Re: [PATCH] dt-bindings: power: reset: cortina,gemini-power-controller: convert to DT schema

[Linus Walleij]

On Sat, Mar 28, 2026 at 6:31 PM Krzysztof Kozlowski <krzk@xxxxxxxxxx> wrote:
> On 28/03/2026 17:12, Linus Walleij wrote:
> > On Sat, Mar 28, 2026 at 3:26 PM Rob Herring (Arm) <robh@xxxxxxxxxx> wrote:
> >
> >> dtschema/dtc warnings/errors:
> >> /builds/robherring/dt-review-ci/linux/Documentation/devicetree/bindings/power/reset/cortina,gemini-power-controller.example.dtb: power-controller@4b000000 (cortina,gemini-power-controller): '#power-domain-cells' is a required property
> >>         from schema $id: http://devicetree.org/schemas/power/power-domain.yaml
> >
> > Weird, this power controller does not handle power domains whatsoever,
> > it handles the mains power. So it should not have any power domain
> > cells.
> >
> > Is this the result of some regexp gone stray?
>
> The name "power controller" is used for power domain controller, so
> that's why this name must not be used for other use cases. Usual
> replacement is power-management, reboot, restart or poweroff, depending
> on what is the purpose of this device.

So in this case this is just a conversion of the 9 years old text document
which is an as valid binding as any:

commit ba443b5ab454a9b5f49229a94b2dadf06ac8b79e
Author: Linus Walleij <linusw@xxxxxxxxxx>
Date:   Sun Mar 12 23:36:01 2017 +0100

    power: reset: Add Gemini poweroff DT bindings

    This adds device tree bindings to the power management controller
    in the Gemini SoC.

    Cc: devicetree@xxxxxxxxxxxxxxx
    Cc: Janos Laube <janos.dev@xxxxxxxxx>
    Cc: Paulius Zaleckas <paulius.zaleckas@xxxxxxxxx>
    Cc: Hans Ulli Kroll <ulli.kroll@xxxxxxxxxxxxxx>
    Cc: Florian Fainelli <f.fainelli@xxxxxxxxx>
    Acked-by: Rob Herring <robh@xxxxxxxxxx>
    Signed-off-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
    Signed-off-by: Sebastian Reichel <sre@xxxxxxxxxx>

The text document was conspiciously named "gemini-poweroff.txt" while the
compatible is ""cortina,gemini-power-controller".

I don't know what came first, this binding or the convention of
*-power-controller,
but it's solidly there for a while so we need to accomodate this, I guess worst
case simply special-casing it?

Yours,
Linus Walleij