I have been getting recurring kernel oopses on a machine I recently put together for my parents. I have went through a number of avenues trying to run down the problem but am now somewhat stumped.

The machine in question uses a Gigabyte GA-7VA motherboard- KT400 based- with 512MB RAM and a 1.4GHz Athlon Thunderbird. The hard drive is a Western Digital, 20GB WD200BB-75CAA0. It is currently running SuSE 9 Pro, kernel 2.4.21-166-athlon, but I have had similar results with a Gentoo 1.4 live disk.

I first became aware of the problem when the machine suffered a hard lockup followed by a horribly mauled reiserfs /home filesystem after merely trying to start kmail. Since then I have managed to provoke a number of oopses mostly by stressing the disk.

I have tested the memory with the suse supplied memtest program and have interrogated the drive's SMART system and done a full badblocks test without finding anything. Although while running a badblocks -vw while running off of the Gentoo bootdisk I did experience an oops.

Worried that the system was running too hot I have fitted a better copper heatsink and added a second chassis fan and the system seems to be running fairly stable.

However, in an attempt to provoke the it I decided to run a few bonnie tests. Running bonnie -s 1000 immediately provoked the following oops, though without bringing down anything else.

kernel BUG at page_alloc.c:263!
invalid operand: 0000 2.4.21-166-athlon #1 Thu Dec 18 18:24:05 UTC 2003
CPU: 0
EIP: 0010:[<c0140690>] Tainted: P
EFLAGS: 00010082
eax: 00000009 ebx: 00000004 ecx: 0001f77c edx: c0304e58
esi: c0304ea8 edi: 00000002 ebp: c15e6760 esp: cd671e50
ds: 0018 es: 0018 ss: 0018
Process bonnie (pid: 2122, stackpage=cd671000)
Stack: 00001000 c01363fb 0001e77c 0001e778 00000286 00000000 c0304e58 c0304e58
c030502c 00000001 cdf754a8 c0140992 00001000 0003e800 cd673700 df2fa440
cd6fb7c0 c15ef040 c0304e58 c0305028 00000000 000001d2 00000001 cd6737b4
Call Trace: [<c01363fb>] (40) [<c0140992>] (64) [<c01362cb>] (08)
[<dfd88e90>] (32) [<c01369a7>] (28) [<c0136e35>] (44) [<c01371d0>] (32)
[<c01374a8>] (16) [<c01371d0>] (56) [<c01489d3>] (44) [<c010904b>] (60)
Modules: [(reiserfs:<dfd80060>:<dfdb5094>)]
Code: 0f 0b 07 01 26 6d 2c c0 83 ee 0c 4f d1 eb 8b 06 8d 4f 01 89

I'm about out of ideas, any help greatly appreciated.

Ian