[PATCH 1/2] dt-bindings: memory-controllers: samsung,exynos4210-srom: Enforce child props

[PATCH 1/2] dt-bindings: memory-controllers: samsung,exynos4210-srom: Enforce child props

[Krzysztof Kozlowski]

Samsung Exynos SROM peripheral properties were moved from the device
schema to separate "peripheral-props" schema for child node, but the
device schema does not reference the new one.

Reference the peripheral-props schema so the child nodes will be
properly validated from the device schema.

Fixes: 67bf606fcf18 ("dt-bindings: memory-controllers: samsung,exynos4210-srom: Split out child node properties")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>

---

Fix for Rob's tree.
---
 .../devicetree/bindings/memory-controllers/exynos-srom.yaml      | 1 +
 1 file changed, 1 insertion(+)

diff --git a/Documentation/devicetree/bindings/memory-controllers/exynos-srom.yaml b/Documentation/devicetree/bindings/memory-controllers/exynos-srom.yaml
index 2267c5107d60..1578514ec58d 100644
--- a/Documentation/devicetree/bindings/memory-controllers/exynos-srom.yaml
+++ b/Documentation/devicetree/bindings/memory-controllers/exynos-srom.yaml
@@ -38,6 +38,7 @@ properties:
 patternProperties:
   "^.*@[0-3],[a-f0-9]+$":
     type: object
+    $ref: mc-peripheral-props.yaml#
     additionalProperties: true
 
     properties:
-- 
2.43.0

[PATCH 2/2] dt-bindings: memory-controllers: qcom,ebi2: Enforce child props

[Krzysztof Kozlowski]

Qualcomm EBI2 peripheral properties were moved from the device schema to
separate "peripheral-props" schema for child node, but the device schema
does not reference the new one.

Reference the peripheral-props schema so the child nodes will be
properly validated from the device schema.

Fixes: 06652f348f28 ("dt-bindings: memory-controllers: qcom,ebi2: Split out child node properties")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>

---

Fix for Rob's tree.
---
 .../devicetree/bindings/memory-controllers/qcom,ebi2.yaml        | 1 +
 1 file changed, 1 insertion(+)

diff --git a/Documentation/devicetree/bindings/memory-controllers/qcom,ebi2.yaml b/Documentation/devicetree/bindings/memory-controllers/qcom,ebi2.yaml
index 3e6da1ba460e..423d7a75134f 100644
--- a/Documentation/devicetree/bindings/memory-controllers/qcom,ebi2.yaml
+++ b/Documentation/devicetree/bindings/memory-controllers/qcom,ebi2.yaml
@@ -104,6 +104,7 @@ required:
 patternProperties:
   "^.*@[0-5],[0-9a-f]+$":
     type: object
+    $ref: mc-peripheral-props.yaml#
     additionalProperties: true
 
 additionalProperties: false
-- 
2.43.0

Re: [PATCH 1/2] dt-bindings: memory-controllers: samsung,exynos4210-srom: Enforce child props

[Rob Herring]

On Thu, Mar 6, 2025 at 2:58 AM Krzysztof Kozlowski
<krzysztof.kozlowski@linaro.org> wrote:
>
> Samsung Exynos SROM peripheral properties were moved from the device
> schema to separate "peripheral-props" schema for child node, but the
> device schema does not reference the new one.
>
> Reference the peripheral-props schema so the child nodes will be
> properly validated from the device schema.
>
> Fixes: 67bf606fcf18 ("dt-bindings: memory-controllers: samsung,exynos4210-srom: Split out child node properties")
> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
>
> ---
>
> Fix for Rob's tree.
> ---
>  .../devicetree/bindings/memory-controllers/exynos-srom.yaml      | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/Documentation/devicetree/bindings/memory-controllers/exynos-srom.yaml b/Documentation/devicetree/bindings/memory-controllers/exynos-srom.yaml
> index 2267c5107d60..1578514ec58d 100644
> --- a/Documentation/devicetree/bindings/memory-controllers/exynos-srom.yaml
> +++ b/Documentation/devicetree/bindings/memory-controllers/exynos-srom.yaml
> @@ -38,6 +38,7 @@ properties:
>  patternProperties:
>    "^.*@[0-3],[a-f0-9]+$":
>      type: object
> +    $ref: mc-peripheral-props.yaml#

Why do we need this? Any child node should have a schema which needs
to include mc-peripheral-props.yaml if any properties from it can be
used.

Rob

Re: [PATCH 1/2] dt-bindings: memory-controllers: samsung,exynos4210-srom: Enforce child props

[Krzysztof Kozlowski]

On 06/03/2025 14:39, Rob Herring wrote:
> Why do we need this? Any child node should have a schema which needs
> to include mc-peripheral-props.yaml if any properties from it can be
> used.

From correctness point of view, you are right, we don't need it. However:

1. Convention was so far to have in every controller. I think this also
is easier to understand whenever one reads the bindings - clear
documentation what children on this bus should look like.

2. To clearly document from where samsung,srom-timing comes in the
required block:

https://web.git.kernel.org/pub/scm/linux/kernel/git/robh/linux.git/tree/Documentation/devicetree/bindings/memory-controllers/exynos-srom.yaml?h=dt/next#n50
Otherwise for me it is a bit confusing to require a property which is
nowhere here defined.


Best regards,
Krzysztof

Re: [PATCH 1/2] dt-bindings: memory-controllers: samsung,exynos4210-srom: Enforce child props

[Rob Herring]

On Thu, Mar 6, 2025 at 7:51 AM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>
> On 06/03/2025 14:39, Rob Herring wrote:
> > Why do we need this? Any child node should have a schema which needs
> > to include mc-peripheral-props.yaml if any properties from it can be
> > used.
>
> From correctness point of view, you are right, we don't need it. However:
>
> 1. Convention was so far to have in every controller. I think this also
> is easier to understand whenever one reads the bindings - clear
> documentation what children on this bus should look like.
>
> 2. To clearly document from where samsung,srom-timing comes in the
> required block:

I could be wrong, but I want to say that's pretty much an exception.
They are usually optional (other than reg). Though I guess we want to
enforce 'reg'...

> https://web.git.kernel.org/pub/scm/linux/kernel/git/robh/linux.git/tree/Documentation/devicetree/bindings/memory-controllers/exynos-srom.yaml?h=dt/next#n50
> Otherwise for me it is a bit confusing to require a property which is
> nowhere here defined.

It's a bit weird that non-exynos properties are allowed too, though
that's always the case for these properties.

I'll apply these.

Rob