Re: [PATCH] ASoC: rt1320-sdw: don't poll PDE state on power-down

[Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>]

On 4/28/26 09:37, Aaron Ma wrote:
> rt1320_pde23_event(PRE_PMD) writes REQ_POWER_STATE=PS3 then polls
> ACTUAL_POWER_STATE. DAPM fires PRE_PMD while the SoundWire data port
> is still streaming — the codec cannot reach PS3 until the port stops,
> so the poll always times out during playback. This blocks
> snd_ctl_elem_write() for 2-3s making mute unresponsive.
> 
> Remove the poll from PRE_PMD in rt1320_pde23_event() and
> rt1320_pde11_event(). The codec transitions to PS3 on its own once the
> data port becomes inactive. Keep the POST_PMU poll — on power-up the
> domain must reach PS0 before audio can flow.
> 
> Fixes: f465d10cd731 ("ASoC: rt1320: Add support for version C")
> Signed-off-by: Aaron Ma <aaron.ma@canonical.com>
> ---
>  sound/soc/codecs/rt1320-sdw.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/sound/soc/codecs/rt1320-sdw.c b/sound/soc/codecs/rt1320-sdw.c
> index 192faa431b5e9..e42e8b6de853e 100644
> --- a/sound/soc/codecs/rt1320-sdw.c
> +++ b/sound/soc/codecs/rt1320-sdw.c
> @@ -2000,7 +2000,6 @@ static int rt1320_pde11_event(struct snd_soc_dapm_widget *w,
>  		regmap_write(rt1320->regmap,
>  			SDW_SDCA_CTL(FUNC_NUM_MIC, RT1320_SDCA_ENT_PDE11,
>  				RT1320_SDCA_CTL_REQ_POWER_STATE, 0), ps3);
> -		rt1320_pde_transition_delay(rt1320, FUNC_NUM_MIC, RT1320_SDCA_ENT_PDE11, ps3);
>  		break;
>  	default:
>  		break;
> @@ -2028,7 +2027,6 @@ static int rt1320_pde23_event(struct snd_soc_dapm_widget *w,
>  		regmap_write(rt1320->regmap,
>  			SDW_SDCA_CTL(FUNC_NUM_AMP, RT1320_SDCA_ENT_PDE23,
>  				RT1320_SDCA_CTL_REQ_POWER_STATE, 0), ps3);
> -		rt1320_pde_transition_delay(rt1320, FUNC_NUM_AMP, RT1320_SDCA_ENT_PDE23, ps3);

no, sorry, that's clearly wrong.

When we touch the requested power state, we *shall* wait for the power state to be reached.
removing the transition delay is silly.

You would really need to find out why streaming still occurs when you get a DAPM event. That doesn't seem right in the first place.


>  		break;
>  	default:
>  		break;