[PATCH] xfs: do not discard page cache data on EAGAIN

[PATCH] xfs: do not discard page cache data on EAGAIN

[Christoph Hellwig]

If xfs_map_blocks returns EAGAIN because of lock contention we must redirty
the page and not disard the pagecache content and return an error from
writepage.  We used to do this correctly, but the logic got lost during
the recent reshuffle of the writepage code.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reported-by: Mike Gao <ygao.linux@gmail.com> 
Tested-by: Mike Gao <ygao.linux@gmail.com> 

Index: xfs/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.c	2010-08-18 18:10:48.140261091 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_aops.c	2010-08-18 18:12:41.167005888 +0200
@@ -1068,7 +1068,7 @@ xfs_vm_writepage(
 	 * by themselves.
 	 */
 	if ((current->flags & (PF_MEMALLOC|PF_KSWAPD)) == PF_MEMALLOC)
-		goto out_fail;
+		goto redirty;
 
 	/*
 	 * We need a transaction if there are delalloc or unwritten buffers
@@ -1080,7 +1080,7 @@ xfs_vm_writepage(
 	 */
 	xfs_count_page_state(page, &delalloc, &unwritten);
 	if ((current->flags & PF_FSTRANS) && (delalloc || unwritten))
-		goto out_fail;
+		goto redirty;
 
 	/* Is this page beyond the end of the file? */
 	offset = i_size_read(inode);
@@ -1245,12 +1245,15 @@ error:
 	if (iohead)
 		xfs_cancel_ioend(iohead);
 
+	if (err == -EAGAIN)
+		goto redirty;
+
 	xfs_aops_discard_page(page);
 	ClearPageUptodate(page);
 	unlock_page(page);
 	return err;
 
-out_fail:
+redirty:
 	redirty_page_for_writepage(wbc, page);
 	unlock_page(page);
 	return 0;

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

Re: [PATCH] xfs: do not discard page cache data on EAGAIN

[Dave Chinner]

On Mon, Aug 23, 2010 at 05:10:15AM -0400, Christoph Hellwig wrote:
> If xfs_map_blocks returns EAGAIN because of lock contention we must redirty
> the page and not disard the pagecache content and return an error from
> writepage.  We used to do this correctly, but the logic got lost during
> the recent reshuffle of the writepage code.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> Reported-by: Mike Gao <ygao.linux@gmail.com> 
> Tested-by: Mike Gao <ygao.linux@gmail.com> 

Looks good.

Reviewed-by: Dave Chinner <dchinner@redhat.com>

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

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