2020-10-30 08:00:11

by ajye huang

[permalink] [raw]
Subject: [PATCH v2 2/2] ASoC: qcom: sc7180: Modify machine driver for 2mic

In addition, having mixer control to switch between DMICs by
using "dmic-gpios" property.

Refer to this one as an example,
commit b7a742cff3f6 ("ASoC: AMD: Use mixer control to switch between DMICs")

Signed-off-by: Ajye Huang <[email protected]>
---
sound/soc/qcom/sc7180.c | 59 +++++++++++++++++++++++++++++++++++++++++
1 file changed, 59 insertions(+)

diff --git a/sound/soc/qcom/sc7180.c b/sound/soc/qcom/sc7180.c
index b391f64c3a80..e9366bb5f3c2 100644
--- a/sound/soc/qcom/sc7180.c
+++ b/sound/soc/qcom/sc7180.c
@@ -5,6 +5,8 @@
// sc7180.c -- ALSA SoC Machine driver for SC7180

#include <dt-bindings/sound/sc7180-lpass.h>
+#include <linux/gpio.h>
+#include <linux/gpio/consumer.h>
#include <linux/module.h>
#include <linux/of_device.h>
#include <linux/platform_device.h>
@@ -23,6 +25,9 @@

#define DRIVER_NAME "SC7180"

+static struct gpio_desc *dmic_sel;
+static int dmic_switch;
+
struct sc7180_snd_data {
struct snd_soc_card card;
u32 pri_mi2s_clk_count;
@@ -169,6 +174,23 @@ static int sc7180_snd_startup(struct snd_pcm_substream *substream)
return 0;
}

+static int dmic_get(struct snd_kcontrol *kcontrol,
+ struct snd_ctl_elem_value *ucontrol)
+{
+ ucontrol->value.integer.value[0] = dmic_switch;
+ return 0;
+}
+
+static int dmic_set(struct snd_kcontrol *kcontrol,
+ struct snd_ctl_elem_value *ucontrol)
+{
+ if (dmic_sel) {
+ dmic_switch = ucontrol->value.integer.value[0];
+ gpiod_set_value(dmic_sel, dmic_switch);
+ }
+ return 0;
+}
+
static void sc7180_snd_shutdown(struct snd_pcm_substream *substream)
{
struct snd_soc_pcm_runtime *rtd = substream->private_data;
@@ -206,6 +228,30 @@ static const struct snd_soc_dapm_widget sc7180_snd_widgets[] = {
SND_SOC_DAPM_MIC("Headset Mic", NULL),
};

+static const char * const dmic_mux_text[] = {
+ "Front Mic",
+ "Rear Mic",
+};
+
+static SOC_ENUM_SINGLE_DECL(sc7180_dmic_enum,
+ SND_SOC_NOPM, 0, dmic_mux_text);
+
+static const struct snd_kcontrol_new sc7180_dmic_mux_control =
+ SOC_DAPM_ENUM_EXT("DMIC Select Mux", sc7180_dmic_enum,
+ dmic_get, dmic_set);
+
+static const struct snd_soc_dapm_widget sc7180_snd_dual_mic_widgets[] = {
+ SND_SOC_DAPM_HP("Headphone Jack", NULL),
+ SND_SOC_DAPM_MIC("Headset Mic", NULL),
+ SND_SOC_DAPM_MIC("DMIC", NULL),
+ SND_SOC_DAPM_MUX("Dmic Mux", SND_SOC_NOPM, 0, 0, &sc7180_dmic_mux_control),
+};
+
+static const struct snd_soc_dapm_route sc7180_snd_dual_mic_audio_route[] = {
+ {"Dmic Mux", "Front Mic", "DMIC"},
+ {"Dmic Mux", "Rear Mic", "DMIC"},
+};
+
static void sc7180_add_ops(struct snd_soc_card *card)
{
struct snd_soc_dai_link *link;
@@ -238,6 +284,19 @@ static int sc7180_snd_platform_probe(struct platform_device *pdev)
card->dapm_widgets = sc7180_snd_widgets;
card->num_dapm_widgets = ARRAY_SIZE(sc7180_snd_widgets);

+ if (of_property_read_bool(dev->of_node, "dmic-gpios")) {
+ card->dapm_widgets = sc7180_snd_dual_mic_widgets,
+ card->num_dapm_widgets = ARRAY_SIZE(sc7180_snd_dual_mic_widgets),
+ card->dapm_routes = sc7180_snd_dual_mic_audio_route,
+ card->num_dapm_routes = ARRAY_SIZE(sc7180_snd_dual_mic_audio_route),
+ dmic_sel = devm_gpiod_get(&pdev->dev, "dmic", GPIOD_OUT_LOW);
+ if (IS_ERR(dmic_sel)) {
+ dev_err(&pdev->dev, "DMIC gpio failed err=%d\n",
+ PTR_ERR(dmic_sel));
+ return PTR_ERR(dmic_sel);
+ }
+ }
+
ret = qcom_snd_parse_of(card);
if (ret)
return ret;
--
2.25.1


2020-10-30 09:25:30

by Tzung-Bi Shih

[permalink] [raw]
Subject: Re: [PATCH v2 2/2] ASoC: qcom: sc7180: Modify machine driver for 2mic

On Fri, Oct 30, 2020 at 3:57 PM Ajye Huang <[email protected]> wrote:
> +static struct gpio_desc *dmic_sel;
> +static int dmic_switch;

If you really need them, you should put them in struct sc7180_snd_data.

> +static int dmic_set(struct snd_kcontrol *kcontrol,
> + struct snd_ctl_elem_value *ucontrol)
> +{
> + if (dmic_sel) {

if (IS_ERR(dmic_sel))
But I think you don't need to check dmic_sel. Suppose your _probe()
already returned error, the code here shouldn't be called.

> + dmic_switch = ucontrol->value.integer.value[0];

Looks like it can be a local variable. You don't need to save dmic_switch.

2020-10-30 10:57:59

by Ajye Huang

[permalink] [raw]
Subject: Re: [PATCH v2 2/2] ASoC: qcom: sc7180: Modify machine driver for 2mic

On Fri, Oct 30, 2020 at 5:23 PM Tzung-Bi Shih <[email protected]> wrote:
>
> On Fri, Oct 30, 2020 at 3:57 PM Ajye Huang <[email protected]> wrote:
> > +static struct gpio_desc *dmic_sel;
> > +static int dmic_switch;
>
> If you really need them, you should put them in struct sc7180_snd_data.
>
Thank you, I will do it.
> > +static int dmic_set(struct snd_kcontrol *kcontrol,
> > + struct snd_ctl_elem_value *ucontrol)
> > +{
> > + if (dmic_sel) {
>
> if (IS_ERR(dmic_sel))
> But I think you don't need to check dmic_sel. Suppose your _probe()
> already returned error, the code here shouldn't be called.
>
OK, I will remove the discriminant if (dmic_sel) {
> > + dmic_switch = ucontrol->value.integer.value[0];
>
> Looks like it can be a local variable. You don't need to save dmic_switch.
But dmic_get() will need dmic_switch, should i keep dmic_switch?
static int dmic_get(struct snd_kcontrol *kcontrol,
struct snd_ctl_elem_value *ucontrol)
{
ucontrol->value.integer.value[0] = dmic_switch;
return 0;
}

2020-10-30 11:23:43

by Tzung-Bi Shih

[permalink] [raw]
Subject: Re: [PATCH v2 2/2] ASoC: qcom: sc7180: Modify machine driver for 2mic

On Fri, Oct 30, 2020 at 6:55 PM Ajye Huang
<[email protected]> wrote:
> But dmic_get() will need dmic_switch, should i keep dmic_switch?

I see. I overlooked it. You can keep the dmic_switch for this
purpose or just call gpiod_get_value_cansleep().

2020-11-02 05:08:33

by ajye huang

[permalink] [raw]
Subject: Re: [PATCH v2 2/2] ASoC: qcom: sc7180: Modify machine driver for 2mic

Hi, Tzung-Bi
I will release v3 version base on your suggestion

Thank you for your reply
Ajye

Tzung-Bi Shih <[email protected]> 於 2020年10月30日 週五 下午7:19寫道:
>
> On Fri, Oct 30, 2020 at 6:55 PM Ajye Huang
> <[email protected]> wrote:
> > But dmic_get() will need dmic_switch, should i keep dmic_switch?
>
> I see. I overlooked it. You can keep the dmic_switch for this
> purpose or just call gpiod_get_value_cansleep().