Bug 15804 - BUG: soft lockup - CPU#0 stuck
Summary: BUG: soft lockup - CPU#0 stuck
Status: CLOSED DOCUMENTED
Alias: None
Product: Process Management
Classification: Unclassified
Component: Other (show other bugs)
Hardware: All Linux
: P1 high
Assignee: process_other
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-04-18 02:13 UTC by Tony Mugan
Modified: 2012-05-14 17:24 UTC (History)
2 users (show)

See Also:
Kernel Version: 2.6.34-020634rc4-generic
Subsystem:
Regression: No
Bisected commit-id:


Attachments
kern.log (269.88 KB, text/x-log)
2010-04-20 12:06 UTC, Tony Mugan
Details

Description Tony Mugan 2010-04-18 02:13:00 UTC
I have been testing Ubuntu 10.04 Lucid with all the latest updates and have periodically experienced a soft lockup.  This happens on multiple Ubuntu-released kernels.

They recommend that their packaged "mainline" kernels are tried so I have installed a couple to see if the issue is related to particular kernel releases.

I am using 2.6.33-020633-generic now which has not locked up after 30 minutes.

I had been using the later version 2.6.34-020634rc4-generic which caused lockups within about 5 minutes of a restart or cold boot. When locked up, the screen freezes and caps-lock will not register. If I wait for a few minutes, the lockup will end and everything starts to work again.

I am using an Asus M2V-MX motherboard.

Previously reported issues when running the Ubuntu Kernels to the Ubuntu Launchpad site are shown here with diagnostic files attached.

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/565202

https://bugs.launchpad.net/ubuntu/+source/linux-backports-modules-2.6.32/+bug/560676

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/560493

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/559781

I am happy to help with testing this, just let me know what you would like me to attempt.
Comment 1 Tony Mugan 2010-04-18 04:22:48 UTC
After leaving the machine idle for a couple of hours, I started to use it again and within 15 minutes had a soft lockup for about a minute with Ubuntu-packaged mainline kernel 2.6.33-020633-generic.

I am using only Evolution, Firefox and Terminator (Terminal replacement).
Comment 2 Andrew Morton 2010-04-19 22:20:52 UTC
Please add a copy of the kernel's soft-lockup diagnostic messages to this report.
Comment 3 Tony Mugan 2010-04-20 11:50:48 UTC
Andrew, 

Would be glad to.
Can you direct me to any documentation on how to get that?

Regards,

Tony.
Comment 4 Tony Mugan 2010-04-20 12:06:29 UTC
Created attachment 26060 [details]
kern.log
Comment 5 Tony Mugan 2010-04-20 12:07:33 UTC
There was a line on startup I noticed as being odd (in attached kern.log), not sure if it's relevant

[    0.356714] pci 0000:00:02.0: address space collision: [mem 0xd0000000-0xdfffffff 64bit pref] conflicts with GART [mem 0xd0000000-0xd7ffffff]
Comment 6 Tony Mugan 2010-04-20 18:10:01 UTC
I have recently replaced my motherboard going from an ASUS M2V-MX SE (which was running very well with Ubuntu 10.04 Lucid in the Alpha stages) to its predecessor ASUS M2V-MX.
This replacement motherboard has suffered from CPU soft lockups periodically but noticeably since Ubuntu kernel 2.6.32-21-generic.

I am using the following now (uname -a) for about 30 minutes without a lockup.
On the later kernel (or the "mainline" kernels up to vmlinuz-2.6.34-020634rc5-generic, this would have certainly locked up within 10 minutes of usage.

Linux orac 2.6.32-20-generic #30-Ubuntu SMP Mon Apr 12 15:20:57 UTC 2010 x86_64 GNU/Linux

I am keen to help with diagnosing this and am happy to learn what is required to assist in debugging the issue.
Comment 7 Tony Mugan 2010-04-26 12:15:04 UTC
I have disabled APIC and ACPI in the BIOS and have not had any lockups since.

The lockups were so regular that I am sure this is what was causing the issue.

If anyone wants more information, let me know.
Comment 8 Tony Mugan 2010-05-02 12:14:47 UTC
Several days later and still no lockups.  The BIOS changes I made have definitely resolved the issue.

Is there any benefit in further investigation on this?

Note You need to log in before you can comment on or make changes to this bug.