Hi Nathan,
On 11/08/06, Nathan Scott <nathans@xxxxxxx> wrote:
On Fri, Aug 11, 2006 at 01:08:44AM +0200, Michal Piotrowski wrote:
> On 10/08/06, Michal Piotrowski <michal.k.k.piotrowski@xxxxxxxxx> wrote:
> > ...
> > Aug 10 16:50:53 ltg01-fedora kernel: EFLAGS: 00010002 (2.6.18-rc4 #119)
> > Aug 10 16:50:53 ltg01-fedora kernel: EIP is at __lock_acquire+0x2ca/0x9b6
> > Aug 10 16:50:53 ltg01-fedora kernel: Call Trace:
> > Aug 10 16:50:53 ltg01-fedora kernel: [<c0135ccb>]
> > lock_release_non_nested+0xc5/0x11d
> > Aug 10 16:50:53 ltg01-fedora kernel: [<c0136026>] lock_release+0x12f/0x159
> > Aug 10 16:50:53 ltg01-fedora kernel: [<c02d579e>]
> > __mutex_unlock_slowpath+0x99/0xff
> > Aug 10 16:50:53 ltg01-fedora kernel: [<c02d580c>] mutex_unlock+0x8/0xa
> > Aug 10 16:50:53 ltg01-fedora kernel: [<c016d3b2>]
> > generic_shutdown_super+0xbb/0x113
> > Aug 10 16:50:53 ltg01-fedora kernel: [<c016d42a>]
kill_block_super+0x20/0x32
> > Aug 10 16:50:53 ltg01-fedora kernel: [<c016d4ea>]
deactivate_super+0x5d/0x6f
> > Aug 10 16:50:53 ltg01-fedora kernel: [<c0180336>]
mntput_no_expire+0x42/0x72
> > Aug 10 16:50:53 ltg01-fedora kernel: [<c0173147>]
> > path_release_on_umount+0x15/0x18
> > Aug 10 16:50:53 ltg01-fedora kernel: [<c018147c>] sys_umount+0x1e7/0x21b
> ...
> It's fixed in latest 2.6.18-rc4-git*.
We've not changed anything mount related in current git* kernels,
are you sure its fixed (did it happen every time)? I'm not sure
how its XFS related too ... looks possibly lockdep related (from
*non_nested back there?)
You are absolutely right. Oops appears only with CONFIG_DEBUG_LOCKDEP enabled.
and perhaps an issue with sb->s_lock in
the kill_block_super unlock_super call?
cheers.
--
Nathan
Regards,
Michal
--
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
|