Index: stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/trim_map.c =================================================================== --- stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/trim_map.c (revision 274799) +++ stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/trim_map.c (revision 274800) @@ -1,649 +1,647 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2012 Pawel Jakub Dawidek . * All rights reserved. */ #include #include #include #include #include /* * Calculate the zio end, upgrading based on ashift which would be * done by zio_vdev_io_start. * * This makes free range consolidation much more effective * than it would otherwise be as well as ensuring that entire * blocks are invalidated by writes. */ #define TRIM_ZIO_END(vd, offset, size) (offset + \ P2ROUNDUP(size, 1ULL << vd->vdev_top->vdev_ashift)) #define TRIM_MAP_SINC(tm, size) \ atomic_add_64(&(tm)->tm_bytes, (size)) #define TRIM_MAP_SDEC(tm, size) \ atomic_add_64(&(tm)->tm_bytes, -(size)) #define TRIM_MAP_QINC(tm) \ atomic_inc_64(&(tm)->tm_pending); \ #define TRIM_MAP_QDEC(tm) \ atomic_dec_64(&(tm)->tm_pending); typedef struct trim_map { list_t tm_head; /* List of segments sorted by txg. */ avl_tree_t tm_queued_frees; /* AVL tree of segments waiting for TRIM. */ avl_tree_t tm_inflight_frees; /* AVL tree of in-flight TRIMs. */ avl_tree_t tm_inflight_writes; /* AVL tree of in-flight writes. */ list_t tm_pending_writes; /* Writes blocked on in-flight frees. */ kmutex_t tm_lock; uint64_t tm_pending; /* Count of pending TRIMs. */ uint64_t tm_bytes; /* Total size in bytes of queued TRIMs. */ } trim_map_t; typedef struct trim_seg { avl_node_t ts_node; /* AVL node. */ list_node_t ts_next; /* List element. */ uint64_t ts_start; /* Starting offset of this segment. */ uint64_t ts_end; /* Ending offset (non-inclusive). */ uint64_t ts_txg; /* Segment creation txg. */ hrtime_t ts_time; /* Segment creation time. */ } trim_seg_t; extern boolean_t zfs_trim_enabled; static u_int trim_txg_delay = 32; static u_int trim_timeout = 30; static u_int trim_max_interval = 1; /* Limit outstanding TRIMs to 2G (max size for a single TRIM request) */ static uint64_t trim_vdev_max_bytes = 2147483648; /* Limit outstanding TRIMs to 64 (max ranges for a single TRIM request) */ static u_int trim_vdev_max_pending = 64; SYSCTL_DECL(_vfs_zfs); SYSCTL_NODE(_vfs_zfs, OID_AUTO, trim, CTLFLAG_RD, 0, "ZFS TRIM"); TUNABLE_INT("vfs.zfs.trim.txg_delay", &trim_txg_delay); SYSCTL_UINT(_vfs_zfs_trim, OID_AUTO, txg_delay, CTLFLAG_RWTUN, &trim_txg_delay, 0, "Delay TRIMs by up to this many TXGs"); TUNABLE_INT("vfs.zfs.trim.timeout", &trim_timeout); SYSCTL_UINT(_vfs_zfs_trim, OID_AUTO, timeout, CTLFLAG_RWTUN, &trim_timeout, 0, "Delay TRIMs by up to this many seconds"); TUNABLE_INT("vfs.zfs.trim.max_interval", &trim_max_interval); SYSCTL_UINT(_vfs_zfs_trim, OID_AUTO, max_interval, CTLFLAG_RWTUN, &trim_max_interval, 0, "Maximum interval between TRIM queue processing (seconds)"); SYSCTL_DECL(_vfs_zfs_vdev); TUNABLE_QUAD("vfs.zfs.vdev.trim_max_bytes", &trim_vdev_max_bytes); SYSCTL_QUAD(_vfs_zfs_vdev, OID_AUTO, trim_max_bytes, CTLFLAG_RWTUN, &trim_vdev_max_bytes, 0, "Maximum pending TRIM bytes for a vdev"); TUNABLE_INT("vfs.zfs.vdev.trim_max_pending", &trim_vdev_max_pending); SYSCTL_UINT(_vfs_zfs_vdev, OID_AUTO, trim_max_pending, CTLFLAG_RWTUN, &trim_vdev_max_pending, 0, "Maximum pending TRIM segments for a vdev"); static void trim_map_vdev_commit_done(spa_t *spa, vdev_t *vd); static int trim_map_seg_compare(const void *x1, const void *x2) { const trim_seg_t *s1 = x1; const trim_seg_t *s2 = x2; if (s1->ts_start < s2->ts_start) { if (s1->ts_end > s2->ts_start) return (0); return (-1); } if (s1->ts_start > s2->ts_start) { if (s1->ts_start < s2->ts_end) return (0); return (1); } return (0); } static int trim_map_zio_compare(const void *x1, const void *x2) { const zio_t *z1 = x1; const zio_t *z2 = x2; if (z1->io_offset < z2->io_offset) { if (z1->io_offset + z1->io_size > z2->io_offset) return (0); return (-1); } if (z1->io_offset > z2->io_offset) { if (z1->io_offset < z2->io_offset + z2->io_size) return (0); return (1); } return (0); } void trim_map_create(vdev_t *vd) { trim_map_t *tm; - ASSERT(vd->vdev_ops->vdev_op_leaf); - - if (!zfs_trim_enabled) - return; + ASSERT(zfs_trim_enabled && !vd->vdev_notrim && + vd->vdev_ops->vdev_op_leaf); tm = kmem_zalloc(sizeof (*tm), KM_SLEEP); mutex_init(&tm->tm_lock, NULL, MUTEX_DEFAULT, NULL); list_create(&tm->tm_head, sizeof (trim_seg_t), offsetof(trim_seg_t, ts_next)); list_create(&tm->tm_pending_writes, sizeof (zio_t), offsetof(zio_t, io_trim_link)); avl_create(&tm->tm_queued_frees, trim_map_seg_compare, sizeof (trim_seg_t), offsetof(trim_seg_t, ts_node)); avl_create(&tm->tm_inflight_frees, trim_map_seg_compare, sizeof (trim_seg_t), offsetof(trim_seg_t, ts_node)); avl_create(&tm->tm_inflight_writes, trim_map_zio_compare, sizeof (zio_t), offsetof(zio_t, io_trim_node)); vd->vdev_trimmap = tm; } void trim_map_destroy(vdev_t *vd) { trim_map_t *tm; trim_seg_t *ts; ASSERT(vd->vdev_ops->vdev_op_leaf); if (!zfs_trim_enabled) return; tm = vd->vdev_trimmap; if (tm == NULL) return; /* * We may have been called before trim_map_vdev_commit_done() * had a chance to run, so do it now to prune the remaining * inflight frees. */ trim_map_vdev_commit_done(vd->vdev_spa, vd); mutex_enter(&tm->tm_lock); while ((ts = list_head(&tm->tm_head)) != NULL) { avl_remove(&tm->tm_queued_frees, ts); list_remove(&tm->tm_head, ts); kmem_free(ts, sizeof (*ts)); TRIM_MAP_SDEC(tm, ts->ts_end - ts->ts_start); TRIM_MAP_QDEC(tm); } mutex_exit(&tm->tm_lock); avl_destroy(&tm->tm_queued_frees); avl_destroy(&tm->tm_inflight_frees); avl_destroy(&tm->tm_inflight_writes); list_destroy(&tm->tm_pending_writes); list_destroy(&tm->tm_head); mutex_destroy(&tm->tm_lock); kmem_free(tm, sizeof (*tm)); vd->vdev_trimmap = NULL; } static void trim_map_segment_add(trim_map_t *tm, uint64_t start, uint64_t end, uint64_t txg) { avl_index_t where; trim_seg_t tsearch, *ts_before, *ts_after, *ts; boolean_t merge_before, merge_after; hrtime_t time; ASSERT(MUTEX_HELD(&tm->tm_lock)); VERIFY(start < end); time = gethrtime(); tsearch.ts_start = start; tsearch.ts_end = end; ts = avl_find(&tm->tm_queued_frees, &tsearch, &where); if (ts != NULL) { if (start < ts->ts_start) trim_map_segment_add(tm, start, ts->ts_start, txg); if (end > ts->ts_end) trim_map_segment_add(tm, ts->ts_end, end, txg); return; } ts_before = avl_nearest(&tm->tm_queued_frees, where, AVL_BEFORE); ts_after = avl_nearest(&tm->tm_queued_frees, where, AVL_AFTER); merge_before = (ts_before != NULL && ts_before->ts_end == start); merge_after = (ts_after != NULL && ts_after->ts_start == end); if (merge_before && merge_after) { TRIM_MAP_SINC(tm, ts_after->ts_start - ts_before->ts_end); TRIM_MAP_QDEC(tm); avl_remove(&tm->tm_queued_frees, ts_before); list_remove(&tm->tm_head, ts_before); ts_after->ts_start = ts_before->ts_start; ts_after->ts_txg = txg; ts_after->ts_time = time; kmem_free(ts_before, sizeof (*ts_before)); } else if (merge_before) { TRIM_MAP_SINC(tm, end - ts_before->ts_end); ts_before->ts_end = end; ts_before->ts_txg = txg; ts_before->ts_time = time; } else if (merge_after) { TRIM_MAP_SINC(tm, ts_after->ts_start - start); ts_after->ts_start = start; ts_after->ts_txg = txg; ts_after->ts_time = time; } else { TRIM_MAP_SINC(tm, end - start); TRIM_MAP_QINC(tm); ts = kmem_alloc(sizeof (*ts), KM_SLEEP); ts->ts_start = start; ts->ts_end = end; ts->ts_txg = txg; ts->ts_time = time; avl_insert(&tm->tm_queued_frees, ts, where); list_insert_tail(&tm->tm_head, ts); } } static void trim_map_segment_remove(trim_map_t *tm, trim_seg_t *ts, uint64_t start, uint64_t end) { trim_seg_t *nts; boolean_t left_over, right_over; ASSERT(MUTEX_HELD(&tm->tm_lock)); left_over = (ts->ts_start < start); right_over = (ts->ts_end > end); TRIM_MAP_SDEC(tm, end - start); if (left_over && right_over) { nts = kmem_alloc(sizeof (*nts), KM_SLEEP); nts->ts_start = end; nts->ts_end = ts->ts_end; nts->ts_txg = ts->ts_txg; nts->ts_time = ts->ts_time; ts->ts_end = start; avl_insert_here(&tm->tm_queued_frees, nts, ts, AVL_AFTER); list_insert_after(&tm->tm_head, ts, nts); TRIM_MAP_QINC(tm); } else if (left_over) { ts->ts_end = start; } else if (right_over) { ts->ts_start = end; } else { avl_remove(&tm->tm_queued_frees, ts); list_remove(&tm->tm_head, ts); TRIM_MAP_QDEC(tm); kmem_free(ts, sizeof (*ts)); } } static void trim_map_free_locked(trim_map_t *tm, uint64_t start, uint64_t end, uint64_t txg) { zio_t zsearch, *zs; ASSERT(MUTEX_HELD(&tm->tm_lock)); zsearch.io_offset = start; zsearch.io_size = end - start; zs = avl_find(&tm->tm_inflight_writes, &zsearch, NULL); if (zs == NULL) { trim_map_segment_add(tm, start, end, txg); return; } if (start < zs->io_offset) trim_map_free_locked(tm, start, zs->io_offset, txg); if (zs->io_offset + zs->io_size < end) trim_map_free_locked(tm, zs->io_offset + zs->io_size, end, txg); } void trim_map_free(vdev_t *vd, uint64_t offset, uint64_t size, uint64_t txg) { trim_map_t *tm = vd->vdev_trimmap; if (!zfs_trim_enabled || vd->vdev_notrim || tm == NULL) return; mutex_enter(&tm->tm_lock); trim_map_free_locked(tm, offset, TRIM_ZIO_END(vd, offset, size), txg); mutex_exit(&tm->tm_lock); } boolean_t trim_map_write_start(zio_t *zio) { vdev_t *vd = zio->io_vd; trim_map_t *tm = vd->vdev_trimmap; trim_seg_t tsearch, *ts; boolean_t left_over, right_over; uint64_t start, end; if (!zfs_trim_enabled || vd->vdev_notrim || tm == NULL) return (B_TRUE); start = zio->io_offset; end = TRIM_ZIO_END(zio->io_vd, start, zio->io_size); tsearch.ts_start = start; tsearch.ts_end = end; mutex_enter(&tm->tm_lock); /* * Checking for colliding in-flight frees. */ ts = avl_find(&tm->tm_inflight_frees, &tsearch, NULL); if (ts != NULL) { list_insert_tail(&tm->tm_pending_writes, zio); mutex_exit(&tm->tm_lock); return (B_FALSE); } ts = avl_find(&tm->tm_queued_frees, &tsearch, NULL); if (ts != NULL) { /* * Loop until all overlapping segments are removed. */ do { trim_map_segment_remove(tm, ts, start, end); ts = avl_find(&tm->tm_queued_frees, &tsearch, NULL); } while (ts != NULL); } avl_add(&tm->tm_inflight_writes, zio); mutex_exit(&tm->tm_lock); return (B_TRUE); } void trim_map_write_done(zio_t *zio) { vdev_t *vd = zio->io_vd; trim_map_t *tm = vd->vdev_trimmap; /* * Don't check for vdev_notrim, since the write could have * started before vdev_notrim was set. */ if (!zfs_trim_enabled || tm == NULL) return; mutex_enter(&tm->tm_lock); /* * Don't fail if the write isn't in the tree, since the write * could have started after vdev_notrim was set. */ if (zio->io_trim_node.avl_child[0] || zio->io_trim_node.avl_child[1] || AVL_XPARENT(&zio->io_trim_node) || tm->tm_inflight_writes.avl_root == &zio->io_trim_node) avl_remove(&tm->tm_inflight_writes, zio); mutex_exit(&tm->tm_lock); } /* * Return the oldest segment (the one with the lowest txg / time) or NULL if: * 1. The list is empty * 2. The first element's txg is greater than txgsafe * 3. The first element's txg is not greater than the txg argument and the * the first element's time is not greater than time argument */ static trim_seg_t * trim_map_first(trim_map_t *tm, uint64_t txg, uint64_t txgsafe, hrtime_t time) { trim_seg_t *ts; ASSERT(MUTEX_HELD(&tm->tm_lock)); VERIFY(txgsafe >= txg); ts = list_head(&tm->tm_head); if (ts != NULL && ts->ts_txg <= txgsafe && (ts->ts_txg <= txg || ts->ts_time <= time || tm->tm_bytes > trim_vdev_max_bytes || tm->tm_pending > trim_vdev_max_pending)) return (ts); return (NULL); } static void trim_map_vdev_commit(spa_t *spa, zio_t *zio, vdev_t *vd) { trim_map_t *tm = vd->vdev_trimmap; trim_seg_t *ts; uint64_t size, offset, txgtarget, txgsafe; hrtime_t timelimit; ASSERT(vd->vdev_ops->vdev_op_leaf); if (tm == NULL) return; timelimit = gethrtime() - trim_timeout * NANOSEC; if (vd->vdev_isl2cache) { txgsafe = UINT64_MAX; txgtarget = UINT64_MAX; } else { txgsafe = MIN(spa_last_synced_txg(spa), spa_freeze_txg(spa)); if (txgsafe > trim_txg_delay) txgtarget = txgsafe - trim_txg_delay; else txgtarget = 0; } mutex_enter(&tm->tm_lock); /* Loop until we have sent all outstanding free's */ while ((ts = trim_map_first(tm, txgtarget, txgsafe, timelimit)) != NULL) { list_remove(&tm->tm_head, ts); avl_remove(&tm->tm_queued_frees, ts); avl_add(&tm->tm_inflight_frees, ts); size = ts->ts_end - ts->ts_start; offset = ts->ts_start; TRIM_MAP_SDEC(tm, size); TRIM_MAP_QDEC(tm); /* * We drop the lock while we call zio_nowait as the IO * scheduler can result in a different IO being run e.g. * a write which would result in a recursive lock. */ mutex_exit(&tm->tm_lock); zio_nowait(zio_trim(zio, spa, vd, offset, size)); mutex_enter(&tm->tm_lock); ts = trim_map_first(tm, txgtarget, txgsafe, timelimit); } mutex_exit(&tm->tm_lock); } static void trim_map_vdev_commit_done(spa_t *spa, vdev_t *vd) { trim_map_t *tm = vd->vdev_trimmap; trim_seg_t *ts; list_t pending_writes; zio_t *zio; uint64_t start, size; void *cookie; ASSERT(vd->vdev_ops->vdev_op_leaf); if (tm == NULL) return; mutex_enter(&tm->tm_lock); if (!avl_is_empty(&tm->tm_inflight_frees)) { cookie = NULL; while ((ts = avl_destroy_nodes(&tm->tm_inflight_frees, &cookie)) != NULL) { kmem_free(ts, sizeof (*ts)); } } list_create(&pending_writes, sizeof (zio_t), offsetof(zio_t, io_trim_link)); list_move_tail(&pending_writes, &tm->tm_pending_writes); mutex_exit(&tm->tm_lock); while ((zio = list_remove_head(&pending_writes)) != NULL) { zio_vdev_io_reissue(zio); zio_execute(zio); } list_destroy(&pending_writes); } static void trim_map_commit(spa_t *spa, zio_t *zio, vdev_t *vd) { int c; if (vd == NULL) return; if (vd->vdev_ops->vdev_op_leaf) { trim_map_vdev_commit(spa, zio, vd); } else { for (c = 0; c < vd->vdev_children; c++) trim_map_commit(spa, zio, vd->vdev_child[c]); } } static void trim_map_commit_done(spa_t *spa, vdev_t *vd) { int c; if (vd == NULL) return; if (vd->vdev_ops->vdev_op_leaf) { trim_map_vdev_commit_done(spa, vd); } else { for (c = 0; c < vd->vdev_children; c++) trim_map_commit_done(spa, vd->vdev_child[c]); } } static void trim_thread(void *arg) { spa_t *spa = arg; zio_t *zio; #ifdef _KERNEL (void) snprintf(curthread->td_name, sizeof(curthread->td_name), "trim %s", spa_name(spa)); #endif for (;;) { mutex_enter(&spa->spa_trim_lock); if (spa->spa_trim_thread == NULL) { spa->spa_trim_thread = curthread; cv_signal(&spa->spa_trim_cv); mutex_exit(&spa->spa_trim_lock); thread_exit(); } (void) cv_timedwait(&spa->spa_trim_cv, &spa->spa_trim_lock, hz * trim_max_interval); mutex_exit(&spa->spa_trim_lock); zio = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL); spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); trim_map_commit(spa, zio, spa->spa_root_vdev); (void) zio_wait(zio); trim_map_commit_done(spa, spa->spa_root_vdev); spa_config_exit(spa, SCL_STATE, FTAG); } } void trim_thread_create(spa_t *spa) { if (!zfs_trim_enabled) return; mutex_init(&spa->spa_trim_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&spa->spa_trim_cv, NULL, CV_DEFAULT, NULL); mutex_enter(&spa->spa_trim_lock); spa->spa_trim_thread = thread_create(NULL, 0, trim_thread, spa, 0, &p0, TS_RUN, minclsyspri); mutex_exit(&spa->spa_trim_lock); } void trim_thread_destroy(spa_t *spa) { if (!zfs_trim_enabled) return; if (spa->spa_trim_thread == NULL) return; mutex_enter(&spa->spa_trim_lock); /* Setting spa_trim_thread to NULL tells the thread to stop. */ spa->spa_trim_thread = NULL; cv_signal(&spa->spa_trim_cv); /* The thread will set it back to != NULL on exit. */ while (spa->spa_trim_thread == NULL) cv_wait(&spa->spa_trim_cv, &spa->spa_trim_lock); spa->spa_trim_thread = NULL; mutex_exit(&spa->spa_trim_lock); cv_destroy(&spa->spa_trim_cv); mutex_destroy(&spa->spa_trim_lock); } void trim_thread_wakeup(spa_t *spa) { if (!zfs_trim_enabled) return; if (spa->spa_trim_thread == NULL) return; mutex_enter(&spa->spa_trim_lock); cv_signal(&spa->spa_trim_cv); mutex_exit(&spa->spa_trim_lock); } Index: stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev.c =================================================================== --- stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev.c (revision 274799) +++ stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev.c (revision 274800) @@ -1,3503 +1,3502 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2011, 2014 by Delphix. All rights reserved. * Copyright 2013 Martin Matuska . All rights reserved. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include SYSCTL_DECL(_vfs_zfs); SYSCTL_NODE(_vfs_zfs, OID_AUTO, vdev, CTLFLAG_RW, 0, "ZFS VDEV"); /* * Virtual device management. */ /* * The limit for ZFS to automatically increase a top-level vdev's ashift * from logical ashift to physical ashift. * * Example: one or more 512B emulation child vdevs * child->vdev_ashift = 9 (512 bytes) * child->vdev_physical_ashift = 12 (4096 bytes) * zfs_max_auto_ashift = 11 (2048 bytes) * zfs_min_auto_ashift = 9 (512 bytes) * * On pool creation or the addition of a new top-level vdev, ZFS will * increase the ashift of the top-level vdev to 2048 as limited by * zfs_max_auto_ashift. * * Example: one or more 512B emulation child vdevs * child->vdev_ashift = 9 (512 bytes) * child->vdev_physical_ashift = 12 (4096 bytes) * zfs_max_auto_ashift = 13 (8192 bytes) * zfs_min_auto_ashift = 9 (512 bytes) * * On pool creation or the addition of a new top-level vdev, ZFS will * increase the ashift of the top-level vdev to 4096 to match the * max vdev_physical_ashift. * * Example: one or more 512B emulation child vdevs * child->vdev_ashift = 9 (512 bytes) * child->vdev_physical_ashift = 9 (512 bytes) * zfs_max_auto_ashift = 13 (8192 bytes) * zfs_min_auto_ashift = 12 (4096 bytes) * * On pool creation or the addition of a new top-level vdev, ZFS will * increase the ashift of the top-level vdev to 4096 to match the * zfs_min_auto_ashift. */ static uint64_t zfs_max_auto_ashift = SPA_MAXASHIFT; static uint64_t zfs_min_auto_ashift = SPA_MINASHIFT; static int sysctl_vfs_zfs_max_auto_ashift(SYSCTL_HANDLER_ARGS) { uint64_t val; int err; val = zfs_max_auto_ashift; err = sysctl_handle_64(oidp, &val, 0, req); if (err != 0 || req->newptr == NULL) return (err); if (val > SPA_MAXASHIFT || val < zfs_min_auto_ashift) return (EINVAL); zfs_max_auto_ashift = val; return (0); } SYSCTL_PROC(_vfs_zfs, OID_AUTO, max_auto_ashift, CTLTYPE_U64 | CTLFLAG_MPSAFE | CTLFLAG_RW, 0, sizeof(uint64_t), sysctl_vfs_zfs_max_auto_ashift, "QU", "Max ashift used when optimising for logical -> physical sectors size on " "new top-level vdevs."); static int sysctl_vfs_zfs_min_auto_ashift(SYSCTL_HANDLER_ARGS) { uint64_t val; int err; val = zfs_min_auto_ashift; err = sysctl_handle_64(oidp, &val, 0, req); if (err != 0 || req->newptr == NULL) return (err); if (val < SPA_MINASHIFT || val > zfs_max_auto_ashift) return (EINVAL); zfs_min_auto_ashift = val; return (0); } SYSCTL_PROC(_vfs_zfs, OID_AUTO, min_auto_ashift, CTLTYPE_U64 | CTLFLAG_MPSAFE | CTLFLAG_RW, 0, sizeof(uint64_t), sysctl_vfs_zfs_min_auto_ashift, "QU", "Min ashift used when creating new top-level vdevs."); static vdev_ops_t *vdev_ops_table[] = { &vdev_root_ops, &vdev_raidz_ops, &vdev_mirror_ops, &vdev_replacing_ops, &vdev_spare_ops, #ifdef _KERNEL &vdev_geom_ops, #else &vdev_disk_ops, #endif &vdev_file_ops, &vdev_missing_ops, &vdev_hole_ops, NULL }; /* * When a vdev is added, it will be divided into approximately (but no * more than) this number of metaslabs. */ int metaslabs_per_vdev = 200; SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, metaslabs_per_vdev, CTLFLAG_RDTUN, &metaslabs_per_vdev, 0, "When a vdev is added, how many metaslabs the vdev should be divided into"); /* * Given a vdev type, return the appropriate ops vector. */ static vdev_ops_t * vdev_getops(const char *type) { vdev_ops_t *ops, **opspp; for (opspp = vdev_ops_table; (ops = *opspp) != NULL; opspp++) if (strcmp(ops->vdev_op_type, type) == 0) break; return (ops); } /* * Default asize function: return the MAX of psize with the asize of * all children. This is what's used by anything other than RAID-Z. */ uint64_t vdev_default_asize(vdev_t *vd, uint64_t psize) { uint64_t asize = P2ROUNDUP(psize, 1ULL << vd->vdev_top->vdev_ashift); uint64_t csize; for (int c = 0; c < vd->vdev_children; c++) { csize = vdev_psize_to_asize(vd->vdev_child[c], psize); asize = MAX(asize, csize); } return (asize); } /* * Get the minimum allocatable size. We define the allocatable size as * the vdev's asize rounded to the nearest metaslab. This allows us to * replace or attach devices which don't have the same physical size but * can still satisfy the same number of allocations. */ uint64_t vdev_get_min_asize(vdev_t *vd) { vdev_t *pvd = vd->vdev_parent; /* * If our parent is NULL (inactive spare or cache) or is the root, * just return our own asize. */ if (pvd == NULL) return (vd->vdev_asize); /* * The top-level vdev just returns the allocatable size rounded * to the nearest metaslab. */ if (vd == vd->vdev_top) return (P2ALIGN(vd->vdev_asize, 1ULL << vd->vdev_ms_shift)); /* * The allocatable space for a raidz vdev is N * sizeof(smallest child), * so each child must provide at least 1/Nth of its asize. */ if (pvd->vdev_ops == &vdev_raidz_ops) return (pvd->vdev_min_asize / pvd->vdev_children); return (pvd->vdev_min_asize); } void vdev_set_min_asize(vdev_t *vd) { vd->vdev_min_asize = vdev_get_min_asize(vd); for (int c = 0; c < vd->vdev_children; c++) vdev_set_min_asize(vd->vdev_child[c]); } vdev_t * vdev_lookup_top(spa_t *spa, uint64_t vdev) { vdev_t *rvd = spa->spa_root_vdev; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0); if (vdev < rvd->vdev_children) { ASSERT(rvd->vdev_child[vdev] != NULL); return (rvd->vdev_child[vdev]); } return (NULL); } vdev_t * vdev_lookup_by_guid(vdev_t *vd, uint64_t guid) { vdev_t *mvd; if (vd->vdev_guid == guid) return (vd); for (int c = 0; c < vd->vdev_children; c++) if ((mvd = vdev_lookup_by_guid(vd->vdev_child[c], guid)) != NULL) return (mvd); return (NULL); } void vdev_add_child(vdev_t *pvd, vdev_t *cvd) { size_t oldsize, newsize; uint64_t id = cvd->vdev_id; vdev_t **newchild; ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); ASSERT(cvd->vdev_parent == NULL); cvd->vdev_parent = pvd; if (pvd == NULL) return; ASSERT(id >= pvd->vdev_children || pvd->vdev_child[id] == NULL); oldsize = pvd->vdev_children * sizeof (vdev_t *); pvd->vdev_children = MAX(pvd->vdev_children, id + 1); newsize = pvd->vdev_children * sizeof (vdev_t *); newchild = kmem_zalloc(newsize, KM_SLEEP); if (pvd->vdev_child != NULL) { bcopy(pvd->vdev_child, newchild, oldsize); kmem_free(pvd->vdev_child, oldsize); } pvd->vdev_child = newchild; pvd->vdev_child[id] = cvd; cvd->vdev_top = (pvd->vdev_top ? pvd->vdev_top: cvd); ASSERT(cvd->vdev_top->vdev_parent->vdev_parent == NULL); /* * Walk up all ancestors to update guid sum. */ for (; pvd != NULL; pvd = pvd->vdev_parent) pvd->vdev_guid_sum += cvd->vdev_guid_sum; } void vdev_remove_child(vdev_t *pvd, vdev_t *cvd) { int c; uint_t id = cvd->vdev_id; ASSERT(cvd->vdev_parent == pvd); if (pvd == NULL) return; ASSERT(id < pvd->vdev_children); ASSERT(pvd->vdev_child[id] == cvd); pvd->vdev_child[id] = NULL; cvd->vdev_parent = NULL; for (c = 0; c < pvd->vdev_children; c++) if (pvd->vdev_child[c]) break; if (c == pvd->vdev_children) { kmem_free(pvd->vdev_child, c * sizeof (vdev_t *)); pvd->vdev_child = NULL; pvd->vdev_children = 0; } /* * Walk up all ancestors to update guid sum. */ for (; pvd != NULL; pvd = pvd->vdev_parent) pvd->vdev_guid_sum -= cvd->vdev_guid_sum; } /* * Remove any holes in the child array. */ void vdev_compact_children(vdev_t *pvd) { vdev_t **newchild, *cvd; int oldc = pvd->vdev_children; int newc; ASSERT(spa_config_held(pvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); for (int c = newc = 0; c < oldc; c++) if (pvd->vdev_child[c]) newc++; newchild = kmem_alloc(newc * sizeof (vdev_t *), KM_SLEEP); for (int c = newc = 0; c < oldc; c++) { if ((cvd = pvd->vdev_child[c]) != NULL) { newchild[newc] = cvd; cvd->vdev_id = newc++; } } kmem_free(pvd->vdev_child, oldc * sizeof (vdev_t *)); pvd->vdev_child = newchild; pvd->vdev_children = newc; } /* * Allocate and minimally initialize a vdev_t. */ vdev_t * vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid, vdev_ops_t *ops) { vdev_t *vd; vd = kmem_zalloc(sizeof (vdev_t), KM_SLEEP); 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) { if (spa->spa_root_vdev == vd) { /* * The root vdev's guid will also be the pool guid, * which must be unique among all pools. */ guid = spa_generate_guid(NULL); } else { /* * Any other vdev's guid must be unique within the pool. */ guid = spa_generate_guid(spa); } ASSERT(!spa_guid_exists(spa_guid(spa), guid)); } vd->vdev_spa = spa; vd->vdev_id = id; vd->vdev_guid = guid; vd->vdev_guid_sum = guid; vd->vdev_ops = ops; vd->vdev_state = VDEV_STATE_CLOSED; vd->vdev_ishole = (ops == &vdev_hole_ops); 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++) { vd->vdev_dtl[t] = range_tree_create(NULL, NULL, &vd->vdev_dtl_lock); } txg_list_create(&vd->vdev_ms_list, offsetof(struct metaslab, ms_txg_node)); txg_list_create(&vd->vdev_dtl_list, offsetof(struct vdev, vdev_dtl_node)); vd->vdev_stat.vs_timestamp = gethrtime(); vdev_queue_init(vd); vdev_cache_init(vd); return (vd); } /* * Allocate a new vdev. The 'alloctype' is used to control whether we are * creating a new vdev or loading an existing one - the behavior is slightly * different for each case. */ int vdev_alloc(spa_t *spa, vdev_t **vdp, nvlist_t *nv, vdev_t *parent, uint_t id, int alloctype) { vdev_ops_t *ops; char *type; uint64_t guid = 0, islog, nparity; vdev_t *vd; ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) != 0) return (SET_ERROR(EINVAL)); if ((ops = vdev_getops(type)) == NULL) return (SET_ERROR(EINVAL)); /* * If this is a load, get the vdev guid from the nvlist. * Otherwise, vdev_alloc_common() will generate one for us. */ if (alloctype == VDEV_ALLOC_LOAD) { uint64_t label_id; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID, &label_id) || label_id != id) return (SET_ERROR(EINVAL)); if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } else if (alloctype == VDEV_ALLOC_SPARE) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } else if (alloctype == VDEV_ALLOC_L2CACHE) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } else if (alloctype == VDEV_ALLOC_ROOTPOOL) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } /* * The first allocated vdev must be of type 'root'. */ if (ops != &vdev_root_ops && spa->spa_root_vdev == NULL) return (SET_ERROR(EINVAL)); /* * Determine whether we're a log vdev. */ islog = 0; (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_LOG, &islog); if (islog && spa_version(spa) < SPA_VERSION_SLOGS) return (SET_ERROR(ENOTSUP)); if (ops == &vdev_hole_ops && spa_version(spa) < SPA_VERSION_HOLES) return (SET_ERROR(ENOTSUP)); /* * Set the nparity property for RAID-Z vdevs. */ nparity = -1ULL; if (ops == &vdev_raidz_ops) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY, &nparity) == 0) { if (nparity == 0 || nparity > VDEV_RAIDZ_MAXPARITY) return (SET_ERROR(EINVAL)); /* * Previous versions could only support 1 or 2 parity * device. */ if (nparity > 1 && spa_version(spa) < SPA_VERSION_RAIDZ2) return (SET_ERROR(ENOTSUP)); if (nparity > 2 && spa_version(spa) < SPA_VERSION_RAIDZ3) return (SET_ERROR(ENOTSUP)); } else { /* * We require the parity to be specified for SPAs that * support multiple parity levels. */ if (spa_version(spa) >= SPA_VERSION_RAIDZ2) return (SET_ERROR(EINVAL)); /* * Otherwise, we default to 1 parity device for RAID-Z. */ nparity = 1; } } else { nparity = 0; } ASSERT(nparity != -1ULL); vd = vdev_alloc_common(spa, id, guid, ops); vd->vdev_islog = islog; vd->vdev_nparity = nparity; if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &vd->vdev_path) == 0) vd->vdev_path = spa_strdup(vd->vdev_path); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_DEVID, &vd->vdev_devid) == 0) vd->vdev_devid = spa_strdup(vd->vdev_devid); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PHYS_PATH, &vd->vdev_physpath) == 0) vd->vdev_physpath = spa_strdup(vd->vdev_physpath); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_FRU, &vd->vdev_fru) == 0) vd->vdev_fru = spa_strdup(vd->vdev_fru); /* * Set the whole_disk property. If it's not specified, leave the value * as -1. */ if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, &vd->vdev_wholedisk) != 0) vd->vdev_wholedisk = -1ULL; /* * Look for the 'not present' flag. This will only be set if the device * was not present at the time of import. */ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, &vd->vdev_not_present); /* * Get the alignment requirement. */ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASHIFT, &vd->vdev_ashift); /* * Retrieve the vdev creation time. */ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_CREATE_TXG, &vd->vdev_crtxg); /* * If we're a top-level vdev, try to load the allocation parameters. */ if (parent && !parent->vdev_parent && (alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_SPLIT)) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_METASLAB_ARRAY, &vd->vdev_ms_array); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_METASLAB_SHIFT, &vd->vdev_ms_shift); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASIZE, &vd->vdev_asize); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVING, &vd->vdev_removing); } if (parent && !parent->vdev_parent && alloctype != VDEV_ALLOC_ATTACH) { ASSERT(alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_ADD || alloctype == VDEV_ALLOC_SPLIT || alloctype == VDEV_ALLOC_ROOTPOOL); vd->vdev_mg = metaslab_group_create(islog ? spa_log_class(spa) : spa_normal_class(spa), vd); } /* * If we're a leaf vdev, try to load the DTL object and other state. */ if (vd->vdev_ops->vdev_op_leaf && (alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_L2CACHE || alloctype == VDEV_ALLOC_ROOTPOOL)) { if (alloctype == VDEV_ALLOC_LOAD) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DTL, &vd->vdev_dtl_object); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_UNSPARE, &vd->vdev_unspare); } if (alloctype == VDEV_ALLOC_ROOTPOOL) { uint64_t spare = 0; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_SPARE, &spare) == 0 && spare) spa_spare_add(vd); } (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_OFFLINE, &vd->vdev_offline); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_RESILVER_TXG, &vd->vdev_resilver_txg); /* * When importing a pool, we want to ignore the persistent fault * state, as the diagnosis made on another system may not be * valid in the current context. Local vdevs will * remain in the faulted state. */ if (spa_load_state(spa) == SPA_LOAD_OPEN) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_FAULTED, &vd->vdev_faulted); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DEGRADED, &vd->vdev_degraded); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVED, &vd->vdev_removed); if (vd->vdev_faulted || vd->vdev_degraded) { char *aux; vd->vdev_label_aux = VDEV_AUX_ERR_EXCEEDED; if (nvlist_lookup_string(nv, ZPOOL_CONFIG_AUX_STATE, &aux) == 0 && strcmp(aux, "external") == 0) vd->vdev_label_aux = VDEV_AUX_EXTERNAL; } } } /* * Add ourselves to the parent's list of children. */ vdev_add_child(parent, vd); *vdp = vd; return (0); } void vdev_free(vdev_t *vd) { spa_t *spa = vd->vdev_spa; /* * vdev_free() implies closing the vdev first. This is simpler than * trying to ensure complicated semantics for all callers. */ vdev_close(vd); ASSERT(!list_link_active(&vd->vdev_config_dirty_node)); ASSERT(!list_link_active(&vd->vdev_state_dirty_node)); /* * Free all children. */ for (int c = 0; c < vd->vdev_children; c++) vdev_free(vd->vdev_child[c]); ASSERT(vd->vdev_child == NULL); ASSERT(vd->vdev_guid_sum == vd->vdev_guid); /* * Discard allocation state. */ if (vd->vdev_mg != NULL) { vdev_metaslab_fini(vd); metaslab_group_destroy(vd->vdev_mg); } ASSERT0(vd->vdev_stat.vs_space); ASSERT0(vd->vdev_stat.vs_dspace); ASSERT0(vd->vdev_stat.vs_alloc); /* * Remove this vdev from its parent's child list. */ vdev_remove_child(vd->vdev_parent, vd); ASSERT(vd->vdev_parent == NULL); /* * Clean up vdev structure. */ vdev_queue_fini(vd); vdev_cache_fini(vd); if (vd->vdev_path) spa_strfree(vd->vdev_path); if (vd->vdev_devid) spa_strfree(vd->vdev_devid); if (vd->vdev_physpath) spa_strfree(vd->vdev_physpath); if (vd->vdev_fru) spa_strfree(vd->vdev_fru); if (vd->vdev_isspare) spa_spare_remove(vd); if (vd->vdev_isl2cache) spa_l2cache_remove(vd); txg_list_destroy(&vd->vdev_ms_list); txg_list_destroy(&vd->vdev_dtl_list); mutex_enter(&vd->vdev_dtl_lock); space_map_close(vd->vdev_dtl_sm); for (int t = 0; t < DTL_TYPES; t++) { range_tree_vacate(vd->vdev_dtl[t], NULL, NULL); range_tree_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); if (vd == spa->spa_root_vdev) spa->spa_root_vdev = NULL; kmem_free(vd, sizeof (vdev_t)); } /* * Transfer top-level vdev state from svd to tvd. */ static void vdev_top_transfer(vdev_t *svd, vdev_t *tvd) { spa_t *spa = svd->vdev_spa; metaslab_t *msp; vdev_t *vd; int t; ASSERT(tvd == tvd->vdev_top); tvd->vdev_ms_array = svd->vdev_ms_array; tvd->vdev_ms_shift = svd->vdev_ms_shift; tvd->vdev_ms_count = svd->vdev_ms_count; svd->vdev_ms_array = 0; 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; svd->vdev_mg = NULL; svd->vdev_ms = NULL; if (tvd->vdev_mg != NULL) tvd->vdev_mg->mg_vd = tvd; tvd->vdev_stat.vs_alloc = svd->vdev_stat.vs_alloc; tvd->vdev_stat.vs_space = svd->vdev_stat.vs_space; tvd->vdev_stat.vs_dspace = svd->vdev_stat.vs_dspace; svd->vdev_stat.vs_alloc = 0; svd->vdev_stat.vs_space = 0; svd->vdev_stat.vs_dspace = 0; for (t = 0; t < TXG_SIZE; t++) { while ((msp = txg_list_remove(&svd->vdev_ms_list, t)) != NULL) (void) txg_list_add(&tvd->vdev_ms_list, msp, t); while ((vd = txg_list_remove(&svd->vdev_dtl_list, t)) != NULL) (void) txg_list_add(&tvd->vdev_dtl_list, vd, t); if (txg_list_remove_this(&spa->spa_vdev_txg_list, svd, t)) (void) txg_list_add(&spa->spa_vdev_txg_list, tvd, t); } if (list_link_active(&svd->vdev_config_dirty_node)) { vdev_config_clean(svd); vdev_config_dirty(tvd); } if (list_link_active(&svd->vdev_state_dirty_node)) { vdev_state_clean(svd); vdev_state_dirty(tvd); } tvd->vdev_deflate_ratio = svd->vdev_deflate_ratio; svd->vdev_deflate_ratio = 0; tvd->vdev_islog = svd->vdev_islog; svd->vdev_islog = 0; } static void vdev_top_update(vdev_t *tvd, vdev_t *vd) { if (vd == NULL) return; vd->vdev_top = tvd; for (int c = 0; c < vd->vdev_children; c++) vdev_top_update(tvd, vd->vdev_child[c]); } /* * Add a mirror/replacing vdev above an existing vdev. */ vdev_t * vdev_add_parent(vdev_t *cvd, vdev_ops_t *ops) { spa_t *spa = cvd->vdev_spa; vdev_t *pvd = cvd->vdev_parent; vdev_t *mvd; ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); mvd = vdev_alloc_common(spa, cvd->vdev_id, 0, ops); 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_logical_ashift = cvd->vdev_logical_ashift; mvd->vdev_physical_ashift = cvd->vdev_physical_ashift; mvd->vdev_state = cvd->vdev_state; mvd->vdev_crtxg = cvd->vdev_crtxg; vdev_remove_child(pvd, cvd); vdev_add_child(pvd, mvd); cvd->vdev_id = mvd->vdev_children; vdev_add_child(mvd, cvd); vdev_top_update(cvd->vdev_top, cvd->vdev_top); if (mvd == mvd->vdev_top) vdev_top_transfer(cvd, mvd); return (mvd); } /* * Remove a 1-way mirror/replacing vdev from the tree. */ void vdev_remove_parent(vdev_t *cvd) { vdev_t *mvd = cvd->vdev_parent; vdev_t *pvd = mvd->vdev_parent; ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); ASSERT(mvd->vdev_children == 1); ASSERT(mvd->vdev_ops == &vdev_mirror_ops || mvd->vdev_ops == &vdev_replacing_ops || mvd->vdev_ops == &vdev_spare_ops); cvd->vdev_ashift = mvd->vdev_ashift; cvd->vdev_logical_ashift = mvd->vdev_logical_ashift; cvd->vdev_physical_ashift = mvd->vdev_physical_ashift; 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) { uint64_t guid_delta = mvd->vdev_guid - cvd->vdev_guid; cvd->vdev_orig_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); if (cvd == cvd->vdev_top) vdev_top_transfer(mvd, cvd); ASSERT(mvd->vdev_children == 0); vdev_free(mvd); } int vdev_metaslab_init(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; uint64_t m; uint64_t oldc = vd->vdev_ms_count; uint64_t newc = vd->vdev_asize >> vd->vdev_ms_shift; metaslab_t **mspp; int error; ASSERT(txg == 0 || spa_config_held(spa, SCL_ALLOC, RW_WRITER)); /* * This vdev is not being allocated from yet or is a hole. */ if (vd->vdev_ms_shift == 0) return (0); ASSERT(!vd->vdev_ishole); /* * Compute the raidz-deflation ratio. Note, we hard-code * in 128k (1 << 17) because it is the current "typical" blocksize. * Even if SPA_MAXBLOCKSIZE changes, this algorithm must never change, * or we will inconsistently account for existing bp's. */ vd->vdev_deflate_ratio = (1 << 17) / (vdev_psize_to_asize(vd, 1 << 17) >> SPA_MINBLOCKSHIFT); ASSERT(oldc <= newc); mspp = kmem_zalloc(newc * sizeof (*mspp), KM_SLEEP); if (oldc != 0) { bcopy(vd->vdev_ms, mspp, oldc * sizeof (*mspp)); kmem_free(vd->vdev_ms, oldc * sizeof (*mspp)); } vd->vdev_ms = mspp; vd->vdev_ms_count = newc; for (m = oldc; m < newc; m++) { uint64_t object = 0; if (txg == 0) { error = dmu_read(mos, vd->vdev_ms_array, m * sizeof (uint64_t), sizeof (uint64_t), &object, DMU_READ_PREFETCH); if (error) return (error); } vd->vdev_ms[m] = metaslab_init(vd->vdev_mg, m, object, txg); } if (txg == 0) spa_config_enter(spa, SCL_ALLOC, FTAG, RW_WRITER); /* * If the vdev is being removed we don't activate * the metaslabs since we want to ensure that no new * allocations are performed on this device. */ if (oldc == 0 && !vd->vdev_removing) metaslab_group_activate(vd->vdev_mg); if (txg == 0) spa_config_exit(spa, SCL_ALLOC, FTAG); return (0); } void vdev_metaslab_fini(vdev_t *vd) { uint64_t m; uint64_t count = vd->vdev_ms_count; if (vd->vdev_ms != NULL) { metaslab_group_passivate(vd->vdev_mg); for (m = 0; m < count; m++) { metaslab_t *msp = vd->vdev_ms[m]; if (msp != NULL) metaslab_fini(msp); } kmem_free(vd->vdev_ms, count * sizeof (metaslab_t *)); vd->vdev_ms = NULL; } } typedef struct vdev_probe_stats { boolean_t vps_readable; boolean_t vps_writeable; int vps_flags; } vdev_probe_stats_t; static void vdev_probe_done(zio_t *zio) { spa_t *spa = zio->io_spa; vdev_t *vd = zio->io_vd; vdev_probe_stats_t *vps = zio->io_private; ASSERT(vd->vdev_probe_zio != NULL); if (zio->io_type == ZIO_TYPE_READ) { if (zio->io_error == 0) vps->vps_readable = 1; if (zio->io_error == 0 && spa_writeable(spa)) { zio_nowait(zio_write_phys(vd->vdev_probe_zio, vd, zio->io_offset, zio->io_size, zio->io_data, ZIO_CHECKSUM_OFF, vdev_probe_done, vps, ZIO_PRIORITY_SYNC_WRITE, vps->vps_flags, B_TRUE)); } else { zio_buf_free(zio->io_data, zio->io_size); } } else if (zio->io_type == ZIO_TYPE_WRITE) { if (zio->io_error == 0) vps->vps_writeable = 1; zio_buf_free(zio->io_data, zio->io_size); } else if (zio->io_type == ZIO_TYPE_NULL) { zio_t *pio; vd->vdev_cant_read |= !vps->vps_readable; vd->vdev_cant_write |= !vps->vps_writeable; if (vdev_readable(vd) && (vdev_writeable(vd) || !spa_writeable(spa))) { zio->io_error = 0; } else { ASSERT(zio->io_error != 0); zfs_ereport_post(FM_EREPORT_ZFS_PROBE_FAILURE, spa, vd, NULL, 0, 0); zio->io_error = SET_ERROR(ENXIO); } mutex_enter(&vd->vdev_probe_lock); ASSERT(vd->vdev_probe_zio == zio); vd->vdev_probe_zio = NULL; mutex_exit(&vd->vdev_probe_lock); while ((pio = zio_walk_parents(zio)) != NULL) if (!vdev_accessible(vd, pio)) pio->io_error = SET_ERROR(ENXIO); kmem_free(vps, sizeof (*vps)); } } /* * Determine whether this device is accessible. * * Read and write to several known locations: the pad regions of each * vdev label but the first, which we leave alone in case it contains * a VTOC. */ zio_t * vdev_probe(vdev_t *vd, zio_t *zio) { spa_t *spa = vd->vdev_spa; vdev_probe_stats_t *vps = NULL; zio_t *pio; ASSERT(vd->vdev_ops->vdev_op_leaf); /* * Don't probe the probe. */ if (zio && (zio->io_flags & ZIO_FLAG_PROBE)) return (NULL); /* * To prevent 'probe storms' when a device fails, we create * just one probe i/o at a time. All zios that want to probe * this vdev will become parents of the probe io. */ mutex_enter(&vd->vdev_probe_lock); if ((pio = vd->vdev_probe_zio) == NULL) { vps = kmem_zalloc(sizeof (*vps), KM_SLEEP); vps->vps_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_PROBE | ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_AGGREGATE | ZIO_FLAG_TRYHARD; if (spa_config_held(spa, SCL_ZIO, RW_WRITER)) { /* * vdev_cant_read and vdev_cant_write can only * transition from TRUE to FALSE when we have the * SCL_ZIO lock as writer; otherwise they can only * transition from FALSE to TRUE. This ensures that * any zio looking at these values can assume that * failures persist for the life of the I/O. That's * important because when a device has intermittent * connectivity problems, we want to ensure that * they're ascribed to the device (ENXIO) and not * the zio (EIO). * * Since we hold SCL_ZIO as writer here, clear both * values so the probe can reevaluate from first * principles. */ vps->vps_flags |= ZIO_FLAG_CONFIG_WRITER; vd->vdev_cant_read = B_FALSE; vd->vdev_cant_write = B_FALSE; } vd->vdev_probe_zio = pio = zio_null(NULL, spa, vd, vdev_probe_done, vps, vps->vps_flags | ZIO_FLAG_DONT_PROPAGATE); /* * We can't change the vdev state in this context, so we * kick off an async task to do it on our behalf. */ if (zio != NULL) { vd->vdev_probe_wanted = B_TRUE; spa_async_request(spa, SPA_ASYNC_PROBE); } } if (zio != NULL) zio_add_child(zio, pio); mutex_exit(&vd->vdev_probe_lock); if (vps == NULL) { ASSERT(zio != NULL); return (NULL); } for (int 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)), VDEV_PAD_SIZE, zio_buf_alloc(VDEV_PAD_SIZE), ZIO_CHECKSUM_OFF, vdev_probe_done, vps, ZIO_PRIORITY_SYNC_READ, vps->vps_flags, B_TRUE)); } if (zio == NULL) return (pio); zio_nowait(pio); return (NULL); } static void vdev_open_child(void *arg) { vdev_t *vd = arg; vd->vdev_open_thread = curthread; vd->vdev_open_error = vdev_open(vd); vd->vdev_open_thread = NULL; } boolean_t vdev_uses_zvols(vdev_t *vd) { if (vd->vdev_path && strncmp(vd->vdev_path, ZVOL_DIR, strlen(ZVOL_DIR)) == 0) return (B_TRUE); for (int c = 0; c < vd->vdev_children; c++) if (vdev_uses_zvols(vd->vdev_child[c])) return (B_TRUE); return (B_FALSE); } void vdev_open_children(vdev_t *vd) { taskq_t *tq; int children = vd->vdev_children; /* * in order to handle pools on top of zvols, do the opens * in a single thread so that the same thread holds the * spa_namespace_lock */ if (B_TRUE || vdev_uses_zvols(vd)) { for (int 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++) VERIFY(taskq_dispatch(tq, vdev_open_child, vd->vdev_child[c], TQ_SLEEP) != 0); taskq_destroy(tq); } /* * Prepare a virtual device for access. */ int vdev_open(vdev_t *vd) { spa_t *spa = vd->vdev_spa; int error; uint64_t osize = 0; uint64_t max_osize = 0; uint64_t asize, max_asize, psize; uint64_t logical_ashift = 0; uint64_t physical_ashift = 0; ASSERT(vd->vdev_open_thread == curthread || 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); vd->vdev_stat.vs_aux = VDEV_AUX_NONE; vd->vdev_cant_read = B_FALSE; vd->vdev_cant_write = B_FALSE; + vd->vdev_notrim = B_FALSE; vd->vdev_min_asize = vdev_get_min_asize(vd); /* * If this vdev is not removed, check its fault status. If it's * faulted, bail out of the open. */ if (!vd->vdev_removed && vd->vdev_faulted) { ASSERT(vd->vdev_children == 0); ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED || vd->vdev_label_aux == VDEV_AUX_EXTERNAL); vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED, vd->vdev_label_aux); return (SET_ERROR(ENXIO)); } else if (vd->vdev_offline) { ASSERT(vd->vdev_children == 0); vdev_set_state(vd, B_TRUE, VDEV_STATE_OFFLINE, VDEV_AUX_NONE); return (SET_ERROR(ENXIO)); } error = vd->vdev_ops->vdev_op_open(vd, &osize, &max_osize, &logical_ashift, &physical_ashift); /* * Reset the vdev_reopening flag so that we actually close * the vdev on error. */ vd->vdev_reopening = B_FALSE; if (zio_injection_enabled && error == 0) error = zio_handle_device_injection(vd, NULL, ENXIO); if (error) { if (vd->vdev_removed && vd->vdev_stat.vs_aux != VDEV_AUX_OPEN_FAILED) vd->vdev_removed = B_FALSE; vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, vd->vdev_stat.vs_aux); return (error); } vd->vdev_removed = B_FALSE; /* * Recheck the faulted flag now that we have confirmed that * the vdev is accessible. If we're faulted, bail. */ if (vd->vdev_faulted) { ASSERT(vd->vdev_children == 0); ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED || vd->vdev_label_aux == VDEV_AUX_EXTERNAL); vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED, vd->vdev_label_aux); return (SET_ERROR(ENXIO)); } if (vd->vdev_degraded) { ASSERT(vd->vdev_children == 0); vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED, VDEV_AUX_ERR_EXCEEDED); } else { vdev_set_state(vd, B_TRUE, VDEV_STATE_HEALTHY, 0); } /* * For hole or missing vdevs we just return success. */ if (vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops) return (0); - if (vd->vdev_ops->vdev_op_leaf) { - vd->vdev_notrim = B_FALSE; + if (zfs_trim_enabled && !vd->vdev_notrim && vd->vdev_ops->vdev_op_leaf) trim_map_create(vd); - } for (int 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); break; } } 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) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_TOO_SMALL); return (SET_ERROR(EOVERFLOW)); } 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)) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_TOO_SMALL); return (SET_ERROR(EOVERFLOW)); } psize = 0; asize = osize; max_asize = max_osize; } vd->vdev_psize = psize; /* * Make sure the allocatable size hasn't shrunk. */ if (asize < vd->vdev_min_asize) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_LABEL); return (SET_ERROR(EINVAL)); } vd->vdev_physical_ashift = MAX(physical_ashift, vd->vdev_physical_ashift); vd->vdev_logical_ashift = MAX(logical_ashift, vd->vdev_logical_ashift); vd->vdev_ashift = MAX(vd->vdev_logical_ashift, vd->vdev_ashift); if (vd->vdev_logical_ashift > SPA_MAXASHIFT) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_ASHIFT_TOO_BIG); return (EINVAL); } 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. */ vd->vdev_asize = asize; vd->vdev_max_asize = max_asize; } else { /* * Make sure the alignment requirement hasn't increased. */ if (vd->vdev_ashift > vd->vdev_top->vdev_ashift && vd->vdev_ops->vdev_op_leaf) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_LABEL); return (EINVAL); } vd->vdev_max_asize = max_asize; } /* * If all children are healthy and the asize has increased, * then we've experienced dynamic LUN growth. If automatic * expansion is enabled then use the additional space. */ if (vd->vdev_state == VDEV_STATE_HEALTHY && asize > vd->vdev_asize && (vd->vdev_expanding || spa->spa_autoexpand)) vd->vdev_asize = asize; vdev_set_min_asize(vd); /* * Ensure we can issue some IO before declaring the * vdev open for business. */ if (vd->vdev_ops->vdev_op_leaf && (error = zio_wait(vdev_probe(vd, NULL))) != 0) { vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED, VDEV_AUX_ERR_EXCEEDED); return (error); } /* * 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. */ 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); } /* * Called once the vdevs are all opened, this routine validates the label * 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, boolean_t strict) { spa_t *spa = vd->vdev_spa; nvlist_t *label; uint64_t guid = 0, top_guid; uint64_t state; for (int c = 0; c < vd->vdev_children; c++) if (vdev_validate(vd->vdev_child[c], strict) != 0) return (SET_ERROR(EBADF)); /* * If the device has already failed, or was marked offline, don't do * any further validation. Otherwise, label I/O will fail and we will * overwrite the previous state. */ 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, txg)) == NULL) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_LABEL); return (0); } /* * Determine if this vdev has been split off into another * pool. If so, then refuse to open it. */ if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_SPLIT_GUID, &aux_guid) == 0 && aux_guid == spa_guid(spa)) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_SPLIT_POOL); nvlist_free(label); return (0); } 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); return (0); } if (nvlist_lookup_nvlist(label, ZPOOL_CONFIG_VDEV_TREE, &nvl) != 0 || nvlist_lookup_uint64(nvl, ZPOOL_CONFIG_ORIG_GUID, &aux_guid) != 0) aux_guid = 0; /* * If this vdev just became a top-level vdev because its * sibling was detached, it will have adopted the parent's * vdev guid -- but the label may or may not be on disk yet. * Fortunately, either version of the label will have the * same top guid, so if we're a top-level vdev, we can * safely compare to that instead. * * If we split this vdev off instead, then we also check the * original pool's guid. We don't want to consider the vdev * corrupt if it is partway through a split operation. */ if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) != 0 || nvlist_lookup_uint64(label, ZPOOL_CONFIG_TOP_GUID, &top_guid) != 0 || ((vd->vdev_guid != guid && vd->vdev_guid != aux_guid) && (vd->vdev_guid != top_guid || vd != vd->vdev_top))) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); return (0); } if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE, &state) != 0) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); return (0); } nvlist_free(label); /* * If this is a verbatim import, no need to check the * state of the pool. */ if (!(spa->spa_import_flags & ZFS_IMPORT_VERBATIM) && spa_load_state(spa) == SPA_LOAD_OPEN && state != POOL_STATE_ACTIVE) return (SET_ERROR(EBADF)); /* * If we were able to open and validate a vdev that was * previously marked permanently unavailable, clear that state * now. */ if (vd->vdev_not_present) vd->vdev_not_present = 0; } return (0); } /* * Close a virtual device. */ void vdev_close(vdev_t *vd) { spa_t *spa = vd->vdev_spa; vdev_t *pvd = vd->vdev_parent; ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); /* * If our parent is reopening, then we are as well, unless we are * going offline. */ if (pvd != NULL && pvd->vdev_reopening) vd->vdev_reopening = (pvd->vdev_reopening && !vd->vdev_offline); vd->vdev_ops->vdev_op_close(vd); vdev_cache_purge(vd); if (vd->vdev_ops->vdev_op_leaf) trim_map_destroy(vd); /* * We record the previous state before we close it, so that if we are * doing a reopen(), we don't generate FMA ereports if we notice that * it's still faulted. */ vd->vdev_prevstate = vd->vdev_state; if (vd->vdev_offline) vd->vdev_state = VDEV_STATE_OFFLINE; else vd->vdev_state = VDEV_STATE_CLOSED; vd->vdev_stat.vs_aux = VDEV_AUX_NONE; } void vdev_hold(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_is_root(spa)); if (spa->spa_state == POOL_STATE_UNINITIALIZED) return; for (int c = 0; c < vd->vdev_children; c++) vdev_hold(vd->vdev_child[c]); if (vd->vdev_ops->vdev_op_leaf) vd->vdev_ops->vdev_op_hold(vd); } void vdev_rele(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_is_root(spa)); for (int c = 0; c < vd->vdev_children; c++) vdev_rele(vd->vdev_child[c]); if (vd->vdev_ops->vdev_op_leaf) vd->vdev_ops->vdev_op_rele(vd); } /* * Reopen all interior vdevs and any unopened leaves. We don't actually * reopen leaf vdevs which had previously been opened as they might deadlock * on the spa_config_lock. Instead we only obtain the leaf's physical size. * If the leaf has never been opened then open it, as usual. */ void vdev_reopen(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); /* set the reopening flag unless we're taking the vdev offline */ vd->vdev_reopening = !vd->vdev_offline; vdev_close(vd); (void) vdev_open(vd); /* * Call vdev_validate() here to make sure we have the same device. * Otherwise, a device with an invalid label could be successfully * opened in response to vdev_reopen(). */ if (vd->vdev_aux) { (void) vdev_validate_aux(vd); if (vdev_readable(vd) && vdev_writeable(vd) && vd->vdev_aux == &spa->spa_l2cache && !l2arc_vdev_present(vd)) l2arc_add_vdev(spa, vd); } else { (void) vdev_validate(vd, B_TRUE); } /* * Reassess parent vdev's health. */ vdev_propagate_state(vd); } int vdev_create(vdev_t *vd, uint64_t txg, boolean_t isreplacing) { int error; /* * Normally, partial opens (e.g. of a mirror) are allowed. * For a create, however, we want to fail the request if * there are any components we can't open. */ error = vdev_open(vd); if (error || vd->vdev_state != VDEV_STATE_HEALTHY) { vdev_close(vd); return (error ? error : ENXIO); } /* * Recursively load DTLs and initialize all labels. */ if ((error = vdev_dtl_load(vd)) != 0 || (error = vdev_label_init(vd, txg, isreplacing ? VDEV_LABEL_REPLACE : VDEV_LABEL_CREATE)) != 0) { vdev_close(vd); return (error); } return (0); } void vdev_metaslab_set_size(vdev_t *vd) { /* * Aim for roughly metaslabs_per_vdev (default 200) metaslabs per vdev. */ vd->vdev_ms_shift = highbit64(vd->vdev_asize / metaslabs_per_vdev); vd->vdev_ms_shift = MAX(vd->vdev_ms_shift, SPA_MAXBLOCKSHIFT); } /* * Maximize performance by inflating the configured ashift for top level * vdevs to be as close to the physical ashift as possible while maintaining * administrator defined limits and ensuring it doesn't go below the * logical ashift. */ void vdev_ashift_optimize(vdev_t *vd) { if (vd == vd->vdev_top) { if (vd->vdev_ashift < vd->vdev_physical_ashift) { vd->vdev_ashift = MIN( MAX(zfs_max_auto_ashift, vd->vdev_ashift), MAX(zfs_min_auto_ashift, vd->vdev_physical_ashift)); } else { /* * Unusual case where logical ashift > physical ashift * so we can't cap the calculated ashift based on max * ashift as that would cause failures. * We still check if we need to increase it to match * the min ashift. */ vd->vdev_ashift = MAX(zfs_min_auto_ashift, vd->vdev_ashift); } } } void vdev_dirty(vdev_t *vd, int flags, void *arg, uint64_t txg) { 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); if (flags & VDD_DTL) (void) txg_list_add(&vd->vdev_dtl_list, arg, txg); (void) txg_list_add(&vd->vdev_spa->spa_vdev_txg_list, vd, txg); } void vdev_dirty_leaves(vdev_t *vd, int flags, uint64_t txg) { for (int c = 0; c < vd->vdev_children; c++) vdev_dirty_leaves(vd->vdev_child[c], flags, txg); if (vd->vdev_ops->vdev_op_leaf) vdev_dirty(vd->vdev_top, flags, 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 four 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(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size) { range_tree_t *rt = vd->vdev_dtl[t]; ASSERT(t < DTL_TYPES); ASSERT(vd != vd->vdev_spa->spa_root_vdev); ASSERT(spa_writeable(vd->vdev_spa)); mutex_enter(rt->rt_lock); if (!range_tree_contains(rt, txg, size)) range_tree_add(rt, txg, size); mutex_exit(rt->rt_lock); } boolean_t vdev_dtl_contains(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size) { range_tree_t *rt = vd->vdev_dtl[t]; boolean_t dirty = B_FALSE; ASSERT(t < DTL_TYPES); ASSERT(vd != vd->vdev_spa->spa_root_vdev); mutex_enter(rt->rt_lock); if (range_tree_space(rt) != 0) dirty = range_tree_contains(rt, txg, size); mutex_exit(rt->rt_lock); return (dirty); } boolean_t vdev_dtl_empty(vdev_t *vd, vdev_dtl_type_t t) { range_tree_t *rt = vd->vdev_dtl[t]; boolean_t empty; mutex_enter(rt->rt_lock); empty = (range_tree_space(rt) == 0); mutex_exit(rt->rt_lock); return (empty); } /* * Returns the lowest txg in the DTL range. */ static uint64_t vdev_dtl_min(vdev_t *vd) { range_seg_t *rs; ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock)); ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0); ASSERT0(vd->vdev_children); rs = avl_first(&vd->vdev_dtl[DTL_MISSING]->rt_root); return (rs->rs_start - 1); } /* * Returns the highest txg in the DTL. */ static uint64_t vdev_dtl_max(vdev_t *vd) { range_seg_t *rs; ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock)); ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0); ASSERT0(vd->vdev_children); rs = avl_last(&vd->vdev_dtl[DTL_MISSING]->rt_root); return (rs->rs_end); } /* * Determine if a resilvering vdev should remove any DTL entries from * its range. If the vdev was resilvering for the entire duration of the * scan then it should excise that range from its DTLs. Otherwise, this * vdev is considered partially resilvered and should leave its DTL * entries intact. The comment in vdev_dtl_reassess() describes how we * excise the DTLs. */ static boolean_t vdev_dtl_should_excise(vdev_t *vd) { spa_t *spa = vd->vdev_spa; dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan; ASSERT0(scn->scn_phys.scn_errors); ASSERT0(vd->vdev_children); if (vd->vdev_resilver_txg == 0 || range_tree_space(vd->vdev_dtl[DTL_MISSING]) == 0) return (B_TRUE); /* * When a resilver is initiated the scan will assign the scn_max_txg * value to the highest txg value that exists in all DTLs. If this * device's max DTL is not part of this scan (i.e. it is not in * the range (scn_min_txg, scn_max_txg] then it is not eligible * for excision. */ if (vdev_dtl_max(vd) <= scn->scn_phys.scn_max_txg) { ASSERT3U(scn->scn_phys.scn_min_txg, <=, vdev_dtl_min(vd)); ASSERT3U(scn->scn_phys.scn_min_txg, <, vd->vdev_resilver_txg); ASSERT3U(vd->vdev_resilver_txg, <=, scn->scn_phys.scn_max_txg); return (B_TRUE); } return (B_FALSE); } /* * Reassess DTLs after a config change or scrub completion. */ void vdev_dtl_reassess(vdev_t *vd, uint64_t txg, uint64_t scrub_txg, int scrub_done) { spa_t *spa = vd->vdev_spa; avl_tree_t reftree; int minref; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 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 || vd->vdev_ishole || vd->vdev_aux) return; if (vd->vdev_ops->vdev_op_leaf) { dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan; mutex_enter(&vd->vdev_dtl_lock); /* * If we've completed a scan cleanly then determine * if this vdev should remove any DTLs. We only want to * excise regions on vdevs that were available during * the entire duration of this scan. */ if (scrub_txg != 0 && (spa->spa_scrub_started || (scn != NULL && scn->scn_phys.scn_errors == 0)) && vdev_dtl_should_excise(vd)) { /* * We completed a scrub up to scrub_txg. If we * did it without rebooting, then the scrub dtl * 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_reftree_create(&reftree); space_reftree_add_map(&reftree, vd->vdev_dtl[DTL_MISSING], 1); space_reftree_add_seg(&reftree, 0, scrub_txg, -1); space_reftree_add_map(&reftree, vd->vdev_dtl[DTL_SCRUB], 2); space_reftree_generate_map(&reftree, vd->vdev_dtl[DTL_MISSING], 1); space_reftree_destroy(&reftree); } range_tree_vacate(vd->vdev_dtl[DTL_PARTIAL], NULL, NULL); range_tree_walk(vd->vdev_dtl[DTL_MISSING], range_tree_add, vd->vdev_dtl[DTL_PARTIAL]); if (scrub_done) range_tree_vacate(vd->vdev_dtl[DTL_SCRUB], NULL, NULL); range_tree_vacate(vd->vdev_dtl[DTL_OUTAGE], NULL, NULL); if (!vdev_readable(vd)) range_tree_add(vd->vdev_dtl[DTL_OUTAGE], 0, -1ULL); else range_tree_walk(vd->vdev_dtl[DTL_MISSING], range_tree_add, vd->vdev_dtl[DTL_OUTAGE]); /* * If the vdev was resilvering and no longer has any * DTLs then reset its resilvering flag and dirty * the top level so that we persist the change. */ if (vd->vdev_resilver_txg != 0 && range_tree_space(vd->vdev_dtl[DTL_MISSING]) == 0 && range_tree_space(vd->vdev_dtl[DTL_OUTAGE]) == 0) { vd->vdev_resilver_txg = 0; vdev_config_dirty(vd->vdev_top); } mutex_exit(&vd->vdev_dtl_lock); if (txg != 0) vdev_dirty(vd->vdev_top, VDD_DTL, vd, txg); return; } mutex_enter(&vd->vdev_dtl_lock); for (int 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) 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_reftree_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_reftree_add_map(&reftree, cvd->vdev_dtl[s], 1); mutex_exit(&cvd->vdev_dtl_lock); } space_reftree_generate_map(&reftree, vd->vdev_dtl[t], minref); space_reftree_destroy(&reftree); } mutex_exit(&vd->vdev_dtl_lock); } int vdev_dtl_load(vdev_t *vd) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; int error = 0; if (vd->vdev_ops->vdev_op_leaf && vd->vdev_dtl_object != 0) { ASSERT(!vd->vdev_ishole); error = space_map_open(&vd->vdev_dtl_sm, mos, vd->vdev_dtl_object, 0, -1ULL, 0, &vd->vdev_dtl_lock); if (error) return (error); ASSERT(vd->vdev_dtl_sm != NULL); mutex_enter(&vd->vdev_dtl_lock); /* * Now that we've opened the space_map we need to update * the in-core DTL. */ space_map_update(vd->vdev_dtl_sm); error = space_map_load(vd->vdev_dtl_sm, vd->vdev_dtl[DTL_MISSING], SM_ALLOC); mutex_exit(&vd->vdev_dtl_lock); return (error); } for (int c = 0; c < vd->vdev_children; c++) { error = vdev_dtl_load(vd->vdev_child[c]); if (error != 0) break; } return (error); } void vdev_dtl_sync(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; range_tree_t *rt = vd->vdev_dtl[DTL_MISSING]; objset_t *mos = spa->spa_meta_objset; range_tree_t *rtsync; kmutex_t rtlock; dmu_tx_t *tx; uint64_t object = space_map_object(vd->vdev_dtl_sm); ASSERT(!vd->vdev_ishole); ASSERT(vd->vdev_ops->vdev_op_leaf); tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg); if (vd->vdev_detached || vd->vdev_top->vdev_removing) { mutex_enter(&vd->vdev_dtl_lock); space_map_free(vd->vdev_dtl_sm, tx); space_map_close(vd->vdev_dtl_sm); vd->vdev_dtl_sm = NULL; mutex_exit(&vd->vdev_dtl_lock); dmu_tx_commit(tx); return; } if (vd->vdev_dtl_sm == NULL) { uint64_t new_object; new_object = space_map_alloc(mos, tx); VERIFY3U(new_object, !=, 0); VERIFY0(space_map_open(&vd->vdev_dtl_sm, mos, new_object, 0, -1ULL, 0, &vd->vdev_dtl_lock)); ASSERT(vd->vdev_dtl_sm != NULL); } bzero(&rtlock, sizeof(rtlock)); mutex_init(&rtlock, NULL, MUTEX_DEFAULT, NULL); rtsync = range_tree_create(NULL, NULL, &rtlock); mutex_enter(&rtlock); mutex_enter(&vd->vdev_dtl_lock); range_tree_walk(rt, range_tree_add, rtsync); mutex_exit(&vd->vdev_dtl_lock); space_map_truncate(vd->vdev_dtl_sm, tx); space_map_write(vd->vdev_dtl_sm, rtsync, SM_ALLOC, tx); range_tree_vacate(rtsync, NULL, NULL); range_tree_destroy(rtsync); mutex_exit(&rtlock); mutex_destroy(&rtlock); /* * If the object for the space map has changed then dirty * the top level so that we update the config. */ if (object != space_map_object(vd->vdev_dtl_sm)) { zfs_dbgmsg("txg %llu, spa %s, DTL old object %llu, " "new object %llu", txg, spa_name(spa), object, space_map_object(vd->vdev_dtl_sm)); vdev_config_dirty(vd->vdev_top); } dmu_tx_commit(tx); mutex_enter(&vd->vdev_dtl_lock); space_map_update(vd->vdev_dtl_sm); mutex_exit(&vd->vdev_dtl_lock); } /* * 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); if (!required && zio_injection_enabled) required = !!zio_handle_device_injection(vd, NULL, ECHILD); return (required); } /* * Determine if resilver is needed, and if so the txg range. */ boolean_t vdev_resilver_needed(vdev_t *vd, uint64_t *minp, uint64_t *maxp) { boolean_t needed = B_FALSE; uint64_t thismin = UINT64_MAX; uint64_t thismax = 0; if (vd->vdev_children == 0) { mutex_enter(&vd->vdev_dtl_lock); if (range_tree_space(vd->vdev_dtl[DTL_MISSING]) != 0 && vdev_writeable(vd)) { thismin = vdev_dtl_min(vd); thismax = vdev_dtl_max(vd); needed = B_TRUE; } mutex_exit(&vd->vdev_dtl_lock); } else { for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; uint64_t cmin, cmax; if (vdev_resilver_needed(cvd, &cmin, &cmax)) { thismin = MIN(thismin, cmin); thismax = MAX(thismax, cmax); needed = B_TRUE; } } } if (needed && minp) { *minp = thismin; *maxp = thismax; } return (needed); } void vdev_load(vdev_t *vd) { /* * Recursively load all children. */ for (int c = 0; c < vd->vdev_children; c++) vdev_load(vd->vdev_child[c]); /* * If this is a top-level vdev, initialize its metaslabs. */ if (vd == vd->vdev_top && !vd->vdev_ishole && (vd->vdev_ashift == 0 || vd->vdev_asize == 0 || vdev_metaslab_init(vd, 0) != 0)) vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); /* * If this is a leaf vdev, load its DTL. */ if (vd->vdev_ops->vdev_op_leaf && vdev_dtl_load(vd) != 0) vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); } /* * The special vdev case is used for hot spares and l2cache devices. Its * sole purpose it to set the vdev state for the associated vdev. To do this, * we make sure that we can open the underlying device, then try to read the * label, and make sure that the label is sane and that it hasn't been * repurposed to another pool. */ int vdev_validate_aux(vdev_t *vd) { nvlist_t *label; uint64_t guid, version; uint64_t state; if (!vdev_readable(vd)) return (0); 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 || !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) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); return (-1); } /* * We don't actually check the pool state here. If it's in fact in * use by another pool, we update this fact on the fly when requested. */ nvlist_free(label); return (0); } void vdev_remove(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; dmu_tx_t *tx; tx = dmu_tx_create_assigned(spa_get_dsl(spa), txg); if (vd->vdev_ms != NULL) { metaslab_group_t *mg = vd->vdev_mg; metaslab_group_histogram_verify(mg); metaslab_class_histogram_verify(mg->mg_class); for (int m = 0; m < vd->vdev_ms_count; m++) { metaslab_t *msp = vd->vdev_ms[m]; if (msp == NULL || msp->ms_sm == NULL) continue; mutex_enter(&msp->ms_lock); /* * If the metaslab was not loaded when the vdev * was removed then the histogram accounting may * not be accurate. Update the histogram information * here so that we ensure that the metaslab group * and metaslab class are up-to-date. */ metaslab_group_histogram_remove(mg, msp); VERIFY0(space_map_allocated(msp->ms_sm)); space_map_free(msp->ms_sm, tx); space_map_close(msp->ms_sm); msp->ms_sm = NULL; mutex_exit(&msp->ms_lock); } metaslab_group_histogram_verify(mg); metaslab_class_histogram_verify(mg->mg_class); for (int i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++) ASSERT0(mg->mg_histogram[i]); } if (vd->vdev_ms_array) { (void) dmu_object_free(mos, vd->vdev_ms_array, tx); vd->vdev_ms_array = 0; } dmu_tx_commit(tx); } void vdev_sync_done(vdev_t *vd, uint64_t txg) { metaslab_t *msp; boolean_t reassess = !txg_list_empty(&vd->vdev_ms_list, TXG_CLEAN(txg)); ASSERT(!vd->vdev_ishole); while (msp = txg_list_remove(&vd->vdev_ms_list, TXG_CLEAN(txg))) metaslab_sync_done(msp, txg); if (reassess) metaslab_sync_reassess(vd->vdev_mg); } void vdev_sync(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; vdev_t *lvd; metaslab_t *msp; dmu_tx_t *tx; ASSERT(!vd->vdev_ishole); if (vd->vdev_ms_array == 0 && vd->vdev_ms_shift != 0) { ASSERT(vd == vd->vdev_top); tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg); vd->vdev_ms_array = dmu_object_alloc(spa->spa_meta_objset, DMU_OT_OBJECT_ARRAY, 0, DMU_OT_NONE, 0, tx); ASSERT(vd->vdev_ms_array != 0); vdev_config_dirty(vd); dmu_tx_commit(tx); } /* * Remove the metadata associated with this vdev once it's empty. */ if (vd->vdev_stat.vs_alloc == 0 && vd->vdev_removing) vdev_remove(vd, txg); while ((msp = txg_list_remove(&vd->vdev_ms_list, txg)) != NULL) { metaslab_sync(msp, txg); (void) txg_list_add(&vd->vdev_ms_list, msp, TXG_CLEAN(txg)); } while ((lvd = txg_list_remove(&vd->vdev_dtl_list, txg)) != NULL) vdev_dtl_sync(lvd, txg); (void) txg_list_add(&spa->spa_vdev_txg_list, vd, TXG_CLEAN(txg)); } uint64_t vdev_psize_to_asize(vdev_t *vd, uint64_t psize) { return (vd->vdev_ops->vdev_op_asize(vd, psize)); } /* * Mark the given vdev faulted. A faulted vdev behaves as if the device could * not be opened, and no I/O is attempted. */ int vdev_fault(spa_t *spa, uint64_t guid, vdev_aux_t aux) { vdev_t *vd, *tvd; spa_vdev_state_enter(spa, SCL_NONE); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, ENODEV)); 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 * were faulted. */ vd->vdev_label_aux = aux; /* * Faulted state takes precedence over degraded. */ vd->vdev_delayed_close = B_FALSE; vd->vdev_faulted = 1ULL; vd->vdev_degraded = 0ULL; vdev_set_state(vd, B_FALSE, VDEV_STATE_FAULTED, aux); /* * If this device has the only valid copy of the data, then * back off and simply mark the vdev as degraded instead. */ 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(tvd); if (vdev_readable(vd)) vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, aux); } return (spa_vdev_state_exit(spa, vd, 0)); } /* * Mark the given vdev degraded. A degraded vdev is purely an indication to the * user that something is wrong. The vdev continues to operate as normal as far * as I/O is concerned. */ int vdev_degrade(spa_t *spa, uint64_t guid, vdev_aux_t aux) { vdev_t *vd; spa_vdev_state_enter(spa, SCL_NONE); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); /* * If the vdev is already faulted, then don't do anything. */ if (vd->vdev_faulted || vd->vdev_degraded) return (spa_vdev_state_exit(spa, NULL, 0)); vd->vdev_degraded = 1ULL; if (!vdev_is_dead(vd)) vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, aux); return (spa_vdev_state_exit(spa, vd, 0)); } /* * Online the given vdev. * * If 'ZFS_ONLINE_UNSPARE' is set, it implies two things. First, any attached * spare device should be detached when the device finishes resilvering. * Second, the online should be treated like a 'test' online case, so no FMA * events are generated if the device fails to open. */ int vdev_online(spa_t *spa, uint64_t guid, uint64_t flags, vdev_state_t *newstate) { vdev_t *vd, *tvd, *pvd, *rvd = spa->spa_root_vdev; spa_vdev_state_enter(spa, SCL_NONE); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); tvd = vd->vdev_top; vd->vdev_offline = B_FALSE; vd->vdev_tmpoffline = B_FALSE; vd->vdev_checkremove = !!(flags & ZFS_ONLINE_CHECKREMOVE); vd->vdev_forcefault = !!(flags & ZFS_ONLINE_FORCEFAULT); /* XXX - L2ARC 1.0 does not support expansion */ if (!vd->vdev_aux) { for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent) pvd->vdev_expanding = !!(flags & ZFS_ONLINE_EXPAND); } vdev_reopen(tvd); vd->vdev_checkremove = vd->vdev_forcefault = B_FALSE; if (!vd->vdev_aux) { for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent) pvd->vdev_expanding = B_FALSE; } if (newstate) *newstate = vd->vdev_state; if ((flags & ZFS_ONLINE_UNSPARE) && !vdev_is_dead(vd) && vd->vdev_parent && vd->vdev_parent->vdev_ops == &vdev_spare_ops && vd->vdev_parent->vdev_child[0] == vd) vd->vdev_unspare = B_TRUE; if ((flags & ZFS_ONLINE_EXPAND) || spa->spa_autoexpand) { /* XXX - L2ARC 1.0 does not support expansion */ if (vd->vdev_aux) return (spa_vdev_state_exit(spa, vd, ENOTSUP)); spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE); } return (spa_vdev_state_exit(spa, vd, 0)); } static int vdev_offline_locked(spa_t *spa, uint64_t guid, uint64_t flags) { vdev_t *vd, *tvd; int error = 0; uint64_t generation; metaslab_group_t *mg; top: spa_vdev_state_enter(spa, SCL_ALLOC); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); tvd = vd->vdev_top; mg = tvd->vdev_mg; generation = spa->spa_config_generation + 1; /* * If the device isn't already offline, try to offline it. */ if (!vd->vdev_offline) { /* * If this device has the only valid copy of some data, * don't allow it to be offlined. Log devices are always * expendable. */ if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_dtl_required(vd)) return (spa_vdev_state_exit(spa, NULL, EBUSY)); /* * If the top-level is a slog and it has had allocations * then proceed. We check that the vdev's metaslab group * is not NULL since it's possible that we may have just * added this vdev but not yet initialized its metaslabs. */ if (tvd->vdev_islog && mg != NULL) { /* * Prevent any future allocations. */ metaslab_group_passivate(mg); (void) spa_vdev_state_exit(spa, vd, 0); error = spa_offline_log(spa); spa_vdev_state_enter(spa, SCL_ALLOC); /* * Check to see if the config has changed. */ if (error || generation != spa->spa_config_generation) { metaslab_group_activate(mg); if (error) return (spa_vdev_state_exit(spa, vd, error)); (void) spa_vdev_state_exit(spa, vd, 0); goto top; } ASSERT0(tvd->vdev_stat.vs_alloc); } /* * Offline this device and reopen its top-level vdev. * If the top-level vdev is a log device then just offline * it. Otherwise, if this action results in the top-level * vdev becoming unusable, undo it and fail the request. */ vd->vdev_offline = B_TRUE; vdev_reopen(tvd); if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_is_dead(tvd)) { vd->vdev_offline = B_FALSE; vdev_reopen(tvd); return (spa_vdev_state_exit(spa, NULL, EBUSY)); } /* * Add the device back into the metaslab rotor so that * once we online the device it's open for business. */ if (tvd->vdev_islog && mg != NULL) metaslab_group_activate(mg); } vd->vdev_tmpoffline = !!(flags & ZFS_OFFLINE_TEMPORARY); return (spa_vdev_state_exit(spa, vd, 0)); } int vdev_offline(spa_t *spa, uint64_t guid, uint64_t flags) { int error; mutex_enter(&spa->spa_vdev_top_lock); error = vdev_offline_locked(spa, guid, flags); mutex_exit(&spa->spa_vdev_top_lock); return (error); } /* * Clear the error counts associated with this vdev. Unlike vdev_online() and * vdev_offline(), we assume the spa config is locked. We also clear all * children. If 'vd' is NULL, then the user wants to clear all vdevs. */ void vdev_clear(spa_t *spa, vdev_t *vd) { vdev_t *rvd = spa->spa_root_vdev; ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); if (vd == NULL) vd = rvd; vd->vdev_stat.vs_read_errors = 0; vd->vdev_stat.vs_write_errors = 0; vd->vdev_stat.vs_checksum_errors = 0; for (int c = 0; c < vd->vdev_children; c++) vdev_clear(spa, vd->vdev_child[c]); if (vd == rvd) { for (int c = 0; c < spa->spa_l2cache.sav_count; c++) vdev_clear(spa, spa->spa_l2cache.sav_vdevs[c]); for (int c = 0; c < spa->spa_spares.sav_count; c++) vdev_clear(spa, spa->spa_spares.sav_vdevs[c]); } /* * If we're in the FAULTED state or have experienced failed I/O, then * clear the persistent state and attempt to reopen the device. We * also mark the vdev config dirty, so that the new faulted state is * written out to disk. */ if (vd->vdev_faulted || vd->vdev_degraded || !vdev_readable(vd) || !vdev_writeable(vd)) { /* * When reopening in reponse to a clear event, it may be due to * a fmadm repair request. In this case, if the device is * still broken, we want to still post the ereport again. */ vd->vdev_forcefault = B_TRUE; vd->vdev_faulted = vd->vdev_degraded = 0ULL; vd->vdev_cant_read = B_FALSE; vd->vdev_cant_write = B_FALSE; vdev_reopen(vd == rvd ? rvd : vd->vdev_top); vd->vdev_forcefault = B_FALSE; 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); } /* * When clearing a FMA-diagnosed fault, we always want to * unspare the device, as we assume that the original spare was * done in response to the FMA fault. */ if (!vdev_is_dead(vd) && vd->vdev_parent != NULL && vd->vdev_parent->vdev_ops == &vdev_spare_ops && vd->vdev_parent->vdev_child[0] == vd) vd->vdev_unspare = B_TRUE; } boolean_t vdev_is_dead(vdev_t *vd) { /* * Holes and missing devices are always considered "dead". * This simplifies the code since we don't have to check for * these types of devices in the various code paths. * Instead we rely on the fact that we skip over dead devices * before issuing I/O to them. */ return (vd->vdev_state < VDEV_STATE_DEGRADED || vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops); } boolean_t vdev_readable(vdev_t *vd) { return (!vdev_is_dead(vd) && !vd->vdev_cant_read); } boolean_t vdev_writeable(vdev_t *vd) { return (!vdev_is_dead(vd) && !vd->vdev_cant_write); } boolean_t vdev_allocatable(vdev_t *vd) { uint64_t state = vd->vdev_state; /* * 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. 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 (!(state < VDEV_STATE_DEGRADED && state != VDEV_STATE_CLOSED) && !vd->vdev_cant_write && !vd->vdev_ishole); } boolean_t vdev_accessible(vdev_t *vd, zio_t *zio) { ASSERT(zio->io_vd == vd); if (vdev_is_dead(vd) || vd->vdev_remove_wanted) return (B_FALSE); if (zio->io_type == ZIO_TYPE_READ) return (!vd->vdev_cant_read); if (zio->io_type == ZIO_TYPE_WRITE) return (!vd->vdev_cant_write); return (B_TRUE); } /* * Get statistics for the given vdev. */ void vdev_get_stats(vdev_t *vd, vdev_stat_t *vs) { spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0); mutex_enter(&vd->vdev_stat_lock); bcopy(&vd->vdev_stat, vs, sizeof (*vs)); vs->vs_timestamp = gethrtime() - vs->vs_timestamp; vs->vs_state = vd->vdev_state; 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; vs->vs_configured_ashift = vd->vdev_top != NULL ? vd->vdev_top->vdev_ashift : vd->vdev_ashift; vs->vs_logical_ashift = vd->vdev_logical_ashift; vs->vs_physical_ashift = vd->vdev_physical_ashift; if (vd->vdev_aux == NULL && vd == vd->vdev_top && !vd->vdev_ishole) { vs->vs_fragmentation = vd->vdev_mg->mg_fragmentation; } /* * If we're getting stats on the root vdev, aggregate the I/O counts * 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++) { vdev_t *cvd = rvd->vdev_child[c]; vdev_stat_t *cvs = &cvd->vdev_stat; for (int t = 0; t < ZIO_TYPES; t++) { vs->vs_ops[t] += cvs->vs_ops[t]; vs->vs_bytes[t] += cvs->vs_bytes[t]; } cvs->vs_scan_removing = cvd->vdev_removing; } } mutex_exit(&vd->vdev_stat_lock); } void vdev_clear_stats(vdev_t *vd) { mutex_enter(&vd->vdev_stat_lock); vd->vdev_stat.vs_space = 0; vd->vdev_stat.vs_dspace = 0; vd->vdev_stat.vs_alloc = 0; mutex_exit(&vd->vdev_stat_lock); } void vdev_scan_stat_init(vdev_t *vd) { vdev_stat_t *vs = &vd->vdev_stat; for (int c = 0; c < vd->vdev_children; c++) vdev_scan_stat_init(vd->vdev_child[c]); mutex_enter(&vd->vdev_stat_lock); vs->vs_scan_processed = 0; mutex_exit(&vd->vdev_stat_lock); } void vdev_stat_update(zio_t *zio, uint64_t psize) { 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; vdev_stat_t *vs = &vd->vdev_stat; zio_type_t type = zio->io_type; int flags = zio->io_flags; /* * If this i/o is a gang leader, it didn't do any actual work. */ if (zio->io_gang_tree) return; if (zio->io_error == 0) { /* * If this is a root i/o, don't count it -- we've already * counted the top-level vdevs, and vdev_get_stats() will * aggregate them when asked. This reduces contention on * the root vdev_stat_lock and implicitly handles blocks * that compress away to holes, for which there is no i/o. * (Holes never create vdev children, so all the counters * remain zero, which is what we want.) * * Note: this only applies to successful i/o (io_error == 0) * because unlike i/o counts, errors are not additive. * When reading a ditto block, for example, failure of * one top-level vdev does not imply a root-level error. */ if (vd == rvd) return; ASSERT(vd == zio->io_vd); if (flags & ZIO_FLAG_IO_BYPASS) return; mutex_enter(&vd->vdev_stat_lock); if (flags & ZIO_FLAG_IO_REPAIR) { 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; /* XXX cleanup? */ if (vd->vdev_ops->vdev_op_leaf) atomic_add_64(processed, psize); vs->vs_scan_processed += psize; } if (flags & ZIO_FLAG_SELF_HEAL) vs->vs_self_healed += psize; } vs->vs_ops[type]++; vs->vs_bytes[type] += psize; mutex_exit(&vd->vdev_stat_lock); return; } if (flags & ZIO_FLAG_SPECULATIVE) return; /* * If this is an I/O error that is going to be retried, then ignore the * error. Otherwise, the user may interpret B_FAILFAST I/O errors as * hard errors, when in reality they can happen for any number of * innocuous reasons (bus resets, MPxIO link failure, etc). */ if (zio->io_error == EIO && !(zio->io_flags & ZIO_FLAG_IO_RETRY)) return; /* * Intent logs writes won't propagate their error to the root * I/O so don't mark these types of failures as pool-level * errors. */ if (zio->io_vd == NULL && (zio->io_flags & ZIO_FLAG_DONT_PROPAGATE)) return; mutex_enter(&vd->vdev_stat_lock); if (type == ZIO_TYPE_READ && !vdev_is_dead(vd)) { if (zio->io_error == ECKSUM) vs->vs_checksum_errors++; else vs->vs_read_errors++; } if (type == ZIO_TYPE_WRITE && !vdev_is_dead(vd)) vs->vs_write_errors++; mutex_exit(&vd->vdev_stat_lock); if (type == ZIO_TYPE_WRITE && txg != 0 && (!(flags & ZIO_FLAG_IO_REPAIR) || (flags & ZIO_FLAG_SCAN_THREAD) || spa->spa_claiming)) { /* * This is either a normal write (not a repair), or it's * a repair induced by the scrub thread, or it's a repair * made by zil_claim() during spa_load() in the first txg. * 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_syncing_txg(spa). * In the zil_claim() case, we commit in spa_first_txg(spa). * * 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_SCAN_THREAD) { ASSERT(flags & ZIO_FLAG_IO_REPAIR); ASSERT(spa_sync_pass(spa) == 1); vdev_dtl_dirty(vd, DTL_SCRUB, txg, 1); commit_txg = spa_syncing_txg(spa); } else if (spa->spa_claiming) { ASSERT(flags & ZIO_FLAG_IO_REPAIR); commit_txg = spa_first_txg(spa); } ASSERT(commit_txg >= spa_syncing_txg(spa)); if (vdev_dtl_contains(vd, DTL_MISSING, txg, 1)) return; 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); } } /* * Update the in-core space usage stats for this vdev, its metaslab class, * and the root vdev. */ void vdev_space_update(vdev_t *vd, int64_t alloc_delta, int64_t defer_delta, int64_t space_delta) { int64_t dspace_delta = space_delta; spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; metaslab_group_t *mg = vd->vdev_mg; metaslab_class_t *mc = mg ? mg->mg_class : NULL; ASSERT(vd == vd->vdev_top); /* * Apply the inverse of the psize-to-asize (ie. RAID-Z) space-expansion * factor. We must calculate this here and not at the root vdev * because the root vdev's psize-to-asize is simply the max of its * childrens', thus not accurate enough for us. */ ASSERT((dspace_delta & (SPA_MINBLOCKSIZE-1)) == 0); ASSERT(vd->vdev_deflate_ratio != 0 || vd->vdev_isl2cache); dspace_delta = (dspace_delta >> SPA_MINBLOCKSHIFT) * vd->vdev_deflate_ratio; mutex_enter(&vd->vdev_stat_lock); vd->vdev_stat.vs_alloc += alloc_delta; vd->vdev_stat.vs_space += space_delta; vd->vdev_stat.vs_dspace += dspace_delta; mutex_exit(&vd->vdev_stat_lock); if (mc == spa_normal_class(spa)) { mutex_enter(&rvd->vdev_stat_lock); rvd->vdev_stat.vs_alloc += alloc_delta; rvd->vdev_stat.vs_space += space_delta; rvd->vdev_stat.vs_dspace += dspace_delta; mutex_exit(&rvd->vdev_stat_lock); } if (mc != NULL) { ASSERT(rvd == vd->vdev_parent); ASSERT(vd->vdev_ms_count != 0); metaslab_class_space_update(mc, alloc_delta, defer_delta, space_delta, dspace_delta); } } /* * Mark a top-level vdev's config as dirty, placing it on the dirty list * so that it will be written out next time the vdev configuration is synced. * If the root vdev is specified (vdev_top == NULL), dirty all top-level vdevs. */ void vdev_config_dirty(vdev_t *vd) { spa_t *spa = vd->vdev_spa; 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. */ if (vd->vdev_aux != NULL) { spa_aux_vdev_t *sav = vd->vdev_aux; nvlist_t **aux; uint_t naux; for (c = 0; c < sav->sav_count; c++) { if (sav->sav_vdevs[c] == vd) break; } if (c == sav->sav_count) { /* * We're being removed. There's nothing more to do. */ ASSERT(sav->sav_sync == B_TRUE); return; } sav->sav_sync = B_TRUE; if (nvlist_lookup_nvlist_array(sav->sav_config, ZPOOL_CONFIG_L2CACHE, &aux, &naux) != 0) { VERIFY(nvlist_lookup_nvlist_array(sav->sav_config, ZPOOL_CONFIG_SPARES, &aux, &naux) == 0); } ASSERT(c < naux); /* * Setting the nvlist in the middle if the array is a little * sketchy, but it will work. */ nvlist_free(aux[c]); aux[c] = vdev_config_generate(spa, vd, B_TRUE, 0); return; } /* * The dirty list is protected by the SCL_CONFIG lock. The caller * must either hold SCL_CONFIG as writer, or must be the sync thread * (which holds SCL_CONFIG as reader). There's only one sync thread, * so this is sufficient to ensure mutual exclusion. */ ASSERT(spa_config_held(spa, SCL_CONFIG, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_CONFIG, RW_READER))); if (vd == rvd) { for (c = 0; c < rvd->vdev_children; c++) vdev_config_dirty(rvd->vdev_child[c]); } else { ASSERT(vd == vd->vdev_top); if (!list_link_active(&vd->vdev_config_dirty_node) && !vd->vdev_ishole) list_insert_head(&spa->spa_config_dirty_list, vd); } } void vdev_config_clean(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_config_held(spa, SCL_CONFIG, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_CONFIG, RW_READER))); ASSERT(list_link_active(&vd->vdev_config_dirty_node)); list_remove(&spa->spa_config_dirty_list, vd); } /* * Mark a top-level vdev's state as dirty, so that the next pass of * spa_sync() can convert this into vdev_config_dirty(). We distinguish * the state changes from larger config changes because they require * much less locking, and are often needed for administrative actions. */ void vdev_state_dirty(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_writeable(spa)); ASSERT(vd == vd->vdev_top); /* * The state list is protected by the SCL_STATE lock. The caller * must either hold SCL_STATE as writer, or must be the sync thread * (which holds SCL_STATE as reader). There's only one sync thread, * so this is sufficient to ensure mutual exclusion. */ ASSERT(spa_config_held(spa, SCL_STATE, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_STATE, RW_READER))); if (!list_link_active(&vd->vdev_state_dirty_node) && !vd->vdev_ishole) list_insert_head(&spa->spa_state_dirty_list, vd); } void vdev_state_clean(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_config_held(spa, SCL_STATE, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_STATE, RW_READER))); ASSERT(list_link_active(&vd->vdev_state_dirty_node)); list_remove(&spa->spa_state_dirty_list, vd); } /* * Propagate vdev state up from children to parent. */ void vdev_propagate_state(vdev_t *vd) { spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; int degraded = 0, faulted = 0; int corrupted = 0; vdev_t *child; if (vd->vdev_children > 0) { for (int c = 0; c < vd->vdev_children; c++) { child = vd->vdev_child[c]; /* * Don't factor holes into the decision. */ if (child->vdev_ishole) continue; if (!vdev_readable(child) || (!vdev_writeable(child) && spa_writeable(spa))) { /* * Root special: if there is a top-level log * device, treat the root vdev as if it were * degraded. */ if (child->vdev_islog && vd == rvd) degraded++; else faulted++; } else if (child->vdev_state <= VDEV_STATE_DEGRADED) { degraded++; } if (child->vdev_stat.vs_aux == VDEV_AUX_CORRUPT_DATA) corrupted++; } vd->vdev_ops->vdev_op_state_change(vd, faulted, degraded); /* * Root special: if there is a top-level vdev that cannot be * opened due to corrupted metadata, then propagate the root * vdev's aux state as 'corrupt' rather than 'insufficient * replicas'. */ if (corrupted && vd == rvd && rvd->vdev_state == VDEV_STATE_CANT_OPEN) vdev_set_state(rvd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); } if (vd->vdev_parent) vdev_propagate_state(vd->vdev_parent); } /* * Set a vdev's state. If this is during an open, we don't update the parent * state, because we're in the process of opening children depth-first. * Otherwise, we propagate the change to the parent. * * If this routine places a device in a faulted state, an appropriate ereport is * generated. */ void vdev_set_state(vdev_t *vd, boolean_t isopen, vdev_state_t state, vdev_aux_t aux) { uint64_t save_state; spa_t *spa = vd->vdev_spa; if (state == vd->vdev_state) { vd->vdev_stat.vs_aux = aux; return; } save_state = vd->vdev_state; vd->vdev_state = state; vd->vdev_stat.vs_aux = aux; /* * If we are setting the vdev state to anything but an open state, then * always close the underlying device unless the device has requested * a delayed close (i.e. we're about to remove or fault the device). * Otherwise, we keep accessible but invalid devices open forever. * We don't call vdev_close() itself, because that implies some extra * checks (offline, etc) that we don't want here. This is limited to * leaf devices, because otherwise closing the device will affect other * children. */ if (!vd->vdev_delayed_close && vdev_is_dead(vd) && vd->vdev_ops->vdev_op_leaf) vd->vdev_ops->vdev_op_close(vd); /* * If we have brought this vdev back into service, we need * to notify fmd so that it can gracefully repair any outstanding * cases due to a missing device. We do this in all cases, even those * that probably don't correlate to a repaired fault. This is sure to * catch all cases, and we let the zfs-retire agent sort it out. If * this is a transient state it's OK, as the retire agent will * double-check the state of the vdev before repairing it. */ if (state == VDEV_STATE_HEALTHY && vd->vdev_ops->vdev_op_leaf && vd->vdev_prevstate != state) zfs_post_state_change(spa, vd); if (vd->vdev_removed && state == VDEV_STATE_CANT_OPEN && (aux == VDEV_AUX_OPEN_FAILED || vd->vdev_checkremove)) { /* * If the previous state is set to VDEV_STATE_REMOVED, then this * device was previously marked removed and someone attempted to * reopen it. If this failed due to a nonexistent device, then * keep the device in the REMOVED state. We also let this be if * it is one of our special test online cases, which is only * attempting to online the device and shouldn't generate an FMA * fault. */ vd->vdev_state = VDEV_STATE_REMOVED; vd->vdev_stat.vs_aux = VDEV_AUX_NONE; } else if (state == VDEV_STATE_REMOVED) { vd->vdev_removed = B_TRUE; } else if (state == VDEV_STATE_CANT_OPEN) { /* * 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 || spa_load_state(spa) == SPA_LOAD_RECOVER) && vd->vdev_ops->vdev_op_leaf) vd->vdev_not_present = 1; /* * Post the appropriate ereport. If the 'prevstate' field is * set to something other than VDEV_STATE_UNKNOWN, it indicates * that this is part of a vdev_reopen(). In this case, we don't * want to post the ereport if the device was already in the * CANT_OPEN state beforehand. * * If the 'checkremove' flag is set, then this is an attempt to * online the device in response to an insertion event. If we * hit this case, then we have detected an insertion event for a * faulted or offline device that wasn't in the removed state. * In this scenario, we don't post an ereport because we are * about to replace the device, or attempt an online with * vdev_forcefault, which will generate the fault for us. */ if ((vd->vdev_prevstate != state || vd->vdev_forcefault) && !vd->vdev_not_present && !vd->vdev_checkremove && vd != spa->spa_root_vdev) { const char *class; switch (aux) { case VDEV_AUX_OPEN_FAILED: class = FM_EREPORT_ZFS_DEVICE_OPEN_FAILED; break; case VDEV_AUX_CORRUPT_DATA: class = FM_EREPORT_ZFS_DEVICE_CORRUPT_DATA; break; case VDEV_AUX_NO_REPLICAS: class = FM_EREPORT_ZFS_DEVICE_NO_REPLICAS; break; case VDEV_AUX_BAD_GUID_SUM: class = FM_EREPORT_ZFS_DEVICE_BAD_GUID_SUM; break; case VDEV_AUX_TOO_SMALL: class = FM_EREPORT_ZFS_DEVICE_TOO_SMALL; break; case VDEV_AUX_BAD_LABEL: class = FM_EREPORT_ZFS_DEVICE_BAD_LABEL; break; default: class = FM_EREPORT_ZFS_DEVICE_UNKNOWN; } zfs_ereport_post(class, spa, vd, NULL, save_state, 0); } /* Erase any notion of persistent removed state */ vd->vdev_removed = B_FALSE; } else { vd->vdev_removed = B_FALSE; } if (!isopen && vd->vdev_parent) vdev_propagate_state(vd->vdev_parent); } /* * Check the vdev configuration to ensure that it's capable of supporting * a root pool. * * On Solaris, 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. * * For FreeBSD, we can boot from any configuration. There is a * limitation that the boot filesystem must be either uncompressed or * compresses with lzjb compression but I'm not sure how to enforce * that here. */ boolean_t vdev_is_bootable(vdev_t *vd) { #ifdef sun if (!vd->vdev_ops->vdev_op_leaf) { char *vdev_type = vd->vdev_ops->vdev_op_type; if (strcmp(vdev_type, VDEV_TYPE_ROOT) == 0 && vd->vdev_children > 1) { return (B_FALSE); } else if (strcmp(vdev_type, VDEV_TYPE_RAIDZ) == 0 || strcmp(vdev_type, VDEV_TYPE_MISSING) == 0) { return (B_FALSE); } } else if (vd->vdev_wholedisk == 1) { return (B_FALSE); } for (int c = 0; c < vd->vdev_children; c++) { if (!vdev_is_bootable(vd->vdev_child[c])) return (B_FALSE); } #endif /* 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 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; ASSERT(nvd->vdev_top->vdev_islog); ASSERT(spa_config_held(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++) vdev_load_log_state(nvd->vdev_child[c], ovd->vdev_child[c]); if (nvd->vdev_ops->vdev_op_leaf) { /* * Restore the persistent vdev state */ nvd->vdev_offline = ovd->vdev_offline; 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) { if (vd->vdev_ops->vdev_op_leaf && !vd->vdev_faulted && !vd->vdev_removed) return (B_TRUE); for (int 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_expand(vdev_t *vd, uint64_t txg) { ASSERT(vd->vdev_top == vd); ASSERT(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); if ((vd->vdev_asize >> vd->vdev_ms_shift) > vd->vdev_ms_count) { VERIFY(vdev_metaslab_init(vd, txg) == 0); vdev_config_dirty(vd); } } /* * Split a vdev. */ void vdev_split(vdev_t *vd) { vdev_t *cvd, *pvd = vd->vdev_parent; vdev_remove_child(pvd, vd); vdev_compact_children(pvd); cvd = pvd->vdev_child[0]; if (pvd->vdev_children == 1) { vdev_remove_parent(cvd); cvd->vdev_splitting = B_TRUE; } vdev_propagate_state(cvd); } void vdev_deadman(vdev_t *vd) { for (int 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_active_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 panic the system. */ fio = avl_first(&vq->vq_active_tree); delta = gethrtime() - fio->io_timestamp; if (delta > spa_deadman_synctime(spa)) { zfs_dbgmsg("SLOW IO: zio timestamp %lluns, " "delta %lluns, last io %lluns", fio->io_timestamp, delta, vq->vq_io_complete_ts); fm_panic("I/O to pool '%s' appears to be " "hung on vdev guid %llu at '%s'.", spa_name(spa), (long long unsigned int) vd->vdev_guid, vd->vdev_path); } } mutex_exit(&vq->vq_lock); } } Index: stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_disk.c =================================================================== --- stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_disk.c (revision 274799) +++ stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_disk.c (revision 274800) @@ -1,943 +1,945 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, 2014 by Delphix. All rights reserved. * Copyright 2013 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2013 Joyent, Inc. All rights reserved. */ #include #include #include #include #include #include #include #include #include #include /* * Virtual device vector for disks. */ extern ldi_ident_t zfs_li; static void vdev_disk_close(vdev_t *); typedef struct vdev_disk_ldi_cb { list_node_t lcb_next; ldi_callback_id_t lcb_id; } vdev_disk_ldi_cb_t; static void vdev_disk_alloc(vdev_t *vd) { vdev_disk_t *dvd; dvd = vd->vdev_tsd = kmem_zalloc(sizeof (vdev_disk_t), KM_SLEEP); /* * Create the LDI event callback list. */ list_create(&dvd->vd_ldi_cbs, sizeof (vdev_disk_ldi_cb_t), offsetof(vdev_disk_ldi_cb_t, lcb_next)); } static void vdev_disk_free(vdev_t *vd) { vdev_disk_t *dvd = vd->vdev_tsd; vdev_disk_ldi_cb_t *lcb; if (dvd == NULL) return; /* * We have already closed the LDI handle. Clean up the LDI event * callbacks and free vd->vdev_tsd. */ while ((lcb = list_head(&dvd->vd_ldi_cbs)) != NULL) { list_remove(&dvd->vd_ldi_cbs, lcb); (void) ldi_ev_remove_callbacks(lcb->lcb_id); kmem_free(lcb, sizeof (vdev_disk_ldi_cb_t)); } list_destroy(&dvd->vd_ldi_cbs); kmem_free(dvd, sizeof (vdev_disk_t)); vd->vdev_tsd = NULL; } /* ARGSUSED */ static int vdev_disk_off_notify(ldi_handle_t lh, ldi_ev_cookie_t ecookie, void *arg, void *ev_data) { vdev_t *vd = (vdev_t *)arg; vdev_disk_t *dvd = vd->vdev_tsd; /* * Ignore events other than offline. */ if (strcmp(ldi_ev_get_type(ecookie), LDI_EV_OFFLINE) != 0) return (LDI_EV_SUCCESS); /* * All LDI handles must be closed for the state change to succeed, so * call on vdev_disk_close() to do this. * * We inform vdev_disk_close that it is being called from offline * notify context so it will defer cleanup of LDI event callbacks and * freeing of vd->vdev_tsd to the offline finalize or a reopen. */ dvd->vd_ldi_offline = B_TRUE; vdev_disk_close(vd); /* * Now that the device is closed, request that the spa_async_thread * mark the device as REMOVED and notify FMA of the removal. */ zfs_post_remove(vd->vdev_spa, vd); vd->vdev_remove_wanted = B_TRUE; spa_async_request(vd->vdev_spa, SPA_ASYNC_REMOVE); return (LDI_EV_SUCCESS); } /* ARGSUSED */ static void vdev_disk_off_finalize(ldi_handle_t lh, ldi_ev_cookie_t ecookie, int ldi_result, void *arg, void *ev_data) { vdev_t *vd = (vdev_t *)arg; /* * Ignore events other than offline. */ if (strcmp(ldi_ev_get_type(ecookie), LDI_EV_OFFLINE) != 0) return; /* * We have already closed the LDI handle in notify. * Clean up the LDI event callbacks and free vd->vdev_tsd. */ vdev_disk_free(vd); /* * Request that the vdev be reopened if the offline state change was * unsuccessful. */ if (ldi_result != LDI_EV_SUCCESS) { vd->vdev_probe_wanted = B_TRUE; spa_async_request(vd->vdev_spa, SPA_ASYNC_PROBE); } } static ldi_ev_callback_t vdev_disk_off_callb = { .cb_vers = LDI_EV_CB_VERS, .cb_notify = vdev_disk_off_notify, .cb_finalize = vdev_disk_off_finalize }; /* ARGSUSED */ static void vdev_disk_dgrd_finalize(ldi_handle_t lh, ldi_ev_cookie_t ecookie, int ldi_result, void *arg, void *ev_data) { vdev_t *vd = (vdev_t *)arg; /* * Ignore events other than degrade. */ if (strcmp(ldi_ev_get_type(ecookie), LDI_EV_DEGRADE) != 0) return; /* * Degrade events always succeed. Mark the vdev as degraded. * This status is purely informative for the user. */ (void) vdev_degrade(vd->vdev_spa, vd->vdev_guid, 0); } static ldi_ev_callback_t vdev_disk_dgrd_callb = { .cb_vers = LDI_EV_CB_VERS, .cb_notify = NULL, .cb_finalize = vdev_disk_dgrd_finalize }; static void vdev_disk_hold(vdev_t *vd) { ddi_devid_t devid; char *minor; ASSERT(spa_config_held(vd->vdev_spa, SCL_STATE, RW_WRITER)); /* * We must have a pathname, and it must be absolute. */ if (vd->vdev_path == NULL || vd->vdev_path[0] != '/') return; /* * Only prefetch path and devid info if the device has * never been opened. */ if (vd->vdev_tsd != NULL) return; if (vd->vdev_wholedisk == -1ULL) { size_t len = strlen(vd->vdev_path) + 3; char *buf = kmem_alloc(len, KM_SLEEP); (void) snprintf(buf, len, "%ss0", vd->vdev_path); (void) ldi_vp_from_name(buf, &vd->vdev_name_vp); kmem_free(buf, len); } if (vd->vdev_name_vp == NULL) (void) ldi_vp_from_name(vd->vdev_path, &vd->vdev_name_vp); if (vd->vdev_devid != NULL && ddi_devid_str_decode(vd->vdev_devid, &devid, &minor) == 0) { (void) ldi_vp_from_devid(devid, minor, &vd->vdev_devid_vp); ddi_devid_str_free(minor); ddi_devid_free(devid); } } static void vdev_disk_rele(vdev_t *vd) { ASSERT(spa_config_held(vd->vdev_spa, SCL_STATE, RW_WRITER)); if (vd->vdev_name_vp) { VN_RELE_ASYNC(vd->vdev_name_vp, dsl_pool_vnrele_taskq(vd->vdev_spa->spa_dsl_pool)); vd->vdev_name_vp = NULL; } if (vd->vdev_devid_vp) { VN_RELE_ASYNC(vd->vdev_devid_vp, dsl_pool_vnrele_taskq(vd->vdev_spa->spa_dsl_pool)); vd->vdev_devid_vp = NULL; } } static uint64_t vdev_disk_get_space(vdev_t *vd, uint64_t capacity, uint_t blksz) { ASSERT(vd->vdev_wholedisk); vdev_disk_t *dvd = vd->vdev_tsd; dk_efi_t dk_ioc; efi_gpt_t *efi; uint64_t avail_space = 0; int efisize = EFI_LABEL_SIZE * 2; dk_ioc.dki_data = kmem_alloc(efisize, KM_SLEEP); dk_ioc.dki_lba = 1; dk_ioc.dki_length = efisize; dk_ioc.dki_data_64 = (uint64_t)(uintptr_t)dk_ioc.dki_data; efi = dk_ioc.dki_data; if (ldi_ioctl(dvd->vd_lh, DKIOCGETEFI, (intptr_t)&dk_ioc, FKIOCTL, kcred, NULL) == 0) { uint64_t efi_altern_lba = LE_64(efi->efi_gpt_AlternateLBA); if (capacity > efi_altern_lba) avail_space = (capacity - efi_altern_lba) * blksz; } kmem_free(dk_ioc.dki_data, efisize); return (avail_space); } /* * We want to be loud in DEBUG kernels when DKIOCGMEDIAINFOEXT fails, or when * even a fallback to DKIOCGMEDIAINFO fails. */ #ifdef DEBUG #define VDEV_DEBUG(...) cmn_err(CE_NOTE, __VA_ARGS__) #else #define VDEV_DEBUG(...) /* Nothing... */ #endif static int vdev_disk_open(vdev_t *vd, uint64_t *psize, uint64_t *max_psize, uint64_t *ashift) { spa_t *spa = vd->vdev_spa; vdev_disk_t *dvd = vd->vdev_tsd; ldi_ev_cookie_t ecookie; vdev_disk_ldi_cb_t *lcb; union { struct dk_minfo_ext ude; struct dk_minfo ud; } dks; struct dk_minfo_ext *dkmext = &dks.ude; struct dk_minfo *dkm = &dks.ud; int error; dev_t dev; int otyp; boolean_t validate_devid = B_FALSE; ddi_devid_t devid; uint64_t capacity = 0, blksz = 0, pbsize; /* * We must have a pathname, and it must be absolute. */ if (vd->vdev_path == NULL || vd->vdev_path[0] != '/') { vd->vdev_stat.vs_aux = VDEV_AUX_BAD_LABEL; return (SET_ERROR(EINVAL)); } /* * Reopen the device if it's not currently open. Otherwise, * just update the physical size of the device. */ if (dvd != NULL) { if (dvd->vd_ldi_offline && dvd->vd_lh == NULL) { /* * If we are opening a device in its offline notify * context, the LDI handle was just closed. Clean * up the LDI event callbacks and free vd->vdev_tsd. */ vdev_disk_free(vd); } else { ASSERT(vd->vdev_reopening); goto skip_open; } } /* * Create vd->vdev_tsd. */ vdev_disk_alloc(vd); dvd = vd->vdev_tsd; /* * When opening a disk device, we want to preserve the user's original * intent. We always want to open the device by the path the user gave * us, even if it is one of multiple paths to the save device. But we * also want to be able to survive disks being removed/recabled. * Therefore the sequence of opening devices is: * * 1. Try opening the device by path. For legacy pools without the * 'whole_disk' property, attempt to fix the path by appending 's0'. * * 2. If the devid of the device matches the stored value, return * success. * * 3. Otherwise, the device may have moved. Try opening the device * by the devid instead. */ if (vd->vdev_devid != NULL) { if (ddi_devid_str_decode(vd->vdev_devid, &dvd->vd_devid, &dvd->vd_minor) != 0) { vd->vdev_stat.vs_aux = VDEV_AUX_BAD_LABEL; return (SET_ERROR(EINVAL)); } } error = EINVAL; /* presume failure */ if (vd->vdev_path != NULL) { if (vd->vdev_wholedisk == -1ULL) { size_t len = strlen(vd->vdev_path) + 3; char *buf = kmem_alloc(len, KM_SLEEP); (void) snprintf(buf, len, "%ss0", vd->vdev_path); error = ldi_open_by_name(buf, spa_mode(spa), kcred, &dvd->vd_lh, zfs_li); if (error == 0) { spa_strfree(vd->vdev_path); vd->vdev_path = buf; vd->vdev_wholedisk = 1ULL; } else { kmem_free(buf, len); } } /* * If we have not yet opened the device, try to open it by the * specified path. */ if (error != 0) { error = ldi_open_by_name(vd->vdev_path, spa_mode(spa), kcred, &dvd->vd_lh, zfs_li); } /* * Compare the devid to the stored value. */ if (error == 0 && vd->vdev_devid != NULL && ldi_get_devid(dvd->vd_lh, &devid) == 0) { if (ddi_devid_compare(devid, dvd->vd_devid) != 0) { error = SET_ERROR(EINVAL); (void) ldi_close(dvd->vd_lh, spa_mode(spa), kcred); dvd->vd_lh = NULL; } ddi_devid_free(devid); } /* * If we succeeded in opening the device, but 'vdev_wholedisk' * is not yet set, then this must be a slice. */ if (error == 0 && vd->vdev_wholedisk == -1ULL) vd->vdev_wholedisk = 0; } /* * If we were unable to open by path, or the devid check fails, open by * devid instead. */ if (error != 0 && vd->vdev_devid != NULL) { error = ldi_open_by_devid(dvd->vd_devid, dvd->vd_minor, spa_mode(spa), kcred, &dvd->vd_lh, zfs_li); } /* * If all else fails, then try opening by physical path (if available) * or the logical path (if we failed due to the devid check). While not * as reliable as the devid, this will give us something, and the higher * level vdev validation will prevent us from opening the wrong device. */ if (error) { if (vd->vdev_devid != NULL) validate_devid = B_TRUE; if (vd->vdev_physpath != NULL && (dev = ddi_pathname_to_dev_t(vd->vdev_physpath)) != NODEV) error = ldi_open_by_dev(&dev, OTYP_BLK, spa_mode(spa), kcred, &dvd->vd_lh, zfs_li); /* * Note that we don't support the legacy auto-wholedisk support * as above. This hasn't been used in a very long time and we * don't need to propagate its oddities to this edge condition. */ if (error && vd->vdev_path != NULL) error = ldi_open_by_name(vd->vdev_path, spa_mode(spa), kcred, &dvd->vd_lh, zfs_li); } if (error) { vd->vdev_stat.vs_aux = VDEV_AUX_OPEN_FAILED; return (error); } /* * Now that the device has been successfully opened, update the devid * if necessary. */ if (validate_devid && spa_writeable(spa) && ldi_get_devid(dvd->vd_lh, &devid) == 0) { if (ddi_devid_compare(devid, dvd->vd_devid) != 0) { char *vd_devid; vd_devid = ddi_devid_str_encode(devid, dvd->vd_minor); zfs_dbgmsg("vdev %s: update devid from %s, " "to %s", vd->vdev_path, vd->vdev_devid, vd_devid); spa_strfree(vd->vdev_devid); vd->vdev_devid = spa_strdup(vd_devid); ddi_devid_str_free(vd_devid); } ddi_devid_free(devid); } /* * Once a device is opened, verify that the physical device path (if * available) is up to date. */ if (ldi_get_dev(dvd->vd_lh, &dev) == 0 && ldi_get_otyp(dvd->vd_lh, &otyp) == 0) { char *physpath, *minorname; physpath = kmem_alloc(MAXPATHLEN, KM_SLEEP); minorname = NULL; if (ddi_dev_pathname(dev, otyp, physpath) == 0 && ldi_get_minor_name(dvd->vd_lh, &minorname) == 0 && (vd->vdev_physpath == NULL || strcmp(vd->vdev_physpath, physpath) != 0)) { if (vd->vdev_physpath) spa_strfree(vd->vdev_physpath); (void) strlcat(physpath, ":", MAXPATHLEN); (void) strlcat(physpath, minorname, MAXPATHLEN); vd->vdev_physpath = spa_strdup(physpath); } if (minorname) kmem_free(minorname, strlen(minorname) + 1); kmem_free(physpath, MAXPATHLEN); } /* * Register callbacks for the LDI offline event. */ if (ldi_ev_get_cookie(dvd->vd_lh, LDI_EV_OFFLINE, &ecookie) == LDI_EV_SUCCESS) { lcb = kmem_zalloc(sizeof (vdev_disk_ldi_cb_t), KM_SLEEP); list_insert_tail(&dvd->vd_ldi_cbs, lcb); (void) ldi_ev_register_callbacks(dvd->vd_lh, ecookie, &vdev_disk_off_callb, (void *) vd, &lcb->lcb_id); } /* * Register callbacks for the LDI degrade event. */ if (ldi_ev_get_cookie(dvd->vd_lh, LDI_EV_DEGRADE, &ecookie) == LDI_EV_SUCCESS) { lcb = kmem_zalloc(sizeof (vdev_disk_ldi_cb_t), KM_SLEEP); list_insert_tail(&dvd->vd_ldi_cbs, lcb); (void) ldi_ev_register_callbacks(dvd->vd_lh, ecookie, &vdev_disk_dgrd_callb, (void *) vd, &lcb->lcb_id); } skip_open: /* * Determine the actual size of the device. */ if (ldi_get_size(dvd->vd_lh, psize) != 0) { vd->vdev_stat.vs_aux = VDEV_AUX_OPEN_FAILED; return (SET_ERROR(EINVAL)); } *max_psize = *psize; /* * Determine the device's minimum transfer size. * If the ioctl isn't supported, assume DEV_BSIZE. */ if ((error = ldi_ioctl(dvd->vd_lh, DKIOCGMEDIAINFOEXT, (intptr_t)dkmext, FKIOCTL, kcred, NULL)) == 0) { capacity = dkmext->dki_capacity - 1; blksz = dkmext->dki_lbsize; pbsize = dkmext->dki_pbsize; } else if ((error = ldi_ioctl(dvd->vd_lh, DKIOCGMEDIAINFO, (intptr_t)dkm, FKIOCTL, kcred, NULL)) == 0) { VDEV_DEBUG( "vdev_disk_open(\"%s\"): fallback to DKIOCGMEDIAINFO\n", vd->vdev_path); capacity = dkm->dki_capacity - 1; blksz = dkm->dki_lbsize; pbsize = blksz; } else { VDEV_DEBUG("vdev_disk_open(\"%s\"): " "both DKIOCGMEDIAINFO{,EXT} calls failed, %d\n", vd->vdev_path, error); pbsize = DEV_BSIZE; } *ashift = highbit64(MAX(pbsize, SPA_MINBLOCKSIZE)) - 1; if (vd->vdev_wholedisk == 1) { int wce = 1; if (error == 0) { /* * If we have the capability to expand, we'd have * found out via success from DKIOCGMEDIAINFO{,EXT}. * Adjust max_psize upward accordingly since we know * we own the whole disk now. */ *max_psize += vdev_disk_get_space(vd, capacity, blksz); zfs_dbgmsg("capacity change: vdev %s, psize %llu, " "max_psize %llu", vd->vdev_path, *psize, *max_psize); } /* * Since we own the whole disk, try to enable disk write * caching. We ignore errors because it's OK if we can't do it. */ (void) ldi_ioctl(dvd->vd_lh, DKIOCSETWCE, (intptr_t)&wce, FKIOCTL, kcred, NULL); } /* * Clear the nowritecache bit, so that on a vdev_reopen() we will * try again. */ vd->vdev_nowritecache = B_FALSE; return (0); } static void vdev_disk_close(vdev_t *vd) { vdev_disk_t *dvd = vd->vdev_tsd; if (vd->vdev_reopening || dvd == NULL) return; if (dvd->vd_minor != NULL) { ddi_devid_str_free(dvd->vd_minor); dvd->vd_minor = NULL; } if (dvd->vd_devid != NULL) { ddi_devid_free(dvd->vd_devid); dvd->vd_devid = NULL; } if (dvd->vd_lh != NULL) { (void) ldi_close(dvd->vd_lh, spa_mode(vd->vdev_spa), kcred); dvd->vd_lh = NULL; } vd->vdev_delayed_close = B_FALSE; /* * If we closed the LDI handle due to an offline notify from LDI, * don't free vd->vdev_tsd or unregister the callbacks here; * the offline finalize callback or a reopen will take care of it. */ if (dvd->vd_ldi_offline) return; vdev_disk_free(vd); } int vdev_disk_physio(vdev_t *vd, caddr_t data, size_t size, uint64_t offset, int flags, boolean_t isdump) { vdev_disk_t *dvd = vd->vdev_tsd; /* * If the vdev is closed, it's likely in the REMOVED or FAULTED state. * Nothing to be done here but return failure. */ if (dvd == NULL || (dvd->vd_ldi_offline && dvd->vd_lh == NULL)) return (EIO); ASSERT(vd->vdev_ops == &vdev_disk_ops); /* * If in the context of an active crash dump, use the ldi_dump(9F) * call instead of ldi_strategy(9F) as usual. */ if (isdump) { ASSERT3P(dvd, !=, NULL); return (ldi_dump(dvd->vd_lh, data, lbtodb(offset), lbtodb(size))); } return (vdev_disk_ldi_physio(dvd->vd_lh, data, size, offset, flags)); } int vdev_disk_ldi_physio(ldi_handle_t vd_lh, caddr_t data, size_t size, uint64_t offset, int flags) { buf_t *bp; int error = 0; if (vd_lh == NULL) return (SET_ERROR(EINVAL)); ASSERT(flags & B_READ || flags & B_WRITE); bp = getrbuf(KM_SLEEP); bp->b_flags = flags | B_BUSY | B_NOCACHE | B_FAILFAST; bp->b_bcount = size; bp->b_un.b_addr = (void *)data; bp->b_lblkno = lbtodb(offset); bp->b_bufsize = size; error = ldi_strategy(vd_lh, bp); ASSERT(error == 0); if ((error = biowait(bp)) == 0 && bp->b_resid != 0) error = SET_ERROR(EIO); freerbuf(bp); return (error); } static void vdev_disk_io_intr(buf_t *bp) { vdev_buf_t *vb = (vdev_buf_t *)bp; zio_t *zio = vb->vb_io; /* * The rest of the zio stack only deals with EIO, ECKSUM, and ENXIO. * Rather than teach the rest of the stack about other error * possibilities (EFAULT, etc), we normalize the error value here. */ zio->io_error = (geterror(bp) != 0 ? SET_ERROR(EIO) : 0); if (zio->io_error == 0 && bp->b_resid != 0) zio->io_error = SET_ERROR(EIO); kmem_free(vb, sizeof (vdev_buf_t)); zio_interrupt(zio); } static void vdev_disk_ioctl_free(zio_t *zio) { kmem_free(zio->io_vsd, sizeof (struct dk_callback)); } static const zio_vsd_ops_t vdev_disk_vsd_ops = { vdev_disk_ioctl_free, zio_vsd_default_cksum_report }; static void vdev_disk_ioctl_done(void *zio_arg, int error) { zio_t *zio = zio_arg; zio->io_error = error; zio_interrupt(zio); } static int vdev_disk_io_start(zio_t *zio) { vdev_t *vd = zio->io_vd; vdev_disk_t *dvd = vd->vdev_tsd; vdev_buf_t *vb; struct dk_callback *dkc; buf_t *bp; int error; /* * If the vdev is closed, it's likely in the REMOVED or FAULTED state. * Nothing to be done here but return failure. */ if (dvd == NULL || (dvd->vd_ldi_offline && dvd->vd_lh == NULL)) { zio->io_error = SET_ERROR(ENXIO); zio_interrupt(zio); return (ZIO_PIPELINE_STOP); } if (zio->io_type == ZIO_TYPE_IOCTL) { /* XXPOLICY */ if (!vdev_readable(vd)) { zio->io_error = SET_ERROR(ENXIO); zio_interrupt(zio); return (ZIO_PIPELINE_STOP); } switch (zio->io_cmd) { case DKIOCFLUSHWRITECACHE: if (zfs_nocacheflush) break; if (vd->vdev_nowritecache) { zio->io_error = SET_ERROR(ENOTSUP); break; } zio->io_vsd = dkc = kmem_alloc(sizeof (*dkc), KM_SLEEP); zio->io_vsd_ops = &vdev_disk_vsd_ops; dkc->dkc_callback = vdev_disk_ioctl_done; dkc->dkc_flag = FLUSH_VOLATILE; dkc->dkc_cookie = zio; error = ldi_ioctl(dvd->vd_lh, zio->io_cmd, (uintptr_t)dkc, FKIOCTL, kcred, NULL); if (error == 0) { /* * The ioctl will be done asychronously, * and will call vdev_disk_ioctl_done() * upon completion. */ return (ZIO_PIPELINE_STOP); } if (error == ENOTSUP || error == ENOTTY) { /* * If we get ENOTSUP or ENOTTY, we know that * no future attempts will ever succeed. * In this case we set a persistent bit so * that we don't bother with the ioctl in the * future. */ vd->vdev_nowritecache = B_TRUE; } zio->io_error = error; break; default: zio->io_error = SET_ERROR(ENOTSUP); } zio_interrupt(zio); return (ZIO_PIPELINE_STOP); } + ASSERT(zio->io_type == ZIO_TYPE_READ || zio->io_type == ZIO_TYPE_WRITE); + vb = kmem_alloc(sizeof (vdev_buf_t), KM_SLEEP); vb->vb_io = zio; bp = &vb->vb_buf; bioinit(bp); bp->b_flags = B_BUSY | B_NOCACHE | (zio->io_type == ZIO_TYPE_READ ? B_READ : B_WRITE); if (!(zio->io_flags & (ZIO_FLAG_IO_RETRY | ZIO_FLAG_TRYHARD))) bp->b_flags |= B_FAILFAST; bp->b_bcount = zio->io_size; bp->b_un.b_addr = zio->io_data; bp->b_lblkno = lbtodb(zio->io_offset); bp->b_bufsize = zio->io_size; bp->b_iodone = (int (*)())vdev_disk_io_intr; /* ldi_strategy() will return non-zero only on programming errors */ VERIFY(ldi_strategy(dvd->vd_lh, bp) == 0); return (ZIO_PIPELINE_STOP); } static void vdev_disk_io_done(zio_t *zio) { vdev_t *vd = zio->io_vd; /* * If the device returned EIO, then attempt a DKIOCSTATE ioctl to see if * the device has been removed. If this is the case, then we trigger an * asynchronous removal of the device. Otherwise, probe the device and * make sure it's still accessible. */ if (zio->io_error == EIO && !vd->vdev_remove_wanted) { vdev_disk_t *dvd = vd->vdev_tsd; int state = DKIO_NONE; if (ldi_ioctl(dvd->vd_lh, DKIOCSTATE, (intptr_t)&state, FKIOCTL, kcred, NULL) == 0 && state != DKIO_INSERTED) { /* * We post the resource as soon as possible, instead of * when the async removal actually happens, because the * DE is using this information to discard previous I/O * errors. */ zfs_post_remove(zio->io_spa, vd); vd->vdev_remove_wanted = B_TRUE; spa_async_request(zio->io_spa, SPA_ASYNC_REMOVE); } else if (!vd->vdev_delayed_close) { vd->vdev_delayed_close = B_TRUE; } } } vdev_ops_t vdev_disk_ops = { vdev_disk_open, vdev_disk_close, vdev_default_asize, vdev_disk_io_start, vdev_disk_io_done, NULL, vdev_disk_hold, vdev_disk_rele, VDEV_TYPE_DISK, /* name of this vdev type */ B_TRUE /* leaf vdev */ }; /* * Given the root disk device devid or pathname, read the label from * the device, and construct a configuration nvlist. */ int vdev_disk_read_rootlabel(char *devpath, char *devid, nvlist_t **config) { ldi_handle_t vd_lh; vdev_label_t *label; uint64_t s, size; int l; ddi_devid_t tmpdevid; int error = -1; char *minor_name; /* * Read the device label and build the nvlist. */ if (devid != NULL && ddi_devid_str_decode(devid, &tmpdevid, &minor_name) == 0) { error = ldi_open_by_devid(tmpdevid, minor_name, FREAD, kcred, &vd_lh, zfs_li); ddi_devid_free(tmpdevid); ddi_devid_str_free(minor_name); } if (error && (error = ldi_open_by_name(devpath, FREAD, kcred, &vd_lh, zfs_li))) return (error); if (ldi_get_size(vd_lh, &s)) { (void) ldi_close(vd_lh, FREAD, kcred); return (SET_ERROR(EIO)); } size = P2ALIGN_TYPED(s, sizeof (vdev_label_t), uint64_t); label = kmem_alloc(sizeof (vdev_label_t), KM_SLEEP); *config = NULL; for (l = 0; l < VDEV_LABELS; l++) { uint64_t offset, state, txg = 0; /* read vdev label */ offset = vdev_label_offset(size, l, 0); if (vdev_disk_ldi_physio(vd_lh, (caddr_t)label, VDEV_SKIP_SIZE + VDEV_PHYS_SIZE, offset, B_READ) != 0) continue; if (nvlist_unpack(label->vl_vdev_phys.vp_nvlist, sizeof (label->vl_vdev_phys.vp_nvlist), config, 0) != 0) { *config = NULL; continue; } if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_STATE, &state) != 0 || state >= POOL_STATE_DESTROYED) { nvlist_free(*config); *config = NULL; continue; } if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_TXG, &txg) != 0 || txg == 0) { nvlist_free(*config); *config = NULL; continue; } break; } kmem_free(label, sizeof (vdev_label_t)); (void) ldi_close(vd_lh, FREAD, kcred); if (*config == NULL) error = SET_ERROR(EIDRM); return (error); } Index: stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_file.c =================================================================== --- stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_file.c (revision 274799) +++ stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_file.c (revision 274800) @@ -1,237 +1,241 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2013 by Delphix. All rights reserved. */ #include #include #include #include #include #include #include /* * Virtual device vector for files. */ static void vdev_file_hold(vdev_t *vd) { ASSERT(vd->vdev_path != NULL); } static void vdev_file_rele(vdev_t *vd) { ASSERT(vd->vdev_path != NULL); } static int vdev_file_open(vdev_t *vd, uint64_t *psize, uint64_t *max_psize, uint64_t *logical_ashift, uint64_t *physical_ashift) { vdev_file_t *vf; vnode_t *vp; vattr_t vattr; int error; /* * We must have a pathname, and it must be absolute. */ if (vd->vdev_path == NULL || vd->vdev_path[0] != '/') { vd->vdev_stat.vs_aux = VDEV_AUX_BAD_LABEL; return (SET_ERROR(EINVAL)); } /* * Reopen the device if it's not currently open. Otherwise, * just update the physical size of the device. */ if (vd->vdev_tsd != NULL) { ASSERT(vd->vdev_reopening); vf = vd->vdev_tsd; vp = vf->vf_vnode; goto skip_open; } vf = vd->vdev_tsd = kmem_zalloc(sizeof (vdev_file_t), KM_SLEEP); /* * We always open the files from the root of the global zone, even if * we're in a local zone. If the user has gotten to this point, the * administrator has already decided that the pool should be available * to local zone users, so the underlying devices should be as well. */ ASSERT(vd->vdev_path != NULL && vd->vdev_path[0] == '/'); error = vn_openat(vd->vdev_path + 1, UIO_SYSSPACE, spa_mode(vd->vdev_spa) | FOFFMAX, 0, &vp, 0, 0, rootdir, -1); if (error) { vd->vdev_stat.vs_aux = VDEV_AUX_OPEN_FAILED; kmem_free(vd->vdev_tsd, sizeof (vdev_file_t)); vd->vdev_tsd = NULL; return (error); } vf->vf_vnode = vp; #ifdef _KERNEL /* * Make sure it's a regular file. */ if (vp->v_type != VREG) { #ifdef __FreeBSD__ (void) VOP_CLOSE(vp, spa_mode(vd->vdev_spa), 1, 0, kcred, NULL); #endif vd->vdev_stat.vs_aux = VDEV_AUX_OPEN_FAILED; #ifdef __FreeBSD__ kmem_free(vd->vdev_tsd, sizeof (vdev_file_t)); vd->vdev_tsd = NULL; #endif return (SET_ERROR(ENODEV)); } #endif /* _KERNEL */ skip_open: /* * Determine the physical size of the file. */ vattr.va_mask = AT_SIZE; vn_lock(vp, LK_SHARED | LK_RETRY); error = VOP_GETATTR(vp, &vattr, kcred); VOP_UNLOCK(vp, 0); if (error) { (void) VOP_CLOSE(vp, spa_mode(vd->vdev_spa), 1, 0, kcred, NULL); vd->vdev_stat.vs_aux = VDEV_AUX_OPEN_FAILED; kmem_free(vd->vdev_tsd, sizeof (vdev_file_t)); vd->vdev_tsd = NULL; return (error); } + vd->vdev_notrim = B_TRUE; + *max_psize = *psize = vattr.va_size; *logical_ashift = SPA_MINBLOCKSHIFT; *physical_ashift = SPA_MINBLOCKSHIFT; return (0); } static void vdev_file_close(vdev_t *vd) { vdev_file_t *vf = vd->vdev_tsd; if (vd->vdev_reopening || vf == NULL) return; if (vf->vf_vnode != NULL) { (void) VOP_CLOSE(vf->vf_vnode, spa_mode(vd->vdev_spa), 1, 0, kcred, NULL); } vd->vdev_delayed_close = B_FALSE; kmem_free(vf, sizeof (vdev_file_t)); vd->vdev_tsd = NULL; } static int vdev_file_io_start(zio_t *zio) { vdev_t *vd = zio->io_vd; vdev_file_t *vf; vnode_t *vp; ssize_t resid; if (!vdev_readable(vd)) { zio->io_error = SET_ERROR(ENXIO); zio_interrupt(zio); return (ZIO_PIPELINE_STOP); } vf = vd->vdev_tsd; vp = vf->vf_vnode; if (zio->io_type == ZIO_TYPE_IOCTL) { switch (zio->io_cmd) { case DKIOCFLUSHWRITECACHE: zio->io_error = VOP_FSYNC(vp, FSYNC | FDSYNC, kcred, NULL); break; default: zio->io_error = SET_ERROR(ENOTSUP); } zio_interrupt(zio); return (ZIO_PIPELINE_STOP); } + + ASSERT(zio->io_type == ZIO_TYPE_READ || zio->io_type == ZIO_TYPE_WRITE); zio->io_error = vn_rdwr(zio->io_type == ZIO_TYPE_READ ? UIO_READ : UIO_WRITE, vp, zio->io_data, zio->io_size, zio->io_offset, UIO_SYSSPACE, 0, RLIM64_INFINITY, kcred, &resid); if (resid != 0 && zio->io_error == 0) zio->io_error = ENOSPC; zio_interrupt(zio); return (ZIO_PIPELINE_STOP); } /* ARGSUSED */ static void vdev_file_io_done(zio_t *zio) { } vdev_ops_t vdev_file_ops = { vdev_file_open, vdev_file_close, vdev_default_asize, vdev_file_io_start, vdev_file_io_done, NULL, vdev_file_hold, vdev_file_rele, VDEV_TYPE_FILE, /* name of this vdev type */ B_TRUE /* leaf vdev */ }; /* * From userland we access disks just like files. */ #ifndef _KERNEL vdev_ops_t vdev_disk_ops = { vdev_file_open, vdev_file_close, vdev_default_asize, vdev_file_io_start, vdev_file_io_done, NULL, vdev_file_hold, vdev_file_rele, VDEV_TYPE_DISK, /* name of this vdev type */ B_TRUE /* leaf vdev */ }; #endif Index: stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_geom.c =================================================================== --- stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_geom.c (revision 274799) +++ stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_geom.c (revision 274800) @@ -1,898 +1,903 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2006 Pawel Jakub Dawidek * All rights reserved. * * Portions Copyright (c) 2012 Martin Matuska */ #include #include #include #include #include #include #include #include #include #include #include #include /* * Virtual device vector for GEOM. */ static g_attrchanged_t vdev_geom_attrchanged; struct g_class zfs_vdev_class = { .name = "ZFS::VDEV", .version = G_VERSION, .attrchanged = vdev_geom_attrchanged, }; DECLARE_GEOM_CLASS(zfs_vdev_class, zfs_vdev); SYSCTL_DECL(_vfs_zfs_vdev); /* Don't send BIO_FLUSH. */ static int vdev_geom_bio_flush_disable = 0; TUNABLE_INT("vfs.zfs.vdev.bio_flush_disable", &vdev_geom_bio_flush_disable); SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, bio_flush_disable, CTLFLAG_RW, &vdev_geom_bio_flush_disable, 0, "Disable BIO_FLUSH"); /* Don't send BIO_DELETE. */ static int vdev_geom_bio_delete_disable = 0; TUNABLE_INT("vfs.zfs.vdev.bio_delete_disable", &vdev_geom_bio_delete_disable); SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, bio_delete_disable, CTLFLAG_RW, &vdev_geom_bio_delete_disable, 0, "Disable BIO_DELETE"); static void vdev_geom_set_rotation_rate(vdev_t *vd, struct g_consumer *cp) { int error; uint16_t rate; error = g_getattr("GEOM::rotation_rate", cp, &rate); if (error == 0) vd->vdev_rotation_rate = rate; else vd->vdev_rotation_rate = VDEV_RATE_UNKNOWN; } static void vdev_geom_attrchanged(struct g_consumer *cp, const char *attr) { vdev_t *vd; vd = cp->private; if (vd == NULL) return; if (strcmp(attr, "GEOM::rotation_rate") == 0) { vdev_geom_set_rotation_rate(vd, cp); return; } } static void vdev_geom_orphan(struct g_consumer *cp) { vdev_t *vd; g_topology_assert(); vd = cp->private; if (vd == NULL) return; /* * Orphan callbacks occur from the GEOM event thread. * Concurrent with this call, new I/O requests may be * working their way through GEOM about to find out * (only once executed by the g_down thread) that we've * been orphaned from our disk provider. These I/Os * must be retired before we can detach our consumer. * This is most easily achieved by acquiring the * SPA ZIO configuration lock as a writer, but doing * so with the GEOM topology lock held would cause * a lock order reversal. Instead, rely on the SPA's * async removal support to invoke a close on this * vdev once it is safe to do so. */ zfs_post_remove(vd->vdev_spa, vd); vd->vdev_remove_wanted = B_TRUE; spa_async_request(vd->vdev_spa, SPA_ASYNC_REMOVE); } static struct g_consumer * vdev_geom_attach(struct g_provider *pp) { struct g_geom *gp; struct g_consumer *cp; g_topology_assert(); ZFS_LOG(1, "Attaching to %s.", pp->name); /* Do we have geom already? No? Create one. */ LIST_FOREACH(gp, &zfs_vdev_class.geom, geom) { if (gp->flags & G_GEOM_WITHER) continue; if (strcmp(gp->name, "zfs::vdev") != 0) continue; break; } if (gp == NULL) { gp = g_new_geomf(&zfs_vdev_class, "zfs::vdev"); gp->orphan = vdev_geom_orphan; cp = g_new_consumer(gp); if (g_attach(cp, pp) != 0) { g_wither_geom(gp, ENXIO); return (NULL); } if (g_access(cp, 1, 0, 1) != 0) { g_wither_geom(gp, ENXIO); return (NULL); } ZFS_LOG(1, "Created geom and consumer for %s.", pp->name); } else { /* Check if we are already connected to this provider. */ LIST_FOREACH(cp, &gp->consumer, consumer) { if (cp->provider == pp) { ZFS_LOG(1, "Found consumer for %s.", pp->name); break; } } if (cp == NULL) { cp = g_new_consumer(gp); if (g_attach(cp, pp) != 0) { g_destroy_consumer(cp); return (NULL); } if (g_access(cp, 1, 0, 1) != 0) { g_detach(cp); g_destroy_consumer(cp); return (NULL); } ZFS_LOG(1, "Created consumer for %s.", pp->name); } else { if (g_access(cp, 1, 0, 1) != 0) return (NULL); ZFS_LOG(1, "Used existing consumer for %s.", pp->name); } } cp->flags |= G_CF_DIRECT_SEND | G_CF_DIRECT_RECEIVE; return (cp); } static void vdev_geom_detach(void *arg, int flag __unused) { struct g_geom *gp; struct g_consumer *cp; g_topology_assert(); cp = arg; gp = cp->geom; ZFS_LOG(1, "Closing access to %s.", cp->provider->name); g_access(cp, -1, 0, -1); /* Destroy consumer on last close. */ if (cp->acr == 0 && cp->ace == 0) { ZFS_LOG(1, "Destroyed consumer to %s.", cp->provider->name); if (cp->acw > 0) g_access(cp, 0, -cp->acw, 0); g_detach(cp); g_destroy_consumer(cp); } /* Destroy geom if there are no consumers left. */ if (LIST_EMPTY(&gp->consumer)) { ZFS_LOG(1, "Destroyed geom %s.", gp->name); g_wither_geom(gp, ENXIO); } } static uint64_t nvlist_get_guid(nvlist_t *list) { uint64_t value; value = 0; nvlist_lookup_uint64(list, ZPOOL_CONFIG_GUID, &value); return (value); } static int vdev_geom_io(struct g_consumer *cp, int cmd, void *data, off_t offset, off_t size) { struct bio *bp; u_char *p; off_t off, maxio; int error; ASSERT((offset % cp->provider->sectorsize) == 0); ASSERT((size % cp->provider->sectorsize) == 0); bp = g_alloc_bio(); off = offset; offset += size; p = data; maxio = MAXPHYS - (MAXPHYS % cp->provider->sectorsize); error = 0; for (; off < offset; off += maxio, p += maxio, size -= maxio) { bzero(bp, sizeof(*bp)); bp->bio_cmd = cmd; bp->bio_done = NULL; bp->bio_offset = off; bp->bio_length = MIN(size, maxio); bp->bio_data = p; g_io_request(bp, cp); error = biowait(bp, "vdev_geom_io"); if (error != 0) break; } g_destroy_bio(bp); return (error); } static void vdev_geom_taste_orphan(struct g_consumer *cp) { KASSERT(1 == 0, ("%s called while tasting %s.", __func__, cp->provider->name)); } static int vdev_geom_read_config(struct g_consumer *cp, nvlist_t **config) { struct g_provider *pp; vdev_label_t *label; char *p, *buf; size_t buflen; uint64_t psize; off_t offset, size; uint64_t guid, state, txg; int error, l, len; g_topology_assert_not(); pp = cp->provider; ZFS_LOG(1, "Reading config from %s...", pp->name); psize = pp->mediasize; psize = P2ALIGN(psize, (uint64_t)sizeof(vdev_label_t)); size = sizeof(*label) + pp->sectorsize - ((sizeof(*label) - 1) % pp->sectorsize) - 1; guid = 0; label = kmem_alloc(size, KM_SLEEP); buflen = sizeof(label->vl_vdev_phys.vp_nvlist); *config = NULL; for (l = 0; l < VDEV_LABELS; l++) { offset = vdev_label_offset(psize, l, 0); if ((offset % pp->sectorsize) != 0) continue; if (vdev_geom_io(cp, BIO_READ, label, offset, size) != 0) continue; buf = label->vl_vdev_phys.vp_nvlist; if (nvlist_unpack(buf, buflen, config, 0) != 0) continue; if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_STATE, &state) != 0 || state > POOL_STATE_L2CACHE) { nvlist_free(*config); *config = NULL; continue; } if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE && (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_TXG, &txg) != 0 || txg == 0)) { nvlist_free(*config); *config = NULL; continue; } break; } kmem_free(label, size); return (*config == NULL ? ENOENT : 0); } static void resize_configs(nvlist_t ***configs, uint64_t *count, uint64_t id) { nvlist_t **new_configs; uint64_t i; if (id < *count) return; new_configs = kmem_zalloc((id + 1) * sizeof(nvlist_t *), KM_SLEEP); for (i = 0; i < *count; i++) new_configs[i] = (*configs)[i]; if (*configs != NULL) kmem_free(*configs, *count * sizeof(void *)); *configs = new_configs; *count = id + 1; } static void process_vdev_config(nvlist_t ***configs, uint64_t *count, nvlist_t *cfg, const char *name, uint64_t* known_pool_guid) { nvlist_t *vdev_tree; uint64_t pool_guid; uint64_t vdev_guid, known_guid; uint64_t id, txg, known_txg; char *pname; int i; if (nvlist_lookup_string(cfg, ZPOOL_CONFIG_POOL_NAME, &pname) != 0 || strcmp(pname, name) != 0) goto ignore; if (nvlist_lookup_uint64(cfg, ZPOOL_CONFIG_POOL_GUID, &pool_guid) != 0) goto ignore; if (nvlist_lookup_uint64(cfg, ZPOOL_CONFIG_TOP_GUID, &vdev_guid) != 0) goto ignore; if (nvlist_lookup_nvlist(cfg, ZPOOL_CONFIG_VDEV_TREE, &vdev_tree) != 0) goto ignore; if (nvlist_lookup_uint64(vdev_tree, ZPOOL_CONFIG_ID, &id) != 0) goto ignore; VERIFY(nvlist_lookup_uint64(cfg, ZPOOL_CONFIG_POOL_TXG, &txg) == 0); if (*known_pool_guid != 0) { if (pool_guid != *known_pool_guid) goto ignore; } else *known_pool_guid = pool_guid; resize_configs(configs, count, id); if ((*configs)[id] != NULL) { VERIFY(nvlist_lookup_uint64((*configs)[id], ZPOOL_CONFIG_POOL_TXG, &known_txg) == 0); if (txg <= known_txg) goto ignore; nvlist_free((*configs)[id]); } (*configs)[id] = cfg; return; ignore: nvlist_free(cfg); } static int vdev_geom_attach_taster(struct g_consumer *cp, struct g_provider *pp) { int error; if (pp->flags & G_PF_WITHER) return (EINVAL); g_attach(cp, pp); error = g_access(cp, 1, 0, 0); if (error == 0) { if (pp->sectorsize > VDEV_PAD_SIZE || !ISP2(pp->sectorsize)) error = EINVAL; else if (pp->mediasize < SPA_MINDEVSIZE) error = EINVAL; if (error != 0) g_access(cp, -1, 0, 0); } if (error != 0) g_detach(cp); return (error); } static void vdev_geom_detach_taster(struct g_consumer *cp) { g_access(cp, -1, 0, 0); g_detach(cp); } int vdev_geom_read_pool_label(const char *name, nvlist_t ***configs, uint64_t *count) { struct g_class *mp; struct g_geom *gp, *zgp; struct g_provider *pp; struct g_consumer *zcp; nvlist_t *vdev_cfg; uint64_t pool_guid; int error; DROP_GIANT(); g_topology_lock(); zgp = g_new_geomf(&zfs_vdev_class, "zfs::vdev::taste"); /* This orphan function should be never called. */ zgp->orphan = vdev_geom_taste_orphan; zcp = g_new_consumer(zgp); *configs = NULL; *count = 0; pool_guid = 0; LIST_FOREACH(mp, &g_classes, class) { if (mp == &zfs_vdev_class) continue; LIST_FOREACH(gp, &mp->geom, geom) { if (gp->flags & G_GEOM_WITHER) continue; LIST_FOREACH(pp, &gp->provider, provider) { if (pp->flags & G_PF_WITHER) continue; if (vdev_geom_attach_taster(zcp, pp) != 0) continue; g_topology_unlock(); error = vdev_geom_read_config(zcp, &vdev_cfg); g_topology_lock(); vdev_geom_detach_taster(zcp); if (error) continue; ZFS_LOG(1, "successfully read vdev config"); process_vdev_config(configs, count, vdev_cfg, name, &pool_guid); } } } g_destroy_consumer(zcp); g_destroy_geom(zgp); g_topology_unlock(); PICKUP_GIANT(); return (*count > 0 ? 0 : ENOENT); } static uint64_t vdev_geom_read_guid(struct g_consumer *cp) { nvlist_t *config; uint64_t guid; g_topology_assert_not(); guid = 0; if (vdev_geom_read_config(cp, &config) == 0) { guid = nvlist_get_guid(config); nvlist_free(config); } return (guid); } static struct g_consumer * vdev_geom_attach_by_guid(uint64_t guid) { struct g_class *mp; struct g_geom *gp, *zgp; struct g_provider *pp; struct g_consumer *cp, *zcp; uint64_t pguid; g_topology_assert(); zgp = g_new_geomf(&zfs_vdev_class, "zfs::vdev::taste"); /* This orphan function should be never called. */ zgp->orphan = vdev_geom_taste_orphan; zcp = g_new_consumer(zgp); cp = NULL; LIST_FOREACH(mp, &g_classes, class) { if (mp == &zfs_vdev_class) continue; LIST_FOREACH(gp, &mp->geom, geom) { if (gp->flags & G_GEOM_WITHER) continue; LIST_FOREACH(pp, &gp->provider, provider) { if (vdev_geom_attach_taster(zcp, pp) != 0) continue; g_topology_unlock(); pguid = vdev_geom_read_guid(zcp); g_topology_lock(); vdev_geom_detach_taster(zcp); if (pguid != guid) continue; cp = vdev_geom_attach(pp); if (cp == NULL) { printf("ZFS WARNING: Unable to attach to %s.\n", pp->name); continue; } break; } if (cp != NULL) break; } if (cp != NULL) break; } end: g_destroy_consumer(zcp); g_destroy_geom(zgp); return (cp); } static struct g_consumer * vdev_geom_open_by_guid(vdev_t *vd) { struct g_consumer *cp; char *buf; size_t len; g_topology_assert(); ZFS_LOG(1, "Searching by guid [%ju].", (uintmax_t)vd->vdev_guid); cp = vdev_geom_attach_by_guid(vd->vdev_guid); if (cp != NULL) { len = strlen(cp->provider->name) + strlen("/dev/") + 1; buf = kmem_alloc(len, KM_SLEEP); snprintf(buf, len, "/dev/%s", cp->provider->name); spa_strfree(vd->vdev_path); vd->vdev_path = buf; ZFS_LOG(1, "Attach by guid [%ju] succeeded, provider %s.", (uintmax_t)vd->vdev_guid, vd->vdev_path); } else { ZFS_LOG(1, "Search by guid [%ju] failed.", (uintmax_t)vd->vdev_guid); } return (cp); } static struct g_consumer * vdev_geom_open_by_path(vdev_t *vd, int check_guid) { struct g_provider *pp; struct g_consumer *cp; uint64_t guid; g_topology_assert(); cp = NULL; pp = g_provider_by_name(vd->vdev_path + sizeof("/dev/") - 1); if (pp != NULL) { ZFS_LOG(1, "Found provider by name %s.", vd->vdev_path); cp = vdev_geom_attach(pp); if (cp != NULL && check_guid && ISP2(pp->sectorsize) && pp->sectorsize <= VDEV_PAD_SIZE) { g_topology_unlock(); guid = vdev_geom_read_guid(cp); g_topology_lock(); if (guid != vd->vdev_guid) { vdev_geom_detach(cp, 0); cp = NULL; ZFS_LOG(1, "guid mismatch for provider %s: " "%ju != %ju.", vd->vdev_path, (uintmax_t)vd->vdev_guid, (uintmax_t)guid); } else { ZFS_LOG(1, "guid match for provider %s.", vd->vdev_path); } } } return (cp); } static int vdev_geom_open(vdev_t *vd, uint64_t *psize, uint64_t *max_psize, uint64_t *logical_ashift, uint64_t *physical_ashift) { struct g_provider *pp; struct g_consumer *cp; size_t bufsize; int error; /* * We must have a pathname, and it must be absolute. */ if (vd->vdev_path == NULL || vd->vdev_path[0] != '/') { vd->vdev_stat.vs_aux = VDEV_AUX_BAD_LABEL; return (EINVAL); } vd->vdev_tsd = NULL; DROP_GIANT(); g_topology_lock(); error = 0; /* * If we're creating or splitting a pool, just find the GEOM provider * by its name and ignore GUID mismatches. */ if (vd->vdev_spa->spa_load_state == SPA_LOAD_NONE || vd->vdev_spa->spa_splitting_newspa == B_TRUE) cp = vdev_geom_open_by_path(vd, 0); else { cp = vdev_geom_open_by_path(vd, 1); if (cp == NULL) { /* * The device at vd->vdev_path doesn't have the * expected guid. The disks might have merely * moved around so try all other GEOM providers * to find one with the right guid. */ cp = vdev_geom_open_by_guid(vd); } } if (cp == NULL) { ZFS_LOG(1, "Provider %s not found.", vd->vdev_path); error = ENOENT; } else if (cp->provider->sectorsize > VDEV_PAD_SIZE || !ISP2(cp->provider->sectorsize)) { ZFS_LOG(1, "Provider %s has unsupported sectorsize.", vd->vdev_path); vdev_geom_detach(cp, 0); error = EINVAL; cp = NULL; } else if (cp->acw == 0 && (spa_mode(vd->vdev_spa) & FWRITE) != 0) { int i; for (i = 0; i < 5; i++) { error = g_access(cp, 0, 1, 0); if (error == 0) break; g_topology_unlock(); tsleep(vd, 0, "vdev", hz / 2); g_topology_lock(); } if (error != 0) { printf("ZFS WARNING: Unable to open %s for writing (error=%d).\n", vd->vdev_path, error); vdev_geom_detach(cp, 0); cp = NULL; } } g_topology_unlock(); PICKUP_GIANT(); if (cp == NULL) { vd->vdev_stat.vs_aux = VDEV_AUX_OPEN_FAILED; return (error); } cp->private = vd; vd->vdev_tsd = cp; pp = cp->provider; /* * Determine the actual size of the device. */ *max_psize = *psize = pp->mediasize; /* * Determine the device's minimum transfer size and preferred * transfer size. */ *logical_ashift = highbit(MAX(pp->sectorsize, SPA_MINBLOCKSIZE)) - 1; *physical_ashift = 0; if (pp->stripesize) *physical_ashift = highbit(pp->stripesize) - 1; /* * Clear the nowritecache settings, so that on a vdev_reopen() * we will try again. */ vd->vdev_nowritecache = B_FALSE; if (vd->vdev_physpath != NULL) spa_strfree(vd->vdev_physpath); bufsize = sizeof("/dev/") + strlen(pp->name); vd->vdev_physpath = kmem_alloc(bufsize, KM_SLEEP); snprintf(vd->vdev_physpath, bufsize, "/dev/%s", pp->name); /* * Determine the device's rotation rate. */ vdev_geom_set_rotation_rate(vd, cp); return (0); } static void vdev_geom_close(vdev_t *vd) { struct g_consumer *cp; cp = vd->vdev_tsd; if (cp == NULL) return; vd->vdev_tsd = NULL; vd->vdev_delayed_close = B_FALSE; cp->private = NULL; /* XXX locking */ g_post_event(vdev_geom_detach, cp, M_WAITOK, NULL); } static void vdev_geom_io_intr(struct bio *bp) { vdev_t *vd; zio_t *zio; zio = bp->bio_caller1; vd = zio->io_vd; zio->io_error = bp->bio_error; if (zio->io_error == 0 && bp->bio_resid != 0) zio->io_error = SET_ERROR(EIO); switch(zio->io_error) { case ENOTSUP: /* * If we get ENOTSUP for BIO_FLUSH or BIO_DELETE we know * that future attempts will never succeed. In this case * we set a persistent flag so that we don't bother with * requests in the future. */ switch(bp->bio_cmd) { case BIO_FLUSH: vd->vdev_nowritecache = B_TRUE; break; case BIO_DELETE: vd->vdev_notrim = B_TRUE; break; } break; case ENXIO: if (!vd->vdev_remove_wanted) { /* * If provider's error is set we assume it is being * removed. */ if (bp->bio_to->error != 0) { vd->vdev_remove_wanted = B_TRUE; spa_async_request(zio->io_spa, SPA_ASYNC_REMOVE); } else if (!vd->vdev_delayed_close) { vd->vdev_delayed_close = B_TRUE; } } break; } g_destroy_bio(bp); zio_interrupt(zio); } static int vdev_geom_io_start(zio_t *zio) { vdev_t *vd; struct g_consumer *cp; struct bio *bp; int error; vd = zio->io_vd; switch (zio->io_type) { case ZIO_TYPE_IOCTL: /* XXPOLICY */ if (!vdev_readable(vd)) { zio->io_error = SET_ERROR(ENXIO); } else { switch (zio->io_cmd) { case DKIOCFLUSHWRITECACHE: if (zfs_nocacheflush || vdev_geom_bio_flush_disable) break; if (vd->vdev_nowritecache) { zio->io_error = SET_ERROR(ENOTSUP); break; } goto sendreq; default: zio->io_error = SET_ERROR(ENOTSUP); } } zio_interrupt(zio); return (ZIO_PIPELINE_STOP); case ZIO_TYPE_FREE: if (vd->vdev_notrim) { zio->io_error = SET_ERROR(ENOTSUP); } else if (!vdev_geom_bio_delete_disable) { goto sendreq; } zio_interrupt(zio); return (ZIO_PIPELINE_STOP); } sendreq: + ASSERT(zio->io_type == ZIO_TYPE_READ || + zio->io_type == ZIO_TYPE_WRITE || + zio->io_type == ZIO_TYPE_FREE || + zio->io_type == ZIO_TYPE_IOCTL); + cp = vd->vdev_tsd; if (cp == NULL) { zio->io_error = SET_ERROR(ENXIO); zio_interrupt(zio); return (ZIO_PIPELINE_STOP); } bp = g_alloc_bio(); bp->bio_caller1 = zio; switch (zio->io_type) { case ZIO_TYPE_READ: case ZIO_TYPE_WRITE: bp->bio_cmd = zio->io_type == ZIO_TYPE_READ ? BIO_READ : BIO_WRITE; bp->bio_data = zio->io_data; bp->bio_offset = zio->io_offset; bp->bio_length = zio->io_size; break; case ZIO_TYPE_FREE: bp->bio_cmd = BIO_DELETE; bp->bio_data = NULL; bp->bio_offset = zio->io_offset; bp->bio_length = zio->io_size; break; case ZIO_TYPE_IOCTL: bp->bio_cmd = BIO_FLUSH; bp->bio_flags |= BIO_ORDERED; bp->bio_data = NULL; bp->bio_offset = cp->provider->mediasize; bp->bio_length = 0; break; } bp->bio_done = vdev_geom_io_intr; g_io_request(bp, cp); return (ZIO_PIPELINE_STOP); } static void vdev_geom_io_done(zio_t *zio) { } static void vdev_geom_hold(vdev_t *vd) { } static void vdev_geom_rele(vdev_t *vd) { } vdev_ops_t vdev_geom_ops = { vdev_geom_open, vdev_geom_close, vdev_default_asize, vdev_geom_io_start, vdev_geom_io_done, NULL, vdev_geom_hold, vdev_geom_rele, VDEV_TYPE_DISK, /* name of this vdev type */ B_TRUE /* leaf vdev */ }; Index: stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_label.c =================================================================== --- stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_label.c (revision 274799) +++ stable/10/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_label.c (revision 274800) @@ -1,1292 +1,1293 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2013 by Delphix. All rights reserved. */ /* * Virtual Device Labels * --------------------- * * The vdev label serves several distinct purposes: * * 1. Uniquely identify this device as part of a ZFS pool and confirm its * identity within the pool. * * 2. Verify that all the devices given in a configuration are present * within the pool. * * 3. Determine the uberblock for the pool. * * 4. In case of an import operation, determine the configuration of the * toplevel vdev of which it is a part. * * 5. If an import operation cannot find all the devices in the pool, * provide enough information to the administrator to determine which * devices are missing. * * It is important to note that while the kernel is responsible for writing the * label, it only consumes the information in the first three cases. The * latter information is only consumed in userland when determining the * configuration to import a pool. * * * Label Organization * ------------------ * * Before describing the contents of the label, it's important to understand how * the labels are written and updated with respect to the uberblock. * * When the pool configuration is altered, either because it was newly created * or a device was added, we want to update all the labels such that we can deal * with fatal failure at any point. To this end, each disk has two labels which * are updated before and after the uberblock is synced. Assuming we have * labels and an uberblock with the following transaction groups: * * L1 UB L2 * +------+ +------+ +------+ * | | | | | | * | t10 | | t10 | | t10 | * | | | | | | * +------+ +------+ +------+ * * In this stable state, the labels and the uberblock were all updated within * the same transaction group (10). Each label is mirrored and checksummed, so * that we can detect when we fail partway through writing the label. * * In order to identify which labels are valid, the labels are written in the * following manner: * * 1. For each vdev, update 'L1' to the new label * 2. Update the uberblock * 3. For each vdev, update 'L2' to the new label * * Given arbitrary failure, we can determine the correct label to use based on * the transaction group. If we fail after updating L1 but before updating the * UB, we will notice that L1's transaction group is greater than the uberblock, * so L2 must be valid. If we fail after writing the uberblock but before * writing L2, we will notice that L2's transaction group is less than L1, and * therefore L1 is valid. * * Another added complexity is that not every label is updated when the config * is synced. If we add a single device, we do not want to have to re-write * every label for every device in the pool. This means that both L1 and L2 may * be older than the pool uberblock, because the necessary information is stored * on another vdev. * * * On-disk Format * -------------- * * The vdev label consists of two distinct parts, and is wrapped within the * vdev_label_t structure. The label includes 8k of padding to permit legacy * VTOC disk labels, but is otherwise ignored. * * The first half of the label is a packed nvlist which contains pool wide * properties, per-vdev properties, and configuration information. It is * described in more detail below. * * The latter half of the label consists of a redundant array of uberblocks. * These uberblocks are updated whenever a transaction group is committed, * or when the configuration is updated. When a pool is loaded, we scan each * vdev for the 'best' uberblock. * * * Configuration Information * ------------------------- * * The nvlist describing the pool and vdev contains the following elements: * * version ZFS on-disk version * name Pool name * state Pool state * txg Transaction group in which this label was written * pool_guid Unique identifier for this pool * vdev_tree An nvlist describing vdev tree. * features_for_read * An nvlist of the features necessary for reading the MOS. * * Each leaf device label also contains the following: * * top_guid Unique ID for top-level vdev in which this is contained * guid Unique ID for the leaf vdev * * The 'vs' configuration follows the format described in 'spa_config.c'. */ #include #include #include #include #include #include #include #include #include #include #include #include #include static boolean_t vdev_trim_on_init = B_TRUE; SYSCTL_DECL(_vfs_zfs_vdev); SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, trim_on_init, CTLFLAG_RW, &vdev_trim_on_init, 0, "Enable/disable full vdev trim on initialisation"); /* * Basic routines to read and write from a vdev label. * Used throughout the rest of this file. */ uint64_t vdev_label_offset(uint64_t psize, int l, uint64_t offset) { ASSERT(offset < sizeof (vdev_label_t)); ASSERT(P2PHASE_TYPED(psize, sizeof (vdev_label_t), uint64_t) == 0); return (offset + l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ? 0 : psize - VDEV_LABELS * sizeof (vdev_label_t))); } /* * Returns back the vdev label associated with the passed in offset. */ int vdev_label_number(uint64_t psize, uint64_t offset) { int l; if (offset >= psize - VDEV_LABEL_END_SIZE) { offset -= psize - VDEV_LABEL_END_SIZE; offset += (VDEV_LABELS / 2) * sizeof (vdev_label_t); } l = offset / sizeof (vdev_label_t); return (l < VDEV_LABELS ? l : -1); } 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) { ASSERT(spa_config_held(zio->io_spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); ASSERT(flags & ZIO_FLAG_CONFIG_WRITER); zio_nowait(zio_read_phys(zio, vd, vdev_label_offset(vd->vdev_psize, l, offset), size, buf, ZIO_CHECKSUM_LABEL, done, private, 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) { 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(flags & ZIO_FLAG_CONFIG_WRITER); zio_nowait(zio_write_phys(zio, vd, vdev_label_offset(vd->vdev_psize, l, offset), size, buf, ZIO_CHECKSUM_LABEL, done, private, ZIO_PRIORITY_SYNC_WRITE, flags, B_TRUE)); } /* * Generate the nvlist representing this vdev's config. */ nvlist_t * vdev_config_generate(spa_t *spa, vdev_t *vd, boolean_t getstats, vdev_config_flag_t flags) { nvlist_t *nv = NULL; 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_uint64(nv, ZPOOL_CONFIG_ID, vd->vdev_id); fnvlist_add_uint64(nv, ZPOOL_CONFIG_GUID, vd->vdev_guid); if (vd->vdev_path != NULL) fnvlist_add_string(nv, ZPOOL_CONFIG_PATH, vd->vdev_path); if (vd->vdev_devid != NULL) fnvlist_add_string(nv, ZPOOL_CONFIG_DEVID, vd->vdev_devid); if (vd->vdev_physpath != NULL) fnvlist_add_string(nv, ZPOOL_CONFIG_PHYS_PATH, vd->vdev_physpath); if (vd->vdev_fru != NULL) fnvlist_add_string(nv, ZPOOL_CONFIG_FRU, vd->vdev_fru); if (vd->vdev_nparity != 0) { ASSERT(strcmp(vd->vdev_ops->vdev_op_type, VDEV_TYPE_RAIDZ) == 0); /* * Make sure someone hasn't managed to sneak a fancy new vdev * into a crufty old storage pool. */ ASSERT(vd->vdev_nparity == 1 || (vd->vdev_nparity <= 2 && spa_version(spa) >= SPA_VERSION_RAIDZ2) || (vd->vdev_nparity <= 3 && spa_version(spa) >= SPA_VERSION_RAIDZ3)); /* * Note that we'll add the nparity tag even on storage pools * that only support a single parity device -- older software * will just ignore it. */ fnvlist_add_uint64(nv, ZPOOL_CONFIG_NPARITY, vd->vdev_nparity); } if (vd->vdev_wholedisk != -1ULL) fnvlist_add_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, vd->vdev_wholedisk); if (vd->vdev_not_present) fnvlist_add_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, 1); if (vd->vdev_isspare) fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_SPARE, 1); if (!(flags & (VDEV_CONFIG_SPARE | VDEV_CONFIG_L2CACHE)) && vd == vd->vdev_top) { fnvlist_add_uint64(nv, ZPOOL_CONFIG_METASLAB_ARRAY, vd->vdev_ms_array); fnvlist_add_uint64(nv, ZPOOL_CONFIG_METASLAB_SHIFT, vd->vdev_ms_shift); fnvlist_add_uint64(nv, ZPOOL_CONFIG_ASHIFT, vd->vdev_ashift); fnvlist_add_uint64(nv, ZPOOL_CONFIG_ASIZE, vd->vdev_asize); fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_LOG, vd->vdev_islog); if (vd->vdev_removing) fnvlist_add_uint64(nv, ZPOOL_CONFIG_REMOVING, vd->vdev_removing); } if (vd->vdev_dtl_sm != NULL) { fnvlist_add_uint64(nv, ZPOOL_CONFIG_DTL, space_map_object(vd->vdev_dtl_sm)); } if (vd->vdev_crtxg) fnvlist_add_uint64(nv, ZPOOL_CONFIG_CREATE_TXG, vd->vdev_crtxg); 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)); /* provide either current or previous scan information */ if (spa_scan_get_stats(spa, &ps) == 0) { fnvlist_add_uint64_array(nv, ZPOOL_CONFIG_SCAN_STATS, (uint64_t *)&ps, sizeof (pool_scan_stat_t) / sizeof (uint64_t)); } } if (!vd->vdev_ops->vdev_op_leaf) { nvlist_t **child; int c, idx; ASSERT(!vd->vdev_ishole); child = kmem_alloc(vd->vdev_children * sizeof (nvlist_t *), KM_SLEEP); for (c = 0, idx = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; /* * If we're generating an nvlist of removing * vdevs then skip over any device which is * not being removed. */ if ((flags & VDEV_CONFIG_REMOVING) && !cvd->vdev_removing) continue; child[idx++] = vdev_config_generate(spa, cvd, getstats, flags); } if (idx) { fnvlist_add_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, child, idx); } for (c = 0; c < idx; c++) nvlist_free(child[c]); kmem_free(child, vd->vdev_children * sizeof (nvlist_t *)); } else { const char *aux = NULL; if (vd->vdev_offline && !vd->vdev_tmpoffline) fnvlist_add_uint64(nv, ZPOOL_CONFIG_OFFLINE, B_TRUE); if (vd->vdev_resilver_txg != 0) fnvlist_add_uint64(nv, ZPOOL_CONFIG_RESILVER_TXG, vd->vdev_resilver_txg); if (vd->vdev_faulted) fnvlist_add_uint64(nv, ZPOOL_CONFIG_FAULTED, B_TRUE); if (vd->vdev_degraded) fnvlist_add_uint64(nv, ZPOOL_CONFIG_DEGRADED, B_TRUE); if (vd->vdev_removed) fnvlist_add_uint64(nv, ZPOOL_CONFIG_REMOVED, B_TRUE); if (vd->vdev_unspare) fnvlist_add_uint64(nv, ZPOOL_CONFIG_UNSPARE, B_TRUE); if (vd->vdev_ishole) fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_HOLE, B_TRUE); switch (vd->vdev_stat.vs_aux) { case VDEV_AUX_ERR_EXCEEDED: aux = "err_exceeded"; break; case VDEV_AUX_EXTERNAL: aux = "external"; break; } if (aux != NULL) fnvlist_add_string(nv, ZPOOL_CONFIG_AUX_STATE, aux); if (vd->vdev_splitting && vd->vdev_orig_guid != 0LL) { fnvlist_add_uint64(nv, ZPOOL_CONFIG_ORIG_GUID, vd->vdev_orig_guid); } } return (nv); } /* * Generate a view of the top-level vdevs. If we currently have holes * in the namespace, then generate an array which contains a list of holey * vdevs. Additionally, add the number of top-level children that currently * exist. */ void vdev_top_config_generate(spa_t *spa, nvlist_t *config) { vdev_t *rvd = spa->spa_root_vdev; uint64_t *array; uint_t c, idx; 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]; if (tvd->vdev_ishole) array[idx++] = c; } if (idx) { VERIFY(nvlist_add_uint64_array(config, ZPOOL_CONFIG_HOLE_ARRAY, array, idx) == 0); } VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_VDEV_CHILDREN, rvd->vdev_children) == 0); kmem_free(array, rvd->vdev_children * sizeof (uint64_t)); } /* * Returns the configuration from the label of the given vdev. For vdevs * which don't have a txg value stored on their label (i.e. spares/cache) * or have not been completely initialized (txg = 0) just return * the configuration from the first valid label we find. Otherwise, * find the most up-to-date label that does not exceed the specified * 'txg' value. */ nvlist_t * vdev_label_read_config(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; nvlist_t *config = NULL; vdev_phys_t *vp; zio_t *zio; uint64_t best_txg = 0; int error = 0; int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE; 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)); retry: 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, offsetof(vdev_label_t, vl_vdev_phys), sizeof (vdev_phys_t), NULL, NULL, flags); if (zio_wait(zio) == 0 && nvlist_unpack(vp->vp_nvlist, sizeof (vp->vp_nvlist), &label, 0) == 0) { uint64_t label_txg = 0; /* * Auxiliary vdevs won't have txg values in their * labels and newly added vdevs may not have been * completely initialized so just return the * configuration from the first valid label we * encounter. */ error = nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_TXG, &label_txg); if ((error || label_txg == 0) && !config) { config = label; break; } else if (label_txg <= txg && label_txg > best_txg) { best_txg = label_txg; nvlist_free(config); config = fnvlist_dup(label); } } if (label != NULL) { nvlist_free(label); label = NULL; } } if (config == NULL && !(flags & ZIO_FLAG_TRYHARD)) { flags |= ZIO_FLAG_TRYHARD; goto retry; } zio_buf_free(vp, sizeof (vdev_phys_t)); return (config); } /* * Determine if a device is in use. The 'spare_guid' parameter will be filled * in with the device guid if this spare is active elsewhere on the system. */ static boolean_t vdev_inuse(vdev_t *vd, uint64_t crtxg, vdev_labeltype_t reason, uint64_t *spare_guid, uint64_t *l2cache_guid) { spa_t *spa = vd->vdev_spa; uint64_t state, pool_guid, device_guid, txg, spare_pool; uint64_t vdtxg = 0; nvlist_t *label; if (spare_guid) *spare_guid = 0ULL; if (l2cache_guid) *l2cache_guid = 0ULL; /* * Read the label, if any, and perform some basic sanity checks. */ if ((label = vdev_label_read_config(vd, -1ULL)) == NULL) return (B_FALSE); (void) nvlist_lookup_uint64(label, ZPOOL_CONFIG_CREATE_TXG, &vdtxg); if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE, &state) != 0 || nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &device_guid) != 0) { nvlist_free(label); return (B_FALSE); } if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE && (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID, &pool_guid) != 0 || nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_TXG, &txg) != 0)) { nvlist_free(label); return (B_FALSE); } nvlist_free(label); /* * Check to see if this device indeed belongs to the pool it claims to * be a part of. The only way this is allowed is if the device is a hot * spare (which we check for later on). */ if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE && !spa_guid_exists(pool_guid, device_guid) && !spa_spare_exists(device_guid, NULL, NULL) && !spa_l2cache_exists(device_guid, NULL)) return (B_FALSE); /* * If the transaction group is zero, then this an initialized (but * unused) label. This is only an error if the create transaction * on-disk is the same as the one we're using now, in which case the * user has attempted to add the same vdev multiple times in the same * transaction. */ if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE && txg == 0 && vdtxg == crtxg) return (B_TRUE); /* * Check to see if this is a spare device. We do an explicit check for * spa_has_spare() here because it may be on our pending list of spares * to add. We also check if it is an l2cache device. */ if (spa_spare_exists(device_guid, &spare_pool, NULL) || spa_has_spare(spa, device_guid)) { if (spare_guid) *spare_guid = device_guid; switch (reason) { case VDEV_LABEL_CREATE: case VDEV_LABEL_L2CACHE: return (B_TRUE); case VDEV_LABEL_REPLACE: return (!spa_has_spare(spa, device_guid) || spare_pool != 0ULL); case VDEV_LABEL_SPARE: return (spa_has_spare(spa, device_guid)); } } /* * Check to see if this is an l2cache device. */ if (spa_l2cache_exists(device_guid, NULL)) return (B_TRUE); /* * We can't rely on a pool's state if it's been imported * 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 && spa_mode(spa) == FREAD) state = POOL_STATE_ACTIVE; /* * If the device is marked ACTIVE, then this device is in use by another * pool on the system. */ return (state == POOL_STATE_ACTIVE); } /* * Initialize a vdev label. We check to make sure each leaf device is not in * use, and writable. We put down an initial label which we will later * overwrite with a complete label. Note that it's important to do this * sequentially, not in parallel, so that we catch cases of multiple use of the * same leaf vdev in the vdev we're creating -- e.g. mirroring a disk with * itself. */ int vdev_label_init(vdev_t *vd, uint64_t crtxg, vdev_labeltype_t reason) { spa_t *spa = vd->vdev_spa; nvlist_t *label; vdev_phys_t *vp; char *pad2; uberblock_t *ub; zio_t *zio; char *buf; size_t buflen; int error; uint64_t spare_guid, l2cache_guid; int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL; ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); for (int c = 0; c < vd->vdev_children; c++) if ((error = vdev_label_init(vd->vdev_child[c], crtxg, reason)) != 0) return (error); /* Track the creation time for this vdev */ vd->vdev_crtxg = crtxg; if (!vd->vdev_ops->vdev_op_leaf || !spa_writeable(spa)) return (0); /* * Dead vdevs cannot be initialized. */ if (vdev_is_dead(vd)) return (SET_ERROR(EIO)); /* * Determine if the vdev is in use. */ if (reason != VDEV_LABEL_REMOVE && reason != VDEV_LABEL_SPLIT && vdev_inuse(vd, crtxg, reason, &spare_guid, &l2cache_guid)) return (SET_ERROR(EBUSY)); /* * If this is a request to add or replace a spare or l2cache device * that is in use elsewhere on the system, then we must update the * guid (which was initialized to a random value) to reflect the * actual GUID (which is shared between multiple pools). */ if (reason != VDEV_LABEL_REMOVE && reason != VDEV_LABEL_L2CACHE && spare_guid != 0ULL) { uint64_t guid_delta = spare_guid - vd->vdev_guid; vd->vdev_guid += guid_delta; for (vdev_t *pvd = vd; pvd != NULL; pvd = pvd->vdev_parent) pvd->vdev_guid_sum += guid_delta; /* * If this is a replacement, then we want to fallthrough to the * rest of the code. If we're adding a spare, then it's already * labeled appropriately and we can just return. */ if (reason == VDEV_LABEL_SPARE) return (0); ASSERT(reason == VDEV_LABEL_REPLACE || reason == VDEV_LABEL_SPLIT); } if (reason != VDEV_LABEL_REMOVE && reason != VDEV_LABEL_SPARE && l2cache_guid != 0ULL) { uint64_t guid_delta = l2cache_guid - vd->vdev_guid; vd->vdev_guid += guid_delta; for (vdev_t *pvd = vd; pvd != NULL; pvd = pvd->vdev_parent) pvd->vdev_guid_sum += guid_delta; /* * If this is a replacement, then we want to fallthrough to the * rest of the code. If we're adding an l2cache, then it's * already labeled appropriately and we can just return. */ if (reason == VDEV_LABEL_L2CACHE) return (0); ASSERT(reason == VDEV_LABEL_REPLACE); } /* * TRIM the whole thing so that we start with a clean slate. * It's just an optimization, so we don't care if it fails. * Don't TRIM if removing so that we don't interfere with zpool * disaster recovery. */ - if (zfs_trim_enabled && vdev_trim_on_init && (reason == VDEV_LABEL_CREATE || - reason == VDEV_LABEL_SPARE || reason == VDEV_LABEL_L2CACHE)) + if (zfs_trim_enabled && vdev_trim_on_init && !vd->vdev_notrim && + (reason == VDEV_LABEL_CREATE || reason == VDEV_LABEL_SPARE || + reason == VDEV_LABEL_L2CACHE)) zio_wait(zio_trim(NULL, spa, vd, 0, vd->vdev_psize)); /* * Initialize its label. */ vp = zio_buf_alloc(sizeof (vdev_phys_t)); bzero(vp, sizeof (vdev_phys_t)); /* * Generate a label describing the pool and our top-level vdev. * We mark it as being from txg 0 to indicate that it's not * really part of an active pool just yet. The labels will * be written again with a meaningful txg by spa_sync(). */ if (reason == VDEV_LABEL_SPARE || (reason == VDEV_LABEL_REMOVE && vd->vdev_isspare)) { /* * For inactive hot spares, we generate a special label that * identifies as a mutually shared hot spare. We write the * label if we are adding a hot spare, or if we are removing an * active hot spare (in which case we want to revert the * labels). */ VERIFY(nvlist_alloc(&label, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_VERSION, spa_version(spa)) == 0); VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_POOL_STATE, POOL_STATE_SPARE) == 0); VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_GUID, vd->vdev_guid) == 0); } else if (reason == VDEV_LABEL_L2CACHE || (reason == VDEV_LABEL_REMOVE && vd->vdev_isl2cache)) { /* * For level 2 ARC devices, add a special label. */ VERIFY(nvlist_alloc(&label, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_VERSION, spa_version(spa)) == 0); VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_POOL_STATE, POOL_STATE_L2CACHE) == 0); VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_GUID, vd->vdev_guid) == 0); } else { uint64_t txg = 0ULL; if (reason == VDEV_LABEL_SPLIT) txg = spa->spa_uberblock.ub_txg; label = spa_config_generate(spa, vd, txg, B_FALSE); /* * Add our creation time. This allows us to detect multiple * vdev uses as described above, and automatically expires if we * fail. */ VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_CREATE_TXG, crtxg) == 0); } buf = vp->vp_nvlist; buflen = sizeof (vp->vp_nvlist); 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)); /* EFAULT means nvlist_pack ran out of room */ return (error == EFAULT ? ENAMETOOLONG : EINVAL); } /* * Initialize uberblock template. */ ub = zio_buf_alloc(VDEV_UBERBLOCK_RING); bzero(ub, VDEV_UBERBLOCK_RING); *ub = spa->spa_uberblock; ub->ub_txg = 0; /* Initialize the 2nd padding area. */ pad2 = zio_buf_alloc(VDEV_PAD_SIZE); bzero(pad2, VDEV_PAD_SIZE); /* * Write everything in parallel. */ retry: zio = zio_root(spa, NULL, NULL, flags); for (int l = 0; l < VDEV_LABELS; l++) { vdev_label_write(zio, vd, l, vp, offsetof(vdev_label_t, vl_vdev_phys), sizeof (vdev_phys_t), NULL, NULL, flags); /* * Skip the 1st padding area. * Zero out the 2nd padding area where it might have * left over data from previous filesystem format. */ vdev_label_write(zio, vd, l, pad2, offsetof(vdev_label_t, vl_pad2), VDEV_PAD_SIZE, NULL, NULL, flags); vdev_label_write(zio, vd, l, ub, offsetof(vdev_label_t, vl_uberblock), VDEV_UBERBLOCK_RING, NULL, NULL, flags); } error = zio_wait(zio); if (error != 0 && !(flags & ZIO_FLAG_TRYHARD)) { flags |= ZIO_FLAG_TRYHARD; goto retry; } 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)); /* * If this vdev hasn't been previously identified as a spare, then we * mark it as such only if a) we are labeling it as a spare, or b) it * exists as a spare elsewhere in the system. Do the same for * level 2 ARC devices. */ if (error == 0 && !vd->vdev_isspare && (reason == VDEV_LABEL_SPARE || spa_spare_exists(vd->vdev_guid, NULL, NULL))) spa_spare_add(vd); if (error == 0 && !vd->vdev_isl2cache && (reason == VDEV_LABEL_L2CACHE || spa_l2cache_exists(vd->vdev_guid, NULL))) spa_l2cache_add(vd); return (error); } /* * ========================================================================== * uberblock load/sync * ========================================================================== */ /* * Consider the following situation: txg is safely synced to disk. We've * written the first uberblock for txg + 1, and then we lose power. When we * come back up, we fail to see the uberblock for txg + 1 because, say, * it was on a mirrored device and the replica to which we wrote txg + 1 * is now offline. If we then make some changes and sync txg + 1, and then * the missing replica comes back, then for a few seconds we'll have two * conflicting uberblocks on disk with the same txg. The solution is simple: * among uberblocks with equal txg, choose the one with the latest timestamp. */ static int vdev_uberblock_compare(uberblock_t *ub1, 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); return (0); } struct ubl_cbdata { uberblock_t *ubl_ubbest; /* Best uberblock */ vdev_t *ubl_vd; /* vdev associated with the above */ }; static void vdev_uberblock_load_done(zio_t *zio) { vdev_t *vd = zio->io_vd; spa_t *spa = zio->io_spa; zio_t *rio = zio->io_private; uberblock_t *ub = zio->io_data; struct ubl_cbdata *cbp = rio->io_private; ASSERT3U(zio->io_size, ==, VDEV_UBERBLOCK_SIZE(vd)); if (zio->io_error == 0 && uberblock_verify(ub) == 0) { mutex_enter(&rio->io_lock); if (ub->ub_txg <= spa->spa_load_max_txg && vdev_uberblock_compare(ub, cbp->ubl_ubbest) > 0) { /* * Keep track of the vdev in which this uberblock * was found. We will use this information later * to obtain the config nvlist associated with * this uberblock. */ *cbp->ubl_ubbest = *ub; cbp->ubl_vd = vd; } mutex_exit(&rio->io_lock); } zio_buf_free(zio->io_data, zio->io_size); } static void vdev_uberblock_load_impl(zio_t *zio, vdev_t *vd, int flags, struct ubl_cbdata *cbp) { 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 (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), VDEV_UBERBLOCK_SIZE(vd), vdev_uberblock_load_done, zio, flags); } } } } /* * Reads the 'best' uberblock from disk along with its associated * configuration. First, we read the uberblock array of each label of each * vdev, keeping track of the uberblock with the highest txg in each array. * Then, we read the configuration from the same vdev as the best uberblock. */ void vdev_uberblock_load(vdev_t *rvd, uberblock_t *ub, nvlist_t **config) { zio_t *zio; spa_t *spa = rvd->vdev_spa; struct ubl_cbdata cb; int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE | ZIO_FLAG_TRYHARD; ASSERT(ub); ASSERT(config); bzero(ub, sizeof (uberblock_t)); *config = NULL; cb.ubl_ubbest = ub; cb.ubl_vd = NULL; spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); zio = zio_root(spa, NULL, &cb, flags); vdev_uberblock_load_impl(zio, rvd, flags, &cb); (void) zio_wait(zio); /* * It's possible that the best uberblock was discovered on a label * that has a configuration which was written in a future txg. * Search all labels on this vdev to find the configuration that * matches the txg for our uberblock. */ if (cb.ubl_vd != NULL) *config = vdev_label_read_config(cb.ubl_vd, ub->ub_txg); spa_config_exit(spa, SCL_ALL, FTAG); } /* * On success, increment root zio's count of good writes. * We only get credit for writes to known-visible vdevs; see spa_vdev_add(). */ static void vdev_uberblock_sync_done(zio_t *zio) { uint64_t *good_writes = zio->io_private; if (zio->io_error == 0 && zio->io_vd->vdev_top->vdev_ms_array != 0) atomic_inc_64(good_writes); } /* * Write the uberblock to all labels of all leaves of the specified vdev. */ static void vdev_uberblock_sync(zio_t *zio, uberblock_t *ub, vdev_t *vd, int flags) { uberblock_t *ubbuf; int n; 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) return; if (!vdev_writeable(vd)) return; n = ub->ub_txg & (VDEV_UBERBLOCK_COUNT(vd) - 1); ubbuf = zio_buf_alloc(VDEV_UBERBLOCK_SIZE(vd)); bzero(ubbuf, VDEV_UBERBLOCK_SIZE(vd)); *ubbuf = *ub; for (int l = 0; l < VDEV_LABELS; l++) vdev_label_write(zio, vd, l, ubbuf, 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)); } /* Sync the uberblocks to all vdevs in svd[] */ int vdev_uberblock_sync_list(vdev_t **svd, int svdcount, uberblock_t *ub, int flags) { spa_t *spa = svd[0]->vdev_spa; zio_t *zio; uint64_t good_writes = 0; zio = zio_root(spa, NULL, &good_writes, flags); for (int v = 0; v < svdcount; v++) vdev_uberblock_sync(zio, ub, svd[v], flags); (void) zio_wait(zio); /* * Flush the uberblocks to disk. This ensures that the odd labels * are no longer needed (because the new uberblocks and the even * labels are safely on disk), so it is safe to overwrite them. */ zio = zio_root(spa, NULL, NULL, flags); for (int v = 0; v < svdcount; v++) zio_flush(zio, svd[v]); (void) zio_wait(zio); return (good_writes >= 1 ? 0 : EIO); } /* * On success, increment the count of good writes for our top-level vdev. */ static void vdev_label_sync_done(zio_t *zio) { uint64_t *good_writes = zio->io_private; if (zio->io_error == 0) atomic_inc_64(good_writes); } /* * If there weren't enough good writes, indicate failure to the parent. */ static void vdev_label_sync_top_done(zio_t *zio) { uint64_t *good_writes = zio->io_private; if (*good_writes == 0) zio->io_error = SET_ERROR(EIO); kmem_free(good_writes, sizeof (uint64_t)); } /* * We ignore errors for log and cache devices, simply free the private data. */ static void vdev_label_sync_ignore_done(zio_t *zio) { kmem_free(zio->io_private, sizeof (uint64_t)); } /* * Write all even or odd labels to all leaves of the specified vdev. */ static void vdev_label_sync(zio_t *zio, vdev_t *vd, int l, uint64_t txg, int flags) { nvlist_t *label; vdev_phys_t *vp; char *buf; size_t buflen; 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) return; if (!vdev_writeable(vd)) return; /* * Generate a label describing the top-level config to which we belong. */ 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)); buf = vp->vp_nvlist; buflen = sizeof (vp->vp_nvlist); if (nvlist_pack(label, &buf, &buflen, NV_ENCODE_XDR, KM_SLEEP) == 0) { for (; l < VDEV_LABELS; l += 2) { vdev_label_write(zio, vd, l, vp, offsetof(vdev_label_t, vl_vdev_phys), sizeof (vdev_phys_t), vdev_label_sync_done, zio->io_private, flags | ZIO_FLAG_DONT_PROPAGATE); } } zio_buf_free(vp, sizeof (vdev_phys_t)); nvlist_free(label); } int vdev_label_sync_list(spa_t *spa, int l, uint64_t txg, int flags) { list_t *dl = &spa->spa_config_dirty_list; vdev_t *vd; zio_t *zio; int error; /* * Write the new labels to disk. */ zio = zio_root(spa, NULL, NULL, flags); for (vd = list_head(dl); vd != NULL; vd = list_next(dl, vd)) { uint64_t *good_writes = kmem_zalloc(sizeof (uint64_t), KM_SLEEP); ASSERT(!vd->vdev_ishole); 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); vdev_label_sync(vio, vd, l, txg, flags); zio_nowait(vio); } error = zio_wait(zio); /* * Flush the new labels to disk. */ zio = zio_root(spa, NULL, NULL, flags); for (vd = list_head(dl); vd != NULL; vd = list_next(dl, vd)) zio_flush(zio, vd); (void) zio_wait(zio); return (error); } /* * Sync the uberblock and any changes to the vdev configuration. * * The order of operations is carefully crafted to ensure that * if the system panics or loses power at any time, the state on disk * is still transactionally consistent. The in-line comments below * describe the failure semantics at each stage. * * Moreover, vdev_config_sync() is designed to be idempotent: if it fails * 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) { spa_t *spa = svd[0]->vdev_spa; uberblock_t *ub = &spa->spa_uberblock; vdev_t *vd; zio_t *zio; int error; int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL; /* * 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 * sync process to block while we retry. But if we can't write a * single label out, we should retry with ZIO_FLAG_TRYHARD before * bailing out and declaring the pool faulted. */ if (tryhard) flags |= ZIO_FLAG_TRYHARD; ASSERT(ub->ub_txg <= txg); /* * If this isn't a resync due to I/O errors, * and nothing changed in this transaction group, * 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 (txg > spa_freeze_txg(spa)) return (0); ASSERT(txg <= spa->spa_final_txg); /* * Flush the write cache of every disk that's been written to * in this transaction group. This ensures that all blocks * written in this txg will be committed to stable storage * before any uberblock that references them. */ zio = zio_root(spa, NULL, NULL, flags); for (vd = txg_list_head(&spa->spa_vdev_txg_list, TXG_CLEAN(txg)); vd; vd = txg_list_next(&spa->spa_vdev_txg_list, vd, TXG_CLEAN(txg))) zio_flush(zio, vd); (void) zio_wait(zio); /* * Sync out the even labels (L0, L2) for every dirty vdev. If the * system dies in the middle of this process, that's OK: all of the * even labels that made it to disk will be newer than any uberblock, * and will therefore be considered invalid. The odd labels (L1, L3), * which have not yet been touched, will still be valid. We flush * the new labels to disk to ensure that all even-label updates * are committed to stable storage before the uberblock update. */ if ((error = vdev_label_sync_list(spa, 0, txg, flags)) != 0) return (error); /* * Sync the uberblocks to all vdevs in svd[]. * If the system dies in the middle of this step, there are two cases * to consider, and the on-disk state is consistent either way: * * (1) If none of the new uberblocks made it to disk, then the * previous uberblock will be the newest, and the odd labels * (which had not yet been touched) will be valid with respect * to that uberblock. * * (2) If one or more new uberblocks made it to disk, then they * will be the newest, and the even labels (which had all * been successfully committed) will be valid with respect * to the new uberblocks. */ if ((error = vdev_uberblock_sync_list(svd, svdcount, ub, flags)) != 0) return (error); /* * Sync out odd labels for every dirty vdev. If the system dies * in the middle of this process, the even labels and the new * uberblocks will suffice to open the pool. The next time * the pool is opened, the first thing we'll do -- before any * user data is modified -- is mark every vdev dirty so that * all labels will be brought up to date. We flush the new labels * to disk to ensure that all odd-label updates are committed to * stable storage before the next transaction group begins. */ if ((error = vdev_label_sync_list(spa, 1, txg, flags)) != 0) return (error); trim_thread_wakeup(spa); return (0); } Index: stable/10 =================================================================== --- stable/10 (revision 274799) +++ stable/10 (revision 274800) Property changes on: stable/10 ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head:r274619