Le Fri, 26 Nov 2021 17:50:05 +0000,
Vladimir Oltean [off-list ref] a écrit :

On Fri, Nov 26, 2021 at 06:27:36PM +0100, Clément Léger wrote:

quoted

+  ethernet-ports:
+    type: object
+
+    properties:
+      '#address-cells':
+        const: 1
+      '#size-cells':
+        const: 0
+
+    additionalProperties: false
+
+    patternProperties:
+      "^port@[0-9a-f]+$":
+        type: object
+        description: Ethernet ports handled by the switch
+
+        $ref: ethernet-controller.yaml#
+
+        unevaluatedProperties: false
+
+        properties:
+          reg:
+            description: Switch port number
+
+          phy-handle: true
+
+          phy-mode: true
+
+          fixed-link: true
+
+          mac-address: true
+
+        required:
+          - reg
+
+        oneOf:
+          - required:
+              - phy-handle
+              - phy-mode
+          - required:
+              - fixed-link  

Are you practically saying that a phy-mode would not be required with
fixed-link? Because it still is...

I tried to get it right by looking at a binding you probably know
(dsa.yaml), but none of them are using a oneOf property for these
properties so I tried to guess what was really required or not. I will
add the phy-mode property in the required field since it seems always
needed:

+        required:
+          - reg
+          - phy-mode
+
+        oneOf:
+          - required:
+              - phy-handle
+          - required:
+              - fixed-link  

Does it looks good to you ?

Thanks,

quoted

+
+required:
+  - compatible
+  - reg
+  - reg-names
+  - interrupts
+  - interrupt-names
+  - ethernet-port  

-- 
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com