Oopsing Athlong systems

From: Sebastian Mika (Sebastian.Mika@first.fraunhofer.de)
Date: Tue Jun 04 2002 - 11:57:10 EST


Hi,

I have a problem with several Athlon XP/MSI systems for quite a while now. For
some systems the problems almost completely disappeared with newer kernel
version or bios updates. However: not on all and not completely.

I would be very thankful if someone had at least a suggestion what to try
next. There seem to be some people on the web experiencing similar problems.
But the only answers they got were to improve cooling and/or power supply
which are already over-dimensional for our system in question.

More detailed problem and system descriptions follow below. Please ask
questions if you need to know more!

Thanks a lot,
Sebastian Mika.

Problem description:
==========================
* The system does not run stable neither with linux-2.4.[2,13,16,17] nor with
2.4.19-pre10.
* The system crash appears quite randomly, usually under heavier load but
sometimes also during boot up or if the system is (more or less) idle.
* The milder effects are a killing of a running process often with the
following or similar message in the system log:

Jun 4 17:05:05 calculon kernel: swap_dup: Bad swap file entry 00000020
Jun 4 17:05:05 calculon kernel: VM: killing process matlab
Jun 4 17:05:05 calculon kernel: swap_free: Bad swap file entry 00000020

Interestingly this also happens if there is **NO** swap space configured!
Afterwards the system is usually still usable although not very stable.
* Alternatively, the system produces an Oops message. I attached the output of
ksymoops for the two latest occurrence in OOPS.txt and OOPS-2.txt.
* The system seems to be slightly more stable when being operated at 100MHz
front side bus rather than 133MHz.
* This is the fifth system (all Athlon with MSI mainboard and VIA chipset)
having this sort of problem. The other systems are relatively stable now with
the 2.4.18 kernel. All systems have the latest BIOS updates from MSI.
* We did extensive testing to sort out hardware problems. Especially RAM and
CPU seem to be fine.
* SPECIAL: The system boots a local kernel and has local swap/tmp but the rest
of the OS is on a NFS server. However, 20 other systems do not have any
problem with absolutely identical software setup.

System
==========================
Processor: Athlon XP 1800+ (with very good cooling system)
Mainboard: MSI K7T266 Pro2 (VIA KT266 chipset)
RAM: 3*256 DDR-RAM CL2.5
Video: Elsa Gladiac 511 AGP Geforce2MX
Network: 3Com Corporation 3c905C-TX PCI
Power Supply: Enermax 350Watt
Disks/CD:
- hda: Maxtor 4G120J6, ATA DISK drive
- hdc: CD-950E/TKU, ATAPI CD/DVD-ROM drive
both on on-board VIA vt8233 UDMA 100 controller
Kernel: 2.4.19-pre10
Kernel-Boot-Options: "nfsroot=BLABLA disableapic" (disapleapic helped on a
similar system).
gcc version 2.95.2 19991024 (release)
binutils-2.10.0.33-24

-- 
Sebastian Mika             http://www.first.fraunhofer.de/~mika/
Fraunhofer - First
Kekulestr. 7                                      Tel: +49 (30) 6392 1906
12489 Berlin, Germany                   Fax: +49 (30) 6392 1805



- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Fri Jun 07 2002 - 22:00:21 EST