Re: [PATCH 3/4] ASoC: amd: acp: Add machine driver that enables sound for systems with a ES8336 codec

[Pierre-Louis Bossart]

>>> +static int acp3x_es83xx_speaker_power_event(struct snd_soc_dapm_widget *w,
>>> +					    struct snd_kcontrol *kcontrol, int event)
>>> +{
>>> +	struct acp3x_es83xx_private *priv = get_mach_priv(w->dapm->card);
>>> +
>>> +	dev_dbg(priv->codec_dev, "speaker power event: %d\n", event);
>>> +	if (SND_SOC_DAPM_EVENT_ON(event))
>>> +		acp3x_es83xx_set_gpios_values(priv, 1, 0);
>>> +	else
>>> +		acp3x_es83xx_set_gpios_values(priv, 0, 1);
>>
>> Why are these two GPIOs tied together like this?
>>
> 
> These GPIOs represent the speaker and the headphone switches. When
> activating the speaker GPIO you have to deactivate the headphone GPIO
> and vice versa. The logic is taken from the discussion on the sofproject
> pull request:
> https://github.com/thesofproject/linux/pull/4112/commits/810d03e0aecdf0caf580a5179ee6873fb33485ab
> and
> https://github.com/thesofproject/linux/pull/4066

These threads didn't exactly lead to a firm conclusion on how the GPIOs
should be used, IIRC there are cases where the levels are inverted and
all kinds of issues still not clear at all even after reading the tables
from ACPI _DSM methods.

I personally gave up, and I would recommend you take these threads as
inputs rather than firm directions.