/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2013 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2017 by Delphix. All rights reserved.
* Copyright (c) 2015, Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2013, 2014, 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 (c) 2017 Datto Inc.
+ * Copyright 2017 Joyent, Inc.
*/
/*
#include <sys/zil.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_disk.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/zvol.h>
#ifdef _KERNEL
+#include <sys/fm/protocol.h>
+#include <sys/fm/util.h>
#include <sys/bootprops.h>
#include <sys/callb.h>
#include <sys/cpupart.h>
* The interval, in seconds, at which failed configuration cache file writes
* should be retried.
*/
-static int zfs_ccw_retry_interval = 300;
+int zfs_ccw_retry_interval = 300;
typedef enum zti_modes {
ZTI_MODE_FIXED, /* value is # of threads (min 1) */
static void spa_sync_props(void *arg, dmu_tx_t *tx);
static boolean_t spa_has_active_shared_spare(spa_t *spa);
static inline int spa_load_impl(spa_t *spa, uint64_t, nvlist_t *config,
- spa_load_state_t state, spa_import_type_t type, boolean_t mosconfig,
+ spa_load_state_t state, spa_import_type_t type, boolean_t trust_config,
char **ereport);
static void spa_vdev_resilver_done(spa_t *spa);
boolean_t spa_create_process = B_TRUE; /* no process ==> no sysdc */
+/*
+ * Report any spa_load_verify errors found, but do not fail spa_load.
+ * This is used by zdb to analyze non-idle pools.
+ */
+boolean_t spa_load_verify_dryrun = B_FALSE;
+
/*
* This (illegal) pool name is used when temporarily importing a spa_t in order
* to get the vdev stats associated with the imported devices.
zprop_source_t src = ZPROP_SRC_DEFAULT;
zpool_prop_t prop;
- if ((prop = zpool_name_to_prop(za.za_name)) == ZPROP_INVAL)
+ if ((prop = zpool_name_to_prop(za.za_name)) == ZPOOL_PROP_INVAL)
continue;
switch (za.za_integer_length) {
const char *propname = nvpair_name(elem);
zpool_prop_t prop = zpool_name_to_prop(propname);
- switch ((int)prop) {
- case ZPROP_INVAL:
+ switch (prop) {
+ case ZPOOL_PROP_INVAL:
if (!zpool_prop_feature(propname)) {
error = SET_ERROR(EINVAL);
break;
error = SET_ERROR(EINVAL);
break;
+ case ZPOOL_PROP_MULTIHOST:
+ error = nvpair_value_uint64(elem, &intval);
+ if (!error && intval > 1)
+ error = SET_ERROR(EINVAL);
+
+ if (!error && !spa_get_hostid())
+ error = SET_ERROR(ENOTSUP);
+
+ break;
+
case ZPOOL_PROP_BOOTFS:
/*
* If the pool version is less than SPA_VERSION_BOOTFS,
&propval)) == 0 &&
!BOOTFS_COMPRESS_VALID(propval)) {
error = SET_ERROR(ENOTSUP);
- } else if ((error =
- dsl_prop_get_int_ds(dmu_objset_ds(os),
- zfs_prop_to_name(ZFS_PROP_RECORDSIZE),
- &propval)) == 0 &&
- propval > SPA_OLD_MAXBLOCKSIZE) {
- error = SET_ERROR(ENOTSUP);
} else if ((error =
dsl_prop_get_int_ds(dmu_objset_ds(os),
zfs_prop_to_name(ZFS_PROP_DNODESIZE),
case ZPOOL_PROP_FAILUREMODE:
error = nvpair_value_uint64(elem, &intval);
- if (!error && (intval < ZIO_FAILURE_MODE_WAIT ||
- intval > ZIO_FAILURE_MODE_PANIC))
+ if (!error && intval > ZIO_FAILURE_MODE_PANIC)
error = SET_ERROR(EINVAL);
/*
prop == ZPOOL_PROP_READONLY)
continue;
- if (prop == ZPOOL_PROP_VERSION || prop == ZPROP_INVAL) {
+ if (prop == ZPOOL_PROP_VERSION || prop == ZPOOL_PROP_INVAL) {
uint64_t ver;
if (prop == ZPOOL_PROP_VERSION) {
static int
spa_change_guid_check(void *arg, dmu_tx_t *tx)
{
+ ASSERTV(uint64_t *newguid = arg);
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
vdev_t *rvd = spa->spa_root_vdev;
uint64_t vdev_state;
- ASSERTV(uint64_t *newguid = arg);
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
vdev_state = rvd->vdev_state;
spa_change_guid_sync, &guid, 5, ZFS_SPACE_CHECK_RESERVED);
if (error == 0) {
- spa_config_sync(spa, B_FALSE, B_TRUE);
- spa_event_notify(spa, NULL, ESC_ZFS_POOL_REGUID);
+ spa_write_cachefile(spa, B_FALSE, B_TRUE);
+ spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_REGUID);
}
mutex_exit(&spa_namespace_lock);
uint_t count = ztip->zti_count;
spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q];
char name[32];
- uint_t i, flags = TASKQ_DYNAMIC;
+ uint_t flags = 0;
boolean_t batch = B_FALSE;
if (mode == ZTI_MODE_NULL) {
case ZTI_MODE_FIXED:
ASSERT3U(value, >=, 1);
value = MAX(value, 1);
+ flags |= TASKQ_DYNAMIC;
break;
case ZTI_MODE_BATCH:
break;
}
- for (i = 0; i < count; i++) {
+ for (uint_t i = 0; i < count; i++) {
taskq_t *tq;
if (count > 1) {
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];
- uint_t i;
if (tqs->stqs_taskq == NULL) {
ASSERT3U(tqs->stqs_count, ==, 0);
return;
}
- for (i = 0; i < tqs->stqs_count; i++) {
+ for (uint_t i = 0; i < tqs->stqs_count; i++) {
ASSERT3P(tqs->stqs_taskq[i], !=, NULL);
taskq_destroy(tqs->stqs_taskq[i]);
}
static void
spa_create_zio_taskqs(spa_t *spa)
{
- int t, q;
-
- for (t = 0; t < ZIO_TYPES; t++) {
- for (q = 0; q < ZIO_TASKQ_TYPES; q++) {
+ for (int t = 0; t < ZIO_TYPES; t++) {
+ for (int q = 0; q < ZIO_TASKQ_TYPES; q++) {
spa_taskqs_init(spa, t, q);
}
}
}
+/*
+ * Disabled until spa_thread() can be adapted for Linux.
+ */
+#undef HAVE_SPA_THREAD
+
#if defined(_KERNEL) && defined(HAVE_SPA_THREAD)
static void
spa_thread(void *arg)
spa_create_zio_taskqs(spa);
}
+ for (size_t i = 0; i < TXG_SIZE; i++)
+ spa->spa_txg_zio[i] = zio_root(spa, NULL, NULL, 0);
+
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),
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,
+ 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));
+ spa_keystore_init(&spa->spa_keystore);
+
/*
* This taskq is used to perform zvol-minor-related tasks
* asynchronously. This has several advantages, including easy
*/
spa->spa_zvol_taskq = taskq_create("z_zvol", 1, defclsyspri,
1, INT_MAX, 0);
+
+ /*
+ * Taskq dedicated to prefetcher threads: this is used to prevent the
+ * pool traverse code from monopolizing the global (and limited)
+ * system_taskq by inappropriately scheduling long running tasks on it.
+ */
+ spa->spa_prefetch_taskq = taskq_create("z_prefetch", boot_ncpus,
+ defclsyspri, 1, INT_MAX, TASKQ_DYNAMIC);
+
+ /*
+ * 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", boot_ncpus,
+ defclsyspri, 1, INT_MAX, TASKQ_DYNAMIC);
}
/*
static void
spa_deactivate(spa_t *spa)
{
- int t, q;
-
ASSERT(spa->spa_sync_on == B_FALSE);
ASSERT(spa->spa_dsl_pool == NULL);
ASSERT(spa->spa_root_vdev == NULL);
spa->spa_zvol_taskq = NULL;
}
+ if (spa->spa_prefetch_taskq) {
+ taskq_destroy(spa->spa_prefetch_taskq);
+ spa->spa_prefetch_taskq = NULL;
+ }
+
+ if (spa->spa_upgrade_taskq) {
+ taskq_destroy(spa->spa_upgrade_taskq);
+ spa->spa_upgrade_taskq = NULL;
+ }
+
txg_list_destroy(&spa->spa_vdev_txg_list);
list_destroy(&spa->spa_config_dirty_list);
list_destroy(&spa->spa_evicting_os_list);
list_destroy(&spa->spa_state_dirty_list);
- taskq_cancel_id(system_taskq, spa->spa_deadman_tqid);
+ taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid);
- for (t = 0; t < ZIO_TYPES; t++) {
- for (q = 0; q < ZIO_TASKQ_TYPES; q++) {
+ 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;
* still have errors left in the queues. Empty them just in case.
*/
spa_errlog_drain(spa);
-
avl_destroy(&spa->spa_errlist_scrub);
avl_destroy(&spa->spa_errlist_last);
+ spa_keystore_fini(&spa->spa_keystore);
+
spa->spa_state = POOL_STATE_UNINITIALIZED;
mutex_enter(&spa->spa_proc_lock);
nvlist_t **child;
uint_t children;
int error;
- int c;
if ((error = vdev_alloc(spa, vdp, nv, parent, id, atype)) != 0)
return (error);
return (SET_ERROR(EINVAL));
}
- for (c = 0; c < children; c++) {
+ for (int c = 0; c < children; c++) {
vdev_t *vd;
if ((error = spa_config_parse(spa, &vd, child[c], *vdp, c,
atype)) != 0) {
ASSERT(MUTEX_HELD(&spa_namespace_lock));
+ spa_load_note(spa, "UNLOADING");
+
/*
* Stop async tasks.
*/
spa->spa_sync_on = B_FALSE;
}
+ /*
+ * Even though vdev_free() also calls vdev_metaslab_fini, we need
+ * to call it earlier, before we wait for async i/o to complete.
+ * This ensures that there is no async metaslab prefetching, by
+ * calling taskq_wait(mg_taskq).
+ */
+ if (spa->spa_root_vdev != NULL) {
+ spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
+ for (int c = 0; c < spa->spa_root_vdev->vdev_children; c++)
+ vdev_metaslab_fini(spa->spa_root_vdev->vdev_child[c]);
+ spa_config_exit(spa, SCL_ALL, FTAG);
+ }
+
+ 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 (i = 0; i < max_ncpus; i++)
+ for (int i = 0; i < max_ncpus; i++)
(void) zio_wait(spa->spa_async_zio_root[i]);
kmem_free(spa->spa_async_zio_root, max_ncpus * sizeof (void *));
spa->spa_async_zio_root = NULL;
}
+ if (spa->spa_vdev_removal != NULL) {
+ spa_vdev_removal_destroy(spa->spa_vdev_removal);
+ spa->spa_vdev_removal = NULL;
+ }
+
+ if (spa->spa_condense_zthr != NULL) {
+ ASSERT(!zthr_isrunning(spa->spa_condense_zthr));
+ zthr_destroy(spa->spa_condense_zthr);
+ spa->spa_condense_zthr = NULL;
+ }
+
+ spa_condense_fini(spa);
+
bpobj_close(&spa->spa_deferred_bpobj);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
ddt_unload(spa);
-
/*
* Drop and purge level 2 cache
*/
spa->spa_async_suspended = 0;
+ spa->spa_indirect_vdevs_loaded = B_FALSE;
+
if (spa->spa_comment != NULL) {
spa_strfree(spa->spa_comment);
spa->spa_comment = NULL;
* 'spa_spares.sav_config'. We parse this into vdevs, try to open them, and
* then re-generate a more complete list including status information.
*/
-static void
+void
spa_load_spares(spa_t *spa)
{
nvlist_t **spares;
* Devices which are already active have their details maintained, and are
* not re-opened.
*/
-static void
+void
spa_load_l2cache(spa_t *spa)
{
- nvlist_t **l2cache;
+ nvlist_t **l2cache = NULL;
uint_t nl2cache;
int i, j, oldnvdevs;
uint64_t guid;
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
- if (sav->sav_config != NULL) {
- VERIFY(nvlist_lookup_nvlist_array(sav->sav_config,
- ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0);
- newvdevs = kmem_alloc(nl2cache * sizeof (void *), KM_SLEEP);
- } else {
- nl2cache = 0;
- newvdevs = NULL;
- }
-
oldvdevs = sav->sav_vdevs;
oldnvdevs = sav->sav_count;
sav->sav_vdevs = NULL;
sav->sav_count = 0;
+ if (sav->sav_config == NULL) {
+ nl2cache = 0;
+ newvdevs = NULL;
+ goto out;
+ }
+
+ VERIFY(nvlist_lookup_nvlist_array(sav->sav_config,
+ ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0);
+ newvdevs = kmem_alloc(nl2cache * sizeof (void *), KM_SLEEP);
+
/*
* Process new nvlist of vdevs.
*/
}
}
+ sav->sav_vdevs = newvdevs;
+ sav->sav_count = (int)nl2cache;
+
+ /*
+ * Recompute the stashed list of l2cache devices, with status
+ * information this time.
+ */
+ VERIFY(nvlist_remove(sav->sav_config, ZPOOL_CONFIG_L2CACHE,
+ DATA_TYPE_NVLIST_ARRAY) == 0);
+
+ 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);
+ VERIFY(nvlist_add_nvlist_array(sav->sav_config,
+ ZPOOL_CONFIG_L2CACHE, l2cache, sav->sav_count) == 0);
+
+out:
/*
* Purge vdevs that were dropped
*/
if (oldvdevs)
kmem_free(oldvdevs, oldnvdevs * sizeof (void *));
- if (sav->sav_config == NULL)
- goto out;
-
- sav->sav_vdevs = newvdevs;
- sav->sav_count = (int)nl2cache;
-
- /*
- * Recompute the stashed list of l2cache devices, with status
- * information this time.
- */
- VERIFY(nvlist_remove(sav->sav_config, ZPOOL_CONFIG_L2CACHE,
- DATA_TYPE_NVLIST_ARRAY) == 0);
-
- l2cache = kmem_alloc(sav->sav_count * sizeof (void *), KM_SLEEP);
- for (i = 0; i < sav->sav_count; i++)
- l2cache[i] = vdev_config_generate(spa,
- sav->sav_vdevs[i], B_TRUE, VDEV_CONFIG_L2CACHE);
- VERIFY(nvlist_add_nvlist_array(sav->sav_config,
- ZPOOL_CONFIG_L2CACHE, l2cache, sav->sav_count) == 0);
-out:
for (i = 0; i < sav->sav_count; i++)
nvlist_free(l2cache[i]);
if (sav->sav_count)
static void
spa_check_removed(vdev_t *vd)
{
- int c;
-
- for (c = 0; c < vd->vdev_children; c++)
+ for (int 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) &&
- !vd->vdev_ishole) {
+ vdev_is_concrete(vd)) {
zfs_post_autoreplace(vd->vdev_spa, vd);
- spa_event_notify(vd->vdev_spa, vd, ESC_ZFS_VDEV_CHECK);
+ spa_event_notify(vd->vdev_spa, vd, NULL, ESC_ZFS_VDEV_CHECK);
}
}
static void
spa_config_valid_zaps(vdev_t *vd, vdev_t *mvd)
{
- uint64_t i;
-
ASSERT3U(vd->vdev_children, ==, mvd->vdev_children);
vd->vdev_top_zap = mvd->vdev_top_zap;
vd->vdev_leaf_zap = mvd->vdev_leaf_zap;
- for (i = 0; i < vd->vdev_children; i++) {
+ for (uint64_t i = 0; i < vd->vdev_children; i++) {
spa_config_valid_zaps(vd->vdev_child[i], mvd->vdev_child[i]);
}
}
{
vdev_t *mrvd, *rvd = spa->spa_root_vdev;
nvlist_t *nv;
- int c, i;
VERIFY(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nv) == 0);
nvlist_t **child, *nv;
uint64_t idx = 0;
- child = kmem_alloc(rvd->vdev_children * sizeof (nvlist_t **),
+ child = kmem_alloc(rvd->vdev_children * sizeof (nvlist_t *),
KM_SLEEP);
VERIFY(nvlist_alloc(&nv, NV_UNIQUE_NAME, KM_SLEEP) == 0);
- for (c = 0; c < rvd->vdev_children; c++) {
+ for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
vdev_t *mtvd = mrvd->vdev_child[c];
VERIFY(nvlist_add_nvlist(spa->spa_load_info,
ZPOOL_CONFIG_MISSING_DEVICES, nv) == 0);
- for (i = 0; i < idx; i++)
+ for (int i = 0; i < idx; i++)
nvlist_free(child[i]);
}
nvlist_free(nv);
* from the MOS config (mrvd). Check each top-level vdev
* with the corresponding MOS config top-level (mtvd).
*/
- for (c = 0; c < rvd->vdev_children; c++) {
+ for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
vdev_t *mtvd = mrvd->vdev_child[c];
/*
* Resolve any "missing" vdevs in the current configuration.
+ * Also trust the MOS config about any "indirect" vdevs.
* If we find that the MOS config has more accurate information
* about the top-level vdev then use that vdev instead.
*/
- if (tvd->vdev_ops == &vdev_missing_ops &&
- mtvd->vdev_ops != &vdev_missing_ops) {
-
- if (!(spa->spa_import_flags & ZFS_IMPORT_MISSING_LOG))
- continue;
+ if ((tvd->vdev_ops == &vdev_missing_ops &&
+ mtvd->vdev_ops != &vdev_missing_ops) ||
+ (mtvd->vdev_ops == &vdev_indirect_ops &&
+ tvd->vdev_ops != &vdev_indirect_ops)) {
/*
* Device specific actions.
*/
if (mtvd->vdev_islog) {
+ if (!(spa->spa_import_flags &
+ ZFS_IMPORT_MISSING_LOG)) {
+ continue;
+ }
+
spa_set_log_state(spa, SPA_LOG_CLEAR);
- } else {
- /*
- * XXX - once we have 'readonly' pool
- * support we should be able to handle
- * missing data devices by transitioning
- * the pool to readonly.
- */
+ } else if (mtvd->vdev_ops != &vdev_indirect_ops) {
continue;
}
vdev_add_child(rvd, mtvd);
vdev_add_child(mrvd, tvd);
- spa_config_exit(spa, SCL_ALL, FTAG);
- vdev_load(mtvd);
- spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
-
vdev_reopen(rvd);
} else {
if (mtvd->vdev_islog) {
*/
spa_config_valid_zaps(tvd, mtvd);
}
+
+ /*
+ * Never trust this info from userland; always use what's
+ * in the MOS. This prevents it from getting out of sync
+ * with the rest of the info in the MOS.
+ */
+ tvd->vdev_removing = mtvd->vdev_removing;
+ tvd->vdev_indirect_config = mtvd->vdev_indirect_config;
}
vdev_free(mrvd);
{
vdev_t *rvd = spa->spa_root_vdev;
boolean_t slog_found = B_FALSE;
- int c;
ASSERT(spa_config_held(spa, SCL_ALLOC, RW_WRITER));
if (!spa_has_slogs(spa))
return (B_FALSE);
- for (c = 0; c < rvd->vdev_children; c++) {
+ for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
metaslab_group_t *mg = tvd->vdev_mg;
spa_activate_log(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
- int c;
ASSERT(spa_config_held(spa, SCL_ALLOC, RW_WRITER));
- for (c = 0; c < rvd->vdev_children; c++) {
+ for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
metaslab_group_t *mg = tvd->vdev_mg;
}
int
-spa_offline_log(spa_t *spa)
+spa_reset_logs(spa_t *spa)
{
int error;
- error = dmu_objset_find(spa_name(spa), zil_vdev_offline,
+ error = dmu_objset_find(spa_name(spa), zil_reset,
NULL, DS_FIND_CHILDREN);
if (error == 0) {
/*
static void
spa_aux_check_removed(spa_aux_vdev_t *sav)
{
- int i;
-
- for (i = 0; i < sav->sav_count; i++)
+ for (int i = 0; i < sav->sav_count; i++)
spa_check_removed(sav->sav_vdevs[i]);
}
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)
else
atomic_inc_64(&sle->sle_data_count);
}
- zio_data_buf_free(zio->io_data, zio->io_size);
mutex_enter(&spa->spa_scrub_lock);
- spa->spa_scrub_inflight--;
+ spa->spa_load_verify_ios--;
cv_broadcast(&spa->spa_scrub_io_cv);
mutex_exit(&spa->spa_scrub_lock);
}
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;
- size_t size;
- void *data;
-
if (bp == NULL || BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp))
return (0);
/*
*/
if (!spa_load_verify_metadata)
return (0);
- if (BP_GET_BUFC_TYPE(bp) == ARC_BUFC_DATA && !spa_load_verify_data)
+ if (!BP_IS_METADATA(bp) && !spa_load_verify_data)
return (0);
- rio = arg;
- size = BP_GET_PSIZE(bp);
- data = zio_data_buf_alloc(size);
+ zio_t *rio = arg;
+ size_t size = BP_GET_PSIZE(bp);
mutex_enter(&spa->spa_scrub_lock);
- while (spa->spa_scrub_inflight >= spa_load_verify_maxinflight)
+ while (spa->spa_load_verify_ios >= spa_load_verify_maxinflight)
cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock);
- spa->spa_scrub_inflight++;
+ spa->spa_load_verify_ios++;
mutex_exit(&spa->spa_scrub_lock);
- zio_nowait(zio_read(rio, spa, bp, data, size,
+ 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));
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE);
if (spa_load_verify_metadata) {
+ if (spa->spa_extreme_rewind) {
+ spa_load_note(spa, "performing a complete scan of the "
+ "pool since extreme rewind is on. This may take "
+ "a very long time.\n (spa_load_verify_data=%u, "
+ "spa_load_verify_metadata=%u)",
+ spa_load_verify_data, spa_load_verify_metadata);
+ }
error = traverse_pool(spa, spa->spa_verify_min_txg,
- TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA,
- spa_load_verify_cb, rio);
+ TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA |
+ TRAVERSE_NO_DECRYPT, spa_load_verify_cb, rio);
}
(void) zio_wait(rio);
spa->spa_load_meta_errors = sle.sle_meta_count;
spa->spa_load_data_errors = sle.sle_data_count;
- if (!error && sle.sle_meta_count <= policy.zrp_maxmeta &&
- sle.sle_data_count <= policy.zrp_maxdata) {
+ 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.zrp_maxmeta &&
+ sle.sle_data_count <= policy.zrp_maxdata)) {
int64_t loss = 0;
verify_ok = B_TRUE;
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);
* Find a value in the pool directory object.
*/
static int
-spa_dir_prop(spa_t *spa, const char *name, uint64_t *val)
+spa_dir_prop(spa_t *spa, const char *name, uint64_t *val, boolean_t log_enoent)
{
- return (zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
- name, sizeof (uint64_t), 1, val));
+ 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 (err);
+ return (SET_ERROR(err));
+}
+
+static void
+spa_spawn_aux_threads(spa_t *spa)
+{
+ ASSERT(spa_writeable(spa));
+
+ ASSERT(MUTEX_HELD(&spa_namespace_lock));
+
+ spa_start_indirect_condensing_thread(spa);
}
/*
static int
spa_load(spa_t *spa, spa_load_state_t state, spa_import_type_t type,
- boolean_t mosconfig)
+ boolean_t trust_config)
{
nvlist_t *config = spa->spa_config;
char *ereport = FM_EREPORT_ZFS_POOL;
gethrestime(&spa->spa_loaded_ts);
error = spa_load_impl(spa, pool_guid, config, state, type,
- mosconfig, &ereport);
+ trust_config, &ereport);
}
/*
spa->spa_loaded_ts.tv_nsec = 0;
}
if (error != EBADF) {
- zfs_ereport_post(ereport, spa, NULL, NULL, 0, 0);
+ zfs_ereport_post(ereport, spa, NULL, NULL, NULL, 0, 0);
}
}
spa->spa_load_state = error ? SPA_LOAD_ERROR : SPA_LOAD_NONE;
{
spa_t *spa = vd->vdev_spa;
uint64_t total = 0;
- uint64_t i;
if (vd->vdev_top_zap != 0) {
total++;
spa->spa_all_vdev_zaps, vd->vdev_leaf_zap));
}
- for (i = 0; i < vd->vdev_children; i++) {
+ for (uint64_t i = 0; i < vd->vdev_children; i++) {
total += vdev_count_verify_zaps(vd->vdev_child[i]);
}
#endif
/*
- * Load an existing storage pool, using the pool's builtin spa_config as a
- * source of configuration information.
+ * Determine whether the activity check is required.
*/
-__attribute__((always_inline))
-static inline int
-spa_load_impl(spa_t *spa, uint64_t pool_guid, nvlist_t *config,
- spa_load_state_t state, spa_import_type_t type, boolean_t mosconfig,
- char **ereport)
+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;
+ 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_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 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)
+ return (B_FALSE);
+
+ /*
+ * Allow the activity check to be skipped when importing the pool
+ * on the same host which last imported it. Since the hostid from
+ * configuration may be stale use the one read from the label.
+ */
+ if (nvlist_exists(label, ZPOOL_CONFIG_HOSTID))
+ hostid = fnvlist_lookup_uint64(label, ZPOOL_CONFIG_HOSTID);
+
+ if (hostid == spa_get_hostid())
+ return (B_FALSE);
+
+ /*
+ * Skip the activity test when the pool was cleanly exported.
+ */
+ if (state != POOL_STATE_ACTIVE)
+ return (B_FALSE);
+
+ return (B_TRUE);
+}
+
+/*
+ * 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 import_intervals = MAX(zfs_multihost_import_intervals, 1);
+ uint64_t txg = ub->ub_txg;
+ uint64_t timestamp = ub->ub_timestamp;
+ uint64_t import_delay = NANOSEC;
+ 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;
- nvlist_t *nvroot = NULL;
- nvlist_t *label;
- vdev_t *rvd;
- uberblock_t *ub = &spa->spa_uberblock;
- uint64_t children, config_cache_txg = spa->spa_config_txg;
- int orig_mode = spa->spa_mode;
- int parse, i;
- uint64_t obj;
- boolean_t missing_feat_write = B_FALSE;
- nvlist_t *mos_config;
+
+ cv_init(&cv, NULL, CV_DEFAULT, NULL);
+ mutex_init(&mtx, NULL, MUTEX_DEFAULT, NULL);
+ mutex_enter(&mtx);
/*
- * If this is an untrusted config, access the pool in read-only mode.
- * This prevents things like resilvering recently removed devices.
+ * 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 node. Check for
+ * changes in the uberblocks on disk if necessary.
*/
- if (!mosconfig)
- spa->spa_mode = FREAD;
+ if (nvlist_exists(config, ZPOOL_CONFIG_LOAD_INFO)) {
+ nvlist_t *nvinfo = fnvlist_lookup_nvlist(config,
+ ZPOOL_CONFIG_LOAD_INFO);
- ASSERT(MUTEX_HELD(&spa_namespace_lock));
+ 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;
+ }
+ }
- spa->spa_load_state = state;
+ /*
+ * Preferentially use the zfs_multihost_interval from the node which
+ * last imported the pool. This value is stored in an MMP uberblock as.
+ *
+ * ub_mmp_delay * vdev_count_leaves() == zfs_multihost_interval
+ */
+ if (ub->ub_mmp_magic == MMP_MAGIC && ub->ub_mmp_delay)
+ import_delay = MAX(import_delay, import_intervals *
+ ub->ub_mmp_delay * MAX(vdev_count_leaves(spa), 1));
+
+ /* Apply a floor using the local default values. */
+ import_delay = MAX(import_delay, import_intervals *
+ MSEC2NSEC(MAX(zfs_multihost_interval, MMP_MIN_INTERVAL)));
+
+ zfs_dbgmsg("import_delay=%llu ub_mmp_delay=%llu import_intervals=%u "
+ "leaves=%u", import_delay, ub->ub_mmp_delay, import_intervals,
+ vdev_count_leaves(spa));
+
+ /* Add a small random factor in case of simultaneous imports (0-25%) */
+ import_expire = gethrtime() + import_delay +
+ (import_delay * spa_get_random(250) / 1000);
+
+ while (gethrtime() < import_expire) {
+ vdev_uberblock_load(rvd, ub, &mmp_label);
+
+ if (txg != ub->ub_txg || timestamp != ub->ub_timestamp) {
+ error = SET_ERROR(EREMOTEIO);
+ break;
+ }
+
+ if (mmp_label) {
+ nvlist_free(mmp_label);
+ mmp_label = NULL;
+ }
+
+ error = cv_timedwait_sig(&cv, &mtx, ddi_get_lbolt() + hz);
+ if (error != -1) {
+ error = SET_ERROR(EINTR);
+ break;
+ }
+ error = 0;
+ }
+
+out:
+ mutex_exit(&mtx);
+ mutex_destroy(&mtx);
+ cv_destroy(&cv);
+
+ /*
+ * If the pool is determined to be active store the status in the
+ * spa->spa_load_info nvlist. If the remote hostname or hostid are
+ * available from configuration read from disk store them as well.
+ * This allows 'zpool import' to generate a more useful message.
+ *
+ * ZPOOL_CONFIG_MMP_STATE - observed pool status (mandatory)
+ * ZPOOL_CONFIG_MMP_HOSTNAME - hostname from the active pool
+ * ZPOOL_CONFIG_MMP_HOSTID - hostid from the active pool
+ */
+ if (error == EREMOTEIO) {
+ char *hostname = "<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);
+}
- if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot))
+static int
+spa_ld_parse_config(spa_t *spa, uint64_t pool_guid, nvlist_t *config,
+ spa_import_type_t type)
+{
+ int error = 0;
+ nvlist_t *nvtree = NULL;
+ int parse;
+ vdev_t *rvd;
+
+ 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));
+ }
parse = (type == SPA_IMPORT_EXISTING ?
VDEV_ALLOC_LOAD : VDEV_ALLOC_SPLIT);
*/
spa->spa_async_zio_root = kmem_alloc(max_ncpus * sizeof (void *),
KM_SLEEP);
- for (i = 0; i < max_ncpus; i++) {
+ 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);
* configuration requires knowing the version number.
*/
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
- error = spa_config_parse(spa, &rvd, nvroot, NULL, 0, parse);
+ error = spa_config_parse(spa, &rvd, nvtree, NULL, 0, parse);
spa_config_exit(spa, SCL_ALL, FTAG);
- if (error != 0)
+ 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);
ASSERT(spa_guid(spa) == pool_guid);
}
- /*
- * Try to open all vdevs, loading each label in the process.
- */
+ return (0);
+}
+
+static int
+spa_ld_open_vdevs(spa_t *spa)
+{
+ int error = 0;
+
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
- error = vdev_open(rvd);
+ error = vdev_open(spa->spa_root_vdev);
spa_config_exit(spa, SCL_ALL, FTAG);
- if (error != 0)
- return (error);
+ if (error != 0) {
+ spa_load_failed(spa, "unable to open vdev tree [error=%d]",
+ error);
+ }
- /*
- * We need to validate the vdev labels against the configuration that
- * we have in hand, which is dependent on the setting of mosconfig. If
- * mosconfig is true then we're validating the vdev labels based on
- * that config. Otherwise, we're validating against the cached config
- * (zpool.cache) that was read when we loaded the zfs module, and then
- * later we will recursively call spa_load() and validate against
+ return (error);
+}
+
+static int
+spa_ld_validate_vdevs(spa_t *spa, spa_import_type_t type,
+ boolean_t trust_config)
+{
+ int error = 0;
+ vdev_t *rvd = spa->spa_root_vdev;
+
+ /*
+ * We need to validate the vdev labels against the configuration that
+ * we have in hand, which is dependent on the setting of trust_config.
+ * If trust_config is true then we're validating the vdev labels based
+ * on that config. Otherwise, we're validating against the cached
+ * config (zpool.cache) that was read when we loaded the zfs module, and
+ * then later we will recursively call spa_load() and validate against
* the vdev config.
*
* If we're assembling a new pool that's been split off from an
*/
if (type != SPA_IMPORT_ASSEMBLE) {
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
- error = vdev_validate(rvd, mosconfig);
+ error = vdev_validate(rvd, trust_config);
spa_config_exit(spa, SCL_ALL, FTAG);
- if (error != 0)
+ if (error != 0) {
+ spa_load_failed(spa, "vdev_validate failed [error=%d]",
+ error);
return (error);
+ }
- if (rvd->vdev_state <= VDEV_STATE_CANT_OPEN)
+ if (rvd->vdev_state <= VDEV_STATE_CANT_OPEN) {
+ spa_load_failed(spa, "cannot open vdev tree after "
+ "invalidating some vdevs");
return (SET_ERROR(ENXIO));
+ }
}
+ return (0);
+}
+
+static int
+spa_ld_select_uberblock(spa_t *spa, nvlist_t *config, spa_import_type_t type,
+ boolean_t trust_config)
+{
+ vdev_t *rvd = spa->spa_root_vdev;
+ nvlist_t *label;
+ uberblock_t *ub = &spa->spa_uberblock;
+ uint64_t children;
+ boolean_t activity_check = B_FALSE;
+
/*
* Find the best uberblock.
*/
*/
if (ub->ub_txg == 0) {
nvlist_free(label);
+ spa_load_failed(spa, "no valid uberblock found");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, ENXIO));
}
+ spa_load_note(spa, "using uberblock with txg=%llu",
+ (u_longlong_t)ub->ub_txg);
+
+
+ /*
+ * For pools which have the multihost property on determine if the
+ * pool is truly inactive and can be safely imported. Prevent
+ * hosts which don't have a hostid set from importing the pool.
+ */
+ activity_check = spa_activity_check_required(spa, ub, label, config);
+ if (activity_check) {
+ if (ub->ub_mmp_magic == MMP_MAGIC && ub->ub_mmp_delay &&
+ spa_get_hostid() == 0) {
+ nvlist_free(label);
+ fnvlist_add_uint64(spa->spa_load_info,
+ ZPOOL_CONFIG_MMP_STATE, MMP_STATE_NO_HOSTID);
+ return (spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO));
+ }
+
+ int error = spa_activity_check(spa, ub, 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);
+ }
+
/*
* 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 we weren't able to find what's necessary for reading the
* MOS in the label, return failure.
*/
- if (label == NULL || nvlist_lookup_nvlist(label,
- ZPOOL_CONFIG_FEATURES_FOR_READ, &features) != 0) {
+ 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));
}
*/
if (ub->ub_version >= SPA_VERSION_FEATURES) {
nvlist_t *unsup_feat;
- nvpair_t *nvp;
VERIFY(nvlist_alloc(&unsup_feat, NV_UNIQUE_NAME, KM_SLEEP) ==
0);
- for (nvp = nvlist_next_nvpair(spa->spa_label_features, NULL);
- nvp != NULL;
+ for (nvpair_t *nvp = nvlist_next_nvpair(spa->spa_label_features,
+ NULL); nvp != NULL;
nvp = nvlist_next_nvpair(spa->spa_label_features, nvp)) {
if (!zfeature_is_supported(nvpair_name(nvp))) {
VERIFY(nvlist_add_string(unsup_feat,
VERIFY(nvlist_add_nvlist(spa->spa_load_info,
ZPOOL_CONFIG_UNSUP_FEAT, unsup_feat) == 0);
nvlist_free(unsup_feat);
+ spa_load_failed(spa, "some features are unsupported");
return (spa_vdev_err(rvd, VDEV_AUX_UNSUP_FEAT,
ENOTSUP));
}
* can handle missing vdevs.
*/
if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_VDEV_CHILDREN,
- &children) != 0 && mosconfig && type != SPA_IMPORT_ASSEMBLE &&
- rvd->vdev_guid_sum != ub->ub_guid_sum)
+ &children) != 0 && trust_config && type != SPA_IMPORT_ASSEMBLE &&
+ rvd->vdev_guid_sum != ub->ub_guid_sum) {
+ spa_load_failed(spa, "guid sum in config doesn't match guid "
+ "sum in uberblock (%llu != %llu)",
+ (u_longlong_t)rvd->vdev_guid_sum,
+ (u_longlong_t)ub->ub_guid_sum);
return (spa_vdev_err(rvd, VDEV_AUX_BAD_GUID_SUM, ENXIO));
+ }
if (type != SPA_IMPORT_ASSEMBLE && spa->spa_config_splitting) {
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa->spa_claim_max_txg = spa->spa_first_txg;
spa->spa_prev_software_version = ub->ub_software_version;
+ 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)
+ 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;
- if (spa_dir_prop(spa, DMU_POOL_CONFIG, &spa->spa_config_object) != 0)
+ return (0);
+}
+
+static int
+spa_ld_validate_config(spa_t *spa, spa_import_type_t type)
+{
+ vdev_t *rvd = spa->spa_root_vdev;
+
+ 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));
+ /*
+ * Validate the config, using the MOS config to fill in any
+ * information which might be missing. If we fail to validate
+ * the config then declare the pool unfit for use. If we're
+ * assembling a pool from a split, the log is not transferred
+ * over.
+ */
+ if (type != SPA_IMPORT_ASSEMBLE) {
+ 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));
+ }
+
+ if (!spa_config_valid(spa, mos_config)) {
+ nvlist_free(mos_config);
+ spa_load_failed(spa, "mismatch between config provided "
+ "and config stored in MOS");
+ return (spa_vdev_err(rvd, VDEV_AUX_BAD_GUID_SUM,
+ ENXIO));
+ }
+ nvlist_free(mos_config);
+
+ /*
+ * Now that we've validated the config, check the state of the
+ * root vdev. If it can't be opened, it indicates one or
+ * more toplevel vdevs are faulted.
+ */
+ if (rvd->vdev_state <= VDEV_STATE_CANT_OPEN) {
+ spa_load_failed(spa, "some top vdevs are unavailable");
+ return (SET_ERROR(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;
- spa_feature_t i;
if (spa_dir_prop(spa, DMU_POOL_FEATURES_FOR_READ,
- &spa->spa_feat_for_read_obj) != 0) {
+ &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) != 0) {
+ &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) != 0) {
+ &spa->spa_feat_desc_obj, B_TRUE) != 0) {
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
unsup_feat, enabled_feat))
missing_feat_read = B_TRUE;
- if (spa_writeable(spa) || state == SPA_LOAD_TRYIMPORT) {
+ if (spa_writeable(spa) ||
+ spa->spa_load_state == SPA_LOAD_TRYIMPORT) {
if (!spa_features_check(spa, B_TRUE,
unsup_feat, enabled_feat)) {
- missing_feat_write = B_TRUE;
+ *missing_feat_writep = B_TRUE;
}
}
* userland in order to know whether to display the
* abovementioned note.
*/
- if (missing_feat_read || (missing_feat_write &&
+ 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 (i = 0; i < SPA_FEATURES; i++) {
+ for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
uint64_t refcount;
error = feature_get_refcount_from_disk(spa,
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) != 0)
+ &spa->spa_feat_enabled_txg_obj, B_TRUE) != 0)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
+ return (0);
+}
+
+static int
+spa_ld_load_special_directories(spa_t *spa)
+{
+ int error = 0;
+ vdev_t *rvd = spa->spa_root_vdev;
+
spa->spa_is_initializing = B_TRUE;
error = dsl_pool_open(spa->spa_dsl_pool);
spa->spa_is_initializing = B_FALSE;
- if (error != 0)
+ if (error != 0) {
+ spa_load_failed(spa, "dsl_pool_open failed [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
+ }
- if (!mosconfig) {
- uint64_t hostid;
- nvlist_t *policy = NULL, *nvconfig;
+ return (0);
+}
- if (load_nvlist(spa, spa->spa_config_object, &nvconfig) != 0)
- return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
+static int
+spa_ld_prepare_for_reload(spa_t *spa, int orig_mode)
+{
+ vdev_t *rvd = spa->spa_root_vdev;
- if (!spa_is_root(spa) && nvlist_lookup_uint64(nvconfig,
- ZPOOL_CONFIG_HOSTID, &hostid) == 0) {
- char *hostname;
- unsigned long myhostid = 0;
+ uint64_t hostid;
+ nvlist_t *policy = NULL;
+ nvlist_t *mos_config;
- VERIFY(nvlist_lookup_string(nvconfig,
- ZPOOL_CONFIG_HOSTNAME, &hostname) == 0);
+ 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));
+ }
-#ifdef _KERNEL
- myhostid = zone_get_hostid(NULL);
-#else /* _KERNEL */
- /*
- * We're emulating the system's hostid in userland, so
- * we can't use zone_get_hostid().
- */
- (void) ddi_strtoul(hw_serial, NULL, 10, &myhostid);
-#endif /* _KERNEL */
- if (hostid != 0 && myhostid != 0 &&
- hostid != myhostid) {
- nvlist_free(nvconfig);
- cmn_err(CE_WARN, "pool '%s' could not be "
- "loaded as it was last accessed by another "
- "system (host: %s hostid: 0x%lx). See: "
- "http://zfsonlinux.org/msg/ZFS-8000-EY",
- spa_name(spa), hostname,
- (unsigned long)hostid);
- return (SET_ERROR(EBADF));
- }
+ if (!spa_is_root(spa) && nvlist_lookup_uint64(mos_config,
+ ZPOOL_CONFIG_HOSTID, &hostid) == 0) {
+ char *hostname;
+ unsigned long myhostid = 0;
+
+ VERIFY(nvlist_lookup_string(mos_config,
+ ZPOOL_CONFIG_HOSTNAME, &hostname) == 0);
+
+ myhostid = spa_get_hostid();
+ if (hostid && myhostid && hostid != myhostid) {
+ nvlist_free(mos_config);
+ return (SET_ERROR(EBADF));
}
- if (nvlist_lookup_nvlist(spa->spa_config,
- ZPOOL_REWIND_POLICY, &policy) == 0)
- VERIFY(nvlist_add_nvlist(nvconfig,
- ZPOOL_REWIND_POLICY, policy) == 0);
+ }
+ if (nvlist_lookup_nvlist(spa->spa_config,
+ ZPOOL_REWIND_POLICY, &policy) == 0)
+ VERIFY(nvlist_add_nvlist(mos_config,
+ ZPOOL_REWIND_POLICY, policy) == 0);
- spa_config_set(spa, nvconfig);
- spa_unload(spa);
- spa_deactivate(spa);
- spa_activate(spa, orig_mode);
+ spa_config_set(spa, mos_config);
+ spa_unload(spa);
+ spa_deactivate(spa);
+ spa_activate(spa, orig_mode);
- return (spa_load(spa, state, SPA_IMPORT_EXISTING, B_TRUE));
+ 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) != 0)
+ 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)
+ 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);
+ 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);
+ &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);
+ 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);
+ &spa->spa_errlog_scrub, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
* Load the history object. If we have an older pool, this
* will not be present.
*/
- error = spa_dir_prop(spa, DMU_POOL_HISTORY, &spa->spa_history);
+ 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));
*/
/* The sentinel is only available in the MOS config. */
- if (load_nvlist(spa, spa->spa_config_object, &mos_config) != 0)
+ 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);
+ &spa->spa_all_vdev_zaps, B_FALSE);
- if (error != ENOENT && error != 0) {
+ if (error == ENOENT) {
+ VERIFY(!nvlist_exists(mos_config,
+ ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS));
+ spa->spa_avz_action = AVZ_ACTION_INITIALIZE;
+ ASSERT0(vdev_count_verify_zaps(spa->spa_root_vdev));
+ } else if (error != 0) {
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
- } else if (error == 0 && !nvlist_exists(mos_config,
- ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS)) {
+ } 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
}
nvlist_free(mos_config);
+ spa->spa_delegation = zpool_prop_default_numeric(ZPOOL_PROP_DELEGATION);
+
+ error = spa_dir_prop(spa, DMU_POOL_PROPS, &spa->spa_pool_props_object,
+ B_FALSE);
+ if (error && error != ENOENT)
+ return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
+
+ if (error == 0) {
+ uint64_t autoreplace;
+
+ spa_prop_find(spa, ZPOOL_PROP_BOOTFS, &spa->spa_bootfs);
+ spa_prop_find(spa, ZPOOL_PROP_AUTOREPLACE, &autoreplace);
+ spa_prop_find(spa, ZPOOL_PROP_DELEGATION, &spa->spa_delegation);
+ spa_prop_find(spa, ZPOOL_PROP_FAILUREMODE, &spa->spa_failmode);
+ spa_prop_find(spa, ZPOOL_PROP_AUTOEXPAND, &spa->spa_autoexpand);
+ spa_prop_find(spa, ZPOOL_PROP_MULTIHOST, &spa->spa_multihost);
+ spa_prop_find(spa, ZPOOL_PROP_DEDUPDITTO,
+ &spa->spa_dedup_ditto);
+
+ spa->spa_autoreplace = (autoreplace != 0);
+ }
+
+ 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
/*
* Load any hot spares for this pool.
*/
- error = spa_dir_prop(spa, DMU_POOL_SPARES, &spa->spa_spares.sav_object);
+ 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->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);
* Load any level 2 ARC devices for this pool.
*/
error = spa_dir_prop(spa, DMU_POOL_L2CACHE,
- &spa->spa_l2cache.sav_object);
+ &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->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->spa_l2cache.sav_sync = B_TRUE;
}
- spa->spa_delegation = zpool_prop_default_numeric(ZPOOL_PROP_DELEGATION);
-
- error = spa_dir_prop(spa, DMU_POOL_PROPS, &spa->spa_pool_props_object);
- if (error && error != ENOENT)
- return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
-
- if (error == 0) {
- uint64_t autoreplace = 0;
+ return (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_DEDUPDITTO,
- &spa->spa_dedup_ditto);
+static int
+spa_ld_load_vdev_metadata(spa_t *spa)
+{
+ int error = 0;
+ vdev_t *rvd = spa->spa_root_vdev;
- spa->spa_autoreplace = (autoreplace != 0);
+ /*
+ * If the 'multihost' property is set, then never allow a pool to
+ * be imported when the system hostid is zero. The exception to
+ * this rule is zdb which is always allowed to access pools.
+ */
+ if (spa_multihost(spa) && spa_get_hostid() == 0 &&
+ (spa->spa_import_flags & ZFS_IMPORT_SKIP_MMP) == 0) {
+ fnvlist_add_uint64(spa->spa_load_info,
+ ZPOOL_CONFIG_MMP_STATE, MMP_STATE_NO_HOSTID);
+ return (spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO));
}
/*
* unopenable vdevs so that the normal autoreplace handler can take
* over.
*/
- if (spa->spa_autoreplace && state != SPA_LOAD_TRYIMPORT) {
+ 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 (state != SPA_LOAD_IMPORT) {
+ 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 state for all toplevel vdevs.
+ * Load the vdev metadata such as metaslabs, DTLs, spacemap object, etc.
*/
- vdev_load(rvd);
+ error = vdev_load(rvd);
+ if (error != 0) {
+ spa_load_failed(spa, "vdev_load failed [error=%d]", error);
+ return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error));
+ }
/*
- * Propagate the leaf DTLs we just loaded all the way up the tree.
+ * Propagate the leaf DTLs we just loaded all the way up the vdev tree.
*/
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
vdev_dtl_reassess(rvd, 0, 0, B_FALSE);
spa_config_exit(spa, SCL_ALL, FTAG);
+ return (0);
+}
+
+static int
+spa_ld_load_dedup_tables(spa_t *spa)
+{
+ int error = 0;
+ vdev_t *rvd = spa->spa_root_vdev;
+
+ error = ddt_load(spa);
+ if (error != 0) {
+ spa_load_failed(spa, "ddt_load failed [error=%d]", error);
+ return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
+ }
+
+ return (0);
+}
+
+static int
+spa_ld_verify_logs(spa_t *spa, spa_import_type_t type, char **ereport)
+{
+ vdev_t *rvd = spa->spa_root_vdev;
+
+ if (type != SPA_IMPORT_ASSEMBLE && spa_writeable(spa)) {
+ boolean_t missing = spa_check_logs(spa);
+ if (missing) {
+ *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;
+
/*
- * Load the DDTs (dedup tables).
+ * We've successfully opened the pool, verify that we're ready
+ * to start pushing transactions.
*/
- error = ddt_load(spa);
+ 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)
+{
+ 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 (config_cache_txg != spa->spa_config_txg ||
+ spa->spa_load_state == SPA_LOAD_IMPORT ||
+ spa->spa_load_state == SPA_LOAD_RECOVER ||
+ (spa->spa_import_flags & ZFS_IMPORT_VERBATIM))
+ need_update = B_TRUE;
+
+ for (int c = 0; c < rvd->vdev_children; c++)
+ if (rvd->vdev_child[c]->vdev_ms_array == 0)
+ need_update = B_TRUE;
+
+ /*
+ * Update the config cache asychronously in case we're the
+ * root pool, in which case the config cache isn't writable yet.
+ */
+ if (need_update)
+ spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
+}
+
+/*
+ * 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, uint64_t pool_guid, nvlist_t *config,
+ spa_load_state_t state, spa_import_type_t type, boolean_t trust_config,
+ char **ereport)
+{
+ int error = 0;
+ uint64_t config_cache_txg = spa->spa_config_txg;
+ int orig_mode = spa->spa_mode;
+ boolean_t missing_feat_write = B_FALSE;
+
+ ASSERT(MUTEX_HELD(&spa_namespace_lock));
+
+ spa->spa_load_state = state;
+ spa_load_note(spa, "LOADING");
+
+ /*
+ * If this is an untrusted config, first access the pool in read-only
+ * mode. We will then retrieve a trusted copy of the config from the MOS
+ * and use it to reopen the pool in read-write mode.
+ */
+ if (!trust_config)
+ spa->spa_mode = FREAD;
+
+ /*
+ * Parse the config provided to create a vdev tree.
+ */
+ error = spa_ld_parse_config(spa, pool_guid, config, type);
if (error != 0)
- return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
+ return (error);
- spa_update_dspace(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);
/*
- * Validate the config, using the MOS config to fill in any
- * information which might be missing. If we fail to validate
- * the config then declare the pool unfit for use. If we're
- * assembling a pool from a split, the log is not transferred
- * over.
+ * Read the label of each vdev and make sure that the GUIDs stored
+ * there match the GUIDs in the config provided.
*/
- if (type != SPA_IMPORT_ASSEMBLE) {
- nvlist_t *nvconfig;
+ error = spa_ld_validate_vdevs(spa, type, trust_config);
+ if (error != 0)
+ return (error);
- if (load_nvlist(spa, spa->spa_config_object, &nvconfig) != 0)
- return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
+ /*
+ * Read 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, config, type, trust_config);
+ if (error != 0)
+ return (error);
- if (!spa_config_valid(spa, nvconfig)) {
- nvlist_free(nvconfig);
- return (spa_vdev_err(rvd, VDEV_AUX_BAD_GUID_SUM,
- ENXIO));
- }
- nvlist_free(nvconfig);
+ /*
+ * 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);
- /*
- * Now that we've validated the config, check the state of the
- * root vdev. If it can't be opened, it indicates one or
- * more toplevel vdevs are faulted.
- */
- if (rvd->vdev_state <= VDEV_STATE_CANT_OPEN)
- return (SET_ERROR(ENXIO));
+ /*
+ * Retrieve the config stored in the MOS and use it to validate the
+ * config provided. Also extract some information from the MOS config
+ * to update our vdev tree.
+ */
+ error = spa_ld_validate_config(spa, type);
+ if (error != 0)
+ return (error);
- if (spa_writeable(spa) && spa_check_logs(spa)) {
- *ereport = FM_EREPORT_ZFS_LOG_REPLAY;
- return (spa_vdev_err(rvd, VDEV_AUX_BAD_LOG, ENXIO));
- }
+ /*
+ * 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);
+
+ /*
+ * If the config provided is not trusted, discard it and use the config
+ * from the MOS to reload the pool.
+ */
+ if (!trust_config) {
+ error = spa_ld_prepare_for_reload(spa, orig_mode);
+ if (error != 0)
+ return (error);
+
+ spa_load_note(spa, "RELOADING");
+ return (spa_load(spa, state, SPA_IMPORT_EXISTING, B_TRUE));
}
+ /*
+ * Retrieve pool properties from the MOS.
+ */
+ error = spa_ld_get_props(spa);
+ if (error != 0)
+ return (error);
+
+ /*
+ * Retrieve the list of auxiliary devices - cache devices and spares -
+ * and open them.
+ */
+ error = spa_ld_open_aux_vdevs(spa, type);
+ if (error != 0)
+ return (error);
+
+ /*
+ * Load the metadata for all vdevs. Also check if unopenable devices
+ * should be autoreplaced.
+ */
+ error = spa_ld_load_vdev_metadata(spa);
+ if (error != 0)
+ return (error);
+
+ error = spa_ld_load_dedup_tables(spa);
+ if (error != 0)
+ return (error);
+
+ /*
+ * Verify the logs now to make sure we don't have any unexpected errors
+ * when we claim log blocks later.
+ */
+ error = spa_ld_verify_logs(spa, type, ereport);
+ if (error != 0)
+ return (error);
+
if (missing_feat_write) {
ASSERT(state == SPA_LOAD_TRYIMPORT);
* read-only mode but not read-write mode. We now have enough
* information and can return to userland.
*/
- return (spa_vdev_err(rvd, VDEV_AUX_UNSUP_FEAT, ENOTSUP));
+ return (spa_vdev_err(spa->spa_root_vdev, VDEV_AUX_UNSUP_FEAT,
+ ENOTSUP));
}
/*
- * We've successfully opened the pool, verify that we're ready
- * to start pushing transactions.
+ * 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.
*/
- if (state != SPA_LOAD_TRYIMPORT) {
- if ((error = spa_load_verify(spa)))
- return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA,
- error));
- }
+ 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) && (state == SPA_LOAD_RECOVER ||
spa->spa_load_max_txg == UINT64_MAX)) {
- dmu_tx_t *tx;
- int need_update = B_FALSE;
- dsl_pool_t *dp = spa_get_dsl(spa);
- int c;
-
ASSERT(state != SPA_LOAD_TRYIMPORT);
/*
- * 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.
+ * Traverse the ZIL and claim all blocks.
*/
- 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_ld_claim_log_blocks(spa);
- spa_set_log_state(spa, SPA_LOG_GOOD);
+ /*
+ * 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 either zil_check_log_chain()
- * (invoked from spa_check_logs()) or zil_claim() above.
+ * will have been set for us by ZIL traversal operations
+ * performed above.
*/
txg_wait_synced(spa->spa_dsl_pool, spa->spa_claim_max_txg);
/*
- * 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 (config_cache_txg != spa->spa_config_txg ||
- state == SPA_LOAD_IMPORT ||
- state == SPA_LOAD_RECOVER ||
- (spa->spa_import_flags & ZFS_IMPORT_VERBATIM))
- need_update = B_TRUE;
-
- for (c = 0; c < rvd->vdev_children; c++)
- if (rvd->vdev_child[c]->vdev_ms_array == 0)
- need_update = B_TRUE;
-
- /*
- * Update the config cache asychronously in case we're the
- * root pool, in which case the config cache isn't writable yet.
+ * 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).
*/
- if (need_update)
- spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
+ spa_ld_check_for_config_update(spa, config_cache_txg);
/*
* Check all DTLs to see if anything needs resilvering.
*/
if (!dsl_scan_resilvering(spa->spa_dsl_pool) &&
- vdev_resilver_needed(rvd, NULL, NULL))
+ vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL))
spa_async_request(spa, SPA_ASYNC_RESILVER);
/*
* Log the fact that we booted up (so that we can detect if
* we rebooted in the middle of an operation).
*/
- spa_history_log_version(spa, "open");
+ spa_history_log_version(spa, "open", NULL);
/*
* Delete any inconsistent datasets.
* Clean up any stale temporary dataset userrefs.
*/
dsl_pool_clean_tmp_userrefs(spa->spa_dsl_pool);
+
+ spa_restart_removal(spa);
+
+ spa_spawn_aux_threads(spa);
}
+ spa_load_note(spa, "LOADED");
+
return (0);
}
static int
-spa_load_retry(spa_t *spa, spa_load_state_t state, int mosconfig)
+spa_load_retry(spa_t *spa, spa_load_state_t state, int trust_config)
{
int mode = spa->spa_mode;
spa_activate(spa, mode);
spa_async_suspend(spa);
- return (spa_load(spa, state, SPA_IMPORT_EXISTING, mosconfig));
+ 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, trust_config));
}
/*
* spa_load().
*/
static int
-spa_load_best(spa_t *spa, spa_load_state_t state, int mosconfig,
+spa_load_best(spa_t *spa, spa_load_state_t state, int trust_config,
uint64_t max_request, int rewind_flags)
{
nvlist_t *loadinfo = NULL;
}
load_error = rewind_error = spa_load(spa, state, SPA_IMPORT_EXISTING,
- mosconfig);
+ trust_config);
if (load_error == 0)
return (0);
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, mosconfig);
+ rewind_error = spa_load_retry(spa, state, trust_config);
}
spa->spa_extreme_rewind = B_FALSE;
* up calling spa_open() again. The real fix is to figure out how to
* avoid dsl_dir_open() calling this in the first place.
*/
- if (mutex_owner(&spa_namespace_lock) != curthread) {
+ if (MUTEX_NOT_HELD(&spa_namespace_lock)) {
mutex_enter(&spa_namespace_lock);
locked = B_TRUE;
}
if (state != SPA_LOAD_RECOVER)
spa->spa_last_ubsync_txg = spa->spa_load_txg = 0;
+ zfs_dbgmsg("spa_open_common: opening %s", pool);
error = spa_load_best(spa, state, B_FALSE, policy.zrp_txg,
policy.zrp_request);
*/
spa_unload(spa);
spa_deactivate(spa);
- spa_config_sync(spa, B_TRUE, B_TRUE);
+ spa_write_cachefile(spa, B_TRUE, B_TRUE);
spa_remove(spa);
if (locked)
mutex_exit(&spa_namespace_lock);
ZPOOL_CONFIG_ERRCOUNT,
spa_get_errlog_size(spa)) == 0);
- if (spa_suspended(spa))
+ if (spa_suspended(spa)) {
VERIFY(nvlist_add_uint64(*config,
ZPOOL_CONFIG_SUSPENDED,
spa->spa_failmode) == 0);
+ VERIFY(nvlist_add_uint64(*config,
+ ZPOOL_CONFIG_SUSPENDED_REASON,
+ spa->spa_suspended) == 0);
+ }
spa_add_spares(spa, *config);
spa_add_l2cache(spa, *config);
goto out;
}
- /*
- * The L2ARC currently only supports disk devices in
- * kernel context. For user-level testing, we allow it.
- */
-#ifdef _KERNEL
- if ((strcmp(config, ZPOOL_CONFIG_L2CACHE) == 0) &&
- strcmp(vd->vdev_ops->vdev_op_type, VDEV_TYPE_DISK) != 0) {
- error = SET_ERROR(ENOTBLK);
- vdev_free(vd);
- goto out;
- }
-#endif
vd->vdev_top = vd;
if ((error = vdev_open(vd)) == 0 &&
nvlist_t **newdevs;
/*
- * Generate new dev list by concatentating with the
+ * Generate new dev list by concatenating with the
* current dev list.
*/
VERIFY(nvlist_lookup_nvlist_array(sav->sav_config, config,
}
}
+/*
+ * 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));
+}
+
/*
* Pool Creation
*/
int
spa_create(const char *pool, nvlist_t *nvroot, nvlist_t *props,
- nvlist_t *zplprops)
+ nvlist_t *zplprops, dsl_crypto_params_t *dcp)
{
spa_t *spa;
char *altroot = NULL;
uint64_t txg = TXG_INITIAL;
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
- uint64_t version, obj;
+ uint64_t version, obj, root_dsobj = 0;
boolean_t has_features;
- nvpair_t *elem;
- int c, i;
+ boolean_t has_encryption;
+ spa_feature_t feat;
+ char *feat_name;
char *poolname;
nvlist_t *nvl;
spa->spa_import_flags |= ZFS_IMPORT_TEMP_NAME;
has_features = B_FALSE;
- for (elem = nvlist_next_nvpair(props, NULL);
+ has_encryption = 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)))
+ 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;
+ }
+ }
+
+ /* 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_features || nvlist_lookup_uint64(props,
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;
/*
* Create "The Godfather" zio to hold all async IOs
*/
spa->spa_async_zio_root = kmem_alloc(max_ncpus * sizeof (void *),
KM_SLEEP);
- for (i = 0; i < max_ncpus; i++) {
+ 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);
(error = vdev_create(rvd, txg, B_FALSE)) == 0 &&
(error = spa_validate_aux(spa, nvroot, txg,
VDEV_ALLOC_ADD)) == 0) {
- for (c = 0; c < rvd->vdev_children; c++) {
+ for (int c = 0; c < rvd->vdev_children; c++) {
vdev_metaslab_set_size(rvd->vdev_child[c]);
vdev_expand(rvd->vdev_child[c], txg);
}
}
spa->spa_is_initializing = B_TRUE;
- spa->spa_dsl_pool = dp = dsl_pool_create(spa, zplprops, txg);
- spa->spa_meta_objset = dp->dp_meta_objset;
+ spa->spa_dsl_pool = dp = dsl_pool_create(spa, zplprops, dcp, txg);
spa->spa_is_initializing = B_FALSE;
/*
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.
*/
cmn_err(CE_PANIC, "failed to add pool config");
}
- if (spa_version(spa) >= SPA_VERSION_FEATURES)
- spa_feature_create_zap_objects(spa, tx);
-
if (zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CREATION_VERSION,
sizeof (uint64_t), 1, &version, tx) != 0) {
spa->spa_meta_objset, obj));
/*
- * Create the pool's history object.
+ * Generate some random noise for salted checksums to operate on.
*/
- if (version >= SPA_VERSION_ZPOOL_HISTORY)
- spa_history_create_obj(spa, tx);
+ (void) random_get_pseudo_bytes(spa->spa_cksum_salt.zcs_bytes,
+ sizeof (spa->spa_cksum_salt.zcs_bytes));
/*
* Set pool properties.
spa->spa_delegation = zpool_prop_default_numeric(ZPOOL_PROP_DELEGATION);
spa->spa_failmode = zpool_prop_default_numeric(ZPOOL_PROP_FAILUREMODE);
spa->spa_autoexpand = zpool_prop_default_numeric(ZPOOL_PROP_AUTOEXPAND);
+ spa->spa_multihost = zpool_prop_default_numeric(ZPOOL_PROP_MULTIHOST);
if (props != NULL) {
spa_configfile_set(spa, props, B_FALSE);
dmu_tx_commit(tx);
- spa->spa_sync_on = B_TRUE;
- txg_sync_start(spa->spa_dsl_pool);
-
/*
- * We explicitly wait for the first transaction to complete so that our
- * bean counters are appropriately updated.
+ * If the root dataset is encrypted we will need to create key mappings
+ * for the zio layer before we start to write any data to disk and hold
+ * them until after the first txg has been synced. Waiting for the first
+ * transaction to complete also ensures that our bean counters are
+ * appropriately updated.
*/
- txg_wait_synced(spa->spa_dsl_pool, txg);
+ if (dp->dp_root_dir->dd_crypto_obj != 0) {
+ root_dsobj = dsl_dir_phys(dp->dp_root_dir)->dd_head_dataset_obj;
+ VERIFY0(spa_keystore_create_mapping_impl(spa, root_dsobj,
+ dp->dp_root_dir, FTAG));
+ }
+
+ spa->spa_sync_on = B_TRUE;
+ txg_sync_start(dp);
+ mmp_thread_start(spa);
+ txg_wait_synced(dp, txg);
+
+ if (dp->dp_root_dir->dd_crypto_obj != 0)
+ VERIFY0(spa_keystore_remove_mapping(spa, root_dsobj, FTAG));
- spa_config_sync(spa, B_FALSE, B_TRUE);
- spa_event_notify(spa, NULL, ESC_ZFS_POOL_CREATE);
+ spa_spawn_aux_threads(spa);
- spa_history_log_version(spa, "create");
+ spa_write_cachefile(spa, B_FALSE, B_TRUE);
/*
* Don't count references from objsets that are already closed
*/
spa_evicting_os_wait(spa);
spa->spa_minref = refcount_count(&spa->spa_refcount);
+ spa->spa_load_state = SPA_LOAD_NONE;
mutex_exit(&spa_namespace_lock);
if (props != NULL)
spa_configfile_set(spa, props, B_FALSE);
- spa_config_sync(spa, B_FALSE, B_TRUE);
- spa_event_notify(spa, NULL, ESC_ZFS_POOL_IMPORT);
-
+ spa_write_cachefile(spa, B_FALSE, B_TRUE);
+ spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_IMPORT);
+ zfs_dbgmsg("spa_import: verbatim import of %s", pool);
mutex_exit(&spa_namespace_lock);
return (0);
}
state = SPA_LOAD_RECOVER;
/*
- * Pass off the heavy lifting to spa_load(). Pass TRUE for mosconfig
+ * Pass off the heavy lifting to spa_load(). Pass TRUE for trust_config
* because the user-supplied config is actually the one to trust when
* doing an import.
*/
if (state != SPA_LOAD_RECOVER)
spa->spa_last_ubsync_txg = spa->spa_load_txg = 0;
+ zfs_dbgmsg("spa_import: importing %s%s", pool,
+ (state == SPA_LOAD_RECOVER) ? " (RECOVERY MODE)" : "");
error = spa_load_best(spa, state, B_TRUE, policy.zrp_txg,
policy.zrp_request);
VERIFY(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
- if (error == 0)
- error = spa_validate_aux(spa, nvroot, -1ULL,
- VDEV_ALLOC_SPARE);
- if (error == 0)
- error = spa_validate_aux(spa, nvroot, -1ULL,
- VDEV_ALLOC_L2CACHE);
spa_config_exit(spa, SCL_ALL, FTAG);
if (props != NULL)
*/
spa_async_request(spa, SPA_ASYNC_AUTOEXPAND);
- spa_history_log_version(spa, "import");
+ spa_history_log_version(spa, "import", NULL);
- spa_event_notify(spa, NULL, ESC_ZFS_POOL_IMPORT);
+ spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_IMPORT);
zvol_create_minors(spa, pool, B_TRUE);
spa = spa_add(TRYIMPORT_NAME, tryconfig, NULL);
spa_activate(spa, FREAD);
+ zfs_dbgmsg("spa_tryimport: importing %s", poolname);
+
/*
* Pass off the heavy lifting to spa_load().
- * Pass TRUE for mosconfig because the user-supplied config
+ * Pass TRUE for trust_config because the user-supplied config
* is actually the one to trust when doing an import.
*/
error = spa_load(spa, SPA_LOAD_TRYIMPORT, SPA_IMPORT_EXISTING, B_TRUE);
}
export_spa:
- spa_event_notify(spa, NULL, ESC_ZFS_POOL_DESTROY);
+ 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);
if (new_state != POOL_STATE_UNINITIALIZED) {
if (!hardforce)
- spa_config_sync(spa, B_TRUE, B_TRUE);
+ spa_write_cachefile(spa, B_TRUE, B_TRUE);
spa_remove(spa);
}
mutex_exit(&spa_namespace_lock);
vdev_t *vd, *tvd;
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
- int c;
ASSERT(spa_writeable(spa));
return (spa_vdev_exit(spa, vd, txg, error));
/*
- * Transfer each new top-level vdev from vd to rvd.
+ * If we are in the middle of a device removal, we can only add
+ * devices which match the existing devices in the pool.
+ * If we are in the middle of a removal, or have some indirect
+ * vdevs, we can not add raidz toplevels.
*/
- for (c = 0; c < vd->vdev_children; c++) {
+ if (spa->spa_vdev_removal != NULL ||
+ spa->spa_removing_phys.sr_prev_indirect_vdev != -1) {
+ for (int c = 0; c < vd->vdev_children; c++) {
+ tvd = vd->vdev_child[c];
+ if (spa->spa_vdev_removal != NULL &&
+ tvd->vdev_ashift != spa->spa_max_ashift) {
+ return (spa_vdev_exit(spa, vd, txg, EINVAL));
+ }
+ /* Fail if top level vdev is raidz */
+ if (tvd->vdev_ops == &vdev_raidz_ops) {
+ return (spa_vdev_exit(spa, vd, txg, EINVAL));
+ }
+ /*
+ * Need the top level mirror to be
+ * a mirror of leaf vdevs only
+ */
+ if (tvd->vdev_ops == &vdev_mirror_ops) {
+ for (uint64_t cid = 0;
+ cid < tvd->vdev_children; cid++) {
+ vdev_t *cvd = tvd->vdev_child[cid];
+ if (!cvd->vdev_ops->vdev_op_leaf) {
+ return (spa_vdev_exit(spa, vd,
+ txg, EINVAL));
+ }
+ }
+ }
+ }
+ }
+
+ for (int c = 0; c < vd->vdev_children; c++) {
/*
* Set the vdev id to the first hole, if one exists.
mutex_enter(&spa_namespace_lock);
spa_config_update(spa, SPA_CONFIG_UPDATE_POOL);
- spa_event_notify(spa, NULL, ESC_ZFS_VDEV_ADD);
+ spa_event_notify(spa, NULL, NULL, ESC_ZFS_VDEV_ADD);
mutex_exit(&spa_namespace_lock);
return (0);
spa_vdev_attach(spa_t *spa, uint64_t guid, nvlist_t *nvroot, int replacing)
{
uint64_t txg, dtl_max_txg;
+ ASSERTV(vdev_t *rvd = spa->spa_root_vdev);
vdev_t *oldvd, *newvd, *newrootvd, *pvd, *tvd;
vdev_ops_t *pvops;
char *oldvdpath, *newvdpath;
int newvd_isspare;
int error;
- ASSERTV(vdev_t *rvd = spa->spa_root_vdev);
ASSERT(spa_writeable(spa));
oldvd = spa_lookup_by_guid(spa, guid, B_FALSE);
+ if (spa->spa_vdev_removal != NULL)
+ return (spa_vdev_exit(spa, NULL, txg, EBUSY));
+
if (oldvd == NULL)
return (spa_vdev_exit(spa, NULL, txg, ENODEV));
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);
if (newvd->vdev_isspare) {
spa_spare_activate(newvd);
- spa_event_notify(spa, newvd, ESC_ZFS_VDEV_SPARE);
+ spa_event_notify(spa, newvd, NULL, ESC_ZFS_VDEV_SPARE);
}
oldvdpath = spa_strdup(oldvd->vdev_path);
dsl_resilver_restart(spa->spa_dsl_pool, dtl_max_txg);
if (spa->spa_bootfs)
- spa_event_notify(spa, newvd, ESC_ZFS_BOOTFS_VDEV_ATTACH);
+ spa_event_notify(spa, newvd, NULL, ESC_ZFS_BOOTFS_VDEV_ATTACH);
- spa_event_notify(spa, newvd, ESC_ZFS_VDEV_ATTACH);
+ spa_event_notify(spa, newvd, NULL, ESC_ZFS_VDEV_ATTACH);
/*
* Commit the config
{
uint64_t txg;
int error;
+ ASSERTV(vdev_t *rvd = spa->spa_root_vdev);
vdev_t *vd, *pvd, *cvd, *tvd;
boolean_t unspare = B_FALSE;
uint64_t unspare_guid = 0;
char *vdpath;
- int c, t;
- ASSERTV(vdev_t *rvd = spa->spa_root_vdev);
+
ASSERT(spa_writeable(spa));
txg = spa_vdev_enter(spa);
vd->vdev_path != NULL) {
size_t len = strlen(vd->vdev_path);
- for (c = 0; c < pvd->vdev_children; c++) {
+ for (int c = 0; c < pvd->vdev_children; c++) {
cvd = pvd->vdev_child[c];
if (cvd == vd || cvd->vdev_path == NULL)
* But first make sure we're not on any *other* txg's DTL list, to
* prevent vd from being accessed after it's freed.
*/
- vdpath = spa_strdup(vd->vdev_path);
- for (t = 0; t < TXG_SIZE; t++)
+ 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, ESC_ZFS_VDEV_REMOVE);
+ spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_REMOVE);
/* hang on to the spa before we release the lock */
spa_open_ref(spa, FTAG);
/* 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_offline_log(spa);
+ error = spa_reset_logs(spa);
txg = spa_vdev_config_enter(spa);
if (activate_slog)
vdev_t *vd = rvd->vdev_child[c];
/* don't count the holes & logs as children */
- if (vd->vdev_islog || vd->vdev_ishole) {
+ if (vd->vdev_islog || !vdev_is_concrete(vd)) {
if (lastlog == 0)
lastlog = c;
continue;
/* make sure there's nothing stopping the split */
if (vml[c]->vdev_parent->vdev_ops != &vdev_mirror_ops ||
vml[c]->vdev_islog ||
- vml[c]->vdev_ishole ||
+ !vdev_is_concrete(vml[c]) ||
vml[c]->vdev_isspare ||
vml[c]->vdev_isl2cache ||
!vdev_writeable(vml[c]) ||
return (error);
}
-static nvlist_t *
-spa_nvlist_lookup_by_guid(nvlist_t **nvpp, int count, uint64_t target_guid)
-{
- int i;
-
- for (i = 0; i < count; i++) {
- uint64_t guid;
-
- VERIFY(nvlist_lookup_uint64(nvpp[i], ZPOOL_CONFIG_GUID,
- &guid) == 0);
-
- if (guid == target_guid)
- return (nvpp[i]);
- }
-
- return (NULL);
-}
-
-static void
-spa_vdev_remove_aux(nvlist_t *config, char *name, nvlist_t **dev, int count,
- nvlist_t *dev_to_remove)
-{
- nvlist_t **newdev = NULL;
- int i, j;
-
- if (count > 1)
- newdev = kmem_alloc((count - 1) * sizeof (void *), KM_SLEEP);
-
- for (i = 0, j = 0; i < count; i++) {
- if (dev[i] == dev_to_remove)
- continue;
- VERIFY(nvlist_dup(dev[i], &newdev[j++], KM_SLEEP) == 0);
- }
-
- VERIFY(nvlist_remove(config, name, DATA_TYPE_NVLIST_ARRAY) == 0);
- VERIFY(nvlist_add_nvlist_array(config, name, newdev, count - 1) == 0);
-
- for (i = 0; i < count - 1; i++)
- nvlist_free(newdev[i]);
-
- if (count > 1)
- kmem_free(newdev, (count - 1) * sizeof (void *));
-}
-
-/*
- * Evacuate the device.
- */
-static int
-spa_vdev_remove_evacuate(spa_t *spa, vdev_t *vd)
-{
- uint64_t txg;
- int error = 0;
-
- ASSERT(MUTEX_HELD(&spa_namespace_lock));
- ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0);
- ASSERT(vd == vd->vdev_top);
-
- /*
- * Evacuate the device. We don't hold the config lock as writer
- * since we need to do I/O but we do keep the
- * spa_namespace_lock held. Once this completes the device
- * should no longer have any blocks allocated on it.
- */
- if (vd->vdev_islog) {
- if (vd->vdev_stat.vs_alloc != 0)
- error = spa_offline_log(spa);
- } else {
- error = SET_ERROR(ENOTSUP);
- }
-
- if (error)
- return (error);
-
- /*
- * The evacuation succeeded. Remove any remaining MOS metadata
- * associated with this vdev, and wait for these changes to sync.
- */
- ASSERT0(vd->vdev_stat.vs_alloc);
- txg = spa_vdev_config_enter(spa);
- vd->vdev_removing = B_TRUE;
- vdev_dirty_leaves(vd, VDD_DTL, txg);
- vdev_config_dirty(vd);
- spa_vdev_config_exit(spa, NULL, txg, 0, FTAG);
-
- return (0);
-}
-
-/*
- * Complete the removal by cleaning up the namespace.
- */
-static void
-spa_vdev_remove_from_namespace(spa_t *spa, vdev_t *vd)
-{
- vdev_t *rvd = spa->spa_root_vdev;
- uint64_t id = vd->vdev_id;
- boolean_t last_vdev = (id == (rvd->vdev_children - 1));
-
- ASSERT(MUTEX_HELD(&spa_namespace_lock));
- ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
- ASSERT(vd == vd->vdev_top);
-
- /*
- * Only remove any devices which are empty.
- */
- if (vd->vdev_stat.vs_alloc != 0)
- return;
-
- (void) vdev_label_init(vd, 0, VDEV_LABEL_REMOVE);
-
- if (list_link_active(&vd->vdev_state_dirty_node))
- vdev_state_clean(vd);
- if (list_link_active(&vd->vdev_config_dirty_node))
- vdev_config_clean(vd);
-
- vdev_free(vd);
-
- if (last_vdev) {
- vdev_compact_children(rvd);
- } else {
- vd = vdev_alloc_common(spa, id, 0, &vdev_hole_ops);
- vdev_add_child(rvd, vd);
- }
- vdev_config_dirty(rvd);
-
- /*
- * Reassess the health of our root vdev.
- */
- vdev_reopen(rvd);
-}
-
-/*
- * Remove a device from the pool -
- *
- * Removing a device from the vdev namespace requires several steps
- * and can take a significant amount of time. As a result we use
- * the spa_vdev_config_[enter/exit] functions which allow us to
- * grab and release the spa_config_lock while still holding the namespace
- * lock. During each step the configuration is synced out.
- *
- * Currently, this supports removing only hot spares, slogs, and level 2 ARC
- * devices.
- */
-int
-spa_vdev_remove(spa_t *spa, uint64_t guid, boolean_t unspare)
-{
- vdev_t *vd;
- metaslab_group_t *mg;
- nvlist_t **spares, **l2cache, *nv;
- uint64_t txg = 0;
- uint_t nspares, nl2cache;
- int error = 0;
- boolean_t locked = MUTEX_HELD(&spa_namespace_lock);
-
- ASSERT(spa_writeable(spa));
-
- if (!locked)
- txg = spa_vdev_enter(spa);
-
- vd = spa_lookup_by_guid(spa, guid, B_FALSE);
-
- if (spa->spa_spares.sav_vdevs != NULL &&
- nvlist_lookup_nvlist_array(spa->spa_spares.sav_config,
- ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0 &&
- (nv = spa_nvlist_lookup_by_guid(spares, nspares, guid)) != NULL) {
- /*
- * Only remove the hot spare if it's not currently in use
- * in this pool.
- */
- if (vd == NULL || unspare) {
- spa_vdev_remove_aux(spa->spa_spares.sav_config,
- ZPOOL_CONFIG_SPARES, spares, nspares, nv);
- spa_load_spares(spa);
- spa->spa_spares.sav_sync = B_TRUE;
- } else {
- error = SET_ERROR(EBUSY);
- }
- spa_event_notify(spa, vd, ESC_ZFS_VDEV_REMOVE_AUX);
- } else if (spa->spa_l2cache.sav_vdevs != NULL &&
- nvlist_lookup_nvlist_array(spa->spa_l2cache.sav_config,
- ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0 &&
- (nv = spa_nvlist_lookup_by_guid(l2cache, nl2cache, guid)) != NULL) {
- /*
- * Cache devices can always be removed.
- */
- spa_vdev_remove_aux(spa->spa_l2cache.sav_config,
- ZPOOL_CONFIG_L2CACHE, l2cache, nl2cache, nv);
- spa_load_l2cache(spa);
- spa->spa_l2cache.sav_sync = B_TRUE;
- spa_event_notify(spa, vd, ESC_ZFS_VDEV_REMOVE_AUX);
- } else if (vd != NULL && vd->vdev_islog) {
- ASSERT(!locked);
- ASSERT(vd == vd->vdev_top);
-
- mg = vd->vdev_mg;
-
- /*
- * Stop allocating from this vdev.
- */
- metaslab_group_passivate(mg);
-
- /*
- * Wait for the youngest allocations and frees to sync,
- * and then wait for the deferral of those frees to finish.
- */
- spa_vdev_config_exit(spa, NULL,
- txg + TXG_CONCURRENT_STATES + TXG_DEFER_SIZE, 0, FTAG);
-
- /*
- * Attempt to evacuate the vdev.
- */
- error = spa_vdev_remove_evacuate(spa, vd);
-
- txg = spa_vdev_config_enter(spa);
-
- /*
- * If we couldn't evacuate the vdev, unwind.
- */
- if (error) {
- metaslab_group_activate(mg);
- return (spa_vdev_exit(spa, NULL, txg, error));
- }
-
- /*
- * Clean up the vdev namespace.
- */
- spa_vdev_remove_from_namespace(spa, vd);
-
- spa_event_notify(spa, vd, ESC_ZFS_VDEV_REMOVE_DEV);
- } else if (vd != NULL) {
- /*
- * Normal vdevs cannot be removed (yet).
- */
- error = SET_ERROR(ENOTSUP);
- } else {
- /*
- * There is no vdev of any kind with the specified guid.
- */
- error = SET_ERROR(ENOENT);
- }
-
- if (!locked)
- return (spa_vdev_exit(spa, NULL, txg, error));
-
- return (error);
-}
-
/*
* Find any device that's done replacing, or a vdev marked 'unspare' that's
* currently spared, so we can detach it.
spa_vdev_resilver_done_hunt(vdev_t *vd)
{
vdev_t *newvd, *oldvd;
- int c;
- for (c = 0; c < vd->vdev_children; c++) {
+ for (int c = 0; c < vd->vdev_children; c++) {
oldvd = spa_vdev_resilver_done_hunt(vd->vdev_child[c]);
if (oldvd != NULL)
return (oldvd);
* 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)
static void
spa_async_remove(spa_t *spa, vdev_t *vd)
{
- int c;
-
if (vd->vdev_remove_wanted) {
vd->vdev_remove_wanted = B_FALSE;
vd->vdev_delayed_close = B_FALSE;
vdev_state_dirty(vd->vdev_top);
}
- for (c = 0; c < vd->vdev_children; c++)
+ 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)
{
- int c;
-
if (vd->vdev_probe_wanted) {
vd->vdev_probe_wanted = B_FALSE;
vdev_reopen(vd); /* vdev_open() does the actual probe */
}
- for (c = 0; c < vd->vdev_children; c++)
+ 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)
{
- int c;
-
if (!spa->spa_autoexpand)
return;
- for (c = 0; c < vd->vdev_children; c++) {
+ 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, ESC_ZFS_VDEV_AUTOEXPAND);
+ spa_event_notify(vd->vdev_spa, vd, NULL, ESC_ZFS_VDEV_AUTOEXPAND);
}
static void
-spa_async_thread(spa_t *spa)
+spa_async_thread(void *arg)
{
- int tasks, i;
+ spa_t *spa = (spa_t *)arg;
+ int tasks;
ASSERT(spa->spa_sync_on);
if (tasks & SPA_ASYNC_REMOVE) {
spa_vdev_state_enter(spa, SCL_NONE);
spa_async_remove(spa, spa->spa_root_vdev);
- for (i = 0; i < spa->spa_l2cache.sav_count; i++)
+ for (int i = 0; i < spa->spa_l2cache.sav_count; i++)
spa_async_remove(spa, spa->spa_l2cache.sav_vdevs[i]);
- for (i = 0; i < spa->spa_spares.sav_count; 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);
}
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_isrunning(condense_thread))
+ VERIFY0(zthr_cancel(condense_thread));
}
void
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_isrunning(condense_thread))
+ zthr_resume(condense_thread);
}
static boolean_t
} else {
config_task_suspended =
(gethrtime() - spa->spa_ccw_fail_time) <
- (zfs_ccw_retry_interval * NANOSEC);
+ ((hrtime_t)zfs_ccw_retry_interval * NANOSEC);
}
return (non_config_tasks || (config_task && !config_task_suspended));
spa_avz_build(vdev_t *vd, uint64_t avz, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
- uint64_t i;
if (vd->vdev_top_zap != 0) {
VERIFY0(zap_add_int(spa->spa_meta_objset, avz,
VERIFY0(zap_add_int(spa->spa_meta_objset, avz,
vd->vdev_leaf_zap, tx));
}
- for (i = 0; i < vd->vdev_children; i++) {
+ for (uint64_t i = 0; i < vd->vdev_children; i++) {
spa_avz_build(vd->vdev_child[i], avz, tx);
}
}
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) {
- zap_cursor_t zc;
- zap_attribute_t za;
-
/* 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;
zprop_type_t proptype;
spa_feature_t fid;
- prop = zpool_name_to_prop(nvpair_name(elem));
- switch ((int)prop) {
- case ZPROP_INVAL:
+ switch (prop = zpool_name_to_prop(nvpair_name(elem))) {
+ case ZPOOL_PROP_INVAL:
/*
* We checked this earlier in spa_prop_validate().
*/
case ZPOOL_PROP_VERSION:
intval = fnvpair_value_uint64(elem);
/*
- * The version is synced seperatly before other
+ * The version is synced separately before other
* properties and should be correct by now.
*/
ASSERT3U(spa_version(spa), >=, intval);
* We need to dirty the configuration on all the vdevs
* so that their labels get updated. It's unnecessary
* to do this for pool creation since the vdev's
- * configuratoin has already been dirtied.
+ * configuration has already been dirtied.
*/
if (tx->tx_txg != TXG_INITIAL)
vdev_config_dirty(spa->spa_root_vdev);
spa_async_request(spa,
SPA_ASYNC_AUTOEXPAND);
break;
+ case ZPOOL_PROP_MULTIHOST:
+ spa->spa_multihost = intval;
+ break;
case ZPOOL_PROP_DEDUPDITTO:
spa->spa_dedup_ditto = intval;
break;
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)
+{
+ ASSERTV(vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping);
+ ASSERTV(vdev_indirect_births_t *vib = vd->vdev_indirect_births);
+
+ if (vd->vdev_ops == &vdev_indirect_ops) {
+ ASSERT(vim != NULL);
+ ASSERT(vib != NULL);
+ }
+
+ if (vdev_obsolete_sm_object(vd) != 0) {
+ ASSERT(vd->vdev_obsolete_sm != NULL);
+ ASSERT(vd->vdev_removing ||
+ vd->vdev_ops == &vdev_indirect_ops);
+ ASSERT(vdev_indirect_mapping_num_entries(vim) > 0);
+ ASSERT(vdev_indirect_mapping_bytes_mapped(vim) > 0);
+
+ ASSERT3U(vdev_obsolete_sm_object(vd), ==,
+ space_map_object(vd->vdev_obsolete_sm));
+ ASSERT3U(vdev_indirect_mapping_bytes_mapped(vim), >=,
+ space_map_allocated(vd->vdev_obsolete_sm));
+ }
+ ASSERT(vd->vdev_obsolete_segments != NULL);
+
+ /*
+ * Since frees / remaps to an indirect vdev can only
+ * happen in syncing context, the obsolete segments
+ * tree must be empty when we start syncing.
+ */
+ ASSERT0(range_tree_space(vd->vdev_obsolete_segments));
+}
+
/*
* Sync the specified transaction group. New blocks may be dirtied as
* part of the process, so we iterate until it converges.
vdev_t *vd;
dmu_tx_t *tx;
int error;
- int c;
+ uint32_t max_queue_depth = zfs_vdev_async_write_max_active *
+ zfs_vdev_queue_depth_pct / 100;
VERIFY(spa_writeable(spa));
+ /*
+ * Wait for i/os issued in open context that need to complete
+ * before this txg syncs.
+ */
+ VERIFY0(zio_wait(spa->spa_txg_zio[txg & TXG_MASK]));
+ spa->spa_txg_zio[txg & TXG_MASK] = zio_root(spa, NULL, NULL, 0);
+
/*
* Lock out configuration changes.
*/
spa->spa_syncing_txg = txg;
spa->spa_sync_pass = 0;
+ mutex_enter(&spa->spa_alloc_lock);
+ VERIFY0(avl_numnodes(&spa->spa_alloc_tree));
+ mutex_exit(&spa->spa_alloc_lock);
+
/*
* If there are any pending vdev state changes, convert them
* into config changes that go out with this transaction group.
tx = dmu_tx_create_assigned(dp, txg);
spa->spa_sync_starttime = gethrtime();
- taskq_cancel_id(system_taskq, spa->spa_deadman_tqid);
- spa->spa_deadman_tqid = taskq_dispatch_delay(system_taskq,
+ 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));
}
}
+ /*
+ * 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.
+ */
+ uint64_t queue_depth_total = 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 || mg->mg_class != spa_normal_class(spa) ||
+ !metaslab_group_initialized(mg))
+ 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().
+ */
+ ASSERT0(refcount_count(&mg->mg_alloc_queue_depth));
+ mg->mg_max_alloc_queue_depth = max_queue_depth;
+ queue_depth_total += mg->mg_max_alloc_queue_depth;
+ }
+ metaslab_class_t *mc = spa_normal_class(spa);
+ ASSERT0(refcount_count(&mc->mc_alloc_slots));
+ mc->mc_alloc_max_slots = queue_depth_total;
+ mc->mc_alloc_throttle_enabled = zio_dva_throttle_enabled;
+
+ ASSERT3U(mc->mc_alloc_max_slots, <=,
+ max_queue_depth * rvd->vdev_children);
+
+ 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;
+ }
+ }
+
/*
* Iterate to convergence.
*/
ddt_sync(spa, txg);
dsl_scan_sync(dp, tx);
- while ((vd = txg_list_remove(&spa->spa_vdev_txg_list, txg)))
+ if (spa->spa_vdev_removal != NULL)
+ svr_sync(spa, tx);
+
+ while ((vd = txg_list_remove(&spa->spa_vdev_txg_list, txg))
+ != NULL)
vdev_sync(vd, txg);
if (pass == 1) {
}
#endif
+ if (spa->spa_vdev_removal != NULL) {
+ ASSERT0(spa->spa_vdev_removal->svr_bytes_done[txg & TXG_MASK]);
+ }
+
/*
* Rewrite the vdev configuration (which includes the uberblock)
* to commit the transaction group.
int children = rvd->vdev_children;
int c0 = spa_get_random(children);
- for (c = 0; c < children; c++) {
+ for (int c = 0; c < children; c++) {
vd = rvd->vdev_child[(c0 + c) % children];
- if (vd->vdev_ms_array == 0 || vd->vdev_islog)
+ if (vd->vdev_ms_array == 0 || vd->vdev_islog ||
+ !vdev_is_concrete(vd))
continue;
svd[svdcount++] = vd;
if (svdcount == SPA_DVAS_PER_BP)
if (error == 0)
break;
- zio_suspend(spa, NULL);
+ zio_suspend(spa, NULL, ZIO_SUSPEND_IOERR);
zio_resume_wait(spa);
}
dmu_tx_commit(tx);
- taskq_cancel_id(system_taskq, spa->spa_deadman_tqid);
+ taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid);
spa->spa_deadman_tqid = 0;
/*
spa->spa_config_syncing = NULL;
}
- spa->spa_ubsync = spa->spa_uberblock;
-
dsl_pool_sync_done(dp, txg);
+ mutex_enter(&spa->spa_alloc_lock);
+ VERIFY0(avl_numnodes(&spa->spa_alloc_tree));
+ mutex_exit(&spa->spa_alloc_lock);
+
/*
* Update usable space statistics.
*/
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);
return (B_FALSE);
}
+sysevent_t *
+spa_event_create(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl, const char *name)
+{
+ sysevent_t *ev = NULL;
+#ifdef _KERNEL
+ 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;
+ }
+#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));
+ }
+#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
* or zdb as real changes.
*/
void
-spa_event_notify(spa_t *spa, vdev_t *vd, const char *name)
+spa_event_notify(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl, const char *name)
{
- zfs_post_sysevent(spa, vd, name);
+ spa_event_post(spa_event_create(spa, vd, hist_nvl, name));
}
#if defined(_KERNEL) && defined(HAVE_SPL)
EXPORT_SYMBOL(spa_vdev_add);
EXPORT_SYMBOL(spa_vdev_attach);
EXPORT_SYMBOL(spa_vdev_detach);
-EXPORT_SYMBOL(spa_vdev_remove);
EXPORT_SYMBOL(spa_vdev_setpath);
EXPORT_SYMBOL(spa_vdev_setfru);
EXPORT_SYMBOL(spa_vdev_split_mirror);
MODULE_PARM_DESC(spa_load_verify_data,
"Set to traverse data on pool import");
+/* CSTYLED */
module_param(zio_taskq_batch_pct, uint, 0444);
MODULE_PARM_DESC(zio_taskq_batch_pct,
"Percentage of CPUs to run an IO worker thread");