Re: [PATCH v2] rtc: omap: Support ext_wakeup configuration

[Tony Lindgren <tony@atomide.com>]

* Grygorii Strashko <grygorii.strashko@ti.com> [160408 03:43]:
> On 04/08/2016 01:08 PM, Marcin Niestroj wrote:
> >>>>>>> @@ -18,8 +18,12 @@ Optional properties:
> >>>>>>>      through pmic_power_en
> >>>>>>>    - clocks: Any internal or external clocks feeding in to rtc
> >>>>>>>    - clock-names: Corresponding names of the clocks
> >>>>>>> +- gpio-controller: Mark as gpio controller when using ext_wakeup
> >>>>>>> +- #gpio-cells: Should be set to 2
> >>>>>>> +- ngpios: Number of ext_wakeup sources supported by processor 
> >>>>>>> (board)
> >>>>>>> +- ext-wakeup-gpios: List of ext_wakeup sources to configure

I think the naming standard here should be gpio-*, so in this
case gpio-wakeup.

> >>> We don't have power-button connected right to the processor. It is
> >>> connected to PMIC. During runtime we receive IRQs about power-button
> >>> from PMIC using i2c bus. The only purpose of this patch is to
> >>> configure processor's ext_wakeup line, which is triggered during
> >>> RTC-only mode (for example when power-button is pressed), causing
> >>> device wakeup. On the other hand, it is not possible to use ext_wakeup
> >>> during runtime, as we are only able to read it's status, but it
> >>> cannot trigger any interrupts.
> >>
> >> Sry, but I don't like this approach - it could make sense if RTC
> >> EXT_WAKEUP will be at least partially mapped on gpiolib interface.
> >> But your gpiochip is fake, you do not/can't use GPIO hogging mechanism
> >> and you're even parsing DT on your own (in V3).
> > 
> > With gpio hogging we can't pass polarity to the driver. It is hidden
> > in gpiolib.
> 
> kirkwood-openrd.dtsi-			p2 {
> kirkwood-openrd.dtsi:				gpio-hog;
> kirkwood-openrd.dtsi-				gpios = <2 GPIO_ACTIVE_HIGH>;
> 						[input;]
> 
> Sry, if you can't do smth like above - it's just prove that this approach is not right.

Yes let's stick to the standards. It should be using gpiolib interface.

Regards,

Tony