Re: [PATCH 02/11] MIPS: qi_lb60: Migrate to devicetree

[Paul Cercueil <paul@crapouillou.net>]

Le jeu. 25 juil. 2019 à 19:47, Paul Burton <paul.burton@mips.com> a 
écrit :
> Hi Paul,
> 
> On Thu, Jul 25, 2019 at 06:02:06PM -0400, Paul Cercueil wrote:
>>  Move all the platform data to devicetree.
> 
> Nice! :)
> 
>>  The only bit dropped is the PWM beeper, which requires the PWM 
>> driver
>>  to be updated. I figured it's okay to remove it here since it's 
>> really
>>  a non-critical device, and it'll be re-introduced soon enough.
> 
> OK, I can see that being a price worth paying. Though it's possible to
> include the binding at least for that in this series I'd be even
> happier. Actually I see we already have
> 
>   Documentation/devicetree/bindings/pwm/ingenic,jz47xx-pwm.txt
> 
> in mainline - what needs to change with it?

The PWM driver will be updated to use the TCU clocks and the regmap 
provided
by the TCU driver. The PWM node will be a sub-node of the TCU one.

Additionally, there is this[1] ongoing discussion about PWM which makes
me uneasy about how to write the binding. So I'd rather not rush it,
because once the devicetree is written, it's ABI.

[1]: https://lkml.org/lkml/2019/5/22/607


>>  +	spi {
>>  +		compatible = "spi-gpio";
>>  +		#address-cells = <1>;
>>  +		#size-cells = <0>;
>>  +
>>  +		sck-gpios = <&gpc 23 GPIO_ACTIVE_HIGH>;
>>  +		mosi-gpios = <&gpc 22 GPIO_ACTIVE_HIGH>;
>>  +		cs-gpios = <&gpc 21 GPIO_ACTIVE_LOW>;
>>  +		num-chipselects = <1>;
>>  +
>>  +		spi@0 {
>>  +			compatible = "ili8960";
> 
> Should this be "ilitek,ili8960"?
> 
> Is there a binding & driver for this submitted somewhere? If not then 
> do
> we need this at all? It doesn't look like the existing platform data
> would actually lead to a driver being loaded so I'm wondering if we 
> can
> just drop this until such a driver (or at least a documented DT 
> binding)
> exists.

I can drop it. There is no driver for it, and I'm not even sure the LB60
has a ILI8960 in the first place.


> Thanks,
>     Paul