[PATCH] ASoC: qcom: lpass-cpu: Fix pop noise during audio capture begin
This patch fixes PoP noise of around 15ms observed during audio capture begin. Enables BCLK and LRCLK in snd_soc_dai_ops startup call for introducing some delay before capture start and clock enable.
Signed-off-by: Judy Hsiao judyhsiao@chromium.org Signed-off-by: Srinivasa Rao Mandadapu srivasam@codeaurora.org --- sound/soc/qcom/lpass-cpu.c | 37 +++++++++++++++++++++++++++++++++---- 1 file changed, 33 insertions(+), 4 deletions(-)
diff --git a/sound/soc/qcom/lpass-cpu.c b/sound/soc/qcom/lpass-cpu.c index c62d2612e8f5..3dcfc2f115a7 100644 --- a/sound/soc/qcom/lpass-cpu.c +++ b/sound/soc/qcom/lpass-cpu.c @@ -73,14 +73,28 @@ static int lpass_cpu_daiops_startup(struct snd_pcm_substream *substream, struct snd_soc_dai *dai) { struct lpass_data *drvdata = snd_soc_dai_get_drvdata(dai); - int ret; + struct lpaif_i2sctl *i2sctl = drvdata->i2sctl; + unsigned int id = dai->driver->id; + int ret = -EINVAL;
ret = clk_prepare_enable(drvdata->mi2s_osr_clk[dai->driver->id]); if (ret) { dev_err(dai->dev, "error in enabling mi2s osr clk: %d\n", ret); return ret; } - ret = clk_prepare(drvdata->mi2s_bit_clk[dai->driver->id]); + + if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) { + ret = regmap_fields_write(i2sctl->spken, id, + LPAIF_I2SCTL_SPKEN_ENABLE); + } else { + ret = regmap_fields_write(i2sctl->micen, id, + LPAIF_I2SCTL_MICEN_ENABLE); + } + if (ret) + dev_err(dai->dev, "error writing to i2sctl reg: %d\n", + ret); + + ret = clk_prepare_enable(drvdata->mi2s_bit_clk[dai->driver->id]); if (ret) { dev_err(dai->dev, "error in enabling mi2s bit clk: %d\n", ret); clk_disable_unprepare(drvdata->mi2s_osr_clk[dai->driver->id]); @@ -93,9 +107,23 @@ static void lpass_cpu_daiops_shutdown(struct snd_pcm_substream *substream, struct snd_soc_dai *dai) { struct lpass_data *drvdata = snd_soc_dai_get_drvdata(dai); + struct lpaif_i2sctl *i2sctl = drvdata->i2sctl; + unsigned int id = dai->driver->id; + int ret = -EINVAL;
clk_disable_unprepare(drvdata->mi2s_osr_clk[dai->driver->id]); - clk_unprepare(drvdata->mi2s_bit_clk[dai->driver->id]); + clk_disable_unprepare(drvdata->mi2s_bit_clk[dai->driver->id]); + + if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) { + ret = regmap_fields_write(i2sctl->spken, id, + LPAIF_I2SCTL_SPKEN_DISABLE); + } else { + ret = regmap_fields_write(i2sctl->micen, id, + LPAIF_I2SCTL_MICEN_DISABLE); + } + if (ret) + dev_err(dai->dev, "error writing to i2sctl reg: %d\n", + ret); }
static int lpass_cpu_daiops_hw_params(struct snd_pcm_substream *substream, @@ -308,7 +336,8 @@ static int lpass_cpu_daiops_trigger(struct snd_pcm_substream *substream, ret);
clk_disable(drvdata->mi2s_bit_clk[dai->driver->id]); - + break; + default: break; }
On Thu, May 13, 2021 at 05:15:39PM +0530, Srinivasa Rao Mandadapu wrote:
- if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) {
ret = regmap_fields_write(i2sctl->spken, id,
LPAIF_I2SCTL_SPKEN_ENABLE);
- } else {
ret = regmap_fields_write(i2sctl->micen, id,
LPAIF_I2SCTL_MICEN_ENABLE);
- }
This commit doesn't remove the matching update in triger() so we'd have two redundant updates. I guess it's unlikely to be harmful but it looks wrong/confusing.
Hi Brown,
Thanks for your review comments!!!
On 5/13/2021 7:01 PM, Mark Brown wrote:
On Thu, May 13, 2021 at 05:15:39PM +0530, Srinivasa Rao Mandadapu wrote:
- if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) {
ret = regmap_fields_write(i2sctl->spken, id,
LPAIF_I2SCTL_SPKEN_ENABLE);
- } else {
ret = regmap_fields_write(i2sctl->micen, id,
LPAIF_I2SCTL_MICEN_ENABLE);
- }
This commit doesn't remove the matching update in triger() so we'd have two redundant updates. I guess it's unlikely to be harmful but it looks wrong/confusing.
Yes, It's not harmful, as clk_prepare_enable is enabling clock only once but maintaining count.
As Some times in Suspend/resume Sequence not hitting startup/shutdown, but Trigger, so for maintaining
consistency not removed in trigger.
On Fri, May 14, 2021 at 12:20:46PM +0530, Srinivasa Rao Mandadapu wrote:
On 5/13/2021 7:01 PM, Mark Brown wrote:
This commit doesn't remove the matching update in triger() so we'd have two redundant updates. I guess it's unlikely to be harmful but it looks wrong/confusing.
Yes, It's not harmful, as clk_prepare_enable is enabling clock only once but maintaining count.
As Some times in Suspend/resume Sequence not hitting startup/shutdown, but Trigger, so for maintaining
consistency not removed in trigger.
This at least needs some commenting so that it's clear, it looks buggy at the minute.
participants (2)
-
Mark Brown
-
Srinivasa Rao Mandadapu