Petr Cervenka wrote:
> Hello again,
> I have though, that all my troubles with this kernel panic (see 
> https://mail.gna.org/public/xenomai-help/2008-07/msg00054.html) are over, but 
> suddenly it reappeared.
> The full log is in attachment. I used snapshot of screen, OCR program and 
> manual correction, so expect possible errors in log.
> I haven't seen the error for months, I was believed that it was caused the 
> combination of other two problems (hopefully already solved or evaded). But 
> this kernel panic emerged  during the weekend test, when we tested usability 
> of a new motherboard for our purposes.
> It's perhaps not critical for me, but if you have any advices or questions, 
> I'm listening.

Due to the interrupt deferral the I-pipe enforces, it may well happen that we
log and postpone a timer IRQ while the root stage is stalled during a task
switch, which gets eventually played right after the context switch has taken
place, after a different memory context for the next task is reinstated. The
profiler code then tries to dereference a memory address obtained from the IRQ
stack frame that does not exist in the new context. Too bad.

Ok, we need to think about this.

> Thanks
> Petr
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Xenomai-help mailing list
> [email protected]
> https://mail.gna.org/listinfo/xenomai-help


-- 
Philippe.

_______________________________________________
Xenomai-help mailing list
[email protected]
https://mail.gna.org/listinfo/xenomai-help

Reply via email to