[PATCH] drm/i915/selftests: Use preemption timeout on cleanup

Andi Shyti andi.shyti at linux.intel.com
Tue Dec 17 17:12:08 UTC 2024


Hi Janusz,

...

> > > +
> > >  		cond_resched();
> > >  
> > > -		if (intel_gt_wait_for_idle(gt, HZ * 3) == -ETIME) {
> > > +		if (intel_gt_wait_for_idle(gt, HZ * timeout_ms / 500) == -
> ETIME) {
> > 
> > where is this 500 coming from?
> 
> / 1000 would convert it to seconds as needed, and / 500 used instead was 
> supposed to to mean that we are willing to wait for preempt_timeout_ms * 2.  
> Sorry for that shortcut.  Would you like me to provide a clarifying comment, 
> or maybe better use explicit 2 * preempt_timeout / 1000 ?

It was clear that you were doubling it, but what's more
interesting to know (perhaps in a comment) is why you are
choosing to use the double of the timeout_ms instead of other
values.

Makes sense?

Thanks,
Andi


More information about the dri-devel mailing list