Re: [PATCH v2 1/2] ASoC: codecs: ak4619: Add optional powerdown GPIO

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] ASoC: ak4619: missing power-down GPIO assertion on unbind
- [Low] ASoC: ak4619: missing inclusion of <linux/delay.h> for msleep()
--

commit 1a6c5d3bb52b0196ff6b96586fdeceef456751a9
Author: Alexander Koch <mail@alexanderkoch.net>

ASoC: codecs: ak4619: Add optional powerdown GPIO

This commit adds an optional 'powerdown-gpios' device tree binding. If
defined, the PDN pin is configured during probe with a delay prior to
register access.

> diff --git a/sound/soc/codecs/ak4619.c b/sound/soc/codecs/ak4619.c
> index d9c9f6b200284..c4f0db1e51c8d 100644
> --- a/sound/soc/codecs/ak4619.c
> +++ b/sound/soc/codecs/ak4619.c
[ ... ]
> @@ -875,10 +876,20 @@ static int ak4619_i2c_probe(struct i2c_client *i2c)
>  
>  	i2c_set_clientdata(i2c, ak4619);
>  
> +	ak4619->pdn_gpio = devm_gpiod_get_optional(dev, "powerdown",
> +		GPIOD_OUT_LOW);
> +	if (IS_ERR(ak4619->pdn_gpio))
> +		return dev_err_probe(dev, PTR_ERR(ak4619->pdn_gpio),
> +			"powerdown GPIO request failed\n");
> +	if (ak4619->pdn_gpio)
> +		msleep(10);

[Severity: Low]
Does this code rely on transitive header inclusions for msleep()? It appears
that <linux/delay.h> is not explicitly included in this file, which might
cause build failures if core headers are refactored.

[ ... ]
> @@ -887,6 +898,8 @@ static int ak4619_i2c_probe(struct i2c_client *i2c)
>  	if (ret < 0) {
>  		dev_err(dev, "Failed to register ak4619 component: %d\n",
>  			ret);
> +		if (ak4619->pdn_gpio)
> +			gpiod_set_value_cansleep(ak4619->pdn_gpio, 1);
>  		return ret;
>  	}
>  

[Severity: Medium]
Will the device remain permanently powered on if the module is unloaded or
unbound? The probe error paths correctly revert the GPIO state to 1 to power
down the device, but ak4619_i2c_driver lacks a .remove callback to handle
the unbind path.

Could this state leak be addressed by using devm_add_action_or_reset() during
ak4619_i2c_probe() to ensure teardown symmetry when the device is unbound?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260929090821.1476257-1-mail@alexanderkoch.net?part=1