt_stat' size-in-bits='512' is-struct='yes' visibility='default' id='65242dfe'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='dds_blocks' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='dds_lsize' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='dds_psize' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='dds_dsize' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='dds_ref_blocks' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='dds_ref_lsize' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='dds_ref_psize' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='dds_ref_dsize' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='ddt_stat_t' type-id='65242dfe' id='853fd5dc'/>
<class-decl name='ddt_histogram' size-in-bits='32768' is-struct='yes' visibility='default' id='bc2b3086'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='ddh_stat' type-id='b505fc2f' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='ddt_histogram_t' type-id='bc2b3086' id='2d7fe832'/>
<qualified-type-def type-id='2d7fe832' const='yes' id='ec92d602'/>
<pointer-type-def type-id='ec92d602' size-in-bits='64' id='932720f8'/>
<qualified-type-def type-id='853fd5dc' const='yes' id='764c298c'/>
<pointer-type-def type-id='764c298c' size-in-bits='64' id='dfe59052'/>
<function-decl name='zpool_dump_ddt' mangled-name='zpool_dump_ddt' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_dump_ddt'>
<parameter type-id='dfe59052' name='dds_total'/>
<parameter type-id='932720f8' name='ddh'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/avl/avl.c' language='LANG_C99'>
<function-decl name='avl_last' mangled-name='avl_last' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_last'>
<parameter type-id='a3681dea' name='tree'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='avl_nearest' mangled-name='avl_nearest' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_nearest'>
<parameter type-id='a3681dea' name='tree'/>
<parameter type-id='fba6cb51' name='where'/>
<parameter type-id='95e97e5e' name='direction'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='avl_insert_here' mangled-name='avl_insert_here' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_insert_here'>
<parameter type-id='a3681dea' name='tree'/>
<parameter type-id='eaa32e2f' name='new_data'/>
<parameter type-id='eaa32e2f' name='here'/>
<parameter type-id='95e97e5e' name='direction'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_update_lt' mangled-name='avl_update_lt' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_update_lt'>
<parameter type-id='a3681dea' name='t'/>
<parameter type-id='eaa32e2f' name='obj'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='avl_update_gt' mangled-name='avl_update_gt' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_update_gt'>
<parameter type-id='a3681dea' name='t'/>
<parameter type-id='eaa32e2f' name='obj'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='avl_update' mangled-name='avl_update' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_update'>
<parameter type-id='a3681dea' name='t'/>
<parameter type-id='eaa32e2f' name='obj'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='avl_swap' mangled-name='avl_swap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_swap'>
<parameter type-id='a3681dea' name='tree1'/>
<parameter type-id='a3681dea' name='tree2'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_is_empty' mangled-name='avl_is_empty' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_is_empty'>
<parameter type-id='a3681dea' name='tree'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/cityhash.c' language='LANG_C99'>
<function-decl name='cityhash4' mangled-name='cityhash4' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='cityhash4'>
<parameter type-id='9c313c2d' name='w1'/>
<parameter type-id='9c313c2d' name='w2'/>
<parameter type-id='9c313c2d' name='w3'/>
<parameter type-id='9c313c2d' name='w4'/>
<return type-id='9c313c2d'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfeature_common.c' language='LANG_C99'>
- <array-type-def dimensions='1' type-id='83f29ca2' size-in-bits='17920' id='dd432c71'>
- <subrange length='40' type-id='7359adad' id='ae4a9561'/>
+ <array-type-def dimensions='1' type-id='83f29ca2' size-in-bits='18368' id='b93e4d14'>
+ <subrange length='41' type-id='7359adad' id='cb834f44'/>
</array-type-def>
<enum-decl name='zfeature_flags' id='6db816a4'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFEATURE_FLAG_READONLY_COMPAT' value='1'/>
<enumerator name='ZFEATURE_FLAG_MOS' value='2'/>
<enumerator name='ZFEATURE_FLAG_ACTIVATE_ON_ENABLE' value='4'/>
<enumerator name='ZFEATURE_FLAG_PER_DATASET' value='8'/>
</enum-decl>
<typedef-decl name='zfeature_flags_t' type-id='6db816a4' id='fc329033'/>
<enum-decl name='zfeature_type' id='c4fa2355'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFEATURE_TYPE_BOOLEAN' value='0'/>
<enumerator name='ZFEATURE_TYPE_UINT64_ARRAY' value='1'/>
<enumerator name='ZFEATURE_NUM_TYPES' value='2'/>
</enum-decl>
<typedef-decl name='zfeature_type_t' type-id='c4fa2355' id='732d2bb2'/>
<class-decl name='zfeature_info' size-in-bits='448' is-struct='yes' visibility='default' id='1178d146'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='fi_feature' type-id='d6618c78' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='fi_uname' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='fi_guid' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='fi_desc' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='fi_flags' type-id='fc329033' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='fi_zfs_mod_supported' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='fi_type' type-id='732d2bb2' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='fi_depends' type-id='1acff326' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfeature_info_t' type-id='1178d146' id='83f29ca2'/>
<typedef-decl name='__free_fn_t' type-id='b7f9d8e6' id='3ff5e51e'/>
<class-decl name='dirent' size-in-bits='2240' is-struct='yes' visibility='default' id='611586a1'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='d_ino' type-id='71288a47' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='d_off' type-id='724e4de6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='d_reclen' type-id='8efea9e5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='144'>
<var-decl name='d_type' type-id='002ac4a6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='152'>
<var-decl name='d_name' type-id='d1617432' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='zfs_mod_supported_features' size-in-bits='128' is-struct='yes' visibility='default' id='3eee3342'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tree' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='all_features' type-id='c19b74c3' visibility='default'/>
</data-member>
</class-decl>
<qualified-type-def type-id='d6618c78' const='yes' id='81a65028'/>
<pointer-type-def type-id='81a65028' size-in-bits='64' id='1acff326'/>
<qualified-type-def type-id='3eee3342' const='yes' id='0c1d5bbb'/>
<pointer-type-def type-id='0c1d5bbb' size-in-bits='64' id='a3372543'/>
<pointer-type-def type-id='611586a1' size-in-bits='64' id='2e243169'/>
<qualified-type-def type-id='eaa32e2f' const='yes' id='83be723c'/>
<pointer-type-def type-id='83be723c' size-in-bits='64' id='7acd98a2'/>
- <var-decl name='spa_feature_table' type-id='dd432c71' mangled-name='spa_feature_table' visibility='default' elf-symbol-id='spa_feature_table'/>
+ <var-decl name='spa_feature_table' type-id='b93e4d14' mangled-name='spa_feature_table' visibility='default' elf-symbol-id='spa_feature_table'/>
<var-decl name='zfeature_checks_disable' type-id='c19b74c3' mangled-name='zfeature_checks_disable' visibility='default' elf-symbol-id='zfeature_checks_disable'/>
<function-decl name='opendir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='f09217ba'/>
</function-decl>
<function-decl name='tsearch' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='63e171df'/>
<parameter type-id='aba7edd8'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='tfind' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='7acd98a2'/>
<parameter type-id='aba7edd8'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='tdestroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='3ff5e51e'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfeature_is_valid_guid' mangled-name='zfeature_is_valid_guid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfeature_is_valid_guid'>
<parameter type-id='80f4b756' name='name'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfeature_depends_on' mangled-name='zfeature_depends_on' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfeature_depends_on'>
<parameter type-id='d6618c78' name='fid'/>
<parameter type-id='d6618c78' name='check'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_mod_supported' mangled-name='zfs_mod_supported' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_mod_supported'>
<parameter type-id='80f4b756' name='scope'/>
<parameter type-id='80f4b756' name='name'/>
<parameter type-id='a3372543' name='sfeatures'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_comutil.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='b99c00c9' size-in-bits='2624' id='5ce15418'>
<subrange length='41' type-id='7359adad' id='cb834f44'/>
</array-type-def>
<qualified-type-def type-id='80f4b756' const='yes' id='b99c00c9'/>
<pointer-type-def type-id='8f92235e' size-in-bits='64' id='90421557'/>
<function-decl name='nvpair_value_uint32' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<parameter type-id='90421557'/>
<return type-id='95e97e5e'/>
</function-decl>
<var-decl name='zfs_history_event_names' type-id='5ce15418' mangled-name='zfs_history_event_names' visibility='default' elf-symbol-id='zfs_history_event_names'/>
<function-decl name='strpbrk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zfs_allocatable_devs' mangled-name='zfs_allocatable_devs' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_allocatable_devs'>
<parameter type-id='5ce45b60' name='nv'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_special_devs' mangled-name='zfs_special_devs' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_special_devs'>
<parameter type-id='5ce45b60' name='nv'/>
<parameter type-id='80f4b756' name='type'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_zpl_version_map' mangled-name='zfs_zpl_version_map' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_zpl_version_map'>
<parameter type-id='95e97e5e' name='spa_version'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_spa_version_map' mangled-name='zfs_spa_version_map' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_spa_version_map'>
<parameter type-id='95e97e5e' name='zpl_version'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_dataset_name_hidden' mangled-name='zfs_dataset_name_hidden' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dataset_name_hidden'>
<parameter type-id='80f4b756' name='name'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_deleg.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='fa1870fd' size-in-bits='4096' id='59e94aca'>
<subrange length='32' type-id='7359adad' id='ae5bde82'/>
</array-type-def>
<array-type-def dimensions='1' type-id='fa1870fd' size-in-bits='infinite' id='7c00e69d'>
<subrange length='infinite' id='031f2035'/>
</array-type-def>
<enum-decl name='zfs_deleg_who_type_t' naming-typedef-id='36d4bd5a' id='b5fa5816'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_DELEG_WHO_UNKNOWN' value='0'/>
<enumerator name='ZFS_DELEG_USER' value='117'/>
<enumerator name='ZFS_DELEG_USER_SETS' value='85'/>
<enumerator name='ZFS_DELEG_GROUP' value='103'/>
<enumerator name='ZFS_DELEG_GROUP_SETS' value='71'/>
<enumerator name='ZFS_DELEG_EVERYONE' value='101'/>
<enumerator name='ZFS_DELEG_EVERYONE_SETS' value='69'/>
<enumerator name='ZFS_DELEG_CREATE' value='99'/>
<enumerator name='ZFS_DELEG_CREATE_SETS' value='67'/>
<enumerator name='ZFS_DELEG_NAMED_SET' value='115'/>
<enumerator name='ZFS_DELEG_NAMED_SET_SETS' value='83'/>
</enum-decl>
<typedef-decl name='zfs_deleg_who_type_t' type-id='b5fa5816' id='36d4bd5a'/>
<enum-decl name='zfs_deleg_note_t' naming-typedef-id='4613c173' id='729d4547'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_DELEG_NOTE_CREATE' value='0'/>
<enumerator name='ZFS_DELEG_NOTE_DESTROY' value='1'/>
<enumerator name='ZFS_DELEG_NOTE_SNAPSHOT' value='2'/>
<enumerator name='ZFS_DELEG_NOTE_ROLLBACK' value='3'/>
<enumerator name='ZFS_DELEG_NOTE_CLONE' value='4'/>
<enumerator name='ZFS_DELEG_NOTE_PROMOTE' value='5'/>
<enumerator name='ZFS_DELEG_NOTE_RENAME' value='6'/>
<enumerator name='ZFS_DELEG_NOTE_SEND' value='7'/>
<enumerator name='ZFS_DELEG_NOTE_RECEIVE' value='8'/>
<enumerator name='ZFS_DELEG_NOTE_ALLOW' value='9'/>
<enumerator name='ZFS_DELEG_NOTE_USERPROP' value='10'/>
<enumerator name='ZFS_DELEG_NOTE_MOUNT' value='11'/>
<enumerator name='ZFS_DELEG_NOTE_SHARE' value='12'/>
<enumerator name='ZFS_DELEG_NOTE_USERQUOTA' value='13'/>
<enumerator name='ZFS_DELEG_NOTE_GROUPQUOTA' value='14'/>
<enumerator name='ZFS_DELEG_NOTE_USERUSED' value='15'/>
<enumerator name='ZFS_DELEG_NOTE_GROUPUSED' value='16'/>
<enumerator name='ZFS_DELEG_NOTE_USEROBJQUOTA' value='17'/>
<enumerator name='ZFS_DELEG_NOTE_GROUPOBJQUOTA' value='18'/>
<enumerator name='ZFS_DELEG_NOTE_USEROBJUSED' value='19'/>
<enumerator name='ZFS_DELEG_NOTE_GROUPOBJUSED' value='20'/>
<enumerator name='ZFS_DELEG_NOTE_HOLD' value='21'/>
<enumerator name='ZFS_DELEG_NOTE_RELEASE' value='22'/>
<enumerator name='ZFS_DELEG_NOTE_DIFF' value='23'/>
<enumerator name='ZFS_DELEG_NOTE_BOOKMARK' value='24'/>
<enumerator name='ZFS_DELEG_NOTE_LOAD_KEY' value='25'/>
<enumerator name='ZFS_DELEG_NOTE_CHANGE_KEY' value='26'/>
<enumerator name='ZFS_DELEG_NOTE_PROJECTUSED' value='27'/>
<enumerator name='ZFS_DELEG_NOTE_PROJECTQUOTA' value='28'/>
<enumerator name='ZFS_DELEG_NOTE_PROJECTOBJUSED' value='29'/>
<enumerator name='ZFS_DELEG_NOTE_PROJECTOBJQUOTA' value='30'/>
<enumerator name='ZFS_DELEG_NOTE_NONE' value='31'/>
</enum-decl>
<typedef-decl name='zfs_deleg_note_t' type-id='729d4547' id='4613c173'/>
<class-decl name='zfs_deleg_perm_tab' size-in-bits='128' is-struct='yes' visibility='default' id='5aa05c1f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='z_perm' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='z_note' type-id='4613c173' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_deleg_perm_tab_t' type-id='5aa05c1f' id='f3f851ad'/>
<qualified-type-def type-id='f3f851ad' const='yes' id='fa1870fd'/>
<var-decl name='zfs_deleg_perm_tab' type-id='7c00e69d' mangled-name='zfs_deleg_perm_tab' visibility='default' elf-symbol-id='zfs_deleg_perm_tab'/>
<function-decl name='permset_namecheck' mangled-name='permset_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='permset_namecheck'>
<parameter type-id='80f4b756'/>
<parameter type-id='053457bd'/>
<parameter type-id='26a90f95'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_delegatable' mangled-name='zfs_prop_delegatable' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_delegatable'>
<parameter type-id='58603c44'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_deleg_canonicalize_perm' mangled-name='zfs_deleg_canonicalize_perm' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_deleg_canonicalize_perm'>
<parameter type-id='80f4b756' name='perm'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_deleg_verify_nvlist' mangled-name='zfs_deleg_verify_nvlist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_deleg_verify_nvlist'>
<parameter type-id='5ce45b60' name='nvp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_deleg_whokey' mangled-name='zfs_deleg_whokey' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_deleg_whokey'>
<parameter type-id='26a90f95' name='attr'/>
<parameter type-id='36d4bd5a' name='type'/>
<parameter type-id='a84c031d' name='inheritchr'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='512' id='c5d13f42'>
<subrange length='8' type-id='7359adad' id='56e0c0b1'/>
</array-type-def>
<array-type-def dimensions='1' type-id='90dbb6d6' size-in-bits='2048' id='16582e69'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<array-type-def dimensions='1' type-id='8240361c' size-in-bits='1024' id='481f90b1'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<array-type-def dimensions='1' type-id='7c1ab40c' size-in-bits='512' id='cbd91ec1'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<array-type-def dimensions='1' type-id='6d059eaa' size-in-bits='1024' id='729b6ebb'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<enum-decl name='zio_byteorder_t' naming-typedef-id='595a65ec' id='fc861be0'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZIO_CHECKSUM_NATIVE' value='0'/>
<enumerator name='ZIO_CHECKSUM_BYTESWAP' value='1'/>
</enum-decl>
<typedef-decl name='zio_byteorder_t' type-id='fc861be0' id='595a65ec'/>
<class-decl name='zio_abd_checksum_data' size-in-bits='256' is-struct='yes' visibility='default' id='4bf4b004'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='acd_byteorder' type-id='595a65ec' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='acd_ctx' type-id='0f7df99e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='acd_zcp' type-id='c24fc2ee' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='acd_private' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zio_abd_checksum_data_t' type-id='4bf4b004' id='74e39470'/>
<typedef-decl name='zio_abd_checksum_init_t' type-id='a5444274' id='029a8ebe'/>
<typedef-decl name='zio_abd_checksum_fini_t' type-id='a5444274' id='d6fd5c6c'/>
<typedef-decl name='zio_abd_checksum_iter_t' type-id='f4a1892e' id='cefa0f4a'/>
<class-decl name='zio_abd_checksum_func' size-in-bits='192' is-struct='yes' visibility='default' id='aa14691a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='acf_init' type-id='0bcca125' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='acf_fini' type-id='bfe36153' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='acf_iter' type-id='1e276399' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zio_abd_checksum_func_t' type-id='3f8e8d11' id='c2eb138a'/>
<class-decl name='zfs_fletcher_superscalar' size-in-bits='256' is-struct='yes' visibility='default' id='28efb250'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='v' type-id='85c64d26' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_fletcher_superscalar_t' type-id='28efb250' id='6d059eaa'/>
<class-decl name='zfs_fletcher_sse' size-in-bits='128' is-struct='yes' visibility='default' id='acd4019a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='v' type-id='c1c22e6c' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_fletcher_sse_t' type-id='acd4019a' id='7c1ab40c'/>
<class-decl name='zfs_fletcher_avx' size-in-bits='256' is-struct='yes' visibility='default' id='8c208dfa'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='v' type-id='85c64d26' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_fletcher_avx_t' type-id='8c208dfa' id='8240361c'/>
<class-decl name='zfs_fletcher_avx512' size-in-bits='512' is-struct='yes' visibility='default' id='c6d0c382'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='v' type-id='c5d13f42' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_fletcher_avx512_t' type-id='c6d0c382' id='90dbb6d6'/>
<union-decl name='fletcher_4_ctx' size-in-bits='2048' visibility='default' id='1f951ade'>
<data-member access='public'>
<var-decl name='scalar' type-id='39730d0b' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='superscalar' type-id='729b6ebb' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='sse' type-id='cbd91ec1' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='avx' type-id='481f90b1' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='avx512' type-id='16582e69' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='fletcher_4_ctx_t' type-id='1f951ade' id='4b675395'/>
<qualified-type-def type-id='aa14691a' const='yes' id='3f8e8d11'/>
<pointer-type-def type-id='4b675395' size-in-bits='64' id='0f7df99e'/>
<qualified-type-def type-id='8f92235e' volatile='yes' id='430e0681'/>
<pointer-type-def type-id='430e0681' size-in-bits='64' id='3a147f31'/>
<pointer-type-def type-id='74e39470' size-in-bits='64' id='eefe7427'/>
<pointer-type-def type-id='d6fd5c6c' size-in-bits='64' id='bfe36153'/>
<pointer-type-def type-id='029a8ebe' size-in-bits='64' id='0bcca125'/>
<pointer-type-def type-id='cefa0f4a' size-in-bits='64' id='1e276399'/>
<var-decl name='fletcher_4_abd_ops' type-id='c2eb138a' mangled-name='fletcher_4_abd_ops' visibility='default' elf-symbol-id='fletcher_4_abd_ops'/>
<function-decl name='atomic_swap_32' mangled-name='atomic_swap_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_32'>
<parameter type-id='3a147f31'/>
<parameter type-id='8f92235e'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='membar_producer' mangled-name='membar_producer' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='membar_producer'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_init' mangled-name='fletcher_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_init'>
<parameter type-id='c24fc2ee' name='zcp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_2_incremental_native' mangled-name='fletcher_2_incremental_native' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_2_incremental_native'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='b59d7dce' name='size'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fletcher_2_native' mangled-name='fletcher_2_native' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_2_native'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='9c313c2d' name='size'/>
<parameter type-id='eaa32e2f' name='ctx_template'/>
<parameter type-id='c24fc2ee' name='zcp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_2_incremental_byteswap' mangled-name='fletcher_2_incremental_byteswap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_2_incremental_byteswap'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='b59d7dce' name='size'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fletcher_2_byteswap' mangled-name='fletcher_2_byteswap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_2_byteswap'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='9c313c2d' name='size'/>
<parameter type-id='eaa32e2f' name='ctx_template'/>
<parameter type-id='c24fc2ee' name='zcp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_4_impl_set' mangled-name='fletcher_4_impl_set' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_impl_set'>
<parameter type-id='80f4b756' name='val'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fletcher_4_native' mangled-name='fletcher_4_native' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_native'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='9c313c2d' name='size'/>
<parameter type-id='eaa32e2f' name='ctx_template'/>
<parameter type-id='c24fc2ee' name='zcp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_4_byteswap' mangled-name='fletcher_4_byteswap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_byteswap'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='9c313c2d' name='size'/>
<parameter type-id='eaa32e2f' name='ctx_template'/>
<parameter type-id='c24fc2ee' name='zcp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-type size-in-bits='64' id='f4a1892e'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='a5444274'>
<parameter type-id='eefe7427'/>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher_avx512.c' language='LANG_C99'>
<typedef-decl name='fletcher_4_init_f' type-id='173aa527' id='b9ae1656'/>
<typedef-decl name='fletcher_4_fini_f' type-id='0ad5b8a8' id='c4c1f4fc'/>
<typedef-decl name='fletcher_4_compute_f' type-id='38147eff' id='ad1dc4cb'/>
<class-decl name='fletcher_4_func' size-in-bits='1024' is-struct='yes' visibility='default' id='57f479a0'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='init_native' type-id='b9ae1656' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='fini_native' type-id='c4c1f4fc' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='compute_native' type-id='ad1dc4cb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='init_byteswap' type-id='b9ae1656' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='fini_byteswap' type-id='c4c1f4fc' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='compute_byteswap' type-id='ad1dc4cb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='valid' type-id='297d38bc' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='uses_fpu' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='name' type-id='80f4b756' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='fletcher_4_ops_t' type-id='57f479a0' id='eba91718'/>
<qualified-type-def type-id='eba91718' const='yes' id='9eeabdc8'/>
<pointer-type-def type-id='e9e61702' size-in-bits='64' id='297d38bc'/>
<pointer-type-def type-id='fe40251b' size-in-bits='64' id='173aa527'/>
<pointer-type-def type-id='17fb1f83' size-in-bits='64' id='38147eff'/>
<pointer-type-def type-id='fb39e25e' size-in-bits='64' id='0ad5b8a8'/>
<var-decl name='fletcher_4_avx512f_ops' type-id='9eeabdc8' mangled-name='fletcher_4_avx512f_ops' visibility='default' elf-symbol-id='fletcher_4_avx512f_ops'/>
<var-decl name='fletcher_4_avx512bw_ops' type-id='9eeabdc8' mangled-name='fletcher_4_avx512bw_ops' visibility='default' elf-symbol-id='fletcher_4_avx512bw_ops'/>
<function-type size-in-bits='64' id='e9e61702'>
<return type-id='c19b74c3'/>
</function-type>
<function-type size-in-bits='64' id='fe40251b'>
<parameter type-id='0f7df99e'/>
<return type-id='48b5725f'/>
</function-type>
<function-type size-in-bits='64' id='17fb1f83'>
<parameter type-id='0f7df99e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='9c313c2d'/>
<return type-id='48b5725f'/>
</function-type>
<function-type size-in-bits='64' id='fb39e25e'>
<parameter type-id='0f7df99e'/>
<parameter type-id='c24fc2ee'/>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher_intel.c' language='LANG_C99'>
<var-decl name='fletcher_4_avx2_ops' type-id='9eeabdc8' mangled-name='fletcher_4_avx2_ops' visibility='default' elf-symbol-id='fletcher_4_avx2_ops'/>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher_sse.c' language='LANG_C99'>
<var-decl name='fletcher_4_sse2_ops' type-id='9eeabdc8' mangled-name='fletcher_4_sse2_ops' visibility='default' elf-symbol-id='fletcher_4_sse2_ops'/>
<var-decl name='fletcher_4_ssse3_ops' type-id='9eeabdc8' mangled-name='fletcher_4_ssse3_ops' visibility='default' elf-symbol-id='fletcher_4_ssse3_ops'/>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher_superscalar.c' language='LANG_C99'>
<var-decl name='fletcher_4_superscalar_ops' type-id='9eeabdc8' mangled-name='fletcher_4_superscalar_ops' visibility='default' elf-symbol-id='fletcher_4_superscalar_ops'/>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher_superscalar4.c' language='LANG_C99'>
<var-decl name='fletcher_4_superscalar4_ops' type-id='9eeabdc8' mangled-name='fletcher_4_superscalar4_ops' visibility='default' elf-symbol-id='fletcher_4_superscalar4_ops'/>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_namecheck.c' language='LANG_C99'>
<var-decl name='zfs_max_dataset_nesting' type-id='95e97e5e' mangled-name='zfs_max_dataset_nesting' visibility='default' elf-symbol-id='zfs_max_dataset_nesting'/>
<function-decl name='get_dataset_depth' mangled-name='get_dataset_depth' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='get_dataset_depth'>
<parameter type-id='80f4b756' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_component_namecheck' mangled-name='zfs_component_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_component_namecheck'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='053457bd' name='why'/>
<parameter type-id='26a90f95' name='what'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='dataset_namecheck' mangled-name='dataset_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='dataset_namecheck'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='053457bd' name='why'/>
<parameter type-id='26a90f95' name='what'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='bookmark_namecheck' mangled-name='bookmark_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='bookmark_namecheck'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='053457bd' name='why'/>
<parameter type-id='26a90f95' name='what'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='snapshot_namecheck' mangled-name='snapshot_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='snapshot_namecheck'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='053457bd' name='why'/>
<parameter type-id='26a90f95' name='what'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_prop.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='b99c00c9' size-in-bits='768' id='bcc77e38'>
<subrange length='12' type-id='7359adad' id='84827bdc'/>
</array-type-def>
<pointer-type-def type-id='3eee3342' size-in-bits='64' id='73f8e240'/>
<var-decl name='zfs_userquota_prop_prefixes' type-id='bcc77e38' mangled-name='zfs_userquota_prop_prefixes' visibility='default' elf-symbol-id='zfs_userquota_prop_prefixes'/>
<function-decl name='zfs_mod_list_supported' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='73f8e240'/>
</function-decl>
<function-decl name='zfs_mod_list_supported_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='73f8e240'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_register_impl' mangled-name='zprop_register_impl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_register_impl'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='31429eff'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='80f4b756'/>
<parameter type-id='999701cc'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c8bc397b'/>
<parameter type-id='a3372543'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_register_string' mangled-name='zprop_register_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_register_string'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='999701cc'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='a3372543'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_register_number' mangled-name='zprop_register_number' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_register_number'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='999701cc'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='a3372543'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_register_index' mangled-name='zprop_register_index' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_register_index'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='999701cc'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c8bc397b'/>
<parameter type-id='a3372543'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_register_hidden' mangled-name='zprop_register_hidden' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_register_hidden'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='31429eff'/>
<parameter type-id='999701cc'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='a3372543'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_index_to_string' mangled-name='zprop_index_to_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_index_to_string'>
<parameter type-id='95e97e5e'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='7d3cd834'/>
<parameter type-id='2e45de5d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_random_value' mangled-name='zprop_random_value' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_random_value'>
<parameter type-id='95e97e5e'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='2e45de5d'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zprop_valid_char' mangled-name='zprop_valid_char' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_valid_char'>
<parameter type-id='a84c031d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_string_to_index' mangled-name='zfs_prop_string_to_index' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_string_to_index'>
<parameter type-id='58603c44' name='prop'/>
<parameter type-id='80f4b756' name='string'/>
<parameter type-id='5d6479ae' name='index'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_random_value' mangled-name='zfs_prop_random_value' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_random_value'>
<parameter type-id='58603c44' name='prop'/>
<parameter type-id='9c313c2d' name='seed'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zfs_prop_visible' mangled-name='zfs_prop_visible' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_visible'>
<parameter type-id='58603c44' name='prop'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_values' mangled-name='zfs_prop_values' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_values'>
<parameter type-id='58603c44' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_prop_is_string' mangled-name='zfs_prop_is_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_is_string'>
<parameter type-id='58603c44' name='prop'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_column_name' mangled-name='zfs_prop_column_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_column_name'>
<parameter type-id='58603c44' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_prop_align_right' mangled-name='zfs_prop_align_right' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_align_right'>
<parameter type-id='58603c44' name='prop'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zpool_prop.c' language='LANG_C99'>
<function-decl name='zpool_prop_string_to_index' mangled-name='zpool_prop_string_to_index' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_string_to_index'>
<parameter type-id='5d0c23fb' name='prop'/>
<parameter type-id='80f4b756' name='string'/>
<parameter type-id='5d6479ae' name='index'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_prop_random_value' mangled-name='zpool_prop_random_value' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_random_value'>
<parameter type-id='5d0c23fb' name='prop'/>
<parameter type-id='9c313c2d' name='seed'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zpool_prop_values' mangled-name='zpool_prop_values' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_values'>
<parameter type-id='5d0c23fb' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_prop_column_name' mangled-name='zpool_prop_column_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_column_name'>
<parameter type-id='5d0c23fb' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_prop_align_right' mangled-name='zpool_prop_align_right' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_align_right'>
<parameter type-id='5d0c23fb' name='prop'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='vdev_prop_get_table' mangled-name='vdev_prop_get_table' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_get_table'>
<return type-id='76c8174b'/>
</function-decl>
<function-decl name='vdev_prop_string_to_index' mangled-name='vdev_prop_string_to_index' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_string_to_index'>
<parameter type-id='5aa5c90c' name='prop'/>
<parameter type-id='80f4b756' name='string'/>
<parameter type-id='5d6479ae' name='index'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='vdev_prop_random_value' mangled-name='vdev_prop_random_value' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_random_value'>
<parameter type-id='5aa5c90c' name='prop'/>
<parameter type-id='9c313c2d' name='seed'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='vdev_prop_values' mangled-name='vdev_prop_values' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_values'>
<parameter type-id='5aa5c90c' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='vdev_prop_column_name' mangled-name='vdev_prop_column_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_column_name'>
<parameter type-id='5aa5c90c' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='vdev_prop_align_right' mangled-name='vdev_prop_align_right' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_align_right'>
<parameter type-id='5aa5c90c' name='prop'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zprop_common.c' language='LANG_C99'>
<function-decl name='__ctype_tolower_loc' visibility='default' binding='global' size-in-bits='64'>
<return type-id='24f95ba5'/>
</function-decl>
</abi-instr>
</abi-corpus>
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c b/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c
index 4ebd112f452f..71cf029deff5 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c
@@ -1,5398 +1,5441 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright 2015 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>
* Copyright (c) 2018 Datto Inc.
* Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2018, loli10K <ezomori.nozomu@gmail.com>
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
* Copyright (c) 2021, 2023, Klara Inc.
*/
#include <errno.h>
#include <libintl.h>
#include <stdio.h>
#include <stdlib.h>
#include <strings.h>
#include <unistd.h>
#include <libgen.h>
#include <zone.h>
#include <sys/stat.h>
#include <sys/efi_partition.h>
#include <sys/systeminfo.h>
#include <sys/zfs_ioctl.h>
#include <sys/zfs_sysfs.h>
#include <sys/vdev_disk.h>
#include <sys/types.h>
#include <dlfcn.h>
#include <libzutil.h>
#include <fcntl.h>
#include "zfs_namecheck.h"
#include "zfs_prop.h"
#include "libzfs_impl.h"
#include "zfs_comutil.h"
#include "zfeature_common.h"
static boolean_t zpool_vdev_is_interior(const char *name);
typedef struct prop_flags {
unsigned int create:1; /* Validate property on creation */
unsigned int import:1; /* Validate property on import */
unsigned int vdevprop:1; /* Validate property as a VDEV property */
} prop_flags_t;
/*
* ====================================================================
* zpool property functions
* ====================================================================
*/
static int
zpool_get_all_props(zpool_handle_t *zhp)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zcmd_alloc_dst_nvlist(hdl, &zc, 0);
while (zfs_ioctl(hdl, ZFS_IOC_POOL_GET_PROPS, &zc) != 0) {
if (errno == ENOMEM)
zcmd_expand_dst_nvlist(hdl, &zc);
else {
zcmd_free_nvlists(&zc);
return (-1);
}
}
if (zcmd_read_dst_nvlist(hdl, &zc, &zhp->zpool_props) != 0) {
zcmd_free_nvlists(&zc);
return (-1);
}
zcmd_free_nvlists(&zc);
return (0);
}
int
zpool_props_refresh(zpool_handle_t *zhp)
{
nvlist_t *old_props;
old_props = zhp->zpool_props;
if (zpool_get_all_props(zhp) != 0)
return (-1);
nvlist_free(old_props);
return (0);
}
static const char *
zpool_get_prop_string(zpool_handle_t *zhp, zpool_prop_t prop,
zprop_source_t *src)
{
nvlist_t *nv, *nvl;
const char *value;
zprop_source_t source;
nvl = zhp->zpool_props;
if (nvlist_lookup_nvlist(nvl, zpool_prop_to_name(prop), &nv) == 0) {
source = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
value = fnvlist_lookup_string(nv, ZPROP_VALUE);
} else {
source = ZPROP_SRC_DEFAULT;
if ((value = zpool_prop_default_string(prop)) == NULL)
value = "-";
}
if (src)
*src = source;
return (value);
}
uint64_t
zpool_get_prop_int(zpool_handle_t *zhp, zpool_prop_t prop, zprop_source_t *src)
{
nvlist_t *nv, *nvl;
uint64_t value;
zprop_source_t source;
if (zhp->zpool_props == NULL && zpool_get_all_props(zhp)) {
/*
* zpool_get_all_props() has most likely failed because
* the pool is faulted, but if all we need is the top level
* vdev's guid then get it from the zhp config nvlist.
*/
if ((prop == ZPOOL_PROP_GUID) &&
(nvlist_lookup_nvlist(zhp->zpool_config,
ZPOOL_CONFIG_VDEV_TREE, &nv) == 0) &&
(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &value)
== 0)) {
return (value);
}
return (zpool_prop_default_numeric(prop));
}
nvl = zhp->zpool_props;
if (nvlist_lookup_nvlist(nvl, zpool_prop_to_name(prop), &nv) == 0) {
source = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
value = fnvlist_lookup_uint64(nv, ZPROP_VALUE);
} else {
source = ZPROP_SRC_DEFAULT;
value = zpool_prop_default_numeric(prop);
}
if (src)
*src = source;
return (value);
}
/*
* Map VDEV STATE to printed strings.
*/
const char *
zpool_state_to_name(vdev_state_t state, vdev_aux_t aux)
{
switch (state) {
case VDEV_STATE_CLOSED:
case VDEV_STATE_OFFLINE:
return (gettext("OFFLINE"));
case VDEV_STATE_REMOVED:
return (gettext("REMOVED"));
case VDEV_STATE_CANT_OPEN:
if (aux == VDEV_AUX_CORRUPT_DATA || aux == VDEV_AUX_BAD_LOG)
return (gettext("FAULTED"));
else if (aux == VDEV_AUX_SPLIT_POOL)
return (gettext("SPLIT"));
else
return (gettext("UNAVAIL"));
case VDEV_STATE_FAULTED:
return (gettext("FAULTED"));
case VDEV_STATE_DEGRADED:
return (gettext("DEGRADED"));
case VDEV_STATE_HEALTHY:
return (gettext("ONLINE"));
default:
break;
}
return (gettext("UNKNOWN"));
}
/*
* Map POOL STATE to printed strings.
*/
const char *
zpool_pool_state_to_name(pool_state_t state)
{
switch (state) {
default:
break;
case POOL_STATE_ACTIVE:
return (gettext("ACTIVE"));
case POOL_STATE_EXPORTED:
return (gettext("EXPORTED"));
case POOL_STATE_DESTROYED:
return (gettext("DESTROYED"));
case POOL_STATE_SPARE:
return (gettext("SPARE"));
case POOL_STATE_L2CACHE:
return (gettext("L2CACHE"));
case POOL_STATE_UNINITIALIZED:
return (gettext("UNINITIALIZED"));
case POOL_STATE_UNAVAIL:
return (gettext("UNAVAIL"));
case POOL_STATE_POTENTIALLY_ACTIVE:
return (gettext("POTENTIALLY_ACTIVE"));
}
return (gettext("UNKNOWN"));
}
/*
* Given a pool handle, return the pool health string ("ONLINE", "DEGRADED",
* "SUSPENDED", etc).
*/
const char *
zpool_get_state_str(zpool_handle_t *zhp)
{
zpool_errata_t errata;
zpool_status_t status;
const char *str;
status = zpool_get_status(zhp, NULL, &errata);
if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) {
str = gettext("FAULTED");
} else if (status == ZPOOL_STATUS_IO_FAILURE_WAIT ||
status == ZPOOL_STATUS_IO_FAILURE_CONTINUE ||
status == ZPOOL_STATUS_IO_FAILURE_MMP) {
str = gettext("SUSPENDED");
} else {
nvlist_t *nvroot = fnvlist_lookup_nvlist(
zpool_get_config(zhp, NULL), ZPOOL_CONFIG_VDEV_TREE);
uint_t vsc;
vdev_stat_t *vs = (vdev_stat_t *)fnvlist_lookup_uint64_array(
nvroot, ZPOOL_CONFIG_VDEV_STATS, &vsc);
str = zpool_state_to_name(vs->vs_state, vs->vs_aux);
}
return (str);
}
/*
* Get a zpool property value for 'prop' and return the value in
* a pre-allocated buffer.
*/
int
zpool_get_prop(zpool_handle_t *zhp, zpool_prop_t prop, char *buf,
size_t len, zprop_source_t *srctype, boolean_t literal)
{
uint64_t intval;
const char *strval;
zprop_source_t src = ZPROP_SRC_NONE;
if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) {
switch (prop) {
case ZPOOL_PROP_NAME:
(void) strlcpy(buf, zpool_get_name(zhp), len);
break;
case ZPOOL_PROP_HEALTH:
(void) strlcpy(buf, zpool_get_state_str(zhp), len);
break;
case ZPOOL_PROP_GUID:
intval = zpool_get_prop_int(zhp, prop, &src);
(void) snprintf(buf, len, "%llu", (u_longlong_t)intval);
break;
case ZPOOL_PROP_ALTROOT:
case ZPOOL_PROP_CACHEFILE:
case ZPOOL_PROP_COMMENT:
case ZPOOL_PROP_COMPATIBILITY:
if (zhp->zpool_props != NULL ||
zpool_get_all_props(zhp) == 0) {
(void) strlcpy(buf,
zpool_get_prop_string(zhp, prop, &src),
len);
break;
}
zfs_fallthrough;
default:
(void) strlcpy(buf, "-", len);
break;
}
if (srctype != NULL)
*srctype = src;
return (0);
}
if (zhp->zpool_props == NULL && zpool_get_all_props(zhp) &&
prop != ZPOOL_PROP_NAME)
return (-1);
switch (zpool_prop_get_type(prop)) {
case PROP_TYPE_STRING:
(void) strlcpy(buf, zpool_get_prop_string(zhp, prop, &src),
len);
break;
case PROP_TYPE_NUMBER:
intval = zpool_get_prop_int(zhp, prop, &src);
switch (prop) {
case ZPOOL_PROP_SIZE:
case ZPOOL_PROP_ALLOCATED:
case ZPOOL_PROP_FREE:
case ZPOOL_PROP_FREEING:
case ZPOOL_PROP_LEAKED:
case ZPOOL_PROP_ASHIFT:
case ZPOOL_PROP_MAXBLOCKSIZE:
case ZPOOL_PROP_MAXDNODESIZE:
case ZPOOL_PROP_BCLONESAVED:
case ZPOOL_PROP_BCLONEUSED:
if (literal)
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
else
(void) zfs_nicenum(intval, buf, len);
break;
case ZPOOL_PROP_EXPANDSZ:
case ZPOOL_PROP_CHECKPOINT:
if (intval == 0) {
(void) strlcpy(buf, "-", len);
} else if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) zfs_nicebytes(intval, buf, len);
}
break;
case ZPOOL_PROP_CAPACITY:
if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) snprintf(buf, len, "%llu%%",
(u_longlong_t)intval);
}
break;
case ZPOOL_PROP_FRAGMENTATION:
if (intval == UINT64_MAX) {
(void) strlcpy(buf, "-", len);
} else if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) snprintf(buf, len, "%llu%%",
(u_longlong_t)intval);
}
break;
case ZPOOL_PROP_BCLONERATIO:
case ZPOOL_PROP_DEDUPRATIO:
if (literal)
(void) snprintf(buf, len, "%llu.%02llu",
(u_longlong_t)(intval / 100),
(u_longlong_t)(intval % 100));
else
(void) snprintf(buf, len, "%llu.%02llux",
(u_longlong_t)(intval / 100),
(u_longlong_t)(intval % 100));
break;
case ZPOOL_PROP_HEALTH:
(void) strlcpy(buf, zpool_get_state_str(zhp), len);
break;
case ZPOOL_PROP_VERSION:
if (intval >= SPA_VERSION_FEATURES) {
(void) snprintf(buf, len, "-");
break;
}
zfs_fallthrough;
default:
(void) snprintf(buf, len, "%llu", (u_longlong_t)intval);
}
break;
case PROP_TYPE_INDEX:
intval = zpool_get_prop_int(zhp, prop, &src);
if (zpool_prop_index_to_string(prop, intval, &strval)
!= 0)
return (-1);
(void) strlcpy(buf, strval, len);
break;
default:
abort();
}
if (srctype)
*srctype = src;
return (0);
}
/*
* Get a zpool property value for 'propname' and return the value in
* a pre-allocated buffer.
*/
int
zpool_get_userprop(zpool_handle_t *zhp, const char *propname, char *buf,
size_t len, zprop_source_t *srctype)
{
nvlist_t *nv, *nvl;
uint64_t ival;
const char *value;
zprop_source_t source = ZPROP_SRC_LOCAL;
nvl = zhp->zpool_props;
if (nvlist_lookup_nvlist(nvl, propname, &nv) == 0) {
if (nvlist_lookup_uint64(nv, ZPROP_SOURCE, &ival) == 0)
source = ival;
verify(nvlist_lookup_string(nv, ZPROP_VALUE, &value) == 0);
} else {
source = ZPROP_SRC_DEFAULT;
value = "-";
}
if (srctype)
*srctype = source;
(void) strlcpy(buf, value, len);
return (0);
}
/*
* Check if the bootfs name has the same pool name as it is set to.
* Assuming bootfs is a valid dataset name.
*/
static boolean_t
bootfs_name_valid(const char *pool, const char *bootfs)
{
int len = strlen(pool);
if (bootfs[0] == '\0')
return (B_TRUE);
if (!zfs_name_valid(bootfs, ZFS_TYPE_FILESYSTEM|ZFS_TYPE_SNAPSHOT))
return (B_FALSE);
if (strncmp(pool, bootfs, len) == 0 &&
(bootfs[len] == '/' || bootfs[len] == '\0'))
return (B_TRUE);
return (B_FALSE);
}
/*
* Given an nvlist of zpool properties to be set, validate that they are
* correct, and parse any numeric properties (index, boolean, etc) if they are
* specified as strings.
*/
static nvlist_t *
zpool_valid_proplist(libzfs_handle_t *hdl, const char *poolname,
nvlist_t *props, uint64_t version, prop_flags_t flags, char *errbuf)
{
nvpair_t *elem;
nvlist_t *retprops;
zpool_prop_t prop;
const char *strval;
uint64_t intval;
const char *slash, *check;
struct stat64 statbuf;
zpool_handle_t *zhp;
char report[1024];
if (nvlist_alloc(&retprops, NV_UNIQUE_NAME, 0) != 0) {
(void) no_memory(hdl);
return (NULL);
}
elem = NULL;
while ((elem = nvlist_next_nvpair(props, elem)) != NULL) {
const char *propname = nvpair_name(elem);
if (flags.vdevprop && zpool_prop_vdev(propname)) {
vdev_prop_t vprop = vdev_name_to_prop(propname);
if (vdev_prop_readonly(vprop)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' "
"is readonly"), propname);
(void) zfs_error(hdl, EZFS_PROPREADONLY,
errbuf);
goto error;
}
if (zprop_parse_value(hdl, elem, vprop, ZFS_TYPE_VDEV,
retprops, &strval, &intval, errbuf) != 0)
goto error;
continue;
} else if (flags.vdevprop && vdev_prop_user(propname)) {
if (nvlist_add_nvpair(retprops, elem) != 0) {
(void) no_memory(hdl);
goto error;
}
continue;
} else if (flags.vdevprop) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid property: '%s'"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
prop = zpool_name_to_prop(propname);
if (prop == ZPOOL_PROP_INVAL && zpool_prop_feature(propname)) {
int err;
char *fname = strchr(propname, '@') + 1;
err = zfeature_lookup_name(fname, NULL);
if (err != 0) {
ASSERT3U(err, ==, ENOENT);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"feature '%s' unsupported by kernel"),
fname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (nvpair_type(elem) != DATA_TYPE_STRING) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a string"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
(void) nvpair_value_string(elem, &strval);
if (strcmp(strval, ZFS_FEATURE_ENABLED) != 0 &&
strcmp(strval, ZFS_FEATURE_DISABLED) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' can only be set to "
"'enabled' or 'disabled'"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (!flags.create &&
strcmp(strval, ZFS_FEATURE_DISABLED) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' can only be set to "
"'disabled' at creation time"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (nvlist_add_uint64(retprops, propname, 0) != 0) {
(void) no_memory(hdl);
goto error;
}
continue;
} else if (prop == ZPOOL_PROP_INVAL &&
zfs_prop_user(propname)) {
/*
* This is a user property: make sure it's a
* string, and that it's less than ZAP_MAXNAMELEN.
*/
if (nvpair_type(elem) != DATA_TYPE_STRING) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a string"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (strlen(nvpair_name(elem)) >= ZAP_MAXNAMELEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property name '%s' is too long"),
propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
(void) nvpair_value_string(elem, &strval);
if (strlen(strval) >= ZFS_MAXPROPLEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property value '%s' is too long"),
strval);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (nvlist_add_string(retprops, propname,
strval) != 0) {
(void) no_memory(hdl);
goto error;
}
continue;
}
/*
* Make sure this property is valid and applies to this type.
*/
if (prop == ZPOOL_PROP_INVAL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid property '%s'"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (zpool_prop_readonly(prop)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' "
"is readonly"), propname);
(void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf);
goto error;
}
if (!flags.create && zpool_prop_setonce(prop)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' can only be set at "
"creation time"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (zprop_parse_value(hdl, elem, prop, ZFS_TYPE_POOL, retprops,
&strval, &intval, errbuf) != 0)
goto error;
/*
* Perform additional checking for specific properties.
*/
switch (prop) {
case ZPOOL_PROP_VERSION:
if (intval < version ||
!SPA_VERSION_IS_SUPPORTED(intval)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' number %llu is invalid."),
propname, (unsigned long long)intval);
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
goto error;
}
break;
case ZPOOL_PROP_ASHIFT:
if (intval != 0 &&
(intval < ASHIFT_MIN || intval > ASHIFT_MAX)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' number %llu is invalid, "
"only values between %" PRId32 " and %"
PRId32 " are allowed."),
propname, (unsigned long long)intval,
ASHIFT_MIN, ASHIFT_MAX);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZPOOL_PROP_BOOTFS:
if (flags.create || flags.import) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' cannot be set at creation "
"or import time"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (version < SPA_VERSION_BOOTFS) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded to support "
"'%s' property"), propname);
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
goto error;
}
/*
* bootfs property value has to be a dataset name and
* the dataset has to be in the same pool as it sets to.
*/
if (!bootfs_name_valid(poolname, strval)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' "
"is an invalid name"), strval);
(void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf);
goto error;
}
if ((zhp = zpool_open_canfail(hdl, poolname)) == NULL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"could not open pool '%s'"), poolname);
(void) zfs_error(hdl, EZFS_OPENFAILED, errbuf);
goto error;
}
zpool_close(zhp);
break;
case ZPOOL_PROP_ALTROOT:
if (!flags.create && !flags.import) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' can only be set during pool "
"creation or import"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (strval[0] != '/') {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"bad alternate root '%s'"), strval);
(void) zfs_error(hdl, EZFS_BADPATH, errbuf);
goto error;
}
break;
case ZPOOL_PROP_CACHEFILE:
if (strval[0] == '\0')
break;
if (strcmp(strval, "none") == 0)
break;
if (strval[0] != '/') {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' must be empty, an "
"absolute path, or 'none'"), propname);
(void) zfs_error(hdl, EZFS_BADPATH, errbuf);
goto error;
}
slash = strrchr(strval, '/');
if (slash[1] == '\0' || strcmp(slash, "/.") == 0 ||
strcmp(slash, "/..") == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' is not a valid file"), strval);
(void) zfs_error(hdl, EZFS_BADPATH, errbuf);
goto error;
}
*(char *)slash = '\0';
if (strval[0] != '\0' &&
(stat64(strval, &statbuf) != 0 ||
!S_ISDIR(statbuf.st_mode))) {
*(char *)slash = '/';
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' is not a valid directory"),
strval);
(void) zfs_error(hdl, EZFS_BADPATH, errbuf);
goto error;
}
*(char *)slash = '/';
break;
case ZPOOL_PROP_COMPATIBILITY:
switch (zpool_load_compat(strval, NULL, report, 1024)) {
case ZPOOL_COMPATIBILITY_OK:
case ZPOOL_COMPATIBILITY_WARNTOKEN:
break;
case ZPOOL_COMPATIBILITY_BADFILE:
case ZPOOL_COMPATIBILITY_BADTOKEN:
case ZPOOL_COMPATIBILITY_NOFILES:
zfs_error_aux(hdl, "%s", report);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZPOOL_PROP_COMMENT:
for (check = strval; *check != '\0'; check++) {
if (!isprint(*check)) {
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN,
"comment may only have printable "
"characters"));
(void) zfs_error(hdl, EZFS_BADPROP,
errbuf);
goto error;
}
}
if (strlen(strval) > ZPROP_MAX_COMMENT) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"comment must not exceed %d characters"),
ZPROP_MAX_COMMENT);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZPOOL_PROP_READONLY:
if (!flags.import) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' can only be set at "
"import time"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZPOOL_PROP_MULTIHOST:
if (get_system_hostid() == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"requires a non-zero system hostid"));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZPOOL_PROP_DEDUPDITTO:
printf("Note: property '%s' no longer has "
"any effect\n", propname);
break;
default:
break;
}
}
return (retprops);
error:
nvlist_free(retprops);
return (NULL);
}
/*
* Set zpool property : propname=propval.
*/
int
zpool_set_prop(zpool_handle_t *zhp, const char *propname, const char *propval)
{
zfs_cmd_t zc = {"\0"};
int ret = -1;
char errbuf[ERRBUFLEN];
nvlist_t *nvl = NULL;
nvlist_t *realprops;
uint64_t version;
prop_flags_t flags = { 0 };
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot set property for '%s'"),
zhp->zpool_name);
if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
if (nvlist_add_string(nvl, propname, propval) != 0) {
nvlist_free(nvl);
return (no_memory(zhp->zpool_hdl));
}
version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL);
if ((realprops = zpool_valid_proplist(zhp->zpool_hdl,
zhp->zpool_name, nvl, version, flags, errbuf)) == NULL) {
nvlist_free(nvl);
return (-1);
}
nvlist_free(nvl);
nvl = realprops;
/*
* Execute the corresponding ioctl() to set this property.
*/
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zcmd_write_src_nvlist(zhp->zpool_hdl, &zc, nvl);
ret = zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_SET_PROPS, &zc);
zcmd_free_nvlists(&zc);
nvlist_free(nvl);
if (ret)
(void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf);
else
(void) zpool_props_refresh(zhp);
return (ret);
}
int
zpool_expand_proplist(zpool_handle_t *zhp, zprop_list_t **plp,
zfs_type_t type, boolean_t literal)
{
libzfs_handle_t *hdl = zhp->zpool_hdl;
zprop_list_t *entry;
char buf[ZFS_MAXPROPLEN];
nvlist_t *features = NULL;
nvpair_t *nvp;
zprop_list_t **last;
boolean_t firstexpand = (NULL == *plp);
int i;
if (zprop_expand_list(hdl, plp, type) != 0)
return (-1);
if (type == ZFS_TYPE_VDEV)
return (0);
last = plp;
while (*last != NULL)
last = &(*last)->pl_next;
if ((*plp)->pl_all)
features = zpool_get_features(zhp);
if ((*plp)->pl_all && firstexpand) {
/* Handle userprops in the all properties case */
if (zhp->zpool_props == NULL && zpool_props_refresh(zhp))
return (-1);
nvp = NULL;
while ((nvp = nvlist_next_nvpair(zhp->zpool_props, nvp)) !=
NULL) {
const char *propname = nvpair_name(nvp);
if (!zfs_prop_user(propname))
continue;
entry = zfs_alloc(hdl, sizeof (zprop_list_t));
entry->pl_prop = ZPROP_USERPROP;
entry->pl_user_prop = zfs_strdup(hdl, propname);
entry->pl_width = strlen(entry->pl_user_prop);
entry->pl_all = B_TRUE;
*last = entry;
last = &entry->pl_next;
}
for (i = 0; i < SPA_FEATURES; i++) {
entry = zfs_alloc(hdl, sizeof (zprop_list_t));
entry->pl_prop = ZPROP_USERPROP;
entry->pl_user_prop = zfs_asprintf(hdl, "feature@%s",
spa_feature_table[i].fi_uname);
entry->pl_width = strlen(entry->pl_user_prop);
entry->pl_all = B_TRUE;
*last = entry;
last = &entry->pl_next;
}
}
/* add any unsupported features */
for (nvp = nvlist_next_nvpair(features, NULL);
nvp != NULL; nvp = nvlist_next_nvpair(features, nvp)) {
char *propname;
boolean_t found;
if (zfeature_is_supported(nvpair_name(nvp)))
continue;
propname = zfs_asprintf(hdl, "unsupported@%s",
nvpair_name(nvp));
/*
* Before adding the property to the list make sure that no
* other pool already added the same property.
*/
found = B_FALSE;
entry = *plp;
while (entry != NULL) {
if (entry->pl_user_prop != NULL &&
strcmp(propname, entry->pl_user_prop) == 0) {
found = B_TRUE;
break;
}
entry = entry->pl_next;
}
if (found) {
free(propname);
continue;
}
entry = zfs_alloc(hdl, sizeof (zprop_list_t));
entry->pl_prop = ZPROP_USERPROP;
entry->pl_user_prop = propname;
entry->pl_width = strlen(entry->pl_user_prop);
entry->pl_all = B_TRUE;
*last = entry;
last = &entry->pl_next;
}
for (entry = *plp; entry != NULL; entry = entry->pl_next) {
if (entry->pl_fixed && !literal)
continue;
if (entry->pl_prop != ZPROP_USERPROP &&
zpool_get_prop(zhp, entry->pl_prop, buf, sizeof (buf),
NULL, literal) == 0) {
if (strlen(buf) > entry->pl_width)
entry->pl_width = strlen(buf);
} else if (entry->pl_prop == ZPROP_INVAL &&
zfs_prop_user(entry->pl_user_prop) &&
zpool_get_userprop(zhp, entry->pl_user_prop, buf,
sizeof (buf), NULL) == 0) {
if (strlen(buf) > entry->pl_width)
entry->pl_width = strlen(buf);
}
}
return (0);
}
int
vdev_expand_proplist(zpool_handle_t *zhp, const char *vdevname,
zprop_list_t **plp)
{
zprop_list_t *entry;
char buf[ZFS_MAXPROPLEN];
const char *strval = NULL;
int err = 0;
nvpair_t *elem = NULL;
nvlist_t *vprops = NULL;
nvlist_t *propval = NULL;
const char *propname;
vdev_prop_t prop;
zprop_list_t **last;
for (entry = *plp; entry != NULL; entry = entry->pl_next) {
if (entry->pl_fixed)
continue;
if (zpool_get_vdev_prop(zhp, vdevname, entry->pl_prop,
entry->pl_user_prop, buf, sizeof (buf), NULL,
B_FALSE) == 0) {
if (strlen(buf) > entry->pl_width)
entry->pl_width = strlen(buf);
}
if (entry->pl_prop == VDEV_PROP_NAME &&
strlen(vdevname) > entry->pl_width)
entry->pl_width = strlen(vdevname);
}
/* Handle the all properties case */
last = plp;
if (*last != NULL && (*last)->pl_all == B_TRUE) {
while (*last != NULL)
last = &(*last)->pl_next;
err = zpool_get_all_vdev_props(zhp, vdevname, &vprops);
if (err != 0)
return (err);
while ((elem = nvlist_next_nvpair(vprops, elem)) != NULL) {
propname = nvpair_name(elem);
/* Skip properties that are not user defined */
if ((prop = vdev_name_to_prop(propname)) !=
VDEV_PROP_USERPROP)
continue;
if (nvpair_value_nvlist(elem, &propval) != 0)
continue;
strval = fnvlist_lookup_string(propval, ZPROP_VALUE);
entry = zfs_alloc(zhp->zpool_hdl,
sizeof (zprop_list_t));
entry->pl_prop = prop;
entry->pl_user_prop = zfs_strdup(zhp->zpool_hdl,
propname);
entry->pl_width = strlen(strval);
entry->pl_all = B_TRUE;
*last = entry;
last = &entry->pl_next;
}
}
return (0);
}
/*
* Get the state for the given feature on the given ZFS pool.
*/
int
zpool_prop_get_feature(zpool_handle_t *zhp, const char *propname, char *buf,
size_t len)
{
uint64_t refcount;
boolean_t found = B_FALSE;
nvlist_t *features = zpool_get_features(zhp);
boolean_t supported;
const char *feature = strchr(propname, '@') + 1;
supported = zpool_prop_feature(propname);
ASSERT(supported || zpool_prop_unsupported(propname));
/*
* Convert from feature name to feature guid. This conversion is
* unnecessary for unsupported@... properties because they already
* use guids.
*/
if (supported) {
int ret;
spa_feature_t fid;
ret = zfeature_lookup_name(feature, &fid);
if (ret != 0) {
(void) strlcpy(buf, "-", len);
return (ENOTSUP);
}
feature = spa_feature_table[fid].fi_guid;
}
if (nvlist_lookup_uint64(features, feature, &refcount) == 0)
found = B_TRUE;
if (supported) {
if (!found) {
(void) strlcpy(buf, ZFS_FEATURE_DISABLED, len);
} else {
if (refcount == 0)
(void) strlcpy(buf, ZFS_FEATURE_ENABLED, len);
else
(void) strlcpy(buf, ZFS_FEATURE_ACTIVE, len);
}
} else {
if (found) {
if (refcount == 0) {
(void) strcpy(buf, ZFS_UNSUPPORTED_INACTIVE);
} else {
(void) strcpy(buf, ZFS_UNSUPPORTED_READONLY);
}
} else {
(void) strlcpy(buf, "-", len);
return (ENOTSUP);
}
}
return (0);
}
/*
* Validate the given pool name, optionally putting an extended error message in
* 'buf'.
*/
boolean_t
zpool_name_valid(libzfs_handle_t *hdl, boolean_t isopen, const char *pool)
{
namecheck_err_t why;
char what;
int ret;
ret = pool_namecheck(pool, &why, &what);
/*
* The rules for reserved pool names were extended at a later point.
* But we need to support users with existing pools that may now be
* invalid. So we only check for this expanded set of names during a
* create (or import), and only in userland.
*/
if (ret == 0 && !isopen &&
(strncmp(pool, "mirror", 6) == 0 ||
strncmp(pool, "raidz", 5) == 0 ||
strncmp(pool, "draid", 5) == 0 ||
strncmp(pool, "spare", 5) == 0 ||
strcmp(pool, "log") == 0)) {
if (hdl != NULL)
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "name is reserved"));
return (B_FALSE);
}
if (ret != 0) {
if (hdl != NULL) {
switch (why) {
case NAME_ERR_TOOLONG:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "name is too long"));
break;
case NAME_ERR_INVALCHAR:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "invalid character "
"'%c' in pool name"), what);
break;
case NAME_ERR_NOLETTER:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"name must begin with a letter"));
break;
case NAME_ERR_RESERVED:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"name is reserved"));
break;
case NAME_ERR_DISKLIKE:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool name is reserved"));
break;
case NAME_ERR_LEADING_SLASH:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"leading slash in name"));
break;
case NAME_ERR_EMPTY_COMPONENT:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"empty component in name"));
break;
case NAME_ERR_TRAILING_SLASH:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"trailing slash in name"));
break;
case NAME_ERR_MULTIPLE_DELIMITERS:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"multiple '@' and/or '#' delimiters in "
"name"));
break;
case NAME_ERR_NO_AT:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"permission set is missing '@'"));
break;
default:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"(%d) not defined"), why);
break;
}
}
return (B_FALSE);
}
return (B_TRUE);
}
/*
* Open a handle to the given pool, even if the pool is currently in the FAULTED
* state.
*/
zpool_handle_t *
zpool_open_canfail(libzfs_handle_t *hdl, const char *pool)
{
zpool_handle_t *zhp;
boolean_t missing;
/*
* Make sure the pool name is valid.
*/
if (!zpool_name_valid(hdl, B_TRUE, pool)) {
(void) zfs_error_fmt(hdl, EZFS_INVALIDNAME,
dgettext(TEXT_DOMAIN, "cannot open '%s'"),
pool);
return (NULL);
}
zhp = zfs_alloc(hdl, sizeof (zpool_handle_t));
zhp->zpool_hdl = hdl;
(void) strlcpy(zhp->zpool_name, pool, sizeof (zhp->zpool_name));
if (zpool_refresh_stats(zhp, &missing) != 0) {
zpool_close(zhp);
return (NULL);
}
if (missing) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "no such pool"));
(void) zfs_error_fmt(hdl, EZFS_NOENT,
dgettext(TEXT_DOMAIN, "cannot open '%s'"), pool);
zpool_close(zhp);
return (NULL);
}
return (zhp);
}
/*
* Like the above, but silent on error. Used when iterating over pools (because
* the configuration cache may be out of date).
*/
int
zpool_open_silent(libzfs_handle_t *hdl, const char *pool, zpool_handle_t **ret)
{
zpool_handle_t *zhp;
boolean_t missing;
zhp = zfs_alloc(hdl, sizeof (zpool_handle_t));
zhp->zpool_hdl = hdl;
(void) strlcpy(zhp->zpool_name, pool, sizeof (zhp->zpool_name));
if (zpool_refresh_stats(zhp, &missing) != 0) {
zpool_close(zhp);
return (-1);
}
if (missing) {
zpool_close(zhp);
*ret = NULL;
return (0);
}
*ret = zhp;
return (0);
}
/*
* Similar to zpool_open_canfail(), but refuses to open pools in the faulted
* state.
*/
zpool_handle_t *
zpool_open(libzfs_handle_t *hdl, const char *pool)
{
zpool_handle_t *zhp;
if ((zhp = zpool_open_canfail(hdl, pool)) == NULL)
return (NULL);
if (zhp->zpool_state == POOL_STATE_UNAVAIL) {
(void) zfs_error_fmt(hdl, EZFS_POOLUNAVAIL,
dgettext(TEXT_DOMAIN, "cannot open '%s'"), zhp->zpool_name);
zpool_close(zhp);
return (NULL);
}
return (zhp);
}
/*
* Close the handle. Simply frees the memory associated with the handle.
*/
void
zpool_close(zpool_handle_t *zhp)
{
nvlist_free(zhp->zpool_config);
nvlist_free(zhp->zpool_old_config);
nvlist_free(zhp->zpool_props);
free(zhp);
}
/*
* Return the name of the pool.
*/
const char *
zpool_get_name(zpool_handle_t *zhp)
{
return (zhp->zpool_name);
}
/*
* Return the state of the pool (ACTIVE or UNAVAILABLE)
*/
int
zpool_get_state(zpool_handle_t *zhp)
{
return (zhp->zpool_state);
}
/*
* Check if vdev list contains a special vdev
*/
static boolean_t
zpool_has_special_vdev(nvlist_t *nvroot)
{
nvlist_t **child;
uint_t children;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child,
&children) == 0) {
for (uint_t c = 0; c < children; c++) {
const char *bias;
if (nvlist_lookup_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &bias) == 0 &&
strcmp(bias, VDEV_ALLOC_BIAS_SPECIAL) == 0) {
return (B_TRUE);
}
}
}
return (B_FALSE);
}
/*
* Check if vdev list contains a dRAID vdev
*/
static boolean_t
zpool_has_draid_vdev(nvlist_t *nvroot)
{
nvlist_t **child;
uint_t children;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) == 0) {
for (uint_t c = 0; c < children; c++) {
const char *type;
if (nvlist_lookup_string(child[c],
ZPOOL_CONFIG_TYPE, &type) == 0 &&
strcmp(type, VDEV_TYPE_DRAID) == 0) {
return (B_TRUE);
}
}
}
return (B_FALSE);
}
/*
* Output a dRAID top-level vdev name in to the provided buffer.
*/
static char *
zpool_draid_name(char *name, int len, uint64_t data, uint64_t parity,
uint64_t spares, uint64_t children)
{
snprintf(name, len, "%s%llu:%llud:%lluc:%llus",
VDEV_TYPE_DRAID, (u_longlong_t)parity, (u_longlong_t)data,
(u_longlong_t)children, (u_longlong_t)spares);
return (name);
}
/*
* Return B_TRUE if the provided name is a dRAID spare name.
*/
boolean_t
zpool_is_draid_spare(const char *name)
{
uint64_t spare_id, parity, vdev_id;
if (sscanf(name, VDEV_TYPE_DRAID "%llu-%llu-%llu",
(u_longlong_t *)&parity, (u_longlong_t *)&vdev_id,
(u_longlong_t *)&spare_id) == 3) {
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Create the named pool, using the provided vdev list. It is assumed
* that the consumer has already validated the contents of the nvlist, so we
* don't have to worry about error semantics.
*/
int
zpool_create(libzfs_handle_t *hdl, const char *pool, nvlist_t *nvroot,
nvlist_t *props, nvlist_t *fsprops)
{
zfs_cmd_t zc = {"\0"};
nvlist_t *zc_fsprops = NULL;
nvlist_t *zc_props = NULL;
nvlist_t *hidden_args = NULL;
uint8_t *wkeydata = NULL;
uint_t wkeylen = 0;
char errbuf[ERRBUFLEN];
int ret = -1;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot create '%s'"), pool);
if (!zpool_name_valid(hdl, B_FALSE, pool))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
zcmd_write_conf_nvlist(hdl, &zc, nvroot);
if (props) {
prop_flags_t flags = { .create = B_TRUE, .import = B_FALSE };
if ((zc_props = zpool_valid_proplist(hdl, pool, props,
SPA_VERSION_1, flags, errbuf)) == NULL) {
goto create_failed;
}
}
if (fsprops) {
uint64_t zoned;
const char *zonestr;
zoned = ((nvlist_lookup_string(fsprops,
zfs_prop_to_name(ZFS_PROP_ZONED), &zonestr) == 0) &&
strcmp(zonestr, "on") == 0);
if ((zc_fsprops = zfs_valid_proplist(hdl, ZFS_TYPE_FILESYSTEM,
fsprops, zoned, NULL, NULL, B_TRUE, errbuf)) == NULL) {
goto create_failed;
}
if (nvlist_exists(zc_fsprops,
zfs_prop_to_name(ZFS_PROP_SPECIAL_SMALL_BLOCKS)) &&
!zpool_has_special_vdev(nvroot)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"%s property requires a special vdev"),
zfs_prop_to_name(ZFS_PROP_SPECIAL_SMALL_BLOCKS));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto create_failed;
}
if (!zc_props &&
(nvlist_alloc(&zc_props, NV_UNIQUE_NAME, 0) != 0)) {
goto create_failed;
}
if (zfs_crypto_create(hdl, NULL, zc_fsprops, props, B_TRUE,
&wkeydata, &wkeylen) != 0) {
zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf);
goto create_failed;
}
if (nvlist_add_nvlist(zc_props,
ZPOOL_ROOTFS_PROPS, zc_fsprops) != 0) {
goto create_failed;
}
if (wkeydata != NULL) {
if (nvlist_alloc(&hidden_args, NV_UNIQUE_NAME, 0) != 0)
goto create_failed;
if (nvlist_add_uint8_array(hidden_args, "wkeydata",
wkeydata, wkeylen) != 0)
goto create_failed;
if (nvlist_add_nvlist(zc_props, ZPOOL_HIDDEN_ARGS,
hidden_args) != 0)
goto create_failed;
}
}
if (zc_props)
zcmd_write_src_nvlist(hdl, &zc, zc_props);
(void) strlcpy(zc.zc_name, pool, sizeof (zc.zc_name));
if ((ret = zfs_ioctl(hdl, ZFS_IOC_POOL_CREATE, &zc)) != 0) {
zcmd_free_nvlists(&zc);
nvlist_free(zc_props);
nvlist_free(zc_fsprops);
nvlist_free(hidden_args);
if (wkeydata != NULL)
free(wkeydata);
switch (errno) {
case EBUSY:
/*
* This can happen if the user has specified the same
* device multiple times. We can't reliably detect this
* until we try to add it and see we already have a
* label. This can also happen under if the device is
* part of an active md or lvm device.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more vdevs refer to the same device, or "
"one of\nthe devices is part of an active md or "
"lvm device"));
return (zfs_error(hdl, EZFS_BADDEV, errbuf));
case ERANGE:
/*
* This happens if the record size is smaller or larger
* than the allowed size range, or not a power of 2.
*
* NOTE: although zfs_valid_proplist is called earlier,
* this case may have slipped through since the
* pool does not exist yet and it is therefore
* impossible to read properties e.g. max blocksize
* from the pool.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"record size invalid"));
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
case EOVERFLOW:
/*
* This occurs when one of the devices is below
* SPA_MINDEVSIZE. Unfortunately, we can't detect which
* device was the problem device since there's no
* reliable way to determine device size from userland.
*/
{
char buf[64];
zfs_nicebytes(SPA_MINDEVSIZE, buf,
sizeof (buf));
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices is less than the "
"minimum size (%s)"), buf);
}
return (zfs_error(hdl, EZFS_BADDEV, errbuf));
case ENOSPC:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices is out of space"));
return (zfs_error(hdl, EZFS_BADDEV, errbuf));
case EINVAL:
if (zpool_has_draid_vdev(nvroot) &&
zfeature_lookup_name("draid", NULL) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dRAID vdevs are unsupported by the "
"kernel"));
return (zfs_error(hdl, EZFS_BADDEV, errbuf));
} else {
return (zpool_standard_error(hdl, errno,
errbuf));
}
default:
return (zpool_standard_error(hdl, errno, errbuf));
}
}
create_failed:
zcmd_free_nvlists(&zc);
nvlist_free(zc_props);
nvlist_free(zc_fsprops);
nvlist_free(hidden_args);
if (wkeydata != NULL)
free(wkeydata);
return (ret);
}
/*
* Destroy the given pool. It is up to the caller to ensure that there are no
* datasets left in the pool.
*/
int
zpool_destroy(zpool_handle_t *zhp, const char *log_str)
{
zfs_cmd_t zc = {"\0"};
zfs_handle_t *zfp = NULL;
libzfs_handle_t *hdl = zhp->zpool_hdl;
char errbuf[ERRBUFLEN];
if (zhp->zpool_state == POOL_STATE_ACTIVE &&
(zfp = zfs_open(hdl, zhp->zpool_name, ZFS_TYPE_FILESYSTEM)) == NULL)
return (-1);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_history = (uint64_t)(uintptr_t)log_str;
if (zfs_ioctl(hdl, ZFS_IOC_POOL_DESTROY, &zc) != 0) {
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot destroy '%s'"), zhp->zpool_name);
if (errno == EROFS) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices is read only"));
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
} else {
(void) zpool_standard_error(hdl, errno, errbuf);
}
if (zfp)
zfs_close(zfp);
return (-1);
}
if (zfp) {
remove_mountpoint(zfp);
zfs_close(zfp);
}
return (0);
}
/*
* Create a checkpoint in the given pool.
*/
int
zpool_checkpoint(zpool_handle_t *zhp)
{
libzfs_handle_t *hdl = zhp->zpool_hdl;
char errbuf[ERRBUFLEN];
int error;
error = lzc_pool_checkpoint(zhp->zpool_name);
if (error != 0) {
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot checkpoint '%s'"), zhp->zpool_name);
(void) zpool_standard_error(hdl, error, errbuf);
return (-1);
}
return (0);
}
/*
* Discard the checkpoint from the given pool.
*/
int
zpool_discard_checkpoint(zpool_handle_t *zhp)
{
libzfs_handle_t *hdl = zhp->zpool_hdl;
char errbuf[ERRBUFLEN];
int error;
error = lzc_pool_checkpoint_discard(zhp->zpool_name);
if (error != 0) {
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot discard checkpoint in '%s'"), zhp->zpool_name);
(void) zpool_standard_error(hdl, error, errbuf);
return (-1);
}
return (0);
}
/*
* Add the given vdevs to the pool. The caller must have already performed the
* necessary verification to ensure that the vdev specification is well-formed.
*/
int
zpool_add(zpool_handle_t *zhp, nvlist_t *nvroot)
{
zfs_cmd_t zc = {"\0"};
int ret;
libzfs_handle_t *hdl = zhp->zpool_hdl;
char errbuf[ERRBUFLEN];
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot add to '%s'"), zhp->zpool_name);
if (zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL) <
SPA_VERSION_SPARES &&
nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be "
"upgraded to add hot spares"));
return (zfs_error(hdl, EZFS_BADVERSION, errbuf));
}
if (zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL) <
SPA_VERSION_L2CACHE &&
nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be "
"upgraded to add cache devices"));
return (zfs_error(hdl, EZFS_BADVERSION, errbuf));
}
zcmd_write_conf_nvlist(hdl, &zc, nvroot);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_ADD, &zc) != 0) {
switch (errno) {
case EBUSY:
/*
* This can happen if the user has specified the same
* device multiple times. We can't reliably detect this
* until we try to add it and see we already have a
* label.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more vdevs refer to the same device"));
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case EINVAL:
if (zpool_has_draid_vdev(nvroot) &&
zfeature_lookup_name("draid", NULL) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dRAID vdevs are unsupported by the "
"kernel"));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid config; a pool with removing/"
"removed vdevs does not support adding "
"raidz or dRAID vdevs"));
}
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case EOVERFLOW:
/*
* This occurs when one of the devices is below
* SPA_MINDEVSIZE. Unfortunately, we can't detect which
* device was the problem device since there's no
* reliable way to determine device size from userland.
*/
{
char buf[64];
zfs_nicebytes(SPA_MINDEVSIZE, buf,
sizeof (buf));
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"device is less than the minimum "
"size (%s)"), buf);
}
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case ENOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded to add these vdevs"));
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
break;
default:
(void) zpool_standard_error(hdl, errno, errbuf);
}
ret = -1;
} else {
ret = 0;
}
zcmd_free_nvlists(&zc);
return (ret);
}
/*
* Exports the pool from the system. The caller must ensure that there are no
* mounted datasets in the pool.
*/
static int
zpool_export_common(zpool_handle_t *zhp, boolean_t force, boolean_t hardforce,
const char *log_str)
{
zfs_cmd_t zc = {"\0"};
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_cookie = force;
zc.zc_guid = hardforce;
zc.zc_history = (uint64_t)(uintptr_t)log_str;
if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_EXPORT, &zc) != 0) {
switch (errno) {
case EXDEV:
zfs_error_aux(zhp->zpool_hdl, dgettext(TEXT_DOMAIN,
"use '-f' to override the following errors:\n"
"'%s' has an active shared spare which could be"
" used by other pools once '%s' is exported."),
zhp->zpool_name, zhp->zpool_name);
return (zfs_error_fmt(zhp->zpool_hdl, EZFS_ACTIVE_SPARE,
dgettext(TEXT_DOMAIN, "cannot export '%s'"),
zhp->zpool_name));
default:
return (zpool_standard_error_fmt(zhp->zpool_hdl, errno,
dgettext(TEXT_DOMAIN, "cannot export '%s'"),
zhp->zpool_name));
}
}
return (0);
}
int
zpool_export(zpool_handle_t *zhp, boolean_t force, const char *log_str)
{
return (zpool_export_common(zhp, force, B_FALSE, log_str));
}
int
zpool_export_force(zpool_handle_t *zhp, const char *log_str)
{
return (zpool_export_common(zhp, B_TRUE, B_TRUE, log_str));
}
static void
zpool_rewind_exclaim(libzfs_handle_t *hdl, const char *name, boolean_t dryrun,
nvlist_t *config)
{
nvlist_t *nv = NULL;
uint64_t rewindto;
int64_t loss = -1;
struct tm t;
char timestr[128];
if (!hdl->libzfs_printerr || config == NULL)
return;
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, &nv) != 0 ||
nvlist_lookup_nvlist(nv, ZPOOL_CONFIG_REWIND_INFO, &nv) != 0) {
return;
}
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_LOAD_TIME, &rewindto) != 0)
return;
(void) nvlist_lookup_int64(nv, ZPOOL_CONFIG_REWIND_TIME, &loss);
if (localtime_r((time_t *)&rewindto, &t) != NULL &&
strftime(timestr, 128, "%c", &t) != 0) {
if (dryrun) {
(void) printf(dgettext(TEXT_DOMAIN,
"Would be able to return %s "
"to its state as of %s.\n"),
name, timestr);
} else {
(void) printf(dgettext(TEXT_DOMAIN,
"Pool %s returned to its state as of %s.\n"),
name, timestr);
}
if (loss > 120) {
(void) printf(dgettext(TEXT_DOMAIN,
"%s approximately %lld "),
dryrun ? "Would discard" : "Discarded",
((longlong_t)loss + 30) / 60);
(void) printf(dgettext(TEXT_DOMAIN,
"minutes of transactions.\n"));
} else if (loss > 0) {
(void) printf(dgettext(TEXT_DOMAIN,
"%s approximately %lld "),
dryrun ? "Would discard" : "Discarded",
(longlong_t)loss);
(void) printf(dgettext(TEXT_DOMAIN,
"seconds of transactions.\n"));
}
}
}
void
zpool_explain_recover(libzfs_handle_t *hdl, const char *name, int reason,
nvlist_t *config)
{
nvlist_t *nv = NULL;
int64_t loss = -1;
uint64_t edata = UINT64_MAX;
uint64_t rewindto;
struct tm t;
char timestr[128];
if (!hdl->libzfs_printerr)
return;
if (reason >= 0)
(void) printf(dgettext(TEXT_DOMAIN, "action: "));
else
(void) printf(dgettext(TEXT_DOMAIN, "\t"));
/* All attempted rewinds failed if ZPOOL_CONFIG_LOAD_TIME missing */
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, &nv) != 0 ||
nvlist_lookup_nvlist(nv, ZPOOL_CONFIG_REWIND_INFO, &nv) != 0 ||
nvlist_lookup_uint64(nv, ZPOOL_CONFIG_LOAD_TIME, &rewindto) != 0)
goto no_info;
(void) nvlist_lookup_int64(nv, ZPOOL_CONFIG_REWIND_TIME, &loss);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_LOAD_DATA_ERRORS,
&edata);
(void) printf(dgettext(TEXT_DOMAIN,
"Recovery is possible, but will result in some data loss.\n"));
if (localtime_r((time_t *)&rewindto, &t) != NULL &&
strftime(timestr, 128, "%c", &t) != 0) {
(void) printf(dgettext(TEXT_DOMAIN,
"\tReturning the pool to its state as of %s\n"
"\tshould correct the problem. "),
timestr);
} else {
(void) printf(dgettext(TEXT_DOMAIN,
"\tReverting the pool to an earlier state "
"should correct the problem.\n\t"));
}
if (loss > 120) {
(void) printf(dgettext(TEXT_DOMAIN,
"Approximately %lld minutes of data\n"
"\tmust be discarded, irreversibly. "),
((longlong_t)loss + 30) / 60);
} else if (loss > 0) {
(void) printf(dgettext(TEXT_DOMAIN,
"Approximately %lld seconds of data\n"
"\tmust be discarded, irreversibly. "),
(longlong_t)loss);
}
if (edata != 0 && edata != UINT64_MAX) {
if (edata == 1) {
(void) printf(dgettext(TEXT_DOMAIN,
"After rewind, at least\n"
"\tone persistent user-data error will remain. "));
} else {
(void) printf(dgettext(TEXT_DOMAIN,
"After rewind, several\n"
"\tpersistent user-data errors will remain. "));
}
}
(void) printf(dgettext(TEXT_DOMAIN,
"Recovery can be attempted\n\tby executing 'zpool %s -F %s'. "),
reason >= 0 ? "clear" : "import", name);
(void) printf(dgettext(TEXT_DOMAIN,
"A scrub of the pool\n"
"\tis strongly recommended after recovery.\n"));
return;
no_info:
(void) printf(dgettext(TEXT_DOMAIN,
"Destroy and re-create the pool from\n\ta backup source.\n"));
}
/*
* zpool_import() is a contracted interface. Should be kept the same
* if possible.
*
* Applications should use zpool_import_props() to import a pool with
* new properties value to be set.
*/
int
zpool_import(libzfs_handle_t *hdl, nvlist_t *config, const char *newname,
char *altroot)
{
nvlist_t *props = NULL;
int ret;
if (altroot != NULL) {
if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0) {
return (zfs_error_fmt(hdl, EZFS_NOMEM,
dgettext(TEXT_DOMAIN, "cannot import '%s'"),
newname));
}
if (nvlist_add_string(props,
zpool_prop_to_name(ZPOOL_PROP_ALTROOT), altroot) != 0 ||
nvlist_add_string(props,
zpool_prop_to_name(ZPOOL_PROP_CACHEFILE), "none") != 0) {
nvlist_free(props);
return (zfs_error_fmt(hdl, EZFS_NOMEM,
dgettext(TEXT_DOMAIN, "cannot import '%s'"),
newname));
}
}
ret = zpool_import_props(hdl, config, newname, props,
ZFS_IMPORT_NORMAL);
nvlist_free(props);
return (ret);
}
static void
print_vdev_tree(libzfs_handle_t *hdl, const char *name, nvlist_t *nv,
int indent)
{
nvlist_t **child;
uint_t c, children;
char *vname;
uint64_t is_log = 0;
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_LOG,
&is_log);
if (name != NULL)
(void) printf("\t%*s%s%s\n", indent, "", name,
is_log ? " [log]" : "");
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
return;
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(hdl, NULL, child[c], VDEV_NAME_TYPE_ID);
print_vdev_tree(hdl, vname, child[c], indent + 2);
free(vname);
}
}
void
zpool_print_unsup_feat(nvlist_t *config)
{
nvlist_t *nvinfo, *unsup_feat;
nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
unsup_feat = fnvlist_lookup_nvlist(nvinfo, ZPOOL_CONFIG_UNSUP_FEAT);
for (nvpair_t *nvp = nvlist_next_nvpair(unsup_feat, NULL);
nvp != NULL; nvp = nvlist_next_nvpair(unsup_feat, nvp)) {
const char *desc = fnvpair_value_string(nvp);
if (strlen(desc) > 0)
(void) printf("\t%s (%s)\n", nvpair_name(nvp), desc);
else
(void) printf("\t%s\n", nvpair_name(nvp));
}
}
/*
* Import the given pool using the known configuration and a list of
* properties to be set. The configuration should have come from
* zpool_find_import(). The 'newname' parameters control whether the pool
* is imported with a different name.
*/
int
zpool_import_props(libzfs_handle_t *hdl, nvlist_t *config, const char *newname,
nvlist_t *props, int flags)
{
zfs_cmd_t zc = {"\0"};
zpool_load_policy_t policy;
nvlist_t *nv = NULL;
nvlist_t *nvinfo = NULL;
nvlist_t *missing = NULL;
const char *thename;
const char *origname;
int ret;
int error = 0;
char errbuf[ERRBUFLEN];
origname = fnvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME);
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot import pool '%s'"), origname);
if (newname != NULL) {
if (!zpool_name_valid(hdl, B_FALSE, newname))
return (zfs_error_fmt(hdl, EZFS_INVALIDNAME,
dgettext(TEXT_DOMAIN, "cannot import '%s'"),
newname));
thename = newname;
} else {
thename = origname;
}
if (props != NULL) {
uint64_t version;
prop_flags_t flags = { .create = B_FALSE, .import = B_TRUE };
version = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION);
if ((props = zpool_valid_proplist(hdl, origname,
props, version, flags, errbuf)) == NULL)
return (-1);
zcmd_write_src_nvlist(hdl, &zc, props);
nvlist_free(props);
}
(void) strlcpy(zc.zc_name, thename, sizeof (zc.zc_name));
zc.zc_guid = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID);
zcmd_write_conf_nvlist(hdl, &zc, config);
zcmd_alloc_dst_nvlist(hdl, &zc, zc.zc_nvlist_conf_size * 2);
zc.zc_cookie = flags;
while ((ret = zfs_ioctl(hdl, ZFS_IOC_POOL_IMPORT, &zc)) != 0 &&
errno == ENOMEM)
zcmd_expand_dst_nvlist(hdl, &zc);
if (ret != 0)
error = errno;
(void) zcmd_read_dst_nvlist(hdl, &zc, &nv);
zcmd_free_nvlists(&zc);
zpool_get_load_policy(config, &policy);
if (error) {
char desc[1024];
char aux[256];
/*
* Dry-run failed, but we print out what success
* looks like if we found a best txg
*/
if (policy.zlp_rewind & ZPOOL_TRY_REWIND) {
zpool_rewind_exclaim(hdl, newname ? origname : thename,
B_TRUE, nv);
nvlist_free(nv);
return (-1);
}
if (newname == NULL)
(void) snprintf(desc, sizeof (desc),
dgettext(TEXT_DOMAIN, "cannot import '%s'"),
thename);
else
(void) snprintf(desc, sizeof (desc),
dgettext(TEXT_DOMAIN, "cannot import '%s' as '%s'"),
origname, thename);
switch (error) {
case ENOTSUP:
if (nv != NULL && nvlist_lookup_nvlist(nv,
ZPOOL_CONFIG_LOAD_INFO, &nvinfo) == 0 &&
nvlist_exists(nvinfo, ZPOOL_CONFIG_UNSUP_FEAT)) {
(void) printf(dgettext(TEXT_DOMAIN, "This "
"pool uses the following feature(s) not "
"supported by this system:\n"));
zpool_print_unsup_feat(nv);
if (nvlist_exists(nvinfo,
ZPOOL_CONFIG_CAN_RDONLY)) {
(void) printf(dgettext(TEXT_DOMAIN,
"All unsupported features are only "
"required for writing to the pool."
"\nThe pool can be imported using "
"'-o readonly=on'.\n"));
}
}
/*
* Unsupported version.
*/
(void) zfs_error(hdl, EZFS_BADVERSION, desc);
break;
case EREMOTEIO:
if (nv != NULL && nvlist_lookup_nvlist(nv,
ZPOOL_CONFIG_LOAD_INFO, &nvinfo) == 0) {
const char *hostname = "<unknown>";
uint64_t hostid = 0;
mmp_state_t mmp_state;
mmp_state = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_STATE);
if (nvlist_exists(nvinfo,
ZPOOL_CONFIG_MMP_HOSTNAME))
hostname = fnvlist_lookup_string(nvinfo,
ZPOOL_CONFIG_MMP_HOSTNAME);
if (nvlist_exists(nvinfo,
ZPOOL_CONFIG_MMP_HOSTID))
hostid = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_HOSTID);
if (mmp_state == MMP_STATE_ACTIVE) {
(void) snprintf(aux, sizeof (aux),
dgettext(TEXT_DOMAIN, "pool is imp"
"orted on host '%s' (hostid=%lx).\n"
"Export the pool on the other "
"system, then run 'zpool import'."),
hostname, (unsigned long) hostid);
} else if (mmp_state == MMP_STATE_NO_HOSTID) {
(void) snprintf(aux, sizeof (aux),
dgettext(TEXT_DOMAIN, "pool has "
"the multihost property on and "
"the\nsystem's hostid is not set. "
"Set a unique system hostid with "
"the zgenhostid(8) command.\n"));
}
(void) zfs_error_aux(hdl, "%s", aux);
}
(void) zfs_error(hdl, EZFS_ACTIVE_POOL, desc);
break;
case EINVAL:
(void) zfs_error(hdl, EZFS_INVALCONFIG, desc);
break;
case EROFS:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices is read only"));
(void) zfs_error(hdl, EZFS_BADDEV, desc);
break;
case ENXIO:
if (nv && nvlist_lookup_nvlist(nv,
ZPOOL_CONFIG_LOAD_INFO, &nvinfo) == 0 &&
nvlist_lookup_nvlist(nvinfo,
ZPOOL_CONFIG_MISSING_DEVICES, &missing) == 0) {
(void) printf(dgettext(TEXT_DOMAIN,
"The devices below are missing or "
"corrupted, use '-m' to import the pool "
"anyway:\n"));
print_vdev_tree(hdl, NULL, missing, 2);
(void) printf("\n");
}
(void) zpool_standard_error(hdl, error, desc);
break;
case EEXIST:
(void) zpool_standard_error(hdl, error, desc);
break;
case EBUSY:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices are already in use\n"));
(void) zfs_error(hdl, EZFS_BADDEV, desc);
break;
case ENAMETOOLONG:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"new name of at least one dataset is longer than "
"the maximum allowable length"));
(void) zfs_error(hdl, EZFS_NAMETOOLONG, desc);
break;
default:
(void) zpool_standard_error(hdl, error, desc);
zpool_explain_recover(hdl,
newname ? origname : thename, -error, nv);
break;
}
nvlist_free(nv);
ret = -1;
} else {
zpool_handle_t *zhp;
/*
* This should never fail, but play it safe anyway.
*/
if (zpool_open_silent(hdl, thename, &zhp) != 0)
ret = -1;
else if (zhp != NULL)
zpool_close(zhp);
if (policy.zlp_rewind &
(ZPOOL_DO_REWIND | ZPOOL_TRY_REWIND)) {
zpool_rewind_exclaim(hdl, newname ? origname : thename,
((policy.zlp_rewind & ZPOOL_TRY_REWIND) != 0), nv);
}
nvlist_free(nv);
}
return (ret);
}
/*
* Translate vdev names to guids. If a vdev_path is determined to be
* unsuitable then a vd_errlist is allocated and the vdev path and errno
* are added to it.
*/
static int
zpool_translate_vdev_guids(zpool_handle_t *zhp, nvlist_t *vds,
nvlist_t *vdev_guids, nvlist_t *guids_to_paths, nvlist_t **vd_errlist)
{
nvlist_t *errlist = NULL;
int error = 0;
for (nvpair_t *elem = nvlist_next_nvpair(vds, NULL); elem != NULL;
elem = nvlist_next_nvpair(vds, elem)) {
boolean_t spare, cache;
const char *vd_path = nvpair_name(elem);
nvlist_t *tgt = zpool_find_vdev(zhp, vd_path, &spare, &cache,
NULL);
if ((tgt == NULL) || cache || spare) {
if (errlist == NULL) {
errlist = fnvlist_alloc();
error = EINVAL;
}
uint64_t err = (tgt == NULL) ? EZFS_NODEVICE :
(spare ? EZFS_ISSPARE : EZFS_ISL2CACHE);
fnvlist_add_int64(errlist, vd_path, err);
continue;
}
uint64_t guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
fnvlist_add_uint64(vdev_guids, vd_path, guid);
char msg[MAXNAMELEN];
(void) snprintf(msg, sizeof (msg), "%llu", (u_longlong_t)guid);
fnvlist_add_string(guids_to_paths, msg, vd_path);
}
if (error != 0) {
verify(errlist != NULL);
if (vd_errlist != NULL)
*vd_errlist = errlist;
else
fnvlist_free(errlist);
}
return (error);
}
static int
xlate_init_err(int err)
{
switch (err) {
case ENODEV:
return (EZFS_NODEVICE);
case EINVAL:
case EROFS:
return (EZFS_BADDEV);
case EBUSY:
return (EZFS_INITIALIZING);
case ESRCH:
return (EZFS_NO_INITIALIZE);
}
return (err);
}
/*
* Begin, suspend, cancel, or uninit (clear) the initialization (initializing
* of all free blocks) for the given vdevs in the given pool.
*/
static int
zpool_initialize_impl(zpool_handle_t *zhp, pool_initialize_func_t cmd_type,
nvlist_t *vds, boolean_t wait)
{
int err;
nvlist_t *vdev_guids = fnvlist_alloc();
nvlist_t *guids_to_paths = fnvlist_alloc();
nvlist_t *vd_errlist = NULL;
nvlist_t *errlist;
nvpair_t *elem;
err = zpool_translate_vdev_guids(zhp, vds, vdev_guids,
guids_to_paths, &vd_errlist);
if (err != 0) {
verify(vd_errlist != NULL);
goto list_errors;
}
err = lzc_initialize(zhp->zpool_name, cmd_type,
vdev_guids, &errlist);
if (err != 0) {
if (errlist != NULL && nvlist_lookup_nvlist(errlist,
ZPOOL_INITIALIZE_VDEVS, &vd_errlist) == 0) {
goto list_errors;
}
if (err == EINVAL && cmd_type == POOL_INITIALIZE_UNINIT) {
zfs_error_aux(zhp->zpool_hdl, dgettext(TEXT_DOMAIN,
"uninitialize is not supported by kernel"));
}
(void) zpool_standard_error(zhp->zpool_hdl, err,
dgettext(TEXT_DOMAIN, "operation failed"));
goto out;
}
if (wait) {
for (elem = nvlist_next_nvpair(vdev_guids, NULL); elem != NULL;
elem = nvlist_next_nvpair(vdev_guids, elem)) {
uint64_t guid = fnvpair_value_uint64(elem);
err = lzc_wait_tag(zhp->zpool_name,
ZPOOL_WAIT_INITIALIZE, guid, NULL);
if (err != 0) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl,
err, dgettext(TEXT_DOMAIN, "error "
"waiting for '%s' to initialize"),
nvpair_name(elem));
goto out;
}
}
}
goto out;
list_errors:
for (elem = nvlist_next_nvpair(vd_errlist, NULL); elem != NULL;
elem = nvlist_next_nvpair(vd_errlist, elem)) {
int64_t vd_error = xlate_init_err(fnvpair_value_int64(elem));
const char *path;
if (nvlist_lookup_string(guids_to_paths, nvpair_name(elem),
&path) != 0)
path = nvpair_name(elem);
(void) zfs_error_fmt(zhp->zpool_hdl, vd_error,
"cannot initialize '%s'", path);
}
out:
fnvlist_free(vdev_guids);
fnvlist_free(guids_to_paths);
if (vd_errlist != NULL)
fnvlist_free(vd_errlist);
return (err == 0 ? 0 : -1);
}
int
zpool_initialize(zpool_handle_t *zhp, pool_initialize_func_t cmd_type,
nvlist_t *vds)
{
return (zpool_initialize_impl(zhp, cmd_type, vds, B_FALSE));
}
int
zpool_initialize_wait(zpool_handle_t *zhp, pool_initialize_func_t cmd_type,
nvlist_t *vds)
{
return (zpool_initialize_impl(zhp, cmd_type, vds, B_TRUE));
}
static int
xlate_trim_err(int err)
{
switch (err) {
case ENODEV:
return (EZFS_NODEVICE);
case EINVAL:
case EROFS:
return (EZFS_BADDEV);
case EBUSY:
return (EZFS_TRIMMING);
case ESRCH:
return (EZFS_NO_TRIM);
case EOPNOTSUPP:
return (EZFS_TRIM_NOTSUP);
}
return (err);
}
static int
zpool_trim_wait(zpool_handle_t *zhp, nvlist_t *vdev_guids)
{
int err;
nvpair_t *elem;
for (elem = nvlist_next_nvpair(vdev_guids, NULL); elem != NULL;
elem = nvlist_next_nvpair(vdev_guids, elem)) {
uint64_t guid = fnvpair_value_uint64(elem);
err = lzc_wait_tag(zhp->zpool_name,
ZPOOL_WAIT_TRIM, guid, NULL);
if (err != 0) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl,
err, dgettext(TEXT_DOMAIN, "error "
"waiting to trim '%s'"), nvpair_name(elem));
return (err);
}
}
return (0);
}
/*
* Check errlist and report any errors, omitting ones which should be
* suppressed. Returns B_TRUE if any errors were reported.
*/
static boolean_t
check_trim_errs(zpool_handle_t *zhp, trimflags_t *trim_flags,
nvlist_t *guids_to_paths, nvlist_t *vds, nvlist_t *errlist)
{
nvpair_t *elem;
boolean_t reported_errs = B_FALSE;
int num_vds = 0;
int num_suppressed_errs = 0;
for (elem = nvlist_next_nvpair(vds, NULL);
elem != NULL; elem = nvlist_next_nvpair(vds, elem)) {
num_vds++;
}
for (elem = nvlist_next_nvpair(errlist, NULL);
elem != NULL; elem = nvlist_next_nvpair(errlist, elem)) {
int64_t vd_error = xlate_trim_err(fnvpair_value_int64(elem));
const char *path;
/*
* If only the pool was specified, and it was not a secure
* trim then suppress warnings for individual vdevs which
* do not support trimming.
*/
if (vd_error == EZFS_TRIM_NOTSUP &&
trim_flags->fullpool &&
!trim_flags->secure) {
num_suppressed_errs++;
continue;
}
reported_errs = B_TRUE;
if (nvlist_lookup_string(guids_to_paths, nvpair_name(elem),
&path) != 0)
path = nvpair_name(elem);
(void) zfs_error_fmt(zhp->zpool_hdl, vd_error,
"cannot trim '%s'", path);
}
if (num_suppressed_errs == num_vds) {
(void) zfs_error_aux(zhp->zpool_hdl, dgettext(TEXT_DOMAIN,
"no devices in pool support trim operations"));
(void) (zfs_error(zhp->zpool_hdl, EZFS_TRIM_NOTSUP,
dgettext(TEXT_DOMAIN, "cannot trim")));
reported_errs = B_TRUE;
}
return (reported_errs);
}
/*
* Begin, suspend, or cancel the TRIM (discarding of all free blocks) for
* the given vdevs in the given pool.
*/
int
zpool_trim(zpool_handle_t *zhp, pool_trim_func_t cmd_type, nvlist_t *vds,
trimflags_t *trim_flags)
{
int err;
int retval = 0;
nvlist_t *vdev_guids = fnvlist_alloc();
nvlist_t *guids_to_paths = fnvlist_alloc();
nvlist_t *errlist = NULL;
err = zpool_translate_vdev_guids(zhp, vds, vdev_guids,
guids_to_paths, &errlist);
if (err != 0) {
check_trim_errs(zhp, trim_flags, guids_to_paths, vds, errlist);
retval = -1;
goto out;
}
err = lzc_trim(zhp->zpool_name, cmd_type, trim_flags->rate,
trim_flags->secure, vdev_guids, &errlist);
if (err != 0) {
nvlist_t *vd_errlist;
if (errlist != NULL && nvlist_lookup_nvlist(errlist,
ZPOOL_TRIM_VDEVS, &vd_errlist) == 0) {
if (check_trim_errs(zhp, trim_flags, guids_to_paths,
vds, vd_errlist)) {
retval = -1;
goto out;
}
} else {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "operation failed"));
zpool_standard_error(zhp->zpool_hdl, err, errbuf);
retval = -1;
goto out;
}
}
if (trim_flags->wait)
retval = zpool_trim_wait(zhp, vdev_guids);
out:
if (errlist != NULL)
fnvlist_free(errlist);
fnvlist_free(vdev_guids);
fnvlist_free(guids_to_paths);
return (retval);
}
/*
* Scan the pool.
*/
int
zpool_scan(zpool_handle_t *zhp, pool_scan_func_t func, pool_scrub_cmd_t cmd)
{
char errbuf[ERRBUFLEN];
int err;
libzfs_handle_t *hdl = zhp->zpool_hdl;
nvlist_t *args = fnvlist_alloc();
fnvlist_add_uint64(args, "scan_type", (uint64_t)func);
fnvlist_add_uint64(args, "scan_command", (uint64_t)cmd);
err = lzc_scrub(ZFS_IOC_POOL_SCRUB, zhp->zpool_name, args, NULL);
fnvlist_free(args);
if (err == 0) {
return (0);
} else if (err == ZFS_ERR_IOC_CMD_UNAVAIL) {
zfs_cmd_t zc = {"\0"};
(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);
}
/*
* An ECANCELED on a scrub means one of the following:
* 1. we resumed a paused scrub.
* 2. we resumed a paused error scrub.
* 3. Error scrub is not run because of no error log.
*/
if (err == ECANCELED && (func == POOL_SCAN_SCRUB ||
func == POOL_SCAN_ERRORSCRUB) && cmd == POOL_SCRUB_NORMAL)
return (0);
/*
* The following cases have been handled here:
* 1. Paused a scrub/error scrub if there is none in progress.
*/
if (err == ENOENT && func != POOL_SCAN_NONE && cmd ==
POOL_SCRUB_PAUSE) {
return (0);
}
ASSERT3U(func, >=, POOL_SCAN_NONE);
ASSERT3U(func, <, POOL_SCAN_FUNCS);
if (func == POOL_SCAN_SCRUB || func == POOL_SCAN_ERRORSCRUB) {
if (cmd == POOL_SCRUB_PAUSE) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot pause scrubbing %s"),
zhp->zpool_name);
} else {
assert(cmd == POOL_SCRUB_NORMAL);
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot scrub %s"),
zhp->zpool_name);
}
} else if (func == POOL_SCAN_RESILVER) {
assert(cmd == POOL_SCRUB_NORMAL);
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot restart resilver on %s"), zhp->zpool_name);
} else if (func == POOL_SCAN_NONE) {
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot cancel scrubbing %s"), zhp->zpool_name);
} else {
assert(!"unexpected result");
}
/*
* With EBUSY, five cases are possible:
*
* Current state Requested
* 1. Normal Scrub Running Normal Scrub or Error Scrub
* 2. Normal Scrub Paused Error Scrub
* 3. Normal Scrub Paused Pause Normal Scrub
* 4. Error Scrub Running Normal Scrub or Error Scrub
* 5. Error Scrub Paused Pause Error Scrub
* 6. Resilvering Anything else
*/
if (err == EBUSY) {
nvlist_t *nvroot;
pool_scan_stat_t *ps = NULL;
uint_t psc;
nvroot = fnvlist_lookup_nvlist(zhp->zpool_config,
ZPOOL_CONFIG_VDEV_TREE);
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&ps, &psc);
if (ps && ps->pss_func == POOL_SCAN_SCRUB &&
ps->pss_state == DSS_SCANNING) {
if (ps->pss_pass_scrub_pause == 0) {
/* handles case 1 */
assert(cmd == POOL_SCRUB_NORMAL);
return (zfs_error(hdl, EZFS_SCRUBBING,
errbuf));
} else {
if (func == POOL_SCAN_ERRORSCRUB) {
/* handles case 2 */
ASSERT3U(cmd, ==, POOL_SCRUB_NORMAL);
return (zfs_error(hdl,
EZFS_SCRUB_PAUSED_TO_CANCEL,
errbuf));
} else {
/* handles case 3 */
ASSERT3U(func, ==, POOL_SCAN_SCRUB);
ASSERT3U(cmd, ==, POOL_SCRUB_PAUSE);
return (zfs_error(hdl,
EZFS_SCRUB_PAUSED, errbuf));
}
}
} else if (ps &&
ps->pss_error_scrub_func == POOL_SCAN_ERRORSCRUB &&
ps->pss_error_scrub_state == DSS_ERRORSCRUBBING) {
if (ps->pss_pass_error_scrub_pause == 0) {
/* handles case 4 */
ASSERT3U(cmd, ==, POOL_SCRUB_NORMAL);
return (zfs_error(hdl, EZFS_ERRORSCRUBBING,
errbuf));
} else {
/* handles case 5 */
ASSERT3U(func, ==, POOL_SCAN_ERRORSCRUB);
ASSERT3U(cmd, ==, POOL_SCRUB_PAUSE);
return (zfs_error(hdl, EZFS_ERRORSCRUB_PAUSED,
errbuf));
}
} else {
/* handles case 6 */
return (zfs_error(hdl, EZFS_RESILVERING, errbuf));
}
} else if (err == ENOENT) {
return (zfs_error(hdl, EZFS_NO_SCRUB, errbuf));
} else if (err == ENOTSUP && func == POOL_SCAN_RESILVER) {
return (zfs_error(hdl, EZFS_NO_RESILVER_DEFER, errbuf));
} else {
return (zpool_standard_error(hdl, err, errbuf));
}
}
/*
* Find a vdev that matches the search criteria specified. We use the
* the nvpair name to determine how we should look for the device.
* 'avail_spare' is set to TRUE if the provided guid refers to an AVAIL
* spare; but FALSE if its an INUSE spare.
*/
static nvlist_t *
vdev_to_nvlist_iter(nvlist_t *nv, nvlist_t *search, boolean_t *avail_spare,
boolean_t *l2cache, boolean_t *log)
{
uint_t c, children;
nvlist_t **child;
nvlist_t *ret;
uint64_t is_log;
const char *srchkey;
nvpair_t *pair = nvlist_next_nvpair(search, NULL);
/* Nothing to look for */
if (search == NULL || pair == NULL)
return (NULL);
/* Obtain the key we will use to search */
srchkey = nvpair_name(pair);
switch (nvpair_type(pair)) {
case DATA_TYPE_UINT64:
if (strcmp(srchkey, ZPOOL_CONFIG_GUID) == 0) {
uint64_t srchval = fnvpair_value_uint64(pair);
uint64_t theguid = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_GUID);
if (theguid == srchval)
return (nv);
}
break;
case DATA_TYPE_STRING: {
const char *srchval, *val;
srchval = fnvpair_value_string(pair);
if (nvlist_lookup_string(nv, srchkey, &val) != 0)
break;
/*
* Search for the requested value. Special cases:
*
* - ZPOOL_CONFIG_PATH for whole disk entries. These end in
* "-part1", or "p1". The suffix is hidden from the user,
* but included in the string, so this matches around it.
* - ZPOOL_CONFIG_PATH for short names zfs_strcmp_shortname()
* is used to check all possible expanded paths.
* - looking for a top-level vdev name (i.e. ZPOOL_CONFIG_TYPE).
*
* Otherwise, all other searches are simple string compares.
*/
if (strcmp(srchkey, ZPOOL_CONFIG_PATH) == 0) {
uint64_t wholedisk = 0;
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK,
&wholedisk);
if (zfs_strcmp_pathname(srchval, val, wholedisk) == 0)
return (nv);
} else if (strcmp(srchkey, ZPOOL_CONFIG_TYPE) == 0) {
char *type, *idx, *end, *p;
uint64_t id, vdev_id;
/*
* Determine our vdev type, keeping in mind
* that the srchval is composed of a type and
* vdev id pair (i.e. mirror-4).
*/
if ((type = strdup(srchval)) == NULL)
return (NULL);
if ((p = strrchr(type, '-')) == NULL) {
free(type);
break;
}
idx = p + 1;
*p = '\0';
/*
* If the types don't match then keep looking.
*/
if (strncmp(val, type, strlen(val)) != 0) {
free(type);
break;
}
verify(zpool_vdev_is_interior(type));
id = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID);
errno = 0;
vdev_id = strtoull(idx, &end, 10);
/*
* If we are looking for a raidz and a parity is
* specified, make sure it matches.
*/
int rzlen = strlen(VDEV_TYPE_RAIDZ);
assert(rzlen == strlen(VDEV_TYPE_DRAID));
int typlen = strlen(type);
if ((strncmp(type, VDEV_TYPE_RAIDZ, rzlen) == 0 ||
strncmp(type, VDEV_TYPE_DRAID, rzlen) == 0) &&
typlen != rzlen) {
uint64_t vdev_parity;
int parity = *(type + rzlen) - '0';
if (parity <= 0 || parity > 3 ||
(typlen - rzlen) != 1) {
/*
* Nonsense parity specified, can
* never match
*/
free(type);
return (NULL);
}
vdev_parity = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_NPARITY);
if ((int)vdev_parity != parity) {
free(type);
break;
}
}
free(type);
if (errno != 0)
return (NULL);
/*
* Now verify that we have the correct vdev id.
*/
if (vdev_id == id)
return (nv);
}
/*
* Common case
*/
if (strcmp(srchval, val) == 0)
return (nv);
break;
}
default:
break;
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
return (NULL);
for (c = 0; c < children; c++) {
if ((ret = vdev_to_nvlist_iter(child[c], search,
avail_spare, l2cache, NULL)) != NULL) {
/*
* The 'is_log' value is only set for the toplevel
* vdev, not the leaf vdevs. So we always lookup the
* log device from the root of the vdev tree (where
* 'log' is non-NULL).
*/
if (log != NULL &&
nvlist_lookup_uint64(child[c],
ZPOOL_CONFIG_IS_LOG, &is_log) == 0 &&
is_log) {
*log = B_TRUE;
}
return (ret);
}
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES,
&child, &children) == 0) {
for (c = 0; c < children; c++) {
if ((ret = vdev_to_nvlist_iter(child[c], search,
avail_spare, l2cache, NULL)) != NULL) {
*avail_spare = B_TRUE;
return (ret);
}
}
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE,
&child, &children) == 0) {
for (c = 0; c < children; c++) {
if ((ret = vdev_to_nvlist_iter(child[c], search,
avail_spare, l2cache, NULL)) != NULL) {
*l2cache = B_TRUE;
return (ret);
}
}
}
return (NULL);
}
/*
* Given a physical path or guid, find the associated vdev.
*/
nvlist_t *
zpool_find_vdev_by_physpath(zpool_handle_t *zhp, const char *ppath,
boolean_t *avail_spare, boolean_t *l2cache, boolean_t *log)
{
nvlist_t *search, *nvroot, *ret;
uint64_t guid;
char *end;
search = fnvlist_alloc();
guid = strtoull(ppath, &end, 0);
if (guid != 0 && *end == '\0') {
fnvlist_add_uint64(search, ZPOOL_CONFIG_GUID, guid);
} else {
fnvlist_add_string(search, ZPOOL_CONFIG_PHYS_PATH, ppath);
}
nvroot = fnvlist_lookup_nvlist(zhp->zpool_config,
ZPOOL_CONFIG_VDEV_TREE);
*avail_spare = B_FALSE;
*l2cache = B_FALSE;
if (log != NULL)
*log = B_FALSE;
ret = vdev_to_nvlist_iter(nvroot, search, avail_spare, l2cache, log);
fnvlist_free(search);
return (ret);
}
/*
* Determine if we have an "interior" top-level vdev (i.e mirror/raidz).
*/
static boolean_t
zpool_vdev_is_interior(const char *name)
{
if (strncmp(name, VDEV_TYPE_RAIDZ, strlen(VDEV_TYPE_RAIDZ)) == 0 ||
strncmp(name, VDEV_TYPE_SPARE, strlen(VDEV_TYPE_SPARE)) == 0 ||
strncmp(name,
VDEV_TYPE_REPLACING, strlen(VDEV_TYPE_REPLACING)) == 0 ||
strncmp(name, VDEV_TYPE_ROOT, strlen(VDEV_TYPE_ROOT)) == 0 ||
strncmp(name, VDEV_TYPE_MIRROR, strlen(VDEV_TYPE_MIRROR)) == 0)
return (B_TRUE);
if (strncmp(name, VDEV_TYPE_DRAID, strlen(VDEV_TYPE_DRAID)) == 0 &&
!zpool_is_draid_spare(name))
return (B_TRUE);
return (B_FALSE);
}
nvlist_t *
zpool_find_vdev(zpool_handle_t *zhp, const char *path, boolean_t *avail_spare,
boolean_t *l2cache, boolean_t *log)
{
char *end;
nvlist_t *nvroot, *search, *ret;
uint64_t guid;
search = fnvlist_alloc();
guid = strtoull(path, &end, 0);
if (guid != 0 && *end == '\0') {
fnvlist_add_uint64(search, ZPOOL_CONFIG_GUID, guid);
} else if (zpool_vdev_is_interior(path)) {
fnvlist_add_string(search, ZPOOL_CONFIG_TYPE, path);
} else {
fnvlist_add_string(search, ZPOOL_CONFIG_PATH, path);
}
nvroot = fnvlist_lookup_nvlist(zhp->zpool_config,
ZPOOL_CONFIG_VDEV_TREE);
*avail_spare = B_FALSE;
*l2cache = B_FALSE;
if (log != NULL)
*log = B_FALSE;
ret = vdev_to_nvlist_iter(nvroot, search, avail_spare, l2cache, log);
fnvlist_free(search);
return (ret);
}
/*
* Convert a vdev path to a GUID. Returns GUID or 0 on error.
*
* If is_spare, is_l2cache, or is_log is non-NULL, then store within it
* if the VDEV is a spare, l2cache, or log device. If they're NULL then
* ignore them.
*/
static uint64_t
zpool_vdev_path_to_guid_impl(zpool_handle_t *zhp, const char *path,
boolean_t *is_spare, boolean_t *is_l2cache, boolean_t *is_log)
{
boolean_t spare = B_FALSE, l2cache = B_FALSE, log = B_FALSE;
nvlist_t *tgt;
if ((tgt = zpool_find_vdev(zhp, path, &spare, &l2cache,
&log)) == NULL)
return (0);
if (is_spare != NULL)
*is_spare = spare;
if (is_l2cache != NULL)
*is_l2cache = l2cache;
if (is_log != NULL)
*is_log = log;
return (fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID));
}
/* Convert a vdev path to a GUID. Returns GUID or 0 on error. */
uint64_t
zpool_vdev_path_to_guid(zpool_handle_t *zhp, const char *path)
{
return (zpool_vdev_path_to_guid_impl(zhp, path, NULL, NULL, NULL));
}
/*
* Bring the specified vdev online. The 'flags' parameter is a set of the
* ZFS_ONLINE_* flags.
*/
int
zpool_vdev_online(zpool_handle_t *zhp, const char *path, int flags,
vdev_state_t *newstate)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache, islog;
libzfs_handle_t *hdl = zhp->zpool_hdl;
if (flags & ZFS_ONLINE_EXPAND) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot expand %s"), path);
} else {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot online %s"), path);
}
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
&islog)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
if (!(flags & ZFS_ONLINE_SPARE) && avail_spare)
return (zfs_error(hdl, EZFS_ISSPARE, errbuf));
#ifndef __FreeBSD__
const char *pathname;
if ((flags & ZFS_ONLINE_EXPAND ||
zpool_get_prop_int(zhp, ZPOOL_PROP_AUTOEXPAND, NULL)) &&
nvlist_lookup_string(tgt, ZPOOL_CONFIG_PATH, &pathname) == 0) {
uint64_t wholedisk = 0;
(void) nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_WHOLE_DISK,
&wholedisk);
/*
* XXX - L2ARC 1.0 devices can't support expansion.
*/
if (l2cache) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot expand cache devices"));
return (zfs_error(hdl, EZFS_VDEVNOTSUP, errbuf));
}
if (wholedisk) {
const char *fullpath = path;
char buf[MAXPATHLEN];
int error;
if (path[0] != '/') {
error = zfs_resolve_shortname(path, buf,
sizeof (buf));
if (error != 0)
return (zfs_error(hdl, EZFS_NODEVICE,
errbuf));
fullpath = buf;
}
error = zpool_relabel_disk(hdl, fullpath, errbuf);
if (error != 0)
return (error);
}
}
#endif
zc.zc_cookie = VDEV_STATE_ONLINE;
zc.zc_obj = flags;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) != 0) {
if (errno == EINVAL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "was split "
"from this pool into a new one. Use '%s' "
"instead"), "zpool detach");
return (zfs_error(hdl, EZFS_POSTSPLIT_ONLINE, errbuf));
}
return (zpool_standard_error(hdl, errno, errbuf));
}
*newstate = zc.zc_cookie;
return (0);
}
/*
* Take the specified vdev offline
*/
int
zpool_vdev_offline(zpool_handle_t *zhp, const char *path, boolean_t istmp)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache;
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot offline %s"), path);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
NULL)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
if (avail_spare)
return (zfs_error(hdl, EZFS_ISSPARE, errbuf));
zc.zc_cookie = VDEV_STATE_OFFLINE;
zc.zc_obj = istmp ? ZFS_OFFLINE_TEMPORARY : 0;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0)
return (0);
switch (errno) {
case EBUSY:
/*
* There are no other replicas of this device.
*/
return (zfs_error(hdl, EZFS_NOREPLICAS, errbuf));
case EEXIST:
/*
* The log device has unplayed logs
*/
return (zfs_error(hdl, EZFS_UNPLAYED_LOGS, errbuf));
default:
return (zpool_standard_error(hdl, errno, errbuf));
}
}
/*
* Remove the specified vdev asynchronously from the configuration, so
* that it may come ONLINE if reinserted. This is called from zed on
* Udev remove event.
* Note: We also have a similar function zpool_vdev_remove() that
* removes the vdev from the pool.
*/
int
zpool_vdev_remove_wanted(zpool_handle_t *zhp, const char *path)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache;
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot remove %s"), path);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
NULL)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
zc.zc_cookie = VDEV_STATE_REMOVED;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0)
return (0);
return (zpool_standard_error(hdl, errno, errbuf));
}
/*
* Mark the given vdev faulted.
*/
int
zpool_vdev_fault(zpool_handle_t *zhp, uint64_t guid, vdev_aux_t aux)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot fault %llu"), (u_longlong_t)guid);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_guid = guid;
zc.zc_cookie = VDEV_STATE_FAULTED;
zc.zc_obj = aux;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0)
return (0);
switch (errno) {
case EBUSY:
/*
* There are no other replicas of this device.
*/
return (zfs_error(hdl, EZFS_NOREPLICAS, errbuf));
default:
return (zpool_standard_error(hdl, errno, errbuf));
}
}
/*
* Mark the given vdev degraded.
*/
int
zpool_vdev_degrade(zpool_handle_t *zhp, uint64_t guid, vdev_aux_t aux)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot degrade %llu"), (u_longlong_t)guid);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_guid = guid;
zc.zc_cookie = VDEV_STATE_DEGRADED;
zc.zc_obj = aux;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0)
return (0);
return (zpool_standard_error(hdl, errno, errbuf));
}
/*
* Returns TRUE if the given nvlist is a vdev that was originally swapped in as
* a hot spare.
*/
static boolean_t
is_replacing_spare(nvlist_t *search, nvlist_t *tgt, int which)
{
nvlist_t **child;
uint_t c, children;
if (nvlist_lookup_nvlist_array(search, ZPOOL_CONFIG_CHILDREN, &child,
&children) == 0) {
const char *type = fnvlist_lookup_string(search,
ZPOOL_CONFIG_TYPE);
if ((strcmp(type, VDEV_TYPE_SPARE) == 0 ||
strcmp(type, VDEV_TYPE_DRAID_SPARE) == 0) &&
children == 2 && child[which] == tgt)
return (B_TRUE);
for (c = 0; c < children; c++)
if (is_replacing_spare(child[c], tgt, which))
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Attach new_disk (fully described by nvroot) to old_disk.
* If 'replacing' is specified, the new disk will replace the old one.
*/
int
zpool_vdev_attach(zpool_handle_t *zhp, const char *old_disk,
const char *new_disk, nvlist_t *nvroot, int replacing, boolean_t rebuild)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
int ret;
nvlist_t *tgt;
boolean_t avail_spare, l2cache, islog;
uint64_t val;
char *newname;
+ const char *type;
nvlist_t **child;
uint_t children;
nvlist_t *config_root;
libzfs_handle_t *hdl = zhp->zpool_hdl;
if (replacing)
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot replace %s with %s"), old_disk, new_disk);
else
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot attach %s to %s"), new_disk, old_disk);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, old_disk, &avail_spare, &l2cache,
&islog)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
if (avail_spare)
return (zfs_error(hdl, EZFS_ISSPARE, errbuf));
if (l2cache)
return (zfs_error(hdl, EZFS_ISL2CACHE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
zc.zc_cookie = replacing;
zc.zc_simple = rebuild;
if (rebuild &&
zfeature_lookup_guid("org.openzfs:device_rebuild", NULL) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"the loaded zfs module doesn't support device rebuilds"));
return (zfs_error(hdl, EZFS_POOL_NOTSUP, errbuf));
}
+ type = fnvlist_lookup_string(tgt, ZPOOL_CONFIG_TYPE);
+ if (strcmp(type, VDEV_TYPE_RAIDZ) == 0 &&
+ zfeature_lookup_guid("org.openzfs:raidz_expansion", NULL) != 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "the loaded zfs module doesn't support raidz expansion"));
+ return (zfs_error(hdl, EZFS_POOL_NOTSUP, errbuf));
+ }
+
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0 || children != 1) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"new device must be a single disk"));
return (zfs_error(hdl, EZFS_INVALCONFIG, errbuf));
}
config_root = fnvlist_lookup_nvlist(zpool_get_config(zhp, NULL),
ZPOOL_CONFIG_VDEV_TREE);
if ((newname = zpool_vdev_name(NULL, NULL, child[0], 0)) == NULL)
return (-1);
/*
* If the target is a hot spare that has been swapped in, we can only
* replace it with another hot spare.
*/
if (replacing &&
nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_IS_SPARE, &val) == 0 &&
(zpool_find_vdev(zhp, newname, &avail_spare, &l2cache,
NULL) == NULL || !avail_spare) &&
is_replacing_spare(config_root, tgt, 1)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"can only be replaced by another hot spare"));
free(newname);
return (zfs_error(hdl, EZFS_BADTARGET, errbuf));
}
free(newname);
zcmd_write_conf_nvlist(hdl, &zc, nvroot);
ret = zfs_ioctl(hdl, ZFS_IOC_VDEV_ATTACH, &zc);
zcmd_free_nvlists(&zc);
if (ret == 0)
return (0);
switch (errno) {
case ENOTSUP:
/*
* Can't attach to or replace this type of vdev.
*/
if (replacing) {
uint64_t version = zpool_get_prop_int(zhp,
ZPOOL_PROP_VERSION, NULL);
if (islog) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot replace a log with a spare"));
} else if (rebuild) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"only mirror and dRAID vdevs support "
"sequential reconstruction"));
} else if (zpool_is_draid_spare(new_disk)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dRAID spares can only replace child "
"devices in their parent's dRAID vdev"));
} else if (version >= SPA_VERSION_MULTI_REPLACE) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"already in replacing/spare config; wait "
"for completion or use 'zpool detach'"));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot replace a replacing device"));
}
+ } else if (strcmp(type, VDEV_TYPE_RAIDZ) == 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "raidz_expansion feature must be enabled "
+ "in order to attach a device to raidz"));
} else {
char status[64] = {0};
zpool_prop_get_feature(zhp,
"feature@device_rebuild", status, 63);
if (rebuild &&
strncmp(status, ZFS_FEATURE_DISABLED, 64) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"device_rebuild feature must be enabled "
"in order to use sequential "
"reconstruction"));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"can only attach to mirrors and top-level "
"disks"));
}
}
(void) zfs_error(hdl, EZFS_BADTARGET, errbuf);
break;
case EINVAL:
/*
* The new device must be a single disk.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"new device must be a single disk"));
(void) zfs_error(hdl, EZFS_INVALCONFIG, errbuf);
break;
case EBUSY:
- zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "%s is busy, "
- "or device removal is in progress"),
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "%s is busy"),
new_disk);
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case EOVERFLOW:
/*
* The new device is too small.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"device is too small"));
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case EDOM:
/*
* The new device has a different optimal sector size.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"new device has a different optimal sector size; use the "
"option '-o ashift=N' to override the optimal size"));
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case ENAMETOOLONG:
/*
* The resulting top-level vdev spec won't fit in the label.
*/
(void) zfs_error(hdl, EZFS_DEVOVERFLOW, errbuf);
break;
+ case ENXIO:
+ /*
+ * The existing raidz vdev has offline children
+ */
+ if (strcmp(type, VDEV_TYPE_RAIDZ) == 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "raidz vdev has devices that are are offline or "
+ "being replaced"));
+ (void) zfs_error(hdl, EZFS_BADDEV, errbuf);
+ break;
+ } else {
+ (void) zpool_standard_error(hdl, errno, errbuf);
+ }
+ break;
+
+ case EADDRINUSE:
+ /*
+ * The boot reserved area is already being used (FreeBSD)
+ */
+ if (strcmp(type, VDEV_TYPE_RAIDZ) == 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "the reserved boot area needed for the expansion "
+ "is already being used by a boot loader"));
+ (void) zfs_error(hdl, EZFS_BADDEV, errbuf);
+ } else {
+ (void) zpool_standard_error(hdl, errno, errbuf);
+ }
+ break;
default:
(void) zpool_standard_error(hdl, errno, errbuf);
}
return (-1);
}
/*
* Detach the specified device.
*/
int
zpool_vdev_detach(zpool_handle_t *zhp, const char *path)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache;
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot detach %s"), path);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
NULL)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
if (avail_spare)
return (zfs_error(hdl, EZFS_ISSPARE, errbuf));
if (l2cache)
return (zfs_error(hdl, EZFS_ISL2CACHE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_DETACH, &zc) == 0)
return (0);
switch (errno) {
case ENOTSUP:
/*
* Can't detach from this type of vdev.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "only "
"applicable to mirror and replacing vdevs"));
(void) zfs_error(hdl, EZFS_BADTARGET, errbuf);
break;
case EBUSY:
/*
* There are no other replicas of this device.
*/
(void) zfs_error(hdl, EZFS_NOREPLICAS, errbuf);
break;
default:
(void) zpool_standard_error(hdl, errno, errbuf);
}
return (-1);
}
/*
* Find a mirror vdev in the source nvlist.
*
* The mchild array contains a list of disks in one of the top-level mirrors
* of the source pool. The schild array contains a list of disks that the
* user specified on the command line. We loop over the mchild array to
* see if any entry in the schild array matches.
*
* If a disk in the mchild array is found in the schild array, we return
* the index of that entry. Otherwise we return -1.
*/
static int
find_vdev_entry(zpool_handle_t *zhp, nvlist_t **mchild, uint_t mchildren,
nvlist_t **schild, uint_t schildren)
{
uint_t mc;
for (mc = 0; mc < mchildren; mc++) {
uint_t sc;
char *mpath = zpool_vdev_name(zhp->zpool_hdl, zhp,
mchild[mc], 0);
for (sc = 0; sc < schildren; sc++) {
char *spath = zpool_vdev_name(zhp->zpool_hdl, zhp,
schild[sc], 0);
boolean_t result = (strcmp(mpath, spath) == 0);
free(spath);
if (result) {
free(mpath);
return (mc);
}
}
free(mpath);
}
return (-1);
}
/*
* Split a mirror pool. If newroot points to null, then a new nvlist
* is generated and it is the responsibility of the caller to free it.
*/
int
zpool_vdev_split(zpool_handle_t *zhp, char *newname, nvlist_t **newroot,
nvlist_t *props, splitflags_t flags)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
const char *bias;
nvlist_t *tree, *config, **child, **newchild, *newconfig = NULL;
nvlist_t **varray = NULL, *zc_props = NULL;
uint_t c, children, newchildren, lastlog = 0, vcount, found = 0;
libzfs_handle_t *hdl = zhp->zpool_hdl;
uint64_t vers, readonly = B_FALSE;
boolean_t freelist = B_FALSE, memory_err = B_TRUE;
int retval = 0;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "Unable to split %s"), zhp->zpool_name);
if (!zpool_name_valid(hdl, B_FALSE, newname))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
if ((config = zpool_get_config(zhp, NULL)) == NULL) {
(void) fprintf(stderr, gettext("Internal error: unable to "
"retrieve pool configuration\n"));
return (-1);
}
tree = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
vers = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION);
if (props) {
prop_flags_t flags = { .create = B_FALSE, .import = B_TRUE };
if ((zc_props = zpool_valid_proplist(hdl, zhp->zpool_name,
props, vers, flags, errbuf)) == NULL)
return (-1);
(void) nvlist_lookup_uint64(zc_props,
zpool_prop_to_name(ZPOOL_PROP_READONLY), &readonly);
if (readonly) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property %s can only be set at import time"),
zpool_prop_to_name(ZPOOL_PROP_READONLY));
return (-1);
}
}
if (nvlist_lookup_nvlist_array(tree, ZPOOL_CONFIG_CHILDREN, &child,
&children) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Source pool is missing vdev tree"));
nvlist_free(zc_props);
return (-1);
}
varray = zfs_alloc(hdl, children * sizeof (nvlist_t *));
vcount = 0;
if (*newroot == NULL ||
nvlist_lookup_nvlist_array(*newroot, ZPOOL_CONFIG_CHILDREN,
&newchild, &newchildren) != 0)
newchildren = 0;
for (c = 0; c < children; c++) {
uint64_t is_log = B_FALSE, is_hole = B_FALSE;
boolean_t is_special = B_FALSE, is_dedup = B_FALSE;
const char *type;
nvlist_t **mchild, *vdev;
uint_t mchildren;
int entry;
/*
* Unlike cache & spares, slogs are stored in the
* ZPOOL_CONFIG_CHILDREN array. We filter them out here.
*/
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&is_log);
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE,
&is_hole);
if (is_log || is_hole) {
/*
* Create a hole vdev and put it in the config.
*/
if (nvlist_alloc(&vdev, NV_UNIQUE_NAME, 0) != 0)
goto out;
if (nvlist_add_string(vdev, ZPOOL_CONFIG_TYPE,
VDEV_TYPE_HOLE) != 0)
goto out;
if (nvlist_add_uint64(vdev, ZPOOL_CONFIG_IS_HOLE,
1) != 0)
goto out;
if (lastlog == 0)
lastlog = vcount;
varray[vcount++] = vdev;
continue;
}
lastlog = 0;
type = fnvlist_lookup_string(child[c], ZPOOL_CONFIG_TYPE);
if (strcmp(type, VDEV_TYPE_INDIRECT) == 0) {
vdev = child[c];
if (nvlist_dup(vdev, &varray[vcount++], 0) != 0)
goto out;
continue;
} else if (strcmp(type, VDEV_TYPE_MIRROR) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Source pool must be composed only of mirrors\n"));
retval = zfs_error(hdl, EZFS_INVALCONFIG, errbuf);
goto out;
}
if (nvlist_lookup_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &bias) == 0) {
if (strcmp(bias, VDEV_ALLOC_BIAS_SPECIAL) == 0)
is_special = B_TRUE;
else if (strcmp(bias, VDEV_ALLOC_BIAS_DEDUP) == 0)
is_dedup = B_TRUE;
}
verify(nvlist_lookup_nvlist_array(child[c],
ZPOOL_CONFIG_CHILDREN, &mchild, &mchildren) == 0);
/* find or add an entry for this top-level vdev */
if (newchildren > 0 &&
(entry = find_vdev_entry(zhp, mchild, mchildren,
newchild, newchildren)) >= 0) {
/* We found a disk that the user specified. */
vdev = mchild[entry];
++found;
} else {
/* User didn't specify a disk for this vdev. */
vdev = mchild[mchildren - 1];
}
if (nvlist_dup(vdev, &varray[vcount++], 0) != 0)
goto out;
if (flags.dryrun != 0) {
if (is_dedup == B_TRUE) {
if (nvlist_add_string(varray[vcount - 1],
ZPOOL_CONFIG_ALLOCATION_BIAS,
VDEV_ALLOC_BIAS_DEDUP) != 0)
goto out;
} else if (is_special == B_TRUE) {
if (nvlist_add_string(varray[vcount - 1],
ZPOOL_CONFIG_ALLOCATION_BIAS,
VDEV_ALLOC_BIAS_SPECIAL) != 0)
goto out;
}
}
}
/* did we find every disk the user specified? */
if (found != newchildren) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Device list must "
"include at most one disk from each mirror"));
retval = zfs_error(hdl, EZFS_INVALCONFIG, errbuf);
goto out;
}
/* Prepare the nvlist for populating. */
if (*newroot == NULL) {
if (nvlist_alloc(newroot, NV_UNIQUE_NAME, 0) != 0)
goto out;
freelist = B_TRUE;
if (nvlist_add_string(*newroot, ZPOOL_CONFIG_TYPE,
VDEV_TYPE_ROOT) != 0)
goto out;
} else {
verify(nvlist_remove_all(*newroot, ZPOOL_CONFIG_CHILDREN) == 0);
}
/* Add all the children we found */
if (nvlist_add_nvlist_array(*newroot, ZPOOL_CONFIG_CHILDREN,
(const nvlist_t **)varray, lastlog == 0 ? vcount : lastlog) != 0)
goto out;
/*
* If we're just doing a dry run, exit now with success.
*/
if (flags.dryrun) {
memory_err = B_FALSE;
freelist = B_FALSE;
goto out;
}
/* now build up the config list & call the ioctl */
if (nvlist_alloc(&newconfig, NV_UNIQUE_NAME, 0) != 0)
goto out;
if (nvlist_add_nvlist(newconfig,
ZPOOL_CONFIG_VDEV_TREE, *newroot) != 0 ||
nvlist_add_string(newconfig,
ZPOOL_CONFIG_POOL_NAME, newname) != 0 ||
nvlist_add_uint64(newconfig, ZPOOL_CONFIG_VERSION, vers) != 0)
goto out;
/*
* The new pool is automatically part of the namespace unless we
* explicitly export it.
*/
if (!flags.import)
zc.zc_cookie = ZPOOL_EXPORT_AFTER_SPLIT;
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
(void) strlcpy(zc.zc_string, newname, sizeof (zc.zc_string));
zcmd_write_conf_nvlist(hdl, &zc, newconfig);
if (zc_props != NULL)
zcmd_write_src_nvlist(hdl, &zc, zc_props);
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SPLIT, &zc) != 0) {
retval = zpool_standard_error(hdl, errno, errbuf);
goto out;
}
freelist = B_FALSE;
memory_err = B_FALSE;
out:
if (varray != NULL) {
int v;
for (v = 0; v < vcount; v++)
nvlist_free(varray[v]);
free(varray);
}
zcmd_free_nvlists(&zc);
nvlist_free(zc_props);
nvlist_free(newconfig);
if (freelist) {
nvlist_free(*newroot);
*newroot = NULL;
}
if (retval != 0)
return (retval);
if (memory_err)
return (no_memory(hdl));
return (0);
}
/*
* Remove the given device.
*/
int
zpool_vdev_remove(zpool_handle_t *zhp, const char *path)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache, islog;
libzfs_handle_t *hdl = zhp->zpool_hdl;
uint64_t version;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot remove %s"), path);
if (zpool_is_draid_spare(path)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dRAID spares cannot be removed"));
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
}
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
&islog)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL);
if (islog && version < SPA_VERSION_HOLES) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded to support log removal"));
return (zfs_error(hdl, EZFS_BADVERSION, errbuf));
}
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_REMOVE, &zc) == 0)
return (0);
switch (errno) {
case EALREADY:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"removal for this vdev is already in progress."));
(void) zfs_error(hdl, EZFS_BUSY, errbuf);
break;
case EINVAL:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid config; all top-level vdevs must "
"have the same sector size and not be raidz."));
(void) zfs_error(hdl, EZFS_INVALCONFIG, errbuf);
break;
case EBUSY:
if (islog) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Mount encrypted datasets to replay logs."));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Pool busy; removal may already be in progress"));
}
(void) zfs_error(hdl, EZFS_BUSY, errbuf);
break;
case EACCES:
if (islog) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Mount encrypted datasets to replay logs."));
(void) zfs_error(hdl, EZFS_BUSY, errbuf);
} else {
(void) zpool_standard_error(hdl, errno, errbuf);
}
break;
default:
(void) zpool_standard_error(hdl, errno, errbuf);
}
return (-1);
}
int
zpool_vdev_remove_cancel(zpool_handle_t *zhp)
{
zfs_cmd_t zc = {{0}};
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot cancel removal"));
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_cookie = 1;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_REMOVE, &zc) == 0)
return (0);
return (zpool_standard_error(hdl, errno, errbuf));
}
int
zpool_vdev_indirect_size(zpool_handle_t *zhp, const char *path,
uint64_t *sizep)
{
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache, islog;
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot determine indirect size of %s"),
path);
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
&islog)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
if (avail_spare || l2cache || islog) {
*sizep = 0;
return (0);
}
if (nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_INDIRECT_SIZE, sizep) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"indirect size not available"));
return (zfs_error(hdl, EINVAL, errbuf));
}
return (0);
}
/*
* Clear the errors for the pool, or the particular device if specified.
*/
int
zpool_clear(zpool_handle_t *zhp, const char *path, nvlist_t *rewindnvl)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
zpool_load_policy_t policy;
boolean_t avail_spare, l2cache;
libzfs_handle_t *hdl = zhp->zpool_hdl;
nvlist_t *nvi = NULL;
int error;
if (path)
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot clear errors for %s"),
path);
else
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot clear errors for %s"),
zhp->zpool_name);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if (path) {
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare,
&l2cache, NULL)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
/*
* Don't allow error clearing for hot spares. Do allow
* error clearing for l2cache devices.
*/
if (avail_spare)
return (zfs_error(hdl, EZFS_ISSPARE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
}
zpool_get_load_policy(rewindnvl, &policy);
zc.zc_cookie = policy.zlp_rewind;
zcmd_alloc_dst_nvlist(hdl, &zc, zhp->zpool_config_size * 2);
zcmd_write_src_nvlist(hdl, &zc, rewindnvl);
while ((error = zfs_ioctl(hdl, ZFS_IOC_CLEAR, &zc)) != 0 &&
errno == ENOMEM)
zcmd_expand_dst_nvlist(hdl, &zc);
if (!error || ((policy.zlp_rewind & ZPOOL_TRY_REWIND) &&
errno != EPERM && errno != EACCES)) {
if (policy.zlp_rewind &
(ZPOOL_DO_REWIND | ZPOOL_TRY_REWIND)) {
(void) zcmd_read_dst_nvlist(hdl, &zc, &nvi);
zpool_rewind_exclaim(hdl, zc.zc_name,
((policy.zlp_rewind & ZPOOL_TRY_REWIND) != 0),
nvi);
nvlist_free(nvi);
}
zcmd_free_nvlists(&zc);
return (0);
}
zcmd_free_nvlists(&zc);
return (zpool_standard_error(hdl, errno, errbuf));
}
/*
* Similar to zpool_clear(), but takes a GUID (used by fmd).
*/
int
zpool_vdev_clear(zpool_handle_t *zhp, uint64_t guid)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot clear errors for %llx"),
(u_longlong_t)guid);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_guid = guid;
zc.zc_cookie = ZPOOL_NO_REWIND;
if (zfs_ioctl(hdl, ZFS_IOC_CLEAR, &zc) == 0)
return (0);
return (zpool_standard_error(hdl, errno, errbuf));
}
/*
* Change the GUID for a pool.
*/
int
zpool_reguid(zpool_handle_t *zhp)
{
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zpool_hdl;
zfs_cmd_t zc = {"\0"};
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot reguid '%s'"), zhp->zpool_name);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if (zfs_ioctl(hdl, ZFS_IOC_POOL_REGUID, &zc) == 0)
return (0);
return (zpool_standard_error(hdl, errno, errbuf));
}
/*
* Reopen the pool.
*/
int
zpool_reopen_one(zpool_handle_t *zhp, void *data)
{
libzfs_handle_t *hdl = zpool_get_handle(zhp);
const char *pool_name = zpool_get_name(zhp);
boolean_t *scrub_restart = data;
int error;
error = lzc_reopen(pool_name, *scrub_restart);
if (error) {
return (zpool_standard_error_fmt(hdl, error,
dgettext(TEXT_DOMAIN, "cannot reopen '%s'"), pool_name));
}
return (0);
}
/* call into libzfs_core to execute the sync IOCTL per pool */
int
zpool_sync_one(zpool_handle_t *zhp, void *data)
{
int ret;
libzfs_handle_t *hdl = zpool_get_handle(zhp);
const char *pool_name = zpool_get_name(zhp);
boolean_t *force = data;
nvlist_t *innvl = fnvlist_alloc();
fnvlist_add_boolean_value(innvl, "force", *force);
if ((ret = lzc_sync(pool_name, innvl, NULL)) != 0) {
nvlist_free(innvl);
return (zpool_standard_error_fmt(hdl, ret,
dgettext(TEXT_DOMAIN, "sync '%s' failed"), pool_name));
}
nvlist_free(innvl);
return (0);
}
#define PATH_BUF_LEN 64
/*
* Given a vdev, return the name to display in iostat. If the vdev has a path,
* we use that, stripping off any leading "/dev/dsk/"; if not, we use the type.
* We also check if this is a whole disk, in which case we strip off the
* trailing 's0' slice name.
*
* This routine is also responsible for identifying when disks have been
* reconfigured in a new location. The kernel will have opened the device by
* devid, but the path will still refer to the old location. To catch this, we
* first do a path -> devid translation (which is fast for the common case). If
* the devid matches, we're done. If not, we do a reverse devid -> path
* translation and issue the appropriate ioctl() to update the path of the vdev.
* If 'zhp' is NULL, then this is an exported pool, and we don't need to do any
* of these checks.
*/
char *
zpool_vdev_name(libzfs_handle_t *hdl, zpool_handle_t *zhp, nvlist_t *nv,
int name_flags)
{
const char *type, *tpath;
const char *path;
uint64_t value;
char buf[PATH_BUF_LEN];
char tmpbuf[PATH_BUF_LEN * 2];
/*
* vdev_name will be "root"/"root-0" for the root vdev, but it is the
* zpool name that will be displayed to the user.
*/
type = fnvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE);
if (zhp != NULL && strcmp(type, "root") == 0)
return (zfs_strdup(hdl, zpool_get_name(zhp)));
if (libzfs_envvar_is_set("ZPOOL_VDEV_NAME_PATH"))
name_flags |= VDEV_NAME_PATH;
if (libzfs_envvar_is_set("ZPOOL_VDEV_NAME_GUID"))
name_flags |= VDEV_NAME_GUID;
if (libzfs_envvar_is_set("ZPOOL_VDEV_NAME_FOLLOW_LINKS"))
name_flags |= VDEV_NAME_FOLLOW_LINKS;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, &value) == 0 ||
name_flags & VDEV_NAME_GUID) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &value);
(void) snprintf(buf, sizeof (buf), "%llu", (u_longlong_t)value);
path = buf;
} else if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &tpath) == 0) {
path = tpath;
if (name_flags & VDEV_NAME_FOLLOW_LINKS) {
char *rp = realpath(path, NULL);
if (rp) {
strlcpy(buf, rp, sizeof (buf));
path = buf;
free(rp);
}
}
/*
* For a block device only use the name.
*/
if ((strcmp(type, VDEV_TYPE_DISK) == 0) &&
!(name_flags & VDEV_NAME_PATH)) {
path = zfs_strip_path(path);
}
/*
* Remove the partition from the path if this is a whole disk.
*/
if (strcmp(type, VDEV_TYPE_DRAID_SPARE) != 0 &&
nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, &value)
== 0 && value && !(name_flags & VDEV_NAME_PATH)) {
return (zfs_strip_partition(path));
}
} else {
path = type;
/*
* If it's a raidz device, we need to stick in the parity level.
*/
if (strcmp(path, VDEV_TYPE_RAIDZ) == 0) {
value = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY);
(void) snprintf(buf, sizeof (buf), "%s%llu", path,
(u_longlong_t)value);
path = buf;
}
/*
* If it's a dRAID device, we add parity, groups, and spares.
*/
if (strcmp(path, VDEV_TYPE_DRAID) == 0) {
uint64_t ndata, nparity, nspares;
nvlist_t **child;
uint_t children;
verify(nvlist_lookup_nvlist_array(nv,
ZPOOL_CONFIG_CHILDREN, &child, &children) == 0);
nparity = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_NPARITY);
ndata = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_DRAID_NDATA);
nspares = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_DRAID_NSPARES);
path = zpool_draid_name(buf, sizeof (buf), ndata,
nparity, nspares, children);
}
/*
* We identify each top-level vdev by using a <type-id>
* naming convention.
*/
if (name_flags & VDEV_NAME_TYPE_ID) {
uint64_t id = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_ID);
(void) snprintf(tmpbuf, sizeof (tmpbuf), "%s-%llu",
path, (u_longlong_t)id);
path = tmpbuf;
}
}
return (zfs_strdup(hdl, path));
}
static int
zbookmark_mem_compare(const void *a, const void *b)
{
return (memcmp(a, b, sizeof (zbookmark_phys_t)));
}
/*
* Retrieve the persistent error log, uniquify the members, and return to the
* caller.
*/
int
zpool_get_errlog(zpool_handle_t *zhp, nvlist_t **nverrlistp)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zpool_hdl;
zbookmark_phys_t *buf;
uint64_t buflen = 10000; /* approx. 1MB of RAM */
if (fnvlist_lookup_uint64(zhp->zpool_config,
ZPOOL_CONFIG_ERRCOUNT) == 0)
return (0);
/*
* Retrieve the raw error list from the kernel. If it doesn't fit,
* allocate a larger buffer and retry.
*/
(void) strcpy(zc.zc_name, zhp->zpool_name);
for (;;) {
buf = zfs_alloc(zhp->zpool_hdl,
buflen * sizeof (zbookmark_phys_t));
zc.zc_nvlist_dst = (uintptr_t)buf;
zc.zc_nvlist_dst_size = buflen;
if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_ERROR_LOG,
&zc) != 0) {
free(buf);
if (errno == ENOMEM) {
buflen *= 2;
} else {
return (zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN, "errors: List of "
"errors unavailable")));
}
} else {
break;
}
}
/*
* Sort the resulting bookmarks. This is a little confusing due to the
* implementation of ZFS_IOC_ERROR_LOG. The bookmarks are copied last
* to first, and 'zc_nvlist_dst_size' indicates the number of bookmarks
* _not_ copied as part of the process. So we point the start of our
* array appropriate and decrement the total number of elements.
*/
zbookmark_phys_t *zb = buf + zc.zc_nvlist_dst_size;
uint64_t zblen = buflen - zc.zc_nvlist_dst_size;
qsort(zb, zblen, sizeof (zbookmark_phys_t), zbookmark_mem_compare);
verify(nvlist_alloc(nverrlistp, 0, KM_SLEEP) == 0);
/*
* Fill in the nverrlistp with nvlist's of dataset and object numbers.
*/
for (uint64_t i = 0; i < zblen; i++) {
nvlist_t *nv;
/* ignoring zb_blkid and zb_level for now */
if (i > 0 && zb[i-1].zb_objset == zb[i].zb_objset &&
zb[i-1].zb_object == zb[i].zb_object)
continue;
if (nvlist_alloc(&nv, NV_UNIQUE_NAME, KM_SLEEP) != 0)
goto nomem;
if (nvlist_add_uint64(nv, ZPOOL_ERR_DATASET,
zb[i].zb_objset) != 0) {
nvlist_free(nv);
goto nomem;
}
if (nvlist_add_uint64(nv, ZPOOL_ERR_OBJECT,
zb[i].zb_object) != 0) {
nvlist_free(nv);
goto nomem;
}
if (nvlist_add_nvlist(*nverrlistp, "ejk", nv) != 0) {
nvlist_free(nv);
goto nomem;
}
nvlist_free(nv);
}
free(buf);
return (0);
nomem:
free(buf);
return (no_memory(zhp->zpool_hdl));
}
/*
* Upgrade a ZFS pool to the latest on-disk version.
*/
int
zpool_upgrade(zpool_handle_t *zhp, uint64_t new_version)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) strcpy(zc.zc_name, zhp->zpool_name);
zc.zc_cookie = new_version;
if (zfs_ioctl(hdl, ZFS_IOC_POOL_UPGRADE, &zc) != 0)
return (zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN, "cannot upgrade '%s'"),
zhp->zpool_name));
return (0);
}
void
zfs_save_arguments(int argc, char **argv, char *string, int len)
{
int i;
(void) strlcpy(string, zfs_basename(argv[0]), len);
for (i = 1; i < argc; i++) {
(void) strlcat(string, " ", len);
(void) strlcat(string, argv[i], len);
}
}
int
zpool_log_history(libzfs_handle_t *hdl, const char *message)
{
zfs_cmd_t zc = {"\0"};
nvlist_t *args;
args = fnvlist_alloc();
fnvlist_add_string(args, "message", message);
zcmd_write_src_nvlist(hdl, &zc, args);
int err = zfs_ioctl(hdl, ZFS_IOC_LOG_HISTORY, &zc);
nvlist_free(args);
zcmd_free_nvlists(&zc);
return (err);
}
/*
* Perform ioctl to get some command history of a pool.
*
* 'buf' is the buffer to fill up to 'len' bytes. 'off' is the
* logical offset of the history buffer to start reading from.
*
* Upon return, 'off' is the next logical offset to read from and
* 'len' is the actual amount of bytes read into 'buf'.
*/
static int
get_history(zpool_handle_t *zhp, char *buf, uint64_t *off, uint64_t *len)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_history = (uint64_t)(uintptr_t)buf;
zc.zc_history_len = *len;
zc.zc_history_offset = *off;
if (zfs_ioctl(hdl, ZFS_IOC_POOL_GET_HISTORY, &zc) != 0) {
switch (errno) {
case EPERM:
return (zfs_error_fmt(hdl, EZFS_PERM,
dgettext(TEXT_DOMAIN,
"cannot show history for pool '%s'"),
zhp->zpool_name));
case ENOENT:
return (zfs_error_fmt(hdl, EZFS_NOHISTORY,
dgettext(TEXT_DOMAIN, "cannot get history for pool "
"'%s'"), zhp->zpool_name));
case ENOTSUP:
return (zfs_error_fmt(hdl, EZFS_BADVERSION,
dgettext(TEXT_DOMAIN, "cannot get history for pool "
"'%s', pool must be upgraded"), zhp->zpool_name));
default:
return (zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN,
"cannot get history for '%s'"), zhp->zpool_name));
}
}
*len = zc.zc_history_len;
*off = zc.zc_history_offset;
return (0);
}
/*
* Retrieve the command history of a pool.
*/
int
zpool_get_history(zpool_handle_t *zhp, nvlist_t **nvhisp, uint64_t *off,
boolean_t *eof)
{
libzfs_handle_t *hdl = zhp->zpool_hdl;
char *buf;
int buflen = 128 * 1024;
nvlist_t **records = NULL;
uint_t numrecords = 0;
int err = 0, i;
uint64_t start = *off;
buf = zfs_alloc(hdl, buflen);
/* process about 1MiB a time */
while (*off - start < 1024 * 1024) {
uint64_t bytes_read = buflen;
uint64_t leftover;
if ((err = get_history(zhp, buf, off, &bytes_read)) != 0)
break;
/* if nothing else was read in, we're at EOF, just return */
if (!bytes_read) {
*eof = B_TRUE;
break;
}
if ((err = zpool_history_unpack(buf, bytes_read,
&leftover, &records, &numrecords)) != 0) {
zpool_standard_error_fmt(hdl, err,
dgettext(TEXT_DOMAIN,
"cannot get history for '%s'"), zhp->zpool_name);
break;
}
*off -= leftover;
if (leftover == bytes_read) {
/*
* no progress made, because buffer is not big enough
* to hold this record; resize and retry.
*/
buflen *= 2;
free(buf);
buf = zfs_alloc(hdl, buflen);
}
}
free(buf);
if (!err) {
*nvhisp = fnvlist_alloc();
fnvlist_add_nvlist_array(*nvhisp, ZPOOL_HIST_RECORD,
(const nvlist_t **)records, numrecords);
}
for (i = 0; i < numrecords; i++)
nvlist_free(records[i]);
free(records);
return (err);
}
/*
* Retrieve the next event given the passed 'zevent_fd' file descriptor.
* If there is a new event available 'nvp' will contain a newly allocated
* nvlist and 'dropped' will be set to the number of missed events since
* the last call to this function. When 'nvp' is set to NULL it indicates
* no new events are available. In either case the function returns 0 and
* it is up to the caller to free 'nvp'. In the case of a fatal error the
* function will return a non-zero value. When the function is called in
* blocking mode (the default, unless the ZEVENT_NONBLOCK flag is passed),
* it will not return until a new event is available.
*/
int
zpool_events_next(libzfs_handle_t *hdl, nvlist_t **nvp,
int *dropped, unsigned flags, int zevent_fd)
{
zfs_cmd_t zc = {"\0"};
int error = 0;
*nvp = NULL;
*dropped = 0;
zc.zc_cleanup_fd = zevent_fd;
if (flags & ZEVENT_NONBLOCK)
zc.zc_guid = ZEVENT_NONBLOCK;
zcmd_alloc_dst_nvlist(hdl, &zc, ZEVENT_SIZE);
retry:
if (zfs_ioctl(hdl, ZFS_IOC_EVENTS_NEXT, &zc) != 0) {
switch (errno) {
case ESHUTDOWN:
error = zfs_error_fmt(hdl, EZFS_POOLUNAVAIL,
dgettext(TEXT_DOMAIN, "zfs shutdown"));
goto out;
case ENOENT:
/* Blocking error case should not occur */
if (!(flags & ZEVENT_NONBLOCK))
error = zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN, "cannot get event"));
goto out;
case ENOMEM:
zcmd_expand_dst_nvlist(hdl, &zc);
goto retry;
default:
error = zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN, "cannot get event"));
goto out;
}
}
error = zcmd_read_dst_nvlist(hdl, &zc, nvp);
if (error != 0)
goto out;
*dropped = (int)zc.zc_cookie;
out:
zcmd_free_nvlists(&zc);
return (error);
}
/*
* Clear all events.
*/
int
zpool_events_clear(libzfs_handle_t *hdl, int *count)
{
zfs_cmd_t zc = {"\0"};
if (zfs_ioctl(hdl, ZFS_IOC_EVENTS_CLEAR, &zc) != 0)
return (zpool_standard_error(hdl, errno,
dgettext(TEXT_DOMAIN, "cannot clear events")));
if (count != NULL)
*count = (int)zc.zc_cookie; /* # of events cleared */
return (0);
}
/*
* Seek to a specific EID, ZEVENT_SEEK_START, or ZEVENT_SEEK_END for
* the passed zevent_fd file handle. On success zero is returned,
* otherwise -1 is returned and hdl->libzfs_error is set to the errno.
*/
int
zpool_events_seek(libzfs_handle_t *hdl, uint64_t eid, int zevent_fd)
{
zfs_cmd_t zc = {"\0"};
int error = 0;
zc.zc_guid = eid;
zc.zc_cleanup_fd = zevent_fd;
if (zfs_ioctl(hdl, ZFS_IOC_EVENTS_SEEK, &zc) != 0) {
switch (errno) {
case ENOENT:
error = zfs_error_fmt(hdl, EZFS_NOENT,
dgettext(TEXT_DOMAIN, "cannot get event"));
break;
case ENOMEM:
error = zfs_error_fmt(hdl, EZFS_NOMEM,
dgettext(TEXT_DOMAIN, "cannot get event"));
break;
default:
error = zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN, "cannot get event"));
break;
}
}
return (error);
}
static void
zpool_obj_to_path_impl(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj,
char *pathname, size_t len, boolean_t always_unmounted)
{
zfs_cmd_t zc = {"\0"};
boolean_t mounted = B_FALSE;
char *mntpnt = NULL;
char dsname[ZFS_MAX_DATASET_NAME_LEN];
if (dsobj == 0) {
/* special case for the MOS */
(void) snprintf(pathname, len, "<metadata>:<0x%llx>",
(longlong_t)obj);
return;
}
/* get the dataset's name */
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_obj = dsobj;
if (zfs_ioctl(zhp->zpool_hdl,
ZFS_IOC_DSOBJ_TO_DSNAME, &zc) != 0) {
/* just write out a path of two object numbers */
(void) snprintf(pathname, len, "<0x%llx>:<0x%llx>",
(longlong_t)dsobj, (longlong_t)obj);
return;
}
(void) strlcpy(dsname, zc.zc_value, sizeof (dsname));
/* find out if the dataset is mounted */
mounted = !always_unmounted && is_mounted(zhp->zpool_hdl, dsname,
&mntpnt);
/* get the corrupted object's path */
(void) strlcpy(zc.zc_name, dsname, sizeof (zc.zc_name));
zc.zc_obj = obj;
if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_OBJ_TO_PATH,
&zc) == 0) {
if (mounted) {
(void) snprintf(pathname, len, "%s%s", mntpnt,
zc.zc_value);
} else {
(void) snprintf(pathname, len, "%s:%s",
dsname, zc.zc_value);
}
} else {
(void) snprintf(pathname, len, "%s:<0x%llx>", dsname,
(longlong_t)obj);
}
free(mntpnt);
}
void
zpool_obj_to_path(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj,
char *pathname, size_t len)
{
zpool_obj_to_path_impl(zhp, dsobj, obj, pathname, len, B_FALSE);
}
void
zpool_obj_to_path_ds(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj,
char *pathname, size_t len)
{
zpool_obj_to_path_impl(zhp, dsobj, obj, pathname, len, B_TRUE);
}
/*
* Wait while the specified activity is in progress in the pool.
*/
int
zpool_wait(zpool_handle_t *zhp, zpool_wait_activity_t activity)
{
boolean_t missing;
int error = zpool_wait_status(zhp, activity, &missing, NULL);
if (missing) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl, ENOENT,
dgettext(TEXT_DOMAIN, "error waiting in pool '%s'"),
zhp->zpool_name);
return (ENOENT);
} else {
return (error);
}
}
/*
* Wait for the given activity and return the status of the wait (whether or not
* any waiting was done) in the 'waited' parameter. Non-existent pools are
* reported via the 'missing' parameter, rather than by printing an error
* message. This is convenient when this function is called in a loop over a
* long period of time (as it is, for example, by zpool's wait cmd). In that
* scenario, a pool being exported or destroyed should be considered a normal
* event, so we don't want to print an error when we find that the pool doesn't
* exist.
*/
int
zpool_wait_status(zpool_handle_t *zhp, zpool_wait_activity_t activity,
boolean_t *missing, boolean_t *waited)
{
int error = lzc_wait(zhp->zpool_name, activity, waited);
*missing = (error == ENOENT);
if (*missing)
return (0);
if (error != 0) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl, error,
dgettext(TEXT_DOMAIN, "error waiting in pool '%s'"),
zhp->zpool_name);
}
return (error);
}
int
zpool_set_bootenv(zpool_handle_t *zhp, const nvlist_t *envmap)
{
int error = lzc_set_bootenv(zhp->zpool_name, envmap);
if (error != 0) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl, error,
dgettext(TEXT_DOMAIN,
"error setting bootenv in pool '%s'"), zhp->zpool_name);
}
return (error);
}
int
zpool_get_bootenv(zpool_handle_t *zhp, nvlist_t **nvlp)
{
nvlist_t *nvl;
int error;
nvl = NULL;
error = lzc_get_bootenv(zhp->zpool_name, &nvl);
if (error != 0) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl, error,
dgettext(TEXT_DOMAIN,
"error getting bootenv in pool '%s'"), zhp->zpool_name);
} else {
*nvlp = nvl;
}
return (error);
}
/*
* Attempt to read and parse feature file(s) (from "compatibility" property).
* Files contain zpool feature names, comma or whitespace-separated.
* Comments (# character to next newline) are discarded.
*
* Arguments:
* compatibility : string containing feature filenames
* features : either NULL or pointer to array of boolean
* report : either NULL or pointer to string buffer
* rlen : length of "report" buffer
*
* compatibility is NULL (unset), "", "off", "legacy", or list of
* comma-separated filenames. filenames should either be absolute,
* or relative to:
* 1) ZPOOL_SYSCONF_COMPAT_D (eg: /etc/zfs/compatibility.d) or
* 2) ZPOOL_DATA_COMPAT_D (eg: /usr/share/zfs/compatibility.d).
* (Unset), "" or "off" => enable all features
* "legacy" => disable all features
*
* Any feature names read from files which match unames in spa_feature_table
* will have the corresponding boolean set in the features array (if non-NULL).
* If more than one feature set specified, only features present in *all* of
* them will be set.
*
* "report" if not NULL will be populated with a suitable status message.
*
* Return values:
* ZPOOL_COMPATIBILITY_OK : files read and parsed ok
* ZPOOL_COMPATIBILITY_BADFILE : file too big or not a text file
* ZPOOL_COMPATIBILITY_BADTOKEN : SYSCONF file contains invalid feature name
* ZPOOL_COMPATIBILITY_WARNTOKEN : DATA file contains invalid feature name
* ZPOOL_COMPATIBILITY_NOFILES : no feature files found
*/
zpool_compat_status_t
zpool_load_compat(const char *compat, boolean_t *features, char *report,
size_t rlen)
{
int sdirfd, ddirfd, featfd;
struct stat fs;
char *fc;
char *ps, *ls, *ws;
char *file, *line, *word;
char l_compat[ZFS_MAXPROPLEN];
boolean_t ret_nofiles = B_TRUE;
boolean_t ret_badfile = B_FALSE;
boolean_t ret_badtoken = B_FALSE;
boolean_t ret_warntoken = B_FALSE;
/* special cases (unset), "" and "off" => enable all features */
if (compat == NULL || compat[0] == '\0' ||
strcmp(compat, ZPOOL_COMPAT_OFF) == 0) {
if (features != NULL)
for (uint_t i = 0; i < SPA_FEATURES; i++)
features[i] = B_TRUE;
if (report != NULL)
strlcpy(report, gettext("all features enabled"), rlen);
return (ZPOOL_COMPATIBILITY_OK);
}
/* Final special case "legacy" => disable all features */
if (strcmp(compat, ZPOOL_COMPAT_LEGACY) == 0) {
if (features != NULL)
for (uint_t i = 0; i < SPA_FEATURES; i++)
features[i] = B_FALSE;
if (report != NULL)
strlcpy(report, gettext("all features disabled"), rlen);
return (ZPOOL_COMPATIBILITY_OK);
}
/*
* Start with all true; will be ANDed with results from each file
*/
if (features != NULL)
for (uint_t i = 0; i < SPA_FEATURES; i++)
features[i] = B_TRUE;
char err_badfile[ZFS_MAXPROPLEN] = "";
char err_badtoken[ZFS_MAXPROPLEN] = "";
/*
* We ignore errors from the directory open()
* as they're only needed if the filename is relative
* which will be checked during the openat().
*/
/* O_PATH safer than O_RDONLY if system allows it */
#if defined(O_PATH)
#define ZC_DIR_FLAGS (O_DIRECTORY | O_CLOEXEC | O_PATH)
#else
#define ZC_DIR_FLAGS (O_DIRECTORY | O_CLOEXEC | O_RDONLY)
#endif
sdirfd = open(ZPOOL_SYSCONF_COMPAT_D, ZC_DIR_FLAGS);
ddirfd = open(ZPOOL_DATA_COMPAT_D, ZC_DIR_FLAGS);
(void) strlcpy(l_compat, compat, ZFS_MAXPROPLEN);
for (file = strtok_r(l_compat, ",", &ps);
file != NULL;
file = strtok_r(NULL, ",", &ps)) {
boolean_t l_features[SPA_FEATURES];
enum { Z_SYSCONF, Z_DATA } source;
/* try sysconfdir first, then datadir */
source = Z_SYSCONF;
if ((featfd = openat(sdirfd, file, O_RDONLY | O_CLOEXEC)) < 0) {
featfd = openat(ddirfd, file, O_RDONLY | O_CLOEXEC);
source = Z_DATA;
}
/* File readable and correct size? */
if (featfd < 0 ||
fstat(featfd, &fs) < 0 ||
fs.st_size < 1 ||
fs.st_size > ZPOOL_COMPAT_MAXSIZE) {
(void) close(featfd);
strlcat(err_badfile, file, ZFS_MAXPROPLEN);
strlcat(err_badfile, " ", ZFS_MAXPROPLEN);
ret_badfile = B_TRUE;
continue;
}
/* Prefault the file if system allows */
#if defined(MAP_POPULATE)
#define ZC_MMAP_FLAGS (MAP_PRIVATE | MAP_POPULATE)
#elif defined(MAP_PREFAULT_READ)
#define ZC_MMAP_FLAGS (MAP_PRIVATE | MAP_PREFAULT_READ)
#else
#define ZC_MMAP_FLAGS (MAP_PRIVATE)
#endif
/* private mmap() so we can strtok safely */
fc = (char *)mmap(NULL, fs.st_size, PROT_READ | PROT_WRITE,
ZC_MMAP_FLAGS, featfd, 0);
(void) close(featfd);
/* map ok, and last character == newline? */
if (fc == MAP_FAILED || fc[fs.st_size - 1] != '\n') {
(void) munmap((void *) fc, fs.st_size);
strlcat(err_badfile, file, ZFS_MAXPROPLEN);
strlcat(err_badfile, " ", ZFS_MAXPROPLEN);
ret_badfile = B_TRUE;
continue;
}
ret_nofiles = B_FALSE;
for (uint_t i = 0; i < SPA_FEATURES; i++)
l_features[i] = B_FALSE;
/* replace final newline with NULL to ensure string ends */
fc[fs.st_size - 1] = '\0';
for (line = strtok_r(fc, "\n", &ls);
line != NULL;
line = strtok_r(NULL, "\n", &ls)) {
/* discard comments */
char *r = strchr(line, '#');
if (r != NULL)
*r = '\0';
for (word = strtok_r(line, ", \t", &ws);
word != NULL;
word = strtok_r(NULL, ", \t", &ws)) {
/* Find matching feature name */
uint_t f;
for (f = 0; f < SPA_FEATURES; f++) {
zfeature_info_t *fi =
&spa_feature_table[f];
if (strcmp(word, fi->fi_uname) == 0) {
l_features[f] = B_TRUE;
break;
}
}
if (f < SPA_FEATURES)
continue;
/* found an unrecognized word */
/* lightly sanitize it */
if (strlen(word) > 32)
word[32] = '\0';
for (char *c = word; *c != '\0'; c++)
if (!isprint(*c))
*c = '?';
strlcat(err_badtoken, word, ZFS_MAXPROPLEN);
strlcat(err_badtoken, " ", ZFS_MAXPROPLEN);
if (source == Z_SYSCONF)
ret_badtoken = B_TRUE;
else
ret_warntoken = B_TRUE;
}
}
(void) munmap((void *) fc, fs.st_size);
if (features != NULL)
for (uint_t i = 0; i < SPA_FEATURES; i++)
features[i] &= l_features[i];
}
(void) close(sdirfd);
(void) close(ddirfd);
/* Return the most serious error */
if (ret_badfile) {
if (report != NULL)
snprintf(report, rlen, gettext("could not read/"
"parse feature file(s): %s"), err_badfile);
return (ZPOOL_COMPATIBILITY_BADFILE);
}
if (ret_nofiles) {
if (report != NULL)
strlcpy(report,
gettext("no valid compatibility files specified"),
rlen);
return (ZPOOL_COMPATIBILITY_NOFILES);
}
if (ret_badtoken) {
if (report != NULL)
snprintf(report, rlen, gettext("invalid feature "
"name(s) in local compatibility files: %s"),
err_badtoken);
return (ZPOOL_COMPATIBILITY_BADTOKEN);
}
if (ret_warntoken) {
if (report != NULL)
snprintf(report, rlen, gettext("unrecognized feature "
"name(s) in distribution compatibility files: %s"),
err_badtoken);
return (ZPOOL_COMPATIBILITY_WARNTOKEN);
}
if (report != NULL)
strlcpy(report, gettext("compatibility set ok"), rlen);
return (ZPOOL_COMPATIBILITY_OK);
}
static int
zpool_vdev_guid(zpool_handle_t *zhp, const char *vdevname, uint64_t *vdev_guid)
{
nvlist_t *tgt;
boolean_t avail_spare, l2cache;
verify(zhp != NULL);
if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "pool is in an unavailable state"));
return (zfs_error(zhp->zpool_hdl, EZFS_POOLUNAVAIL, errbuf));
}
if ((tgt = zpool_find_vdev(zhp, vdevname, &avail_spare, &l2cache,
NULL)) == NULL) {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "can not find %s in %s"),
vdevname, zhp->zpool_name);
return (zfs_error(zhp->zpool_hdl, EZFS_NODEVICE, errbuf));
}
*vdev_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
return (0);
}
/*
* Get a vdev property value for 'prop' and return the value in
* a pre-allocated buffer.
*/
int
zpool_get_vdev_prop_value(nvlist_t *nvprop, vdev_prop_t prop, char *prop_name,
char *buf, size_t len, zprop_source_t *srctype, boolean_t literal)
{
nvlist_t *nv;
const char *strval;
uint64_t intval;
zprop_source_t src = ZPROP_SRC_NONE;
if (prop == VDEV_PROP_USERPROP) {
/* user property, prop_name must contain the property name */
assert(prop_name != NULL);
if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) {
src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
strval = fnvlist_lookup_string(nv, ZPROP_VALUE);
} else {
/* user prop not found */
return (-1);
}
(void) strlcpy(buf, strval, len);
if (srctype)
*srctype = src;
return (0);
}
if (prop_name == NULL)
prop_name = (char *)vdev_prop_to_name(prop);
switch (vdev_prop_get_type(prop)) {
case PROP_TYPE_STRING:
if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) {
src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
strval = fnvlist_lookup_string(nv, ZPROP_VALUE);
} else {
src = ZPROP_SRC_DEFAULT;
if ((strval = vdev_prop_default_string(prop)) == NULL)
strval = "-";
}
(void) strlcpy(buf, strval, len);
break;
case PROP_TYPE_NUMBER:
if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) {
src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
intval = fnvlist_lookup_uint64(nv, ZPROP_VALUE);
} else {
src = ZPROP_SRC_DEFAULT;
intval = vdev_prop_default_numeric(prop);
}
switch (prop) {
case VDEV_PROP_ASIZE:
case VDEV_PROP_PSIZE:
case VDEV_PROP_SIZE:
case VDEV_PROP_BOOTSIZE:
case VDEV_PROP_ALLOCATED:
case VDEV_PROP_FREE:
case VDEV_PROP_READ_ERRORS:
case VDEV_PROP_WRITE_ERRORS:
case VDEV_PROP_CHECKSUM_ERRORS:
case VDEV_PROP_INITIALIZE_ERRORS:
case VDEV_PROP_OPS_NULL:
case VDEV_PROP_OPS_READ:
case VDEV_PROP_OPS_WRITE:
case VDEV_PROP_OPS_FREE:
case VDEV_PROP_OPS_CLAIM:
case VDEV_PROP_OPS_TRIM:
case VDEV_PROP_BYTES_NULL:
case VDEV_PROP_BYTES_READ:
case VDEV_PROP_BYTES_WRITE:
case VDEV_PROP_BYTES_FREE:
case VDEV_PROP_BYTES_CLAIM:
case VDEV_PROP_BYTES_TRIM:
if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) zfs_nicenum(intval, buf, len);
}
break;
case VDEV_PROP_EXPANDSZ:
if (intval == 0) {
(void) strlcpy(buf, "-", len);
} else if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) zfs_nicenum(intval, buf, len);
}
break;
case VDEV_PROP_CAPACITY:
if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) snprintf(buf, len, "%llu%%",
(u_longlong_t)intval);
}
break;
case VDEV_PROP_CHECKSUM_N:
case VDEV_PROP_CHECKSUM_T:
case VDEV_PROP_IO_N:
case VDEV_PROP_IO_T:
if (intval == UINT64_MAX) {
(void) strlcpy(buf, "-", len);
} else {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
}
break;
case VDEV_PROP_FRAGMENTATION:
if (intval == UINT64_MAX) {
(void) strlcpy(buf, "-", len);
} else {
(void) snprintf(buf, len, "%llu%%",
(u_longlong_t)intval);
}
break;
case VDEV_PROP_STATE:
if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) strlcpy(buf, zpool_state_to_name(intval,
VDEV_AUX_NONE), len);
}
break;
default:
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
}
break;
case PROP_TYPE_INDEX:
if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) {
src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
intval = fnvlist_lookup_uint64(nv, ZPROP_VALUE);
} else {
src = ZPROP_SRC_DEFAULT;
intval = vdev_prop_default_numeric(prop);
+ /* Only use if provided by the RAIDZ VDEV above */
+ if (prop == VDEV_PROP_RAIDZ_EXPANDING)
+ return (ENOENT);
}
if (vdev_prop_index_to_string(prop, intval,
(const char **)&strval) != 0)
return (-1);
(void) strlcpy(buf, strval, len);
break;
default:
abort();
}
if (srctype)
*srctype = src;
return (0);
}
/*
* Get a vdev property value for 'prop_name' and return the value in
* a pre-allocated buffer.
*/
int
zpool_get_vdev_prop(zpool_handle_t *zhp, const char *vdevname, vdev_prop_t prop,
char *prop_name, char *buf, size_t len, zprop_source_t *srctype,
boolean_t literal)
{
nvlist_t *reqnvl, *reqprops;
nvlist_t *retprops = NULL;
uint64_t vdev_guid = 0;
int ret;
if ((ret = zpool_vdev_guid(zhp, vdevname, &vdev_guid)) != 0)
return (ret);
if (nvlist_alloc(&reqnvl, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
if (nvlist_alloc(&reqprops, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
fnvlist_add_uint64(reqnvl, ZPOOL_VDEV_PROPS_GET_VDEV, vdev_guid);
if (prop != VDEV_PROP_USERPROP) {
/* prop_name overrides prop value */
if (prop_name != NULL)
prop = vdev_name_to_prop(prop_name);
else
prop_name = (char *)vdev_prop_to_name(prop);
assert(prop < VDEV_NUM_PROPS);
}
assert(prop_name != NULL);
if (nvlist_add_uint64(reqprops, prop_name, prop) != 0) {
nvlist_free(reqnvl);
nvlist_free(reqprops);
return (no_memory(zhp->zpool_hdl));
}
fnvlist_add_nvlist(reqnvl, ZPOOL_VDEV_PROPS_GET_PROPS, reqprops);
ret = lzc_get_vdev_prop(zhp->zpool_name, reqnvl, &retprops);
if (ret == 0) {
ret = zpool_get_vdev_prop_value(retprops, prop, prop_name, buf,
len, srctype, literal);
} else {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot get vdev property %s from"
" %s in %s"), prop_name, vdevname, zhp->zpool_name);
(void) zpool_standard_error(zhp->zpool_hdl, ret, errbuf);
}
nvlist_free(reqnvl);
nvlist_free(reqprops);
nvlist_free(retprops);
return (ret);
}
/*
* Get all vdev properties
*/
int
zpool_get_all_vdev_props(zpool_handle_t *zhp, const char *vdevname,
nvlist_t **outnvl)
{
nvlist_t *nvl = NULL;
uint64_t vdev_guid = 0;
int ret;
if ((ret = zpool_vdev_guid(zhp, vdevname, &vdev_guid)) != 0)
return (ret);
if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
fnvlist_add_uint64(nvl, ZPOOL_VDEV_PROPS_GET_VDEV, vdev_guid);
ret = lzc_get_vdev_prop(zhp->zpool_name, nvl, outnvl);
nvlist_free(nvl);
if (ret) {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot get vdev properties for"
" %s in %s"), vdevname, zhp->zpool_name);
(void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf);
}
return (ret);
}
/*
* Set vdev property
*/
int
zpool_set_vdev_prop(zpool_handle_t *zhp, const char *vdevname,
const char *propname, const char *propval)
{
int ret;
nvlist_t *nvl = NULL;
nvlist_t *outnvl = NULL;
nvlist_t *props;
nvlist_t *realprops;
prop_flags_t flags = { 0 };
uint64_t version;
uint64_t vdev_guid;
if ((ret = zpool_vdev_guid(zhp, vdevname, &vdev_guid)) != 0)
return (ret);
if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
fnvlist_add_uint64(nvl, ZPOOL_VDEV_PROPS_SET_VDEV, vdev_guid);
if (nvlist_add_string(props, propname, propval) != 0) {
nvlist_free(props);
return (no_memory(zhp->zpool_hdl));
}
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot set property %s for %s on %s"),
propname, vdevname, zhp->zpool_name);
flags.vdevprop = 1;
version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL);
if ((realprops = zpool_valid_proplist(zhp->zpool_hdl,
zhp->zpool_name, props, version, flags, errbuf)) == NULL) {
nvlist_free(props);
nvlist_free(nvl);
return (-1);
}
nvlist_free(props);
props = realprops;
fnvlist_add_nvlist(nvl, ZPOOL_VDEV_PROPS_SET_PROPS, props);
ret = lzc_set_vdev_prop(zhp->zpool_name, nvl, &outnvl);
nvlist_free(props);
nvlist_free(nvl);
nvlist_free(outnvl);
if (ret)
(void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf);
return (ret);
}
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_util.c b/sys/contrib/openzfs/lib/libzfs/libzfs_util.c
index fdd1975fa677..778e511e0b32 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_util.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_util.c
@@ -1,2266 +1,2271 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright 2020 Joyent, Inc. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>
* Copyright (c) 2017 Datto Inc.
* Copyright (c) 2020 The FreeBSD Foundation
*
* Portions of this software were developed by Allan Jude
* under sponsorship from the FreeBSD Foundation.
*/
/*
* Internal utility routines for the ZFS library.
*/
#include <errno.h>
#include <fcntl.h>
#include <libintl.h>
#include <stdarg.h>
#include <stdio.h>
#include <stdlib.h>
#include <strings.h>
#include <unistd.h>
#include <math.h>
#if LIBFETCH_DYNAMIC
#include <dlfcn.h>
#endif
#include <sys/stat.h>
#include <sys/mnttab.h>
#include <sys/mntent.h>
#include <sys/types.h>
#include <sys/wait.h>
#include <libzfs.h>
#include <libzfs_core.h>
#include "libzfs_impl.h"
#include "zfs_prop.h"
#include "zfeature_common.h"
#include <zfs_fletcher.h>
#include <libzutil.h>
/*
* We only care about the scheme in order to match the scheme
* with the handler. Each handler should validate the full URI
* as necessary.
*/
#define URI_REGEX "^\\([A-Za-z][A-Za-z0-9+.\\-]*\\):"
int
libzfs_errno(libzfs_handle_t *hdl)
{
return (hdl->libzfs_error);
}
const char *
libzfs_error_action(libzfs_handle_t *hdl)
{
return (hdl->libzfs_action);
}
const char *
libzfs_error_description(libzfs_handle_t *hdl)
{
if (hdl->libzfs_desc[0] != '\0')
return (hdl->libzfs_desc);
switch (hdl->libzfs_error) {
case EZFS_NOMEM:
return (dgettext(TEXT_DOMAIN, "out of memory"));
case EZFS_BADPROP:
return (dgettext(TEXT_DOMAIN, "invalid property value"));
case EZFS_PROPREADONLY:
return (dgettext(TEXT_DOMAIN, "read-only property"));
case EZFS_PROPTYPE:
return (dgettext(TEXT_DOMAIN, "property doesn't apply to "
"datasets of this type"));
case EZFS_PROPNONINHERIT:
return (dgettext(TEXT_DOMAIN, "property cannot be inherited"));
case EZFS_PROPSPACE:
return (dgettext(TEXT_DOMAIN, "invalid quota or reservation"));
case EZFS_BADTYPE:
return (dgettext(TEXT_DOMAIN, "operation not applicable to "
"datasets of this type"));
case EZFS_BUSY:
return (dgettext(TEXT_DOMAIN, "pool or dataset is busy"));
case EZFS_EXISTS:
return (dgettext(TEXT_DOMAIN, "pool or dataset exists"));
case EZFS_NOENT:
return (dgettext(TEXT_DOMAIN, "no such pool or dataset"));
case EZFS_BADSTREAM:
return (dgettext(TEXT_DOMAIN, "invalid backup stream"));
case EZFS_DSREADONLY:
return (dgettext(TEXT_DOMAIN, "dataset is read-only"));
case EZFS_VOLTOOBIG:
return (dgettext(TEXT_DOMAIN, "volume size exceeds limit for "
"this system"));
case EZFS_INVALIDNAME:
return (dgettext(TEXT_DOMAIN, "invalid name"));
case EZFS_BADRESTORE:
return (dgettext(TEXT_DOMAIN, "unable to restore to "
"destination"));
case EZFS_BADBACKUP:
return (dgettext(TEXT_DOMAIN, "backup failed"));
case EZFS_BADTARGET:
return (dgettext(TEXT_DOMAIN, "invalid target vdev"));
case EZFS_NODEVICE:
return (dgettext(TEXT_DOMAIN, "no such device in pool"));
case EZFS_BADDEV:
return (dgettext(TEXT_DOMAIN, "invalid device"));
case EZFS_NOREPLICAS:
return (dgettext(TEXT_DOMAIN, "no valid replicas"));
case EZFS_RESILVERING:
return (dgettext(TEXT_DOMAIN, "currently resilvering"));
case EZFS_BADVERSION:
return (dgettext(TEXT_DOMAIN, "unsupported version or "
"feature"));
case EZFS_POOLUNAVAIL:
return (dgettext(TEXT_DOMAIN, "pool is unavailable"));
case EZFS_DEVOVERFLOW:
return (dgettext(TEXT_DOMAIN, "too many devices in one vdev"));
case EZFS_BADPATH:
return (dgettext(TEXT_DOMAIN, "must be an absolute path"));
case EZFS_CROSSTARGET:
return (dgettext(TEXT_DOMAIN, "operation crosses datasets or "
"pools"));
case EZFS_ZONED:
return (dgettext(TEXT_DOMAIN, "dataset in use by local zone"));
case EZFS_MOUNTFAILED:
return (dgettext(TEXT_DOMAIN, "mount failed"));
case EZFS_UMOUNTFAILED:
return (dgettext(TEXT_DOMAIN, "unmount failed"));
case EZFS_UNSHARENFSFAILED:
return (dgettext(TEXT_DOMAIN, "NFS share removal failed"));
case EZFS_SHARENFSFAILED:
return (dgettext(TEXT_DOMAIN, "NFS share creation failed"));
case EZFS_UNSHARESMBFAILED:
return (dgettext(TEXT_DOMAIN, "SMB share removal failed"));
case EZFS_SHARESMBFAILED:
return (dgettext(TEXT_DOMAIN, "SMB share creation failed"));
case EZFS_PERM:
return (dgettext(TEXT_DOMAIN, "permission denied"));
case EZFS_NOSPC:
return (dgettext(TEXT_DOMAIN, "out of space"));
case EZFS_FAULT:
return (dgettext(TEXT_DOMAIN, "bad address"));
case EZFS_IO:
return (dgettext(TEXT_DOMAIN, "I/O error"));
case EZFS_INTR:
return (dgettext(TEXT_DOMAIN, "signal received"));
case EZFS_CKSUM:
return (dgettext(TEXT_DOMAIN, "insufficient replicas"));
case EZFS_ISSPARE:
return (dgettext(TEXT_DOMAIN, "device is reserved as a hot "
"spare"));
case EZFS_INVALCONFIG:
return (dgettext(TEXT_DOMAIN, "invalid vdev configuration"));
case EZFS_RECURSIVE:
return (dgettext(TEXT_DOMAIN, "recursive dataset dependency"));
case EZFS_NOHISTORY:
return (dgettext(TEXT_DOMAIN, "no history available"));
case EZFS_POOLPROPS:
return (dgettext(TEXT_DOMAIN, "failed to retrieve "
"pool properties"));
case EZFS_POOL_NOTSUP:
return (dgettext(TEXT_DOMAIN, "operation not supported "
"on this type of pool"));
case EZFS_POOL_INVALARG:
return (dgettext(TEXT_DOMAIN, "invalid argument for "
"this pool operation"));
case EZFS_NAMETOOLONG:
return (dgettext(TEXT_DOMAIN, "dataset name is too long"));
case EZFS_OPENFAILED:
return (dgettext(TEXT_DOMAIN, "open failed"));
case EZFS_NOCAP:
return (dgettext(TEXT_DOMAIN,
"disk capacity information could not be retrieved"));
case EZFS_LABELFAILED:
return (dgettext(TEXT_DOMAIN, "write of label failed"));
case EZFS_BADWHO:
return (dgettext(TEXT_DOMAIN, "invalid user/group"));
case EZFS_BADPERM:
return (dgettext(TEXT_DOMAIN, "invalid permission"));
case EZFS_BADPERMSET:
return (dgettext(TEXT_DOMAIN, "invalid permission set name"));
case EZFS_NODELEGATION:
return (dgettext(TEXT_DOMAIN, "delegated administration is "
"disabled on pool"));
case EZFS_BADCACHE:
return (dgettext(TEXT_DOMAIN, "invalid or missing cache file"));
case EZFS_ISL2CACHE:
return (dgettext(TEXT_DOMAIN, "device is in use as a cache"));
case EZFS_VDEVNOTSUP:
return (dgettext(TEXT_DOMAIN, "vdev specification is not "
"supported"));
case EZFS_NOTSUP:
return (dgettext(TEXT_DOMAIN, "operation not supported "
"on this dataset"));
case EZFS_IOC_NOTSUPPORTED:
return (dgettext(TEXT_DOMAIN, "operation not supported by "
"zfs kernel module"));
case EZFS_ACTIVE_SPARE:
return (dgettext(TEXT_DOMAIN, "pool has active shared spare "
"device"));
case EZFS_UNPLAYED_LOGS:
return (dgettext(TEXT_DOMAIN, "log device has unplayed intent "
"logs"));
case EZFS_REFTAG_RELE:
return (dgettext(TEXT_DOMAIN, "no such tag on this dataset"));
case EZFS_REFTAG_HOLD:
return (dgettext(TEXT_DOMAIN, "tag already exists on this "
"dataset"));
case EZFS_TAGTOOLONG:
return (dgettext(TEXT_DOMAIN, "tag too long"));
case EZFS_PIPEFAILED:
return (dgettext(TEXT_DOMAIN, "pipe create failed"));
case EZFS_THREADCREATEFAILED:
return (dgettext(TEXT_DOMAIN, "thread create failed"));
case EZFS_POSTSPLIT_ONLINE:
return (dgettext(TEXT_DOMAIN, "disk was split from this pool "
"into a new one"));
case EZFS_SCRUB_PAUSED:
return (dgettext(TEXT_DOMAIN, "scrub is paused; "
"use 'zpool scrub' to resume scrub"));
case EZFS_SCRUB_PAUSED_TO_CANCEL:
return (dgettext(TEXT_DOMAIN, "scrub is paused; "
"use 'zpool scrub' to resume or 'zpool scrub -s' to "
"cancel scrub"));
case EZFS_SCRUBBING:
return (dgettext(TEXT_DOMAIN, "currently scrubbing; "
"use 'zpool scrub -s' to cancel scrub"));
case EZFS_ERRORSCRUBBING:
return (dgettext(TEXT_DOMAIN, "currently error scrubbing; "
"use 'zpool scrub -s' to cancel error scrub"));
case EZFS_ERRORSCRUB_PAUSED:
return (dgettext(TEXT_DOMAIN, "error scrub is paused; "
"use 'zpool scrub -e' to resume error scrub"));
case EZFS_NO_SCRUB:
return (dgettext(TEXT_DOMAIN, "there is no active scrub"));
case EZFS_DIFF:
return (dgettext(TEXT_DOMAIN, "unable to generate diffs"));
case EZFS_DIFFDATA:
return (dgettext(TEXT_DOMAIN, "invalid diff data"));
case EZFS_POOLREADONLY:
return (dgettext(TEXT_DOMAIN, "pool is read-only"));
case EZFS_NO_PENDING:
return (dgettext(TEXT_DOMAIN, "operation is not "
"in progress"));
case EZFS_CHECKPOINT_EXISTS:
return (dgettext(TEXT_DOMAIN, "checkpoint exists"));
case EZFS_DISCARDING_CHECKPOINT:
return (dgettext(TEXT_DOMAIN, "currently discarding "
"checkpoint"));
case EZFS_NO_CHECKPOINT:
return (dgettext(TEXT_DOMAIN, "checkpoint does not exist"));
case EZFS_DEVRM_IN_PROGRESS:
return (dgettext(TEXT_DOMAIN, "device removal in progress"));
case EZFS_VDEV_TOO_BIG:
return (dgettext(TEXT_DOMAIN, "device exceeds supported size"));
case EZFS_ACTIVE_POOL:
return (dgettext(TEXT_DOMAIN, "pool is imported on a "
"different host"));
case EZFS_CRYPTOFAILED:
return (dgettext(TEXT_DOMAIN, "encryption failure"));
case EZFS_TOOMANY:
return (dgettext(TEXT_DOMAIN, "argument list too long"));
case EZFS_INITIALIZING:
return (dgettext(TEXT_DOMAIN, "currently initializing"));
case EZFS_NO_INITIALIZE:
return (dgettext(TEXT_DOMAIN, "there is no active "
"initialization"));
case EZFS_WRONG_PARENT:
return (dgettext(TEXT_DOMAIN, "invalid parent dataset"));
case EZFS_TRIMMING:
return (dgettext(TEXT_DOMAIN, "currently trimming"));
case EZFS_NO_TRIM:
return (dgettext(TEXT_DOMAIN, "there is no active trim"));
case EZFS_TRIM_NOTSUP:
return (dgettext(TEXT_DOMAIN, "trim operations are not "
"supported by this device"));
case EZFS_NO_RESILVER_DEFER:
return (dgettext(TEXT_DOMAIN, "this action requires the "
"resilver_defer feature"));
case EZFS_EXPORT_IN_PROGRESS:
return (dgettext(TEXT_DOMAIN, "pool export in progress"));
case EZFS_REBUILDING:
return (dgettext(TEXT_DOMAIN, "currently sequentially "
"resilvering"));
case EZFS_VDEV_NOTSUP:
return (dgettext(TEXT_DOMAIN, "operation not supported "
"on this type of vdev"));
case EZFS_NOT_USER_NAMESPACE:
return (dgettext(TEXT_DOMAIN, "the provided file "
"was not a user namespace file"));
case EZFS_RESUME_EXISTS:
return (dgettext(TEXT_DOMAIN, "Resuming recv on existing "
"dataset without force"));
+ case EZFS_RAIDZ_EXPAND_IN_PROGRESS:
+ return (dgettext(TEXT_DOMAIN, "raidz expansion in progress"));
case EZFS_UNKNOWN:
return (dgettext(TEXT_DOMAIN, "unknown error"));
default:
assert(hdl->libzfs_error == 0);
return (dgettext(TEXT_DOMAIN, "no error"));
}
}
void
zfs_error_aux(libzfs_handle_t *hdl, const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
(void) vsnprintf(hdl->libzfs_desc, sizeof (hdl->libzfs_desc),
fmt, ap);
hdl->libzfs_desc_active = 1;
va_end(ap);
}
static void
zfs_verror(libzfs_handle_t *hdl, int error, const char *fmt, va_list ap)
{
(void) vsnprintf(hdl->libzfs_action, sizeof (hdl->libzfs_action),
fmt, ap);
hdl->libzfs_error = error;
if (hdl->libzfs_desc_active)
hdl->libzfs_desc_active = 0;
else
hdl->libzfs_desc[0] = '\0';
if (hdl->libzfs_printerr) {
if (error == EZFS_UNKNOWN) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN, "internal "
"error: %s: %s\n"), hdl->libzfs_action,
libzfs_error_description(hdl));
abort();
}
(void) fprintf(stderr, "%s: %s\n", hdl->libzfs_action,
libzfs_error_description(hdl));
if (error == EZFS_NOMEM)
exit(1);
}
}
int
zfs_error(libzfs_handle_t *hdl, int error, const char *msg)
{
return (zfs_error_fmt(hdl, error, "%s", msg));
}
int
zfs_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
zfs_verror(hdl, error, fmt, ap);
va_end(ap);
return (-1);
}
static int
zfs_common_error(libzfs_handle_t *hdl, int error, const char *fmt,
va_list ap)
{
switch (error) {
case EPERM:
case EACCES:
zfs_verror(hdl, EZFS_PERM, fmt, ap);
return (-1);
case ECANCELED:
zfs_verror(hdl, EZFS_NODELEGATION, fmt, ap);
return (-1);
case EIO:
zfs_verror(hdl, EZFS_IO, fmt, ap);
return (-1);
case EFAULT:
zfs_verror(hdl, EZFS_FAULT, fmt, ap);
return (-1);
case EINTR:
zfs_verror(hdl, EZFS_INTR, fmt, ap);
return (-1);
case ECKSUM:
zfs_verror(hdl, EZFS_CKSUM, fmt, ap);
return (-1);
}
return (0);
}
int
zfs_standard_error(libzfs_handle_t *hdl, int error, const char *msg)
{
return (zfs_standard_error_fmt(hdl, error, "%s", msg));
}
int
zfs_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
if (zfs_common_error(hdl, error, fmt, ap) != 0) {
va_end(ap);
return (-1);
}
switch (error) {
case ENXIO:
case ENODEV:
case EPIPE:
zfs_verror(hdl, EZFS_IO, fmt, ap);
break;
case ENOENT:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset does not exist"));
zfs_verror(hdl, EZFS_NOENT, fmt, ap);
break;
case ENOSPC:
case EDQUOT:
zfs_verror(hdl, EZFS_NOSPC, fmt, ap);
break;
case EEXIST:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset already exists"));
zfs_verror(hdl, EZFS_EXISTS, fmt, ap);
break;
case EBUSY:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset is busy"));
zfs_verror(hdl, EZFS_BUSY, fmt, ap);
break;
case EROFS:
zfs_verror(hdl, EZFS_POOLREADONLY, fmt, ap);
break;
case ENAMETOOLONG:
zfs_verror(hdl, EZFS_NAMETOOLONG, fmt, ap);
break;
case ENOTSUP:
zfs_verror(hdl, EZFS_BADVERSION, fmt, ap);
break;
case EAGAIN:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool I/O is currently suspended"));
zfs_verror(hdl, EZFS_POOLUNAVAIL, fmt, ap);
break;
case EREMOTEIO:
zfs_verror(hdl, EZFS_ACTIVE_POOL, fmt, ap);
break;
case ZFS_ERR_UNKNOWN_SEND_STREAM_FEATURE:
case ZFS_ERR_IOC_CMD_UNAVAIL:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs "
"module does not support this operation. A reboot may "
"be required to enable this operation."));
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
case ZFS_ERR_IOC_ARG_UNAVAIL:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs "
"module does not support an option for this operation. "
"A reboot may be required to enable this option."));
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
case ZFS_ERR_IOC_ARG_REQUIRED:
case ZFS_ERR_IOC_ARG_BADTYPE:
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
case ZFS_ERR_WRONG_PARENT:
zfs_verror(hdl, EZFS_WRONG_PARENT, fmt, ap);
break;
case ZFS_ERR_BADPROP:
zfs_verror(hdl, EZFS_BADPROP, fmt, ap);
break;
case ZFS_ERR_NOT_USER_NAMESPACE:
zfs_verror(hdl, EZFS_NOT_USER_NAMESPACE, fmt, ap);
break;
default:
zfs_error_aux(hdl, "%s", strerror(error));
zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap);
break;
}
va_end(ap);
return (-1);
}
void
zfs_setprop_error(libzfs_handle_t *hdl, zfs_prop_t prop, int err,
char *errbuf)
{
switch (err) {
case ENOSPC:
/*
* For quotas and reservations, ENOSPC indicates
* something different; setting a quota or reservation
* doesn't use any disk space.
*/
switch (prop) {
case ZFS_PROP_QUOTA:
case ZFS_PROP_REFQUOTA:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"size is less than current used or "
"reserved space"));
(void) zfs_error(hdl, EZFS_PROPSPACE, errbuf);
break;
case ZFS_PROP_RESERVATION:
case ZFS_PROP_REFRESERVATION:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"size is greater than available space"));
(void) zfs_error(hdl, EZFS_PROPSPACE, errbuf);
break;
default:
(void) zfs_standard_error(hdl, err, errbuf);
break;
}
break;
case EBUSY:
(void) zfs_standard_error(hdl, EBUSY, errbuf);
break;
case EROFS:
(void) zfs_error(hdl, EZFS_DSREADONLY, errbuf);
break;
case E2BIG:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property value too long"));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
break;
case ENOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool and or dataset must be upgraded to set this "
"property or value"));
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
break;
case ERANGE:
if (prop == ZFS_PROP_COMPRESSION ||
prop == ZFS_PROP_DNODESIZE ||
prop == ZFS_PROP_RECORDSIZE) {
(void) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property setting is not allowed on "
"bootable datasets"));
(void) zfs_error(hdl, EZFS_NOTSUP, errbuf);
} else if (prop == ZFS_PROP_CHECKSUM ||
prop == ZFS_PROP_DEDUP) {
(void) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property setting is not allowed on "
"root pools"));
(void) zfs_error(hdl, EZFS_NOTSUP, errbuf);
} else {
(void) zfs_standard_error(hdl, err, errbuf);
}
break;
case EINVAL:
if (prop == ZPROP_INVAL) {
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
} else {
(void) zfs_standard_error(hdl, err, errbuf);
}
break;
case ZFS_ERR_BADPROP:
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
break;
case EACCES:
if (prop == ZFS_PROP_KEYLOCATION) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"keylocation may only be set on encryption roots"));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
} else {
(void) zfs_standard_error(hdl, err, errbuf);
}
break;
case EOVERFLOW:
/*
* This platform can't address a volume this big.
*/
#ifdef _ILP32
if (prop == ZFS_PROP_VOLSIZE) {
(void) zfs_error(hdl, EZFS_VOLTOOBIG, errbuf);
break;
}
zfs_fallthrough;
#endif
default:
(void) zfs_standard_error(hdl, err, errbuf);
}
}
int
zpool_standard_error(libzfs_handle_t *hdl, int error, const char *msg)
{
return (zpool_standard_error_fmt(hdl, error, "%s", msg));
}
int
zpool_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
if (zfs_common_error(hdl, error, fmt, ap) != 0) {
va_end(ap);
return (-1);
}
switch (error) {
case ENODEV:
zfs_verror(hdl, EZFS_NODEVICE, fmt, ap);
break;
case ENOENT:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "no such pool or dataset"));
zfs_verror(hdl, EZFS_NOENT, fmt, ap);
break;
case EEXIST:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool already exists"));
zfs_verror(hdl, EZFS_EXISTS, fmt, ap);
break;
case EBUSY:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool is busy"));
zfs_verror(hdl, EZFS_BUSY, fmt, ap);
break;
/* There is no pending operation to cancel */
case ENOTACTIVE:
zfs_verror(hdl, EZFS_NO_PENDING, fmt, ap);
break;
case ENXIO:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices is currently unavailable"));
zfs_verror(hdl, EZFS_BADDEV, fmt, ap);
break;
case ENAMETOOLONG:
zfs_verror(hdl, EZFS_DEVOVERFLOW, fmt, ap);
break;
case ENOTSUP:
zfs_verror(hdl, EZFS_POOL_NOTSUP, fmt, ap);
break;
case EINVAL:
zfs_verror(hdl, EZFS_POOL_INVALARG, fmt, ap);
break;
case ENOSPC:
case EDQUOT:
zfs_verror(hdl, EZFS_NOSPC, fmt, ap);
break;
case EAGAIN:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool I/O is currently suspended"));
zfs_verror(hdl, EZFS_POOLUNAVAIL, fmt, ap);
break;
case EROFS:
zfs_verror(hdl, EZFS_POOLREADONLY, fmt, ap);
break;
case EDOM:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"block size out of range or does not match"));
zfs_verror(hdl, EZFS_BADPROP, fmt, ap);
break;
case EREMOTEIO:
zfs_verror(hdl, EZFS_ACTIVE_POOL, fmt, ap);
break;
case ZFS_ERR_CHECKPOINT_EXISTS:
zfs_verror(hdl, EZFS_CHECKPOINT_EXISTS, fmt, ap);
break;
case ZFS_ERR_DISCARDING_CHECKPOINT:
zfs_verror(hdl, EZFS_DISCARDING_CHECKPOINT, fmt, ap);
break;
case ZFS_ERR_NO_CHECKPOINT:
zfs_verror(hdl, EZFS_NO_CHECKPOINT, fmt, ap);
break;
case ZFS_ERR_DEVRM_IN_PROGRESS:
zfs_verror(hdl, EZFS_DEVRM_IN_PROGRESS, fmt, ap);
break;
case ZFS_ERR_VDEV_TOO_BIG:
zfs_verror(hdl, EZFS_VDEV_TOO_BIG, fmt, ap);
break;
case ZFS_ERR_EXPORT_IN_PROGRESS:
zfs_verror(hdl, EZFS_EXPORT_IN_PROGRESS, fmt, ap);
break;
case ZFS_ERR_RESILVER_IN_PROGRESS:
zfs_verror(hdl, EZFS_RESILVERING, fmt, ap);
break;
case ZFS_ERR_REBUILD_IN_PROGRESS:
zfs_verror(hdl, EZFS_REBUILDING, fmt, ap);
break;
case ZFS_ERR_BADPROP:
zfs_verror(hdl, EZFS_BADPROP, fmt, ap);
break;
case ZFS_ERR_VDEV_NOTSUP:
zfs_verror(hdl, EZFS_VDEV_NOTSUP, fmt, ap);
break;
case ZFS_ERR_IOC_CMD_UNAVAIL:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs "
"module does not support this operation. A reboot may "
"be required to enable this operation."));
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
case ZFS_ERR_IOC_ARG_UNAVAIL:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs "
"module does not support an option for this operation. "
"A reboot may be required to enable this option."));
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
case ZFS_ERR_IOC_ARG_REQUIRED:
case ZFS_ERR_IOC_ARG_BADTYPE:
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
+ case ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS:
+ zfs_verror(hdl, EZFS_RAIDZ_EXPAND_IN_PROGRESS, fmt, ap);
+ break;
default:
zfs_error_aux(hdl, "%s", strerror(error));
zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap);
}
va_end(ap);
return (-1);
}
/*
* Display an out of memory error message and abort the current program.
*/
int
no_memory(libzfs_handle_t *hdl)
{
return (zfs_error(hdl, EZFS_NOMEM, "internal error"));
}
/*
* A safe form of malloc() which will die if the allocation fails.
*/
void *
zfs_alloc(libzfs_handle_t *hdl, size_t size)
{
void *data;
if ((data = calloc(1, size)) == NULL)
(void) no_memory(hdl);
return (data);
}
/*
* A safe form of asprintf() which will die if the allocation fails.
*/
char *
zfs_asprintf(libzfs_handle_t *hdl, const char *fmt, ...)
{
va_list ap;
char *ret;
int err;
va_start(ap, fmt);
err = vasprintf(&ret, fmt, ap);
va_end(ap);
if (err < 0) {
(void) no_memory(hdl);
ret = NULL;
}
return (ret);
}
/*
* A safe form of realloc(), which also zeroes newly allocated space.
*/
void *
zfs_realloc(libzfs_handle_t *hdl, void *ptr, size_t oldsize, size_t newsize)
{
void *ret;
if ((ret = realloc(ptr, newsize)) == NULL) {
(void) no_memory(hdl);
return (NULL);
}
memset((char *)ret + oldsize, 0, newsize - oldsize);
return (ret);
}
/*
* A safe form of strdup() which will die if the allocation fails.
*/
char *
zfs_strdup(libzfs_handle_t *hdl, const char *str)
{
char *ret;
if ((ret = strdup(str)) == NULL)
(void) no_memory(hdl);
return (ret);
}
void
libzfs_print_on_error(libzfs_handle_t *hdl, boolean_t printerr)
{
hdl->libzfs_printerr = printerr;
}
/*
* Read lines from an open file descriptor and store them in an array of
* strings until EOF. lines[] will be allocated and populated with all the
* lines read. All newlines are replaced with NULL terminators for
* convenience. lines[] must be freed after use with libzfs_free_str_array().
*
* Returns the number of lines read.
*/
static int
libzfs_read_stdout_from_fd(int fd, char **lines[])
{
FILE *fp;
int lines_cnt = 0;
size_t len = 0;
char *line = NULL;
char **tmp_lines = NULL, **tmp;
fp = fdopen(fd, "r");
if (fp == NULL) {
close(fd);
return (0);
}
while (getline(&line, &len, fp) != -1) {
tmp = realloc(tmp_lines, sizeof (*tmp_lines) * (lines_cnt + 1));
if (tmp == NULL) {
/* Return the lines we were able to process */
break;
}
tmp_lines = tmp;
/* Remove newline if not EOF */
if (line[strlen(line) - 1] == '\n')
line[strlen(line) - 1] = '\0';
tmp_lines[lines_cnt] = strdup(line);
if (tmp_lines[lines_cnt] == NULL)
break;
++lines_cnt;
}
free(line);
fclose(fp);
*lines = tmp_lines;
return (lines_cnt);
}
static int
libzfs_run_process_impl(const char *path, char *argv[], char *env[], int flags,
char **lines[], int *lines_cnt)
{
pid_t pid;
int error, devnull_fd;
int link[2];
/*
* Setup a pipe between our child and parent process if we're
* reading stdout.
*/
if (lines != NULL && pipe2(link, O_NONBLOCK | O_CLOEXEC) == -1)
return (-EPIPE);
pid = fork();
if (pid == 0) {
/* Child process */
devnull_fd = open("/dev/null", O_WRONLY | O_CLOEXEC);
if (devnull_fd < 0)
_exit(-1);
if (!(flags & STDOUT_VERBOSE) && (lines == NULL))
(void) dup2(devnull_fd, STDOUT_FILENO);
else if (lines != NULL) {
/* Save the output to lines[] */
dup2(link[1], STDOUT_FILENO);
}
if (!(flags & STDERR_VERBOSE))
(void) dup2(devnull_fd, STDERR_FILENO);
if (flags & NO_DEFAULT_PATH) {
if (env == NULL)
execv(path, argv);
else
execve(path, argv, env);
} else {
if (env == NULL)
execvp(path, argv);
else
execvpe(path, argv, env);
}
_exit(-1);
} else if (pid > 0) {
/* Parent process */
int status;
while ((error = waitpid(pid, &status, 0)) == -1 &&
errno == EINTR)
;
if (error < 0 || !WIFEXITED(status))
return (-1);
if (lines != NULL) {
close(link[1]);
*lines_cnt = libzfs_read_stdout_from_fd(link[0], lines);
}
return (WEXITSTATUS(status));
}
return (-1);
}
int
libzfs_run_process(const char *path, char *argv[], int flags)
{
return (libzfs_run_process_impl(path, argv, NULL, flags, NULL, NULL));
}
/*
* Run a command and store its stdout lines in an array of strings (lines[]).
* lines[] is allocated and populated for you, and the number of lines is set in
* lines_cnt. lines[] must be freed after use with libzfs_free_str_array().
* All newlines (\n) in lines[] are terminated for convenience.
*/
int
libzfs_run_process_get_stdout(const char *path, char *argv[], char *env[],
char **lines[], int *lines_cnt)
{
return (libzfs_run_process_impl(path, argv, env, 0, lines, lines_cnt));
}
/*
* Same as libzfs_run_process_get_stdout(), but run without $PATH set. This
* means that *path needs to be the full path to the executable.
*/
int
libzfs_run_process_get_stdout_nopath(const char *path, char *argv[],
char *env[], char **lines[], int *lines_cnt)
{
return (libzfs_run_process_impl(path, argv, env, NO_DEFAULT_PATH,
lines, lines_cnt));
}
/*
* Free an array of strings. Free both the strings contained in the array and
* the array itself.
*/
void
libzfs_free_str_array(char **strs, int count)
{
while (--count >= 0)
free(strs[count]);
free(strs);
}
/*
* Returns 1 if environment variable is set to "YES", "yes", "ON", "on", or
* a non-zero number.
*
* Returns 0 otherwise.
*/
boolean_t
libzfs_envvar_is_set(const char *envvar)
{
char *env = getenv(envvar);
return (env && (strtoul(env, NULL, 0) > 0 ||
(!strncasecmp(env, "YES", 3) && strnlen(env, 4) == 3) ||
(!strncasecmp(env, "ON", 2) && strnlen(env, 3) == 2)));
}
libzfs_handle_t *
libzfs_init(void)
{
libzfs_handle_t *hdl;
int error;
char *env;
if ((error = libzfs_load_module()) != 0) {
errno = error;
return (NULL);
}
if ((hdl = calloc(1, sizeof (libzfs_handle_t))) == NULL) {
return (NULL);
}
if (regcomp(&hdl->libzfs_urire, URI_REGEX, 0) != 0) {
free(hdl);
return (NULL);
}
if ((hdl->libzfs_fd = open(ZFS_DEV, O_RDWR|O_EXCL|O_CLOEXEC)) < 0) {
free(hdl);
return (NULL);
}
if (libzfs_core_init() != 0) {
(void) close(hdl->libzfs_fd);
free(hdl);
return (NULL);
}
zfs_prop_init();
zpool_prop_init();
zpool_feature_init();
vdev_prop_init();
libzfs_mnttab_init(hdl);
fletcher_4_init();
if (getenv("ZFS_PROP_DEBUG") != NULL) {
hdl->libzfs_prop_debug = B_TRUE;
}
if ((env = getenv("ZFS_SENDRECV_MAX_NVLIST")) != NULL) {
if ((error = zfs_nicestrtonum(hdl, env,
&hdl->libzfs_max_nvlist))) {
errno = error;
(void) close(hdl->libzfs_fd);
free(hdl);
return (NULL);
}
} else {
hdl->libzfs_max_nvlist = (SPA_MAXBLOCKSIZE * 4);
}
/*
* For testing, remove some settable properties and features
*/
if (libzfs_envvar_is_set("ZFS_SYSFS_PROP_SUPPORT_TEST")) {
zprop_desc_t *proptbl;
proptbl = zpool_prop_get_table();
proptbl[ZPOOL_PROP_COMMENT].pd_zfs_mod_supported = B_FALSE;
proptbl = zfs_prop_get_table();
proptbl[ZFS_PROP_DNODESIZE].pd_zfs_mod_supported = B_FALSE;
zfeature_info_t *ftbl = spa_feature_table;
ftbl[SPA_FEATURE_LARGE_BLOCKS].fi_zfs_mod_supported = B_FALSE;
}
return (hdl);
}
void
libzfs_fini(libzfs_handle_t *hdl)
{
(void) close(hdl->libzfs_fd);
zpool_free_handles(hdl);
namespace_clear(hdl);
libzfs_mnttab_fini(hdl);
libzfs_core_fini();
regfree(&hdl->libzfs_urire);
fletcher_4_fini();
#if LIBFETCH_DYNAMIC
if (hdl->libfetch != (void *)-1 && hdl->libfetch != NULL)
(void) dlclose(hdl->libfetch);
free(hdl->libfetch_load_error);
#endif
free(hdl);
}
libzfs_handle_t *
zpool_get_handle(zpool_handle_t *zhp)
{
return (zhp->zpool_hdl);
}
libzfs_handle_t *
zfs_get_handle(zfs_handle_t *zhp)
{
return (zhp->zfs_hdl);
}
zpool_handle_t *
zfs_get_pool_handle(const zfs_handle_t *zhp)
{
return (zhp->zpool_hdl);
}
/*
* Given a name, determine whether or not it's a valid path
* (starts with '/' or "./"). If so, walk the mnttab trying
* to match the device number. If not, treat the path as an
* fs/vol/snap/bkmark name.
*/
zfs_handle_t *
zfs_path_to_zhandle(libzfs_handle_t *hdl, const char *path, zfs_type_t argtype)
{
struct stat64 statbuf;
struct extmnttab entry;
if (path[0] != '/' && strncmp(path, "./", strlen("./")) != 0) {
/*
* It's not a valid path, assume it's a name of type 'argtype'.
*/
return (zfs_open(hdl, path, argtype));
}
if (getextmntent(path, &entry, &statbuf) != 0)
return (NULL);
if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0) {
(void) fprintf(stderr, gettext("'%s': not a ZFS filesystem\n"),
path);
return (NULL);
}
return (zfs_open(hdl, entry.mnt_special, ZFS_TYPE_FILESYSTEM));
}
/*
* Initialize the zc_nvlist_dst member to prepare for receiving an nvlist from
* an ioctl().
*/
void
zcmd_alloc_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, size_t len)
{
if (len == 0)
len = 256 * 1024;
zc->zc_nvlist_dst_size = len;
zc->zc_nvlist_dst =
(uint64_t)(uintptr_t)zfs_alloc(hdl, zc->zc_nvlist_dst_size);
}
/*
* Called when an ioctl() which returns an nvlist fails with ENOMEM. This will
* expand the nvlist to the size specified in 'zc_nvlist_dst_size', which was
* filled in by the kernel to indicate the actual required size.
*/
void
zcmd_expand_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc)
{
free((void *)(uintptr_t)zc->zc_nvlist_dst);
zc->zc_nvlist_dst =
(uint64_t)(uintptr_t)zfs_alloc(hdl, zc->zc_nvlist_dst_size);
}
/*
* Called to free the src and dst nvlists stored in the command structure.
*/
void
zcmd_free_nvlists(zfs_cmd_t *zc)
{
free((void *)(uintptr_t)zc->zc_nvlist_conf);
free((void *)(uintptr_t)zc->zc_nvlist_src);
free((void *)(uintptr_t)zc->zc_nvlist_dst);
zc->zc_nvlist_conf = 0;
zc->zc_nvlist_src = 0;
zc->zc_nvlist_dst = 0;
}
static void
zcmd_write_nvlist_com(libzfs_handle_t *hdl, uint64_t *outnv, uint64_t *outlen,
nvlist_t *nvl)
{
char *packed;
size_t len = fnvlist_size(nvl);
packed = zfs_alloc(hdl, len);
verify(nvlist_pack(nvl, &packed, &len, NV_ENCODE_NATIVE, 0) == 0);
*outnv = (uint64_t)(uintptr_t)packed;
*outlen = len;
}
void
zcmd_write_conf_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl)
{
zcmd_write_nvlist_com(hdl, &zc->zc_nvlist_conf,
&zc->zc_nvlist_conf_size, nvl);
}
void
zcmd_write_src_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl)
{
zcmd_write_nvlist_com(hdl, &zc->zc_nvlist_src,
&zc->zc_nvlist_src_size, nvl);
}
/*
* Unpacks an nvlist from the ZFS ioctl command structure.
*/
int
zcmd_read_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t **nvlp)
{
if (nvlist_unpack((void *)(uintptr_t)zc->zc_nvlist_dst,
zc->zc_nvlist_dst_size, nvlp, 0) != 0)
return (no_memory(hdl));
return (0);
}
/*
* ================================================================
* API shared by zfs and zpool property management
* ================================================================
*/
static void
zprop_print_headers(zprop_get_cbdata_t *cbp, zfs_type_t type)
{
zprop_list_t *pl;
int i;
char *title;
size_t len;
cbp->cb_first = B_FALSE;
if (cbp->cb_scripted)
return;
/*
* Start with the length of the column headers.
*/
cbp->cb_colwidths[GET_COL_NAME] = strlen(dgettext(TEXT_DOMAIN, "NAME"));
cbp->cb_colwidths[GET_COL_PROPERTY] = strlen(dgettext(TEXT_DOMAIN,
"PROPERTY"));
cbp->cb_colwidths[GET_COL_VALUE] = strlen(dgettext(TEXT_DOMAIN,
"VALUE"));
cbp->cb_colwidths[GET_COL_RECVD] = strlen(dgettext(TEXT_DOMAIN,
"RECEIVED"));
cbp->cb_colwidths[GET_COL_SOURCE] = strlen(dgettext(TEXT_DOMAIN,
"SOURCE"));
/* first property is always NAME */
assert(cbp->cb_proplist->pl_prop ==
((type == ZFS_TYPE_POOL) ? ZPOOL_PROP_NAME :
((type == ZFS_TYPE_VDEV) ? VDEV_PROP_NAME : ZFS_PROP_NAME)));
/*
* Go through and calculate the widths for each column. For the
* 'source' column, we kludge it up by taking the worst-case scenario of
* inheriting from the longest name. This is acceptable because in the
* majority of cases 'SOURCE' is the last column displayed, and we don't
* use the width anyway. Note that the 'VALUE' column can be oversized,
* if the name of the property is much longer than any values we find.
*/
for (pl = cbp->cb_proplist; pl != NULL; pl = pl->pl_next) {
/*
* 'PROPERTY' column
*/
if (pl->pl_prop != ZPROP_USERPROP) {
const char *propname = (type == ZFS_TYPE_POOL) ?
zpool_prop_to_name(pl->pl_prop) :
((type == ZFS_TYPE_VDEV) ?
vdev_prop_to_name(pl->pl_prop) :
zfs_prop_to_name(pl->pl_prop));
assert(propname != NULL);
len = strlen(propname);
if (len > cbp->cb_colwidths[GET_COL_PROPERTY])
cbp->cb_colwidths[GET_COL_PROPERTY] = len;
} else {
assert(pl->pl_user_prop != NULL);
len = strlen(pl->pl_user_prop);
if (len > cbp->cb_colwidths[GET_COL_PROPERTY])
cbp->cb_colwidths[GET_COL_PROPERTY] = len;
}
/*
* 'VALUE' column. The first property is always the 'name'
* property that was tacked on either by /sbin/zfs's
* zfs_do_get() or when calling zprop_expand_list(), so we
* ignore its width. If the user specified the name property
* to display, then it will be later in the list in any case.
*/
if (pl != cbp->cb_proplist &&
pl->pl_width > cbp->cb_colwidths[GET_COL_VALUE])
cbp->cb_colwidths[GET_COL_VALUE] = pl->pl_width;
/* 'RECEIVED' column. */
if (pl != cbp->cb_proplist &&
pl->pl_recvd_width > cbp->cb_colwidths[GET_COL_RECVD])
cbp->cb_colwidths[GET_COL_RECVD] = pl->pl_recvd_width;
/*
* 'NAME' and 'SOURCE' columns
*/
if (pl->pl_prop == ((type == ZFS_TYPE_POOL) ? ZPOOL_PROP_NAME :
((type == ZFS_TYPE_VDEV) ? VDEV_PROP_NAME :
ZFS_PROP_NAME)) && pl->pl_width >
cbp->cb_colwidths[GET_COL_NAME]) {
cbp->cb_colwidths[GET_COL_NAME] = pl->pl_width;
cbp->cb_colwidths[GET_COL_SOURCE] = pl->pl_width +
strlen(dgettext(TEXT_DOMAIN, "inherited from"));
}
}
/*
* Now go through and print the headers.
*/
for (i = 0; i < ZFS_GET_NCOLS; i++) {
switch (cbp->cb_columns[i]) {
case GET_COL_NAME:
title = dgettext(TEXT_DOMAIN, "NAME");
break;
case GET_COL_PROPERTY:
title = dgettext(TEXT_DOMAIN, "PROPERTY");
break;
case GET_COL_VALUE:
title = dgettext(TEXT_DOMAIN, "VALUE");
break;
case GET_COL_RECVD:
title = dgettext(TEXT_DOMAIN, "RECEIVED");
break;
case GET_COL_SOURCE:
title = dgettext(TEXT_DOMAIN, "SOURCE");
break;
default:
title = NULL;
}
if (title != NULL) {
if (i == (ZFS_GET_NCOLS - 1) ||
cbp->cb_columns[i + 1] == GET_COL_NONE)
(void) printf("%s", title);
else
(void) printf("%-*s ",
cbp->cb_colwidths[cbp->cb_columns[i]],
title);
}
}
(void) printf("\n");
}
/*
* Display a single line of output, according to the settings in the callback
* structure.
*/
void
zprop_print_one_property(const char *name, zprop_get_cbdata_t *cbp,
const char *propname, const char *value, zprop_source_t sourcetype,
const char *source, const char *recvd_value)
{
int i;
const char *str = NULL;
char buf[128];
/*
* Ignore those source types that the user has chosen to ignore.
*/
if ((sourcetype & cbp->cb_sources) == 0)
return;
if (cbp->cb_first)
zprop_print_headers(cbp, cbp->cb_type);
for (i = 0; i < ZFS_GET_NCOLS; i++) {
switch (cbp->cb_columns[i]) {
case GET_COL_NAME:
str = name;
break;
case GET_COL_PROPERTY:
str = propname;
break;
case GET_COL_VALUE:
str = value;
break;
case GET_COL_SOURCE:
switch (sourcetype) {
case ZPROP_SRC_NONE:
str = "-";
break;
case ZPROP_SRC_DEFAULT:
str = "default";
break;
case ZPROP_SRC_LOCAL:
str = "local";
break;
case ZPROP_SRC_TEMPORARY:
str = "temporary";
break;
case ZPROP_SRC_INHERITED:
(void) snprintf(buf, sizeof (buf),
"inherited from %s", source);
str = buf;
break;
case ZPROP_SRC_RECEIVED:
str = "received";
break;
default:
str = NULL;
assert(!"unhandled zprop_source_t");
}
break;
case GET_COL_RECVD:
str = (recvd_value == NULL ? "-" : recvd_value);
break;
default:
continue;
}
if (i == (ZFS_GET_NCOLS - 1) ||
cbp->cb_columns[i + 1] == GET_COL_NONE)
(void) printf("%s", str);
else if (cbp->cb_scripted)
(void) printf("%s\t", str);
else
(void) printf("%-*s ",
cbp->cb_colwidths[cbp->cb_columns[i]],
str);
}
(void) printf("\n");
}
/*
* Given a numeric suffix, convert the value into a number of bits that the
* resulting value must be shifted.
*/
static int
str2shift(libzfs_handle_t *hdl, const char *buf)
{
const char *ends = "BKMGTPEZ";
int i;
if (buf[0] == '\0')
return (0);
for (i = 0; i < strlen(ends); i++) {
if (toupper(buf[0]) == ends[i])
break;
}
if (i == strlen(ends)) {
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid numeric suffix '%s'"), buf);
return (-1);
}
/*
* Allow 'G' = 'GB' = 'GiB', case-insensitively.
* However, 'BB' and 'BiB' are disallowed.
*/
if (buf[1] == '\0' ||
(toupper(buf[0]) != 'B' &&
((toupper(buf[1]) == 'B' && buf[2] == '\0') ||
(toupper(buf[1]) == 'I' && toupper(buf[2]) == 'B' &&
buf[3] == '\0'))))
return (10 * i);
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid numeric suffix '%s'"), buf);
return (-1);
}
/*
* Convert a string of the form '100G' into a real number. Used when setting
* properties or creating a volume. 'buf' is used to place an extended error
* message for the caller to use.
*/
int
zfs_nicestrtonum(libzfs_handle_t *hdl, const char *value, uint64_t *num)
{
char *end;
int shift;
*num = 0;
/* Check to see if this looks like a number. */
if ((value[0] < '0' || value[0] > '9') && value[0] != '.') {
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"bad numeric value '%s'"), value);
return (-1);
}
/* Rely on strtoull() to process the numeric portion. */
errno = 0;
*num = strtoull(value, &end, 10);
/*
* Check for ERANGE, which indicates that the value is too large to fit
* in a 64-bit value.
*/
if (errno == ERANGE) {
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"numeric value is too large"));
return (-1);
}
/*
* If we have a decimal value, then do the computation with floating
* point arithmetic. Otherwise, use standard arithmetic.
*/
if (*end == '.') {
double fval = strtod(value, &end);
if ((shift = str2shift(hdl, end)) == -1)
return (-1);
fval *= pow(2, shift);
/*
* UINT64_MAX is not exactly representable as a double.
* The closest representation is UINT64_MAX + 1, so we
* use a >= comparison instead of > for the bounds check.
*/
if (fval >= (double)UINT64_MAX) {
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"numeric value is too large"));
return (-1);
}
*num = (uint64_t)fval;
} else {
if ((shift = str2shift(hdl, end)) == -1)
return (-1);
/* Check for overflow */
if (shift >= 64 || (*num << shift) >> shift != *num) {
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"numeric value is too large"));
return (-1);
}
*num <<= shift;
}
return (0);
}
/*
* Given a propname=value nvpair to set, parse any numeric properties
* (index, boolean, etc) if they are specified as strings and add the
* resulting nvpair to the returned nvlist.
*
* At the DSL layer, all properties are either 64-bit numbers or strings.
* We want the user to be able to ignore this fact and specify properties
* as native values (numbers, for example) or as strings (to simplify
* command line utilities). This also handles converting index types
* (compression, checksum, etc) from strings to their on-disk index.
*/
int
zprop_parse_value(libzfs_handle_t *hdl, nvpair_t *elem, int prop,
zfs_type_t type, nvlist_t *ret, const char **svalp, uint64_t *ivalp,
const char *errbuf)
{
data_type_t datatype = nvpair_type(elem);
zprop_type_t proptype;
const char *propname;
const char *value;
boolean_t isnone = B_FALSE;
boolean_t isauto = B_FALSE;
int err = 0;
if (type == ZFS_TYPE_POOL) {
proptype = zpool_prop_get_type(prop);
propname = zpool_prop_to_name(prop);
} else if (type == ZFS_TYPE_VDEV) {
proptype = vdev_prop_get_type(prop);
propname = vdev_prop_to_name(prop);
} else {
proptype = zfs_prop_get_type(prop);
propname = zfs_prop_to_name(prop);
}
/*
* Convert any properties to the internal DSL value types.
*/
*svalp = NULL;
*ivalp = 0;
switch (proptype) {
case PROP_TYPE_STRING:
if (datatype != DATA_TYPE_STRING) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a string"), nvpair_name(elem));
goto error;
}
err = nvpair_value_string(elem, svalp);
if (err != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' is invalid"), nvpair_name(elem));
goto error;
}
if (strlen(*svalp) >= ZFS_MAXPROPLEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' is too long"), nvpair_name(elem));
goto error;
}
break;
case PROP_TYPE_NUMBER:
if (datatype == DATA_TYPE_STRING) {
(void) nvpair_value_string(elem, &value);
if (strcmp(value, "none") == 0) {
isnone = B_TRUE;
} else if (strcmp(value, "auto") == 0) {
isauto = B_TRUE;
} else if (zfs_nicestrtonum(hdl, value, ivalp) != 0) {
goto error;
}
} else if (datatype == DATA_TYPE_UINT64) {
(void) nvpair_value_uint64(elem, ivalp);
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a number"), nvpair_name(elem));
goto error;
}
/*
* Quota special: force 'none' and don't allow 0.
*/
if ((type & ZFS_TYPE_DATASET) && *ivalp == 0 && !isnone &&
(prop == ZFS_PROP_QUOTA || prop == ZFS_PROP_REFQUOTA)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"use 'none' to disable quota/refquota"));
goto error;
}
/*
* Special handling for "*_limit=none". In this case it's not
* 0 but UINT64_MAX.
*/
if ((type & ZFS_TYPE_DATASET) && isnone &&
(prop == ZFS_PROP_FILESYSTEM_LIMIT ||
prop == ZFS_PROP_SNAPSHOT_LIMIT)) {
*ivalp = UINT64_MAX;
}
/*
* Special handling for "checksum_*=none". In this case it's not
* 0 but UINT64_MAX.
*/
if ((type & ZFS_TYPE_VDEV) && isnone &&
(prop == VDEV_PROP_CHECKSUM_N ||
prop == VDEV_PROP_CHECKSUM_T ||
prop == VDEV_PROP_IO_N ||
prop == VDEV_PROP_IO_T)) {
*ivalp = UINT64_MAX;
}
/*
* Special handling for setting 'refreservation' to 'auto'. Use
* UINT64_MAX to tell the caller to use zfs_fix_auto_resv().
* 'auto' is only allowed on volumes.
*/
if (isauto) {
switch (prop) {
case ZFS_PROP_REFRESERVATION:
if ((type & ZFS_TYPE_VOLUME) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s=auto' only allowed on "
"volumes"), nvpair_name(elem));
goto error;
}
*ivalp = UINT64_MAX;
break;
default:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'auto' is invalid value for '%s'"),
nvpair_name(elem));
goto error;
}
}
break;
case PROP_TYPE_INDEX:
if (datatype != DATA_TYPE_STRING) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a string"), nvpair_name(elem));
goto error;
}
(void) nvpair_value_string(elem, &value);
if (zprop_string_to_index(prop, value, ivalp, type) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be one of '%s'"), propname,
zprop_values(prop, type));
goto error;
}
break;
default:
abort();
}
/*
* Add the result to our return set of properties.
*/
if (*svalp != NULL) {
if (nvlist_add_string(ret, propname, *svalp) != 0) {
(void) no_memory(hdl);
return (-1);
}
} else {
if (nvlist_add_uint64(ret, propname, *ivalp) != 0) {
(void) no_memory(hdl);
return (-1);
}
}
return (0);
error:
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
return (-1);
}
static int
addlist(libzfs_handle_t *hdl, const char *propname, zprop_list_t **listp,
zfs_type_t type)
{
int prop = zprop_name_to_prop(propname, type);
if (prop != ZPROP_INVAL && !zprop_valid_for_type(prop, type, B_FALSE))
prop = ZPROP_INVAL;
/*
* Return failure if no property table entry was found and this isn't
* a user-defined property.
*/
if (prop == ZPROP_USERPROP && ((type == ZFS_TYPE_POOL &&
!zfs_prop_user(propname) &&
!zpool_prop_feature(propname) &&
!zpool_prop_unsupported(propname)) ||
((type == ZFS_TYPE_DATASET) && !zfs_prop_user(propname) &&
!zfs_prop_userquota(propname) && !zfs_prop_written(propname)) ||
((type == ZFS_TYPE_VDEV) && !vdev_prop_user(propname)))) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid property '%s'"), propname);
return (zfs_error(hdl, EZFS_BADPROP,
dgettext(TEXT_DOMAIN, "bad property list")));
}
zprop_list_t *entry = zfs_alloc(hdl, sizeof (*entry));
entry->pl_prop = prop;
if (prop == ZPROP_USERPROP) {
entry->pl_user_prop = zfs_strdup(hdl, propname);
entry->pl_width = strlen(propname);
} else {
entry->pl_width = zprop_width(prop, &entry->pl_fixed,
type);
}
*listp = entry;
return (0);
}
/*
* Given a comma-separated list of properties, construct a property list
* containing both user-defined and native properties. This function will
* return a NULL list if 'all' is specified, which can later be expanded
* by zprop_expand_list().
*/
int
zprop_get_list(libzfs_handle_t *hdl, char *props, zprop_list_t **listp,
zfs_type_t type)
{
*listp = NULL;
/*
* If 'all' is specified, return a NULL list.
*/
if (strcmp(props, "all") == 0)
return (0);
/*
* If no props were specified, return an error.
*/
if (props[0] == '\0') {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"no properties specified"));
return (zfs_error(hdl, EZFS_BADPROP, dgettext(TEXT_DOMAIN,
"bad property list")));
}
for (char *p; (p = strsep(&props, ",")); )
if (strcmp(p, "space") == 0) {
static const char *const spaceprops[] = {
"name", "avail", "used", "usedbysnapshots",
"usedbydataset", "usedbyrefreservation",
"usedbychildren"
};
for (int i = 0; i < ARRAY_SIZE(spaceprops); i++) {
if (addlist(hdl, spaceprops[i], listp, type))
return (-1);
listp = &(*listp)->pl_next;
}
} else {
if (addlist(hdl, p, listp, type))
return (-1);
listp = &(*listp)->pl_next;
}
return (0);
}
void
zprop_free_list(zprop_list_t *pl)
{
zprop_list_t *next;
while (pl != NULL) {
next = pl->pl_next;
free(pl->pl_user_prop);
free(pl);
pl = next;
}
}
typedef struct expand_data {
zprop_list_t **last;
libzfs_handle_t *hdl;
zfs_type_t type;
} expand_data_t;
static int
zprop_expand_list_cb(int prop, void *cb)
{
zprop_list_t *entry;
expand_data_t *edp = cb;
entry = zfs_alloc(edp->hdl, sizeof (zprop_list_t));
entry->pl_prop = prop;
entry->pl_width = zprop_width(prop, &entry->pl_fixed, edp->type);
entry->pl_all = B_TRUE;
*(edp->last) = entry;
edp->last = &entry->pl_next;
return (ZPROP_CONT);
}
int
zprop_expand_list(libzfs_handle_t *hdl, zprop_list_t **plp, zfs_type_t type)
{
zprop_list_t *entry;
zprop_list_t **last;
expand_data_t exp;
if (*plp == NULL) {
/*
* If this is the very first time we've been called for an 'all'
* specification, expand the list to include all native
* properties.
*/
last = plp;
exp.last = last;
exp.hdl = hdl;
exp.type = type;
if (zprop_iter_common(zprop_expand_list_cb, &exp, B_FALSE,
B_FALSE, type) == ZPROP_INVAL)
return (-1);
/*
* Add 'name' to the beginning of the list, which is handled
* specially.
*/
entry = zfs_alloc(hdl, sizeof (zprop_list_t));
entry->pl_prop = ((type == ZFS_TYPE_POOL) ? ZPOOL_PROP_NAME :
((type == ZFS_TYPE_VDEV) ? VDEV_PROP_NAME : ZFS_PROP_NAME));
entry->pl_width = zprop_width(entry->pl_prop,
&entry->pl_fixed, type);
entry->pl_all = B_TRUE;
entry->pl_next = *plp;
*plp = entry;
}
return (0);
}
int
zprop_iter(zprop_func func, void *cb, boolean_t show_all, boolean_t ordered,
zfs_type_t type)
{
return (zprop_iter_common(func, cb, show_all, ordered, type));
}
const char *
zfs_version_userland(void)
{
return (ZFS_META_ALIAS);
}
/*
* Prints both zfs userland and kernel versions
* Returns 0 on success, and -1 on error
*/
int
zfs_version_print(void)
{
(void) puts(ZFS_META_ALIAS);
char *kver = zfs_version_kernel();
if (kver == NULL) {
fprintf(stderr, "zfs_version_kernel() failed: %s\n",
strerror(errno));
return (-1);
}
(void) printf("zfs-kmod-%s\n", kver);
free(kver);
return (0);
}
/*
* Return 1 if the user requested ANSI color output, and our terminal supports
* it. Return 0 for no color.
*/
int
use_color(void)
{
static int use_color = -1;
char *term;
/*
* Optimization:
*
* For each zpool invocation, we do a single check to see if we should
* be using color or not, and cache that value for the lifetime of the
* the zpool command. That makes it cheap to call use_color() when
* we're printing with color. We assume that the settings are not going
* to change during the invocation of a zpool command (the user isn't
* going to change the ZFS_COLOR value while zpool is running, for
* example).
*/
if (use_color != -1) {
/*
* We've already figured out if we should be using color or
* not. Return the cached value.
*/
return (use_color);
}
term = getenv("TERM");
/*
* The user sets the ZFS_COLOR env var set to enable zpool ANSI color
* output. However if NO_COLOR is set (https://no-color.org/) then
* don't use it. Also, don't use color if terminal doesn't support
* it.
*/
if (libzfs_envvar_is_set("ZFS_COLOR") &&
!libzfs_envvar_is_set("NO_COLOR") &&
isatty(STDOUT_FILENO) && term && strcmp("dumb", term) != 0 &&
strcmp("unknown", term) != 0) {
/* Color supported */
use_color = 1;
} else {
use_color = 0;
}
return (use_color);
}
/*
* The functions color_start() and color_end() are used for when you want
* to colorize a block of text.
*
* For example:
* color_start(ANSI_RED)
* printf("hello");
* printf("world");
* color_end();
*/
void
color_start(const char *color)
{
if (color && use_color()) {
fputs(color, stdout);
fflush(stdout);
}
}
void
color_end(void)
{
if (use_color()) {
fputs(ANSI_RESET, stdout);
fflush(stdout);
}
}
/*
* printf() with a color. If color is NULL, then do a normal printf.
*/
int
printf_color(const char *color, const char *format, ...)
{
va_list aptr;
int rc;
if (color)
color_start(color);
va_start(aptr, format);
rc = vprintf(format, aptr);
va_end(aptr);
if (color)
color_end();
return (rc);
}
/* PATH + 5 env vars + a NULL entry = 7 */
#define ZPOOL_VDEV_SCRIPT_ENV_COUNT 7
/*
* There's a few places where ZFS will call external scripts (like the script
* in zpool.d/ and `zfs_prepare_disk`). These scripts are called with a
* reduced $PATH, and some vdev specific environment vars set. This function
* will allocate an populate the environment variable array that is passed to
* these scripts. The user must free the arrays with zpool_vdev_free_env() when
* they are done.
*
* The following env vars will be set (but value could be blank):
*
* POOL_NAME
* VDEV_PATH
* VDEV_UPATH
* VDEV_ENC_SYSFS_PATH
*
* In addition, you can set an optional environment variable named 'opt_key'
* to 'opt_val' if you want.
*
* Returns allocated env[] array on success, NULL otherwise.
*/
char **
zpool_vdev_script_alloc_env(const char *pool_name,
const char *vdev_path, const char *vdev_upath,
const char *vdev_enc_sysfs_path, const char *opt_key, const char *opt_val)
{
char **env = NULL;
int rc;
env = calloc(ZPOOL_VDEV_SCRIPT_ENV_COUNT, sizeof (*env));
if (!env)
return (NULL);
env[0] = strdup("PATH=/bin:/sbin:/usr/bin:/usr/sbin");
if (!env[0])
goto error;
/* Setup our custom environment variables */
rc = asprintf(&env[1], "POOL_NAME=%s", pool_name ? pool_name : "");
if (rc == -1) {
env[1] = NULL;
goto error;
}
rc = asprintf(&env[2], "VDEV_PATH=%s", vdev_path ? vdev_path : "");
if (rc == -1) {
env[2] = NULL;
goto error;
}
rc = asprintf(&env[3], "VDEV_UPATH=%s", vdev_upath ? vdev_upath : "");
if (rc == -1) {
env[3] = NULL;
goto error;
}
rc = asprintf(&env[4], "VDEV_ENC_SYSFS_PATH=%s",
vdev_enc_sysfs_path ? vdev_enc_sysfs_path : "");
if (rc == -1) {
env[4] = NULL;
goto error;
}
if (opt_key != NULL) {
rc = asprintf(&env[5], "%s=%s", opt_key,
opt_val ? opt_val : "");
if (rc == -1) {
env[5] = NULL;
goto error;
}
}
return (env);
error:
for (int i = 0; i < ZPOOL_VDEV_SCRIPT_ENV_COUNT; i++)
free(env[i]);
free(env);
return (NULL);
}
/*
* Free the env[] array that was allocated by zpool_vdev_script_alloc_env().
*/
void
zpool_vdev_script_free_env(char **env)
{
for (int i = 0; i < ZPOOL_VDEV_SCRIPT_ENV_COUNT; i++)
free(env[i]);
free(env);
}
/*
* Prepare a disk by (optionally) running a program before labeling the disk.
* This can be useful for installing disk firmware or doing some pre-flight
* checks on the disk before it becomes part of the pool. The program run is
* located at ZFSEXECDIR/zfs_prepare_disk
* (E.x: /usr/local/libexec/zfs/zfs_prepare_disk).
*
* Return 0 on success, non-zero on failure.
*/
int
zpool_prepare_disk(zpool_handle_t *zhp, nvlist_t *vdev_nv,
const char *prepare_str, char **lines[], int *lines_cnt)
{
const char *script_path = ZFSEXECDIR "/zfs_prepare_disk";
const char *pool_name;
int rc = 0;
/* Path to script and a NULL entry */
char *argv[2] = {(char *)script_path};
char **env = NULL;
const char *path = NULL, *enc_sysfs_path = NULL;
char *upath;
*lines_cnt = 0;
if (access(script_path, X_OK) != 0) {
/* No script, nothing to do */
return (0);
}
(void) nvlist_lookup_string(vdev_nv, ZPOOL_CONFIG_PATH, &path);
(void) nvlist_lookup_string(vdev_nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH,
&enc_sysfs_path);
upath = zfs_get_underlying_path(path);
pool_name = zhp ? zpool_get_name(zhp) : NULL;
env = zpool_vdev_script_alloc_env(pool_name, path, upath,
enc_sysfs_path, "VDEV_PREPARE", prepare_str);
free(upath);
if (env == NULL) {
return (ENOMEM);
}
rc = libzfs_run_process_get_stdout(script_path, argv, env, lines,
lines_cnt);
zpool_vdev_script_free_env(env);
return (rc);
}
/*
* Optionally run a script and then label a disk. The script can be used to
* prepare a disk for inclusion into the pool. For example, it might update
* the disk's firmware or check its health.
*
* The 'name' provided is the short name, stripped of any leading
* /dev path, and is passed to zpool_label_disk. vdev_nv is the nvlist for
* the vdev. prepare_str is a string that gets passed as the VDEV_PREPARE
* env variable to the script.
*
* The following env vars are passed to the script:
*
* POOL_NAME: The pool name (blank during zpool create)
* VDEV_PREPARE: Reason why the disk is being prepared for inclusion:
* "create", "add", "replace", or "autoreplace"
* VDEV_PATH: Path to the disk
* VDEV_UPATH: One of the 'underlying paths' to the disk. This is
* useful for DM devices.
* VDEV_ENC_SYSFS_PATH: Path to the disk's enclosure sysfs path, if available.
*
* Note, some of these values can be blank.
*
* Return 0 on success, non-zero otherwise.
*/
int
zpool_prepare_and_label_disk(libzfs_handle_t *hdl, zpool_handle_t *zhp,
const char *name, nvlist_t *vdev_nv, const char *prepare_str,
char **lines[], int *lines_cnt)
{
int rc;
char vdev_path[MAXPATHLEN];
(void) snprintf(vdev_path, sizeof (vdev_path), "%s/%s", DISK_ROOT,
name);
/* zhp will be NULL when creating a pool */
rc = zpool_prepare_disk(zhp, vdev_nv, prepare_str, lines, lines_cnt);
if (rc != 0)
return (rc);
rc = zpool_label_disk(hdl, zhp, name);
return (rc);
}
diff --git a/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.abi b/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.abi
index f2087186aa44..5b95c8f779db 100644
--- a/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.abi
+++ b/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.abi
@@ -1,2920 +1,2930 @@
<abi-corpus version='2.0' architecture='elf-amd-x86_64' soname='libzfs_core.so.3'>
<elf-needed>
<dependency name='libnvpair.so.3'/>
<dependency name='libc.so.6'/>
<dependency name='ld-linux-x86-64.so.2'/>
</elf-needed>
<elf-function-symbols>
<elf-symbol name='_sol_getmntent' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_char' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_char_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_int_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_long' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_long_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_ptr_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_short' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_short_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_clear_long_excl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_set_long_excl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_char' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_char_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_int_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_long' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_long_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_ptr_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_short' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_short_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='get_system_hostid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getexecname' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getextmntent' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getmntany' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getzoneid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libspl_assertf' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libspl_set_assert_ok' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_core_fini' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_core_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_head' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_after' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_before' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_head' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_is_empty' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_link_active' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_link_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_link_replace' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_move_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_next' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_prev' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_remove_head' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_remove_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_bookmark' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_change_key' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_channel_program' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_channel_program_nosync' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_clone' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_destroy_bookmarks' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_destroy_snaps' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_exists' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_get_bookmark_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_get_bookmarks' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_get_bootenv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_get_holds' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_get_vdev_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_hold' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_initialize' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_ioctl_fd' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_load_key' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_pool_checkpoint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_pool_checkpoint_discard' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_promote' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_receive' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_receive_one' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_receive_resumable' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_receive_with_cmdprops' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_receive_with_header' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_receive_with_heal' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_redact' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_release' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_rename' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_reopen' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_rollback' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_rollback_to' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_scrub' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_send' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_send_redacted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_send_resume' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_send_resume_redacted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_send_space' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_send_space_resume_redacted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_send_wrapper' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_set_bootenv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_set_vdev_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_snaprange_space' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_snapshot' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_sync' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_trim' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_unload_key' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_wait' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_wait_fs' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='lzc_wait_tag' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_consumer' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_enter' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_exit' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_producer' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_sync' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='mkdirp' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='print_timestamp' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='spl_pagesize' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='strlcat' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='strlcpy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
</elf-function-symbols>
<abi-instr address-size='64' path='lib/libspl/assert.c' language='LANG_C99'>
<class-decl name='__va_list_tag' size-in-bits='192' is-struct='yes' visibility='default' id='d5027220'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='gp_offset' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='fp_offset' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='overflow_arg_area' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='reg_save_area' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='d5027220' size-in-bits='64' id='b7f2d5e6'/>
<function-decl name='abort' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__vfprintf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e75a27e9'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='9d26089a'/>
<parameter type-id='b7f2d5e6'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libspl_set_assert_ok' mangled-name='libspl_set_assert_ok' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libspl_set_assert_ok'>
<parameter type-id='c19b74c3' name='val'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/atomic.c' language='LANG_C99'>
<typedef-decl name='ulong_t' type-id='7359adad' id='ee1f298e'/>
<typedef-decl name='int8_t' type-id='2171a512' id='ee31ee44'/>
<typedef-decl name='uint16_t' type-id='253c2d2a' id='149c6638'/>
<typedef-decl name='__int8_t' type-id='28577a57' id='2171a512'/>
<typedef-decl name='__uint16_t' type-id='8efea9e5' id='253c2d2a'/>
<qualified-type-def type-id='149c6638' volatile='yes' id='5120c5f7'/>
<pointer-type-def type-id='5120c5f7' size-in-bits='64' id='93977ae7'/>
<qualified-type-def type-id='8f92235e' volatile='yes' id='430e0681'/>
<pointer-type-def type-id='430e0681' size-in-bits='64' id='3a147f31'/>
<qualified-type-def type-id='b96825af' volatile='yes' id='84ff7d66'/>
<pointer-type-def type-id='84ff7d66' size-in-bits='64' id='aa323ea4'/>
<qualified-type-def type-id='ee1f298e' volatile='yes' id='6f7e09cb'/>
<pointer-type-def type-id='6f7e09cb' size-in-bits='64' id='64698d33'/>
<qualified-type-def type-id='48b5725f' volatile='yes' id='b0b3cbf9'/>
<pointer-type-def type-id='b0b3cbf9' size-in-bits='64' id='fe09dd29'/>
<function-decl name='atomic_inc_8' mangled-name='atomic_inc_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_8'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_16' mangled-name='atomic_inc_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_16'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_32' mangled-name='atomic_inc_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_32'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_ulong' mangled-name='atomic_inc_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_ulong'>
<parameter type-id='64698d33' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_8' mangled-name='atomic_dec_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_8'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_16' mangled-name='atomic_dec_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_16'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_32' mangled-name='atomic_dec_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_32'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_ulong' mangled-name='atomic_dec_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_ulong'>
<parameter type-id='64698d33' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_ptr' mangled-name='atomic_add_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_8' mangled-name='atomic_add_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_16' mangled-name='atomic_add_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_32' mangled-name='atomic_add_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_ptr' mangled-name='atomic_sub_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_8' mangled-name='atomic_sub_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_16' mangled-name='atomic_sub_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_32' mangled-name='atomic_sub_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_8' mangled-name='atomic_or_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_16' mangled-name='atomic_or_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_32' mangled-name='atomic_or_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_ulong' mangled-name='atomic_or_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_8' mangled-name='atomic_and_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_16' mangled-name='atomic_and_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_32' mangled-name='atomic_and_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_ulong' mangled-name='atomic_and_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_8_nv' mangled-name='atomic_inc_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_inc_16_nv' mangled-name='atomic_inc_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_inc_32_nv' mangled-name='atomic_inc_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_inc_ulong_nv' mangled-name='atomic_inc_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_dec_8_nv' mangled-name='atomic_dec_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_dec_16_nv' mangled-name='atomic_dec_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_dec_32_nv' mangled-name='atomic_dec_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_dec_ulong_nv' mangled-name='atomic_dec_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_add_ptr_nv' mangled-name='atomic_add_ptr_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_ptr_nv'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_add_8_nv' mangled-name='atomic_add_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_add_16_nv' mangled-name='atomic_add_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_add_32_nv' mangled-name='atomic_add_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_add_long_nv' mangled-name='atomic_add_long_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_long_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='bd54fe1a' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_sub_ptr_nv' mangled-name='atomic_sub_ptr_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_ptr_nv'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_sub_8_nv' mangled-name='atomic_sub_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_sub_16_nv' mangled-name='atomic_sub_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_sub_32_nv' mangled-name='atomic_sub_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_sub_long_nv' mangled-name='atomic_sub_long_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_long_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='bd54fe1a' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_or_8_nv' mangled-name='atomic_or_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_or_16_nv' mangled-name='atomic_or_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_or_32_nv' mangled-name='atomic_or_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_or_ulong_nv' mangled-name='atomic_or_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_and_8_nv' mangled-name='atomic_and_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_and_16_nv' mangled-name='atomic_and_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_and_32_nv' mangled-name='atomic_and_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_and_ulong_nv' mangled-name='atomic_and_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_cas_ptr' mangled-name='atomic_cas_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='eaa32e2f' name='exp'/>
<parameter type-id='eaa32e2f' name='des'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_cas_8' mangled-name='atomic_cas_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='exp'/>
<parameter type-id='b96825af' name='des'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_cas_16' mangled-name='atomic_cas_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='exp'/>
<parameter type-id='149c6638' name='des'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_cas_32' mangled-name='atomic_cas_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='exp'/>
<parameter type-id='8f92235e' name='des'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_cas_ulong' mangled-name='atomic_cas_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='exp'/>
<parameter type-id='ee1f298e' name='des'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_swap_8' mangled-name='atomic_swap_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_swap_16' mangled-name='atomic_swap_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_swap_32' mangled-name='atomic_swap_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_swap_ulong' mangled-name='atomic_swap_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_swap_ptr' mangled-name='atomic_swap_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='eaa32e2f' name='bits'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_set_long_excl' mangled-name='atomic_set_long_excl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_set_long_excl'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='3502e3ff' name='value'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='atomic_clear_long_excl' mangled-name='atomic_clear_long_excl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_clear_long_excl'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='3502e3ff' name='value'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='membar_enter' mangled-name='membar_enter' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='membar_enter'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='membar_producer' mangled-name='membar_producer' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='membar_producer'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='membar_consumer' mangled-name='membar_consumer' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='membar_consumer'>
<return type-id='48b5725f'/>
</function-decl>
<type-decl name='unsigned long int' size-in-bits='64' id='7359adad'/>
<type-decl name='unsigned short int' size-in-bits='16' id='8efea9e5'/>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/getexecname.c' language='LANG_C99'>
<function-decl name='getexecname' mangled-name='getexecname' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getexecname'>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='getexecname_impl' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='26a90f95'/>
<return type-id='79a0948f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/list.c' language='LANG_C99'>
<typedef-decl name='list_node_t' type-id='b0b5e45e' id='b21843b2'/>
<typedef-decl name='list_t' type-id='e824dae9' id='0899125f'/>
<class-decl name='list_node' size-in-bits='128' is-struct='yes' visibility='default' id='b0b5e45e'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='next' type-id='b03eadb4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='prev' type-id='b03eadb4' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='list' size-in-bits='256' is-struct='yes' visibility='default' id='e824dae9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='list_size' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='list_offset' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='list_head' type-id='b0b5e45e' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='b0b5e45e' size-in-bits='64' id='b03eadb4'/>
<pointer-type-def type-id='b21843b2' size-in-bits='64' id='ccc38265'/>
<pointer-type-def type-id='0899125f' size-in-bits='64' id='352ec160'/>
<function-decl name='list_create' mangled-name='list_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_create'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='b59d7dce' name='size'/>
<parameter type-id='b59d7dce' name='offset'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_destroy' mangled-name='list_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_destroy'>
<parameter type-id='352ec160' name='list'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_after' mangled-name='list_insert_after' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_after'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<parameter type-id='eaa32e2f' name='nobject'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_before' mangled-name='list_insert_before' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_before'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<parameter type-id='eaa32e2f' name='nobject'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_head' mangled-name='list_insert_head' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_head'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_tail' mangled-name='list_insert_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_tail'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_remove' mangled-name='list_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_remove'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_remove_head' mangled-name='list_remove_head' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_remove_head'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_remove_tail' mangled-name='list_remove_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_remove_tail'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_head' mangled-name='list_head' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_head'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_tail' mangled-name='list_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_tail'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_next' mangled-name='list_next' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_next'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_prev' mangled-name='list_prev' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_prev'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_move_tail' mangled-name='list_move_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_move_tail'>
<parameter type-id='352ec160' name='dst'/>
<parameter type-id='352ec160' name='src'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_link_replace' mangled-name='list_link_replace' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_link_replace'>
<parameter type-id='ccc38265' name='lold'/>
<parameter type-id='ccc38265' name='lnew'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_link_init' mangled-name='list_link_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_link_init'>
<parameter type-id='ccc38265' name='ln'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_link_active' mangled-name='list_link_active' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_link_active'>
<parameter type-id='ccc38265' name='ln'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='list_is_empty' mangled-name='list_is_empty' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_is_empty'>
<parameter type-id='352ec160' name='list'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/mkdirp.c' language='LANG_C99'>
<typedef-decl name='mode_t' type-id='e1c52942' id='d50d396c'/>
<typedef-decl name='wchar_t' type-id='95e97e5e' id='928221d2'/>
<qualified-type-def type-id='26a90f95' restrict='yes' id='266fe297'/>
<qualified-type-def type-id='928221d2' const='yes' id='effb3702'/>
<pointer-type-def type-id='effb3702' size-in-bits='64' id='f077d3f8'/>
<qualified-type-def type-id='f077d3f8' restrict='yes' id='598aab80'/>
<pointer-type-def type-id='928221d2' size-in-bits='64' id='323d93c1'/>
<qualified-type-def type-id='323d93c1' restrict='yes' id='f1358bc3'/>
<function-decl name='calloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='strdup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strlen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='access' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__mbstowcs_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f1358bc3'/>
<parameter type-id='9d26089a'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='__wcstombs_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='266fe297'/>
<parameter type-id='598aab80'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='mkdir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='e1c52942'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='mkdirp' mangled-name='mkdirp' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='mkdirp'>
<parameter type-id='80f4b756' name='d'/>
<parameter type-id='d50d396c' name='mode'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/os/linux/getexecname.c' language='LANG_C99'>
<function-decl name='__readlink_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='266fe297'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='79a0948f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/os/linux/gethostid.c' language='LANG_C99'>
<type-decl name='long long unsigned int' size-in-bits='64' id='3a47d82b'/>
<pointer-type-def type-id='26a90f95' size-in-bits='64' id='9b23c9ad'/>
<qualified-type-def type-id='9b23c9ad' restrict='yes' id='8c85230f'/>
<function-decl name='strtoull' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<return type-id='3a47d82b'/>
</function-decl>
<function-decl name='get_system_hostid' mangled-name='get_system_hostid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='get_system_hostid'>
<return type-id='7359adad'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/os/linux/getmntany.c' language='LANG_C99'>
<class-decl name='mnttab' size-in-bits='256' is-struct='yes' visibility='default' id='1b055409'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='mnt_special' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='mnt_mountp' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='mnt_fstype' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='mnt_mntopts' type-id='26a90f95' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='extmnttab' size-in-bits='320' is-struct='yes' visibility='default' id='0c544dc0'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='mnt_special' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='mnt_mountp' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='mnt_fstype' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='mnt_mntopts' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='mnt_major' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='mnt_minor' type-id='3502e3ff' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='mntent' size-in-bits='320' is-struct='yes' visibility='default' id='56fe4a37'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='mnt_fsname' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='mnt_dir' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='mnt_type' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='mnt_opts' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='mnt_freq' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='mnt_passno' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='stat64' size-in-bits='1152' is-struct='yes' visibility='default' id='0bbec9cd'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='st_dev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='st_ino' type-id='71288a47' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='st_nlink' type-id='80f0b9df' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='st_mode' type-id='e1c52942' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='st_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='st_gid' type-id='d94ec6d9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='__pad0' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='st_rdev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='st_size' type-id='79989e9c' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='st_blksize' type-id='d3f10a7f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='st_blocks' type-id='4e711bf1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='st_atim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='st_mtim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='st_ctim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='__glibc_reserved' type-id='083f8d58' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__ino64_t' type-id='7359adad' id='71288a47'/>
<typedef-decl name='__blkcnt64_t' type-id='bd54fe1a' id='4e711bf1'/>
<pointer-type-def type-id='0c544dc0' size-in-bits='64' id='394fc496'/>
<pointer-type-def type-id='56fe4a37' size-in-bits='64' id='b6b61d2f'/>
<qualified-type-def type-id='b6b61d2f' restrict='yes' id='3cad23cd'/>
<pointer-type-def type-id='1b055409' size-in-bits='64' id='9d424d31'/>
<pointer-type-def type-id='0bbec9cd' size-in-bits='64' id='62f7a03d'/>
<qualified-type-def type-id='62f7a03d' restrict='yes' id='f1cadedf'/>
<function-decl name='getmntent_r' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e75a27e9'/>
<parameter type-id='3cad23cd'/>
<parameter type-id='266fe297'/>
<parameter type-id='95e97e5e'/>
<return type-id='b6b61d2f'/>
</function-decl>
<function-decl name='feof' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='strcmp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='strerror' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='__fprintf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e75a27e9'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='9d26089a'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='stat64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='f1cadedf'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getmntany' mangled-name='getmntany' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getmntany'>
<parameter type-id='822cd80b' name='fp'/>
<parameter type-id='9d424d31' name='mgetp'/>
<parameter type-id='9d424d31' name='mrefp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='_sol_getmntent' mangled-name='_sol_getmntent' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='_sol_getmntent'>
<parameter type-id='822cd80b' name='fp'/>
<parameter type-id='9d424d31' name='mgetp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getextmntent' mangled-name='getextmntent' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getextmntent'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='394fc496' name='entry'/>
<parameter type-id='62f7a03d' name='statbuf'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/os/linux/zone.c' language='LANG_C99'>
<typedef-decl name='zoneid_t' type-id='3502e3ff' id='4da03624'/>
<function-decl name='strtoul' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<return type-id='7359adad'/>
</function-decl>
<function-decl name='getzoneid' mangled-name='getzoneid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getzoneid'>
<return type-id='4da03624'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/page.c' language='LANG_C99'>
<function-decl name='sysconf' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='bd54fe1a'/>
</function-decl>
<function-decl name='spl_pagesize' mangled-name='spl_pagesize' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='spl_pagesize'>
<return type-id='b59d7dce'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/strlcat.c' language='LANG_C99'>
<function-decl name='strlcat' mangled-name='strlcat' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='strlcat'>
<parameter type-id='26a90f95' name='dst'/>
<parameter type-id='80f4b756' name='src'/>
<parameter type-id='b59d7dce' name='dstsize'/>
<return type-id='b59d7dce'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/timestamp.c' language='LANG_C99'>
<typedef-decl name='nl_item' type-id='95e97e5e' id='03b79a94'/>
<class-decl name='tm' size-in-bits='448' is-struct='yes' visibility='default' id='dddf6ca2'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tm_sec' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='tm_min' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tm_hour' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='tm_mday' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='tm_mon' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='tm_year' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='tm_wday' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='tm_yday' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='tm_isdst' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='tm_gmtoff' type-id='bd54fe1a' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='tm_zone' type-id='80f4b756' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='time_t' type-id='65eda9c0' id='c9d12d66'/>
<qualified-type-def type-id='c9d12d66' const='yes' id='588b3216'/>
<pointer-type-def type-id='588b3216' size-in-bits='64' id='9f201474'/>
<qualified-type-def type-id='9f201474' restrict='yes' id='d6e2847c'/>
<qualified-type-def type-id='dddf6ca2' const='yes' id='e824a34f'/>
<pointer-type-def type-id='e824a34f' size-in-bits='64' id='d6ad37ff'/>
<qualified-type-def type-id='d6ad37ff' restrict='yes' id='f8c6051d'/>
<pointer-type-def type-id='c9d12d66' size-in-bits='64' id='b2eb2c3f'/>
<pointer-type-def type-id='dddf6ca2' size-in-bits='64' id='d915a820'/>
<qualified-type-def type-id='d915a820' restrict='yes' id='f099ad08'/>
<function-decl name='nl_langinfo' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='03b79a94'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='time' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b2eb2c3f'/>
<return type-id='c9d12d66'/>
</function-decl>
<function-decl name='strftime' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='266fe297'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='9d26089a'/>
<parameter type-id='f8c6051d'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='localtime_r' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='d6e2847c'/>
<parameter type-id='f099ad08'/>
<return type-id='d915a820'/>
</function-decl>
<function-decl name='__printf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='print_timestamp' mangled-name='print_timestamp' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='print_timestamp'>
<parameter type-id='3502e3ff' name='timestamp_fmt'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs_core/libzfs_core.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='03085adc' size-in-bits='192' id='083f8d58'>
<subrange length='3' type-id='7359adad' id='56f209d2'/>
</array-type-def>
<type-decl name='char' size-in-bits='8' id='a84c031d'/>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='8' id='89feb1ec'>
<subrange length='1' type-id='7359adad' id='52f813b4'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='160' id='664ac0b7'>
<subrange length='20' type-id='7359adad' id='fdca39cf'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='2048' id='d1617432'>
<subrange length='256' type-id='7359adad' id='36e5b9fa'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='32768' id='d16c6df4'>
<subrange length='4096' type-id='7359adad' id='bc1b5ddc'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='320' id='36c46961'>
<subrange length='40' type-id='7359adad' id='8f80b239'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='448' id='6093ff7c'>
<subrange length='56' type-id='7359adad' id='f8137894'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='65536' id='163f6aa5'>
<subrange length='8192' type-id='7359adad' id='c88f397d'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='infinite' id='e84913bd'>
<subrange length='infinite' type-id='7359adad' id='031f2035'/>
</array-type-def>
<class-decl name='_IO_codecvt' is-struct='yes' visibility='default' is-declaration-only='yes' id='a4036571'/>
<class-decl name='_IO_marker' is-struct='yes' visibility='default' is-declaration-only='yes' id='010ae0b9'/>
<class-decl name='_IO_wide_data' is-struct='yes' visibility='default' is-declaration-only='yes' id='79bd3751'/>
<type-decl name='int' size-in-bits='32' id='95e97e5e'/>
<type-decl name='long int' size-in-bits='64' id='bd54fe1a'/>
<type-decl name='short int' size-in-bits='16' id='a2185560'/>
<type-decl name='signed char' size-in-bits='8' id='28577a57'/>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='128' id='c1c22e6c'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='2176' id='8c2bcad1'>
<subrange length='34' type-id='7359adad' id='6a6a7e00'/>
</array-type-def>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='256' id='85c64d26'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='96' id='fa8ef949'>
<subrange length='12' type-id='7359adad' id='84827bdc'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='128' id='fa9986a5'>
<subrange length='16' type-id='7359adad' id='848d0938'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='24' id='d3490169'>
<subrange length='3' type-id='7359adad' id='56f209d2'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='40' id='0f4ddd0b'>
<subrange length='5' type-id='7359adad' id='53010e10'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='48' id='0f562bd0'>
<subrange length='6' type-id='7359adad' id='52fa524b'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='64' id='13339fda'>
<subrange length='8' type-id='7359adad' id='56e0c0b1'/>
</array-type-def>
<type-decl name='unnamed-enum-underlying-type-32' is-anonymous='yes' size-in-bits='32' alignment-in-bits='32' id='9cac1fee'/>
<type-decl name='unsigned char' size-in-bits='8' id='002ac4a6'/>
<type-decl name='unsigned int' size-in-bits='32' id='f0981eeb'/>
<type-decl name='unsigned long int' size-in-bits='64' id='7359adad'/>
<type-decl name='unsigned short int' size-in-bits='16' id='8efea9e5'/>
<type-decl name='variadic parameter type' id='2c1145c5'/>
<type-decl name='void' id='48b5725f'/>
<enum-decl name='lzc_dataset_type' id='bc9887f1'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='LZC_DATSET_TYPE_ZFS' value='2'/>
<enumerator name='LZC_DATSET_TYPE_ZVOL' value='3'/>
</enum-decl>
<enum-decl name='lzc_send_flags' id='bfbd3c8e'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='LZC_SEND_FLAG_EMBED_DATA' value='1'/>
<enumerator name='LZC_SEND_FLAG_LARGE_BLOCK' value='2'/>
<enumerator name='LZC_SEND_FLAG_COMPRESS' value='4'/>
<enumerator name='LZC_SEND_FLAG_RAW' value='8'/>
<enumerator name='LZC_SEND_FLAG_SAVED' value='16'/>
</enum-decl>
<class-decl name='ddt_key' size-in-bits='320' is-struct='yes' visibility='default' id='e0a4a1cb'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='ddk_cksum' type-id='39730d0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='ddk_prop' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='ddt_key_t' type-id='e0a4a1cb' id='67f6d2cf'/>
<enum-decl name='dmu_object_type' id='04b3b0b9'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DMU_OT_NONE' value='0'/>
<enumerator name='DMU_OT_OBJECT_DIRECTORY' value='1'/>
<enumerator name='DMU_OT_OBJECT_ARRAY' value='2'/>
<enumerator name='DMU_OT_PACKED_NVLIST' value='3'/>
<enumerator name='DMU_OT_PACKED_NVLIST_SIZE' value='4'/>
<enumerator name='DMU_OT_BPOBJ' value='5'/>
<enumerator name='DMU_OT_BPOBJ_HDR' value='6'/>
<enumerator name='DMU_OT_SPACE_MAP_HEADER' value='7'/>
<enumerator name='DMU_OT_SPACE_MAP' value='8'/>
<enumerator name='DMU_OT_INTENT_LOG' value='9'/>
<enumerator name='DMU_OT_DNODE' value='10'/>
<enumerator name='DMU_OT_OBJSET' value='11'/>
<enumerator name='DMU_OT_DSL_DIR' value='12'/>
<enumerator name='DMU_OT_DSL_DIR_CHILD_MAP' value='13'/>
<enumerator name='DMU_OT_DSL_DS_SNAP_MAP' value='14'/>
<enumerator name='DMU_OT_DSL_PROPS' value='15'/>
<enumerator name='DMU_OT_DSL_DATASET' value='16'/>
<enumerator name='DMU_OT_ZNODE' value='17'/>
<enumerator name='DMU_OT_OLDACL' value='18'/>
<enumerator name='DMU_OT_PLAIN_FILE_CONTENTS' value='19'/>
<enumerator name='DMU_OT_DIRECTORY_CONTENTS' value='20'/>
<enumerator name='DMU_OT_MASTER_NODE' value='21'/>
<enumerator name='DMU_OT_UNLINKED_SET' value='22'/>
<enumerator name='DMU_OT_ZVOL' value='23'/>
<enumerator name='DMU_OT_ZVOL_PROP' value='24'/>
<enumerator name='DMU_OT_PLAIN_OTHER' value='25'/>
<enumerator name='DMU_OT_UINT64_OTHER' value='26'/>
<enumerator name='DMU_OT_ZAP_OTHER' value='27'/>
<enumerator name='DMU_OT_ERROR_LOG' value='28'/>
<enumerator name='DMU_OT_SPA_HISTORY' value='29'/>
<enumerator name='DMU_OT_SPA_HISTORY_OFFSETS' value='30'/>
<enumerator name='DMU_OT_POOL_PROPS' value='31'/>
<enumerator name='DMU_OT_DSL_PERMS' value='32'/>
<enumerator name='DMU_OT_ACL' value='33'/>
<enumerator name='DMU_OT_SYSACL' value='34'/>
<enumerator name='DMU_OT_FUID' value='35'/>
<enumerator name='DMU_OT_FUID_SIZE' value='36'/>
<enumerator name='DMU_OT_NEXT_CLONES' value='37'/>
<enumerator name='DMU_OT_SCAN_QUEUE' value='38'/>
<enumerator name='DMU_OT_USERGROUP_USED' value='39'/>
<enumerator name='DMU_OT_USERGROUP_QUOTA' value='40'/>
<enumerator name='DMU_OT_USERREFS' value='41'/>
<enumerator name='DMU_OT_DDT_ZAP' value='42'/>
<enumerator name='DMU_OT_DDT_STATS' value='43'/>
<enumerator name='DMU_OT_SA' value='44'/>
<enumerator name='DMU_OT_SA_MASTER_NODE' value='45'/>
<enumerator name='DMU_OT_SA_ATTR_REGISTRATION' value='46'/>
<enumerator name='DMU_OT_SA_ATTR_LAYOUTS' value='47'/>
<enumerator name='DMU_OT_SCAN_XLATE' value='48'/>
<enumerator name='DMU_OT_DEDUP' value='49'/>
<enumerator name='DMU_OT_DEADLIST' value='50'/>
<enumerator name='DMU_OT_DEADLIST_HDR' value='51'/>
<enumerator name='DMU_OT_DSL_CLONES' value='52'/>
<enumerator name='DMU_OT_BPOBJ_SUBOBJ' value='53'/>
<enumerator name='DMU_OT_NUMTYPES' value='54'/>
<enumerator name='DMU_OTN_UINT8_DATA' value='128'/>
<enumerator name='DMU_OTN_UINT8_METADATA' value='192'/>
<enumerator name='DMU_OTN_UINT16_DATA' value='129'/>
<enumerator name='DMU_OTN_UINT16_METADATA' value='193'/>
<enumerator name='DMU_OTN_UINT32_DATA' value='130'/>
<enumerator name='DMU_OTN_UINT32_METADATA' value='194'/>
<enumerator name='DMU_OTN_UINT64_DATA' value='131'/>
<enumerator name='DMU_OTN_UINT64_METADATA' value='195'/>
<enumerator name='DMU_OTN_ZAP_DATA' value='132'/>
<enumerator name='DMU_OTN_ZAP_METADATA' value='196'/>
<enumerator name='DMU_OTN_UINT8_ENC_DATA' value='160'/>
<enumerator name='DMU_OTN_UINT8_ENC_METADATA' value='224'/>
<enumerator name='DMU_OTN_UINT16_ENC_DATA' value='161'/>
<enumerator name='DMU_OTN_UINT16_ENC_METADATA' value='225'/>
<enumerator name='DMU_OTN_UINT32_ENC_DATA' value='162'/>
<enumerator name='DMU_OTN_UINT32_ENC_METADATA' value='226'/>
<enumerator name='DMU_OTN_UINT64_ENC_DATA' value='163'/>
<enumerator name='DMU_OTN_UINT64_ENC_METADATA' value='227'/>
<enumerator name='DMU_OTN_ZAP_ENC_DATA' value='164'/>
<enumerator name='DMU_OTN_ZAP_ENC_METADATA' value='228'/>
</enum-decl>
<typedef-decl name='dmu_object_type_t' type-id='04b3b0b9' id='5c9d8906'/>
<class-decl name='dmu_objset_stats' size-in-bits='2304' is-struct='yes' visibility='default' id='098f0221'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='dds_num_clones' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='dds_creation_txg' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='dds_guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='dds_type' type-id='230f1e16' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='dds_is_snapshot' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='232'>
<var-decl name='dds_inconsistent' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='240'>
<var-decl name='dds_redacted' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='248'>
<var-decl name='dds_origin' type-id='d1617432' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='dmu_objset_stats_t' type-id='098f0221' id='b2c14f17'/>
<enum-decl name='dmu_objset_type' id='6b1b19f9'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DMU_OST_NONE' value='0'/>
<enumerator name='DMU_OST_META' value='1'/>
<enumerator name='DMU_OST_ZFS' value='2'/>
<enumerator name='DMU_OST_ZVOL' value='3'/>
<enumerator name='DMU_OST_OTHER' value='4'/>
<enumerator name='DMU_OST_ANY' value='5'/>
<enumerator name='DMU_OST_NUMTYPES' value='6'/>
</enum-decl>
<typedef-decl name='dmu_objset_type_t' type-id='6b1b19f9' id='230f1e16'/>
<enum-decl name='pool_initialize_func' id='5c246ad4'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_INITIALIZE_START' value='0'/>
<enumerator name='POOL_INITIALIZE_CANCEL' value='1'/>
<enumerator name='POOL_INITIALIZE_SUSPEND' value='2'/>
<enumerator name='POOL_INITIALIZE_UNINIT' value='3'/>
<enumerator name='POOL_INITIALIZE_FUNCS' value='4'/>
</enum-decl>
<typedef-decl name='pool_initialize_func_t' type-id='5c246ad4' id='7063e1ab'/>
<enum-decl name='pool_trim_func' id='54ed608a'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_TRIM_START' value='0'/>
<enumerator name='POOL_TRIM_CANCEL' value='1'/>
<enumerator name='POOL_TRIM_SUSPEND' value='2'/>
<enumerator name='POOL_TRIM_FUNCS' value='3'/>
</enum-decl>
<typedef-decl name='pool_trim_func_t' type-id='54ed608a' id='b1146b8d'/>
<enum-decl name='zfs_ioc' id='12033f13'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_IOC_FIRST' value='23040'/>
<enumerator name='ZFS_IOC' value='23040'/>
<enumerator name='ZFS_IOC_POOL_CREATE' value='23040'/>
<enumerator name='ZFS_IOC_POOL_DESTROY' value='23041'/>
<enumerator name='ZFS_IOC_POOL_IMPORT' value='23042'/>
<enumerator name='ZFS_IOC_POOL_EXPORT' value='23043'/>
<enumerator name='ZFS_IOC_POOL_CONFIGS' value='23044'/>
<enumerator name='ZFS_IOC_POOL_STATS' value='23045'/>
<enumerator name='ZFS_IOC_POOL_TRYIMPORT' value='23046'/>
<enumerator name='ZFS_IOC_POOL_SCAN' value='23047'/>
<enumerator name='ZFS_IOC_POOL_FREEZE' value='23048'/>
<enumerator name='ZFS_IOC_POOL_UPGRADE' value='23049'/>
<enumerator name='ZFS_IOC_POOL_GET_HISTORY' value='23050'/>
<enumerator name='ZFS_IOC_VDEV_ADD' value='23051'/>
<enumerator name='ZFS_IOC_VDEV_REMOVE' value='23052'/>
<enumerator name='ZFS_IOC_VDEV_SET_STATE' value='23053'/>
<enumerator name='ZFS_IOC_VDEV_ATTACH' value='23054'/>
<enumerator name='ZFS_IOC_VDEV_DETACH' value='23055'/>
<enumerator name='ZFS_IOC_VDEV_SETPATH' value='23056'/>
<enumerator name='ZFS_IOC_VDEV_SETFRU' value='23057'/>
<enumerator name='ZFS_IOC_OBJSET_STATS' value='23058'/>
<enumerator name='ZFS_IOC_OBJSET_ZPLPROPS' value='23059'/>
<enumerator name='ZFS_IOC_DATASET_LIST_NEXT' value='23060'/>
<enumerator name='ZFS_IOC_SNAPSHOT_LIST_NEXT' value='23061'/>
<enumerator name='ZFS_IOC_SET_PROP' value='23062'/>
<enumerator name='ZFS_IOC_CREATE' value='23063'/>
<enumerator name='ZFS_IOC_DESTROY' value='23064'/>
<enumerator name='ZFS_IOC_ROLLBACK' value='23065'/>
<enumerator name='ZFS_IOC_RENAME' value='23066'/>
<enumerator name='ZFS_IOC_RECV' value='23067'/>
<enumerator name='ZFS_IOC_SEND' value='23068'/>
<enumerator name='ZFS_IOC_INJECT_FAULT' value='23069'/>
<enumerator name='ZFS_IOC_CLEAR_FAULT' value='23070'/>
<enumerator name='ZFS_IOC_INJECT_LIST_NEXT' value='23071'/>
<enumerator name='ZFS_IOC_ERROR_LOG' value='23072'/>
<enumerator name='ZFS_IOC_CLEAR' value='23073'/>
<enumerator name='ZFS_IOC_PROMOTE' value='23074'/>
<enumerator name='ZFS_IOC_SNAPSHOT' value='23075'/>
<enumerator name='ZFS_IOC_DSOBJ_TO_DSNAME' value='23076'/>
<enumerator name='ZFS_IOC_OBJ_TO_PATH' value='23077'/>
<enumerator name='ZFS_IOC_POOL_SET_PROPS' value='23078'/>
<enumerator name='ZFS_IOC_POOL_GET_PROPS' value='23079'/>
<enumerator name='ZFS_IOC_SET_FSACL' value='23080'/>
<enumerator name='ZFS_IOC_GET_FSACL' value='23081'/>
<enumerator name='ZFS_IOC_SHARE' value='23082'/>
<enumerator name='ZFS_IOC_INHERIT_PROP' value='23083'/>
<enumerator name='ZFS_IOC_SMB_ACL' value='23084'/>
<enumerator name='ZFS_IOC_USERSPACE_ONE' value='23085'/>
<enumerator name='ZFS_IOC_USERSPACE_MANY' value='23086'/>
<enumerator name='ZFS_IOC_USERSPACE_UPGRADE' value='23087'/>
<enumerator name='ZFS_IOC_HOLD' value='23088'/>
<enumerator name='ZFS_IOC_RELEASE' value='23089'/>
<enumerator name='ZFS_IOC_GET_HOLDS' value='23090'/>
<enumerator name='ZFS_IOC_OBJSET_RECVD_PROPS' value='23091'/>
<enumerator name='ZFS_IOC_VDEV_SPLIT' value='23092'/>
<enumerator name='ZFS_IOC_NEXT_OBJ' value='23093'/>
<enumerator name='ZFS_IOC_DIFF' value='23094'/>
<enumerator name='ZFS_IOC_TMP_SNAPSHOT' value='23095'/>
<enumerator name='ZFS_IOC_OBJ_TO_STATS' value='23096'/>
<enumerator name='ZFS_IOC_SPACE_WRITTEN' value='23097'/>
<enumerator name='ZFS_IOC_SPACE_SNAPS' value='23098'/>
<enumerator name='ZFS_IOC_DESTROY_SNAPS' value='23099'/>
<enumerator name='ZFS_IOC_POOL_REGUID' value='23100'/>
<enumerator name='ZFS_IOC_POOL_REOPEN' value='23101'/>
<enumerator name='ZFS_IOC_SEND_PROGRESS' value='23102'/>
<enumerator name='ZFS_IOC_LOG_HISTORY' value='23103'/>
<enumerator name='ZFS_IOC_SEND_NEW' value='23104'/>
<enumerator name='ZFS_IOC_SEND_SPACE' value='23105'/>
<enumerator name='ZFS_IOC_CLONE' value='23106'/>
<enumerator name='ZFS_IOC_BOOKMARK' value='23107'/>
<enumerator name='ZFS_IOC_GET_BOOKMARKS' value='23108'/>
<enumerator name='ZFS_IOC_DESTROY_BOOKMARKS' value='23109'/>
<enumerator name='ZFS_IOC_RECV_NEW' value='23110'/>
<enumerator name='ZFS_IOC_POOL_SYNC' value='23111'/>
<enumerator name='ZFS_IOC_CHANNEL_PROGRAM' value='23112'/>
<enumerator name='ZFS_IOC_LOAD_KEY' value='23113'/>
<enumerator name='ZFS_IOC_UNLOAD_KEY' value='23114'/>
<enumerator name='ZFS_IOC_CHANGE_KEY' value='23115'/>
<enumerator name='ZFS_IOC_REMAP' value='23116'/>
<enumerator name='ZFS_IOC_POOL_CHECKPOINT' value='23117'/>
<enumerator name='ZFS_IOC_POOL_DISCARD_CHECKPOINT' value='23118'/>
<enumerator name='ZFS_IOC_POOL_INITIALIZE' value='23119'/>
<enumerator name='ZFS_IOC_POOL_TRIM' value='23120'/>
<enumerator name='ZFS_IOC_REDACT' value='23121'/>
<enumerator name='ZFS_IOC_GET_BOOKMARK_PROPS' value='23122'/>
<enumerator name='ZFS_IOC_WAIT' value='23123'/>
<enumerator name='ZFS_IOC_WAIT_FS' value='23124'/>
<enumerator name='ZFS_IOC_VDEV_GET_PROPS' value='23125'/>
<enumerator name='ZFS_IOC_VDEV_SET_PROPS' value='23126'/>
<enumerator name='ZFS_IOC_POOL_SCRUB' value='23127'/>
<enumerator name='ZFS_IOC_PLATFORM' value='23168'/>
<enumerator name='ZFS_IOC_EVENTS_NEXT' value='23169'/>
<enumerator name='ZFS_IOC_EVENTS_CLEAR' value='23170'/>
<enumerator name='ZFS_IOC_EVENTS_SEEK' value='23171'/>
<enumerator name='ZFS_IOC_NEXTBOOT' value='23172'/>
<enumerator name='ZFS_IOC_JAIL' value='23173'/>
+ <enumerator name='ZFS_IOC_USERNS_ATTACH' value='23173'/>
<enumerator name='ZFS_IOC_UNJAIL' value='23174'/>
+ <enumerator name='ZFS_IOC_USERNS_DETACH' value='23174'/>
<enumerator name='ZFS_IOC_SET_BOOTENV' value='23175'/>
<enumerator name='ZFS_IOC_GET_BOOTENV' value='23176'/>
<enumerator name='ZFS_IOC_LAST' value='23177'/>
</enum-decl>
<typedef-decl name='zfs_ioc_t' type-id='12033f13' id='5b35941c'/>
<enum-decl name='zpool_wait_activity_t' naming-typedef-id='73446457' id='849338e3'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_WAIT_CKPT_DISCARD' value='0'/>
<enumerator name='ZPOOL_WAIT_FREE' value='1'/>
<enumerator name='ZPOOL_WAIT_INITIALIZE' value='2'/>
<enumerator name='ZPOOL_WAIT_REPLACE' value='3'/>
<enumerator name='ZPOOL_WAIT_REMOVE' value='4'/>
<enumerator name='ZPOOL_WAIT_RESILVER' value='5'/>
<enumerator name='ZPOOL_WAIT_SCRUB' value='6'/>
<enumerator name='ZPOOL_WAIT_TRIM' value='7'/>
- <enumerator name='ZPOOL_WAIT_NUM_ACTIVITIES' value='8'/>
+ <enumerator name='ZPOOL_WAIT_RAIDZ_EXPAND' value='8'/>
+ <enumerator name='ZPOOL_WAIT_NUM_ACTIVITIES' value='9'/>
</enum-decl>
<typedef-decl name='zpool_wait_activity_t' type-id='849338e3' id='73446457'/>
<enum-decl name='zfs_wait_activity_t' naming-typedef-id='3024501a' id='527d5dc6'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_WAIT_DELETEQ' value='0'/>
<enumerator name='ZFS_WAIT_NUM_ACTIVITIES' value='1'/>
</enum-decl>
<typedef-decl name='zfs_wait_activity_t' type-id='527d5dc6' id='3024501a'/>
<enum-decl name='data_type_t' naming-typedef-id='8d0687d2' id='aeeae136'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DATA_TYPE_DONTCARE' value='-1'/>
<enumerator name='DATA_TYPE_UNKNOWN' value='0'/>
<enumerator name='DATA_TYPE_BOOLEAN' value='1'/>
<enumerator name='DATA_TYPE_BYTE' value='2'/>
<enumerator name='DATA_TYPE_INT16' value='3'/>
<enumerator name='DATA_TYPE_UINT16' value='4'/>
<enumerator name='DATA_TYPE_INT32' value='5'/>
<enumerator name='DATA_TYPE_UINT32' value='6'/>
<enumerator name='DATA_TYPE_INT64' value='7'/>
<enumerator name='DATA_TYPE_UINT64' value='8'/>
<enumerator name='DATA_TYPE_STRING' value='9'/>
<enumerator name='DATA_TYPE_BYTE_ARRAY' value='10'/>
<enumerator name='DATA_TYPE_INT16_ARRAY' value='11'/>
<enumerator name='DATA_TYPE_UINT16_ARRAY' value='12'/>
<enumerator name='DATA_TYPE_INT32_ARRAY' value='13'/>
<enumerator name='DATA_TYPE_UINT32_ARRAY' value='14'/>
<enumerator name='DATA_TYPE_INT64_ARRAY' value='15'/>
<enumerator name='DATA_TYPE_UINT64_ARRAY' value='16'/>
<enumerator name='DATA_TYPE_STRING_ARRAY' value='17'/>
<enumerator name='DATA_TYPE_HRTIME' value='18'/>
<enumerator name='DATA_TYPE_NVLIST' value='19'/>
<enumerator name='DATA_TYPE_NVLIST_ARRAY' value='20'/>
<enumerator name='DATA_TYPE_BOOLEAN_VALUE' value='21'/>
<enumerator name='DATA_TYPE_INT8' value='22'/>
<enumerator name='DATA_TYPE_UINT8' value='23'/>
<enumerator name='DATA_TYPE_BOOLEAN_ARRAY' value='24'/>
<enumerator name='DATA_TYPE_INT8_ARRAY' value='25'/>
<enumerator name='DATA_TYPE_UINT8_ARRAY' value='26'/>
<enumerator name='DATA_TYPE_DOUBLE' value='27'/>
</enum-decl>
<typedef-decl name='data_type_t' type-id='aeeae136' id='8d0687d2'/>
<class-decl name='nvpair' size-in-bits='128' is-struct='yes' visibility='default' id='1c34e459'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='nvp_size' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='nvp_name_sz' type-id='23bd8cb5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='48'>
<var-decl name='nvp_reserve' type-id='23bd8cb5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='nvp_value_elem' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='nvp_type' type-id='8d0687d2' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='nvp_name' type-id='e84913bd' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='nvpair_t' type-id='1c34e459' id='57928edf'/>
<class-decl name='nvlist' size-in-bits='192' is-struct='yes' visibility='default' id='ac266fd9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='nvl_version' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='nvl_nvflag' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='nvl_priv' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='nvl_flag' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='nvl_pad' type-id='3ff5601b' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='nvlist_t' type-id='ac266fd9' id='8e8d4be3'/>
<class-decl name='zio_cksum' size-in-bits='256' is-struct='yes' visibility='default' id='1d53e28b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zc_word' type-id='85c64d26' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zio_cksum_t' type-id='1d53e28b' id='39730d0b'/>
<class-decl name='drr_begin' size-in-bits='2432' is-struct='yes' visibility='default' id='09fcdc01'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_magic' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_versioninfo' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_creation_time' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_type' type-id='230f1e16' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='drr_flags' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_fromguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_toname' type-id='d1617432' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='dmu_replay_record' size-in-bits='2496' is-struct='yes' visibility='default' id='781a52d7'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_type' type-id='08f5ca17' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='drr_payloadlen' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_u' type-id='ac5ab595' visibility='default'/>
</data-member>
</class-decl>
<enum-decl name='__anonymous_enum__' is-anonymous='yes' id='08f5ca17'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DRR_BEGIN' value='0'/>
<enumerator name='DRR_OBJECT' value='1'/>
<enumerator name='DRR_FREEOBJECTS' value='2'/>
<enumerator name='DRR_WRITE' value='3'/>
<enumerator name='DRR_FREE' value='4'/>
<enumerator name='DRR_END' value='5'/>
<enumerator name='DRR_WRITE_BYREF' value='6'/>
<enumerator name='DRR_SPILL' value='7'/>
<enumerator name='DRR_WRITE_EMBEDDED' value='8'/>
<enumerator name='DRR_OBJECT_RANGE' value='9'/>
<enumerator name='DRR_REDACT' value='10'/>
<enumerator name='DRR_NUMTYPES' value='11'/>
</enum-decl>
<union-decl name='__anonymous_union__' size-in-bits='2432' is-anonymous='yes' visibility='default' id='ac5ab595'>
<data-member access='public'>
<var-decl name='drr_begin' type-id='09fcdc01' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_end' type-id='6ee25631' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_object' type-id='f9ad530b' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_freeobjects' type-id='a27d958e' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_write' type-id='4cc69e4b' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_free' type-id='c836cfd2' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_write_byref' type-id='e511cdce' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_spill' type-id='1e69a80a' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_write_embedded' type-id='98b1345e' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_object_range' type-id='aba1f9e1' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_redact' type-id='50389039' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_checksum' type-id='a5fe3647' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='drr_end' size-in-bits='320' is-struct='yes' visibility='default' id='6ee25631'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_checksum' type-id='39730d0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_object' size-in-bits='448' is-struct='yes' visibility='default' id='f9ad530b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_type' type-id='5c9d8906' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='drr_bonustype' type-id='5c9d8906' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_blksz' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='drr_bonuslen' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_checksumtype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='200'>
<var-decl name='drr_compress' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='208'>
<var-decl name='drr_dn_slots' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='216'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='drr_raw_bonuslen' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_indblkshift' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='328'>
<var-decl name='drr_nlevels' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='336'>
<var-decl name='drr_nblkptr' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='344'>
<var-decl name='drr_pad' type-id='0f4ddd0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_maxblkid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_freeobjects' size-in-bits='192' is-struct='yes' visibility='default' id='a27d958e'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_firstobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_numobjs' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_write' size-in-bits='1088' is-struct='yes' visibility='default' id='4cc69e4b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_type' type-id='5c9d8906' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='drr_pad' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_logical_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_checksumtype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='328'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='336'>
<var-decl name='drr_compressiontype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='344'>
<var-decl name='drr_pad2' type-id='0f4ddd0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_key' type-id='67f6d2cf' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='drr_compressed_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='768'>
<var-decl name='drr_salt' type-id='13339fda' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='drr_iv' type-id='fa8ef949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='928'>
<var-decl name='drr_mac' type-id='fa9986a5' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_free' size-in-bits='256' is-struct='yes' visibility='default' id='c836cfd2'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_write_byref' size-in-bits='832' is-struct='yes' visibility='default' id='e511cdce'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_refguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_refobject' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_refoffset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='drr_checksumtype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='456'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='464'>
<var-decl name='drr_pad2' type-id='0f562bd0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='drr_key' type-id='67f6d2cf' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_spill' size-in-bits='640' is-struct='yes' visibility='default' id='1e69a80a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='200'>
<var-decl name='drr_compressiontype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='208'>
<var-decl name='drr_pad' type-id='0f562bd0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_compressed_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_salt' type-id='13339fda' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_iv' type-id='fa8ef949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='drr_mac' type-id='fa9986a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='608'>
<var-decl name='drr_type' type-id='5c9d8906' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_write_embedded' size-in-bits='384' is-struct='yes' visibility='default' id='98b1345e'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_compression' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='264'>
<var-decl name='drr_etype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='272'>
<var-decl name='drr_pad' type-id='0f562bd0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_lsize' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='drr_psize' type-id='8f92235e' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_object_range' size-in-bits='512' is-struct='yes' visibility='default' id='aba1f9e1'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_firstobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_numslots' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_salt' type-id='13339fda' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_iv' type-id='fa8ef949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='drr_mac' type-id='fa9986a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='488'>
<var-decl name='drr_pad' type-id='d3490169' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_redact' size-in-bits='256' is-struct='yes' visibility='default' id='50389039'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_checksum' size-in-bits='2432' is-struct='yes' visibility='default' id='a5fe3647'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_pad' type-id='8c2bcad1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2176'>
<var-decl name='drr_checksum' type-id='39730d0b' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='dmu_replay_record_t' type-id='781a52d7' id='8b8fc893'/>
<class-decl name='zinject_record' size-in-bits='2816' is-struct='yes' visibility='default' id='3216f820'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zi_objset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zi_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zi_start' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='zi_end' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='zi_guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='zi_level' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='zi_error' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='zi_type' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='zi_freq' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='zi_failfast' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='zi_func' type-id='d1617432' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2560'>
<var-decl name='zi_iotype' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2592'>
<var-decl name='zi_duration' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2624'>
<var-decl name='zi_timer' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2688'>
<var-decl name='zi_nlanes' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2752'>
<var-decl name='zi_cmd' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2784'>
<var-decl name='zi_dvas' type-id='8f92235e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zinject_record_t' type-id='3216f820' id='a4301ca6'/>
<class-decl name='zfs_share' size-in-bits='256' is-struct='yes' visibility='default' id='feb6f2da'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='z_exportdata' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='z_sharedata' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='z_sharetype' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='z_sharemax' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_share_t' type-id='feb6f2da' id='ee5cec36'/>
<class-decl name='zfs_cmd' size-in-bits='109952' is-struct='yes' visibility='default' id='3522cd69'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zc_name' type-id='d16c6df4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32768'>
<var-decl name='zc_nvlist_src' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32832'>
<var-decl name='zc_nvlist_src_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32896'>
<var-decl name='zc_nvlist_dst' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32960'>
<var-decl name='zc_nvlist_dst_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33024'>
<var-decl name='zc_nvlist_dst_filled' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33056'>
<var-decl name='zc_pad2' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33088'>
<var-decl name='zc_history' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33152'>
<var-decl name='zc_value' type-id='163f6aa5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='98688'>
<var-decl name='zc_string' type-id='d1617432' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100736'>
<var-decl name='zc_guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100800'>
<var-decl name='zc_nvlist_conf' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100864'>
<var-decl name='zc_nvlist_conf_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100928'>
<var-decl name='zc_cookie' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100992'>
<var-decl name='zc_objset_type' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101056'>
<var-decl name='zc_perm_action' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101120'>
<var-decl name='zc_history_len' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101184'>
<var-decl name='zc_history_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101248'>
<var-decl name='zc_obj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101312'>
<var-decl name='zc_iflags' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101376'>
<var-decl name='zc_share' type-id='ee5cec36' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101632'>
<var-decl name='zc_objset_stats' type-id='b2c14f17' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='103936'>
<var-decl name='zc_begin_record' type-id='09fcdc01' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='106368'>
<var-decl name='zc_inject_record' type-id='a4301ca6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109184'>
<var-decl name='zc_defer_destroy' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109216'>
<var-decl name='zc_flags' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109248'>
<var-decl name='zc_action_handle' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109312'>
<var-decl name='zc_cleanup_fd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109344'>
<var-decl name='zc_simple' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109352'>
<var-decl name='zc_pad' type-id='d3490169' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109376'>
<var-decl name='zc_sendobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109440'>
<var-decl name='zc_fromobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109504'>
<var-decl name='zc_createtxg' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109568'>
<var-decl name='zc_stat' type-id='0371a9c7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109888'>
<var-decl name='zc_zoneid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='zfs_stat' size-in-bits='320' is-struct='yes' visibility='default' id='6417f0b9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zs_gen' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zs_mode' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zs_links' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='zs_ctime' type-id='c1c22e6c' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_stat_t' type-id='6417f0b9' id='0371a9c7'/>
<enum-decl name='boolean_t' naming-typedef-id='c19b74c3' id='f58c8277'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='B_FALSE' value='0'/>
<enumerator name='B_TRUE' value='1'/>
</enum-decl>
<typedef-decl name='boolean_t' type-id='f58c8277' id='c19b74c3'/>
<typedef-decl name='uchar_t' type-id='002ac4a6' id='d8bf0010'/>
<typedef-decl name='uint_t' type-id='f0981eeb' id='3502e3ff'/>
<typedef-decl name='pthread_t' type-id='7359adad' id='4051f5e7'/>
<union-decl name='pthread_attr_t' size-in-bits='448' visibility='default' id='b63afacd'>
<data-member access='public'>
<var-decl name='__size' type-id='6093ff7c' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='bd54fe1a' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_attr_t' type-id='b63afacd' id='7d8569fd'/>
<union-decl name='pthread_mutex_t' size-in-bits='320' naming-typedef-id='7a6844eb' visibility='default' id='70681f9b'>
<data-member access='public'>
<var-decl name='__data' type-id='4c734837' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__size' type-id='36c46961' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='bd54fe1a' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_mutex_t' type-id='70681f9b' id='7a6844eb'/>
<typedef-decl name='int16_t' type-id='03896e23' id='23bd8cb5'/>
<typedef-decl name='int32_t' type-id='33f57a65' id='3ff5601b'/>
<typedef-decl name='uint8_t' type-id='c51d6389' id='b96825af'/>
<typedef-decl name='uint32_t' type-id='62f1140c' id='8f92235e'/>
<typedef-decl name='uint64_t' type-id='8910171f' id='9c313c2d'/>
<class-decl name='__pthread_mutex_s' size-in-bits='320' is-struct='yes' visibility='default' id='4c734837'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__lock' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='__count' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='__owner' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='__nusers' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='__kind' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='__spins' type-id='a2185560' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='176'>
<var-decl name='__elision' type-id='a2185560' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='__list' type-id='518fb49c' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='stat' size-in-bits='1152' is-struct='yes' visibility='default' id='aafc373f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='st_dev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='st_ino' type-id='e43e523d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='st_nlink' type-id='80f0b9df' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='st_mode' type-id='e1c52942' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='st_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='st_gid' type-id='d94ec6d9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='__pad0' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='st_rdev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='st_size' type-id='79989e9c' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='st_blksize' type-id='d3f10a7f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='st_blocks' type-id='dbc43803' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='st_atim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='st_mtim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='st_ctim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='__glibc_reserved' type-id='083f8d58' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__pthread_internal_list' size-in-bits='128' is-struct='yes' visibility='default' id='0e01899c'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__prev' type-id='4d98cd5a' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='__next' type-id='4d98cd5a' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__pthread_list_t' type-id='0e01899c' id='518fb49c'/>
<typedef-decl name='__uint8_t' type-id='002ac4a6' id='c51d6389'/>
<typedef-decl name='__int16_t' type-id='a2185560' id='03896e23'/>
<typedef-decl name='__int32_t' type-id='95e97e5e' id='33f57a65'/>
<typedef-decl name='__uint32_t' type-id='f0981eeb' id='62f1140c'/>
<typedef-decl name='__uint64_t' type-id='7359adad' id='8910171f'/>
<typedef-decl name='__dev_t' type-id='7359adad' id='35ed8932'/>
<typedef-decl name='__uid_t' type-id='f0981eeb' id='cc5fcceb'/>
<typedef-decl name='__gid_t' type-id='f0981eeb' id='d94ec6d9'/>
<typedef-decl name='__ino_t' type-id='7359adad' id='e43e523d'/>
<typedef-decl name='__mode_t' type-id='f0981eeb' id='e1c52942'/>
<typedef-decl name='__nlink_t' type-id='7359adad' id='80f0b9df'/>
<typedef-decl name='__off_t' type-id='bd54fe1a' id='79989e9c'/>
<typedef-decl name='__off64_t' type-id='bd54fe1a' id='724e4de6'/>
<typedef-decl name='__time_t' type-id='bd54fe1a' id='65eda9c0'/>
<typedef-decl name='__blksize_t' type-id='bd54fe1a' id='d3f10a7f'/>
<typedef-decl name='__blkcnt_t' type-id='bd54fe1a' id='dbc43803'/>
<typedef-decl name='__ssize_t' type-id='bd54fe1a' id='41060289'/>
<typedef-decl name='__syscall_slong_t' type-id='bd54fe1a' id='03085adc'/>
<typedef-decl name='FILE' type-id='ec1ed955' id='aa12d1ba'/>
<typedef-decl name='_IO_lock_t' type-id='48b5725f' id='bb4788fa'/>
<class-decl name='_IO_FILE' size-in-bits='1728' is-struct='yes' visibility='default' id='ec1ed955'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='_IO_read_ptr' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='_IO_read_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='_IO_read_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='_IO_write_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='_IO_write_ptr' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='_IO_write_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='_IO_buf_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='_IO_buf_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='_IO_save_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='640'>
<var-decl name='_IO_backup_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='_IO_save_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='768'>
<var-decl name='_markers' type-id='e4c6fa61' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='_chain' type-id='dca988a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='896'>
<var-decl name='_fileno' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='928'>
<var-decl name='_flags2' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='_old_offset' type-id='79989e9c' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1024'>
<var-decl name='_cur_column' type-id='8efea9e5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1040'>
<var-decl name='_vtable_offset' type-id='28577a57' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1048'>
<var-decl name='_shortbuf' type-id='89feb1ec' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1088'>
<var-decl name='_lock' type-id='cecf4ea7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1152'>
<var-decl name='_offset' type-id='724e4de6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1216'>
<var-decl name='_codecvt' type-id='570f8c59' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1280'>
<var-decl name='_wide_data' type-id='c65a1f29' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1344'>
<var-decl name='_freeres_list' type-id='dca988a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1408'>
<var-decl name='_freeres_buf' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1472'>
<var-decl name='__pad5' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1536'>
<var-decl name='_mode' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1568'>
<var-decl name='_unused2' type-id='664ac0b7' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='timespec' size-in-bits='128' is-struct='yes' visibility='default' id='a9c79a1f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tv_sec' type-id='65eda9c0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tv_nsec' type-id='03085adc' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='ssize_t' type-id='41060289' id='79a0948f'/>
<typedef-decl name='size_t' type-id='7359adad' id='b59d7dce'/>
<pointer-type-def type-id='aa12d1ba' size-in-bits='64' id='822cd80b'/>
<qualified-type-def type-id='822cd80b' restrict='yes' id='e75a27e9'/>
<pointer-type-def type-id='ec1ed955' size-in-bits='64' id='dca988a5'/>
<pointer-type-def type-id='a4036571' size-in-bits='64' id='570f8c59'/>
<pointer-type-def type-id='bb4788fa' size-in-bits='64' id='cecf4ea7'/>
<pointer-type-def type-id='010ae0b9' size-in-bits='64' id='e4c6fa61'/>
<pointer-type-def type-id='79bd3751' size-in-bits='64' id='c65a1f29'/>
<pointer-type-def type-id='724e4de6' size-in-bits='64' id='ecf845f9'/>
<pointer-type-def type-id='0e01899c' size-in-bits='64' id='4d98cd5a'/>
<pointer-type-def type-id='c19b74c3' size-in-bits='64' id='37e3bd22'/>
<pointer-type-def type-id='a84c031d' size-in-bits='64' id='26a90f95'/>
<qualified-type-def type-id='a84c031d' const='yes' id='9b45d938'/>
<pointer-type-def type-id='9b45d938' size-in-bits='64' id='80f4b756'/>
<qualified-type-def type-id='80f4b756' restrict='yes' id='9d26089a'/>
<qualified-type-def type-id='8b8fc893' const='yes' id='9623bc03'/>
<pointer-type-def type-id='9623bc03' size-in-bits='64' id='8341348b'/>
<qualified-type-def type-id='8e8d4be3' const='yes' id='693c3853'/>
<pointer-type-def type-id='693c3853' size-in-bits='64' id='22cce67b'/>
<qualified-type-def type-id='57928edf' const='yes' id='642ee20f'/>
<pointer-type-def type-id='642ee20f' size-in-bits='64' id='dace003f'/>
<qualified-type-def type-id='7d8569fd' const='yes' id='e06dee2d'/>
<pointer-type-def type-id='e06dee2d' size-in-bits='64' id='540db505'/>
<qualified-type-def type-id='540db505' restrict='yes' id='e1815e87'/>
<qualified-type-def type-id='d8bf0010' const='yes' id='a9125480'/>
<pointer-type-def type-id='a9125480' size-in-bits='64' id='d1db479e'/>
<qualified-type-def type-id='b96825af' const='yes' id='2b61797f'/>
<pointer-type-def type-id='2b61797f' size-in-bits='64' id='9f7200cf'/>
<pointer-type-def type-id='c70fa2e8' size-in-bits='64' id='2e711a2a'/>
<pointer-type-def type-id='95e97e5e' size-in-bits='64' id='7292109c'/>
<pointer-type-def type-id='8e8d4be3' size-in-bits='64' id='5ce45b60'/>
<pointer-type-def type-id='5ce45b60' size-in-bits='64' id='857bb57e'/>
<pointer-type-def type-id='57928edf' size-in-bits='64' id='3fa542f0'/>
<pointer-type-def type-id='7a6844eb' size-in-bits='64' id='18c91f9e'/>
<pointer-type-def type-id='4051f5e7' size-in-bits='64' id='e01b5462'/>
<qualified-type-def type-id='e01b5462' restrict='yes' id='cc338b26'/>
<pointer-type-def type-id='b59d7dce' size-in-bits='64' id='78c01427'/>
<pointer-type-def type-id='aafc373f' size-in-bits='64' id='4330df87'/>
<pointer-type-def type-id='9c313c2d' size-in-bits='64' id='5d6479ae'/>
<pointer-type-def type-id='b96825af' size-in-bits='64' id='ae3e8ca6'/>
<pointer-type-def type-id='48b5725f' size-in-bits='64' id='eaa32e2f'/>
<pointer-type-def type-id='cd5d79f4' size-in-bits='64' id='5ad9edb6'/>
<qualified-type-def type-id='eaa32e2f' restrict='yes' id='1b7446cd'/>
<pointer-type-def type-id='eaa32e2f' size-in-bits='64' id='63e171df'/>
<pointer-type-def type-id='3522cd69' size-in-bits='64' id='b65f7fd1'/>
<class-decl name='_IO_codecvt' is-struct='yes' visibility='default' is-declaration-only='yes' id='a4036571'/>
<class-decl name='_IO_marker' is-struct='yes' visibility='default' is-declaration-only='yes' id='010ae0b9'/>
<class-decl name='_IO_wide_data' is-struct='yes' visibility='default' is-declaration-only='yes' id='79bd3751'/>
<function-decl name='lzc_ioctl_fd' mangled-name='lzc_ioctl_fd' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_ioctl_fd'>
<parameter type-id='95e97e5e'/>
<parameter type-id='7359adad'/>
<parameter type-id='b65f7fd1'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='nvlist_unpack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='857bb57e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_next_nvpair' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='dace003f'/>
<return type-id='3fa542f0'/>
</function-decl>
<function-decl name='nvpair_name' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='fnvlist_alloc' visibility='default' binding='global' size-in-bits='64'>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='fnvlist_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_pack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='78c01427'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='fnvlist_pack_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_unpack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='fnvlist_dup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='fnvlist_add_boolean' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_boolean_value' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_int32' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='3ff5601b'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_byte_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='d1db479e'/>
<parameter type-id='3502e3ff'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_uint8_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9f7200cf'/>
<parameter type-id='3502e3ff'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_lookup_boolean_value' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='fnvlist_lookup_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='fnvlist_lookup_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='libspl_assertf' mangled-name='libspl_assertf' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libspl_assertf'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strlcpy' mangled-name='strlcpy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='strlcpy'>
<parameter type-id='26a90f95'/>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='__errno_location' visibility='default' binding='global' size-in-bits='64'>
<return type-id='7292109c'/>
</function-decl>
<function-decl name='pthread_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='cc338b26'/>
<parameter type-id='e1815e87'/>
<parameter type-id='5ad9edb6'/>
<parameter type-id='1b7446cd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_join' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='4051f5e7'/>
<parameter type-id='63e171df'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_mutex_lock' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='18c91f9e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_mutex_unlock' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='18c91f9e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fclose' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='malloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='getenv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strchr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strrchr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strcspn' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='close' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pipe2' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7292109c'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='splice' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='ecf845f9'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='ecf845f9'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='f0981eeb'/>
<return type-id='41060289'/>
</function-decl>
<function-decl name='__open_too_many_args' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__open_missing_mode' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__read_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='79a0948f'/>
</function-decl>
<function-decl name='libzfs_core_init' mangled-name='libzfs_core_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_core_init'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_core_fini' mangled-name='libzfs_core_fini' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_core_fini'>
<return type-id='48b5725f'/>
</function-decl>
+ <function-decl name='lzc_scrub' mangled-name='lzc_scrub' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_scrub'>
+ <parameter type-id='5b35941c' name='ioc'/>
+ <parameter type-id='80f4b756' name='name'/>
+ <parameter type-id='5ce45b60' name='source'/>
+ <parameter type-id='857bb57e' name='resultp'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
<function-decl name='lzc_create' mangled-name='lzc_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_create'>
<parameter type-id='80f4b756' name='fsname'/>
<parameter type-id='bc9887f1' name='type'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='ae3e8ca6' name='wkeydata'/>
<parameter type-id='3502e3ff' name='wkeylen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_clone' mangled-name='lzc_clone' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_clone'>
<parameter type-id='80f4b756' name='fsname'/>
<parameter type-id='80f4b756' name='origin'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_promote' mangled-name='lzc_promote' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_promote'>
<parameter type-id='80f4b756' name='fsname'/>
<parameter type-id='26a90f95' name='snapnamebuf'/>
<parameter type-id='95e97e5e' name='snapnamelen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_rename' mangled-name='lzc_rename' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_rename'>
<parameter type-id='80f4b756' name='source'/>
<parameter type-id='80f4b756' name='target'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_destroy' mangled-name='lzc_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_destroy'>
<parameter type-id='80f4b756' name='fsname'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_snapshot' mangled-name='lzc_snapshot' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_snapshot'>
<parameter type-id='5ce45b60' name='snaps'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='857bb57e' name='errlist'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_destroy_snaps' mangled-name='lzc_destroy_snaps' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_destroy_snaps'>
<parameter type-id='5ce45b60' name='snaps'/>
<parameter type-id='c19b74c3' name='defer'/>
<parameter type-id='857bb57e' name='errlist'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_snaprange_space' mangled-name='lzc_snaprange_space' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_snaprange_space'>
<parameter type-id='80f4b756' name='firstsnap'/>
<parameter type-id='80f4b756' name='lastsnap'/>
<parameter type-id='5d6479ae' name='usedp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_exists' mangled-name='lzc_exists' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_exists'>
<parameter type-id='80f4b756' name='dataset'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='lzc_sync' mangled-name='lzc_sync' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_sync'>
<parameter type-id='80f4b756' name='pool_name'/>
<parameter type-id='5ce45b60' name='innvl'/>
<parameter type-id='857bb57e' name='outnvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_hold' mangled-name='lzc_hold' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_hold'>
<parameter type-id='5ce45b60' name='holds'/>
<parameter type-id='95e97e5e' name='cleanup_fd'/>
<parameter type-id='857bb57e' name='errlist'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_release' mangled-name='lzc_release' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_release'>
<parameter type-id='5ce45b60' name='holds'/>
<parameter type-id='857bb57e' name='errlist'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_holds' mangled-name='lzc_get_holds' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_get_holds'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='857bb57e' name='holdsp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_wrapper' mangled-name='lzc_send_wrapper' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_send_wrapper'>
<parameter type-id='2e711a2a' name='func'/>
<parameter type-id='95e97e5e' name='orig_fd'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send' mangled-name='lzc_send' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_send'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='80f4b756' name='from'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='bfbd3c8e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_redacted' mangled-name='lzc_send_redacted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_send_redacted'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='80f4b756' name='from'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='bfbd3c8e' name='flags'/>
<parameter type-id='80f4b756' name='redactbook'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_resume' mangled-name='lzc_send_resume' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_send_resume'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='80f4b756' name='from'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='bfbd3c8e' name='flags'/>
<parameter type-id='9c313c2d' name='resumeobj'/>
<parameter type-id='9c313c2d' name='resumeoff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_resume_redacted' mangled-name='lzc_send_resume_redacted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_send_resume_redacted'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='80f4b756' name='from'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='bfbd3c8e' name='flags'/>
<parameter type-id='9c313c2d' name='resumeobj'/>
<parameter type-id='9c313c2d' name='resumeoff'/>
<parameter type-id='80f4b756' name='redactbook'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_space_resume_redacted' mangled-name='lzc_send_space_resume_redacted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_send_space_resume_redacted'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='80f4b756' name='from'/>
<parameter type-id='bfbd3c8e' name='flags'/>
<parameter type-id='9c313c2d' name='resumeobj'/>
<parameter type-id='9c313c2d' name='resumeoff'/>
<parameter type-id='9c313c2d' name='resume_bytes'/>
<parameter type-id='80f4b756' name='redactbook'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='5d6479ae' name='spacep'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_space' mangled-name='lzc_send_space' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_send_space'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='80f4b756' name='from'/>
<parameter type-id='bfbd3c8e' name='flags'/>
<parameter type-id='5d6479ae' name='spacep'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_receive' mangled-name='lzc_receive' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_receive'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='80f4b756' name='origin'/>
<parameter type-id='c19b74c3' name='force'/>
<parameter type-id='c19b74c3' name='raw'/>
<parameter type-id='95e97e5e' name='fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_receive_resumable' mangled-name='lzc_receive_resumable' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_receive_resumable'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='80f4b756' name='origin'/>
<parameter type-id='c19b74c3' name='force'/>
<parameter type-id='c19b74c3' name='raw'/>
<parameter type-id='95e97e5e' name='fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_receive_with_header' mangled-name='lzc_receive_with_header' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_receive_with_header'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='80f4b756' name='origin'/>
<parameter type-id='c19b74c3' name='force'/>
<parameter type-id='c19b74c3' name='resumable'/>
<parameter type-id='c19b74c3' name='raw'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='8341348b' name='begin_record'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_receive_one' mangled-name='lzc_receive_one' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_receive_one'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='80f4b756' name='origin'/>
<parameter type-id='c19b74c3' name='force'/>
<parameter type-id='c19b74c3' name='resumable'/>
<parameter type-id='c19b74c3' name='raw'/>
<parameter type-id='95e97e5e' name='input_fd'/>
<parameter type-id='8341348b' name='begin_record'/>
<parameter type-id='95e97e5e' name='cleanup_fd'/>
<parameter type-id='5d6479ae' name='read_bytes'/>
<parameter type-id='5d6479ae' name='errflags'/>
<parameter type-id='5d6479ae' name='action_handle'/>
<parameter type-id='857bb57e' name='errors'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_receive_with_cmdprops' mangled-name='lzc_receive_with_cmdprops' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_receive_with_cmdprops'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='5ce45b60' name='cmdprops'/>
<parameter type-id='ae3e8ca6' name='wkeydata'/>
<parameter type-id='3502e3ff' name='wkeylen'/>
<parameter type-id='80f4b756' name='origin'/>
<parameter type-id='c19b74c3' name='force'/>
<parameter type-id='c19b74c3' name='resumable'/>
<parameter type-id='c19b74c3' name='raw'/>
<parameter type-id='95e97e5e' name='input_fd'/>
<parameter type-id='8341348b' name='begin_record'/>
<parameter type-id='95e97e5e' name='cleanup_fd'/>
<parameter type-id='5d6479ae' name='read_bytes'/>
<parameter type-id='5d6479ae' name='errflags'/>
<parameter type-id='5d6479ae' name='action_handle'/>
<parameter type-id='857bb57e' name='errors'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_receive_with_heal' mangled-name='lzc_receive_with_heal' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_receive_with_heal'>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='5ce45b60' name='cmdprops'/>
<parameter type-id='ae3e8ca6' name='wkeydata'/>
<parameter type-id='3502e3ff' name='wkeylen'/>
<parameter type-id='80f4b756' name='origin'/>
<parameter type-id='c19b74c3' name='force'/>
<parameter type-id='c19b74c3' name='heal'/>
<parameter type-id='c19b74c3' name='resumable'/>
<parameter type-id='c19b74c3' name='raw'/>
<parameter type-id='95e97e5e' name='input_fd'/>
<parameter type-id='8341348b' name='begin_record'/>
<parameter type-id='95e97e5e' name='cleanup_fd'/>
<parameter type-id='5d6479ae' name='read_bytes'/>
<parameter type-id='5d6479ae' name='errflags'/>
<parameter type-id='5d6479ae' name='action_handle'/>
<parameter type-id='857bb57e' name='errors'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_rollback' mangled-name='lzc_rollback' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_rollback'>
<parameter type-id='80f4b756' name='fsname'/>
<parameter type-id='26a90f95' name='snapnamebuf'/>
<parameter type-id='95e97e5e' name='snapnamelen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_rollback_to' mangled-name='lzc_rollback_to' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_rollback_to'>
<parameter type-id='80f4b756' name='fsname'/>
<parameter type-id='80f4b756' name='snapname'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_bookmark' mangled-name='lzc_bookmark' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_bookmark'>
<parameter type-id='5ce45b60' name='bookmarks'/>
<parameter type-id='857bb57e' name='errlist'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_bookmarks' mangled-name='lzc_get_bookmarks' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_get_bookmarks'>
<parameter type-id='80f4b756' name='fsname'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='857bb57e' name='bmarks'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_bookmark_props' mangled-name='lzc_get_bookmark_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_get_bookmark_props'>
<parameter type-id='80f4b756' name='bookmark'/>
<parameter type-id='857bb57e' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_destroy_bookmarks' mangled-name='lzc_destroy_bookmarks' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_destroy_bookmarks'>
<parameter type-id='5ce45b60' name='bmarks'/>
<parameter type-id='857bb57e' name='errlist'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_channel_program' mangled-name='lzc_channel_program' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_channel_program'>
<parameter type-id='80f4b756' name='pool'/>
<parameter type-id='80f4b756' name='program'/>
<parameter type-id='9c313c2d' name='instrlimit'/>
<parameter type-id='9c313c2d' name='memlimit'/>
<parameter type-id='5ce45b60' name='argnvl'/>
<parameter type-id='857bb57e' name='outnvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_pool_checkpoint' mangled-name='lzc_pool_checkpoint' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_pool_checkpoint'>
<parameter type-id='80f4b756' name='pool'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_pool_checkpoint_discard' mangled-name='lzc_pool_checkpoint_discard' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_pool_checkpoint_discard'>
<parameter type-id='80f4b756' name='pool'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_channel_program_nosync' mangled-name='lzc_channel_program_nosync' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_channel_program_nosync'>
<parameter type-id='80f4b756' name='pool'/>
<parameter type-id='80f4b756' name='program'/>
<parameter type-id='9c313c2d' name='timeout'/>
<parameter type-id='9c313c2d' name='memlimit'/>
<parameter type-id='5ce45b60' name='argnvl'/>
<parameter type-id='857bb57e' name='outnvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_vdev_prop' mangled-name='lzc_get_vdev_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_get_vdev_prop'>
<parameter type-id='80f4b756' name='poolname'/>
<parameter type-id='5ce45b60' name='innvl'/>
<parameter type-id='857bb57e' name='outnvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_set_vdev_prop' mangled-name='lzc_set_vdev_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_set_vdev_prop'>
<parameter type-id='80f4b756' name='poolname'/>
<parameter type-id='5ce45b60' name='innvl'/>
<parameter type-id='857bb57e' name='outnvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_load_key' mangled-name='lzc_load_key' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_load_key'>
<parameter type-id='80f4b756' name='fsname'/>
<parameter type-id='c19b74c3' name='noop'/>
<parameter type-id='ae3e8ca6' name='wkeydata'/>
<parameter type-id='3502e3ff' name='wkeylen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_unload_key' mangled-name='lzc_unload_key' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_unload_key'>
<parameter type-id='80f4b756' name='fsname'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_change_key' mangled-name='lzc_change_key' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_change_key'>
<parameter type-id='80f4b756' name='fsname'/>
<parameter type-id='9c313c2d' name='crypt_cmd'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='ae3e8ca6' name='wkeydata'/>
<parameter type-id='3502e3ff' name='wkeylen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_reopen' mangled-name='lzc_reopen' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_reopen'>
<parameter type-id='80f4b756' name='pool_name'/>
<parameter type-id='c19b74c3' name='scrub_restart'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_initialize' mangled-name='lzc_initialize' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_initialize'>
<parameter type-id='80f4b756' name='poolname'/>
<parameter type-id='7063e1ab' name='cmd_type'/>
<parameter type-id='5ce45b60' name='vdevs'/>
<parameter type-id='857bb57e' name='errlist'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_trim' mangled-name='lzc_trim' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_trim'>
<parameter type-id='80f4b756' name='poolname'/>
<parameter type-id='b1146b8d' name='cmd_type'/>
<parameter type-id='9c313c2d' name='rate'/>
<parameter type-id='c19b74c3' name='secure'/>
<parameter type-id='5ce45b60' name='vdevs'/>
<parameter type-id='857bb57e' name='errlist'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_redact' mangled-name='lzc_redact' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_redact'>
<parameter type-id='80f4b756' name='snapshot'/>
<parameter type-id='80f4b756' name='bookname'/>
<parameter type-id='5ce45b60' name='snapnv'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_wait' mangled-name='lzc_wait' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_wait'>
<parameter type-id='80f4b756' name='pool'/>
<parameter type-id='73446457' name='activity'/>
<parameter type-id='37e3bd22' name='waited'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_wait_tag' mangled-name='lzc_wait_tag' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_wait_tag'>
<parameter type-id='80f4b756' name='pool'/>
<parameter type-id='73446457' name='activity'/>
<parameter type-id='9c313c2d' name='tag'/>
<parameter type-id='37e3bd22' name='waited'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_wait_fs' mangled-name='lzc_wait_fs' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_wait_fs'>
<parameter type-id='80f4b756' name='fs'/>
<parameter type-id='3024501a' name='activity'/>
<parameter type-id='37e3bd22' name='waited'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_set_bootenv' mangled-name='lzc_set_bootenv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_set_bootenv'>
<parameter type-id='80f4b756' name='pool'/>
<parameter type-id='22cce67b' name='env'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_bootenv' mangled-name='lzc_get_bootenv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='lzc_get_bootenv'>
<parameter type-id='80f4b756' name='pool'/>
<parameter type-id='857bb57e' name='outnvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-type size-in-bits='64' id='c70fa2e8'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='cd5d79f4'>
<parameter type-id='eaa32e2f'/>
<return type-id='eaa32e2f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs_core/os/linux/libzfs_core_ioctl.c' language='LANG_C99'>
<function-decl name='ioctl' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='7359adad'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
</abi-corpus>
diff --git a/sys/contrib/openzfs/lib/libzpool/Makefile.am b/sys/contrib/openzfs/lib/libzpool/Makefile.am
index 58d7f07527aa..3c986a707d2d 100644
--- a/sys/contrib/openzfs/lib/libzpool/Makefile.am
+++ b/sys/contrib/openzfs/lib/libzpool/Makefile.am
@@ -1,213 +1,214 @@
libzpool_la_CFLAGS = $(AM_CFLAGS) $(KERNEL_CFLAGS) $(LIBRARY_CFLAGS)
libzpool_la_CFLAGS += $(ZLIB_CFLAGS)
libzpool_la_CPPFLAGS = $(AM_CPPFLAGS) $(FORCEDEBUG_CPPFLAGS)
libzpool_la_CPPFLAGS += -I$(srcdir)/include/os/@ac_system_l@/zfs
libzpool_la_CPPFLAGS += -DLIB_ZPOOL_BUILD
lib_LTLIBRARIES += libzpool.la
CPPCHECKTARGETS += libzpool.la
dist_libzpool_la_SOURCES = \
%D%/kernel.c \
%D%/taskq.c \
%D%/util.c
nodist_libzpool_la_SOURCES = \
module/lua/lapi.c \
module/lua/lauxlib.c \
module/lua/lbaselib.c \
module/lua/lcode.c \
module/lua/lcompat.c \
module/lua/lcorolib.c \
module/lua/lctype.c \
module/lua/ldebug.c \
module/lua/ldo.c \
module/lua/lfunc.c \
module/lua/lgc.c \
module/lua/llex.c \
module/lua/lmem.c \
module/lua/lobject.c \
module/lua/lopcodes.c \
module/lua/lparser.c \
module/lua/lstate.c \
module/lua/lstring.c \
module/lua/lstrlib.c \
module/lua/ltable.c \
module/lua/ltablib.c \
module/lua/ltm.c \
module/lua/lvm.c \
module/lua/lzio.c \
\
module/os/linux/zfs/abd_os.c \
module/os/linux/zfs/arc_os.c \
module/os/linux/zfs/trace.c \
module/os/linux/zfs/vdev_file.c \
+ module/os/linux/zfs/vdev_label_os.c \
module/os/linux/zfs/zfs_debug.c \
module/os/linux/zfs/zfs_racct.c \
module/os/linux/zfs/zfs_znode.c \
module/os/linux/zfs/zio_crypt.c \
\
module/zcommon/cityhash.c \
module/zcommon/zfeature_common.c \
module/zcommon/zfs_comutil.c \
module/zcommon/zfs_deleg.c \
module/zcommon/zfs_fletcher.c \
module/zcommon/zfs_fletcher_aarch64_neon.c \
module/zcommon/zfs_fletcher_avx512.c \
module/zcommon/zfs_fletcher_intel.c \
module/zcommon/zfs_fletcher_sse.c \
module/zcommon/zfs_fletcher_superscalar.c \
module/zcommon/zfs_fletcher_superscalar4.c \
module/zcommon/zfs_namecheck.c \
module/zcommon/zfs_prop.c \
module/zcommon/zpool_prop.c \
module/zcommon/zprop_common.c \
\
module/zfs/abd.c \
module/zfs/aggsum.c \
module/zfs/arc.c \
module/zfs/blake3_zfs.c \
module/zfs/blkptr.c \
module/zfs/bplist.c \
module/zfs/bpobj.c \
module/zfs/bptree.c \
module/zfs/bqueue.c \
module/zfs/btree.c \
module/zfs/brt.c \
module/zfs/dbuf.c \
module/zfs/dbuf_stats.c \
module/zfs/ddt.c \
module/zfs/ddt_zap.c \
module/zfs/dmu.c \
module/zfs/dmu_diff.c \
module/zfs/dmu_object.c \
module/zfs/dmu_objset.c \
module/zfs/dmu_recv.c \
module/zfs/dmu_redact.c \
module/zfs/dmu_send.c \
module/zfs/dmu_traverse.c \
module/zfs/dmu_tx.c \
module/zfs/dmu_zfetch.c \
module/zfs/dnode.c \
module/zfs/dnode_sync.c \
module/zfs/dsl_bookmark.c \
module/zfs/dsl_crypt.c \
module/zfs/dsl_dataset.c \
module/zfs/dsl_deadlist.c \
module/zfs/dsl_deleg.c \
module/zfs/dsl_destroy.c \
module/zfs/dsl_dir.c \
module/zfs/dsl_pool.c \
module/zfs/dsl_prop.c \
module/zfs/dsl_scan.c \
module/zfs/dsl_synctask.c \
module/zfs/dsl_userhold.c \
module/zfs/edonr_zfs.c \
module/zfs/fm.c \
module/zfs/gzip.c \
module/zfs/hkdf.c \
module/zfs/lz4.c \
module/zfs/lz4_zfs.c \
module/zfs/lzjb.c \
module/zfs/metaslab.c \
module/zfs/mmp.c \
module/zfs/multilist.c \
module/zfs/objlist.c \
module/zfs/pathname.c \
module/zfs/range_tree.c \
module/zfs/refcount.c \
module/zfs/rrwlock.c \
module/zfs/sa.c \
module/zfs/sha2_zfs.c \
module/zfs/skein_zfs.c \
module/zfs/spa.c \
module/zfs/spa_checkpoint.c \
module/zfs/spa_config.c \
module/zfs/spa_errlog.c \
module/zfs/spa_history.c \
module/zfs/spa_log_spacemap.c \
module/zfs/spa_misc.c \
module/zfs/spa_stats.c \
module/zfs/space_map.c \
module/zfs/space_reftree.c \
module/zfs/txg.c \
module/zfs/uberblock.c \
module/zfs/unique.c \
module/zfs/vdev.c \
module/zfs/vdev_draid.c \
module/zfs/vdev_draid_rand.c \
module/zfs/vdev_indirect.c \
module/zfs/vdev_indirect_births.c \
module/zfs/vdev_indirect_mapping.c \
module/zfs/vdev_initialize.c \
module/zfs/vdev_label.c \
module/zfs/vdev_mirror.c \
module/zfs/vdev_missing.c \
module/zfs/vdev_queue.c \
module/zfs/vdev_raidz.c \
module/zfs/vdev_raidz_math.c \
module/zfs/vdev_raidz_math_aarch64_neon.c \
module/zfs/vdev_raidz_math_aarch64_neonx2.c \
module/zfs/vdev_raidz_math_avx2.c \
module/zfs/vdev_raidz_math_avx512bw.c \
module/zfs/vdev_raidz_math_avx512f.c \
module/zfs/vdev_raidz_math_powerpc_altivec.c \
module/zfs/vdev_raidz_math_scalar.c \
module/zfs/vdev_raidz_math_sse2.c \
module/zfs/vdev_raidz_math_ssse3.c \
module/zfs/vdev_rebuild.c \
module/zfs/vdev_removal.c \
module/zfs/vdev_root.c \
module/zfs/vdev_trim.c \
module/zfs/zap.c \
module/zfs/zap_leaf.c \
module/zfs/zap_micro.c \
module/zfs/zcp.c \
module/zfs/zcp_get.c \
module/zfs/zcp_global.c \
module/zfs/zcp_iter.c \
module/zfs/zcp_set.c \
module/zfs/zcp_synctask.c \
module/zfs/zfeature.c \
module/zfs/zfs_byteswap.c \
module/zfs/zfs_chksum.c \
module/zfs/zfs_fm.c \
module/zfs/zfs_fuid.c \
module/zfs/zfs_ratelimit.c \
module/zfs/zfs_rlock.c \
module/zfs/zfs_sa.c \
module/zfs/zil.c \
module/zfs/zio.c \
module/zfs/zio_checksum.c \
module/zfs/zio_compress.c \
module/zfs/zio_inject.c \
module/zfs/zle.c \
module/zfs/zrlock.c \
module/zfs/zthr.c
libzpool_la_LIBADD = \
libicp.la \
libunicode.la \
libnvpair.la \
libzstd.la \
libzutil.la
libzpool_la_LIBADD += $(LIBCLOCK_GETTIME) $(ZLIB_LIBS) -ldl -lm
libzpool_la_LDFLAGS = -pthread
if !ASAN_ENABLED
libzpool_la_LDFLAGS += -Wl,-z,defs
endif
if BUILD_FREEBSD
libzpool_la_LIBADD += -lgeom
endif
libzpool_la_LDFLAGS += -version-info 5:0:0
if TARGET_CPU_POWERPC
module/zfs/libzpool_la-vdev_raidz_math_powerpc_altivec.$(OBJEXT) : CFLAGS += -maltivec
module/zfs/libzpool_la-vdev_raidz_math_powerpc_altivec.l$(OBJEXT): CFLAGS += -maltivec
endif
diff --git a/sys/contrib/openzfs/man/man1/ztest.1 b/sys/contrib/openzfs/man/man1/ztest.1
index 64514b317275..bbbe751ca01d 100644
--- a/sys/contrib/openzfs/man/man1/ztest.1
+++ b/sys/contrib/openzfs/man/man1/ztest.1
@@ -1,244 +1,262 @@
.\"
.\" CDDL HEADER START
.\"
.\" The contents of this file are subject to the terms of the
.\" Common Development and Distribution License (the "License").
.\" You may not use this file except in compliance with the License.
.\"
.\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
.\" or https://opensource.org/licenses/CDDL-1.0.
.\" See the License for the specific language governing permissions
.\" and limitations under the License.
.\"
.\" When distributing Covered Code, include this CDDL HEADER in each
.\" file and include the License file at usr/src/OPENSOLARIS.LICENSE.
.\" If applicable, add the following below this CDDL HEADER, with the
.\" fields enclosed by brackets "[]" replaced with your own identifying
.\" information: Portions Copyright [yyyy] [name of copyright owner]
.\"
.\" CDDL HEADER END
.\"
.\" Copyright (c) 2009 Oracle and/or its affiliates. All rights reserved.
.\" Copyright (c) 2009 Michael Gebetsroither <michael.geb@gmx.at>. All rights
.\" reserved.
.\" Copyright (c) 2017, Intel Corporation.
.\"
.Dd May 26, 2021
.Dt ZTEST 1
.Os
.
.Sh NAME
.Nm ztest
.Nd was written by the ZFS Developers as a ZFS unit test
.Sh SYNOPSIS
.Nm
.Op Fl VEG
.Op Fl v Ar vdevs
.Op Fl s Ar size_of_each_vdev
.Op Fl a Ar alignment_shift
.Op Fl m Ar mirror_copies
.Op Fl r Ar raidz_disks/draid_disks
.Op Fl R Ar raid_parity
.Op Fl K Ar raid_kind
.Op Fl D Ar draid_data
.Op Fl S Ar draid_spares
.Op Fl C Ar vdev_class_state
.Op Fl d Ar datasets
.Op Fl t Ar threads
.Op Fl g Ar gang_block_threshold
.Op Fl i Ar initialize_pool_i_times
.Op Fl k Ar kill_percentage
.Op Fl p Ar pool_name
.Op Fl T Ar time
.Op Fl z Ar zil_failure_rate
.
+.Nm
+.Fl X
+.Op Fl VG
+.Op Fl s Ar size_of_each_vdev
+.Op Fl a Ar alignment_shift
+.Op Fl r Ar raidz_disks
+.Op Fl R Ar raid_parity
+.Op Fl d Ar datasets
+.Op Fl t Ar threads
+.
.Sh DESCRIPTION
.Nm
was written by the ZFS Developers as a ZFS unit test.
The tool was developed in tandem with the ZFS functionality and was
executed nightly as one of the many regression test against the daily build.
As features were added to ZFS, unit tests were also added to
.Nm .
In addition, a separate test development team wrote and
executed more functional and stress tests.
.
.Pp
By default
.Nm
runs for ten minutes and uses block files
(stored in
.Pa /tmp )
to create pools rather than using physical disks.
Block files afford
.Nm
its flexibility to play around with
zpool components without requiring large hardware configurations.
However, storing the block files in
.Pa /tmp
may not work for you if you
have a small tmp directory.
.
.Pp
By default is non-verbose.
This is why entering the command above will result in
.Nm
quietly executing for 5 minutes.
The
.Fl V
option can be used to increase the verbosity of the tool.
Adding multiple
.Fl V
options is allowed and the more you add the more chatty
.Nm
becomes.
.
.Pp
After the
.Nm
run completes, you should notice many
.Pa ztest.*
files lying around.
Once the run completes you can safely remove these files.
Note that you shouldn't remove these files during a run.
You can re-use these files in your next
.Nm
run by using the
.Fl E
option.
.
.Sh OPTIONS
.Bl -tag -width "-v v"
.It Fl h , \&? , -help
Print a help summary.
.It Fl v , -vdevs Ns = (default: Sy 5 )
Number of vdevs.
.It Fl s , -vdev-size Ns = (default: Sy 64M )
Size of each vdev.
.It Fl a , -alignment-shift Ns = (default: Sy 9 ) No (use Sy 0 No for random )
Alignment shift used in test.
.It Fl m , -mirror-copies Ns = (default: Sy 2 )
Number of mirror copies.
.It Fl r , -raid-disks Ns = (default: Sy 4 No for raidz/ Ns Sy 16 No for draid )
Number of raidz/draid disks.
.It Fl R , -raid-parity Ns = (default: Sy 1 )
Raid parity (raidz & draid).
-.It Fl K , -raid-kind Ns = Ns Sy raidz Ns | Ns Sy draid Ns | Ns Sy random No (default : Sy random )
+.It Xo
+.Fl K , -raid-kind Ns = Ns
+.Sy raidz Ns | Ns Sy eraidz Ns | Ns Sy draid Ns | Ns Sy random
+(default:
+.Sy random Ns
+)
+.Xc
The kind of RAID config to use.
With
.Sy random
-the kind alternates between raidz and draid.
+the kind alternates between raidz, eraidz (expandable raidz) and draid.
.It Fl D , -draid-data Ns = (default: Sy 4 )
Number of data disks in a dRAID redundancy group.
.It Fl S , -draid-spares Ns = (default: Sy 1 )
Number of dRAID distributed spare disks.
.It Fl d , -datasets Ns = (default: Sy 7 )
Number of datasets.
.It Fl t , -threads Ns = (default: Sy 23 )
Number of threads.
.It Fl g , -gang-block-threshold Ns = (default: Sy 32K )
Gang block threshold.
.It Fl i , -init-count Ns = (default: Sy 1 )
Number of pool initializations.
.It Fl k , -kill-percentage Ns = (default: Sy 70% )
Kill percentage.
.It Fl p , -pool-name Ns = (default: Sy ztest )
Pool name.
.It Fl f , -vdev-file-directory Ns = (default: Pa /tmp )
File directory for vdev files.
.It Fl M , -multi-host
Multi-host; simulate pool imported on remote host.
.It Fl E , -use-existing-pool
Use existing pool (use existing pool instead of creating new one).
.It Fl T , -run-time Ns = (default: Sy 300 Ns s)
Total test run time.
.It Fl P , -pass-time Ns = (default: Sy 60 Ns s)
Time per pass.
.It Fl F , -freeze-loops Ns = (default: Sy 50 )
Max loops in
.Fn spa_freeze .
.It Fl B , -alt-ztest Ns =
Path to alternate ("older")
.Nm ztest
to drive, which will be used to initialise the pool, and, a stochastic half the
time, to run the tests.
The parallel
.Pa lib
directory is prepended to
.Ev LD_LIBRARY_PATH ;
i.e. given
.Fl B Pa ./chroots/lenny/usr/bin/ Ns Nm ,
.Pa ./chroots/lenny/usr/lib
will be loaded.
.It Fl C , -vdev-class-state Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Sy random No (default : Sy random )
The vdev allocation class state.
.It Fl o , -option Ns = Ns Ar variable Ns = Ns Ar value
Set global
.Ar variable
to an unsigned 32-bit integer
.Ar value
(little-endian only).
.It Fl G , -dump-debug
Dump zfs_dbgmsg buffer before exiting due to an error.
.It Fl V , -verbose
Verbose (use multiple times for ever more verbosity).
+.It Fl X , -raidz-expansion
+Perform a dedicated raidz expansion test.
.El
.
.Sh EXAMPLES
To override
.Pa /tmp
as your location for block files, you can use the
.Fl f
option:
.Dl # ztest -f /
.Pp
To get an idea of what
.Nm
is actually testing try this:
.Dl # ztest -f / -VVV
.Pp
Maybe you'd like to run
.Nm ztest
for longer? To do so simply use the
.Fl T
option and specify the runlength in seconds like so:
.Dl # ztest -f / -V -T 120
.
.Sh ENVIRONMENT VARIABLES
.Bl -tag -width "ZF"
.It Ev ZFS_HOSTID Ns = Ns Em id
Use
.Em id
instead of the SPL hostid to identify this host.
Intended for use with
.Nm , but this environment variable will affect any utility which uses
libzpool, including
.Xr zpool 8 .
Since the kernel is unaware of this setting,
results with utilities other than ztest are undefined.
.It Ev ZFS_STACK_SIZE Ns = Ns Em stacksize
Limit the default stack size to
.Em stacksize
bytes for the purpose of
detecting and debugging kernel stack overflows.
This value defaults to
.Em 32K
which is double the default
.Em 16K
Linux kernel stack size.
.Pp
In practice, setting the stack size slightly higher is needed because
differences in stack usage between kernel and user space can lead to spurious
stack overflows (especially when debugging is enabled).
The specified value
will be rounded up to a floor of PTHREAD_STACK_MIN which is the minimum stack
required for a NULL procedure in user space.
.Pp
By default the stack size is limited to
.Em 256K .
.El
.
.Sh SEE ALSO
.Xr zdb 1 ,
.Xr zfs 1 ,
.Xr zpool 1 ,
.Xr spl 4
diff --git a/sys/contrib/openzfs/man/man4/zfs.4 b/sys/contrib/openzfs/man/man4/zfs.4
index f9824ac170ea..4a0441ed9469 100644
--- a/sys/contrib/openzfs/man/man4/zfs.4
+++ b/sys/contrib/openzfs/man/man4/zfs.4
@@ -1,2596 +1,2615 @@
.\"
.\" Copyright (c) 2013 by Turbo Fredriksson <turbo@bayour.com>. All rights reserved.
.\" Copyright (c) 2019, 2021 by Delphix. All rights reserved.
.\" Copyright (c) 2019 Datto Inc.
.\" The contents of this file are subject to the terms of the Common Development
.\" and Distribution License (the "License"). You may not use this file except
.\" in compliance with the License. You can obtain a copy of the license at
.\" usr/src/OPENSOLARIS.LICENSE or https://opensource.org/licenses/CDDL-1.0.
.\"
.\" See the License for the specific language governing permissions and
.\" limitations under the License. When distributing Covered Code, include this
.\" CDDL HEADER in each file and include the License file at
.\" usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this
.\" CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your
.\" own identifying information:
.\" Portions Copyright [yyyy] [name of copyright owner]
.\"
.Dd July 21, 2023
.Dt ZFS 4
.Os
.
.Sh NAME
.Nm zfs
.Nd tuning of the ZFS kernel module
.
.Sh DESCRIPTION
The ZFS module supports these parameters:
.Bl -tag -width Ds
.It Sy dbuf_cache_max_bytes Ns = Ns Sy UINT64_MAX Ns B Pq u64
Maximum size in bytes of the dbuf cache.
The target size is determined by the MIN versus
.No 1/2^ Ns Sy dbuf_cache_shift Pq 1/32nd
of the target ARC size.
The behavior of the dbuf cache and its associated settings
can be observed via the
.Pa /proc/spl/kstat/zfs/dbufstats
kstat.
.
.It Sy dbuf_metadata_cache_max_bytes Ns = Ns Sy UINT64_MAX Ns B Pq u64
Maximum size in bytes of the metadata dbuf cache.
The target size is determined by the MIN versus
.No 1/2^ Ns Sy dbuf_metadata_cache_shift Pq 1/64th
of the target ARC size.
The behavior of the metadata dbuf cache and its associated settings
can be observed via the
.Pa /proc/spl/kstat/zfs/dbufstats
kstat.
.
.It Sy dbuf_cache_hiwater_pct Ns = Ns Sy 10 Ns % Pq uint
The percentage over
.Sy dbuf_cache_max_bytes
when dbufs must be evicted directly.
.
.It Sy dbuf_cache_lowater_pct Ns = Ns Sy 10 Ns % Pq uint
The percentage below
.Sy dbuf_cache_max_bytes
when the evict thread stops evicting dbufs.
.
.It Sy dbuf_cache_shift Ns = Ns Sy 5 Pq uint
Set the size of the dbuf cache
.Pq Sy dbuf_cache_max_bytes
to a log2 fraction of the target ARC size.
.
.It Sy dbuf_metadata_cache_shift Ns = Ns Sy 6 Pq uint
Set the size of the dbuf metadata cache
.Pq Sy dbuf_metadata_cache_max_bytes
to a log2 fraction of the target ARC size.
.
.It Sy dbuf_mutex_cache_shift Ns = Ns Sy 0 Pq uint
Set the size of the mutex array for the dbuf cache.
When set to
.Sy 0
the array is dynamically sized based on total system memory.
.
.It Sy dmu_object_alloc_chunk_shift Ns = Ns Sy 7 Po 128 Pc Pq uint
dnode slots allocated in a single operation as a power of 2.
The default value minimizes lock contention for the bulk operation performed.
.
.It Sy dmu_prefetch_max Ns = Ns Sy 134217728 Ns B Po 128 MiB Pc Pq uint
Limit the amount we can prefetch with one call to this amount in bytes.
This helps to limit the amount of memory that can be used by prefetching.
.
.It Sy ignore_hole_birth Pq int
Alias for
.Sy send_holes_without_birth_time .
.
.It Sy l2arc_feed_again Ns = Ns Sy 1 Ns | Ns 0 Pq int
Turbo L2ARC warm-up.
When the L2ARC is cold the fill interval will be set as fast as possible.
.
.It Sy l2arc_feed_min_ms Ns = Ns Sy 200 Pq u64
Min feed interval in milliseconds.
Requires
.Sy l2arc_feed_again Ns = Ns Ar 1
and only applicable in related situations.
.
.It Sy l2arc_feed_secs Ns = Ns Sy 1 Pq u64
Seconds between L2ARC writing.
.
-.It Sy l2arc_headroom Ns = Ns Sy 2 Pq u64
+.It Sy l2arc_headroom Ns = Ns Sy 8 Pq u64
How far through the ARC lists to search for L2ARC cacheable content,
expressed as a multiplier of
.Sy l2arc_write_max .
ARC persistence across reboots can be achieved with persistent L2ARC
by setting this parameter to
.Sy 0 ,
allowing the full length of ARC lists to be searched for cacheable content.
.
.It Sy l2arc_headroom_boost Ns = Ns Sy 200 Ns % Pq u64
Scales
.Sy l2arc_headroom
by this percentage when L2ARC contents are being successfully compressed
before writing.
A value of
.Sy 100
disables this feature.
.
.It Sy l2arc_exclude_special Ns = Ns Sy 0 Ns | Ns 1 Pq int
Controls whether buffers present on special vdevs are eligible for caching
into L2ARC.
If set to 1, exclude dbufs on special vdevs from being cached to L2ARC.
.
.It Sy l2arc_mfuonly Ns = Ns Sy 0 Ns | Ns 1 Pq int
Controls whether only MFU metadata and data are cached from ARC into L2ARC.
This may be desired to avoid wasting space on L2ARC when reading/writing large
amounts of data that are not expected to be accessed more than once.
.Pp
The default is off,
meaning both MRU and MFU data and metadata are cached.
When turning off this feature, some MRU buffers will still be present
in ARC and eventually cached on L2ARC.
.No If Sy l2arc_noprefetch Ns = Ns Sy 0 ,
some prefetched buffers will be cached to L2ARC, and those might later
transition to MRU, in which case the
.Sy l2arc_mru_asize No arcstat will not be Sy 0 .
.Pp
Regardless of
.Sy l2arc_noprefetch ,
some MFU buffers might be evicted from ARC,
accessed later on as prefetches and transition to MRU as prefetches.
If accessed again they are counted as MRU and the
.Sy l2arc_mru_asize No arcstat will not be Sy 0 .
.Pp
The ARC status of L2ARC buffers when they were first cached in
L2ARC can be seen in the
.Sy l2arc_mru_asize , Sy l2arc_mfu_asize , No and Sy l2arc_prefetch_asize
arcstats when importing the pool or onlining a cache
device if persistent L2ARC is enabled.
.Pp
The
.Sy evict_l2_eligible_mru
arcstat does not take into account if this option is enabled as the information
provided by the
.Sy evict_l2_eligible_m[rf]u
arcstats can be used to decide if toggling this option is appropriate
for the current workload.
.
.It Sy l2arc_meta_percent Ns = Ns Sy 33 Ns % Pq uint
Percent of ARC size allowed for L2ARC-only headers.
Since L2ARC buffers are not evicted on memory pressure,
too many headers on a system with an irrationally large L2ARC
can render it slow or unusable.
This parameter limits L2ARC writes and rebuilds to achieve the target.
.
.It Sy l2arc_trim_ahead Ns = Ns Sy 0 Ns % Pq u64
Trims ahead of the current write size
.Pq Sy l2arc_write_max
on L2ARC devices by this percentage of write size if we have filled the device.
If set to
.Sy 100
we TRIM twice the space required to accommodate upcoming writes.
A minimum of
.Sy 64 MiB
will be trimmed.
It also enables TRIM of the whole L2ARC device upon creation
or addition to an existing pool or if the header of the device is
invalid upon importing a pool or onlining a cache device.
A value of
.Sy 0
disables TRIM on L2ARC altogether and is the default as it can put significant
stress on the underlying storage devices.
This will vary depending of how well the specific device handles these commands.
.
.It Sy l2arc_noprefetch Ns = Ns Sy 1 Ns | Ns 0 Pq int
Do not write buffers to L2ARC if they were prefetched but not used by
applications.
In case there are prefetched buffers in L2ARC and this option
is later set, we do not read the prefetched buffers from L2ARC.
Unsetting this option is useful for caching sequential reads from the
disks to L2ARC and serve those reads from L2ARC later on.
This may be beneficial in case the L2ARC device is significantly faster
in sequential reads than the disks of the pool.
.Pp
Use
.Sy 1
to disable and
.Sy 0
to enable caching/reading prefetches to/from L2ARC.
.
.It Sy l2arc_norw Ns = Ns Sy 0 Ns | Ns 1 Pq int
No reads during writes.
.
-.It Sy l2arc_write_boost Ns = Ns Sy 8388608 Ns B Po 8 MiB Pc Pq u64
+.It Sy l2arc_write_boost Ns = Ns Sy 33554432 Ns B Po 32 MiB Pc Pq u64
Cold L2ARC devices will have
.Sy l2arc_write_max
increased by this amount while they remain cold.
.
-.It Sy l2arc_write_max Ns = Ns Sy 8388608 Ns B Po 8 MiB Pc Pq u64
+.It Sy l2arc_write_max Ns = Ns Sy 33554432 Ns B Po 32 MiB Pc Pq u64
Max write bytes per interval.
.
.It Sy l2arc_rebuild_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Rebuild the L2ARC when importing a pool (persistent L2ARC).
This can be disabled if there are problems importing a pool
or attaching an L2ARC device (e.g. the L2ARC device is slow
in reading stored log metadata, or the metadata
has become somehow fragmented/unusable).
.
.It Sy l2arc_rebuild_blocks_min_l2size Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64
Mininum size of an L2ARC device required in order to write log blocks in it.
The log blocks are used upon importing the pool to rebuild the persistent L2ARC.
.Pp
For L2ARC devices less than 1 GiB, the amount of data
.Fn l2arc_evict
evicts is significant compared to the amount of restored L2ARC data.
In this case, do not write log blocks in L2ARC in order not to waste space.
.
.It Sy metaslab_aliquot Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
Metaslab granularity, in bytes.
This is roughly similar to what would be referred to as the "stripe size"
in traditional RAID arrays.
In normal operation, ZFS will try to write this amount of data to each disk
before moving on to the next top-level vdev.
.
.It Sy metaslab_bias_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable metaslab group biasing based on their vdevs' over- or under-utilization
relative to the pool.
.
.It Sy metaslab_force_ganging Ns = Ns Sy 16777217 Ns B Po 16 MiB + 1 B Pc Pq u64
Make some blocks above a certain size be gang blocks.
This option is used by the test suite to facilitate testing.
.
.It Sy metaslab_force_ganging_pct Ns = Ns Sy 3 Ns % Pq uint
For blocks that could be forced to be a gang block (due to
.Sy metaslab_force_ganging ) ,
force this many of them to be gang blocks.
.
.It Sy zfs_ddt_zap_default_bs Ns = Ns Sy 15 Po 32 KiB Pc Pq int
Default DDT ZAP data block size as a power of 2. Note that changing this after
creating a DDT on the pool will not affect existing DDTs, only newly created
ones.
.
.It Sy zfs_ddt_zap_default_ibs Ns = Ns Sy 15 Po 32 KiB Pc Pq int
Default DDT ZAP indirect block size as a power of 2. Note that changing this
after creating a DDT on the pool will not affect existing DDTs, only newly
created ones.
.
.It Sy zfs_default_bs Ns = Ns Sy 9 Po 512 B Pc Pq int
Default dnode block size as a power of 2.
.
.It Sy zfs_default_ibs Ns = Ns Sy 17 Po 128 KiB Pc Pq int
Default dnode indirect block size as a power of 2.
.
.It Sy zfs_history_output_max Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
When attempting to log an output nvlist of an ioctl in the on-disk history,
the output will not be stored if it is larger than this size (in bytes).
This must be less than
.Sy DMU_MAX_ACCESS Pq 64 MiB .
This applies primarily to
.Fn zfs_ioc_channel_program Pq cf. Xr zfs-program 8 .
.
.It Sy zfs_keep_log_spacemaps_at_export Ns = Ns Sy 0 Ns | Ns 1 Pq int
Prevent log spacemaps from being destroyed during pool exports and destroys.
.
.It Sy zfs_metaslab_segment_weight_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable/disable segment-based metaslab selection.
.
.It Sy zfs_metaslab_switch_threshold Ns = Ns Sy 2 Pq int
When using segment-based metaslab selection, continue allocating
from the active metaslab until this option's
worth of buckets have been exhausted.
.
.It Sy metaslab_debug_load Ns = Ns Sy 0 Ns | Ns 1 Pq int
Load all metaslabs during pool import.
.
.It Sy metaslab_debug_unload Ns = Ns Sy 0 Ns | Ns 1 Pq int
Prevent metaslabs from being unloaded.
.
.It Sy metaslab_fragmentation_factor_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable use of the fragmentation metric in computing metaslab weights.
.
.It Sy metaslab_df_max_search Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint
Maximum distance to search forward from the last offset.
Without this limit, fragmented pools can see
.Em >100`000
iterations and
.Fn metaslab_block_picker
becomes the performance limiting factor on high-performance storage.
.Pp
With the default setting of
.Sy 16 MiB ,
we typically see less than
.Em 500
iterations, even with very fragmented
.Sy ashift Ns = Ns Sy 9
pools.
The maximum number of iterations possible is
.Sy metaslab_df_max_search / 2^(ashift+1) .
With the default setting of
.Sy 16 MiB
this is
.Em 16*1024 Pq with Sy ashift Ns = Ns Sy 9
or
.Em 2*1024 Pq with Sy ashift Ns = Ns Sy 12 .
.
.It Sy metaslab_df_use_largest_segment Ns = Ns Sy 0 Ns | Ns 1 Pq int
If not searching forward (due to
.Sy metaslab_df_max_search , metaslab_df_free_pct ,
.No or Sy metaslab_df_alloc_threshold ) ,
this tunable controls which segment is used.
If set, we will use the largest free segment.
If unset, we will use a segment of at least the requested size.
.
.It Sy zfs_metaslab_max_size_cache_sec Ns = Ns Sy 3600 Ns s Po 1 hour Pc Pq u64
When we unload a metaslab, we cache the size of the largest free chunk.
We use that cached size to determine whether or not to load a metaslab
for a given allocation.
As more frees accumulate in that metaslab while it's unloaded,
the cached max size becomes less and less accurate.
After a number of seconds controlled by this tunable,
we stop considering the cached max size and start
considering only the histogram instead.
.
.It Sy zfs_metaslab_mem_limit Ns = Ns Sy 25 Ns % Pq uint
When we are loading a new metaslab, we check the amount of memory being used
to store metaslab range trees.
If it is over a threshold, we attempt to unload the least recently used metaslab
to prevent the system from clogging all of its memory with range trees.
This tunable sets the percentage of total system memory that is the threshold.
.
.It Sy zfs_metaslab_try_hard_before_gang Ns = Ns Sy 0 Ns | Ns 1 Pq int
.Bl -item -compact
.It
If unset, we will first try normal allocation.
.It
If that fails then we will do a gang allocation.
.It
If that fails then we will do a "try hard" gang allocation.
.It
If that fails then we will have a multi-layer gang block.
.El
.Pp
.Bl -item -compact
.It
If set, we will first try normal allocation.
.It
If that fails then we will do a "try hard" allocation.
.It
If that fails we will do a gang allocation.
.It
If that fails we will do a "try hard" gang allocation.
.It
If that fails then we will have a multi-layer gang block.
.El
.
.It Sy zfs_metaslab_find_max_tries Ns = Ns Sy 100 Pq uint
When not trying hard, we only consider this number of the best metaslabs.
This improves performance, especially when there are many metaslabs per vdev
and the allocation can't actually be satisfied
(so we would otherwise iterate all metaslabs).
.
.It Sy zfs_vdev_default_ms_count Ns = Ns Sy 200 Pq uint
When a vdev is added, target this number of metaslabs per top-level vdev.
.
.It Sy zfs_vdev_default_ms_shift Ns = Ns Sy 29 Po 512 MiB Pc Pq uint
Default lower limit for metaslab size.
.
.It Sy zfs_vdev_max_ms_shift Ns = Ns Sy 34 Po 16 GiB Pc Pq uint
Default upper limit for metaslab size.
.
.It Sy zfs_vdev_max_auto_ashift Ns = Ns Sy 14 Pq uint
Maximum ashift used when optimizing for logical \[->] physical sector size on
new
top-level vdevs.
May be increased up to
.Sy ASHIFT_MAX Po 16 Pc ,
but this may negatively impact pool space efficiency.
.
.It Sy zfs_vdev_min_auto_ashift Ns = Ns Sy ASHIFT_MIN Po 9 Pc Pq uint
Minimum ashift used when creating new top-level vdevs.
.
.It Sy zfs_vdev_min_ms_count Ns = Ns Sy 16 Pq uint
Minimum number of metaslabs to create in a top-level vdev.
.
.It Sy vdev_validate_skip Ns = Ns Sy 0 Ns | Ns 1 Pq int
Skip label validation steps during pool import.
Changing is not recommended unless you know what you're doing
and are recovering a damaged label.
.
.It Sy zfs_vdev_ms_count_limit Ns = Ns Sy 131072 Po 128k Pc Pq uint
Practical upper limit of total metaslabs per top-level vdev.
.
.It Sy metaslab_preload_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable metaslab group preloading.
.
.It Sy metaslab_preload_limit Ns = Ns Sy 10 Pq uint
Maximum number of metaslabs per group to preload
.
.It Sy metaslab_preload_pct Ns = Ns Sy 50 Pq uint
Percentage of CPUs to run a metaslab preload taskq
.
.It Sy metaslab_lba_weighting_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Give more weight to metaslabs with lower LBAs,
assuming they have greater bandwidth,
as is typically the case on a modern constant angular velocity disk drive.
.
.It Sy metaslab_unload_delay Ns = Ns Sy 32 Pq uint
After a metaslab is used, we keep it loaded for this many TXGs, to attempt to
reduce unnecessary reloading.
Note that both this many TXGs and
.Sy metaslab_unload_delay_ms
milliseconds must pass before unloading will occur.
.
.It Sy metaslab_unload_delay_ms Ns = Ns Sy 600000 Ns ms Po 10 min Pc Pq uint
After a metaslab is used, we keep it loaded for this many milliseconds,
to attempt to reduce unnecessary reloading.
Note, that both this many milliseconds and
.Sy metaslab_unload_delay
TXGs must pass before unloading will occur.
.
.It Sy reference_history Ns = Ns Sy 3 Pq uint
Maximum reference holders being tracked when reference_tracking_enable is
active.
+.It Sy raidz_expand_max_copy_bytes Ns = Ns Sy 160MB Pq ulong
+Max amount of memory to use for RAID-Z expansion I/O.
+This limits how much I/O can be outstanding at once.
+.
+.It Sy raidz_expand_max_reflow_bytes Ns = Ns Sy 0 Pq ulong
+For testing, pause RAID-Z expansion when reflow amount reaches this value.
+.
+.It Sy raidz_io_aggregate_rows Ns = Ns Sy 4 Pq ulong
+For expanded RAID-Z, aggregate reads that have more rows than this.
+.
+.It Sy reference_history Ns = Ns Sy 3 Pq int
+Maximum reference holders being tracked when reference_tracking_enable is
+active.
.
.It Sy reference_tracking_enable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Track reference holders to
.Sy refcount_t
objects (debug builds only).
.
.It Sy send_holes_without_birth_time Ns = Ns Sy 1 Ns | Ns 0 Pq int
When set, the
.Sy hole_birth
optimization will not be used, and all holes will always be sent during a
.Nm zfs Cm send .
This is useful if you suspect your datasets are affected by a bug in
.Sy hole_birth .
.
.It Sy spa_config_path Ns = Ns Pa /etc/zfs/zpool.cache Pq charp
SPA config file.
.
.It Sy spa_asize_inflation Ns = Ns Sy 24 Pq uint
Multiplication factor used to estimate actual disk consumption from the
size of data being written.
The default value is a worst case estimate,
but lower values may be valid for a given pool depending on its configuration.
Pool administrators who understand the factors involved
may wish to specify a more realistic inflation factor,
particularly if they operate close to quota or capacity limits.
.
.It Sy spa_load_print_vdev_tree Ns = Ns Sy 0 Ns | Ns 1 Pq int
Whether to print the vdev tree in the debugging message buffer during pool
import.
.
.It Sy spa_load_verify_data Ns = Ns Sy 1 Ns | Ns 0 Pq int
Whether to traverse data blocks during an "extreme rewind"
.Pq Fl X
import.
.Pp
An extreme rewind import normally performs a full traversal of all
blocks in the pool for verification.
If this parameter is unset, the traversal skips non-metadata blocks.
It can be toggled once the
import has started to stop or start the traversal of non-metadata blocks.
.
.It Sy spa_load_verify_metadata Ns = Ns Sy 1 Ns | Ns 0 Pq int
Whether to traverse blocks during an "extreme rewind"
.Pq Fl X
pool import.
.Pp
An extreme rewind import normally performs a full traversal of all
blocks in the pool for verification.
If this parameter is unset, the traversal is not performed.
It can be toggled once the import has started to stop or start the traversal.
.
.It Sy spa_load_verify_shift Ns = Ns Sy 4 Po 1/16th Pc Pq uint
Sets the maximum number of bytes to consume during pool import to the log2
fraction of the target ARC size.
.
.It Sy spa_slop_shift Ns = Ns Sy 5 Po 1/32nd Pc Pq int
Normally, we don't allow the last
.Sy 3.2% Pq Sy 1/2^spa_slop_shift
of space in the pool to be consumed.
This ensures that we don't run the pool completely out of space,
due to unaccounted changes (e.g. to the MOS).
It also limits the worst-case time to allocate space.
If we have less than this amount of free space,
most ZPL operations (e.g. write, create) will return
.Sy ENOSPC .
.
.It Sy spa_num_allocators Ns = Ns Sy 4 Pq int
Determines the number of block alloctators to use per spa instance.
Capped by the number of actual CPUs in the system.
.Pp
Note that setting this value too high could result in performance
degredation and/or excess fragmentation.
.
.It Sy spa_upgrade_errlog_limit Ns = Ns Sy 0 Pq uint
Limits the number of on-disk error log entries that will be converted to the
new format when enabling the
.Sy head_errlog
feature.
The default is to convert all log entries.
.
.It Sy vdev_removal_max_span Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint
During top-level vdev removal, chunks of data are copied from the vdev
which may include free space in order to trade bandwidth for IOPS.
This parameter determines the maximum span of free space, in bytes,
which will be included as "unnecessary" data in a chunk of copied data.
.Pp
The default value here was chosen to align with
.Sy zfs_vdev_read_gap_limit ,
which is a similar concept when doing
regular reads (but there's no reason it has to be the same).
.
.It Sy vdev_file_logical_ashift Ns = Ns Sy 9 Po 512 B Pc Pq u64
Logical ashift for file-based devices.
.
.It Sy vdev_file_physical_ashift Ns = Ns Sy 9 Po 512 B Pc Pq u64
Physical ashift for file-based devices.
.
.It Sy zap_iterate_prefetch Ns = Ns Sy 1 Ns | Ns 0 Pq int
If set, when we start iterating over a ZAP object,
prefetch the entire object (all leaf blocks).
However, this is limited by
.Sy dmu_prefetch_max .
.
.It Sy zap_micro_max_size Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq int
Maximum micro ZAP size.
A micro ZAP is upgraded to a fat ZAP, once it grows beyond the specified size.
.
.It Sy zfetch_min_distance Ns = Ns Sy 4194304 Ns B Po 4 MiB Pc Pq uint
Min bytes to prefetch per stream.
Prefetch distance starts from the demand access size and quickly grows to
this value, doubling on each hit.
After that it may grow further by 1/8 per hit, but only if some prefetch
since last time haven't completed in time to satisfy demand request, i.e.
prefetch depth didn't cover the read latency or the pool got saturated.
.
.It Sy zfetch_max_distance Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq uint
Max bytes to prefetch per stream.
.
.It Sy zfetch_max_idistance Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq uint
Max bytes to prefetch indirects for per stream.
.
.It Sy zfetch_max_streams Ns = Ns Sy 8 Pq uint
Max number of streams per zfetch (prefetch streams per file).
.
.It Sy zfetch_min_sec_reap Ns = Ns Sy 1 Pq uint
Min time before inactive prefetch stream can be reclaimed
.
.It Sy zfetch_max_sec_reap Ns = Ns Sy 2 Pq uint
Max time before inactive prefetch stream can be deleted
.
.It Sy zfs_abd_scatter_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enables ARC from using scatter/gather lists and forces all allocations to be
linear in kernel memory.
Disabling can improve performance in some code paths
at the expense of fragmented kernel memory.
.
.It Sy zfs_abd_scatter_max_order Ns = Ns Sy MAX_ORDER\-1 Pq uint
Maximum number of consecutive memory pages allocated in a single block for
scatter/gather lists.
.Pp
The value of
.Sy MAX_ORDER
depends on kernel configuration.
.
.It Sy zfs_abd_scatter_min_size Ns = Ns Sy 1536 Ns B Po 1.5 KiB Pc Pq uint
This is the minimum allocation size that will use scatter (page-based) ABDs.
Smaller allocations will use linear ABDs.
.
.It Sy zfs_arc_dnode_limit Ns = Ns Sy 0 Ns B Pq u64
When the number of bytes consumed by dnodes in the ARC exceeds this number of
bytes, try to unpin some of it in response to demand for non-metadata.
This value acts as a ceiling to the amount of dnode metadata, and defaults to
.Sy 0 ,
which indicates that a percent which is based on
.Sy zfs_arc_dnode_limit_percent
of the ARC meta buffers that may be used for dnodes.
.It Sy zfs_arc_dnode_limit_percent Ns = Ns Sy 10 Ns % Pq u64
Percentage that can be consumed by dnodes of ARC meta buffers.
.Pp
See also
.Sy zfs_arc_dnode_limit ,
which serves a similar purpose but has a higher priority if nonzero.
.
.It Sy zfs_arc_dnode_reduce_percent Ns = Ns Sy 10 Ns % Pq u64
Percentage of ARC dnodes to try to scan in response to demand for non-metadata
when the number of bytes consumed by dnodes exceeds
.Sy zfs_arc_dnode_limit .
.
.It Sy zfs_arc_average_blocksize Ns = Ns Sy 8192 Ns B Po 8 KiB Pc Pq uint
The ARC's buffer hash table is sized based on the assumption of an average
block size of this value.
This works out to roughly 1 MiB of hash table per 1 GiB of physical memory
with 8-byte pointers.
For configurations with a known larger average block size,
this value can be increased to reduce the memory footprint.
.
.It Sy zfs_arc_eviction_pct Ns = Ns Sy 200 Ns % Pq uint
When
.Fn arc_is_overflowing ,
.Fn arc_get_data_impl
waits for this percent of the requested amount of data to be evicted.
For example, by default, for every
.Em 2 KiB
that's evicted,
.Em 1 KiB
of it may be "reused" by a new allocation.
Since this is above
.Sy 100 Ns % ,
it ensures that progress is made towards getting
.Sy arc_size No under Sy arc_c .
Since this is finite, it ensures that allocations can still happen,
even during the potentially long time that
.Sy arc_size No is more than Sy arc_c .
.
.It Sy zfs_arc_evict_batch_limit Ns = Ns Sy 10 Pq uint
Number ARC headers to evict per sub-list before proceeding to another sub-list.
This batch-style operation prevents entire sub-lists from being evicted at once
but comes at a cost of additional unlocking and locking.
.
.It Sy zfs_arc_grow_retry Ns = Ns Sy 0 Ns s Pq uint
If set to a non zero value, it will replace the
.Sy arc_grow_retry
value with this value.
The
.Sy arc_grow_retry
.No value Pq default Sy 5 Ns s
is the number of seconds the ARC will wait before
trying to resume growth after a memory pressure event.
.
.It Sy zfs_arc_lotsfree_percent Ns = Ns Sy 10 Ns % Pq int
Throttle I/O when free system memory drops below this percentage of total
system memory.
Setting this value to
.Sy 0
will disable the throttle.
.
.It Sy zfs_arc_max Ns = Ns Sy 0 Ns B Pq u64
Max size of ARC in bytes.
If
.Sy 0 ,
then the max size of ARC is determined by the amount of system memory installed.
The larger of
.Sy all_system_memory No \- Sy 1 GiB
and
.Sy 5/8 No \(mu Sy all_system_memory
will be used as the limit.
This value must be at least
.Sy 67108864 Ns B Pq 64 MiB .
.Pp
This value can be changed dynamically, with some caveats.
It cannot be set back to
.Sy 0
while running, and reducing it below the current ARC size will not cause
the ARC to shrink without memory pressure to induce shrinking.
.
.It Sy zfs_arc_meta_balance Ns = Ns Sy 500 Pq uint
Balance between metadata and data on ghost hits.
Values above 100 increase metadata caching by proportionally reducing effect
of ghost data hits on target data/metadata rate.
.
.It Sy zfs_arc_min Ns = Ns Sy 0 Ns B Pq u64
Min size of ARC in bytes.
.No If set to Sy 0 , arc_c_min
will default to consuming the larger of
.Sy 32 MiB
and
.Sy all_system_memory No / Sy 32 .
.
.It Sy zfs_arc_min_prefetch_ms Ns = Ns Sy 0 Ns ms Ns Po Ns ≡ Ns 1s Pc Pq uint
Minimum time prefetched blocks are locked in the ARC.
.
.It Sy zfs_arc_min_prescient_prefetch_ms Ns = Ns Sy 0 Ns ms Ns Po Ns ≡ Ns 6s Pc Pq uint
Minimum time "prescient prefetched" blocks are locked in the ARC.
These blocks are meant to be prefetched fairly aggressively ahead of
the code that may use them.
.
.It Sy zfs_arc_prune_task_threads Ns = Ns Sy 1 Pq int
Number of arc_prune threads.
.Fx
does not need more than one.
Linux may theoretically use one per mount point up to number of CPUs,
but that was not proven to be useful.
.
.It Sy zfs_max_missing_tvds Ns = Ns Sy 0 Pq int
Number of missing top-level vdevs which will be allowed during
pool import (only in read-only mode).
.
.It Sy zfs_max_nvlist_src_size Ns = Sy 0 Pq u64
Maximum size in bytes allowed to be passed as
.Sy zc_nvlist_src_size
for ioctls on
.Pa /dev/zfs .
This prevents a user from causing the kernel to allocate
an excessive amount of memory.
When the limit is exceeded, the ioctl fails with
.Sy EINVAL
and a description of the error is sent to the
.Pa zfs-dbgmsg
log.
This parameter should not need to be touched under normal circumstances.
If
.Sy 0 ,
equivalent to a quarter of the user-wired memory limit under
.Fx
and to
.Sy 134217728 Ns B Pq 128 MiB
under Linux.
.
.It Sy zfs_multilist_num_sublists Ns = Ns Sy 0 Pq uint
To allow more fine-grained locking, each ARC state contains a series
of lists for both data and metadata objects.
Locking is performed at the level of these "sub-lists".
This parameters controls the number of sub-lists per ARC state,
and also applies to other uses of the multilist data structure.
.Pp
If
.Sy 0 ,
equivalent to the greater of the number of online CPUs and
.Sy 4 .
.
.It Sy zfs_arc_overflow_shift Ns = Ns Sy 8 Pq int
The ARC size is considered to be overflowing if it exceeds the current
ARC target size
.Pq Sy arc_c
by thresholds determined by this parameter.
Exceeding by
.Sy ( arc_c No >> Sy zfs_arc_overflow_shift ) No / Sy 2
starts ARC reclamation process.
If that appears insufficient, exceeding by
.Sy ( arc_c No >> Sy zfs_arc_overflow_shift ) No \(mu Sy 1.5
blocks new buffer allocation until the reclaim thread catches up.
Started reclamation process continues till ARC size returns below the
target size.
.Pp
The default value of
.Sy 8
causes the ARC to start reclamation if it exceeds the target size by
.Em 0.2%
of the target size, and block allocations by
.Em 0.6% .
.
.It Sy zfs_arc_shrink_shift Ns = Ns Sy 0 Pq uint
If nonzero, this will update
.Sy arc_shrink_shift Pq default Sy 7
with the new value.
.
.It Sy zfs_arc_pc_percent Ns = Ns Sy 0 Ns % Po off Pc Pq uint
Percent of pagecache to reclaim ARC to.
.Pp
This tunable allows the ZFS ARC to play more nicely
with the kernel's LRU pagecache.
It can guarantee that the ARC size won't collapse under scanning
pressure on the pagecache, yet still allows the ARC to be reclaimed down to
.Sy zfs_arc_min
if necessary.
This value is specified as percent of pagecache size (as measured by
.Sy NR_FILE_PAGES ) ,
where that percent may exceed
.Sy 100 .
This
only operates during memory pressure/reclaim.
.
.It Sy zfs_arc_shrinker_limit Ns = Ns Sy 10000 Pq int
This is a limit on how many pages the ARC shrinker makes available for
eviction in response to one page allocation attempt.
Note that in practice, the kernel's shrinker can ask us to evict
up to about four times this for one allocation attempt.
.Pp
The default limit of
.Sy 10000 Pq in practice, Em 160 MiB No per allocation attempt with 4 KiB pages
limits the amount of time spent attempting to reclaim ARC memory to
less than 100 ms per allocation attempt,
even with a small average compressed block size of ~8 KiB.
.Pp
The parameter can be set to 0 (zero) to disable the limit,
and only applies on Linux.
.
.It Sy zfs_arc_sys_free Ns = Ns Sy 0 Ns B Pq u64
The target number of bytes the ARC should leave as free memory on the system.
If zero, equivalent to the bigger of
.Sy 512 KiB No and Sy all_system_memory/64 .
.
.It Sy zfs_autoimport_disable Ns = Ns Sy 1 Ns | Ns 0 Pq int
Disable pool import at module load by ignoring the cache file
.Pq Sy spa_config_path .
.
.It Sy zfs_checksum_events_per_second Ns = Ns Sy 20 Ns /s Pq uint
Rate limit checksum events to this many per second.
Note that this should not be set below the ZED thresholds
(currently 10 checksums over 10 seconds)
or else the daemon may not trigger any action.
.
.It Sy zfs_commit_timeout_pct Ns = Ns Sy 10 Ns % Pq uint
This controls the amount of time that a ZIL block (lwb) will remain "open"
when it isn't "full", and it has a thread waiting for it to be committed to
stable storage.
The timeout is scaled based on a percentage of the last lwb
latency to avoid significantly impacting the latency of each individual
transaction record (itx).
.
.It Sy zfs_condense_indirect_commit_entry_delay_ms Ns = Ns Sy 0 Ns ms Pq int
Vdev indirection layer (used for device removal) sleeps for this many
milliseconds during mapping generation.
Intended for use with the test suite to throttle vdev removal speed.
.
.It Sy zfs_condense_indirect_obsolete_pct Ns = Ns Sy 25 Ns % Pq uint
Minimum percent of obsolete bytes in vdev mapping required to attempt to
condense
.Pq see Sy zfs_condense_indirect_vdevs_enable .
Intended for use with the test suite
to facilitate triggering condensing as needed.
.
.It Sy zfs_condense_indirect_vdevs_enable Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable condensing indirect vdev mappings.
When set, attempt to condense indirect vdev mappings
if the mapping uses more than
.Sy zfs_condense_min_mapping_bytes
bytes of memory and if the obsolete space map object uses more than
.Sy zfs_condense_max_obsolete_bytes
bytes on-disk.
The condensing process is an attempt to save memory by removing obsolete
mappings.
.
.It Sy zfs_condense_max_obsolete_bytes Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64
Only attempt to condense indirect vdev mappings if the on-disk size
of the obsolete space map object is greater than this number of bytes
.Pq see Sy zfs_condense_indirect_vdevs_enable .
.
.It Sy zfs_condense_min_mapping_bytes Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq u64
Minimum size vdev mapping to attempt to condense
.Pq see Sy zfs_condense_indirect_vdevs_enable .
.
.It Sy zfs_dbgmsg_enable Ns = Ns Sy 1 Ns | Ns 0 Pq int
Internally ZFS keeps a small log to facilitate debugging.
The log is enabled by default, and can be disabled by unsetting this option.
The contents of the log can be accessed by reading
.Pa /proc/spl/kstat/zfs/dbgmsg .
Writing
.Sy 0
to the file clears the log.
.Pp
This setting does not influence debug prints due to
.Sy zfs_flags .
.
.It Sy zfs_dbgmsg_maxsize Ns = Ns Sy 4194304 Ns B Po 4 MiB Pc Pq uint
Maximum size of the internal ZFS debug log.
.
.It Sy zfs_dbuf_state_index Ns = Ns Sy 0 Pq int
Historically used for controlling what reporting was available under
.Pa /proc/spl/kstat/zfs .
No effect.
.
.It Sy zfs_deadman_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
When a pool sync operation takes longer than
.Sy zfs_deadman_synctime_ms ,
or when an individual I/O operation takes longer than
.Sy zfs_deadman_ziotime_ms ,
then the operation is considered to be "hung".
If
.Sy zfs_deadman_enabled
is set, then the deadman behavior is invoked as described by
.Sy zfs_deadman_failmode .
By default, the deadman is enabled and set to
.Sy wait
which results in "hung" I/O operations only being logged.
The deadman is automatically disabled when a pool gets suspended.
.
.It Sy zfs_deadman_failmode Ns = Ns Sy wait Pq charp
Controls the failure behavior when the deadman detects a "hung" I/O operation.
Valid values are:
.Bl -tag -compact -offset 4n -width "continue"
.It Sy wait
Wait for a "hung" operation to complete.
For each "hung" operation a "deadman" event will be posted
describing that operation.
.It Sy continue
Attempt to recover from a "hung" operation by re-dispatching it
to the I/O pipeline if possible.
.It Sy panic
Panic the system.
This can be used to facilitate automatic fail-over
to a properly configured fail-over partner.
.El
.
.It Sy zfs_deadman_checktime_ms Ns = Ns Sy 60000 Ns ms Po 1 min Pc Pq u64
Check time in milliseconds.
This defines the frequency at which we check for hung I/O requests
and potentially invoke the
.Sy zfs_deadman_failmode
behavior.
.
.It Sy zfs_deadman_synctime_ms Ns = Ns Sy 600000 Ns ms Po 10 min Pc Pq u64
Interval in milliseconds after which the deadman is triggered and also
the interval after which a pool sync operation is considered to be "hung".
Once this limit is exceeded the deadman will be invoked every
.Sy zfs_deadman_checktime_ms
milliseconds until the pool sync completes.
.
.It Sy zfs_deadman_ziotime_ms Ns = Ns Sy 300000 Ns ms Po 5 min Pc Pq u64
Interval in milliseconds after which the deadman is triggered and an
individual I/O operation is considered to be "hung".
As long as the operation remains "hung",
the deadman will be invoked every
.Sy zfs_deadman_checktime_ms
milliseconds until the operation completes.
.
.It Sy zfs_dedup_prefetch Ns = Ns Sy 0 Ns | Ns 1 Pq int
Enable prefetching dedup-ed blocks which are going to be freed.
.
.It Sy zfs_delay_min_dirty_percent Ns = Ns Sy 60 Ns % Pq uint
Start to delay each transaction once there is this amount of dirty data,
expressed as a percentage of
.Sy zfs_dirty_data_max .
This value should be at least
.Sy zfs_vdev_async_write_active_max_dirty_percent .
.No See Sx ZFS TRANSACTION DELAY .
.
.It Sy zfs_delay_scale Ns = Ns Sy 500000 Pq int
This controls how quickly the transaction delay approaches infinity.
Larger values cause longer delays for a given amount of dirty data.
.Pp
For the smoothest delay, this value should be about 1 billion divided
by the maximum number of operations per second.
This will smoothly handle between ten times and a tenth of this number.
.No See Sx ZFS TRANSACTION DELAY .
.Pp
.Sy zfs_delay_scale No \(mu Sy zfs_dirty_data_max Em must No be smaller than Sy 2^64 .
.
.It Sy zfs_disable_ivset_guid_check Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disables requirement for IVset GUIDs to be present and match when doing a raw
receive of encrypted datasets.
Intended for users whose pools were created with
OpenZFS pre-release versions and now have compatibility issues.
.
.It Sy zfs_key_max_salt_uses Ns = Ns Sy 400000000 Po 4*10^8 Pc Pq ulong
Maximum number of uses of a single salt value before generating a new one for
encrypted datasets.
The default value is also the maximum.
.
.It Sy zfs_object_mutex_size Ns = Ns Sy 64 Pq uint
Size of the znode hashtable used for holds.
.Pp
Due to the need to hold locks on objects that may not exist yet, kernel mutexes
are not created per-object and instead a hashtable is used where collisions
will result in objects waiting when there is not actually contention on the
same object.
.
.It Sy zfs_slow_io_events_per_second Ns = Ns Sy 20 Ns /s Pq int
Rate limit delay and deadman zevents (which report slow I/O operations) to this
many per
second.
.
.It Sy zfs_unflushed_max_mem_amt Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64
Upper-bound limit for unflushed metadata changes to be held by the
log spacemap in memory, in bytes.
.
.It Sy zfs_unflushed_max_mem_ppm Ns = Ns Sy 1000 Ns ppm Po 0.1% Pc Pq u64
Part of overall system memory that ZFS allows to be used
for unflushed metadata changes by the log spacemap, in millionths.
.
.It Sy zfs_unflushed_log_block_max Ns = Ns Sy 131072 Po 128k Pc Pq u64
Describes the maximum number of log spacemap blocks allowed for each pool.
The default value means that the space in all the log spacemaps
can add up to no more than
.Sy 131072
blocks (which means
.Em 16 GiB
of logical space before compression and ditto blocks,
assuming that blocksize is
.Em 128 KiB ) .
.Pp
This tunable is important because it involves a trade-off between import
time after an unclean export and the frequency of flushing metaslabs.
The higher this number is, the more log blocks we allow when the pool is
active which means that we flush metaslabs less often and thus decrease
the number of I/O operations for spacemap updates per TXG.
At the same time though, that means that in the event of an unclean export,
there will be more log spacemap blocks for us to read, inducing overhead
in the import time of the pool.
The lower the number, the amount of flushing increases, destroying log
blocks quicker as they become obsolete faster, which leaves less blocks
to be read during import time after a crash.
.Pp
Each log spacemap block existing during pool import leads to approximately
one extra logical I/O issued.
This is the reason why this tunable is exposed in terms of blocks rather
than space used.
.
.It Sy zfs_unflushed_log_block_min Ns = Ns Sy 1000 Pq u64
If the number of metaslabs is small and our incoming rate is high,
we could get into a situation that we are flushing all our metaslabs every TXG.
Thus we always allow at least this many log blocks.
.
.It Sy zfs_unflushed_log_block_pct Ns = Ns Sy 400 Ns % Pq u64
Tunable used to determine the number of blocks that can be used for
the spacemap log, expressed as a percentage of the total number of
unflushed metaslabs in the pool.
.
.It Sy zfs_unflushed_log_txg_max Ns = Ns Sy 1000 Pq u64
Tunable limiting maximum time in TXGs any metaslab may remain unflushed.
It effectively limits maximum number of unflushed per-TXG spacemap logs
that need to be read after unclean pool export.
.
.It Sy zfs_unlink_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq uint
When enabled, files will not be asynchronously removed from the list of pending
unlinks and the space they consume will be leaked.
Once this option has been disabled and the dataset is remounted,
the pending unlinks will be processed and the freed space returned to the pool.
This option is used by the test suite.
.
.It Sy zfs_delete_blocks Ns = Ns Sy 20480 Pq ulong
This is the used to define a large file for the purposes of deletion.
Files containing more than
.Sy zfs_delete_blocks
will be deleted asynchronously, while smaller files are deleted synchronously.
Decreasing this value will reduce the time spent in an
.Xr unlink 2
system call, at the expense of a longer delay before the freed space is
available.
This only applies on Linux.
.
.It Sy zfs_dirty_data_max Ns = Pq int
Determines the dirty space limit in bytes.
Once this limit is exceeded, new writes are halted until space frees up.
This parameter takes precedence over
.Sy zfs_dirty_data_max_percent .
.No See Sx ZFS TRANSACTION DELAY .
.Pp
Defaults to
.Sy physical_ram/10 ,
capped at
.Sy zfs_dirty_data_max_max .
.
.It Sy zfs_dirty_data_max_max Ns = Pq int
Maximum allowable value of
.Sy zfs_dirty_data_max ,
expressed in bytes.
This limit is only enforced at module load time, and will be ignored if
.Sy zfs_dirty_data_max
is later changed.
This parameter takes precedence over
.Sy zfs_dirty_data_max_max_percent .
.No See Sx ZFS TRANSACTION DELAY .
.Pp
Defaults to
.Sy min(physical_ram/4, 4GiB) ,
or
.Sy min(physical_ram/4, 1GiB)
for 32-bit systems.
.
.It Sy zfs_dirty_data_max_max_percent Ns = Ns Sy 25 Ns % Pq uint
Maximum allowable value of
.Sy zfs_dirty_data_max ,
expressed as a percentage of physical RAM.
This limit is only enforced at module load time, and will be ignored if
.Sy zfs_dirty_data_max
is later changed.
The parameter
.Sy zfs_dirty_data_max_max
takes precedence over this one.
.No See Sx ZFS TRANSACTION DELAY .
.
.It Sy zfs_dirty_data_max_percent Ns = Ns Sy 10 Ns % Pq uint
Determines the dirty space limit, expressed as a percentage of all memory.
Once this limit is exceeded, new writes are halted until space frees up.
The parameter
.Sy zfs_dirty_data_max
takes precedence over this one.
.No See Sx ZFS TRANSACTION DELAY .
.Pp
Subject to
.Sy zfs_dirty_data_max_max .
.
.It Sy zfs_dirty_data_sync_percent Ns = Ns Sy 20 Ns % Pq uint
Start syncing out a transaction group if there's at least this much dirty data
.Pq as a percentage of Sy zfs_dirty_data_max .
This should be less than
.Sy zfs_vdev_async_write_active_min_dirty_percent .
.
.It Sy zfs_wrlog_data_max Ns = Pq int
The upper limit of write-transaction zil log data size in bytes.
Write operations are throttled when approaching the limit until log data is
cleared out after transaction group sync.
Because of some overhead, it should be set at least 2 times the size of
.Sy zfs_dirty_data_max
.No to prevent harming normal write throughput .
It also should be smaller than the size of the slog device if slog is present.
.Pp
Defaults to
.Sy zfs_dirty_data_max*2
.
.It Sy zfs_fallocate_reserve_percent Ns = Ns Sy 110 Ns % Pq uint
Since ZFS is a copy-on-write filesystem with snapshots, blocks cannot be
preallocated for a file in order to guarantee that later writes will not
run out of space.
Instead,
.Xr fallocate 2
space preallocation only checks that sufficient space is currently available
in the pool or the user's project quota allocation,
and then creates a sparse file of the requested size.
The requested space is multiplied by
.Sy zfs_fallocate_reserve_percent
to allow additional space for indirect blocks and other internal metadata.
Setting this to
.Sy 0
disables support for
.Xr fallocate 2
and causes it to return
.Sy EOPNOTSUPP .
.
.It Sy zfs_fletcher_4_impl Ns = Ns Sy fastest Pq string
Select a fletcher 4 implementation.
.Pp
Supported selectors are:
.Sy fastest , scalar , sse2 , ssse3 , avx2 , avx512f , avx512bw ,
.No and Sy aarch64_neon .
All except
.Sy fastest No and Sy scalar
require instruction set extensions to be available,
and will only appear if ZFS detects that they are present at runtime.
If multiple implementations of fletcher 4 are available, the
.Sy fastest
will be chosen using a micro benchmark.
Selecting
.Sy scalar
results in the original CPU-based calculation being used.
Selecting any option other than
.Sy fastest No or Sy scalar
results in vector instructions
from the respective CPU instruction set being used.
.
.It Sy zfs_blake3_impl Ns = Ns Sy fastest Pq string
Select a BLAKE3 implementation.
.Pp
Supported selectors are:
.Sy cycle , fastest , generic , sse2 , sse41 , avx2 , avx512 .
All except
.Sy cycle , fastest No and Sy generic
require instruction set extensions to be available,
and will only appear if ZFS detects that they are present at runtime.
If multiple implementations of BLAKE3 are available, the
.Sy fastest will be chosen using a micro benchmark. You can see the
benchmark results by reading this kstat file:
.Pa /proc/spl/kstat/zfs/chksum_bench .
.
.It Sy zfs_free_bpobj_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable/disable the processing of the free_bpobj object.
.
.It Sy zfs_async_block_max_blocks Ns = Ns Sy UINT64_MAX Po unlimited Pc Pq u64
Maximum number of blocks freed in a single TXG.
.
.It Sy zfs_max_async_dedup_frees Ns = Ns Sy 100000 Po 10^5 Pc Pq u64
Maximum number of dedup blocks freed in a single TXG.
.
.It Sy zfs_vdev_async_read_max_active Ns = Ns Sy 3 Pq uint
Maximum asynchronous read I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_async_read_min_active Ns = Ns Sy 1 Pq uint
Minimum asynchronous read I/O operation active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_async_write_active_max_dirty_percent Ns = Ns Sy 60 Ns % Pq uint
When the pool has more than this much dirty data, use
.Sy zfs_vdev_async_write_max_active
to limit active async writes.
If the dirty data is between the minimum and maximum,
the active I/O limit is linearly interpolated.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_async_write_active_min_dirty_percent Ns = Ns Sy 30 Ns % Pq uint
When the pool has less than this much dirty data, use
.Sy zfs_vdev_async_write_min_active
to limit active async writes.
If the dirty data is between the minimum and maximum,
the active I/O limit is linearly
interpolated.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_async_write_max_active Ns = Ns Sy 10 Pq uint
Maximum asynchronous write I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_async_write_min_active Ns = Ns Sy 2 Pq uint
Minimum asynchronous write I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.Pp
Lower values are associated with better latency on rotational media but poorer
resilver performance.
The default value of
.Sy 2
was chosen as a compromise.
A value of
.Sy 3
has been shown to improve resilver performance further at a cost of
further increasing latency.
.
.It Sy zfs_vdev_initializing_max_active Ns = Ns Sy 1 Pq uint
Maximum initializing I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_initializing_min_active Ns = Ns Sy 1 Pq uint
Minimum initializing I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_max_active Ns = Ns Sy 1000 Pq uint
The maximum number of I/O operations active to each device.
Ideally, this will be at least the sum of each queue's
.Sy max_active .
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_open_timeout_ms Ns = Ns Sy 1000 Pq uint
Timeout value to wait before determining a device is missing
during import.
This is helpful for transient missing paths due
to links being briefly removed and recreated in response to
udev events.
.
.It Sy zfs_vdev_rebuild_max_active Ns = Ns Sy 3 Pq uint
Maximum sequential resilver I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_rebuild_min_active Ns = Ns Sy 1 Pq uint
Minimum sequential resilver I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_removal_max_active Ns = Ns Sy 2 Pq uint
Maximum removal I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_removal_min_active Ns = Ns Sy 1 Pq uint
Minimum removal I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_scrub_max_active Ns = Ns Sy 2 Pq uint
Maximum scrub I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_scrub_min_active Ns = Ns Sy 1 Pq uint
Minimum scrub I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_sync_read_max_active Ns = Ns Sy 10 Pq uint
Maximum synchronous read I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_sync_read_min_active Ns = Ns Sy 10 Pq uint
Minimum synchronous read I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_sync_write_max_active Ns = Ns Sy 10 Pq uint
Maximum synchronous write I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_sync_write_min_active Ns = Ns Sy 10 Pq uint
Minimum synchronous write I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_trim_max_active Ns = Ns Sy 2 Pq uint
Maximum trim/discard I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_trim_min_active Ns = Ns Sy 1 Pq uint
Minimum trim/discard I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_nia_delay Ns = Ns Sy 5 Pq uint
For non-interactive I/O (scrub, resilver, removal, initialize and rebuild),
the number of concurrently-active I/O operations is limited to
.Sy zfs_*_min_active ,
unless the vdev is "idle".
When there are no interactive I/O operations active (synchronous or otherwise),
and
.Sy zfs_vdev_nia_delay
operations have completed since the last interactive operation,
then the vdev is considered to be "idle",
and the number of concurrently-active non-interactive operations is increased to
.Sy zfs_*_max_active .
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_nia_credit Ns = Ns Sy 5 Pq uint
Some HDDs tend to prioritize sequential I/O so strongly, that concurrent
random I/O latency reaches several seconds.
On some HDDs this happens even if sequential I/O operations
are submitted one at a time, and so setting
.Sy zfs_*_max_active Ns = Sy 1
does not help.
To prevent non-interactive I/O, like scrub,
from monopolizing the device, no more than
.Sy zfs_vdev_nia_credit operations can be sent
while there are outstanding incomplete interactive operations.
This enforced wait ensures the HDD services the interactive I/O
within a reasonable amount of time.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_queue_depth_pct Ns = Ns Sy 1000 Ns % Pq uint
Maximum number of queued allocations per top-level vdev expressed as
a percentage of
.Sy zfs_vdev_async_write_max_active ,
which allows the system to detect devices that are more capable
of handling allocations and to allocate more blocks to those devices.
This allows for dynamic allocation distribution when devices are imbalanced,
as fuller devices will tend to be slower than empty devices.
.Pp
Also see
.Sy zio_dva_throttle_enabled .
.
.It Sy zfs_vdev_def_queue_depth Ns = Ns Sy 32 Pq uint
Default queue depth for each vdev IO allocator.
Higher values allow for better coalescing of sequential writes before sending
them to the disk, but can increase transaction commit times.
.
.It Sy zfs_vdev_failfast_mask Ns = Ns Sy 1 Pq uint
Defines if the driver should retire on a given error type.
The following options may be bitwise-ored together:
.TS
box;
lbz r l l .
Value Name Description
_
1 Device No driver retries on device errors
2 Transport No driver retries on transport errors.
4 Driver No driver retries on driver errors.
.TE
.
.It Sy zfs_expire_snapshot Ns = Ns Sy 300 Ns s Pq int
Time before expiring
.Pa .zfs/snapshot .
.
.It Sy zfs_admin_snapshot Ns = Ns Sy 0 Ns | Ns 1 Pq int
Allow the creation, removal, or renaming of entries in the
.Sy .zfs/snapshot
directory to cause the creation, destruction, or renaming of snapshots.
When enabled, this functionality works both locally and over NFS exports
which have the
.Em no_root_squash
option set.
.
.It Sy zfs_flags Ns = Ns Sy 0 Pq int
Set additional debugging flags.
The following flags may be bitwise-ored together:
.TS
box;
lbz r l l .
Value Name Description
_
1 ZFS_DEBUG_DPRINTF Enable dprintf entries in the debug log.
* 2 ZFS_DEBUG_DBUF_VERIFY Enable extra dbuf verifications.
* 4 ZFS_DEBUG_DNODE_VERIFY Enable extra dnode verifications.
8 ZFS_DEBUG_SNAPNAMES Enable snapshot name verification.
* 16 ZFS_DEBUG_MODIFY Check for illegally modified ARC buffers.
64 ZFS_DEBUG_ZIO_FREE Enable verification of block frees.
128 ZFS_DEBUG_HISTOGRAM_VERIFY Enable extra spacemap histogram verifications.
256 ZFS_DEBUG_METASLAB_VERIFY Verify space accounting on disk matches in-memory \fBrange_trees\fP.
512 ZFS_DEBUG_SET_ERROR Enable \fBSET_ERROR\fP and dprintf entries in the debug log.
1024 ZFS_DEBUG_INDIRECT_REMAP Verify split blocks created by device removal.
2048 ZFS_DEBUG_TRIM Verify TRIM ranges are always within the allocatable range tree.
4096 ZFS_DEBUG_LOG_SPACEMAP Verify that the log summary is consistent with the spacemap log
and enable \fBzfs_dbgmsgs\fP for metaslab loading and flushing.
.TE
.Sy \& * No Requires debug build .
.
.It Sy zfs_btree_verify_intensity Ns = Ns Sy 0 Pq uint
Enables btree verification.
The following settings are culminative:
.TS
box;
lbz r l l .
Value Description
1 Verify height.
2 Verify pointers from children to parent.
3 Verify element counts.
4 Verify element order. (expensive)
* 5 Verify unused memory is poisoned. (expensive)
.TE
.Sy \& * No Requires debug build .
.
.It Sy zfs_free_leak_on_eio Ns = Ns Sy 0 Ns | Ns 1 Pq int
If destroy encounters an
.Sy EIO
while reading metadata (e.g. indirect blocks),
space referenced by the missing metadata can not be freed.
Normally this causes the background destroy to become "stalled",
as it is unable to make forward progress.
While in this stalled state, all remaining space to free
from the error-encountering filesystem is "temporarily leaked".
Set this flag to cause it to ignore the
.Sy EIO ,
permanently leak the space from indirect blocks that can not be read,
and continue to free everything else that it can.
.Pp
The default "stalling" behavior is useful if the storage partially
fails (i.e. some but not all I/O operations fail), and then later recovers.
In this case, we will be able to continue pool operations while it is
partially failed, and when it recovers, we can continue to free the
space, with no leaks.
Note, however, that this case is actually fairly rare.
.Pp
Typically pools either
.Bl -enum -compact -offset 4n -width "1."
.It
fail completely (but perhaps temporarily,
e.g. due to a top-level vdev going offline), or
.It
have localized, permanent errors (e.g. disk returns the wrong data
due to bit flip or firmware bug).
.El
In the former case, this setting does not matter because the
pool will be suspended and the sync thread will not be able to make
forward progress regardless.
In the latter, because the error is permanent, the best we can do
is leak the minimum amount of space,
which is what setting this flag will do.
It is therefore reasonable for this flag to normally be set,
but we chose the more conservative approach of not setting it,
so that there is no possibility of
leaking space in the "partial temporary" failure case.
.
.It Sy zfs_free_min_time_ms Ns = Ns Sy 1000 Ns ms Po 1s Pc Pq uint
During a
.Nm zfs Cm destroy
operation using the
.Sy async_destroy
feature,
a minimum of this much time will be spent working on freeing blocks per TXG.
.
.It Sy zfs_obsolete_min_time_ms Ns = Ns Sy 500 Ns ms Pq uint
Similar to
.Sy zfs_free_min_time_ms ,
but for cleanup of old indirection records for removed vdevs.
.
.It Sy zfs_immediate_write_sz Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq s64
Largest data block to write to the ZIL.
Larger blocks will be treated as if the dataset being written to had the
.Sy logbias Ns = Ns Sy throughput
property set.
.
.It Sy zfs_initialize_value Ns = Ns Sy 16045690984833335022 Po 0xDEADBEEFDEADBEEE Pc Pq u64
Pattern written to vdev free space by
.Xr zpool-initialize 8 .
.
.It Sy zfs_initialize_chunk_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
Size of writes used by
.Xr zpool-initialize 8 .
This option is used by the test suite.
.
.It Sy zfs_livelist_max_entries Ns = Ns Sy 500000 Po 5*10^5 Pc Pq u64
The threshold size (in block pointers) at which we create a new sub-livelist.
Larger sublists are more costly from a memory perspective but the fewer
sublists there are, the lower the cost of insertion.
.
.It Sy zfs_livelist_min_percent_shared Ns = Ns Sy 75 Ns % Pq int
If the amount of shared space between a snapshot and its clone drops below
this threshold, the clone turns off the livelist and reverts to the old
deletion method.
This is in place because livelists no long give us a benefit
once a clone has been overwritten enough.
.
.It Sy zfs_livelist_condense_new_alloc Ns = Ns Sy 0 Pq int
Incremented each time an extra ALLOC blkptr is added to a livelist entry while
it is being condensed.
This option is used by the test suite to track race conditions.
.
.It Sy zfs_livelist_condense_sync_cancel Ns = Ns Sy 0 Pq int
Incremented each time livelist condensing is canceled while in
.Fn spa_livelist_condense_sync .
This option is used by the test suite to track race conditions.
.
.It Sy zfs_livelist_condense_sync_pause Ns = Ns Sy 0 Ns | Ns 1 Pq int
When set, the livelist condense process pauses indefinitely before
executing the synctask \(em
.Fn spa_livelist_condense_sync .
This option is used by the test suite to trigger race conditions.
.
.It Sy zfs_livelist_condense_zthr_cancel Ns = Ns Sy 0 Pq int
Incremented each time livelist condensing is canceled while in
.Fn spa_livelist_condense_cb .
This option is used by the test suite to track race conditions.
.
.It Sy zfs_livelist_condense_zthr_pause Ns = Ns Sy 0 Ns | Ns 1 Pq int
When set, the livelist condense process pauses indefinitely before
executing the open context condensing work in
.Fn spa_livelist_condense_cb .
This option is used by the test suite to trigger race conditions.
.
.It Sy zfs_lua_max_instrlimit Ns = Ns Sy 100000000 Po 10^8 Pc Pq u64
The maximum execution time limit that can be set for a ZFS channel program,
specified as a number of Lua instructions.
.
.It Sy zfs_lua_max_memlimit Ns = Ns Sy 104857600 Po 100 MiB Pc Pq u64
The maximum memory limit that can be set for a ZFS channel program, specified
in bytes.
.
.It Sy zfs_max_dataset_nesting Ns = Ns Sy 50 Pq int
The maximum depth of nested datasets.
This value can be tuned temporarily to
fix existing datasets that exceed the predefined limit.
.
.It Sy zfs_max_log_walking Ns = Ns Sy 5 Pq u64
The number of past TXGs that the flushing algorithm of the log spacemap
feature uses to estimate incoming log blocks.
.
.It Sy zfs_max_logsm_summary_length Ns = Ns Sy 10 Pq u64
Maximum number of rows allowed in the summary of the spacemap log.
.
.It Sy zfs_max_recordsize Ns = Ns Sy 16777216 Po 16 MiB Pc Pq uint
We currently support block sizes from
.Em 512 Po 512 B Pc No to Em 16777216 Po 16 MiB Pc .
The benefits of larger blocks, and thus larger I/O,
need to be weighed against the cost of COWing a giant block to modify one byte.
Additionally, very large blocks can have an impact on I/O latency,
and also potentially on the memory allocator.
Therefore, we formerly forbade creating blocks larger than 1M.
Larger blocks could be created by changing it,
and pools with larger blocks can always be imported and used,
regardless of this setting.
.
.It Sy zfs_allow_redacted_dataset_mount Ns = Ns Sy 0 Ns | Ns 1 Pq int
Allow datasets received with redacted send/receive to be mounted.
Normally disabled because these datasets may be missing key data.
.
.It Sy zfs_min_metaslabs_to_flush Ns = Ns Sy 1 Pq u64
Minimum number of metaslabs to flush per dirty TXG.
.
.It Sy zfs_metaslab_fragmentation_threshold Ns = Ns Sy 70 Ns % Pq uint
Allow metaslabs to keep their active state as long as their fragmentation
percentage is no more than this value.
An active metaslab that exceeds this threshold
will no longer keep its active status allowing better metaslabs to be selected.
.
.It Sy zfs_mg_fragmentation_threshold Ns = Ns Sy 95 Ns % Pq uint
Metaslab groups are considered eligible for allocations if their
fragmentation metric (measured as a percentage) is less than or equal to
this value.
If a metaslab group exceeds this threshold then it will be
skipped unless all metaslab groups within the metaslab class have also
crossed this threshold.
.
.It Sy zfs_mg_noalloc_threshold Ns = Ns Sy 0 Ns % Pq uint
Defines a threshold at which metaslab groups should be eligible for allocations.
The value is expressed as a percentage of free space
beyond which a metaslab group is always eligible for allocations.
If a metaslab group's free space is less than or equal to the
threshold, the allocator will avoid allocating to that group
unless all groups in the pool have reached the threshold.
Once all groups have reached the threshold, all groups are allowed to accept
allocations.
The default value of
.Sy 0
disables the feature and causes all metaslab groups to be eligible for
allocations.
.Pp
This parameter allows one to deal with pools having heavily imbalanced
vdevs such as would be the case when a new vdev has been added.
Setting the threshold to a non-zero percentage will stop allocations
from being made to vdevs that aren't filled to the specified percentage
and allow lesser filled vdevs to acquire more allocations than they
otherwise would under the old
.Sy zfs_mg_alloc_failures
facility.
.
.It Sy zfs_ddt_data_is_special Ns = Ns Sy 1 Ns | Ns 0 Pq int
If enabled, ZFS will place DDT data into the special allocation class.
.
.It Sy zfs_user_indirect_is_special Ns = Ns Sy 1 Ns | Ns 0 Pq int
If enabled, ZFS will place user data indirect blocks
into the special allocation class.
.
.It Sy zfs_multihost_history Ns = Ns Sy 0 Pq uint
Historical statistics for this many latest multihost updates will be available
in
.Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /multihost .
.
.It Sy zfs_multihost_interval Ns = Ns Sy 1000 Ns ms Po 1 s Pc Pq u64
Used to control the frequency of multihost writes which are performed when the
.Sy multihost
pool property is on.
This is one of the factors used to determine the
length of the activity check during import.
.Pp
The multihost write period is
.Sy zfs_multihost_interval No / Sy leaf-vdevs .
On average a multihost write will be issued for each leaf vdev
every
.Sy zfs_multihost_interval
milliseconds.
In practice, the observed period can vary with the I/O load
and this observed value is the delay which is stored in the uberblock.
.
.It Sy zfs_multihost_import_intervals Ns = Ns Sy 20 Pq uint
Used to control the duration of the activity test on import.
Smaller values of
.Sy zfs_multihost_import_intervals
will reduce the import time but increase
the risk of failing to detect an active pool.
The total activity check time is never allowed to drop below one second.
.Pp
On import the activity check waits a minimum amount of time determined by
.Sy zfs_multihost_interval No \(mu Sy zfs_multihost_import_intervals ,
or the same product computed on the host which last had the pool imported,
whichever is greater.
The activity check time may be further extended if the value of MMP
delay found in the best uberblock indicates actual multihost updates happened
at longer intervals than
.Sy zfs_multihost_interval .
A minimum of
.Em 100 ms
is enforced.
.Pp
.Sy 0 No is equivalent to Sy 1 .
.
.It Sy zfs_multihost_fail_intervals Ns = Ns Sy 10 Pq uint
Controls the behavior of the pool when multihost write failures or delays are
detected.
.Pp
When
.Sy 0 ,
multihost write failures or delays are ignored.
The failures will still be reported to the ZED which depending on
its configuration may take action such as suspending the pool or offlining a
device.
.Pp
Otherwise, the pool will be suspended if
.Sy zfs_multihost_fail_intervals No \(mu Sy zfs_multihost_interval
milliseconds pass without a successful MMP write.
This guarantees the activity test will see MMP writes if the pool is imported.
.Sy 1 No is equivalent to Sy 2 ;
this is necessary to prevent the pool from being suspended
due to normal, small I/O latency variations.
.
.It Sy zfs_no_scrub_io Ns = Ns Sy 0 Ns | Ns 1 Pq int
Set to disable scrub I/O.
This results in scrubs not actually scrubbing data and
simply doing a metadata crawl of the pool instead.
.
.It Sy zfs_no_scrub_prefetch Ns = Ns Sy 0 Ns | Ns 1 Pq int
Set to disable block prefetching for scrubs.
.
.It Sy zfs_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable cache flush operations on disks when writing.
Setting this will cause pool corruption on power loss
if a volatile out-of-order write cache is enabled.
.
.It Sy zfs_nopwrite_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Allow no-operation writes.
The occurrence of nopwrites will further depend on other pool properties
.Pq i.a. the checksumming and compression algorithms .
.
.It Sy zfs_dmu_offset_next_sync Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable forcing TXG sync to find holes.
When enabled forces ZFS to sync data when
.Sy SEEK_HOLE No or Sy SEEK_DATA
flags are used allowing holes in a file to be accurately reported.
When disabled holes will not be reported in recently dirtied files.
.
.It Sy zfs_pd_bytes_max Ns = Ns Sy 52428800 Ns B Po 50 MiB Pc Pq int
The number of bytes which should be prefetched during a pool traversal, like
.Nm zfs Cm send
or other data crawling operations.
.
.It Sy zfs_traverse_indirect_prefetch_limit Ns = Ns Sy 32 Pq uint
The number of blocks pointed by indirect (non-L0) block which should be
prefetched during a pool traversal, like
.Nm zfs Cm send
or other data crawling operations.
.
.It Sy zfs_per_txg_dirty_frees_percent Ns = Ns Sy 30 Ns % Pq u64
Control percentage of dirtied indirect blocks from frees allowed into one TXG.
After this threshold is crossed, additional frees will wait until the next TXG.
.Sy 0 No disables this throttle .
.
.It Sy zfs_prefetch_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable predictive prefetch.
Note that it leaves "prescient" prefetch
.Pq for, e.g., Nm zfs Cm send
intact.
Unlike predictive prefetch, prescient prefetch never issues I/O
that ends up not being needed, so it can't hurt performance.
.
.It Sy zfs_qat_checksum_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable QAT hardware acceleration for SHA256 checksums.
May be unset after the ZFS modules have been loaded to initialize the QAT
hardware as long as support is compiled in and the QAT driver is present.
.
.It Sy zfs_qat_compress_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable QAT hardware acceleration for gzip compression.
May be unset after the ZFS modules have been loaded to initialize the QAT
hardware as long as support is compiled in and the QAT driver is present.
.
.It Sy zfs_qat_encrypt_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable QAT hardware acceleration for AES-GCM encryption.
May be unset after the ZFS modules have been loaded to initialize the QAT
hardware as long as support is compiled in and the QAT driver is present.
.
.It Sy zfs_vnops_read_chunk_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
Bytes to read per chunk.
.
.It Sy zfs_read_history Ns = Ns Sy 0 Pq uint
Historical statistics for this many latest reads will be available in
.Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /reads .
.
.It Sy zfs_read_history_hits Ns = Ns Sy 0 Ns | Ns 1 Pq int
Include cache hits in read history
.
.It Sy zfs_rebuild_max_segment Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
Maximum read segment size to issue when sequentially resilvering a
top-level vdev.
.
.It Sy zfs_rebuild_scrub_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Automatically start a pool scrub when the last active sequential resilver
completes in order to verify the checksums of all blocks which have been
resilvered.
This is enabled by default and strongly recommended.
.
.It Sy zfs_rebuild_vdev_limit Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq u64
Maximum amount of I/O that can be concurrently issued for a sequential
resilver per leaf device, given in bytes.
.
.It Sy zfs_reconstruct_indirect_combinations_max Ns = Ns Sy 4096 Pq int
If an indirect split block contains more than this many possible unique
combinations when being reconstructed, consider it too computationally
expensive to check them all.
Instead, try at most this many randomly selected
combinations each time the block is accessed.
This allows all segment copies to participate fairly
in the reconstruction when all combinations
cannot be checked and prevents repeated use of one bad copy.
.
.It Sy zfs_recover Ns = Ns Sy 0 Ns | Ns 1 Pq int
Set to attempt to recover from fatal errors.
This should only be used as a last resort,
as it typically results in leaked space, or worse.
.
.It Sy zfs_removal_ignore_errors Ns = Ns Sy 0 Ns | Ns 1 Pq int
Ignore hard I/O errors during device removal.
When set, if a device encounters a hard I/O error during the removal process
the removal will not be cancelled.
This can result in a normally recoverable block becoming permanently damaged
and is hence not recommended.
This should only be used as a last resort when the
pool cannot be returned to a healthy state prior to removing the device.
.
.It Sy zfs_removal_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq uint
This is used by the test suite so that it can ensure that certain actions
happen while in the middle of a removal.
.
.It Sy zfs_remove_max_segment Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint
The largest contiguous segment that we will attempt to allocate when removing
a device.
If there is a performance problem with attempting to allocate large blocks,
consider decreasing this.
The default value is also the maximum.
.
.It Sy zfs_resilver_disable_defer Ns = Ns Sy 0 Ns | Ns 1 Pq int
Ignore the
.Sy resilver_defer
feature, causing an operation that would start a resilver to
immediately restart the one in progress.
.
.It Sy zfs_resilver_min_time_ms Ns = Ns Sy 3000 Ns ms Po 3 s Pc Pq uint
Resilvers are processed by the sync thread.
While resilvering, it will spend at least this much time
working on a resilver between TXG flushes.
.
.It Sy zfs_scan_ignore_errors Ns = Ns Sy 0 Ns | Ns 1 Pq int
If set, remove the DTL (dirty time list) upon completion of a pool scan (scrub),
even if there were unrepairable errors.
Intended to be used during pool repair or recovery to
stop resilvering when the pool is next imported.
.
+.It Sy zfs_scrub_after_expand Ns = Ns Sy 1 Ns | Ns 0 Pq int
+Automatically start a pool scrub after a RAIDZ expansion completes
+in order to verify the checksums of all blocks which have been
+copied during the expansion.
+This is enabled by default and strongly recommended.
+.
.It Sy zfs_scrub_min_time_ms Ns = Ns Sy 1000 Ns ms Po 1 s Pc Pq uint
Scrubs are processed by the sync thread.
While scrubbing, it will spend at least this much time
working on a scrub between TXG flushes.
.
.It Sy zfs_scrub_error_blocks_per_txg Ns = Ns Sy 4096 Pq uint
Error blocks to be scrubbed in one txg.
.
.It Sy zfs_scan_checkpoint_intval Ns = Ns Sy 7200 Ns s Po 2 hour Pc Pq uint
To preserve progress across reboots, the sequential scan algorithm periodically
needs to stop metadata scanning and issue all the verification I/O to disk.
The frequency of this flushing is determined by this tunable.
.
.It Sy zfs_scan_fill_weight Ns = Ns Sy 3 Pq uint
This tunable affects how scrub and resilver I/O segments are ordered.
A higher number indicates that we care more about how filled in a segment is,
while a lower number indicates we care more about the size of the extent without
considering the gaps within a segment.
This value is only tunable upon module insertion.
Changing the value afterwards will have no effect on scrub or resilver
performance.
.
.It Sy zfs_scan_issue_strategy Ns = Ns Sy 0 Pq uint
Determines the order that data will be verified while scrubbing or resilvering:
.Bl -tag -compact -offset 4n -width "a"
.It Sy 1
Data will be verified as sequentially as possible, given the
amount of memory reserved for scrubbing
.Pq see Sy zfs_scan_mem_lim_fact .
This may improve scrub performance if the pool's data is very fragmented.
.It Sy 2
The largest mostly-contiguous chunk of found data will be verified first.
By deferring scrubbing of small segments, we may later find adjacent data
to coalesce and increase the segment size.
.It Sy 0
.No Use strategy Sy 1 No during normal verification
.No and strategy Sy 2 No while taking a checkpoint .
.El
.
.It Sy zfs_scan_legacy Ns = Ns Sy 0 Ns | Ns 1 Pq int
If unset, indicates that scrubs and resilvers will gather metadata in
memory before issuing sequential I/O.
Otherwise indicates that the legacy algorithm will be used,
where I/O is initiated as soon as it is discovered.
Unsetting will not affect scrubs or resilvers that are already in progress.
.
.It Sy zfs_scan_max_ext_gap Ns = Ns Sy 2097152 Ns B Po 2 MiB Pc Pq int
Sets the largest gap in bytes between scrub/resilver I/O operations
that will still be considered sequential for sorting purposes.
Changing this value will not
affect scrubs or resilvers that are already in progress.
.
.It Sy zfs_scan_mem_lim_fact Ns = Ns Sy 20 Ns ^-1 Pq uint
Maximum fraction of RAM used for I/O sorting by sequential scan algorithm.
This tunable determines the hard limit for I/O sorting memory usage.
When the hard limit is reached we stop scanning metadata and start issuing
data verification I/O.
This is done until we get below the soft limit.
.
.It Sy zfs_scan_mem_lim_soft_fact Ns = Ns Sy 20 Ns ^-1 Pq uint
The fraction of the hard limit used to determined the soft limit for I/O sorting
by the sequential scan algorithm.
When we cross this limit from below no action is taken.
When we cross this limit from above it is because we are issuing verification
I/O.
In this case (unless the metadata scan is done) we stop issuing verification I/O
and start scanning metadata again until we get to the hard limit.
.
.It Sy zfs_scan_report_txgs Ns = Ns Sy 0 Ns | Ns 1 Pq uint
When reporting resilver throughput and estimated completion time use the
performance observed over roughly the last
.Sy zfs_scan_report_txgs
TXGs.
When set to zero performance is calculated over the time between checkpoints.
.
.It Sy zfs_scan_strict_mem_lim Ns = Ns Sy 0 Ns | Ns 1 Pq int
Enforce tight memory limits on pool scans when a sequential scan is in progress.
When disabled, the memory limit may be exceeded by fast disks.
.
.It Sy zfs_scan_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq int
Freezes a scrub/resilver in progress without actually pausing it.
Intended for testing/debugging.
.
.It Sy zfs_scan_vdev_limit Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq int
Maximum amount of data that can be concurrently issued at once for scrubs and
resilvers per leaf device, given in bytes.
.
.It Sy zfs_send_corrupt_data Ns = Ns Sy 0 Ns | Ns 1 Pq int
Allow sending of corrupt data (ignore read/checksum errors when sending).
.
.It Sy zfs_send_unmodified_spill_blocks Ns = Ns Sy 1 Ns | Ns 0 Pq int
Include unmodified spill blocks in the send stream.
Under certain circumstances, previous versions of ZFS could incorrectly
remove the spill block from an existing object.
Including unmodified copies of the spill blocks creates a backwards-compatible
stream which will recreate a spill block if it was incorrectly removed.
.
.It Sy zfs_send_no_prefetch_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint
The fill fraction of the
.Nm zfs Cm send
internal queues.
The fill fraction controls the timing with which internal threads are woken up.
.
.It Sy zfs_send_no_prefetch_queue_length Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint
The maximum number of bytes allowed in
.Nm zfs Cm send Ns 's
internal queues.
.
.It Sy zfs_send_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint
The fill fraction of the
.Nm zfs Cm send
prefetch queue.
The fill fraction controls the timing with which internal threads are woken up.
.
.It Sy zfs_send_queue_length Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint
The maximum number of bytes allowed that will be prefetched by
.Nm zfs Cm send .
This value must be at least twice the maximum block size in use.
.
.It Sy zfs_recv_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint
The fill fraction of the
.Nm zfs Cm receive
queue.
The fill fraction controls the timing with which internal threads are woken up.
.
.It Sy zfs_recv_queue_length Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint
The maximum number of bytes allowed in the
.Nm zfs Cm receive
queue.
This value must be at least twice the maximum block size in use.
.
.It Sy zfs_recv_write_batch_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint
The maximum amount of data, in bytes, that
.Nm zfs Cm receive
will write in one DMU transaction.
This is the uncompressed size, even when receiving a compressed send stream.
This setting will not reduce the write size below a single block.
Capped at a maximum of
.Sy 32 MiB .
.
.It Sy zfs_recv_best_effort_corrective Ns = Ns Sy 0 Pq int
When this variable is set to non-zero a corrective receive:
.Bl -enum -compact -offset 4n -width "1."
.It
Does not enforce the restriction of source & destination snapshot GUIDs
matching.
.It
If there is an error during healing, the healing receive is not
terminated instead it moves on to the next record.
.El
.
.It Sy zfs_override_estimate_recordsize Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Setting this variable overrides the default logic for estimating block
sizes when doing a
.Nm zfs Cm send .
The default heuristic is that the average block size
will be the current recordsize.
Override this value if most data in your dataset is not of that size
and you require accurate zfs send size estimates.
.
.It Sy zfs_sync_pass_deferred_free Ns = Ns Sy 2 Pq uint
Flushing of data to disk is done in passes.
Defer frees starting in this pass.
.
.It Sy zfs_spa_discard_memory_limit Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq int
Maximum memory used for prefetching a checkpoint's space map on each
vdev while discarding the checkpoint.
.
.It Sy zfs_special_class_metadata_reserve_pct Ns = Ns Sy 25 Ns % Pq uint
Only allow small data blocks to be allocated on the special and dedup vdev
types when the available free space percentage on these vdevs exceeds this
value.
This ensures reserved space is available for pool metadata as the
special vdevs approach capacity.
.
.It Sy zfs_sync_pass_dont_compress Ns = Ns Sy 8 Pq uint
Starting in this sync pass, disable compression (including of metadata).
With the default setting, in practice, we don't have this many sync passes,
so this has no effect.
.Pp
The original intent was that disabling compression would help the sync passes
to converge.
However, in practice, disabling compression increases
the average number of sync passes; because when we turn compression off,
many blocks' size will change, and thus we have to re-allocate
(not overwrite) them.
It also increases the number of
.Em 128 KiB
allocations (e.g. for indirect blocks and spacemaps)
because these will not be compressed.
The
.Em 128 KiB
allocations are especially detrimental to performance
on highly fragmented systems, which may have very few free segments of this
size,
and may need to load new metaslabs to satisfy these allocations.
.
.It Sy zfs_sync_pass_rewrite Ns = Ns Sy 2 Pq uint
Rewrite new block pointers starting in this pass.
.
.It Sy zfs_trim_extent_bytes_max Ns = Ns Sy 134217728 Ns B Po 128 MiB Pc Pq uint
Maximum size of TRIM command.
Larger ranges will be split into chunks no larger than this value before
issuing.
.
.It Sy zfs_trim_extent_bytes_min Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint
Minimum size of TRIM commands.
TRIM ranges smaller than this will be skipped,
unless they're part of a larger range which was chunked.
This is done because it's common for these small TRIMs
to negatively impact overall performance.
.
.It Sy zfs_trim_metaslab_skip Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Skip uninitialized metaslabs during the TRIM process.
This option is useful for pools constructed from large thinly-provisioned
devices
where TRIM operations are slow.
As a pool ages, an increasing fraction of the pool's metaslabs
will be initialized, progressively degrading the usefulness of this option.
This setting is stored when starting a manual TRIM and will
persist for the duration of the requested TRIM.
.
.It Sy zfs_trim_queue_limit Ns = Ns Sy 10 Pq uint
Maximum number of queued TRIMs outstanding per leaf vdev.
The number of concurrent TRIM commands issued to the device is controlled by
.Sy zfs_vdev_trim_min_active No and Sy zfs_vdev_trim_max_active .
.
.It Sy zfs_trim_txg_batch Ns = Ns Sy 32 Pq uint
The number of transaction groups' worth of frees which should be aggregated
before TRIM operations are issued to the device.
This setting represents a trade-off between issuing larger,
more efficient TRIM operations and the delay
before the recently trimmed space is available for use by the device.
.Pp
Increasing this value will allow frees to be aggregated for a longer time.
This will result is larger TRIM operations and potentially increased memory
usage.
Decreasing this value will have the opposite effect.
The default of
.Sy 32
was determined to be a reasonable compromise.
.
.It Sy zfs_txg_history Ns = Ns Sy 0 Pq uint
Historical statistics for this many latest TXGs will be available in
.Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /TXGs .
.
.It Sy zfs_txg_timeout Ns = Ns Sy 5 Ns s Pq uint
Flush dirty data to disk at least every this many seconds (maximum TXG
duration).
.
.It Sy zfs_vdev_aggregation_limit Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint
Max vdev I/O aggregation size.
.
.It Sy zfs_vdev_aggregation_limit_non_rotating Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint
Max vdev I/O aggregation size for non-rotating media.
.
.It Sy zfs_vdev_mirror_rotating_inc Ns = Ns Sy 0 Pq int
A number by which the balancing algorithm increments the load calculation for
the purpose of selecting the least busy mirror member when an I/O operation
immediately follows its predecessor on rotational vdevs
for the purpose of making decisions based on load.
.
.It Sy zfs_vdev_mirror_rotating_seek_inc Ns = Ns Sy 5 Pq int
A number by which the balancing algorithm increments the load calculation for
the purpose of selecting the least busy mirror member when an I/O operation
lacks locality as defined by
.Sy zfs_vdev_mirror_rotating_seek_offset .
Operations within this that are not immediately following the previous operation
are incremented by half.
.
.It Sy zfs_vdev_mirror_rotating_seek_offset Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq int
The maximum distance for the last queued I/O operation in which
the balancing algorithm considers an operation to have locality.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_mirror_non_rotating_inc Ns = Ns Sy 0 Pq int
A number by which the balancing algorithm increments the load calculation for
the purpose of selecting the least busy mirror member on non-rotational vdevs
when I/O operations do not immediately follow one another.
.
.It Sy zfs_vdev_mirror_non_rotating_seek_inc Ns = Ns Sy 1 Pq int
A number by which the balancing algorithm increments the load calculation for
the purpose of selecting the least busy mirror member when an I/O operation
lacks
locality as defined by the
.Sy zfs_vdev_mirror_rotating_seek_offset .
Operations within this that are not immediately following the previous operation
are incremented by half.
.
.It Sy zfs_vdev_read_gap_limit Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint
Aggregate read I/O operations if the on-disk gap between them is within this
threshold.
.
.It Sy zfs_vdev_write_gap_limit Ns = Ns Sy 4096 Ns B Po 4 KiB Pc Pq uint
Aggregate write I/O operations if the on-disk gap between them is within this
threshold.
.
.It Sy zfs_vdev_raidz_impl Ns = Ns Sy fastest Pq string
Select the raidz parity implementation to use.
.Pp
Variants that don't depend on CPU-specific features
may be selected on module load, as they are supported on all systems.
The remaining options may only be set after the module is loaded,
as they are available only if the implementations are compiled in
and supported on the running system.
.Pp
Once the module is loaded,
.Pa /sys/module/zfs/parameters/zfs_vdev_raidz_impl
will show the available options,
with the currently selected one enclosed in square brackets.
.Pp
.TS
lb l l .
fastest selected by built-in benchmark
original original implementation
scalar scalar implementation
sse2 SSE2 instruction set 64-bit x86
ssse3 SSSE3 instruction set 64-bit x86
avx2 AVX2 instruction set 64-bit x86
avx512f AVX512F instruction set 64-bit x86
avx512bw AVX512F & AVX512BW instruction sets 64-bit x86
aarch64_neon NEON Aarch64/64-bit ARMv8
aarch64_neonx2 NEON with more unrolling Aarch64/64-bit ARMv8
powerpc_altivec Altivec PowerPC
.TE
.
.It Sy zfs_vdev_scheduler Pq charp
.Sy DEPRECATED .
Prints warning to kernel log for compatibility.
.
.It Sy zfs_zevent_len_max Ns = Ns Sy 512 Pq uint
Max event queue length.
Events in the queue can be viewed with
.Xr zpool-events 8 .
.
.It Sy zfs_zevent_retain_max Ns = Ns Sy 2000 Pq int
Maximum recent zevent records to retain for duplicate checking.
Setting this to
.Sy 0
disables duplicate detection.
.
.It Sy zfs_zevent_retain_expire_secs Ns = Ns Sy 900 Ns s Po 15 min Pc Pq int
Lifespan for a recent ereport that was retained for duplicate checking.
.
.It Sy zfs_zil_clean_taskq_maxalloc Ns = Ns Sy 1048576 Pq int
The maximum number of taskq entries that are allowed to be cached.
When this limit is exceeded transaction records (itxs)
will be cleaned synchronously.
.
.It Sy zfs_zil_clean_taskq_minalloc Ns = Ns Sy 1024 Pq int
The number of taskq entries that are pre-populated when the taskq is first
created and are immediately available for use.
.
.It Sy zfs_zil_clean_taskq_nthr_pct Ns = Ns Sy 100 Ns % Pq int
This controls the number of threads used by
.Sy dp_zil_clean_taskq .
The default value of
.Sy 100%
will create a maximum of one thread per cpu.
.
.It Sy zil_maxblocksize Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint
This sets the maximum block size used by the ZIL.
On very fragmented pools, lowering this
.Pq typically to Sy 36 KiB
can improve performance.
.
.It Sy zil_maxcopied Ns = Ns Sy 7680 Ns B Po 7.5 KiB Pc Pq uint
This sets the maximum number of write bytes logged via WR_COPIED.
It tunes a tradeoff between additional memory copy and possibly worse log
space efficiency vs additional range lock/unlock.
.
.It Sy zil_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable the cache flush commands that are normally sent to disk by
the ZIL after an LWB write has completed.
Setting this will cause ZIL corruption on power loss
if a volatile out-of-order write cache is enabled.
.
.It Sy zil_replay_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable intent logging replay.
Can be disabled for recovery from corrupted ZIL.
.
.It Sy zil_slog_bulk Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq u64
Limit SLOG write size per commit executed with synchronous priority.
Any writes above that will be executed with lower (asynchronous) priority
to limit potential SLOG device abuse by single active ZIL writer.
.
.It Sy zfs_zil_saxattr Ns = Ns Sy 1 Ns | Ns 0 Pq int
Setting this tunable to zero disables ZIL logging of new
.Sy xattr Ns = Ns Sy sa
records if the
.Sy org.openzfs:zilsaxattr
feature is enabled on the pool.
This would only be necessary to work around bugs in the ZIL logging or replay
code for this record type.
The tunable has no effect if the feature is disabled.
.
.It Sy zfs_embedded_slog_min_ms Ns = Ns Sy 64 Pq uint
Usually, one metaslab from each normal-class vdev is dedicated for use by
the ZIL to log synchronous writes.
However, if there are fewer than
.Sy zfs_embedded_slog_min_ms
metaslabs in the vdev, this functionality is disabled.
This ensures that we don't set aside an unreasonable amount of space for the
ZIL.
.
.It Sy zstd_earlyabort_pass Ns = Ns Sy 1 Pq uint
Whether heuristic for detection of incompressible data with zstd levels >= 3
using LZ4 and zstd-1 passes is enabled.
.
.It Sy zstd_abort_size Ns = Ns Sy 131072 Pq uint
Minimal uncompressed size (inclusive) of a record before the early abort
heuristic will be attempted.
.
.It Sy zio_deadman_log_all Ns = Ns Sy 0 Ns | Ns 1 Pq int
If non-zero, the zio deadman will produce debugging messages
.Pq see Sy zfs_dbgmsg_enable
for all zios, rather than only for leaf zios possessing a vdev.
This is meant to be used by developers to gain
diagnostic information for hang conditions which don't involve a mutex
or other locking primitive: typically conditions in which a thread in
the zio pipeline is looping indefinitely.
.
.It Sy zio_slow_io_ms Ns = Ns Sy 30000 Ns ms Po 30 s Pc Pq int
When an I/O operation takes more than this much time to complete,
it's marked as slow.
Each slow operation causes a delay zevent.
Slow I/O counters can be seen with
.Nm zpool Cm status Fl s .
.
.It Sy zio_dva_throttle_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Throttle block allocations in the I/O pipeline.
This allows for dynamic allocation distribution when devices are imbalanced.
When enabled, the maximum number of pending allocations per top-level vdev
is limited by
.Sy zfs_vdev_queue_depth_pct .
.
.It Sy zfs_xattr_compat Ns = Ns 0 Ns | Ns 1 Pq int
Control the naming scheme used when setting new xattrs in the user namespace.
If
.Sy 0
.Pq the default on Linux ,
user namespace xattr names are prefixed with the namespace, to be backwards
compatible with previous versions of ZFS on Linux.
If
.Sy 1
.Pq the default on Fx ,
user namespace xattr names are not prefixed, to be backwards compatible with
previous versions of ZFS on illumos and
.Fx .
.Pp
Either naming scheme can be read on this and future versions of ZFS, regardless
of this tunable, but legacy ZFS on illumos or
.Fx
are unable to read user namespace xattrs written in the Linux format, and
legacy versions of ZFS on Linux are unable to read user namespace xattrs written
in the legacy ZFS format.
.Pp
An existing xattr with the alternate naming scheme is removed when overwriting
the xattr so as to not accumulate duplicates.
.
.It Sy zio_requeue_io_start_cut_in_line Ns = Ns Sy 0 Ns | Ns 1 Pq int
Prioritize requeued I/O.
.
.It Sy zio_taskq_batch_pct Ns = Ns Sy 80 Ns % Pq uint
Percentage of online CPUs which will run a worker thread for I/O.
These workers are responsible for I/O work such as compression and
checksum calculations.
Fractional number of CPUs will be rounded down.
.Pp
The default value of
.Sy 80%
was chosen to avoid using all CPUs which can result in
latency issues and inconsistent application performance,
especially when slower compression and/or checksumming is enabled.
.
.It Sy zio_taskq_batch_tpq Ns = Ns Sy 0 Pq uint
Number of worker threads per taskq.
Lower values improve I/O ordering and CPU utilization,
while higher reduces lock contention.
.Pp
If
.Sy 0 ,
generate a system-dependent value close to 6 threads per taskq.
.
.It Sy zio_taskq_wr_iss_ncpus Ns = Ns Sy 0 Pq uint
Determines the number of CPUs to run write issue taskqs.
.Pp
When 0 (the default), the value to use is computed internally
as the number of actual CPUs in the system divided by the
.Sy spa_num_allocators
value.
.
.It Sy zvol_inhibit_dev Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Do not create zvol device nodes.
This may slightly improve startup time on
systems with a very large number of zvols.
.
.It Sy zvol_major Ns = Ns Sy 230 Pq uint
Major number for zvol block devices.
.
.It Sy zvol_max_discard_blocks Ns = Ns Sy 16384 Pq long
Discard (TRIM) operations done on zvols will be done in batches of this
many blocks, where block size is determined by the
.Sy volblocksize
property of a zvol.
.
.It Sy zvol_prefetch_bytes Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint
When adding a zvol to the system, prefetch this many bytes
from the start and end of the volume.
Prefetching these regions of the volume is desirable,
because they are likely to be accessed immediately by
.Xr blkid 8
or the kernel partitioner.
.
.It Sy zvol_request_sync Ns = Ns Sy 0 Ns | Ns 1 Pq uint
When processing I/O requests for a zvol, submit them synchronously.
This effectively limits the queue depth to
.Em 1
for each I/O submitter.
When unset, requests are handled asynchronously by a thread pool.
The number of requests which can be handled concurrently is controlled by
.Sy zvol_threads .
.Sy zvol_request_sync
is ignored when running on a kernel that supports block multiqueue
.Pq Li blk-mq .
.
.It Sy zvol_threads Ns = Ns Sy 0 Pq uint
The number of system wide threads to use for processing zvol block IOs.
If
.Sy 0
(the default) then internally set
.Sy zvol_threads
to the number of CPUs present or 32 (whichever is greater).
.
.It Sy zvol_blk_mq_threads Ns = Ns Sy 0 Pq uint
The number of threads per zvol to use for queuing IO requests.
This parameter will only appear if your kernel supports
.Li blk-mq
and is only read and assigned to a zvol at zvol load time.
If
.Sy 0
(the default) then internally set
.Sy zvol_blk_mq_threads
to the number of CPUs present.
.
.It Sy zvol_use_blk_mq Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Set to
.Sy 1
to use the
.Li blk-mq
API for zvols.
Set to
.Sy 0
(the default) to use the legacy zvol APIs.
This setting can give better or worse zvol performance depending on
the workload.
This parameter will only appear if your kernel supports
.Li blk-mq
and is only read and assigned to a zvol at zvol load time.
.
.It Sy zvol_blk_mq_blocks_per_thread Ns = Ns Sy 8 Pq uint
If
.Sy zvol_use_blk_mq
is enabled, then process this number of
.Sy volblocksize Ns -sized blocks per zvol thread.
This tunable can be use to favor better performance for zvol reads (lower
values) or writes (higher values).
If set to
.Sy 0 ,
then the zvol layer will process the maximum number of blocks
per thread that it can.
This parameter will only appear if your kernel supports
.Li blk-mq
and is only applied at each zvol's load time.
.
.It Sy zvol_blk_mq_queue_depth Ns = Ns Sy 0 Pq uint
The queue_depth value for the zvol
.Li blk-mq
interface.
This parameter will only appear if your kernel supports
.Li blk-mq
and is only applied at each zvol's load time.
If
.Sy 0
(the default) then use the kernel's default queue depth.
Values are clamped to the kernel's
.Dv BLKDEV_MIN_RQ
and
.Dv BLKDEV_MAX_RQ Ns / Ns Dv BLKDEV_DEFAULT_RQ
limits.
.
.It Sy zvol_volmode Ns = Ns Sy 1 Pq uint
Defines zvol block devices behaviour when
.Sy volmode Ns = Ns Sy default :
.Bl -tag -compact -offset 4n -width "a"
.It Sy 1
.No equivalent to Sy full
.It Sy 2
.No equivalent to Sy dev
.It Sy 3
.No equivalent to Sy none
.El
.
.It Sy zvol_enforce_quotas Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Enable strict ZVOL quota enforcement.
The strict quota enforcement may have a performance impact.
.El
.
.Sh ZFS I/O SCHEDULER
ZFS issues I/O operations to leaf vdevs to satisfy and complete I/O operations.
The scheduler determines when and in what order those operations are issued.
The scheduler divides operations into five I/O classes,
prioritized in the following order: sync read, sync write, async read,
async write, and scrub/resilver.
Each queue defines the minimum and maximum number of concurrent operations
that may be issued to the device.
In addition, the device has an aggregate maximum,
.Sy zfs_vdev_max_active .
Note that the sum of the per-queue minima must not exceed the aggregate maximum.
If the sum of the per-queue maxima exceeds the aggregate maximum,
then the number of active operations may reach
.Sy zfs_vdev_max_active ,
in which case no further operations will be issued,
regardless of whether all per-queue minima have been met.
.Pp
For many physical devices, throughput increases with the number of
concurrent operations, but latency typically suffers.
Furthermore, physical devices typically have a limit
at which more concurrent operations have no
effect on throughput or can actually cause it to decrease.
.Pp
The scheduler selects the next operation to issue by first looking for an
I/O class whose minimum has not been satisfied.
Once all are satisfied and the aggregate maximum has not been hit,
the scheduler looks for classes whose maximum has not been satisfied.
Iteration through the I/O classes is done in the order specified above.
No further operations are issued
if the aggregate maximum number of concurrent operations has been hit,
or if there are no operations queued for an I/O class that has not hit its
maximum.
Every time an I/O operation is queued or an operation completes,
the scheduler looks for new operations to issue.
.Pp
In general, smaller
.Sy max_active Ns s
will lead to lower latency of synchronous operations.
Larger
.Sy max_active Ns s
may lead to higher overall throughput, depending on underlying storage.
.Pp
The ratio of the queues'
.Sy max_active Ns s
determines the balance of performance between reads, writes, and scrubs.
For example, increasing
.Sy zfs_vdev_scrub_max_active
will cause the scrub or resilver to complete more quickly,
but reads and writes to have higher latency and lower throughput.
.Pp
All I/O classes have a fixed maximum number of outstanding operations,
except for the async write class.
Asynchronous writes represent the data that is committed to stable storage
during the syncing stage for transaction groups.
Transaction groups enter the syncing state periodically,
so the number of queued async writes will quickly burst up
and then bleed down to zero.
Rather than servicing them as quickly as possible,
the I/O scheduler changes the maximum number of active async write operations
according to the amount of dirty data in the pool.
Since both throughput and latency typically increase with the number of
concurrent operations issued to physical devices, reducing the
burstiness in the number of simultaneous operations also stabilizes the
response time of operations from other queues, in particular synchronous ones.
In broad strokes, the I/O scheduler will issue more concurrent operations
from the async write queue as there is more dirty data in the pool.
.
.Ss Async Writes
The number of concurrent operations issued for the async write I/O class
follows a piece-wise linear function defined by a few adjustable points:
.Bd -literal
| o---------| <-- \fBzfs_vdev_async_write_max_active\fP
^ | /^ |
| | / | |
active | / | |
I/O | / | |
count | / | |
| / | |
|-------o | | <-- \fBzfs_vdev_async_write_min_active\fP
0|_______^______|_________|
0% | | 100% of \fBzfs_dirty_data_max\fP
| |
| `-- \fBzfs_vdev_async_write_active_max_dirty_percent\fP
`--------- \fBzfs_vdev_async_write_active_min_dirty_percent\fP
.Ed
.Pp
Until the amount of dirty data exceeds a minimum percentage of the dirty
data allowed in the pool, the I/O scheduler will limit the number of
concurrent operations to the minimum.
As that threshold is crossed, the number of concurrent operations issued
increases linearly to the maximum at the specified maximum percentage
of the dirty data allowed in the pool.
.Pp
Ideally, the amount of dirty data on a busy pool will stay in the sloped
part of the function between
.Sy zfs_vdev_async_write_active_min_dirty_percent
and
.Sy zfs_vdev_async_write_active_max_dirty_percent .
If it exceeds the maximum percentage,
this indicates that the rate of incoming data is
greater than the rate that the backend storage can handle.
In this case, we must further throttle incoming writes,
as described in the next section.
.
.Sh ZFS TRANSACTION DELAY
We delay transactions when we've determined that the backend storage
isn't able to accommodate the rate of incoming writes.
.Pp
If there is already a transaction waiting, we delay relative to when
that transaction will finish waiting.
This way the calculated delay time
is independent of the number of threads concurrently executing transactions.
.Pp
If we are the only waiter, wait relative to when the transaction started,
rather than the current time.
This credits the transaction for "time already served",
e.g. reading indirect blocks.
.Pp
The minimum time for a transaction to take is calculated as
.D1 min_time = min( Ns Sy zfs_delay_scale No \(mu Po Sy dirty No \- Sy min Pc / Po Sy max No \- Sy dirty Pc , 100ms)
.Pp
The delay has two degrees of freedom that can be adjusted via tunables.
The percentage of dirty data at which we start to delay is defined by
.Sy zfs_delay_min_dirty_percent .
This should typically be at or above
.Sy zfs_vdev_async_write_active_max_dirty_percent ,
so that we only start to delay after writing at full speed
has failed to keep up with the incoming write rate.
The scale of the curve is defined by
.Sy zfs_delay_scale .
Roughly speaking, this variable determines the amount of delay at the midpoint
of the curve.
.Bd -literal
delay
10ms +-------------------------------------------------------------*+
| *|
9ms + *+
| *|
8ms + *+
| * |
7ms + * +
| * |
6ms + * +
| * |
5ms + * +
| * |
4ms + * +
| * |
3ms + * +
| * |
2ms + (midpoint) * +
| | ** |
1ms + v *** +
| \fBzfs_delay_scale\fP ----------> ******** |
0 +-------------------------------------*********----------------+
0% <- \fBzfs_dirty_data_max\fP -> 100%
.Ed
.Pp
Note, that since the delay is added to the outstanding time remaining on the
most recent transaction it's effectively the inverse of IOPS.
Here, the midpoint of
.Em 500 us
translates to
.Em 2000 IOPS .
The shape of the curve
was chosen such that small changes in the amount of accumulated dirty data
in the first three quarters of the curve yield relatively small differences
in the amount of delay.
.Pp
The effects can be easier to understand when the amount of delay is
represented on a logarithmic scale:
.Bd -literal
delay
100ms +-------------------------------------------------------------++
+ +
| |
+ *+
10ms + *+
+ ** +
| (midpoint) ** |
+ | ** +
1ms + v **** +
+ \fBzfs_delay_scale\fP ----------> ***** +
| **** |
+ **** +
100us + ** +
+ * +
| * |
+ * +
10us + * +
+ +
| |
+ +
+--------------------------------------------------------------+
0% <- \fBzfs_dirty_data_max\fP -> 100%
.Ed
.Pp
Note here that only as the amount of dirty data approaches its limit does
the delay start to increase rapidly.
The goal of a properly tuned system should be to keep the amount of dirty data
out of that range by first ensuring that the appropriate limits are set
for the I/O scheduler to reach optimal throughput on the back-end storage,
and then by changing the value of
.Sy zfs_delay_scale
to increase the steepness of the curve.
diff --git a/sys/contrib/openzfs/man/man7/zpool-features.7 b/sys/contrib/openzfs/man/man7/zpool-features.7
index 3c7b0b345d96..01dec61b91b9 100644
--- a/sys/contrib/openzfs/man/man7/zpool-features.7
+++ b/sys/contrib/openzfs/man/man7/zpool-features.7
@@ -1,966 +1,974 @@
.\"
.\" Copyright (c) 2012, 2018 by Delphix. All rights reserved.
.\" Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
.\" Copyright (c) 2014, Joyent, Inc. All rights reserved.
.\" The contents of this file are subject to the terms of the Common Development
.\" and Distribution License (the "License"). You may not use this file except
.\" in compliance with the License. You can obtain a copy of the license at
.\" usr/src/OPENSOLARIS.LICENSE or https://opensource.org/licenses/CDDL-1.0.
.\"
.\" See the License for the specific language governing permissions and
.\" limitations under the License. When distributing Covered Code, include this
.\" CDDL HEADER in each file and include the License file at
.\" usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this
.\" CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your
.\" own identifying information:
.\" Portions Copyright [yyyy] [name of copyright owner]
.\" Copyright (c) 2019, Klara Inc.
.\" Copyright (c) 2019, Allan Jude
.\" Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
.\"
.Dd June 23, 2022
.Dt ZPOOL-FEATURES 7
.Os
.
.Sh NAME
.Nm zpool-features
.Nd description of ZFS pool features
.
.Sh DESCRIPTION
ZFS pool on-disk format versions are specified via
.Dq features
which replace the old on-disk format numbers
.Pq the last supported on-disk format number is 28 .
To enable a feature on a pool use the
.Nm zpool Cm upgrade ,
or set the
.Sy feature Ns @ Ns Ar feature-name
property to
.Sy enabled .
Please also see the
.Sx Compatibility feature sets
section for information on how sets of features may be enabled together.
.Pp
The pool format does not affect file system version compatibility or the ability
to send file systems between pools.
.Pp
Since most features can be enabled independently of each other, the on-disk
format of the pool is specified by the set of all features marked as
.Sy active
on the pool.
If the pool was created by another software version
this set may include unsupported features.
.
.Ss Identifying features
Every feature has a GUID of the form
.Ar com.example : Ns Ar feature-name .
The reversed DNS name ensures that the feature's GUID is unique across all ZFS
implementations.
When unsupported features are encountered on a pool they will
be identified by their GUIDs.
Refer to the documentation for the ZFS
implementation that created the pool for information about those features.
.Pp
Each supported feature also has a short name.
By convention a feature's short name is the portion of its GUID which follows
the
.Sq \&:
.Po
i.e.
.Ar com.example : Ns Ar feature-name
would have the short name
.Ar feature-name
.Pc ,
however a feature's short name may differ across ZFS implementations if
following the convention would result in name conflicts.
.
.Ss Feature states
Features can be in one of three states:
.Bl -tag -width "disabled"
.It Sy active
This feature's on-disk format changes are in effect on the pool.
Support for this feature is required to import the pool in read-write mode.
If this feature is not read-only compatible,
support is also required to import the pool in read-only mode
.Pq see Sx Read-only compatibility .
.It Sy enabled
An administrator has marked this feature as enabled on the pool, but the
feature's on-disk format changes have not been made yet.
The pool can still be imported by software that does not support this feature,
but changes may be made to the on-disk format at any time
which will move the feature to the
.Sy active
state.
Some features may support returning to the
.Sy enabled
state after becoming
.Sy active .
See feature-specific documentation for details.
.It Sy disabled
This feature's on-disk format changes have not been made and will not be made
unless an administrator moves the feature to the
.Sy enabled
state.
Features cannot be disabled once they have been enabled.
.El
.Pp
The state of supported features is exposed through pool properties of the form
.Sy feature Ns @ Ns Ar short-name .
.
.Ss Read-only compatibility
Some features may make on-disk format changes that do not interfere with other
software's ability to read from the pool.
These features are referred to as
.Dq read-only compatible .
If all unsupported features on a pool are read-only compatible,
the pool can be imported in read-only mode by setting the
.Sy readonly
property during import
.Po see
.Xr zpool-import 8
for details on importing pools
.Pc .
.
.Ss Unsupported features
For each unsupported feature enabled on an imported pool, a pool property
named
.Sy unsupported Ns @ Ns Ar feature-name
will indicate why the import was allowed despite the unsupported feature.
Possible values for this property are:
.Bl -tag -width "readonly"
.It Sy inactive
The feature is in the
.Sy enabled
state and therefore the pool's on-disk
format is still compatible with software that does not support this feature.
.It Sy readonly
The feature is read-only compatible and the pool has been imported in
read-only mode.
.El
.
.Ss Feature dependencies
Some features depend on other features being enabled in order to function.
Enabling a feature will automatically enable any features it depends on.
.
.Ss Compatibility feature sets
It is sometimes necessary for a pool to maintain compatibility with a
specific on-disk format, by enabling and disabling particular features.
The
.Sy compatibility
feature facilitates this by allowing feature sets to be read from text files.
When set to
.Sy off
.Pq the default ,
compatibility feature sets are disabled
.Pq i.e. all features are enabled ;
when set to
.Sy legacy ,
no features are enabled.
When set to a comma-separated list of filenames
.Po
each filename may either be an absolute path, or relative to
.Pa /etc/zfs/compatibility.d
or
.Pa /usr/share/zfs/compatibility.d
.Pc ,
the lists of requested features are read from those files,
separated by whitespace and/or commas.
Only features present in all files are enabled.
.Pp
Simple sanity checks are applied to the files:
they must be between 1 B and 16 KiB in size, and must end with a newline
character.
.Pp
The requested features are applied when a pool is created using
.Nm zpool Cm create Fl o Sy compatibility Ns = Ns Ar …
and controls which features are enabled when using
.Nm zpool Cm upgrade .
.Nm zpool Cm status
will not show a warning about disabled features which are not part
of the requested feature set.
.Pp
The special value
.Sy legacy
prevents any features from being enabled, either via
.Nm zpool Cm upgrade
or
.Nm zpool Cm set Sy feature Ns @ Ns Ar feature-name Ns = Ns Sy enabled .
This setting also prevents pools from being upgraded to newer on-disk versions.
This is a safety measure to prevent new features from being
accidentally enabled, breaking compatibility.
.Pp
By convention, compatibility files in
.Pa /usr/share/zfs/compatibility.d
are provided by the distribution, and include feature sets
supported by important versions of popular distributions, and feature
sets commonly supported at the start of each year.
Compatibility files in
.Pa /etc/zfs/compatibility.d ,
if present, will take precedence over files with the same name in
.Pa /usr/share/zfs/compatibility.d .
.Pp
If an unrecognized feature is found in these files, an error message will
be shown.
If the unrecognized feature is in a file in
.Pa /etc/zfs/compatibility.d ,
this is treated as an error and processing will stop.
If the unrecognized feature is under
.Pa /usr/share/zfs/compatibility.d ,
this is treated as a warning and processing will continue.
This difference is to allow distributions to include features
which might not be recognized by the currently-installed binaries.
.Pp
Compatibility files may include comments:
any text from
.Sq #
to the end of the line is ignored.
.Pp
.Sy Example :
.Bd -literal -compact -offset 4n
.No example# Nm cat Pa /usr/share/zfs/compatibility.d/grub2
# Features which are supported by GRUB2
async_destroy
bookmarks
embedded_data
empty_bpobj
enabled_txg
extensible_dataset
filesystem_limits
hole_birth
large_blocks
livelist
lz4_compress
spacemap_histogram
zpool_checkpoint
.No example# Nm zpool Cm create Fl o Sy compatibility Ns = Ns Ar grub2 Ar bootpool Ar vdev
.Ed
.Pp
See
.Xr zpool-create 8
and
.Xr zpool-upgrade 8
for more information on how these commands are affected by feature sets.
.
.de feature
.It Sy \\$2
.Bl -tag -compact -width "READ-ONLY COMPATIBLE"
.It GUID
.Sy \\$1:\\$2
.if !"\\$4"" \{\
.It DEPENDENCIES
\fB\\$4\fP\c
.if !"\\$5"" , \fB\\$5\fP\c
.if !"\\$6"" , \fB\\$6\fP\c
.if !"\\$7"" , \fB\\$7\fP\c
.if !"\\$8"" , \fB\\$8\fP\c
.if !"\\$9"" , \fB\\$9\fP\c
.\}
.It READ-ONLY COMPATIBLE
\\$3
.El
.Pp
..
.
.ds instant-never \
.No This feature becomes Sy active No as soon as it is enabled \
and will never return to being Sy enabled .
.
.ds remount-upgrade \
.No Each filesystem will be upgraded automatically when remounted, \
or when a new file is created under that filesystem. \
The upgrade can also be triggered on filesystems via \
Nm zfs Cm set Sy version Ns = Ns Sy current Ar fs . \
No The upgrade process runs in the background and may take a while to complete \
for filesystems containing large amounts of files .
.
.de checksum-spiel
When the
.Sy \\$1
feature is set to
.Sy enabled ,
the administrator can turn on the
.Sy \\$1
checksum on any dataset using
.Nm zfs Cm set Sy checksum Ns = Ns Sy \\$1 Ar dset
.Po see Xr zfs-set 8 Pc .
This feature becomes
.Sy active
once a
.Sy checksum
property has been set to
.Sy \\$1 ,
and will return to being
.Sy enabled
once all filesystems that have ever had their checksum set to
.Sy \\$1
are destroyed.
..
.
.Sh FEATURES
The following features are supported on this system:
.Bl -tag -width Ds
.feature org.zfsonlinux allocation_classes yes
This feature enables support for separate allocation classes.
.Pp
This feature becomes
.Sy active
when a dedicated allocation class vdev
.Pq dedup or special
is created with the
.Nm zpool Cm create No or Nm zpool Cm add No commands .
With device removal, it can be returned to the
.Sy enabled
state if all the dedicated allocation class vdevs are removed.
.
.feature com.delphix async_destroy yes
Destroying a file system requires traversing all of its data in order to
return its used space to the pool.
Without
.Sy async_destroy ,
the file system is not fully removed until all space has been reclaimed.
If the destroy operation is interrupted by a reboot or power outage,
the next attempt to open the pool will need to complete the destroy
operation synchronously.
.Pp
When
.Sy async_destroy
is enabled, the file system's data will be reclaimed by a background process,
allowing the destroy operation to complete
without traversing the entire file system.
The background process is able to resume
interrupted destroys after the pool has been opened, eliminating the need
to finish interrupted destroys as part of the open operation.
The amount of space remaining to be reclaimed by the background process
is available through the
.Sy freeing
property.
.Pp
This feature is only
.Sy active
while
.Sy freeing
is non-zero.
.
.feature org.openzfs blake3 no extensible_dataset
This feature enables the use of the BLAKE3 hash algorithm for checksum and
dedup.
BLAKE3 is a secure hash algorithm focused on high performance.
.Pp
.checksum-spiel blake3
.
.feature com.fudosecurity block_cloning yes
When this feature is enabled ZFS will use block cloning for operations like
.Fn copy_file_range 2 .
Block cloning allows to create multiple references to a single block.
It is much faster than copying the data (as the actual data is neither read nor
written) and takes no additional space.
Blocks can be cloned across datasets under some conditions (like disabled
encryption and equal
.Nm recordsize ) .
.Pp
This feature becomes
.Sy active
when first block is cloned.
When the last cloned block is freed, it goes back to the enabled state.
.feature com.delphix bookmarks yes extensible_dataset
This feature enables use of the
.Nm zfs Cm bookmark
command.
.Pp
This feature is
.Sy active
while any bookmarks exist in the pool.
All bookmarks in the pool can be listed by running
.Nm zfs Cm list Fl t Sy bookmark Fl r Ar poolname .
.
.feature com.datto bookmark_v2 no bookmark extensible_dataset
This feature enables the creation and management of larger bookmarks which are
needed for other features in ZFS.
.Pp
This feature becomes
.Sy active
when a v2 bookmark is created and will be returned to the
.Sy enabled
state when all v2 bookmarks are destroyed.
.
.feature com.delphix bookmark_written no bookmark extensible_dataset bookmark_v2
This feature enables additional bookmark accounting fields, enabling the
.Sy written Ns # Ns Ar bookmark
property
.Pq space written since a bookmark
and estimates of send stream sizes for incrementals from bookmarks.
.Pp
This feature becomes
.Sy active
when a bookmark is created and will be
returned to the
.Sy enabled
state when all bookmarks with these fields are destroyed.
.
.feature org.openzfs device_rebuild yes
This feature enables the ability for the
.Nm zpool Cm attach
and
.Nm zpool Cm replace
commands to perform sequential reconstruction
.Pq instead of healing reconstruction
when resilvering.
.Pp
Sequential reconstruction resilvers a device in LBA order without immediately
verifying the checksums.
Once complete, a scrub is started, which then verifies the checksums.
This approach allows full redundancy to be restored to the pool
in the minimum amount of time.
This two-phase approach will take longer than a healing resilver
when the time to verify the checksums is included.
However, unless there is additional pool damage,
no checksum errors should be reported by the scrub.
This feature is incompatible with raidz configurations.
.
This feature becomes
.Sy active
while a sequential resilver is in progress, and returns to
.Sy enabled
when the resilver completes.
.
.feature com.delphix device_removal no
This feature enables the
.Nm zpool Cm remove
command to remove top-level vdevs,
evacuating them to reduce the total size of the pool.
.Pp
This feature becomes
.Sy active
when the
.Nm zpool Cm remove
command is used
on a top-level vdev, and will never return to being
.Sy enabled .
.
.feature org.openzfs draid no
This feature enables use of the
.Sy draid
vdev type.
dRAID is a variant of RAID-Z which provides integrated distributed
hot spares that allow faster resilvering while retaining the benefits of RAID-Z.
Data, parity, and spare space are organized in redundancy groups
and distributed evenly over all of the devices.
.Pp
This feature becomes
.Sy active
when creating a pool which uses the
.Sy draid
vdev type, or when adding a new
.Sy draid
vdev to an existing pool.
.
.feature org.illumos edonr no extensible_dataset
This feature enables the use of the Edon-R hash algorithm for checksum,
including for nopwrite
.Po if compression is also enabled, an overwrite of
a block whose checksum matches the data being written will be ignored
.Pc .
In an abundance of caution, Edon-R requires verification when used with
dedup:
.Nm zfs Cm set Sy dedup Ns = Ns Sy edonr , Ns Sy verify
.Po see Xr zfs-set 8 Pc .
.Pp
Edon-R is a very high-performance hash algorithm that was part
of the NIST SHA-3 competition.
It provides extremely high hash performance
.Pq over 350% faster than SHA-256 ,
but was not selected because of its unsuitability
as a general purpose secure hash algorithm.
This implementation utilizes the new salted checksumming functionality
in ZFS, which means that the checksum is pre-seeded with a secret
256-bit random key
.Pq stored on the pool
before being fed the data block to be checksummed.
Thus the produced checksums are unique to a given pool,
preventing hash collision attacks on systems with dedup.
.Pp
.checksum-spiel edonr
.
.feature com.delphix embedded_data no
This feature improves the performance and compression ratio of
highly-compressible blocks.
Blocks whose contents can compress to 112 bytes
or smaller can take advantage of this feature.
.Pp
When this feature is enabled, the contents of highly-compressible blocks are
stored in the block
.Dq pointer
itself
.Po a misnomer in this case, as it contains
the compressed data, rather than a pointer to its location on disk
.Pc .
Thus the space of the block
.Pq one sector, typically 512 B or 4 KiB
is saved, and no additional I/O is needed to read and write the data block.
.
\*[instant-never]
.
.feature com.delphix empty_bpobj yes
This feature increases the performance of creating and using a large
number of snapshots of a single filesystem or volume, and also reduces
the disk space required.
.Pp
When there are many snapshots, each snapshot uses many Block Pointer
Objects
.Pq bpobjs
to track blocks associated with that snapshot.
However, in common use cases, most of these bpobjs are empty.
This feature allows us to create each bpobj on-demand,
thus eliminating the empty bpobjs.
.Pp
This feature is
.Sy active
while there are any filesystems, volumes,
or snapshots which were created after enabling this feature.
.
.feature com.delphix enabled_txg yes
Once this feature is enabled, ZFS records the transaction group number
in which new features are enabled.
This has no user-visible impact, but other features may depend on this feature.
.Pp
This feature becomes
.Sy active
as soon as it is enabled and will never return to being
.Sy enabled .
.
.feature com.datto encryption no bookmark_v2 extensible_dataset
This feature enables the creation and management of natively encrypted datasets.
.Pp
This feature becomes
.Sy active
when an encrypted dataset is created and will be returned to the
.Sy enabled
state when all datasets that use this feature are destroyed.
.
.feature com.delphix extensible_dataset no
This feature allows more flexible use of internal ZFS data structures,
and exists for other features to depend on.
.Pp
This feature will be
.Sy active
when the first dependent feature uses it, and will be returned to the
.Sy enabled
state when all datasets that use this feature are destroyed.
.
.feature com.joyent filesystem_limits yes extensible_dataset
This feature enables filesystem and snapshot limits.
These limits can be used to control how many filesystems and/or snapshots
can be created at the point in the tree on which the limits are set.
.Pp
This feature is
.Sy active
once either of the limit properties has been set on a dataset
and will never return to being
.Sy enabled .
.
.feature com.delphix head_errlog no
This feature enables the upgraded version of errlog, which required an on-disk
error log format change.
Now the error log of each head dataset is stored separately in the zap object
and keyed by the head id.
With this feature enabled, every dataset affected by an error block is listed
in the output of
.Nm zpool Cm status .
In case of encrypted filesystems with unloaded keys we are unable to check
their snapshots or clones for errors and these will not be reported.
An "access denied" error will be reported.
.Pp
\*[instant-never]
.
.feature com.delphix hole_birth no enabled_txg
This feature has/had bugs, the result of which is that, if you do a
.Nm zfs Cm send Fl i
.Pq or Fl R , No since it uses Fl i
from an affected dataset, the receiving party will not see any checksum
or other errors, but the resulting destination snapshot
will not match the source.
Its use by
.Nm zfs Cm send Fl i
has been disabled by default
.Po
see
.Sy send_holes_without_birth_time
in
.Xr zfs 4
.Pc .
.Pp
This feature improves performance of incremental sends
.Pq Nm zfs Cm send Fl i
and receives for objects with many holes.
The most common case of hole-filled objects is zvols.
.Pp
An incremental send stream from snapshot
.Sy A No to snapshot Sy B
contains information about every block that changed between
.Sy A No and Sy B .
Blocks which did not change between those snapshots can be
identified and omitted from the stream using a piece of metadata called
the
.Dq block birth time ,
but birth times are not recorded for holes
.Pq blocks filled only with zeroes .
Since holes created after
.Sy A No cannot be distinguished from holes created before Sy A ,
information about every hole in the entire filesystem or zvol
is included in the send stream.
.Pp
For workloads where holes are rare this is not a problem.
However, when incrementally replicating filesystems or zvols with many holes
.Pq for example a zvol formatted with another filesystem
a lot of time will be spent sending and receiving unnecessary information
about holes that already exist on the receiving side.
.Pp
Once the
.Sy hole_birth
feature has been enabled the block birth times
of all new holes will be recorded.
Incremental sends between snapshots created after this feature is enabled
will use this new metadata to avoid sending information about holes that
already exist on the receiving side.
.Pp
\*[instant-never]
.
.feature org.open-zfs large_blocks no extensible_dataset
This feature allows the record size on a dataset to be set larger than 128 KiB.
.Pp
This feature becomes
.Sy active
once a dataset contains a file with a block size larger than 128 KiB,
and will return to being
.Sy enabled
once all filesystems that have ever had their recordsize larger than 128 KiB
are destroyed.
.
.feature org.zfsonlinux large_dnode no extensible_dataset
This feature allows the size of dnodes in a dataset to be set larger than 512 B.
.
This feature becomes
.Sy active
once a dataset contains an object with a dnode larger than 512 B,
which occurs as a result of setting the
.Sy dnodesize
dataset property to a value other than
.Sy legacy .
The feature will return to being
.Sy enabled
once all filesystems that have ever contained a dnode larger than 512 B
are destroyed.
Large dnodes allow more data to be stored in the bonus buffer,
thus potentially improving performance by avoiding the use of spill blocks.
.
.feature com.delphix livelist yes
This feature allows clones to be deleted faster than the traditional method
when a large number of random/sparse writes have been made to the clone.
All blocks allocated and freed after a clone is created are tracked by the
the clone's livelist which is referenced during the deletion of the clone.
The feature is activated when a clone is created and remains
.Sy active
until all clones have been destroyed.
.
.feature com.delphix log_spacemap yes com.delphix:spacemap_v2
This feature improves performance for heavily-fragmented pools,
especially when workloads are heavy in random-writes.
It does so by logging all the metaslab changes on a single spacemap every TXG
instead of scattering multiple writes to all the metaslab spacemaps.
.Pp
\*[instant-never]
.
.feature org.illumos lz4_compress no
.Sy lz4
is a high-performance real-time compression algorithm that
features significantly faster compression and decompression as well as a
higher compression ratio than the older
.Sy lzjb
compression.
Typically,
.Sy lz4
compression is approximately 50% faster on compressible data and 200% faster
on incompressible data than
.Sy lzjb .
It is also approximately 80% faster on decompression,
while giving approximately a 10% better compression ratio.
.Pp
When the
.Sy lz4_compress
feature is set to
.Sy enabled ,
the administrator can turn on
.Sy lz4
compression on any dataset on the pool using the
.Xr zfs-set 8
command.
All newly written metadata will be compressed with the
.Sy lz4
algorithm.
.Pp
\*[instant-never]
.
.feature com.joyent multi_vdev_crash_dump no
This feature allows a dump device to be configured with a pool comprised
of multiple vdevs.
Those vdevs may be arranged in any mirrored or raidz configuration.
.Pp
When the
.Sy multi_vdev_crash_dump
feature is set to
.Sy enabled ,
the administrator can use
.Xr dumpadm 8
to configure a dump device on a pool comprised of multiple vdevs.
.Pp
Under
.Fx
and Linux this feature is unused, but registered for compatibility.
New pools created on these systems will have the feature
.Sy enabled
but will never transition to
.Sy active ,
as this functionality is not required for crash dump support.
Existing pools where this feature is
.Sy active
can be imported.
.
.feature com.delphix obsolete_counts yes device_removal
This feature is an enhancement of
.Sy device_removal ,
which will over time reduce the memory used to track removed devices.
When indirect blocks are freed or remapped,
we note that their part of the indirect mapping is
.Dq obsolete
– no longer needed.
.Pp
This feature becomes
.Sy active
when the
.Nm zpool Cm remove
command is used on a top-level vdev, and will never return to being
.Sy enabled .
.
.feature org.zfsonlinux project_quota yes extensible_dataset
This feature allows administrators to account the spaces and objects usage
information against the project identifier
.Pq ID .
.Pp
The project ID is an object-based attribute.
When upgrading an existing filesystem,
objects without a project ID will be assigned a zero project ID.
When this feature is enabled, newly created objects inherit
their parent directories' project ID if the parent's inherit flag is set
.Pq via Nm chattr Sy [+-]P No or Nm zfs Cm project Fl s Ns | Ns Fl C .
Otherwise, the new object's project ID will be zero.
An object's project ID can be changed at any time by the owner
.Pq or privileged user
via
.Nm chattr Fl p Ar prjid
or
.Nm zfs Cm project Fl p Ar prjid .
.Pp
This feature will become
.Sy active
as soon as it is enabled and will never return to being
.Sy disabled .
\*[remount-upgrade]
.
+.feature org.openzfs raidz_expansion no none
+This feature enables the
+.Nm zpool Cm attach
+subcommand to attach a new device to a RAID-Z group, expanding the total
+amount usable space in the pool.
+See
+.Xr zpool-attach 8 .
+.
.feature com.delphix redaction_bookmarks no bookmarks extensible_dataset
This feature enables the use of redacted
.Nm zfs Cm send Ns s ,
which create redaction bookmarks storing the list of blocks
redacted by the send that created them.
For more information about redacted sends, see
.Xr zfs-send 8 .
.
.feature com.delphix redacted_datasets no extensible_dataset
This feature enables the receiving of redacted
.Nm zfs Cm send
streams, which create redacted datasets when received.
These datasets are missing some of their blocks,
and so cannot be safely mounted, and their contents cannot be safely read.
For more information about redacted receives, see
.Xr zfs-send 8 .
.
+.feature com.delphix redaction_list_spill no redaction_bookmarks
+This feature enables the redaction list created by zfs redact to store
+many more entries.
+It becomes
+.Sy active
+when a redaction list is created with more than 36 entries,
+and returns to being
+.Sy enabled
+when no long redaction lists remain in the pool.
+For more information about redacted sends, see
+.Xr zfs-send 8 .
+.
.feature com.datto resilver_defer yes
This feature allows ZFS to postpone new resilvers if an existing one is already
in progress.
Without this feature, any new resilvers will cause the currently
running one to be immediately restarted from the beginning.
.Pp
This feature becomes
.Sy active
once a resilver has been deferred, and returns to being
.Sy enabled
when the deferred resilver begins.
.
.feature org.illumos sha512 no extensible_dataset
This feature enables the use of the SHA-512/256 truncated hash algorithm
.Pq FIPS 180-4
for checksum and dedup.
The native 64-bit arithmetic of SHA-512 provides an approximate 50%
performance boost over SHA-256 on 64-bit hardware
and is thus a good minimum-change replacement candidate
for systems where hash performance is important,
but these systems cannot for whatever reason utilize the faster
.Sy skein No and Sy edonr
algorithms.
.Pp
.checksum-spiel sha512
.
.feature org.illumos skein no extensible_dataset
This feature enables the use of the Skein hash algorithm for checksum and dedup.
Skein is a high-performance secure hash algorithm that was a
finalist in the NIST SHA-3 competition.
It provides a very high security margin and high performance on 64-bit hardware
.Pq 80% faster than SHA-256 .
This implementation also utilizes the new salted checksumming
functionality in ZFS, which means that the checksum is pre-seeded with a
secret 256-bit random key
.Pq stored on the pool
before being fed the data block to be checksummed.
Thus the produced checksums are unique to a given pool,
preventing hash collision attacks on systems with dedup.
.Pp
.checksum-spiel skein
.
.feature com.delphix spacemap_histogram yes
This features allows ZFS to maintain more information about how free space
is organized within the pool.
If this feature is
.Sy enabled ,
it will be activated when a new space map object is created, or
an existing space map is upgraded to the new format,
and never returns back to being
.Sy enabled .
.
.feature com.delphix spacemap_v2 yes
This feature enables the use of the new space map encoding which
consists of two words
.Pq instead of one
whenever it is advantageous.
The new encoding allows space maps to represent large regions of
space more efficiently on-disk while also increasing their maximum
addressable offset.
.Pp
This feature becomes
.Sy active
once it is
.Sy enabled ,
and never returns back to being
.Sy enabled .
.
.feature org.zfsonlinux userobj_accounting yes extensible_dataset
This feature allows administrators to account the object usage information
by user and group.
.Pp
\*[instant-never]
\*[remount-upgrade]
.
.feature com.klarasystems vdev_zaps_v2 no
This feature creates a ZAP object for the root vdev.
.Pp
This feature becomes active after the next
.Nm zpool Cm import
or
.Nm zpool reguid .
.
Properties can be retrieved or set on the root vdev using
.Nm zpool Cm get
and
.Nm zpool Cm set
with
.Sy root
as the vdev name which is an alias for
.Sy root-0 .
.feature org.openzfs zilsaxattr yes extensible_dataset
This feature enables
.Sy xattr Ns = Ns Sy sa
extended attribute logging in the ZIL.
If enabled, extended attribute changes
.Pq both Sy xattrdir Ns = Ns Sy dir No and Sy xattr Ns = Ns Sy sa
are guaranteed to be durable if either the dataset had
.Sy sync Ns = Ns Sy always
set at the time the changes were made, or
.Xr sync 2
is called on the dataset after the changes were made.
.Pp
This feature becomes
.Sy active
when a ZIL is created for at least one dataset and will be returned to the
.Sy enabled
state when it is destroyed for all datasets that use this feature.
.
.feature com.delphix zpool_checkpoint yes
This feature enables the
.Nm zpool Cm checkpoint
command that can checkpoint the state of the pool
at the time it was issued and later rewind back to it or discard it.
.Pp
This feature becomes
.Sy active
when the
.Nm zpool Cm checkpoint
command is used to checkpoint the pool.
The feature will only return back to being
.Sy enabled
when the pool is rewound or the checkpoint has been discarded.
.
.feature org.freebsd zstd_compress no extensible_dataset
.Sy zstd
is a high-performance compression algorithm that features a
combination of high compression ratios and high speed.
Compared to
.Sy gzip ,
.Sy zstd
offers slightly better compression at much higher speeds.
Compared to
.Sy lz4 ,
.Sy zstd
offers much better compression while being only modestly slower.
Typically,
.Sy zstd
compression speed ranges from 250 to 500 MB/s per thread
and decompression speed is over 1 GB/s per thread.
.Pp
When the
.Sy zstd
feature is set to
.Sy enabled ,
the administrator can turn on
.Sy zstd
compression of any dataset using
.Nm zfs Cm set Sy compress Ns = Ns Sy zstd Ar dset
.Po see Xr zfs-set 8 Pc .
This feature becomes
.Sy active
once a
.Sy compress
property has been set to
.Sy zstd ,
and will return to being
.Sy enabled
once all filesystems that have ever had their
.Sy compress
property set to
.Sy zstd
are destroyed.
-.
-.feature com.delphix redaction_list_spill no redaction_bookmarks
-This feature enables the redaction list created by zfs redact to store
-many more entries.
-It becomes
-.Sy active
-when a redaction list is created with more than 36 entries,
-and returns to being
-.Sy enabled
-when no long redaction lists remain in the pool.
-For more information about redacted sends, see
-.Xr zfs-send 8 .
.El
.
.Sh SEE ALSO
.Xr zfs 8 ,
.Xr zpool 8
diff --git a/sys/contrib/openzfs/man/man8/zpool-attach.8 b/sys/contrib/openzfs/man/man8/zpool-attach.8
index 73535cbdf108..22b1369b6853 100644
--- a/sys/contrib/openzfs/man/man8/zpool-attach.8
+++ b/sys/contrib/openzfs/man/man8/zpool-attach.8
@@ -1,98 +1,141 @@
.\"
.\" CDDL HEADER START
.\"
.\" The contents of this file are subject to the terms of the
.\" Common Development and Distribution License (the "License").
.\" You may not use this file except in compliance with the License.
.\"
.\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
.\" or https://opensource.org/licenses/CDDL-1.0.
.\" See the License for the specific language governing permissions
.\" and limitations under the License.
.\"
.\" When distributing Covered Code, include this CDDL HEADER in each
.\" file and include the License file at usr/src/OPENSOLARIS.LICENSE.
.\" If applicable, add the following below this CDDL HEADER, with the
.\" fields enclosed by brackets "[]" replaced with your own identifying
.\" information: Portions Copyright [yyyy] [name of copyright owner]
.\"
.\" CDDL HEADER END
.\"
.\" Copyright (c) 2007, Sun Microsystems, Inc. All Rights Reserved.
.\" Copyright (c) 2012, 2018 by Delphix. All rights reserved.
.\" Copyright (c) 2012 Cyril Plisko. All Rights Reserved.
.\" Copyright (c) 2017 Datto Inc.
.\" Copyright (c) 2018 George Melikov. All Rights Reserved.
.\" Copyright 2017 Nexenta Systems, Inc.
.\" Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
.\"
-.Dd May 15, 2020
+.Dd June 28, 2023
.Dt ZPOOL-ATTACH 8
.Os
.
.Sh NAME
.Nm zpool-attach
.Nd attach new device to existing ZFS vdev
.Sh SYNOPSIS
.Nm zpool
.Cm attach
.Op Fl fsw
.Oo Fl o Ar property Ns = Ns Ar value Oc
.Ar pool device new_device
.
.Sh DESCRIPTION
Attaches
.Ar new_device
to the existing
.Ar device .
-The existing device cannot be part of a raidz configuration.
+The behavior differs depending on if the existing
+.Ar device
+is a RAID-Z device, or a mirror/plain device.
+.Pp
+If the existing device is a mirror or plain device
+.Pq e.g. specified as Qo Li sda Qc or Qq Li mirror-7 ,
+the new device will be mirrored with the existing device, a resilver will be
+initiated, and the new device will contribute to additional redundancy once the
+resilver completes.
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 and any running scrub is cancelled.
+.Pp
+If the existing device is a RAID-Z device
+.Pq e.g. specified as Qq Ar raidz2-0 ,
+the new device will become part of that RAID-Z group.
+A "raidz expansion" will be initiated, and once the expansion completes,
+the new device will contribute additional space to the RAID-Z group.
+The expansion entails reading all allocated space from existing disks in the
+RAID-Z group, and rewriting it to the new disks in the RAID-Z group (including
+the newly added
+.Ar device ) .
+Its progress can be monitored with
+.Nm zpool Cm status .
+.Pp
+Data redundancy is maintained during and after the expansion.
+If a disk fails while the expansion is in progress, the expansion pauses until
+the health of the RAID-Z vdev is restored (e.g. by replacing the failed disk
+and waiting for reconstruction to complete).
+Expansion does not change the number of failures that can be tolerated
+without data loss (e.g. a RAID-Z2 is still a RAID-Z2 even after expansion).
+A RAID-Z vdev can be expanded multiple times.
+.Pp
+After the expansion completes, old blocks retain their old data-to-parity
+ratio
+.Pq e.g. 5-wide RAID-Z2 has 3 data and 2 parity
+but distributed among the larger set of disks.
+New blocks will be written with the new data-to-parity ratio (e.g. a 5-wide
+RAID-Z2 which has been expanded once to 6-wide, has 4 data and 2 parity).
+However, the vdev's assumed parity ratio does not change, so slightly less
+space than is expected may be reported for newly-written blocks, according to
+.Nm zfs Cm list ,
+.Nm df ,
+.Nm ls Fl s ,
+and similar tools.
+.Pp
+A pool-wide scrub is initiated at the end of the expansion in order to verify
+the checksums of all blocks which have been copied during the expansion.
.Bl -tag -width Ds
.It Fl f
Forces use of
.Ar new_device ,
even if it appears to be in use.
Not all devices can be overridden in this manner.
.It Fl o Ar property Ns = Ns Ar value
Sets the given pool properties.
See the
.Xr zpoolprops 7
manual page for a list of valid properties that can be set.
The only property supported at the moment is
.Sy ashift .
.It Fl s
-The
+When attaching to a mirror or plain device, the
.Ar new_device
is reconstructed sequentially to restore redundancy as quickly as possible.
Checksums are not verified during sequential reconstruction so a scrub is
started when the resilver completes.
-Sequential reconstruction is not supported for raidz configurations.
.It Fl w
Waits until
.Ar new_device
-has finished resilvering before returning.
+has finished resilvering or expanding before returning.
.El
.
.Sh SEE ALSO
.Xr zpool-add 8 ,
.Xr zpool-detach 8 ,
.Xr zpool-import 8 ,
.Xr zpool-initialize 8 ,
.Xr zpool-online 8 ,
.Xr zpool-replace 8 ,
.Xr zpool-resilver 8
diff --git a/sys/contrib/openzfs/man/man8/zpool-wait.8 b/sys/contrib/openzfs/man/man8/zpool-wait.8
index 683b0141425c..d646d9cc3312 100644
--- a/sys/contrib/openzfs/man/man8/zpool-wait.8
+++ b/sys/contrib/openzfs/man/man8/zpool-wait.8
@@ -1,116 +1,118 @@
.\"
.\" CDDL HEADER START
.\"
.\" The contents of this file are subject to the terms of the
.\" Common Development and Distribution License (the "License").
.\" You may not use this file except in compliance with the License.
.\"
.\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
.\" or https://opensource.org/licenses/CDDL-1.0.
.\" See the License for the specific language governing permissions
.\" and limitations under the License.
.\"
.\" When distributing Covered Code, include this CDDL HEADER in each
.\" file and include the License file at usr/src/OPENSOLARIS.LICENSE.
.\" If applicable, add the following below this CDDL HEADER, with the
.\" fields enclosed by brackets "[]" replaced with your own identifying
.\" information: Portions Copyright [yyyy] [name of copyright owner]
.\"
.\" CDDL HEADER END
.\"
.\"
.\" Copyright (c) 2007, Sun Microsystems, Inc. All Rights Reserved.
-.\" Copyright (c) 2012, 2018 by Delphix. All rights reserved.
+.\" Copyright (c) 2012, 2021 by Delphix. All rights reserved.
.\" Copyright (c) 2012 Cyril Plisko. All Rights Reserved.
.\" Copyright (c) 2017 Datto Inc.
.\" Copyright (c) 2018 George Melikov. All Rights Reserved.
.\" Copyright 2017 Nexenta Systems, Inc.
.\" Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
.\"
.Dd May 27, 2021
.Dt ZPOOL-WAIT 8
.Os
.
.Sh NAME
.Nm zpool-wait
.Nd wait for activity to stop in a ZFS storage pool
.Sh SYNOPSIS
.Nm zpool
.Cm wait
.Op Fl Hp
.Op Fl T Sy u Ns | Ns Sy d
.Op Fl t Ar activity Ns Oo , Ns Ar activity Ns Oc Ns …
.Ar pool
.Op Ar interval
.
.Sh DESCRIPTION
Waits until all background activity of the given types has ceased in the given
pool.
The activity could cease because it has completed, or because it has been
paused or canceled by a user, or because the pool has been exported or
destroyed.
If no activities are specified, the command waits until background activity of
every type listed below has ceased.
If there is no activity of the given types in progress, the command returns
immediately.
.Pp
These are the possible values for
.Ar activity ,
along with what each one waits for:
-.Bl -tag -compact -offset Ds -width "initialize"
+.Bl -tag -compact -offset Ds -width "raidz_expand"
.It Sy discard
Checkpoint to be discarded
.It Sy free
.Sy freeing
property to become
.Sy 0
.It Sy initialize
All initializations to cease
.It Sy replace
All device replacements to cease
.It Sy remove
Device removal to cease
.It Sy resilver
Resilver to cease
.It Sy scrub
Scrub to cease
.It Sy trim
Manual trim to cease
+.It Sy raidz_expand
+Attaching to a RAID-Z vdev to complete
.El
.Pp
If an
.Ar interval
is provided, the amount of work remaining, in bytes, for each activity is
printed every
.Ar interval
seconds.
.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 p
Display numbers in parsable (exact) values.
.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 .
.El
.
.Sh SEE ALSO
.Xr zpool-checkpoint 8 ,
.Xr zpool-initialize 8 ,
.Xr zpool-remove 8 ,
.Xr zpool-replace 8 ,
.Xr zpool-resilver 8 ,
.Xr zpool-scrub 8 ,
.Xr zpool-status 8 ,
.Xr zpool-trim 8
diff --git a/sys/contrib/openzfs/module/Kbuild.in b/sys/contrib/openzfs/module/Kbuild.in
index c132171592a8..6408fb106717 100644
--- a/sys/contrib/openzfs/module/Kbuild.in
+++ b/sys/contrib/openzfs/module/Kbuild.in
@@ -1,498 +1,499 @@
# When integrated in to a monolithic kernel the spl module must appear
# first. This ensures its module initialization function is run before
# any of the other module initialization functions which depend on it.
ZFS_MODULE_CFLAGS += -std=gnu99 -Wno-declaration-after-statement
ZFS_MODULE_CFLAGS += -Wmissing-prototypes
ZFS_MODULE_CFLAGS += @KERNEL_DEBUG_CFLAGS@ @NO_FORMAT_ZERO_LENGTH@
ifneq ($(KBUILD_EXTMOD),)
zfs_include = @abs_top_srcdir@/include
icp_include = @abs_srcdir@/icp/include
zstd_include = @abs_srcdir@/zstd/include
ZFS_MODULE_CFLAGS += -include @abs_top_builddir@/zfs_config.h
ZFS_MODULE_CFLAGS += -I@abs_top_builddir@/include
src = @abs_srcdir@
obj = @abs_builddir@
else
zfs_include = $(srctree)/include/zfs
icp_include = $(srctree)/$(src)/icp/include
zstd_include = $(srctree)/$(src)/zstd/include
ZFS_MODULE_CFLAGS += -include $(zfs_include)/zfs_config.h
endif
ZFS_MODULE_CFLAGS += -I$(zfs_include)/os/linux/kernel
ZFS_MODULE_CFLAGS += -I$(zfs_include)/os/linux/spl
ZFS_MODULE_CFLAGS += -I$(zfs_include)/os/linux/zfs
ZFS_MODULE_CFLAGS += -I$(zfs_include)
ZFS_MODULE_CPPFLAGS += -D_KERNEL
ZFS_MODULE_CPPFLAGS += @KERNEL_DEBUG_CPPFLAGS@
# KASAN enables -Werror=frame-larger-than=1024, which
# breaks oh so many parts of our build.
ifeq ($(CONFIG_KASAN),y)
ZFS_MODULE_CFLAGS += -Wno-error=frame-larger-than=
endif
# Generated binary search code is particularly bad with this optimization.
# Oddly, range_tree.c is not affected when unrolling is not done and dsl_scan.c
# is not affected when unrolling is done.
# Disable it until the following upstream issue is resolved:
# https://github.com/llvm/llvm-project/issues/62790
ifeq ($(CONFIG_X86),y)
ifeq ($(CONFIG_CC_IS_CLANG),y)
CFLAGS_zfs/dsl_scan.o += -mllvm -x86-cmov-converter=false
CFLAGS_zfs/metaslab.o += -mllvm -x86-cmov-converter=false
CFLAGS_zfs/range_tree.o += -mllvm -x86-cmov-converter=false
CFLAGS_zfs/zap_micro.o += -mllvm -x86-cmov-converter=false
endif
endif
ifneq ($(KBUILD_EXTMOD),)
@CONFIG_QAT_TRUE@ZFS_MODULE_CFLAGS += -I@QAT_SRC@/include
@CONFIG_QAT_TRUE@KBUILD_EXTRA_SYMBOLS += @QAT_SYMBOLS@
endif
asflags-y := $(ZFS_MODULE_CFLAGS) $(ZFS_MODULE_CPPFLAGS)
ccflags-y := $(ZFS_MODULE_CFLAGS) $(ZFS_MODULE_CPPFLAGS)
ifeq ($(CONFIG_ARM64),y)
CFLAGS_REMOVE_zcommon/zfs_fletcher_aarch64_neon.o += -mgeneral-regs-only
CFLAGS_REMOVE_zfs/vdev_raidz_math_aarch64_neon.o += -mgeneral-regs-only
CFLAGS_REMOVE_zfs/vdev_raidz_math_aarch64_neonx2.o += -mgeneral-regs-only
endif
# Suppress unused-value warnings in sparc64 architecture headers
ccflags-$(CONFIG_SPARC64) += -Wno-unused-value
obj-$(CONFIG_ZFS) := spl.o zfs.o
SPL_OBJS := \
spl-atomic.o \
spl-condvar.o \
spl-cred.o \
spl-err.o \
spl-generic.o \
spl-kmem-cache.o \
spl-kmem.o \
spl-kstat.o \
spl-proc.o \
spl-procfs-list.o \
spl-taskq.o \
spl-thread.o \
spl-trace.o \
spl-tsd.o \
spl-vmem.o \
spl-xdr.o \
spl-zlib.o \
spl-zone.o
spl-objs += $(addprefix os/linux/spl/,$(SPL_OBJS))
zfs-objs += avl/avl.o
ICP_OBJS := \
algs/aes/aes_impl.o \
algs/aes/aes_impl_generic.o \
algs/aes/aes_modes.o \
algs/blake3/blake3.o \
algs/blake3/blake3_generic.o \
algs/blake3/blake3_impl.o \
algs/edonr/edonr.o \
algs/modes/cbc.o \
algs/modes/ccm.o \
algs/modes/ctr.o \
algs/modes/ecb.o \
algs/modes/gcm.o \
algs/modes/gcm_generic.o \
algs/modes/modes.o \
algs/sha2/sha2_generic.o \
algs/sha2/sha256_impl.o \
algs/sha2/sha512_impl.o \
algs/skein/skein.o \
algs/skein/skein_block.o \
algs/skein/skein_iv.o \
api/kcf_cipher.o \
api/kcf_ctxops.o \
api/kcf_mac.o \
core/kcf_callprov.o \
core/kcf_mech_tabs.o \
core/kcf_prov_lib.o \
core/kcf_prov_tabs.o \
core/kcf_sched.o \
illumos-crypto.o \
io/aes.o \
io/sha2_mod.o \
io/skein_mod.o \
spi/kcf_spi.o
ICP_OBJS_X86_64 := \
asm-x86_64/aes/aes_aesni.o \
asm-x86_64/aes/aes_amd64.o \
asm-x86_64/aes/aeskey.o \
asm-x86_64/blake3/blake3_avx2.o \
asm-x86_64/blake3/blake3_avx512.o \
asm-x86_64/blake3/blake3_sse2.o \
asm-x86_64/blake3/blake3_sse41.o \
asm-x86_64/sha2/sha256-x86_64.o \
asm-x86_64/sha2/sha512-x86_64.o \
asm-x86_64/modes/aesni-gcm-x86_64.o \
asm-x86_64/modes/gcm_pclmulqdq.o \
asm-x86_64/modes/ghash-x86_64.o
ICP_OBJS_X86 := \
algs/aes/aes_impl_aesni.o \
algs/aes/aes_impl_x86-64.o \
algs/modes/gcm_pclmulqdq.o
ICP_OBJS_ARM := \
asm-arm/sha2/sha256-armv7.o \
asm-arm/sha2/sha512-armv7.o
ICP_OBJS_ARM64 := \
asm-aarch64/blake3/b3_aarch64_sse2.o \
asm-aarch64/blake3/b3_aarch64_sse41.o \
asm-aarch64/sha2/sha256-armv8.o \
asm-aarch64/sha2/sha512-armv8.o
ICP_OBJS_PPC_PPC64 := \
asm-ppc64/blake3/b3_ppc64le_sse2.o \
asm-ppc64/blake3/b3_ppc64le_sse41.o \
asm-ppc64/sha2/sha256-p8.o \
asm-ppc64/sha2/sha512-p8.o \
asm-ppc64/sha2/sha256-ppc.o \
asm-ppc64/sha2/sha512-ppc.o
zfs-objs += $(addprefix icp/,$(ICP_OBJS))
zfs-$(CONFIG_X86) += $(addprefix icp/,$(ICP_OBJS_X86))
zfs-$(CONFIG_UML_X86)+= $(addprefix icp/,$(ICP_OBJS_X86))
zfs-$(CONFIG_X86_64) += $(addprefix icp/,$(ICP_OBJS_X86_64))
zfs-$(CONFIG_ARM) += $(addprefix icp/,$(ICP_OBJS_ARM))
zfs-$(CONFIG_ARM64) += $(addprefix icp/,$(ICP_OBJS_ARM64))
zfs-$(CONFIG_PPC) += $(addprefix icp/,$(ICP_OBJS_PPC_PPC64))
zfs-$(CONFIG_PPC64) += $(addprefix icp/,$(ICP_OBJS_PPC_PPC64))
$(addprefix $(obj)/icp/,$(ICP_OBJS) $(ICP_OBJS_X86) $(ICP_OBJS_X86_64) \
$(ICP_OBJS_ARM64) $(ICP_OBJS_PPC_PPC64)) : asflags-y += -I$(icp_include) -I$(zfs_include)/os/linux/spl -I$(zfs_include)
$(addprefix $(obj)/icp/,$(ICP_OBJS) $(ICP_OBJS_X86) $(ICP_OBJS_X86_64) \
$(ICP_OBJS_ARM64) $(ICP_OBJS_PPC_PPC64)) : ccflags-y += -I$(icp_include) -I$(zfs_include)/os/linux/spl -I$(zfs_include)
# Suppress objtool "return with modified stack frame" warnings.
OBJECT_FILES_NON_STANDARD_aesni-gcm-x86_64.o := y
# Suppress objtool "unsupported stack pointer realignment" warnings.
# See #6950 for the reasoning.
OBJECT_FILES_NON_STANDARD_sha256-x86_64.o := y
OBJECT_FILES_NON_STANDARD_sha512-x86_64.o := y
LUA_OBJS := \
lapi.o \
lauxlib.o \
lbaselib.o \
lcode.o \
lcompat.o \
lcorolib.o \
lctype.o \
ldebug.o \
ldo.o \
lfunc.o \
lgc.o \
llex.o \
lmem.o \
lobject.o \
lopcodes.o \
lparser.o \
lstate.o \
lstring.o \
lstrlib.o \
ltable.o \
ltablib.o \
ltm.o \
lvm.o \
lzio.o \
setjmp/setjmp.o
zfs-objs += $(addprefix lua/,$(LUA_OBJS))
NVPAIR_OBJS := \
fnvpair.o \
nvpair.o \
nvpair_alloc_fixed.o \
nvpair_alloc_spl.o
zfs-objs += $(addprefix nvpair/,$(NVPAIR_OBJS))
UNICODE_OBJS := \
u8_textprep.o \
uconv.o
zfs-objs += $(addprefix unicode/,$(UNICODE_OBJS))
ZCOMMON_OBJS := \
cityhash.o \
zfeature_common.o \
zfs_comutil.o \
zfs_deleg.o \
zfs_fletcher.o \
zfs_fletcher_superscalar.o \
zfs_fletcher_superscalar4.o \
zfs_namecheck.o \
zfs_prop.o \
zpool_prop.o \
zprop_common.o
ZCOMMON_OBJS_X86 := \
zfs_fletcher_avx512.o \
zfs_fletcher_intel.o \
zfs_fletcher_sse.o
ZCOMMON_OBJS_ARM64 := \
zfs_fletcher_aarch64_neon.o
zfs-objs += $(addprefix zcommon/,$(ZCOMMON_OBJS))
zfs-$(CONFIG_X86) += $(addprefix zcommon/,$(ZCOMMON_OBJS_X86))
zfs-$(CONFIG_UML_X86)+= $(addprefix zcommon/,$(ZCOMMON_OBJS_X86))
zfs-$(CONFIG_ARM64) += $(addprefix zcommon/,$(ZCOMMON_OBJS_ARM64))
# Zstd uses -O3 by default, so we should follow
ZFS_ZSTD_FLAGS := -O3
# -fno-tree-vectorize gets set for gcc in zstd/common/compiler.h
# Set it for other compilers, too.
ZFS_ZSTD_FLAGS += -fno-tree-vectorize
# SSE register return with SSE disabled if -march=znverX is passed
ZFS_ZSTD_FLAGS += -U__BMI__
# Quiet warnings about frame size due to unused code in unmodified zstd lib
ZFS_ZSTD_FLAGS += -Wframe-larger-than=20480
ZSTD_OBJS := \
zfs_zstd.o \
zstd_sparc.o
ZSTD_UPSTREAM_OBJS := \
lib/common/entropy_common.o \
lib/common/error_private.o \
lib/common/fse_decompress.o \
lib/common/pool.o \
lib/common/zstd_common.o \
lib/compress/fse_compress.o \
lib/compress/hist.o \
lib/compress/huf_compress.o \
lib/compress/zstd_compress.o \
lib/compress/zstd_compress_literals.o \
lib/compress/zstd_compress_sequences.o \
lib/compress/zstd_compress_superblock.o \
lib/compress/zstd_double_fast.o \
lib/compress/zstd_fast.o \
lib/compress/zstd_lazy.o \
lib/compress/zstd_ldm.o \
lib/compress/zstd_opt.o \
lib/decompress/huf_decompress.o \
lib/decompress/zstd_ddict.o \
lib/decompress/zstd_decompress.o \
lib/decompress/zstd_decompress_block.o
zfs-objs += $(addprefix zstd/,$(ZSTD_OBJS) $(ZSTD_UPSTREAM_OBJS))
# Disable aarch64 neon SIMD instructions for kernel mode
$(addprefix $(obj)/zstd/,$(ZSTD_OBJS) $(ZSTD_UPSTREAM_OBJS)) : ccflags-y += -I$(zstd_include) $(ZFS_ZSTD_FLAGS)
$(addprefix $(obj)/zstd/,$(ZSTD_OBJS) $(ZSTD_UPSTREAM_OBJS)) : asflags-y += -I$(zstd_include)
$(addprefix $(obj)/zstd/,$(ZSTD_UPSTREAM_OBJS)) : ccflags-y += -include $(zstd_include)/aarch64_compat.h -include $(zstd_include)/zstd_compat_wrapper.h -Wp,-w
$(obj)/zstd/zfs_zstd.o : ccflags-y += -include $(zstd_include)/zstd_compat_wrapper.h
ZFS_OBJS := \
abd.o \
aggsum.o \
arc.o \
blake3_zfs.o \
blkptr.o \
bplist.o \
bpobj.o \
bptree.o \
bqueue.o \
brt.o \
btree.o \
dataset_kstats.o \
dbuf.o \
dbuf_stats.o \
ddt.o \
ddt_zap.o \
dmu.o \
dmu_diff.o \
dmu_object.o \
dmu_objset.o \
dmu_recv.o \
dmu_redact.o \
dmu_send.o \
dmu_traverse.o \
dmu_tx.o \
dmu_zfetch.o \
dnode.o \
dnode_sync.o \
dsl_bookmark.o \
dsl_crypt.o \
dsl_dataset.o \
dsl_deadlist.o \
dsl_deleg.o \
dsl_destroy.o \
dsl_dir.o \
dsl_pool.o \
dsl_prop.o \
dsl_scan.o \
dsl_synctask.o \
dsl_userhold.o \
edonr_zfs.o \
fm.o \
gzip.o \
hkdf.o \
lz4.o \
lz4_zfs.o \
lzjb.o \
metaslab.o \
mmp.o \
multilist.o \
objlist.o \
pathname.o \
range_tree.o \
refcount.o \
rrwlock.o \
sa.o \
sha2_zfs.o \
skein_zfs.o \
spa.o \
spa_checkpoint.o \
spa_config.o \
spa_errlog.o \
spa_history.o \
spa_log_spacemap.o \
spa_misc.o \
spa_stats.o \
space_map.o \
space_reftree.o \
txg.o \
uberblock.o \
unique.o \
vdev.o \
vdev_draid.o \
vdev_draid_rand.o \
vdev_indirect.o \
vdev_indirect_births.o \
vdev_indirect_mapping.o \
vdev_initialize.o \
vdev_label.o \
vdev_mirror.o \
vdev_missing.o \
vdev_queue.o \
vdev_raidz.o \
vdev_raidz_math.o \
vdev_raidz_math_scalar.o \
vdev_rebuild.o \
vdev_removal.o \
vdev_root.o \
vdev_trim.o \
zap.o \
zap_leaf.o \
zap_micro.o \
zcp.o \
zcp_get.o \
zcp_global.o \
zcp_iter.o \
zcp_set.o \
zcp_synctask.o \
zfeature.o \
zfs_byteswap.o \
zfs_chksum.o \
zfs_fm.o \
zfs_fuid.o \
zfs_impl.o \
zfs_ioctl.o \
zfs_log.o \
zfs_onexit.o \
zfs_quota.o \
zfs_ratelimit.o \
zfs_replay.o \
zfs_rlock.o \
zfs_sa.o \
zfs_vnops.o \
zil.o \
zio.o \
zio_checksum.o \
zio_compress.o \
zio_inject.o \
zle.o \
zrlock.o \
zthr.o \
zvol.o
ZFS_OBJS_OS := \
abd_os.o \
arc_os.o \
mmp_os.o \
policy.o \
qat.o \
qat_compress.o \
qat_crypt.o \
spa_misc_os.o \
trace.o \
vdev_disk.o \
vdev_file.o \
+ vdev_label_os.o \
zfs_acl.o \
zfs_ctldir.o \
zfs_debug.o \
zfs_dir.o \
zfs_file_os.o \
zfs_ioctl_os.o \
zfs_racct.o \
zfs_sysfs.o \
zfs_uio.o \
zfs_vfsops.o \
zfs_vnops_os.o \
zfs_znode.o \
zio_crypt.o \
zpl_ctldir.o \
zpl_export.o \
zpl_file.o \
zpl_file_range.o \
zpl_inode.o \
zpl_super.o \
zpl_xattr.o \
zvol_os.o
ZFS_OBJS_X86 := \
vdev_raidz_math_avx2.o \
vdev_raidz_math_avx512bw.o \
vdev_raidz_math_avx512f.o \
vdev_raidz_math_sse2.o \
vdev_raidz_math_ssse3.o
ZFS_OBJS_ARM64 := \
vdev_raidz_math_aarch64_neon.o \
vdev_raidz_math_aarch64_neonx2.o
ZFS_OBJS_PPC_PPC64 := \
vdev_raidz_math_powerpc_altivec.o
zfs-objs += $(addprefix zfs/,$(ZFS_OBJS)) $(addprefix os/linux/zfs/,$(ZFS_OBJS_OS))
zfs-$(CONFIG_X86) += $(addprefix zfs/,$(ZFS_OBJS_X86))
zfs-$(CONFIG_UML_X86)+= $(addprefix zfs/,$(ZFS_OBJS_X86))
zfs-$(CONFIG_ARM64) += $(addprefix zfs/,$(ZFS_OBJS_ARM64))
zfs-$(CONFIG_PPC) += $(addprefix zfs/,$(ZFS_OBJS_PPC_PPC64))
zfs-$(CONFIG_PPC64) += $(addprefix zfs/,$(ZFS_OBJS_PPC_PPC64))
# Suppress incorrect warnings from versions of objtool which are not
# aware of x86 EVEX prefix instructions used for AVX512.
OBJECT_FILES_NON_STANDARD_vdev_raidz_math_avx512bw.o := y
OBJECT_FILES_NON_STANDARD_vdev_raidz_math_avx512f.o := y
ifeq ($(CONFIG_ALTIVEC),y)
$(obj)/zfs/vdev_raidz_math_powerpc_altivec.o : c_flags += -maltivec
endif
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/vdev_label_os.c b/sys/contrib/openzfs/module/os/freebsd/zfs/vdev_label_os.c
index bc856b930235..338982ff6873 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/vdev_label_os.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/vdev_label_os.c
@@ -1,74 +1,133 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/dmu.h>
#include <sys/zap.h>
#include <sys/vdev.h>
#include <sys/vdev_os.h>
#include <sys/vdev_impl.h>
#include <sys/uberblock_impl.h>
#include <sys/metaslab.h>
#include <sys/metaslab_impl.h>
#include <sys/zio.h>
#include <sys/dsl_scan.h>
#include <sys/abd.h>
#include <sys/fs/zfs.h>
int
vdev_label_write_pad2(vdev_t *vd, const char *buf, size_t size)
{
spa_t *spa = vd->vdev_spa;
zio_t *zio;
abd_t *pad2;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL;
int error;
if (size > VDEV_PAD_SIZE)
return (EINVAL);
if (!vd->vdev_ops->vdev_op_leaf)
return (ENODEV);
if (vdev_is_dead(vd))
return (ENXIO);
ASSERT3U(spa_config_held(spa, SCL_ALL, RW_WRITER), ==, SCL_ALL);
pad2 = abd_alloc_for_io(VDEV_PAD_SIZE, B_TRUE);
abd_zero(pad2, VDEV_PAD_SIZE);
abd_copy_from_buf(pad2, buf, size);
retry:
zio = zio_root(spa, NULL, NULL, flags);
vdev_label_write(zio, vd, 0, pad2,
offsetof(vdev_label_t, vl_be),
VDEV_PAD_SIZE, NULL, NULL, flags);
error = zio_wait(zio);
if (error != 0 && !(flags & ZIO_FLAG_TRYHARD)) {
flags |= ZIO_FLAG_TRYHARD;
goto retry;
}
abd_free(pad2);
return (error);
}
+
+static void
+vdev_child_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);
+}
+
+/*
+ * Check if the reserved boot area is in-use.
+ *
+ * When booting FreeBSD with an MBR partition with ZFS, the zfsboot file
+ * (which understands the ZFS file system) is written to the ZFS BOOT
+ * reserve area (at offset 512K). We check for that here before attaching
+ * a disk to raidz which would then corrupt this boot data.
+ */
+int
+vdev_check_boot_reserve(spa_t *spa, vdev_t *childvd)
+{
+ ASSERT(childvd->vdev_ops->vdev_op_leaf);
+
+ size_t size = SPA_MINBLOCKSIZE;
+ abd_t *abd = abd_alloc_linear(size, B_FALSE);
+
+ zio_t *pio = zio_root(spa, NULL, NULL, 0);
+ /*
+ * Note: zio_vdev_child_io() adds VDEV_LABEL_START_SIZE to the offset
+ * to calculate the physical offset to write to. Passing in a negative
+ * offset lets us access the boot area.
+ */
+ zio_nowait(zio_vdev_child_io(pio, NULL, childvd,
+ VDEV_BOOT_OFFSET - VDEV_LABEL_START_SIZE, abd, size, ZIO_TYPE_READ,
+ ZIO_PRIORITY_ASYNC_READ, 0, vdev_child_done, pio));
+ zio_wait(pio);
+
+ unsigned char *buf = abd_to_buf(abd);
+
+ /*
+ * The BTX server has a special header at the begining.
+ *
+ * btx_hdr: .byte 0xeb # Machine ID
+ * .byte 0xe # Header size
+ * .ascii "BTX" # Magic
+ * .byte 0x1 # Major version
+ * .byte 0x2 # Minor version
+ * .byte BTX_FLAGS # Flags
+ */
+ if (buf[0] == 0xeb && buf[1] == 0x0e &&
+ buf[2] == 'B' && buf[3] == 'T' && buf[4] == 'X') {
+ abd_free(abd);
+ return (EBUSY);
+ }
+
+ abd_free(abd);
+ return (0);
+}
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/vdev_label_os.c b/sys/contrib/openzfs/module/os/linux/zfs/vdev_label_os.c
new file mode 100644
index 000000000000..3d965b89a962
--- /dev/null
+++ b/sys/contrib/openzfs/module/os/linux/zfs/vdev_label_os.c
@@ -0,0 +1,45 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License (the "License").
+ * You may not use this file except in compliance with the License.
+ *
+ * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+ * or https://opensource.org/licenses/CDDL-1.0.
+ * See the License for the specific language governing permissions
+ * and limitations under the License.
+ *
+ * When distributing Covered Code, include this CDDL HEADER in each
+ * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+ * If applicable, add the following below this CDDL HEADER, with the
+ * fields enclosed by brackets "[]" replaced with your own identifying
+ * information: Portions Copyright [yyyy] [name of copyright owner]
+ *
+ * CDDL HEADER END
+ */
+
+/*
+ * Copyright (c) 2023 by iXsystems, Inc.
+ */
+
+#include <sys/zfs_context.h>
+#include <sys/spa.h>
+#include <sys/spa_impl.h>
+#include <sys/vdev.h>
+#include <sys/vdev_impl.h>
+
+/*
+ * Check if the reserved boot area is in-use.
+ *
+ * This function always returns 0, as there are no known external uses
+ * of the reserved area on Linux.
+ */
+int
+vdev_check_boot_reserve(spa_t *spa, vdev_t *childvd)
+{
+ (void) spa;
+ (void) childvd;
+
+ return (0);
+}
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_debug.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_debug.c
index b090ec684e05..f707959c9445 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_debug.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_debug.c
@@ -1,259 +1,260 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2014 by Delphix. All rights reserved.
*/
#include <sys/zfs_context.h>
#include <sys/trace_zfs.h>
typedef struct zfs_dbgmsg {
procfs_list_node_t zdm_node;
uint64_t zdm_timestamp;
uint_t zdm_size;
char zdm_msg[]; /* variable length allocation */
} zfs_dbgmsg_t;
static procfs_list_t zfs_dbgmsgs;
static uint_t zfs_dbgmsg_size = 0;
static uint_t zfs_dbgmsg_maxsize = 4<<20; /* 4MB */
/*
* Internal ZFS debug messages are enabled by default.
*
* # Print debug messages
* cat /proc/spl/kstat/zfs/dbgmsg
*
* # Disable the kernel debug message log.
* echo 0 > /sys/module/zfs/parameters/zfs_dbgmsg_enable
*
* # Clear the kernel debug message log.
* echo 0 >/proc/spl/kstat/zfs/dbgmsg
*/
int zfs_dbgmsg_enable = B_TRUE;
static int
zfs_dbgmsg_show_header(struct seq_file *f)
{
seq_printf(f, "%-12s %-8s\n", "timestamp", "message");
return (0);
}
static int
zfs_dbgmsg_show(struct seq_file *f, void *p)
{
zfs_dbgmsg_t *zdm = (zfs_dbgmsg_t *)p;
seq_printf(f, "%-12llu %-s\n",
(u_longlong_t)zdm->zdm_timestamp, zdm->zdm_msg);
return (0);
}
static void
zfs_dbgmsg_purge(uint_t max_size)
{
while (zfs_dbgmsg_size > max_size) {
zfs_dbgmsg_t *zdm = list_remove_head(&zfs_dbgmsgs.pl_list);
if (zdm == NULL)
return;
uint_t size = zdm->zdm_size;
kmem_free(zdm, size);
zfs_dbgmsg_size -= size;
}
}
static int
zfs_dbgmsg_clear(procfs_list_t *procfs_list)
{
(void) procfs_list;
mutex_enter(&zfs_dbgmsgs.pl_lock);
zfs_dbgmsg_purge(0);
mutex_exit(&zfs_dbgmsgs.pl_lock);
return (0);
}
void
zfs_dbgmsg_init(void)
{
procfs_list_install("zfs",
NULL,
"dbgmsg",
0600,
&zfs_dbgmsgs,
zfs_dbgmsg_show,
zfs_dbgmsg_show_header,
zfs_dbgmsg_clear,
offsetof(zfs_dbgmsg_t, zdm_node));
}
void
zfs_dbgmsg_fini(void)
{
procfs_list_uninstall(&zfs_dbgmsgs);
zfs_dbgmsg_purge(0);
/*
* TODO - decide how to make this permanent
*/
#ifdef _KERNEL
procfs_list_destroy(&zfs_dbgmsgs);
#endif
}
void
__set_error(const char *file, const char *func, int line, int err)
{
/*
* To enable this:
*
* $ echo 512 >/sys/module/zfs/parameters/zfs_flags
*/
if (zfs_flags & ZFS_DEBUG_SET_ERROR)
__dprintf(B_FALSE, file, func, line, "error %lu",
(ulong_t)err);
}
void
__zfs_dbgmsg(char *buf)
{
uint_t size = sizeof (zfs_dbgmsg_t) + strlen(buf) + 1;
zfs_dbgmsg_t *zdm = kmem_zalloc(size, KM_SLEEP);
zdm->zdm_size = size;
zdm->zdm_timestamp = gethrestime_sec();
strcpy(zdm->zdm_msg, buf);
mutex_enter(&zfs_dbgmsgs.pl_lock);
procfs_list_add(&zfs_dbgmsgs, zdm);
zfs_dbgmsg_size += size;
zfs_dbgmsg_purge(zfs_dbgmsg_maxsize);
mutex_exit(&zfs_dbgmsgs.pl_lock);
}
#ifdef _KERNEL
void
__dprintf(boolean_t dprint, const char *file, const char *func,
int line, const char *fmt, ...)
{
const char *newfile;
va_list adx;
size_t size;
char *buf;
char *nl;
int i;
char *prefix = (dprint) ? "dprintf: " : "";
size = 1024;
buf = kmem_alloc(size, KM_SLEEP);
/*
* Get rid of annoying prefix to filename.
*/
newfile = strrchr(file, '/');
if (newfile != NULL) {
newfile = newfile + 1; /* Get rid of leading / */
} else {
newfile = file;
}
- i = snprintf(buf, size, "%s%s:%d:%s(): ", prefix, newfile, line, func);
+ i = snprintf(buf, size, "%px %s%s:%d:%s(): ",
+ curthread, prefix, newfile, line, func);
if (i < size) {
va_start(adx, fmt);
(void) vsnprintf(buf + i, size - i, fmt, adx);
va_end(adx);
}
/*
* Get rid of trailing newline for dprintf logs.
*/
if (dprint && buf[0] != '\0') {
nl = &buf[strlen(buf) - 1];
if (*nl == '\n')
*nl = '\0';
}
/*
* To get this data enable the zfs__dprintf trace point as shown:
*
* # Enable zfs__dprintf tracepoint, clear the tracepoint ring buffer
* $ echo 1 > /sys/kernel/debug/tracing/events/zfs/enable
* $ echo 0 > /sys/kernel/debug/tracing/trace
*
* # Dump the ring buffer.
* $ cat /sys/kernel/debug/tracing/trace
*/
DTRACE_PROBE1(zfs__dprintf, char *, buf);
/*
* To get this data:
*
* $ cat /proc/spl/kstat/zfs/dbgmsg
*
* To clear the buffer:
* $ echo 0 > /proc/spl/kstat/zfs/dbgmsg
*/
__zfs_dbgmsg(buf);
kmem_free(buf, size);
}
#else
void
zfs_dbgmsg_print(const char *tag)
{
ssize_t ret __attribute__((unused));
/*
* We use write() in this function instead of printf()
* so it is safe to call from a signal handler.
*/
ret = write(STDOUT_FILENO, "ZFS_DBGMSG(", 11);
ret = write(STDOUT_FILENO, tag, strlen(tag));
ret = write(STDOUT_FILENO, ") START:\n", 9);
mutex_enter(&zfs_dbgmsgs.pl_lock);
for (zfs_dbgmsg_t *zdm = list_head(&zfs_dbgmsgs.pl_list); zdm != NULL;
zdm = list_next(&zfs_dbgmsgs.pl_list, zdm)) {
ret = write(STDOUT_FILENO, zdm->zdm_msg,
strlen(zdm->zdm_msg));
ret = write(STDOUT_FILENO, "\n", 1);
}
ret = write(STDOUT_FILENO, "ZFS_DBGMSG(", 11);
ret = write(STDOUT_FILENO, tag, strlen(tag));
ret = write(STDOUT_FILENO, ") END\n", 6);
mutex_exit(&zfs_dbgmsgs.pl_lock);
}
#endif /* _KERNEL */
#ifdef _KERNEL
module_param(zfs_dbgmsg_enable, int, 0644);
MODULE_PARM_DESC(zfs_dbgmsg_enable, "Enable ZFS debug message log");
/* BEGIN CSTYLED */
module_param(zfs_dbgmsg_maxsize, uint, 0644);
/* END CSTYLED */
MODULE_PARM_DESC(zfs_dbgmsg_maxsize, "Maximum ZFS debug log size");
#endif
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c
index 6ad75ace0b5c..1ae8023adc32 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c
@@ -1,4252 +1,4252 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2018 by Delphix. All rights reserved.
* Copyright (c) 2015 by Chunwei Chen. All rights reserved.
* Copyright 2017 Nexenta Systems, Inc.
*/
/* Portions Copyright 2007 Jeremy Teo */
/* Portions Copyright 2010 Robert Milkowski */
#include <sys/types.h>
#include <sys/param.h>
#include <sys/time.h>
#include <sys/sysmacros.h>
#include <sys/vfs.h>
#include <sys/file.h>
#include <sys/stat.h>
#include <sys/kmem.h>
#include <sys/taskq.h>
#include <sys/uio.h>
#include <sys/vmsystm.h>
#include <sys/atomic.h>
#include <sys/pathname.h>
#include <sys/cmn_err.h>
#include <sys/errno.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_acl.h>
#include <sys/zfs_ioctl.h>
#include <sys/fs/zfs.h>
#include <sys/dmu.h>
#include <sys/dmu_objset.h>
#include <sys/spa.h>
#include <sys/txg.h>
#include <sys/dbuf.h>
#include <sys/zap.h>
#include <sys/sa.h>
#include <sys/policy.h>
#include <sys/sunddi.h>
#include <sys/sid.h>
#include <sys/zfs_ctldir.h>
#include <sys/zfs_fuid.h>
#include <sys/zfs_quota.h>
#include <sys/zfs_sa.h>
#include <sys/zfs_vnops.h>
#include <sys/zfs_rlock.h>
#include <sys/cred.h>
#include <sys/zpl.h>
#include <sys/zil.h>
#include <sys/sa_impl.h>
/*
* Programming rules.
*
* Each vnode op performs some logical unit of work. To do this, the ZPL must
* properly lock its in-core state, create a DMU transaction, do the work,
* record this work in the intent log (ZIL), commit the DMU transaction,
* and wait for the intent log to commit if it is a synchronous operation.
* Moreover, the vnode ops must work in both normal and log replay context.
* The ordering of events is important to avoid deadlocks and references
* to freed memory. The example below illustrates the following Big Rules:
*
* (1) A check must be made in each zfs thread for a mounted file system.
* This is done avoiding races using zfs_enter(zfsvfs).
* A zfs_exit(zfsvfs) is needed before all returns. Any znodes
* must be checked with zfs_verify_zp(zp). Both of these macros
* can return EIO from the calling function.
*
* (2) zrele() should always be the last thing except for zil_commit() (if
* necessary) and zfs_exit(). This is for 3 reasons: First, if it's the
* last reference, the vnode/znode can be freed, so the zp may point to
* freed memory. Second, the last reference will call zfs_zinactive(),
* which may induce a lot of work -- pushing cached pages (which acquires
* range locks) and syncing out cached atime changes. Third,
* zfs_zinactive() may require a new tx, which could deadlock the system
* if you were already holding one. This deadlock occurs because the tx
* currently being operated on prevents a txg from syncing, which
* prevents the new tx from progressing, resulting in a deadlock. If you
* must call zrele() within a tx, use zfs_zrele_async(). Note that iput()
* is a synonym for zrele().
*
* (3) All range locks must be grabbed before calling dmu_tx_assign(),
* as they can span dmu_tx_assign() calls.
*
* (4) If ZPL locks are held, pass TXG_NOWAIT as the second argument to
* dmu_tx_assign(). This is critical because we don't want to block
* while holding locks.
*
* If no ZPL locks are held (aside from zfs_enter()), use TXG_WAIT. This
* reduces lock contention and CPU usage when we must wait (note that if
* throughput is constrained by the storage, nearly every transaction
* must wait).
*
* Note, in particular, that if a lock is sometimes acquired before
* the tx assigns, and sometimes after (e.g. z_lock), then failing
* to use a non-blocking assign can deadlock the system. The scenario:
*
* Thread A has grabbed a lock before calling dmu_tx_assign().
* Thread B is in an already-assigned tx, and blocks for this lock.
* Thread A calls dmu_tx_assign(TXG_WAIT) and blocks in txg_wait_open()
* forever, because the previous txg can't quiesce until B's tx commits.
*
* If dmu_tx_assign() returns ERESTART and zfsvfs->z_assign is TXG_NOWAIT,
* then drop all locks, call dmu_tx_wait(), and try again. On subsequent
* calls to dmu_tx_assign(), pass TXG_NOTHROTTLE in addition to TXG_NOWAIT,
* to indicate that this operation has already called dmu_tx_wait().
* This will ensure that we don't retry forever, waiting a short bit
* each time.
*
* (5) If the operation succeeded, generate the intent log entry for it
* before dropping locks. This ensures that the ordering of events
* in the intent log matches the order in which they actually occurred.
* During ZIL replay the zfs_log_* functions will update the sequence
* number to indicate the zil transaction has replayed.
*
* (6) At the end of each vnode op, the DMU tx must always commit,
* regardless of whether there were any errors.
*
* (7) After dropping all locks, invoke zil_commit(zilog, foid)
* to ensure that synchronous semantics are provided when necessary.
*
* In general, this is how things should be ordered in each vnode op:
*
* zfs_enter(zfsvfs); // exit if unmounted
* top:
* zfs_dirent_lock(&dl, ...) // lock directory entry (may igrab())
* rw_enter(...); // grab any other locks you need
* tx = dmu_tx_create(...); // get DMU tx
* dmu_tx_hold_*(); // hold each object you might modify
* error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
* if (error) {
* rw_exit(...); // drop locks
* zfs_dirent_unlock(dl); // unlock directory entry
* zrele(...); // release held znodes
* if (error == ERESTART) {
* waited = B_TRUE;
* dmu_tx_wait(tx);
* dmu_tx_abort(tx);
* goto top;
* }
* dmu_tx_abort(tx); // abort DMU tx
* zfs_exit(zfsvfs); // finished in zfs
* return (error); // really out of space
* }
* error = do_real_work(); // do whatever this VOP does
* if (error == 0)
* zfs_log_*(...); // on success, make ZIL entry
* dmu_tx_commit(tx); // commit DMU tx -- error or not
* rw_exit(...); // drop locks
* zfs_dirent_unlock(dl); // unlock directory entry
* zrele(...); // release held znodes
* zil_commit(zilog, foid); // synchronous when necessary
* zfs_exit(zfsvfs); // finished in zfs
* return (error); // done, report error
*/
int
zfs_open(struct inode *ip, int mode, int flag, cred_t *cr)
{
(void) cr;
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
/* Honor ZFS_APPENDONLY file attribute */
if (blk_mode_is_open_write(mode) && (zp->z_pflags & ZFS_APPENDONLY) &&
((flag & O_APPEND) == 0)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
/*
* Keep a count of the synchronous opens in the znode. On first
* synchronous open we must convert all previous async transactions
* into sync to keep correct ordering.
*/
if (flag & O_SYNC) {
if (atomic_inc_32_nv(&zp->z_sync_cnt) == 1)
zil_async_to_sync(zfsvfs->z_log, zp->z_id);
}
zfs_exit(zfsvfs, FTAG);
return (0);
}
int
zfs_close(struct inode *ip, int flag, cred_t *cr)
{
(void) cr;
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
/* Decrement the synchronous opens in the znode */
if (flag & O_SYNC)
atomic_dec_32(&zp->z_sync_cnt);
zfs_exit(zfsvfs, FTAG);
return (0);
}
#if defined(_KERNEL)
static int zfs_fillpage(struct inode *ip, struct page *pp);
/*
* When a file is memory mapped, we must keep the IO data synchronized
* between the DMU cache and the memory mapped pages. Update all mapped
* pages with the contents of the coresponding dmu buffer.
*/
void
update_pages(znode_t *zp, int64_t start, int len, objset_t *os)
{
struct address_space *mp = ZTOI(zp)->i_mapping;
int64_t off = start & (PAGE_SIZE - 1);
for (start &= PAGE_MASK; len > 0; start += PAGE_SIZE) {
uint64_t nbytes = MIN(PAGE_SIZE - off, len);
struct page *pp = find_lock_page(mp, start >> PAGE_SHIFT);
if (pp) {
if (mapping_writably_mapped(mp))
flush_dcache_page(pp);
void *pb = kmap(pp);
int error = dmu_read(os, zp->z_id, start + off,
nbytes, pb + off, DMU_READ_PREFETCH);
kunmap(pp);
if (error) {
SetPageError(pp);
ClearPageUptodate(pp);
} else {
ClearPageError(pp);
SetPageUptodate(pp);
if (mapping_writably_mapped(mp))
flush_dcache_page(pp);
mark_page_accessed(pp);
}
unlock_page(pp);
put_page(pp);
}
len -= nbytes;
off = 0;
}
}
/*
* When a file is memory mapped, we must keep the I/O data synchronized
* between the DMU cache and the memory mapped pages. Preferentially read
* from memory mapped pages, otherwise fallback to reading through the dmu.
*/
int
mappedread(znode_t *zp, int nbytes, zfs_uio_t *uio)
{
struct inode *ip = ZTOI(zp);
struct address_space *mp = ip->i_mapping;
int64_t start = uio->uio_loffset;
int64_t off = start & (PAGE_SIZE - 1);
int len = nbytes;
int error = 0;
for (start &= PAGE_MASK; len > 0; start += PAGE_SIZE) {
uint64_t bytes = MIN(PAGE_SIZE - off, len);
struct page *pp = find_lock_page(mp, start >> PAGE_SHIFT);
if (pp) {
/*
* If filemap_fault() retries there exists a window
* where the page will be unlocked and not up to date.
* In this case we must try and fill the page.
*/
if (unlikely(!PageUptodate(pp))) {
error = zfs_fillpage(ip, pp);
if (error) {
unlock_page(pp);
put_page(pp);
return (error);
}
}
ASSERT(PageUptodate(pp) || PageDirty(pp));
unlock_page(pp);
void *pb = kmap(pp);
error = zfs_uiomove(pb + off, bytes, UIO_READ, uio);
kunmap(pp);
if (mapping_writably_mapped(mp))
flush_dcache_page(pp);
mark_page_accessed(pp);
put_page(pp);
} else {
error = dmu_read_uio_dbuf(sa_get_db(zp->z_sa_hdl),
uio, bytes);
}
len -= bytes;
off = 0;
if (error)
break;
}
return (error);
}
#endif /* _KERNEL */
static unsigned long zfs_delete_blocks = DMU_MAX_DELETEBLKCNT;
/*
* Write the bytes to a file.
*
* IN: zp - znode of file to be written to
* data - bytes to write
* len - number of bytes to write
* pos - offset to start writing at
*
* OUT: resid - remaining bytes to write
*
* RETURN: 0 if success
* positive error code if failure. EIO is returned
* for a short write when residp isn't provided.
*
* Timestamps:
* zp - ctime|mtime updated if byte count > 0
*/
int
zfs_write_simple(znode_t *zp, const void *data, size_t len,
loff_t pos, size_t *residp)
{
fstrans_cookie_t cookie;
int error;
struct iovec iov;
iov.iov_base = (void *)data;
iov.iov_len = len;
zfs_uio_t uio;
zfs_uio_iovec_init(&uio, &iov, 1, pos, UIO_SYSSPACE, len, 0);
cookie = spl_fstrans_mark();
error = zfs_write(zp, &uio, 0, kcred);
spl_fstrans_unmark(cookie);
if (error == 0) {
if (residp != NULL)
*residp = zfs_uio_resid(&uio);
else if (zfs_uio_resid(&uio) != 0)
error = SET_ERROR(EIO);
}
return (error);
}
static void
zfs_rele_async_task(void *arg)
{
iput(arg);
}
void
zfs_zrele_async(znode_t *zp)
{
struct inode *ip = ZTOI(zp);
objset_t *os = ITOZSB(ip)->z_os;
ASSERT(atomic_read(&ip->i_count) > 0);
ASSERT(os != NULL);
/*
* If decrementing the count would put us at 0, we can't do it inline
* here, because that would be synchronous. Instead, dispatch an iput
* to run later.
*
* For more information on the dangers of a synchronous iput, see the
* header comment of this file.
*/
if (!atomic_add_unless(&ip->i_count, -1, 1)) {
VERIFY(taskq_dispatch(dsl_pool_zrele_taskq(dmu_objset_pool(os)),
zfs_rele_async_task, ip, TQ_SLEEP) != TASKQID_INVALID);
}
}
/*
* Lookup an entry in a directory, or an extended attribute directory.
* If it exists, return a held inode reference for it.
*
* IN: zdp - znode of directory to search.
* nm - name of entry to lookup.
* flags - LOOKUP_XATTR set if looking for an attribute.
* cr - credentials of caller.
* direntflags - directory lookup flags
* realpnp - returned pathname.
*
* OUT: zpp - znode of located entry, NULL if not found.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* NA
*/
int
zfs_lookup(znode_t *zdp, char *nm, znode_t **zpp, int flags, cred_t *cr,
int *direntflags, pathname_t *realpnp)
{
zfsvfs_t *zfsvfs = ZTOZSB(zdp);
int error = 0;
/*
* Fast path lookup, however we must skip DNLC lookup
* for case folding or normalizing lookups because the
* DNLC code only stores the passed in name. This means
* creating 'a' and removing 'A' on a case insensitive
* file system would work, but DNLC still thinks 'a'
* exists and won't let you create it again on the next
* pass through fast path.
*/
if (!(flags & (LOOKUP_XATTR | FIGNORECASE))) {
if (!S_ISDIR(ZTOI(zdp)->i_mode)) {
return (SET_ERROR(ENOTDIR));
} else if (zdp->z_sa_hdl == NULL) {
return (SET_ERROR(EIO));
}
if (nm[0] == 0 || (nm[0] == '.' && nm[1] == '\0')) {
error = zfs_fastaccesschk_execute(zdp, cr);
if (!error) {
*zpp = zdp;
zhold(*zpp);
return (0);
}
return (error);
}
}
if ((error = zfs_enter_verify_zp(zfsvfs, zdp, FTAG)) != 0)
return (error);
*zpp = NULL;
if (flags & LOOKUP_XATTR) {
/*
* We don't allow recursive attributes..
* Maybe someday we will.
*/
if (zdp->z_pflags & ZFS_XATTR) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if ((error = zfs_get_xattrdir(zdp, zpp, cr, flags))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Do we have permission to get into attribute directory?
*/
if ((error = zfs_zaccess(*zpp, ACE_EXECUTE, 0,
B_TRUE, cr, zfs_init_idmap))) {
zrele(*zpp);
*zpp = NULL;
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (!S_ISDIR(ZTOI(zdp)->i_mode)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENOTDIR));
}
/*
* Check accessibility of directory.
*/
if ((error = zfs_zaccess(zdp, ACE_EXECUTE, 0, B_FALSE, cr,
zfs_init_idmap))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (zfsvfs->z_utf8 && u8_validate(nm, strlen(nm),
NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
error = zfs_dirlook(zdp, nm, zpp, flags, direntflags, realpnp);
if ((error == 0) && (*zpp))
zfs_znode_update_vfs(*zpp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Attempt to create a new entry in a directory. If the entry
* already exists, truncate the file if permissible, else return
* an error. Return the ip of the created or trunc'd file.
*
* IN: dzp - znode of directory to put new file entry in.
* name - name of new file entry.
* vap - attributes of new file.
* excl - flag indicating exclusive or non-exclusive mode.
* mode - mode to open file with.
* cr - credentials of caller.
* flag - file flag.
* vsecp - ACL to be set
* mnt_ns - user namespace of the mount
*
* OUT: zpp - znode of created or trunc'd entry.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dzp - ctime|mtime updated if new entry created
* zp - ctime|mtime always, atime if new
*/
int
zfs_create(znode_t *dzp, char *name, vattr_t *vap, int excl,
int mode, znode_t **zpp, cred_t *cr, int flag, vsecattr_t *vsecp,
zidmap_t *mnt_ns)
{
znode_t *zp;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zilog_t *zilog;
objset_t *os;
zfs_dirlock_t *dl;
dmu_tx_t *tx;
int error;
uid_t uid;
gid_t gid;
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
boolean_t have_acl = B_FALSE;
boolean_t waited = B_FALSE;
boolean_t skip_acl = (flag & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE;
/*
* If we have an ephemeral id, ACL, or XVATTR then
* make sure file system is at proper version
*/
gid = crgetgid(cr);
uid = crgetuid(cr);
if (zfsvfs->z_use_fuids == B_FALSE &&
(vsecp || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid)))
return (SET_ERROR(EINVAL));
if (name == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
os = zfsvfs->z_os;
zilog = zfsvfs->z_log;
if (zfsvfs->z_utf8 && u8_validate(name, strlen(name),
NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (vap->va_mask & ATTR_XVATTR) {
if ((error = secpolicy_xvattr((xvattr_t *)vap,
crgetuid(cr), cr, vap->va_mode)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
top:
*zpp = NULL;
if (*name == '\0') {
/*
* Null component name refers to the directory itself.
*/
zhold(dzp);
zp = dzp;
dl = NULL;
error = 0;
} else {
/* possible igrab(zp) */
int zflg = 0;
if (flag & FIGNORECASE)
zflg |= ZCILOOK;
error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg,
NULL, NULL);
if (error) {
if (have_acl)
zfs_acl_ids_free(&acl_ids);
if (strcmp(name, "..") == 0)
error = SET_ERROR(EISDIR);
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
if (zp == NULL) {
uint64_t txtype;
uint64_t projid = ZFS_DEFAULT_PROJID;
/*
* Create a new file object and update the directory
* to reference it.
*/
if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, skip_acl, cr,
mnt_ns))) {
if (have_acl)
zfs_acl_ids_free(&acl_ids);
goto out;
}
/*
* We only support the creation of regular files in
* extended attribute directories.
*/
if ((dzp->z_pflags & ZFS_XATTR) && !S_ISREG(vap->va_mode)) {
if (have_acl)
zfs_acl_ids_free(&acl_ids);
error = SET_ERROR(EINVAL);
goto out;
}
if (!have_acl && (error = zfs_acl_ids_create(dzp, 0, vap,
cr, vsecp, &acl_ids, mnt_ns)) != 0)
goto out;
have_acl = B_TRUE;
if (S_ISREG(vap->va_mode) || S_ISDIR(vap->va_mode))
projid = zfs_inherit_projid(dzp);
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, projid)) {
zfs_acl_ids_free(&acl_ids);
error = SET_ERROR(EDQUOT);
goto out;
}
tx = dmu_tx_create(os);
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
dmu_tx_hold_zap(tx, dzp->z_id, TRUE, name);
dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE);
if (!zfsvfs->z_use_sa &&
acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, acl_ids.z_aclp->z_acl_bytes);
}
error = dmu_tx_assign(tx,
(waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids);
error = zfs_link_create(dl, zp, tx, ZNEW);
if (error != 0) {
/*
* Since, we failed to add the directory entry for it,
* delete the newly created dnode.
*/
zfs_znode_delete(zp, tx);
remove_inode_hash(ZTOI(zp));
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
goto out;
}
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
txtype = zfs_log_create_txtype(Z_FILE, vsecp, vap);
if (flag & FIGNORECASE)
txtype |= TX_CI;
zfs_log_create(zilog, tx, txtype, dzp, zp, name,
vsecp, acl_ids.z_fuidp, vap);
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
} else {
int aflags = (flag & O_APPEND) ? V_APPEND : 0;
if (have_acl)
zfs_acl_ids_free(&acl_ids);
/*
* A directory entry already exists for this name.
*/
/*
* Can't truncate an existing file if in exclusive mode.
*/
if (excl) {
error = SET_ERROR(EEXIST);
goto out;
}
/*
* Can't open a directory for writing.
*/
if (S_ISDIR(ZTOI(zp)->i_mode)) {
error = SET_ERROR(EISDIR);
goto out;
}
/*
* Verify requested access to file.
*/
if (mode && (error = zfs_zaccess_rwx(zp, mode, aflags, cr,
mnt_ns))) {
goto out;
}
mutex_enter(&dzp->z_lock);
dzp->z_seq++;
mutex_exit(&dzp->z_lock);
/*
* Truncate regular files if requested.
*/
if (S_ISREG(ZTOI(zp)->i_mode) &&
(vap->va_mask & ATTR_SIZE) && (vap->va_size == 0)) {
/* we can't hold any locks when calling zfs_freesp() */
if (dl) {
zfs_dirent_unlock(dl);
dl = NULL;
}
error = zfs_freesp(zp, 0, 0, mode, TRUE);
}
}
out:
if (dl)
zfs_dirent_unlock(dl);
if (error) {
if (zp)
zrele(zp);
} else {
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
*zpp = zp;
}
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
int
zfs_tmpfile(struct inode *dip, vattr_t *vap, int excl,
int mode, struct inode **ipp, cred_t *cr, int flag, vsecattr_t *vsecp,
zidmap_t *mnt_ns)
{
(void) excl, (void) mode, (void) flag;
znode_t *zp = NULL, *dzp = ITOZ(dip);
zfsvfs_t *zfsvfs = ITOZSB(dip);
objset_t *os;
dmu_tx_t *tx;
int error;
uid_t uid;
gid_t gid;
zfs_acl_ids_t acl_ids;
uint64_t projid = ZFS_DEFAULT_PROJID;
boolean_t fuid_dirtied;
boolean_t have_acl = B_FALSE;
boolean_t waited = B_FALSE;
/*
* If we have an ephemeral id, ACL, or XVATTR then
* make sure file system is at proper version
*/
gid = crgetgid(cr);
uid = crgetuid(cr);
if (zfsvfs->z_use_fuids == B_FALSE &&
(vsecp || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid)))
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
os = zfsvfs->z_os;
if (vap->va_mask & ATTR_XVATTR) {
if ((error = secpolicy_xvattr((xvattr_t *)vap,
crgetuid(cr), cr, vap->va_mode)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
top:
*ipp = NULL;
/*
* Create a new file object and update the directory
* to reference it.
*/
if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns))) {
if (have_acl)
zfs_acl_ids_free(&acl_ids);
goto out;
}
if (!have_acl && (error = zfs_acl_ids_create(dzp, 0, vap,
cr, vsecp, &acl_ids, mnt_ns)) != 0)
goto out;
have_acl = B_TRUE;
if (S_ISREG(vap->va_mode) || S_ISDIR(vap->va_mode))
projid = zfs_inherit_projid(dzp);
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, projid)) {
zfs_acl_ids_free(&acl_ids);
error = SET_ERROR(EDQUOT);
goto out;
}
tx = dmu_tx_create(os);
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
if (!zfsvfs->z_use_sa &&
acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, acl_ids.z_aclp->z_acl_bytes);
}
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
zfs_mknode(dzp, vap, tx, cr, IS_TMPFILE, &zp, &acl_ids);
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
/* Add to unlinked set */
zp->z_unlinked = B_TRUE;
zfs_unlinked_add(zp, tx);
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
out:
if (error) {
if (zp)
zrele(zp);
} else {
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
*ipp = ZTOI(zp);
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Remove an entry from a directory.
*
* IN: dzp - znode of directory to remove entry from.
* name - name of entry to remove.
* cr - credentials of caller.
* flags - case flags.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* dzp - ctime|mtime
* ip - ctime (if nlink > 0)
*/
static uint64_t null_xattr = 0;
int
zfs_remove(znode_t *dzp, char *name, cred_t *cr, int flags)
{
znode_t *zp;
znode_t *xzp;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zilog_t *zilog;
uint64_t acl_obj, xattr_obj;
uint64_t xattr_obj_unlinked = 0;
uint64_t obj = 0;
uint64_t links;
zfs_dirlock_t *dl;
dmu_tx_t *tx;
boolean_t may_delete_now, delete_now = FALSE;
boolean_t unlinked, toobig = FALSE;
uint64_t txtype;
pathname_t *realnmp = NULL;
pathname_t realnm;
int error;
int zflg = ZEXISTS;
boolean_t waited = B_FALSE;
if (name == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if (flags & FIGNORECASE) {
zflg |= ZCILOOK;
pn_alloc(&realnm);
realnmp = &realnm;
}
top:
xattr_obj = 0;
xzp = NULL;
/*
* Attempt to lock directory; fail if entry doesn't exist.
*/
if ((error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg,
NULL, realnmp))) {
if (realnmp)
pn_free(realnmp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if ((error = zfs_zaccess_delete(dzp, zp, cr, zfs_init_idmap))) {
goto out;
}
/*
* Need to use rmdir for removing directories.
*/
if (S_ISDIR(ZTOI(zp)->i_mode)) {
error = SET_ERROR(EPERM);
goto out;
}
mutex_enter(&zp->z_lock);
may_delete_now = atomic_read(&ZTOI(zp)->i_count) == 1 &&
!zn_has_cached_data(zp, 0, LLONG_MAX);
mutex_exit(&zp->z_lock);
/*
* We may delete the znode now, or we may put it in the unlinked set;
* it depends on whether we're the last link, and on whether there are
* other holds on the inode. So we dmu_tx_hold() the right things to
* allow for either case.
*/
obj = zp->z_id;
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_zap(tx, dzp->z_id, FALSE, name);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
zfs_sa_upgrade_txholds(tx, dzp);
if (may_delete_now) {
toobig = zp->z_size > zp->z_blksz * zfs_delete_blocks;
/* if the file is too big, only hold_free a token amount */
dmu_tx_hold_free(tx, zp->z_id, 0,
(toobig ? DMU_MAX_ACCESS : DMU_OBJECT_END));
}
/* are there any extended attributes? */
error = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs),
&xattr_obj, sizeof (xattr_obj));
if (error == 0 && xattr_obj) {
error = zfs_zget(zfsvfs, xattr_obj, &xzp);
ASSERT0(error);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
dmu_tx_hold_sa(tx, xzp->z_sa_hdl, B_FALSE);
}
mutex_enter(&zp->z_lock);
if ((acl_obj = zfs_external_acl(zp)) != 0 && may_delete_now)
dmu_tx_hold_free(tx, acl_obj, 0, DMU_OBJECT_END);
mutex_exit(&zp->z_lock);
/* charge as an update -- would be nice not to charge at all */
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
/*
* Mark this transaction as typically resulting in a net free of space
*/
dmu_tx_mark_netfree(tx);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
zrele(zp);
if (xzp)
zrele(xzp);
goto top;
}
if (realnmp)
pn_free(realnmp);
dmu_tx_abort(tx);
zrele(zp);
if (xzp)
zrele(xzp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Remove the directory entry.
*/
error = zfs_link_destroy(dl, zp, tx, zflg, &unlinked);
if (error) {
dmu_tx_commit(tx);
goto out;
}
if (unlinked) {
/*
* Hold z_lock so that we can make sure that the ACL obj
* hasn't changed. Could have been deleted due to
* zfs_sa_upgrade().
*/
mutex_enter(&zp->z_lock);
(void) sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs),
&xattr_obj_unlinked, sizeof (xattr_obj_unlinked));
delete_now = may_delete_now && !toobig &&
atomic_read(&ZTOI(zp)->i_count) == 1 &&
!zn_has_cached_data(zp, 0, LLONG_MAX) &&
xattr_obj == xattr_obj_unlinked &&
zfs_external_acl(zp) == acl_obj;
VERIFY_IMPLY(xattr_obj_unlinked, xzp);
}
if (delete_now) {
if (xattr_obj_unlinked) {
ASSERT3U(ZTOI(xzp)->i_nlink, ==, 2);
mutex_enter(&xzp->z_lock);
xzp->z_unlinked = B_TRUE;
clear_nlink(ZTOI(xzp));
links = 0;
error = sa_update(xzp->z_sa_hdl, SA_ZPL_LINKS(zfsvfs),
&links, sizeof (links), tx);
ASSERT3U(error, ==, 0);
mutex_exit(&xzp->z_lock);
zfs_unlinked_add(xzp, tx);
if (zp->z_is_sa)
error = sa_remove(zp->z_sa_hdl,
SA_ZPL_XATTR(zfsvfs), tx);
else
error = sa_update(zp->z_sa_hdl,
SA_ZPL_XATTR(zfsvfs), &null_xattr,
sizeof (uint64_t), tx);
ASSERT0(error);
}
/*
* Add to the unlinked set because a new reference could be
* taken concurrently resulting in a deferred destruction.
*/
zfs_unlinked_add(zp, tx);
mutex_exit(&zp->z_lock);
} else if (unlinked) {
mutex_exit(&zp->z_lock);
zfs_unlinked_add(zp, tx);
}
txtype = TX_REMOVE;
if (flags & FIGNORECASE)
txtype |= TX_CI;
zfs_log_remove(zilog, tx, txtype, dzp, name, obj, unlinked);
dmu_tx_commit(tx);
out:
if (realnmp)
pn_free(realnmp);
zfs_dirent_unlock(dl);
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
if (delete_now)
zrele(zp);
else
zfs_zrele_async(zp);
if (xzp) {
zfs_znode_update_vfs(xzp);
zfs_zrele_async(xzp);
}
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Create a new directory and insert it into dzp using the name
* provided. Return a pointer to the inserted directory.
*
* IN: dzp - znode of directory to add subdir to.
* dirname - name of new directory.
* vap - attributes of new directory.
* cr - credentials of caller.
* flags - case flags.
* vsecp - ACL to be set
* mnt_ns - user namespace of the mount
*
* OUT: zpp - znode of created directory.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* dzp - ctime|mtime updated
* zpp - ctime|mtime|atime updated
*/
int
zfs_mkdir(znode_t *dzp, char *dirname, vattr_t *vap, znode_t **zpp,
cred_t *cr, int flags, vsecattr_t *vsecp, zidmap_t *mnt_ns)
{
znode_t *zp;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zilog_t *zilog;
zfs_dirlock_t *dl;
uint64_t txtype;
dmu_tx_t *tx;
int error;
int zf = ZNEW;
uid_t uid;
gid_t gid = crgetgid(cr);
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
boolean_t waited = B_FALSE;
ASSERT(S_ISDIR(vap->va_mode));
/*
* If we have an ephemeral id, ACL, or XVATTR then
* make sure file system is at proper version
*/
uid = crgetuid(cr);
if (zfsvfs->z_use_fuids == B_FALSE &&
(vsecp || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid)))
return (SET_ERROR(EINVAL));
if (dirname == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if (dzp->z_pflags & ZFS_XATTR) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if (zfsvfs->z_utf8 && u8_validate(dirname,
strlen(dirname), NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (flags & FIGNORECASE)
zf |= ZCILOOK;
if (vap->va_mask & ATTR_XVATTR) {
if ((error = secpolicy_xvattr((xvattr_t *)vap,
crgetuid(cr), cr, vap->va_mode)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
if ((error = zfs_acl_ids_create(dzp, 0, vap, cr,
vsecp, &acl_ids, mnt_ns)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* First make sure the new directory doesn't exist.
*
* Existence is checked first to make sure we don't return
* EACCES instead of EEXIST which can cause some applications
* to fail.
*/
top:
*zpp = NULL;
if ((error = zfs_dirent_lock(&dl, dzp, dirname, &zp, zf,
NULL, NULL))) {
zfs_acl_ids_free(&acl_ids);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if ((error = zfs_zaccess(dzp, ACE_ADD_SUBDIRECTORY, 0, B_FALSE, cr,
mnt_ns))) {
zfs_acl_ids_free(&acl_ids);
zfs_dirent_unlock(dl);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, zfs_inherit_projid(dzp))) {
zfs_acl_ids_free(&acl_ids);
zfs_dirent_unlock(dl);
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EDQUOT));
}
/*
* Add a new entry to the directory.
*/
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_zap(tx, dzp->z_id, TRUE, dirname);
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, FALSE, NULL);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0,
acl_ids.z_aclp->z_acl_bytes);
}
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Create new node.
*/
zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids);
/*
* Now put new name in parent dir.
*/
error = zfs_link_create(dl, zp, tx, ZNEW);
if (error != 0) {
zfs_znode_delete(zp, tx);
remove_inode_hash(ZTOI(zp));
goto out;
}
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
*zpp = zp;
txtype = zfs_log_create_txtype(Z_DIR, vsecp, vap);
if (flags & FIGNORECASE)
txtype |= TX_CI;
zfs_log_create(zilog, tx, txtype, dzp, zp, dirname, vsecp,
acl_ids.z_fuidp, vap);
out:
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
zfs_dirent_unlock(dl);
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
if (error != 0) {
zrele(zp);
} else {
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Remove a directory subdir entry. If the current working
* directory is the same as the subdir to be removed, the
* remove will fail.
*
* IN: dzp - znode of directory to remove from.
* name - name of directory to be removed.
* cwd - inode of current working directory.
* cr - credentials of caller.
* flags - case flags
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dzp - ctime|mtime updated
*/
int
zfs_rmdir(znode_t *dzp, char *name, znode_t *cwd, cred_t *cr,
int flags)
{
znode_t *zp;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zilog_t *zilog;
zfs_dirlock_t *dl;
dmu_tx_t *tx;
int error;
int zflg = ZEXISTS;
boolean_t waited = B_FALSE;
if (name == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if (flags & FIGNORECASE)
zflg |= ZCILOOK;
top:
zp = NULL;
/*
* Attempt to lock directory; fail if entry doesn't exist.
*/
if ((error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg,
NULL, NULL))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
if ((error = zfs_zaccess_delete(dzp, zp, cr, zfs_init_idmap))) {
goto out;
}
if (!S_ISDIR(ZTOI(zp)->i_mode)) {
error = SET_ERROR(ENOTDIR);
goto out;
}
if (zp == cwd) {
error = SET_ERROR(EINVAL);
goto out;
}
/*
* Grab a lock on the directory to make sure that no one is
* trying to add (or lookup) entries while we are removing it.
*/
rw_enter(&zp->z_name_lock, RW_WRITER);
/*
* Grab a lock on the parent pointer to make sure we play well
* with the treewalk and directory rename code.
*/
rw_enter(&zp->z_parent_lock, RW_WRITER);
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_zap(tx, dzp->z_id, FALSE, name);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
zfs_sa_upgrade_txholds(tx, zp);
zfs_sa_upgrade_txholds(tx, dzp);
dmu_tx_mark_netfree(tx);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
rw_exit(&zp->z_parent_lock);
rw_exit(&zp->z_name_lock);
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
zrele(zp);
goto top;
}
dmu_tx_abort(tx);
zrele(zp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
error = zfs_link_destroy(dl, zp, tx, zflg, NULL);
if (error == 0) {
uint64_t txtype = TX_RMDIR;
if (flags & FIGNORECASE)
txtype |= TX_CI;
zfs_log_remove(zilog, tx, txtype, dzp, name, ZFS_NO_OBJECT,
B_FALSE);
}
dmu_tx_commit(tx);
rw_exit(&zp->z_parent_lock);
rw_exit(&zp->z_name_lock);
out:
zfs_dirent_unlock(dl);
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
zrele(zp);
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Read directory entries from the given directory cursor position and emit
* name and position for each entry.
*
* IN: ip - inode of directory to read.
* ctx - directory entry context.
* cr - credentials of caller.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* ip - atime updated
*
* Note that the low 4 bits of the cookie returned by zap is always zero.
* This allows us to use the low range for "special" directory entries:
* We use 0 for '.', and 1 for '..'. If this is the root of the filesystem,
* we use the offset 2 for the '.zfs' directory.
*/
int
zfs_readdir(struct inode *ip, zpl_dir_context_t *ctx, cred_t *cr)
{
(void) cr;
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
objset_t *os;
zap_cursor_t zc;
zap_attribute_t zap;
int error;
uint8_t prefetch;
uint8_t type;
int done = 0;
uint64_t parent;
uint64_t offset; /* must be unsigned; checks for < 1 */
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs),
&parent, sizeof (parent))) != 0)
goto out;
/*
* Quit if directory has been removed (posix)
*/
if (zp->z_unlinked)
goto out;
error = 0;
os = zfsvfs->z_os;
offset = ctx->pos;
prefetch = zp->z_zn_prefetch;
/*
* Initialize the iterator cursor.
*/
if (offset <= 3) {
/*
* Start iteration from the beginning of the directory.
*/
zap_cursor_init(&zc, os, zp->z_id);
} else {
/*
* The offset is a serialized cursor.
*/
zap_cursor_init_serialized(&zc, os, zp->z_id, offset);
}
/*
* Transform to file-system independent format
*/
while (!done) {
uint64_t objnum;
/*
* Special case `.', `..', and `.zfs'.
*/
if (offset == 0) {
(void) strcpy(zap.za_name, ".");
zap.za_normalization_conflict = 0;
objnum = zp->z_id;
type = DT_DIR;
} else if (offset == 1) {
(void) strcpy(zap.za_name, "..");
zap.za_normalization_conflict = 0;
objnum = parent;
type = DT_DIR;
} else if (offset == 2 && zfs_show_ctldir(zp)) {
(void) strcpy(zap.za_name, ZFS_CTLDIR_NAME);
zap.za_normalization_conflict = 0;
objnum = ZFSCTL_INO_ROOT;
type = DT_DIR;
} else {
/*
* Grab next entry.
*/
if ((error = zap_cursor_retrieve(&zc, &zap))) {
if (error == ENOENT)
break;
else
goto update;
}
/*
* Allow multiple entries provided the first entry is
* the object id. Non-zpl consumers may safely make
* use of the additional space.
*
* XXX: This should be a feature flag for compatibility
*/
if (zap.za_integer_length != 8 ||
zap.za_num_integers == 0) {
cmn_err(CE_WARN, "zap_readdir: bad directory "
"entry, obj = %lld, offset = %lld, "
"length = %d, num = %lld\n",
(u_longlong_t)zp->z_id,
(u_longlong_t)offset,
zap.za_integer_length,
(u_longlong_t)zap.za_num_integers);
error = SET_ERROR(ENXIO);
goto update;
}
objnum = ZFS_DIRENT_OBJ(zap.za_first_integer);
type = ZFS_DIRENT_TYPE(zap.za_first_integer);
}
done = !zpl_dir_emit(ctx, zap.za_name, strlen(zap.za_name),
objnum, type);
if (done)
break;
if (prefetch)
dmu_prefetch_dnode(os, objnum, ZIO_PRIORITY_SYNC_READ);
/*
* Move to the next entry, fill in the previous offset.
*/
if (offset > 2 || (offset == 2 && !zfs_show_ctldir(zp))) {
zap_cursor_advance(&zc);
offset = zap_cursor_serialize(&zc);
} else {
offset += 1;
}
ctx->pos = offset;
}
zp->z_zn_prefetch = B_FALSE; /* a lookup will re-enable pre-fetching */
update:
zap_cursor_fini(&zc);
if (error == ENOENT)
error = 0;
out:
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Get the basic file attributes and place them in the provided kstat
* structure. The inode is assumed to be the authoritative source
* for most of the attributes. However, the znode currently has the
* authoritative atime, blksize, and block count.
*
* IN: ip - inode of file.
*
* OUT: sp - kstat values.
*
* RETURN: 0 (always succeeds)
*/
int
#ifdef HAVE_GENERIC_FILLATTR_IDMAP_REQMASK
zfs_getattr_fast(zidmap_t *user_ns, u32 request_mask, struct inode *ip,
struct kstat *sp)
#else
zfs_getattr_fast(zidmap_t *user_ns, struct inode *ip, struct kstat *sp)
#endif
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
uint32_t blksize;
u_longlong_t nblocks;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
mutex_enter(&zp->z_lock);
#ifdef HAVE_GENERIC_FILLATTR_IDMAP_REQMASK
zpl_generic_fillattr(user_ns, request_mask, ip, sp);
#else
zpl_generic_fillattr(user_ns, ip, sp);
#endif
/*
* +1 link count for root inode with visible '.zfs' directory.
*/
if ((zp->z_id == zfsvfs->z_root) && zfs_show_ctldir(zp))
if (sp->nlink < ZFS_LINK_MAX)
sp->nlink++;
sa_object_size(zp->z_sa_hdl, &blksize, &nblocks);
sp->blksize = blksize;
sp->blocks = nblocks;
if (unlikely(zp->z_blksz == 0)) {
/*
* Block size hasn't been set; suggest maximal I/O transfers.
*/
sp->blksize = zfsvfs->z_max_blksz;
}
mutex_exit(&zp->z_lock);
/*
* Required to prevent NFS client from detecting different inode
* numbers of snapshot root dentry before and after snapshot mount.
*/
if (zfsvfs->z_issnap) {
if (ip->i_sb->s_root->d_inode == ip)
sp->ino = ZFSCTL_INO_SNAPDIRS -
dmu_objset_id(zfsvfs->z_os);
}
zfs_exit(zfsvfs, FTAG);
return (0);
}
/*
* For the operation of changing file's user/group/project, we need to
* handle not only the main object that is assigned to the file directly,
* but also the ones that are used by the file via hidden xattr directory.
*
* Because the xattr directory may contains many EA entries, as to it may
* be impossible to change all of them via the transaction of changing the
* main object's user/group/project attributes. Then we have to change them
* via other multiple independent transactions one by one. It may be not good
* solution, but we have no better idea yet.
*/
static int
zfs_setattr_dir(znode_t *dzp)
{
struct inode *dxip = ZTOI(dzp);
struct inode *xip = NULL;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
objset_t *os = zfsvfs->z_os;
zap_cursor_t zc;
zap_attribute_t zap;
zfs_dirlock_t *dl;
znode_t *zp = NULL;
dmu_tx_t *tx = NULL;
uint64_t uid, gid;
sa_bulk_attr_t bulk[4];
int count;
int err;
zap_cursor_init(&zc, os, dzp->z_id);
while ((err = zap_cursor_retrieve(&zc, &zap)) == 0) {
count = 0;
if (zap.za_integer_length != 8 || zap.za_num_integers != 1) {
err = ENXIO;
break;
}
err = zfs_dirent_lock(&dl, dzp, (char *)zap.za_name, &zp,
ZEXISTS, NULL, NULL);
if (err == ENOENT)
goto next;
if (err)
break;
xip = ZTOI(zp);
if (KUID_TO_SUID(xip->i_uid) == KUID_TO_SUID(dxip->i_uid) &&
KGID_TO_SGID(xip->i_gid) == KGID_TO_SGID(dxip->i_gid) &&
zp->z_projid == dzp->z_projid)
goto next;
tx = dmu_tx_create(os);
if (!(zp->z_pflags & ZFS_PROJID))
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
else
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
err = dmu_tx_assign(tx, TXG_WAIT);
if (err)
break;
mutex_enter(&dzp->z_lock);
if (KUID_TO_SUID(xip->i_uid) != KUID_TO_SUID(dxip->i_uid)) {
xip->i_uid = dxip->i_uid;
uid = zfs_uid_read(dxip);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL,
&uid, sizeof (uid));
}
if (KGID_TO_SGID(xip->i_gid) != KGID_TO_SGID(dxip->i_gid)) {
xip->i_gid = dxip->i_gid;
gid = zfs_gid_read(dxip);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL,
&gid, sizeof (gid));
}
if (zp->z_projid != dzp->z_projid) {
if (!(zp->z_pflags & ZFS_PROJID)) {
zp->z_pflags |= ZFS_PROJID;
SA_ADD_BULK_ATTR(bulk, count,
SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags,
sizeof (zp->z_pflags));
}
zp->z_projid = dzp->z_projid;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_PROJID(zfsvfs),
NULL, &zp->z_projid, sizeof (zp->z_projid));
}
mutex_exit(&dzp->z_lock);
if (likely(count > 0)) {
err = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
dmu_tx_commit(tx);
} else {
dmu_tx_abort(tx);
}
tx = NULL;
if (err != 0 && err != ENOENT)
break;
next:
if (zp) {
zrele(zp);
zp = NULL;
zfs_dirent_unlock(dl);
}
zap_cursor_advance(&zc);
}
if (tx)
dmu_tx_abort(tx);
if (zp) {
zrele(zp);
zfs_dirent_unlock(dl);
}
zap_cursor_fini(&zc);
return (err == ENOENT ? 0 : err);
}
/*
* Set the file attributes to the values contained in the
* vattr structure.
*
* IN: zp - znode of file to be modified.
* vap - new attribute values.
* If ATTR_XVATTR set, then optional attrs are being set
* flags - ATTR_UTIME set if non-default time values provided.
* - ATTR_NOACLCHECK (CIFS context only).
* cr - credentials of caller.
* mnt_ns - user namespace of the mount
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* ip - ctime updated, mtime updated if size changed.
*/
int
zfs_setattr(znode_t *zp, vattr_t *vap, int flags, cred_t *cr, zidmap_t *mnt_ns)
{
struct inode *ip;
zfsvfs_t *zfsvfs = ZTOZSB(zp);
objset_t *os = zfsvfs->z_os;
zilog_t *zilog;
dmu_tx_t *tx;
vattr_t oldva;
xvattr_t *tmpxvattr;
uint_t mask = vap->va_mask;
uint_t saved_mask = 0;
int trim_mask = 0;
uint64_t new_mode;
uint64_t new_kuid = 0, new_kgid = 0, new_uid, new_gid;
uint64_t xattr_obj;
uint64_t mtime[2], ctime[2], atime[2];
uint64_t projid = ZFS_INVALID_PROJID;
znode_t *attrzp;
int need_policy = FALSE;
int err, err2 = 0;
zfs_fuid_info_t *fuidp = NULL;
xvattr_t *xvap = (xvattr_t *)vap; /* vap may be an xvattr_t * */
xoptattr_t *xoap;
zfs_acl_t *aclp;
boolean_t skipaclchk = (flags & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE;
boolean_t fuid_dirtied = B_FALSE;
boolean_t handle_eadir = B_FALSE;
sa_bulk_attr_t *bulk, *xattr_bulk;
int count = 0, xattr_count = 0, bulks = 8;
if (mask == 0)
return (0);
if ((err = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (err);
ip = ZTOI(zp);
/*
* If this is a xvattr_t, then get a pointer to the structure of
* optional attributes. If this is NULL, then we have a vattr_t.
*/
xoap = xva_getxoptattr(xvap);
if (xoap != NULL && (mask & ATTR_XVATTR)) {
if (XVA_ISSET_REQ(xvap, XAT_PROJID)) {
if (!dmu_objset_projectquota_enabled(os) ||
(!S_ISREG(ip->i_mode) && !S_ISDIR(ip->i_mode))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENOTSUP));
}
projid = xoap->xoa_projid;
if (unlikely(projid == ZFS_INVALID_PROJID)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if (projid == zp->z_projid && zp->z_pflags & ZFS_PROJID)
projid = ZFS_INVALID_PROJID;
else
need_policy = TRUE;
}
if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT) &&
(xoap->xoa_projinherit !=
((zp->z_pflags & ZFS_PROJINHERIT) != 0)) &&
(!dmu_objset_projectquota_enabled(os) ||
(!S_ISREG(ip->i_mode) && !S_ISDIR(ip->i_mode)))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENOTSUP));
}
}
zilog = zfsvfs->z_log;
/*
* Make sure that if we have ephemeral uid/gid or xvattr specified
* that file system is at proper version level
*/
if (zfsvfs->z_use_fuids == B_FALSE &&
(((mask & ATTR_UID) && IS_EPHEMERAL(vap->va_uid)) ||
((mask & ATTR_GID) && IS_EPHEMERAL(vap->va_gid)) ||
(mask & ATTR_XVATTR))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if (mask & ATTR_SIZE && S_ISDIR(ip->i_mode)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EISDIR));
}
if (mask & ATTR_SIZE && !S_ISREG(ip->i_mode) && !S_ISFIFO(ip->i_mode)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
tmpxvattr = kmem_alloc(sizeof (xvattr_t), KM_SLEEP);
xva_init(tmpxvattr);
bulk = kmem_alloc(sizeof (sa_bulk_attr_t) * bulks, KM_SLEEP);
xattr_bulk = kmem_alloc(sizeof (sa_bulk_attr_t) * bulks, KM_SLEEP);
/*
* Immutable files can only alter immutable bit and atime
*/
if ((zp->z_pflags & ZFS_IMMUTABLE) &&
((mask & (ATTR_SIZE|ATTR_UID|ATTR_GID|ATTR_MTIME|ATTR_MODE)) ||
((mask & ATTR_XVATTR) && XVA_ISSET_REQ(xvap, XAT_CREATETIME)))) {
err = SET_ERROR(EPERM);
goto out3;
}
if ((mask & ATTR_SIZE) && (zp->z_pflags & ZFS_READONLY)) {
err = SET_ERROR(EPERM);
goto out3;
}
/*
* Verify timestamps doesn't overflow 32 bits.
* ZFS can handle large timestamps, but 32bit syscalls can't
* handle times greater than 2039. This check should be removed
* once large timestamps are fully supported.
*/
if (mask & (ATTR_ATIME | ATTR_MTIME)) {
if (((mask & ATTR_ATIME) &&
TIMESPEC_OVERFLOW(&vap->va_atime)) ||
((mask & ATTR_MTIME) &&
TIMESPEC_OVERFLOW(&vap->va_mtime))) {
err = SET_ERROR(EOVERFLOW);
goto out3;
}
}
top:
attrzp = NULL;
aclp = NULL;
/* Can this be moved to before the top label? */
if (zfs_is_readonly(zfsvfs)) {
err = SET_ERROR(EROFS);
goto out3;
}
/*
* First validate permissions
*/
if (mask & ATTR_SIZE) {
err = zfs_zaccess(zp, ACE_WRITE_DATA, 0, skipaclchk, cr,
mnt_ns);
if (err)
goto out3;
/*
* XXX - Note, we are not providing any open
* mode flags here (like FNDELAY), so we may
* block if there are locks present... this
* should be addressed in openat().
*/
/* XXX - would it be OK to generate a log record here? */
err = zfs_freesp(zp, vap->va_size, 0, 0, FALSE);
if (err)
goto out3;
}
if (mask & (ATTR_ATIME|ATTR_MTIME) ||
((mask & ATTR_XVATTR) && (XVA_ISSET_REQ(xvap, XAT_HIDDEN) ||
XVA_ISSET_REQ(xvap, XAT_READONLY) ||
XVA_ISSET_REQ(xvap, XAT_ARCHIVE) ||
XVA_ISSET_REQ(xvap, XAT_OFFLINE) ||
XVA_ISSET_REQ(xvap, XAT_SPARSE) ||
XVA_ISSET_REQ(xvap, XAT_CREATETIME) ||
XVA_ISSET_REQ(xvap, XAT_SYSTEM)))) {
need_policy = zfs_zaccess(zp, ACE_WRITE_ATTRIBUTES, 0,
skipaclchk, cr, mnt_ns);
}
if (mask & (ATTR_UID|ATTR_GID)) {
int idmask = (mask & (ATTR_UID|ATTR_GID));
int take_owner;
int take_group;
uid_t uid;
gid_t gid;
/*
* NOTE: even if a new mode is being set,
* we may clear S_ISUID/S_ISGID bits.
*/
if (!(mask & ATTR_MODE))
vap->va_mode = zp->z_mode;
/*
* Take ownership or chgrp to group we are a member of
*/
uid = zfs_uid_to_vfsuid(mnt_ns, zfs_i_user_ns(ip),
vap->va_uid);
gid = zfs_gid_to_vfsgid(mnt_ns, zfs_i_user_ns(ip),
vap->va_gid);
take_owner = (mask & ATTR_UID) && (uid == crgetuid(cr));
take_group = (mask & ATTR_GID) &&
zfs_groupmember(zfsvfs, gid, cr);
/*
* If both ATTR_UID and ATTR_GID are set then take_owner and
* take_group must both be set in order to allow taking
* ownership.
*
* Otherwise, send the check through secpolicy_vnode_setattr()
*
*/
if (((idmask == (ATTR_UID|ATTR_GID)) &&
take_owner && take_group) ||
((idmask == ATTR_UID) && take_owner) ||
((idmask == ATTR_GID) && take_group)) {
if (zfs_zaccess(zp, ACE_WRITE_OWNER, 0,
skipaclchk, cr, mnt_ns) == 0) {
/*
* Remove setuid/setgid for non-privileged users
*/
(void) secpolicy_setid_clear(vap, cr);
trim_mask = (mask & (ATTR_UID|ATTR_GID));
} else {
need_policy = TRUE;
}
} else {
need_policy = TRUE;
}
}
mutex_enter(&zp->z_lock);
oldva.va_mode = zp->z_mode;
zfs_fuid_map_ids(zp, cr, &oldva.va_uid, &oldva.va_gid);
if (mask & ATTR_XVATTR) {
/*
* Update xvattr mask to include only those attributes
* that are actually changing.
*
* the bits will be restored prior to actually setting
* the attributes so the caller thinks they were set.
*/
if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY)) {
if (xoap->xoa_appendonly !=
((zp->z_pflags & ZFS_APPENDONLY) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_APPENDONLY);
XVA_SET_REQ(tmpxvattr, XAT_APPENDONLY);
}
}
if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT)) {
if (xoap->xoa_projinherit !=
((zp->z_pflags & ZFS_PROJINHERIT) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_PROJINHERIT);
XVA_SET_REQ(tmpxvattr, XAT_PROJINHERIT);
}
}
if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK)) {
if (xoap->xoa_nounlink !=
((zp->z_pflags & ZFS_NOUNLINK) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_NOUNLINK);
XVA_SET_REQ(tmpxvattr, XAT_NOUNLINK);
}
}
if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE)) {
if (xoap->xoa_immutable !=
((zp->z_pflags & ZFS_IMMUTABLE) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_IMMUTABLE);
XVA_SET_REQ(tmpxvattr, XAT_IMMUTABLE);
}
}
if (XVA_ISSET_REQ(xvap, XAT_NODUMP)) {
if (xoap->xoa_nodump !=
((zp->z_pflags & ZFS_NODUMP) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_NODUMP);
XVA_SET_REQ(tmpxvattr, XAT_NODUMP);
}
}
if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED)) {
if (xoap->xoa_av_modified !=
((zp->z_pflags & ZFS_AV_MODIFIED) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_AV_MODIFIED);
XVA_SET_REQ(tmpxvattr, XAT_AV_MODIFIED);
}
}
if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED)) {
if ((!S_ISREG(ip->i_mode) &&
xoap->xoa_av_quarantined) ||
xoap->xoa_av_quarantined !=
((zp->z_pflags & ZFS_AV_QUARANTINED) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_AV_QUARANTINED);
XVA_SET_REQ(tmpxvattr, XAT_AV_QUARANTINED);
}
}
if (XVA_ISSET_REQ(xvap, XAT_REPARSE)) {
mutex_exit(&zp->z_lock);
err = SET_ERROR(EPERM);
goto out3;
}
if (need_policy == FALSE &&
(XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP) ||
XVA_ISSET_REQ(xvap, XAT_OPAQUE))) {
need_policy = TRUE;
}
}
mutex_exit(&zp->z_lock);
if (mask & ATTR_MODE) {
if (zfs_zaccess(zp, ACE_WRITE_ACL, 0, skipaclchk, cr,
mnt_ns) == 0) {
err = secpolicy_setid_setsticky_clear(ip, vap,
&oldva, cr, mnt_ns, zfs_i_user_ns(ip));
if (err)
goto out3;
trim_mask |= ATTR_MODE;
} else {
need_policy = TRUE;
}
}
if (need_policy) {
/*
* If trim_mask is set then take ownership
* has been granted or write_acl is present and user
* has the ability to modify mode. In that case remove
* UID|GID and or MODE from mask so that
* secpolicy_vnode_setattr() doesn't revoke it.
*/
if (trim_mask) {
saved_mask = vap->va_mask;
vap->va_mask &= ~trim_mask;
}
err = secpolicy_vnode_setattr(cr, ip, vap, &oldva, flags,
zfs_zaccess_unix, zp);
if (err)
goto out3;
if (trim_mask)
vap->va_mask |= saved_mask;
}
/*
* secpolicy_vnode_setattr, or take ownership may have
* changed va_mask
*/
mask = vap->va_mask;
if ((mask & (ATTR_UID | ATTR_GID)) || projid != ZFS_INVALID_PROJID) {
handle_eadir = B_TRUE;
err = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs),
&xattr_obj, sizeof (xattr_obj));
if (err == 0 && xattr_obj) {
err = zfs_zget(ZTOZSB(zp), xattr_obj, &attrzp);
if (err)
goto out2;
}
if (mask & ATTR_UID) {
new_kuid = zfs_fuid_create(zfsvfs,
(uint64_t)vap->va_uid, cr, ZFS_OWNER, &fuidp);
if (new_kuid != KUID_TO_SUID(ZTOI(zp)->i_uid) &&
zfs_id_overquota(zfsvfs, DMU_USERUSED_OBJECT,
new_kuid)) {
if (attrzp)
zrele(attrzp);
err = SET_ERROR(EDQUOT);
goto out2;
}
}
if (mask & ATTR_GID) {
new_kgid = zfs_fuid_create(zfsvfs,
(uint64_t)vap->va_gid, cr, ZFS_GROUP, &fuidp);
if (new_kgid != KGID_TO_SGID(ZTOI(zp)->i_gid) &&
zfs_id_overquota(zfsvfs, DMU_GROUPUSED_OBJECT,
new_kgid)) {
if (attrzp)
zrele(attrzp);
err = SET_ERROR(EDQUOT);
goto out2;
}
}
if (projid != ZFS_INVALID_PROJID &&
zfs_id_overquota(zfsvfs, DMU_PROJECTUSED_OBJECT, projid)) {
if (attrzp)
zrele(attrzp);
err = EDQUOT;
goto out2;
}
}
tx = dmu_tx_create(os);
if (mask & ATTR_MODE) {
uint64_t pmode = zp->z_mode;
uint64_t acl_obj;
new_mode = (pmode & S_IFMT) | (vap->va_mode & ~S_IFMT);
if (ZTOZSB(zp)->z_acl_mode == ZFS_ACL_RESTRICTED &&
!(zp->z_pflags & ZFS_ACL_TRIVIAL)) {
err = EPERM;
goto out;
}
if ((err = zfs_acl_chmod_setattr(zp, &aclp, new_mode)))
goto out;
mutex_enter(&zp->z_lock);
if (!zp->z_is_sa && ((acl_obj = zfs_external_acl(zp)) != 0)) {
/*
* Are we upgrading ACL from old V0 format
* to V1 format?
*/
if (zfsvfs->z_version >= ZPL_VERSION_FUID &&
zfs_znode_acl_version(zp) ==
ZFS_ACL_VERSION_INITIAL) {
dmu_tx_hold_free(tx, acl_obj, 0,
DMU_OBJECT_END);
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, aclp->z_acl_bytes);
} else {
dmu_tx_hold_write(tx, acl_obj, 0,
aclp->z_acl_bytes);
}
} else if (!zp->z_is_sa && aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, aclp->z_acl_bytes);
}
mutex_exit(&zp->z_lock);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
} else {
if (((mask & ATTR_XVATTR) &&
XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) ||
(projid != ZFS_INVALID_PROJID &&
!(zp->z_pflags & ZFS_PROJID)))
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
else
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
}
if (attrzp) {
dmu_tx_hold_sa(tx, attrzp->z_sa_hdl, B_FALSE);
}
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
zfs_sa_upgrade_txholds(tx, zp);
err = dmu_tx_assign(tx, TXG_WAIT);
if (err)
goto out;
count = 0;
/*
* Set each attribute requested.
* We group settings according to the locks they need to acquire.
*
* Note: you cannot set ctime directly, although it will be
* updated as a side-effect of calling this function.
*/
if (projid != ZFS_INVALID_PROJID && !(zp->z_pflags & ZFS_PROJID)) {
/*
* For the existed object that is upgraded from old system,
* its on-disk layout has no slot for the project ID attribute.
* But quota accounting logic needs to access related slots by
* offset directly. So we need to adjust old objects' layout
* to make the project ID to some unified and fixed offset.
*/
if (attrzp)
err = sa_add_projid(attrzp->z_sa_hdl, tx, projid);
if (err == 0)
err = sa_add_projid(zp->z_sa_hdl, tx, projid);
if (unlikely(err == EEXIST))
err = 0;
else if (err != 0)
goto out;
else
projid = ZFS_INVALID_PROJID;
}
if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE))
mutex_enter(&zp->z_acl_lock);
mutex_enter(&zp->z_lock);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, sizeof (zp->z_pflags));
if (attrzp) {
if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE))
mutex_enter(&attrzp->z_acl_lock);
mutex_enter(&attrzp->z_lock);
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_FLAGS(zfsvfs), NULL, &attrzp->z_pflags,
sizeof (attrzp->z_pflags));
if (projid != ZFS_INVALID_PROJID) {
attrzp->z_projid = projid;
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_PROJID(zfsvfs), NULL, &attrzp->z_projid,
sizeof (attrzp->z_projid));
}
}
if (mask & (ATTR_UID|ATTR_GID)) {
if (mask & ATTR_UID) {
ZTOI(zp)->i_uid = SUID_TO_KUID(new_kuid);
new_uid = zfs_uid_read(ZTOI(zp));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL,
&new_uid, sizeof (new_uid));
if (attrzp) {
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_UID(zfsvfs), NULL, &new_uid,
sizeof (new_uid));
ZTOI(attrzp)->i_uid = SUID_TO_KUID(new_uid);
}
}
if (mask & ATTR_GID) {
ZTOI(zp)->i_gid = SGID_TO_KGID(new_kgid);
new_gid = zfs_gid_read(ZTOI(zp));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs),
NULL, &new_gid, sizeof (new_gid));
if (attrzp) {
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_GID(zfsvfs), NULL, &new_gid,
sizeof (new_gid));
ZTOI(attrzp)->i_gid = SGID_TO_KGID(new_kgid);
}
}
if (!(mask & ATTR_MODE)) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs),
NULL, &new_mode, sizeof (new_mode));
new_mode = zp->z_mode;
}
err = zfs_acl_chown_setattr(zp);
ASSERT(err == 0);
if (attrzp) {
err = zfs_acl_chown_setattr(attrzp);
ASSERT(err == 0);
}
}
if (mask & ATTR_MODE) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL,
&new_mode, sizeof (new_mode));
zp->z_mode = ZTOI(zp)->i_mode = new_mode;
ASSERT3P(aclp, !=, NULL);
err = zfs_aclset_common(zp, aclp, cr, tx);
ASSERT0(err);
if (zp->z_acl_cached)
zfs_acl_free(zp->z_acl_cached);
zp->z_acl_cached = aclp;
aclp = NULL;
}
if ((mask & ATTR_ATIME) || zp->z_atime_dirty) {
zp->z_atime_dirty = B_FALSE;
ZFS_TIME_ENCODE(&ip->i_atime, atime);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL,
&atime, sizeof (atime));
}
if (mask & (ATTR_MTIME | ATTR_SIZE)) {
ZFS_TIME_ENCODE(&vap->va_mtime, mtime);
ZTOI(zp)->i_mtime = zpl_inode_timestamp_truncate(
vap->va_mtime, ZTOI(zp));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL,
mtime, sizeof (mtime));
}
if (mask & (ATTR_CTIME | ATTR_SIZE)) {
ZFS_TIME_ENCODE(&vap->va_ctime, ctime);
zpl_inode_set_ctime_to_ts(ZTOI(zp),
zpl_inode_timestamp_truncate(vap->va_ctime, ZTOI(zp)));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
ctime, sizeof (ctime));
}
if (projid != ZFS_INVALID_PROJID) {
zp->z_projid = projid;
SA_ADD_BULK_ATTR(bulk, count,
SA_ZPL_PROJID(zfsvfs), NULL, &zp->z_projid,
sizeof (zp->z_projid));
}
if (attrzp && mask) {
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_CTIME(zfsvfs), NULL, &ctime,
sizeof (ctime));
}
/*
* Do this after setting timestamps to prevent timestamp
* update from toggling bit
*/
if (xoap && (mask & ATTR_XVATTR)) {
/*
* restore trimmed off masks
* so that return masks can be set for caller.
*/
if (XVA_ISSET_REQ(tmpxvattr, XAT_APPENDONLY)) {
XVA_SET_REQ(xvap, XAT_APPENDONLY);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_NOUNLINK)) {
XVA_SET_REQ(xvap, XAT_NOUNLINK);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_IMMUTABLE)) {
XVA_SET_REQ(xvap, XAT_IMMUTABLE);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_NODUMP)) {
XVA_SET_REQ(xvap, XAT_NODUMP);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_AV_MODIFIED)) {
XVA_SET_REQ(xvap, XAT_AV_MODIFIED);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_AV_QUARANTINED)) {
XVA_SET_REQ(xvap, XAT_AV_QUARANTINED);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_PROJINHERIT)) {
XVA_SET_REQ(xvap, XAT_PROJINHERIT);
}
if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP))
ASSERT(S_ISREG(ip->i_mode));
zfs_xvattr_set(zp, xvap, tx);
}
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
if (mask != 0)
zfs_log_setattr(zilog, tx, TX_SETATTR, zp, vap, mask, fuidp);
mutex_exit(&zp->z_lock);
if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE))
mutex_exit(&zp->z_acl_lock);
if (attrzp) {
if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE))
mutex_exit(&attrzp->z_acl_lock);
mutex_exit(&attrzp->z_lock);
}
out:
if (err == 0 && xattr_count > 0) {
err2 = sa_bulk_update(attrzp->z_sa_hdl, xattr_bulk,
xattr_count, tx);
ASSERT(err2 == 0);
}
if (aclp)
zfs_acl_free(aclp);
if (fuidp) {
zfs_fuid_info_free(fuidp);
fuidp = NULL;
}
if (err) {
dmu_tx_abort(tx);
if (attrzp)
zrele(attrzp);
if (err == ERESTART)
goto top;
} else {
if (count > 0)
err2 = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
dmu_tx_commit(tx);
if (attrzp) {
if (err2 == 0 && handle_eadir)
err = zfs_setattr_dir(attrzp);
zrele(attrzp);
}
zfs_znode_update_vfs(zp);
}
out2:
if (os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
out3:
kmem_free(xattr_bulk, sizeof (sa_bulk_attr_t) * bulks);
kmem_free(bulk, sizeof (sa_bulk_attr_t) * bulks);
kmem_free(tmpxvattr, sizeof (xvattr_t));
zfs_exit(zfsvfs, FTAG);
return (err);
}
typedef struct zfs_zlock {
krwlock_t *zl_rwlock; /* lock we acquired */
znode_t *zl_znode; /* znode we held */
struct zfs_zlock *zl_next; /* next in list */
} zfs_zlock_t;
/*
* Drop locks and release vnodes that were held by zfs_rename_lock().
*/
static void
zfs_rename_unlock(zfs_zlock_t **zlpp)
{
zfs_zlock_t *zl;
while ((zl = *zlpp) != NULL) {
if (zl->zl_znode != NULL)
zfs_zrele_async(zl->zl_znode);
rw_exit(zl->zl_rwlock);
*zlpp = zl->zl_next;
kmem_free(zl, sizeof (*zl));
}
}
/*
* Search back through the directory tree, using the ".." entries.
* Lock each directory in the chain to prevent concurrent renames.
* Fail any attempt to move a directory into one of its own descendants.
* XXX - z_parent_lock can overlap with map or grow locks
*/
static int
zfs_rename_lock(znode_t *szp, znode_t *tdzp, znode_t *sdzp, zfs_zlock_t **zlpp)
{
zfs_zlock_t *zl;
znode_t *zp = tdzp;
uint64_t rootid = ZTOZSB(zp)->z_root;
uint64_t oidp = zp->z_id;
krwlock_t *rwlp = &szp->z_parent_lock;
krw_t rw = RW_WRITER;
/*
* First pass write-locks szp and compares to zp->z_id.
* Later passes read-lock zp and compare to zp->z_parent.
*/
do {
if (!rw_tryenter(rwlp, rw)) {
/*
* Another thread is renaming in this path.
* Note that if we are a WRITER, we don't have any
* parent_locks held yet.
*/
if (rw == RW_READER && zp->z_id > szp->z_id) {
/*
* Drop our locks and restart
*/
zfs_rename_unlock(&zl);
*zlpp = NULL;
zp = tdzp;
oidp = zp->z_id;
rwlp = &szp->z_parent_lock;
rw = RW_WRITER;
continue;
} else {
/*
* Wait for other thread to drop its locks
*/
rw_enter(rwlp, rw);
}
}
zl = kmem_alloc(sizeof (*zl), KM_SLEEP);
zl->zl_rwlock = rwlp;
zl->zl_znode = NULL;
zl->zl_next = *zlpp;
*zlpp = zl;
if (oidp == szp->z_id) /* We're a descendant of szp */
return (SET_ERROR(EINVAL));
if (oidp == rootid) /* We've hit the top */
return (0);
if (rw == RW_READER) { /* i.e. not the first pass */
int error = zfs_zget(ZTOZSB(zp), oidp, &zp);
if (error)
return (error);
zl->zl_znode = zp;
}
(void) sa_lookup(zp->z_sa_hdl, SA_ZPL_PARENT(ZTOZSB(zp)),
&oidp, sizeof (oidp));
rwlp = &zp->z_parent_lock;
rw = RW_READER;
} while (zp->z_id != sdzp->z_id);
return (0);
}
/*
* Move an entry from the provided source directory to the target
* directory. Change the entry name as indicated.
*
* IN: sdzp - Source directory containing the "old entry".
* snm - Old entry name.
* tdzp - Target directory to contain the "new entry".
* tnm - New entry name.
* cr - credentials of caller.
* flags - case flags
* rflags - RENAME_* flags
* wa_vap - attributes for RENAME_WHITEOUT (must be a char 0:0).
* mnt_ns - user namespace of the mount
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* sdzp,tdzp - ctime|mtime updated
*/
int
zfs_rename(znode_t *sdzp, char *snm, znode_t *tdzp, char *tnm,
cred_t *cr, int flags, uint64_t rflags, vattr_t *wo_vap, zidmap_t *mnt_ns)
{
znode_t *szp, *tzp;
zfsvfs_t *zfsvfs = ZTOZSB(sdzp);
zilog_t *zilog;
zfs_dirlock_t *sdl, *tdl;
dmu_tx_t *tx;
zfs_zlock_t *zl;
int cmp, serr, terr;
int error = 0;
int zflg = 0;
boolean_t waited = B_FALSE;
/* Needed for whiteout inode creation. */
boolean_t fuid_dirtied;
zfs_acl_ids_t acl_ids;
boolean_t have_acl = B_FALSE;
znode_t *wzp = NULL;
if (snm == NULL || tnm == NULL)
return (SET_ERROR(EINVAL));
if (rflags & ~(RENAME_NOREPLACE | RENAME_EXCHANGE | RENAME_WHITEOUT))
return (SET_ERROR(EINVAL));
/* Already checked by Linux VFS, but just to make sure. */
if (rflags & RENAME_EXCHANGE &&
(rflags & (RENAME_NOREPLACE | RENAME_WHITEOUT)))
return (SET_ERROR(EINVAL));
/*
* Make sure we only get wo_vap iff. RENAME_WHITEOUT and that it's the
* right kind of vattr_t for the whiteout file. These are set
* internally by ZFS so should never be incorrect.
*/
VERIFY_EQUIV(rflags & RENAME_WHITEOUT, wo_vap != NULL);
VERIFY_IMPLY(wo_vap, wo_vap->va_mode == S_IFCHR);
VERIFY_IMPLY(wo_vap, wo_vap->va_rdev == makedevice(0, 0));
if ((error = zfs_enter_verify_zp(zfsvfs, sdzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if ((error = zfs_verify_zp(tdzp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* We check i_sb because snapshots and the ctldir must have different
* super blocks.
*/
if (ZTOI(tdzp)->i_sb != ZTOI(sdzp)->i_sb ||
zfsctl_is_node(ZTOI(tdzp))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EXDEV));
}
if (zfsvfs->z_utf8 && u8_validate(tnm,
strlen(tnm), NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (flags & FIGNORECASE)
zflg |= ZCILOOK;
top:
szp = NULL;
tzp = NULL;
zl = NULL;
/*
* This is to prevent the creation of links into attribute space
* by renaming a linked file into/outof an attribute directory.
* See the comment in zfs_link() for why this is considered bad.
*/
if ((tdzp->z_pflags & ZFS_XATTR) != (sdzp->z_pflags & ZFS_XATTR)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* Lock source and target directory entries. To prevent deadlock,
* a lock ordering must be defined. We lock the directory with
* the smallest object id first, or if it's a tie, the one with
* the lexically first name.
*/
if (sdzp->z_id < tdzp->z_id) {
cmp = -1;
} else if (sdzp->z_id > tdzp->z_id) {
cmp = 1;
} else {
/*
* First compare the two name arguments without
* considering any case folding.
*/
int nofold = (zfsvfs->z_norm & ~U8_TEXTPREP_TOUPPER);
cmp = u8_strcmp(snm, tnm, 0, nofold, U8_UNICODE_LATEST, &error);
ASSERT(error == 0 || !zfsvfs->z_utf8);
if (cmp == 0) {
/*
* POSIX: "If the old argument and the new argument
* both refer to links to the same existing file,
* the rename() function shall return successfully
* and perform no other action."
*/
zfs_exit(zfsvfs, FTAG);
return (0);
}
/*
* If the file system is case-folding, then we may
* have some more checking to do. A case-folding file
* system is either supporting mixed case sensitivity
* access or is completely case-insensitive. Note
* that the file system is always case preserving.
*
* In mixed sensitivity mode case sensitive behavior
* is the default. FIGNORECASE must be used to
* explicitly request case insensitive behavior.
*
* If the source and target names provided differ only
* by case (e.g., a request to rename 'tim' to 'Tim'),
* we will treat this as a special case in the
* case-insensitive mode: as long as the source name
* is an exact match, we will allow this to proceed as
* a name-change request.
*/
if ((zfsvfs->z_case == ZFS_CASE_INSENSITIVE ||
(zfsvfs->z_case == ZFS_CASE_MIXED &&
flags & FIGNORECASE)) &&
u8_strcmp(snm, tnm, 0, zfsvfs->z_norm, U8_UNICODE_LATEST,
&error) == 0) {
/*
* case preserving rename request, require exact
* name matches
*/
zflg |= ZCIEXACT;
zflg &= ~ZCILOOK;
}
}
/*
* If the source and destination directories are the same, we should
* grab the z_name_lock of that directory only once.
*/
if (sdzp == tdzp) {
zflg |= ZHAVELOCK;
rw_enter(&sdzp->z_name_lock, RW_READER);
}
if (cmp < 0) {
serr = zfs_dirent_lock(&sdl, sdzp, snm, &szp,
ZEXISTS | zflg, NULL, NULL);
terr = zfs_dirent_lock(&tdl,
tdzp, tnm, &tzp, ZRENAMING | zflg, NULL, NULL);
} else {
terr = zfs_dirent_lock(&tdl,
tdzp, tnm, &tzp, zflg, NULL, NULL);
serr = zfs_dirent_lock(&sdl,
sdzp, snm, &szp, ZEXISTS | ZRENAMING | zflg,
NULL, NULL);
}
if (serr) {
/*
* Source entry invalid or not there.
*/
if (!terr) {
zfs_dirent_unlock(tdl);
if (tzp)
zrele(tzp);
}
if (sdzp == tdzp)
rw_exit(&sdzp->z_name_lock);
if (strcmp(snm, "..") == 0)
serr = EINVAL;
zfs_exit(zfsvfs, FTAG);
return (serr);
}
if (terr) {
zfs_dirent_unlock(sdl);
zrele(szp);
if (sdzp == tdzp)
rw_exit(&sdzp->z_name_lock);
if (strcmp(tnm, "..") == 0)
terr = EINVAL;
zfs_exit(zfsvfs, FTAG);
return (terr);
}
/*
* If we are using project inheritance, means if the directory has
* ZFS_PROJINHERIT set, then its descendant directories will inherit
* not only the project ID, but also the ZFS_PROJINHERIT flag. Under
* such case, we only allow renames into our tree when the project
* IDs are the same.
*/
if (tdzp->z_pflags & ZFS_PROJINHERIT &&
tdzp->z_projid != szp->z_projid) {
error = SET_ERROR(EXDEV);
goto out;
}
/*
* Must have write access at the source to remove the old entry
* and write access at the target to create the new entry.
* Note that if target and source are the same, this can be
* done in a single check.
*/
if ((error = zfs_zaccess_rename(sdzp, szp, tdzp, tzp, cr, mnt_ns)))
goto out;
if (S_ISDIR(ZTOI(szp)->i_mode)) {
/*
* Check to make sure rename is valid.
* Can't do a move like this: /usr/a/b to /usr/a/b/c/d
*/
if ((error = zfs_rename_lock(szp, tdzp, sdzp, &zl)))
goto out;
}
/*
* Does target exist?
*/
if (tzp) {
if (rflags & RENAME_NOREPLACE) {
error = SET_ERROR(EEXIST);
goto out;
}
/*
* Source and target must be the same type (unless exchanging).
*/
if (!(rflags & RENAME_EXCHANGE)) {
boolean_t s_is_dir = S_ISDIR(ZTOI(szp)->i_mode) != 0;
boolean_t t_is_dir = S_ISDIR(ZTOI(tzp)->i_mode) != 0;
if (s_is_dir != t_is_dir) {
error = SET_ERROR(s_is_dir ? ENOTDIR : EISDIR);
goto out;
}
}
/*
* POSIX dictates that when the source and target
* entries refer to the same file object, rename
* must do nothing and exit without error.
*/
if (szp->z_id == tzp->z_id) {
error = 0;
goto out;
}
} else if (rflags & RENAME_EXCHANGE) {
/* Target must exist for RENAME_EXCHANGE. */
error = SET_ERROR(ENOENT);
goto out;
}
/* Set up inode creation for RENAME_WHITEOUT. */
if (rflags & RENAME_WHITEOUT) {
/*
* Whiteout files are not regular files or directories, so to
* match zfs_create() we do not inherit the project id.
*/
uint64_t wo_projid = ZFS_DEFAULT_PROJID;
error = zfs_zaccess(sdzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns);
if (error)
goto out;
if (!have_acl) {
error = zfs_acl_ids_create(sdzp, 0, wo_vap, cr, NULL,
&acl_ids, mnt_ns);
if (error)
goto out;
have_acl = B_TRUE;
}
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, wo_projid)) {
error = SET_ERROR(EDQUOT);
goto out;
}
}
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, szp->z_sa_hdl, B_FALSE);
dmu_tx_hold_sa(tx, sdzp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, sdzp->z_id,
(rflags & RENAME_EXCHANGE) ? TRUE : FALSE, snm);
dmu_tx_hold_zap(tx, tdzp->z_id, TRUE, tnm);
if (sdzp != tdzp) {
dmu_tx_hold_sa(tx, tdzp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, tdzp);
}
if (tzp) {
dmu_tx_hold_sa(tx, tzp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, tzp);
}
if (rflags & RENAME_WHITEOUT) {
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
dmu_tx_hold_zap(tx, sdzp->z_id, TRUE, snm);
dmu_tx_hold_sa(tx, sdzp->z_sa_hdl, B_FALSE);
if (!zfsvfs->z_use_sa &&
acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, acl_ids.z_aclp->z_acl_bytes);
}
}
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
zfs_sa_upgrade_txholds(tx, szp);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
if (zl != NULL)
zfs_rename_unlock(&zl);
zfs_dirent_unlock(sdl);
zfs_dirent_unlock(tdl);
if (sdzp == tdzp)
rw_exit(&sdzp->z_name_lock);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
zrele(szp);
if (tzp)
zrele(tzp);
goto top;
}
dmu_tx_abort(tx);
zrele(szp);
if (tzp)
zrele(tzp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Unlink the source.
*/
szp->z_pflags |= ZFS_AV_MODIFIED;
if (tdzp->z_pflags & ZFS_PROJINHERIT)
szp->z_pflags |= ZFS_PROJINHERIT;
error = sa_update(szp->z_sa_hdl, SA_ZPL_FLAGS(zfsvfs),
(void *)&szp->z_pflags, sizeof (uint64_t), tx);
VERIFY0(error);
error = zfs_link_destroy(sdl, szp, tx, ZRENAMING, NULL);
if (error)
goto commit;
/*
* Unlink the target.
*/
if (tzp) {
int tzflg = zflg;
if (rflags & RENAME_EXCHANGE) {
/* This inode will be re-linked soon. */
tzflg |= ZRENAMING;
tzp->z_pflags |= ZFS_AV_MODIFIED;
if (sdzp->z_pflags & ZFS_PROJINHERIT)
tzp->z_pflags |= ZFS_PROJINHERIT;
error = sa_update(tzp->z_sa_hdl, SA_ZPL_FLAGS(zfsvfs),
(void *)&tzp->z_pflags, sizeof (uint64_t), tx);
ASSERT0(error);
}
error = zfs_link_destroy(tdl, tzp, tx, tzflg, NULL);
if (error)
goto commit_link_szp;
}
/*
* Create the new target links:
* * We always link the target.
* * RENAME_EXCHANGE: Link the old target to the source.
* * RENAME_WHITEOUT: Create a whiteout inode in-place of the source.
*/
error = zfs_link_create(tdl, szp, tx, ZRENAMING);
if (error) {
/*
* If we have removed the existing target, a subsequent call to
* zfs_link_create() to add back the same entry, but with a new
* dnode (szp), should not fail.
*/
ASSERT3P(tzp, ==, NULL);
goto commit_link_tzp;
}
switch (rflags & (RENAME_EXCHANGE | RENAME_WHITEOUT)) {
case RENAME_EXCHANGE:
error = zfs_link_create(sdl, tzp, tx, ZRENAMING);
/*
* The same argument as zfs_link_create() failing for
* szp applies here, since the source directory must
* have had an entry we are replacing.
*/
ASSERT0(error);
if (error)
goto commit_unlink_td_szp;
break;
case RENAME_WHITEOUT:
zfs_mknode(sdzp, wo_vap, tx, cr, 0, &wzp, &acl_ids);
error = zfs_link_create(sdl, wzp, tx, ZNEW);
if (error) {
zfs_znode_delete(wzp, tx);
remove_inode_hash(ZTOI(wzp));
goto commit_unlink_td_szp;
}
break;
}
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
switch (rflags & (RENAME_EXCHANGE | RENAME_WHITEOUT)) {
case RENAME_EXCHANGE:
zfs_log_rename_exchange(zilog, tx,
(flags & FIGNORECASE ? TX_CI : 0), sdzp, sdl->dl_name,
tdzp, tdl->dl_name, szp);
break;
case RENAME_WHITEOUT:
zfs_log_rename_whiteout(zilog, tx,
(flags & FIGNORECASE ? TX_CI : 0), sdzp, sdl->dl_name,
tdzp, tdl->dl_name, szp, wzp);
break;
default:
ASSERT0(rflags & ~RENAME_NOREPLACE);
zfs_log_rename(zilog, tx, (flags & FIGNORECASE ? TX_CI : 0),
sdzp, sdl->dl_name, tdzp, tdl->dl_name, szp);
break;
}
commit:
dmu_tx_commit(tx);
out:
if (have_acl)
zfs_acl_ids_free(&acl_ids);
zfs_znode_update_vfs(sdzp);
if (sdzp == tdzp)
rw_exit(&sdzp->z_name_lock);
if (sdzp != tdzp)
zfs_znode_update_vfs(tdzp);
zfs_znode_update_vfs(szp);
zrele(szp);
if (wzp) {
zfs_znode_update_vfs(wzp);
zrele(wzp);
}
if (tzp) {
zfs_znode_update_vfs(tzp);
zrele(tzp);
}
if (zl != NULL)
zfs_rename_unlock(&zl);
zfs_dirent_unlock(sdl);
zfs_dirent_unlock(tdl);
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
/*
* Clean-up path for broken link state.
*
* At this point we are in a (very) bad state, so we need to do our
* best to correct the state. In particular, all of the nlinks are
* wrong because we were destroying and creating links with ZRENAMING.
*
* In some form, all of these operations have to resolve the state:
*
* * link_destroy() *must* succeed. Fortunately, this is very likely
* since we only just created it.
*
* * link_create()s are allowed to fail (though they shouldn't because
* we only just unlinked them and are putting the entries back
* during clean-up). But if they fail, we can just forcefully drop
* the nlink value to (at the very least) avoid broken nlink values
* -- though in the case of non-empty directories we will have to
* panic (otherwise we'd have a leaked directory with a broken ..).
*/
commit_unlink_td_szp:
VERIFY0(zfs_link_destroy(tdl, szp, tx, ZRENAMING, NULL));
commit_link_tzp:
if (tzp) {
if (zfs_link_create(tdl, tzp, tx, ZRENAMING))
VERIFY0(zfs_drop_nlink(tzp, tx, NULL));
}
commit_link_szp:
if (zfs_link_create(sdl, szp, tx, ZRENAMING))
VERIFY0(zfs_drop_nlink(szp, tx, NULL));
goto commit;
}
/*
* Insert the indicated symbolic reference entry into the directory.
*
* IN: dzp - Directory to contain new symbolic link.
* name - Name of directory entry in dip.
* vap - Attributes of new entry.
* link - Name for new symlink entry.
* cr - credentials of caller.
* flags - case flags
* mnt_ns - user namespace of the mount
*
* OUT: zpp - Znode for new symbolic link.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dip - ctime|mtime updated
*/
int
zfs_symlink(znode_t *dzp, char *name, vattr_t *vap, char *link,
znode_t **zpp, cred_t *cr, int flags, zidmap_t *mnt_ns)
{
znode_t *zp;
zfs_dirlock_t *dl;
dmu_tx_t *tx;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zilog_t *zilog;
uint64_t len = strlen(link);
int error;
int zflg = ZNEW;
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
uint64_t txtype = TX_SYMLINK;
boolean_t waited = B_FALSE;
ASSERT(S_ISLNK(vap->va_mode));
if (name == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if (zfsvfs->z_utf8 && u8_validate(name, strlen(name),
NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (flags & FIGNORECASE)
zflg |= ZCILOOK;
if (len > MAXPATHLEN) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENAMETOOLONG));
}
if ((error = zfs_acl_ids_create(dzp, 0,
vap, cr, NULL, &acl_ids, mnt_ns)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
top:
*zpp = NULL;
/*
* Attempt to lock directory; fail if entry already exists.
*/
error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg, NULL, NULL);
if (error) {
zfs_acl_ids_free(&acl_ids);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns))) {
zfs_acl_ids_free(&acl_ids);
zfs_dirent_unlock(dl);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, ZFS_DEFAULT_PROJID)) {
zfs_acl_ids_free(&acl_ids);
zfs_dirent_unlock(dl);
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EDQUOT));
}
tx = dmu_tx_create(zfsvfs->z_os);
fuid_dirtied = zfsvfs->z_fuid_dirty;
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, MAX(1, len));
dmu_tx_hold_zap(tx, dzp->z_id, TRUE, name);
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE + len);
dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE);
if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0,
acl_ids.z_aclp->z_acl_bytes);
}
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Create a new object for the symlink.
* for version 4 ZPL datasets the symlink will be an SA attribute
*/
zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids);
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
mutex_enter(&zp->z_lock);
if (zp->z_is_sa)
error = sa_update(zp->z_sa_hdl, SA_ZPL_SYMLINK(zfsvfs),
link, len, tx);
else
zfs_sa_symlink(zp, link, len, tx);
mutex_exit(&zp->z_lock);
zp->z_size = len;
(void) sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(zfsvfs),
&zp->z_size, sizeof (zp->z_size), tx);
/*
* Insert the new object into the directory.
*/
error = zfs_link_create(dl, zp, tx, ZNEW);
if (error != 0) {
zfs_znode_delete(zp, tx);
remove_inode_hash(ZTOI(zp));
} else {
if (flags & FIGNORECASE)
txtype |= TX_CI;
zfs_log_symlink(zilog, tx, txtype, dzp, zp, name, link);
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
}
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
zfs_dirent_unlock(dl);
if (error == 0) {
*zpp = zp;
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
} else {
zrele(zp);
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Return, in the buffer contained in the provided uio structure,
* the symbolic path referred to by ip.
*
* IN: ip - inode of symbolic link
* uio - structure to contain the link path.
* cr - credentials of caller.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* ip - atime updated
*/
int
zfs_readlink(struct inode *ip, zfs_uio_t *uio, cred_t *cr)
{
(void) cr;
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
mutex_enter(&zp->z_lock);
if (zp->z_is_sa)
error = sa_lookup_uio(zp->z_sa_hdl,
SA_ZPL_SYMLINK(zfsvfs), uio);
else
error = zfs_sa_readlink(zp, uio);
mutex_exit(&zp->z_lock);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Insert a new entry into directory tdzp referencing szp.
*
* IN: tdzp - Directory to contain new entry.
* szp - znode of new entry.
* name - name of new entry.
* cr - credentials of caller.
* flags - case flags.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* tdzp - ctime|mtime updated
* szp - ctime updated
*/
int
zfs_link(znode_t *tdzp, znode_t *szp, char *name, cred_t *cr,
int flags)
{
struct inode *sip = ZTOI(szp);
znode_t *tzp;
zfsvfs_t *zfsvfs = ZTOZSB(tdzp);
zilog_t *zilog;
zfs_dirlock_t *dl;
dmu_tx_t *tx;
int error;
int zf = ZNEW;
uint64_t parent;
uid_t owner;
boolean_t waited = B_FALSE;
boolean_t is_tmpfile = 0;
uint64_t txg;
#ifdef HAVE_TMPFILE
is_tmpfile = (sip->i_nlink == 0 && (sip->i_state & I_LINKABLE));
#endif
ASSERT(S_ISDIR(ZTOI(tdzp)->i_mode));
if (name == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, tdzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
/*
* POSIX dictates that we return EPERM here.
* Better choices include ENOTSUP or EISDIR.
*/
if (S_ISDIR(sip->i_mode)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if ((error = zfs_verify_zp(szp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* If we are using project inheritance, means if the directory has
* ZFS_PROJINHERIT set, then its descendant directories will inherit
* not only the project ID, but also the ZFS_PROJINHERIT flag. Under
* such case, we only allow hard link creation in our tree when the
* project IDs are the same.
*/
if (tdzp->z_pflags & ZFS_PROJINHERIT &&
tdzp->z_projid != szp->z_projid) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EXDEV));
}
/*
* We check i_sb because snapshots and the ctldir must have different
* super blocks.
*/
if (sip->i_sb != ZTOI(tdzp)->i_sb || zfsctl_is_node(sip)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EXDEV));
}
/* Prevent links to .zfs/shares files */
if ((error = sa_lookup(szp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs),
&parent, sizeof (uint64_t))) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (parent == zfsvfs->z_shares_dir) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if (zfsvfs->z_utf8 && u8_validate(name,
strlen(name), NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (flags & FIGNORECASE)
zf |= ZCILOOK;
/*
* We do not support links between attributes and non-attributes
* because of the potential security risk of creating links
* into "normal" file space in order to circumvent restrictions
* imposed in attribute space.
*/
if ((szp->z_pflags & ZFS_XATTR) != (tdzp->z_pflags & ZFS_XATTR)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
owner = zfs_fuid_map_id(zfsvfs, KUID_TO_SUID(sip->i_uid),
cr, ZFS_OWNER);
if (owner != crgetuid(cr) && secpolicy_basic_link(cr) != 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if ((error = zfs_zaccess(tdzp, ACE_ADD_FILE, 0, B_FALSE, cr,
zfs_init_idmap))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
top:
/*
* Attempt to lock directory; fail if entry already exists.
*/
error = zfs_dirent_lock(&dl, tdzp, name, &tzp, zf, NULL, NULL);
if (error) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, szp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, tdzp->z_id, TRUE, name);
if (is_tmpfile)
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
zfs_sa_upgrade_txholds(tx, szp);
zfs_sa_upgrade_txholds(tx, tdzp);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/* unmark z_unlinked so zfs_link_create will not reject */
if (is_tmpfile)
szp->z_unlinked = B_FALSE;
error = zfs_link_create(dl, szp, tx, 0);
if (error == 0) {
uint64_t txtype = TX_LINK;
/*
* tmpfile is created to be in z_unlinkedobj, so remove it.
* Also, we don't log in ZIL, because all previous file
* operation on the tmpfile are ignored by ZIL. Instead we
* always wait for txg to sync to make sure all previous
* operation are sync safe.
*/
if (is_tmpfile) {
VERIFY(zap_remove_int(zfsvfs->z_os,
zfsvfs->z_unlinkedobj, szp->z_id, tx) == 0);
} else {
if (flags & FIGNORECASE)
txtype |= TX_CI;
zfs_log_link(zilog, tx, txtype, tdzp, szp, name);
}
} else if (is_tmpfile) {
/* restore z_unlinked since when linking failed */
szp->z_unlinked = B_TRUE;
}
txg = dmu_tx_get_txg(tx);
dmu_tx_commit(tx);
zfs_dirent_unlock(dl);
if (!is_tmpfile && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
if (is_tmpfile && zfsvfs->z_os->os_sync != ZFS_SYNC_DISABLED)
txg_wait_synced(dmu_objset_pool(zfsvfs->z_os), txg);
zfs_znode_update_vfs(tdzp);
zfs_znode_update_vfs(szp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
static void
zfs_putpage_sync_commit_cb(void *arg)
{
struct page *pp = arg;
ClearPageError(pp);
end_page_writeback(pp);
}
static void
zfs_putpage_async_commit_cb(void *arg)
{
struct page *pp = arg;
znode_t *zp = ITOZ(pp->mapping->host);
ClearPageError(pp);
end_page_writeback(pp);
atomic_dec_32(&zp->z_async_writes_cnt);
}
/*
* Push a page out to disk, once the page is on stable storage the
* registered commit callback will be run as notification of completion.
*
* IN: ip - page mapped for inode.
* pp - page to push (page is locked)
* wbc - writeback control data
* for_sync - does the caller intend to wait synchronously for the
* page writeback to complete?
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* ip - ctime|mtime updated
*/
int
zfs_putpage(struct inode *ip, struct page *pp, struct writeback_control *wbc,
boolean_t for_sync)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
loff_t offset;
loff_t pgoff;
unsigned int pglen;
dmu_tx_t *tx;
caddr_t va;
int err = 0;
uint64_t mtime[2], ctime[2];
inode_timespec_t tmp_ctime;
sa_bulk_attr_t bulk[3];
int cnt = 0;
struct address_space *mapping;
if ((err = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (err);
ASSERT(PageLocked(pp));
pgoff = page_offset(pp); /* Page byte-offset in file */
offset = i_size_read(ip); /* File length in bytes */
pglen = MIN(PAGE_SIZE, /* Page length in bytes */
P2ROUNDUP(offset, PAGE_SIZE)-pgoff);
/* Page is beyond end of file */
if (pgoff >= offset) {
unlock_page(pp);
zfs_exit(zfsvfs, FTAG);
return (0);
}
/* Truncate page length to end of file */
if (pgoff + pglen > offset)
pglen = offset - pgoff;
#if 0
/*
* FIXME: Allow mmap writes past its quota. The correct fix
* is to register a page_mkwrite() handler to count the page
* against its quota when it is about to be dirtied.
*/
if (zfs_id_overblockquota(zfsvfs, DMU_USERUSED_OBJECT,
KUID_TO_SUID(ip->i_uid)) ||
zfs_id_overblockquota(zfsvfs, DMU_GROUPUSED_OBJECT,
KGID_TO_SGID(ip->i_gid)) ||
(zp->z_projid != ZFS_DEFAULT_PROJID &&
zfs_id_overblockquota(zfsvfs, DMU_PROJECTUSED_OBJECT,
zp->z_projid))) {
err = EDQUOT;
}
#endif
/*
* The ordering here is critical and must adhere to the following
* rules in order to avoid deadlocking in either zfs_read() or
* zfs_free_range() due to a lock inversion.
*
* 1) The page must be unlocked prior to acquiring the range lock.
* This is critical because zfs_read() calls find_lock_page()
* which may block on the page lock while holding the range lock.
*
* 2) Before setting or clearing write back on a page the range lock
* must be held in order to prevent a lock inversion with the
* zfs_free_range() function.
*
* This presents a problem because upon entering this function the
* page lock is already held. To safely acquire the range lock the
* page lock must be dropped. This creates a window where another
* process could truncate, invalidate, dirty, or write out the page.
*
* Therefore, after successfully reacquiring the range and page locks
* the current page state is checked. In the common case everything
* will be as is expected and it can be written out. However, if
* the page state has changed it must be handled accordingly.
*/
mapping = pp->mapping;
redirty_page_for_writepage(wbc, pp);
unlock_page(pp);
zfs_locked_range_t *lr = zfs_rangelock_enter(&zp->z_rangelock,
pgoff, pglen, RL_WRITER);
lock_page(pp);
/* Page mapping changed or it was no longer dirty, we're done */
if (unlikely((mapping != pp->mapping) || !PageDirty(pp))) {
unlock_page(pp);
zfs_rangelock_exit(lr);
zfs_exit(zfsvfs, FTAG);
return (0);
}
/* Another process started write block if required */
if (PageWriteback(pp)) {
unlock_page(pp);
zfs_rangelock_exit(lr);
if (wbc->sync_mode != WB_SYNC_NONE) {
/*
* Speed up any non-sync page writebacks since
* they may take several seconds to complete.
* Refer to the comment in zpl_fsync() (when
* HAVE_FSYNC_RANGE is defined) for details.
*/
if (atomic_load_32(&zp->z_async_writes_cnt) > 0) {
zil_commit(zfsvfs->z_log, zp->z_id);
}
if (PageWriteback(pp))
#ifdef HAVE_PAGEMAP_FOLIO_WAIT_BIT
folio_wait_bit(page_folio(pp), PG_writeback);
#else
wait_on_page_bit(pp, PG_writeback);
#endif
}
zfs_exit(zfsvfs, FTAG);
return (0);
}
/* Clear the dirty flag the required locks are held */
if (!clear_page_dirty_for_io(pp)) {
unlock_page(pp);
zfs_rangelock_exit(lr);
zfs_exit(zfsvfs, FTAG);
return (0);
}
/*
* Counterpart for redirty_page_for_writepage() above. This page
* was in fact not skipped and should not be counted as if it were.
*/
wbc->pages_skipped--;
if (!for_sync)
atomic_inc_32(&zp->z_async_writes_cnt);
set_page_writeback(pp);
unlock_page(pp);
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_write(tx, zp->z_id, pgoff, pglen);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
err = dmu_tx_assign(tx, TXG_NOWAIT);
if (err != 0) {
if (err == ERESTART)
dmu_tx_wait(tx);
dmu_tx_abort(tx);
#ifdef HAVE_VFS_FILEMAP_DIRTY_FOLIO
filemap_dirty_folio(page_mapping(pp), page_folio(pp));
#else
__set_page_dirty_nobuffers(pp);
#endif
ClearPageError(pp);
end_page_writeback(pp);
if (!for_sync)
atomic_dec_32(&zp->z_async_writes_cnt);
zfs_rangelock_exit(lr);
zfs_exit(zfsvfs, FTAG);
return (err);
}
va = kmap(pp);
ASSERT3U(pglen, <=, PAGE_SIZE);
dmu_write(zfsvfs->z_os, zp->z_id, pgoff, pglen, va, tx);
kunmap(pp);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, 8);
/* Preserve the mtime and ctime provided by the inode */
ZFS_TIME_ENCODE(&ip->i_mtime, mtime);
tmp_ctime = zpl_inode_get_ctime(ip);
ZFS_TIME_ENCODE(&tmp_ctime, ctime);
zp->z_atime_dirty = B_FALSE;
zp->z_seq++;
err = sa_bulk_update(zp->z_sa_hdl, bulk, cnt, tx);
boolean_t commit = B_FALSE;
if (wbc->sync_mode != WB_SYNC_NONE) {
/*
* Note that this is rarely called under writepages(), because
* writepages() normally handles the entire commit for
* performance reasons.
*/
commit = B_TRUE;
} else if (!for_sync && atomic_load_32(&zp->z_sync_writes_cnt) > 0) {
/*
* If the caller does not intend to wait synchronously
* for this page writeback to complete and there are active
* synchronous calls on this file, do a commit so that
* the latter don't accidentally end up waiting for
* our writeback to complete. Refer to the comment in
* zpl_fsync() (when HAVE_FSYNC_RANGE is defined) for details.
*/
commit = B_TRUE;
}
zfs_log_write(zfsvfs->z_log, tx, TX_WRITE, zp, pgoff, pglen, commit,
for_sync ? zfs_putpage_sync_commit_cb :
zfs_putpage_async_commit_cb, pp);
dmu_tx_commit(tx);
zfs_rangelock_exit(lr);
if (commit)
zil_commit(zfsvfs->z_log, zp->z_id);
dataset_kstats_update_write_kstats(&zfsvfs->z_kstat, pglen);
zfs_exit(zfsvfs, FTAG);
return (err);
}
/*
* Update the system attributes when the inode has been dirtied. For the
* moment we only update the mode, atime, mtime, and ctime.
*/
int
zfs_dirty_inode(struct inode *ip, int flags)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
dmu_tx_t *tx;
uint64_t mode, atime[2], mtime[2], ctime[2];
inode_timespec_t tmp_ctime;
sa_bulk_attr_t bulk[4];
int error = 0;
int cnt = 0;
if (zfs_is_readonly(zfsvfs) || dmu_objset_is_snapshot(zfsvfs->z_os))
return (0);
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
#ifdef I_DIRTY_TIME
/*
* This is the lazytime semantic introduced in Linux 4.0
* This flag will only be called from update_time when lazytime is set.
* (Note, I_DIRTY_SYNC will also set if not lazytime)
* Fortunately mtime and ctime are managed within ZFS itself, so we
* only need to dirty atime.
*/
if (flags == I_DIRTY_TIME) {
zp->z_atime_dirty = B_TRUE;
goto out;
}
#endif
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
goto out;
}
mutex_enter(&zp->z_lock);
zp->z_atime_dirty = B_FALSE;
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_MODE(zfsvfs), NULL, &mode, 8);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_ATIME(zfsvfs), NULL, &atime, 16);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16);
/* Preserve the mode, mtime and ctime provided by the inode */
ZFS_TIME_ENCODE(&ip->i_atime, atime);
ZFS_TIME_ENCODE(&ip->i_mtime, mtime);
tmp_ctime = zpl_inode_get_ctime(ip);
ZFS_TIME_ENCODE(&tmp_ctime, ctime);
mode = ip->i_mode;
zp->z_mode = mode;
error = sa_bulk_update(zp->z_sa_hdl, bulk, cnt, tx);
mutex_exit(&zp->z_lock);
dmu_tx_commit(tx);
out:
zfs_exit(zfsvfs, FTAG);
return (error);
}
void
zfs_inactive(struct inode *ip)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
uint64_t atime[2];
int error;
int need_unlock = 0;
/* Only read lock if we haven't already write locked, e.g. rollback */
if (!RW_WRITE_HELD(&zfsvfs->z_teardown_inactive_lock)) {
need_unlock = 1;
rw_enter(&zfsvfs->z_teardown_inactive_lock, RW_READER);
}
if (zp->z_sa_hdl == NULL) {
if (need_unlock)
rw_exit(&zfsvfs->z_teardown_inactive_lock);
return;
}
if (zp->z_atime_dirty && zp->z_unlinked == B_FALSE) {
dmu_tx_t *tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
} else {
ZFS_TIME_ENCODE(&ip->i_atime, atime);
mutex_enter(&zp->z_lock);
(void) sa_update(zp->z_sa_hdl, SA_ZPL_ATIME(zfsvfs),
(void *)&atime, sizeof (atime), tx);
zp->z_atime_dirty = B_FALSE;
mutex_exit(&zp->z_lock);
dmu_tx_commit(tx);
}
}
zfs_zinactive(zp);
if (need_unlock)
rw_exit(&zfsvfs->z_teardown_inactive_lock);
}
/*
* Fill pages with data from the disk.
*/
static int
zfs_fillpage(struct inode *ip, struct page *pp)
{
zfsvfs_t *zfsvfs = ITOZSB(ip);
loff_t i_size = i_size_read(ip);
u_offset_t io_off = page_offset(pp);
size_t io_len = PAGE_SIZE;
ASSERT3U(io_off, <, i_size);
if (io_off + io_len > i_size)
io_len = i_size - io_off;
void *va = kmap(pp);
int error = dmu_read(zfsvfs->z_os, ITOZ(ip)->z_id, io_off,
io_len, va, DMU_READ_PREFETCH);
if (io_len != PAGE_SIZE)
memset((char *)va + io_len, 0, PAGE_SIZE - io_len);
kunmap(pp);
if (error) {
/* convert checksum errors into IO errors */
if (error == ECKSUM)
error = SET_ERROR(EIO);
SetPageError(pp);
ClearPageUptodate(pp);
} else {
ClearPageError(pp);
SetPageUptodate(pp);
}
return (error);
}
/*
* Uses zfs_fillpage to read data from the file and fill the page.
*
* IN: ip - inode of file to get data from.
* pp - page to read
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* vp - atime updated
*/
int
zfs_getpage(struct inode *ip, struct page *pp)
{
zfsvfs_t *zfsvfs = ITOZSB(ip);
znode_t *zp = ITOZ(ip);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
error = zfs_fillpage(ip, pp);
if (error == 0)
dataset_kstats_update_read_kstats(&zfsvfs->z_kstat, PAGE_SIZE);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Check ZFS specific permissions to memory map a section of a file.
*
* IN: ip - inode of the file to mmap
* off - file offset
* addrp - start address in memory region
* len - length of memory region
* vm_flags- address flags
*
* RETURN: 0 if success
* error code if failure
*/
int
zfs_map(struct inode *ip, offset_t off, caddr_t *addrp, size_t len,
unsigned long vm_flags)
{
(void) addrp;
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
- if ((vm_flags & VM_WRITE) && (zp->z_pflags &
- (ZFS_IMMUTABLE | ZFS_READONLY | ZFS_APPENDONLY))) {
+ if ((vm_flags & VM_WRITE) && (vm_flags & VM_SHARED) &&
+ (zp->z_pflags & (ZFS_IMMUTABLE | ZFS_READONLY | ZFS_APPENDONLY))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if ((vm_flags & (VM_READ | VM_EXEC)) &&
(zp->z_pflags & ZFS_AV_QUARANTINED)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EACCES));
}
if (off < 0 || len > MAXOFFSET_T - off) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENXIO));
}
zfs_exit(zfsvfs, FTAG);
return (0);
}
/*
* Free or allocate space in a file. Currently, this function only
* supports the `F_FREESP' command. However, this command is somewhat
* misnamed, as its functionality includes the ability to allocate as
* well as free space.
*
* IN: zp - znode of file to free data in.
* cmd - action to take (only F_FREESP supported).
* bfp - section of file to free/alloc.
* flag - current file open mode flags.
* offset - current file offset.
* cr - credentials of caller.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* zp - ctime|mtime updated
*/
int
zfs_space(znode_t *zp, int cmd, flock64_t *bfp, int flag,
offset_t offset, cred_t *cr)
{
(void) offset;
zfsvfs_t *zfsvfs = ZTOZSB(zp);
uint64_t off, len;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if (cmd != F_FREESP) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* Callers might not be able to detect properly that we are read-only,
* so check it explicitly here.
*/
if (zfs_is_readonly(zfsvfs)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EROFS));
}
if (bfp->l_len < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* Permissions aren't checked on Solaris because on this OS
* zfs_space() can only be called with an opened file handle.
* On Linux we can get here through truncate_range() which
* operates directly on inodes, so we need to check access rights.
*/
if ((error = zfs_zaccess(zp, ACE_WRITE_DATA, 0, B_FALSE, cr,
zfs_init_idmap))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
off = bfp->l_start;
len = bfp->l_len; /* 0 means from off to end of file */
error = zfs_freesp(zp, off, len, flag, TRUE);
zfs_exit(zfsvfs, FTAG);
return (error);
}
int
zfs_fid(struct inode *ip, fid_t *fidp)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
uint32_t gen;
uint64_t gen64;
uint64_t object = zp->z_id;
zfid_short_t *zfid;
int size, i, error;
if ((error = zfs_enter(zfsvfs, FTAG)) != 0)
return (error);
if (fidp->fid_len < SHORT_FID_LEN) {
fidp->fid_len = SHORT_FID_LEN;
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENOSPC));
}
if ((error = zfs_verify_zp(zp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_GEN(zfsvfs),
&gen64, sizeof (uint64_t))) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
gen = (uint32_t)gen64;
size = SHORT_FID_LEN;
zfid = (zfid_short_t *)fidp;
zfid->zf_len = size;
for (i = 0; i < sizeof (zfid->zf_object); i++)
zfid->zf_object[i] = (uint8_t)(object >> (8 * i));
/* Must have a non-zero generation number to distinguish from .zfs */
if (gen == 0)
gen = 1;
for (i = 0; i < sizeof (zfid->zf_gen); i++)
zfid->zf_gen[i] = (uint8_t)(gen >> (8 * i));
zfs_exit(zfsvfs, FTAG);
return (0);
}
#if defined(_KERNEL)
EXPORT_SYMBOL(zfs_open);
EXPORT_SYMBOL(zfs_close);
EXPORT_SYMBOL(zfs_lookup);
EXPORT_SYMBOL(zfs_create);
EXPORT_SYMBOL(zfs_tmpfile);
EXPORT_SYMBOL(zfs_remove);
EXPORT_SYMBOL(zfs_mkdir);
EXPORT_SYMBOL(zfs_rmdir);
EXPORT_SYMBOL(zfs_readdir);
EXPORT_SYMBOL(zfs_getattr_fast);
EXPORT_SYMBOL(zfs_setattr);
EXPORT_SYMBOL(zfs_rename);
EXPORT_SYMBOL(zfs_symlink);
EXPORT_SYMBOL(zfs_readlink);
EXPORT_SYMBOL(zfs_link);
EXPORT_SYMBOL(zfs_inactive);
EXPORT_SYMBOL(zfs_space);
EXPORT_SYMBOL(zfs_fid);
EXPORT_SYMBOL(zfs_getpage);
EXPORT_SYMBOL(zfs_putpage);
EXPORT_SYMBOL(zfs_dirty_inode);
EXPORT_SYMBOL(zfs_map);
/* CSTYLED */
module_param(zfs_delete_blocks, ulong, 0644);
MODULE_PARM_DESC(zfs_delete_blocks, "Delete files larger than N blocks async");
#endif
diff --git a/sys/contrib/openzfs/module/zcommon/zfeature_common.c b/sys/contrib/openzfs/module/zcommon/zfeature_common.c
index 2c74d10f43ff..309d9bf14cd4 100644
--- a/sys/contrib/openzfs/module/zcommon/zfeature_common.c
+++ b/sys/contrib/openzfs/module/zcommon/zfeature_common.c
@@ -1,763 +1,768 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2011, 2018 by Delphix. All rights reserved.
* Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
* Copyright (c) 2013, Joyent, Inc. All rights reserved.
* Copyright (c) 2014, Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019, Klara Inc.
* Copyright (c) 2019, Allan Jude
*/
#ifndef _KERNEL
#include <errno.h>
#include <string.h>
#include <dirent.h>
#include <search.h>
#include <sys/stat.h>
#endif
#include <sys/debug.h>
#include <sys/fs/zfs.h>
#include <sys/inttypes.h>
#include <sys/types.h>
#include <sys/param.h>
#include <sys/zfs_sysfs.h>
#include "zfeature_common.h"
/*
* Set to disable all feature checks while opening pools, allowing pools with
* unsupported features to be opened. Set for testing only.
*/
boolean_t zfeature_checks_disable = B_FALSE;
zfeature_info_t spa_feature_table[SPA_FEATURES];
/*
* Valid characters for feature guids. This list is mainly for aesthetic
* purposes and could be expanded in the future. There are different allowed
* characters in the guids reverse dns portion (before the colon) and its
* short name (after the colon).
*/
static int
valid_char(char c, boolean_t after_colon)
{
return ((c >= 'a' && c <= 'z') ||
(c >= '0' && c <= '9') ||
(after_colon && c == '_') ||
(!after_colon && (c == '.' || c == '-')));
}
/*
* Every feature guid must contain exactly one colon which separates a reverse
* dns organization name from the feature's "short" name (e.g.
* "com.company:feature_name").
*/
boolean_t
zfeature_is_valid_guid(const char *name)
{
int i;
boolean_t has_colon = B_FALSE;
i = 0;
while (name[i] != '\0') {
char c = name[i++];
if (c == ':') {
if (has_colon)
return (B_FALSE);
has_colon = B_TRUE;
continue;
}
if (!valid_char(c, has_colon))
return (B_FALSE);
}
return (has_colon);
}
boolean_t
zfeature_is_supported(const char *guid)
{
if (zfeature_checks_disable)
return (B_TRUE);
for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
zfeature_info_t *feature = &spa_feature_table[i];
if (!feature->fi_zfs_mod_supported)
continue;
if (strcmp(guid, feature->fi_guid) == 0)
return (B_TRUE);
}
return (B_FALSE);
}
int
zfeature_lookup_guid(const char *guid, spa_feature_t *res)
{
for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
zfeature_info_t *feature = &spa_feature_table[i];
if (!feature->fi_zfs_mod_supported)
continue;
if (strcmp(guid, feature->fi_guid) == 0) {
if (res != NULL)
*res = i;
return (0);
}
}
return (ENOENT);
}
int
zfeature_lookup_name(const char *name, spa_feature_t *res)
{
for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
zfeature_info_t *feature = &spa_feature_table[i];
if (!feature->fi_zfs_mod_supported)
continue;
if (strcmp(name, feature->fi_uname) == 0) {
if (res != NULL)
*res = i;
return (0);
}
}
return (ENOENT);
}
boolean_t
zfeature_depends_on(spa_feature_t fid, spa_feature_t check)
{
zfeature_info_t *feature = &spa_feature_table[fid];
for (int i = 0; feature->fi_depends[i] != SPA_FEATURE_NONE; i++) {
if (feature->fi_depends[i] == check)
return (B_TRUE);
}
return (B_FALSE);
}
static boolean_t
deps_contains_feature(const spa_feature_t *deps, const spa_feature_t feature)
{
for (int i = 0; deps[i] != SPA_FEATURE_NONE; i++)
if (deps[i] == feature)
return (B_TRUE);
return (B_FALSE);
}
#define STRCMP ((int(*)(const void *, const void *))&strcmp)
struct zfs_mod_supported_features {
void *tree;
boolean_t all_features;
};
struct zfs_mod_supported_features *
zfs_mod_list_supported(const char *scope)
{
#if defined(__FreeBSD__) || defined(_KERNEL) || defined(LIB_ZPOOL_BUILD)
(void) scope;
return (NULL);
#else
struct zfs_mod_supported_features *ret = calloc(1, sizeof (*ret));
if (ret == NULL)
return (NULL);
DIR *sysfs_dir = NULL;
char path[128];
if (snprintf(path, sizeof (path), "%s/%s",
ZFS_SYSFS_DIR, scope) < sizeof (path))
sysfs_dir = opendir(path);
if (sysfs_dir == NULL && errno == ENOENT) {
if (snprintf(path, sizeof (path), "%s/%s",
ZFS_SYSFS_ALT_DIR, scope) < sizeof (path))
sysfs_dir = opendir(path);
}
if (sysfs_dir == NULL) {
ret->all_features = errno == ENOENT &&
(access(ZFS_SYSFS_DIR, F_OK) == 0 ||
access(ZFS_SYSFS_ALT_DIR, F_OK) == 0);
return (ret);
}
struct dirent *node;
while ((node = readdir(sysfs_dir)) != NULL) {
if (strcmp(node->d_name, ".") == 0 ||
strcmp(node->d_name, "..") == 0)
continue;
char *name = strdup(node->d_name);
if (name == NULL) {
goto nomem;
}
if (tsearch(name, &ret->tree, STRCMP) == NULL) {
/*
* Don't bother checking for duplicate entries:
* we're iterating a single directory.
*/
free(name);
goto nomem;
}
}
end:
closedir(sysfs_dir);
return (ret);
nomem:
zfs_mod_list_supported_free(ret);
ret = NULL;
goto end;
#endif
}
void
zfs_mod_list_supported_free(struct zfs_mod_supported_features *list)
{
#if !defined(__FreeBSD__) && !defined(_KERNEL) && !defined(LIB_ZPOOL_BUILD)
if (list) {
tdestroy(list->tree, free);
free(list);
}
#else
(void) list;
#endif
}
#if !defined(_KERNEL) && !defined(LIB_ZPOOL_BUILD)
static boolean_t
zfs_mod_supported_impl(const char *scope, const char *name, const char *sysfs)
{
char path[128];
if (snprintf(path, sizeof (path), "%s%s%s%s%s", sysfs,
scope == NULL ? "" : "/", scope ?: "",
name == NULL ? "" : "/", name ?: "") < sizeof (path))
return (access(path, F_OK) == 0);
else
return (B_FALSE);
}
boolean_t
zfs_mod_supported(const char *scope, const char *name,
const struct zfs_mod_supported_features *sfeatures)
{
boolean_t supported;
if (sfeatures != NULL)
return (sfeatures->all_features ||
tfind(name, &sfeatures->tree, STRCMP));
/*
* Check both the primary and alternate sysfs locations to determine
* if the required functionality is supported.
*/
supported = (zfs_mod_supported_impl(scope, name, ZFS_SYSFS_DIR) ||
zfs_mod_supported_impl(scope, name, ZFS_SYSFS_ALT_DIR));
/*
* For backwards compatibility with kernel modules that predate
* supported feature/property checking. Report the feature/property
* as supported if the kernel module is loaded but the requested
* scope directory does not exist.
*/
if (supported == B_FALSE) {
if ((access(ZFS_SYSFS_DIR, F_OK) == 0 &&
!zfs_mod_supported_impl(scope, NULL, ZFS_SYSFS_DIR)) ||
(access(ZFS_SYSFS_ALT_DIR, F_OK) == 0 &&
!zfs_mod_supported_impl(scope, NULL, ZFS_SYSFS_ALT_DIR))) {
supported = B_TRUE;
}
}
return (supported);
}
#endif
static boolean_t
zfs_mod_supported_feature(const char *name,
const struct zfs_mod_supported_features *sfeatures)
{
/*
* The zfs module spa_feature_table[], whether in-kernel or in
* libzpool, always supports all the features. libzfs needs to
* query the running module, via sysfs, to determine which
* features are supported.
*
* The equivalent _can_ be done on FreeBSD by way of the sysctl
* tree, but this has not been done yet. Therefore, we return
* that all features are supported.
*/
#if defined(_KERNEL) || defined(LIB_ZPOOL_BUILD) || defined(__FreeBSD__)
(void) name, (void) sfeatures;
return (B_TRUE);
#else
return (zfs_mod_supported(ZFS_SYSFS_POOL_FEATURES, name, sfeatures));
#endif
}
static void
zfeature_register(spa_feature_t fid, const char *guid, const char *name,
const char *desc, zfeature_flags_t flags, zfeature_type_t type,
const spa_feature_t *deps,
const struct zfs_mod_supported_features *sfeatures)
{
zfeature_info_t *feature = &spa_feature_table[fid];
static const spa_feature_t nodeps[] = { SPA_FEATURE_NONE };
ASSERT(name != NULL);
ASSERT(desc != NULL);
ASSERT((flags & ZFEATURE_FLAG_READONLY_COMPAT) == 0 ||
(flags & ZFEATURE_FLAG_MOS) == 0);
ASSERT3U(fid, <, SPA_FEATURES);
ASSERT(zfeature_is_valid_guid(guid));
if (deps == NULL)
deps = nodeps;
VERIFY(((flags & ZFEATURE_FLAG_PER_DATASET) == 0) ||
(deps_contains_feature(deps, SPA_FEATURE_EXTENSIBLE_DATASET)));
feature->fi_feature = fid;
feature->fi_guid = guid;
feature->fi_uname = name;
feature->fi_desc = desc;
feature->fi_flags = flags;
feature->fi_type = type;
feature->fi_depends = deps;
feature->fi_zfs_mod_supported =
zfs_mod_supported_feature(guid, sfeatures);
}
/*
* Every feature has a GUID of the form com.example:feature_name. The
* reversed DNS name ensures that the feature's GUID is unique across all ZFS
* implementations. This allows companies to independently develop and
* release features. Examples include org.delphix and org.datto. Previously,
* features developed on one implementation have used that implementation's
* domain name (e.g. org.illumos and org.zfsonlinux). Use of the org.openzfs
* domain name is recommended for new features which are developed by the
* OpenZFS community and its platforms. This domain may optionally be used by
* companies developing features for initial release through an OpenZFS
* implementation. Use of the org.openzfs domain requires reserving the
* feature name in advance with the OpenZFS project.
*/
void
zpool_feature_init(void)
{
struct zfs_mod_supported_features *sfeatures =
zfs_mod_list_supported(ZFS_SYSFS_POOL_FEATURES);
zfeature_register(SPA_FEATURE_ASYNC_DESTROY,
"com.delphix:async_destroy", "async_destroy",
"Destroy filesystems asynchronously.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_EMPTY_BPOBJ,
"com.delphix:empty_bpobj", "empty_bpobj",
"Snapshots use less space.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_LZ4_COMPRESS,
"org.illumos:lz4_compress", "lz4_compress",
"LZ4 compression algorithm support.",
ZFEATURE_FLAG_ACTIVATE_ON_ENABLE, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_MULTI_VDEV_CRASH_DUMP,
"com.joyent:multi_vdev_crash_dump", "multi_vdev_crash_dump",
"Crash dumps to multiple vdev pools.",
0, ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
zfeature_register(SPA_FEATURE_SPACEMAP_HISTOGRAM,
"com.delphix:spacemap_histogram", "spacemap_histogram",
"Spacemaps maintain space histograms.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_ENABLED_TXG,
"com.delphix:enabled_txg", "enabled_txg",
"Record txg at which a feature is enabled",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
{
static const spa_feature_t hole_birth_deps[] = {
SPA_FEATURE_ENABLED_TXG,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_HOLE_BIRTH,
"com.delphix:hole_birth", "hole_birth",
"Retain hole birth txg for more precise zfs send",
ZFEATURE_FLAG_MOS | ZFEATURE_FLAG_ACTIVATE_ON_ENABLE,
ZFEATURE_TYPE_BOOLEAN, hole_birth_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_POOL_CHECKPOINT,
"com.delphix:zpool_checkpoint", "zpool_checkpoint",
"Pool state can be checkpointed, allowing rewind later.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_SPACEMAP_V2,
"com.delphix:spacemap_v2", "spacemap_v2",
"Space maps representing large segments are more efficient.",
ZFEATURE_FLAG_READONLY_COMPAT | ZFEATURE_FLAG_ACTIVATE_ON_ENABLE,
ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
zfeature_register(SPA_FEATURE_EXTENSIBLE_DATASET,
"com.delphix:extensible_dataset", "extensible_dataset",
"Enhanced dataset functionality, used by other features.",
0, ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
{
static const spa_feature_t bookmarks_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_BOOKMARKS,
"com.delphix:bookmarks", "bookmarks",
"\"zfs bookmark\" command",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN,
bookmarks_deps, sfeatures);
}
{
static const spa_feature_t filesystem_limits_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_FS_SS_LIMIT,
"com.joyent:filesystem_limits", "filesystem_limits",
"Filesystem and snapshot limits.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN,
filesystem_limits_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_EMBEDDED_DATA,
"com.delphix:embedded_data", "embedded_data",
"Blocks which compress very well use even less space.",
ZFEATURE_FLAG_MOS | ZFEATURE_FLAG_ACTIVATE_ON_ENABLE,
ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
{
static const spa_feature_t livelist_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_LIVELIST,
"com.delphix:livelist", "livelist",
"Improved clone deletion performance.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN,
livelist_deps, sfeatures);
}
{
static const spa_feature_t log_spacemap_deps[] = {
SPA_FEATURE_SPACEMAP_V2,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_LOG_SPACEMAP,
"com.delphix:log_spacemap", "log_spacemap",
"Log metaslab changes on a single spacemap and "
"flush them periodically.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN,
log_spacemap_deps, sfeatures);
}
{
static const spa_feature_t large_blocks_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_LARGE_BLOCKS,
"org.open-zfs:large_blocks", "large_blocks",
"Support for blocks larger than 128KB.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
large_blocks_deps, sfeatures);
}
{
static const spa_feature_t large_dnode_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_LARGE_DNODE,
"org.zfsonlinux:large_dnode", "large_dnode",
"Variable on-disk size of dnodes.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
large_dnode_deps, sfeatures);
}
{
static const spa_feature_t sha512_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_SHA512,
"org.illumos:sha512", "sha512",
"SHA-512/256 hash algorithm.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
sha512_deps, sfeatures);
}
{
static const spa_feature_t skein_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_SKEIN,
"org.illumos:skein", "skein",
"Skein hash algorithm.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
skein_deps, sfeatures);
}
{
static const spa_feature_t edonr_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_EDONR,
"org.illumos:edonr", "edonr",
"Edon-R hash algorithm.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
edonr_deps, sfeatures);
}
{
static const spa_feature_t redact_books_deps[] = {
SPA_FEATURE_BOOKMARK_V2,
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_BOOKMARKS,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_REDACTION_BOOKMARKS,
"com.delphix:redaction_bookmarks", "redaction_bookmarks",
"Support for bookmarks which store redaction lists for zfs "
"redacted send/recv.", 0, ZFEATURE_TYPE_BOOLEAN,
redact_books_deps, sfeatures);
}
{
static const spa_feature_t redact_datasets_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_REDACTED_DATASETS,
"com.delphix:redacted_datasets", "redacted_datasets",
"Support for redacted datasets, produced by receiving "
"a redacted zfs send stream.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_UINT64_ARRAY,
redact_datasets_deps, sfeatures);
}
{
static const spa_feature_t bookmark_written_deps[] = {
SPA_FEATURE_BOOKMARK_V2,
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_BOOKMARKS,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_BOOKMARK_WRITTEN,
"com.delphix:bookmark_written", "bookmark_written",
"Additional accounting, enabling the written#<bookmark> "
"property (space written since a bookmark), "
"and estimates of send stream sizes for incrementals from "
"bookmarks.",
0, ZFEATURE_TYPE_BOOLEAN, bookmark_written_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_DEVICE_REMOVAL,
"com.delphix:device_removal", "device_removal",
"Top-level vdevs can be removed, reducing logical pool size.",
ZFEATURE_FLAG_MOS, ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
{
static const spa_feature_t obsolete_counts_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_DEVICE_REMOVAL,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_OBSOLETE_COUNTS,
"com.delphix:obsolete_counts", "obsolete_counts",
"Reduce memory used by removed devices when their blocks "
"are freed or remapped.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN,
obsolete_counts_deps, sfeatures);
}
{
static const spa_feature_t userobj_accounting_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_USEROBJ_ACCOUNTING,
"org.zfsonlinux:userobj_accounting", "userobj_accounting",
"User/Group object accounting.",
ZFEATURE_FLAG_READONLY_COMPAT | ZFEATURE_FLAG_PER_DATASET,
ZFEATURE_TYPE_BOOLEAN, userobj_accounting_deps, sfeatures);
}
{
static const spa_feature_t bookmark_v2_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_BOOKMARKS,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_BOOKMARK_V2,
"com.datto:bookmark_v2", "bookmark_v2",
"Support for larger bookmarks",
0, ZFEATURE_TYPE_BOOLEAN, bookmark_v2_deps, sfeatures);
}
{
static const spa_feature_t encryption_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_BOOKMARK_V2,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_ENCRYPTION,
"com.datto:encryption", "encryption",
"Support for dataset level encryption",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
encryption_deps, sfeatures);
}
{
static const spa_feature_t project_quota_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_PROJECT_QUOTA,
"org.zfsonlinux:project_quota", "project_quota",
"space/object accounting based on project ID.",
ZFEATURE_FLAG_READONLY_COMPAT | ZFEATURE_FLAG_PER_DATASET,
ZFEATURE_TYPE_BOOLEAN, project_quota_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_ALLOCATION_CLASSES,
"org.zfsonlinux:allocation_classes", "allocation_classes",
"Support for separate allocation classes.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_RESILVER_DEFER,
"com.datto:resilver_defer", "resilver_defer",
"Support for deferring new resilvers when one is already running.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_DEVICE_REBUILD,
"org.openzfs:device_rebuild", "device_rebuild",
"Support for sequential mirror/dRAID device rebuilds",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
{
static const spa_feature_t zstd_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_ZSTD_COMPRESS,
"org.freebsd:zstd_compress", "zstd_compress",
"zstd compression algorithm support.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN, zstd_deps,
sfeatures);
}
zfeature_register(SPA_FEATURE_DRAID,
"org.openzfs:draid", "draid", "Support for distributed spare RAID",
ZFEATURE_FLAG_MOS, ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
{
static const spa_feature_t zilsaxattr_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_ZILSAXATTR,
"org.openzfs:zilsaxattr", "zilsaxattr",
"Support for xattr=sa extended attribute logging in ZIL.",
ZFEATURE_FLAG_PER_DATASET | ZFEATURE_FLAG_READONLY_COMPAT,
ZFEATURE_TYPE_BOOLEAN, zilsaxattr_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_HEAD_ERRLOG,
"com.delphix:head_errlog", "head_errlog",
"Support for per-dataset on-disk error logs.",
ZFEATURE_FLAG_ACTIVATE_ON_ENABLE, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
{
static const spa_feature_t blake3_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_BLAKE3,
"org.openzfs:blake3", "blake3",
"BLAKE3 hash algorithm.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
blake3_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_BLOCK_CLONING,
"com.fudosecurity:block_cloning", "block_cloning",
"Support for block cloning via Block Reference Table.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_AVZ_V2,
"com.klarasystems:vdev_zaps_v2", "vdev_zaps_v2",
"Support for root vdev ZAP.",
ZFEATURE_FLAG_MOS, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
{
static const spa_feature_t redact_list_spill_deps[] = {
SPA_FEATURE_REDACTION_BOOKMARKS,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_REDACTION_LIST_SPILL,
"com.delphix:redaction_list_spill", "redaction_list_spill",
"Support for increased number of redaction_snapshot "
"arguments in zfs redact.", 0, ZFEATURE_TYPE_BOOLEAN,
redact_list_spill_deps, sfeatures);
}
+ zfeature_register(SPA_FEATURE_RAIDZ_EXPANSION,
+ "org.openzfs:raidz_expansion", "raidz_expansion",
+ "Support for raidz expansion",
+ ZFEATURE_FLAG_MOS, ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
+
zfs_mod_list_supported_free(sfeatures);
}
#if defined(_KERNEL)
EXPORT_SYMBOL(zfeature_lookup_guid);
EXPORT_SYMBOL(zfeature_lookup_name);
EXPORT_SYMBOL(zfeature_is_supported);
EXPORT_SYMBOL(zfeature_is_valid_guid);
EXPORT_SYMBOL(zfeature_depends_on);
EXPORT_SYMBOL(zpool_feature_init);
EXPORT_SYMBOL(spa_feature_table);
#endif
diff --git a/sys/contrib/openzfs/module/zcommon/zpool_prop.c b/sys/contrib/openzfs/module/zcommon/zpool_prop.c
index c4aca04a96bd..e98063e8bc57 100644
--- a/sys/contrib/openzfs/module/zcommon/zpool_prop.c
+++ b/sys/contrib/openzfs/module/zcommon/zpool_prop.c
@@ -1,600 +1,603 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright 2011 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2012, 2018 by Delphix. All rights reserved.
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
* Copyright (c) 2021, Klara Inc.
*/
#include <sys/zio.h>
#include <sys/spa.h>
#include <sys/zfs_acl.h>
#include <sys/zfs_ioctl.h>
#include <sys/fs/zfs.h>
#include "zfs_prop.h"
#if !defined(_KERNEL)
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#endif
static zprop_desc_t zpool_prop_table[ZPOOL_NUM_PROPS];
static zprop_desc_t vdev_prop_table[VDEV_NUM_PROPS];
zprop_desc_t *
zpool_prop_get_table(void)
{
return (zpool_prop_table);
}
void
zpool_prop_init(void)
{
static const zprop_index_t boolean_table[] = {
{ "off", 0},
{ "on", 1},
{ NULL }
};
static const zprop_index_t failuremode_table[] = {
{ "wait", ZIO_FAILURE_MODE_WAIT },
{ "continue", ZIO_FAILURE_MODE_CONTINUE },
{ "panic", ZIO_FAILURE_MODE_PANIC },
{ NULL }
};
struct zfs_mod_supported_features *sfeatures =
zfs_mod_list_supported(ZFS_SYSFS_POOL_PROPERTIES);
/* string properties */
zprop_register_string(ZPOOL_PROP_ALTROOT, "altroot", NULL, PROP_DEFAULT,
ZFS_TYPE_POOL, "<path>", "ALTROOT", sfeatures);
zprop_register_string(ZPOOL_PROP_BOOTFS, "bootfs", NULL, PROP_DEFAULT,
ZFS_TYPE_POOL, "<filesystem>", "BOOTFS", sfeatures);
zprop_register_string(ZPOOL_PROP_CACHEFILE, "cachefile", NULL,
PROP_DEFAULT, ZFS_TYPE_POOL, "<file> | none", "CACHEFILE",
sfeatures);
zprop_register_string(ZPOOL_PROP_COMMENT, "comment", NULL,
PROP_DEFAULT, ZFS_TYPE_POOL, "<comment-string>", "COMMENT",
sfeatures);
zprop_register_string(ZPOOL_PROP_COMPATIBILITY, "compatibility",
"off", PROP_DEFAULT, ZFS_TYPE_POOL,
"<file[,file...]> | off | legacy", "COMPATIBILITY", sfeatures);
/* readonly number properties */
zprop_register_number(ZPOOL_PROP_SIZE, "size", 0, PROP_READONLY,
ZFS_TYPE_POOL, "<size>", "SIZE", B_FALSE, sfeatures);
zprop_register_number(ZPOOL_PROP_FREE, "free", 0, PROP_READONLY,
ZFS_TYPE_POOL, "<size>", "FREE", B_FALSE, sfeatures);
zprop_register_number(ZPOOL_PROP_FREEING, "freeing", 0, PROP_READONLY,
ZFS_TYPE_POOL, "<size>", "FREEING", B_FALSE, sfeatures);
zprop_register_number(ZPOOL_PROP_CHECKPOINT, "checkpoint", 0,
PROP_READONLY, ZFS_TYPE_POOL, "<size>", "CKPOINT", B_FALSE,
sfeatures);
zprop_register_number(ZPOOL_PROP_LEAKED, "leaked", 0, PROP_READONLY,
ZFS_TYPE_POOL, "<size>", "LEAKED", B_FALSE, sfeatures);
zprop_register_number(ZPOOL_PROP_ALLOCATED, "allocated", 0,
PROP_READONLY, ZFS_TYPE_POOL, "<size>", "ALLOC", B_FALSE,
sfeatures);
zprop_register_number(ZPOOL_PROP_EXPANDSZ, "expandsize", 0,
PROP_READONLY, ZFS_TYPE_POOL, "<size>", "EXPANDSZ", B_FALSE,
sfeatures);
zprop_register_number(ZPOOL_PROP_FRAGMENTATION, "fragmentation", 0,
PROP_READONLY, ZFS_TYPE_POOL, "<percent>", "FRAG", B_FALSE,
sfeatures);
zprop_register_number(ZPOOL_PROP_CAPACITY, "capacity", 0, PROP_READONLY,
ZFS_TYPE_POOL, "<size>", "CAP", B_FALSE, sfeatures);
zprop_register_number(ZPOOL_PROP_GUID, "guid", 0, PROP_READONLY,
ZFS_TYPE_POOL, "<guid>", "GUID", B_TRUE, sfeatures);
zprop_register_number(ZPOOL_PROP_LOAD_GUID, "load_guid", 0,
PROP_READONLY, ZFS_TYPE_POOL, "<load_guid>", "LOAD_GUID",
B_TRUE, sfeatures);
zprop_register_number(ZPOOL_PROP_HEALTH, "health", 0, PROP_READONLY,
ZFS_TYPE_POOL, "<state>", "HEALTH", B_FALSE, sfeatures);
zprop_register_number(ZPOOL_PROP_DEDUPRATIO, "dedupratio", 0,
PROP_READONLY, ZFS_TYPE_POOL, "<1.00x or higher if deduped>",
"DEDUP", B_FALSE, sfeatures);
zprop_register_number(ZPOOL_PROP_BCLONEUSED, "bcloneused", 0,
PROP_READONLY, ZFS_TYPE_POOL, "<size>",
"BCLONE_USED", B_FALSE, sfeatures);
zprop_register_number(ZPOOL_PROP_BCLONESAVED, "bclonesaved", 0,
PROP_READONLY, ZFS_TYPE_POOL, "<size>",
"BCLONE_SAVED", B_FALSE, sfeatures);
zprop_register_number(ZPOOL_PROP_BCLONERATIO, "bcloneratio", 0,
PROP_READONLY, ZFS_TYPE_POOL, "<1.00x or higher if cloned>",
"BCLONE_RATIO", B_FALSE, sfeatures);
/* default number properties */
zprop_register_number(ZPOOL_PROP_VERSION, "version", SPA_VERSION,
PROP_DEFAULT, ZFS_TYPE_POOL, "<version>", "VERSION", B_FALSE,
sfeatures);
zprop_register_number(ZPOOL_PROP_ASHIFT, "ashift", 0, PROP_DEFAULT,
ZFS_TYPE_POOL, "<ashift, 9-16, or 0=default>", "ASHIFT", B_FALSE,
sfeatures);
/* default index (boolean) properties */
zprop_register_index(ZPOOL_PROP_DELEGATION, "delegation", 1,
PROP_DEFAULT, ZFS_TYPE_POOL, "on | off", "DELEGATION",
boolean_table, sfeatures);
zprop_register_index(ZPOOL_PROP_AUTOREPLACE, "autoreplace", 0,
PROP_DEFAULT, ZFS_TYPE_POOL, "on | off", "REPLACE", boolean_table,
sfeatures);
zprop_register_index(ZPOOL_PROP_LISTSNAPS, "listsnapshots", 0,
PROP_DEFAULT, ZFS_TYPE_POOL, "on | off", "LISTSNAPS",
boolean_table, sfeatures);
zprop_register_index(ZPOOL_PROP_AUTOEXPAND, "autoexpand", 0,
PROP_DEFAULT, ZFS_TYPE_POOL, "on | off", "EXPAND", boolean_table,
sfeatures);
zprop_register_index(ZPOOL_PROP_READONLY, "readonly", 0,
PROP_DEFAULT, ZFS_TYPE_POOL, "on | off", "RDONLY", boolean_table,
sfeatures);
zprop_register_index(ZPOOL_PROP_MULTIHOST, "multihost", 0,
PROP_DEFAULT, ZFS_TYPE_POOL, "on | off", "MULTIHOST",
boolean_table, sfeatures);
/* default index properties */
zprop_register_index(ZPOOL_PROP_FAILUREMODE, "failmode",
ZIO_FAILURE_MODE_WAIT, PROP_DEFAULT, ZFS_TYPE_POOL,
"wait | continue | panic", "FAILMODE", failuremode_table,
sfeatures);
zprop_register_index(ZPOOL_PROP_AUTOTRIM, "autotrim",
SPA_AUTOTRIM_OFF, PROP_DEFAULT, ZFS_TYPE_POOL,
"on | off", "AUTOTRIM", boolean_table, sfeatures);
/* hidden properties */
zprop_register_hidden(ZPOOL_PROP_NAME, "name", PROP_TYPE_STRING,
PROP_READONLY, ZFS_TYPE_POOL, "NAME", B_TRUE, sfeatures);
zprop_register_hidden(ZPOOL_PROP_MAXBLOCKSIZE, "maxblocksize",
PROP_TYPE_NUMBER, PROP_READONLY, ZFS_TYPE_POOL, "MAXBLOCKSIZE",
B_FALSE, sfeatures);
zprop_register_hidden(ZPOOL_PROP_TNAME, "tname", PROP_TYPE_STRING,
PROP_ONETIME, ZFS_TYPE_POOL, "TNAME", B_TRUE, sfeatures);
zprop_register_hidden(ZPOOL_PROP_MAXDNODESIZE, "maxdnodesize",
PROP_TYPE_NUMBER, PROP_READONLY, ZFS_TYPE_POOL, "MAXDNODESIZE",
B_FALSE, sfeatures);
zprop_register_hidden(ZPOOL_PROP_DEDUPDITTO, "dedupditto",
PROP_TYPE_NUMBER, PROP_DEFAULT, ZFS_TYPE_POOL, "DEDUPDITTO",
B_FALSE, sfeatures);
zfs_mod_list_supported_free(sfeatures);
}
/*
* Given a property name and its type, returns the corresponding property ID.
*/
zpool_prop_t
zpool_name_to_prop(const char *propname)
{
return (zprop_name_to_prop(propname, ZFS_TYPE_POOL));
}
/*
* Given a pool property ID, returns the corresponding name.
* Assuming the pool property ID is valid.
*/
const char *
zpool_prop_to_name(zpool_prop_t prop)
{
return (zpool_prop_table[prop].pd_name);
}
zprop_type_t
zpool_prop_get_type(zpool_prop_t prop)
{
return (zpool_prop_table[prop].pd_proptype);
}
boolean_t
zpool_prop_readonly(zpool_prop_t prop)
{
return (zpool_prop_table[prop].pd_attr == PROP_READONLY);
}
boolean_t
zpool_prop_setonce(zpool_prop_t prop)
{
return (zpool_prop_table[prop].pd_attr == PROP_ONETIME);
}
const char *
zpool_prop_default_string(zpool_prop_t prop)
{
return (zpool_prop_table[prop].pd_strdefault);
}
uint64_t
zpool_prop_default_numeric(zpool_prop_t prop)
{
return (zpool_prop_table[prop].pd_numdefault);
}
/*
* Returns true if this is a valid feature@ property.
*/
boolean_t
zpool_prop_feature(const char *name)
{
static const char *prefix = "feature@";
return (strncmp(name, prefix, strlen(prefix)) == 0);
}
/*
* Returns true if this is a valid unsupported@ property.
*/
boolean_t
zpool_prop_unsupported(const char *name)
{
static const char *prefix = "unsupported@";
return (strncmp(name, prefix, strlen(prefix)) == 0);
}
int
zpool_prop_string_to_index(zpool_prop_t prop, const char *string,
uint64_t *index)
{
return (zprop_string_to_index(prop, string, index, ZFS_TYPE_POOL));
}
int
zpool_prop_index_to_string(zpool_prop_t prop, uint64_t index,
const char **string)
{
return (zprop_index_to_string(prop, index, string, ZFS_TYPE_POOL));
}
uint64_t
zpool_prop_random_value(zpool_prop_t prop, uint64_t seed)
{
return (zprop_random_value(prop, seed, ZFS_TYPE_POOL));
}
#ifndef _KERNEL
#include <libzfs.h>
const char *
zpool_prop_values(zpool_prop_t prop)
{
return (zpool_prop_table[prop].pd_values);
}
const char *
zpool_prop_column_name(zpool_prop_t prop)
{
return (zpool_prop_table[prop].pd_colname);
}
boolean_t
zpool_prop_align_right(zpool_prop_t prop)
{
return (zpool_prop_table[prop].pd_rightalign);
}
#endif
zprop_desc_t *
vdev_prop_get_table(void)
{
return (vdev_prop_table);
}
void
vdev_prop_init(void)
{
static const zprop_index_t boolean_table[] = {
{ "off", 0},
{ "on", 1},
{ NULL }
};
static const zprop_index_t boolean_na_table[] = {
{ "off", 0},
{ "on", 1},
{ "-", 2}, /* ZPROP_BOOLEAN_NA */
{ NULL }
};
struct zfs_mod_supported_features *sfeatures =
zfs_mod_list_supported(ZFS_SYSFS_VDEV_PROPERTIES);
/* string properties */
zprop_register_string(VDEV_PROP_COMMENT, "comment", NULL,
PROP_DEFAULT, ZFS_TYPE_VDEV, "<comment-string>", "COMMENT",
sfeatures);
zprop_register_string(VDEV_PROP_PATH, "path", NULL,
PROP_DEFAULT, ZFS_TYPE_VDEV, "<device-path>", "PATH", sfeatures);
zprop_register_string(VDEV_PROP_DEVID, "devid", NULL,
PROP_READONLY, ZFS_TYPE_VDEV, "<devid>", "DEVID", sfeatures);
zprop_register_string(VDEV_PROP_PHYS_PATH, "physpath", NULL,
PROP_READONLY, ZFS_TYPE_VDEV, "<physpath>", "PHYSPATH", sfeatures);
zprop_register_string(VDEV_PROP_ENC_PATH, "encpath", NULL,
PROP_READONLY, ZFS_TYPE_VDEV, "<encpath>", "ENCPATH", sfeatures);
zprop_register_string(VDEV_PROP_FRU, "fru", NULL,
PROP_READONLY, ZFS_TYPE_VDEV, "<fru>", "FRU", sfeatures);
zprop_register_string(VDEV_PROP_PARENT, "parent", NULL,
PROP_READONLY, ZFS_TYPE_VDEV, "<parent>", "PARENT", sfeatures);
zprop_register_string(VDEV_PROP_CHILDREN, "children", NULL,
PROP_READONLY, ZFS_TYPE_VDEV, "<child[,...]>", "CHILDREN",
sfeatures);
/* readonly number properties */
zprop_register_number(VDEV_PROP_SIZE, "size", 0, PROP_READONLY,
ZFS_TYPE_VDEV, "<size>", "SIZE", B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_FREE, "free", 0, PROP_READONLY,
ZFS_TYPE_VDEV, "<size>", "FREE", B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_ALLOCATED, "allocated", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<size>", "ALLOC", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_EXPANDSZ, "expandsize", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<size>", "EXPANDSZ", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_FRAGMENTATION, "fragmentation", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<percent>", "FRAG", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_CAPACITY, "capacity", 0, PROP_READONLY,
ZFS_TYPE_VDEV, "<size>", "CAP", B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_GUID, "guid", 0, PROP_READONLY,
ZFS_TYPE_VDEV, "<guid>", "GUID", B_TRUE, sfeatures);
zprop_register_number(VDEV_PROP_STATE, "state", 0, PROP_READONLY,
ZFS_TYPE_VDEV, "<state>", "STATE", B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_BOOTSIZE, "bootsize", 0, PROP_READONLY,
ZFS_TYPE_VDEV, "<size>", "BOOTSIZE", B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_ASIZE, "asize", 0, PROP_READONLY,
ZFS_TYPE_VDEV, "<asize>", "ASIZE", B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_PSIZE, "psize", 0, PROP_READONLY,
ZFS_TYPE_VDEV, "<psize>", "PSIZE", B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_ASHIFT, "ashift", 0, PROP_READONLY,
ZFS_TYPE_VDEV, "<ashift>", "ASHIFT", B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_PARITY, "parity", 0, PROP_READONLY,
ZFS_TYPE_VDEV, "<parity>", "PARITY", B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_NUMCHILDREN, "numchildren", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<number-of-children>", "NUMCHILD",
B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_READ_ERRORS, "read_errors", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<errors>", "RDERR", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_WRITE_ERRORS, "write_errors", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<errors>", "WRERR", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_CHECKSUM_ERRORS, "checksum_errors", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<errors>", "CKERR", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_INITIALIZE_ERRORS,
"initialize_errors", 0, PROP_READONLY, ZFS_TYPE_VDEV, "<errors>",
"INITERR", B_FALSE, sfeatures);
zprop_register_number(VDEV_PROP_OPS_NULL, "null_ops", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<operations>", "NULLOP", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_OPS_READ, "read_ops", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<operations>", "READOP", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_OPS_WRITE, "write_ops", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<operations>", "WRITEOP", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_OPS_FREE, "free_ops", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<operations>", "FREEOP", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_OPS_CLAIM, "claim_ops", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<operations>", "CLAIMOP", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_OPS_TRIM, "trim_ops", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<operations>", "TRIMOP", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_BYTES_NULL, "null_bytes", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<bytes>", "NULLBYTE", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_BYTES_READ, "read_bytes", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<bytes>", "READBYTE", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_BYTES_WRITE, "write_bytes", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<bytes>", "WRITEBYTE", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_BYTES_FREE, "free_bytes", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<bytes>", "FREEBYTE", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_BYTES_CLAIM, "claim_bytes", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<bytes>", "CLAIMBYTE", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_BYTES_TRIM, "trim_bytes", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "<bytes>", "TRIMBYTE", B_FALSE,
sfeatures);
/* default numeric properties */
zprop_register_number(VDEV_PROP_CHECKSUM_N, "checksum_n", UINT64_MAX,
PROP_DEFAULT, ZFS_TYPE_VDEV, "<events>", "CKSUM_N", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_CHECKSUM_T, "checksum_t", UINT64_MAX,
PROP_DEFAULT, ZFS_TYPE_VDEV, "<seconds>", "CKSUM_T", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_IO_N, "io_n", UINT64_MAX,
PROP_DEFAULT, ZFS_TYPE_VDEV, "<events>", "IO_N", B_FALSE,
sfeatures);
zprop_register_number(VDEV_PROP_IO_T, "io_t", UINT64_MAX,
PROP_DEFAULT, ZFS_TYPE_VDEV, "<seconds>", "IO_T", B_FALSE,
sfeatures);
/* default index (boolean) properties */
zprop_register_index(VDEV_PROP_REMOVING, "removing", 0,
PROP_READONLY, ZFS_TYPE_VDEV, "on | off", "REMOVING",
boolean_table, sfeatures);
zprop_register_index(VDEV_PROP_ALLOCATING, "allocating", 1,
PROP_DEFAULT, ZFS_TYPE_VDEV, "on | off", "ALLOCATING",
boolean_na_table, sfeatures);
+ zprop_register_index(VDEV_PROP_RAIDZ_EXPANDING, "raidz_expanding", 0,
+ PROP_READONLY, ZFS_TYPE_VDEV, "on | off", "RAIDZ_EXPANDING",
+ boolean_table, sfeatures);
/* default index properties */
zprop_register_index(VDEV_PROP_FAILFAST, "failfast", B_TRUE,
PROP_DEFAULT, ZFS_TYPE_VDEV, "on | off", "FAILFAST", boolean_table,
sfeatures);
/* hidden properties */
zprop_register_hidden(VDEV_PROP_NAME, "name", PROP_TYPE_STRING,
PROP_READONLY, ZFS_TYPE_VDEV, "NAME", B_TRUE, sfeatures);
zfs_mod_list_supported_free(sfeatures);
}
/*
* Given a property name and its type, returns the corresponding property ID.
*/
vdev_prop_t
vdev_name_to_prop(const char *propname)
{
return (zprop_name_to_prop(propname, ZFS_TYPE_VDEV));
}
/*
* Returns true if this is a valid user-defined property (one with a ':').
*/
boolean_t
vdev_prop_user(const char *name)
{
int i;
char c;
boolean_t foundsep = B_FALSE;
for (i = 0; i < strlen(name); i++) {
c = name[i];
if (!zprop_valid_char(c))
return (B_FALSE);
if (c == ':')
foundsep = B_TRUE;
}
return (foundsep);
}
/*
* Given a pool property ID, returns the corresponding name.
* Assuming the pool property ID is valid.
*/
const char *
vdev_prop_to_name(vdev_prop_t prop)
{
return (vdev_prop_table[prop].pd_name);
}
zprop_type_t
vdev_prop_get_type(vdev_prop_t prop)
{
return (vdev_prop_table[prop].pd_proptype);
}
boolean_t
vdev_prop_readonly(vdev_prop_t prop)
{
return (vdev_prop_table[prop].pd_attr == PROP_READONLY);
}
const char *
vdev_prop_default_string(vdev_prop_t prop)
{
return (vdev_prop_table[prop].pd_strdefault);
}
uint64_t
vdev_prop_default_numeric(vdev_prop_t prop)
{
return (vdev_prop_table[prop].pd_numdefault);
}
int
vdev_prop_string_to_index(vdev_prop_t prop, const char *string,
uint64_t *index)
{
return (zprop_string_to_index(prop, string, index, ZFS_TYPE_VDEV));
}
int
vdev_prop_index_to_string(vdev_prop_t prop, uint64_t index,
const char **string)
{
return (zprop_index_to_string(prop, index, string, ZFS_TYPE_VDEV));
}
/*
* Returns true if this is a valid vdev property.
*/
boolean_t
zpool_prop_vdev(const char *name)
{
return (vdev_name_to_prop(name) != VDEV_PROP_INVAL);
}
uint64_t
vdev_prop_random_value(vdev_prop_t prop, uint64_t seed)
{
return (zprop_random_value(prop, seed, ZFS_TYPE_VDEV));
}
#ifndef _KERNEL
const char *
vdev_prop_values(vdev_prop_t prop)
{
return (vdev_prop_table[prop].pd_values);
}
const char *
vdev_prop_column_name(vdev_prop_t prop)
{
return (vdev_prop_table[prop].pd_colname);
}
boolean_t
vdev_prop_align_right(vdev_prop_t prop)
{
return (vdev_prop_table[prop].pd_rightalign);
}
#endif
#if defined(_KERNEL)
/* zpool property functions */
EXPORT_SYMBOL(zpool_prop_init);
EXPORT_SYMBOL(zpool_prop_get_type);
EXPORT_SYMBOL(zpool_prop_get_table);
/* vdev property functions */
EXPORT_SYMBOL(vdev_prop_init);
EXPORT_SYMBOL(vdev_prop_get_type);
EXPORT_SYMBOL(vdev_prop_get_table);
/* Pool property functions shared between libzfs and kernel. */
EXPORT_SYMBOL(zpool_name_to_prop);
EXPORT_SYMBOL(zpool_prop_to_name);
EXPORT_SYMBOL(zpool_prop_default_string);
EXPORT_SYMBOL(zpool_prop_default_numeric);
EXPORT_SYMBOL(zpool_prop_readonly);
EXPORT_SYMBOL(zpool_prop_feature);
EXPORT_SYMBOL(zpool_prop_unsupported);
EXPORT_SYMBOL(zpool_prop_index_to_string);
EXPORT_SYMBOL(zpool_prop_string_to_index);
EXPORT_SYMBOL(zpool_prop_vdev);
/* vdev property functions shared between libzfs and kernel. */
EXPORT_SYMBOL(vdev_name_to_prop);
EXPORT_SYMBOL(vdev_prop_user);
EXPORT_SYMBOL(vdev_prop_to_name);
EXPORT_SYMBOL(vdev_prop_default_string);
EXPORT_SYMBOL(vdev_prop_default_numeric);
EXPORT_SYMBOL(vdev_prop_readonly);
EXPORT_SYMBOL(vdev_prop_index_to_string);
EXPORT_SYMBOL(vdev_prop_string_to_index);
#endif
diff --git a/sys/contrib/openzfs/module/zfs/arc.c b/sys/contrib/openzfs/module/zfs/arc.c
index dfea15b74394..2d08cc5e7240 100644
--- a/sys/contrib/openzfs/module/zfs/arc.c
+++ b/sys/contrib/openzfs/module/zfs/arc.c
@@ -1,10778 +1,10782 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2018, Joyent, Inc.
* Copyright (c) 2011, 2020, Delphix. All rights reserved.
* Copyright (c) 2014, Saso Kiselkov. All rights reserved.
* Copyright (c) 2017, Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2019, loli10K <ezomori.nozomu@gmail.com>. All rights reserved.
* Copyright (c) 2020, George Amanakis. All rights reserved.
* Copyright (c) 2019, Klara Inc.
* Copyright (c) 2019, Allan Jude
* Copyright (c) 2020, The FreeBSD Foundation [1]
*
* [1] Portions of this software were developed by Allan Jude
* under sponsorship from the FreeBSD Foundation.
*/
/*
* DVA-based Adjustable Replacement Cache
*
* While much of the theory of operation used here is
* based on the self-tuning, low overhead replacement cache
* presented by Megiddo and Modha at FAST 2003, there are some
* significant differences:
*
* 1. The Megiddo and Modha model assumes any page is evictable.
* Pages in its cache cannot be "locked" into memory. This makes
* the eviction algorithm simple: evict the last page in the list.
* This also make the performance characteristics easy to reason
* about. Our cache is not so simple. At any given moment, some
* subset of the blocks in the cache are un-evictable because we
* have handed out a reference to them. Blocks are only evictable
* when there are no external references active. This makes
* eviction far more problematic: we choose to evict the evictable
* blocks that are the "lowest" in the list.
*
* There are times when it is not possible to evict the requested
* space. In these circumstances we are unable to adjust the cache
* size. To prevent the cache growing unbounded at these times we
* implement a "cache throttle" that slows the flow of new data
* into the cache until we can make space available.
*
* 2. The Megiddo and Modha model assumes a fixed cache size.
* Pages are evicted when the cache is full and there is a cache
* miss. Our model has a variable sized cache. It grows with
* high use, but also tries to react to memory pressure from the
* operating system: decreasing its size when system memory is
* tight.
*
* 3. The Megiddo and Modha model assumes a fixed page size. All
* elements of the cache are therefore exactly the same size. So
* when adjusting the cache size following a cache miss, its simply
* a matter of choosing a single page to evict. In our model, we
* have variable sized cache blocks (ranging from 512 bytes to
* 128K bytes). We therefore choose a set of blocks to evict to make
* space for a cache miss that approximates as closely as possible
* the space used by the new block.
*
* See also: "ARC: A Self-Tuning, Low Overhead Replacement Cache"
* by N. Megiddo & D. Modha, FAST 2003
*/
/*
* The locking model:
*
* A new reference to a cache buffer can be obtained in two
* ways: 1) via a hash table lookup using the DVA as a key,
* or 2) via one of the ARC lists. The arc_read() interface
* uses method 1, while the internal ARC algorithms for
* adjusting the cache use method 2. We therefore provide two
* types of locks: 1) the hash table lock array, and 2) the
* ARC list locks.
*
* Buffers do not have their own mutexes, rather they rely on the
* hash table mutexes for the bulk of their protection (i.e. most
* fields in the arc_buf_hdr_t are protected by these mutexes).
*
* buf_hash_find() returns the appropriate mutex (held) when it
* locates the requested buffer in the hash table. It returns
* NULL for the mutex if the buffer was not in the table.
*
* buf_hash_remove() expects the appropriate hash mutex to be
* already held before it is invoked.
*
* Each ARC state also has a mutex which is used to protect the
* buffer list associated with the state. When attempting to
* obtain a hash table lock while holding an ARC list lock you
* must use: mutex_tryenter() to avoid deadlock. Also note that
* the active state mutex must be held before the ghost state mutex.
*
* It as also possible to register a callback which is run when the
* metadata limit is reached and no buffers can be safely evicted. In
* this case the arc user should drop a reference on some arc buffers so
* they can be reclaimed. For example, when using the ZPL each dentry
* holds a references on a znode. These dentries must be pruned before
* the arc buffer holding the znode can be safely evicted.
*
* Note that the majority of the performance stats are manipulated
* with atomic operations.
*
* The L2ARC uses the l2ad_mtx on each vdev for the following:
*
* - L2ARC buflist creation
* - L2ARC buflist eviction
* - L2ARC write completion, which walks L2ARC buflists
* - ARC header destruction, as it removes from L2ARC buflists
* - ARC header release, as it removes from L2ARC buflists
*/
/*
* ARC operation:
*
* Every block that is in the ARC is tracked by an arc_buf_hdr_t structure.
* This structure can point either to a block that is still in the cache or to
* one that is only accessible in an L2 ARC device, or it can provide
* information about a block that was recently evicted. If a block is
* only accessible in the L2ARC, then the arc_buf_hdr_t only has enough
* information to retrieve it from the L2ARC device. This information is
* stored in the l2arc_buf_hdr_t sub-structure of the arc_buf_hdr_t. A block
* that is in this state cannot access the data directly.
*
* Blocks that are actively being referenced or have not been evicted
* are cached in the L1ARC. The L1ARC (l1arc_buf_hdr_t) is a structure within
* the arc_buf_hdr_t that will point to the data block in memory. A block can
* only be read by a consumer if it has an l1arc_buf_hdr_t. The L1ARC
* caches data in two ways -- in a list of ARC buffers (arc_buf_t) and
* also in the arc_buf_hdr_t's private physical data block pointer (b_pabd).
*
* The L1ARC's data pointer may or may not be uncompressed. The ARC has the
* ability to store the physical data (b_pabd) associated with the DVA of the
* arc_buf_hdr_t. Since the b_pabd is a copy of the on-disk physical block,
* it will match its on-disk compression characteristics. This behavior can be
* disabled by setting 'zfs_compressed_arc_enabled' to B_FALSE. When the
* compressed ARC functionality is disabled, the b_pabd will point to an
* uncompressed version of the on-disk data.
*
* Data in the L1ARC is not accessed by consumers of the ARC directly. Each
* arc_buf_hdr_t can have multiple ARC buffers (arc_buf_t) which reference it.
* Each ARC buffer (arc_buf_t) is being actively accessed by a specific ARC
* consumer. The ARC will provide references to this data and will keep it
* cached until it is no longer in use. The ARC caches only the L1ARC's physical
* data block and will evict any arc_buf_t that is no longer referenced. The
* amount of memory consumed by the arc_buf_ts' data buffers can be seen via the
* "overhead_size" kstat.
*
* Depending on the consumer, an arc_buf_t can be requested in uncompressed or
* compressed form. The typical case is that consumers will want uncompressed
* data, and when that happens a new data buffer is allocated where the data is
* decompressed for them to use. Currently the only consumer who wants
* compressed arc_buf_t's is "zfs send", when it streams data exactly as it
* exists on disk. When this happens, the arc_buf_t's data buffer is shared
* with the arc_buf_hdr_t.
*
* Here is a diagram showing an arc_buf_hdr_t referenced by two arc_buf_t's. The
* first one is owned by a compressed send consumer (and therefore references
* the same compressed data buffer as the arc_buf_hdr_t) and the second could be
* used by any other consumer (and has its own uncompressed copy of the data
* buffer).
*
* arc_buf_hdr_t
* +-----------+
* | fields |
* | common to |
* | L1- and |
* | L2ARC |
* +-----------+
* | l2arc_buf_hdr_t
* | |
* +-----------+
* | l1arc_buf_hdr_t
* | | arc_buf_t
* | b_buf +------------>+-----------+ arc_buf_t
* | b_pabd +-+ |b_next +---->+-----------+
* +-----------+ | |-----------| |b_next +-->NULL
* | |b_comp = T | +-----------+
* | |b_data +-+ |b_comp = F |
* | +-----------+ | |b_data +-+
* +->+------+ | +-----------+ |
* compressed | | | |
* data | |<--------------+ | uncompressed
* +------+ compressed, | data
* shared +-->+------+
* data | |
* | |
* +------+
*
* When a consumer reads a block, the ARC must first look to see if the
* arc_buf_hdr_t is cached. If the hdr is cached then the ARC allocates a new
* arc_buf_t and either copies uncompressed data into a new data buffer from an
* existing uncompressed arc_buf_t, decompresses the hdr's b_pabd buffer into a
* new data buffer, or shares the hdr's b_pabd buffer, depending on whether the
* hdr is compressed and the desired compression characteristics of the
* arc_buf_t consumer. If the arc_buf_t ends up sharing data with the
* arc_buf_hdr_t and both of them are uncompressed then the arc_buf_t must be
* the last buffer in the hdr's b_buf list, however a shared compressed buf can
* be anywhere in the hdr's list.
*
* The diagram below shows an example of an uncompressed ARC hdr that is
* sharing its data with an arc_buf_t (note that the shared uncompressed buf is
* the last element in the buf list):
*
* arc_buf_hdr_t
* +-----------+
* | |
* | |
* | |
* +-----------+
* l2arc_buf_hdr_t| |
* | |
* +-----------+
* l1arc_buf_hdr_t| |
* | | arc_buf_t (shared)
* | b_buf +------------>+---------+ arc_buf_t
* | | |b_next +---->+---------+
* | b_pabd +-+ |---------| |b_next +-->NULL
* +-----------+ | | | +---------+
* | |b_data +-+ | |
* | +---------+ | |b_data +-+
* +->+------+ | +---------+ |
* | | | |
* uncompressed | | | |
* data +------+ | |
* ^ +->+------+ |
* | uncompressed | | |
* | data | | |
* | +------+ |
* +---------------------------------+
*
* Writing to the ARC requires that the ARC first discard the hdr's b_pabd
* since the physical block is about to be rewritten. The new data contents
* will be contained in the arc_buf_t. As the I/O pipeline performs the write,
* it may compress the data before writing it to disk. The ARC will be called
* with the transformed data and will memcpy the transformed on-disk block into
* a newly allocated b_pabd. Writes are always done into buffers which have
* either been loaned (and hence are new and don't have other readers) or
* buffers which have been released (and hence have their own hdr, if there
* were originally other readers of the buf's original hdr). This ensures that
* the ARC only needs to update a single buf and its hdr after a write occurs.
*
* When the L2ARC is in use, it will also take advantage of the b_pabd. The
* L2ARC will always write the contents of b_pabd to the L2ARC. This means
* that when compressed ARC is enabled that the L2ARC blocks are identical
* to the on-disk block in the main data pool. This provides a significant
* advantage since the ARC can leverage the bp's checksum when reading from the
* L2ARC to determine if the contents are valid. However, if the compressed
* ARC is disabled, then the L2ARC's block must be transformed to look
* like the physical block in the main data pool before comparing the
* checksum and determining its validity.
*
* The L1ARC has a slightly different system for storing encrypted data.
* Raw (encrypted + possibly compressed) data has a few subtle differences from
* data that is just compressed. The biggest difference is that it is not
* possible to decrypt encrypted data (or vice-versa) if the keys aren't loaded.
* The other difference is that encryption cannot be treated as a suggestion.
* If a caller would prefer compressed data, but they actually wind up with
* uncompressed data the worst thing that could happen is there might be a
* performance hit. If the caller requests encrypted data, however, we must be
* sure they actually get it or else secret information could be leaked. Raw
* data is stored in hdr->b_crypt_hdr.b_rabd. An encrypted header, therefore,
* may have both an encrypted version and a decrypted version of its data at
* once. When a caller needs a raw arc_buf_t, it is allocated and the data is
* copied out of this header. To avoid complications with b_pabd, raw buffers
* cannot be shared.
*/
#include <sys/spa.h>
#include <sys/zio.h>
#include <sys/spa_impl.h>
#include <sys/zio_compress.h>
#include <sys/zio_checksum.h>
#include <sys/zfs_context.h>
#include <sys/arc.h>
#include <sys/zfs_refcount.h>
#include <sys/vdev.h>
#include <sys/vdev_impl.h>
#include <sys/dsl_pool.h>
#include <sys/multilist.h>
#include <sys/abd.h>
#include <sys/zil.h>
#include <sys/fm/fs/zfs.h>
#include <sys/callb.h>
#include <sys/kstat.h>
#include <sys/zthr.h>
#include <zfs_fletcher.h>
#include <sys/arc_impl.h>
#include <sys/trace_zfs.h>
#include <sys/aggsum.h>
#include <sys/wmsum.h>
#include <cityhash.h>
#include <sys/vdev_trim.h>
#include <sys/zfs_racct.h>
#include <sys/zstd/zstd.h>
#ifndef _KERNEL
/* set with ZFS_DEBUG=watch, to enable watchpoints on frozen buffers */
boolean_t arc_watch = B_FALSE;
#endif
/*
* This thread's job is to keep enough free memory in the system, by
* calling arc_kmem_reap_soon() plus arc_reduce_target_size(), which improves
* arc_available_memory().
*/
static zthr_t *arc_reap_zthr;
/*
* This thread's job is to keep arc_size under arc_c, by calling
* arc_evict(), which improves arc_is_overflowing().
*/
static zthr_t *arc_evict_zthr;
static arc_buf_hdr_t **arc_state_evict_markers;
static int arc_state_evict_marker_count;
static kmutex_t arc_evict_lock;
static boolean_t arc_evict_needed = B_FALSE;
static clock_t arc_last_uncached_flush;
/*
* Count of bytes evicted since boot.
*/
static uint64_t arc_evict_count;
/*
* List of arc_evict_waiter_t's, representing threads waiting for the
* arc_evict_count to reach specific values.
*/
static list_t arc_evict_waiters;
/*
* When arc_is_overflowing(), arc_get_data_impl() waits for this percent of
* the requested amount of data to be evicted. For example, by default for
* every 2KB that's evicted, 1KB of it may be "reused" by a new allocation.
* Since this is above 100%, it ensures that progress is made towards getting
* arc_size under arc_c. Since this is finite, it ensures that allocations
* can still happen, even during the potentially long time that arc_size is
* more than arc_c.
*/
static uint_t zfs_arc_eviction_pct = 200;
/*
* The number of headers to evict in arc_evict_state_impl() before
* dropping the sublist lock and evicting from another sublist. A lower
* value means we're more likely to evict the "correct" header (i.e. the
* oldest header in the arc state), but comes with higher overhead
* (i.e. more invocations of arc_evict_state_impl()).
*/
static uint_t zfs_arc_evict_batch_limit = 10;
/* number of seconds before growing cache again */
uint_t arc_grow_retry = 5;
/*
* Minimum time between calls to arc_kmem_reap_soon().
*/
static const int arc_kmem_cache_reap_retry_ms = 1000;
/* shift of arc_c for calculating overflow limit in arc_get_data_impl */
static int zfs_arc_overflow_shift = 8;
/* log2(fraction of arc to reclaim) */
uint_t arc_shrink_shift = 7;
/* percent of pagecache to reclaim arc to */
#ifdef _KERNEL
uint_t zfs_arc_pc_percent = 0;
#endif
/*
* log2(fraction of ARC which must be free to allow growing).
* I.e. If there is less than arc_c >> arc_no_grow_shift free memory,
* when reading a new block into the ARC, we will evict an equal-sized block
* from the ARC.
*
* This must be less than arc_shrink_shift, so that when we shrink the ARC,
* we will still not allow it to grow.
*/
uint_t arc_no_grow_shift = 5;
/*
* minimum lifespan of a prefetch block in clock ticks
* (initialized in arc_init())
*/
static uint_t arc_min_prefetch_ms;
static uint_t arc_min_prescient_prefetch_ms;
/*
* If this percent of memory is free, don't throttle.
*/
uint_t arc_lotsfree_percent = 10;
/*
* The arc has filled available memory and has now warmed up.
*/
boolean_t arc_warm;
/*
* These tunables are for performance analysis.
*/
uint64_t zfs_arc_max = 0;
uint64_t zfs_arc_min = 0;
static uint64_t zfs_arc_dnode_limit = 0;
static uint_t zfs_arc_dnode_reduce_percent = 10;
static uint_t zfs_arc_grow_retry = 0;
static uint_t zfs_arc_shrink_shift = 0;
uint_t zfs_arc_average_blocksize = 8 * 1024; /* 8KB */
/*
* ARC dirty data constraints for arc_tempreserve_space() throttle:
* * total dirty data limit
* * anon block dirty limit
* * each pool's anon allowance
*/
static const unsigned long zfs_arc_dirty_limit_percent = 50;
static const unsigned long zfs_arc_anon_limit_percent = 25;
static const unsigned long zfs_arc_pool_dirty_percent = 20;
/*
* Enable or disable compressed arc buffers.
*/
int zfs_compressed_arc_enabled = B_TRUE;
/*
* Balance between metadata and data on ghost hits. Values above 100
* increase metadata caching by proportionally reducing effect of ghost
* data hits on target data/metadata rate.
*/
static uint_t zfs_arc_meta_balance = 500;
/*
* Percentage that can be consumed by dnodes of ARC meta buffers.
*/
static uint_t zfs_arc_dnode_limit_percent = 10;
/*
* These tunables are Linux-specific
*/
static uint64_t zfs_arc_sys_free = 0;
static uint_t zfs_arc_min_prefetch_ms = 0;
static uint_t zfs_arc_min_prescient_prefetch_ms = 0;
static uint_t zfs_arc_lotsfree_percent = 10;
/*
* Number of arc_prune threads
*/
static int zfs_arc_prune_task_threads = 1;
/* The 7 states: */
arc_state_t ARC_anon;
arc_state_t ARC_mru;
arc_state_t ARC_mru_ghost;
arc_state_t ARC_mfu;
arc_state_t ARC_mfu_ghost;
arc_state_t ARC_l2c_only;
arc_state_t ARC_uncached;
arc_stats_t arc_stats = {
{ "hits", KSTAT_DATA_UINT64 },
{ "iohits", KSTAT_DATA_UINT64 },
{ "misses", KSTAT_DATA_UINT64 },
{ "demand_data_hits", KSTAT_DATA_UINT64 },
{ "demand_data_iohits", KSTAT_DATA_UINT64 },
{ "demand_data_misses", KSTAT_DATA_UINT64 },
{ "demand_metadata_hits", KSTAT_DATA_UINT64 },
{ "demand_metadata_iohits", KSTAT_DATA_UINT64 },
{ "demand_metadata_misses", KSTAT_DATA_UINT64 },
{ "prefetch_data_hits", KSTAT_DATA_UINT64 },
{ "prefetch_data_iohits", KSTAT_DATA_UINT64 },
{ "prefetch_data_misses", KSTAT_DATA_UINT64 },
{ "prefetch_metadata_hits", KSTAT_DATA_UINT64 },
{ "prefetch_metadata_iohits", KSTAT_DATA_UINT64 },
{ "prefetch_metadata_misses", KSTAT_DATA_UINT64 },
{ "mru_hits", KSTAT_DATA_UINT64 },
{ "mru_ghost_hits", KSTAT_DATA_UINT64 },
{ "mfu_hits", KSTAT_DATA_UINT64 },
{ "mfu_ghost_hits", KSTAT_DATA_UINT64 },
{ "uncached_hits", KSTAT_DATA_UINT64 },
{ "deleted", KSTAT_DATA_UINT64 },
{ "mutex_miss", KSTAT_DATA_UINT64 },
{ "access_skip", KSTAT_DATA_UINT64 },
{ "evict_skip", KSTAT_DATA_UINT64 },
{ "evict_not_enough", KSTAT_DATA_UINT64 },
{ "evict_l2_cached", KSTAT_DATA_UINT64 },
{ "evict_l2_eligible", KSTAT_DATA_UINT64 },
{ "evict_l2_eligible_mfu", KSTAT_DATA_UINT64 },
{ "evict_l2_eligible_mru", KSTAT_DATA_UINT64 },
{ "evict_l2_ineligible", KSTAT_DATA_UINT64 },
{ "evict_l2_skip", KSTAT_DATA_UINT64 },
{ "hash_elements", KSTAT_DATA_UINT64 },
{ "hash_elements_max", KSTAT_DATA_UINT64 },
{ "hash_collisions", KSTAT_DATA_UINT64 },
{ "hash_chains", KSTAT_DATA_UINT64 },
{ "hash_chain_max", KSTAT_DATA_UINT64 },
{ "meta", KSTAT_DATA_UINT64 },
{ "pd", KSTAT_DATA_UINT64 },
{ "pm", KSTAT_DATA_UINT64 },
{ "c", KSTAT_DATA_UINT64 },
{ "c_min", KSTAT_DATA_UINT64 },
{ "c_max", KSTAT_DATA_UINT64 },
{ "size", KSTAT_DATA_UINT64 },
{ "compressed_size", KSTAT_DATA_UINT64 },
{ "uncompressed_size", KSTAT_DATA_UINT64 },
{ "overhead_size", KSTAT_DATA_UINT64 },
{ "hdr_size", KSTAT_DATA_UINT64 },
{ "data_size", KSTAT_DATA_UINT64 },
{ "metadata_size", KSTAT_DATA_UINT64 },
{ "dbuf_size", KSTAT_DATA_UINT64 },
{ "dnode_size", KSTAT_DATA_UINT64 },
{ "bonus_size", KSTAT_DATA_UINT64 },
#if defined(COMPAT_FREEBSD11)
{ "other_size", KSTAT_DATA_UINT64 },
#endif
{ "anon_size", KSTAT_DATA_UINT64 },
{ "anon_data", KSTAT_DATA_UINT64 },
{ "anon_metadata", KSTAT_DATA_UINT64 },
{ "anon_evictable_data", KSTAT_DATA_UINT64 },
{ "anon_evictable_metadata", KSTAT_DATA_UINT64 },
{ "mru_size", KSTAT_DATA_UINT64 },
{ "mru_data", KSTAT_DATA_UINT64 },
{ "mru_metadata", KSTAT_DATA_UINT64 },
{ "mru_evictable_data", KSTAT_DATA_UINT64 },
{ "mru_evictable_metadata", KSTAT_DATA_UINT64 },
{ "mru_ghost_size", KSTAT_DATA_UINT64 },
{ "mru_ghost_data", KSTAT_DATA_UINT64 },
{ "mru_ghost_metadata", KSTAT_DATA_UINT64 },
{ "mru_ghost_evictable_data", KSTAT_DATA_UINT64 },
{ "mru_ghost_evictable_metadata", KSTAT_DATA_UINT64 },
{ "mfu_size", KSTAT_DATA_UINT64 },
{ "mfu_data", KSTAT_DATA_UINT64 },
{ "mfu_metadata", KSTAT_DATA_UINT64 },
{ "mfu_evictable_data", KSTAT_DATA_UINT64 },
{ "mfu_evictable_metadata", KSTAT_DATA_UINT64 },
{ "mfu_ghost_size", KSTAT_DATA_UINT64 },
{ "mfu_ghost_data", KSTAT_DATA_UINT64 },
{ "mfu_ghost_metadata", KSTAT_DATA_UINT64 },
{ "mfu_ghost_evictable_data", KSTAT_DATA_UINT64 },
{ "mfu_ghost_evictable_metadata", KSTAT_DATA_UINT64 },
{ "uncached_size", KSTAT_DATA_UINT64 },
{ "uncached_data", KSTAT_DATA_UINT64 },
{ "uncached_metadata", KSTAT_DATA_UINT64 },
{ "uncached_evictable_data", KSTAT_DATA_UINT64 },
{ "uncached_evictable_metadata", KSTAT_DATA_UINT64 },
{ "l2_hits", KSTAT_DATA_UINT64 },
{ "l2_misses", KSTAT_DATA_UINT64 },
{ "l2_prefetch_asize", KSTAT_DATA_UINT64 },
{ "l2_mru_asize", KSTAT_DATA_UINT64 },
{ "l2_mfu_asize", KSTAT_DATA_UINT64 },
{ "l2_bufc_data_asize", KSTAT_DATA_UINT64 },
{ "l2_bufc_metadata_asize", KSTAT_DATA_UINT64 },
{ "l2_feeds", KSTAT_DATA_UINT64 },
{ "l2_rw_clash", KSTAT_DATA_UINT64 },
{ "l2_read_bytes", KSTAT_DATA_UINT64 },
{ "l2_write_bytes", KSTAT_DATA_UINT64 },
{ "l2_writes_sent", KSTAT_DATA_UINT64 },
{ "l2_writes_done", KSTAT_DATA_UINT64 },
{ "l2_writes_error", KSTAT_DATA_UINT64 },
{ "l2_writes_lock_retry", KSTAT_DATA_UINT64 },
{ "l2_evict_lock_retry", KSTAT_DATA_UINT64 },
{ "l2_evict_reading", KSTAT_DATA_UINT64 },
{ "l2_evict_l1cached", KSTAT_DATA_UINT64 },
{ "l2_free_on_write", KSTAT_DATA_UINT64 },
{ "l2_abort_lowmem", KSTAT_DATA_UINT64 },
{ "l2_cksum_bad", KSTAT_DATA_UINT64 },
{ "l2_io_error", KSTAT_DATA_UINT64 },
{ "l2_size", KSTAT_DATA_UINT64 },
{ "l2_asize", KSTAT_DATA_UINT64 },
{ "l2_hdr_size", KSTAT_DATA_UINT64 },
{ "l2_log_blk_writes", KSTAT_DATA_UINT64 },
{ "l2_log_blk_avg_asize", KSTAT_DATA_UINT64 },
{ "l2_log_blk_asize", KSTAT_DATA_UINT64 },
{ "l2_log_blk_count", KSTAT_DATA_UINT64 },
{ "l2_data_to_meta_ratio", KSTAT_DATA_UINT64 },
{ "l2_rebuild_success", KSTAT_DATA_UINT64 },
{ "l2_rebuild_unsupported", KSTAT_DATA_UINT64 },
{ "l2_rebuild_io_errors", KSTAT_DATA_UINT64 },
{ "l2_rebuild_dh_errors", KSTAT_DATA_UINT64 },
{ "l2_rebuild_cksum_lb_errors", KSTAT_DATA_UINT64 },
{ "l2_rebuild_lowmem", KSTAT_DATA_UINT64 },
{ "l2_rebuild_size", KSTAT_DATA_UINT64 },
{ "l2_rebuild_asize", KSTAT_DATA_UINT64 },
{ "l2_rebuild_bufs", KSTAT_DATA_UINT64 },
{ "l2_rebuild_bufs_precached", KSTAT_DATA_UINT64 },
{ "l2_rebuild_log_blks", KSTAT_DATA_UINT64 },
{ "memory_throttle_count", KSTAT_DATA_UINT64 },
{ "memory_direct_count", KSTAT_DATA_UINT64 },
{ "memory_indirect_count", KSTAT_DATA_UINT64 },
{ "memory_all_bytes", KSTAT_DATA_UINT64 },
{ "memory_free_bytes", KSTAT_DATA_UINT64 },
{ "memory_available_bytes", KSTAT_DATA_INT64 },
{ "arc_no_grow", KSTAT_DATA_UINT64 },
{ "arc_tempreserve", KSTAT_DATA_UINT64 },
{ "arc_loaned_bytes", KSTAT_DATA_UINT64 },
{ "arc_prune", KSTAT_DATA_UINT64 },
{ "arc_meta_used", KSTAT_DATA_UINT64 },
{ "arc_dnode_limit", KSTAT_DATA_UINT64 },
{ "async_upgrade_sync", KSTAT_DATA_UINT64 },
{ "predictive_prefetch", KSTAT_DATA_UINT64 },
{ "demand_hit_predictive_prefetch", KSTAT_DATA_UINT64 },
{ "demand_iohit_predictive_prefetch", KSTAT_DATA_UINT64 },
{ "prescient_prefetch", KSTAT_DATA_UINT64 },
{ "demand_hit_prescient_prefetch", KSTAT_DATA_UINT64 },
{ "demand_iohit_prescient_prefetch", KSTAT_DATA_UINT64 },
{ "arc_need_free", KSTAT_DATA_UINT64 },
{ "arc_sys_free", KSTAT_DATA_UINT64 },
{ "arc_raw_size", KSTAT_DATA_UINT64 },
{ "cached_only_in_progress", KSTAT_DATA_UINT64 },
{ "abd_chunk_waste_size", KSTAT_DATA_UINT64 },
};
arc_sums_t arc_sums;
#define ARCSTAT_MAX(stat, val) { \
uint64_t m; \
while ((val) > (m = arc_stats.stat.value.ui64) && \
(m != atomic_cas_64(&arc_stats.stat.value.ui64, m, (val)))) \
continue; \
}
/*
* We define a macro to allow ARC hits/misses to be easily broken down by
* two separate conditions, giving a total of four different subtypes for
* each of hits and misses (so eight statistics total).
*/
#define ARCSTAT_CONDSTAT(cond1, stat1, notstat1, cond2, stat2, notstat2, stat) \
if (cond1) { \
if (cond2) { \
ARCSTAT_BUMP(arcstat_##stat1##_##stat2##_##stat); \
} else { \
ARCSTAT_BUMP(arcstat_##stat1##_##notstat2##_##stat); \
} \
} else { \
if (cond2) { \
ARCSTAT_BUMP(arcstat_##notstat1##_##stat2##_##stat); \
} else { \
ARCSTAT_BUMP(arcstat_##notstat1##_##notstat2##_##stat);\
} \
}
/*
* This macro allows us to use kstats as floating averages. Each time we
* update this kstat, we first factor it and the update value by
* ARCSTAT_AVG_FACTOR to shrink the new value's contribution to the overall
* average. This macro assumes that integer loads and stores are atomic, but
* is not safe for multiple writers updating the kstat in parallel (only the
* last writer's update will remain).
*/
#define ARCSTAT_F_AVG_FACTOR 3
#define ARCSTAT_F_AVG(stat, value) \
do { \
uint64_t x = ARCSTAT(stat); \
x = x - x / ARCSTAT_F_AVG_FACTOR + \
(value) / ARCSTAT_F_AVG_FACTOR; \
ARCSTAT(stat) = x; \
} while (0)
static kstat_t *arc_ksp;
/*
* There are several ARC variables that are critical to export as kstats --
* but we don't want to have to grovel around in the kstat whenever we wish to
* manipulate them. For these variables, we therefore define them to be in
* terms of the statistic variable. This assures that we are not introducing
* the possibility of inconsistency by having shadow copies of the variables,
* while still allowing the code to be readable.
*/
#define arc_tempreserve ARCSTAT(arcstat_tempreserve)
#define arc_loaned_bytes ARCSTAT(arcstat_loaned_bytes)
#define arc_dnode_limit ARCSTAT(arcstat_dnode_limit) /* max size for dnodes */
#define arc_need_free ARCSTAT(arcstat_need_free) /* waiting to be evicted */
hrtime_t arc_growtime;
list_t arc_prune_list;
kmutex_t arc_prune_mtx;
taskq_t *arc_prune_taskq;
#define GHOST_STATE(state) \
((state) == arc_mru_ghost || (state) == arc_mfu_ghost || \
(state) == arc_l2c_only)
#define HDR_IN_HASH_TABLE(hdr) ((hdr)->b_flags & ARC_FLAG_IN_HASH_TABLE)
#define HDR_IO_IN_PROGRESS(hdr) ((hdr)->b_flags & ARC_FLAG_IO_IN_PROGRESS)
#define HDR_IO_ERROR(hdr) ((hdr)->b_flags & ARC_FLAG_IO_ERROR)
#define HDR_PREFETCH(hdr) ((hdr)->b_flags & ARC_FLAG_PREFETCH)
#define HDR_PRESCIENT_PREFETCH(hdr) \
((hdr)->b_flags & ARC_FLAG_PRESCIENT_PREFETCH)
#define HDR_COMPRESSION_ENABLED(hdr) \
((hdr)->b_flags & ARC_FLAG_COMPRESSED_ARC)
#define HDR_L2CACHE(hdr) ((hdr)->b_flags & ARC_FLAG_L2CACHE)
#define HDR_UNCACHED(hdr) ((hdr)->b_flags & ARC_FLAG_UNCACHED)
#define HDR_L2_READING(hdr) \
(((hdr)->b_flags & ARC_FLAG_IO_IN_PROGRESS) && \
((hdr)->b_flags & ARC_FLAG_HAS_L2HDR))
#define HDR_L2_WRITING(hdr) ((hdr)->b_flags & ARC_FLAG_L2_WRITING)
#define HDR_L2_EVICTED(hdr) ((hdr)->b_flags & ARC_FLAG_L2_EVICTED)
#define HDR_L2_WRITE_HEAD(hdr) ((hdr)->b_flags & ARC_FLAG_L2_WRITE_HEAD)
#define HDR_PROTECTED(hdr) ((hdr)->b_flags & ARC_FLAG_PROTECTED)
#define HDR_NOAUTH(hdr) ((hdr)->b_flags & ARC_FLAG_NOAUTH)
#define HDR_SHARED_DATA(hdr) ((hdr)->b_flags & ARC_FLAG_SHARED_DATA)
#define HDR_ISTYPE_METADATA(hdr) \
((hdr)->b_flags & ARC_FLAG_BUFC_METADATA)
#define HDR_ISTYPE_DATA(hdr) (!HDR_ISTYPE_METADATA(hdr))
#define HDR_HAS_L1HDR(hdr) ((hdr)->b_flags & ARC_FLAG_HAS_L1HDR)
#define HDR_HAS_L2HDR(hdr) ((hdr)->b_flags & ARC_FLAG_HAS_L2HDR)
#define HDR_HAS_RABD(hdr) \
(HDR_HAS_L1HDR(hdr) && HDR_PROTECTED(hdr) && \
(hdr)->b_crypt_hdr.b_rabd != NULL)
#define HDR_ENCRYPTED(hdr) \
(HDR_PROTECTED(hdr) && DMU_OT_IS_ENCRYPTED((hdr)->b_crypt_hdr.b_ot))
#define HDR_AUTHENTICATED(hdr) \
(HDR_PROTECTED(hdr) && !DMU_OT_IS_ENCRYPTED((hdr)->b_crypt_hdr.b_ot))
/* For storing compression mode in b_flags */
#define HDR_COMPRESS_OFFSET (highbit64(ARC_FLAG_COMPRESS_0) - 1)
#define HDR_GET_COMPRESS(hdr) ((enum zio_compress)BF32_GET((hdr)->b_flags, \
HDR_COMPRESS_OFFSET, SPA_COMPRESSBITS))
#define HDR_SET_COMPRESS(hdr, cmp) BF32_SET((hdr)->b_flags, \
HDR_COMPRESS_OFFSET, SPA_COMPRESSBITS, (cmp));
#define ARC_BUF_LAST(buf) ((buf)->b_next == NULL)
#define ARC_BUF_SHARED(buf) ((buf)->b_flags & ARC_BUF_FLAG_SHARED)
#define ARC_BUF_COMPRESSED(buf) ((buf)->b_flags & ARC_BUF_FLAG_COMPRESSED)
#define ARC_BUF_ENCRYPTED(buf) ((buf)->b_flags & ARC_BUF_FLAG_ENCRYPTED)
/*
* Other sizes
*/
#define HDR_FULL_SIZE ((int64_t)sizeof (arc_buf_hdr_t))
#define HDR_L2ONLY_SIZE ((int64_t)offsetof(arc_buf_hdr_t, b_l1hdr))
/*
* Hash table routines
*/
#define BUF_LOCKS 2048
typedef struct buf_hash_table {
uint64_t ht_mask;
arc_buf_hdr_t **ht_table;
kmutex_t ht_locks[BUF_LOCKS] ____cacheline_aligned;
} buf_hash_table_t;
static buf_hash_table_t buf_hash_table;
#define BUF_HASH_INDEX(spa, dva, birth) \
(buf_hash(spa, dva, birth) & buf_hash_table.ht_mask)
#define BUF_HASH_LOCK(idx) (&buf_hash_table.ht_locks[idx & (BUF_LOCKS-1)])
#define HDR_LOCK(hdr) \
(BUF_HASH_LOCK(BUF_HASH_INDEX(hdr->b_spa, &hdr->b_dva, hdr->b_birth)))
uint64_t zfs_crc64_table[256];
/*
* Level 2 ARC
*/
-#define L2ARC_WRITE_SIZE (8 * 1024 * 1024) /* initial write max */
-#define L2ARC_HEADROOM 2 /* num of writes */
+#define L2ARC_WRITE_SIZE (32 * 1024 * 1024) /* initial write max */
+#define L2ARC_HEADROOM 8 /* num of writes */
/*
* If we discover during ARC scan any buffers to be compressed, we boost
* our headroom for the next scanning cycle by this percentage multiple.
*/
#define L2ARC_HEADROOM_BOOST 200
#define L2ARC_FEED_SECS 1 /* caching interval secs */
#define L2ARC_FEED_MIN_MS 200 /* min caching interval ms */
/*
* We can feed L2ARC from two states of ARC buffers, mru and mfu,
* and each of the state has two types: data and metadata.
*/
#define L2ARC_FEED_TYPES 4
/* L2ARC Performance Tunables */
uint64_t l2arc_write_max = L2ARC_WRITE_SIZE; /* def max write size */
uint64_t l2arc_write_boost = L2ARC_WRITE_SIZE; /* extra warmup write */
uint64_t l2arc_headroom = L2ARC_HEADROOM; /* # of dev writes */
uint64_t l2arc_headroom_boost = L2ARC_HEADROOM_BOOST;
uint64_t l2arc_feed_secs = L2ARC_FEED_SECS; /* interval seconds */
uint64_t l2arc_feed_min_ms = L2ARC_FEED_MIN_MS; /* min interval msecs */
int l2arc_noprefetch = B_TRUE; /* don't cache prefetch bufs */
int l2arc_feed_again = B_TRUE; /* turbo warmup */
int l2arc_norw = B_FALSE; /* no reads during writes */
static uint_t l2arc_meta_percent = 33; /* limit on headers size */
/*
* L2ARC Internals
*/
static list_t L2ARC_dev_list; /* device list */
static list_t *l2arc_dev_list; /* device list pointer */
static kmutex_t l2arc_dev_mtx; /* device list mutex */
static l2arc_dev_t *l2arc_dev_last; /* last device used */
static list_t L2ARC_free_on_write; /* free after write buf list */
static list_t *l2arc_free_on_write; /* free after write list ptr */
static kmutex_t l2arc_free_on_write_mtx; /* mutex for list */
static uint64_t l2arc_ndev; /* number of devices */
typedef struct l2arc_read_callback {
arc_buf_hdr_t *l2rcb_hdr; /* read header */
blkptr_t l2rcb_bp; /* original blkptr */
zbookmark_phys_t l2rcb_zb; /* original bookmark */
int l2rcb_flags; /* original flags */
abd_t *l2rcb_abd; /* temporary buffer */
} l2arc_read_callback_t;
typedef struct l2arc_data_free {
/* protected by l2arc_free_on_write_mtx */
abd_t *l2df_abd;
size_t l2df_size;
arc_buf_contents_t l2df_type;
list_node_t l2df_list_node;
} l2arc_data_free_t;
typedef enum arc_fill_flags {
ARC_FILL_LOCKED = 1 << 0, /* hdr lock is held */
ARC_FILL_COMPRESSED = 1 << 1, /* fill with compressed data */
ARC_FILL_ENCRYPTED = 1 << 2, /* fill with encrypted data */
ARC_FILL_NOAUTH = 1 << 3, /* don't attempt to authenticate */
ARC_FILL_IN_PLACE = 1 << 4 /* fill in place (special case) */
} arc_fill_flags_t;
typedef enum arc_ovf_level {
ARC_OVF_NONE, /* ARC within target size. */
ARC_OVF_SOME, /* ARC is slightly overflowed. */
ARC_OVF_SEVERE /* ARC is severely overflowed. */
} arc_ovf_level_t;
static kmutex_t l2arc_feed_thr_lock;
static kcondvar_t l2arc_feed_thr_cv;
static uint8_t l2arc_thread_exit;
static kmutex_t l2arc_rebuild_thr_lock;
static kcondvar_t l2arc_rebuild_thr_cv;
enum arc_hdr_alloc_flags {
ARC_HDR_ALLOC_RDATA = 0x1,
ARC_HDR_USE_RESERVE = 0x4,
ARC_HDR_ALLOC_LINEAR = 0x8,
};
static abd_t *arc_get_data_abd(arc_buf_hdr_t *, uint64_t, const void *, int);
static void *arc_get_data_buf(arc_buf_hdr_t *, uint64_t, const void *);
static void arc_get_data_impl(arc_buf_hdr_t *, uint64_t, const void *, int);
static void arc_free_data_abd(arc_buf_hdr_t *, abd_t *, uint64_t, const void *);
static void arc_free_data_buf(arc_buf_hdr_t *, void *, uint64_t, const void *);
static void arc_free_data_impl(arc_buf_hdr_t *hdr, uint64_t size,
const void *tag);
static void arc_hdr_free_abd(arc_buf_hdr_t *, boolean_t);
static void arc_hdr_alloc_abd(arc_buf_hdr_t *, int);
static void arc_hdr_destroy(arc_buf_hdr_t *);
static void arc_access(arc_buf_hdr_t *, arc_flags_t, boolean_t);
static void arc_buf_watch(arc_buf_t *);
static void arc_change_state(arc_state_t *, arc_buf_hdr_t *);
static arc_buf_contents_t arc_buf_type(arc_buf_hdr_t *);
static uint32_t arc_bufc_to_flags(arc_buf_contents_t);
static inline void arc_hdr_set_flags(arc_buf_hdr_t *hdr, arc_flags_t flags);
static inline void arc_hdr_clear_flags(arc_buf_hdr_t *hdr, arc_flags_t flags);
static boolean_t l2arc_write_eligible(uint64_t, arc_buf_hdr_t *);
static void l2arc_read_done(zio_t *);
static void l2arc_do_free_on_write(void);
static void l2arc_hdr_arcstats_update(arc_buf_hdr_t *hdr, boolean_t incr,
boolean_t state_only);
static void arc_prune_async(uint64_t adjust);
#define l2arc_hdr_arcstats_increment(hdr) \
l2arc_hdr_arcstats_update((hdr), B_TRUE, B_FALSE)
#define l2arc_hdr_arcstats_decrement(hdr) \
l2arc_hdr_arcstats_update((hdr), B_FALSE, B_FALSE)
#define l2arc_hdr_arcstats_increment_state(hdr) \
l2arc_hdr_arcstats_update((hdr), B_TRUE, B_TRUE)
#define l2arc_hdr_arcstats_decrement_state(hdr) \
l2arc_hdr_arcstats_update((hdr), B_FALSE, B_TRUE)
/*
* l2arc_exclude_special : A zfs module parameter that controls whether buffers
* present on special vdevs are eligibile for caching in L2ARC. If
* set to 1, exclude dbufs on special vdevs from being cached to
* L2ARC.
*/
int l2arc_exclude_special = 0;
/*
* l2arc_mfuonly : A ZFS module parameter that controls whether only MFU
* metadata and data are cached from ARC into L2ARC.
*/
static int l2arc_mfuonly = 0;
/*
* L2ARC TRIM
* l2arc_trim_ahead : A ZFS module parameter that controls how much ahead of
* the current write size (l2arc_write_max) we should TRIM if we
* have filled the device. It is defined as a percentage of the
* write size. If set to 100 we trim twice the space required to
* accommodate upcoming writes. A minimum of 64MB will be trimmed.
* It also enables TRIM of the whole L2ARC device upon creation or
* addition to an existing pool or if the header of the device is
* invalid upon importing a pool or onlining a cache device. The
* default is 0, which disables TRIM on L2ARC altogether as it can
* put significant stress on the underlying storage devices. This
* will vary depending of how well the specific device handles
* these commands.
*/
static uint64_t l2arc_trim_ahead = 0;
/*
* Performance tuning of L2ARC persistence:
*
* l2arc_rebuild_enabled : A ZFS module parameter that controls whether adding
* an L2ARC device (either at pool import or later) will attempt
* to rebuild L2ARC buffer contents.
* l2arc_rebuild_blocks_min_l2size : A ZFS module parameter that controls
* whether log blocks are written to the L2ARC device. If the L2ARC
* device is less than 1GB, the amount of data l2arc_evict()
* evicts is significant compared to the amount of restored L2ARC
* data. In this case do not write log blocks in L2ARC in order
* not to waste space.
*/
static int l2arc_rebuild_enabled = B_TRUE;
static uint64_t l2arc_rebuild_blocks_min_l2size = 1024 * 1024 * 1024;
/* L2ARC persistence rebuild control routines. */
void l2arc_rebuild_vdev(vdev_t *vd, boolean_t reopen);
static __attribute__((noreturn)) void l2arc_dev_rebuild_thread(void *arg);
static int l2arc_rebuild(l2arc_dev_t *dev);
/* L2ARC persistence read I/O routines. */
static int l2arc_dev_hdr_read(l2arc_dev_t *dev);
static int l2arc_log_blk_read(l2arc_dev_t *dev,
const l2arc_log_blkptr_t *this_lp, const l2arc_log_blkptr_t *next_lp,
l2arc_log_blk_phys_t *this_lb, l2arc_log_blk_phys_t *next_lb,
zio_t *this_io, zio_t **next_io);
static zio_t *l2arc_log_blk_fetch(vdev_t *vd,
const l2arc_log_blkptr_t *lp, l2arc_log_blk_phys_t *lb);
static void l2arc_log_blk_fetch_abort(zio_t *zio);
/* L2ARC persistence block restoration routines. */
static void l2arc_log_blk_restore(l2arc_dev_t *dev,
const l2arc_log_blk_phys_t *lb, uint64_t lb_asize);
static void l2arc_hdr_restore(const l2arc_log_ent_phys_t *le,
l2arc_dev_t *dev);
/* L2ARC persistence write I/O routines. */
static uint64_t l2arc_log_blk_commit(l2arc_dev_t *dev, zio_t *pio,
l2arc_write_callback_t *cb);
/* L2ARC persistence auxiliary routines. */
boolean_t l2arc_log_blkptr_valid(l2arc_dev_t *dev,
const l2arc_log_blkptr_t *lbp);
static boolean_t l2arc_log_blk_insert(l2arc_dev_t *dev,
const arc_buf_hdr_t *ab);
boolean_t l2arc_range_check_overlap(uint64_t bottom,
uint64_t top, uint64_t check);
static void l2arc_blk_fetch_done(zio_t *zio);
static inline uint64_t
l2arc_log_blk_overhead(uint64_t write_sz, l2arc_dev_t *dev);
/*
* We use Cityhash for this. It's fast, and has good hash properties without
* requiring any large static buffers.
*/
static uint64_t
buf_hash(uint64_t spa, const dva_t *dva, uint64_t birth)
{
return (cityhash4(spa, dva->dva_word[0], dva->dva_word[1], birth));
}
#define HDR_EMPTY(hdr) \
((hdr)->b_dva.dva_word[0] == 0 && \
(hdr)->b_dva.dva_word[1] == 0)
#define HDR_EMPTY_OR_LOCKED(hdr) \
(HDR_EMPTY(hdr) || MUTEX_HELD(HDR_LOCK(hdr)))
#define HDR_EQUAL(spa, dva, birth, hdr) \
((hdr)->b_dva.dva_word[0] == (dva)->dva_word[0]) && \
((hdr)->b_dva.dva_word[1] == (dva)->dva_word[1]) && \
((hdr)->b_birth == birth) && ((hdr)->b_spa == spa)
static void
buf_discard_identity(arc_buf_hdr_t *hdr)
{
hdr->b_dva.dva_word[0] = 0;
hdr->b_dva.dva_word[1] = 0;
hdr->b_birth = 0;
}
static arc_buf_hdr_t *
buf_hash_find(uint64_t spa, const blkptr_t *bp, kmutex_t **lockp)
{
const dva_t *dva = BP_IDENTITY(bp);
uint64_t birth = BP_PHYSICAL_BIRTH(bp);
uint64_t idx = BUF_HASH_INDEX(spa, dva, birth);
kmutex_t *hash_lock = BUF_HASH_LOCK(idx);
arc_buf_hdr_t *hdr;
mutex_enter(hash_lock);
for (hdr = buf_hash_table.ht_table[idx]; hdr != NULL;
hdr = hdr->b_hash_next) {
if (HDR_EQUAL(spa, dva, birth, hdr)) {
*lockp = hash_lock;
return (hdr);
}
}
mutex_exit(hash_lock);
*lockp = NULL;
return (NULL);
}
/*
* Insert an entry into the hash table. If there is already an element
* equal to elem in the hash table, then the already existing element
* will be returned and the new element will not be inserted.
* Otherwise returns NULL.
* If lockp == NULL, the caller is assumed to already hold the hash lock.
*/
static arc_buf_hdr_t *
buf_hash_insert(arc_buf_hdr_t *hdr, kmutex_t **lockp)
{
uint64_t idx = BUF_HASH_INDEX(hdr->b_spa, &hdr->b_dva, hdr->b_birth);
kmutex_t *hash_lock = BUF_HASH_LOCK(idx);
arc_buf_hdr_t *fhdr;
uint32_t i;
ASSERT(!DVA_IS_EMPTY(&hdr->b_dva));
ASSERT(hdr->b_birth != 0);
ASSERT(!HDR_IN_HASH_TABLE(hdr));
if (lockp != NULL) {
*lockp = hash_lock;
mutex_enter(hash_lock);
} else {
ASSERT(MUTEX_HELD(hash_lock));
}
for (fhdr = buf_hash_table.ht_table[idx], i = 0; fhdr != NULL;
fhdr = fhdr->b_hash_next, i++) {
if (HDR_EQUAL(hdr->b_spa, &hdr->b_dva, hdr->b_birth, fhdr))
return (fhdr);
}
hdr->b_hash_next = buf_hash_table.ht_table[idx];
buf_hash_table.ht_table[idx] = hdr;
arc_hdr_set_flags(hdr, ARC_FLAG_IN_HASH_TABLE);
/* collect some hash table performance data */
if (i > 0) {
ARCSTAT_BUMP(arcstat_hash_collisions);
if (i == 1)
ARCSTAT_BUMP(arcstat_hash_chains);
ARCSTAT_MAX(arcstat_hash_chain_max, i);
}
uint64_t he = atomic_inc_64_nv(
&arc_stats.arcstat_hash_elements.value.ui64);
ARCSTAT_MAX(arcstat_hash_elements_max, he);
return (NULL);
}
static void
buf_hash_remove(arc_buf_hdr_t *hdr)
{
arc_buf_hdr_t *fhdr, **hdrp;
uint64_t idx = BUF_HASH_INDEX(hdr->b_spa, &hdr->b_dva, hdr->b_birth);
ASSERT(MUTEX_HELD(BUF_HASH_LOCK(idx)));
ASSERT(HDR_IN_HASH_TABLE(hdr));
hdrp = &buf_hash_table.ht_table[idx];
while ((fhdr = *hdrp) != hdr) {
ASSERT3P(fhdr, !=, NULL);
hdrp = &fhdr->b_hash_next;
}
*hdrp = hdr->b_hash_next;
hdr->b_hash_next = NULL;
arc_hdr_clear_flags(hdr, ARC_FLAG_IN_HASH_TABLE);
/* collect some hash table performance data */
atomic_dec_64(&arc_stats.arcstat_hash_elements.value.ui64);
if (buf_hash_table.ht_table[idx] &&
buf_hash_table.ht_table[idx]->b_hash_next == NULL)
ARCSTAT_BUMPDOWN(arcstat_hash_chains);
}
/*
* Global data structures and functions for the buf kmem cache.
*/
static kmem_cache_t *hdr_full_cache;
static kmem_cache_t *hdr_l2only_cache;
static kmem_cache_t *buf_cache;
static void
buf_fini(void)
{
#if defined(_KERNEL)
/*
* Large allocations which do not require contiguous pages
* should be using vmem_free() in the linux kernel\
*/
vmem_free(buf_hash_table.ht_table,
(buf_hash_table.ht_mask + 1) * sizeof (void *));
#else
kmem_free(buf_hash_table.ht_table,
(buf_hash_table.ht_mask + 1) * sizeof (void *));
#endif
for (int i = 0; i < BUF_LOCKS; i++)
mutex_destroy(BUF_HASH_LOCK(i));
kmem_cache_destroy(hdr_full_cache);
kmem_cache_destroy(hdr_l2only_cache);
kmem_cache_destroy(buf_cache);
}
/*
* Constructor callback - called when the cache is empty
* and a new buf is requested.
*/
static int
hdr_full_cons(void *vbuf, void *unused, int kmflag)
{
(void) unused, (void) kmflag;
arc_buf_hdr_t *hdr = vbuf;
memset(hdr, 0, HDR_FULL_SIZE);
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_NUMFUNCS;
zfs_refcount_create(&hdr->b_l1hdr.b_refcnt);
#ifdef ZFS_DEBUG
mutex_init(&hdr->b_l1hdr.b_freeze_lock, NULL, MUTEX_DEFAULT, NULL);
#endif
multilist_link_init(&hdr->b_l1hdr.b_arc_node);
list_link_init(&hdr->b_l2hdr.b_l2node);
arc_space_consume(HDR_FULL_SIZE, ARC_SPACE_HDRS);
return (0);
}
static int
hdr_l2only_cons(void *vbuf, void *unused, int kmflag)
{
(void) unused, (void) kmflag;
arc_buf_hdr_t *hdr = vbuf;
memset(hdr, 0, HDR_L2ONLY_SIZE);
arc_space_consume(HDR_L2ONLY_SIZE, ARC_SPACE_L2HDRS);
return (0);
}
static int
buf_cons(void *vbuf, void *unused, int kmflag)
{
(void) unused, (void) kmflag;
arc_buf_t *buf = vbuf;
memset(buf, 0, sizeof (arc_buf_t));
arc_space_consume(sizeof (arc_buf_t), ARC_SPACE_HDRS);
return (0);
}
/*
* Destructor callback - called when a cached buf is
* no longer required.
*/
static void
hdr_full_dest(void *vbuf, void *unused)
{
(void) unused;
arc_buf_hdr_t *hdr = vbuf;
ASSERT(HDR_EMPTY(hdr));
zfs_refcount_destroy(&hdr->b_l1hdr.b_refcnt);
#ifdef ZFS_DEBUG
mutex_destroy(&hdr->b_l1hdr.b_freeze_lock);
#endif
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
arc_space_return(HDR_FULL_SIZE, ARC_SPACE_HDRS);
}
static void
hdr_l2only_dest(void *vbuf, void *unused)
{
(void) unused;
arc_buf_hdr_t *hdr = vbuf;
ASSERT(HDR_EMPTY(hdr));
arc_space_return(HDR_L2ONLY_SIZE, ARC_SPACE_L2HDRS);
}
static void
buf_dest(void *vbuf, void *unused)
{
(void) unused;
(void) vbuf;
arc_space_return(sizeof (arc_buf_t), ARC_SPACE_HDRS);
}
static void
buf_init(void)
{
uint64_t *ct = NULL;
uint64_t hsize = 1ULL << 12;
int i, j;
/*
* The hash table is big enough to fill all of physical memory
* with an average block size of zfs_arc_average_blocksize (default 8K).
* By default, the table will take up
* totalmem * sizeof(void*) / 8K (1MB per GB with 8-byte pointers).
*/
while (hsize * zfs_arc_average_blocksize < arc_all_memory())
hsize <<= 1;
retry:
buf_hash_table.ht_mask = hsize - 1;
#if defined(_KERNEL)
/*
* Large allocations which do not require contiguous pages
* should be using vmem_alloc() in the linux kernel
*/
buf_hash_table.ht_table =
vmem_zalloc(hsize * sizeof (void*), KM_SLEEP);
#else
buf_hash_table.ht_table =
kmem_zalloc(hsize * sizeof (void*), KM_NOSLEEP);
#endif
if (buf_hash_table.ht_table == NULL) {
ASSERT(hsize > (1ULL << 8));
hsize >>= 1;
goto retry;
}
hdr_full_cache = kmem_cache_create("arc_buf_hdr_t_full", HDR_FULL_SIZE,
0, hdr_full_cons, hdr_full_dest, NULL, NULL, NULL, 0);
hdr_l2only_cache = kmem_cache_create("arc_buf_hdr_t_l2only",
HDR_L2ONLY_SIZE, 0, hdr_l2only_cons, hdr_l2only_dest, NULL,
NULL, NULL, 0);
buf_cache = kmem_cache_create("arc_buf_t", sizeof (arc_buf_t),
0, buf_cons, buf_dest, NULL, NULL, NULL, 0);
for (i = 0; i < 256; i++)
for (ct = zfs_crc64_table + i, *ct = i, j = 8; j > 0; j--)
*ct = (*ct >> 1) ^ (-(*ct & 1) & ZFS_CRC64_POLY);
for (i = 0; i < BUF_LOCKS; i++)
mutex_init(BUF_HASH_LOCK(i), NULL, MUTEX_DEFAULT, NULL);
}
#define ARC_MINTIME (hz>>4) /* 62 ms */
/*
* This is the size that the buf occupies in memory. If the buf is compressed,
* it will correspond to the compressed size. You should use this method of
* getting the buf size unless you explicitly need the logical size.
*/
uint64_t
arc_buf_size(arc_buf_t *buf)
{
return (ARC_BUF_COMPRESSED(buf) ?
HDR_GET_PSIZE(buf->b_hdr) : HDR_GET_LSIZE(buf->b_hdr));
}
uint64_t
arc_buf_lsize(arc_buf_t *buf)
{
return (HDR_GET_LSIZE(buf->b_hdr));
}
/*
* This function will return B_TRUE if the buffer is encrypted in memory.
* This buffer can be decrypted by calling arc_untransform().
*/
boolean_t
arc_is_encrypted(arc_buf_t *buf)
{
return (ARC_BUF_ENCRYPTED(buf) != 0);
}
/*
* Returns B_TRUE if the buffer represents data that has not had its MAC
* verified yet.
*/
boolean_t
arc_is_unauthenticated(arc_buf_t *buf)
{
return (HDR_NOAUTH(buf->b_hdr) != 0);
}
void
arc_get_raw_params(arc_buf_t *buf, boolean_t *byteorder, uint8_t *salt,
uint8_t *iv, uint8_t *mac)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT(HDR_PROTECTED(hdr));
memcpy(salt, hdr->b_crypt_hdr.b_salt, ZIO_DATA_SALT_LEN);
memcpy(iv, hdr->b_crypt_hdr.b_iv, ZIO_DATA_IV_LEN);
memcpy(mac, hdr->b_crypt_hdr.b_mac, ZIO_DATA_MAC_LEN);
*byteorder = (hdr->b_l1hdr.b_byteswap == DMU_BSWAP_NUMFUNCS) ?
ZFS_HOST_BYTEORDER : !ZFS_HOST_BYTEORDER;
}
/*
* Indicates how this buffer is compressed in memory. If it is not compressed
* the value will be ZIO_COMPRESS_OFF. It can be made normally readable with
* arc_untransform() as long as it is also unencrypted.
*/
enum zio_compress
arc_get_compression(arc_buf_t *buf)
{
return (ARC_BUF_COMPRESSED(buf) ?
HDR_GET_COMPRESS(buf->b_hdr) : ZIO_COMPRESS_OFF);
}
/*
* Return the compression algorithm used to store this data in the ARC. If ARC
* compression is enabled or this is an encrypted block, this will be the same
* as what's used to store it on-disk. Otherwise, this will be ZIO_COMPRESS_OFF.
*/
static inline enum zio_compress
arc_hdr_get_compress(arc_buf_hdr_t *hdr)
{
return (HDR_COMPRESSION_ENABLED(hdr) ?
HDR_GET_COMPRESS(hdr) : ZIO_COMPRESS_OFF);
}
uint8_t
arc_get_complevel(arc_buf_t *buf)
{
return (buf->b_hdr->b_complevel);
}
static inline boolean_t
arc_buf_is_shared(arc_buf_t *buf)
{
boolean_t shared = (buf->b_data != NULL &&
buf->b_hdr->b_l1hdr.b_pabd != NULL &&
abd_is_linear(buf->b_hdr->b_l1hdr.b_pabd) &&
buf->b_data == abd_to_buf(buf->b_hdr->b_l1hdr.b_pabd));
IMPLY(shared, HDR_SHARED_DATA(buf->b_hdr));
EQUIV(shared, ARC_BUF_SHARED(buf));
IMPLY(shared, ARC_BUF_COMPRESSED(buf) || ARC_BUF_LAST(buf));
/*
* It would be nice to assert arc_can_share() too, but the "hdr isn't
* already being shared" requirement prevents us from doing that.
*/
return (shared);
}
/*
* Free the checksum associated with this header. If there is no checksum, this
* is a no-op.
*/
static inline void
arc_cksum_free(arc_buf_hdr_t *hdr)
{
#ifdef ZFS_DEBUG
ASSERT(HDR_HAS_L1HDR(hdr));
mutex_enter(&hdr->b_l1hdr.b_freeze_lock);
if (hdr->b_l1hdr.b_freeze_cksum != NULL) {
kmem_free(hdr->b_l1hdr.b_freeze_cksum, sizeof (zio_cksum_t));
hdr->b_l1hdr.b_freeze_cksum = NULL;
}
mutex_exit(&hdr->b_l1hdr.b_freeze_lock);
#endif
}
/*
* Return true iff at least one of the bufs on hdr is not compressed.
* Encrypted buffers count as compressed.
*/
static boolean_t
arc_hdr_has_uncompressed_buf(arc_buf_hdr_t *hdr)
{
ASSERT(hdr->b_l1hdr.b_state == arc_anon || HDR_EMPTY_OR_LOCKED(hdr));
for (arc_buf_t *b = hdr->b_l1hdr.b_buf; b != NULL; b = b->b_next) {
if (!ARC_BUF_COMPRESSED(b)) {
return (B_TRUE);
}
}
return (B_FALSE);
}
/*
* If we've turned on the ZFS_DEBUG_MODIFY flag, verify that the buf's data
* matches the checksum that is stored in the hdr. If there is no checksum,
* or if the buf is compressed, this is a no-op.
*/
static void
arc_cksum_verify(arc_buf_t *buf)
{
#ifdef ZFS_DEBUG
arc_buf_hdr_t *hdr = buf->b_hdr;
zio_cksum_t zc;
if (!(zfs_flags & ZFS_DEBUG_MODIFY))
return;
if (ARC_BUF_COMPRESSED(buf))
return;
ASSERT(HDR_HAS_L1HDR(hdr));
mutex_enter(&hdr->b_l1hdr.b_freeze_lock);
if (hdr->b_l1hdr.b_freeze_cksum == NULL || HDR_IO_ERROR(hdr)) {
mutex_exit(&hdr->b_l1hdr.b_freeze_lock);
return;
}
fletcher_2_native(buf->b_data, arc_buf_size(buf), NULL, &zc);
if (!ZIO_CHECKSUM_EQUAL(*hdr->b_l1hdr.b_freeze_cksum, zc))
panic("buffer modified while frozen!");
mutex_exit(&hdr->b_l1hdr.b_freeze_lock);
#endif
}
/*
* This function makes the assumption that data stored in the L2ARC
* will be transformed exactly as it is in the main pool. Because of
* this we can verify the checksum against the reading process's bp.
*/
static boolean_t
arc_cksum_is_equal(arc_buf_hdr_t *hdr, zio_t *zio)
{
ASSERT(!BP_IS_EMBEDDED(zio->io_bp));
VERIFY3U(BP_GET_PSIZE(zio->io_bp), ==, HDR_GET_PSIZE(hdr));
/*
* Block pointers always store the checksum for the logical data.
* If the block pointer has the gang bit set, then the checksum
* it represents is for the reconstituted data and not for an
* individual gang member. The zio pipeline, however, must be able to
* determine the checksum of each of the gang constituents so it
* treats the checksum comparison differently than what we need
* for l2arc blocks. This prevents us from using the
* zio_checksum_error() interface directly. Instead we must call the
* zio_checksum_error_impl() so that we can ensure the checksum is
* generated using the correct checksum algorithm and accounts for the
* logical I/O size and not just a gang fragment.
*/
return (zio_checksum_error_impl(zio->io_spa, zio->io_bp,
BP_GET_CHECKSUM(zio->io_bp), zio->io_abd, zio->io_size,
zio->io_offset, NULL) == 0);
}
/*
* Given a buf full of data, if ZFS_DEBUG_MODIFY is enabled this computes a
* checksum and attaches it to the buf's hdr so that we can ensure that the buf
* isn't modified later on. If buf is compressed or there is already a checksum
* on the hdr, this is a no-op (we only checksum uncompressed bufs).
*/
static void
arc_cksum_compute(arc_buf_t *buf)
{
if (!(zfs_flags & ZFS_DEBUG_MODIFY))
return;
#ifdef ZFS_DEBUG
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT(HDR_HAS_L1HDR(hdr));
mutex_enter(&hdr->b_l1hdr.b_freeze_lock);
if (hdr->b_l1hdr.b_freeze_cksum != NULL || ARC_BUF_COMPRESSED(buf)) {
mutex_exit(&hdr->b_l1hdr.b_freeze_lock);
return;
}
ASSERT(!ARC_BUF_ENCRYPTED(buf));
ASSERT(!ARC_BUF_COMPRESSED(buf));
hdr->b_l1hdr.b_freeze_cksum = kmem_alloc(sizeof (zio_cksum_t),
KM_SLEEP);
fletcher_2_native(buf->b_data, arc_buf_size(buf), NULL,
hdr->b_l1hdr.b_freeze_cksum);
mutex_exit(&hdr->b_l1hdr.b_freeze_lock);
#endif
arc_buf_watch(buf);
}
#ifndef _KERNEL
void
arc_buf_sigsegv(int sig, siginfo_t *si, void *unused)
{
(void) sig, (void) unused;
panic("Got SIGSEGV at address: 0x%lx\n", (long)si->si_addr);
}
#endif
static void
arc_buf_unwatch(arc_buf_t *buf)
{
#ifndef _KERNEL
if (arc_watch) {
ASSERT0(mprotect(buf->b_data, arc_buf_size(buf),
PROT_READ | PROT_WRITE));
}
#else
(void) buf;
#endif
}
static void
arc_buf_watch(arc_buf_t *buf)
{
#ifndef _KERNEL
if (arc_watch)
ASSERT0(mprotect(buf->b_data, arc_buf_size(buf),
PROT_READ));
#else
(void) buf;
#endif
}
static arc_buf_contents_t
arc_buf_type(arc_buf_hdr_t *hdr)
{
arc_buf_contents_t type;
if (HDR_ISTYPE_METADATA(hdr)) {
type = ARC_BUFC_METADATA;
} else {
type = ARC_BUFC_DATA;
}
VERIFY3U(hdr->b_type, ==, type);
return (type);
}
boolean_t
arc_is_metadata(arc_buf_t *buf)
{
return (HDR_ISTYPE_METADATA(buf->b_hdr) != 0);
}
static uint32_t
arc_bufc_to_flags(arc_buf_contents_t type)
{
switch (type) {
case ARC_BUFC_DATA:
/* metadata field is 0 if buffer contains normal data */
return (0);
case ARC_BUFC_METADATA:
return (ARC_FLAG_BUFC_METADATA);
default:
break;
}
panic("undefined ARC buffer type!");
return ((uint32_t)-1);
}
void
arc_buf_thaw(arc_buf_t *buf)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT3P(hdr->b_l1hdr.b_state, ==, arc_anon);
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
arc_cksum_verify(buf);
/*
* Compressed buffers do not manipulate the b_freeze_cksum.
*/
if (ARC_BUF_COMPRESSED(buf))
return;
ASSERT(HDR_HAS_L1HDR(hdr));
arc_cksum_free(hdr);
arc_buf_unwatch(buf);
}
void
arc_buf_freeze(arc_buf_t *buf)
{
if (!(zfs_flags & ZFS_DEBUG_MODIFY))
return;
if (ARC_BUF_COMPRESSED(buf))
return;
ASSERT(HDR_HAS_L1HDR(buf->b_hdr));
arc_cksum_compute(buf);
}
/*
* The arc_buf_hdr_t's b_flags should never be modified directly. Instead,
* the following functions should be used to ensure that the flags are
* updated in a thread-safe way. When manipulating the flags either
* the hash_lock must be held or the hdr must be undiscoverable. This
* ensures that we're not racing with any other threads when updating
* the flags.
*/
static inline void
arc_hdr_set_flags(arc_buf_hdr_t *hdr, arc_flags_t flags)
{
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
hdr->b_flags |= flags;
}
static inline void
arc_hdr_clear_flags(arc_buf_hdr_t *hdr, arc_flags_t flags)
{
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
hdr->b_flags &= ~flags;
}
/*
* Setting the compression bits in the arc_buf_hdr_t's b_flags is
* done in a special way since we have to clear and set bits
* at the same time. Consumers that wish to set the compression bits
* must use this function to ensure that the flags are updated in
* thread-safe manner.
*/
static void
arc_hdr_set_compress(arc_buf_hdr_t *hdr, enum zio_compress cmp)
{
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
/*
* Holes and embedded blocks will always have a psize = 0 so
* we ignore the compression of the blkptr and set the
* want to uncompress them. Mark them as uncompressed.
*/
if (!zfs_compressed_arc_enabled || HDR_GET_PSIZE(hdr) == 0) {
arc_hdr_clear_flags(hdr, ARC_FLAG_COMPRESSED_ARC);
ASSERT(!HDR_COMPRESSION_ENABLED(hdr));
} else {
arc_hdr_set_flags(hdr, ARC_FLAG_COMPRESSED_ARC);
ASSERT(HDR_COMPRESSION_ENABLED(hdr));
}
HDR_SET_COMPRESS(hdr, cmp);
ASSERT3U(HDR_GET_COMPRESS(hdr), ==, cmp);
}
/*
* Looks for another buf on the same hdr which has the data decompressed, copies
* from it, and returns true. If no such buf exists, returns false.
*/
static boolean_t
arc_buf_try_copy_decompressed_data(arc_buf_t *buf)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
boolean_t copied = B_FALSE;
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT3P(buf->b_data, !=, NULL);
ASSERT(!ARC_BUF_COMPRESSED(buf));
for (arc_buf_t *from = hdr->b_l1hdr.b_buf; from != NULL;
from = from->b_next) {
/* can't use our own data buffer */
if (from == buf) {
continue;
}
if (!ARC_BUF_COMPRESSED(from)) {
memcpy(buf->b_data, from->b_data, arc_buf_size(buf));
copied = B_TRUE;
break;
}
}
#ifdef ZFS_DEBUG
/*
* There were no decompressed bufs, so there should not be a
* checksum on the hdr either.
*/
if (zfs_flags & ZFS_DEBUG_MODIFY)
EQUIV(!copied, hdr->b_l1hdr.b_freeze_cksum == NULL);
#endif
return (copied);
}
/*
* Allocates an ARC buf header that's in an evicted & L2-cached state.
* This is used during l2arc reconstruction to make empty ARC buffers
* which circumvent the regular disk->arc->l2arc path and instead come
* into being in the reverse order, i.e. l2arc->arc.
*/
static arc_buf_hdr_t *
arc_buf_alloc_l2only(size_t size, arc_buf_contents_t type, l2arc_dev_t *dev,
dva_t dva, uint64_t daddr, int32_t psize, uint64_t birth,
enum zio_compress compress, uint8_t complevel, boolean_t protected,
boolean_t prefetch, arc_state_type_t arcs_state)
{
arc_buf_hdr_t *hdr;
ASSERT(size != 0);
hdr = kmem_cache_alloc(hdr_l2only_cache, KM_SLEEP);
hdr->b_birth = birth;
hdr->b_type = type;
hdr->b_flags = 0;
arc_hdr_set_flags(hdr, arc_bufc_to_flags(type) | ARC_FLAG_HAS_L2HDR);
HDR_SET_LSIZE(hdr, size);
HDR_SET_PSIZE(hdr, psize);
arc_hdr_set_compress(hdr, compress);
hdr->b_complevel = complevel;
if (protected)
arc_hdr_set_flags(hdr, ARC_FLAG_PROTECTED);
if (prefetch)
arc_hdr_set_flags(hdr, ARC_FLAG_PREFETCH);
hdr->b_spa = spa_load_guid(dev->l2ad_vdev->vdev_spa);
hdr->b_dva = dva;
hdr->b_l2hdr.b_dev = dev;
hdr->b_l2hdr.b_daddr = daddr;
hdr->b_l2hdr.b_arcs_state = arcs_state;
return (hdr);
}
/*
* Return the size of the block, b_pabd, that is stored in the arc_buf_hdr_t.
*/
static uint64_t
arc_hdr_size(arc_buf_hdr_t *hdr)
{
uint64_t size;
if (arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF &&
HDR_GET_PSIZE(hdr) > 0) {
size = HDR_GET_PSIZE(hdr);
} else {
ASSERT3U(HDR_GET_LSIZE(hdr), !=, 0);
size = HDR_GET_LSIZE(hdr);
}
return (size);
}
static int
arc_hdr_authenticate(arc_buf_hdr_t *hdr, spa_t *spa, uint64_t dsobj)
{
int ret;
uint64_t csize;
uint64_t lsize = HDR_GET_LSIZE(hdr);
uint64_t psize = HDR_GET_PSIZE(hdr);
void *tmpbuf = NULL;
abd_t *abd = hdr->b_l1hdr.b_pabd;
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
ASSERT(HDR_AUTHENTICATED(hdr));
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
/*
* The MAC is calculated on the compressed data that is stored on disk.
* However, if compressed arc is disabled we will only have the
* decompressed data available to us now. Compress it into a temporary
* abd so we can verify the MAC. The performance overhead of this will
* be relatively low, since most objects in an encrypted objset will
* be encrypted (instead of authenticated) anyway.
*/
if (HDR_GET_COMPRESS(hdr) != ZIO_COMPRESS_OFF &&
!HDR_COMPRESSION_ENABLED(hdr)) {
csize = zio_compress_data(HDR_GET_COMPRESS(hdr),
hdr->b_l1hdr.b_pabd, &tmpbuf, lsize, hdr->b_complevel);
ASSERT3P(tmpbuf, !=, NULL);
ASSERT3U(csize, <=, psize);
abd = abd_get_from_buf(tmpbuf, lsize);
abd_take_ownership_of_buf(abd, B_TRUE);
abd_zero_off(abd, csize, psize - csize);
}
/*
* Authentication is best effort. We authenticate whenever the key is
* available. If we succeed we clear ARC_FLAG_NOAUTH.
*/
if (hdr->b_crypt_hdr.b_ot == DMU_OT_OBJSET) {
ASSERT3U(HDR_GET_COMPRESS(hdr), ==, ZIO_COMPRESS_OFF);
ASSERT3U(lsize, ==, psize);
ret = spa_do_crypt_objset_mac_abd(B_FALSE, spa, dsobj, abd,
psize, hdr->b_l1hdr.b_byteswap != DMU_BSWAP_NUMFUNCS);
} else {
ret = spa_do_crypt_mac_abd(B_FALSE, spa, dsobj, abd, psize,
hdr->b_crypt_hdr.b_mac);
}
if (ret == 0)
arc_hdr_clear_flags(hdr, ARC_FLAG_NOAUTH);
else if (ret != ENOENT)
goto error;
if (tmpbuf != NULL)
abd_free(abd);
return (0);
error:
if (tmpbuf != NULL)
abd_free(abd);
return (ret);
}
/*
* This function will take a header that only has raw encrypted data in
* b_crypt_hdr.b_rabd and decrypt it into a new buffer which is stored in
* b_l1hdr.b_pabd. If designated in the header flags, this function will
* also decompress the data.
*/
static int
arc_hdr_decrypt(arc_buf_hdr_t *hdr, spa_t *spa, const zbookmark_phys_t *zb)
{
int ret;
abd_t *cabd = NULL;
void *tmp = NULL;
boolean_t no_crypt = B_FALSE;
boolean_t bswap = (hdr->b_l1hdr.b_byteswap != DMU_BSWAP_NUMFUNCS);
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
ASSERT(HDR_ENCRYPTED(hdr));
arc_hdr_alloc_abd(hdr, 0);
ret = spa_do_crypt_abd(B_FALSE, spa, zb, hdr->b_crypt_hdr.b_ot,
B_FALSE, bswap, hdr->b_crypt_hdr.b_salt, hdr->b_crypt_hdr.b_iv,
hdr->b_crypt_hdr.b_mac, HDR_GET_PSIZE(hdr), hdr->b_l1hdr.b_pabd,
hdr->b_crypt_hdr.b_rabd, &no_crypt);
if (ret != 0)
goto error;
if (no_crypt) {
abd_copy(hdr->b_l1hdr.b_pabd, hdr->b_crypt_hdr.b_rabd,
HDR_GET_PSIZE(hdr));
}
/*
* If this header has disabled arc compression but the b_pabd is
* compressed after decrypting it, we need to decompress the newly
* decrypted data.
*/
if (HDR_GET_COMPRESS(hdr) != ZIO_COMPRESS_OFF &&
!HDR_COMPRESSION_ENABLED(hdr)) {
/*
* We want to make sure that we are correctly honoring the
* zfs_abd_scatter_enabled setting, so we allocate an abd here
* and then loan a buffer from it, rather than allocating a
* linear buffer and wrapping it in an abd later.
*/
cabd = arc_get_data_abd(hdr, arc_hdr_size(hdr), hdr, 0);
tmp = abd_borrow_buf(cabd, arc_hdr_size(hdr));
ret = zio_decompress_data(HDR_GET_COMPRESS(hdr),
hdr->b_l1hdr.b_pabd, tmp, HDR_GET_PSIZE(hdr),
HDR_GET_LSIZE(hdr), &hdr->b_complevel);
if (ret != 0) {
abd_return_buf(cabd, tmp, arc_hdr_size(hdr));
goto error;
}
abd_return_buf_copy(cabd, tmp, arc_hdr_size(hdr));
arc_free_data_abd(hdr, hdr->b_l1hdr.b_pabd,
arc_hdr_size(hdr), hdr);
hdr->b_l1hdr.b_pabd = cabd;
}
return (0);
error:
arc_hdr_free_abd(hdr, B_FALSE);
if (cabd != NULL)
arc_free_data_buf(hdr, cabd, arc_hdr_size(hdr), hdr);
return (ret);
}
/*
* This function is called during arc_buf_fill() to prepare the header's
* abd plaintext pointer for use. This involves authenticated protected
* data and decrypting encrypted data into the plaintext abd.
*/
static int
arc_fill_hdr_crypt(arc_buf_hdr_t *hdr, kmutex_t *hash_lock, spa_t *spa,
const zbookmark_phys_t *zb, boolean_t noauth)
{
int ret;
ASSERT(HDR_PROTECTED(hdr));
if (hash_lock != NULL)
mutex_enter(hash_lock);
if (HDR_NOAUTH(hdr) && !noauth) {
/*
* The caller requested authenticated data but our data has
* not been authenticated yet. Verify the MAC now if we can.
*/
ret = arc_hdr_authenticate(hdr, spa, zb->zb_objset);
if (ret != 0)
goto error;
} else if (HDR_HAS_RABD(hdr) && hdr->b_l1hdr.b_pabd == NULL) {
/*
* If we only have the encrypted version of the data, but the
* unencrypted version was requested we take this opportunity
* to store the decrypted version in the header for future use.
*/
ret = arc_hdr_decrypt(hdr, spa, zb);
if (ret != 0)
goto error;
}
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
if (hash_lock != NULL)
mutex_exit(hash_lock);
return (0);
error:
if (hash_lock != NULL)
mutex_exit(hash_lock);
return (ret);
}
/*
* This function is used by the dbuf code to decrypt bonus buffers in place.
* The dbuf code itself doesn't have any locking for decrypting a shared dnode
* block, so we use the hash lock here to protect against concurrent calls to
* arc_buf_fill().
*/
static void
arc_buf_untransform_in_place(arc_buf_t *buf)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT(HDR_ENCRYPTED(hdr));
ASSERT3U(hdr->b_crypt_hdr.b_ot, ==, DMU_OT_DNODE);
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
zio_crypt_copy_dnode_bonus(hdr->b_l1hdr.b_pabd, buf->b_data,
arc_buf_size(buf));
buf->b_flags &= ~ARC_BUF_FLAG_ENCRYPTED;
buf->b_flags &= ~ARC_BUF_FLAG_COMPRESSED;
}
/*
* Given a buf that has a data buffer attached to it, this function will
* efficiently fill the buf with data of the specified compression setting from
* the hdr and update the hdr's b_freeze_cksum if necessary. If the buf and hdr
* are already sharing a data buf, no copy is performed.
*
* If the buf is marked as compressed but uncompressed data was requested, this
* will allocate a new data buffer for the buf, remove that flag, and fill the
* buf with uncompressed data. You can't request a compressed buf on a hdr with
* uncompressed data, and (since we haven't added support for it yet) if you
* want compressed data your buf must already be marked as compressed and have
* the correct-sized data buffer.
*/
static int
arc_buf_fill(arc_buf_t *buf, spa_t *spa, const zbookmark_phys_t *zb,
arc_fill_flags_t flags)
{
int error = 0;
arc_buf_hdr_t *hdr = buf->b_hdr;
boolean_t hdr_compressed =
(arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF);
boolean_t compressed = (flags & ARC_FILL_COMPRESSED) != 0;
boolean_t encrypted = (flags & ARC_FILL_ENCRYPTED) != 0;
dmu_object_byteswap_t bswap = hdr->b_l1hdr.b_byteswap;
kmutex_t *hash_lock = (flags & ARC_FILL_LOCKED) ? NULL : HDR_LOCK(hdr);
ASSERT3P(buf->b_data, !=, NULL);
IMPLY(compressed, hdr_compressed || ARC_BUF_ENCRYPTED(buf));
IMPLY(compressed, ARC_BUF_COMPRESSED(buf));
IMPLY(encrypted, HDR_ENCRYPTED(hdr));
IMPLY(encrypted, ARC_BUF_ENCRYPTED(buf));
IMPLY(encrypted, ARC_BUF_COMPRESSED(buf));
IMPLY(encrypted, !arc_buf_is_shared(buf));
/*
* If the caller wanted encrypted data we just need to copy it from
* b_rabd and potentially byteswap it. We won't be able to do any
* further transforms on it.
*/
if (encrypted) {
ASSERT(HDR_HAS_RABD(hdr));
abd_copy_to_buf(buf->b_data, hdr->b_crypt_hdr.b_rabd,
HDR_GET_PSIZE(hdr));
goto byteswap;
}
/*
* Adjust encrypted and authenticated headers to accommodate
* the request if needed. Dnode blocks (ARC_FILL_IN_PLACE) are
* allowed to fail decryption due to keys not being loaded
* without being marked as an IO error.
*/
if (HDR_PROTECTED(hdr)) {
error = arc_fill_hdr_crypt(hdr, hash_lock, spa,
zb, !!(flags & ARC_FILL_NOAUTH));
if (error == EACCES && (flags & ARC_FILL_IN_PLACE) != 0) {
return (error);
} else if (error != 0) {
if (hash_lock != NULL)
mutex_enter(hash_lock);
arc_hdr_set_flags(hdr, ARC_FLAG_IO_ERROR);
if (hash_lock != NULL)
mutex_exit(hash_lock);
return (error);
}
}
/*
* There is a special case here for dnode blocks which are
* decrypting their bonus buffers. These blocks may request to
* be decrypted in-place. This is necessary because there may
* be many dnodes pointing into this buffer and there is
* currently no method to synchronize replacing the backing
* b_data buffer and updating all of the pointers. Here we use
* the hash lock to ensure there are no races. If the need
* arises for other types to be decrypted in-place, they must
* add handling here as well.
*/
if ((flags & ARC_FILL_IN_PLACE) != 0) {
ASSERT(!hdr_compressed);
ASSERT(!compressed);
ASSERT(!encrypted);
if (HDR_ENCRYPTED(hdr) && ARC_BUF_ENCRYPTED(buf)) {
ASSERT3U(hdr->b_crypt_hdr.b_ot, ==, DMU_OT_DNODE);
if (hash_lock != NULL)
mutex_enter(hash_lock);
arc_buf_untransform_in_place(buf);
if (hash_lock != NULL)
mutex_exit(hash_lock);
/* Compute the hdr's checksum if necessary */
arc_cksum_compute(buf);
}
return (0);
}
if (hdr_compressed == compressed) {
if (ARC_BUF_SHARED(buf)) {
ASSERT(arc_buf_is_shared(buf));
} else {
abd_copy_to_buf(buf->b_data, hdr->b_l1hdr.b_pabd,
arc_buf_size(buf));
}
} else {
ASSERT(hdr_compressed);
ASSERT(!compressed);
/*
* If the buf is sharing its data with the hdr, unlink it and
* allocate a new data buffer for the buf.
*/
if (ARC_BUF_SHARED(buf)) {
ASSERT(ARC_BUF_COMPRESSED(buf));
/* We need to give the buf its own b_data */
buf->b_flags &= ~ARC_BUF_FLAG_SHARED;
buf->b_data =
arc_get_data_buf(hdr, HDR_GET_LSIZE(hdr), buf);
arc_hdr_clear_flags(hdr, ARC_FLAG_SHARED_DATA);
/* Previously overhead was 0; just add new overhead */
ARCSTAT_INCR(arcstat_overhead_size, HDR_GET_LSIZE(hdr));
} else if (ARC_BUF_COMPRESSED(buf)) {
ASSERT(!arc_buf_is_shared(buf));
/* We need to reallocate the buf's b_data */
arc_free_data_buf(hdr, buf->b_data, HDR_GET_PSIZE(hdr),
buf);
buf->b_data =
arc_get_data_buf(hdr, HDR_GET_LSIZE(hdr), buf);
/* We increased the size of b_data; update overhead */
ARCSTAT_INCR(arcstat_overhead_size,
HDR_GET_LSIZE(hdr) - HDR_GET_PSIZE(hdr));
}
/*
* Regardless of the buf's previous compression settings, it
* should not be compressed at the end of this function.
*/
buf->b_flags &= ~ARC_BUF_FLAG_COMPRESSED;
/*
* Try copying the data from another buf which already has a
* decompressed version. If that's not possible, it's time to
* bite the bullet and decompress the data from the hdr.
*/
if (arc_buf_try_copy_decompressed_data(buf)) {
/* Skip byteswapping and checksumming (already done) */
return (0);
} else {
error = zio_decompress_data(HDR_GET_COMPRESS(hdr),
hdr->b_l1hdr.b_pabd, buf->b_data,
HDR_GET_PSIZE(hdr), HDR_GET_LSIZE(hdr),
&hdr->b_complevel);
/*
* Absent hardware errors or software bugs, this should
* be impossible, but log it anyway so we can debug it.
*/
if (error != 0) {
zfs_dbgmsg(
"hdr %px, compress %d, psize %d, lsize %d",
hdr, arc_hdr_get_compress(hdr),
HDR_GET_PSIZE(hdr), HDR_GET_LSIZE(hdr));
if (hash_lock != NULL)
mutex_enter(hash_lock);
arc_hdr_set_flags(hdr, ARC_FLAG_IO_ERROR);
if (hash_lock != NULL)
mutex_exit(hash_lock);
return (SET_ERROR(EIO));
}
}
}
byteswap:
/* Byteswap the buf's data if necessary */
if (bswap != DMU_BSWAP_NUMFUNCS) {
ASSERT(!HDR_SHARED_DATA(hdr));
ASSERT3U(bswap, <, DMU_BSWAP_NUMFUNCS);
dmu_ot_byteswap[bswap].ob_func(buf->b_data, HDR_GET_LSIZE(hdr));
}
/* Compute the hdr's checksum if necessary */
arc_cksum_compute(buf);
return (0);
}
/*
* If this function is being called to decrypt an encrypted buffer or verify an
* authenticated one, the key must be loaded and a mapping must be made
* available in the keystore via spa_keystore_create_mapping() or one of its
* callers.
*/
int
arc_untransform(arc_buf_t *buf, spa_t *spa, const zbookmark_phys_t *zb,
boolean_t in_place)
{
int ret;
arc_fill_flags_t flags = 0;
if (in_place)
flags |= ARC_FILL_IN_PLACE;
ret = arc_buf_fill(buf, spa, zb, flags);
if (ret == ECKSUM) {
/*
* Convert authentication and decryption errors to EIO
* (and generate an ereport) before leaving the ARC.
*/
ret = SET_ERROR(EIO);
spa_log_error(spa, zb, &buf->b_hdr->b_birth);
(void) zfs_ereport_post(FM_EREPORT_ZFS_AUTHENTICATION,
spa, NULL, zb, NULL, 0);
}
return (ret);
}
/*
* Increment the amount of evictable space in the arc_state_t's refcount.
* We account for the space used by the hdr and the arc buf individually
* so that we can add and remove them from the refcount individually.
*/
static void
arc_evictable_space_increment(arc_buf_hdr_t *hdr, arc_state_t *state)
{
arc_buf_contents_t type = arc_buf_type(hdr);
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(state)) {
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
(void) zfs_refcount_add_many(&state->arcs_esize[type],
HDR_GET_LSIZE(hdr), hdr);
return;
}
if (hdr->b_l1hdr.b_pabd != NULL) {
(void) zfs_refcount_add_many(&state->arcs_esize[type],
arc_hdr_size(hdr), hdr);
}
if (HDR_HAS_RABD(hdr)) {
(void) zfs_refcount_add_many(&state->arcs_esize[type],
HDR_GET_PSIZE(hdr), hdr);
}
for (arc_buf_t *buf = hdr->b_l1hdr.b_buf; buf != NULL;
buf = buf->b_next) {
if (ARC_BUF_SHARED(buf))
continue;
(void) zfs_refcount_add_many(&state->arcs_esize[type],
arc_buf_size(buf), buf);
}
}
/*
* Decrement the amount of evictable space in the arc_state_t's refcount.
* We account for the space used by the hdr and the arc buf individually
* so that we can add and remove them from the refcount individually.
*/
static void
arc_evictable_space_decrement(arc_buf_hdr_t *hdr, arc_state_t *state)
{
arc_buf_contents_t type = arc_buf_type(hdr);
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(state)) {
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
HDR_GET_LSIZE(hdr), hdr);
return;
}
if (hdr->b_l1hdr.b_pabd != NULL) {
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
arc_hdr_size(hdr), hdr);
}
if (HDR_HAS_RABD(hdr)) {
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
HDR_GET_PSIZE(hdr), hdr);
}
for (arc_buf_t *buf = hdr->b_l1hdr.b_buf; buf != NULL;
buf = buf->b_next) {
if (ARC_BUF_SHARED(buf))
continue;
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
arc_buf_size(buf), buf);
}
}
/*
* Add a reference to this hdr indicating that someone is actively
* referencing that memory. When the refcount transitions from 0 to 1,
* we remove it from the respective arc_state_t list to indicate that
* it is not evictable.
*/
static void
add_reference(arc_buf_hdr_t *hdr, const void *tag)
{
arc_state_t *state = hdr->b_l1hdr.b_state;
ASSERT(HDR_HAS_L1HDR(hdr));
if (!HDR_EMPTY(hdr) && !MUTEX_HELD(HDR_LOCK(hdr))) {
ASSERT(state == arc_anon);
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
}
if ((zfs_refcount_add(&hdr->b_l1hdr.b_refcnt, tag) == 1) &&
state != arc_anon && state != arc_l2c_only) {
/* We don't use the L2-only state list. */
multilist_remove(&state->arcs_list[arc_buf_type(hdr)], hdr);
arc_evictable_space_decrement(hdr, state);
}
}
/*
* Remove a reference from this hdr. When the reference transitions from
* 1 to 0 and we're not anonymous, then we add this hdr to the arc_state_t's
* list making it eligible for eviction.
*/
static int
remove_reference(arc_buf_hdr_t *hdr, const void *tag)
{
int cnt;
arc_state_t *state = hdr->b_l1hdr.b_state;
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(state == arc_anon || MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT(!GHOST_STATE(state)); /* arc_l2c_only counts as a ghost. */
if ((cnt = zfs_refcount_remove(&hdr->b_l1hdr.b_refcnt, tag)) != 0)
return (cnt);
if (state == arc_anon) {
arc_hdr_destroy(hdr);
return (0);
}
if (state == arc_uncached && !HDR_PREFETCH(hdr)) {
arc_change_state(arc_anon, hdr);
arc_hdr_destroy(hdr);
return (0);
}
multilist_insert(&state->arcs_list[arc_buf_type(hdr)], hdr);
arc_evictable_space_increment(hdr, state);
return (0);
}
/*
* Returns detailed information about a specific arc buffer. When the
* state_index argument is set the function will calculate the arc header
* list position for its arc state. Since this requires a linear traversal
* callers are strongly encourage not to do this. However, it can be helpful
* for targeted analysis so the functionality is provided.
*/
void
arc_buf_info(arc_buf_t *ab, arc_buf_info_t *abi, int state_index)
{
(void) state_index;
arc_buf_hdr_t *hdr = ab->b_hdr;
l1arc_buf_hdr_t *l1hdr = NULL;
l2arc_buf_hdr_t *l2hdr = NULL;
arc_state_t *state = NULL;
memset(abi, 0, sizeof (arc_buf_info_t));
if (hdr == NULL)
return;
abi->abi_flags = hdr->b_flags;
if (HDR_HAS_L1HDR(hdr)) {
l1hdr = &hdr->b_l1hdr;
state = l1hdr->b_state;
}
if (HDR_HAS_L2HDR(hdr))
l2hdr = &hdr->b_l2hdr;
if (l1hdr) {
abi->abi_bufcnt = 0;
for (arc_buf_t *buf = l1hdr->b_buf; buf; buf = buf->b_next)
abi->abi_bufcnt++;
abi->abi_access = l1hdr->b_arc_access;
abi->abi_mru_hits = l1hdr->b_mru_hits;
abi->abi_mru_ghost_hits = l1hdr->b_mru_ghost_hits;
abi->abi_mfu_hits = l1hdr->b_mfu_hits;
abi->abi_mfu_ghost_hits = l1hdr->b_mfu_ghost_hits;
abi->abi_holds = zfs_refcount_count(&l1hdr->b_refcnt);
}
if (l2hdr) {
abi->abi_l2arc_dattr = l2hdr->b_daddr;
abi->abi_l2arc_hits = l2hdr->b_hits;
}
abi->abi_state_type = state ? state->arcs_state : ARC_STATE_ANON;
abi->abi_state_contents = arc_buf_type(hdr);
abi->abi_size = arc_hdr_size(hdr);
}
/*
* Move the supplied buffer to the indicated state. The hash lock
* for the buffer must be held by the caller.
*/
static void
arc_change_state(arc_state_t *new_state, arc_buf_hdr_t *hdr)
{
arc_state_t *old_state;
int64_t refcnt;
boolean_t update_old, update_new;
arc_buf_contents_t type = arc_buf_type(hdr);
/*
* We almost always have an L1 hdr here, since we call arc_hdr_realloc()
* in arc_read() when bringing a buffer out of the L2ARC. However, the
* L1 hdr doesn't always exist when we change state to arc_anon before
* destroying a header, in which case reallocating to add the L1 hdr is
* pointless.
*/
if (HDR_HAS_L1HDR(hdr)) {
old_state = hdr->b_l1hdr.b_state;
refcnt = zfs_refcount_count(&hdr->b_l1hdr.b_refcnt);
update_old = (hdr->b_l1hdr.b_buf != NULL ||
hdr->b_l1hdr.b_pabd != NULL || HDR_HAS_RABD(hdr));
IMPLY(GHOST_STATE(old_state), hdr->b_l1hdr.b_buf == NULL);
IMPLY(GHOST_STATE(new_state), hdr->b_l1hdr.b_buf == NULL);
IMPLY(old_state == arc_anon, hdr->b_l1hdr.b_buf == NULL ||
ARC_BUF_LAST(hdr->b_l1hdr.b_buf));
} else {
old_state = arc_l2c_only;
refcnt = 0;
update_old = B_FALSE;
}
update_new = update_old;
if (GHOST_STATE(old_state))
update_old = B_TRUE;
if (GHOST_STATE(new_state))
update_new = B_TRUE;
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT3P(new_state, !=, old_state);
/*
* If this buffer is evictable, transfer it from the
* old state list to the new state list.
*/
if (refcnt == 0) {
if (old_state != arc_anon && old_state != arc_l2c_only) {
ASSERT(HDR_HAS_L1HDR(hdr));
/* remove_reference() saves on insert. */
if (multilist_link_active(&hdr->b_l1hdr.b_arc_node)) {
multilist_remove(&old_state->arcs_list[type],
hdr);
arc_evictable_space_decrement(hdr, old_state);
}
}
if (new_state != arc_anon && new_state != arc_l2c_only) {
/*
* An L1 header always exists here, since if we're
* moving to some L1-cached state (i.e. not l2c_only or
* anonymous), we realloc the header to add an L1hdr
* beforehand.
*/
ASSERT(HDR_HAS_L1HDR(hdr));
multilist_insert(&new_state->arcs_list[type], hdr);
arc_evictable_space_increment(hdr, new_state);
}
}
ASSERT(!HDR_EMPTY(hdr));
if (new_state == arc_anon && HDR_IN_HASH_TABLE(hdr))
buf_hash_remove(hdr);
/* adjust state sizes (ignore arc_l2c_only) */
if (update_new && new_state != arc_l2c_only) {
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(new_state)) {
/*
* When moving a header to a ghost state, we first
* remove all arc buffers. Thus, we'll have no arc
* buffer to use for the reference. As a result, we
* use the arc header pointer for the reference.
*/
(void) zfs_refcount_add_many(
&new_state->arcs_size[type],
HDR_GET_LSIZE(hdr), hdr);
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
} else {
/*
* Each individual buffer holds a unique reference,
* thus we must remove each of these references one
* at a time.
*/
for (arc_buf_t *buf = hdr->b_l1hdr.b_buf; buf != NULL;
buf = buf->b_next) {
/*
* When the arc_buf_t is sharing the data
* block with the hdr, the owner of the
* reference belongs to the hdr. Only
* add to the refcount if the arc_buf_t is
* not shared.
*/
if (ARC_BUF_SHARED(buf))
continue;
(void) zfs_refcount_add_many(
&new_state->arcs_size[type],
arc_buf_size(buf), buf);
}
if (hdr->b_l1hdr.b_pabd != NULL) {
(void) zfs_refcount_add_many(
&new_state->arcs_size[type],
arc_hdr_size(hdr), hdr);
}
if (HDR_HAS_RABD(hdr)) {
(void) zfs_refcount_add_many(
&new_state->arcs_size[type],
HDR_GET_PSIZE(hdr), hdr);
}
}
}
if (update_old && old_state != arc_l2c_only) {
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(old_state)) {
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
/*
* When moving a header off of a ghost state,
* the header will not contain any arc buffers.
* We use the arc header pointer for the reference
* which is exactly what we did when we put the
* header on the ghost state.
*/
(void) zfs_refcount_remove_many(
&old_state->arcs_size[type],
HDR_GET_LSIZE(hdr), hdr);
} else {
/*
* Each individual buffer holds a unique reference,
* thus we must remove each of these references one
* at a time.
*/
for (arc_buf_t *buf = hdr->b_l1hdr.b_buf; buf != NULL;
buf = buf->b_next) {
/*
* When the arc_buf_t is sharing the data
* block with the hdr, the owner of the
* reference belongs to the hdr. Only
* add to the refcount if the arc_buf_t is
* not shared.
*/
if (ARC_BUF_SHARED(buf))
continue;
(void) zfs_refcount_remove_many(
&old_state->arcs_size[type],
arc_buf_size(buf), buf);
}
ASSERT(hdr->b_l1hdr.b_pabd != NULL ||
HDR_HAS_RABD(hdr));
if (hdr->b_l1hdr.b_pabd != NULL) {
(void) zfs_refcount_remove_many(
&old_state->arcs_size[type],
arc_hdr_size(hdr), hdr);
}
if (HDR_HAS_RABD(hdr)) {
(void) zfs_refcount_remove_many(
&old_state->arcs_size[type],
HDR_GET_PSIZE(hdr), hdr);
}
}
}
if (HDR_HAS_L1HDR(hdr)) {
hdr->b_l1hdr.b_state = new_state;
if (HDR_HAS_L2HDR(hdr) && new_state != arc_l2c_only) {
l2arc_hdr_arcstats_decrement_state(hdr);
hdr->b_l2hdr.b_arcs_state = new_state->arcs_state;
l2arc_hdr_arcstats_increment_state(hdr);
}
}
}
void
arc_space_consume(uint64_t space, arc_space_type_t type)
{
ASSERT(type >= 0 && type < ARC_SPACE_NUMTYPES);
switch (type) {
default:
break;
case ARC_SPACE_DATA:
ARCSTAT_INCR(arcstat_data_size, space);
break;
case ARC_SPACE_META:
ARCSTAT_INCR(arcstat_metadata_size, space);
break;
case ARC_SPACE_BONUS:
ARCSTAT_INCR(arcstat_bonus_size, space);
break;
case ARC_SPACE_DNODE:
ARCSTAT_INCR(arcstat_dnode_size, space);
break;
case ARC_SPACE_DBUF:
ARCSTAT_INCR(arcstat_dbuf_size, space);
break;
case ARC_SPACE_HDRS:
ARCSTAT_INCR(arcstat_hdr_size, space);
break;
case ARC_SPACE_L2HDRS:
aggsum_add(&arc_sums.arcstat_l2_hdr_size, space);
break;
case ARC_SPACE_ABD_CHUNK_WASTE:
/*
* Note: this includes space wasted by all scatter ABD's, not
* just those allocated by the ARC. But the vast majority of
* scatter ABD's come from the ARC, because other users are
* very short-lived.
*/
ARCSTAT_INCR(arcstat_abd_chunk_waste_size, space);
break;
}
if (type != ARC_SPACE_DATA && type != ARC_SPACE_ABD_CHUNK_WASTE)
ARCSTAT_INCR(arcstat_meta_used, space);
aggsum_add(&arc_sums.arcstat_size, space);
}
void
arc_space_return(uint64_t space, arc_space_type_t type)
{
ASSERT(type >= 0 && type < ARC_SPACE_NUMTYPES);
switch (type) {
default:
break;
case ARC_SPACE_DATA:
ARCSTAT_INCR(arcstat_data_size, -space);
break;
case ARC_SPACE_META:
ARCSTAT_INCR(arcstat_metadata_size, -space);
break;
case ARC_SPACE_BONUS:
ARCSTAT_INCR(arcstat_bonus_size, -space);
break;
case ARC_SPACE_DNODE:
ARCSTAT_INCR(arcstat_dnode_size, -space);
break;
case ARC_SPACE_DBUF:
ARCSTAT_INCR(arcstat_dbuf_size, -space);
break;
case ARC_SPACE_HDRS:
ARCSTAT_INCR(arcstat_hdr_size, -space);
break;
case ARC_SPACE_L2HDRS:
aggsum_add(&arc_sums.arcstat_l2_hdr_size, -space);
break;
case ARC_SPACE_ABD_CHUNK_WASTE:
ARCSTAT_INCR(arcstat_abd_chunk_waste_size, -space);
break;
}
if (type != ARC_SPACE_DATA && type != ARC_SPACE_ABD_CHUNK_WASTE)
ARCSTAT_INCR(arcstat_meta_used, -space);
ASSERT(aggsum_compare(&arc_sums.arcstat_size, space) >= 0);
aggsum_add(&arc_sums.arcstat_size, -space);
}
/*
* Given a hdr and a buf, returns whether that buf can share its b_data buffer
* with the hdr's b_pabd.
*/
static boolean_t
arc_can_share(arc_buf_hdr_t *hdr, arc_buf_t *buf)
{
/*
* The criteria for sharing a hdr's data are:
* 1. the buffer is not encrypted
* 2. the hdr's compression matches the buf's compression
* 3. the hdr doesn't need to be byteswapped
* 4. the hdr isn't already being shared
* 5. the buf is either compressed or it is the last buf in the hdr list
*
* Criterion #5 maintains the invariant that shared uncompressed
* bufs must be the final buf in the hdr's b_buf list. Reading this, you
* might ask, "if a compressed buf is allocated first, won't that be the
* last thing in the list?", but in that case it's impossible to create
* a shared uncompressed buf anyway (because the hdr must be compressed
* to have the compressed buf). You might also think that #3 is
* sufficient to make this guarantee, however it's possible
* (specifically in the rare L2ARC write race mentioned in
* arc_buf_alloc_impl()) there will be an existing uncompressed buf that
* is shareable, but wasn't at the time of its allocation. Rather than
* allow a new shared uncompressed buf to be created and then shuffle
* the list around to make it the last element, this simply disallows
* sharing if the new buf isn't the first to be added.
*/
ASSERT3P(buf->b_hdr, ==, hdr);
boolean_t hdr_compressed =
arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF;
boolean_t buf_compressed = ARC_BUF_COMPRESSED(buf) != 0;
return (!ARC_BUF_ENCRYPTED(buf) &&
buf_compressed == hdr_compressed &&
hdr->b_l1hdr.b_byteswap == DMU_BSWAP_NUMFUNCS &&
!HDR_SHARED_DATA(hdr) &&
(ARC_BUF_LAST(buf) || ARC_BUF_COMPRESSED(buf)));
}
/*
* Allocate a buf for this hdr. If you care about the data that's in the hdr,
* or if you want a compressed buffer, pass those flags in. Returns 0 if the
* copy was made successfully, or an error code otherwise.
*/
static int
arc_buf_alloc_impl(arc_buf_hdr_t *hdr, spa_t *spa, const zbookmark_phys_t *zb,
const void *tag, boolean_t encrypted, boolean_t compressed,
boolean_t noauth, boolean_t fill, arc_buf_t **ret)
{
arc_buf_t *buf;
arc_fill_flags_t flags = ARC_FILL_LOCKED;
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT3U(HDR_GET_LSIZE(hdr), >, 0);
VERIFY(hdr->b_type == ARC_BUFC_DATA ||
hdr->b_type == ARC_BUFC_METADATA);
ASSERT3P(ret, !=, NULL);
ASSERT3P(*ret, ==, NULL);
IMPLY(encrypted, compressed);
buf = *ret = kmem_cache_alloc(buf_cache, KM_PUSHPAGE);
buf->b_hdr = hdr;
buf->b_data = NULL;
buf->b_next = hdr->b_l1hdr.b_buf;
buf->b_flags = 0;
add_reference(hdr, tag);
/*
* We're about to change the hdr's b_flags. We must either
* hold the hash_lock or be undiscoverable.
*/
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
/*
* Only honor requests for compressed bufs if the hdr is actually
* compressed. This must be overridden if the buffer is encrypted since
* encrypted buffers cannot be decompressed.
*/
if (encrypted) {
buf->b_flags |= ARC_BUF_FLAG_COMPRESSED;
buf->b_flags |= ARC_BUF_FLAG_ENCRYPTED;
flags |= ARC_FILL_COMPRESSED | ARC_FILL_ENCRYPTED;
} else if (compressed &&
arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF) {
buf->b_flags |= ARC_BUF_FLAG_COMPRESSED;
flags |= ARC_FILL_COMPRESSED;
}
if (noauth) {
ASSERT0(encrypted);
flags |= ARC_FILL_NOAUTH;
}
/*
* If the hdr's data can be shared then we share the data buffer and
* set the appropriate bit in the hdr's b_flags to indicate the hdr is
* sharing it's b_pabd with the arc_buf_t. Otherwise, we allocate a new
* buffer to store the buf's data.
*
* There are two additional restrictions here because we're sharing
* hdr -> buf instead of the usual buf -> hdr. First, the hdr can't be
* actively involved in an L2ARC write, because if this buf is used by
* an arc_write() then the hdr's data buffer will be released when the
* write completes, even though the L2ARC write might still be using it.
* Second, the hdr's ABD must be linear so that the buf's user doesn't
* need to be ABD-aware. It must be allocated via
* zio_[data_]buf_alloc(), not as a page, because we need to be able
* to abd_release_ownership_of_buf(), which isn't allowed on "linear
* page" buffers because the ABD code needs to handle freeing them
* specially.
*/
boolean_t can_share = arc_can_share(hdr, buf) &&
!HDR_L2_WRITING(hdr) &&
hdr->b_l1hdr.b_pabd != NULL &&
abd_is_linear(hdr->b_l1hdr.b_pabd) &&
!abd_is_linear_page(hdr->b_l1hdr.b_pabd);
/* Set up b_data and sharing */
if (can_share) {
buf->b_data = abd_to_buf(hdr->b_l1hdr.b_pabd);
buf->b_flags |= ARC_BUF_FLAG_SHARED;
arc_hdr_set_flags(hdr, ARC_FLAG_SHARED_DATA);
} else {
buf->b_data =
arc_get_data_buf(hdr, arc_buf_size(buf), buf);
ARCSTAT_INCR(arcstat_overhead_size, arc_buf_size(buf));
}
VERIFY3P(buf->b_data, !=, NULL);
hdr->b_l1hdr.b_buf = buf;
/*
* If the user wants the data from the hdr, we need to either copy or
* decompress the data.
*/
if (fill) {
ASSERT3P(zb, !=, NULL);
return (arc_buf_fill(buf, spa, zb, flags));
}
return (0);
}
static const char *arc_onloan_tag = "onloan";
static inline void
arc_loaned_bytes_update(int64_t delta)
{
atomic_add_64(&arc_loaned_bytes, delta);
/* assert that it did not wrap around */
ASSERT3S(atomic_add_64_nv(&arc_loaned_bytes, 0), >=, 0);
}
/*
* Loan out an anonymous arc buffer. Loaned buffers are not counted as in
* flight data by arc_tempreserve_space() until they are "returned". Loaned
* buffers must be returned to the arc before they can be used by the DMU or
* freed.
*/
arc_buf_t *
arc_loan_buf(spa_t *spa, boolean_t is_metadata, int size)
{
arc_buf_t *buf = arc_alloc_buf(spa, arc_onloan_tag,
is_metadata ? ARC_BUFC_METADATA : ARC_BUFC_DATA, size);
arc_loaned_bytes_update(arc_buf_size(buf));
return (buf);
}
arc_buf_t *
arc_loan_compressed_buf(spa_t *spa, uint64_t psize, uint64_t lsize,
enum zio_compress compression_type, uint8_t complevel)
{
arc_buf_t *buf = arc_alloc_compressed_buf(spa, arc_onloan_tag,
psize, lsize, compression_type, complevel);
arc_loaned_bytes_update(arc_buf_size(buf));
return (buf);
}
arc_buf_t *
arc_loan_raw_buf(spa_t *spa, uint64_t dsobj, boolean_t byteorder,
const uint8_t *salt, const uint8_t *iv, const uint8_t *mac,
dmu_object_type_t ot, uint64_t psize, uint64_t lsize,
enum zio_compress compression_type, uint8_t complevel)
{
arc_buf_t *buf = arc_alloc_raw_buf(spa, arc_onloan_tag, dsobj,
byteorder, salt, iv, mac, ot, psize, lsize, compression_type,
complevel);
atomic_add_64(&arc_loaned_bytes, psize);
return (buf);
}
/*
* Return a loaned arc buffer to the arc.
*/
void
arc_return_buf(arc_buf_t *buf, const void *tag)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT3P(buf->b_data, !=, NULL);
ASSERT(HDR_HAS_L1HDR(hdr));
(void) zfs_refcount_add(&hdr->b_l1hdr.b_refcnt, tag);
(void) zfs_refcount_remove(&hdr->b_l1hdr.b_refcnt, arc_onloan_tag);
arc_loaned_bytes_update(-arc_buf_size(buf));
}
/* Detach an arc_buf from a dbuf (tag) */
void
arc_loan_inuse_buf(arc_buf_t *buf, const void *tag)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT3P(buf->b_data, !=, NULL);
ASSERT(HDR_HAS_L1HDR(hdr));
(void) zfs_refcount_add(&hdr->b_l1hdr.b_refcnt, arc_onloan_tag);
(void) zfs_refcount_remove(&hdr->b_l1hdr.b_refcnt, tag);
arc_loaned_bytes_update(arc_buf_size(buf));
}
static void
l2arc_free_abd_on_write(abd_t *abd, size_t size, arc_buf_contents_t type)
{
l2arc_data_free_t *df = kmem_alloc(sizeof (*df), KM_SLEEP);
df->l2df_abd = abd;
df->l2df_size = size;
df->l2df_type = type;
mutex_enter(&l2arc_free_on_write_mtx);
list_insert_head(l2arc_free_on_write, df);
mutex_exit(&l2arc_free_on_write_mtx);
}
static void
arc_hdr_free_on_write(arc_buf_hdr_t *hdr, boolean_t free_rdata)
{
arc_state_t *state = hdr->b_l1hdr.b_state;
arc_buf_contents_t type = arc_buf_type(hdr);
uint64_t size = (free_rdata) ? HDR_GET_PSIZE(hdr) : arc_hdr_size(hdr);
/* protected by hash lock, if in the hash table */
if (multilist_link_active(&hdr->b_l1hdr.b_arc_node)) {
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
ASSERT(state != arc_anon && state != arc_l2c_only);
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
size, hdr);
}
(void) zfs_refcount_remove_many(&state->arcs_size[type], size, hdr);
if (type == ARC_BUFC_METADATA) {
arc_space_return(size, ARC_SPACE_META);
} else {
ASSERT(type == ARC_BUFC_DATA);
arc_space_return(size, ARC_SPACE_DATA);
}
if (free_rdata) {
l2arc_free_abd_on_write(hdr->b_crypt_hdr.b_rabd, size, type);
} else {
l2arc_free_abd_on_write(hdr->b_l1hdr.b_pabd, size, type);
}
}
/*
* Share the arc_buf_t's data with the hdr. Whenever we are sharing the
* data buffer, we transfer the refcount ownership to the hdr and update
* the appropriate kstats.
*/
static void
arc_share_buf(arc_buf_hdr_t *hdr, arc_buf_t *buf)
{
ASSERT(arc_can_share(hdr, buf));
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!ARC_BUF_ENCRYPTED(buf));
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
/*
* Start sharing the data buffer. We transfer the
* refcount ownership to the hdr since it always owns
* the refcount whenever an arc_buf_t is shared.
*/
zfs_refcount_transfer_ownership_many(
&hdr->b_l1hdr.b_state->arcs_size[arc_buf_type(hdr)],
arc_hdr_size(hdr), buf, hdr);
hdr->b_l1hdr.b_pabd = abd_get_from_buf(buf->b_data, arc_buf_size(buf));
abd_take_ownership_of_buf(hdr->b_l1hdr.b_pabd,
HDR_ISTYPE_METADATA(hdr));
arc_hdr_set_flags(hdr, ARC_FLAG_SHARED_DATA);
buf->b_flags |= ARC_BUF_FLAG_SHARED;
/*
* Since we've transferred ownership to the hdr we need
* to increment its compressed and uncompressed kstats and
* decrement the overhead size.
*/
ARCSTAT_INCR(arcstat_compressed_size, arc_hdr_size(hdr));
ARCSTAT_INCR(arcstat_uncompressed_size, HDR_GET_LSIZE(hdr));
ARCSTAT_INCR(arcstat_overhead_size, -arc_buf_size(buf));
}
static void
arc_unshare_buf(arc_buf_hdr_t *hdr, arc_buf_t *buf)
{
ASSERT(arc_buf_is_shared(buf));
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
/*
* We are no longer sharing this buffer so we need
* to transfer its ownership to the rightful owner.
*/
zfs_refcount_transfer_ownership_many(
&hdr->b_l1hdr.b_state->arcs_size[arc_buf_type(hdr)],
arc_hdr_size(hdr), hdr, buf);
arc_hdr_clear_flags(hdr, ARC_FLAG_SHARED_DATA);
abd_release_ownership_of_buf(hdr->b_l1hdr.b_pabd);
abd_free(hdr->b_l1hdr.b_pabd);
hdr->b_l1hdr.b_pabd = NULL;
buf->b_flags &= ~ARC_BUF_FLAG_SHARED;
/*
* Since the buffer is no longer shared between
* the arc buf and the hdr, count it as overhead.
*/
ARCSTAT_INCR(arcstat_compressed_size, -arc_hdr_size(hdr));
ARCSTAT_INCR(arcstat_uncompressed_size, -HDR_GET_LSIZE(hdr));
ARCSTAT_INCR(arcstat_overhead_size, arc_buf_size(buf));
}
/*
* Remove an arc_buf_t from the hdr's buf list and return the last
* arc_buf_t on the list. If no buffers remain on the list then return
* NULL.
*/
static arc_buf_t *
arc_buf_remove(arc_buf_hdr_t *hdr, arc_buf_t *buf)
{
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
arc_buf_t **bufp = &hdr->b_l1hdr.b_buf;
arc_buf_t *lastbuf = NULL;
/*
* Remove the buf from the hdr list and locate the last
* remaining buffer on the list.
*/
while (*bufp != NULL) {
if (*bufp == buf)
*bufp = buf->b_next;
/*
* If we've removed a buffer in the middle of
* the list then update the lastbuf and update
* bufp.
*/
if (*bufp != NULL) {
lastbuf = *bufp;
bufp = &(*bufp)->b_next;
}
}
buf->b_next = NULL;
ASSERT3P(lastbuf, !=, buf);
IMPLY(lastbuf != NULL, ARC_BUF_LAST(lastbuf));
return (lastbuf);
}
/*
* Free up buf->b_data and pull the arc_buf_t off of the arc_buf_hdr_t's
* list and free it.
*/
static void
arc_buf_destroy_impl(arc_buf_t *buf)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
/*
* Free up the data associated with the buf but only if we're not
* sharing this with the hdr. If we are sharing it with the hdr, the
* hdr is responsible for doing the free.
*/
if (buf->b_data != NULL) {
/*
* We're about to change the hdr's b_flags. We must either
* hold the hash_lock or be undiscoverable.
*/
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
arc_cksum_verify(buf);
arc_buf_unwatch(buf);
if (ARC_BUF_SHARED(buf)) {
arc_hdr_clear_flags(hdr, ARC_FLAG_SHARED_DATA);
} else {
ASSERT(!arc_buf_is_shared(buf));
uint64_t size = arc_buf_size(buf);
arc_free_data_buf(hdr, buf->b_data, size, buf);
ARCSTAT_INCR(arcstat_overhead_size, -size);
}
buf->b_data = NULL;
/*
* If we have no more encrypted buffers and we've already
* gotten a copy of the decrypted data we can free b_rabd
* to save some space.
*/
if (ARC_BUF_ENCRYPTED(buf) && HDR_HAS_RABD(hdr) &&
hdr->b_l1hdr.b_pabd != NULL && !HDR_IO_IN_PROGRESS(hdr)) {
arc_buf_t *b;
for (b = hdr->b_l1hdr.b_buf; b; b = b->b_next) {
if (b != buf && ARC_BUF_ENCRYPTED(b))
break;
}
if (b == NULL)
arc_hdr_free_abd(hdr, B_TRUE);
}
}
arc_buf_t *lastbuf = arc_buf_remove(hdr, buf);
if (ARC_BUF_SHARED(buf) && !ARC_BUF_COMPRESSED(buf)) {
/*
* If the current arc_buf_t is sharing its data buffer with the
* hdr, then reassign the hdr's b_pabd to share it with the new
* buffer at the end of the list. The shared buffer is always
* the last one on the hdr's buffer list.
*
* There is an equivalent case for compressed bufs, but since
* they aren't guaranteed to be the last buf in the list and
* that is an exceedingly rare case, we just allow that space be
* wasted temporarily. We must also be careful not to share
* encrypted buffers, since they cannot be shared.
*/
if (lastbuf != NULL && !ARC_BUF_ENCRYPTED(lastbuf)) {
/* Only one buf can be shared at once */
ASSERT(!arc_buf_is_shared(lastbuf));
/* hdr is uncompressed so can't have compressed buf */
ASSERT(!ARC_BUF_COMPRESSED(lastbuf));
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
arc_hdr_free_abd(hdr, B_FALSE);
/*
* We must setup a new shared block between the
* last buffer and the hdr. The data would have
* been allocated by the arc buf so we need to transfer
* ownership to the hdr since it's now being shared.
*/
arc_share_buf(hdr, lastbuf);
}
} else if (HDR_SHARED_DATA(hdr)) {
/*
* Uncompressed shared buffers are always at the end
* of the list. Compressed buffers don't have the
* same requirements. This makes it hard to
* simply assert that the lastbuf is shared so
* we rely on the hdr's compression flags to determine
* if we have a compressed, shared buffer.
*/
ASSERT3P(lastbuf, !=, NULL);
ASSERT(arc_buf_is_shared(lastbuf) ||
arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF);
}
/*
* Free the checksum if we're removing the last uncompressed buf from
* this hdr.
*/
if (!arc_hdr_has_uncompressed_buf(hdr)) {
arc_cksum_free(hdr);
}
/* clean up the buf */
buf->b_hdr = NULL;
kmem_cache_free(buf_cache, buf);
}
static void
arc_hdr_alloc_abd(arc_buf_hdr_t *hdr, int alloc_flags)
{
uint64_t size;
boolean_t alloc_rdata = ((alloc_flags & ARC_HDR_ALLOC_RDATA) != 0);
ASSERT3U(HDR_GET_LSIZE(hdr), >, 0);
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(!HDR_SHARED_DATA(hdr) || alloc_rdata);
IMPLY(alloc_rdata, HDR_PROTECTED(hdr));
if (alloc_rdata) {
size = HDR_GET_PSIZE(hdr);
ASSERT3P(hdr->b_crypt_hdr.b_rabd, ==, NULL);
hdr->b_crypt_hdr.b_rabd = arc_get_data_abd(hdr, size, hdr,
alloc_flags);
ASSERT3P(hdr->b_crypt_hdr.b_rabd, !=, NULL);
ARCSTAT_INCR(arcstat_raw_size, size);
} else {
size = arc_hdr_size(hdr);
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
hdr->b_l1hdr.b_pabd = arc_get_data_abd(hdr, size, hdr,
alloc_flags);
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
}
ARCSTAT_INCR(arcstat_compressed_size, size);
ARCSTAT_INCR(arcstat_uncompressed_size, HDR_GET_LSIZE(hdr));
}
static void
arc_hdr_free_abd(arc_buf_hdr_t *hdr, boolean_t free_rdata)
{
uint64_t size = (free_rdata) ? HDR_GET_PSIZE(hdr) : arc_hdr_size(hdr);
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(hdr->b_l1hdr.b_pabd != NULL || HDR_HAS_RABD(hdr));
IMPLY(free_rdata, HDR_HAS_RABD(hdr));
/*
* If the hdr is currently being written to the l2arc then
* we defer freeing the data by adding it to the l2arc_free_on_write
* list. The l2arc will free the data once it's finished
* writing it to the l2arc device.
*/
if (HDR_L2_WRITING(hdr)) {
arc_hdr_free_on_write(hdr, free_rdata);
ARCSTAT_BUMP(arcstat_l2_free_on_write);
} else if (free_rdata) {
arc_free_data_abd(hdr, hdr->b_crypt_hdr.b_rabd, size, hdr);
} else {
arc_free_data_abd(hdr, hdr->b_l1hdr.b_pabd, size, hdr);
}
if (free_rdata) {
hdr->b_crypt_hdr.b_rabd = NULL;
ARCSTAT_INCR(arcstat_raw_size, -size);
} else {
hdr->b_l1hdr.b_pabd = NULL;
}
if (hdr->b_l1hdr.b_pabd == NULL && !HDR_HAS_RABD(hdr))
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_NUMFUNCS;
ARCSTAT_INCR(arcstat_compressed_size, -size);
ARCSTAT_INCR(arcstat_uncompressed_size, -HDR_GET_LSIZE(hdr));
}
/*
* Allocate empty anonymous ARC header. The header will get its identity
* assigned and buffers attached later as part of read or write operations.
*
* In case of read arc_read() assigns header its identify (b_dva + b_birth),
* inserts it into ARC hash to become globally visible and allocates physical
* (b_pabd) or raw (b_rabd) ABD buffer to read into from disk. On disk read
* completion arc_read_done() allocates ARC buffer(s) as needed, potentially
* sharing one of them with the physical ABD buffer.
*
* In case of write arc_alloc_buf() allocates ARC buffer to be filled with
* data. Then after compression and/or encryption arc_write_ready() allocates
* and fills (or potentially shares) physical (b_pabd) or raw (b_rabd) ABD
* buffer. On disk write completion arc_write_done() assigns the header its
* new identity (b_dva + b_birth) and inserts into ARC hash.
*
* In case of partial overwrite the old data is read first as described. Then
* arc_release() either allocates new anonymous ARC header and moves the ARC
* buffer to it, or reuses the old ARC header by discarding its identity and
* removing it from ARC hash. After buffer modification normal write process
* follows as described.
*/
static arc_buf_hdr_t *
arc_hdr_alloc(uint64_t spa, int32_t psize, int32_t lsize,
boolean_t protected, enum zio_compress compression_type, uint8_t complevel,
arc_buf_contents_t type)
{
arc_buf_hdr_t *hdr;
VERIFY(type == ARC_BUFC_DATA || type == ARC_BUFC_METADATA);
hdr = kmem_cache_alloc(hdr_full_cache, KM_PUSHPAGE);
ASSERT(HDR_EMPTY(hdr));
#ifdef ZFS_DEBUG
ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
#endif
HDR_SET_PSIZE(hdr, psize);
HDR_SET_LSIZE(hdr, lsize);
hdr->b_spa = spa;
hdr->b_type = type;
hdr->b_flags = 0;
arc_hdr_set_flags(hdr, arc_bufc_to_flags(type) | ARC_FLAG_HAS_L1HDR);
arc_hdr_set_compress(hdr, compression_type);
hdr->b_complevel = complevel;
if (protected)
arc_hdr_set_flags(hdr, ARC_FLAG_PROTECTED);
hdr->b_l1hdr.b_state = arc_anon;
hdr->b_l1hdr.b_arc_access = 0;
hdr->b_l1hdr.b_mru_hits = 0;
hdr->b_l1hdr.b_mru_ghost_hits = 0;
hdr->b_l1hdr.b_mfu_hits = 0;
hdr->b_l1hdr.b_mfu_ghost_hits = 0;
hdr->b_l1hdr.b_buf = NULL;
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
return (hdr);
}
/*
* Transition between the two allocation states for the arc_buf_hdr struct.
* The arc_buf_hdr struct can be allocated with (hdr_full_cache) or without
* (hdr_l2only_cache) the fields necessary for the L1 cache - the smaller
* version is used when a cache buffer is only in the L2ARC in order to reduce
* memory usage.
*/
static arc_buf_hdr_t *
arc_hdr_realloc(arc_buf_hdr_t *hdr, kmem_cache_t *old, kmem_cache_t *new)
{
ASSERT(HDR_HAS_L2HDR(hdr));
arc_buf_hdr_t *nhdr;
l2arc_dev_t *dev = hdr->b_l2hdr.b_dev;
ASSERT((old == hdr_full_cache && new == hdr_l2only_cache) ||
(old == hdr_l2only_cache && new == hdr_full_cache));
nhdr = kmem_cache_alloc(new, KM_PUSHPAGE);
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
buf_hash_remove(hdr);
memcpy(nhdr, hdr, HDR_L2ONLY_SIZE);
if (new == hdr_full_cache) {
arc_hdr_set_flags(nhdr, ARC_FLAG_HAS_L1HDR);
/*
* arc_access and arc_change_state need to be aware that a
* header has just come out of L2ARC, so we set its state to
* l2c_only even though it's about to change.
*/
nhdr->b_l1hdr.b_state = arc_l2c_only;
/* Verify previous threads set to NULL before freeing */
ASSERT3P(nhdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
} else {
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
#ifdef ZFS_DEBUG
ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
#endif
/*
* If we've reached here, We must have been called from
* arc_evict_hdr(), as such we should have already been
* removed from any ghost list we were previously on
* (which protects us from racing with arc_evict_state),
* thus no locking is needed during this check.
*/
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
/*
* A buffer must not be moved into the arc_l2c_only
* state if it's not finished being written out to the
* l2arc device. Otherwise, the b_l1hdr.b_pabd field
* might try to be accessed, even though it was removed.
*/
VERIFY(!HDR_L2_WRITING(hdr));
VERIFY3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
arc_hdr_clear_flags(nhdr, ARC_FLAG_HAS_L1HDR);
}
/*
* The header has been reallocated so we need to re-insert it into any
* lists it was on.
*/
(void) buf_hash_insert(nhdr, NULL);
ASSERT(list_link_active(&hdr->b_l2hdr.b_l2node));
mutex_enter(&dev->l2ad_mtx);
/*
* We must place the realloc'ed header back into the list at
* the same spot. Otherwise, if it's placed earlier in the list,
* l2arc_write_buffers() could find it during the function's
* write phase, and try to write it out to the l2arc.
*/
list_insert_after(&dev->l2ad_buflist, hdr, nhdr);
list_remove(&dev->l2ad_buflist, hdr);
mutex_exit(&dev->l2ad_mtx);
/*
* Since we're using the pointer address as the tag when
* incrementing and decrementing the l2ad_alloc refcount, we
* must remove the old pointer (that we're about to destroy) and
* add the new pointer to the refcount. Otherwise we'd remove
* the wrong pointer address when calling arc_hdr_destroy() later.
*/
(void) zfs_refcount_remove_many(&dev->l2ad_alloc,
arc_hdr_size(hdr), hdr);
(void) zfs_refcount_add_many(&dev->l2ad_alloc,
arc_hdr_size(nhdr), nhdr);
buf_discard_identity(hdr);
kmem_cache_free(old, hdr);
return (nhdr);
}
/*
* This function is used by the send / receive code to convert a newly
* allocated arc_buf_t to one that is suitable for a raw encrypted write. It
* is also used to allow the root objset block to be updated without altering
* its embedded MACs. Both block types will always be uncompressed so we do not
* have to worry about compression type or psize.
*/
void
arc_convert_to_raw(arc_buf_t *buf, uint64_t dsobj, boolean_t byteorder,
dmu_object_type_t ot, const uint8_t *salt, const uint8_t *iv,
const uint8_t *mac)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT(ot == DMU_OT_DNODE || ot == DMU_OT_OBJSET);
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT3P(hdr->b_l1hdr.b_state, ==, arc_anon);
buf->b_flags |= (ARC_BUF_FLAG_COMPRESSED | ARC_BUF_FLAG_ENCRYPTED);
arc_hdr_set_flags(hdr, ARC_FLAG_PROTECTED);
hdr->b_crypt_hdr.b_dsobj = dsobj;
hdr->b_crypt_hdr.b_ot = ot;
hdr->b_l1hdr.b_byteswap = (byteorder == ZFS_HOST_BYTEORDER) ?
DMU_BSWAP_NUMFUNCS : DMU_OT_BYTESWAP(ot);
if (!arc_hdr_has_uncompressed_buf(hdr))
arc_cksum_free(hdr);
if (salt != NULL)
memcpy(hdr->b_crypt_hdr.b_salt, salt, ZIO_DATA_SALT_LEN);
if (iv != NULL)
memcpy(hdr->b_crypt_hdr.b_iv, iv, ZIO_DATA_IV_LEN);
if (mac != NULL)
memcpy(hdr->b_crypt_hdr.b_mac, mac, ZIO_DATA_MAC_LEN);
}
/*
* Allocate a new arc_buf_hdr_t and arc_buf_t and return the buf to the caller.
* The buf is returned thawed since we expect the consumer to modify it.
*/
arc_buf_t *
arc_alloc_buf(spa_t *spa, const void *tag, arc_buf_contents_t type,
int32_t size)
{
arc_buf_hdr_t *hdr = arc_hdr_alloc(spa_load_guid(spa), size, size,
B_FALSE, ZIO_COMPRESS_OFF, 0, type);
arc_buf_t *buf = NULL;
VERIFY0(arc_buf_alloc_impl(hdr, spa, NULL, tag, B_FALSE, B_FALSE,
B_FALSE, B_FALSE, &buf));
arc_buf_thaw(buf);
return (buf);
}
/*
* Allocate a compressed buf in the same manner as arc_alloc_buf. Don't use this
* for bufs containing metadata.
*/
arc_buf_t *
arc_alloc_compressed_buf(spa_t *spa, const void *tag, uint64_t psize,
uint64_t lsize, enum zio_compress compression_type, uint8_t complevel)
{
ASSERT3U(lsize, >, 0);
ASSERT3U(lsize, >=, psize);
ASSERT3U(compression_type, >, ZIO_COMPRESS_OFF);
ASSERT3U(compression_type, <, ZIO_COMPRESS_FUNCTIONS);
arc_buf_hdr_t *hdr = arc_hdr_alloc(spa_load_guid(spa), psize, lsize,
B_FALSE, compression_type, complevel, ARC_BUFC_DATA);
arc_buf_t *buf = NULL;
VERIFY0(arc_buf_alloc_impl(hdr, spa, NULL, tag, B_FALSE,
B_TRUE, B_FALSE, B_FALSE, &buf));
arc_buf_thaw(buf);
/*
* To ensure that the hdr has the correct data in it if we call
* arc_untransform() on this buf before it's been written to disk,
* it's easiest if we just set up sharing between the buf and the hdr.
*/
arc_share_buf(hdr, buf);
return (buf);
}
arc_buf_t *
arc_alloc_raw_buf(spa_t *spa, const void *tag, uint64_t dsobj,
boolean_t byteorder, const uint8_t *salt, const uint8_t *iv,
const uint8_t *mac, dmu_object_type_t ot, uint64_t psize, uint64_t lsize,
enum zio_compress compression_type, uint8_t complevel)
{
arc_buf_hdr_t *hdr;
arc_buf_t *buf;
arc_buf_contents_t type = DMU_OT_IS_METADATA(ot) ?
ARC_BUFC_METADATA : ARC_BUFC_DATA;
ASSERT3U(lsize, >, 0);
ASSERT3U(lsize, >=, psize);
ASSERT3U(compression_type, >=, ZIO_COMPRESS_OFF);
ASSERT3U(compression_type, <, ZIO_COMPRESS_FUNCTIONS);
hdr = arc_hdr_alloc(spa_load_guid(spa), psize, lsize, B_TRUE,
compression_type, complevel, type);
hdr->b_crypt_hdr.b_dsobj = dsobj;
hdr->b_crypt_hdr.b_ot = ot;
hdr->b_l1hdr.b_byteswap = (byteorder == ZFS_HOST_BYTEORDER) ?
DMU_BSWAP_NUMFUNCS : DMU_OT_BYTESWAP(ot);
memcpy(hdr->b_crypt_hdr.b_salt, salt, ZIO_DATA_SALT_LEN);
memcpy(hdr->b_crypt_hdr.b_iv, iv, ZIO_DATA_IV_LEN);
memcpy(hdr->b_crypt_hdr.b_mac, mac, ZIO_DATA_MAC_LEN);
/*
* This buffer will be considered encrypted even if the ot is not an
* encrypted type. It will become authenticated instead in
* arc_write_ready().
*/
buf = NULL;
VERIFY0(arc_buf_alloc_impl(hdr, spa, NULL, tag, B_TRUE, B_TRUE,
B_FALSE, B_FALSE, &buf));
arc_buf_thaw(buf);
return (buf);
}
static void
l2arc_hdr_arcstats_update(arc_buf_hdr_t *hdr, boolean_t incr,
boolean_t state_only)
{
l2arc_buf_hdr_t *l2hdr = &hdr->b_l2hdr;
l2arc_dev_t *dev = l2hdr->b_dev;
uint64_t lsize = HDR_GET_LSIZE(hdr);
uint64_t psize = HDR_GET_PSIZE(hdr);
uint64_t asize = vdev_psize_to_asize(dev->l2ad_vdev, psize);
arc_buf_contents_t type = hdr->b_type;
int64_t lsize_s;
int64_t psize_s;
int64_t asize_s;
if (incr) {
lsize_s = lsize;
psize_s = psize;
asize_s = asize;
} else {
lsize_s = -lsize;
psize_s = -psize;
asize_s = -asize;
}
/* If the buffer is a prefetch, count it as such. */
if (HDR_PREFETCH(hdr)) {
ARCSTAT_INCR(arcstat_l2_prefetch_asize, asize_s);
} else {
/*
* We use the value stored in the L2 header upon initial
* caching in L2ARC. This value will be updated in case
* an MRU/MRU_ghost buffer transitions to MFU but the L2ARC
* metadata (log entry) cannot currently be updated. Having
* the ARC state in the L2 header solves the problem of a
* possibly absent L1 header (apparent in buffers restored
* from persistent L2ARC).
*/
switch (hdr->b_l2hdr.b_arcs_state) {
case ARC_STATE_MRU_GHOST:
case ARC_STATE_MRU:
ARCSTAT_INCR(arcstat_l2_mru_asize, asize_s);
break;
case ARC_STATE_MFU_GHOST:
case ARC_STATE_MFU:
ARCSTAT_INCR(arcstat_l2_mfu_asize, asize_s);
break;
default:
break;
}
}
if (state_only)
return;
ARCSTAT_INCR(arcstat_l2_psize, psize_s);
ARCSTAT_INCR(arcstat_l2_lsize, lsize_s);
switch (type) {
case ARC_BUFC_DATA:
ARCSTAT_INCR(arcstat_l2_bufc_data_asize, asize_s);
break;
case ARC_BUFC_METADATA:
ARCSTAT_INCR(arcstat_l2_bufc_metadata_asize, asize_s);
break;
default:
break;
}
}
static void
arc_hdr_l2hdr_destroy(arc_buf_hdr_t *hdr)
{
l2arc_buf_hdr_t *l2hdr = &hdr->b_l2hdr;
l2arc_dev_t *dev = l2hdr->b_dev;
uint64_t psize = HDR_GET_PSIZE(hdr);
uint64_t asize = vdev_psize_to_asize(dev->l2ad_vdev, psize);
ASSERT(MUTEX_HELD(&dev->l2ad_mtx));
ASSERT(HDR_HAS_L2HDR(hdr));
list_remove(&dev->l2ad_buflist, hdr);
l2arc_hdr_arcstats_decrement(hdr);
vdev_space_update(dev->l2ad_vdev, -asize, 0, 0);
(void) zfs_refcount_remove_many(&dev->l2ad_alloc, arc_hdr_size(hdr),
hdr);
arc_hdr_clear_flags(hdr, ARC_FLAG_HAS_L2HDR);
}
static void
arc_hdr_destroy(arc_buf_hdr_t *hdr)
{
if (HDR_HAS_L1HDR(hdr)) {
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
ASSERT3P(hdr->b_l1hdr.b_state, ==, arc_anon);
}
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT(!HDR_IN_HASH_TABLE(hdr));
if (HDR_HAS_L2HDR(hdr)) {
l2arc_dev_t *dev = hdr->b_l2hdr.b_dev;
boolean_t buflist_held = MUTEX_HELD(&dev->l2ad_mtx);
if (!buflist_held)
mutex_enter(&dev->l2ad_mtx);
/*
* Even though we checked this conditional above, we
* need to check this again now that we have the
* l2ad_mtx. This is because we could be racing with
* another thread calling l2arc_evict() which might have
* destroyed this header's L2 portion as we were waiting
* to acquire the l2ad_mtx. If that happens, we don't
* want to re-destroy the header's L2 portion.
*/
if (HDR_HAS_L2HDR(hdr)) {
if (!HDR_EMPTY(hdr))
buf_discard_identity(hdr);
arc_hdr_l2hdr_destroy(hdr);
}
if (!buflist_held)
mutex_exit(&dev->l2ad_mtx);
}
/*
* The header's identify can only be safely discarded once it is no
* longer discoverable. This requires removing it from the hash table
* and the l2arc header list. After this point the hash lock can not
* be used to protect the header.
*/
if (!HDR_EMPTY(hdr))
buf_discard_identity(hdr);
if (HDR_HAS_L1HDR(hdr)) {
arc_cksum_free(hdr);
while (hdr->b_l1hdr.b_buf != NULL)
arc_buf_destroy_impl(hdr->b_l1hdr.b_buf);
if (hdr->b_l1hdr.b_pabd != NULL)
arc_hdr_free_abd(hdr, B_FALSE);
if (HDR_HAS_RABD(hdr))
arc_hdr_free_abd(hdr, B_TRUE);
}
ASSERT3P(hdr->b_hash_next, ==, NULL);
if (HDR_HAS_L1HDR(hdr)) {
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
#ifdef ZFS_DEBUG
ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
#endif
kmem_cache_free(hdr_full_cache, hdr);
} else {
kmem_cache_free(hdr_l2only_cache, hdr);
}
}
void
arc_buf_destroy(arc_buf_t *buf, const void *tag)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
if (hdr->b_l1hdr.b_state == arc_anon) {
ASSERT3P(hdr->b_l1hdr.b_buf, ==, buf);
ASSERT(ARC_BUF_LAST(buf));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
VERIFY0(remove_reference(hdr, tag));
return;
}
kmutex_t *hash_lock = HDR_LOCK(hdr);
mutex_enter(hash_lock);
ASSERT3P(hdr, ==, buf->b_hdr);
ASSERT3P(hdr->b_l1hdr.b_buf, !=, NULL);
ASSERT3P(hash_lock, ==, HDR_LOCK(hdr));
ASSERT3P(hdr->b_l1hdr.b_state, !=, arc_anon);
ASSERT3P(buf->b_data, !=, NULL);
arc_buf_destroy_impl(buf);
(void) remove_reference(hdr, tag);
mutex_exit(hash_lock);
}
/*
* Evict the arc_buf_hdr that is provided as a parameter. The resultant
* state of the header is dependent on its state prior to entering this
* function. The following transitions are possible:
*
* - arc_mru -> arc_mru_ghost
* - arc_mfu -> arc_mfu_ghost
* - arc_mru_ghost -> arc_l2c_only
* - arc_mru_ghost -> deleted
* - arc_mfu_ghost -> arc_l2c_only
* - arc_mfu_ghost -> deleted
* - arc_uncached -> deleted
*
* Return total size of evicted data buffers for eviction progress tracking.
* When evicting from ghost states return logical buffer size to make eviction
* progress at the same (or at least comparable) rate as from non-ghost states.
*
* Return *real_evicted for actual ARC size reduction to wake up threads
* waiting for it. For non-ghost states it includes size of evicted data
* buffers (the headers are not freed there). For ghost states it includes
* only the evicted headers size.
*/
static int64_t
arc_evict_hdr(arc_buf_hdr_t *hdr, uint64_t *real_evicted)
{
arc_state_t *evicted_state, *state;
int64_t bytes_evicted = 0;
uint_t min_lifetime = HDR_PRESCIENT_PREFETCH(hdr) ?
arc_min_prescient_prefetch_ms : arc_min_prefetch_ms;
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
ASSERT0(zfs_refcount_count(&hdr->b_l1hdr.b_refcnt));
*real_evicted = 0;
state = hdr->b_l1hdr.b_state;
if (GHOST_STATE(state)) {
/*
* l2arc_write_buffers() relies on a header's L1 portion
* (i.e. its b_pabd field) during it's write phase.
* Thus, we cannot push a header onto the arc_l2c_only
* state (removing its L1 piece) until the header is
* done being written to the l2arc.
*/
if (HDR_HAS_L2HDR(hdr) && HDR_L2_WRITING(hdr)) {
ARCSTAT_BUMP(arcstat_evict_l2_skip);
return (bytes_evicted);
}
ARCSTAT_BUMP(arcstat_deleted);
bytes_evicted += HDR_GET_LSIZE(hdr);
DTRACE_PROBE1(arc__delete, arc_buf_hdr_t *, hdr);
if (HDR_HAS_L2HDR(hdr)) {
ASSERT(hdr->b_l1hdr.b_pabd == NULL);
ASSERT(!HDR_HAS_RABD(hdr));
/*
* This buffer is cached on the 2nd Level ARC;
* don't destroy the header.
*/
arc_change_state(arc_l2c_only, hdr);
/*
* dropping from L1+L2 cached to L2-only,
* realloc to remove the L1 header.
*/
(void) arc_hdr_realloc(hdr, hdr_full_cache,
hdr_l2only_cache);
*real_evicted += HDR_FULL_SIZE - HDR_L2ONLY_SIZE;
} else {
arc_change_state(arc_anon, hdr);
arc_hdr_destroy(hdr);
*real_evicted += HDR_FULL_SIZE;
}
return (bytes_evicted);
}
ASSERT(state == arc_mru || state == arc_mfu || state == arc_uncached);
evicted_state = (state == arc_uncached) ? arc_anon :
((state == arc_mru) ? arc_mru_ghost : arc_mfu_ghost);
/* prefetch buffers have a minimum lifespan */
if ((hdr->b_flags & (ARC_FLAG_PREFETCH | ARC_FLAG_INDIRECT)) &&
ddi_get_lbolt() - hdr->b_l1hdr.b_arc_access <
MSEC_TO_TICK(min_lifetime)) {
ARCSTAT_BUMP(arcstat_evict_skip);
return (bytes_evicted);
}
if (HDR_HAS_L2HDR(hdr)) {
ARCSTAT_INCR(arcstat_evict_l2_cached, HDR_GET_LSIZE(hdr));
} else {
if (l2arc_write_eligible(hdr->b_spa, hdr)) {
ARCSTAT_INCR(arcstat_evict_l2_eligible,
HDR_GET_LSIZE(hdr));
switch (state->arcs_state) {
case ARC_STATE_MRU:
ARCSTAT_INCR(
arcstat_evict_l2_eligible_mru,
HDR_GET_LSIZE(hdr));
break;
case ARC_STATE_MFU:
ARCSTAT_INCR(
arcstat_evict_l2_eligible_mfu,
HDR_GET_LSIZE(hdr));
break;
default:
break;
}
} else {
ARCSTAT_INCR(arcstat_evict_l2_ineligible,
HDR_GET_LSIZE(hdr));
}
}
bytes_evicted += arc_hdr_size(hdr);
*real_evicted += arc_hdr_size(hdr);
/*
* If this hdr is being evicted and has a compressed buffer then we
* discard it here before we change states. This ensures that the
* accounting is updated correctly in arc_free_data_impl().
*/
if (hdr->b_l1hdr.b_pabd != NULL)
arc_hdr_free_abd(hdr, B_FALSE);
if (HDR_HAS_RABD(hdr))
arc_hdr_free_abd(hdr, B_TRUE);
arc_change_state(evicted_state, hdr);
DTRACE_PROBE1(arc__evict, arc_buf_hdr_t *, hdr);
if (evicted_state == arc_anon) {
arc_hdr_destroy(hdr);
*real_evicted += HDR_FULL_SIZE;
} else {
ASSERT(HDR_IN_HASH_TABLE(hdr));
}
return (bytes_evicted);
}
static void
arc_set_need_free(void)
{
ASSERT(MUTEX_HELD(&arc_evict_lock));
int64_t remaining = arc_free_memory() - arc_sys_free / 2;
arc_evict_waiter_t *aw = list_tail(&arc_evict_waiters);
if (aw == NULL) {
arc_need_free = MAX(-remaining, 0);
} else {
arc_need_free =
MAX(-remaining, (int64_t)(aw->aew_count - arc_evict_count));
}
}
static uint64_t
arc_evict_state_impl(multilist_t *ml, int idx, arc_buf_hdr_t *marker,
uint64_t spa, uint64_t bytes)
{
multilist_sublist_t *mls;
uint64_t bytes_evicted = 0, real_evicted = 0;
arc_buf_hdr_t *hdr;
kmutex_t *hash_lock;
uint_t evict_count = zfs_arc_evict_batch_limit;
ASSERT3P(marker, !=, NULL);
mls = multilist_sublist_lock(ml, idx);
for (hdr = multilist_sublist_prev(mls, marker); likely(hdr != NULL);
hdr = multilist_sublist_prev(mls, marker)) {
if ((evict_count == 0) || (bytes_evicted >= bytes))
break;
/*
* To keep our iteration location, move the marker
* forward. Since we're not holding hdr's hash lock, we
* must be very careful and not remove 'hdr' from the
* sublist. Otherwise, other consumers might mistake the
* 'hdr' as not being on a sublist when they call the
* multilist_link_active() function (they all rely on
* the hash lock protecting concurrent insertions and
* removals). multilist_sublist_move_forward() was
* specifically implemented to ensure this is the case
* (only 'marker' will be removed and re-inserted).
*/
multilist_sublist_move_forward(mls, marker);
/*
* The only case where the b_spa field should ever be
* zero, is the marker headers inserted by
* arc_evict_state(). It's possible for multiple threads
* to be calling arc_evict_state() concurrently (e.g.
* dsl_pool_close() and zio_inject_fault()), so we must
* skip any markers we see from these other threads.
*/
if (hdr->b_spa == 0)
continue;
/* we're only interested in evicting buffers of a certain spa */
if (spa != 0 && hdr->b_spa != spa) {
ARCSTAT_BUMP(arcstat_evict_skip);
continue;
}
hash_lock = HDR_LOCK(hdr);
/*
* We aren't calling this function from any code path
* that would already be holding a hash lock, so we're
* asserting on this assumption to be defensive in case
* this ever changes. Without this check, it would be
* possible to incorrectly increment arcstat_mutex_miss
* below (e.g. if the code changed such that we called
* this function with a hash lock held).
*/
ASSERT(!MUTEX_HELD(hash_lock));
if (mutex_tryenter(hash_lock)) {
uint64_t revicted;
uint64_t evicted = arc_evict_hdr(hdr, &revicted);
mutex_exit(hash_lock);
bytes_evicted += evicted;
real_evicted += revicted;
/*
* If evicted is zero, arc_evict_hdr() must have
* decided to skip this header, don't increment
* evict_count in this case.
*/
if (evicted != 0)
evict_count--;
} else {
ARCSTAT_BUMP(arcstat_mutex_miss);
}
}
multilist_sublist_unlock(mls);
/*
* Increment the count of evicted bytes, and wake up any threads that
* are waiting for the count to reach this value. Since the list is
* ordered by ascending aew_count, we pop off the beginning of the
* list until we reach the end, or a waiter that's past the current
* "count". Doing this outside the loop reduces the number of times
* we need to acquire the global arc_evict_lock.
*
* Only wake when there's sufficient free memory in the system
* (specifically, arc_sys_free/2, which by default is a bit more than
* 1/64th of RAM). See the comments in arc_wait_for_eviction().
*/
mutex_enter(&arc_evict_lock);
arc_evict_count += real_evicted;
if (arc_free_memory() > arc_sys_free / 2) {
arc_evict_waiter_t *aw;
while ((aw = list_head(&arc_evict_waiters)) != NULL &&
aw->aew_count <= arc_evict_count) {
list_remove(&arc_evict_waiters, aw);
cv_broadcast(&aw->aew_cv);
}
}
arc_set_need_free();
mutex_exit(&arc_evict_lock);
/*
* If the ARC size is reduced from arc_c_max to arc_c_min (especially
* if the average cached block is small), eviction can be on-CPU for
* many seconds. To ensure that other threads that may be bound to
* this CPU are able to make progress, make a voluntary preemption
* call here.
*/
kpreempt(KPREEMPT_SYNC);
return (bytes_evicted);
}
/*
* Allocate an array of buffer headers used as placeholders during arc state
* eviction.
*/
static arc_buf_hdr_t **
arc_state_alloc_markers(int count)
{
arc_buf_hdr_t **markers;
markers = kmem_zalloc(sizeof (*markers) * count, KM_SLEEP);
for (int i = 0; i < count; i++) {
markers[i] = kmem_cache_alloc(hdr_full_cache, KM_SLEEP);
/*
* A b_spa of 0 is used to indicate that this header is
* a marker. This fact is used in arc_evict_state_impl().
*/
markers[i]->b_spa = 0;
}
return (markers);
}
static void
arc_state_free_markers(arc_buf_hdr_t **markers, int count)
{
for (int i = 0; i < count; i++)
kmem_cache_free(hdr_full_cache, markers[i]);
kmem_free(markers, sizeof (*markers) * count);
}
/*
* Evict buffers from the given arc state, until we've removed the
* specified number of bytes. Move the removed buffers to the
* appropriate evict state.
*
* This function makes a "best effort". It skips over any buffers
* it can't get a hash_lock on, and so, may not catch all candidates.
* It may also return without evicting as much space as requested.
*
* If bytes is specified using the special value ARC_EVICT_ALL, this
* will evict all available (i.e. unlocked and evictable) buffers from
* the given arc state; which is used by arc_flush().
*/
static uint64_t
arc_evict_state(arc_state_t *state, arc_buf_contents_t type, uint64_t spa,
uint64_t bytes)
{
uint64_t total_evicted = 0;
multilist_t *ml = &state->arcs_list[type];
int num_sublists;
arc_buf_hdr_t **markers;
num_sublists = multilist_get_num_sublists(ml);
/*
* If we've tried to evict from each sublist, made some
* progress, but still have not hit the target number of bytes
* to evict, we want to keep trying. The markers allow us to
* pick up where we left off for each individual sublist, rather
* than starting from the tail each time.
*/
if (zthr_iscurthread(arc_evict_zthr)) {
markers = arc_state_evict_markers;
ASSERT3S(num_sublists, <=, arc_state_evict_marker_count);
} else {
markers = arc_state_alloc_markers(num_sublists);
}
for (int i = 0; i < num_sublists; i++) {
multilist_sublist_t *mls;
mls = multilist_sublist_lock(ml, i);
multilist_sublist_insert_tail(mls, markers[i]);
multilist_sublist_unlock(mls);
}
/*
* While we haven't hit our target number of bytes to evict, or
* we're evicting all available buffers.
*/
while (total_evicted < bytes) {
int sublist_idx = multilist_get_random_index(ml);
uint64_t scan_evicted = 0;
/*
* Start eviction using a randomly selected sublist,
* this is to try and evenly balance eviction across all
* sublists. Always starting at the same sublist
* (e.g. index 0) would cause evictions to favor certain
* sublists over others.
*/
for (int i = 0; i < num_sublists; i++) {
uint64_t bytes_remaining;
uint64_t bytes_evicted;
if (total_evicted < bytes)
bytes_remaining = bytes - total_evicted;
else
break;
bytes_evicted = arc_evict_state_impl(ml, sublist_idx,
markers[sublist_idx], spa, bytes_remaining);
scan_evicted += bytes_evicted;
total_evicted += bytes_evicted;
/* we've reached the end, wrap to the beginning */
if (++sublist_idx >= num_sublists)
sublist_idx = 0;
}
/*
* If we didn't evict anything during this scan, we have
* no reason to believe we'll evict more during another
* scan, so break the loop.
*/
if (scan_evicted == 0) {
/* This isn't possible, let's make that obvious */
ASSERT3S(bytes, !=, 0);
/*
* When bytes is ARC_EVICT_ALL, the only way to
* break the loop is when scan_evicted is zero.
* In that case, we actually have evicted enough,
* so we don't want to increment the kstat.
*/
if (bytes != ARC_EVICT_ALL) {
ASSERT3S(total_evicted, <, bytes);
ARCSTAT_BUMP(arcstat_evict_not_enough);
}
break;
}
}
for (int i = 0; i < num_sublists; i++) {
multilist_sublist_t *mls = multilist_sublist_lock(ml, i);
multilist_sublist_remove(mls, markers[i]);
multilist_sublist_unlock(mls);
}
if (markers != arc_state_evict_markers)
arc_state_free_markers(markers, num_sublists);
return (total_evicted);
}
/*
* Flush all "evictable" data of the given type from the arc state
* specified. This will not evict any "active" buffers (i.e. referenced).
*
* When 'retry' is set to B_FALSE, the function will make a single pass
* over the state and evict any buffers that it can. Since it doesn't
* continually retry the eviction, it might end up leaving some buffers
* in the ARC due to lock misses.
*
* When 'retry' is set to B_TRUE, the function will continually retry the
* eviction until *all* evictable buffers have been removed from the
* state. As a result, if concurrent insertions into the state are
* allowed (e.g. if the ARC isn't shutting down), this function might
* wind up in an infinite loop, continually trying to evict buffers.
*/
static uint64_t
arc_flush_state(arc_state_t *state, uint64_t spa, arc_buf_contents_t type,
boolean_t retry)
{
uint64_t evicted = 0;
while (zfs_refcount_count(&state->arcs_esize[type]) != 0) {
evicted += arc_evict_state(state, type, spa, ARC_EVICT_ALL);
if (!retry)
break;
}
return (evicted);
}
/*
* Evict the specified number of bytes from the state specified. This
* function prevents us from trying to evict more from a state's list
* than is "evictable", and to skip evicting altogether when passed a
* negative value for "bytes". In contrast, arc_evict_state() will
* evict everything it can, when passed a negative value for "bytes".
*/
static uint64_t
arc_evict_impl(arc_state_t *state, arc_buf_contents_t type, int64_t bytes)
{
uint64_t delta;
if (bytes > 0 && zfs_refcount_count(&state->arcs_esize[type]) > 0) {
delta = MIN(zfs_refcount_count(&state->arcs_esize[type]),
bytes);
return (arc_evict_state(state, type, 0, delta));
}
return (0);
}
/*
* Adjust specified fraction, taking into account initial ghost state(s) size,
* ghost hit bytes towards increasing the fraction, ghost hit bytes towards
* decreasing it, plus a balance factor, controlling the decrease rate, used
* to balance metadata vs data.
*/
static uint64_t
arc_evict_adj(uint64_t frac, uint64_t total, uint64_t up, uint64_t down,
uint_t balance)
{
if (total < 8 || up + down == 0)
return (frac);
/*
* We should not have more ghost hits than ghost size, but they
* may get close. Restrict maximum adjustment in that case.
*/
if (up + down >= total / 4) {
uint64_t scale = (up + down) / (total / 8);
up /= scale;
down /= scale;
}
/* Get maximal dynamic range by choosing optimal shifts. */
int s = highbit64(total);
s = MIN(64 - s, 32);
uint64_t ofrac = (1ULL << 32) - frac;
if (frac >= 4 * ofrac)
up /= frac / (2 * ofrac + 1);
up = (up << s) / (total >> (32 - s));
if (ofrac >= 4 * frac)
down /= ofrac / (2 * frac + 1);
down = (down << s) / (total >> (32 - s));
down = down * 100 / balance;
return (frac + up - down);
}
/*
* Evict buffers from the cache, such that arcstat_size is capped by arc_c.
*/
static uint64_t
arc_evict(void)
{
uint64_t asize, bytes, total_evicted = 0;
int64_t e, mrud, mrum, mfud, mfum, w;
static uint64_t ogrd, ogrm, ogfd, ogfm;
static uint64_t gsrd, gsrm, gsfd, gsfm;
uint64_t ngrd, ngrm, ngfd, ngfm;
/* Get current size of ARC states we can evict from. */
mrud = zfs_refcount_count(&arc_mru->arcs_size[ARC_BUFC_DATA]) +
zfs_refcount_count(&arc_anon->arcs_size[ARC_BUFC_DATA]);
mrum = zfs_refcount_count(&arc_mru->arcs_size[ARC_BUFC_METADATA]) +
zfs_refcount_count(&arc_anon->arcs_size[ARC_BUFC_METADATA]);
mfud = zfs_refcount_count(&arc_mfu->arcs_size[ARC_BUFC_DATA]);
mfum = zfs_refcount_count(&arc_mfu->arcs_size[ARC_BUFC_METADATA]);
uint64_t d = mrud + mfud;
uint64_t m = mrum + mfum;
uint64_t t = d + m;
/* Get ARC ghost hits since last eviction. */
ngrd = wmsum_value(&arc_mru_ghost->arcs_hits[ARC_BUFC_DATA]);
uint64_t grd = ngrd - ogrd;
ogrd = ngrd;
ngrm = wmsum_value(&arc_mru_ghost->arcs_hits[ARC_BUFC_METADATA]);
uint64_t grm = ngrm - ogrm;
ogrm = ngrm;
ngfd = wmsum_value(&arc_mfu_ghost->arcs_hits[ARC_BUFC_DATA]);
uint64_t gfd = ngfd - ogfd;
ogfd = ngfd;
ngfm = wmsum_value(&arc_mfu_ghost->arcs_hits[ARC_BUFC_METADATA]);
uint64_t gfm = ngfm - ogfm;
ogfm = ngfm;
/* Adjust ARC states balance based on ghost hits. */
arc_meta = arc_evict_adj(arc_meta, gsrd + gsrm + gsfd + gsfm,
grm + gfm, grd + gfd, zfs_arc_meta_balance);
arc_pd = arc_evict_adj(arc_pd, gsrd + gsfd, grd, gfd, 100);
arc_pm = arc_evict_adj(arc_pm, gsrm + gsfm, grm, gfm, 100);
asize = aggsum_value(&arc_sums.arcstat_size);
int64_t wt = t - (asize - arc_c);
/*
* Try to reduce pinned dnodes if more than 3/4 of wanted metadata
* target is not evictable or if they go over arc_dnode_limit.
*/
int64_t prune = 0;
int64_t dn = wmsum_value(&arc_sums.arcstat_dnode_size);
w = wt * (int64_t)(arc_meta >> 16) >> 16;
if (zfs_refcount_count(&arc_mru->arcs_size[ARC_BUFC_METADATA]) +
zfs_refcount_count(&arc_mfu->arcs_size[ARC_BUFC_METADATA]) -
zfs_refcount_count(&arc_mru->arcs_esize[ARC_BUFC_METADATA]) -
zfs_refcount_count(&arc_mfu->arcs_esize[ARC_BUFC_METADATA]) >
w * 3 / 4) {
prune = dn / sizeof (dnode_t) *
zfs_arc_dnode_reduce_percent / 100;
} else if (dn > arc_dnode_limit) {
prune = (dn - arc_dnode_limit) / sizeof (dnode_t) *
zfs_arc_dnode_reduce_percent / 100;
}
if (prune > 0)
arc_prune_async(prune);
/* Evict MRU metadata. */
w = wt * (int64_t)(arc_meta * arc_pm >> 48) >> 16;
e = MIN((int64_t)(asize - arc_c), (int64_t)(mrum - w));
bytes = arc_evict_impl(arc_mru, ARC_BUFC_METADATA, e);
total_evicted += bytes;
mrum -= bytes;
asize -= bytes;
/* Evict MFU metadata. */
w = wt * (int64_t)(arc_meta >> 16) >> 16;
e = MIN((int64_t)(asize - arc_c), (int64_t)(m - w));
bytes = arc_evict_impl(arc_mfu, ARC_BUFC_METADATA, e);
total_evicted += bytes;
mfum -= bytes;
asize -= bytes;
/* Evict MRU data. */
wt -= m - total_evicted;
w = wt * (int64_t)(arc_pd >> 16) >> 16;
e = MIN((int64_t)(asize - arc_c), (int64_t)(mrud - w));
bytes = arc_evict_impl(arc_mru, ARC_BUFC_DATA, e);
total_evicted += bytes;
mrud -= bytes;
asize -= bytes;
/* Evict MFU data. */
e = asize - arc_c;
bytes = arc_evict_impl(arc_mfu, ARC_BUFC_DATA, e);
mfud -= bytes;
total_evicted += bytes;
/*
* Evict ghost lists
*
* Size of each state's ghost list represents how much that state
* may grow by shrinking the other states. Would it need to shrink
* other states to zero (that is unlikely), its ghost size would be
* equal to sum of other three state sizes. But excessive ghost
* size may result in false ghost hits (too far back), that may
* never result in real cache hits if several states are competing.
* So choose some arbitraty point of 1/2 of other state sizes.
*/
gsrd = (mrum + mfud + mfum) / 2;
e = zfs_refcount_count(&arc_mru_ghost->arcs_size[ARC_BUFC_DATA]) -
gsrd;
(void) arc_evict_impl(arc_mru_ghost, ARC_BUFC_DATA, e);
gsrm = (mrud + mfud + mfum) / 2;
e = zfs_refcount_count(&arc_mru_ghost->arcs_size[ARC_BUFC_METADATA]) -
gsrm;
(void) arc_evict_impl(arc_mru_ghost, ARC_BUFC_METADATA, e);
gsfd = (mrud + mrum + mfum) / 2;
e = zfs_refcount_count(&arc_mfu_ghost->arcs_size[ARC_BUFC_DATA]) -
gsfd;
(void) arc_evict_impl(arc_mfu_ghost, ARC_BUFC_DATA, e);
gsfm = (mrud + mrum + mfud) / 2;
e = zfs_refcount_count(&arc_mfu_ghost->arcs_size[ARC_BUFC_METADATA]) -
gsfm;
(void) arc_evict_impl(arc_mfu_ghost, ARC_BUFC_METADATA, e);
return (total_evicted);
}
void
arc_flush(spa_t *spa, boolean_t retry)
{
uint64_t guid = 0;
/*
* If retry is B_TRUE, a spa must not be specified since we have
* no good way to determine if all of a spa's buffers have been
* evicted from an arc state.
*/
ASSERT(!retry || spa == NULL);
if (spa != NULL)
guid = spa_load_guid(spa);
(void) arc_flush_state(arc_mru, guid, ARC_BUFC_DATA, retry);
(void) arc_flush_state(arc_mru, guid, ARC_BUFC_METADATA, retry);
(void) arc_flush_state(arc_mfu, guid, ARC_BUFC_DATA, retry);
(void) arc_flush_state(arc_mfu, guid, ARC_BUFC_METADATA, retry);
(void) arc_flush_state(arc_mru_ghost, guid, ARC_BUFC_DATA, retry);
(void) arc_flush_state(arc_mru_ghost, guid, ARC_BUFC_METADATA, retry);
(void) arc_flush_state(arc_mfu_ghost, guid, ARC_BUFC_DATA, retry);
(void) arc_flush_state(arc_mfu_ghost, guid, ARC_BUFC_METADATA, retry);
(void) arc_flush_state(arc_uncached, guid, ARC_BUFC_DATA, retry);
(void) arc_flush_state(arc_uncached, guid, ARC_BUFC_METADATA, retry);
}
void
arc_reduce_target_size(int64_t to_free)
{
uint64_t c = arc_c;
if (c <= arc_c_min)
return;
/*
* All callers want the ARC to actually evict (at least) this much
* memory. Therefore we reduce from the lower of the current size and
* the target size. This way, even if arc_c is much higher than
* arc_size (as can be the case after many calls to arc_freed(), we will
* immediately have arc_c < arc_size and therefore the arc_evict_zthr
* will evict.
*/
uint64_t asize = aggsum_value(&arc_sums.arcstat_size);
if (asize < c)
to_free += c - asize;
arc_c = MAX((int64_t)c - to_free, (int64_t)arc_c_min);
/* See comment in arc_evict_cb_check() on why lock+flag */
mutex_enter(&arc_evict_lock);
arc_evict_needed = B_TRUE;
mutex_exit(&arc_evict_lock);
zthr_wakeup(arc_evict_zthr);
}
/*
* Determine if the system is under memory pressure and is asking
* to reclaim memory. A return value of B_TRUE indicates that the system
* is under memory pressure and that the arc should adjust accordingly.
*/
boolean_t
arc_reclaim_needed(void)
{
return (arc_available_memory() < 0);
}
void
arc_kmem_reap_soon(void)
{
size_t i;
kmem_cache_t *prev_cache = NULL;
kmem_cache_t *prev_data_cache = NULL;
#ifdef _KERNEL
#if defined(_ILP32)
/*
* Reclaim unused memory from all kmem caches.
*/
kmem_reap();
#endif
#endif
for (i = 0; i < SPA_MAXBLOCKSIZE >> SPA_MINBLOCKSHIFT; i++) {
#if defined(_ILP32)
/* reach upper limit of cache size on 32-bit */
if (zio_buf_cache[i] == NULL)
break;
#endif
if (zio_buf_cache[i] != prev_cache) {
prev_cache = zio_buf_cache[i];
kmem_cache_reap_now(zio_buf_cache[i]);
}
if (zio_data_buf_cache[i] != prev_data_cache) {
prev_data_cache = zio_data_buf_cache[i];
kmem_cache_reap_now(zio_data_buf_cache[i]);
}
}
kmem_cache_reap_now(buf_cache);
kmem_cache_reap_now(hdr_full_cache);
kmem_cache_reap_now(hdr_l2only_cache);
kmem_cache_reap_now(zfs_btree_leaf_cache);
abd_cache_reap_now();
}
static boolean_t
arc_evict_cb_check(void *arg, zthr_t *zthr)
{
(void) arg, (void) zthr;
#ifdef ZFS_DEBUG
/*
* This is necessary in order to keep the kstat information
* up to date for tools that display kstat data such as the
* mdb ::arc dcmd and the Linux crash utility. These tools
* typically do not call kstat's update function, but simply
* dump out stats from the most recent update. Without
* this call, these commands may show stale stats for the
* anon, mru, mru_ghost, mfu, and mfu_ghost lists. Even
* with this call, the data might be out of date if the
* evict thread hasn't been woken recently; but that should
* suffice. The arc_state_t structures can be queried
* directly if more accurate information is needed.
*/
if (arc_ksp != NULL)
arc_ksp->ks_update(arc_ksp, KSTAT_READ);
#endif
/*
* We have to rely on arc_wait_for_eviction() to tell us when to
* evict, rather than checking if we are overflowing here, so that we
* are sure to not leave arc_wait_for_eviction() waiting on aew_cv.
* If we have become "not overflowing" since arc_wait_for_eviction()
* checked, we need to wake it up. We could broadcast the CV here,
* but arc_wait_for_eviction() may have not yet gone to sleep. We
* would need to use a mutex to ensure that this function doesn't
* broadcast until arc_wait_for_eviction() has gone to sleep (e.g.
* the arc_evict_lock). However, the lock ordering of such a lock
* would necessarily be incorrect with respect to the zthr_lock,
* which is held before this function is called, and is held by
* arc_wait_for_eviction() when it calls zthr_wakeup().
*/
if (arc_evict_needed)
return (B_TRUE);
/*
* If we have buffers in uncached state, evict them periodically.
*/
return ((zfs_refcount_count(&arc_uncached->arcs_esize[ARC_BUFC_DATA]) +
zfs_refcount_count(&arc_uncached->arcs_esize[ARC_BUFC_METADATA]) &&
ddi_get_lbolt() - arc_last_uncached_flush >
MSEC_TO_TICK(arc_min_prefetch_ms / 2)));
}
/*
* Keep arc_size under arc_c by running arc_evict which evicts data
* from the ARC.
*/
static void
arc_evict_cb(void *arg, zthr_t *zthr)
{
- (void) arg, (void) zthr;
+ (void) arg;
uint64_t evicted = 0;
fstrans_cookie_t cookie = spl_fstrans_mark();
/* Always try to evict from uncached state. */
arc_last_uncached_flush = ddi_get_lbolt();
evicted += arc_flush_state(arc_uncached, 0, ARC_BUFC_DATA, B_FALSE);
evicted += arc_flush_state(arc_uncached, 0, ARC_BUFC_METADATA, B_FALSE);
/* Evict from other states only if told to. */
if (arc_evict_needed)
evicted += arc_evict();
/*
* If evicted is zero, we couldn't evict anything
* via arc_evict(). This could be due to hash lock
* collisions, but more likely due to the majority of
* arc buffers being unevictable. Therefore, even if
* arc_size is above arc_c, another pass is unlikely to
* be helpful and could potentially cause us to enter an
* infinite loop. Additionally, zthr_iscancelled() is
* checked here so that if the arc is shutting down, the
* broadcast will wake any remaining arc evict waiters.
+ *
+ * Note we cancel using zthr instead of arc_evict_zthr
+ * because the latter may not yet be initializd when the
+ * callback is first invoked.
*/
mutex_enter(&arc_evict_lock);
- arc_evict_needed = !zthr_iscancelled(arc_evict_zthr) &&
+ arc_evict_needed = !zthr_iscancelled(zthr) &&
evicted > 0 && aggsum_compare(&arc_sums.arcstat_size, arc_c) > 0;
if (!arc_evict_needed) {
/*
* We're either no longer overflowing, or we
* can't evict anything more, so we should wake
* arc_get_data_impl() sooner.
*/
arc_evict_waiter_t *aw;
while ((aw = list_remove_head(&arc_evict_waiters)) != NULL) {
cv_broadcast(&aw->aew_cv);
}
arc_set_need_free();
}
mutex_exit(&arc_evict_lock);
spl_fstrans_unmark(cookie);
}
static boolean_t
arc_reap_cb_check(void *arg, zthr_t *zthr)
{
(void) arg, (void) zthr;
int64_t free_memory = arc_available_memory();
static int reap_cb_check_counter = 0;
/*
* If a kmem reap is already active, don't schedule more. We must
* check for this because kmem_cache_reap_soon() won't actually
* block on the cache being reaped (this is to prevent callers from
* becoming implicitly blocked by a system-wide kmem reap -- which,
* on a system with many, many full magazines, can take minutes).
*/
if (!kmem_cache_reap_active() && free_memory < 0) {
arc_no_grow = B_TRUE;
arc_warm = B_TRUE;
/*
* Wait at least zfs_grow_retry (default 5) seconds
* before considering growing.
*/
arc_growtime = gethrtime() + SEC2NSEC(arc_grow_retry);
return (B_TRUE);
} else if (free_memory < arc_c >> arc_no_grow_shift) {
arc_no_grow = B_TRUE;
} else if (gethrtime() >= arc_growtime) {
arc_no_grow = B_FALSE;
}
/*
* Called unconditionally every 60 seconds to reclaim unused
* zstd compression and decompression context. This is done
* here to avoid the need for an independent thread.
*/
if (!((reap_cb_check_counter++) % 60))
zfs_zstd_cache_reap_now();
return (B_FALSE);
}
/*
* Keep enough free memory in the system by reaping the ARC's kmem
* caches. To cause more slabs to be reapable, we may reduce the
* target size of the cache (arc_c), causing the arc_evict_cb()
* to free more buffers.
*/
static void
arc_reap_cb(void *arg, zthr_t *zthr)
{
(void) arg, (void) zthr;
int64_t free_memory;
fstrans_cookie_t cookie = spl_fstrans_mark();
/*
* Kick off asynchronous kmem_reap()'s of all our caches.
*/
arc_kmem_reap_soon();
/*
* Wait at least arc_kmem_cache_reap_retry_ms between
* arc_kmem_reap_soon() calls. Without this check it is possible to
* end up in a situation where we spend lots of time reaping
* caches, while we're near arc_c_min. Waiting here also gives the
* subsequent free memory check a chance of finding that the
* asynchronous reap has already freed enough memory, and we don't
* need to call arc_reduce_target_size().
*/
delay((hz * arc_kmem_cache_reap_retry_ms + 999) / 1000);
/*
* Reduce the target size as needed to maintain the amount of free
* memory in the system at a fraction of the arc_size (1/128th by
* default). If oversubscribed (free_memory < 0) then reduce the
* target arc_size by the deficit amount plus the fractional
* amount. If free memory is positive but less than the fractional
* amount, reduce by what is needed to hit the fractional amount.
*/
free_memory = arc_available_memory();
int64_t can_free = arc_c - arc_c_min;
if (can_free > 0) {
int64_t to_free = (can_free >> arc_shrink_shift) - free_memory;
if (to_free > 0)
arc_reduce_target_size(to_free);
}
spl_fstrans_unmark(cookie);
}
#ifdef _KERNEL
/*
* Determine the amount of memory eligible for eviction contained in the
* ARC. All clean data reported by the ghost lists can always be safely
* evicted. Due to arc_c_min, the same does not hold for all clean data
* contained by the regular mru and mfu lists.
*
* In the case of the regular mru and mfu lists, we need to report as
* much clean data as possible, such that evicting that same reported
* data will not bring arc_size below arc_c_min. Thus, in certain
* circumstances, the total amount of clean data in the mru and mfu
* lists might not actually be evictable.
*
* The following two distinct cases are accounted for:
*
* 1. The sum of the amount of dirty data contained by both the mru and
* mfu lists, plus the ARC's other accounting (e.g. the anon list),
* is greater than or equal to arc_c_min.
* (i.e. amount of dirty data >= arc_c_min)
*
* This is the easy case; all clean data contained by the mru and mfu
* lists is evictable. Evicting all clean data can only drop arc_size
* to the amount of dirty data, which is greater than arc_c_min.
*
* 2. The sum of the amount of dirty data contained by both the mru and
* mfu lists, plus the ARC's other accounting (e.g. the anon list),
* is less than arc_c_min.
* (i.e. arc_c_min > amount of dirty data)
*
* 2.1. arc_size is greater than or equal arc_c_min.
* (i.e. arc_size >= arc_c_min > amount of dirty data)
*
* In this case, not all clean data from the regular mru and mfu
* lists is actually evictable; we must leave enough clean data
* to keep arc_size above arc_c_min. Thus, the maximum amount of
* evictable data from the two lists combined, is exactly the
* difference between arc_size and arc_c_min.
*
* 2.2. arc_size is less than arc_c_min
* (i.e. arc_c_min > arc_size > amount of dirty data)
*
* In this case, none of the data contained in the mru and mfu
* lists is evictable, even if it's clean. Since arc_size is
* already below arc_c_min, evicting any more would only
* increase this negative difference.
*/
#endif /* _KERNEL */
/*
* Adapt arc info given the number of bytes we are trying to add and
* the state that we are coming from. This function is only called
* when we are adding new content to the cache.
*/
static void
arc_adapt(uint64_t bytes)
{
/*
* Wake reap thread if we do not have any available memory
*/
if (arc_reclaim_needed()) {
zthr_wakeup(arc_reap_zthr);
return;
}
if (arc_no_grow)
return;
if (arc_c >= arc_c_max)
return;
/*
* If we're within (2 * maxblocksize) bytes of the target
* cache size, increment the target cache size
*/
if (aggsum_upper_bound(&arc_sums.arcstat_size) +
2 * SPA_MAXBLOCKSIZE >= arc_c) {
uint64_t dc = MAX(bytes, SPA_OLD_MAXBLOCKSIZE);
if (atomic_add_64_nv(&arc_c, dc) > arc_c_max)
arc_c = arc_c_max;
}
}
/*
* Check if arc_size has grown past our upper threshold, determined by
* zfs_arc_overflow_shift.
*/
static arc_ovf_level_t
arc_is_overflowing(boolean_t use_reserve)
{
/* Always allow at least one block of overflow */
int64_t overflow = MAX(SPA_MAXBLOCKSIZE,
arc_c >> zfs_arc_overflow_shift);
/*
* We just compare the lower bound here for performance reasons. Our
* primary goals are to make sure that the arc never grows without
* bound, and that it can reach its maximum size. This check
* accomplishes both goals. The maximum amount we could run over by is
* 2 * aggsum_borrow_multiplier * NUM_CPUS * the average size of a block
* in the ARC. In practice, that's in the tens of MB, which is low
* enough to be safe.
*/
int64_t over = aggsum_lower_bound(&arc_sums.arcstat_size) -
arc_c - overflow / 2;
if (!use_reserve)
overflow /= 2;
return (over < 0 ? ARC_OVF_NONE :
over < overflow ? ARC_OVF_SOME : ARC_OVF_SEVERE);
}
static abd_t *
arc_get_data_abd(arc_buf_hdr_t *hdr, uint64_t size, const void *tag,
int alloc_flags)
{
arc_buf_contents_t type = arc_buf_type(hdr);
arc_get_data_impl(hdr, size, tag, alloc_flags);
if (alloc_flags & ARC_HDR_ALLOC_LINEAR)
return (abd_alloc_linear(size, type == ARC_BUFC_METADATA));
else
return (abd_alloc(size, type == ARC_BUFC_METADATA));
}
static void *
arc_get_data_buf(arc_buf_hdr_t *hdr, uint64_t size, const void *tag)
{
arc_buf_contents_t type = arc_buf_type(hdr);
arc_get_data_impl(hdr, size, tag, 0);
if (type == ARC_BUFC_METADATA) {
return (zio_buf_alloc(size));
} else {
ASSERT(type == ARC_BUFC_DATA);
return (zio_data_buf_alloc(size));
}
}
/*
* Wait for the specified amount of data (in bytes) to be evicted from the
* ARC, and for there to be sufficient free memory in the system. Waiting for
* eviction ensures that the memory used by the ARC decreases. Waiting for
* free memory ensures that the system won't run out of free pages, regardless
* of ARC behavior and settings. See arc_lowmem_init().
*/
void
arc_wait_for_eviction(uint64_t amount, boolean_t use_reserve)
{
switch (arc_is_overflowing(use_reserve)) {
case ARC_OVF_NONE:
return;
case ARC_OVF_SOME:
/*
* This is a bit racy without taking arc_evict_lock, but the
* worst that can happen is we either call zthr_wakeup() extra
* time due to race with other thread here, or the set flag
* get cleared by arc_evict_cb(), which is unlikely due to
* big hysteresis, but also not important since at this level
* of overflow the eviction is purely advisory. Same time
* taking the global lock here every time without waiting for
* the actual eviction creates a significant lock contention.
*/
if (!arc_evict_needed) {
arc_evict_needed = B_TRUE;
zthr_wakeup(arc_evict_zthr);
}
return;
case ARC_OVF_SEVERE:
default:
{
arc_evict_waiter_t aw;
list_link_init(&aw.aew_node);
cv_init(&aw.aew_cv, NULL, CV_DEFAULT, NULL);
uint64_t last_count = 0;
mutex_enter(&arc_evict_lock);
if (!list_is_empty(&arc_evict_waiters)) {
arc_evict_waiter_t *last =
list_tail(&arc_evict_waiters);
last_count = last->aew_count;
} else if (!arc_evict_needed) {
arc_evict_needed = B_TRUE;
zthr_wakeup(arc_evict_zthr);
}
/*
* Note, the last waiter's count may be less than
* arc_evict_count if we are low on memory in which
* case arc_evict_state_impl() may have deferred
* wakeups (but still incremented arc_evict_count).
*/
aw.aew_count = MAX(last_count, arc_evict_count) + amount;
list_insert_tail(&arc_evict_waiters, &aw);
arc_set_need_free();
DTRACE_PROBE3(arc__wait__for__eviction,
uint64_t, amount,
uint64_t, arc_evict_count,
uint64_t, aw.aew_count);
/*
* We will be woken up either when arc_evict_count reaches
* aew_count, or when the ARC is no longer overflowing and
* eviction completes.
* In case of "false" wakeup, we will still be on the list.
*/
do {
cv_wait(&aw.aew_cv, &arc_evict_lock);
} while (list_link_active(&aw.aew_node));
mutex_exit(&arc_evict_lock);
cv_destroy(&aw.aew_cv);
}
}
}
/*
* Allocate a block and return it to the caller. If we are hitting the
* hard limit for the cache size, we must sleep, waiting for the eviction
* thread to catch up. If we're past the target size but below the hard
* limit, we'll only signal the reclaim thread and continue on.
*/
static void
arc_get_data_impl(arc_buf_hdr_t *hdr, uint64_t size, const void *tag,
int alloc_flags)
{
arc_adapt(size);
/*
* If arc_size is currently overflowing, we must be adding data
* faster than we are evicting. To ensure we don't compound the
* problem by adding more data and forcing arc_size to grow even
* further past it's target size, we wait for the eviction thread to
* make some progress. We also wait for there to be sufficient free
* memory in the system, as measured by arc_free_memory().
*
* Specifically, we wait for zfs_arc_eviction_pct percent of the
* requested size to be evicted. This should be more than 100%, to
* ensure that that progress is also made towards getting arc_size
* under arc_c. See the comment above zfs_arc_eviction_pct.
*/
arc_wait_for_eviction(size * zfs_arc_eviction_pct / 100,
alloc_flags & ARC_HDR_USE_RESERVE);
arc_buf_contents_t type = arc_buf_type(hdr);
if (type == ARC_BUFC_METADATA) {
arc_space_consume(size, ARC_SPACE_META);
} else {
arc_space_consume(size, ARC_SPACE_DATA);
}
/*
* Update the state size. Note that ghost states have a
* "ghost size" and so don't need to be updated.
*/
arc_state_t *state = hdr->b_l1hdr.b_state;
if (!GHOST_STATE(state)) {
(void) zfs_refcount_add_many(&state->arcs_size[type], size,
tag);
/*
* If this is reached via arc_read, the link is
* protected by the hash lock. If reached via
* arc_buf_alloc, the header should not be accessed by
* any other thread. And, if reached via arc_read_done,
* the hash lock will protect it if it's found in the
* hash table; otherwise no other thread should be
* trying to [add|remove]_reference it.
*/
if (multilist_link_active(&hdr->b_l1hdr.b_arc_node)) {
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
(void) zfs_refcount_add_many(&state->arcs_esize[type],
size, tag);
}
}
}
static void
arc_free_data_abd(arc_buf_hdr_t *hdr, abd_t *abd, uint64_t size,
const void *tag)
{
arc_free_data_impl(hdr, size, tag);
abd_free(abd);
}
static void
arc_free_data_buf(arc_buf_hdr_t *hdr, void *buf, uint64_t size, const void *tag)
{
arc_buf_contents_t type = arc_buf_type(hdr);
arc_free_data_impl(hdr, size, tag);
if (type == ARC_BUFC_METADATA) {
zio_buf_free(buf, size);
} else {
ASSERT(type == ARC_BUFC_DATA);
zio_data_buf_free(buf, size);
}
}
/*
* Free the arc data buffer.
*/
static void
arc_free_data_impl(arc_buf_hdr_t *hdr, uint64_t size, const void *tag)
{
arc_state_t *state = hdr->b_l1hdr.b_state;
arc_buf_contents_t type = arc_buf_type(hdr);
/* protected by hash lock, if in the hash table */
if (multilist_link_active(&hdr->b_l1hdr.b_arc_node)) {
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
ASSERT(state != arc_anon && state != arc_l2c_only);
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
size, tag);
}
(void) zfs_refcount_remove_many(&state->arcs_size[type], size, tag);
VERIFY3U(hdr->b_type, ==, type);
if (type == ARC_BUFC_METADATA) {
arc_space_return(size, ARC_SPACE_META);
} else {
ASSERT(type == ARC_BUFC_DATA);
arc_space_return(size, ARC_SPACE_DATA);
}
}
/*
* This routine is called whenever a buffer is accessed.
*/
static void
arc_access(arc_buf_hdr_t *hdr, arc_flags_t arc_flags, boolean_t hit)
{
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT(HDR_HAS_L1HDR(hdr));
/*
* Update buffer prefetch status.
*/
boolean_t was_prefetch = HDR_PREFETCH(hdr);
boolean_t now_prefetch = arc_flags & ARC_FLAG_PREFETCH;
if (was_prefetch != now_prefetch) {
if (was_prefetch) {
ARCSTAT_CONDSTAT(hit, demand_hit, demand_iohit,
HDR_PRESCIENT_PREFETCH(hdr), prescient, predictive,
prefetch);
}
if (HDR_HAS_L2HDR(hdr))
l2arc_hdr_arcstats_decrement_state(hdr);
if (was_prefetch) {
arc_hdr_clear_flags(hdr,
ARC_FLAG_PREFETCH | ARC_FLAG_PRESCIENT_PREFETCH);
} else {
arc_hdr_set_flags(hdr, ARC_FLAG_PREFETCH);
}
if (HDR_HAS_L2HDR(hdr))
l2arc_hdr_arcstats_increment_state(hdr);
}
if (now_prefetch) {
if (arc_flags & ARC_FLAG_PRESCIENT_PREFETCH) {
arc_hdr_set_flags(hdr, ARC_FLAG_PRESCIENT_PREFETCH);
ARCSTAT_BUMP(arcstat_prescient_prefetch);
} else {
ARCSTAT_BUMP(arcstat_predictive_prefetch);
}
}
if (arc_flags & ARC_FLAG_L2CACHE)
arc_hdr_set_flags(hdr, ARC_FLAG_L2CACHE);
clock_t now = ddi_get_lbolt();
if (hdr->b_l1hdr.b_state == arc_anon) {
arc_state_t *new_state;
/*
* This buffer is not in the cache, and does not appear in
* our "ghost" lists. Add it to the MRU or uncached state.
*/
ASSERT0(hdr->b_l1hdr.b_arc_access);
hdr->b_l1hdr.b_arc_access = now;
if (HDR_UNCACHED(hdr)) {
new_state = arc_uncached;
DTRACE_PROBE1(new_state__uncached, arc_buf_hdr_t *,
hdr);
} else {
new_state = arc_mru;
DTRACE_PROBE1(new_state__mru, arc_buf_hdr_t *, hdr);
}
arc_change_state(new_state, hdr);
} else if (hdr->b_l1hdr.b_state == arc_mru) {
/*
* This buffer has been accessed once recently and either
* its read is still in progress or it is in the cache.
*/
if (HDR_IO_IN_PROGRESS(hdr)) {
hdr->b_l1hdr.b_arc_access = now;
return;
}
hdr->b_l1hdr.b_mru_hits++;
ARCSTAT_BUMP(arcstat_mru_hits);
/*
* If the previous access was a prefetch, then it already
* handled possible promotion, so nothing more to do for now.
*/
if (was_prefetch) {
hdr->b_l1hdr.b_arc_access = now;
return;
}
/*
* If more than ARC_MINTIME have passed from the previous
* hit, promote the buffer to the MFU state.
*/
if (ddi_time_after(now, hdr->b_l1hdr.b_arc_access +
ARC_MINTIME)) {
hdr->b_l1hdr.b_arc_access = now;
DTRACE_PROBE1(new_state__mfu, arc_buf_hdr_t *, hdr);
arc_change_state(arc_mfu, hdr);
}
} else if (hdr->b_l1hdr.b_state == arc_mru_ghost) {
arc_state_t *new_state;
/*
* This buffer has been accessed once recently, but was
* evicted from the cache. Would we have bigger MRU, it
* would be an MRU hit, so handle it the same way, except
* we don't need to check the previous access time.
*/
hdr->b_l1hdr.b_mru_ghost_hits++;
ARCSTAT_BUMP(arcstat_mru_ghost_hits);
hdr->b_l1hdr.b_arc_access = now;
wmsum_add(&arc_mru_ghost->arcs_hits[arc_buf_type(hdr)],
arc_hdr_size(hdr));
if (was_prefetch) {
new_state = arc_mru;
DTRACE_PROBE1(new_state__mru, arc_buf_hdr_t *, hdr);
} else {
new_state = arc_mfu;
DTRACE_PROBE1(new_state__mfu, arc_buf_hdr_t *, hdr);
}
arc_change_state(new_state, hdr);
} else if (hdr->b_l1hdr.b_state == arc_mfu) {
/*
* This buffer has been accessed more than once and either
* still in the cache or being restored from one of ghosts.
*/
if (!HDR_IO_IN_PROGRESS(hdr)) {
hdr->b_l1hdr.b_mfu_hits++;
ARCSTAT_BUMP(arcstat_mfu_hits);
}
hdr->b_l1hdr.b_arc_access = now;
} else if (hdr->b_l1hdr.b_state == arc_mfu_ghost) {
/*
* This buffer has been accessed more than once recently, but
* has been evicted from the cache. Would we have bigger MFU
* it would stay in cache, so move it back to MFU state.
*/
hdr->b_l1hdr.b_mfu_ghost_hits++;
ARCSTAT_BUMP(arcstat_mfu_ghost_hits);
hdr->b_l1hdr.b_arc_access = now;
wmsum_add(&arc_mfu_ghost->arcs_hits[arc_buf_type(hdr)],
arc_hdr_size(hdr));
DTRACE_PROBE1(new_state__mfu, arc_buf_hdr_t *, hdr);
arc_change_state(arc_mfu, hdr);
} else if (hdr->b_l1hdr.b_state == arc_uncached) {
/*
* This buffer is uncacheable, but we got a hit. Probably
* a demand read after prefetch. Nothing more to do here.
*/
if (!HDR_IO_IN_PROGRESS(hdr))
ARCSTAT_BUMP(arcstat_uncached_hits);
hdr->b_l1hdr.b_arc_access = now;
} else if (hdr->b_l1hdr.b_state == arc_l2c_only) {
/*
* This buffer is on the 2nd Level ARC and was not accessed
* for a long time, so treat it as new and put into MRU.
*/
hdr->b_l1hdr.b_arc_access = now;
DTRACE_PROBE1(new_state__mru, arc_buf_hdr_t *, hdr);
arc_change_state(arc_mru, hdr);
} else {
cmn_err(CE_PANIC, "invalid arc state 0x%p",
hdr->b_l1hdr.b_state);
}
}
/*
* This routine is called by dbuf_hold() to update the arc_access() state
* which otherwise would be skipped for entries in the dbuf cache.
*/
void
arc_buf_access(arc_buf_t *buf)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
/*
* Avoid taking the hash_lock when possible as an optimization.
* The header must be checked again under the hash_lock in order
* to handle the case where it is concurrently being released.
*/
if (hdr->b_l1hdr.b_state == arc_anon || HDR_EMPTY(hdr))
return;
kmutex_t *hash_lock = HDR_LOCK(hdr);
mutex_enter(hash_lock);
if (hdr->b_l1hdr.b_state == arc_anon || HDR_EMPTY(hdr)) {
mutex_exit(hash_lock);
ARCSTAT_BUMP(arcstat_access_skip);
return;
}
ASSERT(hdr->b_l1hdr.b_state == arc_mru ||
hdr->b_l1hdr.b_state == arc_mfu ||
hdr->b_l1hdr.b_state == arc_uncached);
DTRACE_PROBE1(arc__hit, arc_buf_hdr_t *, hdr);
arc_access(hdr, 0, B_TRUE);
mutex_exit(hash_lock);
ARCSTAT_BUMP(arcstat_hits);
ARCSTAT_CONDSTAT(B_TRUE /* demand */, demand, prefetch,
!HDR_ISTYPE_METADATA(hdr), data, metadata, hits);
}
/* a generic arc_read_done_func_t which you can use */
void
arc_bcopy_func(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp,
arc_buf_t *buf, void *arg)
{
(void) zio, (void) zb, (void) bp;
if (buf == NULL)
return;
memcpy(arg, buf->b_data, arc_buf_size(buf));
arc_buf_destroy(buf, arg);
}
/* a generic arc_read_done_func_t */
void
arc_getbuf_func(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp,
arc_buf_t *buf, void *arg)
{
(void) zb, (void) bp;
arc_buf_t **bufp = arg;
if (buf == NULL) {
ASSERT(zio == NULL || zio->io_error != 0);
*bufp = NULL;
} else {
ASSERT(zio == NULL || zio->io_error == 0);
*bufp = buf;
ASSERT(buf->b_data != NULL);
}
}
static void
arc_hdr_verify(arc_buf_hdr_t *hdr, blkptr_t *bp)
{
if (BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp)) {
ASSERT3U(HDR_GET_PSIZE(hdr), ==, 0);
ASSERT3U(arc_hdr_get_compress(hdr), ==, ZIO_COMPRESS_OFF);
} else {
if (HDR_COMPRESSION_ENABLED(hdr)) {
ASSERT3U(arc_hdr_get_compress(hdr), ==,
BP_GET_COMPRESS(bp));
}
ASSERT3U(HDR_GET_LSIZE(hdr), ==, BP_GET_LSIZE(bp));
ASSERT3U(HDR_GET_PSIZE(hdr), ==, BP_GET_PSIZE(bp));
ASSERT3U(!!HDR_PROTECTED(hdr), ==, BP_IS_PROTECTED(bp));
}
}
static void
arc_read_done(zio_t *zio)
{
blkptr_t *bp = zio->io_bp;
arc_buf_hdr_t *hdr = zio->io_private;
kmutex_t *hash_lock = NULL;
arc_callback_t *callback_list;
arc_callback_t *acb;
/*
* The hdr was inserted into hash-table and removed from lists
* prior to starting I/O. We should find this header, since
* it's in the hash table, and it should be legit since it's
* not possible to evict it during the I/O. The only possible
* reason for it not to be found is if we were freed during the
* read.
*/
if (HDR_IN_HASH_TABLE(hdr)) {
arc_buf_hdr_t *found;
ASSERT3U(hdr->b_birth, ==, BP_PHYSICAL_BIRTH(zio->io_bp));
ASSERT3U(hdr->b_dva.dva_word[0], ==,
BP_IDENTITY(zio->io_bp)->dva_word[0]);
ASSERT3U(hdr->b_dva.dva_word[1], ==,
BP_IDENTITY(zio->io_bp)->dva_word[1]);
found = buf_hash_find(hdr->b_spa, zio->io_bp, &hash_lock);
ASSERT((found == hdr &&
DVA_EQUAL(&hdr->b_dva, BP_IDENTITY(zio->io_bp))) ||
(found == hdr && HDR_L2_READING(hdr)));
ASSERT3P(hash_lock, !=, NULL);
}
if (BP_IS_PROTECTED(bp)) {
hdr->b_crypt_hdr.b_ot = BP_GET_TYPE(bp);
hdr->b_crypt_hdr.b_dsobj = zio->io_bookmark.zb_objset;
zio_crypt_decode_params_bp(bp, hdr->b_crypt_hdr.b_salt,
hdr->b_crypt_hdr.b_iv);
if (zio->io_error == 0) {
if (BP_GET_TYPE(bp) == DMU_OT_INTENT_LOG) {
void *tmpbuf;
tmpbuf = abd_borrow_buf_copy(zio->io_abd,
sizeof (zil_chain_t));
zio_crypt_decode_mac_zil(tmpbuf,
hdr->b_crypt_hdr.b_mac);
abd_return_buf(zio->io_abd, tmpbuf,
sizeof (zil_chain_t));
} else {
zio_crypt_decode_mac_bp(bp,
hdr->b_crypt_hdr.b_mac);
}
}
}
if (zio->io_error == 0) {
/* byteswap if necessary */
if (BP_SHOULD_BYTESWAP(zio->io_bp)) {
if (BP_GET_LEVEL(zio->io_bp) > 0) {
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_UINT64;
} else {
hdr->b_l1hdr.b_byteswap =
DMU_OT_BYTESWAP(BP_GET_TYPE(zio->io_bp));
}
} else {
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_NUMFUNCS;
}
if (!HDR_L2_READING(hdr)) {
hdr->b_complevel = zio->io_prop.zp_complevel;
}
}
arc_hdr_clear_flags(hdr, ARC_FLAG_L2_EVICTED);
if (l2arc_noprefetch && HDR_PREFETCH(hdr))
arc_hdr_clear_flags(hdr, ARC_FLAG_L2CACHE);
callback_list = hdr->b_l1hdr.b_acb;
ASSERT3P(callback_list, !=, NULL);
hdr->b_l1hdr.b_acb = NULL;
/*
* If a read request has a callback (i.e. acb_done is not NULL), then we
* make a buf containing the data according to the parameters which were
* passed in. The implementation of arc_buf_alloc_impl() ensures that we
* aren't needlessly decompressing the data multiple times.
*/
int callback_cnt = 0;
for (acb = callback_list; acb != NULL; acb = acb->acb_next) {
/* We need the last one to call below in original order. */
callback_list = acb;
if (!acb->acb_done || acb->acb_nobuf)
continue;
callback_cnt++;
if (zio->io_error != 0)
continue;
int error = arc_buf_alloc_impl(hdr, zio->io_spa,
&acb->acb_zb, acb->acb_private, acb->acb_encrypted,
acb->acb_compressed, acb->acb_noauth, B_TRUE,
&acb->acb_buf);
/*
* Assert non-speculative zios didn't fail because an
* encryption key wasn't loaded
*/
ASSERT((zio->io_flags & ZIO_FLAG_SPECULATIVE) ||
error != EACCES);
/*
* If we failed to decrypt, report an error now (as the zio
* layer would have done if it had done the transforms).
*/
if (error == ECKSUM) {
ASSERT(BP_IS_PROTECTED(bp));
error = SET_ERROR(EIO);
if ((zio->io_flags & ZIO_FLAG_SPECULATIVE) == 0) {
spa_log_error(zio->io_spa, &acb->acb_zb,
&zio->io_bp->blk_birth);
(void) zfs_ereport_post(
FM_EREPORT_ZFS_AUTHENTICATION,
zio->io_spa, NULL, &acb->acb_zb, zio, 0);
}
}
if (error != 0) {
/*
* Decompression or decryption failed. Set
* io_error so that when we call acb_done
* (below), we will indicate that the read
* failed. Note that in the unusual case
* where one callback is compressed and another
* uncompressed, we will mark all of them
* as failed, even though the uncompressed
* one can't actually fail. In this case,
* the hdr will not be anonymous, because
* if there are multiple callbacks, it's
* because multiple threads found the same
* arc buf in the hash table.
*/
zio->io_error = error;
}
}
/*
* If there are multiple callbacks, we must have the hash lock,
* because the only way for multiple threads to find this hdr is
* in the hash table. This ensures that if there are multiple
* callbacks, the hdr is not anonymous. If it were anonymous,
* we couldn't use arc_buf_destroy() in the error case below.
*/
ASSERT(callback_cnt < 2 || hash_lock != NULL);
if (zio->io_error == 0) {
arc_hdr_verify(hdr, zio->io_bp);
} else {
arc_hdr_set_flags(hdr, ARC_FLAG_IO_ERROR);
if (hdr->b_l1hdr.b_state != arc_anon)
arc_change_state(arc_anon, hdr);
if (HDR_IN_HASH_TABLE(hdr))
buf_hash_remove(hdr);
}
arc_hdr_clear_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
(void) remove_reference(hdr, hdr);
if (hash_lock != NULL)
mutex_exit(hash_lock);
/* execute each callback and free its structure */
while ((acb = callback_list) != NULL) {
if (acb->acb_done != NULL) {
if (zio->io_error != 0 && acb->acb_buf != NULL) {
/*
* If arc_buf_alloc_impl() fails during
* decompression, the buf will still be
* allocated, and needs to be freed here.
*/
arc_buf_destroy(acb->acb_buf,
acb->acb_private);
acb->acb_buf = NULL;
}
acb->acb_done(zio, &zio->io_bookmark, zio->io_bp,
acb->acb_buf, acb->acb_private);
}
if (acb->acb_zio_dummy != NULL) {
acb->acb_zio_dummy->io_error = zio->io_error;
zio_nowait(acb->acb_zio_dummy);
}
callback_list = acb->acb_prev;
if (acb->acb_wait) {
mutex_enter(&acb->acb_wait_lock);
acb->acb_wait_error = zio->io_error;
acb->acb_wait = B_FALSE;
cv_signal(&acb->acb_wait_cv);
mutex_exit(&acb->acb_wait_lock);
/* acb will be freed by the waiting thread. */
} else {
kmem_free(acb, sizeof (arc_callback_t));
}
}
}
/*
* "Read" the block at the specified DVA (in bp) via the
* cache. If the block is found in the cache, invoke the provided
* callback immediately and return. Note that the `zio' parameter
* in the callback will be NULL in this case, since no IO was
* required. If the block is not in the cache pass the read request
* on to the spa with a substitute callback function, so that the
* requested block will be added to the cache.
*
* If a read request arrives for a block that has a read in-progress,
* either wait for the in-progress read to complete (and return the
* results); or, if this is a read with a "done" func, add a record
* to the read to invoke the "done" func when the read completes,
* and return; or just return.
*
* arc_read_done() will invoke all the requested "done" functions
* for readers of this block.
*/
int
arc_read(zio_t *pio, spa_t *spa, const blkptr_t *bp,
arc_read_done_func_t *done, void *private, zio_priority_t priority,
int zio_flags, arc_flags_t *arc_flags, const zbookmark_phys_t *zb)
{
arc_buf_hdr_t *hdr = NULL;
kmutex_t *hash_lock = NULL;
zio_t *rzio;
uint64_t guid = spa_load_guid(spa);
boolean_t compressed_read = (zio_flags & ZIO_FLAG_RAW_COMPRESS) != 0;
boolean_t encrypted_read = BP_IS_ENCRYPTED(bp) &&
(zio_flags & ZIO_FLAG_RAW_ENCRYPT) != 0;
boolean_t noauth_read = BP_IS_AUTHENTICATED(bp) &&
(zio_flags & ZIO_FLAG_RAW_ENCRYPT) != 0;
boolean_t embedded_bp = !!BP_IS_EMBEDDED(bp);
boolean_t no_buf = *arc_flags & ARC_FLAG_NO_BUF;
arc_buf_t *buf = NULL;
int rc = 0;
ASSERT(!embedded_bp ||
BPE_GET_ETYPE(bp) == BP_EMBEDDED_TYPE_DATA);
ASSERT(!BP_IS_HOLE(bp));
ASSERT(!BP_IS_REDACTED(bp));
/*
* Normally SPL_FSTRANS will already be set since kernel threads which
* expect to call the DMU interfaces will set it when created. System
* calls are similarly handled by setting/cleaning the bit in the
* registered callback (module/os/.../zfs/zpl_*).
*
* External consumers such as Lustre which call the exported DMU
* interfaces may not have set SPL_FSTRANS. To avoid a deadlock
* on the hash_lock always set and clear the bit.
*/
fstrans_cookie_t cookie = spl_fstrans_mark();
top:
/*
* Verify the block pointer contents are reasonable. This should
* always be the case since the blkptr is protected by a checksum.
* However, if there is damage it's desirable to detect this early
* and treat it as a checksum error. This allows an alternate blkptr
* to be tried when one is available (e.g. ditto blocks).
*/
if (!zfs_blkptr_verify(spa, bp, (zio_flags & ZIO_FLAG_CONFIG_WRITER) ?
BLK_CONFIG_HELD : BLK_CONFIG_NEEDED, BLK_VERIFY_LOG)) {
rc = SET_ERROR(ECKSUM);
goto done;
}
if (!embedded_bp) {
/*
* Embedded BP's have no DVA and require no I/O to "read".
* Create an anonymous arc buf to back it.
*/
hdr = buf_hash_find(guid, bp, &hash_lock);
}
/*
* Determine if we have an L1 cache hit or a cache miss. For simplicity
* we maintain encrypted data separately from compressed / uncompressed
* data. If the user is requesting raw encrypted data and we don't have
* that in the header we will read from disk to guarantee that we can
* get it even if the encryption keys aren't loaded.
*/
if (hdr != NULL && HDR_HAS_L1HDR(hdr) && (HDR_HAS_RABD(hdr) ||
(hdr->b_l1hdr.b_pabd != NULL && !encrypted_read))) {
boolean_t is_data = !HDR_ISTYPE_METADATA(hdr);
if (HDR_IO_IN_PROGRESS(hdr)) {
if (*arc_flags & ARC_FLAG_CACHED_ONLY) {
mutex_exit(hash_lock);
ARCSTAT_BUMP(arcstat_cached_only_in_progress);
rc = SET_ERROR(ENOENT);
goto done;
}
zio_t *head_zio = hdr->b_l1hdr.b_acb->acb_zio_head;
ASSERT3P(head_zio, !=, NULL);
if ((hdr->b_flags & ARC_FLAG_PRIO_ASYNC_READ) &&
priority == ZIO_PRIORITY_SYNC_READ) {
/*
* This is a sync read that needs to wait for
* an in-flight async read. Request that the
* zio have its priority upgraded.
*/
zio_change_priority(head_zio, priority);
DTRACE_PROBE1(arc__async__upgrade__sync,
arc_buf_hdr_t *, hdr);
ARCSTAT_BUMP(arcstat_async_upgrade_sync);
}
DTRACE_PROBE1(arc__iohit, arc_buf_hdr_t *, hdr);
arc_access(hdr, *arc_flags, B_FALSE);
/*
* If there are multiple threads reading the same block
* and that block is not yet in the ARC, then only one
* thread will do the physical I/O and all other
* threads will wait until that I/O completes.
* Synchronous reads use the acb_wait_cv whereas nowait
* reads register a callback. Both are signalled/called
* in arc_read_done.
*
* Errors of the physical I/O may need to be propagated.
* Synchronous read errors are returned here from
* arc_read_done via acb_wait_error. Nowait reads
* attach the acb_zio_dummy zio to pio and
* arc_read_done propagates the physical I/O's io_error
* to acb_zio_dummy, and thereby to pio.
*/
arc_callback_t *acb = NULL;
if (done || pio || *arc_flags & ARC_FLAG_WAIT) {
acb = kmem_zalloc(sizeof (arc_callback_t),
KM_SLEEP);
acb->acb_done = done;
acb->acb_private = private;
acb->acb_compressed = compressed_read;
acb->acb_encrypted = encrypted_read;
acb->acb_noauth = noauth_read;
acb->acb_nobuf = no_buf;
if (*arc_flags & ARC_FLAG_WAIT) {
acb->acb_wait = B_TRUE;
mutex_init(&acb->acb_wait_lock, NULL,
MUTEX_DEFAULT, NULL);
cv_init(&acb->acb_wait_cv, NULL,
CV_DEFAULT, NULL);
}
acb->acb_zb = *zb;
if (pio != NULL) {
acb->acb_zio_dummy = zio_null(pio,
spa, NULL, NULL, NULL, zio_flags);
}
acb->acb_zio_head = head_zio;
acb->acb_next = hdr->b_l1hdr.b_acb;
hdr->b_l1hdr.b_acb->acb_prev = acb;
hdr->b_l1hdr.b_acb = acb;
}
mutex_exit(hash_lock);
ARCSTAT_BUMP(arcstat_iohits);
ARCSTAT_CONDSTAT(!(*arc_flags & ARC_FLAG_PREFETCH),
demand, prefetch, is_data, data, metadata, iohits);
if (*arc_flags & ARC_FLAG_WAIT) {
mutex_enter(&acb->acb_wait_lock);
while (acb->acb_wait) {
cv_wait(&acb->acb_wait_cv,
&acb->acb_wait_lock);
}
rc = acb->acb_wait_error;
mutex_exit(&acb->acb_wait_lock);
mutex_destroy(&acb->acb_wait_lock);
cv_destroy(&acb->acb_wait_cv);
kmem_free(acb, sizeof (arc_callback_t));
}
goto out;
}
ASSERT(hdr->b_l1hdr.b_state == arc_mru ||
hdr->b_l1hdr.b_state == arc_mfu ||
hdr->b_l1hdr.b_state == arc_uncached);
DTRACE_PROBE1(arc__hit, arc_buf_hdr_t *, hdr);
arc_access(hdr, *arc_flags, B_TRUE);
if (done && !no_buf) {
ASSERT(!embedded_bp || !BP_IS_HOLE(bp));
/* Get a buf with the desired data in it. */
rc = arc_buf_alloc_impl(hdr, spa, zb, private,
encrypted_read, compressed_read, noauth_read,
B_TRUE, &buf);
if (rc == ECKSUM) {
/*
* Convert authentication and decryption errors
* to EIO (and generate an ereport if needed)
* before leaving the ARC.
*/
rc = SET_ERROR(EIO);
if ((zio_flags & ZIO_FLAG_SPECULATIVE) == 0) {
spa_log_error(spa, zb, &hdr->b_birth);
(void) zfs_ereport_post(
FM_EREPORT_ZFS_AUTHENTICATION,
spa, NULL, zb, NULL, 0);
}
}
if (rc != 0) {
arc_buf_destroy_impl(buf);
buf = NULL;
(void) remove_reference(hdr, private);
}
/* assert any errors weren't due to unloaded keys */
ASSERT((zio_flags & ZIO_FLAG_SPECULATIVE) ||
rc != EACCES);
}
mutex_exit(hash_lock);
ARCSTAT_BUMP(arcstat_hits);
ARCSTAT_CONDSTAT(!(*arc_flags & ARC_FLAG_PREFETCH),
demand, prefetch, is_data, data, metadata, hits);
*arc_flags |= ARC_FLAG_CACHED;
goto done;
} else {
uint64_t lsize = BP_GET_LSIZE(bp);
uint64_t psize = BP_GET_PSIZE(bp);
arc_callback_t *acb;
vdev_t *vd = NULL;
uint64_t addr = 0;
boolean_t devw = B_FALSE;
uint64_t size;
abd_t *hdr_abd;
int alloc_flags = encrypted_read ? ARC_HDR_ALLOC_RDATA : 0;
arc_buf_contents_t type = BP_GET_BUFC_TYPE(bp);
if (*arc_flags & ARC_FLAG_CACHED_ONLY) {
if (hash_lock != NULL)
mutex_exit(hash_lock);
rc = SET_ERROR(ENOENT);
goto done;
}
if (hdr == NULL) {
/*
* This block is not in the cache or it has
* embedded data.
*/
arc_buf_hdr_t *exists = NULL;
hdr = arc_hdr_alloc(spa_load_guid(spa), psize, lsize,
BP_IS_PROTECTED(bp), BP_GET_COMPRESS(bp), 0, type);
if (!embedded_bp) {
hdr->b_dva = *BP_IDENTITY(bp);
hdr->b_birth = BP_PHYSICAL_BIRTH(bp);
exists = buf_hash_insert(hdr, &hash_lock);
}
if (exists != NULL) {
/* somebody beat us to the hash insert */
mutex_exit(hash_lock);
buf_discard_identity(hdr);
arc_hdr_destroy(hdr);
goto top; /* restart the IO request */
}
} else {
/*
* This block is in the ghost cache or encrypted data
* was requested and we didn't have it. If it was
* L2-only (and thus didn't have an L1 hdr),
* we realloc the header to add an L1 hdr.
*/
if (!HDR_HAS_L1HDR(hdr)) {
hdr = arc_hdr_realloc(hdr, hdr_l2only_cache,
hdr_full_cache);
}
if (GHOST_STATE(hdr->b_l1hdr.b_state)) {
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT0(zfs_refcount_count(
&hdr->b_l1hdr.b_refcnt));
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
#ifdef ZFS_DEBUG
ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
#endif
} else if (HDR_IO_IN_PROGRESS(hdr)) {
/*
* If this header already had an IO in progress
* and we are performing another IO to fetch
* encrypted data we must wait until the first
* IO completes so as not to confuse
* arc_read_done(). This should be very rare
* and so the performance impact shouldn't
* matter.
*/
arc_callback_t *acb = kmem_zalloc(
sizeof (arc_callback_t), KM_SLEEP);
acb->acb_wait = B_TRUE;
mutex_init(&acb->acb_wait_lock, NULL,
MUTEX_DEFAULT, NULL);
cv_init(&acb->acb_wait_cv, NULL, CV_DEFAULT,
NULL);
acb->acb_zio_head =
hdr->b_l1hdr.b_acb->acb_zio_head;
acb->acb_next = hdr->b_l1hdr.b_acb;
hdr->b_l1hdr.b_acb->acb_prev = acb;
hdr->b_l1hdr.b_acb = acb;
mutex_exit(hash_lock);
mutex_enter(&acb->acb_wait_lock);
while (acb->acb_wait) {
cv_wait(&acb->acb_wait_cv,
&acb->acb_wait_lock);
}
mutex_exit(&acb->acb_wait_lock);
mutex_destroy(&acb->acb_wait_lock);
cv_destroy(&acb->acb_wait_cv);
kmem_free(acb, sizeof (arc_callback_t));
goto top;
}
}
if (*arc_flags & ARC_FLAG_UNCACHED) {
arc_hdr_set_flags(hdr, ARC_FLAG_UNCACHED);
if (!encrypted_read)
alloc_flags |= ARC_HDR_ALLOC_LINEAR;
}
/*
* Take additional reference for IO_IN_PROGRESS. It stops
* arc_access() from putting this header without any buffers
* and so other references but obviously nonevictable onto
* the evictable list of MRU or MFU state.
*/
add_reference(hdr, hdr);
if (!embedded_bp)
arc_access(hdr, *arc_flags, B_FALSE);
arc_hdr_set_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
arc_hdr_alloc_abd(hdr, alloc_flags);
if (encrypted_read) {
ASSERT(HDR_HAS_RABD(hdr));
size = HDR_GET_PSIZE(hdr);
hdr_abd = hdr->b_crypt_hdr.b_rabd;
zio_flags |= ZIO_FLAG_RAW;
} else {
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
size = arc_hdr_size(hdr);
hdr_abd = hdr->b_l1hdr.b_pabd;
if (arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF) {
zio_flags |= ZIO_FLAG_RAW_COMPRESS;
}
/*
* For authenticated bp's, we do not ask the ZIO layer
* to authenticate them since this will cause the entire
* IO to fail if the key isn't loaded. Instead, we
* defer authentication until arc_buf_fill(), which will
* verify the data when the key is available.
*/
if (BP_IS_AUTHENTICATED(bp))
zio_flags |= ZIO_FLAG_RAW_ENCRYPT;
}
if (BP_IS_AUTHENTICATED(bp))
arc_hdr_set_flags(hdr, ARC_FLAG_NOAUTH);
if (BP_GET_LEVEL(bp) > 0)
arc_hdr_set_flags(hdr, ARC_FLAG_INDIRECT);
ASSERT(!GHOST_STATE(hdr->b_l1hdr.b_state));
acb = kmem_zalloc(sizeof (arc_callback_t), KM_SLEEP);
acb->acb_done = done;
acb->acb_private = private;
acb->acb_compressed = compressed_read;
acb->acb_encrypted = encrypted_read;
acb->acb_noauth = noauth_read;
acb->acb_zb = *zb;
ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
hdr->b_l1hdr.b_acb = acb;
if (HDR_HAS_L2HDR(hdr) &&
(vd = hdr->b_l2hdr.b_dev->l2ad_vdev) != NULL) {
devw = hdr->b_l2hdr.b_dev->l2ad_writing;
addr = hdr->b_l2hdr.b_daddr;
/*
* Lock out L2ARC device removal.
*/
if (vdev_is_dead(vd) ||
!spa_config_tryenter(spa, SCL_L2ARC, vd, RW_READER))
vd = NULL;
}
/*
* We count both async reads and scrub IOs as asynchronous so
* that both can be upgraded in the event of a cache hit while
* the read IO is still in-flight.
*/
if (priority == ZIO_PRIORITY_ASYNC_READ ||
priority == ZIO_PRIORITY_SCRUB)
arc_hdr_set_flags(hdr, ARC_FLAG_PRIO_ASYNC_READ);
else
arc_hdr_clear_flags(hdr, ARC_FLAG_PRIO_ASYNC_READ);
/*
* At this point, we have a level 1 cache miss or a blkptr
* with embedded data. Try again in L2ARC if possible.
*/
ASSERT3U(HDR_GET_LSIZE(hdr), ==, lsize);
/*
* Skip ARC stat bump for block pointers with embedded
* data. The data are read from the blkptr itself via
* decode_embedded_bp_compressed().
*/
if (!embedded_bp) {
DTRACE_PROBE4(arc__miss, arc_buf_hdr_t *, hdr,
blkptr_t *, bp, uint64_t, lsize,
zbookmark_phys_t *, zb);
ARCSTAT_BUMP(arcstat_misses);
ARCSTAT_CONDSTAT(!(*arc_flags & ARC_FLAG_PREFETCH),
demand, prefetch, !HDR_ISTYPE_METADATA(hdr), data,
metadata, misses);
zfs_racct_read(size, 1);
}
/* Check if the spa even has l2 configured */
const boolean_t spa_has_l2 = l2arc_ndev != 0 &&
spa->spa_l2cache.sav_count > 0;
if (vd != NULL && spa_has_l2 && !(l2arc_norw && devw)) {
/*
* Read from the L2ARC if the following are true:
* 1. The L2ARC vdev was previously cached.
* 2. This buffer still has L2ARC metadata.
* 3. This buffer isn't currently writing to the L2ARC.
* 4. The L2ARC entry wasn't evicted, which may
* also have invalidated the vdev.
*/
if (HDR_HAS_L2HDR(hdr) &&
!HDR_L2_WRITING(hdr) && !HDR_L2_EVICTED(hdr)) {
l2arc_read_callback_t *cb;
abd_t *abd;
uint64_t asize;
DTRACE_PROBE1(l2arc__hit, arc_buf_hdr_t *, hdr);
ARCSTAT_BUMP(arcstat_l2_hits);
hdr->b_l2hdr.b_hits++;
cb = kmem_zalloc(sizeof (l2arc_read_callback_t),
KM_SLEEP);
cb->l2rcb_hdr = hdr;
cb->l2rcb_bp = *bp;
cb->l2rcb_zb = *zb;
cb->l2rcb_flags = zio_flags;
/*
* When Compressed ARC is disabled, but the
* L2ARC block is compressed, arc_hdr_size()
* will have returned LSIZE rather than PSIZE.
*/
if (HDR_GET_COMPRESS(hdr) != ZIO_COMPRESS_OFF &&
!HDR_COMPRESSION_ENABLED(hdr) &&
HDR_GET_PSIZE(hdr) != 0) {
size = HDR_GET_PSIZE(hdr);
}
asize = vdev_psize_to_asize(vd, size);
if (asize != size) {
abd = abd_alloc_for_io(asize,
HDR_ISTYPE_METADATA(hdr));
cb->l2rcb_abd = abd;
} else {
abd = hdr_abd;
}
ASSERT(addr >= VDEV_LABEL_START_SIZE &&
addr + asize <= vd->vdev_psize -
VDEV_LABEL_END_SIZE);
/*
* l2arc read. The SCL_L2ARC lock will be
* released by l2arc_read_done().
* Issue a null zio if the underlying buffer
* was squashed to zero size by compression.
*/
ASSERT3U(arc_hdr_get_compress(hdr), !=,
ZIO_COMPRESS_EMPTY);
rzio = zio_read_phys(pio, vd, addr,
asize, abd,
ZIO_CHECKSUM_OFF,
l2arc_read_done, cb, priority,
zio_flags | ZIO_FLAG_CANFAIL |
ZIO_FLAG_DONT_PROPAGATE |
ZIO_FLAG_DONT_RETRY, B_FALSE);
acb->acb_zio_head = rzio;
if (hash_lock != NULL)
mutex_exit(hash_lock);
DTRACE_PROBE2(l2arc__read, vdev_t *, vd,
zio_t *, rzio);
ARCSTAT_INCR(arcstat_l2_read_bytes,
HDR_GET_PSIZE(hdr));
if (*arc_flags & ARC_FLAG_NOWAIT) {
zio_nowait(rzio);
goto out;
}
ASSERT(*arc_flags & ARC_FLAG_WAIT);
if (zio_wait(rzio) == 0)
goto out;
/* l2arc read error; goto zio_read() */
if (hash_lock != NULL)
mutex_enter(hash_lock);
} else {
DTRACE_PROBE1(l2arc__miss,
arc_buf_hdr_t *, hdr);
ARCSTAT_BUMP(arcstat_l2_misses);
if (HDR_L2_WRITING(hdr))
ARCSTAT_BUMP(arcstat_l2_rw_clash);
spa_config_exit(spa, SCL_L2ARC, vd);
}
} else {
if (vd != NULL)
spa_config_exit(spa, SCL_L2ARC, vd);
/*
* Only a spa with l2 should contribute to l2
* miss stats. (Including the case of having a
* faulted cache device - that's also a miss.)
*/
if (spa_has_l2) {
/*
* Skip ARC stat bump for block pointers with
* embedded data. The data are read from the
* blkptr itself via
* decode_embedded_bp_compressed().
*/
if (!embedded_bp) {
DTRACE_PROBE1(l2arc__miss,
arc_buf_hdr_t *, hdr);
ARCSTAT_BUMP(arcstat_l2_misses);
}
}
}
rzio = zio_read(pio, spa, bp, hdr_abd, size,
arc_read_done, hdr, priority, zio_flags, zb);
acb->acb_zio_head = rzio;
if (hash_lock != NULL)
mutex_exit(hash_lock);
if (*arc_flags & ARC_FLAG_WAIT) {
rc = zio_wait(rzio);
goto out;
}
ASSERT(*arc_flags & ARC_FLAG_NOWAIT);
zio_nowait(rzio);
}
out:
/* embedded bps don't actually go to disk */
if (!embedded_bp)
spa_read_history_add(spa, zb, *arc_flags);
spl_fstrans_unmark(cookie);
return (rc);
done:
if (done)
done(NULL, zb, bp, buf, private);
if (pio && rc != 0) {
zio_t *zio = zio_null(pio, spa, NULL, NULL, NULL, zio_flags);
zio->io_error = rc;
zio_nowait(zio);
}
goto out;
}
arc_prune_t *
arc_add_prune_callback(arc_prune_func_t *func, void *private)
{
arc_prune_t *p;
p = kmem_alloc(sizeof (*p), KM_SLEEP);
p->p_pfunc = func;
p->p_private = private;
list_link_init(&p->p_node);
zfs_refcount_create(&p->p_refcnt);
mutex_enter(&arc_prune_mtx);
zfs_refcount_add(&p->p_refcnt, &arc_prune_list);
list_insert_head(&arc_prune_list, p);
mutex_exit(&arc_prune_mtx);
return (p);
}
void
arc_remove_prune_callback(arc_prune_t *p)
{
boolean_t wait = B_FALSE;
mutex_enter(&arc_prune_mtx);
list_remove(&arc_prune_list, p);
if (zfs_refcount_remove(&p->p_refcnt, &arc_prune_list) > 0)
wait = B_TRUE;
mutex_exit(&arc_prune_mtx);
/* wait for arc_prune_task to finish */
if (wait)
taskq_wait_outstanding(arc_prune_taskq, 0);
ASSERT0(zfs_refcount_count(&p->p_refcnt));
zfs_refcount_destroy(&p->p_refcnt);
kmem_free(p, sizeof (*p));
}
/*
* Helper function for arc_prune_async() it is responsible for safely
* handling the execution of a registered arc_prune_func_t.
*/
static void
arc_prune_task(void *ptr)
{
arc_prune_t *ap = (arc_prune_t *)ptr;
arc_prune_func_t *func = ap->p_pfunc;
if (func != NULL)
func(ap->p_adjust, ap->p_private);
zfs_refcount_remove(&ap->p_refcnt, func);
}
/*
* Notify registered consumers they must drop holds on a portion of the ARC
* buffers they reference. This provides a mechanism to ensure the ARC can
* honor the metadata limit and reclaim otherwise pinned ARC buffers.
*
* This operation is performed asynchronously so it may be safely called
* in the context of the arc_reclaim_thread(). A reference is taken here
* for each registered arc_prune_t and the arc_prune_task() is responsible
* for releasing it once the registered arc_prune_func_t has completed.
*/
static void
arc_prune_async(uint64_t adjust)
{
arc_prune_t *ap;
mutex_enter(&arc_prune_mtx);
for (ap = list_head(&arc_prune_list); ap != NULL;
ap = list_next(&arc_prune_list, ap)) {
if (zfs_refcount_count(&ap->p_refcnt) >= 2)
continue;
zfs_refcount_add(&ap->p_refcnt, ap->p_pfunc);
ap->p_adjust = adjust;
if (taskq_dispatch(arc_prune_taskq, arc_prune_task,
ap, TQ_SLEEP) == TASKQID_INVALID) {
zfs_refcount_remove(&ap->p_refcnt, ap->p_pfunc);
continue;
}
ARCSTAT_BUMP(arcstat_prune);
}
mutex_exit(&arc_prune_mtx);
}
/*
* Notify the arc that a block was freed, and thus will never be used again.
*/
void
arc_freed(spa_t *spa, const blkptr_t *bp)
{
arc_buf_hdr_t *hdr;
kmutex_t *hash_lock;
uint64_t guid = spa_load_guid(spa);
ASSERT(!BP_IS_EMBEDDED(bp));
hdr = buf_hash_find(guid, bp, &hash_lock);
if (hdr == NULL)
return;
/*
* We might be trying to free a block that is still doing I/O
* (i.e. prefetch) or has some other reference (i.e. a dedup-ed,
* dmu_sync-ed block). A block may also have a reference if it is
* part of a dedup-ed, dmu_synced write. The dmu_sync() function would
* have written the new block to its final resting place on disk but
* without the dedup flag set. This would have left the hdr in the MRU
* state and discoverable. When the txg finally syncs it detects that
* the block was overridden in open context and issues an override I/O.
* Since this is a dedup block, the override I/O will determine if the
* block is already in the DDT. If so, then it will replace the io_bp
* with the bp from the DDT and allow the I/O to finish. When the I/O
* reaches the done callback, dbuf_write_override_done, it will
* check to see if the io_bp and io_bp_override are identical.
* If they are not, then it indicates that the bp was replaced with
* the bp in the DDT and the override bp is freed. This allows
* us to arrive here with a reference on a block that is being
* freed. So if we have an I/O in progress, or a reference to
* this hdr, then we don't destroy the hdr.
*/
if (!HDR_HAS_L1HDR(hdr) ||
zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt)) {
arc_change_state(arc_anon, hdr);
arc_hdr_destroy(hdr);
mutex_exit(hash_lock);
} else {
mutex_exit(hash_lock);
}
}
/*
* Release this buffer from the cache, making it an anonymous buffer. This
* must be done after a read and prior to modifying the buffer contents.
* If the buffer has more than one reference, we must make
* a new hdr for the buffer.
*/
void
arc_release(arc_buf_t *buf, const void *tag)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
/*
* It would be nice to assert that if its DMU metadata (level >
* 0 || it's the dnode file), then it must be syncing context.
* But we don't know that information at this level.
*/
ASSERT(HDR_HAS_L1HDR(hdr));
/*
* We don't grab the hash lock prior to this check, because if
* the buffer's header is in the arc_anon state, it won't be
* linked into the hash table.
*/
if (hdr->b_l1hdr.b_state == arc_anon) {
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT(!HDR_IN_HASH_TABLE(hdr));
ASSERT(!HDR_HAS_L2HDR(hdr));
ASSERT3P(hdr->b_l1hdr.b_buf, ==, buf);
ASSERT(ARC_BUF_LAST(buf));
ASSERT3S(zfs_refcount_count(&hdr->b_l1hdr.b_refcnt), ==, 1);
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
hdr->b_l1hdr.b_arc_access = 0;
/*
* If the buf is being overridden then it may already
* have a hdr that is not empty.
*/
buf_discard_identity(hdr);
arc_buf_thaw(buf);
return;
}
kmutex_t *hash_lock = HDR_LOCK(hdr);
mutex_enter(hash_lock);
/*
* This assignment is only valid as long as the hash_lock is
* held, we must be careful not to reference state or the
* b_state field after dropping the lock.
*/
arc_state_t *state = hdr->b_l1hdr.b_state;
ASSERT3P(hash_lock, ==, HDR_LOCK(hdr));
ASSERT3P(state, !=, arc_anon);
/* this buffer is not on any list */
ASSERT3S(zfs_refcount_count(&hdr->b_l1hdr.b_refcnt), >, 0);
if (HDR_HAS_L2HDR(hdr)) {
mutex_enter(&hdr->b_l2hdr.b_dev->l2ad_mtx);
/*
* We have to recheck this conditional again now that
* we're holding the l2ad_mtx to prevent a race with
* another thread which might be concurrently calling
* l2arc_evict(). In that case, l2arc_evict() might have
* destroyed the header's L2 portion as we were waiting
* to acquire the l2ad_mtx.
*/
if (HDR_HAS_L2HDR(hdr))
arc_hdr_l2hdr_destroy(hdr);
mutex_exit(&hdr->b_l2hdr.b_dev->l2ad_mtx);
}
/*
* Do we have more than one buf?
*/
if (hdr->b_l1hdr.b_buf != buf || !ARC_BUF_LAST(buf)) {
arc_buf_hdr_t *nhdr;
uint64_t spa = hdr->b_spa;
uint64_t psize = HDR_GET_PSIZE(hdr);
uint64_t lsize = HDR_GET_LSIZE(hdr);
boolean_t protected = HDR_PROTECTED(hdr);
enum zio_compress compress = arc_hdr_get_compress(hdr);
arc_buf_contents_t type = arc_buf_type(hdr);
VERIFY3U(hdr->b_type, ==, type);
ASSERT(hdr->b_l1hdr.b_buf != buf || buf->b_next != NULL);
VERIFY3S(remove_reference(hdr, tag), >, 0);
if (ARC_BUF_SHARED(buf) && !ARC_BUF_COMPRESSED(buf)) {
ASSERT3P(hdr->b_l1hdr.b_buf, !=, buf);
ASSERT(ARC_BUF_LAST(buf));
}
/*
* Pull the data off of this hdr and attach it to
* a new anonymous hdr. Also find the last buffer
* in the hdr's buffer list.
*/
arc_buf_t *lastbuf = arc_buf_remove(hdr, buf);
ASSERT3P(lastbuf, !=, NULL);
/*
* If the current arc_buf_t and the hdr are sharing their data
* buffer, then we must stop sharing that block.
*/
if (ARC_BUF_SHARED(buf)) {
ASSERT3P(hdr->b_l1hdr.b_buf, !=, buf);
ASSERT(!arc_buf_is_shared(lastbuf));
/*
* First, sever the block sharing relationship between
* buf and the arc_buf_hdr_t.
*/
arc_unshare_buf(hdr, buf);
/*
* Now we need to recreate the hdr's b_pabd. Since we
* have lastbuf handy, we try to share with it, but if
* we can't then we allocate a new b_pabd and copy the
* data from buf into it.
*/
if (arc_can_share(hdr, lastbuf)) {
arc_share_buf(hdr, lastbuf);
} else {
arc_hdr_alloc_abd(hdr, 0);
abd_copy_from_buf(hdr->b_l1hdr.b_pabd,
buf->b_data, psize);
}
VERIFY3P(lastbuf->b_data, !=, NULL);
} else if (HDR_SHARED_DATA(hdr)) {
/*
* Uncompressed shared buffers are always at the end
* of the list. Compressed buffers don't have the
* same requirements. This makes it hard to
* simply assert that the lastbuf is shared so
* we rely on the hdr's compression flags to determine
* if we have a compressed, shared buffer.
*/
ASSERT(arc_buf_is_shared(lastbuf) ||
arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF);
ASSERT(!arc_buf_is_shared(buf));
}
ASSERT(hdr->b_l1hdr.b_pabd != NULL || HDR_HAS_RABD(hdr));
ASSERT3P(state, !=, arc_l2c_only);
(void) zfs_refcount_remove_many(&state->arcs_size[type],
arc_buf_size(buf), buf);
if (zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt)) {
ASSERT3P(state, !=, arc_l2c_only);
(void) zfs_refcount_remove_many(
&state->arcs_esize[type],
arc_buf_size(buf), buf);
}
arc_cksum_verify(buf);
arc_buf_unwatch(buf);
/* if this is the last uncompressed buf free the checksum */
if (!arc_hdr_has_uncompressed_buf(hdr))
arc_cksum_free(hdr);
mutex_exit(hash_lock);
nhdr = arc_hdr_alloc(spa, psize, lsize, protected,
compress, hdr->b_complevel, type);
ASSERT3P(nhdr->b_l1hdr.b_buf, ==, NULL);
ASSERT0(zfs_refcount_count(&nhdr->b_l1hdr.b_refcnt));
VERIFY3U(nhdr->b_type, ==, type);
ASSERT(!HDR_SHARED_DATA(nhdr));
nhdr->b_l1hdr.b_buf = buf;
(void) zfs_refcount_add(&nhdr->b_l1hdr.b_refcnt, tag);
buf->b_hdr = nhdr;
(void) zfs_refcount_add_many(&arc_anon->arcs_size[type],
arc_buf_size(buf), buf);
} else {
ASSERT(zfs_refcount_count(&hdr->b_l1hdr.b_refcnt) == 1);
/* protected by hash lock, or hdr is on arc_anon */
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
hdr->b_l1hdr.b_mru_hits = 0;
hdr->b_l1hdr.b_mru_ghost_hits = 0;
hdr->b_l1hdr.b_mfu_hits = 0;
hdr->b_l1hdr.b_mfu_ghost_hits = 0;
arc_change_state(arc_anon, hdr);
hdr->b_l1hdr.b_arc_access = 0;
mutex_exit(hash_lock);
buf_discard_identity(hdr);
arc_buf_thaw(buf);
}
}
int
arc_released(arc_buf_t *buf)
{
return (buf->b_data != NULL &&
buf->b_hdr->b_l1hdr.b_state == arc_anon);
}
#ifdef ZFS_DEBUG
int
arc_referenced(arc_buf_t *buf)
{
return (zfs_refcount_count(&buf->b_hdr->b_l1hdr.b_refcnt));
}
#endif
static void
arc_write_ready(zio_t *zio)
{
arc_write_callback_t *callback = zio->io_private;
arc_buf_t *buf = callback->awcb_buf;
arc_buf_hdr_t *hdr = buf->b_hdr;
blkptr_t *bp = zio->io_bp;
uint64_t psize = BP_IS_HOLE(bp) ? 0 : BP_GET_PSIZE(bp);
fstrans_cookie_t cookie = spl_fstrans_mark();
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(!zfs_refcount_is_zero(&buf->b_hdr->b_l1hdr.b_refcnt));
ASSERT3P(hdr->b_l1hdr.b_buf, !=, NULL);
/*
* If we're reexecuting this zio because the pool suspended, then
* cleanup any state that was previously set the first time the
* callback was invoked.
*/
if (zio->io_flags & ZIO_FLAG_REEXECUTED) {
arc_cksum_free(hdr);
arc_buf_unwatch(buf);
if (hdr->b_l1hdr.b_pabd != NULL) {
if (ARC_BUF_SHARED(buf)) {
arc_unshare_buf(hdr, buf);
} else {
ASSERT(!arc_buf_is_shared(buf));
arc_hdr_free_abd(hdr, B_FALSE);
}
}
if (HDR_HAS_RABD(hdr))
arc_hdr_free_abd(hdr, B_TRUE);
}
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
ASSERT(!HDR_SHARED_DATA(hdr));
ASSERT(!arc_buf_is_shared(buf));
callback->awcb_ready(zio, buf, callback->awcb_private);
if (HDR_IO_IN_PROGRESS(hdr)) {
ASSERT(zio->io_flags & ZIO_FLAG_REEXECUTED);
} else {
arc_hdr_set_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
add_reference(hdr, hdr); /* For IO_IN_PROGRESS. */
}
if (BP_IS_PROTECTED(bp)) {
/* ZIL blocks are written through zio_rewrite */
ASSERT3U(BP_GET_TYPE(bp), !=, DMU_OT_INTENT_LOG);
if (BP_SHOULD_BYTESWAP(bp)) {
if (BP_GET_LEVEL(bp) > 0) {
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_UINT64;
} else {
hdr->b_l1hdr.b_byteswap =
DMU_OT_BYTESWAP(BP_GET_TYPE(bp));
}
} else {
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_NUMFUNCS;
}
arc_hdr_set_flags(hdr, ARC_FLAG_PROTECTED);
hdr->b_crypt_hdr.b_ot = BP_GET_TYPE(bp);
hdr->b_crypt_hdr.b_dsobj = zio->io_bookmark.zb_objset;
zio_crypt_decode_params_bp(bp, hdr->b_crypt_hdr.b_salt,
hdr->b_crypt_hdr.b_iv);
zio_crypt_decode_mac_bp(bp, hdr->b_crypt_hdr.b_mac);
} else {
arc_hdr_clear_flags(hdr, ARC_FLAG_PROTECTED);
}
/*
* If this block was written for raw encryption but the zio layer
* ended up only authenticating it, adjust the buffer flags now.
*/
if (BP_IS_AUTHENTICATED(bp) && ARC_BUF_ENCRYPTED(buf)) {
arc_hdr_set_flags(hdr, ARC_FLAG_NOAUTH);
buf->b_flags &= ~ARC_BUF_FLAG_ENCRYPTED;
if (BP_GET_COMPRESS(bp) == ZIO_COMPRESS_OFF)
buf->b_flags &= ~ARC_BUF_FLAG_COMPRESSED;
} else if (BP_IS_HOLE(bp) && ARC_BUF_ENCRYPTED(buf)) {
buf->b_flags &= ~ARC_BUF_FLAG_ENCRYPTED;
buf->b_flags &= ~ARC_BUF_FLAG_COMPRESSED;
}
/* this must be done after the buffer flags are adjusted */
arc_cksum_compute(buf);
enum zio_compress compress;
if (BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp)) {
compress = ZIO_COMPRESS_OFF;
} else {
ASSERT3U(HDR_GET_LSIZE(hdr), ==, BP_GET_LSIZE(bp));
compress = BP_GET_COMPRESS(bp);
}
HDR_SET_PSIZE(hdr, psize);
arc_hdr_set_compress(hdr, compress);
hdr->b_complevel = zio->io_prop.zp_complevel;
if (zio->io_error != 0 || psize == 0)
goto out;
/*
* Fill the hdr with data. If the buffer is encrypted we have no choice
* but to copy the data into b_radb. If the hdr is compressed, the data
* we want is available from the zio, otherwise we can take it from
* the buf.
*
* We might be able to share the buf's data with the hdr here. However,
* doing so would cause the ARC to be full of linear ABDs if we write a
* lot of shareable data. As a compromise, we check whether scattered
* ABDs are allowed, and assume that if they are then the user wants
* the ARC to be primarily filled with them regardless of the data being
* written. Therefore, if they're allowed then we allocate one and copy
* the data into it; otherwise, we share the data directly if we can.
*/
if (ARC_BUF_ENCRYPTED(buf)) {
ASSERT3U(psize, >, 0);
ASSERT(ARC_BUF_COMPRESSED(buf));
arc_hdr_alloc_abd(hdr, ARC_HDR_ALLOC_RDATA |
ARC_HDR_USE_RESERVE);
abd_copy(hdr->b_crypt_hdr.b_rabd, zio->io_abd, psize);
} else if (!(HDR_UNCACHED(hdr) ||
abd_size_alloc_linear(arc_buf_size(buf))) ||
!arc_can_share(hdr, buf)) {
/*
* Ideally, we would always copy the io_abd into b_pabd, but the
* user may have disabled compressed ARC, thus we must check the
* hdr's compression setting rather than the io_bp's.
*/
if (BP_IS_ENCRYPTED(bp)) {
ASSERT3U(psize, >, 0);
arc_hdr_alloc_abd(hdr, ARC_HDR_ALLOC_RDATA |
ARC_HDR_USE_RESERVE);
abd_copy(hdr->b_crypt_hdr.b_rabd, zio->io_abd, psize);
} else if (arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF &&
!ARC_BUF_COMPRESSED(buf)) {
ASSERT3U(psize, >, 0);
arc_hdr_alloc_abd(hdr, ARC_HDR_USE_RESERVE);
abd_copy(hdr->b_l1hdr.b_pabd, zio->io_abd, psize);
} else {
ASSERT3U(zio->io_orig_size, ==, arc_hdr_size(hdr));
arc_hdr_alloc_abd(hdr, ARC_HDR_USE_RESERVE);
abd_copy_from_buf(hdr->b_l1hdr.b_pabd, buf->b_data,
arc_buf_size(buf));
}
} else {
ASSERT3P(buf->b_data, ==, abd_to_buf(zio->io_orig_abd));
ASSERT3U(zio->io_orig_size, ==, arc_buf_size(buf));
ASSERT3P(hdr->b_l1hdr.b_buf, ==, buf);
ASSERT(ARC_BUF_LAST(buf));
arc_share_buf(hdr, buf);
}
out:
arc_hdr_verify(hdr, bp);
spl_fstrans_unmark(cookie);
}
static void
arc_write_children_ready(zio_t *zio)
{
arc_write_callback_t *callback = zio->io_private;
arc_buf_t *buf = callback->awcb_buf;
callback->awcb_children_ready(zio, buf, callback->awcb_private);
}
static void
arc_write_done(zio_t *zio)
{
arc_write_callback_t *callback = zio->io_private;
arc_buf_t *buf = callback->awcb_buf;
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
if (zio->io_error == 0) {
arc_hdr_verify(hdr, zio->io_bp);
if (BP_IS_HOLE(zio->io_bp) || BP_IS_EMBEDDED(zio->io_bp)) {
buf_discard_identity(hdr);
} else {
hdr->b_dva = *BP_IDENTITY(zio->io_bp);
hdr->b_birth = BP_PHYSICAL_BIRTH(zio->io_bp);
}
} else {
ASSERT(HDR_EMPTY(hdr));
}
/*
* If the block to be written was all-zero or compressed enough to be
* embedded in the BP, no write was performed so there will be no
* dva/birth/checksum. The buffer must therefore remain anonymous
* (and uncached).
*/
if (!HDR_EMPTY(hdr)) {
arc_buf_hdr_t *exists;
kmutex_t *hash_lock;
ASSERT3U(zio->io_error, ==, 0);
arc_cksum_verify(buf);
exists = buf_hash_insert(hdr, &hash_lock);
if (exists != NULL) {
/*
* This can only happen if we overwrite for
* sync-to-convergence, because we remove
* buffers from the hash table when we arc_free().
*/
if (zio->io_flags & ZIO_FLAG_IO_REWRITE) {
if (!BP_EQUAL(&zio->io_bp_orig, zio->io_bp))
panic("bad overwrite, hdr=%p exists=%p",
(void *)hdr, (void *)exists);
ASSERT(zfs_refcount_is_zero(
&exists->b_l1hdr.b_refcnt));
arc_change_state(arc_anon, exists);
arc_hdr_destroy(exists);
mutex_exit(hash_lock);
exists = buf_hash_insert(hdr, &hash_lock);
ASSERT3P(exists, ==, NULL);
} else if (zio->io_flags & ZIO_FLAG_NOPWRITE) {
/* nopwrite */
ASSERT(zio->io_prop.zp_nopwrite);
if (!BP_EQUAL(&zio->io_bp_orig, zio->io_bp))
panic("bad nopwrite, hdr=%p exists=%p",
(void *)hdr, (void *)exists);
} else {
/* Dedup */
ASSERT3P(hdr->b_l1hdr.b_buf, !=, NULL);
ASSERT(ARC_BUF_LAST(hdr->b_l1hdr.b_buf));
ASSERT(hdr->b_l1hdr.b_state == arc_anon);
ASSERT(BP_GET_DEDUP(zio->io_bp));
ASSERT(BP_GET_LEVEL(zio->io_bp) == 0);
}
}
arc_hdr_clear_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
VERIFY3S(remove_reference(hdr, hdr), >, 0);
/* if it's not anon, we are doing a scrub */
if (exists == NULL && hdr->b_l1hdr.b_state == arc_anon)
arc_access(hdr, 0, B_FALSE);
mutex_exit(hash_lock);
} else {
arc_hdr_clear_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
VERIFY3S(remove_reference(hdr, hdr), >, 0);
}
callback->awcb_done(zio, buf, callback->awcb_private);
abd_free(zio->io_abd);
kmem_free(callback, sizeof (arc_write_callback_t));
}
zio_t *
arc_write(zio_t *pio, spa_t *spa, uint64_t txg,
blkptr_t *bp, arc_buf_t *buf, boolean_t uncached, boolean_t l2arc,
const zio_prop_t *zp, arc_write_done_func_t *ready,
arc_write_done_func_t *children_ready, arc_write_done_func_t *done,
void *private, zio_priority_t priority, int zio_flags,
const zbookmark_phys_t *zb)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
arc_write_callback_t *callback;
zio_t *zio;
zio_prop_t localprop = *zp;
ASSERT3P(ready, !=, NULL);
ASSERT3P(done, !=, NULL);
ASSERT(!HDR_IO_ERROR(hdr));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
ASSERT3P(hdr->b_l1hdr.b_buf, !=, NULL);
if (uncached)
arc_hdr_set_flags(hdr, ARC_FLAG_UNCACHED);
else if (l2arc)
arc_hdr_set_flags(hdr, ARC_FLAG_L2CACHE);
if (ARC_BUF_ENCRYPTED(buf)) {
ASSERT(ARC_BUF_COMPRESSED(buf));
localprop.zp_encrypt = B_TRUE;
localprop.zp_compress = HDR_GET_COMPRESS(hdr);
localprop.zp_complevel = hdr->b_complevel;
localprop.zp_byteorder =
(hdr->b_l1hdr.b_byteswap == DMU_BSWAP_NUMFUNCS) ?
ZFS_HOST_BYTEORDER : !ZFS_HOST_BYTEORDER;
memcpy(localprop.zp_salt, hdr->b_crypt_hdr.b_salt,
ZIO_DATA_SALT_LEN);
memcpy(localprop.zp_iv, hdr->b_crypt_hdr.b_iv,
ZIO_DATA_IV_LEN);
memcpy(localprop.zp_mac, hdr->b_crypt_hdr.b_mac,
ZIO_DATA_MAC_LEN);
if (DMU_OT_IS_ENCRYPTED(localprop.zp_type)) {
localprop.zp_nopwrite = B_FALSE;
localprop.zp_copies =
MIN(localprop.zp_copies, SPA_DVAS_PER_BP - 1);
}
zio_flags |= ZIO_FLAG_RAW;
} else if (ARC_BUF_COMPRESSED(buf)) {
ASSERT3U(HDR_GET_LSIZE(hdr), !=, arc_buf_size(buf));
localprop.zp_compress = HDR_GET_COMPRESS(hdr);
localprop.zp_complevel = hdr->b_complevel;
zio_flags |= ZIO_FLAG_RAW_COMPRESS;
}
callback = kmem_zalloc(sizeof (arc_write_callback_t), KM_SLEEP);
callback->awcb_ready = ready;
callback->awcb_children_ready = children_ready;
callback->awcb_done = done;
callback->awcb_private = private;
callback->awcb_buf = buf;
/*
* The hdr's b_pabd is now stale, free it now. A new data block
* will be allocated when the zio pipeline calls arc_write_ready().
*/
if (hdr->b_l1hdr.b_pabd != NULL) {
/*
* If the buf is currently sharing the data block with
* the hdr then we need to break that relationship here.
* The hdr will remain with a NULL data pointer and the
* buf will take sole ownership of the block.
*/
if (ARC_BUF_SHARED(buf)) {
arc_unshare_buf(hdr, buf);
} else {
ASSERT(!arc_buf_is_shared(buf));
arc_hdr_free_abd(hdr, B_FALSE);
}
VERIFY3P(buf->b_data, !=, NULL);
}
if (HDR_HAS_RABD(hdr))
arc_hdr_free_abd(hdr, B_TRUE);
if (!(zio_flags & ZIO_FLAG_RAW))
arc_hdr_set_compress(hdr, ZIO_COMPRESS_OFF);
ASSERT(!arc_buf_is_shared(buf));
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
zio = zio_write(pio, spa, txg, bp,
abd_get_from_buf(buf->b_data, HDR_GET_LSIZE(hdr)),
HDR_GET_LSIZE(hdr), arc_buf_size(buf), &localprop, arc_write_ready,
(children_ready != NULL) ? arc_write_children_ready : NULL,
arc_write_done, callback, priority, zio_flags, zb);
return (zio);
}
void
arc_tempreserve_clear(uint64_t reserve)
{
atomic_add_64(&arc_tempreserve, -reserve);
ASSERT((int64_t)arc_tempreserve >= 0);
}
int
arc_tempreserve_space(spa_t *spa, uint64_t reserve, uint64_t txg)
{
int error;
uint64_t anon_size;
if (!arc_no_grow &&
reserve > arc_c/4 &&
reserve * 4 > (2ULL << SPA_MAXBLOCKSHIFT))
arc_c = MIN(arc_c_max, reserve * 4);
/*
* Throttle when the calculated memory footprint for the TXG
* exceeds the target ARC size.
*/
if (reserve > arc_c) {
DMU_TX_STAT_BUMP(dmu_tx_memory_reserve);
return (SET_ERROR(ERESTART));
}
/*
* Don't count loaned bufs as in flight dirty data to prevent long
* network delays from blocking transactions that are ready to be
* assigned to a txg.
*/
/* assert that it has not wrapped around */
ASSERT3S(atomic_add_64_nv(&arc_loaned_bytes, 0), >=, 0);
anon_size = MAX((int64_t)
(zfs_refcount_count(&arc_anon->arcs_size[ARC_BUFC_DATA]) +
zfs_refcount_count(&arc_anon->arcs_size[ARC_BUFC_METADATA]) -
arc_loaned_bytes), 0);
/*
* Writes will, almost always, require additional memory allocations
* in order to compress/encrypt/etc the data. We therefore need to
* make sure that there is sufficient available memory for this.
*/
error = arc_memory_throttle(spa, reserve, txg);
if (error != 0)
return (error);
/*
* Throttle writes when the amount of dirty data in the cache
* gets too large. We try to keep the cache less than half full
* of dirty blocks so that our sync times don't grow too large.
*
* In the case of one pool being built on another pool, we want
* to make sure we don't end up throttling the lower (backing)
* pool when the upper pool is the majority contributor to dirty
* data. To insure we make forward progress during throttling, we
* also check the current pool's net dirty data and only throttle
* if it exceeds zfs_arc_pool_dirty_percent of the anonymous dirty
* data in the cache.
*
* Note: if two requests come in concurrently, we might let them
* both succeed, when one of them should fail. Not a huge deal.
*/
uint64_t total_dirty = reserve + arc_tempreserve + anon_size;
uint64_t spa_dirty_anon = spa_dirty_data(spa);
uint64_t rarc_c = arc_warm ? arc_c : arc_c_max;
if (total_dirty > rarc_c * zfs_arc_dirty_limit_percent / 100 &&
anon_size > rarc_c * zfs_arc_anon_limit_percent / 100 &&
spa_dirty_anon > anon_size * zfs_arc_pool_dirty_percent / 100) {
#ifdef ZFS_DEBUG
uint64_t meta_esize = zfs_refcount_count(
&arc_anon->arcs_esize[ARC_BUFC_METADATA]);
uint64_t data_esize =
zfs_refcount_count(&arc_anon->arcs_esize[ARC_BUFC_DATA]);
dprintf("failing, arc_tempreserve=%lluK anon_meta=%lluK "
"anon_data=%lluK tempreserve=%lluK rarc_c=%lluK\n",
(u_longlong_t)arc_tempreserve >> 10,
(u_longlong_t)meta_esize >> 10,
(u_longlong_t)data_esize >> 10,
(u_longlong_t)reserve >> 10,
(u_longlong_t)rarc_c >> 10);
#endif
DMU_TX_STAT_BUMP(dmu_tx_dirty_throttle);
return (SET_ERROR(ERESTART));
}
atomic_add_64(&arc_tempreserve, reserve);
return (0);
}
static void
arc_kstat_update_state(arc_state_t *state, kstat_named_t *size,
kstat_named_t *data, kstat_named_t *metadata,
kstat_named_t *evict_data, kstat_named_t *evict_metadata)
{
data->value.ui64 =
zfs_refcount_count(&state->arcs_size[ARC_BUFC_DATA]);
metadata->value.ui64 =
zfs_refcount_count(&state->arcs_size[ARC_BUFC_METADATA]);
size->value.ui64 = data->value.ui64 + metadata->value.ui64;
evict_data->value.ui64 =
zfs_refcount_count(&state->arcs_esize[ARC_BUFC_DATA]);
evict_metadata->value.ui64 =
zfs_refcount_count(&state->arcs_esize[ARC_BUFC_METADATA]);
}
static int
arc_kstat_update(kstat_t *ksp, int rw)
{
arc_stats_t *as = ksp->ks_data;
if (rw == KSTAT_WRITE)
return (SET_ERROR(EACCES));
as->arcstat_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_hits);
as->arcstat_iohits.value.ui64 =
wmsum_value(&arc_sums.arcstat_iohits);
as->arcstat_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_misses);
as->arcstat_demand_data_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_data_hits);
as->arcstat_demand_data_iohits.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_data_iohits);
as->arcstat_demand_data_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_data_misses);
as->arcstat_demand_metadata_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_metadata_hits);
as->arcstat_demand_metadata_iohits.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_metadata_iohits);
as->arcstat_demand_metadata_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_metadata_misses);
as->arcstat_prefetch_data_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_data_hits);
as->arcstat_prefetch_data_iohits.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_data_iohits);
as->arcstat_prefetch_data_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_data_misses);
as->arcstat_prefetch_metadata_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_metadata_hits);
as->arcstat_prefetch_metadata_iohits.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_metadata_iohits);
as->arcstat_prefetch_metadata_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_metadata_misses);
as->arcstat_mru_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_mru_hits);
as->arcstat_mru_ghost_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_mru_ghost_hits);
as->arcstat_mfu_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_mfu_hits);
as->arcstat_mfu_ghost_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_mfu_ghost_hits);
as->arcstat_uncached_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_uncached_hits);
as->arcstat_deleted.value.ui64 =
wmsum_value(&arc_sums.arcstat_deleted);
as->arcstat_mutex_miss.value.ui64 =
wmsum_value(&arc_sums.arcstat_mutex_miss);
as->arcstat_access_skip.value.ui64 =
wmsum_value(&arc_sums.arcstat_access_skip);
as->arcstat_evict_skip.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_skip);
as->arcstat_evict_not_enough.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_not_enough);
as->arcstat_evict_l2_cached.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_cached);
as->arcstat_evict_l2_eligible.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_eligible);
as->arcstat_evict_l2_eligible_mfu.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_eligible_mfu);
as->arcstat_evict_l2_eligible_mru.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_eligible_mru);
as->arcstat_evict_l2_ineligible.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_ineligible);
as->arcstat_evict_l2_skip.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_skip);
as->arcstat_hash_collisions.value.ui64 =
wmsum_value(&arc_sums.arcstat_hash_collisions);
as->arcstat_hash_chains.value.ui64 =
wmsum_value(&arc_sums.arcstat_hash_chains);
as->arcstat_size.value.ui64 =
aggsum_value(&arc_sums.arcstat_size);
as->arcstat_compressed_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_compressed_size);
as->arcstat_uncompressed_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_uncompressed_size);
as->arcstat_overhead_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_overhead_size);
as->arcstat_hdr_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_hdr_size);
as->arcstat_data_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_data_size);
as->arcstat_metadata_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_metadata_size);
as->arcstat_dbuf_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_dbuf_size);
#if defined(COMPAT_FREEBSD11)
as->arcstat_other_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_bonus_size) +
wmsum_value(&arc_sums.arcstat_dnode_size) +
wmsum_value(&arc_sums.arcstat_dbuf_size);
#endif
arc_kstat_update_state(arc_anon,
&as->arcstat_anon_size,
&as->arcstat_anon_data,
&as->arcstat_anon_metadata,
&as->arcstat_anon_evictable_data,
&as->arcstat_anon_evictable_metadata);
arc_kstat_update_state(arc_mru,
&as->arcstat_mru_size,
&as->arcstat_mru_data,
&as->arcstat_mru_metadata,
&as->arcstat_mru_evictable_data,
&as->arcstat_mru_evictable_metadata);
arc_kstat_update_state(arc_mru_ghost,
&as->arcstat_mru_ghost_size,
&as->arcstat_mru_ghost_data,
&as->arcstat_mru_ghost_metadata,
&as->arcstat_mru_ghost_evictable_data,
&as->arcstat_mru_ghost_evictable_metadata);
arc_kstat_update_state(arc_mfu,
&as->arcstat_mfu_size,
&as->arcstat_mfu_data,
&as->arcstat_mfu_metadata,
&as->arcstat_mfu_evictable_data,
&as->arcstat_mfu_evictable_metadata);
arc_kstat_update_state(arc_mfu_ghost,
&as->arcstat_mfu_ghost_size,
&as->arcstat_mfu_ghost_data,
&as->arcstat_mfu_ghost_metadata,
&as->arcstat_mfu_ghost_evictable_data,
&as->arcstat_mfu_ghost_evictable_metadata);
arc_kstat_update_state(arc_uncached,
&as->arcstat_uncached_size,
&as->arcstat_uncached_data,
&as->arcstat_uncached_metadata,
&as->arcstat_uncached_evictable_data,
&as->arcstat_uncached_evictable_metadata);
as->arcstat_dnode_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_dnode_size);
as->arcstat_bonus_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_bonus_size);
as->arcstat_l2_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_hits);
as->arcstat_l2_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_misses);
as->arcstat_l2_prefetch_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_prefetch_asize);
as->arcstat_l2_mru_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_mru_asize);
as->arcstat_l2_mfu_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_mfu_asize);
as->arcstat_l2_bufc_data_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_bufc_data_asize);
as->arcstat_l2_bufc_metadata_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_bufc_metadata_asize);
as->arcstat_l2_feeds.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_feeds);
as->arcstat_l2_rw_clash.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rw_clash);
as->arcstat_l2_read_bytes.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_read_bytes);
as->arcstat_l2_write_bytes.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_write_bytes);
as->arcstat_l2_writes_sent.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_writes_sent);
as->arcstat_l2_writes_done.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_writes_done);
as->arcstat_l2_writes_error.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_writes_error);
as->arcstat_l2_writes_lock_retry.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_writes_lock_retry);
as->arcstat_l2_evict_lock_retry.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_evict_lock_retry);
as->arcstat_l2_evict_reading.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_evict_reading);
as->arcstat_l2_evict_l1cached.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_evict_l1cached);
as->arcstat_l2_free_on_write.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_free_on_write);
as->arcstat_l2_abort_lowmem.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_abort_lowmem);
as->arcstat_l2_cksum_bad.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_cksum_bad);
as->arcstat_l2_io_error.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_io_error);
as->arcstat_l2_lsize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_lsize);
as->arcstat_l2_psize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_psize);
as->arcstat_l2_hdr_size.value.ui64 =
aggsum_value(&arc_sums.arcstat_l2_hdr_size);
as->arcstat_l2_log_blk_writes.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_log_blk_writes);
as->arcstat_l2_log_blk_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_log_blk_asize);
as->arcstat_l2_log_blk_count.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_log_blk_count);
as->arcstat_l2_rebuild_success.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_success);
as->arcstat_l2_rebuild_abort_unsupported.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_abort_unsupported);
as->arcstat_l2_rebuild_abort_io_errors.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_abort_io_errors);
as->arcstat_l2_rebuild_abort_dh_errors.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_abort_dh_errors);
as->arcstat_l2_rebuild_abort_cksum_lb_errors.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_abort_cksum_lb_errors);
as->arcstat_l2_rebuild_abort_lowmem.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_abort_lowmem);
as->arcstat_l2_rebuild_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_size);
as->arcstat_l2_rebuild_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_asize);
as->arcstat_l2_rebuild_bufs.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_bufs);
as->arcstat_l2_rebuild_bufs_precached.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_bufs_precached);
as->arcstat_l2_rebuild_log_blks.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_log_blks);
as->arcstat_memory_throttle_count.value.ui64 =
wmsum_value(&arc_sums.arcstat_memory_throttle_count);
as->arcstat_memory_direct_count.value.ui64 =
wmsum_value(&arc_sums.arcstat_memory_direct_count);
as->arcstat_memory_indirect_count.value.ui64 =
wmsum_value(&arc_sums.arcstat_memory_indirect_count);
as->arcstat_memory_all_bytes.value.ui64 =
arc_all_memory();
as->arcstat_memory_free_bytes.value.ui64 =
arc_free_memory();
as->arcstat_memory_available_bytes.value.i64 =
arc_available_memory();
as->arcstat_prune.value.ui64 =
wmsum_value(&arc_sums.arcstat_prune);
as->arcstat_meta_used.value.ui64 =
wmsum_value(&arc_sums.arcstat_meta_used);
as->arcstat_async_upgrade_sync.value.ui64 =
wmsum_value(&arc_sums.arcstat_async_upgrade_sync);
as->arcstat_predictive_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_predictive_prefetch);
as->arcstat_demand_hit_predictive_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_hit_predictive_prefetch);
as->arcstat_demand_iohit_predictive_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_iohit_predictive_prefetch);
as->arcstat_prescient_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_prescient_prefetch);
as->arcstat_demand_hit_prescient_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_hit_prescient_prefetch);
as->arcstat_demand_iohit_prescient_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_iohit_prescient_prefetch);
as->arcstat_raw_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_raw_size);
as->arcstat_cached_only_in_progress.value.ui64 =
wmsum_value(&arc_sums.arcstat_cached_only_in_progress);
as->arcstat_abd_chunk_waste_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_abd_chunk_waste_size);
return (0);
}
/*
* This function *must* return indices evenly distributed between all
* sublists of the multilist. This is needed due to how the ARC eviction
* code is laid out; arc_evict_state() assumes ARC buffers are evenly
* distributed between all sublists and uses this assumption when
* deciding which sublist to evict from and how much to evict from it.
*/
static unsigned int
arc_state_multilist_index_func(multilist_t *ml, void *obj)
{
arc_buf_hdr_t *hdr = obj;
/*
* We rely on b_dva to generate evenly distributed index
* numbers using buf_hash below. So, as an added precaution,
* let's make sure we never add empty buffers to the arc lists.
*/
ASSERT(!HDR_EMPTY(hdr));
/*
* The assumption here, is the hash value for a given
* arc_buf_hdr_t will remain constant throughout its lifetime
* (i.e. its b_spa, b_dva, and b_birth fields don't change).
* Thus, we don't need to store the header's sublist index
* on insertion, as this index can be recalculated on removal.
*
* Also, the low order bits of the hash value are thought to be
* distributed evenly. Otherwise, in the case that the multilist
* has a power of two number of sublists, each sublists' usage
* would not be evenly distributed. In this context full 64bit
* division would be a waste of time, so limit it to 32 bits.
*/
return ((unsigned int)buf_hash(hdr->b_spa, &hdr->b_dva, hdr->b_birth) %
multilist_get_num_sublists(ml));
}
static unsigned int
arc_state_l2c_multilist_index_func(multilist_t *ml, void *obj)
{
panic("Header %p insert into arc_l2c_only %p", obj, ml);
}
#define WARN_IF_TUNING_IGNORED(tuning, value, do_warn) do { \
if ((do_warn) && (tuning) && ((tuning) != (value))) { \
cmn_err(CE_WARN, \
"ignoring tunable %s (using %llu instead)", \
(#tuning), (u_longlong_t)(value)); \
} \
} while (0)
/*
* Called during module initialization and periodically thereafter to
* apply reasonable changes to the exposed performance tunings. Can also be
* called explicitly by param_set_arc_*() functions when ARC tunables are
* updated manually. Non-zero zfs_* values which differ from the currently set
* values will be applied.
*/
void
arc_tuning_update(boolean_t verbose)
{
uint64_t allmem = arc_all_memory();
/* Valid range: 32M - <arc_c_max> */
if ((zfs_arc_min) && (zfs_arc_min != arc_c_min) &&
(zfs_arc_min >= 2ULL << SPA_MAXBLOCKSHIFT) &&
(zfs_arc_min <= arc_c_max)) {
arc_c_min = zfs_arc_min;
arc_c = MAX(arc_c, arc_c_min);
}
WARN_IF_TUNING_IGNORED(zfs_arc_min, arc_c_min, verbose);
/* Valid range: 64M - <all physical memory> */
if ((zfs_arc_max) && (zfs_arc_max != arc_c_max) &&
(zfs_arc_max >= MIN_ARC_MAX) && (zfs_arc_max < allmem) &&
(zfs_arc_max > arc_c_min)) {
arc_c_max = zfs_arc_max;
arc_c = MIN(arc_c, arc_c_max);
if (arc_dnode_limit > arc_c_max)
arc_dnode_limit = arc_c_max;
}
WARN_IF_TUNING_IGNORED(zfs_arc_max, arc_c_max, verbose);
/* Valid range: 0 - <all physical memory> */
arc_dnode_limit = zfs_arc_dnode_limit ? zfs_arc_dnode_limit :
MIN(zfs_arc_dnode_limit_percent, 100) * arc_c_max / 100;
WARN_IF_TUNING_IGNORED(zfs_arc_dnode_limit, arc_dnode_limit, verbose);
/* Valid range: 1 - N */
if (zfs_arc_grow_retry)
arc_grow_retry = zfs_arc_grow_retry;
/* Valid range: 1 - N */
if (zfs_arc_shrink_shift) {
arc_shrink_shift = zfs_arc_shrink_shift;
arc_no_grow_shift = MIN(arc_no_grow_shift, arc_shrink_shift -1);
}
/* Valid range: 1 - N ms */
if (zfs_arc_min_prefetch_ms)
arc_min_prefetch_ms = zfs_arc_min_prefetch_ms;
/* Valid range: 1 - N ms */
if (zfs_arc_min_prescient_prefetch_ms) {
arc_min_prescient_prefetch_ms =
zfs_arc_min_prescient_prefetch_ms;
}
/* Valid range: 0 - 100 */
if (zfs_arc_lotsfree_percent <= 100)
arc_lotsfree_percent = zfs_arc_lotsfree_percent;
WARN_IF_TUNING_IGNORED(zfs_arc_lotsfree_percent, arc_lotsfree_percent,
verbose);
/* Valid range: 0 - <all physical memory> */
if ((zfs_arc_sys_free) && (zfs_arc_sys_free != arc_sys_free))
arc_sys_free = MIN(zfs_arc_sys_free, allmem);
WARN_IF_TUNING_IGNORED(zfs_arc_sys_free, arc_sys_free, verbose);
}
static void
arc_state_multilist_init(multilist_t *ml,
multilist_sublist_index_func_t *index_func, int *maxcountp)
{
multilist_create(ml, sizeof (arc_buf_hdr_t),
offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), index_func);
*maxcountp = MAX(*maxcountp, multilist_get_num_sublists(ml));
}
static void
arc_state_init(void)
{
int num_sublists = 0;
arc_state_multilist_init(&arc_mru->arcs_list[ARC_BUFC_METADATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mru->arcs_list[ARC_BUFC_DATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mru_ghost->arcs_list[ARC_BUFC_METADATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mru_ghost->arcs_list[ARC_BUFC_DATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mfu->arcs_list[ARC_BUFC_METADATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mfu->arcs_list[ARC_BUFC_DATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mfu_ghost->arcs_list[ARC_BUFC_METADATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mfu_ghost->arcs_list[ARC_BUFC_DATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_uncached->arcs_list[ARC_BUFC_METADATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_uncached->arcs_list[ARC_BUFC_DATA],
arc_state_multilist_index_func, &num_sublists);
/*
* L2 headers should never be on the L2 state list since they don't
* have L1 headers allocated. Special index function asserts that.
*/
arc_state_multilist_init(&arc_l2c_only->arcs_list[ARC_BUFC_METADATA],
arc_state_l2c_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_l2c_only->arcs_list[ARC_BUFC_DATA],
arc_state_l2c_multilist_index_func, &num_sublists);
/*
* Keep track of the number of markers needed to reclaim buffers from
* any ARC state. The markers will be pre-allocated so as to minimize
* the number of memory allocations performed by the eviction thread.
*/
arc_state_evict_marker_count = num_sublists;
zfs_refcount_create(&arc_anon->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_anon->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mru->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mru->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mru_ghost->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mru_ghost->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mfu->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mfu->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mfu_ghost->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mfu_ghost->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_l2c_only->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_l2c_only->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_uncached->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_uncached->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_anon->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_anon->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mru->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mru->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mru_ghost->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mru_ghost->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mfu->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mfu->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mfu_ghost->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mfu_ghost->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_l2c_only->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_l2c_only->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_uncached->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_uncached->arcs_size[ARC_BUFC_METADATA]);
wmsum_init(&arc_mru_ghost->arcs_hits[ARC_BUFC_DATA], 0);
wmsum_init(&arc_mru_ghost->arcs_hits[ARC_BUFC_METADATA], 0);
wmsum_init(&arc_mfu_ghost->arcs_hits[ARC_BUFC_DATA], 0);
wmsum_init(&arc_mfu_ghost->arcs_hits[ARC_BUFC_METADATA], 0);
wmsum_init(&arc_sums.arcstat_hits, 0);
wmsum_init(&arc_sums.arcstat_iohits, 0);
wmsum_init(&arc_sums.arcstat_misses, 0);
wmsum_init(&arc_sums.arcstat_demand_data_hits, 0);
wmsum_init(&arc_sums.arcstat_demand_data_iohits, 0);
wmsum_init(&arc_sums.arcstat_demand_data_misses, 0);
wmsum_init(&arc_sums.arcstat_demand_metadata_hits, 0);
wmsum_init(&arc_sums.arcstat_demand_metadata_iohits, 0);
wmsum_init(&arc_sums.arcstat_demand_metadata_misses, 0);
wmsum_init(&arc_sums.arcstat_prefetch_data_hits, 0);
wmsum_init(&arc_sums.arcstat_prefetch_data_iohits, 0);
wmsum_init(&arc_sums.arcstat_prefetch_data_misses, 0);
wmsum_init(&arc_sums.arcstat_prefetch_metadata_hits, 0);
wmsum_init(&arc_sums.arcstat_prefetch_metadata_iohits, 0);
wmsum_init(&arc_sums.arcstat_prefetch_metadata_misses, 0);
wmsum_init(&arc_sums.arcstat_mru_hits, 0);
wmsum_init(&arc_sums.arcstat_mru_ghost_hits, 0);
wmsum_init(&arc_sums.arcstat_mfu_hits, 0);
wmsum_init(&arc_sums.arcstat_mfu_ghost_hits, 0);
wmsum_init(&arc_sums.arcstat_uncached_hits, 0);
wmsum_init(&arc_sums.arcstat_deleted, 0);
wmsum_init(&arc_sums.arcstat_mutex_miss, 0);
wmsum_init(&arc_sums.arcstat_access_skip, 0);
wmsum_init(&arc_sums.arcstat_evict_skip, 0);
wmsum_init(&arc_sums.arcstat_evict_not_enough, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_cached, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_eligible, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_eligible_mfu, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_eligible_mru, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_ineligible, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_skip, 0);
wmsum_init(&arc_sums.arcstat_hash_collisions, 0);
wmsum_init(&arc_sums.arcstat_hash_chains, 0);
aggsum_init(&arc_sums.arcstat_size, 0);
wmsum_init(&arc_sums.arcstat_compressed_size, 0);
wmsum_init(&arc_sums.arcstat_uncompressed_size, 0);
wmsum_init(&arc_sums.arcstat_overhead_size, 0);
wmsum_init(&arc_sums.arcstat_hdr_size, 0);
wmsum_init(&arc_sums.arcstat_data_size, 0);
wmsum_init(&arc_sums.arcstat_metadata_size, 0);
wmsum_init(&arc_sums.arcstat_dbuf_size, 0);
wmsum_init(&arc_sums.arcstat_dnode_size, 0);
wmsum_init(&arc_sums.arcstat_bonus_size, 0);
wmsum_init(&arc_sums.arcstat_l2_hits, 0);
wmsum_init(&arc_sums.arcstat_l2_misses, 0);
wmsum_init(&arc_sums.arcstat_l2_prefetch_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_mru_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_mfu_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_bufc_data_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_bufc_metadata_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_feeds, 0);
wmsum_init(&arc_sums.arcstat_l2_rw_clash, 0);
wmsum_init(&arc_sums.arcstat_l2_read_bytes, 0);
wmsum_init(&arc_sums.arcstat_l2_write_bytes, 0);
wmsum_init(&arc_sums.arcstat_l2_writes_sent, 0);
wmsum_init(&arc_sums.arcstat_l2_writes_done, 0);
wmsum_init(&arc_sums.arcstat_l2_writes_error, 0);
wmsum_init(&arc_sums.arcstat_l2_writes_lock_retry, 0);
wmsum_init(&arc_sums.arcstat_l2_evict_lock_retry, 0);
wmsum_init(&arc_sums.arcstat_l2_evict_reading, 0);
wmsum_init(&arc_sums.arcstat_l2_evict_l1cached, 0);
wmsum_init(&arc_sums.arcstat_l2_free_on_write, 0);
wmsum_init(&arc_sums.arcstat_l2_abort_lowmem, 0);
wmsum_init(&arc_sums.arcstat_l2_cksum_bad, 0);
wmsum_init(&arc_sums.arcstat_l2_io_error, 0);
wmsum_init(&arc_sums.arcstat_l2_lsize, 0);
wmsum_init(&arc_sums.arcstat_l2_psize, 0);
aggsum_init(&arc_sums.arcstat_l2_hdr_size, 0);
wmsum_init(&arc_sums.arcstat_l2_log_blk_writes, 0);
wmsum_init(&arc_sums.arcstat_l2_log_blk_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_log_blk_count, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_success, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_abort_unsupported, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_abort_io_errors, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_abort_dh_errors, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_abort_cksum_lb_errors, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_abort_lowmem, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_size, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_bufs, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_bufs_precached, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_log_blks, 0);
wmsum_init(&arc_sums.arcstat_memory_throttle_count, 0);
wmsum_init(&arc_sums.arcstat_memory_direct_count, 0);
wmsum_init(&arc_sums.arcstat_memory_indirect_count, 0);
wmsum_init(&arc_sums.arcstat_prune, 0);
wmsum_init(&arc_sums.arcstat_meta_used, 0);
wmsum_init(&arc_sums.arcstat_async_upgrade_sync, 0);
wmsum_init(&arc_sums.arcstat_predictive_prefetch, 0);
wmsum_init(&arc_sums.arcstat_demand_hit_predictive_prefetch, 0);
wmsum_init(&arc_sums.arcstat_demand_iohit_predictive_prefetch, 0);
wmsum_init(&arc_sums.arcstat_prescient_prefetch, 0);
wmsum_init(&arc_sums.arcstat_demand_hit_prescient_prefetch, 0);
wmsum_init(&arc_sums.arcstat_demand_iohit_prescient_prefetch, 0);
wmsum_init(&arc_sums.arcstat_raw_size, 0);
wmsum_init(&arc_sums.arcstat_cached_only_in_progress, 0);
wmsum_init(&arc_sums.arcstat_abd_chunk_waste_size, 0);
arc_anon->arcs_state = ARC_STATE_ANON;
arc_mru->arcs_state = ARC_STATE_MRU;
arc_mru_ghost->arcs_state = ARC_STATE_MRU_GHOST;
arc_mfu->arcs_state = ARC_STATE_MFU;
arc_mfu_ghost->arcs_state = ARC_STATE_MFU_GHOST;
arc_l2c_only->arcs_state = ARC_STATE_L2C_ONLY;
arc_uncached->arcs_state = ARC_STATE_UNCACHED;
}
static void
arc_state_fini(void)
{
zfs_refcount_destroy(&arc_anon->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_anon->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mru->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mru->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mru_ghost->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mru_ghost->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mfu->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mfu->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mfu_ghost->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mfu_ghost->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_l2c_only->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_l2c_only->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_uncached->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_uncached->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_anon->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_anon->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mru->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mru->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mru_ghost->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mru_ghost->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mfu->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mfu->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mfu_ghost->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mfu_ghost->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_l2c_only->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_l2c_only->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_uncached->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_uncached->arcs_size[ARC_BUFC_METADATA]);
multilist_destroy(&arc_mru->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_mru_ghost->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_mfu->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_mfu_ghost->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_mru->arcs_list[ARC_BUFC_DATA]);
multilist_destroy(&arc_mru_ghost->arcs_list[ARC_BUFC_DATA]);
multilist_destroy(&arc_mfu->arcs_list[ARC_BUFC_DATA]);
multilist_destroy(&arc_mfu_ghost->arcs_list[ARC_BUFC_DATA]);
multilist_destroy(&arc_l2c_only->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_l2c_only->arcs_list[ARC_BUFC_DATA]);
multilist_destroy(&arc_uncached->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_uncached->arcs_list[ARC_BUFC_DATA]);
wmsum_fini(&arc_mru_ghost->arcs_hits[ARC_BUFC_DATA]);
wmsum_fini(&arc_mru_ghost->arcs_hits[ARC_BUFC_METADATA]);
wmsum_fini(&arc_mfu_ghost->arcs_hits[ARC_BUFC_DATA]);
wmsum_fini(&arc_mfu_ghost->arcs_hits[ARC_BUFC_METADATA]);
wmsum_fini(&arc_sums.arcstat_hits);
wmsum_fini(&arc_sums.arcstat_iohits);
wmsum_fini(&arc_sums.arcstat_misses);
wmsum_fini(&arc_sums.arcstat_demand_data_hits);
wmsum_fini(&arc_sums.arcstat_demand_data_iohits);
wmsum_fini(&arc_sums.arcstat_demand_data_misses);
wmsum_fini(&arc_sums.arcstat_demand_metadata_hits);
wmsum_fini(&arc_sums.arcstat_demand_metadata_iohits);
wmsum_fini(&arc_sums.arcstat_demand_metadata_misses);
wmsum_fini(&arc_sums.arcstat_prefetch_data_hits);
wmsum_fini(&arc_sums.arcstat_prefetch_data_iohits);
wmsum_fini(&arc_sums.arcstat_prefetch_data_misses);
wmsum_fini(&arc_sums.arcstat_prefetch_metadata_hits);
wmsum_fini(&arc_sums.arcstat_prefetch_metadata_iohits);
wmsum_fini(&arc_sums.arcstat_prefetch_metadata_misses);
wmsum_fini(&arc_sums.arcstat_mru_hits);
wmsum_fini(&arc_sums.arcstat_mru_ghost_hits);
wmsum_fini(&arc_sums.arcstat_mfu_hits);
wmsum_fini(&arc_sums.arcstat_mfu_ghost_hits);
wmsum_fini(&arc_sums.arcstat_uncached_hits);
wmsum_fini(&arc_sums.arcstat_deleted);
wmsum_fini(&arc_sums.arcstat_mutex_miss);
wmsum_fini(&arc_sums.arcstat_access_skip);
wmsum_fini(&arc_sums.arcstat_evict_skip);
wmsum_fini(&arc_sums.arcstat_evict_not_enough);
wmsum_fini(&arc_sums.arcstat_evict_l2_cached);
wmsum_fini(&arc_sums.arcstat_evict_l2_eligible);
wmsum_fini(&arc_sums.arcstat_evict_l2_eligible_mfu);
wmsum_fini(&arc_sums.arcstat_evict_l2_eligible_mru);
wmsum_fini(&arc_sums.arcstat_evict_l2_ineligible);
wmsum_fini(&arc_sums.arcstat_evict_l2_skip);
wmsum_fini(&arc_sums.arcstat_hash_collisions);
wmsum_fini(&arc_sums.arcstat_hash_chains);
aggsum_fini(&arc_sums.arcstat_size);
wmsum_fini(&arc_sums.arcstat_compressed_size);
wmsum_fini(&arc_sums.arcstat_uncompressed_size);
wmsum_fini(&arc_sums.arcstat_overhead_size);
wmsum_fini(&arc_sums.arcstat_hdr_size);
wmsum_fini(&arc_sums.arcstat_data_size);
wmsum_fini(&arc_sums.arcstat_metadata_size);
wmsum_fini(&arc_sums.arcstat_dbuf_size);
wmsum_fini(&arc_sums.arcstat_dnode_size);
wmsum_fini(&arc_sums.arcstat_bonus_size);
wmsum_fini(&arc_sums.arcstat_l2_hits);
wmsum_fini(&arc_sums.arcstat_l2_misses);
wmsum_fini(&arc_sums.arcstat_l2_prefetch_asize);
wmsum_fini(&arc_sums.arcstat_l2_mru_asize);
wmsum_fini(&arc_sums.arcstat_l2_mfu_asize);
wmsum_fini(&arc_sums.arcstat_l2_bufc_data_asize);
wmsum_fini(&arc_sums.arcstat_l2_bufc_metadata_asize);
wmsum_fini(&arc_sums.arcstat_l2_feeds);
wmsum_fini(&arc_sums.arcstat_l2_rw_clash);
wmsum_fini(&arc_sums.arcstat_l2_read_bytes);
wmsum_fini(&arc_sums.arcstat_l2_write_bytes);
wmsum_fini(&arc_sums.arcstat_l2_writes_sent);
wmsum_fini(&arc_sums.arcstat_l2_writes_done);
wmsum_fini(&arc_sums.arcstat_l2_writes_error);
wmsum_fini(&arc_sums.arcstat_l2_writes_lock_retry);
wmsum_fini(&arc_sums.arcstat_l2_evict_lock_retry);
wmsum_fini(&arc_sums.arcstat_l2_evict_reading);
wmsum_fini(&arc_sums.arcstat_l2_evict_l1cached);
wmsum_fini(&arc_sums.arcstat_l2_free_on_write);
wmsum_fini(&arc_sums.arcstat_l2_abort_lowmem);
wmsum_fini(&arc_sums.arcstat_l2_cksum_bad);
wmsum_fini(&arc_sums.arcstat_l2_io_error);
wmsum_fini(&arc_sums.arcstat_l2_lsize);
wmsum_fini(&arc_sums.arcstat_l2_psize);
aggsum_fini(&arc_sums.arcstat_l2_hdr_size);
wmsum_fini(&arc_sums.arcstat_l2_log_blk_writes);
wmsum_fini(&arc_sums.arcstat_l2_log_blk_asize);
wmsum_fini(&arc_sums.arcstat_l2_log_blk_count);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_success);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_abort_unsupported);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_abort_io_errors);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_abort_dh_errors);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_abort_cksum_lb_errors);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_abort_lowmem);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_size);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_asize);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_bufs);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_bufs_precached);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_log_blks);
wmsum_fini(&arc_sums.arcstat_memory_throttle_count);
wmsum_fini(&arc_sums.arcstat_memory_direct_count);
wmsum_fini(&arc_sums.arcstat_memory_indirect_count);
wmsum_fini(&arc_sums.arcstat_prune);
wmsum_fini(&arc_sums.arcstat_meta_used);
wmsum_fini(&arc_sums.arcstat_async_upgrade_sync);
wmsum_fini(&arc_sums.arcstat_predictive_prefetch);
wmsum_fini(&arc_sums.arcstat_demand_hit_predictive_prefetch);
wmsum_fini(&arc_sums.arcstat_demand_iohit_predictive_prefetch);
wmsum_fini(&arc_sums.arcstat_prescient_prefetch);
wmsum_fini(&arc_sums.arcstat_demand_hit_prescient_prefetch);
wmsum_fini(&arc_sums.arcstat_demand_iohit_prescient_prefetch);
wmsum_fini(&arc_sums.arcstat_raw_size);
wmsum_fini(&arc_sums.arcstat_cached_only_in_progress);
wmsum_fini(&arc_sums.arcstat_abd_chunk_waste_size);
}
uint64_t
arc_target_bytes(void)
{
return (arc_c);
}
void
arc_set_limits(uint64_t allmem)
{
/* Set min cache to 1/32 of all memory, or 32MB, whichever is more. */
arc_c_min = MAX(allmem / 32, 2ULL << SPA_MAXBLOCKSHIFT);
/* How to set default max varies by platform. */
arc_c_max = arc_default_max(arc_c_min, allmem);
}
void
arc_init(void)
{
uint64_t percent, allmem = arc_all_memory();
mutex_init(&arc_evict_lock, NULL, MUTEX_DEFAULT, NULL);
list_create(&arc_evict_waiters, sizeof (arc_evict_waiter_t),
offsetof(arc_evict_waiter_t, aew_node));
arc_min_prefetch_ms = 1000;
arc_min_prescient_prefetch_ms = 6000;
#if defined(_KERNEL)
arc_lowmem_init();
#endif
arc_set_limits(allmem);
#ifdef _KERNEL
/*
* If zfs_arc_max is non-zero at init, meaning it was set in the kernel
* environment before the module was loaded, don't block setting the
* maximum because it is less than arc_c_min, instead, reset arc_c_min
* to a lower value.
* zfs_arc_min will be handled by arc_tuning_update().
*/
if (zfs_arc_max != 0 && zfs_arc_max >= MIN_ARC_MAX &&
zfs_arc_max < allmem) {
arc_c_max = zfs_arc_max;
if (arc_c_min >= arc_c_max) {
arc_c_min = MAX(zfs_arc_max / 2,
2ULL << SPA_MAXBLOCKSHIFT);
}
}
#else
/*
* In userland, there's only the memory pressure that we artificially
* create (see arc_available_memory()). Don't let arc_c get too
* small, because it can cause transactions to be larger than
* arc_c, causing arc_tempreserve_space() to fail.
*/
arc_c_min = MAX(arc_c_max / 2, 2ULL << SPA_MAXBLOCKSHIFT);
#endif
arc_c = arc_c_min;
/*
* 32-bit fixed point fractions of metadata from total ARC size,
* MRU data from all data and MRU metadata from all metadata.
*/
arc_meta = (1ULL << 32) / 4; /* Metadata is 25% of arc_c. */
arc_pd = (1ULL << 32) / 2; /* Data MRU is 50% of data. */
arc_pm = (1ULL << 32) / 2; /* Metadata MRU is 50% of metadata. */
percent = MIN(zfs_arc_dnode_limit_percent, 100);
arc_dnode_limit = arc_c_max * percent / 100;
/* Apply user specified tunings */
arc_tuning_update(B_TRUE);
/* if kmem_flags are set, lets try to use less memory */
if (kmem_debugging())
arc_c = arc_c / 2;
if (arc_c < arc_c_min)
arc_c = arc_c_min;
arc_register_hotplug();
arc_state_init();
buf_init();
list_create(&arc_prune_list, sizeof (arc_prune_t),
offsetof(arc_prune_t, p_node));
mutex_init(&arc_prune_mtx, NULL, MUTEX_DEFAULT, NULL);
arc_prune_taskq = taskq_create("arc_prune", zfs_arc_prune_task_threads,
defclsyspri, 100, INT_MAX, TASKQ_PREPOPULATE | TASKQ_DYNAMIC);
arc_ksp = kstat_create("zfs", 0, "arcstats", "misc", KSTAT_TYPE_NAMED,
sizeof (arc_stats) / sizeof (kstat_named_t), KSTAT_FLAG_VIRTUAL);
if (arc_ksp != NULL) {
arc_ksp->ks_data = &arc_stats;
arc_ksp->ks_update = arc_kstat_update;
kstat_install(arc_ksp);
}
arc_state_evict_markers =
arc_state_alloc_markers(arc_state_evict_marker_count);
arc_evict_zthr = zthr_create_timer("arc_evict",
arc_evict_cb_check, arc_evict_cb, NULL, SEC2NSEC(1), defclsyspri);
arc_reap_zthr = zthr_create_timer("arc_reap",
arc_reap_cb_check, arc_reap_cb, NULL, SEC2NSEC(1), minclsyspri);
arc_warm = B_FALSE;
/*
* Calculate maximum amount of dirty data per pool.
*
* If it has been set by a module parameter, take that.
* Otherwise, use a percentage of physical memory defined by
* zfs_dirty_data_max_percent (default 10%) with a cap at
* zfs_dirty_data_max_max (default 4G or 25% of physical memory).
*/
#ifdef __LP64__
if (zfs_dirty_data_max_max == 0)
zfs_dirty_data_max_max = MIN(4ULL * 1024 * 1024 * 1024,
allmem * zfs_dirty_data_max_max_percent / 100);
#else
if (zfs_dirty_data_max_max == 0)
zfs_dirty_data_max_max = MIN(1ULL * 1024 * 1024 * 1024,
allmem * zfs_dirty_data_max_max_percent / 100);
#endif
if (zfs_dirty_data_max == 0) {
zfs_dirty_data_max = allmem *
zfs_dirty_data_max_percent / 100;
zfs_dirty_data_max = MIN(zfs_dirty_data_max,
zfs_dirty_data_max_max);
}
if (zfs_wrlog_data_max == 0) {
/*
* dp_wrlog_total is reduced for each txg at the end of
* spa_sync(). However, dp_dirty_total is reduced every time
* a block is written out. Thus under normal operation,
* dp_wrlog_total could grow 2 times as big as
* zfs_dirty_data_max.
*/
zfs_wrlog_data_max = zfs_dirty_data_max * 2;
}
}
void
arc_fini(void)
{
arc_prune_t *p;
#ifdef _KERNEL
arc_lowmem_fini();
#endif /* _KERNEL */
/* Use B_TRUE to ensure *all* buffers are evicted */
arc_flush(NULL, B_TRUE);
if (arc_ksp != NULL) {
kstat_delete(arc_ksp);
arc_ksp = NULL;
}
taskq_wait(arc_prune_taskq);
taskq_destroy(arc_prune_taskq);
mutex_enter(&arc_prune_mtx);
while ((p = list_remove_head(&arc_prune_list)) != NULL) {
zfs_refcount_remove(&p->p_refcnt, &arc_prune_list);
zfs_refcount_destroy(&p->p_refcnt);
kmem_free(p, sizeof (*p));
}
mutex_exit(&arc_prune_mtx);
list_destroy(&arc_prune_list);
mutex_destroy(&arc_prune_mtx);
(void) zthr_cancel(arc_evict_zthr);
(void) zthr_cancel(arc_reap_zthr);
arc_state_free_markers(arc_state_evict_markers,
arc_state_evict_marker_count);
mutex_destroy(&arc_evict_lock);
list_destroy(&arc_evict_waiters);
/*
* Free any buffers that were tagged for destruction. This needs
* to occur before arc_state_fini() runs and destroys the aggsum
* values which are updated when freeing scatter ABDs.
*/
l2arc_do_free_on_write();
/*
* buf_fini() must proceed arc_state_fini() because buf_fin() may
* trigger the release of kmem magazines, which can callback to
* arc_space_return() which accesses aggsums freed in act_state_fini().
*/
buf_fini();
arc_state_fini();
arc_unregister_hotplug();
/*
* We destroy the zthrs after all the ARC state has been
* torn down to avoid the case of them receiving any
* wakeup() signals after they are destroyed.
*/
zthr_destroy(arc_evict_zthr);
zthr_destroy(arc_reap_zthr);
ASSERT0(arc_loaned_bytes);
}
/*
* Level 2 ARC
*
* The level 2 ARC (L2ARC) is a cache layer in-between main memory and disk.
* It uses dedicated storage devices to hold cached data, which are populated
* using large infrequent writes. The main role of this cache is to boost
* the performance of random read workloads. The intended L2ARC devices
* include short-stroked disks, solid state disks, and other media with
* substantially faster read latency than disk.
*
* +-----------------------+
* | ARC |
* +-----------------------+
* | ^ ^
* | | |
* l2arc_feed_thread() arc_read()
* | | |
* | l2arc read |
* V | |
* +---------------+ |
* | L2ARC | |
* +---------------+ |
* | ^ |
* l2arc_write() | |
* | | |
* V | |
* +-------+ +-------+
* | vdev | | vdev |
* | cache | | cache |
* +-------+ +-------+
* +=========+ .-----.
* : L2ARC : |-_____-|
* : devices : | Disks |
* +=========+ `-_____-'
*
* Read requests are satisfied from the following sources, in order:
*
* 1) ARC
* 2) vdev cache of L2ARC devices
* 3) L2ARC devices
* 4) vdev cache of disks
* 5) disks
*
* Some L2ARC device types exhibit extremely slow write performance.
* To accommodate for this there are some significant differences between
* the L2ARC and traditional cache design:
*
* 1. There is no eviction path from the ARC to the L2ARC. Evictions from
* the ARC behave as usual, freeing buffers and placing headers on ghost
* lists. The ARC does not send buffers to the L2ARC during eviction as
* this would add inflated write latencies for all ARC memory pressure.
*
* 2. The L2ARC attempts to cache data from the ARC before it is evicted.
* It does this by periodically scanning buffers from the eviction-end of
* the MFU and MRU ARC lists, copying them to the L2ARC devices if they are
* not already there. It scans until a headroom of buffers is satisfied,
* which itself is a buffer for ARC eviction. If a compressible buffer is
* found during scanning and selected for writing to an L2ARC device, we
* temporarily boost scanning headroom during the next scan cycle to make
* sure we adapt to compression effects (which might significantly reduce
* the data volume we write to L2ARC). The thread that does this is
* l2arc_feed_thread(), illustrated below; example sizes are included to
* provide a better sense of ratio than this diagram:
*
* head --> tail
* +---------------------+----------+
* ARC_mfu |:::::#:::::::::::::::|o#o###o###|-->. # already on L2ARC
* +---------------------+----------+ | o L2ARC eligible
* ARC_mru |:#:::::::::::::::::::|#o#ooo####|-->| : ARC buffer
* +---------------------+----------+ |
* 15.9 Gbytes ^ 32 Mbytes |
* headroom |
* l2arc_feed_thread()
* |
* l2arc write hand <--[oooo]--'
* | 8 Mbyte
* | write max
* V
* +==============================+
* L2ARC dev |####|#|###|###| |####| ... |
* +==============================+
* 32 Gbytes
*
* 3. If an ARC buffer is copied to the L2ARC but then hit instead of
* evicted, then the L2ARC has cached a buffer much sooner than it probably
* needed to, potentially wasting L2ARC device bandwidth and storage. It is
* safe to say that this is an uncommon case, since buffers at the end of
* the ARC lists have moved there due to inactivity.
*
* 4. If the ARC evicts faster than the L2ARC can maintain a headroom,
* then the L2ARC simply misses copying some buffers. This serves as a
* pressure valve to prevent heavy read workloads from both stalling the ARC
* with waits and clogging the L2ARC with writes. This also helps prevent
* the potential for the L2ARC to churn if it attempts to cache content too
* quickly, such as during backups of the entire pool.
*
* 5. After system boot and before the ARC has filled main memory, there are
* no evictions from the ARC and so the tails of the ARC_mfu and ARC_mru
* lists can remain mostly static. Instead of searching from tail of these
* lists as pictured, the l2arc_feed_thread() will search from the list heads
* for eligible buffers, greatly increasing its chance of finding them.
*
* The L2ARC device write speed is also boosted during this time so that
* the L2ARC warms up faster. Since there have been no ARC evictions yet,
* there are no L2ARC reads, and no fear of degrading read performance
* through increased writes.
*
* 6. Writes to the L2ARC devices are grouped and sent in-sequence, so that
* the vdev queue can aggregate them into larger and fewer writes. Each
* device is written to in a rotor fashion, sweeping writes through
* available space then repeating.
*
* 7. The L2ARC does not store dirty content. It never needs to flush
* write buffers back to disk based storage.
*
* 8. If an ARC buffer is written (and dirtied) which also exists in the
* L2ARC, the now stale L2ARC buffer is immediately dropped.
*
* The performance of the L2ARC can be tweaked by a number of tunables, which
* may be necessary for different workloads:
*
* l2arc_write_max max write bytes per interval
* l2arc_write_boost extra write bytes during device warmup
* l2arc_noprefetch skip caching prefetched buffers
* l2arc_headroom number of max device writes to precache
* l2arc_headroom_boost when we find compressed buffers during ARC
* scanning, we multiply headroom by this
* percentage factor for the next scan cycle,
* since more compressed buffers are likely to
* be present
* l2arc_feed_secs seconds between L2ARC writing
*
* Tunables may be removed or added as future performance improvements are
* integrated, and also may become zpool properties.
*
* There are three key functions that control how the L2ARC warms up:
*
* l2arc_write_eligible() check if a buffer is eligible to cache
* l2arc_write_size() calculate how much to write
* l2arc_write_interval() calculate sleep delay between writes
*
* These three functions determine what to write, how much, and how quickly
* to send writes.
*
* L2ARC persistence:
*
* When writing buffers to L2ARC, we periodically add some metadata to
* make sure we can pick them up after reboot, thus dramatically reducing
* the impact that any downtime has on the performance of storage systems
* with large caches.
*
* The implementation works fairly simply by integrating the following two
* modifications:
*
* *) When writing to the L2ARC, we occasionally write a "l2arc log block",
* which is an additional piece of metadata which describes what's been
* written. This allows us to rebuild the arc_buf_hdr_t structures of the
* main ARC buffers. There are 2 linked-lists of log blocks headed by
* dh_start_lbps[2]. We alternate which chain we append to, so they are
* time-wise and offset-wise interleaved, but that is an optimization rather
* than for correctness. The log block also includes a pointer to the
* previous block in its chain.
*
* *) We reserve SPA_MINBLOCKSIZE of space at the start of each L2ARC device
* for our header bookkeeping purposes. This contains a device header,
* which contains our top-level reference structures. We update it each
* time we write a new log block, so that we're able to locate it in the
* L2ARC device. If this write results in an inconsistent device header
* (e.g. due to power failure), we detect this by verifying the header's
* checksum and simply fail to reconstruct the L2ARC after reboot.
*
* Implementation diagram:
*
* +=== L2ARC device (not to scale) ======================================+
* | ___two newest log block pointers__.__________ |
* | / \dh_start_lbps[1] |
* | / \ \dh_start_lbps[0]|
* |.___/__. V V |
* ||L2 dev|....|lb |bufs |lb |bufs |lb |bufs |lb |bufs |lb |---(empty)---|
* || hdr| ^ /^ /^ / / |
* |+------+ ...--\-------/ \-----/--\------/ / |
* | \--------------/ \--------------/ |
* +======================================================================+
*
* As can be seen on the diagram, rather than using a simple linked list,
* we use a pair of linked lists with alternating elements. This is a
* performance enhancement due to the fact that we only find out the
* address of the next log block access once the current block has been
* completely read in. Obviously, this hurts performance, because we'd be
* keeping the device's I/O queue at only a 1 operation deep, thus
* incurring a large amount of I/O round-trip latency. Having two lists
* allows us to fetch two log blocks ahead of where we are currently
* rebuilding L2ARC buffers.
*
* On-device data structures:
*
* L2ARC device header: l2arc_dev_hdr_phys_t
* L2ARC log block: l2arc_log_blk_phys_t
*
* L2ARC reconstruction:
*
* When writing data, we simply write in the standard rotary fashion,
* evicting buffers as we go and simply writing new data over them (writing
* a new log block every now and then). This obviously means that once we
* loop around the end of the device, we will start cutting into an already
* committed log block (and its referenced data buffers), like so:
*
* current write head__ __old tail
* \ /
* V V
* <--|bufs |lb |bufs |lb | |bufs |lb |bufs |lb |-->
* ^ ^^^^^^^^^___________________________________
* | \
* <<nextwrite>> may overwrite this blk and/or its bufs --'
*
* When importing the pool, we detect this situation and use it to stop
* our scanning process (see l2arc_rebuild).
*
* There is one significant caveat to consider when rebuilding ARC contents
* from an L2ARC device: what about invalidated buffers? Given the above
* construction, we cannot update blocks which we've already written to amend
* them to remove buffers which were invalidated. Thus, during reconstruction,
* we might be populating the cache with buffers for data that's not on the
* main pool anymore, or may have been overwritten!
*
* As it turns out, this isn't a problem. Every arc_read request includes
* both the DVA and, crucially, the birth TXG of the BP the caller is
* looking for. So even if the cache were populated by completely rotten
* blocks for data that had been long deleted and/or overwritten, we'll
* never actually return bad data from the cache, since the DVA with the
* birth TXG uniquely identify a block in space and time - once created,
* a block is immutable on disk. The worst thing we have done is wasted
* some time and memory at l2arc rebuild to reconstruct outdated ARC
* entries that will get dropped from the l2arc as it is being updated
* with new blocks.
*
* L2ARC buffers that have been evicted by l2arc_evict() ahead of the write
* hand are not restored. This is done by saving the offset (in bytes)
* l2arc_evict() has evicted to in the L2ARC device header and taking it
* into account when restoring buffers.
*/
static boolean_t
l2arc_write_eligible(uint64_t spa_guid, arc_buf_hdr_t *hdr)
{
/*
* A buffer is *not* eligible for the L2ARC if it:
* 1. belongs to a different spa.
* 2. is already cached on the L2ARC.
* 3. has an I/O in progress (it may be an incomplete read).
* 4. is flagged not eligible (zfs property).
*/
if (hdr->b_spa != spa_guid || HDR_HAS_L2HDR(hdr) ||
HDR_IO_IN_PROGRESS(hdr) || !HDR_L2CACHE(hdr))
return (B_FALSE);
return (B_TRUE);
}
static uint64_t
l2arc_write_size(l2arc_dev_t *dev)
{
uint64_t size;
/*
* Make sure our globals have meaningful values in case the user
* altered them.
*/
size = l2arc_write_max;
if (size == 0) {
cmn_err(CE_NOTE, "Bad value for l2arc_write_max, value must "
"be greater than zero, resetting it to the default (%d)",
L2ARC_WRITE_SIZE);
size = l2arc_write_max = L2ARC_WRITE_SIZE;
}
if (arc_warm == B_FALSE)
size += l2arc_write_boost;
/* We need to add in the worst case scenario of log block overhead. */
size += l2arc_log_blk_overhead(size, dev);
if (dev->l2ad_vdev->vdev_has_trim && l2arc_trim_ahead > 0) {
/*
* Trim ahead of the write size 64MB or (l2arc_trim_ahead/100)
* times the writesize, whichever is greater.
*/
size += MAX(64 * 1024 * 1024,
(size * l2arc_trim_ahead) / 100);
}
/*
* Make sure the write size does not exceed the size of the cache
* device. This is important in l2arc_evict(), otherwise infinite
* iteration can occur.
*/
if (size > dev->l2ad_end - dev->l2ad_start) {
cmn_err(CE_NOTE, "l2arc_write_max or l2arc_write_boost "
"plus the overhead of log blocks (persistent L2ARC, "
"%llu bytes) exceeds the size of the cache device "
"(guid %llu), resetting them to the default (%d)",
(u_longlong_t)l2arc_log_blk_overhead(size, dev),
(u_longlong_t)dev->l2ad_vdev->vdev_guid, L2ARC_WRITE_SIZE);
size = l2arc_write_max = l2arc_write_boost = L2ARC_WRITE_SIZE;
if (l2arc_trim_ahead > 1) {
cmn_err(CE_NOTE, "l2arc_trim_ahead set to 1");
l2arc_trim_ahead = 1;
}
if (arc_warm == B_FALSE)
size += l2arc_write_boost;
size += l2arc_log_blk_overhead(size, dev);
if (dev->l2ad_vdev->vdev_has_trim && l2arc_trim_ahead > 0) {
size += MAX(64 * 1024 * 1024,
(size * l2arc_trim_ahead) / 100);
}
}
return (size);
}
static clock_t
l2arc_write_interval(clock_t began, uint64_t wanted, uint64_t wrote)
{
clock_t interval, next, now;
/*
* If the ARC lists are busy, increase our write rate; if the
* lists are stale, idle back. This is achieved by checking
* how much we previously wrote - if it was more than half of
* what we wanted, schedule the next write much sooner.
*/
if (l2arc_feed_again && wrote > (wanted / 2))
interval = (hz * l2arc_feed_min_ms) / 1000;
else
interval = hz * l2arc_feed_secs;
now = ddi_get_lbolt();
next = MAX(now, MIN(now + interval, began + interval));
return (next);
}
/*
* Cycle through L2ARC devices. This is how L2ARC load balances.
* If a device is returned, this also returns holding the spa config lock.
*/
static l2arc_dev_t *
l2arc_dev_get_next(void)
{
l2arc_dev_t *first, *next = NULL;
/*
* Lock out the removal of spas (spa_namespace_lock), then removal
* of cache devices (l2arc_dev_mtx). Once a device has been selected,
* both locks will be dropped and a spa config lock held instead.
*/
mutex_enter(&spa_namespace_lock);
mutex_enter(&l2arc_dev_mtx);
/* if there are no vdevs, there is nothing to do */
if (l2arc_ndev == 0)
goto out;
first = NULL;
next = l2arc_dev_last;
do {
/* loop around the list looking for a non-faulted vdev */
if (next == NULL) {
next = list_head(l2arc_dev_list);
} else {
next = list_next(l2arc_dev_list, next);
if (next == NULL)
next = list_head(l2arc_dev_list);
}
/* if we have come back to the start, bail out */
if (first == NULL)
first = next;
else if (next == first)
break;
ASSERT3P(next, !=, NULL);
} while (vdev_is_dead(next->l2ad_vdev) || next->l2ad_rebuild ||
next->l2ad_trim_all);
/* if we were unable to find any usable vdevs, return NULL */
if (vdev_is_dead(next->l2ad_vdev) || next->l2ad_rebuild ||
next->l2ad_trim_all)
next = NULL;
l2arc_dev_last = next;
out:
mutex_exit(&l2arc_dev_mtx);
/*
* Grab the config lock to prevent the 'next' device from being
* removed while we are writing to it.
*/
if (next != NULL)
spa_config_enter(next->l2ad_spa, SCL_L2ARC, next, RW_READER);
mutex_exit(&spa_namespace_lock);
return (next);
}
/*
* Free buffers that were tagged for destruction.
*/
static void
l2arc_do_free_on_write(void)
{
l2arc_data_free_t *df;
mutex_enter(&l2arc_free_on_write_mtx);
while ((df = list_remove_head(l2arc_free_on_write)) != NULL) {
ASSERT3P(df->l2df_abd, !=, NULL);
abd_free(df->l2df_abd);
kmem_free(df, sizeof (l2arc_data_free_t));
}
mutex_exit(&l2arc_free_on_write_mtx);
}
/*
* A write to a cache device has completed. Update all headers to allow
* reads from these buffers to begin.
*/
static void
l2arc_write_done(zio_t *zio)
{
l2arc_write_callback_t *cb;
l2arc_lb_abd_buf_t *abd_buf;
l2arc_lb_ptr_buf_t *lb_ptr_buf;
l2arc_dev_t *dev;
l2arc_dev_hdr_phys_t *l2dhdr;
list_t *buflist;
arc_buf_hdr_t *head, *hdr, *hdr_prev;
kmutex_t *hash_lock;
int64_t bytes_dropped = 0;
cb = zio->io_private;
ASSERT3P(cb, !=, NULL);
dev = cb->l2wcb_dev;
l2dhdr = dev->l2ad_dev_hdr;
ASSERT3P(dev, !=, NULL);
head = cb->l2wcb_head;
ASSERT3P(head, !=, NULL);
buflist = &dev->l2ad_buflist;
ASSERT3P(buflist, !=, NULL);
DTRACE_PROBE2(l2arc__iodone, zio_t *, zio,
l2arc_write_callback_t *, cb);
/*
* All writes completed, or an error was hit.
*/
top:
mutex_enter(&dev->l2ad_mtx);
for (hdr = list_prev(buflist, head); hdr; hdr = hdr_prev) {
hdr_prev = list_prev(buflist, hdr);
hash_lock = HDR_LOCK(hdr);
/*
* We cannot use mutex_enter or else we can deadlock
* with l2arc_write_buffers (due to swapping the order
* the hash lock and l2ad_mtx are taken).
*/
if (!mutex_tryenter(hash_lock)) {
/*
* Missed the hash lock. We must retry so we
* don't leave the ARC_FLAG_L2_WRITING bit set.
*/
ARCSTAT_BUMP(arcstat_l2_writes_lock_retry);
/*
* We don't want to rescan the headers we've
* already marked as having been written out, so
* we reinsert the head node so we can pick up
* where we left off.
*/
list_remove(buflist, head);
list_insert_after(buflist, hdr, head);
mutex_exit(&dev->l2ad_mtx);
/*
* We wait for the hash lock to become available
* to try and prevent busy waiting, and increase
* the chance we'll be able to acquire the lock
* the next time around.
*/
mutex_enter(hash_lock);
mutex_exit(hash_lock);
goto top;
}
/*
* We could not have been moved into the arc_l2c_only
* state while in-flight due to our ARC_FLAG_L2_WRITING
* bit being set. Let's just ensure that's being enforced.
*/
ASSERT(HDR_HAS_L1HDR(hdr));
/*
* Skipped - drop L2ARC entry and mark the header as no
* longer L2 eligibile.
*/
if (zio->io_error != 0) {
/*
* Error - drop L2ARC entry.
*/
list_remove(buflist, hdr);
arc_hdr_clear_flags(hdr, ARC_FLAG_HAS_L2HDR);
uint64_t psize = HDR_GET_PSIZE(hdr);
l2arc_hdr_arcstats_decrement(hdr);
bytes_dropped +=
vdev_psize_to_asize(dev->l2ad_vdev, psize);
(void) zfs_refcount_remove_many(&dev->l2ad_alloc,
arc_hdr_size(hdr), hdr);
}
/*
* Allow ARC to begin reads and ghost list evictions to
* this L2ARC entry.
*/
arc_hdr_clear_flags(hdr, ARC_FLAG_L2_WRITING);
mutex_exit(hash_lock);
}
/*
* Free the allocated abd buffers for writing the log blocks.
* If the zio failed reclaim the allocated space and remove the
* pointers to these log blocks from the log block pointer list
* of the L2ARC device.
*/
while ((abd_buf = list_remove_tail(&cb->l2wcb_abd_list)) != NULL) {
abd_free(abd_buf->abd);
zio_buf_free(abd_buf, sizeof (*abd_buf));
if (zio->io_error != 0) {
lb_ptr_buf = list_remove_head(&dev->l2ad_lbptr_list);
/*
* L2BLK_GET_PSIZE returns aligned size for log
* blocks.
*/
uint64_t asize =
L2BLK_GET_PSIZE((lb_ptr_buf->lb_ptr)->lbp_prop);
bytes_dropped += asize;
ARCSTAT_INCR(arcstat_l2_log_blk_asize, -asize);
ARCSTAT_BUMPDOWN(arcstat_l2_log_blk_count);
zfs_refcount_remove_many(&dev->l2ad_lb_asize, asize,
lb_ptr_buf);
zfs_refcount_remove(&dev->l2ad_lb_count, lb_ptr_buf);
kmem_free(lb_ptr_buf->lb_ptr,
sizeof (l2arc_log_blkptr_t));
kmem_free(lb_ptr_buf, sizeof (l2arc_lb_ptr_buf_t));
}
}
list_destroy(&cb->l2wcb_abd_list);
if (zio->io_error != 0) {
ARCSTAT_BUMP(arcstat_l2_writes_error);
/*
* Restore the lbps array in the header to its previous state.
* If the list of log block pointers is empty, zero out the
* log block pointers in the device header.
*/
lb_ptr_buf = list_head(&dev->l2ad_lbptr_list);
for (int i = 0; i < 2; i++) {
if (lb_ptr_buf == NULL) {
/*
* If the list is empty zero out the device
* header. Otherwise zero out the second log
* block pointer in the header.
*/
if (i == 0) {
memset(l2dhdr, 0,
dev->l2ad_dev_hdr_asize);
} else {
memset(&l2dhdr->dh_start_lbps[i], 0,
sizeof (l2arc_log_blkptr_t));
}
break;
}
memcpy(&l2dhdr->dh_start_lbps[i], lb_ptr_buf->lb_ptr,
sizeof (l2arc_log_blkptr_t));
lb_ptr_buf = list_next(&dev->l2ad_lbptr_list,
lb_ptr_buf);
}
}
ARCSTAT_BUMP(arcstat_l2_writes_done);
list_remove(buflist, head);
ASSERT(!HDR_HAS_L1HDR(head));
kmem_cache_free(hdr_l2only_cache, head);
mutex_exit(&dev->l2ad_mtx);
ASSERT(dev->l2ad_vdev != NULL);
vdev_space_update(dev->l2ad_vdev, -bytes_dropped, 0, 0);
l2arc_do_free_on_write();
kmem_free(cb, sizeof (l2arc_write_callback_t));
}
static int
l2arc_untransform(zio_t *zio, l2arc_read_callback_t *cb)
{
int ret;
spa_t *spa = zio->io_spa;
arc_buf_hdr_t *hdr = cb->l2rcb_hdr;
blkptr_t *bp = zio->io_bp;
uint8_t salt[ZIO_DATA_SALT_LEN];
uint8_t iv[ZIO_DATA_IV_LEN];
uint8_t mac[ZIO_DATA_MAC_LEN];
boolean_t no_crypt = B_FALSE;
/*
* ZIL data is never be written to the L2ARC, so we don't need
* special handling for its unique MAC storage.
*/
ASSERT3U(BP_GET_TYPE(bp), !=, DMU_OT_INTENT_LOG);
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
/*
* If the data was encrypted, decrypt it now. Note that
* we must check the bp here and not the hdr, since the
* hdr does not have its encryption parameters updated
* until arc_read_done().
*/
if (BP_IS_ENCRYPTED(bp)) {
abd_t *eabd = arc_get_data_abd(hdr, arc_hdr_size(hdr), hdr,
ARC_HDR_USE_RESERVE);
zio_crypt_decode_params_bp(bp, salt, iv);
zio_crypt_decode_mac_bp(bp, mac);
ret = spa_do_crypt_abd(B_FALSE, spa, &cb->l2rcb_zb,
BP_GET_TYPE(bp), BP_GET_DEDUP(bp), BP_SHOULD_BYTESWAP(bp),
salt, iv, mac, HDR_GET_PSIZE(hdr), eabd,
hdr->b_l1hdr.b_pabd, &no_crypt);
if (ret != 0) {
arc_free_data_abd(hdr, eabd, arc_hdr_size(hdr), hdr);
goto error;
}
/*
* If we actually performed decryption, replace b_pabd
* with the decrypted data. Otherwise we can just throw
* our decryption buffer away.
*/
if (!no_crypt) {
arc_free_data_abd(hdr, hdr->b_l1hdr.b_pabd,
arc_hdr_size(hdr), hdr);
hdr->b_l1hdr.b_pabd = eabd;
zio->io_abd = eabd;
} else {
arc_free_data_abd(hdr, eabd, arc_hdr_size(hdr), hdr);
}
}
/*
* If the L2ARC block was compressed, but ARC compression
* is disabled we decompress the data into a new buffer and
* replace the existing data.
*/
if (HDR_GET_COMPRESS(hdr) != ZIO_COMPRESS_OFF &&
!HDR_COMPRESSION_ENABLED(hdr)) {
abd_t *cabd = arc_get_data_abd(hdr, arc_hdr_size(hdr), hdr,
ARC_HDR_USE_RESERVE);
void *tmp = abd_borrow_buf(cabd, arc_hdr_size(hdr));
ret = zio_decompress_data(HDR_GET_COMPRESS(hdr),
hdr->b_l1hdr.b_pabd, tmp, HDR_GET_PSIZE(hdr),
HDR_GET_LSIZE(hdr), &hdr->b_complevel);
if (ret != 0) {
abd_return_buf_copy(cabd, tmp, arc_hdr_size(hdr));
arc_free_data_abd(hdr, cabd, arc_hdr_size(hdr), hdr);
goto error;
}
abd_return_buf_copy(cabd, tmp, arc_hdr_size(hdr));
arc_free_data_abd(hdr, hdr->b_l1hdr.b_pabd,
arc_hdr_size(hdr), hdr);
hdr->b_l1hdr.b_pabd = cabd;
zio->io_abd = cabd;
zio->io_size = HDR_GET_LSIZE(hdr);
}
return (0);
error:
return (ret);
}
/*
* A read to a cache device completed. Validate buffer contents before
* handing over to the regular ARC routines.
*/
static void
l2arc_read_done(zio_t *zio)
{
int tfm_error = 0;
l2arc_read_callback_t *cb = zio->io_private;
arc_buf_hdr_t *hdr;
kmutex_t *hash_lock;
boolean_t valid_cksum;
boolean_t using_rdata = (BP_IS_ENCRYPTED(&cb->l2rcb_bp) &&
(cb->l2rcb_flags & ZIO_FLAG_RAW_ENCRYPT));
ASSERT3P(zio->io_vd, !=, NULL);
ASSERT(zio->io_flags & ZIO_FLAG_DONT_PROPAGATE);
spa_config_exit(zio->io_spa, SCL_L2ARC, zio->io_vd);
ASSERT3P(cb, !=, NULL);
hdr = cb->l2rcb_hdr;
ASSERT3P(hdr, !=, NULL);
hash_lock = HDR_LOCK(hdr);
mutex_enter(hash_lock);
ASSERT3P(hash_lock, ==, HDR_LOCK(hdr));
/*
* If the data was read into a temporary buffer,
* move it and free the buffer.
*/
if (cb->l2rcb_abd != NULL) {
ASSERT3U(arc_hdr_size(hdr), <, zio->io_size);
if (zio->io_error == 0) {
if (using_rdata) {
abd_copy(hdr->b_crypt_hdr.b_rabd,
cb->l2rcb_abd, arc_hdr_size(hdr));
} else {
abd_copy(hdr->b_l1hdr.b_pabd,
cb->l2rcb_abd, arc_hdr_size(hdr));
}
}
/*
* The following must be done regardless of whether
* there was an error:
* - free the temporary buffer
* - point zio to the real ARC buffer
* - set zio size accordingly
* These are required because zio is either re-used for
* an I/O of the block in the case of the error
* or the zio is passed to arc_read_done() and it
* needs real data.
*/
abd_free(cb->l2rcb_abd);
zio->io_size = zio->io_orig_size = arc_hdr_size(hdr);
if (using_rdata) {
ASSERT(HDR_HAS_RABD(hdr));
zio->io_abd = zio->io_orig_abd =
hdr->b_crypt_hdr.b_rabd;
} else {
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
zio->io_abd = zio->io_orig_abd = hdr->b_l1hdr.b_pabd;
}
}
ASSERT3P(zio->io_abd, !=, NULL);
/*
* Check this survived the L2ARC journey.
*/
ASSERT(zio->io_abd == hdr->b_l1hdr.b_pabd ||
(HDR_HAS_RABD(hdr) && zio->io_abd == hdr->b_crypt_hdr.b_rabd));
zio->io_bp_copy = cb->l2rcb_bp; /* XXX fix in L2ARC 2.0 */
zio->io_bp = &zio->io_bp_copy; /* XXX fix in L2ARC 2.0 */
zio->io_prop.zp_complevel = hdr->b_complevel;
valid_cksum = arc_cksum_is_equal(hdr, zio);
/*
* b_rabd will always match the data as it exists on disk if it is
* being used. Therefore if we are reading into b_rabd we do not
* attempt to untransform the data.
*/
if (valid_cksum && !using_rdata)
tfm_error = l2arc_untransform(zio, cb);
if (valid_cksum && tfm_error == 0 && zio->io_error == 0 &&
!HDR_L2_EVICTED(hdr)) {
mutex_exit(hash_lock);
zio->io_private = hdr;
arc_read_done(zio);
} else {
/*
* Buffer didn't survive caching. Increment stats and
* reissue to the original storage device.
*/
if (zio->io_error != 0) {
ARCSTAT_BUMP(arcstat_l2_io_error);
} else {
zio->io_error = SET_ERROR(EIO);
}
if (!valid_cksum || tfm_error != 0)
ARCSTAT_BUMP(arcstat_l2_cksum_bad);
/*
* If there's no waiter, issue an async i/o to the primary
* storage now. If there *is* a waiter, the caller must
* issue the i/o in a context where it's OK to block.
*/
if (zio->io_waiter == NULL) {
zio_t *pio = zio_unique_parent(zio);
void *abd = (using_rdata) ?
hdr->b_crypt_hdr.b_rabd : hdr->b_l1hdr.b_pabd;
ASSERT(!pio || pio->io_child_type == ZIO_CHILD_LOGICAL);
zio = zio_read(pio, zio->io_spa, zio->io_bp,
abd, zio->io_size, arc_read_done,
hdr, zio->io_priority, cb->l2rcb_flags,
&cb->l2rcb_zb);
/*
* Original ZIO will be freed, so we need to update
* ARC header with the new ZIO pointer to be used
* by zio_change_priority() in arc_read().
*/
for (struct arc_callback *acb = hdr->b_l1hdr.b_acb;
acb != NULL; acb = acb->acb_next)
acb->acb_zio_head = zio;
mutex_exit(hash_lock);
zio_nowait(zio);
} else {
mutex_exit(hash_lock);
}
}
kmem_free(cb, sizeof (l2arc_read_callback_t));
}
/*
* This is the list priority from which the L2ARC will search for pages to
* cache. This is used within loops (0..3) to cycle through lists in the
* desired order. This order can have a significant effect on cache
* performance.
*
* Currently the metadata lists are hit first, MFU then MRU, followed by
* the data lists. This function returns a locked list, and also returns
* the lock pointer.
*/
static multilist_sublist_t *
l2arc_sublist_lock(int list_num)
{
multilist_t *ml = NULL;
unsigned int idx;
ASSERT(list_num >= 0 && list_num < L2ARC_FEED_TYPES);
switch (list_num) {
case 0:
ml = &arc_mfu->arcs_list[ARC_BUFC_METADATA];
break;
case 1:
ml = &arc_mru->arcs_list[ARC_BUFC_METADATA];
break;
case 2:
ml = &arc_mfu->arcs_list[ARC_BUFC_DATA];
break;
case 3:
ml = &arc_mru->arcs_list[ARC_BUFC_DATA];
break;
default:
return (NULL);
}
/*
* Return a randomly-selected sublist. This is acceptable
* because the caller feeds only a little bit of data for each
* call (8MB). Subsequent calls will result in different
* sublists being selected.
*/
idx = multilist_get_random_index(ml);
return (multilist_sublist_lock(ml, idx));
}
/*
* Calculates the maximum overhead of L2ARC metadata log blocks for a given
* L2ARC write size. l2arc_evict and l2arc_write_size need to include this
* overhead in processing to make sure there is enough headroom available
* when writing buffers.
*/
static inline uint64_t
l2arc_log_blk_overhead(uint64_t write_sz, l2arc_dev_t *dev)
{
if (dev->l2ad_log_entries == 0) {
return (0);
} else {
uint64_t log_entries = write_sz >> SPA_MINBLOCKSHIFT;
uint64_t log_blocks = (log_entries +
dev->l2ad_log_entries - 1) /
dev->l2ad_log_entries;
return (vdev_psize_to_asize(dev->l2ad_vdev,
sizeof (l2arc_log_blk_phys_t)) * log_blocks);
}
}
/*
* Evict buffers from the device write hand to the distance specified in
* bytes. This distance may span populated buffers, it may span nothing.
* This is clearing a region on the L2ARC device ready for writing.
* If the 'all' boolean is set, every buffer is evicted.
*/
static void
l2arc_evict(l2arc_dev_t *dev, uint64_t distance, boolean_t all)
{
list_t *buflist;
arc_buf_hdr_t *hdr, *hdr_prev;
kmutex_t *hash_lock;
uint64_t taddr;
l2arc_lb_ptr_buf_t *lb_ptr_buf, *lb_ptr_buf_prev;
vdev_t *vd = dev->l2ad_vdev;
boolean_t rerun;
buflist = &dev->l2ad_buflist;
top:
rerun = B_FALSE;
if (dev->l2ad_hand + distance > dev->l2ad_end) {
/*
* When there is no space to accommodate upcoming writes,
* evict to the end. Then bump the write and evict hands
* to the start and iterate. This iteration does not
* happen indefinitely as we make sure in
* l2arc_write_size() that when the write hand is reset,
* the write size does not exceed the end of the device.
*/
rerun = B_TRUE;
taddr = dev->l2ad_end;
} else {
taddr = dev->l2ad_hand + distance;
}
DTRACE_PROBE4(l2arc__evict, l2arc_dev_t *, dev, list_t *, buflist,
uint64_t, taddr, boolean_t, all);
if (!all) {
/*
* This check has to be placed after deciding whether to
* iterate (rerun).
*/
if (dev->l2ad_first) {
/*
* This is the first sweep through the device. There is
* nothing to evict. We have already trimmmed the
* whole device.
*/
goto out;
} else {
/*
* Trim the space to be evicted.
*/
if (vd->vdev_has_trim && dev->l2ad_evict < taddr &&
l2arc_trim_ahead > 0) {
/*
* We have to drop the spa_config lock because
* vdev_trim_range() will acquire it.
* l2ad_evict already accounts for the label
* size. To prevent vdev_trim_ranges() from
* adding it again, we subtract it from
* l2ad_evict.
*/
spa_config_exit(dev->l2ad_spa, SCL_L2ARC, dev);
vdev_trim_simple(vd,
dev->l2ad_evict - VDEV_LABEL_START_SIZE,
taddr - dev->l2ad_evict);
spa_config_enter(dev->l2ad_spa, SCL_L2ARC, dev,
RW_READER);
}
/*
* When rebuilding L2ARC we retrieve the evict hand
* from the header of the device. Of note, l2arc_evict()
* does not actually delete buffers from the cache
* device, but trimming may do so depending on the
* hardware implementation. Thus keeping track of the
* evict hand is useful.
*/
dev->l2ad_evict = MAX(dev->l2ad_evict, taddr);
}
}
retry:
mutex_enter(&dev->l2ad_mtx);
/*
* We have to account for evicted log blocks. Run vdev_space_update()
* on log blocks whose offset (in bytes) is before the evicted offset
* (in bytes) by searching in the list of pointers to log blocks
* present in the L2ARC device.
*/
for (lb_ptr_buf = list_tail(&dev->l2ad_lbptr_list); lb_ptr_buf;
lb_ptr_buf = lb_ptr_buf_prev) {
lb_ptr_buf_prev = list_prev(&dev->l2ad_lbptr_list, lb_ptr_buf);
/* L2BLK_GET_PSIZE returns aligned size for log blocks */
uint64_t asize = L2BLK_GET_PSIZE(
(lb_ptr_buf->lb_ptr)->lbp_prop);
/*
* We don't worry about log blocks left behind (ie
* lbp_payload_start < l2ad_hand) because l2arc_write_buffers()
* will never write more than l2arc_evict() evicts.
*/
if (!all && l2arc_log_blkptr_valid(dev, lb_ptr_buf->lb_ptr)) {
break;
} else {
vdev_space_update(vd, -asize, 0, 0);
ARCSTAT_INCR(arcstat_l2_log_blk_asize, -asize);
ARCSTAT_BUMPDOWN(arcstat_l2_log_blk_count);
zfs_refcount_remove_many(&dev->l2ad_lb_asize, asize,
lb_ptr_buf);
zfs_refcount_remove(&dev->l2ad_lb_count, lb_ptr_buf);
list_remove(&dev->l2ad_lbptr_list, lb_ptr_buf);
kmem_free(lb_ptr_buf->lb_ptr,
sizeof (l2arc_log_blkptr_t));
kmem_free(lb_ptr_buf, sizeof (l2arc_lb_ptr_buf_t));
}
}
for (hdr = list_tail(buflist); hdr; hdr = hdr_prev) {
hdr_prev = list_prev(buflist, hdr);
ASSERT(!HDR_EMPTY(hdr));
hash_lock = HDR_LOCK(hdr);
/*
* We cannot use mutex_enter or else we can deadlock
* with l2arc_write_buffers (due to swapping the order
* the hash lock and l2ad_mtx are taken).
*/
if (!mutex_tryenter(hash_lock)) {
/*
* Missed the hash lock. Retry.
*/
ARCSTAT_BUMP(arcstat_l2_evict_lock_retry);
mutex_exit(&dev->l2ad_mtx);
mutex_enter(hash_lock);
mutex_exit(hash_lock);
goto retry;
}
/*
* A header can't be on this list if it doesn't have L2 header.
*/
ASSERT(HDR_HAS_L2HDR(hdr));
/* Ensure this header has finished being written. */
ASSERT(!HDR_L2_WRITING(hdr));
ASSERT(!HDR_L2_WRITE_HEAD(hdr));
if (!all && (hdr->b_l2hdr.b_daddr >= dev->l2ad_evict ||
hdr->b_l2hdr.b_daddr < dev->l2ad_hand)) {
/*
* We've evicted to the target address,
* or the end of the device.
*/
mutex_exit(hash_lock);
break;
}
if (!HDR_HAS_L1HDR(hdr)) {
ASSERT(!HDR_L2_READING(hdr));
/*
* This doesn't exist in the ARC. Destroy.
* arc_hdr_destroy() will call list_remove()
* and decrement arcstat_l2_lsize.
*/
arc_change_state(arc_anon, hdr);
arc_hdr_destroy(hdr);
} else {
ASSERT(hdr->b_l1hdr.b_state != arc_l2c_only);
ARCSTAT_BUMP(arcstat_l2_evict_l1cached);
/*
* Invalidate issued or about to be issued
* reads, since we may be about to write
* over this location.
*/
if (HDR_L2_READING(hdr)) {
ARCSTAT_BUMP(arcstat_l2_evict_reading);
arc_hdr_set_flags(hdr, ARC_FLAG_L2_EVICTED);
}
arc_hdr_l2hdr_destroy(hdr);
}
mutex_exit(hash_lock);
}
mutex_exit(&dev->l2ad_mtx);
out:
/*
* We need to check if we evict all buffers, otherwise we may iterate
* unnecessarily.
*/
if (!all && rerun) {
/*
* Bump device hand to the device start if it is approaching the
* end. l2arc_evict() has already evicted ahead for this case.
*/
dev->l2ad_hand = dev->l2ad_start;
dev->l2ad_evict = dev->l2ad_start;
dev->l2ad_first = B_FALSE;
goto top;
}
if (!all) {
/*
* In case of cache device removal (all) the following
* assertions may be violated without functional consequences
* as the device is about to be removed.
*/
ASSERT3U(dev->l2ad_hand + distance, <, dev->l2ad_end);
if (!dev->l2ad_first)
ASSERT3U(dev->l2ad_hand, <=, dev->l2ad_evict);
}
}
/*
* Handle any abd transforms that might be required for writing to the L2ARC.
* If successful, this function will always return an abd with the data
* transformed as it is on disk in a new abd of asize bytes.
*/
static int
l2arc_apply_transforms(spa_t *spa, arc_buf_hdr_t *hdr, uint64_t asize,
abd_t **abd_out)
{
int ret;
void *tmp = NULL;
abd_t *cabd = NULL, *eabd = NULL, *to_write = hdr->b_l1hdr.b_pabd;
enum zio_compress compress = HDR_GET_COMPRESS(hdr);
uint64_t psize = HDR_GET_PSIZE(hdr);
uint64_t size = arc_hdr_size(hdr);
boolean_t ismd = HDR_ISTYPE_METADATA(hdr);
boolean_t bswap = (hdr->b_l1hdr.b_byteswap != DMU_BSWAP_NUMFUNCS);
dsl_crypto_key_t *dck = NULL;
uint8_t mac[ZIO_DATA_MAC_LEN] = { 0 };
boolean_t no_crypt = B_FALSE;
ASSERT((HDR_GET_COMPRESS(hdr) != ZIO_COMPRESS_OFF &&
!HDR_COMPRESSION_ENABLED(hdr)) ||
HDR_ENCRYPTED(hdr) || HDR_SHARED_DATA(hdr) || psize != asize);
ASSERT3U(psize, <=, asize);
/*
* If this data simply needs its own buffer, we simply allocate it
* and copy the data. This may be done to eliminate a dependency on a
* shared buffer or to reallocate the buffer to match asize.
*/
if (HDR_HAS_RABD(hdr) && asize != psize) {
ASSERT3U(asize, >=, psize);
to_write = abd_alloc_for_io(asize, ismd);
abd_copy(to_write, hdr->b_crypt_hdr.b_rabd, psize);
if (psize != asize)
abd_zero_off(to_write, psize, asize - psize);
goto out;
}
if ((compress == ZIO_COMPRESS_OFF || HDR_COMPRESSION_ENABLED(hdr)) &&
!HDR_ENCRYPTED(hdr)) {
ASSERT3U(size, ==, psize);
to_write = abd_alloc_for_io(asize, ismd);
abd_copy(to_write, hdr->b_l1hdr.b_pabd, size);
if (size != asize)
abd_zero_off(to_write, size, asize - size);
goto out;
}
if (compress != ZIO_COMPRESS_OFF && !HDR_COMPRESSION_ENABLED(hdr)) {
/*
* In some cases, we can wind up with size > asize, so
* we need to opt for the larger allocation option here.
*
* (We also need abd_return_buf_copy in all cases because
* it's an ASSERT() to modify the buffer before returning it
* with arc_return_buf(), and all the compressors
* write things before deciding to fail compression in nearly
* every case.)
*/
uint64_t bufsize = MAX(size, asize);
cabd = abd_alloc_for_io(bufsize, ismd);
tmp = abd_borrow_buf(cabd, bufsize);
psize = zio_compress_data(compress, to_write, &tmp, size,
hdr->b_complevel);
if (psize >= asize) {
psize = HDR_GET_PSIZE(hdr);
abd_return_buf_copy(cabd, tmp, bufsize);
HDR_SET_COMPRESS(hdr, ZIO_COMPRESS_OFF);
to_write = cabd;
abd_copy(to_write, hdr->b_l1hdr.b_pabd, psize);
if (psize != asize)
abd_zero_off(to_write, psize, asize - psize);
goto encrypt;
}
ASSERT3U(psize, <=, HDR_GET_PSIZE(hdr));
if (psize < asize)
memset((char *)tmp + psize, 0, bufsize - psize);
psize = HDR_GET_PSIZE(hdr);
abd_return_buf_copy(cabd, tmp, bufsize);
to_write = cabd;
}
encrypt:
if (HDR_ENCRYPTED(hdr)) {
eabd = abd_alloc_for_io(asize, ismd);
/*
* If the dataset was disowned before the buffer
* made it to this point, the key to re-encrypt
* it won't be available. In this case we simply
* won't write the buffer to the L2ARC.
*/
ret = spa_keystore_lookup_key(spa, hdr->b_crypt_hdr.b_dsobj,
FTAG, &dck);
if (ret != 0)
goto error;
ret = zio_do_crypt_abd(B_TRUE, &dck->dck_key,
hdr->b_crypt_hdr.b_ot, bswap, hdr->b_crypt_hdr.b_salt,
hdr->b_crypt_hdr.b_iv, mac, psize, to_write, eabd,
&no_crypt);
if (ret != 0)
goto error;
if (no_crypt)
abd_copy(eabd, to_write, psize);
if (psize != asize)
abd_zero_off(eabd, psize, asize - psize);
/* assert that the MAC we got here matches the one we saved */
ASSERT0(memcmp(mac, hdr->b_crypt_hdr.b_mac, ZIO_DATA_MAC_LEN));
spa_keystore_dsl_key_rele(spa, dck, FTAG);
if (to_write == cabd)
abd_free(cabd);
to_write = eabd;
}
out:
ASSERT3P(to_write, !=, hdr->b_l1hdr.b_pabd);
*abd_out = to_write;
return (0);
error:
if (dck != NULL)
spa_keystore_dsl_key_rele(spa, dck, FTAG);
if (cabd != NULL)
abd_free(cabd);
if (eabd != NULL)
abd_free(eabd);
*abd_out = NULL;
return (ret);
}
static void
l2arc_blk_fetch_done(zio_t *zio)
{
l2arc_read_callback_t *cb;
cb = zio->io_private;
if (cb->l2rcb_abd != NULL)
abd_free(cb->l2rcb_abd);
kmem_free(cb, sizeof (l2arc_read_callback_t));
}
/*
* Find and write ARC buffers to the L2ARC device.
*
* An ARC_FLAG_L2_WRITING flag is set so that the L2ARC buffers are not valid
* for reading until they have completed writing.
* The headroom_boost is an in-out parameter used to maintain headroom boost
* state between calls to this function.
*
* Returns the number of bytes actually written (which may be smaller than
* the delta by which the device hand has changed due to alignment and the
* writing of log blocks).
*/
static uint64_t
l2arc_write_buffers(spa_t *spa, l2arc_dev_t *dev, uint64_t target_sz)
{
arc_buf_hdr_t *hdr, *hdr_prev, *head;
uint64_t write_asize, write_psize, write_lsize, headroom;
boolean_t full;
l2arc_write_callback_t *cb = NULL;
zio_t *pio, *wzio;
uint64_t guid = spa_load_guid(spa);
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
ASSERT3P(dev->l2ad_vdev, !=, NULL);
pio = NULL;
write_lsize = write_asize = write_psize = 0;
full = B_FALSE;
head = kmem_cache_alloc(hdr_l2only_cache, KM_PUSHPAGE);
arc_hdr_set_flags(head, ARC_FLAG_L2_WRITE_HEAD | ARC_FLAG_HAS_L2HDR);
/*
* Copy buffers for L2ARC writing.
*/
for (int pass = 0; pass < L2ARC_FEED_TYPES; pass++) {
/*
* If pass == 1 or 3, we cache MRU metadata and data
* respectively.
*/
if (l2arc_mfuonly) {
if (pass == 1 || pass == 3)
continue;
}
multilist_sublist_t *mls = l2arc_sublist_lock(pass);
uint64_t passed_sz = 0;
VERIFY3P(mls, !=, NULL);
/*
* L2ARC fast warmup.
*
* Until the ARC is warm and starts to evict, read from the
* head of the ARC lists rather than the tail.
*/
if (arc_warm == B_FALSE)
hdr = multilist_sublist_head(mls);
else
hdr = multilist_sublist_tail(mls);
headroom = target_sz * l2arc_headroom;
if (zfs_compressed_arc_enabled)
headroom = (headroom * l2arc_headroom_boost) / 100;
for (; hdr; hdr = hdr_prev) {
kmutex_t *hash_lock;
abd_t *to_write = NULL;
if (arc_warm == B_FALSE)
hdr_prev = multilist_sublist_next(mls, hdr);
else
hdr_prev = multilist_sublist_prev(mls, hdr);
hash_lock = HDR_LOCK(hdr);
if (!mutex_tryenter(hash_lock)) {
/*
* Skip this buffer rather than waiting.
*/
continue;
}
passed_sz += HDR_GET_LSIZE(hdr);
if (l2arc_headroom != 0 && passed_sz > headroom) {
/*
* Searched too far.
*/
mutex_exit(hash_lock);
break;
}
if (!l2arc_write_eligible(guid, hdr)) {
mutex_exit(hash_lock);
continue;
}
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT3U(HDR_GET_PSIZE(hdr), >, 0);
ASSERT3U(arc_hdr_size(hdr), >, 0);
ASSERT(hdr->b_l1hdr.b_pabd != NULL ||
HDR_HAS_RABD(hdr));
uint64_t psize = HDR_GET_PSIZE(hdr);
uint64_t asize = vdev_psize_to_asize(dev->l2ad_vdev,
psize);
/*
* If the allocated size of this buffer plus the max
* size for the pending log block exceeds the evicted
* target size, terminate writing buffers for this run.
*/
if (write_asize + asize +
sizeof (l2arc_log_blk_phys_t) > target_sz) {
full = B_TRUE;
mutex_exit(hash_lock);
break;
}
/*
* We rely on the L1 portion of the header below, so
* it's invalid for this header to have been evicted out
* of the ghost cache, prior to being written out. The
* ARC_FLAG_L2_WRITING bit ensures this won't happen.
*/
arc_hdr_set_flags(hdr, ARC_FLAG_L2_WRITING);
/*
* If this header has b_rabd, we can use this since it
* must always match the data exactly as it exists on
* disk. Otherwise, the L2ARC can normally use the
* hdr's data, but if we're sharing data between the
* hdr and one of its bufs, L2ARC needs its own copy of
* the data so that the ZIO below can't race with the
* buf consumer. To ensure that this copy will be
* available for the lifetime of the ZIO and be cleaned
* up afterwards, we add it to the l2arc_free_on_write
* queue. If we need to apply any transforms to the
* data (compression, encryption) we will also need the
* extra buffer.
*/
if (HDR_HAS_RABD(hdr) && psize == asize) {
to_write = hdr->b_crypt_hdr.b_rabd;
} else if ((HDR_COMPRESSION_ENABLED(hdr) ||
HDR_GET_COMPRESS(hdr) == ZIO_COMPRESS_OFF) &&
!HDR_ENCRYPTED(hdr) && !HDR_SHARED_DATA(hdr) &&
psize == asize) {
to_write = hdr->b_l1hdr.b_pabd;
} else {
int ret;
arc_buf_contents_t type = arc_buf_type(hdr);
ret = l2arc_apply_transforms(spa, hdr, asize,
&to_write);
if (ret != 0) {
arc_hdr_clear_flags(hdr,
ARC_FLAG_L2_WRITING);
mutex_exit(hash_lock);
continue;
}
l2arc_free_abd_on_write(to_write, asize, type);
}
if (pio == NULL) {
/*
* Insert a dummy header on the buflist so
* l2arc_write_done() can find where the
* write buffers begin without searching.
*/
mutex_enter(&dev->l2ad_mtx);
list_insert_head(&dev->l2ad_buflist, head);
mutex_exit(&dev->l2ad_mtx);
cb = kmem_alloc(
sizeof (l2arc_write_callback_t), KM_SLEEP);
cb->l2wcb_dev = dev;
cb->l2wcb_head = head;
/*
* Create a list to save allocated abd buffers
* for l2arc_log_blk_commit().
*/
list_create(&cb->l2wcb_abd_list,
sizeof (l2arc_lb_abd_buf_t),
offsetof(l2arc_lb_abd_buf_t, node));
pio = zio_root(spa, l2arc_write_done, cb,
ZIO_FLAG_CANFAIL);
}
hdr->b_l2hdr.b_dev = dev;
hdr->b_l2hdr.b_hits = 0;
hdr->b_l2hdr.b_daddr = dev->l2ad_hand;
hdr->b_l2hdr.b_arcs_state =
hdr->b_l1hdr.b_state->arcs_state;
arc_hdr_set_flags(hdr, ARC_FLAG_HAS_L2HDR);
mutex_enter(&dev->l2ad_mtx);
list_insert_head(&dev->l2ad_buflist, hdr);
mutex_exit(&dev->l2ad_mtx);
(void) zfs_refcount_add_many(&dev->l2ad_alloc,
arc_hdr_size(hdr), hdr);
wzio = zio_write_phys(pio, dev->l2ad_vdev,
hdr->b_l2hdr.b_daddr, asize, to_write,
ZIO_CHECKSUM_OFF, NULL, hdr,
ZIO_PRIORITY_ASYNC_WRITE,
ZIO_FLAG_CANFAIL, B_FALSE);
write_lsize += HDR_GET_LSIZE(hdr);
DTRACE_PROBE2(l2arc__write, vdev_t *, dev->l2ad_vdev,
zio_t *, wzio);
write_psize += psize;
write_asize += asize;
dev->l2ad_hand += asize;
l2arc_hdr_arcstats_increment(hdr);
vdev_space_update(dev->l2ad_vdev, asize, 0, 0);
mutex_exit(hash_lock);
/*
* Append buf info to current log and commit if full.
* arcstat_l2_{size,asize} kstats are updated
* internally.
*/
if (l2arc_log_blk_insert(dev, hdr)) {
/*
* l2ad_hand will be adjusted in
* l2arc_log_blk_commit().
*/
write_asize +=
l2arc_log_blk_commit(dev, pio, cb);
}
zio_nowait(wzio);
}
multilist_sublist_unlock(mls);
if (full == B_TRUE)
break;
}
/* No buffers selected for writing? */
if (pio == NULL) {
ASSERT0(write_lsize);
ASSERT(!HDR_HAS_L1HDR(head));
kmem_cache_free(hdr_l2only_cache, head);
/*
* Although we did not write any buffers l2ad_evict may
* have advanced.
*/
if (dev->l2ad_evict != l2dhdr->dh_evict)
l2arc_dev_hdr_update(dev);
return (0);
}
if (!dev->l2ad_first)
ASSERT3U(dev->l2ad_hand, <=, dev->l2ad_evict);
ASSERT3U(write_asize, <=, target_sz);
ARCSTAT_BUMP(arcstat_l2_writes_sent);
ARCSTAT_INCR(arcstat_l2_write_bytes, write_psize);
dev->l2ad_writing = B_TRUE;
(void) zio_wait(pio);
dev->l2ad_writing = B_FALSE;
/*
* Update the device header after the zio completes as
* l2arc_write_done() may have updated the memory holding the log block
* pointers in the device header.
*/
l2arc_dev_hdr_update(dev);
return (write_asize);
}
static boolean_t
l2arc_hdr_limit_reached(void)
{
int64_t s = aggsum_upper_bound(&arc_sums.arcstat_l2_hdr_size);
return (arc_reclaim_needed() ||
(s > (arc_warm ? arc_c : arc_c_max) * l2arc_meta_percent / 100));
}
/*
* This thread feeds the L2ARC at regular intervals. This is the beating
* heart of the L2ARC.
*/
static __attribute__((noreturn)) void
l2arc_feed_thread(void *unused)
{
(void) unused;
callb_cpr_t cpr;
l2arc_dev_t *dev;
spa_t *spa;
uint64_t size, wrote;
clock_t begin, next = ddi_get_lbolt();
fstrans_cookie_t cookie;
CALLB_CPR_INIT(&cpr, &l2arc_feed_thr_lock, callb_generic_cpr, FTAG);
mutex_enter(&l2arc_feed_thr_lock);
cookie = spl_fstrans_mark();
while (l2arc_thread_exit == 0) {
CALLB_CPR_SAFE_BEGIN(&cpr);
(void) cv_timedwait_idle(&l2arc_feed_thr_cv,
&l2arc_feed_thr_lock, next);
CALLB_CPR_SAFE_END(&cpr, &l2arc_feed_thr_lock);
next = ddi_get_lbolt() + hz;
/*
* Quick check for L2ARC devices.
*/
mutex_enter(&l2arc_dev_mtx);
if (l2arc_ndev == 0) {
mutex_exit(&l2arc_dev_mtx);
continue;
}
mutex_exit(&l2arc_dev_mtx);
begin = ddi_get_lbolt();
/*
* This selects the next l2arc device to write to, and in
* doing so the next spa to feed from: dev->l2ad_spa. This
* will return NULL if there are now no l2arc devices or if
* they are all faulted.
*
* If a device is returned, its spa's config lock is also
* held to prevent device removal. l2arc_dev_get_next()
* will grab and release l2arc_dev_mtx.
*/
if ((dev = l2arc_dev_get_next()) == NULL)
continue;
spa = dev->l2ad_spa;
ASSERT3P(spa, !=, NULL);
/*
* If the pool is read-only then force the feed thread to
* sleep a little longer.
*/
if (!spa_writeable(spa)) {
next = ddi_get_lbolt() + 5 * l2arc_feed_secs * hz;
spa_config_exit(spa, SCL_L2ARC, dev);
continue;
}
/*
* Avoid contributing to memory pressure.
*/
if (l2arc_hdr_limit_reached()) {
ARCSTAT_BUMP(arcstat_l2_abort_lowmem);
spa_config_exit(spa, SCL_L2ARC, dev);
continue;
}
ARCSTAT_BUMP(arcstat_l2_feeds);
size = l2arc_write_size(dev);
/*
* Evict L2ARC buffers that will be overwritten.
*/
l2arc_evict(dev, size, B_FALSE);
/*
* Write ARC buffers.
*/
wrote = l2arc_write_buffers(spa, dev, size);
/*
* Calculate interval between writes.
*/
next = l2arc_write_interval(begin, size, wrote);
spa_config_exit(spa, SCL_L2ARC, dev);
}
spl_fstrans_unmark(cookie);
l2arc_thread_exit = 0;
cv_broadcast(&l2arc_feed_thr_cv);
CALLB_CPR_EXIT(&cpr); /* drops l2arc_feed_thr_lock */
thread_exit();
}
boolean_t
l2arc_vdev_present(vdev_t *vd)
{
return (l2arc_vdev_get(vd) != NULL);
}
/*
* Returns the l2arc_dev_t associated with a particular vdev_t or NULL if
* the vdev_t isn't an L2ARC device.
*/
l2arc_dev_t *
l2arc_vdev_get(vdev_t *vd)
{
l2arc_dev_t *dev;
mutex_enter(&l2arc_dev_mtx);
for (dev = list_head(l2arc_dev_list); dev != NULL;
dev = list_next(l2arc_dev_list, dev)) {
if (dev->l2ad_vdev == vd)
break;
}
mutex_exit(&l2arc_dev_mtx);
return (dev);
}
static void
l2arc_rebuild_dev(l2arc_dev_t *dev, boolean_t reopen)
{
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
uint64_t l2dhdr_asize = dev->l2ad_dev_hdr_asize;
spa_t *spa = dev->l2ad_spa;
/*
* The L2ARC has to hold at least the payload of one log block for
* them to be restored (persistent L2ARC). The payload of a log block
* depends on the amount of its log entries. We always write log blocks
* with 1022 entries. How many of them are committed or restored depends
* on the size of the L2ARC device. Thus the maximum payload of
* one log block is 1022 * SPA_MAXBLOCKSIZE = 16GB. If the L2ARC device
* is less than that, we reduce the amount of committed and restored
* log entries per block so as to enable persistence.
*/
if (dev->l2ad_end < l2arc_rebuild_blocks_min_l2size) {
dev->l2ad_log_entries = 0;
} else {
dev->l2ad_log_entries = MIN((dev->l2ad_end -
dev->l2ad_start) >> SPA_MAXBLOCKSHIFT,
L2ARC_LOG_BLK_MAX_ENTRIES);
}
/*
* Read the device header, if an error is returned do not rebuild L2ARC.
*/
if (l2arc_dev_hdr_read(dev) == 0 && dev->l2ad_log_entries > 0) {
/*
* If we are onlining a cache device (vdev_reopen) that was
* still present (l2arc_vdev_present()) and rebuild is enabled,
* we should evict all ARC buffers and pointers to log blocks
* and reclaim their space before restoring its contents to
* L2ARC.
*/
if (reopen) {
if (!l2arc_rebuild_enabled) {
return;
} else {
l2arc_evict(dev, 0, B_TRUE);
/* start a new log block */
dev->l2ad_log_ent_idx = 0;
dev->l2ad_log_blk_payload_asize = 0;
dev->l2ad_log_blk_payload_start = 0;
}
}
/*
* Just mark the device as pending for a rebuild. We won't
* be starting a rebuild in line here as it would block pool
* import. Instead spa_load_impl will hand that off to an
* async task which will call l2arc_spa_rebuild_start.
*/
dev->l2ad_rebuild = B_TRUE;
} else if (spa_writeable(spa)) {
/*
* In this case TRIM the whole device if l2arc_trim_ahead > 0,
* otherwise create a new header. We zero out the memory holding
* the header to reset dh_start_lbps. If we TRIM the whole
* device the new header will be written by
* vdev_trim_l2arc_thread() at the end of the TRIM to update the
* trim_state in the header too. When reading the header, if
* trim_state is not VDEV_TRIM_COMPLETE and l2arc_trim_ahead > 0
* we opt to TRIM the whole device again.
*/
if (l2arc_trim_ahead > 0) {
dev->l2ad_trim_all = B_TRUE;
} else {
memset(l2dhdr, 0, l2dhdr_asize);
l2arc_dev_hdr_update(dev);
}
}
}
/*
* Add a vdev for use by the L2ARC. By this point the spa has already
* validated the vdev and opened it.
*/
void
l2arc_add_vdev(spa_t *spa, vdev_t *vd)
{
l2arc_dev_t *adddev;
uint64_t l2dhdr_asize;
ASSERT(!l2arc_vdev_present(vd));
/*
* Create a new l2arc device entry.
*/
adddev = vmem_zalloc(sizeof (l2arc_dev_t), KM_SLEEP);
adddev->l2ad_spa = spa;
adddev->l2ad_vdev = vd;
/* leave extra size for an l2arc device header */
l2dhdr_asize = adddev->l2ad_dev_hdr_asize =
MAX(sizeof (*adddev->l2ad_dev_hdr), 1 << vd->vdev_ashift);
adddev->l2ad_start = VDEV_LABEL_START_SIZE + l2dhdr_asize;
adddev->l2ad_end = VDEV_LABEL_START_SIZE + vdev_get_min_asize(vd);
ASSERT3U(adddev->l2ad_start, <, adddev->l2ad_end);
adddev->l2ad_hand = adddev->l2ad_start;
adddev->l2ad_evict = adddev->l2ad_start;
adddev->l2ad_first = B_TRUE;
adddev->l2ad_writing = B_FALSE;
adddev->l2ad_trim_all = B_FALSE;
list_link_init(&adddev->l2ad_node);
adddev->l2ad_dev_hdr = kmem_zalloc(l2dhdr_asize, KM_SLEEP);
mutex_init(&adddev->l2ad_mtx, NULL, MUTEX_DEFAULT, NULL);
/*
* This is a list of all ARC buffers that are still valid on the
* device.
*/
list_create(&adddev->l2ad_buflist, sizeof (arc_buf_hdr_t),
offsetof(arc_buf_hdr_t, b_l2hdr.b_l2node));
/*
* This is a list of pointers to log blocks that are still present
* on the device.
*/
list_create(&adddev->l2ad_lbptr_list, sizeof (l2arc_lb_ptr_buf_t),
offsetof(l2arc_lb_ptr_buf_t, node));
vdev_space_update(vd, 0, 0, adddev->l2ad_end - adddev->l2ad_hand);
zfs_refcount_create(&adddev->l2ad_alloc);
zfs_refcount_create(&adddev->l2ad_lb_asize);
zfs_refcount_create(&adddev->l2ad_lb_count);
/*
* Decide if dev is eligible for L2ARC rebuild or whole device
* trimming. This has to happen before the device is added in the
* cache device list and l2arc_dev_mtx is released. Otherwise
* l2arc_feed_thread() might already start writing on the
* device.
*/
l2arc_rebuild_dev(adddev, B_FALSE);
/*
* Add device to global list
*/
mutex_enter(&l2arc_dev_mtx);
list_insert_head(l2arc_dev_list, adddev);
atomic_inc_64(&l2arc_ndev);
mutex_exit(&l2arc_dev_mtx);
}
/*
* Decide if a vdev is eligible for L2ARC rebuild, called from vdev_reopen()
* in case of onlining a cache device.
*/
void
l2arc_rebuild_vdev(vdev_t *vd, boolean_t reopen)
{
l2arc_dev_t *dev = NULL;
dev = l2arc_vdev_get(vd);
ASSERT3P(dev, !=, NULL);
/*
* In contrast to l2arc_add_vdev() we do not have to worry about
* l2arc_feed_thread() invalidating previous content when onlining a
* cache device. The device parameters (l2ad*) are not cleared when
* offlining the device and writing new buffers will not invalidate
* all previous content. In worst case only buffers that have not had
* their log block written to the device will be lost.
* When onlining the cache device (ie offline->online without exporting
* the pool in between) this happens:
* vdev_reopen() -> vdev_open() -> l2arc_rebuild_vdev()
* | |
* vdev_is_dead() = B_FALSE l2ad_rebuild = B_TRUE
* During the time where vdev_is_dead = B_FALSE and until l2ad_rebuild
* is set to B_TRUE we might write additional buffers to the device.
*/
l2arc_rebuild_dev(dev, reopen);
}
/*
* Remove a vdev from the L2ARC.
*/
void
l2arc_remove_vdev(vdev_t *vd)
{
l2arc_dev_t *remdev = NULL;
/*
* Find the device by vdev
*/
remdev = l2arc_vdev_get(vd);
ASSERT3P(remdev, !=, NULL);
/*
* Cancel any ongoing or scheduled rebuild.
*/
mutex_enter(&l2arc_rebuild_thr_lock);
if (remdev->l2ad_rebuild_began == B_TRUE) {
remdev->l2ad_rebuild_cancel = B_TRUE;
while (remdev->l2ad_rebuild == B_TRUE)
cv_wait(&l2arc_rebuild_thr_cv, &l2arc_rebuild_thr_lock);
}
mutex_exit(&l2arc_rebuild_thr_lock);
/*
* Remove device from global list
*/
mutex_enter(&l2arc_dev_mtx);
list_remove(l2arc_dev_list, remdev);
l2arc_dev_last = NULL; /* may have been invalidated */
atomic_dec_64(&l2arc_ndev);
mutex_exit(&l2arc_dev_mtx);
/*
* Clear all buflists and ARC references. L2ARC device flush.
*/
l2arc_evict(remdev, 0, B_TRUE);
list_destroy(&remdev->l2ad_buflist);
ASSERT(list_is_empty(&remdev->l2ad_lbptr_list));
list_destroy(&remdev->l2ad_lbptr_list);
mutex_destroy(&remdev->l2ad_mtx);
zfs_refcount_destroy(&remdev->l2ad_alloc);
zfs_refcount_destroy(&remdev->l2ad_lb_asize);
zfs_refcount_destroy(&remdev->l2ad_lb_count);
kmem_free(remdev->l2ad_dev_hdr, remdev->l2ad_dev_hdr_asize);
vmem_free(remdev, sizeof (l2arc_dev_t));
}
void
l2arc_init(void)
{
l2arc_thread_exit = 0;
l2arc_ndev = 0;
mutex_init(&l2arc_feed_thr_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&l2arc_feed_thr_cv, NULL, CV_DEFAULT, NULL);
mutex_init(&l2arc_rebuild_thr_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&l2arc_rebuild_thr_cv, NULL, CV_DEFAULT, NULL);
mutex_init(&l2arc_dev_mtx, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&l2arc_free_on_write_mtx, NULL, MUTEX_DEFAULT, NULL);
l2arc_dev_list = &L2ARC_dev_list;
l2arc_free_on_write = &L2ARC_free_on_write;
list_create(l2arc_dev_list, sizeof (l2arc_dev_t),
offsetof(l2arc_dev_t, l2ad_node));
list_create(l2arc_free_on_write, sizeof (l2arc_data_free_t),
offsetof(l2arc_data_free_t, l2df_list_node));
}
void
l2arc_fini(void)
{
mutex_destroy(&l2arc_feed_thr_lock);
cv_destroy(&l2arc_feed_thr_cv);
mutex_destroy(&l2arc_rebuild_thr_lock);
cv_destroy(&l2arc_rebuild_thr_cv);
mutex_destroy(&l2arc_dev_mtx);
mutex_destroy(&l2arc_free_on_write_mtx);
list_destroy(l2arc_dev_list);
list_destroy(l2arc_free_on_write);
}
void
l2arc_start(void)
{
if (!(spa_mode_global & SPA_MODE_WRITE))
return;
(void) thread_create(NULL, 0, l2arc_feed_thread, NULL, 0, &p0,
TS_RUN, defclsyspri);
}
void
l2arc_stop(void)
{
if (!(spa_mode_global & SPA_MODE_WRITE))
return;
mutex_enter(&l2arc_feed_thr_lock);
cv_signal(&l2arc_feed_thr_cv); /* kick thread out of startup */
l2arc_thread_exit = 1;
while (l2arc_thread_exit != 0)
cv_wait(&l2arc_feed_thr_cv, &l2arc_feed_thr_lock);
mutex_exit(&l2arc_feed_thr_lock);
}
/*
* Punches out rebuild threads for the L2ARC devices in a spa. This should
* be called after pool import from the spa async thread, since starting
* these threads directly from spa_import() will make them part of the
* "zpool import" context and delay process exit (and thus pool import).
*/
void
l2arc_spa_rebuild_start(spa_t *spa)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
/*
* Locate the spa's l2arc devices and kick off rebuild threads.
*/
for (int i = 0; i < spa->spa_l2cache.sav_count; i++) {
l2arc_dev_t *dev =
l2arc_vdev_get(spa->spa_l2cache.sav_vdevs[i]);
if (dev == NULL) {
/* Don't attempt a rebuild if the vdev is UNAVAIL */
continue;
}
mutex_enter(&l2arc_rebuild_thr_lock);
if (dev->l2ad_rebuild && !dev->l2ad_rebuild_cancel) {
dev->l2ad_rebuild_began = B_TRUE;
(void) thread_create(NULL, 0, l2arc_dev_rebuild_thread,
dev, 0, &p0, TS_RUN, minclsyspri);
}
mutex_exit(&l2arc_rebuild_thr_lock);
}
}
/*
* Main entry point for L2ARC rebuilding.
*/
static __attribute__((noreturn)) void
l2arc_dev_rebuild_thread(void *arg)
{
l2arc_dev_t *dev = arg;
VERIFY(!dev->l2ad_rebuild_cancel);
VERIFY(dev->l2ad_rebuild);
(void) l2arc_rebuild(dev);
mutex_enter(&l2arc_rebuild_thr_lock);
dev->l2ad_rebuild_began = B_FALSE;
dev->l2ad_rebuild = B_FALSE;
mutex_exit(&l2arc_rebuild_thr_lock);
thread_exit();
}
/*
* This function implements the actual L2ARC metadata rebuild. It:
* starts reading the log block chain and restores each block's contents
* to memory (reconstructing arc_buf_hdr_t's).
*
* Operation stops under any of the following conditions:
*
* 1) We reach the end of the log block chain.
* 2) We encounter *any* error condition (cksum errors, io errors)
*/
static int
l2arc_rebuild(l2arc_dev_t *dev)
{
vdev_t *vd = dev->l2ad_vdev;
spa_t *spa = vd->vdev_spa;
int err = 0;
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
l2arc_log_blk_phys_t *this_lb, *next_lb;
zio_t *this_io = NULL, *next_io = NULL;
l2arc_log_blkptr_t lbps[2];
l2arc_lb_ptr_buf_t *lb_ptr_buf;
boolean_t lock_held;
this_lb = vmem_zalloc(sizeof (*this_lb), KM_SLEEP);
next_lb = vmem_zalloc(sizeof (*next_lb), KM_SLEEP);
/*
* We prevent device removal while issuing reads to the device,
* then during the rebuilding phases we drop this lock again so
* that a spa_unload or device remove can be initiated - this is
* safe, because the spa will signal us to stop before removing
* our device and wait for us to stop.
*/
spa_config_enter(spa, SCL_L2ARC, vd, RW_READER);
lock_held = B_TRUE;
/*
* Retrieve the persistent L2ARC device state.
* L2BLK_GET_PSIZE returns aligned size for log blocks.
*/
dev->l2ad_evict = MAX(l2dhdr->dh_evict, dev->l2ad_start);
dev->l2ad_hand = MAX(l2dhdr->dh_start_lbps[0].lbp_daddr +
L2BLK_GET_PSIZE((&l2dhdr->dh_start_lbps[0])->lbp_prop),
dev->l2ad_start);
dev->l2ad_first = !!(l2dhdr->dh_flags & L2ARC_DEV_HDR_EVICT_FIRST);
vd->vdev_trim_action_time = l2dhdr->dh_trim_action_time;
vd->vdev_trim_state = l2dhdr->dh_trim_state;
/*
* In case the zfs module parameter l2arc_rebuild_enabled is false
* we do not start the rebuild process.
*/
if (!l2arc_rebuild_enabled)
goto out;
/* Prepare the rebuild process */
memcpy(lbps, l2dhdr->dh_start_lbps, sizeof (lbps));
/* Start the rebuild process */
for (;;) {
if (!l2arc_log_blkptr_valid(dev, &lbps[0]))
break;
if ((err = l2arc_log_blk_read(dev, &lbps[0], &lbps[1],
this_lb, next_lb, this_io, &next_io)) != 0)
goto out;
/*
* Our memory pressure valve. If the system is running low
* on memory, rather than swamping memory with new ARC buf
* hdrs, we opt not to rebuild the L2ARC. At this point,
* however, we have already set up our L2ARC dev to chain in
* new metadata log blocks, so the user may choose to offline/
* online the L2ARC dev at a later time (or re-import the pool)
* to reconstruct it (when there's less memory pressure).
*/
if (l2arc_hdr_limit_reached()) {
ARCSTAT_BUMP(arcstat_l2_rebuild_abort_lowmem);
cmn_err(CE_NOTE, "System running low on memory, "
"aborting L2ARC rebuild.");
err = SET_ERROR(ENOMEM);
goto out;
}
spa_config_exit(spa, SCL_L2ARC, vd);
lock_held = B_FALSE;
/*
* Now that we know that the next_lb checks out alright, we
* can start reconstruction from this log block.
* L2BLK_GET_PSIZE returns aligned size for log blocks.
*/
uint64_t asize = L2BLK_GET_PSIZE((&lbps[0])->lbp_prop);
l2arc_log_blk_restore(dev, this_lb, asize);
/*
* log block restored, include its pointer in the list of
* pointers to log blocks present in the L2ARC device.
*/
lb_ptr_buf = kmem_zalloc(sizeof (l2arc_lb_ptr_buf_t), KM_SLEEP);
lb_ptr_buf->lb_ptr = kmem_zalloc(sizeof (l2arc_log_blkptr_t),
KM_SLEEP);
memcpy(lb_ptr_buf->lb_ptr, &lbps[0],
sizeof (l2arc_log_blkptr_t));
mutex_enter(&dev->l2ad_mtx);
list_insert_tail(&dev->l2ad_lbptr_list, lb_ptr_buf);
ARCSTAT_INCR(arcstat_l2_log_blk_asize, asize);
ARCSTAT_BUMP(arcstat_l2_log_blk_count);
zfs_refcount_add_many(&dev->l2ad_lb_asize, asize, lb_ptr_buf);
zfs_refcount_add(&dev->l2ad_lb_count, lb_ptr_buf);
mutex_exit(&dev->l2ad_mtx);
vdev_space_update(vd, asize, 0, 0);
/*
* Protection against loops of log blocks:
*
* l2ad_hand l2ad_evict
* V V
* l2ad_start |=======================================| l2ad_end
* -----|||----|||---|||----|||
* (3) (2) (1) (0)
* ---|||---|||----|||---|||
* (7) (6) (5) (4)
*
* In this situation the pointer of log block (4) passes
* l2arc_log_blkptr_valid() but the log block should not be
* restored as it is overwritten by the payload of log block
* (0). Only log blocks (0)-(3) should be restored. We check
* whether l2ad_evict lies in between the payload starting
* offset of the next log block (lbps[1].lbp_payload_start)
* and the payload starting offset of the present log block
* (lbps[0].lbp_payload_start). If true and this isn't the
* first pass, we are looping from the beginning and we should
* stop.
*/
if (l2arc_range_check_overlap(lbps[1].lbp_payload_start,
lbps[0].lbp_payload_start, dev->l2ad_evict) &&
!dev->l2ad_first)
goto out;
kpreempt(KPREEMPT_SYNC);
for (;;) {
mutex_enter(&l2arc_rebuild_thr_lock);
if (dev->l2ad_rebuild_cancel) {
dev->l2ad_rebuild = B_FALSE;
cv_signal(&l2arc_rebuild_thr_cv);
mutex_exit(&l2arc_rebuild_thr_lock);
err = SET_ERROR(ECANCELED);
goto out;
}
mutex_exit(&l2arc_rebuild_thr_lock);
if (spa_config_tryenter(spa, SCL_L2ARC, vd,
RW_READER)) {
lock_held = B_TRUE;
break;
}
/*
* L2ARC config lock held by somebody in writer,
* possibly due to them trying to remove us. They'll
* likely to want us to shut down, so after a little
* delay, we check l2ad_rebuild_cancel and retry
* the lock again.
*/
delay(1);
}
/*
* Continue with the next log block.
*/
lbps[0] = lbps[1];
lbps[1] = this_lb->lb_prev_lbp;
PTR_SWAP(this_lb, next_lb);
this_io = next_io;
next_io = NULL;
}
if (this_io != NULL)
l2arc_log_blk_fetch_abort(this_io);
out:
if (next_io != NULL)
l2arc_log_blk_fetch_abort(next_io);
vmem_free(this_lb, sizeof (*this_lb));
vmem_free(next_lb, sizeof (*next_lb));
if (!l2arc_rebuild_enabled) {
spa_history_log_internal(spa, "L2ARC rebuild", NULL,
"disabled");
} else if (err == 0 && zfs_refcount_count(&dev->l2ad_lb_count) > 0) {
ARCSTAT_BUMP(arcstat_l2_rebuild_success);
spa_history_log_internal(spa, "L2ARC rebuild", NULL,
"successful, restored %llu blocks",
(u_longlong_t)zfs_refcount_count(&dev->l2ad_lb_count));
} else if (err == 0 && zfs_refcount_count(&dev->l2ad_lb_count) == 0) {
/*
* No error but also nothing restored, meaning the lbps array
* in the device header points to invalid/non-present log
* blocks. Reset the header.
*/
spa_history_log_internal(spa, "L2ARC rebuild", NULL,
"no valid log blocks");
memset(l2dhdr, 0, dev->l2ad_dev_hdr_asize);
l2arc_dev_hdr_update(dev);
} else if (err == ECANCELED) {
/*
* In case the rebuild was canceled do not log to spa history
* log as the pool may be in the process of being removed.
*/
zfs_dbgmsg("L2ARC rebuild aborted, restored %llu blocks",
(u_longlong_t)zfs_refcount_count(&dev->l2ad_lb_count));
} else if (err != 0) {
spa_history_log_internal(spa, "L2ARC rebuild", NULL,
"aborted, restored %llu blocks",
(u_longlong_t)zfs_refcount_count(&dev->l2ad_lb_count));
}
if (lock_held)
spa_config_exit(spa, SCL_L2ARC, vd);
return (err);
}
/*
* Attempts to read the device header on the provided L2ARC device and writes
* it to `hdr'. On success, this function returns 0, otherwise the appropriate
* error code is returned.
*/
static int
l2arc_dev_hdr_read(l2arc_dev_t *dev)
{
int err;
uint64_t guid;
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
const uint64_t l2dhdr_asize = dev->l2ad_dev_hdr_asize;
abd_t *abd;
guid = spa_guid(dev->l2ad_vdev->vdev_spa);
abd = abd_get_from_buf(l2dhdr, l2dhdr_asize);
err = zio_wait(zio_read_phys(NULL, dev->l2ad_vdev,
VDEV_LABEL_START_SIZE, l2dhdr_asize, abd,
ZIO_CHECKSUM_LABEL, NULL, NULL, ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_CANFAIL | ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY |
ZIO_FLAG_SPECULATIVE, B_FALSE));
abd_free(abd);
if (err != 0) {
ARCSTAT_BUMP(arcstat_l2_rebuild_abort_dh_errors);
zfs_dbgmsg("L2ARC IO error (%d) while reading device header, "
"vdev guid: %llu", err,
(u_longlong_t)dev->l2ad_vdev->vdev_guid);
return (err);
}
if (l2dhdr->dh_magic == BSWAP_64(L2ARC_DEV_HDR_MAGIC))
byteswap_uint64_array(l2dhdr, sizeof (*l2dhdr));
if (l2dhdr->dh_magic != L2ARC_DEV_HDR_MAGIC ||
l2dhdr->dh_spa_guid != guid ||
l2dhdr->dh_vdev_guid != dev->l2ad_vdev->vdev_guid ||
l2dhdr->dh_version != L2ARC_PERSISTENT_VERSION ||
l2dhdr->dh_log_entries != dev->l2ad_log_entries ||
l2dhdr->dh_end != dev->l2ad_end ||
!l2arc_range_check_overlap(dev->l2ad_start, dev->l2ad_end,
l2dhdr->dh_evict) ||
(l2dhdr->dh_trim_state != VDEV_TRIM_COMPLETE &&
l2arc_trim_ahead > 0)) {
/*
* Attempt to rebuild a device containing no actual dev hdr
* or containing a header from some other pool or from another
* version of persistent L2ARC.
*/
ARCSTAT_BUMP(arcstat_l2_rebuild_abort_unsupported);
return (SET_ERROR(ENOTSUP));
}
return (0);
}
/*
* Reads L2ARC log blocks from storage and validates their contents.
*
* This function implements a simple fetcher to make sure that while
* we're processing one buffer the L2ARC is already fetching the next
* one in the chain.
*
* The arguments this_lp and next_lp point to the current and next log block
* address in the block chain. Similarly, this_lb and next_lb hold the
* l2arc_log_blk_phys_t's of the current and next L2ARC blk.
*
* The `this_io' and `next_io' arguments are used for block fetching.
* When issuing the first blk IO during rebuild, you should pass NULL for
* `this_io'. This function will then issue a sync IO to read the block and
* also issue an async IO to fetch the next block in the block chain. The
* fetched IO is returned in `next_io'. On subsequent calls to this
* function, pass the value returned in `next_io' from the previous call
* as `this_io' and a fresh `next_io' pointer to hold the next fetch IO.
* Prior to the call, you should initialize your `next_io' pointer to be
* NULL. If no fetch IO was issued, the pointer is left set at NULL.
*
* On success, this function returns 0, otherwise it returns an appropriate
* error code. On error the fetching IO is aborted and cleared before
* returning from this function. Therefore, if we return `success', the
* caller can assume that we have taken care of cleanup of fetch IOs.
*/
static int
l2arc_log_blk_read(l2arc_dev_t *dev,
const l2arc_log_blkptr_t *this_lbp, const l2arc_log_blkptr_t *next_lbp,
l2arc_log_blk_phys_t *this_lb, l2arc_log_blk_phys_t *next_lb,
zio_t *this_io, zio_t **next_io)
{
int err = 0;
zio_cksum_t cksum;
abd_t *abd = NULL;
uint64_t asize;
ASSERT(this_lbp != NULL && next_lbp != NULL);
ASSERT(this_lb != NULL && next_lb != NULL);
ASSERT(next_io != NULL && *next_io == NULL);
ASSERT(l2arc_log_blkptr_valid(dev, this_lbp));
/*
* Check to see if we have issued the IO for this log block in a
* previous run. If not, this is the first call, so issue it now.
*/
if (this_io == NULL) {
this_io = l2arc_log_blk_fetch(dev->l2ad_vdev, this_lbp,
this_lb);
}
/*
* Peek to see if we can start issuing the next IO immediately.
*/
if (l2arc_log_blkptr_valid(dev, next_lbp)) {
/*
* Start issuing IO for the next log block early - this
* should help keep the L2ARC device busy while we
* decompress and restore this log block.
*/
*next_io = l2arc_log_blk_fetch(dev->l2ad_vdev, next_lbp,
next_lb);
}
/* Wait for the IO to read this log block to complete */
if ((err = zio_wait(this_io)) != 0) {
ARCSTAT_BUMP(arcstat_l2_rebuild_abort_io_errors);
zfs_dbgmsg("L2ARC IO error (%d) while reading log block, "
"offset: %llu, vdev guid: %llu", err,
(u_longlong_t)this_lbp->lbp_daddr,
(u_longlong_t)dev->l2ad_vdev->vdev_guid);
goto cleanup;
}
/*
* Make sure the buffer checks out.
* L2BLK_GET_PSIZE returns aligned size for log blocks.
*/
asize = L2BLK_GET_PSIZE((this_lbp)->lbp_prop);
fletcher_4_native(this_lb, asize, NULL, &cksum);
if (!ZIO_CHECKSUM_EQUAL(cksum, this_lbp->lbp_cksum)) {
ARCSTAT_BUMP(arcstat_l2_rebuild_abort_cksum_lb_errors);
zfs_dbgmsg("L2ARC log block cksum failed, offset: %llu, "
"vdev guid: %llu, l2ad_hand: %llu, l2ad_evict: %llu",
(u_longlong_t)this_lbp->lbp_daddr,
(u_longlong_t)dev->l2ad_vdev->vdev_guid,
(u_longlong_t)dev->l2ad_hand,
(u_longlong_t)dev->l2ad_evict);
err = SET_ERROR(ECKSUM);
goto cleanup;
}
/* Now we can take our time decoding this buffer */
switch (L2BLK_GET_COMPRESS((this_lbp)->lbp_prop)) {
case ZIO_COMPRESS_OFF:
break;
case ZIO_COMPRESS_LZ4:
abd = abd_alloc_for_io(asize, B_TRUE);
abd_copy_from_buf_off(abd, this_lb, 0, asize);
if ((err = zio_decompress_data(
L2BLK_GET_COMPRESS((this_lbp)->lbp_prop),
abd, this_lb, asize, sizeof (*this_lb), NULL)) != 0) {
err = SET_ERROR(EINVAL);
goto cleanup;
}
break;
default:
err = SET_ERROR(EINVAL);
goto cleanup;
}
if (this_lb->lb_magic == BSWAP_64(L2ARC_LOG_BLK_MAGIC))
byteswap_uint64_array(this_lb, sizeof (*this_lb));
if (this_lb->lb_magic != L2ARC_LOG_BLK_MAGIC) {
err = SET_ERROR(EINVAL);
goto cleanup;
}
cleanup:
/* Abort an in-flight fetch I/O in case of error */
if (err != 0 && *next_io != NULL) {
l2arc_log_blk_fetch_abort(*next_io);
*next_io = NULL;
}
if (abd != NULL)
abd_free(abd);
return (err);
}
/*
* Restores the payload of a log block to ARC. This creates empty ARC hdr
* entries which only contain an l2arc hdr, essentially restoring the
* buffers to their L2ARC evicted state. This function also updates space
* usage on the L2ARC vdev to make sure it tracks restored buffers.
*/
static void
l2arc_log_blk_restore(l2arc_dev_t *dev, const l2arc_log_blk_phys_t *lb,
uint64_t lb_asize)
{
uint64_t size = 0, asize = 0;
uint64_t log_entries = dev->l2ad_log_entries;
/*
* Usually arc_adapt() is called only for data, not headers, but
* since we may allocate significant amount of memory here, let ARC
* grow its arc_c.
*/
arc_adapt(log_entries * HDR_L2ONLY_SIZE);
for (int i = log_entries - 1; i >= 0; i--) {
/*
* Restore goes in the reverse temporal direction to preserve
* correct temporal ordering of buffers in the l2ad_buflist.
* l2arc_hdr_restore also does a list_insert_tail instead of
* list_insert_head on the l2ad_buflist:
*
* LIST l2ad_buflist LIST
* HEAD <------ (time) ------ TAIL
* direction +-----+-----+-----+-----+-----+ direction
* of l2arc <== | buf | buf | buf | buf | buf | ===> of rebuild
* fill +-----+-----+-----+-----+-----+
* ^ ^
* | |
* | |
* l2arc_feed_thread l2arc_rebuild
* will place new bufs here restores bufs here
*
* During l2arc_rebuild() the device is not used by
* l2arc_feed_thread() as dev->l2ad_rebuild is set to true.
*/
size += L2BLK_GET_LSIZE((&lb->lb_entries[i])->le_prop);
asize += vdev_psize_to_asize(dev->l2ad_vdev,
L2BLK_GET_PSIZE((&lb->lb_entries[i])->le_prop));
l2arc_hdr_restore(&lb->lb_entries[i], dev);
}
/*
* Record rebuild stats:
* size Logical size of restored buffers in the L2ARC
* asize Aligned size of restored buffers in the L2ARC
*/
ARCSTAT_INCR(arcstat_l2_rebuild_size, size);
ARCSTAT_INCR(arcstat_l2_rebuild_asize, asize);
ARCSTAT_INCR(arcstat_l2_rebuild_bufs, log_entries);
ARCSTAT_F_AVG(arcstat_l2_log_blk_avg_asize, lb_asize);
ARCSTAT_F_AVG(arcstat_l2_data_to_meta_ratio, asize / lb_asize);
ARCSTAT_BUMP(arcstat_l2_rebuild_log_blks);
}
/*
* Restores a single ARC buf hdr from a log entry. The ARC buffer is put
* into a state indicating that it has been evicted to L2ARC.
*/
static void
l2arc_hdr_restore(const l2arc_log_ent_phys_t *le, l2arc_dev_t *dev)
{
arc_buf_hdr_t *hdr, *exists;
kmutex_t *hash_lock;
arc_buf_contents_t type = L2BLK_GET_TYPE((le)->le_prop);
uint64_t asize;
/*
* Do all the allocation before grabbing any locks, this lets us
* sleep if memory is full and we don't have to deal with failed
* allocations.
*/
hdr = arc_buf_alloc_l2only(L2BLK_GET_LSIZE((le)->le_prop), type,
dev, le->le_dva, le->le_daddr,
L2BLK_GET_PSIZE((le)->le_prop), le->le_birth,
L2BLK_GET_COMPRESS((le)->le_prop), le->le_complevel,
L2BLK_GET_PROTECTED((le)->le_prop),
L2BLK_GET_PREFETCH((le)->le_prop),
L2BLK_GET_STATE((le)->le_prop));
asize = vdev_psize_to_asize(dev->l2ad_vdev,
L2BLK_GET_PSIZE((le)->le_prop));
/*
* vdev_space_update() has to be called before arc_hdr_destroy() to
* avoid underflow since the latter also calls vdev_space_update().
*/
l2arc_hdr_arcstats_increment(hdr);
vdev_space_update(dev->l2ad_vdev, asize, 0, 0);
mutex_enter(&dev->l2ad_mtx);
list_insert_tail(&dev->l2ad_buflist, hdr);
(void) zfs_refcount_add_many(&dev->l2ad_alloc, arc_hdr_size(hdr), hdr);
mutex_exit(&dev->l2ad_mtx);
exists = buf_hash_insert(hdr, &hash_lock);
if (exists) {
/* Buffer was already cached, no need to restore it. */
arc_hdr_destroy(hdr);
/*
* If the buffer is already cached, check whether it has
* L2ARC metadata. If not, enter them and update the flag.
* This is important is case of onlining a cache device, since
* we previously evicted all L2ARC metadata from ARC.
*/
if (!HDR_HAS_L2HDR(exists)) {
arc_hdr_set_flags(exists, ARC_FLAG_HAS_L2HDR);
exists->b_l2hdr.b_dev = dev;
exists->b_l2hdr.b_daddr = le->le_daddr;
exists->b_l2hdr.b_arcs_state =
L2BLK_GET_STATE((le)->le_prop);
mutex_enter(&dev->l2ad_mtx);
list_insert_tail(&dev->l2ad_buflist, exists);
(void) zfs_refcount_add_many(&dev->l2ad_alloc,
arc_hdr_size(exists), exists);
mutex_exit(&dev->l2ad_mtx);
l2arc_hdr_arcstats_increment(exists);
vdev_space_update(dev->l2ad_vdev, asize, 0, 0);
}
ARCSTAT_BUMP(arcstat_l2_rebuild_bufs_precached);
}
mutex_exit(hash_lock);
}
/*
* Starts an asynchronous read IO to read a log block. This is used in log
* block reconstruction to start reading the next block before we are done
* decoding and reconstructing the current block, to keep the l2arc device
* nice and hot with read IO to process.
* The returned zio will contain a newly allocated memory buffers for the IO
* data which should then be freed by the caller once the zio is no longer
* needed (i.e. due to it having completed). If you wish to abort this
* zio, you should do so using l2arc_log_blk_fetch_abort, which takes
* care of disposing of the allocated buffers correctly.
*/
static zio_t *
l2arc_log_blk_fetch(vdev_t *vd, const l2arc_log_blkptr_t *lbp,
l2arc_log_blk_phys_t *lb)
{
uint32_t asize;
zio_t *pio;
l2arc_read_callback_t *cb;
/* L2BLK_GET_PSIZE returns aligned size for log blocks */
asize = L2BLK_GET_PSIZE((lbp)->lbp_prop);
ASSERT(asize <= sizeof (l2arc_log_blk_phys_t));
cb = kmem_zalloc(sizeof (l2arc_read_callback_t), KM_SLEEP);
cb->l2rcb_abd = abd_get_from_buf(lb, asize);
pio = zio_root(vd->vdev_spa, l2arc_blk_fetch_done, cb,
ZIO_FLAG_CANFAIL | ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY);
(void) zio_nowait(zio_read_phys(pio, vd, lbp->lbp_daddr, asize,
cb->l2rcb_abd, ZIO_CHECKSUM_OFF, NULL, NULL,
ZIO_PRIORITY_ASYNC_READ, ZIO_FLAG_CANFAIL |
ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY, B_FALSE));
return (pio);
}
/*
* Aborts a zio returned from l2arc_log_blk_fetch and frees the data
* buffers allocated for it.
*/
static void
l2arc_log_blk_fetch_abort(zio_t *zio)
{
(void) zio_wait(zio);
}
/*
* Creates a zio to update the device header on an l2arc device.
*/
void
l2arc_dev_hdr_update(l2arc_dev_t *dev)
{
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
const uint64_t l2dhdr_asize = dev->l2ad_dev_hdr_asize;
abd_t *abd;
int err;
VERIFY(spa_config_held(dev->l2ad_spa, SCL_STATE_ALL, RW_READER));
l2dhdr->dh_magic = L2ARC_DEV_HDR_MAGIC;
l2dhdr->dh_version = L2ARC_PERSISTENT_VERSION;
l2dhdr->dh_spa_guid = spa_guid(dev->l2ad_vdev->vdev_spa);
l2dhdr->dh_vdev_guid = dev->l2ad_vdev->vdev_guid;
l2dhdr->dh_log_entries = dev->l2ad_log_entries;
l2dhdr->dh_evict = dev->l2ad_evict;
l2dhdr->dh_start = dev->l2ad_start;
l2dhdr->dh_end = dev->l2ad_end;
l2dhdr->dh_lb_asize = zfs_refcount_count(&dev->l2ad_lb_asize);
l2dhdr->dh_lb_count = zfs_refcount_count(&dev->l2ad_lb_count);
l2dhdr->dh_flags = 0;
l2dhdr->dh_trim_action_time = dev->l2ad_vdev->vdev_trim_action_time;
l2dhdr->dh_trim_state = dev->l2ad_vdev->vdev_trim_state;
if (dev->l2ad_first)
l2dhdr->dh_flags |= L2ARC_DEV_HDR_EVICT_FIRST;
abd = abd_get_from_buf(l2dhdr, l2dhdr_asize);
err = zio_wait(zio_write_phys(NULL, dev->l2ad_vdev,
VDEV_LABEL_START_SIZE, l2dhdr_asize, abd, ZIO_CHECKSUM_LABEL, NULL,
NULL, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_CANFAIL, B_FALSE));
abd_free(abd);
if (err != 0) {
zfs_dbgmsg("L2ARC IO error (%d) while writing device header, "
"vdev guid: %llu", err,
(u_longlong_t)dev->l2ad_vdev->vdev_guid);
}
}
/*
* Commits a log block to the L2ARC device. This routine is invoked from
* l2arc_write_buffers when the log block fills up.
* This function allocates some memory to temporarily hold the serialized
* buffer to be written. This is then released in l2arc_write_done.
*/
static uint64_t
l2arc_log_blk_commit(l2arc_dev_t *dev, zio_t *pio, l2arc_write_callback_t *cb)
{
l2arc_log_blk_phys_t *lb = &dev->l2ad_log_blk;
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
uint64_t psize, asize;
zio_t *wzio;
l2arc_lb_abd_buf_t *abd_buf;
uint8_t *tmpbuf = NULL;
l2arc_lb_ptr_buf_t *lb_ptr_buf;
VERIFY3S(dev->l2ad_log_ent_idx, ==, dev->l2ad_log_entries);
abd_buf = zio_buf_alloc(sizeof (*abd_buf));
abd_buf->abd = abd_get_from_buf(lb, sizeof (*lb));
lb_ptr_buf = kmem_zalloc(sizeof (l2arc_lb_ptr_buf_t), KM_SLEEP);
lb_ptr_buf->lb_ptr = kmem_zalloc(sizeof (l2arc_log_blkptr_t), KM_SLEEP);
/* link the buffer into the block chain */
lb->lb_prev_lbp = l2dhdr->dh_start_lbps[1];
lb->lb_magic = L2ARC_LOG_BLK_MAGIC;
/*
* l2arc_log_blk_commit() may be called multiple times during a single
* l2arc_write_buffers() call. Save the allocated abd buffers in a list
* so we can free them in l2arc_write_done() later on.
*/
list_insert_tail(&cb->l2wcb_abd_list, abd_buf);
/* try to compress the buffer */
psize = zio_compress_data(ZIO_COMPRESS_LZ4,
abd_buf->abd, (void **) &tmpbuf, sizeof (*lb), 0);
/* a log block is never entirely zero */
ASSERT(psize != 0);
asize = vdev_psize_to_asize(dev->l2ad_vdev, psize);
ASSERT(asize <= sizeof (*lb));
/*
* Update the start log block pointer in the device header to point
* to the log block we're about to write.
*/
l2dhdr->dh_start_lbps[1] = l2dhdr->dh_start_lbps[0];
l2dhdr->dh_start_lbps[0].lbp_daddr = dev->l2ad_hand;
l2dhdr->dh_start_lbps[0].lbp_payload_asize =
dev->l2ad_log_blk_payload_asize;
l2dhdr->dh_start_lbps[0].lbp_payload_start =
dev->l2ad_log_blk_payload_start;
L2BLK_SET_LSIZE(
(&l2dhdr->dh_start_lbps[0])->lbp_prop, sizeof (*lb));
L2BLK_SET_PSIZE(
(&l2dhdr->dh_start_lbps[0])->lbp_prop, asize);
L2BLK_SET_CHECKSUM(
(&l2dhdr->dh_start_lbps[0])->lbp_prop,
ZIO_CHECKSUM_FLETCHER_4);
if (asize < sizeof (*lb)) {
/* compression succeeded */
memset(tmpbuf + psize, 0, asize - psize);
L2BLK_SET_COMPRESS(
(&l2dhdr->dh_start_lbps[0])->lbp_prop,
ZIO_COMPRESS_LZ4);
} else {
/* compression failed */
memcpy(tmpbuf, lb, sizeof (*lb));
L2BLK_SET_COMPRESS(
(&l2dhdr->dh_start_lbps[0])->lbp_prop,
ZIO_COMPRESS_OFF);
}
/* checksum what we're about to write */
fletcher_4_native(tmpbuf, asize, NULL,
&l2dhdr->dh_start_lbps[0].lbp_cksum);
abd_free(abd_buf->abd);
/* perform the write itself */
abd_buf->abd = abd_get_from_buf(tmpbuf, sizeof (*lb));
abd_take_ownership_of_buf(abd_buf->abd, B_TRUE);
wzio = zio_write_phys(pio, dev->l2ad_vdev, dev->l2ad_hand,
asize, abd_buf->abd, ZIO_CHECKSUM_OFF, NULL, NULL,
ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_CANFAIL, B_FALSE);
DTRACE_PROBE2(l2arc__write, vdev_t *, dev->l2ad_vdev, zio_t *, wzio);
(void) zio_nowait(wzio);
dev->l2ad_hand += asize;
/*
* Include the committed log block's pointer in the list of pointers
* to log blocks present in the L2ARC device.
*/
memcpy(lb_ptr_buf->lb_ptr, &l2dhdr->dh_start_lbps[0],
sizeof (l2arc_log_blkptr_t));
mutex_enter(&dev->l2ad_mtx);
list_insert_head(&dev->l2ad_lbptr_list, lb_ptr_buf);
ARCSTAT_INCR(arcstat_l2_log_blk_asize, asize);
ARCSTAT_BUMP(arcstat_l2_log_blk_count);
zfs_refcount_add_many(&dev->l2ad_lb_asize, asize, lb_ptr_buf);
zfs_refcount_add(&dev->l2ad_lb_count, lb_ptr_buf);
mutex_exit(&dev->l2ad_mtx);
vdev_space_update(dev->l2ad_vdev, asize, 0, 0);
/* bump the kstats */
ARCSTAT_INCR(arcstat_l2_write_bytes, asize);
ARCSTAT_BUMP(arcstat_l2_log_blk_writes);
ARCSTAT_F_AVG(arcstat_l2_log_blk_avg_asize, asize);
ARCSTAT_F_AVG(arcstat_l2_data_to_meta_ratio,
dev->l2ad_log_blk_payload_asize / asize);
/* start a new log block */
dev->l2ad_log_ent_idx = 0;
dev->l2ad_log_blk_payload_asize = 0;
dev->l2ad_log_blk_payload_start = 0;
return (asize);
}
/*
* Validates an L2ARC log block address to make sure that it can be read
* from the provided L2ARC device.
*/
boolean_t
l2arc_log_blkptr_valid(l2arc_dev_t *dev, const l2arc_log_blkptr_t *lbp)
{
/* L2BLK_GET_PSIZE returns aligned size for log blocks */
uint64_t asize = L2BLK_GET_PSIZE((lbp)->lbp_prop);
uint64_t end = lbp->lbp_daddr + asize - 1;
uint64_t start = lbp->lbp_payload_start;
boolean_t evicted = B_FALSE;
/*
* A log block is valid if all of the following conditions are true:
* - it fits entirely (including its payload) between l2ad_start and
* l2ad_end
* - it has a valid size
* - neither the log block itself nor part of its payload was evicted
* by l2arc_evict():
*
* l2ad_hand l2ad_evict
* | | lbp_daddr
* | start | | end
* | | | | |
* V V V V V
* l2ad_start ============================================ l2ad_end
* --------------------------||||
* ^ ^
* | log block
* payload
*/
evicted =
l2arc_range_check_overlap(start, end, dev->l2ad_hand) ||
l2arc_range_check_overlap(start, end, dev->l2ad_evict) ||
l2arc_range_check_overlap(dev->l2ad_hand, dev->l2ad_evict, start) ||
l2arc_range_check_overlap(dev->l2ad_hand, dev->l2ad_evict, end);
return (start >= dev->l2ad_start && end <= dev->l2ad_end &&
asize > 0 && asize <= sizeof (l2arc_log_blk_phys_t) &&
(!evicted || dev->l2ad_first));
}
/*
* Inserts ARC buffer header `hdr' into the current L2ARC log block on
* the device. The buffer being inserted must be present in L2ARC.
* Returns B_TRUE if the L2ARC log block is full and needs to be committed
* to L2ARC, or B_FALSE if it still has room for more ARC buffers.
*/
static boolean_t
l2arc_log_blk_insert(l2arc_dev_t *dev, const arc_buf_hdr_t *hdr)
{
l2arc_log_blk_phys_t *lb = &dev->l2ad_log_blk;
l2arc_log_ent_phys_t *le;
if (dev->l2ad_log_entries == 0)
return (B_FALSE);
int index = dev->l2ad_log_ent_idx++;
ASSERT3S(index, <, dev->l2ad_log_entries);
ASSERT(HDR_HAS_L2HDR(hdr));
le = &lb->lb_entries[index];
memset(le, 0, sizeof (*le));
le->le_dva = hdr->b_dva;
le->le_birth = hdr->b_birth;
le->le_daddr = hdr->b_l2hdr.b_daddr;
if (index == 0)
dev->l2ad_log_blk_payload_start = le->le_daddr;
L2BLK_SET_LSIZE((le)->le_prop, HDR_GET_LSIZE(hdr));
L2BLK_SET_PSIZE((le)->le_prop, HDR_GET_PSIZE(hdr));
L2BLK_SET_COMPRESS((le)->le_prop, HDR_GET_COMPRESS(hdr));
le->le_complevel = hdr->b_complevel;
L2BLK_SET_TYPE((le)->le_prop, hdr->b_type);
L2BLK_SET_PROTECTED((le)->le_prop, !!(HDR_PROTECTED(hdr)));
L2BLK_SET_PREFETCH((le)->le_prop, !!(HDR_PREFETCH(hdr)));
L2BLK_SET_STATE((le)->le_prop, hdr->b_l1hdr.b_state->arcs_state);
dev->l2ad_log_blk_payload_asize += vdev_psize_to_asize(dev->l2ad_vdev,
HDR_GET_PSIZE(hdr));
return (dev->l2ad_log_ent_idx == dev->l2ad_log_entries);
}
/*
* Checks whether a given L2ARC device address sits in a time-sequential
* range. The trick here is that the L2ARC is a rotary buffer, so we can't
* just do a range comparison, we need to handle the situation in which the
* range wraps around the end of the L2ARC device. Arguments:
* bottom -- Lower end of the range to check (written to earlier).
* top -- Upper end of the range to check (written to later).
* check -- The address for which we want to determine if it sits in
* between the top and bottom.
*
* The 3-way conditional below represents the following cases:
*
* bottom < top : Sequentially ordered case:
* <check>--------+-------------------+
* | (overlap here?) |
* L2ARC dev V V
* |---------------<bottom>============<top>--------------|
*
* bottom > top: Looped-around case:
* <check>--------+------------------+
* | (overlap here?) |
* L2ARC dev V V
* |===============<top>---------------<bottom>===========|
* ^ ^
* | (or here?) |
* +---------------+---------<check>
*
* top == bottom : Just a single address comparison.
*/
boolean_t
l2arc_range_check_overlap(uint64_t bottom, uint64_t top, uint64_t check)
{
if (bottom < top)
return (bottom <= check && check <= top);
else if (bottom > top)
return (check <= top || bottom <= check);
else
return (check == top);
}
EXPORT_SYMBOL(arc_buf_size);
EXPORT_SYMBOL(arc_write);
EXPORT_SYMBOL(arc_read);
EXPORT_SYMBOL(arc_buf_info);
EXPORT_SYMBOL(arc_getbuf_func);
EXPORT_SYMBOL(arc_add_prune_callback);
EXPORT_SYMBOL(arc_remove_prune_callback);
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, min, param_set_arc_min,
spl_param_get_u64, ZMOD_RW, "Minimum ARC size in bytes");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, max, param_set_arc_max,
spl_param_get_u64, ZMOD_RW, "Maximum ARC size in bytes");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, meta_balance, UINT, ZMOD_RW,
"Balance between metadata and data on ghost hits.");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, grow_retry, param_set_arc_int,
param_get_uint, ZMOD_RW, "Seconds before growing ARC size");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, shrink_shift, param_set_arc_int,
param_get_uint, ZMOD_RW, "log2(fraction of ARC to reclaim)");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, pc_percent, UINT, ZMOD_RW,
"Percent of pagecache to reclaim ARC to");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, average_blocksize, UINT, ZMOD_RD,
"Target average block size");
ZFS_MODULE_PARAM(zfs, zfs_, compressed_arc_enabled, INT, ZMOD_RW,
"Disable compressed ARC buffers");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, min_prefetch_ms, param_set_arc_int,
param_get_uint, ZMOD_RW, "Min life of prefetch block in ms");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, min_prescient_prefetch_ms,
param_set_arc_int, param_get_uint, ZMOD_RW,
"Min life of prescient prefetched block in ms");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, write_max, U64, ZMOD_RW,
"Max write bytes per interval");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, write_boost, U64, ZMOD_RW,
"Extra write bytes during device warmup");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, headroom, U64, ZMOD_RW,
"Number of max device writes to precache");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, headroom_boost, U64, ZMOD_RW,
"Compressed l2arc_headroom multiplier");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, trim_ahead, U64, ZMOD_RW,
"TRIM ahead L2ARC write size multiplier");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, feed_secs, U64, ZMOD_RW,
"Seconds between L2ARC writing");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, feed_min_ms, U64, ZMOD_RW,
"Min feed interval in milliseconds");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, noprefetch, INT, ZMOD_RW,
"Skip caching prefetched buffers");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, feed_again, INT, ZMOD_RW,
"Turbo L2ARC warmup");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, norw, INT, ZMOD_RW,
"No reads during writes");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, meta_percent, UINT, ZMOD_RW,
"Percent of ARC size allowed for L2ARC-only headers");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, rebuild_enabled, INT, ZMOD_RW,
"Rebuild the L2ARC when importing a pool");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, rebuild_blocks_min_l2size, U64, ZMOD_RW,
"Min size in bytes to write rebuild log blocks in L2ARC");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, mfuonly, INT, ZMOD_RW,
"Cache only MFU data from ARC into L2ARC");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, exclude_special, INT, ZMOD_RW,
"Exclude dbufs on special vdevs from being cached to L2ARC if set.");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, lotsfree_percent, param_set_arc_int,
param_get_uint, ZMOD_RW, "System free memory I/O throttle in bytes");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, sys_free, param_set_arc_u64,
spl_param_get_u64, ZMOD_RW, "System free memory target size in bytes");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, dnode_limit, param_set_arc_u64,
spl_param_get_u64, ZMOD_RW, "Minimum bytes of dnodes in ARC");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, dnode_limit_percent,
param_set_arc_int, param_get_uint, ZMOD_RW,
"Percent of ARC meta buffers for dnodes");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, dnode_reduce_percent, UINT, ZMOD_RW,
"Percentage of excess dnodes to try to unpin");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, eviction_pct, UINT, ZMOD_RW,
"When full, ARC allocation waits for eviction of this % of alloc size");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, evict_batch_limit, UINT, ZMOD_RW,
"The number of headers to evict per sublist before moving to the next");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, prune_task_threads, INT, ZMOD_RW,
"Number of arc_prune threads");
diff --git a/sys/contrib/openzfs/module/zfs/dsl_scan.c b/sys/contrib/openzfs/module/zfs/dsl_scan.c
index 34012db82dee..e16128fdff87 100644
--- a/sys/contrib/openzfs/module/zfs/dsl_scan.c
+++ b/sys/contrib/openzfs/module/zfs/dsl_scan.c
@@ -1,5271 +1,5270 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2021 by Delphix. All rights reserved.
* Copyright 2016 Gary Mills
* Copyright (c) 2017, 2019, Datto Inc. All rights reserved.
* Copyright (c) 2015, Nexenta Systems, Inc. All rights reserved.
* Copyright 2019 Joyent, Inc.
*/
#include <sys/dsl_scan.h>
#include <sys/dsl_pool.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_prop.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_synctask.h>
#include <sys/dnode.h>
#include <sys/dmu_tx.h>
#include <sys/dmu_objset.h>
#include <sys/arc.h>
#include <sys/arc_impl.h>
#include <sys/zap.h>
#include <sys/zio.h>
#include <sys/zfs_context.h>
#include <sys/fs/zfs.h>
#include <sys/zfs_znode.h>
#include <sys/spa_impl.h>
#include <sys/vdev_impl.h>
#include <sys/zil_impl.h>
#include <sys/zio_checksum.h>
#include <sys/brt.h>
#include <sys/ddt.h>
#include <sys/sa.h>
#include <sys/sa_impl.h>
#include <sys/zfeature.h>
#include <sys/abd.h>
#include <sys/range_tree.h>
#include <sys/dbuf.h>
#ifdef _KERNEL
#include <sys/zfs_vfsops.h>
#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 void scan_ds_prefetch_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);
static uint64_t dsl_scan_count_data_disks(spa_t *spa);
static void read_by_block_level(dsl_scan_t *scn, zbookmark_phys_t zb);
extern uint_t zfs_vdev_async_write_active_min_dirty_percent;
static int zfs_scan_blkstats = 0;
/*
* 'zpool status' uses bytes processed per pass to report throughput and
* estimate time remaining. We define a pass to start when the scanning
* phase completes for a sequential resilver. Optionally, this value
* may be used to reset the pass statistics every N txgs to provide an
* estimated completion time based on currently observed performance.
*/
static uint_t zfs_scan_report_txgs = 0;
/*
* 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.
*/
static int zfs_scan_strict_mem_lim = B_FALSE;
/*
* 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.
*/
static uint64_t zfs_scan_vdev_limit = 16 << 20;
static uint_t zfs_scan_issue_strategy = 0;
/* don't queue & sort zios, go direct */
static int zfs_scan_legacy = B_FALSE;
static uint64_t zfs_scan_max_ext_gap = 2 << 20; /* in bytes */
/*
* 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.
*/
static uint_t zfs_scan_fill_weight = 3;
static uint64_t fill_weight;
/* See dsl_scan_should_clear() for details on the memory limit tunables */
static const uint64_t zfs_scan_mem_lim_min = 16 << 20; /* bytes */
static const uint64_t zfs_scan_mem_lim_soft_max = 128 << 20; /* bytes */
/* fraction of physmem */
static uint_t zfs_scan_mem_lim_fact = 20;
/* fraction of mem lim above */
static uint_t zfs_scan_mem_lim_soft_fact = 20;
/* minimum milliseconds to scrub per txg */
static uint_t zfs_scrub_min_time_ms = 1000;
/* minimum milliseconds to obsolete per txg */
static uint_t zfs_obsolete_min_time_ms = 500;
/* minimum milliseconds to free per txg */
static uint_t zfs_free_min_time_ms = 1000;
/* minimum milliseconds to resilver per txg */
static uint_t zfs_resilver_min_time_ms = 3000;
static uint_t zfs_scan_checkpoint_intval = 7200; /* in seconds */
int zfs_scan_suspend_progress = 0; /* set to prevent scans from progressing */
static int zfs_no_scrub_io = B_FALSE; /* set to disable scrub i/o */
static int zfs_no_scrub_prefetch = B_FALSE; /* set to disable scrub prefetch */
static const enum ddt_class zfs_scrub_ddt_class_max = DDT_CLASS_DUPLICATE;
/* max number of blocks to free in a single TXG */
static uint64_t zfs_async_block_max_blocks = UINT64_MAX;
/* max number of dedup blocks to free in a single TXG */
static uint64_t zfs_max_async_dedup_frees = 100000;
/* set to disable resilver deferring */
static int zfs_resilver_disable_defer = B_FALSE;
/*
* 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)
/*
* Enable/disable the processing of the free_bpobj object.
*/
static int zfs_free_bpobj_enabled = 1;
/* Error blocks to be scrubbed in one txg. */
static uint_t zfs_scrub_error_blocks_per_txg = 1 << 12;
/* 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_queues_pending == 0
* SYNC_MANDATORY) write out scn_phys always. scn_queues_pending must be 0.
* SYNC_CACHED) if scn_queues_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[];
} 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 */
zio_t *q_zio; /* scn_zio_root child for waiting on IO */
/* trees used for sorting I/Os and extents of I/Os */
range_tree_t *q_exts_by_addr;
zfs_btree_t q_exts_by_size;
avl_tree_t q_sios_by_addr;
uint64_t q_sio_memused;
uint64_t q_last_ext_addr;
/* 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 mechanism for passing additional context to the
* compare functions. Thus we store this value globally and
* we only allow it to be set at module initialization time
*/
fill_weight = zfs_scan_fill_weight;
for (int i = 0; i < SPA_DVAS_PER_BP; i++) {
char name[36];
(void) snprintf(name, sizeof (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)
{
memset(bp, 0, 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);
memcpy(bp->blk_dva, sio->sio_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);
/*
* Calculate the max number of in-flight bytes for pool-wide
* scanning operations (minimum 1MB, maximum 1/4 of arc_c_max).
* Limits for the issuing phase are done per top-level vdev and
* are handled separately.
*/
scn->scn_maxinflight_bytes = MIN(arc_c_max / 4, MAX(1ULL << 20,
zfs_scan_vdev_limit * dsl_scan_count_data_disks(spa)));
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 for %s; "
"restarting new-style scrub in txg %llu",
spa->spa_name,
(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_ERRORSCRUB, sizeof (uint64_t),
ERRORSCRUB_PHYS_NUMINTS, &scn->errorscrub_phys);
if (err != 0 && err != ENOENT)
return (err);
err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS,
&scn->scn_phys);
/*
* Detect if the pool contains the signature of #2094. If it
* does properly update the scn->scn_phys structure and notify
* the administrator by setting an errata for the pool.
*/
if (err == EOVERFLOW) {
uint64_t zaptmp[SCAN_PHYS_NUMINTS + 1];
VERIFY3S(SCAN_PHYS_NUMINTS, ==, 24);
VERIFY3S(offsetof(dsl_scan_phys_t, scn_flags), ==,
(23 * sizeof (uint64_t)));
err = zap_lookup(dp->dp_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN,
sizeof (uint64_t), SCAN_PHYS_NUMINTS + 1, &zaptmp);
if (err == 0) {
uint64_t overflow = zaptmp[SCAN_PHYS_NUMINTS];
if (overflow & ~DSL_SCAN_FLAGS_MASK ||
scn->scn_async_destroying) {
spa->spa_errata =
ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY;
return (EOVERFLOW);
}
memcpy(&scn->scn_phys, zaptmp,
SCAN_PHYS_NUMINTS * sizeof (uint64_t));
scn->scn_phys.scn_flags = overflow;
/* Required scrub already in progress. */
if (scn->scn_phys.scn_state == DSS_FINISHED ||
scn->scn_phys.scn_state == DSS_CANCELED)
spa->spa_errata =
ZPOOL_ERRATA_ZOL_2094_SCRUB;
}
}
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 -
scn->scn_phys.scn_skipped;
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 for %s was modified "
"by old software; restarting in txg %llu",
spa->spa_name,
(longlong_t)scn->scn_restart_txg);
} else if (dsl_scan_resilvering(dp)) {
/*
* If a resilver is in progress and there are already
* errors, restart it instead of finishing this scan and
* then restarting it. If there haven't been any errors
* then remember that the incore DTL is valid.
*/
if (scn->scn_phys.scn_errors > 0) {
scn->scn_restart_txg = txg;
zfs_dbgmsg("resilver can't excise DTL_MISSING "
"when finished; restarting on %s in txg "
"%llu",
spa->spa_name,
(u_longlong_t)scn->scn_restart_txg);
} else {
/* it's safe to excise DTL when finished */
spa->spa_scrub_started = B_TRUE;
}
}
}
memcpy(&scn->scn_phys_cached, &scn->scn_phys, 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);
vdev_scan_stat_init(spa->spa_root_vdev);
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);
scan_ds_prefetch_queue_clear(scn);
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_resilver_scheduled(dsl_pool_t *dp)
{
return ((dp->dp_scan && dp->dp_scan->scn_restart_txg != 0) ||
(spa_async_tasks(dp->dp_spa) & SPA_ASYNC_RESILVER));
}
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_errorscrubbing(const dsl_pool_t *dp)
{
dsl_errorscrub_phys_t *errorscrub_phys = &dp->dp_scan->errorscrub_phys;
return (errorscrub_phys->dep_state == DSS_ERRORSCRUBBING &&
errorscrub_phys->dep_func == POOL_SCAN_ERRORSCRUB);
}
boolean_t
dsl_errorscrub_is_paused(const dsl_scan_t *scn)
{
return (dsl_errorscrubbing(scn->scn_dp) &&
scn->errorscrub_phys.dep_paused_flags);
}
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);
}
static void
dsl_errorscrub_sync_state(dsl_scan_t *scn, dmu_tx_t *tx)
{
scn->errorscrub_phys.dep_cursor =
zap_cursor_serialize(&scn->errorscrub_cursor);
VERIFY0(zap_update(scn->scn_dp->dp_meta_objset,
DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ERRORSCRUB, sizeof (uint64_t), ERRORSCRUB_PHYS_NUMINTS,
&scn->errorscrub_phys, tx));
}
static void
dsl_errorscrub_setup_sync(void *arg, dmu_tx_t *tx)
{
dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan;
pool_scan_func_t *funcp = arg;
dsl_pool_t *dp = scn->scn_dp;
spa_t *spa = dp->dp_spa;
ASSERT(!dsl_scan_is_running(scn));
ASSERT(!dsl_errorscrubbing(scn->scn_dp));
ASSERT(*funcp > POOL_SCAN_NONE && *funcp < POOL_SCAN_FUNCS);
memset(&scn->errorscrub_phys, 0, sizeof (scn->errorscrub_phys));
scn->errorscrub_phys.dep_func = *funcp;
scn->errorscrub_phys.dep_state = DSS_ERRORSCRUBBING;
scn->errorscrub_phys.dep_start_time = gethrestime_sec();
scn->errorscrub_phys.dep_to_examine = spa_get_last_errlog_size(spa);
scn->errorscrub_phys.dep_examined = 0;
scn->errorscrub_phys.dep_errors = 0;
scn->errorscrub_phys.dep_cursor = 0;
zap_cursor_init_serialized(&scn->errorscrub_cursor,
spa->spa_meta_objset, spa->spa_errlog_last,
scn->errorscrub_phys.dep_cursor);
vdev_config_dirty(spa->spa_root_vdev);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_ERRORSCRUB_START);
dsl_errorscrub_sync_state(scn, tx);
spa_history_log_internal(spa, "error scrub setup", tx,
"func=%u mintxg=%u maxtxg=%llu",
*funcp, 0, (u_longlong_t)tx->tx_txg);
}
static int
dsl_errorscrub_setup_check(void *arg, dmu_tx_t *tx)
{
(void) arg;
dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan;
if (dsl_scan_is_running(scn) || (dsl_errorscrubbing(scn->scn_dp))) {
return (SET_ERROR(EBUSY));
}
if (spa_get_last_errlog_size(scn->scn_dp->dp_spa) == 0) {
return (ECANCELED);
}
return (0);
}
/*
* 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_queues_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_queues_pending == 0);
if (scn->scn_queues_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(zfs_btree_first(&q->q_exts_by_size, NULL), ==,
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));
memcpy(&scn->scn_phys_cached, &scn->scn_phys,
sizeof (scn->scn_phys));
if (scn->scn_checkpointing)
zfs_dbgmsg("finish scan checkpoint for %s",
spa->spa_name);
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));
}
}
int
dsl_scan_setup_check(void *arg, dmu_tx_t *tx)
{
(void) arg;
dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan;
vdev_t *rvd = scn->scn_dp->dp_spa->spa_root_vdev;
if (dsl_scan_is_running(scn) || vdev_rebuild_active(rvd) ||
dsl_errorscrubbing(scn->scn_dp))
return (SET_ERROR(EBUSY));
return (0);
}
void
dsl_scan_setup_sync(void *arg, dmu_tx_t *tx)
{
(void) arg;
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);
memset(&scn->scn_phys, 0, sizeof (scn->scn_phys));
/*
* If we are starting a fresh scrub, we erase the error scrub
* information from disk.
*/
memset(&scn->errorscrub_phys, 0, sizeof (scn->errorscrub_phys));
dsl_errorscrub_sync_state(scn, tx);
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);
vdev_scan_stat_init(spa->spa_root_vdev);
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)) {
nvlist_t *aux = fnvlist_alloc();
fnvlist_add_string(aux, ZFS_EV_RESILVER_TYPE,
"healing");
spa_event_notify(spa, NULL, aux,
ESC_ZFS_RESILVER_START);
nvlist_free(aux);
} 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;
/*
* When starting a resilver clear any existing rebuild state.
* This is required to prevent stale rebuild status from
* being reported when a rebuild is run, then a resilver and
* finally a scrub. In which case only the scrub status
* should be reported by 'zpool status'.
*/
if (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) {
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t i = 0; i < rvd->vdev_children; i++) {
vdev_t *vd = rvd->vdev_child[i];
vdev_rebuild_clear_sync(
(void *)(uintptr_t)vd->vdev_id, tx);
}
}
}
/* back to the generic stuff */
if (zfs_scan_blkstats) {
if (dp->dp_blkstats == NULL) {
dp->dp_blkstats =
vmem_alloc(sizeof (zfs_all_blkstats_t), KM_SLEEP);
}
memset(&dp->dp_blkstats->zab_type, 0,
sizeof (dp->dp_blkstats->zab_type));
} else {
if (dp->dp_blkstats) {
vmem_free(dp->dp_blkstats, sizeof (zfs_all_blkstats_t));
dp->dp_blkstats = NULL;
}
}
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);
memcpy(&scn->scn_phys_cached, &scn->scn_phys, 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, (u_longlong_t)scn->scn_phys.scn_min_txg,
(u_longlong_t)scn->scn_phys.scn_max_txg);
}
/*
* Called by ZFS_IOC_POOL_SCRUB and ZFS_IOC_POOL_SCAN ioctl to start a scrub,
* error scrub or resilver. Can also be called to resume a paused scrub or
* error 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_scan_restart_resilver(spa->spa_dsl_pool, 0);
return (0);
}
if (func == POOL_SCAN_ERRORSCRUB) {
if (dsl_errorscrub_is_paused(dp->dp_scan)) {
/*
* got error scrub start cmd, resume paused error 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_ERRORSCRUB_RESUME);
return (ECANCELED);
}
return (SET_ERROR(err));
}
return (dsl_sync_task(spa_name(dp->dp_spa),
dsl_errorscrub_setup_check, dsl_errorscrub_setup_sync,
&func, 0, ZFS_SPACE_CHECK_RESERVED));
}
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 (SET_ERROR(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));
}
static void
dsl_errorscrub_done(dsl_scan_t *scn, boolean_t complete, dmu_tx_t *tx)
{
dsl_pool_t *dp = scn->scn_dp;
spa_t *spa = dp->dp_spa;
if (complete) {
spa_event_notify(spa, NULL, NULL, ESC_ZFS_ERRORSCRUB_FINISH);
spa_history_log_internal(spa, "error scrub done", tx,
"errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa));
} else {
spa_history_log_internal(spa, "error scrub canceled", tx,
"errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa));
}
scn->errorscrub_phys.dep_state = complete ? DSS_FINISHED : DSS_CANCELED;
spa->spa_scrub_active = B_FALSE;
spa_errlog_rotate(spa);
scn->errorscrub_phys.dep_end_time = gethrestime_sec();
zap_cursor_fini(&scn->errorscrub_cursor);
if (spa->spa_errata == ZPOOL_ERRATA_ZOL_2094_SCRUB)
spa->spa_errata = 0;
ASSERT(!dsl_errorscrubbing(scn->scn_dp));
}
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);
scan_ds_prefetch_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;
spa_notify_waiters(spa);
if (dsl_scan_restarting(scn, tx))
spa_history_log_internal(spa, "scan aborted, restarting", tx,
"errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa));
else if (!complete)
spa_history_log_internal(spa, "scan cancelled", tx,
"errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa));
else
spa_history_log_internal(spa, "scan done", tx,
"errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa));
if (DSL_SCAN_IS_SCRUB_RESILVER(scn)) {
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, B_FALSE);
if (scn->scn_phys.scn_min_txg) {
nvlist_t *aux = fnvlist_alloc();
fnvlist_add_string(aux, ZFS_EV_RESILVER_TYPE,
"healing");
spa_event_notify(spa, NULL, aux,
ESC_ZFS_RESILVER_FINISH);
nvlist_free(aux);
} else {
spa_event_notify(spa, NULL, NULL,
ESC_ZFS_SCRUB_FINISH);
}
} else {
vdev_dtl_reassess(spa->spa_root_vdev, tx->tx_txg,
0, B_TRUE, B_FALSE);
}
spa_errlog_rotate(spa);
/*
* Don't clear flag until after vdev_dtl_reassess to ensure that
* DTL_MISSING will get updated when possible.
*/
spa->spa_scrub_started = B_FALSE;
/*
* 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 resilver_deferred flags in the config.
* If there are drives that need resilvering, kick
* off an asynchronous request to start resilver.
* vdev_clear_resilver_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
* and start the resilver then.
*/
if (spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER) &&
vdev_clear_resilver_deferred(spa->spa_root_vdev, tx)) {
spa_history_log_internal(spa,
"starting deferred resilver", tx, "errors=%llu",
(u_longlong_t)spa_approx_errlog_size(spa));
spa_async_request(spa, SPA_ASYNC_RESILVER);
}
/* Clear recent error events (i.e. duplicate events tracking) */
if (complete)
zfs_ereport_clear(spa, NULL);
}
scn->scn_phys.scn_end_time = gethrestime_sec();
if (spa->spa_errata == ZPOOL_ERRATA_ZOL_2094_SCRUB)
spa->spa_errata = 0;
ASSERT(!dsl_scan_is_running(scn));
}
static int
dsl_errorscrub_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 error scrub when there is no in-progress
* error scrub.
*/
if (!dsl_errorscrubbing(dp))
return (SET_ERROR(ENOENT));
/* can't pause a paused error scrub */
if (dsl_errorscrub_is_paused(scn))
return (SET_ERROR(EBUSY));
} else if (*cmd != POOL_SCRUB_NORMAL) {
return (SET_ERROR(ENOTSUP));
}
return (0);
}
static void
dsl_errorscrub_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) {
spa->spa_scan_pass_errorscrub_pause = gethrestime_sec();
scn->errorscrub_phys.dep_paused_flags = B_TRUE;
dsl_errorscrub_sync_state(scn, tx);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_ERRORSCRUB_PAUSED);
} else {
ASSERT3U(*cmd, ==, POOL_SCRUB_NORMAL);
if (dsl_errorscrub_is_paused(scn)) {
/*
* We need to keep track of how much time we spend
* paused per pass so that we can adjust the error scrub
* rate shown in the output of 'zpool status'.
*/
spa->spa_scan_pass_errorscrub_spent_paused +=
gethrestime_sec() -
spa->spa_scan_pass_errorscrub_pause;
spa->spa_scan_pass_errorscrub_pause = 0;
scn->errorscrub_phys.dep_paused_flags = B_FALSE;
zap_cursor_init_serialized(
&scn->errorscrub_cursor,
spa->spa_meta_objset, spa->spa_errlog_last,
scn->errorscrub_phys.dep_cursor);
dsl_errorscrub_sync_state(scn, tx);
}
}
}
static int
dsl_errorscrub_cancel_check(void *arg, dmu_tx_t *tx)
{
(void) arg;
dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan;
/* can't cancel a error scrub when there is no one in-progress */
if (!dsl_errorscrubbing(scn->scn_dp))
return (SET_ERROR(ENOENT));
return (0);
}
static void
dsl_errorscrub_cancel_sync(void *arg, dmu_tx_t *tx)
{
(void) arg;
dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan;
dsl_errorscrub_done(scn, B_FALSE, tx);
dsl_errorscrub_sync_state(scn, tx);
spa_event_notify(scn->scn_dp->dp_spa, NULL, NULL,
ESC_ZFS_ERRORSCRUB_ABORT);
}
static int
dsl_scan_cancel_check(void *arg, dmu_tx_t *tx)
{
(void) arg;
dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan;
if (!dsl_scan_is_running(scn))
return (SET_ERROR(ENOENT));
return (0);
}
static void
dsl_scan_cancel_sync(void *arg, dmu_tx_t *tx)
{
(void) arg;
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)
{
if (dsl_errorscrubbing(dp)) {
return (dsl_sync_task(spa_name(dp->dp_spa),
dsl_errorscrub_cancel_check, dsl_errorscrub_cancel_sync,
NULL, 3, ZFS_SPACE_CHECK_RESERVED));
}
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);
spa_notify_waiters(spa);
} 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)
{
if (dsl_errorscrubbing(dp)) {
return (dsl_sync_task(spa_name(dp->dp_spa),
dsl_errorscrub_pause_resume_check,
dsl_errorscrub_pause_resume_sync, &cmd, 3,
ZFS_SPACE_CHECK_RESERVED));
}
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_scan_restart_resilver(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 for %s at txg=%llu",
dp->dp_spa->spa_name, (longlong_t)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_queues_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)
{
spa_t *spa = scn->scn_dp->dp_spa;
vdev_t *rvd = scn->scn_dp->dp_spa->spa_root_vdev;
uint64_t alloc, mlim_hard, mlim_soft, mused;
alloc = metaslab_class_get_alloc(spa_normal_class(spa));
alloc += metaslab_class_get_alloc(spa_special_class(spa));
alloc += metaslab_class_get_alloc(spa_dedup_class(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) {
/*
* # of extents in exts_by_addr = # in exts_by_size.
* B-tree efficiency is ~75%, but can be as low as 50%.
*/
mused += zfs_btree_numnodes(&queue->q_exts_by_size) *
((sizeof (range_seg_gap_t) + sizeof (uint64_t)) *
3 / 2) + 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_queues_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 and objset blocks. */
if (zb && (zb->zb_level != 0 && zb->zb_level != ZB_ROOT_LEVEL))
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%), someone is explicitly waiting for this txg
* to complete, or we have used up all of the time in the txg
* timeout (default 5 sec).
* 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
*/
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;
uint64_t dirty_min_bytes = zfs_dirty_data_max *
zfs_vdev_async_write_active_min_dirty_percent / 100;
uint_t 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 &&
(scn->scn_dp->dp_dirty_total >= dirty_min_bytes ||
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 && zb->zb_level == ZB_ROOT_LEVEL) {
dprintf("suspending at first available 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);
SET_BOOKMARK(&scn->scn_phys.scn_bookmark,
zb->zb_objset, 0, 0, 0);
} else if (zb != NULL) {
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 {
#ifdef ZFS_DEBUG
dsl_scan_phys_t *scnp = &scn->scn_phys;
dprintf("suspending at 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);
#endif
}
scn->scn_suspending = B_TRUE;
return (B_TRUE);
}
return (B_FALSE);
}
static boolean_t
dsl_error_scrub_check_suspend(dsl_scan_t *scn, const zbookmark_phys_t *zb)
{
/*
* We suspend if:
* - we have scrubbed for at least the minimum time (default 1 sec
* for error scrub), someone is explicitly waiting for this txg
* to complete, or we have used up all of the time in the txg
* timeout (default 5 sec).
* or
* - the spa is shutting down because this pool is being exported
* or the machine is rebooting.
*/
uint64_t curr_time_ns = gethrtime();
uint64_t error_scrub_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 mintime = zfs_scrub_min_time_ms;
if ((NSEC2MSEC(error_scrub_time_ns) > mintime &&
(txg_sync_waiting(scn->scn_dp) ||
NSEC2SEC(sync_time_ns) >= zfs_txg_timeout)) ||
spa_shutting_down(scn->scn_dp->dp_spa)) {
if (zb) {
dprintf("error scrub 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);
}
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;
static int
dsl_scan_zil_block(zilog_t *zilog, const blkptr_t *bp, void *arg,
uint64_t claim_txg)
{
(void) zilog;
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;
ASSERT(!BP_IS_REDACTED(bp));
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);
}
static int
dsl_scan_zil_record(zilog_t *zilog, const lr_t *lrc, void *arg,
uint64_t claim_txg)
{
(void) zilog;
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;
const lr_write_t *lr = (const lr_write_t *)lrc;
const blkptr_t *bp = &lr->lr_blkptr;
zbookmark_phys_t zb;
ASSERT(!BP_IS_REDACTED(bp));
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);
ASSERT3U(BP_GET_LSIZE(bp), !=, 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 (or, in read-only mode, blocks that *would* be claimed).
*/
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, B_FALSE);
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, const 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, const 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, const void *tag)
{
zfs_refcount_add(&spc->spc_refcnt, tag);
}
static void
scan_ds_prefetch_queue_clear(dsl_scan_t *scn)
{
spa_t *spa = scn->scn_dp->dp_spa;
void *cookie = NULL;
scan_prefetch_issue_ctx_t *spic = NULL;
mutex_enter(&spa->spa_scrub_lock);
while ((spic = avl_destroy_nodes(&scn->scn_prefetch_queue,
&cookie)) != NULL) {
scan_prefetch_ctx_rele(spic->spic_spc, scn);
kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t));
}
mutex_exit(&spa->spa_scrub_lock);
}
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 || BP_IS_REDACTED(bp))
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, DN_SPILL_BLKPTR(dnp), &zb);
}
scan_prefetch_ctx_rele(spc, FTAG);
}
static void
dsl_scan_prefetch_cb(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp,
arc_buf_t *buf, void *private)
{
(void) zio;
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 limiting 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_prefetch_stop)
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;
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)) {
if (OBJSET_BUF_HAS_PROJECTUSED(buf)) {
dsl_scan_prefetch_dnode(scn,
&osp->os_projectused_dnode, zb->zb_objset,
DMU_PROJECTUSED_OBJECT);
}
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);
}
static void
dsl_scan_prefetch_thread(void *arg)
{
dsl_scan_t *scn = arg;
spa_t *spa = scn->scn_dp->dp_spa;
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);
if (BP_IS_PROTECTED(&spic->spic_bp)) {
ASSERT(BP_GET_TYPE(&spic->spic_bp) == DMU_OT_DNODE ||
BP_GET_TYPE(&spic->spic_bp) == DMU_OT_OBJSET);
ASSERT3U(BP_GET_LEVEL(&spic->spic_bp), ==, 0);
zio_flags |= ZIO_FLAG_RAW;
}
/* We don't need data L1 buffer since we do not prefetch L0. */
blkptr_t *bp = &spic->spic_bp;
if (BP_GET_LEVEL(bp) == 1 && BP_GET_TYPE(bp) != DMU_OT_DNODE &&
BP_GET_TYPE(bp) != DMU_OT_OBJSET)
flags |= ARC_FLAG_NO_BUF;
/* issue the prefetch asynchronously */
(void) arc_read(scn->scn_zio_root, spa, 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, zero it out to indicate that it's OK
* to start checking for suspending again.
*/
if (zbookmark_subtree_tbd(dnp, zb,
&scn->scn_phys.scn_bookmark)) {
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);
memset(&scn->scn_phys.scn_bookmark, 0, 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);
inline __attribute__((always_inline)) 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.
*/
inline __attribute__((always_inline)) 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;
spa_t *spa = dp->dp_spa;
int zio_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCAN_THREAD;
int err;
ASSERT(!BP_IS_REDACTED(bp));
/*
* There is an unlikely case of encountering dnodes with contradicting
* dn_bonuslen and DNODE_FLAG_SPILL_BLKPTR flag before in files created
* or modified before commit 4254acb was merged. As it is not possible
* to know which of the two is correct, report an error.
*/
if (dnp != NULL &&
dnp->dn_bonuslen > DN_MAX_BONUS_LEN(dnp)) {
scn->scn_phys.scn_errors++;
spa_log_error(spa, zb, &bp->blk_birth);
return (SET_ERROR(EINVAL));
}
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, 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;
if (BP_IS_PROTECTED(bp)) {
ASSERT3U(BP_GET_COMPRESS(bp), ==, ZIO_COMPRESS_OFF);
zio_flags |= ZIO_FLAG_RAW;
}
err = arc_read(NULL, 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, 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/project accounting
* objects, and never skip them, even if we are
* suspending. This is necessary so that the
* space deltas from this txg get integrated.
*/
if (OBJSET_BUF_HAS_PROJECTUSED(buf))
dsl_scan_visitdnode(scn, ds, osp->os_type,
&osp->os_projectused_dnode,
DMU_PROJECTUSED_OBJECT, tx);
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);
} else if (!zfs_blkptr_verify(spa, bp,
BLK_CONFIG_NEEDED, BLK_VERIFY_LOG)) {
/*
* Sanity check the block pointer contents, this is handled
* by arc_read() for the cases above.
*/
scn->scn_phys.scn_errors++;
spa_log_error(spa, zb, &bp->blk_birth);
return (SET_ERROR(EINVAL));
}
return (0);
}
inline __attribute__((always_inline)) 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++;
if (BP_IS_HOLE(bp)) {
scn->scn_holes_this_txg++;
return;
}
if (BP_IS_REDACTED(bp)) {
ASSERT(dsl_dataset_feature_is_active(ds,
SPA_FEATURE_REDACTED_DATASETS));
return;
}
/*
* Check if this block contradicts any filesystem flags.
*/
spa_feature_t f = SPA_FEATURE_LARGE_BLOCKS;
if (BP_GET_LSIZE(bp) > SPA_OLD_MAXBLOCKSIZE)
ASSERT(dsl_dataset_feature_is_active(ds, f));
f = zio_checksum_to_feature(BP_GET_CHECKSUM(bp));
if (f != SPA_FEATURE_NONE)
ASSERT(dsl_dataset_feature_is_active(ds, f));
f = zio_compress_to_feature(BP_GET_COMPRESS(bp));
if (f != SPA_FEATURE_NONE)
ASSERT(dsl_dataset_feature_is_active(ds, f));
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 on %s; currently "
"traversing; reset zb_objset to %llu",
(u_longlong_t)ds->ds_object,
ds->ds_dir->dd_pool->dp_spa->spa_name,
(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 on %s; currently "
"traversing; reset bookmark to -1,0,0,0",
(u_longlong_t)ds->ds_object,
ds->ds_dir->dd_pool->dp_spa->spa_name);
}
}
}
/*
* 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 on %s; in queue; "
"replacing with %llu",
(u_longlong_t)ds->ds_object,
dp->dp_spa->spa_name,
(u_longlong_t)dsl_dataset_phys(ds)->
ds_next_snap_obj);
} else {
zfs_dbgmsg("destroying ds %llu on %s; in queue; "
"removing",
(u_longlong_t)ds->ds_object,
dp->dp_spa->spa_name);
}
}
/*
* 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 on %s; currently traversing; "
"reset zb_objset to %llu",
(u_longlong_t)ds->ds_object,
ds->ds_dir->dd_pool->dp_spa->spa_name,
(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 on %s; in queue; "
"replacing with %llu",
(u_longlong_t)ds->ds_object,
dp->dp_spa->spa_name,
(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 on %s; currently traversing; "
"reset zb_objset to %llu",
(u_longlong_t)ds1->ds_object,
ds1->ds_dir->dd_pool->dp_spa->spa_name,
(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 on %s; currently traversing; "
"reset zb_objset to %llu",
(u_longlong_t)ds2->ds_object,
ds2->ds_dir->dd_pool->dp_spa->spa_name,
(u_longlong_t)ds1->ds_object);
}
}
/*
* Called when an origin dataset and its clone are swapped. If we were
* currently traversing the dataset, we need to switch to traversing the
* newly promoted clone.
*/
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 mintxg1, mintxg2;
boolean_t ds1_queued, ds2_queued;
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);
/*
* Handle the in-memory scan queue.
*/
ds1_queued = scan_ds_queue_contains(scn, ds1->ds_object, &mintxg1);
ds2_queued = scan_ds_queue_contains(scn, ds2->ds_object, &mintxg2);
/* Sanity checking. */
if (ds1_queued) {
ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg);
ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg);
}
if (ds2_queued) {
ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg);
ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg);
}
if (ds1_queued && ds2_queued) {
/*
* If both are queued, we don't need to do anything.
* The swapping code below would not handle this case correctly,
* since we can't insert ds2 if it is already there. That's
* because scan_ds_queue_insert() prohibits a duplicate insert
* and panics.
*/
} else if (ds1_queued) {
scan_ds_queue_remove(scn, ds1->ds_object);
scan_ds_queue_insert(scn, ds2->ds_object, mintxg1);
} else if (ds2_queued) {
scan_ds_queue_remove(scn, ds2->ds_object);
scan_ds_queue_insert(scn, ds1->ds_object, mintxg2);
}
/*
* Handle the on-disk scan queue.
* The on-disk state is an out-of-date version of the in-memory state,
* so the in-memory and on-disk values for ds1_queued and ds2_queued may
* be different. Therefore we need to apply the swap logic to the
* on-disk state independently of the in-memory state.
*/
ds1_queued = zap_lookup_int_key(dp->dp_meta_objset,
scn->scn_phys.scn_queue_obj, ds1->ds_object, &mintxg1) == 0;
ds2_queued = zap_lookup_int_key(dp->dp_meta_objset,
scn->scn_phys.scn_queue_obj, ds2->ds_object, &mintxg2) == 0;
/* Sanity checking. */
if (ds1_queued) {
ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg);
ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg);
}
if (ds2_queued) {
ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg);
ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg);
}
if (ds1_queued && ds2_queued) {
/*
* If both are queued, we don't need to do anything.
* Alternatively, we could check for EEXIST from
* zap_add_int_key() and back out to the original state, but
* that would be more work than checking for this case upfront.
*/
} else if (ds1_queued) {
VERIFY3S(0, ==, zap_remove_int(dp->dp_meta_objset,
scn->scn_phys.scn_queue_obj, ds1->ds_object, tx));
VERIFY3S(0, ==, zap_add_int_key(dp->dp_meta_objset,
scn->scn_phys.scn_queue_obj, ds2->ds_object, mintxg1, tx));
zfs_dbgmsg("clone_swap ds %llu on %s; in queue; "
"replacing with %llu",
(u_longlong_t)ds1->ds_object,
dp->dp_spa->spa_name,
(u_longlong_t)ds2->ds_object);
} else if (ds2_queued) {
VERIFY3S(0, ==, zap_remove_int(dp->dp_meta_objset,
scn->scn_phys.scn_queue_obj, ds2->ds_object, tx));
VERIFY3S(0, ==, zap_add_int_key(dp->dp_meta_objset,
scn->scn_phys.scn_queue_obj, ds1->ds_object, mintxg2, tx));
zfs_dbgmsg("clone_swap ds %llu on %s; in queue; "
"replacing with %llu",
(u_longlong_t)ds2->ds_object,
dp->dp_spa->spa_name,
(u_longlong_t)ds1->ds_object);
}
dsl_scan_sync_state(scn, tx, SYNC_CACHED);
}
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(ZFS_MAX_DATASET_NAME_LEN, 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_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 on %s; visiting again",
dp->dp_spa->spa_name);
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 descendant 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);
}
static int
enqueue_cb(dsl_pool_t *dp, dsl_dataset_t *hds, void *arg)
{
(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);
}
void
dsl_scan_ddt_entry(dsl_scan_t *scn, enum zio_checksum checksum,
ddt_entry_t *dde, dmu_tx_t *tx)
{
(void) tx;
const ddt_key_t *ddk = &dde->dde_key;
ddt_phys_t *ddp = dde->dde_phys;
blkptr_t bp;
zbookmark_phys_t zb = { 0 };
if (!dsl_scan_is_running(scn))
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 (int 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 on %s with class_max = %u; "
"suspending=%u", (longlong_t)n, scn->scn_dp->dp_spa->spa_name,
(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.
*/
memset(&scn->scn_phys.scn_bookmark, 0, 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;
/* 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_data_disks(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
uint64_t i, leaves = 0;
for (i = 0; i < rvd->vdev_children; i++) {
vdev_t *vd = rvd->vdev_child[i];
if (vd->vdev_islog || vd->vdev_isspare || vd->vdev_isl2cache)
continue;
leaves += vdev_get_ndisks(vd) - vdev_get_nparity(vd);
}
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;
uint64_t dirty_min_bytes = zfs_dirty_data_max *
zfs_vdev_async_write_active_min_dirty_percent / 100;
uint_t 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 &&
(scn->scn_dp->dp_dirty_total >= dirty_min_bytes ||
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 I/Os 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 sios 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;
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);
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);
}
return (suspended);
}
/*
* This function removes sios from an IO queue which reside within a given
* range_seg_t and inserts them (in offset order) into a list. Note that
* we only ever return a maximum of 32 sios at once. If there are more sios
* to process within this segment that did not make it onto the list we
* return B_TRUE and otherwise B_FALSE.
*/
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_get_start(rs, queue->q_exts_by_addr));
/*
* 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_get_end(rs,
queue->q_exts_by_addr) && num_sios <= 32) {
ASSERT3U(SIO_GET_OFFSET(sio), >=, rs_get_start(rs,
queue->q_exts_by_addr));
ASSERT3U(SIO_GET_END_OFFSET(sio), <=, rs_get_end(rs,
queue->q_exts_by_addr));
next_sio = AVL_NEXT(&queue->q_sios_by_addr, sio);
avl_remove(&queue->q_sios_by_addr, sio);
if (avl_is_empty(&queue->q_sios_by_addr))
atomic_add_64(&queue->q_scn->scn_queues_pending, -1);
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_get_end(rs,
queue->q_exts_by_addr)) {
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_get_end(rs,
queue->q_exts_by_addr) - SIO_GET_OFFSET(sio));
queue->q_last_ext_addr = SIO_GET_OFFSET(sio);
return (B_TRUE);
} else {
uint64_t rstart = rs_get_start(rs, queue->q_exts_by_addr);
uint64_t rend = rs_get_end(rs, queue->q_exts_by_addr);
range_tree_remove(queue->q_exts_by_addr, rstart, rend - rstart);
queue->q_last_ext_addr = -1;
return (B_FALSE);
}
}
/*
* This is called from the queue emptying thread and selects the next
* extent from which we are to issue I/Os. 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 range_seg_t *
scan_io_queue_fetch_ext(dsl_scan_io_queue_t *queue)
{
dsl_scan_t *scn = queue->q_scn;
range_tree_t *rt = queue->q_exts_by_addr;
ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock));
ASSERT(scn->scn_is_sorted);
if (!scn->scn_checkpointing && !scn->scn_clearing)
return (NULL);
/*
* 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 ((zfs_scan_issue_strategy < 1 && scn->scn_checkpointing) ||
zfs_scan_issue_strategy == 1)
return (range_tree_first(rt));
/*
* Try to continue previous extent if it is not completed yet. After
* shrink in scan_io_queue_gather() it may no longer be the best, but
* otherwise we leave shorter remnant every txg.
*/
uint64_t start;
uint64_t size = 1ULL << rt->rt_shift;
range_seg_t *addr_rs;
if (queue->q_last_ext_addr != -1) {
start = queue->q_last_ext_addr;
addr_rs = range_tree_find(rt, start, size);
if (addr_rs != NULL)
return (addr_rs);
}
/*
* Nothing to continue, so find new best extent.
*/
uint64_t *v = zfs_btree_first(&queue->q_exts_by_size, NULL);
if (v == NULL)
return (NULL);
queue->q_last_ext_addr = start = *v << rt->rt_shift;
/*
* We need to get the original entry in the by_addr tree so we can
* modify it.
*/
addr_rs = range_tree_find(rt, start, size);
ASSERT3P(addr_rs, !=, NULL);
ASSERT3U(rs_get_start(addr_rs, rt), ==, start);
ASSERT3U(rs_get_end(addr_rs, rt), >, start);
return (addr_rs);
}
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;
scan_io_t *sio;
zio_t *zio;
list_t sio_list;
ASSERT(queue->q_scn->scn_is_sorted);
list_create(&sio_list, sizeof (scan_io_t),
offsetof(scan_io_t, sio_nodes.sio_list_node));
zio = zio_null(queue->q_scn->scn_zio_root, queue->q_scn->scn_dp->dp_spa,
NULL, NULL, NULL, ZIO_FLAG_CANFAIL);
mutex_enter(q_lock);
queue->q_zio = zio;
/* Calculate maximum in-flight bytes for this vdev. */
queue->q_maxinflight_bytes = MAX(1, zfs_scan_vdev_limit *
(vdev_get_ndisks(queue->q_vd) - vdev_get_nparity(queue->q_vd)));
/* 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 run out of time or sios */
while ((rs = scan_io_queue_fetch_ext(queue)) != NULL) {
uint64_t seg_start = 0, seg_end = 0;
boolean_t more_left;
ASSERT(list_is_empty(&sio_list));
/* loop while we still have sios left to process in this rs */
do {
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;
} while (more_left);
/* 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_remove_head(&sio_list)) != NULL)
scan_io_queue_insert_impl(queue, sio);
queue->q_zio = NULL;
mutex_exit(q_lock);
zio_nowait(zio);
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 I/Os 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_queues_pending == 0)
return;
if (scn->scn_taskq == NULL) {
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.
*/
scn->scn_taskq = taskq_create("dsl_scan_iss", nthreads,
minclsyspri, nthreads, nthreads, TASKQ_PREPOPULATE);
}
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 their 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 (zfs_async_block_max_blocks != 0 &&
scn->scn_visited_this_txg >= zfs_async_block_max_blocks) {
return (B_TRUE);
}
if (zfs_max_async_dedup_frees != 0 &&
scn->scn_dedup_frees_this_txg >= zfs_max_async_dedup_frees) {
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++;
if (BP_GET_DEDUP(bp))
scn->scn_dedup_frees_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
bpobj_dsl_scan_free_block_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
ASSERT(!bp_freed);
return (dsl_scan_free_block_cb(arg, bp, tx));
}
static int
dsl_scan_obsolete_block_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
ASSERT(!bp_freed);
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;
boolean_t clones_left;
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);
}
clones_left = spa_livelist_delete_check(spa);
return ((used != 0) || (clones_left));
}
boolean_t
dsl_errorscrub_active(dsl_scan_t *scn)
{
spa_t *spa = scn->scn_dp->dp_spa;
if (spa->spa_load_state != SPA_LOAD_NONE)
return (B_FALSE);
if (spa_shutting_down(spa))
return (B_FALSE);
if (dsl_errorscrubbing(scn->scn_dp))
return (B_TRUE);
return (B_FALSE);
}
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 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, psize, phys_birth))
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)
{
dsl_scan_t *scn = dp->dp_scan;
spa_t *spa = dp->dp_spa;
int err = 0;
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,
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 on %s in txg %llu; err=%u",
(longlong_t)scn->scn_visited_this_txg,
(longlong_t)
NSEC2MSEC(gethrtime() - scn->scn_sync_start_time),
spa->spa_name, (longlong_t)tx->tx_txg, err);
scn->scn_visited_this_txg = 0;
scn->scn_dedup_frees_this_txg = 0;
/*
* Write out changes to the DDT and the BRT that may be required
* as a result of the blocks freed. This ensures that the DDT
* and the BRT are clean when a scrub/resilver runs.
*/
ddt_sync(spa, tx->tx_txg);
brt_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 &&
!spa_livelist_delete_check(spa)) {
/* 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);
}
spa_notify_waiters(spa);
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);
}
static void
name_to_bookmark(char *buf, zbookmark_phys_t *zb)
{
zb->zb_objset = zfs_strtonum(buf, &buf);
ASSERT(*buf == ':');
zb->zb_object = zfs_strtonum(buf + 1, &buf);
ASSERT(*buf == ':');
zb->zb_level = (int)zfs_strtonum(buf + 1, &buf);
ASSERT(*buf == ':');
zb->zb_blkid = zfs_strtonum(buf + 1, &buf);
ASSERT(*buf == '\0');
}
static void
name_to_object(char *buf, uint64_t *obj)
{
*obj = zfs_strtonum(buf, &buf);
ASSERT(*buf == '\0');
}
static void
read_by_block_level(dsl_scan_t *scn, zbookmark_phys_t zb)
{
dsl_pool_t *dp = scn->scn_dp;
dsl_dataset_t *ds;
objset_t *os;
if (dsl_dataset_hold_obj(dp, zb.zb_objset, FTAG, &ds) != 0)
return;
if (dmu_objset_from_ds(ds, &os) != 0) {
dsl_dataset_rele(ds, FTAG);
return;
}
/*
* If the key is not loaded dbuf_dnode_findbp() will error out with
* EACCES. However in that case dnode_hold() will eventually call
* dbuf_read()->zio_wait() which may call spa_log_error(). This will
* lead to a deadlock due to us holding the mutex spa_errlist_lock.
* Avoid this by checking here if the keys are loaded, if not return.
* If the keys are not loaded the head_errlog feature is meaningless
* as we cannot figure out the birth txg of the block pointer.
*/
if (dsl_dataset_get_keystatus(ds->ds_dir) ==
ZFS_KEYSTATUS_UNAVAILABLE) {
dsl_dataset_rele(ds, FTAG);
return;
}
dnode_t *dn;
blkptr_t bp;
if (dnode_hold(os, zb.zb_object, FTAG, &dn) != 0) {
dsl_dataset_rele(ds, FTAG);
return;
}
rw_enter(&dn->dn_struct_rwlock, RW_READER);
int error = dbuf_dnode_findbp(dn, zb.zb_level, zb.zb_blkid, &bp, NULL,
NULL);
if (error) {
rw_exit(&dn->dn_struct_rwlock);
dnode_rele(dn, FTAG);
dsl_dataset_rele(ds, FTAG);
return;
}
if (!error && BP_IS_HOLE(&bp)) {
rw_exit(&dn->dn_struct_rwlock);
dnode_rele(dn, FTAG);
dsl_dataset_rele(ds, FTAG);
return;
}
int zio_flags = ZIO_FLAG_SCAN_THREAD | ZIO_FLAG_RAW |
ZIO_FLAG_CANFAIL | ZIO_FLAG_SCRUB;
/* If it's an intent log block, failure is expected. */
if (zb.zb_level == ZB_ZIL_LEVEL)
zio_flags |= ZIO_FLAG_SPECULATIVE;
ASSERT(!BP_IS_EMBEDDED(&bp));
scan_exec_io(dp, &bp, zio_flags, &zb, NULL);
rw_exit(&dn->dn_struct_rwlock);
dnode_rele(dn, FTAG);
dsl_dataset_rele(ds, FTAG);
}
/*
* We keep track of the scrubbed error blocks in "count". This will be used
* when deciding whether we exceeded zfs_scrub_error_blocks_per_txg. This
* function is modelled after check_filesystem().
*/
static int
scrub_filesystem(spa_t *spa, uint64_t fs, zbookmark_err_phys_t *zep,
int *count)
{
dsl_dataset_t *ds;
dsl_pool_t *dp = spa->spa_dsl_pool;
dsl_scan_t *scn = dp->dp_scan;
int error = dsl_dataset_hold_obj(dp, fs, FTAG, &ds);
if (error != 0)
return (error);
uint64_t latest_txg;
uint64_t txg_to_consider = spa->spa_syncing_txg;
boolean_t check_snapshot = B_TRUE;
error = find_birth_txg(ds, zep, &latest_txg);
/*
* If find_birth_txg() errors out, then err on the side of caution and
* proceed. In worst case scenario scrub all objects. If zep->zb_birth
* is 0 (e.g. in case of encryption with unloaded keys) also proceed to
* scrub all objects.
*/
if (error == 0 && zep->zb_birth == latest_txg) {
/* Block neither free nor re written. */
zbookmark_phys_t zb;
zep_to_zb(fs, zep, &zb);
scn->scn_zio_root = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL);
/* We have already acquired the config lock for spa */
read_by_block_level(scn, zb);
(void) zio_wait(scn->scn_zio_root);
scn->scn_zio_root = NULL;
scn->errorscrub_phys.dep_examined++;
scn->errorscrub_phys.dep_to_examine--;
(*count)++;
if ((*count) == zfs_scrub_error_blocks_per_txg ||
dsl_error_scrub_check_suspend(scn, &zb)) {
dsl_dataset_rele(ds, FTAG);
return (SET_ERROR(EFAULT));
}
check_snapshot = B_FALSE;
} else if (error == 0) {
txg_to_consider = latest_txg;
}
/*
* Retrieve the number of snapshots if the dataset is not a snapshot.
*/
uint64_t snap_count = 0;
if (dsl_dataset_phys(ds)->ds_snapnames_zapobj != 0) {
error = zap_count(spa->spa_meta_objset,
dsl_dataset_phys(ds)->ds_snapnames_zapobj, &snap_count);
if (error != 0) {
dsl_dataset_rele(ds, FTAG);
return (error);
}
}
if (snap_count == 0) {
/* Filesystem without snapshots. */
dsl_dataset_rele(ds, FTAG);
return (0);
}
uint64_t snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj;
uint64_t snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg;
dsl_dataset_rele(ds, FTAG);
/* Check only snapshots created from this file system. */
while (snap_obj != 0 && zep->zb_birth < snap_obj_txg &&
snap_obj_txg <= txg_to_consider) {
error = dsl_dataset_hold_obj(dp, snap_obj, FTAG, &ds);
if (error != 0)
return (error);
if (dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj != fs) {
snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj;
snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg;
dsl_dataset_rele(ds, FTAG);
continue;
}
boolean_t affected = B_TRUE;
if (check_snapshot) {
uint64_t blk_txg;
error = find_birth_txg(ds, zep, &blk_txg);
/*
* Scrub the snapshot also when zb_birth == 0 or when
* find_birth_txg() returns an error.
*/
affected = (error == 0 && zep->zb_birth == blk_txg) ||
(error != 0) || (zep->zb_birth == 0);
}
/* Scrub snapshots. */
if (affected) {
zbookmark_phys_t zb;
zep_to_zb(snap_obj, zep, &zb);
scn->scn_zio_root = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL);
/* We have already acquired the config lock for spa */
read_by_block_level(scn, zb);
(void) zio_wait(scn->scn_zio_root);
scn->scn_zio_root = NULL;
scn->errorscrub_phys.dep_examined++;
scn->errorscrub_phys.dep_to_examine--;
(*count)++;
if ((*count) == zfs_scrub_error_blocks_per_txg ||
dsl_error_scrub_check_suspend(scn, &zb)) {
dsl_dataset_rele(ds, FTAG);
return (EFAULT);
}
}
snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg;
snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj;
dsl_dataset_rele(ds, FTAG);
}
return (0);
}
void
dsl_errorscrub_sync(dsl_pool_t *dp, dmu_tx_t *tx)
{
spa_t *spa = dp->dp_spa;
dsl_scan_t *scn = dp->dp_scan;
/*
* Only process scans in sync pass 1.
*/
if (spa_sync_pass(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 (!dsl_errorscrub_active(scn) || dsl_errorscrub_is_paused(scn)) {
return;
}
if (dsl_scan_resilvering(scn->scn_dp)) {
/* cancel the error scrub if resilver started */
dsl_scan_cancel(scn->scn_dp);
return;
}
spa->spa_scrub_active = B_TRUE;
scn->scn_sync_start_time = gethrtime();
/*
* zfs_scan_suspend_progress can be set to disable scrub progress.
* See more detailed comment in dsl_scan_sync().
*/
if (zfs_scan_suspend_progress) {
uint64_t scan_time_ns = gethrtime() - scn->scn_sync_start_time;
int mintime = 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;
}
int i = 0;
zap_attribute_t *za;
zbookmark_phys_t *zb;
boolean_t limit_exceeded = B_FALSE;
za = kmem_zalloc(sizeof (zap_attribute_t), KM_SLEEP);
zb = kmem_zalloc(sizeof (zbookmark_phys_t), KM_SLEEP);
if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) {
for (; zap_cursor_retrieve(&scn->errorscrub_cursor, za) == 0;
zap_cursor_advance(&scn->errorscrub_cursor)) {
name_to_bookmark(za->za_name, zb);
scn->scn_zio_root = zio_root(dp->dp_spa, NULL,
NULL, ZIO_FLAG_CANFAIL);
dsl_pool_config_enter(dp, FTAG);
read_by_block_level(scn, *zb);
dsl_pool_config_exit(dp, FTAG);
(void) zio_wait(scn->scn_zio_root);
scn->scn_zio_root = NULL;
scn->errorscrub_phys.dep_examined += 1;
scn->errorscrub_phys.dep_to_examine -= 1;
i++;
if (i == zfs_scrub_error_blocks_per_txg ||
dsl_error_scrub_check_suspend(scn, zb)) {
limit_exceeded = B_TRUE;
break;
}
}
if (!limit_exceeded)
dsl_errorscrub_done(scn, B_TRUE, tx);
dsl_errorscrub_sync_state(scn, tx);
kmem_free(za, sizeof (*za));
kmem_free(zb, sizeof (*zb));
return;
}
int error = 0;
for (; zap_cursor_retrieve(&scn->errorscrub_cursor, za) == 0;
zap_cursor_advance(&scn->errorscrub_cursor)) {
zap_cursor_t *head_ds_cursor;
zap_attribute_t *head_ds_attr;
zbookmark_err_phys_t head_ds_block;
head_ds_cursor = kmem_zalloc(sizeof (zap_cursor_t), KM_SLEEP);
head_ds_attr = kmem_zalloc(sizeof (zap_attribute_t), KM_SLEEP);
uint64_t head_ds_err_obj = za->za_first_integer;
uint64_t head_ds;
name_to_object(za->za_name, &head_ds);
boolean_t config_held = B_FALSE;
uint64_t top_affected_fs;
for (zap_cursor_init(head_ds_cursor, spa->spa_meta_objset,
head_ds_err_obj); zap_cursor_retrieve(head_ds_cursor,
head_ds_attr) == 0; zap_cursor_advance(head_ds_cursor)) {
name_to_errphys(head_ds_attr->za_name, &head_ds_block);
/*
* In case we are called from spa_sync the pool
* config is already held.
*/
if (!dsl_pool_config_held(dp)) {
dsl_pool_config_enter(dp, FTAG);
config_held = B_TRUE;
}
error = find_top_affected_fs(spa,
head_ds, &head_ds_block, &top_affected_fs);
if (error)
break;
error = scrub_filesystem(spa, top_affected_fs,
&head_ds_block, &i);
if (error == SET_ERROR(EFAULT)) {
limit_exceeded = B_TRUE;
break;
}
}
zap_cursor_fini(head_ds_cursor);
kmem_free(head_ds_cursor, sizeof (*head_ds_cursor));
kmem_free(head_ds_attr, sizeof (*head_ds_attr));
if (config_held)
dsl_pool_config_exit(dp, FTAG);
}
kmem_free(za, sizeof (*za));
kmem_free(zb, sizeof (*zb));
if (!limit_exceeded)
dsl_errorscrub_done(scn, B_TRUE, tx);
dsl_errorscrub_sync_state(scn, tx);
}
/*
* This is the primary entry point for scans that is called from syncing
* context. Scans must happen entirely during syncing context so that we
* can guarantee that blocks we are currently scanning will not change out
* from under us. While a scan is active, this function 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)
{
int err = 0;
dsl_scan_t *scn = dp->dp_scan;
spa_t *spa = dp->dp_spa;
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). 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) ||
(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 on %s txg=%llu",
func, dp->dp_spa->spa_name, (longlong_t)tx->tx_txg);
dsl_scan_setup_sync(&func, tx);
}
/*
* Only process scans in sync pass 1.
*/
if (spa_sync_pass(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_dedup_frees_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 suspend, 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;
uint_t 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;
}
/*
* Disabled by default, set zfs_scan_report_txgs to report
* average performance over the last zfs_scan_report_txgs TXGs.
*/
if (zfs_scan_report_txgs != 0 &&
tx->tx_txg % zfs_scan_report_txgs == 0) {
scn->scn_issued_before_pass += spa->spa_scan_pass_issued;
spa_scan_stat_init(spa);
}
/*
* 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 consistent 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 for %s",
spa->spa_name);
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 for %s",
spa->spa_name);
scn->scn_clearing = B_TRUE;
} else if (!should_clear && scn->scn_clearing) {
zfs_dbgmsg("finish scan clearing for %s",
spa->spa_name);
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;
/*
* Calculate the max number of in-flight bytes for pool-wide
* scanning operations (minimum 1MB, maximum 1/4 of arc_c_max).
* Limits for the issuing phase are done per top-level vdev and
* are handled separately.
*/
scn->scn_maxinflight_bytes = MIN(arc_c_max / 4, MAX(1ULL << 20,
zfs_scan_vdev_limit * dsl_scan_count_data_disks(spa)));
if (scnp->scn_ddt_bookmark.ddb_class <=
scnp->scn_ddt_class_max) {
ASSERT(ZB_IS_ZERO(&scnp->scn_bookmark));
zfs_dbgmsg("doing scan sync for %s txg %llu; "
"ddt bm=%llu/%llu/%llu/%llx",
spa->spa_name,
(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 for %s txg %llu; "
"bm=%llu/%llu/%llu/%llu",
spa->spa_name,
(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 of %s in %llums "
"(%llu os's, %llu holes, %llu < mintxg, "
"%llu in ddt, %llu > maxtxg)",
(longlong_t)scn->scn_visited_this_txg,
spa->spa_name,
(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;
scn->scn_issued_before_pass +=
spa->spa_scan_pass_issued;
spa_scan_stat_init(spa);
}
zfs_dbgmsg("scan complete for %s txg %llu",
spa->spa_name,
(longlong_t)tx->tx_txg);
}
} else if (scn->scn_is_sorted && scn->scn_queues_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("scan issued %llu blocks for %s (%llu segs) "
"in %llums (avg_block_size = %llu, avg_seg_size = %llu)",
(longlong_t)scn->scn_zios_this_txg,
spa->spa_name,
(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 for %s",
(longlong_t)tx->tx_txg,
spa->spa_name);
ASSERT3U(scn->scn_done_txg, !=, 0);
ASSERT0(spa->spa_scrub_inflight);
ASSERT0(scn->scn_queues_pending);
dsl_scan_done(scn, B_TRUE, tx);
sync_type = SYNC_MANDATORY;
}
dsl_scan_sync_state(scn, tx, sync_type);
}
static void
count_block_issued(spa_t *spa, const blkptr_t *bp, boolean_t all)
{
/*
* 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.
*/
atomic_add_64(&spa->spa_scan_pass_issued,
all ? BP_GET_ASIZE(bp) : DVA_GET_ASIZE(&bp->blk_dva[0]));
}
static void
count_block_skipped(dsl_scan_t *scn, const blkptr_t *bp, boolean_t all)
{
if (BP_IS_EMBEDDED(bp))
return;
atomic_add_64(&scn->scn_phys.scn_skipped,
all ? BP_GET_ASIZE(bp) : DVA_GET_ASIZE(&bp->blk_dva[0]));
}
static void
count_block(zfs_all_blkstats_t *zab, const blkptr_t *bp)
{
/*
* If we resume after a reboot, zab will be NULL; don't record
* incomplete stats in that case.
*/
if (zab == NULL)
return;
for (int 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;
}
}
}
static void
scan_io_queue_insert_impl(dsl_scan_io_queue_t *queue, scan_io_t *sio)
{
avl_index_t idx;
dsl_scan_t *scn = queue->q_scn;
ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock));
if (unlikely(avl_is_empty(&queue->q_sios_by_addr)))
atomic_add_64(&scn->scn_queues_pending, 1);
if (avl_find(&queue->q_sios_by_addr, sio, &idx) != NULL) {
/* block is already scheduled for reading */
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),
SIO_GET_ASIZE(sio));
}
/*
* 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)
{
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;
queue->q_last_ext_addr = -1;
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 = B_FALSE;
int zio_flags = ZIO_FLAG_SCAN_THREAD | ZIO_FLAG_RAW | ZIO_FLAG_CANFAIL;
count_block(dp->dp_blkstats, bp);
if (phys_birth <= scn->scn_phys.scn_min_txg ||
phys_birth >= scn->scn_phys.scn_max_txg) {
count_block_skipped(scn, bp, B_TRUE);
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 (int 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(8) status can make useful progress reports.
*/
uint64_t asize = DVA_GET_ASIZE(dva);
scn->scn_phys.scn_examined += asize;
spa->spa_scan_pass_exam += asize;
/* 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_skipped(scn, bp, B_TRUE);
}
/* 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))) {
if (dsl_errorscrubbing(spa->spa_dsl_pool) &&
!dsl_errorscrub_is_paused(spa->spa_dsl_pool->dp_scan)) {
atomic_inc_64(&spa->spa_dsl_pool->dp_scan
->errorscrub_phys.dep_errors);
} else {
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);
zio_t *pio;
if (queue == NULL) {
ASSERT3U(scn->scn_maxinflight_bytes, >, 0);
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);
pio = scn->scn_zio_root;
} else {
kmutex_t *q_lock = &queue->q_vd->vdev_scan_io_queue_lock;
ASSERT3U(queue->q_maxinflight_bytes, >, 0);
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);
pio = queue->q_zio;
mutex_exit(q_lock);
}
ASSERT(pio != NULL);
count_block_issued(spa, bp, queue == NULL);
zio_nowait(zio_read(pio, 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 effectively multiplies and divides by 128).
*
* Since we do not care if one extent is only few percent better than another,
* compress the score into 6 bits via binary logarithm AKA highbit64() and
* put into otherwise unused due to ashift high bits of offset. This allows
* to reduce q_exts_by_size B-tree elements to only 64 bits and compare them
* with single operation. Plus it makes scrubs more sequential and reduces
* chances that minor extent change move it within the B-tree.
*/
__attribute__((always_inline)) inline
static int
ext_size_compare(const void *x, const void *y)
{
const uint64_t *a = x, *b = y;
return (TREE_CMP(*a, *b));
}
ZFS_BTREE_FIND_IN_BUF_FUNC(ext_size_find_in_buf, uint64_t,
ext_size_compare)
static void
ext_size_create(range_tree_t *rt, void *arg)
{
(void) rt;
zfs_btree_t *size_tree = arg;
zfs_btree_create(size_tree, ext_size_compare, ext_size_find_in_buf,
sizeof (uint64_t));
}
static void
ext_size_destroy(range_tree_t *rt, void *arg)
{
(void) rt;
zfs_btree_t *size_tree = arg;
ASSERT0(zfs_btree_numnodes(size_tree));
zfs_btree_destroy(size_tree);
}
static uint64_t
ext_size_value(range_tree_t *rt, range_seg_gap_t *rsg)
{
(void) rt;
uint64_t size = rsg->rs_end - rsg->rs_start;
uint64_t score = rsg->rs_fill + ((((rsg->rs_fill << 7) / size) *
fill_weight * rsg->rs_fill) >> 7);
ASSERT3U(rt->rt_shift, >=, 8);
return (((uint64_t)(64 - highbit64(score)) << 56) | rsg->rs_start);
}
static void
ext_size_add(range_tree_t *rt, range_seg_t *rs, void *arg)
{
zfs_btree_t *size_tree = arg;
ASSERT3U(rt->rt_type, ==, RANGE_SEG_GAP);
uint64_t v = ext_size_value(rt, (range_seg_gap_t *)rs);
zfs_btree_add(size_tree, &v);
}
static void
ext_size_remove(range_tree_t *rt, range_seg_t *rs, void *arg)
{
zfs_btree_t *size_tree = arg;
ASSERT3U(rt->rt_type, ==, RANGE_SEG_GAP);
uint64_t v = ext_size_value(rt, (range_seg_gap_t *)rs);
zfs_btree_remove(size_tree, &v);
}
static void
ext_size_vacate(range_tree_t *rt, void *arg)
{
zfs_btree_t *size_tree = arg;
zfs_btree_clear(size_tree);
zfs_btree_destroy(size_tree);
ext_size_create(rt, arg);
}
static const range_tree_ops_t ext_size_ops = {
.rtop_create = ext_size_create,
.rtop_destroy = ext_size_destroy,
.rtop_add = ext_size_add,
.rtop_remove = ext_size_remove,
.rtop_vacate = ext_size_vacate
};
/*
* 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 (TREE_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;
q->q_last_ext_addr = -1;
cv_init(&q->q_zio_cv, NULL, CV_DEFAULT, NULL);
q->q_exts_by_addr = range_tree_create_gap(&ext_size_ops, RANGE_SEG_GAP,
&q->q_exts_by_size, 0, vd->vdev_ashift, 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;
ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock));
if (!avl_is_empty(&queue->q_sios_by_addr))
atomic_add_64(&scn->scn_queues_pending, -1);
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)));
queue->q_sio_memused -= SIO_GET_MUSED(sio);
sio_free(sio);
}
ASSERT0(queue->q_sio_memused);
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) {
blkptr_t tmpbp;
/* Got it while it was cold in the queue */
ASSERT3U(start, ==, SIO_GET_OFFSET(sio));
ASSERT3U(size, ==, SIO_GET_ASIZE(sio));
avl_remove(&queue->q_sios_by_addr, sio);
if (avl_is_empty(&queue->q_sios_by_addr))
atomic_add_64(&scn->scn_queues_pending, -1);
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);
/* count the block as though we skipped it */
sio2bp(sio, &tmpbp);
count_block_skipped(scn, &tmpbp, B_FALSE);
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);
}
/*
* Check if a vdev needs resilvering (non-empty DTL), if so, and resilver has
* not started, start it. Otherwise, only restart if max txg in DTL range is
* greater than the max txg in the current scan. If the DTL max is less than
* the scan max, then the vdev has not missed any new data since the resilver
* started, so a restart is not needed.
*/
void
dsl_scan_assess_vdev(dsl_pool_t *dp, vdev_t *vd)
{
uint64_t min, max;
if (!vdev_resilver_needed(vd, &min, &max))
return;
if (!dsl_scan_resilvering(dp)) {
spa_async_request(dp->dp_spa, SPA_ASYNC_RESILVER);
return;
}
if (max <= dp->dp_scan->scn_phys.scn_max_txg)
return;
/* restart is needed, check if it can be deferred */
if (spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_RESILVER_DEFER))
vdev_defer_resilver(vd);
else
spa_async_request(dp->dp_spa, SPA_ASYNC_RESILVER);
}
ZFS_MODULE_PARAM(zfs, zfs_, scan_vdev_limit, U64, ZMOD_RW,
"Max bytes in flight per leaf vdev for scrubs and resilvers");
ZFS_MODULE_PARAM(zfs, zfs_, scrub_min_time_ms, UINT, ZMOD_RW,
"Min millisecs to scrub per txg");
ZFS_MODULE_PARAM(zfs, zfs_, obsolete_min_time_ms, UINT, ZMOD_RW,
"Min millisecs to obsolete per txg");
ZFS_MODULE_PARAM(zfs, zfs_, free_min_time_ms, UINT, ZMOD_RW,
"Min millisecs to free per txg");
ZFS_MODULE_PARAM(zfs, zfs_, resilver_min_time_ms, UINT, ZMOD_RW,
"Min millisecs to resilver per txg");
ZFS_MODULE_PARAM(zfs, zfs_, scan_suspend_progress, INT, ZMOD_RW,
"Set to prevent scans from progressing");
ZFS_MODULE_PARAM(zfs, zfs_, no_scrub_io, INT, ZMOD_RW,
"Set to disable scrub I/O");
ZFS_MODULE_PARAM(zfs, zfs_, no_scrub_prefetch, INT, ZMOD_RW,
"Set to disable scrub prefetching");
ZFS_MODULE_PARAM(zfs, zfs_, async_block_max_blocks, U64, ZMOD_RW,
"Max number of blocks freed in one txg");
ZFS_MODULE_PARAM(zfs, zfs_, max_async_dedup_frees, U64, ZMOD_RW,
"Max number of dedup blocks freed in one txg");
ZFS_MODULE_PARAM(zfs, zfs_, free_bpobj_enabled, INT, ZMOD_RW,
"Enable processing of the free_bpobj");
ZFS_MODULE_PARAM(zfs, zfs_, scan_blkstats, INT, ZMOD_RW,
"Enable block statistics calculation during scrub");
ZFS_MODULE_PARAM(zfs, zfs_, scan_mem_lim_fact, UINT, ZMOD_RW,
"Fraction of RAM for scan hard limit");
ZFS_MODULE_PARAM(zfs, zfs_, scan_issue_strategy, UINT, ZMOD_RW,
"IO issuing strategy during scrubbing. 0 = default, 1 = LBA, 2 = size");
ZFS_MODULE_PARAM(zfs, zfs_, scan_legacy, INT, ZMOD_RW,
"Scrub using legacy non-sequential method");
ZFS_MODULE_PARAM(zfs, zfs_, scan_checkpoint_intval, UINT, ZMOD_RW,
"Scan progress on-disk checkpointing interval");
ZFS_MODULE_PARAM(zfs, zfs_, scan_max_ext_gap, U64, ZMOD_RW,
"Max gap in bytes between sequential scrub / resilver I/Os");
ZFS_MODULE_PARAM(zfs, zfs_, scan_mem_lim_soft_fact, UINT, ZMOD_RW,
"Fraction of hard limit used as soft limit");
ZFS_MODULE_PARAM(zfs, zfs_, scan_strict_mem_lim, INT, ZMOD_RW,
"Tunable to attempt to reduce lock contention");
ZFS_MODULE_PARAM(zfs, zfs_, scan_fill_weight, UINT, ZMOD_RW,
"Tunable to adjust bias towards more filled segments during scans");
ZFS_MODULE_PARAM(zfs, zfs_, scan_report_txgs, UINT, ZMOD_RW,
"Tunable to report resilver performance over the last N txgs");
ZFS_MODULE_PARAM(zfs, zfs_, resilver_disable_defer, INT, ZMOD_RW,
"Process all resilvers immediately");
ZFS_MODULE_PARAM(zfs, zfs_, scrub_error_blocks_per_txg, UINT, ZMOD_RW,
"Error blocks to be scrubbed in one txg");
/* END CSTYLED */
diff --git a/sys/contrib/openzfs/module/zfs/metaslab.c b/sys/contrib/openzfs/module/zfs/metaslab.c
index e0d4a6a63508..0983ba143a1d 100644
--- a/sys/contrib/openzfs/module/zfs/metaslab.c
+++ b/sys/contrib/openzfs/module/zfs/metaslab.c
@@ -1,6284 +1,6286 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2019 by Delphix. All rights reserved.
* Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
* Copyright (c) 2015, Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
*/
#include <sys/zfs_context.h>
#include <sys/dmu.h>
#include <sys/dmu_tx.h>
#include <sys/space_map.h>
#include <sys/metaslab_impl.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_draid.h>
#include <sys/zio.h>
#include <sys/spa_impl.h>
#include <sys/zfeature.h>
#include <sys/vdev_indirect_mapping.h>
#include <sys/zap.h>
#include <sys/btree.h>
#define GANG_ALLOCATION(flags) \
((flags) & (METASLAB_GANG_CHILD | METASLAB_GANG_HEADER))
/*
* Metaslab granularity, in bytes. This is roughly similar to what would be
* referred to as the "stripe size" in traditional RAID arrays. In normal
* operation, we will try to write this amount of data to each disk before
* moving on to the next top-level vdev.
*/
static uint64_t metaslab_aliquot = 1024 * 1024;
/*
* For testing, make some blocks above a certain size be gang blocks.
*/
uint64_t metaslab_force_ganging = SPA_MAXBLOCKSIZE + 1;
/*
* Of blocks of size >= metaslab_force_ganging, actually gang them this often.
*/
uint_t metaslab_force_ganging_pct = 3;
/*
* In pools where the log space map feature is not enabled we touch
* multiple metaslabs (and their respective space maps) with each
* transaction group. Thus, we benefit from having a small space map
* block size since it allows us to issue more I/O operations scattered
* around the disk. So a sane default for the space map block size
* is 8~16K.
*/
int zfs_metaslab_sm_blksz_no_log = (1 << 14);
/*
* When the log space map feature is enabled, we accumulate a lot of
* changes per metaslab that are flushed once in a while so we benefit
* from a bigger block size like 128K for the metaslab space maps.
*/
int zfs_metaslab_sm_blksz_with_log = (1 << 17);
/*
* The in-core space map representation is more compact than its on-disk form.
* The zfs_condense_pct determines how much more compact the in-core
* space map representation must be before we compact it on-disk.
* Values should be greater than or equal to 100.
*/
uint_t zfs_condense_pct = 200;
/*
* Condensing a metaslab is not guaranteed to actually reduce the amount of
* space used on disk. In particular, a space map uses data in increments of
* MAX(1 << ashift, space_map_blksz), so a metaslab might use the
* same number of blocks after condensing. Since the goal of condensing is to
* reduce the number of IOPs required to read the space map, we only want to
* condense when we can be sure we will reduce the number of blocks used by the
* space map. Unfortunately, we cannot precisely compute whether or not this is
* the case in metaslab_should_condense since we are holding ms_lock. Instead,
* we apply the following heuristic: do not condense a spacemap unless the
* uncondensed size consumes greater than zfs_metaslab_condense_block_threshold
* blocks.
*/
static const int zfs_metaslab_condense_block_threshold = 4;
/*
* The zfs_mg_noalloc_threshold defines which metaslab groups should
* be eligible for allocation. The value is defined as a percentage of
* free space. Metaslab groups that have more free space than
* zfs_mg_noalloc_threshold are always eligible for allocations. Once
* a metaslab group's free space is less than or equal to the
* zfs_mg_noalloc_threshold the allocator will avoid allocating to that
* group unless all groups in the pool have reached zfs_mg_noalloc_threshold.
* Once all groups in the pool reach zfs_mg_noalloc_threshold then all
* groups are allowed to accept allocations. Gang blocks are always
* eligible to allocate on any metaslab group. The default value of 0 means
* no metaslab group will be excluded based on this criterion.
*/
static uint_t zfs_mg_noalloc_threshold = 0;
/*
* Metaslab groups are considered eligible for allocations if their
* fragmentation metric (measured as a percentage) is less than or
* equal to zfs_mg_fragmentation_threshold. If a metaslab group
* exceeds this threshold then it will be skipped unless all metaslab
* groups within the metaslab class have also crossed this threshold.
*
* This tunable was introduced to avoid edge cases where we continue
* allocating from very fragmented disks in our pool while other, less
* fragmented disks, exists. On the other hand, if all disks in the
* pool are uniformly approaching the threshold, the threshold can
* be a speed bump in performance, where we keep switching the disks
* that we allocate from (e.g. we allocate some segments from disk A
* making it bypassing the threshold while freeing segments from disk
* B getting its fragmentation below the threshold).
*
* Empirically, we've seen that our vdev selection for allocations is
* good enough that fragmentation increases uniformly across all vdevs
* the majority of the time. Thus we set the threshold percentage high
* enough to avoid hitting the speed bump on pools that are being pushed
* to the edge.
*/
static uint_t zfs_mg_fragmentation_threshold = 95;
/*
* Allow metaslabs to keep their active state as long as their fragmentation
* percentage is less than or equal to zfs_metaslab_fragmentation_threshold. An
* active metaslab that exceeds this threshold will no longer keep its active
* status allowing better metaslabs to be selected.
*/
static uint_t zfs_metaslab_fragmentation_threshold = 70;
/*
* When set will load all metaslabs when pool is first opened.
*/
int metaslab_debug_load = B_FALSE;
/*
* When set will prevent metaslabs from being unloaded.
*/
static int metaslab_debug_unload = B_FALSE;
/*
* Minimum size which forces the dynamic allocator to change
* it's allocation strategy. Once the space map cannot satisfy
* an allocation of this size then it switches to using more
* aggressive strategy (i.e search by size rather than offset).
*/
uint64_t metaslab_df_alloc_threshold = SPA_OLD_MAXBLOCKSIZE;
/*
* The minimum free space, in percent, which must be available
* in a space map to continue allocations in a first-fit fashion.
* Once the space map's free space drops below this level we dynamically
* switch to using best-fit allocations.
*/
uint_t metaslab_df_free_pct = 4;
/*
* Maximum distance to search forward from the last offset. Without this
* limit, fragmented pools can see >100,000 iterations and
* metaslab_block_picker() becomes the performance limiting factor on
* high-performance storage.
*
* With the default setting of 16MB, we typically see less than 500
* iterations, even with very fragmented, ashift=9 pools. The maximum number
* of iterations possible is:
* metaslab_df_max_search / (2 * (1<<ashift))
* With the default setting of 16MB this is 16*1024 (with ashift=9) or
* 2048 (with ashift=12).
*/
static uint_t metaslab_df_max_search = 16 * 1024 * 1024;
/*
* Forces the metaslab_block_picker function to search for at least this many
* segments forwards until giving up on finding a segment that the allocation
* will fit into.
*/
static const uint32_t metaslab_min_search_count = 100;
/*
* If we are not searching forward (due to metaslab_df_max_search,
* metaslab_df_free_pct, or metaslab_df_alloc_threshold), this tunable
* controls what segment is used. If it is set, we will use the largest free
* segment. If it is not set, we will use a segment of exactly the requested
* size (or larger).
*/
static int metaslab_df_use_largest_segment = B_FALSE;
/*
* These tunables control how long a metaslab will remain loaded after the
* last allocation from it. A metaslab can't be unloaded until at least
* metaslab_unload_delay TXG's and metaslab_unload_delay_ms milliseconds
* have elapsed. However, zfs_metaslab_mem_limit may cause it to be
* unloaded sooner. These settings are intended to be generous -- to keep
* metaslabs loaded for a long time, reducing the rate of metaslab loading.
*/
static uint_t metaslab_unload_delay = 32;
static uint_t metaslab_unload_delay_ms = 10 * 60 * 1000; /* ten minutes */
/*
* Max number of metaslabs per group to preload.
*/
uint_t metaslab_preload_limit = 10;
/*
* Enable/disable preloading of metaslab.
*/
static int metaslab_preload_enabled = B_TRUE;
/*
* Enable/disable fragmentation weighting on metaslabs.
*/
static int metaslab_fragmentation_factor_enabled = B_TRUE;
/*
* Enable/disable lba weighting (i.e. outer tracks are given preference).
*/
static int metaslab_lba_weighting_enabled = B_TRUE;
/*
* Enable/disable metaslab group biasing.
*/
static int metaslab_bias_enabled = B_TRUE;
/*
* Enable/disable remapping of indirect DVAs to their concrete vdevs.
*/
static const boolean_t zfs_remap_blkptr_enable = B_TRUE;
/*
* Enable/disable segment-based metaslab selection.
*/
static int zfs_metaslab_segment_weight_enabled = B_TRUE;
/*
* When using segment-based metaslab selection, we will continue
* allocating from the active metaslab until we have exhausted
* zfs_metaslab_switch_threshold of its buckets.
*/
static int zfs_metaslab_switch_threshold = 2;
/*
* Internal switch to enable/disable the metaslab allocation tracing
* facility.
*/
static const boolean_t metaslab_trace_enabled = B_FALSE;
/*
* Maximum entries that the metaslab allocation tracing facility will keep
* in a given list when running in non-debug mode. We limit the number
* of entries in non-debug mode to prevent us from using up too much memory.
* The limit should be sufficiently large that we don't expect any allocation
* to every exceed this value. In debug mode, the system will panic if this
* limit is ever reached allowing for further investigation.
*/
static const uint64_t metaslab_trace_max_entries = 5000;
/*
* Maximum number of metaslabs per group that can be disabled
* simultaneously.
*/
static const int max_disabled_ms = 3;
/*
* Time (in seconds) to respect ms_max_size when the metaslab is not loaded.
* To avoid 64-bit overflow, don't set above UINT32_MAX.
*/
static uint64_t zfs_metaslab_max_size_cache_sec = 1 * 60 * 60; /* 1 hour */
/*
* Maximum percentage of memory to use on storing loaded metaslabs. If loading
* a metaslab would take it over this percentage, the oldest selected metaslab
* is automatically unloaded.
*/
static uint_t zfs_metaslab_mem_limit = 25;
/*
* Force the per-metaslab range trees to use 64-bit integers to store
* segments. Used for debugging purposes.
*/
static const boolean_t zfs_metaslab_force_large_segs = B_FALSE;
/*
* By default we only store segments over a certain size in the size-sorted
* metaslab trees (ms_allocatable_by_size and
* ms_unflushed_frees_by_size). This dramatically reduces memory usage and
* improves load and unload times at the cost of causing us to use slightly
* larger segments than we would otherwise in some cases.
*/
static const uint32_t metaslab_by_size_min_shift = 14;
/*
* If not set, we will first try normal allocation. If that fails then
* we will do a gang allocation. If that fails then we will do a "try hard"
* gang allocation. If that fails then we will have a multi-layer gang
* block.
*
* If set, we will first try normal allocation. If that fails then
* we will do a "try hard" allocation. If that fails we will do a gang
* allocation. If that fails we will do a "try hard" gang allocation. If
* that fails then we will have a multi-layer gang block.
*/
static int zfs_metaslab_try_hard_before_gang = B_FALSE;
/*
* When not trying hard, we only consider the best zfs_metaslab_find_max_tries
* metaslabs. This improves performance, especially when there are many
* metaslabs per vdev and the allocation can't actually be satisfied (so we
* would otherwise iterate all the metaslabs). If there is a metaslab with a
* worse weight but it can actually satisfy the allocation, we won't find it
* until trying hard. This may happen if the worse metaslab is not loaded
* (and the true weight is better than we have calculated), or due to weight
* bucketization. E.g. we are looking for a 60K segment, and the best
* metaslabs all have free segments in the 32-63K bucket, but the best
* zfs_metaslab_find_max_tries metaslabs have ms_max_size <60KB, and a
* subsequent metaslab has ms_max_size >60KB (but fewer segments in this
* bucket, and therefore a lower weight).
*/
static uint_t zfs_metaslab_find_max_tries = 100;
static uint64_t metaslab_weight(metaslab_t *, boolean_t);
static void metaslab_set_fragmentation(metaslab_t *, boolean_t);
static void metaslab_free_impl(vdev_t *, uint64_t, uint64_t, boolean_t);
static void metaslab_check_free_impl(vdev_t *, uint64_t, uint64_t);
static void metaslab_passivate(metaslab_t *msp, uint64_t weight);
static uint64_t metaslab_weight_from_range_tree(metaslab_t *msp);
static void metaslab_flush_update(metaslab_t *, dmu_tx_t *);
static unsigned int metaslab_idx_func(multilist_t *, void *);
static void metaslab_evict(metaslab_t *, uint64_t);
static void metaslab_rt_add(range_tree_t *rt, range_seg_t *rs, void *arg);
kmem_cache_t *metaslab_alloc_trace_cache;
typedef struct metaslab_stats {
kstat_named_t metaslabstat_trace_over_limit;
kstat_named_t metaslabstat_reload_tree;
kstat_named_t metaslabstat_too_many_tries;
kstat_named_t metaslabstat_try_hard;
} metaslab_stats_t;
static metaslab_stats_t metaslab_stats = {
{ "trace_over_limit", KSTAT_DATA_UINT64 },
{ "reload_tree", KSTAT_DATA_UINT64 },
{ "too_many_tries", KSTAT_DATA_UINT64 },
{ "try_hard", KSTAT_DATA_UINT64 },
};
#define METASLABSTAT_BUMP(stat) \
atomic_inc_64(&metaslab_stats.stat.value.ui64);
static kstat_t *metaslab_ksp;
void
metaslab_stat_init(void)
{
ASSERT(metaslab_alloc_trace_cache == NULL);
metaslab_alloc_trace_cache = kmem_cache_create(
"metaslab_alloc_trace_cache", sizeof (metaslab_alloc_trace_t),
0, NULL, NULL, NULL, NULL, NULL, 0);
metaslab_ksp = kstat_create("zfs", 0, "metaslab_stats",
"misc", KSTAT_TYPE_NAMED, sizeof (metaslab_stats) /
sizeof (kstat_named_t), KSTAT_FLAG_VIRTUAL);
if (metaslab_ksp != NULL) {
metaslab_ksp->ks_data = &metaslab_stats;
kstat_install(metaslab_ksp);
}
}
void
metaslab_stat_fini(void)
{
if (metaslab_ksp != NULL) {
kstat_delete(metaslab_ksp);
metaslab_ksp = NULL;
}
kmem_cache_destroy(metaslab_alloc_trace_cache);
metaslab_alloc_trace_cache = NULL;
}
/*
* ==========================================================================
* Metaslab classes
* ==========================================================================
*/
metaslab_class_t *
metaslab_class_create(spa_t *spa, const metaslab_ops_t *ops)
{
metaslab_class_t *mc;
mc = kmem_zalloc(offsetof(metaslab_class_t,
mc_allocator[spa->spa_alloc_count]), KM_SLEEP);
mc->mc_spa = spa;
mc->mc_ops = ops;
mutex_init(&mc->mc_lock, NULL, MUTEX_DEFAULT, NULL);
multilist_create(&mc->mc_metaslab_txg_list, sizeof (metaslab_t),
offsetof(metaslab_t, ms_class_txg_node), metaslab_idx_func);
for (int i = 0; i < spa->spa_alloc_count; i++) {
metaslab_class_allocator_t *mca = &mc->mc_allocator[i];
mca->mca_rotor = NULL;
zfs_refcount_create_tracked(&mca->mca_alloc_slots);
}
return (mc);
}
void
metaslab_class_destroy(metaslab_class_t *mc)
{
spa_t *spa = mc->mc_spa;
ASSERT(mc->mc_alloc == 0);
ASSERT(mc->mc_deferred == 0);
ASSERT(mc->mc_space == 0);
ASSERT(mc->mc_dspace == 0);
for (int i = 0; i < spa->spa_alloc_count; i++) {
metaslab_class_allocator_t *mca = &mc->mc_allocator[i];
ASSERT(mca->mca_rotor == NULL);
zfs_refcount_destroy(&mca->mca_alloc_slots);
}
mutex_destroy(&mc->mc_lock);
multilist_destroy(&mc->mc_metaslab_txg_list);
kmem_free(mc, offsetof(metaslab_class_t,
mc_allocator[spa->spa_alloc_count]));
}
int
metaslab_class_validate(metaslab_class_t *mc)
{
metaslab_group_t *mg;
vdev_t *vd;
/*
* Must hold one of the spa_config locks.
*/
ASSERT(spa_config_held(mc->mc_spa, SCL_ALL, RW_READER) ||
spa_config_held(mc->mc_spa, SCL_ALL, RW_WRITER));
if ((mg = mc->mc_allocator[0].mca_rotor) == NULL)
return (0);
do {
vd = mg->mg_vd;
ASSERT(vd->vdev_mg != NULL);
ASSERT3P(vd->vdev_top, ==, vd);
ASSERT3P(mg->mg_class, ==, mc);
ASSERT3P(vd->vdev_ops, !=, &vdev_hole_ops);
} while ((mg = mg->mg_next) != mc->mc_allocator[0].mca_rotor);
return (0);
}
static void
metaslab_class_space_update(metaslab_class_t *mc, int64_t alloc_delta,
int64_t defer_delta, int64_t space_delta, int64_t dspace_delta)
{
atomic_add_64(&mc->mc_alloc, alloc_delta);
atomic_add_64(&mc->mc_deferred, defer_delta);
atomic_add_64(&mc->mc_space, space_delta);
atomic_add_64(&mc->mc_dspace, dspace_delta);
}
uint64_t
metaslab_class_get_alloc(metaslab_class_t *mc)
{
return (mc->mc_alloc);
}
uint64_t
metaslab_class_get_deferred(metaslab_class_t *mc)
{
return (mc->mc_deferred);
}
uint64_t
metaslab_class_get_space(metaslab_class_t *mc)
{
return (mc->mc_space);
}
uint64_t
metaslab_class_get_dspace(metaslab_class_t *mc)
{
return (spa_deflate(mc->mc_spa) ? mc->mc_dspace : mc->mc_space);
}
void
metaslab_class_histogram_verify(metaslab_class_t *mc)
{
spa_t *spa = mc->mc_spa;
vdev_t *rvd = spa->spa_root_vdev;
uint64_t *mc_hist;
int i;
if ((zfs_flags & ZFS_DEBUG_HISTOGRAM_VERIFY) == 0)
return;
mc_hist = kmem_zalloc(sizeof (uint64_t) * RANGE_TREE_HISTOGRAM_SIZE,
KM_SLEEP);
mutex_enter(&mc->mc_lock);
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
metaslab_group_t *mg = vdev_get_mg(tvd, mc);
/*
* Skip any holes, uninitialized top-levels, or
* vdevs that are not in this metalab class.
*/
if (!vdev_is_concrete(tvd) || tvd->vdev_ms_shift == 0 ||
mg->mg_class != mc) {
continue;
}
IMPLY(mg == mg->mg_vd->vdev_log_mg,
mc == spa_embedded_log_class(mg->mg_vd->vdev_spa));
for (i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++)
mc_hist[i] += mg->mg_histogram[i];
}
for (i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++) {
VERIFY3U(mc_hist[i], ==, mc->mc_histogram[i]);
}
mutex_exit(&mc->mc_lock);
kmem_free(mc_hist, sizeof (uint64_t) * RANGE_TREE_HISTOGRAM_SIZE);
}
/*
* Calculate the metaslab class's fragmentation metric. The metric
* is weighted based on the space contribution of each metaslab group.
* The return value will be a number between 0 and 100 (inclusive), or
* ZFS_FRAG_INVALID if the metric has not been set. See comment above the
* zfs_frag_table for more information about the metric.
*/
uint64_t
metaslab_class_fragmentation(metaslab_class_t *mc)
{
vdev_t *rvd = mc->mc_spa->spa_root_vdev;
uint64_t fragmentation = 0;
spa_config_enter(mc->mc_spa, SCL_VDEV, FTAG, RW_READER);
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
metaslab_group_t *mg = tvd->vdev_mg;
/*
* Skip any holes, uninitialized top-levels,
* or vdevs that are not in this metalab class.
*/
if (!vdev_is_concrete(tvd) || tvd->vdev_ms_shift == 0 ||
mg->mg_class != mc) {
continue;
}
/*
* If a metaslab group does not contain a fragmentation
* metric then just bail out.
*/
if (mg->mg_fragmentation == ZFS_FRAG_INVALID) {
spa_config_exit(mc->mc_spa, SCL_VDEV, FTAG);
return (ZFS_FRAG_INVALID);
}
/*
* Determine how much this metaslab_group is contributing
* to the overall pool fragmentation metric.
*/
fragmentation += mg->mg_fragmentation *
metaslab_group_get_space(mg);
}
fragmentation /= metaslab_class_get_space(mc);
ASSERT3U(fragmentation, <=, 100);
spa_config_exit(mc->mc_spa, SCL_VDEV, FTAG);
return (fragmentation);
}
/*
* Calculate the amount of expandable space that is available in
* this metaslab class. If a device is expanded then its expandable
* space will be the amount of allocatable space that is currently not
* part of this metaslab class.
*/
uint64_t
metaslab_class_expandable_space(metaslab_class_t *mc)
{
vdev_t *rvd = mc->mc_spa->spa_root_vdev;
uint64_t space = 0;
spa_config_enter(mc->mc_spa, SCL_VDEV, FTAG, RW_READER);
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
metaslab_group_t *mg = tvd->vdev_mg;
if (!vdev_is_concrete(tvd) || tvd->vdev_ms_shift == 0 ||
mg->mg_class != mc) {
continue;
}
/*
* Calculate if we have enough space to add additional
* metaslabs. We report the expandable space in terms
* of the metaslab size since that's the unit of expansion.
*/
space += P2ALIGN(tvd->vdev_max_asize - tvd->vdev_asize,
1ULL << tvd->vdev_ms_shift);
}
spa_config_exit(mc->mc_spa, SCL_VDEV, FTAG);
return (space);
}
void
metaslab_class_evict_old(metaslab_class_t *mc, uint64_t txg)
{
multilist_t *ml = &mc->mc_metaslab_txg_list;
for (int i = 0; i < multilist_get_num_sublists(ml); i++) {
multilist_sublist_t *mls = multilist_sublist_lock(ml, i);
metaslab_t *msp = multilist_sublist_head(mls);
multilist_sublist_unlock(mls);
while (msp != NULL) {
mutex_enter(&msp->ms_lock);
/*
* If the metaslab has been removed from the list
* (which could happen if we were at the memory limit
* and it was evicted during this loop), then we can't
* proceed and we should restart the sublist.
*/
if (!multilist_link_active(&msp->ms_class_txg_node)) {
mutex_exit(&msp->ms_lock);
i--;
break;
}
mls = multilist_sublist_lock(ml, i);
metaslab_t *next_msp = multilist_sublist_next(mls, msp);
multilist_sublist_unlock(mls);
if (txg >
msp->ms_selected_txg + metaslab_unload_delay &&
gethrtime() > msp->ms_selected_time +
(uint64_t)MSEC2NSEC(metaslab_unload_delay_ms)) {
metaslab_evict(msp, txg);
} else {
/*
* Once we've hit a metaslab selected too
* recently to evict, we're done evicting for
* now.
*/
mutex_exit(&msp->ms_lock);
break;
}
mutex_exit(&msp->ms_lock);
msp = next_msp;
}
}
}
static int
metaslab_compare(const void *x1, const void *x2)
{
const metaslab_t *m1 = (const metaslab_t *)x1;
const metaslab_t *m2 = (const metaslab_t *)x2;
int sort1 = 0;
int sort2 = 0;
if (m1->ms_allocator != -1 && m1->ms_primary)
sort1 = 1;
else if (m1->ms_allocator != -1 && !m1->ms_primary)
sort1 = 2;
if (m2->ms_allocator != -1 && m2->ms_primary)
sort2 = 1;
else if (m2->ms_allocator != -1 && !m2->ms_primary)
sort2 = 2;
/*
* Sort inactive metaslabs first, then primaries, then secondaries. When
* selecting a metaslab to allocate from, an allocator first tries its
* primary, then secondary active metaslab. If it doesn't have active
* metaslabs, or can't allocate from them, it searches for an inactive
* metaslab to activate. If it can't find a suitable one, it will steal
* a primary or secondary metaslab from another allocator.
*/
if (sort1 < sort2)
return (-1);
if (sort1 > sort2)
return (1);
int cmp = TREE_CMP(m2->ms_weight, m1->ms_weight);
if (likely(cmp))
return (cmp);
IMPLY(TREE_CMP(m1->ms_start, m2->ms_start) == 0, m1 == m2);
return (TREE_CMP(m1->ms_start, m2->ms_start));
}
/*
* ==========================================================================
* Metaslab groups
* ==========================================================================
*/
/*
* Update the allocatable flag and the metaslab group's capacity.
* The allocatable flag is set to true if the capacity is below
* the zfs_mg_noalloc_threshold or has a fragmentation value that is
* greater than zfs_mg_fragmentation_threshold. If a metaslab group
* transitions from allocatable to non-allocatable or vice versa then the
* metaslab group's class is updated to reflect the transition.
*/
static void
metaslab_group_alloc_update(metaslab_group_t *mg)
{
vdev_t *vd = mg->mg_vd;
metaslab_class_t *mc = mg->mg_class;
vdev_stat_t *vs = &vd->vdev_stat;
boolean_t was_allocatable;
boolean_t was_initialized;
ASSERT(vd == vd->vdev_top);
ASSERT3U(spa_config_held(mc->mc_spa, SCL_ALLOC, RW_READER), ==,
SCL_ALLOC);
mutex_enter(&mg->mg_lock);
was_allocatable = mg->mg_allocatable;
was_initialized = mg->mg_initialized;
mg->mg_free_capacity = ((vs->vs_space - vs->vs_alloc) * 100) /
(vs->vs_space + 1);
mutex_enter(&mc->mc_lock);
/*
* If the metaslab group was just added then it won't
* have any space until we finish syncing out this txg.
* At that point we will consider it initialized and available
* for allocations. We also don't consider non-activated
* metaslab groups (e.g. vdevs that are in the middle of being removed)
* to be initialized, because they can't be used for allocation.
*/
mg->mg_initialized = metaslab_group_initialized(mg);
if (!was_initialized && mg->mg_initialized) {
mc->mc_groups++;
} else if (was_initialized && !mg->mg_initialized) {
ASSERT3U(mc->mc_groups, >, 0);
mc->mc_groups--;
}
if (mg->mg_initialized)
mg->mg_no_free_space = B_FALSE;
/*
* A metaslab group is considered allocatable if it has plenty
* of free space or is not heavily fragmented. We only take
* fragmentation into account if the metaslab group has a valid
* fragmentation metric (i.e. a value between 0 and 100).
*/
mg->mg_allocatable = (mg->mg_activation_count > 0 &&
mg->mg_free_capacity > zfs_mg_noalloc_threshold &&
(mg->mg_fragmentation == ZFS_FRAG_INVALID ||
mg->mg_fragmentation <= zfs_mg_fragmentation_threshold));
/*
* The mc_alloc_groups maintains a count of the number of
* groups in this metaslab class that are still above the
* zfs_mg_noalloc_threshold. This is used by the allocating
* threads to determine if they should avoid allocations to
* a given group. The allocator will avoid allocations to a group
* if that group has reached or is below the zfs_mg_noalloc_threshold
* and there are still other groups that are above the threshold.
* When a group transitions from allocatable to non-allocatable or
* vice versa we update the metaslab class to reflect that change.
* When the mc_alloc_groups value drops to 0 that means that all
* groups have reached the zfs_mg_noalloc_threshold making all groups
* eligible for allocations. This effectively means that all devices
* are balanced again.
*/
if (was_allocatable && !mg->mg_allocatable)
mc->mc_alloc_groups--;
else if (!was_allocatable && mg->mg_allocatable)
mc->mc_alloc_groups++;
mutex_exit(&mc->mc_lock);
mutex_exit(&mg->mg_lock);
}
int
metaslab_sort_by_flushed(const void *va, const void *vb)
{
const metaslab_t *a = va;
const metaslab_t *b = vb;
int cmp = TREE_CMP(a->ms_unflushed_txg, b->ms_unflushed_txg);
if (likely(cmp))
return (cmp);
uint64_t a_vdev_id = a->ms_group->mg_vd->vdev_id;
uint64_t b_vdev_id = b->ms_group->mg_vd->vdev_id;
cmp = TREE_CMP(a_vdev_id, b_vdev_id);
if (cmp)
return (cmp);
return (TREE_CMP(a->ms_id, b->ms_id));
}
metaslab_group_t *
metaslab_group_create(metaslab_class_t *mc, vdev_t *vd, int allocators)
{
metaslab_group_t *mg;
mg = kmem_zalloc(offsetof(metaslab_group_t,
mg_allocator[allocators]), KM_SLEEP);
mutex_init(&mg->mg_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&mg->mg_ms_disabled_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&mg->mg_ms_disabled_cv, NULL, CV_DEFAULT, NULL);
avl_create(&mg->mg_metaslab_tree, metaslab_compare,
sizeof (metaslab_t), offsetof(metaslab_t, ms_group_node));
mg->mg_vd = vd;
mg->mg_class = mc;
mg->mg_activation_count = 0;
mg->mg_initialized = B_FALSE;
mg->mg_no_free_space = B_TRUE;
mg->mg_allocators = allocators;
for (int i = 0; i < allocators; i++) {
metaslab_group_allocator_t *mga = &mg->mg_allocator[i];
zfs_refcount_create_tracked(&mga->mga_alloc_queue_depth);
}
return (mg);
}
void
metaslab_group_destroy(metaslab_group_t *mg)
{
ASSERT(mg->mg_prev == NULL);
ASSERT(mg->mg_next == NULL);
/*
* We may have gone below zero with the activation count
* either because we never activated in the first place or
* because we're done, and possibly removing the vdev.
*/
ASSERT(mg->mg_activation_count <= 0);
avl_destroy(&mg->mg_metaslab_tree);
mutex_destroy(&mg->mg_lock);
mutex_destroy(&mg->mg_ms_disabled_lock);
cv_destroy(&mg->mg_ms_disabled_cv);
for (int i = 0; i < mg->mg_allocators; i++) {
metaslab_group_allocator_t *mga = &mg->mg_allocator[i];
zfs_refcount_destroy(&mga->mga_alloc_queue_depth);
}
kmem_free(mg, offsetof(metaslab_group_t,
mg_allocator[mg->mg_allocators]));
}
void
metaslab_group_activate(metaslab_group_t *mg)
{
metaslab_class_t *mc = mg->mg_class;
spa_t *spa = mc->mc_spa;
metaslab_group_t *mgprev, *mgnext;
ASSERT3U(spa_config_held(spa, SCL_ALLOC, RW_WRITER), !=, 0);
ASSERT(mg->mg_prev == NULL);
ASSERT(mg->mg_next == NULL);
ASSERT(mg->mg_activation_count <= 0);
if (++mg->mg_activation_count <= 0)
return;
mg->mg_aliquot = metaslab_aliquot * MAX(1,
vdev_get_ndisks(mg->mg_vd) - vdev_get_nparity(mg->mg_vd));
metaslab_group_alloc_update(mg);
if ((mgprev = mc->mc_allocator[0].mca_rotor) == NULL) {
mg->mg_prev = mg;
mg->mg_next = mg;
} else {
mgnext = mgprev->mg_next;
mg->mg_prev = mgprev;
mg->mg_next = mgnext;
mgprev->mg_next = mg;
mgnext->mg_prev = mg;
}
for (int i = 0; i < spa->spa_alloc_count; i++) {
mc->mc_allocator[i].mca_rotor = mg;
mg = mg->mg_next;
}
}
/*
* Passivate a metaslab group and remove it from the allocation rotor.
* Callers must hold both the SCL_ALLOC and SCL_ZIO lock prior to passivating
* a metaslab group. This function will momentarily drop spa_config_locks
* that are lower than the SCL_ALLOC lock (see comment below).
*/
void
metaslab_group_passivate(metaslab_group_t *mg)
{
metaslab_class_t *mc = mg->mg_class;
spa_t *spa = mc->mc_spa;
metaslab_group_t *mgprev, *mgnext;
int locks = spa_config_held(spa, SCL_ALL, RW_WRITER);
ASSERT3U(spa_config_held(spa, SCL_ALLOC | SCL_ZIO, RW_WRITER), ==,
(SCL_ALLOC | SCL_ZIO));
if (--mg->mg_activation_count != 0) {
for (int i = 0; i < spa->spa_alloc_count; i++)
ASSERT(mc->mc_allocator[i].mca_rotor != mg);
ASSERT(mg->mg_prev == NULL);
ASSERT(mg->mg_next == NULL);
ASSERT(mg->mg_activation_count < 0);
return;
}
/*
* The spa_config_lock is an array of rwlocks, ordered as
* follows (from highest to lowest):
* SCL_CONFIG > SCL_STATE > SCL_L2ARC > SCL_ALLOC >
* SCL_ZIO > SCL_FREE > SCL_VDEV
* (For more information about the spa_config_lock see spa_misc.c)
* The higher the lock, the broader its coverage. When we passivate
* a metaslab group, we must hold both the SCL_ALLOC and the SCL_ZIO
* config locks. However, the metaslab group's taskq might be trying
* to preload metaslabs so we must drop the SCL_ZIO lock and any
* lower locks to allow the I/O to complete. At a minimum,
* we continue to hold the SCL_ALLOC lock, which prevents any future
* allocations from taking place and any changes to the vdev tree.
*/
spa_config_exit(spa, locks & ~(SCL_ZIO - 1), spa);
taskq_wait_outstanding(spa->spa_metaslab_taskq, 0);
spa_config_enter(spa, locks & ~(SCL_ZIO - 1), spa, RW_WRITER);
metaslab_group_alloc_update(mg);
for (int i = 0; i < mg->mg_allocators; i++) {
metaslab_group_allocator_t *mga = &mg->mg_allocator[i];
metaslab_t *msp = mga->mga_primary;
if (msp != NULL) {
mutex_enter(&msp->ms_lock);
metaslab_passivate(msp,
metaslab_weight_from_range_tree(msp));
mutex_exit(&msp->ms_lock);
}
msp = mga->mga_secondary;
if (msp != NULL) {
mutex_enter(&msp->ms_lock);
metaslab_passivate(msp,
metaslab_weight_from_range_tree(msp));
mutex_exit(&msp->ms_lock);
}
}
mgprev = mg->mg_prev;
mgnext = mg->mg_next;
if (mg == mgnext) {
mgnext = NULL;
} else {
mgprev->mg_next = mgnext;
mgnext->mg_prev = mgprev;
}
for (int i = 0; i < spa->spa_alloc_count; i++) {
if (mc->mc_allocator[i].mca_rotor == mg)
mc->mc_allocator[i].mca_rotor = mgnext;
}
mg->mg_prev = NULL;
mg->mg_next = NULL;
}
boolean_t
metaslab_group_initialized(metaslab_group_t *mg)
{
vdev_t *vd = mg->mg_vd;
vdev_stat_t *vs = &vd->vdev_stat;
return (vs->vs_space != 0 && mg->mg_activation_count > 0);
}
uint64_t
metaslab_group_get_space(metaslab_group_t *mg)
{
/*
* Note that the number of nodes in mg_metaslab_tree may be one less
* than vdev_ms_count, due to the embedded log metaslab.
*/
mutex_enter(&mg->mg_lock);
uint64_t ms_count = avl_numnodes(&mg->mg_metaslab_tree);
mutex_exit(&mg->mg_lock);
return ((1ULL << mg->mg_vd->vdev_ms_shift) * ms_count);
}
void
metaslab_group_histogram_verify(metaslab_group_t *mg)
{
uint64_t *mg_hist;
avl_tree_t *t = &mg->mg_metaslab_tree;
uint64_t ashift = mg->mg_vd->vdev_ashift;
if ((zfs_flags & ZFS_DEBUG_HISTOGRAM_VERIFY) == 0)
return;
mg_hist = kmem_zalloc(sizeof (uint64_t) * RANGE_TREE_HISTOGRAM_SIZE,
KM_SLEEP);
ASSERT3U(RANGE_TREE_HISTOGRAM_SIZE, >=,
SPACE_MAP_HISTOGRAM_SIZE + ashift);
mutex_enter(&mg->mg_lock);
for (metaslab_t *msp = avl_first(t);
msp != NULL; msp = AVL_NEXT(t, msp)) {
VERIFY3P(msp->ms_group, ==, mg);
/* skip if not active */
if (msp->ms_sm == NULL)
continue;
for (int i = 0; i < SPACE_MAP_HISTOGRAM_SIZE; i++) {
mg_hist[i + ashift] +=
msp->ms_sm->sm_phys->smp_histogram[i];
}
}
for (int i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i ++)
VERIFY3U(mg_hist[i], ==, mg->mg_histogram[i]);
mutex_exit(&mg->mg_lock);
kmem_free(mg_hist, sizeof (uint64_t) * RANGE_TREE_HISTOGRAM_SIZE);
}
static void
metaslab_group_histogram_add(metaslab_group_t *mg, metaslab_t *msp)
{
metaslab_class_t *mc = mg->mg_class;
uint64_t ashift = mg->mg_vd->vdev_ashift;
ASSERT(MUTEX_HELD(&msp->ms_lock));
if (msp->ms_sm == NULL)
return;
mutex_enter(&mg->mg_lock);
mutex_enter(&mc->mc_lock);
for (int i = 0; i < SPACE_MAP_HISTOGRAM_SIZE; i++) {
IMPLY(mg == mg->mg_vd->vdev_log_mg,
mc == spa_embedded_log_class(mg->mg_vd->vdev_spa));
mg->mg_histogram[i + ashift] +=
msp->ms_sm->sm_phys->smp_histogram[i];
mc->mc_histogram[i + ashift] +=
msp->ms_sm->sm_phys->smp_histogram[i];
}
mutex_exit(&mc->mc_lock);
mutex_exit(&mg->mg_lock);
}
void
metaslab_group_histogram_remove(metaslab_group_t *mg, metaslab_t *msp)
{
metaslab_class_t *mc = mg->mg_class;
uint64_t ashift = mg->mg_vd->vdev_ashift;
ASSERT(MUTEX_HELD(&msp->ms_lock));
if (msp->ms_sm == NULL)
return;
mutex_enter(&mg->mg_lock);
mutex_enter(&mc->mc_lock);
for (int i = 0; i < SPACE_MAP_HISTOGRAM_SIZE; i++) {
ASSERT3U(mg->mg_histogram[i + ashift], >=,
msp->ms_sm->sm_phys->smp_histogram[i]);
ASSERT3U(mc->mc_histogram[i + ashift], >=,
msp->ms_sm->sm_phys->smp_histogram[i]);
IMPLY(mg == mg->mg_vd->vdev_log_mg,
mc == spa_embedded_log_class(mg->mg_vd->vdev_spa));
mg->mg_histogram[i + ashift] -=
msp->ms_sm->sm_phys->smp_histogram[i];
mc->mc_histogram[i + ashift] -=
msp->ms_sm->sm_phys->smp_histogram[i];
}
mutex_exit(&mc->mc_lock);
mutex_exit(&mg->mg_lock);
}
static void
metaslab_group_add(metaslab_group_t *mg, metaslab_t *msp)
{
ASSERT(msp->ms_group == NULL);
mutex_enter(&mg->mg_lock);
msp->ms_group = mg;
msp->ms_weight = 0;
avl_add(&mg->mg_metaslab_tree, msp);
mutex_exit(&mg->mg_lock);
mutex_enter(&msp->ms_lock);
metaslab_group_histogram_add(mg, msp);
mutex_exit(&msp->ms_lock);
}
static void
metaslab_group_remove(metaslab_group_t *mg, metaslab_t *msp)
{
mutex_enter(&msp->ms_lock);
metaslab_group_histogram_remove(mg, msp);
mutex_exit(&msp->ms_lock);
mutex_enter(&mg->mg_lock);
ASSERT(msp->ms_group == mg);
avl_remove(&mg->mg_metaslab_tree, msp);
metaslab_class_t *mc = msp->ms_group->mg_class;
multilist_sublist_t *mls =
multilist_sublist_lock_obj(&mc->mc_metaslab_txg_list, msp);
if (multilist_link_active(&msp->ms_class_txg_node))
multilist_sublist_remove(mls, msp);
multilist_sublist_unlock(mls);
msp->ms_group = NULL;
mutex_exit(&mg->mg_lock);
}
static void
metaslab_group_sort_impl(metaslab_group_t *mg, metaslab_t *msp, uint64_t weight)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
ASSERT(MUTEX_HELD(&mg->mg_lock));
ASSERT(msp->ms_group == mg);
avl_remove(&mg->mg_metaslab_tree, msp);
msp->ms_weight = weight;
avl_add(&mg->mg_metaslab_tree, msp);
}
static void
metaslab_group_sort(metaslab_group_t *mg, metaslab_t *msp, uint64_t weight)
{
/*
* Although in principle the weight can be any value, in
* practice we do not use values in the range [1, 511].
*/
ASSERT(weight >= SPA_MINBLOCKSIZE || weight == 0);
ASSERT(MUTEX_HELD(&msp->ms_lock));
mutex_enter(&mg->mg_lock);
metaslab_group_sort_impl(mg, msp, weight);
mutex_exit(&mg->mg_lock);
}
/*
* Calculate the fragmentation for a given metaslab group. We can use
* a simple average here since all metaslabs within the group must have
* the same size. The return value will be a value between 0 and 100
* (inclusive), or ZFS_FRAG_INVALID if less than half of the metaslab in this
* group have a fragmentation metric.
*/
uint64_t
metaslab_group_fragmentation(metaslab_group_t *mg)
{
vdev_t *vd = mg->mg_vd;
uint64_t fragmentation = 0;
uint64_t valid_ms = 0;
for (int m = 0; m < vd->vdev_ms_count; m++) {
metaslab_t *msp = vd->vdev_ms[m];
if (msp->ms_fragmentation == ZFS_FRAG_INVALID)
continue;
if (msp->ms_group != mg)
continue;
valid_ms++;
fragmentation += msp->ms_fragmentation;
}
if (valid_ms <= mg->mg_vd->vdev_ms_count / 2)
return (ZFS_FRAG_INVALID);
fragmentation /= valid_ms;
ASSERT3U(fragmentation, <=, 100);
return (fragmentation);
}
/*
* Determine if a given metaslab group should skip allocations. A metaslab
* group should avoid allocations if its free capacity is less than the
* zfs_mg_noalloc_threshold or its fragmentation metric is greater than
* zfs_mg_fragmentation_threshold and there is at least one metaslab group
* that can still handle allocations. If the allocation throttle is enabled
* then we skip allocations to devices that have reached their maximum
* allocation queue depth unless the selected metaslab group is the only
* eligible group remaining.
*/
static boolean_t
metaslab_group_allocatable(metaslab_group_t *mg, metaslab_group_t *rotor,
int flags, uint64_t psize, int allocator, int d)
{
spa_t *spa = mg->mg_vd->vdev_spa;
metaslab_class_t *mc = mg->mg_class;
/*
* We can only consider skipping this metaslab group if it's
* in the normal metaslab class and there are other metaslab
* groups to select from. Otherwise, we always consider it eligible
* for allocations.
*/
if ((mc != spa_normal_class(spa) &&
mc != spa_special_class(spa) &&
mc != spa_dedup_class(spa)) ||
mc->mc_groups <= 1)
return (B_TRUE);
/*
* If the metaslab group's mg_allocatable flag is set (see comments
* in metaslab_group_alloc_update() for more information) and
* the allocation throttle is disabled then allow allocations to this
* device. However, if the allocation throttle is enabled then
* check if we have reached our allocation limit (mga_alloc_queue_depth)
* to determine if we should allow allocations to this metaslab group.
* If all metaslab groups are no longer considered allocatable
* (mc_alloc_groups == 0) or we're trying to allocate the smallest
* gang block size then we allow allocations on this metaslab group
* regardless of the mg_allocatable or throttle settings.
*/
if (mg->mg_allocatable) {
metaslab_group_allocator_t *mga = &mg->mg_allocator[allocator];
int64_t qdepth;
uint64_t qmax = mga->mga_cur_max_alloc_queue_depth;
if (!mc->mc_alloc_throttle_enabled)
return (B_TRUE);
/*
* If this metaslab group does not have any free space, then
* there is no point in looking further.
*/
if (mg->mg_no_free_space)
return (B_FALSE);
/*
* Some allocations (e.g., those coming from device removal
* where the * allocations are not even counted in the
* metaslab * allocation queues) are allowed to bypass
* the throttle.
*/
if (flags & METASLAB_DONT_THROTTLE)
return (B_TRUE);
/*
* Relax allocation throttling for ditto blocks. Due to
* random imbalances in allocation it tends to push copies
* to one vdev, that looks a bit better at the moment.
*/
qmax = qmax * (4 + d) / 4;
qdepth = zfs_refcount_count(&mga->mga_alloc_queue_depth);
/*
* If this metaslab group is below its qmax or it's
* the only allocatable metaslab group, then attempt
* to allocate from it.
*/
if (qdepth < qmax || mc->mc_alloc_groups == 1)
return (B_TRUE);
ASSERT3U(mc->mc_alloc_groups, >, 1);
/*
* Since this metaslab group is at or over its qmax, we
* need to determine if there are metaslab groups after this
* one that might be able to handle this allocation. This is
* racy since we can't hold the locks for all metaslab
* groups at the same time when we make this check.
*/
for (metaslab_group_t *mgp = mg->mg_next;
mgp != rotor; mgp = mgp->mg_next) {
metaslab_group_allocator_t *mgap =
&mgp->mg_allocator[allocator];
qmax = mgap->mga_cur_max_alloc_queue_depth;
qmax = qmax * (4 + d) / 4;
qdepth =
zfs_refcount_count(&mgap->mga_alloc_queue_depth);
/*
* If there is another metaslab group that
* might be able to handle the allocation, then
* we return false so that we skip this group.
*/
if (qdepth < qmax && !mgp->mg_no_free_space)
return (B_FALSE);
}
/*
* We didn't find another group to handle the allocation
* so we can't skip this metaslab group even though
* we are at or over our qmax.
*/
return (B_TRUE);
} else if (mc->mc_alloc_groups == 0 || psize == SPA_MINBLOCKSIZE) {
return (B_TRUE);
}
return (B_FALSE);
}
/*
* ==========================================================================
* Range tree callbacks
* ==========================================================================
*/
/*
* Comparison function for the private size-ordered tree using 32-bit
* ranges. Tree is sorted by size, larger sizes at the end of the tree.
*/
__attribute__((always_inline)) inline
static int
metaslab_rangesize32_compare(const void *x1, const void *x2)
{
const range_seg32_t *r1 = x1;
const range_seg32_t *r2 = x2;
uint64_t rs_size1 = r1->rs_end - r1->rs_start;
uint64_t rs_size2 = r2->rs_end - r2->rs_start;
int cmp = TREE_CMP(rs_size1, rs_size2);
return (cmp + !cmp * TREE_CMP(r1->rs_start, r2->rs_start));
}
/*
* Comparison function for the private size-ordered tree using 64-bit
* ranges. Tree is sorted by size, larger sizes at the end of the tree.
*/
__attribute__((always_inline)) inline
static int
metaslab_rangesize64_compare(const void *x1, const void *x2)
{
const range_seg64_t *r1 = x1;
const range_seg64_t *r2 = x2;
uint64_t rs_size1 = r1->rs_end - r1->rs_start;
uint64_t rs_size2 = r2->rs_end - r2->rs_start;
int cmp = TREE_CMP(rs_size1, rs_size2);
return (cmp + !cmp * TREE_CMP(r1->rs_start, r2->rs_start));
}
typedef struct metaslab_rt_arg {
zfs_btree_t *mra_bt;
uint32_t mra_floor_shift;
} metaslab_rt_arg_t;
struct mssa_arg {
range_tree_t *rt;
metaslab_rt_arg_t *mra;
};
static void
metaslab_size_sorted_add(void *arg, uint64_t start, uint64_t size)
{
struct mssa_arg *mssap = arg;
range_tree_t *rt = mssap->rt;
metaslab_rt_arg_t *mrap = mssap->mra;
range_seg_max_t seg = {0};
rs_set_start(&seg, rt, start);
rs_set_end(&seg, rt, start + size);
metaslab_rt_add(rt, &seg, mrap);
}
static void
metaslab_size_tree_full_load(range_tree_t *rt)
{
metaslab_rt_arg_t *mrap = rt->rt_arg;
METASLABSTAT_BUMP(metaslabstat_reload_tree);
ASSERT0(zfs_btree_numnodes(mrap->mra_bt));
mrap->mra_floor_shift = 0;
struct mssa_arg arg = {0};
arg.rt = rt;
arg.mra = mrap;
range_tree_walk(rt, metaslab_size_sorted_add, &arg);
}
ZFS_BTREE_FIND_IN_BUF_FUNC(metaslab_rt_find_rangesize32_in_buf,
range_seg32_t, metaslab_rangesize32_compare)
ZFS_BTREE_FIND_IN_BUF_FUNC(metaslab_rt_find_rangesize64_in_buf,
range_seg64_t, metaslab_rangesize64_compare)
/*
* Create any block allocator specific components. The current allocators
* rely on using both a size-ordered range_tree_t and an array of uint64_t's.
*/
static void
metaslab_rt_create(range_tree_t *rt, void *arg)
{
metaslab_rt_arg_t *mrap = arg;
zfs_btree_t *size_tree = mrap->mra_bt;
size_t size;
int (*compare) (const void *, const void *);
bt_find_in_buf_f bt_find;
switch (rt->rt_type) {
case RANGE_SEG32:
size = sizeof (range_seg32_t);
compare = metaslab_rangesize32_compare;
bt_find = metaslab_rt_find_rangesize32_in_buf;
break;
case RANGE_SEG64:
size = sizeof (range_seg64_t);
compare = metaslab_rangesize64_compare;
bt_find = metaslab_rt_find_rangesize64_in_buf;
break;
default:
panic("Invalid range seg type %d", rt->rt_type);
}
zfs_btree_create(size_tree, compare, bt_find, size);
mrap->mra_floor_shift = metaslab_by_size_min_shift;
}
static void
metaslab_rt_destroy(range_tree_t *rt, void *arg)
{
(void) rt;
metaslab_rt_arg_t *mrap = arg;
zfs_btree_t *size_tree = mrap->mra_bt;
zfs_btree_destroy(size_tree);
kmem_free(mrap, sizeof (*mrap));
}
static void
metaslab_rt_add(range_tree_t *rt, range_seg_t *rs, void *arg)
{
metaslab_rt_arg_t *mrap = arg;
zfs_btree_t *size_tree = mrap->mra_bt;
if (rs_get_end(rs, rt) - rs_get_start(rs, rt) <
(1ULL << mrap->mra_floor_shift))
return;
zfs_btree_add(size_tree, rs);
}
static void
metaslab_rt_remove(range_tree_t *rt, range_seg_t *rs, void *arg)
{
metaslab_rt_arg_t *mrap = arg;
zfs_btree_t *size_tree = mrap->mra_bt;
if (rs_get_end(rs, rt) - rs_get_start(rs, rt) < (1ULL <<
mrap->mra_floor_shift))
return;
zfs_btree_remove(size_tree, rs);
}
static void
metaslab_rt_vacate(range_tree_t *rt, void *arg)
{
metaslab_rt_arg_t *mrap = arg;
zfs_btree_t *size_tree = mrap->mra_bt;
zfs_btree_clear(size_tree);
zfs_btree_destroy(size_tree);
metaslab_rt_create(rt, arg);
}
static const range_tree_ops_t metaslab_rt_ops = {
.rtop_create = metaslab_rt_create,
.rtop_destroy = metaslab_rt_destroy,
.rtop_add = metaslab_rt_add,
.rtop_remove = metaslab_rt_remove,
.rtop_vacate = metaslab_rt_vacate
};
/*
* ==========================================================================
* Common allocator routines
* ==========================================================================
*/
/*
* Return the maximum contiguous segment within the metaslab.
*/
uint64_t
metaslab_largest_allocatable(metaslab_t *msp)
{
zfs_btree_t *t = &msp->ms_allocatable_by_size;
range_seg_t *rs;
if (t == NULL)
return (0);
if (zfs_btree_numnodes(t) == 0)
metaslab_size_tree_full_load(msp->ms_allocatable);
rs = zfs_btree_last(t, NULL);
if (rs == NULL)
return (0);
return (rs_get_end(rs, msp->ms_allocatable) - rs_get_start(rs,
msp->ms_allocatable));
}
/*
* Return the maximum contiguous segment within the unflushed frees of this
* metaslab.
*/
static uint64_t
metaslab_largest_unflushed_free(metaslab_t *msp)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
if (msp->ms_unflushed_frees == NULL)
return (0);
if (zfs_btree_numnodes(&msp->ms_unflushed_frees_by_size) == 0)
metaslab_size_tree_full_load(msp->ms_unflushed_frees);
range_seg_t *rs = zfs_btree_last(&msp->ms_unflushed_frees_by_size,
NULL);
if (rs == NULL)
return (0);
/*
* When a range is freed from the metaslab, that range is added to
* both the unflushed frees and the deferred frees. While the block
* will eventually be usable, if the metaslab were loaded the range
* would not be added to the ms_allocatable tree until TXG_DEFER_SIZE
* txgs had passed. As a result, when attempting to estimate an upper
* bound for the largest currently-usable free segment in the
* metaslab, we need to not consider any ranges currently in the defer
* trees. This algorithm approximates the largest available chunk in
* the largest range in the unflushed_frees tree by taking the first
* chunk. While this may be a poor estimate, it should only remain so
* briefly and should eventually self-correct as frees are no longer
* deferred. Similar logic applies to the ms_freed tree. See
* metaslab_load() for more details.
*
* There are two primary sources of inaccuracy in this estimate. Both
* are tolerated for performance reasons. The first source is that we
* only check the largest segment for overlaps. Smaller segments may
* have more favorable overlaps with the other trees, resulting in
* larger usable chunks. Second, we only look at the first chunk in
* the largest segment; there may be other usable chunks in the
* largest segment, but we ignore them.
*/
uint64_t rstart = rs_get_start(rs, msp->ms_unflushed_frees);
uint64_t rsize = rs_get_end(rs, msp->ms_unflushed_frees) - rstart;
for (int t = 0; t < TXG_DEFER_SIZE; t++) {
uint64_t start = 0;
uint64_t size = 0;
boolean_t found = range_tree_find_in(msp->ms_defer[t], rstart,
rsize, &start, &size);
if (found) {
if (rstart == start)
return (0);
rsize = start - rstart;
}
}
uint64_t start = 0;
uint64_t size = 0;
boolean_t found = range_tree_find_in(msp->ms_freed, rstart,
rsize, &start, &size);
if (found)
rsize = start - rstart;
return (rsize);
}
static range_seg_t *
metaslab_block_find(zfs_btree_t *t, range_tree_t *rt, uint64_t start,
uint64_t size, zfs_btree_index_t *where)
{
range_seg_t *rs;
range_seg_max_t rsearch;
rs_set_start(&rsearch, rt, start);
rs_set_end(&rsearch, rt, start + size);
rs = zfs_btree_find(t, &rsearch, where);
if (rs == NULL) {
rs = zfs_btree_next(t, where, where);
}
return (rs);
}
/*
* This is a helper function that can be used by the allocator to find a
* suitable block to allocate. This will search the specified B-tree looking
* for a block that matches the specified criteria.
*/
static uint64_t
metaslab_block_picker(range_tree_t *rt, uint64_t *cursor, uint64_t size,
uint64_t max_search)
{
if (*cursor == 0)
*cursor = rt->rt_start;
zfs_btree_t *bt = &rt->rt_root;
zfs_btree_index_t where;
range_seg_t *rs = metaslab_block_find(bt, rt, *cursor, size, &where);
uint64_t first_found;
int count_searched = 0;
if (rs != NULL)
first_found = rs_get_start(rs, rt);
while (rs != NULL && (rs_get_start(rs, rt) - first_found <=
max_search || count_searched < metaslab_min_search_count)) {
uint64_t offset = rs_get_start(rs, rt);
if (offset + size <= rs_get_end(rs, rt)) {
*cursor = offset + size;
return (offset);
}
rs = zfs_btree_next(bt, &where, &where);
count_searched++;
}
*cursor = 0;
return (-1ULL);
}
static uint64_t metaslab_df_alloc(metaslab_t *msp, uint64_t size);
static uint64_t metaslab_cf_alloc(metaslab_t *msp, uint64_t size);
static uint64_t metaslab_ndf_alloc(metaslab_t *msp, uint64_t size);
metaslab_ops_t *metaslab_allocator(spa_t *spa);
static metaslab_ops_t metaslab_allocators[] = {
{ "dynamic", metaslab_df_alloc },
{ "cursor", metaslab_cf_alloc },
{ "new-dynamic", metaslab_ndf_alloc },
};
static int
spa_find_allocator_byname(const char *val)
{
int a = ARRAY_SIZE(metaslab_allocators) - 1;
if (strcmp("new-dynamic", val) == 0)
return (-1); /* remove when ndf is working */
for (; a >= 0; a--) {
if (strcmp(val, metaslab_allocators[a].msop_name) == 0)
return (a);
}
return (-1);
}
void
spa_set_allocator(spa_t *spa, const char *allocator)
{
int a = spa_find_allocator_byname(allocator);
if (a < 0) a = 0;
spa->spa_active_allocator = a;
zfs_dbgmsg("spa allocator: %s\n", metaslab_allocators[a].msop_name);
}
int
spa_get_allocator(spa_t *spa)
{
return (spa->spa_active_allocator);
}
#if defined(_KERNEL)
int
param_set_active_allocator_common(const char *val)
{
char *p;
if (val == NULL)
return (SET_ERROR(EINVAL));
if ((p = strchr(val, '\n')) != NULL)
*p = '\0';
int a = spa_find_allocator_byname(val);
if (a < 0)
return (SET_ERROR(EINVAL));
zfs_active_allocator = metaslab_allocators[a].msop_name;
return (0);
}
#endif
metaslab_ops_t *
metaslab_allocator(spa_t *spa)
{
int allocator = spa_get_allocator(spa);
return (&metaslab_allocators[allocator]);
}
/*
* ==========================================================================
* Dynamic Fit (df) block allocator
*
* Search for a free chunk of at least this size, starting from the last
* offset (for this alignment of block) looking for up to
* metaslab_df_max_search bytes (16MB). If a large enough free chunk is not
* found within 16MB, then return a free chunk of exactly the requested size (or
* larger).
*
* If it seems like searching from the last offset will be unproductive, skip
* that and just return a free chunk of exactly the requested size (or larger).
* This is based on metaslab_df_alloc_threshold and metaslab_df_free_pct. This
* mechanism is probably not very useful and may be removed in the future.
*
* The behavior when not searching can be changed to return the largest free
* chunk, instead of a free chunk of exactly the requested size, by setting
* metaslab_df_use_largest_segment.
* ==========================================================================
*/
static uint64_t
metaslab_df_alloc(metaslab_t *msp, uint64_t size)
{
/*
* Find the largest power of 2 block size that evenly divides the
* requested size. This is used to try to allocate blocks with similar
* alignment from the same area of the metaslab (i.e. same cursor
* bucket) but it does not guarantee that other allocations sizes
* may exist in the same region.
*/
uint64_t align = size & -size;
uint64_t *cursor = &msp->ms_lbas[highbit64(align) - 1];
range_tree_t *rt = msp->ms_allocatable;
uint_t free_pct = range_tree_space(rt) * 100 / msp->ms_size;
uint64_t offset;
ASSERT(MUTEX_HELD(&msp->ms_lock));
/*
* If we're running low on space, find a segment based on size,
* rather than iterating based on offset.
*/
if (metaslab_largest_allocatable(msp) < metaslab_df_alloc_threshold ||
free_pct < metaslab_df_free_pct) {
offset = -1;
} else {
offset = metaslab_block_picker(rt,
cursor, size, metaslab_df_max_search);
}
if (offset == -1) {
range_seg_t *rs;
if (zfs_btree_numnodes(&msp->ms_allocatable_by_size) == 0)
metaslab_size_tree_full_load(msp->ms_allocatable);
if (metaslab_df_use_largest_segment) {
/* use largest free segment */
rs = zfs_btree_last(&msp->ms_allocatable_by_size, NULL);
} else {
zfs_btree_index_t where;
/* use segment of this size, or next largest */
rs = metaslab_block_find(&msp->ms_allocatable_by_size,
rt, msp->ms_start, size, &where);
}
if (rs != NULL && rs_get_start(rs, rt) + size <= rs_get_end(rs,
rt)) {
offset = rs_get_start(rs, rt);
*cursor = offset + size;
}
}
return (offset);
}
/*
* ==========================================================================
* Cursor fit block allocator -
* Select the largest region in the metaslab, set the cursor to the beginning
* of the range and the cursor_end to the end of the range. As allocations
* are made advance the cursor. Continue allocating from the cursor until
* the range is exhausted and then find a new range.
* ==========================================================================
*/
static uint64_t
metaslab_cf_alloc(metaslab_t *msp, uint64_t size)
{
range_tree_t *rt = msp->ms_allocatable;
zfs_btree_t *t = &msp->ms_allocatable_by_size;
uint64_t *cursor = &msp->ms_lbas[0];
uint64_t *cursor_end = &msp->ms_lbas[1];
uint64_t offset = 0;
ASSERT(MUTEX_HELD(&msp->ms_lock));
ASSERT3U(*cursor_end, >=, *cursor);
if ((*cursor + size) > *cursor_end) {
range_seg_t *rs;
if (zfs_btree_numnodes(t) == 0)
metaslab_size_tree_full_load(msp->ms_allocatable);
rs = zfs_btree_last(t, NULL);
if (rs == NULL || (rs_get_end(rs, rt) - rs_get_start(rs, rt)) <
size)
return (-1ULL);
*cursor = rs_get_start(rs, rt);
*cursor_end = rs_get_end(rs, rt);
}
offset = *cursor;
*cursor += size;
return (offset);
}
/*
* ==========================================================================
* New dynamic fit allocator -
* Select a region that is large enough to allocate 2^metaslab_ndf_clump_shift
* contiguous blocks. If no region is found then just use the largest segment
* that remains.
* ==========================================================================
*/
/*
* Determines desired number of contiguous blocks (2^metaslab_ndf_clump_shift)
* to request from the allocator.
*/
uint64_t metaslab_ndf_clump_shift = 4;
static uint64_t
metaslab_ndf_alloc(metaslab_t *msp, uint64_t size)
{
zfs_btree_t *t = &msp->ms_allocatable->rt_root;
range_tree_t *rt = msp->ms_allocatable;
zfs_btree_index_t where;
range_seg_t *rs;
range_seg_max_t rsearch;
uint64_t hbit = highbit64(size);
uint64_t *cursor = &msp->ms_lbas[hbit - 1];
uint64_t max_size = metaslab_largest_allocatable(msp);
ASSERT(MUTEX_HELD(&msp->ms_lock));
if (max_size < size)
return (-1ULL);
rs_set_start(&rsearch, rt, *cursor);
rs_set_end(&rsearch, rt, *cursor + size);
rs = zfs_btree_find(t, &rsearch, &where);
if (rs == NULL || (rs_get_end(rs, rt) - rs_get_start(rs, rt)) < size) {
t = &msp->ms_allocatable_by_size;
rs_set_start(&rsearch, rt, 0);
rs_set_end(&rsearch, rt, MIN(max_size, 1ULL << (hbit +
metaslab_ndf_clump_shift)));
rs = zfs_btree_find(t, &rsearch, &where);
if (rs == NULL)
rs = zfs_btree_next(t, &where, &where);
ASSERT(rs != NULL);
}
if ((rs_get_end(rs, rt) - rs_get_start(rs, rt)) >= size) {
*cursor = rs_get_start(rs, rt) + size;
return (rs_get_start(rs, rt));
}
return (-1ULL);
}
/*
* ==========================================================================
* Metaslabs
* ==========================================================================
*/
/*
* Wait for any in-progress metaslab loads to complete.
*/
static void
metaslab_load_wait(metaslab_t *msp)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
while (msp->ms_loading) {
ASSERT(!msp->ms_loaded);
cv_wait(&msp->ms_load_cv, &msp->ms_lock);
}
}
/*
* Wait for any in-progress flushing to complete.
*/
static void
metaslab_flush_wait(metaslab_t *msp)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
while (msp->ms_flushing)
cv_wait(&msp->ms_flush_cv, &msp->ms_lock);
}
static unsigned int
metaslab_idx_func(multilist_t *ml, void *arg)
{
metaslab_t *msp = arg;
/*
* ms_id values are allocated sequentially, so full 64bit
* division would be a waste of time, so limit it to 32 bits.
*/
return ((unsigned int)msp->ms_id % multilist_get_num_sublists(ml));
}
uint64_t
metaslab_allocated_space(metaslab_t *msp)
{
return (msp->ms_allocated_space);
}
/*
* Verify that the space accounting on disk matches the in-core range_trees.
*/
static void
metaslab_verify_space(metaslab_t *msp, uint64_t txg)
{
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
uint64_t allocating = 0;
uint64_t sm_free_space, msp_free_space;
ASSERT(MUTEX_HELD(&msp->ms_lock));
ASSERT(!msp->ms_condensing);
if ((zfs_flags & ZFS_DEBUG_METASLAB_VERIFY) == 0)
return;
/*
* We can only verify the metaslab space when we're called
* from syncing context with a loaded metaslab that has an
* allocated space map. Calling this in non-syncing context
* does not provide a consistent view of the metaslab since
* we're performing allocations in the future.
*/
if (txg != spa_syncing_txg(spa) || msp->ms_sm == NULL ||
!msp->ms_loaded)
return;
/*
* Even though the smp_alloc field can get negative,
* when it comes to a metaslab's space map, that should
* never be the case.
*/
ASSERT3S(space_map_allocated(msp->ms_sm), >=, 0);
ASSERT3U(space_map_allocated(msp->ms_sm), >=,
range_tree_space(msp->ms_unflushed_frees));
ASSERT3U(metaslab_allocated_space(msp), ==,
space_map_allocated(msp->ms_sm) +
range_tree_space(msp->ms_unflushed_allocs) -
range_tree_space(msp->ms_unflushed_frees));
sm_free_space = msp->ms_size - metaslab_allocated_space(msp);
/*
* Account for future allocations since we would have
* already deducted that space from the ms_allocatable.
*/
for (int t = 0; t < TXG_CONCURRENT_STATES; t++) {
allocating +=
range_tree_space(msp->ms_allocating[(txg + t) & TXG_MASK]);
}
ASSERT3U(allocating + msp->ms_allocated_this_txg, ==,
msp->ms_allocating_total);
ASSERT3U(msp->ms_deferspace, ==,
range_tree_space(msp->ms_defer[0]) +
range_tree_space(msp->ms_defer[1]));
msp_free_space = range_tree_space(msp->ms_allocatable) + allocating +
msp->ms_deferspace + range_tree_space(msp->ms_freed);
VERIFY3U(sm_free_space, ==, msp_free_space);
}
static void
metaslab_aux_histograms_clear(metaslab_t *msp)
{
/*
* Auxiliary histograms are only cleared when resetting them,
* which can only happen while the metaslab is loaded.
*/
ASSERT(msp->ms_loaded);
memset(msp->ms_synchist, 0, sizeof (msp->ms_synchist));
for (int t = 0; t < TXG_DEFER_SIZE; t++)
memset(msp->ms_deferhist[t], 0, sizeof (msp->ms_deferhist[t]));
}
static void
metaslab_aux_histogram_add(uint64_t *histogram, uint64_t shift,
range_tree_t *rt)
{
/*
* This is modeled after space_map_histogram_add(), so refer to that
* function for implementation details. We want this to work like
* the space map histogram, and not the range tree histogram, as we
* are essentially constructing a delta that will be later subtracted
* from the space map histogram.
*/
int idx = 0;
for (int i = shift; i < RANGE_TREE_HISTOGRAM_SIZE; i++) {
ASSERT3U(i, >=, idx + shift);
histogram[idx] += rt->rt_histogram[i] << (i - idx - shift);
if (idx < SPACE_MAP_HISTOGRAM_SIZE - 1) {
ASSERT3U(idx + shift, ==, i);
idx++;
ASSERT3U(idx, <, SPACE_MAP_HISTOGRAM_SIZE);
}
}
}
/*
* Called at every sync pass that the metaslab gets synced.
*
* The reason is that we want our auxiliary histograms to be updated
* wherever the metaslab's space map histogram is updated. This way
* we stay consistent on which parts of the metaslab space map's
* histogram are currently not available for allocations (e.g because
* they are in the defer, freed, and freeing trees).
*/
static void
metaslab_aux_histograms_update(metaslab_t *msp)
{
space_map_t *sm = msp->ms_sm;
ASSERT(sm != NULL);
/*
* This is similar to the metaslab's space map histogram updates
* that take place in metaslab_sync(). The only difference is that
* we only care about segments that haven't made it into the
* ms_allocatable tree yet.
*/
if (msp->ms_loaded) {
metaslab_aux_histograms_clear(msp);
metaslab_aux_histogram_add(msp->ms_synchist,
sm->sm_shift, msp->ms_freed);
for (int t = 0; t < TXG_DEFER_SIZE; t++) {
metaslab_aux_histogram_add(msp->ms_deferhist[t],
sm->sm_shift, msp->ms_defer[t]);
}
}
metaslab_aux_histogram_add(msp->ms_synchist,
sm->sm_shift, msp->ms_freeing);
}
/*
* Called every time we are done syncing (writing to) the metaslab,
* i.e. at the end of each sync pass.
* [see the comment in metaslab_impl.h for ms_synchist, ms_deferhist]
*/
static void
metaslab_aux_histograms_update_done(metaslab_t *msp, boolean_t defer_allowed)
{
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
space_map_t *sm = msp->ms_sm;
if (sm == NULL) {
/*
* We came here from metaslab_init() when creating/opening a
* pool, looking at a metaslab that hasn't had any allocations
* yet.
*/
return;
}
/*
* This is similar to the actions that we take for the ms_freed
* and ms_defer trees in metaslab_sync_done().
*/
uint64_t hist_index = spa_syncing_txg(spa) % TXG_DEFER_SIZE;
if (defer_allowed) {
memcpy(msp->ms_deferhist[hist_index], msp->ms_synchist,
sizeof (msp->ms_synchist));
} else {
memset(msp->ms_deferhist[hist_index], 0,
sizeof (msp->ms_deferhist[hist_index]));
}
memset(msp->ms_synchist, 0, sizeof (msp->ms_synchist));
}
/*
* Ensure that the metaslab's weight and fragmentation are consistent
* with the contents of the histogram (either the range tree's histogram
* or the space map's depending whether the metaslab is loaded).
*/
static void
metaslab_verify_weight_and_frag(metaslab_t *msp)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
if ((zfs_flags & ZFS_DEBUG_METASLAB_VERIFY) == 0)
return;
/*
* We can end up here from vdev_remove_complete(), in which case we
* cannot do these assertions because we hold spa config locks and
* thus we are not allowed to read from the DMU.
*
* We check if the metaslab group has been removed and if that's
* the case we return immediately as that would mean that we are
* here from the aforementioned code path.
*/
if (msp->ms_group == NULL)
return;
/*
* Devices being removed always return a weight of 0 and leave
* fragmentation and ms_max_size as is - there is nothing for
* us to verify here.
*/
vdev_t *vd = msp->ms_group->mg_vd;
if (vd->vdev_removing)
return;
/*
* If the metaslab is dirty it probably means that we've done
* some allocations or frees that have changed our histograms
* and thus the weight.
*/
for (int t = 0; t < TXG_SIZE; t++) {
if (txg_list_member(&vd->vdev_ms_list, msp, t))
return;
}
/*
* This verification checks that our in-memory state is consistent
* with what's on disk. If the pool is read-only then there aren't
* any changes and we just have the initially-loaded state.
*/
if (!spa_writeable(msp->ms_group->mg_vd->vdev_spa))
return;
/* some extra verification for in-core tree if you can */
if (msp->ms_loaded) {
range_tree_stat_verify(msp->ms_allocatable);
VERIFY(space_map_histogram_verify(msp->ms_sm,
msp->ms_allocatable));
}
uint64_t weight = msp->ms_weight;
uint64_t was_active = msp->ms_weight & METASLAB_ACTIVE_MASK;
boolean_t space_based = WEIGHT_IS_SPACEBASED(msp->ms_weight);
uint64_t frag = msp->ms_fragmentation;
uint64_t max_segsize = msp->ms_max_size;
msp->ms_weight = 0;
msp->ms_fragmentation = 0;
/*
* This function is used for verification purposes and thus should
* not introduce any side-effects/mutations on the system's state.
*
* Regardless of whether metaslab_weight() thinks this metaslab
* should be active or not, we want to ensure that the actual weight
* (and therefore the value of ms_weight) would be the same if it
* was to be recalculated at this point.
*
* In addition we set the nodirty flag so metaslab_weight() does
* not dirty the metaslab for future TXGs (e.g. when trying to
* force condensing to upgrade the metaslab spacemaps).
*/
msp->ms_weight = metaslab_weight(msp, B_TRUE) | was_active;
VERIFY3U(max_segsize, ==, msp->ms_max_size);
/*
* If the weight type changed then there is no point in doing
* verification. Revert fields to their original values.
*/
if ((space_based && !WEIGHT_IS_SPACEBASED(msp->ms_weight)) ||
(!space_based && WEIGHT_IS_SPACEBASED(msp->ms_weight))) {
msp->ms_fragmentation = frag;
msp->ms_weight = weight;
return;
}
VERIFY3U(msp->ms_fragmentation, ==, frag);
VERIFY3U(msp->ms_weight, ==, weight);
}
/*
* If we're over the zfs_metaslab_mem_limit, select the loaded metaslab from
* this class that was used longest ago, and attempt to unload it. We don't
* want to spend too much time in this loop to prevent performance
* degradation, and we expect that most of the time this operation will
* succeed. Between that and the normal unloading processing during txg sync,
* we expect this to keep the metaslab memory usage under control.
*/
static void
metaslab_potentially_evict(metaslab_class_t *mc)
{
#ifdef _KERNEL
uint64_t allmem = arc_all_memory();
uint64_t inuse = spl_kmem_cache_inuse(zfs_btree_leaf_cache);
uint64_t size = spl_kmem_cache_entry_size(zfs_btree_leaf_cache);
uint_t tries = 0;
for (; allmem * zfs_metaslab_mem_limit / 100 < inuse * size &&
tries < multilist_get_num_sublists(&mc->mc_metaslab_txg_list) * 2;
tries++) {
unsigned int idx = multilist_get_random_index(
&mc->mc_metaslab_txg_list);
multilist_sublist_t *mls =
multilist_sublist_lock(&mc->mc_metaslab_txg_list, idx);
metaslab_t *msp = multilist_sublist_head(mls);
multilist_sublist_unlock(mls);
while (msp != NULL && allmem * zfs_metaslab_mem_limit / 100 <
inuse * size) {
VERIFY3P(mls, ==, multilist_sublist_lock(
&mc->mc_metaslab_txg_list, idx));
ASSERT3U(idx, ==,
metaslab_idx_func(&mc->mc_metaslab_txg_list, msp));
if (!multilist_link_active(&msp->ms_class_txg_node)) {
multilist_sublist_unlock(mls);
break;
}
metaslab_t *next_msp = multilist_sublist_next(mls, msp);
multilist_sublist_unlock(mls);
/*
* If the metaslab is currently loading there are two
* cases. If it's the metaslab we're evicting, we
* can't continue on or we'll panic when we attempt to
* recursively lock the mutex. If it's another
* metaslab that's loading, it can be safely skipped,
* since we know it's very new and therefore not a
* good eviction candidate. We check later once the
* lock is held that the metaslab is fully loaded
* before actually unloading it.
*/
if (msp->ms_loading) {
msp = next_msp;
inuse =
spl_kmem_cache_inuse(zfs_btree_leaf_cache);
continue;
}
/*
* We can't unload metaslabs with no spacemap because
* they're not ready to be unloaded yet. We can't
* unload metaslabs with outstanding allocations
* because doing so could cause the metaslab's weight
* to decrease while it's unloaded, which violates an
* invariant that we use to prevent unnecessary
* loading. We also don't unload metaslabs that are
* currently active because they are high-weight
* metaslabs that are likely to be used in the near
* future.
*/
mutex_enter(&msp->ms_lock);
if (msp->ms_allocator == -1 && msp->ms_sm != NULL &&
msp->ms_allocating_total == 0) {
metaslab_unload(msp);
}
mutex_exit(&msp->ms_lock);
msp = next_msp;
inuse = spl_kmem_cache_inuse(zfs_btree_leaf_cache);
}
}
#else
(void) mc, (void) zfs_metaslab_mem_limit;
#endif
}
static int
metaslab_load_impl(metaslab_t *msp)
{
int error = 0;
ASSERT(MUTEX_HELD(&msp->ms_lock));
ASSERT(msp->ms_loading);
ASSERT(!msp->ms_condensing);
/*
* We temporarily drop the lock to unblock other operations while we
* are reading the space map. Therefore, metaslab_sync() and
* metaslab_sync_done() can run at the same time as we do.
*
* If we are using the log space maps, metaslab_sync() can't write to
* the metaslab's space map while we are loading as we only write to
* it when we are flushing the metaslab, and that can't happen while
* we are loading it.
*
* If we are not using log space maps though, metaslab_sync() can
* append to the space map while we are loading. Therefore we load
* only entries that existed when we started the load. Additionally,
* metaslab_sync_done() has to wait for the load to complete because
* there are potential races like metaslab_load() loading parts of the
* space map that are currently being appended by metaslab_sync(). If
* we didn't, the ms_allocatable would have entries that
* metaslab_sync_done() would try to re-add later.
*
* That's why before dropping the lock we remember the synced length
* of the metaslab and read up to that point of the space map,
* ignoring entries appended by metaslab_sync() that happen after we
* drop the lock.
*/
uint64_t length = msp->ms_synced_length;
mutex_exit(&msp->ms_lock);
hrtime_t load_start = gethrtime();
metaslab_rt_arg_t *mrap;
if (msp->ms_allocatable->rt_arg == NULL) {
mrap = kmem_zalloc(sizeof (*mrap), KM_SLEEP);
} else {
mrap = msp->ms_allocatable->rt_arg;
msp->ms_allocatable->rt_ops = NULL;
msp->ms_allocatable->rt_arg = NULL;
}
mrap->mra_bt = &msp->ms_allocatable_by_size;
mrap->mra_floor_shift = metaslab_by_size_min_shift;
if (msp->ms_sm != NULL) {
error = space_map_load_length(msp->ms_sm, msp->ms_allocatable,
SM_FREE, length);
/* Now, populate the size-sorted tree. */
metaslab_rt_create(msp->ms_allocatable, mrap);
msp->ms_allocatable->rt_ops = &metaslab_rt_ops;
msp->ms_allocatable->rt_arg = mrap;
struct mssa_arg arg = {0};
arg.rt = msp->ms_allocatable;
arg.mra = mrap;
range_tree_walk(msp->ms_allocatable, metaslab_size_sorted_add,
&arg);
} else {
/*
* Add the size-sorted tree first, since we don't need to load
* the metaslab from the spacemap.
*/
metaslab_rt_create(msp->ms_allocatable, mrap);
msp->ms_allocatable->rt_ops = &metaslab_rt_ops;
msp->ms_allocatable->rt_arg = mrap;
/*
* The space map has not been allocated yet, so treat
* all the space in the metaslab as free and add it to the
* ms_allocatable tree.
*/
range_tree_add(msp->ms_allocatable,
msp->ms_start, msp->ms_size);
if (msp->ms_new) {
/*
* If the ms_sm doesn't exist, this means that this
* metaslab hasn't gone through metaslab_sync() and
* thus has never been dirtied. So we shouldn't
* expect any unflushed allocs or frees from previous
* TXGs.
*/
ASSERT(range_tree_is_empty(msp->ms_unflushed_allocs));
ASSERT(range_tree_is_empty(msp->ms_unflushed_frees));
}
}
/*
* We need to grab the ms_sync_lock to prevent metaslab_sync() from
* changing the ms_sm (or log_sm) and the metaslab's range trees
* while we are about to use them and populate the ms_allocatable.
* The ms_lock is insufficient for this because metaslab_sync() doesn't
* hold the ms_lock while writing the ms_checkpointing tree to disk.
*/
mutex_enter(&msp->ms_sync_lock);
mutex_enter(&msp->ms_lock);
ASSERT(!msp->ms_condensing);
ASSERT(!msp->ms_flushing);
if (error != 0) {
mutex_exit(&msp->ms_sync_lock);
return (error);
}
ASSERT3P(msp->ms_group, !=, NULL);
msp->ms_loaded = B_TRUE;
/*
* Apply all the unflushed changes to ms_allocatable right
* away so any manipulations we do below have a clear view
* of what is allocated and what is free.
*/
range_tree_walk(msp->ms_unflushed_allocs,
range_tree_remove, msp->ms_allocatable);
range_tree_walk(msp->ms_unflushed_frees,
range_tree_add, msp->ms_allocatable);
ASSERT3P(msp->ms_group, !=, NULL);
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
if (spa_syncing_log_sm(spa) != NULL) {
ASSERT(spa_feature_is_enabled(spa,
SPA_FEATURE_LOG_SPACEMAP));
/*
* If we use a log space map we add all the segments
* that are in ms_unflushed_frees so they are available
* for allocation.
*
* ms_allocatable needs to contain all free segments
* that are ready for allocations (thus not segments
* from ms_freeing, ms_freed, and the ms_defer trees).
* But if we grab the lock in this code path at a sync
* pass later that 1, then it also contains the
* segments of ms_freed (they were added to it earlier
* in this path through ms_unflushed_frees). So we
* need to remove all the segments that exist in
* ms_freed from ms_allocatable as they will be added
* later in metaslab_sync_done().
*
* When there's no log space map, the ms_allocatable
* correctly doesn't contain any segments that exist
* in ms_freed [see ms_synced_length].
*/
range_tree_walk(msp->ms_freed,
range_tree_remove, msp->ms_allocatable);
}
/*
* If we are not using the log space map, ms_allocatable
* contains the segments that exist in the ms_defer trees
* [see ms_synced_length]. Thus we need to remove them
* from ms_allocatable as they will be added again in
* metaslab_sync_done().
*
* If we are using the log space map, ms_allocatable still
* contains the segments that exist in the ms_defer trees.
* Not because it read them through the ms_sm though. But
* because these segments are part of ms_unflushed_frees
* whose segments we add to ms_allocatable earlier in this
* code path.
*/
for (int t = 0; t < TXG_DEFER_SIZE; t++) {
range_tree_walk(msp->ms_defer[t],
range_tree_remove, msp->ms_allocatable);
}
/*
* Call metaslab_recalculate_weight_and_sort() now that the
* metaslab is loaded so we get the metaslab's real weight.
*
* Unless this metaslab was created with older software and
* has not yet been converted to use segment-based weight, we
* expect the new weight to be better or equal to the weight
* that the metaslab had while it was not loaded. This is
* because the old weight does not take into account the
* consolidation of adjacent segments between TXGs. [see
* comment for ms_synchist and ms_deferhist[] for more info]
*/
uint64_t weight = msp->ms_weight;
uint64_t max_size = msp->ms_max_size;
metaslab_recalculate_weight_and_sort(msp);
if (!WEIGHT_IS_SPACEBASED(weight))
ASSERT3U(weight, <=, msp->ms_weight);
msp->ms_max_size = metaslab_largest_allocatable(msp);
ASSERT3U(max_size, <=, msp->ms_max_size);
hrtime_t load_end = gethrtime();
msp->ms_load_time = load_end;
zfs_dbgmsg("metaslab_load: txg %llu, spa %s, vdev_id %llu, "
"ms_id %llu, smp_length %llu, "
"unflushed_allocs %llu, unflushed_frees %llu, "
"freed %llu, defer %llu + %llu, unloaded time %llu ms, "
"loading_time %lld ms, ms_max_size %llu, "
"max size error %lld, "
"old_weight %llx, new_weight %llx",
(u_longlong_t)spa_syncing_txg(spa), spa_name(spa),
(u_longlong_t)msp->ms_group->mg_vd->vdev_id,
(u_longlong_t)msp->ms_id,
(u_longlong_t)space_map_length(msp->ms_sm),
(u_longlong_t)range_tree_space(msp->ms_unflushed_allocs),
(u_longlong_t)range_tree_space(msp->ms_unflushed_frees),
(u_longlong_t)range_tree_space(msp->ms_freed),
(u_longlong_t)range_tree_space(msp->ms_defer[0]),
(u_longlong_t)range_tree_space(msp->ms_defer[1]),
(longlong_t)((load_start - msp->ms_unload_time) / 1000000),
(longlong_t)((load_end - load_start) / 1000000),
(u_longlong_t)msp->ms_max_size,
(u_longlong_t)msp->ms_max_size - max_size,
(u_longlong_t)weight, (u_longlong_t)msp->ms_weight);
metaslab_verify_space(msp, spa_syncing_txg(spa));
mutex_exit(&msp->ms_sync_lock);
return (0);
}
int
metaslab_load(metaslab_t *msp)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
/*
* There may be another thread loading the same metaslab, if that's
* the case just wait until the other thread is done and return.
*/
metaslab_load_wait(msp);
if (msp->ms_loaded)
return (0);
VERIFY(!msp->ms_loading);
ASSERT(!msp->ms_condensing);
/*
* We set the loading flag BEFORE potentially dropping the lock to
* wait for an ongoing flush (see ms_flushing below). This way other
* threads know that there is already a thread that is loading this
* metaslab.
*/
msp->ms_loading = B_TRUE;
/*
* Wait for any in-progress flushing to finish as we drop the ms_lock
* both here (during space_map_load()) and in metaslab_flush() (when
* we flush our changes to the ms_sm).
*/
if (msp->ms_flushing)
metaslab_flush_wait(msp);
/*
* In the possibility that we were waiting for the metaslab to be
* flushed (where we temporarily dropped the ms_lock), ensure that
* no one else loaded the metaslab somehow.
*/
ASSERT(!msp->ms_loaded);
/*
* If we're loading a metaslab in the normal class, consider evicting
* another one to keep our memory usage under the limit defined by the
* zfs_metaslab_mem_limit tunable.
*/
if (spa_normal_class(msp->ms_group->mg_class->mc_spa) ==
msp->ms_group->mg_class) {
metaslab_potentially_evict(msp->ms_group->mg_class);
}
int error = metaslab_load_impl(msp);
ASSERT(MUTEX_HELD(&msp->ms_lock));
msp->ms_loading = B_FALSE;
cv_broadcast(&msp->ms_load_cv);
return (error);
}
void
metaslab_unload(metaslab_t *msp)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
/*
* This can happen if a metaslab is selected for eviction (in
* metaslab_potentially_evict) and then unloaded during spa_sync (via
* metaslab_class_evict_old).
*/
if (!msp->ms_loaded)
return;
range_tree_vacate(msp->ms_allocatable, NULL, NULL);
msp->ms_loaded = B_FALSE;
msp->ms_unload_time = gethrtime();
msp->ms_activation_weight = 0;
msp->ms_weight &= ~METASLAB_ACTIVE_MASK;
if (msp->ms_group != NULL) {
metaslab_class_t *mc = msp->ms_group->mg_class;
multilist_sublist_t *mls =
multilist_sublist_lock_obj(&mc->mc_metaslab_txg_list, msp);
if (multilist_link_active(&msp->ms_class_txg_node))
multilist_sublist_remove(mls, msp);
multilist_sublist_unlock(mls);
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
zfs_dbgmsg("metaslab_unload: txg %llu, spa %s, vdev_id %llu, "
"ms_id %llu, weight %llx, "
"selected txg %llu (%llu ms ago), alloc_txg %llu, "
"loaded %llu ms ago, max_size %llu",
(u_longlong_t)spa_syncing_txg(spa), spa_name(spa),
(u_longlong_t)msp->ms_group->mg_vd->vdev_id,
(u_longlong_t)msp->ms_id,
(u_longlong_t)msp->ms_weight,
(u_longlong_t)msp->ms_selected_txg,
(u_longlong_t)(msp->ms_unload_time -
msp->ms_selected_time) / 1000 / 1000,
(u_longlong_t)msp->ms_alloc_txg,
(u_longlong_t)(msp->ms_unload_time -
msp->ms_load_time) / 1000 / 1000,
(u_longlong_t)msp->ms_max_size);
}
/*
* We explicitly recalculate the metaslab's weight based on its space
* map (as it is now not loaded). We want unload metaslabs to always
* have their weights calculated from the space map histograms, while
* loaded ones have it calculated from their in-core range tree
* [see metaslab_load()]. This way, the weight reflects the information
* available in-core, whether it is loaded or not.
*
* If ms_group == NULL means that we came here from metaslab_fini(),
* at which point it doesn't make sense for us to do the recalculation
* and the sorting.
*/
if (msp->ms_group != NULL)
metaslab_recalculate_weight_and_sort(msp);
}
/*
* We want to optimize the memory use of the per-metaslab range
* trees. To do this, we store the segments in the range trees in
* units of sectors, zero-indexing from the start of the metaslab. If
* the vdev_ms_shift - the vdev_ashift is less than 32, we can store
* the ranges using two uint32_ts, rather than two uint64_ts.
*/
range_seg_type_t
metaslab_calculate_range_tree_type(vdev_t *vdev, metaslab_t *msp,
uint64_t *start, uint64_t *shift)
{
if (vdev->vdev_ms_shift - vdev->vdev_ashift < 32 &&
!zfs_metaslab_force_large_segs) {
*shift = vdev->vdev_ashift;
*start = msp->ms_start;
return (RANGE_SEG32);
} else {
*shift = 0;
*start = 0;
return (RANGE_SEG64);
}
}
void
metaslab_set_selected_txg(metaslab_t *msp, uint64_t txg)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
metaslab_class_t *mc = msp->ms_group->mg_class;
multilist_sublist_t *mls =
multilist_sublist_lock_obj(&mc->mc_metaslab_txg_list, msp);
if (multilist_link_active(&msp->ms_class_txg_node))
multilist_sublist_remove(mls, msp);
msp->ms_selected_txg = txg;
msp->ms_selected_time = gethrtime();
multilist_sublist_insert_tail(mls, msp);
multilist_sublist_unlock(mls);
}
void
metaslab_space_update(vdev_t *vd, metaslab_class_t *mc, int64_t alloc_delta,
int64_t defer_delta, int64_t space_delta)
{
vdev_space_update(vd, alloc_delta, defer_delta, space_delta);
ASSERT3P(vd->vdev_spa->spa_root_vdev, ==, vd->vdev_parent);
ASSERT(vd->vdev_ms_count != 0);
metaslab_class_space_update(mc, alloc_delta, defer_delta, space_delta,
vdev_deflated_space(vd, space_delta));
}
int
metaslab_init(metaslab_group_t *mg, uint64_t id, uint64_t object,
uint64_t txg, metaslab_t **msp)
{
vdev_t *vd = mg->mg_vd;
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa->spa_meta_objset;
metaslab_t *ms;
int error;
ms = kmem_zalloc(sizeof (metaslab_t), KM_SLEEP);
mutex_init(&ms->ms_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&ms->ms_sync_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&ms->ms_load_cv, NULL, CV_DEFAULT, NULL);
cv_init(&ms->ms_flush_cv, NULL, CV_DEFAULT, NULL);
multilist_link_init(&ms->ms_class_txg_node);
ms->ms_id = id;
ms->ms_start = id << vd->vdev_ms_shift;
ms->ms_size = 1ULL << vd->vdev_ms_shift;
ms->ms_allocator = -1;
ms->ms_new = B_TRUE;
vdev_ops_t *ops = vd->vdev_ops;
if (ops->vdev_op_metaslab_init != NULL)
ops->vdev_op_metaslab_init(vd, &ms->ms_start, &ms->ms_size);
/*
* We only open space map objects that already exist. All others
* will be opened when we finally allocate an object for it. For
* readonly pools there is no need to open the space map object.
*
* Note:
* When called from vdev_expand(), we can't call into the DMU as
* we are holding the spa_config_lock as a writer and we would
* deadlock [see relevant comment in vdev_metaslab_init()]. in
* that case, the object parameter is zero though, so we won't
* call into the DMU.
*/
if (object != 0 && !(spa->spa_mode == SPA_MODE_READ &&
!spa->spa_read_spacemaps)) {
error = space_map_open(&ms->ms_sm, mos, object, ms->ms_start,
ms->ms_size, vd->vdev_ashift);
if (error != 0) {
kmem_free(ms, sizeof (metaslab_t));
return (error);
}
ASSERT(ms->ms_sm != NULL);
ms->ms_allocated_space = space_map_allocated(ms->ms_sm);
}
uint64_t shift, start;
range_seg_type_t type =
metaslab_calculate_range_tree_type(vd, ms, &start, &shift);
ms->ms_allocatable = range_tree_create(NULL, type, NULL, start, shift);
for (int t = 0; t < TXG_SIZE; t++) {
ms->ms_allocating[t] = range_tree_create(NULL, type,
NULL, start, shift);
}
ms->ms_freeing = range_tree_create(NULL, type, NULL, start, shift);
ms->ms_freed = range_tree_create(NULL, type, NULL, start, shift);
for (int t = 0; t < TXG_DEFER_SIZE; t++) {
ms->ms_defer[t] = range_tree_create(NULL, type, NULL,
start, shift);
}
ms->ms_checkpointing =
range_tree_create(NULL, type, NULL, start, shift);
ms->ms_unflushed_allocs =
range_tree_create(NULL, type, NULL, start, shift);
metaslab_rt_arg_t *mrap = kmem_zalloc(sizeof (*mrap), KM_SLEEP);
mrap->mra_bt = &ms->ms_unflushed_frees_by_size;
mrap->mra_floor_shift = metaslab_by_size_min_shift;
ms->ms_unflushed_frees = range_tree_create(&metaslab_rt_ops,
type, mrap, start, shift);
ms->ms_trim = range_tree_create(NULL, type, NULL, start, shift);
metaslab_group_add(mg, ms);
metaslab_set_fragmentation(ms, B_FALSE);
/*
* If we're opening an existing pool (txg == 0) or creating
* a new one (txg == TXG_INITIAL), all space is available now.
* If we're adding space to an existing pool, the new space
* does not become available until after this txg has synced.
* The metaslab's weight will also be initialized when we sync
* out this txg. This ensures that we don't attempt to allocate
* from it before we have initialized it completely.
*/
if (txg <= TXG_INITIAL) {
metaslab_sync_done(ms, 0);
metaslab_space_update(vd, mg->mg_class,
metaslab_allocated_space(ms), 0, 0);
}
if (txg != 0) {
vdev_dirty(vd, 0, NULL, txg);
vdev_dirty(vd, VDD_METASLAB, ms, txg);
}
*msp = ms;
return (0);
}
static void
metaslab_fini_flush_data(metaslab_t *msp)
{
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
if (metaslab_unflushed_txg(msp) == 0) {
ASSERT3P(avl_find(&spa->spa_metaslabs_by_flushed, msp, NULL),
==, NULL);
return;
}
ASSERT(spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP));
mutex_enter(&spa->spa_flushed_ms_lock);
avl_remove(&spa->spa_metaslabs_by_flushed, msp);
mutex_exit(&spa->spa_flushed_ms_lock);
spa_log_sm_decrement_mscount(spa, metaslab_unflushed_txg(msp));
spa_log_summary_decrement_mscount(spa, metaslab_unflushed_txg(msp),
metaslab_unflushed_dirty(msp));
}
uint64_t
metaslab_unflushed_changes_memused(metaslab_t *ms)
{
return ((range_tree_numsegs(ms->ms_unflushed_allocs) +
range_tree_numsegs(ms->ms_unflushed_frees)) *
ms->ms_unflushed_allocs->rt_root.bt_elem_size);
}
void
metaslab_fini(metaslab_t *msp)
{
metaslab_group_t *mg = msp->ms_group;
vdev_t *vd = mg->mg_vd;
spa_t *spa = vd->vdev_spa;
metaslab_fini_flush_data(msp);
metaslab_group_remove(mg, msp);
mutex_enter(&msp->ms_lock);
VERIFY(msp->ms_group == NULL);
/*
* If this metaslab hasn't been through metaslab_sync_done() yet its
* space hasn't been accounted for in its vdev and doesn't need to be
* subtracted.
*/
if (!msp->ms_new) {
metaslab_space_update(vd, mg->mg_class,
-metaslab_allocated_space(msp), 0, -msp->ms_size);
}
space_map_close(msp->ms_sm);
msp->ms_sm = NULL;
metaslab_unload(msp);
range_tree_destroy(msp->ms_allocatable);
range_tree_destroy(msp->ms_freeing);
range_tree_destroy(msp->ms_freed);
ASSERT3U(spa->spa_unflushed_stats.sus_memused, >=,
metaslab_unflushed_changes_memused(msp));
spa->spa_unflushed_stats.sus_memused -=
metaslab_unflushed_changes_memused(msp);
range_tree_vacate(msp->ms_unflushed_allocs, NULL, NULL);
range_tree_destroy(msp->ms_unflushed_allocs);
range_tree_destroy(msp->ms_checkpointing);
range_tree_vacate(msp->ms_unflushed_frees, NULL, NULL);
range_tree_destroy(msp->ms_unflushed_frees);
for (int t = 0; t < TXG_SIZE; t++) {
range_tree_destroy(msp->ms_allocating[t]);
}
for (int t = 0; t < TXG_DEFER_SIZE; t++) {
range_tree_destroy(msp->ms_defer[t]);
}
ASSERT0(msp->ms_deferspace);
for (int t = 0; t < TXG_SIZE; t++)
ASSERT(!txg_list_member(&vd->vdev_ms_list, msp, t));
range_tree_vacate(msp->ms_trim, NULL, NULL);
range_tree_destroy(msp->ms_trim);
mutex_exit(&msp->ms_lock);
cv_destroy(&msp->ms_load_cv);
cv_destroy(&msp->ms_flush_cv);
mutex_destroy(&msp->ms_lock);
mutex_destroy(&msp->ms_sync_lock);
ASSERT3U(msp->ms_allocator, ==, -1);
kmem_free(msp, sizeof (metaslab_t));
}
#define FRAGMENTATION_TABLE_SIZE 17
/*
* This table defines a segment size based fragmentation metric that will
* allow each metaslab to derive its own fragmentation value. This is done
* by calculating the space in each bucket of the spacemap histogram and
* multiplying that by the fragmentation metric in this table. Doing
* this for all buckets and dividing it by the total amount of free
* space in this metaslab (i.e. the total free space in all buckets) gives
* us the fragmentation metric. This means that a high fragmentation metric
* equates to most of the free space being comprised of small segments.
* Conversely, if the metric is low, then most of the free space is in
* large segments. A 10% change in fragmentation equates to approximately
* double the number of segments.
*
* This table defines 0% fragmented space using 16MB segments. Testing has
* shown that segments that are greater than or equal to 16MB do not suffer
* from drastic performance problems. Using this value, we derive the rest
* of the table. Since the fragmentation value is never stored on disk, it
* is possible to change these calculations in the future.
*/
static const int zfs_frag_table[FRAGMENTATION_TABLE_SIZE] = {
100, /* 512B */
100, /* 1K */
98, /* 2K */
95, /* 4K */
90, /* 8K */
80, /* 16K */
70, /* 32K */
60, /* 64K */
50, /* 128K */
40, /* 256K */
30, /* 512K */
20, /* 1M */
15, /* 2M */
10, /* 4M */
5, /* 8M */
0 /* 16M */
};
/*
* Calculate the metaslab's fragmentation metric and set ms_fragmentation.
* Setting this value to ZFS_FRAG_INVALID means that the metaslab has not
* been upgraded and does not support this metric. Otherwise, the return
* value should be in the range [0, 100].
*/
static void
metaslab_set_fragmentation(metaslab_t *msp, boolean_t nodirty)
{
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
uint64_t fragmentation = 0;
uint64_t total = 0;
boolean_t feature_enabled = spa_feature_is_enabled(spa,
SPA_FEATURE_SPACEMAP_HISTOGRAM);
if (!feature_enabled) {
msp->ms_fragmentation = ZFS_FRAG_INVALID;
return;
}
/*
* A null space map means that the entire metaslab is free
* and thus is not fragmented.
*/
if (msp->ms_sm == NULL) {
msp->ms_fragmentation = 0;
return;
}
/*
* If this metaslab's space map has not been upgraded, flag it
* so that we upgrade next time we encounter it.
*/
if (msp->ms_sm->sm_dbuf->db_size != sizeof (space_map_phys_t)) {
uint64_t txg = spa_syncing_txg(spa);
vdev_t *vd = msp->ms_group->mg_vd;
/*
* If we've reached the final dirty txg, then we must
* be shutting down the pool. We don't want to dirty
* any data past this point so skip setting the condense
* flag. We can retry this action the next time the pool
* is imported. We also skip marking this metaslab for
* condensing if the caller has explicitly set nodirty.
*/
if (!nodirty &&
spa_writeable(spa) && txg < spa_final_dirty_txg(spa)) {
msp->ms_condense_wanted = B_TRUE;
vdev_dirty(vd, VDD_METASLAB, msp, txg + 1);
zfs_dbgmsg("txg %llu, requesting force condense: "
"ms_id %llu, vdev_id %llu", (u_longlong_t)txg,
(u_longlong_t)msp->ms_id,
(u_longlong_t)vd->vdev_id);
}
msp->ms_fragmentation = ZFS_FRAG_INVALID;
return;
}
for (int i = 0; i < SPACE_MAP_HISTOGRAM_SIZE; i++) {
uint64_t space = 0;
uint8_t shift = msp->ms_sm->sm_shift;
int idx = MIN(shift - SPA_MINBLOCKSHIFT + i,
FRAGMENTATION_TABLE_SIZE - 1);
if (msp->ms_sm->sm_phys->smp_histogram[i] == 0)
continue;
space = msp->ms_sm->sm_phys->smp_histogram[i] << (i + shift);
total += space;
ASSERT3U(idx, <, FRAGMENTATION_TABLE_SIZE);
fragmentation += space * zfs_frag_table[idx];
}
if (total > 0)
fragmentation /= total;
ASSERT3U(fragmentation, <=, 100);
msp->ms_fragmentation = fragmentation;
}
/*
* Compute a weight -- a selection preference value -- for the given metaslab.
* This is based on the amount of free space, the level of fragmentation,
* the LBA range, and whether the metaslab is loaded.
*/
static uint64_t
metaslab_space_weight(metaslab_t *msp)
{
metaslab_group_t *mg = msp->ms_group;
vdev_t *vd = mg->mg_vd;
uint64_t weight, space;
ASSERT(MUTEX_HELD(&msp->ms_lock));
/*
* The baseline weight is the metaslab's free space.
*/
space = msp->ms_size - metaslab_allocated_space(msp);
if (metaslab_fragmentation_factor_enabled &&
msp->ms_fragmentation != ZFS_FRAG_INVALID) {
/*
* Use the fragmentation information to inversely scale
* down the baseline weight. We need to ensure that we
* don't exclude this metaslab completely when it's 100%
* fragmented. To avoid this we reduce the fragmented value
* by 1.
*/
space = (space * (100 - (msp->ms_fragmentation - 1))) / 100;
/*
* If space < SPA_MINBLOCKSIZE, then we will not allocate from
* this metaslab again. The fragmentation metric may have
* decreased the space to something smaller than
* SPA_MINBLOCKSIZE, so reset the space to SPA_MINBLOCKSIZE
* so that we can consume any remaining space.
*/
if (space > 0 && space < SPA_MINBLOCKSIZE)
space = SPA_MINBLOCKSIZE;
}
weight = space;
/*
* Modern disks have uniform bit density and constant angular velocity.
* Therefore, the outer recording zones are faster (higher bandwidth)
* than the inner zones by the ratio of outer to inner track diameter,
* which is typically around 2:1. We account for this by assigning
* higher weight to lower metaslabs (multiplier ranging from 2x to 1x).
* In effect, this means that we'll select the metaslab with the most
* free bandwidth rather than simply the one with the most free space.
*/
if (!vd->vdev_nonrot && metaslab_lba_weighting_enabled) {
weight = 2 * weight - (msp->ms_id * weight) / vd->vdev_ms_count;
ASSERT(weight >= space && weight <= 2 * space);
}
/*
* If this metaslab is one we're actively using, adjust its
* weight to make it preferable to any inactive metaslab so
* we'll polish it off. If the fragmentation on this metaslab
* has exceed our threshold, then don't mark it active.
*/
if (msp->ms_loaded && msp->ms_fragmentation != ZFS_FRAG_INVALID &&
msp->ms_fragmentation <= zfs_metaslab_fragmentation_threshold) {
weight |= (msp->ms_weight & METASLAB_ACTIVE_MASK);
}
WEIGHT_SET_SPACEBASED(weight);
return (weight);
}
/*
* Return the weight of the specified metaslab, according to the segment-based
* weighting algorithm. The metaslab must be loaded. This function can
* be called within a sync pass since it relies only on the metaslab's
* range tree which is always accurate when the metaslab is loaded.
*/
static uint64_t
metaslab_weight_from_range_tree(metaslab_t *msp)
{
uint64_t weight = 0;
uint32_t segments = 0;
ASSERT(msp->ms_loaded);
for (int i = RANGE_TREE_HISTOGRAM_SIZE - 1; i >= SPA_MINBLOCKSHIFT;
i--) {
uint8_t shift = msp->ms_group->mg_vd->vdev_ashift;
int max_idx = SPACE_MAP_HISTOGRAM_SIZE + shift - 1;
segments <<= 1;
segments += msp->ms_allocatable->rt_histogram[i];
/*
* The range tree provides more precision than the space map
* and must be downgraded so that all values fit within the
* space map's histogram. This allows us to compare loaded
* vs. unloaded metaslabs to determine which metaslab is
* considered "best".
*/
if (i > max_idx)
continue;
if (segments != 0) {
WEIGHT_SET_COUNT(weight, segments);
WEIGHT_SET_INDEX(weight, i);
WEIGHT_SET_ACTIVE(weight, 0);
break;
}
}
return (weight);
}
/*
* Calculate the weight based on the on-disk histogram. Should be applied
* only to unloaded metaslabs (i.e no incoming allocations) in-order to
* give results consistent with the on-disk state
*/
static uint64_t
metaslab_weight_from_spacemap(metaslab_t *msp)
{
space_map_t *sm = msp->ms_sm;
ASSERT(!msp->ms_loaded);
ASSERT(sm != NULL);
ASSERT3U(space_map_object(sm), !=, 0);
ASSERT3U(sm->sm_dbuf->db_size, ==, sizeof (space_map_phys_t));
/*
* Create a joint histogram from all the segments that have made
* it to the metaslab's space map histogram, that are not yet
* available for allocation because they are still in the freeing
* pipeline (e.g. freeing, freed, and defer trees). Then subtract
* these segments from the space map's histogram to get a more
* accurate weight.
*/
uint64_t deferspace_histogram[SPACE_MAP_HISTOGRAM_SIZE] = {0};
for (int i = 0; i < SPACE_MAP_HISTOGRAM_SIZE; i++)
deferspace_histogram[i] += msp->ms_synchist[i];
for (int t = 0; t < TXG_DEFER_SIZE; t++) {
for (int i = 0; i < SPACE_MAP_HISTOGRAM_SIZE; i++) {
deferspace_histogram[i] += msp->ms_deferhist[t][i];
}
}
uint64_t weight = 0;
for (int i = SPACE_MAP_HISTOGRAM_SIZE - 1; i >= 0; i--) {
ASSERT3U(sm->sm_phys->smp_histogram[i], >=,
deferspace_histogram[i]);
uint64_t count =
sm->sm_phys->smp_histogram[i] - deferspace_histogram[i];
if (count != 0) {
WEIGHT_SET_COUNT(weight, count);
WEIGHT_SET_INDEX(weight, i + sm->sm_shift);
WEIGHT_SET_ACTIVE(weight, 0);
break;
}
}
return (weight);
}
/*
* Compute a segment-based weight for the specified metaslab. The weight
* is determined by highest bucket in the histogram. The information
* for the highest bucket is encoded into the weight value.
*/
static uint64_t
metaslab_segment_weight(metaslab_t *msp)
{
metaslab_group_t *mg = msp->ms_group;
uint64_t weight = 0;
uint8_t shift = mg->mg_vd->vdev_ashift;
ASSERT(MUTEX_HELD(&msp->ms_lock));
/*
* The metaslab is completely free.
*/
if (metaslab_allocated_space(msp) == 0) {
int idx = highbit64(msp->ms_size) - 1;
int max_idx = SPACE_MAP_HISTOGRAM_SIZE + shift - 1;
if (idx < max_idx) {
WEIGHT_SET_COUNT(weight, 1ULL);
WEIGHT_SET_INDEX(weight, idx);
} else {
WEIGHT_SET_COUNT(weight, 1ULL << (idx - max_idx));
WEIGHT_SET_INDEX(weight, max_idx);
}
WEIGHT_SET_ACTIVE(weight, 0);
ASSERT(!WEIGHT_IS_SPACEBASED(weight));
return (weight);
}
ASSERT3U(msp->ms_sm->sm_dbuf->db_size, ==, sizeof (space_map_phys_t));
/*
* If the metaslab is fully allocated then just make the weight 0.
*/
if (metaslab_allocated_space(msp) == msp->ms_size)
return (0);
/*
* If the metaslab is already loaded, then use the range tree to
* determine the weight. Otherwise, we rely on the space map information
* to generate the weight.
*/
if (msp->ms_loaded) {
weight = metaslab_weight_from_range_tree(msp);
} else {
weight = metaslab_weight_from_spacemap(msp);
}
/*
* If the metaslab was active the last time we calculated its weight
* then keep it active. We want to consume the entire region that
* is associated with this weight.
*/
if (msp->ms_activation_weight != 0 && weight != 0)
WEIGHT_SET_ACTIVE(weight, WEIGHT_GET_ACTIVE(msp->ms_weight));
return (weight);
}
/*
* Determine if we should attempt to allocate from this metaslab. If the
* metaslab is loaded, then we can determine if the desired allocation
* can be satisfied by looking at the size of the maximum free segment
* on that metaslab. Otherwise, we make our decision based on the metaslab's
* weight. For segment-based weighting we can determine the maximum
* allocation based on the index encoded in its value. For space-based
* weights we rely on the entire weight (excluding the weight-type bit).
*/
static boolean_t
metaslab_should_allocate(metaslab_t *msp, uint64_t asize, boolean_t try_hard)
{
/*
* This case will usually but not always get caught by the checks below;
* metaslabs can be loaded by various means, including the trim and
* initialize code. Once that happens, without this check they are
* allocatable even before they finish their first txg sync.
*/
if (unlikely(msp->ms_new))
return (B_FALSE);
/*
* If the metaslab is loaded, ms_max_size is definitive and we can use
* the fast check. If it's not, the ms_max_size is a lower bound (once
* set), and we should use the fast check as long as we're not in
* try_hard and it's been less than zfs_metaslab_max_size_cache_sec
* seconds since the metaslab was unloaded.
*/
if (msp->ms_loaded ||
(msp->ms_max_size != 0 && !try_hard && gethrtime() <
msp->ms_unload_time + SEC2NSEC(zfs_metaslab_max_size_cache_sec)))
return (msp->ms_max_size >= asize);
boolean_t should_allocate;
if (!WEIGHT_IS_SPACEBASED(msp->ms_weight)) {
/*
* The metaslab segment weight indicates segments in the
* range [2^i, 2^(i+1)), where i is the index in the weight.
* Since the asize might be in the middle of the range, we
* should attempt the allocation if asize < 2^(i+1).
*/
should_allocate = (asize <
1ULL << (WEIGHT_GET_INDEX(msp->ms_weight) + 1));
} else {
should_allocate = (asize <=
(msp->ms_weight & ~METASLAB_WEIGHT_TYPE));
}
return (should_allocate);
}
static uint64_t
metaslab_weight(metaslab_t *msp, boolean_t nodirty)
{
vdev_t *vd = msp->ms_group->mg_vd;
spa_t *spa = vd->vdev_spa;
uint64_t weight;
ASSERT(MUTEX_HELD(&msp->ms_lock));
metaslab_set_fragmentation(msp, nodirty);
/*
* Update the maximum size. If the metaslab is loaded, this will
* ensure that we get an accurate maximum size if newly freed space
* has been added back into the free tree. If the metaslab is
* unloaded, we check if there's a larger free segment in the
* unflushed frees. This is a lower bound on the largest allocatable
* segment size. Coalescing of adjacent entries may reveal larger
* allocatable segments, but we aren't aware of those until loading
* the space map into a range tree.
*/
if (msp->ms_loaded) {
msp->ms_max_size = metaslab_largest_allocatable(msp);
} else {
msp->ms_max_size = MAX(msp->ms_max_size,
metaslab_largest_unflushed_free(msp));
}
/*
* Segment-based weighting requires space map histogram support.
*/
if (zfs_metaslab_segment_weight_enabled &&
spa_feature_is_enabled(spa, SPA_FEATURE_SPACEMAP_HISTOGRAM) &&
(msp->ms_sm == NULL || msp->ms_sm->sm_dbuf->db_size ==
sizeof (space_map_phys_t))) {
weight = metaslab_segment_weight(msp);
} else {
weight = metaslab_space_weight(msp);
}
return (weight);
}
void
metaslab_recalculate_weight_and_sort(metaslab_t *msp)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
/* note: we preserve the mask (e.g. indication of primary, etc..) */
uint64_t was_active = msp->ms_weight & METASLAB_ACTIVE_MASK;
metaslab_group_sort(msp->ms_group, msp,
metaslab_weight(msp, B_FALSE) | was_active);
}
static int
metaslab_activate_allocator(metaslab_group_t *mg, metaslab_t *msp,
int allocator, uint64_t activation_weight)
{
metaslab_group_allocator_t *mga = &mg->mg_allocator[allocator];
ASSERT(MUTEX_HELD(&msp->ms_lock));
/*
* If we're activating for the claim code, we don't want to actually
* set the metaslab up for a specific allocator.
*/
if (activation_weight == METASLAB_WEIGHT_CLAIM) {
ASSERT0(msp->ms_activation_weight);
msp->ms_activation_weight = msp->ms_weight;
metaslab_group_sort(mg, msp, msp->ms_weight |
activation_weight);
return (0);
}
metaslab_t **mspp = (activation_weight == METASLAB_WEIGHT_PRIMARY ?
&mga->mga_primary : &mga->mga_secondary);
mutex_enter(&mg->mg_lock);
if (*mspp != NULL) {
mutex_exit(&mg->mg_lock);
return (EEXIST);
}
*mspp = msp;
ASSERT3S(msp->ms_allocator, ==, -1);
msp->ms_allocator = allocator;
msp->ms_primary = (activation_weight == METASLAB_WEIGHT_PRIMARY);
ASSERT0(msp->ms_activation_weight);
msp->ms_activation_weight = msp->ms_weight;
metaslab_group_sort_impl(mg, msp,
msp->ms_weight | activation_weight);
mutex_exit(&mg->mg_lock);
return (0);
}
static int
metaslab_activate(metaslab_t *msp, int allocator, uint64_t activation_weight)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
/*
* The current metaslab is already activated for us so there
* is nothing to do. Already activated though, doesn't mean
* that this metaslab is activated for our allocator nor our
* requested activation weight. The metaslab could have started
* as an active one for our allocator but changed allocators
* while we were waiting to grab its ms_lock or we stole it
* [see find_valid_metaslab()]. This means that there is a
* possibility of passivating a metaslab of another allocator
* or from a different activation mask, from this thread.
*/
if ((msp->ms_weight & METASLAB_ACTIVE_MASK) != 0) {
ASSERT(msp->ms_loaded);
return (0);
}
int error = metaslab_load(msp);
if (error != 0) {
metaslab_group_sort(msp->ms_group, msp, 0);
return (error);
}
/*
* When entering metaslab_load() we may have dropped the
* ms_lock because we were loading this metaslab, or we
* were waiting for another thread to load it for us. In
* that scenario, we recheck the weight of the metaslab
* to see if it was activated by another thread.
*
* If the metaslab was activated for another allocator or
* it was activated with a different activation weight (e.g.
* we wanted to make it a primary but it was activated as
* secondary) we return error (EBUSY).
*
* If the metaslab was activated for the same allocator
* and requested activation mask, skip activating it.
*/
if ((msp->ms_weight & METASLAB_ACTIVE_MASK) != 0) {
if (msp->ms_allocator != allocator)
return (EBUSY);
if ((msp->ms_weight & activation_weight) == 0)
return (SET_ERROR(EBUSY));
EQUIV((activation_weight == METASLAB_WEIGHT_PRIMARY),
msp->ms_primary);
return (0);
}
/*
* If the metaslab has literally 0 space, it will have weight 0. In
* that case, don't bother activating it. This can happen if the
* metaslab had space during find_valid_metaslab, but another thread
* loaded it and used all that space while we were waiting to grab the
* lock.
*/
if (msp->ms_weight == 0) {
ASSERT0(range_tree_space(msp->ms_allocatable));
return (SET_ERROR(ENOSPC));
}
if ((error = metaslab_activate_allocator(msp->ms_group, msp,
allocator, activation_weight)) != 0) {
return (error);
}
ASSERT(msp->ms_loaded);
ASSERT(msp->ms_weight & METASLAB_ACTIVE_MASK);
return (0);
}
static void
metaslab_passivate_allocator(metaslab_group_t *mg, metaslab_t *msp,
uint64_t weight)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
ASSERT(msp->ms_loaded);
if (msp->ms_weight & METASLAB_WEIGHT_CLAIM) {
metaslab_group_sort(mg, msp, weight);
return;
}
mutex_enter(&mg->mg_lock);
ASSERT3P(msp->ms_group, ==, mg);
ASSERT3S(0, <=, msp->ms_allocator);
ASSERT3U(msp->ms_allocator, <, mg->mg_allocators);
metaslab_group_allocator_t *mga = &mg->mg_allocator[msp->ms_allocator];
if (msp->ms_primary) {
ASSERT3P(mga->mga_primary, ==, msp);
ASSERT(msp->ms_weight & METASLAB_WEIGHT_PRIMARY);
mga->mga_primary = NULL;
} else {
ASSERT3P(mga->mga_secondary, ==, msp);
ASSERT(msp->ms_weight & METASLAB_WEIGHT_SECONDARY);
mga->mga_secondary = NULL;
}
msp->ms_allocator = -1;
metaslab_group_sort_impl(mg, msp, weight);
mutex_exit(&mg->mg_lock);
}
static void
metaslab_passivate(metaslab_t *msp, uint64_t weight)
{
uint64_t size __maybe_unused = weight & ~METASLAB_WEIGHT_TYPE;
/*
* If size < SPA_MINBLOCKSIZE, then we will not allocate from
* this metaslab again. In that case, it had better be empty,
* or we would be leaving space on the table.
*/
ASSERT(!WEIGHT_IS_SPACEBASED(msp->ms_weight) ||
size >= SPA_MINBLOCKSIZE ||
range_tree_space(msp->ms_allocatable) == 0);
ASSERT0(weight & METASLAB_ACTIVE_MASK);
ASSERT(msp->ms_activation_weight != 0);
msp->ms_activation_weight = 0;
metaslab_passivate_allocator(msp->ms_group, msp, weight);
ASSERT0(msp->ms_weight & METASLAB_ACTIVE_MASK);
}
/*
* Segment-based metaslabs are activated once and remain active until
* we either fail an allocation attempt (similar to space-based metaslabs)
* or have exhausted the free space in zfs_metaslab_switch_threshold
* buckets since the metaslab was activated. This function checks to see
* if we've exhausted the zfs_metaslab_switch_threshold buckets in the
* metaslab and passivates it proactively. This will allow us to select a
* metaslab with a larger contiguous region, if any, remaining within this
* metaslab group. If we're in sync pass > 1, then we continue using this
* metaslab so that we don't dirty more block and cause more sync passes.
*/
static void
metaslab_segment_may_passivate(metaslab_t *msp)
{
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
if (WEIGHT_IS_SPACEBASED(msp->ms_weight) || spa_sync_pass(spa) > 1)
return;
/*
* Since we are in the middle of a sync pass, the most accurate
* information that is accessible to us is the in-core range tree
* histogram; calculate the new weight based on that information.
*/
uint64_t weight = metaslab_weight_from_range_tree(msp);
int activation_idx = WEIGHT_GET_INDEX(msp->ms_activation_weight);
int current_idx = WEIGHT_GET_INDEX(weight);
if (current_idx <= activation_idx - zfs_metaslab_switch_threshold)
metaslab_passivate(msp, weight);
}
static void
metaslab_preload(void *arg)
{
metaslab_t *msp = arg;
metaslab_class_t *mc = msp->ms_group->mg_class;
spa_t *spa = mc->mc_spa;
fstrans_cookie_t cookie = spl_fstrans_mark();
ASSERT(!MUTEX_HELD(&msp->ms_group->mg_lock));
mutex_enter(&msp->ms_lock);
(void) metaslab_load(msp);
metaslab_set_selected_txg(msp, spa_syncing_txg(spa));
mutex_exit(&msp->ms_lock);
spl_fstrans_unmark(cookie);
}
static void
metaslab_group_preload(metaslab_group_t *mg)
{
spa_t *spa = mg->mg_vd->vdev_spa;
metaslab_t *msp;
avl_tree_t *t = &mg->mg_metaslab_tree;
int m = 0;
if (spa_shutting_down(spa) || !metaslab_preload_enabled)
return;
mutex_enter(&mg->mg_lock);
/*
* Load the next potential metaslabs
*/
for (msp = avl_first(t); msp != NULL; msp = AVL_NEXT(t, msp)) {
ASSERT3P(msp->ms_group, ==, mg);
/*
* We preload only the maximum number of metaslabs specified
* by metaslab_preload_limit. If a metaslab is being forced
* to condense then we preload it too. This will ensure
* that force condensing happens in the next txg.
*/
if (++m > metaslab_preload_limit && !msp->ms_condense_wanted) {
continue;
}
VERIFY(taskq_dispatch(spa->spa_metaslab_taskq, metaslab_preload,
msp, TQ_SLEEP | (m <= mg->mg_allocators ? TQ_FRONT : 0))
!= TASKQID_INVALID);
}
mutex_exit(&mg->mg_lock);
}
/*
* Determine if the space map's on-disk footprint is past our tolerance for
* inefficiency. We would like to use the following criteria to make our
* decision:
*
* 1. Do not condense if the size of the space map object would dramatically
* increase as a result of writing out the free space range tree.
*
* 2. Condense if the on on-disk space map representation is at least
* zfs_condense_pct/100 times the size of the optimal representation
* (i.e. zfs_condense_pct = 110 and in-core = 1MB, optimal = 1.1MB).
*
* 3. Do not condense if the on-disk size of the space map does not actually
* decrease.
*
* Unfortunately, we cannot compute the on-disk size of the space map in this
* context because we cannot accurately compute the effects of compression, etc.
* Instead, we apply the heuristic described in the block comment for
* zfs_metaslab_condense_block_threshold - we only condense if the space used
* is greater than a threshold number of blocks.
*/
static boolean_t
metaslab_should_condense(metaslab_t *msp)
{
space_map_t *sm = msp->ms_sm;
vdev_t *vd = msp->ms_group->mg_vd;
uint64_t vdev_blocksize = 1ULL << vd->vdev_ashift;
ASSERT(MUTEX_HELD(&msp->ms_lock));
ASSERT(msp->ms_loaded);
ASSERT(sm != NULL);
ASSERT3U(spa_sync_pass(vd->vdev_spa), ==, 1);
/*
* We always condense metaslabs that are empty and metaslabs for
* which a condense request has been made.
*/
if (range_tree_numsegs(msp->ms_allocatable) == 0 ||
msp->ms_condense_wanted)
return (B_TRUE);
uint64_t record_size = MAX(sm->sm_blksz, vdev_blocksize);
uint64_t object_size = space_map_length(sm);
uint64_t optimal_size = space_map_estimate_optimal_size(sm,
msp->ms_allocatable, SM_NO_VDEVID);
return (object_size >= (optimal_size * zfs_condense_pct / 100) &&
object_size > zfs_metaslab_condense_block_threshold * record_size);
}
/*
* Condense the on-disk space map representation to its minimized form.
* The minimized form consists of a small number of allocations followed
* by the entries of the free range tree (ms_allocatable). The condensed
* spacemap contains all the entries of previous TXGs (including those in
* the pool-wide log spacemaps; thus this is effectively a superset of
* metaslab_flush()), but this TXG's entries still need to be written.
*/
static void
metaslab_condense(metaslab_t *msp, dmu_tx_t *tx)
{
range_tree_t *condense_tree;
space_map_t *sm = msp->ms_sm;
uint64_t txg = dmu_tx_get_txg(tx);
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
ASSERT(MUTEX_HELD(&msp->ms_lock));
ASSERT(msp->ms_loaded);
ASSERT(msp->ms_sm != NULL);
/*
* In order to condense the space map, we need to change it so it
* only describes which segments are currently allocated and free.
*
* All the current free space resides in the ms_allocatable, all
* the ms_defer trees, and all the ms_allocating trees. We ignore
* ms_freed because it is empty because we're in sync pass 1. We
* ignore ms_freeing because these changes are not yet reflected
* in the spacemap (they will be written later this txg).
*
* So to truncate the space map to represent all the entries of
* previous TXGs we do the following:
*
* 1] We create a range tree (condense tree) that is 100% empty.
* 2] We add to it all segments found in the ms_defer trees
* as those segments are marked as free in the original space
* map. We do the same with the ms_allocating trees for the same
* reason. Adding these segments should be a relatively
* inexpensive operation since we expect these trees to have a
* small number of nodes.
* 3] We vacate any unflushed allocs, since they are not frees we
* need to add to the condense tree. Then we vacate any
* unflushed frees as they should already be part of ms_allocatable.
* 4] At this point, we would ideally like to add all segments
* in the ms_allocatable tree from the condense tree. This way
* we would write all the entries of the condense tree as the
* condensed space map, which would only contain freed
* segments with everything else assumed to be allocated.
*
* Doing so can be prohibitively expensive as ms_allocatable can
* be large, and therefore computationally expensive to add to
* the condense_tree. Instead we first sync out an entry marking
* everything as allocated, then the condense_tree and then the
* ms_allocatable, in the condensed space map. While this is not
* optimal, it is typically close to optimal and more importantly
* much cheaper to compute.
*
* 5] Finally, as both of the unflushed trees were written to our
* new and condensed metaslab space map, we basically flushed
* all the unflushed changes to disk, thus we call
* metaslab_flush_update().
*/
ASSERT3U(spa_sync_pass(spa), ==, 1);
ASSERT(range_tree_is_empty(msp->ms_freed)); /* since it is pass 1 */
zfs_dbgmsg("condensing: txg %llu, msp[%llu] %px, vdev id %llu, "
"spa %s, smp size %llu, segments %llu, forcing condense=%s",
(u_longlong_t)txg, (u_longlong_t)msp->ms_id, msp,
(u_longlong_t)msp->ms_group->mg_vd->vdev_id,
spa->spa_name, (u_longlong_t)space_map_length(msp->ms_sm),
(u_longlong_t)range_tree_numsegs(msp->ms_allocatable),
msp->ms_condense_wanted ? "TRUE" : "FALSE");
msp->ms_condense_wanted = B_FALSE;
range_seg_type_t type;
uint64_t shift, start;
type = metaslab_calculate_range_tree_type(msp->ms_group->mg_vd, msp,
&start, &shift);
condense_tree = range_tree_create(NULL, type, NULL, start, shift);
for (int t = 0; t < TXG_DEFER_SIZE; t++) {
range_tree_walk(msp->ms_defer[t],
range_tree_add, condense_tree);
}
for (int t = 0; t < TXG_CONCURRENT_STATES; t++) {
range_tree_walk(msp->ms_allocating[(txg + t) & TXG_MASK],
range_tree_add, condense_tree);
}
ASSERT3U(spa->spa_unflushed_stats.sus_memused, >=,
metaslab_unflushed_changes_memused(msp));
spa->spa_unflushed_stats.sus_memused -=
metaslab_unflushed_changes_memused(msp);
range_tree_vacate(msp->ms_unflushed_allocs, NULL, NULL);
range_tree_vacate(msp->ms_unflushed_frees, NULL, NULL);
/*
* We're about to drop the metaslab's lock thus allowing other
* consumers to change it's content. Set the metaslab's ms_condensing
* flag to ensure that allocations on this metaslab do not occur
* while we're in the middle of committing it to disk. This is only
* critical for ms_allocatable as all other range trees use per TXG
* views of their content.
*/
msp->ms_condensing = B_TRUE;
mutex_exit(&msp->ms_lock);
uint64_t object = space_map_object(msp->ms_sm);
space_map_truncate(sm,
spa_feature_is_enabled(spa, SPA_FEATURE_LOG_SPACEMAP) ?
zfs_metaslab_sm_blksz_with_log : zfs_metaslab_sm_blksz_no_log, tx);
/*
* space_map_truncate() may have reallocated the spacemap object.
* If so, update the vdev_ms_array.
*/
if (space_map_object(msp->ms_sm) != object) {
object = space_map_object(msp->ms_sm);
dmu_write(spa->spa_meta_objset,
msp->ms_group->mg_vd->vdev_ms_array, sizeof (uint64_t) *
msp->ms_id, sizeof (uint64_t), &object, tx);
}
/*
* Note:
* When the log space map feature is enabled, each space map will
* always have ALLOCS followed by FREES for each sync pass. This is
* typically true even when the log space map feature is disabled,
* except from the case where a metaslab goes through metaslab_sync()
* and gets condensed. In that case the metaslab's space map will have
* ALLOCS followed by FREES (due to condensing) followed by ALLOCS
* followed by FREES (due to space_map_write() in metaslab_sync()) for
* sync pass 1.
*/
range_tree_t *tmp_tree = range_tree_create(NULL, type, NULL, start,
shift);
range_tree_add(tmp_tree, msp->ms_start, msp->ms_size);
space_map_write(sm, tmp_tree, SM_ALLOC, SM_NO_VDEVID, tx);
space_map_write(sm, msp->ms_allocatable, SM_FREE, SM_NO_VDEVID, tx);
space_map_write(sm, condense_tree, SM_FREE, SM_NO_VDEVID, tx);
range_tree_vacate(condense_tree, NULL, NULL);
range_tree_destroy(condense_tree);
range_tree_vacate(tmp_tree, NULL, NULL);
range_tree_destroy(tmp_tree);
mutex_enter(&msp->ms_lock);
msp->ms_condensing = B_FALSE;
metaslab_flush_update(msp, tx);
}
static void
metaslab_unflushed_add(metaslab_t *msp, dmu_tx_t *tx)
{
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
ASSERT(spa_syncing_log_sm(spa) != NULL);
ASSERT(msp->ms_sm != NULL);
ASSERT(range_tree_is_empty(msp->ms_unflushed_allocs));
ASSERT(range_tree_is_empty(msp->ms_unflushed_frees));
mutex_enter(&spa->spa_flushed_ms_lock);
metaslab_set_unflushed_txg(msp, spa_syncing_txg(spa), tx);
metaslab_set_unflushed_dirty(msp, B_TRUE);
avl_add(&spa->spa_metaslabs_by_flushed, msp);
mutex_exit(&spa->spa_flushed_ms_lock);
spa_log_sm_increment_current_mscount(spa);
spa_log_summary_add_flushed_metaslab(spa, B_TRUE);
}
void
metaslab_unflushed_bump(metaslab_t *msp, dmu_tx_t *tx, boolean_t dirty)
{
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
ASSERT(spa_syncing_log_sm(spa) != NULL);
ASSERT(msp->ms_sm != NULL);
ASSERT(metaslab_unflushed_txg(msp) != 0);
ASSERT3P(avl_find(&spa->spa_metaslabs_by_flushed, msp, NULL), ==, msp);
ASSERT(range_tree_is_empty(msp->ms_unflushed_allocs));
ASSERT(range_tree_is_empty(msp->ms_unflushed_frees));
VERIFY3U(tx->tx_txg, <=, spa_final_dirty_txg(spa));
/* update metaslab's position in our flushing tree */
uint64_t ms_prev_flushed_txg = metaslab_unflushed_txg(msp);
boolean_t ms_prev_flushed_dirty = metaslab_unflushed_dirty(msp);
mutex_enter(&spa->spa_flushed_ms_lock);
avl_remove(&spa->spa_metaslabs_by_flushed, msp);
metaslab_set_unflushed_txg(msp, spa_syncing_txg(spa), tx);
metaslab_set_unflushed_dirty(msp, dirty);
avl_add(&spa->spa_metaslabs_by_flushed, msp);
mutex_exit(&spa->spa_flushed_ms_lock);
/* update metaslab counts of spa_log_sm_t nodes */
spa_log_sm_decrement_mscount(spa, ms_prev_flushed_txg);
spa_log_sm_increment_current_mscount(spa);
/* update log space map summary */
spa_log_summary_decrement_mscount(spa, ms_prev_flushed_txg,
ms_prev_flushed_dirty);
spa_log_summary_add_flushed_metaslab(spa, dirty);
/* cleanup obsolete logs if any */
spa_cleanup_old_sm_logs(spa, tx);
}
/*
* Called when the metaslab has been flushed (its own spacemap now reflects
* all the contents of the pool-wide spacemap log). Updates the metaslab's
* metadata and any pool-wide related log space map data (e.g. summary,
* obsolete logs, etc..) to reflect that.
*/
static void
metaslab_flush_update(metaslab_t *msp, dmu_tx_t *tx)
{
metaslab_group_t *mg = msp->ms_group;
spa_t *spa = mg->mg_vd->vdev_spa;
ASSERT(MUTEX_HELD(&msp->ms_lock));
ASSERT3U(spa_sync_pass(spa), ==, 1);
/*
* Just because a metaslab got flushed, that doesn't mean that
* it will pass through metaslab_sync_done(). Thus, make sure to
* update ms_synced_length here in case it doesn't.
*/
msp->ms_synced_length = space_map_length(msp->ms_sm);
/*
* We may end up here from metaslab_condense() without the
* feature being active. In that case this is a no-op.
*/
if (!spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP) ||
metaslab_unflushed_txg(msp) == 0)
return;
metaslab_unflushed_bump(msp, tx, B_FALSE);
}
boolean_t
metaslab_flush(metaslab_t *msp, dmu_tx_t *tx)
{
spa_t *spa = msp->ms_group->mg_vd->vdev_spa;
ASSERT(MUTEX_HELD(&msp->ms_lock));
ASSERT3U(spa_sync_pass(spa), ==, 1);
ASSERT(spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP));
ASSERT(msp->ms_sm != NULL);
ASSERT(metaslab_unflushed_txg(msp) != 0);
ASSERT(avl_find(&spa->spa_metaslabs_by_flushed, msp, NULL) != NULL);
/*
* There is nothing wrong with flushing the same metaslab twice, as
* this codepath should work on that case. However, the current
* flushing scheme makes sure to avoid this situation as we would be
* making all these calls without having anything meaningful to write
* to disk. We assert this behavior here.
*/
ASSERT3U(metaslab_unflushed_txg(msp), <, dmu_tx_get_txg(tx));
/*
* We can not flush while loading, because then we would
* not load the ms_unflushed_{allocs,frees}.
*/
if (msp->ms_loading)
return (B_FALSE);
metaslab_verify_space(msp, dmu_tx_get_txg(tx));
metaslab_verify_weight_and_frag(msp);
/*
* Metaslab condensing is effectively flushing. Therefore if the
* metaslab can be condensed we can just condense it instead of
* flushing it.
*
* Note that metaslab_condense() does call metaslab_flush_update()
* so we can just return immediately after condensing. We also
* don't need to care about setting ms_flushing or broadcasting
* ms_flush_cv, even if we temporarily drop the ms_lock in
* metaslab_condense(), as the metaslab is already loaded.
*/
if (msp->ms_loaded && metaslab_should_condense(msp)) {
metaslab_group_t *mg = msp->ms_group;
/*
* For all histogram operations below refer to the
* comments of metaslab_sync() where we follow a
* similar procedure.
*/
metaslab_group_histogram_verify(mg);
metaslab_class_histogram_verify(mg->mg_class);
metaslab_group_histogram_remove(mg, msp);
metaslab_condense(msp, tx);
space_map_histogram_clear(msp->ms_sm);
space_map_histogram_add(msp->ms_sm, msp->ms_allocatable, tx);
ASSERT(range_tree_is_empty(msp->ms_freed));
for (int t = 0; t < TXG_DEFER_SIZE; t++) {
space_map_histogram_add(msp->ms_sm,
msp->ms_defer[t], tx);
}
metaslab_aux_histograms_update(msp);
metaslab_group_histogram_add(mg, msp);
metaslab_group_histogram_verify(mg);
metaslab_class_histogram_verify(mg->mg_class);
metaslab_verify_space(msp, dmu_tx_get_txg(tx));
/*
* Since we recreated the histogram (and potentially
* the ms_sm too while condensing) ensure that the
* weight is updated too because we are not guaranteed
* that this metaslab is dirty and will go through
* metaslab_sync_done().
*/
metaslab_recalculate_weight_and_sort(msp);
return (B_TRUE);
}
msp->ms_flushing = B_TRUE;
uint64_t sm_len_before = space_map_length(msp->ms_sm);
mutex_exit(&msp->ms_lock);
space_map_write(msp->ms_sm, msp->ms_unflushed_allocs, SM_ALLOC,
SM_NO_VDEVID, tx);
space_map_write(msp->ms_sm, msp->ms_unflushed_frees, SM_FREE,
SM_NO_VDEVID, tx);
mutex_enter(&msp->ms_lock);
uint64_t sm_len_after = space_map_length(msp->ms_sm);
if (zfs_flags & ZFS_DEBUG_LOG_SPACEMAP) {
zfs_dbgmsg("flushing: txg %llu, spa %s, vdev_id %llu, "
"ms_id %llu, unflushed_allocs %llu, unflushed_frees %llu, "
"appended %llu bytes", (u_longlong_t)dmu_tx_get_txg(tx),
spa_name(spa),
(u_longlong_t)msp->ms_group->mg_vd->vdev_id,
(u_longlong_t)msp->ms_id,
(u_longlong_t)range_tree_space(msp->ms_unflushed_allocs),
(u_longlong_t)range_tree_space(msp->ms_unflushed_frees),
(u_longlong_t)(sm_len_after - sm_len_before));
}
ASSERT3U(spa->spa_unflushed_stats.sus_memused, >=,
metaslab_unflushed_changes_memused(msp));
spa->spa_unflushed_stats.sus_memused -=
metaslab_unflushed_changes_memused(msp);
range_tree_vacate(msp->ms_unflushed_allocs, NULL, NULL);
range_tree_vacate(msp->ms_unflushed_frees, NULL, NULL);
metaslab_verify_space(msp, dmu_tx_get_txg(tx));
metaslab_verify_weight_and_frag(msp);
metaslab_flush_update(msp, tx);
metaslab_verify_space(msp, dmu_tx_get_txg(tx));
metaslab_verify_weight_and_frag(msp);
msp->ms_flushing = B_FALSE;
cv_broadcast(&msp->ms_flush_cv);
return (B_TRUE);
}
/*
* Write a metaslab to disk in the context of the specified transaction group.
*/
void
metaslab_sync(metaslab_t *msp, uint64_t txg)
{
metaslab_group_t *mg = msp->ms_group;
vdev_t *vd = mg->mg_vd;
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa_meta_objset(spa);
range_tree_t *alloctree = msp->ms_allocating[txg & TXG_MASK];
dmu_tx_t *tx;
ASSERT(!vd->vdev_ishole);
/*
* This metaslab has just been added so there's no work to do now.
*/
if (msp->ms_new) {
ASSERT0(range_tree_space(alloctree));
ASSERT0(range_tree_space(msp->ms_freeing));
ASSERT0(range_tree_space(msp->ms_freed));
ASSERT0(range_tree_space(msp->ms_checkpointing));
ASSERT0(range_tree_space(msp->ms_trim));
return;
}
/*
* Normally, we don't want to process a metaslab if there are no
* allocations or frees to perform. However, if the metaslab is being
* forced to condense, it's loaded and we're not beyond the final
* dirty txg, we need to let it through. Not condensing beyond the
* final dirty txg prevents an issue where metaslabs that need to be
* condensed but were loaded for other reasons could cause a panic
* here. By only checking the txg in that branch of the conditional,
* we preserve the utility of the VERIFY statements in all other
* cases.
*/
if (range_tree_is_empty(alloctree) &&
range_tree_is_empty(msp->ms_freeing) &&
range_tree_is_empty(msp->ms_checkpointing) &&
!(msp->ms_loaded && msp->ms_condense_wanted &&
txg <= spa_final_dirty_txg(spa)))
return;
VERIFY3U(txg, <=, spa_final_dirty_txg(spa));
/*
* The only state that can actually be changing concurrently
* with metaslab_sync() is the metaslab's ms_allocatable. No
* other thread can be modifying this txg's alloc, freeing,
* freed, or space_map_phys_t. We drop ms_lock whenever we
* could call into the DMU, because the DMU can call down to
* us (e.g. via zio_free()) at any time.
*
* The spa_vdev_remove_thread() can be reading metaslab state
* concurrently, and it is locked out by the ms_sync_lock.
* Note that the ms_lock is insufficient for this, because it
* is dropped by space_map_write().
*/
tx = dmu_tx_create_assigned(spa_get_dsl(spa), txg);
/*
* Generate a log space map if one doesn't exist already.
*/
spa_generate_syncing_log_sm(spa, tx);
if (msp->ms_sm == NULL) {
uint64_t new_object = space_map_alloc(mos,
spa_feature_is_enabled(spa, SPA_FEATURE_LOG_SPACEMAP) ?
zfs_metaslab_sm_blksz_with_log :
zfs_metaslab_sm_blksz_no_log, tx);
VERIFY3U(new_object, !=, 0);
dmu_write(mos, vd->vdev_ms_array, sizeof (uint64_t) *
msp->ms_id, sizeof (uint64_t), &new_object, tx);
VERIFY0(space_map_open(&msp->ms_sm, mos, new_object,
msp->ms_start, msp->ms_size, vd->vdev_ashift));
ASSERT(msp->ms_sm != NULL);
ASSERT(range_tree_is_empty(msp->ms_unflushed_allocs));
ASSERT(range_tree_is_empty(msp->ms_unflushed_frees));
ASSERT0(metaslab_allocated_space(msp));
}
if (!range_tree_is_empty(msp->ms_checkpointing) &&
vd->vdev_checkpoint_sm == NULL) {
ASSERT(spa_has_checkpoint(spa));
uint64_t new_object = space_map_alloc(mos,
zfs_vdev_standard_sm_blksz, tx);
VERIFY3U(new_object, !=, 0);
VERIFY0(space_map_open(&vd->vdev_checkpoint_sm,
mos, new_object, 0, vd->vdev_asize, vd->vdev_ashift));
ASSERT3P(vd->vdev_checkpoint_sm, !=, NULL);
/*
* We save the space map object as an entry in vdev_top_zap
* so it can be retrieved when the pool is reopened after an
* export or through zdb.
*/
VERIFY0(zap_add(vd->vdev_spa->spa_meta_objset,
vd->vdev_top_zap, VDEV_TOP_ZAP_POOL_CHECKPOINT_SM,
sizeof (new_object), 1, &new_object, tx));
}
mutex_enter(&msp->ms_sync_lock);
mutex_enter(&msp->ms_lock);
/*
* Note: metaslab_condense() clears the space map's histogram.
* Therefore we must verify and remove this histogram before
* condensing.
*/
metaslab_group_histogram_verify(mg);
metaslab_class_histogram_verify(mg->mg_class);
metaslab_group_histogram_remove(mg, msp);
if (spa->spa_sync_pass == 1 && msp->ms_loaded &&
metaslab_should_condense(msp))
metaslab_condense(msp, tx);
/*
* We'll be going to disk to sync our space accounting, thus we
* drop the ms_lock during that time so allocations coming from
* open-context (ZIL) for future TXGs do not block.
*/
mutex_exit(&msp->ms_lock);
space_map_t *log_sm = spa_syncing_log_sm(spa);
if (log_sm != NULL) {
ASSERT(spa_feature_is_enabled(spa, SPA_FEATURE_LOG_SPACEMAP));
if (metaslab_unflushed_txg(msp) == 0)
metaslab_unflushed_add(msp, tx);
else if (!metaslab_unflushed_dirty(msp))
metaslab_unflushed_bump(msp, tx, B_TRUE);
space_map_write(log_sm, alloctree, SM_ALLOC,
vd->vdev_id, tx);
space_map_write(log_sm, msp->ms_freeing, SM_FREE,
vd->vdev_id, tx);
mutex_enter(&msp->ms_lock);
ASSERT3U(spa->spa_unflushed_stats.sus_memused, >=,
metaslab_unflushed_changes_memused(msp));
spa->spa_unflushed_stats.sus_memused -=
metaslab_unflushed_changes_memused(msp);
range_tree_remove_xor_add(alloctree,
msp->ms_unflushed_frees, msp->ms_unflushed_allocs);
range_tree_remove_xor_add(msp->ms_freeing,
msp->ms_unflushed_allocs, msp->ms_unflushed_frees);
spa->spa_unflushed_stats.sus_memused +=
metaslab_unflushed_changes_memused(msp);
} else {
ASSERT(!spa_feature_is_enabled(spa, SPA_FEATURE_LOG_SPACEMAP));
space_map_write(msp->ms_sm, alloctree, SM_ALLOC,
SM_NO_VDEVID, tx);
space_map_write(msp->ms_sm, msp->ms_freeing, SM_FREE,
SM_NO_VDEVID, tx);
mutex_enter(&msp->ms_lock);
}
msp->ms_allocated_space += range_tree_space(alloctree);
ASSERT3U(msp->ms_allocated_space, >=,
range_tree_space(msp->ms_freeing));
msp->ms_allocated_space -= range_tree_space(msp->ms_freeing);
if (!range_tree_is_empty(msp->ms_checkpointing)) {
ASSERT(spa_has_checkpoint(spa));
ASSERT3P(vd->vdev_checkpoint_sm, !=, NULL);
/*
* Since we are doing writes to disk and the ms_checkpointing
* tree won't be changing during that time, we drop the
* ms_lock while writing to the checkpoint space map, for the
* same reason mentioned above.
*/
mutex_exit(&msp->ms_lock);
space_map_write(vd->vdev_checkpoint_sm,
msp->ms_checkpointing, SM_FREE, SM_NO_VDEVID, tx);
mutex_enter(&msp->ms_lock);
spa->spa_checkpoint_info.sci_dspace +=
range_tree_space(msp->ms_checkpointing);
vd->vdev_stat.vs_checkpoint_space +=
range_tree_space(msp->ms_checkpointing);
ASSERT3U(vd->vdev_stat.vs_checkpoint_space, ==,
-space_map_allocated(vd->vdev_checkpoint_sm));
range_tree_vacate(msp->ms_checkpointing, NULL, NULL);
}
if (msp->ms_loaded) {
/*
* When the space map is loaded, we have an accurate
* histogram in the range tree. This gives us an opportunity
* to bring the space map's histogram up-to-date so we clear
* it first before updating it.
*/
space_map_histogram_clear(msp->ms_sm);
space_map_histogram_add(msp->ms_sm, msp->ms_allocatable, tx);
/*
* Since we've cleared the histogram we need to add back
* any free space that has already been processed, plus
* any deferred space. This allows the on-disk histogram
* to accurately reflect all free space even if some space
* is not yet available for allocation (i.e. deferred).
*/
space_map_histogram_add(msp->ms_sm, msp->ms_freed, tx);
/*
* Add back any deferred free space that has not been
* added back into the in-core free tree yet. This will
* ensure that we don't end up with a space map histogram
* that is completely empty unless the metaslab is fully
* allocated.
*/
for (int t = 0; t < TXG_DEFER_SIZE; t++) {
space_map_histogram_add(msp->ms_sm,
msp->ms_defer[t], tx);
}
}
/*
* Always add the free space from this sync pass to the space
* map histogram. We want to make sure that the on-disk histogram
* accounts for all free space. If the space map is not loaded,
* then we will lose some accuracy but will correct it the next
* time we load the space map.
*/
space_map_histogram_add(msp->ms_sm, msp->ms_freeing, tx);
metaslab_aux_histograms_update(msp);
metaslab_group_histogram_add(mg, msp);
metaslab_group_histogram_verify(mg);
metaslab_class_histogram_verify(mg->mg_class);
/*
* For sync pass 1, we avoid traversing this txg's free range tree
* and instead will just swap the pointers for freeing and freed.
* We can safely do this since the freed_tree is guaranteed to be
* empty on the initial pass.
*
* Keep in mind that even if we are currently using a log spacemap
* we want current frees to end up in the ms_allocatable (but not
* get appended to the ms_sm) so their ranges can be reused as usual.
*/
if (spa_sync_pass(spa) == 1) {
range_tree_swap(&msp->ms_freeing, &msp->ms_freed);
ASSERT0(msp->ms_allocated_this_txg);
} else {
range_tree_vacate(msp->ms_freeing,
range_tree_add, msp->ms_freed);
}
msp->ms_allocated_this_txg += range_tree_space(alloctree);
range_tree_vacate(alloctree, NULL, NULL);
ASSERT0(range_tree_space(msp->ms_allocating[txg & TXG_MASK]));
ASSERT0(range_tree_space(msp->ms_allocating[TXG_CLEAN(txg)
& TXG_MASK]));
ASSERT0(range_tree_space(msp->ms_freeing));
ASSERT0(range_tree_space(msp->ms_checkpointing));
mutex_exit(&msp->ms_lock);
/*
* Verify that the space map object ID has been recorded in the
* vdev_ms_array.
*/
uint64_t object;
VERIFY0(dmu_read(mos, vd->vdev_ms_array,
msp->ms_id * sizeof (uint64_t), sizeof (uint64_t), &object, 0));
VERIFY3U(object, ==, space_map_object(msp->ms_sm));
mutex_exit(&msp->ms_sync_lock);
dmu_tx_commit(tx);
}
static void
metaslab_evict(metaslab_t *msp, uint64_t txg)
{
if (!msp->ms_loaded || msp->ms_disabled != 0)
return;
for (int t = 1; t < TXG_CONCURRENT_STATES; t++) {
VERIFY0(range_tree_space(
msp->ms_allocating[(txg + t) & TXG_MASK]));
}
if (msp->ms_allocator != -1)
metaslab_passivate(msp, msp->ms_weight & ~METASLAB_ACTIVE_MASK);
if (!metaslab_debug_unload)
metaslab_unload(msp);
}
/*
* Called after a transaction group has completely synced to mark
* all of the metaslab's free space as usable.
*/
void
metaslab_sync_done(metaslab_t *msp, uint64_t txg)
{
metaslab_group_t *mg = msp->ms_group;
vdev_t *vd = mg->mg_vd;
spa_t *spa = vd->vdev_spa;
range_tree_t **defer_tree;
int64_t alloc_delta, defer_delta;
boolean_t defer_allowed = B_TRUE;
ASSERT(!vd->vdev_ishole);
mutex_enter(&msp->ms_lock);
if (msp->ms_new) {
/* this is a new metaslab, add its capacity to the vdev */
metaslab_space_update(vd, mg->mg_class, 0, 0, msp->ms_size);
/* there should be no allocations nor frees at this point */
VERIFY0(msp->ms_allocated_this_txg);
VERIFY0(range_tree_space(msp->ms_freed));
}
ASSERT0(range_tree_space(msp->ms_freeing));
ASSERT0(range_tree_space(msp->ms_checkpointing));
defer_tree = &msp->ms_defer[txg % TXG_DEFER_SIZE];
uint64_t free_space = metaslab_class_get_space(spa_normal_class(spa)) -
metaslab_class_get_alloc(spa_normal_class(spa));
- if (free_space <= spa_get_slop_space(spa) || vd->vdev_removing) {
+ if (free_space <= spa_get_slop_space(spa) || vd->vdev_removing ||
+ vd->vdev_rz_expanding) {
defer_allowed = B_FALSE;
}
defer_delta = 0;
alloc_delta = msp->ms_allocated_this_txg -
range_tree_space(msp->ms_freed);
if (defer_allowed) {
defer_delta = range_tree_space(msp->ms_freed) -
range_tree_space(*defer_tree);
} else {
defer_delta -= range_tree_space(*defer_tree);
}
metaslab_space_update(vd, mg->mg_class, alloc_delta + defer_delta,
defer_delta, 0);
if (spa_syncing_log_sm(spa) == NULL) {
/*
* If there's a metaslab_load() in progress and we don't have
* a log space map, it means that we probably wrote to the
* metaslab's space map. If this is the case, we need to
* make sure that we wait for the load to complete so that we
* have a consistent view at the in-core side of the metaslab.
*/
metaslab_load_wait(msp);
} else {
ASSERT(spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP));
}
/*
* When auto-trimming is enabled, free ranges which are added to
* ms_allocatable are also be added to ms_trim. The ms_trim tree is
* periodically consumed by the vdev_autotrim_thread() which issues
* trims for all ranges and then vacates the tree. The ms_trim tree
* can be discarded at any time with the sole consequence of recent
* frees not being trimmed.
*/
if (spa_get_autotrim(spa) == SPA_AUTOTRIM_ON) {
range_tree_walk(*defer_tree, range_tree_add, msp->ms_trim);
if (!defer_allowed) {
range_tree_walk(msp->ms_freed, range_tree_add,
msp->ms_trim);
}
} else {
range_tree_vacate(msp->ms_trim, NULL, NULL);
}
/*
* Move the frees from the defer_tree back to the free
* range tree (if it's loaded). Swap the freed_tree and
* the defer_tree -- this is safe to do because we've
* just emptied out the defer_tree.
*/
range_tree_vacate(*defer_tree,
msp->ms_loaded ? range_tree_add : NULL, msp->ms_allocatable);
if (defer_allowed) {
range_tree_swap(&msp->ms_freed, defer_tree);
} else {
range_tree_vacate(msp->ms_freed,
msp->ms_loaded ? range_tree_add : NULL,
msp->ms_allocatable);
}
msp->ms_synced_length = space_map_length(msp->ms_sm);
msp->ms_deferspace += defer_delta;
ASSERT3S(msp->ms_deferspace, >=, 0);
ASSERT3S(msp->ms_deferspace, <=, msp->ms_size);
if (msp->ms_deferspace != 0) {
/*
* Keep syncing this metaslab until all deferred frees
* are back in circulation.
*/
vdev_dirty(vd, VDD_METASLAB, msp, txg + 1);
}
metaslab_aux_histograms_update_done(msp, defer_allowed);
if (msp->ms_new) {
msp->ms_new = B_FALSE;
mutex_enter(&mg->mg_lock);
mg->mg_ms_ready++;
mutex_exit(&mg->mg_lock);
}
/*
* Re-sort metaslab within its group now that we've adjusted
* its allocatable space.
*/
metaslab_recalculate_weight_and_sort(msp);
ASSERT0(range_tree_space(msp->ms_allocating[txg & TXG_MASK]));
ASSERT0(range_tree_space(msp->ms_freeing));
ASSERT0(range_tree_space(msp->ms_freed));
ASSERT0(range_tree_space(msp->ms_checkpointing));
msp->ms_allocating_total -= msp->ms_allocated_this_txg;
msp->ms_allocated_this_txg = 0;
mutex_exit(&msp->ms_lock);
}
void
metaslab_sync_reassess(metaslab_group_t *mg)
{
spa_t *spa = mg->mg_class->mc_spa;
spa_config_enter(spa, SCL_ALLOC, FTAG, RW_READER);
metaslab_group_alloc_update(mg);
mg->mg_fragmentation = metaslab_group_fragmentation(mg);
/*
* Preload the next potential metaslabs but only on active
* metaslab groups. We can get into a state where the metaslab
* is no longer active since we dirty metaslabs as we remove a
* a device, thus potentially making the metaslab group eligible
* for preloading.
*/
if (mg->mg_activation_count > 0) {
metaslab_group_preload(mg);
}
spa_config_exit(spa, SCL_ALLOC, FTAG);
}
/*
* When writing a ditto block (i.e. more than one DVA for a given BP) on
* the same vdev as an existing DVA of this BP, then try to allocate it
* on a different metaslab than existing DVAs (i.e. a unique metaslab).
*/
static boolean_t
metaslab_is_unique(metaslab_t *msp, dva_t *dva)
{
uint64_t dva_ms_id;
if (DVA_GET_ASIZE(dva) == 0)
return (B_TRUE);
if (msp->ms_group->mg_vd->vdev_id != DVA_GET_VDEV(dva))
return (B_TRUE);
dva_ms_id = DVA_GET_OFFSET(dva) >> msp->ms_group->mg_vd->vdev_ms_shift;
return (msp->ms_id != dva_ms_id);
}
/*
* ==========================================================================
* Metaslab allocation tracing facility
* ==========================================================================
*/
/*
* Add an allocation trace element to the allocation tracing list.
*/
static void
metaslab_trace_add(zio_alloc_list_t *zal, metaslab_group_t *mg,
metaslab_t *msp, uint64_t psize, uint32_t dva_id, uint64_t offset,
int allocator)
{
metaslab_alloc_trace_t *mat;
if (!metaslab_trace_enabled)
return;
/*
* When the tracing list reaches its maximum we remove
* the second element in the list before adding a new one.
* By removing the second element we preserve the original
* entry as a clue to what allocations steps have already been
* performed.
*/
if (zal->zal_size == metaslab_trace_max_entries) {
metaslab_alloc_trace_t *mat_next;
#ifdef ZFS_DEBUG
panic("too many entries in allocation list");
#endif
METASLABSTAT_BUMP(metaslabstat_trace_over_limit);
zal->zal_size--;
mat_next = list_next(&zal->zal_list, list_head(&zal->zal_list));
list_remove(&zal->zal_list, mat_next);
kmem_cache_free(metaslab_alloc_trace_cache, mat_next);
}
mat = kmem_cache_alloc(metaslab_alloc_trace_cache, KM_SLEEP);
list_link_init(&mat->mat_list_node);
mat->mat_mg = mg;
mat->mat_msp = msp;
mat->mat_size = psize;
mat->mat_dva_id = dva_id;
mat->mat_offset = offset;
mat->mat_weight = 0;
mat->mat_allocator = allocator;
if (msp != NULL)
mat->mat_weight = msp->ms_weight;
/*
* The list is part of the zio so locking is not required. Only
* a single thread will perform allocations for a given zio.
*/
list_insert_tail(&zal->zal_list, mat);
zal->zal_size++;
ASSERT3U(zal->zal_size, <=, metaslab_trace_max_entries);
}
void
metaslab_trace_init(zio_alloc_list_t *zal)
{
list_create(&zal->zal_list, sizeof (metaslab_alloc_trace_t),
offsetof(metaslab_alloc_trace_t, mat_list_node));
zal->zal_size = 0;
}
void
metaslab_trace_fini(zio_alloc_list_t *zal)
{
metaslab_alloc_trace_t *mat;
while ((mat = list_remove_head(&zal->zal_list)) != NULL)
kmem_cache_free(metaslab_alloc_trace_cache, mat);
list_destroy(&zal->zal_list);
zal->zal_size = 0;
}
/*
* ==========================================================================
* Metaslab block operations
* ==========================================================================
*/
static void
metaslab_group_alloc_increment(spa_t *spa, uint64_t vdev, const void *tag,
int flags, int allocator)
{
if (!(flags & METASLAB_ASYNC_ALLOC) ||
(flags & METASLAB_DONT_THROTTLE))
return;
metaslab_group_t *mg = vdev_lookup_top(spa, vdev)->vdev_mg;
if (!mg->mg_class->mc_alloc_throttle_enabled)
return;
metaslab_group_allocator_t *mga = &mg->mg_allocator[allocator];
(void) zfs_refcount_add(&mga->mga_alloc_queue_depth, tag);
}
static void
metaslab_group_increment_qdepth(metaslab_group_t *mg, int allocator)
{
metaslab_group_allocator_t *mga = &mg->mg_allocator[allocator];
metaslab_class_allocator_t *mca =
&mg->mg_class->mc_allocator[allocator];
uint64_t max = mg->mg_max_alloc_queue_depth;
uint64_t cur = mga->mga_cur_max_alloc_queue_depth;
while (cur < max) {
if (atomic_cas_64(&mga->mga_cur_max_alloc_queue_depth,
cur, cur + 1) == cur) {
atomic_inc_64(&mca->mca_alloc_max_slots);
return;
}
cur = mga->mga_cur_max_alloc_queue_depth;
}
}
void
metaslab_group_alloc_decrement(spa_t *spa, uint64_t vdev, const void *tag,
int flags, int allocator, boolean_t io_complete)
{
if (!(flags & METASLAB_ASYNC_ALLOC) ||
(flags & METASLAB_DONT_THROTTLE))
return;
metaslab_group_t *mg = vdev_lookup_top(spa, vdev)->vdev_mg;
if (!mg->mg_class->mc_alloc_throttle_enabled)
return;
metaslab_group_allocator_t *mga = &mg->mg_allocator[allocator];
(void) zfs_refcount_remove(&mga->mga_alloc_queue_depth, tag);
if (io_complete)
metaslab_group_increment_qdepth(mg, allocator);
}
void
metaslab_group_alloc_verify(spa_t *spa, const blkptr_t *bp, const void *tag,
int allocator)
{
#ifdef ZFS_DEBUG
const dva_t *dva = bp->blk_dva;
int ndvas = BP_GET_NDVAS(bp);
for (int d = 0; d < ndvas; d++) {
uint64_t vdev = DVA_GET_VDEV(&dva[d]);
metaslab_group_t *mg = vdev_lookup_top(spa, vdev)->vdev_mg;
metaslab_group_allocator_t *mga = &mg->mg_allocator[allocator];
VERIFY(zfs_refcount_not_held(&mga->mga_alloc_queue_depth, tag));
}
#endif
}
static uint64_t
metaslab_block_alloc(metaslab_t *msp, uint64_t size, uint64_t txg)
{
uint64_t start;
range_tree_t *rt = msp->ms_allocatable;
metaslab_class_t *mc = msp->ms_group->mg_class;
ASSERT(MUTEX_HELD(&msp->ms_lock));
VERIFY(!msp->ms_condensing);
VERIFY0(msp->ms_disabled);
+ VERIFY0(msp->ms_new);
start = mc->mc_ops->msop_alloc(msp, size);
if (start != -1ULL) {
metaslab_group_t *mg = msp->ms_group;
vdev_t *vd = mg->mg_vd;
VERIFY0(P2PHASE(start, 1ULL << vd->vdev_ashift));
VERIFY0(P2PHASE(size, 1ULL << vd->vdev_ashift));
VERIFY3U(range_tree_space(rt) - size, <=, msp->ms_size);
range_tree_remove(rt, start, size);
range_tree_clear(msp->ms_trim, start, size);
if (range_tree_is_empty(msp->ms_allocating[txg & TXG_MASK]))
vdev_dirty(mg->mg_vd, VDD_METASLAB, msp, txg);
range_tree_add(msp->ms_allocating[txg & TXG_MASK], start, size);
msp->ms_allocating_total += size;
/* Track the last successful allocation */
msp->ms_alloc_txg = txg;
metaslab_verify_space(msp, txg);
}
/*
* Now that we've attempted the allocation we need to update the
* metaslab's maximum block size since it may have changed.
*/
msp->ms_max_size = metaslab_largest_allocatable(msp);
return (start);
}
/*
* Find the metaslab with the highest weight that is less than what we've
* already tried. In the common case, this means that we will examine each
* metaslab at most once. Note that concurrent callers could reorder metaslabs
* by activation/passivation once we have dropped the mg_lock. If a metaslab is
* activated by another thread, and we fail to allocate from the metaslab we
* have selected, we may not try the newly-activated metaslab, and instead
* activate another metaslab. This is not optimal, but generally does not cause
* any problems (a possible exception being if every metaslab is completely full
* except for the newly-activated metaslab which we fail to examine).
*/
static metaslab_t *
find_valid_metaslab(metaslab_group_t *mg, uint64_t activation_weight,
dva_t *dva, int d, boolean_t want_unique, uint64_t asize, int allocator,
boolean_t try_hard, zio_alloc_list_t *zal, metaslab_t *search,
boolean_t *was_active)
{
avl_index_t idx;
avl_tree_t *t = &mg->mg_metaslab_tree;
metaslab_t *msp = avl_find(t, search, &idx);
if (msp == NULL)
msp = avl_nearest(t, idx, AVL_AFTER);
uint_t tries = 0;
for (; msp != NULL; msp = AVL_NEXT(t, msp)) {
int i;
if (!try_hard && tries > zfs_metaslab_find_max_tries) {
METASLABSTAT_BUMP(metaslabstat_too_many_tries);
return (NULL);
}
tries++;
if (!metaslab_should_allocate(msp, asize, try_hard)) {
metaslab_trace_add(zal, mg, msp, asize, d,
TRACE_TOO_SMALL, allocator);
continue;
}
/*
- * If the selected metaslab is condensing or disabled,
- * skip it.
+ * If the selected metaslab is condensing or disabled, or
+ * hasn't gone through a metaslab_sync_done(), then skip it.
*/
- if (msp->ms_condensing || msp->ms_disabled > 0)
+ if (msp->ms_condensing || msp->ms_disabled > 0 || msp->ms_new)
continue;
*was_active = msp->ms_allocator != -1;
/*
* If we're activating as primary, this is our first allocation
* from this disk, so we don't need to check how close we are.
* If the metaslab under consideration was already active,
* we're getting desperate enough to steal another allocator's
* metaslab, so we still don't care about distances.
*/
if (activation_weight == METASLAB_WEIGHT_PRIMARY || *was_active)
break;
for (i = 0; i < d; i++) {
if (want_unique &&
!metaslab_is_unique(msp, &dva[i]))
break; /* try another metaslab */
}
if (i == d)
break;
}
if (msp != NULL) {
search->ms_weight = msp->ms_weight;
search->ms_start = msp->ms_start + 1;
search->ms_allocator = msp->ms_allocator;
search->ms_primary = msp->ms_primary;
}
return (msp);
}
static void
metaslab_active_mask_verify(metaslab_t *msp)
{
ASSERT(MUTEX_HELD(&msp->ms_lock));
if ((zfs_flags & ZFS_DEBUG_METASLAB_VERIFY) == 0)
return;
if ((msp->ms_weight & METASLAB_ACTIVE_MASK) == 0)
return;
if (msp->ms_weight & METASLAB_WEIGHT_PRIMARY) {
VERIFY0(msp->ms_weight & METASLAB_WEIGHT_SECONDARY);
VERIFY0(msp->ms_weight & METASLAB_WEIGHT_CLAIM);
VERIFY3S(msp->ms_allocator, !=, -1);
VERIFY(msp->ms_primary);
return;
}
if (msp->ms_weight & METASLAB_WEIGHT_SECONDARY) {
VERIFY0(msp->ms_weight & METASLAB_WEIGHT_PRIMARY);
VERIFY0(msp->ms_weight & METASLAB_WEIGHT_CLAIM);
VERIFY3S(msp->ms_allocator, !=, -1);
VERIFY(!msp->ms_primary);
return;
}
if (msp->ms_weight & METASLAB_WEIGHT_CLAIM) {
VERIFY0(msp->ms_weight & METASLAB_WEIGHT_PRIMARY);
VERIFY0(msp->ms_weight & METASLAB_WEIGHT_SECONDARY);
VERIFY3S(msp->ms_allocator, ==, -1);
return;
}
}
static uint64_t
metaslab_group_alloc_normal(metaslab_group_t *mg, zio_alloc_list_t *zal,
uint64_t asize, uint64_t txg, boolean_t want_unique, dva_t *dva, int d,
int allocator, boolean_t try_hard)
{
metaslab_t *msp = NULL;
uint64_t offset = -1ULL;
uint64_t activation_weight = METASLAB_WEIGHT_PRIMARY;
for (int i = 0; i < d; i++) {
if (activation_weight == METASLAB_WEIGHT_PRIMARY &&
DVA_GET_VDEV(&dva[i]) == mg->mg_vd->vdev_id) {
activation_weight = METASLAB_WEIGHT_SECONDARY;
} else if (activation_weight == METASLAB_WEIGHT_SECONDARY &&
DVA_GET_VDEV(&dva[i]) == mg->mg_vd->vdev_id) {
activation_weight = METASLAB_WEIGHT_CLAIM;
break;
}
}
/*
* If we don't have enough metaslabs active to fill the entire array, we
* just use the 0th slot.
*/
if (mg->mg_ms_ready < mg->mg_allocators * 3)
allocator = 0;
metaslab_group_allocator_t *mga = &mg->mg_allocator[allocator];
ASSERT3U(mg->mg_vd->vdev_ms_count, >=, 2);
metaslab_t *search = kmem_alloc(sizeof (*search), KM_SLEEP);
search->ms_weight = UINT64_MAX;
search->ms_start = 0;
/*
* At the end of the metaslab tree are the already-active metaslabs,
* first the primaries, then the secondaries. When we resume searching
* through the tree, we need to consider ms_allocator and ms_primary so
* we start in the location right after where we left off, and don't
* accidentally loop forever considering the same metaslabs.
*/
search->ms_allocator = -1;
search->ms_primary = B_TRUE;
for (;;) {
boolean_t was_active = B_FALSE;
mutex_enter(&mg->mg_lock);
if (activation_weight == METASLAB_WEIGHT_PRIMARY &&
mga->mga_primary != NULL) {
msp = mga->mga_primary;
/*
* Even though we don't hold the ms_lock for the
* primary metaslab, those fields should not
* change while we hold the mg_lock. Thus it is
* safe to make assertions on them.
*/
ASSERT(msp->ms_primary);
ASSERT3S(msp->ms_allocator, ==, allocator);
ASSERT(msp->ms_loaded);
was_active = B_TRUE;
ASSERT(msp->ms_weight & METASLAB_ACTIVE_MASK);
} else if (activation_weight == METASLAB_WEIGHT_SECONDARY &&
mga->mga_secondary != NULL) {
msp = mga->mga_secondary;
/*
* See comment above about the similar assertions
* for the primary metaslab.
*/
ASSERT(!msp->ms_primary);
ASSERT3S(msp->ms_allocator, ==, allocator);
ASSERT(msp->ms_loaded);
was_active = B_TRUE;
ASSERT(msp->ms_weight & METASLAB_ACTIVE_MASK);
} else {
msp = find_valid_metaslab(mg, activation_weight, dva, d,
want_unique, asize, allocator, try_hard, zal,
search, &was_active);
}
mutex_exit(&mg->mg_lock);
if (msp == NULL) {
kmem_free(search, sizeof (*search));
return (-1ULL);
}
mutex_enter(&msp->ms_lock);
metaslab_active_mask_verify(msp);
/*
* This code is disabled out because of issues with
* tracepoints in non-gpl kernel modules.
*/
#if 0
DTRACE_PROBE3(ms__activation__attempt,
metaslab_t *, msp, uint64_t, activation_weight,
boolean_t, was_active);
#endif
/*
* Ensure that the metaslab we have selected is still
* capable of handling our request. It's possible that
* another thread may have changed the weight while we
* were blocked on the metaslab lock. We check the
* active status first to see if we need to set_selected_txg
* a new metaslab.
*/
if (was_active && !(msp->ms_weight & METASLAB_ACTIVE_MASK)) {
ASSERT3S(msp->ms_allocator, ==, -1);
mutex_exit(&msp->ms_lock);
continue;
}
/*
* If the metaslab was activated for another allocator
* while we were waiting in the ms_lock above, or it's
* a primary and we're seeking a secondary (or vice versa),
* we go back and select a new metaslab.
*/
if (!was_active && (msp->ms_weight & METASLAB_ACTIVE_MASK) &&
(msp->ms_allocator != -1) &&
(msp->ms_allocator != allocator || ((activation_weight ==
METASLAB_WEIGHT_PRIMARY) != msp->ms_primary))) {
ASSERT(msp->ms_loaded);
ASSERT((msp->ms_weight & METASLAB_WEIGHT_CLAIM) ||
msp->ms_allocator != -1);
mutex_exit(&msp->ms_lock);
continue;
}
/*
* This metaslab was used for claiming regions allocated
* by the ZIL during pool import. Once these regions are
* claimed we don't need to keep the CLAIM bit set
* anymore. Passivate this metaslab to zero its activation
* mask.
*/
if (msp->ms_weight & METASLAB_WEIGHT_CLAIM &&
activation_weight != METASLAB_WEIGHT_CLAIM) {
ASSERT(msp->ms_loaded);
ASSERT3S(msp->ms_allocator, ==, -1);
metaslab_passivate(msp, msp->ms_weight &
~METASLAB_WEIGHT_CLAIM);
mutex_exit(&msp->ms_lock);
continue;
}
metaslab_set_selected_txg(msp, txg);
int activation_error =
metaslab_activate(msp, allocator, activation_weight);
metaslab_active_mask_verify(msp);
/*
* If the metaslab was activated by another thread for
* another allocator or activation_weight (EBUSY), or it
* failed because another metaslab was assigned as primary
* for this allocator (EEXIST) we continue using this
* metaslab for our allocation, rather than going on to a
* worse metaslab (we waited for that metaslab to be loaded
* after all).
*
* If the activation failed due to an I/O error or ENOSPC we
* skip to the next metaslab.
*/
boolean_t activated;
if (activation_error == 0) {
activated = B_TRUE;
} else if (activation_error == EBUSY ||
activation_error == EEXIST) {
activated = B_FALSE;
} else {
mutex_exit(&msp->ms_lock);
continue;
}
ASSERT(msp->ms_loaded);
/*
* Now that we have the lock, recheck to see if we should
* continue to use this metaslab for this allocation. The
* the metaslab is now loaded so metaslab_should_allocate()
* can accurately determine if the allocation attempt should
* proceed.
*/
if (!metaslab_should_allocate(msp, asize, try_hard)) {
/* Passivate this metaslab and select a new one. */
metaslab_trace_add(zal, mg, msp, asize, d,
TRACE_TOO_SMALL, allocator);
goto next;
}
/*
* If this metaslab is currently condensing then pick again
* as we can't manipulate this metaslab until it's committed
* to disk. If this metaslab is being initialized, we shouldn't
* allocate from it since the allocated region might be
* overwritten after allocation.
*/
if (msp->ms_condensing) {
metaslab_trace_add(zal, mg, msp, asize, d,
TRACE_CONDENSING, allocator);
if (activated) {
metaslab_passivate(msp, msp->ms_weight &
~METASLAB_ACTIVE_MASK);
}
mutex_exit(&msp->ms_lock);
continue;
} else if (msp->ms_disabled > 0) {
metaslab_trace_add(zal, mg, msp, asize, d,
TRACE_DISABLED, allocator);
if (activated) {
metaslab_passivate(msp, msp->ms_weight &
~METASLAB_ACTIVE_MASK);
}
mutex_exit(&msp->ms_lock);
continue;
}
offset = metaslab_block_alloc(msp, asize, txg);
metaslab_trace_add(zal, mg, msp, asize, d, offset, allocator);
if (offset != -1ULL) {
/* Proactively passivate the metaslab, if needed */
if (activated)
metaslab_segment_may_passivate(msp);
break;
}
next:
ASSERT(msp->ms_loaded);
/*
* This code is disabled out because of issues with
* tracepoints in non-gpl kernel modules.
*/
#if 0
DTRACE_PROBE2(ms__alloc__failure, metaslab_t *, msp,
uint64_t, asize);
#endif
/*
* We were unable to allocate from this metaslab so determine
* a new weight for this metaslab. Now that we have loaded
* the metaslab we can provide a better hint to the metaslab
* selector.
*
* For space-based metaslabs, we use the maximum block size.
* This information is only available when the metaslab
* is loaded and is more accurate than the generic free
* space weight that was calculated by metaslab_weight().
* This information allows us to quickly compare the maximum
* available allocation in the metaslab to the allocation
* size being requested.
*
* For segment-based metaslabs, determine the new weight
* based on the highest bucket in the range tree. We
* explicitly use the loaded segment weight (i.e. the range
* tree histogram) since it contains the space that is
* currently available for allocation and is accurate
* even within a sync pass.
*/
uint64_t weight;
if (WEIGHT_IS_SPACEBASED(msp->ms_weight)) {
weight = metaslab_largest_allocatable(msp);
WEIGHT_SET_SPACEBASED(weight);
} else {
weight = metaslab_weight_from_range_tree(msp);
}
if (activated) {
metaslab_passivate(msp, weight);
} else {
/*
* For the case where we use the metaslab that is
* active for another allocator we want to make
* sure that we retain the activation mask.
*
* Note that we could attempt to use something like
* metaslab_recalculate_weight_and_sort() that
* retains the activation mask here. That function
* uses metaslab_weight() to set the weight though
* which is not as accurate as the calculations
* above.
*/
weight |= msp->ms_weight & METASLAB_ACTIVE_MASK;
metaslab_group_sort(mg, msp, weight);
}
metaslab_active_mask_verify(msp);
/*
* We have just failed an allocation attempt, check
* that metaslab_should_allocate() agrees. Otherwise,
* we may end up in an infinite loop retrying the same
* metaslab.
*/
ASSERT(!metaslab_should_allocate(msp, asize, try_hard));
mutex_exit(&msp->ms_lock);
}
mutex_exit(&msp->ms_lock);
kmem_free(search, sizeof (*search));
return (offset);
}
static uint64_t
metaslab_group_alloc(metaslab_group_t *mg, zio_alloc_list_t *zal,
uint64_t asize, uint64_t txg, boolean_t want_unique, dva_t *dva, int d,
int allocator, boolean_t try_hard)
{
uint64_t offset;
ASSERT(mg->mg_initialized);
offset = metaslab_group_alloc_normal(mg, zal, asize, txg, want_unique,
dva, d, allocator, try_hard);
mutex_enter(&mg->mg_lock);
if (offset == -1ULL) {
mg->mg_failed_allocations++;
metaslab_trace_add(zal, mg, NULL, asize, d,
TRACE_GROUP_FAILURE, allocator);
if (asize == SPA_GANGBLOCKSIZE) {
/*
* This metaslab group was unable to allocate
* the minimum gang block size so it must be out of
* space. We must notify the allocation throttle
* to start skipping allocation attempts to this
* metaslab group until more space becomes available.
* Note: this failure cannot be caused by the
* allocation throttle since the allocation throttle
* is only responsible for skipping devices and
* not failing block allocations.
*/
mg->mg_no_free_space = B_TRUE;
}
}
mg->mg_allocations++;
mutex_exit(&mg->mg_lock);
return (offset);
}
/*
* Allocate a block for the specified i/o.
*/
int
metaslab_alloc_dva(spa_t *spa, metaslab_class_t *mc, uint64_t psize,
dva_t *dva, int d, dva_t *hintdva, uint64_t txg, int flags,
zio_alloc_list_t *zal, int allocator)
{
metaslab_class_allocator_t *mca = &mc->mc_allocator[allocator];
metaslab_group_t *mg, *rotor;
vdev_t *vd;
boolean_t try_hard = B_FALSE;
ASSERT(!DVA_IS_VALID(&dva[d]));
/*
* For testing, make some blocks above a certain size be gang blocks.
* This will result in more split blocks when using device removal,
* and a large number of split blocks coupled with ztest-induced
* damage can result in extremely long reconstruction times. This
* will also test spilling from special to normal.
*/
if (psize >= metaslab_force_ganging &&
metaslab_force_ganging_pct > 0 &&
(random_in_range(100) < MIN(metaslab_force_ganging_pct, 100))) {
metaslab_trace_add(zal, NULL, NULL, psize, d, TRACE_FORCE_GANG,
allocator);
return (SET_ERROR(ENOSPC));
}
/*
* Start at the rotor and loop through all mgs until we find something.
* Note that there's no locking on mca_rotor or mca_aliquot because
* nothing actually breaks if we miss a few updates -- we just won't
* allocate quite as evenly. It all balances out over time.
*
* If we are doing ditto or log blocks, try to spread them across
* consecutive vdevs. If we're forced to reuse a vdev before we've
* allocated all of our ditto blocks, then try and spread them out on
* that vdev as much as possible. If it turns out to not be possible,
* gradually lower our standards until anything becomes acceptable.
* Also, allocating on consecutive vdevs (as opposed to random vdevs)
* gives us hope of containing our fault domains to something we're
* able to reason about. Otherwise, any two top-level vdev failures
* will guarantee the loss of data. With consecutive allocation,
* only two adjacent top-level vdev failures will result in data loss.
*
* If we are doing gang blocks (hintdva is non-NULL), try to keep
* ourselves on the same vdev as our gang block header. That
* way, we can hope for locality in vdev_cache, plus it makes our
* fault domains something tractable.
*/
if (hintdva) {
vd = vdev_lookup_top(spa, DVA_GET_VDEV(&hintdva[d]));
/*
* It's possible the vdev we're using as the hint no
* longer exists or its mg has been closed (e.g. by
* device removal). Consult the rotor when
* all else fails.
*/
if (vd != NULL && vd->vdev_mg != NULL) {
mg = vdev_get_mg(vd, mc);
if (flags & METASLAB_HINTBP_AVOID)
mg = mg->mg_next;
} else {
mg = mca->mca_rotor;
}
} else if (d != 0) {
vd = vdev_lookup_top(spa, DVA_GET_VDEV(&dva[d - 1]));
mg = vd->vdev_mg->mg_next;
} else {
ASSERT(mca->mca_rotor != NULL);
mg = mca->mca_rotor;
}
/*
* If the hint put us into the wrong metaslab class, or into a
* metaslab group that has been passivated, just follow the rotor.
*/
if (mg->mg_class != mc || mg->mg_activation_count <= 0)
mg = mca->mca_rotor;
rotor = mg;
top:
do {
boolean_t allocatable;
ASSERT(mg->mg_activation_count == 1);
vd = mg->mg_vd;
/*
* Don't allocate from faulted devices.
*/
if (try_hard) {
spa_config_enter(spa, SCL_ZIO, FTAG, RW_READER);
allocatable = vdev_allocatable(vd);
spa_config_exit(spa, SCL_ZIO, FTAG);
} else {
allocatable = vdev_allocatable(vd);
}
/*
* Determine if the selected metaslab group is eligible
* for allocations. If we're ganging then don't allow
* this metaslab group to skip allocations since that would
* inadvertently return ENOSPC and suspend the pool
* even though space is still available.
*/
if (allocatable && !GANG_ALLOCATION(flags) && !try_hard) {
allocatable = metaslab_group_allocatable(mg, rotor,
flags, psize, allocator, d);
}
if (!allocatable) {
metaslab_trace_add(zal, mg, NULL, psize, d,
TRACE_NOT_ALLOCATABLE, allocator);
goto next;
}
ASSERT(mg->mg_initialized);
/*
* Avoid writing single-copy data to an unhealthy,
* non-redundant vdev, unless we've already tried all
* other vdevs.
*/
if (vd->vdev_state < VDEV_STATE_HEALTHY &&
d == 0 && !try_hard && vd->vdev_children == 0) {
metaslab_trace_add(zal, mg, NULL, psize, d,
TRACE_VDEV_ERROR, allocator);
goto next;
}
ASSERT(mg->mg_class == mc);
- uint64_t asize = vdev_psize_to_asize(vd, psize);
+ uint64_t asize = vdev_psize_to_asize_txg(vd, psize, txg);
ASSERT(P2PHASE(asize, 1ULL << vd->vdev_ashift) == 0);
/*
* If we don't need to try hard, then require that the
* block be on a different metaslab from any other DVAs
* in this BP (unique=true). If we are trying hard, then
* allow any metaslab to be used (unique=false).
*/
uint64_t offset = metaslab_group_alloc(mg, zal, asize, txg,
!try_hard, dva, d, allocator, try_hard);
if (offset != -1ULL) {
/*
* If we've just selected this metaslab group,
* figure out whether the corresponding vdev is
* over- or under-used relative to the pool,
* and set an allocation bias to even it out.
*
* Bias is also used to compensate for unequally
* sized vdevs so that space is allocated fairly.
*/
if (mca->mca_aliquot == 0 && metaslab_bias_enabled) {
vdev_stat_t *vs = &vd->vdev_stat;
int64_t vs_free = vs->vs_space - vs->vs_alloc;
int64_t mc_free = mc->mc_space - mc->mc_alloc;
int64_t ratio;
/*
* Calculate how much more or less we should
* try to allocate from this device during
* this iteration around the rotor.
*
* This basically introduces a zero-centered
* bias towards the devices with the most
* free space, while compensating for vdev
* size differences.
*
* Examples:
* vdev V1 = 16M/128M
* vdev V2 = 16M/128M
* ratio(V1) = 100% ratio(V2) = 100%
*
* vdev V1 = 16M/128M
* vdev V2 = 64M/128M
* ratio(V1) = 127% ratio(V2) = 72%
*
* vdev V1 = 16M/128M
* vdev V2 = 64M/512M
* ratio(V1) = 40% ratio(V2) = 160%
*/
ratio = (vs_free * mc->mc_alloc_groups * 100) /
(mc_free + 1);
mg->mg_bias = ((ratio - 100) *
(int64_t)mg->mg_aliquot) / 100;
} else if (!metaslab_bias_enabled) {
mg->mg_bias = 0;
}
if ((flags & METASLAB_ZIL) ||
atomic_add_64_nv(&mca->mca_aliquot, asize) >=
mg->mg_aliquot + mg->mg_bias) {
mca->mca_rotor = mg->mg_next;
mca->mca_aliquot = 0;
}
DVA_SET_VDEV(&dva[d], vd->vdev_id);
DVA_SET_OFFSET(&dva[d], offset);
DVA_SET_GANG(&dva[d],
((flags & METASLAB_GANG_HEADER) ? 1 : 0));
DVA_SET_ASIZE(&dva[d], asize);
return (0);
}
next:
mca->mca_rotor = mg->mg_next;
mca->mca_aliquot = 0;
} while ((mg = mg->mg_next) != rotor);
/*
* If we haven't tried hard, perhaps do so now.
*/
if (!try_hard && (zfs_metaslab_try_hard_before_gang ||
GANG_ALLOCATION(flags) || (flags & METASLAB_ZIL) != 0 ||
psize <= 1 << spa->spa_min_ashift)) {
METASLABSTAT_BUMP(metaslabstat_try_hard);
try_hard = B_TRUE;
goto top;
}
memset(&dva[d], 0, sizeof (dva_t));
metaslab_trace_add(zal, rotor, NULL, psize, d, TRACE_ENOSPC, allocator);
return (SET_ERROR(ENOSPC));
}
void
metaslab_free_concrete(vdev_t *vd, uint64_t offset, uint64_t asize,
boolean_t checkpoint)
{
metaslab_t *msp;
spa_t *spa = vd->vdev_spa;
ASSERT(vdev_is_concrete(vd));
ASSERT3U(spa_config_held(spa, SCL_ALL, RW_READER), !=, 0);
ASSERT3U(offset >> vd->vdev_ms_shift, <, vd->vdev_ms_count);
msp = vd->vdev_ms[offset >> vd->vdev_ms_shift];
VERIFY(!msp->ms_condensing);
VERIFY3U(offset, >=, msp->ms_start);
VERIFY3U(offset + asize, <=, msp->ms_start + msp->ms_size);
VERIFY0(P2PHASE(offset, 1ULL << vd->vdev_ashift));
VERIFY0(P2PHASE(asize, 1ULL << vd->vdev_ashift));
metaslab_check_free_impl(vd, offset, asize);
mutex_enter(&msp->ms_lock);
if (range_tree_is_empty(msp->ms_freeing) &&
range_tree_is_empty(msp->ms_checkpointing)) {
vdev_dirty(vd, VDD_METASLAB, msp, spa_syncing_txg(spa));
}
if (checkpoint) {
ASSERT(spa_has_checkpoint(spa));
range_tree_add(msp->ms_checkpointing, offset, asize);
} else {
range_tree_add(msp->ms_freeing, offset, asize);
}
mutex_exit(&msp->ms_lock);
}
void
metaslab_free_impl_cb(uint64_t inner_offset, vdev_t *vd, uint64_t offset,
uint64_t size, void *arg)
{
(void) inner_offset;
boolean_t *checkpoint = arg;
ASSERT3P(checkpoint, !=, NULL);
if (vd->vdev_ops->vdev_op_remap != NULL)
vdev_indirect_mark_obsolete(vd, offset, size);
else
metaslab_free_impl(vd, offset, size, *checkpoint);
}
static void
metaslab_free_impl(vdev_t *vd, uint64_t offset, uint64_t size,
boolean_t checkpoint)
{
spa_t *spa = vd->vdev_spa;
ASSERT3U(spa_config_held(spa, SCL_ALL, RW_READER), !=, 0);
if (spa_syncing_txg(spa) > spa_freeze_txg(spa))
return;
if (spa->spa_vdev_removal != NULL &&
spa->spa_vdev_removal->svr_vdev_id == vd->vdev_id &&
vdev_is_concrete(vd)) {
/*
* Note: we check if the vdev is concrete because when
* we complete the removal, we first change the vdev to be
* an indirect vdev (in open context), and then (in syncing
* context) clear spa_vdev_removal.
*/
free_from_removing_vdev(vd, offset, size);
} else if (vd->vdev_ops->vdev_op_remap != NULL) {
vdev_indirect_mark_obsolete(vd, offset, size);
vd->vdev_ops->vdev_op_remap(vd, offset, size,
metaslab_free_impl_cb, &checkpoint);
} else {
metaslab_free_concrete(vd, offset, size, checkpoint);
}
}
typedef struct remap_blkptr_cb_arg {
blkptr_t *rbca_bp;
spa_remap_cb_t rbca_cb;
vdev_t *rbca_remap_vd;
uint64_t rbca_remap_offset;
void *rbca_cb_arg;
} remap_blkptr_cb_arg_t;
static void
remap_blkptr_cb(uint64_t inner_offset, vdev_t *vd, uint64_t offset,
uint64_t size, void *arg)
{
remap_blkptr_cb_arg_t *rbca = arg;
blkptr_t *bp = rbca->rbca_bp;
/* We can not remap split blocks. */
if (size != DVA_GET_ASIZE(&bp->blk_dva[0]))
return;
ASSERT0(inner_offset);
if (rbca->rbca_cb != NULL) {
/*
* At this point we know that we are not handling split
* blocks and we invoke the callback on the previous
* vdev which must be indirect.
*/
ASSERT3P(rbca->rbca_remap_vd->vdev_ops, ==, &vdev_indirect_ops);
rbca->rbca_cb(rbca->rbca_remap_vd->vdev_id,
rbca->rbca_remap_offset, size, rbca->rbca_cb_arg);
/* set up remap_blkptr_cb_arg for the next call */
rbca->rbca_remap_vd = vd;
rbca->rbca_remap_offset = offset;
}
/*
* The phys birth time is that of dva[0]. This ensures that we know
* when each dva was written, so that resilver can determine which
* blocks need to be scrubbed (i.e. those written during the time
* the vdev was offline). It also ensures that the key used in
* the ARC hash table is unique (i.e. dva[0] + phys_birth). If
* we didn't change the phys_birth, a lookup in the ARC for a
* remapped BP could find the data that was previously stored at
* this vdev + offset.
*/
vdev_t *oldvd = vdev_lookup_top(vd->vdev_spa,
DVA_GET_VDEV(&bp->blk_dva[0]));
vdev_indirect_births_t *vib = oldvd->vdev_indirect_births;
bp->blk_phys_birth = vdev_indirect_births_physbirth(vib,
DVA_GET_OFFSET(&bp->blk_dva[0]), DVA_GET_ASIZE(&bp->blk_dva[0]));
DVA_SET_VDEV(&bp->blk_dva[0], vd->vdev_id);
DVA_SET_OFFSET(&bp->blk_dva[0], offset);
}
/*
* If the block pointer contains any indirect DVAs, modify them to refer to
* concrete DVAs. Note that this will sometimes not be possible, leaving
* the indirect DVA in place. This happens if the indirect DVA spans multiple
* segments in the mapping (i.e. it is a "split block").
*
* If the BP was remapped, calls the callback on the original dva (note the
* callback can be called multiple times if the original indirect DVA refers
* to another indirect DVA, etc).
*
* Returns TRUE if the BP was remapped.
*/
boolean_t
spa_remap_blkptr(spa_t *spa, blkptr_t *bp, spa_remap_cb_t callback, void *arg)
{
remap_blkptr_cb_arg_t rbca;
if (!zfs_remap_blkptr_enable)
return (B_FALSE);
if (!spa_feature_is_enabled(spa, SPA_FEATURE_OBSOLETE_COUNTS))
return (B_FALSE);
/*
* Dedup BP's can not be remapped, because ddt_phys_select() depends
* on DVA[0] being the same in the BP as in the DDT (dedup table).
*/
if (BP_GET_DEDUP(bp))
return (B_FALSE);
/*
* Gang blocks can not be remapped, because
* zio_checksum_gang_verifier() depends on the DVA[0] that's in
* the BP used to read the gang block header (GBH) being the same
* as the DVA[0] that we allocated for the GBH.
*/
if (BP_IS_GANG(bp))
return (B_FALSE);
/*
* Embedded BP's have no DVA to remap.
*/
if (BP_GET_NDVAS(bp) < 1)
return (B_FALSE);
/*
* Note: we only remap dva[0]. If we remapped other dvas, we
* would no longer know what their phys birth txg is.
*/
dva_t *dva = &bp->blk_dva[0];
uint64_t offset = DVA_GET_OFFSET(dva);
uint64_t size = DVA_GET_ASIZE(dva);
vdev_t *vd = vdev_lookup_top(spa, DVA_GET_VDEV(dva));
if (vd->vdev_ops->vdev_op_remap == NULL)
return (B_FALSE);
rbca.rbca_bp = bp;
rbca.rbca_cb = callback;
rbca.rbca_remap_vd = vd;
rbca.rbca_remap_offset = offset;
rbca.rbca_cb_arg = arg;
/*
* remap_blkptr_cb() will be called in order for each level of
* indirection, until a concrete vdev is reached or a split block is
* encountered. old_vd and old_offset are updated within the callback
* as we go from the one indirect vdev to the next one (either concrete
* or indirect again) in that order.
*/
vd->vdev_ops->vdev_op_remap(vd, offset, size, remap_blkptr_cb, &rbca);
/* Check if the DVA wasn't remapped because it is a split block */
if (DVA_GET_VDEV(&rbca.rbca_bp->blk_dva[0]) == vd->vdev_id)
return (B_FALSE);
return (B_TRUE);
}
/*
* Undo the allocation of a DVA which happened in the given transaction group.
*/
void
metaslab_unalloc_dva(spa_t *spa, const dva_t *dva, uint64_t txg)
{
metaslab_t *msp;
vdev_t *vd;
uint64_t vdev = DVA_GET_VDEV(dva);
uint64_t offset = DVA_GET_OFFSET(dva);
uint64_t size = DVA_GET_ASIZE(dva);
ASSERT(DVA_IS_VALID(dva));
ASSERT3U(spa_config_held(spa, SCL_ALL, RW_READER), !=, 0);
if (txg > spa_freeze_txg(spa))
return;
if ((vd = vdev_lookup_top(spa, vdev)) == NULL || !DVA_IS_VALID(dva) ||
(offset >> vd->vdev_ms_shift) >= vd->vdev_ms_count) {
zfs_panic_recover("metaslab_free_dva(): bad DVA %llu:%llu:%llu",
(u_longlong_t)vdev, (u_longlong_t)offset,
(u_longlong_t)size);
return;
}
ASSERT(!vd->vdev_removing);
ASSERT(vdev_is_concrete(vd));
ASSERT0(vd->vdev_indirect_config.vic_mapping_object);
ASSERT3P(vd->vdev_indirect_mapping, ==, NULL);
if (DVA_GET_GANG(dva))
size = vdev_gang_header_asize(vd);
msp = vd->vdev_ms[offset >> vd->vdev_ms_shift];
mutex_enter(&msp->ms_lock);
range_tree_remove(msp->ms_allocating[txg & TXG_MASK],
offset, size);
msp->ms_allocating_total -= size;
VERIFY(!msp->ms_condensing);
VERIFY3U(offset, >=, msp->ms_start);
VERIFY3U(offset + size, <=, msp->ms_start + msp->ms_size);
VERIFY3U(range_tree_space(msp->ms_allocatable) + size, <=,
msp->ms_size);
VERIFY0(P2PHASE(offset, 1ULL << vd->vdev_ashift));
VERIFY0(P2PHASE(size, 1ULL << vd->vdev_ashift));
range_tree_add(msp->ms_allocatable, offset, size);
mutex_exit(&msp->ms_lock);
}
/*
* Free the block represented by the given DVA.
*/
void
metaslab_free_dva(spa_t *spa, const dva_t *dva, boolean_t checkpoint)
{
uint64_t vdev = DVA_GET_VDEV(dva);
uint64_t offset = DVA_GET_OFFSET(dva);
uint64_t size = DVA_GET_ASIZE(dva);
vdev_t *vd = vdev_lookup_top(spa, vdev);
ASSERT(DVA_IS_VALID(dva));
ASSERT3U(spa_config_held(spa, SCL_ALL, RW_READER), !=, 0);
if (DVA_GET_GANG(dva)) {
size = vdev_gang_header_asize(vd);
}
metaslab_free_impl(vd, offset, size, checkpoint);
}
/*
* Reserve some allocation slots. The reservation system must be called
* before we call into the allocator. If there aren't any available slots
* then the I/O will be throttled until an I/O completes and its slots are
* freed up. The function returns true if it was successful in placing
* the reservation.
*/
boolean_t
metaslab_class_throttle_reserve(metaslab_class_t *mc, int slots, int allocator,
zio_t *zio, int flags)
{
metaslab_class_allocator_t *mca = &mc->mc_allocator[allocator];
uint64_t max = mca->mca_alloc_max_slots;
ASSERT(mc->mc_alloc_throttle_enabled);
if (GANG_ALLOCATION(flags) || (flags & METASLAB_MUST_RESERVE) ||
zfs_refcount_count(&mca->mca_alloc_slots) + slots <= max) {
/*
* The potential race between _count() and _add() is covered
* by the allocator lock in most cases, or irrelevant due to
* GANG_ALLOCATION() or METASLAB_MUST_RESERVE set in others.
* But even if we assume some other non-existing scenario, the
* worst that can happen is few more I/Os get to allocation
* earlier, that is not a problem.
*
* We reserve the slots individually so that we can unreserve
* them individually when an I/O completes.
*/
zfs_refcount_add_few(&mca->mca_alloc_slots, slots, zio);
zio->io_flags |= ZIO_FLAG_IO_ALLOCATING;
return (B_TRUE);
}
return (B_FALSE);
}
void
metaslab_class_throttle_unreserve(metaslab_class_t *mc, int slots,
int allocator, zio_t *zio)
{
metaslab_class_allocator_t *mca = &mc->mc_allocator[allocator];
ASSERT(mc->mc_alloc_throttle_enabled);
zfs_refcount_remove_few(&mca->mca_alloc_slots, slots, zio);
}
static int
metaslab_claim_concrete(vdev_t *vd, uint64_t offset, uint64_t size,
uint64_t txg)
{
metaslab_t *msp;
spa_t *spa = vd->vdev_spa;
int error = 0;
if (offset >> vd->vdev_ms_shift >= vd->vdev_ms_count)
return (SET_ERROR(ENXIO));
ASSERT3P(vd->vdev_ms, !=, NULL);
msp = vd->vdev_ms[offset >> vd->vdev_ms_shift];
mutex_enter(&msp->ms_lock);
if ((txg != 0 && spa_writeable(spa)) || !msp->ms_loaded) {
error = metaslab_activate(msp, 0, METASLAB_WEIGHT_CLAIM);
if (error == EBUSY) {
ASSERT(msp->ms_loaded);
ASSERT(msp->ms_weight & METASLAB_ACTIVE_MASK);
error = 0;
}
}
if (error == 0 &&
!range_tree_contains(msp->ms_allocatable, offset, size))
error = SET_ERROR(ENOENT);
if (error || txg == 0) { /* txg == 0 indicates dry run */
mutex_exit(&msp->ms_lock);
return (error);
}
VERIFY(!msp->ms_condensing);
VERIFY0(P2PHASE(offset, 1ULL << vd->vdev_ashift));
VERIFY0(P2PHASE(size, 1ULL << vd->vdev_ashift));
VERIFY3U(range_tree_space(msp->ms_allocatable) - size, <=,
msp->ms_size);
range_tree_remove(msp->ms_allocatable, offset, size);
range_tree_clear(msp->ms_trim, offset, size);
if (spa_writeable(spa)) { /* don't dirty if we're zdb(8) */
metaslab_class_t *mc = msp->ms_group->mg_class;
multilist_sublist_t *mls =
multilist_sublist_lock_obj(&mc->mc_metaslab_txg_list, msp);
if (!multilist_link_active(&msp->ms_class_txg_node)) {
msp->ms_selected_txg = txg;
multilist_sublist_insert_head(mls, msp);
}
multilist_sublist_unlock(mls);
if (range_tree_is_empty(msp->ms_allocating[txg & TXG_MASK]))
vdev_dirty(vd, VDD_METASLAB, msp, txg);
range_tree_add(msp->ms_allocating[txg & TXG_MASK],
offset, size);
msp->ms_allocating_total += size;
}
mutex_exit(&msp->ms_lock);
return (0);
}
typedef struct metaslab_claim_cb_arg_t {
uint64_t mcca_txg;
int mcca_error;
} metaslab_claim_cb_arg_t;
static void
metaslab_claim_impl_cb(uint64_t inner_offset, vdev_t *vd, uint64_t offset,
uint64_t size, void *arg)
{
(void) inner_offset;
metaslab_claim_cb_arg_t *mcca_arg = arg;
if (mcca_arg->mcca_error == 0) {
mcca_arg->mcca_error = metaslab_claim_concrete(vd, offset,
size, mcca_arg->mcca_txg);
}
}
int
metaslab_claim_impl(vdev_t *vd, uint64_t offset, uint64_t size, uint64_t txg)
{
if (vd->vdev_ops->vdev_op_remap != NULL) {
metaslab_claim_cb_arg_t arg;
/*
* Only zdb(8) can claim on indirect vdevs. This is used
* to detect leaks of mapped space (that are not accounted
* for in the obsolete counts, spacemap, or bpobj).
*/
ASSERT(!spa_writeable(vd->vdev_spa));
arg.mcca_error = 0;
arg.mcca_txg = txg;
vd->vdev_ops->vdev_op_remap(vd, offset, size,
metaslab_claim_impl_cb, &arg);
if (arg.mcca_error == 0) {
arg.mcca_error = metaslab_claim_concrete(vd,
offset, size, txg);
}
return (arg.mcca_error);
} else {
return (metaslab_claim_concrete(vd, offset, size, txg));
}
}
/*
* Intent log support: upon opening the pool after a crash, notify the SPA
* of blocks that the intent log has allocated for immediate write, but
* which are still considered free by the SPA because the last transaction
* group didn't commit yet.
*/
static int
metaslab_claim_dva(spa_t *spa, const dva_t *dva, uint64_t txg)
{
uint64_t vdev = DVA_GET_VDEV(dva);
uint64_t offset = DVA_GET_OFFSET(dva);
uint64_t size = DVA_GET_ASIZE(dva);
vdev_t *vd;
if ((vd = vdev_lookup_top(spa, vdev)) == NULL) {
return (SET_ERROR(ENXIO));
}
ASSERT(DVA_IS_VALID(dva));
if (DVA_GET_GANG(dva))
size = vdev_gang_header_asize(vd);
return (metaslab_claim_impl(vd, offset, size, txg));
}
int
metaslab_alloc(spa_t *spa, metaslab_class_t *mc, uint64_t psize, blkptr_t *bp,
int ndvas, uint64_t txg, blkptr_t *hintbp, int flags,
zio_alloc_list_t *zal, zio_t *zio, int allocator)
{
dva_t *dva = bp->blk_dva;
dva_t *hintdva = (hintbp != NULL) ? hintbp->blk_dva : NULL;
int error = 0;
ASSERT(bp->blk_birth == 0);
ASSERT(BP_PHYSICAL_BIRTH(bp) == 0);
spa_config_enter(spa, SCL_ALLOC, FTAG, RW_READER);
if (mc->mc_allocator[allocator].mca_rotor == NULL) {
/* no vdevs in this class */
spa_config_exit(spa, SCL_ALLOC, FTAG);
return (SET_ERROR(ENOSPC));
}
ASSERT(ndvas > 0 && ndvas <= spa_max_replication(spa));
ASSERT(BP_GET_NDVAS(bp) == 0);
ASSERT(hintbp == NULL || ndvas <= BP_GET_NDVAS(hintbp));
ASSERT3P(zal, !=, NULL);
for (int d = 0; d < ndvas; d++) {
error = metaslab_alloc_dva(spa, mc, psize, dva, d, hintdva,
txg, flags, zal, allocator);
if (error != 0) {
for (d--; d >= 0; d--) {
metaslab_unalloc_dva(spa, &dva[d], txg);
metaslab_group_alloc_decrement(spa,
DVA_GET_VDEV(&dva[d]), zio, flags,
allocator, B_FALSE);
memset(&dva[d], 0, sizeof (dva_t));
}
spa_config_exit(spa, SCL_ALLOC, FTAG);
return (error);
} else {
/*
* Update the metaslab group's queue depth
* based on the newly allocated dva.
*/
metaslab_group_alloc_increment(spa,
DVA_GET_VDEV(&dva[d]), zio, flags, allocator);
}
}
ASSERT(error == 0);
ASSERT(BP_GET_NDVAS(bp) == ndvas);
spa_config_exit(spa, SCL_ALLOC, FTAG);
BP_SET_BIRTH(bp, txg, 0);
return (0);
}
void
metaslab_free(spa_t *spa, const blkptr_t *bp, uint64_t txg, boolean_t now)
{
const dva_t *dva = bp->blk_dva;
int ndvas = BP_GET_NDVAS(bp);
ASSERT(!BP_IS_HOLE(bp));
ASSERT(!now || bp->blk_birth >= spa_syncing_txg(spa));
/*
* If we have a checkpoint for the pool we need to make sure that
* the blocks that we free that are part of the checkpoint won't be
* reused until the checkpoint is discarded or we revert to it.
*
* The checkpoint flag is passed down the metaslab_free code path
* and is set whenever we want to add a block to the checkpoint's
* accounting. That is, we "checkpoint" blocks that existed at the
* time the checkpoint was created and are therefore referenced by
* the checkpointed uberblock.
*
* Note that, we don't checkpoint any blocks if the current
* syncing txg <= spa_checkpoint_txg. We want these frees to sync
* normally as they will be referenced by the checkpointed uberblock.
*/
boolean_t checkpoint = B_FALSE;
if (bp->blk_birth <= spa->spa_checkpoint_txg &&
spa_syncing_txg(spa) > spa->spa_checkpoint_txg) {
/*
* At this point, if the block is part of the checkpoint
* there is no way it was created in the current txg.
*/
ASSERT(!now);
ASSERT3U(spa_syncing_txg(spa), ==, txg);
checkpoint = B_TRUE;
}
spa_config_enter(spa, SCL_FREE, FTAG, RW_READER);
for (int d = 0; d < ndvas; d++) {
if (now) {
metaslab_unalloc_dva(spa, &dva[d], txg);
} else {
ASSERT3U(txg, ==, spa_syncing_txg(spa));
metaslab_free_dva(spa, &dva[d], checkpoint);
}
}
spa_config_exit(spa, SCL_FREE, FTAG);
}
int
metaslab_claim(spa_t *spa, const blkptr_t *bp, uint64_t txg)
{
const dva_t *dva = bp->blk_dva;
int ndvas = BP_GET_NDVAS(bp);
int error = 0;
ASSERT(!BP_IS_HOLE(bp));
if (txg != 0) {
/*
* First do a dry run to make sure all DVAs are claimable,
* so we don't have to unwind from partial failures below.
*/
if ((error = metaslab_claim(spa, bp, 0)) != 0)
return (error);
}
spa_config_enter(spa, SCL_ALLOC, FTAG, RW_READER);
for (int d = 0; d < ndvas; d++) {
error = metaslab_claim_dva(spa, &dva[d], txg);
if (error != 0)
break;
}
spa_config_exit(spa, SCL_ALLOC, FTAG);
ASSERT(error == 0 || txg == 0);
return (error);
}
static void
metaslab_check_free_impl_cb(uint64_t inner, vdev_t *vd, uint64_t offset,
uint64_t size, void *arg)
{
(void) inner, (void) arg;
if (vd->vdev_ops == &vdev_indirect_ops)
return;
metaslab_check_free_impl(vd, offset, size);
}
static void
metaslab_check_free_impl(vdev_t *vd, uint64_t offset, uint64_t size)
{
metaslab_t *msp;
spa_t *spa __maybe_unused = vd->vdev_spa;
if ((zfs_flags & ZFS_DEBUG_ZIO_FREE) == 0)
return;
if (vd->vdev_ops->vdev_op_remap != NULL) {
vd->vdev_ops->vdev_op_remap(vd, offset, size,
metaslab_check_free_impl_cb, NULL);
return;
}
ASSERT(vdev_is_concrete(vd));
ASSERT3U(offset >> vd->vdev_ms_shift, <, vd->vdev_ms_count);
ASSERT3U(spa_config_held(spa, SCL_ALL, RW_READER), !=, 0);
msp = vd->vdev_ms[offset >> vd->vdev_ms_shift];
mutex_enter(&msp->ms_lock);
if (msp->ms_loaded) {
range_tree_verify_not_present(msp->ms_allocatable,
offset, size);
}
/*
* Check all segments that currently exist in the freeing pipeline.
*
* It would intuitively make sense to also check the current allocating
* tree since metaslab_unalloc_dva() exists for extents that are
* allocated and freed in the same sync pass within the same txg.
* Unfortunately there are places (e.g. the ZIL) where we allocate a
* segment but then we free part of it within the same txg
* [see zil_sync()]. Thus, we don't call range_tree_verify() in the
* current allocating tree.
*/
range_tree_verify_not_present(msp->ms_freeing, offset, size);
range_tree_verify_not_present(msp->ms_checkpointing, offset, size);
range_tree_verify_not_present(msp->ms_freed, offset, size);
for (int j = 0; j < TXG_DEFER_SIZE; j++)
range_tree_verify_not_present(msp->ms_defer[j], offset, size);
range_tree_verify_not_present(msp->ms_trim, offset, size);
mutex_exit(&msp->ms_lock);
}
void
metaslab_check_free(spa_t *spa, const blkptr_t *bp)
{
if ((zfs_flags & ZFS_DEBUG_ZIO_FREE) == 0)
return;
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
for (int i = 0; i < BP_GET_NDVAS(bp); i++) {
uint64_t vdev = DVA_GET_VDEV(&bp->blk_dva[i]);
vdev_t *vd = vdev_lookup_top(spa, vdev);
uint64_t offset = DVA_GET_OFFSET(&bp->blk_dva[i]);
uint64_t size = DVA_GET_ASIZE(&bp->blk_dva[i]);
if (DVA_GET_GANG(&bp->blk_dva[i]))
size = vdev_gang_header_asize(vd);
ASSERT3P(vd, !=, NULL);
metaslab_check_free_impl(vd, offset, size);
}
spa_config_exit(spa, SCL_VDEV, FTAG);
}
static void
metaslab_group_disable_wait(metaslab_group_t *mg)
{
ASSERT(MUTEX_HELD(&mg->mg_ms_disabled_lock));
while (mg->mg_disabled_updating) {
cv_wait(&mg->mg_ms_disabled_cv, &mg->mg_ms_disabled_lock);
}
}
static void
metaslab_group_disabled_increment(metaslab_group_t *mg)
{
ASSERT(MUTEX_HELD(&mg->mg_ms_disabled_lock));
ASSERT(mg->mg_disabled_updating);
while (mg->mg_ms_disabled >= max_disabled_ms) {
cv_wait(&mg->mg_ms_disabled_cv, &mg->mg_ms_disabled_lock);
}
mg->mg_ms_disabled++;
ASSERT3U(mg->mg_ms_disabled, <=, max_disabled_ms);
}
/*
* Mark the metaslab as disabled to prevent any allocations on this metaslab.
* We must also track how many metaslabs are currently disabled within a
* metaslab group and limit them to prevent allocation failures from
* occurring because all metaslabs are disabled.
*/
void
metaslab_disable(metaslab_t *msp)
{
ASSERT(!MUTEX_HELD(&msp->ms_lock));
metaslab_group_t *mg = msp->ms_group;
mutex_enter(&mg->mg_ms_disabled_lock);
/*
* To keep an accurate count of how many threads have disabled
* a specific metaslab group, we only allow one thread to mark
* the metaslab group at a time. This ensures that the value of
* ms_disabled will be accurate when we decide to mark a metaslab
* group as disabled. To do this we force all other threads
* to wait till the metaslab's mg_disabled_updating flag is no
* longer set.
*/
metaslab_group_disable_wait(mg);
mg->mg_disabled_updating = B_TRUE;
if (msp->ms_disabled == 0) {
metaslab_group_disabled_increment(mg);
}
mutex_enter(&msp->ms_lock);
msp->ms_disabled++;
mutex_exit(&msp->ms_lock);
mg->mg_disabled_updating = B_FALSE;
cv_broadcast(&mg->mg_ms_disabled_cv);
mutex_exit(&mg->mg_ms_disabled_lock);
}
void
metaslab_enable(metaslab_t *msp, boolean_t sync, boolean_t unload)
{
metaslab_group_t *mg = msp->ms_group;
spa_t *spa = mg->mg_vd->vdev_spa;
/*
* Wait for the outstanding IO to be synced to prevent newly
* allocated blocks from being overwritten. This used by
* initialize and TRIM which are modifying unallocated space.
*/
if (sync)
txg_wait_synced(spa_get_dsl(spa), 0);
mutex_enter(&mg->mg_ms_disabled_lock);
mutex_enter(&msp->ms_lock);
if (--msp->ms_disabled == 0) {
mg->mg_ms_disabled--;
cv_broadcast(&mg->mg_ms_disabled_cv);
if (unload)
metaslab_unload(msp);
}
mutex_exit(&msp->ms_lock);
mutex_exit(&mg->mg_ms_disabled_lock);
}
void
metaslab_set_unflushed_dirty(metaslab_t *ms, boolean_t dirty)
{
ms->ms_unflushed_dirty = dirty;
}
static void
metaslab_update_ondisk_flush_data(metaslab_t *ms, dmu_tx_t *tx)
{
vdev_t *vd = ms->ms_group->mg_vd;
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa_meta_objset(spa);
ASSERT(spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP));
metaslab_unflushed_phys_t entry = {
.msp_unflushed_txg = metaslab_unflushed_txg(ms),
};
uint64_t entry_size = sizeof (entry);
uint64_t entry_offset = ms->ms_id * entry_size;
uint64_t object = 0;
int err = zap_lookup(mos, vd->vdev_top_zap,
VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS, sizeof (uint64_t), 1,
&object);
if (err == ENOENT) {
object = dmu_object_alloc(mos, DMU_OTN_UINT64_METADATA,
SPA_OLD_MAXBLOCKSIZE, DMU_OT_NONE, 0, tx);
VERIFY0(zap_add(mos, vd->vdev_top_zap,
VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS, sizeof (uint64_t), 1,
&object, tx));
} else {
VERIFY0(err);
}
dmu_write(spa_meta_objset(spa), object, entry_offset, entry_size,
&entry, tx);
}
void
metaslab_set_unflushed_txg(metaslab_t *ms, uint64_t txg, dmu_tx_t *tx)
{
ms->ms_unflushed_txg = txg;
metaslab_update_ondisk_flush_data(ms, tx);
}
boolean_t
metaslab_unflushed_dirty(metaslab_t *ms)
{
return (ms->ms_unflushed_dirty);
}
uint64_t
metaslab_unflushed_txg(metaslab_t *ms)
{
return (ms->ms_unflushed_txg);
}
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, aliquot, U64, ZMOD_RW,
"Allocation granularity (a.k.a. stripe size)");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, debug_load, INT, ZMOD_RW,
"Load all metaslabs when pool is first opened");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, debug_unload, INT, ZMOD_RW,
"Prevent metaslabs from being unloaded");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, preload_enabled, INT, ZMOD_RW,
"Preload potential metaslabs during reassessment");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, preload_limit, UINT, ZMOD_RW,
"Max number of metaslabs per group to preload");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, unload_delay, UINT, ZMOD_RW,
"Delay in txgs after metaslab was last used before unloading");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, unload_delay_ms, UINT, ZMOD_RW,
"Delay in milliseconds after metaslab was last used before unloading");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs_mg, zfs_mg_, noalloc_threshold, UINT, ZMOD_RW,
"Percentage of metaslab group size that should be free to make it "
"eligible for allocation");
ZFS_MODULE_PARAM(zfs_mg, zfs_mg_, fragmentation_threshold, UINT, ZMOD_RW,
"Percentage of metaslab group size that should be considered eligible "
"for allocations unless all metaslab groups within the metaslab class "
"have also crossed this threshold");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, fragmentation_factor_enabled, INT,
ZMOD_RW,
"Use the fragmentation metric to prefer less fragmented metaslabs");
/* END CSTYLED */
ZFS_MODULE_PARAM(zfs_metaslab, zfs_metaslab_, fragmentation_threshold, UINT,
ZMOD_RW, "Fragmentation for metaslab to allow allocation");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, lba_weighting_enabled, INT, ZMOD_RW,
"Prefer metaslabs with lower LBAs");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, bias_enabled, INT, ZMOD_RW,
"Enable metaslab group biasing");
ZFS_MODULE_PARAM(zfs_metaslab, zfs_metaslab_, segment_weight_enabled, INT,
ZMOD_RW, "Enable segment-based metaslab selection");
ZFS_MODULE_PARAM(zfs_metaslab, zfs_metaslab_, switch_threshold, INT, ZMOD_RW,
"Segment-based metaslab selection maximum buckets before switching");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, force_ganging, U64, ZMOD_RW,
"Blocks larger than this size are sometimes forced to be gang blocks");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, force_ganging_pct, UINT, ZMOD_RW,
"Percentage of large blocks that will be forced to be gang blocks");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, df_max_search, UINT, ZMOD_RW,
"Max distance (bytes) to search forward before using size tree");
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, df_use_largest_segment, INT, ZMOD_RW,
"When looking in size tree, use largest segment instead of exact fit");
ZFS_MODULE_PARAM(zfs_metaslab, zfs_metaslab_, max_size_cache_sec, U64,
ZMOD_RW, "How long to trust the cached max chunk size of a metaslab");
ZFS_MODULE_PARAM(zfs_metaslab, zfs_metaslab_, mem_limit, UINT, ZMOD_RW,
"Percentage of memory that can be used to store metaslab range trees");
ZFS_MODULE_PARAM(zfs_metaslab, zfs_metaslab_, try_hard_before_gang, INT,
ZMOD_RW, "Try hard to allocate before ganging");
ZFS_MODULE_PARAM(zfs_metaslab, zfs_metaslab_, find_max_tries, UINT, ZMOD_RW,
"Normally only consider this many of the best metaslabs in each vdev");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM_CALL(zfs, zfs_, active_allocator,
param_set_active_allocator, param_get_charp, ZMOD_RW,
"SPA active allocator");
/* END CSTYLED */
diff --git a/sys/contrib/openzfs/module/zfs/spa.c b/sys/contrib/openzfs/module/zfs/spa.c
index 68f367c1c744..20225640f8c5 100644
--- a/sys/contrib/openzfs/module/zfs/spa.c
+++ b/sys/contrib/openzfs/module/zfs/spa.c
@@ -1,10360 +1,10508 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright (c) 2018, Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
* Copyright 2013 Saso Kiselkov. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright 2016 Toomas Soome <tsoome@me.com>
* Copyright (c) 2016 Actifio, Inc. All rights reserved.
* Copyright 2018 Joyent, Inc.
* Copyright (c) 2017, 2019, Datto Inc. All rights reserved.
* Copyright 2017 Joyent, Inc.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
* Copyright (c) 2023 Hewlett Packard Enterprise Development LP.
*/
/*
* SPA: Storage Pool Allocator
*
* This file contains all the routines used when modifying on-disk SPA state.
* This includes opening, importing, destroying, exporting a pool, and syncing a
* pool.
*/
#include <sys/zfs_context.h>
#include <sys/fm/fs/zfs.h>
#include <sys/spa_impl.h>
#include <sys/zio.h>
#include <sys/zio_checksum.h>
#include <sys/dmu.h>
#include <sys/dmu_tx.h>
#include <sys/zap.h>
#include <sys/zil.h>
#include <sys/brt.h>
#include <sys/ddt.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_removal.h>
#include <sys/vdev_indirect_mapping.h>
#include <sys/vdev_indirect_births.h>
#include <sys/vdev_initialize.h>
#include <sys/vdev_rebuild.h>
#include <sys/vdev_trim.h>
#include <sys/vdev_disk.h>
+#include <sys/vdev_raidz.h>
#include <sys/vdev_draid.h>
#include <sys/metaslab.h>
#include <sys/metaslab_impl.h>
#include <sys/mmp.h>
#include <sys/uberblock_impl.h>
#include <sys/txg.h>
#include <sys/avl.h>
#include <sys/bpobj.h>
#include <sys/dmu_traverse.h>
#include <sys/dmu_objset.h>
#include <sys/unique.h>
#include <sys/dsl_pool.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_prop.h>
#include <sys/dsl_synctask.h>
#include <sys/fs/zfs.h>
#include <sys/arc.h>
#include <sys/callb.h>
#include <sys/systeminfo.h>
#include <sys/zfs_ioctl.h>
#include <sys/dsl_scan.h>
#include <sys/zfeature.h>
#include <sys/dsl_destroy.h>
#include <sys/zvol.h>
#ifdef _KERNEL
#include <sys/fm/protocol.h>
#include <sys/fm/util.h>
#include <sys/callb.h>
#include <sys/zone.h>
#include <sys/vmsystm.h>
#endif /* _KERNEL */
#include "zfs_prop.h"
#include "zfs_comutil.h"
#include <cityhash.h>
/*
* spa_thread() existed on Illumos as a parent thread for the various worker
* threads that actually run the pool, as a way to both reference the entire
* pool work as a single object, and to share properties like scheduling
* options. It has not yet been adapted to Linux or FreeBSD. This define is
* used to mark related parts of the code to make things easier for the reader,
* and to compile this code out. It can be removed when someone implements it,
* moves it to some Illumos-specific place, or removes it entirely.
*/
#undef HAVE_SPA_THREAD
/*
* The "System Duty Cycle" scheduling class is an Illumos feature to help
* prevent CPU-intensive kernel threads from affecting latency on interactive
* threads. It doesn't exist on Linux or FreeBSD, so the supporting code is
* gated behind a define. On Illumos SDC depends on spa_thread(), but
* spa_thread() also has other uses, so this is a separate define.
*/
#undef HAVE_SYSDC
/*
* 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_SCALE, /* Taskqs scale with CPUs. */
ZTI_MODE_SYNC, /* sync thread assigned */
ZTI_MODE_NULL, /* don't create a taskq */
ZTI_NMODES
} zti_modes_t;
#define ZTI_P(n, q) { ZTI_MODE_FIXED, (n), (q) }
#define ZTI_PCT(n) { ZTI_MODE_ONLINE_PERCENT, (n), 1 }
#define ZTI_SCALE { ZTI_MODE_SCALE, 0, 1 }
#define ZTI_SYNC { ZTI_MODE_SYNC, 0, 1 }
#define ZTI_NULL { ZTI_MODE_NULL, 0, 0 }
#define ZTI_N(n) ZTI_P(n, 1)
#define ZTI_ONE ZTI_N(1)
typedef struct zio_taskq_info {
zti_modes_t zti_mode;
uint_t zti_value;
uint_t zti_count;
} zio_taskq_info_t;
static const char *const zio_taskq_types[ZIO_TASKQ_TYPES] = {
"iss", "iss_h", "int", "int_h"
};
/*
* This table defines the taskq settings for each ZFS I/O type. When
* initializing a pool, we use this table to create an appropriately sized
* taskq. Some operations are low volume and therefore have a small, static
* number of threads assigned to their taskqs using the ZTI_N(#) or ZTI_ONE
* macros. Other operations process a large amount of data; the ZTI_SCALE
* macro causes us to create a taskq oriented for throughput. Some operations
* are so high frequency and short-lived that the taskq itself can become a
* point of lock contention. The ZTI_P(#, #) macro indicates that we need an
* additional degree of parallelism specified by the number of threads per-
* taskq and the number of taskqs; when dispatching an event in this case, the
* particular taskq is chosen at random. ZTI_SCALE uses a number of taskqs
* that scales with the number of CPUs.
*
* The different taskq priorities are to handle the different contexts (issue
* and interrupt) and then to reserve threads for ZIO_PRIORITY_NOW I/Os that
* need to be handled with minimum delay.
*/
static const zio_taskq_info_t zio_taskqs[ZIO_TYPES][ZIO_TASKQ_TYPES] = {
/* ISSUE ISSUE_HIGH INTR INTR_HIGH */
{ ZTI_ONE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* NULL */
{ ZTI_N(8), ZTI_NULL, ZTI_SCALE, ZTI_NULL }, /* READ */
{ ZTI_SYNC, ZTI_N(5), ZTI_SCALE, ZTI_N(5) }, /* WRITE */
{ ZTI_SCALE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* FREE */
{ ZTI_ONE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* CLAIM */
{ ZTI_ONE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* IOCTL */
{ ZTI_N(4), ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* TRIM */
};
static void spa_sync_version(void *arg, dmu_tx_t *tx);
static void spa_sync_props(void *arg, dmu_tx_t *tx);
static boolean_t spa_has_active_shared_spare(spa_t *spa);
static int spa_load_impl(spa_t *spa, spa_import_type_t type,
const char **ereport);
static void spa_vdev_resilver_done(spa_t *spa);
/*
* Percentage of all CPUs that can be used by the metaslab preload taskq.
*/
static uint_t metaslab_preload_pct = 50;
static uint_t zio_taskq_batch_pct = 80; /* 1 thread per cpu in pset */
static uint_t zio_taskq_batch_tpq; /* threads per taskq */
#ifdef HAVE_SYSDC
static const boolean_t zio_taskq_sysdc = B_TRUE; /* use SDC scheduling class */
static const uint_t zio_taskq_basedc = 80; /* base duty cycle */
#endif
#ifdef HAVE_SPA_THREAD
static const boolean_t spa_create_process = B_TRUE; /* no process => no sysdc */
#endif
static uint_t zio_taskq_wr_iss_ncpus = 0;
/*
* Report any spa_load_verify errors found, but do not fail spa_load.
* This is used by zdb to analyze non-idle pools.
*/
boolean_t spa_load_verify_dryrun = B_FALSE;
/*
* Allow read spacemaps in case of readonly import (spa_mode == SPA_MODE_READ).
* This is used by zdb for spacemaps verification.
*/
boolean_t spa_mode_readable_spacemaps = B_FALSE;
/*
* This (illegal) pool name is used when temporarily importing a spa_t in order
* to get the vdev stats associated with the imported devices.
*/
#define TRYIMPORT_NAME "$import"
/*
* For debugging purposes: print out vdev tree during pool import.
*/
static int spa_load_print_vdev_tree = B_FALSE;
/*
* A non-zero value for zfs_max_missing_tvds means that we allow importing
* pools with missing top-level vdevs. This is strictly intended for advanced
* pool recovery cases since missing data is almost inevitable. Pools with
* missing devices can only be imported read-only for safety reasons, and their
* fail-mode will be automatically set to "continue".
*
* With 1 missing vdev we should be able to import the pool and mount all
* datasets. User data that was not modified after the missing device has been
* added should be recoverable. This means that snapshots created prior to the
* addition of that device should be completely intact.
*
* With 2 missing vdevs, some datasets may fail to mount since there are
* dataset statistics that are stored as regular metadata. Some data might be
* recoverable if those vdevs were added recently.
*
* With 3 or more missing vdevs, the pool is severely damaged and MOS entries
* may be missing entirely. Chances of data recovery are very low. Note that
* there are also risks of performing an inadvertent rewind as we might be
* missing all the vdevs with the latest uberblocks.
*/
uint64_t zfs_max_missing_tvds = 0;
/*
* The parameters below are similar to zfs_max_missing_tvds but are only
* intended for a preliminary open of the pool with an untrusted config which
* might be incomplete or out-dated.
*
* We are more tolerant for pools opened from a cachefile since we could have
* an out-dated cachefile where a device removal was not registered.
* We could have set the limit arbitrarily high but in the case where devices
* are really missing we would want to return the proper error codes; we chose
* SPA_DVAS_PER_BP - 1 so that some copies of the MOS would still be available
* and we get a chance to retrieve the trusted config.
*/
uint64_t zfs_max_missing_tvds_cachefile = SPA_DVAS_PER_BP - 1;
/*
* In the case where config was assembled by scanning device paths (/dev/dsks
* by default) we are less tolerant since all the existing devices should have
* been detected and we want spa_load to return the right error codes.
*/
uint64_t zfs_max_missing_tvds_scan = 0;
/*
* Debugging aid that pauses spa_sync() towards the end.
*/
static const boolean_t zfs_pause_spa_sync = B_FALSE;
/*
* Variables to indicate the livelist condense zthr func should wait at certain
* points for the livelist to be removed - used to test condense/destroy races
*/
static int zfs_livelist_condense_zthr_pause = 0;
static int zfs_livelist_condense_sync_pause = 0;
/*
* Variables to track whether or not condense cancellation has been
* triggered in testing.
*/
static int zfs_livelist_condense_sync_cancel = 0;
static int zfs_livelist_condense_zthr_cancel = 0;
/*
* Variable to track whether or not extra ALLOC blkptrs were added to a
* livelist entry while it was being condensed (caused by the way we track
* remapped blkptrs in dbuf_remap_impl)
*/
static int zfs_livelist_condense_new_alloc = 0;
/*
* ==========================================================================
* SPA properties routines
* ==========================================================================
*/
/*
* Add a (source=src, propname=propval) list to an nvlist.
*/
static void
spa_prop_add_list(nvlist_t *nvl, zpool_prop_t prop, const char *strval,
uint64_t intval, zprop_source_t src)
{
const char *propname = zpool_prop_to_name(prop);
nvlist_t *propval;
propval = fnvlist_alloc();
fnvlist_add_uint64(propval, ZPROP_SOURCE, src);
if (strval != NULL)
fnvlist_add_string(propval, ZPROP_VALUE, strval);
else
fnvlist_add_uint64(propval, ZPROP_VALUE, intval);
fnvlist_add_nvlist(nvl, propname, propval);
nvlist_free(propval);
}
/*
* Add a user property (source=src, propname=propval) to an nvlist.
*/
static void
spa_prop_add_user(nvlist_t *nvl, const char *propname, char *strval,
zprop_source_t src)
{
nvlist_t *propval;
VERIFY(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP) == 0);
VERIFY(nvlist_add_uint64(propval, ZPROP_SOURCE, src) == 0);
VERIFY(nvlist_add_string(propval, ZPROP_VALUE, strval) == 0);
VERIFY(nvlist_add_nvlist(nvl, propname, propval) == 0);
nvlist_free(propval);
}
/*
* Get property values from the spa configuration.
*/
static void
spa_prop_get_config(spa_t *spa, nvlist_t **nvp)
{
vdev_t *rvd = spa->spa_root_vdev;
dsl_pool_t *pool = spa->spa_dsl_pool;
uint64_t size, alloc, cap, version;
const zprop_source_t src = ZPROP_SRC_NONE;
spa_config_dirent_t *dp;
metaslab_class_t *mc = spa_normal_class(spa);
ASSERT(MUTEX_HELD(&spa->spa_props_lock));
if (rvd != NULL) {
alloc = metaslab_class_get_alloc(mc);
alloc += metaslab_class_get_alloc(spa_special_class(spa));
alloc += metaslab_class_get_alloc(spa_dedup_class(spa));
alloc += metaslab_class_get_alloc(spa_embedded_log_class(spa));
size = metaslab_class_get_space(mc);
size += metaslab_class_get_space(spa_special_class(spa));
size += metaslab_class_get_space(spa_dedup_class(spa));
size += metaslab_class_get_space(spa_embedded_log_class(spa));
spa_prop_add_list(*nvp, ZPOOL_PROP_NAME, spa_name(spa), 0, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_SIZE, NULL, size, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_ALLOCATED, NULL, alloc, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_FREE, NULL,
size - alloc, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_CHECKPOINT, NULL,
spa->spa_checkpoint_info.sci_dspace, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_FRAGMENTATION, NULL,
metaslab_class_fragmentation(mc), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_EXPANDSZ, NULL,
metaslab_class_expandable_space(mc), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_READONLY, NULL,
(spa_mode(spa) == SPA_MODE_READ), src);
cap = (size == 0) ? 0 : (alloc * 100 / size);
spa_prop_add_list(*nvp, ZPOOL_PROP_CAPACITY, NULL, cap, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_DEDUPRATIO, NULL,
ddt_get_pool_dedup_ratio(spa), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_BCLONEUSED, NULL,
brt_get_used(spa), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_BCLONESAVED, NULL,
brt_get_saved(spa), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_BCLONERATIO, NULL,
brt_get_ratio(spa), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_HEALTH, NULL,
rvd->vdev_state, src);
version = spa_version(spa);
if (version == zpool_prop_default_numeric(ZPOOL_PROP_VERSION)) {
spa_prop_add_list(*nvp, ZPOOL_PROP_VERSION, NULL,
version, ZPROP_SRC_DEFAULT);
} else {
spa_prop_add_list(*nvp, ZPOOL_PROP_VERSION, NULL,
version, ZPROP_SRC_LOCAL);
}
spa_prop_add_list(*nvp, ZPOOL_PROP_LOAD_GUID,
NULL, spa_load_guid(spa), src);
}
if (pool != NULL) {
/*
* The $FREE directory was introduced in SPA_VERSION_DEADLISTS,
* when opening pools before this version freedir will be NULL.
*/
if (pool->dp_free_dir != NULL) {
spa_prop_add_list(*nvp, ZPOOL_PROP_FREEING, NULL,
dsl_dir_phys(pool->dp_free_dir)->dd_used_bytes,
src);
} else {
spa_prop_add_list(*nvp, ZPOOL_PROP_FREEING,
NULL, 0, src);
}
if (pool->dp_leak_dir != NULL) {
spa_prop_add_list(*nvp, ZPOOL_PROP_LEAKED, NULL,
dsl_dir_phys(pool->dp_leak_dir)->dd_used_bytes,
src);
} else {
spa_prop_add_list(*nvp, ZPOOL_PROP_LEAKED,
NULL, 0, src);
}
}
spa_prop_add_list(*nvp, ZPOOL_PROP_GUID, NULL, spa_guid(spa), src);
if (spa->spa_comment != NULL) {
spa_prop_add_list(*nvp, ZPOOL_PROP_COMMENT, spa->spa_comment,
0, ZPROP_SRC_LOCAL);
}
if (spa->spa_compatibility != NULL) {
spa_prop_add_list(*nvp, ZPOOL_PROP_COMPATIBILITY,
spa->spa_compatibility, 0, ZPROP_SRC_LOCAL);
}
if (spa->spa_root != NULL)
spa_prop_add_list(*nvp, ZPOOL_PROP_ALTROOT, spa->spa_root,
0, ZPROP_SRC_LOCAL);
if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_BLOCKS)) {
spa_prop_add_list(*nvp, ZPOOL_PROP_MAXBLOCKSIZE, NULL,
MIN(zfs_max_recordsize, SPA_MAXBLOCKSIZE), ZPROP_SRC_NONE);
} else {
spa_prop_add_list(*nvp, ZPOOL_PROP_MAXBLOCKSIZE, NULL,
SPA_OLD_MAXBLOCKSIZE, ZPROP_SRC_NONE);
}
if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_DNODE)) {
spa_prop_add_list(*nvp, ZPOOL_PROP_MAXDNODESIZE, NULL,
DNODE_MAX_SIZE, ZPROP_SRC_NONE);
} else {
spa_prop_add_list(*nvp, ZPOOL_PROP_MAXDNODESIZE, NULL,
DNODE_MIN_SIZE, ZPROP_SRC_NONE);
}
if ((dp = list_head(&spa->spa_config_list)) != NULL) {
if (dp->scd_path == NULL) {
spa_prop_add_list(*nvp, ZPOOL_PROP_CACHEFILE,
"none", 0, ZPROP_SRC_LOCAL);
} else if (strcmp(dp->scd_path, spa_config_path) != 0) {
spa_prop_add_list(*nvp, ZPOOL_PROP_CACHEFILE,
dp->scd_path, 0, ZPROP_SRC_LOCAL);
}
}
}
/*
* Get zpool property values.
*/
int
spa_prop_get(spa_t *spa, nvlist_t **nvp)
{
objset_t *mos = spa->spa_meta_objset;
zap_cursor_t zc;
zap_attribute_t za;
dsl_pool_t *dp;
int err;
err = nvlist_alloc(nvp, NV_UNIQUE_NAME, KM_SLEEP);
if (err)
return (err);
dp = spa_get_dsl(spa);
dsl_pool_config_enter(dp, FTAG);
mutex_enter(&spa->spa_props_lock);
/*
* Get properties from the spa config.
*/
spa_prop_get_config(spa, nvp);
/* If no pool property object, no more prop to get. */
if (mos == NULL || spa->spa_pool_props_object == 0)
goto out;
/*
* Get properties from the MOS pool property object.
*/
for (zap_cursor_init(&zc, mos, spa->spa_pool_props_object);
(err = zap_cursor_retrieve(&zc, &za)) == 0;
zap_cursor_advance(&zc)) {
uint64_t intval = 0;
char *strval = NULL;
zprop_source_t src = ZPROP_SRC_DEFAULT;
zpool_prop_t prop;
if ((prop = zpool_name_to_prop(za.za_name)) ==
ZPOOL_PROP_INVAL && !zfs_prop_user(za.za_name))
continue;
switch (za.za_integer_length) {
case 8:
/* integer property */
if (za.za_first_integer !=
zpool_prop_default_numeric(prop))
src = ZPROP_SRC_LOCAL;
if (prop == ZPOOL_PROP_BOOTFS) {
dsl_dataset_t *ds = NULL;
err = dsl_dataset_hold_obj(dp,
za.za_first_integer, FTAG, &ds);
if (err != 0)
break;
strval = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN,
KM_SLEEP);
dsl_dataset_name(ds, strval);
dsl_dataset_rele(ds, FTAG);
} else {
strval = NULL;
intval = za.za_first_integer;
}
spa_prop_add_list(*nvp, prop, strval, intval, src);
if (strval != NULL)
kmem_free(strval, ZFS_MAX_DATASET_NAME_LEN);
break;
case 1:
/* string property */
strval = kmem_alloc(za.za_num_integers, KM_SLEEP);
err = zap_lookup(mos, spa->spa_pool_props_object,
za.za_name, 1, za.za_num_integers, strval);
if (err) {
kmem_free(strval, za.za_num_integers);
break;
}
if (prop != ZPOOL_PROP_INVAL) {
spa_prop_add_list(*nvp, prop, strval, 0, src);
} else {
src = ZPROP_SRC_LOCAL;
spa_prop_add_user(*nvp, za.za_name, strval,
src);
}
kmem_free(strval, za.za_num_integers);
break;
default:
break;
}
}
zap_cursor_fini(&zc);
out:
mutex_exit(&spa->spa_props_lock);
dsl_pool_config_exit(dp, FTAG);
if (err && err != ENOENT) {
nvlist_free(*nvp);
*nvp = NULL;
return (err);
}
return (0);
}
/*
* Validate the given pool properties nvlist and modify the list
* for the property values to be set.
*/
static int
spa_prop_validate(spa_t *spa, nvlist_t *props)
{
nvpair_t *elem;
int error = 0, reset_bootfs = 0;
uint64_t objnum = 0;
boolean_t has_feature = B_FALSE;
elem = NULL;
while ((elem = nvlist_next_nvpair(props, elem)) != NULL) {
uint64_t intval;
const char *strval, *slash, *check, *fname;
const char *propname = nvpair_name(elem);
zpool_prop_t prop = zpool_name_to_prop(propname);
switch (prop) {
case ZPOOL_PROP_INVAL:
/*
* Sanitize the input.
*/
if (zfs_prop_user(propname)) {
if (strlen(propname) >= ZAP_MAXNAMELEN) {
error = SET_ERROR(ENAMETOOLONG);
break;
}
if (strlen(fnvpair_value_string(elem)) >=
ZAP_MAXVALUELEN) {
error = SET_ERROR(E2BIG);
break;
}
} else if (zpool_prop_feature(propname)) {
if (nvpair_type(elem) != DATA_TYPE_UINT64) {
error = SET_ERROR(EINVAL);
break;
}
if (nvpair_value_uint64(elem, &intval) != 0) {
error = SET_ERROR(EINVAL);
break;
}
if (intval != 0) {
error = SET_ERROR(EINVAL);
break;
}
fname = strchr(propname, '@') + 1;
if (zfeature_lookup_name(fname, NULL) != 0) {
error = SET_ERROR(EINVAL);
break;
}
has_feature = B_TRUE;
} else {
error = SET_ERROR(EINVAL);
break;
}
break;
case ZPOOL_PROP_VERSION:
error = nvpair_value_uint64(elem, &intval);
if (!error &&
(intval < spa_version(spa) ||
intval > SPA_VERSION_BEFORE_FEATURES ||
has_feature))
error = SET_ERROR(EINVAL);
break;
case ZPOOL_PROP_DELEGATION:
case ZPOOL_PROP_AUTOREPLACE:
case ZPOOL_PROP_LISTSNAPS:
case ZPOOL_PROP_AUTOEXPAND:
case ZPOOL_PROP_AUTOTRIM:
error = nvpair_value_uint64(elem, &intval);
if (!error && intval > 1)
error = SET_ERROR(EINVAL);
break;
case ZPOOL_PROP_MULTIHOST:
error = nvpair_value_uint64(elem, &intval);
if (!error && intval > 1)
error = SET_ERROR(EINVAL);
if (!error) {
uint32_t hostid = zone_get_hostid(NULL);
if (hostid)
spa->spa_hostid = hostid;
else
error = SET_ERROR(ENOTSUP);
}
break;
case ZPOOL_PROP_BOOTFS:
/*
* If the pool version is less than SPA_VERSION_BOOTFS,
* or the pool is still being created (version == 0),
* the bootfs property cannot be set.
*/
if (spa_version(spa) < SPA_VERSION_BOOTFS) {
error = SET_ERROR(ENOTSUP);
break;
}
/*
* Make sure the vdev config is bootable
*/
if (!vdev_is_bootable(spa->spa_root_vdev)) {
error = SET_ERROR(ENOTSUP);
break;
}
reset_bootfs = 1;
error = nvpair_value_string(elem, &strval);
if (!error) {
objset_t *os;
if (strval == NULL || strval[0] == '\0') {
objnum = zpool_prop_default_numeric(
ZPOOL_PROP_BOOTFS);
break;
}
error = dmu_objset_hold(strval, FTAG, &os);
if (error != 0)
break;
/* Must be ZPL. */
if (dmu_objset_type(os) != DMU_OST_ZFS) {
error = SET_ERROR(ENOTSUP);
} else {
objnum = dmu_objset_id(os);
}
dmu_objset_rele(os, FTAG);
}
break;
case ZPOOL_PROP_FAILUREMODE:
error = nvpair_value_uint64(elem, &intval);
if (!error && intval > ZIO_FAILURE_MODE_PANIC)
error = SET_ERROR(EINVAL);
/*
* This is a special case which only occurs when
* the pool has completely failed. This allows
* the user to change the in-core failmode property
* without syncing it out to disk (I/Os might
* currently be blocked). We do this by returning
* EIO to the caller (spa_prop_set) to trick it
* into thinking we encountered a property validation
* error.
*/
if (!error && spa_suspended(spa)) {
spa->spa_failmode = intval;
error = SET_ERROR(EIO);
}
break;
case ZPOOL_PROP_CACHEFILE:
if ((error = nvpair_value_string(elem, &strval)) != 0)
break;
if (strval[0] == '\0')
break;
if (strcmp(strval, "none") == 0)
break;
if (strval[0] != '/') {
error = SET_ERROR(EINVAL);
break;
}
slash = strrchr(strval, '/');
ASSERT(slash != NULL);
if (slash[1] == '\0' || strcmp(slash, "/.") == 0 ||
strcmp(slash, "/..") == 0)
error = SET_ERROR(EINVAL);
break;
case ZPOOL_PROP_COMMENT:
if ((error = nvpair_value_string(elem, &strval)) != 0)
break;
for (check = strval; *check != '\0'; check++) {
if (!isprint(*check)) {
error = SET_ERROR(EINVAL);
break;
}
}
if (strlen(strval) > ZPROP_MAX_COMMENT)
error = SET_ERROR(E2BIG);
break;
default:
break;
}
if (error)
break;
}
(void) nvlist_remove_all(props,
zpool_prop_to_name(ZPOOL_PROP_DEDUPDITTO));
if (!error && reset_bootfs) {
error = nvlist_remove(props,
zpool_prop_to_name(ZPOOL_PROP_BOOTFS), DATA_TYPE_STRING);
if (!error) {
error = nvlist_add_uint64(props,
zpool_prop_to_name(ZPOOL_PROP_BOOTFS), objnum);
}
}
return (error);
}
void
spa_configfile_set(spa_t *spa, nvlist_t *nvp, boolean_t need_sync)
{
const char *cachefile;
spa_config_dirent_t *dp;
if (nvlist_lookup_string(nvp, zpool_prop_to_name(ZPOOL_PROP_CACHEFILE),
&cachefile) != 0)
return;
dp = kmem_alloc(sizeof (spa_config_dirent_t),
KM_SLEEP);
if (cachefile[0] == '\0')
dp->scd_path = spa_strdup(spa_config_path);
else if (strcmp(cachefile, "none") == 0)
dp->scd_path = NULL;
else
dp->scd_path = spa_strdup(cachefile);
list_insert_head(&spa->spa_config_list, dp);
if (need_sync)
spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
}
int
spa_prop_set(spa_t *spa, nvlist_t *nvp)
{
int error;
nvpair_t *elem = NULL;
boolean_t need_sync = B_FALSE;
if ((error = spa_prop_validate(spa, nvp)) != 0)
return (error);
while ((elem = nvlist_next_nvpair(nvp, elem)) != NULL) {
zpool_prop_t prop = zpool_name_to_prop(nvpair_name(elem));
if (prop == ZPOOL_PROP_CACHEFILE ||
prop == ZPOOL_PROP_ALTROOT ||
prop == ZPOOL_PROP_READONLY)
continue;
if (prop == ZPOOL_PROP_INVAL &&
zfs_prop_user(nvpair_name(elem))) {
need_sync = B_TRUE;
break;
}
if (prop == ZPOOL_PROP_VERSION || prop == ZPOOL_PROP_INVAL) {
uint64_t ver = 0;
if (prop == ZPOOL_PROP_VERSION) {
VERIFY(nvpair_value_uint64(elem, &ver) == 0);
} else {
ASSERT(zpool_prop_feature(nvpair_name(elem)));
ver = SPA_VERSION_FEATURES;
need_sync = B_TRUE;
}
/* Save time if the version is already set. */
if (ver == spa_version(spa))
continue;
/*
* In addition to the pool directory object, we might
* create the pool properties object, the features for
* read object, the features for write object, or the
* feature descriptions object.
*/
error = dsl_sync_task(spa->spa_name, NULL,
spa_sync_version, &ver,
6, ZFS_SPACE_CHECK_RESERVED);
if (error)
return (error);
continue;
}
need_sync = B_TRUE;
break;
}
if (need_sync) {
return (dsl_sync_task(spa->spa_name, NULL, spa_sync_props,
nvp, 6, ZFS_SPACE_CHECK_RESERVED));
}
return (0);
}
/*
* If the bootfs property value is dsobj, clear it.
*/
void
spa_prop_clear_bootfs(spa_t *spa, uint64_t dsobj, dmu_tx_t *tx)
{
if (spa->spa_bootfs == dsobj && spa->spa_pool_props_object != 0) {
VERIFY(zap_remove(spa->spa_meta_objset,
spa->spa_pool_props_object,
zpool_prop_to_name(ZPOOL_PROP_BOOTFS), tx) == 0);
spa->spa_bootfs = 0;
}
}
static int
spa_change_guid_check(void *arg, dmu_tx_t *tx)
{
uint64_t *newguid __maybe_unused = arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
vdev_t *rvd = spa->spa_root_vdev;
uint64_t vdev_state;
if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) {
int error = (spa_has_checkpoint(spa)) ?
ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT;
return (SET_ERROR(error));
}
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
vdev_state = rvd->vdev_state;
spa_config_exit(spa, SCL_STATE, FTAG);
if (vdev_state != VDEV_STATE_HEALTHY)
return (SET_ERROR(ENXIO));
ASSERT3U(spa_guid(spa), !=, *newguid);
return (0);
}
static void
spa_change_guid_sync(void *arg, dmu_tx_t *tx)
{
uint64_t *newguid = arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
uint64_t oldguid;
vdev_t *rvd = spa->spa_root_vdev;
oldguid = spa_guid(spa);
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
rvd->vdev_guid = *newguid;
rvd->vdev_guid_sum += (*newguid - oldguid);
vdev_config_dirty(rvd);
spa_config_exit(spa, SCL_STATE, FTAG);
spa_history_log_internal(spa, "guid change", tx, "old=%llu new=%llu",
(u_longlong_t)oldguid, (u_longlong_t)*newguid);
}
/*
* Change the GUID for the pool. This is done so that we can later
* re-import a pool built from a clone of our own vdevs. We will modify
* the root vdev's guid, our own pool guid, and then mark all of our
* vdevs dirty. Note that we must make sure that all our vdevs are
* online when we do this, or else any vdevs that weren't present
* would be orphaned from our pool. We are also going to issue a
* sysevent to update any watchers.
*/
int
spa_change_guid(spa_t *spa)
{
int error;
uint64_t guid;
mutex_enter(&spa->spa_vdev_top_lock);
mutex_enter(&spa_namespace_lock);
guid = spa_generate_guid(NULL);
error = dsl_sync_task(spa->spa_name, spa_change_guid_check,
spa_change_guid_sync, &guid, 5, ZFS_SPACE_CHECK_RESERVED);
if (error == 0) {
/*
* Clear the kobj flag from all the vdevs to allow
* vdev_cache_process_kobj_evt() to post events to all the
* vdevs since GUID is updated.
*/
vdev_clear_kobj_evt(spa->spa_root_vdev);
for (int i = 0; i < spa->spa_l2cache.sav_count; i++)
vdev_clear_kobj_evt(spa->spa_l2cache.sav_vdevs[i]);
spa_write_cachefile(spa, B_FALSE, B_TRUE, B_TRUE);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_REGUID);
}
mutex_exit(&spa_namespace_lock);
mutex_exit(&spa->spa_vdev_top_lock);
return (error);
}
/*
* ==========================================================================
* SPA state manipulation (open/create/destroy/import/export)
* ==========================================================================
*/
static int
spa_error_entry_compare(const void *a, const void *b)
{
const spa_error_entry_t *sa = (const spa_error_entry_t *)a;
const spa_error_entry_t *sb = (const spa_error_entry_t *)b;
int ret;
ret = memcmp(&sa->se_bookmark, &sb->se_bookmark,
sizeof (zbookmark_phys_t));
return (TREE_ISIGN(ret));
}
/*
* Utility function which retrieves copies of the current logs and
* re-initializes them in the process.
*/
void
spa_get_errlists(spa_t *spa, avl_tree_t *last, avl_tree_t *scrub)
{
ASSERT(MUTEX_HELD(&spa->spa_errlist_lock));
memcpy(last, &spa->spa_errlist_last, sizeof (avl_tree_t));
memcpy(scrub, &spa->spa_errlist_scrub, sizeof (avl_tree_t));
avl_create(&spa->spa_errlist_scrub,
spa_error_entry_compare, sizeof (spa_error_entry_t),
offsetof(spa_error_entry_t, se_avl));
avl_create(&spa->spa_errlist_last,
spa_error_entry_compare, sizeof (spa_error_entry_t),
offsetof(spa_error_entry_t, se_avl));
}
static void
spa_taskqs_init(spa_t *spa, zio_type_t t, zio_taskq_type_t q)
{
const zio_taskq_info_t *ztip = &zio_taskqs[t][q];
enum zti_modes mode = ztip->zti_mode;
uint_t value = ztip->zti_value;
uint_t count = ztip->zti_count;
spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q];
uint_t cpus, flags = TASKQ_DYNAMIC;
switch (mode) {
case ZTI_MODE_FIXED:
ASSERT3U(value, >, 0);
break;
case ZTI_MODE_SYNC:
/*
* Create one wr_iss taskq for every 'zio_taskq_wr_iss_ncpus',
* not to exceed the number of spa allocators.
*/
if (zio_taskq_wr_iss_ncpus == 0) {
count = MAX(boot_ncpus / spa->spa_alloc_count, 1);
} else {
count = MAX(1,
boot_ncpus / MAX(1, zio_taskq_wr_iss_ncpus));
}
count = MAX(count, (zio_taskq_batch_pct + 99) / 100);
count = MIN(count, spa->spa_alloc_count);
/*
* zio_taskq_batch_pct is unbounded and may exceed 100%, but no
* single taskq may have more threads than 100% of online cpus.
*/
value = (zio_taskq_batch_pct + count / 2) / count;
value = MIN(value, 100);
flags |= TASKQ_THREADS_CPU_PCT;
break;
case ZTI_MODE_SCALE:
flags |= TASKQ_THREADS_CPU_PCT;
/*
* We want more taskqs to reduce lock contention, but we want
* less for better request ordering and CPU utilization.
*/
cpus = MAX(1, boot_ncpus * zio_taskq_batch_pct / 100);
if (zio_taskq_batch_tpq > 0) {
count = MAX(1, (cpus + zio_taskq_batch_tpq / 2) /
zio_taskq_batch_tpq);
} else {
/*
* Prefer 6 threads per taskq, but no more taskqs
* than threads in them on large systems. For 80%:
*
* taskq taskq total
* cpus taskqs percent threads threads
* ------- ------- ------- ------- -------
* 1 1 80% 1 1
* 2 1 80% 1 1
* 4 1 80% 3 3
* 8 2 40% 3 6
* 16 3 27% 4 12
* 32 5 16% 5 25
* 64 7 11% 7 49
* 128 10 8% 10 100
* 256 14 6% 15 210
*/
count = 1 + cpus / 6;
while (count * count > cpus)
count--;
}
/* Limit each taskq within 100% to not trigger assertion. */
count = MAX(count, (zio_taskq_batch_pct + 99) / 100);
value = (zio_taskq_batch_pct + count / 2) / count;
break;
case ZTI_MODE_NULL:
tqs->stqs_count = 0;
tqs->stqs_taskq = NULL;
return;
default:
panic("unrecognized mode for %s_%s taskq (%u:%u) in "
"spa_taskqs_init()",
zio_type_name[t], zio_taskq_types[q], mode, value);
break;
}
ASSERT3U(count, >, 0);
tqs->stqs_count = count;
tqs->stqs_taskq = kmem_alloc(count * sizeof (taskq_t *), KM_SLEEP);
for (uint_t i = 0; i < count; i++) {
taskq_t *tq;
char name[32];
if (count > 1)
(void) snprintf(name, sizeof (name), "%s_%s_%u",
zio_type_name[t], zio_taskq_types[q], i);
else
(void) snprintf(name, sizeof (name), "%s_%s",
zio_type_name[t], zio_taskq_types[q]);
#ifdef HAVE_SYSDC
if (zio_taskq_sysdc && spa->spa_proc != &p0) {
(void) zio_taskq_basedc;
tq = taskq_create_sysdc(name, value, 50, INT_MAX,
spa->spa_proc, zio_taskq_basedc, flags);
} else {
#endif
pri_t pri = maxclsyspri;
/*
* The write issue taskq can be extremely CPU
* intensive. Run it at slightly less important
* priority than the other taskqs.
*
* Under Linux and FreeBSD this means incrementing
* the priority value as opposed to platforms like
* illumos where it should be decremented.
*
* On FreeBSD, if priorities divided by four (RQ_PPQ)
* are equal then a difference between them is
* insignificant.
*/
if (t == ZIO_TYPE_WRITE && q == ZIO_TASKQ_ISSUE) {
#if defined(__linux__)
pri++;
#elif defined(__FreeBSD__)
pri += 4;
#else
#error "unknown OS"
#endif
}
tq = taskq_create_proc(name, value, pri, 50,
INT_MAX, spa->spa_proc, flags);
#ifdef HAVE_SYSDC
}
#endif
tqs->stqs_taskq[i] = tq;
}
}
static void
spa_taskqs_fini(spa_t *spa, zio_type_t t, zio_taskq_type_t q)
{
spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q];
if (tqs->stqs_taskq == NULL) {
ASSERT3U(tqs->stqs_count, ==, 0);
return;
}
for (uint_t i = 0; i < tqs->stqs_count; i++) {
ASSERT3P(tqs->stqs_taskq[i], !=, NULL);
taskq_destroy(tqs->stqs_taskq[i]);
}
kmem_free(tqs->stqs_taskq, tqs->stqs_count * sizeof (taskq_t *));
tqs->stqs_taskq = NULL;
}
/*
* Dispatch a task to the appropriate taskq for the ZFS I/O type and priority.
* Note that a type may have multiple discrete taskqs to avoid lock contention
* on the taskq itself.
*/
static taskq_t *
spa_taskq_dispatch_select(spa_t *spa, zio_type_t t, zio_taskq_type_t q,
zio_t *zio)
{
spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q];
taskq_t *tq;
ASSERT3P(tqs->stqs_taskq, !=, NULL);
ASSERT3U(tqs->stqs_count, !=, 0);
if ((t == ZIO_TYPE_WRITE) && (q == ZIO_TASKQ_ISSUE) &&
(zio != NULL) && (zio->io_wr_iss_tq != NULL)) {
/* dispatch to assigned write issue taskq */
tq = zio->io_wr_iss_tq;
return (tq);
}
if (tqs->stqs_count == 1) {
tq = tqs->stqs_taskq[0];
} else {
tq = tqs->stqs_taskq[((uint64_t)gethrtime()) % tqs->stqs_count];
}
return (tq);
}
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,
zio_t *zio)
{
taskq_t *tq = spa_taskq_dispatch_select(spa, t, q, zio);
taskq_dispatch_ent(tq, func, arg, flags, ent);
}
/*
* Same as spa_taskq_dispatch_ent() but block on the task until completion.
*/
void
spa_taskq_dispatch_sync(spa_t *spa, zio_type_t t, zio_taskq_type_t q,
task_func_t *func, void *arg, uint_t flags)
{
taskq_t *tq = spa_taskq_dispatch_select(spa, t, q, NULL);
taskqid_t id = taskq_dispatch(tq, func, arg, flags);
if (id)
taskq_wait_id(tq, id);
}
static void
spa_create_zio_taskqs(spa_t *spa)
{
for (int t = 0; t < ZIO_TYPES; t++) {
for (int q = 0; q < ZIO_TASKQ_TYPES; q++) {
spa_taskqs_init(spa, t, q);
}
}
}
#if defined(_KERNEL) && defined(HAVE_SPA_THREAD)
static void
spa_thread(void *arg)
{
psetid_t zio_taskq_psrset_bind = PS_NONE;
callb_cpr_t cprinfo;
spa_t *spa = arg;
user_t *pu = PTOU(curproc);
CALLB_CPR_INIT(&cprinfo, &spa->spa_proc_lock, callb_generic_cpr,
spa->spa_name);
ASSERT(curproc != &p0);
(void) snprintf(pu->u_psargs, sizeof (pu->u_psargs),
"zpool-%s", spa->spa_name);
(void) strlcpy(pu->u_comm, pu->u_psargs, sizeof (pu->u_comm));
/* bind this thread to the requested psrset */
if (zio_taskq_psrset_bind != PS_NONE) {
pool_lock();
mutex_enter(&cpu_lock);
mutex_enter(&pidlock);
mutex_enter(&curproc->p_lock);
if (cpupart_bind_thread(curthread, zio_taskq_psrset_bind,
0, NULL, NULL) == 0) {
curthread->t_bind_pset = zio_taskq_psrset_bind;
} else {
cmn_err(CE_WARN,
"Couldn't bind process for zfs pool \"%s\" to "
"pset %d\n", spa->spa_name, zio_taskq_psrset_bind);
}
mutex_exit(&curproc->p_lock);
mutex_exit(&pidlock);
mutex_exit(&cpu_lock);
pool_unlock();
}
#ifdef HAVE_SYSDC
if (zio_taskq_sysdc) {
sysdc_thread_enter(curthread, 100, 0);
}
#endif
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
extern metaslab_ops_t *metaslab_allocator(spa_t *spa);
/*
* Activate an uninitialized pool.
*/
static void
spa_activate(spa_t *spa, spa_mode_t mode)
{
metaslab_ops_t *msp = metaslab_allocator(spa);
ASSERT(spa->spa_state == POOL_STATE_UNINITIALIZED);
spa->spa_state = POOL_STATE_ACTIVE;
spa->spa_mode = mode;
spa->spa_read_spacemaps = spa_mode_readable_spacemaps;
spa->spa_normal_class = metaslab_class_create(spa, msp);
spa->spa_log_class = metaslab_class_create(spa, msp);
spa->spa_embedded_log_class = metaslab_class_create(spa, msp);
spa->spa_special_class = metaslab_class_create(spa, msp);
spa->spa_dedup_class = metaslab_class_create(spa, msp);
/* 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;
#ifdef HAVE_SPA_THREAD
/* Only create a process if we're going to be around a while. */
if (spa_create_process && strcmp(spa->spa_name, TRYIMPORT_NAME) != 0) {
if (newproc(spa_thread, (caddr_t)spa, syscid, maxclsyspri,
NULL, 0) == 0) {
spa->spa_proc_state = SPA_PROC_CREATED;
while (spa->spa_proc_state == SPA_PROC_CREATED) {
cv_wait(&spa->spa_proc_cv,
&spa->spa_proc_lock);
}
ASSERT(spa->spa_proc_state == SPA_PROC_ACTIVE);
ASSERT(spa->spa_proc != &p0);
ASSERT(spa->spa_did != 0);
} else {
#ifdef _KERNEL
cmn_err(CE_WARN,
"Couldn't create process for zfs pool \"%s\"\n",
spa->spa_name);
#endif
}
}
#endif /* HAVE_SPA_THREAD */
mutex_exit(&spa->spa_proc_lock);
/* If we didn't create a process, we need to create our taskqs. */
if (spa->spa_proc == &p0) {
spa_create_zio_taskqs(spa);
}
for (size_t i = 0; i < TXG_SIZE; i++) {
spa->spa_txg_zio[i] = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL);
}
list_create(&spa->spa_config_dirty_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_config_dirty_node));
list_create(&spa->spa_evicting_os_list, sizeof (objset_t),
offsetof(objset_t, os_evicting_node));
list_create(&spa->spa_state_dirty_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_state_dirty_node));
txg_list_create(&spa->spa_vdev_txg_list, spa,
offsetof(struct vdev, vdev_txg_node));
avl_create(&spa->spa_errlist_scrub,
spa_error_entry_compare, sizeof (spa_error_entry_t),
offsetof(spa_error_entry_t, se_avl));
avl_create(&spa->spa_errlist_last,
spa_error_entry_compare, sizeof (spa_error_entry_t),
offsetof(spa_error_entry_t, se_avl));
avl_create(&spa->spa_errlist_healed,
spa_error_entry_compare, sizeof (spa_error_entry_t),
offsetof(spa_error_entry_t, se_avl));
spa_activate_os(spa);
spa_keystore_init(&spa->spa_keystore);
/*
* This taskq is used to perform zvol-minor-related tasks
* asynchronously. This has several advantages, including easy
* resolution of various deadlocks.
*
* The taskq must be single threaded to ensure tasks are always
* processed in the order in which they were dispatched.
*
* A taskq per pool allows one to keep the pools independent.
* This way if one pool is suspended, it will not impact another.
*
* The preferred location to dispatch a zvol minor task is a sync
* task. In this context, there is easy access to the spa_t and minimal
* error handling is required because the sync task must succeed.
*/
spa->spa_zvol_taskq = taskq_create("z_zvol", 1, defclsyspri,
1, INT_MAX, 0);
/*
* The taskq to preload metaslabs.
*/
spa->spa_metaslab_taskq = taskq_create("z_metaslab",
metaslab_preload_pct, maxclsyspri, 1, INT_MAX,
TASKQ_DYNAMIC | TASKQ_THREADS_CPU_PCT);
/*
* Taskq dedicated to prefetcher threads: this is used to prevent the
* pool traverse code from monopolizing the global (and limited)
* system_taskq by inappropriately scheduling long running tasks on it.
*/
spa->spa_prefetch_taskq = taskq_create("z_prefetch", 100,
defclsyspri, 1, INT_MAX, TASKQ_DYNAMIC | TASKQ_THREADS_CPU_PCT);
/*
* The taskq to upgrade datasets in this pool. Currently used by
* feature SPA_FEATURE_USEROBJ_ACCOUNTING/SPA_FEATURE_PROJECT_QUOTA.
*/
spa->spa_upgrade_taskq = taskq_create("z_upgrade", 100,
defclsyspri, 1, INT_MAX, TASKQ_DYNAMIC | TASKQ_THREADS_CPU_PCT);
}
/*
* Opposite of spa_activate().
*/
static void
spa_deactivate(spa_t *spa)
{
ASSERT(spa->spa_sync_on == B_FALSE);
ASSERT(spa->spa_dsl_pool == NULL);
ASSERT(spa->spa_root_vdev == NULL);
ASSERT(spa->spa_async_zio_root == NULL);
ASSERT(spa->spa_state != POOL_STATE_UNINITIALIZED);
spa_evicting_os_wait(spa);
if (spa->spa_zvol_taskq) {
taskq_destroy(spa->spa_zvol_taskq);
spa->spa_zvol_taskq = NULL;
}
if (spa->spa_metaslab_taskq) {
taskq_destroy(spa->spa_metaslab_taskq);
spa->spa_metaslab_taskq = NULL;
}
if (spa->spa_prefetch_taskq) {
taskq_destroy(spa->spa_prefetch_taskq);
spa->spa_prefetch_taskq = NULL;
}
if (spa->spa_upgrade_taskq) {
taskq_destroy(spa->spa_upgrade_taskq);
spa->spa_upgrade_taskq = NULL;
}
txg_list_destroy(&spa->spa_vdev_txg_list);
list_destroy(&spa->spa_config_dirty_list);
list_destroy(&spa->spa_evicting_os_list);
list_destroy(&spa->spa_state_dirty_list);
taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid);
for (int t = 0; t < ZIO_TYPES; t++) {
for (int q = 0; q < ZIO_TASKQ_TYPES; q++) {
spa_taskqs_fini(spa, t, q);
}
}
for (size_t i = 0; i < TXG_SIZE; i++) {
ASSERT3P(spa->spa_txg_zio[i], !=, NULL);
VERIFY0(zio_wait(spa->spa_txg_zio[i]));
spa->spa_txg_zio[i] = NULL;
}
metaslab_class_destroy(spa->spa_normal_class);
spa->spa_normal_class = NULL;
metaslab_class_destroy(spa->spa_log_class);
spa->spa_log_class = NULL;
metaslab_class_destroy(spa->spa_embedded_log_class);
spa->spa_embedded_log_class = NULL;
metaslab_class_destroy(spa->spa_special_class);
spa->spa_special_class = NULL;
metaslab_class_destroy(spa->spa_dedup_class);
spa->spa_dedup_class = NULL;
/*
* If this was part of an import or the open otherwise failed, we may
* still have errors left in the queues. Empty them just in case.
*/
spa_errlog_drain(spa);
avl_destroy(&spa->spa_errlist_scrub);
avl_destroy(&spa->spa_errlist_last);
avl_destroy(&spa->spa_errlist_healed);
spa_keystore_fini(&spa->spa_keystore);
spa->spa_state = POOL_STATE_UNINITIALIZED;
mutex_enter(&spa->spa_proc_lock);
if (spa->spa_proc_state != SPA_PROC_NONE) {
ASSERT(spa->spa_proc_state == SPA_PROC_ACTIVE);
spa->spa_proc_state = SPA_PROC_DEACTIVATE;
cv_broadcast(&spa->spa_proc_cv);
while (spa->spa_proc_state == SPA_PROC_DEACTIVATE) {
ASSERT(spa->spa_proc != &p0);
cv_wait(&spa->spa_proc_cv, &spa->spa_proc_lock);
}
ASSERT(spa->spa_proc_state == SPA_PROC_GONE);
spa->spa_proc_state = SPA_PROC_NONE;
}
ASSERT(spa->spa_proc == &p0);
mutex_exit(&spa->spa_proc_lock);
/*
* We want to make sure spa_thread() has actually exited the ZFS
* module, so that the module can't be unloaded out from underneath
* it.
*/
if (spa->spa_did != 0) {
thread_join(spa->spa_did);
spa->spa_did = 0;
}
spa_deactivate_os(spa);
}
/*
* Verify a pool configuration, and construct the vdev tree appropriately. This
* will create all the necessary vdevs in the appropriate layout, with each vdev
* in the CLOSED state. This will prep the pool before open/creation/import.
* All vdev validation is done by the vdev_alloc() routine.
*/
int
spa_config_parse(spa_t *spa, vdev_t **vdp, nvlist_t *nv, vdev_t *parent,
uint_t id, int atype)
{
nvlist_t **child;
uint_t children;
int error;
if ((error = vdev_alloc(spa, vdp, nv, parent, id, atype)) != 0)
return (error);
if ((*vdp)->vdev_ops->vdev_op_leaf)
return (0);
error = nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children);
if (error == ENOENT)
return (0);
if (error) {
vdev_free(*vdp);
*vdp = NULL;
return (SET_ERROR(EINVAL));
}
for (int c = 0; c < children; c++) {
vdev_t *vd;
if ((error = spa_config_parse(spa, &vd, child[c], *vdp, c,
atype)) != 0) {
vdev_free(*vdp);
*vdp = NULL;
return (error);
}
}
ASSERT(*vdp != NULL);
return (0);
}
static boolean_t
spa_should_flush_logs_on_unload(spa_t *spa)
{
if (!spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP))
return (B_FALSE);
if (!spa_writeable(spa))
return (B_FALSE);
if (!spa->spa_sync_on)
return (B_FALSE);
if (spa_state(spa) != POOL_STATE_EXPORTED)
return (B_FALSE);
if (zfs_keep_log_spacemaps_at_export)
return (B_FALSE);
return (B_TRUE);
}
/*
* Opens a transaction that will set the flag that will instruct
* spa_sync to attempt to flush all the metaslabs for that txg.
*/
static void
spa_unload_log_sm_flush_all(spa_t *spa)
{
dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir);
VERIFY0(dmu_tx_assign(tx, TXG_WAIT));
ASSERT3U(spa->spa_log_flushall_txg, ==, 0);
spa->spa_log_flushall_txg = dmu_tx_get_txg(tx);
dmu_tx_commit(tx);
txg_wait_synced(spa_get_dsl(spa), spa->spa_log_flushall_txg);
}
static void
spa_unload_log_sm_metadata(spa_t *spa)
{
void *cookie = NULL;
spa_log_sm_t *sls;
log_summary_entry_t *e;
while ((sls = avl_destroy_nodes(&spa->spa_sm_logs_by_txg,
&cookie)) != NULL) {
VERIFY0(sls->sls_mscount);
kmem_free(sls, sizeof (spa_log_sm_t));
}
while ((e = list_remove_head(&spa->spa_log_summary)) != NULL) {
VERIFY0(e->lse_mscount);
kmem_free(e, sizeof (log_summary_entry_t));
}
spa->spa_unflushed_stats.sus_nblocks = 0;
spa->spa_unflushed_stats.sus_memused = 0;
spa->spa_unflushed_stats.sus_blocklimit = 0;
}
static void
spa_destroy_aux_threads(spa_t *spa)
{
if (spa->spa_condense_zthr != NULL) {
zthr_destroy(spa->spa_condense_zthr);
spa->spa_condense_zthr = NULL;
}
if (spa->spa_checkpoint_discard_zthr != NULL) {
zthr_destroy(spa->spa_checkpoint_discard_zthr);
spa->spa_checkpoint_discard_zthr = NULL;
}
if (spa->spa_livelist_delete_zthr != NULL) {
zthr_destroy(spa->spa_livelist_delete_zthr);
spa->spa_livelist_delete_zthr = NULL;
}
if (spa->spa_livelist_condense_zthr != NULL) {
zthr_destroy(spa->spa_livelist_condense_zthr);
spa->spa_livelist_condense_zthr = NULL;
}
+ if (spa->spa_raidz_expand_zthr != NULL) {
+ zthr_destroy(spa->spa_raidz_expand_zthr);
+ spa->spa_raidz_expand_zthr = NULL;
+ }
}
/*
* Opposite of spa_load().
*/
static void
spa_unload(spa_t *spa)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(spa_state(spa) != POOL_STATE_UNINITIALIZED);
spa_import_progress_remove(spa_guid(spa));
spa_load_note(spa, "UNLOADING");
spa_wake_waiters(spa);
/*
* If we have set the spa_final_txg, we have already performed the
* tasks below in spa_export_common(). We should not redo it here since
* we delay the final TXGs beyond what spa_final_txg is set at.
*/
if (spa->spa_final_txg == UINT64_MAX) {
/*
* If the log space map feature is enabled and the pool is
* getting exported (but not destroyed), we want to spend some
* time flushing as many metaslabs as we can in an attempt to
* destroy log space maps and save import time.
*/
if (spa_should_flush_logs_on_unload(spa))
spa_unload_log_sm_flush_all(spa);
/*
* Stop async tasks.
*/
spa_async_suspend(spa);
if (spa->spa_root_vdev) {
vdev_t *root_vdev = spa->spa_root_vdev;
vdev_initialize_stop_all(root_vdev,
VDEV_INITIALIZE_ACTIVE);
vdev_trim_stop_all(root_vdev, VDEV_TRIM_ACTIVE);
vdev_autotrim_stop_all(spa);
vdev_rebuild_stop_all(spa);
}
}
/*
* Stop syncing.
*/
if (spa->spa_sync_on) {
txg_sync_stop(spa->spa_dsl_pool);
spa->spa_sync_on = B_FALSE;
}
/*
* This ensures that there is no async metaslab prefetching
* while we attempt to unload the spa.
*/
taskq_wait(spa->spa_metaslab_taskq);
if (spa->spa_mmp.mmp_thread)
mmp_thread_stop(spa);
/*
* Wait for any outstanding async I/O to complete.
*/
if (spa->spa_async_zio_root != NULL) {
for (int i = 0; i < max_ncpus; i++)
(void) zio_wait(spa->spa_async_zio_root[i]);
kmem_free(spa->spa_async_zio_root, max_ncpus * sizeof (void *));
spa->spa_async_zio_root = NULL;
}
if (spa->spa_vdev_removal != NULL) {
spa_vdev_removal_destroy(spa->spa_vdev_removal);
spa->spa_vdev_removal = NULL;
}
spa_destroy_aux_threads(spa);
spa_condense_fini(spa);
bpobj_close(&spa->spa_deferred_bpobj);
spa_config_enter(spa, SCL_ALL, spa, RW_WRITER);
/*
* Close all vdevs.
*/
if (spa->spa_root_vdev)
vdev_free(spa->spa_root_vdev);
ASSERT(spa->spa_root_vdev == NULL);
/*
* Close the dsl pool.
*/
if (spa->spa_dsl_pool) {
dsl_pool_close(spa->spa_dsl_pool);
spa->spa_dsl_pool = NULL;
spa->spa_meta_objset = NULL;
}
ddt_unload(spa);
brt_unload(spa);
spa_unload_log_sm_metadata(spa);
/*
* Drop and purge level 2 cache
*/
spa_l2cache_drop(spa);
if (spa->spa_spares.sav_vdevs) {
for (int i = 0; i < spa->spa_spares.sav_count; i++)
vdev_free(spa->spa_spares.sav_vdevs[i]);
kmem_free(spa->spa_spares.sav_vdevs,
spa->spa_spares.sav_count * sizeof (void *));
spa->spa_spares.sav_vdevs = NULL;
}
if (spa->spa_spares.sav_config) {
nvlist_free(spa->spa_spares.sav_config);
spa->spa_spares.sav_config = NULL;
}
spa->spa_spares.sav_count = 0;
if (spa->spa_l2cache.sav_vdevs) {
for (int i = 0; i < spa->spa_l2cache.sav_count; i++) {
vdev_clear_stats(spa->spa_l2cache.sav_vdevs[i]);
vdev_free(spa->spa_l2cache.sav_vdevs[i]);
}
kmem_free(spa->spa_l2cache.sav_vdevs,
spa->spa_l2cache.sav_count * sizeof (void *));
spa->spa_l2cache.sav_vdevs = NULL;
}
if (spa->spa_l2cache.sav_config) {
nvlist_free(spa->spa_l2cache.sav_config);
spa->spa_l2cache.sav_config = NULL;
}
spa->spa_l2cache.sav_count = 0;
spa->spa_async_suspended = 0;
spa->spa_indirect_vdevs_loaded = B_FALSE;
if (spa->spa_comment != NULL) {
spa_strfree(spa->spa_comment);
spa->spa_comment = NULL;
}
if (spa->spa_compatibility != NULL) {
spa_strfree(spa->spa_compatibility);
spa->spa_compatibility = NULL;
}
+ spa->spa_raidz_expand = NULL;
+
spa_config_exit(spa, SCL_ALL, spa);
}
/*
* Load (or re-load) the current list of vdevs describing the active spares for
* this pool. When this is called, we have some form of basic information in
* 'spa_spares.sav_config'. We parse this into vdevs, try to open them, and
* then re-generate a more complete list including status information.
*/
void
spa_load_spares(spa_t *spa)
{
nvlist_t **spares;
uint_t nspares;
int i;
vdev_t *vd, *tvd;
#ifndef _KERNEL
/*
* zdb opens both the current state of the pool and the
* checkpointed state (if present), with a different spa_t.
*
* As spare vdevs are shared among open pools, we skip loading
* them when we load the checkpointed state of the pool.
*/
if (!spa_writeable(spa))
return;
#endif
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
/*
* First, close and free any existing spare vdevs.
*/
if (spa->spa_spares.sav_vdevs) {
for (i = 0; i < spa->spa_spares.sav_count; i++) {
vd = spa->spa_spares.sav_vdevs[i];
/* Undo the call to spa_activate() below */
if ((tvd = spa_lookup_by_guid(spa, vd->vdev_guid,
B_FALSE)) != NULL && tvd->vdev_isspare)
spa_spare_remove(tvd);
vdev_close(vd);
vdev_free(vd);
}
kmem_free(spa->spa_spares.sav_vdevs,
spa->spa_spares.sav_count * sizeof (void *));
}
if (spa->spa_spares.sav_config == NULL)
nspares = 0;
else
VERIFY0(nvlist_lookup_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, &spares, &nspares));
spa->spa_spares.sav_count = (int)nspares;
spa->spa_spares.sav_vdevs = NULL;
if (nspares == 0)
return;
/*
* Construct the array of vdevs, opening them to get status in the
* process. For each spare, there is potentially two different vdev_t
* structures associated with it: one in the list of spares (used only
* for basic validation purposes) and one in the active vdev
* configuration (if it's spared in). During this phase we open and
* validate each vdev on the spare list. If the vdev also exists in the
* active configuration, then we also mark this vdev as an active spare.
*/
spa->spa_spares.sav_vdevs = kmem_zalloc(nspares * sizeof (void *),
KM_SLEEP);
for (i = 0; i < spa->spa_spares.sav_count; i++) {
VERIFY(spa_config_parse(spa, &vd, spares[i], NULL, 0,
VDEV_ALLOC_SPARE) == 0);
ASSERT(vd != NULL);
spa->spa_spares.sav_vdevs[i] = vd;
if ((tvd = spa_lookup_by_guid(spa, vd->vdev_guid,
B_FALSE)) != NULL) {
if (!tvd->vdev_isspare)
spa_spare_add(tvd);
/*
* We only mark the spare active if we were successfully
* able to load the vdev. Otherwise, importing a pool
* with a bad active spare would result in strange
* behavior, because multiple pool would think the spare
* is actively in use.
*
* There is a vulnerability here to an equally bizarre
* circumstance, where a dead active spare is later
* brought back to life (onlined or otherwise). Given
* the rarity of this scenario, and the extra complexity
* it adds, we ignore the possibility.
*/
if (!vdev_is_dead(tvd))
spa_spare_activate(tvd);
}
vd->vdev_top = vd;
vd->vdev_aux = &spa->spa_spares;
if (vdev_open(vd) != 0)
continue;
if (vdev_validate_aux(vd) == 0)
spa_spare_add(vd);
}
/*
* Recompute the stashed list of spares, with status information
* this time.
*/
fnvlist_remove(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES);
spares = kmem_alloc(spa->spa_spares.sav_count * sizeof (void *),
KM_SLEEP);
for (i = 0; i < spa->spa_spares.sav_count; i++)
spares[i] = vdev_config_generate(spa,
spa->spa_spares.sav_vdevs[i], B_TRUE, VDEV_CONFIG_SPARE);
fnvlist_add_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, (const nvlist_t * const *)spares,
spa->spa_spares.sav_count);
for (i = 0; i < spa->spa_spares.sav_count; i++)
nvlist_free(spares[i]);
kmem_free(spares, spa->spa_spares.sav_count * sizeof (void *));
}
/*
* Load (or re-load) the current list of vdevs describing the active l2cache for
* this pool. When this is called, we have some form of basic information in
* 'spa_l2cache.sav_config'. We parse this into vdevs, try to open them, and
* then re-generate a more complete list including status information.
* Devices which are already active have their details maintained, and are
* not re-opened.
*/
void
spa_load_l2cache(spa_t *spa)
{
nvlist_t **l2cache = NULL;
uint_t nl2cache;
int i, j, oldnvdevs;
uint64_t guid;
vdev_t *vd, **oldvdevs, **newvdevs;
spa_aux_vdev_t *sav = &spa->spa_l2cache;
#ifndef _KERNEL
/*
* zdb opens both the current state of the pool and the
* checkpointed state (if present), with a different spa_t.
*
* As L2 caches are part of the ARC which is shared among open
* pools, we skip loading them when we load the checkpointed
* state of the pool.
*/
if (!spa_writeable(spa))
return;
#endif
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
oldvdevs = sav->sav_vdevs;
oldnvdevs = sav->sav_count;
sav->sav_vdevs = NULL;
sav->sav_count = 0;
if (sav->sav_config == NULL) {
nl2cache = 0;
newvdevs = NULL;
goto out;
}
VERIFY0(nvlist_lookup_nvlist_array(sav->sav_config,
ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache));
newvdevs = kmem_alloc(nl2cache * sizeof (void *), KM_SLEEP);
/*
* Process new nvlist of vdevs.
*/
for (i = 0; i < nl2cache; i++) {
guid = fnvlist_lookup_uint64(l2cache[i], ZPOOL_CONFIG_GUID);
newvdevs[i] = NULL;
for (j = 0; j < oldnvdevs; j++) {
vd = oldvdevs[j];
if (vd != NULL && guid == vd->vdev_guid) {
/*
* Retain previous vdev for add/remove ops.
*/
newvdevs[i] = vd;
oldvdevs[j] = NULL;
break;
}
}
if (newvdevs[i] == NULL) {
/*
* Create new vdev
*/
VERIFY(spa_config_parse(spa, &vd, l2cache[i], NULL, 0,
VDEV_ALLOC_L2CACHE) == 0);
ASSERT(vd != NULL);
newvdevs[i] = vd;
/*
* Commit this vdev as an l2cache device,
* even if it fails to open.
*/
spa_l2cache_add(vd);
vd->vdev_top = vd;
vd->vdev_aux = sav;
spa_l2cache_activate(vd);
if (vdev_open(vd) != 0)
continue;
(void) vdev_validate_aux(vd);
if (!vdev_is_dead(vd))
l2arc_add_vdev(spa, vd);
/*
* Upon cache device addition to a pool or pool
* creation with a cache device or if the header
* of the device is invalid we issue an async
* TRIM command for the whole device which will
* execute if l2arc_trim_ahead > 0.
*/
spa_async_request(spa, SPA_ASYNC_L2CACHE_TRIM);
}
}
sav->sav_vdevs = newvdevs;
sav->sav_count = (int)nl2cache;
/*
* Recompute the stashed list of l2cache devices, with status
* information this time.
*/
fnvlist_remove(sav->sav_config, ZPOOL_CONFIG_L2CACHE);
if (sav->sav_count > 0)
l2cache = kmem_alloc(sav->sav_count * sizeof (void *),
KM_SLEEP);
for (i = 0; i < sav->sav_count; i++)
l2cache[i] = vdev_config_generate(spa,
sav->sav_vdevs[i], B_TRUE, VDEV_CONFIG_L2CACHE);
fnvlist_add_nvlist_array(sav->sav_config, ZPOOL_CONFIG_L2CACHE,
(const nvlist_t * const *)l2cache, sav->sav_count);
out:
/*
* Purge vdevs that were dropped
*/
if (oldvdevs) {
for (i = 0; i < oldnvdevs; i++) {
uint64_t pool;
vd = oldvdevs[i];
if (vd != NULL) {
ASSERT(vd->vdev_isl2cache);
if (spa_l2cache_exists(vd->vdev_guid, &pool) &&
pool != 0ULL && l2arc_vdev_present(vd))
l2arc_remove_vdev(vd);
vdev_clear_stats(vd);
vdev_free(vd);
}
}
kmem_free(oldvdevs, oldnvdevs * sizeof (void *));
}
for (i = 0; i < sav->sav_count; i++)
nvlist_free(l2cache[i]);
if (sav->sav_count)
kmem_free(l2cache, sav->sav_count * sizeof (void *));
}
static int
load_nvlist(spa_t *spa, uint64_t obj, nvlist_t **value)
{
dmu_buf_t *db;
char *packed = NULL;
size_t nvsize = 0;
int error;
*value = NULL;
error = dmu_bonus_hold(spa->spa_meta_objset, obj, FTAG, &db);
if (error)
return (error);
nvsize = *(uint64_t *)db->db_data;
dmu_buf_rele(db, FTAG);
packed = vmem_alloc(nvsize, KM_SLEEP);
error = dmu_read(spa->spa_meta_objset, obj, 0, nvsize, packed,
DMU_READ_PREFETCH);
if (error == 0)
error = nvlist_unpack(packed, nvsize, value, 0);
vmem_free(packed, nvsize);
return (error);
}
/*
* Concrete top-level vdevs that are not missing and are not logs. At every
* spa_sync we write new uberblocks to at least SPA_SYNC_MIN_VDEVS core tvds.
*/
static uint64_t
spa_healthy_core_tvds(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
uint64_t tvds = 0;
for (uint64_t i = 0; i < rvd->vdev_children; i++) {
vdev_t *vd = rvd->vdev_child[i];
if (vd->vdev_islog)
continue;
if (vdev_is_concrete(vd) && !vdev_is_dead(vd))
tvds++;
}
return (tvds);
}
/*
* Checks to see if the given vdev could not be opened, in which case we post a
* sysevent to notify the autoreplace code that the device has been removed.
*/
static void
spa_check_removed(vdev_t *vd)
{
for (uint64_t c = 0; c < vd->vdev_children; c++)
spa_check_removed(vd->vdev_child[c]);
if (vd->vdev_ops->vdev_op_leaf && vdev_is_dead(vd) &&
vdev_is_concrete(vd)) {
zfs_post_autoreplace(vd->vdev_spa, vd);
spa_event_notify(vd->vdev_spa, vd, NULL, ESC_ZFS_VDEV_CHECK);
}
}
static int
spa_check_for_missing_logs(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
/*
* If we're doing a normal import, then build up any additional
* diagnostic information about missing log devices.
* We'll pass this up to the user for further processing.
*/
if (!(spa->spa_import_flags & ZFS_IMPORT_MISSING_LOG)) {
nvlist_t **child, *nv;
uint64_t idx = 0;
child = kmem_alloc(rvd->vdev_children * sizeof (nvlist_t *),
KM_SLEEP);
nv = fnvlist_alloc();
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
/*
* We consider a device as missing only if it failed
* to open (i.e. offline or faulted is not considered
* as missing).
*/
if (tvd->vdev_islog &&
tvd->vdev_state == VDEV_STATE_CANT_OPEN) {
child[idx++] = vdev_config_generate(spa, tvd,
B_FALSE, VDEV_CONFIG_MISSING);
}
}
if (idx > 0) {
fnvlist_add_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
(const nvlist_t * const *)child, idx);
fnvlist_add_nvlist(spa->spa_load_info,
ZPOOL_CONFIG_MISSING_DEVICES, nv);
for (uint64_t i = 0; i < idx; i++)
nvlist_free(child[i]);
}
nvlist_free(nv);
kmem_free(child, rvd->vdev_children * sizeof (char **));
if (idx > 0) {
spa_load_failed(spa, "some log devices are missing");
vdev_dbgmsg_print_tree(rvd, 2);
return (SET_ERROR(ENXIO));
}
} else {
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
if (tvd->vdev_islog &&
tvd->vdev_state == VDEV_STATE_CANT_OPEN) {
spa_set_log_state(spa, SPA_LOG_CLEAR);
spa_load_note(spa, "some log devices are "
"missing, ZIL is dropped.");
vdev_dbgmsg_print_tree(rvd, 2);
break;
}
}
}
return (0);
}
/*
* Check for missing log devices
*/
static boolean_t
spa_check_logs(spa_t *spa)
{
boolean_t rv = B_FALSE;
dsl_pool_t *dp = spa_get_dsl(spa);
switch (spa->spa_log_state) {
default:
break;
case SPA_LOG_MISSING:
/* need to recheck in case slog has been restored */
case SPA_LOG_UNKNOWN:
rv = (dmu_objset_find_dp(dp, dp->dp_root_dir_obj,
zil_check_log_chain, NULL, DS_FIND_CHILDREN) != 0);
if (rv)
spa_set_log_state(spa, SPA_LOG_MISSING);
break;
}
return (rv);
}
/*
* Passivate any log vdevs (note, does not apply to embedded log metaslabs).
*/
static boolean_t
spa_passivate_log(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
boolean_t slog_found = B_FALSE;
ASSERT(spa_config_held(spa, SCL_ALLOC, RW_WRITER));
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
if (tvd->vdev_islog) {
ASSERT3P(tvd->vdev_log_mg, ==, NULL);
metaslab_group_passivate(tvd->vdev_mg);
slog_found = B_TRUE;
}
}
return (slog_found);
}
/*
* Activate any log vdevs (note, does not apply to embedded log metaslabs).
*/
static void
spa_activate_log(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
ASSERT(spa_config_held(spa, SCL_ALLOC, RW_WRITER));
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
if (tvd->vdev_islog) {
ASSERT3P(tvd->vdev_log_mg, ==, NULL);
metaslab_group_activate(tvd->vdev_mg);
}
}
}
int
spa_reset_logs(spa_t *spa)
{
int error;
error = dmu_objset_find(spa_name(spa), zil_reset,
NULL, DS_FIND_CHILDREN);
if (error == 0) {
/*
* We successfully offlined the log device, sync out the
* current txg so that the "stubby" block can be removed
* by zil_sync().
*/
txg_wait_synced(spa->spa_dsl_pool, 0);
}
return (error);
}
static void
spa_aux_check_removed(spa_aux_vdev_t *sav)
{
for (int i = 0; i < sav->sav_count; i++)
spa_check_removed(sav->sav_vdevs[i]);
}
void
spa_claim_notify(zio_t *zio)
{
spa_t *spa = zio->io_spa;
if (zio->io_error)
return;
mutex_enter(&spa->spa_props_lock); /* any mutex will do */
if (spa->spa_claim_max_txg < zio->io_bp->blk_birth)
spa->spa_claim_max_txg = zio->io_bp->blk_birth;
mutex_exit(&spa->spa_props_lock);
}
typedef struct spa_load_error {
boolean_t sle_verify_data;
uint64_t sle_meta_count;
uint64_t sle_data_count;
} spa_load_error_t;
static void
spa_load_verify_done(zio_t *zio)
{
blkptr_t *bp = zio->io_bp;
spa_load_error_t *sle = zio->io_private;
dmu_object_type_t type = BP_GET_TYPE(bp);
int error = zio->io_error;
spa_t *spa = zio->io_spa;
abd_free(zio->io_abd);
if (error) {
if ((BP_GET_LEVEL(bp) != 0 || DMU_OT_IS_METADATA(type)) &&
type != DMU_OT_INTENT_LOG)
atomic_inc_64(&sle->sle_meta_count);
else
atomic_inc_64(&sle->sle_data_count);
}
mutex_enter(&spa->spa_scrub_lock);
spa->spa_load_verify_bytes -= BP_GET_PSIZE(bp);
cv_broadcast(&spa->spa_scrub_io_cv);
mutex_exit(&spa->spa_scrub_lock);
}
/*
* Maximum number of inflight bytes is the log2 fraction of the arc size.
* By default, we set it to 1/16th of the arc.
*/
static uint_t spa_load_verify_shift = 4;
static int spa_load_verify_metadata = B_TRUE;
static int spa_load_verify_data = B_TRUE;
static int
spa_load_verify_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
const zbookmark_phys_t *zb, const dnode_phys_t *dnp, void *arg)
{
zio_t *rio = arg;
spa_load_error_t *sle = rio->io_private;
(void) zilog, (void) dnp;
/*
* Note: normally this routine will not be called if
* spa_load_verify_metadata is not set. However, it may be useful
* to manually set the flag after the traversal has begun.
*/
if (!spa_load_verify_metadata)
return (0);
/*
* Sanity check the block pointer in order to detect obvious damage
* before using the contents in subsequent checks or in zio_read().
* When damaged consider it to be a metadata error since we cannot
* trust the BP_GET_TYPE and BP_GET_LEVEL values.
*/
if (!zfs_blkptr_verify(spa, bp, BLK_CONFIG_NEEDED, BLK_VERIFY_LOG)) {
atomic_inc_64(&sle->sle_meta_count);
return (0);
}
if (zb->zb_level == ZB_DNODE_LEVEL || BP_IS_HOLE(bp) ||
BP_IS_EMBEDDED(bp) || BP_IS_REDACTED(bp))
return (0);
if (!BP_IS_METADATA(bp) &&
(!spa_load_verify_data || !sle->sle_verify_data))
return (0);
uint64_t maxinflight_bytes =
arc_target_bytes() >> spa_load_verify_shift;
size_t size = BP_GET_PSIZE(bp);
mutex_enter(&spa->spa_scrub_lock);
while (spa->spa_load_verify_bytes >= maxinflight_bytes)
cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock);
spa->spa_load_verify_bytes += size;
mutex_exit(&spa->spa_scrub_lock);
zio_nowait(zio_read(rio, spa, bp, abd_alloc_for_io(size, B_FALSE), size,
spa_load_verify_done, rio->io_private, ZIO_PRIORITY_SCRUB,
ZIO_FLAG_SPECULATIVE | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SCRUB | ZIO_FLAG_RAW, zb));
return (0);
}
static int
verify_dataset_name_len(dsl_pool_t *dp, dsl_dataset_t *ds, void *arg)
{
(void) dp, (void) arg;
if (dsl_dataset_namelen(ds) >= ZFS_MAX_DATASET_NAME_LEN)
return (SET_ERROR(ENAMETOOLONG));
return (0);
}
static int
spa_load_verify(spa_t *spa)
{
zio_t *rio;
spa_load_error_t sle = { 0 };
zpool_load_policy_t policy;
boolean_t verify_ok = B_FALSE;
int error = 0;
zpool_get_load_policy(spa->spa_config, &policy);
if (policy.zlp_rewind & ZPOOL_NEVER_REWIND ||
policy.zlp_maxmeta == UINT64_MAX)
return (0);
dsl_pool_config_enter(spa->spa_dsl_pool, FTAG);
error = dmu_objset_find_dp(spa->spa_dsl_pool,
spa->spa_dsl_pool->dp_root_dir_obj, verify_dataset_name_len, NULL,
DS_FIND_CHILDREN);
dsl_pool_config_exit(spa->spa_dsl_pool, FTAG);
if (error != 0)
return (error);
/*
* Verify data only if we are rewinding or error limit was set.
* Otherwise nothing except dbgmsg care about it to waste time.
*/
sle.sle_verify_data = (policy.zlp_rewind & ZPOOL_REWIND_MASK) ||
(policy.zlp_maxdata < UINT64_MAX);
rio = zio_root(spa, NULL, &sle,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE);
if (spa_load_verify_metadata) {
if (spa->spa_extreme_rewind) {
spa_load_note(spa, "performing a complete scan of the "
"pool since extreme rewind is on. This may take "
"a very long time.\n (spa_load_verify_data=%u, "
"spa_load_verify_metadata=%u)",
spa_load_verify_data, spa_load_verify_metadata);
}
error = traverse_pool(spa, spa->spa_verify_min_txg,
TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA |
TRAVERSE_NO_DECRYPT, spa_load_verify_cb, rio);
}
(void) zio_wait(rio);
ASSERT0(spa->spa_load_verify_bytes);
spa->spa_load_meta_errors = sle.sle_meta_count;
spa->spa_load_data_errors = sle.sle_data_count;
if (sle.sle_meta_count != 0 || sle.sle_data_count != 0) {
spa_load_note(spa, "spa_load_verify found %llu metadata errors "
"and %llu data errors", (u_longlong_t)sle.sle_meta_count,
(u_longlong_t)sle.sle_data_count);
}
if (spa_load_verify_dryrun ||
(!error && sle.sle_meta_count <= policy.zlp_maxmeta &&
sle.sle_data_count <= policy.zlp_maxdata)) {
int64_t loss = 0;
verify_ok = B_TRUE;
spa->spa_load_txg = spa->spa_uberblock.ub_txg;
spa->spa_load_txg_ts = spa->spa_uberblock.ub_timestamp;
loss = spa->spa_last_ubsync_txg_ts - spa->spa_load_txg_ts;
fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_LOAD_TIME,
spa->spa_load_txg_ts);
fnvlist_add_int64(spa->spa_load_info, ZPOOL_CONFIG_REWIND_TIME,
loss);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_LOAD_META_ERRORS, sle.sle_meta_count);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_LOAD_DATA_ERRORS, sle.sle_data_count);
} else {
spa->spa_load_max_txg = spa->spa_uberblock.ub_txg;
}
if (spa_load_verify_dryrun)
return (0);
if (error) {
if (error != ENXIO && error != EIO)
error = SET_ERROR(EIO);
return (error);
}
return (verify_ok ? 0 : EIO);
}
/*
* Find a value in the pool props object.
*/
static void
spa_prop_find(spa_t *spa, zpool_prop_t prop, uint64_t *val)
{
(void) zap_lookup(spa->spa_meta_objset, spa->spa_pool_props_object,
zpool_prop_to_name(prop), sizeof (uint64_t), 1, val);
}
/*
* Find a value in the pool directory object.
*/
static int
spa_dir_prop(spa_t *spa, const char *name, uint64_t *val, boolean_t log_enoent)
{
int error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
name, sizeof (uint64_t), 1, val);
if (error != 0 && (error != ENOENT || log_enoent)) {
spa_load_failed(spa, "couldn't get '%s' value in MOS directory "
"[error=%d]", name, error);
}
return (error);
}
static int
spa_vdev_err(vdev_t *vdev, vdev_aux_t aux, int err)
{
vdev_set_state(vdev, B_TRUE, VDEV_STATE_CANT_OPEN, aux);
return (SET_ERROR(err));
}
boolean_t
spa_livelist_delete_check(spa_t *spa)
{
return (spa->spa_livelists_to_delete != 0);
}
static boolean_t
spa_livelist_delete_cb_check(void *arg, zthr_t *z)
{
(void) z;
spa_t *spa = arg;
return (spa_livelist_delete_check(spa));
}
static int
delete_blkptr_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
spa_t *spa = arg;
zio_free(spa, tx->tx_txg, bp);
dsl_dir_diduse_space(tx->tx_pool->dp_free_dir, DD_USED_HEAD,
-bp_get_dsize_sync(spa, bp),
-BP_GET_PSIZE(bp), -BP_GET_UCSIZE(bp), tx);
return (0);
}
static int
dsl_get_next_livelist_obj(objset_t *os, uint64_t zap_obj, uint64_t *llp)
{
int err;
zap_cursor_t zc;
zap_attribute_t za;
zap_cursor_init(&zc, os, zap_obj);
err = zap_cursor_retrieve(&zc, &za);
zap_cursor_fini(&zc);
if (err == 0)
*llp = za.za_first_integer;
return (err);
}
/*
* Components of livelist deletion that must be performed in syncing
* context: freeing block pointers and updating the pool-wide data
* structures to indicate how much work is left to do
*/
typedef struct sublist_delete_arg {
spa_t *spa;
dsl_deadlist_t *ll;
uint64_t key;
bplist_t *to_free;
} sublist_delete_arg_t;
static void
sublist_delete_sync(void *arg, dmu_tx_t *tx)
{
sublist_delete_arg_t *sda = arg;
spa_t *spa = sda->spa;
dsl_deadlist_t *ll = sda->ll;
uint64_t key = sda->key;
bplist_t *to_free = sda->to_free;
bplist_iterate(to_free, delete_blkptr_cb, spa, tx);
dsl_deadlist_remove_entry(ll, key, tx);
}
typedef struct livelist_delete_arg {
spa_t *spa;
uint64_t ll_obj;
uint64_t zap_obj;
} livelist_delete_arg_t;
static void
livelist_delete_sync(void *arg, dmu_tx_t *tx)
{
livelist_delete_arg_t *lda = arg;
spa_t *spa = lda->spa;
uint64_t ll_obj = lda->ll_obj;
uint64_t zap_obj = lda->zap_obj;
objset_t *mos = spa->spa_meta_objset;
uint64_t count;
/* free the livelist and decrement the feature count */
VERIFY0(zap_remove_int(mos, zap_obj, ll_obj, tx));
dsl_deadlist_free(mos, ll_obj, tx);
spa_feature_decr(spa, SPA_FEATURE_LIVELIST, tx);
VERIFY0(zap_count(mos, zap_obj, &count));
if (count == 0) {
/* no more livelists to delete */
VERIFY0(zap_remove(mos, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_DELETED_CLONES, tx));
VERIFY0(zap_destroy(mos, zap_obj, tx));
spa->spa_livelists_to_delete = 0;
spa_notify_waiters(spa);
}
}
/*
* Load in the value for the livelist to be removed and open it. Then,
* load its first sublist and determine which block pointers should actually
* be freed. Then, call a synctask which performs the actual frees and updates
* the pool-wide livelist data.
*/
static void
spa_livelist_delete_cb(void *arg, zthr_t *z)
{
spa_t *spa = arg;
uint64_t ll_obj = 0, count;
objset_t *mos = spa->spa_meta_objset;
uint64_t zap_obj = spa->spa_livelists_to_delete;
/*
* Determine the next livelist to delete. This function should only
* be called if there is at least one deleted clone.
*/
VERIFY0(dsl_get_next_livelist_obj(mos, zap_obj, &ll_obj));
VERIFY0(zap_count(mos, ll_obj, &count));
if (count > 0) {
dsl_deadlist_t *ll;
dsl_deadlist_entry_t *dle;
bplist_t to_free;
ll = kmem_zalloc(sizeof (dsl_deadlist_t), KM_SLEEP);
dsl_deadlist_open(ll, mos, ll_obj);
dle = dsl_deadlist_first(ll);
ASSERT3P(dle, !=, NULL);
bplist_create(&to_free);
int err = dsl_process_sub_livelist(&dle->dle_bpobj, &to_free,
z, NULL);
if (err == 0) {
sublist_delete_arg_t sync_arg = {
.spa = spa,
.ll = ll,
.key = dle->dle_mintxg,
.to_free = &to_free
};
zfs_dbgmsg("deleting sublist (id %llu) from"
" livelist %llu, %lld remaining",
(u_longlong_t)dle->dle_bpobj.bpo_object,
(u_longlong_t)ll_obj, (longlong_t)count - 1);
VERIFY0(dsl_sync_task(spa_name(spa), NULL,
sublist_delete_sync, &sync_arg, 0,
ZFS_SPACE_CHECK_DESTROY));
} else {
VERIFY3U(err, ==, EINTR);
}
bplist_clear(&to_free);
bplist_destroy(&to_free);
dsl_deadlist_close(ll);
kmem_free(ll, sizeof (dsl_deadlist_t));
} else {
livelist_delete_arg_t sync_arg = {
.spa = spa,
.ll_obj = ll_obj,
.zap_obj = zap_obj
};
zfs_dbgmsg("deletion of livelist %llu completed",
(u_longlong_t)ll_obj);
VERIFY0(dsl_sync_task(spa_name(spa), NULL, livelist_delete_sync,
&sync_arg, 0, ZFS_SPACE_CHECK_DESTROY));
}
}
static void
spa_start_livelist_destroy_thread(spa_t *spa)
{
ASSERT3P(spa->spa_livelist_delete_zthr, ==, NULL);
spa->spa_livelist_delete_zthr =
zthr_create("z_livelist_destroy",
spa_livelist_delete_cb_check, spa_livelist_delete_cb, spa,
minclsyspri);
}
typedef struct livelist_new_arg {
bplist_t *allocs;
bplist_t *frees;
} livelist_new_arg_t;
static int
livelist_track_new_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
ASSERT(tx == NULL);
livelist_new_arg_t *lna = arg;
if (bp_freed) {
bplist_append(lna->frees, bp);
} else {
bplist_append(lna->allocs, bp);
zfs_livelist_condense_new_alloc++;
}
return (0);
}
typedef struct livelist_condense_arg {
spa_t *spa;
bplist_t to_keep;
uint64_t first_size;
uint64_t next_size;
} livelist_condense_arg_t;
static void
spa_livelist_condense_sync(void *arg, dmu_tx_t *tx)
{
livelist_condense_arg_t *lca = arg;
spa_t *spa = lca->spa;
bplist_t new_frees;
dsl_dataset_t *ds = spa->spa_to_condense.ds;
/* Have we been cancelled? */
if (spa->spa_to_condense.cancelled) {
zfs_livelist_condense_sync_cancel++;
goto out;
}
dsl_deadlist_entry_t *first = spa->spa_to_condense.first;
dsl_deadlist_entry_t *next = spa->spa_to_condense.next;
dsl_deadlist_t *ll = &ds->ds_dir->dd_livelist;
/*
* It's possible that the livelist was changed while the zthr was
* running. Therefore, we need to check for new blkptrs in the two
* entries being condensed and continue to track them in the livelist.
* Because of the way we handle remapped blkptrs (see dbuf_remap_impl),
* it's possible that the newly added blkptrs are FREEs or ALLOCs so
* we need to sort them into two different bplists.
*/
uint64_t first_obj = first->dle_bpobj.bpo_object;
uint64_t next_obj = next->dle_bpobj.bpo_object;
uint64_t cur_first_size = first->dle_bpobj.bpo_phys->bpo_num_blkptrs;
uint64_t cur_next_size = next->dle_bpobj.bpo_phys->bpo_num_blkptrs;
bplist_create(&new_frees);
livelist_new_arg_t new_bps = {
.allocs = &lca->to_keep,
.frees = &new_frees,
};
if (cur_first_size > lca->first_size) {
VERIFY0(livelist_bpobj_iterate_from_nofree(&first->dle_bpobj,
livelist_track_new_cb, &new_bps, lca->first_size));
}
if (cur_next_size > lca->next_size) {
VERIFY0(livelist_bpobj_iterate_from_nofree(&next->dle_bpobj,
livelist_track_new_cb, &new_bps, lca->next_size));
}
dsl_deadlist_clear_entry(first, ll, tx);
ASSERT(bpobj_is_empty(&first->dle_bpobj));
dsl_deadlist_remove_entry(ll, next->dle_mintxg, tx);
bplist_iterate(&lca->to_keep, dsl_deadlist_insert_alloc_cb, ll, tx);
bplist_iterate(&new_frees, dsl_deadlist_insert_free_cb, ll, tx);
bplist_destroy(&new_frees);
char dsname[ZFS_MAX_DATASET_NAME_LEN];
dsl_dataset_name(ds, dsname);
zfs_dbgmsg("txg %llu condensing livelist of %s (id %llu), bpobj %llu "
"(%llu blkptrs) and bpobj %llu (%llu blkptrs) -> bpobj %llu "
"(%llu blkptrs)", (u_longlong_t)tx->tx_txg, dsname,
(u_longlong_t)ds->ds_object, (u_longlong_t)first_obj,
(u_longlong_t)cur_first_size, (u_longlong_t)next_obj,
(u_longlong_t)cur_next_size,
(u_longlong_t)first->dle_bpobj.bpo_object,
(u_longlong_t)first->dle_bpobj.bpo_phys->bpo_num_blkptrs);
out:
dmu_buf_rele(ds->ds_dbuf, spa);
spa->spa_to_condense.ds = NULL;
bplist_clear(&lca->to_keep);
bplist_destroy(&lca->to_keep);
kmem_free(lca, sizeof (livelist_condense_arg_t));
spa->spa_to_condense.syncing = B_FALSE;
}
static void
spa_livelist_condense_cb(void *arg, zthr_t *t)
{
while (zfs_livelist_condense_zthr_pause &&
!(zthr_has_waiters(t) || zthr_iscancelled(t)))
delay(1);
spa_t *spa = arg;
dsl_deadlist_entry_t *first = spa->spa_to_condense.first;
dsl_deadlist_entry_t *next = spa->spa_to_condense.next;
uint64_t first_size, next_size;
livelist_condense_arg_t *lca =
kmem_alloc(sizeof (livelist_condense_arg_t), KM_SLEEP);
bplist_create(&lca->to_keep);
/*
* Process the livelists (matching FREEs and ALLOCs) in open context
* so we have minimal work in syncing context to condense.
*
* We save bpobj sizes (first_size and next_size) to use later in
* syncing context to determine if entries were added to these sublists
* while in open context. This is possible because the clone is still
* active and open for normal writes and we want to make sure the new,
* unprocessed blockpointers are inserted into the livelist normally.
*
* Note that dsl_process_sub_livelist() both stores the size number of
* blockpointers and iterates over them while the bpobj's lock held, so
* the sizes returned to us are consistent which what was actually
* processed.
*/
int err = dsl_process_sub_livelist(&first->dle_bpobj, &lca->to_keep, t,
&first_size);
if (err == 0)
err = dsl_process_sub_livelist(&next->dle_bpobj, &lca->to_keep,
t, &next_size);
if (err == 0) {
while (zfs_livelist_condense_sync_pause &&
!(zthr_has_waiters(t) || zthr_iscancelled(t)))
delay(1);
dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir);
dmu_tx_mark_netfree(tx);
dmu_tx_hold_space(tx, 1);
err = dmu_tx_assign(tx, TXG_NOWAIT | TXG_NOTHROTTLE);
if (err == 0) {
/*
* Prevent the condense zthr restarting before
* the synctask completes.
*/
spa->spa_to_condense.syncing = B_TRUE;
lca->spa = spa;
lca->first_size = first_size;
lca->next_size = next_size;
dsl_sync_task_nowait(spa_get_dsl(spa),
spa_livelist_condense_sync, lca, tx);
dmu_tx_commit(tx);
return;
}
}
/*
* Condensing can not continue: either it was externally stopped or
* we were unable to assign to a tx because the pool has run out of
* space. In the second case, we'll just end up trying to condense
* again in a later txg.
*/
ASSERT(err != 0);
bplist_clear(&lca->to_keep);
bplist_destroy(&lca->to_keep);
kmem_free(lca, sizeof (livelist_condense_arg_t));
dmu_buf_rele(spa->spa_to_condense.ds->ds_dbuf, spa);
spa->spa_to_condense.ds = NULL;
if (err == EINTR)
zfs_livelist_condense_zthr_cancel++;
}
/*
* Check that there is something to condense but that a condense is not
* already in progress and that condensing has not been cancelled.
*/
static boolean_t
spa_livelist_condense_cb_check(void *arg, zthr_t *z)
{
(void) z;
spa_t *spa = arg;
if ((spa->spa_to_condense.ds != NULL) &&
(spa->spa_to_condense.syncing == B_FALSE) &&
(spa->spa_to_condense.cancelled == B_FALSE)) {
return (B_TRUE);
}
return (B_FALSE);
}
static void
spa_start_livelist_condensing_thread(spa_t *spa)
{
spa->spa_to_condense.ds = NULL;
spa->spa_to_condense.first = NULL;
spa->spa_to_condense.next = NULL;
spa->spa_to_condense.syncing = B_FALSE;
spa->spa_to_condense.cancelled = B_FALSE;
ASSERT3P(spa->spa_livelist_condense_zthr, ==, NULL);
spa->spa_livelist_condense_zthr =
zthr_create("z_livelist_condense",
spa_livelist_condense_cb_check,
spa_livelist_condense_cb, spa, minclsyspri);
}
static void
spa_spawn_aux_threads(spa_t *spa)
{
ASSERT(spa_writeable(spa));
ASSERT(MUTEX_HELD(&spa_namespace_lock));
+ spa_start_raidz_expansion_thread(spa);
spa_start_indirect_condensing_thread(spa);
spa_start_livelist_destroy_thread(spa);
spa_start_livelist_condensing_thread(spa);
ASSERT3P(spa->spa_checkpoint_discard_zthr, ==, NULL);
spa->spa_checkpoint_discard_zthr =
zthr_create("z_checkpoint_discard",
spa_checkpoint_discard_thread_check,
spa_checkpoint_discard_thread, spa, minclsyspri);
}
/*
* Fix up config after a partly-completed split. This is done with the
* ZPOOL_CONFIG_SPLIT nvlist. Both the splitting pool and the split-off
* pool have that entry in their config, but only the splitting one contains
* a list of all the guids of the vdevs that are being split off.
*
* This function determines what to do with that list: either rejoin
* all the disks to the pool, or complete the splitting process. To attempt
* the rejoin, each disk that is offlined is marked online again, and
* we do a reopen() call. If the vdev label for every disk that was
* marked online indicates it was successfully split off (VDEV_AUX_SPLIT_POOL)
* then we call vdev_split() on each disk, and complete the split.
*
* Otherwise we leave the config alone, with all the vdevs in place in
* the original pool.
*/
static void
spa_try_repair(spa_t *spa, nvlist_t *config)
{
uint_t extracted;
uint64_t *glist;
uint_t i, gcount;
nvlist_t *nvl;
vdev_t **vd;
boolean_t attempt_reopen;
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_SPLIT, &nvl) != 0)
return;
/* check that the config is complete */
if (nvlist_lookup_uint64_array(nvl, ZPOOL_CONFIG_SPLIT_LIST,
&glist, &gcount) != 0)
return;
vd = kmem_zalloc(gcount * sizeof (vdev_t *), KM_SLEEP);
/* attempt to online all the vdevs & validate */
attempt_reopen = B_TRUE;
for (i = 0; i < gcount; i++) {
if (glist[i] == 0) /* vdev is hole */
continue;
vd[i] = spa_lookup_by_guid(spa, glist[i], B_FALSE);
if (vd[i] == NULL) {
/*
* Don't bother attempting to reopen the disks;
* just do the split.
*/
attempt_reopen = B_FALSE;
} else {
/* attempt to re-online it */
vd[i]->vdev_offline = B_FALSE;
}
}
if (attempt_reopen) {
vdev_reopen(spa->spa_root_vdev);
/* check each device to see what state it's in */
for (extracted = 0, i = 0; i < gcount; i++) {
if (vd[i] != NULL &&
vd[i]->vdev_stat.vs_aux != VDEV_AUX_SPLIT_POOL)
break;
++extracted;
}
}
/*
* If every disk has been moved to the new pool, or if we never
* even attempted to look at them, then we split them off for
* good.
*/
if (!attempt_reopen || gcount == extracted) {
for (i = 0; i < gcount; i++)
if (vd[i] != NULL)
vdev_split(vd[i]);
vdev_reopen(spa->spa_root_vdev);
}
kmem_free(vd, gcount * sizeof (vdev_t *));
}
static int
spa_load(spa_t *spa, spa_load_state_t state, spa_import_type_t type)
{
const char *ereport = FM_EREPORT_ZFS_POOL;
int error;
spa->spa_load_state = state;
(void) spa_import_progress_set_state(spa_guid(spa),
spa_load_state(spa));
gethrestime(&spa->spa_loaded_ts);
error = spa_load_impl(spa, type, &ereport);
/*
* Don't count references from objsets that are already closed
* and are making their way through the eviction process.
*/
spa_evicting_os_wait(spa);
spa->spa_minref = zfs_refcount_count(&spa->spa_refcount);
if (error) {
if (error != EEXIST) {
spa->spa_loaded_ts.tv_sec = 0;
spa->spa_loaded_ts.tv_nsec = 0;
}
if (error != EBADF) {
(void) zfs_ereport_post(ereport, spa,
NULL, NULL, NULL, 0);
}
}
spa->spa_load_state = error ? SPA_LOAD_ERROR : SPA_LOAD_NONE;
spa->spa_ena = 0;
(void) spa_import_progress_set_state(spa_guid(spa),
spa_load_state(spa));
return (error);
}
#ifdef ZFS_DEBUG
/*
* Count the number of per-vdev ZAPs associated with all of the vdevs in the
* vdev tree rooted in the given vd, and ensure that each ZAP is present in the
* spa's per-vdev ZAP list.
*/
static uint64_t
vdev_count_verify_zaps(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
uint64_t total = 0;
if (spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_AVZ_V2) &&
vd->vdev_root_zap != 0) {
total++;
ASSERT0(zap_lookup_int(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, vd->vdev_root_zap));
}
if (vd->vdev_top_zap != 0) {
total++;
ASSERT0(zap_lookup_int(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, vd->vdev_top_zap));
}
if (vd->vdev_leaf_zap != 0) {
total++;
ASSERT0(zap_lookup_int(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, vd->vdev_leaf_zap));
}
for (uint64_t i = 0; i < vd->vdev_children; i++) {
total += vdev_count_verify_zaps(vd->vdev_child[i]);
}
return (total);
}
#else
#define vdev_count_verify_zaps(vd) ((void) sizeof (vd), 0)
#endif
/*
* Determine whether the activity check is required.
*/
static boolean_t
spa_activity_check_required(spa_t *spa, uberblock_t *ub, nvlist_t *label,
nvlist_t *config)
{
uint64_t state = 0;
uint64_t hostid = 0;
uint64_t tryconfig_txg = 0;
uint64_t tryconfig_timestamp = 0;
uint16_t tryconfig_mmp_seq = 0;
nvlist_t *nvinfo;
if (nvlist_exists(config, ZPOOL_CONFIG_LOAD_INFO)) {
nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
(void) nvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_TXG,
&tryconfig_txg);
(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_TIMESTAMP,
&tryconfig_timestamp);
(void) nvlist_lookup_uint16(nvinfo, ZPOOL_CONFIG_MMP_SEQ,
&tryconfig_mmp_seq);
}
(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE, &state);
/*
* Disable the MMP activity check - This is used by zdb which
* is intended to be used on potentially active pools.
*/
if (spa->spa_import_flags & ZFS_IMPORT_SKIP_MMP)
return (B_FALSE);
/*
* Skip the activity check when the MMP feature is disabled.
*/
if (ub->ub_mmp_magic == MMP_MAGIC && ub->ub_mmp_delay == 0)
return (B_FALSE);
/*
* If the tryconfig_ values are nonzero, they are the results of an
* earlier tryimport. If they all match the uberblock we just found,
* then the pool has not changed and we return false so we do not test
* a second time.
*/
if (tryconfig_txg && tryconfig_txg == ub->ub_txg &&
tryconfig_timestamp && tryconfig_timestamp == ub->ub_timestamp &&
tryconfig_mmp_seq && tryconfig_mmp_seq ==
(MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0))
return (B_FALSE);
/*
* Allow the activity check to be skipped when importing the pool
* on the same host which last imported it. Since the hostid from
* configuration may be stale use the one read from the label.
*/
if (nvlist_exists(label, ZPOOL_CONFIG_HOSTID))
hostid = fnvlist_lookup_uint64(label, ZPOOL_CONFIG_HOSTID);
if (hostid == spa_get_hostid(spa))
return (B_FALSE);
/*
* Skip the activity test when the pool was cleanly exported.
*/
if (state != POOL_STATE_ACTIVE)
return (B_FALSE);
return (B_TRUE);
}
/*
* Nanoseconds the activity check must watch for changes on-disk.
*/
static uint64_t
spa_activity_check_duration(spa_t *spa, uberblock_t *ub)
{
uint64_t import_intervals = MAX(zfs_multihost_import_intervals, 1);
uint64_t multihost_interval = MSEC2NSEC(
MMP_INTERVAL_OK(zfs_multihost_interval));
uint64_t import_delay = MAX(NANOSEC, import_intervals *
multihost_interval);
/*
* Local tunables determine a minimum duration except for the case
* where we know when the remote host will suspend the pool if MMP
* writes do not land.
*
* See Big Theory comment at the top of mmp.c for the reasoning behind
* these cases and times.
*/
ASSERT(MMP_IMPORT_SAFETY_FACTOR >= 100);
if (MMP_INTERVAL_VALID(ub) && MMP_FAIL_INT_VALID(ub) &&
MMP_FAIL_INT(ub) > 0) {
/* MMP on remote host will suspend pool after failed writes */
import_delay = MMP_FAIL_INT(ub) * MSEC2NSEC(MMP_INTERVAL(ub)) *
MMP_IMPORT_SAFETY_FACTOR / 100;
zfs_dbgmsg("fail_intvals>0 import_delay=%llu ub_mmp "
"mmp_fails=%llu ub_mmp mmp_interval=%llu "
"import_intervals=%llu", (u_longlong_t)import_delay,
(u_longlong_t)MMP_FAIL_INT(ub),
(u_longlong_t)MMP_INTERVAL(ub),
(u_longlong_t)import_intervals);
} else if (MMP_INTERVAL_VALID(ub) && MMP_FAIL_INT_VALID(ub) &&
MMP_FAIL_INT(ub) == 0) {
/* MMP on remote host will never suspend pool */
import_delay = MAX(import_delay, (MSEC2NSEC(MMP_INTERVAL(ub)) +
ub->ub_mmp_delay) * import_intervals);
zfs_dbgmsg("fail_intvals=0 import_delay=%llu ub_mmp "
"mmp_interval=%llu ub_mmp_delay=%llu "
"import_intervals=%llu", (u_longlong_t)import_delay,
(u_longlong_t)MMP_INTERVAL(ub),
(u_longlong_t)ub->ub_mmp_delay,
(u_longlong_t)import_intervals);
} else if (MMP_VALID(ub)) {
/*
* zfs-0.7 compatibility case
*/
import_delay = MAX(import_delay, (multihost_interval +
ub->ub_mmp_delay) * import_intervals);
zfs_dbgmsg("import_delay=%llu ub_mmp_delay=%llu "
"import_intervals=%llu leaves=%u",
(u_longlong_t)import_delay,
(u_longlong_t)ub->ub_mmp_delay,
(u_longlong_t)import_intervals,
vdev_count_leaves(spa));
} else {
/* Using local tunings is the only reasonable option */
zfs_dbgmsg("pool last imported on non-MMP aware "
"host using import_delay=%llu multihost_interval=%llu "
"import_intervals=%llu", (u_longlong_t)import_delay,
(u_longlong_t)multihost_interval,
(u_longlong_t)import_intervals);
}
return (import_delay);
}
/*
* Perform the import activity check. If the user canceled the import or
* we detected activity then fail.
*/
static int
spa_activity_check(spa_t *spa, uberblock_t *ub, nvlist_t *config)
{
uint64_t txg = ub->ub_txg;
uint64_t timestamp = ub->ub_timestamp;
uint64_t mmp_config = ub->ub_mmp_config;
uint16_t mmp_seq = MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0;
uint64_t import_delay;
hrtime_t import_expire;
nvlist_t *mmp_label = NULL;
vdev_t *rvd = spa->spa_root_vdev;
kcondvar_t cv;
kmutex_t mtx;
int error = 0;
cv_init(&cv, NULL, CV_DEFAULT, NULL);
mutex_init(&mtx, NULL, MUTEX_DEFAULT, NULL);
mutex_enter(&mtx);
/*
* If ZPOOL_CONFIG_MMP_TXG is present an activity check was performed
* during the earlier tryimport. If the txg recorded there is 0 then
* the pool is known to be active on another host.
*
* Otherwise, the pool might be in use on another host. Check for
* changes in the uberblocks on disk if necessary.
*/
if (nvlist_exists(config, ZPOOL_CONFIG_LOAD_INFO)) {
nvlist_t *nvinfo = fnvlist_lookup_nvlist(config,
ZPOOL_CONFIG_LOAD_INFO);
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_TXG) &&
fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_TXG) == 0) {
vdev_uberblock_load(rvd, ub, &mmp_label);
error = SET_ERROR(EREMOTEIO);
goto out;
}
}
import_delay = spa_activity_check_duration(spa, ub);
/* Add a small random factor in case of simultaneous imports (0-25%) */
import_delay += import_delay * random_in_range(250) / 1000;
import_expire = gethrtime() + import_delay;
while (gethrtime() < import_expire) {
(void) spa_import_progress_set_mmp_check(spa_guid(spa),
NSEC2SEC(import_expire - gethrtime()));
vdev_uberblock_load(rvd, ub, &mmp_label);
if (txg != ub->ub_txg || timestamp != ub->ub_timestamp ||
mmp_seq != (MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0)) {
zfs_dbgmsg("multihost activity detected "
"txg %llu ub_txg %llu "
"timestamp %llu ub_timestamp %llu "
"mmp_config %#llx ub_mmp_config %#llx",
(u_longlong_t)txg, (u_longlong_t)ub->ub_txg,
(u_longlong_t)timestamp,
(u_longlong_t)ub->ub_timestamp,
(u_longlong_t)mmp_config,
(u_longlong_t)ub->ub_mmp_config);
error = SET_ERROR(EREMOTEIO);
break;
}
if (mmp_label) {
nvlist_free(mmp_label);
mmp_label = NULL;
}
error = cv_timedwait_sig(&cv, &mtx, ddi_get_lbolt() + hz);
if (error != -1) {
error = SET_ERROR(EINTR);
break;
}
error = 0;
}
out:
mutex_exit(&mtx);
mutex_destroy(&mtx);
cv_destroy(&cv);
/*
* If the pool is determined to be active store the status in the
* spa->spa_load_info nvlist. If the remote hostname or hostid are
* available from configuration read from disk store them as well.
* This allows 'zpool import' to generate a more useful message.
*
* ZPOOL_CONFIG_MMP_STATE - observed pool status (mandatory)
* ZPOOL_CONFIG_MMP_HOSTNAME - hostname from the active pool
* ZPOOL_CONFIG_MMP_HOSTID - hostid from the active pool
*/
if (error == EREMOTEIO) {
const char *hostname = "<unknown>";
uint64_t hostid = 0;
if (mmp_label) {
if (nvlist_exists(mmp_label, ZPOOL_CONFIG_HOSTNAME)) {
hostname = fnvlist_lookup_string(mmp_label,
ZPOOL_CONFIG_HOSTNAME);
fnvlist_add_string(spa->spa_load_info,
ZPOOL_CONFIG_MMP_HOSTNAME, hostname);
}
if (nvlist_exists(mmp_label, ZPOOL_CONFIG_HOSTID)) {
hostid = fnvlist_lookup_uint64(mmp_label,
ZPOOL_CONFIG_HOSTID);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_HOSTID, hostid);
}
}
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_STATE, MMP_STATE_ACTIVE);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_TXG, 0);
error = spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO);
}
if (mmp_label)
nvlist_free(mmp_label);
return (error);
}
static int
spa_verify_host(spa_t *spa, nvlist_t *mos_config)
{
uint64_t hostid;
const char *hostname;
uint64_t myhostid = 0;
if (!spa_is_root(spa) && nvlist_lookup_uint64(mos_config,
ZPOOL_CONFIG_HOSTID, &hostid) == 0) {
hostname = fnvlist_lookup_string(mos_config,
ZPOOL_CONFIG_HOSTNAME);
myhostid = zone_get_hostid(NULL);
if (hostid != 0 && myhostid != 0 && hostid != myhostid) {
cmn_err(CE_WARN, "pool '%s' could not be "
"loaded as it was last accessed by "
"another system (host: %s hostid: 0x%llx). "
"See: https://openzfs.github.io/openzfs-docs/msg/"
"ZFS-8000-EY",
spa_name(spa), hostname, (u_longlong_t)hostid);
spa_load_failed(spa, "hostid verification failed: pool "
"last accessed by host: %s (hostid: 0x%llx)",
hostname, (u_longlong_t)hostid);
return (SET_ERROR(EBADF));
}
}
return (0);
}
static int
spa_ld_parse_config(spa_t *spa, spa_import_type_t type)
{
int error = 0;
nvlist_t *nvtree, *nvl, *config = spa->spa_config;
int parse;
vdev_t *rvd;
uint64_t pool_guid;
const char *comment;
const char *compatibility;
/*
* Versioning wasn't explicitly added to the label until later, so if
* it's not present treat it as the initial version.
*/
if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
&spa->spa_ubsync.ub_version) != 0)
spa->spa_ubsync.ub_version = SPA_VERSION_INITIAL;
if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &pool_guid)) {
spa_load_failed(spa, "invalid config provided: '%s' missing",
ZPOOL_CONFIG_POOL_GUID);
return (SET_ERROR(EINVAL));
}
/*
* If we are doing an import, ensure that the pool is not already
* imported by checking if its pool guid already exists in the
* spa namespace.
*
* The only case that we allow an already imported pool to be
* imported again, is when the pool is checkpointed and we want to
* look at its checkpointed state from userland tools like zdb.
*/
#ifdef _KERNEL
if ((spa->spa_load_state == SPA_LOAD_IMPORT ||
spa->spa_load_state == SPA_LOAD_TRYIMPORT) &&
spa_guid_exists(pool_guid, 0)) {
#else
if ((spa->spa_load_state == SPA_LOAD_IMPORT ||
spa->spa_load_state == SPA_LOAD_TRYIMPORT) &&
spa_guid_exists(pool_guid, 0) &&
!spa_importing_readonly_checkpoint(spa)) {
#endif
spa_load_failed(spa, "a pool with guid %llu is already open",
(u_longlong_t)pool_guid);
return (SET_ERROR(EEXIST));
}
spa->spa_config_guid = pool_guid;
nvlist_free(spa->spa_load_info);
spa->spa_load_info = fnvlist_alloc();
ASSERT(spa->spa_comment == NULL);
if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMMENT, &comment) == 0)
spa->spa_comment = spa_strdup(comment);
ASSERT(spa->spa_compatibility == NULL);
if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMPATIBILITY,
&compatibility) == 0)
spa->spa_compatibility = spa_strdup(compatibility);
(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_TXG,
&spa->spa_config_txg);
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_SPLIT, &nvl) == 0)
spa->spa_config_splitting = fnvlist_dup(nvl);
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvtree)) {
spa_load_failed(spa, "invalid config provided: '%s' missing",
ZPOOL_CONFIG_VDEV_TREE);
return (SET_ERROR(EINVAL));
}
/*
* Create "The Godfather" zio to hold all async IOs
*/
spa->spa_async_zio_root = kmem_alloc(max_ncpus * sizeof (void *),
KM_SLEEP);
for (int i = 0; i < max_ncpus; i++) {
spa->spa_async_zio_root[i] = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE |
ZIO_FLAG_GODFATHER);
}
/*
* Parse the configuration into a vdev tree. We explicitly set the
* value that will be returned by spa_version() since parsing the
* configuration requires knowing the version number.
*/
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
parse = (type == SPA_IMPORT_EXISTING ?
VDEV_ALLOC_LOAD : VDEV_ALLOC_SPLIT);
error = spa_config_parse(spa, &rvd, nvtree, NULL, 0, parse);
spa_config_exit(spa, SCL_ALL, FTAG);
if (error != 0) {
spa_load_failed(spa, "unable to parse config [error=%d]",
error);
return (error);
}
ASSERT(spa->spa_root_vdev == rvd);
ASSERT3U(spa->spa_min_ashift, >=, SPA_MINBLOCKSHIFT);
ASSERT3U(spa->spa_max_ashift, <=, SPA_MAXBLOCKSHIFT);
if (type != SPA_IMPORT_ASSEMBLE) {
ASSERT(spa_guid(spa) == pool_guid);
}
return (0);
}
/*
* Recursively open all vdevs in the vdev tree. This function is called twice:
* first with the untrusted config, then with the trusted config.
*/
static int
spa_ld_open_vdevs(spa_t *spa)
{
int error = 0;
/*
* spa_missing_tvds_allowed defines how many top-level vdevs can be
* missing/unopenable for the root vdev to be still considered openable.
*/
if (spa->spa_trust_config) {
spa->spa_missing_tvds_allowed = zfs_max_missing_tvds;
} else if (spa->spa_config_source == SPA_CONFIG_SRC_CACHEFILE) {
spa->spa_missing_tvds_allowed = zfs_max_missing_tvds_cachefile;
} else if (spa->spa_config_source == SPA_CONFIG_SRC_SCAN) {
spa->spa_missing_tvds_allowed = zfs_max_missing_tvds_scan;
} else {
spa->spa_missing_tvds_allowed = 0;
}
spa->spa_missing_tvds_allowed =
MAX(zfs_max_missing_tvds, spa->spa_missing_tvds_allowed);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
error = vdev_open(spa->spa_root_vdev);
spa_config_exit(spa, SCL_ALL, FTAG);
if (spa->spa_missing_tvds != 0) {
spa_load_note(spa, "vdev tree has %lld missing top-level "
"vdevs.", (u_longlong_t)spa->spa_missing_tvds);
if (spa->spa_trust_config && (spa->spa_mode & SPA_MODE_WRITE)) {
/*
* Although theoretically we could allow users to open
* incomplete pools in RW mode, we'd need to add a lot
* of extra logic (e.g. adjust pool space to account
* for missing vdevs).
* This limitation also prevents users from accidentally
* opening the pool in RW mode during data recovery and
* damaging it further.
*/
spa_load_note(spa, "pools with missing top-level "
"vdevs can only be opened in read-only mode.");
error = SET_ERROR(ENXIO);
} else {
spa_load_note(spa, "current settings allow for maximum "
"%lld missing top-level vdevs at this stage.",
(u_longlong_t)spa->spa_missing_tvds_allowed);
}
}
if (error != 0) {
spa_load_failed(spa, "unable to open vdev tree [error=%d]",
error);
}
if (spa->spa_missing_tvds != 0 || error != 0)
vdev_dbgmsg_print_tree(spa->spa_root_vdev, 2);
return (error);
}
/*
* We need to validate the vdev labels against the configuration that
* we have in hand. This function is called twice: first with an untrusted
* config, then with a trusted config. The validation is more strict when the
* config is trusted.
*/
static int
spa_ld_validate_vdevs(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
error = vdev_validate(rvd);
spa_config_exit(spa, SCL_ALL, FTAG);
if (error != 0) {
spa_load_failed(spa, "vdev_validate failed [error=%d]", error);
return (error);
}
if (rvd->vdev_state <= VDEV_STATE_CANT_OPEN) {
spa_load_failed(spa, "cannot open vdev tree after invalidating "
"some vdevs");
vdev_dbgmsg_print_tree(rvd, 2);
return (SET_ERROR(ENXIO));
}
return (0);
}
static void
spa_ld_select_uberblock_done(spa_t *spa, uberblock_t *ub)
{
spa->spa_state = POOL_STATE_ACTIVE;
spa->spa_ubsync = spa->spa_uberblock;
spa->spa_verify_min_txg = spa->spa_extreme_rewind ?
TXG_INITIAL - 1 : spa_last_synced_txg(spa) - TXG_DEFER_SIZE - 1;
spa->spa_first_txg = spa->spa_last_ubsync_txg ?
spa->spa_last_ubsync_txg : spa_last_synced_txg(spa) + 1;
spa->spa_claim_max_txg = spa->spa_first_txg;
spa->spa_prev_software_version = ub->ub_software_version;
}
static int
spa_ld_select_uberblock(spa_t *spa, spa_import_type_t type)
{
vdev_t *rvd = spa->spa_root_vdev;
nvlist_t *label;
uberblock_t *ub = &spa->spa_uberblock;
boolean_t activity_check = B_FALSE;
/*
* If we are opening the checkpointed state of the pool by
* rewinding to it, at this point we will have written the
* checkpointed uberblock to the vdev labels, so searching
* the labels will find the right uberblock. However, if
* we are opening the checkpointed state read-only, we have
* not modified the labels. Therefore, we must ignore the
* labels and continue using the spa_uberblock that was set
* by spa_ld_checkpoint_rewind.
*
* Note that it would be fine to ignore the labels when
* rewinding (opening writeable) as well. However, if we
* crash just after writing the labels, we will end up
* searching the labels. Doing so in the common case means
* that this code path gets exercised normally, rather than
* just in the edge case.
*/
if (ub->ub_checkpoint_txg != 0 &&
spa_importing_readonly_checkpoint(spa)) {
spa_ld_select_uberblock_done(spa, ub);
return (0);
}
/*
* Find the best uberblock.
*/
vdev_uberblock_load(rvd, ub, &label);
/*
* If we weren't able to find a single valid uberblock, return failure.
*/
if (ub->ub_txg == 0) {
nvlist_free(label);
spa_load_failed(spa, "no valid uberblock found");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, ENXIO));
}
if (spa->spa_load_max_txg != UINT64_MAX) {
(void) spa_import_progress_set_max_txg(spa_guid(spa),
(u_longlong_t)spa->spa_load_max_txg);
}
spa_load_note(spa, "using uberblock with txg=%llu",
(u_longlong_t)ub->ub_txg);
+ if (ub->ub_raidz_reflow_info != 0) {
+ spa_load_note(spa, "uberblock raidz_reflow_info: "
+ "state=%u offset=%llu",
+ (int)RRSS_GET_STATE(ub),
+ (u_longlong_t)RRSS_GET_OFFSET(ub));
+ }
/*
* For pools which have the multihost property on determine if the
* pool is truly inactive and can be safely imported. Prevent
* hosts which don't have a hostid set from importing the pool.
*/
activity_check = spa_activity_check_required(spa, ub, label,
spa->spa_config);
if (activity_check) {
if (ub->ub_mmp_magic == MMP_MAGIC && ub->ub_mmp_delay &&
spa_get_hostid(spa) == 0) {
nvlist_free(label);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_STATE, MMP_STATE_NO_HOSTID);
return (spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO));
}
int error = spa_activity_check(spa, ub, spa->spa_config);
if (error) {
nvlist_free(label);
return (error);
}
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_STATE, MMP_STATE_INACTIVE);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_TXG, ub->ub_txg);
fnvlist_add_uint16(spa->spa_load_info,
ZPOOL_CONFIG_MMP_SEQ,
(MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0));
}
/*
* If the pool has an unsupported version we can't open it.
*/
if (!SPA_VERSION_IS_SUPPORTED(ub->ub_version)) {
nvlist_free(label);
spa_load_failed(spa, "version %llu is not supported",
(u_longlong_t)ub->ub_version);
return (spa_vdev_err(rvd, VDEV_AUX_VERSION_NEWER, ENOTSUP));
}
if (ub->ub_version >= SPA_VERSION_FEATURES) {
nvlist_t *features;
/*
* If we weren't able to find what's necessary for reading the
* MOS in the label, return failure.
*/
if (label == NULL) {
spa_load_failed(spa, "label config unavailable");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA,
ENXIO));
}
if (nvlist_lookup_nvlist(label, ZPOOL_CONFIG_FEATURES_FOR_READ,
&features) != 0) {
nvlist_free(label);
spa_load_failed(spa, "invalid label: '%s' missing",
ZPOOL_CONFIG_FEATURES_FOR_READ);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA,
ENXIO));
}
/*
* Update our in-core representation with the definitive values
* from the label.
*/
nvlist_free(spa->spa_label_features);
spa->spa_label_features = fnvlist_dup(features);
}
nvlist_free(label);
/*
* Look through entries in the label nvlist's features_for_read. If
* there is a feature listed there which we don't understand then we
* cannot open a pool.
*/
if (ub->ub_version >= SPA_VERSION_FEATURES) {
nvlist_t *unsup_feat;
unsup_feat = fnvlist_alloc();
for (nvpair_t *nvp = nvlist_next_nvpair(spa->spa_label_features,
NULL); nvp != NULL;
nvp = nvlist_next_nvpair(spa->spa_label_features, nvp)) {
if (!zfeature_is_supported(nvpair_name(nvp))) {
fnvlist_add_string(unsup_feat,
nvpair_name(nvp), "");
}
}
if (!nvlist_empty(unsup_feat)) {
fnvlist_add_nvlist(spa->spa_load_info,
ZPOOL_CONFIG_UNSUP_FEAT, unsup_feat);
nvlist_free(unsup_feat);
spa_load_failed(spa, "some features are unsupported");
return (spa_vdev_err(rvd, VDEV_AUX_UNSUP_FEAT,
ENOTSUP));
}
nvlist_free(unsup_feat);
}
if (type != SPA_IMPORT_ASSEMBLE && spa->spa_config_splitting) {
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_try_repair(spa, spa->spa_config);
spa_config_exit(spa, SCL_ALL, FTAG);
nvlist_free(spa->spa_config_splitting);
spa->spa_config_splitting = NULL;
}
/*
* Initialize internal SPA structures.
*/
spa_ld_select_uberblock_done(spa, ub);
return (0);
}
static int
spa_ld_open_rootbp(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
error = dsl_pool_init(spa, spa->spa_first_txg, &spa->spa_dsl_pool);
if (error != 0) {
spa_load_failed(spa, "unable to open rootbp in dsl_pool_init "
"[error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
spa->spa_meta_objset = spa->spa_dsl_pool->dp_meta_objset;
return (0);
}
static int
spa_ld_trusted_config(spa_t *spa, spa_import_type_t type,
boolean_t reloading)
{
vdev_t *mrvd, *rvd = spa->spa_root_vdev;
nvlist_t *nv, *mos_config, *policy;
int error = 0, copy_error;
uint64_t healthy_tvds, healthy_tvds_mos;
uint64_t mos_config_txg;
if (spa_dir_prop(spa, DMU_POOL_CONFIG, &spa->spa_config_object, B_TRUE)
!= 0)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
/*
* If we're assembling a pool from a split, the config provided is
* already trusted so there is nothing to do.
*/
if (type == SPA_IMPORT_ASSEMBLE)
return (0);
healthy_tvds = spa_healthy_core_tvds(spa);
if (load_nvlist(spa, spa->spa_config_object, &mos_config)
!= 0) {
spa_load_failed(spa, "unable to retrieve MOS config");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
/*
* If we are doing an open, pool owner wasn't verified yet, thus do
* the verification here.
*/
if (spa->spa_load_state == SPA_LOAD_OPEN) {
error = spa_verify_host(spa, mos_config);
if (error != 0) {
nvlist_free(mos_config);
return (error);
}
}
nv = fnvlist_lookup_nvlist(mos_config, ZPOOL_CONFIG_VDEV_TREE);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
/*
* Build a new vdev tree from the trusted config
*/
error = spa_config_parse(spa, &mrvd, nv, NULL, 0, VDEV_ALLOC_LOAD);
if (error != 0) {
nvlist_free(mos_config);
spa_config_exit(spa, SCL_ALL, FTAG);
spa_load_failed(spa, "spa_config_parse failed [error=%d]",
error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error));
}
/*
* Vdev paths in the MOS may be obsolete. If the untrusted config was
* obtained by scanning /dev/dsk, then it will have the right vdev
* paths. We update the trusted MOS config with this information.
* We first try to copy the paths with vdev_copy_path_strict, which
* succeeds only when both configs have exactly the same vdev tree.
* If that fails, we fall back to a more flexible method that has a
* best effort policy.
*/
copy_error = vdev_copy_path_strict(rvd, mrvd);
if (copy_error != 0 || spa_load_print_vdev_tree) {
spa_load_note(spa, "provided vdev tree:");
vdev_dbgmsg_print_tree(rvd, 2);
spa_load_note(spa, "MOS vdev tree:");
vdev_dbgmsg_print_tree(mrvd, 2);
}
if (copy_error != 0) {
spa_load_note(spa, "vdev_copy_path_strict failed, falling "
"back to vdev_copy_path_relaxed");
vdev_copy_path_relaxed(rvd, mrvd);
}
vdev_close(rvd);
vdev_free(rvd);
spa->spa_root_vdev = mrvd;
rvd = mrvd;
spa_config_exit(spa, SCL_ALL, FTAG);
/*
* If 'zpool import' used a cached config, then the on-disk hostid and
* hostname may be different to the cached config in ways that should
* prevent import. Userspace can't discover this without a scan, but
* we know, so we add these values to LOAD_INFO so the caller can know
* the difference.
*
* Note that we have to do this before the config is regenerated,
* because the new config will have the hostid and hostname for this
* host, in readiness for import.
*/
if (nvlist_exists(mos_config, ZPOOL_CONFIG_HOSTID))
fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_HOSTID,
fnvlist_lookup_uint64(mos_config, ZPOOL_CONFIG_HOSTID));
if (nvlist_exists(mos_config, ZPOOL_CONFIG_HOSTNAME))
fnvlist_add_string(spa->spa_load_info, ZPOOL_CONFIG_HOSTNAME,
fnvlist_lookup_string(mos_config, ZPOOL_CONFIG_HOSTNAME));
/*
* We will use spa_config if we decide to reload the spa or if spa_load
* fails and we rewind. We must thus regenerate the config using the
* MOS information with the updated paths. ZPOOL_LOAD_POLICY is used to
* pass settings on how to load the pool and is not stored in the MOS.
* We copy it over to our new, trusted config.
*/
mos_config_txg = fnvlist_lookup_uint64(mos_config,
ZPOOL_CONFIG_POOL_TXG);
nvlist_free(mos_config);
mos_config = spa_config_generate(spa, NULL, mos_config_txg, B_FALSE);
if (nvlist_lookup_nvlist(spa->spa_config, ZPOOL_LOAD_POLICY,
&policy) == 0)
fnvlist_add_nvlist(mos_config, ZPOOL_LOAD_POLICY, policy);
spa_config_set(spa, mos_config);
spa->spa_config_source = SPA_CONFIG_SRC_MOS;
/*
* Now that we got the config from the MOS, we should be more strict
* in checking blkptrs and can make assumptions about the consistency
* of the vdev tree. spa_trust_config must be set to true before opening
* vdevs in order for them to be writeable.
*/
spa->spa_trust_config = B_TRUE;
/*
* Open and validate the new vdev tree
*/
error = spa_ld_open_vdevs(spa);
if (error != 0)
return (error);
error = spa_ld_validate_vdevs(spa);
if (error != 0)
return (error);
if (copy_error != 0 || spa_load_print_vdev_tree) {
spa_load_note(spa, "final vdev tree:");
vdev_dbgmsg_print_tree(rvd, 2);
}
if (spa->spa_load_state != SPA_LOAD_TRYIMPORT &&
!spa->spa_extreme_rewind && zfs_max_missing_tvds == 0) {
/*
* Sanity check to make sure that we are indeed loading the
* latest uberblock. If we missed SPA_SYNC_MIN_VDEVS tvds
* in the config provided and they happened to be the only ones
* to have the latest uberblock, we could involuntarily perform
* an extreme rewind.
*/
healthy_tvds_mos = spa_healthy_core_tvds(spa);
if (healthy_tvds_mos - healthy_tvds >=
SPA_SYNC_MIN_VDEVS) {
spa_load_note(spa, "config provided misses too many "
"top-level vdevs compared to MOS (%lld vs %lld). ",
(u_longlong_t)healthy_tvds,
(u_longlong_t)healthy_tvds_mos);
spa_load_note(spa, "vdev tree:");
vdev_dbgmsg_print_tree(rvd, 2);
if (reloading) {
spa_load_failed(spa, "config was already "
"provided from MOS. Aborting.");
return (spa_vdev_err(rvd,
VDEV_AUX_CORRUPT_DATA, EIO));
}
spa_load_note(spa, "spa must be reloaded using MOS "
"config");
return (SET_ERROR(EAGAIN));
}
}
error = spa_check_for_missing_logs(spa);
if (error != 0)
return (spa_vdev_err(rvd, VDEV_AUX_BAD_GUID_SUM, ENXIO));
if (rvd->vdev_guid_sum != spa->spa_uberblock.ub_guid_sum) {
spa_load_failed(spa, "uberblock guid sum doesn't match MOS "
"guid sum (%llu != %llu)",
(u_longlong_t)spa->spa_uberblock.ub_guid_sum,
(u_longlong_t)rvd->vdev_guid_sum);
return (spa_vdev_err(rvd, VDEV_AUX_BAD_GUID_SUM,
ENXIO));
}
return (0);
}
static int
spa_ld_open_indirect_vdev_metadata(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
/*
* Everything that we read before spa_remove_init() must be stored
* on concreted vdevs. Therefore we do this as early as possible.
*/
error = spa_remove_init(spa);
if (error != 0) {
spa_load_failed(spa, "spa_remove_init failed [error=%d]",
error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
/*
* Retrieve information needed to condense indirect vdev mappings.
*/
error = spa_condense_init(spa);
if (error != 0) {
spa_load_failed(spa, "spa_condense_init failed [error=%d]",
error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error));
}
return (0);
}
static int
spa_ld_check_features(spa_t *spa, boolean_t *missing_feat_writep)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
if (spa_version(spa) >= SPA_VERSION_FEATURES) {
boolean_t missing_feat_read = B_FALSE;
nvlist_t *unsup_feat, *enabled_feat;
if (spa_dir_prop(spa, DMU_POOL_FEATURES_FOR_READ,
&spa->spa_feat_for_read_obj, B_TRUE) != 0) {
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
if (spa_dir_prop(spa, DMU_POOL_FEATURES_FOR_WRITE,
&spa->spa_feat_for_write_obj, B_TRUE) != 0) {
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
if (spa_dir_prop(spa, DMU_POOL_FEATURE_DESCRIPTIONS,
&spa->spa_feat_desc_obj, B_TRUE) != 0) {
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
enabled_feat = fnvlist_alloc();
unsup_feat = fnvlist_alloc();
if (!spa_features_check(spa, B_FALSE,
unsup_feat, enabled_feat))
missing_feat_read = B_TRUE;
if (spa_writeable(spa) ||
spa->spa_load_state == SPA_LOAD_TRYIMPORT) {
if (!spa_features_check(spa, B_TRUE,
unsup_feat, enabled_feat)) {
*missing_feat_writep = B_TRUE;
}
}
fnvlist_add_nvlist(spa->spa_load_info,
ZPOOL_CONFIG_ENABLED_FEAT, enabled_feat);
if (!nvlist_empty(unsup_feat)) {
fnvlist_add_nvlist(spa->spa_load_info,
ZPOOL_CONFIG_UNSUP_FEAT, unsup_feat);
}
fnvlist_free(enabled_feat);
fnvlist_free(unsup_feat);
if (!missing_feat_read) {
fnvlist_add_boolean(spa->spa_load_info,
ZPOOL_CONFIG_CAN_RDONLY);
}
/*
* If the state is SPA_LOAD_TRYIMPORT, our objective is
* twofold: to determine whether the pool is available for
* import in read-write mode and (if it is not) whether the
* pool is available for import in read-only mode. If the pool
* is available for import in read-write mode, it is displayed
* as available in userland; if it is not available for import
* in read-only mode, it is displayed as unavailable in
* userland. If the pool is available for import in read-only
* mode but not read-write mode, it is displayed as unavailable
* in userland with a special note that the pool is actually
* available for open in read-only mode.
*
* As a result, if the state is SPA_LOAD_TRYIMPORT and we are
* missing a feature for write, we must first determine whether
* the pool can be opened read-only before returning to
* userland in order to know whether to display the
* abovementioned note.
*/
if (missing_feat_read || (*missing_feat_writep &&
spa_writeable(spa))) {
spa_load_failed(spa, "pool uses unsupported features");
return (spa_vdev_err(rvd, VDEV_AUX_UNSUP_FEAT,
ENOTSUP));
}
/*
* Load refcounts for ZFS features from disk into an in-memory
* cache during SPA initialization.
*/
for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
uint64_t refcount;
error = feature_get_refcount_from_disk(spa,
&spa_feature_table[i], &refcount);
if (error == 0) {
spa->spa_feat_refcount_cache[i] = refcount;
} else if (error == ENOTSUP) {
spa->spa_feat_refcount_cache[i] =
SPA_FEATURE_DISABLED;
} else {
spa_load_failed(spa, "error getting refcount "
"for feature %s [error=%d]",
spa_feature_table[i].fi_guid, error);
return (spa_vdev_err(rvd,
VDEV_AUX_CORRUPT_DATA, EIO));
}
}
}
if (spa_feature_is_active(spa, SPA_FEATURE_ENABLED_TXG)) {
if (spa_dir_prop(spa, DMU_POOL_FEATURE_ENABLED_TXG,
&spa->spa_feat_enabled_txg_obj, B_TRUE) != 0)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
/*
* Encryption was added before bookmark_v2, even though bookmark_v2
* is now a dependency. If this pool has encryption enabled without
* bookmark_v2, trigger an errata message.
*/
if (spa_feature_is_enabled(spa, SPA_FEATURE_ENCRYPTION) &&
!spa_feature_is_enabled(spa, SPA_FEATURE_BOOKMARK_V2)) {
spa->spa_errata = ZPOOL_ERRATA_ZOL_8308_ENCRYPTION;
}
return (0);
}
static int
spa_ld_load_special_directories(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
spa->spa_is_initializing = B_TRUE;
error = dsl_pool_open(spa->spa_dsl_pool);
spa->spa_is_initializing = B_FALSE;
if (error != 0) {
spa_load_failed(spa, "dsl_pool_open failed [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
return (0);
}
static int
spa_ld_get_props(spa_t *spa)
{
int error = 0;
uint64_t obj;
vdev_t *rvd = spa->spa_root_vdev;
/* Grab the checksum salt from the MOS. */
error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_CHECKSUM_SALT, 1,
sizeof (spa->spa_cksum_salt.zcs_bytes),
spa->spa_cksum_salt.zcs_bytes);
if (error == ENOENT) {
/* Generate a new salt for subsequent use */
(void) random_get_pseudo_bytes(spa->spa_cksum_salt.zcs_bytes,
sizeof (spa->spa_cksum_salt.zcs_bytes));
} else if (error != 0) {
spa_load_failed(spa, "unable to retrieve checksum salt from "
"MOS [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
if (spa_dir_prop(spa, DMU_POOL_SYNC_BPOBJ, &obj, B_TRUE) != 0)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
error = bpobj_open(&spa->spa_deferred_bpobj, spa->spa_meta_objset, obj);
if (error != 0) {
spa_load_failed(spa, "error opening deferred-frees bpobj "
"[error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
/*
* Load the bit that tells us to use the new accounting function
* (raid-z deflation). If we have an older pool, this will not
* be present.
*/
error = spa_dir_prop(spa, DMU_POOL_DEFLATE, &spa->spa_deflate, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
error = spa_dir_prop(spa, DMU_POOL_CREATION_VERSION,
&spa->spa_creation_version, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
/*
* Load the persistent error log. If we have an older pool, this will
* not be present.
*/
error = spa_dir_prop(spa, DMU_POOL_ERRLOG_LAST, &spa->spa_errlog_last,
B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
error = spa_dir_prop(spa, DMU_POOL_ERRLOG_SCRUB,
&spa->spa_errlog_scrub, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
/*
* Load the livelist deletion field. If a livelist is queued for
* deletion, indicate that in the spa
*/
error = spa_dir_prop(spa, DMU_POOL_DELETED_CLONES,
&spa->spa_livelists_to_delete, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
/*
* Load the history object. If we have an older pool, this
* will not be present.
*/
error = spa_dir_prop(spa, DMU_POOL_HISTORY, &spa->spa_history, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
/*
* Load the per-vdev ZAP map. If we have an older pool, this will not
* be present; in this case, defer its creation to a later time to
* avoid dirtying the MOS this early / out of sync context. See
* spa_sync_config_object.
*/
/* The sentinel is only available in the MOS config. */
nvlist_t *mos_config;
if (load_nvlist(spa, spa->spa_config_object, &mos_config) != 0) {
spa_load_failed(spa, "unable to retrieve MOS config");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
error = spa_dir_prop(spa, DMU_POOL_VDEV_ZAP_MAP,
&spa->spa_all_vdev_zaps, B_FALSE);
if (error == ENOENT) {
VERIFY(!nvlist_exists(mos_config,
ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS));
spa->spa_avz_action = AVZ_ACTION_INITIALIZE;
ASSERT0(vdev_count_verify_zaps(spa->spa_root_vdev));
} else if (error != 0) {
nvlist_free(mos_config);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
} else if (!nvlist_exists(mos_config, ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS)) {
/*
* An older version of ZFS overwrote the sentinel value, so
* we have orphaned per-vdev ZAPs in the MOS. Defer their
* destruction to later; see spa_sync_config_object.
*/
spa->spa_avz_action = AVZ_ACTION_DESTROY;
/*
* We're assuming that no vdevs have had their ZAPs created
* before this. Better be sure of it.
*/
ASSERT0(vdev_count_verify_zaps(spa->spa_root_vdev));
}
nvlist_free(mos_config);
spa->spa_delegation = zpool_prop_default_numeric(ZPOOL_PROP_DELEGATION);
error = spa_dir_prop(spa, DMU_POOL_PROPS, &spa->spa_pool_props_object,
B_FALSE);
if (error && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
if (error == 0) {
uint64_t autoreplace = 0;
spa_prop_find(spa, ZPOOL_PROP_BOOTFS, &spa->spa_bootfs);
spa_prop_find(spa, ZPOOL_PROP_AUTOREPLACE, &autoreplace);
spa_prop_find(spa, ZPOOL_PROP_DELEGATION, &spa->spa_delegation);
spa_prop_find(spa, ZPOOL_PROP_FAILUREMODE, &spa->spa_failmode);
spa_prop_find(spa, ZPOOL_PROP_AUTOEXPAND, &spa->spa_autoexpand);
spa_prop_find(spa, ZPOOL_PROP_MULTIHOST, &spa->spa_multihost);
spa_prop_find(spa, ZPOOL_PROP_AUTOTRIM, &spa->spa_autotrim);
spa->spa_autoreplace = (autoreplace != 0);
}
/*
* If we are importing a pool with missing top-level vdevs,
* we enforce that the pool doesn't panic or get suspended on
* error since the likelihood of missing data is extremely high.
*/
if (spa->spa_missing_tvds > 0 &&
spa->spa_failmode != ZIO_FAILURE_MODE_CONTINUE &&
spa->spa_load_state != SPA_LOAD_TRYIMPORT) {
spa_load_note(spa, "forcing failmode to 'continue' "
"as some top level vdevs are missing");
spa->spa_failmode = ZIO_FAILURE_MODE_CONTINUE;
}
return (0);
}
static int
spa_ld_open_aux_vdevs(spa_t *spa, spa_import_type_t type)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
/*
* If we're assembling the pool from the split-off vdevs of
* an existing pool, we don't want to attach the spares & cache
* devices.
*/
/*
* Load any hot spares for this pool.
*/
error = spa_dir_prop(spa, DMU_POOL_SPARES, &spa->spa_spares.sav_object,
B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
if (error == 0 && type != SPA_IMPORT_ASSEMBLE) {
ASSERT(spa_version(spa) >= SPA_VERSION_SPARES);
if (load_nvlist(spa, spa->spa_spares.sav_object,
&spa->spa_spares.sav_config) != 0) {
spa_load_failed(spa, "error loading spares nvlist");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_spares(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
} else if (error == 0) {
spa->spa_spares.sav_sync = B_TRUE;
}
/*
* Load any level 2 ARC devices for this pool.
*/
error = spa_dir_prop(spa, DMU_POOL_L2CACHE,
&spa->spa_l2cache.sav_object, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
if (error == 0 && type != SPA_IMPORT_ASSEMBLE) {
ASSERT(spa_version(spa) >= SPA_VERSION_L2CACHE);
if (load_nvlist(spa, spa->spa_l2cache.sav_object,
&spa->spa_l2cache.sav_config) != 0) {
spa_load_failed(spa, "error loading l2cache nvlist");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_l2cache(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
} else if (error == 0) {
spa->spa_l2cache.sav_sync = B_TRUE;
}
return (0);
}
static int
spa_ld_load_vdev_metadata(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
/*
* If the 'multihost' property is set, then never allow a pool to
* be imported when the system hostid is zero. The exception to
* this rule is zdb which is always allowed to access pools.
*/
if (spa_multihost(spa) && spa_get_hostid(spa) == 0 &&
(spa->spa_import_flags & ZFS_IMPORT_SKIP_MMP) == 0) {
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_STATE, MMP_STATE_NO_HOSTID);
return (spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO));
}
/*
* If the 'autoreplace' property is set, then post a resource notifying
* the ZFS DE that it should not issue any faults for unopenable
* devices. We also iterate over the vdevs, and post a sysevent for any
* unopenable vdevs so that the normal autoreplace handler can take
* over.
*/
if (spa->spa_autoreplace && spa->spa_load_state != SPA_LOAD_TRYIMPORT) {
spa_check_removed(spa->spa_root_vdev);
/*
* For the import case, this is done in spa_import(), because
* at this point we're using the spare definitions from
* the MOS config, not necessarily from the userland config.
*/
if (spa->spa_load_state != SPA_LOAD_IMPORT) {
spa_aux_check_removed(&spa->spa_spares);
spa_aux_check_removed(&spa->spa_l2cache);
}
}
/*
* Load the vdev metadata such as metaslabs, DTLs, spacemap object, etc.
*/
error = vdev_load(rvd);
if (error != 0) {
spa_load_failed(spa, "vdev_load failed [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error));
}
error = spa_ld_log_spacemaps(spa);
if (error != 0) {
spa_load_failed(spa, "spa_ld_log_spacemaps failed [error=%d]",
error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error));
}
/*
* Propagate the leaf DTLs we just loaded all the way up the vdev tree.
*/
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
vdev_dtl_reassess(rvd, 0, 0, B_FALSE, B_FALSE);
spa_config_exit(spa, SCL_ALL, FTAG);
return (0);
}
static int
spa_ld_load_dedup_tables(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
error = ddt_load(spa);
if (error != 0) {
spa_load_failed(spa, "ddt_load failed [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
return (0);
}
static int
spa_ld_load_brt(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
error = brt_load(spa);
if (error != 0) {
spa_load_failed(spa, "brt_load failed [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
return (0);
}
static int
spa_ld_verify_logs(spa_t *spa, spa_import_type_t type, const char **ereport)
{
vdev_t *rvd = spa->spa_root_vdev;
if (type != SPA_IMPORT_ASSEMBLE && spa_writeable(spa)) {
boolean_t missing = spa_check_logs(spa);
if (missing) {
if (spa->spa_missing_tvds != 0) {
spa_load_note(spa, "spa_check_logs failed "
"so dropping the logs");
} else {
*ereport = FM_EREPORT_ZFS_LOG_REPLAY;
spa_load_failed(spa, "spa_check_logs failed");
return (spa_vdev_err(rvd, VDEV_AUX_BAD_LOG,
ENXIO));
}
}
}
return (0);
}
static int
spa_ld_verify_pool_data(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
/*
* We've successfully opened the pool, verify that we're ready
* to start pushing transactions.
*/
if (spa->spa_load_state != SPA_LOAD_TRYIMPORT) {
error = spa_load_verify(spa);
if (error != 0) {
spa_load_failed(spa, "spa_load_verify failed "
"[error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA,
error));
}
}
return (0);
}
static void
spa_ld_claim_log_blocks(spa_t *spa)
{
dmu_tx_t *tx;
dsl_pool_t *dp = spa_get_dsl(spa);
/*
* Claim log blocks that haven't been committed yet.
* This must all happen in a single txg.
* Note: spa_claim_max_txg is updated by spa_claim_notify(),
* invoked from zil_claim_log_block()'s i/o done callback.
* Price of rollback is that we abandon the log.
*/
spa->spa_claiming = B_TRUE;
tx = dmu_tx_create_assigned(dp, spa_first_txg(spa));
(void) dmu_objset_find_dp(dp, dp->dp_root_dir_obj,
zil_claim, tx, DS_FIND_CHILDREN);
dmu_tx_commit(tx);
spa->spa_claiming = B_FALSE;
spa_set_log_state(spa, SPA_LOG_GOOD);
}
static void
spa_ld_check_for_config_update(spa_t *spa, uint64_t config_cache_txg,
boolean_t update_config_cache)
{
vdev_t *rvd = spa->spa_root_vdev;
int need_update = B_FALSE;
/*
* If the config cache is stale, or we have uninitialized
* metaslabs (see spa_vdev_add()), then update the config.
*
* If this is a verbatim import, trust the current
* in-core spa_config and update the disk labels.
*/
if (update_config_cache || config_cache_txg != spa->spa_config_txg ||
spa->spa_load_state == SPA_LOAD_IMPORT ||
spa->spa_load_state == SPA_LOAD_RECOVER ||
(spa->spa_import_flags & ZFS_IMPORT_VERBATIM))
need_update = B_TRUE;
for (int c = 0; c < rvd->vdev_children; c++)
if (rvd->vdev_child[c]->vdev_ms_array == 0)
need_update = B_TRUE;
/*
* Update the config cache asynchronously in case we're the
* root pool, in which case the config cache isn't writable yet.
*/
if (need_update)
spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
}
static void
spa_ld_prepare_for_reload(spa_t *spa)
{
spa_mode_t mode = spa->spa_mode;
int async_suspended = spa->spa_async_suspended;
spa_unload(spa);
spa_deactivate(spa);
spa_activate(spa, mode);
/*
* We save the value of spa_async_suspended as it gets reset to 0 by
* spa_unload(). We want to restore it back to the original value before
* returning as we might be calling spa_async_resume() later.
*/
spa->spa_async_suspended = async_suspended;
}
static int
spa_ld_read_checkpoint_txg(spa_t *spa)
{
uberblock_t checkpoint;
int error = 0;
ASSERT0(spa->spa_checkpoint_txg);
ASSERT(MUTEX_HELD(&spa_namespace_lock));
error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ZPOOL_CHECKPOINT, sizeof (uint64_t),
sizeof (uberblock_t) / sizeof (uint64_t), &checkpoint);
if (error == ENOENT)
return (0);
if (error != 0)
return (error);
ASSERT3U(checkpoint.ub_txg, !=, 0);
ASSERT3U(checkpoint.ub_checkpoint_txg, !=, 0);
ASSERT3U(checkpoint.ub_timestamp, !=, 0);
spa->spa_checkpoint_txg = checkpoint.ub_txg;
spa->spa_checkpoint_info.sci_timestamp = checkpoint.ub_timestamp;
return (0);
}
static int
spa_ld_mos_init(spa_t *spa, spa_import_type_t type)
{
int error = 0;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(spa->spa_config_source != SPA_CONFIG_SRC_NONE);
/*
* Never trust the config that is provided unless we are assembling
* a pool following a split.
* This means don't trust blkptrs and the vdev tree in general. This
* also effectively puts the spa in read-only mode since
* spa_writeable() checks for spa_trust_config to be true.
* We will later load a trusted config from the MOS.
*/
if (type != SPA_IMPORT_ASSEMBLE)
spa->spa_trust_config = B_FALSE;
/*
* Parse the config provided to create a vdev tree.
*/
error = spa_ld_parse_config(spa, type);
if (error != 0)
return (error);
spa_import_progress_add(spa);
/*
* Now that we have the vdev tree, try to open each vdev. This involves
* opening the underlying physical device, retrieving its geometry and
* probing the vdev with a dummy I/O. The state of each vdev will be set
* based on the success of those operations. After this we'll be ready
* to read from the vdevs.
*/
error = spa_ld_open_vdevs(spa);
if (error != 0)
return (error);
/*
* Read the label of each vdev and make sure that the GUIDs stored
* there match the GUIDs in the config provided.
* If we're assembling a new pool that's been split off from an
* existing pool, the labels haven't yet been updated so we skip
* validation for now.
*/
if (type != SPA_IMPORT_ASSEMBLE) {
error = spa_ld_validate_vdevs(spa);
if (error != 0)
return (error);
}
/*
* Read all vdev labels to find the best uberblock (i.e. latest,
* unless spa_load_max_txg is set) and store it in spa_uberblock. We
* get the list of features required to read blkptrs in the MOS from
* the vdev label with the best uberblock and verify that our version
* of zfs supports them all.
*/
error = spa_ld_select_uberblock(spa, type);
if (error != 0)
return (error);
/*
* Pass that uberblock to the dsl_pool layer which will open the root
* blkptr. This blkptr points to the latest version of the MOS and will
* allow us to read its contents.
*/
error = spa_ld_open_rootbp(spa);
if (error != 0)
return (error);
return (0);
}
static int
spa_ld_checkpoint_rewind(spa_t *spa)
{
uberblock_t checkpoint;
int error = 0;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT);
error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ZPOOL_CHECKPOINT, sizeof (uint64_t),
sizeof (uberblock_t) / sizeof (uint64_t), &checkpoint);
if (error != 0) {
spa_load_failed(spa, "unable to retrieve checkpointed "
"uberblock from the MOS config [error=%d]", error);
if (error == ENOENT)
error = ZFS_ERR_NO_CHECKPOINT;
return (error);
}
ASSERT3U(checkpoint.ub_txg, <, spa->spa_uberblock.ub_txg);
ASSERT3U(checkpoint.ub_txg, ==, checkpoint.ub_checkpoint_txg);
/*
* We need to update the txg and timestamp of the checkpointed
* uberblock to be higher than the latest one. This ensures that
* the checkpointed uberblock is selected if we were to close and
* reopen the pool right after we've written it in the vdev labels.
* (also see block comment in vdev_uberblock_compare)
*/
checkpoint.ub_txg = spa->spa_uberblock.ub_txg + 1;
checkpoint.ub_timestamp = gethrestime_sec();
/*
* Set current uberblock to be the checkpointed uberblock.
*/
spa->spa_uberblock = checkpoint;
/*
* If we are doing a normal rewind, then the pool is open for
* writing and we sync the "updated" checkpointed uberblock to
* disk. Once this is done, we've basically rewound the whole
* pool and there is no way back.
*
* There are cases when we don't want to attempt and sync the
* checkpointed uberblock to disk because we are opening a
* pool as read-only. Specifically, verifying the checkpointed
* state with zdb, and importing the checkpointed state to get
* a "preview" of its content.
*/
if (spa_writeable(spa)) {
vdev_t *rvd = spa->spa_root_vdev;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
vdev_t *svd[SPA_SYNC_MIN_VDEVS] = { NULL };
int svdcount = 0;
int children = rvd->vdev_children;
int c0 = random_in_range(children);
for (int c = 0; c < children; c++) {
vdev_t *vd = rvd->vdev_child[(c0 + c) % children];
/* Stop when revisiting the first vdev */
if (c > 0 && svd[0] == vd)
break;
if (vd->vdev_ms_array == 0 || vd->vdev_islog ||
!vdev_is_concrete(vd))
continue;
svd[svdcount++] = vd;
if (svdcount == SPA_SYNC_MIN_VDEVS)
break;
}
error = vdev_config_sync(svd, svdcount, spa->spa_first_txg);
if (error == 0)
spa->spa_last_synced_guid = rvd->vdev_guid;
spa_config_exit(spa, SCL_ALL, FTAG);
if (error != 0) {
spa_load_failed(spa, "failed to write checkpointed "
"uberblock to the vdev labels [error=%d]", error);
return (error);
}
}
return (0);
}
static int
spa_ld_mos_with_trusted_config(spa_t *spa, spa_import_type_t type,
boolean_t *update_config_cache)
{
int error;
/*
* Parse the config for pool, open and validate vdevs,
* select an uberblock, and use that uberblock to open
* the MOS.
*/
error = spa_ld_mos_init(spa, type);
if (error != 0)
return (error);
/*
* Retrieve the trusted config stored in the MOS and use it to create
* a new, exact version of the vdev tree, then reopen all vdevs.
*/
error = spa_ld_trusted_config(spa, type, B_FALSE);
if (error == EAGAIN) {
if (update_config_cache != NULL)
*update_config_cache = B_TRUE;
/*
* Redo the loading process with the trusted config if it is
* too different from the untrusted config.
*/
spa_ld_prepare_for_reload(spa);
spa_load_note(spa, "RELOADING");
error = spa_ld_mos_init(spa, type);
if (error != 0)
return (error);
error = spa_ld_trusted_config(spa, type, B_TRUE);
if (error != 0)
return (error);
} else if (error != 0) {
return (error);
}
return (0);
}
/*
* Load an existing storage pool, using the config provided. This config
* describes which vdevs are part of the pool and is later validated against
* partial configs present in each vdev's label and an entire copy of the
* config stored in the MOS.
*/
static int
spa_load_impl(spa_t *spa, spa_import_type_t type, const char **ereport)
{
int error = 0;
boolean_t missing_feat_write = B_FALSE;
boolean_t checkpoint_rewind =
(spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT);
boolean_t update_config_cache = B_FALSE;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(spa->spa_config_source != SPA_CONFIG_SRC_NONE);
spa_load_note(spa, "LOADING");
error = spa_ld_mos_with_trusted_config(spa, type, &update_config_cache);
if (error != 0)
return (error);
/*
* If we are rewinding to the checkpoint then we need to repeat
* everything we've done so far in this function but this time
* selecting the checkpointed uberblock and using that to open
* the MOS.
*/
if (checkpoint_rewind) {
/*
* If we are rewinding to the checkpoint update config cache
* anyway.
*/
update_config_cache = B_TRUE;
/*
* Extract the checkpointed uberblock from the current MOS
* and use this as the pool's uberblock from now on. If the
* pool is imported as writeable we also write the checkpoint
* uberblock to the labels, making the rewind permanent.
*/
error = spa_ld_checkpoint_rewind(spa);
if (error != 0)
return (error);
/*
* Redo the loading process again with the
* checkpointed uberblock.
*/
spa_ld_prepare_for_reload(spa);
spa_load_note(spa, "LOADING checkpointed uberblock");
error = spa_ld_mos_with_trusted_config(spa, type, NULL);
if (error != 0)
return (error);
}
/*
* Retrieve the checkpoint txg if the pool has a checkpoint.
*/
error = spa_ld_read_checkpoint_txg(spa);
if (error != 0)
return (error);
/*
* Retrieve the mapping of indirect vdevs. Those vdevs were removed
* from the pool and their contents were re-mapped to other vdevs. Note
* that everything that we read before this step must have been
* rewritten on concrete vdevs after the last device removal was
* initiated. Otherwise we could be reading from indirect vdevs before
* we have loaded their mappings.
*/
error = spa_ld_open_indirect_vdev_metadata(spa);
if (error != 0)
return (error);
/*
* Retrieve the full list of active features from the MOS and check if
* they are all supported.
*/
error = spa_ld_check_features(spa, &missing_feat_write);
if (error != 0)
return (error);
/*
* Load several special directories from the MOS needed by the dsl_pool
* layer.
*/
error = spa_ld_load_special_directories(spa);
if (error != 0)
return (error);
/*
* Retrieve pool properties from the MOS.
*/
error = spa_ld_get_props(spa);
if (error != 0)
return (error);
/*
* Retrieve the list of auxiliary devices - cache devices and spares -
* and open them.
*/
error = spa_ld_open_aux_vdevs(spa, type);
if (error != 0)
return (error);
/*
* Load the metadata for all vdevs. Also check if unopenable devices
* should be autoreplaced.
*/
error = spa_ld_load_vdev_metadata(spa);
if (error != 0)
return (error);
error = spa_ld_load_dedup_tables(spa);
if (error != 0)
return (error);
error = spa_ld_load_brt(spa);
if (error != 0)
return (error);
/*
* Verify the logs now to make sure we don't have any unexpected errors
* when we claim log blocks later.
*/
error = spa_ld_verify_logs(spa, type, ereport);
if (error != 0)
return (error);
if (missing_feat_write) {
ASSERT(spa->spa_load_state == SPA_LOAD_TRYIMPORT);
/*
* At this point, we know that we can open the pool in
* read-only mode but not read-write mode. We now have enough
* information and can return to userland.
*/
return (spa_vdev_err(spa->spa_root_vdev, VDEV_AUX_UNSUP_FEAT,
ENOTSUP));
}
/*
* Traverse the last txgs to make sure the pool was left off in a safe
* state. When performing an extreme rewind, we verify the whole pool,
* which can take a very long time.
*/
error = spa_ld_verify_pool_data(spa);
if (error != 0)
return (error);
/*
* Calculate the deflated space for the pool. This must be done before
* we write anything to the pool because we'd need to update the space
* accounting using the deflated sizes.
*/
spa_update_dspace(spa);
/*
* We have now retrieved all the information we needed to open the
* pool. If we are importing the pool in read-write mode, a few
* additional steps must be performed to finish the import.
*/
if (spa_writeable(spa) && (spa->spa_load_state == SPA_LOAD_RECOVER ||
spa->spa_load_max_txg == UINT64_MAX)) {
uint64_t config_cache_txg = spa->spa_config_txg;
ASSERT(spa->spa_load_state != SPA_LOAD_TRYIMPORT);
+ /*
+ * Before we do any zio_write's, complete the raidz expansion
+ * scratch space copying, if necessary.
+ */
+ if (RRSS_GET_STATE(&spa->spa_uberblock) == RRSS_SCRATCH_VALID)
+ vdev_raidz_reflow_copy_scratch(spa);
+
/*
* In case of a checkpoint rewind, log the original txg
* of the checkpointed uberblock.
*/
if (checkpoint_rewind) {
spa_history_log_internal(spa, "checkpoint rewind",
NULL, "rewound state to txg=%llu",
(u_longlong_t)spa->spa_uberblock.ub_checkpoint_txg);
}
/*
* Traverse the ZIL and claim all blocks.
*/
spa_ld_claim_log_blocks(spa);
/*
* Kick-off the syncing thread.
*/
spa->spa_sync_on = B_TRUE;
txg_sync_start(spa->spa_dsl_pool);
mmp_thread_start(spa);
/*
* Wait for all claims to sync. We sync up to the highest
* claimed log block birth time so that claimed log blocks
* don't appear to be from the future. spa_claim_max_txg
* will have been set for us by ZIL traversal operations
* performed above.
*/
txg_wait_synced(spa->spa_dsl_pool, spa->spa_claim_max_txg);
/*
* Check if we need to request an update of the config. On the
* next sync, we would update the config stored in vdev labels
* and the cachefile (by default /etc/zfs/zpool.cache).
*/
spa_ld_check_for_config_update(spa, config_cache_txg,
update_config_cache);
/*
* Check if a rebuild was in progress and if so resume it.
* Then check all DTLs to see if anything needs resilvering.
* The resilver will be deferred if a rebuild was started.
*/
if (vdev_rebuild_active(spa->spa_root_vdev)) {
vdev_rebuild_restart(spa);
} else if (!dsl_scan_resilvering(spa->spa_dsl_pool) &&
vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL)) {
spa_async_request(spa, SPA_ASYNC_RESILVER);
}
/*
* Log the fact that we booted up (so that we can detect if
* we rebooted in the middle of an operation).
*/
spa_history_log_version(spa, "open", NULL);
spa_restart_removal(spa);
spa_spawn_aux_threads(spa);
/*
* Delete any inconsistent datasets.
*
* Note:
* Since we may be issuing deletes for clones here,
* we make sure to do so after we've spawned all the
* auxiliary threads above (from which the livelist
* deletion zthr is part of).
*/
(void) dmu_objset_find(spa_name(spa),
dsl_destroy_inconsistent, NULL, DS_FIND_CHILDREN);
/*
* Clean up any stale temporary dataset userrefs.
*/
dsl_pool_clean_tmp_userrefs(spa->spa_dsl_pool);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_initialize_restart(spa->spa_root_vdev);
vdev_trim_restart(spa->spa_root_vdev);
vdev_autotrim_restart(spa);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
spa_import_progress_remove(spa_guid(spa));
spa_async_request(spa, SPA_ASYNC_L2CACHE_REBUILD);
spa_load_note(spa, "LOADED");
return (0);
}
static int
spa_load_retry(spa_t *spa, spa_load_state_t state)
{
spa_mode_t mode = spa->spa_mode;
spa_unload(spa);
spa_deactivate(spa);
spa->spa_load_max_txg = spa->spa_uberblock.ub_txg - 1;
spa_activate(spa, mode);
spa_async_suspend(spa);
spa_load_note(spa, "spa_load_retry: rewind, max txg: %llu",
(u_longlong_t)spa->spa_load_max_txg);
return (spa_load(spa, state, SPA_IMPORT_EXISTING));
}
/*
* If spa_load() fails this function will try loading prior txg's. If
* 'state' is SPA_LOAD_RECOVER and one of these loads succeeds the pool
* will be rewound to that txg. If 'state' is not SPA_LOAD_RECOVER this
* function will not rewind the pool and will return the same error as
* spa_load().
*/
static int
spa_load_best(spa_t *spa, spa_load_state_t state, uint64_t max_request,
int rewind_flags)
{
nvlist_t *loadinfo = NULL;
nvlist_t *config = NULL;
int load_error, rewind_error;
uint64_t safe_rewind_txg;
uint64_t min_txg;
if (spa->spa_load_txg && state == SPA_LOAD_RECOVER) {
spa->spa_load_max_txg = spa->spa_load_txg;
spa_set_log_state(spa, SPA_LOG_CLEAR);
} else {
spa->spa_load_max_txg = max_request;
if (max_request != UINT64_MAX)
spa->spa_extreme_rewind = B_TRUE;
}
load_error = rewind_error = spa_load(spa, state, SPA_IMPORT_EXISTING);
if (load_error == 0)
return (0);
if (load_error == ZFS_ERR_NO_CHECKPOINT) {
/*
* When attempting checkpoint-rewind on a pool with no
* checkpoint, we should not attempt to load uberblocks
* from previous txgs when spa_load fails.
*/
ASSERT(spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT);
spa_import_progress_remove(spa_guid(spa));
return (load_error);
}
if (spa->spa_root_vdev != NULL)
config = spa_config_generate(spa, NULL, -1ULL, B_TRUE);
spa->spa_last_ubsync_txg = spa->spa_uberblock.ub_txg;
spa->spa_last_ubsync_txg_ts = spa->spa_uberblock.ub_timestamp;
if (rewind_flags & ZPOOL_NEVER_REWIND) {
nvlist_free(config);
spa_import_progress_remove(spa_guid(spa));
return (load_error);
}
if (state == SPA_LOAD_RECOVER) {
/* Price of rolling back is discarding txgs, including log */
spa_set_log_state(spa, SPA_LOG_CLEAR);
} else {
/*
* If we aren't rolling back save the load info from our first
* import attempt so that we can restore it after attempting
* to rewind.
*/
loadinfo = spa->spa_load_info;
spa->spa_load_info = fnvlist_alloc();
}
spa->spa_load_max_txg = spa->spa_last_ubsync_txg;
safe_rewind_txg = spa->spa_last_ubsync_txg - TXG_DEFER_SIZE;
min_txg = (rewind_flags & ZPOOL_EXTREME_REWIND) ?
TXG_INITIAL : safe_rewind_txg;
/*
* Continue as long as we're finding errors, we're still within
* the acceptable rewind range, and we're still finding uberblocks
*/
while (rewind_error && spa->spa_uberblock.ub_txg >= min_txg &&
spa->spa_uberblock.ub_txg <= spa->spa_load_max_txg) {
if (spa->spa_load_max_txg < safe_rewind_txg)
spa->spa_extreme_rewind = B_TRUE;
rewind_error = spa_load_retry(spa, state);
}
spa->spa_extreme_rewind = B_FALSE;
spa->spa_load_max_txg = UINT64_MAX;
if (config && (rewind_error || state != SPA_LOAD_RECOVER))
spa_config_set(spa, config);
else
nvlist_free(config);
if (state == SPA_LOAD_RECOVER) {
ASSERT3P(loadinfo, ==, NULL);
spa_import_progress_remove(spa_guid(spa));
return (rewind_error);
} else {
/* Store the rewind info as part of the initial load info */
fnvlist_add_nvlist(loadinfo, ZPOOL_CONFIG_REWIND_INFO,
spa->spa_load_info);
/* Restore the initial load info */
fnvlist_free(spa->spa_load_info);
spa->spa_load_info = loadinfo;
spa_import_progress_remove(spa_guid(spa));
return (load_error);
}
}
/*
* Pool Open/Import
*
* The import case is identical to an open except that the configuration is sent
* down from userland, instead of grabbed from the configuration cache. For the
* case of an open, the pool configuration will exist in the
* POOL_STATE_UNINITIALIZED state.
*
* The stats information (gen/count/ustats) is used to gather vdev statistics at
* the same time open the pool, without having to keep around the spa_t in some
* ambiguous state.
*/
static int
spa_open_common(const char *pool, spa_t **spapp, const void *tag,
nvlist_t *nvpolicy, nvlist_t **config)
{
spa_t *spa;
spa_load_state_t state = SPA_LOAD_OPEN;
int error;
int locked = B_FALSE;
int firstopen = B_FALSE;
*spapp = NULL;
/*
* As disgusting as this is, we need to support recursive calls to this
* function because dsl_dir_open() is called during spa_load(), and ends
* up calling spa_open() again. The real fix is to figure out how to
* avoid dsl_dir_open() calling this in the first place.
*/
if (MUTEX_NOT_HELD(&spa_namespace_lock)) {
mutex_enter(&spa_namespace_lock);
locked = B_TRUE;
}
if ((spa = spa_lookup(pool)) == NULL) {
if (locked)
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(ENOENT));
}
if (spa->spa_state == POOL_STATE_UNINITIALIZED) {
zpool_load_policy_t policy;
firstopen = B_TRUE;
zpool_get_load_policy(nvpolicy ? nvpolicy : spa->spa_config,
&policy);
if (policy.zlp_rewind & ZPOOL_DO_REWIND)
state = SPA_LOAD_RECOVER;
spa_activate(spa, spa_mode_global);
if (state != SPA_LOAD_RECOVER)
spa->spa_last_ubsync_txg = spa->spa_load_txg = 0;
spa->spa_config_source = SPA_CONFIG_SRC_CACHEFILE;
zfs_dbgmsg("spa_open_common: opening %s", pool);
error = spa_load_best(spa, state, policy.zlp_txg,
policy.zlp_rewind);
if (error == EBADF) {
/*
* If vdev_validate() returns failure (indicated by
* EBADF), it indicates that one of the vdevs indicates
* that the pool has been exported or destroyed. If
* this is the case, the config cache is out of sync and
* we should remove the pool from the namespace.
*/
spa_unload(spa);
spa_deactivate(spa);
spa_write_cachefile(spa, B_TRUE, B_TRUE, B_FALSE);
spa_remove(spa);
if (locked)
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(ENOENT));
}
if (error) {
/*
* We can't open the pool, but we still have useful
* information: the state of each vdev after the
* attempted vdev_open(). Return this to the user.
*/
if (config != NULL && spa->spa_config) {
*config = fnvlist_dup(spa->spa_config);
fnvlist_add_nvlist(*config,
ZPOOL_CONFIG_LOAD_INFO,
spa->spa_load_info);
}
spa_unload(spa);
spa_deactivate(spa);
spa->spa_last_open_failed = error;
if (locked)
mutex_exit(&spa_namespace_lock);
*spapp = NULL;
return (error);
}
}
spa_open_ref(spa, tag);
if (config != NULL)
*config = spa_config_generate(spa, NULL, -1ULL, B_TRUE);
/*
* If we've recovered the pool, pass back any information we
* gathered while doing the load.
*/
if (state == SPA_LOAD_RECOVER && config != NULL) {
fnvlist_add_nvlist(*config, ZPOOL_CONFIG_LOAD_INFO,
spa->spa_load_info);
}
if (locked) {
spa->spa_last_open_failed = 0;
spa->spa_last_ubsync_txg = 0;
spa->spa_load_txg = 0;
mutex_exit(&spa_namespace_lock);
}
if (firstopen)
zvol_create_minors_recursive(spa_name(spa));
*spapp = spa;
return (0);
}
int
spa_open_rewind(const char *name, spa_t **spapp, const void *tag,
nvlist_t *policy, nvlist_t **config)
{
return (spa_open_common(name, spapp, tag, policy, config));
}
int
spa_open(const char *name, spa_t **spapp, const void *tag)
{
return (spa_open_common(name, spapp, tag, NULL, NULL));
}
/*
* Lookup the given spa_t, incrementing the inject count in the process,
* preventing it from being exported or destroyed.
*/
spa_t *
spa_inject_addref(char *name)
{
spa_t *spa;
mutex_enter(&spa_namespace_lock);
if ((spa = spa_lookup(name)) == NULL) {
mutex_exit(&spa_namespace_lock);
return (NULL);
}
spa->spa_inject_ref++;
mutex_exit(&spa_namespace_lock);
return (spa);
}
void
spa_inject_delref(spa_t *spa)
{
mutex_enter(&spa_namespace_lock);
spa->spa_inject_ref--;
mutex_exit(&spa_namespace_lock);
}
/*
* Add spares device information to the nvlist.
*/
static void
spa_add_spares(spa_t *spa, nvlist_t *config)
{
nvlist_t **spares;
uint_t i, nspares;
nvlist_t *nvroot;
uint64_t guid;
vdev_stat_t *vs;
uint_t vsc;
uint64_t pool;
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER));
if (spa->spa_spares.sav_count == 0)
return;
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
VERIFY0(nvlist_lookup_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, &spares, &nspares));
if (nspares != 0) {
fnvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
(const nvlist_t * const *)spares, nspares);
VERIFY0(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares));
/*
* Go through and find any spares which have since been
* repurposed as an active spare. If this is the case, update
* their status appropriately.
*/
for (i = 0; i < nspares; i++) {
guid = fnvlist_lookup_uint64(spares[i],
ZPOOL_CONFIG_GUID);
VERIFY0(nvlist_lookup_uint64_array(spares[i],
ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc));
if (spa_spare_exists(guid, &pool, NULL) &&
pool != 0ULL) {
vs->vs_state = VDEV_STATE_CANT_OPEN;
vs->vs_aux = VDEV_AUX_SPARED;
} else {
vs->vs_state =
spa->spa_spares.sav_vdevs[i]->vdev_state;
}
}
}
}
/*
* Add l2cache device information to the nvlist, including vdev stats.
*/
static void
spa_add_l2cache(spa_t *spa, nvlist_t *config)
{
nvlist_t **l2cache;
uint_t i, j, nl2cache;
nvlist_t *nvroot;
uint64_t guid;
vdev_t *vd;
vdev_stat_t *vs;
uint_t vsc;
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER));
if (spa->spa_l2cache.sav_count == 0)
return;
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
VERIFY0(nvlist_lookup_nvlist_array(spa->spa_l2cache.sav_config,
ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache));
if (nl2cache != 0) {
fnvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
(const nvlist_t * const *)l2cache, nl2cache);
VERIFY0(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache));
/*
* Update level 2 cache device stats.
*/
for (i = 0; i < nl2cache; i++) {
guid = fnvlist_lookup_uint64(l2cache[i],
ZPOOL_CONFIG_GUID);
vd = NULL;
for (j = 0; j < spa->spa_l2cache.sav_count; j++) {
if (guid ==
spa->spa_l2cache.sav_vdevs[j]->vdev_guid) {
vd = spa->spa_l2cache.sav_vdevs[j];
break;
}
}
ASSERT(vd != NULL);
VERIFY0(nvlist_lookup_uint64_array(l2cache[i],
ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc));
vdev_get_stats(vd, vs);
vdev_config_generate_stats(vd, l2cache[i]);
}
}
}
static void
spa_feature_stats_from_disk(spa_t *spa, nvlist_t *features)
{
zap_cursor_t zc;
zap_attribute_t za;
if (spa->spa_feat_for_read_obj != 0) {
for (zap_cursor_init(&zc, spa->spa_meta_objset,
spa->spa_feat_for_read_obj);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
ASSERT(za.za_integer_length == sizeof (uint64_t) &&
za.za_num_integers == 1);
VERIFY0(nvlist_add_uint64(features, za.za_name,
za.za_first_integer));
}
zap_cursor_fini(&zc);
}
if (spa->spa_feat_for_write_obj != 0) {
for (zap_cursor_init(&zc, spa->spa_meta_objset,
spa->spa_feat_for_write_obj);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
ASSERT(za.za_integer_length == sizeof (uint64_t) &&
za.za_num_integers == 1);
VERIFY0(nvlist_add_uint64(features, za.za_name,
za.za_first_integer));
}
zap_cursor_fini(&zc);
}
}
static void
spa_feature_stats_from_cache(spa_t *spa, nvlist_t *features)
{
int i;
for (i = 0; i < SPA_FEATURES; i++) {
zfeature_info_t feature = spa_feature_table[i];
uint64_t refcount;
if (feature_get_refcount(spa, &feature, &refcount) != 0)
continue;
VERIFY0(nvlist_add_uint64(features, feature.fi_guid, refcount));
}
}
/*
* Store a list of pool features and their reference counts in the
* config.
*
* The first time this is called on a spa, allocate a new nvlist, fetch
* the pool features and reference counts from disk, then save the list
* in the spa. In subsequent calls on the same spa use the saved nvlist
* and refresh its values from the cached reference counts. This
* ensures we don't block here on I/O on a suspended pool so 'zpool
* clear' can resume the pool.
*/
static void
spa_add_feature_stats(spa_t *spa, nvlist_t *config)
{
nvlist_t *features;
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER));
mutex_enter(&spa->spa_feat_stats_lock);
features = spa->spa_feat_stats;
if (features != NULL) {
spa_feature_stats_from_cache(spa, features);
} else {
VERIFY0(nvlist_alloc(&features, NV_UNIQUE_NAME, KM_SLEEP));
spa->spa_feat_stats = features;
spa_feature_stats_from_disk(spa, features);
}
VERIFY0(nvlist_add_nvlist(config, ZPOOL_CONFIG_FEATURE_STATS,
features));
mutex_exit(&spa->spa_feat_stats_lock);
}
int
spa_get_stats(const char *name, nvlist_t **config,
char *altroot, size_t buflen)
{
int error;
spa_t *spa;
*config = NULL;
error = spa_open_common(name, &spa, FTAG, NULL, config);
if (spa != NULL) {
/*
* This still leaves a window of inconsistency where the spares
* or l2cache devices could change and the config would be
* self-inconsistent.
*/
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
if (*config != NULL) {
uint64_t loadtimes[2];
loadtimes[0] = spa->spa_loaded_ts.tv_sec;
loadtimes[1] = spa->spa_loaded_ts.tv_nsec;
fnvlist_add_uint64_array(*config,
ZPOOL_CONFIG_LOADED_TIME, loadtimes, 2);
fnvlist_add_uint64(*config,
ZPOOL_CONFIG_ERRCOUNT,
spa_approx_errlog_size(spa));
if (spa_suspended(spa)) {
fnvlist_add_uint64(*config,
ZPOOL_CONFIG_SUSPENDED,
spa->spa_failmode);
fnvlist_add_uint64(*config,
ZPOOL_CONFIG_SUSPENDED_REASON,
spa->spa_suspended);
}
spa_add_spares(spa, *config);
spa_add_l2cache(spa, *config);
spa_add_feature_stats(spa, *config);
}
}
/*
* We want to get the alternate root even for faulted pools, so we cheat
* and call spa_lookup() directly.
*/
if (altroot) {
if (spa == NULL) {
mutex_enter(&spa_namespace_lock);
spa = spa_lookup(name);
if (spa)
spa_altroot(spa, altroot, buflen);
else
altroot[0] = '\0';
spa = NULL;
mutex_exit(&spa_namespace_lock);
} else {
spa_altroot(spa, altroot, buflen);
}
}
if (spa != NULL) {
spa_config_exit(spa, SCL_CONFIG, FTAG);
spa_close(spa, FTAG);
}
return (error);
}
/*
* Validate that the auxiliary device array is well formed. We must have an
* array of nvlists, each which describes a valid leaf vdev. If this is an
* import (mode is VDEV_ALLOC_SPARE), then we allow corrupted spares to be
* specified, as long as they are well-formed.
*/
static int
spa_validate_aux_devs(spa_t *spa, nvlist_t *nvroot, uint64_t crtxg, int mode,
spa_aux_vdev_t *sav, const char *config, uint64_t version,
vdev_labeltype_t label)
{
nvlist_t **dev;
uint_t i, ndev;
vdev_t *vd;
int error;
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
/*
* It's acceptable to have no devs specified.
*/
if (nvlist_lookup_nvlist_array(nvroot, config, &dev, &ndev) != 0)
return (0);
if (ndev == 0)
return (SET_ERROR(EINVAL));
/*
* Make sure the pool is formatted with a version that supports this
* device type.
*/
if (spa_version(spa) < version)
return (SET_ERROR(ENOTSUP));
/*
* Set the pending device list so we correctly handle device in-use
* checking.
*/
sav->sav_pending = dev;
sav->sav_npending = ndev;
for (i = 0; i < ndev; i++) {
if ((error = spa_config_parse(spa, &vd, dev[i], NULL, 0,
mode)) != 0)
goto out;
if (!vd->vdev_ops->vdev_op_leaf) {
vdev_free(vd);
error = SET_ERROR(EINVAL);
goto out;
}
vd->vdev_top = vd;
if ((error = vdev_open(vd)) == 0 &&
(error = vdev_label_init(vd, crtxg, label)) == 0) {
fnvlist_add_uint64(dev[i], ZPOOL_CONFIG_GUID,
vd->vdev_guid);
}
vdev_free(vd);
if (error &&
(mode != VDEV_ALLOC_SPARE && mode != VDEV_ALLOC_L2CACHE))
goto out;
else
error = 0;
}
out:
sav->sav_pending = NULL;
sav->sav_npending = 0;
return (error);
}
static int
spa_validate_aux(spa_t *spa, nvlist_t *nvroot, uint64_t crtxg, int mode)
{
int error;
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
if ((error = spa_validate_aux_devs(spa, nvroot, crtxg, mode,
&spa->spa_spares, ZPOOL_CONFIG_SPARES, SPA_VERSION_SPARES,
VDEV_LABEL_SPARE)) != 0) {
return (error);
}
return (spa_validate_aux_devs(spa, nvroot, crtxg, mode,
&spa->spa_l2cache, ZPOOL_CONFIG_L2CACHE, SPA_VERSION_L2CACHE,
VDEV_LABEL_L2CACHE));
}
static void
spa_set_aux_vdevs(spa_aux_vdev_t *sav, nvlist_t **devs, int ndevs,
const char *config)
{
int i;
if (sav->sav_config != NULL) {
nvlist_t **olddevs;
uint_t oldndevs;
nvlist_t **newdevs;
/*
* Generate new dev list by concatenating with the
* current dev list.
*/
VERIFY0(nvlist_lookup_nvlist_array(sav->sav_config, config,
&olddevs, &oldndevs));
newdevs = kmem_alloc(sizeof (void *) *
(ndevs + oldndevs), KM_SLEEP);
for (i = 0; i < oldndevs; i++)
newdevs[i] = fnvlist_dup(olddevs[i]);
for (i = 0; i < ndevs; i++)
newdevs[i + oldndevs] = fnvlist_dup(devs[i]);
fnvlist_remove(sav->sav_config, config);
fnvlist_add_nvlist_array(sav->sav_config, config,
(const nvlist_t * const *)newdevs, ndevs + oldndevs);
for (i = 0; i < oldndevs + ndevs; i++)
nvlist_free(newdevs[i]);
kmem_free(newdevs, (oldndevs + ndevs) * sizeof (void *));
} else {
/*
* Generate a new dev list.
*/
sav->sav_config = fnvlist_alloc();
fnvlist_add_nvlist_array(sav->sav_config, config,
(const nvlist_t * const *)devs, ndevs);
}
}
/*
* Stop and drop level 2 ARC devices
*/
void
spa_l2cache_drop(spa_t *spa)
{
vdev_t *vd;
int i;
spa_aux_vdev_t *sav = &spa->spa_l2cache;
for (i = 0; i < sav->sav_count; i++) {
uint64_t pool;
vd = sav->sav_vdevs[i];
ASSERT(vd != NULL);
if (spa_l2cache_exists(vd->vdev_guid, &pool) &&
pool != 0ULL && l2arc_vdev_present(vd))
l2arc_remove_vdev(vd);
}
}
/*
* Verify encryption parameters for spa creation. If we are encrypting, we must
* have the encryption feature flag enabled.
*/
static int
spa_create_check_encryption_params(dsl_crypto_params_t *dcp,
boolean_t has_encryption)
{
if (dcp->cp_crypt != ZIO_CRYPT_OFF &&
dcp->cp_crypt != ZIO_CRYPT_INHERIT &&
!has_encryption)
return (SET_ERROR(ENOTSUP));
return (dmu_objset_create_crypt_check(NULL, dcp, NULL));
}
/*
* Pool Creation
*/
int
spa_create(const char *pool, nvlist_t *nvroot, nvlist_t *props,
nvlist_t *zplprops, dsl_crypto_params_t *dcp)
{
spa_t *spa;
const char *altroot = NULL;
vdev_t *rvd;
dsl_pool_t *dp;
dmu_tx_t *tx;
int error = 0;
uint64_t txg = TXG_INITIAL;
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
uint64_t version, obj, ndraid = 0;
boolean_t has_features;
boolean_t has_encryption;
boolean_t has_allocclass;
spa_feature_t feat;
const char *feat_name;
const char *poolname;
nvlist_t *nvl;
if (props == NULL ||
nvlist_lookup_string(props, "tname", &poolname) != 0)
poolname = (char *)pool;
/*
* If this pool already exists, return failure.
*/
mutex_enter(&spa_namespace_lock);
if (spa_lookup(poolname) != NULL) {
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(EEXIST));
}
/*
* Allocate a new spa_t structure.
*/
nvl = fnvlist_alloc();
fnvlist_add_string(nvl, ZPOOL_CONFIG_POOL_NAME, pool);
(void) nvlist_lookup_string(props,
zpool_prop_to_name(ZPOOL_PROP_ALTROOT), &altroot);
spa = spa_add(poolname, nvl, altroot);
fnvlist_free(nvl);
spa_activate(spa, spa_mode_global);
if (props && (error = spa_prop_validate(spa, props))) {
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (error);
}
/*
* Temporary pool names should never be written to disk.
*/
if (poolname != pool)
spa->spa_import_flags |= ZFS_IMPORT_TEMP_NAME;
has_features = B_FALSE;
has_encryption = B_FALSE;
has_allocclass = B_FALSE;
for (nvpair_t *elem = nvlist_next_nvpair(props, NULL);
elem != NULL; elem = nvlist_next_nvpair(props, elem)) {
if (zpool_prop_feature(nvpair_name(elem))) {
has_features = B_TRUE;
feat_name = strchr(nvpair_name(elem), '@') + 1;
VERIFY0(zfeature_lookup_name(feat_name, &feat));
if (feat == SPA_FEATURE_ENCRYPTION)
has_encryption = B_TRUE;
if (feat == SPA_FEATURE_ALLOCATION_CLASSES)
has_allocclass = B_TRUE;
}
}
/* verify encryption params, if they were provided */
if (dcp != NULL) {
error = spa_create_check_encryption_params(dcp, has_encryption);
if (error != 0) {
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (error);
}
}
if (!has_allocclass && zfs_special_devs(nvroot, NULL)) {
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (ENOTSUP);
}
if (has_features || nvlist_lookup_uint64(props,
zpool_prop_to_name(ZPOOL_PROP_VERSION), &version) != 0) {
version = SPA_VERSION;
}
ASSERT(SPA_VERSION_IS_SUPPORTED(version));
spa->spa_first_txg = txg;
spa->spa_uberblock.ub_txg = txg - 1;
spa->spa_uberblock.ub_version = version;
spa->spa_ubsync = spa->spa_uberblock;
spa->spa_load_state = SPA_LOAD_CREATE;
spa->spa_removing_phys.sr_state = DSS_NONE;
spa->spa_removing_phys.sr_removing_vdev = -1;
spa->spa_removing_phys.sr_prev_indirect_vdev = -1;
spa->spa_indirect_vdevs_loaded = B_TRUE;
/*
* Create "The Godfather" zio to hold all async IOs
*/
spa->spa_async_zio_root = kmem_alloc(max_ncpus * sizeof (void *),
KM_SLEEP);
for (int i = 0; i < max_ncpus; i++) {
spa->spa_async_zio_root[i] = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE |
ZIO_FLAG_GODFATHER);
}
/*
* Create the root vdev.
*/
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
error = spa_config_parse(spa, &rvd, nvroot, NULL, 0, VDEV_ALLOC_ADD);
ASSERT(error != 0 || rvd != NULL);
ASSERT(error != 0 || spa->spa_root_vdev == rvd);
if (error == 0 && !zfs_allocatable_devs(nvroot))
error = SET_ERROR(EINVAL);
if (error == 0 &&
(error = vdev_create(rvd, txg, B_FALSE)) == 0 &&
(error = vdev_draid_spare_create(nvroot, rvd, &ndraid, 0)) == 0 &&
(error = spa_validate_aux(spa, nvroot, txg, VDEV_ALLOC_ADD)) == 0) {
/*
* instantiate the metaslab groups (this will dirty the vdevs)
* we can no longer error exit past this point
*/
for (int c = 0; error == 0 && c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
vdev_metaslab_set_size(vd);
vdev_expand(vd, txg);
}
}
spa_config_exit(spa, SCL_ALL, FTAG);
if (error != 0) {
spa_unload(spa);
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (error);
}
/*
* Get the list of spares, if specified.
*/
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares) == 0) {
spa->spa_spares.sav_config = fnvlist_alloc();
fnvlist_add_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, (const nvlist_t * const *)spares,
nspares);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_spares(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
spa->spa_spares.sav_sync = B_TRUE;
}
/*
* Get the list of level 2 cache devices, if specified.
*/
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache) == 0) {
VERIFY0(nvlist_alloc(&spa->spa_l2cache.sav_config,
NV_UNIQUE_NAME, KM_SLEEP));
fnvlist_add_nvlist_array(spa->spa_l2cache.sav_config,
ZPOOL_CONFIG_L2CACHE, (const nvlist_t * const *)l2cache,
nl2cache);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_l2cache(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
spa->spa_l2cache.sav_sync = B_TRUE;
}
spa->spa_is_initializing = B_TRUE;
spa->spa_dsl_pool = dp = dsl_pool_create(spa, zplprops, dcp, txg);
spa->spa_is_initializing = B_FALSE;
/*
* Create DDTs (dedup tables).
*/
ddt_create(spa);
/*
* Create BRT table and BRT table object.
*/
brt_create(spa);
spa_update_dspace(spa);
tx = dmu_tx_create_assigned(dp, txg);
/*
* Create the pool's history object.
*/
if (version >= SPA_VERSION_ZPOOL_HISTORY && !spa->spa_history)
spa_history_create_obj(spa, tx);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_CREATE);
spa_history_log_version(spa, "create", tx);
/*
* Create the pool config object.
*/
spa->spa_config_object = dmu_object_alloc(spa->spa_meta_objset,
DMU_OT_PACKED_NVLIST, SPA_CONFIG_BLOCKSIZE,
DMU_OT_PACKED_NVLIST_SIZE, sizeof (uint64_t), tx);
if (zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CONFIG,
sizeof (uint64_t), 1, &spa->spa_config_object, tx) != 0) {
cmn_err(CE_PANIC, "failed to add pool config");
}
if (zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CREATION_VERSION,
sizeof (uint64_t), 1, &version, tx) != 0) {
cmn_err(CE_PANIC, "failed to add pool version");
}
/* Newly created pools with the right version are always deflated. */
if (version >= SPA_VERSION_RAIDZ_DEFLATE) {
spa->spa_deflate = TRUE;
if (zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_DEFLATE,
sizeof (uint64_t), 1, &spa->spa_deflate, tx) != 0) {
cmn_err(CE_PANIC, "failed to add deflate");
}
}
/*
* Create the deferred-free bpobj. Turn off compression
* because sync-to-convergence takes longer if the blocksize
* keeps changing.
*/
obj = bpobj_alloc(spa->spa_meta_objset, 1 << 14, tx);
dmu_object_set_compress(spa->spa_meta_objset, obj,
ZIO_COMPRESS_OFF, tx);
if (zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SYNC_BPOBJ,
sizeof (uint64_t), 1, &obj, tx) != 0) {
cmn_err(CE_PANIC, "failed to add bpobj");
}
VERIFY3U(0, ==, bpobj_open(&spa->spa_deferred_bpobj,
spa->spa_meta_objset, obj));
/*
* Generate some random noise for salted checksums to operate on.
*/
(void) random_get_pseudo_bytes(spa->spa_cksum_salt.zcs_bytes,
sizeof (spa->spa_cksum_salt.zcs_bytes));
/*
* Set pool properties.
*/
spa->spa_bootfs = zpool_prop_default_numeric(ZPOOL_PROP_BOOTFS);
spa->spa_delegation = zpool_prop_default_numeric(ZPOOL_PROP_DELEGATION);
spa->spa_failmode = zpool_prop_default_numeric(ZPOOL_PROP_FAILUREMODE);
spa->spa_autoexpand = zpool_prop_default_numeric(ZPOOL_PROP_AUTOEXPAND);
spa->spa_multihost = zpool_prop_default_numeric(ZPOOL_PROP_MULTIHOST);
spa->spa_autotrim = zpool_prop_default_numeric(ZPOOL_PROP_AUTOTRIM);
if (props != NULL) {
spa_configfile_set(spa, props, B_FALSE);
spa_sync_props(props, tx);
}
for (int i = 0; i < ndraid; i++)
spa_feature_incr(spa, SPA_FEATURE_DRAID, tx);
dmu_tx_commit(tx);
spa->spa_sync_on = B_TRUE;
txg_sync_start(dp);
mmp_thread_start(spa);
txg_wait_synced(dp, txg);
spa_spawn_aux_threads(spa);
spa_write_cachefile(spa, B_FALSE, B_TRUE, B_TRUE);
/*
* Don't count references from objsets that are already closed
* and are making their way through the eviction process.
*/
spa_evicting_os_wait(spa);
spa->spa_minref = zfs_refcount_count(&spa->spa_refcount);
spa->spa_load_state = SPA_LOAD_NONE;
spa_import_os(spa);
mutex_exit(&spa_namespace_lock);
return (0);
}
/*
* Import a non-root pool into the system.
*/
int
spa_import(char *pool, nvlist_t *config, nvlist_t *props, uint64_t flags)
{
spa_t *spa;
const char *altroot = NULL;
spa_load_state_t state = SPA_LOAD_IMPORT;
zpool_load_policy_t policy;
spa_mode_t mode = spa_mode_global;
uint64_t readonly = B_FALSE;
int error;
nvlist_t *nvroot;
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
/*
* If a pool with this name exists, return failure.
*/
mutex_enter(&spa_namespace_lock);
if (spa_lookup(pool) != NULL) {
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(EEXIST));
}
/*
* Create and initialize the spa structure.
*/
(void) nvlist_lookup_string(props,
zpool_prop_to_name(ZPOOL_PROP_ALTROOT), &altroot);
(void) nvlist_lookup_uint64(props,
zpool_prop_to_name(ZPOOL_PROP_READONLY), &readonly);
if (readonly)
mode = SPA_MODE_READ;
spa = spa_add(pool, config, altroot);
spa->spa_import_flags = flags;
/*
* Verbatim import - Take a pool and insert it into the namespace
* as if it had been loaded at boot.
*/
if (spa->spa_import_flags & ZFS_IMPORT_VERBATIM) {
if (props != NULL)
spa_configfile_set(spa, props, B_FALSE);
spa_write_cachefile(spa, B_FALSE, B_TRUE, B_FALSE);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_IMPORT);
zfs_dbgmsg("spa_import: verbatim import of %s", pool);
mutex_exit(&spa_namespace_lock);
return (0);
}
spa_activate(spa, mode);
/*
* Don't start async tasks until we know everything is healthy.
*/
spa_async_suspend(spa);
zpool_get_load_policy(config, &policy);
if (policy.zlp_rewind & ZPOOL_DO_REWIND)
state = SPA_LOAD_RECOVER;
spa->spa_config_source = SPA_CONFIG_SRC_TRYIMPORT;
if (state != SPA_LOAD_RECOVER) {
spa->spa_last_ubsync_txg = spa->spa_load_txg = 0;
zfs_dbgmsg("spa_import: importing %s", pool);
} else {
zfs_dbgmsg("spa_import: importing %s, max_txg=%lld "
"(RECOVERY MODE)", pool, (longlong_t)policy.zlp_txg);
}
error = spa_load_best(spa, state, policy.zlp_txg, policy.zlp_rewind);
/*
* Propagate anything learned while loading the pool and pass it
* back to caller (i.e. rewind info, missing devices, etc).
*/
fnvlist_add_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, spa->spa_load_info);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
/*
* Toss any existing sparelist, as it doesn't have any validity
* anymore, and conflicts with spa_has_spare().
*/
if (spa->spa_spares.sav_config) {
nvlist_free(spa->spa_spares.sav_config);
spa->spa_spares.sav_config = NULL;
spa_load_spares(spa);
}
if (spa->spa_l2cache.sav_config) {
nvlist_free(spa->spa_l2cache.sav_config);
spa->spa_l2cache.sav_config = NULL;
spa_load_l2cache(spa);
}
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
spa_config_exit(spa, SCL_ALL, FTAG);
if (props != NULL)
spa_configfile_set(spa, props, B_FALSE);
if (error != 0 || (props && spa_writeable(spa) &&
(error = spa_prop_set(spa, props)))) {
spa_unload(spa);
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (error);
}
spa_async_resume(spa);
/*
* Override any spares and level 2 cache devices as specified by
* the user, as these may have correct device names/devids, etc.
*/
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares) == 0) {
if (spa->spa_spares.sav_config)
fnvlist_remove(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES);
else
spa->spa_spares.sav_config = fnvlist_alloc();
fnvlist_add_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, (const nvlist_t * const *)spares,
nspares);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_spares(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
spa->spa_spares.sav_sync = B_TRUE;
}
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache) == 0) {
if (spa->spa_l2cache.sav_config)
fnvlist_remove(spa->spa_l2cache.sav_config,
ZPOOL_CONFIG_L2CACHE);
else
spa->spa_l2cache.sav_config = fnvlist_alloc();
fnvlist_add_nvlist_array(spa->spa_l2cache.sav_config,
ZPOOL_CONFIG_L2CACHE, (const nvlist_t * const *)l2cache,
nl2cache);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_l2cache(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
spa->spa_l2cache.sav_sync = B_TRUE;
}
/*
* Check for any removed devices.
*/
if (spa->spa_autoreplace) {
spa_aux_check_removed(&spa->spa_spares);
spa_aux_check_removed(&spa->spa_l2cache);
}
if (spa_writeable(spa)) {
/*
* Update the config cache to include the newly-imported pool.
*/
spa_config_update(spa, SPA_CONFIG_UPDATE_POOL);
}
/*
* It's possible that the pool was expanded while it was exported.
* We kick off an async task to handle this for us.
*/
spa_async_request(spa, SPA_ASYNC_AUTOEXPAND);
spa_history_log_version(spa, "import", NULL);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_IMPORT);
mutex_exit(&spa_namespace_lock);
zvol_create_minors_recursive(pool);
spa_import_os(spa);
return (0);
}
nvlist_t *
spa_tryimport(nvlist_t *tryconfig)
{
nvlist_t *config = NULL;
const char *poolname, *cachefile;
spa_t *spa;
uint64_t state;
int error;
zpool_load_policy_t policy;
if (nvlist_lookup_string(tryconfig, ZPOOL_CONFIG_POOL_NAME, &poolname))
return (NULL);
if (nvlist_lookup_uint64(tryconfig, ZPOOL_CONFIG_POOL_STATE, &state))
return (NULL);
/*
* Create and initialize the spa structure.
*/
mutex_enter(&spa_namespace_lock);
spa = spa_add(TRYIMPORT_NAME, tryconfig, NULL);
spa_activate(spa, SPA_MODE_READ);
/*
* Rewind pool if a max txg was provided.
*/
zpool_get_load_policy(spa->spa_config, &policy);
if (policy.zlp_txg != UINT64_MAX) {
spa->spa_load_max_txg = policy.zlp_txg;
spa->spa_extreme_rewind = B_TRUE;
zfs_dbgmsg("spa_tryimport: importing %s, max_txg=%lld",
poolname, (longlong_t)policy.zlp_txg);
} else {
zfs_dbgmsg("spa_tryimport: importing %s", poolname);
}
if (nvlist_lookup_string(tryconfig, ZPOOL_CONFIG_CACHEFILE, &cachefile)
== 0) {
zfs_dbgmsg("spa_tryimport: using cachefile '%s'", cachefile);
spa->spa_config_source = SPA_CONFIG_SRC_CACHEFILE;
} else {
spa->spa_config_source = SPA_CONFIG_SRC_SCAN;
}
/*
* spa_import() relies on a pool config fetched by spa_try_import()
* for spare/cache devices. Import flags are not passed to
* spa_tryimport(), which makes it return early due to a missing log
* device and missing retrieving the cache device and spare eventually.
* Passing ZFS_IMPORT_MISSING_LOG to spa_tryimport() makes it fetch
* the correct configuration regardless of the missing log device.
*/
spa->spa_import_flags |= ZFS_IMPORT_MISSING_LOG;
error = spa_load(spa, SPA_LOAD_TRYIMPORT, SPA_IMPORT_EXISTING);
/*
* If 'tryconfig' was at least parsable, return the current config.
*/
if (spa->spa_root_vdev != NULL) {
config = spa_config_generate(spa, NULL, -1ULL, B_TRUE);
fnvlist_add_string(config, ZPOOL_CONFIG_POOL_NAME, poolname);
fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_STATE, state);
fnvlist_add_uint64(config, ZPOOL_CONFIG_TIMESTAMP,
spa->spa_uberblock.ub_timestamp);
fnvlist_add_nvlist(config, ZPOOL_CONFIG_LOAD_INFO,
spa->spa_load_info);
fnvlist_add_uint64(config, ZPOOL_CONFIG_ERRATA,
spa->spa_errata);
/*
* If the bootfs property exists on this pool then we
* copy it out so that external consumers can tell which
* pools are bootable.
*/
if ((!error || error == EEXIST) && spa->spa_bootfs) {
char *tmpname = kmem_alloc(MAXPATHLEN, KM_SLEEP);
/*
* We have to play games with the name since the
* pool was opened as TRYIMPORT_NAME.
*/
if (dsl_dsobj_to_dsname(spa_name(spa),
spa->spa_bootfs, tmpname) == 0) {
char *cp;
char *dsname;
dsname = kmem_alloc(MAXPATHLEN, KM_SLEEP);
cp = strchr(tmpname, '/');
if (cp == NULL) {
(void) strlcpy(dsname, tmpname,
MAXPATHLEN);
} else {
(void) snprintf(dsname, MAXPATHLEN,
"%s/%s", poolname, ++cp);
}
fnvlist_add_string(config, ZPOOL_CONFIG_BOOTFS,
dsname);
kmem_free(dsname, MAXPATHLEN);
}
kmem_free(tmpname, MAXPATHLEN);
}
/*
* Add the list of hot spares and level 2 cache devices.
*/
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
spa_add_spares(spa, config);
spa_add_l2cache(spa, config);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
spa_unload(spa);
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (config);
}
/*
* Pool export/destroy
*
* The act of destroying or exporting a pool is very simple. We make sure there
* is no more pending I/O and any references to the pool are gone. Then, we
* update the pool state and sync all the labels to disk, removing the
* configuration from the cache afterwards. If the 'hardforce' flag is set, then
* we don't sync the labels or remove the configuration cache.
*/
static int
spa_export_common(const char *pool, int new_state, nvlist_t **oldconfig,
boolean_t force, boolean_t hardforce)
{
int error;
spa_t *spa;
if (oldconfig)
*oldconfig = NULL;
if (!(spa_mode_global & SPA_MODE_WRITE))
return (SET_ERROR(EROFS));
mutex_enter(&spa_namespace_lock);
if ((spa = spa_lookup(pool)) == NULL) {
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(ENOENT));
}
if (spa->spa_is_exporting) {
/* the pool is being exported by another thread */
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(ZFS_ERR_EXPORT_IN_PROGRESS));
}
spa->spa_is_exporting = B_TRUE;
/*
* Put a hold on the pool, drop the namespace lock, stop async tasks,
* reacquire the namespace lock, and see if we can export.
*/
spa_open_ref(spa, FTAG);
mutex_exit(&spa_namespace_lock);
spa_async_suspend(spa);
if (spa->spa_zvol_taskq) {
zvol_remove_minors(spa, spa_name(spa), B_TRUE);
taskq_wait(spa->spa_zvol_taskq);
}
mutex_enter(&spa_namespace_lock);
spa_close(spa, FTAG);
if (spa->spa_state == POOL_STATE_UNINITIALIZED)
goto export_spa;
/*
* The pool will be in core if it's openable, in which case we can
* modify its state. Objsets may be open only because they're dirty,
* so we have to force it to sync before checking spa_refcnt.
*/
if (spa->spa_sync_on) {
txg_wait_synced(spa->spa_dsl_pool, 0);
spa_evicting_os_wait(spa);
}
/*
* A pool cannot be exported or destroyed if there are active
* references. If we are resetting a pool, allow references by
* fault injection handlers.
*/
if (!spa_refcount_zero(spa) || (spa->spa_inject_ref != 0)) {
error = SET_ERROR(EBUSY);
goto fail;
}
if (spa->spa_sync_on) {
vdev_t *rvd = spa->spa_root_vdev;
/*
* A pool cannot be exported if it has an active shared spare.
* This is to prevent other pools stealing the active spare
* from an exported pool. At user's own will, such pool can
* be forcedly exported.
*/
if (!force && new_state == POOL_STATE_EXPORTED &&
spa_has_active_shared_spare(spa)) {
error = SET_ERROR(EXDEV);
goto fail;
}
/*
* We're about to export or destroy this pool. Make sure
* we stop all initialization and trim activity here before
* we set the spa_final_txg. This will ensure that all
* dirty data resulting from the initialization is
* committed to disk before we unload the pool.
*/
vdev_initialize_stop_all(rvd, VDEV_INITIALIZE_ACTIVE);
vdev_trim_stop_all(rvd, VDEV_TRIM_ACTIVE);
vdev_autotrim_stop_all(spa);
vdev_rebuild_stop_all(spa);
/*
* We want this to be reflected on every label,
* so mark them all dirty. spa_unload() will do the
* final sync that pushes these changes out.
*/
if (new_state != POOL_STATE_UNINITIALIZED && !hardforce) {
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa->spa_state = new_state;
vdev_config_dirty(rvd);
spa_config_exit(spa, SCL_ALL, FTAG);
}
/*
* If the log space map feature is enabled and the pool is
* getting exported (but not destroyed), we want to spend some
* time flushing as many metaslabs as we can in an attempt to
* destroy log space maps and save import time. This has to be
* done before we set the spa_final_txg, otherwise
* spa_sync() -> spa_flush_metaslabs() may dirty the final TXGs.
* spa_should_flush_logs_on_unload() should be called after
* spa_state has been set to the new_state.
*/
if (spa_should_flush_logs_on_unload(spa))
spa_unload_log_sm_flush_all(spa);
if (new_state != POOL_STATE_UNINITIALIZED && !hardforce) {
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa->spa_final_txg = spa_last_synced_txg(spa) +
TXG_DEFER_SIZE + 1;
spa_config_exit(spa, SCL_ALL, FTAG);
}
}
export_spa:
spa_export_os(spa);
if (new_state == POOL_STATE_DESTROYED)
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_DESTROY);
else if (new_state == POOL_STATE_EXPORTED)
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_EXPORT);
if (spa->spa_state != POOL_STATE_UNINITIALIZED) {
spa_unload(spa);
spa_deactivate(spa);
}
if (oldconfig && spa->spa_config)
*oldconfig = fnvlist_dup(spa->spa_config);
if (new_state != POOL_STATE_UNINITIALIZED) {
if (!hardforce)
spa_write_cachefile(spa, B_TRUE, B_TRUE, B_FALSE);
spa_remove(spa);
} else {
/*
* If spa_remove() is not called for this spa_t and
* there is any possibility that it can be reused,
* we make sure to reset the exporting flag.
*/
spa->spa_is_exporting = B_FALSE;
}
mutex_exit(&spa_namespace_lock);
return (0);
fail:
spa->spa_is_exporting = B_FALSE;
spa_async_resume(spa);
mutex_exit(&spa_namespace_lock);
return (error);
}
/*
* Destroy a storage pool.
*/
int
spa_destroy(const char *pool)
{
return (spa_export_common(pool, POOL_STATE_DESTROYED, NULL,
B_FALSE, B_FALSE));
}
/*
* Export a storage pool.
*/
int
spa_export(const char *pool, nvlist_t **oldconfig, boolean_t force,
boolean_t hardforce)
{
return (spa_export_common(pool, POOL_STATE_EXPORTED, oldconfig,
force, hardforce));
}
/*
* Similar to spa_export(), this unloads the spa_t without actually removing it
* from the namespace in any way.
*/
int
spa_reset(const char *pool)
{
return (spa_export_common(pool, POOL_STATE_UNINITIALIZED, NULL,
B_FALSE, B_FALSE));
}
/*
* ==========================================================================
* Device manipulation
* ==========================================================================
*/
/*
* This is called as a synctask to increment the draid feature flag
*/
static void
spa_draid_feature_incr(void *arg, dmu_tx_t *tx)
{
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
int draid = (int)(uintptr_t)arg;
for (int c = 0; c < draid; c++)
spa_feature_incr(spa, SPA_FEATURE_DRAID, tx);
}
/*
* Add a device to a storage pool.
*/
int
spa_vdev_add(spa_t *spa, nvlist_t *nvroot)
{
uint64_t txg, ndraid = 0;
int error;
vdev_t *rvd = spa->spa_root_vdev;
vdev_t *vd, *tvd;
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
ASSERT(spa_writeable(spa));
txg = spa_vdev_enter(spa);
if ((error = spa_config_parse(spa, &vd, nvroot, NULL, 0,
VDEV_ALLOC_ADD)) != 0)
return (spa_vdev_exit(spa, NULL, txg, error));
spa->spa_pending_vdev = vd; /* spa_vdev_exit() will clear this */
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares,
&nspares) != 0)
nspares = 0;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache,
&nl2cache) != 0)
nl2cache = 0;
if (vd->vdev_children == 0 && nspares == 0 && nl2cache == 0)
return (spa_vdev_exit(spa, vd, txg, EINVAL));
if (vd->vdev_children != 0 &&
(error = vdev_create(vd, txg, B_FALSE)) != 0) {
return (spa_vdev_exit(spa, vd, txg, error));
}
/*
* The virtual dRAID spares must be added after vdev tree is created
* and the vdev guids are generated. The guid of their associated
* dRAID is stored in the config and used when opening the spare.
*/
if ((error = vdev_draid_spare_create(nvroot, vd, &ndraid,
rvd->vdev_children)) == 0) {
if (ndraid > 0 && nvlist_lookup_nvlist_array(nvroot,
ZPOOL_CONFIG_SPARES, &spares, &nspares) != 0)
nspares = 0;
} else {
return (spa_vdev_exit(spa, vd, txg, error));
}
/*
* We must validate the spares and l2cache devices after checking the
* children. Otherwise, vdev_inuse() will blindly overwrite the spare.
*/
if ((error = spa_validate_aux(spa, nvroot, txg, VDEV_ALLOC_ADD)) != 0)
return (spa_vdev_exit(spa, vd, txg, error));
/*
* If we are in the middle of a device removal, we can only add
* devices which match the existing devices in the pool.
* If we are in the middle of a removal, or have some indirect
* vdevs, we can not add raidz or dRAID top levels.
*/
if (spa->spa_vdev_removal != NULL ||
spa->spa_removing_phys.sr_prev_indirect_vdev != -1) {
for (int c = 0; c < vd->vdev_children; c++) {
tvd = vd->vdev_child[c];
if (spa->spa_vdev_removal != NULL &&
tvd->vdev_ashift != spa->spa_max_ashift) {
return (spa_vdev_exit(spa, vd, txg, EINVAL));
}
/* Fail if top level vdev is raidz or a dRAID */
if (vdev_get_nparity(tvd) != 0)
return (spa_vdev_exit(spa, vd, txg, EINVAL));
/*
* Need the top level mirror to be
* a mirror of leaf vdevs only
*/
if (tvd->vdev_ops == &vdev_mirror_ops) {
for (uint64_t cid = 0;
cid < tvd->vdev_children; cid++) {
vdev_t *cvd = tvd->vdev_child[cid];
if (!cvd->vdev_ops->vdev_op_leaf) {
return (spa_vdev_exit(spa, vd,
txg, EINVAL));
}
}
}
}
}
for (int c = 0; c < vd->vdev_children; c++) {
tvd = vd->vdev_child[c];
vdev_remove_child(vd, tvd);
tvd->vdev_id = rvd->vdev_children;
vdev_add_child(rvd, tvd);
vdev_config_dirty(tvd);
}
if (nspares != 0) {
spa_set_aux_vdevs(&spa->spa_spares, spares, nspares,
ZPOOL_CONFIG_SPARES);
spa_load_spares(spa);
spa->spa_spares.sav_sync = B_TRUE;
}
if (nl2cache != 0) {
spa_set_aux_vdevs(&spa->spa_l2cache, l2cache, nl2cache,
ZPOOL_CONFIG_L2CACHE);
spa_load_l2cache(spa);
spa->spa_l2cache.sav_sync = B_TRUE;
}
/*
* We can't increment a feature while holding spa_vdev so we
* have to do it in a synctask.
*/
if (ndraid != 0) {
dmu_tx_t *tx;
tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg);
dsl_sync_task_nowait(spa->spa_dsl_pool, spa_draid_feature_incr,
(void *)(uintptr_t)ndraid, tx);
dmu_tx_commit(tx);
}
/*
* We have to be careful when adding new vdevs to an existing pool.
* If other threads start allocating from these vdevs before we
* sync the config cache, and we lose power, then upon reboot we may
* fail to open the pool because there are DVAs that the config cache
* can't translate. Therefore, we first add the vdevs without
* initializing metaslabs; sync the config cache (via spa_vdev_exit());
* and then let spa_config_update() initialize the new metaslabs.
*
* spa_load() checks for added-but-not-initialized vdevs, so that
* if we lose power at any point in this sequence, the remaining
* steps will be completed the next time we load the pool.
*/
(void) spa_vdev_exit(spa, vd, txg, 0);
mutex_enter(&spa_namespace_lock);
spa_config_update(spa, SPA_CONFIG_UPDATE_POOL);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_VDEV_ADD);
mutex_exit(&spa_namespace_lock);
return (0);
}
/*
- * Attach a device to a mirror. The arguments are the path to any device
- * in the mirror, and the nvroot for the new device. If the path specifies
- * a device that is not mirrored, we automatically insert the mirror vdev.
+ * Attach a device to a vdev specified by its guid. The vdev type can be
+ * a mirror, a raidz, or a leaf device that is also a top-level (e.g. a
+ * single device). When the vdev is a single device, a mirror vdev will be
+ * automatically inserted.
*
* If 'replacing' is specified, the new device is intended to replace the
* existing device; in this case the two devices are made into their own
* mirror using the 'replacing' vdev, which is functionally identical to
* the mirror vdev (it actually reuses all the same ops) but has a few
* extra rules: you can't attach to it after it's been created, and upon
* completion of resilvering, the first disk (the one being replaced)
* is automatically detached.
*
* If 'rebuild' is specified, then sequential reconstruction (a.ka. rebuild)
* should be performed instead of traditional healing reconstruction. From
* an administrators perspective these are both resilver operations.
*/
int
spa_vdev_attach(spa_t *spa, uint64_t guid, nvlist_t *nvroot, int replacing,
int rebuild)
{
uint64_t txg, dtl_max_txg;
vdev_t *rvd = spa->spa_root_vdev;
vdev_t *oldvd, *newvd, *newrootvd, *pvd, *tvd;
vdev_ops_t *pvops;
char *oldvdpath, *newvdpath;
- int newvd_isspare;
+ int newvd_isspare = B_FALSE;
int error;
ASSERT(spa_writeable(spa));
txg = spa_vdev_enter(spa);
oldvd = spa_lookup_by_guid(spa, guid, B_FALSE);
ASSERT(MUTEX_HELD(&spa_namespace_lock));
if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) {
error = (spa_has_checkpoint(spa)) ?
ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT;
return (spa_vdev_exit(spa, NULL, txg, error));
}
if (rebuild) {
if (!spa_feature_is_enabled(spa, SPA_FEATURE_DEVICE_REBUILD))
return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
if (dsl_scan_resilvering(spa_get_dsl(spa)) ||
dsl_scan_resilver_scheduled(spa_get_dsl(spa))) {
return (spa_vdev_exit(spa, NULL, txg,
ZFS_ERR_RESILVER_IN_PROGRESS));
}
} else {
if (vdev_rebuild_active(rvd))
return (spa_vdev_exit(spa, NULL, txg,
ZFS_ERR_REBUILD_IN_PROGRESS));
}
- if (spa->spa_vdev_removal != NULL)
- return (spa_vdev_exit(spa, NULL, txg, EBUSY));
+ if (spa->spa_vdev_removal != NULL) {
+ return (spa_vdev_exit(spa, NULL, txg,
+ ZFS_ERR_DEVRM_IN_PROGRESS));
+ }
if (oldvd == NULL)
return (spa_vdev_exit(spa, NULL, txg, ENODEV));
- if (!oldvd->vdev_ops->vdev_op_leaf)
+ boolean_t raidz = oldvd->vdev_ops == &vdev_raidz_ops;
+
+ if (raidz) {
+ if (!spa_feature_is_enabled(spa, SPA_FEATURE_RAIDZ_EXPANSION))
+ return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
+
+ /*
+ * Can't expand a raidz while prior expand is in progress.
+ */
+ if (spa->spa_raidz_expand != NULL) {
+ return (spa_vdev_exit(spa, NULL, txg,
+ ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS));
+ }
+ } else if (!oldvd->vdev_ops->vdev_op_leaf) {
return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
+ }
- pvd = oldvd->vdev_parent;
+ if (raidz)
+ pvd = oldvd;
+ else
+ pvd = oldvd->vdev_parent;
if (spa_config_parse(spa, &newrootvd, nvroot, NULL, 0,
VDEV_ALLOC_ATTACH) != 0)
return (spa_vdev_exit(spa, NULL, txg, EINVAL));
if (newrootvd->vdev_children != 1)
return (spa_vdev_exit(spa, newrootvd, txg, EINVAL));
newvd = newrootvd->vdev_child[0];
if (!newvd->vdev_ops->vdev_op_leaf)
return (spa_vdev_exit(spa, newrootvd, txg, EINVAL));
if ((error = vdev_create(newrootvd, txg, replacing)) != 0)
return (spa_vdev_exit(spa, newrootvd, txg, error));
/*
* log, dedup and special vdevs should not be replaced by spares.
*/
if ((oldvd->vdev_top->vdev_alloc_bias != VDEV_BIAS_NONE ||
oldvd->vdev_top->vdev_islog) && newvd->vdev_isspare) {
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
}
/*
* A dRAID spare can only replace a child of its parent dRAID vdev.
*/
if (newvd->vdev_ops == &vdev_draid_spare_ops &&
oldvd->vdev_top != vdev_draid_spare_get_parent(newvd)) {
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
}
if (rebuild) {
/*
* For rebuilds, the top vdev must support reconstruction
* using only space maps. This means the only allowable
* vdevs types are the root vdev, a mirror, or dRAID.
*/
tvd = pvd;
if (pvd->vdev_top != NULL)
tvd = pvd->vdev_top;
if (tvd->vdev_ops != &vdev_mirror_ops &&
tvd->vdev_ops != &vdev_root_ops &&
tvd->vdev_ops != &vdev_draid_ops) {
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
}
}
if (!replacing) {
/*
* For attach, the only allowable parent is a mirror or the root
* vdev.
*/
if (pvd->vdev_ops != &vdev_mirror_ops &&
+ pvd->vdev_ops != &vdev_raidz_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))
+ vdev_t *min_vdev = raidz ? oldvd->vdev_child[0] : oldvd;
+ if (newvd->vdev_asize < vdev_get_min_asize(min_vdev))
return (spa_vdev_exit(spa, newrootvd, txg, EOVERFLOW));
/*
* The new device cannot have a higher alignment requirement
* than the top-level vdev.
*/
if (newvd->vdev_ashift > oldvd->vdev_top->vdev_ashift)
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
+ /*
+ * RAIDZ-expansion-specific checks.
+ */
+ if (raidz) {
+ if (vdev_raidz_attach_check(newvd) != 0)
+ return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
+
+ /*
+ * Fail early if a child is not healthy or being replaced
+ */
+ for (int i = 0; i < oldvd->vdev_children; i++) {
+ if (vdev_is_dead(oldvd->vdev_child[i]) ||
+ !oldvd->vdev_child[i]->vdev_ops->vdev_op_leaf) {
+ return (spa_vdev_exit(spa, newrootvd, txg,
+ ENXIO));
+ }
+ /* Also fail if reserved boot area is in-use */
+ if (vdev_check_boot_reserve(spa, oldvd->vdev_child[i])
+ != 0) {
+ return (spa_vdev_exit(spa, newrootvd, txg,
+ EADDRINUSE));
+ }
+ }
+ }
+
+ if (raidz) {
+ /*
+ * Note: oldvdpath is freed by spa_strfree(), but
+ * kmem_asprintf() is freed by kmem_strfree(), so we have to
+ * move it to a spa_strdup-ed string.
+ */
+ char *tmp = kmem_asprintf("raidz%u-%u",
+ (uint_t)vdev_get_nparity(oldvd), (uint_t)oldvd->vdev_id);
+ oldvdpath = spa_strdup(tmp);
+ kmem_strfree(tmp);
+ } else {
+ oldvdpath = spa_strdup(oldvd->vdev_path);
+ }
+ newvdpath = spa_strdup(newvd->vdev_path);
+
/*
* 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) {
+ if (strcmp(oldvdpath, newvdpath) == 0) {
spa_strfree(oldvd->vdev_path);
- oldvd->vdev_path = kmem_alloc(strlen(newvd->vdev_path) + 5,
+ oldvd->vdev_path = kmem_alloc(strlen(newvdpath) + 5,
KM_SLEEP);
- (void) snprintf(oldvd->vdev_path, strlen(newvd->vdev_path) + 5,
- "%s/%s", newvd->vdev_path, "old");
+ (void) sprintf(oldvd->vdev_path, "%s/old",
+ newvdpath);
if (oldvd->vdev_devid != NULL) {
spa_strfree(oldvd->vdev_devid);
oldvd->vdev_devid = NULL;
}
+ spa_strfree(oldvdpath);
+ oldvdpath = spa_strdup(oldvd->vdev_path);
}
/*
* 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)
+ if (!raidz && pvd->vdev_ops != pvops) {
pvd = vdev_add_parent(oldvd, pvops);
+ ASSERT(pvd->vdev_ops == pvops);
+ ASSERT(oldvd->vdev_parent == pvd);
+ }
ASSERT(pvd->vdev_top->vdev_parent == rvd);
- ASSERT(pvd->vdev_ops == pvops);
- ASSERT(oldvd->vdev_parent == pvd);
/*
* Extract the new device from its root and add it to pvd.
*/
vdev_remove_child(newrootvd, newvd);
newvd->vdev_id = pvd->vdev_children;
newvd->vdev_crtxg = oldvd->vdev_crtxg;
vdev_add_child(pvd, newvd);
/*
* Reevaluate the parent vdev state.
*/
vdev_propagate_state(pvd);
tvd = newvd->vdev_top;
ASSERT(pvd->vdev_top == tvd);
ASSERT(tvd->vdev_parent == rvd);
vdev_config_dirty(tvd);
/*
* Set newvd's DTL to [TXG_INITIAL, dtl_max_txg) so that we account
* for any dmu_sync-ed blocks. It will propagate upward when
* spa_vdev_exit() calls vdev_dtl_reassess().
*/
dtl_max_txg = txg + TXG_CONCURRENT_STATES;
- vdev_dtl_dirty(newvd, DTL_MISSING,
- TXG_INITIAL, dtl_max_txg - TXG_INITIAL);
+ if (raidz) {
+ /*
+ * 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);
- if (newvd->vdev_isspare) {
- spa_spare_activate(newvd);
- spa_event_notify(spa, newvd, NULL, ESC_ZFS_VDEV_SPARE);
- }
+ vdev_initialize_stop_all(tvd, VDEV_INITIALIZE_ACTIVE);
+ vdev_trim_stop_all(tvd, VDEV_TRIM_ACTIVE);
+ vdev_autotrim_stop_wait(tvd);
- oldvdpath = spa_strdup(oldvd->vdev_path);
- newvdpath = spa_strdup(newvd->vdev_path);
- newvd_isspare = newvd->vdev_isspare;
+ dtl_max_txg = spa_vdev_config_enter(spa);
- /*
- * Mark newvd's DTL dirty in this txg.
- */
- vdev_dirty(tvd, VDD_DTL, newvd, txg);
+ tvd->vdev_rz_expanding = B_TRUE;
- /*
- * Schedule the resilver or rebuild to restart in the future. We do
- * this to ensure that dmu_sync-ed blocks have been stitched into the
- * respective datasets.
- */
- if (rebuild) {
- newvd->vdev_rebuild_txg = txg;
+ vdev_dirty_leaves(tvd, VDD_DTL, dtl_max_txg);
+ vdev_config_dirty(tvd);
- vdev_rebuild(tvd);
+ dmu_tx_t *tx = dmu_tx_create_assigned(spa->spa_dsl_pool,
+ dtl_max_txg);
+ dsl_sync_task_nowait(spa->spa_dsl_pool, vdev_raidz_attach_sync,
+ newvd, tx);
+ dmu_tx_commit(tx);
} else {
- newvd->vdev_resilver_txg = txg;
+ 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);
+ }
+
+ newvd_isspare = newvd->vdev_isspare;
+
+ /*
+ * Mark newvd's DTL dirty in this txg.
+ */
+ vdev_dirty(tvd, VDD_DTL, newvd, txg);
- if (dsl_scan_resilvering(spa_get_dsl(spa)) &&
- spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER)) {
- vdev_defer_resilver(newvd);
+ /*
+ * Schedule the resilver or rebuild to restart in the future.
+ * We do this to ensure that dmu_sync-ed blocks have been
+ * stitched into the respective datasets.
+ */
+ if (rebuild) {
+ newvd->vdev_rebuild_txg = txg;
+
+ vdev_rebuild(tvd);
} else {
- dsl_scan_restart_resilver(spa->spa_dsl_pool,
- dtl_max_txg);
+ newvd->vdev_resilver_txg = txg;
+
+ if (dsl_scan_resilvering(spa_get_dsl(spa)) &&
+ spa_feature_is_enabled(spa,
+ SPA_FEATURE_RESILVER_DEFER)) {
+ vdev_defer_resilver(newvd);
+ } else {
+ dsl_scan_restart_resilver(spa->spa_dsl_pool,
+ dtl_max_txg);
+ }
}
}
if (spa->spa_bootfs)
spa_event_notify(spa, newvd, NULL, ESC_ZFS_BOOTFS_VDEV_ATTACH);
spa_event_notify(spa, newvd, NULL, ESC_ZFS_VDEV_ATTACH);
/*
* Commit the config
*/
(void) spa_vdev_exit(spa, newrootvd, dtl_max_txg, 0);
spa_history_log_internal(spa, "vdev attach", NULL,
"%s vdev=%s %s vdev=%s",
replacing && newvd_isspare ? "spare in" :
replacing ? "replace" : "attach", newvdpath,
replacing ? "for" : "to", oldvdpath);
spa_strfree(oldvdpath);
spa_strfree(newvdpath);
return (0);
}
/*
* Detach a device from a mirror or replacing vdev.
*
* If 'replace_done' is specified, only detach if the parent
* is a replacing or a spare vdev.
*/
int
spa_vdev_detach(spa_t *spa, uint64_t guid, uint64_t pguid, int replace_done)
{
uint64_t txg;
int error;
vdev_t *rvd __maybe_unused = spa->spa_root_vdev;
vdev_t *vd, *pvd, *cvd, *tvd;
boolean_t unspare = B_FALSE;
uint64_t unspare_guid = 0;
char *vdpath;
ASSERT(spa_writeable(spa));
txg = spa_vdev_detach_enter(spa, guid);
vd = spa_lookup_by_guid(spa, guid, B_FALSE);
/*
* Besides being called directly from the userland through the
* ioctl interface, spa_vdev_detach() can be potentially called
* at the end of spa_vdev_resilver_done().
*
* In the regular case, when we have a checkpoint this shouldn't
* happen as we never empty the DTLs of a vdev during the scrub
* [see comment in dsl_scan_done()]. Thus spa_vdev_resilvering_done()
* should never get here when we have a checkpoint.
*
* That said, even in a case when we checkpoint the pool exactly
* as spa_vdev_resilver_done() calls this function everything
* should be fine as the resilver will return right away.
*/
ASSERT(MUTEX_HELD(&spa_namespace_lock));
if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) {
error = (spa_has_checkpoint(spa)) ?
ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT;
return (spa_vdev_exit(spa, NULL, txg, error));
}
if (vd == NULL)
return (spa_vdev_exit(spa, NULL, txg, ENODEV));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
pvd = vd->vdev_parent;
/*
* If the parent/child relationship is not as expected, don't do it.
* Consider M(A,R(B,C)) -- that is, a mirror of A with a replacing
* vdev that's replacing B with C. The user's intent in replacing
* is to go from M(A,B) to M(A,C). If the user decides to cancel
* the replace by detaching C, the expected behavior is to end up
* M(A,B). But suppose that right after deciding to detach C,
* the replacement of B completes. We would have M(A,C), and then
* ask to detach C, which would leave us with just A -- not what
* the user wanted. To prevent this, we make sure that the
* parent/child relationship hasn't changed -- in this example,
* that C's parent is still the replacing vdev R.
*/
if (pvd->vdev_guid != pguid && pguid != 0)
return (spa_vdev_exit(spa, NULL, txg, EBUSY));
/*
* Only 'replacing' or 'spare' vdevs can be replaced.
*/
if (replace_done && pvd->vdev_ops != &vdev_replacing_ops &&
pvd->vdev_ops != &vdev_spare_ops)
return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
ASSERT(pvd->vdev_ops != &vdev_spare_ops ||
spa_version(spa) >= SPA_VERSION_SPARES);
/*
* Only mirror, replacing, and spare vdevs support detach.
*/
if (pvd->vdev_ops != &vdev_replacing_ops &&
pvd->vdev_ops != &vdev_mirror_ops &&
pvd->vdev_ops != &vdev_spare_ops)
return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
/*
* If this device has the only valid copy of some data,
* we cannot safely detach it.
*/
if (vdev_dtl_required(vd))
return (spa_vdev_exit(spa, NULL, txg, EBUSY));
ASSERT(pvd->vdev_children >= 2);
/*
* If we are detaching the second disk from a replacing vdev, then
* check to see if we changed the original vdev's path to have "/old"
* at the end in spa_vdev_attach(). If so, undo that change now.
*/
if (pvd->vdev_ops == &vdev_replacing_ops && vd->vdev_id > 0 &&
vd->vdev_path != NULL) {
size_t len = strlen(vd->vdev_path);
for (int c = 0; c < pvd->vdev_children; c++) {
cvd = pvd->vdev_child[c];
if (cvd == vd || cvd->vdev_path == NULL)
continue;
if (strncmp(cvd->vdev_path, vd->vdev_path, len) == 0 &&
strcmp(cvd->vdev_path + len, "/old") == 0) {
spa_strfree(cvd->vdev_path);
cvd->vdev_path = spa_strdup(vd->vdev_path);
break;
}
}
}
/*
* If we are detaching the original disk from a normal spare, then it
* implies that the spare should become a real disk, and be removed
* from the active spare list for the pool. dRAID spares on the
* other hand are coupled to the pool and thus should never be removed
* from the spares list.
*/
if (pvd->vdev_ops == &vdev_spare_ops && vd->vdev_id == 0) {
vdev_t *last_cvd = pvd->vdev_child[pvd->vdev_children - 1];
if (last_cvd->vdev_isspare &&
last_cvd->vdev_ops != &vdev_draid_spare_ops) {
unspare = B_TRUE;
}
}
/*
* Erase the disk labels so the disk can be used for other things.
* This must be done after all other error cases are handled,
* but before we disembowel vd (so we can still do I/O to it).
* But if we can't do it, don't treat the error as fatal --
* it may be that the unwritability of the disk is the reason
* it's being detached!
*/
(void) vdev_label_init(vd, 0, VDEV_LABEL_REMOVE);
/*
* Remove vd from its parent and compact the parent's children.
*/
vdev_remove_child(pvd, vd);
vdev_compact_children(pvd);
/*
* Remember one of the remaining children so we can get tvd below.
*/
cvd = pvd->vdev_child[pvd->vdev_children - 1];
/*
* If we need to remove the remaining child from the list of hot spares,
* do it now, marking the vdev as no longer a spare in the process.
* We must do this before vdev_remove_parent(), because that can
* change the GUID if it creates a new toplevel GUID. For a similar
* reason, we must remove the spare now, in the same txg as the detach;
* otherwise someone could attach a new sibling, change the GUID, and
* the subsequent attempt to spa_vdev_remove(unspare_guid) would fail.
*/
if (unspare) {
ASSERT(cvd->vdev_isspare);
spa_spare_remove(cvd);
unspare_guid = cvd->vdev_guid;
(void) spa_vdev_remove(spa, unspare_guid, B_TRUE);
cvd->vdev_unspare = B_TRUE;
}
/*
* If the parent mirror/replacing vdev only has one child,
* the parent is no longer needed. Remove it from the tree.
*/
if (pvd->vdev_children == 1) {
if (pvd->vdev_ops == &vdev_spare_ops)
cvd->vdev_unspare = B_FALSE;
vdev_remove_parent(cvd);
}
/*
* We don't set tvd until now because the parent we just removed
* may have been the previous top-level vdev.
*/
tvd = cvd->vdev_top;
ASSERT(tvd->vdev_parent == rvd);
/*
* Reevaluate the parent vdev state.
*/
vdev_propagate_state(cvd);
/*
* If the 'autoexpand' property is set on the pool then automatically
* try to expand the size of the pool. For example if the device we
* just detached was smaller than the others, it may be possible to
* add metaslabs (i.e. grow the pool). We need to reopen the vdev
* first so that we can obtain the updated sizes of the leaf vdevs.
*/
if (spa->spa_autoexpand) {
vdev_reopen(tvd);
vdev_expand(tvd, txg);
}
vdev_config_dirty(tvd);
/*
* Mark vd's DTL as dirty in this txg. vdev_dtl_sync() will see that
* vd->vdev_detached is set and free vd's DTL object in syncing context.
* But first make sure we're not on any *other* txg's DTL list, to
* prevent vd from being accessed after it's freed.
*/
vdpath = spa_strdup(vd->vdev_path ? vd->vdev_path : "none");
for (int t = 0; t < TXG_SIZE; t++)
(void) txg_list_remove_this(&tvd->vdev_dtl_list, vd, t);
vd->vdev_detached = B_TRUE;
vdev_dirty(tvd, VDD_DTL, vd, txg);
spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_REMOVE);
spa_notify_waiters(spa);
/* hang on to the spa before we release the lock */
spa_open_ref(spa, FTAG);
error = spa_vdev_exit(spa, vd, txg, 0);
spa_history_log_internal(spa, "detach", NULL,
"vdev=%s", vdpath);
spa_strfree(vdpath);
/*
* If this was the removal of the original device in a hot spare vdev,
* then we want to go through and remove the device from the hot spare
* list of every other pool.
*/
if (unspare) {
spa_t *altspa = NULL;
mutex_enter(&spa_namespace_lock);
while ((altspa = spa_next(altspa)) != NULL) {
if (altspa->spa_state != POOL_STATE_ACTIVE ||
altspa == spa)
continue;
spa_open_ref(altspa, FTAG);
mutex_exit(&spa_namespace_lock);
(void) spa_vdev_remove(altspa, unspare_guid, B_TRUE);
mutex_enter(&spa_namespace_lock);
spa_close(altspa, FTAG);
}
mutex_exit(&spa_namespace_lock);
/* search the rest of the vdevs for spares to remove */
spa_vdev_resilver_done(spa);
}
/* all done with the spa; OK to release */
mutex_enter(&spa_namespace_lock);
spa_close(spa, FTAG);
mutex_exit(&spa_namespace_lock);
return (error);
}
static int
spa_vdev_initialize_impl(spa_t *spa, uint64_t guid, uint64_t cmd_type,
list_t *vd_list)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
/* Look up vdev and ensure it's a leaf. */
vdev_t *vd = spa_lookup_by_guid(spa, guid, B_FALSE);
if (vd == NULL || vd->vdev_detached) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(ENODEV));
} else if (!vd->vdev_ops->vdev_op_leaf || !vdev_is_concrete(vd)) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EINVAL));
} else if (!vdev_writeable(vd)) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EROFS));
}
mutex_enter(&vd->vdev_initialize_lock);
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
/*
* When we activate an initialize action we check to see
* if the vdev_initialize_thread is NULL. We do this instead
* of using the vdev_initialize_state since there might be
* a previous initialization process which has completed but
* the thread is not exited.
*/
if (cmd_type == POOL_INITIALIZE_START &&
(vd->vdev_initialize_thread != NULL ||
- vd->vdev_top->vdev_removing)) {
+ vd->vdev_top->vdev_removing || vd->vdev_top->vdev_rz_expanding)) {
mutex_exit(&vd->vdev_initialize_lock);
return (SET_ERROR(EBUSY));
} else if (cmd_type == POOL_INITIALIZE_CANCEL &&
(vd->vdev_initialize_state != VDEV_INITIALIZE_ACTIVE &&
vd->vdev_initialize_state != VDEV_INITIALIZE_SUSPENDED)) {
mutex_exit(&vd->vdev_initialize_lock);
return (SET_ERROR(ESRCH));
} else if (cmd_type == POOL_INITIALIZE_SUSPEND &&
vd->vdev_initialize_state != VDEV_INITIALIZE_ACTIVE) {
mutex_exit(&vd->vdev_initialize_lock);
return (SET_ERROR(ESRCH));
} else if (cmd_type == POOL_INITIALIZE_UNINIT &&
vd->vdev_initialize_thread != NULL) {
mutex_exit(&vd->vdev_initialize_lock);
return (SET_ERROR(EBUSY));
}
switch (cmd_type) {
case POOL_INITIALIZE_START:
vdev_initialize(vd);
break;
case POOL_INITIALIZE_CANCEL:
vdev_initialize_stop(vd, VDEV_INITIALIZE_CANCELED, vd_list);
break;
case POOL_INITIALIZE_SUSPEND:
vdev_initialize_stop(vd, VDEV_INITIALIZE_SUSPENDED, vd_list);
break;
case POOL_INITIALIZE_UNINIT:
vdev_uninitialize(vd);
break;
default:
panic("invalid cmd_type %llu", (unsigned long long)cmd_type);
}
mutex_exit(&vd->vdev_initialize_lock);
return (0);
}
int
spa_vdev_initialize(spa_t *spa, nvlist_t *nv, uint64_t cmd_type,
nvlist_t *vdev_errlist)
{
int total_errors = 0;
list_t vd_list;
list_create(&vd_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_initialize_node));
/*
* We hold the namespace lock through the whole function
* to prevent any changes to the pool while we're starting or
* stopping initialization. The config and state locks are held so that
* we can properly assess the vdev state before we commit to
* the initializing operation.
*/
mutex_enter(&spa_namespace_lock);
for (nvpair_t *pair = nvlist_next_nvpair(nv, NULL);
pair != NULL; pair = nvlist_next_nvpair(nv, pair)) {
uint64_t vdev_guid = fnvpair_value_uint64(pair);
int error = spa_vdev_initialize_impl(spa, vdev_guid, cmd_type,
&vd_list);
if (error != 0) {
char guid_as_str[MAXNAMELEN];
(void) snprintf(guid_as_str, sizeof (guid_as_str),
"%llu", (unsigned long long)vdev_guid);
fnvlist_add_int64(vdev_errlist, guid_as_str, error);
total_errors++;
}
}
/* Wait for all initialize threads to stop. */
vdev_initialize_stop_wait(spa, &vd_list);
/* Sync out the initializing state */
txg_wait_synced(spa->spa_dsl_pool, 0);
mutex_exit(&spa_namespace_lock);
list_destroy(&vd_list);
return (total_errors);
}
static int
spa_vdev_trim_impl(spa_t *spa, uint64_t guid, uint64_t cmd_type,
uint64_t rate, boolean_t partial, boolean_t secure, list_t *vd_list)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
/* Look up vdev and ensure it's a leaf. */
vdev_t *vd = spa_lookup_by_guid(spa, guid, B_FALSE);
if (vd == NULL || vd->vdev_detached) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(ENODEV));
} else if (!vd->vdev_ops->vdev_op_leaf || !vdev_is_concrete(vd)) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EINVAL));
} else if (!vdev_writeable(vd)) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EROFS));
} else if (!vd->vdev_has_trim) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EOPNOTSUPP));
} else if (secure && !vd->vdev_has_securetrim) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EOPNOTSUPP));
}
mutex_enter(&vd->vdev_trim_lock);
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
/*
* When we activate a TRIM action we check to see if the
* vdev_trim_thread is NULL. We do this instead of using the
* vdev_trim_state since there might be a previous TRIM process
* which has completed but the thread is not exited.
*/
if (cmd_type == POOL_TRIM_START &&
- (vd->vdev_trim_thread != NULL || vd->vdev_top->vdev_removing)) {
+ (vd->vdev_trim_thread != NULL || vd->vdev_top->vdev_removing ||
+ vd->vdev_top->vdev_rz_expanding)) {
mutex_exit(&vd->vdev_trim_lock);
return (SET_ERROR(EBUSY));
} else if (cmd_type == POOL_TRIM_CANCEL &&
(vd->vdev_trim_state != VDEV_TRIM_ACTIVE &&
vd->vdev_trim_state != VDEV_TRIM_SUSPENDED)) {
mutex_exit(&vd->vdev_trim_lock);
return (SET_ERROR(ESRCH));
} else if (cmd_type == POOL_TRIM_SUSPEND &&
vd->vdev_trim_state != VDEV_TRIM_ACTIVE) {
mutex_exit(&vd->vdev_trim_lock);
return (SET_ERROR(ESRCH));
}
switch (cmd_type) {
case POOL_TRIM_START:
vdev_trim(vd, rate, partial, secure);
break;
case POOL_TRIM_CANCEL:
vdev_trim_stop(vd, VDEV_TRIM_CANCELED, vd_list);
break;
case POOL_TRIM_SUSPEND:
vdev_trim_stop(vd, VDEV_TRIM_SUSPENDED, vd_list);
break;
default:
panic("invalid cmd_type %llu", (unsigned long long)cmd_type);
}
mutex_exit(&vd->vdev_trim_lock);
return (0);
}
/*
* Initiates a manual TRIM for the requested vdevs. This kicks off individual
* TRIM threads for each child vdev. These threads pass over all of the free
* space in the vdev's metaslabs and issues TRIM commands for that space.
*/
int
spa_vdev_trim(spa_t *spa, nvlist_t *nv, uint64_t cmd_type, uint64_t rate,
boolean_t partial, boolean_t secure, nvlist_t *vdev_errlist)
{
int total_errors = 0;
list_t vd_list;
list_create(&vd_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_trim_node));
/*
* We hold the namespace lock through the whole function
* to prevent any changes to the pool while we're starting or
* stopping TRIM. The config and state locks are held so that
* we can properly assess the vdev state before we commit to
* the TRIM operation.
*/
mutex_enter(&spa_namespace_lock);
for (nvpair_t *pair = nvlist_next_nvpair(nv, NULL);
pair != NULL; pair = nvlist_next_nvpair(nv, pair)) {
uint64_t vdev_guid = fnvpair_value_uint64(pair);
int error = spa_vdev_trim_impl(spa, vdev_guid, cmd_type,
rate, partial, secure, &vd_list);
if (error != 0) {
char guid_as_str[MAXNAMELEN];
(void) snprintf(guid_as_str, sizeof (guid_as_str),
"%llu", (unsigned long long)vdev_guid);
fnvlist_add_int64(vdev_errlist, guid_as_str, error);
total_errors++;
}
}
/* Wait for all TRIM threads to stop. */
vdev_trim_stop_wait(spa, &vd_list);
/* Sync out the TRIM state */
txg_wait_synced(spa->spa_dsl_pool, 0);
mutex_exit(&spa_namespace_lock);
list_destroy(&vd_list);
return (total_errors);
}
/*
* Split a set of devices from their mirrors, and create a new pool from them.
*/
int
spa_vdev_split_mirror(spa_t *spa, const char *newname, nvlist_t *config,
nvlist_t *props, boolean_t exp)
{
int error = 0;
uint64_t txg, *glist;
spa_t *newspa;
uint_t c, children, lastlog;
nvlist_t **child, *nvl, *tmp;
dmu_tx_t *tx;
const char *altroot = NULL;
vdev_t *rvd, **vml = NULL; /* vdev modify list */
boolean_t activate_slog;
ASSERT(spa_writeable(spa));
txg = spa_vdev_enter(spa);
ASSERT(MUTEX_HELD(&spa_namespace_lock));
if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) {
error = (spa_has_checkpoint(spa)) ?
ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT;
return (spa_vdev_exit(spa, NULL, txg, error));
}
/* clear the log and flush everything up to now */
activate_slog = spa_passivate_log(spa);
(void) spa_vdev_config_exit(spa, NULL, txg, 0, FTAG);
error = spa_reset_logs(spa);
txg = spa_vdev_config_enter(spa);
if (activate_slog)
spa_activate_log(spa);
if (error != 0)
return (spa_vdev_exit(spa, NULL, txg, error));
/* check new spa name before going any further */
if (spa_lookup(newname) != NULL)
return (spa_vdev_exit(spa, NULL, txg, EEXIST));
/*
* scan through all the children to ensure they're all mirrors
*/
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvl) != 0 ||
nvlist_lookup_nvlist_array(nvl, ZPOOL_CONFIG_CHILDREN, &child,
&children) != 0)
return (spa_vdev_exit(spa, NULL, txg, EINVAL));
/* first, check to ensure we've got the right child count */
rvd = spa->spa_root_vdev;
lastlog = 0;
for (c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
/* don't count the holes & logs as children */
if (vd->vdev_islog || (vd->vdev_ops != &vdev_indirect_ops &&
!vdev_is_concrete(vd))) {
if (lastlog == 0)
lastlog = c;
continue;
}
lastlog = 0;
}
if (children != (lastlog != 0 ? lastlog : rvd->vdev_children))
return (spa_vdev_exit(spa, NULL, txg, EINVAL));
/* next, ensure no spare or cache devices are part of the split */
if (nvlist_lookup_nvlist(nvl, ZPOOL_CONFIG_SPARES, &tmp) == 0 ||
nvlist_lookup_nvlist(nvl, ZPOOL_CONFIG_L2CACHE, &tmp) == 0)
return (spa_vdev_exit(spa, NULL, txg, EINVAL));
vml = kmem_zalloc(children * sizeof (vdev_t *), KM_SLEEP);
glist = kmem_zalloc(children * sizeof (uint64_t), KM_SLEEP);
/* then, loop over each vdev and validate it */
for (c = 0; c < children; c++) {
uint64_t is_hole = 0;
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE,
&is_hole);
if (is_hole != 0) {
if (spa->spa_root_vdev->vdev_child[c]->vdev_ishole ||
spa->spa_root_vdev->vdev_child[c]->vdev_islog) {
continue;
} else {
error = SET_ERROR(EINVAL);
break;
}
}
/* deal with indirect vdevs */
if (spa->spa_root_vdev->vdev_child[c]->vdev_ops ==
&vdev_indirect_ops)
continue;
/* which disk is going to be split? */
if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_GUID,
&glist[c]) != 0) {
error = SET_ERROR(EINVAL);
break;
}
/* look it up in the spa */
vml[c] = spa_lookup_by_guid(spa, glist[c], B_FALSE);
if (vml[c] == NULL) {
error = SET_ERROR(ENODEV);
break;
}
/* make sure there's nothing stopping the split */
if (vml[c]->vdev_parent->vdev_ops != &vdev_mirror_ops ||
vml[c]->vdev_islog ||
!vdev_is_concrete(vml[c]) ||
vml[c]->vdev_isspare ||
vml[c]->vdev_isl2cache ||
!vdev_writeable(vml[c]) ||
vml[c]->vdev_children != 0 ||
vml[c]->vdev_state != VDEV_STATE_HEALTHY ||
c != spa->spa_root_vdev->vdev_child[c]->vdev_id) {
error = SET_ERROR(EINVAL);
break;
}
if (vdev_dtl_required(vml[c]) ||
vdev_resilver_needed(vml[c], NULL, NULL)) {
error = SET_ERROR(EBUSY);
break;
}
/* we need certain info from the top level */
fnvlist_add_uint64(child[c], ZPOOL_CONFIG_METASLAB_ARRAY,
vml[c]->vdev_top->vdev_ms_array);
fnvlist_add_uint64(child[c], ZPOOL_CONFIG_METASLAB_SHIFT,
vml[c]->vdev_top->vdev_ms_shift);
fnvlist_add_uint64(child[c], ZPOOL_CONFIG_ASIZE,
vml[c]->vdev_top->vdev_asize);
fnvlist_add_uint64(child[c], ZPOOL_CONFIG_ASHIFT,
vml[c]->vdev_top->vdev_ashift);
/* transfer per-vdev ZAPs */
ASSERT3U(vml[c]->vdev_leaf_zap, !=, 0);
VERIFY0(nvlist_add_uint64(child[c],
ZPOOL_CONFIG_VDEV_LEAF_ZAP, vml[c]->vdev_leaf_zap));
ASSERT3U(vml[c]->vdev_top->vdev_top_zap, !=, 0);
VERIFY0(nvlist_add_uint64(child[c],
ZPOOL_CONFIG_VDEV_TOP_ZAP,
vml[c]->vdev_parent->vdev_top_zap));
}
if (error != 0) {
kmem_free(vml, children * sizeof (vdev_t *));
kmem_free(glist, children * sizeof (uint64_t));
return (spa_vdev_exit(spa, NULL, txg, error));
}
/* stop writers from using the disks */
for (c = 0; c < children; c++) {
if (vml[c] != NULL)
vml[c]->vdev_offline = B_TRUE;
}
vdev_reopen(spa->spa_root_vdev);
/*
* Temporarily record the splitting vdevs in the spa config. This
* will disappear once the config is regenerated.
*/
nvl = fnvlist_alloc();
fnvlist_add_uint64_array(nvl, ZPOOL_CONFIG_SPLIT_LIST, glist, children);
kmem_free(glist, children * sizeof (uint64_t));
mutex_enter(&spa->spa_props_lock);
fnvlist_add_nvlist(spa->spa_config, ZPOOL_CONFIG_SPLIT, nvl);
mutex_exit(&spa->spa_props_lock);
spa->spa_config_splitting = nvl;
vdev_config_dirty(spa->spa_root_vdev);
/* configure and create the new pool */
fnvlist_add_string(config, ZPOOL_CONFIG_POOL_NAME, newname);
fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_STATE,
exp ? POOL_STATE_EXPORTED : POOL_STATE_ACTIVE);
fnvlist_add_uint64(config, ZPOOL_CONFIG_VERSION, spa_version(spa));
fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_TXG, spa->spa_config_txg);
fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_GUID,
spa_generate_guid(NULL));
VERIFY0(nvlist_add_boolean(config, ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS));
(void) nvlist_lookup_string(props,
zpool_prop_to_name(ZPOOL_PROP_ALTROOT), &altroot);
/* add the new pool to the namespace */
newspa = spa_add(newname, config, altroot);
newspa->spa_avz_action = AVZ_ACTION_REBUILD;
newspa->spa_config_txg = spa->spa_config_txg;
spa_set_log_state(newspa, SPA_LOG_CLEAR);
/* release the spa config lock, retaining the namespace lock */
spa_vdev_config_exit(spa, NULL, txg, 0, FTAG);
if (zio_injection_enabled)
zio_handle_panic_injection(spa, FTAG, 1);
spa_activate(newspa, spa_mode_global);
spa_async_suspend(newspa);
/*
* Temporarily stop the initializing and TRIM activity. We set the
* state to ACTIVE so that we know to resume initializing or TRIM
* once the split has completed.
*/
list_t vd_initialize_list;
list_create(&vd_initialize_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_initialize_node));
list_t vd_trim_list;
list_create(&vd_trim_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_trim_node));
for (c = 0; c < children; c++) {
if (vml[c] != NULL && vml[c]->vdev_ops != &vdev_indirect_ops) {
mutex_enter(&vml[c]->vdev_initialize_lock);
vdev_initialize_stop(vml[c],
VDEV_INITIALIZE_ACTIVE, &vd_initialize_list);
mutex_exit(&vml[c]->vdev_initialize_lock);
mutex_enter(&vml[c]->vdev_trim_lock);
vdev_trim_stop(vml[c], VDEV_TRIM_ACTIVE, &vd_trim_list);
mutex_exit(&vml[c]->vdev_trim_lock);
}
}
vdev_initialize_stop_wait(spa, &vd_initialize_list);
vdev_trim_stop_wait(spa, &vd_trim_list);
list_destroy(&vd_initialize_list);
list_destroy(&vd_trim_list);
newspa->spa_config_source = SPA_CONFIG_SRC_SPLIT;
newspa->spa_is_splitting = B_TRUE;
/* create the new pool from the disks of the original pool */
error = spa_load(newspa, SPA_LOAD_IMPORT, SPA_IMPORT_ASSEMBLE);
if (error)
goto out;
/* if that worked, generate a real config for the new pool */
if (newspa->spa_root_vdev != NULL) {
newspa->spa_config_splitting = fnvlist_alloc();
fnvlist_add_uint64(newspa->spa_config_splitting,
ZPOOL_CONFIG_SPLIT_GUID, spa_guid(spa));
spa_config_set(newspa, spa_config_generate(newspa, NULL, -1ULL,
B_TRUE));
}
/* set the props */
if (props != NULL) {
spa_configfile_set(newspa, props, B_FALSE);
error = spa_prop_set(newspa, props);
if (error)
goto out;
}
/* flush everything */
txg = spa_vdev_config_enter(newspa);
vdev_config_dirty(newspa->spa_root_vdev);
(void) spa_vdev_config_exit(newspa, NULL, txg, 0, FTAG);
if (zio_injection_enabled)
zio_handle_panic_injection(spa, FTAG, 2);
spa_async_resume(newspa);
/* finally, update the original pool's config */
txg = spa_vdev_config_enter(spa);
tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error != 0)
dmu_tx_abort(tx);
for (c = 0; c < children; c++) {
if (vml[c] != NULL && vml[c]->vdev_ops != &vdev_indirect_ops) {
vdev_t *tvd = vml[c]->vdev_top;
/*
* Need to be sure the detachable VDEV is not
* on any *other* txg's DTL list to prevent it
* from being accessed after it's freed.
*/
for (int t = 0; t < TXG_SIZE; t++) {
(void) txg_list_remove_this(
&tvd->vdev_dtl_list, vml[c], t);
}
vdev_split(vml[c]);
if (error == 0)
spa_history_log_internal(spa, "detach", tx,
"vdev=%s", vml[c]->vdev_path);
vdev_free(vml[c]);
}
}
spa->spa_avz_action = AVZ_ACTION_REBUILD;
vdev_config_dirty(spa->spa_root_vdev);
spa->spa_config_splitting = NULL;
nvlist_free(nvl);
if (error == 0)
dmu_tx_commit(tx);
(void) spa_vdev_exit(spa, NULL, txg, 0);
if (zio_injection_enabled)
zio_handle_panic_injection(spa, FTAG, 3);
/* split is complete; log a history record */
spa_history_log_internal(newspa, "split", NULL,
"from pool %s", spa_name(spa));
newspa->spa_is_splitting = B_FALSE;
kmem_free(vml, children * sizeof (vdev_t *));
/* if we're not going to mount the filesystems in userland, export */
if (exp)
error = spa_export_common(newname, POOL_STATE_EXPORTED, NULL,
B_FALSE, B_FALSE);
return (error);
out:
spa_unload(newspa);
spa_deactivate(newspa);
spa_remove(newspa);
txg = spa_vdev_config_enter(spa);
/* re-online all offlined disks */
for (c = 0; c < children; c++) {
if (vml[c] != NULL)
vml[c]->vdev_offline = B_FALSE;
}
/* restart initializing or trimming disks as necessary */
spa_async_request(spa, SPA_ASYNC_INITIALIZE_RESTART);
spa_async_request(spa, SPA_ASYNC_TRIM_RESTART);
spa_async_request(spa, SPA_ASYNC_AUTOTRIM_RESTART);
vdev_reopen(spa->spa_root_vdev);
nvlist_free(spa->spa_config_splitting);
spa->spa_config_splitting = NULL;
(void) spa_vdev_exit(spa, NULL, txg, error);
kmem_free(vml, children * sizeof (vdev_t *));
return (error);
}
/*
* Find any device that's done replacing, or a vdev marked 'unspare' that's
* currently spared, so we can detach it.
*/
static vdev_t *
spa_vdev_resilver_done_hunt(vdev_t *vd)
{
vdev_t *newvd, *oldvd;
for (int c = 0; c < vd->vdev_children; c++) {
oldvd = spa_vdev_resilver_done_hunt(vd->vdev_child[c]);
if (oldvd != NULL)
return (oldvd);
}
/*
* Check for a completed replacement. We always consider the first
* vdev in the list to be the oldest vdev, and the last one to be
* the newest (see spa_vdev_attach() for how that works). In
* the case where the newest vdev is faulted, we will not automatically
* remove it after a resilver completes. This is OK as it will require
* user intervention to determine which disk the admin wishes to keep.
*/
if (vd->vdev_ops == &vdev_replacing_ops) {
ASSERT(vd->vdev_children > 1);
newvd = vd->vdev_child[vd->vdev_children - 1];
oldvd = vd->vdev_child[0];
if (vdev_dtl_empty(newvd, DTL_MISSING) &&
vdev_dtl_empty(newvd, DTL_OUTAGE) &&
!vdev_dtl_required(oldvd))
return (oldvd);
}
/*
* Check for a completed resilver with the 'unspare' flag set.
* Also potentially update faulted state.
*/
if (vd->vdev_ops == &vdev_spare_ops) {
vdev_t *first = vd->vdev_child[0];
vdev_t *last = vd->vdev_child[vd->vdev_children - 1];
if (last->vdev_unspare) {
oldvd = first;
newvd = last;
} else if (first->vdev_unspare) {
oldvd = last;
newvd = first;
} else {
oldvd = NULL;
}
if (oldvd != NULL &&
vdev_dtl_empty(newvd, DTL_MISSING) &&
vdev_dtl_empty(newvd, DTL_OUTAGE) &&
!vdev_dtl_required(oldvd))
return (oldvd);
vdev_propagate_state(vd);
/*
* If there are more than two spares attached to a disk,
* and those spares are not required, then we want to
* attempt to free them up now so that they can be used
* by other pools. Once we're back down to a single
* disk+spare, we stop removing them.
*/
if (vd->vdev_children > 2) {
newvd = vd->vdev_child[1];
if (newvd->vdev_isspare && last->vdev_isspare &&
vdev_dtl_empty(last, DTL_MISSING) &&
vdev_dtl_empty(last, DTL_OUTAGE) &&
!vdev_dtl_required(newvd))
return (newvd);
}
}
return (NULL);
}
static void
spa_vdev_resilver_done(spa_t *spa)
{
vdev_t *vd, *pvd, *ppvd;
uint64_t guid, sguid, pguid, ppguid;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
while ((vd = spa_vdev_resilver_done_hunt(spa->spa_root_vdev)) != NULL) {
pvd = vd->vdev_parent;
ppvd = pvd->vdev_parent;
guid = vd->vdev_guid;
pguid = pvd->vdev_guid;
ppguid = ppvd->vdev_guid;
sguid = 0;
/*
* If we have just finished replacing a hot spared device, then
* we need to detach the parent's first child (the original hot
* spare) as well.
*/
if (ppvd->vdev_ops == &vdev_spare_ops && pvd->vdev_id == 0 &&
ppvd->vdev_children == 2) {
ASSERT(pvd->vdev_ops == &vdev_replacing_ops);
sguid = ppvd->vdev_child[1]->vdev_guid;
}
ASSERT(vd->vdev_resilver_txg == 0 || !vdev_dtl_required(vd));
spa_config_exit(spa, SCL_ALL, FTAG);
if (spa_vdev_detach(spa, guid, pguid, B_TRUE) != 0)
return;
if (sguid && spa_vdev_detach(spa, sguid, ppguid, B_TRUE) != 0)
return;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
}
spa_config_exit(spa, SCL_ALL, FTAG);
/*
* If a detach was not performed above replace waiters will not have
* been notified. In which case we must do so now.
*/
spa_notify_waiters(spa);
}
/*
* Update the stored path or FRU for this vdev.
*/
static int
spa_vdev_set_common(spa_t *spa, uint64_t guid, const char *value,
boolean_t ispath)
{
vdev_t *vd;
boolean_t sync = B_FALSE;
ASSERT(spa_writeable(spa));
spa_vdev_state_enter(spa, SCL_ALL);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, ENOENT));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_state_exit(spa, NULL, ENOTSUP));
if (ispath) {
if (strcmp(value, vd->vdev_path) != 0) {
spa_strfree(vd->vdev_path);
vd->vdev_path = spa_strdup(value);
sync = B_TRUE;
}
} else {
if (vd->vdev_fru == NULL) {
vd->vdev_fru = spa_strdup(value);
sync = B_TRUE;
} else if (strcmp(value, vd->vdev_fru) != 0) {
spa_strfree(vd->vdev_fru);
vd->vdev_fru = spa_strdup(value);
sync = B_TRUE;
}
}
return (spa_vdev_state_exit(spa, sync ? vd : NULL, 0));
}
int
spa_vdev_setpath(spa_t *spa, uint64_t guid, const char *newpath)
{
return (spa_vdev_set_common(spa, guid, newpath, B_TRUE));
}
int
spa_vdev_setfru(spa_t *spa, uint64_t guid, const char *newfru)
{
return (spa_vdev_set_common(spa, guid, newfru, B_FALSE));
}
/*
* ==========================================================================
* SPA Scanning
* ==========================================================================
*/
int
spa_scrub_pause_resume(spa_t *spa, pool_scrub_cmd_t cmd)
{
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0);
if (dsl_scan_resilvering(spa->spa_dsl_pool))
return (SET_ERROR(EBUSY));
return (dsl_scrub_set_pause_resume(spa->spa_dsl_pool, cmd));
}
int
spa_scan_stop(spa_t *spa)
{
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0);
if (dsl_scan_resilvering(spa->spa_dsl_pool))
return (SET_ERROR(EBUSY));
return (dsl_scan_cancel(spa->spa_dsl_pool));
}
int
spa_scan(spa_t *spa, pool_scan_func_t func)
{
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0);
if (func >= POOL_SCAN_FUNCS || func == POOL_SCAN_NONE)
return (SET_ERROR(ENOTSUP));
if (func == POOL_SCAN_RESILVER &&
!spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER))
return (SET_ERROR(ENOTSUP));
/*
* If a resilver was requested, but there is no DTL on a
* writeable leaf device, we have nothing to do.
*/
if (func == POOL_SCAN_RESILVER &&
!vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL)) {
spa_async_request(spa, SPA_ASYNC_RESILVER_DONE);
return (0);
}
if (func == POOL_SCAN_ERRORSCRUB &&
!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG))
return (SET_ERROR(ENOTSUP));
return (dsl_scan(spa->spa_dsl_pool, func));
}
/*
* ==========================================================================
* SPA async task processing
* ==========================================================================
*/
static void
spa_async_remove(spa_t *spa, vdev_t *vd)
{
if (vd->vdev_remove_wanted) {
vd->vdev_remove_wanted = B_FALSE;
vd->vdev_delayed_close = B_FALSE;
vdev_set_state(vd, B_FALSE, VDEV_STATE_REMOVED, VDEV_AUX_NONE);
/*
* We want to clear the stats, but we don't want to do a full
* vdev_clear() as that will cause us to throw away
* degraded/faulted state as well as attempt to reopen the
* device, all of which is a waste.
*/
vd->vdev_stat.vs_read_errors = 0;
vd->vdev_stat.vs_write_errors = 0;
vd->vdev_stat.vs_checksum_errors = 0;
vdev_state_dirty(vd->vdev_top);
/* Tell userspace that the vdev is gone. */
zfs_post_remove(spa, vd);
}
for (int c = 0; c < vd->vdev_children; c++)
spa_async_remove(spa, vd->vdev_child[c]);
}
static void
spa_async_probe(spa_t *spa, vdev_t *vd)
{
if (vd->vdev_probe_wanted) {
vd->vdev_probe_wanted = B_FALSE;
vdev_reopen(vd); /* vdev_open() does the actual probe */
}
for (int c = 0; c < vd->vdev_children; c++)
spa_async_probe(spa, vd->vdev_child[c]);
}
static void
spa_async_autoexpand(spa_t *spa, vdev_t *vd)
{
if (!spa->spa_autoexpand)
return;
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
spa_async_autoexpand(spa, cvd);
}
if (!vd->vdev_ops->vdev_op_leaf || vd->vdev_physpath == NULL)
return;
spa_event_notify(vd->vdev_spa, vd, NULL, ESC_ZFS_VDEV_AUTOEXPAND);
}
static __attribute__((noreturn)) void
spa_async_thread(void *arg)
{
spa_t *spa = (spa_t *)arg;
dsl_pool_t *dp = spa->spa_dsl_pool;
int tasks;
ASSERT(spa->spa_sync_on);
mutex_enter(&spa->spa_async_lock);
tasks = spa->spa_async_tasks;
spa->spa_async_tasks = 0;
mutex_exit(&spa->spa_async_lock);
/*
* See if the config needs to be updated.
*/
if (tasks & SPA_ASYNC_CONFIG_UPDATE) {
uint64_t old_space, new_space;
mutex_enter(&spa_namespace_lock);
old_space = metaslab_class_get_space(spa_normal_class(spa));
old_space += metaslab_class_get_space(spa_special_class(spa));
old_space += metaslab_class_get_space(spa_dedup_class(spa));
old_space += metaslab_class_get_space(
spa_embedded_log_class(spa));
spa_config_update(spa, SPA_CONFIG_UPDATE_POOL);
new_space = metaslab_class_get_space(spa_normal_class(spa));
new_space += metaslab_class_get_space(spa_special_class(spa));
new_space += metaslab_class_get_space(spa_dedup_class(spa));
new_space += metaslab_class_get_space(
spa_embedded_log_class(spa));
mutex_exit(&spa_namespace_lock);
/*
* If the pool grew as a result of the config update,
* then log an internal history event.
*/
if (new_space != old_space) {
spa_history_log_internal(spa, "vdev online", NULL,
"pool '%s' size: %llu(+%llu)",
spa_name(spa), (u_longlong_t)new_space,
(u_longlong_t)(new_space - old_space));
}
}
/*
* See if any devices need to be marked REMOVED.
*/
if (tasks & SPA_ASYNC_REMOVE) {
spa_vdev_state_enter(spa, SCL_NONE);
spa_async_remove(spa, spa->spa_root_vdev);
for (int i = 0; i < spa->spa_l2cache.sav_count; i++)
spa_async_remove(spa, spa->spa_l2cache.sav_vdevs[i]);
for (int i = 0; i < spa->spa_spares.sav_count; i++)
spa_async_remove(spa, spa->spa_spares.sav_vdevs[i]);
(void) spa_vdev_state_exit(spa, NULL, 0);
}
if ((tasks & SPA_ASYNC_AUTOEXPAND) && !spa_suspended(spa)) {
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
spa_async_autoexpand(spa, spa->spa_root_vdev);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
/*
* See if any devices need to be probed.
*/
if (tasks & SPA_ASYNC_PROBE) {
spa_vdev_state_enter(spa, SCL_NONE);
spa_async_probe(spa, spa->spa_root_vdev);
(void) spa_vdev_state_exit(spa, NULL, 0);
}
/*
* If any devices are done replacing, detach them.
*/
if (tasks & SPA_ASYNC_RESILVER_DONE ||
tasks & SPA_ASYNC_REBUILD_DONE ||
tasks & SPA_ASYNC_DETACH_SPARE) {
spa_vdev_resilver_done(spa);
}
/*
* Kick off a resilver.
*/
if (tasks & SPA_ASYNC_RESILVER &&
!vdev_rebuild_active(spa->spa_root_vdev) &&
(!dsl_scan_resilvering(dp) ||
!spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_RESILVER_DEFER)))
dsl_scan_restart_resilver(dp, 0);
if (tasks & SPA_ASYNC_INITIALIZE_RESTART) {
mutex_enter(&spa_namespace_lock);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_initialize_restart(spa->spa_root_vdev);
spa_config_exit(spa, SCL_CONFIG, FTAG);
mutex_exit(&spa_namespace_lock);
}
if (tasks & SPA_ASYNC_TRIM_RESTART) {
mutex_enter(&spa_namespace_lock);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_trim_restart(spa->spa_root_vdev);
spa_config_exit(spa, SCL_CONFIG, FTAG);
mutex_exit(&spa_namespace_lock);
}
if (tasks & SPA_ASYNC_AUTOTRIM_RESTART) {
mutex_enter(&spa_namespace_lock);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_autotrim_restart(spa);
spa_config_exit(spa, SCL_CONFIG, FTAG);
mutex_exit(&spa_namespace_lock);
}
/*
* Kick off L2 cache whole device TRIM.
*/
if (tasks & SPA_ASYNC_L2CACHE_TRIM) {
mutex_enter(&spa_namespace_lock);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_trim_l2arc(spa);
spa_config_exit(spa, SCL_CONFIG, FTAG);
mutex_exit(&spa_namespace_lock);
}
/*
* Kick off L2 cache rebuilding.
*/
if (tasks & SPA_ASYNC_L2CACHE_REBUILD) {
mutex_enter(&spa_namespace_lock);
spa_config_enter(spa, SCL_L2ARC, FTAG, RW_READER);
l2arc_spa_rebuild_start(spa);
spa_config_exit(spa, SCL_L2ARC, FTAG);
mutex_exit(&spa_namespace_lock);
}
/*
* Let the world know that we're done.
*/
mutex_enter(&spa->spa_async_lock);
spa->spa_async_thread = NULL;
cv_broadcast(&spa->spa_async_cv);
mutex_exit(&spa->spa_async_lock);
thread_exit();
}
void
spa_async_suspend(spa_t *spa)
{
mutex_enter(&spa->spa_async_lock);
spa->spa_async_suspended++;
while (spa->spa_async_thread != NULL)
cv_wait(&spa->spa_async_cv, &spa->spa_async_lock);
mutex_exit(&spa->spa_async_lock);
spa_vdev_remove_suspend(spa);
zthr_t *condense_thread = spa->spa_condense_zthr;
if (condense_thread != NULL)
zthr_cancel(condense_thread);
+ zthr_t *raidz_expand_thread = spa->spa_raidz_expand_zthr;
+ if (raidz_expand_thread != NULL)
+ zthr_cancel(raidz_expand_thread);
+
zthr_t *discard_thread = spa->spa_checkpoint_discard_zthr;
if (discard_thread != NULL)
zthr_cancel(discard_thread);
zthr_t *ll_delete_thread = spa->spa_livelist_delete_zthr;
if (ll_delete_thread != NULL)
zthr_cancel(ll_delete_thread);
zthr_t *ll_condense_thread = spa->spa_livelist_condense_zthr;
if (ll_condense_thread != NULL)
zthr_cancel(ll_condense_thread);
}
void
spa_async_resume(spa_t *spa)
{
mutex_enter(&spa->spa_async_lock);
ASSERT(spa->spa_async_suspended != 0);
spa->spa_async_suspended--;
mutex_exit(&spa->spa_async_lock);
spa_restart_removal(spa);
zthr_t *condense_thread = spa->spa_condense_zthr;
if (condense_thread != NULL)
zthr_resume(condense_thread);
+ zthr_t *raidz_expand_thread = spa->spa_raidz_expand_zthr;
+ if (raidz_expand_thread != NULL)
+ zthr_resume(raidz_expand_thread);
+
zthr_t *discard_thread = spa->spa_checkpoint_discard_zthr;
if (discard_thread != NULL)
zthr_resume(discard_thread);
zthr_t *ll_delete_thread = spa->spa_livelist_delete_zthr;
if (ll_delete_thread != NULL)
zthr_resume(ll_delete_thread);
zthr_t *ll_condense_thread = spa->spa_livelist_condense_zthr;
if (ll_condense_thread != NULL)
zthr_resume(ll_condense_thread);
}
static boolean_t
spa_async_tasks_pending(spa_t *spa)
{
uint_t non_config_tasks;
uint_t config_task;
boolean_t config_task_suspended;
non_config_tasks = spa->spa_async_tasks & ~SPA_ASYNC_CONFIG_UPDATE;
config_task = spa->spa_async_tasks & SPA_ASYNC_CONFIG_UPDATE;
if (spa->spa_ccw_fail_time == 0) {
config_task_suspended = B_FALSE;
} else {
config_task_suspended =
(gethrtime() - spa->spa_ccw_fail_time) <
((hrtime_t)zfs_ccw_retry_interval * NANOSEC);
}
return (non_config_tasks || (config_task && !config_task_suspended));
}
static void
spa_async_dispatch(spa_t *spa)
{
mutex_enter(&spa->spa_async_lock);
if (spa_async_tasks_pending(spa) &&
!spa->spa_async_suspended &&
spa->spa_async_thread == NULL)
spa->spa_async_thread = thread_create(NULL, 0,
spa_async_thread, spa, 0, &p0, TS_RUN, maxclsyspri);
mutex_exit(&spa->spa_async_lock);
}
void
spa_async_request(spa_t *spa, int task)
{
zfs_dbgmsg("spa=%s async request task=%u", spa->spa_name, task);
mutex_enter(&spa->spa_async_lock);
spa->spa_async_tasks |= task;
mutex_exit(&spa->spa_async_lock);
}
int
spa_async_tasks(spa_t *spa)
{
return (spa->spa_async_tasks);
}
/*
* ==========================================================================
* SPA syncing routines
* ==========================================================================
*/
static int
bpobj_enqueue_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
bpobj_t *bpo = arg;
bpobj_enqueue(bpo, bp, bp_freed, tx);
return (0);
}
int
bpobj_enqueue_alloc_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
return (bpobj_enqueue_cb(arg, bp, B_FALSE, tx));
}
int
bpobj_enqueue_free_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
return (bpobj_enqueue_cb(arg, bp, B_TRUE, tx));
}
static int
spa_free_sync_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
zio_t *pio = arg;
zio_nowait(zio_free_sync(pio, pio->io_spa, dmu_tx_get_txg(tx), bp,
pio->io_flags));
return (0);
}
static int
bpobj_spa_free_sync_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
ASSERT(!bp_freed);
return (spa_free_sync_cb(arg, bp, tx));
}
/*
* Note: this simple function is not inlined to make it easier to dtrace the
* amount of time spent syncing frees.
*/
static void
spa_sync_frees(spa_t *spa, bplist_t *bpl, dmu_tx_t *tx)
{
zio_t *zio = zio_root(spa, NULL, NULL, 0);
bplist_iterate(bpl, spa_free_sync_cb, zio, tx);
VERIFY(zio_wait(zio) == 0);
}
/*
* Note: this simple function is not inlined to make it easier to dtrace the
* amount of time spent syncing deferred frees.
*/
static void
spa_sync_deferred_frees(spa_t *spa, dmu_tx_t *tx)
{
if (spa_sync_pass(spa) != 1)
return;
/*
* Note:
* If the log space map feature is active, we stop deferring
* frees to the next TXG and therefore running this function
* would be considered a no-op as spa_deferred_bpobj should
* not have any entries.
*
* That said we run this function anyway (instead of returning
* immediately) for the edge-case scenario where we just
* activated the log space map feature in this TXG but we have
* deferred frees from the previous TXG.
*/
zio_t *zio = zio_root(spa, NULL, NULL, 0);
VERIFY3U(bpobj_iterate(&spa->spa_deferred_bpobj,
bpobj_spa_free_sync_cb, zio, tx), ==, 0);
VERIFY0(zio_wait(zio));
}
static void
spa_sync_nvlist(spa_t *spa, uint64_t obj, nvlist_t *nv, dmu_tx_t *tx)
{
char *packed = NULL;
size_t bufsize;
size_t nvsize = 0;
dmu_buf_t *db;
VERIFY(nvlist_size(nv, &nvsize, NV_ENCODE_XDR) == 0);
/*
* Write full (SPA_CONFIG_BLOCKSIZE) blocks of configuration
* information. This avoids the dmu_buf_will_dirty() path and
* saves us a pre-read to get data we don't actually care about.
*/
bufsize = P2ROUNDUP((uint64_t)nvsize, SPA_CONFIG_BLOCKSIZE);
packed = vmem_alloc(bufsize, KM_SLEEP);
VERIFY(nvlist_pack(nv, &packed, &nvsize, NV_ENCODE_XDR,
KM_SLEEP) == 0);
memset(packed + nvsize, 0, bufsize - nvsize);
dmu_write(spa->spa_meta_objset, obj, 0, bufsize, packed, tx);
vmem_free(packed, bufsize);
VERIFY(0 == dmu_bonus_hold(spa->spa_meta_objset, obj, FTAG, &db));
dmu_buf_will_dirty(db, tx);
*(uint64_t *)db->db_data = nvsize;
dmu_buf_rele(db, FTAG);
}
static void
spa_sync_aux_dev(spa_t *spa, spa_aux_vdev_t *sav, dmu_tx_t *tx,
const char *config, const char *entry)
{
nvlist_t *nvroot;
nvlist_t **list;
int i;
if (!sav->sav_sync)
return;
/*
* Update the MOS nvlist describing the list of available devices.
* spa_validate_aux() will have already made sure this nvlist is
* valid and the vdevs are labeled appropriately.
*/
if (sav->sav_object == 0) {
sav->sav_object = dmu_object_alloc(spa->spa_meta_objset,
DMU_OT_PACKED_NVLIST, 1 << 14, DMU_OT_PACKED_NVLIST_SIZE,
sizeof (uint64_t), tx);
VERIFY(zap_update(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, entry, sizeof (uint64_t), 1,
&sav->sav_object, tx) == 0);
}
nvroot = fnvlist_alloc();
if (sav->sav_count == 0) {
fnvlist_add_nvlist_array(nvroot, config,
(const nvlist_t * const *)NULL, 0);
} else {
list = kmem_alloc(sav->sav_count*sizeof (void *), KM_SLEEP);
for (i = 0; i < sav->sav_count; i++)
list[i] = vdev_config_generate(spa, sav->sav_vdevs[i],
B_FALSE, VDEV_CONFIG_L2CACHE);
fnvlist_add_nvlist_array(nvroot, config,
(const nvlist_t * const *)list, sav->sav_count);
for (i = 0; i < sav->sav_count; i++)
nvlist_free(list[i]);
kmem_free(list, sav->sav_count * sizeof (void *));
}
spa_sync_nvlist(spa, sav->sav_object, nvroot, tx);
nvlist_free(nvroot);
sav->sav_sync = B_FALSE;
}
/*
* Rebuild spa's all-vdev ZAP from the vdev ZAPs indicated in each vdev_t.
* The all-vdev ZAP must be empty.
*/
static void
spa_avz_build(vdev_t *vd, uint64_t avz, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
if (vd->vdev_root_zap != 0 &&
spa_feature_is_active(spa, SPA_FEATURE_AVZ_V2)) {
VERIFY0(zap_add_int(spa->spa_meta_objset, avz,
vd->vdev_root_zap, tx));
}
if (vd->vdev_top_zap != 0) {
VERIFY0(zap_add_int(spa->spa_meta_objset, avz,
vd->vdev_top_zap, tx));
}
if (vd->vdev_leaf_zap != 0) {
VERIFY0(zap_add_int(spa->spa_meta_objset, avz,
vd->vdev_leaf_zap, tx));
}
for (uint64_t i = 0; i < vd->vdev_children; i++) {
spa_avz_build(vd->vdev_child[i], avz, tx);
}
}
static void
spa_sync_config_object(spa_t *spa, dmu_tx_t *tx)
{
nvlist_t *config;
/*
* If the pool is being imported from a pre-per-vdev-ZAP version of ZFS,
* its config may not be dirty but we still need to build per-vdev ZAPs.
* Similarly, if the pool is being assembled (e.g. after a split), we
* need to rebuild the AVZ although the config may not be dirty.
*/
if (list_is_empty(&spa->spa_config_dirty_list) &&
spa->spa_avz_action == AVZ_ACTION_NONE)
return;
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
ASSERT(spa->spa_avz_action == AVZ_ACTION_NONE ||
spa->spa_avz_action == AVZ_ACTION_INITIALIZE ||
spa->spa_all_vdev_zaps != 0);
if (spa->spa_avz_action == AVZ_ACTION_REBUILD) {
/* Make and build the new AVZ */
uint64_t new_avz = zap_create(spa->spa_meta_objset,
DMU_OTN_ZAP_METADATA, DMU_OT_NONE, 0, tx);
spa_avz_build(spa->spa_root_vdev, new_avz, tx);
/* Diff old AVZ with new one */
zap_cursor_t zc;
zap_attribute_t za;
for (zap_cursor_init(&zc, spa->spa_meta_objset,
spa->spa_all_vdev_zaps);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
uint64_t vdzap = za.za_first_integer;
if (zap_lookup_int(spa->spa_meta_objset, new_avz,
vdzap) == ENOENT) {
/*
* ZAP is listed in old AVZ but not in new one;
* destroy it
*/
VERIFY0(zap_destroy(spa->spa_meta_objset, vdzap,
tx));
}
}
zap_cursor_fini(&zc);
/* Destroy the old AVZ */
VERIFY0(zap_destroy(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, tx));
/* Replace the old AVZ in the dir obj with the new one */
VERIFY0(zap_update(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_VDEV_ZAP_MAP,
sizeof (new_avz), 1, &new_avz, tx));
spa->spa_all_vdev_zaps = new_avz;
} else if (spa->spa_avz_action == AVZ_ACTION_DESTROY) {
zap_cursor_t zc;
zap_attribute_t za;
/* Walk through the AVZ and destroy all listed ZAPs */
for (zap_cursor_init(&zc, spa->spa_meta_objset,
spa->spa_all_vdev_zaps);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
uint64_t zap = za.za_first_integer;
VERIFY0(zap_destroy(spa->spa_meta_objset, zap, tx));
}
zap_cursor_fini(&zc);
/* Destroy and unlink the AVZ itself */
VERIFY0(zap_destroy(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, tx));
VERIFY0(zap_remove(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_VDEV_ZAP_MAP, tx));
spa->spa_all_vdev_zaps = 0;
}
if (spa->spa_all_vdev_zaps == 0) {
spa->spa_all_vdev_zaps = zap_create_link(spa->spa_meta_objset,
DMU_OTN_ZAP_METADATA, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_VDEV_ZAP_MAP, tx);
}
spa->spa_avz_action = AVZ_ACTION_NONE;
/* Create ZAPs for vdevs that don't have them. */
vdev_construct_zaps(spa->spa_root_vdev, tx);
config = spa_config_generate(spa, spa->spa_root_vdev,
dmu_tx_get_txg(tx), B_FALSE);
/*
* If we're upgrading the spa version then make sure that
* the config object gets updated with the correct version.
*/
if (spa->spa_ubsync.ub_version < spa->spa_uberblock.ub_version)
fnvlist_add_uint64(config, ZPOOL_CONFIG_VERSION,
spa->spa_uberblock.ub_version);
spa_config_exit(spa, SCL_STATE, FTAG);
nvlist_free(spa->spa_config_syncing);
spa->spa_config_syncing = config;
spa_sync_nvlist(spa, spa->spa_config_object, config, tx);
}
static void
spa_sync_version(void *arg, dmu_tx_t *tx)
{
uint64_t *versionp = arg;
uint64_t version = *versionp;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
/*
* Setting the version is special cased when first creating the pool.
*/
ASSERT(tx->tx_txg != TXG_INITIAL);
ASSERT(SPA_VERSION_IS_SUPPORTED(version));
ASSERT(version >= spa_version(spa));
spa->spa_uberblock.ub_version = version;
vdev_config_dirty(spa->spa_root_vdev);
spa_history_log_internal(spa, "set", tx, "version=%lld",
(longlong_t)version);
}
/*
* Set zpool properties.
*/
static void
spa_sync_props(void *arg, dmu_tx_t *tx)
{
nvlist_t *nvp = arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
objset_t *mos = spa->spa_meta_objset;
nvpair_t *elem = NULL;
mutex_enter(&spa->spa_props_lock);
while ((elem = nvlist_next_nvpair(nvp, elem))) {
uint64_t intval;
const char *strval, *fname;
zpool_prop_t prop;
const char *propname;
const char *elemname = nvpair_name(elem);
zprop_type_t proptype;
spa_feature_t fid;
switch (prop = zpool_name_to_prop(elemname)) {
case ZPOOL_PROP_VERSION:
intval = fnvpair_value_uint64(elem);
/*
* The version is synced separately before other
* properties and should be correct by now.
*/
ASSERT3U(spa_version(spa), >=, intval);
break;
case ZPOOL_PROP_ALTROOT:
/*
* 'altroot' is a non-persistent property. It should
* have been set temporarily at creation or import time.
*/
ASSERT(spa->spa_root != NULL);
break;
case ZPOOL_PROP_READONLY:
case ZPOOL_PROP_CACHEFILE:
/*
* 'readonly' and 'cachefile' are also non-persistent
* properties.
*/
break;
case ZPOOL_PROP_COMMENT:
strval = fnvpair_value_string(elem);
if (spa->spa_comment != NULL)
spa_strfree(spa->spa_comment);
spa->spa_comment = spa_strdup(strval);
/*
* We need to dirty the configuration on all the vdevs
* so that their labels get updated. We also need to
* update the cache file to keep it in sync with the
* MOS version. It's unnecessary to do this for pool
* creation since the vdev's configuration has already
* been dirtied.
*/
if (tx->tx_txg != TXG_INITIAL) {
vdev_config_dirty(spa->spa_root_vdev);
spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
}
spa_history_log_internal(spa, "set", tx,
"%s=%s", elemname, strval);
break;
case ZPOOL_PROP_COMPATIBILITY:
strval = fnvpair_value_string(elem);
if (spa->spa_compatibility != NULL)
spa_strfree(spa->spa_compatibility);
spa->spa_compatibility = spa_strdup(strval);
/*
* Dirty the configuration on vdevs as above.
*/
if (tx->tx_txg != TXG_INITIAL) {
vdev_config_dirty(spa->spa_root_vdev);
spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
}
spa_history_log_internal(spa, "set", tx,
"%s=%s", nvpair_name(elem), strval);
break;
case ZPOOL_PROP_INVAL:
if (zpool_prop_feature(elemname)) {
fname = strchr(elemname, '@') + 1;
VERIFY0(zfeature_lookup_name(fname, &fid));
spa_feature_enable(spa, fid, tx);
spa_history_log_internal(spa, "set", tx,
"%s=enabled", elemname);
break;
} else if (!zfs_prop_user(elemname)) {
ASSERT(zpool_prop_feature(elemname));
break;
}
zfs_fallthrough;
default:
/*
* Set pool property values in the poolprops mos object.
*/
if (spa->spa_pool_props_object == 0) {
spa->spa_pool_props_object =
zap_create_link(mos, DMU_OT_POOL_PROPS,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_PROPS,
tx);
}
/* normalize the property name */
if (prop == ZPOOL_PROP_INVAL) {
propname = elemname;
proptype = PROP_TYPE_STRING;
} else {
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", elemname, strval);
} else if (nvpair_type(elem) == DATA_TYPE_UINT64) {
intval = fnvpair_value_uint64(elem);
if (proptype == PROP_TYPE_INDEX) {
const char *unused;
VERIFY0(zpool_prop_index_to_string(
prop, intval, &unused));
}
VERIFY0(zap_update(mos,
spa->spa_pool_props_object, propname,
8, 1, &intval, tx));
spa_history_log_internal(spa, "set", tx,
"%s=%lld", elemname,
(longlong_t)intval);
switch (prop) {
case ZPOOL_PROP_DELEGATION:
spa->spa_delegation = intval;
break;
case ZPOOL_PROP_BOOTFS:
spa->spa_bootfs = intval;
break;
case ZPOOL_PROP_FAILUREMODE:
spa->spa_failmode = intval;
break;
case ZPOOL_PROP_AUTOTRIM:
spa->spa_autotrim = intval;
spa_async_request(spa,
SPA_ASYNC_AUTOTRIM_RESTART);
break;
case ZPOOL_PROP_AUTOEXPAND:
spa->spa_autoexpand = intval;
if (tx->tx_txg != TXG_INITIAL)
spa_async_request(spa,
SPA_ASYNC_AUTOEXPAND);
break;
case ZPOOL_PROP_MULTIHOST:
spa->spa_multihost = intval;
break;
default:
break;
}
} else {
ASSERT(0); /* not allowed */
}
}
}
mutex_exit(&spa->spa_props_lock);
}
/*
* Perform one-time upgrade on-disk changes. spa_version() does not
* reflect the new version this txg, so there must be no changes this
* txg to anything that the upgrade code depends on after it executes.
* Therefore this must be called after dsl_pool_sync() does the sync
* tasks.
*/
static void
spa_sync_upgrades(spa_t *spa, dmu_tx_t *tx)
{
if (spa_sync_pass(spa) != 1)
return;
dsl_pool_t *dp = spa->spa_dsl_pool;
rrw_enter(&dp->dp_config_rwlock, RW_WRITER, FTAG);
if (spa->spa_ubsync.ub_version < SPA_VERSION_ORIGIN &&
spa->spa_uberblock.ub_version >= SPA_VERSION_ORIGIN) {
dsl_pool_create_origin(dp, tx);
/* Keeping the origin open increases spa_minref */
spa->spa_minref += 3;
}
if (spa->spa_ubsync.ub_version < SPA_VERSION_NEXT_CLONES &&
spa->spa_uberblock.ub_version >= SPA_VERSION_NEXT_CLONES) {
dsl_pool_upgrade_clones(dp, tx);
}
if (spa->spa_ubsync.ub_version < SPA_VERSION_DIR_CLONES &&
spa->spa_uberblock.ub_version >= SPA_VERSION_DIR_CLONES) {
dsl_pool_upgrade_dir_clones(dp, tx);
/* Keeping the freedir open increases spa_minref */
spa->spa_minref += 3;
}
if (spa->spa_ubsync.ub_version < SPA_VERSION_FEATURES &&
spa->spa_uberblock.ub_version >= SPA_VERSION_FEATURES) {
spa_feature_create_zap_objects(spa, tx);
}
/*
* LZ4_COMPRESS feature's behaviour was changed to activate_on_enable
* when possibility to use lz4 compression for metadata was added
* Old pools that have this feature enabled must be upgraded to have
* this feature active
*/
if (spa->spa_uberblock.ub_version >= SPA_VERSION_FEATURES) {
boolean_t lz4_en = spa_feature_is_enabled(spa,
SPA_FEATURE_LZ4_COMPRESS);
boolean_t lz4_ac = spa_feature_is_active(spa,
SPA_FEATURE_LZ4_COMPRESS);
if (lz4_en && !lz4_ac)
spa_feature_incr(spa, SPA_FEATURE_LZ4_COMPRESS, tx);
}
/*
* If we haven't written the salt, do so now. Note that the
* feature may not be activated yet, but that's fine since
* the presence of this ZAP entry is backwards compatible.
*/
if (zap_contains(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_CHECKSUM_SALT) == ENOENT) {
VERIFY0(zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CHECKSUM_SALT, 1,
sizeof (spa->spa_cksum_salt.zcs_bytes),
spa->spa_cksum_salt.zcs_bytes, tx));
}
rrw_exit(&dp->dp_config_rwlock, FTAG);
}
static void
vdev_indirect_state_sync_verify(vdev_t *vd)
{
vdev_indirect_mapping_t *vim __maybe_unused = vd->vdev_indirect_mapping;
vdev_indirect_births_t *vib __maybe_unused = vd->vdev_indirect_births;
if (vd->vdev_ops == &vdev_indirect_ops) {
ASSERT(vim != NULL);
ASSERT(vib != NULL);
}
uint64_t obsolete_sm_object = 0;
ASSERT0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
if (obsolete_sm_object != 0) {
ASSERT(vd->vdev_obsolete_sm != NULL);
ASSERT(vd->vdev_removing ||
vd->vdev_ops == &vdev_indirect_ops);
ASSERT(vdev_indirect_mapping_num_entries(vim) > 0);
ASSERT(vdev_indirect_mapping_bytes_mapped(vim) > 0);
ASSERT3U(obsolete_sm_object, ==,
space_map_object(vd->vdev_obsolete_sm));
ASSERT3U(vdev_indirect_mapping_bytes_mapped(vim), >=,
space_map_allocated(vd->vdev_obsolete_sm));
}
ASSERT(vd->vdev_obsolete_segments != NULL);
/*
* Since frees / remaps to an indirect vdev can only
* happen in syncing context, the obsolete segments
* tree must be empty when we start syncing.
*/
ASSERT0(range_tree_space(vd->vdev_obsolete_segments));
}
/*
* Set the top-level vdev's max queue depth. Evaluate each top-level's
* async write queue depth in case it changed. The max queue depth will
* not change in the middle of syncing out this txg.
*/
static void
spa_sync_adjust_vdev_max_queue_depth(spa_t *spa)
{
ASSERT(spa_writeable(spa));
vdev_t *rvd = spa->spa_root_vdev;
uint32_t max_queue_depth = zfs_vdev_async_write_max_active *
zfs_vdev_queue_depth_pct / 100;
metaslab_class_t *normal = spa_normal_class(spa);
metaslab_class_t *special = spa_special_class(spa);
metaslab_class_t *dedup = spa_dedup_class(spa);
uint64_t slots_per_allocator = 0;
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
metaslab_group_t *mg = tvd->vdev_mg;
if (mg == NULL || !metaslab_group_initialized(mg))
continue;
metaslab_class_t *mc = mg->mg_class;
if (mc != normal && mc != special && mc != dedup)
continue;
/*
* It is safe to do a lock-free check here because only async
* allocations look at mg_max_alloc_queue_depth, and async
* allocations all happen from spa_sync().
*/
for (int i = 0; i < mg->mg_allocators; i++) {
ASSERT0(zfs_refcount_count(
&(mg->mg_allocator[i].mga_alloc_queue_depth)));
}
mg->mg_max_alloc_queue_depth = max_queue_depth;
for (int i = 0; i < mg->mg_allocators; i++) {
mg->mg_allocator[i].mga_cur_max_alloc_queue_depth =
zfs_vdev_def_queue_depth;
}
slots_per_allocator += zfs_vdev_def_queue_depth;
}
for (int i = 0; i < spa->spa_alloc_count; i++) {
ASSERT0(zfs_refcount_count(&normal->mc_allocator[i].
mca_alloc_slots));
ASSERT0(zfs_refcount_count(&special->mc_allocator[i].
mca_alloc_slots));
ASSERT0(zfs_refcount_count(&dedup->mc_allocator[i].
mca_alloc_slots));
normal->mc_allocator[i].mca_alloc_max_slots =
slots_per_allocator;
special->mc_allocator[i].mca_alloc_max_slots =
slots_per_allocator;
dedup->mc_allocator[i].mca_alloc_max_slots =
slots_per_allocator;
}
normal->mc_alloc_throttle_enabled = zio_dva_throttle_enabled;
special->mc_alloc_throttle_enabled = zio_dva_throttle_enabled;
dedup->mc_alloc_throttle_enabled = zio_dva_throttle_enabled;
}
static void
spa_sync_condense_indirect(spa_t *spa, dmu_tx_t *tx)
{
ASSERT(spa_writeable(spa));
vdev_t *rvd = spa->spa_root_vdev;
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
vdev_indirect_state_sync_verify(vd);
if (vdev_indirect_should_condense(vd)) {
spa_condense_indirect_start_sync(vd, tx);
break;
}
}
}
static void
spa_sync_iterate_to_convergence(spa_t *spa, dmu_tx_t *tx)
{
objset_t *mos = spa->spa_meta_objset;
dsl_pool_t *dp = spa->spa_dsl_pool;
uint64_t txg = tx->tx_txg;
bplist_t *free_bpl = &spa->spa_free_bplist[txg & TXG_MASK];
do {
int pass = ++spa->spa_sync_pass;
spa_sync_config_object(spa, tx);
spa_sync_aux_dev(spa, &spa->spa_spares, tx,
ZPOOL_CONFIG_SPARES, DMU_POOL_SPARES);
spa_sync_aux_dev(spa, &spa->spa_l2cache, tx,
ZPOOL_CONFIG_L2CACHE, DMU_POOL_L2CACHE);
spa_errlog_sync(spa, txg);
dsl_pool_sync(dp, txg);
if (pass < zfs_sync_pass_deferred_free ||
spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP)) {
/*
* If the log space map feature is active we don't
* care about deferred frees and the deferred bpobj
* as the log space map should effectively have the
* same results (i.e. appending only to one object).
*/
spa_sync_frees(spa, free_bpl, tx);
} else {
/*
* We can not defer frees in pass 1, because
* we sync the deferred frees later in pass 1.
*/
ASSERT3U(pass, >, 1);
bplist_iterate(free_bpl, bpobj_enqueue_alloc_cb,
&spa->spa_deferred_bpobj, tx);
}
brt_sync(spa, txg);
ddt_sync(spa, txg);
dsl_scan_sync(dp, tx);
dsl_errorscrub_sync(dp, tx);
svr_sync(spa, tx);
spa_sync_upgrades(spa, tx);
spa_flush_metaslabs(spa, tx);
vdev_t *vd = NULL;
while ((vd = txg_list_remove(&spa->spa_vdev_txg_list, txg))
!= NULL)
vdev_sync(vd, txg);
+ if (pass == 1) {
+ /*
+ * dsl_pool_sync() -> dp_sync_tasks may have dirtied
+ * the config. If that happens, this txg should not
+ * be a no-op. So we must sync the config to the MOS
+ * before checking for no-op.
+ *
+ * Note that when the config is dirty, it will
+ * be written to the MOS (i.e. the MOS will be
+ * dirtied) every time we call spa_sync_config_object()
+ * in this txg. Therefore we can't call this after
+ * dsl_pool_sync() every pass, because it would
+ * prevent us from converging, since we'd dirty
+ * the MOS every pass.
+ *
+ * Sync tasks can only be processed in pass 1, so
+ * there's no need to do this in later passes.
+ */
+ spa_sync_config_object(spa, tx);
+ }
+
/*
* Note: We need to check if the MOS is dirty because we could
* have marked the MOS dirty without updating the uberblock
* (e.g. if we have sync tasks but no dirty user data). We need
* to check the uberblock's rootbp because it is updated if we
* have synced out dirty data (though in this case the MOS will
* most likely also be dirty due to second order effects, we
* don't want to rely on that here).
*/
if (pass == 1 &&
spa->spa_uberblock.ub_rootbp.blk_birth < txg &&
!dmu_objset_is_dirty(mos, txg)) {
/*
* Nothing changed on the first pass, therefore this
* TXG is a no-op. Avoid syncing deferred frees, so
* that we can keep this TXG as a no-op.
*/
ASSERT(txg_list_empty(&dp->dp_dirty_datasets, txg));
ASSERT(txg_list_empty(&dp->dp_dirty_dirs, txg));
ASSERT(txg_list_empty(&dp->dp_sync_tasks, txg));
ASSERT(txg_list_empty(&dp->dp_early_sync_tasks, txg));
break;
}
spa_sync_deferred_frees(spa, tx);
} while (dmu_objset_is_dirty(mos, txg));
}
/*
* Rewrite the vdev configuration (which includes the uberblock) to
* commit the transaction group.
*
* If there are no dirty vdevs, we sync the uberblock to a few random
* top-level vdevs that are known to be visible in the config cache
* (see spa_vdev_add() for a complete description). If there *are* dirty
* vdevs, sync the uberblock to all vdevs.
*/
static void
spa_sync_rewrite_vdev_config(spa_t *spa, dmu_tx_t *tx)
{
vdev_t *rvd = spa->spa_root_vdev;
uint64_t txg = tx->tx_txg;
for (;;) {
int error = 0;
/*
* We hold SCL_STATE to prevent vdev open/close/etc.
* while we're attempting to write the vdev labels.
*/
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
if (list_is_empty(&spa->spa_config_dirty_list)) {
vdev_t *svd[SPA_SYNC_MIN_VDEVS] = { NULL };
int svdcount = 0;
int children = rvd->vdev_children;
int c0 = random_in_range(children);
for (int c = 0; c < children; c++) {
vdev_t *vd =
rvd->vdev_child[(c0 + c) % children];
/* Stop when revisiting the first vdev */
if (c > 0 && svd[0] == vd)
break;
if (vd->vdev_ms_array == 0 ||
vd->vdev_islog ||
!vdev_is_concrete(vd))
continue;
svd[svdcount++] = vd;
if (svdcount == SPA_SYNC_MIN_VDEVS)
break;
}
error = vdev_config_sync(svd, svdcount, txg);
} else {
error = vdev_config_sync(rvd->vdev_child,
rvd->vdev_children, txg);
}
if (error == 0)
spa->spa_last_synced_guid = rvd->vdev_guid;
spa_config_exit(spa, SCL_STATE, FTAG);
if (error == 0)
break;
zio_suspend(spa, NULL, ZIO_SUSPEND_IOERR);
zio_resume_wait(spa);
}
}
/*
* Sync the specified transaction group. New blocks may be dirtied as
* part of the process, so we iterate until it converges.
*/
void
spa_sync(spa_t *spa, uint64_t txg)
{
vdev_t *vd = NULL;
VERIFY(spa_writeable(spa));
/*
* Wait for i/os issued in open context that need to complete
* before this txg syncs.
*/
(void) zio_wait(spa->spa_txg_zio[txg & TXG_MASK]);
spa->spa_txg_zio[txg & TXG_MASK] = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL);
/*
* Now that there can be no more cloning in this transaction group,
* but we are still before issuing frees, we can process pending BRT
* updates.
*/
brt_pending_apply(spa, txg);
/*
* Lock out configuration changes.
*/
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
spa->spa_syncing_txg = txg;
spa->spa_sync_pass = 0;
for (int i = 0; i < spa->spa_alloc_count; i++) {
mutex_enter(&spa->spa_allocs[i].spaa_lock);
VERIFY0(avl_numnodes(&spa->spa_allocs[i].spaa_tree));
mutex_exit(&spa->spa_allocs[i].spaa_lock);
}
/*
* If there are any pending vdev state changes, convert them
* into config changes that go out with this transaction group.
*/
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
while ((vd = list_head(&spa->spa_state_dirty_list)) != NULL) {
/* Avoid holding the write lock unless actually necessary */
if (vd->vdev_aux == NULL) {
vdev_state_clean(vd);
vdev_config_dirty(vd);
continue;
}
/*
* We need the write lock here because, for aux vdevs,
* calling vdev_config_dirty() modifies sav_config.
* This is ugly and will become unnecessary when we
* eliminate the aux vdev wart by integrating all vdevs
* into the root vdev tree.
*/
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_WRITER);
while ((vd = list_head(&spa->spa_state_dirty_list)) != NULL) {
vdev_state_clean(vd);
vdev_config_dirty(vd);
}
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
}
spa_config_exit(spa, SCL_STATE, FTAG);
dsl_pool_t *dp = spa->spa_dsl_pool;
dmu_tx_t *tx = dmu_tx_create_assigned(dp, txg);
spa->spa_sync_starttime = gethrtime();
taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid);
spa->spa_deadman_tqid = taskq_dispatch_delay(system_delay_taskq,
spa_deadman, spa, TQ_SLEEP, ddi_get_lbolt() +
NSEC_TO_TICK(spa->spa_deadman_synctime));
/*
* If we are upgrading to SPA_VERSION_RAIDZ_DEFLATE this txg,
* set spa_deflate if we have no raid-z vdevs.
*/
if (spa->spa_ubsync.ub_version < SPA_VERSION_RAIDZ_DEFLATE &&
spa->spa_uberblock.ub_version >= SPA_VERSION_RAIDZ_DEFLATE) {
vdev_t *rvd = spa->spa_root_vdev;
int i;
for (i = 0; i < rvd->vdev_children; i++) {
vd = rvd->vdev_child[i];
if (vd->vdev_deflate_ratio != SPA_MINBLOCKSIZE)
break;
}
if (i == rvd->vdev_children) {
spa->spa_deflate = TRUE;
VERIFY0(zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_DEFLATE,
sizeof (uint64_t), 1, &spa->spa_deflate, tx));
}
}
spa_sync_adjust_vdev_max_queue_depth(spa);
spa_sync_condense_indirect(spa, tx);
spa_sync_iterate_to_convergence(spa, tx);
#ifdef ZFS_DEBUG
if (!list_is_empty(&spa->spa_config_dirty_list)) {
/*
* Make sure that the number of ZAPs for all the vdevs matches
* the number of ZAPs in the per-vdev ZAP list. This only gets
* called if the config is dirty; otherwise there may be
* outstanding AVZ operations that weren't completed in
* spa_sync_config_object.
*/
uint64_t all_vdev_zap_entry_count;
ASSERT0(zap_count(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, &all_vdev_zap_entry_count));
ASSERT3U(vdev_count_verify_zaps(spa->spa_root_vdev), ==,
all_vdev_zap_entry_count);
}
#endif
if (spa->spa_vdev_removal != NULL) {
ASSERT0(spa->spa_vdev_removal->svr_bytes_done[txg & TXG_MASK]);
}
spa_sync_rewrite_vdev_config(spa, tx);
dmu_tx_commit(tx);
taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid);
spa->spa_deadman_tqid = 0;
/*
* Clear the dirty config list.
*/
while ((vd = list_head(&spa->spa_config_dirty_list)) != NULL)
vdev_config_clean(vd);
/*
* Now that the new config has synced transactionally,
* let it become visible to the config cache.
*/
if (spa->spa_config_syncing != NULL) {
spa_config_set(spa, spa->spa_config_syncing);
spa->spa_config_txg = txg;
spa->spa_config_syncing = NULL;
}
dsl_pool_sync_done(dp, txg);
for (int i = 0; i < spa->spa_alloc_count; i++) {
mutex_enter(&spa->spa_allocs[i].spaa_lock);
VERIFY0(avl_numnodes(&spa->spa_allocs[i].spaa_tree));
mutex_exit(&spa->spa_allocs[i].spaa_lock);
}
/*
* Update usable space statistics.
*/
while ((vd = txg_list_remove(&spa->spa_vdev_txg_list, TXG_CLEAN(txg)))
!= NULL)
vdev_sync_done(vd, txg);
metaslab_class_evict_old(spa->spa_normal_class, txg);
metaslab_class_evict_old(spa->spa_log_class, txg);
spa_sync_close_syncing_log_sm(spa);
spa_update_dspace(spa);
if (spa_get_autotrim(spa) == SPA_AUTOTRIM_ON)
vdev_autotrim_kick(spa);
/*
* It had better be the case that we didn't dirty anything
* since vdev_config_sync().
*/
ASSERT(txg_list_empty(&dp->dp_dirty_datasets, txg));
ASSERT(txg_list_empty(&dp->dp_dirty_dirs, txg));
ASSERT(txg_list_empty(&spa->spa_vdev_txg_list, txg));
while (zfs_pause_spa_sync)
delay(1);
spa->spa_sync_pass = 0;
/*
* Update the last synced uberblock here. We want to do this at
* the end of spa_sync() so that consumers of spa_last_synced_txg()
* will be guaranteed that all the processing associated with
* that txg has been completed.
*/
spa->spa_ubsync = spa->spa_uberblock;
spa_config_exit(spa, SCL_CONFIG, FTAG);
spa_handle_ignored_writes(spa);
/*
* If any async tasks have been requested, kick them off.
*/
spa_async_dispatch(spa);
}
/*
* Sync all pools. We don't want to hold the namespace lock across these
* operations, so we take a reference on the spa_t and drop the lock during the
* sync.
*/
void
spa_sync_allpools(void)
{
spa_t *spa = NULL;
mutex_enter(&spa_namespace_lock);
while ((spa = spa_next(spa)) != NULL) {
if (spa_state(spa) != POOL_STATE_ACTIVE ||
!spa_writeable(spa) || spa_suspended(spa))
continue;
spa_open_ref(spa, FTAG);
mutex_exit(&spa_namespace_lock);
txg_wait_synced(spa_get_dsl(spa), 0);
mutex_enter(&spa_namespace_lock);
spa_close(spa, FTAG);
}
mutex_exit(&spa_namespace_lock);
}
taskq_t *
spa_sync_tq_create(spa_t *spa, const char *name)
{
kthread_t **kthreads;
ASSERT(spa->spa_sync_tq == NULL);
ASSERT3S(spa->spa_alloc_count, <=, boot_ncpus);
/*
* - do not allow more allocators than cpus.
* - there may be more cpus than allocators.
* - do not allow more sync taskq threads than allocators or cpus.
*/
int nthreads = spa->spa_alloc_count;
spa->spa_syncthreads = kmem_zalloc(sizeof (spa_syncthread_info_t) *
nthreads, KM_SLEEP);
spa->spa_sync_tq = taskq_create_synced(name, nthreads, minclsyspri,
nthreads, INT_MAX, TASKQ_PREPOPULATE, &kthreads);
VERIFY(spa->spa_sync_tq != NULL);
VERIFY(kthreads != NULL);
spa_taskqs_t *tqs =
&spa->spa_zio_taskq[ZIO_TYPE_WRITE][ZIO_TASKQ_ISSUE];
spa_syncthread_info_t *ti = spa->spa_syncthreads;
for (int i = 0, w = 0; i < nthreads; i++, w++, ti++) {
ti->sti_thread = kthreads[i];
if (w == tqs->stqs_count) {
w = 0;
}
ti->sti_wr_iss_tq = tqs->stqs_taskq[w];
}
kmem_free(kthreads, sizeof (*kthreads) * nthreads);
return (spa->spa_sync_tq);
}
void
spa_sync_tq_destroy(spa_t *spa)
{
ASSERT(spa->spa_sync_tq != NULL);
taskq_wait(spa->spa_sync_tq);
taskq_destroy(spa->spa_sync_tq);
kmem_free(spa->spa_syncthreads,
sizeof (spa_syncthread_info_t) * spa->spa_alloc_count);
spa->spa_sync_tq = NULL;
}
void
spa_select_allocator(zio_t *zio)
{
zbookmark_phys_t *bm = &zio->io_bookmark;
spa_t *spa = zio->io_spa;
ASSERT(zio->io_type == ZIO_TYPE_WRITE);
/*
* A gang block (for example) may have inherited its parent's
* allocator, in which case there is nothing further to do here.
*/
if (ZIO_HAS_ALLOCATOR(zio))
return;
ASSERT(spa != NULL);
ASSERT(bm != NULL);
/*
* First try to use an allocator assigned to the syncthread, and set
* the corresponding write issue taskq for the allocator.
* Note, we must have an open pool to do this.
*/
if (spa->spa_sync_tq != NULL) {
spa_syncthread_info_t *ti = spa->spa_syncthreads;
for (int i = 0; i < spa->spa_alloc_count; i++, ti++) {
if (ti->sti_thread == curthread) {
zio->io_allocator = i;
zio->io_wr_iss_tq = ti->sti_wr_iss_tq;
return;
}
}
}
/*
* We want to try to use as many allocators as possible to help improve
* performance, but we also want logically adjacent IOs to be physically
* adjacent to improve sequential read performance. We chunk each object
* into 2^20 block regions, and then hash based on the objset, object,
* level, and region to accomplish both of these goals.
*/
uint64_t hv = cityhash4(bm->zb_objset, bm->zb_object, bm->zb_level,
bm->zb_blkid >> 20);
zio->io_allocator = (uint_t)hv % spa->spa_alloc_count;
zio->io_wr_iss_tq = NULL;
}
/*
* ==========================================================================
* Miscellaneous routines
* ==========================================================================
*/
/*
* Remove all pools in the system.
*/
void
spa_evict_all(void)
{
spa_t *spa;
/*
* Remove all cached state. All pools should be closed now,
* so every spa in the AVL tree should be unreferenced.
*/
mutex_enter(&spa_namespace_lock);
while ((spa = spa_next(NULL)) != NULL) {
/*
* Stop async tasks. The async thread may need to detach
* a device that's been replaced, which requires grabbing
* spa_namespace_lock, so we must drop it here.
*/
spa_open_ref(spa, FTAG);
mutex_exit(&spa_namespace_lock);
spa_async_suspend(spa);
mutex_enter(&spa_namespace_lock);
spa_close(spa, FTAG);
if (spa->spa_state != POOL_STATE_UNINITIALIZED) {
spa_unload(spa);
spa_deactivate(spa);
}
spa_remove(spa);
}
mutex_exit(&spa_namespace_lock);
}
vdev_t *
spa_lookup_by_guid(spa_t *spa, uint64_t guid, boolean_t aux)
{
vdev_t *vd;
int i;
if ((vd = vdev_lookup_by_guid(spa->spa_root_vdev, guid)) != NULL)
return (vd);
if (aux) {
for (i = 0; i < spa->spa_l2cache.sav_count; i++) {
vd = spa->spa_l2cache.sav_vdevs[i];
if (vd->vdev_guid == guid)
return (vd);
}
for (i = 0; i < spa->spa_spares.sav_count; i++) {
vd = spa->spa_spares.sav_vdevs[i];
if (vd->vdev_guid == guid)
return (vd);
}
}
return (NULL);
}
void
spa_upgrade(spa_t *spa, uint64_t version)
{
ASSERT(spa_writeable(spa));
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
/*
* This should only be called for a non-faulted pool, and since a
* future version would result in an unopenable pool, this shouldn't be
* possible.
*/
ASSERT(SPA_VERSION_IS_SUPPORTED(spa->spa_uberblock.ub_version));
ASSERT3U(version, >=, spa->spa_uberblock.ub_version);
spa->spa_uberblock.ub_version = version;
vdev_config_dirty(spa->spa_root_vdev);
spa_config_exit(spa, SCL_ALL, FTAG);
txg_wait_synced(spa_get_dsl(spa), 0);
}
static boolean_t
spa_has_aux_vdev(spa_t *spa, uint64_t guid, spa_aux_vdev_t *sav)
{
(void) spa;
int i;
uint64_t vdev_guid;
for (i = 0; i < sav->sav_count; i++)
if (sav->sav_vdevs[i]->vdev_guid == guid)
return (B_TRUE);
for (i = 0; i < sav->sav_npending; i++) {
if (nvlist_lookup_uint64(sav->sav_pending[i], ZPOOL_CONFIG_GUID,
&vdev_guid) == 0 && vdev_guid == guid)
return (B_TRUE);
}
return (B_FALSE);
}
boolean_t
spa_has_l2cache(spa_t *spa, uint64_t guid)
{
return (spa_has_aux_vdev(spa, guid, &spa->spa_l2cache));
}
boolean_t
spa_has_spare(spa_t *spa, uint64_t guid)
{
return (spa_has_aux_vdev(spa, guid, &spa->spa_spares));
}
/*
* Check if a pool has an active shared spare device.
* Note: reference count of an active spare is 2, as a spare and as a replace
*/
static boolean_t
spa_has_active_shared_spare(spa_t *spa)
{
int i, refcnt;
uint64_t pool;
spa_aux_vdev_t *sav = &spa->spa_spares;
for (i = 0; i < sav->sav_count; i++) {
if (spa_spare_exists(sav->sav_vdevs[i]->vdev_guid, &pool,
&refcnt) && pool != 0ULL && pool == spa_guid(spa) &&
refcnt > 2)
return (B_TRUE);
}
return (B_FALSE);
}
uint64_t
spa_total_metaslabs(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
uint64_t m = 0;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
if (!vdev_is_concrete(vd))
continue;
m += vd->vdev_ms_count;
}
return (m);
}
/*
* Notify any waiting threads that some activity has switched from being in-
* progress to not-in-progress so that the thread can wake up and determine
* whether it is finished waiting.
*/
void
spa_notify_waiters(spa_t *spa)
{
/*
* Acquiring spa_activities_lock here prevents the cv_broadcast from
* happening between the waiting thread's check and cv_wait.
*/
mutex_enter(&spa->spa_activities_lock);
cv_broadcast(&spa->spa_activities_cv);
mutex_exit(&spa->spa_activities_lock);
}
/*
* Notify any waiting threads that the pool is exporting, and then block until
* they are finished using the spa_t.
*/
void
spa_wake_waiters(spa_t *spa)
{
mutex_enter(&spa->spa_activities_lock);
spa->spa_waiters_cancel = B_TRUE;
cv_broadcast(&spa->spa_activities_cv);
while (spa->spa_waiters != 0)
cv_wait(&spa->spa_waiters_cv, &spa->spa_activities_lock);
spa->spa_waiters_cancel = B_FALSE;
mutex_exit(&spa->spa_activities_lock);
}
/* Whether the vdev or any of its descendants are being initialized/trimmed. */
static boolean_t
spa_vdev_activity_in_progress_impl(vdev_t *vd, zpool_wait_activity_t activity)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_config_held(spa, SCL_CONFIG | SCL_STATE, RW_READER));
ASSERT(MUTEX_HELD(&spa->spa_activities_lock));
ASSERT(activity == ZPOOL_WAIT_INITIALIZE ||
activity == ZPOOL_WAIT_TRIM);
kmutex_t *lock = activity == ZPOOL_WAIT_INITIALIZE ?
&vd->vdev_initialize_lock : &vd->vdev_trim_lock;
mutex_exit(&spa->spa_activities_lock);
mutex_enter(lock);
mutex_enter(&spa->spa_activities_lock);
boolean_t in_progress = (activity == ZPOOL_WAIT_INITIALIZE) ?
(vd->vdev_initialize_state == VDEV_INITIALIZE_ACTIVE) :
(vd->vdev_trim_state == VDEV_TRIM_ACTIVE);
mutex_exit(lock);
if (in_progress)
return (B_TRUE);
for (int i = 0; i < vd->vdev_children; i++) {
if (spa_vdev_activity_in_progress_impl(vd->vdev_child[i],
activity))
return (B_TRUE);
}
return (B_FALSE);
}
/*
* If use_guid is true, this checks whether the vdev specified by guid is
* being initialized/trimmed. Otherwise, it checks whether any vdev in the pool
* is being initialized/trimmed. The caller must hold the config lock and
* spa_activities_lock.
*/
static int
spa_vdev_activity_in_progress(spa_t *spa, boolean_t use_guid, uint64_t guid,
zpool_wait_activity_t activity, boolean_t *in_progress)
{
mutex_exit(&spa->spa_activities_lock);
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
mutex_enter(&spa->spa_activities_lock);
vdev_t *vd;
if (use_guid) {
vd = spa_lookup_by_guid(spa, guid, B_FALSE);
if (vd == NULL || !vd->vdev_ops->vdev_op_leaf) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (EINVAL);
}
} else {
vd = spa->spa_root_vdev;
}
*in_progress = spa_vdev_activity_in_progress_impl(vd, activity);
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (0);
}
/*
* Locking for waiting threads
* ---------------------------
*
* Waiting threads need a way to check whether a given activity is in progress,
* and then, if it is, wait for it to complete. Each activity will have some
* in-memory representation of the relevant on-disk state which can be used to
* determine whether or not the activity is in progress. The in-memory state and
* the locking used to protect it will be different for each activity, and may
* not be suitable for use with a cvar (e.g., some state is protected by the
* config lock). To allow waiting threads to wait without any races, another
* lock, spa_activities_lock, is used.
*
* When the state is checked, both the activity-specific lock (if there is one)
* and spa_activities_lock are held. In some cases, the activity-specific lock
* is acquired explicitly (e.g. the config lock). In others, the locking is
* internal to some check (e.g. bpobj_is_empty). After checking, the waiting
* thread releases the activity-specific lock and, if the activity is in
* progress, then cv_waits using spa_activities_lock.
*
* The waiting thread is woken when another thread, one completing some
* activity, updates the state of the activity and then calls
* spa_notify_waiters, which will cv_broadcast. This 'completing' thread only
* needs to hold its activity-specific lock when updating the state, and this
* lock can (but doesn't have to) be dropped before calling spa_notify_waiters.
*
* Because spa_notify_waiters acquires spa_activities_lock before broadcasting,
* and because it is held when the waiting thread checks the state of the
* activity, it can never be the case that the completing thread both updates
* the activity state and cv_broadcasts in between the waiting thread's check
* and cv_wait. Thus, a waiting thread can never miss a wakeup.
*
* In order to prevent deadlock, when the waiting thread does its check, in some
* cases it will temporarily drop spa_activities_lock in order to acquire the
* activity-specific lock. The order in which spa_activities_lock and the
* activity specific lock are acquired in the waiting thread is determined by
* the order in which they are acquired in the completing thread; if the
* completing thread calls spa_notify_waiters with the activity-specific lock
* held, then the waiting thread must also acquire the activity-specific lock
* first.
*/
static int
spa_activity_in_progress(spa_t *spa, zpool_wait_activity_t activity,
boolean_t use_tag, uint64_t tag, boolean_t *in_progress)
{
int error = 0;
ASSERT(MUTEX_HELD(&spa->spa_activities_lock));
switch (activity) {
case ZPOOL_WAIT_CKPT_DISCARD:
*in_progress =
(spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT) &&
zap_contains(spa_meta_objset(spa),
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ZPOOL_CHECKPOINT) ==
ENOENT);
break;
case ZPOOL_WAIT_FREE:
*in_progress = ((spa_version(spa) >= SPA_VERSION_DEADLISTS &&
!bpobj_is_empty(&spa->spa_dsl_pool->dp_free_bpobj)) ||
spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY) ||
spa_livelist_delete_check(spa));
break;
case ZPOOL_WAIT_INITIALIZE:
case ZPOOL_WAIT_TRIM:
error = spa_vdev_activity_in_progress(spa, use_tag, tag,
activity, in_progress);
break;
case ZPOOL_WAIT_REPLACE:
mutex_exit(&spa->spa_activities_lock);
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
mutex_enter(&spa->spa_activities_lock);
*in_progress = vdev_replace_in_progress(spa->spa_root_vdev);
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
break;
case ZPOOL_WAIT_REMOVE:
*in_progress = (spa->spa_removing_phys.sr_state ==
DSS_SCANNING);
break;
case ZPOOL_WAIT_RESILVER:
- if ((*in_progress = vdev_rebuild_active(spa->spa_root_vdev)))
+ *in_progress = vdev_rebuild_active(spa->spa_root_vdev);
+ if (*in_progress)
break;
zfs_fallthrough;
case ZPOOL_WAIT_SCRUB:
{
boolean_t scanning, paused, is_scrub;
dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan;
is_scrub = (scn->scn_phys.scn_func == POOL_SCAN_SCRUB);
scanning = (scn->scn_phys.scn_state == DSS_SCANNING);
paused = dsl_scan_is_paused_scrub(scn);
*in_progress = (scanning && !paused &&
is_scrub == (activity == ZPOOL_WAIT_SCRUB));
break;
}
+ case ZPOOL_WAIT_RAIDZ_EXPAND:
+ {
+ vdev_raidz_expand_t *vre = spa->spa_raidz_expand;
+ *in_progress = (vre != NULL && vre->vre_state == DSS_SCANNING);
+ break;
+ }
default:
panic("unrecognized value for activity %d", activity);
}
return (error);
}
static int
spa_wait_common(const char *pool, zpool_wait_activity_t activity,
boolean_t use_tag, uint64_t tag, boolean_t *waited)
{
/*
* The tag is used to distinguish between instances of an activity.
* 'initialize' and 'trim' are the only activities that we use this for.
* The other activities can only have a single instance in progress in a
* pool at one time, making the tag unnecessary.
*
* There can be multiple devices being replaced at once, but since they
* all finish once resilvering finishes, we don't bother keeping track
* of them individually, we just wait for them all to finish.
*/
if (use_tag && activity != ZPOOL_WAIT_INITIALIZE &&
activity != ZPOOL_WAIT_TRIM)
return (EINVAL);
if (activity < 0 || activity >= ZPOOL_WAIT_NUM_ACTIVITIES)
return (EINVAL);
spa_t *spa;
int error = spa_open(pool, &spa, FTAG);
if (error != 0)
return (error);
/*
* Increment the spa's waiter count so that we can call spa_close and
* still ensure that the spa_t doesn't get freed before this thread is
* finished with it when the pool is exported. We want to call spa_close
* before we start waiting because otherwise the additional ref would
* prevent the pool from being exported or destroyed throughout the
* potentially long wait.
*/
mutex_enter(&spa->spa_activities_lock);
spa->spa_waiters++;
spa_close(spa, FTAG);
*waited = B_FALSE;
for (;;) {
boolean_t in_progress;
error = spa_activity_in_progress(spa, activity, use_tag, tag,
&in_progress);
if (error || !in_progress || spa->spa_waiters_cancel)
break;
*waited = B_TRUE;
if (cv_wait_sig(&spa->spa_activities_cv,
&spa->spa_activities_lock) == 0) {
error = EINTR;
break;
}
}
spa->spa_waiters--;
cv_signal(&spa->spa_waiters_cv);
mutex_exit(&spa->spa_activities_lock);
return (error);
}
/*
* Wait for a particular instance of the specified activity to complete, where
* the instance is identified by 'tag'
*/
int
spa_wait_tag(const char *pool, zpool_wait_activity_t activity, uint64_t tag,
boolean_t *waited)
{
return (spa_wait_common(pool, activity, B_TRUE, tag, waited));
}
/*
* Wait for all instances of the specified activity complete
*/
int
spa_wait(const char *pool, zpool_wait_activity_t activity, boolean_t *waited)
{
return (spa_wait_common(pool, activity, B_FALSE, 0, waited));
}
sysevent_t *
spa_event_create(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl, const char *name)
{
sysevent_t *ev = NULL;
#ifdef _KERNEL
nvlist_t *resource;
resource = zfs_event_create(spa, vd, FM_SYSEVENT_CLASS, name, hist_nvl);
if (resource) {
ev = kmem_alloc(sizeof (sysevent_t), KM_SLEEP);
ev->resource = resource;
}
#else
(void) spa, (void) vd, (void) hist_nvl, (void) name;
#endif
return (ev);
}
void
spa_event_post(sysevent_t *ev)
{
#ifdef _KERNEL
if (ev) {
zfs_zevent_post(ev->resource, NULL, zfs_zevent_post_cb);
kmem_free(ev, sizeof (*ev));
}
#else
(void) ev;
#endif
}
/*
* Post a zevent corresponding to the given sysevent. The 'name' must be one
* of the event definitions in sys/sysevent/eventdefs.h. The payload will be
* filled in from the spa and (optionally) the vdev. This doesn't do anything
* in the userland libzpool, as we don't want consumers to misinterpret ztest
* or zdb as real changes.
*/
void
spa_event_notify(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl, const char *name)
{
spa_event_post(spa_event_create(spa, vd, hist_nvl, name));
}
/* state manipulation functions */
EXPORT_SYMBOL(spa_open);
EXPORT_SYMBOL(spa_open_rewind);
EXPORT_SYMBOL(spa_get_stats);
EXPORT_SYMBOL(spa_create);
EXPORT_SYMBOL(spa_import);
EXPORT_SYMBOL(spa_tryimport);
EXPORT_SYMBOL(spa_destroy);
EXPORT_SYMBOL(spa_export);
EXPORT_SYMBOL(spa_reset);
EXPORT_SYMBOL(spa_async_request);
EXPORT_SYMBOL(spa_async_suspend);
EXPORT_SYMBOL(spa_async_resume);
EXPORT_SYMBOL(spa_inject_addref);
EXPORT_SYMBOL(spa_inject_delref);
EXPORT_SYMBOL(spa_scan_stat_init);
EXPORT_SYMBOL(spa_scan_get_stats);
/* device manipulation */
EXPORT_SYMBOL(spa_vdev_add);
EXPORT_SYMBOL(spa_vdev_attach);
EXPORT_SYMBOL(spa_vdev_detach);
EXPORT_SYMBOL(spa_vdev_setpath);
EXPORT_SYMBOL(spa_vdev_setfru);
EXPORT_SYMBOL(spa_vdev_split_mirror);
/* spare statech is global across all pools) */
EXPORT_SYMBOL(spa_spare_add);
EXPORT_SYMBOL(spa_spare_remove);
EXPORT_SYMBOL(spa_spare_exists);
EXPORT_SYMBOL(spa_spare_activate);
/* L2ARC statech is global across all pools) */
EXPORT_SYMBOL(spa_l2cache_add);
EXPORT_SYMBOL(spa_l2cache_remove);
EXPORT_SYMBOL(spa_l2cache_exists);
EXPORT_SYMBOL(spa_l2cache_activate);
EXPORT_SYMBOL(spa_l2cache_drop);
/* scanning */
EXPORT_SYMBOL(spa_scan);
EXPORT_SYMBOL(spa_scan_stop);
/* spa syncing */
EXPORT_SYMBOL(spa_sync); /* only for DMU use */
EXPORT_SYMBOL(spa_sync_allpools);
/* properties */
EXPORT_SYMBOL(spa_prop_set);
EXPORT_SYMBOL(spa_prop_get);
EXPORT_SYMBOL(spa_prop_clear_bootfs);
/* asynchronous event notification */
EXPORT_SYMBOL(spa_event_notify);
ZFS_MODULE_PARAM(zfs_metaslab, metaslab_, preload_pct, UINT, ZMOD_RW,
"Percentage of CPUs to run a metaslab preload taskq");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs_spa, spa_, load_verify_shift, UINT, ZMOD_RW,
"log2 fraction of arc that can be used by inflight I/Os when "
"verifying pool during import");
/* END CSTYLED */
ZFS_MODULE_PARAM(zfs_spa, spa_, load_verify_metadata, INT, ZMOD_RW,
"Set to traverse metadata on pool import");
ZFS_MODULE_PARAM(zfs_spa, spa_, load_verify_data, INT, ZMOD_RW,
"Set to traverse data on pool import");
ZFS_MODULE_PARAM(zfs_spa, spa_, load_print_vdev_tree, INT, ZMOD_RW,
"Print vdev tree to zfs_dbgmsg during pool import");
ZFS_MODULE_PARAM(zfs_zio, zio_, taskq_batch_pct, UINT, ZMOD_RD,
"Percentage of CPUs to run an IO worker thread");
ZFS_MODULE_PARAM(zfs_zio, zio_, taskq_batch_tpq, UINT, ZMOD_RD,
"Number of threads per IO worker taskqueue");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs, zfs_, max_missing_tvds, U64, ZMOD_RW,
"Allow importing pool with up to this number of missing top-level "
"vdevs (in read-only mode)");
/* END CSTYLED */
ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, zthr_pause, INT,
ZMOD_RW, "Set the livelist condense zthr to pause");
ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, sync_pause, INT,
ZMOD_RW, "Set the livelist condense synctask to pause");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, sync_cancel,
INT, ZMOD_RW,
"Whether livelist condensing was canceled in the synctask");
ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, zthr_cancel,
INT, ZMOD_RW,
"Whether livelist condensing was canceled in the zthr function");
ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, new_alloc, INT,
ZMOD_RW,
"Whether extra ALLOC blkptrs were added to a livelist entry while it "
"was being condensed");
/* END CSTYLED */
ZFS_MODULE_PARAM(zfs_zio, zio_, taskq_wr_iss_ncpus, UINT, ZMOD_RW,
"Number of CPUs to run write issue taskqs");
diff --git a/sys/contrib/openzfs/module/zfs/spa_checkpoint.c b/sys/contrib/openzfs/module/zfs/spa_checkpoint.c
index b588f7041e5c..1efff47f87a0 100644
--- a/sys/contrib/openzfs/module/zfs/spa_checkpoint.c
+++ b/sys/contrib/openzfs/module/zfs/spa_checkpoint.c
@@ -1,637 +1,640 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2017 by Delphix. All rights reserved.
*/
/*
* Storage Pool Checkpoint
*
* A storage pool checkpoint can be thought of as a pool-wide snapshot or
* a stable version of extreme rewind that guarantees no blocks from the
* checkpointed state will have been overwritten. It remembers the entire
* state of the storage pool (e.g. snapshots, dataset names, etc..) from the
* point that it was taken and the user can rewind back to that point even if
* they applied destructive operations on their datasets or even enabled new
* zpool on-disk features. If a pool has a checkpoint that is no longer
* needed, the user can discard it.
*
* == On disk data structures used ==
*
* - The pool has a new feature flag and a new entry in the MOS. The feature
* flag is set to active when we create the checkpoint and remains active
* until the checkpoint is fully discarded. The entry in the MOS config
* (DMU_POOL_ZPOOL_CHECKPOINT) is populated with the uberblock that
* references the state of the pool when we take the checkpoint. The entry
* remains populated until we start discarding the checkpoint or we rewind
* back to it.
*
* - Each vdev contains a vdev-wide space map while the pool has a checkpoint,
* which persists until the checkpoint is fully discarded. The space map
* contains entries that have been freed in the current state of the pool
* but we want to keep around in case we decide to rewind to the checkpoint.
* [see vdev_checkpoint_sm]
*
* - Each metaslab's ms_sm space map behaves the same as without the
* checkpoint, with the only exception being the scenario when we free
* blocks that belong to the checkpoint. In this case, these blocks remain
* ALLOCATED in the metaslab's space map and they are added as FREE in the
* vdev's checkpoint space map.
*
* - Each uberblock has a field (ub_checkpoint_txg) which holds the txg that
* the uberblock was checkpointed. For normal uberblocks this field is 0.
*
* == Overview of operations ==
*
* - To create a checkpoint, we first wait for the current TXG to be synced,
* so we can use the most recently synced uberblock (spa_ubsync) as the
* checkpointed uberblock. Then we use an early synctask to place that
* uberblock in MOS config, increment the feature flag for the checkpoint
* (marking it active), and setting spa_checkpoint_txg (see its use below)
* to the TXG of the checkpointed uberblock. We use an early synctask for
* the aforementioned operations to ensure that no blocks were dirtied
* between the current TXG and the TXG of the checkpointed uberblock
* (e.g the previous txg).
*
* - When a checkpoint exists, we need to ensure that the blocks that
* belong to the checkpoint are freed but never reused. This means that
* these blocks should never end up in the ms_allocatable or the ms_freeing
* trees of a metaslab. Therefore, whenever there is a checkpoint the new
* ms_checkpointing tree is used in addition to the aforementioned ones.
*
* Whenever a block is freed and we find out that it is referenced by the
* checkpoint (we find out by comparing its birth to spa_checkpoint_txg),
* we place it in the ms_checkpointing tree instead of the ms_freeingtree.
* This way, we divide the blocks that are being freed into checkpointed
* and not-checkpointed blocks.
*
* In order to persist these frees, we write the extents from the
* ms_freeingtree to the ms_sm as usual, and the extents from the
* ms_checkpointing tree to the vdev_checkpoint_sm. This way, these
* checkpointed extents will remain allocated in the metaslab's ms_sm space
* map, and therefore won't be reused [see metaslab_sync()]. In addition,
* when we discard the checkpoint, we can find the entries that have
* actually been freed in vdev_checkpoint_sm.
* [see spa_checkpoint_discard_thread_sync()]
*
* - To discard the checkpoint we use an early synctask to delete the
* checkpointed uberblock from the MOS config, set spa_checkpoint_txg to 0,
* and wakeup the discarding zthr thread (an open-context async thread).
* We use an early synctask to ensure that the operation happens before any
* new data end up in the checkpoint's data structures.
*
* Once the synctask is done and the discarding zthr is awake, we discard
* the checkpointed data over multiple TXGs by having the zthr prefetching
* entries from vdev_checkpoint_sm and then starting a synctask that places
* them as free blocks into their respective ms_allocatable and ms_sm
* structures.
* [see spa_checkpoint_discard_thread()]
*
* When there are no entries left in the vdev_checkpoint_sm of all
* top-level vdevs, a final synctask runs that decrements the feature flag.
*
* - To rewind to the checkpoint, we first use the current uberblock and
* open the MOS so we can access the checkpointed uberblock from the MOS
* config. After we retrieve the checkpointed uberblock, we use it as the
* current uberblock for the pool by writing it to disk with an updated
* TXG, opening its version of the MOS, and moving on as usual from there.
* [see spa_ld_checkpoint_rewind()]
*
* An important note on rewinding to the checkpoint has to do with how we
* handle ZIL blocks. In the scenario of a rewind, we clear out any ZIL
* blocks that have not been claimed by the time we took the checkpoint
* as they should no longer be valid.
* [see comment in zil_claim()]
*
* == Miscellaneous information ==
*
* - In the hypothetical event that we take a checkpoint, remove a vdev,
* and attempt to rewind, the rewind would fail as the checkpointed
* uberblock would reference data in the removed device. For this reason
* and others of similar nature, we disallow the following operations that
* can change the config:
* vdev removal and attach/detach, mirror splitting, and pool reguid.
*
* - As most of the checkpoint logic is implemented in the SPA and doesn't
* distinguish datasets when it comes to space accounting, having a
* checkpoint can potentially break the boundaries set by dataset
* reservations.
*/
#include <sys/dmu_tx.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_synctask.h>
#include <sys/metaslab_impl.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/spa_checkpoint.h>
#include <sys/vdev_impl.h>
#include <sys/zap.h>
#include <sys/zfeature.h>
/*
* The following parameter limits the amount of memory to be used for the
* prefetching of the checkpoint space map done on each vdev while
* discarding the checkpoint.
*
* The reason it exists is because top-level vdevs with long checkpoint
* space maps can potentially take up a lot of memory depending on the
* amount of checkpointed data that has been freed within them while
* the pool had a checkpoint.
*/
static uint64_t zfs_spa_discard_memory_limit = 16 * 1024 * 1024;
int
spa_checkpoint_get_stats(spa_t *spa, pool_checkpoint_stat_t *pcs)
{
if (!spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT))
return (SET_ERROR(ZFS_ERR_NO_CHECKPOINT));
memset(pcs, 0, sizeof (pool_checkpoint_stat_t));
int error = zap_contains(spa_meta_objset(spa),
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ZPOOL_CHECKPOINT);
ASSERT(error == 0 || error == ENOENT);
if (error == ENOENT)
pcs->pcs_state = CS_CHECKPOINT_DISCARDING;
else
pcs->pcs_state = CS_CHECKPOINT_EXISTS;
pcs->pcs_space = spa->spa_checkpoint_info.sci_dspace;
pcs->pcs_start_time = spa->spa_checkpoint_info.sci_timestamp;
return (0);
}
static void
spa_checkpoint_discard_complete_sync(void *arg, dmu_tx_t *tx)
{
spa_t *spa = arg;
spa->spa_checkpoint_info.sci_timestamp = 0;
spa_feature_decr(spa, SPA_FEATURE_POOL_CHECKPOINT, tx);
spa_notify_waiters(spa);
spa_history_log_internal(spa, "spa discard checkpoint", tx,
"finished discarding checkpointed state from the pool");
}
typedef struct spa_checkpoint_discard_sync_callback_arg {
vdev_t *sdc_vd;
uint64_t sdc_txg;
uint64_t sdc_entry_limit;
} spa_checkpoint_discard_sync_callback_arg_t;
static int
spa_checkpoint_discard_sync_callback(space_map_entry_t *sme, void *arg)
{
spa_checkpoint_discard_sync_callback_arg_t *sdc = arg;
vdev_t *vd = sdc->sdc_vd;
metaslab_t *ms = vd->vdev_ms[sme->sme_offset >> vd->vdev_ms_shift];
uint64_t end = sme->sme_offset + sme->sme_run;
if (sdc->sdc_entry_limit == 0)
return (SET_ERROR(EINTR));
/*
* Since the space map is not condensed, we know that
* none of its entries is crossing the boundaries of
* its respective metaslab.
*
* That said, there is no fundamental requirement that
* the checkpoint's space map entries should not cross
* metaslab boundaries. So if needed we could add code
* that handles metaslab-crossing segments in the future.
*/
VERIFY3U(sme->sme_type, ==, SM_FREE);
VERIFY3U(sme->sme_offset, >=, ms->ms_start);
VERIFY3U(end, <=, ms->ms_start + ms->ms_size);
/*
* At this point we should not be processing any
* other frees concurrently, so the lock is technically
* unnecessary. We use the lock anyway though to
* potentially save ourselves from future headaches.
*/
mutex_enter(&ms->ms_lock);
if (range_tree_is_empty(ms->ms_freeing))
vdev_dirty(vd, VDD_METASLAB, ms, sdc->sdc_txg);
range_tree_add(ms->ms_freeing, sme->sme_offset, sme->sme_run);
mutex_exit(&ms->ms_lock);
ASSERT3U(vd->vdev_spa->spa_checkpoint_info.sci_dspace, >=,
sme->sme_run);
ASSERT3U(vd->vdev_stat.vs_checkpoint_space, >=, sme->sme_run);
vd->vdev_spa->spa_checkpoint_info.sci_dspace -= sme->sme_run;
vd->vdev_stat.vs_checkpoint_space -= sme->sme_run;
sdc->sdc_entry_limit--;
return (0);
}
#ifdef ZFS_DEBUG
static void
spa_checkpoint_accounting_verify(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
uint64_t ckpoint_sm_space_sum = 0;
uint64_t vs_ckpoint_space_sum = 0;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
if (vd->vdev_checkpoint_sm != NULL) {
ckpoint_sm_space_sum +=
-space_map_allocated(vd->vdev_checkpoint_sm);
vs_ckpoint_space_sum +=
vd->vdev_stat.vs_checkpoint_space;
ASSERT3U(ckpoint_sm_space_sum, ==,
vs_ckpoint_space_sum);
} else {
ASSERT0(vd->vdev_stat.vs_checkpoint_space);
}
}
ASSERT3U(spa->spa_checkpoint_info.sci_dspace, ==, ckpoint_sm_space_sum);
}
#endif
static void
spa_checkpoint_discard_thread_sync(void *arg, dmu_tx_t *tx)
{
vdev_t *vd = arg;
int error;
/*
* The space map callback is applied only to non-debug entries.
* Because the number of debug entries is less or equal to the
* number of non-debug entries, we want to ensure that we only
* read what we prefetched from open-context.
*
* Thus, we set the maximum entries that the space map callback
* will be applied to be half the entries that could fit in the
* imposed memory limit.
*
* Note that since this is a conservative estimate we also
* assume the worst case scenario in our computation where each
* entry is two-word.
*/
uint64_t max_entry_limit =
(zfs_spa_discard_memory_limit / (2 * sizeof (uint64_t))) >> 1;
/*
* Iterate from the end of the space map towards the beginning,
* placing its entries on ms_freeing and removing them from the
* space map. The iteration stops if one of the following
* conditions is true:
*
* 1] We reached the beginning of the space map. At this point
* the space map should be completely empty and
* space_map_incremental_destroy should have returned 0.
* The next step would be to free and close the space map
* and remove its entry from its vdev's top zap. This allows
* spa_checkpoint_discard_thread() to move on to the next vdev.
*
* 2] We reached the memory limit (amount of memory used to hold
* space map entries in memory) and space_map_incremental_destroy
* returned EINTR. This means that there are entries remaining
* in the space map that will be cleared in a future invocation
* of this function by spa_checkpoint_discard_thread().
*/
spa_checkpoint_discard_sync_callback_arg_t sdc;
sdc.sdc_vd = vd;
sdc.sdc_txg = tx->tx_txg;
sdc.sdc_entry_limit = max_entry_limit;
uint64_t words_before =
space_map_length(vd->vdev_checkpoint_sm) / sizeof (uint64_t);
error = space_map_incremental_destroy(vd->vdev_checkpoint_sm,
spa_checkpoint_discard_sync_callback, &sdc, tx);
uint64_t words_after =
space_map_length(vd->vdev_checkpoint_sm) / sizeof (uint64_t);
#ifdef ZFS_DEBUG
spa_checkpoint_accounting_verify(vd->vdev_spa);
#endif
zfs_dbgmsg("discarding checkpoint: txg %llu, vdev id %lld, "
"deleted %llu words - %llu words are left",
(u_longlong_t)tx->tx_txg, (longlong_t)vd->vdev_id,
(u_longlong_t)(words_before - words_after),
(u_longlong_t)words_after);
if (error != EINTR) {
if (error != 0) {
zfs_panic_recover("zfs: error %lld was returned "
"while incrementally destroying the checkpoint "
"space map of vdev %llu\n",
(longlong_t)error, vd->vdev_id);
}
ASSERT0(words_after);
ASSERT0(space_map_allocated(vd->vdev_checkpoint_sm));
ASSERT0(space_map_length(vd->vdev_checkpoint_sm));
space_map_free(vd->vdev_checkpoint_sm, tx);
space_map_close(vd->vdev_checkpoint_sm);
vd->vdev_checkpoint_sm = NULL;
VERIFY0(zap_remove(spa_meta_objset(vd->vdev_spa),
vd->vdev_top_zap, VDEV_TOP_ZAP_POOL_CHECKPOINT_SM, tx));
}
}
static boolean_t
spa_checkpoint_discard_is_done(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
ASSERT(!spa_has_checkpoint(spa));
ASSERT(spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT));
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
if (rvd->vdev_child[c]->vdev_checkpoint_sm != NULL)
return (B_FALSE);
ASSERT0(rvd->vdev_child[c]->vdev_stat.vs_checkpoint_space);
}
return (B_TRUE);
}
boolean_t
spa_checkpoint_discard_thread_check(void *arg, zthr_t *zthr)
{
(void) zthr;
spa_t *spa = arg;
if (!spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT))
return (B_FALSE);
if (spa_has_checkpoint(spa))
return (B_FALSE);
return (B_TRUE);
}
void
spa_checkpoint_discard_thread(void *arg, zthr_t *zthr)
{
spa_t *spa = arg;
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
while (vd->vdev_checkpoint_sm != NULL) {
space_map_t *checkpoint_sm = vd->vdev_checkpoint_sm;
int numbufs;
dmu_buf_t **dbp;
if (zthr_iscancelled(zthr))
return;
ASSERT3P(vd->vdev_ops, !=, &vdev_indirect_ops);
uint64_t size = MIN(space_map_length(checkpoint_sm),
zfs_spa_discard_memory_limit);
uint64_t offset =
space_map_length(checkpoint_sm) - size;
/*
* Ensure that the part of the space map that will
* be destroyed by the synctask, is prefetched in
* memory before the synctask runs.
*/
int error = dmu_buf_hold_array_by_bonus(
checkpoint_sm->sm_dbuf, offset, size,
B_TRUE, FTAG, &numbufs, &dbp);
if (error != 0) {
zfs_panic_recover("zfs: error %d was returned "
"while prefetching checkpoint space map "
"entries of vdev %llu\n",
error, vd->vdev_id);
}
VERIFY0(dsl_sync_task(spa->spa_name, NULL,
spa_checkpoint_discard_thread_sync, vd,
0, ZFS_SPACE_CHECK_NONE));
dmu_buf_rele_array(dbp, numbufs, FTAG);
}
}
VERIFY(spa_checkpoint_discard_is_done(spa));
VERIFY0(spa->spa_checkpoint_info.sci_dspace);
VERIFY0(dsl_sync_task(spa->spa_name, NULL,
spa_checkpoint_discard_complete_sync, spa,
0, ZFS_SPACE_CHECK_NONE));
}
static int
spa_checkpoint_check(void *arg, dmu_tx_t *tx)
{
(void) arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
if (!spa_feature_is_enabled(spa, SPA_FEATURE_POOL_CHECKPOINT))
return (SET_ERROR(ENOTSUP));
if (!spa_top_vdevs_spacemap_addressable(spa))
return (SET_ERROR(ZFS_ERR_VDEV_TOO_BIG));
if (spa->spa_removing_phys.sr_state == DSS_SCANNING)
return (SET_ERROR(ZFS_ERR_DEVRM_IN_PROGRESS));
+ if (spa->spa_raidz_expand != NULL)
+ return (SET_ERROR(ZFS_ERR_RAIDZ_EXPAND_IN_PROGRESS));
+
if (spa->spa_checkpoint_txg != 0)
return (SET_ERROR(ZFS_ERR_CHECKPOINT_EXISTS));
if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT))
return (SET_ERROR(ZFS_ERR_DISCARDING_CHECKPOINT));
return (0);
}
static void
spa_checkpoint_sync(void *arg, dmu_tx_t *tx)
{
(void) arg;
dsl_pool_t *dp = dmu_tx_pool(tx);
spa_t *spa = dp->dp_spa;
uberblock_t checkpoint = spa->spa_ubsync;
/*
* At this point, there should not be a checkpoint in the MOS.
*/
ASSERT3U(zap_contains(spa_meta_objset(spa), DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ZPOOL_CHECKPOINT), ==, ENOENT);
ASSERT0(spa->spa_checkpoint_info.sci_timestamp);
ASSERT0(spa->spa_checkpoint_info.sci_dspace);
/*
* Since the checkpointed uberblock is the one that just got synced
* (we use spa_ubsync), its txg must be equal to the txg number of
* the txg we are syncing, minus 1.
*/
ASSERT3U(checkpoint.ub_txg, ==, spa->spa_syncing_txg - 1);
/*
* Once the checkpoint is in place, we need to ensure that none of
* its blocks will be marked for reuse after it has been freed.
* When there is a checkpoint and a block is freed, we compare its
* birth txg to the txg of the checkpointed uberblock to see if the
* block is part of the checkpoint or not. Therefore, we have to set
* spa_checkpoint_txg before any frees happen in this txg (which is
* why this is done as an early_synctask as explained in the comment
* in spa_checkpoint()).
*/
spa->spa_checkpoint_txg = checkpoint.ub_txg;
spa->spa_checkpoint_info.sci_timestamp = checkpoint.ub_timestamp;
checkpoint.ub_checkpoint_txg = checkpoint.ub_txg;
VERIFY0(zap_add(spa->spa_dsl_pool->dp_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ZPOOL_CHECKPOINT,
sizeof (uint64_t), sizeof (uberblock_t) / sizeof (uint64_t),
&checkpoint, tx));
/*
* Increment the feature refcount and thus activate the feature.
* Note that the feature will be deactivated when we've
* completely discarded all checkpointed state (both vdev
* space maps and uberblock).
*/
spa_feature_incr(spa, SPA_FEATURE_POOL_CHECKPOINT, tx);
spa_history_log_internal(spa, "spa checkpoint", tx,
"checkpointed uberblock txg=%llu", (u_longlong_t)checkpoint.ub_txg);
}
/*
* Create a checkpoint for the pool.
*/
int
spa_checkpoint(const char *pool)
{
int error;
spa_t *spa;
error = spa_open(pool, &spa, FTAG);
if (error != 0)
return (error);
mutex_enter(&spa->spa_vdev_top_lock);
/*
* Wait for current syncing txg to finish so the latest synced
* uberblock (spa_ubsync) has all the changes that we expect
* to see if we were to revert later to the checkpoint. In other
* words we want the checkpointed uberblock to include/reference
* all the changes that were pending at the time that we issued
* the checkpoint command.
*/
txg_wait_synced(spa_get_dsl(spa), 0);
/*
* As the checkpointed uberblock references blocks from the previous
* txg (spa_ubsync) we want to ensure that are not freeing any of
* these blocks in the same txg that the following synctask will
* run. Thus, we run it as an early synctask, so the dirty changes
* that are synced to disk afterwards during zios and other synctasks
* do not reuse checkpointed blocks.
*/
error = dsl_early_sync_task(pool, spa_checkpoint_check,
spa_checkpoint_sync, NULL, 0, ZFS_SPACE_CHECK_NORMAL);
mutex_exit(&spa->spa_vdev_top_lock);
spa_close(spa, FTAG);
return (error);
}
static int
spa_checkpoint_discard_check(void *arg, dmu_tx_t *tx)
{
(void) arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
if (!spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT))
return (SET_ERROR(ZFS_ERR_NO_CHECKPOINT));
if (spa->spa_checkpoint_txg == 0)
return (SET_ERROR(ZFS_ERR_DISCARDING_CHECKPOINT));
VERIFY0(zap_contains(spa_meta_objset(spa),
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ZPOOL_CHECKPOINT));
return (0);
}
static void
spa_checkpoint_discard_sync(void *arg, dmu_tx_t *tx)
{
(void) arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
VERIFY0(zap_remove(spa_meta_objset(spa), DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ZPOOL_CHECKPOINT, tx));
spa->spa_checkpoint_txg = 0;
zthr_wakeup(spa->spa_checkpoint_discard_zthr);
spa_history_log_internal(spa, "spa discard checkpoint", tx,
"started discarding checkpointed state from the pool");
}
/*
* Discard the checkpoint from a pool.
*/
int
spa_checkpoint_discard(const char *pool)
{
/*
* Similarly to spa_checkpoint(), we want our synctask to run
* before any pending dirty data are written to disk so they
* won't end up in the checkpoint's data structures (e.g.
* ms_checkpointing and vdev_checkpoint_sm) and re-create any
* space maps that the discarding open-context thread has
* deleted.
* [see spa_discard_checkpoint_sync and spa_discard_checkpoint_thread]
*/
return (dsl_early_sync_task(pool, spa_checkpoint_discard_check,
spa_checkpoint_discard_sync, NULL, 0,
ZFS_SPACE_CHECK_DISCARD_CHECKPOINT));
}
EXPORT_SYMBOL(spa_checkpoint_get_stats);
EXPORT_SYMBOL(spa_checkpoint_discard_thread);
EXPORT_SYMBOL(spa_checkpoint_discard_thread_check);
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs_spa, zfs_spa_, discard_memory_limit, U64, ZMOD_RW,
"Limit for memory used in prefetching the checkpoint space map done "
"on each vdev while discarding the checkpoint");
/* END CSTYLED */
diff --git a/sys/contrib/openzfs/module/zfs/vdev.c b/sys/contrib/openzfs/module/zfs/vdev.c
index afb01c0ef7fd..c10c78ebf6db 100644
--- a/sys/contrib/openzfs/module/zfs/vdev.c
+++ b/sys/contrib/openzfs/module/zfs/vdev.c
@@ -1,6413 +1,6457 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2021 by Delphix. All rights reserved.
* Copyright 2017 Nexenta Systems, Inc.
* Copyright (c) 2014 Integros [integros.com]
* Copyright 2016 Toomas Soome <tsoome@me.com>
* Copyright 2017 Joyent, Inc.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019, Datto Inc. All rights reserved.
* Copyright (c) 2021, Klara Inc.
* Copyright (c) 2021, 2023 Hewlett Packard Enterprise Development LP.
*/
#include <sys/zfs_context.h>
#include <sys/fm/fs/zfs.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/bpobj.h>
#include <sys/dmu.h>
#include <sys/dmu_tx.h>
#include <sys/dsl_dir.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_rebuild.h>
#include <sys/vdev_draid.h>
#include <sys/uberblock_impl.h>
#include <sys/metaslab.h>
#include <sys/metaslab_impl.h>
#include <sys/space_map.h>
#include <sys/space_reftree.h>
#include <sys/zio.h>
#include <sys/zap.h>
#include <sys/fs/zfs.h>
#include <sys/arc.h>
#include <sys/zil.h>
#include <sys/dsl_scan.h>
#include <sys/vdev_raidz.h>
#include <sys/abd.h>
#include <sys/vdev_initialize.h>
#include <sys/vdev_trim.h>
+#include <sys/vdev_raidz.h>
#include <sys/zvol.h>
#include <sys/zfs_ratelimit.h>
#include "zfs_prop.h"
/*
* One metaslab from each (normal-class) vdev is used by the ZIL. These are
* called "embedded slog metaslabs", are referenced by vdev_log_mg, and are
* part of the spa_embedded_log_class. The metaslab with the most free space
* in each vdev is selected for this purpose when the pool is opened (or a
* vdev is added). See vdev_metaslab_init().
*
* Log blocks can be allocated from the following locations. Each one is tried
* in order until the allocation succeeds:
* 1. dedicated log vdevs, aka "slog" (spa_log_class)
* 2. embedded slog metaslabs (spa_embedded_log_class)
* 3. other metaslabs in normal vdevs (spa_normal_class)
*
* zfs_embedded_slog_min_ms disables the embedded slog if there are fewer
* than this number of metaslabs in the vdev. This ensures that we don't set
* aside an unreasonable amount of space for the ZIL. If set to less than
* 1 << (spa_slop_shift + 1), on small pools the usable space may be reduced
* (by more than 1<<spa_slop_shift) due to the embedded slog metaslab.
*/
static uint_t zfs_embedded_slog_min_ms = 64;
/* default target for number of metaslabs per top-level vdev */
static uint_t zfs_vdev_default_ms_count = 200;
/* minimum number of metaslabs per top-level vdev */
static uint_t zfs_vdev_min_ms_count = 16;
/* practical upper limit of total metaslabs per top-level vdev */
static uint_t zfs_vdev_ms_count_limit = 1ULL << 17;
/* lower limit for metaslab size (512M) */
static uint_t zfs_vdev_default_ms_shift = 29;
/* upper limit for metaslab size (16G) */
static uint_t zfs_vdev_max_ms_shift = 34;
int vdev_validate_skip = B_FALSE;
/*
* Since the DTL space map of a vdev is not expected to have a lot of
* entries, we default its block size to 4K.
*/
int zfs_vdev_dtl_sm_blksz = (1 << 12);
/*
* Rate limit slow IO (delay) events to this many per second.
*/
static unsigned int zfs_slow_io_events_per_second = 20;
/*
* Rate limit checksum events after this many checksum errors per second.
*/
static unsigned int zfs_checksum_events_per_second = 20;
/*
* Ignore errors during scrub/resilver. Allows to work around resilver
* upon import when there are pool errors.
*/
static int zfs_scan_ignore_errors = 0;
/*
* vdev-wide space maps that have lots of entries written to them at
* the end of each transaction can benefit from a higher I/O bandwidth
* (e.g. vdev_obsolete_sm), thus we default their block size to 128K.
*/
int zfs_vdev_standard_sm_blksz = (1 << 17);
/*
* Tunable parameter for debugging or performance analysis. Setting this
* will cause pool corruption on power loss if a volatile out-of-order
* write cache is enabled.
*/
int zfs_nocacheflush = 0;
/*
* Maximum and minimum ashift values that can be automatically set based on
* vdev's physical ashift (disk's physical sector size). While ASHIFT_MAX
* is higher than the maximum value, it is intentionally limited here to not
* excessively impact pool space efficiency. Higher ashift values may still
* be forced by vdev logical ashift or by user via ashift property, but won't
* be set automatically as a performance optimization.
*/
uint_t zfs_vdev_max_auto_ashift = 14;
uint_t zfs_vdev_min_auto_ashift = ASHIFT_MIN;
void
vdev_dbgmsg(vdev_t *vd, const char *fmt, ...)
{
va_list adx;
char buf[256];
va_start(adx, fmt);
(void) vsnprintf(buf, sizeof (buf), fmt, adx);
va_end(adx);
if (vd->vdev_path != NULL) {
zfs_dbgmsg("%s vdev '%s': %s", vd->vdev_ops->vdev_op_type,
vd->vdev_path, buf);
} else {
zfs_dbgmsg("%s-%llu vdev (guid %llu): %s",
vd->vdev_ops->vdev_op_type,
(u_longlong_t)vd->vdev_id,
(u_longlong_t)vd->vdev_guid, buf);
}
}
void
vdev_dbgmsg_print_tree(vdev_t *vd, int indent)
{
char state[20];
if (vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops) {
zfs_dbgmsg("%*svdev %llu: %s", indent, "",
(u_longlong_t)vd->vdev_id,
vd->vdev_ops->vdev_op_type);
return;
}
switch (vd->vdev_state) {
case VDEV_STATE_UNKNOWN:
(void) snprintf(state, sizeof (state), "unknown");
break;
case VDEV_STATE_CLOSED:
(void) snprintf(state, sizeof (state), "closed");
break;
case VDEV_STATE_OFFLINE:
(void) snprintf(state, sizeof (state), "offline");
break;
case VDEV_STATE_REMOVED:
(void) snprintf(state, sizeof (state), "removed");
break;
case VDEV_STATE_CANT_OPEN:
(void) snprintf(state, sizeof (state), "can't open");
break;
case VDEV_STATE_FAULTED:
(void) snprintf(state, sizeof (state), "faulted");
break;
case VDEV_STATE_DEGRADED:
(void) snprintf(state, sizeof (state), "degraded");
break;
case VDEV_STATE_HEALTHY:
(void) snprintf(state, sizeof (state), "healthy");
break;
default:
(void) snprintf(state, sizeof (state), "<state %u>",
(uint_t)vd->vdev_state);
}
zfs_dbgmsg("%*svdev %u: %s%s, guid: %llu, path: %s, %s", indent,
"", (int)vd->vdev_id, vd->vdev_ops->vdev_op_type,
vd->vdev_islog ? " (log)" : "",
(u_longlong_t)vd->vdev_guid,
vd->vdev_path ? vd->vdev_path : "N/A", state);
for (uint64_t i = 0; i < vd->vdev_children; i++)
vdev_dbgmsg_print_tree(vd->vdev_child[i], indent + 2);
}
/*
* Virtual device management.
*/
static vdev_ops_t *const vdev_ops_table[] = {
&vdev_root_ops,
&vdev_raidz_ops,
&vdev_draid_ops,
&vdev_draid_spare_ops,
&vdev_mirror_ops,
&vdev_replacing_ops,
&vdev_spare_ops,
&vdev_disk_ops,
&vdev_file_ops,
&vdev_missing_ops,
&vdev_hole_ops,
&vdev_indirect_ops,
NULL
};
/*
* Given a vdev type, return the appropriate ops vector.
*/
static vdev_ops_t *
vdev_getops(const char *type)
{
vdev_ops_t *ops, *const *opspp;
for (opspp = vdev_ops_table; (ops = *opspp) != NULL; opspp++)
if (strcmp(ops->vdev_op_type, type) == 0)
break;
return (ops);
}
/*
* Given a vdev and a metaslab class, find which metaslab group we're
* interested in. All vdevs may belong to two different metaslab classes.
* Dedicated slog devices use only the primary metaslab group, rather than a
* separate log group. For embedded slogs, the vdev_log_mg will be non-NULL.
*/
metaslab_group_t *
vdev_get_mg(vdev_t *vd, metaslab_class_t *mc)
{
if (mc == spa_embedded_log_class(vd->vdev_spa) &&
vd->vdev_log_mg != NULL)
return (vd->vdev_log_mg);
else
return (vd->vdev_mg);
}
void
vdev_default_xlate(vdev_t *vd, const range_seg64_t *logical_rs,
range_seg64_t *physical_rs, range_seg64_t *remain_rs)
{
(void) vd, (void) remain_rs;
physical_rs->rs_start = logical_rs->rs_start;
physical_rs->rs_end = logical_rs->rs_end;
}
/*
* Derive the enumerated allocation bias from string input.
* String origin is either the per-vdev zap or zpool(8).
*/
static vdev_alloc_bias_t
vdev_derive_alloc_bias(const char *bias)
{
vdev_alloc_bias_t alloc_bias = VDEV_BIAS_NONE;
if (strcmp(bias, VDEV_ALLOC_BIAS_LOG) == 0)
alloc_bias = VDEV_BIAS_LOG;
else if (strcmp(bias, VDEV_ALLOC_BIAS_SPECIAL) == 0)
alloc_bias = VDEV_BIAS_SPECIAL;
else if (strcmp(bias, VDEV_ALLOC_BIAS_DEDUP) == 0)
alloc_bias = VDEV_BIAS_DEDUP;
return (alloc_bias);
}
/*
* Default asize function: return the MAX of psize with the asize of
* all children. This is what's used by anything other than RAID-Z.
*/
uint64_t
-vdev_default_asize(vdev_t *vd, uint64_t psize)
+vdev_default_asize(vdev_t *vd, uint64_t psize, uint64_t txg)
{
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);
+ csize = vdev_psize_to_asize_txg(vd->vdev_child[c], psize, txg);
asize = MAX(asize, csize);
}
return (asize);
}
uint64_t
vdev_default_min_asize(vdev_t *vd)
{
return (vd->vdev_min_asize);
}
/*
* Get the minimum allocatable size. We define the allocatable size as
* the vdev's asize rounded to the nearest metaslab. This allows us to
* replace or attach devices which don't have the same physical size but
* can still satisfy the same number of allocations.
*/
uint64_t
vdev_get_min_asize(vdev_t *vd)
{
vdev_t *pvd = vd->vdev_parent;
/*
* If our parent is NULL (inactive spare or cache) or is the root,
* just return our own asize.
*/
if (pvd == NULL)
return (vd->vdev_asize);
/*
* The top-level vdev just returns the allocatable size rounded
* to the nearest metaslab.
*/
if (vd == vd->vdev_top)
return (P2ALIGN(vd->vdev_asize, 1ULL << vd->vdev_ms_shift));
return (pvd->vdev_ops->vdev_op_min_asize(pvd));
}
void
vdev_set_min_asize(vdev_t *vd)
{
vd->vdev_min_asize = vdev_get_min_asize(vd);
for (int c = 0; c < vd->vdev_children; c++)
vdev_set_min_asize(vd->vdev_child[c]);
}
/*
* Get the minimal allocation size for the top-level vdev.
*/
uint64_t
vdev_get_min_alloc(vdev_t *vd)
{
uint64_t min_alloc = 1ULL << vd->vdev_ashift;
if (vd->vdev_ops->vdev_op_min_alloc != NULL)
min_alloc = vd->vdev_ops->vdev_op_min_alloc(vd);
return (min_alloc);
}
/*
* Get the parity level for a top-level vdev.
*/
uint64_t
vdev_get_nparity(vdev_t *vd)
{
uint64_t nparity = 0;
if (vd->vdev_ops->vdev_op_nparity != NULL)
nparity = vd->vdev_ops->vdev_op_nparity(vd);
return (nparity);
}
static int
vdev_prop_get_int(vdev_t *vd, vdev_prop_t prop, uint64_t *value)
{
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa->spa_meta_objset;
uint64_t objid;
int err;
if (vd->vdev_root_zap != 0) {
objid = vd->vdev_root_zap;
} else if (vd->vdev_top_zap != 0) {
objid = vd->vdev_top_zap;
} else if (vd->vdev_leaf_zap != 0) {
objid = vd->vdev_leaf_zap;
} else {
return (EINVAL);
}
err = zap_lookup(mos, objid, vdev_prop_to_name(prop),
sizeof (uint64_t), 1, value);
if (err == ENOENT)
*value = vdev_prop_default_numeric(prop);
return (err);
}
/*
* Get the number of data disks for a top-level vdev.
*/
uint64_t
vdev_get_ndisks(vdev_t *vd)
{
uint64_t ndisks = 1;
if (vd->vdev_ops->vdev_op_ndisks != NULL)
ndisks = vd->vdev_ops->vdev_op_ndisks(vd);
return (ndisks);
}
vdev_t *
vdev_lookup_top(spa_t *spa, uint64_t vdev)
{
vdev_t *rvd = spa->spa_root_vdev;
ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0);
if (vdev < rvd->vdev_children) {
ASSERT(rvd->vdev_child[vdev] != NULL);
return (rvd->vdev_child[vdev]);
}
return (NULL);
}
vdev_t *
vdev_lookup_by_guid(vdev_t *vd, uint64_t guid)
{
vdev_t *mvd;
if (vd->vdev_guid == guid)
return (vd);
for (int c = 0; c < vd->vdev_children; c++)
if ((mvd = vdev_lookup_by_guid(vd->vdev_child[c], guid)) !=
NULL)
return (mvd);
return (NULL);
}
static int
vdev_count_leaves_impl(vdev_t *vd)
{
int n = 0;
if (vd->vdev_ops->vdev_op_leaf)
return (1);
for (int c = 0; c < vd->vdev_children; c++)
n += vdev_count_leaves_impl(vd->vdev_child[c]);
return (n);
}
int
vdev_count_leaves(spa_t *spa)
{
int rc;
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
rc = vdev_count_leaves_impl(spa->spa_root_vdev);
spa_config_exit(spa, SCL_VDEV, FTAG);
return (rc);
}
void
vdev_add_child(vdev_t *pvd, vdev_t *cvd)
{
size_t oldsize, newsize;
uint64_t id = cvd->vdev_id;
vdev_t **newchild;
ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL);
ASSERT(cvd->vdev_parent == NULL);
cvd->vdev_parent = pvd;
if (pvd == NULL)
return;
ASSERT(id >= pvd->vdev_children || pvd->vdev_child[id] == NULL);
oldsize = pvd->vdev_children * sizeof (vdev_t *);
pvd->vdev_children = MAX(pvd->vdev_children, id + 1);
newsize = pvd->vdev_children * sizeof (vdev_t *);
newchild = kmem_alloc(newsize, KM_SLEEP);
if (pvd->vdev_child != NULL) {
memcpy(newchild, pvd->vdev_child, oldsize);
kmem_free(pvd->vdev_child, oldsize);
}
pvd->vdev_child = newchild;
pvd->vdev_child[id] = cvd;
cvd->vdev_top = (pvd->vdev_top ? pvd->vdev_top: cvd);
ASSERT(cvd->vdev_top->vdev_parent->vdev_parent == NULL);
/*
* Walk up all ancestors to update guid sum.
*/
for (; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum += cvd->vdev_guid_sum;
if (cvd->vdev_ops->vdev_op_leaf) {
list_insert_head(&cvd->vdev_spa->spa_leaf_list, cvd);
cvd->vdev_spa->spa_leaf_list_gen++;
}
}
void
vdev_remove_child(vdev_t *pvd, vdev_t *cvd)
{
int c;
uint_t id = cvd->vdev_id;
ASSERT(cvd->vdev_parent == pvd);
if (pvd == NULL)
return;
ASSERT(id < pvd->vdev_children);
ASSERT(pvd->vdev_child[id] == cvd);
pvd->vdev_child[id] = NULL;
cvd->vdev_parent = NULL;
for (c = 0; c < pvd->vdev_children; c++)
if (pvd->vdev_child[c])
break;
if (c == pvd->vdev_children) {
kmem_free(pvd->vdev_child, c * sizeof (vdev_t *));
pvd->vdev_child = NULL;
pvd->vdev_children = 0;
}
if (cvd->vdev_ops->vdev_op_leaf) {
spa_t *spa = cvd->vdev_spa;
list_remove(&spa->spa_leaf_list, cvd);
spa->spa_leaf_list_gen++;
}
/*
* Walk up all ancestors to update guid sum.
*/
for (; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum -= cvd->vdev_guid_sum;
}
/*
* Remove any holes in the child array.
*/
void
vdev_compact_children(vdev_t *pvd)
{
vdev_t **newchild, *cvd;
int oldc = pvd->vdev_children;
int newc;
ASSERT(spa_config_held(pvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL);
if (oldc == 0)
return;
for (int c = newc = 0; c < oldc; c++)
if (pvd->vdev_child[c])
newc++;
if (newc > 0) {
newchild = kmem_zalloc(newc * sizeof (vdev_t *), KM_SLEEP);
for (int c = newc = 0; c < oldc; c++) {
if ((cvd = pvd->vdev_child[c]) != NULL) {
newchild[newc] = cvd;
cvd->vdev_id = newc++;
}
}
} else {
newchild = NULL;
}
kmem_free(pvd->vdev_child, oldc * sizeof (vdev_t *));
pvd->vdev_child = newchild;
pvd->vdev_children = newc;
}
/*
* Allocate and minimally initialize a vdev_t.
*/
vdev_t *
vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid, vdev_ops_t *ops)
{
vdev_t *vd;
vdev_indirect_config_t *vic;
vd = kmem_zalloc(sizeof (vdev_t), KM_SLEEP);
vic = &vd->vdev_indirect_config;
if (spa->spa_root_vdev == NULL) {
ASSERT(ops == &vdev_root_ops);
spa->spa_root_vdev = vd;
spa->spa_load_guid = spa_generate_guid(NULL);
}
if (guid == 0 && ops != &vdev_hole_ops) {
if (spa->spa_root_vdev == vd) {
/*
* The root vdev's guid will also be the pool guid,
* which must be unique among all pools.
*/
guid = spa_generate_guid(NULL);
} else {
/*
* Any other vdev's guid must be unique within the pool.
*/
guid = spa_generate_guid(spa);
}
ASSERT(!spa_guid_exists(spa_guid(spa), guid));
}
vd->vdev_spa = spa;
vd->vdev_id = id;
vd->vdev_guid = guid;
vd->vdev_guid_sum = guid;
vd->vdev_ops = ops;
vd->vdev_state = VDEV_STATE_CLOSED;
vd->vdev_ishole = (ops == &vdev_hole_ops);
vic->vic_prev_indirect_vdev = UINT64_MAX;
rw_init(&vd->vdev_indirect_rwlock, NULL, RW_DEFAULT, NULL);
mutex_init(&vd->vdev_obsolete_lock, NULL, MUTEX_DEFAULT, NULL);
vd->vdev_obsolete_segments = range_tree_create(NULL, RANGE_SEG64, NULL,
0, 0);
/*
* Initialize rate limit structs for events. We rate limit ZIO delay
* and checksum events so that we don't overwhelm ZED with thousands
* of events when a disk is acting up.
*/
zfs_ratelimit_init(&vd->vdev_delay_rl, &zfs_slow_io_events_per_second,
1);
zfs_ratelimit_init(&vd->vdev_deadman_rl, &zfs_slow_io_events_per_second,
1);
zfs_ratelimit_init(&vd->vdev_checksum_rl,
&zfs_checksum_events_per_second, 1);
/*
* Default Thresholds for tuning ZED
*/
vd->vdev_checksum_n = vdev_prop_default_numeric(VDEV_PROP_CHECKSUM_N);
vd->vdev_checksum_t = vdev_prop_default_numeric(VDEV_PROP_CHECKSUM_T);
vd->vdev_io_n = vdev_prop_default_numeric(VDEV_PROP_IO_N);
vd->vdev_io_t = vdev_prop_default_numeric(VDEV_PROP_IO_T);
list_link_init(&vd->vdev_config_dirty_node);
list_link_init(&vd->vdev_state_dirty_node);
list_link_init(&vd->vdev_initialize_node);
list_link_init(&vd->vdev_leaf_node);
list_link_init(&vd->vdev_trim_node);
mutex_init(&vd->vdev_dtl_lock, NULL, MUTEX_NOLOCKDEP, NULL);
mutex_init(&vd->vdev_stat_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_probe_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_scan_io_queue_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_initialize_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_initialize_io_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&vd->vdev_initialize_cv, NULL, CV_DEFAULT, NULL);
cv_init(&vd->vdev_initialize_io_cv, NULL, CV_DEFAULT, NULL);
mutex_init(&vd->vdev_trim_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_autotrim_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_trim_io_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&vd->vdev_trim_cv, NULL, CV_DEFAULT, NULL);
cv_init(&vd->vdev_autotrim_cv, NULL, CV_DEFAULT, NULL);
cv_init(&vd->vdev_autotrim_kick_cv, NULL, CV_DEFAULT, NULL);
cv_init(&vd->vdev_trim_io_cv, NULL, CV_DEFAULT, NULL);
mutex_init(&vd->vdev_rebuild_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&vd->vdev_rebuild_cv, NULL, CV_DEFAULT, NULL);
for (int t = 0; t < DTL_TYPES; t++) {
vd->vdev_dtl[t] = range_tree_create(NULL, RANGE_SEG64, NULL, 0,
0);
}
txg_list_create(&vd->vdev_ms_list, spa,
offsetof(struct metaslab, ms_txg_node));
txg_list_create(&vd->vdev_dtl_list, spa,
offsetof(struct vdev, vdev_dtl_node));
vd->vdev_stat.vs_timestamp = gethrtime();
vdev_queue_init(vd);
return (vd);
}
/*
* Allocate a new vdev. The 'alloctype' is used to control whether we are
* creating a new vdev or loading an existing one - the behavior is slightly
* different for each case.
*/
int
vdev_alloc(spa_t *spa, vdev_t **vdp, nvlist_t *nv, vdev_t *parent, uint_t id,
int alloctype)
{
vdev_ops_t *ops;
const char *type;
uint64_t guid = 0, islog;
vdev_t *vd;
vdev_indirect_config_t *vic;
const char *tmp = NULL;
int rc;
vdev_alloc_bias_t alloc_bias = VDEV_BIAS_NONE;
boolean_t top_level = (parent && !parent->vdev_parent);
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) != 0)
return (SET_ERROR(EINVAL));
if ((ops = vdev_getops(type)) == NULL)
return (SET_ERROR(EINVAL));
/*
* If this is a load, get the vdev guid from the nvlist.
* Otherwise, vdev_alloc_common() will generate one for us.
*/
if (alloctype == VDEV_ALLOC_LOAD) {
uint64_t label_id;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID, &label_id) ||
label_id != id)
return (SET_ERROR(EINVAL));
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0)
return (SET_ERROR(EINVAL));
} else if (alloctype == VDEV_ALLOC_SPARE) {
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0)
return (SET_ERROR(EINVAL));
} else if (alloctype == VDEV_ALLOC_L2CACHE) {
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0)
return (SET_ERROR(EINVAL));
} else if (alloctype == VDEV_ALLOC_ROOTPOOL) {
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0)
return (SET_ERROR(EINVAL));
}
/*
* The first allocated vdev must be of type 'root'.
*/
if (ops != &vdev_root_ops && spa->spa_root_vdev == NULL)
return (SET_ERROR(EINVAL));
/*
* Determine whether we're a log vdev.
*/
islog = 0;
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_LOG, &islog);
if (islog && spa_version(spa) < SPA_VERSION_SLOGS)
return (SET_ERROR(ENOTSUP));
if (ops == &vdev_hole_ops && spa_version(spa) < SPA_VERSION_HOLES)
return (SET_ERROR(ENOTSUP));
if (top_level && alloctype == VDEV_ALLOC_ADD) {
const char *bias;
/*
* If creating a top-level vdev, check for allocation
* classes input.
*/
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_ALLOCATION_BIAS,
&bias) == 0) {
alloc_bias = vdev_derive_alloc_bias(bias);
/* spa_vdev_add() expects feature to be enabled */
if (spa->spa_load_state != SPA_LOAD_CREATE &&
!spa_feature_is_enabled(spa,
SPA_FEATURE_ALLOCATION_CLASSES)) {
return (SET_ERROR(ENOTSUP));
}
}
/* spa_vdev_add() expects feature to be enabled */
if (ops == &vdev_draid_ops &&
spa->spa_load_state != SPA_LOAD_CREATE &&
!spa_feature_is_enabled(spa, SPA_FEATURE_DRAID)) {
return (SET_ERROR(ENOTSUP));
}
}
/*
* Initialize the vdev specific data. This is done before calling
* vdev_alloc_common() since it may fail and this simplifies the
* error reporting and cleanup code paths.
*/
void *tsd = NULL;
if (ops->vdev_op_init != NULL) {
rc = ops->vdev_op_init(spa, nv, &tsd);
if (rc != 0) {
return (rc);
}
}
vd = vdev_alloc_common(spa, id, guid, ops);
vd->vdev_tsd = tsd;
vd->vdev_islog = islog;
if (top_level && alloc_bias != VDEV_BIAS_NONE)
vd->vdev_alloc_bias = alloc_bias;
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &tmp) == 0)
vd->vdev_path = spa_strdup(tmp);
/*
* ZPOOL_CONFIG_AUX_STATE = "external" means we previously forced a
* fault on a vdev and want it to persist across imports (like with
* zpool offline -f).
*/
rc = nvlist_lookup_string(nv, ZPOOL_CONFIG_AUX_STATE, &tmp);
if (rc == 0 && tmp != NULL && strcmp(tmp, "external") == 0) {
vd->vdev_stat.vs_aux = VDEV_AUX_EXTERNAL;
vd->vdev_faulted = 1;
vd->vdev_label_aux = VDEV_AUX_EXTERNAL;
}
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_DEVID, &tmp) == 0)
vd->vdev_devid = spa_strdup(tmp);
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PHYS_PATH, &tmp) == 0)
vd->vdev_physpath = spa_strdup(tmp);
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH,
&tmp) == 0)
vd->vdev_enc_sysfs_path = spa_strdup(tmp);
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_FRU, &tmp) == 0)
vd->vdev_fru = spa_strdup(tmp);
/*
* Set the whole_disk property. If it's not specified, leave the value
* as -1.
*/
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK,
&vd->vdev_wholedisk) != 0)
vd->vdev_wholedisk = -1ULL;
vic = &vd->vdev_indirect_config;
ASSERT0(vic->vic_mapping_object);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_INDIRECT_OBJECT,
&vic->vic_mapping_object);
ASSERT0(vic->vic_births_object);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_INDIRECT_BIRTHS,
&vic->vic_births_object);
ASSERT3U(vic->vic_prev_indirect_vdev, ==, UINT64_MAX);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_PREV_INDIRECT_VDEV,
&vic->vic_prev_indirect_vdev);
/*
* Look for the 'not present' flag. This will only be set if the device
* was not present at the time of import.
*/
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT,
&vd->vdev_not_present);
/*
* Get the alignment requirement. Ignore pool ashift for vdev
* attach case.
*/
if (alloctype != VDEV_ALLOC_ATTACH) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASHIFT,
&vd->vdev_ashift);
} else {
vd->vdev_attaching = B_TRUE;
}
/*
* Retrieve the vdev creation time.
*/
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_CREATE_TXG,
&vd->vdev_crtxg);
if (vd->vdev_ops == &vdev_root_ops &&
(alloctype == VDEV_ALLOC_LOAD ||
alloctype == VDEV_ALLOC_SPLIT ||
alloctype == VDEV_ALLOC_ROOTPOOL)) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_VDEV_ROOT_ZAP,
&vd->vdev_root_zap);
}
/*
* If we're a top-level vdev, try to load the allocation parameters.
*/
if (top_level &&
(alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_SPLIT)) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_METASLAB_ARRAY,
&vd->vdev_ms_array);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_METASLAB_SHIFT,
&vd->vdev_ms_shift);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASIZE,
&vd->vdev_asize);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NONALLOCATING,
&vd->vdev_noalloc);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVING,
&vd->vdev_removing);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_VDEV_TOP_ZAP,
&vd->vdev_top_zap);
+ vd->vdev_rz_expanding = nvlist_exists(nv,
+ ZPOOL_CONFIG_RAIDZ_EXPANDING);
} else {
ASSERT0(vd->vdev_top_zap);
}
if (top_level && alloctype != VDEV_ALLOC_ATTACH) {
ASSERT(alloctype == VDEV_ALLOC_LOAD ||
alloctype == VDEV_ALLOC_ADD ||
alloctype == VDEV_ALLOC_SPLIT ||
alloctype == VDEV_ALLOC_ROOTPOOL);
/* Note: metaslab_group_create() is now deferred */
}
if (vd->vdev_ops->vdev_op_leaf &&
(alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_SPLIT)) {
(void) nvlist_lookup_uint64(nv,
ZPOOL_CONFIG_VDEV_LEAF_ZAP, &vd->vdev_leaf_zap);
} else {
ASSERT0(vd->vdev_leaf_zap);
}
/*
* If we're a leaf vdev, try to load the DTL object and other state.
*/
if (vd->vdev_ops->vdev_op_leaf &&
(alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_L2CACHE ||
alloctype == VDEV_ALLOC_ROOTPOOL)) {
if (alloctype == VDEV_ALLOC_LOAD) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DTL,
&vd->vdev_dtl_object);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_UNSPARE,
&vd->vdev_unspare);
}
if (alloctype == VDEV_ALLOC_ROOTPOOL) {
uint64_t spare = 0;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_SPARE,
&spare) == 0 && spare)
spa_spare_add(vd);
}
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_OFFLINE,
&vd->vdev_offline);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_RESILVER_TXG,
&vd->vdev_resilver_txg);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REBUILD_TXG,
&vd->vdev_rebuild_txg);
if (nvlist_exists(nv, ZPOOL_CONFIG_RESILVER_DEFER))
vdev_defer_resilver(vd);
/*
* In general, when importing a pool we want to ignore the
* persistent fault state, as the diagnosis made on another
* system may not be valid in the current context. The only
* exception is if we forced a vdev to a persistently faulted
* state with 'zpool offline -f'. The persistent fault will
* remain across imports until cleared.
*
* Local vdevs will remain in the faulted state.
*/
if (spa_load_state(spa) == SPA_LOAD_OPEN ||
spa_load_state(spa) == SPA_LOAD_IMPORT) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_FAULTED,
&vd->vdev_faulted);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DEGRADED,
&vd->vdev_degraded);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVED,
&vd->vdev_removed);
if (vd->vdev_faulted || vd->vdev_degraded) {
const char *aux;
vd->vdev_label_aux =
VDEV_AUX_ERR_EXCEEDED;
if (nvlist_lookup_string(nv,
ZPOOL_CONFIG_AUX_STATE, &aux) == 0 &&
strcmp(aux, "external") == 0)
vd->vdev_label_aux = VDEV_AUX_EXTERNAL;
else
vd->vdev_faulted = 0ULL;
}
}
}
/*
* Add ourselves to the parent's list of children.
*/
vdev_add_child(parent, vd);
*vdp = vd;
return (0);
}
void
vdev_free(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT3P(vd->vdev_initialize_thread, ==, NULL);
ASSERT3P(vd->vdev_trim_thread, ==, NULL);
ASSERT3P(vd->vdev_autotrim_thread, ==, NULL);
ASSERT3P(vd->vdev_rebuild_thread, ==, NULL);
/*
* Scan queues are normally destroyed at the end of a scan. If the
* queue exists here, that implies the vdev is being removed while
* the scan is still running.
*/
if (vd->vdev_scan_io_queue != NULL) {
mutex_enter(&vd->vdev_scan_io_queue_lock);
dsl_scan_io_queue_destroy(vd->vdev_scan_io_queue);
vd->vdev_scan_io_queue = NULL;
mutex_exit(&vd->vdev_scan_io_queue_lock);
}
/*
* vdev_free() implies closing the vdev first. This is simpler than
* trying to ensure complicated semantics for all callers.
*/
vdev_close(vd);
ASSERT(!list_link_active(&vd->vdev_config_dirty_node));
ASSERT(!list_link_active(&vd->vdev_state_dirty_node));
/*
* Free all children.
*/
for (int c = 0; c < vd->vdev_children; c++)
vdev_free(vd->vdev_child[c]);
ASSERT(vd->vdev_child == NULL);
ASSERT(vd->vdev_guid_sum == vd->vdev_guid);
if (vd->vdev_ops->vdev_op_fini != NULL)
vd->vdev_ops->vdev_op_fini(vd);
/*
* Discard allocation state.
*/
if (vd->vdev_mg != NULL) {
vdev_metaslab_fini(vd);
metaslab_group_destroy(vd->vdev_mg);
vd->vdev_mg = NULL;
}
if (vd->vdev_log_mg != NULL) {
ASSERT0(vd->vdev_ms_count);
metaslab_group_destroy(vd->vdev_log_mg);
vd->vdev_log_mg = NULL;
}
ASSERT0(vd->vdev_stat.vs_space);
ASSERT0(vd->vdev_stat.vs_dspace);
ASSERT0(vd->vdev_stat.vs_alloc);
/*
* Remove this vdev from its parent's child list.
*/
vdev_remove_child(vd->vdev_parent, vd);
ASSERT(vd->vdev_parent == NULL);
ASSERT(!list_link_active(&vd->vdev_leaf_node));
/*
* Clean up vdev structure.
*/
vdev_queue_fini(vd);
if (vd->vdev_path)
spa_strfree(vd->vdev_path);
if (vd->vdev_devid)
spa_strfree(vd->vdev_devid);
if (vd->vdev_physpath)
spa_strfree(vd->vdev_physpath);
if (vd->vdev_enc_sysfs_path)
spa_strfree(vd->vdev_enc_sysfs_path);
if (vd->vdev_fru)
spa_strfree(vd->vdev_fru);
if (vd->vdev_isspare)
spa_spare_remove(vd);
if (vd->vdev_isl2cache)
spa_l2cache_remove(vd);
txg_list_destroy(&vd->vdev_ms_list);
txg_list_destroy(&vd->vdev_dtl_list);
mutex_enter(&vd->vdev_dtl_lock);
space_map_close(vd->vdev_dtl_sm);
for (int t = 0; t < DTL_TYPES; t++) {
range_tree_vacate(vd->vdev_dtl[t], NULL, NULL);
range_tree_destroy(vd->vdev_dtl[t]);
}
mutex_exit(&vd->vdev_dtl_lock);
EQUIV(vd->vdev_indirect_births != NULL,
vd->vdev_indirect_mapping != NULL);
if (vd->vdev_indirect_births != NULL) {
vdev_indirect_mapping_close(vd->vdev_indirect_mapping);
vdev_indirect_births_close(vd->vdev_indirect_births);
}
if (vd->vdev_obsolete_sm != NULL) {
ASSERT(vd->vdev_removing ||
vd->vdev_ops == &vdev_indirect_ops);
space_map_close(vd->vdev_obsolete_sm);
vd->vdev_obsolete_sm = NULL;
}
range_tree_destroy(vd->vdev_obsolete_segments);
rw_destroy(&vd->vdev_indirect_rwlock);
mutex_destroy(&vd->vdev_obsolete_lock);
mutex_destroy(&vd->vdev_dtl_lock);
mutex_destroy(&vd->vdev_stat_lock);
mutex_destroy(&vd->vdev_probe_lock);
mutex_destroy(&vd->vdev_scan_io_queue_lock);
mutex_destroy(&vd->vdev_initialize_lock);
mutex_destroy(&vd->vdev_initialize_io_lock);
cv_destroy(&vd->vdev_initialize_io_cv);
cv_destroy(&vd->vdev_initialize_cv);
mutex_destroy(&vd->vdev_trim_lock);
mutex_destroy(&vd->vdev_autotrim_lock);
mutex_destroy(&vd->vdev_trim_io_lock);
cv_destroy(&vd->vdev_trim_cv);
cv_destroy(&vd->vdev_autotrim_cv);
cv_destroy(&vd->vdev_autotrim_kick_cv);
cv_destroy(&vd->vdev_trim_io_cv);
mutex_destroy(&vd->vdev_rebuild_lock);
cv_destroy(&vd->vdev_rebuild_cv);
zfs_ratelimit_fini(&vd->vdev_delay_rl);
zfs_ratelimit_fini(&vd->vdev_deadman_rl);
zfs_ratelimit_fini(&vd->vdev_checksum_rl);
if (vd == spa->spa_root_vdev)
spa->spa_root_vdev = NULL;
kmem_free(vd, sizeof (vdev_t));
}
/*
* Transfer top-level vdev state from svd to tvd.
*/
static void
vdev_top_transfer(vdev_t *svd, vdev_t *tvd)
{
spa_t *spa = svd->vdev_spa;
metaslab_t *msp;
vdev_t *vd;
int t;
ASSERT(tvd == tvd->vdev_top);
tvd->vdev_ms_array = svd->vdev_ms_array;
tvd->vdev_ms_shift = svd->vdev_ms_shift;
tvd->vdev_ms_count = svd->vdev_ms_count;
tvd->vdev_top_zap = svd->vdev_top_zap;
svd->vdev_ms_array = 0;
svd->vdev_ms_shift = 0;
svd->vdev_ms_count = 0;
svd->vdev_top_zap = 0;
if (tvd->vdev_mg)
ASSERT3P(tvd->vdev_mg, ==, svd->vdev_mg);
if (tvd->vdev_log_mg)
ASSERT3P(tvd->vdev_log_mg, ==, svd->vdev_log_mg);
tvd->vdev_mg = svd->vdev_mg;
tvd->vdev_log_mg = svd->vdev_log_mg;
tvd->vdev_ms = svd->vdev_ms;
svd->vdev_mg = NULL;
svd->vdev_log_mg = NULL;
svd->vdev_ms = NULL;
if (tvd->vdev_mg != NULL)
tvd->vdev_mg->mg_vd = tvd;
if (tvd->vdev_log_mg != NULL)
tvd->vdev_log_mg->mg_vd = tvd;
tvd->vdev_checkpoint_sm = svd->vdev_checkpoint_sm;
svd->vdev_checkpoint_sm = NULL;
tvd->vdev_alloc_bias = svd->vdev_alloc_bias;
svd->vdev_alloc_bias = VDEV_BIAS_NONE;
tvd->vdev_stat.vs_alloc = svd->vdev_stat.vs_alloc;
tvd->vdev_stat.vs_space = svd->vdev_stat.vs_space;
tvd->vdev_stat.vs_dspace = svd->vdev_stat.vs_dspace;
svd->vdev_stat.vs_alloc = 0;
svd->vdev_stat.vs_space = 0;
svd->vdev_stat.vs_dspace = 0;
/*
* State which may be set on a top-level vdev that's in the
* process of being removed.
*/
ASSERT0(tvd->vdev_indirect_config.vic_births_object);
ASSERT0(tvd->vdev_indirect_config.vic_mapping_object);
ASSERT3U(tvd->vdev_indirect_config.vic_prev_indirect_vdev, ==, -1ULL);
ASSERT3P(tvd->vdev_indirect_mapping, ==, NULL);
ASSERT3P(tvd->vdev_indirect_births, ==, NULL);
ASSERT3P(tvd->vdev_obsolete_sm, ==, NULL);
ASSERT0(tvd->vdev_noalloc);
ASSERT0(tvd->vdev_removing);
ASSERT0(tvd->vdev_rebuilding);
tvd->vdev_noalloc = svd->vdev_noalloc;
tvd->vdev_removing = svd->vdev_removing;
tvd->vdev_rebuilding = svd->vdev_rebuilding;
tvd->vdev_rebuild_config = svd->vdev_rebuild_config;
tvd->vdev_indirect_config = svd->vdev_indirect_config;
tvd->vdev_indirect_mapping = svd->vdev_indirect_mapping;
tvd->vdev_indirect_births = svd->vdev_indirect_births;
range_tree_swap(&svd->vdev_obsolete_segments,
&tvd->vdev_obsolete_segments);
tvd->vdev_obsolete_sm = svd->vdev_obsolete_sm;
svd->vdev_indirect_config.vic_mapping_object = 0;
svd->vdev_indirect_config.vic_births_object = 0;
svd->vdev_indirect_config.vic_prev_indirect_vdev = -1ULL;
svd->vdev_indirect_mapping = NULL;
svd->vdev_indirect_births = NULL;
svd->vdev_obsolete_sm = NULL;
svd->vdev_noalloc = 0;
svd->vdev_removing = 0;
svd->vdev_rebuilding = 0;
for (t = 0; t < TXG_SIZE; t++) {
while ((msp = txg_list_remove(&svd->vdev_ms_list, t)) != NULL)
(void) txg_list_add(&tvd->vdev_ms_list, msp, t);
while ((vd = txg_list_remove(&svd->vdev_dtl_list, t)) != NULL)
(void) txg_list_add(&tvd->vdev_dtl_list, vd, t);
if (txg_list_remove_this(&spa->spa_vdev_txg_list, svd, t))
(void) txg_list_add(&spa->spa_vdev_txg_list, tvd, t);
}
if (list_link_active(&svd->vdev_config_dirty_node)) {
vdev_config_clean(svd);
vdev_config_dirty(tvd);
}
if (list_link_active(&svd->vdev_state_dirty_node)) {
vdev_state_clean(svd);
vdev_state_dirty(tvd);
}
tvd->vdev_deflate_ratio = svd->vdev_deflate_ratio;
svd->vdev_deflate_ratio = 0;
tvd->vdev_islog = svd->vdev_islog;
svd->vdev_islog = 0;
dsl_scan_io_queue_vdev_xfer(svd, tvd);
}
static void
vdev_top_update(vdev_t *tvd, vdev_t *vd)
{
if (vd == NULL)
return;
vd->vdev_top = tvd;
for (int c = 0; c < vd->vdev_children; c++)
vdev_top_update(tvd, vd->vdev_child[c]);
}
/*
* Add a mirror/replacing vdev above an existing vdev. There is no need to
* call .vdev_op_init() since mirror/replacing vdevs do not have private state.
*/
vdev_t *
vdev_add_parent(vdev_t *cvd, vdev_ops_t *ops)
{
spa_t *spa = cvd->vdev_spa;
vdev_t *pvd = cvd->vdev_parent;
vdev_t *mvd;
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
mvd = vdev_alloc_common(spa, cvd->vdev_id, 0, ops);
mvd->vdev_asize = cvd->vdev_asize;
mvd->vdev_min_asize = cvd->vdev_min_asize;
mvd->vdev_max_asize = cvd->vdev_max_asize;
mvd->vdev_psize = cvd->vdev_psize;
mvd->vdev_ashift = cvd->vdev_ashift;
mvd->vdev_logical_ashift = cvd->vdev_logical_ashift;
mvd->vdev_physical_ashift = cvd->vdev_physical_ashift;
mvd->vdev_state = cvd->vdev_state;
mvd->vdev_crtxg = cvd->vdev_crtxg;
vdev_remove_child(pvd, cvd);
vdev_add_child(pvd, mvd);
cvd->vdev_id = mvd->vdev_children;
vdev_add_child(mvd, cvd);
vdev_top_update(cvd->vdev_top, cvd->vdev_top);
if (mvd == mvd->vdev_top)
vdev_top_transfer(cvd, mvd);
return (mvd);
}
/*
* Remove a 1-way mirror/replacing vdev from the tree.
*/
void
vdev_remove_parent(vdev_t *cvd)
{
vdev_t *mvd = cvd->vdev_parent;
vdev_t *pvd = mvd->vdev_parent;
ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL);
ASSERT(mvd->vdev_children == 1);
ASSERT(mvd->vdev_ops == &vdev_mirror_ops ||
mvd->vdev_ops == &vdev_replacing_ops ||
mvd->vdev_ops == &vdev_spare_ops);
cvd->vdev_ashift = mvd->vdev_ashift;
cvd->vdev_logical_ashift = mvd->vdev_logical_ashift;
cvd->vdev_physical_ashift = mvd->vdev_physical_ashift;
vdev_remove_child(mvd, cvd);
vdev_remove_child(pvd, mvd);
/*
* If cvd will replace mvd as a top-level vdev, preserve mvd's guid.
* Otherwise, we could have detached an offline device, and when we
* go to import the pool we'll think we have two top-level vdevs,
* instead of a different version of the same top-level vdev.
*/
if (mvd->vdev_top == mvd) {
uint64_t guid_delta = mvd->vdev_guid - cvd->vdev_guid;
cvd->vdev_orig_guid = cvd->vdev_guid;
cvd->vdev_guid += guid_delta;
cvd->vdev_guid_sum += guid_delta;
/*
* If pool not set for autoexpand, we need to also preserve
* mvd's asize to prevent automatic expansion of cvd.
* Otherwise if we are adjusting the mirror by attaching and
* detaching children of non-uniform sizes, the mirror could
* autoexpand, unexpectedly requiring larger devices to
* re-establish the mirror.
*/
if (!cvd->vdev_spa->spa_autoexpand)
cvd->vdev_asize = mvd->vdev_asize;
}
cvd->vdev_id = mvd->vdev_id;
vdev_add_child(pvd, cvd);
vdev_top_update(cvd->vdev_top, cvd->vdev_top);
if (cvd == cvd->vdev_top)
vdev_top_transfer(mvd, cvd);
ASSERT(mvd->vdev_children == 0);
vdev_free(mvd);
}
/*
* Choose GCD for spa_gcd_alloc.
*/
static uint64_t
vdev_gcd(uint64_t a, uint64_t b)
{
while (b != 0) {
uint64_t t = b;
b = a % b;
a = t;
}
return (a);
}
/*
* Set spa_min_alloc and spa_gcd_alloc.
*/
static void
vdev_spa_set_alloc(spa_t *spa, uint64_t min_alloc)
{
if (min_alloc < spa->spa_min_alloc)
spa->spa_min_alloc = min_alloc;
if (spa->spa_gcd_alloc == INT_MAX) {
spa->spa_gcd_alloc = min_alloc;
} else {
spa->spa_gcd_alloc = vdev_gcd(min_alloc,
spa->spa_gcd_alloc);
}
}
void
vdev_metaslab_group_create(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
/*
* metaslab_group_create was delayed until allocation bias was available
*/
if (vd->vdev_mg == NULL) {
metaslab_class_t *mc;
if (vd->vdev_islog && vd->vdev_alloc_bias == VDEV_BIAS_NONE)
vd->vdev_alloc_bias = VDEV_BIAS_LOG;
ASSERT3U(vd->vdev_islog, ==,
(vd->vdev_alloc_bias == VDEV_BIAS_LOG));
switch (vd->vdev_alloc_bias) {
case VDEV_BIAS_LOG:
mc = spa_log_class(spa);
break;
case VDEV_BIAS_SPECIAL:
mc = spa_special_class(spa);
break;
case VDEV_BIAS_DEDUP:
mc = spa_dedup_class(spa);
break;
default:
mc = spa_normal_class(spa);
}
vd->vdev_mg = metaslab_group_create(mc, vd,
spa->spa_alloc_count);
if (!vd->vdev_islog) {
vd->vdev_log_mg = metaslab_group_create(
spa_embedded_log_class(spa), vd, 1);
}
/*
* The spa ashift min/max only apply for the normal metaslab
* class. Class destination is late binding so ashift boundary
* setting had to wait until now.
*/
if (vd->vdev_top == vd && vd->vdev_ashift != 0 &&
mc == spa_normal_class(spa) && vd->vdev_aux == NULL) {
if (vd->vdev_ashift > spa->spa_max_ashift)
spa->spa_max_ashift = vd->vdev_ashift;
if (vd->vdev_ashift < spa->spa_min_ashift)
spa->spa_min_ashift = vd->vdev_ashift;
uint64_t min_alloc = vdev_get_min_alloc(vd);
vdev_spa_set_alloc(spa, min_alloc);
}
}
}
int
vdev_metaslab_init(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
uint64_t oldc = vd->vdev_ms_count;
uint64_t newc = vd->vdev_asize >> vd->vdev_ms_shift;
metaslab_t **mspp;
int error;
boolean_t expanding = (oldc != 0);
ASSERT(txg == 0 || spa_config_held(spa, SCL_ALLOC, RW_WRITER));
/*
* This vdev is not being allocated from yet or is a hole.
*/
if (vd->vdev_ms_shift == 0)
return (0);
ASSERT(!vd->vdev_ishole);
ASSERT(oldc <= newc);
mspp = vmem_zalloc(newc * sizeof (*mspp), KM_SLEEP);
if (expanding) {
memcpy(mspp, vd->vdev_ms, oldc * sizeof (*mspp));
vmem_free(vd->vdev_ms, oldc * sizeof (*mspp));
}
vd->vdev_ms = mspp;
vd->vdev_ms_count = newc;
for (uint64_t m = oldc; m < newc; m++) {
uint64_t object = 0;
/*
* vdev_ms_array may be 0 if we are creating the "fake"
* metaslabs for an indirect vdev for zdb's leak detection.
* See zdb_leak_init().
*/
if (txg == 0 && vd->vdev_ms_array != 0) {
error = dmu_read(spa->spa_meta_objset,
vd->vdev_ms_array,
m * sizeof (uint64_t), sizeof (uint64_t), &object,
DMU_READ_PREFETCH);
if (error != 0) {
vdev_dbgmsg(vd, "unable to read the metaslab "
"array [error=%d]", error);
return (error);
}
}
error = metaslab_init(vd->vdev_mg, m, object, txg,
&(vd->vdev_ms[m]));
if (error != 0) {
vdev_dbgmsg(vd, "metaslab_init failed [error=%d]",
error);
return (error);
}
}
/*
* Find the emptiest metaslab on the vdev and mark it for use for
* embedded slog by moving it from the regular to the log metaslab
* group.
*/
if (vd->vdev_mg->mg_class == spa_normal_class(spa) &&
vd->vdev_ms_count > zfs_embedded_slog_min_ms &&
avl_is_empty(&vd->vdev_log_mg->mg_metaslab_tree)) {
uint64_t slog_msid = 0;
uint64_t smallest = UINT64_MAX;
/*
* Note, we only search the new metaslabs, because the old
* (pre-existing) ones may be active (e.g. have non-empty
* range_tree's), and we don't move them to the new
* metaslab_t.
*/
for (uint64_t m = oldc; m < newc; m++) {
uint64_t alloc =
space_map_allocated(vd->vdev_ms[m]->ms_sm);
if (alloc < smallest) {
slog_msid = m;
smallest = alloc;
}
}
metaslab_t *slog_ms = vd->vdev_ms[slog_msid];
/*
* The metaslab was marked as dirty at the end of
* metaslab_init(). Remove it from the dirty list so that we
* can uninitialize and reinitialize it to the new class.
*/
if (txg != 0) {
(void) txg_list_remove_this(&vd->vdev_ms_list,
slog_ms, txg);
}
uint64_t sm_obj = space_map_object(slog_ms->ms_sm);
metaslab_fini(slog_ms);
VERIFY0(metaslab_init(vd->vdev_log_mg, slog_msid, sm_obj, txg,
&vd->vdev_ms[slog_msid]));
}
if (txg == 0)
spa_config_enter(spa, SCL_ALLOC, FTAG, RW_WRITER);
/*
* If the vdev is marked as non-allocating then don't
* activate the metaslabs since we want to ensure that
* no allocations are performed on this device.
*/
if (vd->vdev_noalloc) {
/* track non-allocating vdev space */
spa->spa_nonallocating_dspace += spa_deflate(spa) ?
vd->vdev_stat.vs_dspace : vd->vdev_stat.vs_space;
} else if (!expanding) {
metaslab_group_activate(vd->vdev_mg);
if (vd->vdev_log_mg != NULL)
metaslab_group_activate(vd->vdev_log_mg);
}
if (txg == 0)
spa_config_exit(spa, SCL_ALLOC, FTAG);
return (0);
}
void
vdev_metaslab_fini(vdev_t *vd)
{
if (vd->vdev_checkpoint_sm != NULL) {
ASSERT(spa_feature_is_active(vd->vdev_spa,
SPA_FEATURE_POOL_CHECKPOINT));
space_map_close(vd->vdev_checkpoint_sm);
/*
* Even though we close the space map, we need to set its
* pointer to NULL. The reason is that vdev_metaslab_fini()
* may be called multiple times for certain operations
* (i.e. when destroying a pool) so we need to ensure that
* this clause never executes twice. This logic is similar
* to the one used for the vdev_ms clause below.
*/
vd->vdev_checkpoint_sm = NULL;
}
if (vd->vdev_ms != NULL) {
metaslab_group_t *mg = vd->vdev_mg;
metaslab_group_passivate(mg);
if (vd->vdev_log_mg != NULL) {
ASSERT(!vd->vdev_islog);
metaslab_group_passivate(vd->vdev_log_mg);
}
uint64_t count = vd->vdev_ms_count;
for (uint64_t m = 0; m < count; m++) {
metaslab_t *msp = vd->vdev_ms[m];
if (msp != NULL)
metaslab_fini(msp);
}
vmem_free(vd->vdev_ms, count * sizeof (metaslab_t *));
vd->vdev_ms = NULL;
vd->vdev_ms_count = 0;
for (int i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++) {
ASSERT0(mg->mg_histogram[i]);
if (vd->vdev_log_mg != NULL)
ASSERT0(vd->vdev_log_mg->mg_histogram[i]);
}
}
ASSERT0(vd->vdev_ms_count);
}
typedef struct vdev_probe_stats {
boolean_t vps_readable;
boolean_t vps_writeable;
int vps_flags;
} vdev_probe_stats_t;
static void
vdev_probe_done(zio_t *zio)
{
spa_t *spa = zio->io_spa;
vdev_t *vd = zio->io_vd;
vdev_probe_stats_t *vps = zio->io_private;
ASSERT(vd->vdev_probe_zio != NULL);
if (zio->io_type == ZIO_TYPE_READ) {
if (zio->io_error == 0)
vps->vps_readable = 1;
if (zio->io_error == 0 && spa_writeable(spa)) {
zio_nowait(zio_write_phys(vd->vdev_probe_zio, vd,
zio->io_offset, zio->io_size, zio->io_abd,
ZIO_CHECKSUM_OFF, vdev_probe_done, vps,
ZIO_PRIORITY_SYNC_WRITE, vps->vps_flags, B_TRUE));
} else {
abd_free(zio->io_abd);
}
} else if (zio->io_type == ZIO_TYPE_WRITE) {
if (zio->io_error == 0)
vps->vps_writeable = 1;
abd_free(zio->io_abd);
} else if (zio->io_type == ZIO_TYPE_NULL) {
zio_t *pio;
zio_link_t *zl;
vd->vdev_cant_read |= !vps->vps_readable;
vd->vdev_cant_write |= !vps->vps_writeable;
+ vdev_dbgmsg(vd, "probe done, cant_read=%u cant_write=%u",
+ vd->vdev_cant_read, vd->vdev_cant_write);
if (vdev_readable(vd) &&
(vdev_writeable(vd) || !spa_writeable(spa))) {
zio->io_error = 0;
} else {
ASSERT(zio->io_error != 0);
vdev_dbgmsg(vd, "failed probe");
(void) zfs_ereport_post(FM_EREPORT_ZFS_PROBE_FAILURE,
spa, vd, NULL, NULL, 0);
zio->io_error = SET_ERROR(ENXIO);
}
mutex_enter(&vd->vdev_probe_lock);
ASSERT(vd->vdev_probe_zio == zio);
vd->vdev_probe_zio = NULL;
mutex_exit(&vd->vdev_probe_lock);
zl = NULL;
while ((pio = zio_walk_parents(zio, &zl)) != NULL)
if (!vdev_accessible(vd, pio))
pio->io_error = SET_ERROR(ENXIO);
kmem_free(vps, sizeof (*vps));
}
}
/*
* Determine whether this device is accessible.
*
* Read and write to several known locations: the pad regions of each
* vdev label but the first, which we leave alone in case it contains
* a VTOC.
*/
zio_t *
vdev_probe(vdev_t *vd, zio_t *zio)
{
spa_t *spa = vd->vdev_spa;
vdev_probe_stats_t *vps = NULL;
zio_t *pio;
ASSERT(vd->vdev_ops->vdev_op_leaf);
/*
* Don't probe the probe.
*/
if (zio && (zio->io_flags & ZIO_FLAG_PROBE))
return (NULL);
/*
* To prevent 'probe storms' when a device fails, we create
* just one probe i/o at a time. All zios that want to probe
* this vdev will become parents of the probe io.
*/
mutex_enter(&vd->vdev_probe_lock);
if ((pio = vd->vdev_probe_zio) == NULL) {
vps = kmem_zalloc(sizeof (*vps), KM_SLEEP);
vps->vps_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_PROBE |
ZIO_FLAG_DONT_AGGREGATE | ZIO_FLAG_TRYHARD;
if (spa_config_held(spa, SCL_ZIO, RW_WRITER)) {
/*
* vdev_cant_read and vdev_cant_write can only
* transition from TRUE to FALSE when we have the
* SCL_ZIO lock as writer; otherwise they can only
* transition from FALSE to TRUE. This ensures that
* any zio looking at these values can assume that
* failures persist for the life of the I/O. That's
* important because when a device has intermittent
* connectivity problems, we want to ensure that
* they're ascribed to the device (ENXIO) and not
* the zio (EIO).
*
* Since we hold SCL_ZIO as writer here, clear both
* values so the probe can reevaluate from first
* principles.
*/
vps->vps_flags |= ZIO_FLAG_CONFIG_WRITER;
vd->vdev_cant_read = B_FALSE;
vd->vdev_cant_write = B_FALSE;
}
vd->vdev_probe_zio = pio = zio_null(NULL, spa, vd,
vdev_probe_done, vps,
vps->vps_flags | ZIO_FLAG_DONT_PROPAGATE);
/*
* We can't change the vdev state in this context, so we
* kick off an async task to do it on our behalf.
*/
if (zio != NULL) {
vd->vdev_probe_wanted = B_TRUE;
spa_async_request(spa, SPA_ASYNC_PROBE);
}
}
if (zio != NULL)
zio_add_child(zio, pio);
mutex_exit(&vd->vdev_probe_lock);
if (vps == NULL) {
ASSERT(zio != NULL);
return (NULL);
}
for (int l = 1; l < VDEV_LABELS; l++) {
zio_nowait(zio_read_phys(pio, vd,
vdev_label_offset(vd->vdev_psize, l,
offsetof(vdev_label_t, vl_be)), VDEV_PAD_SIZE,
abd_alloc_for_io(VDEV_PAD_SIZE, B_TRUE),
ZIO_CHECKSUM_OFF, vdev_probe_done, vps,
ZIO_PRIORITY_SYNC_READ, vps->vps_flags, B_TRUE));
}
if (zio == NULL)
return (pio);
zio_nowait(pio);
return (NULL);
}
static void
vdev_load_child(void *arg)
{
vdev_t *vd = arg;
vd->vdev_load_error = vdev_load(vd);
}
static void
vdev_open_child(void *arg)
{
vdev_t *vd = arg;
vd->vdev_open_thread = curthread;
vd->vdev_open_error = vdev_open(vd);
vd->vdev_open_thread = NULL;
}
static boolean_t
vdev_uses_zvols(vdev_t *vd)
{
#ifdef _KERNEL
if (zvol_is_zvol(vd->vdev_path))
return (B_TRUE);
#endif
for (int c = 0; c < vd->vdev_children; c++)
if (vdev_uses_zvols(vd->vdev_child[c]))
return (B_TRUE);
return (B_FALSE);
}
/*
* Returns B_TRUE if the passed child should be opened.
*/
static boolean_t
vdev_default_open_children_func(vdev_t *vd)
{
(void) vd;
return (B_TRUE);
}
/*
* Open the requested child vdevs. If any of the leaf vdevs are using
* a ZFS volume then do the opens in a single thread. This avoids a
* deadlock when the current thread is holding the spa_namespace_lock.
*/
static void
vdev_open_children_impl(vdev_t *vd, vdev_open_children_func_t *open_func)
{
int children = vd->vdev_children;
taskq_t *tq = taskq_create("vdev_open", children, minclsyspri,
children, children, TASKQ_PREPOPULATE);
vd->vdev_nonrot = B_TRUE;
for (int c = 0; c < children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (open_func(cvd) == B_FALSE)
continue;
if (tq == NULL || vdev_uses_zvols(vd)) {
cvd->vdev_open_error = vdev_open(cvd);
} else {
VERIFY(taskq_dispatch(tq, vdev_open_child,
cvd, TQ_SLEEP) != TASKQID_INVALID);
}
vd->vdev_nonrot &= cvd->vdev_nonrot;
}
if (tq != NULL) {
taskq_wait(tq);
taskq_destroy(tq);
}
}
/*
* Open all child vdevs.
*/
void
vdev_open_children(vdev_t *vd)
{
vdev_open_children_impl(vd, vdev_default_open_children_func);
}
/*
* Conditionally open a subset of child vdevs.
*/
void
vdev_open_children_subset(vdev_t *vd, vdev_open_children_func_t *open_func)
{
vdev_open_children_impl(vd, open_func);
}
/*
- * Compute the raidz-deflation ratio. Note, we hard-code
- * in 128k (1 << 17) because it is the "typical" blocksize.
- * Even though SPA_MAXBLOCKSIZE changed, this algorithm can not change,
- * otherwise it would inconsistently account for existing bp's.
+ * Compute the raidz-deflation ratio. Note, we hard-code 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. We also hard-code txg 0 for the same reason
+ * since expanded RAIDZ vdevs can use a different asize for different birth
+ * txg'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);
+ (vdev_psize_to_asize_txg(vd, 1 << 17, 0) >>
+ SPA_MINBLOCKSHIFT);
}
}
/*
* Choose the best of two ashifts, preferring one between logical ashift
* (absolute minimum) and administrator defined maximum, otherwise take
* the biggest of the two.
*/
uint64_t
vdev_best_ashift(uint64_t logical, uint64_t a, uint64_t b)
{
if (a > logical && a <= zfs_vdev_max_auto_ashift) {
if (b <= logical || b > zfs_vdev_max_auto_ashift)
return (a);
else
return (MAX(a, b));
} else if (b <= logical || b > zfs_vdev_max_auto_ashift)
return (MAX(a, b));
return (b);
}
/*
* Maximize performance by inflating the configured ashift for top level
* vdevs to be as close to the physical ashift as possible while maintaining
* administrator defined limits and ensuring it doesn't go below the
* logical ashift.
*/
static void
vdev_ashift_optimize(vdev_t *vd)
{
ASSERT(vd == vd->vdev_top);
if (vd->vdev_ashift < vd->vdev_physical_ashift &&
vd->vdev_physical_ashift <= zfs_vdev_max_auto_ashift) {
vd->vdev_ashift = MIN(
MAX(zfs_vdev_max_auto_ashift, vd->vdev_ashift),
MAX(zfs_vdev_min_auto_ashift,
vd->vdev_physical_ashift));
} else {
/*
* If the logical and physical ashifts are the same, then
* we ensure that the top-level vdev's ashift is not smaller
* than our minimum ashift value. For the unusual case
* where logical ashift > physical ashift, we can't cap
* the calculated ashift based on max ashift as that
* would cause failures.
* We still check if we need to increase it to match
* the min ashift.
*/
vd->vdev_ashift = MAX(zfs_vdev_min_auto_ashift,
vd->vdev_ashift);
}
}
/*
* Prepare a virtual device for access.
*/
int
vdev_open(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
int error;
uint64_t osize = 0;
uint64_t max_osize = 0;
uint64_t asize, max_asize, psize;
uint64_t logical_ashift = 0;
uint64_t physical_ashift = 0;
ASSERT(vd->vdev_open_thread == curthread ||
spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
ASSERT(vd->vdev_state == VDEV_STATE_CLOSED ||
vd->vdev_state == VDEV_STATE_CANT_OPEN ||
vd->vdev_state == VDEV_STATE_OFFLINE);
vd->vdev_stat.vs_aux = VDEV_AUX_NONE;
vd->vdev_cant_read = B_FALSE;
vd->vdev_cant_write = B_FALSE;
vd->vdev_min_asize = vdev_get_min_asize(vd);
/*
* If this vdev is not removed, check its fault status. If it's
* faulted, bail out of the open.
*/
if (!vd->vdev_removed && vd->vdev_faulted) {
ASSERT(vd->vdev_children == 0);
ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED ||
vd->vdev_label_aux == VDEV_AUX_EXTERNAL);
vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED,
vd->vdev_label_aux);
return (SET_ERROR(ENXIO));
} else if (vd->vdev_offline) {
ASSERT(vd->vdev_children == 0);
vdev_set_state(vd, B_TRUE, VDEV_STATE_OFFLINE, VDEV_AUX_NONE);
return (SET_ERROR(ENXIO));
}
error = vd->vdev_ops->vdev_op_open(vd, &osize, &max_osize,
&logical_ashift, &physical_ashift);
/* Keep the device in removed state if unplugged */
if (error == ENOENT && vd->vdev_removed) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_REMOVED,
VDEV_AUX_NONE);
return (error);
}
/*
* Physical volume size should never be larger than its max size, unless
* the disk has shrunk while we were reading it or the device is buggy
* or damaged: either way it's not safe for use, bail out of the open.
*/
if (osize > max_osize) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_OPEN_FAILED);
return (SET_ERROR(ENXIO));
}
/*
* Reset the vdev_reopening flag so that we actually close
* the vdev on error.
*/
vd->vdev_reopening = B_FALSE;
if (zio_injection_enabled && error == 0)
error = zio_handle_device_injection(vd, NULL, SET_ERROR(ENXIO));
if (error) {
if (vd->vdev_removed &&
vd->vdev_stat.vs_aux != VDEV_AUX_OPEN_FAILED)
vd->vdev_removed = B_FALSE;
if (vd->vdev_stat.vs_aux == VDEV_AUX_CHILDREN_OFFLINE) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_OFFLINE,
vd->vdev_stat.vs_aux);
} else {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
vd->vdev_stat.vs_aux);
}
return (error);
}
vd->vdev_removed = B_FALSE;
/*
* Recheck the faulted flag now that we have confirmed that
* the vdev is accessible. If we're faulted, bail.
*/
if (vd->vdev_faulted) {
ASSERT(vd->vdev_children == 0);
ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED ||
vd->vdev_label_aux == VDEV_AUX_EXTERNAL);
vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED,
vd->vdev_label_aux);
return (SET_ERROR(ENXIO));
}
if (vd->vdev_degraded) {
ASSERT(vd->vdev_children == 0);
vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED,
VDEV_AUX_ERR_EXCEEDED);
} else {
vdev_set_state(vd, B_TRUE, VDEV_STATE_HEALTHY, 0);
}
/*
* For hole or missing vdevs we just return success.
*/
if (vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops)
return (0);
for (int c = 0; c < vd->vdev_children; c++) {
if (vd->vdev_child[c]->vdev_state != VDEV_STATE_HEALTHY) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED,
VDEV_AUX_NONE);
break;
}
}
osize = P2ALIGN(osize, (uint64_t)sizeof (vdev_label_t));
max_osize = P2ALIGN(max_osize, (uint64_t)sizeof (vdev_label_t));
if (vd->vdev_children == 0) {
if (osize < SPA_MINDEVSIZE) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_TOO_SMALL);
return (SET_ERROR(EOVERFLOW));
}
psize = osize;
asize = osize - (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE);
max_asize = max_osize - (VDEV_LABEL_START_SIZE +
VDEV_LABEL_END_SIZE);
} else {
if (vd->vdev_parent != NULL && osize < SPA_MINDEVSIZE -
(VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE)) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_TOO_SMALL);
return (SET_ERROR(EOVERFLOW));
}
psize = 0;
asize = osize;
max_asize = max_osize;
}
/*
* If the vdev was expanded, record this so that we can re-create the
* uberblock rings in labels {2,3}, during the next sync.
*/
if ((psize > vd->vdev_psize) && (vd->vdev_psize != 0))
vd->vdev_copy_uberblocks = B_TRUE;
vd->vdev_psize = psize;
/*
* Make sure the allocatable size hasn't shrunk too much.
*/
if (asize < vd->vdev_min_asize) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_BAD_LABEL);
return (SET_ERROR(EINVAL));
}
/*
* We can always set the logical/physical ashift members since
* their values are only used to calculate the vdev_ashift when
* the device is first added to the config. These values should
* not be used for anything else since they may change whenever
* the device is reopened and we don't store them in the label.
*/
vd->vdev_physical_ashift =
MAX(physical_ashift, vd->vdev_physical_ashift);
vd->vdev_logical_ashift = MAX(logical_ashift,
vd->vdev_logical_ashift);
if (vd->vdev_asize == 0) {
/*
* This is the first-ever open, so use the computed values.
* For compatibility, a different ashift can be requested.
*/
vd->vdev_asize = asize;
vd->vdev_max_asize = max_asize;
/*
* If the vdev_ashift was not overridden at creation time,
* then set it the logical ashift and optimize the ashift.
*/
if (vd->vdev_ashift == 0) {
vd->vdev_ashift = vd->vdev_logical_ashift;
if (vd->vdev_logical_ashift > ASHIFT_MAX) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_ASHIFT_TOO_BIG);
return (SET_ERROR(EDOM));
}
if (vd->vdev_top == vd && vd->vdev_attaching == B_FALSE)
vdev_ashift_optimize(vd);
vd->vdev_attaching = B_FALSE;
}
if (vd->vdev_ashift != 0 && (vd->vdev_ashift < ASHIFT_MIN ||
vd->vdev_ashift > ASHIFT_MAX)) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_BAD_ASHIFT);
return (SET_ERROR(EDOM));
}
} else {
/*
* Make sure the alignment required hasn't increased.
*/
if (vd->vdev_ashift > vd->vdev_top->vdev_ashift &&
vd->vdev_ops->vdev_op_leaf) {
(void) zfs_ereport_post(
FM_EREPORT_ZFS_DEVICE_BAD_ASHIFT,
spa, vd, NULL, NULL, 0);
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_BAD_LABEL);
return (SET_ERROR(EDOM));
}
vd->vdev_max_asize = max_asize;
}
/*
* If all children are healthy we update asize if either:
* The asize has increased, due to a device expansion caused by dynamic
* LUN growth or vdev replacement, and automatic expansion is enabled;
* making the additional space available.
*
* The asize has decreased, due to a device shrink usually caused by a
* vdev replace with a smaller device. This ensures that calculations
* based of max_asize and asize e.g. esize are always valid. It's safe
* to do this as we've already validated that asize is greater than
* vdev_min_asize.
*/
if (vd->vdev_state == VDEV_STATE_HEALTHY &&
((asize > vd->vdev_asize &&
(vd->vdev_expanding || spa->spa_autoexpand)) ||
(asize < vd->vdev_asize)))
vd->vdev_asize = asize;
vdev_set_min_asize(vd);
/*
* Ensure we can issue some IO before declaring the
* vdev open for business.
*/
if (vd->vdev_ops->vdev_op_leaf &&
(error = zio_wait(vdev_probe(vd, NULL))) != 0) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED,
VDEV_AUX_ERR_EXCEEDED);
return (error);
}
/*
* Track the minimum allocation size.
*/
if (vd->vdev_top == vd && vd->vdev_ashift != 0 &&
vd->vdev_islog == 0 && vd->vdev_aux == NULL) {
uint64_t min_alloc = vdev_get_min_alloc(vd);
vdev_spa_set_alloc(spa, min_alloc);
}
/*
* If this is a leaf vdev, assess whether a resilver is needed.
* But don't do this if we are doing a reopen for a scrub, since
* this would just restart the scrub we are already doing.
*/
if (vd->vdev_ops->vdev_op_leaf && !spa->spa_scrub_reopen)
dsl_scan_assess_vdev(spa->spa_dsl_pool, vd);
return (0);
}
static void
vdev_validate_child(void *arg)
{
vdev_t *vd = arg;
vd->vdev_validate_thread = curthread;
vd->vdev_validate_error = vdev_validate(vd);
vd->vdev_validate_thread = NULL;
}
/*
* Called once the vdevs are all opened, this routine validates the label
* contents. This needs to be done before vdev_load() so that we don't
* inadvertently do repair I/Os to the wrong device.
*
* This function will only return failure if one of the vdevs indicates that it
* has since been destroyed or exported. This is only possible if
* /etc/zfs/zpool.cache was readonly at the time. Otherwise, the vdev state
* will be updated but the function will return 0.
*/
int
vdev_validate(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
taskq_t *tq = NULL;
nvlist_t *label;
uint64_t guid = 0, aux_guid = 0, top_guid;
uint64_t state;
nvlist_t *nvl;
uint64_t txg;
int children = vd->vdev_children;
if (vdev_validate_skip)
return (0);
if (children > 0) {
tq = taskq_create("vdev_validate", children, minclsyspri,
children, children, TASKQ_PREPOPULATE);
}
for (uint64_t c = 0; c < children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (tq == NULL || vdev_uses_zvols(cvd)) {
vdev_validate_child(cvd);
} else {
VERIFY(taskq_dispatch(tq, vdev_validate_child, cvd,
TQ_SLEEP) != TASKQID_INVALID);
}
}
if (tq != NULL) {
taskq_wait(tq);
taskq_destroy(tq);
}
for (int c = 0; c < children; c++) {
int error = vd->vdev_child[c]->vdev_validate_error;
if (error != 0)
return (SET_ERROR(EBADF));
}
/*
* If the device has already failed, or was marked offline, don't do
* any further validation. Otherwise, label I/O will fail and we will
* overwrite the previous state.
*/
if (!vd->vdev_ops->vdev_op_leaf || !vdev_readable(vd))
return (0);
/*
* If we are performing an extreme rewind, we allow for a label that
* was modified at a point after the current txg.
* If config lock is not held do not check for the txg. spa_sync could
* be updating the vdev's label before updating spa_last_synced_txg.
*/
if (spa->spa_extreme_rewind || spa_last_synced_txg(spa) == 0 ||
spa_config_held(spa, SCL_CONFIG, RW_WRITER) != SCL_CONFIG)
txg = UINT64_MAX;
else
txg = spa_last_synced_txg(spa);
if ((label = vdev_label_read_config(vd, txg)) == NULL) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_BAD_LABEL);
vdev_dbgmsg(vd, "vdev_validate: failed reading config for "
"txg %llu", (u_longlong_t)txg);
return (0);
}
/*
* Determine if this vdev has been split off into another
* pool. If so, then refuse to open it.
*/
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_SPLIT_GUID,
&aux_guid) == 0 && aux_guid == spa_guid(spa)) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_SPLIT_POOL);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: vdev split into other pool");
return (0);
}
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID, &guid) != 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label",
ZPOOL_CONFIG_POOL_GUID);
return (0);
}
/*
* If config is not trusted then ignore the spa guid check. This is
* necessary because if the machine crashed during a re-guid the new
* guid might have been written to all of the vdev labels, but not the
* cached config. The check will be performed again once we have the
* trusted config from the MOS.
*/
if (spa->spa_trust_config && guid != spa_guid(spa)) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: vdev label pool_guid doesn't "
"match config (%llu != %llu)", (u_longlong_t)guid,
(u_longlong_t)spa_guid(spa));
return (0);
}
if (nvlist_lookup_nvlist(label, ZPOOL_CONFIG_VDEV_TREE, &nvl)
!= 0 || nvlist_lookup_uint64(nvl, ZPOOL_CONFIG_ORIG_GUID,
&aux_guid) != 0)
aux_guid = 0;
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) != 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label",
ZPOOL_CONFIG_GUID);
return (0);
}
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_TOP_GUID, &top_guid)
!= 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label",
ZPOOL_CONFIG_TOP_GUID);
return (0);
}
/*
* If this vdev just became a top-level vdev because its sibling was
* detached, it will have adopted the parent's vdev guid -- but the
* label may or may not be on disk yet. Fortunately, either version
* of the label will have the same top guid, so if we're a top-level
* vdev, we can safely compare to that instead.
* However, if the config comes from a cachefile that failed to update
* after the detach, a top-level vdev will appear as a non top-level
* vdev in the config. Also relax the constraints if we perform an
* extreme rewind.
*
* If we split this vdev off instead, then we also check the
* original pool's guid. We don't want to consider the vdev
* corrupt if it is partway through a split operation.
*/
if (vd->vdev_guid != guid && vd->vdev_guid != aux_guid) {
boolean_t mismatch = B_FALSE;
if (spa->spa_trust_config && !spa->spa_extreme_rewind) {
if (vd != vd->vdev_top || vd->vdev_guid != top_guid)
mismatch = B_TRUE;
} else {
if (vd->vdev_guid != top_guid &&
vd->vdev_top->vdev_guid != guid)
mismatch = B_TRUE;
}
if (mismatch) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: config guid "
"doesn't match label guid");
vdev_dbgmsg(vd, "CONFIG: guid %llu, top_guid %llu",
(u_longlong_t)vd->vdev_guid,
(u_longlong_t)vd->vdev_top->vdev_guid);
vdev_dbgmsg(vd, "LABEL: guid %llu, top_guid %llu, "
"aux_guid %llu", (u_longlong_t)guid,
(u_longlong_t)top_guid, (u_longlong_t)aux_guid);
return (0);
}
}
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE,
&state) != 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label",
ZPOOL_CONFIG_POOL_STATE);
return (0);
}
nvlist_free(label);
/*
* If this is a verbatim import, no need to check the
* state of the pool.
*/
if (!(spa->spa_import_flags & ZFS_IMPORT_VERBATIM) &&
spa_load_state(spa) == SPA_LOAD_OPEN &&
state != POOL_STATE_ACTIVE) {
vdev_dbgmsg(vd, "vdev_validate: invalid pool state (%llu) "
"for spa %s", (u_longlong_t)state, spa->spa_name);
return (SET_ERROR(EBADF));
}
/*
* If we were able to open and validate a vdev that was
* previously marked permanently unavailable, clear that state
* now.
*/
if (vd->vdev_not_present)
vd->vdev_not_present = 0;
return (0);
}
static void
vdev_copy_path_impl(vdev_t *svd, vdev_t *dvd)
{
char *old, *new;
if (svd->vdev_path != NULL && dvd->vdev_path != NULL) {
if (strcmp(svd->vdev_path, dvd->vdev_path) != 0) {
zfs_dbgmsg("vdev_copy_path: vdev %llu: path changed "
"from '%s' to '%s'", (u_longlong_t)dvd->vdev_guid,
dvd->vdev_path, svd->vdev_path);
spa_strfree(dvd->vdev_path);
dvd->vdev_path = spa_strdup(svd->vdev_path);
}
} else if (svd->vdev_path != NULL) {
dvd->vdev_path = spa_strdup(svd->vdev_path);
zfs_dbgmsg("vdev_copy_path: vdev %llu: path set to '%s'",
(u_longlong_t)dvd->vdev_guid, dvd->vdev_path);
}
/*
* Our enclosure sysfs path may have changed between imports
*/
old = dvd->vdev_enc_sysfs_path;
new = svd->vdev_enc_sysfs_path;
if ((old != NULL && new == NULL) ||
(old == NULL && new != NULL) ||
((old != NULL && new != NULL) && strcmp(new, old) != 0)) {
zfs_dbgmsg("vdev_copy_path: vdev %llu: vdev_enc_sysfs_path "
"changed from '%s' to '%s'", (u_longlong_t)dvd->vdev_guid,
old, new);
if (dvd->vdev_enc_sysfs_path)
spa_strfree(dvd->vdev_enc_sysfs_path);
if (svd->vdev_enc_sysfs_path) {
dvd->vdev_enc_sysfs_path = spa_strdup(
svd->vdev_enc_sysfs_path);
} else {
dvd->vdev_enc_sysfs_path = NULL;
}
}
}
/*
* Recursively copy vdev paths from one vdev to another. Source and destination
* vdev trees must have same geometry otherwise return error. Intended to copy
* paths from userland config into MOS config.
*/
int
vdev_copy_path_strict(vdev_t *svd, vdev_t *dvd)
{
if ((svd->vdev_ops == &vdev_missing_ops) ||
(svd->vdev_ishole && dvd->vdev_ishole) ||
(dvd->vdev_ops == &vdev_indirect_ops))
return (0);
if (svd->vdev_ops != dvd->vdev_ops) {
vdev_dbgmsg(svd, "vdev_copy_path: vdev type mismatch: %s != %s",
svd->vdev_ops->vdev_op_type, dvd->vdev_ops->vdev_op_type);
return (SET_ERROR(EINVAL));
}
if (svd->vdev_guid != dvd->vdev_guid) {
vdev_dbgmsg(svd, "vdev_copy_path: guids mismatch (%llu != "
"%llu)", (u_longlong_t)svd->vdev_guid,
(u_longlong_t)dvd->vdev_guid);
return (SET_ERROR(EINVAL));
}
if (svd->vdev_children != dvd->vdev_children) {
vdev_dbgmsg(svd, "vdev_copy_path: children count mismatch: "
"%llu != %llu", (u_longlong_t)svd->vdev_children,
(u_longlong_t)dvd->vdev_children);
return (SET_ERROR(EINVAL));
}
for (uint64_t i = 0; i < svd->vdev_children; i++) {
int error = vdev_copy_path_strict(svd->vdev_child[i],
dvd->vdev_child[i]);
if (error != 0)
return (error);
}
if (svd->vdev_ops->vdev_op_leaf)
vdev_copy_path_impl(svd, dvd);
return (0);
}
static void
vdev_copy_path_search(vdev_t *stvd, vdev_t *dvd)
{
ASSERT(stvd->vdev_top == stvd);
ASSERT3U(stvd->vdev_id, ==, dvd->vdev_top->vdev_id);
for (uint64_t i = 0; i < dvd->vdev_children; i++) {
vdev_copy_path_search(stvd, dvd->vdev_child[i]);
}
if (!dvd->vdev_ops->vdev_op_leaf || !vdev_is_concrete(dvd))
return;
/*
* The idea here is that while a vdev can shift positions within
* a top vdev (when replacing, attaching mirror, etc.) it cannot
* step outside of it.
*/
vdev_t *vd = vdev_lookup_by_guid(stvd, dvd->vdev_guid);
if (vd == NULL || vd->vdev_ops != dvd->vdev_ops)
return;
ASSERT(vd->vdev_ops->vdev_op_leaf);
vdev_copy_path_impl(vd, dvd);
}
/*
* Recursively copy vdev paths from one root vdev to another. Source and
* destination vdev trees may differ in geometry. For each destination leaf
* vdev, search a vdev with the same guid and top vdev id in the source.
* Intended to copy paths from userland config into MOS config.
*/
void
vdev_copy_path_relaxed(vdev_t *srvd, vdev_t *drvd)
{
uint64_t children = MIN(srvd->vdev_children, drvd->vdev_children);
ASSERT(srvd->vdev_ops == &vdev_root_ops);
ASSERT(drvd->vdev_ops == &vdev_root_ops);
for (uint64_t i = 0; i < children; i++) {
vdev_copy_path_search(srvd->vdev_child[i],
drvd->vdev_child[i]);
}
}
/*
* Close a virtual device.
*/
void
vdev_close(vdev_t *vd)
{
vdev_t *pvd = vd->vdev_parent;
spa_t *spa __maybe_unused = vd->vdev_spa;
ASSERT(vd != NULL);
ASSERT(vd->vdev_open_thread == curthread ||
spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
/*
* If our parent is reopening, then we are as well, unless we are
* going offline.
*/
if (pvd != NULL && pvd->vdev_reopening)
vd->vdev_reopening = (pvd->vdev_reopening && !vd->vdev_offline);
vd->vdev_ops->vdev_op_close(vd);
/*
* We record the previous state before we close it, so that if we are
* doing a reopen(), we don't generate FMA ereports if we notice that
* it's still faulted.
*/
vd->vdev_prevstate = vd->vdev_state;
if (vd->vdev_offline)
vd->vdev_state = VDEV_STATE_OFFLINE;
else
vd->vdev_state = VDEV_STATE_CLOSED;
vd->vdev_stat.vs_aux = VDEV_AUX_NONE;
}
void
vdev_hold(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_is_root(spa));
if (spa->spa_state == POOL_STATE_UNINITIALIZED)
return;
for (int c = 0; c < vd->vdev_children; c++)
vdev_hold(vd->vdev_child[c]);
if (vd->vdev_ops->vdev_op_leaf && vd->vdev_ops->vdev_op_hold != NULL)
vd->vdev_ops->vdev_op_hold(vd);
}
void
vdev_rele(vdev_t *vd)
{
ASSERT(spa_is_root(vd->vdev_spa));
for (int c = 0; c < vd->vdev_children; c++)
vdev_rele(vd->vdev_child[c]);
if (vd->vdev_ops->vdev_op_leaf && vd->vdev_ops->vdev_op_rele != NULL)
vd->vdev_ops->vdev_op_rele(vd);
}
/*
* Reopen all interior vdevs and any unopened leaves. We don't actually
* reopen leaf vdevs which had previously been opened as they might deadlock
* on the spa_config_lock. Instead we only obtain the leaf's physical size.
* If the leaf has never been opened then open it, as usual.
*/
void
vdev_reopen(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
/* set the reopening flag unless we're taking the vdev offline */
vd->vdev_reopening = !vd->vdev_offline;
vdev_close(vd);
(void) vdev_open(vd);
/*
* Call vdev_validate() here to make sure we have the same device.
* Otherwise, a device with an invalid label could be successfully
* opened in response to vdev_reopen().
*/
if (vd->vdev_aux) {
(void) vdev_validate_aux(vd);
if (vdev_readable(vd) && vdev_writeable(vd) &&
vd->vdev_aux == &spa->spa_l2cache) {
/*
* In case the vdev is present we should evict all ARC
* buffers and pointers to log blocks and reclaim their
* space before restoring its contents to L2ARC.
*/
if (l2arc_vdev_present(vd)) {
l2arc_rebuild_vdev(vd, B_TRUE);
} else {
l2arc_add_vdev(spa, vd);
}
spa_async_request(spa, SPA_ASYNC_L2CACHE_REBUILD);
spa_async_request(spa, SPA_ASYNC_L2CACHE_TRIM);
}
} else {
(void) vdev_validate(vd);
}
/*
* Recheck if resilver is still needed and cancel any
* scheduled resilver if resilver is unneeded.
*/
if (!vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL) &&
spa->spa_async_tasks & SPA_ASYNC_RESILVER) {
mutex_enter(&spa->spa_async_lock);
spa->spa_async_tasks &= ~SPA_ASYNC_RESILVER;
mutex_exit(&spa->spa_async_lock);
}
/*
* Reassess parent vdev's health.
*/
vdev_propagate_state(vd);
}
int
vdev_create(vdev_t *vd, uint64_t txg, boolean_t isreplacing)
{
int error;
/*
* Normally, partial opens (e.g. of a mirror) are allowed.
* For a create, however, we want to fail the request if
* there are any components we can't open.
*/
error = vdev_open(vd);
if (error || vd->vdev_state != VDEV_STATE_HEALTHY) {
vdev_close(vd);
return (error ? error : SET_ERROR(ENXIO));
}
/*
* Recursively load DTLs and initialize all labels.
*/
if ((error = vdev_dtl_load(vd)) != 0 ||
(error = vdev_label_init(vd, txg, isreplacing ?
VDEV_LABEL_REPLACE : VDEV_LABEL_CREATE)) != 0) {
vdev_close(vd);
return (error);
}
return (0);
}
void
vdev_metaslab_set_size(vdev_t *vd)
{
uint64_t asize = vd->vdev_asize;
uint64_t ms_count = asize >> zfs_vdev_default_ms_shift;
uint64_t ms_shift;
/*
* There are two dimensions to the metaslab sizing calculation:
* the size of the metaslab and the count of metaslabs per vdev.
*
* The default values used below are a good balance between memory
* usage (larger metaslab size means more memory needed for loaded
* metaslabs; more metaslabs means more memory needed for the
* metaslab_t structs), metaslab load time (larger metaslabs take
* longer to load), and metaslab sync time (more metaslabs means
* more time spent syncing all of them).
*
* In general, we aim for zfs_vdev_default_ms_count (200) metaslabs.
* The range of the dimensions are as follows:
*
* 2^29 <= ms_size <= 2^34
* 16 <= ms_count <= 131,072
*
* On the lower end of vdev sizes, we aim for metaslabs sizes of
* at least 512MB (2^29) to minimize fragmentation effects when
* testing with smaller devices. However, the count constraint
* of at least 16 metaslabs will override this minimum size goal.
*
* On the upper end of vdev sizes, we aim for a maximum metaslab
* size of 16GB. However, we will cap the total count to 2^17
* metaslabs to keep our memory footprint in check and let the
* metaslab size grow from there if that limit is hit.
*
* The net effect of applying above constrains is summarized below.
*
* vdev size metaslab count
* --------------|-----------------
* < 8GB ~16
* 8GB - 100GB one per 512MB
* 100GB - 3TB ~200
* 3TB - 2PB one per 16GB
* > 2PB ~131,072
* --------------------------------
*
* Finally, note that all of the above calculate the initial
* number of metaslabs. Expanding a top-level vdev will result
* in additional metaslabs being allocated making it possible
* to exceed the zfs_vdev_ms_count_limit.
*/
if (ms_count < zfs_vdev_min_ms_count)
ms_shift = highbit64(asize / zfs_vdev_min_ms_count);
else if (ms_count > zfs_vdev_default_ms_count)
ms_shift = highbit64(asize / zfs_vdev_default_ms_count);
else
ms_shift = zfs_vdev_default_ms_shift;
if (ms_shift < SPA_MAXBLOCKSHIFT) {
ms_shift = SPA_MAXBLOCKSHIFT;
} else if (ms_shift > zfs_vdev_max_ms_shift) {
ms_shift = zfs_vdev_max_ms_shift;
/* cap the total count to constrain memory footprint */
if ((asize >> ms_shift) > zfs_vdev_ms_count_limit)
ms_shift = highbit64(asize / zfs_vdev_ms_count_limit);
}
vd->vdev_ms_shift = ms_shift;
ASSERT3U(vd->vdev_ms_shift, >=, SPA_MAXBLOCKSHIFT);
}
void
vdev_dirty(vdev_t *vd, int flags, void *arg, uint64_t txg)
{
ASSERT(vd == vd->vdev_top);
/* indirect vdevs don't have metaslabs or dtls */
ASSERT(vdev_is_concrete(vd) || flags == 0);
ASSERT(ISP2(flags));
ASSERT(spa_writeable(vd->vdev_spa));
if (flags & VDD_METASLAB)
(void) txg_list_add(&vd->vdev_ms_list, arg, txg);
if (flags & VDD_DTL)
(void) txg_list_add(&vd->vdev_dtl_list, arg, txg);
(void) txg_list_add(&vd->vdev_spa->spa_vdev_txg_list, vd, txg);
}
void
vdev_dirty_leaves(vdev_t *vd, int flags, uint64_t txg)
{
for (int c = 0; c < vd->vdev_children; c++)
vdev_dirty_leaves(vd->vdev_child[c], flags, txg);
if (vd->vdev_ops->vdev_op_leaf)
vdev_dirty(vd->vdev_top, flags, vd, txg);
}
/*
* DTLs.
*
* A vdev's DTL (dirty time log) is the set of transaction groups for which
* the vdev has less than perfect replication. There are four kinds of DTL:
*
* DTL_MISSING: txgs for which the vdev has no valid copies of the data
*
* DTL_PARTIAL: txgs for which data is available, but not fully replicated
*
* DTL_SCRUB: the txgs that could not be repaired by the last scrub; upon
* scrub completion, DTL_SCRUB replaces DTL_MISSING in the range of
* txgs that was scrubbed.
*
* DTL_OUTAGE: txgs which cannot currently be read, whether due to
* persistent errors or just some device being offline.
* Unlike the other three, the DTL_OUTAGE map is not generally
* maintained; it's only computed when needed, typically to
* determine whether a device can be detached.
*
* For leaf vdevs, DTL_MISSING and DTL_PARTIAL are identical: the device
* either has the data or it doesn't.
*
* For interior vdevs such as mirror and RAID-Z the picture is more complex.
* A vdev's DTL_PARTIAL is the union of its children's DTL_PARTIALs, because
* if any child is less than fully replicated, then so is its parent.
* A vdev's DTL_MISSING is a modified union of its children's DTL_MISSINGs,
* comprising only those txgs which appear in 'maxfaults' or more children;
* those are the txgs we don't have enough replication to read. For example,
* double-parity RAID-Z can tolerate up to two missing devices (maxfaults == 2);
* thus, its DTL_MISSING consists of the set of txgs that appear in more than
* two child DTL_MISSING maps.
*
* It should be clear from the above that to compute the DTLs and outage maps
* for all vdevs, it suffices to know just the leaf vdevs' DTL_MISSING maps.
* Therefore, that is all we keep on disk. When loading the pool, or after
* a configuration change, we generate all other DTLs from first principles.
*/
void
vdev_dtl_dirty(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size)
{
range_tree_t *rt = vd->vdev_dtl[t];
ASSERT(t < DTL_TYPES);
ASSERT(vd != vd->vdev_spa->spa_root_vdev);
ASSERT(spa_writeable(vd->vdev_spa));
mutex_enter(&vd->vdev_dtl_lock);
if (!range_tree_contains(rt, txg, size))
range_tree_add(rt, txg, size);
mutex_exit(&vd->vdev_dtl_lock);
}
boolean_t
vdev_dtl_contains(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size)
{
range_tree_t *rt = vd->vdev_dtl[t];
boolean_t dirty = B_FALSE;
ASSERT(t < DTL_TYPES);
ASSERT(vd != vd->vdev_spa->spa_root_vdev);
/*
* While we are loading the pool, the DTLs have not been loaded yet.
* This isn't a problem but it can result in devices being tried
* which are known to not have the data. In which case, the import
* is relying on the checksum to ensure that we get the right data.
* Note that while importing we are only reading the MOS, which is
* always checksummed.
*/
mutex_enter(&vd->vdev_dtl_lock);
if (!range_tree_is_empty(rt))
dirty = range_tree_contains(rt, txg, size);
mutex_exit(&vd->vdev_dtl_lock);
return (dirty);
}
boolean_t
vdev_dtl_empty(vdev_t *vd, vdev_dtl_type_t t)
{
range_tree_t *rt = vd->vdev_dtl[t];
boolean_t empty;
mutex_enter(&vd->vdev_dtl_lock);
empty = range_tree_is_empty(rt);
mutex_exit(&vd->vdev_dtl_lock);
return (empty);
}
/*
* Check if the txg falls within the range which must be
* resilvered. DVAs outside this range can always be skipped.
*/
boolean_t
vdev_default_need_resilver(vdev_t *vd, const dva_t *dva, size_t psize,
uint64_t phys_birth)
{
(void) dva, (void) psize;
/* Set by sequential resilver. */
if (phys_birth == TXG_UNKNOWN)
return (B_TRUE);
return (vdev_dtl_contains(vd, DTL_PARTIAL, phys_birth, 1));
}
/*
* Returns B_TRUE if the vdev determines the DVA needs to be resilvered.
*/
boolean_t
vdev_dtl_need_resilver(vdev_t *vd, const dva_t *dva, size_t psize,
uint64_t phys_birth)
{
ASSERT(vd != vd->vdev_spa->spa_root_vdev);
if (vd->vdev_ops->vdev_op_need_resilver == NULL ||
vd->vdev_ops->vdev_op_leaf)
return (B_TRUE);
return (vd->vdev_ops->vdev_op_need_resilver(vd, dva, psize,
phys_birth));
}
/*
* Returns the lowest txg in the DTL range.
*/
static uint64_t
vdev_dtl_min(vdev_t *vd)
{
ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock));
ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0);
ASSERT0(vd->vdev_children);
return (range_tree_min(vd->vdev_dtl[DTL_MISSING]) - 1);
}
/*
* Returns the highest txg in the DTL.
*/
static uint64_t
vdev_dtl_max(vdev_t *vd)
{
ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock));
ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0);
ASSERT0(vd->vdev_children);
return (range_tree_max(vd->vdev_dtl[DTL_MISSING]));
}
/*
* Determine if a resilvering vdev should remove any DTL entries from
* its range. If the vdev was resilvering for the entire duration of the
* scan then it should excise that range from its DTLs. Otherwise, this
* vdev is considered partially resilvered and should leave its DTL
* entries intact. The comment in vdev_dtl_reassess() describes how we
* excise the DTLs.
*/
static boolean_t
vdev_dtl_should_excise(vdev_t *vd, boolean_t rebuild_done)
{
ASSERT0(vd->vdev_children);
if (vd->vdev_state < VDEV_STATE_DEGRADED)
return (B_FALSE);
if (vd->vdev_resilver_deferred)
return (B_FALSE);
if (range_tree_is_empty(vd->vdev_dtl[DTL_MISSING]))
return (B_TRUE);
if (rebuild_done) {
vdev_rebuild_t *vr = &vd->vdev_top->vdev_rebuild_config;
vdev_rebuild_phys_t *vrp = &vr->vr_rebuild_phys;
/* Rebuild not initiated by attach */
if (vd->vdev_rebuild_txg == 0)
return (B_TRUE);
/*
* When a rebuild completes without error then all missing data
* up to the rebuild max txg has been reconstructed and the DTL
* is eligible for excision.
*/
if (vrp->vrp_rebuild_state == VDEV_REBUILD_COMPLETE &&
vdev_dtl_max(vd) <= vrp->vrp_max_txg) {
ASSERT3U(vrp->vrp_min_txg, <=, vdev_dtl_min(vd));
ASSERT3U(vrp->vrp_min_txg, <, vd->vdev_rebuild_txg);
ASSERT3U(vd->vdev_rebuild_txg, <=, vrp->vrp_max_txg);
return (B_TRUE);
}
} else {
dsl_scan_t *scn = vd->vdev_spa->spa_dsl_pool->dp_scan;
dsl_scan_phys_t *scnp __maybe_unused = &scn->scn_phys;
/* Resilver not initiated by attach */
if (vd->vdev_resilver_txg == 0)
return (B_TRUE);
/*
* When a resilver is initiated the scan will assign the
* scn_max_txg value to the highest txg value that exists
* in all DTLs. If this device's max DTL is not part of this
* scan (i.e. it is not in the range (scn_min_txg, scn_max_txg]
* then it is not eligible for excision.
*/
if (vdev_dtl_max(vd) <= scn->scn_phys.scn_max_txg) {
ASSERT3U(scnp->scn_min_txg, <=, vdev_dtl_min(vd));
ASSERT3U(scnp->scn_min_txg, <, vd->vdev_resilver_txg);
ASSERT3U(vd->vdev_resilver_txg, <=, scnp->scn_max_txg);
return (B_TRUE);
}
}
return (B_FALSE);
}
/*
* Reassess DTLs after a config change or scrub completion. If txg == 0 no
* write operations will be issued to the pool.
*/
void
vdev_dtl_reassess(vdev_t *vd, uint64_t txg, uint64_t scrub_txg,
boolean_t scrub_done, boolean_t rebuild_done)
{
spa_t *spa = vd->vdev_spa;
avl_tree_t reftree;
int minref;
ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0);
for (int c = 0; c < vd->vdev_children; c++)
vdev_dtl_reassess(vd->vdev_child[c], txg,
scrub_txg, scrub_done, rebuild_done);
if (vd == spa->spa_root_vdev || !vdev_is_concrete(vd) || vd->vdev_aux)
return;
if (vd->vdev_ops->vdev_op_leaf) {
dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan;
vdev_rebuild_t *vr = &vd->vdev_top->vdev_rebuild_config;
boolean_t check_excise = B_FALSE;
boolean_t wasempty = B_TRUE;
mutex_enter(&vd->vdev_dtl_lock);
/*
* If requested, pretend the scan or rebuild completed cleanly.
*/
if (zfs_scan_ignore_errors) {
if (scn != NULL)
scn->scn_phys.scn_errors = 0;
if (vr != NULL)
vr->vr_rebuild_phys.vrp_errors = 0;
}
if (scrub_txg != 0 &&
!range_tree_is_empty(vd->vdev_dtl[DTL_MISSING])) {
wasempty = B_FALSE;
zfs_dbgmsg("guid:%llu txg:%llu scrub:%llu started:%d "
"dtl:%llu/%llu errors:%llu",
(u_longlong_t)vd->vdev_guid, (u_longlong_t)txg,
(u_longlong_t)scrub_txg, spa->spa_scrub_started,
(u_longlong_t)vdev_dtl_min(vd),
(u_longlong_t)vdev_dtl_max(vd),
(u_longlong_t)(scn ? scn->scn_phys.scn_errors : 0));
}
/*
* If we've completed a scrub/resilver or a rebuild cleanly
* then determine if this vdev should remove any DTLs. We
* only want to excise regions on vdevs that were available
* during the entire duration of this scan.
*/
if (rebuild_done &&
vr != NULL && vr->vr_rebuild_phys.vrp_errors == 0) {
check_excise = B_TRUE;
} else {
if (spa->spa_scrub_started ||
(scn != NULL && scn->scn_phys.scn_errors == 0)) {
check_excise = B_TRUE;
}
}
if (scrub_txg && check_excise &&
vdev_dtl_should_excise(vd, rebuild_done)) {
/*
* We completed a scrub, resilver or rebuild up to
* scrub_txg. If we did it without rebooting, then
* the scrub dtl will be valid, so excise the old
* region and fold in the scrub dtl. Otherwise,
* leave the dtl as-is if there was an error.
*
* There's little trick here: to excise the beginning
* of the DTL_MISSING map, we put it into a reference
* tree and then add a segment with refcnt -1 that
* covers the range [0, scrub_txg). This means
* that each txg in that range has refcnt -1 or 0.
* We then add DTL_SCRUB with a refcnt of 2, so that
* entries in the range [0, scrub_txg) will have a
* positive refcnt -- either 1 or 2. We then convert
* the reference tree into the new DTL_MISSING map.
*/
space_reftree_create(&reftree);
space_reftree_add_map(&reftree,
vd->vdev_dtl[DTL_MISSING], 1);
space_reftree_add_seg(&reftree, 0, scrub_txg, -1);
space_reftree_add_map(&reftree,
vd->vdev_dtl[DTL_SCRUB], 2);
space_reftree_generate_map(&reftree,
vd->vdev_dtl[DTL_MISSING], 1);
space_reftree_destroy(&reftree);
if (!range_tree_is_empty(vd->vdev_dtl[DTL_MISSING])) {
zfs_dbgmsg("update DTL_MISSING:%llu/%llu",
(u_longlong_t)vdev_dtl_min(vd),
(u_longlong_t)vdev_dtl_max(vd));
} else if (!wasempty) {
zfs_dbgmsg("DTL_MISSING is now empty");
}
}
range_tree_vacate(vd->vdev_dtl[DTL_PARTIAL], NULL, NULL);
range_tree_walk(vd->vdev_dtl[DTL_MISSING],
range_tree_add, vd->vdev_dtl[DTL_PARTIAL]);
if (scrub_done)
range_tree_vacate(vd->vdev_dtl[DTL_SCRUB], NULL, NULL);
range_tree_vacate(vd->vdev_dtl[DTL_OUTAGE], NULL, NULL);
if (!vdev_readable(vd))
range_tree_add(vd->vdev_dtl[DTL_OUTAGE], 0, -1ULL);
else
range_tree_walk(vd->vdev_dtl[DTL_MISSING],
range_tree_add, vd->vdev_dtl[DTL_OUTAGE]);
/*
* If the vdev was resilvering or rebuilding and no longer
* has any DTLs then reset the appropriate flag and dirty
* the top level so that we persist the change.
*/
if (txg != 0 &&
range_tree_is_empty(vd->vdev_dtl[DTL_MISSING]) &&
range_tree_is_empty(vd->vdev_dtl[DTL_OUTAGE])) {
if (vd->vdev_rebuild_txg != 0) {
vd->vdev_rebuild_txg = 0;
vdev_config_dirty(vd->vdev_top);
} else if (vd->vdev_resilver_txg != 0) {
vd->vdev_resilver_txg = 0;
vdev_config_dirty(vd->vdev_top);
}
}
mutex_exit(&vd->vdev_dtl_lock);
if (txg != 0)
vdev_dirty(vd->vdev_top, VDD_DTL, vd, txg);
- return;
+ } else {
+ 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) {
+ /* leaf vdevs only */
+ continue;
+ }
+ if (t == DTL_PARTIAL) {
+ /* i.e. non-zero */
+ minref = 1;
+ } else if (vdev_get_nparity(vd) != 0) {
+ /* RAIDZ, DRAID */
+ minref = vdev_get_nparity(vd) + 1;
+ } else {
+ /* any kind of mirror */
+ minref = vd->vdev_children;
+ }
+ 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);
}
- mutex_enter(&vd->vdev_dtl_lock);
- for (int t = 0; t < DTL_TYPES; t++) {
- /* account for child's outage in parent's missing map */
- int s = (t == DTL_MISSING) ? DTL_OUTAGE: t;
- if (t == DTL_SCRUB)
- continue; /* leaf vdevs only */
- if (t == DTL_PARTIAL)
- minref = 1; /* i.e. non-zero */
- else if (vdev_get_nparity(vd) != 0)
- minref = vdev_get_nparity(vd) + 1; /* RAID-Z, dRAID */
- else
- minref = vd->vdev_children; /* any kind of mirror */
- space_reftree_create(&reftree);
- for (int c = 0; c < vd->vdev_children; c++) {
- vdev_t *cvd = vd->vdev_child[c];
- mutex_enter(&cvd->vdev_dtl_lock);
- space_reftree_add_map(&reftree, cvd->vdev_dtl[s], 1);
- mutex_exit(&cvd->vdev_dtl_lock);
- }
- space_reftree_generate_map(&reftree, vd->vdev_dtl[t], minref);
- space_reftree_destroy(&reftree);
+ if (vd->vdev_top->vdev_ops == &vdev_raidz_ops) {
+ raidz_dtl_reassessed(vd);
}
- mutex_exit(&vd->vdev_dtl_lock);
}
/*
* Iterate over all the vdevs except spare, and post kobj events
*/
void
vdev_post_kobj_evt(vdev_t *vd)
{
if (vd->vdev_ops->vdev_op_kobj_evt_post &&
vd->vdev_kobj_flag == B_FALSE) {
vd->vdev_kobj_flag = B_TRUE;
vd->vdev_ops->vdev_op_kobj_evt_post(vd);
}
for (int c = 0; c < vd->vdev_children; c++)
vdev_post_kobj_evt(vd->vdev_child[c]);
}
/*
* Iterate over all the vdevs except spare, and clear kobj events
*/
void
vdev_clear_kobj_evt(vdev_t *vd)
{
vd->vdev_kobj_flag = B_FALSE;
for (int c = 0; c < vd->vdev_children; c++)
vdev_clear_kobj_evt(vd->vdev_child[c]);
}
int
vdev_dtl_load(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa->spa_meta_objset;
range_tree_t *rt;
int error = 0;
if (vd->vdev_ops->vdev_op_leaf && vd->vdev_dtl_object != 0) {
ASSERT(vdev_is_concrete(vd));
/*
* If the dtl cannot be sync'd there is no need to open it.
*/
if (spa->spa_mode == SPA_MODE_READ && !spa->spa_read_spacemaps)
return (0);
error = space_map_open(&vd->vdev_dtl_sm, mos,
vd->vdev_dtl_object, 0, -1ULL, 0);
if (error)
return (error);
ASSERT(vd->vdev_dtl_sm != NULL);
rt = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0);
error = space_map_load(vd->vdev_dtl_sm, rt, SM_ALLOC);
if (error == 0) {
mutex_enter(&vd->vdev_dtl_lock);
range_tree_walk(rt, range_tree_add,
vd->vdev_dtl[DTL_MISSING]);
mutex_exit(&vd->vdev_dtl_lock);
}
range_tree_vacate(rt, NULL, NULL);
range_tree_destroy(rt);
return (error);
}
for (int c = 0; c < vd->vdev_children; c++) {
error = vdev_dtl_load(vd->vdev_child[c]);
if (error != 0)
break;
}
return (error);
}
static void
vdev_zap_allocation_data(vdev_t *vd, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa->spa_meta_objset;
vdev_alloc_bias_t alloc_bias = vd->vdev_alloc_bias;
const char *string;
ASSERT(alloc_bias != VDEV_BIAS_NONE);
string =
(alloc_bias == VDEV_BIAS_LOG) ? VDEV_ALLOC_BIAS_LOG :
(alloc_bias == VDEV_BIAS_SPECIAL) ? VDEV_ALLOC_BIAS_SPECIAL :
(alloc_bias == VDEV_BIAS_DEDUP) ? VDEV_ALLOC_BIAS_DEDUP : NULL;
ASSERT(string != NULL);
VERIFY0(zap_add(mos, vd->vdev_top_zap, VDEV_TOP_ZAP_ALLOCATION_BIAS,
1, strlen(string) + 1, string, tx));
if (alloc_bias == VDEV_BIAS_SPECIAL || alloc_bias == VDEV_BIAS_DEDUP) {
spa_activate_allocation_classes(spa, tx);
}
}
void
vdev_destroy_unlink_zap(vdev_t *vd, uint64_t zapobj, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
VERIFY0(zap_destroy(spa->spa_meta_objset, zapobj, tx));
VERIFY0(zap_remove_int(spa->spa_meta_objset, spa->spa_all_vdev_zaps,
zapobj, tx));
}
uint64_t
vdev_create_link_zap(vdev_t *vd, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
uint64_t zap = zap_create(spa->spa_meta_objset, DMU_OTN_ZAP_METADATA,
DMU_OT_NONE, 0, tx);
ASSERT(zap != 0);
VERIFY0(zap_add_int(spa->spa_meta_objset, spa->spa_all_vdev_zaps,
zap, tx));
return (zap);
}
void
vdev_construct_zaps(vdev_t *vd, dmu_tx_t *tx)
{
if (vd->vdev_ops != &vdev_hole_ops &&
vd->vdev_ops != &vdev_missing_ops &&
vd->vdev_ops != &vdev_root_ops &&
!vd->vdev_top->vdev_removing) {
if (vd->vdev_ops->vdev_op_leaf && vd->vdev_leaf_zap == 0) {
vd->vdev_leaf_zap = vdev_create_link_zap(vd, tx);
}
if (vd == vd->vdev_top && vd->vdev_top_zap == 0) {
vd->vdev_top_zap = vdev_create_link_zap(vd, tx);
if (vd->vdev_alloc_bias != VDEV_BIAS_NONE)
vdev_zap_allocation_data(vd, tx);
}
}
if (vd->vdev_ops == &vdev_root_ops && vd->vdev_root_zap == 0 &&
spa_feature_is_enabled(vd->vdev_spa, SPA_FEATURE_AVZ_V2)) {
if (!spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_AVZ_V2))
spa_feature_incr(vd->vdev_spa, SPA_FEATURE_AVZ_V2, tx);
vd->vdev_root_zap = vdev_create_link_zap(vd, tx);
}
for (uint64_t i = 0; i < vd->vdev_children; i++) {
vdev_construct_zaps(vd->vdev_child[i], tx);
}
}
static void
vdev_dtl_sync(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
range_tree_t *rt = vd->vdev_dtl[DTL_MISSING];
objset_t *mos = spa->spa_meta_objset;
range_tree_t *rtsync;
dmu_tx_t *tx;
uint64_t object = space_map_object(vd->vdev_dtl_sm);
ASSERT(vdev_is_concrete(vd));
ASSERT(vd->vdev_ops->vdev_op_leaf);
tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg);
if (vd->vdev_detached || vd->vdev_top->vdev_removing) {
mutex_enter(&vd->vdev_dtl_lock);
space_map_free(vd->vdev_dtl_sm, tx);
space_map_close(vd->vdev_dtl_sm);
vd->vdev_dtl_sm = NULL;
mutex_exit(&vd->vdev_dtl_lock);
/*
* We only destroy the leaf ZAP for detached leaves or for
* removed log devices. Removed data devices handle leaf ZAP
* cleanup later, once cancellation is no longer possible.
*/
if (vd->vdev_leaf_zap != 0 && (vd->vdev_detached ||
vd->vdev_top->vdev_islog)) {
vdev_destroy_unlink_zap(vd, vd->vdev_leaf_zap, tx);
vd->vdev_leaf_zap = 0;
}
dmu_tx_commit(tx);
return;
}
if (vd->vdev_dtl_sm == NULL) {
uint64_t new_object;
new_object = space_map_alloc(mos, zfs_vdev_dtl_sm_blksz, tx);
VERIFY3U(new_object, !=, 0);
VERIFY0(space_map_open(&vd->vdev_dtl_sm, mos, new_object,
0, -1ULL, 0));
ASSERT(vd->vdev_dtl_sm != NULL);
}
rtsync = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0);
mutex_enter(&vd->vdev_dtl_lock);
range_tree_walk(rt, range_tree_add, rtsync);
mutex_exit(&vd->vdev_dtl_lock);
space_map_truncate(vd->vdev_dtl_sm, zfs_vdev_dtl_sm_blksz, tx);
space_map_write(vd->vdev_dtl_sm, rtsync, SM_ALLOC, SM_NO_VDEVID, tx);
range_tree_vacate(rtsync, NULL, NULL);
range_tree_destroy(rtsync);
/*
* If the object for the space map has changed then dirty
* the top level so that we update the config.
*/
if (object != space_map_object(vd->vdev_dtl_sm)) {
vdev_dbgmsg(vd, "txg %llu, spa %s, DTL old object %llu, "
"new object %llu", (u_longlong_t)txg, spa_name(spa),
(u_longlong_t)object,
(u_longlong_t)space_map_object(vd->vdev_dtl_sm));
vdev_config_dirty(vd->vdev_top);
}
dmu_tx_commit(tx);
}
/*
* Determine whether the specified vdev can be offlined/detached/removed
* without losing data.
*/
boolean_t
vdev_dtl_required(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
vdev_t *tvd = vd->vdev_top;
uint8_t cant_read = vd->vdev_cant_read;
boolean_t required;
ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
if (vd == spa->spa_root_vdev || vd == tvd)
return (B_TRUE);
/*
* Temporarily mark the device as unreadable, and then determine
* whether this results in any DTL outages in the top-level vdev.
* If not, we can safely offline/detach/remove the device.
*/
vd->vdev_cant_read = B_TRUE;
vdev_dtl_reassess(tvd, 0, 0, B_FALSE, B_FALSE);
required = !vdev_dtl_empty(tvd, DTL_OUTAGE);
vd->vdev_cant_read = cant_read;
vdev_dtl_reassess(tvd, 0, 0, B_FALSE, B_FALSE);
if (!required && zio_injection_enabled) {
required = !!zio_handle_device_injection(vd, NULL,
SET_ERROR(ECHILD));
}
return (required);
}
/*
* Determine if resilver is needed, and if so the txg range.
*/
boolean_t
vdev_resilver_needed(vdev_t *vd, uint64_t *minp, uint64_t *maxp)
{
boolean_t needed = B_FALSE;
uint64_t thismin = UINT64_MAX;
uint64_t thismax = 0;
if (vd->vdev_children == 0) {
mutex_enter(&vd->vdev_dtl_lock);
if (!range_tree_is_empty(vd->vdev_dtl[DTL_MISSING]) &&
vdev_writeable(vd)) {
thismin = vdev_dtl_min(vd);
thismax = vdev_dtl_max(vd);
needed = B_TRUE;
}
mutex_exit(&vd->vdev_dtl_lock);
} else {
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
uint64_t cmin, cmax;
if (vdev_resilver_needed(cvd, &cmin, &cmax)) {
thismin = MIN(thismin, cmin);
thismax = MAX(thismax, cmax);
needed = B_TRUE;
}
}
}
if (needed && minp) {
*minp = thismin;
*maxp = thismax;
}
return (needed);
}
/*
* Gets the checkpoint space map object from the vdev's ZAP. On success sm_obj
* will contain either the checkpoint spacemap object or zero if none exists.
* All other errors are returned to the caller.
*/
int
vdev_checkpoint_sm_object(vdev_t *vd, uint64_t *sm_obj)
{
ASSERT0(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER));
if (vd->vdev_top_zap == 0) {
*sm_obj = 0;
return (0);
}
int error = zap_lookup(spa_meta_objset(vd->vdev_spa), vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM, sizeof (uint64_t), 1, sm_obj);
if (error == ENOENT) {
*sm_obj = 0;
error = 0;
}
return (error);
}
int
vdev_load(vdev_t *vd)
{
int children = vd->vdev_children;
int error = 0;
taskq_t *tq = NULL;
/*
* It's only worthwhile to use the taskq for the root vdev, because the
* slow part is metaslab_init, and that only happens for top-level
* vdevs.
*/
if (vd->vdev_ops == &vdev_root_ops && vd->vdev_children > 0) {
tq = taskq_create("vdev_load", children, minclsyspri,
children, children, TASKQ_PREPOPULATE);
}
/*
* Recursively load all children.
*/
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (tq == NULL || vdev_uses_zvols(cvd)) {
cvd->vdev_load_error = vdev_load(cvd);
} else {
VERIFY(taskq_dispatch(tq, vdev_load_child,
cvd, TQ_SLEEP) != TASKQID_INVALID);
}
}
if (tq != NULL) {
taskq_wait(tq);
taskq_destroy(tq);
}
for (int c = 0; c < vd->vdev_children; c++) {
int error = vd->vdev_child[c]->vdev_load_error;
if (error != 0)
return (error);
}
vdev_set_deflate_ratio(vd);
+ if (vd->vdev_ops == &vdev_raidz_ops) {
+ error = vdev_raidz_load(vd);
+ if (error != 0)
+ return (error);
+ }
+
/*
* On spa_load path, grab the allocation bias from our zap
*/
if (vd == vd->vdev_top && vd->vdev_top_zap != 0) {
spa_t *spa = vd->vdev_spa;
char bias_str[64];
error = zap_lookup(spa->spa_meta_objset, vd->vdev_top_zap,
VDEV_TOP_ZAP_ALLOCATION_BIAS, 1, sizeof (bias_str),
bias_str);
if (error == 0) {
ASSERT(vd->vdev_alloc_bias == VDEV_BIAS_NONE);
vd->vdev_alloc_bias = vdev_derive_alloc_bias(bias_str);
} else if (error != ENOENT) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
vdev_dbgmsg(vd, "vdev_load: zap_lookup(top_zap=%llu) "
"failed [error=%d]",
(u_longlong_t)vd->vdev_top_zap, error);
return (error);
}
}
if (vd == vd->vdev_top && vd->vdev_top_zap != 0) {
spa_t *spa = vd->vdev_spa;
uint64_t failfast;
error = zap_lookup(spa->spa_meta_objset, vd->vdev_top_zap,
vdev_prop_to_name(VDEV_PROP_FAILFAST), sizeof (failfast),
1, &failfast);
if (error == 0) {
vd->vdev_failfast = failfast & 1;
} else if (error == ENOENT) {
vd->vdev_failfast = vdev_prop_default_numeric(
VDEV_PROP_FAILFAST);
} else {
vdev_dbgmsg(vd,
"vdev_load: zap_lookup(top_zap=%llu) "
"failed [error=%d]",
(u_longlong_t)vd->vdev_top_zap, error);
}
}
/*
* Load any rebuild state from the top-level vdev zap.
*/
if (vd == vd->vdev_top && vd->vdev_top_zap != 0) {
error = vdev_rebuild_load(vd);
if (error && error != ENOTSUP) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
vdev_dbgmsg(vd, "vdev_load: vdev_rebuild_load "
"failed [error=%d]", error);
return (error);
}
}
if (vd->vdev_top_zap != 0 || vd->vdev_leaf_zap != 0) {
uint64_t zapobj;
if (vd->vdev_top_zap != 0)
zapobj = vd->vdev_top_zap;
else
zapobj = vd->vdev_leaf_zap;
error = vdev_prop_get_int(vd, VDEV_PROP_CHECKSUM_N,
&vd->vdev_checksum_n);
if (error && error != ENOENT)
vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) "
"failed [error=%d]", (u_longlong_t)zapobj, error);
error = vdev_prop_get_int(vd, VDEV_PROP_CHECKSUM_T,
&vd->vdev_checksum_t);
if (error && error != ENOENT)
vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) "
"failed [error=%d]", (u_longlong_t)zapobj, error);
error = vdev_prop_get_int(vd, VDEV_PROP_IO_N,
&vd->vdev_io_n);
if (error && error != ENOENT)
vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) "
"failed [error=%d]", (u_longlong_t)zapobj, error);
error = vdev_prop_get_int(vd, VDEV_PROP_IO_T,
&vd->vdev_io_t);
if (error && error != ENOENT)
vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) "
"failed [error=%d]", (u_longlong_t)zapobj, error);
}
/*
* If this is a top-level vdev, initialize its metaslabs.
*/
if (vd == vd->vdev_top && vdev_is_concrete(vd)) {
vdev_metaslab_group_create(vd);
if (vd->vdev_ashift == 0 || vd->vdev_asize == 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
vdev_dbgmsg(vd, "vdev_load: invalid size. ashift=%llu, "
"asize=%llu", (u_longlong_t)vd->vdev_ashift,
(u_longlong_t)vd->vdev_asize);
return (SET_ERROR(ENXIO));
}
error = vdev_metaslab_init(vd, 0);
if (error != 0) {
vdev_dbgmsg(vd, "vdev_load: metaslab_init failed "
"[error=%d]", error);
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
return (error);
}
uint64_t checkpoint_sm_obj;
error = vdev_checkpoint_sm_object(vd, &checkpoint_sm_obj);
if (error == 0 && checkpoint_sm_obj != 0) {
objset_t *mos = spa_meta_objset(vd->vdev_spa);
ASSERT(vd->vdev_asize != 0);
ASSERT3P(vd->vdev_checkpoint_sm, ==, NULL);
error = space_map_open(&vd->vdev_checkpoint_sm,
mos, checkpoint_sm_obj, 0, vd->vdev_asize,
vd->vdev_ashift);
if (error != 0) {
vdev_dbgmsg(vd, "vdev_load: space_map_open "
"failed for checkpoint spacemap (obj %llu) "
"[error=%d]",
(u_longlong_t)checkpoint_sm_obj, error);
return (error);
}
ASSERT3P(vd->vdev_checkpoint_sm, !=, NULL);
/*
* Since the checkpoint_sm contains free entries
* exclusively we can use space_map_allocated() to
* indicate the cumulative checkpointed space that
* has been freed.
*/
vd->vdev_stat.vs_checkpoint_space =
-space_map_allocated(vd->vdev_checkpoint_sm);
vd->vdev_spa->spa_checkpoint_info.sci_dspace +=
vd->vdev_stat.vs_checkpoint_space;
} else if (error != 0) {
vdev_dbgmsg(vd, "vdev_load: failed to retrieve "
"checkpoint space map object from vdev ZAP "
"[error=%d]", error);
return (error);
}
}
/*
* If this is a leaf vdev, load its DTL.
*/
if (vd->vdev_ops->vdev_op_leaf && (error = vdev_dtl_load(vd)) != 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
vdev_dbgmsg(vd, "vdev_load: vdev_dtl_load failed "
"[error=%d]", error);
return (error);
}
uint64_t obsolete_sm_object;
error = vdev_obsolete_sm_object(vd, &obsolete_sm_object);
if (error == 0 && obsolete_sm_object != 0) {
objset_t *mos = vd->vdev_spa->spa_meta_objset;
ASSERT(vd->vdev_asize != 0);
ASSERT3P(vd->vdev_obsolete_sm, ==, NULL);
if ((error = space_map_open(&vd->vdev_obsolete_sm, mos,
obsolete_sm_object, 0, vd->vdev_asize, 0))) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
vdev_dbgmsg(vd, "vdev_load: space_map_open failed for "
"obsolete spacemap (obj %llu) [error=%d]",
(u_longlong_t)obsolete_sm_object, error);
return (error);
}
} else if (error != 0) {
vdev_dbgmsg(vd, "vdev_load: failed to retrieve obsolete "
"space map object from vdev ZAP [error=%d]", error);
return (error);
}
return (0);
}
/*
* The special vdev case is used for hot spares and l2cache devices. Its
* sole purpose it to set the vdev state for the associated vdev. To do this,
* we make sure that we can open the underlying device, then try to read the
* label, and make sure that the label is sane and that it hasn't been
* repurposed to another pool.
*/
int
vdev_validate_aux(vdev_t *vd)
{
nvlist_t *label;
uint64_t guid, version;
uint64_t state;
if (!vdev_readable(vd))
return (0);
if ((label = vdev_label_read_config(vd, -1ULL)) == NULL) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
return (-1);
}
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_VERSION, &version) != 0 ||
!SPA_VERSION_IS_SUPPORTED(version) ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) != 0 ||
guid != vd->vdev_guid ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE, &state) != 0) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
return (-1);
}
/*
* We don't actually check the pool state here. If it's in fact in
* use by another pool, we update this fact on the fly when requested.
*/
nvlist_free(label);
return (0);
}
static void
vdev_destroy_ms_flush_data(vdev_t *vd, dmu_tx_t *tx)
{
objset_t *mos = spa_meta_objset(vd->vdev_spa);
if (vd->vdev_top_zap == 0)
return;
uint64_t object = 0;
int err = zap_lookup(mos, vd->vdev_top_zap,
VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS, sizeof (uint64_t), 1, &object);
if (err == ENOENT)
return;
VERIFY0(err);
VERIFY0(dmu_object_free(mos, object, tx));
VERIFY0(zap_remove(mos, vd->vdev_top_zap,
VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS, tx));
}
/*
* Free the objects used to store this vdev's spacemaps, and the array
* that points to them.
*/
void
vdev_destroy_spacemaps(vdev_t *vd, dmu_tx_t *tx)
{
if (vd->vdev_ms_array == 0)
return;
objset_t *mos = vd->vdev_spa->spa_meta_objset;
uint64_t array_count = vd->vdev_asize >> vd->vdev_ms_shift;
size_t array_bytes = array_count * sizeof (uint64_t);
uint64_t *smobj_array = kmem_alloc(array_bytes, KM_SLEEP);
VERIFY0(dmu_read(mos, vd->vdev_ms_array, 0,
array_bytes, smobj_array, 0));
for (uint64_t i = 0; i < array_count; i++) {
uint64_t smobj = smobj_array[i];
if (smobj == 0)
continue;
space_map_free_obj(mos, smobj, tx);
}
kmem_free(smobj_array, array_bytes);
VERIFY0(dmu_object_free(mos, vd->vdev_ms_array, tx));
vdev_destroy_ms_flush_data(vd, tx);
vd->vdev_ms_array = 0;
}
static void
vdev_remove_empty_log(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
ASSERT(vd->vdev_islog);
ASSERT(vd == vd->vdev_top);
ASSERT3U(txg, ==, spa_syncing_txg(spa));
dmu_tx_t *tx = dmu_tx_create_assigned(spa_get_dsl(spa), txg);
vdev_destroy_spacemaps(vd, tx);
if (vd->vdev_top_zap != 0) {
vdev_destroy_unlink_zap(vd, vd->vdev_top_zap, tx);
vd->vdev_top_zap = 0;
}
dmu_tx_commit(tx);
}
void
vdev_sync_done(vdev_t *vd, uint64_t txg)
{
metaslab_t *msp;
boolean_t reassess = !txg_list_empty(&vd->vdev_ms_list, TXG_CLEAN(txg));
ASSERT(vdev_is_concrete(vd));
while ((msp = txg_list_remove(&vd->vdev_ms_list, TXG_CLEAN(txg)))
!= NULL)
metaslab_sync_done(msp, txg);
if (reassess) {
metaslab_sync_reassess(vd->vdev_mg);
if (vd->vdev_log_mg != NULL)
metaslab_sync_reassess(vd->vdev_log_mg);
}
}
void
vdev_sync(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
vdev_t *lvd;
metaslab_t *msp;
ASSERT3U(txg, ==, spa->spa_syncing_txg);
dmu_tx_t *tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg);
if (range_tree_space(vd->vdev_obsolete_segments) > 0) {
ASSERT(vd->vdev_removing ||
vd->vdev_ops == &vdev_indirect_ops);
vdev_indirect_sync_obsolete(vd, tx);
/*
* If the vdev is indirect, it can't have dirty
* metaslabs or DTLs.
*/
if (vd->vdev_ops == &vdev_indirect_ops) {
ASSERT(txg_list_empty(&vd->vdev_ms_list, txg));
ASSERT(txg_list_empty(&vd->vdev_dtl_list, txg));
dmu_tx_commit(tx);
return;
}
}
ASSERT(vdev_is_concrete(vd));
if (vd->vdev_ms_array == 0 && vd->vdev_ms_shift != 0 &&
!vd->vdev_removing) {
ASSERT(vd == vd->vdev_top);
ASSERT0(vd->vdev_indirect_config.vic_mapping_object);
vd->vdev_ms_array = dmu_object_alloc(spa->spa_meta_objset,
DMU_OT_OBJECT_ARRAY, 0, DMU_OT_NONE, 0, tx);
ASSERT(vd->vdev_ms_array != 0);
vdev_config_dirty(vd);
}
while ((msp = txg_list_remove(&vd->vdev_ms_list, txg)) != NULL) {
metaslab_sync(msp, txg);
(void) txg_list_add(&vd->vdev_ms_list, msp, TXG_CLEAN(txg));
}
while ((lvd = txg_list_remove(&vd->vdev_dtl_list, txg)) != NULL)
vdev_dtl_sync(lvd, txg);
/*
* If this is an empty log device being removed, destroy the
* metadata associated with it.
*/
if (vd->vdev_islog && vd->vdev_stat.vs_alloc == 0 && vd->vdev_removing)
vdev_remove_empty_log(vd, txg);
(void) txg_list_add(&spa->spa_vdev_txg_list, vd, TXG_CLEAN(txg));
dmu_tx_commit(tx);
}
+/*
+ * Return the amount of space that should be (or was) allocated for the given
+ * psize (compressed block size) in the given TXG. Note that for expanded
+ * RAIDZ vdevs, the size allocated for older BP's may be larger. See
+ * vdev_raidz_asize().
+ */
+uint64_t
+vdev_psize_to_asize_txg(vdev_t *vd, uint64_t psize, uint64_t txg)
+{
+ return (vd->vdev_ops->vdev_op_asize(vd, psize, txg));
+}
+
uint64_t
vdev_psize_to_asize(vdev_t *vd, uint64_t psize)
{
- return (vd->vdev_ops->vdev_op_asize(vd, psize));
+ return (vdev_psize_to_asize_txg(vd, psize, 0));
}
/*
* Mark the given vdev faulted. A faulted vdev behaves as if the device could
* not be opened, and no I/O is attempted.
*/
int
vdev_fault(spa_t *spa, uint64_t guid, vdev_aux_t aux)
{
vdev_t *vd, *tvd;
spa_vdev_state_enter(spa, SCL_NONE);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV)));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP)));
tvd = vd->vdev_top;
/*
* If user did a 'zpool offline -f' then make the fault persist across
* reboots.
*/
if (aux == VDEV_AUX_EXTERNAL_PERSIST) {
/*
* There are two kinds of forced faults: temporary and
* persistent. Temporary faults go away at pool import, while
* persistent faults stay set. Both types of faults can be
* cleared with a zpool clear.
*
* We tell if a vdev is persistently faulted by looking at the
* ZPOOL_CONFIG_AUX_STATE nvpair. If it's set to "external" at
* import then it's a persistent fault. Otherwise, it's
* temporary. We get ZPOOL_CONFIG_AUX_STATE set to "external"
* by setting vd.vdev_stat.vs_aux to VDEV_AUX_EXTERNAL. This
* tells vdev_config_generate() (which gets run later) to set
* ZPOOL_CONFIG_AUX_STATE to "external" in the nvlist.
*/
vd->vdev_stat.vs_aux = VDEV_AUX_EXTERNAL;
vd->vdev_tmpoffline = B_FALSE;
aux = VDEV_AUX_EXTERNAL;
} else {
vd->vdev_tmpoffline = B_TRUE;
}
/*
* We don't directly use the aux state here, but if we do a
* vdev_reopen(), we need this value to be present to remember why we
* were faulted.
*/
vd->vdev_label_aux = aux;
/*
* Faulted state takes precedence over degraded.
*/
vd->vdev_delayed_close = B_FALSE;
vd->vdev_faulted = 1ULL;
vd->vdev_degraded = 0ULL;
vdev_set_state(vd, B_FALSE, VDEV_STATE_FAULTED, aux);
/*
* If this device has the only valid copy of the data, then
* back off and simply mark the vdev as degraded instead.
*/
if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_dtl_required(vd)) {
vd->vdev_degraded = 1ULL;
vd->vdev_faulted = 0ULL;
/*
* If we reopen the device and it's not dead, only then do we
* mark it degraded.
*/
vdev_reopen(tvd);
if (vdev_readable(vd))
vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, aux);
}
return (spa_vdev_state_exit(spa, vd, 0));
}
/*
* Mark the given vdev degraded. A degraded vdev is purely an indication to the
* user that something is wrong. The vdev continues to operate as normal as far
* as I/O is concerned.
*/
int
vdev_degrade(spa_t *spa, uint64_t guid, vdev_aux_t aux)
{
vdev_t *vd;
spa_vdev_state_enter(spa, SCL_NONE);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV)));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP)));
/*
* If the vdev is already faulted, then don't do anything.
*/
if (vd->vdev_faulted || vd->vdev_degraded)
return (spa_vdev_state_exit(spa, NULL, 0));
vd->vdev_degraded = 1ULL;
if (!vdev_is_dead(vd))
vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED,
aux);
return (spa_vdev_state_exit(spa, vd, 0));
}
int
vdev_remove_wanted(spa_t *spa, uint64_t guid)
{
vdev_t *vd;
spa_vdev_state_enter(spa, SCL_NONE);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV)));
/*
* If the vdev is already removed, or expanding which can trigger
* repartition add/remove events, then don't do anything.
*/
if (vd->vdev_removed || vd->vdev_expanding)
return (spa_vdev_state_exit(spa, NULL, 0));
/*
* Confirm the vdev has been removed, otherwise don't do anything.
*/
if (vd->vdev_ops->vdev_op_leaf && !zio_wait(vdev_probe(vd, NULL)))
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(EEXIST)));
vd->vdev_remove_wanted = B_TRUE;
spa_async_request(spa, SPA_ASYNC_REMOVE);
return (spa_vdev_state_exit(spa, vd, 0));
}
/*
* Online the given vdev.
*
* If 'ZFS_ONLINE_UNSPARE' is set, it implies two things. First, any attached
* spare device should be detached when the device finishes resilvering.
* Second, the online should be treated like a 'test' online case, so no FMA
* events are generated if the device fails to open.
*/
int
vdev_online(spa_t *spa, uint64_t guid, uint64_t flags, vdev_state_t *newstate)
{
vdev_t *vd, *tvd, *pvd, *rvd = spa->spa_root_vdev;
boolean_t wasoffline;
vdev_state_t oldstate;
spa_vdev_state_enter(spa, SCL_NONE);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV)));
- if (!vd->vdev_ops->vdev_op_leaf)
- return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP)));
-
wasoffline = (vd->vdev_offline || vd->vdev_tmpoffline);
oldstate = vd->vdev_state;
tvd = vd->vdev_top;
vd->vdev_offline = B_FALSE;
vd->vdev_tmpoffline = B_FALSE;
vd->vdev_checkremove = !!(flags & ZFS_ONLINE_CHECKREMOVE);
vd->vdev_forcefault = !!(flags & ZFS_ONLINE_FORCEFAULT);
/* XXX - L2ARC 1.0 does not support expansion */
if (!vd->vdev_aux) {
for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent)
pvd->vdev_expanding = !!((flags & ZFS_ONLINE_EXPAND) ||
spa->spa_autoexpand);
vd->vdev_expansion_time = gethrestime_sec();
}
vdev_reopen(tvd);
vd->vdev_checkremove = vd->vdev_forcefault = B_FALSE;
if (!vd->vdev_aux) {
for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent)
pvd->vdev_expanding = B_FALSE;
}
if (newstate)
*newstate = vd->vdev_state;
if ((flags & ZFS_ONLINE_UNSPARE) &&
!vdev_is_dead(vd) && vd->vdev_parent &&
vd->vdev_parent->vdev_ops == &vdev_spare_ops &&
vd->vdev_parent->vdev_child[0] == vd)
vd->vdev_unspare = B_TRUE;
if ((flags & ZFS_ONLINE_EXPAND) || spa->spa_autoexpand) {
/* XXX - L2ARC 1.0 does not support expansion */
if (vd->vdev_aux)
return (spa_vdev_state_exit(spa, vd, ENOTSUP));
spa->spa_ccw_fail_time = 0;
spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
}
/* Restart initializing if necessary */
mutex_enter(&vd->vdev_initialize_lock);
if (vdev_writeable(vd) &&
vd->vdev_initialize_thread == NULL &&
vd->vdev_initialize_state == VDEV_INITIALIZE_ACTIVE) {
(void) vdev_initialize(vd);
}
mutex_exit(&vd->vdev_initialize_lock);
/*
* Restart trimming if necessary. We do not restart trimming for cache
* devices here. This is triggered by l2arc_rebuild_vdev()
* asynchronously for the whole device or in l2arc_evict() as it evicts
* space for upcoming writes.
*/
mutex_enter(&vd->vdev_trim_lock);
if (vdev_writeable(vd) && !vd->vdev_isl2cache &&
vd->vdev_trim_thread == NULL &&
vd->vdev_trim_state == VDEV_TRIM_ACTIVE) {
(void) vdev_trim(vd, vd->vdev_trim_rate, vd->vdev_trim_partial,
vd->vdev_trim_secure);
}
mutex_exit(&vd->vdev_trim_lock);
if (wasoffline ||
(oldstate < VDEV_STATE_DEGRADED &&
vd->vdev_state >= VDEV_STATE_DEGRADED)) {
spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_ONLINE);
/*
* Asynchronously detach spare vdev if resilver or
* rebuild is not required
*/
if (vd->vdev_unspare &&
!dsl_scan_resilvering(spa->spa_dsl_pool) &&
!dsl_scan_resilver_scheduled(spa->spa_dsl_pool) &&
!vdev_rebuild_active(tvd))
spa_async_request(spa, SPA_ASYNC_DETACH_SPARE);
}
return (spa_vdev_state_exit(spa, vd, 0));
}
static int
vdev_offline_locked(spa_t *spa, uint64_t guid, uint64_t flags)
{
vdev_t *vd, *tvd;
int error = 0;
uint64_t generation;
metaslab_group_t *mg;
top:
spa_vdev_state_enter(spa, SCL_ALLOC);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV)));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP)));
if (vd->vdev_ops == &vdev_draid_spare_ops)
return (spa_vdev_state_exit(spa, NULL, ENOTSUP));
tvd = vd->vdev_top;
mg = tvd->vdev_mg;
generation = spa->spa_config_generation + 1;
/*
* If the device isn't already offline, try to offline it.
*/
if (!vd->vdev_offline) {
/*
* If this device has the only valid copy of some data,
* don't allow it to be offlined. Log devices are always
* expendable.
*/
if (!tvd->vdev_islog && vd->vdev_aux == NULL &&
vdev_dtl_required(vd))
return (spa_vdev_state_exit(spa, NULL,
SET_ERROR(EBUSY)));
/*
* If the top-level is a slog and it has had allocations
* then proceed. We check that the vdev's metaslab group
* is not NULL since it's possible that we may have just
* added this vdev but not yet initialized its metaslabs.
*/
if (tvd->vdev_islog && mg != NULL) {
/*
* Prevent any future allocations.
*/
ASSERT3P(tvd->vdev_log_mg, ==, NULL);
metaslab_group_passivate(mg);
(void) spa_vdev_state_exit(spa, vd, 0);
error = spa_reset_logs(spa);
/*
* If the log device was successfully reset but has
* checkpointed data, do not offline it.
*/
if (error == 0 &&
tvd->vdev_checkpoint_sm != NULL) {
ASSERT3U(space_map_allocated(
tvd->vdev_checkpoint_sm), !=, 0);
error = ZFS_ERR_CHECKPOINT_EXISTS;
}
spa_vdev_state_enter(spa, SCL_ALLOC);
/*
* Check to see if the config has changed.
*/
if (error || generation != spa->spa_config_generation) {
metaslab_group_activate(mg);
if (error)
return (spa_vdev_state_exit(spa,
vd, error));
(void) spa_vdev_state_exit(spa, vd, 0);
goto top;
}
ASSERT0(tvd->vdev_stat.vs_alloc);
}
/*
* Offline this device and reopen its top-level vdev.
* If the top-level vdev is a log device then just offline
* it. Otherwise, if this action results in the top-level
* vdev becoming unusable, undo it and fail the request.
*/
vd->vdev_offline = B_TRUE;
vdev_reopen(tvd);
if (!tvd->vdev_islog && vd->vdev_aux == NULL &&
vdev_is_dead(tvd)) {
vd->vdev_offline = B_FALSE;
vdev_reopen(tvd);
return (spa_vdev_state_exit(spa, NULL,
SET_ERROR(EBUSY)));
}
/*
* Add the device back into the metaslab rotor so that
* once we online the device it's open for business.
*/
if (tvd->vdev_islog && mg != NULL)
metaslab_group_activate(mg);
}
vd->vdev_tmpoffline = !!(flags & ZFS_OFFLINE_TEMPORARY);
return (spa_vdev_state_exit(spa, vd, 0));
}
int
vdev_offline(spa_t *spa, uint64_t guid, uint64_t flags)
{
int error;
mutex_enter(&spa->spa_vdev_top_lock);
error = vdev_offline_locked(spa, guid, flags);
mutex_exit(&spa->spa_vdev_top_lock);
return (error);
}
/*
* Clear the error counts associated with this vdev. Unlike vdev_online() and
* vdev_offline(), we assume the spa config is locked. We also clear all
* children. If 'vd' is NULL, then the user wants to clear all vdevs.
*/
void
vdev_clear(spa_t *spa, vdev_t *vd)
{
vdev_t *rvd = spa->spa_root_vdev;
ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
if (vd == NULL)
vd = rvd;
vd->vdev_stat.vs_read_errors = 0;
vd->vdev_stat.vs_write_errors = 0;
vd->vdev_stat.vs_checksum_errors = 0;
vd->vdev_stat.vs_slow_ios = 0;
for (int c = 0; c < vd->vdev_children; c++)
vdev_clear(spa, vd->vdev_child[c]);
/*
* It makes no sense to "clear" an indirect or removed vdev.
*/
if (!vdev_is_concrete(vd) || vd->vdev_removed)
return;
/*
* If we're in the FAULTED state or have experienced failed I/O, then
* clear the persistent state and attempt to reopen the device. We
* also mark the vdev config dirty, so that the new faulted state is
* written out to disk.
*/
if (vd->vdev_faulted || vd->vdev_degraded ||
!vdev_readable(vd) || !vdev_writeable(vd)) {
/*
* When reopening in response to a clear event, it may be due to
* a fmadm repair request. In this case, if the device is
* still broken, we want to still post the ereport again.
*/
vd->vdev_forcefault = B_TRUE;
vd->vdev_faulted = vd->vdev_degraded = 0ULL;
vd->vdev_cant_read = B_FALSE;
vd->vdev_cant_write = B_FALSE;
vd->vdev_stat.vs_aux = 0;
vdev_reopen(vd == rvd ? rvd : vd->vdev_top);
vd->vdev_forcefault = B_FALSE;
if (vd != rvd && vdev_writeable(vd->vdev_top))
vdev_state_dirty(vd->vdev_top);
/* If a resilver isn't required, check if vdevs can be culled */
if (vd->vdev_aux == NULL && !vdev_is_dead(vd) &&
!dsl_scan_resilvering(spa->spa_dsl_pool) &&
!dsl_scan_resilver_scheduled(spa->spa_dsl_pool))
spa_async_request(spa, SPA_ASYNC_RESILVER_DONE);
spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_CLEAR);
}
/*
* When clearing a FMA-diagnosed fault, we always want to
* unspare the device, as we assume that the original spare was
* done in response to the FMA fault.
*/
if (!vdev_is_dead(vd) && vd->vdev_parent != NULL &&
vd->vdev_parent->vdev_ops == &vdev_spare_ops &&
vd->vdev_parent->vdev_child[0] == vd)
vd->vdev_unspare = B_TRUE;
/* Clear recent error events cache (i.e. duplicate events tracking) */
zfs_ereport_clear(spa, vd);
}
boolean_t
vdev_is_dead(vdev_t *vd)
{
/*
* Holes and missing devices are always considered "dead".
* This simplifies the code since we don't have to check for
* these types of devices in the various code paths.
* Instead we rely on the fact that we skip over dead devices
* before issuing I/O to them.
*/
return (vd->vdev_state < VDEV_STATE_DEGRADED ||
vd->vdev_ops == &vdev_hole_ops ||
vd->vdev_ops == &vdev_missing_ops);
}
boolean_t
vdev_readable(vdev_t *vd)
{
return (!vdev_is_dead(vd) && !vd->vdev_cant_read);
}
boolean_t
vdev_writeable(vdev_t *vd)
{
return (!vdev_is_dead(vd) && !vd->vdev_cant_write &&
vdev_is_concrete(vd));
}
boolean_t
vdev_allocatable(vdev_t *vd)
{
uint64_t state = vd->vdev_state;
/*
* We currently allow allocations from vdevs which may be in the
* process of reopening (i.e. VDEV_STATE_CLOSED). If the device
* fails to reopen then we'll catch it later when we're holding
* the proper locks. Note that we have to get the vdev state
* in a local variable because although it changes atomically,
* we're asking two separate questions about it.
*/
return (!(state < VDEV_STATE_DEGRADED && state != VDEV_STATE_CLOSED) &&
!vd->vdev_cant_write && vdev_is_concrete(vd) &&
vd->vdev_mg->mg_initialized);
}
boolean_t
vdev_accessible(vdev_t *vd, zio_t *zio)
{
ASSERT(zio->io_vd == vd);
if (vdev_is_dead(vd) || vd->vdev_remove_wanted)
return (B_FALSE);
if (zio->io_type == ZIO_TYPE_READ)
return (!vd->vdev_cant_read);
if (zio->io_type == ZIO_TYPE_WRITE)
return (!vd->vdev_cant_write);
return (B_TRUE);
}
static void
vdev_get_child_stat(vdev_t *cvd, vdev_stat_t *vs, vdev_stat_t *cvs)
{
/*
* Exclude the dRAID spare when aggregating to avoid double counting
* the ops and bytes. These IOs are counted by the physical leaves.
*/
if (cvd->vdev_ops == &vdev_draid_spare_ops)
return;
for (int t = 0; t < VS_ZIO_TYPES; t++) {
vs->vs_ops[t] += cvs->vs_ops[t];
vs->vs_bytes[t] += cvs->vs_bytes[t];
}
cvs->vs_scan_removing = cvd->vdev_removing;
}
/*
* Get extended stats
*/
static void
vdev_get_child_stat_ex(vdev_t *cvd, vdev_stat_ex_t *vsx, vdev_stat_ex_t *cvsx)
{
(void) cvd;
int t, b;
for (t = 0; t < ZIO_TYPES; t++) {
for (b = 0; b < ARRAY_SIZE(vsx->vsx_disk_histo[0]); b++)
vsx->vsx_disk_histo[t][b] += cvsx->vsx_disk_histo[t][b];
for (b = 0; b < ARRAY_SIZE(vsx->vsx_total_histo[0]); b++) {
vsx->vsx_total_histo[t][b] +=
cvsx->vsx_total_histo[t][b];
}
}
for (t = 0; t < ZIO_PRIORITY_NUM_QUEUEABLE; t++) {
for (b = 0; b < ARRAY_SIZE(vsx->vsx_queue_histo[0]); b++) {
vsx->vsx_queue_histo[t][b] +=
cvsx->vsx_queue_histo[t][b];
}
vsx->vsx_active_queue[t] += cvsx->vsx_active_queue[t];
vsx->vsx_pend_queue[t] += cvsx->vsx_pend_queue[t];
for (b = 0; b < ARRAY_SIZE(vsx->vsx_ind_histo[0]); b++)
vsx->vsx_ind_histo[t][b] += cvsx->vsx_ind_histo[t][b];
for (b = 0; b < ARRAY_SIZE(vsx->vsx_agg_histo[0]); b++)
vsx->vsx_agg_histo[t][b] += cvsx->vsx_agg_histo[t][b];
}
}
boolean_t
vdev_is_spacemap_addressable(vdev_t *vd)
{
if (spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_SPACEMAP_V2))
return (B_TRUE);
/*
* If double-word space map entries are not enabled we assume
* 47 bits of the space map entry are dedicated to the entry's
* offset (see SM_OFFSET_BITS in space_map.h). We then use that
* to calculate the maximum address that can be described by a
* space map entry for the given device.
*/
uint64_t shift = vd->vdev_ashift + SM_OFFSET_BITS;
if (shift >= 63) /* detect potential overflow */
return (B_TRUE);
return (vd->vdev_asize < (1ULL << shift));
}
/*
* Get statistics for the given vdev.
*/
static void
vdev_get_stats_ex_impl(vdev_t *vd, vdev_stat_t *vs, vdev_stat_ex_t *vsx)
{
int t;
/*
* If we're getting stats on the root vdev, aggregate the I/O counts
* over all top-level vdevs (i.e. the direct children of the root).
*/
if (!vd->vdev_ops->vdev_op_leaf) {
if (vs) {
memset(vs->vs_ops, 0, sizeof (vs->vs_ops));
memset(vs->vs_bytes, 0, sizeof (vs->vs_bytes));
}
if (vsx)
memset(vsx, 0, sizeof (*vsx));
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
vdev_stat_t *cvs = &cvd->vdev_stat;
vdev_stat_ex_t *cvsx = &cvd->vdev_stat_ex;
vdev_get_stats_ex_impl(cvd, cvs, cvsx);
if (vs)
vdev_get_child_stat(cvd, vs, cvs);
if (vsx)
vdev_get_child_stat_ex(cvd, vsx, cvsx);
}
} else {
/*
* We're a leaf. Just copy our ZIO active queue stats in. The
* other leaf stats are updated in vdev_stat_update().
*/
if (!vsx)
return;
memcpy(vsx, &vd->vdev_stat_ex, sizeof (vd->vdev_stat_ex));
for (t = 0; t < ZIO_PRIORITY_NUM_QUEUEABLE; t++) {
vsx->vsx_active_queue[t] = vd->vdev_queue.vq_cactive[t];
vsx->vsx_pend_queue[t] = vdev_queue_class_length(vd, t);
}
}
}
void
vdev_get_stats_ex(vdev_t *vd, vdev_stat_t *vs, vdev_stat_ex_t *vsx)
{
vdev_t *tvd = vd->vdev_top;
mutex_enter(&vd->vdev_stat_lock);
if (vs) {
memcpy(vs, &vd->vdev_stat, sizeof (*vs));
vs->vs_timestamp = gethrtime() - vs->vs_timestamp;
vs->vs_state = vd->vdev_state;
vs->vs_rsize = vdev_get_min_asize(vd);
if (vd->vdev_ops->vdev_op_leaf) {
vs->vs_pspace = vd->vdev_psize;
vs->vs_rsize += VDEV_LABEL_START_SIZE +
VDEV_LABEL_END_SIZE;
/*
* Report initializing progress. Since we don't
* have the initializing locks held, this is only
* an estimate (although a fairly accurate one).
*/
vs->vs_initialize_bytes_done =
vd->vdev_initialize_bytes_done;
vs->vs_initialize_bytes_est =
vd->vdev_initialize_bytes_est;
vs->vs_initialize_state = vd->vdev_initialize_state;
vs->vs_initialize_action_time =
vd->vdev_initialize_action_time;
/*
* Report manual TRIM progress. Since we don't have
* the manual TRIM locks held, this is only an
* estimate (although fairly accurate one).
*/
vs->vs_trim_notsup = !vd->vdev_has_trim;
vs->vs_trim_bytes_done = vd->vdev_trim_bytes_done;
vs->vs_trim_bytes_est = vd->vdev_trim_bytes_est;
vs->vs_trim_state = vd->vdev_trim_state;
vs->vs_trim_action_time = vd->vdev_trim_action_time;
/* Set when there is a deferred resilver. */
vs->vs_resilver_deferred = vd->vdev_resilver_deferred;
}
/*
* Report expandable space on top-level, non-auxiliary devices
* only. The expandable space is reported in terms of metaslab
* sized units since that determines how much space the pool
* can expand.
*/
if (vd->vdev_aux == NULL && tvd != NULL) {
vs->vs_esize = P2ALIGN(
vd->vdev_max_asize - vd->vdev_asize,
1ULL << tvd->vdev_ms_shift);
}
vs->vs_configured_ashift = vd->vdev_top != NULL
? vd->vdev_top->vdev_ashift : vd->vdev_ashift;
vs->vs_logical_ashift = vd->vdev_logical_ashift;
if (vd->vdev_physical_ashift <= ASHIFT_MAX)
vs->vs_physical_ashift = vd->vdev_physical_ashift;
else
vs->vs_physical_ashift = 0;
/*
* Report fragmentation and rebuild progress for top-level,
* non-auxiliary, concrete devices.
*/
if (vd->vdev_aux == NULL && vd == vd->vdev_top &&
vdev_is_concrete(vd)) {
/*
* The vdev fragmentation rating doesn't take into
* account the embedded slog metaslab (vdev_log_mg).
* Since it's only one metaslab, it would have a tiny
* impact on the overall fragmentation.
*/
vs->vs_fragmentation = (vd->vdev_mg != NULL) ?
vd->vdev_mg->mg_fragmentation : 0;
}
vs->vs_noalloc = MAX(vd->vdev_noalloc,
tvd ? tvd->vdev_noalloc : 0);
}
vdev_get_stats_ex_impl(vd, vs, vsx);
mutex_exit(&vd->vdev_stat_lock);
}
void
vdev_get_stats(vdev_t *vd, vdev_stat_t *vs)
{
return (vdev_get_stats_ex(vd, vs, NULL));
}
void
vdev_clear_stats(vdev_t *vd)
{
mutex_enter(&vd->vdev_stat_lock);
vd->vdev_stat.vs_space = 0;
vd->vdev_stat.vs_dspace = 0;
vd->vdev_stat.vs_alloc = 0;
mutex_exit(&vd->vdev_stat_lock);
}
void
vdev_scan_stat_init(vdev_t *vd)
{
vdev_stat_t *vs = &vd->vdev_stat;
for (int c = 0; c < vd->vdev_children; c++)
vdev_scan_stat_init(vd->vdev_child[c]);
mutex_enter(&vd->vdev_stat_lock);
vs->vs_scan_processed = 0;
mutex_exit(&vd->vdev_stat_lock);
}
void
vdev_stat_update(zio_t *zio, uint64_t psize)
{
spa_t *spa = zio->io_spa;
vdev_t *rvd = spa->spa_root_vdev;
vdev_t *vd = zio->io_vd ? zio->io_vd : rvd;
vdev_t *pvd;
uint64_t txg = zio->io_txg;
/* Suppress ASAN false positive */
#ifdef __SANITIZE_ADDRESS__
vdev_stat_t *vs = vd ? &vd->vdev_stat : NULL;
vdev_stat_ex_t *vsx = vd ? &vd->vdev_stat_ex : NULL;
#else
vdev_stat_t *vs = &vd->vdev_stat;
vdev_stat_ex_t *vsx = &vd->vdev_stat_ex;
#endif
zio_type_t type = zio->io_type;
int flags = zio->io_flags;
/*
* If this i/o is a gang leader, it didn't do any actual work.
*/
if (zio->io_gang_tree)
return;
if (zio->io_error == 0) {
/*
* If this is a root i/o, don't count it -- we've already
* counted the top-level vdevs, and vdev_get_stats() will
* aggregate them when asked. This reduces contention on
* the root vdev_stat_lock and implicitly handles blocks
* that compress away to holes, for which there is no i/o.
* (Holes never create vdev children, so all the counters
* remain zero, which is what we want.)
*
* Note: this only applies to successful i/o (io_error == 0)
* because unlike i/o counts, errors are not additive.
* When reading a ditto block, for example, failure of
* one top-level vdev does not imply a root-level error.
*/
if (vd == rvd)
return;
ASSERT(vd == zio->io_vd);
if (flags & ZIO_FLAG_IO_BYPASS)
return;
mutex_enter(&vd->vdev_stat_lock);
if (flags & ZIO_FLAG_IO_REPAIR) {
/*
* Repair is the result of a resilver issued by the
* scan thread (spa_sync).
*/
if (flags & ZIO_FLAG_SCAN_THREAD) {
dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan;
dsl_scan_phys_t *scn_phys = &scn->scn_phys;
uint64_t *processed = &scn_phys->scn_processed;
if (vd->vdev_ops->vdev_op_leaf)
atomic_add_64(processed, psize);
vs->vs_scan_processed += psize;
}
/*
* Repair is the result of a rebuild issued by the
* rebuild thread (vdev_rebuild_thread). To avoid
* double counting repaired bytes the virtual dRAID
* spare vdev is excluded from the processed bytes.
*/
if (zio->io_priority == ZIO_PRIORITY_REBUILD) {
vdev_t *tvd = vd->vdev_top;
vdev_rebuild_t *vr = &tvd->vdev_rebuild_config;
vdev_rebuild_phys_t *vrp = &vr->vr_rebuild_phys;
uint64_t *rebuilt = &vrp->vrp_bytes_rebuilt;
if (vd->vdev_ops->vdev_op_leaf &&
vd->vdev_ops != &vdev_draid_spare_ops) {
atomic_add_64(rebuilt, psize);
}
vs->vs_rebuild_processed += psize;
}
if (flags & ZIO_FLAG_SELF_HEAL)
vs->vs_self_healed += psize;
}
/*
* The bytes/ops/histograms are recorded at the leaf level and
* aggregated into the higher level vdevs in vdev_get_stats().
*/
if (vd->vdev_ops->vdev_op_leaf &&
(zio->io_priority < ZIO_PRIORITY_NUM_QUEUEABLE)) {
zio_type_t vs_type = type;
zio_priority_t priority = zio->io_priority;
/*
* TRIM ops and bytes are reported to user space as
* ZIO_TYPE_IOCTL. This is done to preserve the
* vdev_stat_t structure layout for user space.
*/
if (type == ZIO_TYPE_TRIM)
vs_type = ZIO_TYPE_IOCTL;
/*
* Solely for the purposes of 'zpool iostat -lqrw'
* reporting use the priority to categorize the IO.
* Only the following are reported to user space:
*
* ZIO_PRIORITY_SYNC_READ,
* ZIO_PRIORITY_SYNC_WRITE,
* ZIO_PRIORITY_ASYNC_READ,
* ZIO_PRIORITY_ASYNC_WRITE,
* ZIO_PRIORITY_SCRUB,
* ZIO_PRIORITY_TRIM,
* ZIO_PRIORITY_REBUILD.
*/
if (priority == ZIO_PRIORITY_INITIALIZING) {
ASSERT3U(type, ==, ZIO_TYPE_WRITE);
priority = ZIO_PRIORITY_ASYNC_WRITE;
} else if (priority == ZIO_PRIORITY_REMOVAL) {
priority = ((type == ZIO_TYPE_WRITE) ?
ZIO_PRIORITY_ASYNC_WRITE :
ZIO_PRIORITY_ASYNC_READ);
}
vs->vs_ops[vs_type]++;
vs->vs_bytes[vs_type] += psize;
if (flags & ZIO_FLAG_DELEGATED) {
vsx->vsx_agg_histo[priority]
[RQ_HISTO(zio->io_size)]++;
} else {
vsx->vsx_ind_histo[priority]
[RQ_HISTO(zio->io_size)]++;
}
if (zio->io_delta && zio->io_delay) {
vsx->vsx_queue_histo[priority]
[L_HISTO(zio->io_delta - zio->io_delay)]++;
vsx->vsx_disk_histo[type]
[L_HISTO(zio->io_delay)]++;
vsx->vsx_total_histo[type]
[L_HISTO(zio->io_delta)]++;
}
}
mutex_exit(&vd->vdev_stat_lock);
return;
}
if (flags & ZIO_FLAG_SPECULATIVE)
return;
/*
* If this is an I/O error that is going to be retried, then ignore the
* error. Otherwise, the user may interpret B_FAILFAST I/O errors as
* hard errors, when in reality they can happen for any number of
* innocuous reasons (bus resets, MPxIO link failure, etc).
*/
if (zio->io_error == EIO &&
!(zio->io_flags & ZIO_FLAG_IO_RETRY))
return;
/*
* Intent logs writes won't propagate their error to the root
* I/O so don't mark these types of failures as pool-level
* errors.
*/
if (zio->io_vd == NULL && (zio->io_flags & ZIO_FLAG_DONT_PROPAGATE))
return;
if (type == ZIO_TYPE_WRITE && txg != 0 &&
(!(flags & ZIO_FLAG_IO_REPAIR) ||
(flags & ZIO_FLAG_SCAN_THREAD) ||
spa->spa_claiming)) {
/*
* This is either a normal write (not a repair), or it's
* a repair induced by the scrub thread, or it's a repair
* made by zil_claim() during spa_load() in the first txg.
* In the normal case, we commit the DTL change in the same
* txg as the block was born. In the scrub-induced repair
* case, we know that scrubs run in first-pass syncing context,
* so we commit the DTL change in spa_syncing_txg(spa).
* In the zil_claim() case, we commit in spa_first_txg(spa).
*
* We currently do not make DTL entries for failed spontaneous
* self-healing writes triggered by normal (non-scrubbing)
* reads, because we have no transactional context in which to
* do so -- and it's not clear that it'd be desirable anyway.
*/
if (vd->vdev_ops->vdev_op_leaf) {
uint64_t commit_txg = txg;
if (flags & ZIO_FLAG_SCAN_THREAD) {
ASSERT(flags & ZIO_FLAG_IO_REPAIR);
ASSERT(spa_sync_pass(spa) == 1);
vdev_dtl_dirty(vd, DTL_SCRUB, txg, 1);
commit_txg = spa_syncing_txg(spa);
} else if (spa->spa_claiming) {
ASSERT(flags & ZIO_FLAG_IO_REPAIR);
commit_txg = spa_first_txg(spa);
}
ASSERT(commit_txg >= spa_syncing_txg(spa));
if (vdev_dtl_contains(vd, DTL_MISSING, txg, 1))
return;
for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent)
vdev_dtl_dirty(pvd, DTL_PARTIAL, txg, 1);
vdev_dirty(vd->vdev_top, VDD_DTL, vd, commit_txg);
}
if (vd != rvd)
vdev_dtl_dirty(vd, DTL_MISSING, txg, 1);
}
}
int64_t
vdev_deflated_space(vdev_t *vd, int64_t space)
{
ASSERT((space & (SPA_MINBLOCKSIZE-1)) == 0);
ASSERT(vd->vdev_deflate_ratio != 0 || vd->vdev_isl2cache);
return ((space >> SPA_MINBLOCKSHIFT) * vd->vdev_deflate_ratio);
}
/*
* Update the in-core space usage stats for this vdev, its metaslab class,
* and the root vdev.
*/
void
vdev_space_update(vdev_t *vd, int64_t alloc_delta, int64_t defer_delta,
int64_t space_delta)
{
(void) defer_delta;
int64_t dspace_delta;
spa_t *spa = vd->vdev_spa;
vdev_t *rvd = spa->spa_root_vdev;
ASSERT(vd == vd->vdev_top);
/*
* Apply the inverse of the psize-to-asize (ie. RAID-Z) space-expansion
* factor. We must calculate this here and not at the root vdev
* because the root vdev's psize-to-asize is simply the max of its
* children's, thus not accurate enough for us.
*/
dspace_delta = vdev_deflated_space(vd, space_delta);
mutex_enter(&vd->vdev_stat_lock);
/* ensure we won't underflow */
if (alloc_delta < 0) {
ASSERT3U(vd->vdev_stat.vs_alloc, >=, -alloc_delta);
}
vd->vdev_stat.vs_alloc += alloc_delta;
vd->vdev_stat.vs_space += space_delta;
vd->vdev_stat.vs_dspace += dspace_delta;
mutex_exit(&vd->vdev_stat_lock);
/* every class but log contributes to root space stats */
if (vd->vdev_mg != NULL && !vd->vdev_islog) {
ASSERT(!vd->vdev_isl2cache);
mutex_enter(&rvd->vdev_stat_lock);
rvd->vdev_stat.vs_alloc += alloc_delta;
rvd->vdev_stat.vs_space += space_delta;
rvd->vdev_stat.vs_dspace += dspace_delta;
mutex_exit(&rvd->vdev_stat_lock);
}
/* Note: metaslab_class_space_update moved to metaslab_space_update */
}
/*
* Mark a top-level vdev's config as dirty, placing it on the dirty list
* so that it will be written out next time the vdev configuration is synced.
* If the root vdev is specified (vdev_top == NULL), dirty all top-level vdevs.
*/
void
vdev_config_dirty(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
vdev_t *rvd = spa->spa_root_vdev;
int c;
ASSERT(spa_writeable(spa));
/*
* If this is an aux vdev (as with l2cache and spare devices), then we
* update the vdev config manually and set the sync flag.
*/
if (vd->vdev_aux != NULL) {
spa_aux_vdev_t *sav = vd->vdev_aux;
nvlist_t **aux;
uint_t naux;
for (c = 0; c < sav->sav_count; c++) {
if (sav->sav_vdevs[c] == vd)
break;
}
if (c == sav->sav_count) {
/*
* We're being removed. There's nothing more to do.
*/
ASSERT(sav->sav_sync == B_TRUE);
return;
}
sav->sav_sync = B_TRUE;
if (nvlist_lookup_nvlist_array(sav->sav_config,
ZPOOL_CONFIG_L2CACHE, &aux, &naux) != 0) {
VERIFY(nvlist_lookup_nvlist_array(sav->sav_config,
ZPOOL_CONFIG_SPARES, &aux, &naux) == 0);
}
ASSERT(c < naux);
/*
* Setting the nvlist in the middle if the array is a little
* sketchy, but it will work.
*/
nvlist_free(aux[c]);
aux[c] = vdev_config_generate(spa, vd, B_TRUE, 0);
return;
}
/*
* The dirty list is protected by the SCL_CONFIG lock. The caller
* must either hold SCL_CONFIG as writer, or must be the sync thread
* (which holds SCL_CONFIG as reader). There's only one sync thread,
* so this is sufficient to ensure mutual exclusion.
*/
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_WRITER) ||
(dsl_pool_sync_context(spa_get_dsl(spa)) &&
spa_config_held(spa, SCL_CONFIG, RW_READER)));
if (vd == rvd) {
for (c = 0; c < rvd->vdev_children; c++)
vdev_config_dirty(rvd->vdev_child[c]);
} else {
ASSERT(vd == vd->vdev_top);
if (!list_link_active(&vd->vdev_config_dirty_node) &&
vdev_is_concrete(vd)) {
list_insert_head(&spa->spa_config_dirty_list, vd);
}
}
}
void
vdev_config_clean(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_WRITER) ||
(dsl_pool_sync_context(spa_get_dsl(spa)) &&
spa_config_held(spa, SCL_CONFIG, RW_READER)));
ASSERT(list_link_active(&vd->vdev_config_dirty_node));
list_remove(&spa->spa_config_dirty_list, vd);
}
/*
* Mark a top-level vdev's state as dirty, so that the next pass of
* spa_sync() can convert this into vdev_config_dirty(). We distinguish
* the state changes from larger config changes because they require
* much less locking, and are often needed for administrative actions.
*/
void
vdev_state_dirty(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_writeable(spa));
ASSERT(vd == vd->vdev_top);
/*
* The state list is protected by the SCL_STATE lock. The caller
* must either hold SCL_STATE as writer, or must be the sync thread
* (which holds SCL_STATE as reader). There's only one sync thread,
* so this is sufficient to ensure mutual exclusion.
*/
ASSERT(spa_config_held(spa, SCL_STATE, RW_WRITER) ||
(dsl_pool_sync_context(spa_get_dsl(spa)) &&
spa_config_held(spa, SCL_STATE, RW_READER)));
if (!list_link_active(&vd->vdev_state_dirty_node) &&
vdev_is_concrete(vd))
list_insert_head(&spa->spa_state_dirty_list, vd);
}
void
vdev_state_clean(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_config_held(spa, SCL_STATE, RW_WRITER) ||
(dsl_pool_sync_context(spa_get_dsl(spa)) &&
spa_config_held(spa, SCL_STATE, RW_READER)));
ASSERT(list_link_active(&vd->vdev_state_dirty_node));
list_remove(&spa->spa_state_dirty_list, vd);
}
/*
* Propagate vdev state up from children to parent.
*/
void
vdev_propagate_state(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
vdev_t *rvd = spa->spa_root_vdev;
int degraded = 0, faulted = 0;
int corrupted = 0;
vdev_t *child;
if (vd->vdev_children > 0) {
for (int c = 0; c < vd->vdev_children; c++) {
child = vd->vdev_child[c];
/*
* Don't factor holes or indirect vdevs into the
* decision.
*/
if (!vdev_is_concrete(child))
continue;
if (!vdev_readable(child) ||
(!vdev_writeable(child) && spa_writeable(spa))) {
/*
* Root special: if there is a top-level log
* device, treat the root vdev as if it were
* degraded.
*/
if (child->vdev_islog && vd == rvd)
degraded++;
else
faulted++;
} else if (child->vdev_state <= VDEV_STATE_DEGRADED) {
degraded++;
}
if (child->vdev_stat.vs_aux == VDEV_AUX_CORRUPT_DATA)
corrupted++;
}
vd->vdev_ops->vdev_op_state_change(vd, faulted, degraded);
/*
* Root special: if there is a top-level vdev that cannot be
* opened due to corrupted metadata, then propagate the root
* vdev's aux state as 'corrupt' rather than 'insufficient
* replicas'.
*/
if (corrupted && vd == rvd &&
rvd->vdev_state == VDEV_STATE_CANT_OPEN)
vdev_set_state(rvd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
}
if (vd->vdev_parent)
vdev_propagate_state(vd->vdev_parent);
}
/*
* Set a vdev's state. If this is during an open, we don't update the parent
* state, because we're in the process of opening children depth-first.
* Otherwise, we propagate the change to the parent.
*
* If this routine places a device in a faulted state, an appropriate ereport is
* generated.
*/
void
vdev_set_state(vdev_t *vd, boolean_t isopen, vdev_state_t state, vdev_aux_t aux)
{
uint64_t save_state;
spa_t *spa = vd->vdev_spa;
if (state == vd->vdev_state) {
/*
* Since vdev_offline() code path is already in an offline
* state we can miss a statechange event to OFFLINE. Check
* the previous state to catch this condition.
*/
if (vd->vdev_ops->vdev_op_leaf &&
(state == VDEV_STATE_OFFLINE) &&
(vd->vdev_prevstate >= VDEV_STATE_FAULTED)) {
/* post an offline state change */
zfs_post_state_change(spa, vd, vd->vdev_prevstate);
}
vd->vdev_stat.vs_aux = aux;
return;
}
save_state = vd->vdev_state;
vd->vdev_state = state;
vd->vdev_stat.vs_aux = aux;
/*
* If we are setting the vdev state to anything but an open state, then
* always close the underlying device unless the device has requested
* a delayed close (i.e. we're about to remove or fault the device).
* Otherwise, we keep accessible but invalid devices open forever.
* We don't call vdev_close() itself, because that implies some extra
* checks (offline, etc) that we don't want here. This is limited to
* leaf devices, because otherwise closing the device will affect other
* children.
*/
if (!vd->vdev_delayed_close && vdev_is_dead(vd) &&
vd->vdev_ops->vdev_op_leaf)
vd->vdev_ops->vdev_op_close(vd);
if (vd->vdev_removed &&
state == VDEV_STATE_CANT_OPEN &&
(aux == VDEV_AUX_OPEN_FAILED || vd->vdev_checkremove)) {
/*
* If the previous state is set to VDEV_STATE_REMOVED, then this
* device was previously marked removed and someone attempted to
* reopen it. If this failed due to a nonexistent device, then
* keep the device in the REMOVED state. We also let this be if
* it is one of our special test online cases, which is only
* attempting to online the device and shouldn't generate an FMA
* fault.
*/
vd->vdev_state = VDEV_STATE_REMOVED;
vd->vdev_stat.vs_aux = VDEV_AUX_NONE;
} else if (state == VDEV_STATE_REMOVED) {
vd->vdev_removed = B_TRUE;
} else if (state == VDEV_STATE_CANT_OPEN) {
/*
* If we fail to open a vdev during an import or recovery, we
* mark it as "not available", which signifies that it was
* never there to begin with. Failure to open such a device
* is not considered an error.
*/
if ((spa_load_state(spa) == SPA_LOAD_IMPORT ||
spa_load_state(spa) == SPA_LOAD_RECOVER) &&
vd->vdev_ops->vdev_op_leaf)
vd->vdev_not_present = 1;
/*
* Post the appropriate ereport. If the 'prevstate' field is
* set to something other than VDEV_STATE_UNKNOWN, it indicates
* that this is part of a vdev_reopen(). In this case, we don't
* want to post the ereport if the device was already in the
* CANT_OPEN state beforehand.
*
* If the 'checkremove' flag is set, then this is an attempt to
* online the device in response to an insertion event. If we
* hit this case, then we have detected an insertion event for a
* faulted or offline device that wasn't in the removed state.
* In this scenario, we don't post an ereport because we are
* about to replace the device, or attempt an online with
* vdev_forcefault, which will generate the fault for us.
*/
if ((vd->vdev_prevstate != state || vd->vdev_forcefault) &&
!vd->vdev_not_present && !vd->vdev_checkremove &&
vd != spa->spa_root_vdev) {
const char *class;
switch (aux) {
case VDEV_AUX_OPEN_FAILED:
class = FM_EREPORT_ZFS_DEVICE_OPEN_FAILED;
break;
case VDEV_AUX_CORRUPT_DATA:
class = FM_EREPORT_ZFS_DEVICE_CORRUPT_DATA;
break;
case VDEV_AUX_NO_REPLICAS:
class = FM_EREPORT_ZFS_DEVICE_NO_REPLICAS;
break;
case VDEV_AUX_BAD_GUID_SUM:
class = FM_EREPORT_ZFS_DEVICE_BAD_GUID_SUM;
break;
case VDEV_AUX_TOO_SMALL:
class = FM_EREPORT_ZFS_DEVICE_TOO_SMALL;
break;
case VDEV_AUX_BAD_LABEL:
class = FM_EREPORT_ZFS_DEVICE_BAD_LABEL;
break;
case VDEV_AUX_BAD_ASHIFT:
class = FM_EREPORT_ZFS_DEVICE_BAD_ASHIFT;
break;
default:
class = FM_EREPORT_ZFS_DEVICE_UNKNOWN;
}
(void) zfs_ereport_post(class, spa, vd, NULL, NULL,
save_state);
}
/* Erase any notion of persistent removed state */
vd->vdev_removed = B_FALSE;
} else {
vd->vdev_removed = B_FALSE;
}
/*
* Notify ZED of any significant state-change on a leaf vdev.
*
*/
if (vd->vdev_ops->vdev_op_leaf) {
/* preserve original state from a vdev_reopen() */
if ((vd->vdev_prevstate != VDEV_STATE_UNKNOWN) &&
(vd->vdev_prevstate != vd->vdev_state) &&
(save_state <= VDEV_STATE_CLOSED))
save_state = vd->vdev_prevstate;
/* filter out state change due to initial vdev_open */
if (save_state > VDEV_STATE_CLOSED)
zfs_post_state_change(spa, vd, save_state);
}
if (!isopen && vd->vdev_parent)
vdev_propagate_state(vd->vdev_parent);
}
boolean_t
vdev_children_are_offline(vdev_t *vd)
{
ASSERT(!vd->vdev_ops->vdev_op_leaf);
for (uint64_t i = 0; i < vd->vdev_children; i++) {
if (vd->vdev_child[i]->vdev_state != VDEV_STATE_OFFLINE)
return (B_FALSE);
}
return (B_TRUE);
}
/*
* Check the vdev configuration to ensure that it's capable of supporting
* a root pool. We do not support partial configuration.
*/
boolean_t
vdev_is_bootable(vdev_t *vd)
{
if (!vd->vdev_ops->vdev_op_leaf) {
const char *vdev_type = vd->vdev_ops->vdev_op_type;
if (strcmp(vdev_type, VDEV_TYPE_MISSING) == 0)
return (B_FALSE);
}
for (int c = 0; c < vd->vdev_children; c++) {
if (!vdev_is_bootable(vd->vdev_child[c]))
return (B_FALSE);
}
return (B_TRUE);
}
boolean_t
vdev_is_concrete(vdev_t *vd)
{
vdev_ops_t *ops = vd->vdev_ops;
if (ops == &vdev_indirect_ops || ops == &vdev_hole_ops ||
ops == &vdev_missing_ops || ops == &vdev_root_ops) {
return (B_FALSE);
} else {
return (B_TRUE);
}
}
/*
* Determine if a log device has valid content. If the vdev was
* removed or faulted in the MOS config then we know that
* the content on the log device has already been written to the pool.
*/
boolean_t
vdev_log_state_valid(vdev_t *vd)
{
if (vd->vdev_ops->vdev_op_leaf && !vd->vdev_faulted &&
!vd->vdev_removed)
return (B_TRUE);
for (int c = 0; c < vd->vdev_children; c++)
if (vdev_log_state_valid(vd->vdev_child[c]))
return (B_TRUE);
return (B_FALSE);
}
/*
* Expand a vdev if possible.
*/
void
vdev_expand(vdev_t *vd, uint64_t txg)
{
ASSERT(vd->vdev_top == vd);
ASSERT(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL);
ASSERT(vdev_is_concrete(vd));
vdev_set_deflate_ratio(vd);
- if ((vd->vdev_asize >> vd->vdev_ms_shift) > vd->vdev_ms_count &&
+ if ((vd->vdev_spa->spa_raidz_expand == NULL ||
+ vd->vdev_spa->spa_raidz_expand->vre_vdev_id != vd->vdev_id) &&
+ (vd->vdev_asize >> vd->vdev_ms_shift) > vd->vdev_ms_count &&
vdev_is_concrete(vd)) {
vdev_metaslab_group_create(vd);
VERIFY(vdev_metaslab_init(vd, txg) == 0);
vdev_config_dirty(vd);
}
}
/*
* Split a vdev.
*/
void
vdev_split(vdev_t *vd)
{
vdev_t *cvd, *pvd = vd->vdev_parent;
VERIFY3U(pvd->vdev_children, >, 1);
vdev_remove_child(pvd, vd);
vdev_compact_children(pvd);
ASSERT3P(pvd->vdev_child, !=, NULL);
cvd = pvd->vdev_child[0];
if (pvd->vdev_children == 1) {
vdev_remove_parent(cvd);
cvd->vdev_splitting = B_TRUE;
}
vdev_propagate_state(cvd);
}
void
vdev_deadman(vdev_t *vd, const char *tag)
{
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
vdev_deadman(cvd, tag);
}
if (vd->vdev_ops->vdev_op_leaf) {
vdev_queue_t *vq = &vd->vdev_queue;
mutex_enter(&vq->vq_lock);
if (vq->vq_active > 0) {
spa_t *spa = vd->vdev_spa;
zio_t *fio;
uint64_t delta;
zfs_dbgmsg("slow vdev: %s has %u active IOs",
vd->vdev_path, vq->vq_active);
/*
* Look at the head of all the pending queues,
* if any I/O has been outstanding for longer than
* the spa_deadman_synctime invoke the deadman logic.
*/
fio = list_head(&vq->vq_active_list);
delta = gethrtime() - fio->io_timestamp;
if (delta > spa_deadman_synctime(spa))
zio_deadman(fio, tag);
}
mutex_exit(&vq->vq_lock);
}
}
void
vdev_defer_resilver(vdev_t *vd)
{
ASSERT(vd->vdev_ops->vdev_op_leaf);
vd->vdev_resilver_deferred = B_TRUE;
vd->vdev_spa->spa_resilver_deferred = B_TRUE;
}
/*
* Clears the resilver deferred flag on all leaf devs under vd. Returns
* B_TRUE if we have devices that need to be resilvered and are available to
* accept resilver I/Os.
*/
boolean_t
vdev_clear_resilver_deferred(vdev_t *vd, dmu_tx_t *tx)
{
boolean_t resilver_needed = B_FALSE;
spa_t *spa = vd->vdev_spa;
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
resilver_needed |= vdev_clear_resilver_deferred(cvd, tx);
}
if (vd == spa->spa_root_vdev &&
spa_feature_is_active(spa, SPA_FEATURE_RESILVER_DEFER)) {
spa_feature_decr(spa, SPA_FEATURE_RESILVER_DEFER, tx);
vdev_config_dirty(vd);
spa->spa_resilver_deferred = B_FALSE;
return (resilver_needed);
}
if (!vdev_is_concrete(vd) || vd->vdev_aux ||
!vd->vdev_ops->vdev_op_leaf)
return (resilver_needed);
vd->vdev_resilver_deferred = B_FALSE;
return (!vdev_is_dead(vd) && !vd->vdev_offline &&
vdev_resilver_needed(vd, NULL, NULL));
}
boolean_t
vdev_xlate_is_empty(range_seg64_t *rs)
{
return (rs->rs_start == rs->rs_end);
}
/*
* Translate a logical range to the first contiguous physical range for the
* specified vdev_t. This function is initially called with a leaf vdev and
* will walk each parent vdev until it reaches a top-level vdev. Once the
* top-level is reached the physical range is initialized and the recursive
* function begins to unwind. As it unwinds it calls the parent's vdev
* specific translation function to do the real conversion.
*/
void
vdev_xlate(vdev_t *vd, const range_seg64_t *logical_rs,
range_seg64_t *physical_rs, range_seg64_t *remain_rs)
{
/*
* Walk up the vdev tree
*/
if (vd != vd->vdev_top) {
vdev_xlate(vd->vdev_parent, logical_rs, physical_rs,
remain_rs);
} else {
/*
* We've reached the top-level vdev, initialize the physical
* range to the logical range and set an empty remaining
* range then start to unwind.
*/
physical_rs->rs_start = logical_rs->rs_start;
physical_rs->rs_end = logical_rs->rs_end;
remain_rs->rs_start = logical_rs->rs_start;
remain_rs->rs_end = logical_rs->rs_start;
return;
}
vdev_t *pvd = vd->vdev_parent;
ASSERT3P(pvd, !=, NULL);
ASSERT3P(pvd->vdev_ops->vdev_op_xlate, !=, NULL);
/*
* As this recursive function unwinds, translate the logical
* range into its physical and any remaining components by calling
* the vdev specific translate function.
*/
range_seg64_t intermediate = { 0 };
pvd->vdev_ops->vdev_op_xlate(vd, physical_rs, &intermediate, remain_rs);
physical_rs->rs_start = intermediate.rs_start;
physical_rs->rs_end = intermediate.rs_end;
}
void
vdev_xlate_walk(vdev_t *vd, const range_seg64_t *logical_rs,
vdev_xlate_func_t *func, void *arg)
{
range_seg64_t iter_rs = *logical_rs;
range_seg64_t physical_rs;
range_seg64_t remain_rs;
while (!vdev_xlate_is_empty(&iter_rs)) {
vdev_xlate(vd, &iter_rs, &physical_rs, &remain_rs);
/*
* With raidz and dRAID, it's possible that the logical range
* does not live on this leaf vdev. Only when there is a non-
* zero physical size call the provided function.
*/
if (!vdev_xlate_is_empty(&physical_rs))
func(arg, &physical_rs);
iter_rs = remain_rs;
}
}
static char *
vdev_name(vdev_t *vd, char *buf, int buflen)
{
if (vd->vdev_path == NULL) {
if (strcmp(vd->vdev_ops->vdev_op_type, "root") == 0) {
strlcpy(buf, vd->vdev_spa->spa_name, buflen);
} else if (!vd->vdev_ops->vdev_op_leaf) {
snprintf(buf, buflen, "%s-%llu",
vd->vdev_ops->vdev_op_type,
(u_longlong_t)vd->vdev_id);
}
} else {
strlcpy(buf, vd->vdev_path, buflen);
}
return (buf);
}
/*
* Look at the vdev tree and determine whether any devices are currently being
* replaced.
*/
boolean_t
vdev_replace_in_progress(vdev_t *vdev)
{
ASSERT(spa_config_held(vdev->vdev_spa, SCL_ALL, RW_READER) != 0);
if (vdev->vdev_ops == &vdev_replacing_ops)
return (B_TRUE);
/*
* A 'spare' vdev indicates that we have a replace in progress, unless
* it has exactly two children, and the second, the hot spare, has
* finished being resilvered.
*/
if (vdev->vdev_ops == &vdev_spare_ops && (vdev->vdev_children > 2 ||
!vdev_dtl_empty(vdev->vdev_child[1], DTL_MISSING)))
return (B_TRUE);
for (int i = 0; i < vdev->vdev_children; i++) {
if (vdev_replace_in_progress(vdev->vdev_child[i]))
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Add a (source=src, propname=propval) list to an nvlist.
*/
static void
vdev_prop_add_list(nvlist_t *nvl, const char *propname, const char *strval,
uint64_t intval, zprop_source_t src)
{
nvlist_t *propval;
propval = fnvlist_alloc();
fnvlist_add_uint64(propval, ZPROP_SOURCE, src);
if (strval != NULL)
fnvlist_add_string(propval, ZPROP_VALUE, strval);
else
fnvlist_add_uint64(propval, ZPROP_VALUE, intval);
fnvlist_add_nvlist(nvl, propname, propval);
nvlist_free(propval);
}
static void
vdev_props_set_sync(void *arg, dmu_tx_t *tx)
{
vdev_t *vd;
nvlist_t *nvp = arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
objset_t *mos = spa->spa_meta_objset;
nvpair_t *elem = NULL;
uint64_t vdev_guid;
uint64_t objid;
nvlist_t *nvprops;
vdev_guid = fnvlist_lookup_uint64(nvp, ZPOOL_VDEV_PROPS_SET_VDEV);
nvprops = fnvlist_lookup_nvlist(nvp, ZPOOL_VDEV_PROPS_SET_PROPS);
vd = spa_lookup_by_guid(spa, vdev_guid, B_TRUE);
/* this vdev could get removed while waiting for this sync task */
if (vd == NULL)
return;
/*
* Set vdev property values in the vdev props mos object.
*/
if (vd->vdev_root_zap != 0) {
objid = vd->vdev_root_zap;
} else if (vd->vdev_top_zap != 0) {
objid = vd->vdev_top_zap;
} else if (vd->vdev_leaf_zap != 0) {
objid = vd->vdev_leaf_zap;
} else {
panic("unexpected vdev type");
}
mutex_enter(&spa->spa_props_lock);
while ((elem = nvlist_next_nvpair(nvprops, elem)) != NULL) {
uint64_t intval;
const char *strval;
vdev_prop_t prop;
const char *propname = nvpair_name(elem);
zprop_type_t proptype;
switch (prop = vdev_name_to_prop(propname)) {
case VDEV_PROP_USERPROP:
if (vdev_prop_user(propname)) {
strval = fnvpair_value_string(elem);
if (strlen(strval) == 0) {
/* remove the property if value == "" */
(void) zap_remove(mos, objid, propname,
tx);
} else {
VERIFY0(zap_update(mos, objid, propname,
1, strlen(strval) + 1, strval, tx));
}
spa_history_log_internal(spa, "vdev set", tx,
"vdev_guid=%llu: %s=%s",
(u_longlong_t)vdev_guid, nvpair_name(elem),
strval);
}
break;
default:
/* normalize the property name */
propname = vdev_prop_to_name(prop);
proptype = vdev_prop_get_type(prop);
if (nvpair_type(elem) == DATA_TYPE_STRING) {
ASSERT(proptype == PROP_TYPE_STRING);
strval = fnvpair_value_string(elem);
VERIFY0(zap_update(mos, objid, propname,
1, strlen(strval) + 1, strval, tx));
spa_history_log_internal(spa, "vdev set", tx,
"vdev_guid=%llu: %s=%s",
(u_longlong_t)vdev_guid, nvpair_name(elem),
strval);
} else if (nvpair_type(elem) == DATA_TYPE_UINT64) {
intval = fnvpair_value_uint64(elem);
if (proptype == PROP_TYPE_INDEX) {
const char *unused;
VERIFY0(vdev_prop_index_to_string(
prop, intval, &unused));
}
VERIFY0(zap_update(mos, objid, propname,
sizeof (uint64_t), 1, &intval, tx));
spa_history_log_internal(spa, "vdev set", tx,
"vdev_guid=%llu: %s=%lld",
(u_longlong_t)vdev_guid,
nvpair_name(elem), (longlong_t)intval);
} else {
panic("invalid vdev property type %u",
nvpair_type(elem));
}
}
}
mutex_exit(&spa->spa_props_lock);
}
int
vdev_prop_set(vdev_t *vd, nvlist_t *innvl, nvlist_t *outnvl)
{
spa_t *spa = vd->vdev_spa;
nvpair_t *elem = NULL;
uint64_t vdev_guid;
nvlist_t *nvprops;
int error = 0;
ASSERT(vd != NULL);
/* Check that vdev has a zap we can use */
if (vd->vdev_root_zap == 0 &&
vd->vdev_top_zap == 0 &&
vd->vdev_leaf_zap == 0)
return (SET_ERROR(EINVAL));
if (nvlist_lookup_uint64(innvl, ZPOOL_VDEV_PROPS_SET_VDEV,
&vdev_guid) != 0)
return (SET_ERROR(EINVAL));
if (nvlist_lookup_nvlist(innvl, ZPOOL_VDEV_PROPS_SET_PROPS,
&nvprops) != 0)
return (SET_ERROR(EINVAL));
if ((vd = spa_lookup_by_guid(spa, vdev_guid, B_TRUE)) == NULL)
return (SET_ERROR(EINVAL));
while ((elem = nvlist_next_nvpair(nvprops, elem)) != NULL) {
const char *propname = nvpair_name(elem);
vdev_prop_t prop = vdev_name_to_prop(propname);
uint64_t intval = 0;
const char *strval = NULL;
if (prop == VDEV_PROP_USERPROP && !vdev_prop_user(propname)) {
error = EINVAL;
goto end;
}
if (vdev_prop_readonly(prop)) {
error = EROFS;
goto end;
}
/* Special Processing */
switch (prop) {
case VDEV_PROP_PATH:
if (vd->vdev_path == NULL) {
error = EROFS;
break;
}
if (nvpair_value_string(elem, &strval) != 0) {
error = EINVAL;
break;
}
/* New path must start with /dev/ */
if (strncmp(strval, "/dev/", 5)) {
error = EINVAL;
break;
}
error = spa_vdev_setpath(spa, vdev_guid, strval);
break;
case VDEV_PROP_ALLOCATING:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
if (intval != vd->vdev_noalloc)
break;
if (intval == 0)
error = spa_vdev_noalloc(spa, vdev_guid);
else
error = spa_vdev_alloc(spa, vdev_guid);
break;
case VDEV_PROP_FAILFAST:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
vd->vdev_failfast = intval & 1;
break;
case VDEV_PROP_CHECKSUM_N:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
vd->vdev_checksum_n = intval;
break;
case VDEV_PROP_CHECKSUM_T:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
vd->vdev_checksum_t = intval;
break;
case VDEV_PROP_IO_N:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
vd->vdev_io_n = intval;
break;
case VDEV_PROP_IO_T:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
vd->vdev_io_t = intval;
break;
default:
/* Most processing is done in vdev_props_set_sync */
break;
}
end:
if (error != 0) {
intval = error;
vdev_prop_add_list(outnvl, propname, strval, intval, 0);
return (error);
}
}
return (dsl_sync_task(spa->spa_name, NULL, vdev_props_set_sync,
innvl, 6, ZFS_SPACE_CHECK_EXTRA_RESERVED));
}
int
vdev_prop_get(vdev_t *vd, nvlist_t *innvl, nvlist_t *outnvl)
{
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa->spa_meta_objset;
int err = 0;
uint64_t objid;
uint64_t vdev_guid;
nvpair_t *elem = NULL;
nvlist_t *nvprops = NULL;
uint64_t intval = 0;
char *strval = NULL;
const char *propname = NULL;
vdev_prop_t prop;
ASSERT(vd != NULL);
ASSERT(mos != NULL);
if (nvlist_lookup_uint64(innvl, ZPOOL_VDEV_PROPS_GET_VDEV,
&vdev_guid) != 0)
return (SET_ERROR(EINVAL));
nvlist_lookup_nvlist(innvl, ZPOOL_VDEV_PROPS_GET_PROPS, &nvprops);
if (vd->vdev_root_zap != 0) {
objid = vd->vdev_root_zap;
} else if (vd->vdev_top_zap != 0) {
objid = vd->vdev_top_zap;
} else if (vd->vdev_leaf_zap != 0) {
objid = vd->vdev_leaf_zap;
} else {
return (SET_ERROR(EINVAL));
}
ASSERT(objid != 0);
mutex_enter(&spa->spa_props_lock);
if (nvprops != NULL) {
char namebuf[64] = { 0 };
while ((elem = nvlist_next_nvpair(nvprops, elem)) != NULL) {
intval = 0;
strval = NULL;
propname = nvpair_name(elem);
prop = vdev_name_to_prop(propname);
zprop_source_t src = ZPROP_SRC_DEFAULT;
uint64_t integer_size, num_integers;
switch (prop) {
/* Special Read-only Properties */
case VDEV_PROP_NAME:
strval = vdev_name(vd, namebuf,
sizeof (namebuf));
if (strval == NULL)
continue;
vdev_prop_add_list(outnvl, propname, strval, 0,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_CAPACITY:
/* percent used */
intval = (vd->vdev_stat.vs_dspace == 0) ? 0 :
(vd->vdev_stat.vs_alloc * 100 /
vd->vdev_stat.vs_dspace);
vdev_prop_add_list(outnvl, propname, NULL,
intval, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_STATE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_state, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_GUID:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_guid, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_ASIZE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_asize, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_PSIZE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_psize, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_ASHIFT:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_ashift, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_SIZE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_dspace, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_FREE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_dspace -
vd->vdev_stat.vs_alloc, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_ALLOCATED:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_alloc, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_EXPANDSZ:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_esize, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_FRAGMENTATION:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_fragmentation,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_PARITY:
vdev_prop_add_list(outnvl, propname, NULL,
vdev_get_nparity(vd), ZPROP_SRC_NONE);
continue;
case VDEV_PROP_PATH:
if (vd->vdev_path == NULL)
continue;
vdev_prop_add_list(outnvl, propname,
vd->vdev_path, 0, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_DEVID:
if (vd->vdev_devid == NULL)
continue;
vdev_prop_add_list(outnvl, propname,
vd->vdev_devid, 0, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_PHYS_PATH:
if (vd->vdev_physpath == NULL)
continue;
vdev_prop_add_list(outnvl, propname,
vd->vdev_physpath, 0, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_ENC_PATH:
if (vd->vdev_enc_sysfs_path == NULL)
continue;
vdev_prop_add_list(outnvl, propname,
vd->vdev_enc_sysfs_path, 0, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_FRU:
if (vd->vdev_fru == NULL)
continue;
vdev_prop_add_list(outnvl, propname,
vd->vdev_fru, 0, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_PARENT:
if (vd->vdev_parent != NULL) {
strval = vdev_name(vd->vdev_parent,
namebuf, sizeof (namebuf));
vdev_prop_add_list(outnvl, propname,
strval, 0, ZPROP_SRC_NONE);
}
continue;
case VDEV_PROP_CHILDREN:
if (vd->vdev_children > 0)
strval = kmem_zalloc(ZAP_MAXVALUELEN,
KM_SLEEP);
for (uint64_t i = 0; i < vd->vdev_children;
i++) {
const char *vname;
vname = vdev_name(vd->vdev_child[i],
namebuf, sizeof (namebuf));
if (vname == NULL)
vname = "(unknown)";
if (strlen(strval) > 0)
strlcat(strval, ",",
ZAP_MAXVALUELEN);
strlcat(strval, vname, ZAP_MAXVALUELEN);
}
if (strval != NULL) {
vdev_prop_add_list(outnvl, propname,
strval, 0, ZPROP_SRC_NONE);
kmem_free(strval, ZAP_MAXVALUELEN);
}
continue;
case VDEV_PROP_NUMCHILDREN:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_children, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_READ_ERRORS:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_read_errors,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_WRITE_ERRORS:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_write_errors,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_CHECKSUM_ERRORS:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_checksum_errors,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_INITIALIZE_ERRORS:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_initialize_errors,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_NULL:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_NULL],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_READ:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_READ],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_WRITE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_WRITE],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_FREE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_FREE],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_CLAIM:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_CLAIM],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_TRIM:
/*
* TRIM ops and bytes are reported to user
* space as ZIO_TYPE_IOCTL. This is done to
* preserve the vdev_stat_t structure layout
* for user space.
*/
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_IOCTL],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_NULL:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_NULL],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_READ:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_READ],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_WRITE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_WRITE],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_FREE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_FREE],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_CLAIM:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_CLAIM],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_TRIM:
/*
* TRIM ops and bytes are reported to user
* space as ZIO_TYPE_IOCTL. This is done to
* preserve the vdev_stat_t structure layout
* for user space.
*/
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_IOCTL],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_REMOVING:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_removing, ZPROP_SRC_NONE);
continue;
+ case VDEV_PROP_RAIDZ_EXPANDING:
+ /* Only expose this for raidz */
+ if (vd->vdev_ops == &vdev_raidz_ops) {
+ vdev_prop_add_list(outnvl, propname,
+ NULL, vd->vdev_rz_expanding,
+ ZPROP_SRC_NONE);
+ }
+ continue;
/* Numeric Properites */
case VDEV_PROP_ALLOCATING:
/* Leaf vdevs cannot have this property */
if (vd->vdev_mg == NULL &&
vd->vdev_top != NULL) {
src = ZPROP_SRC_NONE;
intval = ZPROP_BOOLEAN_NA;
} else {
err = vdev_prop_get_int(vd, prop,
&intval);
if (err && err != ENOENT)
break;
if (intval ==
vdev_prop_default_numeric(prop))
src = ZPROP_SRC_DEFAULT;
else
src = ZPROP_SRC_LOCAL;
}
vdev_prop_add_list(outnvl, propname, NULL,
intval, src);
break;
case VDEV_PROP_FAILFAST:
src = ZPROP_SRC_LOCAL;
strval = NULL;
err = zap_lookup(mos, objid, nvpair_name(elem),
sizeof (uint64_t), 1, &intval);
if (err == ENOENT) {
intval = vdev_prop_default_numeric(
prop);
err = 0;
} else if (err) {
break;
}
if (intval == vdev_prop_default_numeric(prop))
src = ZPROP_SRC_DEFAULT;
vdev_prop_add_list(outnvl, propname, strval,
intval, src);
break;
case VDEV_PROP_CHECKSUM_N:
case VDEV_PROP_CHECKSUM_T:
case VDEV_PROP_IO_N:
case VDEV_PROP_IO_T:
err = vdev_prop_get_int(vd, prop, &intval);
if (err && err != ENOENT)
break;
if (intval == vdev_prop_default_numeric(prop))
src = ZPROP_SRC_DEFAULT;
else
src = ZPROP_SRC_LOCAL;
vdev_prop_add_list(outnvl, propname, NULL,
intval, src);
break;
/* Text Properties */
case VDEV_PROP_COMMENT:
/* Exists in the ZAP below */
/* FALLTHRU */
case VDEV_PROP_USERPROP:
/* User Properites */
src = ZPROP_SRC_LOCAL;
err = zap_length(mos, objid, nvpair_name(elem),
&integer_size, &num_integers);
if (err)
break;
switch (integer_size) {
case 8:
/* User properties cannot be integers */
err = EINVAL;
break;
case 1:
/* string property */
strval = kmem_alloc(num_integers,
KM_SLEEP);
err = zap_lookup(mos, objid,
nvpair_name(elem), 1,
num_integers, strval);
if (err) {
kmem_free(strval,
num_integers);
break;
}
vdev_prop_add_list(outnvl, propname,
strval, 0, src);
kmem_free(strval, num_integers);
break;
}
break;
default:
err = ENOENT;
break;
}
if (err)
break;
}
} else {
/*
* Get all properties from the MOS vdev property object.
*/
zap_cursor_t zc;
zap_attribute_t za;
for (zap_cursor_init(&zc, mos, objid);
(err = zap_cursor_retrieve(&zc, &za)) == 0;
zap_cursor_advance(&zc)) {
intval = 0;
strval = NULL;
zprop_source_t src = ZPROP_SRC_DEFAULT;
propname = za.za_name;
switch (za.za_integer_length) {
case 8:
/* We do not allow integer user properties */
/* This is likely an internal value */
break;
case 1:
/* string property */
strval = kmem_alloc(za.za_num_integers,
KM_SLEEP);
err = zap_lookup(mos, objid, za.za_name, 1,
za.za_num_integers, strval);
if (err) {
kmem_free(strval, za.za_num_integers);
break;
}
vdev_prop_add_list(outnvl, propname, strval, 0,
src);
kmem_free(strval, za.za_num_integers);
break;
default:
break;
}
}
zap_cursor_fini(&zc);
}
mutex_exit(&spa->spa_props_lock);
if (err && err != ENOENT) {
return (err);
}
return (0);
}
EXPORT_SYMBOL(vdev_fault);
EXPORT_SYMBOL(vdev_degrade);
EXPORT_SYMBOL(vdev_online);
EXPORT_SYMBOL(vdev_offline);
EXPORT_SYMBOL(vdev_clear);
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, default_ms_count, UINT, ZMOD_RW,
"Target number of metaslabs per top-level vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, default_ms_shift, UINT, ZMOD_RW,
"Default lower limit for metaslab size");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, max_ms_shift, UINT, ZMOD_RW,
"Default upper limit for metaslab size");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, min_ms_count, UINT, ZMOD_RW,
"Minimum number of metaslabs per top-level vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, ms_count_limit, UINT, ZMOD_RW,
"Practical upper limit of total metaslabs per top-level vdev");
ZFS_MODULE_PARAM(zfs, zfs_, slow_io_events_per_second, UINT, ZMOD_RW,
"Rate limit slow IO (delay) events to this many per second");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs, zfs_, checksum_events_per_second, UINT, ZMOD_RW,
"Rate limit checksum events to this many checksum errors per second "
"(do not set below ZED threshold).");
/* END CSTYLED */
ZFS_MODULE_PARAM(zfs, zfs_, scan_ignore_errors, INT, ZMOD_RW,
"Ignore errors during resilver/scrub");
ZFS_MODULE_PARAM(zfs_vdev, vdev_, validate_skip, INT, ZMOD_RW,
"Bypass vdev_validate()");
ZFS_MODULE_PARAM(zfs, zfs_, nocacheflush, INT, ZMOD_RW,
"Disable cache flushes");
ZFS_MODULE_PARAM(zfs, zfs_, embedded_slog_min_ms, UINT, ZMOD_RW,
"Minimum number of metaslabs required to dedicate one for log blocks");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM_CALL(zfs_vdev, zfs_vdev_, min_auto_ashift,
param_set_min_auto_ashift, param_get_uint, ZMOD_RW,
"Minimum ashift used when creating new top-level vdevs");
ZFS_MODULE_PARAM_CALL(zfs_vdev, zfs_vdev_, max_auto_ashift,
param_set_max_auto_ashift, param_get_uint, ZMOD_RW,
"Maximum ashift used when optimizing for logical -> physical sector "
"size on new top-level vdevs");
/* END CSTYLED */
diff --git a/sys/contrib/openzfs/module/zfs/vdev_draid.c b/sys/contrib/openzfs/module/zfs/vdev_draid.c
index 307e2353d020..ec961255fd64 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_draid.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_draid.c
@@ -1,2839 +1,2825 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2018 Intel Corporation.
* Copyright (c) 2020 by Lawrence Livermore National Security, LLC.
*/
#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_draid.h>
#include <sys/vdev_raidz.h>
#include <sys/vdev_rebuild.h>
#include <sys/abd.h>
#include <sys/zio.h>
#include <sys/nvpair.h>
#include <sys/zio_checksum.h>
#include <sys/fs/zfs.h>
#include <sys/fm/fs/zfs.h>
#include <zfs_fletcher.h>
#ifdef ZFS_DEBUG
#include <sys/vdev.h> /* For vdev_xlate() in vdev_draid_io_verify() */
#endif
/*
* dRAID is a distributed spare implementation for ZFS. A dRAID vdev is
* comprised of multiple raidz redundancy groups which are spread over the
* dRAID children. To ensure an even distribution, and avoid hot spots, a
* permutation mapping is applied to the order of the dRAID children.
* This mixing effectively distributes the parity columns evenly over all
* of the disks in the dRAID.
*
* This is beneficial because it means when resilvering all of the disks
* can participate thereby increasing the available IOPs and bandwidth.
* Furthermore, by reserving a small fraction of each child's total capacity
* virtual distributed spare disks can be created. These spares similarly
* benefit from the performance gains of spanning all of the children. The
* consequence of which is that resilvering to a distributed spare can
* substantially reduce the time required to restore full parity to pool
* with a failed disks.
*
* === dRAID group layout ===
*
* First, let's define a "row" in the configuration to be a 16M chunk from
* each physical drive at the same offset. This is the minimum allowable
* size since it must be possible to store a full 16M block when there is
* only a single data column. Next, we define a "group" to be a set of
* sequential disks containing both the parity and data columns. We allow
* groups to span multiple rows in order to align any group size to any
* number of physical drives. Finally, a "slice" is comprised of the rows
* which contain the target number of groups. The permutation mappings
* are applied in a round robin fashion to each slice.
*
* Given D+P drives in a group (including parity drives) and C-S physical
* drives (not including the spare drives), we can distribute the groups
* across R rows without remainder by selecting the least common multiple
* of D+P and C-S as the number of groups; i.e. ngroups = LCM(D+P, C-S).
*
* In the example below, there are C=14 physical drives in the configuration
* with S=2 drives worth of spare capacity. Each group has a width of 9
* which includes D=8 data and P=1 parity drive. There are 4 groups and
* 3 rows per slice. Each group has a size of 144M (16M * 9) and a slice
* size is 576M (144M * 4). When allocating from a dRAID each group is
* filled before moving on to the next as show in slice0 below.
*
* data disks (8 data + 1 parity) spares (2)
* +===+===+===+===+===+===+===+===+===+===+===+===+===+===+
* ^ | 2 | 6 | 1 | 11| 4 | 0 | 7 | 10| 8 | 9 | 13| 5 | 12| 3 | device map 0
* | +===+===+===+===+===+===+===+===+===+===+===+===+===+===+
* | | group 0 | group 1..| |
* | +-----------------------------------+-----------+-------|
* | | 0 1 2 3 4 5 6 7 8 | 36 37 38| | r
* | | 9 10 11 12 13 14 15 16 17| 45 46 47| | o
* | | 18 19 20 21 22 23 24 25 26| 54 55 56| | w
* | 27 28 29 30 31 32 33 34 35| 63 64 65| | 0
* s +-----------------------+-----------------------+-------+
* l | ..group 1 | group 2.. | |
* i +-----------------------+-----------------------+-------+
* c | 39 40 41 42 43 44| 72 73 74 75 76 77| | r
* e | 48 49 50 51 52 53| 81 82 83 84 85 86| | o
* 0 | 57 58 59 60 61 62| 90 91 92 93 94 95| | w
* | 66 67 68 69 70 71| 99 100 101 102 103 104| | 1
* | +-----------+-----------+-----------------------+-------+
* | |..group 2 | group 3 | |
* | +-----------+-----------+-----------------------+-------+
* | | 78 79 80|108 109 110 111 112 113 114 115 116| | r
* | | 87 88 89|117 118 119 120 121 122 123 124 125| | o
* | | 96 97 98|126 127 128 129 130 131 132 133 134| | w
* v |105 106 107|135 136 137 138 139 140 141 142 143| | 2
* +===+===+===+===+===+===+===+===+===+===+===+===+===+===+
* | 9 | 11| 12| 2 | 4 | 1 | 3 | 0 | 10| 13| 8 | 5 | 6 | 7 | device map 1
* s +===+===+===+===+===+===+===+===+===+===+===+===+===+===+
* l | group 4 | group 5..| | row 3
* i +-----------------------+-----------+-----------+-------|
* c | ..group 5 | group 6.. | | row 4
* e +-----------+-----------+-----------------------+-------+
* 1 |..group 6 | group 7 | | row 5
* +===+===+===+===+===+===+===+===+===+===+===+===+===+===+
* | 3 | 5 | 10| 8 | 6 | 11| 12| 0 | 2 | 4 | 7 | 1 | 9 | 13| device map 2
* s +===+===+===+===+===+===+===+===+===+===+===+===+===+===+
* l | group 8 | group 9..| | row 6
* i +-----------------------------------------------+-------|
* c | ..group 9 | group 10.. | | row 7
* e +-----------------------+-----------------------+-------+
* 2 |..group 10 | group 11 | | row 8
* +-----------+-----------------------------------+-------+
*
* This layout has several advantages over requiring that each row contain
* a whole number of groups.
*
* 1. The group count is not a relevant parameter when defining a dRAID
* layout. Only the group width is needed, and *all* groups will have
* the desired size.
*
* 2. All possible group widths (<= physical disk count) can be supported.
*
* 3. The logic within vdev_draid.c is simplified when the group width is
* the same for all groups (although some of the logic around computing
* permutation numbers and drive offsets is more complicated).
*
* N.B. The following array describes all valid dRAID permutation maps.
* Each row is used to generate a permutation map for a different number
* of children from a unique seed. The seeds were generated and carefully
* evaluated by the 'draid' utility in order to provide balanced mappings.
* In addition to the seed a checksum of the in-memory mapping is stored
* for verification.
*
* The imbalance ratio of a given failure (e.g. 5 disks wide, child 3 failed,
* with a given permutation map) is the ratio of the amounts of I/O that will
* be sent to the least and most busy disks when resilvering. The average
* imbalance ratio (of a given number of disks and permutation map) is the
* average of the ratios of all possible single and double disk failures.
*
* In order to achieve a low imbalance ratio the number of permutations in
* the mapping must be significantly larger than the number of children.
* For dRAID the number of permutations has been limited to 512 to minimize
* the map size. This does result in a gradually increasing imbalance ratio
* as seen in the table below. Increasing the number of permutations for
* larger child counts would reduce the imbalance ratio. However, in practice
* when there are a large number of children each child is responsible for
* fewer total IOs so it's less of a concern.
*
* Note these values are hard coded and must never be changed. Existing
* pools depend on the same mapping always being generated in order to
* read and write from the correct locations. Any change would make
* existing pools completely inaccessible.
*/
static const draid_map_t draid_maps[VDEV_DRAID_MAX_MAPS] = {
{ 2, 256, 0x89ef3dabbcc7de37, 0x00000000433d433d }, /* 1.000 */
{ 3, 256, 0x89a57f3de98121b4, 0x00000000bcd8b7b5 }, /* 1.000 */
{ 4, 256, 0xc9ea9ec82340c885, 0x00000001819d7c69 }, /* 1.000 */
{ 5, 256, 0xf46733b7f4d47dfd, 0x00000002a1648d74 }, /* 1.010 */
{ 6, 256, 0x88c3c62d8585b362, 0x00000003d3b0c2c4 }, /* 1.031 */
{ 7, 256, 0x3a65d809b4d1b9d5, 0x000000055c4183ee }, /* 1.043 */
{ 8, 256, 0xe98930e3c5d2e90a, 0x00000006edfb0329 }, /* 1.059 */
{ 9, 256, 0x5a5430036b982ccb, 0x00000008ceaf6934 }, /* 1.056 */
{ 10, 256, 0x92bf389e9eadac74, 0x0000000b26668c09 }, /* 1.072 */
{ 11, 256, 0x74ccebf1dcf3ae80, 0x0000000dd691358c }, /* 1.083 */
{ 12, 256, 0x8847e41a1a9f5671, 0x00000010a0c63c8e }, /* 1.097 */
{ 13, 256, 0x7481b56debf0e637, 0x0000001424121fe4 }, /* 1.100 */
{ 14, 256, 0x559b8c44065f8967, 0x00000016ab2ff079 }, /* 1.121 */
{ 15, 256, 0x34c49545a2ee7f01, 0x0000001a6028efd6 }, /* 1.103 */
{ 16, 256, 0xb85f4fa81a7698f7, 0x0000001e95ff5e66 }, /* 1.111 */
{ 17, 256, 0x6353e47b7e47aba0, 0x00000021a81fa0fe }, /* 1.133 */
{ 18, 256, 0xaa549746b1cbb81c, 0x00000026f02494c9 }, /* 1.131 */
{ 19, 256, 0x892e343f2f31d690, 0x00000029eb392835 }, /* 1.130 */
{ 20, 256, 0x76914824db98cc3f, 0x0000003004f31a7c }, /* 1.141 */
{ 21, 256, 0x4b3cbabf9cfb1d0f, 0x00000036363a2408 }, /* 1.139 */
{ 22, 256, 0xf45c77abb4f035d4, 0x00000038dd0f3e84 }, /* 1.150 */
{ 23, 256, 0x5e18bd7f3fd4baf4, 0x0000003f0660391f }, /* 1.174 */
{ 24, 256, 0xa7b3a4d285d6503b, 0x000000443dfc9ff6 }, /* 1.168 */
{ 25, 256, 0x56ac7dd967521f5a, 0x0000004b03a87eb7 }, /* 1.180 */
{ 26, 256, 0x3a42dfda4eb880f7, 0x000000522c719bba }, /* 1.226 */
{ 27, 256, 0xd200d2fc6b54bf60, 0x0000005760b4fdf5 }, /* 1.228 */
{ 28, 256, 0xc52605bbd486c546, 0x0000005e00d8f74c }, /* 1.217 */
{ 29, 256, 0xc761779e63cd762f, 0x00000067be3cd85c }, /* 1.239 */
{ 30, 256, 0xca577b1e07f85ca5, 0x0000006f5517f3e4 }, /* 1.238 */
{ 31, 256, 0xfd50a593c518b3d4, 0x0000007370e7778f }, /* 1.273 */
{ 32, 512, 0xc6c87ba5b042650b, 0x000000f7eb08a156 }, /* 1.191 */
{ 33, 512, 0xc3880d0c9d458304, 0x0000010734b5d160 }, /* 1.199 */
{ 34, 512, 0xe920927e4d8b2c97, 0x00000118c1edbce0 }, /* 1.195 */
{ 35, 512, 0x8da7fcda87bde316, 0x0000012a3e9f9110 }, /* 1.201 */
{ 36, 512, 0xcf09937491514a29, 0x0000013bd6a24bef }, /* 1.194 */
{ 37, 512, 0x9b5abbf345cbd7cc, 0x0000014b9d90fac3 }, /* 1.237 */
{ 38, 512, 0x506312a44668d6a9, 0x0000015e1b5f6148 }, /* 1.242 */
{ 39, 512, 0x71659ede62b4755f, 0x00000173ef029bcd }, /* 1.231 */
{ 40, 512, 0xa7fde73fb74cf2d7, 0x000001866fb72748 }, /* 1.233 */
{ 41, 512, 0x19e8b461a1dea1d3, 0x000001a046f76b23 }, /* 1.271 */
{ 42, 512, 0x031c9b868cc3e976, 0x000001afa64c49d3 }, /* 1.263 */
{ 43, 512, 0xbaa5125faa781854, 0x000001c76789e278 }, /* 1.270 */
{ 44, 512, 0x4ed55052550d721b, 0x000001d800ccd8eb }, /* 1.281 */
{ 45, 512, 0x0fd63ddbdff90677, 0x000001f08ad59ed2 }, /* 1.282 */
{ 46, 512, 0x36d66546de7fdd6f, 0x000002016f09574b }, /* 1.286 */
{ 47, 512, 0x99f997e7eafb69d7, 0x0000021e42e47cb6 }, /* 1.329 */
{ 48, 512, 0xbecd9c2571312c5d, 0x000002320fe2872b }, /* 1.286 */
{ 49, 512, 0xd97371329e488a32, 0x0000024cd73f2ca7 }, /* 1.322 */
{ 50, 512, 0x30e9b136670749ee, 0x000002681c83b0e0 }, /* 1.335 */
{ 51, 512, 0x11ad6bc8f47aaeb4, 0x0000027e9261b5d5 }, /* 1.305 */
{ 52, 512, 0x68e445300af432c1, 0x0000029aa0eb7dbf }, /* 1.330 */
{ 53, 512, 0x910fb561657ea98c, 0x000002b3dca04853 }, /* 1.365 */
{ 54, 512, 0xd619693d8ce5e7a5, 0x000002cc280e9c97 }, /* 1.334 */
{ 55, 512, 0x24e281f564dbb60a, 0x000002e9fa842713 }, /* 1.364 */
{ 56, 512, 0x947a7d3bdaab44c5, 0x000003046680f72e }, /* 1.374 */
{ 57, 512, 0x2d44fec9c093e0de, 0x00000324198ba810 }, /* 1.363 */
{ 58, 512, 0x87743c272d29bb4c, 0x0000033ec48c9ac9 }, /* 1.401 */
{ 59, 512, 0x96aa3b6f67f5d923, 0x0000034faead902c }, /* 1.392 */
{ 60, 512, 0x94a4f1faf520b0d3, 0x0000037d713ab005 }, /* 1.360 */
{ 61, 512, 0xb13ed3a272f711a2, 0x00000397368f3cbd }, /* 1.396 */
{ 62, 512, 0x3b1b11805fa4a64a, 0x000003b8a5e2840c }, /* 1.453 */
{ 63, 512, 0x4c74caad9172ba71, 0x000003d4be280290 }, /* 1.437 */
{ 64, 512, 0x035ff643923dd29e, 0x000003fad6c355e1 }, /* 1.402 */
{ 65, 512, 0x768e9171b11abd3c, 0x0000040eb07fed20 }, /* 1.459 */
{ 66, 512, 0x75880e6f78a13ddd, 0x000004433d6acf14 }, /* 1.423 */
{ 67, 512, 0x910b9714f698a877, 0x00000451ea65d5db }, /* 1.447 */
{ 68, 512, 0x87f5db6f9fdcf5c7, 0x000004732169e3f7 }, /* 1.450 */
{ 69, 512, 0x836d4968fbaa3706, 0x000004954068a380 }, /* 1.455 */
{ 70, 512, 0xc567d73a036421ab, 0x000004bd7cb7bd3d }, /* 1.463 */
{ 71, 512, 0x619df40f240b8fed, 0x000004e376c2e972 }, /* 1.463 */
{ 72, 512, 0x42763a680d5bed8e, 0x000005084275c680 }, /* 1.452 */
{ 73, 512, 0x5866f064b3230431, 0x0000052906f2c9ab }, /* 1.498 */
{ 74, 512, 0x9fa08548b1621a44, 0x0000054708019247 }, /* 1.526 */
{ 75, 512, 0xb6053078ce0fc303, 0x00000572cc5c72b0 }, /* 1.491 */
{ 76, 512, 0x4a7aad7bf3890923, 0x0000058e987bc8e9 }, /* 1.470 */
{ 77, 512, 0xe165613fd75b5a53, 0x000005c20473a211 }, /* 1.527 */
{ 78, 512, 0x3ff154ac878163a6, 0x000005d659194bf3 }, /* 1.509 */
{ 79, 512, 0x24b93ade0aa8a532, 0x0000060a201c4f8e }, /* 1.569 */
{ 80, 512, 0xc18e2d14cd9bb554, 0x0000062c55cfe48c }, /* 1.555 */
{ 81, 512, 0x98cc78302feb58b6, 0x0000066656a07194 }, /* 1.509 */
{ 82, 512, 0xc6c5fd5a2abc0543, 0x0000067cff94fbf8 }, /* 1.596 */
{ 83, 512, 0xa7962f514acbba21, 0x000006ab7b5afa2e }, /* 1.568 */
{ 84, 512, 0xba02545069ddc6dc, 0x000006d19861364f }, /* 1.541 */
{ 85, 512, 0x447c73192c35073e, 0x000006fce315ce35 }, /* 1.623 */
{ 86, 512, 0x48beef9e2d42b0c2, 0x00000720a8e38b6b }, /* 1.620 */
{ 87, 512, 0x4874cf98541a35e0, 0x00000758382a2273 }, /* 1.597 */
{ 88, 512, 0xad4cf8333a31127a, 0x00000781e1651b1b }, /* 1.575 */
{ 89, 512, 0x47ae4859d57888c1, 0x000007b27edbe5bc }, /* 1.627 */
{ 90, 512, 0x06f7723cfe5d1891, 0x000007dc2a96d8eb }, /* 1.596 */
{ 91, 512, 0xd4e44218d660576d, 0x0000080ac46f02d5 }, /* 1.622 */
{ 92, 512, 0x7066702b0d5be1f2, 0x00000832c96d154e }, /* 1.695 */
{ 93, 512, 0x011209b4f9e11fb9, 0x0000085eefda104c }, /* 1.605 */
{ 94, 512, 0x47ffba30a0b35708, 0x00000899badc32dc }, /* 1.625 */
{ 95, 512, 0x1a95a6ac4538aaa8, 0x000008b6b69a42b2 }, /* 1.687 */
{ 96, 512, 0xbda2b239bb2008eb, 0x000008f22d2de38a }, /* 1.621 */
{ 97, 512, 0x7ffa0bea90355c6c, 0x0000092e5b23b816 }, /* 1.699 */
{ 98, 512, 0x1d56ba34be426795, 0x0000094f482e5d1b }, /* 1.688 */
{ 99, 512, 0x0aa89d45c502e93d, 0x00000977d94a98ce }, /* 1.642 */
{ 100, 512, 0x54369449f6857774, 0x000009c06c9b34cc }, /* 1.683 */
{ 101, 512, 0xf7d4dd8445b46765, 0x000009e5dc542259 }, /* 1.755 */
{ 102, 512, 0xfa8866312f169469, 0x00000a16b54eae93 }, /* 1.692 */
{ 103, 512, 0xd8a5aea08aef3ff9, 0x00000a381d2cbfe7 }, /* 1.747 */
{ 104, 512, 0x66bcd2c3d5f9ef0e, 0x00000a8191817be7 }, /* 1.751 */
{ 105, 512, 0x3fb13a47a012ec81, 0x00000ab562b9a254 }, /* 1.751 */
{ 106, 512, 0x43100f01c9e5e3ca, 0x00000aeee84c185f }, /* 1.726 */
{ 107, 512, 0xca09c50ccee2d054, 0x00000b1c359c047d }, /* 1.788 */
{ 108, 512, 0xd7176732ac503f9b, 0x00000b578bc52a73 }, /* 1.740 */
{ 109, 512, 0xed206e51f8d9422d, 0x00000b8083e0d960 }, /* 1.780 */
{ 110, 512, 0x17ead5dc6ba0dcd6, 0x00000bcfb1a32ca8 }, /* 1.836 */
{ 111, 512, 0x5f1dc21e38a969eb, 0x00000c0171becdd6 }, /* 1.778 */
{ 112, 512, 0xddaa973de33ec528, 0x00000c3edaba4b95 }, /* 1.831 */
{ 113, 512, 0x2a5eccd7735a3630, 0x00000c630664e7df }, /* 1.825 */
{ 114, 512, 0xafcccee5c0b71446, 0x00000cb65392f6e4 }, /* 1.826 */
{ 115, 512, 0x8fa30c5e7b147e27, 0x00000cd4db391e55 }, /* 1.843 */
{ 116, 512, 0x5afe0711fdfafd82, 0x00000d08cb4ec35d }, /* 1.826 */
{ 117, 512, 0x533a6090238afd4c, 0x00000d336f115d1b }, /* 1.803 */
{ 118, 512, 0x90cf11b595e39a84, 0x00000d8e041c2048 }, /* 1.857 */
{ 119, 512, 0x0d61a3b809444009, 0x00000dcb798afe35 }, /* 1.877 */
{ 120, 512, 0x7f34da0f54b0d114, 0x00000df3922664e1 }, /* 1.849 */
{ 121, 512, 0xa52258d5b72f6551, 0x00000e4d37a9872d }, /* 1.867 */
{ 122, 512, 0xc1de54d7672878db, 0x00000e6583a94cf6 }, /* 1.978 */
{ 123, 512, 0x1d03354316a414ab, 0x00000ebffc50308d }, /* 1.947 */
{ 124, 512, 0xcebdcc377665412c, 0x00000edee1997cea }, /* 1.865 */
{ 125, 512, 0x4ddd4c04b1a12344, 0x00000f21d64b373f }, /* 1.881 */
{ 126, 512, 0x64fc8f94e3973658, 0x00000f8f87a8896b }, /* 1.882 */
{ 127, 512, 0x68765f78034a334e, 0x00000fb8fe62197e }, /* 1.867 */
{ 128, 512, 0xaf36b871a303e816, 0x00000fec6f3afb1e }, /* 1.972 */
{ 129, 512, 0x2a4cbf73866c3a28, 0x00001027febfe4e5 }, /* 1.896 */
{ 130, 512, 0x9cb128aacdcd3b2f, 0x0000106aa8ac569d }, /* 1.965 */
{ 131, 512, 0x5511d41c55869124, 0x000010bbd755ddf1 }, /* 1.963 */
{ 132, 512, 0x42f92461937f284a, 0x000010fb8bceb3b5 }, /* 1.925 */
{ 133, 512, 0xe2d89a1cf6f1f287, 0x0000114cf5331e34 }, /* 1.862 */
{ 134, 512, 0xdc631a038956200e, 0x0000116428d2adc5 }, /* 2.042 */
{ 135, 512, 0xb2e5ac222cd236be, 0x000011ca88e4d4d2 }, /* 1.935 */
{ 136, 512, 0xbc7d8236655d88e7, 0x000011e39cb94e66 }, /* 2.005 */
{ 137, 512, 0x073e02d88d2d8e75, 0x0000123136c7933c }, /* 2.041 */
{ 138, 512, 0x3ddb9c3873166be0, 0x00001280e4ec6d52 }, /* 1.997 */
{ 139, 512, 0x7d3b1a845420e1b5, 0x000012c2e7cd6a44 }, /* 1.996 */
{ 140, 512, 0x60102308aa7b2a6c, 0x000012fc490e6c7d }, /* 2.053 */
{ 141, 512, 0xdb22bb2f9eb894aa, 0x00001343f5a85a1a }, /* 1.971 */
{ 142, 512, 0xd853f879a13b1606, 0x000013bb7d5f9048 }, /* 2.018 */
{ 143, 512, 0x001620a03f804b1d, 0x000013e74cc794fd }, /* 1.961 */
{ 144, 512, 0xfdb52dda76fbf667, 0x00001442d2f22480 }, /* 2.046 */
{ 145, 512, 0xa9160110f66e24ff, 0x0000144b899f9dbb }, /* 1.968 */
{ 146, 512, 0x77306a30379ae03b, 0x000014cb98eb1f81 }, /* 2.143 */
{ 147, 512, 0x14f5985d2752319d, 0x000014feab821fc9 }, /* 2.064 */
{ 148, 512, 0xa4b8ff11de7863f8, 0x0000154a0e60b9c9 }, /* 2.023 */
{ 149, 512, 0x44b345426455c1b3, 0x000015999c3c569c }, /* 2.136 */
{ 150, 512, 0x272677826049b46c, 0x000015c9697f4b92 }, /* 2.063 */
{ 151, 512, 0x2f9216e2cd74fe40, 0x0000162b1f7bbd39 }, /* 1.974 */
{ 152, 512, 0x706ae3e763ad8771, 0x00001661371c55e1 }, /* 2.210 */
{ 153, 512, 0xf7fd345307c2480e, 0x000016e251f28b6a }, /* 2.006 */
{ 154, 512, 0x6e94e3d26b3139eb, 0x000016f2429bb8c6 }, /* 2.193 */
{ 155, 512, 0x5458bbfbb781fcba, 0x0000173efdeca1b9 }, /* 2.163 */
{ 156, 512, 0xa80e2afeccd93b33, 0x000017bfdcb78adc }, /* 2.046 */
{ 157, 512, 0x1e4ccbb22796cf9d, 0x00001826fdcc39c9 }, /* 2.084 */
{ 158, 512, 0x8fba4b676aaa3663, 0x00001841a1379480 }, /* 2.264 */
{ 159, 512, 0xf82b843814b315fa, 0x000018886e19b8a3 }, /* 2.074 */
{ 160, 512, 0x7f21e920ecf753a3, 0x0000191812ca0ea7 }, /* 2.282 */
{ 161, 512, 0x48bb8ea2c4caa620, 0x0000192f310faccf }, /* 2.148 */
{ 162, 512, 0x5cdb652b4952c91b, 0x0000199e1d7437c7 }, /* 2.355 */
{ 163, 512, 0x6ac1ba6f78c06cd4, 0x000019cd11f82c70 }, /* 2.164 */
{ 164, 512, 0x9faf5f9ca2669a56, 0x00001a18d5431f6a }, /* 2.393 */
{ 165, 512, 0xaa57e9383eb01194, 0x00001a9e7d253d85 }, /* 2.178 */
{ 166, 512, 0x896967bf495c34d2, 0x00001afb8319b9fc }, /* 2.334 */
{ 167, 512, 0xdfad5f05de225f1b, 0x00001b3a59c3093b }, /* 2.266 */
{ 168, 512, 0xfd299a99f9f2abdd, 0x00001bb6f1a10799 }, /* 2.304 */
{ 169, 512, 0xdda239e798fe9fd4, 0x00001bfae0c9692d }, /* 2.218 */
{ 170, 512, 0x5fca670414a32c3e, 0x00001c22129dbcff }, /* 2.377 */
{ 171, 512, 0x1bb8934314b087de, 0x00001c955db36cd0 }, /* 2.155 */
{ 172, 512, 0xd96394b4b082200d, 0x00001cfc8619b7e6 }, /* 2.404 */
{ 173, 512, 0xb612a7735b1c8cbc, 0x00001d303acdd585 }, /* 2.205 */
{ 174, 512, 0x28e7430fe5875fe1, 0x00001d7ed5b3697d }, /* 2.359 */
{ 175, 512, 0x5038e89efdd981b9, 0x00001dc40ec35c59 }, /* 2.158 */
{ 176, 512, 0x075fd78f1d14db7c, 0x00001e31c83b4a2b }, /* 2.614 */
{ 177, 512, 0xc50fafdb5021be15, 0x00001e7cdac82fbc }, /* 2.239 */
{ 178, 512, 0xe6dc7572ce7b91c7, 0x00001edd8bb454fc }, /* 2.493 */
{ 179, 512, 0x21f7843e7beda537, 0x00001f3a8e019d6c }, /* 2.327 */
{ 180, 512, 0xc83385e20b43ec82, 0x00001f70735ec137 }, /* 2.231 */
{ 181, 512, 0xca818217dddb21fd, 0x0000201ca44c5a3c }, /* 2.237 */
{ 182, 512, 0xe6035defea48f933, 0x00002038e3346658 }, /* 2.691 */
{ 183, 512, 0x47262a4f953dac5a, 0x000020c2e554314e }, /* 2.170 */
{ 184, 512, 0xe24c7246260873ea, 0x000021197e618d64 }, /* 2.600 */
{ 185, 512, 0xeef6b57c9b58e9e1, 0x0000217ea48ecddc }, /* 2.391 */
{ 186, 512, 0x2becd3346e386142, 0x000021c496d4a5f9 }, /* 2.677 */
{ 187, 512, 0x63c6207bdf3b40a3, 0x0000220e0f2eec0c }, /* 2.410 */
{ 188, 512, 0x3056ce8989767d4b, 0x0000228eb76cd137 }, /* 2.776 */
{ 189, 512, 0x91af61c307cee780, 0x000022e17e2ea501 }, /* 2.266 */
{ 190, 512, 0xda359da225f6d54f, 0x00002358a2debc19 }, /* 2.717 */
{ 191, 512, 0x0a5f7a2a55607ba0, 0x0000238a79dac18c }, /* 2.474 */
{ 192, 512, 0x27bb75bf5224638a, 0x00002403a58e2351 }, /* 2.673 */
{ 193, 512, 0x1ebfdb94630f5d0f, 0x00002492a10cb339 }, /* 2.420 */
{ 194, 512, 0x6eae5e51d9c5f6fb, 0x000024ce4bf98715 }, /* 2.898 */
{ 195, 512, 0x08d903b4daedc2e0, 0x0000250d1e15886c }, /* 2.363 */
{ 196, 512, 0xc722a2f7fa7cd686, 0x0000258a99ed0c9e }, /* 2.747 */
{ 197, 512, 0x8f71faf0e54e361d, 0x000025dee11976f5 }, /* 2.531 */
{ 198, 512, 0x87f64695c91a54e7, 0x0000264e00a43da0 }, /* 2.707 */
{ 199, 512, 0xc719cbac2c336b92, 0x000026d327277ac1 }, /* 2.315 */
{ 200, 512, 0xe7e647afaf771ade, 0x000027523a5c44bf }, /* 3.012 */
{ 201, 512, 0x12d4b5c38ce8c946, 0x0000273898432545 }, /* 2.378 */
{ 202, 512, 0xf2e0cd4067bdc94a, 0x000027e47bb2c935 }, /* 2.969 */
{ 203, 512, 0x21b79f14d6d947d3, 0x0000281e64977f0d }, /* 2.594 */
{ 204, 512, 0x515093f952f18cd6, 0x0000289691a473fd }, /* 2.763 */
{ 205, 512, 0xd47b160a1b1022c8, 0x00002903e8b52411 }, /* 2.457 */
{ 206, 512, 0xc02fc96684715a16, 0x0000297515608601 }, /* 3.057 */
{ 207, 512, 0xef51e68efba72ed0, 0x000029ef73604804 }, /* 2.590 */
{ 208, 512, 0x9e3be6e5448b4f33, 0x00002a2846ed074b }, /* 3.047 */
{ 209, 512, 0x81d446c6d5fec063, 0x00002a92ca693455 }, /* 2.676 */
{ 210, 512, 0xff215de8224e57d5, 0x00002b2271fe3729 }, /* 2.993 */
{ 211, 512, 0xe2524d9ba8f69796, 0x00002b64b99c3ba2 }, /* 2.457 */
{ 212, 512, 0xf6b28e26097b7e4b, 0x00002bd768b6e068 }, /* 3.182 */
{ 213, 512, 0x893a487f30ce1644, 0x00002c67f722b4b2 }, /* 2.563 */
{ 214, 512, 0x386566c3fc9871df, 0x00002cc1cf8b4037 }, /* 3.025 */
{ 215, 512, 0x1e0ed78edf1f558a, 0x00002d3948d36c7f }, /* 2.730 */
{ 216, 512, 0xe3bc20c31e61f113, 0x00002d6d6b12e025 }, /* 3.036 */
{ 217, 512, 0xd6c3ad2e23021882, 0x00002deff7572241 }, /* 2.722 */
{ 218, 512, 0xb4a9f95cf0f69c5a, 0x00002e67d537aa36 }, /* 3.356 */
{ 219, 512, 0x6e98ed6f6c38e82f, 0x00002e9720626789 }, /* 2.697 */
{ 220, 512, 0x2e01edba33fddac7, 0x00002f407c6b0198 }, /* 2.979 */
{ 221, 512, 0x559d02e1f5f57ccc, 0x00002fb6a5ab4f24 }, /* 2.858 */
{ 222, 512, 0xac18f5a916adcd8e, 0x0000304ae1c5c57e }, /* 3.258 */
{ 223, 512, 0x15789fbaddb86f4b, 0x0000306f6e019c78 }, /* 2.693 */
{ 224, 512, 0xf4a9c36d5bc4c408, 0x000030da40434213 }, /* 3.259 */
{ 225, 512, 0xf640f90fd2727f44, 0x00003189ed37b90c }, /* 2.733 */
{ 226, 512, 0xb5313d390d61884a, 0x000031e152616b37 }, /* 3.235 */
{ 227, 512, 0x4bae6b3ce9160939, 0x0000321f40aeac42 }, /* 2.983 */
{ 228, 512, 0x838c34480f1a66a1, 0x000032f389c0f78e }, /* 3.308 */
{ 229, 512, 0xb1c4a52c8e3d6060, 0x0000330062a40284 }, /* 2.715 */
{ 230, 512, 0xe0f1110c6d0ed822, 0x0000338be435644f }, /* 3.540 */
{ 231, 512, 0x9f1a8ccdcea68d4b, 0x000034045a4e97e1 }, /* 2.779 */
{ 232, 512, 0x3261ed62223f3099, 0x000034702cfc401c }, /* 3.084 */
{ 233, 512, 0xf2191e2311022d65, 0x00003509dd19c9fc }, /* 2.987 */
{ 234, 512, 0xf102a395c2033abc, 0x000035654dc96fae }, /* 3.341 */
{ 235, 512, 0x11fe378f027906b6, 0x000035b5193b0264 }, /* 2.793 */
{ 236, 512, 0xf777f2c026b337aa, 0x000036704f5d9297 }, /* 3.518 */
{ 237, 512, 0x1b04e9c2ee143f32, 0x000036dfbb7af218 }, /* 2.962 */
{ 238, 512, 0x2fcec95266f9352c, 0x00003785c8df24a9 }, /* 3.196 */
{ 239, 512, 0xfe2b0e47e427dd85, 0x000037cbdf5da729 }, /* 2.914 */
{ 240, 512, 0x72b49bf2225f6c6d, 0x0000382227c15855 }, /* 3.408 */
{ 241, 512, 0x50486b43df7df9c7, 0x0000389b88be6453 }, /* 2.903 */
{ 242, 512, 0x5192a3e53181c8ab, 0x000038ddf3d67263 }, /* 3.778 */
{ 243, 512, 0xe9f5d8365296fd5e, 0x0000399f1c6c9e9c }, /* 3.026 */
{ 244, 512, 0xc740263f0301efa8, 0x00003a147146512d }, /* 3.347 */
{ 245, 512, 0x23cd0f2b5671e67d, 0x00003ab10bcc0d9d }, /* 3.212 */
{ 246, 512, 0x002ccc7e5cd41390, 0x00003ad6cd14a6c0 }, /* 3.482 */
{ 247, 512, 0x9aafb3c02544b31b, 0x00003b8cb8779fb0 }, /* 3.146 */
{ 248, 512, 0x72ba07a78b121999, 0x00003c24142a5a3f }, /* 3.626 */
{ 249, 512, 0x3d784aa58edfc7b4, 0x00003cd084817d99 }, /* 2.952 */
{ 250, 512, 0xaab750424d8004af, 0x00003d506a8e098e }, /* 3.463 */
{ 251, 512, 0x84403fcf8e6b5ca2, 0x00003d4c54c2aec4 }, /* 3.131 */
{ 252, 512, 0x71eb7455ec98e207, 0x00003e655715cf2c }, /* 3.538 */
{ 253, 512, 0xd752b4f19301595b, 0x00003ecd7b2ca5ac }, /* 2.974 */
{ 254, 512, 0xc4674129750499de, 0x00003e99e86d3e95 }, /* 3.843 */
{ 255, 512, 0x9772baff5cd12ef5, 0x00003f895c019841 }, /* 3.088 */
};
/*
* Verify the map is valid. Each device index must appear exactly
* once in every row, and the permutation array checksum must match.
*/
static int
verify_perms(uint8_t *perms, uint64_t children, uint64_t nperms,
uint64_t checksum)
{
int countssz = sizeof (uint16_t) * children;
uint16_t *counts = kmem_zalloc(countssz, KM_SLEEP);
for (int i = 0; i < nperms; i++) {
for (int j = 0; j < children; j++) {
uint8_t val = perms[(i * children) + j];
if (val >= children || counts[val] != i) {
kmem_free(counts, countssz);
return (EINVAL);
}
counts[val]++;
}
}
if (checksum != 0) {
int permssz = sizeof (uint8_t) * children * nperms;
zio_cksum_t cksum;
fletcher_4_native_varsize(perms, permssz, &cksum);
if (checksum != cksum.zc_word[0]) {
kmem_free(counts, countssz);
return (ECKSUM);
}
}
kmem_free(counts, countssz);
return (0);
}
/*
* Generate the permutation array for the draid_map_t. These maps control
* the placement of all data in a dRAID. Therefore it's critical that the
* seed always generates the same mapping. We provide our own pseudo-random
* number generator for this purpose.
*/
int
vdev_draid_generate_perms(const draid_map_t *map, uint8_t **permsp)
{
VERIFY3U(map->dm_children, >=, VDEV_DRAID_MIN_CHILDREN);
VERIFY3U(map->dm_children, <=, VDEV_DRAID_MAX_CHILDREN);
VERIFY3U(map->dm_seed, !=, 0);
VERIFY3U(map->dm_nperms, !=, 0);
VERIFY3P(map->dm_perms, ==, NULL);
#ifdef _KERNEL
/*
* The kernel code always provides both a map_seed and checksum.
* Only the tests/zfs-tests/cmd/draid/draid.c utility will provide
* a zero checksum when generating new candidate maps.
*/
VERIFY3U(map->dm_checksum, !=, 0);
#endif
uint64_t children = map->dm_children;
uint64_t nperms = map->dm_nperms;
int rowsz = sizeof (uint8_t) * children;
int permssz = rowsz * nperms;
uint8_t *perms;
/* Allocate the permutation array */
perms = vmem_alloc(permssz, KM_SLEEP);
/* Setup an initial row with a known pattern */
uint8_t *initial_row = kmem_alloc(rowsz, KM_SLEEP);
for (int i = 0; i < children; i++)
initial_row[i] = i;
uint64_t draid_seed[2] = { VDEV_DRAID_SEED, map->dm_seed };
uint8_t *current_row, *previous_row = initial_row;
/*
* Perform a Fisher-Yates shuffle of each row using the previous
* row as the starting point. An initial_row with known pattern
* is used as the input for the first row.
*/
for (int i = 0; i < nperms; i++) {
current_row = &perms[i * children];
memcpy(current_row, previous_row, rowsz);
for (int j = children - 1; j > 0; j--) {
uint64_t k = vdev_draid_rand(draid_seed) % (j + 1);
uint8_t val = current_row[j];
current_row[j] = current_row[k];
current_row[k] = val;
}
previous_row = current_row;
}
kmem_free(initial_row, rowsz);
int error = verify_perms(perms, children, nperms, map->dm_checksum);
if (error) {
vmem_free(perms, permssz);
return (error);
}
*permsp = perms;
return (0);
}
/*
* Lookup the fixed draid_map_t for the requested number of children.
*/
int
vdev_draid_lookup_map(uint64_t children, const draid_map_t **mapp)
{
for (int i = 0; i < VDEV_DRAID_MAX_MAPS; i++) {
if (draid_maps[i].dm_children == children) {
*mapp = &draid_maps[i];
return (0);
}
}
return (ENOENT);
}
/*
* Lookup the permutation array and iteration id for the provided offset.
*/
static void
vdev_draid_get_perm(vdev_draid_config_t *vdc, uint64_t pindex,
uint8_t **base, uint64_t *iter)
{
uint64_t ncols = vdc->vdc_children;
uint64_t poff = pindex % (vdc->vdc_nperms * ncols);
*base = vdc->vdc_perms + (poff / ncols) * ncols;
*iter = poff % ncols;
}
static inline uint64_t
vdev_draid_permute_id(vdev_draid_config_t *vdc,
uint8_t *base, uint64_t iter, uint64_t index)
{
return ((base[index] + iter) % vdc->vdc_children);
}
/*
* Return the asize which is the psize rounded up to a full group width.
* i.e. vdev_draid_psize_to_asize().
*/
static uint64_t
-vdev_draid_asize(vdev_t *vd, uint64_t psize)
+vdev_draid_asize(vdev_t *vd, uint64_t psize, uint64_t txg)
{
+ (void) txg;
vdev_draid_config_t *vdc = vd->vdev_tsd;
uint64_t ashift = vd->vdev_ashift;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
uint64_t rows = ((psize - 1) / (vdc->vdc_ndata << ashift)) + 1;
uint64_t asize = (rows * vdc->vdc_groupwidth) << ashift;
ASSERT3U(asize, !=, 0);
ASSERT3U(asize % (vdc->vdc_groupwidth), ==, 0);
return (asize);
}
/*
* Deflate the asize to the psize, this includes stripping parity.
*/
uint64_t
vdev_draid_asize_to_psize(vdev_t *vd, uint64_t asize)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT0(asize % vdc->vdc_groupwidth);
return ((asize / vdc->vdc_groupwidth) * vdc->vdc_ndata);
}
/*
* Convert a logical offset to the corresponding group number.
*/
static uint64_t
vdev_draid_offset_to_group(vdev_t *vd, uint64_t offset)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
return (offset / vdc->vdc_groupsz);
}
/*
* Convert a group number to the logical starting offset for that group.
*/
static uint64_t
vdev_draid_group_to_offset(vdev_t *vd, uint64_t group)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
return (group * vdc->vdc_groupsz);
}
/*
* Full stripe writes. When writing, all columns (D+P) are required. Parity
* is calculated over all the columns, including empty zero filled sectors,
* and each is written to disk. While only the data columns are needed for
* a normal read, all of the columns are required for reconstruction when
* performing a sequential resilver.
*
* For "big columns" it's sufficient to map the correct range of the zio ABD.
* Partial columns require allocating a gang ABD in order to zero fill the
* empty sectors. When the column is empty a zero filled sector must be
* mapped. In all cases the data ABDs must be the same size as the parity
* ABDs (e.g. rc->rc_size == parity_size).
*/
static void
vdev_draid_map_alloc_write(zio_t *zio, uint64_t abd_offset, raidz_row_t *rr)
{
uint64_t skip_size = 1ULL << zio->io_vd->vdev_top->vdev_ashift;
uint64_t parity_size = rr->rr_col[0].rc_size;
uint64_t abd_off = abd_offset;
ASSERT3U(zio->io_type, ==, ZIO_TYPE_WRITE);
ASSERT3U(parity_size, ==, abd_get_size(rr->rr_col[0].rc_abd));
for (uint64_t c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_size == 0) {
/* empty data column (small write), add a skip sector */
ASSERT3U(skip_size, ==, parity_size);
rc->rc_abd = abd_get_zeros(skip_size);
} else if (rc->rc_size == parity_size) {
/* this is a "big column" */
rc->rc_abd = abd_get_offset_struct(&rc->rc_abdstruct,
zio->io_abd, abd_off, rc->rc_size);
} else {
/* short data column, add a skip sector */
ASSERT3U(rc->rc_size + skip_size, ==, parity_size);
rc->rc_abd = abd_alloc_gang();
abd_gang_add(rc->rc_abd, abd_get_offset_size(
zio->io_abd, abd_off, rc->rc_size), B_TRUE);
abd_gang_add(rc->rc_abd, abd_get_zeros(skip_size),
B_TRUE);
}
ASSERT3U(abd_get_size(rc->rc_abd), ==, parity_size);
abd_off += rc->rc_size;
rc->rc_size = parity_size;
}
IMPLY(abd_offset != 0, abd_off == zio->io_size);
}
/*
* Scrub/resilver reads. In order to store the contents of the skip sectors
* an additional ABD is allocated. The columns are handled in the same way
* as a full stripe write except instead of using the zero ABD the newly
* allocated skip ABD is used to back the skip sectors. In all cases the
* data ABD must be the same size as the parity ABDs.
*/
static void
vdev_draid_map_alloc_scrub(zio_t *zio, uint64_t abd_offset, raidz_row_t *rr)
{
uint64_t skip_size = 1ULL << zio->io_vd->vdev_top->vdev_ashift;
uint64_t parity_size = rr->rr_col[0].rc_size;
uint64_t abd_off = abd_offset;
uint64_t skip_off = 0;
ASSERT3U(zio->io_type, ==, ZIO_TYPE_READ);
ASSERT3P(rr->rr_abd_empty, ==, NULL);
if (rr->rr_nempty > 0) {
rr->rr_abd_empty = abd_alloc_linear(rr->rr_nempty * skip_size,
B_FALSE);
}
for (uint64_t c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_size == 0) {
/* empty data column (small read), add a skip sector */
ASSERT3U(skip_size, ==, parity_size);
ASSERT3U(rr->rr_nempty, !=, 0);
rc->rc_abd = abd_get_offset_size(rr->rr_abd_empty,
skip_off, skip_size);
skip_off += skip_size;
} else if (rc->rc_size == parity_size) {
/* this is a "big column" */
rc->rc_abd = abd_get_offset_struct(&rc->rc_abdstruct,
zio->io_abd, abd_off, rc->rc_size);
} else {
/* short data column, add a skip sector */
ASSERT3U(rc->rc_size + skip_size, ==, parity_size);
ASSERT3U(rr->rr_nempty, !=, 0);
rc->rc_abd = abd_alloc_gang();
abd_gang_add(rc->rc_abd, abd_get_offset_size(
zio->io_abd, abd_off, rc->rc_size), B_TRUE);
abd_gang_add(rc->rc_abd, abd_get_offset_size(
rr->rr_abd_empty, skip_off, skip_size), B_TRUE);
skip_off += skip_size;
}
uint64_t abd_size = abd_get_size(rc->rc_abd);
ASSERT3U(abd_size, ==, abd_get_size(rr->rr_col[0].rc_abd));
/*
* Increase rc_size so the skip ABD is included in subsequent
* parity calculations.
*/
abd_off += rc->rc_size;
rc->rc_size = abd_size;
}
IMPLY(abd_offset != 0, abd_off == zio->io_size);
ASSERT3U(skip_off, ==, rr->rr_nempty * skip_size);
}
/*
* Normal reads. In this common case only the columns containing data
* are read in to the zio ABDs. Neither the parity columns or empty skip
* sectors are read unless the checksum fails verification. In which case
* vdev_raidz_read_all() will call vdev_draid_map_alloc_empty() to expand
* the raid map in order to allow reconstruction using the parity data and
* skip sectors.
*/
static void
vdev_draid_map_alloc_read(zio_t *zio, uint64_t abd_offset, raidz_row_t *rr)
{
uint64_t abd_off = abd_offset;
ASSERT3U(zio->io_type, ==, ZIO_TYPE_READ);
for (uint64_t c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_size > 0) {
rc->rc_abd = abd_get_offset_struct(&rc->rc_abdstruct,
zio->io_abd, abd_off, rc->rc_size);
abd_off += rc->rc_size;
}
}
IMPLY(abd_offset != 0, abd_off == zio->io_size);
}
/*
* Converts a normal "read" raidz_row_t to a "scrub" raidz_row_t. The key
* difference is that an ABD is allocated to back skip sectors so they may
* be read in to memory, verified, and repaired if needed.
*/
void
vdev_draid_map_alloc_empty(zio_t *zio, raidz_row_t *rr)
{
uint64_t skip_size = 1ULL << zio->io_vd->vdev_top->vdev_ashift;
uint64_t parity_size = rr->rr_col[0].rc_size;
uint64_t skip_off = 0;
ASSERT3U(zio->io_type, ==, ZIO_TYPE_READ);
ASSERT3P(rr->rr_abd_empty, ==, NULL);
if (rr->rr_nempty > 0) {
rr->rr_abd_empty = abd_alloc_linear(rr->rr_nempty * skip_size,
B_FALSE);
}
for (uint64_t c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_size == 0) {
/* empty data column (small read), add a skip sector */
ASSERT3U(skip_size, ==, parity_size);
ASSERT3U(rr->rr_nempty, !=, 0);
ASSERT3P(rc->rc_abd, ==, NULL);
rc->rc_abd = abd_get_offset_size(rr->rr_abd_empty,
skip_off, skip_size);
skip_off += skip_size;
} else if (rc->rc_size == parity_size) {
/* this is a "big column", nothing to add */
ASSERT3P(rc->rc_abd, !=, NULL);
} else {
/*
* short data column, add a skip sector and clear
* rc_tried to force the entire column to be re-read
* thereby including the missing skip sector data
* which is needed for reconstruction.
*/
ASSERT3U(rc->rc_size + skip_size, ==, parity_size);
ASSERT3U(rr->rr_nempty, !=, 0);
ASSERT3P(rc->rc_abd, !=, NULL);
ASSERT(!abd_is_gang(rc->rc_abd));
abd_t *read_abd = rc->rc_abd;
rc->rc_abd = abd_alloc_gang();
abd_gang_add(rc->rc_abd, read_abd, B_TRUE);
abd_gang_add(rc->rc_abd, abd_get_offset_size(
rr->rr_abd_empty, skip_off, skip_size), B_TRUE);
skip_off += skip_size;
rc->rc_tried = 0;
}
/*
* Increase rc_size so the empty ABD is included in subsequent
* parity calculations.
*/
rc->rc_size = parity_size;
}
ASSERT3U(skip_off, ==, rr->rr_nempty * skip_size);
}
/*
* Verify that all empty sectors are zero filled before using them to
* calculate parity. Otherwise, silent corruption in an empty sector will
* result in bad parity being generated. That bad parity will then be
* considered authoritative and overwrite the good parity on disk. This
* is possible because the checksum is only calculated over the data,
* thus it cannot be used to detect damage in empty sectors.
*/
int
vdev_draid_map_verify_empty(zio_t *zio, raidz_row_t *rr)
{
uint64_t skip_size = 1ULL << zio->io_vd->vdev_top->vdev_ashift;
uint64_t parity_size = rr->rr_col[0].rc_size;
uint64_t skip_off = parity_size - skip_size;
uint64_t empty_off = 0;
int ret = 0;
ASSERT3U(zio->io_type, ==, ZIO_TYPE_READ);
ASSERT3P(rr->rr_abd_empty, !=, NULL);
ASSERT3U(rr->rr_bigcols, >, 0);
void *zero_buf = kmem_zalloc(skip_size, KM_SLEEP);
for (int c = rr->rr_bigcols; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
ASSERT3P(rc->rc_abd, !=, NULL);
ASSERT3U(rc->rc_size, ==, parity_size);
if (abd_cmp_buf_off(rc->rc_abd, zero_buf, skip_off,
skip_size) != 0) {
vdev_raidz_checksum_error(zio, rc, rc->rc_abd);
abd_zero_off(rc->rc_abd, skip_off, skip_size);
rc->rc_error = SET_ERROR(ECKSUM);
ret++;
}
empty_off += skip_size;
}
ASSERT3U(empty_off, ==, abd_get_size(rr->rr_abd_empty));
kmem_free(zero_buf, skip_size);
return (ret);
}
/*
* Given a logical address within a dRAID configuration, return the physical
* address on the first drive in the group that this address maps to
* (at position 'start' in permutation number 'perm').
*/
static uint64_t
vdev_draid_logical_to_physical(vdev_t *vd, uint64_t logical_offset,
uint64_t *perm, uint64_t *start)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
/* b is the dRAID (parent) sector offset. */
uint64_t ashift = vd->vdev_top->vdev_ashift;
uint64_t b_offset = logical_offset >> ashift;
/*
* The height of a row in units of the vdev's minimum sector size.
* This is the amount of data written to each disk of each group
* in a given permutation.
*/
uint64_t rowheight_sectors = VDEV_DRAID_ROWHEIGHT >> ashift;
/*
* We cycle through a disk permutation every groupsz * ngroups chunk
* of address space. Note that ngroups * groupsz must be a multiple
* of the number of data drives (ndisks) in order to guarantee
* alignment. So, for example, if our row height is 16MB, our group
* size is 10, and there are 13 data drives in the draid, then ngroups
* will be 13, we will change permutation every 2.08GB and each
* disk will have 160MB of data per chunk.
*/
uint64_t groupwidth = vdc->vdc_groupwidth;
uint64_t ngroups = vdc->vdc_ngroups;
uint64_t ndisks = vdc->vdc_ndisks;
/*
* groupstart is where the group this IO will land in "starts" in
* the permutation array.
*/
uint64_t group = logical_offset / vdc->vdc_groupsz;
uint64_t groupstart = (group * groupwidth) % ndisks;
ASSERT3U(groupstart + groupwidth, <=, ndisks + groupstart);
*start = groupstart;
/* b_offset is the sector offset within a group chunk */
b_offset = b_offset % (rowheight_sectors * groupwidth);
ASSERT0(b_offset % groupwidth);
/*
* Find the starting byte offset on each child vdev:
* - within a permutation there are ngroups groups spread over the
* rows, where each row covers a slice portion of the disk
* - each permutation has (groupwidth * ngroups) / ndisks rows
* - so each permutation covers rows * slice portion of the disk
* - so we need to find the row where this IO group target begins
*/
*perm = group / ngroups;
uint64_t row = (*perm * ((groupwidth * ngroups) / ndisks)) +
(((group % ngroups) * groupwidth) / ndisks);
return (((rowheight_sectors * row) +
(b_offset / groupwidth)) << ashift);
}
static uint64_t
vdev_draid_map_alloc_row(zio_t *zio, raidz_row_t **rrp, uint64_t io_offset,
uint64_t abd_offset, uint64_t abd_size)
{
vdev_t *vd = zio->io_vd;
vdev_draid_config_t *vdc = vd->vdev_tsd;
uint64_t ashift = vd->vdev_top->vdev_ashift;
uint64_t io_size = abd_size;
- uint64_t io_asize = vdev_draid_asize(vd, io_size);
+ uint64_t io_asize = vdev_draid_asize(vd, io_size, 0);
uint64_t group = vdev_draid_offset_to_group(vd, io_offset);
uint64_t start_offset = vdev_draid_group_to_offset(vd, group + 1);
/*
* Limit the io_size to the space remaining in the group. A second
* row in the raidz_map_t is created for the remainder.
*/
if (io_offset + io_asize > start_offset) {
io_size = vdev_draid_asize_to_psize(vd,
start_offset - io_offset);
}
/*
* At most a block may span the logical end of one group and the start
* of the next group. Therefore, at the end of a group the io_size must
* span the group width evenly and the remainder must be aligned to the
* start of the next group.
*/
IMPLY(abd_offset == 0 && io_size < zio->io_size,
(io_asize >> ashift) % vdc->vdc_groupwidth == 0);
IMPLY(abd_offset != 0,
vdev_draid_group_to_offset(vd, group) == io_offset);
/* Lookup starting byte offset on each child vdev */
uint64_t groupstart, perm;
uint64_t physical_offset = vdev_draid_logical_to_physical(vd,
io_offset, &perm, &groupstart);
/*
* If there is less than groupwidth drives available after the group
* start, the group is going to wrap onto the next row. 'wrap' is the
* group disk number that starts on the next row.
*/
uint64_t ndisks = vdc->vdc_ndisks;
uint64_t groupwidth = vdc->vdc_groupwidth;
uint64_t wrap = groupwidth;
if (groupstart + groupwidth > ndisks)
wrap = ndisks - groupstart;
/* The io size in units of the vdev's minimum sector size. */
const uint64_t psize = io_size >> ashift;
/*
* "Quotient": The number of data sectors for this stripe on all but
* the "big column" child vdevs that also contain "remainder" data.
*/
uint64_t q = psize / vdc->vdc_ndata;
/*
* "Remainder": The number of partial stripe data sectors in this I/O.
* This will add a sector to some, but not all, child vdevs.
*/
uint64_t r = psize - q * vdc->vdc_ndata;
/* The number of "big columns" - those which contain remainder data. */
uint64_t bc = (r == 0 ? 0 : r + vdc->vdc_nparity);
ASSERT3U(bc, <, groupwidth);
/* The total number of data and parity sectors for this I/O. */
uint64_t tot = psize + (vdc->vdc_nparity * (q + (r == 0 ? 0 : 1)));
ASSERT3U(vdc->vdc_nparity, >, 0);
- raidz_row_t *rr;
- rr = kmem_alloc(offsetof(raidz_row_t, rr_col[groupwidth]), KM_SLEEP);
- rr->rr_cols = groupwidth;
- rr->rr_scols = groupwidth;
+ raidz_row_t *rr = vdev_raidz_row_alloc(groupwidth);
rr->rr_bigcols = bc;
- rr->rr_missingdata = 0;
- rr->rr_missingparity = 0;
rr->rr_firstdatacol = vdc->vdc_nparity;
- rr->rr_abd_empty = NULL;
#ifdef ZFS_DEBUG
rr->rr_offset = io_offset;
rr->rr_size = io_size;
#endif
*rrp = rr;
uint8_t *base;
uint64_t iter, asize = 0;
vdev_draid_get_perm(vdc, perm, &base, &iter);
for (uint64_t i = 0; i < groupwidth; i++) {
raidz_col_t *rc = &rr->rr_col[i];
uint64_t c = (groupstart + i) % ndisks;
/* increment the offset if we wrap to the next row */
if (i == wrap)
physical_offset += VDEV_DRAID_ROWHEIGHT;
rc->rc_devidx = vdev_draid_permute_id(vdc, base, iter, c);
rc->rc_offset = physical_offset;
- rc->rc_abd = NULL;
- rc->rc_orig_data = NULL;
- rc->rc_error = 0;
- rc->rc_tried = 0;
- rc->rc_skipped = 0;
- rc->rc_force_repair = 0;
- rc->rc_allow_repair = 1;
- rc->rc_need_orig_restore = B_FALSE;
if (q == 0 && i >= bc)
rc->rc_size = 0;
else if (i < bc)
rc->rc_size = (q + 1) << ashift;
else
rc->rc_size = q << ashift;
asize += rc->rc_size;
}
ASSERT3U(asize, ==, tot << ashift);
rr->rr_nempty = roundup(tot, groupwidth) - tot;
IMPLY(bc > 0, rr->rr_nempty == groupwidth - bc);
/* Allocate buffers for the parity columns */
for (uint64_t c = 0; c < rr->rr_firstdatacol; c++) {
raidz_col_t *rc = &rr->rr_col[c];
rc->rc_abd = abd_alloc_linear(rc->rc_size, B_FALSE);
}
/*
* Map buffers for data columns and allocate/map buffers for skip
* sectors. There are three distinct cases for dRAID which are
* required to support sequential rebuild.
*/
if (zio->io_type == ZIO_TYPE_WRITE) {
vdev_draid_map_alloc_write(zio, abd_offset, rr);
} else if ((rr->rr_nempty > 0) &&
(zio->io_flags & (ZIO_FLAG_SCRUB | ZIO_FLAG_RESILVER))) {
vdev_draid_map_alloc_scrub(zio, abd_offset, rr);
} else {
ASSERT3U(zio->io_type, ==, ZIO_TYPE_READ);
vdev_draid_map_alloc_read(zio, abd_offset, rr);
}
return (io_size);
}
/*
* Allocate the raidz mapping to be applied to the dRAID I/O. The parity
* calculations for dRAID are identical to raidz however there are a few
* differences in the layout.
*
* - dRAID always allocates a full stripe width. Any extra sectors due
* this padding are zero filled and written to disk. They will be read
* back during a scrub or repair operation since they are included in
* the parity calculation. This property enables sequential resilvering.
*
* - When the block at the logical offset spans redundancy groups then two
* rows are allocated in the raidz_map_t. One row resides at the end of
* the first group and the other at the start of the following group.
*/
static raidz_map_t *
vdev_draid_map_alloc(zio_t *zio)
{
raidz_row_t *rr[2];
uint64_t abd_offset = 0;
uint64_t abd_size = zio->io_size;
uint64_t io_offset = zio->io_offset;
uint64_t size;
int nrows = 1;
size = vdev_draid_map_alloc_row(zio, &rr[0], io_offset,
abd_offset, abd_size);
if (size < abd_size) {
vdev_t *vd = zio->io_vd;
- io_offset += vdev_draid_asize(vd, size);
+ io_offset += vdev_draid_asize(vd, size, 0);
abd_offset += size;
abd_size -= size;
nrows++;
ASSERT3U(io_offset, ==, vdev_draid_group_to_offset(
vd, vdev_draid_offset_to_group(vd, io_offset)));
ASSERT3U(abd_offset, <, zio->io_size);
ASSERT3U(abd_size, !=, 0);
size = vdev_draid_map_alloc_row(zio, &rr[1],
io_offset, abd_offset, abd_size);
VERIFY3U(size, ==, abd_size);
}
raidz_map_t *rm;
rm = kmem_zalloc(offsetof(raidz_map_t, rm_row[nrows]), KM_SLEEP);
rm->rm_ops = vdev_raidz_math_get_ops();
rm->rm_nrows = nrows;
rm->rm_row[0] = rr[0];
if (nrows == 2)
rm->rm_row[1] = rr[1];
-
return (rm);
}
/*
* Given an offset into a dRAID return the next group width aligned offset
* which can be used to start an allocation.
*/
static uint64_t
vdev_draid_get_astart(vdev_t *vd, const uint64_t start)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
return (roundup(start, vdc->vdc_groupwidth << vd->vdev_ashift));
}
/*
* Allocatable space for dRAID is (children - nspares) * sizeof(smallest child)
* rounded down to the last full slice. So each child must provide at least
* 1 / (children - nspares) of its asize.
*/
static uint64_t
vdev_draid_min_asize(vdev_t *vd)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
return (VDEV_DRAID_REFLOW_RESERVE +
(vd->vdev_min_asize + vdc->vdc_ndisks - 1) / (vdc->vdc_ndisks));
}
/*
* When using dRAID the minimum allocation size is determined by the number
* of data disks in the redundancy group. Full stripes are always used.
*/
static uint64_t
vdev_draid_min_alloc(vdev_t *vd)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
return (vdc->vdc_ndata << vd->vdev_ashift);
}
/*
* Returns true if the txg range does not exist on any leaf vdev.
*
* A dRAID spare does not fit into the DTL model. While it has child vdevs
* there is no redundancy among them, and the effective child vdev is
* determined by offset. Essentially we do a vdev_dtl_reassess() on the
* fly by replacing a dRAID spare with the child vdev under the offset.
* Note that it is a recursive process because the child vdev can be
* another dRAID spare and so on.
*/
boolean_t
vdev_draid_missing(vdev_t *vd, uint64_t physical_offset, uint64_t txg,
uint64_t size)
{
if (vd->vdev_ops == &vdev_spare_ops ||
vd->vdev_ops == &vdev_replacing_ops) {
/*
* Check all of the readable children, if any child
* contains the txg range the data it is not missing.
*/
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (!vdev_readable(cvd))
continue;
if (!vdev_draid_missing(cvd, physical_offset,
txg, size))
return (B_FALSE);
}
return (B_TRUE);
}
if (vd->vdev_ops == &vdev_draid_spare_ops) {
/*
* When sequentially resilvering we don't have a proper
* txg range so instead we must presume all txgs are
* missing on this vdev until the resilver completes.
*/
if (vd->vdev_rebuild_txg != 0)
return (B_TRUE);
/*
* DTL_MISSING is set for all prior txgs when a resilver
* is started in spa_vdev_attach().
*/
if (vdev_dtl_contains(vd, DTL_MISSING, txg, size))
return (B_TRUE);
/*
* Consult the DTL on the relevant vdev. Either a vdev
* leaf or spare/replace mirror child may be returned so
* we must recursively call vdev_draid_missing_impl().
*/
vd = vdev_draid_spare_get_child(vd, physical_offset);
if (vd == NULL)
return (B_TRUE);
return (vdev_draid_missing(vd, physical_offset,
txg, size));
}
return (vdev_dtl_contains(vd, DTL_MISSING, txg, size));
}
/*
* Returns true if the txg is only partially replicated on the leaf vdevs.
*/
static boolean_t
vdev_draid_partial(vdev_t *vd, uint64_t physical_offset, uint64_t txg,
uint64_t size)
{
if (vd->vdev_ops == &vdev_spare_ops ||
vd->vdev_ops == &vdev_replacing_ops) {
/*
* Check all of the readable children, if any child is
* missing the txg range then it is partially replicated.
*/
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (!vdev_readable(cvd))
continue;
if (vdev_draid_partial(cvd, physical_offset, txg, size))
return (B_TRUE);
}
return (B_FALSE);
}
if (vd->vdev_ops == &vdev_draid_spare_ops) {
/*
* When sequentially resilvering we don't have a proper
* txg range so instead we must presume all txgs are
* missing on this vdev until the resilver completes.
*/
if (vd->vdev_rebuild_txg != 0)
return (B_TRUE);
/*
* DTL_MISSING is set for all prior txgs when a resilver
* is started in spa_vdev_attach().
*/
if (vdev_dtl_contains(vd, DTL_MISSING, txg, size))
return (B_TRUE);
/*
* Consult the DTL on the relevant vdev. Either a vdev
* leaf or spare/replace mirror child may be returned so
* we must recursively call vdev_draid_missing_impl().
*/
vd = vdev_draid_spare_get_child(vd, physical_offset);
if (vd == NULL)
return (B_TRUE);
return (vdev_draid_partial(vd, physical_offset, txg, size));
}
return (vdev_dtl_contains(vd, DTL_MISSING, txg, size));
}
/*
* Determine if the vdev is readable at the given offset.
*/
boolean_t
vdev_draid_readable(vdev_t *vd, uint64_t physical_offset)
{
if (vd->vdev_ops == &vdev_draid_spare_ops) {
vd = vdev_draid_spare_get_child(vd, physical_offset);
if (vd == NULL)
return (B_FALSE);
}
if (vd->vdev_ops == &vdev_spare_ops ||
vd->vdev_ops == &vdev_replacing_ops) {
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (!vdev_readable(cvd))
continue;
if (vdev_draid_readable(cvd, physical_offset))
return (B_TRUE);
}
return (B_FALSE);
}
return (vdev_readable(vd));
}
/*
* Returns the first distributed spare found under the provided vdev tree.
*/
static vdev_t *
vdev_draid_find_spare(vdev_t *vd)
{
if (vd->vdev_ops == &vdev_draid_spare_ops)
return (vd);
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *svd = vdev_draid_find_spare(vd->vdev_child[c]);
if (svd != NULL)
return (svd);
}
return (NULL);
}
/*
* Returns B_TRUE if the passed in vdev is currently "faulted".
* Faulted, in this context, means that the vdev represents a
* replacing or sparing vdev tree.
*/
static boolean_t
vdev_draid_faulted(vdev_t *vd, uint64_t physical_offset)
{
if (vd->vdev_ops == &vdev_draid_spare_ops) {
vd = vdev_draid_spare_get_child(vd, physical_offset);
if (vd == NULL)
return (B_FALSE);
/*
* After resolving the distributed spare to a leaf vdev
* check the parent to determine if it's "faulted".
*/
vd = vd->vdev_parent;
}
return (vd->vdev_ops == &vdev_replacing_ops ||
vd->vdev_ops == &vdev_spare_ops);
}
/*
* Determine if the dRAID block at the logical offset is degraded.
* Used by sequential resilver.
*/
static boolean_t
vdev_draid_group_degraded(vdev_t *vd, uint64_t offset)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
ASSERT3U(vdev_draid_get_astart(vd, offset), ==, offset);
uint64_t groupstart, perm;
uint64_t physical_offset = vdev_draid_logical_to_physical(vd,
offset, &perm, &groupstart);
uint8_t *base;
uint64_t iter;
vdev_draid_get_perm(vdc, perm, &base, &iter);
for (uint64_t i = 0; i < vdc->vdc_groupwidth; i++) {
uint64_t c = (groupstart + i) % vdc->vdc_ndisks;
uint64_t cid = vdev_draid_permute_id(vdc, base, iter, c);
vdev_t *cvd = vd->vdev_child[cid];
/* Group contains a faulted vdev. */
if (vdev_draid_faulted(cvd, physical_offset))
return (B_TRUE);
/*
* Always check groups with active distributed spares
* because any vdev failure in the pool will affect them.
*/
if (vdev_draid_find_spare(cvd) != NULL)
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Determine if the txg is missing. Used by healing resilver.
*/
static boolean_t
vdev_draid_group_missing(vdev_t *vd, uint64_t offset, uint64_t txg,
uint64_t size)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
ASSERT3U(vdev_draid_get_astart(vd, offset), ==, offset);
uint64_t groupstart, perm;
uint64_t physical_offset = vdev_draid_logical_to_physical(vd,
offset, &perm, &groupstart);
uint8_t *base;
uint64_t iter;
vdev_draid_get_perm(vdc, perm, &base, &iter);
for (uint64_t i = 0; i < vdc->vdc_groupwidth; i++) {
uint64_t c = (groupstart + i) % vdc->vdc_ndisks;
uint64_t cid = vdev_draid_permute_id(vdc, base, iter, c);
vdev_t *cvd = vd->vdev_child[cid];
/* Transaction group is known to be partially replicated. */
if (vdev_draid_partial(cvd, physical_offset, txg, size))
return (B_TRUE);
/*
* Always check groups with active distributed spares
* because any vdev failure in the pool will affect them.
*/
if (vdev_draid_find_spare(cvd) != NULL)
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Find the smallest child asize and largest sector size to calculate the
* available capacity. Distributed spares are ignored since their capacity
* is also based of the minimum child size in the top-level dRAID.
*/
static void
vdev_draid_calculate_asize(vdev_t *vd, uint64_t *asizep, uint64_t *max_asizep,
uint64_t *logical_ashiftp, uint64_t *physical_ashiftp)
{
uint64_t logical_ashift = 0, physical_ashift = 0;
uint64_t asize = 0, max_asize = 0;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (cvd->vdev_ops == &vdev_draid_spare_ops)
continue;
asize = MIN(asize - 1, cvd->vdev_asize - 1) + 1;
max_asize = MIN(max_asize - 1, cvd->vdev_max_asize - 1) + 1;
logical_ashift = MAX(logical_ashift, cvd->vdev_ashift);
}
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (cvd->vdev_ops == &vdev_draid_spare_ops)
continue;
physical_ashift = vdev_best_ashift(logical_ashift,
physical_ashift, cvd->vdev_physical_ashift);
}
*asizep = asize;
*max_asizep = max_asize;
*logical_ashiftp = logical_ashift;
*physical_ashiftp = physical_ashift;
}
/*
* Open spare vdevs.
*/
static boolean_t
vdev_draid_open_spares(vdev_t *vd)
{
return (vd->vdev_ops == &vdev_draid_spare_ops ||
vd->vdev_ops == &vdev_replacing_ops ||
vd->vdev_ops == &vdev_spare_ops);
}
/*
* Open all children, excluding spares.
*/
static boolean_t
vdev_draid_open_children(vdev_t *vd)
{
return (!vdev_draid_open_spares(vd));
}
/*
* Open a top-level dRAID vdev.
*/
static int
vdev_draid_open(vdev_t *vd, uint64_t *asize, uint64_t *max_asize,
uint64_t *logical_ashift, uint64_t *physical_ashift)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
uint64_t nparity = vdc->vdc_nparity;
int open_errors = 0;
if (nparity > VDEV_DRAID_MAXPARITY ||
vd->vdev_children < nparity + 1) {
vd->vdev_stat.vs_aux = VDEV_AUX_BAD_LABEL;
return (SET_ERROR(EINVAL));
}
/*
* First open the normal children then the distributed spares. This
* ordering is important to ensure the distributed spares calculate
* the correct psize in the event that the dRAID vdevs were expanded.
*/
vdev_open_children_subset(vd, vdev_draid_open_children);
vdev_open_children_subset(vd, vdev_draid_open_spares);
/* Verify enough of the children are available to continue. */
for (int c = 0; c < vd->vdev_children; c++) {
if (vd->vdev_child[c]->vdev_open_error != 0) {
if ((++open_errors) > nparity) {
vd->vdev_stat.vs_aux = VDEV_AUX_NO_REPLICAS;
return (SET_ERROR(ENXIO));
}
}
}
/*
* Allocatable capacity is the sum of the space on all children less
* the number of distributed spares rounded down to last full row
* and then to the last full group. An additional 32MB of scratch
* space is reserved at the end of each child for use by the dRAID
* expansion feature.
*/
uint64_t child_asize, child_max_asize;
vdev_draid_calculate_asize(vd, &child_asize, &child_max_asize,
logical_ashift, physical_ashift);
/*
* Should be unreachable since the minimum child size is 64MB, but
* we want to make sure an underflow absolutely cannot occur here.
*/
if (child_asize < VDEV_DRAID_REFLOW_RESERVE ||
child_max_asize < VDEV_DRAID_REFLOW_RESERVE) {
return (SET_ERROR(ENXIO));
}
child_asize = ((child_asize - VDEV_DRAID_REFLOW_RESERVE) /
VDEV_DRAID_ROWHEIGHT) * VDEV_DRAID_ROWHEIGHT;
child_max_asize = ((child_max_asize - VDEV_DRAID_REFLOW_RESERVE) /
VDEV_DRAID_ROWHEIGHT) * VDEV_DRAID_ROWHEIGHT;
*asize = (((child_asize * vdc->vdc_ndisks) / vdc->vdc_groupsz) *
vdc->vdc_groupsz);
*max_asize = (((child_max_asize * vdc->vdc_ndisks) / vdc->vdc_groupsz) *
vdc->vdc_groupsz);
return (0);
}
/*
* Close a top-level dRAID vdev.
*/
static void
vdev_draid_close(vdev_t *vd)
{
for (int c = 0; c < vd->vdev_children; c++) {
if (vd->vdev_child[c] != NULL)
vdev_close(vd->vdev_child[c]);
}
}
/*
* Return the maximum asize for a rebuild zio in the provided range
* given the following constraints. A dRAID chunks may not:
*
* - Exceed the maximum allowed block size (SPA_MAXBLOCKSIZE), or
* - Span dRAID redundancy groups.
*/
static uint64_t
vdev_draid_rebuild_asize(vdev_t *vd, uint64_t start, uint64_t asize,
uint64_t max_segment)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
uint64_t ashift = vd->vdev_ashift;
uint64_t ndata = vdc->vdc_ndata;
uint64_t psize = MIN(P2ROUNDUP(max_segment * ndata, 1 << ashift),
SPA_MAXBLOCKSIZE);
ASSERT3U(vdev_draid_get_astart(vd, start), ==, start);
ASSERT3U(asize % (vdc->vdc_groupwidth << ashift), ==, 0);
/* Chunks must evenly span all data columns in the group. */
psize = (((psize >> ashift) / ndata) * ndata) << ashift;
uint64_t chunk_size = MIN(asize, vdev_psize_to_asize(vd, psize));
/* Reduce the chunk size to the group space remaining. */
uint64_t group = vdev_draid_offset_to_group(vd, start);
uint64_t left = vdev_draid_group_to_offset(vd, group + 1) - start;
chunk_size = MIN(chunk_size, left);
ASSERT3U(chunk_size % (vdc->vdc_groupwidth << ashift), ==, 0);
ASSERT3U(vdev_draid_offset_to_group(vd, start), ==,
vdev_draid_offset_to_group(vd, start + chunk_size - 1));
return (chunk_size);
}
/*
* Align the start of the metaslab to the group width and slightly reduce
* its size to a multiple of the group width. Since full stripe writes are
* required by dRAID this space is unallocable. Furthermore, aligning the
* metaslab start is important for vdev initialize and TRIM which both operate
* on metaslab boundaries which vdev_xlate() expects to be aligned.
*/
static void
vdev_draid_metaslab_init(vdev_t *vd, uint64_t *ms_start, uint64_t *ms_size)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
uint64_t sz = vdc->vdc_groupwidth << vd->vdev_ashift;
uint64_t astart = vdev_draid_get_astart(vd, *ms_start);
uint64_t asize = ((*ms_size - (astart - *ms_start)) / sz) * sz;
*ms_start = astart;
*ms_size = asize;
ASSERT0(*ms_start % sz);
ASSERT0(*ms_size % sz);
}
/*
* Add virtual dRAID spares to the list of valid spares. In order to accomplish
* this the existing array must be freed and reallocated with the additional
* entries.
*/
int
vdev_draid_spare_create(nvlist_t *nvroot, vdev_t *vd, uint64_t *ndraidp,
uint64_t next_vdev_id)
{
uint64_t draid_nspares = 0;
uint64_t ndraid = 0;
int error;
for (uint64_t i = 0; i < vd->vdev_children; i++) {
vdev_t *cvd = vd->vdev_child[i];
if (cvd->vdev_ops == &vdev_draid_ops) {
vdev_draid_config_t *vdc = cvd->vdev_tsd;
draid_nspares += vdc->vdc_nspares;
ndraid++;
}
}
if (draid_nspares == 0) {
*ndraidp = ndraid;
return (0);
}
nvlist_t **old_spares, **new_spares;
uint_t old_nspares;
error = nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&old_spares, &old_nspares);
if (error)
old_nspares = 0;
/* Allocate memory and copy of the existing spares. */
new_spares = kmem_alloc(sizeof (nvlist_t *) *
(draid_nspares + old_nspares), KM_SLEEP);
for (uint_t i = 0; i < old_nspares; i++)
new_spares[i] = fnvlist_dup(old_spares[i]);
/* Add new distributed spares to ZPOOL_CONFIG_SPARES. */
uint64_t n = old_nspares;
for (uint64_t vdev_id = 0; vdev_id < vd->vdev_children; vdev_id++) {
vdev_t *cvd = vd->vdev_child[vdev_id];
char path[64];
if (cvd->vdev_ops != &vdev_draid_ops)
continue;
vdev_draid_config_t *vdc = cvd->vdev_tsd;
uint64_t nspares = vdc->vdc_nspares;
uint64_t nparity = vdc->vdc_nparity;
for (uint64_t spare_id = 0; spare_id < nspares; spare_id++) {
memset(path, 0, sizeof (path));
(void) snprintf(path, sizeof (path) - 1,
"%s%llu-%llu-%llu", VDEV_TYPE_DRAID,
(u_longlong_t)nparity,
(u_longlong_t)next_vdev_id + vdev_id,
(u_longlong_t)spare_id);
nvlist_t *spare = fnvlist_alloc();
fnvlist_add_string(spare, ZPOOL_CONFIG_PATH, path);
fnvlist_add_string(spare, ZPOOL_CONFIG_TYPE,
VDEV_TYPE_DRAID_SPARE);
fnvlist_add_uint64(spare, ZPOOL_CONFIG_TOP_GUID,
cvd->vdev_guid);
fnvlist_add_uint64(spare, ZPOOL_CONFIG_SPARE_ID,
spare_id);
fnvlist_add_uint64(spare, ZPOOL_CONFIG_IS_LOG, 0);
fnvlist_add_uint64(spare, ZPOOL_CONFIG_IS_SPARE, 1);
fnvlist_add_uint64(spare, ZPOOL_CONFIG_WHOLE_DISK, 1);
fnvlist_add_uint64(spare, ZPOOL_CONFIG_ASHIFT,
cvd->vdev_ashift);
new_spares[n] = spare;
n++;
}
}
if (n > 0) {
(void) nvlist_remove_all(nvroot, ZPOOL_CONFIG_SPARES);
fnvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
(const nvlist_t **)new_spares, n);
}
for (int i = 0; i < n; i++)
nvlist_free(new_spares[i]);
kmem_free(new_spares, sizeof (*new_spares) * n);
*ndraidp = ndraid;
return (0);
}
/*
* Determine if any portion of the provided block resides on a child vdev
* with a dirty DTL and therefore needs to be resilvered.
*/
static boolean_t
vdev_draid_need_resilver(vdev_t *vd, const dva_t *dva, size_t psize,
uint64_t phys_birth)
{
uint64_t offset = DVA_GET_OFFSET(dva);
- uint64_t asize = vdev_draid_asize(vd, psize);
+ uint64_t asize = vdev_draid_asize(vd, psize, 0);
if (phys_birth == TXG_UNKNOWN) {
/*
* Sequential resilver. There is no meaningful phys_birth
* for this block, we can only determine if block resides
* in a degraded group in which case it must be resilvered.
*/
ASSERT3U(vdev_draid_offset_to_group(vd, offset), ==,
vdev_draid_offset_to_group(vd, offset + asize - 1));
return (vdev_draid_group_degraded(vd, offset));
} else {
/*
* Healing resilver. TXGs not in DTL_PARTIAL are intact,
* as are blocks in non-degraded groups.
*/
if (!vdev_dtl_contains(vd, DTL_PARTIAL, phys_birth, 1))
return (B_FALSE);
if (vdev_draid_group_missing(vd, offset, phys_birth, 1))
return (B_TRUE);
/* The block may span groups in which case check both. */
if (vdev_draid_offset_to_group(vd, offset) !=
vdev_draid_offset_to_group(vd, offset + asize - 1)) {
if (vdev_draid_group_missing(vd,
offset + asize, phys_birth, 1))
return (B_TRUE);
}
return (B_FALSE);
}
}
static boolean_t
vdev_draid_rebuilding(vdev_t *vd)
{
if (vd->vdev_ops->vdev_op_leaf && vd->vdev_rebuild_txg)
return (B_TRUE);
for (int i = 0; i < vd->vdev_children; i++) {
if (vdev_draid_rebuilding(vd->vdev_child[i])) {
return (B_TRUE);
}
}
return (B_FALSE);
}
static void
vdev_draid_io_verify(vdev_t *vd, raidz_row_t *rr, int col)
{
#ifdef ZFS_DEBUG
range_seg64_t logical_rs, physical_rs, remain_rs;
logical_rs.rs_start = rr->rr_offset;
logical_rs.rs_end = logical_rs.rs_start +
- vdev_draid_asize(vd, rr->rr_size);
+ vdev_draid_asize(vd, rr->rr_size, 0);
raidz_col_t *rc = &rr->rr_col[col];
vdev_t *cvd = vd->vdev_child[rc->rc_devidx];
vdev_xlate(cvd, &logical_rs, &physical_rs, &remain_rs);
ASSERT(vdev_xlate_is_empty(&remain_rs));
ASSERT3U(rc->rc_offset, ==, physical_rs.rs_start);
ASSERT3U(rc->rc_offset, <, physical_rs.rs_end);
ASSERT3U(rc->rc_offset + rc->rc_size, ==, physical_rs.rs_end);
#endif
}
/*
* For write operations:
* 1. Generate the parity data
* 2. Create child zio write operations to each column's vdev, for both
* data and parity. A gang ABD is allocated by vdev_draid_map_alloc()
* if a skip sector needs to be added to a column.
*/
static void
vdev_draid_io_start_write(zio_t *zio, raidz_row_t *rr)
{
vdev_t *vd = zio->io_vd;
raidz_map_t *rm = zio->io_vsd;
vdev_raidz_generate_parity_row(rm, rr);
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
/*
* Empty columns are zero filled and included in the parity
* calculation and therefore must be written.
*/
ASSERT3U(rc->rc_size, !=, 0);
/* Verify physical to logical translation */
vdev_draid_io_verify(vd, rr, c);
zio_nowait(zio_vdev_child_io(zio, NULL,
vd->vdev_child[rc->rc_devidx], rc->rc_offset,
rc->rc_abd, rc->rc_size, zio->io_type, zio->io_priority,
0, vdev_raidz_child_done, rc));
}
}
/*
* For read operations:
* 1. The vdev_draid_map_alloc() function will create a minimal raidz
* mapping for the read based on the zio->io_flags. There are two
* possible mappings either 1) a normal read, or 2) a scrub/resilver.
* 2. Create the zio read operations. This will include all parity
* columns and skip sectors for a scrub/resilver.
*/
static void
vdev_draid_io_start_read(zio_t *zio, raidz_row_t *rr)
{
vdev_t *vd = zio->io_vd;
/* Sequential rebuild must do IO at redundancy group boundary. */
IMPLY(zio->io_priority == ZIO_PRIORITY_REBUILD, rr->rr_nempty == 0);
/*
* Iterate over the columns in reverse order so that we hit the parity
* last. Any errors along the way will force us to read the parity.
* For scrub/resilver IOs which verify skip sectors, a gang ABD will
* have been allocated to store them and rc->rc_size is increased.
*/
for (int c = rr->rr_cols - 1; c >= 0; c--) {
raidz_col_t *rc = &rr->rr_col[c];
vdev_t *cvd = vd->vdev_child[rc->rc_devidx];
if (!vdev_draid_readable(cvd, rc->rc_offset)) {
if (c >= rr->rr_firstdatacol)
rr->rr_missingdata++;
else
rr->rr_missingparity++;
rc->rc_error = SET_ERROR(ENXIO);
rc->rc_tried = 1;
rc->rc_skipped = 1;
continue;
}
if (vdev_draid_missing(cvd, rc->rc_offset, zio->io_txg, 1)) {
if (c >= rr->rr_firstdatacol)
rr->rr_missingdata++;
else
rr->rr_missingparity++;
rc->rc_error = SET_ERROR(ESTALE);
rc->rc_skipped = 1;
continue;
}
/*
* Empty columns may be read during vdev_draid_io_done().
* Only skip them after the readable and missing checks
* verify they are available.
*/
if (rc->rc_size == 0) {
rc->rc_skipped = 1;
continue;
}
if (zio->io_flags & ZIO_FLAG_RESILVER) {
vdev_t *svd;
/*
* Sequential rebuilds need to always consider the data
* on the child being rebuilt to be stale. This is
* important when all columns are available to aid
* known reconstruction in identifing which columns
* contain incorrect data.
*
* Furthermore, all repairs need to be constrained to
* the devices being rebuilt because without a checksum
* we cannot verify the data is actually correct and
* performing an incorrect repair could result in
* locking in damage and making the data unrecoverable.
*/
if (zio->io_priority == ZIO_PRIORITY_REBUILD) {
if (vdev_draid_rebuilding(cvd)) {
if (c >= rr->rr_firstdatacol)
rr->rr_missingdata++;
else
rr->rr_missingparity++;
rc->rc_error = SET_ERROR(ESTALE);
rc->rc_skipped = 1;
rc->rc_allow_repair = 1;
continue;
} else {
rc->rc_allow_repair = 0;
}
} else {
rc->rc_allow_repair = 1;
}
/*
* If this child is a distributed spare then the
* offset might reside on the vdev being replaced.
* In which case this data must be written to the
* new device. Failure to do so would result in
* checksum errors when the old device is detached
* and the pool is scrubbed.
*/
if ((svd = vdev_draid_find_spare(cvd)) != NULL) {
svd = vdev_draid_spare_get_child(svd,
rc->rc_offset);
if (svd && (svd->vdev_ops == &vdev_spare_ops ||
svd->vdev_ops == &vdev_replacing_ops)) {
rc->rc_force_repair = 1;
if (vdev_draid_rebuilding(svd))
rc->rc_allow_repair = 1;
}
}
/*
* Always issue a repair IO to this child when its
* a spare or replacing vdev with an active rebuild.
*/
if ((cvd->vdev_ops == &vdev_spare_ops ||
cvd->vdev_ops == &vdev_replacing_ops) &&
vdev_draid_rebuilding(cvd)) {
rc->rc_force_repair = 1;
rc->rc_allow_repair = 1;
}
}
}
/*
* Either a parity or data column is missing this means a repair
* may be attempted by vdev_draid_io_done(). Expand the raid map
* to read in empty columns which are needed along with the parity
* during reconstruction.
*/
if ((rr->rr_missingdata > 0 || rr->rr_missingparity > 0) &&
rr->rr_nempty > 0 && rr->rr_abd_empty == NULL) {
vdev_draid_map_alloc_empty(zio, rr);
}
for (int c = rr->rr_cols - 1; c >= 0; c--) {
raidz_col_t *rc = &rr->rr_col[c];
vdev_t *cvd = vd->vdev_child[rc->rc_devidx];
if (rc->rc_error || rc->rc_size == 0)
continue;
if (c >= rr->rr_firstdatacol || rr->rr_missingdata > 0 ||
(zio->io_flags & (ZIO_FLAG_SCRUB | ZIO_FLAG_RESILVER))) {
zio_nowait(zio_vdev_child_io(zio, NULL, cvd,
rc->rc_offset, rc->rc_abd, rc->rc_size,
zio->io_type, zio->io_priority, 0,
vdev_raidz_child_done, rc));
}
}
}
/*
* Start an IO operation to a dRAID vdev.
*/
static void
vdev_draid_io_start(zio_t *zio)
{
vdev_t *vd __maybe_unused = zio->io_vd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
ASSERT3U(zio->io_offset, ==, vdev_draid_get_astart(vd, zio->io_offset));
raidz_map_t *rm = vdev_draid_map_alloc(zio);
zio->io_vsd = rm;
zio->io_vsd_ops = &vdev_raidz_vsd_ops;
if (zio->io_type == ZIO_TYPE_WRITE) {
for (int i = 0; i < rm->rm_nrows; i++) {
vdev_draid_io_start_write(zio, rm->rm_row[i]);
}
} else {
ASSERT(zio->io_type == ZIO_TYPE_READ);
for (int i = 0; i < rm->rm_nrows; i++) {
vdev_draid_io_start_read(zio, rm->rm_row[i]);
}
}
zio_execute(zio);
}
/*
* Complete an IO operation on a dRAID vdev. The raidz logic can be applied
* to dRAID since the layout is fully described by the raidz_map_t.
*/
static void
vdev_draid_io_done(zio_t *zio)
{
vdev_raidz_io_done(zio);
}
static void
vdev_draid_state_change(vdev_t *vd, int faulted, int degraded)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
ASSERT(vd->vdev_ops == &vdev_draid_ops);
if (faulted > vdc->vdc_nparity)
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_NO_REPLICAS);
else if (degraded + faulted != 0)
vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, VDEV_AUX_NONE);
else
vdev_set_state(vd, B_FALSE, VDEV_STATE_HEALTHY, VDEV_AUX_NONE);
}
static void
vdev_draid_xlate(vdev_t *cvd, const range_seg64_t *logical_rs,
range_seg64_t *physical_rs, range_seg64_t *remain_rs)
{
vdev_t *raidvd = cvd->vdev_parent;
ASSERT(raidvd->vdev_ops == &vdev_draid_ops);
vdev_draid_config_t *vdc = raidvd->vdev_tsd;
uint64_t ashift = raidvd->vdev_top->vdev_ashift;
/* Make sure the offsets are block-aligned */
ASSERT0(logical_rs->rs_start % (1 << ashift));
ASSERT0(logical_rs->rs_end % (1 << ashift));
uint64_t logical_start = logical_rs->rs_start;
uint64_t logical_end = logical_rs->rs_end;
/*
* Unaligned ranges must be skipped. All metaslabs are correctly
* aligned so this should not happen, but this case is handled in
* case it's needed by future callers.
*/
uint64_t astart = vdev_draid_get_astart(raidvd, logical_start);
if (astart != logical_start) {
physical_rs->rs_start = logical_start;
physical_rs->rs_end = logical_start;
remain_rs->rs_start = MIN(astart, logical_end);
remain_rs->rs_end = logical_end;
return;
}
/*
* Unlike with mirrors and raidz a dRAID logical range can map
* to multiple non-contiguous physical ranges. This is handled by
* limiting the size of the logical range to a single group and
* setting the remain argument such that it describes the remaining
* unmapped logical range. This is stricter than absolutely
* necessary but helps simplify the logic below.
*/
uint64_t group = vdev_draid_offset_to_group(raidvd, logical_start);
uint64_t nextstart = vdev_draid_group_to_offset(raidvd, group + 1);
if (logical_end > nextstart)
logical_end = nextstart;
/* Find the starting offset for each vdev in the group */
uint64_t perm, groupstart;
uint64_t start = vdev_draid_logical_to_physical(raidvd,
logical_start, &perm, &groupstart);
uint64_t end = start;
uint8_t *base;
uint64_t iter, id;
vdev_draid_get_perm(vdc, perm, &base, &iter);
/*
* Check if the passed child falls within the group. If it does
* update the start and end to reflect the physical range.
* Otherwise, leave them unmodified which will result in an empty
* (zero-length) physical range being returned.
*/
for (uint64_t i = 0; i < vdc->vdc_groupwidth; i++) {
uint64_t c = (groupstart + i) % vdc->vdc_ndisks;
if (c == 0 && i != 0) {
/* the group wrapped, increment the start */
start += VDEV_DRAID_ROWHEIGHT;
end = start;
}
id = vdev_draid_permute_id(vdc, base, iter, c);
if (id == cvd->vdev_id) {
uint64_t b_size = (logical_end >> ashift) -
(logical_start >> ashift);
ASSERT3U(b_size, >, 0);
end = start + ((((b_size - 1) /
vdc->vdc_groupwidth) + 1) << ashift);
break;
}
}
physical_rs->rs_start = start;
physical_rs->rs_end = end;
/*
* Only top-level vdevs are allowed to set remain_rs because
* when .vdev_op_xlate() is called for their children the full
* logical range is not provided by vdev_xlate().
*/
remain_rs->rs_start = logical_end;
remain_rs->rs_end = logical_rs->rs_end;
ASSERT3U(physical_rs->rs_start, <=, logical_start);
ASSERT3U(physical_rs->rs_end - physical_rs->rs_start, <=,
logical_end - logical_start);
}
/*
* Add dRAID specific fields to the config nvlist.
*/
static void
vdev_draid_config_generate(vdev_t *vd, nvlist_t *nv)
{
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_ops);
vdev_draid_config_t *vdc = vd->vdev_tsd;
fnvlist_add_uint64(nv, ZPOOL_CONFIG_NPARITY, vdc->vdc_nparity);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_DRAID_NDATA, vdc->vdc_ndata);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_DRAID_NSPARES, vdc->vdc_nspares);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_DRAID_NGROUPS, vdc->vdc_ngroups);
}
/*
* Initialize private dRAID specific fields from the nvlist.
*/
static int
vdev_draid_init(spa_t *spa, nvlist_t *nv, void **tsd)
{
(void) spa;
uint64_t ndata, nparity, nspares, ngroups;
int error;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DRAID_NDATA, &ndata))
return (SET_ERROR(EINVAL));
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY, &nparity) ||
nparity == 0 || nparity > VDEV_DRAID_MAXPARITY) {
return (SET_ERROR(EINVAL));
}
uint_t children;
nvlist_t **child;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0 || children == 0 ||
children > VDEV_DRAID_MAX_CHILDREN) {
return (SET_ERROR(EINVAL));
}
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DRAID_NSPARES, &nspares) ||
nspares > 100 || nspares > (children - (ndata + nparity))) {
return (SET_ERROR(EINVAL));
}
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DRAID_NGROUPS, &ngroups) ||
ngroups == 0 || ngroups > VDEV_DRAID_MAX_CHILDREN) {
return (SET_ERROR(EINVAL));
}
/*
* Validate the minimum number of children exist per group for the
* specified parity level (draid1 >= 2, draid2 >= 3, draid3 >= 4).
*/
if (children < (ndata + nparity + nspares))
return (SET_ERROR(EINVAL));
/*
* Create the dRAID configuration using the pool nvlist configuration
* and the fixed mapping for the correct number of children.
*/
vdev_draid_config_t *vdc;
const draid_map_t *map;
error = vdev_draid_lookup_map(children, &map);
if (error)
return (SET_ERROR(EINVAL));
vdc = kmem_zalloc(sizeof (*vdc), KM_SLEEP);
vdc->vdc_ndata = ndata;
vdc->vdc_nparity = nparity;
vdc->vdc_nspares = nspares;
vdc->vdc_children = children;
vdc->vdc_ngroups = ngroups;
vdc->vdc_nperms = map->dm_nperms;
error = vdev_draid_generate_perms(map, &vdc->vdc_perms);
if (error) {
kmem_free(vdc, sizeof (*vdc));
return (SET_ERROR(EINVAL));
}
/*
* Derived constants.
*/
vdc->vdc_groupwidth = vdc->vdc_ndata + vdc->vdc_nparity;
vdc->vdc_ndisks = vdc->vdc_children - vdc->vdc_nspares;
vdc->vdc_groupsz = vdc->vdc_groupwidth * VDEV_DRAID_ROWHEIGHT;
vdc->vdc_devslicesz = (vdc->vdc_groupsz * vdc->vdc_ngroups) /
vdc->vdc_ndisks;
ASSERT3U(vdc->vdc_groupwidth, >=, 2);
ASSERT3U(vdc->vdc_groupwidth, <=, vdc->vdc_ndisks);
ASSERT3U(vdc->vdc_groupsz, >=, 2 * VDEV_DRAID_ROWHEIGHT);
ASSERT3U(vdc->vdc_devslicesz, >=, VDEV_DRAID_ROWHEIGHT);
ASSERT3U(vdc->vdc_devslicesz % VDEV_DRAID_ROWHEIGHT, ==, 0);
ASSERT3U((vdc->vdc_groupwidth * vdc->vdc_ngroups) %
vdc->vdc_ndisks, ==, 0);
*tsd = vdc;
return (0);
}
static void
vdev_draid_fini(vdev_t *vd)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
vmem_free(vdc->vdc_perms, sizeof (uint8_t) *
vdc->vdc_children * vdc->vdc_nperms);
kmem_free(vdc, sizeof (*vdc));
}
static uint64_t
vdev_draid_nparity(vdev_t *vd)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
return (vdc->vdc_nparity);
}
static uint64_t
vdev_draid_ndisks(vdev_t *vd)
{
vdev_draid_config_t *vdc = vd->vdev_tsd;
return (vdc->vdc_ndisks);
}
vdev_ops_t vdev_draid_ops = {
.vdev_op_init = vdev_draid_init,
.vdev_op_fini = vdev_draid_fini,
.vdev_op_open = vdev_draid_open,
.vdev_op_close = vdev_draid_close,
.vdev_op_asize = vdev_draid_asize,
.vdev_op_min_asize = vdev_draid_min_asize,
.vdev_op_min_alloc = vdev_draid_min_alloc,
.vdev_op_io_start = vdev_draid_io_start,
.vdev_op_io_done = vdev_draid_io_done,
.vdev_op_state_change = vdev_draid_state_change,
.vdev_op_need_resilver = vdev_draid_need_resilver,
.vdev_op_hold = NULL,
.vdev_op_rele = NULL,
.vdev_op_remap = NULL,
.vdev_op_xlate = vdev_draid_xlate,
.vdev_op_rebuild_asize = vdev_draid_rebuild_asize,
.vdev_op_metaslab_init = vdev_draid_metaslab_init,
.vdev_op_config_generate = vdev_draid_config_generate,
.vdev_op_nparity = vdev_draid_nparity,
.vdev_op_ndisks = vdev_draid_ndisks,
.vdev_op_type = VDEV_TYPE_DRAID,
.vdev_op_leaf = B_FALSE,
};
/*
* A dRAID distributed spare is a virtual leaf vdev which is included in the
* parent dRAID configuration. The last N columns of the dRAID permutation
* table are used to determine on which dRAID children a specific offset
* should be written. These spare leaf vdevs can only be used to replace
* faulted children in the same dRAID configuration.
*/
/*
* Distributed spare state. All fields are set when the distributed spare is
* first opened and are immutable.
*/
typedef struct {
vdev_t *vds_draid_vdev; /* top-level parent dRAID vdev */
uint64_t vds_top_guid; /* top-level parent dRAID guid */
uint64_t vds_spare_id; /* spare id (0 - vdc->vdc_nspares-1) */
} vdev_draid_spare_t;
/*
* Returns the parent dRAID vdev to which the distributed spare belongs.
* This may be safely called even when the vdev is not open.
*/
vdev_t *
vdev_draid_spare_get_parent(vdev_t *vd)
{
vdev_draid_spare_t *vds = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_spare_ops);
if (vds->vds_draid_vdev != NULL)
return (vds->vds_draid_vdev);
return (vdev_lookup_by_guid(vd->vdev_spa->spa_root_vdev,
vds->vds_top_guid));
}
/*
* A dRAID space is active when it's the child of a vdev using the
* vdev_spare_ops, vdev_replacing_ops or vdev_draid_ops.
*/
static boolean_t
vdev_draid_spare_is_active(vdev_t *vd)
{
vdev_t *pvd = vd->vdev_parent;
if (pvd != NULL && (pvd->vdev_ops == &vdev_spare_ops ||
pvd->vdev_ops == &vdev_replacing_ops ||
pvd->vdev_ops == &vdev_draid_ops)) {
return (B_TRUE);
} else {
return (B_FALSE);
}
}
/*
* Given a dRAID distribute spare vdev, returns the physical child vdev
* on which the provided offset resides. This may involve recursing through
* multiple layers of distributed spares. Note that offset is relative to
* this vdev.
*/
vdev_t *
vdev_draid_spare_get_child(vdev_t *vd, uint64_t physical_offset)
{
vdev_draid_spare_t *vds = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_spare_ops);
/* The vdev is closed */
if (vds->vds_draid_vdev == NULL)
return (NULL);
vdev_t *tvd = vds->vds_draid_vdev;
vdev_draid_config_t *vdc = tvd->vdev_tsd;
ASSERT3P(tvd->vdev_ops, ==, &vdev_draid_ops);
ASSERT3U(vds->vds_spare_id, <, vdc->vdc_nspares);
uint8_t *base;
uint64_t iter;
uint64_t perm = physical_offset / vdc->vdc_devslicesz;
vdev_draid_get_perm(vdc, perm, &base, &iter);
uint64_t cid = vdev_draid_permute_id(vdc, base, iter,
(tvd->vdev_children - 1) - vds->vds_spare_id);
vdev_t *cvd = tvd->vdev_child[cid];
if (cvd->vdev_ops == &vdev_draid_spare_ops)
return (vdev_draid_spare_get_child(cvd, physical_offset));
return (cvd);
}
static void
vdev_draid_spare_close(vdev_t *vd)
{
vdev_draid_spare_t *vds = vd->vdev_tsd;
vds->vds_draid_vdev = NULL;
}
/*
* Opening a dRAID spare device is done by looking up the associated dRAID
* top-level vdev guid from the spare configuration.
*/
static int
vdev_draid_spare_open(vdev_t *vd, uint64_t *psize, uint64_t *max_psize,
uint64_t *logical_ashift, uint64_t *physical_ashift)
{
vdev_draid_spare_t *vds = vd->vdev_tsd;
vdev_t *rvd = vd->vdev_spa->spa_root_vdev;
uint64_t asize, max_asize;
vdev_t *tvd = vdev_lookup_by_guid(rvd, vds->vds_top_guid);
if (tvd == NULL) {
/*
* When spa_vdev_add() is labeling new spares the
* associated dRAID is not attached to the root vdev
* nor does this spare have a parent. Simulate a valid
* device in order to allow the label to be initialized
* and the distributed spare added to the configuration.
*/
if (vd->vdev_parent == NULL) {
*psize = *max_psize = SPA_MINDEVSIZE;
*logical_ashift = *physical_ashift = ASHIFT_MIN;
return (0);
}
return (SET_ERROR(EINVAL));
}
vdev_draid_config_t *vdc = tvd->vdev_tsd;
if (tvd->vdev_ops != &vdev_draid_ops || vdc == NULL)
return (SET_ERROR(EINVAL));
if (vds->vds_spare_id >= vdc->vdc_nspares)
return (SET_ERROR(EINVAL));
/*
* Neither tvd->vdev_asize or tvd->vdev_max_asize can be used here
* because the caller may be vdev_draid_open() in which case the
* values are stale as they haven't yet been updated by vdev_open().
* To avoid this always recalculate the dRAID asize and max_asize.
*/
vdev_draid_calculate_asize(tvd, &asize, &max_asize,
logical_ashift, physical_ashift);
*psize = asize + VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE;
*max_psize = max_asize + VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE;
vds->vds_draid_vdev = tvd;
return (0);
}
/*
* Completed distributed spare IO. Store the result in the parent zio
* as if it had performed the operation itself. Only the first error is
* preserved if there are multiple errors.
*/
static void
vdev_draid_spare_child_done(zio_t *zio)
{
zio_t *pio = zio->io_private;
/*
* IOs are issued to non-writable vdevs in order to keep their
* DTLs accurate. However, we don't want to propagate the
* error in to the distributed spare's DTL. When resilvering
* vdev_draid_need_resilver() will consult the relevant DTL
* to determine if the data is missing and must be repaired.
*/
if (!vdev_writeable(zio->io_vd))
return;
if (pio->io_error == 0)
pio->io_error = zio->io_error;
}
/*
* Returns a valid label nvlist for the distributed spare vdev. This is
* used to bypass the IO pipeline to avoid the complexity of constructing
* a complete label with valid checksum to return when read.
*/
nvlist_t *
vdev_draid_read_config_spare(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
spa_aux_vdev_t *sav = &spa->spa_spares;
uint64_t guid = vd->vdev_guid;
nvlist_t *nv = fnvlist_alloc();
fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_SPARE, 1);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_CREATE_TXG, vd->vdev_crtxg);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_VERSION, spa_version(spa));
fnvlist_add_string(nv, ZPOOL_CONFIG_POOL_NAME, spa_name(spa));
fnvlist_add_uint64(nv, ZPOOL_CONFIG_POOL_GUID, spa_guid(spa));
fnvlist_add_uint64(nv, ZPOOL_CONFIG_POOL_TXG, spa->spa_config_txg);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_TOP_GUID, vd->vdev_top->vdev_guid);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_POOL_STATE,
vdev_draid_spare_is_active(vd) ?
POOL_STATE_ACTIVE : POOL_STATE_SPARE);
/* Set the vdev guid based on the vdev list in sav_count. */
for (int i = 0; i < sav->sav_count; i++) {
if (sav->sav_vdevs[i]->vdev_ops == &vdev_draid_spare_ops &&
strcmp(sav->sav_vdevs[i]->vdev_path, vd->vdev_path) == 0) {
guid = sav->sav_vdevs[i]->vdev_guid;
break;
}
}
fnvlist_add_uint64(nv, ZPOOL_CONFIG_GUID, guid);
return (nv);
}
/*
* Handle any ioctl requested of the distributed spare. Only flushes
* are supported in which case all children must be flushed.
*/
static int
vdev_draid_spare_ioctl(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
int error = 0;
if (zio->io_cmd == DKIOCFLUSHWRITECACHE) {
for (int c = 0; c < vd->vdev_children; c++) {
zio_nowait(zio_vdev_child_io(zio, NULL,
vd->vdev_child[c], zio->io_offset, zio->io_abd,
zio->io_size, zio->io_type, zio->io_priority, 0,
vdev_draid_spare_child_done, zio));
}
} else {
error = SET_ERROR(ENOTSUP);
}
return (error);
}
/*
* Initiate an IO to the distributed spare. For normal IOs this entails using
* the zio->io_offset and permutation table to calculate which child dRAID vdev
* is responsible for the data. Then passing along the zio to that child to
* perform the actual IO. The label ranges are not stored on disk and require
* some special handling which is described below.
*/
static void
vdev_draid_spare_io_start(zio_t *zio)
{
vdev_t *cvd = NULL, *vd = zio->io_vd;
vdev_draid_spare_t *vds = vd->vdev_tsd;
uint64_t offset = zio->io_offset - VDEV_LABEL_START_SIZE;
/*
* If the vdev is closed, it's likely in the REMOVED or FAULTED state.
* Nothing to be done here but return failure.
*/
if (vds == NULL) {
zio->io_error = ENXIO;
zio_interrupt(zio);
return;
}
switch (zio->io_type) {
case ZIO_TYPE_IOCTL:
zio->io_error = vdev_draid_spare_ioctl(zio);
break;
case ZIO_TYPE_WRITE:
if (VDEV_OFFSET_IS_LABEL(vd, zio->io_offset)) {
/*
* Accept probe IOs and config writers to simulate the
* existence of an on disk label. vdev_label_sync(),
* vdev_uberblock_sync() and vdev_copy_uberblocks()
* skip the distributed spares. This only leaves
* vdev_label_init() which is allowed to succeed to
* avoid adding special cases the function.
*/
if (zio->io_flags & ZIO_FLAG_PROBE ||
zio->io_flags & ZIO_FLAG_CONFIG_WRITER) {
zio->io_error = 0;
} else {
zio->io_error = SET_ERROR(EIO);
}
} else {
cvd = vdev_draid_spare_get_child(vd, offset);
if (cvd == NULL) {
zio->io_error = SET_ERROR(ENXIO);
} else {
zio_nowait(zio_vdev_child_io(zio, NULL, cvd,
offset, zio->io_abd, zio->io_size,
zio->io_type, zio->io_priority, 0,
vdev_draid_spare_child_done, zio));
}
}
break;
case ZIO_TYPE_READ:
if (VDEV_OFFSET_IS_LABEL(vd, zio->io_offset)) {
/*
* Accept probe IOs to simulate the existence of a
* label. vdev_label_read_config() bypasses the
* pipeline to read the label configuration and
* vdev_uberblock_load() skips distributed spares
* when attempting to locate the best uberblock.
*/
if (zio->io_flags & ZIO_FLAG_PROBE) {
zio->io_error = 0;
} else {
zio->io_error = SET_ERROR(EIO);
}
} else {
cvd = vdev_draid_spare_get_child(vd, offset);
if (cvd == NULL || !vdev_readable(cvd)) {
zio->io_error = SET_ERROR(ENXIO);
} else {
zio_nowait(zio_vdev_child_io(zio, NULL, cvd,
offset, zio->io_abd, zio->io_size,
zio->io_type, zio->io_priority, 0,
vdev_draid_spare_child_done, zio));
}
}
break;
case ZIO_TYPE_TRIM:
/* The vdev label ranges are never trimmed */
ASSERT0(VDEV_OFFSET_IS_LABEL(vd, zio->io_offset));
cvd = vdev_draid_spare_get_child(vd, offset);
if (cvd == NULL || !cvd->vdev_has_trim) {
zio->io_error = SET_ERROR(ENXIO);
} else {
zio_nowait(zio_vdev_child_io(zio, NULL, cvd,
offset, zio->io_abd, zio->io_size,
zio->io_type, zio->io_priority, 0,
vdev_draid_spare_child_done, zio));
}
break;
default:
zio->io_error = SET_ERROR(ENOTSUP);
break;
}
zio_execute(zio);
}
static void
vdev_draid_spare_io_done(zio_t *zio)
{
(void) zio;
}
/*
* Lookup the full spare config in spa->spa_spares.sav_config and
* return the top_guid and spare_id for the named spare.
*/
static int
vdev_draid_spare_lookup(spa_t *spa, nvlist_t *nv, uint64_t *top_guidp,
uint64_t *spare_idp)
{
nvlist_t **spares;
uint_t nspares;
int error;
if ((spa->spa_spares.sav_config == NULL) ||
(nvlist_lookup_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, &spares, &nspares) != 0)) {
return (SET_ERROR(ENOENT));
}
const char *spare_name;
error = nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &spare_name);
if (error != 0)
return (SET_ERROR(EINVAL));
for (int i = 0; i < nspares; i++) {
nvlist_t *spare = spares[i];
uint64_t top_guid, spare_id;
const char *type, *path;
/* Skip non-distributed spares */
error = nvlist_lookup_string(spare, ZPOOL_CONFIG_TYPE, &type);
if (error != 0 || strcmp(type, VDEV_TYPE_DRAID_SPARE) != 0)
continue;
/* Skip spares with the wrong name */
error = nvlist_lookup_string(spare, ZPOOL_CONFIG_PATH, &path);
if (error != 0 || strcmp(path, spare_name) != 0)
continue;
/* Found the matching spare */
error = nvlist_lookup_uint64(spare,
ZPOOL_CONFIG_TOP_GUID, &top_guid);
if (error == 0) {
error = nvlist_lookup_uint64(spare,
ZPOOL_CONFIG_SPARE_ID, &spare_id);
}
if (error != 0) {
return (SET_ERROR(EINVAL));
} else {
*top_guidp = top_guid;
*spare_idp = spare_id;
return (0);
}
}
return (SET_ERROR(ENOENT));
}
/*
* Initialize private dRAID spare specific fields from the nvlist.
*/
static int
vdev_draid_spare_init(spa_t *spa, nvlist_t *nv, void **tsd)
{
vdev_draid_spare_t *vds;
uint64_t top_guid = 0;
uint64_t spare_id;
/*
* In the normal case check the list of spares stored in the spa
* to lookup the top_guid and spare_id for provided spare config.
* When creating a new pool or adding vdevs the spare list is not
* yet populated and the values are provided in the passed config.
*/
if (vdev_draid_spare_lookup(spa, nv, &top_guid, &spare_id) != 0) {
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_TOP_GUID,
&top_guid) != 0)
return (SET_ERROR(EINVAL));
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_SPARE_ID,
&spare_id) != 0)
return (SET_ERROR(EINVAL));
}
vds = kmem_alloc(sizeof (vdev_draid_spare_t), KM_SLEEP);
vds->vds_draid_vdev = NULL;
vds->vds_top_guid = top_guid;
vds->vds_spare_id = spare_id;
*tsd = vds;
return (0);
}
static void
vdev_draid_spare_fini(vdev_t *vd)
{
kmem_free(vd->vdev_tsd, sizeof (vdev_draid_spare_t));
}
static void
vdev_draid_spare_config_generate(vdev_t *vd, nvlist_t *nv)
{
vdev_draid_spare_t *vds = vd->vdev_tsd;
ASSERT3P(vd->vdev_ops, ==, &vdev_draid_spare_ops);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_TOP_GUID, vds->vds_top_guid);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_SPARE_ID, vds->vds_spare_id);
}
vdev_ops_t vdev_draid_spare_ops = {
.vdev_op_init = vdev_draid_spare_init,
.vdev_op_fini = vdev_draid_spare_fini,
.vdev_op_open = vdev_draid_spare_open,
.vdev_op_close = vdev_draid_spare_close,
.vdev_op_asize = vdev_default_asize,
.vdev_op_min_asize = vdev_default_min_asize,
.vdev_op_min_alloc = NULL,
.vdev_op_io_start = vdev_draid_spare_io_start,
.vdev_op_io_done = vdev_draid_spare_io_done,
.vdev_op_state_change = NULL,
.vdev_op_need_resilver = NULL,
.vdev_op_hold = NULL,
.vdev_op_rele = NULL,
.vdev_op_remap = NULL,
.vdev_op_xlate = vdev_default_xlate,
.vdev_op_rebuild_asize = NULL,
.vdev_op_metaslab_init = NULL,
.vdev_op_config_generate = vdev_draid_spare_config_generate,
.vdev_op_nparity = NULL,
.vdev_op_ndisks = NULL,
.vdev_op_type = VDEV_TYPE_DRAID_SPARE,
.vdev_op_leaf = B_TRUE,
};
diff --git a/sys/contrib/openzfs/module/zfs/vdev_initialize.c b/sys/contrib/openzfs/module/zfs/vdev_initialize.c
index ffdcef1972c3..5aaef1a69986 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_initialize.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_initialize.c
@@ -1,824 +1,828 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2016, 2019 by Delphix. All rights reserved.
*/
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/txg.h>
#include <sys/vdev_impl.h>
#include <sys/metaslab_impl.h>
#include <sys/dsl_synctask.h>
#include <sys/zap.h>
#include <sys/dmu_tx.h>
#include <sys/vdev_initialize.h>
/*
* Value that is written to disk during initialization.
*/
static uint64_t zfs_initialize_value = 0xdeadbeefdeadbeeeULL;
/* maximum number of I/Os outstanding per leaf vdev */
static const int zfs_initialize_limit = 1;
/* size of initializing writes; default 1MiB, see zfs_remove_max_segment */
static uint64_t zfs_initialize_chunk_size = 1024 * 1024;
static boolean_t
vdev_initialize_should_stop(vdev_t *vd)
{
return (vd->vdev_initialize_exit_wanted || !vdev_writeable(vd) ||
- vd->vdev_detached || vd->vdev_top->vdev_removing);
+ vd->vdev_detached || vd->vdev_top->vdev_removing ||
+ vd->vdev_top->vdev_rz_expanding);
}
static void
vdev_initialize_zap_update_sync(void *arg, dmu_tx_t *tx)
{
/*
* We pass in the guid instead of the vdev_t since the vdev may
* have been freed prior to the sync task being processed. This
* happens when a vdev is detached as we call spa_config_vdev_exit(),
* stop the initializing thread, schedule the sync task, and free
* the vdev. Later when the scheduled sync task is invoked, it would
* find that the vdev has been freed.
*/
uint64_t guid = *(uint64_t *)arg;
uint64_t txg = dmu_tx_get_txg(tx);
kmem_free(arg, sizeof (uint64_t));
vdev_t *vd = spa_lookup_by_guid(tx->tx_pool->dp_spa, guid, B_FALSE);
- if (vd == NULL || vd->vdev_top->vdev_removing || !vdev_is_concrete(vd))
+ if (vd == NULL || vd->vdev_top->vdev_removing ||
+ !vdev_is_concrete(vd) || vd->vdev_top->vdev_rz_expanding)
return;
uint64_t last_offset = vd->vdev_initialize_offset[txg & TXG_MASK];
vd->vdev_initialize_offset[txg & TXG_MASK] = 0;
VERIFY(vd->vdev_leaf_zap != 0);
objset_t *mos = vd->vdev_spa->spa_meta_objset;
if (last_offset > 0) {
vd->vdev_initialize_last_offset = last_offset;
VERIFY0(zap_update(mos, vd->vdev_leaf_zap,
VDEV_LEAF_ZAP_INITIALIZE_LAST_OFFSET,
sizeof (last_offset), 1, &last_offset, tx));
}
if (vd->vdev_initialize_action_time > 0) {
uint64_t val = (uint64_t)vd->vdev_initialize_action_time;
VERIFY0(zap_update(mos, vd->vdev_leaf_zap,
VDEV_LEAF_ZAP_INITIALIZE_ACTION_TIME, sizeof (val),
1, &val, tx));
}
uint64_t initialize_state = vd->vdev_initialize_state;
VERIFY0(zap_update(mos, vd->vdev_leaf_zap,
VDEV_LEAF_ZAP_INITIALIZE_STATE, sizeof (initialize_state), 1,
&initialize_state, tx));
}
static void
vdev_initialize_zap_remove_sync(void *arg, dmu_tx_t *tx)
{
uint64_t guid = *(uint64_t *)arg;
kmem_free(arg, sizeof (uint64_t));
vdev_t *vd = spa_lookup_by_guid(tx->tx_pool->dp_spa, guid, B_FALSE);
if (vd == NULL || vd->vdev_top->vdev_removing || !vdev_is_concrete(vd))
return;
ASSERT3S(vd->vdev_initialize_state, ==, VDEV_INITIALIZE_NONE);
ASSERT3U(vd->vdev_leaf_zap, !=, 0);
vd->vdev_initialize_last_offset = 0;
vd->vdev_initialize_action_time = 0;
objset_t *mos = vd->vdev_spa->spa_meta_objset;
int error;
error = zap_remove(mos, vd->vdev_leaf_zap,
VDEV_LEAF_ZAP_INITIALIZE_LAST_OFFSET, tx);
VERIFY(error == 0 || error == ENOENT);
error = zap_remove(mos, vd->vdev_leaf_zap,
VDEV_LEAF_ZAP_INITIALIZE_STATE, tx);
VERIFY(error == 0 || error == ENOENT);
error = zap_remove(mos, vd->vdev_leaf_zap,
VDEV_LEAF_ZAP_INITIALIZE_ACTION_TIME, tx);
VERIFY(error == 0 || error == ENOENT);
}
static void
vdev_initialize_change_state(vdev_t *vd, vdev_initializing_state_t new_state)
{
ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock));
spa_t *spa = vd->vdev_spa;
if (new_state == vd->vdev_initialize_state)
return;
/*
* Copy the vd's guid, this will be freed by the sync task.
*/
uint64_t *guid = kmem_zalloc(sizeof (uint64_t), KM_SLEEP);
*guid = vd->vdev_guid;
/*
* If we're suspending, then preserving the original start time.
*/
if (vd->vdev_initialize_state != VDEV_INITIALIZE_SUSPENDED) {
vd->vdev_initialize_action_time = gethrestime_sec();
}
vdev_initializing_state_t old_state = vd->vdev_initialize_state;
vd->vdev_initialize_state = new_state;
dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir);
VERIFY0(dmu_tx_assign(tx, TXG_WAIT));
if (new_state != VDEV_INITIALIZE_NONE) {
dsl_sync_task_nowait(spa_get_dsl(spa),
vdev_initialize_zap_update_sync, guid, tx);
} else {
dsl_sync_task_nowait(spa_get_dsl(spa),
vdev_initialize_zap_remove_sync, guid, tx);
}
switch (new_state) {
case VDEV_INITIALIZE_ACTIVE:
spa_history_log_internal(spa, "initialize", tx,
"vdev=%s activated", vd->vdev_path);
break;
case VDEV_INITIALIZE_SUSPENDED:
spa_history_log_internal(spa, "initialize", tx,
"vdev=%s suspended", vd->vdev_path);
break;
case VDEV_INITIALIZE_CANCELED:
if (old_state == VDEV_INITIALIZE_ACTIVE ||
old_state == VDEV_INITIALIZE_SUSPENDED)
spa_history_log_internal(spa, "initialize", tx,
"vdev=%s canceled", vd->vdev_path);
break;
case VDEV_INITIALIZE_COMPLETE:
spa_history_log_internal(spa, "initialize", tx,
"vdev=%s complete", vd->vdev_path);
break;
case VDEV_INITIALIZE_NONE:
spa_history_log_internal(spa, "uninitialize", tx,
"vdev=%s", vd->vdev_path);
break;
default:
panic("invalid state %llu", (unsigned long long)new_state);
}
dmu_tx_commit(tx);
if (new_state != VDEV_INITIALIZE_ACTIVE)
spa_notify_waiters(spa);
}
static void
vdev_initialize_cb(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
mutex_enter(&vd->vdev_initialize_io_lock);
if (zio->io_error == ENXIO && !vdev_writeable(vd)) {
/*
* The I/O failed because the vdev was unavailable; roll the
* last offset back. (This works because spa_sync waits on
* spa_txg_zio before it runs sync tasks.)
*/
uint64_t *off =
&vd->vdev_initialize_offset[zio->io_txg & TXG_MASK];
*off = MIN(*off, zio->io_offset);
} else {
/*
* Since initializing is best-effort, we ignore I/O errors and
* rely on vdev_probe to determine if the errors are more
* critical.
*/
if (zio->io_error != 0)
vd->vdev_stat.vs_initialize_errors++;
vd->vdev_initialize_bytes_done += zio->io_orig_size;
}
ASSERT3U(vd->vdev_initialize_inflight, >, 0);
vd->vdev_initialize_inflight--;
cv_broadcast(&vd->vdev_initialize_io_cv);
mutex_exit(&vd->vdev_initialize_io_lock);
spa_config_exit(vd->vdev_spa, SCL_STATE_ALL, vd);
}
/* Takes care of physical writing and limiting # of concurrent ZIOs. */
static int
vdev_initialize_write(vdev_t *vd, uint64_t start, uint64_t size, abd_t *data)
{
spa_t *spa = vd->vdev_spa;
/* Limit inflight initializing I/Os */
mutex_enter(&vd->vdev_initialize_io_lock);
while (vd->vdev_initialize_inflight >= zfs_initialize_limit) {
cv_wait(&vd->vdev_initialize_io_cv,
&vd->vdev_initialize_io_lock);
}
vd->vdev_initialize_inflight++;
mutex_exit(&vd->vdev_initialize_io_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);
spa_config_enter(spa, SCL_STATE_ALL, vd, RW_READER);
mutex_enter(&vd->vdev_initialize_lock);
if (vd->vdev_initialize_offset[txg & TXG_MASK] == 0) {
uint64_t *guid = kmem_zalloc(sizeof (uint64_t), KM_SLEEP);
*guid = vd->vdev_guid;
/* This is the first write of this txg. */
dsl_sync_task_nowait(spa_get_dsl(spa),
vdev_initialize_zap_update_sync, guid, tx);
}
/*
* We know the vdev struct will still be around since all
* consumers of vdev_free must stop the initialization first.
*/
if (vdev_initialize_should_stop(vd)) {
mutex_enter(&vd->vdev_initialize_io_lock);
ASSERT3U(vd->vdev_initialize_inflight, >, 0);
vd->vdev_initialize_inflight--;
mutex_exit(&vd->vdev_initialize_io_lock);
spa_config_exit(vd->vdev_spa, SCL_STATE_ALL, vd);
mutex_exit(&vd->vdev_initialize_lock);
dmu_tx_commit(tx);
return (SET_ERROR(EINTR));
}
mutex_exit(&vd->vdev_initialize_lock);
vd->vdev_initialize_offset[txg & TXG_MASK] = start + size;
zio_nowait(zio_write_phys(spa->spa_txg_zio[txg & TXG_MASK], vd, start,
size, data, ZIO_CHECKSUM_OFF, vdev_initialize_cb, NULL,
ZIO_PRIORITY_INITIALIZING, ZIO_FLAG_CANFAIL, B_FALSE));
/* vdev_initialize_cb releases SCL_STATE_ALL */
dmu_tx_commit(tx);
return (0);
}
/*
* Callback to fill each ABD chunk with zfs_initialize_value. len must be
* divisible by sizeof (uint64_t), and buf must be 8-byte aligned. The ABD
* allocation will guarantee these for us.
*/
static int
vdev_initialize_block_fill(void *buf, size_t len, void *unused)
{
(void) unused;
ASSERT0(len % sizeof (uint64_t));
for (uint64_t i = 0; i < len; i += sizeof (uint64_t)) {
*(uint64_t *)((char *)(buf) + i) = zfs_initialize_value;
}
return (0);
}
static abd_t *
vdev_initialize_block_alloc(void)
{
/* Allocate ABD for filler data */
abd_t *data = abd_alloc_for_io(zfs_initialize_chunk_size, B_FALSE);
ASSERT0(zfs_initialize_chunk_size % sizeof (uint64_t));
(void) abd_iterate_func(data, 0, zfs_initialize_chunk_size,
vdev_initialize_block_fill, NULL);
return (data);
}
static void
vdev_initialize_block_free(abd_t *data)
{
abd_free(data);
}
static int
vdev_initialize_ranges(vdev_t *vd, abd_t *data)
{
range_tree_t *rt = vd->vdev_initialize_tree;
zfs_btree_t *bt = &rt->rt_root;
zfs_btree_index_t where;
for (range_seg_t *rs = zfs_btree_first(bt, &where); rs != NULL;
rs = zfs_btree_next(bt, &where, &where)) {
uint64_t size = rs_get_end(rs, rt) - rs_get_start(rs, rt);
/* Split range into legally-sized physical chunks */
uint64_t writes_required =
((size - 1) / zfs_initialize_chunk_size) + 1;
for (uint64_t w = 0; w < writes_required; w++) {
int error;
error = vdev_initialize_write(vd,
VDEV_LABEL_START_SIZE + rs_get_start(rs, rt) +
(w * zfs_initialize_chunk_size),
MIN(size - (w * zfs_initialize_chunk_size),
zfs_initialize_chunk_size), data);
if (error != 0)
return (error);
}
}
return (0);
}
static void
vdev_initialize_xlate_last_rs_end(void *arg, range_seg64_t *physical_rs)
{
uint64_t *last_rs_end = (uint64_t *)arg;
if (physical_rs->rs_end > *last_rs_end)
*last_rs_end = physical_rs->rs_end;
}
static void
vdev_initialize_xlate_progress(void *arg, range_seg64_t *physical_rs)
{
vdev_t *vd = (vdev_t *)arg;
uint64_t size = physical_rs->rs_end - physical_rs->rs_start;
vd->vdev_initialize_bytes_est += size;
if (vd->vdev_initialize_last_offset > physical_rs->rs_end) {
vd->vdev_initialize_bytes_done += size;
} else if (vd->vdev_initialize_last_offset > physical_rs->rs_start &&
vd->vdev_initialize_last_offset < physical_rs->rs_end) {
vd->vdev_initialize_bytes_done +=
vd->vdev_initialize_last_offset - physical_rs->rs_start;
}
}
static void
vdev_initialize_calculate_progress(vdev_t *vd)
{
ASSERT(spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_READER) ||
spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_WRITER));
ASSERT(vd->vdev_leaf_zap != 0);
vd->vdev_initialize_bytes_est = 0;
vd->vdev_initialize_bytes_done = 0;
for (uint64_t i = 0; i < vd->vdev_top->vdev_ms_count; i++) {
metaslab_t *msp = vd->vdev_top->vdev_ms[i];
mutex_enter(&msp->ms_lock);
uint64_t ms_free = (msp->ms_size -
metaslab_allocated_space(msp)) /
vdev_get_ndisks(vd->vdev_top);
/*
* Convert the metaslab range to a physical range
* on our vdev. We use this to determine if we are
* in the middle of this metaslab range.
*/
range_seg64_t logical_rs, physical_rs, remain_rs;
logical_rs.rs_start = msp->ms_start;
logical_rs.rs_end = msp->ms_start + msp->ms_size;
/* Metaslab space after this offset has not been initialized */
vdev_xlate(vd, &logical_rs, &physical_rs, &remain_rs);
if (vd->vdev_initialize_last_offset <= physical_rs.rs_start) {
vd->vdev_initialize_bytes_est += ms_free;
mutex_exit(&msp->ms_lock);
continue;
}
/* Metaslab space before this offset has been initialized */
uint64_t last_rs_end = physical_rs.rs_end;
if (!vdev_xlate_is_empty(&remain_rs)) {
vdev_xlate_walk(vd, &remain_rs,
vdev_initialize_xlate_last_rs_end, &last_rs_end);
}
if (vd->vdev_initialize_last_offset > last_rs_end) {
vd->vdev_initialize_bytes_done += ms_free;
vd->vdev_initialize_bytes_est += ms_free;
mutex_exit(&msp->ms_lock);
continue;
}
/*
* If we get here, we're in the middle of initializing this
* metaslab. Load it and walk the free tree for more accurate
* progress estimation.
*/
VERIFY0(metaslab_load(msp));
zfs_btree_index_t where;
range_tree_t *rt = msp->ms_allocatable;
for (range_seg_t *rs =
zfs_btree_first(&rt->rt_root, &where); rs;
rs = zfs_btree_next(&rt->rt_root, &where,
&where)) {
logical_rs.rs_start = rs_get_start(rs, rt);
logical_rs.rs_end = rs_get_end(rs, rt);
vdev_xlate_walk(vd, &logical_rs,
vdev_initialize_xlate_progress, vd);
}
mutex_exit(&msp->ms_lock);
}
}
static int
vdev_initialize_load(vdev_t *vd)
{
int err = 0;
ASSERT(spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_READER) ||
spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_WRITER));
ASSERT(vd->vdev_leaf_zap != 0);
if (vd->vdev_initialize_state == VDEV_INITIALIZE_ACTIVE ||
vd->vdev_initialize_state == VDEV_INITIALIZE_SUSPENDED) {
err = zap_lookup(vd->vdev_spa->spa_meta_objset,
vd->vdev_leaf_zap, VDEV_LEAF_ZAP_INITIALIZE_LAST_OFFSET,
sizeof (vd->vdev_initialize_last_offset), 1,
&vd->vdev_initialize_last_offset);
if (err == ENOENT) {
vd->vdev_initialize_last_offset = 0;
err = 0;
}
}
vdev_initialize_calculate_progress(vd);
return (err);
}
static void
vdev_initialize_xlate_range_add(void *arg, range_seg64_t *physical_rs)
{
vdev_t *vd = arg;
/* Only add segments that we have not visited yet */
if (physical_rs->rs_end <= vd->vdev_initialize_last_offset)
return;
/* Pick up where we left off mid-range. */
if (vd->vdev_initialize_last_offset > physical_rs->rs_start) {
zfs_dbgmsg("range write: vd %s changed (%llu, %llu) to "
"(%llu, %llu)", vd->vdev_path,
(u_longlong_t)physical_rs->rs_start,
(u_longlong_t)physical_rs->rs_end,
(u_longlong_t)vd->vdev_initialize_last_offset,
(u_longlong_t)physical_rs->rs_end);
ASSERT3U(physical_rs->rs_end, >,
vd->vdev_initialize_last_offset);
physical_rs->rs_start = vd->vdev_initialize_last_offset;
}
ASSERT3U(physical_rs->rs_end, >, physical_rs->rs_start);
range_tree_add(vd->vdev_initialize_tree, physical_rs->rs_start,
physical_rs->rs_end - physical_rs->rs_start);
}
/*
* Convert the logical range into a physical range and add it to our
* avl tree.
*/
static void
vdev_initialize_range_add(void *arg, uint64_t start, uint64_t size)
{
vdev_t *vd = arg;
range_seg64_t logical_rs;
logical_rs.rs_start = start;
logical_rs.rs_end = start + size;
ASSERT(vd->vdev_ops->vdev_op_leaf);
vdev_xlate_walk(vd, &logical_rs, vdev_initialize_xlate_range_add, arg);
}
static __attribute__((noreturn)) void
vdev_initialize_thread(void *arg)
{
vdev_t *vd = arg;
spa_t *spa = vd->vdev_spa;
int error = 0;
uint64_t ms_count = 0;
ASSERT(vdev_is_concrete(vd));
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vd->vdev_initialize_last_offset = 0;
VERIFY0(vdev_initialize_load(vd));
abd_t *deadbeef = vdev_initialize_block_alloc();
vd->vdev_initialize_tree = range_tree_create(NULL, RANGE_SEG64, NULL,
0, 0);
for (uint64_t i = 0; !vd->vdev_detached &&
i < vd->vdev_top->vdev_ms_count; i++) {
metaslab_t *msp = vd->vdev_top->vdev_ms[i];
boolean_t unload_when_done = B_FALSE;
/*
* If we've expanded the top-level vdev or it's our
* first pass, calculate our progress.
*/
if (vd->vdev_top->vdev_ms_count != ms_count) {
vdev_initialize_calculate_progress(vd);
ms_count = vd->vdev_top->vdev_ms_count;
}
spa_config_exit(spa, SCL_CONFIG, FTAG);
metaslab_disable(msp);
mutex_enter(&msp->ms_lock);
if (!msp->ms_loaded && !msp->ms_loading)
unload_when_done = B_TRUE;
VERIFY0(metaslab_load(msp));
range_tree_walk(msp->ms_allocatable, vdev_initialize_range_add,
vd);
mutex_exit(&msp->ms_lock);
error = vdev_initialize_ranges(vd, deadbeef);
metaslab_enable(msp, B_TRUE, unload_when_done);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
range_tree_vacate(vd->vdev_initialize_tree, NULL, NULL);
if (error != 0)
break;
}
spa_config_exit(spa, SCL_CONFIG, FTAG);
mutex_enter(&vd->vdev_initialize_io_lock);
while (vd->vdev_initialize_inflight > 0) {
cv_wait(&vd->vdev_initialize_io_cv,
&vd->vdev_initialize_io_lock);
}
mutex_exit(&vd->vdev_initialize_io_lock);
range_tree_destroy(vd->vdev_initialize_tree);
vdev_initialize_block_free(deadbeef);
vd->vdev_initialize_tree = NULL;
mutex_enter(&vd->vdev_initialize_lock);
if (!vd->vdev_initialize_exit_wanted) {
if (vdev_writeable(vd)) {
vdev_initialize_change_state(vd,
VDEV_INITIALIZE_COMPLETE);
} else if (vd->vdev_faulted) {
vdev_initialize_change_state(vd,
VDEV_INITIALIZE_CANCELED);
}
}
ASSERT(vd->vdev_initialize_thread != NULL ||
vd->vdev_initialize_inflight == 0);
/*
* Drop the vdev_initialize_lock while we sync out the
* txg since it's possible that a device might be trying to
* come online and must check to see if it needs to restart an
* initialization. That thread will be holding the spa_config_lock
* which would prevent the txg_wait_synced from completing.
*/
mutex_exit(&vd->vdev_initialize_lock);
txg_wait_synced(spa_get_dsl(spa), 0);
mutex_enter(&vd->vdev_initialize_lock);
vd->vdev_initialize_thread = NULL;
cv_broadcast(&vd->vdev_initialize_cv);
mutex_exit(&vd->vdev_initialize_lock);
thread_exit();
}
/*
* Initiates a device. Caller must hold vdev_initialize_lock.
* Device must be a leaf and not already be initializing.
*/
void
vdev_initialize(vdev_t *vd)
{
ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock));
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(vdev_is_concrete(vd));
ASSERT3P(vd->vdev_initialize_thread, ==, NULL);
ASSERT(!vd->vdev_detached);
ASSERT(!vd->vdev_initialize_exit_wanted);
ASSERT(!vd->vdev_top->vdev_removing);
+ ASSERT(!vd->vdev_top->vdev_rz_expanding);
vdev_initialize_change_state(vd, VDEV_INITIALIZE_ACTIVE);
vd->vdev_initialize_thread = thread_create(NULL, 0,
vdev_initialize_thread, vd, 0, &p0, TS_RUN, maxclsyspri);
}
/*
* Uninitializes a device. Caller must hold vdev_initialize_lock.
* Device must be a leaf and not already be initializing.
*/
void
vdev_uninitialize(vdev_t *vd)
{
ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock));
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(vdev_is_concrete(vd));
ASSERT3P(vd->vdev_initialize_thread, ==, NULL);
ASSERT(!vd->vdev_detached);
ASSERT(!vd->vdev_initialize_exit_wanted);
ASSERT(!vd->vdev_top->vdev_removing);
vdev_initialize_change_state(vd, VDEV_INITIALIZE_NONE);
}
/*
* Wait for the initialize thread to be terminated (cancelled or stopped).
*/
static void
vdev_initialize_stop_wait_impl(vdev_t *vd)
{
ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock));
while (vd->vdev_initialize_thread != NULL)
cv_wait(&vd->vdev_initialize_cv, &vd->vdev_initialize_lock);
ASSERT3P(vd->vdev_initialize_thread, ==, NULL);
vd->vdev_initialize_exit_wanted = B_FALSE;
}
/*
* Wait for vdev initialize threads which were either to cleanly exit.
*/
void
vdev_initialize_stop_wait(spa_t *spa, list_t *vd_list)
{
(void) spa;
vdev_t *vd;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
while ((vd = list_remove_head(vd_list)) != NULL) {
mutex_enter(&vd->vdev_initialize_lock);
vdev_initialize_stop_wait_impl(vd);
mutex_exit(&vd->vdev_initialize_lock);
}
}
/*
* Stop initializing a device, with the resultant initializing state being
* tgt_state. For blocking behavior pass NULL for vd_list. Otherwise, when
* a list_t is provided the stopping vdev is inserted in to the list. Callers
* are then required to call vdev_initialize_stop_wait() to block for all the
* initialization threads to exit. The caller must hold vdev_initialize_lock
* and must not be writing to the spa config, as the initializing thread may
* try to enter the config as a reader before exiting.
*/
void
vdev_initialize_stop(vdev_t *vd, vdev_initializing_state_t tgt_state,
list_t *vd_list)
{
ASSERT(!spa_config_held(vd->vdev_spa, SCL_CONFIG|SCL_STATE, RW_WRITER));
ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock));
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(vdev_is_concrete(vd));
/*
* Allow cancel requests to proceed even if the initialize thread
* has stopped.
*/
if (vd->vdev_initialize_thread == NULL &&
tgt_state != VDEV_INITIALIZE_CANCELED) {
return;
}
vdev_initialize_change_state(vd, tgt_state);
vd->vdev_initialize_exit_wanted = B_TRUE;
if (vd_list == NULL) {
vdev_initialize_stop_wait_impl(vd);
} else {
ASSERT(MUTEX_HELD(&spa_namespace_lock));
list_insert_tail(vd_list, vd);
}
}
static void
vdev_initialize_stop_all_impl(vdev_t *vd, vdev_initializing_state_t tgt_state,
list_t *vd_list)
{
if (vd->vdev_ops->vdev_op_leaf && vdev_is_concrete(vd)) {
mutex_enter(&vd->vdev_initialize_lock);
vdev_initialize_stop(vd, tgt_state, vd_list);
mutex_exit(&vd->vdev_initialize_lock);
return;
}
for (uint64_t i = 0; i < vd->vdev_children; i++) {
vdev_initialize_stop_all_impl(vd->vdev_child[i], tgt_state,
vd_list);
}
}
/*
* Convenience function to stop initializing of a vdev tree and set all
* initialize thread pointers to NULL.
*/
void
vdev_initialize_stop_all(vdev_t *vd, vdev_initializing_state_t tgt_state)
{
spa_t *spa = vd->vdev_spa;
list_t vd_list;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
list_create(&vd_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_initialize_node));
vdev_initialize_stop_all_impl(vd, tgt_state, &vd_list);
vdev_initialize_stop_wait(spa, &vd_list);
if (vd->vdev_spa->spa_sync_on) {
/* Make sure that our state has been synced to disk */
txg_wait_synced(spa_get_dsl(vd->vdev_spa), 0);
}
list_destroy(&vd_list);
}
void
vdev_initialize_restart(vdev_t *vd)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(!spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER));
if (vd->vdev_leaf_zap != 0) {
mutex_enter(&vd->vdev_initialize_lock);
uint64_t initialize_state = VDEV_INITIALIZE_NONE;
int err = zap_lookup(vd->vdev_spa->spa_meta_objset,
vd->vdev_leaf_zap, VDEV_LEAF_ZAP_INITIALIZE_STATE,
sizeof (initialize_state), 1, &initialize_state);
ASSERT(err == 0 || err == ENOENT);
vd->vdev_initialize_state = initialize_state;
uint64_t timestamp = 0;
err = zap_lookup(vd->vdev_spa->spa_meta_objset,
vd->vdev_leaf_zap, VDEV_LEAF_ZAP_INITIALIZE_ACTION_TIME,
sizeof (timestamp), 1, &timestamp);
ASSERT(err == 0 || err == ENOENT);
vd->vdev_initialize_action_time = timestamp;
- if (vd->vdev_initialize_state == VDEV_INITIALIZE_SUSPENDED ||
- vd->vdev_offline) {
+ if ((vd->vdev_initialize_state == VDEV_INITIALIZE_SUSPENDED ||
+ vd->vdev_offline) && !vd->vdev_top->vdev_rz_expanding) {
/* load progress for reporting, but don't resume */
VERIFY0(vdev_initialize_load(vd));
} else if (vd->vdev_initialize_state ==
VDEV_INITIALIZE_ACTIVE && vdev_writeable(vd) &&
!vd->vdev_top->vdev_removing &&
+ !vd->vdev_top->vdev_rz_expanding &&
vd->vdev_initialize_thread == NULL) {
vdev_initialize(vd);
}
mutex_exit(&vd->vdev_initialize_lock);
}
for (uint64_t i = 0; i < vd->vdev_children; i++) {
vdev_initialize_restart(vd->vdev_child[i]);
}
}
EXPORT_SYMBOL(vdev_initialize);
EXPORT_SYMBOL(vdev_uninitialize);
EXPORT_SYMBOL(vdev_initialize_stop);
EXPORT_SYMBOL(vdev_initialize_stop_all);
EXPORT_SYMBOL(vdev_initialize_stop_wait);
EXPORT_SYMBOL(vdev_initialize_restart);
ZFS_MODULE_PARAM(zfs, zfs_, initialize_value, U64, ZMOD_RW,
"Value written during zpool initialize");
ZFS_MODULE_PARAM(zfs, zfs_, initialize_chunk_size, U64, ZMOD_RW,
"Size in bytes of writes by zpool initialize");
diff --git a/sys/contrib/openzfs/module/zfs/vdev_label.c b/sys/contrib/openzfs/module/zfs/vdev_label.c
index a2e5524a8391..e8f562a1a6a2 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_label.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_label.c
@@ -1,2041 +1,2084 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2020 by Delphix. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
*/
/*
* Virtual Device Labels
* ---------------------
*
* The vdev label serves several distinct purposes:
*
* 1. Uniquely identify this device as part of a ZFS pool and confirm its
* identity within the pool.
*
* 2. Verify that all the devices given in a configuration are present
* within the pool.
*
* 3. Determine the uberblock for the pool.
*
* 4. In case of an import operation, determine the configuration of the
* toplevel vdev of which it is a part.
*
* 5. If an import operation cannot find all the devices in the pool,
* provide enough information to the administrator to determine which
* devices are missing.
*
* It is important to note that while the kernel is responsible for writing the
* label, it only consumes the information in the first three cases. The
* latter information is only consumed in userland when determining the
* configuration to import a pool.
*
*
* Label Organization
* ------------------
*
* Before describing the contents of the label, it's important to understand how
* the labels are written and updated with respect to the uberblock.
*
* When the pool configuration is altered, either because it was newly created
* or a device was added, we want to update all the labels such that we can deal
* with fatal failure at any point. To this end, each disk has two labels which
* are updated before and after the uberblock is synced. Assuming we have
* labels and an uberblock with the following transaction groups:
*
* L1 UB L2
* +------+ +------+ +------+
* | | | | | |
* | t10 | | t10 | | t10 |
* | | | | | |
* +------+ +------+ +------+
*
* In this stable state, the labels and the uberblock were all updated within
* the same transaction group (10). Each label is mirrored and checksummed, so
* that we can detect when we fail partway through writing the label.
*
* In order to identify which labels are valid, the labels are written in the
* following manner:
*
* 1. For each vdev, update 'L1' to the new label
* 2. Update the uberblock
* 3. For each vdev, update 'L2' to the new label
*
* Given arbitrary failure, we can determine the correct label to use based on
* the transaction group. If we fail after updating L1 but before updating the
* UB, we will notice that L1's transaction group is greater than the uberblock,
* so L2 must be valid. If we fail after writing the uberblock but before
* writing L2, we will notice that L2's transaction group is less than L1, and
* therefore L1 is valid.
*
* Another added complexity is that not every label is updated when the config
* is synced. If we add a single device, we do not want to have to re-write
* every label for every device in the pool. This means that both L1 and L2 may
* be older than the pool uberblock, because the necessary information is stored
* on another vdev.
*
*
* On-disk Format
* --------------
*
* The vdev label consists of two distinct parts, and is wrapped within the
* vdev_label_t structure. The label includes 8k of padding to permit legacy
* VTOC disk labels, but is otherwise ignored.
*
* The first half of the label is a packed nvlist which contains pool wide
* properties, per-vdev properties, and configuration information. It is
* described in more detail below.
*
* The latter half of the label consists of a redundant array of uberblocks.
* These uberblocks are updated whenever a transaction group is committed,
* or when the configuration is updated. When a pool is loaded, we scan each
* vdev for the 'best' uberblock.
*
*
* Configuration Information
* -------------------------
*
* The nvlist describing the pool and vdev contains the following elements:
*
* version ZFS on-disk version
* name Pool name
* state Pool state
* txg Transaction group in which this label was written
* pool_guid Unique identifier for this pool
* vdev_tree An nvlist describing vdev tree.
* features_for_read
* An nvlist of the features necessary for reading the MOS.
*
* Each leaf device label also contains the following:
*
* top_guid Unique ID for top-level vdev in which this is contained
* guid Unique ID for the leaf vdev
*
* The 'vs' configuration follows the format described in 'spa_config.c'.
*/
#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/dmu.h>
#include <sys/zap.h>
#include <sys/vdev.h>
#include <sys/vdev_impl.h>
+#include <sys/vdev_raidz.h>
#include <sys/vdev_draid.h>
#include <sys/uberblock_impl.h>
#include <sys/metaslab.h>
#include <sys/metaslab_impl.h>
#include <sys/zio.h>
#include <sys/dsl_scan.h>
#include <sys/abd.h>
#include <sys/fs/zfs.h>
#include <sys/byteorder.h>
#include <sys/zfs_bootenv.h>
/*
* Basic routines to read and write from a vdev label.
* Used throughout the rest of this file.
*/
uint64_t
vdev_label_offset(uint64_t psize, int l, uint64_t offset)
{
ASSERT(offset < sizeof (vdev_label_t));
ASSERT(P2PHASE_TYPED(psize, sizeof (vdev_label_t), uint64_t) == 0);
return (offset + l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ?
0 : psize - VDEV_LABELS * sizeof (vdev_label_t)));
}
/*
* Returns back the vdev label associated with the passed in offset.
*/
int
vdev_label_number(uint64_t psize, uint64_t offset)
{
int l;
if (offset >= psize - VDEV_LABEL_END_SIZE) {
offset -= psize - VDEV_LABEL_END_SIZE;
offset += (VDEV_LABELS / 2) * sizeof (vdev_label_t);
}
l = offset / sizeof (vdev_label_t);
return (l < VDEV_LABELS ? l : -1);
}
static void
vdev_label_read(zio_t *zio, vdev_t *vd, int l, abd_t *buf, uint64_t offset,
uint64_t size, zio_done_func_t *done, void *private, int flags)
{
ASSERT(
spa_config_held(zio->io_spa, SCL_STATE, RW_READER) == SCL_STATE ||
spa_config_held(zio->io_spa, SCL_STATE, RW_WRITER) == SCL_STATE);
ASSERT(flags & ZIO_FLAG_CONFIG_WRITER);
zio_nowait(zio_read_phys(zio, vd,
vdev_label_offset(vd->vdev_psize, l, offset),
size, buf, ZIO_CHECKSUM_LABEL, done, private,
ZIO_PRIORITY_SYNC_READ, flags, B_TRUE));
}
void
vdev_label_write(zio_t *zio, vdev_t *vd, int l, abd_t *buf, uint64_t offset,
uint64_t size, zio_done_func_t *done, void *private, int flags)
{
ASSERT(
spa_config_held(zio->io_spa, SCL_STATE, RW_READER) == SCL_STATE ||
spa_config_held(zio->io_spa, SCL_STATE, RW_WRITER) == SCL_STATE);
ASSERT(flags & ZIO_FLAG_CONFIG_WRITER);
zio_nowait(zio_write_phys(zio, vd,
vdev_label_offset(vd->vdev_psize, l, offset),
size, buf, ZIO_CHECKSUM_LABEL, done, private,
ZIO_PRIORITY_SYNC_WRITE, flags, B_TRUE));
}
/*
* Generate the nvlist representing this vdev's stats
*/
void
vdev_config_generate_stats(vdev_t *vd, nvlist_t *nv)
{
nvlist_t *nvx;
vdev_stat_t *vs;
vdev_stat_ex_t *vsx;
vs = kmem_alloc(sizeof (*vs), KM_SLEEP);
vsx = kmem_alloc(sizeof (*vsx), KM_SLEEP);
vdev_get_stats_ex(vd, vs, vsx);
fnvlist_add_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t *)vs, sizeof (*vs) / sizeof (uint64_t));
/*
* Add extended stats into a special extended stats nvlist. This keeps
* all the extended stats nicely grouped together. The extended stats
* nvlist is then added to the main nvlist.
*/
nvx = fnvlist_alloc();
/* ZIOs in flight to disk */
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_SYNC_READ]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_SYNC_WRITE]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_ASYNC_READ]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_ASYNC_WRITE]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_SCRUB]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_TRIM]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_REBUILD]);
/* ZIOs pending */
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_R_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_SYNC_READ]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_W_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_SYNC_WRITE]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_R_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_ASYNC_READ]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_W_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_ASYNC_WRITE]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SCRUB_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_SCRUB]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_TRIM_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_TRIM]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_REBUILD_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_REBUILD]);
/* Histograms */
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO,
vsx->vsx_total_histo[ZIO_TYPE_READ],
ARRAY_SIZE(vsx->vsx_total_histo[ZIO_TYPE_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO,
vsx->vsx_total_histo[ZIO_TYPE_WRITE],
ARRAY_SIZE(vsx->vsx_total_histo[ZIO_TYPE_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO,
vsx->vsx_disk_histo[ZIO_TYPE_READ],
ARRAY_SIZE(vsx->vsx_disk_histo[ZIO_TYPE_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO,
vsx->vsx_disk_histo[ZIO_TYPE_WRITE],
ARRAY_SIZE(vsx->vsx_disk_histo[ZIO_TYPE_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_READ],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_WRITE],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_READ],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_WRITE],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_SCRUB],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_SCRUB]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_TRIM],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_TRIM]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_REBUILD],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_REBUILD]));
/* Request sizes */
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_IND_R_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_READ],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_IND_W_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_WRITE],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_IND_R_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_READ],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_IND_W_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_WRITE],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_IND_SCRUB_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_SCRUB],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_SCRUB]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_IND_TRIM_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_TRIM],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_TRIM]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_IND_REBUILD_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_REBUILD],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_REBUILD]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_AGG_R_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_READ],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_AGG_W_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_WRITE],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_AGG_R_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_READ],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_AGG_W_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_WRITE],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_AGG_SCRUB_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_SCRUB],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_SCRUB]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_AGG_TRIM_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_TRIM],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_TRIM]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_AGG_REBUILD_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_REBUILD],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_REBUILD]));
/* IO delays */
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SLOW_IOS, vs->vs_slow_ios);
/* Add extended stats nvlist to main nvlist */
fnvlist_add_nvlist(nv, ZPOOL_CONFIG_VDEV_STATS_EX, nvx);
fnvlist_free(nvx);
kmem_free(vs, sizeof (*vs));
kmem_free(vsx, sizeof (*vsx));
}
static void
root_vdev_actions_getprogress(vdev_t *vd, nvlist_t *nvl)
{
spa_t *spa = vd->vdev_spa;
if (vd != spa->spa_root_vdev)
return;
/* provide either current or previous scan information */
pool_scan_stat_t ps;
if (spa_scan_get_stats(spa, &ps) == 0) {
fnvlist_add_uint64_array(nvl,
ZPOOL_CONFIG_SCAN_STATS, (uint64_t *)&ps,
sizeof (pool_scan_stat_t) / sizeof (uint64_t));
}
pool_removal_stat_t prs;
if (spa_removal_get_stats(spa, &prs) == 0) {
fnvlist_add_uint64_array(nvl,
ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t *)&prs,
sizeof (prs) / sizeof (uint64_t));
}
pool_checkpoint_stat_t pcs;
if (spa_checkpoint_get_stats(spa, &pcs) == 0) {
fnvlist_add_uint64_array(nvl,
ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t *)&pcs,
sizeof (pcs) / sizeof (uint64_t));
}
+
+ pool_raidz_expand_stat_t pres;
+ if (spa_raidz_expand_get_stats(spa, &pres) == 0) {
+ fnvlist_add_uint64_array(nvl,
+ ZPOOL_CONFIG_RAIDZ_EXPAND_STATS, (uint64_t *)&pres,
+ sizeof (pres) / sizeof (uint64_t));
+ }
}
static void
top_vdev_actions_getprogress(vdev_t *vd, nvlist_t *nvl)
{
if (vd == vd->vdev_top) {
vdev_rebuild_stat_t vrs;
if (vdev_rebuild_get_stats(vd, &vrs) == 0) {
fnvlist_add_uint64_array(nvl,
ZPOOL_CONFIG_REBUILD_STATS, (uint64_t *)&vrs,
sizeof (vrs) / sizeof (uint64_t));
}
}
}
/*
* Generate the nvlist representing this vdev's config.
*/
nvlist_t *
vdev_config_generate(spa_t *spa, vdev_t *vd, boolean_t getstats,
vdev_config_flag_t flags)
{
nvlist_t *nv = NULL;
vdev_indirect_config_t *vic = &vd->vdev_indirect_config;
nv = fnvlist_alloc();
fnvlist_add_string(nv, ZPOOL_CONFIG_TYPE, vd->vdev_ops->vdev_op_type);
if (!(flags & (VDEV_CONFIG_SPARE | VDEV_CONFIG_L2CACHE)))
fnvlist_add_uint64(nv, ZPOOL_CONFIG_ID, vd->vdev_id);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_GUID, vd->vdev_guid);
if (vd->vdev_path != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_PATH, vd->vdev_path);
if (vd->vdev_devid != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_DEVID, vd->vdev_devid);
if (vd->vdev_physpath != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_PHYS_PATH,
vd->vdev_physpath);
if (vd->vdev_enc_sysfs_path != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH,
vd->vdev_enc_sysfs_path);
if (vd->vdev_fru != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_FRU, vd->vdev_fru);
if (vd->vdev_ops->vdev_op_config_generate != NULL)
vd->vdev_ops->vdev_op_config_generate(vd, nv);
if (vd->vdev_wholedisk != -1ULL) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK,
vd->vdev_wholedisk);
}
if (vd->vdev_not_present && !(flags & VDEV_CONFIG_MISSING))
fnvlist_add_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, 1);
if (vd->vdev_isspare)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_SPARE, 1);
if (flags & VDEV_CONFIG_L2CACHE)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_ASHIFT, vd->vdev_ashift);
if (!(flags & (VDEV_CONFIG_SPARE | VDEV_CONFIG_L2CACHE)) &&
vd == vd->vdev_top) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_METASLAB_ARRAY,
vd->vdev_ms_array);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_METASLAB_SHIFT,
vd->vdev_ms_shift);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_ASHIFT, vd->vdev_ashift);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_ASIZE,
vd->vdev_asize);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_LOG, vd->vdev_islog);
if (vd->vdev_noalloc) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_NONALLOCATING,
vd->vdev_noalloc);
}
/*
* Slog devices are removed synchronously so don't
* persist the vdev_removing flag to the label.
*/
if (vd->vdev_removing && !vd->vdev_islog) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_REMOVING,
vd->vdev_removing);
}
/* zpool command expects alloc class data */
if (getstats && vd->vdev_alloc_bias != VDEV_BIAS_NONE) {
const char *bias = NULL;
switch (vd->vdev_alloc_bias) {
case VDEV_BIAS_LOG:
bias = VDEV_ALLOC_BIAS_LOG;
break;
case VDEV_BIAS_SPECIAL:
bias = VDEV_ALLOC_BIAS_SPECIAL;
break;
case VDEV_BIAS_DEDUP:
bias = VDEV_ALLOC_BIAS_DEDUP;
break;
default:
ASSERT3U(vd->vdev_alloc_bias, ==,
VDEV_BIAS_NONE);
}
fnvlist_add_string(nv, ZPOOL_CONFIG_ALLOCATION_BIAS,
bias);
}
}
if (vd->vdev_dtl_sm != NULL) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_DTL,
space_map_object(vd->vdev_dtl_sm));
}
if (vic->vic_mapping_object != 0) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_INDIRECT_OBJECT,
vic->vic_mapping_object);
}
if (vic->vic_births_object != 0) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_INDIRECT_BIRTHS,
vic->vic_births_object);
}
if (vic->vic_prev_indirect_vdev != UINT64_MAX) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_PREV_INDIRECT_VDEV,
vic->vic_prev_indirect_vdev);
}
if (vd->vdev_crtxg)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_CREATE_TXG, vd->vdev_crtxg);
if (vd->vdev_expansion_time)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_EXPANSION_TIME,
vd->vdev_expansion_time);
if (flags & VDEV_CONFIG_MOS) {
if (vd->vdev_leaf_zap != 0) {
ASSERT(vd->vdev_ops->vdev_op_leaf);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_VDEV_LEAF_ZAP,
vd->vdev_leaf_zap);
}
if (vd->vdev_top_zap != 0) {
ASSERT(vd == vd->vdev_top);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_VDEV_TOP_ZAP,
vd->vdev_top_zap);
}
if (vd->vdev_ops == &vdev_root_ops && vd->vdev_root_zap != 0 &&
spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_AVZ_V2)) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_VDEV_ROOT_ZAP,
vd->vdev_root_zap);
}
if (vd->vdev_resilver_deferred) {
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(spa->spa_resilver_deferred);
fnvlist_add_boolean(nv, ZPOOL_CONFIG_RESILVER_DEFER);
}
}
if (getstats) {
vdev_config_generate_stats(vd, nv);
root_vdev_actions_getprogress(vd, nv);
top_vdev_actions_getprogress(vd, nv);
/*
* Note: this can be called from open context
* (spa_get_stats()), so we need the rwlock to prevent
* the mapping from being changed by condensing.
*/
rw_enter(&vd->vdev_indirect_rwlock, RW_READER);
if (vd->vdev_indirect_mapping != NULL) {
ASSERT(vd->vdev_indirect_births != NULL);
vdev_indirect_mapping_t *vim =
vd->vdev_indirect_mapping;
fnvlist_add_uint64(nv, ZPOOL_CONFIG_INDIRECT_SIZE,
vdev_indirect_mapping_size(vim));
}
rw_exit(&vd->vdev_indirect_rwlock);
if (vd->vdev_mg != NULL &&
vd->vdev_mg->mg_fragmentation != ZFS_FRAG_INVALID) {
/*
* Compute approximately how much memory would be used
* for the indirect mapping if this device were to
* be removed.
*
* Note: If the frag metric is invalid, then not
* enough metaslabs have been converted to have
* histograms.
*/
uint64_t seg_count = 0;
uint64_t to_alloc = vd->vdev_stat.vs_alloc;
/*
* There are the same number of allocated segments
* as free segments, so we will have at least one
* entry per free segment. However, small free
* segments (smaller than vdev_removal_max_span)
* will be combined with adjacent allocated segments
* as a single mapping.
*/
for (int i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++) {
if (i + 1 < highbit64(vdev_removal_max_span)
- 1) {
to_alloc +=
vd->vdev_mg->mg_histogram[i] <<
(i + 1);
} else {
seg_count +=
vd->vdev_mg->mg_histogram[i];
}
}
/*
* The maximum length of a mapping is
* zfs_remove_max_segment, so we need at least one entry
* per zfs_remove_max_segment of allocated data.
*/
seg_count += to_alloc / spa_remove_max_segment(spa);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_INDIRECT_SIZE,
seg_count *
sizeof (vdev_indirect_mapping_entry_phys_t));
}
}
if (!vd->vdev_ops->vdev_op_leaf) {
nvlist_t **child;
uint64_t c;
ASSERT(!vd->vdev_ishole);
child = kmem_alloc(vd->vdev_children * sizeof (nvlist_t *),
KM_SLEEP);
for (c = 0; c < vd->vdev_children; c++) {
child[c] = vdev_config_generate(spa, vd->vdev_child[c],
getstats, flags);
}
fnvlist_add_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
(const nvlist_t * const *)child, vd->vdev_children);
for (c = 0; c < vd->vdev_children; c++)
nvlist_free(child[c]);
kmem_free(child, vd->vdev_children * sizeof (nvlist_t *));
} else {
const char *aux = NULL;
if (vd->vdev_offline && !vd->vdev_tmpoffline)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_OFFLINE, B_TRUE);
if (vd->vdev_resilver_txg != 0)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_RESILVER_TXG,
vd->vdev_resilver_txg);
if (vd->vdev_rebuild_txg != 0)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_REBUILD_TXG,
vd->vdev_rebuild_txg);
if (vd->vdev_faulted)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_FAULTED, B_TRUE);
if (vd->vdev_degraded)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_DEGRADED, B_TRUE);
if (vd->vdev_removed)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_REMOVED, B_TRUE);
if (vd->vdev_unspare)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_UNSPARE, B_TRUE);
if (vd->vdev_ishole)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_HOLE, B_TRUE);
/* Set the reason why we're FAULTED/DEGRADED. */
switch (vd->vdev_stat.vs_aux) {
case VDEV_AUX_ERR_EXCEEDED:
aux = "err_exceeded";
break;
case VDEV_AUX_EXTERNAL:
aux = "external";
break;
}
if (aux != NULL && !vd->vdev_tmpoffline) {
fnvlist_add_string(nv, ZPOOL_CONFIG_AUX_STATE, aux);
} else {
/*
* We're healthy - clear any previous AUX_STATE values.
*/
if (nvlist_exists(nv, ZPOOL_CONFIG_AUX_STATE))
nvlist_remove_all(nv, ZPOOL_CONFIG_AUX_STATE);
}
if (vd->vdev_splitting && vd->vdev_orig_guid != 0LL) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_ORIG_GUID,
vd->vdev_orig_guid);
}
}
return (nv);
}
/*
* Generate a view of the top-level vdevs. If we currently have holes
* in the namespace, then generate an array which contains a list of holey
* vdevs. Additionally, add the number of top-level children that currently
* exist.
*/
void
vdev_top_config_generate(spa_t *spa, nvlist_t *config)
{
vdev_t *rvd = spa->spa_root_vdev;
uint64_t *array;
uint_t c, idx;
array = kmem_alloc(rvd->vdev_children * sizeof (uint64_t), KM_SLEEP);
for (c = 0, idx = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
if (tvd->vdev_ishole) {
array[idx++] = c;
}
}
if (idx) {
VERIFY(nvlist_add_uint64_array(config, ZPOOL_CONFIG_HOLE_ARRAY,
array, idx) == 0);
}
VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_VDEV_CHILDREN,
rvd->vdev_children) == 0);
kmem_free(array, rvd->vdev_children * sizeof (uint64_t));
}
/*
* Returns the configuration from the label of the given vdev. For vdevs
* which don't have a txg value stored on their label (i.e. spares/cache)
* or have not been completely initialized (txg = 0) just return
* the configuration from the first valid label we find. Otherwise,
* find the most up-to-date label that does not exceed the specified
* 'txg' value.
*/
nvlist_t *
vdev_label_read_config(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
nvlist_t *config = NULL;
vdev_phys_t *vp[VDEV_LABELS];
abd_t *vp_abd[VDEV_LABELS];
zio_t *zio[VDEV_LABELS];
uint64_t best_txg = 0;
uint64_t label_txg = 0;
int error = 0;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SPECULATIVE;
ASSERT(vd->vdev_validate_thread == curthread ||
spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
if (!vdev_readable(vd))
return (NULL);
/*
* The label for a dRAID distributed spare is not stored on disk.
* Instead it is generated when needed which allows us to bypass
* the pipeline when reading the config from the label.
*/
if (vd->vdev_ops == &vdev_draid_spare_ops)
return (vdev_draid_read_config_spare(vd));
for (int l = 0; l < VDEV_LABELS; l++) {
vp_abd[l] = abd_alloc_linear(sizeof (vdev_phys_t), B_TRUE);
vp[l] = abd_to_buf(vp_abd[l]);
}
retry:
for (int l = 0; l < VDEV_LABELS; l++) {
zio[l] = zio_root(spa, NULL, NULL, flags);
vdev_label_read(zio[l], vd, l, vp_abd[l],
offsetof(vdev_label_t, vl_vdev_phys), sizeof (vdev_phys_t),
NULL, NULL, flags);
}
for (int l = 0; l < VDEV_LABELS; l++) {
nvlist_t *label = NULL;
if (zio_wait(zio[l]) == 0 &&
nvlist_unpack(vp[l]->vp_nvlist, sizeof (vp[l]->vp_nvlist),
&label, 0) == 0) {
/*
* Auxiliary vdevs won't have txg values in their
* labels and newly added vdevs may not have been
* completely initialized so just return the
* configuration from the first valid label we
* encounter.
*/
error = nvlist_lookup_uint64(label,
ZPOOL_CONFIG_POOL_TXG, &label_txg);
if ((error || label_txg == 0) && !config) {
config = label;
for (l++; l < VDEV_LABELS; l++)
zio_wait(zio[l]);
break;
} else if (label_txg <= txg && label_txg > best_txg) {
best_txg = label_txg;
nvlist_free(config);
config = fnvlist_dup(label);
}
}
if (label != NULL) {
nvlist_free(label);
label = NULL;
}
}
if (config == NULL && !(flags & ZIO_FLAG_TRYHARD)) {
flags |= ZIO_FLAG_TRYHARD;
goto retry;
}
/*
* We found a valid label but it didn't pass txg restrictions.
*/
if (config == NULL && label_txg != 0) {
vdev_dbgmsg(vd, "label discarded as txg is too large "
"(%llu > %llu)", (u_longlong_t)label_txg,
(u_longlong_t)txg);
}
for (int l = 0; l < VDEV_LABELS; l++) {
abd_free(vp_abd[l]);
}
return (config);
}
/*
* Determine if a device is in use. The 'spare_guid' parameter will be filled
* in with the device guid if this spare is active elsewhere on the system.
*/
static boolean_t
vdev_inuse(vdev_t *vd, uint64_t crtxg, vdev_labeltype_t reason,
uint64_t *spare_guid, uint64_t *l2cache_guid)
{
spa_t *spa = vd->vdev_spa;
uint64_t state, pool_guid, device_guid, txg, spare_pool;
uint64_t vdtxg = 0;
nvlist_t *label;
if (spare_guid)
*spare_guid = 0ULL;
if (l2cache_guid)
*l2cache_guid = 0ULL;
/*
* Read the label, if any, and perform some basic sanity checks.
*/
if ((label = vdev_label_read_config(vd, -1ULL)) == NULL)
return (B_FALSE);
(void) nvlist_lookup_uint64(label, ZPOOL_CONFIG_CREATE_TXG,
&vdtxg);
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE,
&state) != 0 ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID,
&device_guid) != 0) {
nvlist_free(label);
return (B_FALSE);
}
if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
(nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID,
&pool_guid) != 0 ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_TXG,
&txg) != 0)) {
nvlist_free(label);
return (B_FALSE);
}
nvlist_free(label);
/*
* Check to see if this device indeed belongs to the pool it claims to
* be a part of. The only way this is allowed is if the device is a hot
* spare (which we check for later on).
*/
if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
!spa_guid_exists(pool_guid, device_guid) &&
!spa_spare_exists(device_guid, NULL, NULL) &&
!spa_l2cache_exists(device_guid, NULL))
return (B_FALSE);
/*
* If the transaction group is zero, then this an initialized (but
* unused) label. This is only an error if the create transaction
* on-disk is the same as the one we're using now, in which case the
* user has attempted to add the same vdev multiple times in the same
* transaction.
*/
if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
txg == 0 && vdtxg == crtxg)
return (B_TRUE);
/*
* Check to see if this is a spare device. We do an explicit check for
* spa_has_spare() here because it may be on our pending list of spares
* to add.
*/
if (spa_spare_exists(device_guid, &spare_pool, NULL) ||
spa_has_spare(spa, device_guid)) {
if (spare_guid)
*spare_guid = device_guid;
switch (reason) {
case VDEV_LABEL_CREATE:
return (B_TRUE);
case VDEV_LABEL_REPLACE:
return (!spa_has_spare(spa, device_guid) ||
spare_pool != 0ULL);
case VDEV_LABEL_SPARE:
return (spa_has_spare(spa, device_guid));
default:
break;
}
}
/*
* Check to see if this is an l2cache device.
*/
if (spa_l2cache_exists(device_guid, NULL) ||
spa_has_l2cache(spa, device_guid)) {
if (l2cache_guid)
*l2cache_guid = device_guid;
switch (reason) {
case VDEV_LABEL_CREATE:
return (B_TRUE);
case VDEV_LABEL_REPLACE:
return (!spa_has_l2cache(spa, device_guid));
case VDEV_LABEL_L2CACHE:
return (spa_has_l2cache(spa, device_guid));
default:
break;
}
}
/*
* We can't rely on a pool's state if it's been imported
* read-only. Instead we look to see if the pools is marked
* read-only in the namespace and set the state to active.
*/
if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
(spa = spa_by_guid(pool_guid, device_guid)) != NULL &&
spa_mode(spa) == SPA_MODE_READ)
state = POOL_STATE_ACTIVE;
/*
* If the device is marked ACTIVE, then this device is in use by another
* pool on the system.
*/
return (state == POOL_STATE_ACTIVE);
}
/*
* Initialize a vdev label. We check to make sure each leaf device is not in
* use, and writable. We put down an initial label which we will later
* overwrite with a complete label. Note that it's important to do this
* sequentially, not in parallel, so that we catch cases of multiple use of the
* same leaf vdev in the vdev we're creating -- e.g. mirroring a disk with
* itself.
*/
int
vdev_label_init(vdev_t *vd, uint64_t crtxg, vdev_labeltype_t reason)
{
spa_t *spa = vd->vdev_spa;
nvlist_t *label;
vdev_phys_t *vp;
abd_t *vp_abd;
abd_t *bootenv;
uberblock_t *ub;
abd_t *ub_abd;
zio_t *zio;
char *buf;
size_t buflen;
int error;
uint64_t spare_guid = 0, l2cache_guid = 0;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL;
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
for (int c = 0; c < vd->vdev_children; c++)
if ((error = vdev_label_init(vd->vdev_child[c],
crtxg, reason)) != 0)
return (error);
/* Track the creation time for this vdev */
vd->vdev_crtxg = crtxg;
if (!vd->vdev_ops->vdev_op_leaf || !spa_writeable(spa))
return (0);
/*
* Dead vdevs cannot be initialized.
*/
if (vdev_is_dead(vd))
return (SET_ERROR(EIO));
/*
* Determine if the vdev is in use.
*/
if (reason != VDEV_LABEL_REMOVE && reason != VDEV_LABEL_SPLIT &&
vdev_inuse(vd, crtxg, reason, &spare_guid, &l2cache_guid))
return (SET_ERROR(EBUSY));
/*
* If this is a request to add or replace a spare or l2cache device
* that is in use elsewhere on the system, then we must update the
* guid (which was initialized to a random value) to reflect the
* actual GUID (which is shared between multiple pools).
*/
if (reason != VDEV_LABEL_REMOVE && reason != VDEV_LABEL_L2CACHE &&
spare_guid != 0ULL) {
uint64_t guid_delta = spare_guid - vd->vdev_guid;
vd->vdev_guid += guid_delta;
for (vdev_t *pvd = vd; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum += guid_delta;
/*
* If this is a replacement, then we want to fallthrough to the
* rest of the code. If we're adding a spare, then it's already
* labeled appropriately and we can just return.
*/
if (reason == VDEV_LABEL_SPARE)
return (0);
ASSERT(reason == VDEV_LABEL_REPLACE ||
reason == VDEV_LABEL_SPLIT);
}
if (reason != VDEV_LABEL_REMOVE && reason != VDEV_LABEL_SPARE &&
l2cache_guid != 0ULL) {
uint64_t guid_delta = l2cache_guid - vd->vdev_guid;
vd->vdev_guid += guid_delta;
for (vdev_t *pvd = vd; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum += guid_delta;
/*
* If this is a replacement, then we want to fallthrough to the
* rest of the code. If we're adding an l2cache, then it's
* already labeled appropriately and we can just return.
*/
if (reason == VDEV_LABEL_L2CACHE)
return (0);
ASSERT(reason == VDEV_LABEL_REPLACE);
}
/*
* Initialize its label.
*/
vp_abd = abd_alloc_linear(sizeof (vdev_phys_t), B_TRUE);
abd_zero(vp_abd, sizeof (vdev_phys_t));
vp = abd_to_buf(vp_abd);
/*
* Generate a label describing the pool and our top-level vdev.
* We mark it as being from txg 0 to indicate that it's not
* really part of an active pool just yet. The labels will
* be written again with a meaningful txg by spa_sync().
*/
if (reason == VDEV_LABEL_SPARE ||
(reason == VDEV_LABEL_REMOVE && vd->vdev_isspare)) {
/*
* For inactive hot spares, we generate a special label that
* identifies as a mutually shared hot spare. We write the
* label if we are adding a hot spare, or if we are removing an
* active hot spare (in which case we want to revert the
* labels).
*/
VERIFY(nvlist_alloc(&label, NV_UNIQUE_NAME, KM_SLEEP) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_VERSION,
spa_version(spa)) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_POOL_STATE,
POOL_STATE_SPARE) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_GUID,
vd->vdev_guid) == 0);
} else if (reason == VDEV_LABEL_L2CACHE ||
(reason == VDEV_LABEL_REMOVE && vd->vdev_isl2cache)) {
/*
* For level 2 ARC devices, add a special label.
*/
VERIFY(nvlist_alloc(&label, NV_UNIQUE_NAME, KM_SLEEP) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_VERSION,
spa_version(spa)) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_POOL_STATE,
POOL_STATE_L2CACHE) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_GUID,
vd->vdev_guid) == 0);
/*
* This is merely to facilitate reporting the ashift of the
* cache device through zdb. The actual retrieval of the
* ashift (in vdev_alloc()) uses the nvlist
* spa->spa_l2cache->sav_config (populated in
* spa_ld_open_aux_vdevs()).
*/
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_ASHIFT,
vd->vdev_ashift) == 0);
} else {
uint64_t txg = 0ULL;
if (reason == VDEV_LABEL_SPLIT)
txg = spa->spa_uberblock.ub_txg;
label = spa_config_generate(spa, vd, txg, B_FALSE);
/*
* Add our creation time. This allows us to detect multiple
* vdev uses as described above, and automatically expires if we
* fail.
*/
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_CREATE_TXG,
crtxg) == 0);
}
buf = vp->vp_nvlist;
buflen = sizeof (vp->vp_nvlist);
error = nvlist_pack(label, &buf, &buflen, NV_ENCODE_XDR, KM_SLEEP);
if (error != 0) {
nvlist_free(label);
abd_free(vp_abd);
/* EFAULT means nvlist_pack ran out of room */
return (SET_ERROR(error == EFAULT ? ENAMETOOLONG : EINVAL));
}
/*
* Initialize uberblock template.
*/
ub_abd = abd_alloc_linear(VDEV_UBERBLOCK_RING, B_TRUE);
abd_zero(ub_abd, VDEV_UBERBLOCK_RING);
abd_copy_from_buf(ub_abd, &spa->spa_uberblock, sizeof (uberblock_t));
ub = abd_to_buf(ub_abd);
ub->ub_txg = 0;
/* Initialize the 2nd padding area. */
bootenv = abd_alloc_for_io(VDEV_PAD_SIZE, B_TRUE);
abd_zero(bootenv, VDEV_PAD_SIZE);
/*
* Write everything in parallel.
*/
retry:
zio = zio_root(spa, NULL, NULL, flags);
for (int l = 0; l < VDEV_LABELS; l++) {
vdev_label_write(zio, vd, l, vp_abd,
offsetof(vdev_label_t, vl_vdev_phys),
sizeof (vdev_phys_t), NULL, NULL, flags);
/*
* Skip the 1st padding area.
* Zero out the 2nd padding area where it might have
* left over data from previous filesystem format.
*/
vdev_label_write(zio, vd, l, bootenv,
offsetof(vdev_label_t, vl_be),
VDEV_PAD_SIZE, NULL, NULL, flags);
vdev_label_write(zio, vd, l, ub_abd,
offsetof(vdev_label_t, vl_uberblock),
VDEV_UBERBLOCK_RING, NULL, NULL, flags);
}
error = zio_wait(zio);
if (error != 0 && !(flags & ZIO_FLAG_TRYHARD)) {
flags |= ZIO_FLAG_TRYHARD;
goto retry;
}
nvlist_free(label);
abd_free(bootenv);
abd_free(ub_abd);
abd_free(vp_abd);
/*
* If this vdev hasn't been previously identified as a spare, then we
* mark it as such only if a) we are labeling it as a spare, or b) it
* exists as a spare elsewhere in the system. Do the same for
* level 2 ARC devices.
*/
if (error == 0 && !vd->vdev_isspare &&
(reason == VDEV_LABEL_SPARE ||
spa_spare_exists(vd->vdev_guid, NULL, NULL)))
spa_spare_add(vd);
if (error == 0 && !vd->vdev_isl2cache &&
(reason == VDEV_LABEL_L2CACHE ||
spa_l2cache_exists(vd->vdev_guid, NULL)))
spa_l2cache_add(vd);
return (error);
}
/*
* Done callback for vdev_label_read_bootenv_impl. If this is the first
* callback to finish, store our abd in the callback pointer. Otherwise, we
* just free our abd and return.
*/
static void
vdev_label_read_bootenv_done(zio_t *zio)
{
zio_t *rio = zio->io_private;
abd_t **cbp = rio->io_private;
ASSERT3U(zio->io_size, ==, VDEV_PAD_SIZE);
if (zio->io_error == 0) {
mutex_enter(&rio->io_lock);
if (*cbp == NULL) {
/* Will free this buffer in vdev_label_read_bootenv. */
*cbp = zio->io_abd;
} else {
abd_free(zio->io_abd);
}
mutex_exit(&rio->io_lock);
} else {
abd_free(zio->io_abd);
}
}
static void
vdev_label_read_bootenv_impl(zio_t *zio, vdev_t *vd, int flags)
{
for (int c = 0; c < vd->vdev_children; c++)
vdev_label_read_bootenv_impl(zio, vd->vdev_child[c], flags);
/*
* We just use the first label that has a correct checksum; the
* bootloader should have rewritten them all to be the same on boot,
* and any changes we made since boot have been the same across all
* labels.
*/
if (vd->vdev_ops->vdev_op_leaf && vdev_readable(vd)) {
for (int l = 0; l < VDEV_LABELS; l++) {
vdev_label_read(zio, vd, l,
abd_alloc_linear(VDEV_PAD_SIZE, B_FALSE),
offsetof(vdev_label_t, vl_be), VDEV_PAD_SIZE,
vdev_label_read_bootenv_done, zio, flags);
}
}
}
int
vdev_label_read_bootenv(vdev_t *rvd, nvlist_t *bootenv)
{
nvlist_t *config;
spa_t *spa = rvd->vdev_spa;
abd_t *abd = NULL;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SPECULATIVE | ZIO_FLAG_TRYHARD;
ASSERT(bootenv);
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
zio_t *zio = zio_root(spa, NULL, &abd, flags);
vdev_label_read_bootenv_impl(zio, rvd, flags);
int err = zio_wait(zio);
if (abd != NULL) {
char *buf;
vdev_boot_envblock_t *vbe = abd_to_buf(abd);
vbe->vbe_version = ntohll(vbe->vbe_version);
switch (vbe->vbe_version) {
case VB_RAW:
/*
* if we have textual data in vbe_bootenv, create nvlist
* with key "envmap".
*/
fnvlist_add_uint64(bootenv, BOOTENV_VERSION, VB_RAW);
vbe->vbe_bootenv[sizeof (vbe->vbe_bootenv) - 1] = '\0';
fnvlist_add_string(bootenv, GRUB_ENVMAP,
vbe->vbe_bootenv);
break;
case VB_NVLIST:
err = nvlist_unpack(vbe->vbe_bootenv,
sizeof (vbe->vbe_bootenv), &config, 0);
if (err == 0) {
fnvlist_merge(bootenv, config);
nvlist_free(config);
break;
}
zfs_fallthrough;
default:
/* Check for FreeBSD zfs bootonce command string */
buf = abd_to_buf(abd);
if (*buf == '\0') {
fnvlist_add_uint64(bootenv, BOOTENV_VERSION,
VB_NVLIST);
break;
}
fnvlist_add_string(bootenv, FREEBSD_BOOTONCE, buf);
}
/*
* abd was allocated in vdev_label_read_bootenv_impl()
*/
abd_free(abd);
/*
* If we managed to read any successfully,
* return success.
*/
return (0);
}
return (err);
}
int
vdev_label_write_bootenv(vdev_t *vd, nvlist_t *env)
{
zio_t *zio;
spa_t *spa = vd->vdev_spa;
vdev_boot_envblock_t *bootenv;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL;
int error;
size_t nvsize;
char *nvbuf;
const char *tmp;
error = nvlist_size(env, &nvsize, NV_ENCODE_XDR);
if (error != 0)
return (SET_ERROR(error));
if (nvsize >= sizeof (bootenv->vbe_bootenv)) {
return (SET_ERROR(E2BIG));
}
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
error = ENXIO;
for (int c = 0; c < vd->vdev_children; c++) {
int child_err;
child_err = vdev_label_write_bootenv(vd->vdev_child[c], env);
/*
* As long as any of the disks managed to write all of their
* labels successfully, return success.
*/
if (child_err == 0)
error = child_err;
}
if (!vd->vdev_ops->vdev_op_leaf || vdev_is_dead(vd) ||
!vdev_writeable(vd)) {
return (error);
}
ASSERT3U(sizeof (*bootenv), ==, VDEV_PAD_SIZE);
abd_t *abd = abd_alloc_for_io(VDEV_PAD_SIZE, B_TRUE);
abd_zero(abd, VDEV_PAD_SIZE);
bootenv = abd_borrow_buf_copy(abd, VDEV_PAD_SIZE);
nvbuf = bootenv->vbe_bootenv;
nvsize = sizeof (bootenv->vbe_bootenv);
bootenv->vbe_version = fnvlist_lookup_uint64(env, BOOTENV_VERSION);
switch (bootenv->vbe_version) {
case VB_RAW:
if (nvlist_lookup_string(env, GRUB_ENVMAP, &tmp) == 0) {
(void) strlcpy(bootenv->vbe_bootenv, tmp, nvsize);
}
error = 0;
break;
case VB_NVLIST:
error = nvlist_pack(env, &nvbuf, &nvsize, NV_ENCODE_XDR,
KM_SLEEP);
break;
default:
error = EINVAL;
break;
}
if (error == 0) {
bootenv->vbe_version = htonll(bootenv->vbe_version);
abd_return_buf_copy(abd, bootenv, VDEV_PAD_SIZE);
} else {
abd_free(abd);
return (SET_ERROR(error));
}
retry:
zio = zio_root(spa, NULL, NULL, flags);
for (int l = 0; l < VDEV_LABELS; l++) {
vdev_label_write(zio, vd, l, abd,
offsetof(vdev_label_t, vl_be),
VDEV_PAD_SIZE, NULL, NULL, flags);
}
error = zio_wait(zio);
if (error != 0 && !(flags & ZIO_FLAG_TRYHARD)) {
flags |= ZIO_FLAG_TRYHARD;
goto retry;
}
abd_free(abd);
return (error);
}
/*
* ==========================================================================
* uberblock load/sync
* ==========================================================================
*/
/*
* Consider the following situation: txg is safely synced to disk. We've
* written the first uberblock for txg + 1, and then we lose power. When we
* come back up, we fail to see the uberblock for txg + 1 because, say,
* it was on a mirrored device and the replica to which we wrote txg + 1
* is now offline. If we then make some changes and sync txg + 1, and then
* the missing replica comes back, then for a few seconds we'll have two
* conflicting uberblocks on disk with the same txg. The solution is simple:
* among uberblocks with equal txg, choose the one with the latest timestamp.
*/
static int
vdev_uberblock_compare(const uberblock_t *ub1, const uberblock_t *ub2)
{
int cmp = TREE_CMP(ub1->ub_txg, ub2->ub_txg);
if (likely(cmp))
return (cmp);
cmp = TREE_CMP(ub1->ub_timestamp, ub2->ub_timestamp);
if (likely(cmp))
return (cmp);
/*
* If MMP_VALID(ub) && MMP_SEQ_VALID(ub) then the host has an MMP-aware
* ZFS, e.g. OpenZFS >= 0.7.
*
* If one ub has MMP and the other does not, they were written by
* different hosts, which matters for MMP. So we treat no MMP/no SEQ as
* a 0 value.
*
* Since timestamp and txg are the same if we get this far, either is
* acceptable for importing the pool.
*/
unsigned int seq1 = 0;
unsigned int seq2 = 0;
if (MMP_VALID(ub1) && MMP_SEQ_VALID(ub1))
seq1 = MMP_SEQ(ub1);
if (MMP_VALID(ub2) && MMP_SEQ_VALID(ub2))
seq2 = MMP_SEQ(ub2);
return (TREE_CMP(seq1, seq2));
}
struct ubl_cbdata {
- uberblock_t *ubl_ubbest; /* Best uberblock */
+ uberblock_t ubl_latest; /* Most recent uberblock */
+ uberblock_t *ubl_ubbest; /* Best uberblock (w/r/t max_txg) */
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 (vdev_uberblock_compare(ub, &cbp->ubl_latest) > 0) {
+ cbp->ubl_latest = *ub;
+ }
if (ub->ub_txg <= spa->spa_load_max_txg &&
vdev_uberblock_compare(ub, cbp->ubl_ubbest) > 0) {
/*
* Keep track of the vdev in which this uberblock
* was found. We will use this information later
* to obtain the config nvlist associated with
* this uberblock.
*/
*cbp->ubl_ubbest = *ub;
cbp->ubl_vd = vd;
}
mutex_exit(&rio->io_lock);
}
abd_free(zio->io_abd);
}
static void
vdev_uberblock_load_impl(zio_t *zio, vdev_t *vd, int flags,
struct ubl_cbdata *cbp)
{
for (int c = 0; c < vd->vdev_children; c++)
vdev_uberblock_load_impl(zio, vd->vdev_child[c], flags, cbp);
if (vd->vdev_ops->vdev_op_leaf && vdev_readable(vd) &&
vd->vdev_ops != &vdev_draid_spare_ops) {
for (int l = 0; l < VDEV_LABELS; l++) {
for (int n = 0; n < VDEV_UBERBLOCK_COUNT(vd); n++) {
vdev_label_read(zio, vd, l,
abd_alloc_linear(VDEV_UBERBLOCK_SIZE(vd),
B_TRUE), VDEV_UBERBLOCK_OFFSET(vd, n),
VDEV_UBERBLOCK_SIZE(vd),
vdev_uberblock_load_done, zio, flags);
}
}
}
}
/*
* Reads the 'best' uberblock from disk along with its associated
* configuration. First, we read the uberblock array of each label of each
* vdev, keeping track of the uberblock with the highest txg in each array.
* Then, we read the configuration from the same vdev as the best uberblock.
*/
void
vdev_uberblock_load(vdev_t *rvd, uberblock_t *ub, nvlist_t **config)
{
zio_t *zio;
spa_t *spa = rvd->vdev_spa;
struct ubl_cbdata cb;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SPECULATIVE | ZIO_FLAG_TRYHARD;
ASSERT(ub);
ASSERT(config);
memset(ub, 0, sizeof (uberblock_t));
+ memset(&cb, 0, sizeof (cb));
*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);
+ if (ub->ub_raidz_reflow_info !=
+ cb.ubl_latest.ub_raidz_reflow_info) {
+ vdev_dbgmsg(cb.ubl_vd,
+ "spa=%s best uberblock (txg=%llu info=0x%llx) "
+ "has different raidz_reflow_info than latest "
+ "uberblock (txg=%llu info=0x%llx)",
+ spa->spa_name,
+ (u_longlong_t)ub->ub_txg,
+ (u_longlong_t)ub->ub_raidz_reflow_info,
+ (u_longlong_t)cb.ubl_latest.ub_txg,
+ (u_longlong_t)cb.ubl_latest.ub_raidz_reflow_info);
+ memset(ub, 0, sizeof (uberblock_t));
+ spa_config_exit(spa, SCL_ALL, FTAG);
+ return;
+ }
+
*config = vdev_label_read_config(cb.ubl_vd, ub->ub_txg);
if (*config == NULL && spa->spa_extreme_rewind) {
vdev_dbgmsg(cb.ubl_vd, "failed to read label config. "
"Trying again without txg restrictions.");
*config = vdev_label_read_config(cb.ubl_vd, UINT64_MAX);
}
if (*config == NULL) {
vdev_dbgmsg(cb.ubl_vd, "failed to read label config");
}
}
spa_config_exit(spa, SCL_ALL, FTAG);
}
/*
* For use when a leaf vdev is expanded.
* The location of labels 2 and 3 changed, and at the new location the
* uberblock rings are either empty or contain garbage. The sync will write
* new configs there because the vdev is dirty, but expansion also needs the
* uberblock rings copied. Read them from label 0 which did not move.
*
* Since the point is to populate labels {2,3} with valid uberblocks,
* we zero uberblocks we fail to read or which are not valid.
*/
static void
vdev_copy_uberblocks(vdev_t *vd)
{
abd_t *ub_abd;
zio_t *write_zio;
int locks = (SCL_L2ARC | SCL_ZIO);
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SPECULATIVE;
ASSERT(spa_config_held(vd->vdev_spa, SCL_STATE, RW_READER) ==
SCL_STATE);
ASSERT(vd->vdev_ops->vdev_op_leaf);
/*
* No uberblocks are stored on distributed spares, they may be
* safely skipped when expanding a leaf vdev.
*/
if (vd->vdev_ops == &vdev_draid_spare_ops)
return;
spa_config_enter(vd->vdev_spa, locks, FTAG, RW_READER);
ub_abd = abd_alloc_linear(VDEV_UBERBLOCK_SIZE(vd), B_TRUE);
write_zio = zio_root(vd->vdev_spa, NULL, NULL, flags);
for (int n = 0; n < VDEV_UBERBLOCK_COUNT(vd); n++) {
const int src_label = 0;
zio_t *zio;
zio = zio_root(vd->vdev_spa, NULL, NULL, flags);
vdev_label_read(zio, vd, src_label, ub_abd,
VDEV_UBERBLOCK_OFFSET(vd, n), VDEV_UBERBLOCK_SIZE(vd),
NULL, NULL, flags);
if (zio_wait(zio) || uberblock_verify(abd_to_buf(ub_abd)))
abd_zero(ub_abd, VDEV_UBERBLOCK_SIZE(vd));
for (int l = 2; l < VDEV_LABELS; l++)
vdev_label_write(write_zio, vd, l, ub_abd,
VDEV_UBERBLOCK_OFFSET(vd, n),
VDEV_UBERBLOCK_SIZE(vd), NULL, NULL,
flags | ZIO_FLAG_DONT_PROPAGATE);
}
(void) zio_wait(write_zio);
spa_config_exit(vd->vdev_spa, locks, FTAG);
abd_free(ub_abd);
}
/*
* On success, increment root zio's count of good writes.
* We only get credit for writes to known-visible vdevs; see spa_vdev_add().
*/
static void
vdev_uberblock_sync_done(zio_t *zio)
{
uint64_t *good_writes = zio->io_private;
if (zio->io_error == 0 && zio->io_vd->vdev_top->vdev_ms_array != 0)
atomic_inc_64(good_writes);
}
/*
* Write the uberblock to all labels of all leaves of the specified vdev.
*/
static void
vdev_uberblock_sync(zio_t *zio, uint64_t *good_writes,
uberblock_t *ub, vdev_t *vd, int flags)
{
for (uint64_t c = 0; c < vd->vdev_children; c++) {
vdev_uberblock_sync(zio, good_writes,
ub, vd->vdev_child[c], flags);
}
if (!vd->vdev_ops->vdev_op_leaf)
return;
if (!vdev_writeable(vd))
return;
/*
* There's no need to write uberblocks to a distributed spare, they
* are already stored on all the leaves of the parent dRAID. For
* this same reason vdev_uberblock_load_impl() skips distributed
* spares when reading uberblocks.
*/
if (vd->vdev_ops == &vdev_draid_spare_ops)
return;
/* If the vdev was expanded, need to copy uberblock rings. */
if (vd->vdev_state == VDEV_STATE_HEALTHY &&
vd->vdev_copy_uberblocks == B_TRUE) {
vdev_copy_uberblocks(vd);
vd->vdev_copy_uberblocks = B_FALSE;
}
+ /*
+ * We chose a slot based on the txg. If this uberblock has a special
+ * RAIDZ expansion state, then it is essentially an update of the
+ * current uberblock (it has the same txg). However, the current
+ * state is committed, so we want to write it to a different slot. If
+ * we overwrote the same slot, and we lose power during the uberblock
+ * write, and the disk does not do single-sector overwrites
+ * atomically (even though it is required to - i.e. we should see
+ * either the old or the new uberblock), then we could lose this
+ * txg's uberblock. Rewinding to the previous txg's uberblock may not
+ * be possible because RAIDZ expansion may have already overwritten
+ * some of the data, so we need the progress indicator in the
+ * uberblock.
+ */
int m = spa_multihost(vd->vdev_spa) ? MMP_BLOCKS_PER_LABEL : 0;
- int n = ub->ub_txg % (VDEV_UBERBLOCK_COUNT(vd) - m);
+ int n = (ub->ub_txg - (RRSS_GET_STATE(ub) == RRSS_SCRATCH_VALID)) %
+ (VDEV_UBERBLOCK_COUNT(vd) - m);
/* Copy the uberblock_t into the ABD */
abd_t *ub_abd = abd_alloc_for_io(VDEV_UBERBLOCK_SIZE(vd), B_TRUE);
abd_zero(ub_abd, VDEV_UBERBLOCK_SIZE(vd));
abd_copy_from_buf(ub_abd, ub, sizeof (uberblock_t));
for (int l = 0; l < VDEV_LABELS; l++)
vdev_label_write(zio, vd, l, ub_abd,
VDEV_UBERBLOCK_OFFSET(vd, n), VDEV_UBERBLOCK_SIZE(vd),
vdev_uberblock_sync_done, good_writes,
flags | ZIO_FLAG_DONT_PROPAGATE);
abd_free(ub_abd);
}
/* Sync the uberblocks to all vdevs in svd[] */
-static int
+int
vdev_uberblock_sync_list(vdev_t **svd, int svdcount, uberblock_t *ub, int flags)
{
spa_t *spa = svd[0]->vdev_spa;
zio_t *zio;
uint64_t good_writes = 0;
zio = zio_root(spa, NULL, NULL, flags);
for (int v = 0; v < svdcount; v++)
vdev_uberblock_sync(zio, &good_writes, ub, svd[v], flags);
(void) zio_wait(zio);
/*
* Flush the uberblocks to disk. This ensures that the odd labels
* are no longer needed (because the new uberblocks and the even
* labels are safely on disk), so it is safe to overwrite them.
*/
zio = zio_root(spa, NULL, NULL, flags);
for (int v = 0; v < svdcount; v++) {
if (vdev_writeable(svd[v])) {
zio_flush(zio, svd[v]);
}
}
(void) zio_wait(zio);
return (good_writes >= 1 ? 0 : EIO);
}
/*
* On success, increment the count of good writes for our top-level vdev.
*/
static void
vdev_label_sync_done(zio_t *zio)
{
uint64_t *good_writes = zio->io_private;
if (zio->io_error == 0)
atomic_inc_64(good_writes);
}
/*
* If there weren't enough good writes, indicate failure to the parent.
*/
static void
vdev_label_sync_top_done(zio_t *zio)
{
uint64_t *good_writes = zio->io_private;
if (*good_writes == 0)
zio->io_error = SET_ERROR(EIO);
kmem_free(good_writes, sizeof (uint64_t));
}
/*
* We ignore errors for log and cache devices, simply free the private data.
*/
static void
vdev_label_sync_ignore_done(zio_t *zio)
{
kmem_free(zio->io_private, sizeof (uint64_t));
}
/*
* Write all even or odd labels to all leaves of the specified vdev.
*/
static void
vdev_label_sync(zio_t *zio, uint64_t *good_writes,
vdev_t *vd, int l, uint64_t txg, int flags)
{
nvlist_t *label;
vdev_phys_t *vp;
abd_t *vp_abd;
char *buf;
size_t buflen;
for (int c = 0; c < vd->vdev_children; c++) {
vdev_label_sync(zio, good_writes,
vd->vdev_child[c], l, txg, flags);
}
if (!vd->vdev_ops->vdev_op_leaf)
return;
if (!vdev_writeable(vd))
return;
/*
* The top-level config never needs to be written to a distributed
* spare. When read vdev_dspare_label_read_config() will generate
* the config for the vdev_label_read_config().
*/
if (vd->vdev_ops == &vdev_draid_spare_ops)
return;
/*
* Generate a label describing the top-level config to which we belong.
*/
label = spa_config_generate(vd->vdev_spa, vd, txg, B_FALSE);
vp_abd = abd_alloc_linear(sizeof (vdev_phys_t), B_TRUE);
abd_zero(vp_abd, sizeof (vdev_phys_t));
vp = abd_to_buf(vp_abd);
buf = vp->vp_nvlist;
buflen = sizeof (vp->vp_nvlist);
if (!nvlist_pack(label, &buf, &buflen, NV_ENCODE_XDR, KM_SLEEP)) {
for (; l < VDEV_LABELS; l += 2) {
vdev_label_write(zio, vd, l, vp_abd,
offsetof(vdev_label_t, vl_vdev_phys),
sizeof (vdev_phys_t),
vdev_label_sync_done, good_writes,
flags | ZIO_FLAG_DONT_PROPAGATE);
}
}
abd_free(vp_abd);
nvlist_free(label);
}
static int
vdev_label_sync_list(spa_t *spa, int l, uint64_t txg, int flags)
{
list_t *dl = &spa->spa_config_dirty_list;
vdev_t *vd;
zio_t *zio;
int error;
/*
* Write the new labels to disk.
*/
zio = zio_root(spa, NULL, NULL, flags);
for (vd = list_head(dl); vd != NULL; vd = list_next(dl, vd)) {
uint64_t *good_writes;
ASSERT(!vd->vdev_ishole);
good_writes = kmem_zalloc(sizeof (uint64_t), KM_SLEEP);
zio_t *vio = zio_null(zio, spa, NULL,
(vd->vdev_islog || vd->vdev_aux != NULL) ?
vdev_label_sync_ignore_done : vdev_label_sync_top_done,
good_writes, flags);
vdev_label_sync(vio, good_writes, vd, l, txg, flags);
zio_nowait(vio);
}
error = zio_wait(zio);
/*
* Flush the new labels to disk.
*/
zio = zio_root(spa, NULL, NULL, flags);
for (vd = list_head(dl); vd != NULL; vd = list_next(dl, vd))
zio_flush(zio, vd);
(void) zio_wait(zio);
return (error);
}
/*
* Sync the uberblock and any changes to the vdev configuration.
*
* The order of operations is carefully crafted to ensure that
* if the system panics or loses power at any time, the state on disk
* is still transactionally consistent. The in-line comments below
* describe the failure semantics at each stage.
*
* Moreover, vdev_config_sync() is designed to be idempotent: if it fails
* at any time, you can just call it again, and it will resume its work.
*/
int
vdev_config_sync(vdev_t **svd, int svdcount, uint64_t txg)
{
spa_t *spa = svd[0]->vdev_spa;
uberblock_t *ub = &spa->spa_uberblock;
int error = 0;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL;
ASSERT(svdcount != 0);
retry:
/*
* Normally, we don't want to try too hard to write every label and
* uberblock. If there is a flaky disk, we don't want the rest of the
* sync process to block while we retry. But if we can't write a
* single label out, we should retry with ZIO_FLAG_TRYHARD before
* bailing out and declaring the pool faulted.
*/
if (error != 0) {
if ((flags & ZIO_FLAG_TRYHARD) != 0)
return (error);
flags |= ZIO_FLAG_TRYHARD;
}
ASSERT(ub->ub_txg <= txg);
/*
* If this isn't a resync due to I/O errors,
* and nothing changed in this transaction group,
* and the vdev configuration hasn't changed,
* then there's nothing to do.
*/
if (ub->ub_txg < txg) {
boolean_t changed = uberblock_update(ub, spa->spa_root_vdev,
txg, spa->spa_mmp.mmp_delay);
if (!changed && list_is_empty(&spa->spa_config_dirty_list))
return (0);
}
if (txg > spa_freeze_txg(spa))
return (0);
ASSERT(txg <= spa->spa_final_txg);
/*
* Flush the write cache of every disk that's been written to
* in this transaction group. This ensures that all blocks
* written in this txg will be committed to stable storage
* before any uberblock that references them.
*/
zio_t *zio = zio_root(spa, NULL, NULL, flags);
for (vdev_t *vd =
txg_list_head(&spa->spa_vdev_txg_list, TXG_CLEAN(txg)); vd != NULL;
vd = txg_list_next(&spa->spa_vdev_txg_list, vd, TXG_CLEAN(txg)))
zio_flush(zio, vd);
(void) zio_wait(zio);
/*
* Sync out the even labels (L0, L2) for every dirty vdev. If the
* system dies in the middle of this process, that's OK: all of the
* even labels that made it to disk will be newer than any uberblock,
* and will therefore be considered invalid. The odd labels (L1, L3),
* which have not yet been touched, will still be valid. We flush
* the new labels to disk to ensure that all even-label updates
* are committed to stable storage before the uberblock update.
*/
if ((error = vdev_label_sync_list(spa, 0, txg, flags)) != 0) {
if ((flags & ZIO_FLAG_TRYHARD) != 0) {
zfs_dbgmsg("vdev_label_sync_list() returned error %d "
"for pool '%s' when syncing out the even labels "
"of dirty vdevs", error, spa_name(spa));
}
goto retry;
}
/*
* Sync the uberblocks to all vdevs in svd[].
* If the system dies in the middle of this step, there are two cases
* to consider, and the on-disk state is consistent either way:
*
* (1) If none of the new uberblocks made it to disk, then the
* previous uberblock will be the newest, and the odd labels
* (which had not yet been touched) will be valid with respect
* to that uberblock.
*
* (2) If one or more new uberblocks made it to disk, then they
* will be the newest, and the even labels (which had all
* been successfully committed) will be valid with respect
* to the new uberblocks.
*/
if ((error = vdev_uberblock_sync_list(svd, svdcount, ub, flags)) != 0) {
if ((flags & ZIO_FLAG_TRYHARD) != 0) {
zfs_dbgmsg("vdev_uberblock_sync_list() returned error "
"%d for pool '%s'", error, spa_name(spa));
}
goto retry;
}
if (spa_multihost(spa))
mmp_update_uberblock(spa, ub);
/*
* Sync out odd labels for every dirty vdev. If the system dies
* in the middle of this process, the even labels and the new
* uberblocks will suffice to open the pool. The next time
* the pool is opened, the first thing we'll do -- before any
* user data is modified -- is mark every vdev dirty so that
* all labels will be brought up to date. We flush the new labels
* to disk to ensure that all odd-label updates are committed to
* stable storage before the next transaction group begins.
*/
if ((error = vdev_label_sync_list(spa, 1, txg, flags)) != 0) {
if ((flags & ZIO_FLAG_TRYHARD) != 0) {
zfs_dbgmsg("vdev_label_sync_list() returned error %d "
"for pool '%s' when syncing out the odd labels of "
"dirty vdevs", error, spa_name(spa));
}
goto retry;
}
return (0);
}
diff --git a/sys/contrib/openzfs/module/zfs/vdev_raidz.c b/sys/contrib/openzfs/module/zfs/vdev_raidz.c
index 3445fa9d35d5..9d0b8763f16f 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_raidz.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_raidz.c
@@ -1,2673 +1,5021 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2020 by Delphix. All rights reserved.
* Copyright (c) 2016 Gvozden Nešković. All rights reserved.
*/
#include <sys/zfs_context.h>
#include <sys/spa.h>
+#include <sys/spa_impl.h>
+#include <sys/zap.h>
#include <sys/vdev_impl.h>
+#include <sys/metaslab_impl.h>
#include <sys/zio.h>
#include <sys/zio_checksum.h>
+#include <sys/dmu_tx.h>
#include <sys/abd.h>
+#include <sys/zfs_rlock.h>
#include <sys/fs/zfs.h>
#include <sys/fm/fs/zfs.h>
#include <sys/vdev_raidz.h>
#include <sys/vdev_raidz_impl.h>
#include <sys/vdev_draid.h>
+#include <sys/uberblock_impl.h>
+#include <sys/dsl_scan.h>
#ifdef ZFS_DEBUG
#include <sys/vdev.h> /* For vdev_xlate() in vdev_raidz_io_verify() */
#endif
/*
* Virtual device vector for RAID-Z.
*
* This vdev supports single, double, and triple parity. For single parity,
* we use a simple XOR of all the data columns. For double or triple parity,
* we use a special case of Reed-Solomon coding. This extends the
* technique described in "The mathematics of RAID-6" by H. Peter Anvin by
* drawing on the system described in "A Tutorial on Reed-Solomon Coding for
* Fault-Tolerance in RAID-like Systems" by James S. Plank on which the
* former is also based. The latter is designed to provide higher performance
* for writes.
*
* Note that the Plank paper claimed to support arbitrary N+M, but was then
* amended six years later identifying a critical flaw that invalidates its
* claims. Nevertheless, the technique can be adapted to work for up to
* triple parity. For additional parity, the amendment "Note: Correction to
* the 1997 Tutorial on Reed-Solomon Coding" by James S. Plank and Ying Ding
* is viable, but the additional complexity means that write performance will
* suffer.
*
* All of the methods above operate on a Galois field, defined over the
* integers mod 2^N. In our case we choose N=8 for GF(8) so that all elements
* can be expressed with a single byte. Briefly, the operations on the
* field are defined as follows:
*
* o addition (+) is represented by a bitwise XOR
* o subtraction (-) is therefore identical to addition: A + B = A - B
* o multiplication of A by 2 is defined by the following bitwise expression:
*
* (A * 2)_7 = A_6
* (A * 2)_6 = A_5
* (A * 2)_5 = A_4
* (A * 2)_4 = A_3 + A_7
* (A * 2)_3 = A_2 + A_7
* (A * 2)_2 = A_1 + A_7
* (A * 2)_1 = A_0
* (A * 2)_0 = A_7
*
* In C, multiplying by 2 is therefore ((a << 1) ^ ((a & 0x80) ? 0x1d : 0)).
* As an aside, this multiplication is derived from the error correcting
* primitive polynomial x^8 + x^4 + x^3 + x^2 + 1.
*
* Observe that any number in the field (except for 0) can be expressed as a
* power of 2 -- a generator for the field. We store a table of the powers of
* 2 and logs base 2 for quick look ups, and exploit the fact that A * B can
* be rewritten as 2^(log_2(A) + log_2(B)) (where '+' is normal addition rather
* than field addition). The inverse of a field element A (A^-1) is therefore
* A ^ (255 - 1) = A^254.
*
* The up-to-three parity columns, P, Q, R over several data columns,
* D_0, ... D_n-1, can be expressed by field operations:
*
* P = D_0 + D_1 + ... + D_n-2 + D_n-1
* Q = 2^n-1 * D_0 + 2^n-2 * D_1 + ... + 2^1 * D_n-2 + 2^0 * D_n-1
* = ((...((D_0) * 2 + D_1) * 2 + ...) * 2 + D_n-2) * 2 + D_n-1
* R = 4^n-1 * D_0 + 4^n-2 * D_1 + ... + 4^1 * D_n-2 + 4^0 * D_n-1
* = ((...((D_0) * 4 + D_1) * 4 + ...) * 4 + D_n-2) * 4 + D_n-1
*
* We chose 1, 2, and 4 as our generators because 1 corresponds to the trivial
* XOR operation, and 2 and 4 can be computed quickly and generate linearly-
* independent coefficients. (There are no additional coefficients that have
* this property which is why the uncorrected Plank method breaks down.)
*
* See the reconstruction code below for how P, Q and R can used individually
* or in concert to recover missing data columns.
*/
#define VDEV_RAIDZ_P 0
#define VDEV_RAIDZ_Q 1
#define VDEV_RAIDZ_R 2
#define VDEV_RAIDZ_MUL_2(x) (((x) << 1) ^ (((x) & 0x80) ? 0x1d : 0))
#define VDEV_RAIDZ_MUL_4(x) (VDEV_RAIDZ_MUL_2(VDEV_RAIDZ_MUL_2(x)))
/*
* We provide a mechanism to perform the field multiplication operation on a
* 64-bit value all at once rather than a byte at a time. This works by
* creating a mask from the top bit in each byte and using that to
* conditionally apply the XOR of 0x1d.
*/
#define VDEV_RAIDZ_64MUL_2(x, mask) \
{ \
(mask) = (x) & 0x8080808080808080ULL; \
(mask) = ((mask) << 1) - ((mask) >> 7); \
(x) = (((x) << 1) & 0xfefefefefefefefeULL) ^ \
((mask) & 0x1d1d1d1d1d1d1d1dULL); \
}
#define VDEV_RAIDZ_64MUL_4(x, mask) \
{ \
VDEV_RAIDZ_64MUL_2((x), mask); \
VDEV_RAIDZ_64MUL_2((x), mask); \
}
+
+/*
+ * Big Theory Statement for how a RAIDZ VDEV is expanded
+ *
+ * An existing RAIDZ VDEV can be expanded by attaching a new disk. Expansion
+ * works with all three RAIDZ parity choices, including RAIDZ1, 2, or 3. VDEVs
+ * that have been previously expanded can be expanded again.
+ *
+ * The RAIDZ VDEV must be healthy (must be able to write to all the drives in
+ * the VDEV) when an expansion starts. And the expansion will pause if any
+ * disk in the VDEV fails, and resume once the VDEV is healthy again. All other
+ * operations on the pool can continue while an expansion is in progress (e.g.
+ * read/write, snapshot, zpool add, etc). Except zpool checkpoint, zpool trim,
+ * and zpool initialize which can't be run during an expansion. Following a
+ * reboot or export/import, the expansion resumes where it left off.
+ *
+ * == Reflowing the Data ==
+ *
+ * The expansion involves reflowing (copying) the data from the current set
+ * of disks to spread it across the new set which now has one more disk. This
+ * reflow operation is similar to reflowing text when the column width of a
+ * text editor window is expanded. The text doesn’t change but the location of
+ * the text changes to accommodate the new width. An example reflow result for
+ * a 4-wide RAIDZ1 to a 5-wide is shown below.
+ *
+ * Reflow End State
+ * Each letter indicates a parity group (logical stripe)
+ *
+ * Before expansion After Expansion
+ * D1 D2 D3 D4 D1 D2 D3 D4 D5
+ * +------+------+------+------+ +------+------+------+------+------+
+ * | | | | | | | | | | |
+ * | A | A | A | A | | A | A | A | A | B |
+ * | 1| 2| 3| 4| | 1| 2| 3| 4| 5|
+ * +------+------+------+------+ +------+------+------+------+------+
+ * | | | | | | | | | | |
+ * | B | B | C | C | | B | C | C | C | C |
+ * | 5| 6| 7| 8| | 6| 7| 8| 9| 10|
+ * +------+------+------+------+ +------+------+------+------+------+
+ * | | | | | | | | | | |
+ * | C | C | D | D | | D | D | E | E | E |
+ * | 9| 10| 11| 12| | 11| 12| 13| 14| 15|
+ * +------+------+------+------+ +------+------+------+------+------+
+ * | | | | | | | | | | |
+ * | E | E | E | E | --> | E | F | F | G | G |
+ * | 13| 14| 15| 16| | 16| 17| 18|p 19| 20|
+ * +------+------+------+------+ +------+------+------+------+------+
+ * | | | | | | | | | | |
+ * | F | F | G | G | | G | G | H | H | H |
+ * | 17| 18| 19| 20| | 21| 22| 23| 24| 25|
+ * +------+------+------+------+ +------+------+------+------+------+
+ * | | | | | | | | | | |
+ * | G | G | H | H | | H | I | I | J | J |
+ * | 21| 22| 23| 24| | 26| 27| 28| 29| 30|
+ * +------+------+------+------+ +------+------+------+------+------+
+ * | | | | | | | | | | |
+ * | H | H | I | I | | J | J | | | K |
+ * | 25| 26| 27| 28| | 31| 32| 33| 34| 35|
+ * +------+------+------+------+ +------+------+------+------+------+
+ *
+ * This reflow approach has several advantages. There is no need to read or
+ * modify the block pointers or recompute any block checksums. The reflow
+ * doesn’t need to know where the parity sectors reside. We can read and write
+ * data sequentially and the copy can occur in a background thread in open
+ * context. The design also allows for fast discovery of what data to copy.
+ *
+ * The VDEV metaslabs are processed, one at a time, to copy the block data to
+ * have it flow across all the disks. The metaslab is disabled for allocations
+ * during the copy. As an optimization, we only copy the allocated data which
+ * can be determined by looking at the metaslab range tree. During the copy we
+ * must maintain the redundancy guarantees of the RAIDZ VDEV (i.e., we still
+ * need to be able to survive losing parity count disks). This means we
+ * cannot overwrite data during the reflow that would be needed if a disk is
+ * lost.
+ *
+ * After the reflow completes, all newly-written blocks will have the new
+ * layout, i.e., they will have the parity to data ratio implied by the new
+ * number of disks in the RAIDZ group. Even though the reflow copies all of
+ * the allocated space (data and parity), it is only rearranged, not changed.
+ *
+ * This act of reflowing the data has a few implications about blocks
+ * that were written before the reflow completes:
+ *
+ * - Old blocks will still use the same amount of space (i.e., they will have
+ * the parity to data ratio implied by the old number of disks in the RAIDZ
+ * group).
+ * - Reading old blocks will be slightly slower than before the reflow, for
+ * two reasons. First, we will have to read from all disks in the RAIDZ
+ * VDEV, rather than being able to skip the children that contain only
+ * parity of this block (because the data of a single block is now spread
+ * out across all the disks). Second, in most cases there will be an extra
+ * bcopy, needed to rearrange the data back to its original layout in memory.
+ *
+ * == Scratch Area ==
+ *
+ * As we copy the block data, we can only progress to the point that writes
+ * will not overlap with blocks whose progress has not yet been recorded on
+ * disk. Since partially-copied rows are always read from the old location,
+ * we need to stop one row before the sector-wise overlap, to prevent any
+ * row-wise overlap. For example, in the diagram above, when we reflow sector
+ * B6 it will overwite the original location for B5.
+ *
+ * To get around this, a scratch space is used so that we can start copying
+ * without risking data loss by overlapping the row. As an added benefit, it
+ * improves performance at the beginning of the reflow, but that small perf
+ * boost wouldn't be worth the complexity on its own.
+ *
+ * Ideally we want to copy at least 2 * (new_width)^2 so that we have a
+ * separation of 2*(new_width+1) and a chunk size of new_width+2. With the max
+ * RAIDZ width of 255 and 4K sectors this would be 2MB per disk. In practice
+ * the widths will likely be single digits so we can get a substantial chuck
+ * size using only a few MB of scratch per disk.
+ *
+ * The scratch area is persisted to disk which holds a large amount of reflowed
+ * state. We can always read the partially written stripes when a disk fails or
+ * the copy is interrupted (crash) during the initial copying phase and also
+ * get past a small chunk size restriction. At a minimum, the scratch space
+ * must be large enough to get us to the point that one row does not overlap
+ * itself when moved (i.e new_width^2). But going larger is even better. We
+ * use the 3.5 MiB reserved "boot" space that resides after the ZFS disk labels
+ * as our scratch space to handle overwriting the initial part of the VDEV.
+ *
+ * 0 256K 512K 4M
+ * +------+------+-----------------------+-----------------------------
+ * | VDEV | VDEV | Boot Block (3.5M) | Allocatable space ...
+ * | L0 | L1 | Reserved | (Metaslabs)
+ * +------+------+-----------------------+-------------------------------
+ * Scratch Area
+ *
+ * == Reflow Progress Updates ==
+ * After the initial scratch-based reflow, the expansion process works
+ * similarly to device removal. We create a new open context thread which
+ * reflows the data, and periodically kicks off sync tasks to update logical
+ * state. In this case, state is the committed progress (offset of next data
+ * to copy). We need to persist the completed offset on disk, so that if we
+ * crash we know which format each VDEV offset is in.
+ *
+ * == Time Dependent Geometry ==
+ *
+ * In non-expanded RAIDZ, blocks are read from disk in a column by column
+ * fashion. For a multi-row block, the second sector is in the first column
+ * not in the second column. This allows us to issue full reads for each
+ * column directly into the request buffer. The block data is thus laid out
+ * sequentially in a column-by-column fashion.
+ *
+ * For example, in the before expansion diagram above, one logical block might
+ * be sectors G19-H26. The parity is in G19,H23; and the data is in
+ * G20,H24,G21,H25,G22,H26.
+ *
+ * After a block is reflowed, the sectors that were all in the original column
+ * data can now reside in different columns. When reading from an expanded
+ * VDEV, we need to know the logical stripe width for each block so we can
+ * reconstitute the block’s data after the reads are completed. Likewise,
+ * when we perform the combinatorial reconstruction we need to know the
+ * original width so we can retry combinations from the past layouts.
+ *
+ * Time dependent geometry is what we call having blocks with different layouts
+ * (stripe widths) in the same VDEV. This time-dependent geometry uses the
+ * block’s birth time (+ the time expansion ended) to establish the correct
+ * width for a given block. After an expansion completes, we record the time
+ * for blocks written with a particular width (geometry).
+ *
+ * == On Disk Format Changes ==
+ *
+ * New pool feature flag, 'raidz_expansion' whose reference count is the number
+ * of RAIDZ VDEVs that have been expanded.
+ *
+ * The blocks on expanded RAIDZ VDEV can have different logical stripe widths.
+ *
+ * Since the uberblock can point to arbitrary blocks, which might be on the
+ * expanding RAIDZ, and might or might not have been expanded. We need to know
+ * which way a block is laid out before reading it. This info is the next
+ * offset that needs to be reflowed and we persist that in the uberblock, in
+ * the new ub_raidz_reflow_info field, as opposed to the MOS or the vdev label.
+ * After the expansion is complete, we then use the raidz_expand_txgs array
+ * (see below) to determine how to read a block and the ub_raidz_reflow_info
+ * field no longer required.
+ *
+ * The uberblock's ub_raidz_reflow_info field also holds the scratch space
+ * state (i.e., active or not) which is also required before reading a block
+ * during the initial phase of reflowing the data.
+ *
+ * The top-level RAIDZ VDEV has two new entries in the nvlist:
+ *
+ * 'raidz_expand_txgs' array: logical stripe widths by txg are recorded here
+ * and used after the expansion is complete to
+ * determine how to read a raidz block
+ * 'raidz_expanding' boolean: present during reflow and removed after completion
+ * used during a spa import to resume an unfinished
+ * expansion
+ *
+ * And finally the VDEVs top zap adds the following informational entries:
+ * VDEV_TOP_ZAP_RAIDZ_EXPAND_STATE
+ * VDEV_TOP_ZAP_RAIDZ_EXPAND_START_TIME
+ * VDEV_TOP_ZAP_RAIDZ_EXPAND_END_TIME
+ * VDEV_TOP_ZAP_RAIDZ_EXPAND_BYTES_COPIED
+ */
+
+/*
+ * For testing only: pause the raidz expansion after reflowing this amount.
+ * (accessed by ZTS and ztest)
+ */
+#ifdef _KERNEL
+static
+#endif /* _KERNEL */
+unsigned long raidz_expand_max_reflow_bytes = 0;
+
+/*
+ * For testing only: pause the raidz expansion at a certain point.
+ */
+uint_t raidz_expand_pause_point = 0;
+
+/*
+ * Maximum amount of copy io's outstanding at once.
+ */
+static unsigned long raidz_expand_max_copy_bytes = 10 * SPA_MAXBLOCKSIZE;
+
+/*
+ * Apply raidz map abds aggregation if the number of rows in the map is equal
+ * or greater than the value below.
+ */
+static unsigned long raidz_io_aggregate_rows = 4;
+
+/*
+ * Automatically start a pool scrub when a RAIDZ expansion completes in
+ * order to verify the checksums of all blocks which have been copied
+ * during the expansion. Automatic scrubbing is enabled by default and
+ * is strongly recommended.
+ */
+static int zfs_scrub_after_expand = 1;
+
static void
vdev_raidz_row_free(raidz_row_t *rr)
{
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_size != 0)
abd_free(rc->rc_abd);
if (rc->rc_orig_data != NULL)
abd_free(rc->rc_orig_data);
}
if (rr->rr_abd_empty != NULL)
abd_free(rr->rr_abd_empty);
kmem_free(rr, offsetof(raidz_row_t, rr_col[rr->rr_scols]));
}
void
vdev_raidz_map_free(raidz_map_t *rm)
{
for (int i = 0; i < rm->rm_nrows; i++)
vdev_raidz_row_free(rm->rm_row[i]);
+ if (rm->rm_nphys_cols) {
+ for (int i = 0; i < rm->rm_nphys_cols; i++) {
+ if (rm->rm_phys_col[i].rc_abd != NULL)
+ abd_free(rm->rm_phys_col[i].rc_abd);
+ }
+
+ kmem_free(rm->rm_phys_col, sizeof (raidz_col_t) *
+ rm->rm_nphys_cols);
+ }
+
+ ASSERT3P(rm->rm_lr, ==, NULL);
kmem_free(rm, offsetof(raidz_map_t, rm_row[rm->rm_nrows]));
}
static void
vdev_raidz_map_free_vsd(zio_t *zio)
{
raidz_map_t *rm = zio->io_vsd;
vdev_raidz_map_free(rm);
}
+static int
+vdev_raidz_reflow_compare(const void *x1, const void *x2)
+{
+ const reflow_node_t *l = x1;
+ const reflow_node_t *r = x2;
+
+ return (TREE_CMP(l->re_txg, r->re_txg));
+}
+
const zio_vsd_ops_t vdev_raidz_vsd_ops = {
.vsd_free = vdev_raidz_map_free_vsd,
};
+raidz_row_t *
+vdev_raidz_row_alloc(int cols)
+{
+ raidz_row_t *rr =
+ kmem_zalloc(offsetof(raidz_row_t, rr_col[cols]), KM_SLEEP);
+
+ rr->rr_cols = cols;
+ rr->rr_scols = cols;
+
+ for (int c = 0; c < cols; c++) {
+ raidz_col_t *rc = &rr->rr_col[c];
+ rc->rc_shadow_devidx = INT_MAX;
+ rc->rc_shadow_offset = UINT64_MAX;
+ rc->rc_allow_repair = 1;
+ }
+ return (rr);
+}
+
static void
vdev_raidz_map_alloc_write(zio_t *zio, raidz_map_t *rm, uint64_t ashift)
{
int c;
int nwrapped = 0;
uint64_t off = 0;
raidz_row_t *rr = rm->rm_row[0];
ASSERT3U(zio->io_type, ==, ZIO_TYPE_WRITE);
ASSERT3U(rm->rm_nrows, ==, 1);
/*
* Pad any parity columns with additional space to account for skip
* sectors.
*/
if (rm->rm_skipstart < rr->rr_firstdatacol) {
ASSERT0(rm->rm_skipstart);
nwrapped = rm->rm_nskip;
} else if (rr->rr_scols < (rm->rm_skipstart + rm->rm_nskip)) {
nwrapped =
(rm->rm_skipstart + rm->rm_nskip) % rr->rr_scols;
}
/*
* Optional single skip sectors (rc_size == 0) will be handled in
* vdev_raidz_io_start_write().
*/
int skipped = rr->rr_scols - rr->rr_cols;
/* Allocate buffers for the parity columns */
for (c = 0; c < rr->rr_firstdatacol; c++) {
raidz_col_t *rc = &rr->rr_col[c];
/*
* Parity columns will pad out a linear ABD to account for
* the skip sector. A linear ABD is used here because
* parity calculations use the ABD buffer directly to calculate
* parity. This avoids doing a memcpy back to the ABD after the
* parity has been calculated. By issuing the parity column
* with the skip sector we can reduce contention on the child
* VDEV queue locks (vq_lock).
*/
if (c < nwrapped) {
rc->rc_abd = abd_alloc_linear(
rc->rc_size + (1ULL << ashift), B_FALSE);
abd_zero_off(rc->rc_abd, rc->rc_size, 1ULL << ashift);
skipped++;
} else {
rc->rc_abd = abd_alloc_linear(rc->rc_size, B_FALSE);
}
}
for (off = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
abd_t *abd = abd_get_offset_struct(&rc->rc_abdstruct,
zio->io_abd, off, rc->rc_size);
/*
* Generate I/O for skip sectors to improve aggregation
* continuity. We will use gang ABD's to reduce contention
* on the child VDEV queue locks (vq_lock) by issuing
* a single I/O that contains the data and skip sector.
*
* It is important to make sure that rc_size is not updated
* even though we are adding a skip sector to the ABD. When
* calculating the parity in vdev_raidz_generate_parity_row()
* the rc_size is used to iterate through the ABD's. We can
* not have zero'd out skip sectors used for calculating
* parity for raidz, because those same sectors are not used
* during reconstruction.
*/
if (c >= rm->rm_skipstart && skipped < rm->rm_nskip) {
rc->rc_abd = abd_alloc_gang();
abd_gang_add(rc->rc_abd, abd, B_TRUE);
abd_gang_add(rc->rc_abd,
abd_get_zeros(1ULL << ashift), B_TRUE);
skipped++;
} else {
rc->rc_abd = abd;
}
off += rc->rc_size;
}
ASSERT3U(off, ==, zio->io_size);
ASSERT3S(skipped, ==, rm->rm_nskip);
}
static void
vdev_raidz_map_alloc_read(zio_t *zio, raidz_map_t *rm)
{
int c;
raidz_row_t *rr = rm->rm_row[0];
ASSERT3U(rm->rm_nrows, ==, 1);
/* Allocate buffers for the parity columns */
for (c = 0; c < rr->rr_firstdatacol; c++)
rr->rr_col[c].rc_abd =
abd_alloc_linear(rr->rr_col[c].rc_size, B_FALSE);
for (uint64_t off = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
rc->rc_abd = abd_get_offset_struct(&rc->rc_abdstruct,
zio->io_abd, off, rc->rc_size);
off += rc->rc_size;
}
}
/*
* Divides the IO evenly across all child vdevs; usually, dcols is
* the number of children in the target vdev.
*
* Avoid inlining the function to keep vdev_raidz_io_start(), which
* is this functions only caller, as small as possible on the stack.
*/
noinline raidz_map_t *
vdev_raidz_map_alloc(zio_t *zio, uint64_t ashift, uint64_t dcols,
uint64_t nparity)
{
raidz_row_t *rr;
/* The starting RAIDZ (parent) vdev sector of the block. */
uint64_t b = zio->io_offset >> ashift;
/* The zio's size in units of the vdev's minimum sector size. */
uint64_t s = zio->io_size >> ashift;
/* The first column for this stripe. */
uint64_t f = b % dcols;
/* The starting byte offset on each child vdev. */
uint64_t o = (b / dcols) << ashift;
- uint64_t q, r, c, bc, col, acols, scols, coff, devidx, asize, tot;
+ uint64_t acols, scols;
raidz_map_t *rm =
kmem_zalloc(offsetof(raidz_map_t, rm_row[1]), KM_SLEEP);
rm->rm_nrows = 1;
/*
* "Quotient": The number of data sectors for this stripe on all but
* the "big column" child vdevs that also contain "remainder" data.
*/
- q = s / (dcols - nparity);
+ uint64_t q = s / (dcols - nparity);
/*
* "Remainder": The number of partial stripe data sectors in this I/O.
* This will add a sector to some, but not all, child vdevs.
*/
- r = s - q * (dcols - nparity);
+ uint64_t r = s - q * (dcols - nparity);
/* The number of "big columns" - those which contain remainder data. */
- bc = (r == 0 ? 0 : r + nparity);
+ uint64_t bc = (r == 0 ? 0 : r + nparity);
/*
* The total number of data and parity sectors associated with
* this I/O.
*/
- tot = s + nparity * (q + (r == 0 ? 0 : 1));
+ uint64_t tot = s + nparity * (q + (r == 0 ? 0 : 1));
/*
* acols: The columns that will be accessed.
* scols: The columns that will be accessed or skipped.
*/
if (q == 0) {
/* Our I/O request doesn't span all child vdevs. */
acols = bc;
scols = MIN(dcols, roundup(bc, nparity + 1));
} else {
acols = dcols;
scols = dcols;
}
ASSERT3U(acols, <=, scols);
-
- rr = kmem_alloc(offsetof(raidz_row_t, rr_col[scols]), KM_SLEEP);
+ rr = vdev_raidz_row_alloc(scols);
rm->rm_row[0] = rr;
-
rr->rr_cols = acols;
- rr->rr_scols = scols;
rr->rr_bigcols = bc;
- rr->rr_missingdata = 0;
- rr->rr_missingparity = 0;
rr->rr_firstdatacol = nparity;
- rr->rr_abd_empty = NULL;
- rr->rr_nempty = 0;
#ifdef ZFS_DEBUG
rr->rr_offset = zio->io_offset;
rr->rr_size = zio->io_size;
#endif
- asize = 0;
+ uint64_t asize = 0;
- for (c = 0; c < scols; c++) {
+ for (uint64_t c = 0; c < scols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
- col = f + c;
- coff = o;
+ uint64_t col = f + c;
+ uint64_t coff = o;
if (col >= dcols) {
col -= dcols;
coff += 1ULL << ashift;
}
rc->rc_devidx = col;
rc->rc_offset = coff;
- rc->rc_abd = NULL;
- rc->rc_orig_data = NULL;
- rc->rc_error = 0;
- rc->rc_tried = 0;
- rc->rc_skipped = 0;
- rc->rc_force_repair = 0;
- rc->rc_allow_repair = 1;
- rc->rc_need_orig_restore = B_FALSE;
if (c >= acols)
rc->rc_size = 0;
else if (c < bc)
rc->rc_size = (q + 1) << ashift;
else
rc->rc_size = q << ashift;
asize += rc->rc_size;
}
ASSERT3U(asize, ==, tot << ashift);
rm->rm_nskip = roundup(tot, nparity + 1) - tot;
rm->rm_skipstart = bc;
/*
* If all data stored spans all columns, there's a danger that parity
* will always be on the same device and, since parity isn't read
* during normal operation, that device's I/O bandwidth won't be
* used effectively. We therefore switch the parity every 1MB.
*
* ... at least that was, ostensibly, the theory. As a practical
* matter unless we juggle the parity between all devices evenly, we
* won't see any benefit. Further, occasional writes that aren't a
* multiple of the LCM of the number of children and the minimum
* stripe width are sufficient to avoid pessimal behavior.
* Unfortunately, this decision created an implicit on-disk format
* requirement that we need to support for all eternity, but only
* for single-parity RAID-Z.
*
* If we intend to skip a sector in the zeroth column for padding
* we must make sure to note this swap. We will never intend to
* skip the first column since at least one data and one parity
* column must appear in each row.
*/
ASSERT(rr->rr_cols >= 2);
ASSERT(rr->rr_col[0].rc_size == rr->rr_col[1].rc_size);
if (rr->rr_firstdatacol == 1 && (zio->io_offset & (1ULL << 20))) {
- devidx = rr->rr_col[0].rc_devidx;
+ uint64_t devidx = rr->rr_col[0].rc_devidx;
o = rr->rr_col[0].rc_offset;
rr->rr_col[0].rc_devidx = rr->rr_col[1].rc_devidx;
rr->rr_col[0].rc_offset = rr->rr_col[1].rc_offset;
rr->rr_col[1].rc_devidx = devidx;
rr->rr_col[1].rc_offset = o;
-
if (rm->rm_skipstart == 0)
rm->rm_skipstart = 1;
}
if (zio->io_type == ZIO_TYPE_WRITE) {
vdev_raidz_map_alloc_write(zio, rm, ashift);
} else {
vdev_raidz_map_alloc_read(zio, rm);
}
+ /* init RAIDZ parity ops */
+ rm->rm_ops = vdev_raidz_math_get_ops();
+
+ return (rm);
+}
+
+/*
+ * Everything before reflow_offset_synced should have been moved to the new
+ * location (read and write completed). However, this may not yet be reflected
+ * in the on-disk format (e.g. raidz_reflow_sync() has been called but the
+ * uberblock has not yet been written). If reflow is not in progress,
+ * reflow_offset_synced should be UINT64_MAX. For each row, if the row is
+ * entirely before reflow_offset_synced, it will come from the new location.
+ * Otherwise this row will come from the old location. Therefore, rows that
+ * straddle the reflow_offset_synced will come from the old location.
+ *
+ * For writes, reflow_offset_next is the next offset to copy. If a sector has
+ * been copied, but not yet reflected in the on-disk progress
+ * (reflow_offset_synced), it will also be written to the new (already copied)
+ * offset.
+ */
+noinline raidz_map_t *
+vdev_raidz_map_alloc_expanded(zio_t *zio,
+ uint64_t ashift, uint64_t physical_cols, uint64_t logical_cols,
+ uint64_t nparity, uint64_t reflow_offset_synced,
+ uint64_t reflow_offset_next, boolean_t use_scratch)
+{
+ abd_t *abd = zio->io_abd;
+ uint64_t offset = zio->io_offset;
+ uint64_t size = zio->io_size;
+
+ /* The zio's size in units of the vdev's minimum sector size. */
+ uint64_t s = size >> ashift;
+
+ /*
+ * "Quotient": The number of data sectors for this stripe on all but
+ * the "big column" child vdevs that also contain "remainder" data.
+ * AKA "full rows"
+ */
+ uint64_t q = s / (logical_cols - nparity);
+
+ /*
+ * "Remainder": The number of partial stripe data sectors in this I/O.
+ * This will add a sector to some, but not all, child vdevs.
+ */
+ uint64_t r = s - q * (logical_cols - nparity);
+
+ /* The number of "big columns" - those which contain remainder data. */
+ uint64_t bc = (r == 0 ? 0 : r + nparity);
+
+ /*
+ * The total number of data and parity sectors associated with
+ * this I/O.
+ */
+ uint64_t tot = s + nparity * (q + (r == 0 ? 0 : 1));
+
+ /* How many rows contain data (not skip) */
+ uint64_t rows = howmany(tot, logical_cols);
+ int cols = MIN(tot, logical_cols);
+
+ raidz_map_t *rm =
+ kmem_zalloc(offsetof(raidz_map_t, rm_row[rows]),
+ KM_SLEEP);
+ rm->rm_nrows = rows;
+ rm->rm_nskip = roundup(tot, nparity + 1) - tot;
+ rm->rm_skipstart = bc;
+ uint64_t asize = 0;
+
+ for (uint64_t row = 0; row < rows; row++) {
+ boolean_t row_use_scratch = B_FALSE;
+ raidz_row_t *rr = vdev_raidz_row_alloc(cols);
+ rm->rm_row[row] = rr;
+
+ /* The starting RAIDZ (parent) vdev sector of the row. */
+ uint64_t b = (offset >> ashift) + row * logical_cols;
+
+ /*
+ * If we are in the middle of a reflow, and the copying has
+ * not yet completed for any part of this row, then use the
+ * old location of this row. Note that reflow_offset_synced
+ * reflects the i/o that's been completed, because it's
+ * updated by a synctask, after zio_wait(spa_txg_zio[]).
+ * This is sufficient for our check, even if that progress
+ * has not yet been recorded to disk (reflected in
+ * spa_ubsync). Also note that we consider the last row to
+ * be "full width" (`cols`-wide rather than `bc`-wide) for
+ * this calculation. This causes a tiny bit of unnecessary
+ * double-writes but is safe and simpler to calculate.
+ */
+ int row_phys_cols = physical_cols;
+ if (b + cols > reflow_offset_synced >> ashift)
+ row_phys_cols--;
+ else if (use_scratch)
+ row_use_scratch = B_TRUE;
+
+ /* starting child of this row */
+ uint64_t child_id = b % row_phys_cols;
+ /* The starting byte offset on each child vdev. */
+ uint64_t child_offset = (b / row_phys_cols) << ashift;
+
+ /*
+ * Note, rr_cols is the entire width of the block, even
+ * if this row is shorter. This is needed because parity
+ * generation (for Q and R) needs to know the entire width,
+ * because it treats the short row as though it was
+ * full-width (and the "phantom" sectors were zero-filled).
+ *
+ * Another approach to this would be to set cols shorter
+ * (to just the number of columns that we might do i/o to)
+ * and have another mechanism to tell the parity generation
+ * about the "entire width". Reconstruction (at least
+ * vdev_raidz_reconstruct_general()) would also need to
+ * know about the "entire width".
+ */
+ rr->rr_firstdatacol = nparity;
+#ifdef ZFS_DEBUG
+ /*
+ * note: rr_size is PSIZE, not ASIZE
+ */
+ rr->rr_offset = b << ashift;
+ rr->rr_size = (rr->rr_cols - rr->rr_firstdatacol) << ashift;
+#endif
+
+ for (int c = 0; c < rr->rr_cols; c++, child_id++) {
+ if (child_id >= row_phys_cols) {
+ child_id -= row_phys_cols;
+ child_offset += 1ULL << ashift;
+ }
+ raidz_col_t *rc = &rr->rr_col[c];
+ rc->rc_devidx = child_id;
+ rc->rc_offset = child_offset;
+
+ /*
+ * Get this from the scratch space if appropriate.
+ * This only happens if we crashed in the middle of
+ * raidz_reflow_scratch_sync() (while it's running,
+ * the rangelock prevents us from doing concurrent
+ * io), and even then only during zpool import or
+ * when the pool is imported readonly.
+ */
+ if (row_use_scratch)
+ rc->rc_offset -= VDEV_BOOT_SIZE;
+
+ uint64_t dc = c - rr->rr_firstdatacol;
+ if (c < rr->rr_firstdatacol) {
+ rc->rc_size = 1ULL << ashift;
+
+ /*
+ * Parity sectors' rc_abd's are set below
+ * after determining if this is an aggregation.
+ */
+ } else if (row == rows - 1 && bc != 0 && c >= bc) {
+ /*
+ * Past the end of the block (even including
+ * skip sectors). This sector is part of the
+ * map so that we have full rows for p/q parity
+ * generation.
+ */
+ rc->rc_size = 0;
+ rc->rc_abd = NULL;
+ } else {
+ /* "data column" (col excluding parity) */
+ uint64_t off;
+
+ if (c < bc || r == 0) {
+ off = dc * rows + row;
+ } else {
+ off = r * rows +
+ (dc - r) * (rows - 1) + row;
+ }
+ rc->rc_size = 1ULL << ashift;
+ rc->rc_abd = abd_get_offset_struct(
+ &rc->rc_abdstruct, abd, off << ashift,
+ rc->rc_size);
+ }
+
+ if (rc->rc_size == 0)
+ continue;
+
+ /*
+ * If any part of this row is in both old and new
+ * locations, the primary location is the old
+ * location. If this sector was already copied to the
+ * new location, we need to also write to the new,
+ * "shadow" location.
+ *
+ * Note, `row_phys_cols != physical_cols` indicates
+ * that the primary location is the old location.
+ * `b+c < reflow_offset_next` indicates that the copy
+ * to the new location has been initiated. We know
+ * that the copy has completed because we have the
+ * rangelock, which is held exclusively while the
+ * copy is in progress.
+ */
+ if (row_use_scratch ||
+ (row_phys_cols != physical_cols &&
+ b + c < reflow_offset_next >> ashift)) {
+ rc->rc_shadow_devidx = (b + c) % physical_cols;
+ rc->rc_shadow_offset =
+ ((b + c) / physical_cols) << ashift;
+ if (row_use_scratch)
+ rc->rc_shadow_offset -= VDEV_BOOT_SIZE;
+ }
+
+ asize += rc->rc_size;
+ }
+
+ /*
+ * See comment in vdev_raidz_map_alloc()
+ */
+ if (rr->rr_firstdatacol == 1 && rr->rr_cols > 1 &&
+ (offset & (1ULL << 20))) {
+ ASSERT(rr->rr_cols >= 2);
+ ASSERT(rr->rr_col[0].rc_size == rr->rr_col[1].rc_size);
+
+ int devidx0 = rr->rr_col[0].rc_devidx;
+ uint64_t offset0 = rr->rr_col[0].rc_offset;
+ int shadow_devidx0 = rr->rr_col[0].rc_shadow_devidx;
+ uint64_t shadow_offset0 =
+ rr->rr_col[0].rc_shadow_offset;
+
+ rr->rr_col[0].rc_devidx = rr->rr_col[1].rc_devidx;
+ rr->rr_col[0].rc_offset = rr->rr_col[1].rc_offset;
+ rr->rr_col[0].rc_shadow_devidx =
+ rr->rr_col[1].rc_shadow_devidx;
+ rr->rr_col[0].rc_shadow_offset =
+ rr->rr_col[1].rc_shadow_offset;
+
+ rr->rr_col[1].rc_devidx = devidx0;
+ rr->rr_col[1].rc_offset = offset0;
+ rr->rr_col[1].rc_shadow_devidx = shadow_devidx0;
+ rr->rr_col[1].rc_shadow_offset = shadow_offset0;
+ }
+ }
+ ASSERT3U(asize, ==, tot << ashift);
+
+ /*
+ * Determine if the block is contiguous, in which case we can use
+ * an aggregation.
+ */
+ if (rows >= raidz_io_aggregate_rows) {
+ rm->rm_nphys_cols = physical_cols;
+ rm->rm_phys_col =
+ kmem_zalloc(sizeof (raidz_col_t) * rm->rm_nphys_cols,
+ KM_SLEEP);
+
+ /*
+ * Determine the aggregate io's offset and size, and check
+ * that the io is contiguous.
+ */
+ for (int i = 0;
+ i < rm->rm_nrows && rm->rm_phys_col != NULL; i++) {
+ raidz_row_t *rr = rm->rm_row[i];
+ for (int c = 0; c < rr->rr_cols; c++) {
+ raidz_col_t *rc = &rr->rr_col[c];
+ raidz_col_t *prc =
+ &rm->rm_phys_col[rc->rc_devidx];
+
+ if (rc->rc_size == 0)
+ continue;
+
+ if (prc->rc_size == 0) {
+ ASSERT0(prc->rc_offset);
+ prc->rc_offset = rc->rc_offset;
+ } else if (prc->rc_offset + prc->rc_size !=
+ rc->rc_offset) {
+ /*
+ * This block is not contiguous and
+ * therefore can't be aggregated.
+ * This is expected to be rare, so
+ * the cost of allocating and then
+ * freeing rm_phys_col is not
+ * significant.
+ */
+ kmem_free(rm->rm_phys_col,
+ sizeof (raidz_col_t) *
+ rm->rm_nphys_cols);
+ rm->rm_phys_col = NULL;
+ rm->rm_nphys_cols = 0;
+ break;
+ }
+ prc->rc_size += rc->rc_size;
+ }
+ }
+ }
+ if (rm->rm_phys_col != NULL) {
+ /*
+ * Allocate aggregate ABD's.
+ */
+ for (int i = 0; i < rm->rm_nphys_cols; i++) {
+ raidz_col_t *prc = &rm->rm_phys_col[i];
+
+ prc->rc_devidx = i;
+ if (prc->rc_size == 0)
+ continue;
+
+ prc->rc_abd =
+ abd_alloc_linear(rm->rm_phys_col[i].rc_size,
+ B_FALSE);
+ }
+
+ /*
+ * Point the parity abd's into the aggregate abd's.
+ */
+ for (int i = 0; i < rm->rm_nrows; i++) {
+ raidz_row_t *rr = rm->rm_row[i];
+ for (int c = 0; c < rr->rr_firstdatacol; c++) {
+ raidz_col_t *rc = &rr->rr_col[c];
+ raidz_col_t *prc =
+ &rm->rm_phys_col[rc->rc_devidx];
+ rc->rc_abd =
+ abd_get_offset_struct(&rc->rc_abdstruct,
+ prc->rc_abd,
+ rc->rc_offset - prc->rc_offset,
+ rc->rc_size);
+ }
+ }
+ } else {
+ /*
+ * Allocate new abd's for the parity sectors.
+ */
+ for (int i = 0; i < rm->rm_nrows; i++) {
+ raidz_row_t *rr = rm->rm_row[i];
+ for (int c = 0; c < rr->rr_firstdatacol; c++) {
+ raidz_col_t *rc = &rr->rr_col[c];
+ rc->rc_abd =
+ abd_alloc_linear(rc->rc_size,
+ B_TRUE);
+ }
+ }
+ }
/* init RAIDZ parity ops */
rm->rm_ops = vdev_raidz_math_get_ops();
return (rm);
}
struct pqr_struct {
uint64_t *p;
uint64_t *q;
uint64_t *r;
};
static int
vdev_raidz_p_func(void *buf, size_t size, void *private)
{
struct pqr_struct *pqr = private;
const uint64_t *src = buf;
- int i, cnt = size / sizeof (src[0]);
+ int cnt = size / sizeof (src[0]);
ASSERT(pqr->p && !pqr->q && !pqr->r);
- for (i = 0; i < cnt; i++, src++, pqr->p++)
+ for (int i = 0; i < cnt; i++, src++, pqr->p++)
*pqr->p ^= *src;
return (0);
}
static int
vdev_raidz_pq_func(void *buf, size_t size, void *private)
{
struct pqr_struct *pqr = private;
const uint64_t *src = buf;
uint64_t mask;
- int i, cnt = size / sizeof (src[0]);
+ int cnt = size / sizeof (src[0]);
ASSERT(pqr->p && pqr->q && !pqr->r);
- for (i = 0; i < cnt; i++, src++, pqr->p++, pqr->q++) {
+ for (int i = 0; i < cnt; i++, src++, pqr->p++, pqr->q++) {
*pqr->p ^= *src;
VDEV_RAIDZ_64MUL_2(*pqr->q, mask);
*pqr->q ^= *src;
}
return (0);
}
static int
vdev_raidz_pqr_func(void *buf, size_t size, void *private)
{
struct pqr_struct *pqr = private;
const uint64_t *src = buf;
uint64_t mask;
- int i, cnt = size / sizeof (src[0]);
+ int cnt = size / sizeof (src[0]);
ASSERT(pqr->p && pqr->q && pqr->r);
- for (i = 0; i < cnt; i++, src++, pqr->p++, pqr->q++, pqr->r++) {
+ for (int i = 0; i < cnt; i++, src++, pqr->p++, pqr->q++, pqr->r++) {
*pqr->p ^= *src;
VDEV_RAIDZ_64MUL_2(*pqr->q, mask);
*pqr->q ^= *src;
VDEV_RAIDZ_64MUL_4(*pqr->r, mask);
*pqr->r ^= *src;
}
return (0);
}
static void
vdev_raidz_generate_parity_p(raidz_row_t *rr)
{
uint64_t *p = abd_to_buf(rr->rr_col[VDEV_RAIDZ_P].rc_abd);
for (int c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
abd_t *src = rr->rr_col[c].rc_abd;
if (c == rr->rr_firstdatacol) {
abd_copy_to_buf(p, src, rr->rr_col[c].rc_size);
} else {
struct pqr_struct pqr = { p, NULL, NULL };
(void) abd_iterate_func(src, 0, rr->rr_col[c].rc_size,
vdev_raidz_p_func, &pqr);
}
}
}
static void
vdev_raidz_generate_parity_pq(raidz_row_t *rr)
{
uint64_t *p = abd_to_buf(rr->rr_col[VDEV_RAIDZ_P].rc_abd);
uint64_t *q = abd_to_buf(rr->rr_col[VDEV_RAIDZ_Q].rc_abd);
uint64_t pcnt = rr->rr_col[VDEV_RAIDZ_P].rc_size / sizeof (p[0]);
ASSERT(rr->rr_col[VDEV_RAIDZ_P].rc_size ==
rr->rr_col[VDEV_RAIDZ_Q].rc_size);
for (int c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
abd_t *src = rr->rr_col[c].rc_abd;
uint64_t ccnt = rr->rr_col[c].rc_size / sizeof (p[0]);
if (c == rr->rr_firstdatacol) {
ASSERT(ccnt == pcnt || ccnt == 0);
abd_copy_to_buf(p, src, rr->rr_col[c].rc_size);
(void) memcpy(q, p, rr->rr_col[c].rc_size);
for (uint64_t i = ccnt; i < pcnt; i++) {
p[i] = 0;
q[i] = 0;
}
} else {
struct pqr_struct pqr = { p, q, NULL };
ASSERT(ccnt <= pcnt);
(void) abd_iterate_func(src, 0, rr->rr_col[c].rc_size,
vdev_raidz_pq_func, &pqr);
/*
* Treat short columns as though they are full of 0s.
* Note that there's therefore nothing needed for P.
*/
uint64_t mask;
for (uint64_t i = ccnt; i < pcnt; i++) {
VDEV_RAIDZ_64MUL_2(q[i], mask);
}
}
}
}
static void
vdev_raidz_generate_parity_pqr(raidz_row_t *rr)
{
uint64_t *p = abd_to_buf(rr->rr_col[VDEV_RAIDZ_P].rc_abd);
uint64_t *q = abd_to_buf(rr->rr_col[VDEV_RAIDZ_Q].rc_abd);
uint64_t *r = abd_to_buf(rr->rr_col[VDEV_RAIDZ_R].rc_abd);
uint64_t pcnt = rr->rr_col[VDEV_RAIDZ_P].rc_size / sizeof (p[0]);
ASSERT(rr->rr_col[VDEV_RAIDZ_P].rc_size ==
rr->rr_col[VDEV_RAIDZ_Q].rc_size);
ASSERT(rr->rr_col[VDEV_RAIDZ_P].rc_size ==
rr->rr_col[VDEV_RAIDZ_R].rc_size);
for (int c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
abd_t *src = rr->rr_col[c].rc_abd;
uint64_t ccnt = rr->rr_col[c].rc_size / sizeof (p[0]);
if (c == rr->rr_firstdatacol) {
ASSERT(ccnt == pcnt || ccnt == 0);
abd_copy_to_buf(p, src, rr->rr_col[c].rc_size);
(void) memcpy(q, p, rr->rr_col[c].rc_size);
(void) memcpy(r, p, rr->rr_col[c].rc_size);
for (uint64_t i = ccnt; i < pcnt; i++) {
p[i] = 0;
q[i] = 0;
r[i] = 0;
}
} else {
struct pqr_struct pqr = { p, q, r };
ASSERT(ccnt <= pcnt);
(void) abd_iterate_func(src, 0, rr->rr_col[c].rc_size,
vdev_raidz_pqr_func, &pqr);
/*
* Treat short columns as though they are full of 0s.
* Note that there's therefore nothing needed for P.
*/
uint64_t mask;
for (uint64_t i = ccnt; i < pcnt; i++) {
VDEV_RAIDZ_64MUL_2(q[i], mask);
VDEV_RAIDZ_64MUL_4(r[i], mask);
}
}
}
}
/*
* Generate RAID parity in the first virtual columns according to the number of
* parity columns available.
*/
void
vdev_raidz_generate_parity_row(raidz_map_t *rm, raidz_row_t *rr)
{
- ASSERT3U(rr->rr_cols, !=, 0);
+ if (rr->rr_cols == 0) {
+ /*
+ * We are handling this block one row at a time (because
+ * this block has a different logical vs physical width,
+ * due to RAIDZ expansion), and this is a pad-only row,
+ * which has no parity.
+ */
+ return;
+ }
/* Generate using the new math implementation */
if (vdev_raidz_math_generate(rm, rr) != RAIDZ_ORIGINAL_IMPL)
return;
switch (rr->rr_firstdatacol) {
case 1:
vdev_raidz_generate_parity_p(rr);
break;
case 2:
vdev_raidz_generate_parity_pq(rr);
break;
case 3:
vdev_raidz_generate_parity_pqr(rr);
break;
default:
cmn_err(CE_PANIC, "invalid RAID-Z configuration");
}
}
void
vdev_raidz_generate_parity(raidz_map_t *rm)
{
for (int i = 0; i < rm->rm_nrows; i++) {
raidz_row_t *rr = rm->rm_row[i];
vdev_raidz_generate_parity_row(rm, rr);
}
}
static int
vdev_raidz_reconst_p_func(void *dbuf, void *sbuf, size_t size, void *private)
{
(void) private;
uint64_t *dst = dbuf;
uint64_t *src = sbuf;
int cnt = size / sizeof (src[0]);
for (int i = 0; i < cnt; i++) {
dst[i] ^= src[i];
}
return (0);
}
static int
vdev_raidz_reconst_q_pre_func(void *dbuf, void *sbuf, size_t size,
void *private)
{
(void) private;
uint64_t *dst = dbuf;
uint64_t *src = sbuf;
uint64_t mask;
int cnt = size / sizeof (dst[0]);
for (int i = 0; i < cnt; i++, dst++, src++) {
VDEV_RAIDZ_64MUL_2(*dst, mask);
*dst ^= *src;
}
return (0);
}
static int
vdev_raidz_reconst_q_pre_tail_func(void *buf, size_t size, void *private)
{
(void) private;
uint64_t *dst = buf;
uint64_t mask;
int cnt = size / sizeof (dst[0]);
for (int i = 0; i < cnt; i++, dst++) {
/* same operation as vdev_raidz_reconst_q_pre_func() on dst */
VDEV_RAIDZ_64MUL_2(*dst, mask);
}
return (0);
}
struct reconst_q_struct {
uint64_t *q;
int exp;
};
static int
vdev_raidz_reconst_q_post_func(void *buf, size_t size, void *private)
{
struct reconst_q_struct *rq = private;
uint64_t *dst = buf;
int cnt = size / sizeof (dst[0]);
for (int i = 0; i < cnt; i++, dst++, rq->q++) {
int j;
uint8_t *b;
*dst ^= *rq->q;
for (j = 0, b = (uint8_t *)dst; j < 8; j++, b++) {
*b = vdev_raidz_exp2(*b, rq->exp);
}
}
return (0);
}
struct reconst_pq_struct {
uint8_t *p;
uint8_t *q;
uint8_t *pxy;
uint8_t *qxy;
int aexp;
int bexp;
};
static int
vdev_raidz_reconst_pq_func(void *xbuf, void *ybuf, size_t size, void *private)
{
struct reconst_pq_struct *rpq = private;
uint8_t *xd = xbuf;
uint8_t *yd = ybuf;
for (int i = 0; i < size;
i++, rpq->p++, rpq->q++, rpq->pxy++, rpq->qxy++, xd++, yd++) {
*xd = vdev_raidz_exp2(*rpq->p ^ *rpq->pxy, rpq->aexp) ^
vdev_raidz_exp2(*rpq->q ^ *rpq->qxy, rpq->bexp);
*yd = *rpq->p ^ *rpq->pxy ^ *xd;
}
return (0);
}
static int
vdev_raidz_reconst_pq_tail_func(void *xbuf, size_t size, void *private)
{
struct reconst_pq_struct *rpq = private;
uint8_t *xd = xbuf;
for (int i = 0; i < size;
i++, rpq->p++, rpq->q++, rpq->pxy++, rpq->qxy++, xd++) {
/* same operation as vdev_raidz_reconst_pq_func() on xd */
*xd = vdev_raidz_exp2(*rpq->p ^ *rpq->pxy, rpq->aexp) ^
vdev_raidz_exp2(*rpq->q ^ *rpq->qxy, rpq->bexp);
}
return (0);
}
static void
vdev_raidz_reconstruct_p(raidz_row_t *rr, int *tgts, int ntgts)
{
int x = tgts[0];
abd_t *dst, *src;
+ if (zfs_flags & ZFS_DEBUG_RAIDZ_RECONSTRUCT)
+ zfs_dbgmsg("reconstruct_p(rm=%px x=%u)", rr, x);
+
ASSERT3U(ntgts, ==, 1);
ASSERT3U(x, >=, rr->rr_firstdatacol);
ASSERT3U(x, <, rr->rr_cols);
ASSERT3U(rr->rr_col[x].rc_size, <=, rr->rr_col[VDEV_RAIDZ_P].rc_size);
src = rr->rr_col[VDEV_RAIDZ_P].rc_abd;
dst = rr->rr_col[x].rc_abd;
abd_copy_from_buf(dst, abd_to_buf(src), rr->rr_col[x].rc_size);
for (int c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
uint64_t size = MIN(rr->rr_col[x].rc_size,
rr->rr_col[c].rc_size);
src = rr->rr_col[c].rc_abd;
if (c == x)
continue;
(void) abd_iterate_func2(dst, src, 0, 0, size,
vdev_raidz_reconst_p_func, NULL);
}
}
static void
vdev_raidz_reconstruct_q(raidz_row_t *rr, int *tgts, int ntgts)
{
int x = tgts[0];
int c, exp;
abd_t *dst, *src;
+ if (zfs_flags & ZFS_DEBUG_RAIDZ_RECONSTRUCT)
+ zfs_dbgmsg("reconstruct_q(rm=%px x=%u)", rr, x);
+
ASSERT(ntgts == 1);
ASSERT(rr->rr_col[x].rc_size <= rr->rr_col[VDEV_RAIDZ_Q].rc_size);
for (c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
uint64_t size = (c == x) ? 0 : MIN(rr->rr_col[x].rc_size,
rr->rr_col[c].rc_size);
src = rr->rr_col[c].rc_abd;
dst = rr->rr_col[x].rc_abd;
if (c == rr->rr_firstdatacol) {
abd_copy(dst, src, size);
if (rr->rr_col[x].rc_size > size) {
abd_zero_off(dst, size,
rr->rr_col[x].rc_size - size);
}
} else {
ASSERT3U(size, <=, rr->rr_col[x].rc_size);
(void) abd_iterate_func2(dst, src, 0, 0, size,
vdev_raidz_reconst_q_pre_func, NULL);
(void) abd_iterate_func(dst,
size, rr->rr_col[x].rc_size - size,
vdev_raidz_reconst_q_pre_tail_func, NULL);
}
}
src = rr->rr_col[VDEV_RAIDZ_Q].rc_abd;
dst = rr->rr_col[x].rc_abd;
exp = 255 - (rr->rr_cols - 1 - x);
struct reconst_q_struct rq = { abd_to_buf(src), exp };
(void) abd_iterate_func(dst, 0, rr->rr_col[x].rc_size,
vdev_raidz_reconst_q_post_func, &rq);
}
static void
vdev_raidz_reconstruct_pq(raidz_row_t *rr, int *tgts, int ntgts)
{
uint8_t *p, *q, *pxy, *qxy, tmp, a, b, aexp, bexp;
abd_t *pdata, *qdata;
uint64_t xsize, ysize;
int x = tgts[0];
int y = tgts[1];
abd_t *xd, *yd;
+ if (zfs_flags & ZFS_DEBUG_RAIDZ_RECONSTRUCT)
+ zfs_dbgmsg("reconstruct_pq(rm=%px x=%u y=%u)", rr, x, y);
+
ASSERT(ntgts == 2);
ASSERT(x < y);
ASSERT(x >= rr->rr_firstdatacol);
ASSERT(y < rr->rr_cols);
ASSERT(rr->rr_col[x].rc_size >= rr->rr_col[y].rc_size);
/*
* Move the parity data aside -- we're going to compute parity as
* though columns x and y were full of zeros -- Pxy and Qxy. We want to
* reuse the parity generation mechanism without trashing the actual
* parity so we make those columns appear to be full of zeros by
* setting their lengths to zero.
*/
pdata = rr->rr_col[VDEV_RAIDZ_P].rc_abd;
qdata = rr->rr_col[VDEV_RAIDZ_Q].rc_abd;
xsize = rr->rr_col[x].rc_size;
ysize = rr->rr_col[y].rc_size;
rr->rr_col[VDEV_RAIDZ_P].rc_abd =
abd_alloc_linear(rr->rr_col[VDEV_RAIDZ_P].rc_size, B_TRUE);
rr->rr_col[VDEV_RAIDZ_Q].rc_abd =
abd_alloc_linear(rr->rr_col[VDEV_RAIDZ_Q].rc_size, B_TRUE);
rr->rr_col[x].rc_size = 0;
rr->rr_col[y].rc_size = 0;
vdev_raidz_generate_parity_pq(rr);
rr->rr_col[x].rc_size = xsize;
rr->rr_col[y].rc_size = ysize;
p = abd_to_buf(pdata);
q = abd_to_buf(qdata);
pxy = abd_to_buf(rr->rr_col[VDEV_RAIDZ_P].rc_abd);
qxy = abd_to_buf(rr->rr_col[VDEV_RAIDZ_Q].rc_abd);
xd = rr->rr_col[x].rc_abd;
yd = rr->rr_col[y].rc_abd;
/*
* We now have:
* Pxy = P + D_x + D_y
* Qxy = Q + 2^(ndevs - 1 - x) * D_x + 2^(ndevs - 1 - y) * D_y
*
* We can then solve for D_x:
* D_x = A * (P + Pxy) + B * (Q + Qxy)
* where
* A = 2^(x - y) * (2^(x - y) + 1)^-1
* B = 2^(ndevs - 1 - x) * (2^(x - y) + 1)^-1
*
* With D_x in hand, we can easily solve for D_y:
* D_y = P + Pxy + D_x
*/
a = vdev_raidz_pow2[255 + x - y];
b = vdev_raidz_pow2[255 - (rr->rr_cols - 1 - x)];
tmp = 255 - vdev_raidz_log2[a ^ 1];
aexp = vdev_raidz_log2[vdev_raidz_exp2(a, tmp)];
bexp = vdev_raidz_log2[vdev_raidz_exp2(b, tmp)];
ASSERT3U(xsize, >=, ysize);
struct reconst_pq_struct rpq = { p, q, pxy, qxy, aexp, bexp };
(void) abd_iterate_func2(xd, yd, 0, 0, ysize,
vdev_raidz_reconst_pq_func, &rpq);
(void) abd_iterate_func(xd, ysize, xsize - ysize,
vdev_raidz_reconst_pq_tail_func, &rpq);
abd_free(rr->rr_col[VDEV_RAIDZ_P].rc_abd);
abd_free(rr->rr_col[VDEV_RAIDZ_Q].rc_abd);
/*
* Restore the saved parity data.
*/
rr->rr_col[VDEV_RAIDZ_P].rc_abd = pdata;
rr->rr_col[VDEV_RAIDZ_Q].rc_abd = qdata;
}
/*
* In the general case of reconstruction, we must solve the system of linear
* equations defined by the coefficients used to generate parity as well as
* the contents of the data and parity disks. This can be expressed with
* vectors for the original data (D) and the actual data (d) and parity (p)
* and a matrix composed of the identity matrix (I) and a dispersal matrix (V):
*
* __ __ __ __
* | | __ __ | p_0 |
* | V | | D_0 | | p_m-1 |
* | | x | : | = | d_0 |
* | I | | D_n-1 | | : |
* | | ~~ ~~ | d_n-1 |
* ~~ ~~ ~~ ~~
*
* I is simply a square identity matrix of size n, and V is a vandermonde
* matrix defined by the coefficients we chose for the various parity columns
* (1, 2, 4). Note that these values were chosen both for simplicity, speedy
* computation as well as linear separability.
*
* __ __ __ __
* | 1 .. 1 1 1 | | p_0 |
* | 2^n-1 .. 4 2 1 | __ __ | : |
* | 4^n-1 .. 16 4 1 | | D_0 | | p_m-1 |
* | 1 .. 0 0 0 | | D_1 | | d_0 |
* | 0 .. 0 0 0 | x | D_2 | = | d_1 |
* | : : : : | | : | | d_2 |
* | 0 .. 1 0 0 | | D_n-1 | | : |
* | 0 .. 0 1 0 | ~~ ~~ | : |
* | 0 .. 0 0 1 | | d_n-1 |
* ~~ ~~ ~~ ~~
*
* Note that I, V, d, and p are known. To compute D, we must invert the
* matrix and use the known data and parity values to reconstruct the unknown
* data values. We begin by removing the rows in V|I and d|p that correspond
* to failed or missing columns; we then make V|I square (n x n) and d|p
* sized n by removing rows corresponding to unused parity from the bottom up
* to generate (V|I)' and (d|p)'. We can then generate the inverse of (V|I)'
* using Gauss-Jordan elimination. In the example below we use m=3 parity
* columns, n=8 data columns, with errors in d_1, d_2, and p_1:
* __ __
* | 1 1 1 1 1 1 1 1 |
* | 128 64 32 16 8 4 2 1 | <-----+-+-- missing disks
* | 19 205 116 29 64 16 4 1 | / /
* | 1 0 0 0 0 0 0 0 | / /
* | 0 1 0 0 0 0 0 0 | <--' /
* (V|I) = | 0 0 1 0 0 0 0 0 | <---'
* | 0 0 0 1 0 0 0 0 |
* | 0 0 0 0 1 0 0 0 |
* | 0 0 0 0 0 1 0 0 |
* | 0 0 0 0 0 0 1 0 |
* | 0 0 0 0 0 0 0 1 |
* ~~ ~~
* __ __
* | 1 1 1 1 1 1 1 1 |
* | 128 64 32 16 8 4 2 1 |
* | 19 205 116 29 64 16 4 1 |
* | 1 0 0 0 0 0 0 0 |
* | 0 1 0 0 0 0 0 0 |
* (V|I)' = | 0 0 1 0 0 0 0 0 |
* | 0 0 0 1 0 0 0 0 |
* | 0 0 0 0 1 0 0 0 |
* | 0 0 0 0 0 1 0 0 |
* | 0 0 0 0 0 0 1 0 |
* | 0 0 0 0 0 0 0 1 |
* ~~ ~~
*
* Here we employ Gauss-Jordan elimination to find the inverse of (V|I)'. We
* have carefully chosen the seed values 1, 2, and 4 to ensure that this
* matrix is not singular.
* __ __
* | 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 |
* | 19 205 116 29 64 16 4 1 0 1 0 0 0 0 0 0 |
* | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 |
* | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 |
* | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 |
* | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 |
* | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 |
* | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 |
* ~~ ~~
* __ __
* | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 |
* | 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 |
* | 19 205 116 29 64 16 4 1 0 1 0 0 0 0 0 0 |
* | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 |
* | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 |
* | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 |
* | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 |
* | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 |
* ~~ ~~
* __ __
* | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 |
* | 0 1 1 0 0 0 0 0 1 0 1 1 1 1 1 1 |
* | 0 205 116 0 0 0 0 0 0 1 19 29 64 16 4 1 |
* | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 |
* | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 |
* | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 |
* | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 |
* | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 |
* ~~ ~~
* __ __
* | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 |
* | 0 1 1 0 0 0 0 0 1 0 1 1 1 1 1 1 |
* | 0 0 185 0 0 0 0 0 205 1 222 208 141 221 201 204 |
* | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 |
* | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 |
* | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 |
* | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 |
* | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 |
* ~~ ~~
* __ __
* | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 |
* | 0 1 1 0 0 0 0 0 1 0 1 1 1 1 1 1 |
* | 0 0 1 0 0 0 0 0 166 100 4 40 158 168 216 209 |
* | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 |
* | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 |
* | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 |
* | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 |
* | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 |
* ~~ ~~
* __ __
* | 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 |
* | 0 1 0 0 0 0 0 0 167 100 5 41 159 169 217 208 |
* | 0 0 1 0 0 0 0 0 166 100 4 40 158 168 216 209 |
* | 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 |
* | 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 |
* | 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 |
* | 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 |
* | 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 |
* ~~ ~~
* __ __
* | 0 0 1 0 0 0 0 0 |
* | 167 100 5 41 159 169 217 208 |
* | 166 100 4 40 158 168 216 209 |
* (V|I)'^-1 = | 0 0 0 1 0 0 0 0 |
* | 0 0 0 0 1 0 0 0 |
* | 0 0 0 0 0 1 0 0 |
* | 0 0 0 0 0 0 1 0 |
* | 0 0 0 0 0 0 0 1 |
* ~~ ~~
*
* We can then simply compute D = (V|I)'^-1 x (d|p)' to discover the values
* of the missing data.
*
* As is apparent from the example above, the only non-trivial rows in the
* inverse matrix correspond to the data disks that we're trying to
* reconstruct. Indeed, those are the only rows we need as the others would
* only be useful for reconstructing data known or assumed to be valid. For
* that reason, we only build the coefficients in the rows that correspond to
* targeted columns.
*/
static void
vdev_raidz_matrix_init(raidz_row_t *rr, int n, int nmap, int *map,
uint8_t **rows)
{
int i, j;
int pow;
ASSERT(n == rr->rr_cols - rr->rr_firstdatacol);
/*
* Fill in the missing rows of interest.
*/
for (i = 0; i < nmap; i++) {
ASSERT3S(0, <=, map[i]);
ASSERT3S(map[i], <=, 2);
pow = map[i] * n;
if (pow > 255)
pow -= 255;
ASSERT(pow <= 255);
for (j = 0; j < n; j++) {
pow -= map[i];
if (pow < 0)
pow += 255;
rows[i][j] = vdev_raidz_pow2[pow];
}
}
}
static void
vdev_raidz_matrix_invert(raidz_row_t *rr, int n, int nmissing, int *missing,
uint8_t **rows, uint8_t **invrows, const uint8_t *used)
{
int i, j, ii, jj;
uint8_t log;
/*
* Assert that the first nmissing entries from the array of used
* columns correspond to parity columns and that subsequent entries
* correspond to data columns.
*/
for (i = 0; i < nmissing; i++) {
ASSERT3S(used[i], <, rr->rr_firstdatacol);
}
for (; i < n; i++) {
ASSERT3S(used[i], >=, rr->rr_firstdatacol);
}
/*
* First initialize the storage where we'll compute the inverse rows.
*/
for (i = 0; i < nmissing; i++) {
for (j = 0; j < n; j++) {
invrows[i][j] = (i == j) ? 1 : 0;
}
}
/*
* Subtract all trivial rows from the rows of consequence.
*/
for (i = 0; i < nmissing; i++) {
for (j = nmissing; j < n; j++) {
ASSERT3U(used[j], >=, rr->rr_firstdatacol);
jj = used[j] - rr->rr_firstdatacol;
ASSERT3S(jj, <, n);
invrows[i][j] = rows[i][jj];
rows[i][jj] = 0;
}
}
/*
* For each of the rows of interest, we must normalize it and subtract
* a multiple of it from the other rows.
*/
for (i = 0; i < nmissing; i++) {
for (j = 0; j < missing[i]; j++) {
ASSERT0(rows[i][j]);
}
ASSERT3U(rows[i][missing[i]], !=, 0);
/*
* Compute the inverse of the first element and multiply each
* element in the row by that value.
*/
log = 255 - vdev_raidz_log2[rows[i][missing[i]]];
for (j = 0; j < n; j++) {
rows[i][j] = vdev_raidz_exp2(rows[i][j], log);
invrows[i][j] = vdev_raidz_exp2(invrows[i][j], log);
}
for (ii = 0; ii < nmissing; ii++) {
if (i == ii)
continue;
ASSERT3U(rows[ii][missing[i]], !=, 0);
log = vdev_raidz_log2[rows[ii][missing[i]]];
for (j = 0; j < n; j++) {
rows[ii][j] ^=
vdev_raidz_exp2(rows[i][j], log);
invrows[ii][j] ^=
vdev_raidz_exp2(invrows[i][j], log);
}
}
}
/*
* Verify that the data that is left in the rows are properly part of
* an identity matrix.
*/
for (i = 0; i < nmissing; i++) {
for (j = 0; j < n; j++) {
if (j == missing[i]) {
ASSERT3U(rows[i][j], ==, 1);
} else {
ASSERT0(rows[i][j]);
}
}
}
}
static void
vdev_raidz_matrix_reconstruct(raidz_row_t *rr, int n, int nmissing,
int *missing, uint8_t **invrows, const uint8_t *used)
{
int i, j, x, cc, c;
uint8_t *src;
uint64_t ccount;
uint8_t *dst[VDEV_RAIDZ_MAXPARITY] = { NULL };
uint64_t dcount[VDEV_RAIDZ_MAXPARITY] = { 0 };
uint8_t log = 0;
uint8_t val;
int ll;
uint8_t *invlog[VDEV_RAIDZ_MAXPARITY];
uint8_t *p, *pp;
size_t psize;
psize = sizeof (invlog[0][0]) * n * nmissing;
p = kmem_alloc(psize, KM_SLEEP);
for (pp = p, i = 0; i < nmissing; i++) {
invlog[i] = pp;
pp += n;
}
for (i = 0; i < nmissing; i++) {
for (j = 0; j < n; j++) {
ASSERT3U(invrows[i][j], !=, 0);
invlog[i][j] = vdev_raidz_log2[invrows[i][j]];
}
}
for (i = 0; i < n; i++) {
c = used[i];
ASSERT3U(c, <, rr->rr_cols);
ccount = rr->rr_col[c].rc_size;
ASSERT(ccount >= rr->rr_col[missing[0]].rc_size || i > 0);
if (ccount == 0)
continue;
src = abd_to_buf(rr->rr_col[c].rc_abd);
for (j = 0; j < nmissing; j++) {
cc = missing[j] + rr->rr_firstdatacol;
ASSERT3U(cc, >=, rr->rr_firstdatacol);
ASSERT3U(cc, <, rr->rr_cols);
ASSERT3U(cc, !=, c);
dcount[j] = rr->rr_col[cc].rc_size;
if (dcount[j] != 0)
dst[j] = abd_to_buf(rr->rr_col[cc].rc_abd);
}
for (x = 0; x < ccount; x++, src++) {
if (*src != 0)
log = vdev_raidz_log2[*src];
for (cc = 0; cc < nmissing; cc++) {
if (x >= dcount[cc])
continue;
if (*src == 0) {
val = 0;
} else {
if ((ll = log + invlog[cc][i]) >= 255)
ll -= 255;
val = vdev_raidz_pow2[ll];
}
if (i == 0)
dst[cc][x] = val;
else
dst[cc][x] ^= val;
}
}
}
kmem_free(p, psize);
}
static void
vdev_raidz_reconstruct_general(raidz_row_t *rr, int *tgts, int ntgts)
{
int n, i, c, t, tt;
int nmissing_rows;
int missing_rows[VDEV_RAIDZ_MAXPARITY];
int parity_map[VDEV_RAIDZ_MAXPARITY];
uint8_t *p, *pp;
size_t psize;
uint8_t *rows[VDEV_RAIDZ_MAXPARITY];
uint8_t *invrows[VDEV_RAIDZ_MAXPARITY];
uint8_t *used;
abd_t **bufs = NULL;
+ if (zfs_flags & ZFS_DEBUG_RAIDZ_RECONSTRUCT)
+ zfs_dbgmsg("reconstruct_general(rm=%px ntgts=%u)", rr, ntgts);
/*
* Matrix reconstruction can't use scatter ABDs yet, so we allocate
* temporary linear ABDs if any non-linear ABDs are found.
*/
for (i = rr->rr_firstdatacol; i < rr->rr_cols; i++) {
+ ASSERT(rr->rr_col[i].rc_abd != NULL);
if (!abd_is_linear(rr->rr_col[i].rc_abd)) {
bufs = kmem_alloc(rr->rr_cols * sizeof (abd_t *),
KM_PUSHPAGE);
for (c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
raidz_col_t *col = &rr->rr_col[c];
bufs[c] = col->rc_abd;
if (bufs[c] != NULL) {
col->rc_abd = abd_alloc_linear(
col->rc_size, B_TRUE);
abd_copy(col->rc_abd, bufs[c],
col->rc_size);
}
}
break;
}
}
n = rr->rr_cols - rr->rr_firstdatacol;
/*
* Figure out which data columns are missing.
*/
nmissing_rows = 0;
for (t = 0; t < ntgts; t++) {
if (tgts[t] >= rr->rr_firstdatacol) {
missing_rows[nmissing_rows++] =
tgts[t] - rr->rr_firstdatacol;
}
}
/*
* Figure out which parity columns to use to help generate the missing
* data columns.
*/
for (tt = 0, c = 0, i = 0; i < nmissing_rows; c++) {
ASSERT(tt < ntgts);
ASSERT(c < rr->rr_firstdatacol);
/*
* Skip any targeted parity columns.
*/
if (c == tgts[tt]) {
tt++;
continue;
}
parity_map[i] = c;
i++;
}
psize = (sizeof (rows[0][0]) + sizeof (invrows[0][0])) *
nmissing_rows * n + sizeof (used[0]) * n;
p = kmem_alloc(psize, KM_SLEEP);
for (pp = p, i = 0; i < nmissing_rows; i++) {
rows[i] = pp;
pp += n;
invrows[i] = pp;
pp += n;
}
used = pp;
for (i = 0; i < nmissing_rows; i++) {
used[i] = parity_map[i];
}
for (tt = 0, c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
if (tt < nmissing_rows &&
c == missing_rows[tt] + rr->rr_firstdatacol) {
tt++;
continue;
}
ASSERT3S(i, <, n);
used[i] = c;
i++;
}
/*
* Initialize the interesting rows of the matrix.
*/
vdev_raidz_matrix_init(rr, n, nmissing_rows, parity_map, rows);
/*
* Invert the matrix.
*/
vdev_raidz_matrix_invert(rr, n, nmissing_rows, missing_rows, rows,
invrows, used);
/*
* Reconstruct the missing data using the generated matrix.
*/
vdev_raidz_matrix_reconstruct(rr, n, nmissing_rows, missing_rows,
invrows, used);
kmem_free(p, psize);
/*
* copy back from temporary linear abds and free them
*/
if (bufs) {
for (c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
raidz_col_t *col = &rr->rr_col[c];
if (bufs[c] != NULL) {
abd_copy(bufs[c], col->rc_abd, col->rc_size);
abd_free(col->rc_abd);
}
col->rc_abd = bufs[c];
}
kmem_free(bufs, rr->rr_cols * sizeof (abd_t *));
}
}
static void
vdev_raidz_reconstruct_row(raidz_map_t *rm, raidz_row_t *rr,
const int *t, int nt)
{
int tgts[VDEV_RAIDZ_MAXPARITY], *dt;
int ntgts;
int i, c, ret;
int nbadparity, nbaddata;
int parity_valid[VDEV_RAIDZ_MAXPARITY];
+ if (zfs_flags & ZFS_DEBUG_RAIDZ_RECONSTRUCT) {
+ zfs_dbgmsg("reconstruct(rm=%px nt=%u cols=%u md=%u mp=%u)",
+ rr, nt, (int)rr->rr_cols, (int)rr->rr_missingdata,
+ (int)rr->rr_missingparity);
+ }
+
nbadparity = rr->rr_firstdatacol;
nbaddata = rr->rr_cols - nbadparity;
ntgts = 0;
for (i = 0, c = 0; c < rr->rr_cols; c++) {
+ if (zfs_flags & ZFS_DEBUG_RAIDZ_RECONSTRUCT) {
+ zfs_dbgmsg("reconstruct(rm=%px col=%u devid=%u "
+ "offset=%llx error=%u)",
+ rr, c, (int)rr->rr_col[c].rc_devidx,
+ (long long)rr->rr_col[c].rc_offset,
+ (int)rr->rr_col[c].rc_error);
+ }
if (c < rr->rr_firstdatacol)
parity_valid[c] = B_FALSE;
if (i < nt && c == t[i]) {
tgts[ntgts++] = c;
i++;
} else if (rr->rr_col[c].rc_error != 0) {
tgts[ntgts++] = c;
} else if (c >= rr->rr_firstdatacol) {
nbaddata--;
} else {
parity_valid[c] = B_TRUE;
nbadparity--;
}
}
ASSERT(ntgts >= nt);
ASSERT(nbaddata >= 0);
ASSERT(nbaddata + nbadparity == ntgts);
dt = &tgts[nbadparity];
/* Reconstruct using the new math implementation */
ret = vdev_raidz_math_reconstruct(rm, rr, parity_valid, dt, nbaddata);
if (ret != RAIDZ_ORIGINAL_IMPL)
return;
/*
* See if we can use any of our optimized reconstruction routines.
*/
switch (nbaddata) {
case 1:
if (parity_valid[VDEV_RAIDZ_P]) {
vdev_raidz_reconstruct_p(rr, dt, 1);
return;
}
ASSERT(rr->rr_firstdatacol > 1);
if (parity_valid[VDEV_RAIDZ_Q]) {
vdev_raidz_reconstruct_q(rr, dt, 1);
return;
}
ASSERT(rr->rr_firstdatacol > 2);
break;
case 2:
ASSERT(rr->rr_firstdatacol > 1);
if (parity_valid[VDEV_RAIDZ_P] &&
parity_valid[VDEV_RAIDZ_Q]) {
vdev_raidz_reconstruct_pq(rr, dt, 2);
return;
}
ASSERT(rr->rr_firstdatacol > 2);
break;
}
vdev_raidz_reconstruct_general(rr, tgts, ntgts);
}
static int
vdev_raidz_open(vdev_t *vd, uint64_t *asize, uint64_t *max_asize,
uint64_t *logical_ashift, uint64_t *physical_ashift)
{
vdev_raidz_t *vdrz = vd->vdev_tsd;
uint64_t nparity = vdrz->vd_nparity;
int c;
int lasterror = 0;
int numerrors = 0;
ASSERT(nparity > 0);
if (nparity > VDEV_RAIDZ_MAXPARITY ||
vd->vdev_children < nparity + 1) {
vd->vdev_stat.vs_aux = VDEV_AUX_BAD_LABEL;
return (SET_ERROR(EINVAL));
}
vdev_open_children(vd);
for (c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (cvd->vdev_open_error != 0) {
lasterror = cvd->vdev_open_error;
numerrors++;
continue;
}
*asize = MIN(*asize - 1, cvd->vdev_asize - 1) + 1;
*max_asize = MIN(*max_asize - 1, cvd->vdev_max_asize - 1) + 1;
*logical_ashift = MAX(*logical_ashift, cvd->vdev_ashift);
}
for (c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (cvd->vdev_open_error != 0)
continue;
*physical_ashift = vdev_best_ashift(*logical_ashift,
*physical_ashift, cvd->vdev_physical_ashift);
}
- *asize *= vd->vdev_children;
- *max_asize *= vd->vdev_children;
+ if (vd->vdev_rz_expanding) {
+ *asize *= vd->vdev_children - 1;
+ *max_asize *= vd->vdev_children - 1;
+
+ vd->vdev_min_asize = *asize;
+ } else {
+ *asize *= vd->vdev_children;
+ *max_asize *= vd->vdev_children;
+ }
if (numerrors > nparity) {
vd->vdev_stat.vs_aux = VDEV_AUX_NO_REPLICAS;
return (lasterror);
}
return (0);
}
static void
vdev_raidz_close(vdev_t *vd)
{
for (int c = 0; c < vd->vdev_children; c++) {
if (vd->vdev_child[c] != NULL)
vdev_close(vd->vdev_child[c]);
}
}
+/*
+ * Return the logical width to use, given the txg in which the allocation
+ * happened. Note that BP_PHYSICAL_BIRTH() is usually the txg in which the
+ * BP was allocated. Remapped BP's (that were relocated due to device
+ * removal, see remap_blkptr_cb()), will have a more recent
+ * BP_PHYSICAL_BIRTH() which reflects when the BP was relocated, but we can
+ * ignore these because they can't be on RAIDZ (device removal doesn't
+ * support RAIDZ).
+ */
+static uint64_t
+vdev_raidz_get_logical_width(vdev_raidz_t *vdrz, uint64_t txg)
+{
+ reflow_node_t lookup = {
+ .re_txg = txg,
+ };
+ avl_index_t where;
+
+ uint64_t width;
+ mutex_enter(&vdrz->vd_expand_lock);
+ reflow_node_t *re = avl_find(&vdrz->vd_expand_txgs, &lookup, &where);
+ if (re != NULL) {
+ width = re->re_logical_width;
+ } else {
+ re = avl_nearest(&vdrz->vd_expand_txgs, where, AVL_BEFORE);
+ if (re != NULL)
+ width = re->re_logical_width;
+ else
+ width = vdrz->vd_original_width;
+ }
+ mutex_exit(&vdrz->vd_expand_lock);
+ return (width);
+}
+
+/*
+ * Note: If the RAIDZ vdev has been expanded, older BP's may have allocated
+ * more space due to the lower data-to-parity ratio. In this case it's
+ * important to pass in the correct txg. Note that vdev_gang_header_asize()
+ * relies on a constant asize for psize=SPA_GANGBLOCKSIZE=SPA_MINBLOCKSIZE,
+ * regardless of txg. This is assured because for a single data sector, we
+ * allocate P+1 sectors regardless of width ("cols", which is at least P+1).
+ */
static uint64_t
-vdev_raidz_asize(vdev_t *vd, uint64_t psize)
+vdev_raidz_asize(vdev_t *vd, uint64_t psize, uint64_t txg)
{
vdev_raidz_t *vdrz = vd->vdev_tsd;
uint64_t asize;
uint64_t ashift = vd->vdev_top->vdev_ashift;
- uint64_t cols = vdrz->vd_logical_width;
+ uint64_t cols = vdrz->vd_original_width;
uint64_t nparity = vdrz->vd_nparity;
+ cols = vdev_raidz_get_logical_width(vdrz, txg);
+
asize = ((psize - 1) >> ashift) + 1;
asize += nparity * ((asize + cols - nparity - 1) / (cols - nparity));
asize = roundup(asize, nparity + 1) << ashift;
+#ifdef ZFS_DEBUG
+ uint64_t asize_new = ((psize - 1) >> ashift) + 1;
+ uint64_t ncols_new = vdrz->vd_physical_width;
+ asize_new += nparity * ((asize_new + ncols_new - nparity - 1) /
+ (ncols_new - nparity));
+ asize_new = roundup(asize_new, nparity + 1) << ashift;
+ VERIFY3U(asize_new, <=, asize);
+#endif
+
return (asize);
}
/*
* The allocatable space for a raidz vdev is N * sizeof(smallest child)
* so each child must provide at least 1/Nth of its asize.
*/
static uint64_t
vdev_raidz_min_asize(vdev_t *vd)
{
return ((vd->vdev_min_asize + vd->vdev_children - 1) /
vd->vdev_children);
}
void
vdev_raidz_child_done(zio_t *zio)
{
raidz_col_t *rc = zio->io_private;
ASSERT3P(rc->rc_abd, !=, NULL);
rc->rc_error = zio->io_error;
rc->rc_tried = 1;
rc->rc_skipped = 0;
}
static void
-vdev_raidz_io_verify(vdev_t *vd, raidz_row_t *rr, int col)
+vdev_raidz_shadow_child_done(zio_t *zio)
{
-#ifdef ZFS_DEBUG
- vdev_t *tvd = vd->vdev_top;
+ raidz_col_t *rc = zio->io_private;
+
+ rc->rc_shadow_error = zio->io_error;
+}
+static void
+vdev_raidz_io_verify(zio_t *zio, raidz_map_t *rm, raidz_row_t *rr, int col)
+{
+ (void) rm;
+#ifdef ZFS_DEBUG
range_seg64_t logical_rs, physical_rs, remain_rs;
logical_rs.rs_start = rr->rr_offset;
logical_rs.rs_end = logical_rs.rs_start +
- vdev_raidz_asize(vd, rr->rr_size);
+ vdev_raidz_asize(zio->io_vd, rr->rr_size,
+ BP_PHYSICAL_BIRTH(zio->io_bp));
raidz_col_t *rc = &rr->rr_col[col];
- vdev_t *cvd = vd->vdev_child[rc->rc_devidx];
+ vdev_t *cvd = zio->io_vd->vdev_child[rc->rc_devidx];
vdev_xlate(cvd, &logical_rs, &physical_rs, &remain_rs);
ASSERT(vdev_xlate_is_empty(&remain_rs));
+ if (vdev_xlate_is_empty(&physical_rs)) {
+ /*
+ * If we are in the middle of expansion, the
+ * physical->logical mapping is changing so vdev_xlate()
+ * can't give us a reliable answer.
+ */
+ return;
+ }
ASSERT3U(rc->rc_offset, ==, physical_rs.rs_start);
ASSERT3U(rc->rc_offset, <, physical_rs.rs_end);
/*
* It would be nice to assert that rs_end is equal
* to rc_offset + rc_size but there might be an
* optional I/O at the end that is not accounted in
* rc_size.
*/
if (physical_rs.rs_end > rc->rc_offset + rc->rc_size) {
ASSERT3U(physical_rs.rs_end, ==, rc->rc_offset +
- rc->rc_size + (1 << tvd->vdev_ashift));
+ rc->rc_size + (1 << zio->io_vd->vdev_top->vdev_ashift));
} else {
ASSERT3U(physical_rs.rs_end, ==, rc->rc_offset + rc->rc_size);
}
#endif
}
static void
-vdev_raidz_io_start_write(zio_t *zio, raidz_row_t *rr, uint64_t ashift)
+vdev_raidz_io_start_write(zio_t *zio, raidz_row_t *rr)
{
vdev_t *vd = zio->io_vd;
raidz_map_t *rm = zio->io_vsd;
vdev_raidz_generate_parity_row(rm, rr);
for (int c = 0; c < rr->rr_scols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
vdev_t *cvd = vd->vdev_child[rc->rc_devidx];
/* Verify physical to logical translation */
- vdev_raidz_io_verify(vd, rr, c);
+ vdev_raidz_io_verify(zio, rm, rr, c);
- if (rc->rc_size > 0) {
- ASSERT3P(rc->rc_abd, !=, NULL);
- zio_nowait(zio_vdev_child_io(zio, NULL, cvd,
- rc->rc_offset, rc->rc_abd,
- abd_get_size(rc->rc_abd), zio->io_type,
- zio->io_priority, 0, vdev_raidz_child_done, rc));
- } else {
- /*
- * Generate optional write for skip sector to improve
- * aggregation contiguity.
- */
- ASSERT3P(rc->rc_abd, ==, NULL);
- zio_nowait(zio_vdev_child_io(zio, NULL, cvd,
- rc->rc_offset, NULL, 1ULL << ashift,
- zio->io_type, zio->io_priority,
- ZIO_FLAG_NODATA | ZIO_FLAG_OPTIONAL, NULL,
- NULL));
+ if (rc->rc_size == 0)
+ continue;
+
+ ASSERT3U(rc->rc_offset + rc->rc_size, <,
+ cvd->vdev_psize - VDEV_LABEL_END_SIZE);
+
+ ASSERT3P(rc->rc_abd, !=, NULL);
+ zio_nowait(zio_vdev_child_io(zio, NULL, cvd,
+ rc->rc_offset, rc->rc_abd,
+ abd_get_size(rc->rc_abd), zio->io_type,
+ zio->io_priority, 0, vdev_raidz_child_done, rc));
+
+ if (rc->rc_shadow_devidx != INT_MAX) {
+ vdev_t *cvd2 = vd->vdev_child[rc->rc_shadow_devidx];
+
+ ASSERT3U(
+ rc->rc_shadow_offset + abd_get_size(rc->rc_abd), <,
+ cvd2->vdev_psize - VDEV_LABEL_END_SIZE);
+
+ zio_nowait(zio_vdev_child_io(zio, NULL, cvd2,
+ rc->rc_shadow_offset, rc->rc_abd,
+ abd_get_size(rc->rc_abd),
+ zio->io_type, zio->io_priority, 0,
+ vdev_raidz_shadow_child_done, rc));
}
}
}
+/*
+ * Generate optional I/Os for skip sectors to improve aggregation contiguity.
+ * This only works for vdev_raidz_map_alloc() (not _expanded()).
+ */
static void
-vdev_raidz_io_start_read(zio_t *zio, raidz_row_t *rr)
+raidz_start_skip_writes(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
-
- /*
+ uint64_t ashift = vd->vdev_top->vdev_ashift;
+ raidz_map_t *rm = zio->io_vsd;
+ ASSERT3U(rm->rm_nrows, ==, 1);
+ raidz_row_t *rr = rm->rm_row[0];
+ for (int c = 0; c < rr->rr_scols; c++) {
+ raidz_col_t *rc = &rr->rr_col[c];
+ vdev_t *cvd = vd->vdev_child[rc->rc_devidx];
+ if (rc->rc_size != 0)
+ continue;
+ ASSERT3P(rc->rc_abd, ==, NULL);
+
+ ASSERT3U(rc->rc_offset, <,
+ cvd->vdev_psize - VDEV_LABEL_END_SIZE);
+
+ zio_nowait(zio_vdev_child_io(zio, NULL, cvd, rc->rc_offset,
+ NULL, 1ULL << ashift, zio->io_type, zio->io_priority,
+ ZIO_FLAG_NODATA | ZIO_FLAG_OPTIONAL, NULL, NULL));
+ }
+}
+
+static void
+vdev_raidz_io_start_read_row(zio_t *zio, raidz_row_t *rr, boolean_t forceparity)
+{
+ vdev_t *vd = zio->io_vd;
+
+ /*
* Iterate over the columns in reverse order so that we hit the parity
* last -- any errors along the way will force us to read the parity.
*/
for (int c = rr->rr_cols - 1; c >= 0; c--) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_size == 0)
continue;
vdev_t *cvd = vd->vdev_child[rc->rc_devidx];
if (!vdev_readable(cvd)) {
if (c >= rr->rr_firstdatacol)
rr->rr_missingdata++;
else
rr->rr_missingparity++;
rc->rc_error = SET_ERROR(ENXIO);
rc->rc_tried = 1; /* don't even try */
rc->rc_skipped = 1;
continue;
}
if (vdev_dtl_contains(cvd, DTL_MISSING, zio->io_txg, 1)) {
if (c >= rr->rr_firstdatacol)
rr->rr_missingdata++;
else
rr->rr_missingparity++;
rc->rc_error = SET_ERROR(ESTALE);
rc->rc_skipped = 1;
continue;
}
- if (c >= rr->rr_firstdatacol || rr->rr_missingdata > 0 ||
+ if (forceparity ||
+ c >= rr->rr_firstdatacol || rr->rr_missingdata > 0 ||
(zio->io_flags & (ZIO_FLAG_SCRUB | ZIO_FLAG_RESILVER))) {
zio_nowait(zio_vdev_child_io(zio, NULL, cvd,
rc->rc_offset, rc->rc_abd, rc->rc_size,
zio->io_type, zio->io_priority, 0,
vdev_raidz_child_done, rc));
}
}
}
+static void
+vdev_raidz_io_start_read_phys_cols(zio_t *zio, raidz_map_t *rm)
+{
+ vdev_t *vd = zio->io_vd;
+
+ for (int i = 0; i < rm->rm_nphys_cols; i++) {
+ raidz_col_t *prc = &rm->rm_phys_col[i];
+ if (prc->rc_size == 0)
+ continue;
+
+ ASSERT3U(prc->rc_devidx, ==, i);
+ vdev_t *cvd = vd->vdev_child[i];
+ if (!vdev_readable(cvd)) {
+ prc->rc_error = SET_ERROR(ENXIO);
+ prc->rc_tried = 1; /* don't even try */
+ prc->rc_skipped = 1;
+ continue;
+ }
+ if (vdev_dtl_contains(cvd, DTL_MISSING, zio->io_txg, 1)) {
+ prc->rc_error = SET_ERROR(ESTALE);
+ prc->rc_skipped = 1;
+ continue;
+ }
+ zio_nowait(zio_vdev_child_io(zio, NULL, cvd,
+ prc->rc_offset, prc->rc_abd, prc->rc_size,
+ zio->io_type, zio->io_priority, 0,
+ vdev_raidz_child_done, prc));
+ }
+}
+
+static void
+vdev_raidz_io_start_read(zio_t *zio, raidz_map_t *rm)
+{
+ /*
+ * If there are multiple rows, we will be hitting
+ * all disks, so go ahead and read the parity so
+ * that we are reading in decent size chunks.
+ */
+ boolean_t forceparity = rm->rm_nrows > 1;
+
+ if (rm->rm_phys_col) {
+ vdev_raidz_io_start_read_phys_cols(zio, rm);
+ } else {
+ for (int i = 0; i < rm->rm_nrows; i++) {
+ raidz_row_t *rr = rm->rm_row[i];
+ vdev_raidz_io_start_read_row(zio, rr, forceparity);
+ }
+ }
+}
+
/*
* Start an IO operation on a RAIDZ VDev
*
* Outline:
* - For write operations:
* 1. Generate the parity data
* 2. Create child zio write operations to each column's vdev, for both
* data and parity.
* 3. If the column skips any sectors for padding, create optional dummy
* write zio children for those areas to improve aggregation continuity.
* - For read operations:
* 1. Create child zio read operations to each data column's vdev to read
* the range of data required for zio.
* 2. If this is a scrub or resilver operation, or if any of the data
* vdevs have had errors, then create zio read operations to the parity
* columns' VDevs as well.
*/
static void
vdev_raidz_io_start(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
vdev_t *tvd = vd->vdev_top;
vdev_raidz_t *vdrz = vd->vdev_tsd;
+ raidz_map_t *rm;
+
+ uint64_t logical_width = vdev_raidz_get_logical_width(vdrz,
+ BP_PHYSICAL_BIRTH(zio->io_bp));
+ if (logical_width != vdrz->vd_physical_width) {
+ zfs_locked_range_t *lr = NULL;
+ uint64_t synced_offset = UINT64_MAX;
+ uint64_t next_offset = UINT64_MAX;
+ boolean_t use_scratch = B_FALSE;
+ /*
+ * Note: when the expansion is completing, we set
+ * vre_state=DSS_FINISHED (in raidz_reflow_complete_sync())
+ * in a later txg than when we last update spa_ubsync's state
+ * (see the end of spa_raidz_expand_thread()). Therefore we
+ * may see vre_state!=SCANNING before
+ * VDEV_TOP_ZAP_RAIDZ_EXPAND_STATE=DSS_FINISHED is reflected
+ * on disk, but the copying progress has been synced to disk
+ * (and reflected in spa_ubsync). In this case it's fine to
+ * treat the expansion as completed, since if we crash there's
+ * no additional copying to do.
+ */
+ if (vdrz->vn_vre.vre_state == DSS_SCANNING) {
+ ASSERT3P(vd->vdev_spa->spa_raidz_expand, ==,
+ &vdrz->vn_vre);
+ lr = zfs_rangelock_enter(&vdrz->vn_vre.vre_rangelock,
+ zio->io_offset, zio->io_size, RL_READER);
+ use_scratch =
+ (RRSS_GET_STATE(&vd->vdev_spa->spa_ubsync) ==
+ RRSS_SCRATCH_VALID);
+ synced_offset =
+ RRSS_GET_OFFSET(&vd->vdev_spa->spa_ubsync);
+ next_offset = vdrz->vn_vre.vre_offset;
+ /*
+ * If we haven't resumed expanding since importing the
+ * pool, vre_offset won't have been set yet. In
+ * this case the next offset to be copied is the same
+ * as what was synced.
+ */
+ if (next_offset == UINT64_MAX) {
+ next_offset = synced_offset;
+ }
+ }
+ if (use_scratch) {
+ zfs_dbgmsg("zio=%px %s io_offset=%llu offset_synced="
+ "%lld next_offset=%lld use_scratch=%u",
+ zio,
+ zio->io_type == ZIO_TYPE_WRITE ? "WRITE" : "READ",
+ (long long)zio->io_offset,
+ (long long)synced_offset,
+ (long long)next_offset,
+ use_scratch);
+ }
+
+ rm = vdev_raidz_map_alloc_expanded(zio,
+ tvd->vdev_ashift, vdrz->vd_physical_width,
+ logical_width, vdrz->vd_nparity,
+ synced_offset, next_offset, use_scratch);
+ rm->rm_lr = lr;
+ } else {
+ rm = vdev_raidz_map_alloc(zio,
+ tvd->vdev_ashift, logical_width, vdrz->vd_nparity);
+ }
+ rm->rm_original_width = vdrz->vd_original_width;
- raidz_map_t *rm = vdev_raidz_map_alloc(zio, tvd->vdev_ashift,
- vdrz->vd_logical_width, vdrz->vd_nparity);
zio->io_vsd = rm;
zio->io_vsd_ops = &vdev_raidz_vsd_ops;
-
- /*
- * Until raidz expansion is implemented all maps for a raidz vdev
- * contain a single row.
- */
- ASSERT3U(rm->rm_nrows, ==, 1);
- raidz_row_t *rr = rm->rm_row[0];
-
if (zio->io_type == ZIO_TYPE_WRITE) {
- vdev_raidz_io_start_write(zio, rr, tvd->vdev_ashift);
+ for (int i = 0; i < rm->rm_nrows; i++) {
+ vdev_raidz_io_start_write(zio, rm->rm_row[i]);
+ }
+
+ if (logical_width == vdrz->vd_physical_width) {
+ raidz_start_skip_writes(zio);
+ }
} else {
ASSERT(zio->io_type == ZIO_TYPE_READ);
- vdev_raidz_io_start_read(zio, rr);
+ vdev_raidz_io_start_read(zio, rm);
}
zio_execute(zio);
}
/*
* Report a checksum error for a child of a RAID-Z device.
*/
void
vdev_raidz_checksum_error(zio_t *zio, raidz_col_t *rc, abd_t *bad_data)
{
vdev_t *vd = zio->io_vd->vdev_child[rc->rc_devidx];
if (!(zio->io_flags & ZIO_FLAG_SPECULATIVE) &&
zio->io_priority != ZIO_PRIORITY_REBUILD) {
zio_bad_cksum_t zbc;
raidz_map_t *rm = zio->io_vsd;
zbc.zbc_has_cksum = 0;
zbc.zbc_injected = rm->rm_ecksuminjected;
mutex_enter(&vd->vdev_stat_lock);
vd->vdev_stat.vs_checksum_errors++;
mutex_exit(&vd->vdev_stat_lock);
(void) zfs_ereport_post_checksum(zio->io_spa, vd,
&zio->io_bookmark, zio, rc->rc_offset, rc->rc_size,
rc->rc_abd, bad_data, &zbc);
}
}
/*
* We keep track of whether or not there were any injected errors, so that
* any ereports we generate can note it.
*/
static int
raidz_checksum_verify(zio_t *zio)
{
zio_bad_cksum_t zbc = {0};
raidz_map_t *rm = zio->io_vsd;
int ret = zio_checksum_error(zio, &zbc);
if (ret != 0 && zbc.zbc_injected != 0)
rm->rm_ecksuminjected = 1;
return (ret);
}
/*
* Generate the parity from the data columns. If we tried and were able to
* read the parity without error, verify that the generated parity matches the
* data we read. If it doesn't, we fire off a checksum error. Return the
* number of such failures.
*/
static int
raidz_parity_verify(zio_t *zio, raidz_row_t *rr)
{
abd_t *orig[VDEV_RAIDZ_MAXPARITY];
int c, ret = 0;
raidz_map_t *rm = zio->io_vsd;
raidz_col_t *rc;
blkptr_t *bp = zio->io_bp;
enum zio_checksum checksum = (bp == NULL ? zio->io_prop.zp_checksum :
(BP_IS_GANG(bp) ? ZIO_CHECKSUM_GANG_HEADER : BP_GET_CHECKSUM(bp)));
if (checksum == ZIO_CHECKSUM_NOPARITY)
return (ret);
for (c = 0; c < rr->rr_firstdatacol; c++) {
rc = &rr->rr_col[c];
if (!rc->rc_tried || rc->rc_error != 0)
continue;
orig[c] = rc->rc_abd;
ASSERT3U(abd_get_size(rc->rc_abd), ==, rc->rc_size);
rc->rc_abd = abd_alloc_linear(rc->rc_size, B_FALSE);
}
/*
* Verify any empty sectors are zero filled to ensure the parity
* is calculated correctly even if these non-data sectors are damaged.
*/
if (rr->rr_nempty && rr->rr_abd_empty != NULL)
ret += vdev_draid_map_verify_empty(zio, rr);
/*
* Regenerates parity even for !tried||rc_error!=0 columns. This
* isn't harmful but it does have the side effect of fixing stuff
* we didn't realize was necessary (i.e. even if we return 0).
*/
vdev_raidz_generate_parity_row(rm, rr);
for (c = 0; c < rr->rr_firstdatacol; c++) {
rc = &rr->rr_col[c];
if (!rc->rc_tried || rc->rc_error != 0)
continue;
if (abd_cmp(orig[c], rc->rc_abd) != 0) {
+ zfs_dbgmsg("found error on col=%u devidx=%u off %llx",
+ c, (int)rc->rc_devidx, (u_longlong_t)rc->rc_offset);
vdev_raidz_checksum_error(zio, rc, orig[c]);
rc->rc_error = SET_ERROR(ECKSUM);
ret++;
}
abd_free(orig[c]);
}
return (ret);
}
static int
vdev_raidz_worst_error(raidz_row_t *rr)
{
int error = 0;
- for (int c = 0; c < rr->rr_cols; c++)
+ for (int c = 0; c < rr->rr_cols; c++) {
error = zio_worst_error(error, rr->rr_col[c].rc_error);
+ error = zio_worst_error(error, rr->rr_col[c].rc_shadow_error);
+ }
return (error);
}
static void
vdev_raidz_io_done_verified(zio_t *zio, raidz_row_t *rr)
{
int unexpected_errors = 0;
int parity_errors = 0;
int parity_untried = 0;
int data_errors = 0;
ASSERT3U(zio->io_type, ==, ZIO_TYPE_READ);
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_error) {
if (c < rr->rr_firstdatacol)
parity_errors++;
else
data_errors++;
if (!rc->rc_skipped)
unexpected_errors++;
} else if (c < rr->rr_firstdatacol && !rc->rc_tried) {
parity_untried++;
}
if (rc->rc_force_repair)
unexpected_errors++;
}
/*
* If we read more parity disks than were used for
* reconstruction, confirm that the other parity disks produced
* correct data.
*
* Note that we also regenerate parity when resilvering so we
* can write it out to failed devices later.
*/
if (parity_errors + parity_untried <
rr->rr_firstdatacol - data_errors ||
(zio->io_flags & ZIO_FLAG_RESILVER)) {
int n = raidz_parity_verify(zio, rr);
unexpected_errors += n;
}
if (zio->io_error == 0 && spa_writeable(zio->io_spa) &&
(unexpected_errors > 0 || (zio->io_flags & ZIO_FLAG_RESILVER))) {
/*
* Use the good data we have in hand to repair damaged children.
*/
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
vdev_t *vd = zio->io_vd;
vdev_t *cvd = vd->vdev_child[rc->rc_devidx];
if (!rc->rc_allow_repair) {
continue;
} else if (!rc->rc_force_repair &&
(rc->rc_error == 0 || rc->rc_size == 0)) {
continue;
}
+ zfs_dbgmsg("zio=%px repairing c=%u devidx=%u "
+ "offset=%llx",
+ zio, c, rc->rc_devidx, (long long)rc->rc_offset);
+
zio_nowait(zio_vdev_child_io(zio, NULL, cvd,
rc->rc_offset, rc->rc_abd, rc->rc_size,
ZIO_TYPE_WRITE,
zio->io_priority == ZIO_PRIORITY_REBUILD ?
ZIO_PRIORITY_REBUILD : ZIO_PRIORITY_ASYNC_WRITE,
ZIO_FLAG_IO_REPAIR | (unexpected_errors ?
ZIO_FLAG_SELF_HEAL : 0), NULL, NULL));
}
}
+
+ /*
+ * Scrub or resilver i/o's: overwrite any shadow locations with the
+ * good data. This ensures that if we've already copied this sector,
+ * it will be corrected if it was damaged. This writes more than is
+ * necessary, but since expansion is paused during scrub/resilver, at
+ * most a single row will have a shadow location.
+ */
+ if (zio->io_error == 0 && spa_writeable(zio->io_spa) &&
+ (zio->io_flags & (ZIO_FLAG_RESILVER | ZIO_FLAG_SCRUB))) {
+ for (int c = 0; c < rr->rr_cols; c++) {
+ raidz_col_t *rc = &rr->rr_col[c];
+ vdev_t *vd = zio->io_vd;
+
+ if (rc->rc_shadow_devidx == INT_MAX || rc->rc_size == 0)
+ continue;
+ vdev_t *cvd = vd->vdev_child[rc->rc_shadow_devidx];
+
+ /*
+ * Note: We don't want to update the repair stats
+ * because that would incorrectly indicate that there
+ * was bad data to repair, which we aren't sure about.
+ * By clearing the SCAN_THREAD flag, we prevent this
+ * from happening, despite having the REPAIR flag set.
+ * We need to set SELF_HEAL so that this i/o can't be
+ * bypassed by zio_vdev_io_start().
+ */
+ zio_t *cio = zio_vdev_child_io(zio, NULL, cvd,
+ rc->rc_shadow_offset, rc->rc_abd, rc->rc_size,
+ ZIO_TYPE_WRITE, ZIO_PRIORITY_ASYNC_WRITE,
+ ZIO_FLAG_IO_REPAIR | ZIO_FLAG_SELF_HEAL,
+ NULL, NULL);
+ cio->io_flags &= ~ZIO_FLAG_SCAN_THREAD;
+ zio_nowait(cio);
+ }
+ }
}
static void
raidz_restore_orig_data(raidz_map_t *rm)
{
for (int i = 0; i < rm->rm_nrows; i++) {
raidz_row_t *rr = rm->rm_row[i];
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_need_orig_restore) {
abd_copy(rc->rc_abd,
rc->rc_orig_data, rc->rc_size);
rc->rc_need_orig_restore = B_FALSE;
}
}
}
}
+/*
+ * During raidz_reconstruct() for expanded VDEV, we need special consideration
+ * failure simulations. See note in raidz_reconstruct() on simulating failure
+ * of a pre-expansion device.
+ *
+ * Treating logical child i as failed, return TRUE if the given column should
+ * be treated as failed. The idea of logical children allows us to imagine
+ * that a disk silently failed before a RAIDZ expansion (reads from this disk
+ * succeed but return the wrong data). Since the expansion doesn't verify
+ * checksums, the incorrect data will be moved to new locations spread among
+ * the children (going diagonally across them).
+ *
+ * Higher "logical child failures" (values of `i`) indicate these
+ * "pre-expansion failures". The first physical_width values imagine that a
+ * current child failed; the next physical_width-1 values imagine that a
+ * child failed before the most recent expansion; the next physical_width-2
+ * values imagine a child failed in the expansion before that, etc.
+ */
+static boolean_t
+raidz_simulate_failure(int physical_width, int original_width, int ashift,
+ int i, raidz_col_t *rc)
+{
+ uint64_t sector_id =
+ physical_width * (rc->rc_offset >> ashift) +
+ rc->rc_devidx;
+
+ for (int w = physical_width; w >= original_width; w--) {
+ if (i < w) {
+ return (sector_id % w == i);
+ } else {
+ i -= w;
+ }
+ }
+ ASSERT(!"invalid logical child id");
+ return (B_FALSE);
+}
+
/*
* returns EINVAL if reconstruction of the block will not be possible
* returns ECKSUM if this specific reconstruction failed
* returns 0 on successful reconstruction
*/
static int
raidz_reconstruct(zio_t *zio, int *ltgts, int ntgts, int nparity)
{
raidz_map_t *rm = zio->io_vsd;
+ int physical_width = zio->io_vd->vdev_children;
+ int original_width = (rm->rm_original_width != 0) ?
+ rm->rm_original_width : physical_width;
+ int dbgmsg = zfs_flags & ZFS_DEBUG_RAIDZ_RECONSTRUCT;
+
+ if (dbgmsg) {
+ zfs_dbgmsg("raidz_reconstruct_expanded(zio=%px ltgts=%u,%u,%u "
+ "ntgts=%u", zio, ltgts[0], ltgts[1], ltgts[2], ntgts);
+ }
/* Reconstruct each row */
for (int r = 0; r < rm->rm_nrows; r++) {
raidz_row_t *rr = rm->rm_row[r];
int my_tgts[VDEV_RAIDZ_MAXPARITY]; /* value is child id */
int t = 0;
int dead = 0;
int dead_data = 0;
+ if (dbgmsg)
+ zfs_dbgmsg("raidz_reconstruct_expanded(row=%u)", r);
+
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
ASSERT0(rc->rc_need_orig_restore);
if (rc->rc_error != 0) {
dead++;
if (c >= nparity)
dead_data++;
continue;
}
if (rc->rc_size == 0)
continue;
for (int lt = 0; lt < ntgts; lt++) {
- if (rc->rc_devidx == ltgts[lt]) {
+ if (raidz_simulate_failure(physical_width,
+ original_width,
+ zio->io_vd->vdev_top->vdev_ashift,
+ ltgts[lt], rc)) {
if (rc->rc_orig_data == NULL) {
rc->rc_orig_data =
abd_alloc_linear(
rc->rc_size, B_TRUE);
abd_copy(rc->rc_orig_data,
rc->rc_abd, rc->rc_size);
}
rc->rc_need_orig_restore = B_TRUE;
dead++;
if (c >= nparity)
dead_data++;
- my_tgts[t++] = c;
+ /*
+ * Note: simulating failure of a
+ * pre-expansion device can hit more
+ * than one column, in which case we
+ * might try to simulate more failures
+ * than can be reconstructed, which is
+ * also more than the size of my_tgts.
+ * This check prevents accessing past
+ * the end of my_tgts. The "dead >
+ * nparity" check below will fail this
+ * reconstruction attempt.
+ */
+ if (t < VDEV_RAIDZ_MAXPARITY) {
+ my_tgts[t++] = c;
+ if (dbgmsg) {
+ zfs_dbgmsg("simulating "
+ "failure of col %u "
+ "devidx %u", c,
+ (int)rc->rc_devidx);
+ }
+ }
break;
}
}
}
if (dead > nparity) {
/* reconstruction not possible */
+ if (dbgmsg) {
+ zfs_dbgmsg("reconstruction not possible; "
+ "too many failures");
+ }
raidz_restore_orig_data(rm);
return (EINVAL);
}
if (dead_data > 0)
vdev_raidz_reconstruct_row(rm, rr, my_tgts, t);
}
/* Check for success */
if (raidz_checksum_verify(zio) == 0) {
/* Reconstruction succeeded - report errors */
for (int i = 0; i < rm->rm_nrows; i++) {
raidz_row_t *rr = rm->rm_row[i];
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_need_orig_restore) {
/*
* Note: if this is a parity column,
* we don't really know if it's wrong.
* We need to let
* vdev_raidz_io_done_verified() check
* it, and if we set rc_error, it will
* think that it is a "known" error
* that doesn't need to be checked
* or corrected.
*/
if (rc->rc_error == 0 &&
c >= rr->rr_firstdatacol) {
vdev_raidz_checksum_error(zio,
rc, rc->rc_orig_data);
rc->rc_error =
SET_ERROR(ECKSUM);
}
rc->rc_need_orig_restore = B_FALSE;
}
}
vdev_raidz_io_done_verified(zio, rr);
}
zio_checksum_verified(zio);
+ if (dbgmsg) {
+ zfs_dbgmsg("reconstruction successful "
+ "(checksum verified)");
+ }
return (0);
}
/* Reconstruction failed - restore original data */
raidz_restore_orig_data(rm);
+ if (dbgmsg) {
+ zfs_dbgmsg("raidz_reconstruct_expanded(zio=%px) checksum "
+ "failed", zio);
+ }
return (ECKSUM);
}
/*
* Iterate over all combinations of N bad vdevs and attempt a reconstruction.
* Note that the algorithm below is non-optimal because it doesn't take into
* account how reconstruction is actually performed. For example, with
* triple-parity RAID-Z the reconstruction procedure is the same if column 4
* is targeted as invalid as if columns 1 and 4 are targeted since in both
* cases we'd only use parity information in column 0.
*
* The order that we find the various possible combinations of failed
* disks is dictated by these rules:
* - Examine each "slot" (the "i" in tgts[i])
- * - Try to increment this slot (tgts[i] = tgts[i] + 1)
+ * - Try to increment this slot (tgts[i] += 1)
* - if we can't increment because it runs into the next slot,
* reset our slot to the minimum, and examine the next slot
*
* For example, with a 6-wide RAIDZ3, and no known errors (so we have to choose
* 3 columns to reconstruct), we will generate the following sequence:
*
* STATE ACTION
* 0 1 2 special case: skip since these are all parity
* 0 1 3 first slot: reset to 0; middle slot: increment to 2
* 0 2 3 first slot: increment to 1
* 1 2 3 first: reset to 0; middle: reset to 1; last: increment to 4
* 0 1 4 first: reset to 0; middle: increment to 2
* 0 2 4 first: increment to 1
* 1 2 4 first: reset to 0; middle: increment to 3
* 0 3 4 first: increment to 1
* 1 3 4 first: increment to 2
* 2 3 4 first: reset to 0; middle: reset to 1; last: increment to 5
* 0 1 5 first: reset to 0; middle: increment to 2
* 0 2 5 first: increment to 1
* 1 2 5 first: reset to 0; middle: increment to 3
* 0 3 5 first: increment to 1
* 1 3 5 first: increment to 2
* 2 3 5 first: reset to 0; middle: increment to 4
* 0 4 5 first: increment to 1
* 1 4 5 first: increment to 2
* 2 4 5 first: increment to 3
* 3 4 5 done
*
* This strategy works for dRAID but is less efficient when there are a large
* number of child vdevs and therefore permutations to check. Furthermore,
- * since the raidz_map_t rows likely do not overlap reconstruction would be
+ * since the raidz_map_t rows likely do not overlap, reconstruction would be
* possible as long as there are no more than nparity data errors per row.
* These additional permutations are not currently checked but could be as
* a future improvement.
+ *
+ * Returns 0 on success, ECKSUM on failure.
*/
static int
vdev_raidz_combrec(zio_t *zio)
{
int nparity = vdev_get_nparity(zio->io_vd);
raidz_map_t *rm = zio->io_vsd;
+ int physical_width = zio->io_vd->vdev_children;
+ int original_width = (rm->rm_original_width != 0) ?
+ rm->rm_original_width : physical_width;
- /* Check if there's enough data to attempt reconstrution. */
for (int i = 0; i < rm->rm_nrows; i++) {
raidz_row_t *rr = rm->rm_row[i];
int total_errors = 0;
for (int c = 0; c < rr->rr_cols; c++) {
if (rr->rr_col[c].rc_error)
total_errors++;
}
if (total_errors > nparity)
return (vdev_raidz_worst_error(rr));
}
for (int num_failures = 1; num_failures <= nparity; num_failures++) {
int tstore[VDEV_RAIDZ_MAXPARITY + 2];
int *ltgts = &tstore[1]; /* value is logical child ID */
- /* Determine number of logical children, n */
- int n = zio->io_vd->vdev_children;
+
+ /*
+ * Determine number of logical children, n. See comment
+ * above raidz_simulate_failure().
+ */
+ int n = 0;
+ for (int w = physical_width;
+ w >= original_width; w--) {
+ n += w;
+ }
ASSERT3U(num_failures, <=, nparity);
ASSERT3U(num_failures, <=, VDEV_RAIDZ_MAXPARITY);
/* Handle corner cases in combrec logic */
ltgts[-1] = -1;
for (int i = 0; i < num_failures; i++) {
ltgts[i] = i;
}
ltgts[num_failures] = n;
for (;;) {
int err = raidz_reconstruct(zio, ltgts, num_failures,
nparity);
if (err == EINVAL) {
/*
* Reconstruction not possible with this #
* failures; try more failures.
*/
break;
} else if (err == 0)
return (0);
/* Compute next targets to try */
for (int t = 0; ; t++) {
ASSERT3U(t, <, num_failures);
ltgts[t]++;
if (ltgts[t] == n) {
/* try more failures */
ASSERT3U(t, ==, num_failures - 1);
+ if (zfs_flags &
+ ZFS_DEBUG_RAIDZ_RECONSTRUCT) {
+ zfs_dbgmsg("reconstruction "
+ "failed for num_failures="
+ "%u; tried all "
+ "combinations",
+ num_failures);
+ }
break;
}
ASSERT3U(ltgts[t], <, n);
ASSERT3U(ltgts[t], <=, ltgts[t + 1]);
/*
* If that spot is available, we're done here.
* Try the next combination.
*/
if (ltgts[t] != ltgts[t + 1])
- break;
+ break; // found next combination
/*
* Otherwise, reset this tgt to the minimum,
* and move on to the next tgt.
*/
ltgts[t] = ltgts[t - 1] + 1;
ASSERT3U(ltgts[t], ==, t);
}
/* Increase the number of failures and keep trying. */
if (ltgts[num_failures - 1] == n)
break;
}
}
-
+ if (zfs_flags & ZFS_DEBUG_RAIDZ_RECONSTRUCT)
+ zfs_dbgmsg("reconstruction failed for all num_failures");
return (ECKSUM);
}
void
vdev_raidz_reconstruct(raidz_map_t *rm, const int *t, int nt)
{
for (uint64_t row = 0; row < rm->rm_nrows; row++) {
raidz_row_t *rr = rm->rm_row[row];
vdev_raidz_reconstruct_row(rm, rr, t, nt);
}
}
/*
* Complete a write IO operation on a RAIDZ VDev
*
* Outline:
* 1. Check for errors on the child IOs.
* 2. Return, setting an error code if too few child VDevs were written
* to reconstruct the data later. Note that partial writes are
* considered successful if they can be reconstructed at all.
*/
static void
vdev_raidz_io_done_write_impl(zio_t *zio, raidz_row_t *rr)
{
- int total_errors = 0;
+ int normal_errors = 0;
+ int shadow_errors = 0;
ASSERT3U(rr->rr_missingparity, <=, rr->rr_firstdatacol);
ASSERT3U(rr->rr_missingdata, <=, rr->rr_cols - rr->rr_firstdatacol);
ASSERT3U(zio->io_type, ==, ZIO_TYPE_WRITE);
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
- if (rc->rc_error) {
+ if (rc->rc_error != 0) {
ASSERT(rc->rc_error != ECKSUM); /* child has no bp */
-
- total_errors++;
+ normal_errors++;
+ }
+ if (rc->rc_shadow_error != 0) {
+ ASSERT(rc->rc_shadow_error != ECKSUM);
+ shadow_errors++;
}
}
/*
* Treat partial writes as a success. If we couldn't write enough
- * columns to reconstruct the data, the I/O failed. Otherwise,
- * good enough.
+ * columns to reconstruct the data, the I/O failed. Otherwise, good
+ * enough. Note that in the case of a shadow write (during raidz
+ * expansion), depending on if we crash, either the normal (old) or
+ * shadow (new) location may become the "real" version of the block,
+ * so both locations must have sufficient redundancy.
*
* Now that we support write reallocation, it would be better
* to treat partial failure as real failure unless there are
* no non-degraded top-level vdevs left, and not update DTLs
* if we intend to reallocate.
*/
- if (total_errors > rr->rr_firstdatacol) {
+ if (normal_errors > rr->rr_firstdatacol ||
+ shadow_errors > rr->rr_firstdatacol) {
zio->io_error = zio_worst_error(zio->io_error,
vdev_raidz_worst_error(rr));
}
}
static void
vdev_raidz_io_done_reconstruct_known_missing(zio_t *zio, raidz_map_t *rm,
raidz_row_t *rr)
{
int parity_errors = 0;
int parity_untried = 0;
int data_errors = 0;
int total_errors = 0;
ASSERT3U(rr->rr_missingparity, <=, rr->rr_firstdatacol);
ASSERT3U(rr->rr_missingdata, <=, rr->rr_cols - rr->rr_firstdatacol);
- ASSERT3U(zio->io_type, ==, ZIO_TYPE_READ);
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
/*
* If scrubbing and a replacing/sparing child vdev determined
* that not all of its children have an identical copy of the
* data, then clear the error so the column is treated like
* any other read and force a repair to correct the damage.
*/
if (rc->rc_error == ECKSUM) {
ASSERT(zio->io_flags & ZIO_FLAG_SCRUB);
vdev_raidz_checksum_error(zio, rc, rc->rc_abd);
rc->rc_force_repair = 1;
rc->rc_error = 0;
}
if (rc->rc_error) {
if (c < rr->rr_firstdatacol)
parity_errors++;
else
data_errors++;
total_errors++;
} else if (c < rr->rr_firstdatacol && !rc->rc_tried) {
parity_untried++;
}
}
/*
* If there were data errors and the number of errors we saw was
* correctable -- less than or equal to the number of parity disks read
* -- reconstruct based on the missing data.
*/
if (data_errors != 0 &&
total_errors <= rr->rr_firstdatacol - parity_untried) {
/*
* We either attempt to read all the parity columns or
* none of them. If we didn't try to read parity, we
* wouldn't be here in the correctable case. There must
* also have been fewer parity errors than parity
* columns or, again, we wouldn't be in this code path.
*/
ASSERT(parity_untried == 0);
ASSERT(parity_errors < rr->rr_firstdatacol);
/*
* Identify the data columns that reported an error.
*/
int n = 0;
int tgts[VDEV_RAIDZ_MAXPARITY];
for (int c = rr->rr_firstdatacol; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_error != 0) {
ASSERT(n < VDEV_RAIDZ_MAXPARITY);
tgts[n++] = c;
}
}
ASSERT(rr->rr_firstdatacol >= n);
vdev_raidz_reconstruct_row(rm, rr, tgts, n);
}
}
/*
* Return the number of reads issued.
*/
static int
vdev_raidz_read_all(zio_t *zio, raidz_row_t *rr)
{
vdev_t *vd = zio->io_vd;
int nread = 0;
rr->rr_missingdata = 0;
rr->rr_missingparity = 0;
/*
* If this rows contains empty sectors which are not required
* for a normal read then allocate an ABD for them now so they
* may be read, verified, and any needed repairs performed.
*/
- if (rr->rr_nempty && rr->rr_abd_empty == NULL)
+ if (rr->rr_nempty != 0 && rr->rr_abd_empty == NULL)
vdev_draid_map_alloc_empty(zio, rr);
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
if (rc->rc_tried || rc->rc_size == 0)
continue;
zio_nowait(zio_vdev_child_io(zio, NULL,
vd->vdev_child[rc->rc_devidx],
rc->rc_offset, rc->rc_abd, rc->rc_size,
zio->io_type, zio->io_priority, 0,
vdev_raidz_child_done, rc));
nread++;
}
return (nread);
}
/*
* We're here because either there were too many errors to even attempt
* reconstruction (total_errors == rm_first_datacol), or vdev_*_combrec()
* failed. In either case, there is enough bad data to prevent reconstruction.
* Start checksum ereports for all children which haven't failed.
*/
static void
vdev_raidz_io_done_unrecoverable(zio_t *zio)
{
raidz_map_t *rm = zio->io_vsd;
for (int i = 0; i < rm->rm_nrows; i++) {
raidz_row_t *rr = rm->rm_row[i];
for (int c = 0; c < rr->rr_cols; c++) {
raidz_col_t *rc = &rr->rr_col[c];
vdev_t *cvd = zio->io_vd->vdev_child[rc->rc_devidx];
if (rc->rc_error != 0)
continue;
zio_bad_cksum_t zbc;
zbc.zbc_has_cksum = 0;
zbc.zbc_injected = rm->rm_ecksuminjected;
mutex_enter(&cvd->vdev_stat_lock);
cvd->vdev_stat.vs_checksum_errors++;
mutex_exit(&cvd->vdev_stat_lock);
(void) zfs_ereport_start_checksum(zio->io_spa,
cvd, &zio->io_bookmark, zio, rc->rc_offset,
rc->rc_size, &zbc);
}
}
}
void
vdev_raidz_io_done(zio_t *zio)
{
raidz_map_t *rm = zio->io_vsd;
+ ASSERT(zio->io_bp != NULL);
if (zio->io_type == ZIO_TYPE_WRITE) {
for (int i = 0; i < rm->rm_nrows; i++) {
vdev_raidz_io_done_write_impl(zio, rm->rm_row[i]);
}
} else {
+ if (rm->rm_phys_col) {
+ /*
+ * This is an aggregated read. Copy the data and status
+ * from the aggregate abd's to the individual rows.
+ */
+ for (int i = 0; i < rm->rm_nrows; i++) {
+ raidz_row_t *rr = rm->rm_row[i];
+
+ for (int c = 0; c < rr->rr_cols; c++) {
+ raidz_col_t *rc = &rr->rr_col[c];
+ if (rc->rc_tried || rc->rc_size == 0)
+ continue;
+
+ raidz_col_t *prc =
+ &rm->rm_phys_col[rc->rc_devidx];
+ rc->rc_error = prc->rc_error;
+ rc->rc_tried = prc->rc_tried;
+ rc->rc_skipped = prc->rc_skipped;
+ if (c >= rr->rr_firstdatacol) {
+ /*
+ * Note: this is slightly faster
+ * than using abd_copy_off().
+ */
+ char *physbuf = abd_to_buf(
+ prc->rc_abd);
+ void *physloc = physbuf +
+ rc->rc_offset -
+ prc->rc_offset;
+
+ abd_copy_from_buf(rc->rc_abd,
+ physloc, rc->rc_size);
+ }
+ }
+ }
+ }
+
for (int i = 0; i < rm->rm_nrows; i++) {
raidz_row_t *rr = rm->rm_row[i];
vdev_raidz_io_done_reconstruct_known_missing(zio,
rm, rr);
}
if (raidz_checksum_verify(zio) == 0) {
for (int i = 0; i < rm->rm_nrows; i++) {
raidz_row_t *rr = rm->rm_row[i];
vdev_raidz_io_done_verified(zio, rr);
}
zio_checksum_verified(zio);
} else {
/*
* A sequential resilver has no checksum which makes
* combinatoral reconstruction impossible. This code
* path is unreachable since raidz_checksum_verify()
* has no checksum to verify and must succeed.
*/
ASSERT3U(zio->io_priority, !=, ZIO_PRIORITY_REBUILD);
/*
* This isn't a typical situation -- either we got a
* read error or a child silently returned bad data.
* Read every block so we can try again with as much
* data and parity as we can track down. If we've
* already been through once before, all children will
* be marked as tried so we'll proceed to combinatorial
* reconstruction.
*/
int nread = 0;
for (int i = 0; i < rm->rm_nrows; i++) {
nread += vdev_raidz_read_all(zio,
rm->rm_row[i]);
}
if (nread != 0) {
/*
* Normally our stage is VDEV_IO_DONE, but if
* we've already called redone(), it will have
* changed to VDEV_IO_START, in which case we
* don't want to call redone() again.
*/
if (zio->io_stage != ZIO_STAGE_VDEV_IO_START)
zio_vdev_io_redone(zio);
return;
}
-
+ /*
+ * It would be too expensive to try every possible
+ * combination of failed sectors in every row, so
+ * instead we try every combination of failed current or
+ * past physical disk. This means that if the incorrect
+ * sectors were all on Nparity disks at any point in the
+ * past, we will find the correct data. The only known
+ * case where this is less durable than a non-expanded
+ * RAIDZ, is if we have a silent failure during
+ * expansion. In that case, one block could be
+ * partially in the old format and partially in the
+ * new format, so we'd lost some sectors from the old
+ * format and some from the new format.
+ *
+ * e.g. logical_width=4 physical_width=6
+ * the 15 (6+5+4) possible failed disks are:
+ * width=6 child=0
+ * width=6 child=1
+ * width=6 child=2
+ * width=6 child=3
+ * width=6 child=4
+ * width=6 child=5
+ * width=5 child=0
+ * width=5 child=1
+ * width=5 child=2
+ * width=5 child=3
+ * width=5 child=4
+ * width=4 child=0
+ * width=4 child=1
+ * width=4 child=2
+ * width=4 child=3
+ * And we will try every combination of Nparity of these
+ * failing.
+ *
+ * As a first pass, we can generate every combo,
+ * and try reconstructing, ignoring any known
+ * failures. If any row has too many known + simulated
+ * failures, then we bail on reconstructing with this
+ * number of simulated failures. As an improvement,
+ * we could detect the number of whole known failures
+ * (i.e. we have known failures on these disks for
+ * every row; the disks never succeeded), and
+ * subtract that from the max # failures to simulate.
+ * We could go even further like the current
+ * combrec code, but that doesn't seem like it
+ * gains us very much. If we simulate a failure
+ * that is also a known failure, that's fine.
+ */
zio->io_error = vdev_raidz_combrec(zio);
if (zio->io_error == ECKSUM &&
!(zio->io_flags & ZIO_FLAG_SPECULATIVE)) {
vdev_raidz_io_done_unrecoverable(zio);
}
}
}
+ if (rm->rm_lr != NULL) {
+ zfs_rangelock_exit(rm->rm_lr);
+ rm->rm_lr = NULL;
+ }
}
static void
vdev_raidz_state_change(vdev_t *vd, int faulted, int degraded)
{
vdev_raidz_t *vdrz = vd->vdev_tsd;
if (faulted > vdrz->vd_nparity)
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_NO_REPLICAS);
else if (degraded + faulted != 0)
vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, VDEV_AUX_NONE);
else
vdev_set_state(vd, B_FALSE, VDEV_STATE_HEALTHY, VDEV_AUX_NONE);
}
/*
* Determine if any portion of the provided block resides on a child vdev
* with a dirty DTL and therefore needs to be resilvered. The function
* assumes that at least one DTL is dirty which implies that full stripe
* width blocks must be resilvered.
*/
static boolean_t
vdev_raidz_need_resilver(vdev_t *vd, const dva_t *dva, size_t psize,
uint64_t phys_birth)
{
vdev_raidz_t *vdrz = vd->vdev_tsd;
+
+ /*
+ * If we're in the middle of a RAIDZ expansion, this block may be in
+ * the old and/or new location. For simplicity, always resilver it.
+ */
+ if (vdrz->vn_vre.vre_state == DSS_SCANNING)
+ return (B_TRUE);
+
uint64_t dcols = vd->vdev_children;
uint64_t nparity = vdrz->vd_nparity;
uint64_t ashift = vd->vdev_top->vdev_ashift;
/* The starting RAIDZ (parent) vdev sector of the block. */
uint64_t b = DVA_GET_OFFSET(dva) >> ashift;
/* The zio's size in units of the vdev's minimum sector size. */
uint64_t s = ((psize - 1) >> ashift) + 1;
/* The first column for this stripe. */
uint64_t f = b % dcols;
/* Unreachable by sequential resilver. */
ASSERT3U(phys_birth, !=, TXG_UNKNOWN);
if (!vdev_dtl_contains(vd, DTL_PARTIAL, phys_birth, 1))
return (B_FALSE);
if (s + nparity >= dcols)
return (B_TRUE);
for (uint64_t c = 0; c < s + nparity; c++) {
uint64_t devidx = (f + c) % dcols;
vdev_t *cvd = vd->vdev_child[devidx];
/*
* dsl_scan_need_resilver() already checked vd with
* vdev_dtl_contains(). So here just check cvd with
* vdev_dtl_empty(), cheaper and a good approximation.
*/
if (!vdev_dtl_empty(cvd, DTL_PARTIAL))
return (B_TRUE);
}
return (B_FALSE);
}
static void
vdev_raidz_xlate(vdev_t *cvd, const range_seg64_t *logical_rs,
range_seg64_t *physical_rs, range_seg64_t *remain_rs)
{
(void) remain_rs;
vdev_t *raidvd = cvd->vdev_parent;
ASSERT(raidvd->vdev_ops == &vdev_raidz_ops);
- uint64_t width = raidvd->vdev_children;
+ vdev_raidz_t *vdrz = raidvd->vdev_tsd;
+
+ if (vdrz->vn_vre.vre_state == DSS_SCANNING) {
+ /*
+ * We're in the middle of expansion, in which case the
+ * translation is in flux. Any answer we give may be wrong
+ * by the time we return, so it isn't safe for the caller to
+ * act on it. Therefore we say that this range isn't present
+ * on any children. The only consumers of this are "zpool
+ * initialize" and trimming, both of which are "best effort"
+ * anyway.
+ */
+ physical_rs->rs_start = physical_rs->rs_end = 0;
+ remain_rs->rs_start = remain_rs->rs_end = 0;
+ return;
+ }
+
+ uint64_t width = vdrz->vd_physical_width;
uint64_t tgt_col = cvd->vdev_id;
uint64_t ashift = raidvd->vdev_top->vdev_ashift;
/* make sure the offsets are block-aligned */
ASSERT0(logical_rs->rs_start % (1 << ashift));
ASSERT0(logical_rs->rs_end % (1 << ashift));
uint64_t b_start = logical_rs->rs_start >> ashift;
uint64_t b_end = logical_rs->rs_end >> ashift;
uint64_t start_row = 0;
if (b_start > tgt_col) /* avoid underflow */
start_row = ((b_start - tgt_col - 1) / width) + 1;
uint64_t end_row = 0;
if (b_end > tgt_col)
end_row = ((b_end - tgt_col - 1) / width) + 1;
physical_rs->rs_start = start_row << ashift;
physical_rs->rs_end = end_row << ashift;
ASSERT3U(physical_rs->rs_start, <=, logical_rs->rs_start);
ASSERT3U(physical_rs->rs_end - physical_rs->rs_start, <=,
logical_rs->rs_end - logical_rs->rs_start);
}
-/*
- * Initialize private RAIDZ specific fields from the nvlist.
- */
-static int
-vdev_raidz_init(spa_t *spa, nvlist_t *nv, void **tsd)
+static void
+raidz_reflow_sync(void *arg, dmu_tx_t *tx)
{
- vdev_raidz_t *vdrz;
- uint64_t nparity;
+ spa_t *spa = arg;
+ int txgoff = dmu_tx_get_txg(tx) & TXG_MASK;
+ vdev_raidz_expand_t *vre = spa->spa_raidz_expand;
- uint_t children;
- nvlist_t **child;
- int error = nvlist_lookup_nvlist_array(nv,
- ZPOOL_CONFIG_CHILDREN, &child, &children);
- if (error != 0)
- return (SET_ERROR(EINVAL));
+ /*
+ * Ensure there are no i/os to the range that is being committed.
+ */
+ uint64_t old_offset = RRSS_GET_OFFSET(&spa->spa_uberblock);
+ ASSERT3U(vre->vre_offset_pertxg[txgoff], >=, old_offset);
- if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY, &nparity) == 0) {
- if (nparity == 0 || nparity > VDEV_RAIDZ_MAXPARITY)
- return (SET_ERROR(EINVAL));
+ mutex_enter(&vre->vre_lock);
+ uint64_t new_offset =
+ MIN(vre->vre_offset_pertxg[txgoff], vre->vre_failed_offset);
+ /*
+ * We should not have committed anything that failed.
+ */
+ VERIFY3U(vre->vre_failed_offset, >=, old_offset);
+ mutex_exit(&vre->vre_lock);
- /*
- * Previous versions could only support 1 or 2 parity
- * device.
- */
- if (nparity > 1 && spa_version(spa) < SPA_VERSION_RAIDZ2)
- return (SET_ERROR(EINVAL));
- else if (nparity > 2 && spa_version(spa) < SPA_VERSION_RAIDZ3)
- return (SET_ERROR(EINVAL));
- } 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));
+ zfs_locked_range_t *lr = zfs_rangelock_enter(&vre->vre_rangelock,
+ old_offset, new_offset - old_offset,
+ RL_WRITER);
- /*
- * Otherwise, we default to 1 parity device for RAID-Z.
- */
- nparity = 1;
- }
+ /*
+ * Update the uberblock that will be written when this txg completes.
+ */
+ RAIDZ_REFLOW_SET(&spa->spa_uberblock,
+ RRSS_SCRATCH_INVALID_SYNCED_REFLOW, new_offset);
+ vre->vre_offset_pertxg[txgoff] = 0;
+ zfs_rangelock_exit(lr);
+
+ mutex_enter(&vre->vre_lock);
+ vre->vre_bytes_copied += vre->vre_bytes_copied_pertxg[txgoff];
+ vre->vre_bytes_copied_pertxg[txgoff] = 0;
+ mutex_exit(&vre->vre_lock);
+
+ vdev_t *vd = vdev_lookup_top(spa, vre->vre_vdev_id);
+ VERIFY0(zap_update(spa->spa_meta_objset,
+ vd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_BYTES_COPIED,
+ sizeof (vre->vre_bytes_copied), 1, &vre->vre_bytes_copied, tx));
+}
- vdrz = kmem_zalloc(sizeof (*vdrz), KM_SLEEP);
- vdrz->vd_logical_width = children;
- vdrz->vd_nparity = nparity;
+static void
+raidz_reflow_complete_sync(void *arg, dmu_tx_t *tx)
+{
+ spa_t *spa = arg;
+ vdev_raidz_expand_t *vre = spa->spa_raidz_expand;
+ vdev_t *raidvd = vdev_lookup_top(spa, vre->vre_vdev_id);
+ vdev_raidz_t *vdrz = raidvd->vdev_tsd;
- *tsd = vdrz;
+ for (int i = 0; i < TXG_SIZE; i++)
+ VERIFY0(vre->vre_offset_pertxg[i]);
- return (0);
+ reflow_node_t *re = kmem_zalloc(sizeof (*re), KM_SLEEP);
+ re->re_txg = tx->tx_txg + TXG_CONCURRENT_STATES;
+ re->re_logical_width = vdrz->vd_physical_width;
+ mutex_enter(&vdrz->vd_expand_lock);
+ avl_add(&vdrz->vd_expand_txgs, re);
+ mutex_exit(&vdrz->vd_expand_lock);
+
+ vdev_t *vd = vdev_lookup_top(spa, vre->vre_vdev_id);
+
+ /*
+ * Dirty the config so that the updated ZPOOL_CONFIG_RAIDZ_EXPAND_TXGS
+ * will get written (based on vd_expand_txgs).
+ */
+ vdev_config_dirty(vd);
+
+ /*
+ * Before we change vre_state, the on-disk state must reflect that we
+ * have completed all copying, so that vdev_raidz_io_start() can use
+ * vre_state to determine if the reflow is in progress. See also the
+ * end of spa_raidz_expand_thread().
+ */
+ VERIFY3U(RRSS_GET_OFFSET(&spa->spa_ubsync), ==,
+ raidvd->vdev_ms_count << raidvd->vdev_ms_shift);
+
+ vre->vre_end_time = gethrestime_sec();
+ vre->vre_state = DSS_FINISHED;
+
+ uint64_t state = vre->vre_state;
+ VERIFY0(zap_update(spa->spa_meta_objset,
+ vd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_STATE,
+ sizeof (state), 1, &state, tx));
+
+ uint64_t end_time = vre->vre_end_time;
+ VERIFY0(zap_update(spa->spa_meta_objset,
+ vd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_END_TIME,
+ sizeof (end_time), 1, &end_time, tx));
+
+ spa->spa_uberblock.ub_raidz_reflow_info = 0;
+
+ spa_history_log_internal(spa, "raidz vdev expansion completed", tx,
+ "%s vdev %llu new width %llu", spa_name(spa),
+ (unsigned long long)vd->vdev_id,
+ (unsigned long long)vd->vdev_children);
+
+ spa->spa_raidz_expand = NULL;
+ raidvd->vdev_rz_expanding = B_FALSE;
+
+ spa_async_request(spa, SPA_ASYNC_INITIALIZE_RESTART);
+ spa_async_request(spa, SPA_ASYNC_TRIM_RESTART);
+ spa_async_request(spa, SPA_ASYNC_AUTOTRIM_RESTART);
+
+ spa_notify_waiters(spa);
+
+ /*
+ * While we're in syncing context take the opportunity to
+ * setup a scrub. All the data has been sucessfully copied
+ * but we have not validated any checksums.
+ */
+ pool_scan_func_t func = POOL_SCAN_SCRUB;
+ if (zfs_scrub_after_expand && dsl_scan_setup_check(&func, tx) == 0)
+ dsl_scan_setup_sync(&func, tx);
}
+/*
+ * Struct for one copy zio.
+ */
+typedef struct raidz_reflow_arg {
+ vdev_raidz_expand_t *rra_vre;
+ zfs_locked_range_t *rra_lr;
+ uint64_t rra_txg;
+} raidz_reflow_arg_t;
+
+/*
+ * The write of the new location is done.
+ */
static void
-vdev_raidz_fini(vdev_t *vd)
+raidz_reflow_write_done(zio_t *zio)
{
- kmem_free(vd->vdev_tsd, sizeof (vdev_raidz_t));
+ raidz_reflow_arg_t *rra = zio->io_private;
+ vdev_raidz_expand_t *vre = rra->rra_vre;
+
+ abd_free(zio->io_abd);
+
+ mutex_enter(&vre->vre_lock);
+ if (zio->io_error != 0) {
+ /* Force a reflow pause on errors */
+ vre->vre_failed_offset =
+ MIN(vre->vre_failed_offset, rra->rra_lr->lr_offset);
+ }
+ ASSERT3U(vre->vre_outstanding_bytes, >=, zio->io_size);
+ vre->vre_outstanding_bytes -= zio->io_size;
+ if (rra->rra_lr->lr_offset + rra->rra_lr->lr_length <
+ vre->vre_failed_offset) {
+ vre->vre_bytes_copied_pertxg[rra->rra_txg & TXG_MASK] +=
+ zio->io_size;
+ }
+ cv_signal(&vre->vre_cv);
+ mutex_exit(&vre->vre_lock);
+
+ zfs_rangelock_exit(rra->rra_lr);
+
+ kmem_free(rra, sizeof (*rra));
+ spa_config_exit(zio->io_spa, SCL_STATE, zio->io_spa);
}
/*
- * Add RAIDZ specific fields to the config nvlist.
+ * The read of the old location is done. The parent zio is the write to
+ * the new location. Allow it to start.
*/
static void
-vdev_raidz_config_generate(vdev_t *vd, nvlist_t *nv)
+raidz_reflow_read_done(zio_t *zio)
{
- ASSERT3P(vd->vdev_ops, ==, &vdev_raidz_ops);
- vdev_raidz_t *vdrz = vd->vdev_tsd;
+ raidz_reflow_arg_t *rra = zio->io_private;
+ vdev_raidz_expand_t *vre = rra->rra_vre;
/*
- * Make sure someone hasn't managed to sneak a fancy new vdev
- * into a crufty old storage pool.
+ * If the read failed, or if it was done on a vdev that is not fully
+ * healthy (e.g. a child that has a resilver in progress), we may not
+ * have the correct data. Note that it's OK if the write proceeds.
+ * It may write garbage but the location is otherwise unused and we
+ * will retry later due to vre_failed_offset.
*/
- ASSERT(vdrz->vd_nparity == 1 ||
- (vdrz->vd_nparity <= 2 &&
- spa_version(vd->vdev_spa) >= SPA_VERSION_RAIDZ2) ||
- (vdrz->vd_nparity <= 3 &&
- spa_version(vd->vdev_spa) >= SPA_VERSION_RAIDZ3));
+ if (zio->io_error != 0 || !vdev_dtl_empty(zio->io_vd, DTL_MISSING)) {
+ zfs_dbgmsg("reflow read failed off=%llu size=%llu txg=%llu "
+ "err=%u partial_dtl_empty=%u missing_dtl_empty=%u",
+ (long long)rra->rra_lr->lr_offset,
+ (long long)rra->rra_lr->lr_length,
+ (long long)rra->rra_txg,
+ zio->io_error,
+ vdev_dtl_empty(zio->io_vd, DTL_PARTIAL),
+ vdev_dtl_empty(zio->io_vd, DTL_MISSING));
+ mutex_enter(&vre->vre_lock);
+ /* Force a reflow pause on errors */
+ vre->vre_failed_offset =
+ MIN(vre->vre_failed_offset, rra->rra_lr->lr_offset);
+ mutex_exit(&vre->vre_lock);
+ }
- /*
- * Note that we'll add these even on storage pools where they
- * aren't strictly required -- older software will just ignore
- * it.
- */
- fnvlist_add_uint64(nv, ZPOOL_CONFIG_NPARITY, vdrz->vd_nparity);
+ zio_nowait(zio_unique_parent(zio));
}
-static uint64_t
-vdev_raidz_nparity(vdev_t *vd)
+static void
+raidz_reflow_record_progress(vdev_raidz_expand_t *vre, uint64_t offset,
+ dmu_tx_t *tx)
{
- vdev_raidz_t *vdrz = vd->vdev_tsd;
- return (vdrz->vd_nparity);
-}
+ int txgoff = dmu_tx_get_txg(tx) & TXG_MASK;
+ spa_t *spa = dmu_tx_pool(tx)->dp_spa;
-static uint64_t
-vdev_raidz_ndisks(vdev_t *vd)
-{
- return (vd->vdev_children);
+ if (offset == 0)
+ return;
+
+ mutex_enter(&vre->vre_lock);
+ ASSERT3U(vre->vre_offset, <=, offset);
+ vre->vre_offset = offset;
+ mutex_exit(&vre->vre_lock);
+
+ if (vre->vre_offset_pertxg[txgoff] == 0) {
+ dsl_sync_task_nowait(dmu_tx_pool(tx), raidz_reflow_sync,
+ spa, tx);
+ }
+ vre->vre_offset_pertxg[txgoff] = offset;
}
-vdev_ops_t vdev_raidz_ops = {
+static boolean_t
+vdev_raidz_expand_child_replacing(vdev_t *raidz_vd)
+{
+ for (int i = 0; i < raidz_vd->vdev_children; i++) {
+ /* Quick check if a child is being replaced */
+ if (!raidz_vd->vdev_child[i]->vdev_ops->vdev_op_leaf)
+ return (B_TRUE);
+ }
+ return (B_FALSE);
+}
+
+static boolean_t
+raidz_reflow_impl(vdev_t *vd, vdev_raidz_expand_t *vre, range_tree_t *rt,
+ dmu_tx_t *tx)
+{
+ spa_t *spa = vd->vdev_spa;
+ int ashift = vd->vdev_top->vdev_ashift;
+ uint64_t offset, size;
+
+ if (!range_tree_find_in(rt, 0, vd->vdev_top->vdev_asize,
+ &offset, &size)) {
+ return (B_FALSE);
+ }
+ ASSERT(IS_P2ALIGNED(offset, 1 << ashift));
+ ASSERT3U(size, >=, 1 << ashift);
+ uint64_t length = 1 << ashift;
+ int txgoff = dmu_tx_get_txg(tx) & TXG_MASK;
+
+ uint64_t blkid = offset >> ashift;
+
+ int old_children = vd->vdev_children - 1;
+
+ /*
+ * We can only progress to the point that writes will not overlap
+ * with blocks whose progress has not yet been recorded on disk.
+ * Since partially-copied rows are still read from the old location,
+ * we need to stop one row before the sector-wise overlap, to prevent
+ * row-wise overlap.
+ *
+ * Note that even if we are skipping over a large unallocated region,
+ * we can't move the on-disk progress to `offset`, because concurrent
+ * writes/allocations could still use the currently-unallocated
+ * region.
+ */
+ uint64_t ubsync_blkid =
+ RRSS_GET_OFFSET(&spa->spa_ubsync) >> ashift;
+ uint64_t next_overwrite_blkid = ubsync_blkid +
+ ubsync_blkid / old_children - old_children;
+ VERIFY3U(next_overwrite_blkid, >, ubsync_blkid);
+
+ if (blkid >= next_overwrite_blkid) {
+ raidz_reflow_record_progress(vre,
+ next_overwrite_blkid << ashift, tx);
+ return (B_TRUE);
+ }
+
+ range_tree_remove(rt, offset, length);
+
+ raidz_reflow_arg_t *rra = kmem_zalloc(sizeof (*rra), KM_SLEEP);
+ rra->rra_vre = vre;
+ rra->rra_lr = zfs_rangelock_enter(&vre->vre_rangelock,
+ offset, length, RL_WRITER);
+ rra->rra_txg = dmu_tx_get_txg(tx);
+
+ raidz_reflow_record_progress(vre, offset + length, tx);
+
+ mutex_enter(&vre->vre_lock);
+ vre->vre_outstanding_bytes += length;
+ mutex_exit(&vre->vre_lock);
+
+ /*
+ * SCL_STATE will be released when the read and write are done,
+ * by raidz_reflow_write_done().
+ */
+ spa_config_enter(spa, SCL_STATE, spa, RW_READER);
+
+ /* check if a replacing vdev was added, if so treat it as an error */
+ if (vdev_raidz_expand_child_replacing(vd)) {
+ zfs_dbgmsg("replacing vdev encountered, reflow paused at "
+ "offset=%llu txg=%llu",
+ (long long)rra->rra_lr->lr_offset,
+ (long long)rra->rra_txg);
+
+ mutex_enter(&vre->vre_lock);
+ vre->vre_failed_offset =
+ MIN(vre->vre_failed_offset, rra->rra_lr->lr_offset);
+ cv_signal(&vre->vre_cv);
+ mutex_exit(&vre->vre_lock);
+
+ /* drop everything we acquired */
+ zfs_rangelock_exit(rra->rra_lr);
+ kmem_free(rra, sizeof (*rra));
+ spa_config_exit(spa, SCL_STATE, spa);
+ return (B_TRUE);
+ }
+
+ zio_t *pio = spa->spa_txg_zio[txgoff];
+ abd_t *abd = abd_alloc_for_io(length, B_FALSE);
+ zio_t *write_zio = zio_vdev_child_io(pio, NULL,
+ vd->vdev_child[blkid % vd->vdev_children],
+ (blkid / vd->vdev_children) << ashift,
+ abd, length,
+ ZIO_TYPE_WRITE, ZIO_PRIORITY_REMOVAL,
+ ZIO_FLAG_CANFAIL,
+ raidz_reflow_write_done, rra);
+
+ zio_nowait(zio_vdev_child_io(write_zio, NULL,
+ vd->vdev_child[blkid % old_children],
+ (blkid / old_children) << ashift,
+ abd, length,
+ ZIO_TYPE_READ, ZIO_PRIORITY_REMOVAL,
+ ZIO_FLAG_CANFAIL,
+ raidz_reflow_read_done, rra));
+
+ return (B_FALSE);
+}
+
+/*
+ * For testing (ztest specific)
+ */
+static void
+raidz_expand_pause(uint_t pause_point)
+{
+ while (raidz_expand_pause_point != 0 &&
+ raidz_expand_pause_point <= pause_point)
+ delay(hz);
+}
+
+static void
+raidz_scratch_child_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);
+}
+
+/*
+ * Reflow the beginning portion of the vdev into an intermediate scratch area
+ * in memory and on disk. This operation must be persisted on disk before we
+ * proceed to overwrite the beginning portion with the reflowed data.
+ *
+ * This multi-step task can fail to complete if disk errors are encountered
+ * and we can return here after a pause (waiting for disk to become healthy).
+ */
+static void
+raidz_reflow_scratch_sync(void *arg, dmu_tx_t *tx)
+{
+ vdev_raidz_expand_t *vre = arg;
+ spa_t *spa = dmu_tx_pool(tx)->dp_spa;
+ zio_t *pio;
+ int error;
+
+ spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
+ vdev_t *raidvd = vdev_lookup_top(spa, vre->vre_vdev_id);
+ int ashift = raidvd->vdev_ashift;
+ uint64_t write_size = P2ALIGN(VDEV_BOOT_SIZE, 1 << ashift);
+ uint64_t logical_size = write_size * raidvd->vdev_children;
+ uint64_t read_size =
+ P2ROUNDUP(DIV_ROUND_UP(logical_size, (raidvd->vdev_children - 1)),
+ 1 << ashift);
+
+ /*
+ * The scratch space must be large enough to get us to the point
+ * that one row does not overlap itself when moved. This is checked
+ * by vdev_raidz_attach_check().
+ */
+ VERIFY3U(write_size, >=, raidvd->vdev_children << ashift);
+ VERIFY3U(write_size, <=, VDEV_BOOT_SIZE);
+ VERIFY3U(write_size, <=, read_size);
+
+ zfs_locked_range_t *lr = zfs_rangelock_enter(&vre->vre_rangelock,
+ 0, logical_size, RL_WRITER);
+
+ abd_t **abds = kmem_alloc(raidvd->vdev_children * sizeof (abd_t *),
+ KM_SLEEP);
+ for (int i = 0; i < raidvd->vdev_children; i++) {
+ abds[i] = abd_alloc_linear(read_size, B_FALSE);
+ }
+
+ raidz_expand_pause(RAIDZ_EXPAND_PAUSE_PRE_SCRATCH_1);
+
+ /*
+ * If we have already written the scratch area then we must read from
+ * there, since new writes were redirected there while we were paused
+ * or the original location may have been partially overwritten with
+ * reflowed data.
+ */
+ if (RRSS_GET_STATE(&spa->spa_ubsync) == RRSS_SCRATCH_VALID) {
+ VERIFY3U(RRSS_GET_OFFSET(&spa->spa_ubsync), ==, logical_size);
+ /*
+ * Read from scratch space.
+ */
+ pio = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL);
+ for (int i = 0; i < raidvd->vdev_children; i++) {
+ /*
+ * Note: zio_vdev_child_io() adds VDEV_LABEL_START_SIZE
+ * to the offset to calculate the physical offset to
+ * write to. Passing in a negative offset makes us
+ * access the scratch area.
+ */
+ zio_nowait(zio_vdev_child_io(pio, NULL,
+ raidvd->vdev_child[i],
+ VDEV_BOOT_OFFSET - VDEV_LABEL_START_SIZE, abds[i],
+ write_size, ZIO_TYPE_READ, ZIO_PRIORITY_ASYNC_READ,
+ ZIO_FLAG_CANFAIL, raidz_scratch_child_done, pio));
+ }
+ error = zio_wait(pio);
+ if (error != 0) {
+ zfs_dbgmsg("reflow: error %d reading scratch location",
+ error);
+ goto io_error_exit;
+ }
+ goto overwrite;
+ }
+
+ /*
+ * Read from original location.
+ */
+ pio = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL);
+ for (int i = 0; i < raidvd->vdev_children - 1; i++) {
+ ASSERT0(vdev_is_dead(raidvd->vdev_child[i]));
+ zio_nowait(zio_vdev_child_io(pio, NULL, raidvd->vdev_child[i],
+ 0, abds[i], read_size, ZIO_TYPE_READ,
+ ZIO_PRIORITY_ASYNC_READ, ZIO_FLAG_CANFAIL,
+ raidz_scratch_child_done, pio));
+ }
+ error = zio_wait(pio);
+ if (error != 0) {
+ zfs_dbgmsg("reflow: error %d reading original location", error);
+io_error_exit:
+ for (int i = 0; i < raidvd->vdev_children; i++)
+ abd_free(abds[i]);
+ kmem_free(abds, raidvd->vdev_children * sizeof (abd_t *));
+ zfs_rangelock_exit(lr);
+ spa_config_exit(spa, SCL_STATE, FTAG);
+ return;
+ }
+
+ raidz_expand_pause(RAIDZ_EXPAND_PAUSE_PRE_SCRATCH_2);
+
+ /*
+ * Reflow in memory.
+ */
+ uint64_t logical_sectors = logical_size >> ashift;
+ for (int i = raidvd->vdev_children - 1; i < logical_sectors; i++) {
+ int oldchild = i % (raidvd->vdev_children - 1);
+ uint64_t oldoff = (i / (raidvd->vdev_children - 1)) << ashift;
+
+ int newchild = i % raidvd->vdev_children;
+ uint64_t newoff = (i / raidvd->vdev_children) << ashift;
+
+ /* a single sector should not be copying over itself */
+ ASSERT(!(newchild == oldchild && newoff == oldoff));
+
+ abd_copy_off(abds[newchild], abds[oldchild],
+ newoff, oldoff, 1 << ashift);
+ }
+
+ /*
+ * Verify that we filled in everything we intended to (write_size on
+ * each child).
+ */
+ VERIFY0(logical_sectors % raidvd->vdev_children);
+ VERIFY3U((logical_sectors / raidvd->vdev_children) << ashift, ==,
+ write_size);
+
+ /*
+ * Write to scratch location (boot area).
+ */
+ pio = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL);
+ for (int i = 0; i < raidvd->vdev_children; i++) {
+ /*
+ * Note: zio_vdev_child_io() adds VDEV_LABEL_START_SIZE to
+ * the offset to calculate the physical offset to write to.
+ * Passing in a negative offset lets us access the boot area.
+ */
+ zio_nowait(zio_vdev_child_io(pio, NULL, raidvd->vdev_child[i],
+ VDEV_BOOT_OFFSET - VDEV_LABEL_START_SIZE, abds[i],
+ write_size, ZIO_TYPE_WRITE, ZIO_PRIORITY_ASYNC_WRITE,
+ ZIO_FLAG_CANFAIL, raidz_scratch_child_done, pio));
+ }
+ error = zio_wait(pio);
+ if (error != 0) {
+ zfs_dbgmsg("reflow: error %d writing scratch location", error);
+ goto io_error_exit;
+ }
+ pio = zio_root(spa, NULL, NULL, 0);
+ zio_flush(pio, raidvd);
+ zio_wait(pio);
+
+ zfs_dbgmsg("reflow: wrote %llu bytes (logical) to scratch area",
+ (long long)logical_size);
+
+ raidz_expand_pause(RAIDZ_EXPAND_PAUSE_PRE_SCRATCH_3);
+
+ /*
+ * Update uberblock to indicate that scratch space is valid. This is
+ * needed because after this point, the real location may be
+ * overwritten. If we crash, we need to get the data from the
+ * scratch space, rather than the real location.
+ *
+ * Note: ub_timestamp is bumped so that vdev_uberblock_compare()
+ * will prefer this uberblock.
+ */
+ RAIDZ_REFLOW_SET(&spa->spa_ubsync, RRSS_SCRATCH_VALID, logical_size);
+ spa->spa_ubsync.ub_timestamp++;
+ ASSERT0(vdev_uberblock_sync_list(&spa->spa_root_vdev, 1,
+ &spa->spa_ubsync, ZIO_FLAG_CONFIG_WRITER));
+ if (spa_multihost(spa))
+ mmp_update_uberblock(spa, &spa->spa_ubsync);
+
+ zfs_dbgmsg("reflow: uberblock updated "
+ "(txg %llu, SCRATCH_VALID, size %llu, ts %llu)",
+ (long long)spa->spa_ubsync.ub_txg,
+ (long long)logical_size,
+ (long long)spa->spa_ubsync.ub_timestamp);
+
+ raidz_expand_pause(RAIDZ_EXPAND_PAUSE_SCRATCH_VALID);
+
+ /*
+ * Overwrite with reflow'ed data.
+ */
+overwrite:
+ pio = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL);
+ for (int i = 0; i < raidvd->vdev_children; i++) {
+ zio_nowait(zio_vdev_child_io(pio, NULL, raidvd->vdev_child[i],
+ 0, abds[i], write_size, ZIO_TYPE_WRITE,
+ ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_CANFAIL,
+ raidz_scratch_child_done, pio));
+ }
+ error = zio_wait(pio);
+ if (error != 0) {
+ /*
+ * When we exit early here and drop the range lock, new
+ * writes will go into the scratch area so we'll need to
+ * read from there when we return after pausing.
+ */
+ zfs_dbgmsg("reflow: error %d writing real location", error);
+ /*
+ * Update the uberblock that is written when this txg completes.
+ */
+ RAIDZ_REFLOW_SET(&spa->spa_uberblock, RRSS_SCRATCH_VALID,
+ logical_size);
+ goto io_error_exit;
+ }
+ pio = zio_root(spa, NULL, NULL, 0);
+ zio_flush(pio, raidvd);
+ zio_wait(pio);
+
+ zfs_dbgmsg("reflow: overwrote %llu bytes (logical) to real location",
+ (long long)logical_size);
+ for (int i = 0; i < raidvd->vdev_children; i++)
+ abd_free(abds[i]);
+ kmem_free(abds, raidvd->vdev_children * sizeof (abd_t *));
+
+ raidz_expand_pause(RAIDZ_EXPAND_PAUSE_SCRATCH_REFLOWED);
+
+ /*
+ * Update uberblock to indicate that the initial part has been
+ * reflow'ed. This is needed because after this point (when we exit
+ * the rangelock), we allow regular writes to this region, which will
+ * be written to the new location only (because reflow_offset_next ==
+ * reflow_offset_synced). If we crashed and re-copied from the
+ * scratch space, we would lose the regular writes.
+ */
+ RAIDZ_REFLOW_SET(&spa->spa_ubsync, RRSS_SCRATCH_INVALID_SYNCED,
+ logical_size);
+ spa->spa_ubsync.ub_timestamp++;
+ ASSERT0(vdev_uberblock_sync_list(&spa->spa_root_vdev, 1,
+ &spa->spa_ubsync, ZIO_FLAG_CONFIG_WRITER));
+ if (spa_multihost(spa))
+ mmp_update_uberblock(spa, &spa->spa_ubsync);
+
+ zfs_dbgmsg("reflow: uberblock updated "
+ "(txg %llu, SCRATCH_NOT_IN_USE, size %llu, ts %llu)",
+ (long long)spa->spa_ubsync.ub_txg,
+ (long long)logical_size,
+ (long long)spa->spa_ubsync.ub_timestamp);
+
+ raidz_expand_pause(RAIDZ_EXPAND_PAUSE_SCRATCH_POST_REFLOW_1);
+
+ /*
+ * Update progress.
+ */
+ vre->vre_offset = logical_size;
+ zfs_rangelock_exit(lr);
+ spa_config_exit(spa, SCL_STATE, FTAG);
+
+ int txgoff = dmu_tx_get_txg(tx) & TXG_MASK;
+ vre->vre_offset_pertxg[txgoff] = vre->vre_offset;
+ vre->vre_bytes_copied_pertxg[txgoff] = vre->vre_bytes_copied;
+ /*
+ * Note - raidz_reflow_sync() will update the uberblock state to
+ * RRSS_SCRATCH_INVALID_SYNCED_REFLOW
+ */
+ raidz_reflow_sync(spa, tx);
+
+ raidz_expand_pause(RAIDZ_EXPAND_PAUSE_SCRATCH_POST_REFLOW_2);
+}
+
+/*
+ * We crashed in the middle of raidz_reflow_scratch_sync(); complete its work
+ * here. No other i/o can be in progress, so we don't need the vre_rangelock.
+ */
+void
+vdev_raidz_reflow_copy_scratch(spa_t *spa)
+{
+ vdev_raidz_expand_t *vre = spa->spa_raidz_expand;
+ uint64_t logical_size = RRSS_GET_OFFSET(&spa->spa_uberblock);
+ ASSERT3U(RRSS_GET_STATE(&spa->spa_uberblock), ==, RRSS_SCRATCH_VALID);
+
+ spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
+ vdev_t *raidvd = vdev_lookup_top(spa, vre->vre_vdev_id);
+ ASSERT0(logical_size % raidvd->vdev_children);
+ uint64_t write_size = logical_size / raidvd->vdev_children;
+
+ zio_t *pio;
+
+ /*
+ * Read from scratch space.
+ */
+ abd_t **abds = kmem_alloc(raidvd->vdev_children * sizeof (abd_t *),
+ KM_SLEEP);
+ for (int i = 0; i < raidvd->vdev_children; i++) {
+ abds[i] = abd_alloc_linear(write_size, B_FALSE);
+ }
+
+ pio = zio_root(spa, NULL, NULL, 0);
+ for (int i = 0; i < raidvd->vdev_children; i++) {
+ /*
+ * Note: zio_vdev_child_io() adds VDEV_LABEL_START_SIZE to
+ * the offset to calculate the physical offset to write to.
+ * Passing in a negative offset lets us access the boot area.
+ */
+ zio_nowait(zio_vdev_child_io(pio, NULL, raidvd->vdev_child[i],
+ VDEV_BOOT_OFFSET - VDEV_LABEL_START_SIZE, abds[i],
+ write_size, ZIO_TYPE_READ,
+ ZIO_PRIORITY_ASYNC_READ, 0,
+ raidz_scratch_child_done, pio));
+ }
+ zio_wait(pio);
+
+ /*
+ * Overwrite real location with reflow'ed data.
+ */
+ pio = zio_root(spa, NULL, NULL, 0);
+ for (int i = 0; i < raidvd->vdev_children; i++) {
+ zio_nowait(zio_vdev_child_io(pio, NULL, raidvd->vdev_child[i],
+ 0, abds[i], write_size, ZIO_TYPE_WRITE,
+ ZIO_PRIORITY_ASYNC_WRITE, 0,
+ raidz_scratch_child_done, pio));
+ }
+ zio_wait(pio);
+ pio = zio_root(spa, NULL, NULL, 0);
+ zio_flush(pio, raidvd);
+ zio_wait(pio);
+
+ zfs_dbgmsg("reflow recovery: overwrote %llu bytes (logical) "
+ "to real location", (long long)logical_size);
+
+ for (int i = 0; i < raidvd->vdev_children; i++)
+ abd_free(abds[i]);
+ kmem_free(abds, raidvd->vdev_children * sizeof (abd_t *));
+
+ /*
+ * Update uberblock.
+ */
+ RAIDZ_REFLOW_SET(&spa->spa_ubsync,
+ RRSS_SCRATCH_INVALID_SYNCED_ON_IMPORT, logical_size);
+ spa->spa_ubsync.ub_timestamp++;
+ VERIFY0(vdev_uberblock_sync_list(&spa->spa_root_vdev, 1,
+ &spa->spa_ubsync, ZIO_FLAG_CONFIG_WRITER));
+ if (spa_multihost(spa))
+ mmp_update_uberblock(spa, &spa->spa_ubsync);
+
+ zfs_dbgmsg("reflow recovery: uberblock updated "
+ "(txg %llu, SCRATCH_NOT_IN_USE, size %llu, ts %llu)",
+ (long long)spa->spa_ubsync.ub_txg,
+ (long long)logical_size,
+ (long long)spa->spa_ubsync.ub_timestamp);
+
+ dmu_tx_t *tx = dmu_tx_create_assigned(spa->spa_dsl_pool,
+ spa_first_txg(spa));
+ int txgoff = dmu_tx_get_txg(tx) & TXG_MASK;
+ vre->vre_offset = logical_size;
+ vre->vre_offset_pertxg[txgoff] = vre->vre_offset;
+ vre->vre_bytes_copied_pertxg[txgoff] = vre->vre_bytes_copied;
+ /*
+ * Note that raidz_reflow_sync() will update the uberblock once more
+ */
+ raidz_reflow_sync(spa, tx);
+
+ dmu_tx_commit(tx);
+
+ spa_config_exit(spa, SCL_STATE, FTAG);
+}
+
+static boolean_t
+spa_raidz_expand_thread_check(void *arg, zthr_t *zthr)
+{
+ (void) zthr;
+ spa_t *spa = arg;
+
+ return (spa->spa_raidz_expand != NULL &&
+ !spa->spa_raidz_expand->vre_waiting_for_resilver);
+}
+
+/*
+ * RAIDZ expansion background thread
+ *
+ * Can be called multiple times if the reflow is paused
+ */
+static void
+spa_raidz_expand_thread(void *arg, zthr_t *zthr)
+{
+ spa_t *spa = arg;
+ vdev_raidz_expand_t *vre = spa->spa_raidz_expand;
+
+ if (RRSS_GET_STATE(&spa->spa_ubsync) == RRSS_SCRATCH_VALID)
+ vre->vre_offset = 0;
+ else
+ vre->vre_offset = RRSS_GET_OFFSET(&spa->spa_ubsync);
+
+ /* Reflow the begining portion using the scratch area */
+ if (vre->vre_offset == 0) {
+ VERIFY0(dsl_sync_task(spa_name(spa),
+ NULL, raidz_reflow_scratch_sync,
+ vre, 0, ZFS_SPACE_CHECK_NONE));
+
+ /* if we encountered errors then pause */
+ if (vre->vre_offset == 0) {
+ mutex_enter(&vre->vre_lock);
+ vre->vre_waiting_for_resilver = B_TRUE;
+ mutex_exit(&vre->vre_lock);
+ return;
+ }
+ }
+
+ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
+ vdev_t *raidvd = vdev_lookup_top(spa, vre->vre_vdev_id);
+
+ uint64_t guid = raidvd->vdev_guid;
+
+ /* Iterate over all the remaining metaslabs */
+ for (uint64_t i = vre->vre_offset >> raidvd->vdev_ms_shift;
+ i < raidvd->vdev_ms_count &&
+ !zthr_iscancelled(zthr) &&
+ vre->vre_failed_offset == UINT64_MAX; i++) {
+ metaslab_t *msp = raidvd->vdev_ms[i];
+
+ metaslab_disable(msp);
+ mutex_enter(&msp->ms_lock);
+
+ /*
+ * The metaslab may be newly created (for the expanded
+ * space), in which case its trees won't exist yet,
+ * so we need to bail out early.
+ */
+ if (msp->ms_new) {
+ mutex_exit(&msp->ms_lock);
+ metaslab_enable(msp, B_FALSE, B_FALSE);
+ continue;
+ }
+
+ VERIFY0(metaslab_load(msp));
+
+ /*
+ * We want to copy everything except the free (allocatable)
+ * space. Note that there may be a little bit more free
+ * space (e.g. in ms_defer), and it's fine to copy that too.
+ */
+ range_tree_t *rt = range_tree_create(NULL, RANGE_SEG64,
+ NULL, 0, 0);
+ range_tree_add(rt, msp->ms_start, msp->ms_size);
+ range_tree_walk(msp->ms_allocatable, range_tree_remove, rt);
+ mutex_exit(&msp->ms_lock);
+
+ /*
+ * Force the last sector of each metaslab to be copied. This
+ * ensures that we advance the on-disk progress to the end of
+ * this metaslab while the metaslab is disabled. Otherwise, we
+ * could move past this metaslab without advancing the on-disk
+ * progress, and then an allocation to this metaslab would not
+ * be copied.
+ */
+ int sectorsz = 1 << raidvd->vdev_ashift;
+ uint64_t ms_last_offset = msp->ms_start +
+ msp->ms_size - sectorsz;
+ if (!range_tree_contains(rt, ms_last_offset, sectorsz)) {
+ range_tree_add(rt, ms_last_offset, sectorsz);
+ }
+
+ /*
+ * When we are resuming from a paused expansion (i.e.
+ * when importing a pool with a expansion in progress),
+ * discard any state that we have already processed.
+ */
+ range_tree_clear(rt, 0, vre->vre_offset);
+
+ while (!zthr_iscancelled(zthr) &&
+ !range_tree_is_empty(rt) &&
+ vre->vre_failed_offset == UINT64_MAX) {
+
+ /*
+ * 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);
+
+ /*
+ * If requested, pause the reflow when the amount
+ * specified by raidz_expand_max_reflow_bytes is reached
+ *
+ * This pause is only used during testing or debugging.
+ */
+ while (raidz_expand_max_reflow_bytes != 0 &&
+ raidz_expand_max_reflow_bytes <=
+ vre->vre_bytes_copied && !zthr_iscancelled(zthr)) {
+ delay(hz);
+ }
+
+ mutex_enter(&vre->vre_lock);
+ while (vre->vre_outstanding_bytes >
+ raidz_expand_max_copy_bytes) {
+ cv_wait(&vre->vre_cv, &vre->vre_lock);
+ }
+ mutex_exit(&vre->vre_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. Theoretically, the
+ * vdev_t that we're expanding may have changed.
+ */
+ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
+ raidvd = vdev_lookup_top(spa, vre->vre_vdev_id);
+
+ boolean_t needsync =
+ raidz_reflow_impl(raidvd, vre, rt, tx);
+
+ dmu_tx_commit(tx);
+
+ if (needsync) {
+ spa_config_exit(spa, SCL_CONFIG, FTAG);
+ txg_wait_synced(spa->spa_dsl_pool, txg);
+ spa_config_enter(spa, SCL_CONFIG, FTAG,
+ RW_READER);
+ }
+ }
+
+ spa_config_exit(spa, SCL_CONFIG, FTAG);
+
+ metaslab_enable(msp, B_FALSE, B_FALSE);
+ range_tree_vacate(rt, NULL, NULL);
+ range_tree_destroy(rt);
+
+ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
+ raidvd = vdev_lookup_top(spa, vre->vre_vdev_id);
+ }
+
+ spa_config_exit(spa, SCL_CONFIG, FTAG);
+
+ /*
+ * The txg_wait_synced() here ensures that all reflow zio's have
+ * completed, and vre_failed_offset has been set if necessary. It
+ * also ensures that the progress of the last raidz_reflow_sync() is
+ * written to disk before raidz_reflow_complete_sync() changes the
+ * in-memory vre_state. vdev_raidz_io_start() uses vre_state to
+ * determine if a reflow is in progress, in which case we may need to
+ * write to both old and new locations. Therefore we can only change
+ * vre_state once this is not necessary, which is once the on-disk
+ * progress (in spa_ubsync) has been set past any possible writes (to
+ * the end of the last metaslab).
+ */
+ txg_wait_synced(spa->spa_dsl_pool, 0);
+
+ if (!zthr_iscancelled(zthr) &&
+ vre->vre_offset == raidvd->vdev_ms_count << raidvd->vdev_ms_shift) {
+ /*
+ * We are not being canceled or paused, so the reflow must be
+ * complete. In that case also mark it as completed on disk.
+ */
+ ASSERT3U(vre->vre_failed_offset, ==, UINT64_MAX);
+ VERIFY0(dsl_sync_task(spa_name(spa), NULL,
+ raidz_reflow_complete_sync, spa,
+ 0, ZFS_SPACE_CHECK_NONE));
+ (void) vdev_online(spa, guid, ZFS_ONLINE_EXPAND, NULL);
+ } else {
+ /*
+ * Wait for all copy zio's to complete and for all the
+ * raidz_reflow_sync() synctasks to be run.
+ */
+ spa_history_log_internal(spa, "reflow pause",
+ NULL, "offset=%llu failed_offset=%lld",
+ (long long)vre->vre_offset,
+ (long long)vre->vre_failed_offset);
+ mutex_enter(&vre->vre_lock);
+ if (vre->vre_failed_offset != UINT64_MAX) {
+ /*
+ * Reset progress so that we will retry everything
+ * after the point that something failed.
+ */
+ vre->vre_offset = vre->vre_failed_offset;
+ vre->vre_failed_offset = UINT64_MAX;
+ vre->vre_waiting_for_resilver = B_TRUE;
+ }
+ mutex_exit(&vre->vre_lock);
+ }
+}
+
+void
+spa_start_raidz_expansion_thread(spa_t *spa)
+{
+ ASSERT3P(spa->spa_raidz_expand_zthr, ==, NULL);
+ spa->spa_raidz_expand_zthr = zthr_create("raidz_expand",
+ spa_raidz_expand_thread_check, spa_raidz_expand_thread,
+ spa, defclsyspri);
+}
+
+void
+raidz_dtl_reassessed(vdev_t *vd)
+{
+ spa_t *spa = vd->vdev_spa;
+ if (spa->spa_raidz_expand != NULL) {
+ vdev_raidz_expand_t *vre = spa->spa_raidz_expand;
+ /*
+ * we get called often from vdev_dtl_reassess() so make
+ * sure it's our vdev and any replacing is complete
+ */
+ if (vd->vdev_top->vdev_id == vre->vre_vdev_id &&
+ !vdev_raidz_expand_child_replacing(vd->vdev_top)) {
+ mutex_enter(&vre->vre_lock);
+ if (vre->vre_waiting_for_resilver) {
+ vdev_dbgmsg(vd, "DTL reassessed, "
+ "continuing raidz expansion");
+ vre->vre_waiting_for_resilver = B_FALSE;
+ zthr_wakeup(spa->spa_raidz_expand_zthr);
+ }
+ mutex_exit(&vre->vre_lock);
+ }
+ }
+}
+
+int
+vdev_raidz_attach_check(vdev_t *new_child)
+{
+ vdev_t *raidvd = new_child->vdev_parent;
+ uint64_t new_children = raidvd->vdev_children;
+
+ /*
+ * We use the "boot" space as scratch space to handle overwriting the
+ * initial part of the vdev. If it is too small, then this expansion
+ * is not allowed. This would be very unusual (e.g. ashift > 13 and
+ * >200 children).
+ */
+ if (new_children << raidvd->vdev_ashift > VDEV_BOOT_SIZE) {
+ return (EINVAL);
+ }
+ return (0);
+}
+
+void
+vdev_raidz_attach_sync(void *arg, dmu_tx_t *tx)
+{
+ vdev_t *new_child = arg;
+ spa_t *spa = new_child->vdev_spa;
+ vdev_t *raidvd = new_child->vdev_parent;
+ vdev_raidz_t *vdrz = raidvd->vdev_tsd;
+ ASSERT3P(raidvd->vdev_ops, ==, &vdev_raidz_ops);
+ ASSERT3P(raidvd->vdev_top, ==, raidvd);
+ ASSERT3U(raidvd->vdev_children, >, vdrz->vd_original_width);
+ ASSERT3U(raidvd->vdev_children, ==, vdrz->vd_physical_width + 1);
+ ASSERT3P(raidvd->vdev_child[raidvd->vdev_children - 1], ==,
+ new_child);
+
+ spa_feature_incr(spa, SPA_FEATURE_RAIDZ_EXPANSION, tx);
+
+ vdrz->vd_physical_width++;
+
+ VERIFY0(spa->spa_uberblock.ub_raidz_reflow_info);
+ vdrz->vn_vre.vre_vdev_id = raidvd->vdev_id;
+ vdrz->vn_vre.vre_offset = 0;
+ vdrz->vn_vre.vre_failed_offset = UINT64_MAX;
+ spa->spa_raidz_expand = &vdrz->vn_vre;
+ zthr_wakeup(spa->spa_raidz_expand_zthr);
+
+ /*
+ * Dirty the config so that ZPOOL_CONFIG_RAIDZ_EXPANDING will get
+ * written to the config.
+ */
+ vdev_config_dirty(raidvd);
+
+ vdrz->vn_vre.vre_start_time = gethrestime_sec();
+ vdrz->vn_vre.vre_end_time = 0;
+ vdrz->vn_vre.vre_state = DSS_SCANNING;
+ vdrz->vn_vre.vre_bytes_copied = 0;
+
+ uint64_t state = vdrz->vn_vre.vre_state;
+ VERIFY0(zap_update(spa->spa_meta_objset,
+ raidvd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_STATE,
+ sizeof (state), 1, &state, tx));
+
+ uint64_t start_time = vdrz->vn_vre.vre_start_time;
+ VERIFY0(zap_update(spa->spa_meta_objset,
+ raidvd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_START_TIME,
+ sizeof (start_time), 1, &start_time, tx));
+
+ (void) zap_remove(spa->spa_meta_objset,
+ raidvd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_END_TIME, tx);
+ (void) zap_remove(spa->spa_meta_objset,
+ raidvd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_BYTES_COPIED, tx);
+
+ spa_history_log_internal(spa, "raidz vdev expansion started", tx,
+ "%s vdev %llu new width %llu", spa_name(spa),
+ (unsigned long long)raidvd->vdev_id,
+ (unsigned long long)raidvd->vdev_children);
+}
+
+int
+vdev_raidz_load(vdev_t *vd)
+{
+ vdev_raidz_t *vdrz = vd->vdev_tsd;
+ int err;
+
+ uint64_t state = DSS_NONE;
+ uint64_t start_time = 0;
+ uint64_t end_time = 0;
+ uint64_t bytes_copied = 0;
+
+ if (vd->vdev_top_zap != 0) {
+ err = zap_lookup(vd->vdev_spa->spa_meta_objset,
+ vd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_STATE,
+ sizeof (state), 1, &state);
+ if (err != 0 && err != ENOENT)
+ return (err);
+
+ err = zap_lookup(vd->vdev_spa->spa_meta_objset,
+ vd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_START_TIME,
+ sizeof (start_time), 1, &start_time);
+ if (err != 0 && err != ENOENT)
+ return (err);
+
+ err = zap_lookup(vd->vdev_spa->spa_meta_objset,
+ vd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_END_TIME,
+ sizeof (end_time), 1, &end_time);
+ if (err != 0 && err != ENOENT)
+ return (err);
+
+ err = zap_lookup(vd->vdev_spa->spa_meta_objset,
+ vd->vdev_top_zap, VDEV_TOP_ZAP_RAIDZ_EXPAND_BYTES_COPIED,
+ sizeof (bytes_copied), 1, &bytes_copied);
+ if (err != 0 && err != ENOENT)
+ return (err);
+ }
+
+ /*
+ * If we are in the middle of expansion, vre_state should have
+ * already been set by vdev_raidz_init().
+ */
+ EQUIV(vdrz->vn_vre.vre_state == DSS_SCANNING, state == DSS_SCANNING);
+ vdrz->vn_vre.vre_state = (dsl_scan_state_t)state;
+ vdrz->vn_vre.vre_start_time = start_time;
+ vdrz->vn_vre.vre_end_time = end_time;
+ vdrz->vn_vre.vre_bytes_copied = bytes_copied;
+
+ return (0);
+}
+
+int
+spa_raidz_expand_get_stats(spa_t *spa, pool_raidz_expand_stat_t *pres)
+{
+ vdev_raidz_expand_t *vre = spa->spa_raidz_expand;
+
+ if (vre == NULL) {
+ /* no removal in progress; find most recent completed */
+ for (int c = 0; c < spa->spa_root_vdev->vdev_children; c++) {
+ vdev_t *vd = spa->spa_root_vdev->vdev_child[c];
+ if (vd->vdev_ops == &vdev_raidz_ops) {
+ vdev_raidz_t *vdrz = vd->vdev_tsd;
+
+ if (vdrz->vn_vre.vre_end_time != 0 &&
+ (vre == NULL ||
+ vdrz->vn_vre.vre_end_time >
+ vre->vre_end_time)) {
+ vre = &vdrz->vn_vre;
+ }
+ }
+ }
+ }
+
+ if (vre == NULL) {
+ return (SET_ERROR(ENOENT));
+ }
+
+ pres->pres_state = vre->vre_state;
+ pres->pres_expanding_vdev = vre->vre_vdev_id;
+
+ vdev_t *vd = vdev_lookup_top(spa, vre->vre_vdev_id);
+ pres->pres_to_reflow = vd->vdev_stat.vs_alloc;
+
+ mutex_enter(&vre->vre_lock);
+ pres->pres_reflowed = vre->vre_bytes_copied;
+ for (int i = 0; i < TXG_SIZE; i++)
+ pres->pres_reflowed += vre->vre_bytes_copied_pertxg[i];
+ mutex_exit(&vre->vre_lock);
+
+ pres->pres_start_time = vre->vre_start_time;
+ pres->pres_end_time = vre->vre_end_time;
+ pres->pres_waiting_for_resilver = vre->vre_waiting_for_resilver;
+
+ return (0);
+}
+
+/*
+ * Initialize private RAIDZ specific fields from the nvlist.
+ */
+static int
+vdev_raidz_init(spa_t *spa, nvlist_t *nv, void **tsd)
+{
+ uint_t children;
+ nvlist_t **child;
+ int error = nvlist_lookup_nvlist_array(nv,
+ ZPOOL_CONFIG_CHILDREN, &child, &children);
+ if (error != 0)
+ return (SET_ERROR(EINVAL));
+
+ uint64_t nparity;
+ 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(EINVAL));
+ else if (nparity > 2 && spa_version(spa) < SPA_VERSION_RAIDZ3)
+ return (SET_ERROR(EINVAL));
+ } 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;
+ }
+
+ vdev_raidz_t *vdrz = kmem_zalloc(sizeof (*vdrz), KM_SLEEP);
+ vdrz->vn_vre.vre_vdev_id = -1;
+ vdrz->vn_vre.vre_offset = UINT64_MAX;
+ vdrz->vn_vre.vre_failed_offset = UINT64_MAX;
+ mutex_init(&vdrz->vn_vre.vre_lock, NULL, MUTEX_DEFAULT, NULL);
+ cv_init(&vdrz->vn_vre.vre_cv, NULL, CV_DEFAULT, NULL);
+ zfs_rangelock_init(&vdrz->vn_vre.vre_rangelock, NULL, NULL);
+ mutex_init(&vdrz->vd_expand_lock, NULL, MUTEX_DEFAULT, NULL);
+ avl_create(&vdrz->vd_expand_txgs, vdev_raidz_reflow_compare,
+ sizeof (reflow_node_t), offsetof(reflow_node_t, re_link));
+
+ vdrz->vd_physical_width = children;
+ vdrz->vd_nparity = nparity;
+
+ /* note, the ID does not exist when creating a pool */
+ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID,
+ &vdrz->vn_vre.vre_vdev_id);
+
+ boolean_t reflow_in_progress =
+ nvlist_exists(nv, ZPOOL_CONFIG_RAIDZ_EXPANDING);
+ if (reflow_in_progress) {
+ spa->spa_raidz_expand = &vdrz->vn_vre;
+ vdrz->vn_vre.vre_state = DSS_SCANNING;
+ }
+
+ vdrz->vd_original_width = children;
+ uint64_t *txgs;
+ unsigned int txgs_size = 0;
+ error = nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_RAIDZ_EXPAND_TXGS,
+ &txgs, &txgs_size);
+ if (error == 0) {
+ for (int i = 0; i < txgs_size; i++) {
+ reflow_node_t *re = kmem_zalloc(sizeof (*re), KM_SLEEP);
+ re->re_txg = txgs[txgs_size - i - 1];
+ re->re_logical_width = vdrz->vd_physical_width - i;
+
+ if (reflow_in_progress)
+ re->re_logical_width--;
+
+ avl_add(&vdrz->vd_expand_txgs, re);
+ }
+
+ vdrz->vd_original_width = vdrz->vd_physical_width - txgs_size;
+ }
+ if (reflow_in_progress) {
+ vdrz->vd_original_width--;
+ zfs_dbgmsg("reflow_in_progress, %u wide, %d prior expansions",
+ children, txgs_size);
+ }
+
+ *tsd = vdrz;
+
+ return (0);
+}
+
+static void
+vdev_raidz_fini(vdev_t *vd)
+{
+ vdev_raidz_t *vdrz = vd->vdev_tsd;
+ if (vd->vdev_spa->spa_raidz_expand == &vdrz->vn_vre)
+ vd->vdev_spa->spa_raidz_expand = NULL;
+ reflow_node_t *re;
+ void *cookie = NULL;
+ avl_tree_t *tree = &vdrz->vd_expand_txgs;
+ while ((re = avl_destroy_nodes(tree, &cookie)) != NULL)
+ kmem_free(re, sizeof (*re));
+ avl_destroy(&vdrz->vd_expand_txgs);
+ mutex_destroy(&vdrz->vd_expand_lock);
+ mutex_destroy(&vdrz->vn_vre.vre_lock);
+ cv_destroy(&vdrz->vn_vre.vre_cv);
+ zfs_rangelock_fini(&vdrz->vn_vre.vre_rangelock);
+ kmem_free(vdrz, sizeof (*vdrz));
+}
+
+/*
+ * Add RAIDZ specific fields to the config nvlist.
+ */
+static void
+vdev_raidz_config_generate(vdev_t *vd, nvlist_t *nv)
+{
+ ASSERT3P(vd->vdev_ops, ==, &vdev_raidz_ops);
+ vdev_raidz_t *vdrz = vd->vdev_tsd;
+
+ /*
+ * Make sure someone hasn't managed to sneak a fancy new vdev
+ * into a crufty old storage pool.
+ */
+ ASSERT(vdrz->vd_nparity == 1 ||
+ (vdrz->vd_nparity <= 2 &&
+ spa_version(vd->vdev_spa) >= SPA_VERSION_RAIDZ2) ||
+ (vdrz->vd_nparity <= 3 &&
+ spa_version(vd->vdev_spa) >= SPA_VERSION_RAIDZ3));
+
+ /*
+ * Note that we'll add these even on storage pools where they
+ * aren't strictly required -- older software will just ignore
+ * it.
+ */
+ fnvlist_add_uint64(nv, ZPOOL_CONFIG_NPARITY, vdrz->vd_nparity);
+
+ if (vdrz->vn_vre.vre_state == DSS_SCANNING) {
+ fnvlist_add_boolean(nv, ZPOOL_CONFIG_RAIDZ_EXPANDING);
+ }
+
+ mutex_enter(&vdrz->vd_expand_lock);
+ if (!avl_is_empty(&vdrz->vd_expand_txgs)) {
+ uint64_t count = avl_numnodes(&vdrz->vd_expand_txgs);
+ uint64_t *txgs = kmem_alloc(sizeof (uint64_t) * count,
+ KM_SLEEP);
+ uint64_t i = 0;
+
+ for (reflow_node_t *re = avl_first(&vdrz->vd_expand_txgs);
+ re != NULL; re = AVL_NEXT(&vdrz->vd_expand_txgs, re)) {
+ txgs[i++] = re->re_txg;
+ }
+
+ fnvlist_add_uint64_array(nv, ZPOOL_CONFIG_RAIDZ_EXPAND_TXGS,
+ txgs, count);
+
+ kmem_free(txgs, sizeof (uint64_t) * count);
+ }
+ mutex_exit(&vdrz->vd_expand_lock);
+}
+
+static uint64_t
+vdev_raidz_nparity(vdev_t *vd)
+{
+ vdev_raidz_t *vdrz = vd->vdev_tsd;
+ return (vdrz->vd_nparity);
+}
+
+static uint64_t
+vdev_raidz_ndisks(vdev_t *vd)
+{
+ return (vd->vdev_children);
+}
+
+vdev_ops_t vdev_raidz_ops = {
.vdev_op_init = vdev_raidz_init,
.vdev_op_fini = vdev_raidz_fini,
.vdev_op_open = vdev_raidz_open,
.vdev_op_close = vdev_raidz_close,
.vdev_op_asize = vdev_raidz_asize,
.vdev_op_min_asize = vdev_raidz_min_asize,
.vdev_op_min_alloc = NULL,
.vdev_op_io_start = vdev_raidz_io_start,
.vdev_op_io_done = vdev_raidz_io_done,
.vdev_op_state_change = vdev_raidz_state_change,
.vdev_op_need_resilver = vdev_raidz_need_resilver,
.vdev_op_hold = NULL,
.vdev_op_rele = NULL,
.vdev_op_remap = NULL,
.vdev_op_xlate = vdev_raidz_xlate,
.vdev_op_rebuild_asize = NULL,
.vdev_op_metaslab_init = NULL,
.vdev_op_config_generate = vdev_raidz_config_generate,
.vdev_op_nparity = vdev_raidz_nparity,
.vdev_op_ndisks = vdev_raidz_ndisks,
.vdev_op_type = VDEV_TYPE_RAIDZ, /* name of this vdev type */
.vdev_op_leaf = B_FALSE /* not a leaf vdev */
};
+
+/* BEGIN CSTYLED */
+ZFS_MODULE_PARAM(zfs_vdev, raidz_, expand_max_reflow_bytes, ULONG, ZMOD_RW,
+ "For testing, pause RAIDZ expansion after reflowing this many bytes");
+ZFS_MODULE_PARAM(zfs_vdev, raidz_, expand_max_copy_bytes, ULONG, ZMOD_RW,
+ "Max amount of concurrent i/o for RAIDZ expansion");
+ZFS_MODULE_PARAM(zfs_vdev, raidz_, io_aggregate_rows, ULONG, ZMOD_RW,
+ "For expanded RAIDZ, aggregate reads that have more rows than this");
+ZFS_MODULE_PARAM(zfs, zfs_, scrub_after_expand, INT, ZMOD_RW,
+ "For expanded RAIDZ, automatically start a pool scrub when expansion "
+ "completes");
+/* END CSTYLED */
diff --git a/sys/contrib/openzfs/module/zfs/vdev_trim.c b/sys/contrib/openzfs/module/zfs/vdev_trim.c
index 03e17db024ea..1c54eae40355 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_trim.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_trim.c
@@ -1,1776 +1,1783 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2016 by Delphix. All rights reserved.
* Copyright (c) 2019 by Lawrence Livermore National Security, LLC.
* Copyright (c) 2021 Hewlett Packard Enterprise Development LP
* Copyright 2023 RackTop Systems, Inc.
*/
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/txg.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_trim.h>
#include <sys/metaslab_impl.h>
#include <sys/dsl_synctask.h>
#include <sys/zap.h>
#include <sys/dmu_tx.h>
#include <sys/arc_impl.h>
/*
* TRIM is a feature which is used to notify a SSD that some previously
* written space is no longer allocated by the pool. This is useful because
* writes to a SSD must be performed to blocks which have first been erased.
* Ensuring the SSD always has a supply of erased blocks for new writes
* helps prevent the performance from deteriorating.
*
* There are two supported TRIM methods; manual and automatic.
*
* Manual TRIM:
*
* A manual TRIM is initiated by running the 'zpool trim' command. A single
* 'vdev_trim' thread is created for each leaf vdev, and it is responsible for
* managing that vdev TRIM process. This involves iterating over all the
* metaslabs, calculating the unallocated space ranges, and then issuing the
* required TRIM I/Os.
*
* While a metaslab is being actively trimmed it is not eligible to perform
* new allocations. After traversing all of the metaslabs the thread is
* terminated. Finally, both the requested options and current progress of
* the TRIM are regularly written to the pool. This allows the TRIM to be
* suspended and resumed as needed.
*
* Automatic TRIM:
*
* An automatic TRIM is enabled by setting the 'autotrim' pool property
* to 'on'. When enabled, a `vdev_autotrim' thread is created for each
* top-level (not leaf) vdev in the pool. These threads perform the same
* core TRIM process as a manual TRIM, but with a few key differences.
*
* 1) Automatic TRIM happens continuously in the background and operates
* solely on recently freed blocks (ms_trim not ms_allocatable).
*
* 2) Each thread is associated with a top-level (not leaf) vdev. This has
* the benefit of simplifying the threading model, it makes it easier
* to coordinate administrative commands, and it ensures only a single
* metaslab is disabled at a time. Unlike manual TRIM, this means each
* 'vdev_autotrim' thread is responsible for issuing TRIM I/Os for its
* children.
*
* 3) There is no automatic TRIM progress information stored on disk, nor
* is it reported by 'zpool status'.
*
* While the automatic TRIM process is highly effective it is more likely
* than a manual TRIM to encounter tiny ranges. Ranges less than or equal to
* 'zfs_trim_extent_bytes_min' (32k) are considered too small to efficiently
* TRIM and are skipped. This means small amounts of freed space may not
* be automatically trimmed.
*
* Furthermore, devices with attached hot spares and devices being actively
* replaced are skipped. This is done to avoid adding additional stress to
* a potentially unhealthy device and to minimize the required rebuild time.
*
* For this reason it may be beneficial to occasionally manually TRIM a pool
* even when automatic TRIM is enabled.
*/
/*
* Maximum size of TRIM I/O, ranges will be chunked in to 128MiB lengths.
*/
static unsigned int zfs_trim_extent_bytes_max = 128 * 1024 * 1024;
/*
* Minimum size of TRIM I/O, extents smaller than 32Kib will be skipped.
*/
static unsigned int zfs_trim_extent_bytes_min = 32 * 1024;
/*
* Skip uninitialized metaslabs during the TRIM process. This option is
* useful for pools constructed from large thinly-provisioned devices where
* TRIM operations are slow. As a pool ages an increasing fraction of
* the pools metaslabs will be initialized progressively degrading the
* usefulness of this option. This setting is stored when starting a
* manual TRIM and will persist for the duration of the requested TRIM.
*/
unsigned int zfs_trim_metaslab_skip = 0;
/*
* Maximum number of queued TRIM I/Os per leaf vdev. The number of
* concurrent TRIM I/Os issued to the device is controlled by the
* zfs_vdev_trim_min_active and zfs_vdev_trim_max_active module options.
*/
static unsigned int zfs_trim_queue_limit = 10;
/*
* The minimum number of transaction groups between automatic trims of a
* metaslab. This setting represents a trade-off between issuing more
* efficient TRIM operations, by allowing them to be aggregated longer,
* and issuing them promptly so the trimmed space is available. Note
* that this value is a minimum; metaslabs can be trimmed less frequently
* when there are a large number of ranges which need to be trimmed.
*
* Increasing this value will allow frees to be aggregated for a longer
* time. This can result is larger TRIM operations, and increased memory
* usage in order to track the ranges to be trimmed. Decreasing this value
* has the opposite effect. The default value of 32 was determined though
* testing to be a reasonable compromise.
*/
static unsigned int zfs_trim_txg_batch = 32;
/*
* The trim_args are a control structure which describe how a leaf vdev
* should be trimmed. The core elements are the vdev, the metaslab being
* trimmed and a range tree containing the extents to TRIM. All provided
* ranges must be within the metaslab.
*/
typedef struct trim_args {
/*
* These fields are set by the caller of vdev_trim_ranges().
*/
vdev_t *trim_vdev; /* Leaf vdev to TRIM */
metaslab_t *trim_msp; /* Disabled metaslab */
range_tree_t *trim_tree; /* TRIM ranges (in metaslab) */
trim_type_t trim_type; /* Manual or auto TRIM */
uint64_t trim_extent_bytes_max; /* Maximum TRIM I/O size */
uint64_t trim_extent_bytes_min; /* Minimum TRIM I/O size */
enum trim_flag trim_flags; /* TRIM flags (secure) */
/*
* These fields are updated by vdev_trim_ranges().
*/
hrtime_t trim_start_time; /* Start time */
uint64_t trim_bytes_done; /* Bytes trimmed */
} trim_args_t;
/*
* Determines whether a vdev_trim_thread() should be stopped.
*/
static boolean_t
vdev_trim_should_stop(vdev_t *vd)
{
return (vd->vdev_trim_exit_wanted || !vdev_writeable(vd) ||
- vd->vdev_detached || vd->vdev_top->vdev_removing);
+ vd->vdev_detached || vd->vdev_top->vdev_removing ||
+ vd->vdev_top->vdev_rz_expanding);
}
/*
* Determines whether a vdev_autotrim_thread() should be stopped.
*/
static boolean_t
vdev_autotrim_should_stop(vdev_t *tvd)
{
return (tvd->vdev_autotrim_exit_wanted ||
!vdev_writeable(tvd) || tvd->vdev_removing ||
+ tvd->vdev_rz_expanding ||
spa_get_autotrim(tvd->vdev_spa) == SPA_AUTOTRIM_OFF);
}
/*
* Wait for given number of kicks, return true if the wait is aborted due to
* vdev_autotrim_exit_wanted.
*/
static boolean_t
vdev_autotrim_wait_kick(vdev_t *vd, int num_of_kick)
{
mutex_enter(&vd->vdev_autotrim_lock);
for (int i = 0; i < num_of_kick; i++) {
if (vd->vdev_autotrim_exit_wanted)
break;
cv_wait(&vd->vdev_autotrim_kick_cv, &vd->vdev_autotrim_lock);
}
boolean_t exit_wanted = vd->vdev_autotrim_exit_wanted;
mutex_exit(&vd->vdev_autotrim_lock);
return (exit_wanted);
}
/*
* The sync task for updating the on-disk state of a manual TRIM. This
* is scheduled by vdev_trim_change_state().
*/
static void
vdev_trim_zap_update_sync(void *arg, dmu_tx_t *tx)
{
/*
* We pass in the guid instead of the vdev_t since the vdev may
* have been freed prior to the sync task being processed. This
* happens when a vdev is detached as we call spa_config_vdev_exit(),
* stop the trimming thread, schedule the sync task, and free
* the vdev. Later when the scheduled sync task is invoked, it would
* find that the vdev has been freed.
*/
uint64_t guid = *(uint64_t *)arg;
uint64_t txg = dmu_tx_get_txg(tx);
kmem_free(arg, sizeof (uint64_t));
vdev_t *vd = spa_lookup_by_guid(tx->tx_pool->dp_spa, guid, B_FALSE);
- if (vd == NULL || vd->vdev_top->vdev_removing || !vdev_is_concrete(vd))
+ if (vd == NULL || vd->vdev_top->vdev_removing ||
+ !vdev_is_concrete(vd) || vd->vdev_top->vdev_rz_expanding)
return;
uint64_t last_offset = vd->vdev_trim_offset[txg & TXG_MASK];
vd->vdev_trim_offset[txg & TXG_MASK] = 0;
VERIFY3U(vd->vdev_leaf_zap, !=, 0);
objset_t *mos = vd->vdev_spa->spa_meta_objset;
if (last_offset > 0 || vd->vdev_trim_last_offset == UINT64_MAX) {
if (vd->vdev_trim_last_offset == UINT64_MAX)
last_offset = 0;
vd->vdev_trim_last_offset = last_offset;
VERIFY0(zap_update(mos, vd->vdev_leaf_zap,
VDEV_LEAF_ZAP_TRIM_LAST_OFFSET,
sizeof (last_offset), 1, &last_offset, tx));
}
if (vd->vdev_trim_action_time > 0) {
uint64_t val = (uint64_t)vd->vdev_trim_action_time;
VERIFY0(zap_update(mos, vd->vdev_leaf_zap,
VDEV_LEAF_ZAP_TRIM_ACTION_TIME, sizeof (val),
1, &val, tx));
}
if (vd->vdev_trim_rate > 0) {
uint64_t rate = (uint64_t)vd->vdev_trim_rate;
if (rate == UINT64_MAX)
rate = 0;
VERIFY0(zap_update(mos, vd->vdev_leaf_zap,
VDEV_LEAF_ZAP_TRIM_RATE, sizeof (rate), 1, &rate, tx));
}
uint64_t partial = vd->vdev_trim_partial;
if (partial == UINT64_MAX)
partial = 0;
VERIFY0(zap_update(mos, vd->vdev_leaf_zap, VDEV_LEAF_ZAP_TRIM_PARTIAL,
sizeof (partial), 1, &partial, tx));
uint64_t secure = vd->vdev_trim_secure;
if (secure == UINT64_MAX)
secure = 0;
VERIFY0(zap_update(mos, vd->vdev_leaf_zap, VDEV_LEAF_ZAP_TRIM_SECURE,
sizeof (secure), 1, &secure, tx));
uint64_t trim_state = vd->vdev_trim_state;
VERIFY0(zap_update(mos, vd->vdev_leaf_zap, VDEV_LEAF_ZAP_TRIM_STATE,
sizeof (trim_state), 1, &trim_state, tx));
}
/*
* Update the on-disk state of a manual TRIM. This is called to request
* that a TRIM be started/suspended/canceled, or to change one of the
* TRIM options (partial, secure, rate).
*/
static void
vdev_trim_change_state(vdev_t *vd, vdev_trim_state_t new_state,
uint64_t rate, boolean_t partial, boolean_t secure)
{
ASSERT(MUTEX_HELD(&vd->vdev_trim_lock));
spa_t *spa = vd->vdev_spa;
if (new_state == vd->vdev_trim_state)
return;
/*
* Copy the vd's guid, this will be freed by the sync task.
*/
uint64_t *guid = kmem_zalloc(sizeof (uint64_t), KM_SLEEP);
*guid = vd->vdev_guid;
/*
* If we're suspending, then preserve the original start time.
*/
if (vd->vdev_trim_state != VDEV_TRIM_SUSPENDED) {
vd->vdev_trim_action_time = gethrestime_sec();
}
/*
* If we're activating, then preserve the requested rate and trim
* method. Setting the last offset and rate to UINT64_MAX is used
* as a sentinel to indicate they should be reset to default values.
*/
if (new_state == VDEV_TRIM_ACTIVE) {
if (vd->vdev_trim_state == VDEV_TRIM_COMPLETE ||
vd->vdev_trim_state == VDEV_TRIM_CANCELED) {
vd->vdev_trim_last_offset = UINT64_MAX;
vd->vdev_trim_rate = UINT64_MAX;
vd->vdev_trim_partial = UINT64_MAX;
vd->vdev_trim_secure = UINT64_MAX;
}
if (rate != 0)
vd->vdev_trim_rate = rate;
if (partial != 0)
vd->vdev_trim_partial = partial;
if (secure != 0)
vd->vdev_trim_secure = secure;
}
vdev_trim_state_t old_state = vd->vdev_trim_state;
boolean_t resumed = (old_state == VDEV_TRIM_SUSPENDED);
vd->vdev_trim_state = new_state;
dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir);
VERIFY0(dmu_tx_assign(tx, TXG_WAIT));
dsl_sync_task_nowait(spa_get_dsl(spa), vdev_trim_zap_update_sync,
guid, tx);
switch (new_state) {
case VDEV_TRIM_ACTIVE:
spa_event_notify(spa, vd, NULL,
resumed ? ESC_ZFS_TRIM_RESUME : ESC_ZFS_TRIM_START);
spa_history_log_internal(spa, "trim", tx,
"vdev=%s activated", vd->vdev_path);
break;
case VDEV_TRIM_SUSPENDED:
spa_event_notify(spa, vd, NULL, ESC_ZFS_TRIM_SUSPEND);
spa_history_log_internal(spa, "trim", tx,
"vdev=%s suspended", vd->vdev_path);
break;
case VDEV_TRIM_CANCELED:
if (old_state == VDEV_TRIM_ACTIVE ||
old_state == VDEV_TRIM_SUSPENDED) {
spa_event_notify(spa, vd, NULL, ESC_ZFS_TRIM_CANCEL);
spa_history_log_internal(spa, "trim", tx,
"vdev=%s canceled", vd->vdev_path);
}
break;
case VDEV_TRIM_COMPLETE:
spa_event_notify(spa, vd, NULL, ESC_ZFS_TRIM_FINISH);
spa_history_log_internal(spa, "trim", tx,
"vdev=%s complete", vd->vdev_path);
break;
default:
panic("invalid state %llu", (unsigned long long)new_state);
}
dmu_tx_commit(tx);
if (new_state != VDEV_TRIM_ACTIVE)
spa_notify_waiters(spa);
}
/*
* The zio_done_func_t done callback for each manual TRIM issued. It is
* responsible for updating the TRIM stats, reissuing failed TRIM I/Os,
* and limiting the number of in flight TRIM I/Os.
*/
static void
vdev_trim_cb(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
mutex_enter(&vd->vdev_trim_io_lock);
if (zio->io_error == ENXIO && !vdev_writeable(vd)) {
/*
* The I/O failed because the vdev was unavailable; roll the
* last offset back. (This works because spa_sync waits on
* spa_txg_zio before it runs sync tasks.)
*/
uint64_t *offset =
&vd->vdev_trim_offset[zio->io_txg & TXG_MASK];
*offset = MIN(*offset, zio->io_offset);
} else {
if (zio->io_error != 0) {
vd->vdev_stat.vs_trim_errors++;
spa_iostats_trim_add(vd->vdev_spa, TRIM_TYPE_MANUAL,
0, 0, 0, 0, 1, zio->io_orig_size);
} else {
spa_iostats_trim_add(vd->vdev_spa, TRIM_TYPE_MANUAL,
1, zio->io_orig_size, 0, 0, 0, 0);
}
vd->vdev_trim_bytes_done += zio->io_orig_size;
}
ASSERT3U(vd->vdev_trim_inflight[TRIM_TYPE_MANUAL], >, 0);
vd->vdev_trim_inflight[TRIM_TYPE_MANUAL]--;
cv_broadcast(&vd->vdev_trim_io_cv);
mutex_exit(&vd->vdev_trim_io_lock);
spa_config_exit(vd->vdev_spa, SCL_STATE_ALL, vd);
}
/*
* The zio_done_func_t done callback for each automatic TRIM issued. It
* is responsible for updating the TRIM stats and limiting the number of
* in flight TRIM I/Os. Automatic TRIM I/Os are best effort and are
* never reissued on failure.
*/
static void
vdev_autotrim_cb(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
mutex_enter(&vd->vdev_trim_io_lock);
if (zio->io_error != 0) {
vd->vdev_stat.vs_trim_errors++;
spa_iostats_trim_add(vd->vdev_spa, TRIM_TYPE_AUTO,
0, 0, 0, 0, 1, zio->io_orig_size);
} else {
spa_iostats_trim_add(vd->vdev_spa, TRIM_TYPE_AUTO,
1, zio->io_orig_size, 0, 0, 0, 0);
}
ASSERT3U(vd->vdev_trim_inflight[TRIM_TYPE_AUTO], >, 0);
vd->vdev_trim_inflight[TRIM_TYPE_AUTO]--;
cv_broadcast(&vd->vdev_trim_io_cv);
mutex_exit(&vd->vdev_trim_io_lock);
spa_config_exit(vd->vdev_spa, SCL_STATE_ALL, vd);
}
/*
* The zio_done_func_t done callback for each TRIM issued via
* vdev_trim_simple(). It is responsible for updating the TRIM stats and
* limiting the number of in flight TRIM I/Os. Simple TRIM I/Os are best
* effort and are never reissued on failure.
*/
static void
vdev_trim_simple_cb(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
mutex_enter(&vd->vdev_trim_io_lock);
if (zio->io_error != 0) {
vd->vdev_stat.vs_trim_errors++;
spa_iostats_trim_add(vd->vdev_spa, TRIM_TYPE_SIMPLE,
0, 0, 0, 0, 1, zio->io_orig_size);
} else {
spa_iostats_trim_add(vd->vdev_spa, TRIM_TYPE_SIMPLE,
1, zio->io_orig_size, 0, 0, 0, 0);
}
ASSERT3U(vd->vdev_trim_inflight[TRIM_TYPE_SIMPLE], >, 0);
vd->vdev_trim_inflight[TRIM_TYPE_SIMPLE]--;
cv_broadcast(&vd->vdev_trim_io_cv);
mutex_exit(&vd->vdev_trim_io_lock);
spa_config_exit(vd->vdev_spa, SCL_STATE_ALL, vd);
}
/*
* Returns the average trim rate in bytes/sec for the ta->trim_vdev.
*/
static uint64_t
vdev_trim_calculate_rate(trim_args_t *ta)
{
return (ta->trim_bytes_done * 1000 /
(NSEC2MSEC(gethrtime() - ta->trim_start_time) + 1));
}
/*
* Issues a physical TRIM and takes care of rate limiting (bytes/sec)
* and number of concurrent TRIM I/Os.
*/
static int
vdev_trim_range(trim_args_t *ta, uint64_t start, uint64_t size)
{
vdev_t *vd = ta->trim_vdev;
spa_t *spa = vd->vdev_spa;
void *cb;
mutex_enter(&vd->vdev_trim_io_lock);
/*
* Limit manual TRIM I/Os to the requested rate. This does not
* apply to automatic TRIM since no per vdev rate can be specified.
*/
if (ta->trim_type == TRIM_TYPE_MANUAL) {
while (vd->vdev_trim_rate != 0 && !vdev_trim_should_stop(vd) &&
vdev_trim_calculate_rate(ta) > vd->vdev_trim_rate) {
cv_timedwait_idle(&vd->vdev_trim_io_cv,
&vd->vdev_trim_io_lock, ddi_get_lbolt() +
MSEC_TO_TICK(10));
}
}
ta->trim_bytes_done += size;
/* Limit in flight trimming I/Os */
while (vd->vdev_trim_inflight[0] + vd->vdev_trim_inflight[1] +
vd->vdev_trim_inflight[2] >= zfs_trim_queue_limit) {
cv_wait(&vd->vdev_trim_io_cv, &vd->vdev_trim_io_lock);
}
vd->vdev_trim_inflight[ta->trim_type]++;
mutex_exit(&vd->vdev_trim_io_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);
spa_config_enter(spa, SCL_STATE_ALL, vd, RW_READER);
mutex_enter(&vd->vdev_trim_lock);
if (ta->trim_type == TRIM_TYPE_MANUAL &&
vd->vdev_trim_offset[txg & TXG_MASK] == 0) {
uint64_t *guid = kmem_zalloc(sizeof (uint64_t), KM_SLEEP);
*guid = vd->vdev_guid;
/* This is the first write of this txg. */
dsl_sync_task_nowait(spa_get_dsl(spa),
vdev_trim_zap_update_sync, guid, tx);
}
/*
* We know the vdev_t will still be around since all consumers of
* vdev_free must stop the trimming first.
*/
if ((ta->trim_type == TRIM_TYPE_MANUAL &&
vdev_trim_should_stop(vd)) ||
(ta->trim_type == TRIM_TYPE_AUTO &&
vdev_autotrim_should_stop(vd->vdev_top))) {
mutex_enter(&vd->vdev_trim_io_lock);
vd->vdev_trim_inflight[ta->trim_type]--;
mutex_exit(&vd->vdev_trim_io_lock);
spa_config_exit(vd->vdev_spa, SCL_STATE_ALL, vd);
mutex_exit(&vd->vdev_trim_lock);
dmu_tx_commit(tx);
return (SET_ERROR(EINTR));
}
mutex_exit(&vd->vdev_trim_lock);
if (ta->trim_type == TRIM_TYPE_MANUAL)
vd->vdev_trim_offset[txg & TXG_MASK] = start + size;
if (ta->trim_type == TRIM_TYPE_MANUAL) {
cb = vdev_trim_cb;
} else if (ta->trim_type == TRIM_TYPE_AUTO) {
cb = vdev_autotrim_cb;
} else {
cb = vdev_trim_simple_cb;
}
zio_nowait(zio_trim(spa->spa_txg_zio[txg & TXG_MASK], vd,
start, size, cb, NULL, ZIO_PRIORITY_TRIM, ZIO_FLAG_CANFAIL,
ta->trim_flags));
/* vdev_trim_cb and vdev_autotrim_cb release SCL_STATE_ALL */
dmu_tx_commit(tx);
return (0);
}
/*
* Issues TRIM I/Os for all ranges in the provided ta->trim_tree range tree.
* Additional parameters describing how the TRIM should be performed must
* be set in the trim_args structure. See the trim_args definition for
* additional information.
*/
static int
vdev_trim_ranges(trim_args_t *ta)
{
vdev_t *vd = ta->trim_vdev;
zfs_btree_t *t = &ta->trim_tree->rt_root;
zfs_btree_index_t idx;
uint64_t extent_bytes_max = ta->trim_extent_bytes_max;
uint64_t extent_bytes_min = ta->trim_extent_bytes_min;
spa_t *spa = vd->vdev_spa;
int error = 0;
ta->trim_start_time = gethrtime();
ta->trim_bytes_done = 0;
for (range_seg_t *rs = zfs_btree_first(t, &idx); rs != NULL;
rs = zfs_btree_next(t, &idx, &idx)) {
uint64_t size = rs_get_end(rs, ta->trim_tree) - rs_get_start(rs,
ta->trim_tree);
if (extent_bytes_min && size < extent_bytes_min) {
spa_iostats_trim_add(spa, ta->trim_type,
0, 0, 1, size, 0, 0);
continue;
}
/* Split range into legally-sized physical chunks */
uint64_t writes_required = ((size - 1) / extent_bytes_max) + 1;
for (uint64_t w = 0; w < writes_required; w++) {
error = vdev_trim_range(ta, VDEV_LABEL_START_SIZE +
rs_get_start(rs, ta->trim_tree) +
(w *extent_bytes_max), MIN(size -
(w * extent_bytes_max), extent_bytes_max));
if (error != 0) {
goto done;
}
}
}
done:
/*
* Make sure all TRIMs for this metaslab have completed before
* returning. TRIM zios have lower priority over regular or syncing
* zios, so all TRIM zios for this metaslab must complete before the
* metaslab is re-enabled. Otherwise it's possible write zios to
* this metaslab could cut ahead of still queued TRIM zios for this
* metaslab causing corruption if the ranges overlap.
*/
mutex_enter(&vd->vdev_trim_io_lock);
while (vd->vdev_trim_inflight[0] > 0) {
cv_wait(&vd->vdev_trim_io_cv, &vd->vdev_trim_io_lock);
}
mutex_exit(&vd->vdev_trim_io_lock);
return (error);
}
static void
vdev_trim_xlate_last_rs_end(void *arg, range_seg64_t *physical_rs)
{
uint64_t *last_rs_end = (uint64_t *)arg;
if (physical_rs->rs_end > *last_rs_end)
*last_rs_end = physical_rs->rs_end;
}
static void
vdev_trim_xlate_progress(void *arg, range_seg64_t *physical_rs)
{
vdev_t *vd = (vdev_t *)arg;
uint64_t size = physical_rs->rs_end - physical_rs->rs_start;
vd->vdev_trim_bytes_est += size;
if (vd->vdev_trim_last_offset >= physical_rs->rs_end) {
vd->vdev_trim_bytes_done += size;
} else if (vd->vdev_trim_last_offset > physical_rs->rs_start &&
vd->vdev_trim_last_offset <= physical_rs->rs_end) {
vd->vdev_trim_bytes_done +=
vd->vdev_trim_last_offset - physical_rs->rs_start;
}
}
/*
* Calculates the completion percentage of a manual TRIM.
*/
static void
vdev_trim_calculate_progress(vdev_t *vd)
{
ASSERT(spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_READER) ||
spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_WRITER));
ASSERT(vd->vdev_leaf_zap != 0);
vd->vdev_trim_bytes_est = 0;
vd->vdev_trim_bytes_done = 0;
for (uint64_t i = 0; i < vd->vdev_top->vdev_ms_count; i++) {
metaslab_t *msp = vd->vdev_top->vdev_ms[i];
mutex_enter(&msp->ms_lock);
uint64_t ms_free = (msp->ms_size -
metaslab_allocated_space(msp)) /
vdev_get_ndisks(vd->vdev_top);
/*
* Convert the metaslab range to a physical range
* on our vdev. We use this to determine if we are
* in the middle of this metaslab range.
*/
range_seg64_t logical_rs, physical_rs, remain_rs;
logical_rs.rs_start = msp->ms_start;
logical_rs.rs_end = msp->ms_start + msp->ms_size;
/* Metaslab space after this offset has not been trimmed. */
vdev_xlate(vd, &logical_rs, &physical_rs, &remain_rs);
if (vd->vdev_trim_last_offset <= physical_rs.rs_start) {
vd->vdev_trim_bytes_est += ms_free;
mutex_exit(&msp->ms_lock);
continue;
}
/* Metaslab space before this offset has been trimmed */
uint64_t last_rs_end = physical_rs.rs_end;
if (!vdev_xlate_is_empty(&remain_rs)) {
vdev_xlate_walk(vd, &remain_rs,
vdev_trim_xlate_last_rs_end, &last_rs_end);
}
if (vd->vdev_trim_last_offset > last_rs_end) {
vd->vdev_trim_bytes_done += ms_free;
vd->vdev_trim_bytes_est += ms_free;
mutex_exit(&msp->ms_lock);
continue;
}
/*
* If we get here, we're in the middle of trimming this
* metaslab. Load it and walk the free tree for more
* accurate progress estimation.
*/
VERIFY0(metaslab_load(msp));
range_tree_t *rt = msp->ms_allocatable;
zfs_btree_t *bt = &rt->rt_root;
zfs_btree_index_t idx;
for (range_seg_t *rs = zfs_btree_first(bt, &idx);
rs != NULL; rs = zfs_btree_next(bt, &idx, &idx)) {
logical_rs.rs_start = rs_get_start(rs, rt);
logical_rs.rs_end = rs_get_end(rs, rt);
vdev_xlate_walk(vd, &logical_rs,
vdev_trim_xlate_progress, vd);
}
mutex_exit(&msp->ms_lock);
}
}
/*
* Load from disk the vdev's manual TRIM information. This includes the
* state, progress, and options provided when initiating the manual TRIM.
*/
static int
vdev_trim_load(vdev_t *vd)
{
int err = 0;
ASSERT(spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_READER) ||
spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_WRITER));
ASSERT(vd->vdev_leaf_zap != 0);
if (vd->vdev_trim_state == VDEV_TRIM_ACTIVE ||
vd->vdev_trim_state == VDEV_TRIM_SUSPENDED) {
err = zap_lookup(vd->vdev_spa->spa_meta_objset,
vd->vdev_leaf_zap, VDEV_LEAF_ZAP_TRIM_LAST_OFFSET,
sizeof (vd->vdev_trim_last_offset), 1,
&vd->vdev_trim_last_offset);
if (err == ENOENT) {
vd->vdev_trim_last_offset = 0;
err = 0;
}
if (err == 0) {
err = zap_lookup(vd->vdev_spa->spa_meta_objset,
vd->vdev_leaf_zap, VDEV_LEAF_ZAP_TRIM_RATE,
sizeof (vd->vdev_trim_rate), 1,
&vd->vdev_trim_rate);
if (err == ENOENT) {
vd->vdev_trim_rate = 0;
err = 0;
}
}
if (err == 0) {
err = zap_lookup(vd->vdev_spa->spa_meta_objset,
vd->vdev_leaf_zap, VDEV_LEAF_ZAP_TRIM_PARTIAL,
sizeof (vd->vdev_trim_partial), 1,
&vd->vdev_trim_partial);
if (err == ENOENT) {
vd->vdev_trim_partial = 0;
err = 0;
}
}
if (err == 0) {
err = zap_lookup(vd->vdev_spa->spa_meta_objset,
vd->vdev_leaf_zap, VDEV_LEAF_ZAP_TRIM_SECURE,
sizeof (vd->vdev_trim_secure), 1,
&vd->vdev_trim_secure);
if (err == ENOENT) {
vd->vdev_trim_secure = 0;
err = 0;
}
}
}
vdev_trim_calculate_progress(vd);
return (err);
}
static void
vdev_trim_xlate_range_add(void *arg, range_seg64_t *physical_rs)
{
trim_args_t *ta = arg;
vdev_t *vd = ta->trim_vdev;
/*
* Only a manual trim will be traversing the vdev sequentially.
* For an auto trim all valid ranges should be added.
*/
if (ta->trim_type == TRIM_TYPE_MANUAL) {
/* Only add segments that we have not visited yet */
if (physical_rs->rs_end <= vd->vdev_trim_last_offset)
return;
/* Pick up where we left off mid-range. */
if (vd->vdev_trim_last_offset > physical_rs->rs_start) {
ASSERT3U(physical_rs->rs_end, >,
vd->vdev_trim_last_offset);
physical_rs->rs_start = vd->vdev_trim_last_offset;
}
}
ASSERT3U(physical_rs->rs_end, >, physical_rs->rs_start);
range_tree_add(ta->trim_tree, physical_rs->rs_start,
physical_rs->rs_end - physical_rs->rs_start);
}
/*
* Convert the logical range into physical ranges and add them to the
* range tree passed in the trim_args_t.
*/
static void
vdev_trim_range_add(void *arg, uint64_t start, uint64_t size)
{
trim_args_t *ta = arg;
vdev_t *vd = ta->trim_vdev;
range_seg64_t logical_rs;
logical_rs.rs_start = start;
logical_rs.rs_end = start + size;
/*
* Every range to be trimmed must be part of ms_allocatable.
* When ZFS_DEBUG_TRIM is set load the metaslab to verify this
* is always the case.
*/
if (zfs_flags & ZFS_DEBUG_TRIM) {
metaslab_t *msp = ta->trim_msp;
VERIFY0(metaslab_load(msp));
VERIFY3B(msp->ms_loaded, ==, B_TRUE);
VERIFY(range_tree_contains(msp->ms_allocatable, start, size));
}
ASSERT(vd->vdev_ops->vdev_op_leaf);
vdev_xlate_walk(vd, &logical_rs, vdev_trim_xlate_range_add, arg);
}
/*
* Each manual TRIM thread is responsible for trimming the unallocated
* space for each leaf vdev. This is accomplished by sequentially iterating
* over its top-level metaslabs and issuing TRIM I/O for the space described
* by its ms_allocatable. While a metaslab is undergoing trimming it is
* not eligible for new allocations.
*/
static __attribute__((noreturn)) void
vdev_trim_thread(void *arg)
{
vdev_t *vd = arg;
spa_t *spa = vd->vdev_spa;
trim_args_t ta;
int error = 0;
/*
* The VDEV_LEAF_ZAP_TRIM_* entries may have been updated by
* vdev_trim(). Wait for the updated values to be reflected
* in the zap in order to start with the requested settings.
*/
txg_wait_synced(spa_get_dsl(vd->vdev_spa), 0);
ASSERT(vdev_is_concrete(vd));
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vd->vdev_trim_last_offset = 0;
vd->vdev_trim_rate = 0;
vd->vdev_trim_partial = 0;
vd->vdev_trim_secure = 0;
VERIFY0(vdev_trim_load(vd));
ta.trim_vdev = vd;
ta.trim_extent_bytes_max = zfs_trim_extent_bytes_max;
ta.trim_extent_bytes_min = zfs_trim_extent_bytes_min;
ta.trim_tree = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0);
ta.trim_type = TRIM_TYPE_MANUAL;
ta.trim_flags = 0;
/*
* When a secure TRIM has been requested infer that the intent
* is that everything must be trimmed. Override the default
* minimum TRIM size to prevent ranges from being skipped.
*/
if (vd->vdev_trim_secure) {
ta.trim_flags |= ZIO_TRIM_SECURE;
ta.trim_extent_bytes_min = SPA_MINBLOCKSIZE;
}
uint64_t ms_count = 0;
for (uint64_t i = 0; !vd->vdev_detached &&
i < vd->vdev_top->vdev_ms_count; i++) {
metaslab_t *msp = vd->vdev_top->vdev_ms[i];
/*
* If we've expanded the top-level vdev or it's our
* first pass, calculate our progress.
*/
if (vd->vdev_top->vdev_ms_count != ms_count) {
vdev_trim_calculate_progress(vd);
ms_count = vd->vdev_top->vdev_ms_count;
}
spa_config_exit(spa, SCL_CONFIG, FTAG);
metaslab_disable(msp);
mutex_enter(&msp->ms_lock);
VERIFY0(metaslab_load(msp));
/*
* If a partial TRIM was requested skip metaslabs which have
* never been initialized and thus have never been written.
*/
if (msp->ms_sm == NULL && vd->vdev_trim_partial) {
mutex_exit(&msp->ms_lock);
metaslab_enable(msp, B_FALSE, B_FALSE);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_trim_calculate_progress(vd);
continue;
}
ta.trim_msp = msp;
range_tree_walk(msp->ms_allocatable, vdev_trim_range_add, &ta);
range_tree_vacate(msp->ms_trim, NULL, NULL);
mutex_exit(&msp->ms_lock);
error = vdev_trim_ranges(&ta);
metaslab_enable(msp, B_TRUE, B_FALSE);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
range_tree_vacate(ta.trim_tree, NULL, NULL);
if (error != 0)
break;
}
spa_config_exit(spa, SCL_CONFIG, FTAG);
range_tree_destroy(ta.trim_tree);
mutex_enter(&vd->vdev_trim_lock);
if (!vd->vdev_trim_exit_wanted) {
if (vdev_writeable(vd)) {
vdev_trim_change_state(vd, VDEV_TRIM_COMPLETE,
vd->vdev_trim_rate, vd->vdev_trim_partial,
vd->vdev_trim_secure);
} else if (vd->vdev_faulted) {
vdev_trim_change_state(vd, VDEV_TRIM_CANCELED,
vd->vdev_trim_rate, vd->vdev_trim_partial,
vd->vdev_trim_secure);
}
}
ASSERT(vd->vdev_trim_thread != NULL || vd->vdev_trim_inflight[0] == 0);
/*
* Drop the vdev_trim_lock while we sync out the txg since it's
* possible that a device might be trying to come online and must
* check to see if it needs to restart a trim. That thread will be
* holding the spa_config_lock which would prevent the txg_wait_synced
* from completing.
*/
mutex_exit(&vd->vdev_trim_lock);
txg_wait_synced(spa_get_dsl(spa), 0);
mutex_enter(&vd->vdev_trim_lock);
vd->vdev_trim_thread = NULL;
cv_broadcast(&vd->vdev_trim_cv);
mutex_exit(&vd->vdev_trim_lock);
thread_exit();
}
/*
* Initiates a manual TRIM for the vdev_t. Callers must hold vdev_trim_lock,
* the vdev_t must be a leaf and cannot already be manually trimming.
*/
void
vdev_trim(vdev_t *vd, uint64_t rate, boolean_t partial, boolean_t secure)
{
ASSERT(MUTEX_HELD(&vd->vdev_trim_lock));
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(vdev_is_concrete(vd));
ASSERT3P(vd->vdev_trim_thread, ==, NULL);
ASSERT(!vd->vdev_detached);
ASSERT(!vd->vdev_trim_exit_wanted);
ASSERT(!vd->vdev_top->vdev_removing);
+ ASSERT(!vd->vdev_rz_expanding);
vdev_trim_change_state(vd, VDEV_TRIM_ACTIVE, rate, partial, secure);
vd->vdev_trim_thread = thread_create(NULL, 0,
vdev_trim_thread, vd, 0, &p0, TS_RUN, maxclsyspri);
}
/*
* Wait for the trimming thread to be terminated (canceled or stopped).
*/
static void
vdev_trim_stop_wait_impl(vdev_t *vd)
{
ASSERT(MUTEX_HELD(&vd->vdev_trim_lock));
while (vd->vdev_trim_thread != NULL)
cv_wait(&vd->vdev_trim_cv, &vd->vdev_trim_lock);
ASSERT3P(vd->vdev_trim_thread, ==, NULL);
vd->vdev_trim_exit_wanted = B_FALSE;
}
/*
* Wait for vdev trim threads which were listed to cleanly exit.
*/
void
vdev_trim_stop_wait(spa_t *spa, list_t *vd_list)
{
(void) spa;
vdev_t *vd;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
while ((vd = list_remove_head(vd_list)) != NULL) {
mutex_enter(&vd->vdev_trim_lock);
vdev_trim_stop_wait_impl(vd);
mutex_exit(&vd->vdev_trim_lock);
}
}
/*
* Stop trimming a device, with the resultant trimming state being tgt_state.
* For blocking behavior pass NULL for vd_list. Otherwise, when a list_t is
* provided the stopping vdev is inserted in to the list. Callers are then
* required to call vdev_trim_stop_wait() to block for all the trim threads
* to exit. The caller must hold vdev_trim_lock and must not be writing to
* the spa config, as the trimming thread may try to enter the config as a
* reader before exiting.
*/
void
vdev_trim_stop(vdev_t *vd, vdev_trim_state_t tgt_state, list_t *vd_list)
{
ASSERT(!spa_config_held(vd->vdev_spa, SCL_CONFIG|SCL_STATE, RW_WRITER));
ASSERT(MUTEX_HELD(&vd->vdev_trim_lock));
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(vdev_is_concrete(vd));
/*
* Allow cancel requests to proceed even if the trim thread has
* stopped.
*/
if (vd->vdev_trim_thread == NULL && tgt_state != VDEV_TRIM_CANCELED)
return;
vdev_trim_change_state(vd, tgt_state, 0, 0, 0);
vd->vdev_trim_exit_wanted = B_TRUE;
if (vd_list == NULL) {
vdev_trim_stop_wait_impl(vd);
} else {
ASSERT(MUTEX_HELD(&spa_namespace_lock));
list_insert_tail(vd_list, vd);
}
}
/*
* Requests that all listed vdevs stop trimming.
*/
static void
vdev_trim_stop_all_impl(vdev_t *vd, vdev_trim_state_t tgt_state,
list_t *vd_list)
{
if (vd->vdev_ops->vdev_op_leaf && vdev_is_concrete(vd)) {
mutex_enter(&vd->vdev_trim_lock);
vdev_trim_stop(vd, tgt_state, vd_list);
mutex_exit(&vd->vdev_trim_lock);
return;
}
for (uint64_t i = 0; i < vd->vdev_children; i++) {
vdev_trim_stop_all_impl(vd->vdev_child[i], tgt_state,
vd_list);
}
}
/*
* Convenience function to stop trimming of a vdev tree and set all trim
* thread pointers to NULL.
*/
void
vdev_trim_stop_all(vdev_t *vd, vdev_trim_state_t tgt_state)
{
spa_t *spa = vd->vdev_spa;
list_t vd_list;
vdev_t *vd_l2cache;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
list_create(&vd_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_trim_node));
vdev_trim_stop_all_impl(vd, tgt_state, &vd_list);
/*
* Iterate over cache devices and request stop trimming the
* whole device in case we export the pool or remove the cache
* device prematurely.
*/
for (int i = 0; i < spa->spa_l2cache.sav_count; i++) {
vd_l2cache = spa->spa_l2cache.sav_vdevs[i];
vdev_trim_stop_all_impl(vd_l2cache, tgt_state, &vd_list);
}
vdev_trim_stop_wait(spa, &vd_list);
if (vd->vdev_spa->spa_sync_on) {
/* Make sure that our state has been synced to disk */
txg_wait_synced(spa_get_dsl(vd->vdev_spa), 0);
}
list_destroy(&vd_list);
}
/*
* Conditionally restarts a manual TRIM given its on-disk state.
*/
void
vdev_trim_restart(vdev_t *vd)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(!spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER));
if (vd->vdev_leaf_zap != 0) {
mutex_enter(&vd->vdev_trim_lock);
uint64_t trim_state = VDEV_TRIM_NONE;
int err = zap_lookup(vd->vdev_spa->spa_meta_objset,
vd->vdev_leaf_zap, VDEV_LEAF_ZAP_TRIM_STATE,
sizeof (trim_state), 1, &trim_state);
ASSERT(err == 0 || err == ENOENT);
vd->vdev_trim_state = trim_state;
uint64_t timestamp = 0;
err = zap_lookup(vd->vdev_spa->spa_meta_objset,
vd->vdev_leaf_zap, VDEV_LEAF_ZAP_TRIM_ACTION_TIME,
sizeof (timestamp), 1, &timestamp);
ASSERT(err == 0 || err == ENOENT);
vd->vdev_trim_action_time = timestamp;
- if (vd->vdev_trim_state == VDEV_TRIM_SUSPENDED ||
- vd->vdev_offline) {
+ if ((vd->vdev_trim_state == VDEV_TRIM_SUSPENDED ||
+ vd->vdev_offline) && !vd->vdev_top->vdev_rz_expanding) {
/* load progress for reporting, but don't resume */
VERIFY0(vdev_trim_load(vd));
} else if (vd->vdev_trim_state == VDEV_TRIM_ACTIVE &&
vdev_writeable(vd) && !vd->vdev_top->vdev_removing &&
+ !vd->vdev_top->vdev_rz_expanding &&
vd->vdev_trim_thread == NULL) {
VERIFY0(vdev_trim_load(vd));
vdev_trim(vd, vd->vdev_trim_rate,
vd->vdev_trim_partial, vd->vdev_trim_secure);
}
mutex_exit(&vd->vdev_trim_lock);
}
for (uint64_t i = 0; i < vd->vdev_children; i++) {
vdev_trim_restart(vd->vdev_child[i]);
}
}
/*
* Used by the automatic TRIM when ZFS_DEBUG_TRIM is set to verify that
* every TRIM range is contained within ms_allocatable.
*/
static void
vdev_trim_range_verify(void *arg, uint64_t start, uint64_t size)
{
trim_args_t *ta = arg;
metaslab_t *msp = ta->trim_msp;
VERIFY3B(msp->ms_loaded, ==, B_TRUE);
VERIFY3U(msp->ms_disabled, >, 0);
VERIFY(range_tree_contains(msp->ms_allocatable, start, size));
}
/*
* Each automatic TRIM thread is responsible for managing the trimming of a
* top-level vdev in the pool. No automatic TRIM state is maintained on-disk.
*
* N.B. This behavior is different from a manual TRIM where a thread
* is created for each leaf vdev, instead of each top-level vdev.
*/
static __attribute__((noreturn)) void
vdev_autotrim_thread(void *arg)
{
vdev_t *vd = arg;
spa_t *spa = vd->vdev_spa;
int shift = 0;
mutex_enter(&vd->vdev_autotrim_lock);
ASSERT3P(vd->vdev_top, ==, vd);
ASSERT3P(vd->vdev_autotrim_thread, !=, NULL);
mutex_exit(&vd->vdev_autotrim_lock);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
while (!vdev_autotrim_should_stop(vd)) {
int txgs_per_trim = MAX(zfs_trim_txg_batch, 1);
uint64_t extent_bytes_max = zfs_trim_extent_bytes_max;
uint64_t extent_bytes_min = zfs_trim_extent_bytes_min;
/*
* All of the metaslabs are divided in to groups of size
* num_metaslabs / zfs_trim_txg_batch. Each of these groups
* is composed of metaslabs which are spread evenly over the
* device.
*
* For example, when zfs_trim_txg_batch = 32 (default) then
* group 0 will contain metaslabs 0, 32, 64, ...;
* group 1 will contain metaslabs 1, 33, 65, ...;
* group 2 will contain metaslabs 2, 34, 66, ...; and so on.
*
* On each pass through the while() loop one of these groups
* is selected. This is accomplished by using a shift value
* to select the starting metaslab, then striding over the
* metaslabs using the zfs_trim_txg_batch size. This is
* done to accomplish two things.
*
* 1) By dividing the metaslabs in to groups, and making sure
* that each group takes a minimum of one txg to process.
* Then zfs_trim_txg_batch controls the minimum number of
* txgs which must occur before a metaslab is revisited.
*
* 2) Selecting non-consecutive metaslabs distributes the
* TRIM commands for a group evenly over the entire device.
* This can be advantageous for certain types of devices.
*/
for (uint64_t i = shift % txgs_per_trim; i < vd->vdev_ms_count;
i += txgs_per_trim) {
metaslab_t *msp = vd->vdev_ms[i];
range_tree_t *trim_tree;
boolean_t issued_trim = B_FALSE;
boolean_t wait_aborted = B_FALSE;
spa_config_exit(spa, SCL_CONFIG, FTAG);
metaslab_disable(msp);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
mutex_enter(&msp->ms_lock);
/*
* Skip the metaslab when it has never been allocated
* or when there are no recent frees to trim.
*/
if (msp->ms_sm == NULL ||
range_tree_is_empty(msp->ms_trim)) {
mutex_exit(&msp->ms_lock);
metaslab_enable(msp, B_FALSE, B_FALSE);
continue;
}
/*
* Skip the metaslab when it has already been disabled.
* This may happen when a manual TRIM or initialize
* operation is running concurrently. In the case
* of a manual TRIM, the ms_trim tree will have been
* vacated. Only ranges added after the manual TRIM
* disabled the metaslab will be included in the tree.
* These will be processed when the automatic TRIM
* next revisits this metaslab.
*/
if (msp->ms_disabled > 1) {
mutex_exit(&msp->ms_lock);
metaslab_enable(msp, B_FALSE, B_FALSE);
continue;
}
/*
* Allocate an empty range tree which is swapped in
* for the existing ms_trim tree while it is processed.
*/
trim_tree = range_tree_create(NULL, RANGE_SEG64, NULL,
0, 0);
range_tree_swap(&msp->ms_trim, &trim_tree);
ASSERT(range_tree_is_empty(msp->ms_trim));
/*
* There are two cases when constructing the per-vdev
* trim trees for a metaslab. If the top-level vdev
* has no children then it is also a leaf and should
* be trimmed. Otherwise our children are the leaves
* and a trim tree should be constructed for each.
*/
trim_args_t *tap;
uint64_t children = vd->vdev_children;
if (children == 0) {
children = 1;
tap = kmem_zalloc(sizeof (trim_args_t) *
children, KM_SLEEP);
tap[0].trim_vdev = vd;
} else {
tap = kmem_zalloc(sizeof (trim_args_t) *
children, KM_SLEEP);
for (uint64_t c = 0; c < children; c++) {
tap[c].trim_vdev = vd->vdev_child[c];
}
}
for (uint64_t c = 0; c < children; c++) {
trim_args_t *ta = &tap[c];
vdev_t *cvd = ta->trim_vdev;
ta->trim_msp = msp;
ta->trim_extent_bytes_max = extent_bytes_max;
ta->trim_extent_bytes_min = extent_bytes_min;
ta->trim_type = TRIM_TYPE_AUTO;
ta->trim_flags = 0;
if (cvd->vdev_detached ||
!vdev_writeable(cvd) ||
!cvd->vdev_has_trim ||
cvd->vdev_trim_thread != NULL) {
continue;
}
/*
* When a device has an attached hot spare, or
* is being replaced it will not be trimmed.
* This is done to avoid adding additional
* stress to a potentially unhealthy device,
* and to minimize the required rebuild time.
*/
if (!cvd->vdev_ops->vdev_op_leaf)
continue;
ta->trim_tree = range_tree_create(NULL,
RANGE_SEG64, NULL, 0, 0);
range_tree_walk(trim_tree,
vdev_trim_range_add, ta);
}
mutex_exit(&msp->ms_lock);
spa_config_exit(spa, SCL_CONFIG, FTAG);
/*
* Issue the TRIM I/Os for all ranges covered by the
* TRIM trees. These ranges are safe to TRIM because
* no new allocations will be performed until the call
* to metaslab_enabled() below.
*/
for (uint64_t c = 0; c < children; c++) {
trim_args_t *ta = &tap[c];
/*
* Always yield to a manual TRIM if one has
* been started for the child vdev.
*/
if (ta->trim_tree == NULL ||
ta->trim_vdev->vdev_trim_thread != NULL) {
continue;
}
/*
* After this point metaslab_enable() must be
* called with the sync flag set. This is done
* here because vdev_trim_ranges() is allowed
* to be interrupted (EINTR) before issuing all
* of the required TRIM I/Os.
*/
issued_trim = B_TRUE;
int error = vdev_trim_ranges(ta);
if (error)
break;
}
/*
* Verify every range which was trimmed is still
* contained within the ms_allocatable tree.
*/
if (zfs_flags & ZFS_DEBUG_TRIM) {
mutex_enter(&msp->ms_lock);
VERIFY0(metaslab_load(msp));
VERIFY3P(tap[0].trim_msp, ==, msp);
range_tree_walk(trim_tree,
vdev_trim_range_verify, &tap[0]);
mutex_exit(&msp->ms_lock);
}
range_tree_vacate(trim_tree, NULL, NULL);
range_tree_destroy(trim_tree);
/*
* Wait for couples of kicks, to ensure the trim io is
* synced. If the wait is aborted due to
* vdev_autotrim_exit_wanted, we need to signal
* metaslab_enable() to wait for sync.
*/
if (issued_trim) {
wait_aborted = vdev_autotrim_wait_kick(vd,
TXG_CONCURRENT_STATES + TXG_DEFER_SIZE);
}
metaslab_enable(msp, wait_aborted, B_FALSE);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
for (uint64_t c = 0; c < children; c++) {
trim_args_t *ta = &tap[c];
if (ta->trim_tree == NULL)
continue;
range_tree_vacate(ta->trim_tree, NULL, NULL);
range_tree_destroy(ta->trim_tree);
}
kmem_free(tap, sizeof (trim_args_t) * children);
if (vdev_autotrim_should_stop(vd))
break;
}
spa_config_exit(spa, SCL_CONFIG, FTAG);
vdev_autotrim_wait_kick(vd, 1);
shift++;
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
}
for (uint64_t c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
mutex_enter(&cvd->vdev_trim_io_lock);
while (cvd->vdev_trim_inflight[1] > 0) {
cv_wait(&cvd->vdev_trim_io_cv,
&cvd->vdev_trim_io_lock);
}
mutex_exit(&cvd->vdev_trim_io_lock);
}
spa_config_exit(spa, SCL_CONFIG, FTAG);
/*
* When exiting because the autotrim property was set to off, then
* abandon any unprocessed ms_trim ranges to reclaim the memory.
*/
if (spa_get_autotrim(spa) == SPA_AUTOTRIM_OFF) {
for (uint64_t i = 0; i < vd->vdev_ms_count; i++) {
metaslab_t *msp = vd->vdev_ms[i];
mutex_enter(&msp->ms_lock);
range_tree_vacate(msp->ms_trim, NULL, NULL);
mutex_exit(&msp->ms_lock);
}
}
mutex_enter(&vd->vdev_autotrim_lock);
ASSERT(vd->vdev_autotrim_thread != NULL);
vd->vdev_autotrim_thread = NULL;
cv_broadcast(&vd->vdev_autotrim_cv);
mutex_exit(&vd->vdev_autotrim_lock);
thread_exit();
}
/*
* Starts an autotrim thread, if needed, for each top-level vdev which can be
* trimmed. A top-level vdev which has been evacuated will never be trimmed.
*/
void
vdev_autotrim(spa_t *spa)
{
vdev_t *root_vd = spa->spa_root_vdev;
for (uint64_t i = 0; i < root_vd->vdev_children; i++) {
vdev_t *tvd = root_vd->vdev_child[i];
mutex_enter(&tvd->vdev_autotrim_lock);
if (vdev_writeable(tvd) && !tvd->vdev_removing &&
- tvd->vdev_autotrim_thread == NULL) {
+ tvd->vdev_autotrim_thread == NULL &&
+ !tvd->vdev_rz_expanding) {
ASSERT3P(tvd->vdev_top, ==, tvd);
tvd->vdev_autotrim_thread = thread_create(NULL, 0,
vdev_autotrim_thread, tvd, 0, &p0, TS_RUN,
maxclsyspri);
ASSERT(tvd->vdev_autotrim_thread != NULL);
}
mutex_exit(&tvd->vdev_autotrim_lock);
}
}
/*
* Wait for the vdev_autotrim_thread associated with the passed top-level
* vdev to be terminated (canceled or stopped).
*/
void
vdev_autotrim_stop_wait(vdev_t *tvd)
{
mutex_enter(&tvd->vdev_autotrim_lock);
if (tvd->vdev_autotrim_thread != NULL) {
tvd->vdev_autotrim_exit_wanted = B_TRUE;
cv_broadcast(&tvd->vdev_autotrim_kick_cv);
cv_wait(&tvd->vdev_autotrim_cv,
&tvd->vdev_autotrim_lock);
ASSERT3P(tvd->vdev_autotrim_thread, ==, NULL);
tvd->vdev_autotrim_exit_wanted = B_FALSE;
}
mutex_exit(&tvd->vdev_autotrim_lock);
}
void
vdev_autotrim_kick(spa_t *spa)
{
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER));
vdev_t *root_vd = spa->spa_root_vdev;
vdev_t *tvd;
for (uint64_t i = 0; i < root_vd->vdev_children; i++) {
tvd = root_vd->vdev_child[i];
mutex_enter(&tvd->vdev_autotrim_lock);
if (tvd->vdev_autotrim_thread != NULL)
cv_broadcast(&tvd->vdev_autotrim_kick_cv);
mutex_exit(&tvd->vdev_autotrim_lock);
}
}
/*
* Wait for all of the vdev_autotrim_thread associated with the pool to
* be terminated (canceled or stopped).
*/
void
vdev_autotrim_stop_all(spa_t *spa)
{
vdev_t *root_vd = spa->spa_root_vdev;
for (uint64_t i = 0; i < root_vd->vdev_children; i++)
vdev_autotrim_stop_wait(root_vd->vdev_child[i]);
}
/*
* Conditionally restart all of the vdev_autotrim_thread's for the pool.
*/
void
vdev_autotrim_restart(spa_t *spa)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
if (spa->spa_autotrim)
vdev_autotrim(spa);
}
static __attribute__((noreturn)) void
vdev_trim_l2arc_thread(void *arg)
{
vdev_t *vd = arg;
spa_t *spa = vd->vdev_spa;
l2arc_dev_t *dev = l2arc_vdev_get(vd);
trim_args_t ta = {0};
range_seg64_t physical_rs;
ASSERT(vdev_is_concrete(vd));
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vd->vdev_trim_last_offset = 0;
vd->vdev_trim_rate = 0;
vd->vdev_trim_partial = 0;
vd->vdev_trim_secure = 0;
ta.trim_vdev = vd;
ta.trim_tree = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0);
ta.trim_type = TRIM_TYPE_MANUAL;
ta.trim_extent_bytes_max = zfs_trim_extent_bytes_max;
ta.trim_extent_bytes_min = SPA_MINBLOCKSIZE;
ta.trim_flags = 0;
physical_rs.rs_start = vd->vdev_trim_bytes_done = 0;
physical_rs.rs_end = vd->vdev_trim_bytes_est =
vdev_get_min_asize(vd);
range_tree_add(ta.trim_tree, physical_rs.rs_start,
physical_rs.rs_end - physical_rs.rs_start);
mutex_enter(&vd->vdev_trim_lock);
vdev_trim_change_state(vd, VDEV_TRIM_ACTIVE, 0, 0, 0);
mutex_exit(&vd->vdev_trim_lock);
(void) vdev_trim_ranges(&ta);
spa_config_exit(spa, SCL_CONFIG, FTAG);
mutex_enter(&vd->vdev_trim_io_lock);
while (vd->vdev_trim_inflight[TRIM_TYPE_MANUAL] > 0) {
cv_wait(&vd->vdev_trim_io_cv, &vd->vdev_trim_io_lock);
}
mutex_exit(&vd->vdev_trim_io_lock);
range_tree_vacate(ta.trim_tree, NULL, NULL);
range_tree_destroy(ta.trim_tree);
mutex_enter(&vd->vdev_trim_lock);
if (!vd->vdev_trim_exit_wanted && vdev_writeable(vd)) {
vdev_trim_change_state(vd, VDEV_TRIM_COMPLETE,
vd->vdev_trim_rate, vd->vdev_trim_partial,
vd->vdev_trim_secure);
}
ASSERT(vd->vdev_trim_thread != NULL ||
vd->vdev_trim_inflight[TRIM_TYPE_MANUAL] == 0);
/*
* Drop the vdev_trim_lock while we sync out the txg since it's
* possible that a device might be trying to come online and
* must check to see if it needs to restart a trim. That thread
* will be holding the spa_config_lock which would prevent the
* txg_wait_synced from completing. Same strategy as in
* vdev_trim_thread().
*/
mutex_exit(&vd->vdev_trim_lock);
txg_wait_synced(spa_get_dsl(vd->vdev_spa), 0);
mutex_enter(&vd->vdev_trim_lock);
/*
* Update the header of the cache device here, before
* broadcasting vdev_trim_cv which may lead to the removal
* of the device. The same applies for setting l2ad_trim_all to
* false.
*/
spa_config_enter(vd->vdev_spa, SCL_L2ARC, vd,
RW_READER);
memset(dev->l2ad_dev_hdr, 0, dev->l2ad_dev_hdr_asize);
l2arc_dev_hdr_update(dev);
spa_config_exit(vd->vdev_spa, SCL_L2ARC, vd);
vd->vdev_trim_thread = NULL;
if (vd->vdev_trim_state == VDEV_TRIM_COMPLETE)
dev->l2ad_trim_all = B_FALSE;
cv_broadcast(&vd->vdev_trim_cv);
mutex_exit(&vd->vdev_trim_lock);
thread_exit();
}
/*
* Punches out TRIM threads for the L2ARC devices in a spa and assigns them
* to vd->vdev_trim_thread variable. This facilitates the management of
* trimming the whole cache device using TRIM_TYPE_MANUAL upon addition
* to a pool or pool creation or when the header of the device is invalid.
*/
void
vdev_trim_l2arc(spa_t *spa)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
/*
* Locate the spa's l2arc devices and kick off TRIM threads.
*/
for (int i = 0; i < spa->spa_l2cache.sav_count; i++) {
vdev_t *vd = spa->spa_l2cache.sav_vdevs[i];
l2arc_dev_t *dev = l2arc_vdev_get(vd);
if (dev == NULL || !dev->l2ad_trim_all) {
/*
* Don't attempt TRIM if the vdev is UNAVAIL or if the
* cache device was not marked for whole device TRIM
* (ie l2arc_trim_ahead = 0, or the L2ARC device header
* is valid with trim_state = VDEV_TRIM_COMPLETE and
* l2ad_log_entries > 0).
*/
continue;
}
mutex_enter(&vd->vdev_trim_lock);
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(vdev_is_concrete(vd));
ASSERT3P(vd->vdev_trim_thread, ==, NULL);
ASSERT(!vd->vdev_detached);
ASSERT(!vd->vdev_trim_exit_wanted);
ASSERT(!vd->vdev_top->vdev_removing);
vdev_trim_change_state(vd, VDEV_TRIM_ACTIVE, 0, 0, 0);
vd->vdev_trim_thread = thread_create(NULL, 0,
vdev_trim_l2arc_thread, vd, 0, &p0, TS_RUN, maxclsyspri);
mutex_exit(&vd->vdev_trim_lock);
}
}
/*
* A wrapper which calls vdev_trim_ranges(). It is intended to be called
* on leaf vdevs.
*/
int
vdev_trim_simple(vdev_t *vd, uint64_t start, uint64_t size)
{
trim_args_t ta = {0};
range_seg64_t physical_rs;
int error;
physical_rs.rs_start = start;
physical_rs.rs_end = start + size;
ASSERT(vdev_is_concrete(vd));
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(!vd->vdev_detached);
ASSERT(!vd->vdev_top->vdev_removing);
+ ASSERT(!vd->vdev_top->vdev_rz_expanding);
ta.trim_vdev = vd;
ta.trim_tree = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0);
ta.trim_type = TRIM_TYPE_SIMPLE;
ta.trim_extent_bytes_max = zfs_trim_extent_bytes_max;
ta.trim_extent_bytes_min = SPA_MINBLOCKSIZE;
ta.trim_flags = 0;
ASSERT3U(physical_rs.rs_end, >=, physical_rs.rs_start);
if (physical_rs.rs_end > physical_rs.rs_start) {
range_tree_add(ta.trim_tree, physical_rs.rs_start,
physical_rs.rs_end - physical_rs.rs_start);
} else {
ASSERT3U(physical_rs.rs_end, ==, physical_rs.rs_start);
}
error = vdev_trim_ranges(&ta);
mutex_enter(&vd->vdev_trim_io_lock);
while (vd->vdev_trim_inflight[TRIM_TYPE_SIMPLE] > 0) {
cv_wait(&vd->vdev_trim_io_cv, &vd->vdev_trim_io_lock);
}
mutex_exit(&vd->vdev_trim_io_lock);
range_tree_vacate(ta.trim_tree, NULL, NULL);
range_tree_destroy(ta.trim_tree);
return (error);
}
EXPORT_SYMBOL(vdev_trim);
EXPORT_SYMBOL(vdev_trim_stop);
EXPORT_SYMBOL(vdev_trim_stop_all);
EXPORT_SYMBOL(vdev_trim_stop_wait);
EXPORT_SYMBOL(vdev_trim_restart);
EXPORT_SYMBOL(vdev_autotrim);
EXPORT_SYMBOL(vdev_autotrim_stop_all);
EXPORT_SYMBOL(vdev_autotrim_stop_wait);
EXPORT_SYMBOL(vdev_autotrim_restart);
EXPORT_SYMBOL(vdev_trim_l2arc);
EXPORT_SYMBOL(vdev_trim_simple);
ZFS_MODULE_PARAM(zfs_trim, zfs_trim_, extent_bytes_max, UINT, ZMOD_RW,
"Max size of TRIM commands, larger will be split");
ZFS_MODULE_PARAM(zfs_trim, zfs_trim_, extent_bytes_min, UINT, ZMOD_RW,
"Min size of TRIM commands, smaller will be skipped");
ZFS_MODULE_PARAM(zfs_trim, zfs_trim_, metaslab_skip, UINT, ZMOD_RW,
"Skip metaslabs which have never been initialized");
ZFS_MODULE_PARAM(zfs_trim, zfs_trim_, txg_batch, UINT, ZMOD_RW,
"Min number of txgs to aggregate frees before issuing TRIM");
ZFS_MODULE_PARAM(zfs_trim, zfs_trim_, queue_limit, UINT, ZMOD_RW,
"Max queued TRIMs outstanding per leaf vdev");
diff --git a/sys/contrib/openzfs/scripts/zloop.sh b/sys/contrib/openzfs/scripts/zloop.sh
index 83160c34ab97..7cda23743382 100755
--- a/sys/contrib/openzfs/scripts/zloop.sh
+++ b/sys/contrib/openzfs/scripts/zloop.sh
@@ -1,334 +1,357 @@
#!/usr/bin/env bash
#
# CDDL HEADER START
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
# CDDL HEADER END
#
#
# Copyright (c) 2015 by Delphix. All rights reserved.
# Copyright (C) 2016 Lawrence Livermore National Security, LLC.
# Copyright (c) 2017, Intel Corporation.
#
BASE_DIR=${0%/*}
SCRIPT_COMMON=common.sh
if [[ -f "${BASE_DIR}/${SCRIPT_COMMON}" ]]; then
. "${BASE_DIR}/${SCRIPT_COMMON}"
else
echo "Missing helper script ${SCRIPT_COMMON}" && exit 1
fi
# shellcheck disable=SC2034
PROG=zloop.sh
GDB=${GDB:-gdb}
DEFAULTWORKDIR=/var/tmp
DEFAULTCOREDIR=/var/tmp/zloop
function usage
{
cat >&2 <<EOF
$0 [-hl] [-c <dump directory>] [-f <vdev directory>]
[-m <max core dumps>] [-s <vdev size>] [-t <timeout>]
[-I <max iterations>] [-- [extra ztest parameters]]
This script runs ztest repeatedly with randomized arguments.
If a crash is encountered, the ztest logs, any associated
vdev files, and core file (if one exists) are moved to the
output directory ($DEFAULTCOREDIR by default). Any options
after the -- end-of-options marker will be passed to ztest.
Options:
-c Specify a core dump directory to use.
-f Specify working directory for ztest vdev files.
-h Print this help message.
-l Create 'ztest.core.N' symlink to core directory.
-m Max number of core dumps to allow before exiting.
-s Size of vdev devices.
-t Total time to loop for, in seconds. If not provided,
zloop runs forever.
-I Max number of iterations to loop before exiting.
EOF
}
function or_die
{
if ! "$@"; then
echo "Command failed: $*"
exit 1
fi
}
case $(uname) in
FreeBSD)
coreglob="z*.core"
;;
Linux)
# core file helpers
read -r origcorepattern </proc/sys/kernel/core_pattern
coreglob="$(grep -E -o '^([^|%[:space:]]*)' /proc/sys/kernel/core_pattern)*"
if [[ $coreglob = "*" ]]; then
echo "Setting core file pattern..."
echo "core" > /proc/sys/kernel/core_pattern
coreglob="$(grep -E -o '^([^|%[:space:]]*)' \
/proc/sys/kernel/core_pattern)*"
fi
;;
*)
exit 1
;;
esac
function core_file
{
# shellcheck disable=SC2012,SC2086
ls -tr1 $coreglob 2>/dev/null | head -1
}
function core_prog
{
# shellcheck disable=SC2154
prog=$ZTEST
core_id=$($GDB --batch -c "$1" | grep "Core was generated by" | \
tr \' ' ')
if [[ "$core_id" == *"zdb "* ]]; then
# shellcheck disable=SC2154
prog=$ZDB
fi
printf "%s" "$prog"
}
function store_core
{
core="$(core_file)"
if [[ $ztrc -ne 0 ]] || [[ -f "$core" ]]; then
df -h "$workdir" >>ztest.out
coreid=$(date "+zloop-%y%m%d-%H%M%S")
foundcrashes=$((foundcrashes + 1))
# zdb debugging
zdbcmd="$ZDB -U "$workdir/zpool.cache" -dddMmDDG ztest"
zdbdebug=$($zdbcmd 2>&1)
echo -e "$zdbcmd\n" >>ztest.zdb
echo "$zdbdebug" >>ztest.zdb
dest=$coredir/$coreid
or_die mkdir -p "$dest/vdev"
if [[ $symlink -ne 0 ]]; then
or_die ln -sf "$dest" "ztest.core.${foundcrashes}"
fi
echo "*** ztest crash found - moving logs to $dest"
or_die mv ztest.history ztest.zdb ztest.out "$dest/"
or_die mv "$workdir/"ztest* "$dest/vdev/"
if [[ -e "$workdir/zpool.cache" ]]; then
or_die mv "$workdir/zpool.cache" "$dest/vdev/"
fi
# check for core
if [[ -f "$core" ]]; then
coreprog=$(core_prog "$core")
coredebug=$($GDB --batch --quiet \
-ex "set print thread-events off" \
-ex "printf \"*\n* Backtrace \n*\n\"" \
-ex "bt" \
-ex "printf \"*\n* Libraries \n*\n\"" \
-ex "info sharedlib" \
-ex "printf \"*\n* Threads (full) \n*\n\"" \
-ex "info threads" \
-ex "printf \"*\n* Backtraces \n*\n\"" \
-ex "thread apply all bt" \
-ex "printf \"*\n* Backtraces (full) \n*\n\"" \
-ex "thread apply all bt full" \
-ex "quit" "$coreprog" "$core" 2>&1 | \
grep -v "New LWP")
# Dump core + logs to stored directory
echo "$coredebug" >>"$dest/ztest.gdb"
or_die mv "$core" "$dest/"
# Record info in cores logfile
echo "*** core @ $coredir/$coreid/$core:" | \
tee -a ztest.cores
fi
if [[ $coremax -gt 0 ]] &&
[[ $foundcrashes -ge $coremax ]]; then
echo "exiting... max $coremax allowed cores"
exit 1
else
echo "continuing..."
fi
fi
}
# parse arguments
# expected format: zloop [-t timeout] [-c coredir] [-- extra ztest args]
coredir=$DEFAULTCOREDIR
basedir=$DEFAULTWORKDIR
rundir="zloop-run"
timeout=0
size="512m"
coremax=0
symlink=0
iterations=0
while getopts ":ht:m:I:s:c:f:l" opt; do
case $opt in
t ) [[ $OPTARG -gt 0 ]] && timeout=$OPTARG ;;
m ) [[ $OPTARG -gt 0 ]] && coremax=$OPTARG ;;
I ) [[ -n $OPTARG ]] && iterations=$OPTARG ;;
s ) [[ -n $OPTARG ]] && size=$OPTARG ;;
c ) [[ -n $OPTARG ]] && coredir=$OPTARG ;;
f ) [[ -n $OPTARG ]] && basedir=$(readlink -f "$OPTARG") ;;
l ) symlink=1 ;;
h ) usage
exit 2
;;
* ) echo "Invalid argument: -$OPTARG";
usage
exit 1
esac
done
# pass remaining arguments on to ztest
shift $((OPTIND - 1))
# enable core dumps
ulimit -c unlimited
export ASAN_OPTIONS=abort_on_error=true:halt_on_error=true:allocator_may_return_null=true:disable_coredump=false:detect_stack_use_after_return=true
export UBSAN_OPTIONS=abort_on_error=true:halt_on_error=true:print_stacktrace=true
if [[ -f "$(core_file)" ]]; then
echo -n "There's a core dump here you might want to look at first... "
core_file
echo
exit 1
fi
if [[ ! -d $coredir ]]; then
echo "core dump directory ($coredir) does not exist, creating it."
or_die mkdir -p "$coredir"
fi
if [[ ! -w $coredir ]]; then
echo "core dump directory ($coredir) is not writable."
exit 1
fi
or_die rm -f ztest.history ztest.zdb ztest.cores
ztrc=0 # ztest return value
foundcrashes=0 # number of crashes found so far
starttime=$(date +%s)
curtime=$starttime
iteration=0
# if no timeout was specified, loop forever.
while (( timeout == 0 )) || (( curtime <= (starttime + timeout) )); do
if (( iterations > 0 )) && (( iteration++ == iterations )); then
break
fi
zopt="-G -VVVVV"
# start each run with an empty directory
workdir="$basedir/$rundir"
or_die rm -rf "$workdir"
or_die mkdir "$workdir"
- # switch between three types of configs
- # 1/3 basic, 1/3 raidz mix, and 1/3 draid mix
- choice=$((RANDOM % 3))
-
# ashift range 9 - 15
align=$(((RANDOM % 2) * 3 + 9))
+ # choose parity value
+ parity=$(((RANDOM % 3) + 1))
+
+ draid_data=0
+ draid_spares=0
+
# randomly use special classes
class="special=random"
- if [[ $choice -eq 0 ]]; then
- # basic mirror only
- parity=1
+ # choose between four types of configs
+ # (basic, raidz mix, raidz expansion, and draid mix)
+ case $((RANDOM % 4)) in
+
+ # basic mirror configuration
+ 0) parity=1
mirrors=2
- draid_data=0
- draid_spares=0
raid_children=0
vdevs=2
raid_type="raidz"
- elif [[ $choice -eq 1 ]]; then
- # fully randomized mirror/raidz (sans dRAID)
- parity=$(((RANDOM % 3) + 1))
- mirrors=$(((RANDOM % 3) * 1))
- draid_data=0
- draid_spares=0
+ ;;
+
+ # fully randomized mirror/raidz (sans dRAID)
+ 1) mirrors=$(((RANDOM % 3) * 1))
raid_children=$((((RANDOM % 9) + parity + 1) * (RANDOM % 2)))
vdevs=$(((RANDOM % 3) + 3))
raid_type="raidz"
- else
- # fully randomized dRAID (sans mirror/raidz)
- parity=$(((RANDOM % 3) + 1))
- mirrors=0
+ ;;
+
+ # randomized raidz expansion (one top-level raidz vdev)
+ 2) mirrors=0
+ vdevs=1
+ # derive initial raidz disk count based on parity choice
+ # P1: 3 - 7 disks
+ # P2: 5 - 9 disks
+ # P3: 7 - 11 disks
+ raid_children=$(((RANDOM % 5) + (parity * 2) + 1))
+
+ # 1/3 of the time use a dedicated '-X' raidz expansion test
+ if [[ $((RANDOM % 3)) -eq 0 ]]; then
+ zopt="$zopt -X -t 16"
+ raid_type="raidz"
+ else
+ raid_type="eraidz"
+ fi
+ ;;
+
+ # fully randomized dRAID (sans mirror/raidz)
+ 3) mirrors=0
draid_data=$(((RANDOM % 8) + 3))
draid_spares=$(((RANDOM % 2) + parity))
stripe=$((draid_data + parity))
extra=$((draid_spares + (RANDOM % 4)))
raid_children=$(((((RANDOM % 4) + 1) * stripe) + extra))
vdevs=$((RANDOM % 3))
raid_type="draid"
- fi
+ ;;
+ *)
+ # avoid shellcheck SC2249
+ ;;
+ esac
zopt="$zopt -K $raid_type"
zopt="$zopt -m $mirrors"
zopt="$zopt -r $raid_children"
zopt="$zopt -D $draid_data"
zopt="$zopt -S $draid_spares"
zopt="$zopt -R $parity"
zopt="$zopt -v $vdevs"
zopt="$zopt -a $align"
zopt="$zopt -C $class"
zopt="$zopt -s $size"
zopt="$zopt -f $workdir"
cmd="$ZTEST $zopt $*"
echo "$(date '+%m/%d %T') $cmd" | tee -a ztest.history ztest.out
$cmd >>ztest.out 2>&1
ztrc=$?
grep -E '===|WARNING' ztest.out >>ztest.history
store_core
curtime=$(date +%s)
done
echo "zloop finished, $foundcrashes crashes found"
# restore core pattern.
case $(uname) in
Linux)
echo "$origcorepattern" > /proc/sys/kernel/core_pattern
;;
*)
;;
esac
uptime >>ztest.out
if [[ $foundcrashes -gt 0 ]]; then
exit 1
fi
diff --git a/sys/contrib/openzfs/tests/runfiles/common.run b/sys/contrib/openzfs/tests/runfiles/common.run
index ef787c65c0f9..20c9b823c648 100644
--- a/sys/contrib/openzfs/tests/runfiles/common.run
+++ b/sys/contrib/openzfs/tests/runfiles/common.run
@@ -1,983 +1,987 @@
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
# This run file contains all of the common functional tests. When
# adding a new test consider also adding it to the sanity.run file
# if the new test runs to completion in only a few seconds.
#
# Approximate run time: 4-5 hours
#
[DEFAULT]
pre = setup
quiet = False
pre_user = root
user = root
timeout = 600
post_user = root
post = cleanup
failsafe_user = root
failsafe = callbacks/zfs_failsafe
outputdir = /var/tmp/test_results
tags = ['functional']
[tests/functional/acl/off]
tests = ['dosmode', 'posixmode']
tags = ['functional', 'acl']
[tests/functional/alloc_class]
tests = ['alloc_class_001_pos', 'alloc_class_002_neg', 'alloc_class_003_pos',
'alloc_class_004_pos', 'alloc_class_005_pos', 'alloc_class_006_pos',
'alloc_class_007_pos', 'alloc_class_008_pos', 'alloc_class_009_pos',
'alloc_class_010_pos', 'alloc_class_011_neg', 'alloc_class_012_pos',
'alloc_class_013_pos', 'alloc_class_014_neg', 'alloc_class_015_pos']
tags = ['functional', 'alloc_class']
[tests/functional/append]
tests = ['file_append', 'threadsappend_001_pos']
tags = ['functional', 'append']
[tests/functional/arc]
tests = ['dbufstats_001_pos', 'dbufstats_002_pos', 'dbufstats_003_pos',
'arcstats_runtime_tuning']
tags = ['functional', 'arc']
[tests/functional/atime]
tests = ['atime_001_pos', 'atime_002_neg', 'root_atime_off', 'root_atime_on']
tags = ['functional', 'atime']
[tests/functional/bootfs]
tests = ['bootfs_001_pos', 'bootfs_002_neg', 'bootfs_003_pos',
'bootfs_004_neg', 'bootfs_005_neg', 'bootfs_006_pos', 'bootfs_007_pos',
'bootfs_008_pos']
tags = ['functional', 'bootfs']
[tests/functional/btree]
tests = ['btree_positive', 'btree_negative']
tags = ['functional', 'btree']
pre =
post =
[tests/functional/cache]
tests = ['cache_001_pos', 'cache_002_pos', 'cache_003_pos', 'cache_004_neg',
'cache_005_neg', 'cache_006_pos', 'cache_007_neg', 'cache_008_neg',
'cache_009_pos', 'cache_010_pos', 'cache_011_pos', 'cache_012_pos']
tags = ['functional', 'cache']
[tests/functional/cachefile]
tests = ['cachefile_001_pos', 'cachefile_002_pos', 'cachefile_003_pos',
'cachefile_004_pos']
tags = ['functional', 'cachefile']
[tests/functional/casenorm]
tests = ['case_all_values', 'norm_all_values', 'mixed_create_failure',
'sensitive_none_lookup', 'sensitive_none_delete',
'sensitive_formd_lookup', 'sensitive_formd_delete',
'insensitive_none_lookup', 'insensitive_none_delete',
'insensitive_formd_lookup', 'insensitive_formd_delete',
'mixed_none_lookup', 'mixed_none_lookup_ci', 'mixed_none_delete',
'mixed_formd_lookup', 'mixed_formd_lookup_ci', 'mixed_formd_delete']
tags = ['functional', 'casenorm']
[tests/functional/channel_program/lua_core]
tests = ['tst.args_to_lua', 'tst.divide_by_zero', 'tst.exists',
'tst.integer_illegal', 'tst.integer_overflow', 'tst.language_functions_neg',
'tst.language_functions_pos', 'tst.large_prog', 'tst.libraries',
'tst.memory_limit', 'tst.nested_neg', 'tst.nested_pos', 'tst.nvlist_to_lua',
'tst.recursive_neg', 'tst.recursive_pos', 'tst.return_large',
'tst.return_nvlist_neg', 'tst.return_nvlist_pos',
'tst.return_recursive_table', 'tst.stack_gsub', 'tst.timeout']
tags = ['functional', 'channel_program', 'lua_core']
[tests/functional/channel_program/synctask_core]
tests = ['tst.destroy_fs', 'tst.destroy_snap', 'tst.get_count_and_limit',
'tst.get_index_props', 'tst.get_mountpoint', 'tst.get_neg',
'tst.get_number_props', 'tst.get_string_props', 'tst.get_type',
'tst.get_userquota', 'tst.get_written', 'tst.inherit', 'tst.list_bookmarks',
'tst.list_children', 'tst.list_clones', 'tst.list_holds',
'tst.list_snapshots', 'tst.list_system_props',
'tst.list_user_props', 'tst.parse_args_neg','tst.promote_conflict',
'tst.promote_multiple', 'tst.promote_simple', 'tst.rollback_mult',
'tst.rollback_one', 'tst.set_props', 'tst.snapshot_destroy', 'tst.snapshot_neg',
'tst.snapshot_recursive', 'tst.snapshot_rename', 'tst.snapshot_simple',
'tst.bookmark.create', 'tst.bookmark.copy',
'tst.terminate_by_signal'
]
tags = ['functional', 'channel_program', 'synctask_core']
[tests/functional/checksum]
tests = ['run_edonr_test', 'run_sha2_test', 'run_skein_test', 'run_blake3_test',
'filetest_001_pos', 'filetest_002_pos']
tags = ['functional', 'checksum']
[tests/functional/clean_mirror]
tests = [ 'clean_mirror_001_pos', 'clean_mirror_002_pos',
'clean_mirror_003_pos', 'clean_mirror_004_pos']
tags = ['functional', 'clean_mirror']
[tests/functional/cli_root/zdb]
tests = ['zdb_002_pos', 'zdb_003_pos', 'zdb_004_pos', 'zdb_005_pos',
'zdb_006_pos', 'zdb_args_neg', 'zdb_args_pos',
'zdb_block_size_histogram', 'zdb_checksum', 'zdb_decompress',
'zdb_display_block', 'zdb_encrypted', 'zdb_label_checksum',
'zdb_object_range_neg', 'zdb_object_range_pos', 'zdb_objset_id',
'zdb_decompress_zstd', 'zdb_recover', 'zdb_recover_2', 'zdb_backup']
pre =
post =
tags = ['functional', 'cli_root', 'zdb']
[tests/functional/cli_root/zfs]
tests = ['zfs_001_neg', 'zfs_002_pos']
tags = ['functional', 'cli_root', 'zfs']
[tests/functional/cli_root/zfs_bookmark]
tests = ['zfs_bookmark_cliargs']
tags = ['functional', 'cli_root', 'zfs_bookmark']
[tests/functional/cli_root/zfs_change-key]
tests = ['zfs_change-key', 'zfs_change-key_child', 'zfs_change-key_format',
'zfs_change-key_inherit', 'zfs_change-key_load', 'zfs_change-key_location',
'zfs_change-key_pbkdf2iters', 'zfs_change-key_clones']
tags = ['functional', 'cli_root', 'zfs_change-key']
[tests/functional/cli_root/zfs_clone]
tests = ['zfs_clone_001_neg', 'zfs_clone_002_pos', 'zfs_clone_003_pos',
'zfs_clone_004_pos', 'zfs_clone_005_pos', 'zfs_clone_006_pos',
'zfs_clone_007_pos', 'zfs_clone_008_neg', 'zfs_clone_009_neg',
'zfs_clone_010_pos', 'zfs_clone_encrypted', 'zfs_clone_deeply_nested',
'zfs_clone_rm_nested']
tags = ['functional', 'cli_root', 'zfs_clone']
[tests/functional/cli_root/zfs_copies]
tests = ['zfs_copies_001_pos', 'zfs_copies_002_pos', 'zfs_copies_003_pos',
'zfs_copies_004_neg', 'zfs_copies_005_neg', 'zfs_copies_006_pos']
tags = ['functional', 'cli_root', 'zfs_copies']
[tests/functional/cli_root/zfs_create]
tests = ['zfs_create_001_pos', 'zfs_create_002_pos', 'zfs_create_003_pos',
'zfs_create_004_pos', 'zfs_create_005_pos', 'zfs_create_006_pos',
'zfs_create_007_pos', 'zfs_create_008_neg', 'zfs_create_009_neg',
'zfs_create_010_neg', 'zfs_create_011_pos', 'zfs_create_012_pos',
'zfs_create_013_pos', 'zfs_create_014_pos', 'zfs_create_encrypted',
'zfs_create_crypt_combos', 'zfs_create_dryrun', 'zfs_create_nomount',
'zfs_create_verbose']
tags = ['functional', 'cli_root', 'zfs_create']
[tests/functional/cli_root/zfs_destroy]
tests = ['zfs_clone_livelist_condense_and_disable',
'zfs_clone_livelist_condense_races', 'zfs_clone_livelist_dedup',
'zfs_destroy_001_pos', 'zfs_destroy_002_pos', 'zfs_destroy_003_pos',
'zfs_destroy_004_pos', 'zfs_destroy_005_neg', 'zfs_destroy_006_neg',
'zfs_destroy_007_neg', 'zfs_destroy_008_pos', 'zfs_destroy_009_pos',
'zfs_destroy_010_pos', 'zfs_destroy_011_pos', 'zfs_destroy_012_pos',
'zfs_destroy_013_neg', 'zfs_destroy_014_pos', 'zfs_destroy_015_pos',
'zfs_destroy_016_pos', 'zfs_destroy_clone_livelist',
'zfs_destroy_dev_removal', 'zfs_destroy_dev_removal_condense']
tags = ['functional', 'cli_root', 'zfs_destroy']
[tests/functional/cli_root/zfs_diff]
tests = ['zfs_diff_changes', 'zfs_diff_cliargs', 'zfs_diff_timestamp',
'zfs_diff_types', 'zfs_diff_encrypted', 'zfs_diff_mangle']
tags = ['functional', 'cli_root', 'zfs_diff']
[tests/functional/cli_root/zfs_get]
tests = ['zfs_get_001_pos', 'zfs_get_002_pos', 'zfs_get_003_pos',
'zfs_get_004_pos', 'zfs_get_005_neg', 'zfs_get_006_neg', 'zfs_get_007_neg',
'zfs_get_008_pos', 'zfs_get_009_pos', 'zfs_get_010_neg']
tags = ['functional', 'cli_root', 'zfs_get']
[tests/functional/cli_root/zfs_ids_to_path]
tests = ['zfs_ids_to_path_001_pos']
tags = ['functional', 'cli_root', 'zfs_ids_to_path']
[tests/functional/cli_root/zfs_inherit]
tests = ['zfs_inherit_001_neg', 'zfs_inherit_002_neg', 'zfs_inherit_003_pos',
'zfs_inherit_mountpoint']
tags = ['functional', 'cli_root', 'zfs_inherit']
[tests/functional/cli_root/zfs_load-key]
tests = ['zfs_load-key', 'zfs_load-key_all', 'zfs_load-key_file',
'zfs_load-key_https', 'zfs_load-key_location', 'zfs_load-key_noop',
'zfs_load-key_recursive']
tags = ['functional', 'cli_root', 'zfs_load-key']
[tests/functional/cli_root/zfs_mount]
tests = ['zfs_mount_001_pos', 'zfs_mount_002_pos', 'zfs_mount_003_pos',
'zfs_mount_004_pos', 'zfs_mount_005_pos', 'zfs_mount_007_pos',
'zfs_mount_009_neg', 'zfs_mount_010_neg', 'zfs_mount_011_neg',
'zfs_mount_012_pos', 'zfs_mount_all_001_pos', 'zfs_mount_encrypted',
'zfs_mount_remount', 'zfs_mount_all_fail', 'zfs_mount_all_mountpoints',
'zfs_mount_test_race']
tags = ['functional', 'cli_root', 'zfs_mount']
[tests/functional/cli_root/zfs_program]
tests = ['zfs_program_json']
tags = ['functional', 'cli_root', 'zfs_program']
[tests/functional/cli_root/zfs_promote]
tests = ['zfs_promote_001_pos', 'zfs_promote_002_pos', 'zfs_promote_003_pos',
'zfs_promote_004_pos', 'zfs_promote_005_pos', 'zfs_promote_006_neg',
'zfs_promote_007_neg', 'zfs_promote_008_pos', 'zfs_promote_encryptionroot']
tags = ['functional', 'cli_root', 'zfs_promote']
[tests/functional/cli_root/zfs_property]
tests = ['zfs_written_property_001_pos']
tags = ['functional', 'cli_root', 'zfs_property']
[tests/functional/cli_root/zfs_receive]
tests = ['zfs_receive_001_pos', 'zfs_receive_002_pos', 'zfs_receive_003_pos',
'zfs_receive_004_neg', 'zfs_receive_005_neg', 'zfs_receive_006_pos',
'zfs_receive_007_neg', 'zfs_receive_008_pos', 'zfs_receive_009_neg',
'zfs_receive_010_pos', 'zfs_receive_011_pos', 'zfs_receive_012_pos',
'zfs_receive_013_pos', 'zfs_receive_014_pos', 'zfs_receive_015_pos',
'zfs_receive_016_pos', 'receive-o-x_props_override',
'receive-o-x_props_aliases',
'zfs_receive_from_encrypted', 'zfs_receive_to_encrypted',
'zfs_receive_raw', 'zfs_receive_raw_incremental', 'zfs_receive_-e',
'zfs_receive_raw_-d', 'zfs_receive_from_zstd', 'zfs_receive_new_props',
'zfs_receive_-wR-encrypted-mix', 'zfs_receive_corrective',
'zfs_receive_compressed_corrective', 'zfs_receive_large_block_corrective']
tags = ['functional', 'cli_root', 'zfs_receive']
[tests/functional/cli_root/zfs_rename]
tests = ['zfs_rename_001_pos', 'zfs_rename_002_pos', 'zfs_rename_003_pos',
'zfs_rename_004_neg', 'zfs_rename_005_neg', 'zfs_rename_006_pos',
'zfs_rename_007_pos', 'zfs_rename_008_pos', 'zfs_rename_009_neg',
'zfs_rename_010_neg', 'zfs_rename_011_pos', 'zfs_rename_012_neg',
'zfs_rename_013_pos', 'zfs_rename_014_neg', 'zfs_rename_encrypted_child',
'zfs_rename_to_encrypted', 'zfs_rename_mountpoint', 'zfs_rename_nounmount']
tags = ['functional', 'cli_root', 'zfs_rename']
[tests/functional/cli_root/zfs_reservation]
tests = ['zfs_reservation_001_pos', 'zfs_reservation_002_pos']
tags = ['functional', 'cli_root', 'zfs_reservation']
[tests/functional/cli_root/zfs_rollback]
tests = ['zfs_rollback_001_pos', 'zfs_rollback_002_pos',
'zfs_rollback_003_neg', 'zfs_rollback_004_neg']
tags = ['functional', 'cli_root', 'zfs_rollback']
[tests/functional/cli_root/zfs_send]
tests = ['zfs_send_001_pos', 'zfs_send_002_pos', 'zfs_send_003_pos',
'zfs_send_004_neg', 'zfs_send_005_pos', 'zfs_send_006_pos',
'zfs_send_007_pos', 'zfs_send_encrypted', 'zfs_send_encrypted_unloaded',
'zfs_send_raw', 'zfs_send_sparse', 'zfs_send-b', 'zfs_send_skip_missing']
tags = ['functional', 'cli_root', 'zfs_send']
[tests/functional/cli_root/zfs_set]
tests = ['cache_001_pos', 'cache_002_neg', 'canmount_001_pos',
'canmount_002_pos', 'canmount_003_pos', 'canmount_004_pos',
'checksum_001_pos', 'compression_001_pos', 'mountpoint_001_pos',
'mountpoint_002_pos', 'reservation_001_neg', 'user_property_002_pos',
'share_mount_001_neg', 'snapdir_001_pos', 'onoffs_001_pos',
'user_property_001_pos', 'user_property_003_neg', 'readonly_001_pos',
'user_property_004_pos', 'version_001_neg', 'zfs_set_001_neg',
'zfs_set_002_neg', 'zfs_set_003_neg', 'property_alias_001_pos',
'mountpoint_003_pos', 'ro_props_001_pos', 'zfs_set_keylocation',
'zfs_set_feature_activation', 'zfs_set_nomount']
tags = ['functional', 'cli_root', 'zfs_set']
[tests/functional/cli_root/zfs_share]
tests = ['zfs_share_001_pos', 'zfs_share_002_pos', 'zfs_share_003_pos',
'zfs_share_004_pos', 'zfs_share_006_pos', 'zfs_share_008_neg',
'zfs_share_010_neg', 'zfs_share_011_pos', 'zfs_share_concurrent_shares']
tags = ['functional', 'cli_root', 'zfs_share']
[tests/functional/cli_root/zfs_snapshot]
tests = ['zfs_snapshot_001_neg', 'zfs_snapshot_002_neg',
'zfs_snapshot_003_neg', 'zfs_snapshot_004_neg', 'zfs_snapshot_005_neg',
'zfs_snapshot_006_pos', 'zfs_snapshot_007_neg', 'zfs_snapshot_008_neg',
'zfs_snapshot_009_pos']
tags = ['functional', 'cli_root', 'zfs_snapshot']
[tests/functional/cli_root/zfs_unload-key]
tests = ['zfs_unload-key', 'zfs_unload-key_all', 'zfs_unload-key_recursive']
tags = ['functional', 'cli_root', 'zfs_unload-key']
[tests/functional/cli_root/zfs_unmount]
tests = ['zfs_unmount_001_pos', 'zfs_unmount_002_pos', 'zfs_unmount_003_pos',
'zfs_unmount_004_pos', 'zfs_unmount_005_pos', 'zfs_unmount_006_pos',
'zfs_unmount_007_neg', 'zfs_unmount_008_neg', 'zfs_unmount_009_pos',
'zfs_unmount_all_001_pos', 'zfs_unmount_nested', 'zfs_unmount_unload_keys']
tags = ['functional', 'cli_root', 'zfs_unmount']
[tests/functional/cli_root/zfs_unshare]
tests = ['zfs_unshare_001_pos', 'zfs_unshare_002_pos', 'zfs_unshare_003_pos',
'zfs_unshare_004_neg', 'zfs_unshare_005_neg', 'zfs_unshare_006_pos',
'zfs_unshare_007_pos']
tags = ['functional', 'cli_root', 'zfs_unshare']
[tests/functional/cli_root/zfs_upgrade]
tests = ['zfs_upgrade_001_pos', 'zfs_upgrade_002_pos', 'zfs_upgrade_003_pos',
'zfs_upgrade_004_pos', 'zfs_upgrade_005_pos', 'zfs_upgrade_006_neg',
'zfs_upgrade_007_neg']
tags = ['functional', 'cli_root', 'zfs_upgrade']
[tests/functional/cli_root/zfs_wait]
tests = ['zfs_wait_deleteq', 'zfs_wait_getsubopt']
tags = ['functional', 'cli_root', 'zfs_wait']
[tests/functional/cli_root/zhack]
tests = ['zhack_label_repair_001', 'zhack_label_repair_002',
'zhack_label_repair_003', 'zhack_label_repair_004']
pre =
post =
tags = ['functional', 'cli_root', 'zhack']
[tests/functional/cli_root/zpool]
tests = ['zpool_001_neg', 'zpool_002_pos', 'zpool_003_pos', 'zpool_colors']
tags = ['functional', 'cli_root', 'zpool']
[tests/functional/cli_root/zpool_add]
tests = ['zpool_add_001_pos', 'zpool_add_002_pos', 'zpool_add_003_pos',
'zpool_add_004_pos', 'zpool_add_006_pos', 'zpool_add_007_neg',
'zpool_add_008_neg', 'zpool_add_009_neg', 'zpool_add_010_pos',
'add-o_ashift', 'add_prop_ashift', 'zpool_add_dryrun_output']
tags = ['functional', 'cli_root', 'zpool_add']
[tests/functional/cli_root/zpool_attach]
tests = ['zpool_attach_001_neg', 'attach-o_ashift']
tags = ['functional', 'cli_root', 'zpool_attach']
[tests/functional/cli_root/zpool_clear]
tests = ['zpool_clear_001_pos', 'zpool_clear_002_neg', 'zpool_clear_003_neg',
'zpool_clear_readonly']
tags = ['functional', 'cli_root', 'zpool_clear']
[tests/functional/cli_root/zpool_create]
tests = ['zpool_create_001_pos', 'zpool_create_002_pos',
'zpool_create_003_pos', 'zpool_create_004_pos', 'zpool_create_005_pos',
'zpool_create_006_pos', 'zpool_create_007_neg', 'zpool_create_008_pos',
'zpool_create_009_neg', 'zpool_create_010_neg', 'zpool_create_011_neg',
'zpool_create_012_neg', 'zpool_create_014_neg', 'zpool_create_015_neg',
'zpool_create_017_neg', 'zpool_create_018_pos', 'zpool_create_019_pos',
'zpool_create_020_pos', 'zpool_create_021_pos', 'zpool_create_022_pos',
'zpool_create_023_neg', 'zpool_create_024_pos',
'zpool_create_encrypted', 'zpool_create_crypt_combos',
'zpool_create_draid_001_pos', 'zpool_create_draid_002_pos',
'zpool_create_draid_003_pos', 'zpool_create_draid_004_pos',
'zpool_create_features_001_pos', 'zpool_create_features_002_pos',
'zpool_create_features_003_pos', 'zpool_create_features_004_neg',
'zpool_create_features_005_pos', 'zpool_create_features_006_pos',
'zpool_create_features_007_pos', 'zpool_create_features_008_pos',
'zpool_create_features_009_pos', 'create-o_ashift',
'zpool_create_tempname', 'zpool_create_dryrun_output']
tags = ['functional', 'cli_root', 'zpool_create']
[tests/functional/cli_root/zpool_destroy]
tests = ['zpool_destroy_001_pos', 'zpool_destroy_002_pos',
'zpool_destroy_003_neg']
pre =
post =
tags = ['functional', 'cli_root', 'zpool_destroy']
[tests/functional/cli_root/zpool_detach]
tests = ['zpool_detach_001_neg']
tags = ['functional', 'cli_root', 'zpool_detach']
[tests/functional/cli_root/zpool_events]
tests = ['zpool_events_clear', 'zpool_events_cliargs', 'zpool_events_follow',
'zpool_events_poolname', 'zpool_events_errors', 'zpool_events_duplicates',
'zpool_events_clear_retained']
tags = ['functional', 'cli_root', 'zpool_events']
[tests/functional/cli_root/zpool_export]
tests = ['zpool_export_001_pos', 'zpool_export_002_pos',
'zpool_export_003_neg', 'zpool_export_004_pos']
tags = ['functional', 'cli_root', 'zpool_export']
[tests/functional/cli_root/zpool_get]
tests = ['zpool_get_001_pos', 'zpool_get_002_pos', 'zpool_get_003_pos',
'zpool_get_004_neg', 'zpool_get_005_pos', 'vdev_get_001_pos']
tags = ['functional', 'cli_root', 'zpool_get']
[tests/functional/cli_root/zpool_history]
tests = ['zpool_history_001_neg', 'zpool_history_002_pos']
tags = ['functional', 'cli_root', 'zpool_history']
[tests/functional/cli_root/zpool_import]
tests = ['zpool_import_001_pos', 'zpool_import_002_pos',
'zpool_import_003_pos', 'zpool_import_004_pos', 'zpool_import_005_pos',
'zpool_import_006_pos', 'zpool_import_007_pos', 'zpool_import_008_pos',
'zpool_import_009_neg', 'zpool_import_010_pos', 'zpool_import_011_neg',
'zpool_import_012_pos', 'zpool_import_013_neg', 'zpool_import_014_pos',
'zpool_import_015_pos', 'zpool_import_016_pos', 'zpool_import_017_pos',
'zpool_import_features_001_pos', 'zpool_import_features_002_neg',
'zpool_import_features_003_pos', 'zpool_import_missing_001_pos',
'zpool_import_missing_002_pos', 'zpool_import_missing_003_pos',
'zpool_import_rename_001_pos', 'zpool_import_all_001_pos',
'zpool_import_encrypted', 'zpool_import_encrypted_load',
'zpool_import_errata3', 'zpool_import_errata4',
'import_cachefile_device_added',
'import_cachefile_device_removed',
'import_cachefile_device_replaced',
'import_cachefile_mirror_attached',
'import_cachefile_mirror_detached',
'import_cachefile_paths_changed',
'import_cachefile_shared_device',
'import_devices_missing', 'import_log_missing',
'import_paths_changed',
'import_rewind_config_changed',
'import_rewind_device_replaced']
tags = ['functional', 'cli_root', 'zpool_import']
timeout = 1200
[tests/functional/cli_root/zpool_labelclear]
tests = ['zpool_labelclear_active', 'zpool_labelclear_exported',
'zpool_labelclear_removed', 'zpool_labelclear_valid']
pre =
post =
tags = ['functional', 'cli_root', 'zpool_labelclear']
[tests/functional/cli_root/zpool_initialize]
tests = ['zpool_initialize_attach_detach_add_remove',
'zpool_initialize_fault_export_import_online',
'zpool_initialize_import_export',
'zpool_initialize_offline_export_import_online',
'zpool_initialize_online_offline',
'zpool_initialize_split',
'zpool_initialize_start_and_cancel_neg',
'zpool_initialize_start_and_cancel_pos',
'zpool_initialize_suspend_resume',
'zpool_initialize_uninit',
'zpool_initialize_unsupported_vdevs',
'zpool_initialize_verify_checksums',
'zpool_initialize_verify_initialized']
pre =
tags = ['functional', 'cli_root', 'zpool_initialize']
[tests/functional/cli_root/zpool_offline]
tests = ['zpool_offline_001_pos', 'zpool_offline_002_neg',
'zpool_offline_003_pos']
tags = ['functional', 'cli_root', 'zpool_offline']
[tests/functional/cli_root/zpool_online]
tests = ['zpool_online_001_pos', 'zpool_online_002_neg']
tags = ['functional', 'cli_root', 'zpool_online']
[tests/functional/cli_root/zpool_remove]
tests = ['zpool_remove_001_neg', 'zpool_remove_002_pos',
'zpool_remove_003_pos']
tags = ['functional', 'cli_root', 'zpool_remove']
[tests/functional/cli_root/zpool_replace]
tests = ['zpool_replace_001_neg', 'replace-o_ashift', 'replace_prop_ashift']
tags = ['functional', 'cli_root', 'zpool_replace']
[tests/functional/cli_root/zpool_resilver]
tests = ['zpool_resilver_bad_args', 'zpool_resilver_restart',
'zpool_resilver_concurrent']
tags = ['functional', 'cli_root', 'zpool_resilver']
[tests/functional/cli_root/zpool_scrub]
tests = ['zpool_scrub_001_neg', 'zpool_scrub_002_pos', 'zpool_scrub_003_pos',
'zpool_scrub_004_pos', 'zpool_scrub_005_pos',
'zpool_scrub_encrypted_unloaded', 'zpool_scrub_print_repairing',
'zpool_scrub_offline_device', 'zpool_scrub_multiple_copies',
'zpool_error_scrub_001_pos', 'zpool_error_scrub_002_pos',
'zpool_error_scrub_003_pos', 'zpool_error_scrub_004_pos']
tags = ['functional', 'cli_root', 'zpool_scrub']
[tests/functional/cli_root/zpool_set]
tests = ['zpool_set_001_pos', 'zpool_set_002_neg', 'zpool_set_003_neg',
'zpool_set_ashift', 'zpool_set_features', 'vdev_set_001_pos',
'user_property_001_pos', 'user_property_002_neg']
tags = ['functional', 'cli_root', 'zpool_set']
[tests/functional/cli_root/zpool_split]
tests = ['zpool_split_cliargs', 'zpool_split_devices',
'zpool_split_encryption', 'zpool_split_props', 'zpool_split_vdevs',
'zpool_split_resilver', 'zpool_split_indirect',
'zpool_split_dryrun_output']
tags = ['functional', 'cli_root', 'zpool_split']
[tests/functional/cli_root/zpool_status]
tests = ['zpool_status_001_pos', 'zpool_status_002_pos',
'zpool_status_003_pos', 'zpool_status_004_pos',
'zpool_status_005_pos', 'zpool_status_006_pos',
'zpool_status_007_pos', 'zpool_status_features_001_pos']
tags = ['functional', 'cli_root', 'zpool_status']
[tests/functional/cli_root/zpool_sync]
tests = ['zpool_sync_001_pos', 'zpool_sync_002_neg']
tags = ['functional', 'cli_root', 'zpool_sync']
[tests/functional/cli_root/zpool_trim]
tests = ['zpool_trim_attach_detach_add_remove',
'zpool_trim_fault_export_import_online',
'zpool_trim_import_export', 'zpool_trim_multiple', 'zpool_trim_neg',
'zpool_trim_offline_export_import_online', 'zpool_trim_online_offline',
'zpool_trim_partial', 'zpool_trim_rate', 'zpool_trim_rate_neg',
'zpool_trim_secure', 'zpool_trim_split', 'zpool_trim_start_and_cancel_neg',
'zpool_trim_start_and_cancel_pos', 'zpool_trim_suspend_resume',
'zpool_trim_unsupported_vdevs', 'zpool_trim_verify_checksums',
'zpool_trim_verify_trimmed']
tags = ['functional', 'zpool_trim']
[tests/functional/cli_root/zpool_upgrade]
tests = ['zpool_upgrade_001_pos', 'zpool_upgrade_002_pos',
'zpool_upgrade_003_pos', 'zpool_upgrade_004_pos',
'zpool_upgrade_005_neg', 'zpool_upgrade_006_neg',
'zpool_upgrade_007_pos', 'zpool_upgrade_008_pos',
'zpool_upgrade_009_neg', 'zpool_upgrade_features_001_pos']
tags = ['functional', 'cli_root', 'zpool_upgrade']
[tests/functional/cli_root/zpool_wait]
tests = ['zpool_wait_discard', 'zpool_wait_freeing',
'zpool_wait_initialize_basic', 'zpool_wait_initialize_cancel',
'zpool_wait_initialize_flag', 'zpool_wait_multiple',
'zpool_wait_no_activity', 'zpool_wait_remove', 'zpool_wait_remove_cancel',
'zpool_wait_trim_basic', 'zpool_wait_trim_cancel', 'zpool_wait_trim_flag',
'zpool_wait_usage']
tags = ['functional', 'cli_root', 'zpool_wait']
[tests/functional/cli_root/zpool_wait/scan]
tests = ['zpool_wait_replace_cancel', 'zpool_wait_rebuild',
'zpool_wait_resilver', 'zpool_wait_scrub_cancel',
'zpool_wait_replace', 'zpool_wait_scrub_basic', 'zpool_wait_scrub_flag']
tags = ['functional', 'cli_root', 'zpool_wait']
[tests/functional/cli_user/misc]
tests = ['zdb_001_neg', 'zfs_001_neg', 'zfs_allow_001_neg',
'zfs_clone_001_neg', 'zfs_create_001_neg', 'zfs_destroy_001_neg',
'zfs_get_001_neg', 'zfs_inherit_001_neg', 'zfs_mount_001_neg',
'zfs_promote_001_neg', 'zfs_receive_001_neg', 'zfs_rename_001_neg',
'zfs_rollback_001_neg', 'zfs_send_001_neg', 'zfs_set_001_neg',
'zfs_share_001_neg', 'zfs_snapshot_001_neg', 'zfs_unallow_001_neg',
'zfs_unmount_001_neg', 'zfs_unshare_001_neg', 'zfs_upgrade_001_neg',
'zpool_001_neg', 'zpool_add_001_neg', 'zpool_attach_001_neg',
'zpool_clear_001_neg', 'zpool_create_001_neg', 'zpool_destroy_001_neg',
'zpool_detach_001_neg', 'zpool_export_001_neg', 'zpool_get_001_neg',
'zpool_history_001_neg', 'zpool_import_001_neg', 'zpool_import_002_neg',
'zpool_offline_001_neg', 'zpool_online_001_neg', 'zpool_remove_001_neg',
'zpool_replace_001_neg', 'zpool_scrub_001_neg', 'zpool_set_001_neg',
'zpool_status_001_neg', 'zpool_upgrade_001_neg', 'arcstat_001_pos',
'arc_summary_001_pos', 'arc_summary_002_neg', 'zpool_wait_privilege',
'zilstat_001_pos']
user =
tags = ['functional', 'cli_user', 'misc']
[tests/functional/cli_user/zfs_list]
tests = ['zfs_list_001_pos', 'zfs_list_002_pos', 'zfs_list_003_pos',
'zfs_list_004_neg', 'zfs_list_005_neg', 'zfs_list_007_pos',
'zfs_list_008_neg']
user =
tags = ['functional', 'cli_user', 'zfs_list']
[tests/functional/cli_user/zpool_iostat]
tests = ['zpool_iostat_001_neg', 'zpool_iostat_002_pos',
'zpool_iostat_003_neg', 'zpool_iostat_004_pos',
'zpool_iostat_005_pos', 'zpool_iostat_-c_disable',
'zpool_iostat_-c_homedir', 'zpool_iostat_-c_searchpath']
user =
tags = ['functional', 'cli_user', 'zpool_iostat']
[tests/functional/cli_user/zpool_list]
tests = ['zpool_list_001_pos', 'zpool_list_002_neg']
user =
tags = ['functional', 'cli_user', 'zpool_list']
[tests/functional/cli_user/zpool_status]
tests = ['zpool_status_003_pos', 'zpool_status_-c_disable',
'zpool_status_-c_homedir', 'zpool_status_-c_searchpath']
user =
tags = ['functional', 'cli_user', 'zpool_status']
[tests/functional/compression]
tests = ['compress_001_pos', 'compress_002_pos', 'compress_003_pos',
'l2arc_compressed_arc', 'l2arc_compressed_arc_disabled',
'l2arc_encrypted', 'l2arc_encrypted_no_compressed_arc']
tags = ['functional', 'compression']
[tests/functional/cp_files]
tests = ['cp_files_001_pos']
tags = ['functional', 'cp_files']
[tests/functional/crtime]
tests = ['crtime_001_pos' ]
tags = ['functional', 'crtime']
[tests/functional/ctime]
tests = ['ctime_001_pos' ]
tags = ['functional', 'ctime']
[tests/functional/deadman]
tests = ['deadman_ratelimit', 'deadman_sync', 'deadman_zio']
pre =
post =
tags = ['functional', 'deadman']
[tests/functional/delegate]
tests = ['zfs_allow_001_pos', 'zfs_allow_002_pos', 'zfs_allow_003_pos',
'zfs_allow_004_pos', 'zfs_allow_005_pos', 'zfs_allow_006_pos',
'zfs_allow_007_pos', 'zfs_allow_008_pos', 'zfs_allow_009_neg',
'zfs_allow_010_pos', 'zfs_allow_011_neg', 'zfs_allow_012_neg',
'zfs_unallow_001_pos', 'zfs_unallow_002_pos', 'zfs_unallow_003_pos',
'zfs_unallow_004_pos', 'zfs_unallow_005_pos', 'zfs_unallow_006_pos',
'zfs_unallow_007_neg', 'zfs_unallow_008_neg']
tags = ['functional', 'delegate']
[tests/functional/exec]
tests = ['exec_001_pos', 'exec_002_neg']
tags = ['functional', 'exec']
[tests/functional/fallocate]
tests = ['fallocate_punch-hole']
tags = ['functional', 'fallocate']
[tests/functional/features/async_destroy]
tests = ['async_destroy_001_pos']
tags = ['functional', 'features', 'async_destroy']
[tests/functional/features/large_dnode]
tests = ['large_dnode_001_pos', 'large_dnode_003_pos', 'large_dnode_004_neg',
'large_dnode_005_pos', 'large_dnode_007_neg', 'large_dnode_009_pos']
tags = ['functional', 'features', 'large_dnode']
[tests/functional/grow]
pre =
post =
tests = ['grow_pool_001_pos', 'grow_replicas_001_pos']
tags = ['functional', 'grow']
[tests/functional/history]
tests = ['history_001_pos', 'history_002_pos', 'history_003_pos',
'history_004_pos', 'history_005_neg', 'history_006_neg',
'history_007_pos', 'history_008_pos', 'history_009_pos',
'history_010_pos']
tags = ['functional', 'history']
[tests/functional/hkdf]
pre =
post =
tests = ['hkdf_test']
tags = ['functional', 'hkdf']
[tests/functional/inheritance]
tests = ['inherit_001_pos']
pre =
tags = ['functional', 'inheritance']
[tests/functional/io]
tests = ['sync', 'psync', 'posixaio', 'mmap']
tags = ['functional', 'io']
[tests/functional/inuse]
tests = ['inuse_004_pos', 'inuse_005_pos', 'inuse_008_pos', 'inuse_009_pos']
post =
tags = ['functional', 'inuse']
[tests/functional/large_files]
tests = ['large_files_001_pos', 'large_files_002_pos']
tags = ['functional', 'large_files']
[tests/functional/limits]
tests = ['filesystem_count', 'filesystem_limit', 'snapshot_count',
'snapshot_limit']
tags = ['functional', 'limits']
[tests/functional/link_count]
tests = ['link_count_001', 'link_count_root_inode']
tags = ['functional', 'link_count']
[tests/functional/migration]
tests = ['migration_001_pos', 'migration_002_pos', 'migration_003_pos',
'migration_004_pos', 'migration_005_pos', 'migration_006_pos',
'migration_007_pos', 'migration_008_pos', 'migration_009_pos',
'migration_010_pos', 'migration_011_pos', 'migration_012_pos']
tags = ['functional', 'migration']
[tests/functional/mmap]
tests = ['mmap_mixed', 'mmap_read_001_pos', 'mmap_seek_001_pos',
'mmap_sync_001_pos', 'mmap_write_001_pos']
tags = ['functional', 'mmap']
[tests/functional/mount]
tests = ['umount_001', 'umountall_001']
tags = ['functional', 'mount']
[tests/functional/mv_files]
tests = ['mv_files_001_pos', 'mv_files_002_pos', 'random_creation']
tags = ['functional', 'mv_files']
[tests/functional/nestedfs]
tests = ['nestedfs_001_pos']
tags = ['functional', 'nestedfs']
[tests/functional/no_space]
tests = ['enospc_001_pos', 'enospc_002_pos', 'enospc_003_pos',
'enospc_df', 'enospc_ganging', 'enospc_rm']
tags = ['functional', 'no_space']
[tests/functional/nopwrite]
tests = ['nopwrite_copies', 'nopwrite_mtime', 'nopwrite_negative',
'nopwrite_promoted_clone', 'nopwrite_recsize', 'nopwrite_sync',
'nopwrite_varying_compression', 'nopwrite_volume']
tags = ['functional', 'nopwrite']
[tests/functional/online_offline]
tests = ['online_offline_001_pos', 'online_offline_002_neg',
'online_offline_003_neg']
tags = ['functional', 'online_offline']
[tests/functional/pool_checkpoint]
tests = ['checkpoint_after_rewind', 'checkpoint_big_rewind',
'checkpoint_capacity', 'checkpoint_conf_change', 'checkpoint_discard',
'checkpoint_discard_busy', 'checkpoint_discard_many',
'checkpoint_indirect', 'checkpoint_invalid', 'checkpoint_lun_expsz',
'checkpoint_open', 'checkpoint_removal', 'checkpoint_rewind',
'checkpoint_ro_rewind', 'checkpoint_sm_scale', 'checkpoint_twice',
'checkpoint_vdev_add', 'checkpoint_zdb', 'checkpoint_zhack_feat']
tags = ['functional', 'pool_checkpoint']
timeout = 1800
[tests/functional/pool_names]
tests = ['pool_names_001_pos', 'pool_names_002_neg']
pre =
post =
tags = ['functional', 'pool_names']
[tests/functional/poolversion]
tests = ['poolversion_001_pos', 'poolversion_002_pos']
tags = ['functional', 'poolversion']
[tests/functional/pyzfs]
tests = ['pyzfs_unittest']
pre =
post =
tags = ['functional', 'pyzfs']
[tests/functional/quota]
tests = ['quota_001_pos', 'quota_002_pos', 'quota_003_pos',
'quota_004_pos', 'quota_005_pos', 'quota_006_neg']
tags = ['functional', 'quota']
[tests/functional/redacted_send]
tests = ['redacted_compressed', 'redacted_contents', 'redacted_deleted',
'redacted_disabled_feature', 'redacted_embedded', 'redacted_holes',
'redacted_incrementals', 'redacted_largeblocks', 'redacted_many_clones',
'redacted_mixed_recsize', 'redacted_mounts', 'redacted_negative',
'redacted_origin', 'redacted_panic', 'redacted_props', 'redacted_resume',
'redacted_size', 'redacted_volume']
tags = ['functional', 'redacted_send']
[tests/functional/raidz]
-tests = ['raidz_001_neg', 'raidz_002_pos', 'raidz_003_pos', 'raidz_004_pos']
+tests = ['raidz_001_neg', 'raidz_002_pos', 'raidz_expand_001_pos',
+ 'raidz_expand_002_pos', 'raidz_expand_003_neg', 'raidz_expand_003_pos',
+ 'raidz_expand_004_pos', 'raidz_expand_005_pos', 'raidz_expand_006_neg',
+ 'raidz_expand_007_neg']
tags = ['functional', 'raidz']
+timeout = 1200
[tests/functional/redundancy]
tests = ['redundancy_draid', 'redundancy_draid1', 'redundancy_draid2',
'redundancy_draid3', 'redundancy_draid_damaged1',
'redundancy_draid_damaged2', 'redundancy_draid_spare1',
'redundancy_draid_spare2', 'redundancy_draid_spare3', 'redundancy_mirror',
'redundancy_raidz', 'redundancy_raidz1', 'redundancy_raidz2',
'redundancy_raidz3', 'redundancy_stripe']
tags = ['functional', 'redundancy']
timeout = 1200
[tests/functional/refquota]
tests = ['refquota_001_pos', 'refquota_002_pos', 'refquota_003_pos',
'refquota_004_pos', 'refquota_005_pos', 'refquota_006_neg',
'refquota_007_neg', 'refquota_008_neg']
tags = ['functional', 'refquota']
[tests/functional/refreserv]
tests = ['refreserv_001_pos', 'refreserv_002_pos', 'refreserv_003_pos',
'refreserv_004_pos', 'refreserv_005_pos', 'refreserv_multi_raidz',
'refreserv_raidz']
tags = ['functional', 'refreserv']
[tests/functional/removal]
pre =
tests = ['removal_all_vdev', 'removal_cancel', 'removal_check_space',
'removal_condense_export', 'removal_multiple_indirection',
'removal_nopwrite', 'removal_remap_deadlists',
'removal_resume_export', 'removal_sanity', 'removal_with_add',
'removal_with_create_fs', 'removal_with_dedup',
'removal_with_errors', 'removal_with_export', 'removal_with_indirect',
'removal_with_ganging', 'removal_with_faulted',
'removal_with_remove', 'removal_with_scrub', 'removal_with_send',
'removal_with_send_recv', 'removal_with_snapshot',
'removal_with_write', 'removal_with_zdb', 'remove_expanded',
'remove_mirror', 'remove_mirror_sanity', 'remove_raidz',
'remove_indirect', 'remove_attach_mirror', 'removal_reservation']
tags = ['functional', 'removal']
[tests/functional/rename_dirs]
tests = ['rename_dirs_001_pos']
tags = ['functional', 'rename_dirs']
[tests/functional/replacement]
tests = ['attach_import', 'attach_multiple', 'attach_rebuild',
'attach_resilver', 'detach', 'rebuild_disabled_feature',
'rebuild_multiple', 'rebuild_raidz', 'replace_import', 'replace_rebuild',
'replace_resilver', 'resilver_restart_001', 'resilver_restart_002',
'scrub_cancel']
tags = ['functional', 'replacement']
[tests/functional/reservation]
tests = ['reservation_001_pos', 'reservation_002_pos', 'reservation_003_pos',
'reservation_004_pos', 'reservation_005_pos', 'reservation_006_pos',
'reservation_007_pos', 'reservation_008_pos', 'reservation_009_pos',
'reservation_010_pos', 'reservation_011_pos', 'reservation_012_pos',
'reservation_013_pos', 'reservation_014_pos', 'reservation_015_pos',
'reservation_016_pos', 'reservation_017_pos', 'reservation_018_pos',
'reservation_019_pos', 'reservation_020_pos', 'reservation_021_neg',
'reservation_022_pos']
tags = ['functional', 'reservation']
[tests/functional/rootpool]
tests = ['rootpool_002_neg', 'rootpool_003_neg', 'rootpool_007_pos']
tags = ['functional', 'rootpool']
[tests/functional/rsend]
tests = ['recv_dedup', 'recv_dedup_encrypted_zvol', 'rsend_001_pos',
'rsend_002_pos', 'rsend_003_pos', 'rsend_004_pos', 'rsend_005_pos',
'rsend_006_pos', 'rsend_007_pos', 'rsend_008_pos', 'rsend_009_pos',
'rsend_010_pos', 'rsend_011_pos', 'rsend_012_pos', 'rsend_013_pos',
'rsend_014_pos', 'rsend_016_neg', 'rsend_019_pos', 'rsend_020_pos',
'rsend_021_pos', 'rsend_022_pos', 'rsend_024_pos', 'rsend_025_pos',
'rsend_026_neg', 'rsend_027_pos', 'rsend_028_neg', 'rsend_029_neg',
'rsend_030_pos', 'rsend_031_pos', 'send-c_verify_ratio',
'send-c_verify_contents', 'send-c_props', 'send-c_incremental',
'send-c_volume', 'send-c_zstream_recompress', 'send-c_zstreamdump',
'send-c_lz4_disabled', 'send-c_recv_lz4_disabled',
'send-c_mixed_compression', 'send-c_stream_size_estimate',
'send-c_embedded_blocks', 'send-c_resume', 'send-cpL_varied_recsize',
'send-c_recv_dedup', 'send-L_toggle', 'send_encrypted_incremental',
'send_encrypted_freeobjects', 'send_encrypted_hierarchy',
'send_encrypted_props', 'send_encrypted_truncated_files',
'send_freeobjects', 'send_realloc_files', 'send_realloc_encrypted_files',
'send_spill_block', 'send_holds', 'send_hole_birth', 'send_mixed_raw',
'send-wR_encrypted_zvol', 'send_partial_dataset', 'send_invalid',
'send_doall', 'send_raw_spill_block', 'send_raw_ashift',
'send_raw_large_blocks']
tags = ['functional', 'rsend']
[tests/functional/scrub_mirror]
tests = ['scrub_mirror_001_pos', 'scrub_mirror_002_pos',
'scrub_mirror_003_pos', 'scrub_mirror_004_pos']
tags = ['functional', 'scrub_mirror']
[tests/functional/slog]
tests = ['slog_001_pos', 'slog_002_pos', 'slog_003_pos', 'slog_004_pos',
'slog_005_pos', 'slog_006_pos', 'slog_007_pos', 'slog_008_neg',
'slog_009_neg', 'slog_010_neg', 'slog_011_neg', 'slog_012_neg',
'slog_013_pos', 'slog_014_pos', 'slog_015_neg', 'slog_replay_fs_001',
'slog_replay_fs_002', 'slog_replay_volume', 'slog_016_pos']
tags = ['functional', 'slog']
[tests/functional/snapshot]
tests = ['clone_001_pos', 'rollback_001_pos', 'rollback_002_pos',
'rollback_003_pos', 'snapshot_001_pos', 'snapshot_002_pos',
'snapshot_003_pos', 'snapshot_004_pos', 'snapshot_005_pos',
'snapshot_006_pos', 'snapshot_007_pos', 'snapshot_008_pos',
'snapshot_009_pos', 'snapshot_010_pos', 'snapshot_011_pos',
'snapshot_012_pos', 'snapshot_013_pos', 'snapshot_014_pos',
'snapshot_017_pos', 'snapshot_018_pos']
tags = ['functional', 'snapshot']
[tests/functional/snapused]
tests = ['snapused_001_pos', 'snapused_002_pos', 'snapused_003_pos',
'snapused_004_pos', 'snapused_005_pos']
tags = ['functional', 'snapused']
[tests/functional/sparse]
tests = ['sparse_001_pos']
tags = ['functional', 'sparse']
[tests/functional/stat]
tests = ['stat_001_pos']
tags = ['functional', 'stat']
[tests/functional/suid]
tests = ['suid_write_to_suid', 'suid_write_to_sgid', 'suid_write_to_suid_sgid',
'suid_write_to_none', 'suid_write_zil_replay']
tags = ['functional', 'suid']
[tests/functional/trim]
tests = ['autotrim_integrity', 'autotrim_config', 'autotrim_trim_integrity',
'trim_integrity', 'trim_config', 'trim_l2arc']
tags = ['functional', 'trim']
[tests/functional/truncate]
tests = ['truncate_001_pos', 'truncate_002_pos', 'truncate_timestamps']
tags = ['functional', 'truncate']
[tests/functional/upgrade]
tests = ['upgrade_userobj_001_pos', 'upgrade_readonly_pool']
tags = ['functional', 'upgrade']
[tests/functional/userquota]
tests = [
'userquota_001_pos', 'userquota_002_pos', 'userquota_003_pos',
'userquota_004_pos', 'userquota_005_neg', 'userquota_006_pos',
'userquota_007_pos', 'userquota_008_pos', 'userquota_009_pos',
'userquota_010_pos', 'userquota_011_pos', 'userquota_012_neg',
'userspace_001_pos', 'userspace_002_pos', 'userspace_encrypted',
'userspace_send_encrypted', 'userspace_encrypted_13709']
tags = ['functional', 'userquota']
[tests/functional/vdev_zaps]
tests = ['vdev_zaps_001_pos', 'vdev_zaps_002_pos', 'vdev_zaps_003_pos',
'vdev_zaps_004_pos', 'vdev_zaps_005_pos', 'vdev_zaps_006_pos',
'vdev_zaps_007_pos']
tags = ['functional', 'vdev_zaps']
[tests/functional/write_dirs]
tests = ['write_dirs_001_pos', 'write_dirs_002_pos']
tags = ['functional', 'write_dirs']
[tests/functional/xattr]
tests = ['xattr_001_pos', 'xattr_002_neg', 'xattr_003_neg', 'xattr_004_pos',
'xattr_005_pos', 'xattr_006_pos', 'xattr_007_neg',
'xattr_011_pos', 'xattr_012_pos', 'xattr_013_pos', 'xattr_compat']
tags = ['functional', 'xattr']
[tests/functional/zvol/zvol_ENOSPC]
tests = ['zvol_ENOSPC_001_pos']
tags = ['functional', 'zvol', 'zvol_ENOSPC']
[tests/functional/zvol/zvol_cli]
tests = ['zvol_cli_001_pos', 'zvol_cli_002_pos', 'zvol_cli_003_neg']
tags = ['functional', 'zvol', 'zvol_cli']
[tests/functional/zvol/zvol_misc]
tests = ['zvol_misc_002_pos', 'zvol_misc_hierarchy', 'zvol_misc_rename_inuse',
'zvol_misc_snapdev', 'zvol_misc_trim', 'zvol_misc_volmode', 'zvol_misc_zil']
tags = ['functional', 'zvol', 'zvol_misc']
[tests/functional/zvol/zvol_stress]
tests = ['zvol_stress']
tags = ['functional', 'zvol', 'zvol_stress']
[tests/functional/zvol/zvol_swap]
tests = ['zvol_swap_001_pos', 'zvol_swap_002_pos', 'zvol_swap_004_pos']
tags = ['functional', 'zvol', 'zvol_swap']
[tests/functional/libzfs]
tests = ['many_fds', 'libzfs_input']
tags = ['functional', 'libzfs']
[tests/functional/log_spacemap]
tests = ['log_spacemap_import_logs']
pre =
post =
tags = ['functional', 'log_spacemap']
[tests/functional/l2arc]
tests = ['l2arc_arcstats_pos', 'l2arc_mfuonly_pos', 'l2arc_l2miss_pos',
'persist_l2arc_001_pos', 'persist_l2arc_002_pos',
'persist_l2arc_003_neg', 'persist_l2arc_004_pos', 'persist_l2arc_005_pos']
tags = ['functional', 'l2arc']
[tests/functional/zpool_influxdb]
tests = ['zpool_influxdb']
tags = ['functional', 'zpool_influxdb']
diff --git a/sys/contrib/openzfs/tests/zfs-tests/include/tunables.cfg b/sys/contrib/openzfs/tests/zfs-tests/include/tunables.cfg
index 80e7bcb3bd09..fb861f1a285c 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/include/tunables.cfg
+++ b/sys/contrib/openzfs/tests/zfs-tests/include/tunables.cfg
@@ -1,101 +1,104 @@
# This file exports variables for each tunable used in the test suite.
#
# Different platforms use different names for most tunables. To avoid littering
# the tests with conditional logic for deciding how to set each tunable, the
# logic is instead consolidated to this one file.
#
# Any use of tunables in tests must use a name defined here. New entries
# should be added to the table as needed. Please keep the table sorted
# alphabetically for ease of maintenance.
#
# Platform-specific tunables should still use a NAME from this table for
# consistency. Enter UNSUPPORTED in the column for platforms on which the
# tunable is not implemented.
UNAME=$(uname)
# NAME FreeBSD tunable Linux tunable
cat <<%%%% |
ADMIN_SNAPSHOT UNSUPPORTED zfs_admin_snapshot
ALLOW_REDACTED_DATASET_MOUNT allow_redacted_dataset_mount zfs_allow_redacted_dataset_mount
ARC_MAX arc.max zfs_arc_max
ARC_MIN arc.min zfs_arc_min
ASYNC_BLOCK_MAX_BLOCKS async_block_max_blocks zfs_async_block_max_blocks
CHECKSUM_EVENTS_PER_SECOND checksum_events_per_second zfs_checksum_events_per_second
COMMIT_TIMEOUT_PCT commit_timeout_pct zfs_commit_timeout_pct
COMPRESSED_ARC_ENABLED compressed_arc_enabled zfs_compressed_arc_enabled
CONDENSE_INDIRECT_COMMIT_ENTRY_DELAY_MS condense.indirect_commit_entry_delay_ms zfs_condense_indirect_commit_entry_delay_ms
CONDENSE_INDIRECT_OBSOLETE_PCT condense.indirect_obsolete_pct zfs_condense_indirect_obsolete_pct
CONDENSE_MIN_MAPPING_BYTES condense.min_mapping_bytes zfs_condense_min_mapping_bytes
DBUF_CACHE_SHIFT dbuf.cache_shift dbuf_cache_shift
DEADMAN_CHECKTIME_MS deadman.checktime_ms zfs_deadman_checktime_ms
DEADMAN_FAILMODE deadman.failmode zfs_deadman_failmode
DEADMAN_SYNCTIME_MS deadman.synctime_ms zfs_deadman_synctime_ms
DEADMAN_ZIOTIME_MS deadman.ziotime_ms zfs_deadman_ziotime_ms
DISABLE_IVSET_GUID_CHECK disable_ivset_guid_check zfs_disable_ivset_guid_check
DMU_OFFSET_NEXT_SYNC dmu_offset_next_sync zfs_dmu_offset_next_sync
+EMBEDDED_SLOG_MIN_MS embedded_slog_min_ms zfs_embedded_slog_min_ms
INITIALIZE_CHUNK_SIZE initialize_chunk_size zfs_initialize_chunk_size
INITIALIZE_VALUE initialize_value zfs_initialize_value
KEEP_LOG_SPACEMAPS_AT_EXPORT keep_log_spacemaps_at_export zfs_keep_log_spacemaps_at_export
LUA_MAX_MEMLIMIT lua.max_memlimit zfs_lua_max_memlimit
L2ARC_MFUONLY l2arc.mfuonly l2arc_mfuonly
L2ARC_NOPREFETCH l2arc.noprefetch l2arc_noprefetch
L2ARC_REBUILD_BLOCKS_MIN_L2SIZE l2arc.rebuild_blocks_min_l2size l2arc_rebuild_blocks_min_l2size
L2ARC_REBUILD_ENABLED l2arc.rebuild_enabled l2arc_rebuild_enabled
L2ARC_TRIM_AHEAD l2arc.trim_ahead l2arc_trim_ahead
L2ARC_WRITE_BOOST l2arc.write_boost l2arc_write_boost
L2ARC_WRITE_MAX l2arc.write_max l2arc_write_max
LIVELIST_CONDENSE_NEW_ALLOC livelist.condense.new_alloc zfs_livelist_condense_new_alloc
LIVELIST_CONDENSE_SYNC_CANCEL livelist.condense.sync_cancel zfs_livelist_condense_sync_cancel
LIVELIST_CONDENSE_SYNC_PAUSE livelist.condense.sync_pause zfs_livelist_condense_sync_pause
LIVELIST_CONDENSE_ZTHR_CANCEL livelist.condense.zthr_cancel zfs_livelist_condense_zthr_cancel
LIVELIST_CONDENSE_ZTHR_PAUSE livelist.condense.zthr_pause zfs_livelist_condense_zthr_pause
LIVELIST_MAX_ENTRIES livelist.max_entries zfs_livelist_max_entries
LIVELIST_MIN_PERCENT_SHARED livelist.min_percent_shared zfs_livelist_min_percent_shared
MAX_DATASET_NESTING max_dataset_nesting zfs_max_dataset_nesting
MAX_MISSING_TVDS max_missing_tvds zfs_max_missing_tvds
METASLAB_DEBUG_LOAD metaslab.debug_load metaslab_debug_load
METASLAB_FORCE_GANGING metaslab.force_ganging metaslab_force_ganging
MULTIHOST_FAIL_INTERVALS multihost.fail_intervals zfs_multihost_fail_intervals
MULTIHOST_HISTORY multihost.history zfs_multihost_history
MULTIHOST_IMPORT_INTERVALS multihost.import_intervals zfs_multihost_import_intervals
MULTIHOST_INTERVAL multihost.interval zfs_multihost_interval
OVERRIDE_ESTIMATE_RECORDSIZE send.override_estimate_recordsize zfs_override_estimate_recordsize
PREFETCH_DISABLE prefetch.disable zfs_prefetch_disable
+RAIDZ_EXPAND_MAX_REFLOW_BYTES vdev.expand_max_reflow_bytes raidz_expand_max_reflow_bytes
REBUILD_SCRUB_ENABLED rebuild_scrub_enabled zfs_rebuild_scrub_enabled
REMOVAL_SUSPEND_PROGRESS removal_suspend_progress zfs_removal_suspend_progress
REMOVE_MAX_SEGMENT remove_max_segment zfs_remove_max_segment
RESILVER_MIN_TIME_MS resilver_min_time_ms zfs_resilver_min_time_ms
SCAN_LEGACY scan_legacy zfs_scan_legacy
SCAN_SUSPEND_PROGRESS scan_suspend_progress zfs_scan_suspend_progress
SCAN_VDEV_LIMIT scan_vdev_limit zfs_scan_vdev_limit
+SCRUB_AFTER_EXPAND scrub_after_expand zfs_scrub_after_expand
SEND_HOLES_WITHOUT_BIRTH_TIME send_holes_without_birth_time send_holes_without_birth_time
SLOW_IO_EVENTS_PER_SECOND slow_io_events_per_second zfs_slow_io_events_per_second
SPA_ASIZE_INFLATION spa.asize_inflation spa_asize_inflation
SPA_DISCARD_MEMORY_LIMIT spa.discard_memory_limit zfs_spa_discard_memory_limit
SPA_LOAD_VERIFY_DATA spa.load_verify_data spa_load_verify_data
SPA_LOAD_VERIFY_METADATA spa.load_verify_metadata spa_load_verify_metadata
TRIM_EXTENT_BYTES_MIN trim.extent_bytes_min zfs_trim_extent_bytes_min
TRIM_METASLAB_SKIP trim.metaslab_skip zfs_trim_metaslab_skip
TRIM_TXG_BATCH trim.txg_batch zfs_trim_txg_batch
TXG_HISTORY txg.history zfs_txg_history
TXG_TIMEOUT txg.timeout zfs_txg_timeout
UNLINK_SUSPEND_PROGRESS UNSUPPORTED zfs_unlink_suspend_progress
VDEV_FILE_LOGICAL_ASHIFT vdev.file.logical_ashift vdev_file_logical_ashift
VDEV_FILE_PHYSICAL_ASHIFT vdev.file.physical_ashift vdev_file_physical_ashift
VDEV_MAX_AUTO_ASHIFT vdev.max_auto_ashift zfs_vdev_max_auto_ashift
VDEV_MIN_MS_COUNT vdev.min_ms_count zfs_vdev_min_ms_count
VDEV_VALIDATE_SKIP vdev.validate_skip vdev_validate_skip
VOL_INHIBIT_DEV UNSUPPORTED zvol_inhibit_dev
VOL_MODE vol.mode zvol_volmode
VOL_RECURSIVE vol.recursive UNSUPPORTED
VOL_USE_BLK_MQ UNSUPPORTED zvol_use_blk_mq
XATTR_COMPAT xattr_compat zfs_xattr_compat
ZEVENT_LEN_MAX zevent.len_max zfs_zevent_len_max
ZEVENT_RETAIN_MAX zevent.retain_max zfs_zevent_retain_max
ZIO_SLOW_IO_MS zio.slow_io_ms zio_slow_io_ms
ZIL_SAXATTR zil_saxattr zfs_zil_saxattr
%%%%
while read name FreeBSD Linux; do
eval "export ${name}=\$${UNAME}"
done
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am b/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am
index 87b50f59ca7a..79aef1900b59 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am
@@ -1,2074 +1,2080 @@
CLEANFILES =
dist_noinst_DATA =
include $(top_srcdir)/config/Substfiles.am
datadir_zfs_tests_testsdir = $(datadir)/$(PACKAGE)/zfs-tests/tests
nobase_dist_datadir_zfs_tests_tests_DATA = \
perf/nfs-sample.cfg \
perf/perf.shlib \
\
perf/fio/mkfiles.fio \
perf/fio/random_reads.fio \
perf/fio/random_readwrite.fio \
perf/fio/random_readwrite_fixed.fio \
perf/fio/random_writes.fio \
perf/fio/sequential_reads.fio \
perf/fio/sequential_readwrite.fio \
perf/fio/sequential_writes.fio
nobase_dist_datadir_zfs_tests_tests_SCRIPTS = \
perf/regression/random_reads.ksh \
perf/regression/random_readwrite.ksh \
perf/regression/random_readwrite_fixed.ksh \
perf/regression/random_writes.ksh \
perf/regression/random_writes_zil.ksh \
perf/regression/sequential_reads_arc_cached_clone.ksh \
perf/regression/sequential_reads_arc_cached.ksh \
perf/regression/sequential_reads_dbuf_cached.ksh \
perf/regression/sequential_reads.ksh \
perf/regression/sequential_writes.ksh \
perf/regression/setup.ksh \
\
perf/scripts/prefetch_io.sh
# These lists can be regenerated by running make regen-tests at the root, or, on a *clean* source:
# find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' ! -executable -name '*.in' | sort | sed 's/\.in$//;s/^/\t/;$!s/$/ \\/'
# find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' -executable -name '*.in' | sort | sed 's/\.in$//;s/^/\t/;$!s/$/ \\/'
# find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' ! -name '*.in' ! -name '*.c' | grep -Fe /simd -e /tmpfile | sort | sed 's/^/\t/;$!s/$/ \\/'
# find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' ! -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$!s/$/ \\/'
# find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$!s/$/ \\/'
#
# simd and tmpfile are Linux-only and not installed elsewhere
#
# C programs are specced in ../Makefile.am above as part of the main Makefile
find_common := find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po'
regen:
@$(MAKE) -C $(top_builddir) clean
@$(MAKE) clean
$(SED) $(ac_inplace) '/^# -- >8 --/q' Makefile.am
echo >> Makefile.am
echo 'nobase_nodist_datadir_zfs_tests_tests_DATA = \' >> Makefile.am
$(find_common) ! -executable -name '*.in' | sort | sed 's/\.in$$//;s/^/\t/;$$!s/$$/ \\/' >> Makefile.am
echo 'nobase_nodist_datadir_zfs_tests_tests_SCRIPTS = \' >> Makefile.am
$(find_common) -executable -name '*.in' | sort | sed 's/\.in$$//;s/^/\t/;$$!s/$$/ \\/' >> Makefile.am
echo >> Makefile.am
echo 'SUBSTFILES += $$(nobase_nodist_datadir_zfs_tests_tests_DATA) $$(nobase_nodist_datadir_zfs_tests_tests_SCRIPTS)' >> Makefile.am
echo >> Makefile.am
echo 'if BUILD_LINUX' >> Makefile.am
echo 'nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \' >> Makefile.am
$(find_common) ! -name '*.in' ! -name '*.c' | grep -Fe /simd -e /tmpfile | sort | sed 's/^/\t/;$$!s/$$/ \\/' >> Makefile.am
echo 'endif' >> Makefile.am
echo >> Makefile.am
echo 'nobase_dist_datadir_zfs_tests_tests_DATA += \' >> Makefile.am
$(find_common) ! -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$$!s/$$/ \\/' >> Makefile.am
echo >> Makefile.am
echo 'nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \' >> Makefile.am
$(find_common) -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$$!s/$$/ \\/' >> Makefile.am
# -- >8 --
nobase_nodist_datadir_zfs_tests_tests_DATA = \
functional/pam/utilities.kshlib
nobase_nodist_datadir_zfs_tests_tests_SCRIPTS = \
functional/pyzfs/pyzfs_unittest.ksh
SUBSTFILES += $(nobase_nodist_datadir_zfs_tests_tests_DATA) $(nobase_nodist_datadir_zfs_tests_tests_SCRIPTS)
if BUILD_LINUX
nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \
functional/simd/simd_supported.ksh \
functional/tmpfile/cleanup.ksh \
functional/tmpfile/setup.ksh
endif
nobase_dist_datadir_zfs_tests_tests_DATA += \
functional/acl/acl.cfg \
functional/acl/acl_common.kshlib \
functional/alloc_class/alloc_class.cfg \
functional/alloc_class/alloc_class.kshlib \
functional/atime/atime.cfg \
functional/atime/atime_common.kshlib \
functional/block_cloning/block_cloning.kshlib \
functional/cache/cache.cfg \
functional/cache/cache.kshlib \
functional/cachefile/cachefile.cfg \
functional/cachefile/cachefile.kshlib \
functional/casenorm/casenorm.cfg \
functional/casenorm/casenorm.kshlib \
functional/channel_program/channel_common.kshlib \
functional/channel_program/lua_core/tst.args_to_lua.out \
functional/channel_program/lua_core/tst.args_to_lua.zcp \
functional/channel_program/lua_core/tst.divide_by_zero.err \
functional/channel_program/lua_core/tst.divide_by_zero.zcp \
functional/channel_program/lua_core/tst.exists.zcp \
functional/channel_program/lua_core/tst.large_prog.out \
functional/channel_program/lua_core/tst.large_prog.zcp \
functional/channel_program/lua_core/tst.lib_base.lua \
functional/channel_program/lua_core/tst.lib_coroutine.lua \
functional/channel_program/lua_core/tst.lib_strings.lua \
functional/channel_program/lua_core/tst.lib_table.lua \
functional/channel_program/lua_core/tst.nested_neg.zcp \
functional/channel_program/lua_core/tst.nested_pos.zcp \
functional/channel_program/lua_core/tst.recursive.zcp \
functional/channel_program/lua_core/tst.return_large.zcp \
functional/channel_program/lua_core/tst.return_recursive_table.zcp \
functional/channel_program/lua_core/tst.stack_gsub.err \
functional/channel_program/lua_core/tst.stack_gsub.zcp \
functional/channel_program/lua_core/tst.timeout.zcp \
functional/channel_program/synctask_core/tst.bookmark.copy.zcp \
functional/channel_program/synctask_core/tst.bookmark.create.zcp \
functional/channel_program/synctask_core/tst.get_index_props.out \
functional/channel_program/synctask_core/tst.get_index_props.zcp \
functional/channel_program/synctask_core/tst.get_number_props.out \
functional/channel_program/synctask_core/tst.get_number_props.zcp \
functional/channel_program/synctask_core/tst.get_string_props.out \
functional/channel_program/synctask_core/tst.get_string_props.zcp \
functional/channel_program/synctask_core/tst.promote_conflict.zcp \
functional/channel_program/synctask_core/tst.set_props.zcp \
functional/channel_program/synctask_core/tst.snapshot_destroy.zcp \
functional/channel_program/synctask_core/tst.snapshot_neg.zcp \
functional/channel_program/synctask_core/tst.snapshot_recursive.zcp \
functional/channel_program/synctask_core/tst.snapshot_rename.zcp \
functional/channel_program/synctask_core/tst.snapshot_simple.zcp \
functional/checksum/default.cfg \
functional/clean_mirror/clean_mirror_common.kshlib \
functional/clean_mirror/default.cfg \
functional/cli_root/cli_common.kshlib \
functional/cli_root/zfs_copies/zfs_copies.cfg \
functional/cli_root/zfs_copies/zfs_copies.kshlib \
functional/cli_root/zfs_create/properties.kshlib \
functional/cli_root/zfs_create/zfs_create.cfg \
functional/cli_root/zfs_create/zfs_create_common.kshlib \
functional/cli_root/zfs_destroy/zfs_destroy.cfg \
functional/cli_root/zfs_destroy/zfs_destroy_common.kshlib \
functional/cli_root/zfs_get/zfs_get_common.kshlib \
functional/cli_root/zfs_get/zfs_get_list_d.kshlib \
functional/cli_root/zfs_jail/jail.conf \
functional/cli_root/zfs_load-key/HEXKEY \
functional/cli_root/zfs_load-key/PASSPHRASE \
functional/cli_root/zfs_load-key/RAWKEY \
functional/cli_root/zfs_load-key/zfs_load-key.cfg \
functional/cli_root/zfs_load-key/zfs_load-key_common.kshlib \
functional/cli_root/zfs_mount/zfs_mount.cfg \
functional/cli_root/zfs_mount/zfs_mount.kshlib \
functional/cli_root/zfs_promote/zfs_promote.cfg \
functional/cli_root/zfs_receive/zstd_test_data.txt \
functional/cli_root/zfs_rename/zfs_rename.cfg \
functional/cli_root/zfs_rename/zfs_rename.kshlib \
functional/cli_root/zfs_rollback/zfs_rollback.cfg \
functional/cli_root/zfs_rollback/zfs_rollback_common.kshlib \
functional/cli_root/zfs_send/zfs_send.cfg \
functional/cli_root/zfs_set/zfs_set_common.kshlib \
functional/cli_root/zfs_share/zfs_share.cfg \
functional/cli_root/zfs_snapshot/zfs_snapshot.cfg \
functional/cli_root/zfs_unmount/zfs_unmount.cfg \
functional/cli_root/zfs_unmount/zfs_unmount.kshlib \
functional/cli_root/zfs_upgrade/zfs_upgrade.kshlib \
functional/cli_root/zfs_wait/zfs_wait.kshlib \
functional/cli_root/zpool_add/zpool_add.cfg \
functional/cli_root/zpool_add/zpool_add.kshlib \
functional/cli_root/zpool_clear/zpool_clear.cfg \
functional/cli_root/zpool_create/draidcfg.gz \
functional/cli_root/zpool_create/zpool_create.cfg \
functional/cli_root/zpool_create/zpool_create.shlib \
functional/cli_root/zpool_destroy/zpool_destroy.cfg \
functional/cli_root/zpool_events/zpool_events.cfg \
functional/cli_root/zpool_events/zpool_events.kshlib \
functional/cli_root/zpool_expand/zpool_expand.cfg \
functional/cli_root/zpool_export/zpool_export.cfg \
functional/cli_root/zpool_export/zpool_export.kshlib \
functional/cli_root/zpool_get/vdev_get.cfg \
functional/cli_root/zpool_get/zpool_get.cfg \
functional/cli_root/zpool_get/zpool_get_parsable.cfg \
functional/cli_root/zpool_import/blockfiles/cryptv0.dat.bz2 \
functional/cli_root/zpool_import/blockfiles/missing_ivset.dat.bz2 \
functional/cli_root/zpool_import/blockfiles/unclean_export.dat.bz2 \
functional/cli_root/zpool_import/zpool_import.cfg \
functional/cli_root/zpool_import/zpool_import.kshlib \
functional/cli_root/zpool_initialize/zpool_initialize.kshlib \
functional/cli_root/zpool_labelclear/labelclear.cfg \
functional/cli_root/zpool_remove/zpool_remove.cfg \
functional/cli_root/zpool_reopen/zpool_reopen.cfg \
functional/cli_root/zpool_reopen/zpool_reopen.shlib \
functional/cli_root/zpool_resilver/zpool_resilver.cfg \
functional/cli_root/zpool_scrub/zpool_scrub.cfg \
functional/cli_root/zpool_split/zpool_split.cfg \
functional/cli_root/zpool_trim/zpool_trim.kshlib \
functional/cli_root/zpool_upgrade/blockfiles/zfs-broken-mirror1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-broken-mirror2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v10.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v11.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v12.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v13.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v14.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v15.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1mirror1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1mirror2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1mirror3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1raidz1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1raidz2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1raidz3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1stripe1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1stripe2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1stripe3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2mirror1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2mirror2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2mirror3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2raidz1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2raidz2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2raidz3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2stripe1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2stripe2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2stripe3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3hotspare1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3hotspare2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3hotspare3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3mirror1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3mirror2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3mirror3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz21.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz22.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz23.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3stripe1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3stripe2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3stripe3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v4.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v5.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v6.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v7.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v8.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v999.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v9.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-vBROKEN.dat.bz2 \
functional/cli_root/zpool_upgrade/zpool_upgrade.cfg \
functional/cli_root/zpool_upgrade/zpool_upgrade.kshlib \
functional/cli_root/zpool_wait/zpool_wait.kshlib \
functional/cli_root/zhack/library.kshlib \
functional/cli_user/misc/misc.cfg \
functional/cli_user/zfs_list/zfs_list.cfg \
functional/cli_user/zfs_list/zfs_list.kshlib \
functional/compression/compress.cfg \
functional/compression/testpool_zstd.tar.gz \
functional/deadman/deadman.cfg \
functional/delegate/delegate.cfg \
functional/delegate/delegate_common.kshlib \
functional/devices/devices.cfg \
functional/devices/devices_common.kshlib \
functional/events/events.cfg \
functional/events/events_common.kshlib \
functional/fault/fault.cfg \
functional/grow/grow.cfg \
functional/history/history.cfg \
functional/history/history_common.kshlib \
functional/history/i386.migratedpool.DAT.Z \
functional/history/i386.orig_history.txt \
functional/history/sparc.migratedpool.DAT.Z \
functional/history/sparc.orig_history.txt \
functional/history/zfs-pool-v4.dat.Z \
functional/inheritance/config001.cfg \
functional/inheritance/config002.cfg \
functional/inheritance/config003.cfg \
functional/inheritance/config004.cfg \
functional/inheritance/config005.cfg \
functional/inheritance/config006.cfg \
functional/inheritance/config007.cfg \
functional/inheritance/config008.cfg \
functional/inheritance/config009.cfg \
functional/inheritance/config010.cfg \
functional/inheritance/config011.cfg \
functional/inheritance/config012.cfg \
functional/inheritance/config013.cfg \
functional/inheritance/config014.cfg \
functional/inheritance/config015.cfg \
functional/inheritance/config016.cfg \
functional/inheritance/config017.cfg \
functional/inheritance/config018.cfg \
functional/inheritance/config019.cfg \
functional/inheritance/config020.cfg \
functional/inheritance/config021.cfg \
functional/inheritance/config022.cfg \
functional/inheritance/config023.cfg \
functional/inheritance/config024.cfg \
functional/inheritance/inherit.kshlib \
functional/inheritance/README.config \
functional/inheritance/README.state \
functional/inheritance/state001.cfg \
functional/inheritance/state002.cfg \
functional/inheritance/state003.cfg \
functional/inheritance/state004.cfg \
functional/inheritance/state005.cfg \
functional/inheritance/state006.cfg \
functional/inheritance/state007.cfg \
functional/inheritance/state008.cfg \
functional/inheritance/state009.cfg \
functional/inheritance/state010.cfg \
functional/inheritance/state011.cfg \
functional/inheritance/state012.cfg \
functional/inheritance/state013.cfg \
functional/inheritance/state014.cfg \
functional/inheritance/state015.cfg \
functional/inheritance/state016.cfg \
functional/inheritance/state017.cfg \
functional/inheritance/state018.cfg \
functional/inheritance/state019.cfg \
functional/inheritance/state020.cfg \
functional/inheritance/state021.cfg \
functional/inheritance/state022.cfg \
functional/inheritance/state023.cfg \
functional/inheritance/state024.cfg \
functional/inuse/inuse.cfg \
functional/io/io.cfg \
functional/l2arc/l2arc.cfg \
functional/largest_pool/largest_pool.cfg \
functional/migration/migration.cfg \
functional/migration/migration.kshlib \
functional/mmap/mmap.cfg \
functional/mmp/mmp.cfg \
functional/mmp/mmp.kshlib \
functional/mv_files/mv_files.cfg \
functional/mv_files/mv_files_common.kshlib \
functional/nopwrite/nopwrite.shlib \
functional/no_space/enospc.cfg \
functional/online_offline/online_offline.cfg \
functional/pool_checkpoint/pool_checkpoint.kshlib \
functional/projectquota/projectquota.cfg \
functional/projectquota/projectquota_common.kshlib \
functional/quota/quota.cfg \
functional/quota/quota.kshlib \
functional/redacted_send/redacted.cfg \
functional/redacted_send/redacted.kshlib \
functional/redundancy/redundancy.cfg \
functional/redundancy/redundancy.kshlib \
functional/refreserv/refreserv.cfg \
functional/removal/removal.kshlib \
functional/replacement/replacement.cfg \
functional/reservation/reservation.cfg \
functional/reservation/reservation.shlib \
functional/rsend/dedup_encrypted_zvol.bz2 \
functional/rsend/dedup_encrypted_zvol.zsend.bz2 \
functional/rsend/dedup.zsend.bz2 \
functional/rsend/fs.tar.gz \
functional/rsend/rsend.cfg \
functional/rsend/rsend.kshlib \
functional/scrub_mirror/default.cfg \
functional/scrub_mirror/scrub_mirror_common.kshlib \
functional/slog/slog.cfg \
functional/slog/slog.kshlib \
functional/snapshot/snapshot.cfg \
functional/snapused/snapused.kshlib \
functional/sparse/sparse.cfg \
functional/trim/trim.cfg \
functional/trim/trim.kshlib \
functional/truncate/truncate.cfg \
functional/upgrade/upgrade_common.kshlib \
functional/user_namespace/user_namespace.cfg \
functional/user_namespace/user_namespace_common.kshlib \
functional/userquota/13709_reproducer.bz2 \
functional/userquota/userquota.cfg \
functional/userquota/userquota_common.kshlib \
functional/vdev_zaps/vdev_zaps.kshlib \
functional/xattr/xattr.cfg \
functional/xattr/xattr_common.kshlib \
functional/zvol/zvol.cfg \
functional/zvol/zvol_cli/zvol_cli.cfg \
functional/zvol/zvol_common.shlib \
functional/zvol/zvol_ENOSPC/zvol_ENOSPC.cfg \
functional/zvol/zvol_misc/zvol_misc_common.kshlib \
functional/zvol/zvol_swap/zvol_swap.cfg \
functional/idmap_mount/idmap_mount.cfg \
functional/idmap_mount/idmap_mount_common.kshlib
nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \
functional/acl/off/cleanup.ksh \
functional/acl/off/dosmode.ksh \
functional/acl/off/posixmode.ksh \
functional/acl/off/setup.ksh \
functional/acl/posix/cleanup.ksh \
functional/acl/posix/posix_001_pos.ksh \
functional/acl/posix/posix_002_pos.ksh \
functional/acl/posix/posix_003_pos.ksh \
functional/acl/posix/posix_004_pos.ksh \
functional/acl/posix-sa/cleanup.ksh \
functional/acl/posix-sa/posix_001_pos.ksh \
functional/acl/posix-sa/posix_002_pos.ksh \
functional/acl/posix-sa/posix_003_pos.ksh \
functional/acl/posix-sa/posix_004_pos.ksh \
functional/acl/posix-sa/setup.ksh \
functional/acl/posix/setup.ksh \
functional/alloc_class/alloc_class_001_pos.ksh \
functional/alloc_class/alloc_class_002_neg.ksh \
functional/alloc_class/alloc_class_003_pos.ksh \
functional/alloc_class/alloc_class_004_pos.ksh \
functional/alloc_class/alloc_class_005_pos.ksh \
functional/alloc_class/alloc_class_006_pos.ksh \
functional/alloc_class/alloc_class_007_pos.ksh \
functional/alloc_class/alloc_class_008_pos.ksh \
functional/alloc_class/alloc_class_009_pos.ksh \
functional/alloc_class/alloc_class_010_pos.ksh \
functional/alloc_class/alloc_class_011_neg.ksh \
functional/alloc_class/alloc_class_012_pos.ksh \
functional/alloc_class/alloc_class_013_pos.ksh \
functional/alloc_class/alloc_class_014_neg.ksh \
functional/alloc_class/alloc_class_015_pos.ksh \
functional/alloc_class/cleanup.ksh \
functional/alloc_class/setup.ksh \
functional/append/file_append.ksh \
functional/append/threadsappend_001_pos.ksh \
functional/append/cleanup.ksh \
functional/append/setup.ksh \
functional/arc/arcstats_runtime_tuning.ksh \
functional/arc/cleanup.ksh \
functional/arc/dbufstats_001_pos.ksh \
functional/arc/dbufstats_002_pos.ksh \
functional/arc/dbufstats_003_pos.ksh \
functional/arc/setup.ksh \
functional/atime/atime_001_pos.ksh \
functional/atime/atime_002_neg.ksh \
functional/atime/atime_003_pos.ksh \
functional/atime/cleanup.ksh \
functional/atime/root_atime_off.ksh \
functional/atime/root_atime_on.ksh \
functional/atime/root_relatime_on.ksh \
functional/atime/setup.ksh \
functional/block_cloning/cleanup.ksh \
functional/block_cloning/setup.ksh \
functional/block_cloning/block_cloning_copyfilerange_cross_dataset.ksh \
functional/block_cloning/block_cloning_copyfilerange_fallback.ksh \
functional/block_cloning/block_cloning_copyfilerange_fallback_same_txg.ksh \
functional/block_cloning/block_cloning_copyfilerange.ksh \
functional/block_cloning/block_cloning_copyfilerange_partial.ksh \
functional/block_cloning/block_cloning_disabled_copyfilerange.ksh \
functional/block_cloning/block_cloning_disabled_ficlone.ksh \
functional/block_cloning/block_cloning_disabled_ficlonerange.ksh \
functional/block_cloning/block_cloning_ficlone.ksh \
functional/block_cloning/block_cloning_ficlonerange.ksh \
functional/block_cloning/block_cloning_ficlonerange_partial.ksh \
functional/bootfs/bootfs_001_pos.ksh \
functional/bootfs/bootfs_002_neg.ksh \
functional/bootfs/bootfs_003_pos.ksh \
functional/bootfs/bootfs_004_neg.ksh \
functional/bootfs/bootfs_005_neg.ksh \
functional/bootfs/bootfs_006_pos.ksh \
functional/bootfs/bootfs_007_pos.ksh \
functional/bootfs/bootfs_008_pos.ksh \
functional/bootfs/cleanup.ksh \
functional/bootfs/setup.ksh \
functional/btree/btree_negative.ksh \
functional/btree/btree_positive.ksh \
functional/cache/cache_001_pos.ksh \
functional/cache/cache_002_pos.ksh \
functional/cache/cache_003_pos.ksh \
functional/cache/cache_004_neg.ksh \
functional/cache/cache_005_neg.ksh \
functional/cache/cache_006_pos.ksh \
functional/cache/cache_007_neg.ksh \
functional/cache/cache_008_neg.ksh \
functional/cache/cache_009_pos.ksh \
functional/cache/cache_010_pos.ksh \
functional/cache/cache_011_pos.ksh \
functional/cache/cache_012_pos.ksh \
functional/cache/cleanup.ksh \
functional/cachefile/cachefile_001_pos.ksh \
functional/cachefile/cachefile_002_pos.ksh \
functional/cachefile/cachefile_003_pos.ksh \
functional/cachefile/cachefile_004_pos.ksh \
functional/cachefile/cleanup.ksh \
functional/cachefile/setup.ksh \
functional/cache/setup.ksh \
functional/casenorm/case_all_values.ksh \
functional/casenorm/cleanup.ksh \
functional/casenorm/insensitive_formd_delete.ksh \
functional/casenorm/insensitive_formd_lookup.ksh \
functional/casenorm/insensitive_none_delete.ksh \
functional/casenorm/insensitive_none_lookup.ksh \
functional/casenorm/mixed_create_failure.ksh \
functional/casenorm/mixed_formd_delete.ksh \
functional/casenorm/mixed_formd_lookup_ci.ksh \
functional/casenorm/mixed_formd_lookup.ksh \
functional/casenorm/mixed_none_delete.ksh \
functional/casenorm/mixed_none_lookup_ci.ksh \
functional/casenorm/mixed_none_lookup.ksh \
functional/casenorm/norm_all_values.ksh \
functional/casenorm/sensitive_formd_delete.ksh \
functional/casenorm/sensitive_formd_lookup.ksh \
functional/casenorm/sensitive_none_delete.ksh \
functional/casenorm/sensitive_none_lookup.ksh \
functional/casenorm/setup.ksh \
functional/channel_program/lua_core/cleanup.ksh \
functional/channel_program/lua_core/setup.ksh \
functional/channel_program/lua_core/tst.args_to_lua.ksh \
functional/channel_program/lua_core/tst.divide_by_zero.ksh \
functional/channel_program/lua_core/tst.exists.ksh \
functional/channel_program/lua_core/tst.integer_illegal.ksh \
functional/channel_program/lua_core/tst.integer_overflow.ksh \
functional/channel_program/lua_core/tst.language_functions_neg.ksh \
functional/channel_program/lua_core/tst.language_functions_pos.ksh \
functional/channel_program/lua_core/tst.large_prog.ksh \
functional/channel_program/lua_core/tst.libraries.ksh \
functional/channel_program/lua_core/tst.memory_limit.ksh \
functional/channel_program/lua_core/tst.nested_neg.ksh \
functional/channel_program/lua_core/tst.nested_pos.ksh \
functional/channel_program/lua_core/tst.nvlist_to_lua.ksh \
functional/channel_program/lua_core/tst.recursive_neg.ksh \
functional/channel_program/lua_core/tst.recursive_pos.ksh \
functional/channel_program/lua_core/tst.return_large.ksh \
functional/channel_program/lua_core/tst.return_nvlist_neg.ksh \
functional/channel_program/lua_core/tst.return_nvlist_pos.ksh \
functional/channel_program/lua_core/tst.return_recursive_table.ksh \
functional/channel_program/lua_core/tst.stack_gsub.ksh \
functional/channel_program/lua_core/tst.timeout.ksh \
functional/channel_program/synctask_core/cleanup.ksh \
functional/channel_program/synctask_core/setup.ksh \
functional/channel_program/synctask_core/tst.bookmark.copy.ksh \
functional/channel_program/synctask_core/tst.bookmark.create.ksh \
functional/channel_program/synctask_core/tst.destroy_fs.ksh \
functional/channel_program/synctask_core/tst.destroy_snap.ksh \
functional/channel_program/synctask_core/tst.get_count_and_limit.ksh \
functional/channel_program/synctask_core/tst.get_index_props.ksh \
functional/channel_program/synctask_core/tst.get_mountpoint.ksh \
functional/channel_program/synctask_core/tst.get_neg.ksh \
functional/channel_program/synctask_core/tst.get_number_props.ksh \
functional/channel_program/synctask_core/tst.get_string_props.ksh \
functional/channel_program/synctask_core/tst.get_type.ksh \
functional/channel_program/synctask_core/tst.get_userquota.ksh \
functional/channel_program/synctask_core/tst.get_written.ksh \
functional/channel_program/synctask_core/tst.inherit.ksh \
functional/channel_program/synctask_core/tst.list_bookmarks.ksh \
functional/channel_program/synctask_core/tst.list_children.ksh \
functional/channel_program/synctask_core/tst.list_clones.ksh \
functional/channel_program/synctask_core/tst.list_holds.ksh \
functional/channel_program/synctask_core/tst.list_snapshots.ksh \
functional/channel_program/synctask_core/tst.list_system_props.ksh \
functional/channel_program/synctask_core/tst.list_user_props.ksh \
functional/channel_program/synctask_core/tst.parse_args_neg.ksh \
functional/channel_program/synctask_core/tst.promote_conflict.ksh \
functional/channel_program/synctask_core/tst.promote_multiple.ksh \
functional/channel_program/synctask_core/tst.promote_simple.ksh \
functional/channel_program/synctask_core/tst.rollback_mult.ksh \
functional/channel_program/synctask_core/tst.rollback_one.ksh \
functional/channel_program/synctask_core/tst.set_props.ksh \
functional/channel_program/synctask_core/tst.snapshot_destroy.ksh \
functional/channel_program/synctask_core/tst.snapshot_neg.ksh \
functional/channel_program/synctask_core/tst.snapshot_recursive.ksh \
functional/channel_program/synctask_core/tst.snapshot_rename.ksh \
functional/channel_program/synctask_core/tst.snapshot_simple.ksh \
functional/channel_program/synctask_core/tst.terminate_by_signal.ksh \
functional/chattr/chattr_001_pos.ksh \
functional/chattr/chattr_002_neg.ksh \
functional/chattr/cleanup.ksh \
functional/chattr/setup.ksh \
functional/checksum/cleanup.ksh \
functional/checksum/filetest_001_pos.ksh \
functional/checksum/filetest_002_pos.ksh \
functional/checksum/run_blake3_test.ksh \
functional/checksum/run_edonr_test.ksh \
functional/checksum/run_sha2_test.ksh \
functional/checksum/run_skein_test.ksh \
functional/checksum/setup.ksh \
functional/clean_mirror/clean_mirror_001_pos.ksh \
functional/clean_mirror/clean_mirror_002_pos.ksh \
functional/clean_mirror/clean_mirror_003_pos.ksh \
functional/clean_mirror/clean_mirror_004_pos.ksh \
functional/clean_mirror/cleanup.ksh \
functional/clean_mirror/setup.ksh \
functional/cli_root/zdb/zdb_002_pos.ksh \
functional/cli_root/zdb/zdb_003_pos.ksh \
functional/cli_root/zdb/zdb_004_pos.ksh \
functional/cli_root/zdb/zdb_005_pos.ksh \
functional/cli_root/zdb/zdb_006_pos.ksh \
functional/cli_root/zdb/zdb_args_neg.ksh \
functional/cli_root/zdb/zdb_args_pos.ksh \
functional/cli_root/zdb/zdb_backup.ksh \
functional/cli_root/zdb/zdb_block_size_histogram.ksh \
functional/cli_root/zdb/zdb_checksum.ksh \
functional/cli_root/zdb/zdb_decompress.ksh \
functional/cli_root/zdb/zdb_decompress_zstd.ksh \
functional/cli_root/zdb/zdb_display_block.ksh \
functional/cli_root/zdb/zdb_encrypted.ksh \
functional/cli_root/zdb/zdb_label_checksum.ksh \
functional/cli_root/zdb/zdb_object_range_neg.ksh \
functional/cli_root/zdb/zdb_object_range_pos.ksh \
functional/cli_root/zdb/zdb_objset_id.ksh \
functional/cli_root/zdb/zdb_recover_2.ksh \
functional/cli_root/zdb/zdb_recover.ksh \
functional/cli_root/zfs_bookmark/cleanup.ksh \
functional/cli_root/zfs_bookmark/setup.ksh \
functional/cli_root/zfs_bookmark/zfs_bookmark_cliargs.ksh \
functional/cli_root/zfs_change-key/cleanup.ksh \
functional/cli_root/zfs_change-key/setup.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_child.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_clones.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_format.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_inherit.ksh \
functional/cli_root/zfs_change-key/zfs_change-key.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_load.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_location.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_pbkdf2iters.ksh \
functional/cli_root/zfs/cleanup.ksh \
functional/cli_root/zfs_clone/cleanup.ksh \
functional/cli_root/zfs_clone/setup.ksh \
functional/cli_root/zfs_clone/zfs_clone_001_neg.ksh \
functional/cli_root/zfs_clone/zfs_clone_002_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_003_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_004_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_005_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_006_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_007_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_008_neg.ksh \
functional/cli_root/zfs_clone/zfs_clone_009_neg.ksh \
functional/cli_root/zfs_clone/zfs_clone_010_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_deeply_nested.ksh \
functional/cli_root/zfs_clone/zfs_clone_encrypted.ksh \
functional/cli_root/zfs_clone/zfs_clone_rm_nested.ksh \
functional/cli_root/zfs_copies/cleanup.ksh \
functional/cli_root/zfs_copies/setup.ksh \
functional/cli_root/zfs_copies/zfs_copies_001_pos.ksh \
functional/cli_root/zfs_copies/zfs_copies_002_pos.ksh \
functional/cli_root/zfs_copies/zfs_copies_003_pos.ksh \
functional/cli_root/zfs_copies/zfs_copies_004_neg.ksh \
functional/cli_root/zfs_copies/zfs_copies_005_neg.ksh \
functional/cli_root/zfs_copies/zfs_copies_006_pos.ksh \
functional/cli_root/zfs_create/cleanup.ksh \
functional/cli_root/zfs_create/setup.ksh \
functional/cli_root/zfs_create/zfs_create_001_pos.ksh \
functional/cli_root/zfs_create/zfs_create_002_pos.ksh \
functional/cli_root/zfs_create/zfs_create_003_pos.ksh \
functional/cli_root/zfs_create/zfs_create_004_pos.ksh \
functional/cli_root/zfs_create/zfs_create_005_pos.ksh \
functional/cli_root/zfs_create/zfs_create_006_pos.ksh \
functional/cli_root/zfs_create/zfs_create_007_pos.ksh \
functional/cli_root/zfs_create/zfs_create_008_neg.ksh \
functional/cli_root/zfs_create/zfs_create_009_neg.ksh \
functional/cli_root/zfs_create/zfs_create_010_neg.ksh \
functional/cli_root/zfs_create/zfs_create_011_pos.ksh \
functional/cli_root/zfs_create/zfs_create_012_pos.ksh \
functional/cli_root/zfs_create/zfs_create_013_pos.ksh \
functional/cli_root/zfs_create/zfs_create_014_pos.ksh \
functional/cli_root/zfs_create/zfs_create_crypt_combos.ksh \
functional/cli_root/zfs_create/zfs_create_dryrun.ksh \
functional/cli_root/zfs_create/zfs_create_encrypted.ksh \
functional/cli_root/zfs_create/zfs_create_nomount.ksh \
functional/cli_root/zfs_create/zfs_create_verbose.ksh \
functional/cli_root/zfs_destroy/cleanup.ksh \
functional/cli_root/zfs_destroy/setup.ksh \
functional/cli_root/zfs_destroy/zfs_clone_livelist_condense_and_disable.ksh \
functional/cli_root/zfs_destroy/zfs_clone_livelist_condense_races.ksh \
functional/cli_root/zfs_destroy/zfs_clone_livelist_dedup.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_001_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_002_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_003_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_004_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_005_neg.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_006_neg.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_007_neg.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_008_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_009_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_010_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_011_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_012_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_013_neg.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_014_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_015_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_016_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_clone_livelist.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_dev_removal_condense.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_dev_removal.ksh \
functional/cli_root/zfs_diff/cleanup.ksh \
functional/cli_root/zfs_diff/setup.ksh \
functional/cli_root/zfs_diff/zfs_diff_changes.ksh \
functional/cli_root/zfs_diff/zfs_diff_cliargs.ksh \
functional/cli_root/zfs_diff/zfs_diff_encrypted.ksh \
functional/cli_root/zfs_diff/zfs_diff_mangle.ksh \
functional/cli_root/zfs_diff/zfs_diff_timestamp.ksh \
functional/cli_root/zfs_diff/zfs_diff_types.ksh \
functional/cli_root/zfs_get/cleanup.ksh \
functional/cli_root/zfs_get/setup.ksh \
functional/cli_root/zfs_get/zfs_get_001_pos.ksh \
functional/cli_root/zfs_get/zfs_get_002_pos.ksh \
functional/cli_root/zfs_get/zfs_get_003_pos.ksh \
functional/cli_root/zfs_get/zfs_get_004_pos.ksh \
functional/cli_root/zfs_get/zfs_get_005_neg.ksh \
functional/cli_root/zfs_get/zfs_get_006_neg.ksh \
functional/cli_root/zfs_get/zfs_get_007_neg.ksh \
functional/cli_root/zfs_get/zfs_get_008_pos.ksh \
functional/cli_root/zfs_get/zfs_get_009_pos.ksh \
functional/cli_root/zfs_get/zfs_get_010_neg.ksh \
functional/cli_root/zfs_ids_to_path/cleanup.ksh \
functional/cli_root/zfs_ids_to_path/setup.ksh \
functional/cli_root/zfs_ids_to_path/zfs_ids_to_path_001_pos.ksh \
functional/cli_root/zfs_inherit/cleanup.ksh \
functional/cli_root/zfs_inherit/setup.ksh \
functional/cli_root/zfs_inherit/zfs_inherit_001_neg.ksh \
functional/cli_root/zfs_inherit/zfs_inherit_002_neg.ksh \
functional/cli_root/zfs_inherit/zfs_inherit_003_pos.ksh \
functional/cli_root/zfs_inherit/zfs_inherit_mountpoint.ksh \
functional/cli_root/zfs_jail/cleanup.ksh \
functional/cli_root/zfs_jail/setup.ksh \
functional/cli_root/zfs_jail/zfs_jail_001_pos.ksh \
functional/cli_root/zfs_load-key/cleanup.ksh \
functional/cli_root/zfs_load-key/setup.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_all.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_file.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_https.ksh \
functional/cli_root/zfs_load-key/zfs_load-key.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_location.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_noop.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_recursive.ksh \
functional/cli_root/zfs_mount/cleanup.ksh \
functional/cli_root/zfs_mount/setup.ksh \
functional/cli_root/zfs_mount/zfs_mount_001_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_002_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_003_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_004_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_005_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_006_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_007_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_008_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_009_neg.ksh \
functional/cli_root/zfs_mount/zfs_mount_010_neg.ksh \
functional/cli_root/zfs_mount/zfs_mount_011_neg.ksh \
functional/cli_root/zfs_mount/zfs_mount_012_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_013_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_014_neg.ksh \
functional/cli_root/zfs_mount/zfs_mount_all_001_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_all_fail.ksh \
functional/cli_root/zfs_mount/zfs_mount_all_mountpoints.ksh \
functional/cli_root/zfs_mount/zfs_mount_encrypted.ksh \
functional/cli_root/zfs_mount/zfs_mount_remount.ksh \
functional/cli_root/zfs_mount/zfs_mount_test_race.ksh \
functional/cli_root/zfs_mount/zfs_multi_mount.ksh \
functional/cli_root/zfs_program/cleanup.ksh \
functional/cli_root/zfs_program/setup.ksh \
functional/cli_root/zfs_program/zfs_program_json.ksh \
functional/cli_root/zfs_promote/cleanup.ksh \
functional/cli_root/zfs_promote/setup.ksh \
functional/cli_root/zfs_promote/zfs_promote_001_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_002_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_003_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_004_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_005_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_006_neg.ksh \
functional/cli_root/zfs_promote/zfs_promote_007_neg.ksh \
functional/cli_root/zfs_promote/zfs_promote_008_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_encryptionroot.ksh \
functional/cli_root/zfs_property/cleanup.ksh \
functional/cli_root/zfs_property/setup.ksh \
functional/cli_root/zfs_property/zfs_written_property_001_pos.ksh \
functional/cli_root/zfs_receive/cleanup.ksh \
functional/cli_root/zfs_receive/receive-o-x_props_aliases.ksh \
functional/cli_root/zfs_receive/receive-o-x_props_override.ksh \
functional/cli_root/zfs_receive/setup.ksh \
functional/cli_root/zfs_receive/zfs_receive_001_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_002_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_003_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_004_neg.ksh \
functional/cli_root/zfs_receive/zfs_receive_005_neg.ksh \
functional/cli_root/zfs_receive/zfs_receive_006_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_007_neg.ksh \
functional/cli_root/zfs_receive/zfs_receive_008_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_009_neg.ksh \
functional/cli_root/zfs_receive/zfs_receive_010_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_011_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_012_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_013_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_014_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_015_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_016_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_-e.ksh \
functional/cli_root/zfs_receive/zfs_receive_from_encrypted.ksh \
functional/cli_root/zfs_receive/zfs_receive_from_zstd.ksh \
functional/cli_root/zfs_receive/zfs_receive_new_props.ksh \
functional/cli_root/zfs_receive/zfs_receive_raw_-d.ksh \
functional/cli_root/zfs_receive/zfs_receive_raw_incremental.ksh \
functional/cli_root/zfs_receive/zfs_receive_raw.ksh \
functional/cli_root/zfs_receive/zfs_receive_to_encrypted.ksh \
functional/cli_root/zfs_receive/zfs_receive_-wR-encrypted-mix.ksh \
functional/cli_root/zfs_receive/zfs_receive_corrective.ksh \
functional/cli_root/zfs_receive/zfs_receive_compressed_corrective.ksh \
functional/cli_root/zfs_receive/zfs_receive_large_block_corrective.ksh \
functional/cli_root/zfs_rename/cleanup.ksh \
functional/cli_root/zfs_rename/setup.ksh \
functional/cli_root/zfs_rename/zfs_rename_001_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_002_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_003_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_004_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_005_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_006_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_007_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_008_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_009_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_010_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_011_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_012_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_013_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_014_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_encrypted_child.ksh \
functional/cli_root/zfs_rename/zfs_rename_mountpoint.ksh \
functional/cli_root/zfs_rename/zfs_rename_nounmount.ksh \
functional/cli_root/zfs_rename/zfs_rename_to_encrypted.ksh \
functional/cli_root/zfs_reservation/cleanup.ksh \
functional/cli_root/zfs_reservation/setup.ksh \
functional/cli_root/zfs_reservation/zfs_reservation_001_pos.ksh \
functional/cli_root/zfs_reservation/zfs_reservation_002_pos.ksh \
functional/cli_root/zfs_rollback/cleanup.ksh \
functional/cli_root/zfs_rollback/setup.ksh \
functional/cli_root/zfs_rollback/zfs_rollback_001_pos.ksh \
functional/cli_root/zfs_rollback/zfs_rollback_002_pos.ksh \
functional/cli_root/zfs_rollback/zfs_rollback_003_neg.ksh \
functional/cli_root/zfs_rollback/zfs_rollback_004_neg.ksh \
functional/cli_root/zfs_send/cleanup.ksh \
functional/cli_root/zfs_send/setup.ksh \
functional/cli_root/zfs_send/zfs_send_001_pos.ksh \
functional/cli_root/zfs_send/zfs_send_002_pos.ksh \
functional/cli_root/zfs_send/zfs_send_003_pos.ksh \
functional/cli_root/zfs_send/zfs_send_004_neg.ksh \
functional/cli_root/zfs_send/zfs_send_005_pos.ksh \
functional/cli_root/zfs_send/zfs_send_006_pos.ksh \
functional/cli_root/zfs_send/zfs_send_007_pos.ksh \
functional/cli_root/zfs_send/zfs_send-b.ksh \
functional/cli_root/zfs_send/zfs_send_encrypted.ksh \
functional/cli_root/zfs_send/zfs_send_encrypted_unloaded.ksh \
functional/cli_root/zfs_send/zfs_send_raw.ksh \
functional/cli_root/zfs_send/zfs_send_skip_missing.ksh \
functional/cli_root/zfs_send/zfs_send_sparse.ksh \
functional/cli_root/zfs_set/cache_001_pos.ksh \
functional/cli_root/zfs_set/cache_002_neg.ksh \
functional/cli_root/zfs_set/canmount_001_pos.ksh \
functional/cli_root/zfs_set/canmount_002_pos.ksh \
functional/cli_root/zfs_set/canmount_003_pos.ksh \
functional/cli_root/zfs_set/canmount_004_pos.ksh \
functional/cli_root/zfs_set/checksum_001_pos.ksh \
functional/cli_root/zfs_set/cleanup.ksh \
functional/cli_root/zfs_set/compression_001_pos.ksh \
functional/cli_root/zfs_set/mountpoint_001_pos.ksh \
functional/cli_root/zfs_set/mountpoint_002_pos.ksh \
functional/cli_root/zfs_set/mountpoint_003_pos.ksh \
functional/cli_root/zfs_set/onoffs_001_pos.ksh \
functional/cli_root/zfs_set/property_alias_001_pos.ksh \
functional/cli_root/zfs_set/readonly_001_pos.ksh \
functional/cli_root/zfs_set/reservation_001_neg.ksh \
functional/cli_root/zfs_set/ro_props_001_pos.ksh \
functional/cli_root/zfs_set/setup.ksh \
functional/cli_root/zfs_set/share_mount_001_neg.ksh \
functional/cli_root/zfs_set/snapdir_001_pos.ksh \
functional/cli_root/zfs/setup.ksh \
functional/cli_root/zfs_set/user_property_001_pos.ksh \
functional/cli_root/zfs_set/user_property_002_pos.ksh \
functional/cli_root/zfs_set/user_property_003_neg.ksh \
functional/cli_root/zfs_set/user_property_004_pos.ksh \
functional/cli_root/zfs_set/version_001_neg.ksh \
functional/cli_root/zfs_set/zfs_set_001_neg.ksh \
functional/cli_root/zfs_set/zfs_set_002_neg.ksh \
functional/cli_root/zfs_set/zfs_set_003_neg.ksh \
functional/cli_root/zfs_set/zfs_set_feature_activation.ksh \
functional/cli_root/zfs_set/zfs_set_keylocation.ksh \
functional/cli_root/zfs_set/zfs_set_nomount.ksh \
functional/cli_root/zfs_share/cleanup.ksh \
functional/cli_root/zfs_share/setup.ksh \
functional/cli_root/zfs_share/zfs_share_001_pos.ksh \
functional/cli_root/zfs_share/zfs_share_002_pos.ksh \
functional/cli_root/zfs_share/zfs_share_003_pos.ksh \
functional/cli_root/zfs_share/zfs_share_004_pos.ksh \
functional/cli_root/zfs_share/zfs_share_005_pos.ksh \
functional/cli_root/zfs_share/zfs_share_006_pos.ksh \
functional/cli_root/zfs_share/zfs_share_007_neg.ksh \
functional/cli_root/zfs_share/zfs_share_008_neg.ksh \
functional/cli_root/zfs_share/zfs_share_009_neg.ksh \
functional/cli_root/zfs_share/zfs_share_010_neg.ksh \
functional/cli_root/zfs_share/zfs_share_011_pos.ksh \
functional/cli_root/zfs_share/zfs_share_012_pos.ksh \
functional/cli_root/zfs_share/zfs_share_013_pos.ksh \
functional/cli_root/zfs_share/zfs_share_concurrent_shares.ksh \
functional/cli_root/zfs_snapshot/cleanup.ksh \
functional/cli_root/zfs_snapshot/setup.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_001_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_002_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_003_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_004_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_005_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_006_pos.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_007_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_008_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_009_pos.ksh \
functional/cli_root/zfs_sysfs/cleanup.ksh \
functional/cli_root/zfs_sysfs/setup.ksh \
functional/cli_root/zfs_sysfs/zfeature_set_unsupported.ksh \
functional/cli_root/zfs_sysfs/zfs_get_unsupported.ksh \
functional/cli_root/zfs_sysfs/zfs_set_unsupported.ksh \
functional/cli_root/zfs_sysfs/zfs_sysfs_live.ksh \
functional/cli_root/zfs_sysfs/zpool_get_unsupported.ksh \
functional/cli_root/zfs_sysfs/zpool_set_unsupported.ksh \
functional/cli_root/zfs_unload-key/cleanup.ksh \
functional/cli_root/zfs_unload-key/setup.ksh \
functional/cli_root/zfs_unload-key/zfs_unload-key_all.ksh \
functional/cli_root/zfs_unload-key/zfs_unload-key.ksh \
functional/cli_root/zfs_unload-key/zfs_unload-key_recursive.ksh \
functional/cli_root/zfs_unmount/cleanup.ksh \
functional/cli_root/zfs_unmount/setup.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_001_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_002_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_003_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_004_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_005_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_006_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_007_neg.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_008_neg.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_009_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_all_001_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_nested.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_unload_keys.ksh \
functional/cli_root/zfs_unshare/cleanup.ksh \
functional/cli_root/zfs_unshare/setup.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_001_pos.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_002_pos.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_003_pos.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_004_neg.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_005_neg.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_006_pos.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_007_pos.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_008_pos.ksh \
functional/cli_root/zfs_upgrade/cleanup.ksh \
functional/cli_root/zfs_upgrade/setup.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_001_pos.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_002_pos.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_003_pos.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_004_pos.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_005_pos.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_006_neg.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_007_neg.ksh \
functional/cli_root/zfs_wait/cleanup.ksh \
functional/cli_root/zfs_wait/setup.ksh \
functional/cli_root/zfs_wait/zfs_wait_deleteq.ksh \
functional/cli_root/zfs_wait/zfs_wait_getsubopt.ksh \
functional/cli_root/zfs/zfs_001_neg.ksh \
functional/cli_root/zfs/zfs_002_pos.ksh \
functional/cli_root/zfs/zfs_003_neg.ksh \
functional/cli_root/zhack/zhack_label_repair_001.ksh \
functional/cli_root/zhack/zhack_label_repair_002.ksh \
functional/cli_root/zhack/zhack_label_repair_003.ksh \
functional/cli_root/zhack/zhack_label_repair_004.ksh \
functional/cli_root/zpool_add/add_nested_replacing_spare.ksh \
functional/cli_root/zpool_add/add-o_ashift.ksh \
functional/cli_root/zpool_add/add_prop_ashift.ksh \
functional/cli_root/zpool_add/cleanup.ksh \
functional/cli_root/zpool_add/setup.ksh \
functional/cli_root/zpool_add/zpool_add_001_pos.ksh \
functional/cli_root/zpool_add/zpool_add_002_pos.ksh \
functional/cli_root/zpool_add/zpool_add_003_pos.ksh \
functional/cli_root/zpool_add/zpool_add_004_pos.ksh \
functional/cli_root/zpool_add/zpool_add_005_pos.ksh \
functional/cli_root/zpool_add/zpool_add_006_pos.ksh \
functional/cli_root/zpool_add/zpool_add_007_neg.ksh \
functional/cli_root/zpool_add/zpool_add_008_neg.ksh \
functional/cli_root/zpool_add/zpool_add_009_neg.ksh \
functional/cli_root/zpool_add/zpool_add_010_pos.ksh \
functional/cli_root/zpool_add/zpool_add_dryrun_output.ksh \
functional/cli_root/zpool_attach/attach-o_ashift.ksh \
functional/cli_root/zpool_attach/cleanup.ksh \
functional/cli_root/zpool_attach/setup.ksh \
functional/cli_root/zpool_attach/zpool_attach_001_neg.ksh \
functional/cli_root/zpool/cleanup.ksh \
functional/cli_root/zpool_clear/cleanup.ksh \
functional/cli_root/zpool_clear/setup.ksh \
functional/cli_root/zpool_clear/zpool_clear_001_pos.ksh \
functional/cli_root/zpool_clear/zpool_clear_002_neg.ksh \
functional/cli_root/zpool_clear/zpool_clear_003_neg.ksh \
functional/cli_root/zpool_clear/zpool_clear_readonly.ksh \
functional/cli_root/zpool_create/cleanup.ksh \
functional/cli_root/zpool_create/create-o_ashift.ksh \
functional/cli_root/zpool_create/setup.ksh \
functional/cli_root/zpool_create/zpool_create_001_pos.ksh \
functional/cli_root/zpool_create/zpool_create_002_pos.ksh \
functional/cli_root/zpool_create/zpool_create_003_pos.ksh \
functional/cli_root/zpool_create/zpool_create_004_pos.ksh \
functional/cli_root/zpool_create/zpool_create_005_pos.ksh \
functional/cli_root/zpool_create/zpool_create_006_pos.ksh \
functional/cli_root/zpool_create/zpool_create_007_neg.ksh \
functional/cli_root/zpool_create/zpool_create_008_pos.ksh \
functional/cli_root/zpool_create/zpool_create_009_neg.ksh \
functional/cli_root/zpool_create/zpool_create_010_neg.ksh \
functional/cli_root/zpool_create/zpool_create_011_neg.ksh \
functional/cli_root/zpool_create/zpool_create_012_neg.ksh \
functional/cli_root/zpool_create/zpool_create_014_neg.ksh \
functional/cli_root/zpool_create/zpool_create_015_neg.ksh \
functional/cli_root/zpool_create/zpool_create_016_pos.ksh \
functional/cli_root/zpool_create/zpool_create_017_neg.ksh \
functional/cli_root/zpool_create/zpool_create_018_pos.ksh \
functional/cli_root/zpool_create/zpool_create_019_pos.ksh \
functional/cli_root/zpool_create/zpool_create_020_pos.ksh \
functional/cli_root/zpool_create/zpool_create_021_pos.ksh \
functional/cli_root/zpool_create/zpool_create_022_pos.ksh \
functional/cli_root/zpool_create/zpool_create_023_neg.ksh \
functional/cli_root/zpool_create/zpool_create_024_pos.ksh \
functional/cli_root/zpool_create/zpool_create_crypt_combos.ksh \
functional/cli_root/zpool_create/zpool_create_draid_001_pos.ksh \
functional/cli_root/zpool_create/zpool_create_draid_002_pos.ksh \
functional/cli_root/zpool_create/zpool_create_draid_003_pos.ksh \
functional/cli_root/zpool_create/zpool_create_draid_004_pos.ksh \
functional/cli_root/zpool_create/zpool_create_dryrun_output.ksh \
functional/cli_root/zpool_create/zpool_create_encrypted.ksh \
functional/cli_root/zpool_create/zpool_create_features_001_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_002_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_003_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_004_neg.ksh \
functional/cli_root/zpool_create/zpool_create_features_005_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_006_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_007_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_008_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_009_pos.ksh \
functional/cli_root/zpool_create/zpool_create_tempname.ksh \
functional/cli_root/zpool_destroy/zpool_destroy_001_pos.ksh \
functional/cli_root/zpool_destroy/zpool_destroy_002_pos.ksh \
functional/cli_root/zpool_destroy/zpool_destroy_003_neg.ksh \
functional/cli_root/zpool_detach/cleanup.ksh \
functional/cli_root/zpool_detach/setup.ksh \
functional/cli_root/zpool_detach/zpool_detach_001_neg.ksh \
functional/cli_root/zpool_events/cleanup.ksh \
functional/cli_root/zpool_events/setup.ksh \
functional/cli_root/zpool_events/zpool_events_clear.ksh \
functional/cli_root/zpool_events/zpool_events_clear_retained.ksh \
functional/cli_root/zpool_events/zpool_events_cliargs.ksh \
functional/cli_root/zpool_events/zpool_events_duplicates.ksh \
functional/cli_root/zpool_events/zpool_events_errors.ksh \
functional/cli_root/zpool_events/zpool_events_follow.ksh \
functional/cli_root/zpool_events/zpool_events_poolname.ksh \
functional/cli_root/zpool_expand/cleanup.ksh \
functional/cli_root/zpool_expand/setup.ksh \
functional/cli_root/zpool_expand/zpool_expand_001_pos.ksh \
functional/cli_root/zpool_expand/zpool_expand_002_pos.ksh \
functional/cli_root/zpool_expand/zpool_expand_003_neg.ksh \
functional/cli_root/zpool_expand/zpool_expand_004_pos.ksh \
functional/cli_root/zpool_expand/zpool_expand_005_pos.ksh \
functional/cli_root/zpool_export/cleanup.ksh \
functional/cli_root/zpool_export/setup.ksh \
functional/cli_root/zpool_export/zpool_export_001_pos.ksh \
functional/cli_root/zpool_export/zpool_export_002_pos.ksh \
functional/cli_root/zpool_export/zpool_export_003_neg.ksh \
functional/cli_root/zpool_export/zpool_export_004_pos.ksh \
functional/cli_root/zpool_get/cleanup.ksh \
functional/cli_root/zpool_get/setup.ksh \
functional/cli_root/zpool_get/vdev_get_001_pos.ksh \
functional/cli_root/zpool_get/zpool_get_001_pos.ksh \
functional/cli_root/zpool_get/zpool_get_002_pos.ksh \
functional/cli_root/zpool_get/zpool_get_003_pos.ksh \
functional/cli_root/zpool_get/zpool_get_004_neg.ksh \
functional/cli_root/zpool_get/zpool_get_005_pos.ksh \
functional/cli_root/zpool_history/cleanup.ksh \
functional/cli_root/zpool_history/setup.ksh \
functional/cli_root/zpool_history/zpool_history_001_neg.ksh \
functional/cli_root/zpool_history/zpool_history_002_pos.ksh \
functional/cli_root/zpool_import/cleanup.ksh \
functional/cli_root/zpool_import/import_cachefile_device_added.ksh \
functional/cli_root/zpool_import/import_cachefile_device_removed.ksh \
functional/cli_root/zpool_import/import_cachefile_device_replaced.ksh \
functional/cli_root/zpool_import/import_cachefile_mirror_attached.ksh \
functional/cli_root/zpool_import/import_cachefile_mirror_detached.ksh \
functional/cli_root/zpool_import/import_cachefile_paths_changed.ksh \
functional/cli_root/zpool_import/import_cachefile_shared_device.ksh \
functional/cli_root/zpool_import/import_devices_missing.ksh \
functional/cli_root/zpool_import/import_log_missing.ksh \
functional/cli_root/zpool_import/import_paths_changed.ksh \
functional/cli_root/zpool_import/import_rewind_config_changed.ksh \
functional/cli_root/zpool_import/import_rewind_device_replaced.ksh \
functional/cli_root/zpool_import/setup.ksh \
functional/cli_root/zpool_import/zpool_import_001_pos.ksh \
functional/cli_root/zpool_import/zpool_import_002_pos.ksh \
functional/cli_root/zpool_import/zpool_import_003_pos.ksh \
functional/cli_root/zpool_import/zpool_import_004_pos.ksh \
functional/cli_root/zpool_import/zpool_import_005_pos.ksh \
functional/cli_root/zpool_import/zpool_import_006_pos.ksh \
functional/cli_root/zpool_import/zpool_import_007_pos.ksh \
functional/cli_root/zpool_import/zpool_import_008_pos.ksh \
functional/cli_root/zpool_import/zpool_import_009_neg.ksh \
functional/cli_root/zpool_import/zpool_import_010_pos.ksh \
functional/cli_root/zpool_import/zpool_import_011_neg.ksh \
functional/cli_root/zpool_import/zpool_import_012_pos.ksh \
functional/cli_root/zpool_import/zpool_import_013_neg.ksh \
functional/cli_root/zpool_import/zpool_import_014_pos.ksh \
functional/cli_root/zpool_import/zpool_import_015_pos.ksh \
functional/cli_root/zpool_import/zpool_import_016_pos.ksh \
functional/cli_root/zpool_import/zpool_import_017_pos.ksh \
functional/cli_root/zpool_import/zpool_import_all_001_pos.ksh \
functional/cli_root/zpool_import/zpool_import_encrypted.ksh \
functional/cli_root/zpool_import/zpool_import_encrypted_load.ksh \
functional/cli_root/zpool_import/zpool_import_errata3.ksh \
functional/cli_root/zpool_import/zpool_import_errata4.ksh \
functional/cli_root/zpool_import/zpool_import_features_001_pos.ksh \
functional/cli_root/zpool_import/zpool_import_features_002_neg.ksh \
functional/cli_root/zpool_import/zpool_import_features_003_pos.ksh \
functional/cli_root/zpool_import/zpool_import_hostid_changed.ksh \
functional/cli_root/zpool_import/zpool_import_hostid_changed_unclean_export.ksh \
functional/cli_root/zpool_import/zpool_import_hostid_changed_cachefile.ksh \
functional/cli_root/zpool_import/zpool_import_hostid_changed_cachefile_unclean_export.ksh \
functional/cli_root/zpool_import/zpool_import_missing_001_pos.ksh \
functional/cli_root/zpool_import/zpool_import_missing_002_pos.ksh \
functional/cli_root/zpool_import/zpool_import_missing_003_pos.ksh \
functional/cli_root/zpool_import/zpool_import_rename_001_pos.ksh \
functional/cli_root/zpool_initialize/cleanup.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_attach_detach_add_remove.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_fault_export_import_online.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_import_export.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_offline_export_import_online.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_online_offline.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_split.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_start_and_cancel_neg.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_start_and_cancel_pos.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_suspend_resume.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_uninit.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_unsupported_vdevs.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_verify_checksums.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_verify_initialized.ksh \
functional/cli_root/zpool_labelclear/zpool_labelclear_active.ksh \
functional/cli_root/zpool_labelclear/zpool_labelclear_exported.ksh \
functional/cli_root/zpool_labelclear/zpool_labelclear_removed.ksh \
functional/cli_root/zpool_labelclear/zpool_labelclear_valid.ksh \
functional/cli_root/zpool_offline/cleanup.ksh \
functional/cli_root/zpool_offline/setup.ksh \
functional/cli_root/zpool_offline/zpool_offline_001_pos.ksh \
functional/cli_root/zpool_offline/zpool_offline_002_neg.ksh \
functional/cli_root/zpool_offline/zpool_offline_003_pos.ksh \
functional/cli_root/zpool_online/cleanup.ksh \
functional/cli_root/zpool_online/setup.ksh \
functional/cli_root/zpool_online/zpool_online_001_pos.ksh \
functional/cli_root/zpool_online/zpool_online_002_neg.ksh \
functional/cli_root/zpool_remove/cleanup.ksh \
functional/cli_root/zpool_remove/setup.ksh \
functional/cli_root/zpool_remove/zpool_remove_001_neg.ksh \
functional/cli_root/zpool_remove/zpool_remove_002_pos.ksh \
functional/cli_root/zpool_remove/zpool_remove_003_pos.ksh \
functional/cli_root/zpool_reopen/cleanup.ksh \
functional/cli_root/zpool_reopen/setup.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_001_pos.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_002_pos.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_003_pos.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_004_pos.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_005_pos.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_006_neg.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_007_pos.ksh \
functional/cli_root/zpool_replace/cleanup.ksh \
functional/cli_root/zpool_replace/replace-o_ashift.ksh \
functional/cli_root/zpool_replace/replace_prop_ashift.ksh \
functional/cli_root/zpool_replace/setup.ksh \
functional/cli_root/zpool_replace/zpool_replace_001_neg.ksh \
functional/cli_root/zpool_resilver/cleanup.ksh \
functional/cli_root/zpool_resilver/setup.ksh \
functional/cli_root/zpool_resilver/zpool_resilver_bad_args.ksh \
functional/cli_root/zpool_resilver/zpool_resilver_restart.ksh \
functional/cli_root/zpool_resilver/zpool_resilver_concurrent.ksh \
functional/cli_root/zpool_scrub/cleanup.ksh \
functional/cli_root/zpool_scrub/setup.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_001_neg.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_002_pos.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_003_pos.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_004_pos.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_005_pos.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_encrypted_unloaded.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_multiple_copies.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_offline_device.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_print_repairing.ksh \
functional/cli_root/zpool_scrub/zpool_error_scrub_001_pos.ksh \
functional/cli_root/zpool_scrub/zpool_error_scrub_002_pos.ksh \
functional/cli_root/zpool_scrub/zpool_error_scrub_003_pos.ksh \
functional/cli_root/zpool_scrub/zpool_error_scrub_004_pos.ksh \
functional/cli_root/zpool_set/cleanup.ksh \
functional/cli_root/zpool_set/setup.ksh \
functional/cli_root/zpool/setup.ksh \
functional/cli_root/zpool_set/vdev_set_001_pos.ksh \
functional/cli_root/zpool_set/zpool_set_common.kshlib \
functional/cli_root/zpool_set/zpool_set_001_pos.ksh \
functional/cli_root/zpool_set/zpool_set_002_neg.ksh \
functional/cli_root/zpool_set/zpool_set_003_neg.ksh \
functional/cli_root/zpool_set/zpool_set_ashift.ksh \
functional/cli_root/zpool_set/user_property_001_pos.ksh \
functional/cli_root/zpool_set/user_property_002_neg.ksh \
functional/cli_root/zpool_set/zpool_set_features.ksh \
functional/cli_root/zpool_split/cleanup.ksh \
functional/cli_root/zpool_split/setup.ksh \
functional/cli_root/zpool_split/zpool_split_cliargs.ksh \
functional/cli_root/zpool_split/zpool_split_devices.ksh \
functional/cli_root/zpool_split/zpool_split_dryrun_output.ksh \
functional/cli_root/zpool_split/zpool_split_encryption.ksh \
functional/cli_root/zpool_split/zpool_split_indirect.ksh \
functional/cli_root/zpool_split/zpool_split_props.ksh \
functional/cli_root/zpool_split/zpool_split_resilver.ksh \
functional/cli_root/zpool_split/zpool_split_vdevs.ksh \
functional/cli_root/zpool_split/zpool_split_wholedisk.ksh \
functional/cli_root/zpool_status/cleanup.ksh \
functional/cli_root/zpool_status/setup.ksh \
functional/cli_root/zpool_status/zpool_status_001_pos.ksh \
functional/cli_root/zpool_status/zpool_status_002_pos.ksh \
functional/cli_root/zpool_status/zpool_status_003_pos.ksh \
functional/cli_root/zpool_status/zpool_status_004_pos.ksh \
functional/cli_root/zpool_status/zpool_status_005_pos.ksh \
functional/cli_root/zpool_status/zpool_status_006_pos.ksh \
functional/cli_root/zpool_status/zpool_status_007_pos.ksh \
functional/cli_root/zpool_status/zpool_status_features_001_pos.ksh \
functional/cli_root/zpool_sync/cleanup.ksh \
functional/cli_root/zpool_sync/setup.ksh \
functional/cli_root/zpool_sync/zpool_sync_001_pos.ksh \
functional/cli_root/zpool_sync/zpool_sync_002_neg.ksh \
functional/cli_root/zpool_trim/cleanup.ksh \
functional/cli_root/zpool_trim/setup.ksh \
functional/cli_root/zpool_trim/zpool_trim_attach_detach_add_remove.ksh \
functional/cli_root/zpool_trim/zpool_trim_fault_export_import_online.ksh \
functional/cli_root/zpool_trim/zpool_trim_import_export.ksh \
functional/cli_root/zpool_trim/zpool_trim_multiple.ksh \
functional/cli_root/zpool_trim/zpool_trim_neg.ksh \
functional/cli_root/zpool_trim/zpool_trim_offline_export_import_online.ksh \
functional/cli_root/zpool_trim/zpool_trim_online_offline.ksh \
functional/cli_root/zpool_trim/zpool_trim_partial.ksh \
functional/cli_root/zpool_trim/zpool_trim_rate.ksh \
functional/cli_root/zpool_trim/zpool_trim_rate_neg.ksh \
functional/cli_root/zpool_trim/zpool_trim_secure.ksh \
functional/cli_root/zpool_trim/zpool_trim_split.ksh \
functional/cli_root/zpool_trim/zpool_trim_start_and_cancel_neg.ksh \
functional/cli_root/zpool_trim/zpool_trim_start_and_cancel_pos.ksh \
functional/cli_root/zpool_trim/zpool_trim_suspend_resume.ksh \
functional/cli_root/zpool_trim/zpool_trim_unsupported_vdevs.ksh \
functional/cli_root/zpool_trim/zpool_trim_verify_checksums.ksh \
functional/cli_root/zpool_trim/zpool_trim_verify_trimmed.ksh \
functional/cli_root/zpool_upgrade/cleanup.ksh \
functional/cli_root/zpool_upgrade/setup.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_001_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_002_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_003_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_004_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_005_neg.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_006_neg.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_007_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_008_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_009_neg.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_features_001_pos.ksh \
functional/cli_root/zpool_wait/cleanup.ksh \
functional/cli_root/zpool_wait/scan/cleanup.ksh \
functional/cli_root/zpool_wait/scan/setup.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_rebuild.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_replace_cancel.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_replace.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_resilver.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_scrub_basic.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_scrub_cancel.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_scrub_flag.ksh \
functional/cli_root/zpool_wait/setup.ksh \
functional/cli_root/zpool_wait/zpool_wait_discard.ksh \
functional/cli_root/zpool_wait/zpool_wait_freeing.ksh \
functional/cli_root/zpool_wait/zpool_wait_initialize_basic.ksh \
functional/cli_root/zpool_wait/zpool_wait_initialize_cancel.ksh \
functional/cli_root/zpool_wait/zpool_wait_initialize_flag.ksh \
functional/cli_root/zpool_wait/zpool_wait_multiple.ksh \
functional/cli_root/zpool_wait/zpool_wait_no_activity.ksh \
functional/cli_root/zpool_wait/zpool_wait_remove_cancel.ksh \
functional/cli_root/zpool_wait/zpool_wait_remove.ksh \
functional/cli_root/zpool_wait/zpool_wait_trim_basic.ksh \
functional/cli_root/zpool_wait/zpool_wait_trim_cancel.ksh \
functional/cli_root/zpool_wait/zpool_wait_trim_flag.ksh \
functional/cli_root/zpool_wait/zpool_wait_usage.ksh \
functional/cli_root/zpool/zpool_001_neg.ksh \
functional/cli_root/zpool/zpool_002_pos.ksh \
functional/cli_root/zpool/zpool_003_pos.ksh \
functional/cli_root/zpool/zpool_colors.ksh \
functional/cli_user/misc/arcstat_001_pos.ksh \
functional/cli_user/misc/arc_summary_001_pos.ksh \
functional/cli_user/misc/arc_summary_002_neg.ksh \
functional/cli_user/misc/zilstat_001_pos.ksh \
functional/cli_user/misc/cleanup.ksh \
functional/cli_user/misc/setup.ksh \
functional/cli_user/misc/zdb_001_neg.ksh \
functional/cli_user/misc/zfs_001_neg.ksh \
functional/cli_user/misc/zfs_allow_001_neg.ksh \
functional/cli_user/misc/zfs_clone_001_neg.ksh \
functional/cli_user/misc/zfs_create_001_neg.ksh \
functional/cli_user/misc/zfs_destroy_001_neg.ksh \
functional/cli_user/misc/zfs_get_001_neg.ksh \
functional/cli_user/misc/zfs_inherit_001_neg.ksh \
functional/cli_user/misc/zfs_mount_001_neg.ksh \
functional/cli_user/misc/zfs_promote_001_neg.ksh \
functional/cli_user/misc/zfs_receive_001_neg.ksh \
functional/cli_user/misc/zfs_rename_001_neg.ksh \
functional/cli_user/misc/zfs_rollback_001_neg.ksh \
functional/cli_user/misc/zfs_send_001_neg.ksh \
functional/cli_user/misc/zfs_set_001_neg.ksh \
functional/cli_user/misc/zfs_share_001_neg.ksh \
functional/cli_user/misc/zfs_snapshot_001_neg.ksh \
functional/cli_user/misc/zfs_unallow_001_neg.ksh \
functional/cli_user/misc/zfs_unmount_001_neg.ksh \
functional/cli_user/misc/zfs_unshare_001_neg.ksh \
functional/cli_user/misc/zfs_upgrade_001_neg.ksh \
functional/cli_user/misc/zpool_001_neg.ksh \
functional/cli_user/misc/zpool_add_001_neg.ksh \
functional/cli_user/misc/zpool_attach_001_neg.ksh \
functional/cli_user/misc/zpool_clear_001_neg.ksh \
functional/cli_user/misc/zpool_create_001_neg.ksh \
functional/cli_user/misc/zpool_destroy_001_neg.ksh \
functional/cli_user/misc/zpool_detach_001_neg.ksh \
functional/cli_user/misc/zpool_export_001_neg.ksh \
functional/cli_user/misc/zpool_get_001_neg.ksh \
functional/cli_user/misc/zpool_history_001_neg.ksh \
functional/cli_user/misc/zpool_import_001_neg.ksh \
functional/cli_user/misc/zpool_import_002_neg.ksh \
functional/cli_user/misc/zpool_offline_001_neg.ksh \
functional/cli_user/misc/zpool_online_001_neg.ksh \
functional/cli_user/misc/zpool_remove_001_neg.ksh \
functional/cli_user/misc/zpool_replace_001_neg.ksh \
functional/cli_user/misc/zpool_scrub_001_neg.ksh \
functional/cli_user/misc/zpool_set_001_neg.ksh \
functional/cli_user/misc/zpool_status_001_neg.ksh \
functional/cli_user/misc/zpool_upgrade_001_neg.ksh \
functional/cli_user/misc/zpool_wait_privilege.ksh \
functional/cli_user/zfs_list/cleanup.ksh \
functional/cli_user/zfs_list/setup.ksh \
functional/cli_user/zfs_list/zfs_list_001_pos.ksh \
functional/cli_user/zfs_list/zfs_list_002_pos.ksh \
functional/cli_user/zfs_list/zfs_list_003_pos.ksh \
functional/cli_user/zfs_list/zfs_list_004_neg.ksh \
functional/cli_user/zfs_list/zfs_list_005_neg.ksh \
functional/cli_user/zfs_list/zfs_list_007_pos.ksh \
functional/cli_user/zfs_list/zfs_list_008_neg.ksh \
functional/cli_user/zpool_iostat/cleanup.ksh \
functional/cli_user/zpool_iostat/setup.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_001_neg.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_002_pos.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_003_neg.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_004_pos.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_005_pos.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_-c_disable.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_-c_homedir.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_-c_searchpath.ksh \
functional/cli_user/zpool_list/cleanup.ksh \
functional/cli_user/zpool_list/setup.ksh \
functional/cli_user/zpool_list/zpool_list_001_pos.ksh \
functional/cli_user/zpool_list/zpool_list_002_neg.ksh \
functional/cli_user/zpool_status/cleanup.ksh \
functional/cli_user/zpool_status/setup.ksh \
functional/cli_user/zpool_status/zpool_status_003_pos.ksh \
functional/cli_user/zpool_status/zpool_status_-c_disable.ksh \
functional/cli_user/zpool_status/zpool_status_-c_homedir.ksh \
functional/cli_user/zpool_status/zpool_status_-c_searchpath.ksh \
functional/compression/cleanup.ksh \
functional/compression/compress_001_pos.ksh \
functional/compression/compress_002_pos.ksh \
functional/compression/compress_003_pos.ksh \
functional/compression/compress_004_pos.ksh \
functional/compression/compress_zstd_bswap.ksh \
functional/compression/l2arc_compressed_arc_disabled.ksh \
functional/compression/l2arc_compressed_arc.ksh \
functional/compression/l2arc_encrypted.ksh \
functional/compression/l2arc_encrypted_no_compressed_arc.ksh \
functional/compression/setup.ksh \
functional/cp_files/cleanup.ksh \
functional/cp_files/cp_files_001_pos.ksh \
functional/cp_files/setup.ksh \
functional/crtime/cleanup.ksh \
functional/crtime/crtime_001_pos.ksh \
functional/crtime/setup.ksh \
functional/ctime/cleanup.ksh \
functional/ctime/ctime_001_pos.ksh \
functional/ctime/setup.ksh \
functional/deadman/deadman_ratelimit.ksh \
functional/deadman/deadman_sync.ksh \
functional/deadman/deadman_zio.ksh \
functional/delegate/cleanup.ksh \
functional/delegate/setup.ksh \
functional/delegate/zfs_allow_001_pos.ksh \
functional/delegate/zfs_allow_002_pos.ksh \
functional/delegate/zfs_allow_003_pos.ksh \
functional/delegate/zfs_allow_004_pos.ksh \
functional/delegate/zfs_allow_005_pos.ksh \
functional/delegate/zfs_allow_006_pos.ksh \
functional/delegate/zfs_allow_007_pos.ksh \
functional/delegate/zfs_allow_008_pos.ksh \
functional/delegate/zfs_allow_009_neg.ksh \
functional/delegate/zfs_allow_010_pos.ksh \
functional/delegate/zfs_allow_011_neg.ksh \
functional/delegate/zfs_allow_012_neg.ksh \
functional/delegate/zfs_unallow_001_pos.ksh \
functional/delegate/zfs_unallow_002_pos.ksh \
functional/delegate/zfs_unallow_003_pos.ksh \
functional/delegate/zfs_unallow_004_pos.ksh \
functional/delegate/zfs_unallow_005_pos.ksh \
functional/delegate/zfs_unallow_006_pos.ksh \
functional/delegate/zfs_unallow_007_neg.ksh \
functional/delegate/zfs_unallow_008_neg.ksh \
functional/devices/cleanup.ksh \
functional/devices/devices_001_pos.ksh \
functional/devices/devices_002_neg.ksh \
functional/devices/devices_003_pos.ksh \
functional/devices/setup.ksh \
functional/dos_attributes/cleanup.ksh \
functional/dos_attributes/read_dos_attrs_001.ksh \
functional/dos_attributes/setup.ksh \
functional/dos_attributes/write_dos_attrs_001.ksh \
functional/events/cleanup.ksh \
functional/events/events_001_pos.ksh \
functional/events/events_002_pos.ksh \
functional/events/setup.ksh \
functional/events/zed_cksum_config.ksh \
functional/events/zed_cksum_reported.ksh \
functional/events/zed_fd_spill.ksh \
functional/events/zed_io_config.ksh \
functional/events/zed_rc_filter.ksh \
functional/exec/cleanup.ksh \
functional/exec/exec_001_pos.ksh \
functional/exec/exec_002_neg.ksh \
functional/exec/setup.ksh \
functional/fadvise/cleanup.ksh \
functional/fadvise/fadvise_sequential.ksh \
functional/fadvise/setup.ksh \
functional/fallocate/cleanup.ksh \
functional/fallocate/fallocate_prealloc.ksh \
functional/fallocate/fallocate_punch-hole.ksh \
functional/fallocate/fallocate_zero-range.ksh \
functional/fallocate/setup.ksh \
functional/fault/auto_offline_001_pos.ksh \
functional/fault/auto_online_001_pos.ksh \
functional/fault/auto_online_002_pos.ksh \
functional/fault/auto_replace_001_pos.ksh \
functional/fault/auto_replace_002_pos.ksh \
functional/fault/auto_spare_001_pos.ksh \
functional/fault/auto_spare_002_pos.ksh \
functional/fault/auto_spare_ashift.ksh \
functional/fault/auto_spare_multiple.ksh \
functional/fault/auto_spare_shared.ksh \
functional/fault/cleanup.ksh \
functional/fault/decompress_fault.ksh \
functional/fault/decrypt_fault.ksh \
functional/fault/scrub_after_resilver.ksh \
functional/fault/setup.ksh \
functional/fault/zpool_status_-s.ksh \
functional/features/async_destroy/async_destroy_001_pos.ksh \
functional/features/async_destroy/cleanup.ksh \
functional/features/async_destroy/setup.ksh \
functional/features/large_dnode/cleanup.ksh \
functional/features/large_dnode/large_dnode_001_pos.ksh \
functional/features/large_dnode/large_dnode_002_pos.ksh \
functional/features/large_dnode/large_dnode_003_pos.ksh \
functional/features/large_dnode/large_dnode_004_neg.ksh \
functional/features/large_dnode/large_dnode_005_pos.ksh \
functional/features/large_dnode/large_dnode_006_pos.ksh \
functional/features/large_dnode/large_dnode_007_neg.ksh \
functional/features/large_dnode/large_dnode_008_pos.ksh \
functional/features/large_dnode/large_dnode_009_pos.ksh \
functional/features/large_dnode/setup.ksh \
functional/grow/grow_pool_001_pos.ksh \
functional/grow/grow_replicas_001_pos.ksh \
functional/history/cleanup.ksh \
functional/history/history_001_pos.ksh \
functional/history/history_002_pos.ksh \
functional/history/history_003_pos.ksh \
functional/history/history_004_pos.ksh \
functional/history/history_005_neg.ksh \
functional/history/history_006_neg.ksh \
functional/history/history_007_pos.ksh \
functional/history/history_008_pos.ksh \
functional/history/history_009_pos.ksh \
functional/history/history_010_pos.ksh \
functional/history/setup.ksh \
functional/inheritance/cleanup.ksh \
functional/inheritance/inherit_001_pos.ksh \
functional/inuse/inuse_001_pos.ksh \
functional/inuse/inuse_003_pos.ksh \
functional/inuse/inuse_004_pos.ksh \
functional/inuse/inuse_005_pos.ksh \
functional/inuse/inuse_006_pos.ksh \
functional/inuse/inuse_007_pos.ksh \
functional/inuse/inuse_008_pos.ksh \
functional/inuse/inuse_009_pos.ksh \
functional/inuse/setup.ksh \
functional/io/cleanup.ksh \
functional/io/io_uring.ksh \
functional/io/libaio.ksh \
functional/io/mmap.ksh \
functional/io/posixaio.ksh \
functional/io/psync.ksh \
functional/io/setup.ksh \
functional/io/sync.ksh \
functional/l2arc/cleanup.ksh \
functional/l2arc/l2arc_arcstats_pos.ksh \
functional/l2arc/l2arc_l2miss_pos.ksh \
functional/l2arc/l2arc_mfuonly_pos.ksh \
functional/l2arc/persist_l2arc_001_pos.ksh \
functional/l2arc/persist_l2arc_002_pos.ksh \
functional/l2arc/persist_l2arc_003_neg.ksh \
functional/l2arc/persist_l2arc_004_pos.ksh \
functional/l2arc/persist_l2arc_005_pos.ksh \
functional/l2arc/setup.ksh \
functional/large_files/cleanup.ksh \
functional/large_files/large_files_001_pos.ksh \
functional/large_files/large_files_002_pos.ksh \
functional/large_files/setup.ksh \
functional/largest_pool/largest_pool_001_pos.ksh \
functional/libzfs/cleanup.ksh \
functional/libzfs/libzfs_input.ksh \
functional/libzfs/setup.ksh \
functional/limits/cleanup.ksh \
functional/limits/filesystem_count.ksh \
functional/limits/filesystem_limit.ksh \
functional/limits/setup.ksh \
functional/limits/snapshot_count.ksh \
functional/limits/snapshot_limit.ksh \
functional/link_count/cleanup.ksh \
functional/link_count/link_count_001.ksh \
functional/link_count/link_count_root_inode.ksh \
functional/link_count/setup.ksh \
functional/log_spacemap/log_spacemap_import_logs.ksh \
functional/migration/cleanup.ksh \
functional/migration/migration_001_pos.ksh \
functional/migration/migration_002_pos.ksh \
functional/migration/migration_003_pos.ksh \
functional/migration/migration_004_pos.ksh \
functional/migration/migration_005_pos.ksh \
functional/migration/migration_006_pos.ksh \
functional/migration/migration_007_pos.ksh \
functional/migration/migration_008_pos.ksh \
functional/migration/migration_009_pos.ksh \
functional/migration/migration_010_pos.ksh \
functional/migration/migration_011_pos.ksh \
functional/migration/migration_012_pos.ksh \
functional/migration/setup.ksh \
functional/mmap/cleanup.ksh \
functional/mmap/mmap_libaio_001_pos.ksh \
functional/mmap/mmap_mixed.ksh \
functional/mmap/mmap_read_001_pos.ksh \
functional/mmap/mmap_seek_001_pos.ksh \
functional/mmap/mmap_sync_001_pos.ksh \
functional/mmap/mmap_write_001_pos.ksh \
functional/mmap/setup.ksh \
functional/mmp/cleanup.ksh \
functional/mmp/mmp_active_import.ksh \
functional/mmp/mmp_exported_import.ksh \
functional/mmp/mmp_hostid.ksh \
functional/mmp/mmp_inactive_import.ksh \
functional/mmp/mmp_interval.ksh \
functional/mmp/mmp_on_off.ksh \
functional/mmp/mmp_on_thread.ksh \
functional/mmp/mmp_on_uberblocks.ksh \
functional/mmp/mmp_on_zdb.ksh \
functional/mmp/mmp_reset_interval.ksh \
functional/mmp/mmp_write_distribution.ksh \
functional/mmp/mmp_write_uberblocks.ksh \
functional/mmp/multihost_history.ksh \
functional/mmp/setup.ksh \
functional/mount/cleanup.ksh \
functional/mount/setup.ksh \
functional/mount/umount_001.ksh \
functional/mount/umountall_001.ksh \
functional/mount/umount_unlinked_drain.ksh \
functional/mv_files/cleanup.ksh \
functional/mv_files/mv_files_001_pos.ksh \
functional/mv_files/mv_files_002_pos.ksh \
functional/mv_files/random_creation.ksh \
functional/mv_files/setup.ksh \
functional/nestedfs/cleanup.ksh \
functional/nestedfs/nestedfs_001_pos.ksh \
functional/nestedfs/setup.ksh \
functional/nopwrite/cleanup.ksh \
functional/nopwrite/nopwrite_copies.ksh \
functional/nopwrite/nopwrite_mtime.ksh \
functional/nopwrite/nopwrite_negative.ksh \
functional/nopwrite/nopwrite_promoted_clone.ksh \
functional/nopwrite/nopwrite_recsize.ksh \
functional/nopwrite/nopwrite_sync.ksh \
functional/nopwrite/nopwrite_varying_compression.ksh \
functional/nopwrite/nopwrite_volume.ksh \
functional/nopwrite/setup.ksh \
functional/no_space/cleanup.ksh \
functional/no_space/enospc_001_pos.ksh \
functional/no_space/enospc_002_pos.ksh \
functional/no_space/enospc_003_pos.ksh \
functional/no_space/enospc_df.ksh \
functional/no_space/enospc_ganging.ksh \
functional/no_space/enospc_rm.ksh \
functional/no_space/setup.ksh \
functional/online_offline/cleanup.ksh \
functional/online_offline/online_offline_001_pos.ksh \
functional/online_offline/online_offline_002_neg.ksh \
functional/online_offline/online_offline_003_neg.ksh \
functional/online_offline/setup.ksh \
functional/pam/cleanup.ksh \
functional/pam/pam_basic.ksh \
functional/pam/pam_change_unmounted.ksh \
functional/pam/pam_nounmount.ksh \
functional/pam/pam_recursive.ksh \
functional/pam/pam_short_password.ksh \
functional/pam/setup.ksh \
functional/pool_checkpoint/checkpoint_after_rewind.ksh \
functional/pool_checkpoint/checkpoint_big_rewind.ksh \
functional/pool_checkpoint/checkpoint_capacity.ksh \
functional/pool_checkpoint/checkpoint_conf_change.ksh \
functional/pool_checkpoint/checkpoint_discard_busy.ksh \
functional/pool_checkpoint/checkpoint_discard.ksh \
functional/pool_checkpoint/checkpoint_discard_many.ksh \
functional/pool_checkpoint/checkpoint_indirect.ksh \
functional/pool_checkpoint/checkpoint_invalid.ksh \
functional/pool_checkpoint/checkpoint_lun_expsz.ksh \
functional/pool_checkpoint/checkpoint_open.ksh \
functional/pool_checkpoint/checkpoint_removal.ksh \
functional/pool_checkpoint/checkpoint_rewind.ksh \
functional/pool_checkpoint/checkpoint_ro_rewind.ksh \
functional/pool_checkpoint/checkpoint_sm_scale.ksh \
functional/pool_checkpoint/checkpoint_twice.ksh \
functional/pool_checkpoint/checkpoint_vdev_add.ksh \
functional/pool_checkpoint/checkpoint_zdb.ksh \
functional/pool_checkpoint/checkpoint_zhack_feat.ksh \
functional/pool_checkpoint/cleanup.ksh \
functional/pool_checkpoint/setup.ksh \
functional/pool_names/pool_names_001_pos.ksh \
functional/pool_names/pool_names_002_neg.ksh \
functional/poolversion/cleanup.ksh \
functional/poolversion/poolversion_001_pos.ksh \
functional/poolversion/poolversion_002_pos.ksh \
functional/poolversion/setup.ksh \
functional/privilege/cleanup.ksh \
functional/privilege/privilege_001_pos.ksh \
functional/privilege/privilege_002_pos.ksh \
functional/privilege/setup.ksh \
functional/procfs/cleanup.ksh \
functional/procfs/pool_state.ksh \
functional/procfs/procfs_list_basic.ksh \
functional/procfs/procfs_list_concurrent_readers.ksh \
functional/procfs/procfs_list_stale_read.ksh \
functional/procfs/setup.ksh \
functional/projectquota/cleanup.ksh \
functional/projectquota/projectid_001_pos.ksh \
functional/projectquota/projectid_002_pos.ksh \
functional/projectquota/projectid_003_pos.ksh \
functional/projectquota/projectquota_001_pos.ksh \
functional/projectquota/projectquota_002_pos.ksh \
functional/projectquota/projectquota_003_pos.ksh \
functional/projectquota/projectquota_004_neg.ksh \
functional/projectquota/projectquota_005_pos.ksh \
functional/projectquota/projectquota_006_pos.ksh \
functional/projectquota/projectquota_007_pos.ksh \
functional/projectquota/projectquota_008_pos.ksh \
functional/projectquota/projectquota_009_pos.ksh \
functional/projectquota/projectspace_001_pos.ksh \
functional/projectquota/projectspace_002_pos.ksh \
functional/projectquota/projectspace_003_pos.ksh \
functional/projectquota/projectspace_004_pos.ksh \
functional/projectquota/projecttree_001_pos.ksh \
functional/projectquota/projecttree_002_pos.ksh \
functional/projectquota/projecttree_003_neg.ksh \
functional/projectquota/setup.ksh \
functional/quota/cleanup.ksh \
functional/quota/quota_001_pos.ksh \
functional/quota/quota_002_pos.ksh \
functional/quota/quota_003_pos.ksh \
functional/quota/quota_004_pos.ksh \
functional/quota/quota_005_pos.ksh \
functional/quota/quota_006_neg.ksh \
functional/quota/setup.ksh \
functional/raidz/cleanup.ksh \
functional/raidz/raidz_001_neg.ksh \
functional/raidz/raidz_002_pos.ksh \
- functional/raidz/raidz_003_pos.ksh \
- functional/raidz/raidz_004_pos.ksh \
+ functional/raidz/raidz_expand_001_pos.ksh \
+ functional/raidz/raidz_expand_002_pos.ksh \
+ functional/raidz/raidz_expand_003_neg.ksh \
+ functional/raidz/raidz_expand_003_pos.ksh \
+ functional/raidz/raidz_expand_004_pos.ksh \
+ functional/raidz/raidz_expand_005_pos.ksh \
+ functional/raidz/raidz_expand_006_neg.ksh \
+ functional/raidz/raidz_expand_007_neg.ksh \
functional/raidz/setup.ksh \
functional/redacted_send/cleanup.ksh \
functional/redacted_send/redacted_compressed.ksh \
functional/redacted_send/redacted_contents.ksh \
functional/redacted_send/redacted_deleted.ksh \
functional/redacted_send/redacted_disabled_feature.ksh \
functional/redacted_send/redacted_embedded.ksh \
functional/redacted_send/redacted_holes.ksh \
functional/redacted_send/redacted_incrementals.ksh \
functional/redacted_send/redacted_largeblocks.ksh \
functional/redacted_send/redacted_many_clones.ksh \
functional/redacted_send/redacted_mixed_recsize.ksh \
functional/redacted_send/redacted_mounts.ksh \
functional/redacted_send/redacted_negative.ksh \
functional/redacted_send/redacted_origin.ksh \
functional/redacted_send/redacted_panic.ksh \
functional/redacted_send/redacted_props.ksh \
functional/redacted_send/redacted_resume.ksh \
functional/redacted_send/redacted_size.ksh \
functional/redacted_send/redacted_volume.ksh \
functional/redacted_send/setup.ksh \
functional/redundancy/cleanup.ksh \
functional/redundancy/redundancy_draid1.ksh \
functional/redundancy/redundancy_draid2.ksh \
functional/redundancy/redundancy_draid3.ksh \
functional/redundancy/redundancy_draid_damaged1.ksh \
functional/redundancy/redundancy_draid_damaged2.ksh \
functional/redundancy/redundancy_draid.ksh \
functional/redundancy/redundancy_draid_spare1.ksh \
functional/redundancy/redundancy_draid_spare2.ksh \
functional/redundancy/redundancy_draid_spare3.ksh \
functional/redundancy/redundancy_mirror.ksh \
functional/redundancy/redundancy_raidz1.ksh \
functional/redundancy/redundancy_raidz2.ksh \
functional/redundancy/redundancy_raidz3.ksh \
functional/redundancy/redundancy_raidz.ksh \
functional/redundancy/redundancy_stripe.ksh \
functional/redundancy/setup.ksh \
functional/refquota/cleanup.ksh \
functional/refquota/refquota_001_pos.ksh \
functional/refquota/refquota_002_pos.ksh \
functional/refquota/refquota_003_pos.ksh \
functional/refquota/refquota_004_pos.ksh \
functional/refquota/refquota_005_pos.ksh \
functional/refquota/refquota_006_neg.ksh \
functional/refquota/refquota_007_neg.ksh \
functional/refquota/refquota_008_neg.ksh \
functional/refquota/setup.ksh \
functional/refreserv/cleanup.ksh \
functional/refreserv/refreserv_001_pos.ksh \
functional/refreserv/refreserv_002_pos.ksh \
functional/refreserv/refreserv_003_pos.ksh \
functional/refreserv/refreserv_004_pos.ksh \
functional/refreserv/refreserv_005_pos.ksh \
functional/refreserv/refreserv_multi_raidz.ksh \
functional/refreserv/refreserv_raidz.ksh \
functional/refreserv/setup.ksh \
functional/removal/cleanup.ksh \
functional/removal/removal_all_vdev.ksh \
functional/removal/removal_cancel.ksh \
functional/removal/removal_check_space.ksh \
functional/removal/removal_condense_export.ksh \
functional/removal/removal_multiple_indirection.ksh \
functional/removal/removal_nopwrite.ksh \
functional/removal/removal_remap_deadlists.ksh \
functional/removal/removal_reservation.ksh \
functional/removal/removal_resume_export.ksh \
functional/removal/removal_sanity.ksh \
functional/removal/removal_with_add.ksh \
functional/removal/removal_with_create_fs.ksh \
functional/removal/removal_with_dedup.ksh \
functional/removal/removal_with_errors.ksh \
functional/removal/removal_with_export.ksh \
functional/removal/removal_with_faulted.ksh \
functional/removal/removal_with_ganging.ksh \
functional/removal/removal_with_indirect.ksh \
functional/removal/removal_with_remove.ksh \
functional/removal/removal_with_scrub.ksh \
functional/removal/removal_with_send.ksh \
functional/removal/removal_with_send_recv.ksh \
functional/removal/removal_with_snapshot.ksh \
functional/removal/removal_with_write.ksh \
functional/removal/removal_with_zdb.ksh \
functional/removal/remove_attach_mirror.ksh \
functional/removal/remove_expanded.ksh \
functional/removal/remove_indirect.ksh \
functional/removal/remove_mirror.ksh \
functional/removal/remove_mirror_sanity.ksh \
functional/removal/remove_raidz.ksh \
functional/rename_dirs/cleanup.ksh \
functional/rename_dirs/rename_dirs_001_pos.ksh \
functional/rename_dirs/setup.ksh \
functional/renameat2/cleanup.ksh \
functional/renameat2/setup.ksh \
functional/renameat2/renameat2_exchange.ksh \
functional/renameat2/renameat2_noreplace.ksh \
functional/renameat2/renameat2_whiteout.ksh \
functional/replacement/attach_import.ksh \
functional/replacement/attach_multiple.ksh \
functional/replacement/attach_rebuild.ksh \
functional/replacement/attach_resilver.ksh \
functional/replacement/cleanup.ksh \
functional/replacement/detach.ksh \
functional/replacement/rebuild_disabled_feature.ksh \
functional/replacement/rebuild_multiple.ksh \
functional/replacement/rebuild_raidz.ksh \
functional/replacement/replace_import.ksh \
functional/replacement/replace_rebuild.ksh \
functional/replacement/replace_resilver.ksh \
functional/replacement/resilver_restart_001.ksh \
functional/replacement/resilver_restart_002.ksh \
functional/replacement/scrub_cancel.ksh \
functional/replacement/setup.ksh \
functional/reservation/cleanup.ksh \
functional/reservation/reservation_001_pos.ksh \
functional/reservation/reservation_002_pos.ksh \
functional/reservation/reservation_003_pos.ksh \
functional/reservation/reservation_004_pos.ksh \
functional/reservation/reservation_005_pos.ksh \
functional/reservation/reservation_006_pos.ksh \
functional/reservation/reservation_007_pos.ksh \
functional/reservation/reservation_008_pos.ksh \
functional/reservation/reservation_009_pos.ksh \
functional/reservation/reservation_010_pos.ksh \
functional/reservation/reservation_011_pos.ksh \
functional/reservation/reservation_012_pos.ksh \
functional/reservation/reservation_013_pos.ksh \
functional/reservation/reservation_014_pos.ksh \
functional/reservation/reservation_015_pos.ksh \
functional/reservation/reservation_016_pos.ksh \
functional/reservation/reservation_017_pos.ksh \
functional/reservation/reservation_018_pos.ksh \
functional/reservation/reservation_019_pos.ksh \
functional/reservation/reservation_020_pos.ksh \
functional/reservation/reservation_021_neg.ksh \
functional/reservation/reservation_022_pos.ksh \
functional/reservation/setup.ksh \
functional/rootpool/cleanup.ksh \
functional/rootpool/rootpool_002_neg.ksh \
functional/rootpool/rootpool_003_neg.ksh \
functional/rootpool/rootpool_007_pos.ksh \
functional/rootpool/setup.ksh \
functional/rsend/cleanup.ksh \
functional/rsend/recv_dedup_encrypted_zvol.ksh \
functional/rsend/recv_dedup.ksh \
functional/rsend/rsend_001_pos.ksh \
functional/rsend/rsend_002_pos.ksh \
functional/rsend/rsend_003_pos.ksh \
functional/rsend/rsend_004_pos.ksh \
functional/rsend/rsend_005_pos.ksh \
functional/rsend/rsend_006_pos.ksh \
functional/rsend/rsend_007_pos.ksh \
functional/rsend/rsend_008_pos.ksh \
functional/rsend/rsend_009_pos.ksh \
functional/rsend/rsend_010_pos.ksh \
functional/rsend/rsend_011_pos.ksh \
functional/rsend/rsend_012_pos.ksh \
functional/rsend/rsend_013_pos.ksh \
functional/rsend/rsend_014_pos.ksh \
functional/rsend/rsend_016_neg.ksh \
functional/rsend/rsend_019_pos.ksh \
functional/rsend/rsend_020_pos.ksh \
functional/rsend/rsend_021_pos.ksh \
functional/rsend/rsend_022_pos.ksh \
functional/rsend/rsend_024_pos.ksh \
functional/rsend/rsend_025_pos.ksh \
functional/rsend/rsend_026_neg.ksh \
functional/rsend/rsend_027_pos.ksh \
functional/rsend/rsend_028_neg.ksh \
functional/rsend/rsend_029_neg.ksh \
functional/rsend/rsend_030_pos.ksh \
functional/rsend/rsend_031_pos.ksh \
functional/rsend/send-c_embedded_blocks.ksh \
functional/rsend/send-c_incremental.ksh \
functional/rsend/send-c_lz4_disabled.ksh \
functional/rsend/send-c_mixed_compression.ksh \
functional/rsend/send-c_props.ksh \
functional/rsend/send-c_recv_dedup.ksh \
functional/rsend/send-c_recv_lz4_disabled.ksh \
functional/rsend/send-c_resume.ksh \
functional/rsend/send-c_stream_size_estimate.ksh \
functional/rsend/send-c_verify_contents.ksh \
functional/rsend/send-c_verify_ratio.ksh \
functional/rsend/send-c_volume.ksh \
functional/rsend/send-c_zstream_recompress.ksh \
functional/rsend/send-c_zstreamdump.ksh \
functional/rsend/send-cpL_varied_recsize.ksh \
functional/rsend/send_doall.ksh \
functional/rsend/send_encrypted_incremental.ksh \
functional/rsend/send_encrypted_files.ksh \
functional/rsend/send_encrypted_freeobjects.ksh \
functional/rsend/send_encrypted_hierarchy.ksh \
functional/rsend/send_encrypted_props.ksh \
functional/rsend/send_encrypted_truncated_files.ksh \
functional/rsend/send_freeobjects.ksh \
functional/rsend/send_holds.ksh \
functional/rsend/send_hole_birth.ksh \
functional/rsend/send_invalid.ksh \
functional/rsend/send-L_toggle.ksh \
functional/rsend/send_mixed_raw.ksh \
functional/rsend/send_partial_dataset.ksh \
functional/rsend/send_raw_ashift.ksh \
functional/rsend/send_raw_spill_block.ksh \
functional/rsend/send_raw_large_blocks.ksh \
functional/rsend/send_realloc_dnode_size.ksh \
functional/rsend/send_realloc_encrypted_files.ksh \
functional/rsend/send_realloc_files.ksh \
functional/rsend/send_spill_block.ksh \
functional/rsend/send-wR_encrypted_zvol.ksh \
functional/rsend/setup.ksh \
functional/scrub_mirror/cleanup.ksh \
functional/scrub_mirror/scrub_mirror_001_pos.ksh \
functional/scrub_mirror/scrub_mirror_002_pos.ksh \
functional/scrub_mirror/scrub_mirror_003_pos.ksh \
functional/scrub_mirror/scrub_mirror_004_pos.ksh \
functional/scrub_mirror/setup.ksh \
functional/slog/cleanup.ksh \
functional/slog/setup.ksh \
functional/slog/slog_001_pos.ksh \
functional/slog/slog_002_pos.ksh \
functional/slog/slog_003_pos.ksh \
functional/slog/slog_004_pos.ksh \
functional/slog/slog_005_pos.ksh \
functional/slog/slog_006_pos.ksh \
functional/slog/slog_007_pos.ksh \
functional/slog/slog_008_neg.ksh \
functional/slog/slog_009_neg.ksh \
functional/slog/slog_010_neg.ksh \
functional/slog/slog_011_neg.ksh \
functional/slog/slog_012_neg.ksh \
functional/slog/slog_013_pos.ksh \
functional/slog/slog_014_pos.ksh \
functional/slog/slog_015_neg.ksh \
functional/slog/slog_016_pos.ksh \
functional/slog/slog_replay_fs_001.ksh \
functional/slog/slog_replay_fs_002.ksh \
functional/slog/slog_replay_volume.ksh \
functional/snapshot/cleanup.ksh \
functional/snapshot/clone_001_pos.ksh \
functional/snapshot/rollback_001_pos.ksh \
functional/snapshot/rollback_002_pos.ksh \
functional/snapshot/rollback_003_pos.ksh \
functional/snapshot/setup.ksh \
functional/snapshot/snapshot_001_pos.ksh \
functional/snapshot/snapshot_002_pos.ksh \
functional/snapshot/snapshot_003_pos.ksh \
functional/snapshot/snapshot_004_pos.ksh \
functional/snapshot/snapshot_005_pos.ksh \
functional/snapshot/snapshot_006_pos.ksh \
functional/snapshot/snapshot_007_pos.ksh \
functional/snapshot/snapshot_008_pos.ksh \
functional/snapshot/snapshot_009_pos.ksh \
functional/snapshot/snapshot_010_pos.ksh \
functional/snapshot/snapshot_011_pos.ksh \
functional/snapshot/snapshot_012_pos.ksh \
functional/snapshot/snapshot_013_pos.ksh \
functional/snapshot/snapshot_014_pos.ksh \
functional/snapshot/snapshot_015_pos.ksh \
functional/snapshot/snapshot_016_pos.ksh \
functional/snapshot/snapshot_017_pos.ksh \
functional/snapshot/snapshot_018_pos.ksh \
functional/snapused/cleanup.ksh \
functional/snapused/setup.ksh \
functional/snapused/snapused_001_pos.ksh \
functional/snapused/snapused_002_pos.ksh \
functional/snapused/snapused_003_pos.ksh \
functional/snapused/snapused_004_pos.ksh \
functional/snapused/snapused_005_pos.ksh \
functional/sparse/cleanup.ksh \
functional/sparse/setup.ksh \
functional/sparse/sparse_001_pos.ksh \
functional/stat/cleanup.ksh \
functional/stat/setup.ksh \
functional/stat/stat_001_pos.ksh \
functional/suid/cleanup.ksh \
functional/suid/setup.ksh \
functional/suid/suid_write_to_none.ksh \
functional/suid/suid_write_to_sgid.ksh \
functional/suid/suid_write_to_suid.ksh \
functional/suid/suid_write_to_suid_sgid.ksh \
functional/suid/suid_write_zil_replay.ksh \
functional/trim/autotrim_config.ksh \
functional/trim/autotrim_integrity.ksh \
functional/trim/autotrim_trim_integrity.ksh \
functional/trim/cleanup.ksh \
functional/trim/setup.ksh \
functional/trim/trim_config.ksh \
functional/trim/trim_integrity.ksh \
functional/trim/trim_l2arc.ksh \
functional/truncate/cleanup.ksh \
functional/truncate/setup.ksh \
functional/truncate/truncate_001_pos.ksh \
functional/truncate/truncate_002_pos.ksh \
functional/truncate/truncate_timestamps.ksh \
functional/upgrade/cleanup.ksh \
functional/upgrade/setup.ksh \
functional/upgrade/upgrade_projectquota_001_pos.ksh \
functional/upgrade/upgrade_readonly_pool.ksh \
functional/upgrade/upgrade_userobj_001_pos.ksh \
functional/user_namespace/cleanup.ksh \
functional/user_namespace/setup.ksh \
functional/user_namespace/user_namespace_001.ksh \
functional/user_namespace/user_namespace_002.ksh \
functional/user_namespace/user_namespace_003.ksh \
functional/user_namespace/user_namespace_004.ksh \
functional/userquota/cleanup.ksh \
functional/userquota/groupspace_001_pos.ksh \
functional/userquota/groupspace_002_pos.ksh \
functional/userquota/groupspace_003_pos.ksh \
functional/userquota/setup.ksh \
functional/userquota/userquota_001_pos.ksh \
functional/userquota/userquota_002_pos.ksh \
functional/userquota/userquota_003_pos.ksh \
functional/userquota/userquota_004_pos.ksh \
functional/userquota/userquota_005_neg.ksh \
functional/userquota/userquota_006_pos.ksh \
functional/userquota/userquota_007_pos.ksh \
functional/userquota/userquota_008_pos.ksh \
functional/userquota/userquota_009_pos.ksh \
functional/userquota/userquota_010_pos.ksh \
functional/userquota/userquota_011_pos.ksh \
functional/userquota/userquota_012_neg.ksh \
functional/userquota/userquota_013_pos.ksh \
functional/userquota/userspace_001_pos.ksh \
functional/userquota/userspace_002_pos.ksh \
functional/userquota/userspace_003_pos.ksh \
functional/userquota/userspace_encrypted.ksh \
functional/userquota/userspace_send_encrypted.ksh \
functional/userquota/userspace_encrypted_13709.ksh \
functional/vdev_zaps/cleanup.ksh \
functional/vdev_zaps/setup.ksh \
functional/vdev_zaps/vdev_zaps_001_pos.ksh \
functional/vdev_zaps/vdev_zaps_002_pos.ksh \
functional/vdev_zaps/vdev_zaps_003_pos.ksh \
functional/vdev_zaps/vdev_zaps_004_pos.ksh \
functional/vdev_zaps/vdev_zaps_005_pos.ksh \
functional/vdev_zaps/vdev_zaps_006_pos.ksh \
functional/vdev_zaps/vdev_zaps_007_pos.ksh \
functional/write_dirs/cleanup.ksh \
functional/write_dirs/setup.ksh \
functional/write_dirs/write_dirs_001_pos.ksh \
functional/write_dirs/write_dirs_002_pos.ksh \
functional/xattr/cleanup.ksh \
functional/xattr/setup.ksh \
functional/xattr/xattr_001_pos.ksh \
functional/xattr/xattr_002_neg.ksh \
functional/xattr/xattr_003_neg.ksh \
functional/xattr/xattr_004_pos.ksh \
functional/xattr/xattr_005_pos.ksh \
functional/xattr/xattr_006_pos.ksh \
functional/xattr/xattr_007_neg.ksh \
functional/xattr/xattr_008_pos.ksh \
functional/xattr/xattr_009_neg.ksh \
functional/xattr/xattr_010_neg.ksh \
functional/xattr/xattr_011_pos.ksh \
functional/xattr/xattr_012_pos.ksh \
functional/xattr/xattr_013_pos.ksh \
functional/xattr/xattr_compat.ksh \
functional/zpool_influxdb/cleanup.ksh \
functional/zpool_influxdb/setup.ksh \
functional/zpool_influxdb/zpool_influxdb.ksh \
functional/zvol/zvol_cli/cleanup.ksh \
functional/zvol/zvol_cli/setup.ksh \
functional/zvol/zvol_cli/zvol_cli_001_pos.ksh \
functional/zvol/zvol_cli/zvol_cli_002_pos.ksh \
functional/zvol/zvol_cli/zvol_cli_003_neg.ksh \
functional/zvol/zvol_ENOSPC/cleanup.ksh \
functional/zvol/zvol_ENOSPC/setup.ksh \
functional/zvol/zvol_ENOSPC/zvol_ENOSPC_001_pos.ksh \
functional/zvol/zvol_misc/cleanup.ksh \
functional/zvol/zvol_misc/setup.ksh \
functional/zvol/zvol_misc/zvol_misc_001_neg.ksh \
functional/zvol/zvol_misc/zvol_misc_002_pos.ksh \
functional/zvol/zvol_misc/zvol_misc_003_neg.ksh \
functional/zvol/zvol_misc/zvol_misc_004_pos.ksh \
functional/zvol/zvol_misc/zvol_misc_005_neg.ksh \
functional/zvol/zvol_misc/zvol_misc_006_pos.ksh \
functional/zvol/zvol_misc/zvol_misc_fua.ksh \
functional/zvol/zvol_misc/zvol_misc_hierarchy.ksh \
functional/zvol/zvol_misc/zvol_misc_rename_inuse.ksh \
functional/zvol/zvol_misc/zvol_misc_snapdev.ksh \
functional/zvol/zvol_misc/zvol_misc_trim.ksh \
functional/zvol/zvol_misc/zvol_misc_volmode.ksh \
functional/zvol/zvol_misc/zvol_misc_zil.ksh \
functional/zvol/zvol_stress/cleanup.ksh \
functional/zvol/zvol_stress/setup.ksh \
functional/zvol/zvol_stress/zvol_stress.ksh \
functional/zvol/zvol_swap/cleanup.ksh \
functional/zvol/zvol_swap/setup.ksh \
functional/zvol/zvol_swap/zvol_swap_001_pos.ksh \
functional/zvol/zvol_swap/zvol_swap_002_pos.ksh \
functional/zvol/zvol_swap/zvol_swap_003_pos.ksh \
functional/zvol/zvol_swap/zvol_swap_004_pos.ksh \
functional/zvol/zvol_swap/zvol_swap_005_pos.ksh \
functional/zvol/zvol_swap/zvol_swap_006_pos.ksh \
functional/idmap_mount/cleanup.ksh \
functional/idmap_mount/setup.ksh \
functional/idmap_mount/idmap_mount_001.ksh \
functional/idmap_mount/idmap_mount_002.ksh \
functional/idmap_mount/idmap_mount_003.ksh \
functional/idmap_mount/idmap_mount_004.ksh \
functional/idmap_mount/idmap_mount_005.ksh
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/zpool_get.cfg b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/zpool_get.cfg
index 4248578cde16..6ebce9459190 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/zpool_get.cfg
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/zpool_get.cfg
@@ -1,110 +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 https://opensource.org/licenses/CDDL-1.0.
# See the License for the specific language governing permissions
# and limitations under the License.
#
# When distributing Covered Code, include this CDDL HEADER in each
# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
# If applicable, add the following below this CDDL HEADER, with the
# fields enclosed by brackets "[]" replaced with your own identifying
# information: Portions Copyright [yyyy] [name of copyright owner]
#
# CDDL HEADER END
#
#
# Copyright 2009 Sun Microsystems, Inc. All rights reserved.
# Use is subject to license terms.
#
#
# Copyright (c) 2013, 2014 by Delphix. All rights reserved.
# Copyright 2016 Nexenta Systems, Inc. All rights reserved.
#
# Set the expected properties of zpool
typeset -a properties=(
"size"
"capacity"
"altroot"
"health"
"guid"
"load_guid"
"version"
"bootfs"
"delegation"
"autoreplace"
"cachefile"
"checkpoint"
"failmode"
"listsnapshots"
"autoexpand"
"dedupratio"
"free"
"allocated"
"readonly"
"comment"
"expandsize"
"freeing"
"fragmentation"
"leaked"
"multihost"
"autotrim"
"compatibility"
"bcloneused"
"bclonesaved"
"bcloneratio"
"feature@async_destroy"
"feature@empty_bpobj"
"feature@lz4_compress"
"feature@multi_vdev_crash_dump"
"feature@spacemap_histogram"
"feature@enabled_txg"
"feature@hole_birth"
"feature@extensible_dataset"
"feature@embedded_data"
"feature@bookmarks"
"feature@filesystem_limits"
"feature@large_blocks"
"feature@sha512"
"feature@skein"
"feature@edonr"
"feature@device_removal"
"feature@obsolete_counts"
"feature@zpool_checkpoint"
"feature@spacemap_v2"
"feature@redaction_bookmarks"
"feature@redacted_datasets"
"feature@bookmark_written"
"feature@log_spacemap"
"feature@device_rebuild"
"feature@draid"
"feature@redaction_list_spill"
)
if is_linux || is_freebsd; then
properties+=(
"ashift"
"feature@large_dnode"
"feature@userobj_accounting"
"feature@encryption"
"feature@project_quota"
"feature@allocation_classes"
"feature@resilver_defer"
"feature@bookmark_v2"
"feature@livelist"
"feature@zstd_compress"
"feature@zilsaxattr"
"feature@head_errlog"
"feature@blake3"
"feature@block_cloning"
"feature@vdev_zaps_v2"
+ "feature@raidz_expansion"
)
fi
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_002_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_002_pos.ksh
index 746718ad9a58..4bd11a94036d 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_002_pos.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_002_pos.ksh
@@ -1,41 +1,61 @@
#!/bin/ksh -p
#
# CDDL HEADER START
#
# The contents of this file are subject to the terms of the
# Common Development and Distribution License (the "License").
# You may not use this file except in compliance with the License.
#
# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
# or https://opensource.org/licenses/CDDL-1.0.
# See the License for the specific language governing permissions
# and limitations under the License.
#
# When distributing Covered Code, include this CDDL HEADER in each
# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
# If applicable, add the following below this CDDL HEADER, with the
# fields enclosed by brackets "[]" replaced with your own identifying
# information: Portions Copyright [yyyy] [name of copyright owner]
#
# CDDL HEADER END
#
#
# Copyright (c) 2016 by Gvozden Neskovic. All rights reserved.
# Use is subject to license terms.
+# Copyright (c) 2020 by vStack. All rights reserved.
#
. $STF_SUITE/include/libtest.shlib
#
# DESCRIPTION:
-# Call the raidz_test tool with -S to test all supported raidz
-# implementations. This options will test several raidz block geometries
+# Call the raidz_test tool with sweep to test all supported raidz
+# implementations. This will test several raidz block geometries
# and several zio parameters that affect raidz block layout. Data
-# reconstruction performs all combinations of failed disks. Wall time
-# is set to 5min, but actual runtime might be longer.
+# reconstruction performs all combinations of failed disks. Wall
+# time is set to 5 min, but actual runtime might be longer.
#
-log_must raidz_test -S -t 300
-
-log_pass "raidz_test parameter sweep test succeeded."
+case $((RANDOM % 3)) in
+ 0)
+ # Basic sweep test
+ log_must raidz_test -S -t 300
+ log_pass "raidz_test parameter sweep test succeeded."
+ ;;
+ 1)
+ # Using expanded raidz map to test all supported raidz
+ # implementations with expanded map and default reflow offset.
+ log_must raidz_test -S -e -t 300
+ log_pass "raidz_test sweep test with expanded map succeeded."
+ ;;
+ 2)
+ # Using expanded raidz map ('-e') to test all supported raidz
+ # implementations with expanded map and zero reflow offset.
+ log_must raidz_test -S -e -r 0 -t 300
+ log_pass "raidz_test sweep test with expanded map succeeded."
+ ;;
+ *)
+ # avoid shellcheck SC2249
+ ;;
+esac
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_003_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_003_pos.ksh
deleted file mode 100755
index ce44906d5a4f..000000000000
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_003_pos.ksh
+++ /dev/null
@@ -1,41 +0,0 @@
-#!/bin/ksh -p
-#
-# CDDL HEADER START
-#
-# The contents of this file are subject to the terms of the
-# Common Development and Distribution License (the "License").
-# You may not use this file except in compliance with the License.
-#
-# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
-# or https://opensource.org/licenses/CDDL-1.0.
-# See the License for the specific language governing permissions
-# and limitations under the License.
-#
-# When distributing Covered Code, include this CDDL HEADER in each
-# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
-# If applicable, add the following below this CDDL HEADER, with the
-# fields enclosed by brackets "[]" replaced with your own identifying
-# information: Portions Copyright [yyyy] [name of copyright owner]
-#
-# CDDL HEADER END
-#
-
-#
-# Copyright (c) 2020 by vStack. All rights reserved.
-#
-
-. $STF_SUITE/include/libtest.shlib
-
-#
-# DESCRIPTION:
-# Call the raidz_test tool with -S and -e to test all supported raidz
-# implementations with expanded map and default reflow offset.
-# This options will test several raidz block geometries and several zio
-# parameters that affect raidz block layout. Data reconstruction performs
-# all combinations of failed disks. Wall time is set to 5min, but actual
-# runtime might be longer.
-#
-
-log_must raidz_test -S -e -t 60
-
-log_pass "raidz_test parameter sweep test with expanded map succeeded."
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_004_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_004_pos.ksh
deleted file mode 100755
index 0e3affd5143c..000000000000
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_004_pos.ksh
+++ /dev/null
@@ -1,41 +0,0 @@
-#!/bin/ksh -p
-#
-# CDDL HEADER START
-#
-# The contents of this file are subject to the terms of the
-# Common Development and Distribution License (the "License").
-# You may not use this file except in compliance with the License.
-#
-# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
-# or https://opensource.org/licenses/CDDL-1.0.
-# See the License for the specific language governing permissions
-# and limitations under the License.
-#
-# When distributing Covered Code, include this CDDL HEADER in each
-# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
-# If applicable, add the following below this CDDL HEADER, with the
-# fields enclosed by brackets "[]" replaced with your own identifying
-# information: Portions Copyright [yyyy] [name of copyright owner]
-#
-# CDDL HEADER END
-#
-
-#
-# Copyright (c) 2020 by vStack. All rights reserved.
-#
-
-. $STF_SUITE/include/libtest.shlib
-
-#
-# DESCRIPTION:
-# Call the raidz_test tool with -S and -e to test all supported raidz
-# implementations with expanded map and zero reflow offset.
-# This options will test several raidz block geometries and several zio
-# parameters that affect raidz block layout. Data reconstruction performs
-# all combinations of failed disks. Wall time is set to 5min, but actual
-# runtime might be longer.
-#
-
-log_must raidz_test -S -e -r 0 -t 60
-
-log_pass "raidz_test parameter sweep test with expanded map succeeded."
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_001_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_001_pos.ksh
new file mode 100755
index 000000000000..063d7fa735df
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_001_pos.ksh
@@ -0,0 +1,215 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or 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) 2020 by vStack. All rights reserved.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+#
+# DESCRIPTION:
+# 'zpool attach poolname raidz ...' should attach new device to the pool.
+#
+# STRATEGY:
+# 1. Create block device files for the test raidz pool
+# 2. For each parity value [1..3]
+# - create raidz pool
+# - fill it with some directories/files
+# - attach device to the raidz pool
+# - verify that device attached and the raidz pool size increase
+# - verify resilver by replacing parity devices
+# - verify resilver by replacing data devices
+# - verify scrub by zeroing parity devices
+# - verify scrub by zeroing data devices
+# - verify the raidz pool
+# - destroy the raidz pool
+
+typeset -r devs=6
+typeset -r dev_size_mb=128
+
+typeset -a disks
+
+prefetch_disable=$(get_tunable PREFETCH_DISABLE)
+
+function cleanup
+{
+ log_pos zpool status $TESTPOOL
+
+ poolexists "$TESTPOOL" && log_must_busy zpool destroy "$TESTPOOL"
+
+ for i in {0..$devs}; do
+ log_must rm -f "$TEST_BASE_DIR/dev-$i"
+ done
+
+ log_must set_tunable32 PREFETCH_DISABLE $prefetch_disable
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES 0
+}
+
+function wait_expand_paused
+{
+ oldcopied='0'
+ newcopied='1'
+ while [[ $oldcopied != $newcopied ]]; do
+ oldcopied=$newcopied
+ sleep 2
+ newcopied=$(zpool status $TESTPOOL | \
+ grep 'copied out of' | \
+ awk '{print $1}')
+ log_note "newcopied=$newcopied"
+ done
+ log_note "paused at $newcopied"
+}
+
+function test_resilver # <pool> <parity> <dir>
+{
+ typeset pool=$1
+ typeset nparity=$2
+ typeset dir=$3
+
+ for (( i=0; i<$nparity; i=i+1 )); do
+ log_must zpool offline $pool $dir/dev-$i
+ done
+
+ log_must zpool export $pool
+
+ for (( i=0; i<$nparity; i=i+1 )); do
+ log_must zpool labelclear -f $dir/dev-$i
+ done
+
+ log_must zpool import -o cachefile=none -d $dir $pool
+
+ for (( i=0; i<$nparity; i=i+1 )); do
+ log_must zpool replace -f $pool $dir/dev-$i
+ done
+
+ log_must zpool wait -t resilver $pool
+
+ log_must check_pool_status $pool "errors" "No known data errors"
+
+ log_must zpool clear $pool
+
+ for (( i=$nparity; i<$nparity*2; i=i+1 )); do
+ log_must zpool offline $pool $dir/dev-$i
+ done
+
+ log_must zpool export $pool
+
+ for (( i=$nparity; i<$nparity*2; i=i+1 )); do
+ log_must zpool labelclear -f $dir/dev-$i
+ done
+
+ log_must zpool import -o cachefile=none -d $dir $pool
+
+ for (( i=$nparity; i<$nparity*2; i=i+1 )); do
+ log_must zpool replace -f $pool $dir/dev-$i
+ done
+
+ log_must zpool wait -t resilver $pool
+
+ log_must check_pool_status $pool "errors" "No known data errors"
+
+ log_must zpool clear $pool
+}
+
+function test_scrub # <pool> <parity> <dir>
+{
+ typeset pool=$1
+ typeset nparity=$2
+ typeset dir=$3
+ typeset combrec=$4
+
+ reflow_size=$(get_pool_prop allocated $pool)
+ randbyte=$(( ((RANDOM<<15) + RANDOM) % $reflow_size ))
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES $randbyte
+ log_must zpool attach $TESTPOOL ${raid}-0 $dir/dev-$devs
+ wait_expand_paused
+
+ log_must zpool export $pool
+
+ # zero out parity disks
+ for (( i=0; i<$nparity; i=i+1 )); do
+ dd conv=notrunc if=/dev/zero of=$dir/dev-$i \
+ bs=1M seek=4 count=$(($dev_size_mb-4))
+ done
+
+ log_must zpool import -o cachefile=none -d $dir $pool
+
+ log_must zpool scrub -w $pool
+ log_must zpool clear $pool
+ log_must zpool export $pool
+
+ # zero out parity count worth of data disks
+ for (( i=$nparity; i<$nparity*2; i=i+1 )); do
+ dd conv=notrunc if=/dev/zero of=$dir/dev-$i \
+ bs=1M seek=4 count=$(($dev_size_mb-4))
+ done
+
+ log_must zpool import -o cachefile=none -d $dir $pool
+
+ log_must zpool scrub -w $pool
+
+ log_must check_pool_status $pool "errors" "No known data errors"
+
+ log_must zpool clear $pool
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES 0
+ log_must zpool wait -t raidz_expand $TESTPOOL
+}
+
+log_onexit cleanup
+
+log_must set_tunable32 PREFETCH_DISABLE 1
+
+# Disk files which will be used by pool
+for i in {0..$(($devs - 1))}; do
+ device=$TEST_BASE_DIR/dev-$i
+ log_must truncate -s ${dev_size_mb}M $device
+ disks[${#disks[*]}+1]=$device
+done
+
+# Disk file which will be attached
+log_must truncate -s 512M $TEST_BASE_DIR/dev-$devs
+
+nparity=$((RANDOM%(3) + 1))
+raid=raidz$nparity
+dir=$TEST_BASE_DIR
+
+log_must zpool create -f -o cachefile=none $TESTPOOL $raid ${disks[@]}
+log_must zfs set primarycache=metadata $TESTPOOL
+
+log_must zfs create $TESTPOOL/fs
+log_must fill_fs /$TESTPOOL/fs 1 512 100 1024 R
+
+log_must zfs create -o compress=on $TESTPOOL/fs2
+log_must fill_fs /$TESTPOOL/fs2 1 512 100 1024 R
+
+log_must zfs create -o compress=on -o recordsize=8k $TESTPOOL/fs3
+log_must fill_fs /$TESTPOOL/fs3 1 512 100 1024 R
+
+log_must check_pool_status $TESTPOOL "errors" "No known data errors"
+
+test_scrub $TESTPOOL $nparity $dir
+test_resilver $TESTPOOL $nparity $dir
+
+zpool destroy "$TESTPOOL"
+
+log_pass "raidz expansion test succeeded."
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_002_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_002_pos.ksh
new file mode 100755
index 000000000000..004f3d1f9255
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_002_pos.ksh
@@ -0,0 +1,115 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or 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) 2020 by vStack. All rights reserved.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+#
+# DESCRIPTION:
+# 'zpool attach poolname raidz ...' should attach new devive to the pool.
+#
+# STRATEGY:
+# 1. Create block device files for the test raidz pool
+# 2. For each parity value [1..3]
+# - create raidz pool with minimum block device files required
+# - for each free test block device
+# - attach to the pool
+# - verify the raidz pool
+# - destroy the raidz pool
+
+typeset -r devs=6
+typeset -r dev_size_mb=512
+
+typeset -a disks
+
+prefetch_disable=$(get_tunable PREFETCH_DISABLE)
+
+function cleanup
+{
+ poolexists "$TESTPOOL" && log_must_busy zpool destroy "$TESTPOOL"
+
+ for i in {0..$devs}; do
+ log_must rm -f "$TEST_BASE_DIR/dev-$i"
+ done
+
+ log_must set_tunable32 PREFETCH_DISABLE $prefetch_disable
+}
+
+log_onexit cleanup
+
+log_must set_tunable32 PREFETCH_DISABLE 1
+
+# Disk files which will be used by pool
+for i in {0..$(($devs))}; do
+ device=$TEST_BASE_DIR/dev-$i
+ log_must truncate -s ${dev_size_mb}M $device
+ disks[${#disks[*]}+1]=$device
+done
+
+nparity=$((RANDOM%(3) + 1))
+raid=raidz$nparity
+dir=$TEST_BASE_DIR
+pool=$TESTPOOL
+opts="-o cachefile=none"
+
+log_must zpool create -f $opts $pool $raid ${disks[1..$(($nparity+1))]}
+log_must zfs set primarycache=metadata $pool
+
+log_must zfs create $pool/fs
+log_must fill_fs /$pool/fs 1 512 100 1024 R
+
+log_must zfs create -o compress=on $pool/fs2
+log_must fill_fs /$pool/fs2 1 512 100 1024 R
+
+log_must zfs create -o compress=on -o recordsize=8k $pool/fs3
+log_must fill_fs /$pool/fs3 1 512 100 1024 R
+
+typeset pool_size=$(get_pool_prop size $pool)
+
+for disk in ${disks[$(($nparity+2))..$devs]}; do
+ log_must dd if=/dev/urandom of=/${pool}/FILE-$RANDOM bs=1M \
+ count=64
+
+ log_must zpool attach -w $pool ${raid}-0 $disk
+
+ # Wait some time for pool size increase
+ sleep 5
+
+ # Confirm that disk was attached to the pool
+ log_must zpool get -H path $TESTPOOL $disk
+
+ typeset expand_size=$(get_pool_prop size $pool)
+ if [[ "$expand_size" -le "$pool_size" ]]; then
+ log_fail "pool $pool not expanded"
+ fi
+
+ verify_pool $pool
+
+ pool_size=$expand_size
+done
+
+zpool destroy "$pool"
+
+log_pass "raidz expansion test succeeded."
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_003_neg.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_003_neg.ksh
new file mode 100755
index 000000000000..4d85c46897b8
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_003_neg.ksh
@@ -0,0 +1,102 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or 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) 2021 by vStack. All rights reserved.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+#
+# DESCRIPTION:
+# 'zpool attach poolname raidz ...' should reject device attach if pool
+# is in checkpointed state. If checkpoint creation requested on
+# expanding pool, the request should be rejected.
+
+#
+# STRATEGY:
+# 1. Create block device files for the test raidz pool.
+# 2. Create pool and checkpoint it.
+# 3. Try to expand raidz, ensure that request rejected.
+# 4. Recreate the pool.
+# 5. Apply raidz expansion.
+# 6. Ensure, that checkpoint cannot be created.
+
+typeset -r devs=6
+typeset -r dev_size_mb=512
+
+typeset -a disks
+
+prefetch_disable=$(get_tunable PREFETCH_DISABLE)
+
+function cleanup
+{
+ poolexists "$TESTPOOL" && log_must_busy zpool destroy "$TESTPOOL"
+
+ for i in {0..$devs}; do
+ log_must rm -f "$TEST_BASE_DIR/dev-$i"
+ done
+
+ log_must set_tunable32 PREFETCH_DISABLE $prefetch_disable
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES 0
+}
+
+log_onexit cleanup
+
+log_must set_tunable32 PREFETCH_DISABLE 1
+
+# Disk files which will be used by pool
+for i in {0..$(($devs))}; do
+ device=$TEST_BASE_DIR/dev-$i
+ log_must truncate -s ${dev_size_mb}M $device
+ disks[${#disks[*]}+1]=$device
+done
+
+nparity=1
+raid=raidz$nparity
+pool=$TESTPOOL
+opts="-o cachefile=none"
+
+# case 1: checkpoint exist, try to expand
+log_must zpool create -f $opts $pool $raid ${disks[1..$(($devs-1))]}
+log_must zfs set primarycache=metadata $pool
+log_must zpool checkpoint $pool
+log_mustnot zpool attach $pool ${raid}-0 ${disks[$devs]}
+log_must zpool destroy $pool
+
+#
+# case 2: expansion in progress, try to checkpoint
+#
+# Sets pause point at 25% of allocated space so that we know an
+# expansion is still in progress when we attempt the checkpoint
+#
+log_must zpool create -f $opts $pool $raid ${disks[1..$(($devs-1))]}
+log_must zfs set primarycache=metadata $pool
+log_must zfs create $pool/fs
+log_must fill_fs /$pool/fs 1 512 100 1024 R
+allocated=$(zpool list -Hp -o allocated $pool)
+log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES $((allocated / 4))
+log_must zpool attach $pool ${raid}-0 ${disks[$devs]}
+log_mustnot zpool checkpoint $pool
+log_must zpool destroy $pool
+
+log_pass "raidz expansion test succeeded."
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_003_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_003_pos.ksh
new file mode 100755
index 000000000000..712b25261773
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_003_pos.ksh
@@ -0,0 +1,141 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or 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) 2021 by vStack. All rights reserved.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/cli_root/zfs_wait/zfs_wait.kshlib
+
+#
+# DESCRIPTION:
+# Check raidz expansion is able to work correctly under i/o load.
+#
+# STRATEGY:
+# 1. Create block device files for the test raidz pool
+# 2. For each parity value [1..3]
+# - create raidz pool with minimum block device files required
+# - create couple of datasets with different recordsize and fill it
+# - set a max reflow value near pool capacity
+# - wait for reflow to reach this max
+# - verify pool
+# - set reflow bytes to max value to complete the expansion
+
+typeset -r devs=10
+typeset -r dev_size_mb=256
+
+typeset -a disks
+
+embedded_slog_min_ms=$(get_tunable EMBEDDED_SLOG_MIN_MS)
+
+function cleanup
+{
+ poolexists "$TESTPOOL" && zpool status -v "$TESTPOOL"
+ poolexists "$TESTPOOL" && log_must_busy zpool destroy "$TESTPOOL"
+
+ for i in {0..$devs}; do
+ log_must rm -f "$TEST_BASE_DIR/dev-$i"
+ done
+
+ log_must set_tunable32 EMBEDDED_SLOG_MIN_MS $embedded_slog_min_ms
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES 0
+}
+
+function wait_expand_paused
+{
+ oldcopied='0'
+ newcopied='1'
+ # wait until reflow copied value stops changing
+ while [[ $oldcopied != $newcopied ]]; do
+ oldcopied=$newcopied
+ sleep 1
+ newcopied=$(zpool status $TESTPOOL | \
+ grep 'copied out of' | \
+ awk '{print $1}')
+ done
+}
+
+log_onexit cleanup
+
+log_must set_tunable32 EMBEDDED_SLOG_MIN_MS 99999
+
+# Disk files which will be used by pool
+for i in {0..$(($devs))}; do
+ device=$TEST_BASE_DIR/dev-$i
+ log_must truncate -s ${dev_size_mb}M $device
+ disks[${#disks[*]}+1]=$device
+done
+
+nparity=$((RANDOM%(3) + 1))
+raid=raidz$nparity
+pool=$TESTPOOL
+opts="-o cachefile=none"
+
+log_must zpool create -f $opts $pool $raid ${disks[1..$(($nparity+1))]}
+
+log_must zfs create -o recordsize=8k $pool/fs
+log_must fill_fs /$pool/fs 1 256 100 1024 R
+
+log_must zfs create -o recordsize=128k $pool/fs2
+log_must fill_fs /$pool/fs2 1 256 100 1024 R
+
+for disk in ${disks[$(($nparity+2))..$devs]}; do
+ log_must mkfile -n 400m /$pool/fs/file
+ log_bkgrnd randwritecomp /$pool/fs/file 250
+ pid0=$!
+
+ # start some random writes in the background during expansion
+ log_must mkfile -n 400m /$pool/fs2/file2
+ log_bkgrnd randwritecomp /$pool/fs2/file2 250
+ pid1=$!
+ sleep 10
+
+ # Pause at half total bytes to be copied for expansion
+ reflow_size=$(get_pool_prop allocated $pool)
+ log_note need to reflow $reflow_size bytes
+ pause=$((reflow_size/2))
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES $pause
+
+ log_must zpool attach $pool ${raid}-0 $disk
+ wait_expand_paused
+
+ kill_if_running $pid0
+ kill_if_running $pid1
+
+ log_must zpool scrub -w $pool
+
+ log_must check_pool_status $pool "errors" "No known data errors"
+ log_must check_pool_status $pool "scan" "with 0 errors"
+ log_must check_pool_status $pool "scan" "repaired 0B"
+
+ # Set pause past largest possible value for this pool
+ pause=$((devs*dev_size_mb*1024*1024))
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES $pause
+
+ log_must zpool wait -t raidz_expand $pool
+done
+
+log_must zpool destroy "$pool"
+
+log_pass "raidz expansion test succeeded."
+
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_004_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_004_pos.ksh
new file mode 100755
index 000000000000..2be55dae4254
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_004_pos.ksh
@@ -0,0 +1,121 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or 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) 2021 by vStack. All rights reserved.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+#
+# DESCRIPTION:
+# Check device replacement during raidz expansion.
+#
+# STRATEGY:
+# 1. Create block device files for the test raidz pool
+# 2. For each parity value [1..3]
+# - create raidz pool with minimum block device files required
+# - create couple of datasets with different recordsize and fill it
+# - attach new device to the pool
+# - offline and zero vdevs allowed by parity
+# - wait some time and start offlined vdevs replacement
+# - wait replacement completion and verify pool status
+
+typeset -r devs=10
+typeset -r dev_size_mb=128
+
+typeset -a disks
+
+embedded_slog_min_ms=$(get_tunable EMBEDDED_SLOG_MIN_MS)
+original_scrub_after_expand=$(get_tunable SCRUB_AFTER_EXPAND)
+
+function cleanup
+{
+ poolexists "$TESTPOOL" && log_must_busy zpool destroy "$TESTPOOL"
+
+ for i in {0..$devs}; do
+ log_must rm -f "$TEST_BASE_DIR/dev-$i"
+ done
+
+ log_must set_tunable32 EMBEDDED_SLOG_MIN_MS $embedded_slog_min_ms
+ log_must set_tunable32 SCRUB_AFTER_EXPAND $original_scrub_after_expand
+}
+
+log_onexit cleanup
+
+log_must set_tunable32 EMBEDDED_SLOG_MIN_MS 99999
+
+# Disk files which will be used by pool
+for i in {0..$(($devs))}; do
+ device=$TEST_BASE_DIR/dev-$i
+ log_must truncate -s ${dev_size_mb}M $device
+ disks[${#disks[*]}+1]=$device
+done
+
+nparity=$((RANDOM%(3) + 1))
+raid=raidz$nparity
+pool=$TESTPOOL
+opts="-o cachefile=none"
+
+log_must set_tunable32 SCRUB_AFTER_EXPAND 0
+
+log_must zpool create -f $opts $pool $raid ${disks[1..$(($nparity+1))]}
+
+log_must zfs create -o recordsize=8k $pool/fs
+log_must fill_fs /$pool/fs 1 128 100 1024 R
+
+log_must zfs create -o recordsize=128k $pool/fs2
+log_must fill_fs /$pool/fs2 1 128 100 1024 R
+
+for disk in ${disks[$(($nparity+2))..$devs]}; do
+ log_must zpool attach $pool ${raid}-0 $disk
+
+ sleep 10
+
+ for (( i=1; i<=$nparity; i=i+1 )); do
+ log_must zpool offline $pool ${disks[$i]}
+ log_must dd if=/dev/zero of=${disks[$i]} \
+ bs=1024k count=$dev_size_mb conv=notrunc
+ done
+
+ sleep 3
+
+ for (( i=1; i<=$nparity; i=i+1 )); do
+ log_must zpool replace $pool ${disks[$i]}
+ done
+
+ log_must zpool wait -t replace $pool
+ log_must check_pool_status $pool "scan" "with 0 errors"
+
+ log_must zpool wait -t raidz_expand $pool
+
+ log_must zpool clear $pool
+ log_must zpool scrub -w $pool
+
+ log_must zpool status -v
+ log_must check_pool_status $pool "scan" "with 0 errors"
+done
+
+log_must zpool destroy "$pool"
+
+log_pass "raidz expansion test succeeded."
+
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_005_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_005_pos.ksh
new file mode 100755
index 000000000000..a31a7d254bfe
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_005_pos.ksh
@@ -0,0 +1,177 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or 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) 2021 by vStack. All rights reserved.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+#
+# DESCRIPTION:
+# Check device replacement during raidz expansion using expansion pausing.
+#
+# STRATEGY:
+# 1. Create block device files for the test raidz pool
+# 2. For each parity value [1..3]
+# - create raidz pool with minimum block device files required
+# - create couple of datasets with different recordsize and fill it
+# - set raidz expand maximum reflow bytes
+# - attach new device to the pool
+# - wait for reflow bytes to reach the maximum
+# - offline and zero vdevs allowed by parity
+# - wait some time and start offlined vdevs replacement
+# - wait replacement completion and verify pool status
+# - loop thru vdevs replacing with the max reflow bytes increasing
+# - verify pool
+# - set reflow bytes to max value to complete the expansion
+
+typeset -r devs=10
+typeset -r dev_size_mb=128
+
+typeset -a disks
+
+embedded_slog_min_ms=$(get_tunable EMBEDDED_SLOG_MIN_MS)
+original_scrub_after_expand=$(get_tunable SCRUB_AFTER_EXPAND)
+
+function cleanup
+{
+ poolexists "$TESTPOOL" && zpool status -v "$TESTPOOL"
+ poolexists "$TESTPOOL" && log_must_busy zpool destroy "$TESTPOOL"
+
+ for i in {0..$devs}; do
+ log_must rm -f "$TEST_BASE_DIR/dev-$i"
+ done
+
+ log_must set_tunable32 EMBEDDED_SLOG_MIN_MS $embedded_slog_min_ms
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES 0
+ log_must set_tunable32 SCRUB_AFTER_EXPAND $original_scrub_after_expand
+}
+
+function wait_expand_paused
+{
+ oldcopied='0'
+ newcopied='1'
+ while [[ $oldcopied != $newcopied ]]; do
+ oldcopied=$newcopied
+ sleep 1
+ newcopied=$(zpool status $TESTPOOL | \
+ grep 'copied out of' | \
+ awk '{print $1}')
+ done
+}
+
+log_onexit cleanup
+
+function test_replace # <pool> <devices> <parity>
+{
+ pool=${1}
+ devices=${2}
+ nparity=${3}
+ device_count=0
+
+ log_must echo "devices=$devices"
+
+ for dev in ${devices}; do
+ device_count=$((device_count+1))
+ done
+
+ index=$((RANDOM%(device_count-nparity)))
+ for (( j=1; j<=$nparity; j=j+1 )); do
+ log_must zpool offline $pool ${disks[$((index+j))]}
+ log_must dd if=/dev/zero of=${disks[$((index+j))]} \
+ bs=1024k count=$dev_size_mb conv=notrunc
+ done
+
+ for (( j=1; j<=$nparity; j=j+1 )); do
+ log_must zpool replace $pool ${disks[$((index+j))]}
+ done
+
+ log_must zpool wait -t replace $pool
+ log_must check_pool_status $pool "scan" "with 0 errors"
+
+ log_must zpool clear $pool
+ log_must zpool scrub -w $pool
+
+ log_must zpool status -v
+ log_must check_pool_status $pool "scan" "repaired 0B"
+}
+
+log_must set_tunable32 EMBEDDED_SLOG_MIN_MS 99999
+
+# Disk files which will be used by pool
+for i in {0..$(($devs))}; do
+ device=$TEST_BASE_DIR/dev-$i
+ log_must truncate -s ${dev_size_mb}M $device
+ disks[${#disks[*]}+1]=$device
+done
+
+nparity=$((RANDOM%(3) + 1))
+raid=raidz$nparity
+pool=$TESTPOOL
+opts="-o cachefile=none"
+devices=""
+
+log_must set_tunable32 SCRUB_AFTER_EXPAND 0
+
+log_must zpool create -f $opts $pool $raid ${disks[1..$(($nparity+1))]}
+devices="${disks[1..$(($nparity+1))]}"
+
+log_must zfs create -o recordsize=8k $pool/fs
+log_must fill_fs /$pool/fs 1 128 100 1024 R
+
+log_must zfs create -o recordsize=128k $pool/fs2
+log_must fill_fs /$pool/fs2 1 128 100 1024 R
+
+for disk in ${disks[$(($nparity+2))..$devs]}; do
+ # Set pause to some random value near halfway point
+ reflow_size=$(get_pool_prop allocated $pool)
+ pause=$((((RANDOM << 15) + RANDOM) % reflow_size / 2))
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES $pause
+
+ log_must zpool attach $pool ${raid}-0 $disk
+ devices="$devices $disk"
+
+ wait_expand_paused
+
+ for (( i=0; i<2; i++ )); do
+ test_replace $pool "$devices" $nparity
+
+ # Increase pause by about 25%
+ pause=$((pause + (((RANDOM << 15) + RANDOM) % \
+ reflow_size) / 4))
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES $pause
+
+ wait_expand_paused
+ done
+
+ # Set pause past largest possible value for this pool
+ pause=$((devs*dev_size_mb*1024*1024))
+ log_must set_tunable64 RAIDZ_EXPAND_MAX_REFLOW_BYTES $pause
+
+ log_must zpool wait -t raidz_expand $pool
+done
+
+log_must zpool destroy "$pool"
+
+log_pass "raidz expansion test succeeded."
+
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_006_neg.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_006_neg.ksh
new file mode 100755
index 000000000000..35ba8bde2392
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_006_neg.ksh
@@ -0,0 +1,78 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or 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) 2023 by iXsystems, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+#
+# DESCRIPTION:
+# 'zpool attach poolname raidz ...' should fail if raidz_expansion
+# feature is not enabled.
+#
+# STRATEGY:
+# 1. Create raidz pool with raidz_expansion feature disabled
+# 2. Attempt to attach a device to the raidz vdev
+# 3. Verify that device attached failed
+# 4. Destroy the raidz pool
+
+typeset -r devs=4
+typeset -r dev_size_mb=128
+typeset -a disks
+
+function cleanup
+{
+ log_pos zpool status "$TESTPOOL"
+
+ poolexists "$TESTPOOL" && log_must_busy zpool destroy "$TESTPOOL"
+
+ for i in {0..$devs}; do
+ log_must rm -f "$TEST_BASE_DIR/dev-$i"
+ done
+}
+
+log_onexit cleanup
+
+for i in {0..$devs}; do
+ device=$TEST_BASE_DIR/dev-$i
+ log_must truncate -s ${dev_size_mb}M "$device"
+ if [[ $i -ne $devs ]]; then
+ disks[${#disks[*]}+1]=$device
+ fi
+done
+
+# create a pool with raidz_expansion feature disabled
+log_must zpool create -f -o cachefile=none -o feature@raidz_expansion=disabled \
+ "$TESTPOOL" raidz1 "${disks[@]}"
+status=$(zpool list -H -o feature@raidz_expansion "$TESTPOOL")
+if [[ "$status" != "disabled" ]]; then
+ log_fail "raidz_expansion feature was not disabled"
+fi
+
+# expecting attach to fail
+log_mustnot_expect "raidz_expansion feature must be enabled" zpool attach -f \
+ "$TESTPOOL" raidz1-0 "$TEST_BASE_DIR/dev-$devs"
+log_must zpool destroy "$TESTPOOL"
+
+log_pass "raidz attach failed with feature disabled as expected"
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_007_neg.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_007_neg.ksh
new file mode 100755
index 000000000000..78294cb9e516
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/raidz/raidz_expand_007_neg.ksh
@@ -0,0 +1,86 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or 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) 2023 by iXsystems, Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+#
+# DESCRIPTION:
+# Negative for FreeBSD Only
+#
+# Attempting to expand a RAIDZ should fail if the scratch area on the
+# existing disks contains BTX Server binary (used to boot FreeBSD when
+# using MBR partitions with ZFS).
+#
+# STRATEGY:
+# 1. Create raidz pool
+# 2. Add a BTX header to the reserved boot area
+# 3. Attempt to attach a device to the raidz vdev
+# 4. Verify that device attached failed
+# 5. Destroy the raidz pool
+
+typeset -r devs=4
+typeset -r dev_size_mb=128
+typeset -a disks
+
+function cleanup
+{
+ log_pos zpool status "$TESTPOOL"
+
+ poolexists "$TESTPOOL" && log_must_busy zpool destroy "$TESTPOOL"
+
+ for i in {0..$devs}; do
+ log_must rm -f "$TEST_BASE_DIR/dev-$i"
+ done
+}
+
+log_onexit cleanup
+
+for i in {0..$devs}; do
+ device=$TEST_BASE_DIR/dev-$i
+ # simulate active BTX Server data by inserting a BTX header
+ printf "\xeb\x0e%s\x01\x02\x80" "BTX" | dd of="$device" \
+ bs=512 seek=1024 status=none
+ log_must truncate -s ${dev_size_mb}M "$device"
+ if [[ $i -ne $devs ]]; then
+ disks[${#disks[*]}+1]=$device
+ fi
+done
+
+log_must zpool create -f -o cachefile=none "$TESTPOOL" raidz1 "${disks[@]}"
+
+if is_freebsd; then
+ # expecting attach to fail
+ log_mustnot_expect "the reserved boot area" zpool attach -f \
+ "$TESTPOOL" raidz1-0 "$TEST_BASE_DIR/dev-$devs"
+ log_must zpool destroy "$TESTPOOL"
+ log_pass "raidz attach failed with in-use reserved boot area"
+else
+ # expecting attach to pass everywhere else
+ log_must zpool attach -f "$TESTPOOL" raidz1-0 "$TEST_BASE_DIR/dev-$devs"
+ log_must zpool destroy "$TESTPOOL"
+ log_pass "raidz attach passed with in-use reserved boot area"
+fi
+
diff --git a/sys/modules/zfs/zfs_config.h b/sys/modules/zfs/zfs_config.h
index dd1a926d286c..686ab2e3f80d 100644
--- a/sys/modules/zfs/zfs_config.h
+++ b/sys/modules/zfs/zfs_config.h
@@ -1,1149 +1,1149 @@
/*
*/
/* zfs_config.h. Generated from zfs_config.h.in by configure. */
/* zfs_config.h.in. Generated from configure.ac by autoheader. */
/* Define to 1 if translation of program messages to the user's native
language is requested. */
/* #undef ENABLE_NLS */
/* bio_end_io_t wants 1 arg */
/* #undef HAVE_1ARG_BIO_END_IO_T */
/* lookup_bdev() wants 1 arg */
/* #undef HAVE_1ARG_LOOKUP_BDEV */
/* submit_bio() wants 1 arg */
/* #undef HAVE_1ARG_SUBMIT_BIO */
/* bdi_setup_and_register() wants 2 args */
/* #undef HAVE_2ARGS_BDI_SETUP_AND_REGISTER */
/* vfs_getattr wants 2 args */
/* #undef HAVE_2ARGS_VFS_GETATTR */
/* zlib_deflate_workspacesize() wants 2 args */
/* #undef HAVE_2ARGS_ZLIB_DEFLATE_WORKSPACESIZE */
/* bdi_setup_and_register() wants 3 args */
/* #undef HAVE_3ARGS_BDI_SETUP_AND_REGISTER */
/* vfs_getattr wants 3 args */
/* #undef HAVE_3ARGS_VFS_GETATTR */
/* vfs_getattr wants 4 args */
/* #undef HAVE_4ARGS_VFS_GETATTR */
/* kernel has access_ok with 'type' parameter */
/* #undef HAVE_ACCESS_OK_TYPE */
/* posix_acl has refcount_t */
/* #undef HAVE_ACL_REFCOUNT */
/* add_disk() returns int */
/* #undef HAVE_ADD_DISK_RET */
/* Define if host toolchain supports AES */
#define HAVE_AES 1
/* Define if you have [rt] */
#define HAVE_AIO_H 1
#ifdef __amd64__
#ifndef RESCUE
/* Define if host toolchain supports AVX */
#define HAVE_AVX 1
#endif
/* Define if host toolchain supports AVX2 */
#define HAVE_AVX2 1
/* Define if host toolchain supports AVX512BW */
#define HAVE_AVX512BW 1
/* Define if host toolchain supports AVX512CD */
#define HAVE_AVX512CD 1
/* Define if host toolchain supports AVX512DQ */
#define HAVE_AVX512DQ 1
/* Define if host toolchain supports AVX512ER */
#define HAVE_AVX512ER 1
/* Define if host toolchain supports AVX512F */
#define HAVE_AVX512F 1
/* Define if host toolchain supports AVX512IFMA */
#define HAVE_AVX512IFMA 1
/* Define if host toolchain supports AVX512PF */
#define HAVE_AVX512PF 1
/* Define if host toolchain supports AVX512VBMI */
#define HAVE_AVX512VBMI 1
/* Define if host toolchain supports AVX512VL */
#define HAVE_AVX512VL 1
#endif
/* bdevname() is available */
/* #undef HAVE_BDEVNAME */
/* bdev_check_media_change() exists */
/* #undef HAVE_BDEV_CHECK_MEDIA_CHANGE */
/* bdev_*_io_acct() available */
/* #undef HAVE_BDEV_IO_ACCT_63 */
/* bdev_*_io_acct() available */
/* #undef HAVE_BDEV_IO_ACCT_OLD */
/* bdev_kobj() exists */
/* #undef HAVE_BDEV_KOBJ */
/* bdev_max_discard_sectors() is available */
/* #undef HAVE_BDEV_MAX_DISCARD_SECTORS */
/* bdev_max_secure_erase_sectors() is available */
/* #undef HAVE_BDEV_MAX_SECURE_ERASE_SECTORS */
/* block_device_operations->submit_bio() returns void */
/* #undef HAVE_BDEV_SUBMIT_BIO_RETURNS_VOID */
/* bdev_whole() is available */
/* #undef HAVE_BDEV_WHOLE */
/* bio_alloc() takes 4 arguments */
/* #undef HAVE_BIO_ALLOC_4ARG */
/* bio->bi_bdev->bd_disk exists */
/* #undef HAVE_BIO_BDEV_DISK */
/* bio->bi_opf is defined */
/* #undef HAVE_BIO_BI_OPF */
/* bio->bi_status exists */
/* #undef HAVE_BIO_BI_STATUS */
/* bio has bi_iter */
/* #undef HAVE_BIO_BVEC_ITER */
/* bio_*_io_acct() available */
/* #undef HAVE_BIO_IO_ACCT */
/* bio_max_segs() is implemented */
/* #undef HAVE_BIO_MAX_SEGS */
/* bio_set_dev() is available */
/* #undef HAVE_BIO_SET_DEV */
/* bio_set_dev() GPL-only */
/* #undef HAVE_BIO_SET_DEV_GPL_ONLY */
/* bio_set_dev() is a macro */
/* #undef HAVE_BIO_SET_DEV_MACRO */
/* bio_set_op_attrs is available */
/* #undef HAVE_BIO_SET_OP_ATTRS */
/* blkdev_get_by_path() exists and takes 4 args */
/* #undef HAVE_BLKDEV_GET_BY_PATH_4ARG */
/* blkdev_get_by_path() handles ERESTARTSYS */
/* #undef HAVE_BLKDEV_GET_ERESTARTSYS */
/* blkdev_issue_discard() is available */
/* #undef HAVE_BLKDEV_ISSUE_DISCARD */
/* blkdev_issue_secure_erase() is available */
/* #undef HAVE_BLKDEV_ISSUE_SECURE_ERASE */
/* blkdev_put() accepts void* as arg 2 */
/* #undef HAVE_BLKDEV_PUT_HOLDER */
/* blkdev_reread_part() exists */
/* #undef HAVE_BLKDEV_REREAD_PART */
/* blkg_tryget() is available */
/* #undef HAVE_BLKG_TRYGET */
/* blkg_tryget() GPL-only */
/* #undef HAVE_BLKG_TRYGET_GPL_ONLY */
/* blk_alloc_disk() exists */
/* #undef HAVE_BLK_ALLOC_DISK */
/* blk_alloc_queue() expects request function */
/* #undef HAVE_BLK_ALLOC_QUEUE_REQUEST_FN */
/* blk_alloc_queue_rh() expects request function */
/* #undef HAVE_BLK_ALLOC_QUEUE_REQUEST_FN_RH */
/* blk_cleanup_disk() exists */
/* #undef HAVE_BLK_CLEANUP_DISK */
/* blk_mode_t is defined */
/* #undef HAVE_BLK_MODE_T */
/* block multiqueue is available */
/* #undef HAVE_BLK_MQ */
/* blk queue backing_dev_info is dynamic */
/* #undef HAVE_BLK_QUEUE_BDI_DYNAMIC */
/* blk_queue_discard() is available */
/* #undef HAVE_BLK_QUEUE_DISCARD */
/* blk_queue_flag_clear() exists */
/* #undef HAVE_BLK_QUEUE_FLAG_CLEAR */
/* blk_queue_flag_set() exists */
/* #undef HAVE_BLK_QUEUE_FLAG_SET */
/* blk_queue_flush() is available */
/* #undef HAVE_BLK_QUEUE_FLUSH */
/* blk_queue_flush() is GPL-only */
/* #undef HAVE_BLK_QUEUE_FLUSH_GPL_ONLY */
/* blk_queue_secdiscard() is available */
/* #undef HAVE_BLK_QUEUE_SECDISCARD */
/* blk_queue_secure_erase() is available */
/* #undef HAVE_BLK_QUEUE_SECURE_ERASE */
/* blk_queue_update_readahead() exists */
/* #undef HAVE_BLK_QUEUE_UPDATE_READAHEAD */
/* blk_queue_write_cache() exists */
/* #undef HAVE_BLK_QUEUE_WRITE_CACHE */
/* blk_queue_write_cache() is GPL-only */
/* #undef HAVE_BLK_QUEUE_WRITE_CACHE_GPL_ONLY */
/* BLK_STS_RESV_CONFLICT is defined */
/* #undef HAVE_BLK_STS_RESV_CONFLICT */
/* Define if release() in block_device_operations takes 1 arg */
/* #undef HAVE_BLOCK_DEVICE_OPERATIONS_RELEASE_1ARG */
/* Define if revalidate_disk() in block_device_operations */
/* #undef HAVE_BLOCK_DEVICE_OPERATIONS_REVALIDATE_DISK */
/* Define to 1 if you have the Mac OS X function CFLocaleCopyCurrent in the
CoreFoundation framework. */
/* #undef HAVE_CFLOCALECOPYCURRENT */
/* Define to 1 if you have the Mac OS X function
CFLocaleCopyPreferredLanguages in the CoreFoundation framework. */
/* #undef HAVE_CFLOCALECOPYPREFERREDLANGUAGES */
/* Define to 1 if you have the Mac OS X function CFPreferencesCopyAppValue in
the CoreFoundation framework. */
/* #undef HAVE_CFPREFERENCESCOPYAPPVALUE */
/* check_disk_change() exists */
/* #undef HAVE_CHECK_DISK_CHANGE */
/* clear_inode() is available */
/* #undef HAVE_CLEAR_INODE */
/* dentry uses const struct dentry_operations */
/* #undef HAVE_CONST_DENTRY_OPERATIONS */
/* copy_from_iter() is available */
/* #undef HAVE_COPY_FROM_ITER */
/* copy_splice_read exists */
/* #undef HAVE_COPY_SPLICE_READ */
/* copy_to_iter() is available */
/* #undef HAVE_COPY_TO_ITER */
/* cpu_has_feature() is GPL-only */
/* #undef HAVE_CPU_HAS_FEATURE_GPL_ONLY */
/* yes */
/* #undef HAVE_CPU_HOTPLUG */
/* current_time() exists */
/* #undef HAVE_CURRENT_TIME */
/* Define if the GNU dcgettext() function is already present or preinstalled.
*/
/* #undef HAVE_DCGETTEXT */
/* DECLARE_EVENT_CLASS() is available */
/* #undef HAVE_DECLARE_EVENT_CLASS */
/* dentry aliases are in d_u member */
/* #undef HAVE_DENTRY_D_U_ALIASES */
/* dequeue_signal() takes 4 arguments */
/* #undef HAVE_DEQUEUE_SIGNAL_4ARG */
/* lookup_bdev() wants dev_t arg */
/* #undef HAVE_DEVT_LOOKUP_BDEV */
/* sops->dirty_inode() wants flags */
/* #undef HAVE_DIRTY_INODE_WITH_FLAGS */
/* disk_check_media_change() exists */
/* #undef HAVE_DISK_CHECK_MEDIA_CHANGE */
/* disk_*_io_acct() available */
/* #undef HAVE_DISK_IO_ACCT */
/* disk_update_readahead() exists */
/* #undef HAVE_DISK_UPDATE_READAHEAD */
/* Define to 1 if you have the <dlfcn.h> header file. */
#define HAVE_DLFCN_H 1
/* d_make_root() is available */
/* #undef HAVE_D_MAKE_ROOT */
/* d_prune_aliases() is available */
/* #undef HAVE_D_PRUNE_ALIASES */
/* dops->d_revalidate() operation takes nameidata */
/* #undef HAVE_D_REVALIDATE_NAMEIDATA */
/* eops->encode_fh() wants child and parent inodes */
/* #undef HAVE_ENCODE_FH_WITH_INODE */
/* sops->evict_inode() exists */
/* #undef HAVE_EVICT_INODE */
/* FALLOC_FL_ZERO_RANGE is defined */
/* #undef HAVE_FALLOC_FL_ZERO_RANGE */
/* fault_in_iov_iter_readable() is available */
/* #undef HAVE_FAULT_IN_IOV_ITER_READABLE */
/* filemap_range_has_page() is available */
/* #undef HAVE_FILEMAP_RANGE_HAS_PAGE */
/* fops->aio_fsync() exists */
/* #undef HAVE_FILE_AIO_FSYNC */
/* file_dentry() is available */
/* #undef HAVE_FILE_DENTRY */
/* fops->fadvise() exists */
/* #undef HAVE_FILE_FADVISE */
/* file_inode() is available */
/* #undef HAVE_FILE_INODE */
/* flush_dcache_page() is GPL-only */
/* #undef HAVE_FLUSH_DCACHE_PAGE_GPL_ONLY */
/* iops->follow_link() cookie */
/* #undef HAVE_FOLLOW_LINK_COOKIE */
/* iops->follow_link() nameidata */
/* #undef HAVE_FOLLOW_LINK_NAMEIDATA */
/* Define if compiler supports -Wformat-overflow */
/* #undef HAVE_FORMAT_OVERFLOW */
/* fsync_bdev() is declared in include/blkdev.h */
/* #undef HAVE_FSYNC_BDEV */
/* fops->fsync() with range */
/* #undef HAVE_FSYNC_RANGE */
/* fops->fsync() without dentry */
/* #undef HAVE_FSYNC_WITHOUT_DENTRY */
/* yes */
/* #undef HAVE_GENERIC_FADVISE */
/* generic_fillattr requires struct mnt_idmap* */
/* #undef HAVE_GENERIC_FILLATTR_IDMAP */
/* generic_fillattr requires struct mnt_idmap* and u32 request_mask */
/* #undef HAVE_GENERIC_FILLATTR_IDMAP_REQMASK */
/* generic_fillattr requires struct user_namespace* */
/* #undef HAVE_GENERIC_FILLATTR_USERNS */
/* generic_*_io_acct() 3 arg available */
/* #undef HAVE_GENERIC_IO_ACCT_3ARG */
/* generic_*_io_acct() 4 arg available */
/* #undef HAVE_GENERIC_IO_ACCT_4ARG */
/* generic_readlink is global */
/* #undef HAVE_GENERIC_READLINK */
/* generic_setxattr() exists */
/* #undef HAVE_GENERIC_SETXATTR */
/* generic_write_checks() takes kiocb */
/* #undef HAVE_GENERIC_WRITE_CHECKS_KIOCB */
/* Define if the GNU gettext() function is already present or preinstalled. */
/* #undef HAVE_GETTEXT */
/* iops->get_acl() exists */
/* #undef HAVE_GET_ACL */
/* iops->get_acl() takes rcu */
/* #undef HAVE_GET_ACL_RCU */
/* has iops->get_inode_acl() */
/* #undef HAVE_GET_INODE_ACL */
/* iops->get_link() cookie */
/* #undef HAVE_GET_LINK_COOKIE */
/* iops->get_link() delayed */
/* #undef HAVE_GET_LINK_DELAYED */
/* group_info->gid exists */
/* #undef HAVE_GROUP_INFO_GID */
/* has_capability() is available */
/* #undef HAVE_HAS_CAPABILITY */
/* iattr->ia_vfsuid and iattr->ia_vfsgid exist */
/* #undef HAVE_IATTR_VFSID */
/* Define if you have the iconv() function and it works. */
#define HAVE_ICONV 1
/* iops->getattr() takes struct mnt_idmap* */
/* #undef HAVE_IDMAP_IOPS_GETATTR */
/* iops->setattr() takes struct mnt_idmap* */
/* #undef HAVE_IDMAP_IOPS_SETATTR */
/* APIs for idmapped mount are present */
/* #undef HAVE_IDMAP_MNT_API */
/* Define if compiler supports -Wimplicit-fallthrough */
/* #undef HAVE_IMPLICIT_FALLTHROUGH */
/* Define if compiler supports -Winfinite-recursion */
/* #undef HAVE_INFINITE_RECURSION */
/* inode_get_ctime() exists in linux/fs.h */
/* #undef HAVE_INODE_GET_CTIME */
/* yes */
/* #undef HAVE_INODE_LOCK_SHARED */
/* inode_owner_or_capable() exists */
/* #undef HAVE_INODE_OWNER_OR_CAPABLE */
/* inode_owner_or_capable() takes mnt_idmap */
/* #undef HAVE_INODE_OWNER_OR_CAPABLE_IDMAP */
/* inode_owner_or_capable() takes user_ns */
/* #undef HAVE_INODE_OWNER_OR_CAPABLE_USERNS */
/* inode_set_ctime_to_ts() exists in linux/fs.h */
/* #undef HAVE_INODE_SET_CTIME_TO_TS */
/* inode_set_flags() exists */
/* #undef HAVE_INODE_SET_FLAGS */
/* inode_set_iversion() exists */
/* #undef HAVE_INODE_SET_IVERSION */
/* inode->i_*time's are timespec64 */
/* #undef HAVE_INODE_TIMESPEC64_TIMES */
/* timestamp_truncate() exists */
/* #undef HAVE_INODE_TIMESTAMP_TRUNCATE */
/* Define to 1 if you have the <inttypes.h> header file. */
#define HAVE_INTTYPES_H 1
/* in_compat_syscall() is available */
/* #undef HAVE_IN_COMPAT_SYSCALL */
/* iops->create() takes struct mnt_idmap* */
/* #undef HAVE_IOPS_CREATE_IDMAP */
/* iops->create() takes struct user_namespace* */
/* #undef HAVE_IOPS_CREATE_USERNS */
/* iops->mkdir() takes struct mnt_idmap* */
/* #undef HAVE_IOPS_MKDIR_IDMAP */
/* iops->mkdir() takes struct user_namespace* */
/* #undef HAVE_IOPS_MKDIR_USERNS */
/* iops->mknod() takes struct mnt_idmap* */
/* #undef HAVE_IOPS_MKNOD_IDMAP */
/* iops->mknod() takes struct user_namespace* */
/* #undef HAVE_IOPS_MKNOD_USERNS */
/* iops->permission() takes struct mnt_idmap* */
/* #undef HAVE_IOPS_PERMISSION_IDMAP */
/* iops->permission() takes struct user_namespace* */
/* #undef HAVE_IOPS_PERMISSION_USERNS */
/* iops->rename() takes struct mnt_idmap* */
/* #undef HAVE_IOPS_RENAME_IDMAP */
/* iops->rename() takes struct user_namespace* */
/* #undef HAVE_IOPS_RENAME_USERNS */
/* iops->setattr() exists */
/* #undef HAVE_IOPS_SETATTR */
/* iops->symlink() takes struct mnt_idmap* */
/* #undef HAVE_IOPS_SYMLINK_IDMAP */
/* iops->symlink() takes struct user_namespace* */
/* #undef HAVE_IOPS_SYMLINK_USERNS */
/* iov_iter_advance() is available */
/* #undef HAVE_IOV_ITER_ADVANCE */
/* iov_iter_count() is available */
/* #undef HAVE_IOV_ITER_COUNT */
/* iov_iter_fault_in_readable() is available */
/* #undef HAVE_IOV_ITER_FAULT_IN_READABLE */
/* iov_iter_revert() is available */
/* #undef HAVE_IOV_ITER_REVERT */
/* iov_iter_type() is available */
/* #undef HAVE_IOV_ITER_TYPE */
/* iov_iter types are available */
/* #undef HAVE_IOV_ITER_TYPES */
/* yes */
/* #undef HAVE_IO_SCHEDULE_TIMEOUT */
/* Define to 1 if you have the `issetugid' function. */
#define HAVE_ISSETUGID 1
/* iter_iov() is available */
/* #undef HAVE_ITER_IOV */
/* kernel has kernel_fpu_* functions */
/* #undef HAVE_KERNEL_FPU */
/* kernel has asm/fpu/api.h */
/* #undef HAVE_KERNEL_FPU_API_HEADER */
/* kernel fpu internal */
/* #undef HAVE_KERNEL_FPU_INTERNAL */
/* kernel has asm/fpu/internal.h */
/* #undef HAVE_KERNEL_FPU_INTERNAL_HEADER */
/* uncached_acl_sentinel() exists */
/* #undef HAVE_KERNEL_GET_ACL_HANDLE_CACHE */
/* Define if compiler supports -Winfinite-recursion */
/* #undef HAVE_KERNEL_INFINITE_RECURSION */
/* kernel does stack verification */
/* #undef HAVE_KERNEL_OBJTOOL */
/* kernel has linux/objtool.h */
/* #undef HAVE_KERNEL_OBJTOOL_HEADER */
/* kernel_read() take loff_t pointer */
/* #undef HAVE_KERNEL_READ_PPOS */
/* timer_list.function gets a timer_list */
/* #undef HAVE_KERNEL_TIMER_FUNCTION_TIMER_LIST */
/* struct timer_list has a flags member */
/* #undef HAVE_KERNEL_TIMER_LIST_FLAGS */
/* timer_setup() is available */
/* #undef HAVE_KERNEL_TIMER_SETUP */
/* kernel_write() take loff_t pointer */
/* #undef HAVE_KERNEL_WRITE_PPOS */
/* kmem_cache_create_usercopy() exists */
/* #undef HAVE_KMEM_CACHE_CREATE_USERCOPY */
/* kstrtoul() exists */
/* #undef HAVE_KSTRTOUL */
/* ktime_get_coarse_real_ts64() exists */
/* #undef HAVE_KTIME_GET_COARSE_REAL_TS64 */
/* ktime_get_raw_ts64() exists */
/* #undef HAVE_KTIME_GET_RAW_TS64 */
/* kvmalloc exists */
/* #undef HAVE_KVMALLOC */
/* Define if you have [aio] */
/* #undef HAVE_LIBAIO */
/* Define if you have [blkid] */
/* #undef HAVE_LIBBLKID */
/* Define if you have [crypto] */
#define HAVE_LIBCRYPTO 1
/* Define if you have [tirpc] */
/* #undef HAVE_LIBTIRPC */
/* Define if you have [udev] */
/* #undef HAVE_LIBUDEV */
/* Define if you have [uuid] */
/* #undef HAVE_LIBUUID */
/* linux/blk-cgroup.h exists */
/* #undef HAVE_LINUX_BLK_CGROUP_HEADER */
/* lseek_execute() is available */
/* #undef HAVE_LSEEK_EXECUTE */
/* makedev() is declared in sys/mkdev.h */
/* #undef HAVE_MAKEDEV_IN_MKDEV */
/* makedev() is declared in sys/sysmacros.h */
/* #undef HAVE_MAKEDEV_IN_SYSMACROS */
/* Noting that make_request_fn() returns blk_qc_t */
/* #undef HAVE_MAKE_REQUEST_FN_RET_QC */
/* Noting that make_request_fn() returns void */
/* #undef HAVE_MAKE_REQUEST_FN_RET_VOID */
/* iops->mkdir() takes umode_t */
/* #undef HAVE_MKDIR_UMODE_T */
/* Define to 1 if you have the `mlockall' function. */
#define HAVE_MLOCKALL 1
/* lookup_bdev() wants mode arg */
/* #undef HAVE_MODE_LOOKUP_BDEV */
/* Define if host toolchain supports MOVBE */
#define HAVE_MOVBE 1
/* new_sync_read()/new_sync_write() are available */
/* #undef HAVE_NEW_SYNC_READ */
/* folio_wait_bit() exists */
/* #undef HAVE_PAGEMAP_FOLIO_WAIT_BIT */
/* part_to_dev() exists */
/* #undef HAVE_PART_TO_DEV */
/* iops->getattr() takes a path */
/* #undef HAVE_PATH_IOPS_GETATTR */
/* Define if host toolchain supports PCLMULQDQ */
#define HAVE_PCLMULQDQ 1
/* percpu_counter_add_batch() is defined */
/* #undef HAVE_PERCPU_COUNTER_ADD_BATCH */
/* percpu_counter_init() wants gfp_t */
/* #undef HAVE_PERCPU_COUNTER_INIT_WITH_GFP */
/* posix_acl_chmod() exists */
/* #undef HAVE_POSIX_ACL_CHMOD */
/* posix_acl_from_xattr() needs user_ns */
/* #undef HAVE_POSIX_ACL_FROM_XATTR_USERNS */
/* posix_acl_release() is available */
/* #undef HAVE_POSIX_ACL_RELEASE */
/* posix_acl_release() is GPL-only */
/* #undef HAVE_POSIX_ACL_RELEASE_GPL_ONLY */
/* posix_acl_valid() wants user namespace */
/* #undef HAVE_POSIX_ACL_VALID_WITH_NS */
/* proc_ops structure exists */
/* #undef HAVE_PROC_OPS_STRUCT */
/* iops->put_link() cookie */
/* #undef HAVE_PUT_LINK_COOKIE */
/* iops->put_link() delayed */
/* #undef HAVE_PUT_LINK_DELAYED */
/* iops->put_link() nameidata */
/* #undef HAVE_PUT_LINK_NAMEIDATA */
/* If available, contains the Python version number currently in use. */
#define HAVE_PYTHON "3.7"
/* qat is enabled and existed */
/* #undef HAVE_QAT */
/* struct reclaim_state has reclaimed */
/* #undef HAVE_RECLAIM_STATE_RECLAIMED */
/* register_shrinker is vararg */
/* #undef HAVE_REGISTER_SHRINKER_VARARG */
/* register_sysctl_table exists */
/* #undef HAVE_REGISTER_SYSCTL_TABLE */
/* iops->rename2() exists */
/* #undef HAVE_RENAME2 */
/* struct inode_operations_wrapper takes .rename2() */
/* #undef HAVE_RENAME2_OPERATIONS_WRAPPER */
/* iops->rename() wants flags */
/* #undef HAVE_RENAME_WANTS_FLAGS */
/* REQ_DISCARD is defined */
/* #undef HAVE_REQ_DISCARD */
/* REQ_FLUSH is defined */
/* #undef HAVE_REQ_FLUSH */
/* REQ_OP_DISCARD is defined */
/* #undef HAVE_REQ_OP_DISCARD */
/* REQ_OP_FLUSH is defined */
/* #undef HAVE_REQ_OP_FLUSH */
/* REQ_OP_SECURE_ERASE is defined */
/* #undef HAVE_REQ_OP_SECURE_ERASE */
/* REQ_PREFLUSH is defined */
/* #undef HAVE_REQ_PREFLUSH */
/* revalidate_disk() is available */
/* #undef HAVE_REVALIDATE_DISK */
/* revalidate_disk_size() is available */
/* #undef HAVE_REVALIDATE_DISK_SIZE */
/* struct rw_semaphore has member activity */
/* #undef HAVE_RWSEM_ACTIVITY */
/* struct rw_semaphore has atomic_long_t member count */
/* #undef HAVE_RWSEM_ATOMIC_LONG_COUNT */
/* linux/sched/signal.h exists */
/* #undef HAVE_SCHED_SIGNAL_HEADER */
/* Define to 1 if you have the <security/pam_modules.h> header file. */
#define HAVE_SECURITY_PAM_MODULES_H 1
/* setattr_prepare() accepts mnt_idmap */
/* #undef HAVE_SETATTR_PREPARE_IDMAP */
/* setattr_prepare() is available, doesn't accept user_namespace */
/* #undef HAVE_SETATTR_PREPARE_NO_USERNS */
/* setattr_prepare() accepts user_namespace */
/* #undef HAVE_SETATTR_PREPARE_USERNS */
/* iops->set_acl() exists, takes 3 args */
/* #undef HAVE_SET_ACL */
/* iops->set_acl() takes 4 args, arg1 is struct mnt_idmap * */
/* #undef HAVE_SET_ACL_IDMAP_DENTRY */
/* iops->set_acl() takes 4 args */
/* #undef HAVE_SET_ACL_USERNS */
/* iops->set_acl() takes 4 args, arg2 is struct dentry * */
/* #undef HAVE_SET_ACL_USERNS_DENTRY_ARG2 */
/* set_cached_acl() is usable */
/* #undef HAVE_SET_CACHED_ACL_USABLE */
/* set_special_state() exists */
/* #undef HAVE_SET_SPECIAL_STATE */
/* struct shrink_control exists */
/* #undef HAVE_SHRINK_CONTROL_STRUCT */
/* kernel_siginfo_t exists */
/* #undef HAVE_SIGINFO */
/* signal_stop() exists */
/* #undef HAVE_SIGNAL_STOP */
/* new shrinker callback wants 2 args */
/* #undef HAVE_SINGLE_SHRINKER_CALLBACK */
/* cs->count_objects exists */
/* #undef HAVE_SPLIT_SHRINKER_CALLBACK */
#if defined(__amd64__) || defined(__i386__)
/* Define if host toolchain supports SSE */
#define HAVE_SSE 1
/* Define if host toolchain supports SSE2 */
#define HAVE_SSE2 1
/* Define if host toolchain supports SSE3 */
#define HAVE_SSE3 1
/* Define if host toolchain supports SSE4.1 */
#define HAVE_SSE4_1 1
/* Define if host toolchain supports SSE4.2 */
#define HAVE_SSE4_2 1
/* Define if host toolchain supports SSSE3 */
#define HAVE_SSSE3 1
#endif
/* STACK_FRAME_NON_STANDARD is defined */
/* #undef HAVE_STACK_FRAME_NON_STANDARD */
/* standalone <linux/stdarg.h> exists */
/* #undef HAVE_STANDALONE_LINUX_STDARG */
/* Define to 1 if you have the <stdint.h> header file. */
#define HAVE_STDINT_H 1
/* Define to 1 if you have the <stdio.h> header file. */
#define HAVE_STDIO_H 1
/* Define to 1 if you have the <stdlib.h> header file. */
#define HAVE_STDLIB_H 1
/* Define to 1 if you have the <strings.h> header file. */
#define HAVE_STRINGS_H 1
/* Define to 1 if you have the <string.h> header file. */
#define HAVE_STRING_H 1
/* Define to 1 if you have the `strlcat' function. */
#define HAVE_STRLCAT 1
/* Define to 1 if you have the `strlcpy' function. */
#define HAVE_STRLCPY 1
/* submit_bio is member of struct block_device_operations */
/* #undef HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS */
/* super_setup_bdi_name() exits */
/* #undef HAVE_SUPER_SETUP_BDI_NAME */
/* super_block->s_user_ns exists */
/* #undef HAVE_SUPER_USER_NS */
/* sync_blockdev() is declared in include/blkdev.h */
/* #undef HAVE_SYNC_BLOCKDEV */
/* struct kobj_type has default_groups */
/* #undef HAVE_SYSFS_DEFAULT_GROUPS */
/* Define to 1 if you have the <sys/stat.h> header file. */
#define HAVE_SYS_STAT_H 1
/* Define to 1 if you have the <sys/types.h> header file. */
#define HAVE_SYS_TYPES_H 1
/* i_op->tmpfile() exists */
/* #undef HAVE_TMPFILE */
/* i_op->tmpfile() uses old dentry signature */
/* #undef HAVE_TMPFILE_DENTRY */
/* i_op->tmpfile() has mnt_idmap */
/* #undef HAVE_TMPFILE_IDMAP */
/* i_op->tmpfile() has userns */
/* #undef HAVE_TMPFILE_USERNS */
/* totalhigh_pages() exists */
/* #undef HAVE_TOTALHIGH_PAGES */
/* kernel has totalram_pages() */
/* #undef HAVE_TOTALRAM_PAGES_FUNC */
/* Define to 1 if you have the `udev_device_get_is_initialized' function. */
/* #undef HAVE_UDEV_DEVICE_GET_IS_INITIALIZED */
/* kernel has __kernel_fpu_* functions */
/* #undef HAVE_UNDERSCORE_KERNEL_FPU */
/* Define to 1 if you have the <unistd.h> header file. */
#define HAVE_UNISTD_H 1
/* iops->getattr() takes struct user_namespace* */
/* #undef HAVE_USERNS_IOPS_GETATTR */
/* iops->setattr() takes struct user_namespace* */
/* #undef HAVE_USERNS_IOPS_SETATTR */
/* user_namespace->ns.inum exists */
/* #undef HAVE_USER_NS_COMMON_INUM */
/* iops->getattr() takes a vfsmount */
/* #undef HAVE_VFSMOUNT_IOPS_GETATTR */
/* fops->clone_file_range() is available */
/* #undef HAVE_VFS_CLONE_FILE_RANGE */
/* fops->copy_file_range() is available */
/* #undef HAVE_VFS_COPY_FILE_RANGE */
/* fops->dedupe_file_range() is available */
/* #undef HAVE_VFS_DEDUPE_FILE_RANGE */
/* aops->direct_IO() uses iovec */
/* #undef HAVE_VFS_DIRECT_IO_IOVEC */
/* aops->direct_IO() uses iov_iter without rw */
/* #undef HAVE_VFS_DIRECT_IO_ITER */
/* aops->direct_IO() uses iov_iter with offset */
/* #undef HAVE_VFS_DIRECT_IO_ITER_OFFSET */
/* aops->direct_IO() uses iov_iter with rw and offset */
/* #undef HAVE_VFS_DIRECT_IO_ITER_RW_OFFSET */
/* filemap_dirty_folio exists */
/* #undef HAVE_VFS_FILEMAP_DIRTY_FOLIO */
/* file_operations_extend takes .copy_file_range() and .clone_file_range() */
/* #undef HAVE_VFS_FILE_OPERATIONS_EXTEND */
/* generic_copy_file_range() is available */
/* #undef HAVE_VFS_GENERIC_COPY_FILE_RANGE */
/* All required iov_iter interfaces are available */
/* #undef HAVE_VFS_IOV_ITER */
/* fops->iterate() is available */
/* #undef HAVE_VFS_ITERATE */
/* fops->iterate_shared() is available */
/* #undef HAVE_VFS_ITERATE_SHARED */
/* fops->readdir() is available */
/* #undef HAVE_VFS_READDIR */
/* address_space_operations->readpages exists */
/* #undef HAVE_VFS_READPAGES */
/* read_folio exists */
/* #undef HAVE_VFS_READ_FOLIO */
/* fops->remap_file_range() is available */
/* #undef HAVE_VFS_REMAP_FILE_RANGE */
/* fops->read/write_iter() are available */
/* #undef HAVE_VFS_RW_ITERATE */
/* __set_page_dirty_nobuffers exists */
/* #undef HAVE_VFS_SET_PAGE_DIRTY_NOBUFFERS */
/* __vmalloc page flags exists */
/* #undef HAVE_VMALLOC_PAGE_KERNEL */
/* yes */
/* #undef HAVE_WAIT_ON_BIT_ACTION */
/* wait_queue_entry_t exists */
/* #undef HAVE_WAIT_QUEUE_ENTRY_T */
/* wq_head->head and wq_entry->entry exist */
/* #undef HAVE_WAIT_QUEUE_HEAD_ENTRY */
/* int (*writepage_t)() takes struct folio* */
/* #undef HAVE_WRITEPAGE_T_FOLIO */
/* xattr_handler->get() wants dentry */
/* #undef HAVE_XATTR_GET_DENTRY */
/* xattr_handler->get() wants both dentry and inode */
/* #undef HAVE_XATTR_GET_DENTRY_INODE */
/* xattr_handler->get() wants dentry and inode and flags */
/* #undef HAVE_XATTR_GET_DENTRY_INODE_FLAGS */
/* xattr_handler->get() wants xattr_handler */
/* #undef HAVE_XATTR_GET_HANDLER */
/* xattr_handler has name */
/* #undef HAVE_XATTR_HANDLER_NAME */
/* xattr_handler->list() wants dentry */
/* #undef HAVE_XATTR_LIST_DENTRY */
/* xattr_handler->list() wants xattr_handler */
/* #undef HAVE_XATTR_LIST_HANDLER */
/* xattr_handler->list() wants simple */
/* #undef HAVE_XATTR_LIST_SIMPLE */
/* xattr_handler->set() wants dentry */
/* #undef HAVE_XATTR_SET_DENTRY */
/* xattr_handler->set() wants both dentry and inode */
/* #undef HAVE_XATTR_SET_DENTRY_INODE */
/* xattr_handler->set() wants xattr_handler */
/* #undef HAVE_XATTR_SET_HANDLER */
/* xattr_handler->set() takes mnt_idmap */
/* #undef HAVE_XATTR_SET_IDMAP */
/* xattr_handler->set() takes user_namespace */
/* #undef HAVE_XATTR_SET_USERNS */
/* Define if host toolchain supports XSAVE */
#define HAVE_XSAVE 1
/* Define if host toolchain supports XSAVEOPT */
#define HAVE_XSAVEOPT 1
/* Define if host toolchain supports XSAVES */
#define HAVE_XSAVES 1
/* ZERO_PAGE() is GPL-only */
/* #undef HAVE_ZERO_PAGE_GPL_ONLY */
/* Define if you have [z] */
#define HAVE_ZLIB 1
/* __posix_acl_chmod() exists */
/* #undef HAVE___POSIX_ACL_CHMOD */
/* kernel exports FPU functions */
/* #undef KERNEL_EXPORTS_X86_FPU */
/* TBD: fetch(3) support */
#if 0
/* whether the chosen libfetch is to be loaded at run-time */
#define LIBFETCH_DYNAMIC 1
/* libfetch is fetch(3) */
#define LIBFETCH_IS_FETCH 1
/* libfetch is libcurl */
#define LIBFETCH_IS_LIBCURL 0
/* soname of chosen libfetch */
#define LIBFETCH_SONAME "libfetch.so.6"
#endif
/* Define to the sub-directory where libtool stores uninstalled libraries. */
#define LT_OBJDIR ".libs/"
/* make_request_fn() return type */
/* #undef MAKE_REQUEST_FN_RET */
/* struct shrink_control has nid */
/* #undef SHRINK_CONTROL_HAS_NID */
/* using complete_and_exit() instead */
/* #undef SPL_KTHREAD_COMPLETE_AND_EXIT */
/* Defined for legacy compatibility. */
#define SPL_META_ALIAS ZFS_META_ALIAS
/* Defined for legacy compatibility. */
#define SPL_META_RELEASE ZFS_META_RELEASE
/* Defined for legacy compatibility. */
#define SPL_META_VERSION ZFS_META_VERSION
/* pde_data() is PDE_DATA() */
/* #undef SPL_PDE_DATA */
/* Define to 1 if all of the C90 standard headers exist (not just the ones
required in a freestanding environment). This macro is provided for
backward compatibility; new code need not use it. */
#define SYSTEM_FREEBSD 1
/* True if ZFS is to be compiled for a Linux system */
/* #undef SYSTEM_LINUX */
/* Version number of package */
/* #undef ZFS_DEBUG */
/* /dev/zfs minor */
/* #undef ZFS_DEVICE_MINOR */
/* enum node_stat_item contains NR_FILE_PAGES */
/* #undef ZFS_ENUM_NODE_STAT_ITEM_NR_FILE_PAGES */
/* enum node_stat_item contains NR_INACTIVE_ANON */
/* #undef ZFS_ENUM_NODE_STAT_ITEM_NR_INACTIVE_ANON */
/* enum node_stat_item contains NR_INACTIVE_FILE */
/* #undef ZFS_ENUM_NODE_STAT_ITEM_NR_INACTIVE_FILE */
/* enum zone_stat_item contains NR_FILE_PAGES */
/* #undef ZFS_ENUM_ZONE_STAT_ITEM_NR_FILE_PAGES */
/* enum zone_stat_item contains NR_INACTIVE_ANON */
/* #undef ZFS_ENUM_ZONE_STAT_ITEM_NR_INACTIVE_ANON */
/* enum zone_stat_item contains NR_INACTIVE_FILE */
/* #undef ZFS_ENUM_ZONE_STAT_ITEM_NR_INACTIVE_FILE */
/* GENHD_FL_EXT_DEVT flag is not available */
/* #undef ZFS_GENHD_FL_EXT_DEVT */
/* GENHD_FL_NO_PART_SCAN flag is available */
/* #undef ZFS_GENHD_FL_NO_PART */
/* global_node_page_state() exists */
/* #undef ZFS_GLOBAL_NODE_PAGE_STATE */
/* global_zone_page_state() exists */
/* #undef ZFS_GLOBAL_ZONE_PAGE_STATE */
/* Define to 1 if GPL-only symbols can be used */
/* #undef ZFS_IS_GPL_COMPATIBLE */
/* Define the project alias string. */
-#define ZFS_META_ALIAS "zfs-2.2.99-197-FreeBSD_g9198de8f1"
+#define ZFS_META_ALIAS "zfs-2.2.99-202-FreeBSD_g887a3c533"
/* Define the project author. */
#define ZFS_META_AUTHOR "OpenZFS"
/* Define the project release date. */
/* #undef ZFS_META_DATA */
/* Define the maximum compatible kernel version. */
#define ZFS_META_KVER_MAX "6.5"
/* Define the minimum compatible kernel version. */
#define ZFS_META_KVER_MIN "3.10"
/* Define the project license. */
#define ZFS_META_LICENSE "CDDL"
/* Define the libtool library 'age' version information. */
/* #undef ZFS_META_LT_AGE */
/* Define the libtool library 'current' version information. */
/* #undef ZFS_META_LT_CURRENT */
/* Define the libtool library 'revision' version information. */
/* #undef ZFS_META_LT_REVISION */
/* Define the project name. */
#define ZFS_META_NAME "zfs"
/* Define the project release. */
-#define ZFS_META_RELEASE "197-FreeBSD_g9198de8f1"
+#define ZFS_META_RELEASE "202-FreeBSD_g887a3c533"
/* Define the project version. */
#define ZFS_META_VERSION "2.2.99"
/* count is located in percpu_ref.data */
/* #undef ZFS_PERCPU_REF_COUNT_IN_DATA */
diff --git a/sys/modules/zfs/zfs_gitrev.h b/sys/modules/zfs/zfs_gitrev.h
index 86525c162b79..d2aeef5edc18 100644
--- a/sys/modules/zfs/zfs_gitrev.h
+++ b/sys/modules/zfs/zfs_gitrev.h
@@ -1 +1 @@
-#define ZFS_META_GITREV "zfs-2.2.99-197-g9198de8f1"
+#define ZFS_META_GITREV "zfs-2.2.99-202-g887a3c533"

File Metadata

Mime Type
application/octet-stream
Expires
Tue, May 14, 1:01 PM (2 d)
Storage Engine
chunks
Storage Format
Chunks
Storage Handle
UTBIK7wf79iO
Default Alt Text
(4 MB)

Event Timeline