Re: [xfs-masters] Re: 2.6.18-rc6-mm2

[David Chinner <dgc@sgi.com>]

On Thu, Sep 14, 2006 at 10:50:42AM +0200, Michal Piotrowski wrote:
> On 14/09/06, Michal Piotrowski <michal.k.k.piotrowski@gmail.com> wrote:
> >David Chinner wrote:
> >> On Wed, Sep 13, 2006 at 11:43:32AM +0200, Michal Piotrowski wrote:
> >>> On 13/09/06, David Chinner <dgc@sgi.com> wrote:
> >>>> I've booted 2.6.18-rc6-mm2 and mounted and unmounted several xfs
> >>>> filesystems. I'm currently running xfsqa on it, and I haven't seen
> >>>> any failures on unmount yet.
> >>>>
> >>>> That test case would be really handy, Michal.
> >>> http://www.stardust.webpages.pl/files/mm/2.6.18-rc6-mm2/test_mount_fs.sh
> >>>
> >>> ls -hs /home/fs-farm/
> >>> total 3.6G
> >>> 513M ext2.img  513M ext4.img  513M reiser3.img  513M xfs.img
> >>> 513M ext3.img  513M jfs.img   513M reiser4.img
> >>
> >> Ok, so you're using loopback and mounting one of each filesystem, then
> >> unmounting them in the same order. I have mounted and unmounted an
> >> XFS filesystem in isolation in exactly the same way you have been, but
> >> I haven't seen any failures.
> >>
> >> Can you rerun the test with just XFS in your script and see if you
> >> see any failures? If you don't see any failures, can you add each
> >> filesystem back in one at a time until you see failures again?
> >
> >
> >I still get an oops (with xfs only). Maybe it's file system image problem.
> >
> >xfs_info /mnt/fs-farm/xfs/
> >meta-data=/dev/loop1             isize=256    agcount=8, agsize=16384 blks
> >         =                       sectsz=512
> >data     =                       bsize=4096   blocks=131072, imaxpct=25
> >         =                       sunit=0      swidth=0 blks, unwritten=1
> >naming   =version 2              bsize=4096
> >log      =internal               bsize=4096   blocks=1200, version=1
> >         =                       sectsz=512   sunit=0 blks
> >realtime =none                   extsz=65536  blocks=0, rtextents=0
> 
> Can I send to you this fs image? It's only 246KB bz2 file.

I've downloaded it, and I don't see a panic on that fs at all.
I've got it sitting in a tight loop mounting and unmounting the
image you sent me, and nothing has gone wrong. I don't think it's
a corrupted filesystem problem - it seems more like a memory corruption
problem to me.

What arch are you running on and what compiler are you using?
Can you try 2.6.18-rc6 and see if it panics like this on your
machine? there is little difference in xfs between -rc6 and -rc6-mm2
so it would be good to know if this is a problem isolated to
the -mm tree or not....

Cheers,

Dave.
-- 
Dave Chinner
Principal Engineer
SGI Australian Software Group