| Message ID | 20240519-midas-wm1811-gpio-jack-v3-3-0c1736144c0e@gmail.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | ASoC: samsung: midas-audio: Add GPIO-based headset jack detection | expand |
On Sun, May 19, 2024 at 10:17:49AM +0200, Artur Weber wrote: > +static int midas_headset_mic_bias(struct snd_soc_dapm_widget *w, > + struct snd_kcontrol *kcontrol, int event) > +{ > + struct snd_soc_card *card = w->dapm->card; > + struct midas_priv *priv = snd_soc_card_get_drvdata(card); > + > + if (!priv->reg_headset_mic_bias) > + return 0; > + > + switch (event) { > + case SND_SOC_DAPM_PRE_PMU: > + return regulator_enable(priv->reg_headset_mic_bias); > + case SND_SOC_DAPM_POST_PMD: > + return regulator_disable(priv->reg_headset_mic_bias); > + } We have SND_SOC_DAPM_REGULATOR_SUPPLY?
On 20.05.2024 16:35, Mark Brown wrote: > On Sun, May 19, 2024 at 10:17:49AM +0200, Artur Weber wrote: > >> +static int midas_headset_mic_bias(struct snd_soc_dapm_widget *w, >> + struct snd_kcontrol *kcontrol, int event) >> +{ >> + struct snd_soc_card *card = w->dapm->card; >> + struct midas_priv *priv = snd_soc_card_get_drvdata(card); >> + >> + if (!priv->reg_headset_mic_bias) >> + return 0; >> + >> + switch (event) { >> + case SND_SOC_DAPM_PRE_PMU: >> + return regulator_enable(priv->reg_headset_mic_bias); >> + case SND_SOC_DAPM_POST_PMD: >> + return regulator_disable(priv->reg_headset_mic_bias); >> + } > > We have SND_SOC_DAPM_REGULATOR_SUPPLY? This is pretty much copied from the implementation of the mic bias and sub mic bias regulator handling in the same driver; they all use SND_SOC_DAPM_MIC combined with an enable/disable function like this one. What would be the correct thing to do here - add a secondary DAPM widget of type REGULATOR_SUPPLY and connect it to the MIC widget via audio- routing, or replace the entire MIC widget with a REGULATOR_SUPPLY (or something else entirely)? And should this be done with the Main Mic and Sub Mic as well? Best regards Artur
On Wed, May 22, 2024 at 06:20:14PM +0200, Artur Weber wrote: > What would be the correct thing to do here - add a secondary DAPM widget > of type REGULATOR_SUPPLY and connect it to the MIC widget via audio- > routing, or replace the entire MIC widget with a REGULATOR_SUPPLY (or > something else entirely)? The microphone is getting a supply so the first option seems better. > And should this be done with the Main Mic and Sub Mic as well? Probably yes.
diff --git a/sound/soc/samsung/midas_wm1811.c b/sound/soc/samsung/midas_wm1811.c index f31244156ff6..ab0a4804b45e 100644 --- a/sound/soc/samsung/midas_wm1811.c +++ b/sound/soc/samsung/midas_wm1811.c @@ -29,6 +29,7 @@ struct midas_priv { struct regulator *reg_mic_bias; struct regulator *reg_submic_bias; + struct regulator *reg_headset_mic_bias; struct gpio_desc *gpio_fm_sel; struct gpio_desc *gpio_lineout_sel; unsigned int fll1_rate; @@ -201,6 +202,25 @@ static int midas_submic_bias(struct snd_soc_dapm_widget *w, return 0; } +static int midas_headset_mic_bias(struct snd_soc_dapm_widget *w, + struct snd_kcontrol *kcontrol, int event) +{ + struct snd_soc_card *card = w->dapm->card; + struct midas_priv *priv = snd_soc_card_get_drvdata(card); + + if (!priv->reg_headset_mic_bias) + return 0; + + switch (event) { + case SND_SOC_DAPM_PRE_PMU: + return regulator_enable(priv->reg_headset_mic_bias); + case SND_SOC_DAPM_POST_PMD: + return regulator_disable(priv->reg_headset_mic_bias); + } + + return 0; +} + static int midas_fm_set(struct snd_soc_dapm_widget *w, struct snd_kcontrol *kcontrol, int event) { @@ -271,7 +291,7 @@ static const struct snd_soc_dapm_widget midas_dapm_widgets[] = { SND_SOC_DAPM_LINE("FM In", midas_fm_set), SND_SOC_DAPM_HP("Headphone", NULL), - SND_SOC_DAPM_MIC("Headset Mic", NULL), + SND_SOC_DAPM_MIC("Headset Mic", midas_headset_mic_bias), SND_SOC_DAPM_MIC("Main Mic", midas_mic_bias), SND_SOC_DAPM_MIC("Sub Mic", midas_submic_bias), }; @@ -455,6 +475,17 @@ static int midas_probe(struct platform_device *pdev) return PTR_ERR(priv->reg_submic_bias); } + priv->reg_headset_mic_bias = devm_regulator_get_optional(dev, + "headset-mic-bias"); + if (IS_ERR(priv->reg_headset_mic_bias)) { + ret = PTR_ERR(priv->reg_headset_mic_bias); + if (ret == -ENODEV) + priv->reg_headset_mic_bias = NULL; + else + return dev_err_probe(dev, ret, + "Failed to get headset mic bias regulator\n"); + } + priv->gpio_fm_sel = devm_gpiod_get_optional(dev, "fm-sel", GPIOD_OUT_HIGH); if (IS_ERR(priv->gpio_fm_sel)) { dev_err(dev, "Failed to get FM selection GPIO\n");
Some devices use a headset mic bias supply (sometimes referred to as "ear mic bias") to enable/disable the headset mic. Add support for getting the supply from DT and setting it up accordingly to the value of the Headset Mic switch. Signed-off-by: Artur Weber <aweber.kernel@gmail.com> --- Changes in v2: - Added this commit --- sound/soc/samsung/midas_wm1811.c | 33 ++++++++++++++++++++++++++++++++- 1 file changed, 32 insertions(+), 1 deletion(-)