aboutsummaryrefslogtreecommitdiff
path: root/fs/btrfs/dev-replace.c
diff options
context:
space:
mode:
authorMiao Xie <miaox@cn.fujitsu.com>2013-05-14 10:20:40 +0000
committerJosef Bacik <jbacik@fusionio.com>2013-06-14 11:29:32 -0400
commitd027824564c5fcee19109530b87c94c9908e910a (patch)
tree9ee6f6b7c9eaeee8494536061f868e732d607198 /fs/btrfs/dev-replace.c
parent3c64a1aba7cfcb04f79e76f859b3d66660275d59 (diff)
downloadlinux-d027824564c5fcee19109530b87c94c9908e910a.tar.gz
linux-d027824564c5fcee19109530b87c94c9908e910a.tar.bz2
linux-d027824564c5fcee19109530b87c94c9908e910a.zip
Btrfs: remove unnecessary ->s_umount in cleaner_kthread()
In order to avoid the R/O remount, we acquired ->s_umount lock during we deleted the dead snapshots and subvolumes. But it is unnecessary, because we have cleaner_mutex. We use cleaner_mutex to protect the process of the dead snapshots/subvolumes deletion. And when we remount the fs to be R/O, we also acquire this mutex to do cleanup after we change the status of the fs. That is this lock can serialize the above operations, the cleaner can be aware of the status of the fs, and if the cleaner is deleting the dead snapshots/subvolumes, the remount task will wait for it. So it is safe to remove ->s_umount in cleaner_kthread(). Cc: David Sterba <dsterba@suse.cz> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Diffstat (limited to 'fs/btrfs/dev-replace.c')
0 files changed, 0 insertions, 0 deletions