On 15/09/06, Philippe Gerum <[EMAIL PROTECTED]> wrote:
On Fri, 2006-09-15 at 00:12 +0200, Dmitry Adamushko wrote:


> And it's not a good idea to get ipipe_catch_event() buzy spinning as
> (as I understand) ipipe_dispatch_event() may take, in general, an
> unbounded amount of time.

Actually, this is not an issue, since there is no requirement for
bounded execution time in ipipe_catch_event(). This call is part of the
initialization chores, it is not meant as being deterministic.

I meant it's not ok to keep a CPU spinning all this time. But your approach with schedule_timeout() is really better. And probably, we don't need to try being more general that it's really required as we need to get synched only with ipipe_catch_event(..., NULL) from the Linux domain.


Best regards,
Dmitry Adamushko
Xenomai-core mailing list

Reply via email to