Re: [PATCH 1/2] dt-bindings: iio: pressure: honeywell,mprls0025pa

[Conor Dooley]

On Thu, Dec 21, 2023 at 11:04:17AM +0000, Jonathan Cameron wrote:
> On Wed, 20 Dec 2023 19:25:25 +0200
> Petre Rodan <petre.rodan@xxxxxxxxxxxxxxx> wrote:
> 
> > hi Jonathan,
> > 
> > On Wed, Dec 20, 2023 at 03:16:45PM +0000, Jonathan Cameron wrote:
> > > On Tue, 19 Dec 2023 15:02:20 +0200
> > > Petre Rodan <petre.rodan@xxxxxxxxxxxxxxx> wrote:  
> > > >    honeywell,pmin-pascal:
> > > >      description:
> > > >        Minimum pressure value the sensor can measure in pascal.
> > > > +      To be specified only if honeywell,pressure-triplet is set to "NA".  
> > > That just added a backwards compatibility break.  It would be fine
> > > if there was a default: NA for honeywell,pressure-triplet or a check that either
> > > one or the other was supplied (which I'd prefer).  Thus old bindings will work
> > > and new ones also supported.  
> > 
> > ok, I see your reasoning. but in this second scenario that you prefer how can we
> > propery define the 'required:' block? an equivalent to
> > 
> > required:
> >   - compatible
> >   - reg
> >   - (honeywell,pmin-pascal && honeywell,pmax-pascal) || honeywell,pressure-triplet
> >   - honeywell,transfer-function
> 
> Yes, it would end up something like that.  There are exclusive or examples in tree.
> I think something like dac/adi,ad3552r.yaml
>  should work.
> 
> oneOf:
>   - required:
>       - honeywell,pmin-pascal
>       - honeywell,pmax-pascal
>   - required:
>       - honeywell,pressure-triplet
> 
> but you will want to try all the cases to make sure that works (my ability to
> figure these ones out is tricky).
> 
> + you ideally want to exclude them all being set which is fiddlier.
> 
> Some similar examples but they are based on a value in the property. I'm not
> sure how you check for it just being defined.
> 
> Something along lines of.
> 
> allOf:
>   - if:
>       properties:
>         honeywell,pressure-triplet
>     then:
>       properties:
>         honeywell,pmin-pascal: false
>         honeywell,pmax-pascal: false
> 
> Might work?  I always end up trawling the kernel to find a similar example for cases but
> can't find anything closer right now.

I hate to admit it, but I'm not great at expressing these in the minimum
forms either, but I think you're missing a "required" from here, in place
of the "properties":
allOf:
  - if:
      required:
        - honeywell,pressure-triplet
    then:
      properties:
        honeywell,pmin-pascal: false
        honeywell,pmax-pascal: false

Cheers,
Conor.

Attachment: signature.asc
Description: PGP signature