Hi Roger,
> >>>> With 'unevaluatedProperties' support implemented, the TI GPMC example
> >>>> has a warning:
> >>>>
> >>>> Documentation/devicetree/bindings/memory-controllers/ti,gpmc.example.dt.yaml: nand@0,0: Unevaluated properties are not allowed ('rb-gpios' was unexpected)
> >>>>
> >>>> Add the missing definition for 'rb-gpios'.
> >>>
> >>> rb-gpios is already defined in nand-controller.yaml. I seems like the
> >>> real problem is that this file does not refer to it. Can you update the
> >>> fix?
> >>
> >> I don't think we can refer to nand-controller.yaml right now as we are not
> >> fully compatible with it yet. Please see examples below.
> >
> > This is a *very* wrong way of defining a NAND setup. I will take the
> > patch to silence the warning, but please convert this representation to
>
> Thanks!
>
> > the 'new' one. I believe on the driver side it should not be too
> > complicated to support having a few of these properties moved to a NAND
> > chip subnode and still support the below binding. Just be very clear
> > that if the legacy bindings are used, only a single chip is supported.
>
> I agree with you. It has been that way since ages. I will look at cleaning
> it up whenever possible for me. I think we will have to support the old
> bindings as well as the new one going forward as many legacy platforms
> are using it.
Yes of course, this is even a primary requirement.
Thanks,
Miquèl
