Steven Seeger wrote:
> If I want a periodic task to wait on a condition variable whose purpose
> is to wait until the task is needed again, should I set it non-periodic
> first and then set it periodic again after it wakes up?
>
>
>
> If I am running in periodic timebase mode, I notice weird behavior
> without this behavior. It seems ok in aperiodic mode, however.
Could you reduce your code to a minimal example showing this behaviour?
> Also, if I set the native skin to periodic mode, the latency default
> setting of 4200 doesn't change tick types, therefore it causes all my
> calls to rt_task_set_periodic() fail with jiffies less than this.
Try the attached patch. It looks like a condition got negated when
replacing the test by a predicate.
--
Gilles Chanteperdrix.
Index: ksrc/nucleus/pod.c
===================================================================
--- ksrc/nucleus/pod.c (revision 3460)
+++ ksrc/nucleus/pod.c (working copy)
@@ -3305,7 +3305,8 @@ int xnpod_set_thread_periodic(xnthread_t
xntimer_stop(&thread->ptimer);
goto unlock_and_exit;
- } else if (xntbase_periodic_p(xnthread_time_base(thread)) && period <
nklatency) {
+ } else if (!xntbase_periodic_p(xnthread_time_base(thread))
+ && period < nklatency) {
/* LART: detect periods which are shorter than the
* intrinsic latency figure; this must be a joke... */
err = -EINVAL;
_______________________________________________
Xenomai-help mailing list
[email protected]
https://mail.gna.org/listinfo/xenomai-help
