oops / 2.6.11 / run_timer_softirq (mountvirtfs)

oops / 2.6.11 / run_timer_softirq (mountvirtfs)

[jerome lacoste]

On an VIA EPIA board, I got this single oops at boot. Wasn't stored on
file so I had to take a screenshot with a digital camera. Basicallly
goes along those lines:

Process: S36mountvirtfs

Call trace:
     run_timer_softirq+0x16f/0x200
     __do_softirq
     do_softirq
     irq_exit
     do_IRQ
     common_interrupt

Process is found here on my system:

lrwxr-xr-x  1 root root 21 Mar  1 00:29 /etc/rcS.d/S36mountvirtfs ->
../init.d/mountvirtfs

The exact screenshot (500k) can be found here:

http://coffeebreaks.dyndns.org/~jerome/static/images/linux/oops_2.6.11_run_timer_softirq_boot.jpg

I can spend time copying the input into a file and doing the ksymoops
stuff if someone wants to get it.
Otherwise I will go back to try fixing the other problems that happens
much more often...

J

Re: oops / 2.6.11 / run_timer_softirq (mountvirtfs)

[Andrew Morton]

jerome lacoste <jerome.lacoste@gmail.com> wrote:
>
> On an VIA EPIA board, I got this single oops at boot. Wasn't stored on
> file so I had to take a screenshot with a digital camera. Basicallly
> goes along those lines:
> 
> Process: S36mountvirtfs
> 
> Call trace:
>      run_timer_softirq+0x16f/0x200
>      __do_softirq
>      do_softirq
>      irq_exit
>      do_IRQ
>      common_interrupt
> 
> Process is found here on my system:
> 
> lrwxr-xr-x  1 root root 21 Mar  1 00:29 /etc/rcS.d/S36mountvirtfs ->
> ../init.d/mountvirtfs
> 
> The exact screenshot (500k) can be found here:
> 
> http://coffeebreaks.dyndns.org/~jerome/static/images/linux/oops_2.6.11_run_timer_softirq_boot.jpg
> 

An oops in cascade() is tricky.  Normally it means that some piece of code
has done something bad with a kernel timer.  Later, a clock tick happens
and the kernel falls over.  We're left with no hints as to which part of
the kernel misbehaved.

Please try enabling CONFIG_DEBUG_SLAB and CONFIG_DEBUG_PAGEALLOC and see if
that reveals any additional info.

Apart from that, you have a lot of modules configured there.  Please try
disabling them all, see if the oops goes away.  If it does then try
re-enabling them, see if you can narrow it down to the one which is causing
the timer list corruption.

Thanks.

Re: oops / 2.6.11 / run_timer_softirq (mountvirtfs)

[jerome lacoste]

On Thu, 10 Mar 2005 20:59:43 -0800, Andrew Morton <akpm@osdl.org> wrote:
> jerome lacoste <jerome.lacoste@gmail.com> wrote:
> >
> > On an VIA EPIA board, I got this single oops at boot. Wasn't stored on
> > file so I had to take a screenshot with a digital camera. Basicallly
> > goes along those lines:
> >
> > Process: S36mountvirtfs
> >
> > Call trace:
> >      run_timer_softirq+0x16f/0x200
> >      __do_softirq
> >      do_softirq
> >      irq_exit
> >      do_IRQ
> >      common_interrupt
> >
> > Process is found here on my system:
> >
> > lrwxr-xr-x  1 root root 21 Mar  1 00:29 /etc/rcS.d/S36mountvirtfs ->
> > ../init.d/mountvirtfs
> >
> > The exact screenshot (500k) can be found here:
> >
> > http://coffeebreaks.dyndns.org/~jerome/static/images/linux/oops_2.6.11_run_timer_softirq_boot.jpg
> >
> 
> An oops in cascade() is tricky.  Normally it means that some piece of code
> has done something bad with a kernel timer.  Later, a clock tick happens
> and the kernel falls over.  We're left with no hints as to which part of
> the kernel misbehaved.
> 
> Please try enabling CONFIG_DEBUG_SLAB and CONFIG_DEBUG_PAGEALLOC and see if
> that reveals any additional info.

Question; the thing happened once at boot time (out of hundreds) so it
will probably be hard to reproduce.

I you may have seen on the pictures, the screen was completely filled
up with the oops information. How will the new CONFIG_ options help if
I don't have more information on the screen when it oopses?

> Apart from that, you have a lot of modules configured there.  Please try
> disabling them all, see if the oops goes away.  If it does then try
> re-enabling them, see if you can narrow it down to the one which is causing
> the timer list corruption.

If the problem reappears I will see what I can do.

Jerome

> Thanks.

Pareillement

J