On Mon, Mar 22, 2021 at 9:38 AM Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote: > > Hi Sagar, Rob, > > (replying to an old email, as this one seems to be the most appropriate) > > On Tue, Sep 22, 2020 at 10:34 PM Rob Herring <robh@xxxxxxxxxx> wrote: > > On Thu, Sep 10, 2020 at 04:14:03PM +0530, Sagar Kadam wrote: > > > Convert device tree bindings for SiFive's PLIC to YAML format > > > > > > Signed-off-by: Sagar Kadam <sagar.kadam@xxxxxxxxxx> > > > > --- /dev/null > > > +++ b/Documentation/devicetree/bindings/interrupt-controller/sifive,plic-1.0.0.yaml > > > > + > > > + '#address-cells': > > > + const: 0 > > > + description: Should be <0> or more. > > > > Drop. 'or more' is wrong. If there's a case with more, it will need to > > be documented. > > Why do we have the "'#address-cells': const: 0" at all... > > > > +required: > > > + - compatible > > > + - '#address-cells' > > ... and why is it required? It is only required if an 'interrupt-map' points to this node. Currently dtc is warning if it is missing always, but there are plans to relax dtc to only warn when 'interrupt-map' is present. Of course, if you had 'interrupt-map' in an overlay, you'd want #address-cells in the base dt and there's no other way to check that than making it required. Rob
