Re: [PATCH v3 09/17] ASoC: simple-amplifier: Introduce support for gpio-audio-amp

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] Mixing of independent stereo audio paths in the DAPM graph due to the use of a single PGA widget.
- [Low] Preprocessor macro `#define SIMPLE_AUDIO_SUPPORT_PGA` is defined inside a struct definition.
--

commit d8f2282bd54ac87735c5b891e08ae94ea9bfa366
Author: Herve Codina <herve.codina@bootlin.com>

ASoC: simple-amplifier: Introduce support for gpio-audio-amp

This commit introduces preliminary support for GPIO-driven audio amplifiers.
It adds compatible strings and DAPM tables for both mono and stereo variants.

> diff --git a/sound/soc/codecs/simple-amplifier.c b/sound/soc/codecs/simple-amplifier.c
> index 3e644c1c26960..1704cdbb7de5d 100644
> --- a/sound/soc/codecs/simple-amplifier.c
> +++ b/sound/soc/codecs/simple-amplifier.c
> @@ -12,6 +13,9 @@
>  #include <sound/soc.h>
>  
>  struct simple_amp_data {
> +	unsigned int supports;
> +#define SIMPLE_AUDIO_SUPPORT_PGA		BIT(0)
> +

This isn't a bug, but should this macro be defined outside the struct?

Macros are handled by the preprocessor and have file-wide scope, so placing it
inside the struct can be misleading and goes against standard kernel coding
style.

>  	const struct snd_soc_dapm_widget *dapm_widgets;

[ ... ]

> +static const struct snd_soc_dapm_route simple_amp_stereo_pga_dapm_routes[] = {
> +	{ "PGA", NULL, "INL" },
> +	{ "PGA", NULL, "INR" },
> +	{ "PGA", NULL, "vdd" },
> +	{ "OUTL", NULL, "PGA" },
> +	{ "OUTR", NULL, "PGA" },
> +};

Does this routing mix the independent stereo audio paths?

By routing both INL and INR into a single PGA widget, and then routing that
PGA to both OUTL and OUTR, DAPM will see valid audio paths crossing over
(INL to OUTR, and INR to OUTL). 

Could this cause incorrect power management behavior where playing audio on
just the left channel keeps the right channel's downstream widgets powered
up unnecessarily?

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260513081702.317117-1-herve.codina@bootlin.com?part=9