diff options
Diffstat (limited to 'fs/btrfs')
-rw-r--r-- | fs/btrfs/btrfs_inode.h | 10 | ||||
-rw-r--r-- | fs/btrfs/disk-io.c | 10 | ||||
-rw-r--r-- | fs/btrfs/extent-tree.c | 2 | ||||
-rw-r--r-- | fs/btrfs/extent_io.c | 60 | ||||
-rw-r--r-- | fs/btrfs/file.c | 16 | ||||
-rw-r--r-- | fs/btrfs/ordered-data.c | 31 | ||||
-rw-r--r-- | fs/btrfs/qgroup.c | 32 | ||||
-rw-r--r-- | fs/btrfs/scrub.c | 2 | ||||
-rw-r--r-- | fs/btrfs/super.c | 8 | ||||
-rw-r--r-- | fs/btrfs/tree-log.c | 17 | ||||
-rw-r--r-- | fs/btrfs/zoned.c | 13 |
11 files changed, 145 insertions, 56 deletions
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h index 91c994b569f3..6ed495ca7a31 100644 --- a/fs/btrfs/btrfs_inode.h +++ b/fs/btrfs/btrfs_inode.h @@ -89,6 +89,16 @@ enum { BTRFS_INODE_FREE_SPACE_INODE, /* Set when there are no capabilities in XATTs for the inode. */ BTRFS_INODE_NO_CAP_XATTR, + /* + * Set if an error happened when doing a COW write before submitting a + * bio or during writeback. Used for both buffered writes and direct IO + * writes. This is to signal a fast fsync that it has to wait for + * ordered extents to complete and therefore not log extent maps that + * point to unwritten extents (when an ordered extent completes and it + * has the BTRFS_ORDERED_IOERR flag set, it drops extent maps in its + * range). + */ + BTRFS_INODE_COW_WRITE_ERROR, }; /* in memory btrfs inode */ diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 1b20b3e390df..38cdb8875e8e 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -4538,18 +4538,10 @@ static void btrfs_destroy_delayed_refs(struct btrfs_transaction *trans, struct btrfs_fs_info *fs_info) { struct rb_node *node; - struct btrfs_delayed_ref_root *delayed_refs; + struct btrfs_delayed_ref_root *delayed_refs = &trans->delayed_refs; struct btrfs_delayed_ref_node *ref; - delayed_refs = &trans->delayed_refs; - spin_lock(&delayed_refs->lock); - if (atomic_read(&delayed_refs->num_entries) == 0) { - spin_unlock(&delayed_refs->lock); - btrfs_debug(fs_info, "delayed_refs has NO entry"); - return; - } - while ((node = rb_first_cached(&delayed_refs->href_root)) != NULL) { struct btrfs_delayed_ref_head *head; struct rb_node *n; diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 47d48233b592..3774c191e36d 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -477,7 +477,7 @@ again: if (path->slots[0] >= nritems) { ret = btrfs_next_leaf(root, path); if (ret) { - if (ret > 1) + if (ret > 0) return -ENOENT; return ret; } diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 597387e9f040..f688fab55251 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3689,6 +3689,8 @@ static struct extent_buffer *grab_extent_buffer( struct folio *folio = page_folio(page); struct extent_buffer *exists; + lockdep_assert_held(&page->mapping->i_private_lock); + /* * For subpage case, we completely rely on radix tree to ensure we * don't try to insert two ebs for the same bytenr. So here we always @@ -3756,13 +3758,14 @@ static int check_eb_alignment(struct btrfs_fs_info *fs_info, u64 start) * The caller needs to free the existing folios and retry using the same order. */ static int attach_eb_folio_to_filemap(struct extent_buffer *eb, int i, + struct btrfs_subpage *prealloc, struct extent_buffer **found_eb_ret) { struct btrfs_fs_info *fs_info = eb->fs_info; struct address_space *mapping = fs_info->btree_inode->i_mapping; const unsigned long index = eb->start >> PAGE_SHIFT; - struct folio *existing_folio; + struct folio *existing_folio = NULL; int ret; ASSERT(found_eb_ret); @@ -3774,12 +3777,14 @@ retry: ret = filemap_add_folio(mapping, eb->folios[i], index + i, GFP_NOFS | __GFP_NOFAIL); if (!ret) - return 0; + goto finish; existing_folio = filemap_lock_folio(mapping, index + i); /* The page cache only exists for a very short time, just retry. */ - if (IS_ERR(existing_folio)) + if (IS_ERR(existing_folio)) { + existing_folio = NULL; goto retry; + } /* For now, we should only have single-page folios for btree inode. */ ASSERT(folio_nr_pages(existing_folio) == 1); @@ -3790,14 +3795,13 @@ retry: return -EAGAIN; } - if (fs_info->nodesize < PAGE_SIZE) { - /* - * We're going to reuse the existing page, can drop our page - * and subpage structure now. - */ +finish: + spin_lock(&mapping->i_private_lock); + if (existing_folio && fs_info->nodesize < PAGE_SIZE) { + /* We're going to reuse the existing page, can drop our folio now. */ __free_page(folio_page(eb->folios[i], 0)); eb->folios[i] = existing_folio; - } else { + } else if (existing_folio) { struct extent_buffer *existing_eb; existing_eb = grab_extent_buffer(fs_info, @@ -3805,6 +3809,7 @@ retry: if (existing_eb) { /* The extent buffer still exists, we can use it directly. */ *found_eb_ret = existing_eb; + spin_unlock(&mapping->i_private_lock); folio_unlock(existing_folio); folio_put(existing_folio); return 1; @@ -3813,6 +3818,22 @@ retry: __free_page(folio_page(eb->folios[i], 0)); eb->folios[i] = existing_folio; } + eb->folio_size = folio_size(eb->folios[i]); + eb->folio_shift = folio_shift(eb->folios[i]); + /* Should not fail, as we have preallocated the memory. */ + ret = attach_extent_buffer_folio(eb, eb->folios[i], prealloc); + ASSERT(!ret); + /* + * To inform we have an extra eb under allocation, so that + * detach_extent_buffer_page() won't release the folio private when the + * eb hasn't been inserted into radix tree yet. + * + * The ref will be decreased when the eb releases the page, in + * detach_extent_buffer_page(). Thus needs no special handling in the + * error path. + */ + btrfs_folio_inc_eb_refs(fs_info, eb->folios[i]); + spin_unlock(&mapping->i_private_lock); return 0; } @@ -3824,7 +3845,6 @@ struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info, int attached = 0; struct extent_buffer *eb; struct extent_buffer *existing_eb = NULL; - struct address_space *mapping = fs_info->btree_inode->i_mapping; struct btrfs_subpage *prealloc = NULL; u64 lockdep_owner = owner_root; bool page_contig = true; @@ -3890,7 +3910,7 @@ reallocate: for (int i = 0; i < num_folios; i++) { struct folio *folio; - ret = attach_eb_folio_to_filemap(eb, i, &existing_eb); + ret = attach_eb_folio_to_filemap(eb, i, prealloc, &existing_eb); if (ret > 0) { ASSERT(existing_eb); goto out; @@ -3927,24 +3947,6 @@ reallocate: * and free the allocated page. */ folio = eb->folios[i]; - eb->folio_size = folio_size(folio); - eb->folio_shift = folio_shift(folio); - spin_lock(&mapping->i_private_lock); - /* Should not fail, as we have preallocated the memory */ - ret = attach_extent_buffer_folio(eb, folio, prealloc); - ASSERT(!ret); - /* - * To inform we have extra eb under allocation, so that - * detach_extent_buffer_page() won't release the folio private - * when the eb hasn't yet been inserted into radix tree. - * - * The ref will be decreased when the eb released the page, in - * detach_extent_buffer_page(). - * Thus needs no special handling in error path. - */ - btrfs_folio_inc_eb_refs(fs_info, folio); - spin_unlock(&mapping->i_private_lock); - WARN_ON(btrfs_folio_test_dirty(fs_info, folio, eb->start, eb->len)); /* diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index e764ac3f22e2..d90138683a0a 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1885,6 +1885,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) */ if (full_sync || btrfs_is_zoned(fs_info)) { ret = btrfs_wait_ordered_range(inode, start, len); + clear_bit(BTRFS_INODE_COW_WRITE_ERROR, &BTRFS_I(inode)->runtime_flags); } else { /* * Get our ordered extents as soon as possible to avoid doing @@ -1894,6 +1895,21 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) btrfs_get_ordered_extents_for_logging(BTRFS_I(inode), &ctx.ordered_extents); ret = filemap_fdatawait_range(inode->i_mapping, start, end); + if (ret) + goto out_release_extents; + + /* + * Check and clear the BTRFS_INODE_COW_WRITE_ERROR now after + * starting and waiting for writeback, because for buffered IO + * it may have been set during the end IO callback + * (end_bbio_data_write() -> btrfs_finish_ordered_extent()) in + * case an error happened and we need to wait for ordered + * extents to complete so that any extent maps that point to + * unwritten locations are dropped and we don't log them. + */ + if (test_and_clear_bit(BTRFS_INODE_COW_WRITE_ERROR, + &BTRFS_I(inode)->runtime_flags)) + ret = btrfs_wait_ordered_range(inode, start, len); } if (ret) diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c index c5bdd674f55c..35a413ce935d 100644 --- a/fs/btrfs/ordered-data.c +++ b/fs/btrfs/ordered-data.c @@ -388,6 +388,37 @@ bool btrfs_finish_ordered_extent(struct btrfs_ordered_extent *ordered, ret = can_finish_ordered_extent(ordered, page, file_offset, len, uptodate); spin_unlock_irqrestore(&inode->ordered_tree_lock, flags); + /* + * If this is a COW write it means we created new extent maps for the + * range and they point to unwritten locations if we got an error either + * before submitting a bio or during IO. + * + * We have marked the ordered extent with BTRFS_ORDERED_IOERR, and we + * are queuing its completion below. During completion, at + * btrfs_finish_one_ordered(), we will drop the extent maps for the + * unwritten extents. + * + * However because completion runs in a work queue we can end up having + * a fast fsync running before that. In the case of direct IO, once we + * unlock the inode the fsync might start, and we queue the completion + * before unlocking the inode. In the case of buffered IO when writeback + * finishes (end_bbio_data_write()) we queue the completion, so if the + * writeback was triggered by a fast fsync, the fsync might start + * logging before ordered extent completion runs in the work queue. + * + * The fast fsync will log file extent items based on the extent maps it + * finds, so if by the time it collects extent maps the ordered extent + * completion didn't happen yet, it will log file extent items that + * point to unwritten extents, resulting in a corruption if a crash + * happens and the log tree is replayed. Note that a fast fsync does not + * wait for completion of ordered extents in order to reduce latency. + * + * Set a flag in the inode so that the next fast fsync will wait for + * ordered extents to complete before starting to log. + */ + if (!uptodate && !test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags)) + set_bit(BTRFS_INODE_COW_WRITE_ERROR, &inode->runtime_flags); + if (ret) btrfs_queue_ordered_fn(ordered); return ret; diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index eb28141d5c37..fc2a7ea26354 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -468,6 +468,7 @@ int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info) } if (!qgroup) { struct btrfs_qgroup *prealloc; + struct btrfs_root *tree_root = fs_info->tree_root; prealloc = kzalloc(sizeof(*prealloc), GFP_KERNEL); if (!prealloc) { @@ -475,6 +476,25 @@ int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info) goto out; } qgroup = add_qgroup_rb(fs_info, prealloc, found_key.offset); + /* + * If a qgroup exists for a subvolume ID, it is possible + * that subvolume has been deleted, in which case + * re-using that ID would lead to incorrect accounting. + * + * Ensure that we skip any such subvol ids. + * + * We don't need to lock because this is only called + * during mount before we start doing things like creating + * subvolumes. + */ + if (is_fstree(qgroup->qgroupid) && + qgroup->qgroupid > tree_root->free_objectid) + /* + * Don't need to check against BTRFS_LAST_FREE_OBJECTID, + * as it will get checked on the next call to + * btrfs_get_free_objectid. + */ + tree_root->free_objectid = qgroup->qgroupid + 1; } ret = btrfs_sysfs_add_one_qgroup(fs_info, qgroup); if (ret < 0) @@ -3820,14 +3840,14 @@ qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid, /* we're resuming qgroup rescan at mount time */ if (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN)) { - btrfs_warn(fs_info, + btrfs_debug(fs_info, "qgroup rescan init failed, qgroup rescan is not queued"); ret = -EINVAL; } else if (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_ON)) { - btrfs_warn(fs_info, + btrfs_debug(fs_info, "qgroup rescan init failed, qgroup is not enabled"); - ret = -EINVAL; + ret = -ENOTCONN; } if (ret) @@ -3838,14 +3858,12 @@ qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid, if (init_flags) { if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) { - btrfs_warn(fs_info, - "qgroup rescan is already in progress"); ret = -EINPROGRESS; } else if (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_ON)) { - btrfs_warn(fs_info, + btrfs_debug(fs_info, "qgroup rescan init failed, qgroup is not enabled"); - ret = -EINVAL; + ret = -ENOTCONN; } else if (btrfs_qgroup_mode(fs_info) == BTRFS_QGROUP_MODE_DISABLED) { /* Quota disable is in progress */ ret = -EBUSY; diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 4b22cfe9a98c..afd6932f5e89 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -2100,7 +2100,7 @@ static int scrub_simple_mirror(struct scrub_ctx *sctx, struct btrfs_fs_info *fs_info = sctx->fs_info; const u64 logical_end = logical_start + logical_length; u64 cur_logical = logical_start; - int ret; + int ret = 0; /* The range must be inside the bg */ ASSERT(logical_start >= bg->start && logical_end <= bg->start + bg->length); diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 2dbc930a20f7..f05cce7c8b8d 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -119,6 +119,7 @@ enum { Opt_thread_pool, Opt_treelog, Opt_user_subvol_rm_allowed, + Opt_norecovery, /* Rescue options */ Opt_rescue, @@ -245,6 +246,8 @@ static const struct fs_parameter_spec btrfs_fs_parameters[] = { __fsparam(NULL, "nologreplay", Opt_nologreplay, fs_param_deprecated, NULL), /* Deprecated, with alias rescue=usebackuproot */ __fsparam(NULL, "usebackuproot", Opt_usebackuproot, fs_param_deprecated, NULL), + /* For compatibility only, alias for "rescue=nologreplay". */ + fsparam_flag("norecovery", Opt_norecovery), /* Debugging options. */ fsparam_flag_no("enospc_debug", Opt_enospc_debug), @@ -438,6 +441,11 @@ static int btrfs_parse_param(struct fs_context *fc, struct fs_parameter *param) "'nologreplay' is deprecated, use 'rescue=nologreplay' instead"); btrfs_set_opt(ctx->mount_opt, NOLOGREPLAY); break; + case Opt_norecovery: + btrfs_info(NULL, +"'norecovery' is for compatibility only, recommended to use 'rescue=nologreplay'"); + btrfs_set_opt(ctx->mount_opt, NOLOGREPLAY); + break; case Opt_flushoncommit: if (result.negated) btrfs_clear_opt(ctx->mount_opt, FLUSHONCOMMIT); diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 5146387b416b..26a2e5aa08e9 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -4860,18 +4860,23 @@ static int btrfs_log_prealloc_extents(struct btrfs_trans_handle *trans, path->slots[0]++; continue; } - if (!dropped_extents) { - /* - * Avoid logging extent items logged in past fsync calls - * and leading to duplicate keys in the log tree. - */ + /* + * Avoid overlapping items in the log tree. The first time we + * get here, get rid of everything from a past fsync. After + * that, if the current extent starts before the end of the last + * extent we copied, truncate the last one. This can happen if + * an ordered extent completion modifies the subvolume tree + * while btrfs_next_leaf() has the tree unlocked. + */ + if (!dropped_extents || key.offset < truncate_offset) { ret = truncate_inode_items(trans, root->log_root, inode, - truncate_offset, + min(key.offset, truncate_offset), BTRFS_EXTENT_DATA_KEY); if (ret) goto out; dropped_extents = true; } + truncate_offset = btrfs_file_extent_end(path); if (ins_nr == 0) start_slot = slot; ins_nr++; diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c index 9b43fa493219..947a87576f6c 100644 --- a/fs/btrfs/zoned.c +++ b/fs/btrfs/zoned.c @@ -1290,7 +1290,7 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx, struct btrfs_chunk_map *map) { struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace; - struct btrfs_device *device = map->stripes[zone_idx].dev; + struct btrfs_device *device; int dev_replace_is_ongoing = 0; unsigned int nofs_flag; struct blk_zone zone; @@ -1298,7 +1298,11 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx, info->physical = map->stripes[zone_idx].physical; + down_read(&dev_replace->rwsem); + device = map->stripes[zone_idx].dev; + if (!device->bdev) { + up_read(&dev_replace->rwsem); info->alloc_offset = WP_MISSING_DEV; return 0; } @@ -1308,6 +1312,7 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx, __set_bit(zone_idx, active); if (!btrfs_dev_is_sequential(device, info->physical)) { + up_read(&dev_replace->rwsem); info->alloc_offset = WP_CONVENTIONAL; return 0; } @@ -1315,11 +1320,9 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx, /* This zone will be used for allocation, so mark this zone non-empty. */ btrfs_dev_clear_zone_empty(device, info->physical); - down_read(&dev_replace->rwsem); dev_replace_is_ongoing = btrfs_dev_replace_is_ongoing(dev_replace); if (dev_replace_is_ongoing && dev_replace->tgtdev != NULL) btrfs_dev_clear_zone_empty(dev_replace->tgtdev, info->physical); - up_read(&dev_replace->rwsem); /* * The group is mapped to a sequential zone. Get the zone write pointer @@ -1330,6 +1333,7 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx, ret = btrfs_get_dev_zone(device, info->physical, &zone); memalloc_nofs_restore(nofs_flag); if (ret) { + up_read(&dev_replace->rwsem); if (ret != -EIO && ret != -EOPNOTSUPP) return ret; info->alloc_offset = WP_MISSING_DEV; @@ -1341,6 +1345,7 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx, "zoned: unexpected conventional zone %llu on device %s (devid %llu)", zone.start << SECTOR_SHIFT, rcu_str_deref(device->name), device->devid); + up_read(&dev_replace->rwsem); return -EIO; } @@ -1368,6 +1373,8 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx, break; } + up_read(&dev_replace->rwsem); + return 0; } |