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);
- space_map_create(&vd->vdev_dtl_map, 0, -1ULL, 0, &vd->vdev_dtl_lock);
- space_map_create(&vd->vdev_dtl_scrub, 0, -1ULL, 0, &vd->vdev_dtl_lock);
+ for (int t = 0; t < DTL_TYPES; t++) {
+ space_map_create(&vd->vdev_dtl[t], 0, -1ULL, 0,
+ &vd->vdev_dtl_lock);
+ }
txg_list_create(&vd->vdev_ms_list,
offsetof(struct metaslab, ms_txg_node));
txg_list_create(&vd->vdev_dtl_list,
(alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_L2CACHE)) {
if (alloctype == VDEV_ALLOC_LOAD) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DTL,
- &vd->vdev_dtl.smo_object);
+ &vd->vdev_dtl_smo.smo_object);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_UNSPARE,
&vd->vdev_unspare);
}
txg_list_destroy(&vd->vdev_ms_list);
txg_list_destroy(&vd->vdev_dtl_list);
+
mutex_enter(&vd->vdev_dtl_lock);
- space_map_unload(&vd->vdev_dtl_map);
- space_map_destroy(&vd->vdev_dtl_map);
- space_map_vacate(&vd->vdev_dtl_scrub, NULL, NULL);
- space_map_destroy(&vd->vdev_dtl_scrub);
+ for (int t = 0; t < DTL_TYPES; t++) {
+ space_map_unload(&vd->vdev_dtl[t]);
+ space_map_destroy(&vd->vdev_dtl[t]);
+ }
mutex_exit(&vd->vdev_dtl_lock);
+
mutex_destroy(&vd->vdev_dtl_lock);
mutex_destroy(&vd->vdev_stat_lock);
mutex_destroy(&vd->vdev_probe_lock);
vdev_remove_child(mvd, cvd);
vdev_remove_child(pvd, mvd);
+
/*
* If cvd will replace mvd as a top-level vdev, preserve mvd's guid.
* Otherwise, we could have detached an offline device, and when we
* go to import the pool we'll think we have two top-level vdevs,
* instead of a different version of the same top-level vdev.
*/
- if (mvd->vdev_top == mvd)
- cvd->vdev_guid = cvd->vdev_guid_sum = mvd->vdev_guid;
+ if (mvd->vdev_top == mvd) {
+ uint64_t guid_delta = mvd->vdev_guid - cvd->vdev_guid;
+ cvd->vdev_guid += guid_delta;
+ cvd->vdev_guid_sum += guid_delta;
+ }
cvd->vdev_id = mvd->vdev_id;
vdev_add_child(pvd, cvd);
vdev_top_update(cvd->vdev_top, cvd->vdev_top);
static void
vdev_probe_done(zio_t *zio)
{
+ spa_t *spa = zio->io_spa;
vdev_probe_stats_t *vps = zio->io_private;
vdev_t *vd = vps->vps_vd;
ASSERT(zio->io_vd == vd);
if (zio->io_error == 0)
vps->vps_readable = 1;
- if (zio->io_error == 0 && (spa_mode & FWRITE)) {
+ if (zio->io_error == 0 && spa_writeable(spa)) {
zio_nowait(zio_write_phys(vps->vps_root, vd,
zio->io_offset, zio->io_size, zio->io_data,
ZIO_CHECKSUM_OFF, vdev_probe_done, vps,
vd->vdev_cant_write |= !vps->vps_writeable;
if (vdev_readable(vd) &&
- (vdev_writeable(vd) || !(spa_mode & FWRITE))) {
+ (vdev_writeable(vd) || !spa_writeable(spa))) {
zio->io_error = 0;
} else {
ASSERT(zio->io_error != 0);
zfs_ereport_post(FM_EREPORT_ZFS_PROBE_FAILURE,
- zio->io_spa, vd, NULL, 0, 0);
+ spa, vd, NULL, 0, 0);
zio->io_error = ENXIO;
}
kmem_free(vps, sizeof (*vps));
int
vdev_open(vdev_t *vd)
{
+ spa_t *spa = vd->vdev_spa;
int error;
int c;
uint64_t osize = 0;
uint64_t asize, psize;
uint64_t ashift = 0;
+ ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
+
ASSERT(vd->vdev_state == VDEV_STATE_CLOSED ||
vd->vdev_state == VDEV_STATE_CANT_OPEN ||
vd->vdev_state == VDEV_STATE_OFFLINE);
/*
* If a leaf vdev has a DTL, and seems healthy, then kick off a
- * resilver. But don't do this if we are doing a reopen for a
- * scrub, since this would just restart the scrub we are already
- * doing.
+ * resilver. But don't do this if we are doing a reopen for a scrub,
+ * since this would just restart the scrub we are already doing.
*/
- if (vd->vdev_children == 0 && !vd->vdev_spa->spa_scrub_reopen) {
- mutex_enter(&vd->vdev_dtl_lock);
- if (vd->vdev_dtl_map.sm_space != 0 && vdev_writeable(vd))
- spa_async_request(vd->vdev_spa, SPA_ASYNC_RESILVER);
- mutex_exit(&vd->vdev_dtl_lock);
- }
+ if (vd->vdev_ops->vdev_op_leaf && !spa->spa_scrub_reopen &&
+ vdev_resilver_needed(vd, NULL, NULL))
+ spa_async_request(spa, SPA_ASYNC_RESILVER);
return (0);
}
void
vdev_close(vdev_t *vd)
{
+ spa_t *spa = vd->vdev_spa;
+
+ ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
+
vd->vdev_ops->vdev_op_close(vd);
vdev_cache_purge(vd);
(void) txg_list_add(&vd->vdev_spa->spa_vdev_txg_list, vd, txg);
}
+/*
+ * DTLs.
+ *
+ * A vdev's DTL (dirty time log) is the set of transaction groups for which
+ * the vdev has less than perfect replication. There are three kinds of DTL:
+ *
+ * DTL_MISSING: txgs for which the vdev has no valid copies of the data
+ *
+ * DTL_PARTIAL: txgs for which data is available, but not fully replicated
+ *
+ * DTL_SCRUB: the txgs that could not be repaired by the last scrub; upon
+ * scrub completion, DTL_SCRUB replaces DTL_MISSING in the range of
+ * txgs that was scrubbed.
+ *
+ * DTL_OUTAGE: txgs which cannot currently be read, whether due to
+ * persistent errors or just some device being offline.
+ * Unlike the other three, the DTL_OUTAGE map is not generally
+ * maintained; it's only computed when needed, typically to
+ * determine whether a device can be detached.
+ *
+ * For leaf vdevs, DTL_MISSING and DTL_PARTIAL are identical: the device
+ * either has the data or it doesn't.
+ *
+ * For interior vdevs such as mirror and RAID-Z the picture is more complex.
+ * A vdev's DTL_PARTIAL is the union of its children's DTL_PARTIALs, because
+ * if any child is less than fully replicated, then so is its parent.
+ * A vdev's DTL_MISSING is a modified union of its children's DTL_MISSINGs,
+ * comprising only those txgs which appear in 'maxfaults' or more children;
+ * those are the txgs we don't have enough replication to read. For example,
+ * double-parity RAID-Z can tolerate up to two missing devices (maxfaults == 2);
+ * thus, its DTL_MISSING consists of the set of txgs that appear in more than
+ * two child DTL_MISSING maps.
+ *
+ * It should be clear from the above that to compute the DTLs and outage maps
+ * for all vdevs, it suffices to know just the leaf vdevs' DTL_MISSING maps.
+ * Therefore, that is all we keep on disk. When loading the pool, or after
+ * a configuration change, we generate all other DTLs from first principles.
+ */
void
-vdev_dtl_dirty(space_map_t *sm, uint64_t txg, uint64_t size)
+vdev_dtl_dirty(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size)
{
+ space_map_t *sm = &vd->vdev_dtl[t];
+
+ ASSERT(t < DTL_TYPES);
+ ASSERT(vd != vd->vdev_spa->spa_root_vdev);
+
mutex_enter(sm->sm_lock);
if (!space_map_contains(sm, txg, size))
space_map_add(sm, txg, size);
mutex_exit(sm->sm_lock);
}
-int
-vdev_dtl_contains(space_map_t *sm, uint64_t txg, uint64_t size)
+boolean_t
+vdev_dtl_contains(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size)
{
- int dirty;
+ space_map_t *sm = &vd->vdev_dtl[t];
+ boolean_t dirty = B_FALSE;
- /*
- * Quick test without the lock -- covers the common case that
- * there are no dirty time segments.
- */
- if (sm->sm_space == 0)
- return (0);
+ ASSERT(t < DTL_TYPES);
+ ASSERT(vd != vd->vdev_spa->spa_root_vdev);
mutex_enter(sm->sm_lock);
- dirty = space_map_contains(sm, txg, size);
+ if (sm->sm_space != 0)
+ dirty = space_map_contains(sm, txg, size);
mutex_exit(sm->sm_lock);
return (dirty);
}
+boolean_t
+vdev_dtl_empty(vdev_t *vd, vdev_dtl_type_t t)
+{
+ space_map_t *sm = &vd->vdev_dtl[t];
+ boolean_t empty;
+
+ mutex_enter(sm->sm_lock);
+ empty = (sm->sm_space == 0);
+ mutex_exit(sm->sm_lock);
+
+ return (empty);
+}
+
/*
* Reassess DTLs after a config change or scrub completion.
*/
vdev_dtl_reassess(vdev_t *vd, uint64_t txg, uint64_t scrub_txg, int scrub_done)
{
spa_t *spa = vd->vdev_spa;
- int c;
+ avl_tree_t reftree;
+ int minref;
- ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER));
+ ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0);
- if (vd->vdev_children == 0) {
+ for (int c = 0; c < vd->vdev_children; c++)
+ vdev_dtl_reassess(vd->vdev_child[c], txg,
+ scrub_txg, scrub_done);
+
+ if (vd == spa->spa_root_vdev)
+ return;
+
+ if (vd->vdev_ops->vdev_op_leaf) {
mutex_enter(&vd->vdev_dtl_lock);
if (scrub_txg != 0 &&
(spa->spa_scrub_started || spa->spa_scrub_errors == 0)) {
* will be valid, so excise the old region and
* fold in the scrub dtl. Otherwise, leave the
* dtl as-is if there was an error.
+ *
+ * There's little trick here: to excise the beginning
+ * of the DTL_MISSING map, we put it into a reference
+ * tree and then add a segment with refcnt -1 that
+ * covers the range [0, scrub_txg). This means
+ * that each txg in that range has refcnt -1 or 0.
+ * We then add DTL_SCRUB with a refcnt of 2, so that
+ * entries in the range [0, scrub_txg) will have a
+ * positive refcnt -- either 1 or 2. We then convert
+ * the reference tree into the new DTL_MISSING map.
*/
- space_map_excise(&vd->vdev_dtl_map, 0, scrub_txg);
- space_map_union(&vd->vdev_dtl_map, &vd->vdev_dtl_scrub);
+ space_map_ref_create(&reftree);
+ space_map_ref_add_map(&reftree,
+ &vd->vdev_dtl[DTL_MISSING], 1);
+ space_map_ref_add_seg(&reftree, 0, scrub_txg, -1);
+ space_map_ref_add_map(&reftree,
+ &vd->vdev_dtl[DTL_SCRUB], 2);
+ space_map_ref_generate_map(&reftree,
+ &vd->vdev_dtl[DTL_MISSING], 1);
+ space_map_ref_destroy(&reftree);
}
+ space_map_vacate(&vd->vdev_dtl[DTL_PARTIAL], NULL, NULL);
+ space_map_walk(&vd->vdev_dtl[DTL_MISSING],
+ space_map_add, &vd->vdev_dtl[DTL_PARTIAL]);
if (scrub_done)
- space_map_vacate(&vd->vdev_dtl_scrub, NULL, NULL);
+ space_map_vacate(&vd->vdev_dtl[DTL_SCRUB], NULL, NULL);
+ space_map_vacate(&vd->vdev_dtl[DTL_OUTAGE], NULL, NULL);
+ if (!vdev_readable(vd))
+ space_map_add(&vd->vdev_dtl[DTL_OUTAGE], 0, -1ULL);
+ else
+ space_map_walk(&vd->vdev_dtl[DTL_MISSING],
+ space_map_add, &vd->vdev_dtl[DTL_OUTAGE]);
mutex_exit(&vd->vdev_dtl_lock);
if (txg != 0)
return;
}
- /*
- * Make sure the DTLs are always correct under the scrub lock.
- */
- if (vd == spa->spa_root_vdev)
- mutex_enter(&spa->spa_scrub_lock);
-
mutex_enter(&vd->vdev_dtl_lock);
- space_map_vacate(&vd->vdev_dtl_map, NULL, NULL);
- space_map_vacate(&vd->vdev_dtl_scrub, NULL, NULL);
- mutex_exit(&vd->vdev_dtl_lock);
-
- for (c = 0; c < vd->vdev_children; c++) {
- vdev_t *cvd = vd->vdev_child[c];
- vdev_dtl_reassess(cvd, txg, scrub_txg, scrub_done);
- mutex_enter(&vd->vdev_dtl_lock);
- space_map_union(&vd->vdev_dtl_map, &cvd->vdev_dtl_map);
- space_map_union(&vd->vdev_dtl_scrub, &cvd->vdev_dtl_scrub);
- mutex_exit(&vd->vdev_dtl_lock);
+ for (int t = 0; t < DTL_TYPES; t++) {
+ if (t == DTL_SCRUB)
+ continue; /* leaf vdevs only */
+ if (t == DTL_PARTIAL)
+ minref = 1; /* i.e. non-zero */
+ else if (vd->vdev_nparity != 0)
+ minref = vd->vdev_nparity + 1; /* RAID-Z */
+ else
+ minref = vd->vdev_children; /* any kind of mirror */
+ space_map_ref_create(&reftree);
+ for (int c = 0; c < vd->vdev_children; c++) {
+ vdev_t *cvd = vd->vdev_child[c];
+ mutex_enter(&cvd->vdev_dtl_lock);
+ space_map_ref_add_map(&reftree, &cvd->vdev_dtl[t], 1);
+ mutex_exit(&cvd->vdev_dtl_lock);
+ }
+ space_map_ref_generate_map(&reftree, &vd->vdev_dtl[t], minref);
+ space_map_ref_destroy(&reftree);
}
-
- if (vd == spa->spa_root_vdev)
- mutex_exit(&spa->spa_scrub_lock);
+ mutex_exit(&vd->vdev_dtl_lock);
}
static int
vdev_dtl_load(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
- space_map_obj_t *smo = &vd->vdev_dtl;
+ space_map_obj_t *smo = &vd->vdev_dtl_smo;
objset_t *mos = spa->spa_meta_objset;
dmu_buf_t *db;
int error;
dmu_buf_rele(db, FTAG);
mutex_enter(&vd->vdev_dtl_lock);
- error = space_map_load(&vd->vdev_dtl_map, NULL, SM_ALLOC, smo, mos);
+ error = space_map_load(&vd->vdev_dtl[DTL_MISSING],
+ NULL, SM_ALLOC, smo, mos);
mutex_exit(&vd->vdev_dtl_lock);
return (error);
vdev_dtl_sync(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
- space_map_obj_t *smo = &vd->vdev_dtl;
- space_map_t *sm = &vd->vdev_dtl_map;
+ space_map_obj_t *smo = &vd->vdev_dtl_smo;
+ space_map_t *sm = &vd->vdev_dtl[DTL_MISSING];
objset_t *mos = spa->spa_meta_objset;
space_map_t smsync;
kmutex_t smlock;
}
/*
+ * Determine whether the specified vdev can be offlined/detached/removed
+ * without losing data.
+ */
+boolean_t
+vdev_dtl_required(vdev_t *vd)
+{
+ spa_t *spa = vd->vdev_spa;
+ vdev_t *tvd = vd->vdev_top;
+ uint8_t cant_read = vd->vdev_cant_read;
+ boolean_t required;
+
+ ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
+
+ if (vd == spa->spa_root_vdev || vd == tvd)
+ return (B_TRUE);
+
+ /*
+ * Temporarily mark the device as unreadable, and then determine
+ * whether this results in any DTL outages in the top-level vdev.
+ * If not, we can safely offline/detach/remove the device.
+ */
+ vd->vdev_cant_read = B_TRUE;
+ vdev_dtl_reassess(tvd, 0, 0, B_FALSE);
+ required = !vdev_dtl_empty(tvd, DTL_OUTAGE);
+ vd->vdev_cant_read = cant_read;
+ vdev_dtl_reassess(tvd, 0, 0, B_FALSE);
+
+ return (required);
+}
+
+/*
* Determine if resilver is needed, and if so the txg range.
*/
boolean_t
if (vd->vdev_children == 0) {
mutex_enter(&vd->vdev_dtl_lock);
- if (vd->vdev_dtl_map.sm_space != 0 && vdev_writeable(vd)) {
+ if (vd->vdev_dtl[DTL_MISSING].sm_space != 0 &&
+ vdev_writeable(vd)) {
space_seg_t *ss;
- ss = avl_first(&vd->vdev_dtl_map.sm_root);
+ ss = avl_first(&vd->vdev_dtl[DTL_MISSING].sm_root);
thismin = ss->ss_start - 1;
- ss = avl_last(&vd->vdev_dtl_map.sm_root);
+ ss = avl_last(&vd->vdev_dtl[DTL_MISSING].sm_root);
thismax = ss->ss_end;
needed = B_TRUE;
}
mutex_exit(&vd->vdev_dtl_lock);
} else {
- int c;
- for (c = 0; c < vd->vdev_children; c++) {
+ for (int 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 (c = 0; c < vd->vdev_children; c++)
+ for (int c = 0; c < vd->vdev_children; c++)
vdev_load(vd->vdev_child[c]);
/*
vd->vdev_parent->vdev_child[0] == vd)
vd->vdev_unspare = B_TRUE;
- (void) spa_vdev_state_exit(spa, vd, 0);
-
- VERIFY3U(spa_scrub(spa, POOL_SCRUB_RESILVER), ==, 0);
-
- return (0);
+ return (spa_vdev_state_exit(spa, vd, 0));
}
int
*/
if (!vd->vdev_offline) {
/*
- * If this device's top-level vdev has a non-empty DTL,
- * don't allow the device to be offlined.
- *
- * XXX -- make this more precise by allowing the offline
- * as long as the remaining devices don't have any DTL holes.
+ * If this device has the only valid copy of some data,
+ * don't allow it to be offlined.
*/
- if (vd->vdev_top->vdev_dtl_map.sm_space != 0)
+ if (vd->vdev_aux == NULL && vdev_dtl_required(vd))
return (spa_vdev_state_exit(spa, NULL, EBUSY));
/*
*/
vd->vdev_offline = B_TRUE;
vdev_reopen(vd->vdev_top);
- if (vdev_is_dead(vd->vdev_top) && vd->vdev_aux == NULL) {
+ if (vd->vdev_aux == NULL && vdev_is_dead(vd->vdev_top)) {
vd->vdev_offline = B_FALSE;
vdev_reopen(vd->vdev_top);
return (spa_vdev_state_exit(spa, NULL, EBUSY));
boolean_t
vdev_allocatable(vdev_t *vd)
{
+ uint64_t state = vd->vdev_state;
+
/*
- * We currently allow allocations from vdevs which maybe in the
+ * We currently allow allocations from vdevs which may be in the
* process of reopening (i.e. VDEV_STATE_CLOSED). If the device
* fails to reopen then we'll catch it later when we're holding
- * the proper locks.
+ * the proper locks. Note that we have to get the vdev state
+ * in a local variable because although it changes atomically,
+ * we're asking two separate questions about it.
*/
- return (!(vdev_is_dead(vd) && vd->vdev_state != VDEV_STATE_CLOSED) &&
+ return (!(state < VDEV_STATE_DEGRADED && state != VDEV_STATE_CLOSED) &&
!vd->vdev_cant_write);
}
void
vdev_stat_update(zio_t *zio, uint64_t psize)
{
- vdev_t *rvd = zio->io_spa->spa_root_vdev;
+ spa_t *spa = zio->io_spa;
+ vdev_t *rvd = spa->spa_root_vdev;
vdev_t *vd = zio->io_vd ? zio->io_vd : rvd;
vdev_t *pvd;
uint64_t txg = zio->io_txg;
return;
ASSERT(vd == zio->io_vd);
- if (!(flags & ZIO_FLAG_IO_BYPASS)) {
- mutex_enter(&vd->vdev_stat_lock);
- vs->vs_ops[type]++;
- vs->vs_bytes[type] += psize;
- mutex_exit(&vd->vdev_stat_lock);
- }
+
+ if (flags & ZIO_FLAG_IO_BYPASS)
+ return;
+
+ mutex_enter(&vd->vdev_stat_lock);
+
if (flags & ZIO_FLAG_IO_REPAIR) {
- ASSERT(zio->io_delegate_list == NULL);
- mutex_enter(&vd->vdev_stat_lock);
if (flags & ZIO_FLAG_SCRUB_THREAD)
vs->vs_scrub_repaired += psize;
- else
+ if (flags & ZIO_FLAG_SELF_HEAL)
vs->vs_self_healed += psize;
- mutex_exit(&vd->vdev_stat_lock);
}
+
+ vs->vs_ops[type]++;
+ vs->vs_bytes[type] += psize;
+
+ mutex_exit(&vd->vdev_stat_lock);
return;
}
vs->vs_write_errors++;
mutex_exit(&vd->vdev_stat_lock);
- if (type == ZIO_TYPE_WRITE && txg != 0 && vd->vdev_children == 0) {
- if (flags & ZIO_FLAG_SCRUB_THREAD) {
- ASSERT(flags & ZIO_FLAG_IO_REPAIR);
- for (pvd = vd; pvd != NULL; pvd = pvd->vdev_parent)
- vdev_dtl_dirty(&pvd->vdev_dtl_scrub, txg, 1);
- }
- if (!(flags & ZIO_FLAG_IO_REPAIR)) {
- if (vdev_dtl_contains(&vd->vdev_dtl_map, txg, 1))
+ if (type == ZIO_TYPE_WRITE && txg != 0 &&
+ (!(flags & ZIO_FLAG_IO_REPAIR) ||
+ (flags & ZIO_FLAG_SCRUB_THREAD))) {
+ /*
+ * This is either a normal write (not a repair), or it's a
+ * repair induced by the scrub thread. In the normal case,
+ * we commit the DTL change in the same txg as the block
+ * was born. In the scrub-induced repair case, we know that
+ * scrubs run in first-pass syncing context, so we commit
+ * the DTL change in spa->spa_syncing_txg.
+ *
+ * We currently do not make DTL entries for failed spontaneous
+ * self-healing writes triggered by normal (non-scrubbing)
+ * reads, because we have no transactional context in which to
+ * do so -- and it's not clear that it'd be desirable anyway.
+ */
+ if (vd->vdev_ops->vdev_op_leaf) {
+ uint64_t commit_txg = txg;
+ if (flags & ZIO_FLAG_SCRUB_THREAD) {
+ ASSERT(flags & ZIO_FLAG_IO_REPAIR);
+ ASSERT(spa_sync_pass(spa) == 1);
+ vdev_dtl_dirty(vd, DTL_SCRUB, txg, 1);
+ commit_txg = spa->spa_syncing_txg;
+ }
+ ASSERT(commit_txg >= spa->spa_syncing_txg);
+ if (vdev_dtl_contains(vd, DTL_MISSING, txg, 1))
return;
- vdev_dirty(vd->vdev_top, VDD_DTL, vd, txg);
- for (pvd = vd; pvd != NULL; pvd = pvd->vdev_parent)
- vdev_dtl_dirty(&pvd->vdev_dtl_map, txg, 1);
+ for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent)
+ vdev_dtl_dirty(pvd, DTL_PARTIAL, txg, 1);
+ vdev_dirty(vd->vdev_top, VDD_DTL, vd, commit_txg);
}
+ if (vd != rvd)
+ vdev_dtl_dirty(vd, DTL_MISSING, txg, 1);
}
}
void
vdev_propagate_state(vdev_t *vd)
{
- vdev_t *rvd = vd->vdev_spa->spa_root_vdev;
+ spa_t *spa = vd->vdev_spa;
+ vdev_t *rvd = spa->spa_root_vdev;
int degraded = 0, faulted = 0;
int corrupted = 0;
int c;
child = vd->vdev_child[c];
if (!vdev_readable(child) ||
- (!vdev_writeable(child) && (spa_mode & FWRITE))) {
+ (!vdev_writeable(child) && spa_writeable(spa))) {
/*
* Root special: if there is a top-level log
* device, treat the root vdev as if it were
git://git.camperquake.de / zfs.git / blobdiff