Horrible Kernel Oops

Thetargos · 08-12-2003, 11:25 AM

I've compiled a 2.4.21 kernel with Con Kolivas' patch (version 3). Now the problem is this: Whenever I start the computer (yes I actually shut it off when I'm done working with it

), and enable the anacrond service, an error message briefly appears. However when the shut down sequence starts, I get this Oops from the kernel:

Code:

Shutting down daemon APM       [   OK   ]
Unable to handle kernel paging request at virtual adrress e1631c9
Printing EIP:
*pde  =   01677067
*pte   =   00000000
Oops:     0000
CPU:      0
EIP:        0010: [ <e1a631c9> ]      Tainted P
EFLAGS: 00010046
eax:00005305     ebx:00000000     ecx:00000000
edx:00000000      esi:00000000     edi:00000292
ebp:00000001     esp:c014df84
ds:0000    es:0000     ss:0018
Process swapper (pid: 0, stackpage = c014d000
Stack: 00000001     00000292     00000018     00005305
          c0140000     00000000     00c76708     c014c000
          0000000f      00000001     e1|a63377    00005305
          00000000     00000000     c014dfc4      00005305
          e1a634ac     c014c000     c0153000     0008c000
          c01c11d4     00000000
Call tree:   [ <c01c1160> ]   [ <c01c11c4> ]
Code:  Bad EIP value
<0> Kernel Panic: Attpemted to kill ide task!
In idle task - not syncing
Saving random seed                  [   OK   ]

What does this mean?? also what could have caused it, did I compiled into the kernel a faulty feature??

As refference, the official Red Hat kernel that I'm using ATM does not have this problem.
HELP

jailbait · 08-12-2003, 01:39 PM

"What does this mean??"

It means that you have hit a kernel bug, most likely in the ide routine.

"also what could have caused it, did I compiled into the kernel a faulty feature??"

The bug could be the responsibility of either the kernel developers or Con Kolivas. The Con Kolivas patch contains code to speed up swap and code to speed up read latency so the patch is high on the suspect list.

The thing to do is to compile another kernel identical to this one except without the patch. Then by comparing the way the kernels work you can know whether to report the bug to the kernel developers or to Con Kolivas.

Thetargos · 08-13-2003, 12:47 AM

Tnak you for your reply, ajilbait. I have considered that same issue... It's a pitty that this version of the patch has this problem I had a running 2.4.21-ck1 kernel (fast and steady, same config) until this unfortunate bug storm. I have checked several messages both at boot (and from the logs) and these kind of problems are recurrent... I'll have to compile a plain vanilla kernel (I can't go back to the CK patch 1 'cause I lost it

).