Home » Mailing list archives » linux-kernel » 2010 » October » 24

OOPs on loop module removal

Previous thread: [PATCH] mfd: ab3550: fix error check by Nicolas Kaiser on Sunday, October 24, 2010 - 11:57 am. (1 message)

Next thread: Question about Kernel Headers by Lucas Philippidis on Sunday, October 24, 2010 - 12:57 pm. (1 message)
[view original message]
From: Milan Broz
Subject: OOPs on loop module removal
Date: Sunday, October 24, 2010 - 12:14 pm

Hi,
Simple modprobe loop ; rmmod loop
on Linus' tree now produces now OOPS.

Bisected to patch (and reverting fixes problem)

commit 7681bfeeccff5efa9eb29bf09249a3c400b15327
Author: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Date:   Tue Oct 19 09:05:00 2010 +0200

    block: fix accounting bug on cross partition merges

The problem is probably that patch uses directly q->queue_lock
even if queue is being destroyed (QUEUE_FLAG_DEAD is set).
See what loop (and not only this code) does:

        blk_cleanup_queue(lo->lo_queue);
	put_disk(lo->lo_disk);

Milan


OOPS:
[   40.778161] BUG: unable to handle kernel NULL pointer dereference at (null)
[   40.779104] IP: [<c11bf2e5>] do_raw_spin_trylock+0x5/0x37
[   40.779843] *pde = 00000000 
[   40.780113] Oops: 0000 [#1] PREEMPT SMP 
[   40.780410] last sysfs file: /sys/devices/virtual/block/loop7/removable
[   40.780876] Modules linked in: loop(-) usbcore dm_mod
[   40.781442] 
[   40.781740] Pid: 3781, comm: rmmod Not tainted 2.6.36+ #85 440BX Desktop Reference Platform/VMware Virtual Platform
[   40.782177] EIP: 0060:[<c11bf2e5>] EFLAGS: 00010082 CPU: 1
[   40.782437] EIP is at do_raw_spin_trylock+0x5/0x37
[   40.782667] EAX: 00000000 EBX: 00000000 ECX: 00000000 EDX: 00000000
[   40.782906] ESI: 00000010 EDI: cee60800 EBP: cef07ea4 ESP: cef07ea4
[   40.804954]  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
[   40.805080] Process rmmod (pid: 3781, ti=cef06000 task=cf5fe070 task.ti=cef06000)
[   40.805254] Stack:
[   40.805343]  cef07ec4 c130ee92 00000000 00000002 00000000 c11acf48 ce1412e0 ced14070
[   40.805521] <0> cef07ed4 c11acf48 cec08c00 c14a1a50 cef07ee0 c11acfbd ce1f6d40 cef07eec
[   40.805747] <0> c122d14c cec08c58 cef07f00 c11b5982 cec08c74 c11b5942 ce14fe00 cef07f10
[   40.807338] Call Trace:
[   40.807576]  [<c130ee92>] ? _raw_spin_lock_irq+0x35/0x5e
[   40.807681]  [<c11acf48>] ? disk_replace_part_tbl+0x27/0x49
[   40.807771]  [<c11acf48>] ? disk_replace_part_tbl+0x27/0x49
[   40.807858]  ...
[ message continues ]
[view original message]
From: Jens Axboe
Subject: Re: OOPs on loop module removal
Date: Sunday, October 24, 2010 - 1:07 pm

Yep that commit was not the most fortunate one. I will send out a revert
now, which will gives us time to fix this for real without rushing
things.

-- 
Jens Axboe

--


[ message continues ]
Previous thread: [PATCH] mfd: ab3550: fix error check by Nicolas Kaiser on Sunday, October 24, 2010 - 11:57 am. (1 message)

Next thread: Question about Kernel Headers by Lucas Philippidis on Sunday, October 24, 2010 - 12:57 pm. (1 message)