Btrfs: fix extent state leak on transaction abortion
authorLiu Bo <bo.li.liu@oracle.com>
Mon, 13 Jan 2014 11:53:53 +0000 (19:53 +0800)
committerChris Mason <clm@fb.com>
Tue, 28 Jan 2014 21:20:35 +0000 (13:20 -0800)
commit1a4319cc3c495d5b6b8e41f4d4c73b950d54c2be
treefca5ffd2b1b88281063a288552e04c6864a12fa8
parent078025347c8ed43ff330e392476d8866ac1b297f
Btrfs: fix extent state leak on transaction abortion

When transaction is aborted, we fail to commit transaction, instead we do
cleanup work.  After that when we umount btrfs, we get to free fs roots' log
trees respectively, but that happens after we unpin extents, so those extents
pinned by freeing log trees will remain in memory and lead to the leak.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Chris Mason <clm@fb.com>
fs/btrfs/disk-io.c