On Mon, Jan 11, 2021 at 08:02:50AM +0200, Tony Lindgren wrote:
> * Tony Lindgren <tony@xxxxxxxxxxx> [210111 05:13]:
> > * Dmitry Torokhov <dmitry.torokhov@xxxxxxxxx> [210111 05:01]:
> > > Hi Tony,
> > >
> > > On Sun, Jan 10, 2021 at 09:05:28PM +0200, Tony Lindgren wrote:
> > > > @@ -350,15 +379,12 @@ static int omap4_keypad_probe(struct platform_device *pdev)
> > > >
> > > > error = omap4_keypad_check_revision(&pdev->dev,
> > > > keypad_data);
> > > > - if (!error) {
> > > > - /* Ensure device does not raise interrupts */
> > > > - omap4_keypad_stop(keypad_data);
> > > > - }
> > > > -
> > > > - pm_runtime_put_sync(&pdev->dev);
> > >
> > > Why are we moving this down? The idea was to make sure the power usage
> > > counters are correct even if we exit probe early.
> >
> > Yes you are right, omap4_keypad_close() won't help with balancing the
> > get if we exit early.
> >
> > > Can we call pm_runtime_mark_last_busy() and pm_runtime_put_autosuspend()
> > > here?
> >
> > Yes that should work as there's no more register access during the probe.
>
> Below is an updated version. I updated probe to use dev instead of
> &pdev->dev since we have it there. Does this look OK to you?
Yep, looks good, applied.
--
Dmitry
