All of lore.kernel.org
 help / color / mirror / Atom feed
* [Xenomai-help] pthread_make_periodic_np initial start date
@ 2010-02-02 22:29 Steve Deiters
  2010-02-02 22:35 ` Gilles Chanteperdrix
  0 siblings, 1 reply; 2+ messages in thread
From: Steve Deiters @ 2010-02-02 22:29 UTC (permalink / raw)
  To: xenomai

[-- Attachment #1: Type: text/plain, Size: 648 bytes --]

Is there a way to specify no initial delay when calling
pthread_make_periodic_np?  If I pass a value directly obtained from
clock_gettime, it will return a ETIMEDOUT error due to time passing
before I make the call.  I could add a small delay to the value obtained
from clock_gettime, but I would be concerned that due to other
happenings in the system, the absolute date may still have passed by the
time it gets around to calling pthread_make_periodic_np.

What I would be looking for I guess is what happens when I call
rt_task_set_periodic with an idate parameter of TM_NOW.  Is there any
way to simulate this behavior?

Thanks.


[-- Attachment #2: Type: text/html, Size: 1269 bytes --]

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [Xenomai-help] pthread_make_periodic_np initial start date
  2010-02-02 22:29 [Xenomai-help] pthread_make_periodic_np initial start date Steve Deiters
@ 2010-02-02 22:35 ` Gilles Chanteperdrix
  0 siblings, 0 replies; 2+ messages in thread
From: Gilles Chanteperdrix @ 2010-02-02 22:35 UTC (permalink / raw)
  To: Steve Deiters; +Cc: xenomai

Steve Deiters wrote:
> Is there a way to specify no initial delay when calling
> pthread_make_periodic_np?  If I pass a value directly obtained from
> clock_gettime, it will return a ETIMEDOUT error due to time passing
> before I make the call.  I could add a small delay to the value obtained
> from clock_gettime, but I would be concerned that due to other
> happenings in the system, the absolute date may still have passed by the
> time it gets around to calling pthread_make_periodic_np.
> 
> What I would be looking for I guess is what happens when I call
> rt_task_set_periodic with an idate parameter of TM_NOW.  Is there any
> way to simulate this behavior?

pthread_make_periodic_np is a legacy from RTLinux. You should be using
clock_nanosleep instead.

That said, pthread_make_periodic_np internally uses the same service as
rt_task_set_periodic, and TM_NOW is just 0. So, you can probably pass a
timespec with tv_sec and tv_nsec both 0.


-- 
					    Gilles.


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2010-02-02 22:35 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-02-02 22:29 [Xenomai-help] pthread_make_periodic_np initial start date Steve Deiters
2010-02-02 22:35 ` Gilles Chanteperdrix

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.