xfs: fix double unlock in xfs_swap_extents()
authorFelix Blyakher <felixb@sgi.com>
Fri, 8 May 2009 00:49:45 +0000 (19:49 -0500)
committerFelix Blyakher <felixb@sgi.com>
Tue, 2 Jun 2009 03:59:29 +0000 (22:59 -0500)
commit1f23920dbf1377fa9e4aef4f3d20c34a06a71a35
tree0fa80478499f1c42b2dc649659e1d1f22c6862f2
parent28e211700a81b0a934b6c7a4b8e7dda843634d2f
xfs: fix double unlock in xfs_swap_extents()

Regreesion from commit ef8f7fc, which rearranged the code in
xfs_swap_extents() leading to double unlock of xfs inode ilock.
That resulted in xfs_fsr deadlocking itself on platforms, which
don't handle double unlock of rw_semaphore nicely. It caused the
count go negative, which represents the write holder, without
really having one. ia64 is one of the platforms where deadlock
was easily reproduced and the fix was tested.

Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Felix Blyakher <felixb@sgi.com>
fs/xfs/xfs_dfrag.c