Re: [alsa-devel] [PATCH 4/4] sound: codecs: pcm1792a: introduce channel side

From: RaphaÃl Poggi
Date: Tue Nov 24 2015 - 11:05:43 EST


Hi,

2015-11-24 9:27 GMT+01:00 Michael Trimarchi <michael@xxxxxxxxxxxxxxxxxxxx>:
> Hi
>
> On Tue, Nov 24, 2015 at 9:21 AM, RaphaÃl Poggi <poggi.raph@xxxxxxxxx> wrote:
>> Hi,
>>
>> 2015-11-21 10:37 GMT+01:00 Michael Trimarchi <michael@xxxxxxxxxxxxxxxxxxxx>:
>>> Hi
>>>
>>> On Fri, Nov 20, 2015 at 11:09 AM, Raphael Poggi <poggi.raph@xxxxxxxxx> wrote:
>>>> From: Raphael Poggi <poggi.raph@xxxxxxxxx>
>>>>
>>>> Add possibility to choose the channel side using the device tree,
>>>> and also modify it using alsa ctrl.
>>>>
>>>> Signed-off-by: Raphael Poggi <poggi.raph@xxxxxxxxx>
>>>> ---
>>>> sound/soc/codecs/pcm1792a.c | 46 +++++++++++++++++++++++++++++++++++++++++++++
>>>> 1 file changed, 46 insertions(+)
>>>>
>>>> diff --git a/sound/soc/codecs/pcm1792a.c b/sound/soc/codecs/pcm1792a.c
>>>> index febaa48..005b679 100644
>>>> --- a/sound/soc/codecs/pcm1792a.c
>>>> +++ b/sound/soc/codecs/pcm1792a.c
>>>> @@ -65,6 +65,12 @@ static const struct reg_default pcm1792a_reg_defaults[] = {
>>>> { PCM1792A_DEVICE_ID, 0x00 },
>>>> };
>>>>
>>>> +enum __pcm1792_side {
>>>> + STEREO,
>>>> + MONORAL_LEFT,
>>>> + MONORAL_RIGHT
>>>> +};
>>>> +
>>>> static bool pcm1792a_accessible_reg(struct device *dev, unsigned int reg)
>>>> {
>>>> return reg >= 0x10 && reg <= 0x17;
>>>> @@ -83,6 +89,7 @@ struct pcm1792a_private {
>>>> struct regmap *regmap;
>>>> unsigned int format;
>>>> unsigned int rate;
>>>> + unsigned int side;
>>>> };
>>>>
>>>> static int pcm1792a_set_dai_fmt(struct snd_soc_dai *codec_dai,
>>>> @@ -153,6 +160,13 @@ static int pcm1792a_hw_params(struct snd_pcm_substream *substream,
>>>> ret = regmap_update_bits(priv->regmap, PCM1792A_FMT_CONTROL,
>>>> PCM1792A_FMT_MASK | PCM1792A_ATLD_ENABLE, val);
>>>>
>>>> + if (priv->side == MONORAL_LEFT)
>>>> + val = PCM1792A_CHANNEL_MONO_LEFT;
>>>> + else if (priv->side == MONORAL_RIGHT)
>>>> + val = PCM1792A_CHANNEL_MONO_RIGHT;
>>>> +
>>>> + regmap_update_bits(priv->regmap, PCM1792A_CHANNEL, PCM1792A_CHANNEL_MONO_MASK, val);
>>>> +
>>>> return ret;
>>>> }
>>>>
>>>> @@ -222,10 +236,31 @@ static struct snd_soc_codec_driver soc_codec_dev_pcm1792a = {
>>>> .num_dapm_routes = ARRAY_SIZE(pcm1792a_dapm_routes),
>>>> };
>>>>
>>>> +static int pcm1792a_of_init(struct spi_device *spi)
>>>> +{
>>>> + int ret = 0;
>>>> + struct pcm1792a_private *pcm1792a = dev_get_drvdata(&spi->dev);
>>>> + const char *name;
>>>> + struct device_node *np = spi->dev.of_node;
>>>> +
>>>> + name = of_get_property(np, "ti,side", NULL);
>>>> + if (name) {
>>>> + if (!strcmp(name, "left"))
>>>> + pcm1792a->side = MONORAL_LEFT;
>>>> + else
>>>> + pcm1792a->side = MONORAL_RIGHT;
>>>> + }
>>>> + else
>>>> + pcm1792a->side = STEREO;
>>>> +
>>>> + return ret;
>>>> +}
>>>> +
>>>> static int pcm1792a_spi_probe(struct spi_device *spi)
>>>> {
>>>> struct pcm1792a_private *pcm1792a;
>>>> int ret;
>>>> + struct device_node *np = spi->dev.of_node;
>>>>
>>>> pcm1792a = devm_kzalloc(&spi->dev, sizeof(struct pcm1792a_private),
>>>> GFP_KERNEL);
>>>> @@ -241,6 +276,17 @@ static int pcm1792a_spi_probe(struct spi_device *spi)
>>>> return ret;
>>>> }
>>>>
>>>> + if (np) {
>>>> + ret = pcm1792a_of_init(spi);
>>>> + if (ret)
>>>> + return ret;
>>>> + }
>>>> +
>>>> + if (pcm1792a->side == MONORAL_LEFT)
>>>> + regmap_update_bits(pcm1792a->regmap, PCM1792A_CHANNEL, PCM1792A_CHANNEL_MONO_MASK, PCM1792A_CHANNEL_MONO_LEFT);
>>>> + else if (pcm1792a->side == MONORAL_RIGHT)
>>>> + regmap_update_bits(pcm1792a->regmap, PCM1792A_CHANNEL, PCM1792A_CHANNEL_MONO_MASK, PCM1792A_CHANNEL_MONO_RIGHT);
>>>> +
>>>
>>> Why this is a part of dts and not a part of mixer control?
>>>
>>> Michael
>>>
>>
>> The "channel side" bindings should be used when the pcm7192a is used
>> in monaural mode (which requires two DACs, both DACs operate in a
>> balanced mode for one channel of audio input data).
>> This is a hardware level design, it will never change at runtime.
>> I feel that a device tree binding is more appropriate than a control.
>>
>
> Well I know exactly why is used for, but I don't really know if it's
> somenthing that should be
> in dts. Do I want to do channel inversion?
>
> Michael
>

I am not sure to understand. You mean "What if I want to do channel
inversion", right ?
If so, we have two differents "problems", one at hardware level design
because monaural mode requires two DACs, and an another one at
software level
"channel inversion".
I have always thought that hardware description should be in dts,
that's why I implement monaural mode using dts.

Honestly, I don't know what is the best solution between "this is a
hardware design, it should be in dts and shouldn't change" vs "this
should be an alsa controler because we want extra features independent
from hardware design"

RaphaÃl

>>>> return snd_soc_register_codec(&spi->dev,
>>>> &soc_codec_dev_pcm1792a, &pcm1792a_dai, 1);
>>>> }
>>>> --
>>>> 2.1.0
>>>>
>>>> _______________________________________________
>>>> Alsa-devel mailing list
>>>> Alsa-devel@xxxxxxxxxxxxxxxx
>>>> http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
>>>
>>>
>>>
>>> --
>>> | Michael Nazzareno Trimarchi Amarula Solutions BV |
>>> | COO - Founder Cruquiuskade 47 |
>>> | +31(0)851119172 Amsterdam 1018 AM NL |
>>> | [`as] http://www.amarulasolutions.com |
>>> --
>>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>>> the body of a message to majordomo@xxxxxxxxxxxxxxx
>>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>>> Please read the FAQ at http://www.tux.org/lkml/
>
>
>
> --
> | Michael Nazzareno Trimarchi Amarula Solutions BV |
> | COO - Founder Cruquiuskade 47 |
> | +31(0)851119172 Amsterdam 1018 AM NL |
> | [`as] http://www.amarulasolutions.com |
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/