diff options
Diffstat (limited to 'sys/cddl/contrib/opensolaris/uts/common')
12 files changed, 334 insertions, 171 deletions
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c index 9430037..1700194 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c @@ -21,7 +21,7 @@ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, Joyent, Inc. All rights reserved. - * Copyright (c) 2011, 2015 by Delphix. All rights reserved. + * Copyright (c) 2011, 2016 by Delphix. All rights reserved. * Copyright (c) 2014 by Saso Kiselkov. All rights reserved. * Copyright 2015 Nexenta Systems, Inc. All rights reserved. */ @@ -782,6 +782,7 @@ typedef struct arc_write_callback arc_write_callback_t; struct arc_write_callback { void *awcb_private; arc_done_func_t *awcb_ready; + arc_done_func_t *awcb_children_ready; arc_done_func_t *awcb_physdone; arc_done_func_t *awcb_done; arc_buf_t *awcb_buf; @@ -5045,6 +5046,15 @@ arc_write_ready(zio_t *zio) hdr->b_flags |= ARC_FLAG_IO_IN_PROGRESS; } +static void +arc_write_children_ready(zio_t *zio) +{ + arc_write_callback_t *callback = zio->io_private; + arc_buf_t *buf = callback->awcb_buf; + + callback->awcb_children_ready(zio, buf, callback->awcb_private); +} + /* * The SPA calls this callback for each physical write that happens on behalf * of a logical write. See the comment in dbuf_write_physdone() for details. @@ -5141,7 +5151,8 @@ arc_write_done(zio_t *zio) zio_t * arc_write(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp, arc_buf_t *buf, boolean_t l2arc, boolean_t l2arc_compress, - const zio_prop_t *zp, arc_done_func_t *ready, arc_done_func_t *physdone, + const zio_prop_t *zp, arc_done_func_t *ready, + arc_done_func_t *children_ready, arc_done_func_t *physdone, arc_done_func_t *done, void *private, zio_priority_t priority, int zio_flags, const zbookmark_phys_t *zb) { @@ -5161,13 +5172,16 @@ arc_write(zio_t *pio, spa_t *spa, uint64_t txg, hdr->b_flags |= ARC_FLAG_L2COMPRESS; callback = kmem_zalloc(sizeof (arc_write_callback_t), KM_SLEEP); callback->awcb_ready = ready; + callback->awcb_children_ready = children_ready; callback->awcb_physdone = physdone; callback->awcb_done = done; callback->awcb_private = private; callback->awcb_buf = buf; zio = zio_write(pio, spa, txg, bp, buf->b_data, hdr->b_size, zp, - arc_write_ready, arc_write_physdone, arc_write_done, callback, + arc_write_ready, + (children_ready != NULL) ? arc_write_children_ready : NULL, + arc_write_physdone, arc_write_done, callback, priority, zio_flags, zb); return (zio); diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c index 2f26e9fb..9d5c398 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c @@ -481,13 +481,49 @@ dbuf_verify(dmu_buf_impl_t *db) * If the blkptr isn't set but they have nonzero data, * it had better be dirty, otherwise we'll lose that * data when we evict this buffer. + * + * There is an exception to this rule for indirect blocks; in + * this case, if the indirect block is a hole, we fill in a few + * fields on each of the child blocks (importantly, birth time) + * to prevent hole birth times from being lost when you + * partially fill in a hole. */ if (db->db_dirtycnt == 0) { - uint64_t *buf = db->db.db_data; - int i; + if (db->db_level == 0) { + uint64_t *buf = db->db.db_data; + int i; - for (i = 0; i < db->db.db_size >> 3; i++) { - ASSERT(buf[i] == 0); + for (i = 0; i < db->db.db_size >> 3; i++) { + ASSERT(buf[i] == 0); + } + } else { + blkptr_t *bps = db->db.db_data; + ASSERT3U(1 << DB_DNODE(db)->dn_indblkshift, ==, + db->db.db_size); + /* + * We want to verify that all the blkptrs in the + * indirect block are holes, but we may have + * automatically set up a few fields for them. + * We iterate through each blkptr and verify + * they only have those fields set. + */ + for (int i = 0; + i < db->db.db_size / sizeof (blkptr_t); + i++) { + blkptr_t *bp = &bps[i]; + ASSERT(ZIO_CHECKSUM_IS_ZERO( + &bp->blk_cksum)); + ASSERT( + DVA_IS_EMPTY(&bp->blk_dva[0]) && + DVA_IS_EMPTY(&bp->blk_dva[1]) && + DVA_IS_EMPTY(&bp->blk_dva[2])); + ASSERT0(bp->blk_fill); + ASSERT0(bp->blk_pad[0]); + ASSERT0(bp->blk_pad[1]); + ASSERT(!BP_IS_EMBEDDED(bp)); + ASSERT(BP_IS_HOLE(bp)); + ASSERT0(bp->blk_phys_birth); + } } } } @@ -655,10 +691,31 @@ dbuf_read_impl(dmu_buf_impl_t *db, zio_t *zio, uint32_t flags) BP_IS_HOLE(db->db_blkptr)))) { arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db); - DB_DNODE_EXIT(db); dbuf_set_data(db, arc_buf_alloc(db->db_objset->os_spa, db->db.db_size, db, type)); bzero(db->db.db_data, db->db.db_size); + + if (db->db_blkptr != NULL && db->db_level > 0 && + BP_IS_HOLE(db->db_blkptr) && + db->db_blkptr->blk_birth != 0) { + blkptr_t *bps = db->db.db_data; + for (int i = 0; i < ((1 << + DB_DNODE(db)->dn_indblkshift) / sizeof (blkptr_t)); + i++) { + blkptr_t *bp = &bps[i]; + ASSERT3U(BP_GET_LSIZE(db->db_blkptr), ==, + 1 << dn->dn_indblkshift); + BP_SET_LSIZE(bp, + BP_GET_LEVEL(db->db_blkptr) == 1 ? + dn->dn_datablksz : + BP_GET_LSIZE(db->db_blkptr)); + BP_SET_TYPE(bp, BP_GET_TYPE(db->db_blkptr)); + BP_SET_LEVEL(bp, + BP_GET_LEVEL(db->db_blkptr) - 1); + BP_SET_BIRTH(bp, db->db_blkptr->blk_birth, 0); + } + } + DB_DNODE_EXIT(db); db->db_state = DB_CACHED; mutex_exit(&db->db_mtx); return; @@ -2876,7 +2933,8 @@ dbuf_write_ready(zio_t *zio, arc_buf_t *buf, void *vdb) uint64_t fill = 0; int i; - ASSERT3P(db->db_blkptr, ==, bp); + ASSERT3P(db->db_blkptr, !=, NULL); + ASSERT3P(&db->db_data_pending->dr_bp_copy, ==, bp); DB_DNODE_ENTER(db); dn = DB_DNODE(db); @@ -2898,7 +2956,7 @@ dbuf_write_ready(zio_t *zio, arc_buf_t *buf, void *vdb) #ifdef ZFS_DEBUG if (db->db_blkid == DMU_SPILL_BLKID) { ASSERT(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR); - ASSERT(!(BP_IS_HOLE(db->db_blkptr)) && + ASSERT(!(BP_IS_HOLE(bp)) && db->db_blkptr == &dn->dn_phys->dn_spill); } #endif @@ -2939,6 +2997,49 @@ dbuf_write_ready(zio_t *zio, arc_buf_t *buf, void *vdb) bp->blk_fill = fill; mutex_exit(&db->db_mtx); + + rw_enter(&dn->dn_struct_rwlock, RW_WRITER); + *db->db_blkptr = *bp; + rw_exit(&dn->dn_struct_rwlock); +} + +/* ARGSUSED */ +/* + * This function gets called just prior to running through the compression + * stage of the zio pipeline. If we're an indirect block comprised of only + * holes, then we want this indirect to be compressed away to a hole. In + * order to do that we must zero out any information about the holes that + * this indirect points to prior to before we try to compress it. + */ +static void +dbuf_write_children_ready(zio_t *zio, arc_buf_t *buf, void *vdb) +{ + dmu_buf_impl_t *db = vdb; + dnode_t *dn; + blkptr_t *bp; + uint64_t i; + int epbs; + + ASSERT3U(db->db_level, >, 0); + DB_DNODE_ENTER(db); + dn = DB_DNODE(db); + epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT; + + /* Determine if all our children are holes */ + for (i = 0, bp = db->db.db_data; i < 1 << epbs; i++, bp++) { + if (!BP_IS_HOLE(bp)) + break; + } + + /* + * If all the children are holes, then zero them all out so that + * we may get compressed away. + */ + if (i == 1 << epbs) { + /* didn't find any non-holes */ + bzero(db->db.db_data, db->db.db_size); + } + DB_DNODE_EXIT(db); } /* @@ -3117,6 +3218,8 @@ dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx) zio_t *zio; int wp_flag = 0; + ASSERT(dmu_tx_is_syncing(tx)); + DB_DNODE_ENTER(db); dn = DB_DNODE(db); os = dn->dn_objset; @@ -3175,6 +3278,14 @@ dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx) dmu_write_policy(os, dn, db->db_level, wp_flag, &zp); DB_DNODE_EXIT(db); + /* + * We copy the blkptr now (rather than when we instantiate the dirty + * record), because its value can change between open context and + * syncing context. We do not need to hold dn_struct_rwlock to read + * db_blkptr because we are in syncing context. + */ + dr->dr_bp_copy = *db->db_blkptr; + if (db->db_level == 0 && dr->dt.dl.dr_override_state == DR_OVERRIDDEN) { /* @@ -3184,8 +3295,9 @@ dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx) void *contents = (data != NULL) ? data->b_data : NULL; dr->dr_zio = zio_write(zio, os->os_spa, txg, - db->db_blkptr, contents, db->db.db_size, &zp, - dbuf_write_override_ready, NULL, dbuf_write_override_done, + &dr->dr_bp_copy, contents, db->db.db_size, &zp, + dbuf_write_override_ready, NULL, NULL, + dbuf_write_override_done, dr, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb); mutex_enter(&db->db_mtx); dr->dt.dl.dr_override_state = DR_NOT_OVERRIDDEN; @@ -3196,15 +3308,27 @@ dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx) ASSERT(zp.zp_checksum == ZIO_CHECKSUM_OFF || zp.zp_checksum == ZIO_CHECKSUM_NOPARITY); dr->dr_zio = zio_write(zio, os->os_spa, txg, - db->db_blkptr, NULL, db->db.db_size, &zp, - dbuf_write_nofill_ready, NULL, dbuf_write_nofill_done, db, + &dr->dr_bp_copy, NULL, db->db.db_size, &zp, + dbuf_write_nofill_ready, NULL, NULL, + dbuf_write_nofill_done, db, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED | ZIO_FLAG_NODATA, &zb); } else { ASSERT(arc_released(data)); + + /* + * For indirect blocks, we want to setup the children + * ready callback so that we can properly handle an indirect + * block that only contains holes. + */ + arc_done_func_t *children_ready_cb = NULL; + if (db->db_level != 0) + children_ready_cb = dbuf_write_children_ready; + dr->dr_zio = arc_write(zio, os->os_spa, txg, - db->db_blkptr, data, DBUF_IS_L2CACHEABLE(db), + &dr->dr_bp_copy, data, DBUF_IS_L2CACHEABLE(db), DBUF_IS_L2COMPRESSIBLE(db), &zp, dbuf_write_ready, + children_ready_cb, dbuf_write_physdone, dbuf_write_done, db, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb); } diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c index d8d186c..9ce9665 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c @@ -20,7 +20,7 @@ */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. - * Copyright (c) 2011, 2015 by Delphix. All rights reserved. + * Copyright (c) 2011, 2016 by Delphix. All rights reserved. */ /* Copyright (c) 2013 by Saso Kiselkov. All rights reserved. */ /* Copyright (c) 2013, Joyent, Inc. All rights reserved. */ @@ -1590,10 +1590,11 @@ dmu_sync_late_arrival(zio_t *pio, objset_t *os, dmu_sync_cb_t *done, zgd_t *zgd, dsa->dsa_zgd = zgd; dsa->dsa_tx = tx; - zio_nowait(zio_write(pio, os->os_spa, dmu_tx_get_txg(tx), zgd->zgd_bp, - zgd->zgd_db->db_data, zgd->zgd_db->db_size, zp, - dmu_sync_late_arrival_ready, NULL, dmu_sync_late_arrival_done, dsa, - ZIO_PRIORITY_SYNC_WRITE, ZIO_FLAG_CANFAIL, zb)); + zio_nowait(zio_write(pio, os->os_spa, dmu_tx_get_txg(tx), + zgd->zgd_bp, zgd->zgd_db->db_data, zgd->zgd_db->db_size, + zp, dmu_sync_late_arrival_ready, NULL, + NULL, dmu_sync_late_arrival_done, dsa, ZIO_PRIORITY_SYNC_WRITE, + ZIO_FLAG_CANFAIL, zb)); return (0); } @@ -1746,8 +1747,8 @@ dmu_sync(zio_t *pio, uint64_t txg, dmu_sync_cb_t *done, zgd_t *zgd) zio_nowait(arc_write(pio, os->os_spa, txg, bp, dr->dt.dl.dr_data, DBUF_IS_L2CACHEABLE(db), DBUF_IS_L2COMPRESSIBLE(db), &zp, dmu_sync_ready, - NULL, dmu_sync_done, dsa, ZIO_PRIORITY_SYNC_WRITE, - ZIO_FLAG_CANFAIL, &zb)); + NULL, NULL, dmu_sync_done, dsa, + ZIO_PRIORITY_SYNC_WRITE, ZIO_FLAG_CANFAIL, &zb)); return (0); } diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_objset.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_objset.c index 367dbcb..82c5854 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_objset.c +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_objset.c @@ -20,7 +20,7 @@ */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. - * Copyright (c) 2012, 2014 by Delphix. All rights reserved. + * Copyright (c) 2012, 2016 by Delphix. All rights reserved. * Copyright (c) 2013 by Saso Kiselkov. All rights reserved. * Copyright (c) 2013, Joyent, Inc. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. @@ -1116,9 +1116,9 @@ dmu_objset_sync(objset_t *os, zio_t *pio, dmu_tx_t *tx) zio = arc_write(pio, os->os_spa, tx->tx_txg, os->os_rootbp, os->os_phys_buf, DMU_OS_IS_L2CACHEABLE(os), - DMU_OS_IS_L2COMPRESSIBLE(os), &zp, dmu_objset_write_ready, - NULL, dmu_objset_write_done, os, ZIO_PRIORITY_ASYNC_WRITE, - ZIO_FLAG_MUSTSUCCEED, &zb); + DMU_OS_IS_L2COMPRESSIBLE(os), + &zp, dmu_objset_write_ready, NULL, NULL, dmu_objset_write_done, + os, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb); /* * Sync special dnodes - the parent IO for the sync is the root block diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode_sync.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode_sync.c index 9aee513..7179c41 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode_sync.c +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode_sync.c @@ -21,7 +21,7 @@ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. - * Copyright (c) 2012, 2015 by Delphix. All rights reserved. + * Copyright (c) 2012, 2016 by Delphix. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. */ @@ -60,20 +60,14 @@ dnode_increase_indirection(dnode_t *dn, dmu_tx_t *tx) dprintf("os=%p obj=%llu, increase to %d\n", dn->dn_objset, dn->dn_object, dn->dn_phys->dn_nlevels); - /* check for existing blkptrs in the dnode */ - for (i = 0; i < nblkptr; i++) - if (!BP_IS_HOLE(&dn->dn_phys->dn_blkptr[i])) - break; - if (i != nblkptr) { - /* transfer dnode's block pointers to new indirect block */ - (void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED|DB_RF_HAVESTRUCT); - ASSERT(db->db.db_data); - ASSERT(arc_released(db->db_buf)); - ASSERT3U(sizeof (blkptr_t) * nblkptr, <=, db->db.db_size); - bcopy(dn->dn_phys->dn_blkptr, db->db.db_data, - sizeof (blkptr_t) * nblkptr); - arc_buf_freeze(db->db_buf); - } + /* transfer dnode's block pointers to new indirect block */ + (void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED|DB_RF_HAVESTRUCT); + ASSERT(db->db.db_data); + ASSERT(arc_released(db->db_buf)); + ASSERT3U(sizeof (blkptr_t) * nblkptr, <=, db->db.db_size); + bcopy(dn->dn_phys->dn_blkptr, db->db.db_data, + sizeof (blkptr_t) * nblkptr); + arc_buf_freeze(db->db_buf); /* set dbuf's parent pointers to new indirect buf */ for (i = 0; i < nblkptr; i++) { diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_scan.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_scan.c index a157dfe..9880de9 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_scan.c +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_scan.c @@ -56,7 +56,8 @@ typedef int (scan_cb_t)(dsl_pool_t *, const blkptr_t *, static scan_cb_t dsl_scan_scrub_cb; static void dsl_scan_cancel_sync(void *, dmu_tx_t *); -static void dsl_scan_sync_state(dsl_scan_t *, dmu_tx_t *tx); +static void dsl_scan_sync_state(dsl_scan_t *, dmu_tx_t *); +static boolean_t dsl_scan_restarting(dsl_scan_t *, dmu_tx_t *); unsigned int zfs_top_maxinflight = 32; /* maximum I/Os per top-level */ unsigned int zfs_resilver_delay = 2; /* number of ticks to delay resilver */ @@ -329,8 +330,15 @@ dsl_scan_done(dsl_scan_t *scn, boolean_t complete, dmu_tx_t *tx) else scn->scn_phys.scn_state = DSS_CANCELED; - spa_history_log_internal(spa, "scan done", tx, - "complete=%u", complete); + if (dsl_scan_restarting(scn, tx)) + spa_history_log_internal(spa, "scan aborted, restarting", tx, + "errors=%llu", spa_get_errlog_size(spa)); + else if (!complete) + spa_history_log_internal(spa, "scan cancelled", tx, + "errors=%llu", spa_get_errlog_size(spa)); + else + spa_history_log_internal(spa, "scan done", tx, + "errors=%llu", spa_get_errlog_size(spa)); if (DSL_SCAN_IS_SCRUB_RESILVER(scn)) { mutex_enter(&spa->spa_scrub_lock); @@ -1485,8 +1493,7 @@ dsl_scan_sync(dsl_pool_t *dp, dmu_tx_t *tx) * that we can restart an old-style scan while the pool is being * imported (see dsl_scan_init). */ - if (scn->scn_restart_txg != 0 && - scn->scn_restart_txg <= tx->tx_txg) { + if (dsl_scan_restarting(scn, tx)) { pool_scan_func_t func = POOL_SCAN_SCRUB; dsl_scan_done(scn, B_FALSE, tx); if (vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL)) @@ -1913,3 +1920,10 @@ dsl_scan(dsl_pool_t *dp, pool_scan_func_t func) return (dsl_sync_task(spa_name(spa), dsl_scan_setup_check, dsl_scan_setup_sync, &func, 0, ZFS_SPACE_CHECK_NONE)); } + +static boolean_t +dsl_scan_restarting(dsl_scan_t *scn, dmu_tx_t *tx) +{ + return (scn->scn_restart_txg != 0 && + scn->scn_restart_txg <= tx->tx_txg); +} diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/arc.h b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/arc.h index 04a80f7..714c528 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/arc.h +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/arc.h @@ -20,7 +20,7 @@ */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. - * Copyright (c) 2012, 2014 by Delphix. All rights reserved. + * Copyright (c) 2012, 2016 by Delphix. All rights reserved. * Copyright (c) 2013 by Saso Kiselkov. All rights reserved. */ @@ -141,9 +141,11 @@ int arc_read(zio_t *pio, spa_t *spa, const blkptr_t *bp, arc_flags_t *arc_flags, const zbookmark_phys_t *zb); zio_t *arc_write(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp, arc_buf_t *buf, boolean_t l2arc, boolean_t l2arc_compress, - const zio_prop_t *zp, arc_done_func_t *ready, arc_done_func_t *physdone, - arc_done_func_t *done, void *priv, zio_priority_t priority, - int zio_flags, const zbookmark_phys_t *zb); + const zio_prop_t *zp, + arc_done_func_t *ready, arc_done_func_t *child_ready, + arc_done_func_t *physdone, arc_done_func_t *done, + void *priv, zio_priority_t priority, int zio_flags, + const zbookmark_phys_t *zb); void arc_freed(spa_t *spa, const blkptr_t *bp); void arc_set_callback(arc_buf_t *buf, arc_evict_func_t *func, void *priv); diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dbuf.h b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dbuf.h index 233d541..4964126 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dbuf.h +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dbuf.h @@ -121,6 +121,9 @@ typedef struct dbuf_dirty_record { /* How much space was changed to dsl_pool_dirty_space() for this? */ unsigned int dr_accounted; + /* A copy of the bp that points to us */ + blkptr_t dr_bp_copy; + union dirty_types { struct dirty_indirect { diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zio.h b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zio.h index 56c821a..fc7be5b 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zio.h +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/zio.h @@ -22,7 +22,7 @@ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. - * Copyright (c) 2012, 2014 by Delphix. All rights reserved. + * Copyright (c) 2012, 2016 by Delphix. All rights reserved. * Copyright (c) 2013 by Saso Kiselkov. All rights reserved. */ @@ -436,6 +436,7 @@ struct zio { /* Callback info */ zio_done_func_t *io_ready; + zio_done_func_t *io_children_ready; zio_done_func_t *io_physdone; zio_done_func_t *io_done; void *io_private; @@ -503,9 +504,10 @@ extern zio_t *zio_read(zio_t *pio, spa_t *spa, const blkptr_t *bp, void *data, extern zio_t *zio_write(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp, void *data, uint64_t size, const zio_prop_t *zp, - zio_done_func_t *ready, zio_done_func_t *physdone, zio_done_func_t *done, - void *priv, - zio_priority_t priority, enum zio_flag flags, const zbookmark_phys_t *zb); + zio_done_func_t *ready, zio_done_func_t *children_ready, + zio_done_func_t *physdone, zio_done_func_t *done, + void *priv, zio_priority_t priority, enum zio_flag flags, + const zbookmark_phys_t *zb); extern zio_t *zio_rewrite(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp, void *data, uint64_t size, zio_done_func_t *done, void *priv, diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vfsops.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vfsops.c index 9bb6091..8523bc4 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vfsops.c +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vfsops.c @@ -22,7 +22,7 @@ * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011 Pawel Jakub Dawidek <pawel@dawidek.net>. * All rights reserved. - * Copyright (c) 2012, 2014 by Delphix. All rights reserved. + * Copyright (c) 2012, 2015 by Delphix. All rights reserved. * Copyright (c) 2014 Integros [integros.com] */ @@ -847,72 +847,46 @@ zfs_owner_overquota(zfsvfs_t *zfsvfs, znode_t *zp, boolean_t isgroup) return (zfs_fuid_overquota(zfsvfs, isgroup, fuid)); } -int -zfsvfs_create(const char *osname, zfsvfs_t **zfvp) +/* + * Associate this zfsvfs with the given objset, which must be owned. + * This will cache a bunch of on-disk state from the objset in the + * zfsvfs. + */ +static int +zfsvfs_init(zfsvfs_t *zfsvfs, objset_t *os) { - objset_t *os; - zfsvfs_t *zfsvfs; - uint64_t zval; - int i, error; - uint64_t sa_obj; - - /* - * XXX: Fix struct statfs so this isn't necessary! - * - * The 'osname' is used as the filesystem's special node, which means - * it must fit in statfs.f_mntfromname, or else it can't be - * enumerated, so libzfs_mnttab_find() returns NULL, which causes - * 'zfs unmount' to think it's not mounted when it is. - */ - if (strlen(osname) >= MNAMELEN) - return (SET_ERROR(ENAMETOOLONG)); - - zfsvfs = kmem_zalloc(sizeof (zfsvfs_t), KM_SLEEP); - - /* - * We claim to always be readonly so we can open snapshots; - * other ZPL code will prevent us from writing to snapshots. - */ - error = dmu_objset_own(osname, DMU_OST_ZFS, B_TRUE, zfsvfs, &os); - if (error) { - kmem_free(zfsvfs, sizeof (zfsvfs_t)); - return (error); - } + int error; + uint64_t val; - /* - * Initialize the zfs-specific filesystem structure. - * Should probably make this a kmem cache, shuffle fields, - * and just bzero up to z_hold_mtx[]. - */ - zfsvfs->z_vfs = NULL; - zfsvfs->z_parent = zfsvfs; zfsvfs->z_max_blksz = SPA_OLD_MAXBLOCKSIZE; zfsvfs->z_show_ctldir = ZFS_SNAPDIR_VISIBLE; zfsvfs->z_os = os; error = zfs_get_zplprop(os, ZFS_PROP_VERSION, &zfsvfs->z_version); - if (error) { - goto out; - } else if (zfsvfs->z_version > + if (error != 0) + return (error); + if (zfsvfs->z_version > zfs_zpl_version_map(spa_version(dmu_objset_spa(os)))) { (void) printf("Can't mount a version %lld file system " "on a version %lld pool\n. Pool must be upgraded to mount " "this file system.", (u_longlong_t)zfsvfs->z_version, (u_longlong_t)spa_version(dmu_objset_spa(os))); - error = SET_ERROR(ENOTSUP); - goto out; + return (SET_ERROR(ENOTSUP)); } - if ((error = zfs_get_zplprop(os, ZFS_PROP_NORMALIZE, &zval)) != 0) - goto out; - zfsvfs->z_norm = (int)zval; + error = zfs_get_zplprop(os, ZFS_PROP_NORMALIZE, &val); + if (error != 0) + return (error); + zfsvfs->z_norm = (int)val; - if ((error = zfs_get_zplprop(os, ZFS_PROP_UTF8ONLY, &zval)) != 0) - goto out; - zfsvfs->z_utf8 = (zval != 0); + error = zfs_get_zplprop(os, ZFS_PROP_UTF8ONLY, &val); + if (error != 0) + return (error); + zfsvfs->z_utf8 = (val != 0); - if ((error = zfs_get_zplprop(os, ZFS_PROP_CASE, &zval)) != 0) - goto out; - zfsvfs->z_case = (uint_t)zval; + error = zfs_get_zplprop(os, ZFS_PROP_CASE, &val); + if (error != 0) + return (error); + zfsvfs->z_case = (uint_t)val; /* * Fold case on file systems that are always or sometimes case @@ -925,60 +899,99 @@ zfsvfs_create(const char *osname, zfsvfs_t **zfvp) zfsvfs->z_use_fuids = USE_FUIDS(zfsvfs->z_version, zfsvfs->z_os); zfsvfs->z_use_sa = USE_SA(zfsvfs->z_version, zfsvfs->z_os); + uint64_t sa_obj = 0; if (zfsvfs->z_use_sa) { /* should either have both of these objects or none */ error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_SA_ATTRS, 8, 1, &sa_obj); - if (error) - goto out; - } else { - /* - * Pre SA versions file systems should never touch - * either the attribute registration or layout objects. - */ - sa_obj = 0; + if (error != 0) + return (error); } error = sa_setup(os, sa_obj, zfs_attr_table, ZPL_END, &zfsvfs->z_attr_table); - if (error) - goto out; + if (error != 0) + return (error); if (zfsvfs->z_version >= ZPL_VERSION_SA) sa_register_update_callback(os, zfs_sa_upgrade); error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_ROOT_OBJ, 8, 1, &zfsvfs->z_root); - if (error) - goto out; + if (error != 0) + return (error); ASSERT(zfsvfs->z_root != 0); error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_UNLINKED_SET, 8, 1, &zfsvfs->z_unlinkedobj); - if (error) - goto out; + if (error != 0) + return (error); error = zap_lookup(os, MASTER_NODE_OBJ, zfs_userquota_prop_prefixes[ZFS_PROP_USERQUOTA], 8, 1, &zfsvfs->z_userquota_obj); - if (error && error != ENOENT) - goto out; + if (error == ENOENT) + zfsvfs->z_userquota_obj = 0; + else if (error != 0) + return (error); error = zap_lookup(os, MASTER_NODE_OBJ, zfs_userquota_prop_prefixes[ZFS_PROP_GROUPQUOTA], 8, 1, &zfsvfs->z_groupquota_obj); - if (error && error != ENOENT) - goto out; + if (error == ENOENT) + zfsvfs->z_groupquota_obj = 0; + else if (error != 0) + return (error); error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_FUID_TABLES, 8, 1, &zfsvfs->z_fuid_obj); - if (error && error != ENOENT) - goto out; + if (error == ENOENT) + zfsvfs->z_fuid_obj = 0; + else if (error != 0) + return (error); error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_SHARES_DIR, 8, 1, &zfsvfs->z_shares_dir); - if (error && error != ENOENT) - goto out; + if (error == ENOENT) + zfsvfs->z_shares_dir = 0; + else if (error != 0) + return (error); + + return (0); +} + +int +zfsvfs_create(const char *osname, zfsvfs_t **zfvp) +{ + objset_t *os; + zfsvfs_t *zfsvfs; + int error; + + /* + * XXX: Fix struct statfs so this isn't necessary! + * + * The 'osname' is used as the filesystem's special node, which means + * it must fit in statfs.f_mntfromname, or else it can't be + * enumerated, so libzfs_mnttab_find() returns NULL, which causes + * 'zfs unmount' to think it's not mounted when it is. + */ + if (strlen(osname) >= MNAMELEN) + return (SET_ERROR(ENAMETOOLONG)); + + zfsvfs = kmem_zalloc(sizeof (zfsvfs_t), KM_SLEEP); + + /* + * We claim to always be readonly so we can open snapshots; + * other ZPL code will prevent us from writing to snapshots. + */ + error = dmu_objset_own(osname, DMU_OST_ZFS, B_TRUE, zfsvfs, &os); + if (error) { + kmem_free(zfsvfs, sizeof (zfsvfs_t)); + return (error); + } + + zfsvfs->z_vfs = NULL; + zfsvfs->z_parent = zfsvfs; mutex_init(&zfsvfs->z_znodes_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&zfsvfs->z_lock, NULL, MUTEX_DEFAULT, NULL); @@ -987,17 +1000,19 @@ zfsvfs_create(const char *osname, zfsvfs_t **zfvp) rrm_init(&zfsvfs->z_teardown_lock, B_FALSE); rw_init(&zfsvfs->z_teardown_inactive_lock, NULL, RW_DEFAULT, NULL); rw_init(&zfsvfs->z_fuid_lock, NULL, RW_DEFAULT, NULL); - for (i = 0; i != ZFS_OBJ_MTX_SZ; i++) + for (int i = 0; i != ZFS_OBJ_MTX_SZ; i++) mutex_init(&zfsvfs->z_hold_mtx[i], NULL, MUTEX_DEFAULT, NULL); + error = zfsvfs_init(zfsvfs, os); + if (error != 0) { + dmu_objset_disown(os, zfsvfs); + *zfvp = NULL; + kmem_free(zfsvfs, sizeof (zfsvfs_t)); + return (error); + } + *zfvp = zfsvfs; return (0); - -out: - dmu_objset_disown(os, zfsvfs); - *zfvp = NULL; - kmem_free(zfsvfs, sizeof (zfsvfs_t)); - return (error); } static int @@ -2199,7 +2214,6 @@ zfs_resume_fs(zfsvfs_t *zfsvfs, const char *osname) { int err; znode_t *zp; - uint64_t sa_obj = 0; ASSERT(RRM_WRITE_HELD(&zfsvfs->z_teardown_lock)); ASSERT(RW_WRITE_HELD(&zfsvfs->z_teardown_inactive_lock)); @@ -2208,35 +2222,16 @@ zfs_resume_fs(zfsvfs_t *zfsvfs, const char *osname) * We already own this, so just hold and rele it to update the * objset_t, as the one we had before may have been evicted. */ - VERIFY0(dmu_objset_hold(osname, zfsvfs, &zfsvfs->z_os)); - VERIFY3P(zfsvfs->z_os->os_dsl_dataset->ds_owner, ==, zfsvfs); - VERIFY(dsl_dataset_long_held(zfsvfs->z_os->os_dsl_dataset)); - dmu_objset_rele(zfsvfs->z_os, zfsvfs); - - /* - * Make sure version hasn't changed - */ - - err = zfs_get_zplprop(zfsvfs->z_os, ZFS_PROP_VERSION, - &zfsvfs->z_version); - - if (err) - goto bail; - - err = zap_lookup(zfsvfs->z_os, MASTER_NODE_OBJ, - ZFS_SA_ATTRS, 8, 1, &sa_obj); - - if (err && zfsvfs->z_version >= ZPL_VERSION_SA) - goto bail; + objset_t *os; + VERIFY0(dmu_objset_hold(osname, zfsvfs, &os)); + VERIFY3P(os->os_dsl_dataset->ds_owner, ==, zfsvfs); + VERIFY(dsl_dataset_long_held(os->os_dsl_dataset)); + dmu_objset_rele(os, zfsvfs); - if ((err = sa_setup(zfsvfs->z_os, sa_obj, - zfs_attr_table, ZPL_END, &zfsvfs->z_attr_table)) != 0) + err = zfsvfs_init(zfsvfs, os); + if (err != 0) goto bail; - if (zfsvfs->z_version >= ZPL_VERSION_SA) - sa_register_update_callback(zfsvfs->z_os, - zfs_sa_upgrade); - VERIFY(zfsvfs_setup(zfsvfs, B_FALSE) == 0); zfs_set_fuid_feature(zfsvfs); diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c index b0f11ac..17179f6 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c @@ -2438,6 +2438,7 @@ top: dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL); zfs_sa_upgrade_txholds(tx, zp); zfs_sa_upgrade_txholds(tx, dzp); + dmu_tx_mark_netfree(tx); error = dmu_tx_assign(tx, waited ? TXG_WAITED : TXG_NOWAIT); if (error) { rw_exit(&zp->z_parent_lock); diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zio.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zio.c index 616aa79..98fd449 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zio.c +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zio.c @@ -20,7 +20,7 @@ */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. - * Copyright (c) 2011, 2015 by Delphix. All rights reserved. + * Copyright (c) 2011, 2016 by Delphix. All rights reserved. * Copyright (c) 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2014 Integros [integros.com] */ @@ -771,9 +771,10 @@ zio_read(zio_t *pio, spa_t *spa, const blkptr_t *bp, zio_t * zio_write(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp, void *data, uint64_t size, const zio_prop_t *zp, - zio_done_func_t *ready, zio_done_func_t *physdone, zio_done_func_t *done, - void *private, - zio_priority_t priority, enum zio_flag flags, const zbookmark_phys_t *zb) + zio_done_func_t *ready, zio_done_func_t *children_ready, + zio_done_func_t *physdone, zio_done_func_t *done, + void *private, zio_priority_t priority, enum zio_flag flags, + const zbookmark_phys_t *zb) { zio_t *zio; @@ -792,6 +793,7 @@ zio_write(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp, ZIO_DDT_CHILD_WRITE_PIPELINE : ZIO_WRITE_PIPELINE); zio->io_ready = ready; + zio->io_children_ready = children_ready; zio->io_physdone = physdone; zio->io_prop = *zp; @@ -1189,6 +1191,16 @@ zio_write_bp_init(zio_t *zio) if (!IO_IS_ALLOCATING(zio)) return (ZIO_PIPELINE_CONTINUE); + if (zio->io_children_ready != NULL) { + /* + * Now that all our children are ready, run the callback + * associated with this zio in case it wants to modify the + * data to be written. + */ + ASSERT3U(zp->zp_level, >, 0); + zio->io_children_ready(zio); + } + ASSERT(zio->io_child_type != ZIO_CHILD_DDT); if (zio->io_bp_override) { @@ -2118,9 +2130,9 @@ zio_write_gang_block(zio_t *pio) zio_nowait(zio_write(zio, spa, txg, &gbh->zg_blkptr[g], (char *)pio->io_data + (pio->io_size - resid), lsize, &zp, - zio_write_gang_member_ready, NULL, NULL, &gn->gn_child[g], - pio->io_priority, ZIO_GANG_CHILD_FLAGS(pio), - &pio->io_bookmark)); + zio_write_gang_member_ready, NULL, NULL, NULL, + &gn->gn_child[g], pio->io_priority, + ZIO_GANG_CHILD_FLAGS(pio), &pio->io_bookmark)); } /* @@ -2509,7 +2521,7 @@ zio_ddt_write(zio_t *zio) dio = zio_write(zio, spa, txg, bp, zio->io_orig_data, zio->io_orig_size, &czp, NULL, NULL, - zio_ddt_ditto_write_done, dde, zio->io_priority, + NULL, zio_ddt_ditto_write_done, dde, zio->io_priority, ZIO_DDT_CHILD_FLAGS(zio), &zio->io_bookmark); zio_push_transform(dio, zio->io_data, zio->io_size, 0, NULL); @@ -2530,7 +2542,8 @@ zio_ddt_write(zio_t *zio) ddt_phys_addref(ddp); } else { cio = zio_write(zio, spa, txg, bp, zio->io_orig_data, - zio->io_orig_size, zp, zio_ddt_child_write_ready, NULL, + zio->io_orig_size, zp, + zio_ddt_child_write_ready, NULL, NULL, zio_ddt_child_write_done, dde, zio->io_priority, ZIO_DDT_CHILD_FLAGS(zio), &zio->io_bookmark); |
