xfs_getbmap assert

xfs_getbmap assert

[Christoph Hellwig]

With Dmitris fsstress updates I can hit the following assert fairly
regularly:

[11904.943956] XFS: Assertion failed: ((iflags & BMV_IF_DELALLOC) != 0) ||
(map[i].br_startblock != DELAYSTARTBLOCK), file:
/home/hch/work/linux-2.6/fs/xfs/xfs_bmap.c, line: 5604

which means we get an delalloc extent back from FIEMAP/GETBMAP despite
asking for a flush beforehand.  While we hold the iolock over the call
and thus exclude new buffered writers from appearing that doesn't
prevent shared writeable mmaps from creating new delalloc extents.

I don't think this actually is a real issue, and a workaround would
be extremly hard.  For now I've just remove the assert in my tree:


Index: xfs/fs/xfs/xfs_bmap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap.c	2011-11-05 09:02:24.410021945 +0100
+++ xfs/fs/xfs/xfs_bmap.c	2011-11-05 09:03:54.036494235 +0100
@@ -5600,8 +5600,6 @@ xfs_getbmap(
 				XFS_FSB_TO_BB(mp, map[i].br_blockcount);
 			out[cur_ext].bmv_unused1 = 0;
 			out[cur_ext].bmv_unused2 = 0;
-			ASSERT(((iflags & BMV_IF_DELALLOC) != 0) ||
-			      (map[i].br_startblock != DELAYSTARTBLOCK));
                         if (map[i].br_startblock == HOLESTARTBLOCK &&
 			    whichfork == XFS_ATTR_FORK) {
 				/* came to the end of attribute fork */

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

Re: xfs_getbmap assert

[Dave Chinner]

On Mon, Nov 07, 2011 at 05:30:37AM -0500, Christoph Hellwig wrote:
> With Dmitris fsstress updates I can hit the following assert fairly
> regularly:
> 
> [11904.943956] XFS: Assertion failed: ((iflags & BMV_IF_DELALLOC) != 0) ||
> (map[i].br_startblock != DELAYSTARTBLOCK), file:
> /home/hch/work/linux-2.6/fs/xfs/xfs_bmap.c, line: 5604
> 
> which means we get an delalloc extent back from FIEMAP/GETBMAP despite
> asking for a flush beforehand.  While we hold the iolock over the call
> and thus exclude new buffered writers from appearing that doesn't
> prevent shared writeable mmaps from creating new delalloc extents.

Possibly. However, we might actually be tripping over speculative
delalloc regions beyond EOF - flushing will not convert those at
all. There's already a comment earlier on like this:

	/*
	 * even after flushing the inode, there can still be delalloc
	 * blocks on the inode beyond EOF due to speculative
	 * preallocation. These are not removed until the release
	 * function is called or the inode is inactivated. Hence we
	 * cannot assert here that ip->i_delayed_blks == 0.
	 */

Can you see if that is the case that is being triggered?

> I don't think this actually is a real issue, and a workaround would
> be extremly hard.  For now I've just remove the assert in my tree:

Probably the right thing to do, anyway.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

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

Re: xfs_getbmap assert

[Christoph Hellwig]

On Tue, Nov 08, 2011 at 07:15:05AM +1100, Dave Chinner wrote:
> Can you see if that is the case that is being triggered?

I've tried to reproduce the issue with additional debugging for almost
a week now, but I haven't been able to hit the assert again yet.  Which
is strange given how often I managed to hit it just before.

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

Re: xfs_getbmap assert

[Dave Chinner]

On Mon, Nov 14, 2011 at 03:05:54PM -0500, Christoph Hellwig wrote:
> On Tue, Nov 08, 2011 at 07:15:05AM +1100, Dave Chinner wrote:
> > Can you see if that is the case that is being triggered?
> 
> I've tried to reproduce the issue with additional debugging for almost
> a week now, but I haven't been able to hit the assert again yet.  Which
> is strange given how often I managed to hit it just before.

Sounds exactly like the problems I've had trying to track this down
- as soon as you go looking for it, it disappears... :/

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

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

Re: xfs_getbmap assert

[Christoph Hellwig]

On Wed, Nov 16, 2011 at 07:02:46AM +1100, Dave Chinner wrote:
> On Mon, Nov 14, 2011 at 03:05:54PM -0500, Christoph Hellwig wrote:
> > On Tue, Nov 08, 2011 at 07:15:05AM +1100, Dave Chinner wrote:
> > > Can you see if that is the case that is being triggered?
> > 
> > I've tried to reproduce the issue with additional debugging for almost
> > a week now, but I haven't been able to hit the assert again yet.  Which
> > is strange given how often I managed to hit it just before.
> 
> Sounds exactly like the problems I've had trying to track this down
> - as soon as you go looking for it, it disappears... :/

I've actually reproduced it once today - of course that was running a
kernel without the debugging patch in it..

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