| Message ID | 1406807619-20859-1-git-send-email-fdmanana@suse.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 783ea3b..8ca6761 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -5833,7 +5833,7 @@ again: } ret = 0; done: - unlock_up(path, 0, 1, 0, NULL); + btrfs_unlock_up_safe(path, 1); path->leave_spinning = old_spinning; if (!old_spinning) btrfs_set_path_blocking(path);
Calling unlock_up() to release our new path doesn't release the read lock on the node at level 1, because our return path has path->slots[0] == 0, which makes unlock_up() skip unlocking that node. Since we don't need to return that node locked, call btrfs_unlock_up_safe() instead of unlock_up(), which will release all nodes in the path (except the leaf of course). For any level N >= 2, the corresponding node lock isn't released by unlock_up() too if path->slots[N - 1] == 0. Releasing the read lock immediately will allow concurrent writers to write lock that node at level 1 (or higher levels if applicable) while the btrfs_next_leaf() caller processes the leaf. Signed-off-by: Filipe Manana <fdmanana@suse.com> --- fs/btrfs/ctree.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)