Re: [RFC PATCH] ASoC: ak4458: use reset control instead of reset gpio

From: Peter Ujfalusi
Date: Fri Nov 20 2020 - 08:41:14 EST



Hi Viorel,

On 19/11/2020 18.24, Viorel Suman wrote:
> Hi Peter,
>
>> DTS is supposed to look as follows:
>>>
>>> / {
>>> ak4458_reset: gpio-reset {
>>> compatible = "gpio-reset";
>>> reset-gpios = <&pca6416 4 GPIO_ACTIVE_LOW>;
>>> #reset-cells = <0>;
>>> initially-in-reset;
>>
>> I can not find anything resembling to this in next-20201119.
>> Where is the implementation and documentation for this gpio-reset?
>
> The board schematics is not publicly available; some info may be seen in DTS files below:
> https://source.codeaurora.org/external/imx/linux-imx/tree/arch/arm64/boot/dts/freescale/imx8mm-evk.dts?h=imx_5.4.24_2.1.0
> https://source.codeaurora.org/external/imx/linux-imx/tree/arch/arm64/boot/dts/freescale/imx8mm-ab2.dts?h=imx_5.4.24_2.1.0
> https://source.codeaurora.org/external/imx/linux-imx/tree/arch/arm64/boot/dts/freescale/imx8mp-ab2.dts?h=imx_5.4.24_2.1.0
>
> In examples above the GPIO is handled by machine driver - wrong approach given that
> it requires machine driver being probed before codec driver.

Right, so this gpio-reset driver is not in mainline. You are adding
support for something which does not exists ;)

>
>>> - ak4458->reset_gpiod = devm_gpiod_get_optional(ak4458->dev,
>> "reset",
>>> - GPIOD_OUT_LOW);
>>> - if (IS_ERR(ak4458->reset_gpiod))
>>> - return PTR_ERR(ak4458->reset_gpiod);
>>> + ak4458->reset = devm_reset_control_get_optional_shared(ak4458-
>>> dev, NULL);
>>> + if (IS_ERR(ak4458->reset))
>>> + return PTR_ERR(ak4458->reset);
>>
>> The binding documentation must be updated and you must support the gpio
>> way as well.
>
> Sure, make sense.
>
>> When I had this discussion around using the reset framework for shared
>> enable and/or reset pins it was suggested that _if_ such a driver makes
>> sense then it should internally handle (by using magic strings) the fallback
>> and work with pre-reset binding.
>
> Thanks, would appreciate if you point me to the discussion you had.

There were few iterations of it when I finally given up, I can quickly
find rfc v2:
https://lkml.org/lkml/2019/10/30/311

Probably in earlier or later series the reset framework was also discussed.

I ended up using GPIOD_FLAGS_BIT_NONEXCLUSIVE in the pcm3168a driver.
https://lkml.org/lkml/2019/11/13/411
- Péter

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki