Re: Realtime Preemption, 2.6.12, Beginners Guide?

[Alistair John Strachan <s0348365@sms.ed.ac.uk>]

On Thursday 07 Jul 2005 13:15, Alistair John Strachan wrote:
> On Thursday 07 Jul 2005 12:42, Ingo Molnar wrote:
> > * Alistair John Strachan <s0348365@sms.ed.ac.uk> wrote:
> > > > do you have DEBUG_STACKOVERFLOW and latency tracing still enabled? 
> > > > The combination of those two options is pretty good at detecting
> > > > stack overflows. Also, you might want to enable CONFIG_4KSTACKS, that
> > > > too disturbs the stack layout enough so that the error message may
> > > > make it to the console.
> > >
> > > I already have 4KSTACKS on. Latency tracing is enabled, but
> > > STACKOVERFLOW isn't; I'll just reenable everything again until we fix
> > > this. Do you think if I removed the printk() line I might get some
> > > useful information, before it does the stack trace?
> >
> > usually such loops happen if the stack has been overflown and critical
> > information that lies on the bottom of the stack (struct thread_info) is
> > overwritten. Then we often cannot even perform simple printks. Stack
> > overflow debugging wont prevent the crash, but might give a better
> > traceback.
> >
> > 	Ingo
>
> http://devzero.co.uk/~alistair/oops1.jpeg
>
> I disabled the trace and the STACKOVERFLOW option seems to help; I've got a
> (slightly truncated) oops from the kernel. What happens is that I get an
> oops, then I get a BUG: warning me about the softlock, then I get another
> oops. I'm about to reboot to confirm whether the second oops is identical
> to the first (I suspect that it is).

http://devzero.co.uk/~alistair/oops3.jpeg

This shows the first oops (it's slightly different).

http://devzero.co.uk/~alistair/oops2.jpeg

This shows the BUG: after the first oops

-- 
Cheers,
Alistair.

personal:   alistair()devzero!co!uk
university: s0348365()sms!ed!ac!uk
student:    CS/CSim Undergraduate
contact:    1F2 55 South Clerk Street,
            Edinburgh. EH8 9PP.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/