Re: [PATCH v3 6/6] gpio: tps65086: Add GPIO driver for the TPS65086 PMIC

["Andrew F. Davis" <afd@ti.com>]

On 11/18/2015 03:44 AM, Linus Walleij wrote:
> On Tue, Nov 17, 2015 at 5:11 PM, Andrew F. Davis <afd@ti.com> wrote:
>> On 11/17/2015 03:17 AM, Linus Walleij wrote:
>>>
>>> On Wed, Nov 4, 2015 at 6:12 PM, Andrew F. Davis <afd@ti.com> wrote:
>>>
>>>> Add support for the TPS65086 PMIC GPOs.
>>>>
>>>> TPS65086 has four configurable GPOs that can be used for several
>>>> purposes.
>>>>
>>>> Signed-off-by: Andrew F. Davis <afd@ti.com>
>>>
>>>
>>> OK...
>>>
>>>> +static int tps65086_gpio_get(struct gpio_chip *gc, unsigned offset)
>>>> +static void tps65086_gpio_set(struct gpio_chip *gc, unsigned offset,
>>>
>>>
>>> Just get/set and no get_direction/direction_input/direction_output?
>>> Are you sure?
>>>
>>
>> Yeah, these are output only, I could probably add get_direction and just
>> always return output, but setters wouldn't make sense here.
>
> It's important that you note in the driver, commit blurb and maybe even
> Kconfig that these GPIOs are output-only. Someone could get confused.
>
> You should implement .direction_output() and return 0, and implement
> .direction_input and return negative error code. As it is now, it will
> seem like input is working, while it's not, right?
>
> Yours,
> Linus Walleij
>

Right, will add.

Also I would hold off on taking the GPIO bindings just yet. Mark Brown
has stated he would like the compatible string removed from the regulator
sub-node before he takes the regulator components, so for consistency sake
I am also removing the string from GPIO as well. With this, it is difficult
to have the sub-node still be a GPIO controller.

( gpiochip->of_node vs gpiochip->dev->of_node gpiolib.c:694 etc.. )

So the main pmic node will also be the GPIO controller node.

pmic: tps65912@2d {
	compatible = "ti,tps65912";
	reg = <0x2d>;

	interrupt-parent = <&gpio1>;
	interrupts = <28 8>;

	interrupt-controller;
	#interrupt-cells = <2>;

	gpio-controller;
	#gpio-cells = <2>;
};