Re: [PATCH 2/2] mfd: arizona-spi: Add Android board ACPI table handling

[Lee Jones]

On Wed, 23 Feb 2022, Hans de Goede wrote:

> x86/ACPI boards with an arizona WM5102 codec ship with either Windows or
> Android as factory installed OS.
> 
> The ACPI fwnode for the codec on Android boards misses 2 things compared
> to the Windows boards (this is hardcoded in the Android board kernels):
> 
> 1. There is no CLKE ACPI method to enabe the 32 KHz clock the codec needs
>    for jack-detection.
> 
> 2. The GPIOs used by the codec are not listed in the fwnode for the codec.
> 
> The ACPI tables on x86/ACPI boards shipped with Android being incomplete
> happens a lot. The special drivers/platform/x86/x86-android-tablets.c
> module contains DMI based per model handling to compensate for this.
> 
> This module will enable the 32KHz clock through the pinctrl framework
> to fix 1. and it will also register a gpio-lookup table for all GPIOs
> needed by the codec + machine driver, including the GPIOs coming from
> the codec itself.
> 
> Add an arizona_spi_acpi_android_probe() function which waits for the
> x86-android-tablets to have set things up before continue with probing
> the arizona WM5102 codec.
> 
> Signed-off-by: Hans de Goede <hdegoede@xxxxxxxxxx>
> ---
>  drivers/mfd/arizona-spi.c | 34 +++++++++++++++++++++++++++++++++-
>  1 file changed, 33 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/mfd/arizona-spi.c b/drivers/mfd/arizona-spi.c
> index 238355542ab1..2c686e71db21 100644
> --- a/drivers/mfd/arizona-spi.c
> +++ b/drivers/mfd/arizona-spi.c
> @@ -81,6 +81,29 @@ static int arizona_spi_acpi_windows_probe(struct arizona *arizona)
>  	return 0;
>  }
>  
> +/* For ACPI tables from boards which ship with Android as factory OS */
> +static int arizona_spi_acpi_android_probe(struct arizona *arizona)
> +{
> +	int ret;
> +
> +	/*
> +	 * Get the reset GPIO, treating -ENOENT as -EPROBE_DEFER to wait for
> +	 * the x86-android-tablets module to register the board specific GPIO
> +	 * lookup table.
> +	 */
> +	arizona->pdata.reset = devm_gpiod_get(arizona->dev, "reset", GPIOD_OUT_LOW);
> +	if (IS_ERR(arizona->pdata.reset)) {
> +		ret = PTR_ERR(arizona->pdata.reset);
> +		if (ret == -ENOENT) {
> +			dev_info_once(arizona->dev, "Deferring probe till GPIO lookup is registered\n");

Nit: How many chars is this?

I thought we were drawing the line at 100 these days?

Does this patch pass checkpatch.pl?

> +			ret = -EPROBE_DEFER;
> +		}
> +		return dev_err_probe(arizona->dev, ret, "getting reset GPIO\n");
> +	}
> +
> +	return 0;
> +}
> +
>  /*
>   * The AOSP 3.5 mm Headset: Accessory Specification gives the following values:
>   * Function A Play/Pause:           0 ohm
> @@ -102,9 +125,14 @@ static const struct arizona_micd_range arizona_micd_aosp_ranges[] = {
>  
>  static int arizona_spi_acpi_probe(struct arizona *arizona)
>  {
> +	struct acpi_device *adev = ACPI_COMPANION(arizona->dev);
>  	int ret;
>  
> -	ret = arizona_spi_acpi_windows_probe(arizona);
> +	if (acpi_dev_hid_uid_match(adev, "10WM5102", NULL))
> +		ret = arizona_spi_acpi_android_probe(arizona);
> +	else
> +		ret = arizona_spi_acpi_windows_probe(arizona);
> +
>  	if (ret)
>  		return ret;
>  
> @@ -142,6 +170,10 @@ static const struct acpi_device_id arizona_acpi_match[] = {
>  		.id = "WM510205",
>  		.driver_data = WM5102,
>  	},
> +	{
> +		.id = "10WM5102",
> +		.driver_data = WM5102,
> +	},
>  	{ }
>  };
>  MODULE_DEVICE_TABLE(acpi, arizona_acpi_match);

-- 
Lee Jones [李琼斯]
Principal Technical Lead - Developer Services
Linaro.org │ Open source software for Arm SoCs
Follow Linaro: Facebook | Twitter | Blog