| Message ID | 20210906150428.2399128-1-naohiro.aota@wdc.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | btrfs: zoned: fix double counting of split ordered extent | expand |
Looks good,
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
On Tue, Sep 07, 2021 at 12:04:28AM +0900, Naohiro Aota wrote: > btrfs_add_ordered_extent_*() add num_bytes to fs_info->ordered_bytes. > Then, splitting an ordered extent will call btrfs_add_ordered_extent_*() > again for split extents, leading to double counting of the region of > a split extent. These leaked bytes are finally reported at unmount time > as follow. > > BTRFS info (device dm-1): at unmount dio bytes count 364544 > > This commit fixes the double counting by subtracting split extent's size > from fs_info->ordered_bytes. > > Fixes: d22002fd37bd ("btrfs: zoned: split ordered extent when bio is sent") > Cc: stable@vger.kernel.org # 5.12+ > Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> Added to misc-next, thanks.
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c index edb65abf0393..6b51fd2ec5ac 100644 --- a/fs/btrfs/ordered-data.c +++ b/fs/btrfs/ordered-data.c @@ -1049,6 +1049,7 @@ static int clone_ordered_extent(struct btrfs_ordered_extent *ordered, u64 pos, u64 len) { struct inode *inode = ordered->inode; + struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info; u64 file_offset = ordered->file_offset + pos; u64 disk_bytenr = ordered->disk_bytenr + pos; u64 num_bytes = len; @@ -1066,6 +1067,13 @@ static int clone_ordered_extent(struct btrfs_ordered_extent *ordered, u64 pos, else type = __ffs(flags_masked); + /* + * The splitting extent is already counted and will be added again + * in btrfs_add_ordered_extent_*(). Subtract num_bytes to avoid + * double counting. + */ + percpu_counter_add_batch(&fs_info->ordered_bytes, -num_bytes, + fs_info->delalloc_batch); if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered->flags)) { WARN_ON_ONCE(1); ret = btrfs_add_ordered_extent_compress(BTRFS_I(inode),
btrfs_add_ordered_extent_*() add num_bytes to fs_info->ordered_bytes. Then, splitting an ordered extent will call btrfs_add_ordered_extent_*() again for split extents, leading to double counting of the region of a split extent. These leaked bytes are finally reported at unmount time as follow. BTRFS info (device dm-1): at unmount dio bytes count 364544 This commit fixes the double counting by subtracting split extent's size from fs_info->ordered_bytes. Fixes: d22002fd37bd ("btrfs: zoned: split ordered extent when bio is sent") Cc: stable@vger.kernel.org # 5.12+ Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com> --- fs/btrfs/ordered-data.c | 8 ++++++++ 1 file changed, 8 insertions(+)