Re: [PATCH 2/3] dt-bindings: mfd: twl: add charger node also for TWL603x

[Andreas Kemnade <andreas@kemnade.info>]

Am Fri, 20 Sep 2024 19:51:25 -0500
schrieb Rob Herring <robh@kernel.org>:

> On Wed, Sep 18, 2024 at 12:47:22PM +0200, Krzysztof Kozlowski wrote:
> > On 18/09/2024 10:41, Andreas Kemnade wrote:  
> > > Also the TWL603X devices have a charger, so allow to specify it
> > > here.
> > > 
> > > Signed-off-by: Andreas Kemnade <andreas@kemnade.info>
> > > ---
> > >  .../devicetree/bindings/mfd/ti,twl.yaml        | 18
> > > ++++++++++++++++++ 1 file changed, 18 insertions(+)
> > > 
> > > diff --git a/Documentation/devicetree/bindings/mfd/ti,twl.yaml
> > > b/Documentation/devicetree/bindings/mfd/ti,twl.yaml index
> > > e94b0fd7af0f8..4064a228cb0fc 100644 ---
> > > a/Documentation/devicetree/bindings/mfd/ti,twl.yaml +++
> > > b/Documentation/devicetree/bindings/mfd/ti,twl.yaml @@ -105,6
> > > +105,11 @@ allOf: regulator-initial-mode: false
> > >  
> > >        properties:
> > > +        bci:  
> > 
> > charger
> >   
> > > +          type: object  
> > 
> > additionalProperties: true
> > 
Thinking again. Why additionalProperties? unevaluatedProperties
looks more reasonable for me. There are additional properties but
they should be evaluated by another schema.

> > Each node must end with additionalProperties or unevaluated. I
> > think you never tested it, because dtschema reports this.  
> 
> This is under an if/then schema is why there's no errors.
> 
and then it just accepts anything with compatible twl6032-charger e.g.
and does not care about anything in patch 2, because it has a different
compatible.

> This schema probably should have been 3 with a ti,twl-common.yaml
> schema for the common properties, but I'm not sure it is worth
> changing now.
> 
Or a ti,twl4030.yaml and a ti,twl603X.yaml. 6030 and 6032 have more in
common than the 4030.
I would propose that is something for the next more final cleaning
up/conversion round. First I would like to avoid having drained
batteries because of no charging, so allow for more automated testing
and bisecting. 
I think I will prepare a v2 series on monday.

Regards,
Andreas