[PATCH] ASoC: cs42xx8: Move reset operation to runtime resume

[PATCH] ASoC: cs42xx8: Move reset operation to runtime resume

[Shengjiu Wang]

When codec is in idle state, the runtime suspend is executed, if codec
has been reset and user wants to use i2ctransfer command to get the
codec registers, i2c access will fail.

Move reset operation to the runtime resume stage to avoid such issues.

Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>
---
 sound/soc/codecs/cs42xx8.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/sound/soc/codecs/cs42xx8.c b/sound/soc/codecs/cs42xx8.c
index 6a925f3f7137..be215cd2a714 100644
--- a/sound/soc/codecs/cs42xx8.c
+++ b/sound/soc/codecs/cs42xx8.c
@@ -617,6 +617,7 @@ static int cs42xx8_runtime_resume(struct device *dev)
 		return ret;
 	}
 
+	gpiod_set_value_cansleep(cs42xx8->gpiod_reset, 1);
 	gpiod_set_value_cansleep(cs42xx8->gpiod_reset, 0);
 
 	ret = regulator_bulk_enable(ARRAY_SIZE(cs42xx8->supplies),
@@ -658,8 +659,6 @@ static int cs42xx8_runtime_suspend(struct device *dev)
 	regulator_bulk_disable(ARRAY_SIZE(cs42xx8->supplies),
 			       cs42xx8->supplies);
 
-	gpiod_set_value_cansleep(cs42xx8->gpiod_reset, 1);
-
 	clk_disable_unprepare(cs42xx8->clk);
 
 	return 0;
-- 
2.34.1

Re: [PATCH] ASoC: cs42xx8: Move reset operation to runtime resume

[Charles Keepax]

On Tue, Jun 17, 2025 at 03:43:34PM +0800, Shengjiu Wang wrote:
> When codec is in idle state, the runtime suspend is executed, if codec
> has been reset and user wants to use i2ctransfer command to get the
> codec registers, i2c access will fail.
> 
> Move reset operation to the runtime resume stage to avoid such issues.
> 
> Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>
> ---
>  sound/soc/codecs/cs42xx8.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/sound/soc/codecs/cs42xx8.c b/sound/soc/codecs/cs42xx8.c
> index 6a925f3f7137..be215cd2a714 100644
> --- a/sound/soc/codecs/cs42xx8.c
> +++ b/sound/soc/codecs/cs42xx8.c
> @@ -617,6 +617,7 @@ static int cs42xx8_runtime_resume(struct device *dev)
>  		return ret;
>  	}
>  
> +	gpiod_set_value_cansleep(cs42xx8->gpiod_reset, 1);
>  	gpiod_set_value_cansleep(cs42xx8->gpiod_reset, 0);
>  
>  	ret = regulator_bulk_enable(ARRAY_SIZE(cs42xx8->supplies),
> @@ -658,8 +659,6 @@ static int cs42xx8_runtime_suspend(struct device *dev)
>  	regulator_bulk_disable(ARRAY_SIZE(cs42xx8->supplies),
>  			       cs42xx8->supplies);
>  
> -	gpiod_set_value_cansleep(cs42xx8->gpiod_reset, 1);
> -

This makes no sense the line before this is disabling all the
supplies. This code is parking the chip in a turned off state
and it is better for it to be in reset for that. If you want to
access the chip directly from user-space during runtime suspend
you need to runtime resume it first, I believe that can be done
through sysfs.

Thanks,
Charles