diff options
| author | Linus Torvalds <torvalds@linux-foundation.org> | 2018-11-11 16:54:38 -0600 |
|---|---|---|
| committer | Linus Torvalds <torvalds@linux-foundation.org> | 2018-11-11 16:54:38 -0600 |
| commit | 63a42e1a5cb3d01eef2f370c11d8733a32f12f86 (patch) | |
| tree | e28d678bc58570c67242b770a95e09b433d3ecee /fs/btrfs/tree-log.c | |
| parent | c140f8b072d16595c83d4d16a05693e72d9b1973 (diff) | |
| parent | d6fd0ae25c6495674dc5a41a8d16bc8e0073276d (diff) | |
Merge tag 'for-4.20-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux
Pull btrfs fixes from David Sterba:
"Several fixes to recent release (4.19, fixes tagged for stable) and
other fixes"
* tag 'for-4.20-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
Btrfs: fix missing delayed iputs on unmount
Btrfs: fix data corruption due to cloning of eof block
Btrfs: fix infinite loop on inode eviction after deduplication of eof block
Btrfs: fix deadlock on tree root leaf when finding free extent
btrfs: avoid link error with CONFIG_NO_AUTO_INLINE
btrfs: tree-checker: Fix misleading group system information
Btrfs: fix missing data checksums after a ranged fsync (msync)
btrfs: fix pinned underflow after transaction aborted
Btrfs: fix cur_offset in the error case for nocow
Diffstat (limited to 'fs/btrfs/tree-log.c')
| -rw-r--r-- | fs/btrfs/tree-log.c | 17 |
1 files changed, 17 insertions, 0 deletions
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index e07f3376b7df..a5ce99a6c936 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -4396,6 +4396,23 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans, logged_end = end; list_for_each_entry_safe(em, n, &tree->modified_extents, list) { + /* + * Skip extents outside our logging range. It's important to do + * it for correctness because if we don't ignore them, we may + * log them before their ordered extent completes, and therefore + * we could log them without logging their respective checksums + * (the checksum items are added to the csum tree at the very + * end of btrfs_finish_ordered_io()). Also leave such extents + * outside of our range in the list, since we may have another + * ranged fsync in the near future that needs them. If an extent + * outside our range corresponds to a hole, log it to avoid + * leaving gaps between extents (fsck will complain when we are + * not using the NO_HOLES feature). + */ + if ((em->start > end || em->start + em->len <= start) && + em->block_start != EXTENT_MAP_HOLE) + continue; + list_del_init(&em->list); /* * Just an arbitrary number, this can be really CPU intensive |