Hi,
I use the following configuration imx6dl with Kernel 3.10.53 and Xenomai
2.6.4.
I have a periodic workflow with a timebase of 1ms.
The timebase creates a Xenomai task (timer_task) with priority 99.
In the normal case I have an idle time of 666µs before
the timer_task starts the new period.
In the failed case the last Linux Task switching to Gatekeeper 24µs
before the timer_task starts the new period.
Now I see that the Gatekeeper runs 1004µs and then the normal
workflow starts.
First of all is it possible that the timer_task (prio 99) interrupts the
Gatekeeper (prio 99) ?
Is there an option that the interruption of the Gatekepper is the reason for
long runtime of the Gatekeeper in error case ?
Because of the short difference (24µs) between start of Gatekeeper and
start of timer_task can be the scheduler be part of the problem ?
Here a short overview of the normal workflow
Xenomai prio 80 53µs
Xenomai prio 30 24µs
Xenomai prio 29 19µs
Xenomai prio 1 1µs
Linux task 40µs
Gatekeeper 71µs
Linux task 77µs
Gatekeeper 13µs
idle 666µs
-------------------------------------
Sum 954µs
Kind regards
Wolfgang
_______________________________________________
Xenomai mailing list
[email protected]
https://xenomai.org/mailman/listinfo/xenomai
