diff options
| author | Chris Mason <clm@fb.com> | 2016-12-13 09:14:42 -0800 |
|---|---|---|
| committer | Chris Mason <clm@fb.com> | 2016-12-13 09:14:42 -0800 |
| commit | 5f52a2c512a55500349aa261e469d099ede0f256 (patch) | |
| tree | fc1d55c8f31f12e6eb0536de01a90bf7dfbb329a /fs/btrfs/qgroup.c | |
| parent | 7c4c71ac8a72aea595f4cc7e09c2bcc61929c4ac (diff) | |
| parent | 2a7bf53f577e49c43de4ffa7776056de26db65d9 (diff) | |
| download | linux-stable-5f52a2c512a55500349aa261e469d099ede0f256.tar.gz linux-stable-5f52a2c512a55500349aa261e469d099ede0f256.tar.bz2 linux-stable-5f52a2c512a55500349aa261e469d099ede0f256.zip | |
Merge branch 'for-chris-4.10' of git://git.kernel.org/pub/scm/linux/kernel/git/fdmanana/linux into for-linus-4.10
Patches queued up by Filipe:
The most important change is still the fix for the extent tree
corruption that happens due to balance when qgroups are enabled (a
regression introduced in 4.7 by a fix for a regression from the last
qgroups rework). This has been hitting SLE and openSUSE users and QA
very badly, where transactions keep getting aborted when running
delayed references leaving the root filesystem in RO mode and nearly
unusable. There are fixes here that allow us to run xfstests again
with the integrity checker enabled, which has been impossible since 4.8
(apparently I'm the only one running xfstests with the integrity
checker enabled, which is useful to validate dirtied leafs, like
checking if there are keys out of order, etc). The rest are just some
trivial fixes, most of them tagged for stable, and two cleanups.
Signed-off-by: Chris Mason <clm@fb.com>
Diffstat (limited to 'fs/btrfs/qgroup.c')
| -rw-r--r-- | fs/btrfs/qgroup.c | 5 |
1 files changed, 1 insertions, 4 deletions
diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 3e473e9a4844..662821f1252c 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -2555,10 +2555,6 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work) int err = -ENOMEM; int ret = 0; - mutex_lock(&fs_info->qgroup_rescan_lock); - fs_info->qgroup_rescan_running = true; - mutex_unlock(&fs_info->qgroup_rescan_lock); - path = btrfs_alloc_path(); if (!path) goto out; @@ -2669,6 +2665,7 @@ qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid, sizeof(fs_info->qgroup_rescan_progress)); fs_info->qgroup_rescan_progress.objectid = progress_objectid; init_completion(&fs_info->qgroup_rescan_completion); + fs_info->qgroup_rescan_running = true; spin_unlock(&fs_info->qgroup_lock); mutex_unlock(&fs_info->qgroup_rescan_lock); |