Re: Support Dell headset mode for ALC288

Re: Support Dell headset mode for ALC288

[Takashi Iwai]

At Tue, 31 Mar 2015 08:01:57 +0000,
Kailang wrote:
> 
> 
> Hi Takashi,
>   
>  Dell has new platform with ALC288.
>  Attach patch is support headset mode for ALC288.

> @@ -3828,6 +3894,11 @@ static void alc_determine_headset_type(struct hda_codec *codec)
>   conteol) */
>  		{}
>  	};
> +	static struct coef_fw coef0286[] = {
> +		//UPDATE_COEF(0x4f, 0xfcc0, 0x5400), /* Combo Jack auto detect */
> +		UPDATE_COEF(0x4f, 0xfcc0, 0xd400),

Why is this commented out and what's the next line does instead?
And it's 0286?  Not 0288?

> @@ -3949,8 +4027,15 @@ static void alc_update_headset_jack_cb(struct hda_codec *codec,
>  				       struct hda_jack_callback *jack)
>  {
>  	struct alc_spec *spec = codec->spec;
> +	int present;
>  	spec->current_headset_type = ALC_HEADSET_TYPE_UNKNOWN;
>  	snd_hda_gen_hp_automute(codec, jack);
> +	present = spec->gen.hp_jack_present ? 0x40 : 0;
> +	if (codec->bus->pci->subsystem_vendor == 0x1028 &&
> +				codec->core.vendor_id == 0x10ec0288)
> +		/* Headset Mic enable and disable, ony for Dell Dino */
> +		snd_hda_codec_write(codec, 0x01, 0, AC_VERB_SET_GPIO_DATA,
> +					present);

This is too ugly to put in the common function.
For Dell, prepare its own callback function and use it instead.  Or,
put a new hook there to call additionally.  This avoids the PCI ID
check in each place.

So, better to split to a few patches.  One for adding ALC286 (or 288?)
support generically.  And another for adding Dell-specific things.


thanks,

Takashi

Re: Support Dell headset mode for ALC288

[Kailang]

 

> -----Original Message-----
> From: Takashi Iwai [mailto:tiwai@suse.de] 
> Sent: Saturday, April 04, 2015 8:43 PM
> To: Kailang
> Cc: (alsa-devel@alsa-project.org)
> Subject: Re: Support Dell headset mode for ALC288
> 
> At Tue, 31 Mar 2015 08:01:57 +0000,
> Kailang wrote:
> > 
> > 
> > Hi Takashi,
> >   
> >  Dell has new platform with ALC288.
> >  Attach patch is support headset mode for ALC288.
> 
> > @@ -3828,6 +3894,11 @@ static void 
> alc_determine_headset_type(struct hda_codec *codec)
> >   conteol) */
> >  		{}
> >  	};
> > +	static struct coef_fw coef0286[] = {
> > +		//UPDATE_COEF(0x4f, 0xfcc0, 0x5400), /* Combo 
> Jack auto detect */
> > +		UPDATE_COEF(0x4f, 0xfcc0, 0xd400),
> 
> Why is this commented out and what's the next line does instead?
> And it's 0286?  Not 0288?
> 
> > @@ -3949,8 +4027,15 @@ static void 
> alc_update_headset_jack_cb(struct hda_codec *codec,
> >  				       struct hda_jack_callback 
> *jack)  {
> >  	struct alc_spec *spec = codec->spec;
> > +	int present;
> >  	spec->current_headset_type = ALC_HEADSET_TYPE_UNKNOWN;
> >  	snd_hda_gen_hp_automute(codec, jack);
> > +	present = spec->gen.hp_jack_present ? 0x40 : 0;
> > +	if (codec->bus->pci->subsystem_vendor == 0x1028 &&
> > +				codec->core.vendor_id == 0x10ec0288)
> > +		/* Headset Mic enable and disable, ony for Dell Dino */
> > +		snd_hda_codec_write(codec, 0x01, 0, 
> AC_VERB_SET_GPIO_DATA,
> > +					present);
> 
> This is too ugly to put in the common function.
> For Dell, prepare its own callback function and use it 
> instead.  Or, put a new hook there to call additionally.  
> This avoids the PCI ID check in each place.
> 
> So, better to split to a few patches.  One for adding ALC286 
> (or 288?) support generically.  And another for adding 
> Dell-specific things.
>
 
Ok! I will split it for two patches.
But spec->gen.hp_automute_hook = alc_update_headset_jack_cb in alc_fixup_headset_mode().
If I move dell specific part to fixup table by SSID, spec->gen.hp_automute_hook will be reinit to new.
So, I just place the code here.

> 
> thanks,
> 
> Takashi
> 
> ------Please consider the environment before printing this e-mail.
> 

Re: Support Dell headset mode for ALC288

[Takashi Iwai]

At Wed, 8 Apr 2015 07:31:33 +0000,
Kailang wrote:
> 
> 
> Hi Takashi,
> 
> Attach patch is support alc288 headset mode.

This also supports ALC286, too, right?


Takashi

> 
> BR,
> Kailang
> 
> > -----Original Message-----
> > From: Takashi Iwai [mailto:tiwai@suse.de] 
> > Sent: Saturday, April 04, 2015 8:43 PM
> > To: Kailang
> > Cc: (alsa-devel@alsa-project.org)
> > Subject: Re: Support Dell headset mode for ALC288
> > 
> > At Tue, 31 Mar 2015 08:01:57 +0000,
> > Kailang wrote:
> > > 
> > > 
> > > Hi Takashi,
> > >   
> > >  Dell has new platform with ALC288.
> > >  Attach patch is support headset mode for ALC288.
> > 
> > > @@ -3828,6 +3894,11 @@ static void 
> > alc_determine_headset_type(struct hda_codec *codec)
> > >   conteol) */
> > >  		{}
> > >  	};
> > > +	static struct coef_fw coef0286[] = {
> > > +		//UPDATE_COEF(0x4f, 0xfcc0, 0x5400), /* Combo 
> > Jack auto detect */
> > > +		UPDATE_COEF(0x4f, 0xfcc0, 0xd400),
> > 
> > Why is this commented out and what's the next line does instead?
> > And it's 0286?  Not 0288?
> > 
> > > @@ -3949,8 +4027,15 @@ static void 
> > alc_update_headset_jack_cb(struct hda_codec *codec,
> > >  				       struct hda_jack_callback 
> > *jack)  {
> > >  	struct alc_spec *spec = codec->spec;
> > > +	int present;
> > >  	spec->current_headset_type = ALC_HEADSET_TYPE_UNKNOWN;
> > >  	snd_hda_gen_hp_automute(codec, jack);
> > > +	present = spec->gen.hp_jack_present ? 0x40 : 0;
> > > +	if (codec->bus->pci->subsystem_vendor == 0x1028 &&
> > > +				codec->core.vendor_id == 0x10ec0288)
> > > +		/* Headset Mic enable and disable, ony for Dell Dino */
> > > +		snd_hda_codec_write(codec, 0x01, 0, 
> > AC_VERB_SET_GPIO_DATA,
> > > +					present);
> > 
> > This is too ugly to put in the common function.
> > For Dell, prepare its own callback function and use it 
> > instead.  Or, put a new hook there to call additionally.  
> > This avoids the PCI ID check in each place.
> > 
> > So, better to split to a few patches.  One for adding ALC286 
> > (or 288?) support generically.  And another for adding 
> > Dell-specific things.
> > 
> > 
> > thanks,
> > 
> > Takashi
> > 
> > ------Please consider the environment before printing this e-mail.
> > 
> [2 0000-alc288-headset-mode.patch <application/octet-stream (base64)>]
> 

Re: Support Dell headset mode for ALC288

[Kailang]

> -----Original Message-----
> From: Takashi Iwai [mailto:tiwai@suse.de] 
> Sent: Wednesday, April 08, 2015 3:40 PM
> To: Kailang
> Cc: (alsa-devel@alsa-project.org)
> Subject: Re: Support Dell headset mode for ALC288
> 
> At Wed, 8 Apr 2015 07:31:33 +0000,
> Kailang wrote:
> > 
> > 
> > Hi Takashi,
> > 
> > Attach patch is support alc288 headset mode.
> 
> This also supports ALC286, too, right?

Yes, they are brother.

> 
> 
> Takashi
> 
> > 
> > BR,
> > Kailang
> > 
> > > -----Original Message-----
> > > From: Takashi Iwai [mailto:tiwai@suse.de]
> > > Sent: Saturday, April 04, 2015 8:43 PM
> > > To: Kailang
> > > Cc: (alsa-devel@alsa-project.org)
> > > Subject: Re: Support Dell headset mode for ALC288
> > > 
> > > At Tue, 31 Mar 2015 08:01:57 +0000,
> > > Kailang wrote:
> > > > 
> > > > 
> > > > Hi Takashi,
> > > >   
> > > >  Dell has new platform with ALC288.
> > > >  Attach patch is support headset mode for ALC288.
> > > 
> > > > @@ -3828,6 +3894,11 @@ static void
> > > alc_determine_headset_type(struct hda_codec *codec)
> > > >   conteol) */
> > > >  		{}
> > > >  	};
> > > > +	static struct coef_fw coef0286[] = {
> > > > +		//UPDATE_COEF(0x4f, 0xfcc0, 0x5400), /* Combo
> > > Jack auto detect */
> > > > +		UPDATE_COEF(0x4f, 0xfcc0, 0xd400),
> > > 
> > > Why is this commented out and what's the next line does instead?
> > > And it's 0286?  Not 0288?
> > > 
> > > > @@ -3949,8 +4027,15 @@ static void
> > > alc_update_headset_jack_cb(struct hda_codec *codec,
> > > >  				       struct hda_jack_callback
> > > *jack)  {
> > > >  	struct alc_spec *spec = codec->spec;
> > > > +	int present;
> > > >  	spec->current_headset_type = ALC_HEADSET_TYPE_UNKNOWN;
> > > >  	snd_hda_gen_hp_automute(codec, jack);
> > > > +	present = spec->gen.hp_jack_present ? 0x40 : 0;
> > > > +	if (codec->bus->pci->subsystem_vendor == 0x1028 &&
> > > > +				codec->core.vendor_id 
> == 0x10ec0288)
> > > > +		/* Headset Mic enable and disable, ony 
> for Dell Dino */
> > > > +		snd_hda_codec_write(codec, 0x01, 0,
> > > AC_VERB_SET_GPIO_DATA,
> > > > +					present);
> > > 
> > > This is too ugly to put in the common function.
> > > For Dell, prepare its own callback function and use it 
> instead.  Or, 
> > > put a new hook there to call additionally.
> > > This avoids the PCI ID check in each place.
> > > 
> > > So, better to split to a few patches.  One for adding ALC286 (or 
> > > 288?) support generically.  And another for adding Dell-specific 
> > > things.
> > > 
> > > 
> > > thanks,
> > > 
> > > Takashi
> > > 
> > > ------Please consider the environment before printing this e-mail.
> > > 
> > [2 0000-alc288-headset-mode.patch <application/octet-stream 
> (base64)>]
> > 
> 

Re: Support Dell headset mode for ALC288

[Takashi Iwai]

At Wed, 8 Apr 2015 08:08:47 +0000,
Kailang wrote:
> 
> 
> > -----Original Message-----
> > From: Takashi Iwai [mailto:tiwai@suse.de] 
> > Sent: Wednesday, April 08, 2015 3:40 PM
> > To: Kailang
> > Cc: (alsa-devel@alsa-project.org)
> > Subject: Re: Support Dell headset mode for ALC288
> > 
> > At Wed, 8 Apr 2015 07:31:33 +0000,
> > Kailang wrote:
> > > 
> > > 
> > > Hi Takashi,
> > > 
> > > Attach patch is support alc288 headset mode.
> > 
> > This also supports ALC286, too, right?
> 
> Yes, they are brother.

OK, I applied this with rephrasing the changelog mentioning ALC286,
too.

Since this one and Dell patch are relative large changes, I queued
them to for-next branch instead of for-linus, i.e. the merge to Linus
tree will be postponed to 4.1-rc1.


thanks,

Takashi

> 
> > 
> > 
> > Takashi
> > 
> > > 
> > > BR,
> > > Kailang
> > > 
> > > > -----Original Message-----
> > > > From: Takashi Iwai [mailto:tiwai@suse.de]
> > > > Sent: Saturday, April 04, 2015 8:43 PM
> > > > To: Kailang
> > > > Cc: (alsa-devel@alsa-project.org)
> > > > Subject: Re: Support Dell headset mode for ALC288
> > > > 
> > > > At Tue, 31 Mar 2015 08:01:57 +0000,
> > > > Kailang wrote:
> > > > > 
> > > > > 
> > > > > Hi Takashi,
> > > > >   
> > > > >  Dell has new platform with ALC288.
> > > > >  Attach patch is support headset mode for ALC288.
> > > > 
> > > > > @@ -3828,6 +3894,11 @@ static void
> > > > alc_determine_headset_type(struct hda_codec *codec)
> > > > >   conteol) */
> > > > >  		{}
> > > > >  	};
> > > > > +	static struct coef_fw coef0286[] = {
> > > > > +		//UPDATE_COEF(0x4f, 0xfcc0, 0x5400), /* Combo
> > > > Jack auto detect */
> > > > > +		UPDATE_COEF(0x4f, 0xfcc0, 0xd400),
> > > > 
> > > > Why is this commented out and what's the next line does instead?
> > > > And it's 0286?  Not 0288?
> > > > 
> > > > > @@ -3949,8 +4027,15 @@ static void
> > > > alc_update_headset_jack_cb(struct hda_codec *codec,
> > > > >  				       struct hda_jack_callback
> > > > *jack)  {
> > > > >  	struct alc_spec *spec = codec->spec;
> > > > > +	int present;
> > > > >  	spec->current_headset_type = ALC_HEADSET_TYPE_UNKNOWN;
> > > > >  	snd_hda_gen_hp_automute(codec, jack);
> > > > > +	present = spec->gen.hp_jack_present ? 0x40 : 0;
> > > > > +	if (codec->bus->pci->subsystem_vendor == 0x1028 &&
> > > > > +				codec->core.vendor_id 
> > == 0x10ec0288)
> > > > > +		/* Headset Mic enable and disable, ony 
> > for Dell Dino */
> > > > > +		snd_hda_codec_write(codec, 0x01, 0,
> > > > AC_VERB_SET_GPIO_DATA,
> > > > > +					present);
> > > > 
> > > > This is too ugly to put in the common function.
> > > > For Dell, prepare its own callback function and use it 
> > instead.  Or, 
> > > > put a new hook there to call additionally.
> > > > This avoids the PCI ID check in each place.
> > > > 
> > > > So, better to split to a few patches.  One for adding ALC286 (or 
> > > > 288?) support generically.  And another for adding Dell-specific 
> > > > things.
> > > > 
> > > > 
> > > > thanks,
> > > > 
> > > > Takashi
> > > > 
> > > > ------Please consider the environment before printing this e-mail.
> > > > 
> > > [2 0000-alc288-headset-mode.patch <application/octet-stream 
> > (base64)>]
> > > 
> > 

Re: Support Dell headset mode for ALC288

[Takashi Iwai]

At Wed, 8 Apr 2015 08:13:21 +0000,
Kailang wrote:
> 
> 
> Hi Takashi,
> 
> Attach patch is Dell specific for headset mode.
> I splited it and done some modification.

The patch doesn't compile.  I fixed it manually and also did the
simple refactoring as below.  Now it's applied to for-next branch.


thanks,

Takashi

-- 8< --
From: Kailang Yang <kailang@realtek.com>
Subject: [PATCH] ALSA: hda/realtek - Support Dell headset mode for ALC288

Dell create new platform with ALC288 codec.
This patch will enable headset mode for Dino platform.

[slight code refactoring and compile fix by tiwai]

Signed-off-by: Kailang Yang <kailang@realtek.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
---
 sound/pci/hda/patch_realtek.c | 67 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 67 insertions(+)

diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c
index ceb599a87803..7b5c93e0e78c 100644
--- a/sound/pci/hda/patch_realtek.c
+++ b/sound/pci/hda/patch_realtek.c
@@ -4121,6 +4121,29 @@ static void alc_fixup_headset_mode_alc255_no_hp_mic(struct hda_codec *codec,
 		alc_fixup_headset_mode(codec, fix, action);
 }
 
+static void alc288_update_headset_jack_cb(struct hda_codec *codec,
+				       struct hda_jack_callback *jack)
+{
+	struct alc_spec *spec = codec->spec;
+	int present;
+
+	alc_update_headset_jack_cb(codec, jack);
+	/* Headset Mic enable or disable, only for Dell Dino */
+	present = spec->gen.hp_jack_present ? 0x40 : 0;
+	snd_hda_codec_write(codec, 0x01, 0, AC_VERB_SET_GPIO_DATA,
+				present);
+}
+
+static void alc_fixup_headset_mode_dell_alc288(struct hda_codec *codec,
+				const struct hda_fixup *fix, int action)
+{
+	alc_fixup_headset_mode(codec, fix, action);
+	if (action == HDA_FIXUP_ACT_PROBE) {
+		struct alc_spec *spec = codec->spec;
+		spec->gen.hp_automute_hook = alc288_update_headset_jack_cb;
+	}
+}
+
 static void alc_fixup_auto_mute_via_amp(struct hda_codec *codec,
 					const struct hda_fixup *fix, int action)
 {
@@ -4460,6 +4483,9 @@ enum {
 	ALC286_FIXUP_HP_GPIO_LED,
 	ALC280_FIXUP_HP_GPIO2_MIC_HOTKEY,
 	ALC280_FIXUP_HP_DOCK_PINS,
+	ALC288_FIXUP_DELL_HEADSET_MODE,
+	ALC288_FIXUP_DELL1_MIC_NO_PRESENCE,
+	ALC288_FIXUP_DELL_XPS_13_GPIO6,
 };
 
 static const struct hda_fixup alc269_fixups[] = {
@@ -4948,6 +4974,33 @@ static const struct hda_fixup alc269_fixups[] = {
 		.chained = true,
 		.chain_id = ALC280_FIXUP_HP_GPIO4
 	},
+	[ALC288_FIXUP_DELL_HEADSET_MODE] = {
+		.type = HDA_FIXUP_FUNC,
+		.v.func = alc_fixup_headset_mode_dell_alc288,
+		.chained = true,
+		.chain_id = ALC255_FIXUP_DELL_WMI_MIC_MUTE_LED
+	},
+	[ALC288_FIXUP_DELL1_MIC_NO_PRESENCE] = {
+		.type = HDA_FIXUP_PINS,
+		.v.pins = (const struct hda_pintbl[]) {
+			{ 0x18, 0x01a1913c }, /* use as headset mic, without its own jack detect */
+			{ 0x1a, 0x01a1913d }, /* use as headphone mic, without its own jack detect */
+			{ }
+		},
+		.chained = true,
+		.chain_id = ALC288_FIXUP_DELL_HEADSET_MODE
+	},
+	[ALC288_FIXUP_DELL_XPS_13_GPIO6] = {
+		.type = HDA_FIXUP_VERBS,
+		.v.verbs = (const struct hda_verb[]) {
+			{0x01, AC_VERB_SET_GPIO_MASK, 0x40},
+			{0x01, AC_VERB_SET_GPIO_DIRECTION, 0x40},
+			{0x01, AC_VERB_SET_GPIO_DATA, 0x00},
+			{ }
+		},
+		.chained = true,
+		.chain_id = ALC288_FIXUP_DELL1_MIC_NO_PRESENCE
+	},
 };
 
 static const struct snd_pci_quirk alc269_fixup_tbl[] = {
@@ -5174,6 +5227,13 @@ static const struct hda_model_fixup alc269_fixup_models[] = {
 	{0x1b, 0x411111f0}, \
 	{0x1e, 0x411111f0}
 
+#define ALC288_STANDARD_PINS \
+	{0x17, 0x411111f0}, \
+	{0x18, 0x411111f0}, \
+	{0x19, 0x411111f0}, \
+	{0x1a, 0x411111f0}, \
+	{0x1e, 0x411111f0}
+
 #define ALC290_STANDARD_PINS \
 	{0x12, 0x99a30130}, \
 	{0x13, 0x40000000}, \
@@ -5369,6 +5429,13 @@ static const struct snd_hda_pin_quirk alc269_pin_fixup_tbl[] = {
 		{0x19, 0x03a11020},
 		{0x1d, 0x40e00001},
 		{0x21, 0x0321101f}),
+	SND_HDA_PIN_QUIRK(0x10ec0288, 0x1028, "Dell", ALC288_FIXUP_DELL_XPS_13_GPIO6,
+		ALC288_STANDARD_PINS,
+		{0x12, 0x90a60120},
+		{0x13, 0x40000000},
+		{0x14, 0x90170110},
+		{0x1d, 0x4076832d},
+		{0x21, 0x0321101f}),
 	SND_HDA_PIN_QUIRK(0x10ec0290, 0x103c, "HP", ALC269_FIXUP_HP_MUTE_LED_MIC1,
 		ALC290_STANDARD_PINS,
 		{0x14, 0x411111f0},
-- 
2.3.4

Re: Support Dell headset mode for ALC288

[Mario Limonciello]

On 04/08/2015 04:40 AM, Takashi Iwai wrote:
> At Wed, 8 Apr 2015 08:08:47 +0000,
> Kailang wrote:
> OK, I applied this with rephrasing the changelog mentioning ALC286,
> too.
>
> Since this one and Dell patch are relative large changes, I queued
> them to for-next branch instead of for-linus, i.e. the merge to Linus
> tree will be postponed to 4.1-rc1.
>
>
> thanks,
>
> Takashi
>
>
Takashi,

It's a bit unfortunate that these are landing in 4.1 rather than 4.0.  4.1 they actually won't be very useful in the default kernel due to this patch landing:
https://github.com/torvalds/linux/commit/b1ef2972586577e0ca9675254ee141f65a8824e5

With 4.1 the XPS 13 will run in I2S audio mode.

Is there any chance at backporting these to @stable?  Some of the distros (Ubuntu and Fedora) are already backporting this into some of their kernel versions, but I think this should benefit the community at large if the other distributions that tracked @stable could pick this up too.

Thanks,

Re: Support Dell headset mode for ALC288

[Takashi Iwai]

At Mon, 20 Apr 2015 15:12:59 -0500,
Mario Limonciello wrote:
> 
> 
> On 04/08/2015 04:40 AM, Takashi Iwai wrote:
> > At Wed, 8 Apr 2015 08:08:47 +0000,
> > Kailang wrote:
> > OK, I applied this with rephrasing the changelog mentioning ALC286,
> > too.
> >
> > Since this one and Dell patch are relative large changes, I queued
> > them to for-next branch instead of for-linus, i.e. the merge to Linus
> > tree will be postponed to 4.1-rc1.
> >
> >
> > thanks,
> >
> > Takashi
> >
> >
> Takashi,
> 
> It's a bit unfortunate that these are landing in 4.1 rather than 4.0.  4.1 they actually won't be very useful in the default kernel due to this patch landing:
> https://github.com/torvalds/linux/commit/b1ef2972586577e0ca9675254ee141f65a8824e5
> 
> With 4.1 the XPS 13 will run in I2S audio mode.
> 
> Is there any chance at backporting these to @stable?  Some of the distros (Ubuntu and Fedora) are already backporting this into some of their kernel versions, but I think this should benefit the community at large if the other distributions that tracked @stable could pick this up too.

It's easy, you just need to send a post to stable kernel to tell the
commit id to cherry-pick.

At best it should be done by the people who actually test the
backport, as this is no generic fix I can confirm.


Takashi