Re: [PATCH 1/5] dt-bindings: mtd: gpmi-nand: Add 'fsl,imx8qxp-gpmi-nand' compatible string

[Frank Li]

On Fri, May 17, 2024 at 09:50:55PM +0200, Miquel Raynal wrote:
> Hi Frank,
> 
> Frank.li@xxxxxxx wrote on Fri, 17 May 2024 15:15:42 -0400:
> 
> > On Fri, May 17, 2024 at 08:36:21PM +0200, Miquel Raynal wrote:
> > > Hi Frank,
> > > 
> > > Frank.Li@xxxxxxx wrote on Fri, 17 May 2024 14:09:48 -0400:
> > >   
> > > > Add 'fsl,imx8qxp-gpmi-nand' compatible string and clock-names restriction.
> > > > 
> > > > Signed-off-by: Frank Li <Frank.Li@xxxxxxx>
> > > > ---
> > > >  .../devicetree/bindings/mtd/gpmi-nand.yaml         | 22 ++++++++++++++++++++++
> > > >  1 file changed, 22 insertions(+)
> > > > 
> > > > diff --git a/Documentation/devicetree/bindings/mtd/gpmi-nand.yaml b/Documentation/devicetree/bindings/mtd/gpmi-nand.yaml
> > > > index 021c0da0b072f..f9eb1868ca1f4 100644
> > > > --- a/Documentation/devicetree/bindings/mtd/gpmi-nand.yaml
> > > > +++ b/Documentation/devicetree/bindings/mtd/gpmi-nand.yaml
> > > > @@ -24,6 +24,7 @@ properties:
> > > >            - fsl,imx6q-gpmi-nand
> > > >            - fsl,imx6sx-gpmi-nand
> > > >            - fsl,imx7d-gpmi-nand
> > > > +          - fsl,imx8qxp-gpmi-nand
> > > >        - items:
> > > >            - enum:
> > > >                - fsl,imx8mm-gpmi-nand
> > > > @@ -151,6 +152,27 @@ allOf:
> > > >              - const: gpmi_io
> > > >              - const: gpmi_bch_apb
> > > >  
> > > > +  - if:
> > > > +      properties:
> > > > +        compatible:
> > > > +          contains:
> > > > +            enum:
> > > > +              - fsl,imx8qxp-gpmi-nand
> > > > +    then:
> > > > +      properties:
> > > > +        clocks:
> > > > +          items:
> > > > +            - description: SoC gpmi io clock
> > > > +            - description: SoC gpmi apb clock  
> > > 
> > > I believe these two clocks are mandatory?  
> > 
> > minItems default is equal to items numbers, here is 4. So all 4 clock are
> > mandatory.
> > 
> > Anything wrong here?
> 
> I'd say that the two "bch" clocks are only used if you decide to
> configure the on-host hardware ECC engine and thus are not needed with
> software corrections, but I'm fine keeping the fourth described in all
> cases if that's simpler.
> 
> Also,here the diff just shows that "if we provide a clocks property
> with this compatible, then we need to provide 4 members", I believe the
> "required" property is already filled somewhere with the
> clocks/clock-names properties?

yes, before allOf

required:
..                                                                 
  - clocks                                                                 
  - clock-names                                                            
..

> 
> Thanks,
> Miquèl