[2.6.35.x regression] rcu_preempt_state stall warning and machine slow-downs

[2.6.35.x regression] rcu_preempt_state stall warning and machine slow-downs

[Matthias Dahl]

[-- Attachment #1: Type: text/plain, Size: 996 bytes --]

Hi.

I've a recurring problem w/ kernel 2.6.35.x that I cannot pinpoint exactly but
seems to be related to rcu_preempt_state stall warnings I get frequently.

From time to time, the machine slows down to an almost halt (meaning jumpy or
non-responsive mouse cursor, unable to switch to text console, ...) which can
also result in a complete hang w/o any way to recover.   Sometimes the system
catches up again but once this happened,  the system is sluggish and a reboot
is required nevertheless. High load seems to help a lot to trigger this issue.

In every instance, I found stall warnings in my log which might be related. I
have attached three of them along with my config and dmesg.

The system works flawlessly with pre-2.6.35 kernels-  no stall warnings and no
slow-downs at all.

If there is anything I can do to help further diagnose this problem, please let
me know.

So long,
matthias.

PS. Please cc' me because I am not subscribed to the list. Thanks a lot.

[-- Attachment #2: config.gz --]
[-- Type: application/x-gzip, Size: 13554 bytes --]

[-- Attachment #3: backtrace2.gz --]
[-- Type: application/x-gzip, Size: 3572 bytes --]

[-- Attachment #4: backtrace1.gz --]
[-- Type: application/x-gzip, Size: 3737 bytes --]

[-- Attachment #5: backtrace3.gz --]
[-- Type: application/x-gzip, Size: 3630 bytes --]

[-- Attachment #6: dmesg.gz --]
[-- Type: application/x-gzip, Size: 11974 bytes --]

Re: [2.6.35.x regression] rcu_preempt_state stall warning and machine slow-downs

[Maciej Rutecki]

On środa, 1 września 2010 o 08:47:19 Matthias Dahl wrote:
> Hi.
> 
> I've a recurring problem w/ kernel 2.6.35.x that I cannot pinpoint exactly
> but seems to be related to rcu_preempt_state stall warnings I get
> frequently.
> 
> From time to time, the machine slows down to an almost halt (meaning jumpy
> or non-responsive mouse cursor, unable to switch to text console, ...)
> which can also result in a complete hang w/o any way to recover.  
> Sometimes the system catches up again but once this happened,  the system
> is sluggish and a reboot is required nevertheless. High load seems to help
> a lot to trigger this issue.
> 
> In every instance, I found stall warnings in my log which might be related.
> I have attached three of them along with my config and dmesg.
> 
> The system works flawlessly with pre-2.6.35 kernels-  no stall warnings and
> no slow-downs at all.
> 
> If there is anything I can do to help further diagnose this problem, please
> let me know.
> 
> So long,
> matthias.
> 
> PS. Please cc' me because I am not subscribed to the list. Thanks a lot.

I created a Bugzilla entry at 
https://bugzilla.kernel.org/show_bug.cgi?id=17732
for your bug report, please add your address to the CC list in there, thanks!

-- 
Maciej Rutecki
http://www.maciek.unixy.pl