2.6.39 - crash when plugging in a usb memory stick

Hi,

I've just had the following crash after plugging in a usb memory stick. Some diagnostic output may be missing because it was scrolled off screen by a subsequent oops, which I have assumed was induced by this crash and not, therefore, copied down. Also, I skipped writing down the modules linked because it's a long list, but there are no out-of-tree modules of any kind and the kernel was not tainted (until after this crash).

last sysfs file: /sys/devices/virtual/bdi/8:32/uevent
Modules linked in ...
PID: 3, comm: ksoftirqd/0 Not tainted 2.6.39 #1 Gigabyte Techmology Co., Ltd EG41MF-US2H/EG41MF-US2H EIP is at 0x0
EAX: f4500e20 EBX:f4500e20 ECX:ebc88c60 EDX: 00000000
ESI: f4500e20 EDI: c9c9fc00 EBP: ebc2d420 ESP: f4c4be80
DS: 007b ES: 007b FS: 00d8 GS:0000 SS: 0068
Process ksoftirqd/0 (pid 3, ti=f4c4a000 task=f4c44720 task.ti=f4c4a000)

Stack:
 c114d4ec f4500e20 c9c9fc00 ebc2d420 00000000 f4500e20 c9c9fc00 c120abce
 c141e188 c120a944 e2d13d9c fffffffb c102582f c6258840 f4500e20 00000296
 c9c9fc00 e6c2d420 c114ada8 ebc2d400 f4c4beec c12094de cb2580c0 00000246
Call trace:

c114d4ec blk_peek_request+oxfc/0x150
c120abce scsi_request_fn+ox30e.0x370
c120a944 scsi_request_fn+0x84/0x370
c102582f complete+0x3f/-x60
c114ada8 blk_run_queue+0x28/0x40
c12094de scsi_run_queue+0xae/0x230
c120aef0 scsi_next_command+0x30/0x50
c120b0a0 scsi_io_completion+0x150/0x590
c1151722 blk_done_softirq+0x52/0x60
c1037f85 __do_soft_irq+0x75/0xf0
c1075b7f rcu_note_context_switch+0x6f/0x150
c10380b8 run_softirqd+0xb8/0x1a0
c1038000 __do_softirq+0xf0/0xf0
c104b104 kthread+0x74/0x80
c1048090 kthread_worker_fn+0x170/0x170
c13623f6 kernel_thread_helper+0x6/0xd

Code: Bad EIP value

EIP: [<00000000>] 0x0 SS:ESP 0068:f4c4be80
CR2: 00000000000000000

I've been using 2.6.39 since the day it was released, plus -rc version since rc2 and this is the first crash I have experienced. Consequently, recreating the crash might not be possible. Nevertheless, let me know if there's any way I can help to fix this.

Chris

The more I see, the more I know. The more I know, the less I understand. Changing Man - Paul Weller

Hi again,

Peter Alfredsen seems to be experiencing exactly the same crash. At least the call trace appears identical.
http://marc.info/?l=linux-kernel&m=130662291207794&w=4

The more I see, the more I know. The more I know, the less I understand. Changing Man - Paul Weller