Arch Linux

Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines

Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.

REPEAT: Do NOT report bugs for outdated packages!
Tasklist

FS#3649 - Kernel panic after/during? loading Qlogic HBA Driver.

Attached to Project: Arch Linux
Opened by Eric Oden (1270) - Thursday, 22 December 2005, 18:06 GMT
Last edited by Tobias Powalowski (tpowa) - Saturday, 31 December 2005, 15:05 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To Judd Vinet (judd)
Architecture not specified
Severity High
Priority Normal
Reported Version 0.7 Wombat
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

To make a long story short I got a new mobo and CPU. I put them in and tried to boot the default kernel and got the error
"QLogic Fibre Channel HBA Driver
CPU 0: Machine Check Exception: 00000000000004(+- a few zeros)
Bank 4: b200000000070f0f(^)
Kernel panic - not syncing: CPU context corruption". After this I tried a custom kernel and that booted up without network. The custom kernel lacked all support for scsi.

Lacking the kernel source to compile a new kernel with the network module for my Ethernet I did a reinstalling using the 0.7.1-pre1 installer. It didn't work with "arch" but, using "arch-noscsi" the installer started up. After I did a reinstall I rebooted and got the same error using the 2.6.14-2 kernel from current.

Don't know if the following is helpful but, I'll add it anyways.
CPU: AMD Sempron 64 2600+
Mobo: Biostar K8NHA Grand with the nForce3 250Gb chipset
This task depends upon

Closed by  Roman Kyrylych (Romashka)
Wednesday, 03 January 2007, 21:29 GMT
Reason for closing:  None
Comment by Judd Vinet (judd) - Monday, 02 January 2006, 22:01 GMT
Hmmm. If you want, you could try a non-SMP kernel.

Here's one: http://www.archlinux.org/~judd/fixes/kernel26-nosmp-2.6.14.5-1.pkg.tar.gz
Comment by Eric Oden (1270) - Wednesday, 04 January 2006, 06:30 GMT
I will test the non-smp kernel. I did get the system up and running with a 2.6 series kernel. I installed arch with the 2.4 non-scsi kernel, rebooted, copied the kernel source from a CD, and compiled the kernel from source lacking all scsi support. It booted up fine. I've been meaning to do some testing, but the holiday season came up. I'll test the non-smp kernel after I get some work done on this site I have to code.
Comment by William Rea (willysilly) - Sunday, 29 January 2006, 02:47 GMT
I have the same issue and tried the nosmp kernel with no success
Comment by Aaron Griffin (phrakture) - Monday, 30 January 2006, 02:02 GMT
See here:
http://linux.derkeiler.com/Mailing-Lists/Kernel/2004-09/3860.html

Alan Cox reports it as a hardware error, and this tool detects MCE #4 as a PCI error - according to this guy, it's a parity thing. Try mucking with the BIOS. One google find reported this problem when setting some bios setting to 'Aggressive'.
Comment by Eric Oden (1270) - Wednesday, 01 February 2006, 23:06 GMT
That's an interesting read. I have gotten the default kernel working on my hardware. I read a forum post about mkinitrd and I set AUTODETECT to 1. Doing that makes it not load the SCSI drivers that cause the problems. Though I can get the system up and running it's really unstable(system locks and network dies). It seems like a hardware problem. So maybe it some bad component the Biostar and Matsonic mobos share?
Comment by arjan timmerman (blaasvis) - Saturday, 25 March 2006, 21:55 GMT
is this fixed with kernel 2.6.16 ?
Comment by William Rea (willysilly) - Saturday, 25 March 2006, 22:59 GMT
No, unfortunately it still does it with 2.6.16
Comment by Tobias Powalowski (tpowa) - Saturday, 01 April 2006, 10:16 GMT
have you ever tried to run a memtest? last bug with those errors was a broken ecc ram module
Comment by Tobias Powalowski (tpowa) - Saturday, 10 June 2006, 06:33 GMT
still valid bug with 2.6.17?
Comment by William Rea (willysilly) - Saturday, 10 June 2006, 06:34 GMT
Not sure with 2.6.17, but last week I fixed this problem by removing scsi from the initrd
Comment by Tobias Powalowski (tpowa) - Saturday, 10 June 2006, 07:54 GMT
have you tried mkinitpcio? does this fix the issue too?

Loading...