Skip to contentSkip to conversation

CHROMIUM: ASoC: Intel: boards: Use FS as nau8825 sysclk in nau88l25_max98357a machine

PATCH In review Grace Kao 3 participants 4 messages Nov 17, 2016 – Nov 18, 2016

Subsystems (3)
lore
Share & tools
lore

Grace Kao

author
#0 lore raw

PATCH 1/1 CHROMIUM: ASoC: Intel: boards: Use FS as nau8825 sysclk in nau88l25_max98357a machine

256FS sysclk gives the best audio quality per nau8825 datasheet.

Loud headphone pop happens if there is no sysclk during nau8825 playback power up sequence. Currently Skylake does not output MCLK/FS when the back-end DAI op hw_param is called, so we cannot switch to MCLK/FS in hw_param. This patch reduces pop by leting nau8825 keep using its internal VCO clock during widget power up sequence, until SNDRV_PCM_TRIGGER_START when MCLK/FS is available.

This is a temporary fix before Intel ADSP can output MCLK/FS early.

BUG=chrome-os-partner:50471
TEST=emerge-sentry sys-kernel/chromeos-kernel-3_18. Verified a local build that could reduce the pop sound.

Change-Id I4943fffa8eb1c874cc66b50f0897092bf48dd579

Reviewed-on: https://chromium-review.googlesource.com/333180

Series-to: [email protected]
Series-cc: [email protected]
---

sound/soc/intel/boards/skl_nau88l25_max98357a.c +28 -26
index 0f89602..6ad8f52 100644
@@ -70,14 +70,7 @@
 		return -EIO;
 	}
 
-	if (SND_SOC_DAPM_EVENT_ON(event)) {
-		ret = snd_soc_dai_set_sysclk(codec_dai,
-				NAU8825_CLK_MCLK, 24000000, SND_SOC_CLOCK_IN);
-		if (ret < 0) {
-			dev_err(card->dev, "set sysclk err = %d\n", ret);
-			return -EIO;
-		}
-	} else {
+	if (!SND_SOC_DAPM_EVENT_ON(event)) {
 		ret = snd_soc_dai_set_sysclk(codec_dai,
 				NAU8825_CLK_INTERNAL, 0, SND_SOC_CLOCK_IN);
 		if (ret < 0) {
@@ -272,25 +265,7 @@
 	.startup = skl_fe_startup,
 };
 
-static int skylake_nau8825_hw_params(struct snd_pcm_substream *substream,
-	struct snd_pcm_hw_params *params)
-{
-	struct snd_soc_pcm_runtime *rtd = substream->private_data;
-	struct snd_soc_dai *codec_dai = rtd->codec_dai;
-	int ret;
 
-	ret = snd_soc_dai_set_sysclk(codec_dai,
-			NAU8825_CLK_MCLK, 24000000, SND_SOC_CLOCK_IN);
-
-	if (ret < 0)
-		dev_err(rtd->dev, "snd_soc_dai_set_sysclk err = %d\n", ret);
-
-	return ret;
-}
-
-static struct snd_soc_ops skylake_nau8825_ops = {
-	.hw_params = skylake_nau8825_hw_params,
-};
 
 static int skylake_dmic_fixup(struct snd_soc_pcm_runtime *rtd,
 		struct snd_pcm_hw_params *params)
@@ -306,6 +281,33 @@
 	return 0;
 }
 
+static int skylake_nau8825_trigger(struct snd_pcm_substream *substream, int cmd)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct snd_pcm_runtime *runtime = substream->runtime;
+	struct snd_soc_dai *codec_dai = rtd->codec_dai;
+	int ret = 0;
+
+	switch (cmd) {
+	case SNDRV_PCM_TRIGGER_START:
+		ret = snd_soc_dai_set_sysclk(codec_dai,	NAU8825_CLK_FLL_FS, 0,
+			SND_SOC_CLOCK_IN);
+		if (ret < 0)
+			dev_err(codec_dai->dev, "can't set FS clock %d\n", ret);
+		ret = snd_soc_dai_set_pll(codec_dai, 0, 0, runtime->rate,
+			runtime->rate * 256);
+		if (ret < 0)
+			dev_err(codec_dai->dev, "can't set FLL: %d\n", ret);
+		break;
+	}
+
+	return ret;
+}
+
+static struct snd_soc_ops skylake_nau8825_ops = {
+	.trigger = skylake_nau8825_trigger,
+};
+
 static unsigned int channels_dmic[] = {
 	2, 4,
 };
#1 lore raw
Quoted 8 lines · Grace Kao

256FS sysclk gives the best audio quality per nau8825 datasheet.

Loud headphone pop happens if there is no sysclk during nau8825 playback power up sequence. Currently Skylake does not output MCLK/FS when the back-end DAI op hw_param is called, so we cannot switch to MCLK/FS in hw_param. This patch reduces pop by leting nau8825 keep using its internal VCO clock during widget power up sequence, until SNDRV_PCM_TRIGGER_START when MCLK/FS is available.

Please CC maintainers Mark and Liam on ASoC patches

The tag CHROMIUM is not fit for upstream please remove

Quoted 6 lines

This is a temporary fix before Intel ADSP can output MCLK/FS early.

BUG=chrome-os-partner:50471
TEST=emerge-sentry sys-kernel/chromeos-kernel-3_18. Verified a local build that could reduce the pop sound.

This is just noise for upstream, please remove

Quoted 2 lines

Change-Id: I4943fffa8eb1c874cc66b50f0897092bf48dd579

This as well

Quoted 2 lines

Reviewed-on: https://chromium-review.googlesource.com/333180

And this

Quoted 4 lines

Series-to: [email protected]
Series-cc: [email protected]

here too

Quoted 9 lines

---

sound/soc/intel/boards/skl_nau88l25_max98357a.c +0 -0
index 0f89602..6ad8f52 100644
@@ -70,14 +70,7 @@
 		return -EIO;
 	}

sound/soc/intel/boards/skl_nau88l25_max98357a.c line 73

Quoted 10 lines
sound/soc/intel/boards/skl_nau88l25_max98357a.c +1 -8
 
-	if (SND_SOC_DAPM_EVENT_ON(event)) {
-		ret = snd_soc_dai_set_sysclk(codec_dai,
-				NAU8825_CLK_MCLK, 24000000, SND_SOC_CLOCK_IN);
-		if (ret < 0) {
-			dev_err(card->dev, "set sysclk err = %d\n", ret);
-			return -EIO;
-		}
-	} else {
+	if (!SND_SOC_DAPM_EVENT_ON(event)) {

SND_SOC_DAPM_EVENT_OFF ??

Quoted 8 lines
sound/soc/intel/boards/skl_nau88l25_max98357a.c +8 -0
+static int skylake_nau8825_trigger(struct snd_pcm_substream *substream, int cmd)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct snd_pcm_runtime *runtime = substream->runtime;
+	struct snd_soc_dai *codec_dai = rtd->codec_dai;
+	int ret = 0;
+
+	switch (cmd) {

sound/soc/intel/boards/skl_nau88l25_max98357a.c line 301

Quoted 10 lines
sound/soc/intel/boards/skl_nau88l25_max98357a.c +10 -0
+	case SNDRV_PCM_TRIGGER_START:
+		ret = snd_soc_dai_set_sysclk(codec_dai,	NAU8825_CLK_FLL_FS, 0,
+			SND_SOC_CLOCK_IN);
+		if (ret < 0)
+			dev_err(codec_dai->dev, "can't set FS clock %d\n", ret);
+		ret = snd_soc_dai_set_pll(codec_dai, 0, 0, runtime->rate,
+			runtime->rate * 256);
+		if (ret < 0)
+			dev_err(codec_dai->dev, "can't set FLL: %d\n", ret);
+		break;

starting clock on trigger doesnt sound right to me..

Signature · 1 line
~Vinod
#2 lore raw
Quoted 9 lines · Vinod Koul
Quoted 8 lines · Grace Kao

256FS sysclk gives the best audio quality per nau8825 datasheet.

Loud headphone pop happens if there is no sysclk during nau8825 playback power up sequence. Currently Skylake does not output MCLK/FS when the back-end DAI op hw_param is called, so we cannot switch to MCLK/FS in hw_param. This patch reduces pop by leting nau8825 keep using its internal VCO clock during widget power up sequence, until SNDRV_PCM_TRIGGER_START when MCLK/FS is available.

<snip>

Quoted 13 lines
Quoted 13 lines
sound/soc/intel/boards/skl_nau88l25_max98357a.c +12 -0
+static int skylake_nau8825_trigger(struct snd_pcm_substream *substream, int cmd)
+{
+	struct snd_soc_pcm_runtime *rtd = substream->private_data;
+	struct snd_pcm_runtime *runtime = substream->runtime;
+	struct snd_soc_dai *codec_dai = rtd->codec_dai;
+	int ret = 0;
+
+	switch (cmd) {
+	case SNDRV_PCM_TRIGGER_START:
+		ret = snd_soc_dai_set_sysclk(codec_dai,	NAU8825_CLK_FLL_FS, 0,
+			SND_SOC_CLOCK_IN);
+		if (ret < 0)
Quoted 9 lines

sound/soc/intel/boards/skl_nau88l25_max98357a.c line 301

Quoted 6 lines
sound/soc/intel/boards/skl_nau88l25_max98357a.c +6 -0
+			dev_err(codec_dai->dev, "can't set FS clock %d\n", ret);
+		ret = snd_soc_dai_set_pll(codec_dai, 0, 0, runtime->rate,
+			runtime->rate * 256);
+		if (ret < 0)
+			dev_err(codec_dai->dev, "can't set FLL: %d\n", ret);
+		break;

starting clock on trigger doesnt sound right to me..

Me neither can't trigger be called from an atomic context? In which case I would expect operations on an external CODEC to be very likely to cause issues.

Thanks,
Charles

#3 lore raw
Quoted 16 lines · Charles Keepax
Quoted 6 lines
Quoted 6 lines
sound/soc/intel/boards/skl_nau88l25_max98357a.c +6 -0
+
+	switch (cmd) {
+	case SNDRV_PCM_TRIGGER_START:
+		ret = snd_soc_dai_set_sysclk(codec_dai,	NAU8825_CLK_FLL_FS, 0,
+			SND_SOC_CLOCK_IN);
+		if (ret < 0)
Quoted 9 lines

sound/soc/intel/boards/skl_nau88l25_max98357a.c line 301

Quoted 6 lines
sound/soc/intel/boards/skl_nau88l25_max98357a.c +6 -0
+			dev_err(codec_dai->dev, "can't set FS clock %d\n", ret);
+		ret = snd_soc_dai_set_pll(codec_dai, 0, 0, runtime->rate,
+			runtime->rate * 256);
+		if (ret < 0)
+			dev_err(codec_dai->dev, "can't set FLL: %d\n", ret);
+		break;

starting clock on trigger doesnt sound right to me..

Quoted 3 lines

Me neither can't trigger be called from an atomic context? In which case I would expect operations on an external CODEC to be very likely to cause issues.

Not in this case :) We use non atomic triggers.. But that doesnt make this correct :)

Signature · 1 line
~Vinod