Re: [EXT] Re: [PATCH v5 4/8] dt-bindings: usb: ci-hdrc-usb2: add restrictions for reg, interrupts, clock and clock-names properties

On Fri, Feb 02, 2024 at 12:04:51PM +0100, Krzysztof Kozlowski wrote:

On 02/02/2024 10:10, Xu Yang wrote:

quoted

Hi Krzysztof,

quoted

On 31/01/2024 12:43, Xu Yang wrote:

quoted

Change reg, interrupts, clock and clock-names as common properties and add
restrictions on them for different compatibles.

Signed-off-by: Xu Yang <xu.yang_2@nxp.com>

---
Changes in v4:
 - new patch since v3's discussion
 - split the reg, interrupts, clock and clock-names properties into
   common part and device-specific
Changes in v5:
 - keep common property unchanged
 - make if-then more readable
 - remove non imx part
---
 .../devicetree/bindings/usb/ci-hdrc-usb2.yaml | 118 ++++++++++++++++++
 1 file changed, 118 insertions(+)

diff --git a/Documentation/devicetree/bindings/usb/ci-hdrc-usb2.yaml b/Documentation/devicetree/bindings/usb/ci-

hdrc-usb2.yaml

quoted

index 3b56e0edb1c6..6ad3582051b8 100644

--- a/Documentation/devicetree/bindings/usb/ci-hdrc-usb2.yaml
+++ b/Documentation/devicetree/bindings/usb/ci-hdrc-usb2.yaml

@@ -412,6 +412,124 @@ allOf:
         samsung,picophy-pre-emp-curr-control: false
         samsung,picophy-dc-vol-level-adjust: false

+  - if:
+      properties:
+        compatible:
+          const: fsl,imx27-usb
+    then:
+      properties:
+        clocks:
+          minItems: 3
+          maxItems: 3
+        clock-names:
+          minItems: 3
+          maxItems: 3
+          items:
+            anyOf:
+              - const: ipg
+              - const: ahb
+              - const: per

This would be just: enum: [ipg, ahb, per], but in both cases I question
why the order should be flexible? Nothing in commit msg explains it.

The driver will get the clock by clock-name, then the order should not
matter. However, these three clock-names should be present at the same
time. I should use enum then.

quoted

Plus I will repeat myself from your v4. I don't think this is helping,
because the file will soon grow to umnanageable chunk. I prefer to fix
it at beginning, before we reach snps-schema level of complexities.

Please define common schema, reference in this file and move IMX to own
file.

I'm not that familiar with dt-bindings architecture. If I define a common
schema, then should I create imx, qcom, nvidia and other dt-binding files
too? 

No, the rest you can leave here. Someone, maybe me, will move them some
time. The point is to move at least IMX to its own file.

That will only work as long as i.MX doesn't define any extra properties 
which it already has. We have to have at a minium common, imx, and 
everything else schema docs.

Rob

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel