[PATCH 4/6] FS_NOATIME for ocfs

[PATCH 4/6] FS_NOATIME for ocfs

[Christoph Hellwig]

OCFS is only in -mm so far, so here's a separate patch for the
FS_NOATIME in ocfs in case this patchkit goes in before ocfs is merged.

Although I think ocfs should still go into 2.6.15 a an additional driver
that stands on it's own..


Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: linux-2.6.15-rc5/fs/ocfs2/super.c
===================================================================
--- linux-2.6.15-rc5.orig/fs/ocfs2/super.c	2005-12-12 18:51:17.000000000 +0100
+++ linux-2.6.15-rc5/fs/ocfs2/super.c	2005-12-13 11:14:40.000000000 +0100
@@ -677,7 +677,7 @@
 	.kill_sb        = kill_block_super, /* set to the generic one
 					     * right now, but do we
 					     * need to change that? */
-	.fs_flags       = FS_REQUIRES_DEV,
+	.fs_flags       = FS_REQUIRES_DEV | FS_NOATIME,
 	.next           = NULL
 };
 
@@ -1226,7 +1226,6 @@
 	sb->s_fs_info = osb;
 	sb->s_op = &ocfs2_sops;
 	sb->s_export_op = &ocfs2_export_ops;
-	sb->s_flags |= MS_NOATIME;
 	/* this is needed to support O_LARGEFILE */
 	sb->s_maxbytes = ocfs2_max_file_offset(sb->s_blocksize_bits);
 

Re: [PATCH 4/6] FS_NOATIME for ocfs

[Jeff Garzik]

Christoph Hellwig wrote:
> Although I think ocfs should still go into 2.6.15 a an additional driver
> that stands on it's own..

I ACK the inclusion of ocfs as well, but not in 2.6.15.

Standalone or not, we don't need to set an example of including new 
features at the last minute.  That just encourages other folks to submit 
new features at the last minute.  Which leads to lack of review, and 
unnecessary last minute work for some.  Do that enough times, and you'll 
be seeing reiser4 sent to Linus 24 hours before each -final release :)

	Jeff