Re: [PATCH 1/4] dt-bindings: clock: ti: Convert to yaml

[Sukrut Bellary <sbellary@baylibre.com>]

On Sat, Apr 05, 2025 at 09:55:29PM +0200, Andreas Kemnade wrote:
> Am Fri, 4 Apr 2025 12:44:39 +0200
> schrieb Krzysztof Kozlowski <krzk@kernel.org>:
> 
> > On Thu, Apr 03, 2025 at 06:44:57PM GMT, Sukrut Bellary wrote:
> > > +properties:
> > > +  reg:
> > > +    maxItems: 1  
> > 
> > How reg is part of this? Every clock has reg, doesn't it? Otherwise how
> > do you control it? Drop.
> > 
> > > +
> > > +  ti,autoidle-shift:
> > > +    $ref: /schemas/types.yaml#/definitions/uint32
> > > +    description:
> > > +      bit shift of the autoidle enable bit for the clock
> > > +    maximum: 31
> > > +    default: 0
> > > +
> > > +  ti,invert-autoidle-bit:
> > > +    type: boolean
> > > +    description:
> > > +      autoidle is enabled by setting the bit to 0  
> > 
> > required:
> >   - ti,autoidle-shift
> >   - ti,invert-autoidle-bit - although this makes no sense, so probably
> > old binding was not correct here
> > 
> 
> well, the more informal definition in the txt file can be read as: if
> the clock supports autoidle, then ti,autoidle-shift is required. But
> that does not
> translate to the formal definition in the yaml file.
> So we have nothing required here.
> 
> I am a bit wondering whether we should just drop the autoidle.txt. The
> only thing worth there is the description.

Thanks for the review.
IMO, it would be good to keep this as other clocks refer to these
properties.
But yes, I will add details in the description about if the clock supports
autoidle, then these properties are applicable.

> Regards,
> Andreas