From: Dave Chinner <redacted>

lockdep reports splice vs direct-io write lock inversions due to
generic_file_splice_write() taking the inode->i_mutex inside
XFS_IOLOCK_EXCL context. These lock contexts are inverted, hence can
deadlock.  Use generic_file_splice_write_unlocked() because the
i_mutex does not need to be held over the operations that are done
in the XFS splice write path.

Signed-off-by: Dave Chinner <redacted>
---
 fs/xfs/linux-2.6/xfs_file.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c
index f51a384..1e641e6 100644
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ b/fs/xfs/linux-2.6/xfs_file.c

@@ -463,7 +463,8 @@ xfs_file_splice_write(
 
 	trace_xfs_file_splice_write(ip, count, *ppos, ioflags);
 
-	ret = generic_file_splice_write(pipe, outfilp, ppos, count, flags);
+	ret = generic_file_splice_write_unlocked(pipe, outfilp, ppos,
+						count, flags);
 
 	xfs_aio_write_isize_update(inode, ppos, ret);
 	xfs_aio_write_newsize_update(ip);

-- 
1.7.5.1

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs