Problem with rtems_event_send crashing from interrupt routine

Mr. Andrei Chichak groups at chichak.ca
Wed Jun 10 20:07:54 UTC 2015


On 2015-June-10, at 1:37 PM, Sebastian Huber <sebastian.huber at embedded-brains.de> wrote:
> 
> 
> ----- Mr. Andrei Chichak <groups at chichak.ca> schrieb:
>> 
>>> On 2015-June-10, at 1:26 PM, Sebastian Huber <sebastian.huber at embedded-brains.de> wrote:
>>> 
>>> Hello Andrei,
>>> 
>>> which RTEMS version do you use?
>>> 
>> 
>> Sorry about that, 4.11 updated to the head earlier this week, with the tools rebuilt earlier this week using source-builder.
> 
> Ok, in which function is this infinite loop?

I was afraid that you would ask that. I am at the stage of doing “printk” debugging. If I use my JTAG module with GDB, as soon as the interrupts start GDB gets confused and the debugger is useless. I have not been able to find a way around this or I could halt and examine the task control blocks to see which tasks are where.



I assumed that it was an infinite loop within RTEMS (which doesn’t help) as I don’t have any user code left, I can bracket the event_receive call with GPIO set/reset sequences and receive never returns, and when the problem happens the lower priority task does not keep running. I was thinking that the higher priority task was cycle starving the lower. The interrupt routines keep firing though.

I’m currently checking to see if I change the priorities of my two tasks (monitor putting out rtems_cpu_usage_report once per second, and my ADC scanner that just sends and receives events) so that the monitor has a higher priority, it will keep running and the ADC routine will show no time being consumed, just to be sure and to give me a way to probe the task tables with printk.

Andrei








More information about the users mailing list