Hi Krzysztof,
Thanks for the review.
On Fri, 5 Apr 2024 at 08:49, Krzysztof Kozlowski <krzk@kernel.org> wrote:
>
> On 04/04/2024 14:25, Peter Griffin wrote:
> > Add dedicated google,gs101-ufs compatible for Google Tensor gs101
> > SoC.
> >
> > Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
> > ---
> > .../bindings/ufs/samsung,exynos-ufs.yaml | 51 +++++++++++++++----
> > 1 file changed, 42 insertions(+), 9 deletions(-)
> >
> > diff --git a/Documentation/devicetree/bindings/ufs/samsung,exynos-ufs.yaml b/Documentation/devicetree/bindings/ufs/samsung,exynos-ufs.yaml
> > index b2b509b3944d..898da6c0e94f 100644
> > --- a/Documentation/devicetree/bindings/ufs/samsung,exynos-ufs.yaml
> > +++ b/Documentation/devicetree/bindings/ufs/samsung,exynos-ufs.yaml
> > @@ -12,12 +12,10 @@ maintainers:
> > description: |
> > Each Samsung UFS host controller instance should have its own node.
> >
> > -allOf:
> > - - $ref: ufs-common.yaml
> > -
> > properties:
> > compatible:
> > enum:
> > + - google,gs101-ufs
> > - samsung,exynos7-ufs
> > - samsung,exynosautov9-ufs
> > - samsung,exynosautov9-ufs-vh
> > @@ -38,14 +36,12 @@ properties:
> > - const: ufsp
> >
> > clocks:
> > - items:
> > - - description: ufs link core clock
> > - - description: unipro main clock
> > + minItems: 2
> > + maxItems: 5
>
> Keep here minItems and:
>
> + - description: ufs link core clock
> + - description: unipro main clock
> + - description: fmp clock
> + - description: ufs aclk clock
> + - description: ufs pclk clock
>
> >
> > clock-names:
> > - items:
> > - - const: core_clk
> > - - const: sclk_unipro_main
> > + minItems: 2
> > + maxItems: 5
>
> Similarly here
>
> >
> > phys:
> > maxItems: 1
> > @@ -72,6 +68,43 @@ required:
> > - clocks
> > - clock-names
> >
> > +allOf:
> > + - $ref: ufs-common.yaml
> > + - if:
> > + properties:
> > + compatible:
> > + contains:
> > + const: google,gs101-ufs
> > +
> > + then:
> > + properties:
> > + clocks:
>
> Enough is:
> minItems: 5
>
> > + items:
>
> and drop the items since they are defined in top-level.
>
> Your original code is correct, but with my approach we keep the list
> synced between variants, at least part of the list. If another variant
> appears, then maybe it will go back to your approach, but maybe we can
> still have the same clocks and their order.
Will update like you suggest in v2.
Thanks,
Peter