Re: [PATCH] ASoC: meson: g12a-toacodec: add support for SM1 TOACODEC

[Neil Armstrong <narmstrong@baylibre.com>]

On 29/04/2021 19:43, Jerome Brunet wrote:
> 
> On Thu 29 Apr 2021 at 19:01, Neil Armstrong <narmstrong@baylibre.com> wrote:
> 
> 
>>  
>> +static int sm1_toacodec_mux_put_enum(struct snd_kcontrol *kcontrol,
>> +				     struct snd_ctl_elem_value *ucontrol)
>> +{
>> +	struct snd_soc_component *component =
>> +		snd_soc_dapm_kcontrol_component(kcontrol);
>> +	struct snd_soc_dapm_context *dapm =
>> +		snd_soc_dapm_kcontrol_dapm(kcontrol);
>> +	struct soc_enum *e = (struct soc_enum *)kcontrol->private_value;
>> +	unsigned int mux, changed;
>> +
>> +	mux = snd_soc_enum_item_to_val(e, ucontrol->value.enumerated.item[0]);
>> +	changed = snd_soc_component_test_bits(component, e->reg,
>> +					      CTRL0_DAT_SEL_SM1,
>> +					      FIELD_PREP(CTRL0_DAT_SEL_SM1, mux));
>> +
>> +	if (!changed)
>> +		return 0;
>> +
>> +	/* Force disconnect of the mux while updating */
>> +	snd_soc_dapm_mux_update_power(dapm, kcontrol, 0, NULL, NULL);
>> +
>> +	snd_soc_component_update_bits(component, e->reg,
>> +				      CTRL0_DAT_SEL_SM1 |
>> +				      CTRL0_LRCLK_SEL_SM1 |
>> +				      CTRL0_BCLK_SEL_SM1,
>> +				      FIELD_PREP(CTRL0_DAT_SEL_SM1, mux) |
>> +				      FIELD_PREP(CTRL0_LRCLK_SEL_SM1, mux) |
>> +				      FIELD_PREP(CTRL0_BCLK_SEL_SM1, mux));
>> +
>> +	/*
>> +	 * FIXME:
>> +	 * On this soc, the glue gets the MCLK directly from the clock
>> +	 * controller instead of going the through the TDM interface.
>> +	 *
>> +	 * Here we assume interface A uses clock A, etc ... While it is
>> +	 * true for now, it could be different. Instead the glue should
>> +	 * find out the clock used by the interface and select the same
>> +	 * source. For that, we will need regmap backed clock mux which
>> +	 * is a work in progress
>> +	 */
>> +	snd_soc_component_update_bits(component, e->reg,
>> +				      CTRL0_MCLK_SEL,
>> +				      FIELD_PREP(CTRL0_MCLK_SEL, mux));
>> +
>> +	snd_soc_dapm_mux_update_power(dapm, kcontrol, mux, e, NULL);
>> +
>> +	return 0;
>> +}
> 
> Instead of duplicating this function, I'd prefer if you could use regmap fields
> 
> 

Sure, will do