Bug 140873
Summary: | Speed changing problems with VIA C3 | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Alan Cox <alan> |
Component: | kernel | Assignee: | Dave Jones <davej> |
Status: | CLOSED DEFERRED | QA Contact: | |
Severity: | high | Docs Contact: | |
Priority: | medium | ||
Version: | 3 | CC: | adrian, bughunt, buytenh, dtucker, ematt7, pfrields, wtogami |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2005-09-26 22:11:14 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
Alan Cox
2004-11-25 23:15:45 UTC
this isn't the first I've heard of problems like this. Something is very very wrong with the longhaul driver on pre-nehemiah systems. I think for the time being, the safest thing is going to be to disable it on the samuel/ezra era C3s until I/someone has found the time to dig into it. One thing we *should* be doing, but aren't -- the longhaul spec mandates that we must disable pci bus mastering around the speed transitions. I did have some code to disable mastering in the root bridge, which might be enough to satisfy this, but the whole spec needs rereading to find out what went wrong. The bad reports all starting happening circa 2.6.6 iirc. update: I've managed to reproduce this on a Nehemiah now as well, which could highlight a broader problem with the longhaul driver, as it now seems to affect every stepping. I also have a problem with my C3. It crashes every third or fourth day. As I have no monitor connected I have no idea what the reason for the crashes are but I have now disabled the cpuspeed daemon and see if it gets better. processor : 0 vendor_id : CentaurHauls cpu family : 6 model : 7 model name : VIA Samuel 2 stepping : 3 cpu MHz : 601.576 cache size : 64 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu de tsc msr cx8 mtrr pge mmx 3dnow bogomips : 1183.74 *** Bug 146690 has been marked as a duplicate of this bug. *** "Me too." Random hang every few days, which seems fixed by disabling cpuspeed. Switching the CPU frequency 100 times per second while pinging the box or running hdparm -T -t results in an almost certain hang within seconds. This happens on three different VIA PD10000 boards, 1.0GHz VIA Nehemiah. % cat /proc/cpuinfo processor : 0 vendor_id : CentaurHauls cpu family : 6 model : 9 model name : VIA Nehemiah stepping : 5 cpu MHz : 1002.457 cache size : 64 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu de pse tsc msr cx8 mtrr pge cmov mmx fxsr sse rng rng_en bogomips : 1982.46 the latest errata kernels have this driver disabled. I'll leave it disabled until I (or someone else) gets time to fix the driver upstream. I guess I figured out what this is about... The way to reproduce is to change CPU frequency quickly, then copy large files around. But prepare to run fsck as a powercycle will be necessary. This on a VIA Nehemiah, stepping 8 using a recompiled kernel based on 2.6.10-1.770_FC3. For a couple of weeks I thought that recompilation with CONFIG_MVIAC3_2=y had gotten rid of the hangs, but no. Change frequencies: cd /sys/devices/system/cpu/cpu0/cpufreq/ while true; do for freq in `cat scaling_available_frequencies`; do echo $freq > scaling_setspeed done done An update has been released for Fedora Core 3 (kernel-2.6.12-1.1372_FC3) which may contain a fix for your problem. Please update to this new kernel, and report whether or not it fixes your problem. If you have updated to Fedora Core 4 since this bug was opened, and the problem still occurs with the latest updates for that release, please change the version field of this bug to 'fc4'. Thank you. My current theory is that this is a gcc problem. The mass failures started to appear around the time I converted longhaul.c to use bitfields. At some point I'll convert upstream to using less funky C, and maybe this will just resolve itself. Until then, this driver is staying off in Fedora. |