RE: [PATCH V4 1/2] dt-bindings: leds: pwm: add enable-gpios property

[LI Qingwu]

> -----Original Message-----
> From: LI Qingwu <Qing-wu.Li@xxxxxxxxxxxxxxxxxxxxxxx>
> Sent: Friday, July 4, 2025 4:36 PM
> To: Krzysztof Kozlowski <krzk@xxxxxxxxxx>
> Cc: lee@xxxxxxxxxx; pavel@xxxxxxxxxx; robh@xxxxxxxxxx; krzk+dt@xxxxxxxxxx;
> conor+dt@xxxxxxxxxx; linux-leds@xxxxxxxxxxxxxxx; devicetree@xxxxxxxxxxxxxxx;
> linux-kernel@xxxxxxxxxxxxxxx; GEO-CHHER-bsp-development
> <bsp-development.geo@xxxxxxxxxxxxxxxxxxxx>
> Subject: RE: [PATCH V4 1/2] dt-bindings: leds: pwm: add enable-gpios property
> 
> 
> 
> > -----Original Message-----
> > From: Krzysztof Kozlowski <krzk@xxxxxxxxxx>
> > Sent: Friday, July 4, 2025 3:51 PM
> > To: LI Qingwu <Qing-wu.Li@xxxxxxxxxxxxxxxxxxxxxxx>
> > Cc: lee@xxxxxxxxxx; pavel@xxxxxxxxxx; robh@xxxxxxxxxx;
> > krzk+dt@xxxxxxxxxx;
> > conor+dt@xxxxxxxxxx; linux-leds@xxxxxxxxxxxxxxx;
> > conor+devicetree@xxxxxxxxxxxxxxx;
> > linux-kernel@xxxxxxxxxxxxxxx; GEO-CHHER-bsp-development
> > <bsp-development.geo@xxxxxxxxxxxxxxxxxxxx>
> > Subject: Re: [PATCH V4 1/2] dt-bindings: leds: pwm: add enable-gpios
> > property
> > On Thu, Jul 03, 2025 at 11:36:10AM +0000, LI Qingwu wrote:
> > >
> > >
> > > > -----Original Message-----
> > > > From: Krzysztof Kozlowski <krzk@xxxxxxxxxx>
> > > > Sent: Thursday, July 3, 2025 6:55 PM
> > > > To: LI Qingwu <Qing-wu.Li@xxxxxxxxxxxxxxxxxxxxxxx>;
> > > > lee@xxxxxxxxxx; pavel@xxxxxxxxxx; robh@xxxxxxxxxx;
> > > > krzk+dt@xxxxxxxxxx;
> > > > conor+dt@xxxxxxxxxx; linux-leds@xxxxxxxxxxxxxxx;
> > > > conor+devicetree@xxxxxxxxxxxxxxx;
> > > > linux-kernel@xxxxxxxxxxxxxxx
> > > > Cc: GEO-CHHER-bsp-development
> > > > <bsp-development.geo@xxxxxxxxxxxxxxxxxxxx>
> > > > Subject: Re: [PATCH V4 1/2] dt-bindings: leds: pwm: add
> > > > enable-gpios property
> > > >
> > > >
> > > >
> > > > On 03/07/2025 12:27, LI Qingwu wrote:
> > > > >
> > > > >
> > > > >> -----Original Message-----
> > > > >> From: Krzysztof Kozlowski <krzk@xxxxxxxxxx>
> > > > >> Sent: Thursday, July 3, 2025 5:59 PM
> > > > >> To: LI Qingwu <Qing-wu.Li@xxxxxxxxxxxxxxxxxxxxxxx>;
> > > > >> lee@xxxxxxxxxx; pavel@xxxxxxxxxx; robh@xxxxxxxxxx;
> > > > >> krzk+dt@xxxxxxxxxx;
> > > > >> conor+dt@xxxxxxxxxx; linux-leds@xxxxxxxxxxxxxxx;
> > > > >> conor+devicetree@xxxxxxxxxxxxxxx;
> > > > >> linux-kernel@xxxxxxxxxxxxxxx
> > > > >> Cc: GEO-CHHER-bsp-development
> > > > >> <bsp-development.geo@xxxxxxxxxxxxxxxxxxxx>
> > > > >> Subject: Re: [PATCH V4 1/2] dt-bindings: leds: pwm: add
> > > > >> enable-gpios property
> > > > >>
> > > >
> > > >
> > > > >>
> > > > >>
> > > > >> On 03/07/2025 11:34, LI Qingwu wrote:
> > > > >>> some pwm led driver chips like tps92380 require a separate
> > > > >>> enable signal
> > > > >>
> > > > >> Sentence starts with capital letter.
> > > > >>
> > > > >> tps92380 does not have dedicated enable pin. It has VDDIO,
> > > > >> which serves also enable purpose, but it is a supply.
> > > > >
> > > > > So this patch is unacceptable anyway?
> > > >
> > > > If you make this patch for tps92380, I think it is not correct.
> > > > You have entire commit msg to explain the hardware and all unusual
> > > > things. Having VDDIO and EN pin is unusual, because you do not
> > > > supply
> > power directly from GPIOs of a SoC.
> > > > All this should be explained.
> > >
> > > Thank you for the feedback, what about rename to power-supply with
> > regulator support ?
> > > Convert this patch to add regulator support instead of GPIO, using
> > > "power-supply" property to control LED power, or drop this patch and
> > > give up
> > upstream, or what's the better from your point of view?
> > > appreciate your guidance!
> >
> > Does it solve your problem? It is surprising that once you say it is
> > GPIO and once as regulator. How is it in your board?
> >
> > Regulator is looking as correct hardware description, so that patch
> > would be fine.
> 
> 
> Thank you for your feedback and clarification.
> On our hardware, the TPS92380's VDDIO/EN pin is connected to a GPIO, which
> is used to enable or disable the device.
> According to the datasheet, this pin is described as "enable input for the device
> as well as supply input (VDDIO) for digital pins". If describing this as a supply is
> the preferred and correct way for upstream, I can implement as supply
> regulator controlled by GPIO, and to model this behavior in device tree.
> 
> or if you have a better suggestion for such cases.
LT3743 also has also such EN/UVLO pin and it has serval microampere current
https://www.analog.com/en/products/lt3743.html#documentation

> >
> > Best regards,
> > Krzysztof