Re: [PATCH] Btrfs: fix ordered extent leak when failing to start a transaction
From: David Sterba <hidden>
Date: 2012-08-24 14:19:43
On Wed, Aug 22, 2012 at 11:13:25AM +0800, bo.li.liu@oracle.com wrote:
quoted hunk ↗ jump to hunk
We cannot just return error before freeing ordered extent and releasing reserved space when we fail to start a transacion. Signed-off-by: Liu Bo <redacted> --- fs/btrfs/inode.c | 7 +++++-- 1 files changed, 5 insertions(+), 2 deletions(-)diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 7131fac..f6c43ff 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c@@ -1883,8 +1883,11 @@ static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent) trans = btrfs_join_transaction_nolock(root); else trans = btrfs_join_transaction(root); - if (IS_ERR(trans)) - return PTR_ERR(trans); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + trans = NULL; + goto out; + }
The same code is used a few lines below (after locking the range) when transaction fails, so I agree that this should be done here as well. Ack. david