Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 7180 - APIC errors on a GIGABYTE 5DX Dual-P1/MMX system...
Summary: APIC errors on a GIGABYTE 5DX Dual-P1/MMX system...
Status: RESOLVED FIXED
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Core system (show other bugs)
Hardware: x86 Linux
: High normal (vote)
Assignee: Brandon Low (RETIRED)
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2002-08-28 14:17 UTC by Ralf
Modified: 2003-02-04 19:42 UTC (History)
2 users (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Ralf 2002-08-28 14:17:57 UTC
The header tells it all: on my dual-P233MMX system the error messages get 
annoying: "APIC Error on CPU0" (or CPU1), and on one heavy system build the 
system crashed.

I GOOGLEd around just to find that this APIC errors seem to be caused by a buggy 
SMP handling of my board's BIOS. My old slackware & kernel 2.2.something never 
threw this at me, but now... gentoo does :-) (maybe because of the 2.4.19 kernel 
I use)

Just to list the suggested (non working) solutions found elsewhere: 
1) set SMP protocol version to 1.1 (no chance, no BIOS switch to do that)
2) get the newest BIOS for this board (I have the latest - V3.43 from 1997)

Has anyone a clue what I could do to get around this trouble? Any switches to 
try at boot time, secret rituals to perform at the terminal, patches?

I'll try linuxbios.org after this bug report - maybe they have a completely new 
BIOS for this old board. Or I could get an older 2.2 kernel as mentioned in the 
INSTALL GUIDE. Anyway - this is my last hope before wrecking the board :-(
Comment 1 Hanno Böck gentoo-dev 2002-08-28 14:46:36 UTC
Well, it seems to be a kernel-issue.
Alan Cox writes something in his changelog.
o	Hopefully fix the smp boot/apic problem		(James Cleverdon)
http://marc.theaimsgroup.com/?l=linux-kernel&m=103035849204944&w=2

This one is not merged into the main tree until now. You should try an ac-kernel
and see if this fixes the issue.
Comment 2 Brandon Low (RETIRED) gentoo-dev 2002-09-14 14:07:35 UTC
what kernel specifically were you using?

have you tried the latest gentoo-r9 kernel?
Comment 3 Ralf 2002-09-14 16:40:34 UTC
Thanks for the suggestion to try out the AC kernel - as this is not too urgent 
(2 other boxes running gentoo without trouble), I'm going to wait for the 2.4.20 
release.

The kernel I was using was the plain vanilla 2.4.19. I'm going to wait, update, 
re-test and eventually set this bug to FIXED, when 2.4.20 (or up) will cure 
this. I seem the only one with this antique board, anyway :-)

Thanks again - it feels fine to have helping hands around.

Cheers,

Ralf
Comment 4 Hanno Böck gentoo-dev 2002-12-06 07:06:04 UTC
To the reporter: Is this still a problem or can we close this bug?
Comment 5 Ralf 2002-12-07 11:15:34 UTC
Well, I installed 2.4.20 and still have spurious "APIC error on CPU1: 04(00)" or 
"APIC error on CPU0: 02(00)".

Hmmm, how could I apply the AC patches to the kernel? Sorry for this stupid 
question, but I had been patching my way through the 0.x up to 2.2.x kernels - 
and never applied one of the famous AC changes :-) Will I have to rename 
(symlink) my current 2.4.20 directory to linux.vanilla? If anybody cares, please 
send me a pointer to a FAQ, please.

Ho, Ho, Ho,

Ralf
Comment 6 Brandon Low (RETIRED) gentoo-dev 2002-12-07 14:24:13 UTC
emerge ac-sources
we have an ebuild for it to save you the trouble :)
Comment 7 Ralf 2002-12-29 06:29:34 UTC
Hi there,

switching the system to kernel 2.4.20-ac2 seems to have cured the annoying APIC 
bug. I was busy updating my system to gentoo 1.4 (took over one week up 
to now... made me nearly jump out of my window, lucky me - my shack is in the 
basement; gcc 3.2.1 from the 1.4rc package wouldn't compile; this ruined my 
stage1 and stage2 setup attempts. Now I'm using a stage3 tarball (rc2) and will 
tweak the system to optimize it for Dual-233MMX later :-)
Will run some stress tests with heavy CPU load (SETI and/or RC5-72) and I/O over 
the SCSI i/f (DAT copying etc) - but from my POV, this bug is GONE.
Thanks for the help,

Ralf

BTW: happy new year!