On Thu, Dec 08, 2022 at 05:09:32PM -0800, Stephen Hemminger wrote:
> On Tue,  6 Dec 2022 09:28:24 -0800
> Tyler Retzlaff <roret...@linux.microsoft.com> wrote:
> 
> > +   /* Wait for the control thread to initialize successfully */
> > +   while ((ctrl_thread_status =
> > +                   __atomic_load_n(&params->ctrl_thread_status,
> > +                   __ATOMIC_ACQUIRE)) == CTRL_THREAD_LAUNCHING) {
> > +           /* Yield the CPU. Using sched_yield call requires maintaining
> > +            * another implementation for Windows as sched_yield is not
> > +            * supported on Windows.
> > +            */
> > +           rte_delay_us_sleep(1);
> > +   }
> 
> Would it be worth introducing and using rte_thread_yield().
> Rather than waiting 1us which seems like a long time for this.

i'd have to figure out if it can be provided for all platforms, if so
then it seems reasonable. though i would not introduce it in this patch
series since it isn't the central purpose of this change. more than
happy to do it in a new series.

of course if someone else wants to put up a patch to add it i'm more
than happy to review that too. (i have a lot of patches in the pipe
already i'd like to avoid getting too distracted).

i'll put it in my backlog to take a look.

thanks

Reply via email to