ALSA: hda: Disable odd numbered channels on HDMI controllers.
PATCH 1/1 ALSA: hda: Disable odd numbered channels on HDMI controllers.
Added code in hda_intel.c to disable odd numbered channels not supported by HDMI controllers.
--- pci/hda/hda_intel.c | 13 +++++++++++++ 1 files changed, 13 insertions(+), 0 deletions(-)
pci/hda/hda_intel.c +13 -0
@@ -1539,6 +1539,16 @@ struct azx_pcm {struct hda_pcm_stream *hinfo[2 ]; }; +static unsigned int channels_2_4_6_8[] = { +2 ,4 ,6 ,8 +}; + +static struct snd_pcm_hw_constraint_list hw_constraints_2_4_6_8_channels = { + .count =ARRAY_SIZE (channels_2_4_6_8), + .list = channels_2_4_6_8, + .mask =0 , +}; +static int azx_pcm_open (struct snd_pcm_substream *substream) {struct azx_pcm *apcm =snd_pcm_substream_chip (substream);
@@ -1566,6 +1576,9 @@ static int azx_pcm_open(struct snd_pcm_substream *substream) 128);snd_pcm_hw_constraint_step (runtime,0 ,SNDRV_PCM_HW_PARAM_PERIOD_BYTES ,128 ); +snd_pcm_hw_constraint_list (runtime,0 , +SNDRV_PCM_HW_PARAM_CHANNELS , + &hw_constraints_2_4_6_8_channels);snd_hda_power_up (apcm->codec); err = hinfo->ops.open (hinfo, apcm->codec, substream);if (err <0 ) {
Signature · 9 lines
1.7.0.4 ----------------------------------------------------------------------------------- This email message is for the sole use of the intended recipient(s) and may contain confidential information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message. -----------------------------------------------------------------------------------
Quoted 16 lines · Nitin Daga
Added code in
hda_intel.cto disable odd numbered channels not supported by HDMI controllers.Signed-off-by Nitin Daga <[email protected]>Acked-by Stephen Warren <[email protected]>--- pci/hda/hda_intel.c | 13 +++++++++++++ 1 files changed, 13 insertions(+), 0 deletions(-)pci/hda/hda_intel.c +0
-0@@ -1539,6 +1539,16 @@ struct azx_pcm {struct hda_pcm_stream *hinfo[2 ]; };
pci/hda/hda_intel.c struct azx_pcm { · line 1550
Quoted 10 lines
pci/hda/hda_intel.c +9
-0+static unsigned int channels_2_4_6_8[] = { +2 ,4 ,6 ,8 +}; + +static struct snd_pcm_hw_constraint_list hw_constraints_2_4_6_8_channels = { + .count =ARRAY_SIZE (channels_2_4_6_8), + .list = channels_2_4_6_8, + .mask =0 , +};
Hmm, can't these be const? OTOH none of the other drivers use const for snd_pcm_hw_constraint_list either, so I guess it is ok.
pci/hda/hda_intel.c static int azx_pcm_open(struct snd_pcm_substream *substream) · line 1581
Quoted 10 lines
pci/hda/hda_intel.c +3
-0static int azx_pcm_open (struct snd_pcm_substream *substream) {struct azx_pcm *apcm =snd_pcm_substream_chip (substream);@@ -1566,6 +1576,9 @@ static int azx_pcm_open(struct snd_pcm_substream *substream) 128);snd_pcm_hw_constraint_step (runtime,0 ,SNDRV_PCM_HW_PARAM_PERIOD_BYTES ,128 ); +snd_pcm_hw_constraint_list (runtime,0 , +SNDRV_PCM_HW_PARAM_CHANNELS , + &hw_constraints_2_4_6_8_channels);
This would seem simpler and more generic (there are codecs that have over 8 channels):snd_pcm_hw_constraint_step(runtime, 0, SNDRV_PCM_HW_PARAM_CHANNELS, 2);
However, either of those would AFAICS break Si3054 Modem codec which is mono-only.
Maybe just add the constraint to patch_hdmi.c?
Or use the patch as is but add 1, 10, 12, 14, 16 (or so) to the allowed channel counts list.
Not really sure what is the preferred approach, though.
Quoted 3 lines
pci/hda/hda_intel.c +0
-0snd_hda_power_up (apcm->codec); err = hinfo->ops.open (hinfo, apcm->codec, substream);if (err <0 ) {
Signature · 1 line
Anssi Hannula
At Thu, 13 Jan 2011 22:42:39 +0200,
Quoted 47 lines · Anssi Hannula
Quoted 16 lines · Nitin Daga
Added code in
hda_intel.cto disable odd numbered channels not supported by HDMI controllers.Signed-off-by Nitin Daga <[email protected]>Acked-by Stephen Warren <[email protected]>--- pci/hda/hda_intel.c | 13 +++++++++++++ 1 files changed, 13 insertions(+), 0 deletions(-)pci/hda/hda_intel.c +0
-0@@ -1539,6 +1539,16 @@ struct azx_pcm {struct hda_pcm_stream *hinfo[2 ]; };pci/hda/hda_intel.c struct azx_pcm { · line 1550
Quoted 10 lines
pci/hda/hda_intel.c +9
-0+static unsigned int channels_2_4_6_8[] = { +2 ,4 ,6 ,8 +}; + +static struct snd_pcm_hw_constraint_list hw_constraints_2_4_6_8_channels = { + .count =ARRAY_SIZE (channels_2_4_6_8), + .list = channels_2_4_6_8, + .mask =0 , +};Hmm, can't these be const? OTOH none of the other drivers use const for
snd_pcm_hw_constraint_listeither, so I guess it is ok.pci/hda/hda_intel.c static int azx_pcm_open(struct snd_pcm_substream *substream) · line 1581
Quoted 10 lines
pci/hda/hda_intel.c +3
-0static int azx_pcm_open (struct snd_pcm_substream *substream) {struct azx_pcm *apcm =snd_pcm_substream_chip (substream);@@ -1566,6 +1576,9 @@ static int azx_pcm_open(struct snd_pcm_substream *substream) 128);snd_pcm_hw_constraint_step (runtime,0 ,SNDRV_PCM_HW_PARAM_PERIOD_BYTES ,128 ); +snd_pcm_hw_constraint_list (runtime,0 , +SNDRV_PCM_HW_PARAM_CHANNELS , + &hw_constraints_2_4_6_8_channels);This would seem simpler and more generic (there are codecs that have over 8 channels):
snd_pcm_hw_constraint_step(runtime, 0, SNDRV_PCM_HW_PARAM_CHANNELS, 2);
Quoted 4 lines
However, either of those would AFAICS break Si3054 Modem codec which is mono-only.
Maybe just add the constraint to
patch_hdmi.c?
Yes, looks so. I applied the following patch instead.
Alternatively, we can check channels_min and channels_max that are already set in hda_intel.c::azx_pcm_open(), and apply the hw_constraint only when channels_min%2==0 or such.
But, as we aren't 100% sure whether the odd number of channels work with other (analog) codecs, it's safer to limit only for HDMI for the time being.
thanks,
Takashi
---
Quoted 1 line
From
ad09fc9d2156f3d37537b34418a6b79309013d33Mon Sep 17 00:00:00 2001
From: Takashi Iwai <[email protected]>
Date: Fri, 14 Jan 2011 09:42:27 +0100
Subject: [PATCH] ALSA: hda - Suppress the odd number of channels for HDMI
It looks like that HDMI codecs don't support the odd number of channels although HD-audio spec doesn't have the restriction. Add the hw_constraint to limit to only the even number of channels.
--- sound/pci/hda/patch_hdmi.c | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-)
sound/pci/hda/patch_hdmi.c +3 -0
@@ -1238,6 +1238,9 @@ static int simple_playback_pcm_open(struct hda_pcm_stream *hinfo,snd_pcm_hw_constraint_list (substream->runtime,0 ,SNDRV_PCM_HW_PARAM_CHANNELS , hw_constraints_channels); + }else { +snd_pcm_hw_constraint_step (substream->runtime,0 , +SNDRV_PCM_HW_PARAM_CHANNELS ,2 ); }return snd_hda_multi_out_dig_open (codec, &spec->multiout);
Signature · 1 line
1.7.3.4
Quoted 12 lines · Takashi Iwai
At Thu, 13 Jan 2011 22:42:39 +0200,
Quoted 10 lines · Anssi Hannula
Quoted 8 lines · Nitin Daga
Added code in
hda_intel.cto disable odd numbered channels not supported by HDMI controllers.Signed-off-by Nitin Daga <[email protected]>Acked-by Stephen Warren <[email protected]>--- pci/hda/hda_intel.c | 13 +++++++++++++ 1 files changed, 13 insertions(+), 0 deletions(-)
[...]
Quoted 10 lines
Quoted 8 lines
This would seem simpler and more generic (there are codecs that have over 8 channels):
snd_pcm_hw_constraint_step(runtime, 0, SNDRV_PCM_HW_PARAM_CHANNELS, 2);However, either of those would AFAICS break Si3054 Modem codec which is mono-only.
Maybe just add the constraint to
patch_hdmi.c?Yes, looks so. I applied the following patch instead.
It is ineffective, the generic hdmi devices use hdmi_pcm_open() instead of simple_playback_pcm_open().
Quoted 7 lines
Alternatively, we can check
channels_minandchannels_maxthat are already set inhda_intel.c::azx_pcm_open(), and apply thehw_constraintonly when channels_min%2==0 or such.But, as we aren't 100% sure whether the odd number of channels work with other (analog) codecs, it's safer to limit only for HDMI for the time being.
Well, snd_hda_multi_out_analog_open() already sets such a constraint, but I didn't check if all the analog multichannel codecs use that.
Quoted 33 lines
thanks,
Takashi
---
Fromad09fc9d2156f3d37537b34418a6b79309013d33Mon Sep 17 00:00:00 2001
From: Takashi Iwai <[email protected]>
Date: Fri, 14 Jan 2011 09:42:27 +0100
Subject: [PATCH] ALSA: hda - Suppress the odd number of channels for HDMIIt looks like that HDMI codecs don't support the odd number of channels although HD-audio spec doesn't have the restriction. Add the
hw_constraintto limit to only the even number of channels.Signed-off-by Takashi Iwai <[email protected]>--- sound/pci/hda/patch_hdmi.c | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-)sound/pci/hda/patch_hdmi.c +3
-0@@ -1238,6 +1238,9 @@ static int simple_playback_pcm_open(struct hda_pcm_stream *hinfo,snd_pcm_hw_constraint_list (substream->runtime,0 ,SNDRV_PCM_HW_PARAM_CHANNELS , hw_constraints_channels); + }else { +snd_pcm_hw_constraint_step (substream->runtime,0 , +SNDRV_PCM_HW_PARAM_CHANNELS ,2 ); }return snd_hda_multi_out_dig_open (codec, &spec->multiout);
Signature · 1 line
Anssi Hannula
At Fri, 14 Jan 2011 11:14:16 +0200,
Quoted 22 lines · Anssi Hannula
Quoted 12 lines · Takashi Iwai
At Thu, 13 Jan 2011 22:42:39 +0200,
Quoted 10 lines · Anssi Hannula
Quoted 8 lines · Nitin Daga
Added code in
hda_intel.cto disable odd numbered channels not supported by HDMI controllers.Signed-off-by Nitin Daga <[email protected]>Acked-by Stephen Warren <[email protected]>--- pci/hda/hda_intel.c | 13 +++++++++++++ 1 files changed, 13 insertions(+), 0 deletions(-)[...]
Quoted 7 lines
Quoted 7 lines
This would seem simpler and more generic (there are codecs that have over 8 channels):
snd_pcm_hw_constraint_step(runtime, 0, SNDRV_PCM_HW_PARAM_CHANNELS, 2);However, either of those would AFAICS break Si3054 Modem codec which is mono-only.
Quoted 6 lines
Quoted 3 lines
Quoted 1 line
Maybe just add the constraint to
patch_hdmi.c?Yes, looks so. I applied the following patch instead.
It is ineffective, the generic hdmi devices use
hdmi_pcm_open()instead ofsimple_playback_pcm_open().
Ah, indeed. Fixed now.
Meanwhile, I found another bug in patch_hdmi.c. The updated PCM parameters aren't reflected properly at the first time because runtime->hw was already set. Fixed by patch below.
Quoted 10 lines
Quoted 7 lines
Alternatively, we can check
channels_minandchannels_maxthat are already set inhda_intel.c::azx_pcm_open(), and apply thehw_constraintonly when channels_min%2==0 or such.But, as we aren't 100% sure whether the odd number of channels work with other (analog) codecs, it's safer to limit only for HDMI for the time being.
Well,
snd_hda_multi_out_analog_open()already sets such a constraint, but I didn't check if all the analog multichannel codecs use that.
Right. OTOH, the spec doesn't restrict it, so limiting it in the controller driver side (hda_intel.c) doesn't sound right. It's rather likely a codec-specific issue. So, restriction should be applied in the codec side, IMO.
thanks,
Takashi
---
Quoted 1 line
From
639cef0eb6df05d5516520aa89b0c9fe62ee2d3bMon Sep 17 00:00:00 2001
From: Takashi Iwai <[email protected]>
Date: Fri, 14 Jan 2011 10:30:46 +0100
Subject: [PATCH] ALSA: hda - Store PCM parameters properly in HDMI open callback
In hdmi_pcm_open(), the evaluated PCM hw parameters are stored in hinfo, but these aren't properly set back to the current runtime record since these have been set beforehand in azx_pcm_open(). This patch fixes the behavior.
--- sound/pci/hda/patch_hdmi.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-)
sound/pci/hda/patch_hdmi.c +6 -0
@@ -817,6 +817,7 @@ static int hdmi_pcm_open(struct hda_pcm_stream *hinfo,struct hdmi_spec *spec = codec->spec;struct hdmi_eld *eld;struct hda_pcm_stream *codec_pars; +struct snd_pcm_runtime *runtime = substream->runtime;unsigned int idx;for (idx =0 ; idx < spec->num_cvts; idx++)
@@ -844,6 +845,11 @@ static int hdmi_pcm_open(struct hda_pcm_stream *hinfo, hinfo->formats = codec_pars->formats; hinfo->maxbps = codec_pars->maxbps; } +/* store the updated parameters */ + runtime->hw.channels_min = hinfo->channels_min; + runtime->hw.channels_max = hinfo->channels_max; + runtime->hw.formats = hinfo->formats; + runtime->hw.rates = hinfo->rates;return 0 ; }
Signature · 1 line
1.7.3.4