fs: always maintain i_dio_count
authorChristoph Hellwig <hch@infradead.org>
Fri, 24 Jun 2011 18:29:46 +0000 (14:29 -0400)
committerAl Viro <viro@zeniv.linux.org.uk>
Thu, 21 Jul 2011 00:47:48 +0000 (20:47 -0400)
commitdf2d6f26586f12a24f3ae5df4e236dc5c08d6eb4
tree68c6ec96177f766d3b9ab0a48408271ef2af4d89
parent562c72aa57c36b178eacc3500a0215651eca9429
fs: always maintain i_dio_count

Maintain i_dio_count for all filesystems, not just those using DIO_LOCKING.
This these filesystems to also protect truncate against direct I/O requests
by using common code.  Right now the only non-DIO_LOCKING filesystem that
appears to do so is XFS, which uses an opencoded variant of the i_dio_count
scheme.

Behaviour doesn't change for filesystems never calling inode_dio_wait.
For ext4 behaviour changes when using the dioread_nonlock option, which
previously was missing any protection between truncate and direct I/O reads.
For ocfs2 that handcrafted i_dio_count manipulations are replaced with
the common code now enable.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
fs/direct-io.c
fs/ocfs2/aops.c
fs/ocfs2/file.c