Thanks Rodrigo for reviewing this.
On Mon, 2023-08-07 at 13:56 -0400, Vivi, Rodrigo wrote:
> On Wed, Aug 02, 2023 at 04:35:01PM -0700, Alan Previn wrote:
> > When suspending, add a timeout when calling
> > intel_gt_pm_wait_for_idle else if we have a lost
> > G2H event that holds a wakeref (which would be
> > indicating of a bug elsewhere in the driver), we
> > get to complete the suspend-resume cycle, albeit
> > without all the lower power hw counters hitting
> > its targets, instead of hanging in the kernel.
> >
alan:snip
> > -int intel_wakeref_wait_for_idle(struct intel_wakeref *wf)
> > +int intel_wakeref_wait_for_idle(struct intel_wakeref *wf, int timeout_ms)
> > {
> > - int err;
> > + int err = 0;
> >
> > might_sleep();
> >
> > - err = wait_var_event_killable(&wf->wakeref,
> > - !intel_wakeref_is_active(wf));
> > + if (!timeout_ms)
> > + err = wait_var_event_killable(&wf->wakeref,
> > + !intel_wakeref_is_active(wf));
> > + else if (wait_var_event_timeout(&wf->wakeref,
> > + !intel_wakeref_is_active(wf),
> > + msecs_to_jiffies(timeout_ms)) < 1)
> > + err = -ETIME;
>
> it looks to me that -ETIMEDOUT would be a better error.
alan: okay - will do, thanks.
