Skip to contentSkip to conversation

Tests of the hda-jack branch

PATCH David Henningsson 5 participants 11 messages Nov 29, 2011 – Nov 30, 2011 Sound

lore
Share & tools
lore
#0 lore raw

Hi Takashi,

I've done some more tests of your branch now, and the new names seems to work for my machine here.

I'm attaching two patches - one adds a "device index" for mapping against pcm devices on HDMI, the other one is a backport for the driver-build tree.

What are your thoughts on how to proceed, when do you think this is ready to be merged into the topic/hda branch?

// David

Quoted 1 line

From 0f14d27ffebf72b636c765e95003c780a1cb8c07 Mon Sep 17 00:00:00 2001

From: David Henningsson <[email protected]>
Date: Tue, 29 Nov 2011 19:41:27 +0100
Subject: [PATCH] alsa-driver-build: Make new jack API compile on older
kernels

---
 acore/ctljack.c    |    4 ++++
 pci/hda/hda_jack.c |    3 +++
 2 files changed, 7 insertions(+), 0 deletions(-)
 create mode 100644 acore/ctljack.c
 create mode 100644 pci/hda/hda_jack.c
acore/ctljack.c new file +4 -0
new file mode 100644
index 0000000..4d9352f
@@ -0,0 +1,4 @@
+#define __NO_VERSION__
+#include "adriver.h"
+#include "../alsa-kernel/core/ctljack.c"
+
pci/hda/hda_jack.c new file +3 -0
new file mode 100644
index 0000000..160ad73
@@ -0,0 +1,3 @@
+#include "adriver.h"
+#include "../../alsa-kernel/pci/hda/hda_jack.c"
+
Signature · 1 line
1.7.5.4
Quoted 1 line

From 4b4c17927eb05b460a9788d9dfc9a03ace5d160e Mon Sep 17 00:00:00 2001

From: David Henningsson <[email protected]>
Date: Tue, 29 Nov 2011 20:19:56 +0100
Subject: [PATCH] ALSA: HDA/Jack: Make device indices for HDMI

Make sure we can associate HDMI Jacks with the correct PCM devices.

---
 sound/pci/hda/hda_jack.c   |    5 +++--
 sound/pci/hda/hda_jack.h   |    2 +-
 sound/pci/hda/patch_hdmi.c |    3 ++-
 3 files changed, 6 insertions(+), 4 deletions(-)
sound/pci/hda/hda_jack.c +3 -2
index 3949015..0b1f72c 100644
@@ -210,7 +210,7 @@ EXPORT_SYMBOL_HDA(snd_hda_jack_report_sync);
  * will have the given name and index.
  */
 int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-			  const char *name, int idx)
+			  const char *name, int idx, int device)
 {
 	struct hda_jack_tbl *jack;
 	struct snd_kcontrol *kctl;
@@ -223,6 +223,7 @@ int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
 	kctl = snd_kctl_jack_new(name, idx, codec);
 	if (!kctl)
 		return -ENOMEM;
+	kctl->id.device = device;
 	if (snd_hda_ctl_add(codec, nid, kctl) < 0)
 		return -ENOMEM;
 	jack->kctl = kctl;
@@ -249,7 +250,7 @@ static int add_jack_kctl(struct hda_codec *codec, hda_nid_t nid,
 		return 0;
 
 	snd_hda_get_pin_label(codec, nid, cfg, name, sizeof(name), &idx);
-	err = snd_hda_jack_add_kctl(codec, nid, name, idx);
+	err = snd_hda_jack_add_kctl(codec, nid, name, idx, 0);
 	if (err < 0)
 		return err;
 	return snd_hda_jack_detect_enable(codec, nid, 0);
sound/pci/hda/hda_jack.h +1 -1
index f8f97c7..fcca96c 100644
@@ -76,7 +76,7 @@ static inline bool is_jack_detectable(struct hda_codec *codec, hda_nid_t nid)
 }
 
 int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-			  const char *name, int idx);
+			  const char *name, int idx, int device);
 int snd_hda_jack_add_kctls(struct hda_codec *codec,
 			   const struct auto_pin_cfg *cfg);
 
sound/pci/hda/patch_hdmi.c +2 -1
index bb8cfc6..9689c8d 100644
@@ -1277,7 +1277,8 @@ static int generic_hdmi_build_controls(struct hda_codec *codec)
 		if (err < 0)
 			return err;
 		err = snd_hda_jack_add_kctl(codec, per_pin->pin_nid,
-					    "HDMI", pin_idx);
+					    "HDMI", pin_idx,
+					    spec->pcm_rec[pin_idx].device);
 		if (err < 0)
 			return err;
 	}
Signature · 1 line
1.7.5.4
#1 lore raw
Quoted 2 lines · David Henningsson

What are your thoughts on how to proceed, when do you think this is ready to be merged into the topic/hda branch?

I'm still rather sad that this infrastructure is all totally separate to the existing jack infrastructure, one of the goals of having common infrastructure was to insulate drivers from having to deal with changes in the APIs used for reporting and I certainly don't want to have to have all the drivers individually reporting via all the APIs (especially not with the multiswitch API coming along too).

I've not really reviewed the code beyond noticing that, I'd been waiting for it to turn up on the list.

#4 lore raw
In reply to David Henningsson

Review

Quoted 4 lines · David Henningsson

... two patches - one adds a "device index" for mapping against pcm devices on HDMI,

(quoted diff) +1 -0
+	kctl->id.device = device;

In the general case, there is not always a 1:1 mapping between jacks and PCM devices, so I fear that a jack belonging to device 0 cannot be distinguished from a jack that has no specific device.

There is a convention that id.device is a valid PCM device number if and only if id.iface is IFACE_PCM. However, it might be useful to map controls with IFACE_MIXER to PCM devices, so I'd propose a new TLV, like in the incomplete patch below.

Regards,
Clemens

include/sound/tlv.h +6 -0
@@ -37,6 +37,7 @@
 #define SNDRV_CTL_TLVT_DB_RANGE 3	/* dB range container */
 #define SNDRV_CTL_TLVT_DB_MINMAX 4	/* dB scale with min/max */
 #define SNDRV_CTL_TLVT_DB_MINMAX_MUTE 5	/* dB scale with min/max with mute */
+#define SNDRV_CTL_TLVT_PCM_DEVICE 6	/* jack (or other control?) maps to this device */
 
 #define _TLV_ITEM(type, ...) \
 	(type), _TLV_LENGTH(__VA_ARGS__), __VA_ARGS__
@@ -81,4 +82,9 @@
 
 #define TLV_DB_GAIN_MUTE	-9999999
 
+#define TLB_PCM_DEVICE_NUMBER_ITEM(devno) \
+	SNDRV_CTL_TLVT_PCM_DEVICE, 1 * sizeof(unsigned int), (devno)
+#define DECLARE_TLV_PCM_DEVICE_NUMBER(name, devno) \
+	unsigned int name[] = { TLV_PCM_DEVICE_NUMBER_ITEM(devno) }
+
 #endif /* __SOUND_TLV_H */
sound/pci/hda/hda_jack.c +8 -2
@@ -15,6 +15,7 @@
 #include <sound/core.h>
 #include <sound/control.h>
 #include <sound/jack.h>
+#include <sound/tlv.h>
 #include "hda_codec.h"
 #include "hda_local.h"
 #include "hda_jack.h"
@@ -210,7 +211,7 @@ EXPORT_SYMBOL_HDA(snd_hda_jack_report_sync);
  * will have the given name and index.
  */
 int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-			  const char *name, int idx)
+			  const char *name, int idx, int device)
 {
 	struct hda_jack_tbl *jack;
 	struct snd_kcontrol *kctl;
@@ -223,6 +224,11 @@ int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
 	kctl = snd_kctl_jack_new(name, idx, codec);
 	if (!kctl)
 		return -ENOMEM;
+	jack->tlv[0] = SNDRV_CTL_TLVT_PCM_DEVICE;
+	jack->tlv[1] = 1 * sizeof(unsigned int);
+	jack->tlv[2] = device;
+	kctl->tlv.p = jack->tlv;
+	kctl->access |= SNDRV_CTL_ELEM_ACCESS_TLV_READ;
 	if (snd_hda_ctl_add(codec, nid, kctl) < 0)
 		return -ENOMEM;
 	jack->kctl = kctl;
@@ -249,7 +255,7 @@ static int add_jack_kctl(struct hda_codec *codec, hda_nid_t nid,
 		return 0;
 
 	snd_hda_get_pin_label(codec, nid, cfg, name, sizeof(name), &idx);
-	err = snd_hda_jack_add_kctl(codec, nid, name, idx);
+	err = snd_hda_jack_add_kctl(codec, nid, name, idx, 0);
 	if (err < 0)
 		return err;
 	return snd_hda_jack_detect_enable(codec, nid, 0);
sound/pci/hda/hda_jack.h +2 -1
@@ -22,6 +22,7 @@ struct hda_jack_tbl {
 	unsigned int jack_detect:1;	/* capable of jack-detection? */
 	unsigned int jack_dirty:1;	/* needs to update? */
 	struct snd_kcontrol *kctl;	/* assigned kctl for jack-detection */
+	unsigned int tlv[3];
 #ifdef CONFIG_SND_HDA_INPUT_JACK
 	int type;
 	struct snd_jack *jack;
@@ -76,7 +77,7 @@ static inline bool is_jack_detectable(struct hda_codec *codec, hda_nid_t nid)
 }
 
 int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-			  const char *name, int idx);
+			  const char *name, int idx, int device);
 int snd_hda_jack_add_kctls(struct hda_codec *codec,
 			   const struct auto_pin_cfg *cfg);
#5 lore raw

2011/11/30 Clemens Ladisch <[email protected]>:

Quoted 91 lines
Quoted 4 lines · David Henningsson

... two patches - one adds a "device index" for mapping against pcm devices on HDMI,

+     kctl->id.device = device;

In the general case, there is not always a 1:1 mapping between jacks and PCM devices, so I fear that a jack belonging to device 0 cannot be distinguished from a jack that has no specific device.

There is a convention that id.device is a valid PCM device number if and only if id.iface is IFACE_PCM.  However, it might be useful to map controls with IFACE_MIXER to PCM devices, so I'd propose a new TLV, like in the incomplete patch below.

Regards,
Clemens

include/sound/tlv.h +1 -0
@@ -37,6 +37,7 @@
  #define SNDRV_CTL_TLVT_DB_RANGE 3      /* dB range container */
  #define SNDRV_CTL_TLVT_DB_MINMAX 4     /* dB scale with min/max */
  #define SNDRV_CTL_TLVT_DB_MINMAX_MUTE 5        /* dB scale with min/max with mute */
+#define SNDRV_CTL_TLVT_PCM_DEVICE 6    /* jack (or other control?) maps to this device */
 
  #define _TLV_ITEM(type, ...) \

       (type), _TLV_LENGTH(__VA_ARGS__), __VA_ARGS__

include/sound/tlv.h +5 -0
@@ -81,4 +82,9 @@
 
  #define TLV_DB_GAIN_MUTE       -9999999
 
+#define TLB_PCM_DEVICE_NUMBER_ITEM(devno) \
+       SNDRV_CTL_TLVT_PCM_DEVICE, 1 * sizeof(unsigned int), (devno)
+#define DECLARE_TLV_PCM_DEVICE_NUMBER(name, devno) \
+       unsigned int name[] = { TLV_PCM_DEVICE_NUMBER_ITEM(devno) }
+
  #endif /* __SOUND_TLV_H */
sound/pci/hda/hda_jack.c +8 -2
@@ -15,6 +15,7 @@
  #include <sound/core.h>
  #include <sound/control.h>
  #include <sound/jack.h>
+#include <sound/tlv.h>
  #include "hda_codec.h"
  #include "hda_local.h"
  #include "hda_jack.h"
@@ -210,7 +211,7 @@ EXPORT_SYMBOL_HDA(snd_hda_jack_report_sync);
  * will have the given name and index.
  */
  int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-                         const char *name, int idx)
+                         const char *name, int idx, int device)
  {
        struct hda_jack_tbl *jack;
        struct snd_kcontrol *kctl;
@@ -223,6 +224,11 @@ int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
        kctl = snd_kctl_jack_new(name, idx, codec);
        if (!kctl)
                return -ENOMEM;
+       jack->tlv[0] = SNDRV_CTL_TLVT_PCM_DEVICE;
+       jack->tlv[1] = 1 * sizeof(unsigned int);
+       jack->tlv[2] = device;
+       kctl->tlv.p = jack->tlv;
+       kctl->access |= SNDRV_CTL_ELEM_ACCESS_TLV_READ;
        if (snd_hda_ctl_add(codec, nid, kctl) < 0)
                return -ENOMEM;
        jack->kctl = kctl;
@@ -249,7 +255,7 @@ static int add_jack_kctl(struct hda_codec *codec, hda_nid_t nid,
                return 0;
 
        snd_hda_get_pin_label(codec, nid, cfg, name, sizeof(name), &idx);
-       err = snd_hda_jack_add_kctl(codec, nid, name, idx);
+       err = snd_hda_jack_add_kctl(codec, nid, name, idx, 0);
        if (err < 0)
                return err;
        return snd_hda_jack_detect_enable(codec, nid, 0);
sound/pci/hda/hda_jack.h +0 -0
@@ -22,6 +22,7 @@ struct hda_jack_tbl {

       unsigned int jack_detect:1;     /* capable of jack-detection? */ unsigned int jack_dirty:1;      /* needs to update? */ struct snd_kcontrol *kctl;      /* assigned kctl for jack-detection */
+       unsigned int tlv[3];
 #ifdef CONFIG_SND_HDA_INPUT_JACK
       int type;
       struct snd_jack *jack;

sound/pci/hda/hda_jack.h static inline bool is_jack_detectable(struct hda_codec *codec, hda_nid_t nid) · line 83

Quoted 8 lines
sound/pci/hda/hda_jack.h +1 -1
@@ -76,7 +77,7 @@ static inline bool is_jack_detectable(struct hda_codec *codec, hda_nid_t nid)
  }
 
  int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-                         const char *name, int idx);
+                         const char *name, int idx, int device);
  int snd_hda_jack_add_kctls(struct hda_codec *codec,
                           const struct auto_pin_cfg *cfg);
 

Does this solve the problem of selecting the correct capture source when the codec have more than one adc

1) realek codec with two adc , capture and alt_capture device

ARECORD

**** List of CAPTURE Hardware Devices ****
card 0: PCH [HDA Intel PCH], device 0: ALC892 Analog [ALC892 Analog]
Subdevices: 1/1
Subdevice #0: subdevice #0
card 0: PCH [HDA Intel PCH], device 2: ALC892 Analog [ALC892 Analog]
Subdevices: 1/1
Subdevice #0: subdevice #0

Simple mixer control 'Input Source',0
Capabilities: cenum
Items: 'Mic' 'Internal Mic' 'Line'
Item0: 'Internal Mic'
Simple mixer control 'Input Source',1
Capabilities: cenum
Items: 'Mic' 'Internal Mic' 'Line'
Item0: 'Mic'

2) ad198x codec with 3 adc and 3 capture subdevices

Line-Out Front Jack
Line-Out Surround Jack
Line-Out CLFE Jack
Line-Out Side Jack
Front Headphone Jack
Front Mic Jack
Rear Mic Jack
Line Jack

card 1: Intel [HDA Intel], device 0: AD198x Analog [AD198x Analog]
Subdevices: 3/3
Subdevice #0: subdevice #0
Subdevice #1: subdevice #1
Subdevice #2: subdevice #2

Simple mixer control 'Input Source',0
Capabilities: cenum
Items: 'Mix' 'Front Mic' 'Rear Mic' 'Line' 'CD'
Item0: 'Front Mic'
Simple mixer control 'Input Source',1
Capabilities: cenum
Items: 'Mix' 'Front Mic' 'Rear Mic' 'Line' 'CD'
Item0: 'Front Mic'
Simple mixer control 'Input Source',2
Capabilities: cenum
Items: 'Mix' 'Front Mic' 'Rear Mic' 'Line' 'CD'
Item0: 'Line'

#8 lore raw
In reply to David Henningsson

2011/11/30 David Henningsson <[email protected]>:

Quoted 5 lines

Hi Takashi,

I've done some more tests of your branch now, and the new names seems to work for my machine here.

Quoted 8 lines

I'm attaching two patches - one adds a "device index" for mapping against pcm devices on HDMI, the other one is a backport for the driver-build tree.

What are your thoughts on how to proceed, when do you think this is ready to be merged into the topic/hda branch?

// David

http://git.alsa-project.org/?p=alsa-kernel.git;a=blob_plain;f=Documentation/sound/alsa/HD-Audio-Controls.txt

Does it handle "Independent HP" ?

VIA codecs
----------

* Independent HP
When this enum control is enabled, the headphone output is routed from an individual stream (the third PCM such as hw:0,2) instead of the primary stream. In the case the headphone DAC is shared with a side or a CLFE-channel DAC, the DAC is switched to the headphone automatically.

Analog codecs
--------------

* Independent HP
When this enum control is enabled, the headphone output is routed from an individual stream (the third PCM such as hw:0,2) instead of the primary stream.

#10 lore raw
In reply to David Henningsson

At Tue, 29 Nov 2011 20:43:06 +0100,

Quoted 9 lines · David Henningsson

Hi Takashi,

I've done some more tests of your branch now, and the new names seems to work for my machine here.

I'm attaching two patches - one adds a "device index" for mapping against pcm devices on HDMI, the other one is a backport for the driver-build tree.

Thanks, the alsa-driver-build one was already found in alsa-driver-build-unstable tree. Actually the branch is included in sound-unstable tree, too.

Quoted 2 lines

What are your thoughts on how to proceed, when do you think this is ready to be merged into the topic/hda branch?

We still need to think of PCM and control-NID mapping. At least, decide what to be done.

thanks,

Takashi

Quoted 118 lines

// David

[2 0001-alsa-driver-build-Make-new-jack-API-compile-on-older.patch <text/x-patch (7bit)>]

Quoted 1 line

From 0f14d27ffebf72b636c765e95003c780a1cb8c07 Mon Sep 17 00:00:00 2001

From: David Henningsson <[email protected]>
Date: Tue, 29 Nov 2011 19:41:27 +0100
Subject: [PATCH] alsa-driver-build: Make new jack API compile on older
kernels

---
 acore/ctljack.c    |    4 ++++
 pci/hda/hda_jack.c |    3 +++
 2 files changed, 7 insertions(+), 0 deletions(-)
 create mode 100644 acore/ctljack.c
 create mode 100644 pci/hda/hda_jack.c
acore/ctljack.c new file +4 -0
new file mode 100644
index 0000000..4d9352f
@@ -0,0 +1,4 @@
+#define __NO_VERSION__
+#include "adriver.h"
+#include "../alsa-kernel/core/ctljack.c"
+
pci/hda/hda_jack.c new file +3 -0
new file mode 100644
index 0000000..160ad73
@@ -0,0 +1,3 @@
+#include "adriver.h"
+#include "../../alsa-kernel/pci/hda/hda_jack.c"
+
Signature · 5 lines
1.7.5.4



[3 0001-ALSA-HDA-Jack-Make-device-indices-for-HDMI.patch <text/x-patch (7bit)>]
Quoted 1 line

From 4b4c17927eb05b460a9788d9dfc9a03ace5d160e Mon Sep 17 00:00:00 2001

From: David Henningsson <[email protected]>
Date: Tue, 29 Nov 2011 20:19:56 +0100
Subject: [PATCH] ALSA: HDA/Jack: Make device indices for HDMI

Make sure we can associate HDMI Jacks with the correct PCM devices.

---
 sound/pci/hda/hda_jack.c   |    5 +++--
 sound/pci/hda/hda_jack.h   |    2 +-
 sound/pci/hda/patch_hdmi.c |    3 ++-
 3 files changed, 6 insertions(+), 4 deletions(-)
sound/pci/hda/hda_jack.c +3 -2
index 3949015..0b1f72c 100644
@@ -210,7 +210,7 @@ EXPORT_SYMBOL_HDA(snd_hda_jack_report_sync);
  * will have the given name and index.
  */
 int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-			  const char *name, int idx)
+			  const char *name, int idx, int device)
 {
 	struct hda_jack_tbl *jack;
 	struct snd_kcontrol *kctl;
@@ -223,6 +223,7 @@ int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
 	kctl = snd_kctl_jack_new(name, idx, codec);
 	if (!kctl)
 		return -ENOMEM;
+	kctl->id.device = device;
 	if (snd_hda_ctl_add(codec, nid, kctl) < 0)
 		return -ENOMEM;
 	jack->kctl = kctl;
@@ -249,7 +250,7 @@ static int add_jack_kctl(struct hda_codec *codec, hda_nid_t nid,
 		return 0;
 
 	snd_hda_get_pin_label(codec, nid, cfg, name, sizeof(name), &idx);
-	err = snd_hda_jack_add_kctl(codec, nid, name, idx);
+	err = snd_hda_jack_add_kctl(codec, nid, name, idx, 0);
 	if (err < 0)
 		return err;
 	return snd_hda_jack_detect_enable(codec, nid, 0);
sound/pci/hda/hda_jack.h +1 -1
index f8f97c7..fcca96c 100644
@@ -76,7 +76,7 @@ static inline bool is_jack_detectable(struct hda_codec *codec, hda_nid_t nid)
 }
 
 int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-			  const char *name, int idx);
+			  const char *name, int idx, int device);
 int snd_hda_jack_add_kctls(struct hda_codec *codec,
 			   const struct auto_pin_cfg *cfg);
 
sound/pci/hda/patch_hdmi.c +2 -1
index bb8cfc6..9689c8d 100644
@@ -1277,7 +1277,8 @@ static int generic_hdmi_build_controls(struct hda_codec *codec)
 		if (err < 0)
 			return err;
 		err = snd_hda_jack_add_kctl(codec, per_pin->pin_nid,
-					    "HDMI", pin_idx);
+					    "HDMI", pin_idx,
+					    spec->pcm_rec[pin_idx].device);
 		if (err < 0)
 			return err;
 	}
Signature · 1 line
1.7.5.4
#6 lore raw
In reply to Clemens Ladisch
Quoted 10 lines · Clemens Ladisch

Review

Quoted 4 lines · David Henningsson

... two patches - one adds a "device index" for mapping against pcm devices on HDMI,

(quoted diff) +1 -0
+	kctl->id.device = device;

In the general case, there is not always a 1:1 mapping between jacks and PCM devices, so I fear that a jack belonging to device 0 cannot be distinguished from a jack that has no specific device.

Quoted 3 lines

There is a convention that id.device is a valid PCM device number if and only if id.iface is IFACE_PCM. However, it might be useful to map controls with IFACE_MIXER to PCM devices, so I'd propose a new TLV,

Er, the point was for the TLV to be optional; please ignore the previous patch. :)

--8<---------------------------------------------------------------->8--
ALSA: hda/jack: add PCM device TLV for jack controls

Allow to add a new TLV to jack kcontrols so that HDMI jacks can be associated with the correct PCM devices.

---
 include/sound/tlv.h        |    6 ++++++
 sound/pci/hda/hda_jack.c   |   12 ++++++++++--
 sound/pci/hda/hda_jack.h   |    3 ++-
 sound/pci/hda/patch_hdmi.c |    3 ++-
 4 files changed, 20 insertions(+), 4 deletions(-)
include/sound/tlv.h +6 -0
index 7067e2d..809b077 100644
@@ -37,6 +37,7 @@
 #define SNDRV_CTL_TLVT_DB_RANGE 3	/* dB range container */
 #define SNDRV_CTL_TLVT_DB_MINMAX 4	/* dB scale with min/max */
 #define SNDRV_CTL_TLVT_DB_MINMAX_MUTE 5	/* dB scale with min/max with mute */
+#define SNDRV_CTL_TLVT_PCM_DEVICE 6	/* (jack) control maps to this device */
 
 #define TLV_DB_SCALE_MASK	0xffff
 #define TLV_DB_SCALE_MUTE	0x10000
@@ -73,4 +74,9 @@
 
 #define TLV_DB_GAIN_MUTE	-9999999
 
+#define TLV_PCM_DEVICE_NUMBER_ITEM(devno) \
+	SNDRV_CTL_TLVT_PCM_DEVICE, 1 * sizeof(unsigned int), (devno)
+#define DECLARE_TLV_PCM_DEVICE_NUMBER(name, devno) \
+	unsigned int name[] = { TLV_PCM_DEVICE_NUMBER_ITEM(devno) }
+
 #endif /* __SOUND_TLV_H */
sound/pci/hda/hda_jack.c +10 -2
index 3949015..848dd03 100644
@@ -15,6 +15,7 @@
 #include <sound/core.h>
 #include <sound/control.h>
 #include <sound/jack.h>
+#include <sound/tlv.h>
 #include "hda_codec.h"
 #include "hda_local.h"
 #include "hda_jack.h"
@@ -210,7 +211,7 @@ EXPORT_SYMBOL_HDA(snd_hda_jack_report_sync);
  * will have the given name and index.
  */
 int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-			  const char *name, int idx)
+			  const char *name, int idx, int device)
 {
 	struct hda_jack_tbl *jack;
 	struct snd_kcontrol *kctl;
@@ -223,6 +224,13 @@ int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
 	kctl = snd_kctl_jack_new(name, idx, codec);
 	if (!kctl)
 		return -ENOMEM;
+	if (device >= 0) {
+		jack->tlv[0] = SNDRV_CTL_TLVT_PCM_DEVICE;
+		jack->tlv[1] = 1 * sizeof(unsigned int);
+		jack->tlv[2] = device;
+		kctl->tlv.p = jack->tlv;
+		kctl->vd[0].access |= SNDRV_CTL_ELEM_ACCESS_TLV_READ;
+	}
 	if (snd_hda_ctl_add(codec, nid, kctl) < 0)
 		return -ENOMEM;
 	jack->kctl = kctl;
@@ -249,7 +257,7 @@ static int add_jack_kctl(struct hda_codec *codec, hda_nid_t nid,
 		return 0;
 
 	snd_hda_get_pin_label(codec, nid, cfg, name, sizeof(name), &idx);
-	err = snd_hda_jack_add_kctl(codec, nid, name, idx);
+	err = snd_hda_jack_add_kctl(codec, nid, name, idx, -1);
 	if (err < 0)
 		return err;
 	return snd_hda_jack_detect_enable(codec, nid, 0);
sound/pci/hda/hda_jack.h +2 -1
index f8f97c7..a2e69c5 100644
@@ -22,6 +22,7 @@ struct hda_jack_tbl {
 	unsigned int jack_detect:1;	/* capable of jack-detection? */
 	unsigned int jack_dirty:1;	/* needs to update? */
 	struct snd_kcontrol *kctl;	/* assigned kctl for jack-detection */
+	unsigned int tlv[3];
 #ifdef CONFIG_SND_HDA_INPUT_JACK
 	int type;
 	struct snd_jack *jack;
@@ -76,7 +77,7 @@ static inline bool is_jack_detectable(struct hda_codec *codec, hda_nid_t nid)
 }
 
 int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-			  const char *name, int idx);
+			  const char *name, int idx, int device);
 int snd_hda_jack_add_kctls(struct hda_codec *codec,
 			   const struct auto_pin_cfg *cfg);
 
sound/pci/hda/patch_hdmi.c +2 -1
index bb8cfc6..9689c8d 100644
@@ -1277,7 +1277,8 @@ static int generic_hdmi_build_controls(struct hda_codec *codec)
 		if (err < 0)
 			return err;
 		err = snd_hda_jack_add_kctl(codec, per_pin->pin_nid,
-					    "HDMI", pin_idx);
+					    "HDMI", pin_idx,
+					    spec->pcm_rec[pin_idx].device);
 		if (err < 0)
 			return err;
 	}
#2 lore raw
In reply to Mark Brown

At Tue, 29 Nov 2011 20:03:51 +0000,

Quoted 6 lines · Mark Brown
Quoted 2 lines · David Henningsson

What are your thoughts on how to proceed, when do you think this is ready to be merged into the topic/hda branch?

Quoted 6 lines

I'm still rather sad that this infrastructure is all totally separate to the existing jack infrastructure, one of the goals of having common infrastructure was to insulate drivers from having to deal with changes in the APIs used for reporting and I certainly don't want to have to have all the drivers individually reporting via all the APIs (especially not with the multiswitch API coming along too).

That'll happen later. I haven't done yet just because I couldn't
spend my time for this in these weeks. I need to complete the feature in HD-audio side at first before breaking all in ASoC.

BTW, how will the multiswitch API look like?

Takashi

#7 lore raw
In reply to Clemens Ladisch

At Tue, 29 Nov 2011 21:51:45 +0100,

Quoted 10 lines · Clemens Ladisch

Review

Quoted 4 lines · David Henningsson

... two patches - one adds a "device index" for mapping against pcm devices on HDMI,

(quoted diff) +1 -0
+	kctl->id.device = device;

In the general case, there is not always a 1:1 mapping between jacks and PCM devices, so I fear that a jack belonging to device 0 cannot be distinguished from a jack that has no specific device.

Hm, a good point.

Quoted 4 lines

There is a convention that id.device is a valid PCM device number if and only if id.iface is IFACE_PCM. However, it might be useful to map controls with IFACE_MIXER to PCM devices, so I'd propose a new TLV, like in the incomplete patch below.

Yes, I had a similar idea. My original idea was to implement PCM mappings like this and dependency-mapping between control elements (e.g. master -> front -> speaker) via TLV.

Takashi

Quoted 85 lines

Regards,
Clemens

include/sound/tlv.h +6 -0
@@ -37,6 +37,7 @@
 #define SNDRV_CTL_TLVT_DB_RANGE 3	/* dB range container */
 #define SNDRV_CTL_TLVT_DB_MINMAX 4	/* dB scale with min/max */
 #define SNDRV_CTL_TLVT_DB_MINMAX_MUTE 5	/* dB scale with min/max with mute */
+#define SNDRV_CTL_TLVT_PCM_DEVICE 6	/* jack (or other control?) maps to this device */
 
 #define _TLV_ITEM(type, ...) \
 	(type), _TLV_LENGTH(__VA_ARGS__), __VA_ARGS__
@@ -81,4 +82,9 @@
 
 #define TLV_DB_GAIN_MUTE	-9999999
 
+#define TLB_PCM_DEVICE_NUMBER_ITEM(devno) \
+	SNDRV_CTL_TLVT_PCM_DEVICE, 1 * sizeof(unsigned int), (devno)
+#define DECLARE_TLV_PCM_DEVICE_NUMBER(name, devno) \
+	unsigned int name[] = { TLV_PCM_DEVICE_NUMBER_ITEM(devno) }
+
 #endif /* __SOUND_TLV_H */
sound/pci/hda/hda_jack.c +8 -2
@@ -15,6 +15,7 @@
 #include <sound/core.h>
 #include <sound/control.h>
 #include <sound/jack.h>
+#include <sound/tlv.h>
 #include "hda_codec.h"
 #include "hda_local.h"
 #include "hda_jack.h"
@@ -210,7 +211,7 @@ EXPORT_SYMBOL_HDA(snd_hda_jack_report_sync);
  * will have the given name and index.
  */
 int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-			  const char *name, int idx)
+			  const char *name, int idx, int device)
 {
 	struct hda_jack_tbl *jack;
 	struct snd_kcontrol *kctl;
@@ -223,6 +224,11 @@ int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
 	kctl = snd_kctl_jack_new(name, idx, codec);
 	if (!kctl)
 		return -ENOMEM;
+	jack->tlv[0] = SNDRV_CTL_TLVT_PCM_DEVICE;
+	jack->tlv[1] = 1 * sizeof(unsigned int);
+	jack->tlv[2] = device;
+	kctl->tlv.p = jack->tlv;
+	kctl->access |= SNDRV_CTL_ELEM_ACCESS_TLV_READ;
 	if (snd_hda_ctl_add(codec, nid, kctl) < 0)
 		return -ENOMEM;
 	jack->kctl = kctl;
@@ -249,7 +255,7 @@ static int add_jack_kctl(struct hda_codec *codec, hda_nid_t nid,
 		return 0;
 
 	snd_hda_get_pin_label(codec, nid, cfg, name, sizeof(name), &idx);
-	err = snd_hda_jack_add_kctl(codec, nid, name, idx);
+	err = snd_hda_jack_add_kctl(codec, nid, name, idx, 0);
 	if (err < 0)
 		return err;
 	return snd_hda_jack_detect_enable(codec, nid, 0);
sound/pci/hda/hda_jack.h +2 -1
@@ -22,6 +22,7 @@ struct hda_jack_tbl {
 	unsigned int jack_detect:1;	/* capable of jack-detection? */
 	unsigned int jack_dirty:1;	/* needs to update? */
 	struct snd_kcontrol *kctl;	/* assigned kctl for jack-detection */
+	unsigned int tlv[3];
 #ifdef CONFIG_SND_HDA_INPUT_JACK
 	int type;
 	struct snd_jack *jack;
@@ -76,7 +77,7 @@ static inline bool is_jack_detectable(struct hda_codec *codec, hda_nid_t nid)
 }
 
 int snd_hda_jack_add_kctl(struct hda_codec *codec, hda_nid_t nid,
-			  const char *name, int idx);
+			  const char *name, int idx, int device);
 int snd_hda_jack_add_kctls(struct hda_codec *codec,
 			   const struct auto_pin_cfg *cfg);
 
#9 lore raw
In reply to Raymond Yau
Quoted 12 lines · Raymond Yau

2011/11/30 David Henningsson<[email protected]>:

Quoted 11 lines

Hi Takashi,

I've done some more tests of your branch now, and the new names seems to work for my machine here.

I'm attaching two patches - one adds a "device index" for mapping against pcm devices on HDMI, the other one is a backport for the driver-build tree.

What are your thoughts on how to proceed, when do you think this is ready to be merged into the topic/hda branch?

Quoted 6 lines
Quoted 2 lines

// David

http://git.alsa-project.org/?p=alsa-kernel.git;a=blob_plain;f=Documentation/sound/alsa/HD-Audio-Controls.txt

Does it handle "Independent HP" ?

Not currently, my patch is concerned with HDMI only.
You're correct in that if "Independent HP" is on, it wouldn't hurt for the Headphone jack to also have such a device index.

Signature · 2 lines
David Henningsson, Canonical Ltd.
http://launchpad.net/~diwic
#3 lore raw
In reply to Takashi Iwai
Quoted 1 line · Takashi Iwai

BTW, how will the multiswitch API look like?

It's taken the Android switch API as a starting point so the main reporting mechanism is uevents.