probably NFS related Oops during shutdown with 2.6.12-rc3-mm3

probably NFS related Oops during shutdown with 2.6.12-rc3-mm3

[Christian Kujau]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

hi,

i noticed that i get an Oops during shutdown and it says something about
rpciod/0 and i do have NFSv3 volumes mounted (and thus unmounted on
shutdown), full log, dmesg, .config here:

    http://nerdbynature.de/bits/prinz/2.6.12-rc3-mm3/

CPU:    0
EIP:    0060:[<00000000>]    Not tainted VLI
EFLAGS: 00010282   (2.6.12-rc3-mm3)
EIP is at _stext+0x3feffdd8/0x8
eax: c1628ec0   ebx: c1628ec0   ecx: 00000000   edx: dfa600b0
esi: 00000000   edi: c1628f34   ebp: de9327c0   esp: de413f24
ds: 007b   es: 007b   ss: 0068
Process rpciod/0 (pid: 8245, threadinfo=de412000 task=dfa600b0)

Stack:
c039942b
dfa600b0
dfa601d8
00000292
c1628f3c
00000297
c1628f40
c012b31e


00000000
00000000
dffc4550
de412000
de9327d8
de9327c8
de9327d0
de412000


c1628ec0
c0399560
de412000
ffffffff
ffffffff
00000001
00000000
c0117cf0


Call Trace:
[<c039942b>] __rpc_execute+0x14b/0x250
[<c012b31e>] worker_thread+0x1ae/0x280
[<c0399560>] rpc_async_schedule+0x0/0x10
[<c0117cf0>] default_wake_function+0x0/0x10
[<c0117d37>] __wake_up_common+0x37/0x60
[<c0117cf0>] default_wake_function+0x0/0x10
[<c012b170>] worker_thread+0x0/0x280
[<c012f615>] kthread+0x95/0xd0
[<c012f580>] kthread+0x0/0xd0
[<c010136d>] kernel_thread_helper+0x5/0x18


thank you,
Christian.

PS: i'm booting with netconsole which is *really* great for such things,
because the oops did not make it to the disk, as the local-filesystems are
already unmonted (shutdown!). but: the output is really odd. for example
the first line after the Call Trace looked like this:

Call Trace:
[<c039942b>]
 __rpc_execute+0x14b/0x250


ok, easy - just delete the linewrap. but i really wonder if the Stack:
above is readable at all (i haven't touched it)

- --
BOFH excuse #69:

knot in cables caused data stream to become twisted and kinked
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFChpHj+A7rjkF8z0wRAiSwAJ9LwFJT268TJMkNDEQ0asxMsxPfeACeKJzl
0aRaGxGJXfEsWBgjMJVQrHI=
=Vz3W
-----END PGP SIGNATURE-----

Re: probably NFS related Oops during shutdown with 2.6.12-rc3-mm3

[Trond Myklebust]

su den 15.05.2005 Klokka 02:03 (+0200) skreiv Christian Kujau:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> hi,
> 
> i noticed that i get an Oops during shutdown and it says something about
> rpciod/0 and i do have NFSv3 volumes mounted (and thus unmounted on
> shutdown), full log, dmesg, .config here:
> 
>     http://nerdbynature.de/bits/prinz/2.6.12-rc3-mm3/
> 
> CPU:    0
> EIP:    0060:[<00000000>]    Not tainted VLI
> EFLAGS: 00010282   (2.6.12-rc3-mm3)
> EIP is at _stext+0x3feffdd8/0x8
> eax: c1628ec0   ebx: c1628ec0   ecx: 00000000   edx: dfa600b0
> esi: 00000000   edi: c1628f34   ebp: de9327c0   esp: de413f24
> ds: 007b   es: 007b   ss: 0068
> Process rpciod/0 (pid: 8245, threadinfo=de412000 task=dfa600b0)
> 
> Stack:
> c039942b
> dfa600b0
> dfa601d8
> 00000292
> c1628f3c
> 00000297
> c1628f40
> c012b31e
> 
> 
> 00000000
> 00000000
> dffc4550
> de412000
> de9327d8
> de9327c8
> de9327d0
> de412000
> 
> 
> c1628ec0
> c0399560
> de412000
> ffffffff
> ffffffff
> 00000001
> 00000000
> c0117cf0
> 
> 
> Call Trace:
> [<c039942b>] __rpc_execute+0x14b/0x250
> [<c012b31e>] worker_thread+0x1ae/0x280
> [<c0399560>] rpc_async_schedule+0x0/0x10
> [<c0117cf0>] default_wake_function+0x0/0x10
> [<c0117d37>] __wake_up_common+0x37/0x60
> [<c0117cf0>] default_wake_function+0x0/0x10
> [<c012b170>] worker_thread+0x0/0x280
> [<c012f615>] kthread+0x95/0xd0
> [<c012f580>] kthread+0x0/0xd0
> [<c010136d>] kernel_thread_helper+0x5/0x18


That should already have been fixed in 2.6.12-rc4-mm1.

Cheers,
   Trond

Re: probably NFS related Oops during shutdown with 2.6.12-rc3-mm3

[Christian]

On Sun, May 15, 2005 2:26, Trond Myklebust said:
>
> That should already have been fixed in 2.6.12-rc4-mm1.
>

minutes after sending the report, i saw that 2.6.12-rc4-mm1 was out
already. compiled, booted, but today's shutdown produced the same error
:-(

just in case i was unclear: the shutdown completes, the oops shows "only"
up on the tty, no data corruption or the like so far.

are there any fancy debug flags i could set, to provide more information?

thank,
Christian.

-- 
make bzImage, not war

Re: probably NFS related Oops during shutdown with 2.6.12-rc3-mm3

[Christian Kujau]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Maneesh Soni wrote:
> Could you try configuring kdump and get a crashdump when the system 
> crashes(oops). Please refer to Documentation/kdump.txt in -mm kernel tree.
> You might also want to set /proc/sys/kernel/panic_on_oops to 1, so as to
> initiate kernel panic during oops.

i've done so, but..hmm...no panic was triggered, because the Oops is gone.
but it *was* there, i swear ;-)

http://nerdbynature.de/bits/prinz/2.6.12-rc3-mm3/prinz-nc_2.6.12-rc4-mm1.log
(but this time with a tainted kernel)

now i can boot 2.6.12-rc4-mm1 (tainted/not tainted), but the oops does not
show up anymore.

but thanks for the hint, i'll keep that in mind next time [1]

Christian.

[1] when will it be? even the -mm kernel is pretty rock-solid for me.
    great job!
- --
BOFH excuse #415:

Maintenance window broken
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCiJhY+A7rjkF8z0wRAqJbAJ9zIhAoIGfzLqrNgewmgqR7FafdSQCeOQMM
U6VPMiAOlofUVPl64lf7HdM=
=T0dN
-----END PGP SIGNATURE-----