Hi Takashi,

spec->suppress_auto_mute = 1; This will cause spec->gen.autocfg.hp_pins [0] = 0. This will cause alc2??_init() and alc2??_shutup() and dell headset mode not run errors.

Do it has other place to store pin NID? Or just only this place hp_pin = spec->gen.autocfg.hp_pins[0]; to get headphone NID.

BR, Kailang

-----Original Message----- From: Takashi Iwai tiwai@suse.de Sent: Thursday, January 31, 2019 4:23 PM To: Kailang kailang@realtek.com Cc: (alsa-devel@alsa-project.org) alsa-devel@alsa-project.org Subject: Re: auto_mute for early patch issue

On Thu, 31 Jan 2019 09:00:36 +0100, Kailang wrote:

Hi Takashi,

I move check line to below. It will show value for hp_pins[0]. As I know, search from verb table in function snd_hda_parse_pin_defcfg(). It will fill pins at auto_pin_cfg. Why it need to run below function to get cfg->hp_pins?

It's just because hp_pins[] and speaker_pins[] are used for the case with the auto-mute, so far, not for other evaluations :)

The change below looks safe, so if this works for you, feel fee to apply like that.

thanks,

Takashi

static int check_auto_mute_availability(struct hda_codec *codec) { struct hda_gen_spec *spec = codec->spec; struct auto_pin_cfg *cfg = &spec->autocfg; int present = 0; int i, err;

• if (spec->suppress_auto_mute)

• return 0;
  
  

  if (cfg->hp_pins[0]) present++; if (cfg->line_out_pins[0]) present++; if (cfg->speaker_pins[0]) present++; if (present < 2) /* need two different output types */ return 0;

  if (!cfg->speaker_pins[0] && cfg->line_out_type == AUTO_PIN_SPEAKER_OUT) { memcpy(cfg->speaker_pins, cfg->line_out_pins, sizeof(cfg->speaker_pins)); cfg->speaker_outs = cfg->line_outs; }

  if (!cfg->hp_pins[0] && cfg->line_out_type == AUTO_PIN_HP_OUT) { memcpy(cfg->hp_pins, cfg->line_out_pins, sizeof(cfg->hp_pins)); cfg->hp_outs = cfg->line_outs; }

• if (spec->suppress_auto_mute)

• return 0;
  
  

...... ..... ...... }

BR, Kailang

-----Original Message----- From: Kailang Sent: Thursday, January 31, 2019 12:08 PM To: Takashi Iwai (tiwai@suse.de) tiwai@suse.de Cc: (alsa-devel@alsa-project.org) alsa-devel@alsa-project.org Subject: auto_mute for early patch issue

Hi Takashi,

Chrome OS use early patch firmware file to load model and use auto_mute =no hint options. I find an issue for early patch issue about hint option auto_mute = no.

This will cause all spec->init_hook can't get hp_pin values.

For example: alc225_init() alc225_shutup() alc256_init() alc256_shutup() Early patch file hint auto_mute = no equal spec->suppress_auto_mute = 1 It will return function check_auto_mute_availability() in had_generic.c.

hp_pin = spec->gen.autocfg.hp_pins[0]; ==> It will get value 0.

I think this issue was for a long time. Please see the attach dmesg. One is kill hint auto_mute option.

BR, Kailang

------Please consider the environment before printing this e-mail.

On Thu, 31 Jan 2019 11:50:26 +0100, Kailang wrote:

Hi Takashi,

spec->suppress_auto_mute = 1; This will cause spec->gen.autocfg.hp_pins [0] = 0. This will cause alc2??_init() and alc2??_shutup() and dell headset mode not run errors.

That's the result with your suggested change (moving spec->supress_auto_mute_mode in check_auto_mute_availability()) ??

Takashi

Does it has other place to store pin NID? Or just only this place hp_pin = spec->gen.autocfg.hp_pins[0]; to get headphone NID.

BR, Kailang

-----Original Message----- From: Takashi Iwai tiwai@suse.de Sent: Thursday, January 31, 2019 4:23 PM To: Kailang kailang@realtek.com Cc: (alsa-devel@alsa-project.org) alsa-devel@alsa-project.org Subject: Re: auto_mute for early patch issue

On Thu, 31 Jan 2019 09:00:36 +0100, Kailang wrote:

...

Hi Takashi,

I move check line to below. It will show value for hp_pins[0]. As I know, search from verb table in function snd_hda_parse_pin_defcfg(). It will fill pins at auto_pin_cfg. Why it need to run below function to get cfg->hp_pins?

It's just because hp_pins[] and speaker_pins[] are used for the case with the auto-mute, so far, not for other evaluations :)

The change below looks safe, so if this works for you, feel fee to apply like that.

thanks,

Takashi

...

static int check_auto_mute_availability(struct hda_codec *codec) { struct hda_gen_spec *spec = codec->spec; struct auto_pin_cfg *cfg = &spec->autocfg; int present = 0; int i, err;

• if (spec->suppress_auto_mute)

• return 0;
  
  

  if (cfg->hp_pins[0]) present++; if (cfg->line_out_pins[0]) present++; if (cfg->speaker_pins[0]) present++; if (present < 2) /* need two different output types */ return 0;

  if (!cfg->speaker_pins[0] && cfg->line_out_type == AUTO_PIN_SPEAKER_OUT) { memcpy(cfg->speaker_pins, cfg->line_out_pins, sizeof(cfg->speaker_pins)); cfg->speaker_outs = cfg->line_outs; }

  if (!cfg->hp_pins[0] && cfg->line_out_type == AUTO_PIN_HP_OUT) { memcpy(cfg->hp_pins, cfg->line_out_pins, sizeof(cfg->hp_pins)); cfg->hp_outs = cfg->line_outs; }

• if (spec->suppress_auto_mute)

• return 0;
  
  

...... ..... ...... }

BR, Kailang

-----Original Message----- From: Kailang Sent: Thursday, January 31, 2019 12:08 PM To: Takashi Iwai (tiwai@suse.de) tiwai@suse.de Cc: (alsa-devel@alsa-project.org) alsa-devel@alsa-project.org Subject: auto_mute for early patch issue

Hi Takashi,

Chrome OS use early patch firmware file to load model and use auto_mute =no hint options. I find an issue for early patch issue about hint option auto_mute = no.

This will cause all spec->init_hook can't get hp_pin values.

For example: alc225_init() alc225_shutup() alc256_init() alc256_shutup() Early patch file hint auto_mute = no equal spec->suppress_auto_mute = 1 It will return function check_auto_mute_availability() in had_generic.c.

hp_pin = spec->gen.autocfg.hp_pins[0]; ==> It will get value 0.

I think this issue was for a long time. Please see the attach dmesg. One is kill hint auto_mute option.

BR, Kailang

------Please consider the environment before printing this e-mail.

On Fri, 01 Feb 2019 10:18:45 +0100, Kailang wrote:

Hi Takashi,

I forgot it solved this issue for ALC283 before. It was also chrome OS case. I cooked attach patch. It will not need to move spec->suppress_auto_mute check line. Many thanks.

OK, I'll queue this to for-linus branch, as it fixes real bugs.

OTOH, I find filling hp_pins[] unconditionally makes sense. So, for 5.1 development, maybe it'd be better to drop this change again and apply the change to move the spec->suppress_auto_mute check.

But the remaining question is whether moving the check helped at all; it wasn't clear from your previous replies...

thanks,

Takashi

BR, Kailang

-----Original Message----- From: Takashi Iwai tiwai@suse.de Sent: Thursday, January 31, 2019 7:02 PM To: Kailang kailang@realtek.com Cc: (alsa-devel@alsa-project.org) alsa-devel@alsa-project.org Subject: Re: auto_mute for early patch issue

On Thu, 31 Jan 2019 11:50:26 +0100, Kailang wrote:

...

Hi Takashi,

spec->suppress_auto_mute = 1; This will cause spec->gen.autocfg.hp_pins [0] = 0. This will cause alc2??_init() and alc2??_shutup() and dell headset mode not run errors.

That's the result with your suggested change (moving spec->supress_auto_mute_mode in check_auto_mute_availability()) ??

Takashi

...

Does it has other place to store pin NID? Or just only this place hp_pin = spec->gen.autocfg.hp_pins[0]; to get headphone NID.

BR, Kailang

-----Original Message----- From: Takashi Iwai tiwai@suse.de Sent: Thursday, January 31, 2019 4:23 PM To: Kailang kailang@realtek.com Cc: (alsa-devel@alsa-project.org) alsa-devel@alsa-project.org Subject: Re: auto_mute for early patch issue

On Thu, 31 Jan 2019 09:00:36 +0100, Kailang wrote:

...

Hi Takashi,

I move check line to below. It will show value for hp_pins[0]. As I know, search from verb table in function snd_hda_parse_pin_defcfg(). It will fill pins at auto_pin_cfg. Why it need to run below function to get cfg->hp_pins?

It's just because hp_pins[] and speaker_pins[] are used for the case with the auto-mute, so far, not for other evaluations :)

The change below looks safe, so if this works for you, feel fee to apply like that.

thanks,

Takashi

...

static int check_auto_mute_availability(struct hda_codec *codec) { struct hda_gen_spec *spec = codec->spec; struct auto_pin_cfg *cfg = &spec->autocfg; int present = 0; int i, err;

• if (spec->suppress_auto_mute)

• return 0;
  
  

  if (cfg->hp_pins[0]) present++; if (cfg->line_out_pins[0]) present++; if (cfg->speaker_pins[0]) present++; if (present < 2) /* need two different output types */ return 0;

  if (!cfg->speaker_pins[0] && cfg->line_out_type == AUTO_PIN_SPEAKER_OUT) { memcpy(cfg->speaker_pins, cfg->line_out_pins, sizeof(cfg->speaker_pins)); cfg->speaker_outs = cfg->line_outs; }

  if (!cfg->hp_pins[0] && cfg->line_out_type == AUTO_PIN_HP_OUT) { memcpy(cfg->hp_pins, cfg->line_out_pins, sizeof(cfg->hp_pins)); cfg->hp_outs = cfg->line_outs; }

• if (spec->suppress_auto_mute)

• return 0;
  
  

...... ..... ...... }

BR, Kailang

-----Original Message----- From: Kailang Sent: Thursday, January 31, 2019 12:08 PM To: Takashi Iwai (tiwai@suse.de) tiwai@suse.de Cc: (alsa-devel@alsa-project.org) alsa-devel@alsa-project.org Subject: auto_mute for early patch issue

Hi Takashi,

Chrome OS use early patch firmware file to load model and use auto_mute =no hint options. I find an issue for early patch issue about hint option auto_mute = no.

This will cause all spec->init_hook can't get hp_pin values.

For example: alc225_init() alc225_shutup() alc256_init() alc256_shutup() Early patch file hint auto_mute = no equal spec->suppress_auto_mute = 1 It will return function check_auto_mute_availability() in had_generic.c.

hp_pin = spec->gen.autocfg.hp_pins[0]; ==> It will get value 0.

I think this issue was for a long time. Please see the attach dmesg. One is kill hint auto_mute option.

BR, Kailang

------Please consider the environment before printing this e-mail.

[2 0000-fix-lose-hp-pins.patch <application/octet-stream (base64)>]