Re: [PATCH] dt-bindings: xlnx,vcu-settings: fix dt_binding_check warnings

[Michael Tretter <m.tretter@pengutronix.de>]

On Thu, 03 Dec 2020 08:49:01 +0100, Michal Simek wrote:
> On 02. 12. 20 10:05, Michael Tretter wrote:
> > When running make dt_binding_check, the xlnx,vcu-settings binding
> > triggers the following two warnings:
> > 
> > 	'additionalProperties' is a required property
> > 
> > 	example-0: vcu@a0041000:reg:0: [0, 2684620800, 0, 4096] is too long
> > 
> > Fix the binding and make the checker happy.
> > 
> > Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
> > ---
> > 
> > Hi,
> > 
> > The xlnx,vcu-settings binding was reviewed [0] before the bot started to
> > run automated tests on the device tree bindings, but now produces some
> > warnings. The original patch that introduces the binding is still in
> > Michal's tree and I am not entirely sure how to handle it, but here is a
> > patch.
> > 
> > Michael
> > 
> > [0] https://lore.kernel.org/linux-arm-kernel/20200429213659.GA9051@bogus/
> > ---
> >  .../bindings/soc/xilinx/xlnx,vcu-settings.yaml    | 15 ++++++++++++---
> >  1 file changed, 12 insertions(+), 3 deletions(-)
> > 
> > diff --git a/Documentation/devicetree/bindings/soc/xilinx/xlnx,vcu-settings.yaml b/Documentation/devicetree/bindings/soc/xilinx/xlnx,vcu-settings.yaml
> > index 378d0ced43c8..cb245f400287 100644
> > --- a/Documentation/devicetree/bindings/soc/xilinx/xlnx,vcu-settings.yaml
> > +++ b/Documentation/devicetree/bindings/soc/xilinx/xlnx,vcu-settings.yaml
> > @@ -26,9 +26,18 @@ required:
> >    - compatible
> >    - reg
> >  
> > +additionalProperties: false
> > +
> >  examples:
> >    - |
> > -    xlnx_vcu: vcu@a0041000 {
> > -          compatible = "xlnx,vcu-settings", "syscon";
> > -          reg = <0x0 0xa0041000 0x0 0x1000>;
> > +    fpga {
> > +        #address-cells = <2>;
> > +        #size-cells = <2>;
> > +
> > +        xlnx_vcu: vcu@a0041000 {
> > +            compatible = "xlnx,vcu-settings", "syscon";
> > +            reg = <0x0 0xa0041000 0x0 0x1000>;
> > +        };
> 
> IIRC we had been discussing this recently and Rob wanted to have just
> 1/1 mapping here.
> 
> Take a look at 0db958b689ca9.

Thanks for the pointer.

Rob: Is there some kind of rule, when to use a 1/1 mapping and when to add a
bus with more cells? I still see several examples that add a bus with 2 cells.
I assume that they more or less legacy, but I didn't find any discussion going
beyond the commit description of 0db958b689ca9, which "just" fixes the
warnings.

I will send a v2, but I'd like to understand the rationale for having the 1/1
mapping first.

Michael