From 21d48b5eac54ba7b2e4d8bf0d0e747082e0e8577 Mon Sep 17 00:00:00 2001 From: Paul Dagnelie Date: Tue, 20 Sep 2016 10:02:29 -0700 Subject: OpenZFS 9438 - Holes can lose birth time info if a block has a mix of birth times As reported by https://github.com/zfsonlinux/zfs/issues/4996, there is yet another hole birth issue. In this one, if a block is entirely holes, but the birth times are not all the same, we lose that information by creating one hole with the current txg as its birth time. The ZoL PR's fix approach is incorrect. Ultimately, the problem here is that when you truncate and write a file in the same transaction group, the dbuf for the indirect block will be zeroed out to deal with the truncation, and then written for the write. During this process, we will lose hole birth time information for any holes in the range. In the case where a dnode is being freed, we need to determine whether the block should be converted to a higher-level hole in the zio pipeline, and if so do it when the dnode is being synced out. Porting Notes: * The DMU_OBJECT_END change in zfs_znode.c was already applied. * Added test cases from #5675 provided by @rincebrain for hole_birth issues. These test cases should be pushed upstream to OpenZFS. * Updated mk_files which is used by several rsend tests so the files created are a little more interesting and may contain holes. Authored by: Paul Dagnelie Reviewed by: Matt Ahrens Reviewed by: George Wilson Approved by: Robert Mustacchi Ported-by: Brian Behlendorf OpenZFS-issue: https://www.illumos.org/issues/9438 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/738e2a3c External-issue: DLPX-46861 Closes #7746 --- module/zfs/dmu_object.c | 4 +++ module/zfs/dnode.c | 68 +++++++++++++++++++++++++++++++++++++++++++++++++ module/zfs/dnode_sync.c | 54 ++++++++++++++++++++++----------------- 3 files changed, 102 insertions(+), 24 deletions(-) (limited to 'module') diff --git a/module/zfs/dmu_object.c b/module/zfs/dmu_object.c index b9960782e..afee869b6 100644 --- a/module/zfs/dmu_object.c +++ b/module/zfs/dmu_object.c @@ -311,6 +311,10 @@ dmu_object_free(objset_t *os, uint64_t object, dmu_tx_t *tx) return (err); ASSERT(dn->dn_type != DMU_OT_NONE); + /* + * If we don't create this free range, we'll leak indirect blocks when + * we get to freeing the dnode in syncing context. + */ dnode_free_range(dn, 0, DMU_OBJECT_END, tx); dnode_free(dn, tx); dnode_rele(dn, FTAG); diff --git a/module/zfs/dnode.c b/module/zfs/dnode.c index 7939a7ba6..0be72e90a 100644 --- a/module/zfs/dnode.c +++ b/module/zfs/dnode.c @@ -1889,6 +1889,72 @@ dnode_dirty_l1(dnode_t *dn, uint64_t l1blkid, dmu_tx_t *tx) } } +/* + * Dirty all the in-core level-1 dbufs in the range specified by start_blkid + * and end_blkid. + */ +static void +dnode_dirty_l1range(dnode_t *dn, uint64_t start_blkid, uint64_t end_blkid, + dmu_tx_t *tx) +{ + dmu_buf_impl_t db_search; + dmu_buf_impl_t *db; + avl_index_t where; + + mutex_enter(&dn->dn_dbufs_mtx); + + db_search.db_level = 1; + db_search.db_blkid = start_blkid + 1; + db_search.db_state = DB_SEARCH; + for (;;) { + + db = avl_find(&dn->dn_dbufs, &db_search, &where); + if (db == NULL) + db = avl_nearest(&dn->dn_dbufs, where, AVL_AFTER); + + if (db == NULL || db->db_level != 1 || + db->db_blkid >= end_blkid) { + break; + } + + /* + * Setup the next blkid we want to search for. + */ + db_search.db_blkid = db->db_blkid + 1; + ASSERT3U(db->db_blkid, >=, start_blkid); + + /* + * If the dbuf transitions to DB_EVICTING while we're trying + * to dirty it, then we will be unable to discover it in + * the dbuf hash table. This will result in a call to + * dbuf_create() which needs to acquire the dn_dbufs_mtx + * lock. To avoid a deadlock, we drop the lock before + * dirtying the level-1 dbuf. + */ + mutex_exit(&dn->dn_dbufs_mtx); + dnode_dirty_l1(dn, db->db_blkid, tx); + mutex_enter(&dn->dn_dbufs_mtx); + } + +#ifdef ZFS_DEBUG + /* + * Walk all the in-core level-1 dbufs and verify they have been dirtied. + */ + db_search.db_level = 1; + db_search.db_blkid = start_blkid + 1; + db_search.db_state = DB_SEARCH; + db = avl_find(&dn->dn_dbufs, &db_search, &where); + if (db == NULL) + db = avl_nearest(&dn->dn_dbufs, where, AVL_AFTER); + for (; db != NULL; db = AVL_NEXT(&dn->dn_dbufs, db)) { + if (db->db_level != 1 || db->db_blkid >= end_blkid) + break; + ASSERT(db->db_dirtycnt > 0); + } +#endif + mutex_exit(&dn->dn_dbufs_mtx); +} + void dnode_free_range(dnode_t *dn, uint64_t off, uint64_t len, dmu_tx_t *tx) { @@ -2040,6 +2106,8 @@ dnode_free_range(dnode_t *dn, uint64_t off, uint64_t len, dmu_tx_t *tx) if (last != first) dnode_dirty_l1(dn, last, tx); + dnode_dirty_l1range(dn, first, last, tx); + int shift = dn->dn_datablkshift + dn->dn_indblkshift - SPA_BLKPTRSHIFT; for (uint64_t i = first + 1; i < last; i++) { diff --git a/module/zfs/dnode_sync.c b/module/zfs/dnode_sync.c index 22b401ab5..3202faf49 100644 --- a/module/zfs/dnode_sync.c +++ b/module/zfs/dnode_sync.c @@ -230,9 +230,24 @@ free_verify(dmu_buf_impl_t *db, uint64_t start, uint64_t end, dmu_tx_t *tx) } #endif +/* + * We don't usually free the indirect blocks here. If in one txg we have a + * free_range and a write to the same indirect block, it's important that we + * preserve the hole's birth times. Therefore, we don't free any any indirect + * blocks in free_children(). If an indirect block happens to turn into all + * holes, it will be freed by dbuf_write_children_ready, which happens at a + * point in the syncing process where we know for certain the contents of the + * indirect block. + * + * However, if we're freeing a dnode, its space accounting must go to zero + * before we actually try to free the dnode, or we will trip an assertion. In + * addition, we know the case described above cannot occur, because the dnode is + * being freed. Therefore, we free the indirect blocks immediately in that + * case. + */ static void free_children(dmu_buf_impl_t *db, uint64_t blkid, uint64_t nblks, - dmu_tx_t *tx) + boolean_t free_indirects, dmu_tx_t *tx) { dnode_t *dn; blkptr_t *bp; @@ -284,32 +299,16 @@ free_children(dmu_buf_impl_t *db, uint64_t blkid, uint64_t nblks, rw_exit(&dn->dn_struct_rwlock); ASSERT3P(bp, ==, subdb->db_blkptr); - free_children(subdb, blkid, nblks, tx); + free_children(subdb, blkid, nblks, free_indirects, tx); dbuf_rele(subdb, FTAG); } } - /* If this whole block is free, free ourself too. */ - for (i = 0, bp = db->db.db_data; i < 1ULL << epbs; i++, bp++) { - if (!BP_IS_HOLE(bp)) - break; - } - if (i == 1 << epbs) { - /* - * We only found holes. Grab the rwlock to prevent - * anybody from reading the blocks we're about to - * zero out. - */ - rw_enter(&dn->dn_struct_rwlock, RW_WRITER); + if (free_indirects) { + for (i = 0, bp = db->db.db_data; i < 1 << epbs; i++, bp++) + ASSERT(BP_IS_HOLE(bp)); bzero(db->db.db_data, db->db.db_size); - rw_exit(&dn->dn_struct_rwlock); free_blocks(dn, db->db_blkptr, 1, tx); - } else { - /* - * Partial block free; must be marked dirty so that it - * will be written out. - */ - ASSERT(db->db_dirtycnt > 0); } DB_DNODE_EXIT(db); @@ -322,7 +321,7 @@ free_children(dmu_buf_impl_t *db, uint64_t blkid, uint64_t nblks, */ static void dnode_sync_free_range_impl(dnode_t *dn, uint64_t blkid, uint64_t nblks, - dmu_tx_t *tx) + boolean_t free_indirects, dmu_tx_t *tx) { blkptr_t *bp = dn->dn_phys->dn_blkptr; int dnlevel = dn->dn_phys->dn_nlevels; @@ -362,7 +361,7 @@ dnode_sync_free_range_impl(dnode_t *dn, uint64_t blkid, uint64_t nblks, TRUE, FALSE, FTAG, &db)); rw_exit(&dn->dn_struct_rwlock); - free_children(db, blkid, nblks, tx); + free_children(db, blkid, nblks, free_indirects, tx); dbuf_rele(db, FTAG); } } @@ -387,6 +386,7 @@ dnode_sync_free_range_impl(dnode_t *dn, uint64_t blkid, uint64_t nblks, typedef struct dnode_sync_free_range_arg { dnode_t *dsfra_dnode; dmu_tx_t *dsfra_tx; + boolean_t dsfra_free_indirects; } dnode_sync_free_range_arg_t; static void @@ -396,7 +396,8 @@ dnode_sync_free_range(void *arg, uint64_t blkid, uint64_t nblks) dnode_t *dn = dsfra->dsfra_dnode; mutex_exit(&dn->dn_mtx); - dnode_sync_free_range_impl(dn, blkid, nblks, dsfra->dsfra_tx); + dnode_sync_free_range_impl(dn, blkid, nblks, + dsfra->dsfra_free_indirects, dsfra->dsfra_tx); mutex_enter(&dn->dn_mtx); } @@ -712,6 +713,11 @@ dnode_sync(dnode_t *dn, dmu_tx_t *tx) dnode_sync_free_range_arg_t dsfra; dsfra.dsfra_dnode = dn; dsfra.dsfra_tx = tx; + dsfra.dsfra_free_indirects = freeing_dnode; + if (freeing_dnode) { + ASSERT(range_tree_contains(dn->dn_free_ranges[txgoff], + 0, dn->dn_maxblkid + 1)); + } mutex_enter(&dn->dn_mtx); range_tree_vacate(dn->dn_free_ranges[txgoff], dnode_sync_free_range, &dsfra); -- cgit v1.2.3