/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2013 by Delphix. All rights reserved.
+ * Copyright (c) 2012, 2015 by Delphix. All rights reserved.
*/
/*
#include <sys/metaslab.h>
#include <sys/zio.h>
#include <sys/dsl_scan.h>
+#include <sys/abd.h>
#include <sys/fs/zfs.h>
/*
}
static void
-vdev_label_read(zio_t *zio, vdev_t *vd, int l, void *buf, uint64_t offset,
- uint64_t size, zio_done_func_t *done, void *private, int flags)
+vdev_label_read(zio_t *zio, vdev_t *vd, int l, abd_t *buf, uint64_t offset,
+ uint64_t size, zio_done_func_t *done, void *private, int flags)
{
- ASSERT(spa_config_held(zio->io_spa, SCL_STATE_ALL, RW_WRITER) ==
- SCL_STATE_ALL);
+ ASSERT(
+ spa_config_held(zio->io_spa, SCL_STATE, RW_READER) == SCL_STATE ||
+ spa_config_held(zio->io_spa, SCL_STATE, RW_WRITER) == SCL_STATE);
ASSERT(flags & ZIO_FLAG_CONFIG_WRITER);
zio_nowait(zio_read_phys(zio, vd,
ZIO_PRIORITY_SYNC_READ, flags, B_TRUE));
}
-static void
-vdev_label_write(zio_t *zio, vdev_t *vd, int l, void *buf, uint64_t offset,
- uint64_t size, zio_done_func_t *done, void *private, int flags)
+void
+vdev_label_write(zio_t *zio, vdev_t *vd, int l, abd_t *buf, uint64_t offset,
+ uint64_t size, zio_done_func_t *done, void *private, int flags)
{
- ASSERT(spa_config_held(zio->io_spa, SCL_ALL, RW_WRITER) == SCL_ALL ||
- (spa_config_held(zio->io_spa, SCL_CONFIG | SCL_STATE, RW_READER) ==
- (SCL_CONFIG | SCL_STATE) &&
- dsl_pool_sync_context(spa_get_dsl(zio->io_spa))));
+ ASSERT(
+ spa_config_held(zio->io_spa, SCL_STATE, RW_READER) == SCL_STATE ||
+ spa_config_held(zio->io_spa, SCL_STATE, RW_WRITER) == SCL_STATE);
ASSERT(flags & ZIO_FLAG_CONFIG_WRITER);
zio_nowait(zio_write_phys(zio, vd,
ZIO_PRIORITY_SYNC_WRITE, flags, B_TRUE));
}
+/*
+ * Generate the nvlist representing this vdev's stats
+ */
+void
+vdev_config_generate_stats(vdev_t *vd, nvlist_t *nv)
+{
+ nvlist_t *nvx;
+ vdev_stat_t *vs;
+ vdev_stat_ex_t *vsx;
+
+ vs = kmem_alloc(sizeof (*vs), KM_SLEEP);
+ vsx = kmem_alloc(sizeof (*vsx), KM_SLEEP);
+
+ vdev_get_stats_ex(vd, vs, vsx);
+ fnvlist_add_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
+ (uint64_t *)vs, sizeof (*vs) / sizeof (uint64_t));
+
+ kmem_free(vs, sizeof (*vs));
+
+ /*
+ * Add extended stats into a special extended stats nvlist. This keeps
+ * all the extended stats nicely grouped together. The extended stats
+ * nvlist is then added to the main nvlist.
+ */
+ nvx = fnvlist_alloc();
+
+ /* ZIOs in flight to disk */
+ fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE,
+ vsx->vsx_active_queue[ZIO_PRIORITY_SYNC_READ]);
+
+ fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE,
+ vsx->vsx_active_queue[ZIO_PRIORITY_SYNC_WRITE]);
+
+ fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE,
+ vsx->vsx_active_queue[ZIO_PRIORITY_ASYNC_READ]);
+
+ fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE,
+ vsx->vsx_active_queue[ZIO_PRIORITY_ASYNC_WRITE]);
+
+ fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE,
+ vsx->vsx_active_queue[ZIO_PRIORITY_SCRUB]);
+
+ /* ZIOs pending */
+ fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_R_PEND_QUEUE,
+ vsx->vsx_pend_queue[ZIO_PRIORITY_SYNC_READ]);
+
+ fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_W_PEND_QUEUE,
+ vsx->vsx_pend_queue[ZIO_PRIORITY_SYNC_WRITE]);
+
+ fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_R_PEND_QUEUE,
+ vsx->vsx_pend_queue[ZIO_PRIORITY_ASYNC_READ]);
+
+ fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_W_PEND_QUEUE,
+ vsx->vsx_pend_queue[ZIO_PRIORITY_ASYNC_WRITE]);
+
+ fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SCRUB_PEND_QUEUE,
+ vsx->vsx_pend_queue[ZIO_PRIORITY_SCRUB]);
+
+ /* Histograms */
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO,
+ vsx->vsx_total_histo[ZIO_TYPE_READ],
+ ARRAY_SIZE(vsx->vsx_total_histo[ZIO_TYPE_READ]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO,
+ vsx->vsx_total_histo[ZIO_TYPE_WRITE],
+ ARRAY_SIZE(vsx->vsx_total_histo[ZIO_TYPE_WRITE]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO,
+ vsx->vsx_disk_histo[ZIO_TYPE_READ],
+ ARRAY_SIZE(vsx->vsx_disk_histo[ZIO_TYPE_READ]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO,
+ vsx->vsx_disk_histo[ZIO_TYPE_WRITE],
+ ARRAY_SIZE(vsx->vsx_disk_histo[ZIO_TYPE_WRITE]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO,
+ vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_READ],
+ ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_READ]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO,
+ vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_WRITE],
+ ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_WRITE]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO,
+ vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_READ],
+ ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_READ]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO,
+ vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_WRITE],
+ ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_WRITE]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO,
+ vsx->vsx_queue_histo[ZIO_PRIORITY_SCRUB],
+ ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_SCRUB]));
+
+ /* Request sizes */
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_IND_R_HISTO,
+ vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_READ],
+ ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_READ]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_IND_W_HISTO,
+ vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_WRITE],
+ ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_WRITE]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_IND_R_HISTO,
+ vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_READ],
+ ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_READ]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_IND_W_HISTO,
+ vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_WRITE],
+ ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_WRITE]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_IND_SCRUB_HISTO,
+ vsx->vsx_ind_histo[ZIO_PRIORITY_SCRUB],
+ ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_SCRUB]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_AGG_R_HISTO,
+ vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_READ],
+ ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_READ]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_AGG_W_HISTO,
+ vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_WRITE],
+ ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_WRITE]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_AGG_R_HISTO,
+ vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_READ],
+ ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_READ]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_AGG_W_HISTO,
+ vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_WRITE],
+ ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_WRITE]));
+
+ fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_AGG_SCRUB_HISTO,
+ vsx->vsx_agg_histo[ZIO_PRIORITY_SCRUB],
+ ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_SCRUB]));
+
+ /* Add extended stats nvlist to main nvlist */
+ fnvlist_add_nvlist(nv, ZPOOL_CONFIG_VDEV_STATS_EX, nvx);
+
+ fnvlist_free(nvx);
+ kmem_free(vsx, sizeof (*vsx));
+}
+
/*
* Generate the nvlist representing this vdev's config.
*/
vdev_config_flag_t flags)
{
nvlist_t *nv = NULL;
-
- VERIFY(nvlist_alloc(&nv, NV_UNIQUE_NAME, KM_PUSHPAGE) == 0);
+ nv = fnvlist_alloc();
fnvlist_add_string(nv, ZPOOL_CONFIG_TYPE, vd->vdev_ops->vdev_op_type);
if (!(flags & (VDEV_CONFIG_SPARE | VDEV_CONFIG_L2CACHE)))
fnvlist_add_string(nv, ZPOOL_CONFIG_PHYS_PATH,
vd->vdev_physpath);
+ if (vd->vdev_enc_sysfs_path != NULL)
+ fnvlist_add_string(nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH,
+ vd->vdev_enc_sysfs_path);
+
if (vd->vdev_fru != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_FRU, vd->vdev_fru);
vd->vdev_removing);
}
- if (vd->vdev_dtl_smo.smo_object != 0)
+ if (vd->vdev_dtl_sm != NULL) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_DTL,
- vd->vdev_dtl_smo.smo_object);
+ space_map_object(vd->vdev_dtl_sm));
+ }
if (vd->vdev_crtxg)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_CREATE_TXG, vd->vdev_crtxg);
+ if (flags & VDEV_CONFIG_MOS) {
+ if (vd->vdev_leaf_zap != 0) {
+ ASSERT(vd->vdev_ops->vdev_op_leaf);
+ fnvlist_add_uint64(nv, ZPOOL_CONFIG_VDEV_LEAF_ZAP,
+ vd->vdev_leaf_zap);
+ }
+
+ if (vd->vdev_top_zap != 0) {
+ ASSERT(vd == vd->vdev_top);
+ fnvlist_add_uint64(nv, ZPOOL_CONFIG_VDEV_TOP_ZAP,
+ vd->vdev_top_zap);
+ }
+ }
+
if (getstats) {
- vdev_stat_t vs;
pool_scan_stat_t ps;
- vdev_get_stats(vd, &vs);
- fnvlist_add_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
- (uint64_t *)&vs, sizeof (vs) / sizeof (uint64_t));
+ vdev_config_generate_stats(vd, nv);
/* provide either current or previous scan information */
if (spa_scan_get_stats(spa, &ps) == 0) {
ASSERT(!vd->vdev_ishole);
child = kmem_alloc(vd->vdev_children * sizeof (nvlist_t *),
- KM_PUSHPAGE);
+ KM_SLEEP);
for (c = 0, idx = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (vd->vdev_ishole)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_HOLE, B_TRUE);
+ /* Set the reason why we're FAULTED/DEGRADED. */
switch (vd->vdev_stat.vs_aux) {
case VDEV_AUX_ERR_EXCEEDED:
aux = "err_exceeded";
break;
}
- if (aux != NULL)
+ if (aux != NULL && !vd->vdev_tmpoffline) {
fnvlist_add_string(nv, ZPOOL_CONFIG_AUX_STATE, aux);
+ } else {
+ /*
+ * We're healthy - clear any previous AUX_STATE values.
+ */
+ if (nvlist_exists(nv, ZPOOL_CONFIG_AUX_STATE))
+ nvlist_remove_all(nv, ZPOOL_CONFIG_AUX_STATE);
+ }
if (vd->vdev_splitting && vd->vdev_orig_guid != 0LL) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_ORIG_GUID,
uint64_t *array;
uint_t c, idx;
- array = kmem_alloc(rvd->vdev_children * sizeof (uint64_t), KM_PUSHPAGE);
+ array = kmem_alloc(rvd->vdev_children * sizeof (uint64_t), KM_SLEEP);
for (c = 0, idx = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
spa_t *spa = vd->vdev_spa;
nvlist_t *config = NULL;
vdev_phys_t *vp;
+ abd_t *vp_abd;
zio_t *zio;
uint64_t best_txg = 0;
int error = 0;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SPECULATIVE;
- int l;
ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
if (!vdev_readable(vd))
return (NULL);
- vp = zio_buf_alloc(sizeof (vdev_phys_t));
+ vp_abd = abd_alloc_linear(sizeof (vdev_phys_t), B_TRUE);
+ vp = abd_to_buf(vp_abd);
retry:
- for (l = 0; l < VDEV_LABELS; l++) {
+ for (int l = 0; l < VDEV_LABELS; l++) {
nvlist_t *label = NULL;
zio = zio_root(spa, NULL, NULL, flags);
- vdev_label_read(zio, vd, l, vp,
+ vdev_label_read(zio, vd, l, vp_abd,
offsetof(vdev_label_t, vl_vdev_phys),
sizeof (vdev_phys_t), NULL, NULL, flags);
goto retry;
}
- zio_buf_free(vp, sizeof (vdev_phys_t));
+ abd_free(vp_abd);
return (config);
}
* read-only. Instead we look to see if the pools is marked
* read-only in the namespace and set the state to active.
*/
- if ((spa = spa_by_guid(pool_guid, device_guid)) != NULL &&
+ if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
+ (spa = spa_by_guid(pool_guid, device_guid)) != NULL &&
spa_mode(spa) == FREAD)
state = POOL_STATE_ACTIVE;
spa_t *spa = vd->vdev_spa;
nvlist_t *label;
vdev_phys_t *vp;
- char *pad2;
+ abd_t *vp_abd;
+ abd_t *pad2;
uberblock_t *ub;
+ abd_t *ub_abd;
zio_t *zio;
char *buf;
size_t buflen;
int error;
uint64_t spare_guid = 0, l2cache_guid = 0;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL;
- int c, l;
- vdev_t *pvd;
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
- for (c = 0; c < vd->vdev_children; c++)
+ for (int c = 0; c < vd->vdev_children; c++)
if ((error = vdev_label_init(vd->vdev_child[c],
crtxg, reason)) != 0)
return (error);
vd->vdev_guid += guid_delta;
- for (pvd = vd; pvd != NULL; pvd = pvd->vdev_parent)
+ for (vdev_t *pvd = vd; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum += guid_delta;
/*
vd->vdev_guid += guid_delta;
- for (pvd = vd; pvd != NULL; pvd = pvd->vdev_parent)
+ for (vdev_t *pvd = vd; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum += guid_delta;
/*
/*
* Initialize its label.
*/
- vp = zio_buf_alloc(sizeof (vdev_phys_t));
- bzero(vp, sizeof (vdev_phys_t));
+ vp_abd = abd_alloc_linear(sizeof (vdev_phys_t), B_TRUE);
+ abd_zero(vp_abd, sizeof (vdev_phys_t));
+ vp = abd_to_buf(vp_abd);
/*
* Generate a label describing the pool and our top-level vdev.
* active hot spare (in which case we want to revert the
* labels).
*/
- VERIFY(nvlist_alloc(&label, NV_UNIQUE_NAME, KM_PUSHPAGE) == 0);
+ VERIFY(nvlist_alloc(&label, NV_UNIQUE_NAME, KM_SLEEP) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_VERSION,
spa_version(spa)) == 0);
/*
* For level 2 ARC devices, add a special label.
*/
- VERIFY(nvlist_alloc(&label, NV_UNIQUE_NAME, KM_PUSHPAGE) == 0);
+ VERIFY(nvlist_alloc(&label, NV_UNIQUE_NAME, KM_SLEEP) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_VERSION,
spa_version(spa)) == 0);
buf = vp->vp_nvlist;
buflen = sizeof (vp->vp_nvlist);
- error = nvlist_pack(label, &buf, &buflen, NV_ENCODE_XDR, KM_PUSHPAGE);
+ error = nvlist_pack(label, &buf, &buflen, NV_ENCODE_XDR, KM_SLEEP);
if (error != 0) {
nvlist_free(label);
- zio_buf_free(vp, sizeof (vdev_phys_t));
+ abd_free(vp_abd);
/* EFAULT means nvlist_pack ran out of room */
- return (error == EFAULT ? ENAMETOOLONG : EINVAL);
+ return (SET_ERROR(error == EFAULT ? ENAMETOOLONG : EINVAL));
}
/*
* Initialize uberblock template.
*/
- ub = zio_buf_alloc(VDEV_UBERBLOCK_RING);
- bzero(ub, VDEV_UBERBLOCK_RING);
- *ub = spa->spa_uberblock;
+ ub_abd = abd_alloc_linear(VDEV_UBERBLOCK_RING, B_TRUE);
+ abd_zero(ub_abd, VDEV_UBERBLOCK_RING);
+ abd_copy_from_buf(ub_abd, &spa->spa_uberblock, sizeof (uberblock_t));
+ ub = abd_to_buf(ub_abd);
ub->ub_txg = 0;
/* Initialize the 2nd padding area. */
- pad2 = zio_buf_alloc(VDEV_PAD_SIZE);
- bzero(pad2, VDEV_PAD_SIZE);
+ pad2 = abd_alloc_for_io(VDEV_PAD_SIZE, B_TRUE);
+ abd_zero(pad2, VDEV_PAD_SIZE);
/*
* Write everything in parallel.
retry:
zio = zio_root(spa, NULL, NULL, flags);
- for (l = 0; l < VDEV_LABELS; l++) {
+ for (int l = 0; l < VDEV_LABELS; l++) {
- vdev_label_write(zio, vd, l, vp,
+ vdev_label_write(zio, vd, l, vp_abd,
offsetof(vdev_label_t, vl_vdev_phys),
sizeof (vdev_phys_t), NULL, NULL, flags);
offsetof(vdev_label_t, vl_pad2),
VDEV_PAD_SIZE, NULL, NULL, flags);
- vdev_label_write(zio, vd, l, ub,
+ vdev_label_write(zio, vd, l, ub_abd,
offsetof(vdev_label_t, vl_uberblock),
VDEV_UBERBLOCK_RING, NULL, NULL, flags);
}
}
nvlist_free(label);
- zio_buf_free(pad2, VDEV_PAD_SIZE);
- zio_buf_free(ub, VDEV_UBERBLOCK_RING);
- zio_buf_free(vp, sizeof (vdev_phys_t));
+ abd_free(pad2);
+ abd_free(ub_abd);
+ abd_free(vp_abd);
/*
* If this vdev hasn't been previously identified as a spare, then we
* among uberblocks with equal txg, choose the one with the latest timestamp.
*/
static int
-vdev_uberblock_compare(uberblock_t *ub1, uberblock_t *ub2)
+vdev_uberblock_compare(const uberblock_t *ub1, const uberblock_t *ub2)
{
- if (ub1->ub_txg < ub2->ub_txg)
- return (-1);
- if (ub1->ub_txg > ub2->ub_txg)
- return (1);
-
- if (ub1->ub_timestamp < ub2->ub_timestamp)
- return (-1);
- if (ub1->ub_timestamp > ub2->ub_timestamp)
- return (1);
+ int cmp = AVL_CMP(ub1->ub_txg, ub2->ub_txg);
+ if (likely(cmp))
+ return (cmp);
- return (0);
+ return (AVL_CMP(ub1->ub_timestamp, ub2->ub_timestamp));
}
struct ubl_cbdata {
vdev_t *vd = zio->io_vd;
spa_t *spa = zio->io_spa;
zio_t *rio = zio->io_private;
- uberblock_t *ub = zio->io_data;
+ uberblock_t *ub = abd_to_buf(zio->io_abd);
struct ubl_cbdata *cbp = rio->io_private;
ASSERT3U(zio->io_size, ==, VDEV_UBERBLOCK_SIZE(vd));
mutex_exit(&rio->io_lock);
}
- zio_buf_free(zio->io_data, zio->io_size);
+ abd_free(zio->io_abd);
}
static void
vdev_uberblock_load_impl(zio_t *zio, vdev_t *vd, int flags,
struct ubl_cbdata *cbp)
{
- int c, l, n;
-
- for (c = 0; c < vd->vdev_children; c++)
+ for (int c = 0; c < vd->vdev_children; c++)
vdev_uberblock_load_impl(zio, vd->vdev_child[c], flags, cbp);
if (vd->vdev_ops->vdev_op_leaf && vdev_readable(vd)) {
- for (l = 0; l < VDEV_LABELS; l++) {
- for (n = 0; n < VDEV_UBERBLOCK_COUNT(vd); n++) {
+ for (int l = 0; l < VDEV_LABELS; l++) {
+ for (int n = 0; n < VDEV_UBERBLOCK_COUNT(vd); n++) {
vdev_label_read(zio, vd, l,
- zio_buf_alloc(VDEV_UBERBLOCK_SIZE(vd)),
- VDEV_UBERBLOCK_OFFSET(vd, n),
+ abd_alloc_linear(VDEV_UBERBLOCK_SIZE(vd),
+ B_TRUE), VDEV_UBERBLOCK_OFFSET(vd, n),
VDEV_UBERBLOCK_SIZE(vd),
vdev_uberblock_load_done, zio, flags);
}
spa_config_exit(spa, SCL_ALL, FTAG);
}
+/*
+ * For use when a leaf vdev is expanded.
+ * The location of labels 2 and 3 changed, and at the new location the
+ * uberblock rings are either empty or contain garbage. The sync will write
+ * new configs there because the vdev is dirty, but expansion also needs the
+ * uberblock rings copied. Read them from label 0 which did not move.
+ *
+ * Since the point is to populate labels {2,3} with valid uberblocks,
+ * we zero uberblocks we fail to read or which are not valid.
+ */
+
+static void
+vdev_copy_uberblocks(vdev_t *vd)
+{
+ abd_t *ub_abd;
+ zio_t *write_zio;
+ int locks = (SCL_L2ARC | SCL_ZIO);
+ int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL |
+ ZIO_FLAG_SPECULATIVE;
+
+ ASSERT(spa_config_held(vd->vdev_spa, SCL_STATE, RW_READER) ==
+ SCL_STATE);
+ ASSERT(vd->vdev_ops->vdev_op_leaf);
+
+ spa_config_enter(vd->vdev_spa, locks, FTAG, RW_READER);
+
+ ub_abd = abd_alloc_linear(VDEV_UBERBLOCK_SIZE(vd), B_TRUE);
+
+ write_zio = zio_root(vd->vdev_spa, NULL, NULL, flags);
+ for (int n = 0; n < VDEV_UBERBLOCK_COUNT(vd); n++) {
+ const int src_label = 0;
+ zio_t *zio;
+
+ zio = zio_root(vd->vdev_spa, NULL, NULL, flags);
+ vdev_label_read(zio, vd, src_label, ub_abd,
+ VDEV_UBERBLOCK_OFFSET(vd, n), VDEV_UBERBLOCK_SIZE(vd),
+ NULL, NULL, flags);
+
+ if (zio_wait(zio) || uberblock_verify(abd_to_buf(ub_abd)))
+ abd_zero(ub_abd, VDEV_UBERBLOCK_SIZE(vd));
+
+ for (int l = 2; l < VDEV_LABELS; l++)
+ vdev_label_write(write_zio, vd, l, ub_abd,
+ VDEV_UBERBLOCK_OFFSET(vd, n),
+ VDEV_UBERBLOCK_SIZE(vd), NULL, NULL,
+ flags | ZIO_FLAG_DONT_PROPAGATE);
+ }
+ (void) zio_wait(write_zio);
+
+ spa_config_exit(vd->vdev_spa, locks, FTAG);
+
+ abd_free(ub_abd);
+}
+
/*
* On success, increment root zio's count of good writes.
* We only get credit for writes to known-visible vdevs; see spa_vdev_add().
uint64_t *good_writes = zio->io_private;
if (zio->io_error == 0 && zio->io_vd->vdev_top->vdev_ms_array != 0)
- atomic_add_64(good_writes, 1);
+ atomic_inc_64(good_writes);
}
/*
static void
vdev_uberblock_sync(zio_t *zio, uberblock_t *ub, vdev_t *vd, int flags)
{
- uberblock_t *ubbuf;
- int c, l, n;
-
- for (c = 0; c < vd->vdev_children; c++)
+ for (int c = 0; c < vd->vdev_children; c++)
vdev_uberblock_sync(zio, ub, vd->vdev_child[c], flags);
if (!vd->vdev_ops->vdev_op_leaf)
if (!vdev_writeable(vd))
return;
- n = ub->ub_txg & (VDEV_UBERBLOCK_COUNT(vd) - 1);
+ /* If the vdev was expanded, need to copy uberblock rings. */
+ if (vd->vdev_state == VDEV_STATE_HEALTHY &&
+ vd->vdev_copy_uberblocks == B_TRUE) {
+ vdev_copy_uberblocks(vd);
+ vd->vdev_copy_uberblocks = B_FALSE;
+ }
- ubbuf = zio_buf_alloc(VDEV_UBERBLOCK_SIZE(vd));
- bzero(ubbuf, VDEV_UBERBLOCK_SIZE(vd));
- *ubbuf = *ub;
+ int m = spa_multihost(vd->vdev_spa) ? MMP_BLOCKS_PER_LABEL : 0;
+ int n = ub->ub_txg % (VDEV_UBERBLOCK_COUNT(vd) - m);
- for (l = 0; l < VDEV_LABELS; l++)
- vdev_label_write(zio, vd, l, ubbuf,
+ /* Copy the uberblock_t into the ABD */
+ abd_t *ub_abd = abd_alloc_for_io(VDEV_UBERBLOCK_SIZE(vd), B_TRUE);
+ abd_zero(ub_abd, VDEV_UBERBLOCK_SIZE(vd));
+ abd_copy_from_buf(ub_abd, ub, sizeof (uberblock_t));
+
+ for (int l = 0; l < VDEV_LABELS; l++)
+ vdev_label_write(zio, vd, l, ub_abd,
VDEV_UBERBLOCK_OFFSET(vd, n), VDEV_UBERBLOCK_SIZE(vd),
vdev_uberblock_sync_done, zio->io_private,
flags | ZIO_FLAG_DONT_PROPAGATE);
- zio_buf_free(ubbuf, VDEV_UBERBLOCK_SIZE(vd));
+ abd_free(ub_abd);
}
/* Sync the uberblocks to all vdevs in svd[] */
spa_t *spa = svd[0]->vdev_spa;
zio_t *zio;
uint64_t good_writes = 0;
- int v;
zio = zio_root(spa, NULL, &good_writes, flags);
- for (v = 0; v < svdcount; v++)
+ for (int v = 0; v < svdcount; v++)
vdev_uberblock_sync(zio, ub, svd[v], flags);
(void) zio_wait(zio);
*/
zio = zio_root(spa, NULL, NULL, flags);
- for (v = 0; v < svdcount; v++)
+ for (int v = 0; v < svdcount; v++)
zio_flush(zio, svd[v]);
(void) zio_wait(zio);
uint64_t *good_writes = zio->io_private;
if (zio->io_error == 0)
- atomic_add_64(good_writes, 1);
+ atomic_inc_64(good_writes);
}
/*
{
nvlist_t *label;
vdev_phys_t *vp;
+ abd_t *vp_abd;
char *buf;
size_t buflen;
- int c;
- for (c = 0; c < vd->vdev_children; c++)
+ for (int c = 0; c < vd->vdev_children; c++)
vdev_label_sync(zio, vd->vdev_child[c], l, txg, flags);
if (!vd->vdev_ops->vdev_op_leaf)
*/
label = spa_config_generate(vd->vdev_spa, vd, txg, B_FALSE);
- vp = zio_buf_alloc(sizeof (vdev_phys_t));
- bzero(vp, sizeof (vdev_phys_t));
+ vp_abd = abd_alloc_linear(sizeof (vdev_phys_t), B_TRUE);
+ abd_zero(vp_abd, sizeof (vdev_phys_t));
+ vp = abd_to_buf(vp_abd);
buf = vp->vp_nvlist;
buflen = sizeof (vp->vp_nvlist);
- if (!nvlist_pack(label, &buf, &buflen, NV_ENCODE_XDR, KM_PUSHPAGE)) {
+ if (!nvlist_pack(label, &buf, &buflen, NV_ENCODE_XDR, KM_SLEEP)) {
for (; l < VDEV_LABELS; l += 2) {
- vdev_label_write(zio, vd, l, vp,
+ vdev_label_write(zio, vd, l, vp_abd,
offsetof(vdev_label_t, vl_vdev_phys),
sizeof (vdev_phys_t),
vdev_label_sync_done, zio->io_private,
}
}
- zio_buf_free(vp, sizeof (vdev_phys_t));
+ abd_free(vp_abd);
nvlist_free(label);
}
for (vd = list_head(dl); vd != NULL; vd = list_next(dl, vd)) {
uint64_t *good_writes;
- zio_t *vio;
ASSERT(!vd->vdev_ishole);
- good_writes = kmem_zalloc(sizeof (uint64_t), KM_PUSHPAGE);
- vio = zio_null(zio, spa, NULL,
+ good_writes = kmem_zalloc(sizeof (uint64_t), KM_SLEEP);
+ zio_t *vio = zio_null(zio, spa, NULL,
(vd->vdev_islog || vd->vdev_aux != NULL) ?
vdev_label_sync_ignore_done : vdev_label_sync_top_done,
good_writes, flags);
* at any time, you can just call it again, and it will resume its work.
*/
int
-vdev_config_sync(vdev_t **svd, int svdcount, uint64_t txg, boolean_t tryhard)
+vdev_config_sync(vdev_t **svd, int svdcount, uint64_t txg)
{
spa_t *spa = svd[0]->vdev_spa;
uberblock_t *ub = &spa->spa_uberblock;
vdev_t *vd;
zio_t *zio;
- int error;
+ int error = 0;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL;
+retry:
/*
* Normally, we don't want to try too hard to write every label and
* uberblock. If there is a flaky disk, we don't want the rest of the
* single label out, we should retry with ZIO_FLAG_TRYHARD before
* bailing out and declaring the pool faulted.
*/
- if (tryhard)
+ if (error != 0) {
+ if ((flags & ZIO_FLAG_TRYHARD) != 0)
+ return (error);
flags |= ZIO_FLAG_TRYHARD;
+ }
ASSERT(ub->ub_txg <= txg);
* and the vdev configuration hasn't changed,
* then there's nothing to do.
*/
- if (ub->ub_txg < txg &&
- uberblock_update(ub, spa->spa_root_vdev, txg) == B_FALSE &&
- list_is_empty(&spa->spa_config_dirty_list))
- return (0);
+ if (ub->ub_txg < txg) {
+ boolean_t changed = uberblock_update(ub, spa->spa_root_vdev,
+ txg, spa->spa_mmp.mmp_delay);
+
+ if (!changed && list_is_empty(&spa->spa_config_dirty_list))
+ return (0);
+ }
if (txg > spa_freeze_txg(spa))
return (0);
* are committed to stable storage before the uberblock update.
*/
if ((error = vdev_label_sync_list(spa, 0, txg, flags)) != 0)
- return (error);
+ goto retry;
/*
* Sync the uberblocks to all vdevs in svd[].
* to the new uberblocks.
*/
if ((error = vdev_uberblock_sync_list(svd, svdcount, ub, flags)) != 0)
- return (error);
+ goto retry;
+
+
+ if (spa_multihost(spa))
+ mmp_update_uberblock(spa, ub);
/*
* Sync out odd labels for every dirty vdev. If the system dies
* to disk to ensure that all odd-label updates are committed to
* stable storage before the next transaction group begins.
*/
- return (vdev_label_sync_list(spa, 1, txg, flags));
+ if ((error = vdev_label_sync_list(spa, 1, txg, flags)) != 0)
+ goto retry;
+
+ return (0);
}