Re: [PATCH 1/2] dt-bindings: power: supply: Add Richtek RT9471 battery charger

[ChiYuan Huang]

ChiYuan Huang <u0084500@xxxxxxxxx> 於 2022年8月12日 週五 晚上11:57寫道：
>
> Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx> 於 2022年8月12日 週五 下午2:54寫道：
> >
> > On 12/08/2022 04:32, ChiYuan Huang wrote:
> > > Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx> 於 2022年8月11日 週四 晚上10:12寫道：
> > >>
> > >> On 11/08/2022 16:41, cy_huang wrote:
> > >>> From: ChiYuan Huang <cy_huang@xxxxxxxxxxx>
> > >>>
> > >>> Add bindings for the Richtek RT9471 I2C controlled battery charger.
> > >>>
> > >>
> > >> Thank you for your patch. There is something to discuss/improve.
> > >>
> > >>> +properties:
> > >>> +  compatible:
> > >>> +    const: richtek,rt9471
> > >>> +
> > >>> +  reg:
> > >>> +    maxItems: 1
> > >>> +
> > >>> +  ceb-gpios:
> > >>> +    maxItems: 1
> > >>
> > >> This looks not standard, so please provide a description.
> > > It's the external 'charge enable' pin that's used to control battery charging.
> > > The priority is higher than the register 'CHG_EN' control.
> > > In the word, 'b' means it's reverse logic, low to allow charging, high
> > > to force disable charging.
> >
> > Isn't this standard enable-gpios property?
> Not the same thing, this charger includes power patch control.
> This gpio is used to 'force disable' charge the battery.
> >
> > >
> > > description:
> > >   External charge enable pin that can force control not to charge the battery.
> > >   Low to allow charging, high to disable charging.
> > >
> > >>
> > >>> +
> > >>> +  wakeup-source: true
> > >>> +
> > >>> +  interrupts:
> > >>> +    maxItems: 1
> > >>> +
> > >>> +  interrupt-controller: true
> > >>> +
> > >>> +  "#interrupt-cells":
> > >>> +    const: 1
> > >>
> > >> Why a charger driver is a interrupt-controller?
> > > There're 32 nested IRQs from RT9471.
> > > The original thought is to make the user easy to bind the interrupt
> > > into their driver.
> >
> > Bindings are not related to the driver but to hardware...
> >
> Sorry, I mislead  your comment.
> Refer to bq2515x.yaml, I think it's better to change this property to
> 'charge-enable-gpios'.
> It's the same usage like as TI charger.
> > >
> > > For charger driver, does it mean legacy IRQ handler is more preferred?
> >
> > Who is the consumer of these interrupts? Can you show the DTS with the
> > interrupt consumer?
> >
Sorry, I forget to reply this question.
Some battery driver may need to know the 'full', 'recharge' , 'ieoc' status.
The usage will  be like as below

battery {
  interrupts-extended = <&rt9471_chg 2 0>, <&rt9471_chg 3 0>, &(rt9471_chg 5 0>;
  interrupt-names = "chg-done", "chg-recharge", "chg-ieoc";
};

Some gauge HW needs this information to enhance the battery capacity accuracy.

> > >>
> > >>> +
> > >>> +  usb-otg-vbus-regulator:
> > >>> +    type: object
> > >>> +    unevaluatedProperties: false
> > >>> +    $ref: /schemas/regulator/regulator.yaml#
> > >>> +
> > >>> +required:
> > >>> +  - compatible
> > >>> +  - reg
> > >>> +  - wakeup-source
> > >>> +  - interrupts
> > >>> +  - interrupt-controller
> > >>> +  - "#interrupt-cells"
> > >>> +
> > >>> +additionalProperties: false
> > >>> +
> > >>> +examples:
> > >>> +  - |
> > >>> +    #include <dt-bindings/interrupt-controller/irq.h>
> > >>> +    i2c {
> > >>> +      #address-cells = <1>;
> > >>> +      #size-cells = <0>;
> > >>> +
> > >>> +      charger@53 {
> > >>> +        compatible = "richtek,rt9471";
> > >>> +        reg = <0x53>;
> > >>> +        ceb-gpios = <&gpio26 1 0>;
> > >>
> > >> Isn't the last value a GPIO flag? If yes, use appropriate define.
> > > I already specify GPIOD_OUT_LOW in the gpiod_request flag.
> >
> > It is not related to the DTS. Anyway writing "low" for a meaning of high
> > is not correct usually...
> >
> > > Do I need to convert the gpio request code to GPIOD_OUT_HIGH,
> > > and specify here as GPIO_ACTIVE_LOW?
> >
> > You need to properly describe the hardware. The polarity of logical
> > signal is defined by DTS, not by driver. It does not make sense to do it
> > in driver. What if on some board the signal is inverted?
> >
> From our discussion, binding example just keep the active level that the pin is.
> So 'GPIO_ACTIVE_LOW', thanks.
>
> All of the above will be fixed in the next revision.
>
> > Best regards,
> > Krzysztof