Re: [PATCH] ASoC: cs42l43-jack: Remove manual pm_runtime get/put from tip_sense_work

[Charles Keepax <ckeepax@opensource.cirrus.com>]

On Tue, Mar 17, 2026 at 02:07:29PM +0200, Péter Ujfalusi wrote:
> On 17/03/2026 13:44, Charles Keepax wrote:
> > On Tue, Mar 17, 2026 at 08:21:12AM +0200, Péter Ujfalusi wrote:
> >> On 16/03/2026 18:40, Charles Keepax wrote:
> >>> On Mon, Mar 16, 2026 at 04:37:28PM +0200, Péter Ujfalusi wrote:
> >>>> On 16/03/2026 16:27, Charles Keepax wrote:
> > There is probably a discussion to be had here, its far from clear
> > to me this is the wrong place to do this. Generally the codec
> > controls when the codec wants to mark itself as runtime active.
> > For example on our phone devices where far more of the chip
> > powered down in runtime suspend having a jack in would always
> > keep the device powered up so the button detect could run,
> > as the lowest power states disabled that.
> 
> I see, what about this:
> if the inserted accessory is CS42L43_JACK_HEADSET (SND_JACK_HEADSET)
> then do one more pm_runtime_get() to allow the button presses to be handled?
> This would allow the laptop to hit lower power state if a headphone is
> connected, headphones do not have buttons as they don't have mic ring.
> 
> The jack_plugged would be renamed as jack_is_headset or something and
> drop the rpm on jack removal for headset ( and on module remove).

Yeah it might be possible to not hold the reference for
headphones as we don't need the button detect. We will need to
have a little think about the re-running of the jack detection. I
am not super keen on it running a full jack detect each time we
exit a clock stop. But we might be able to cache that and ignore
the new IRQs after the clock stop.

Thanks,
Charles