[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#600656: linux-image-2.6.32-5-amd64: Crash after nullpointer dereference during gparted reading a disk



On Mon, 2010-10-18 at 23:43 +0200, Andreas Feldner wrote:
> Package: linux-2.6
> Version: 2.6.32-23
> Severity: important
> 
> My system crashes reproducibly during the gparted operation "move /dev/sda1 to the right" in the read-only test stage.
> It is very hard to track down as the system freezes (ping is working, no other networking stuff like sshd, screen
> is black, keyboard doesn't react, not even the shift indicator LEDs). After reboot, no indication of the problem
> can be found in the log files anymore.
>
> Running tail -f /var/log/kern.log on a different machine, and running gparted with remote X11 did the trick to
> get to some information. The following lines were the last words to hear from the machine:
> 
> Oct 18 21:44:10 xxxxxxx kernel: [88819.864398] BUG: unable to handle kernel NULL pointer dereference at (null)
> Oct 18 21:44:10 xxxxxxx kernel: [88819.864413] IP: [<ffffffff8110c456>] drop_buffers+0x23/0x9d
> Oct 18 21:44:10 xxxxxxx kernel: [88819.864429] PGD 7d9c4067 PUD 1f542067 PMD 0 
> Oct 18 21:44:10 xxxxxxx kernel: [88819.864439] Oops: 0000 [#1] SMP 
> 
> gparted was attempting to read a 511.94 GiB sized partition on a SATA disk with a block size of 16.00 MiB. The
> crash orccured at 365.60 GiB (in case that matters). The file system on the partition in question is reported
> OK by fsck. A test with dd if=/dev/sda1 of=/dev/null bs=4096 worked out OK (didn't try with 16M block size). 
> SMART status if the hard disk is passed.
> 
> I suspect some rare freezes of the machine to come from the same origin, though I didn't find another way
> to actually reproduce the problem than running the described gparted operation.
> 
> Unfortunately, I have no idea if this bug report contains any useful information, nor how to provide any
> additional.
[...]

It really isn't enough information.  Please use netconsole or a serial
console to capture the full 'oops' message.  Also please test whether
this is reproducible if you don't use the nvidia proprietary driver.

Ben.

-- 
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: