Bug 243083
| Summary: | soft lockup detected on CPU#0 and CPU#1 | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Bernard Fouché <bernard.fouche> | ||||
| Component: | kernel | Assignee: | Kernel Maintainer List <kernel-maint> | ||||
| Status: | CLOSED ERRATA | QA Contact: | Brian Brock <bbrock> | ||||
| Severity: | high | Docs Contact: | |||||
| Priority: | low | ||||||
| Version: | 7 | CC: | matteo, matt | ||||
| Target Milestone: | --- | ||||||
| Target Release: | --- | ||||||
| Hardware: | i386 | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | Environment: | ||||||
| Last Closed: | 2007-08-29 18:38:46 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: | |||||||
| Attachments: |
|
||||||
|
Description
Bernard Fouché
2007-06-07 08:58:06 UTC
please try kernel parameter
clocksource=acpi_pm
What I did: - after my bug report, I stayed in non-hyperthreading mode (set in bios), had the previous reported kernel traces in /var/log/messages but did not experience any problem with the computer for a full days of work (many cross-compilations, find(1) for /, etc): no crawling, no crash. - read your query, set 'clocksource=acpi_pm' in /etc/grub.conf, used the bios to re-activate hyperthreading. System booted finely. I can see in /var/log/messages: Jun 7 19:18:21 linuxbf kernel: Time: acpi_pm clocksource has been installed. (times are local time in France) Now the system works correctly, but I must go back home. I'll let the computer run and report further problems (or lack of!) (FYI Bug #240982 is still present.) Thanks. This morning the computer was still running. However it was much less responsive than yesterday when I rebooted it. I started to fill the bugzilla form while the computer was crawling more and more until it froze. The only solution was to switch it off/on. I went back to single core operation (thru bios) and now it works correctly. Having added 'clocksource=acpi_pm' got rid of 'soft lockup' messages in /var/log/messages. I'll report later, but I think that the freezing problem is linked to bug #240982 and not the present one which has vanished with the 'clocksource' statement. Using 'clocksource=acpi_pm' and hyperthreading disabled for many hours now, I still have no more "soft lockup" nor crawling. Computer ran for a week-end without any problem. IMHO 'clocksource=acpi_pm' fixed the problem. are you still running without hyperthreading enabled ? Could someone explain this parameter in basic detail please I can't find into on it. when I say this parameter I mean the 'clocksource=acpi_pm' option. Yes I'm still in single core mode, set thru bios. I wait for a fix for bug #240982 before trying dual core mode again: I can't afford freezes these days. Went back to hyperthreading set thru bios. Dropped parameter 'clocksource=acpi_pm'. Updated kernel to 2.6.21-1.3228. No more 'soft lockup' in /var/log/messages 40 minutes after reboot. Will report later if this error message is back. Computer froze after one hour. No particular output in /var/log/messages. Was unable to ssh in the computer, lost the mouse pointer when hit ctrl-alt-f1 while trying to get a text terminal. Went back to F7 original kernel, no dual core, clocksource=acpi_pm. Will retry later when I can afford to lose some more time... I am experiencing the same problem but only if I run my folding@home client. I get the messages in /var/log/messages and I cannot start new processes (but all the running one are fine). If I kill the folding@home client everything is fine. Created attachment 156990 [details]
dmesg trace
This is a snippet of the kernel messages. I will maybe try to change the
clocksource when I'll reboot the machine.
I'm running folding@home also. Thanks to have pointed it that to me. If there is a problem at a low level on threads or sockets, then F@H may activate a yet unknown bug! I'll try later to reboot with 3228 and no F@H. I can confirm the same behavior with 3228. Went back to hyperthreading with kernel 3228, no clocksource parameter. I leave the office now and won't be back until Tuesday. I let the computer idling without folding@home. Did not need to reboot since friday with kernel 3228 (computer running 6 days). No more softlockup message. This kernel is fine for me, but I did not try it with F@H. Time to close this bug report? |