Date: Mon, 09 Oct 2006 17:52:48 -0400
From: Steve Dickson <SteveD@redhat.com>
Subject: [RHEL5/FC6] [PATCH] NFS data corruption

The attached patch fixes RHEL5 bz210071 and FC6 bz209420
which are both blockers for each relative release...
(Note for the FC6 version of this patch, please grab it
from the bz because it will apply cleaner)....

The patch is pretty self explanatory, but for NFS to
invalidate its pages correctly, the gfp_mask has to
have at least the __GFP_FS bit set (see nfs_release_page()).

The path was tested by me and a couple upstream people...

steved.



If try_to_release_page() is called with a zero gfp mask, then the
filesystem is effectively denied the possibility of sleeping while
attempting to release the page.  There doesn't appear to be any valid
reason why this should be banned, given that we're not calling this from a
memory allocation context.

For this reason, change the gfp_mask argument of the call to GFP_KERNEL.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Steve Dickson <SteveD@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
---

 mm/truncate.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--- linux-2.6.18.i686/mm/truncate.c.orig	2006-10-09 15:05:42.102684000 -0400
+++ linux-2.6.18.i686/mm/truncate.c	2006-10-09 16:54:20.026163000 -0400
@@ -287,7 +287,7 @@ invalidate_complete_page2(struct address
 	if (page->mapping != mapping)
 		return 0;
 
-	if (PagePrivate(page) && !try_to_release_page(page, 0))
+	if (PagePrivate(page) && !try_to_release_page(page, GFP_KERNEL))
 		return 0;
 
 	write_lock_irq(&mapping->tree_lock);