/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Copyright 2011 Nexenta Systems, Inc. All rights reserved.
+ * Copyright (c) 2012 by Delphix. All rights reserved.
*/
#include <sys/zfs_context.h>
#include <sys/arc.h>
#include <sys/zil.h>
#include <sys/dsl_scan.h>
+#include <sys/zvol.h>
/*
* Virtual device management.
{
uint64_t asize = P2ROUNDUP(psize, 1ULL << vd->vdev_top->vdev_ashift);
uint64_t csize;
+ int c;
- for (int c = 0; c < vd->vdev_children; c++) {
+ for (c = 0; c < vd->vdev_children; c++) {
csize = vdev_psize_to_asize(vd->vdev_child[c], psize);
asize = MAX(asize, csize);
}
vdev_t *pvd = vd->vdev_parent;
/*
- * The our parent is NULL (inactive spare or cache) or is the root,
+ * If our parent is NULL (inactive spare or cache) or is the root,
* just return our own asize.
*/
if (pvd == NULL)
void
vdev_set_min_asize(vdev_t *vd)
{
+ int c;
vd->vdev_min_asize = vdev_get_min_asize(vd);
- for (int c = 0; c < vd->vdev_children; c++)
+ for (c = 0; c < vd->vdev_children; c++)
vdev_set_min_asize(vd->vdev_child[c]);
}
vdev_lookup_by_guid(vdev_t *vd, uint64_t guid)
{
vdev_t *mvd;
+ int c;
if (vd->vdev_guid == guid)
return (vd);
- for (int c = 0; c < vd->vdev_children; c++)
+ for (c = 0; c < vd->vdev_children; c++)
if ((mvd = vdev_lookup_by_guid(vd->vdev_child[c], guid)) !=
NULL)
return (mvd);
pvd->vdev_children = MAX(pvd->vdev_children, id + 1);
newsize = pvd->vdev_children * sizeof (vdev_t *);
- newchild = kmem_zalloc(newsize, KM_SLEEP);
+ newchild = kmem_zalloc(newsize, KM_PUSHPAGE);
if (pvd->vdev_child != NULL) {
bcopy(pvd->vdev_child, newchild, oldsize);
kmem_free(pvd->vdev_child, oldsize);
*/
for (; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum += cvd->vdev_guid_sum;
-
- if (cvd->vdev_ops->vdev_op_leaf)
- cvd->vdev_spa->spa_scrub_maxinflight += zfs_scrub_limit;
}
void
*/
for (; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum -= cvd->vdev_guid_sum;
-
- if (cvd->vdev_ops->vdev_op_leaf)
- cvd->vdev_spa->spa_scrub_maxinflight -= zfs_scrub_limit;
}
/*
vdev_t **newchild, *cvd;
int oldc = pvd->vdev_children;
int newc;
+ int c;
ASSERT(spa_config_held(pvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL);
- for (int c = newc = 0; c < oldc; c++)
+ for (c = newc = 0; c < oldc; c++)
if (pvd->vdev_child[c])
newc++;
- newchild = kmem_alloc(newc * sizeof (vdev_t *), KM_SLEEP);
+ newchild = kmem_alloc(newc * sizeof (vdev_t *), KM_PUSHPAGE);
- for (int c = newc = 0; c < oldc; c++) {
+ for (c = newc = 0; c < oldc; c++) {
if ((cvd = pvd->vdev_child[c]) != NULL) {
newchild[newc] = cvd;
cvd->vdev_id = newc++;
vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid, vdev_ops_t *ops)
{
vdev_t *vd;
+ int t;
- vd = kmem_zalloc(sizeof (vdev_t), KM_SLEEP);
+ vd = kmem_zalloc(sizeof (vdev_t), KM_PUSHPAGE);
if (spa->spa_root_vdev == NULL) {
ASSERT(ops == &vdev_root_ops);
spa->spa_root_vdev = vd;
+ spa->spa_load_guid = spa_generate_guid(NULL);
}
if (guid == 0 && ops != &vdev_hole_ops) {
vd->vdev_state = VDEV_STATE_CLOSED;
vd->vdev_ishole = (ops == &vdev_hole_ops);
+ list_link_init(&vd->vdev_config_dirty_node);
+ list_link_init(&vd->vdev_state_dirty_node);
mutex_init(&vd->vdev_dtl_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_stat_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_probe_lock, NULL, MUTEX_DEFAULT, NULL);
- for (int t = 0; t < DTL_TYPES; t++) {
+ for (t = 0; t < DTL_TYPES; t++) {
space_map_create(&vd->vdev_dtl[t], 0, -1ULL, 0,
&vd->vdev_dtl_lock);
}
&vd->vdev_removing);
}
- if (parent && !parent->vdev_parent) {
+ if (parent && !parent->vdev_parent && alloctype != VDEV_ALLOC_ATTACH) {
ASSERT(alloctype == VDEV_ALLOC_LOAD ||
alloctype == VDEV_ALLOC_ADD ||
alloctype == VDEV_ALLOC_SPLIT ||
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_OFFLINE,
&vd->vdev_offline);
+ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_RESILVERING,
+ &vd->vdev_resilvering);
+
/*
* When importing a pool, we want to ignore the persistent fault
* state, as the diagnosis made on another system may not be
void
vdev_free(vdev_t *vd)
{
+ int c, t;
spa_t *spa = vd->vdev_spa;
/*
/*
* Free all children.
*/
- for (int c = 0; c < vd->vdev_children; c++)
+ for (c = 0; c < vd->vdev_children; c++)
vdev_free(vd->vdev_child[c]);
ASSERT(vd->vdev_child == NULL);
txg_list_destroy(&vd->vdev_dtl_list);
mutex_enter(&vd->vdev_dtl_lock);
- for (int t = 0; t < DTL_TYPES; t++) {
+ for (t = 0; t < DTL_TYPES; t++) {
space_map_unload(&vd->vdev_dtl[t]);
space_map_destroy(&vd->vdev_dtl[t]);
}
svd->vdev_ms_shift = 0;
svd->vdev_ms_count = 0;
+ if (tvd->vdev_mg)
+ ASSERT3P(tvd->vdev_mg, ==, svd->vdev_mg);
tvd->vdev_mg = svd->vdev_mg;
tvd->vdev_ms = svd->vdev_ms;
static void
vdev_top_update(vdev_t *tvd, vdev_t *vd)
{
+ int c;
+
if (vd == NULL)
return;
vd->vdev_top = tvd;
- for (int c = 0; c < vd->vdev_children; c++)
+ for (c = 0; c < vd->vdev_children; c++)
vdev_top_update(tvd, vd->vdev_child[c]);
}
mvd->vdev_asize = cvd->vdev_asize;
mvd->vdev_min_asize = cvd->vdev_min_asize;
+ mvd->vdev_max_asize = cvd->vdev_max_asize;
mvd->vdev_ashift = cvd->vdev_ashift;
mvd->vdev_state = cvd->vdev_state;
mvd->vdev_crtxg = cvd->vdev_crtxg;
ASSERT(oldc <= newc);
- mspp = kmem_zalloc(newc * sizeof (*mspp), KM_SLEEP);
+ mspp = kmem_zalloc(newc * sizeof (*mspp), KM_PUSHPAGE | KM_NODEBUG);
if (oldc != 0) {
bcopy(vd->vdev_ms, mspp, oldc * sizeof (*mspp));
kmem_free(vd->vdev_ms, count * sizeof (metaslab_t *));
vd->vdev_ms = NULL;
}
+
+ ASSERT3U(vd->vdev_pending_fastwrite, ==, 0);
}
typedef struct vdev_probe_stats {
spa_t *spa = vd->vdev_spa;
vdev_probe_stats_t *vps = NULL;
zio_t *pio;
+ int l;
ASSERT(vd->vdev_ops->vdev_op_leaf);
mutex_enter(&vd->vdev_probe_lock);
if ((pio = vd->vdev_probe_zio) == NULL) {
- vps = kmem_zalloc(sizeof (*vps), KM_SLEEP);
+ vps = kmem_zalloc(sizeof (*vps), KM_PUSHPAGE);
vps->vps_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_PROBE |
ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_AGGREGATE |
return (NULL);
}
- for (int l = 1; l < VDEV_LABELS; l++) {
+ for (l = 1; l < VDEV_LABELS; l++) {
zio_nowait(zio_read_phys(pio, vd,
vdev_label_offset(vd->vdev_psize, l,
offsetof(vdev_label_t, vl_pad2)),
vd->vdev_open_thread = NULL;
}
-boolean_t
+static boolean_t
vdev_uses_zvols(vdev_t *vd)
{
- if (vd->vdev_path && strncmp(vd->vdev_path, ZVOL_DIR,
- strlen(ZVOL_DIR)) == 0)
+ int c;
+
+#ifdef _KERNEL
+ if (zvol_is_zvol(vd->vdev_path))
return (B_TRUE);
- for (int c = 0; c < vd->vdev_children; c++)
+#endif
+
+ for (c = 0; c < vd->vdev_children; c++)
if (vdev_uses_zvols(vd->vdev_child[c]))
return (B_TRUE);
+
return (B_FALSE);
}
{
taskq_t *tq;
int children = vd->vdev_children;
+ int c;
/*
* in order to handle pools on top of zvols, do the opens
* spa_namespace_lock
*/
if (vdev_uses_zvols(vd)) {
- for (int c = 0; c < children; c++)
+ for (c = 0; c < children; c++)
vd->vdev_child[c]->vdev_open_error =
vdev_open(vd->vdev_child[c]);
return;
tq = taskq_create("vdev_open", children, minclsyspri,
children, children, TASKQ_PREPOPULATE);
- for (int c = 0; c < children; c++)
+ for (c = 0; c < children; c++)
VERIFY(taskq_dispatch(tq, vdev_open_child, vd->vdev_child[c],
- TQ_SLEEP) != NULL);
+ TQ_SLEEP) != 0);
taskq_destroy(tq);
}
spa_t *spa = vd->vdev_spa;
int error;
uint64_t osize = 0;
- uint64_t asize, psize;
+ uint64_t max_osize = 0;
+ uint64_t asize, max_asize, psize;
uint64_t ashift = 0;
+ int c;
ASSERT(vd->vdev_open_thread == curthread ||
spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
return (ENXIO);
}
- error = vd->vdev_ops->vdev_op_open(vd, &osize, &ashift);
+ error = vd->vdev_ops->vdev_op_open(vd, &osize, &max_osize, &ashift);
/*
* Reset the vdev_reopening flag so that we actually close
if (vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops)
return (0);
- for (int c = 0; c < vd->vdev_children; c++) {
+ for (c = 0; c < vd->vdev_children; c++) {
if (vd->vdev_child[c]->vdev_state != VDEV_STATE_HEALTHY) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED,
VDEV_AUX_NONE);
}
osize = P2ALIGN(osize, (uint64_t)sizeof (vdev_label_t));
+ max_osize = P2ALIGN(max_osize, (uint64_t)sizeof (vdev_label_t));
if (vd->vdev_children == 0) {
if (osize < SPA_MINDEVSIZE) {
}
psize = osize;
asize = osize - (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE);
+ max_asize = max_osize - (VDEV_LABEL_START_SIZE +
+ VDEV_LABEL_END_SIZE);
} else {
if (vd->vdev_parent != NULL && osize < SPA_MINDEVSIZE -
(VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE)) {
}
psize = 0;
asize = osize;
+ max_asize = max_osize;
}
vd->vdev_psize = psize;
if (vd->vdev_asize == 0) {
/*
* This is the first-ever open, so use the computed values.
- * For testing purposes, a higher ashift can be requested.
+ * For compatibility, a different ashift can be requested.
*/
vd->vdev_asize = asize;
- vd->vdev_ashift = MAX(ashift, vd->vdev_ashift);
+ vd->vdev_max_asize = max_asize;
+ if (vd->vdev_ashift == 0)
+ vd->vdev_ashift = ashift;
} else {
/*
- * Make sure the alignment requirement hasn't increased.
+ * Detect if the alignment requirement has increased.
+ * We don't want to make the pool unavailable, just
+ * post an event instead.
*/
- if (ashift > vd->vdev_top->vdev_ashift) {
- vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
- VDEV_AUX_BAD_LABEL);
- return (EINVAL);
+ if (ashift > vd->vdev_top->vdev_ashift &&
+ vd->vdev_ops->vdev_op_leaf) {
+ zfs_ereport_post(FM_EREPORT_ZFS_DEVICE_BAD_ASHIFT,
+ spa, vd, NULL, 0, 0);
}
+
+ vd->vdev_max_asize = max_asize;
}
/*
* contents. This needs to be done before vdev_load() so that we don't
* inadvertently do repair I/Os to the wrong device.
*
+ * If 'strict' is false ignore the spa guid check. This is necessary because
+ * if the machine crashed during a re-guid the new guid might have been written
+ * to all of the vdev labels, but not the cached config. The strict check
+ * will be performed when the pool is opened again using the mos config.
+ *
* This function will only return failure if one of the vdevs indicates that it
* has since been destroyed or exported. This is only possible if
* /etc/zfs/zpool.cache was readonly at the time. Otherwise, the vdev state
* will be updated but the function will return 0.
*/
int
-vdev_validate(vdev_t *vd)
+vdev_validate(vdev_t *vd, boolean_t strict)
{
spa_t *spa = vd->vdev_spa;
nvlist_t *label;
uint64_t guid = 0, top_guid;
uint64_t state;
+ int c;
- for (int c = 0; c < vd->vdev_children; c++)
- if (vdev_validate(vd->vdev_child[c]) != 0)
+ for (c = 0; c < vd->vdev_children; c++)
+ if (vdev_validate(vd->vdev_child[c], strict) != 0)
return (EBADF);
/*
if (vd->vdev_ops->vdev_op_leaf && vdev_readable(vd)) {
uint64_t aux_guid = 0;
nvlist_t *nvl;
+ uint64_t txg = spa_last_synced_txg(spa) != 0 ?
+ spa_last_synced_txg(spa) : -1ULL;
- if ((label = vdev_label_read_config(vd)) == NULL) {
+ if ((label = vdev_label_read_config(vd, txg)) == NULL) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_BAD_LABEL);
return (0);
return (0);
}
- if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID,
- &guid) != 0 || guid != spa_guid(spa)) {
+ if (strict && (nvlist_lookup_uint64(label,
+ ZPOOL_CONFIG_POOL_GUID, &guid) != 0 ||
+ guid != spa_guid(spa))) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
nvlist_free(label);
/*
- * If spa->spa_load_verbatim is true, no need to check the
+ * If this is a verbatim import, no need to check the
* state of the pool.
*/
- if (!spa->spa_load_verbatim &&
+ if (!(spa->spa_import_flags & ZFS_IMPORT_VERBATIM) &&
spa_load_state(spa) == SPA_LOAD_OPEN &&
state != POOL_STATE_ACTIVE)
return (EBADF);
void
vdev_close(vdev_t *vd)
{
- spa_t *spa = vd->vdev_spa;
vdev_t *pvd = vd->vdev_parent;
+ ASSERTV(spa_t *spa = vd->vdev_spa);
ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
vdev_hold(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
+ int c;
ASSERT(spa_is_root(spa));
if (spa->spa_state == POOL_STATE_UNINITIALIZED)
return;
- for (int c = 0; c < vd->vdev_children; c++)
+ for (c = 0; c < vd->vdev_children; c++)
vdev_hold(vd->vdev_child[c]);
if (vd->vdev_ops->vdev_op_leaf)
void
vdev_rele(vdev_t *vd)
{
- spa_t *spa = vd->vdev_spa;
+ int c;
- ASSERT(spa_is_root(spa));
- for (int c = 0; c < vd->vdev_children; c++)
+ ASSERT(spa_is_root(vd->vdev_spa));
+ for (c = 0; c < vd->vdev_children; c++)
vdev_rele(vd->vdev_child[c]);
if (vd->vdev_ops->vdev_op_leaf)
!l2arc_vdev_present(vd))
l2arc_add_vdev(spa, vd);
} else {
- (void) vdev_validate(vd);
+ (void) vdev_validate(vd, B_TRUE);
}
/*
ASSERT(vd == vd->vdev_top);
ASSERT(!vd->vdev_ishole);
ASSERT(ISP2(flags));
+ ASSERT(spa_writeable(vd->vdev_spa));
if (flags & VDD_METASLAB)
(void) txg_list_add(&vd->vdev_ms_list, arg, txg);
ASSERT(t < DTL_TYPES);
ASSERT(vd != vd->vdev_spa->spa_root_vdev);
+ ASSERT(spa_writeable(vd->vdev_spa));
mutex_enter(sm->sm_lock);
if (!space_map_contains(sm, txg, size))
{
spa_t *spa = vd->vdev_spa;
avl_tree_t reftree;
- int minref;
+ int c, t, minref;
ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0);
- for (int c = 0; c < vd->vdev_children; c++)
+ for (c = 0; c < vd->vdev_children; c++)
vdev_dtl_reassess(vd->vdev_child[c], txg,
scrub_txg, scrub_done);
}
mutex_enter(&vd->vdev_dtl_lock);
- for (int t = 0; t < DTL_TYPES; t++) {
+ for (t = 0; t < DTL_TYPES; t++) {
/* account for child's outage in parent's missing map */
int s = (t == DTL_MISSING) ? DTL_OUTAGE: t;
if (t == DTL_SCRUB)
else
minref = vd->vdev_children; /* any kind of mirror */
space_map_ref_create(&reftree);
- for (int c = 0; c < vd->vdev_children; c++) {
+ for (c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
mutex_enter(&cvd->vdev_dtl_lock);
space_map_ref_add_map(&reftree, &cvd->vdev_dtl[s], 1);
if (vd->vdev_detached) {
if (smo->smo_object != 0) {
- int err = dmu_object_free(mos, smo->smo_object, tx);
- ASSERT3U(err, ==, 0);
+ VERIFY(0 == dmu_object_free(mos, smo->smo_object, tx));
smo->smo_object = 0;
}
dmu_tx_commit(tx);
vd->vdev_cant_read = cant_read;
vdev_dtl_reassess(tvd, 0, 0, B_FALSE);
+ if (!required && zio_injection_enabled)
+ required = !!zio_handle_device_injection(vd, NULL, ECHILD);
+
return (required);
}
boolean_t needed = B_FALSE;
uint64_t thismin = UINT64_MAX;
uint64_t thismax = 0;
+ int c;
if (vd->vdev_children == 0) {
mutex_enter(&vd->vdev_dtl_lock);
}
mutex_exit(&vd->vdev_dtl_lock);
} else {
- for (int c = 0; c < vd->vdev_children; c++) {
+ for (c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
uint64_t cmin, cmax;
void
vdev_load(vdev_t *vd)
{
+ int c;
+
/*
* Recursively load all children.
*/
- for (int c = 0; c < vd->vdev_children; c++)
+ for (c = 0; c < vd->vdev_children; c++)
vdev_load(vd->vdev_child[c]);
/*
if (!vdev_readable(vd))
return (0);
- if ((label = vdev_label_read_config(vd)) == NULL) {
+ if ((label = vdev_label_read_config(vd, -1ULL)) == NULL) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
return (-1);
}
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_VERSION, &version) != 0 ||
- version > SPA_VERSION ||
+ !SPA_VERSION_IS_SUPPORTED(version) ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) != 0 ||
guid != vd->vdev_guid ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE, &state) != 0) {
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa->spa_meta_objset;
dmu_tx_t *tx;
+ int m;
tx = dmu_tx_create_assigned(spa_get_dsl(spa), txg);
}
if (vd->vdev_ms != NULL) {
- for (int m = 0; m < vd->vdev_ms_count; m++) {
+ for (m = 0; m < vd->vdev_ms_count; m++) {
metaslab_t *msp = vd->vdev_ms[m];
if (msp == NULL || msp->ms_smo.smo_object == 0)
ASSERT(!vd->vdev_ishole);
- while (msp = txg_list_remove(&vd->vdev_ms_list, TXG_CLEAN(txg)))
+ while ((msp = txg_list_remove(&vd->vdev_ms_list, TXG_CLEAN(txg))))
metaslab_sync_done(msp, txg);
if (reassess)
int
vdev_fault(spa_t *spa, uint64_t guid, vdev_aux_t aux)
{
- vdev_t *vd;
+ vdev_t *vd, *tvd;
spa_vdev_state_enter(spa, SCL_NONE);
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_state_exit(spa, NULL, ENOTSUP));
+ tvd = vd->vdev_top;
+
/*
* We don't directly use the aux state here, but if we do a
* vdev_reopen(), we need this value to be present to remember why we
* If this device has the only valid copy of the data, then
* back off and simply mark the vdev as degraded instead.
*/
- if (!vd->vdev_islog && vd->vdev_aux == NULL && vdev_dtl_required(vd)) {
+ if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_dtl_required(vd)) {
vd->vdev_degraded = 1ULL;
vd->vdev_faulted = 0ULL;
* If we reopen the device and it's not dead, only then do we
* mark it degraded.
*/
- vdev_reopen(vd);
+ vdev_reopen(tvd);
if (vdev_readable(vd))
vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, aux);
vdev_clear(spa_t *spa, vdev_t *vd)
{
vdev_t *rvd = spa->spa_root_vdev;
+ int c;
ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
vd->vdev_stat.vs_write_errors = 0;
vd->vdev_stat.vs_checksum_errors = 0;
- for (int c = 0; c < vd->vdev_children; c++)
+ for (c = 0; c < vd->vdev_children; c++)
vdev_clear(spa, vd->vdev_child[c]);
/*
*/
vd->vdev_forcefault = B_TRUE;
- vd->vdev_faulted = vd->vdev_degraded = 0;
+ vd->vdev_faulted = vd->vdev_degraded = 0ULL;
vd->vdev_cant_read = B_FALSE;
vd->vdev_cant_write = B_FALSE;
- vdev_reopen(vd);
+ vdev_reopen(vd == rvd ? rvd : vd->vdev_top);
vd->vdev_forcefault = B_FALSE;
- if (vd != rvd)
+ if (vd != rvd && vdev_writeable(vd->vdev_top))
vdev_state_dirty(vd->vdev_top);
if (vd->vdev_aux == NULL && !vdev_is_dead(vd))
spa_async_request(spa, SPA_ASYNC_RESILVER);
- spa_event_notify(spa, vd, ESC_ZFS_VDEV_CLEAR);
+ spa_event_notify(spa, vd, FM_EREPORT_ZFS_DEVICE_CLEAR);
}
/*
vdev_get_stats(vdev_t *vd, vdev_stat_t *vs)
{
vdev_t *rvd = vd->vdev_spa->spa_root_vdev;
+ int c, t;
mutex_enter(&vd->vdev_stat_lock);
bcopy(&vd->vdev_stat, vs, sizeof (*vs));
vs->vs_rsize = vdev_get_min_asize(vd);
if (vd->vdev_ops->vdev_op_leaf)
vs->vs_rsize += VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE;
+ vs->vs_esize = vd->vdev_max_asize - vd->vdev_asize;
mutex_exit(&vd->vdev_stat_lock);
/*
* over all top-level vdevs (i.e. the direct children of the root).
*/
if (vd == rvd) {
- for (int c = 0; c < rvd->vdev_children; c++) {
+ for (c = 0; c < rvd->vdev_children; c++) {
vdev_t *cvd = rvd->vdev_child[c];
vdev_stat_t *cvs = &cvd->vdev_stat;
mutex_enter(&vd->vdev_stat_lock);
- for (int t = 0; t < ZIO_TYPES; t++) {
+ for (t = 0; t < ZIO_TYPES; t++) {
vs->vs_ops[t] += cvs->vs_ops[t];
vs->vs_bytes[t] += cvs->vs_bytes[t];
}
vdev_scan_stat_init(vdev_t *vd)
{
vdev_stat_t *vs = &vd->vdev_stat;
+ int c;
- for (int c = 0; c < vd->vdev_children; c++)
+ for (c = 0; c < vd->vdev_children; c++)
vdev_scan_stat_init(vd->vdev_child[c]);
mutex_enter(&vd->vdev_stat_lock);
mutex_enter(&vd->vdev_stat_lock);
if (flags & ZIO_FLAG_IO_REPAIR) {
- if (flags & ZIO_FLAG_SCRUB_THREAD) {
+ if (flags & ZIO_FLAG_SCAN_THREAD) {
dsl_scan_phys_t *scn_phys =
&spa->spa_dsl_pool->dp_scan->scn_phys;
uint64_t *processed = &scn_phys->scn_processed;
if (type == ZIO_TYPE_WRITE && txg != 0 &&
(!(flags & ZIO_FLAG_IO_REPAIR) ||
- (flags & ZIO_FLAG_SCRUB_THREAD) ||
+ (flags & ZIO_FLAG_SCAN_THREAD) ||
spa->spa_claiming)) {
/*
* This is either a normal write (not a repair), or it's
*/
if (vd->vdev_ops->vdev_op_leaf) {
uint64_t commit_txg = txg;
- if (flags & ZIO_FLAG_SCRUB_THREAD) {
+ if (flags & ZIO_FLAG_SCAN_THREAD) {
ASSERT(flags & ZIO_FLAG_IO_REPAIR);
ASSERT(spa_sync_pass(spa) == 1);
vdev_dtl_dirty(vd, DTL_SCRUB, txg, 1);
vdev_t *rvd = spa->spa_root_vdev;
int c;
+ ASSERT(spa_writeable(spa));
+
/*
* If this is an aux vdev (as with l2cache and spare devices), then we
* update the vdev config manually and set the sync flag.
{
spa_t *spa = vd->vdev_spa;
+ ASSERT(spa_writeable(spa));
ASSERT(vd == vd->vdev_top);
/*
int degraded = 0, faulted = 0;
int corrupted = 0;
vdev_t *child;
+ int c;
if (vd->vdev_children > 0) {
- for (int c = 0; c < vd->vdev_children; c++) {
+ for (c = 0; c < vd->vdev_children; c++) {
child = vd->vdev_child[c];
/*
vd->vdev_removed = B_TRUE;
} else if (state == VDEV_STATE_CANT_OPEN) {
/*
- * If we fail to open a vdev during an import, we mark it as
- * "not available", which signifies that it was never there to
- * begin with. Failure to open such a device is not considered
- * an error.
+ * If we fail to open a vdev during an import or recovery, we
+ * mark it as "not available", which signifies that it was
+ * never there to begin with. Failure to open such a device
+ * is not considered an error.
*/
- if (spa_load_state(spa) == SPA_LOAD_IMPORT &&
+ if ((spa_load_state(spa) == SPA_LOAD_IMPORT ||
+ spa_load_state(spa) == SPA_LOAD_RECOVER) &&
vd->vdev_ops->vdev_op_leaf)
vd->vdev_not_present = 1;
/*
* Check the vdev configuration to ensure that it's capable of supporting
- * a root pool. Currently, we do not support RAID-Z or partial configuration.
- * In addition, only a single top-level vdev is allowed and none of the leaves
- * can be wholedisks.
+ * a root pool.
*/
boolean_t
vdev_is_bootable(vdev_t *vd)
{
+#if defined(__sun__) || defined(__sun)
+ /*
+ * Currently, we do not support RAID-Z or partial configuration.
+ * In addition, only a single top-level vdev is allowed and none of the
+ * leaves can be wholedisks.
+ */
+ int c;
+
if (!vd->vdev_ops->vdev_op_leaf) {
char *vdev_type = vd->vdev_ops->vdev_op_type;
return (B_FALSE);
}
- for (int c = 0; c < vd->vdev_children; c++) {
+ for (c = 0; c < vd->vdev_children; c++) {
if (!vdev_is_bootable(vd->vdev_child[c]))
return (B_FALSE);
}
+#endif /* __sun__ || __sun */
return (B_TRUE);
}
/*
* Load the state from the original vdev tree (ovd) which
* we've retrieved from the MOS config object. If the original
- * vdev was offline then we transfer that state to the device
- * in the current vdev tree (nvd).
+ * vdev was offline or faulted then we transfer that state to the
+ * device in the current vdev tree (nvd).
*/
void
vdev_load_log_state(vdev_t *nvd, vdev_t *ovd)
{
- spa_t *spa = nvd->vdev_spa;
+ int c;
- ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
+ ASSERT(nvd->vdev_top->vdev_islog);
+ ASSERT(spa_config_held(nvd->vdev_spa,
+ SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
ASSERT3U(nvd->vdev_guid, ==, ovd->vdev_guid);
- for (int c = 0; c < nvd->vdev_children; c++)
+ for (c = 0; c < nvd->vdev_children; c++)
vdev_load_log_state(nvd->vdev_child[c], ovd->vdev_child[c]);
- if (nvd->vdev_ops->vdev_op_leaf && ovd->vdev_offline) {
+ if (nvd->vdev_ops->vdev_op_leaf) {
/*
- * It would be nice to call vdev_offline()
- * directly but the pool isn't fully loaded and
- * the txg threads have not been started yet.
+ * Restore the persistent vdev state
*/
nvd->vdev_offline = ovd->vdev_offline;
- vdev_reopen(nvd->vdev_top);
+ nvd->vdev_faulted = ovd->vdev_faulted;
+ nvd->vdev_degraded = ovd->vdev_degraded;
+ nvd->vdev_removed = ovd->vdev_removed;
}
}
/*
+ * Determine if a log device has valid content. If the vdev was
+ * removed or faulted in the MOS config then we know that
+ * the content on the log device has already been written to the pool.
+ */
+boolean_t
+vdev_log_state_valid(vdev_t *vd)
+{
+ int c;
+
+ if (vd->vdev_ops->vdev_op_leaf && !vd->vdev_faulted &&
+ !vd->vdev_removed)
+ return (B_TRUE);
+
+ for (c = 0; c < vd->vdev_children; c++)
+ if (vdev_log_state_valid(vd->vdev_child[c]))
+ return (B_TRUE);
+
+ return (B_FALSE);
+}
+
+/*
* Expand a vdev if possible.
*/
void
}
vdev_propagate_state(cvd);
}
+
+void
+vdev_deadman(vdev_t *vd)
+{
+ int c;
+
+ for (c = 0; c < vd->vdev_children; c++) {
+ vdev_t *cvd = vd->vdev_child[c];
+
+ vdev_deadman(cvd);
+ }
+
+ if (vd->vdev_ops->vdev_op_leaf) {
+ vdev_queue_t *vq = &vd->vdev_queue;
+
+ mutex_enter(&vq->vq_lock);
+ if (avl_numnodes(&vq->vq_pending_tree) > 0) {
+ spa_t *spa = vd->vdev_spa;
+ zio_t *fio;
+ uint64_t delta;
+
+ /*
+ * Look at the head of all the pending queues,
+ * if any I/O has been outstanding for longer than
+ * the spa_deadman_synctime we log a zevent.
+ */
+ fio = avl_first(&vq->vq_pending_tree);
+ delta = ddi_get_lbolt64() - fio->io_timestamp;
+ if (delta > NSEC_TO_TICK(spa_deadman_synctime(spa))) {
+ zfs_dbgmsg("SLOW IO: zio timestamp %llu, "
+ "delta %llu, last io %llu",
+ fio->io_timestamp, delta,
+ vq->vq_io_complete_ts);
+ zfs_ereport_post(FM_EREPORT_ZFS_DELAY,
+ spa, vd, fio, 0, 0);
+ }
+ }
+ mutex_exit(&vq->vq_lock);
+ }
+}
+
+#if defined(_KERNEL) && defined(HAVE_SPL)
+EXPORT_SYMBOL(vdev_fault);
+EXPORT_SYMBOL(vdev_degrade);
+EXPORT_SYMBOL(vdev_online);
+EXPORT_SYMBOL(vdev_offline);
+EXPORT_SYMBOL(vdev_clear);
+
+module_param(zfs_scrub_limit, int, 0644);
+MODULE_PARM_DESC(zfs_scrub_limit, "Max scrub/resilver I/O per leaf vdev");
+#endif
git://git.camperquake.de / zfs.git / blobdiff