benoit at tionex.de
Thu Oct 11 09:16:14 CEST 2007
Paolo Mantegazza schrieb:
> I think the best thing to help is that you can reproduce your troubles
> with an example using a standard PC interrupt so that I can try it
> myself. A succesful debugging is not guarrantee but I can try at least.
Thanks for the offer Paolo.
I think i found the problem. There were two problems overlapping...
1.) passing NULL to rt_request_irq_task works perfectly and it was my
app, which later freezed in the irq because of other reasons.
2.) passing the handler into rt_request_irq_task generated the Ooops. It
seems, that the handler returned from rt_task_init_schmod is not
acceptable for rt_request_irq_task and causes:
"unable to handle kernel NULL pointer dereference at virtual address".
(the ref is not null)
Meanwhile i remembered, that 18 months ago, i already had that, and
after looking into the sources I decided to pass NULL. In that case a
call to _rt_whoami is used to get the task ref.
Still unclear to me, why i cannot use RT_SCHED_FIFO, as it is used in
the example. And the example is working on the same machine.
More information about the Rtai