[mythtv-users] OT : Help! -- BUG: soft lockup detected on CPU#0!

[Paul Fine]

Mark Paulus wrote:
>
>
> Paul Fine wrote:
>> Robin Hill wrote:
>>> On Tue Oct 31, 2006 at 01:29:46PM -0500, Paul Fine wrote:
>>>
>>>  
>>>> Each time I boot the computer, it starts up fine.  I start the 
>>>> vncserver.  Some time after, usually around 100 minutes later but 
>>>> the last time it was 20 minutes after reboot, the vnc remote 
>>>> session becomes unresponsive, and the direct display (my television 
>>>> through a FX5200 S-video connection) goes blank.  The computer is 
>>>> then locked up.  I found the very scary  "BUG: soft lockup detected 
>>>> on CPU#0!" message in the /var/log/messages file when I rebooted.
>>>>
>>>> KeStallExecutionProcessor+0x9/0xc [ndiswrapper]  <f9d95bc2>     
>>> I'd guess this is your problem, and also why you're unlikely to get 
>>> much
>>> help.  You've got a binary blob coded for a different operating system
>>> stuck in the middle of your system.  Have you tried the open-source 
>>> WLAN
>>> drivers?  I used to have a Netgear card which worked perfectly well 
>>> with
>>> those.
>>>
>>> Cheers,
>>>         Robin
>>>   
I pulled the Netgear Card out and the computer ran for a reasonable 
amount of time with crashing.  Unfortunately, I have not found a 
suitable open source driver for my WG311v3 WLAN card. 

Last night I uninstalled the ndisdriver package and rebuilt it from the 
latest stable source.  I reinstalled the Windows "blob" and let it run.  
By morning, the system had not crashed. 

One possible explanation I saw on the ndiswrapper lists is that some of 
the Windows drivers demand a larger stack than the default 4K stack in 
my kernel.  It recommends rebuilding the kernel with the 16K stack 
option enabled.  I am not sure this is the problem because my system did 
not crash where the ndiswrapper warning said it would if I had a stack 
problem.

Right now, I think that the stack issue is a big red flag waving in my face.

Thanks again, Robin, for pointing me towards the ndiswrapper.

--Paul

> Another question:  When did you enable the SATA controller in the BIOS?
> Was it enabled all the time, or did you enable it when you installed 
> the Hard Drive?  Also, what happens if you leave the SATA drive out, 
> and you disable the SATA controller in the BIOS?  Do you still get the 
> crashes/lockups?
>
> I was using a SIL3112 SATA controller with a 2.6.8 kernel and was 
> running into wierdnesses.  I upgraded to a 2.6.16 kernel, and the 
> SIL3112 is happy, and my box is happy.
> (Something else to look at and confuse all).
>
I never enabled the SATA controller in BIOS.  As far as I can tell, it 
has always been enabled, but never had anything to control until last 
weekend.  The first thing I tried after things started acting badly was 
to disconnect the SATA drive, which did not seems to effect anything.  I 
did not disable the controller in the BIOS, though.

If it turns out that the ndiswrapper is not the issue, I will go down 
the SATA path next.