1. Patchew
  2. linux
  3. View series

[PATCH v2] ASoC: fsl_sai: Force a software reset when starting in consumer mode

Arun Raghavan posted 1 patch 3 months, 2 weeks ago
There is a newer version of this series
sound/soc/fsl/fsl_sai.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
[PATCH v2] ASoC: fsl_sai: Force a software reset when starting in consumer mode
Posted by Arun Raghavan 3 months, 2 weeks ago 
From: Arun Raghavan <arun@asymptotic.io>

In a setup with an external clock provider, when running the receiver
(arecord) and triggering an xrun with xrun_injection, we see a channel
swap/offset. This happens sometimes when running only the receiver, but
occurs reliably if a transmitter (aplay) is also concurrently running.

The theory is that SAI seems to lose track of frame sync during the
trigger stop -> trigger start cycle that occurs during an xrun. Doing
just a FIFO reset in this case does not suffice, and only a software
reset seems to get it back on track.

Signed-off-by: Arun Raghavan <arun@asymptotic.io>
Reported-by: Pieterjan Camerlynck <p.camerlynck@televic.com>
---

v2:
- Address build warning from kernel test robot

 sound/soc/fsl/fsl_sai.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/sound/soc/fsl/fsl_sai.c b/sound/soc/fsl/fsl_sai.c
index af1a168d35e3..d158352c7640 100644
--- a/sound/soc/fsl/fsl_sai.c
+++ b/sound/soc/fsl/fsl_sai.c
@@ -841,6 +841,18 @@ static int fsl_sai_trigger(struct snd_pcm_substream *substream, int cmd,
 	case SNDRV_PCM_TRIGGER_START:
 	case SNDRV_PCM_TRIGGER_RESUME:
 	case SNDRV_PCM_TRIGGER_PAUSE_RELEASE:
+		/*
+		 * Force a software reset if we are not the clock provider, as we
+		 * might have lost frame sync during xrun recovery.
+		 */
+		if (sai->is_consumer_mode[tx]) {
+			regmap_update_bits(sai->regmap,
+					FSL_SAI_xCSR(tx, ofs), FSL_SAI_CSR_SR,
+					FSL_SAI_CSR_SR);
+			regmap_update_bits(sai->regmap,
+					FSL_SAI_xCSR(tx, ofs), FSL_SAI_CSR_SR,
+					0);
+		}
 		regmap_update_bits(sai->regmap, FSL_SAI_xCSR(tx, ofs),
 				   FSL_SAI_CSR_FRDE, FSL_SAI_CSR_FRDE);
 
-- 
2.49.0
Re: [PATCH v2] ASoC: fsl_sai: Force a software reset when starting in consumer mode
Posted by Mark Brown 3 months, 2 weeks ago 
On Wed, Jun 25, 2025 at 07:56:16PM -0400, Arun Raghavan wrote:
> From: Arun Raghavan <arun@asymptotic.io>
> 
> In a setup with an external clock provider, when running the receiver
> (arecord) and triggering an xrun with xrun_injection, we see a channel
> swap/offset. This happens sometimes when running only the receiver, but
> occurs reliably if a transmitter (aplay) is also concurrently running.

Please don't send new patches or versions in reply to old threads, it
makes it harder to follow what's going on.
Re: [PATCH v2] ASoC: fsl_sai: Force a software reset when starting in consumer mode
Posted by Arun Raghavan 3 months, 2 weeks ago 
On Thu, 26 Jun 2025, at 7:08 AM, Mark Brown wrote:
> On Wed, Jun 25, 2025 at 07:56:16PM -0400, Arun Raghavan wrote:
>> From: Arun Raghavan <arun@asymptotic.io>
>> 
>> In a setup with an external clock provider, when running the receiver
>> (arecord) and triggering an xrun with xrun_injection, we see a channel
>> swap/offset. This happens sometimes when running only the receiver, but
>> occurs reliably if a transmitter (aplay) is also concurrently running.
>
> Please don't send new patches or versions in reply to old threads, it
> makes it harder to follow what's going on.

Ack, sorry about that!

-- Arun
Re: [PATCH v2] ASoC: fsl_sai: Force a software reset when starting in consumer mode
Posted by Shengjiu Wang 3 months, 2 weeks ago 
On Thu, Jun 26, 2025 at 7:58 AM Arun Raghavan <arun@arunraghavan.net> wrote:
>
> From: Arun Raghavan <arun@asymptotic.io>
>
> In a setup with an external clock provider, when running the receiver
> (arecord) and triggering an xrun with xrun_injection, we see a channel
> swap/offset. This happens sometimes when running only the receiver, but
> occurs reliably if a transmitter (aplay) is also concurrently running.
>
> The theory is that SAI seems to lose track of frame sync during the
> trigger stop -> trigger start cycle that occurs during an xrun. Doing
> just a FIFO reset in this case does not suffice, and only a software
> reset seems to get it back on track.
>
> Signed-off-by: Arun Raghavan <arun@asymptotic.io>
> Reported-by: Pieterjan Camerlynck <p.camerlynck@televic.com>
> ---
>
> v2:
> - Address build warning from kernel test robot
>
>  sound/soc/fsl/fsl_sai.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
>
> diff --git a/sound/soc/fsl/fsl_sai.c b/sound/soc/fsl/fsl_sai.c
> index af1a168d35e3..d158352c7640 100644
> --- a/sound/soc/fsl/fsl_sai.c
> +++ b/sound/soc/fsl/fsl_sai.c
> @@ -841,6 +841,18 @@ static int fsl_sai_trigger(struct snd_pcm_substream *substream, int cmd,
>         case SNDRV_PCM_TRIGGER_START:
>         case SNDRV_PCM_TRIGGER_RESUME:
>         case SNDRV_PCM_TRIGGER_PAUSE_RELEASE:
> +               /*
> +                * Force a software reset if we are not the clock provider, as we
> +                * might have lost frame sync during xrun recovery.
> +                */
> +               if (sai->is_consumer_mode[tx]) {
> +                       regmap_update_bits(sai->regmap,
> +                                       FSL_SAI_xCSR(tx, ofs), FSL_SAI_CSR_SR,
> +                                       FSL_SAI_CSR_SR);
> +                       regmap_update_bits(sai->regmap,
> +                                       FSL_SAI_xCSR(tx, ofs), FSL_SAI_CSR_SR,
> +                                       0);
> +               }

Which platform are you using?  and please add chip info in your commit
message.

This change can be moved to fsl_sai_config_disable(). that is:

--- a/sound/soc/fsl/fsl_sai.c
+++ b/sound/soc/fsl/fsl_sai.c
@@ -798,18 +798,16 @@ static void fsl_sai_config_disable(struct
fsl_sai *sai, int dir)
                           FSL_SAI_CSR_FR, FSL_SAI_CSR_FR);

        /*
-        * For sai master mode, after several open/close sai,
+        * For sai master/slave mode, after several open/close sai,
         * there will be no frame clock, and can't recover
         * anymore. Add software reset to fix this issue.
         * This is a hardware bug, and will be fix in the
         * next sai version.
         */
-       if (!sai->is_consumer_mode[tx]) {
-               /* Software Reset */
-               regmap_write(sai->regmap, FSL_SAI_xCSR(tx, ofs),
FSL_SAI_CSR_SR);
-               /* Clear SR bit to finish the reset */
-               regmap_write(sai->regmap, FSL_SAI_xCSR(tx, ofs), 0);
-       }
+       /* Software Reset */
+       regmap_write(sai->regmap, FSL_SAI_xCSR(tx, ofs), FSL_SAI_CSR_SR);
+       /* Clear SR bit to finish the reset */
+       regmap_write(sai->regmap, FSL_SAI_xCSR(tx, ofs), 0);
 }

Could you please try the above change to also work for your case?

Best regards
Shengjiu Wang

>                 regmap_update_bits(sai->regmap, FSL_SAI_xCSR(tx, ofs),
>                                    FSL_SAI_CSR_FRDE, FSL_SAI_CSR_FRDE);
>
> --
> 2.49.0
>

Patchew 2.1-devel-b67e4c2

© 2016 - 2025 Red Hat, Inc.