Re: [PATCH v2 1/2] dt-bindings: regulator: add pf8x00 PMIC

[Adrien Grassein <adrien.grassein@gmail.com>]

Hello,

Le ven. 11 déc. 2020 à 15:04, Mark Brown <broonie@kernel.org> a écrit :
>
> On Thu, Dec 10, 2020 at 11:16:28PM +0100, Adrien Grassein wrote:
> > Add a devicetree binding documentation for the pf8x00 regulator driver.
>
> Please don't send new patches in reply to old threads, it makes it hard
> to keep track of what is going on.

Sorry. Should I create a new mail each time I send a new version of the patch?

>
> > +          regulator-name:
> > +            pattern: "^ldo[1-4]$"
> > +            description:
> > +              should be ldo1", ..., "ldo4"
>
> This is part of the generic regulator binding and since it's for board
> specific usage it should not be constrained by the chip binding.

Ok.
>
> > +          nxp,vselect-en:
> > +            $ref: /schemas/types.yaml#definitions/flag
> > +            description:
> > +              Only available for ldo2. When specified, use the VSELECT
> > +              input pin of the chip to control the output voltage of the
> > +              ldo02 regulator. (3.3V if VSELECT is LOW, 1.8V if HIGH).
>
> How is VSELECT used without a binding specifying some mechanism for
> control?

I think that VSELECT input should be controlled by the sub system that
uses it (via maybe a GPIO).
On my board, it's directly controlled by another chip (so without a GPIO).

>
> > +          nxp,ilim-microamp:
> > +            $ref: /schemas/types.yaml#definitions/uint32
> > +            minimum: 2100
> > +            maximum: 4500
> > +            default: 2100
> > +            enum: [ 2100, 2600, 3000, 4500 ]
> > +            description:
> > +              Defines the maximum current delivered by the regulator in microamp.
>
> Instead of implementing a custom property this should use the already
> existing properties for current limits, this is a common thing for
> hardware to have so we shouldn't have custom bindings.  We'll need to
> relax the check the code currently has for a non-zero minimum limit but
> otherwise everything should already be there.

Ok I now use "regulator-max-microamp" property from the regulator that
acts like my property.
I was wrong with the default value. I re-read the documentation and
the default value is stored in OTP
memory. So if someone skipped this property, it's OK to not send any value.


Thanks again,
Regards,