diff options
author | pjd <pjd@FreeBSD.org> | 2008-11-17 20:49:29 +0000 |
---|---|---|
committer | pjd <pjd@FreeBSD.org> | 2008-11-17 20:49:29 +0000 |
commit | bbe899b96e388a8b82439f81ed3707e0d9c6070d (patch) | |
tree | 81b89fa4ac6467771d5aa291a97f4665981a6108 /cddl/contrib/opensolaris/cmd/ztest/ztest.c | |
parent | d2f579595c362ce27b4d87e2c40e1c4e09b929e3 (diff) | |
download | FreeBSD-src-bbe899b96e388a8b82439f81ed3707e0d9c6070d.zip FreeBSD-src-bbe899b96e388a8b82439f81ed3707e0d9c6070d.tar.gz |
Update ZFS from version 6 to 13 and bring some FreeBSD-specific changes.
This bring huge amount of changes, I'll enumerate only user-visible changes:
- Delegated Administration
Allows regular users to perform ZFS operations, like file system
creation, snapshot creation, etc.
- L2ARC
Level 2 cache for ZFS - allows to use additional disks for cache.
Huge performance improvements mostly for random read of mostly
static content.
- slog
Allow to use additional disks for ZFS Intent Log to speed up
operations like fsync(2).
- vfs.zfs.super_owner
Allows regular users to perform privileged operations on files stored
on ZFS file systems owned by him. Very careful with this one.
- chflags(2)
Not all the flags are supported. This still needs work.
- ZFSBoot
Support to boot off of ZFS pool. Not finished, AFAIK.
Submitted by: dfr
- Snapshot properties
- New failure modes
Before if write requested failed, system paniced. Now one
can select from one of three failure modes:
- panic - panic on write error
- wait - wait for disk to reappear
- continue - serve read requests if possible, block write requests
- Refquota, refreservation properties
Just quota and reservation properties, but don't count space consumed
by children file systems, clones and snapshots.
- Sparse volumes
ZVOLs that don't reserve space in the pool.
- External attributes
Compatible with extattr(2).
- NFSv4-ACLs
Not sure about the status, might not be complete yet.
Submitted by: trasz
- Creation-time properties
- Regression tests for zpool(8) command.
Obtained from: OpenSolaris
Diffstat (limited to 'cddl/contrib/opensolaris/cmd/ztest/ztest.c')
-rw-r--r-- | cddl/contrib/opensolaris/cmd/ztest/ztest.c | 1640 |
1 files changed, 889 insertions, 751 deletions
diff --git a/cddl/contrib/opensolaris/cmd/ztest/ztest.c b/cddl/contrib/opensolaris/cmd/ztest/ztest.c index 5d9f028..b7ca302 100644 --- a/cddl/contrib/opensolaris/cmd/ztest/ztest.c +++ b/cddl/contrib/opensolaris/cmd/ztest/ztest.c @@ -19,12 +19,10 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - /* * 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. @@ -92,6 +90,7 @@ #include <sys/zio_compress.h> #include <sys/zil.h> #include <sys/vdev_impl.h> +#include <sys/vdev_file.h> #include <sys/spa_impl.h> #include <sys/dsl_prop.h> #include <sys/refcount.h> @@ -128,8 +127,18 @@ static char *zopt_dir = "/tmp"; static uint64_t zopt_time = 300; /* 5 minutes */ static int zopt_maxfaults; +typedef struct ztest_block_tag { + uint64_t bt_objset; + uint64_t bt_object; + uint64_t bt_offset; + uint64_t bt_txg; + uint64_t bt_thread; + uint64_t bt_seq; +} ztest_block_tag_t; + typedef struct ztest_args { - char *za_pool; + char za_pool[MAXNAMELEN]; + spa_t *za_spa; objset_t *za_os; zilog_t *za_zilog; thread_t za_thread; @@ -142,6 +151,13 @@ typedef struct ztest_args { hrtime_t za_stop; hrtime_t za_kill; traverse_handle_t *za_th; + /* + * Thread-local variables can go here to aid debugging. + */ + ztest_block_tag_t za_rbt; + ztest_block_tag_t za_wbt; + dmu_object_info_t za_doi; + dmu_buf_t *za_dbuf; } ztest_args_t; typedef void ztest_func_t(ztest_args_t *); @@ -160,14 +176,16 @@ ztest_func_t ztest_dmu_objset_create_destroy; ztest_func_t ztest_dmu_snapshot_create_destroy; ztest_func_t ztest_spa_create_destroy; ztest_func_t ztest_fault_inject; +ztest_func_t ztest_spa_rename; ztest_func_t ztest_vdev_attach_detach; ztest_func_t ztest_vdev_LUN_growth; ztest_func_t ztest_vdev_add_remove; +ztest_func_t ztest_vdev_aux_add_remove; ztest_func_t ztest_scrub; -ztest_func_t ztest_spa_rename; 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 */ uint64_t zi_calls; /* per-pass count */ uint64_t zi_call_time; /* per-pass time */ @@ -181,22 +199,23 @@ uint64_t zopt_sometimes = 10; /* every 10 seconds */ uint64_t zopt_rarely = 60; /* every 60 seconds */ ztest_info_t ztest_info[] = { - { ztest_dmu_read_write, &zopt_always }, - { ztest_dmu_write_parallel, &zopt_always }, - { ztest_dmu_object_alloc_free, &zopt_always }, - { ztest_zap, &zopt_always }, - { ztest_zap_parallel, &zopt_always }, - { ztest_traverse, &zopt_often }, - { ztest_dsl_prop_get_set, &zopt_sometimes }, - { ztest_dmu_objset_create_destroy, &zopt_sometimes }, - { ztest_dmu_snapshot_create_destroy, &zopt_rarely }, - { ztest_spa_create_destroy, &zopt_sometimes }, - { ztest_fault_inject, &zopt_sometimes }, - { ztest_spa_rename, &zopt_rarely }, - { ztest_vdev_attach_detach, &zopt_rarely }, - { ztest_vdev_LUN_growth, &zopt_rarely }, - { ztest_vdev_add_remove, &zopt_vdevtime }, - { ztest_scrub, &zopt_vdevtime }, + { ztest_dmu_read_write, 1, &zopt_always }, + { ztest_dmu_write_parallel, 30, &zopt_always }, + { ztest_dmu_object_alloc_free, 1, &zopt_always }, + { ztest_zap, 30, &zopt_always }, + { ztest_zap_parallel, 100, &zopt_always }, + { ztest_traverse, 1, &zopt_often }, + { ztest_dsl_prop_get_set, 1, &zopt_sometimes }, + { ztest_dmu_objset_create_destroy, 1, &zopt_sometimes }, + { ztest_dmu_snapshot_create_destroy, 1, &zopt_sometimes }, + { ztest_spa_create_destroy, 1, &zopt_sometimes }, + { ztest_fault_inject, 1, &zopt_sometimes }, + { ztest_spa_rename, 1, &zopt_rarely }, + { ztest_vdev_attach_detach, 1, &zopt_rarely }, + { ztest_vdev_LUN_growth, 1, &zopt_rarely }, + { ztest_vdev_add_remove, 1, &zopt_vdevtime }, + { ztest_vdev_aux_add_remove, 1, &zopt_vdevtime }, + { ztest_scrub, 1, &zopt_vdevtime }, }; #define ZTEST_FUNCS (sizeof (ztest_info) / sizeof (ztest_info_t)) @@ -210,34 +229,27 @@ typedef struct ztest_shared { mutex_t zs_vdev_lock; rwlock_t zs_name_lock; uint64_t zs_vdev_primaries; + uint64_t zs_vdev_aux; uint64_t zs_enospc_count; hrtime_t zs_start_time; hrtime_t zs_stop_time; uint64_t zs_alloc; uint64_t zs_space; - uint64_t zs_txg; ztest_info_t zs_info[ZTEST_FUNCS]; mutex_t zs_sync_lock[ZTEST_SYNC_LOCKS]; uint64_t zs_seq[ZTEST_SYNC_LOCKS]; } ztest_shared_t; -typedef struct ztest_block_tag { - uint64_t bt_objset; - uint64_t bt_object; - uint64_t bt_offset; - uint64_t bt_txg; - uint64_t bt_thread; - uint64_t bt_seq; -} ztest_block_tag_t; - static char ztest_dev_template[] = "%s/%s.%llua"; +static char ztest_aux_template[] = "%s/%s.%s.%llu"; static ztest_shared_t *ztest_shared; static int ztest_random_fd; static int ztest_dump_core = 1; -extern uint64_t zio_gang_bang; -extern uint16_t zio_zil_fail_shift; +static boolean_t ztest_exiting; + +extern uint64_t metaslab_gang_bang; #define ZTEST_DIROBJ 1 #define ZTEST_MICROZAP_OBJ 2 @@ -357,7 +369,7 @@ usage(boolean_t requested) FILE *fp = requested ? stdout : stderr; nicenum(zopt_vdev_size, nice_vdev_size); - nicenum(zio_gang_bang, nice_gang_bang); + nicenum(metaslab_gang_bang, nice_gang_bang); (void) fprintf(fp, "Usage: %s\n" "\t[-v vdevs (default: %llu)]\n" @@ -377,26 +389,24 @@ usage(boolean_t requested) "\t[-E(xisting)] (use existing pool instead of creating new one)\n" "\t[-T time] total run time (default: %llu sec)\n" "\t[-P passtime] time per pass (default: %llu sec)\n" - "\t[-z zil failure rate (default: fail every 2^%llu allocs)]\n" "\t[-h] (print help)\n" "", cmdname, - (u_longlong_t)zopt_vdevs, /* -v */ - nice_vdev_size, /* -s */ - zopt_ashift, /* -a */ - zopt_mirrors, /* -m */ - zopt_raidz, /* -r */ - zopt_raidz_parity, /* -R */ - zopt_datasets, /* -d */ - zopt_threads, /* -t */ - nice_gang_bang, /* -g */ - zopt_init, /* -i */ - (u_longlong_t)zopt_killrate, /* -k */ - zopt_pool, /* -p */ - zopt_dir, /* -f */ - (u_longlong_t)zopt_time, /* -T */ - (u_longlong_t)zopt_passtime, /* -P */ - (u_longlong_t)zio_zil_fail_shift); /* -z */ + (u_longlong_t)zopt_vdevs, /* -v */ + nice_vdev_size, /* -s */ + zopt_ashift, /* -a */ + zopt_mirrors, /* -m */ + zopt_raidz, /* -r */ + zopt_raidz_parity, /* -R */ + zopt_datasets, /* -d */ + zopt_threads, /* -t */ + nice_gang_bang, /* -g */ + zopt_init, /* -i */ + (u_longlong_t)zopt_killrate, /* -k */ + zopt_pool, /* -p */ + zopt_dir, /* -f */ + (u_longlong_t)zopt_time, /* -T */ + (u_longlong_t)zopt_passtime); /* -P */ exit(requested ? 0 : 1); } @@ -431,91 +441,84 @@ process_options(int argc, char **argv) progname = argv[0]; /* By default, test gang blocks for blocks 32K and greater */ - zio_gang_bang = 32 << 10; - - /* Default value, fail every 32nd allocation */ - zio_zil_fail_shift = 5; + metaslab_gang_bang = 32 << 10; while ((opt = getopt(argc, argv, - "v:s:a:m:r:R:d:t:g:i:k:p:f:VET:P:z:h")) != EOF) { + "v:s:a:m:r:R:d:t:g:i:k:p:f:VET:P:h")) != EOF) { value = 0; switch (opt) { - case 'v': - case 's': - case 'a': - case 'm': - case 'r': - case 'R': - case 'd': - case 't': - case 'g': - case 'i': - case 'k': - case 'T': - case 'P': - case 'z': + case 'v': + case 's': + case 'a': + case 'm': + case 'r': + case 'R': + case 'd': + case 't': + case 'g': + case 'i': + case 'k': + case 'T': + case 'P': value = nicenumtoull(optarg); } switch (opt) { - case 'v': + case 'v': zopt_vdevs = value; break; - case 's': + case 's': zopt_vdev_size = MAX(SPA_MINDEVSIZE, value); break; - case 'a': + case 'a': zopt_ashift = value; break; - case 'm': + case 'm': zopt_mirrors = value; break; - case 'r': + case 'r': zopt_raidz = MAX(1, value); break; - case 'R': + case 'R': zopt_raidz_parity = MIN(MAX(value, 1), 2); break; - case 'd': + case 'd': zopt_datasets = MAX(1, value); break; - case 't': + case 't': zopt_threads = MAX(1, value); break; - case 'g': - zio_gang_bang = MAX(SPA_MINBLOCKSIZE << 1, value); + case 'g': + metaslab_gang_bang = MAX(SPA_MINBLOCKSIZE << 1, value); break; - case 'i': + case 'i': zopt_init = value; break; - case 'k': + case 'k': zopt_killrate = value; break; - case 'p': + case 'p': zopt_pool = strdup(optarg); break; - case 'f': + case 'f': zopt_dir = strdup(optarg); break; - case 'V': + case 'V': zopt_verbose++; break; - case 'E': + case 'E': zopt_init = 0; break; - case 'T': + case 'T': zopt_time = value; break; - case 'P': + case 'P': zopt_passtime = MAX(1, value); break; - case 'z': - zio_zil_fail_shift = MIN(value, 16); - break; - case 'h': + case 'h': usage(B_TRUE); break; - case '?': - default: + case '?': + default: usage(B_FALSE); break; } @@ -536,51 +539,58 @@ ztest_get_ashift(void) } static nvlist_t * -make_vdev_file(size_t size) +make_vdev_file(char *path, char *aux, size_t size, uint64_t ashift) { - char dev_name[MAXPATHLEN]; + char pathbuf[MAXPATHLEN]; uint64_t vdev; - uint64_t ashift = ztest_get_ashift(); - int fd; nvlist_t *file; - if (size == 0) { - (void) snprintf(dev_name, sizeof (dev_name), "%s", - "/dev/bogus"); - } else { - vdev = ztest_shared->zs_vdev_primaries++; - (void) sprintf(dev_name, ztest_dev_template, - zopt_dir, zopt_pool, vdev); + if (ashift == 0) + ashift = ztest_get_ashift(); + + if (path == NULL) { + path = pathbuf; + + if (aux != NULL) { + vdev = ztest_shared->zs_vdev_aux; + (void) sprintf(path, ztest_aux_template, + zopt_dir, zopt_pool, aux, vdev); + } else { + vdev = ztest_shared->zs_vdev_primaries++; + (void) sprintf(path, ztest_dev_template, + zopt_dir, zopt_pool, vdev); + } + } - fd = open(dev_name, O_RDWR | O_CREAT | O_TRUNC, 0666); + if (size != 0) { + int fd = open(path, O_RDWR | O_CREAT | O_TRUNC, 0666); if (fd == -1) - fatal(1, "can't open %s", dev_name); + fatal(1, "can't open %s", path); if (ftruncate(fd, size) != 0) - fatal(1, "can't ftruncate %s", dev_name); + fatal(1, "can't ftruncate %s", path); (void) close(fd); } VERIFY(nvlist_alloc(&file, NV_UNIQUE_NAME, 0) == 0); VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_TYPE, VDEV_TYPE_FILE) == 0); - VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_PATH, dev_name) == 0); + VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_PATH, path) == 0); VERIFY(nvlist_add_uint64(file, ZPOOL_CONFIG_ASHIFT, ashift) == 0); return (file); } static nvlist_t * -make_vdev_raidz(size_t size, int r) +make_vdev_raidz(char *path, char *aux, size_t size, uint64_t ashift, int r) { nvlist_t *raidz, **child; int c; if (r < 2) - return (make_vdev_file(size)); - + return (make_vdev_file(path, aux, size, ashift)); child = umem_alloc(r * sizeof (nvlist_t *), UMEM_NOFAIL); for (c = 0; c < r; c++) - child[c] = make_vdev_file(size); + child[c] = make_vdev_file(path, aux, size, ashift); VERIFY(nvlist_alloc(&raidz, NV_UNIQUE_NAME, 0) == 0); VERIFY(nvlist_add_string(raidz, ZPOOL_CONFIG_TYPE, @@ -599,18 +609,19 @@ make_vdev_raidz(size_t size, int r) } static nvlist_t * -make_vdev_mirror(size_t size, int r, int m) +make_vdev_mirror(char *path, char *aux, size_t size, uint64_t ashift, + int r, int m) { nvlist_t *mirror, **child; int c; if (m < 1) - return (make_vdev_raidz(size, r)); + return (make_vdev_raidz(path, aux, size, ashift, r)); child = umem_alloc(m * sizeof (nvlist_t *), UMEM_NOFAIL); for (c = 0; c < m; c++) - child[c] = make_vdev_raidz(size, r); + child[c] = make_vdev_raidz(path, aux, size, ashift, r); VERIFY(nvlist_alloc(&mirror, NV_UNIQUE_NAME, 0) == 0); VERIFY(nvlist_add_string(mirror, ZPOOL_CONFIG_TYPE, @@ -627,7 +638,8 @@ make_vdev_mirror(size_t size, int r, int m) } static nvlist_t * -make_vdev_root(size_t size, int r, int m, int t) +make_vdev_root(char *path, char *aux, size_t size, uint64_t ashift, + int log, int r, int m, int t) { nvlist_t *root, **child; int c; @@ -636,12 +648,15 @@ make_vdev_root(size_t size, int r, int m, int t) child = umem_alloc(t * sizeof (nvlist_t *), UMEM_NOFAIL); - for (c = 0; c < t; c++) - child[c] = make_vdev_mirror(size, r, m); + for (c = 0; c < t; c++) { + child[c] = make_vdev_mirror(path, aux, size, ashift, r, m); + VERIFY(nvlist_add_uint64(child[c], ZPOOL_CONFIG_IS_LOG, + log) == 0); + } VERIFY(nvlist_alloc(&root, NV_UNIQUE_NAME, 0) == 0); VERIFY(nvlist_add_string(root, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT) == 0); - VERIFY(nvlist_add_nvlist_array(root, ZPOOL_CONFIG_CHILDREN, + VERIFY(nvlist_add_nvlist_array(root, aux ? aux : ZPOOL_CONFIG_CHILDREN, child, t) == 0); for (c = 0; c < t; c++) @@ -785,8 +800,8 @@ ztest_spa_create_destroy(ztest_args_t *za) /* * Attempt to create using a bad file. */ - nvroot = make_vdev_root(0, 0, 0, 1); - error = spa_create("ztest_bad_file", nvroot, NULL); + nvroot = make_vdev_root("/dev/bogus", NULL, 0, 0, 0, 0, 0, 1); + error = spa_create("ztest_bad_file", nvroot, NULL, NULL, NULL); nvlist_free(nvroot); if (error != ENOENT) fatal(0, "spa_create(bad_file) = %d", error); @@ -794,8 +809,8 @@ ztest_spa_create_destroy(ztest_args_t *za) /* * Attempt to create using a bad mirror. */ - nvroot = make_vdev_root(0, 0, 2, 1); - error = spa_create("ztest_bad_mirror", nvroot, NULL); + nvroot = make_vdev_root("/dev/bogus", NULL, 0, 0, 0, 0, 2, 1); + error = spa_create("ztest_bad_mirror", nvroot, NULL, NULL, NULL); nvlist_free(nvroot); if (error != ENOENT) fatal(0, "spa_create(bad_mirror) = %d", error); @@ -805,8 +820,8 @@ ztest_spa_create_destroy(ztest_args_t *za) * what's in the nvroot; we should fail with EEXIST. */ (void) rw_rdlock(&ztest_shared->zs_name_lock); - nvroot = make_vdev_root(0, 0, 0, 1); - error = spa_create(za->za_pool, nvroot, NULL); + nvroot = make_vdev_root("/dev/bogus", NULL, 0, 0, 0, 0, 0, 1); + error = spa_create(za->za_pool, nvroot, NULL, NULL, NULL); nvlist_free(nvroot); if (error != EEXIST) fatal(0, "spa_create(whatever) = %d", error); @@ -823,30 +838,48 @@ ztest_spa_create_destroy(ztest_args_t *za) (void) rw_unlock(&ztest_shared->zs_name_lock); } +static vdev_t * +vdev_lookup_by_path(vdev_t *vd, const char *path) +{ + vdev_t *mvd; + + if (vd->vdev_path != NULL && strcmp(path, vd->vdev_path) == 0) + return (vd); + + for (int c = 0; c < vd->vdev_children; c++) + if ((mvd = vdev_lookup_by_path(vd->vdev_child[c], path)) != + NULL) + return (mvd); + + return (NULL); +} + /* * Verify that vdev_add() works as expected. */ void ztest_vdev_add_remove(ztest_args_t *za) { - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; uint64_t leaves = MAX(zopt_mirrors, 1) * zopt_raidz; nvlist_t *nvroot; int error; - if (zopt_verbose >= 6) - (void) printf("adding vdev\n"); - (void) mutex_lock(&ztest_shared->zs_vdev_lock); - spa_config_enter(spa, RW_READER, FTAG); + spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); ztest_shared->zs_vdev_primaries = spa->spa_root_vdev->vdev_children * leaves; - spa_config_exit(spa, FTAG); + spa_config_exit(spa, SCL_VDEV, FTAG); + + /* + * Make 1/4 of the devices be log devices. + */ + nvroot = make_vdev_root(NULL, NULL, zopt_vdev_size, 0, + ztest_random(4) == 0, zopt_raidz, zopt_mirrors, 1); - nvroot = make_vdev_root(zopt_vdev_size, zopt_raidz, zopt_mirrors, 1); error = spa_vdev_add(spa, nvroot); nvlist_free(nvroot); @@ -856,37 +889,86 @@ ztest_vdev_add_remove(ztest_args_t *za) ztest_record_enospc("spa_vdev_add"); else if (error != 0) fatal(0, "spa_vdev_add() = %d", error); - - if (zopt_verbose >= 6) - (void) printf("spa_vdev_add = %d, as expected\n", error); } -static vdev_t * -vdev_lookup_by_path(vdev_t *vd, const char *path) +/* + * Verify that adding/removing aux devices (l2arc, hot spare) works as expected. + */ +void +ztest_vdev_aux_add_remove(ztest_args_t *za) { - int c; - vdev_t *mvd; + spa_t *spa = za->za_spa; + vdev_t *rvd = spa->spa_root_vdev; + spa_aux_vdev_t *sav; + char *aux; + uint64_t guid = 0; + int error; - if (vd->vdev_path != NULL) { - if (vd->vdev_wholedisk == 1) { - /* - * For whole disks, the internal path has 's0', but the - * path passed in by the user doesn't. - */ - if (strlen(path) == strlen(vd->vdev_path) - 2 && - strncmp(path, vd->vdev_path, strlen(path)) == 0) - return (vd); - } else if (strcmp(path, vd->vdev_path) == 0) { - return (vd); + if (ztest_random(2) == 0) { + sav = &spa->spa_spares; + aux = ZPOOL_CONFIG_SPARES; + } else { + sav = &spa->spa_l2cache; + aux = ZPOOL_CONFIG_L2CACHE; + } + + (void) mutex_lock(&ztest_shared->zs_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. + */ + guid = sav->sav_vdevs[ztest_random(sav->sav_count)]->vdev_guid; + } else { + /* + * Find an unused device we can add. + */ + ztest_shared->zs_vdev_aux = 0; + for (;;) { + char path[MAXPATHLEN]; + int c; + (void) sprintf(path, ztest_aux_template, zopt_dir, + zopt_pool, aux, ztest_shared->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; + ztest_shared->zs_vdev_aux++; } } - for (c = 0; c < vd->vdev_children; c++) - if ((mvd = vdev_lookup_by_path(vd->vdev_child[c], path)) != - NULL) - return (mvd); + spa_config_exit(spa, SCL_VDEV, FTAG); - return (NULL); + if (guid == 0) { + /* + * Add a new device. + */ + nvlist_t *nvroot = make_vdev_root(NULL, aux, + (zopt_vdev_size * 5) / 4, 0, 0, 0, 0, 1); + error = spa_vdev_add(spa, nvroot); + if (error != 0) + fatal(0, "spa_vdev_add(%p) = %d", nvroot, error); + nvlist_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, B_FALSE, NULL); + + error = spa_vdev_remove(spa, guid, B_FALSE); + if (error != 0 && error != EBUSY) + fatal(0, "spa_vdev_remove(%llu) = %d", guid, error); + } + + (void) mutex_unlock(&ztest_shared->zs_vdev_lock); } /* @@ -895,22 +977,26 @@ vdev_lookup_by_path(vdev_t *vd, const char *path) void ztest_vdev_attach_detach(ztest_args_t *za) { - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; + spa_aux_vdev_t *sav = &spa->spa_spares; vdev_t *rvd = spa->spa_root_vdev; vdev_t *oldvd, *newvd, *pvd; - nvlist_t *root, *file; + nvlist_t *root; uint64_t leaves = MAX(zopt_mirrors, 1) * zopt_raidz; uint64_t leaf, top; uint64_t ashift = ztest_get_ashift(); + uint64_t oldguid; size_t oldsize, newsize; char oldpath[MAXPATHLEN], newpath[MAXPATHLEN]; int replacing; + int oldvd_has_siblings = B_FALSE; + int newvd_is_spare = B_FALSE; + int oldvd_is_log; int error, expected_error; - int fd; (void) mutex_lock(&ztest_shared->zs_vdev_lock); - spa_config_enter(spa, RW_READER, FTAG); + spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); /* * Decide whether to do an attach or a replace. @@ -928,39 +1014,70 @@ ztest_vdev_attach_detach(ztest_args_t *za) leaf = ztest_random(leaves); /* - * Generate the path to this leaf. The filename will end with 'a'. - * We'll alternate replacements with a filename that ends with 'b'. + * Locate this vdev. */ - (void) snprintf(oldpath, sizeof (oldpath), - ztest_dev_template, zopt_dir, zopt_pool, top * leaves + leaf); - - bcopy(oldpath, newpath, MAXPATHLEN); + oldvd = rvd->vdev_child[top]; + if (zopt_mirrors >= 1) + oldvd = oldvd->vdev_child[leaf / zopt_raidz]; + if (zopt_raidz > 1) + oldvd = oldvd->vdev_child[leaf % zopt_raidz]; /* - * If the 'a' file isn't part of the pool, the 'b' file must be. + * If we're already doing an attach or replace, oldvd may be a + * mirror vdev -- in which case, pick a random child. */ - if (vdev_lookup_by_path(rvd, oldpath) == NULL) - oldpath[strlen(oldpath) - 1] = 'b'; - else - newpath[strlen(newpath) - 1] = 'b'; + while (oldvd->vdev_children != 0) { + oldvd_has_siblings = B_TRUE; + ASSERT(oldvd->vdev_children == 2); + oldvd = oldvd->vdev_child[ztest_random(2)]; + } + + oldguid = oldvd->vdev_guid; + oldsize = vdev_get_rsize(oldvd); + oldvd_is_log = oldvd->vdev_top->vdev_islog; + (void) strcpy(oldpath, oldvd->vdev_path); + pvd = oldvd->vdev_parent; /* - * Now oldpath represents something that's already in the pool, - * and newpath is the thing we'll try to attach. + * If oldvd has siblings, then half of the time, detach it. */ - oldvd = vdev_lookup_by_path(rvd, oldpath); - newvd = vdev_lookup_by_path(rvd, newpath); - ASSERT(oldvd != NULL); - pvd = oldvd->vdev_parent; + if (oldvd_has_siblings && ztest_random(2) == 0) { + spa_config_exit(spa, SCL_VDEV, FTAG); + error = spa_vdev_detach(spa, oldguid, B_FALSE); + if (error != 0 && error != ENODEV && error != EBUSY) + fatal(0, "detach (%s) returned %d", + oldpath, error); + (void) mutex_unlock(&ztest_shared->zs_vdev_lock); + return; + } /* - * 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. + * For the new vdev, choose with equal probability between the two + * standard paths (ending in either 'a' or 'b') or a random hot spare. */ - oldsize = vdev_get_rsize(oldvd); - newsize = 10 * oldsize / (9 + ztest_random(3)); + if (sav->sav_count != 0 && ztest_random(3) == 0) { + newvd = sav->sav_vdevs[ztest_random(sav->sav_count)]; + newvd_is_spare = B_TRUE; + (void) strcpy(newpath, newvd->vdev_path); + } else { + (void) snprintf(newpath, sizeof (newpath), ztest_dev_template, + zopt_dir, zopt_pool, top * leaves + leaf); + if (ztest_random(2) == 0) + newpath[strlen(newpath) - 1] = 'b'; + newvd = vdev_lookup_by_path(rvd, newpath); + } + + if (newvd) { + newsize = vdev_get_rsize(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, @@ -970,12 +1087,17 @@ ztest_vdev_attach_detach(ztest_args_t *za) * * If newvd is too small, it should fail with EOVERFLOW. */ - if (newvd != NULL) - expected_error = EBUSY; - else if (pvd->vdev_ops != &vdev_mirror_ops && - pvd->vdev_ops != &vdev_root_ops && - (!replacing || pvd->vdev_ops == &vdev_replacing_ops)) + 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)) 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 (newsize < oldsize) expected_error = EOVERFLOW; else if (ashift > oldvd->vdev_top->vdev_ashift) @@ -983,36 +1105,16 @@ ztest_vdev_attach_detach(ztest_args_t *za) else expected_error = 0; - /* - * If newvd isn't already part of the pool, create it. - */ - if (newvd == NULL) { - fd = open(newpath, O_RDWR | O_CREAT | O_TRUNC, 0666); - if (fd == -1) - fatal(1, "can't open %s", newpath); - if (ftruncate(fd, newsize) != 0) - fatal(1, "can't ftruncate %s", newpath); - (void) close(fd); - } - - spa_config_exit(spa, FTAG); + spa_config_exit(spa, SCL_VDEV, FTAG); /* * Build the nvlist describing newpath. */ - VERIFY(nvlist_alloc(&file, NV_UNIQUE_NAME, 0) == 0); - VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_TYPE, VDEV_TYPE_FILE) == 0); - VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_PATH, newpath) == 0); - VERIFY(nvlist_add_uint64(file, ZPOOL_CONFIG_ASHIFT, ashift) == 0); + root = make_vdev_root(newpath, NULL, newvd == NULL ? newsize : 0, + ashift, 0, 0, 0, 1); - VERIFY(nvlist_alloc(&root, NV_UNIQUE_NAME, 0) == 0); - VERIFY(nvlist_add_string(root, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT) == 0); - VERIFY(nvlist_add_nvlist_array(root, ZPOOL_CONFIG_CHILDREN, - &file, 1) == 0); - - error = spa_vdev_attach(spa, oldvd->vdev_guid, root, replacing); + error = spa_vdev_attach(spa, oldguid, root, replacing); - nvlist_free(file); nvlist_free(root); /* @@ -1027,12 +1129,15 @@ ztest_vdev_attach_detach(ztest_args_t *za) /* * If someone grew the LUN, the replacement may be too small. */ - if (error == EOVERFLOW) + if (error == EOVERFLOW || error == EBUSY) expected_error = error; - if (error != expected_error) { - fatal(0, "attach (%s, %s, %d) returned %d, expected %d", - oldpath, newpath, replacing, error, expected_error); + /* XXX workaround 6690467 */ + if (error != expected_error && expected_error != EBUSY) { + fatal(0, "attach (%s %llu, %s %llu, %d) " + "returned %d, expected %d", + oldpath, (longlong_t)oldsize, newpath, + (longlong_t)newsize, replacing, error, expected_error); } (void) mutex_unlock(&ztest_shared->zs_vdev_lock); @@ -1045,7 +1150,7 @@ ztest_vdev_attach_detach(ztest_args_t *za) void ztest_vdev_LUN_growth(ztest_args_t *za) { - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; char dev_name[MAXPATHLEN]; uint64_t leaves = MAX(zopt_mirrors, 1) * zopt_raidz; uint64_t vdev; @@ -1057,9 +1162,9 @@ ztest_vdev_LUN_growth(ztest_args_t *za) /* * Pick a random leaf vdev. */ - spa_config_enter(spa, RW_READER, FTAG); + spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); vdev = ztest_random(spa->spa_root_vdev->vdev_children * leaves); - spa_config_exit(spa, FTAG); + spa_config_exit(spa, SCL_VDEV, FTAG); (void) sprintf(dev_name, ztest_dev_template, zopt_dir, zopt_pool, vdev); @@ -1088,14 +1193,14 @@ ztest_vdev_LUN_growth(ztest_args_t *za) /* ARGSUSED */ static void -ztest_create_cb(objset_t *os, void *arg, dmu_tx_t *tx) +ztest_create_cb(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx) { /* * Create the directory object. */ VERIFY(dmu_object_claim(os, ZTEST_DIROBJ, DMU_OT_UINT64_OTHER, ZTEST_DIROBJ_BLOCKSIZE, - DMU_OT_UINT64_OTHER, sizeof (ztest_block_tag_t), tx) == 0); + DMU_OT_UINT64_OTHER, 5 * sizeof (ztest_block_tag_t), tx) == 0); VERIFY(zap_create_claim(os, ZTEST_MICROZAP_OBJ, DMU_OT_ZAP_OTHER, DMU_OT_NONE, 0, tx) == 0); @@ -1104,26 +1209,26 @@ ztest_create_cb(objset_t *os, void *arg, dmu_tx_t *tx) DMU_OT_ZAP_OTHER, DMU_OT_NONE, 0, tx) == 0); } -/* ARGSUSED */ static int ztest_destroy_cb(char *name, void *arg) { + ztest_args_t *za = arg; objset_t *os; - dmu_object_info_t doi; + dmu_object_info_t *doi = &za->za_doi; int error; /* * Verify that the dataset contains a directory object. */ error = dmu_objset_open(name, DMU_OST_OTHER, - DS_MODE_STANDARD | DS_MODE_READONLY, &os); + DS_MODE_USER | DS_MODE_READONLY, &os); ASSERT3U(error, ==, 0); - error = dmu_object_info(os, ZTEST_DIROBJ, &doi); + error = dmu_object_info(os, ZTEST_DIROBJ, doi); if (error != ENOENT) { /* We could have crashed in the middle of destroying it */ ASSERT3U(error, ==, 0); - ASSERT3U(doi.doi_type, ==, DMU_OT_UINT64_OTHER); - ASSERT3S(doi.doi_physical_blks, >=, 0); + ASSERT3U(doi->doi_type, ==, DMU_OT_UINT64_OTHER); + ASSERT3S(doi->doi_physical_blks, >=, 0); } dmu_objset_close(os); @@ -1131,7 +1236,11 @@ ztest_destroy_cb(char *name, void *arg) * Destroy the dataset. */ error = dmu_objset_destroy(name); - ASSERT3U(error, ==, 0); + if (error) { + (void) dmu_objset_open(name, DMU_OST_OTHER, + DS_MODE_USER | DS_MODE_READONLY, &os); + fatal(0, "dmu_objset_destroy(os=%p) = %d\n", &os, error); + } return (0); } @@ -1171,9 +1280,9 @@ void ztest_dmu_objset_create_destroy(ztest_args_t *za) { int error; - objset_t *os; + objset_t *os, *os2; char name[100]; - int mode, basemode, expected_error; + int basemode, expected_error; zilog_t *zilog; uint64_t seq; uint64_t objects; @@ -1183,9 +1292,9 @@ ztest_dmu_objset_create_destroy(ztest_args_t *za) (void) snprintf(name, 100, "%s/%s_temp_%llu", za->za_pool, za->za_pool, (u_longlong_t)za->za_instance); - basemode = DS_MODE_LEVEL(za->za_instance); - if (basemode == DS_MODE_NONE) - basemode++; + basemode = DS_MODE_TYPE(za->za_instance); + if (basemode != DS_MODE_USER && basemode != DS_MODE_OWNER) + basemode = DS_MODE_USER; /* * If this dataset exists from a previous run, process its replay log @@ -1193,9 +1302,9 @@ ztest_dmu_objset_create_destroy(ztest_args_t *za) * (invoked from ztest_destroy_cb() below) should just throw it away. */ if (ztest_random(2) == 0 && - dmu_objset_open(name, DMU_OST_OTHER, DS_MODE_PRIMARY, &os) == 0) { + dmu_objset_open(name, DMU_OST_OTHER, DS_MODE_OWNER, &os) == 0) { zr.zr_os = os; - zil_replay(os, &zr, &zr.zr_assign, ztest_replay_vector); + zil_replay(os, &zr, &zr.zr_assign, ztest_replay_vector, NULL); dmu_objset_close(os); } @@ -1204,7 +1313,7 @@ ztest_dmu_objset_create_destroy(ztest_args_t *za) * create lying around from a previous run. If so, destroy it * and all of its snapshots. */ - (void) dmu_objset_find(name, ztest_destroy_cb, NULL, + (void) dmu_objset_find(name, ztest_destroy_cb, za, DS_FIND_CHILDREN | DS_FIND_SNAPSHOTS); /* @@ -1218,8 +1327,8 @@ ztest_dmu_objset_create_destroy(ztest_args_t *za) /* * Verify that we can create a new dataset. */ - error = dmu_objset_create(name, DMU_OST_OTHER, NULL, ztest_create_cb, - NULL); + error = dmu_objset_create(name, DMU_OST_OTHER, NULL, 0, + ztest_create_cb, NULL); if (error) { if (error == ENOSPC) { ztest_record_enospc("dmu_objset_create"); @@ -1274,26 +1383,29 @@ ztest_dmu_objset_create_destroy(ztest_args_t *za) /* * Verify that we cannot create an existing dataset. */ - error = dmu_objset_create(name, DMU_OST_OTHER, NULL, NULL, NULL); + error = dmu_objset_create(name, DMU_OST_OTHER, NULL, 0, NULL, NULL); if (error != EEXIST) fatal(0, "created existing dataset, error = %d", error); /* - * Verify that multiple dataset opens are allowed, but only when + * Verify that multiple dataset holds are allowed, but only when * the new access mode is compatible with the base mode. - * We use a mixture of typed and typeless opens, and when the - * open succeeds, verify that the discovered type is correct. - */ - for (mode = DS_MODE_STANDARD; mode < DS_MODE_LEVELS; mode++) { - objset_t *os2; - error = dmu_objset_open(name, DMU_OST_OTHER, mode, &os2); - expected_error = (basemode + mode < DS_MODE_LEVELS) ? 0 : EBUSY; - if (error != expected_error) - fatal(0, "dmu_objset_open('%s') = %d, expected %d", - name, error, expected_error); - if (error == 0) + */ + if (basemode == DS_MODE_OWNER) { + error = dmu_objset_open(name, DMU_OST_OTHER, DS_MODE_USER, + &os2); + if (error) + fatal(0, "dmu_objset_open('%s') = %d", name, error); + else dmu_objset_close(os2); } + error = dmu_objset_open(name, DMU_OST_OTHER, DS_MODE_OWNER, &os2); + expected_error = (basemode == DS_MODE_OWNER) ? EBUSY : 0; + if (error != expected_error) + fatal(0, "dmu_objset_open('%s') = %d, expected %d", + name, error, expected_error); + if (error == 0) + dmu_objset_close(os2); zil_close(zilog); dmu_objset_close(os); @@ -1417,7 +1529,7 @@ ztest_blk_cb(traverse_blk_cache_t *bc, spa_t *spa, void *arg) void ztest_traverse(ztest_args_t *za) { - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; traverse_handle_t *th = za->za_th; int rc, advance; uint64_t cbstart, cblimit; @@ -1489,7 +1601,7 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) dmu_tx_t *tx; uint64_t batchobj, object, batchsize, endoff, temp; int b, c, error, bonuslen; - dmu_object_info_t doi; + dmu_object_info_t *doi = &za->za_doi; char osname[MAXNAMELEN]; dmu_objset_name(os, osname); @@ -1500,7 +1612,7 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) /* * Create a batch object if necessary, and record it in the directory. */ - VERIFY(0 == dmu_read(os, ZTEST_DIROBJ, za->za_diroff, + VERIFY3U(0, ==, dmu_read(os, ZTEST_DIROBJ, za->za_diroff, sizeof (uint64_t), &batchobj)); if (batchobj == 0) { tx = dmu_tx_create(os); @@ -1525,7 +1637,7 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) * Destroy the previous batch of objects. */ for (b = 0; b < batchsize; b++) { - VERIFY(0 == dmu_read(os, batchobj, b * sizeof (uint64_t), + VERIFY3U(0, ==, dmu_read(os, batchobj, b * sizeof (uint64_t), sizeof (uint64_t), &object)); if (object == 0) continue; @@ -1534,13 +1646,14 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) * We expect the nth byte of the bonus buffer to be n. */ VERIFY(0 == dmu_bonus_hold(os, object, FTAG, &db)); + za->za_dbuf = db; - dmu_object_info_from_db(db, &doi); - ASSERT(doi.doi_type == DMU_OT_UINT64_OTHER); - ASSERT(doi.doi_bonus_type == DMU_OT_PLAIN_OTHER); - ASSERT3S(doi.doi_physical_blks, >=, 0); + dmu_object_info_from_db(db, doi); + ASSERT(doi->doi_type == DMU_OT_UINT64_OTHER); + ASSERT(doi->doi_bonus_type == DMU_OT_PLAIN_OTHER); + ASSERT3S(doi->doi_physical_blks, >=, 0); - bonuslen = db->db_size; + bonuslen = doi->doi_bonus_size; for (c = 0; c < bonuslen; c++) { if (((uint8_t *)db->db_data)[c] != @@ -1554,6 +1667,7 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) } dmu_buf_rele(db, FTAG); + za->za_dbuf = NULL; /* * We expect the word at endoff to be our object number. @@ -1658,8 +1772,9 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) /* * Write to both the bonus buffer and the regular data. */ - VERIFY(0 == dmu_bonus_hold(os, object, FTAG, &db)); - ASSERT3U(bonuslen, ==, db->db_size); + VERIFY(dmu_bonus_hold(os, object, FTAG, &db) == 0); + za->za_dbuf = db; + ASSERT3U(bonuslen, <=, db->db_size); dmu_object_size_from_db(db, &va_blksize, &va_nblocks); ASSERT3S(va_nblocks, >=, 0); @@ -1670,10 +1785,11 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) * See comments above regarding the contents of * the bonus buffer and the word at endoff. */ - for (c = 0; c < db->db_size; c++) + for (c = 0; c < bonuslen; c++) ((uint8_t *)db->db_data)[c] = (uint8_t)(c + bonuslen); dmu_buf_rele(db, FTAG); + za->za_dbuf = NULL; /* * Write to a large offset to increase indirection. @@ -1928,226 +2044,240 @@ ztest_dmu_read_write(ztest_args_t *za) } void -ztest_dmu_check_future_leak(objset_t *os, uint64_t txg) +ztest_dmu_check_future_leak(ztest_args_t *za) { + objset_t *os = za->za_os; dmu_buf_t *db; - ztest_block_tag_t rbt; - - if (zopt_verbose >= 3) { - char osname[MAXNAMELEN]; - dmu_objset_name(os, osname); - (void) printf("checking %s for future leaks in txg %lld...\n", - osname, (u_longlong_t)txg); - } + ztest_block_tag_t *bt; + dmu_object_info_t *doi = &za->za_doi; /* * Make sure that, if there is a write record in the bonus buffer * of the ZTEST_DIROBJ, that the txg for this record is <= the * last synced txg of the pool. */ - - VERIFY(0 == dmu_bonus_hold(os, ZTEST_DIROBJ, FTAG, &db)); - ASSERT3U(db->db_size, ==, sizeof (rbt)); - bcopy(db->db_data, &rbt, db->db_size); - if (rbt.bt_objset != 0) { - ASSERT3U(rbt.bt_objset, ==, dmu_objset_id(os)); - ASSERT3U(rbt.bt_object, ==, ZTEST_DIROBJ); - ASSERT3U(rbt.bt_offset, ==, -1ULL); - if (rbt.bt_txg > txg) { - fatal(0, - "future leak: got %llx, last synced txg is %llx", - rbt.bt_txg, txg); - } + VERIFY(dmu_bonus_hold(os, ZTEST_DIROBJ, FTAG, &db) == 0); + za->za_dbuf = db; + VERIFY(dmu_object_info(os, ZTEST_DIROBJ, doi) == 0); + ASSERT3U(doi->doi_bonus_size, >=, sizeof (*bt)); + ASSERT3U(doi->doi_bonus_size, <=, db->db_size); + ASSERT3U(doi->doi_bonus_size % sizeof (*bt), ==, 0); + bt = (void *)((char *)db->db_data + doi->doi_bonus_size - sizeof (*bt)); + if (bt->bt_objset != 0) { + ASSERT3U(bt->bt_objset, ==, dmu_objset_id(os)); + ASSERT3U(bt->bt_object, ==, ZTEST_DIROBJ); + ASSERT3U(bt->bt_offset, ==, -1ULL); + ASSERT3U(bt->bt_txg, <, spa_first_txg(za->za_spa)); } dmu_buf_rele(db, FTAG); + za->za_dbuf = NULL; } void ztest_dmu_write_parallel(ztest_args_t *za) { objset_t *os = za->za_os; - dmu_tx_t *tx; + ztest_block_tag_t *rbt = &za->za_rbt; + ztest_block_tag_t *wbt = &za->za_wbt; + const size_t btsize = sizeof (ztest_block_tag_t); dmu_buf_t *db; - int i, b, error, do_free, bs; - uint64_t off, txg_how, txg; + int b, error; + int bs = ZTEST_DIROBJ_BLOCKSIZE; + int do_free = 0; + uint64_t off, txg, txg_how; mutex_t *lp; char osname[MAXNAMELEN]; char iobuf[SPA_MAXBLOCKSIZE]; - ztest_block_tag_t rbt, wbt; + blkptr_t blk = { 0 }; + uint64_t blkoff; + zbookmark_t zb; + dmu_tx_t *tx = dmu_tx_create(os); dmu_objset_name(os, osname); - bs = ZTEST_DIROBJ_BLOCKSIZE; /* * Have multiple threads write to large offsets in ZTEST_DIROBJ * to verify that having multiple threads writing to the same object * in parallel doesn't cause any trouble. - * Also do parallel writes to the bonus buffer on occasion. */ - for (i = 0; i < 50; i++) { + if (ztest_random(4) == 0) { + /* + * Do the bonus buffer instead of a regular block. + * We need a lock to serialize resize vs. others, + * so we hash on the objset ID. + */ + b = dmu_objset_id(os) % ZTEST_SYNC_LOCKS; + off = -1ULL; + dmu_tx_hold_bonus(tx, ZTEST_DIROBJ); + } else { b = ztest_random(ZTEST_SYNC_LOCKS); - lp = &ztest_shared->zs_sync_lock[b]; - - do_free = (ztest_random(4) == 0); - - off = za->za_diroff_shared + ((uint64_t)b << SPA_MAXBLOCKSHIFT); - + off = za->za_diroff_shared + (b << SPA_MAXBLOCKSHIFT); if (ztest_random(4) == 0) { - /* - * Do the bonus buffer instead of a regular block. - */ - do_free = 0; - off = -1ULL; - } - - tx = dmu_tx_create(os); - - if (off == -1ULL) - dmu_tx_hold_bonus(tx, ZTEST_DIROBJ); - else if (do_free) + do_free = 1; dmu_tx_hold_free(tx, ZTEST_DIROBJ, off, bs); - else + } else { dmu_tx_hold_write(tx, ZTEST_DIROBJ, off, bs); + } + } - txg_how = ztest_random(2) == 0 ? TXG_WAIT : TXG_NOWAIT; - error = dmu_tx_assign(tx, txg_how); - if (error) { - if (error == ERESTART) { - ASSERT(txg_how == TXG_NOWAIT); - dmu_tx_wait(tx); - dmu_tx_abort(tx); - continue; - } - dmu_tx_abort(tx); + txg_how = ztest_random(2) == 0 ? TXG_WAIT : TXG_NOWAIT; + error = dmu_tx_assign(tx, txg_how); + if (error) { + if (error == ERESTART) { + ASSERT(txg_how == TXG_NOWAIT); + dmu_tx_wait(tx); + } else { ztest_record_enospc("dmu write parallel"); - return; } - txg = dmu_tx_get_txg(tx); + dmu_tx_abort(tx); + return; + } + txg = dmu_tx_get_txg(tx); - if (do_free) { - (void) mutex_lock(lp); - VERIFY(0 == dmu_free_range(os, ZTEST_DIROBJ, off, - bs, tx)); - (void) mutex_unlock(lp); - dmu_tx_commit(tx); - continue; + lp = &ztest_shared->zs_sync_lock[b]; + (void) mutex_lock(lp); + + wbt->bt_objset = dmu_objset_id(os); + wbt->bt_object = ZTEST_DIROBJ; + wbt->bt_offset = off; + wbt->bt_txg = txg; + wbt->bt_thread = za->za_instance; + wbt->bt_seq = ztest_shared->zs_seq[b]++; /* protected by lp */ + + /* + * Occasionally, write an all-zero block to test the behavior + * of blocks that compress into holes. + */ + if (off != -1ULL && ztest_random(8) == 0) + bzero(wbt, btsize); + + if (off == -1ULL) { + dmu_object_info_t *doi = &za->za_doi; + char *dboff; + + VERIFY(dmu_bonus_hold(os, ZTEST_DIROBJ, FTAG, &db) == 0); + za->za_dbuf = db; + dmu_object_info_from_db(db, doi); + ASSERT3U(doi->doi_bonus_size, <=, db->db_size); + ASSERT3U(doi->doi_bonus_size, >=, btsize); + ASSERT3U(doi->doi_bonus_size % btsize, ==, 0); + dboff = (char *)db->db_data + doi->doi_bonus_size - btsize; + bcopy(dboff, rbt, btsize); + if (rbt->bt_objset != 0) { + ASSERT3U(rbt->bt_objset, ==, wbt->bt_objset); + ASSERT3U(rbt->bt_object, ==, wbt->bt_object); + ASSERT3U(rbt->bt_offset, ==, wbt->bt_offset); + ASSERT3U(rbt->bt_txg, <=, wbt->bt_txg); } - - wbt.bt_objset = dmu_objset_id(os); - wbt.bt_object = ZTEST_DIROBJ; - wbt.bt_offset = off; - wbt.bt_txg = txg; - wbt.bt_thread = za->za_instance; - - if (off == -1ULL) { - wbt.bt_seq = 0; - VERIFY(0 == dmu_bonus_hold(os, ZTEST_DIROBJ, - FTAG, &db)); - ASSERT3U(db->db_size, ==, sizeof (wbt)); - bcopy(db->db_data, &rbt, db->db_size); - if (rbt.bt_objset != 0) { - ASSERT3U(rbt.bt_objset, ==, wbt.bt_objset); - ASSERT3U(rbt.bt_object, ==, wbt.bt_object); - ASSERT3U(rbt.bt_offset, ==, wbt.bt_offset); - ASSERT3U(rbt.bt_txg, <=, wbt.bt_txg); - } - dmu_buf_will_dirty(db, tx); - bcopy(&wbt, db->db_data, db->db_size); - dmu_buf_rele(db, FTAG); - dmu_tx_commit(tx); - continue; + if (ztest_random(10) == 0) { + int newsize = (ztest_random(db->db_size / + btsize) + 1) * btsize; + + ASSERT3U(newsize, >=, btsize); + ASSERT3U(newsize, <=, db->db_size); + VERIFY3U(dmu_set_bonus(db, newsize, tx), ==, 0); + dboff = (char *)db->db_data + newsize - btsize; } + dmu_buf_will_dirty(db, tx); + bcopy(wbt, dboff, btsize); + dmu_buf_rele(db, FTAG); + za->za_dbuf = NULL; + } else if (do_free) { + VERIFY(dmu_free_range(os, ZTEST_DIROBJ, off, bs, tx) == 0); + } else { + dmu_write(os, ZTEST_DIROBJ, off, btsize, wbt, tx); + } - (void) mutex_lock(lp); + (void) mutex_unlock(lp); - wbt.bt_seq = ztest_shared->zs_seq[b]++; + if (ztest_random(1000) == 0) + (void) poll(NULL, 0, 1); /* open dn_notxholds window */ - dmu_write(os, ZTEST_DIROBJ, off, sizeof (wbt), &wbt, tx); + dmu_tx_commit(tx); + + if (ztest_random(10000) == 0) + txg_wait_synced(dmu_objset_pool(os), txg); + + if (off == -1ULL || do_free) + return; + if (ztest_random(2) != 0) + return; + + /* + * dmu_sync() the block we just wrote. + */ + (void) mutex_lock(lp); + + blkoff = P2ALIGN_TYPED(off, bs, uint64_t); + error = dmu_buf_hold(os, ZTEST_DIROBJ, blkoff, FTAG, &db); + za->za_dbuf = db; + if (error) { + dprintf("dmu_buf_hold(%s, %d, %llx) = %d\n", + osname, ZTEST_DIROBJ, blkoff, error); (void) mutex_unlock(lp); + return; + } + blkoff = off - blkoff; + error = dmu_sync(NULL, db, &blk, txg, NULL, NULL); + dmu_buf_rele(db, FTAG); + za->za_dbuf = NULL; - if (ztest_random(100) == 0) - (void) poll(NULL, 0, 1); /* open dn_notxholds window */ + (void) mutex_unlock(lp); - dmu_tx_commit(tx); + if (error) { + dprintf("dmu_sync(%s, %d, %llx) = %d\n", + osname, ZTEST_DIROBJ, off, error); + return; + } - if (ztest_random(1000) == 0) - txg_wait_synced(dmu_objset_pool(os), txg); - - if (ztest_random(2) == 0) { - blkptr_t blk = { 0 }; - uint64_t blkoff; - zbookmark_t zb; - - (void) mutex_lock(lp); - blkoff = P2ALIGN_TYPED(off, bs, uint64_t); - error = dmu_buf_hold(os, - ZTEST_DIROBJ, blkoff, FTAG, &db); - if (error) { - dprintf("dmu_buf_hold(%s, %d, %llx) = %d\n", - osname, ZTEST_DIROBJ, blkoff, error); - (void) mutex_unlock(lp); - continue; - } - blkoff = off - blkoff; - error = dmu_sync(NULL, db, &blk, txg, NULL, NULL); - dmu_buf_rele(db, FTAG); - (void) mutex_unlock(lp); - if (error) { - dprintf("dmu_sync(%s, %d, %llx) = %d\n", - osname, ZTEST_DIROBJ, off, error); - continue; - } + if (blk.blk_birth == 0) /* concurrent free */ + return; - if (blk.blk_birth == 0) { /* concurrent free */ - continue; - } - txg_suspend(dmu_objset_pool(os)); + txg_suspend(dmu_objset_pool(os)); - ASSERT(blk.blk_fill == 1); - ASSERT3U(BP_GET_TYPE(&blk), ==, DMU_OT_UINT64_OTHER); - ASSERT3U(BP_GET_LEVEL(&blk), ==, 0); - ASSERT3U(BP_GET_LSIZE(&blk), ==, bs); + ASSERT(blk.blk_fill == 1); + ASSERT3U(BP_GET_TYPE(&blk), ==, DMU_OT_UINT64_OTHER); + ASSERT3U(BP_GET_LEVEL(&blk), ==, 0); + ASSERT3U(BP_GET_LSIZE(&blk), ==, bs); - /* - * Read the block that dmu_sync() returned to - * make sure its contents match what we wrote. - * We do this while still txg_suspend()ed to ensure - * that the block can't be reused before we read it. - */ - zb.zb_objset = dmu_objset_id(os); - zb.zb_object = ZTEST_DIROBJ; - zb.zb_level = 0; - zb.zb_blkid = off / bs; - error = zio_wait(zio_read(NULL, dmu_objset_spa(os), - &blk, iobuf, bs, NULL, NULL, - ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_MUSTSUCCEED, &zb)); - ASSERT(error == 0); + /* + * Read the block that dmu_sync() returned to make sure its contents + * match what we wrote. We do this while still txg_suspend()ed + * to ensure that the block can't be reused before we read it. + */ + zb.zb_objset = dmu_objset_id(os); + zb.zb_object = ZTEST_DIROBJ; + zb.zb_level = 0; + zb.zb_blkid = off / bs; + error = zio_wait(zio_read(NULL, za->za_spa, &blk, iobuf, bs, + NULL, NULL, ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_MUSTSUCCEED, &zb)); + ASSERT3U(error, ==, 0); - txg_resume(dmu_objset_pool(os)); + txg_resume(dmu_objset_pool(os)); - bcopy(&iobuf[blkoff], &rbt, sizeof (rbt)); + bcopy(&iobuf[blkoff], rbt, btsize); - if (rbt.bt_objset == 0) /* concurrent free */ - continue; + if (rbt->bt_objset == 0) /* concurrent free */ + return; - ASSERT3U(rbt.bt_objset, ==, wbt.bt_objset); - ASSERT3U(rbt.bt_object, ==, wbt.bt_object); - ASSERT3U(rbt.bt_offset, ==, wbt.bt_offset); + if (wbt->bt_objset == 0) /* all-zero overwrite */ + return; - /* - * The semantic of dmu_sync() is that we always - * push the most recent version of the data, - * so in the face of concurrent updates we may - * see a newer version of the block. That's OK. - */ - ASSERT3U(rbt.bt_txg, >=, wbt.bt_txg); - if (rbt.bt_thread == wbt.bt_thread) - ASSERT3U(rbt.bt_seq, ==, wbt.bt_seq); - else - ASSERT3U(rbt.bt_seq, >, wbt.bt_seq); - } - } + ASSERT3U(rbt->bt_objset, ==, wbt->bt_objset); + ASSERT3U(rbt->bt_object, ==, wbt->bt_object); + ASSERT3U(rbt->bt_offset, ==, wbt->bt_offset); + + /* + * The semantic of dmu_sync() is that we always push the most recent + * version of the data, so in the face of concurrent updates we may + * see a newer version of the block. That's OK. + */ + ASSERT3U(rbt->bt_txg, >=, wbt->bt_txg); + if (rbt->bt_thread == wbt->bt_thread) + ASSERT3U(rbt->bt_seq, ==, wbt->bt_seq); + else + ASSERT3U(rbt->bt_seq, >, wbt->bt_seq); } /* @@ -2166,7 +2296,6 @@ ztest_zap(ztest_args_t *za) uint64_t value[ZTEST_ZAP_MAX_INTS]; uint64_t zl_ints, zl_intsize, prop; int i, ints; - int iters = 100; dmu_tx_t *tx; char propname[100], txgname[100]; int error; @@ -2230,122 +2359,113 @@ ztest_zap(ztest_args_t *za) ints = MAX(ZTEST_ZAP_MIN_INTS, object % ZTEST_ZAP_MAX_INTS); - while (--iters >= 0) { - prop = ztest_random(ZTEST_ZAP_MAX_PROPS); - (void) sprintf(propname, "prop_%llu", (u_longlong_t)prop); - (void) sprintf(txgname, "txg_%llu", (u_longlong_t)prop); - bzero(value, 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); - - error = zap_lookup(os, object, txgname, zl_intsize, - zl_ints, &last_txg); + prop = ztest_random(ZTEST_ZAP_MAX_PROPS); + (void) sprintf(propname, "prop_%llu", (u_longlong_t)prop); + (void) sprintf(txgname, "txg_%llu", (u_longlong_t)prop); + bzero(value, sizeof (value)); + last_txg = 0; - ASSERT3U(error, ==, 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); - error = zap_length(os, object, propname, &zl_intsize, - &zl_ints); + VERIFY(zap_lookup(os, object, txgname, zl_intsize, + zl_ints, &last_txg) == 0); - ASSERT3U(error, ==, 0); - ASSERT3U(zl_intsize, ==, sizeof (uint64_t)); - ASSERT3U(zl_ints, ==, ints); + VERIFY(zap_length(os, object, propname, &zl_intsize, + &zl_ints) == 0); - error = zap_lookup(os, object, propname, zl_intsize, - zl_ints, value); + ASSERT3U(zl_intsize, ==, sizeof (uint64_t)); + ASSERT3U(zl_ints, ==, ints); - ASSERT3U(error, ==, 0); + VERIFY(zap_lookup(os, object, propname, zl_intsize, + zl_ints, value) == 0); - for (i = 0; i < ints; i++) { - ASSERT3U(value[i], ==, last_txg + object + i); - } - } else { - ASSERT3U(error, ==, ENOENT); + 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, TRUE, NULL); - error = dmu_tx_assign(tx, TXG_WAIT); - if (error) { - ztest_record_enospc("create zap entry"); - dmu_tx_abort(tx); - return; - } - txg = dmu_tx_get_txg(tx); + /* + * 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, TRUE, NULL); + error = dmu_tx_assign(tx, TXG_WAIT); + if (error) { + ztest_record_enospc("create zap entry"); + dmu_tx_abort(tx); + return; + } + txg = dmu_tx_get_txg(tx); - if (last_txg > txg) - fatal(0, "zap future leak: old %llu new %llu", - last_txg, txg); + if (last_txg > txg) + fatal(0, "zap future leak: old %llu new %llu", last_txg, txg); - for (i = 0; i < ints; i++) - value[i] = txg + object + i; + for (i = 0; i < ints; i++) + value[i] = txg + object + i; - error = zap_update(os, object, txgname, sizeof (uint64_t), - 1, &txg, tx); - if (error) - fatal(0, "zap_update('%s', %llu, '%s') = %d", - osname, object, txgname, error); + error = zap_update(os, object, txgname, sizeof (uint64_t), 1, &txg, tx); + if (error) + fatal(0, "zap_update('%s', %llu, '%s') = %d", + osname, object, txgname, error); - error = zap_update(os, object, propname, sizeof (uint64_t), - ints, value, tx); - if (error) - fatal(0, "zap_update('%s', %llu, '%s') = %d", - osname, object, propname, error); + error = zap_update(os, object, propname, sizeof (uint64_t), + ints, value, tx); + if (error) + fatal(0, "zap_update('%s', %llu, '%s') = %d", + osname, object, propname, error); - dmu_tx_commit(tx); + dmu_tx_commit(tx); - /* - * Remove a random pair of entries. - */ - prop = ztest_random(ZTEST_ZAP_MAX_PROPS); - (void) sprintf(propname, "prop_%llu", (u_longlong_t)prop); - (void) sprintf(txgname, "txg_%llu", (u_longlong_t)prop); + /* + * Remove a random pair of entries. + */ + prop = ztest_random(ZTEST_ZAP_MAX_PROPS); + (void) sprintf(propname, "prop_%llu", (u_longlong_t)prop); + (void) sprintf(txgname, "txg_%llu", (u_longlong_t)prop); - error = zap_length(os, object, txgname, &zl_intsize, &zl_ints); + error = zap_length(os, object, txgname, &zl_intsize, &zl_ints); - if (error == ENOENT) - continue; + if (error == ENOENT) + return; - ASSERT3U(error, ==, 0); + ASSERT3U(error, ==, 0); - tx = dmu_tx_create(os); - dmu_tx_hold_zap(tx, object, TRUE, NULL); - error = dmu_tx_assign(tx, TXG_WAIT); - if (error) { - ztest_record_enospc("remove zap entry"); - dmu_tx_abort(tx); - return; - } - error = zap_remove(os, object, txgname, tx); - if (error) - fatal(0, "zap_remove('%s', %llu, '%s') = %d", - osname, object, txgname, error); + tx = dmu_tx_create(os); + dmu_tx_hold_zap(tx, object, TRUE, NULL); + error = dmu_tx_assign(tx, TXG_WAIT); + if (error) { + ztest_record_enospc("remove zap entry"); + dmu_tx_abort(tx); + return; + } + error = zap_remove(os, object, txgname, tx); + if (error) + fatal(0, "zap_remove('%s', %llu, '%s') = %d", + osname, object, txgname, error); - error = zap_remove(os, object, propname, tx); - if (error) - fatal(0, "zap_remove('%s', %llu, '%s') = %d", - osname, object, propname, error); + error = zap_remove(os, object, propname, tx); + if (error) + fatal(0, "zap_remove('%s', %llu, '%s') = %d", + osname, object, propname, error); - dmu_tx_commit(tx); - } + dmu_tx_commit(tx); /* * Once in a while, destroy the object. */ - if (ztest_random(100) != 0) + if (ztest_random(1000) != 0) return; tx = dmu_tx_create(os); @@ -2372,111 +2492,107 @@ ztest_zap_parallel(ztest_args_t *za) { objset_t *os = za->za_os; uint64_t txg, object, count, wsize, wc, zl_wsize, zl_wc; - int iters = 100; dmu_tx_t *tx; int i, namelen, error; char name[20], string_value[20]; void *data; - while (--iters >= 0) { - /* - * 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; + /* + * 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'; + for (i = 0; i < 3; i++) + name[i] = '!' + ztest_random('~' - '!' + 1); + for (; i < namelen - 1; i++) + name[i] = '.'; + name[i] = '\0'; - if (ztest_random(2) == 0) - object = ZTEST_MICROZAP_OBJ; - else - object = ZTEST_FATZAP_OBJ; + if (ztest_random(2) == 0) + object = ZTEST_MICROZAP_OBJ; + else + object = ZTEST_FATZAP_OBJ; - if ((namelen & 1) || object == ZTEST_MICROZAP_OBJ) { - wsize = sizeof (txg); - wc = 1; - data = &txg; - } else { - wsize = 1; - wc = namelen; - data = string_value; - } + if ((namelen & 1) || object == ZTEST_MICROZAP_OBJ) { + wsize = sizeof (txg); + wc = 1; + data = &txg; + } else { + wsize = 1; + wc = namelen; + data = string_value; + } - count = -1ULL; - VERIFY(zap_count(os, object, &count) == 0); - ASSERT(count != -1ULL); + count = -1ULL; + VERIFY(zap_count(os, object, &count) == 0); + ASSERT(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, TRUE, NULL); - error = dmu_tx_assign(tx, TXG_WAIT); - if (error) { - ztest_record_enospc("zap parallel"); - dmu_tx_abort(tx); - return; - } - txg = dmu_tx_get_txg(tx); - bcopy(name, string_value, namelen); - } else { - tx = NULL; - txg = 0; - bzero(string_value, namelen); - } + /* + * Select an operation: length, lookup, add, update, remove. + */ + i = ztest_random(5); - switch (i) { + if (i >= 2) { + tx = dmu_tx_create(os); + dmu_tx_hold_zap(tx, object, TRUE, NULL); + error = dmu_tx_assign(tx, TXG_WAIT); + if (error) { + ztest_record_enospc("zap parallel"); + dmu_tx_abort(tx); + return; + } + txg = dmu_tx_get_txg(tx); + bcopy(name, string_value, namelen); + } else { + tx = NULL; + txg = 0; + bzero(string_value, namelen); + } - 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; + switch (i) { - case 1: - error = zap_lookup(os, object, name, wsize, wc, data); - if (error == 0) { - if (data == string_value && - bcmp(name, data, namelen) != 0) - fatal(0, "name '%s' != val '%s' len %d", - name, data, namelen); - } else { - ASSERT3U(error, ==, ENOENT); - } - break; + 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 2: - error = zap_add(os, object, name, wsize, wc, data, tx); - ASSERT(error == 0 || error == EEXIST); - break; + case 1: + error = zap_lookup(os, object, name, wsize, wc, data); + if (error == 0) { + if (data == string_value && + bcmp(name, data, namelen) != 0) + fatal(0, "name '%s' != val '%s' len %d", + name, data, namelen); + } else { + ASSERT3U(error, ==, ENOENT); + } + break; - case 3: - VERIFY(zap_update(os, object, name, wsize, wc, - data, tx) == 0); - break; + case 2: + error = zap_add(os, object, name, wsize, wc, data, tx); + ASSERT(error == 0 || error == EEXIST); + break; - case 4: - error = zap_remove(os, object, name, tx); - ASSERT(error == 0 || error == ENOENT); - break; - } + case 3: + VERIFY(zap_update(os, object, name, wsize, wc, data, tx) == 0); + break; - if (tx != NULL) - dmu_tx_commit(tx); + case 4: + error = zap_remove(os, object, name, tx); + ASSERT(error == 0 || error == ENOENT); + break; } + + if (tx != NULL) + dmu_tx_commit(tx); } void @@ -2532,21 +2648,6 @@ ztest_dsl_prop_get_set(ztest_args_t *za) (void) rw_unlock(&ztest_shared->zs_name_lock); } -static void -ztest_error_setup(vdev_t *vd, int mode, int mask, uint64_t arg) -{ - int c; - - for (c = 0; c < vd->vdev_children; c++) - ztest_error_setup(vd->vdev_child[c], mode, mask, arg); - - if (vd->vdev_path != NULL) { - vd->vdev_fault_mode = mode; - vd->vdev_fault_mask = mask; - vd->vdev_fault_arg = arg; - } -} - /* * Inject random faults into the on-disk data. */ @@ -2561,67 +2662,97 @@ ztest_fault_inject(ztest_args_t *za) char path0[MAXPATHLEN]; char pathrand[MAXPATHLEN]; size_t fsize; - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; int bshift = SPA_MAXBLOCKSHIFT + 2; /* don't scrog all labels */ int iters = 1000; - vdev_t *vd0; + int maxfaults = zopt_maxfaults; + vdev_t *vd0 = NULL; uint64_t guid0 = 0; - /* - * We can't inject faults when we have no fault tolerance. - */ - if (zopt_maxfaults == 0) - return; - - ASSERT(leaves >= 2); + ASSERT(leaves >= 1); /* - * Pick a random top-level vdev. + * We need SCL_STATE here because we're going to look at vd0->vdev_tsd. */ - spa_config_enter(spa, RW_READER, FTAG); - top = ztest_random(spa->spa_root_vdev->vdev_children); - spa_config_exit(spa, FTAG); + spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); - /* - * Pick a random leaf. - */ - leaf = ztest_random(leaves); + if (ztest_random(2) == 0) { + /* + * Inject errors on a normal data device. + */ + top = ztest_random(spa->spa_root_vdev->vdev_children); + leaf = ztest_random(leaves); - /* - * Generate paths to the first two leaves in this top-level vdev, - * and to the random leaf we selected. We'll induce transient - * I/O errors and random online/offline activity on leaf 0, - * and we'll write random garbage to the randomly chosen leaf. - */ - (void) snprintf(path0, sizeof (path0), - ztest_dev_template, zopt_dir, zopt_pool, top * leaves + 0); - (void) snprintf(pathrand, sizeof (pathrand), - ztest_dev_template, zopt_dir, zopt_pool, top * leaves + leaf); + /* + * 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, sizeof (path0), ztest_dev_template, + zopt_dir, zopt_pool, top * leaves + 0); + (void) snprintf(pathrand, sizeof (pathrand), ztest_dev_template, + zopt_dir, zopt_pool, top * leaves + leaf); - dprintf("damaging %s and %s\n", path0, pathrand); + vd0 = vdev_lookup_by_path(spa->spa_root_vdev, path0); + if (vd0 != NULL && maxfaults != 1) { + /* + * Make vd0 explicitly claim to be unreadable, + * or unwriteable, 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; - spa_config_enter(spa, RW_READER, FTAG); + if (vf != NULL && ztest_random(3) == 0) { + (void) close(vf->vf_vnode->v_fd); + vf->vf_vnode->v_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 we can tolerate two or more faults, make vd0 fail randomly. - */ - vd0 = vdev_lookup_by_path(spa->spa_root_vdev, path0); - if (vd0 != NULL && zopt_maxfaults >= 2) { + if (sav->sav_count == 0) { + spa_config_exit(spa, SCL_STATE, FTAG); + return; + } + vd0 = sav->sav_vdevs[ztest_random(sav->sav_count)]; guid0 = vd0->vdev_guid; - ztest_error_setup(vd0, VDEV_FAULT_COUNT, - (1U << ZIO_TYPE_READ) | (1U << ZIO_TYPE_WRITE), 100); + (void) strcpy(path0, vd0->vdev_path); + (void) strcpy(pathrand, vd0->vdev_path); + + leaf = 0; + leaves = 1; + maxfaults = INT_MAX; /* no limit on cache devices */ } - spa_config_exit(spa, FTAG); + dprintf("damaging %s and %s\n", path0, pathrand); + + spa_config_exit(spa, SCL_STATE, FTAG); + + if (maxfaults == 0) + return; /* * If we can tolerate two or more faults, randomly online/offline vd0. */ - if (zopt_maxfaults >= 2 && guid0 != 0) { + if (maxfaults >= 2 && guid0 != 0) { if (ztest_random(10) < 6) (void) vdev_offline(spa, guid0, B_TRUE); else - (void) vdev_online(spa, guid0); + (void) vdev_online(spa, guid0, B_FALSE, NULL); } /* @@ -2660,11 +2791,11 @@ ztest_fault_inject(ztest_args_t *za) void ztest_scrub(ztest_args_t *za) { - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; - (void) spa_scrub(spa, POOL_SCRUB_EVERYTHING, B_FALSE); + (void) spa_scrub(spa, POOL_SCRUB_EVERYTHING); (void) poll(NULL, 0, 1000); /* wait a second, then force a restart */ - (void) spa_scrub(spa, POOL_SCRUB_EVERYTHING, B_FALSE); + (void) spa_scrub(spa, POOL_SCRUB_EVERYTHING); } /* @@ -2706,7 +2837,7 @@ ztest_spa_rename(ztest_args_t *za) if (error != 0) fatal(0, "spa_open('%s') = %d", newname, error); - ASSERT(spa == dmu_objset_spa(za->za_os)); + ASSERT(spa == za->za_spa); spa_close(spa, FTAG); /* @@ -2724,7 +2855,7 @@ ztest_spa_rename(ztest_args_t *za) if (error != 0) fatal(0, "spa_open('%s') = %d", oldname, error); - ASSERT(spa == dmu_objset_spa(za->za_os)); + ASSERT(spa == za->za_spa); spa_close(spa, FTAG); umem_free(newname, strlen(newname) + 1); @@ -2778,10 +2909,9 @@ static void ztest_replace_one_disk(spa_t *spa, uint64_t vdev) { char dev_name[MAXPATHLEN]; - nvlist_t *file, *root; + nvlist_t *root; int error; uint64_t guid; - uint64_t ashift = ztest_get_ashift(); vdev_t *vd; (void) sprintf(dev_name, ztest_dev_template, zopt_dir, zopt_pool, vdev); @@ -2789,22 +2919,14 @@ ztest_replace_one_disk(spa_t *spa, uint64_t vdev) /* * Build the nvlist describing dev_name. */ - VERIFY(nvlist_alloc(&file, NV_UNIQUE_NAME, 0) == 0); - VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_TYPE, VDEV_TYPE_FILE) == 0); - VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_PATH, dev_name) == 0); - VERIFY(nvlist_add_uint64(file, ZPOOL_CONFIG_ASHIFT, ashift) == 0); - - VERIFY(nvlist_alloc(&root, NV_UNIQUE_NAME, 0) == 0); - VERIFY(nvlist_add_string(root, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT) == 0); - VERIFY(nvlist_add_nvlist_array(root, ZPOOL_CONFIG_CHILDREN, - &file, 1) == 0); + root = make_vdev_root(dev_name, NULL, 0, 0, 0, 0, 0, 1); - spa_config_enter(spa, RW_READER, FTAG); + spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); if ((vd = vdev_lookup_by_path(spa->spa_root_vdev, dev_name)) == NULL) guid = 0; else guid = vd->vdev_guid; - spa_config_exit(spa, FTAG); + spa_config_exit(spa, SCL_VDEV, FTAG); error = spa_vdev_attach(spa, guid, root, B_TRUE); if (error != 0 && error != EBUSY && @@ -2813,7 +2935,6 @@ ztest_replace_one_disk(spa_t *spa, uint64_t vdev) error != EDOM) fatal(0, "spa_vdev_attach(in-place) = %d", error); - nvlist_free(file); nvlist_free(root); } @@ -2824,6 +2945,9 @@ ztest_verify_blocks(char *pool) char zdb[MAXPATHLEN + MAXNAMELEN + 20]; char zbuf[1024]; char *bin; + char *ztest; + char *isa; + int isalen; FILE *fp; if (realpath(progname, zdb) == NULL) @@ -2831,13 +2955,19 @@ ztest_verify_blocks(char *pool) /* zdb lives in /usr/sbin, while ztest lives in /usr/bin */ bin = strstr(zdb, "/usr/bin/"); - if (bin == NULL) - bin = zdb; + ztest = strstr(bin, "/ztest"); + isa = bin + 8; + isalen = ztest - isa; + isa = strdup(isa); /* LINTED */ - (void) sprintf(bin, "/usr/sbin/zdb -bc%s%s -U -O %s %s", + (void) sprintf(bin, + "/usr/sbin%.*s/zdb -bc%s%s -U /tmp/zpool.cache -O %s %s", + isalen, + isa, zopt_verbose >= 3 ? "s" : "", zopt_verbose >= 4 ? "v" : "", ztest_random(2) == 0 ? "pre" : "post", pool); + free(isa); if (zopt_verbose >= 5) (void) printf("Executing %s\n", strstr(zdb, "zdb ")); @@ -2909,7 +3039,7 @@ ztest_spa_import_export(char *oldname, char *newname) /* * Export it. */ - error = spa_export(oldname, &config); + error = spa_export(oldname, &config, B_FALSE); if (error) fatal(0, "spa_export('%s') = %d", oldname, error); @@ -2958,35 +3088,41 @@ ztest_spa_import_export(char *oldname, char *newname) } static void * +ztest_resume(void *arg) +{ + spa_t *spa = arg; + + while (!ztest_exiting) { + (void) poll(NULL, 0, 1000); + + if (!spa_suspended(spa)) + continue; + + spa_vdev_state_enter(spa); + vdev_clear(spa, NULL); + (void) spa_vdev_state_exit(spa, NULL, 0); + + zio_resume(spa); + } + return (NULL); +} + +static void * ztest_thread(void *arg) { ztest_args_t *za = arg; ztest_shared_t *zs = ztest_shared; hrtime_t now, functime; ztest_info_t *zi; - int f; + int f, i; while ((now = gethrtime()) < za->za_stop) { /* * See if it's time to force a crash. */ if (now > za->za_kill) { - dmu_tx_t *tx; - uint64_t txg; - - mutex_enter(&spa_namespace_lock); - tx = dmu_tx_create(za->za_os); - VERIFY(0 == dmu_tx_assign(tx, TXG_NOWAIT)); - txg = dmu_tx_get_txg(tx); - dmu_tx_commit(tx); - zs->zs_txg = txg; - if (zopt_verbose >= 3) - (void) printf( - "killing process after txg %lld\n", - (u_longlong_t)txg); - txg_wait_synced(dmu_objset_pool(za->za_os), txg); - zs->zs_alloc = spa_get_alloc(dmu_objset_spa(za->za_os)); - zs->zs_space = spa_get_space(dmu_objset_spa(za->za_os)); + zs->zs_alloc = spa_get_alloc(za->za_spa); + zs->zs_space = spa_get_space(za->za_spa); (void) kill(getpid(), SIGKILL); } @@ -3011,9 +3147,8 @@ ztest_thread(void *arg) ZTEST_DIRSIZE; za->za_diroff_shared = (1ULL << 63); - ztest_dmu_write_parallel(za); - - zi->zi_func(za); + for (i = 0; i < zi->zi_iters; i++) + zi->zi_func(za); functime = gethrtime() - now; @@ -3047,6 +3182,9 @@ ztest_run(char *pool) ztest_args_t *za; spa_t *spa; char name[100]; + thread_t resume_tid; + + ztest_exiting = B_FALSE; (void) _mutex_init(&zs->zs_vdev_lock, USYNC_THREAD, NULL); (void) rwlock_init(&zs->zs_name_lock, USYNC_THREAD, NULL); @@ -3071,9 +3209,7 @@ ztest_run(char *pool) * Kick off a replacement of the disk we just obliterated. */ kernel_init(FREAD | FWRITE); - error = spa_open(pool, &spa, FTAG); - if (error) - fatal(0, "spa_open(%s) = %d", pool, error); + VERIFY(spa_open(pool, &spa, FTAG) == 0); ztest_replace_one_disk(spa, 0); if (zopt_verbose >= 5) show_pool_stats(spa); @@ -3106,9 +3242,13 @@ ztest_run(char *pool) /* * Open our pool. */ - error = spa_open(pool, &spa, FTAG); - if (error) - fatal(0, "spa_open() = %d", error); + VERIFY(spa_open(pool, &spa, FTAG) == 0); + + /* + * Create a thread to periodically resume suspended I/O. + */ + VERIFY(thr_create(0, 0, ztest_resume, spa, THR_BOUND, + &resume_tid) == 0); /* * Verify that we can safely inquire about about any object, @@ -3144,71 +3284,62 @@ ztest_run(char *pool) for (t = 0; t < zopt_threads; t++) { d = t % zopt_datasets; + + (void) strcpy(za[t].za_pool, pool); + za[t].za_os = za[d].za_os; + za[t].za_spa = spa; + za[t].za_zilog = za[d].za_zilog; + za[t].za_instance = t; + za[t].za_random = ztest_random(-1ULL); + za[t].za_start = za[0].za_start; + za[t].za_stop = za[0].za_stop; + za[t].za_kill = za[0].za_kill; + if (t < zopt_datasets) { ztest_replay_t zr; int test_future = FALSE; (void) rw_rdlock(&ztest_shared->zs_name_lock); (void) snprintf(name, 100, "%s/%s_%d", pool, pool, d); - error = dmu_objset_create(name, DMU_OST_OTHER, NULL, + error = dmu_objset_create(name, DMU_OST_OTHER, NULL, 0, ztest_create_cb, NULL); if (error == EEXIST) { test_future = TRUE; + } else if (error == ENOSPC) { + zs->zs_enospc_count++; + (void) rw_unlock(&ztest_shared->zs_name_lock); + break; } else if (error != 0) { - if (error == ENOSPC) { - zs->zs_enospc_count++; - (void) rw_unlock( - &ztest_shared->zs_name_lock); - break; - } fatal(0, "dmu_objset_create(%s) = %d", name, error); } error = dmu_objset_open(name, DMU_OST_OTHER, - DS_MODE_STANDARD, &za[d].za_os); + DS_MODE_USER, &za[d].za_os); if (error) fatal(0, "dmu_objset_open('%s') = %d", name, error); (void) rw_unlock(&ztest_shared->zs_name_lock); - if (test_future && ztest_shared->zs_txg > 0) - ztest_dmu_check_future_leak(za[d].za_os, - ztest_shared->zs_txg); + if (test_future) + ztest_dmu_check_future_leak(&za[t]); zr.zr_os = za[d].za_os; zil_replay(zr.zr_os, &zr, &zr.zr_assign, - ztest_replay_vector); + ztest_replay_vector, NULL); za[d].za_zilog = zil_open(za[d].za_os, NULL); } - za[t].za_pool = spa_strdup(pool); - za[t].za_os = za[d].za_os; - za[t].za_zilog = za[d].za_zilog; - za[t].za_instance = t; - za[t].za_random = ztest_random(-1ULL); - za[t].za_start = za[0].za_start; - za[t].za_stop = za[0].za_stop; - za[t].za_kill = za[0].za_kill; - error = thr_create(0, 0, ztest_thread, &za[t], THR_BOUND, - &za[t].za_thread); - if (error) - fatal(0, "can't create thread %d: error %d", - t, error); + VERIFY(thr_create(0, 0, ztest_thread, &za[t], THR_BOUND, + &za[t].za_thread) == 0); } - ztest_shared->zs_txg = 0; while (--t >= 0) { - error = thr_join(za[t].za_thread, NULL, NULL); - if (error) - fatal(0, "thr_join(%d) = %d", t, error); + VERIFY(thr_join(za[t].za_thread, NULL, NULL) == 0); if (za[t].za_th) traverse_fini(za[t].za_th); if (t < zopt_datasets) { zil_close(za[t].za_zilog); dmu_objset_close(za[t].za_os); } - spa_strfree(za[t].za_pool); } - umem_free(za, zopt_threads * sizeof (ztest_args_t)); - if (zopt_verbose >= 3) show_pool_stats(spa); @@ -3218,21 +3349,27 @@ ztest_run(char *pool) zs->zs_space = spa_get_space(spa); /* - * Did we have out-of-space errors? If so, destroy a random objset. + * If we had out-of-space errors, destroy a random objset. */ if (zs->zs_enospc_count != 0) { (void) rw_rdlock(&ztest_shared->zs_name_lock); - (void) snprintf(name, 100, "%s/%s_%d", pool, pool, - (int)ztest_random(zopt_datasets)); + d = (int)ztest_random(zopt_datasets); + (void) snprintf(name, 100, "%s/%s_%d", pool, pool, d); if (zopt_verbose >= 3) (void) printf("Destroying %s to free up space\n", name); - (void) dmu_objset_find(name, ztest_destroy_cb, NULL, + (void) dmu_objset_find(name, ztest_destroy_cb, &za[d], DS_FIND_SNAPSHOTS | DS_FIND_CHILDREN); (void) rw_unlock(&ztest_shared->zs_name_lock); } txg_wait_synced(spa_get_dsl(spa), 0); + umem_free(za, zopt_threads * sizeof (ztest_args_t)); + + /* Kill the resume thread */ + ztest_exiting = B_TRUE; + VERIFY(thr_join(resume_tid, NULL, NULL) == 0); + /* * Right before closing the pool, kick off a bunch of async I/O; * spa_close() should wait for it to complete. @@ -3288,8 +3425,9 @@ ztest_init(char *pool) */ (void) spa_destroy(pool); ztest_shared->zs_vdev_primaries = 0; - nvroot = make_vdev_root(zopt_vdev_size, zopt_raidz, zopt_mirrors, 1); - error = spa_create(pool, nvroot, NULL); + nvroot = make_vdev_root(NULL, NULL, zopt_vdev_size, 0, + 0, zopt_raidz, zopt_mirrors, 1); + error = spa_create(pool, nvroot, NULL, NULL, NULL); nvlist_free(nvroot); if (error) @@ -3320,7 +3458,7 @@ main(int argc, char **argv) (void) setvbuf(stdout, NULL, _IOLBF, 0); /* Override location of zpool.cache */ - spa_config_dir = "/tmp"; + spa_config_path = "/tmp/zpool.cache"; ztest_random_fd = open("/dev/urandom", O_RDONLY); |