From owner-svn-src-all@FreeBSD.ORG Thu Jan 16 14:08:14 2014 Return-Path: Delivered-To: svn-src-all@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 5F6F6CB5; Thu, 16 Jan 2014 14:08:14 +0000 (UTC) Received: from svn.freebsd.org (svn.freebsd.org [IPv6:2001:1900:2254:2068::e6a:0]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mx1.freebsd.org (Postfix) with ESMTPS id 4981D1FE9; Thu, 16 Jan 2014 14:08:14 +0000 (UTC) Received: from svn.freebsd.org ([127.0.1.70]) by svn.freebsd.org (8.14.7/8.14.7) with ESMTP id s0GE8Ece024960; Thu, 16 Jan 2014 14:08:14 GMT (envelope-from avg@svn.freebsd.org) Received: (from avg@localhost) by svn.freebsd.org (8.14.7/8.14.7/Submit) id s0GE8C9m024949; Thu, 16 Jan 2014 14:08:12 GMT (envelope-from avg@svn.freebsd.org) Message-Id: <201401161408.s0GE8C9m024949@svn.freebsd.org> From: Andriy Gapon Date: Thu, 16 Jan 2014 14:08:12 +0000 (UTC) To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-9@freebsd.org Subject: svn commit: r260722 - in stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs: . sys X-SVN-Group: stable-9 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.17 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 16 Jan 2014 14:08:14 -0000 Author: avg Date: Thu Jan 16 14:08:11 2014 New Revision: 260722 URL: http://svnweb.freebsd.org/changeset/base/260722 Log: MFC r253821,254753,256259 MFV r253783: 3834 incremental replication of 'holey' file systems is slow MFV r254747:4047 panic from dbuf_free_range() from dmu_free_object() while doing zfs receive MFV r255257: 4082 zfs receive gets EFBIG from dmu_tx_hold_free() Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_send.c stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_tx.c stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode.c stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_destroy.c stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu.h stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_impl.h stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_send.h stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dnode.h Directory Properties: stable/9/sys/ (props changed) stable/9/sys/cddl/contrib/opensolaris/ (props changed) Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c ============================================================================== --- stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c Thu Jan 16 14:05:05 2014 (r260721) +++ stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c Thu Jan 16 14:08:11 2014 (r260722) @@ -27,6 +27,7 @@ #include #include +#include #include #include #include @@ -39,6 +40,12 @@ #include #include +/* + * Number of times that zfs_free_range() took the slow path while doing + * a zfs receive. A nonzero value indicates a potential performance problem. + */ +uint64_t zfs_free_range_recv_miss; + static void dbuf_destroy(dmu_buf_impl_t *db); static boolean_t dbuf_undirty(dmu_buf_impl_t *db, dmu_tx_t *tx); static void dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx); @@ -796,9 +803,12 @@ dbuf_unoverride(dbuf_dirty_record_t *dr) /* * Evict (if its unreferenced) or clear (if its referenced) any level-0 * data blocks in the free range, so that any future readers will find - * empty blocks. Also, if we happen accross any level-1 dbufs in the + * empty blocks. Also, if we happen across any level-1 dbufs in the * range that have not already been marked dirty, mark them dirty so * they stay in memory. + * + * This is a no-op if the dataset is in the middle of an incremental + * receive; see comment below for details. */ void dbuf_free_range(dnode_t *dn, uint64_t start, uint64_t end, dmu_tx_t *tx) @@ -814,7 +824,23 @@ dbuf_free_range(dnode_t *dn, uint64_t st last_l1 = end >> epbs; } dprintf_dnode(dn, "start=%llu end=%llu\n", start, end); + mutex_enter(&dn->dn_dbufs_mtx); + if (start >= dn->dn_unlisted_l0_blkid * dn->dn_datablksz) { + /* There can't be any dbufs in this range; no need to search. */ + mutex_exit(&dn->dn_dbufs_mtx); + return; + } else if (dmu_objset_is_receiving(dn->dn_objset)) { + /* + * If we are receiving, we expect there to be no dbufs in + * the range to be freed, because receive modifies each + * block at most once, and in offset order. If this is + * not the case, it can lead to performance problems, + * so note that we unexpectedly took the slow path. + */ + atomic_inc_64(&zfs_free_range_recv_miss); + } + for (db = list_head(&dn->dn_dbufs); db; db = db_next) { db_next = list_next(&dn->dn_dbufs, db); ASSERT(db->db_blkid != DMU_BONUS_BLKID); @@ -1702,6 +1728,9 @@ dbuf_create(dnode_t *dn, uint8_t level, return (odb); } list_insert_head(&dn->dn_dbufs, db); + if (db->db_level == 0 && db->db_blkid >= + dn->dn_unlisted_l0_blkid) + dn->dn_unlisted_l0_blkid = db->db_blkid + 1; db->db_state = DB_UNCACHED; mutex_exit(&dn->dn_dbufs_mtx); arc_space_consume(sizeof (dmu_buf_impl_t), ARC_SPACE_OTHER); Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c ============================================================================== --- stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c Thu Jan 16 14:05:05 2014 (r260721) +++ stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c Thu Jan 16 14:08:11 2014 (r260722) @@ -575,98 +575,93 @@ dmu_prefetch(objset_t *os, uint64_t obje * the end so that the file gets shorter over time (if we crashes in the * middle, this will leave us in a better state). We find allocated file * data by simply searching the allocated level 1 indirects. + * + * On input, *start should be the first offset that does not need to be + * freed (e.g. "offset + length"). On return, *start will be the first + * offset that should be freed. */ static int -get_next_chunk(dnode_t *dn, uint64_t *start, uint64_t limit) +get_next_chunk(dnode_t *dn, uint64_t *start, uint64_t minimum) { - uint64_t len = *start - limit; - uint64_t blkcnt = 0; - uint64_t maxblks = DMU_MAX_ACCESS / (1ULL << (dn->dn_indblkshift + 1)); + uint64_t maxblks = DMU_MAX_ACCESS >> (dn->dn_indblkshift + 1); + /* bytes of data covered by a level-1 indirect block */ uint64_t iblkrange = dn->dn_datablksz * EPB(dn->dn_indblkshift, SPA_BLKPTRSHIFT); - ASSERT(limit <= *start); + ASSERT3U(minimum, <=, *start); - if (len <= iblkrange * maxblks) { - *start = limit; + if (*start - minimum <= iblkrange * maxblks) { + *start = minimum; return (0); } ASSERT(ISP2(iblkrange)); - while (*start > limit && blkcnt < maxblks) { + for (uint64_t blks = 0; *start > minimum && blks < maxblks; blks++) { int err; - /* find next allocated L1 indirect */ + /* + * dnode_next_offset(BACKWARDS) will find an allocated L1 + * indirect block at or before the input offset. We must + * decrement *start so that it is at the end of the region + * to search. + */ + (*start)--; err = dnode_next_offset(dn, DNODE_FIND_BACKWARDS, start, 2, 1, 0); - /* if there are no more, then we are done */ + /* if there are no indirect blocks before start, we are done */ if (err == ESRCH) { - *start = limit; - return (0); - } else if (err) { + *start = minimum; + break; + } else if (err != 0) { return (err); } - blkcnt += 1; - /* reset offset to end of "next" block back */ + /* set start to the beginning of this L1 indirect */ *start = P2ALIGN(*start, iblkrange); - if (*start <= limit) - *start = limit; - else - *start -= 1; } + if (*start < minimum) + *start = minimum; return (0); } static int dmu_free_long_range_impl(objset_t *os, dnode_t *dn, uint64_t offset, - uint64_t length, boolean_t free_dnode) + uint64_t length) { - dmu_tx_t *tx; - uint64_t object_size, start, end, len; - boolean_t trunc = (length == DMU_OBJECT_END); - int align, err; - - align = 1 << dn->dn_datablkshift; - ASSERT(align > 0); - object_size = align == 1 ? dn->dn_datablksz : - (dn->dn_maxblkid + 1) << dn->dn_datablkshift; - - end = offset + length; - if (trunc || end > object_size) - end = object_size; - if (end <= offset) + uint64_t object_size = (dn->dn_maxblkid + 1) * dn->dn_datablksz; + int err; + + if (offset >= object_size) return (0); - length = end - offset; - while (length) { - start = end; - /* assert(offset <= start) */ - err = get_next_chunk(dn, &start, offset); + if (length == DMU_OBJECT_END || offset + length > object_size) + length = object_size - offset; + + while (length != 0) { + uint64_t chunk_end, chunk_begin; + + chunk_end = chunk_begin = offset + length; + + /* move chunk_begin backwards to the beginning of this chunk */ + err = get_next_chunk(dn, &chunk_begin, offset); if (err) return (err); - len = trunc ? DMU_OBJECT_END : end - start; + ASSERT3U(chunk_begin, >=, offset); + ASSERT3U(chunk_begin, <=, chunk_end); - tx = dmu_tx_create(os); - dmu_tx_hold_free(tx, dn->dn_object, start, len); + dmu_tx_t *tx = dmu_tx_create(os); + dmu_tx_hold_free(tx, dn->dn_object, + chunk_begin, chunk_end - chunk_begin); err = dmu_tx_assign(tx, TXG_WAIT); if (err) { dmu_tx_abort(tx); return (err); } - - dnode_free_range(dn, start, trunc ? -1 : len, tx); - - if (start == 0 && free_dnode) { - ASSERT(trunc); - dnode_free(dn, tx); - } - - length -= end - start; - + dnode_free_range(dn, chunk_begin, chunk_end - chunk_begin, tx); dmu_tx_commit(tx); - end = start; + + length -= chunk_end - chunk_begin; } return (0); } @@ -681,38 +676,42 @@ dmu_free_long_range(objset_t *os, uint64 err = dnode_hold(os, object, FTAG, &dn); if (err != 0) return (err); - err = dmu_free_long_range_impl(os, dn, offset, length, FALSE); + err = dmu_free_long_range_impl(os, dn, offset, length); + + /* + * It is important to zero out the maxblkid when freeing the entire + * file, so that (a) subsequent calls to dmu_free_long_range_impl() + * will take the fast path, and (b) dnode_reallocate() can verify + * that the entire file has been freed. + */ + if (offset == 0 && length == DMU_OBJECT_END) + dn->dn_maxblkid = 0; + dnode_rele(dn, FTAG); return (err); } int -dmu_free_object(objset_t *os, uint64_t object) +dmu_free_long_object(objset_t *os, uint64_t object) { - dnode_t *dn; dmu_tx_t *tx; int err; - err = dnode_hold_impl(os, object, DNODE_MUST_BE_ALLOCATED, - FTAG, &dn); + err = dmu_free_long_range(os, object, 0, DMU_OBJECT_END); if (err != 0) return (err); - if (dn->dn_nlevels == 1) { - tx = dmu_tx_create(os); - dmu_tx_hold_bonus(tx, object); - dmu_tx_hold_free(tx, dn->dn_object, 0, DMU_OBJECT_END); - err = dmu_tx_assign(tx, TXG_WAIT); - if (err == 0) { - dnode_free_range(dn, 0, DMU_OBJECT_END, tx); - dnode_free(dn, tx); - dmu_tx_commit(tx); - } else { - dmu_tx_abort(tx); - } + + tx = dmu_tx_create(os); + dmu_tx_hold_bonus(tx, object); + dmu_tx_hold_free(tx, object, 0, DMU_OBJECT_END); + err = dmu_tx_assign(tx, TXG_WAIT); + if (err == 0) { + err = dmu_object_free(os, object, tx); + dmu_tx_commit(tx); } else { - err = dmu_free_long_range_impl(os, dn, 0, DMU_OBJECT_END, TRUE); + dmu_tx_abort(tx); } - dnode_rele(dn, FTAG); + return (err); } Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_send.c ============================================================================== --- stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_send.c Thu Jan 16 14:05:05 2014 (r260721) +++ stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_send.c Thu Jan 16 14:08:11 2014 (r260722) @@ -96,6 +96,32 @@ dump_free(dmu_sendarg_t *dsp, uint64_t o { struct drr_free *drrf = &(dsp->dsa_drr->drr_u.drr_free); + /* + * When we receive a free record, dbuf_free_range() assumes + * that the receiving system doesn't have any dbufs in the range + * being freed. This is always true because there is a one-record + * constraint: we only send one WRITE record for any given + * object+offset. We know that the one-record constraint is + * true because we always send data in increasing order by + * object,offset. + * + * If the increasing-order constraint ever changes, we should find + * another way to assert that the one-record constraint is still + * satisfied. + */ + ASSERT(object > dsp->dsa_last_data_object || + (object == dsp->dsa_last_data_object && + offset > dsp->dsa_last_data_offset)); + + /* + * If we are doing a non-incremental send, then there can't + * be any data in the dataset we're receiving into. Therefore + * a free record would simply be a no-op. Save space by not + * sending it to begin with. + */ + if (!dsp->dsa_incremental) + return (0); + if (length != -1ULL && offset + length < offset) length = -1ULL; @@ -162,6 +188,15 @@ dump_data(dmu_sendarg_t *dsp, dmu_object { struct drr_write *drrw = &(dsp->dsa_drr->drr_u.drr_write); + /* + * We send data in increasing object, offset order. + * See comment in dump_free() for details. + */ + ASSERT(object > dsp->dsa_last_data_object || + (object == dsp->dsa_last_data_object && + offset > dsp->dsa_last_data_offset)); + dsp->dsa_last_data_object = object; + dsp->dsa_last_data_offset = offset + blksz - 1; /* * If there is any kind of pending aggregation (currently either @@ -229,6 +264,10 @@ dump_freeobjects(dmu_sendarg_t *dsp, uin { struct drr_freeobjects *drrfo = &(dsp->dsa_drr->drr_u.drr_freeobjects); + /* See comment in dump_free(). */ + if (!dsp->dsa_incremental) + return (0); + /* * If there is a pending op, but it's not PENDING_FREEOBJECTS, * push it out, since free block aggregation can only be done for @@ -305,9 +344,9 @@ dump_dnode(dmu_sendarg_t *dsp, uint64_t if (dump_bytes(dsp, DN_BONUS(dnp), P2ROUNDUP(dnp->dn_bonuslen, 8)) != 0) return (SET_ERROR(EINTR)); - /* free anything past the end of the file */ + /* Free anything past the end of the file. */ if (dump_free(dsp, object, (dnp->dn_maxblkid + 1) * - (dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT), -1ULL)) + (dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT), -1ULL) != 0) return (SET_ERROR(EINTR)); if (dsp->dsa_err != 0) return (SET_ERROR(EINTR)); @@ -495,6 +534,7 @@ dmu_send_impl(void *tag, dsl_pool_t *dp, dsp->dsa_toguid = ds->ds_phys->ds_guid; ZIO_SET_CHECKSUM(&dsp->dsa_zc, 0, 0, 0, 0); dsp->dsa_pending_op = PENDING_NONE; + dsp->dsa_incremental = (fromtxg != 0); mutex_enter(&ds->ds_sendstream_lock); list_insert_head(&ds->ds_sendstreams, dsp); @@ -1238,7 +1278,7 @@ restore_freeobjects(struct restorearg *r if (dmu_object_info(os, obj, NULL) != 0) continue; - err = dmu_free_object(os, obj); + err = dmu_free_long_object(os, obj); if (err != 0) return (err); } @@ -1764,3 +1804,13 @@ dmu_recv_end(dmu_recv_cookie_t *drc, voi else return (dmu_recv_existing_end(drc)); } + +/* + * Return TRUE if this objset is currently being received into. + */ +boolean_t +dmu_objset_is_receiving(objset_t *os) +{ + return (os->os_dsl_dataset != NULL && + os->os_dsl_dataset->ds_owner == dmu_recv_tag); +} Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_tx.c ============================================================================== --- stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_tx.c Thu Jan 16 14:05:05 2014 (r260721) +++ stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_tx.c Thu Jan 16 14:08:11 2014 (r260722) @@ -587,8 +587,7 @@ dmu_tx_hold_free(dmu_tx_t *tx, uint64_t { dmu_tx_hold_t *txh; dnode_t *dn; - uint64_t start, end, i; - int err, shift; + int err; zio_t *zio; ASSERT(tx->tx_txg == 0); @@ -599,34 +598,49 @@ dmu_tx_hold_free(dmu_tx_t *tx, uint64_t return; dn = txh->txh_dnode; - /* first block */ - if (off != 0) - dmu_tx_count_write(txh, off, 1); - /* last block */ - if (len != DMU_OBJECT_END) - dmu_tx_count_write(txh, off+len, 1); - - dmu_tx_count_dnode(txh); - if (off >= (dn->dn_maxblkid+1) * dn->dn_datablksz) return; if (len == DMU_OBJECT_END) len = (dn->dn_maxblkid+1) * dn->dn_datablksz - off; + dmu_tx_count_dnode(txh); + /* - * For i/o error checking, read the first and last level-0 - * blocks, and all the level-1 blocks. The above count_write's - * have already taken care of the level-0 blocks. + * For i/o error checking, we read the first and last level-0 + * blocks if they are not aligned, and all the level-1 blocks. + * + * Note: dbuf_free_range() assumes that we have not instantiated + * any level-0 dbufs that will be completely freed. Therefore we must + * exercise care to not read or count the first and last blocks + * if they are blocksize-aligned. + */ + if (dn->dn_datablkshift == 0) { + if (off != 0 || len < dn->dn_datablksz) + dmu_tx_count_write(txh, 0, dn->dn_datablksz); + } else { + /* first block will be modified if it is not aligned */ + if (!IS_P2ALIGNED(off, 1 << dn->dn_datablkshift)) + dmu_tx_count_write(txh, off, 1); + /* last block will be modified if it is not aligned */ + if (!IS_P2ALIGNED(off + len, 1 << dn->dn_datablkshift)) + dmu_tx_count_write(txh, off+len, 1); + } + + /* + * Check level-1 blocks. */ if (dn->dn_nlevels > 1) { - shift = dn->dn_datablkshift + dn->dn_indblkshift - + int shift = dn->dn_datablkshift + dn->dn_indblkshift - SPA_BLKPTRSHIFT; - start = off >> shift; - end = dn->dn_datablkshift ? ((off+len) >> shift) : 0; + uint64_t start = off >> shift; + uint64_t end = (off + len) >> shift; + + ASSERT(dn->dn_datablkshift != 0); + ASSERT(dn->dn_indblkshift != 0); zio = zio_root(tx->tx_pool->dp_spa, NULL, NULL, ZIO_FLAG_CANFAIL); - for (i = start; i <= end; i++) { + for (uint64_t i = start; i <= end; i++) { uint64_t ibyte = i << shift; err = dnode_next_offset(dn, 0, &ibyte, 2, 1, 0); i = ibyte >> shift; Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode.c ============================================================================== --- stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode.c Thu Jan 16 14:05:05 2014 (r260721) +++ stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode.c Thu Jan 16 14:08:11 2014 (r260722) @@ -117,6 +117,7 @@ dnode_cons(void *arg, void *unused, int dn->dn_id_flags = 0; dn->dn_dbufs_count = 0; + dn->dn_unlisted_l0_blkid = 0; list_create(&dn->dn_dbufs, sizeof (dmu_buf_impl_t), offsetof(dmu_buf_impl_t, db_link)); @@ -170,6 +171,7 @@ dnode_dest(void *arg, void *unused) ASSERT0(dn->dn_id_flags); ASSERT0(dn->dn_dbufs_count); + ASSERT0(dn->dn_unlisted_l0_blkid); list_destroy(&dn->dn_dbufs); } @@ -475,6 +477,7 @@ dnode_destroy(dnode_t *dn) dn->dn_newuid = 0; dn->dn_newgid = 0; dn->dn_id_flags = 0; + dn->dn_unlisted_l0_blkid = 0; dmu_zfetch_rele(&dn->dn_zfetch); kmem_cache_free(dnode_cache, dn); @@ -705,6 +708,7 @@ dnode_move_impl(dnode_t *odn, dnode_t *n ASSERT(list_is_empty(&ndn->dn_dbufs)); list_move_tail(&ndn->dn_dbufs, &odn->dn_dbufs); ndn->dn_dbufs_count = odn->dn_dbufs_count; + ndn->dn_unlisted_l0_blkid = odn->dn_unlisted_l0_blkid; ndn->dn_bonus = odn->dn_bonus; ndn->dn_have_spill = odn->dn_have_spill; ndn->dn_zio = odn->dn_zio; @@ -739,6 +743,7 @@ dnode_move_impl(dnode_t *odn, dnode_t *n list_create(&odn->dn_dbufs, sizeof (dmu_buf_impl_t), offsetof(dmu_buf_impl_t, db_link)); odn->dn_dbufs_count = 0; + odn->dn_unlisted_l0_blkid = 0; odn->dn_bonus = NULL; odn->dn_zfetch.zf_dnode = NULL; @@ -1528,7 +1533,7 @@ dnode_free_range(dnode_t *dn, uint64_t o blkshift = dn->dn_datablkshift; epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT; - if (len == -1ULL) { + if (len == DMU_OBJECT_END) { len = UINT64_MAX - off; trunc = TRUE; } Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_destroy.c ============================================================================== --- stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_destroy.c Thu Jan 16 14:05:05 2014 (r260721) +++ stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_destroy.c Thu Jan 16 14:08:11 2014 (r260722) @@ -899,7 +899,7 @@ dsl_destroy_head(const char *name) for (uint64_t obj = 0; error == 0; error = dmu_object_next(os, &obj, FALSE, prev_snap_txg)) - (void) dmu_free_object(os, obj); + (void) dmu_free_long_object(os, obj); /* sync out all frees */ txg_wait_synced(dmu_objset_pool(os), 0); dmu_objset_disown(os, FTAG); Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu.h ============================================================================== --- stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu.h Thu Jan 16 14:05:05 2014 (r260721) +++ stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu.h Thu Jan 16 14:08:11 2014 (r260722) @@ -21,7 +21,7 @@ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. - * Copyright (c) 2012 by Delphix. All rights reserved. + * Copyright (c) 2013 by Delphix. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2012, Joyent, Inc. All rights reserved. */ @@ -583,7 +583,7 @@ int dmu_free_range(objset_t *os, uint64_ uint64_t size, dmu_tx_t *tx); int dmu_free_long_range(objset_t *os, uint64_t object, uint64_t offset, uint64_t size); -int dmu_free_object(objset_t *os, uint64_t object); +int dmu_free_long_object(objset_t *os, uint64_t object); /* * Convenience functions. Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_impl.h ============================================================================== --- stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_impl.h Thu Jan 16 14:05:05 2014 (r260721) +++ stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_impl.h Thu Jan 16 14:08:11 2014 (r260722) @@ -21,8 +21,11 @@ /* * Copyright 2010 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. + */ +/* * Copyright (c) 2012, Joyent, Inc. All rights reserved. * Copyright (c) 2012, Martin Matuska . All rights reserved. + * Copyright (c) 2013 by Delphix. All rights reserved. */ #ifndef _SYS_DMU_IMPL_H @@ -293,6 +296,9 @@ typedef struct dmu_sendarg { uint64_t dsa_toguid; int dsa_err; dmu_pendop_t dsa_pending_op; + boolean_t dsa_incremental; + uint64_t dsa_last_data_object; + uint64_t dsa_last_data_offset; } dmu_sendarg_t; Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_send.h ============================================================================== --- stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_send.h Thu Jan 16 14:05:05 2014 (r260721) +++ stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_send.h Thu Jan 16 14:08:11 2014 (r260722) @@ -21,7 +21,7 @@ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. - * Copyright (c) 2012 by Delphix. All rights reserved. + * Copyright (c) 2013 by Delphix. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2012, Joyent, Inc. All rights reserved. */ @@ -74,5 +74,6 @@ int dmu_recv_stream(dmu_recv_cookie_t *d #endif int cleanup_fd, uint64_t *action_handlep); int dmu_recv_end(dmu_recv_cookie_t *drc, void *owner); +boolean_t dmu_objset_is_receiving(objset_t *os); #endif /* _DMU_SEND_H */ Modified: stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dnode.h ============================================================================== --- stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dnode.h Thu Jan 16 14:05:05 2014 (r260721) +++ stable/9/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dnode.h Thu Jan 16 14:08:11 2014 (r260722) @@ -20,7 +20,7 @@ */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. - * Copyright (c) 2012 by Delphix. All rights reserved. + * Copyright (c) 2013 by Delphix. All rights reserved. */ #ifndef _SYS_DNODE_H @@ -188,6 +188,8 @@ typedef struct dnode { /* protected by dn_dbufs_mtx; declared here to fill 32-bit hole */ uint32_t dn_dbufs_count; /* count of dn_dbufs */ + /* There are no level-0 blocks of this blkid or higher in dn_dbufs */ + uint64_t dn_unlisted_l0_blkid; /* protected by os_lock: */ list_node_t dn_dirty_link[TXG_SIZE]; /* next on dataset's dirty */