Re: [PATCH V1 2/4] ALSA: hda: add no_pin_sense_update flag for jack detection

[Takashi Iwai <tiwai@suse.de>]

On Mon, 23 Jun 2025 12:39:48 +0200,
Joakim  Zhang wrote:
> 
> Hello Takashi,
> 
> > -----Original Message-----
> > From: Joakim Zhang <joakim.zhang@cixtech.com>
> > Sent: Saturday, June 21, 2025 2:15 PM
> > To: Takashi Iwai <tiwai@suse.de>
> > Cc: perex@perex.cz; tiwai@suse.com; linux-sound@vger.kernel.org;
> > chris.chiu@canonical.com; kailang@realtek.com; geans_chen@realsil.com.cn;
> > cix-kernel-upstream <cix-kernel-upstream@cixtech.com>
> > Subject: 回复: [PATCH V1 2/4] ALSA: hda: add no_pin_sense_update flag for
> > jack detection
> >
> >
> > Hello Takashi,
> > ________________________________________
> > 发件人: Takashi Iwai <tiwai@suse.de>
> > 已发送: 2025 年 6 月 20 日 星期五 20:11
> > 收件人: Joakim  Zhang <joakim.zhang@cixtech.com>
> > 抄送: perex@perex.cz <perex@perex.cz>; tiwai@suse.com <tiwai@suse.com>;
> > linux-sound@vger.kernel.org <linux-sound@vger.kernel.org>;
> > chris.chiu@canonical.com <chris.chiu@canonical.com>; kailang@realtek.com
> > <kailang@realtek.com>; geans_chen@realsil.com.cn
> > <geans_chen@realsil.com.cn>; cix-kernel-upstream <cix-kernel-
> > upstream@cixtech.com>
> > 主题: Re: [PATCH V1 2/4] ALSA: hda: add no_pin_sense_update flag for jack
> > detection
> 
> [......]
> 
> > [Joakim] Let me take realtek alc256 as an example,
> > hda_call_codec_resume()
> >     1)codec->patch_ops.resume(codec) // alc269_resume()
> >         codec->patch_ops.init(codec) // alc_init()
> >             spec->init_hook(codec) // alc256_init()
> >             snd_hda_jack_detect(codec, hp_pin)
> >                     ....
> >                     snd_hda_jack_pin_sense()
> >                         jack_detect_update()
> >                             jack->pin_sense = read_pin_sense(codec, jack->nid, jack-
> > >dev_id);  // update jack->pin_sense here, but not report jack status
> >     2)hda_jackpoll_work(&codec->jackpoll_work.work)
> >             snd_hda_jack_poll_all(codec)
> >                  old_sense = get_jack_plug_state(jack->pin_sense);
> >                  jack_detect_update(codec, jack);
> >                  if (changes)
> >                       snd_hda_jack_report_sync(codec);
> >
> > If the headset is not plugin before system suspending, so the jack->pin_sense
> > = 0x0, plugin the headset when system in suspended state, system resume
> > back, the jack->pin_sense update to 0x80000000 from codec
> > patch_ops.init(), then jackpoll work get the old_sense = 0x80000000 (but
> > actually it is not old sense before system suspending), and the current jack-
> > >pin_sense also 0x80000000 when calling jack_detect_update(), as a result
> > there is no changes, finally not report the headset plugin event.
> >
> > Is it logical to report jack status immediately when updating the jack pin
> > sense? If not, the old pin sense can still be trusted?
> >
> > Please correct me if anything misunderstanding, thanks, as a hda newer may
> > not understood very deeply.
> 
> I confirm this issue exist in the original logic, but your PoC also fix this issue, I quite agree with your above point of view:
> "I believe we should simply call  snd_hda_jack_report_sync() unconditionally, then schedule the jackpoll work,  as done in my PoC patch."
> 
> With your PoC patch, three issues(patch 1/2/3) have been solved simultaneously, please feel free to add my tested tag:
>       Tested-by: Joakim Zhang <joakim.zhang@cixtech.com>

Good to hear!  I'm going to submit the proper patch set.


thanks,

Takashi