Random lockups with new hardware.

supated · Posted: Sun Feb 15, 2004 7:58 am Post subject: Random lockups with new hardware.

Hey all,

I've previously been running Gentoo for a few months stably on an AthlonXP 1800, MSI KT333 mobo and 256mb PC2400, Geforce2MX400 box. However, recently I've upgraded to an AthlonXP 2500, ASUS A7V8X-X Nforce II mobo, and 512mb of PC2400 and a GeforceFX5200. I'm not using the onboard sound or NIC.
I've installed the latest 2.6.3 mm-sources Kernel with this configuration and have it running fine, except for the fact that it's locked up 4 times today. Previously running 2.4 with this configuration there were also lockups though there didnt seem to be as many. Running with my old hardware configuration I dont think I can remember it locking up once.
I've also read on this board that there are issues with the Nforce II, similar or the same as those I'm describing, though I havent really found any information on how to fix them.

Any assistance is much appreciated.
Thanks,
SuPaTeD

dsd · Posted: Sun Feb 15, 2004 12:51 pm Post subject:

hi

1. take the patch from this thread (bear in mind that it is repeated twice, you only want it once). there is a direct download link, here

save this patch as: /usr/src/linux/nforce-fix.patch

2. patch the kernel with:

supated · Posted: Mon Feb 16, 2004 5:46 am Post subject:

Thank you so much for the prompt and insightful response, dsd - the Gentoo community surely seems like a fantastic one! While I did encounter an error while trying to make bzImage, a quick recompile with genkernel hopefully did the trick. With any luck I'll now be enjoying "freeze-less operation".
Thanks again,
Ted

m2 · Posted: Mon Feb 16, 2004 8:57 am Post subject:

I get an error...

Hunk #4 FAILED at 210.

then make doesn't work..

/arc/i386/kernel/process.c:225: 'idle' undeclared (first use in this function)
..etc

help.

dsd · Posted: Mon Feb 16, 2004 10:09 am Post subject:

supated, you shouldnt have encountered an error, did it stop the entire compile process?

m2, which kernel are you running?
_________________
http://dev.gentoo.org/~dsd

m2 · Posted: Mon Feb 16, 2004 11:26 am Post subject:

2.6.1 r1otherwise known as gentoo-dev-sources

trying some peoples other tips atm

i.e. unchecked local uniprocessor APIC controller in menuconfig.

stable so far, but i've had it up longer than this normally (up for a few hours, normally stays up for a day or so before freezing while using HDD's hard). But i'm hammering it atm, so i'll see what happens

supated · Posted: Mon Feb 16, 2004 11:55 am Post subject:

dsd, yeah it did. I mean it all patched okay and to my knowledge it's still stayed patched and genkernel made the bzImage fine, but yeah, manually making the bzImage crashed it.

Here is the output I got:

dsd · Posted: Mon Feb 16, 2004 2:09 pm Post subject:

m2: the patch is designed to apply against mm-sources, which is what I use. yes, disabling APIC seems to solve the problem for some people (myself included).

supated: strange, i think running "make clean" before making bzImage again may have fixed that. genkernel probably did that.
is the system running ok now?
_________________
http://dev.gentoo.org/~dsd

supated · Posted: Mon Feb 16, 2004 9:02 pm Post subject:

dsd, Gaah, it seems not - woke up today to a black screen that the comp was unable to wake from :'(

PrakashP · Posted: Mon Feb 16, 2004 11:00 pm Post subject:

I am currently trying to track it down, whether it is a kernel issue or APIC related.

Findings so far: Using Ross' patches seems to stabilize the machine, but unfortunately not 100%. kernels tried:
2.6.3-mm2 (and mm1)
2.6.2-rc2
2.6.2-rc1-bk1
2.6.2-rc1
2.6.1

ALL lock up in APIC mode, disconnect enabled and Ross' patches used. Either I get a hard lock-up (just as i am experiencing in Windows), where everything freezes and after a while the screen gets blanked or I get an inifinite repeating oops of nvidia binary with badness in pci-search.c (or alike). Mouse cursor still moves, but nothing reacts. In this state I can still reach the PC from LAN and reboot it gracefully.

Currently I have tested 2.6.1 and 2.6.2-rc1 WITHOUT APIC, but CPU Disconnect on. I threw out usb and dvb. SO far stable. But I know that from a certain kernel onward all will lock up even without APIC an my board. So I am trying to find out whcih patch causes this. Then I'll reinvest the APIC issue. I dunno whether Ross' patches are still not 100% work-arounds or something else of the system causes lock-ups.

malloc · l33t Joined: 19 Sep 2003 Posts: 762

I think you don't need to patch the kernel anymore. Just upgrade your board bios to the latest one available (1007 for Asus boards) and the APIC lockups will go away.

irasnyd · Posted: Tue Feb 17, 2004 1:30 am Post subject:

I've just tried applying this patch to love-sources. It failed hunk #4, so I fixed it by hand. Now, when I run a kernel compile (make && make modules_install) I get this:

supated · Posted: Tue Feb 17, 2004 6:57 am Post subject: