On Wed, Aug 14, 2024 at 06:49:39AM +0200, Jan Kiszka wrote:

On 13.08.24 17:40, Conor Dooley wrote:

quoted

On Tue, Aug 13, 2024 at 08:04:59AM +0200, Jan Kiszka wrote:

quoted

From: Jan Kiszka <jan.kiszka@siemens.com>

Expand bindings to cover both the MCU and the main usage of the AM654
system controller.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
---
 .../soc/ti/ti,am654-system-controller.yaml    | 25 +++++++++++++++++++
 1 file changed, 25 insertions(+)

diff --git a/Documentation/devicetree/bindings/soc/ti/ti,am654-system-controller.yaml b/Documentation/devicetree/bindings/soc/ti/ti,am654-system-controller.yaml
index e79803e586ca..5a689ec3c5c9 100644
--- a/Documentation/devicetree/bindings/soc/ti/ti,am654-system-controller.yaml
+++ b/Documentation/devicetree/bindings/soc/ti/ti,am654-system-controller.yaml

@@ -29,11 +29,36 @@ properties:
 
   ranges: true
 
+  mux-controller:
+    type: object
+    ref: /schemas/mux/reg-mux.yaml#
+    description:
+      This is the SERDES lane control mux.
+
 patternProperties:
   "^phy@[0-9a-f]+$":
     type: object
     $ref: /schemas/phy/ti,phy-gmii-sel.yaml#
 
+  "^clock@[0-9a-f]+$":

Could you explain to me why these are all patternProperties? Why are the
addresses of these things not fixed for an am654?

I could indeed spell out dss-oldi-io-ctrl@41e0 and
clock-controller@4140, and their addresses are likely fixed, indeed. But
there are also clock@4080 and clock@4090 - should I duplicate their
object descriptions while moving them to the regular properties?

If you're going to itemise the 3 clocks, I think you should mention
what's different about each of them. Otherwise, if they're all
identical, leave them as a patternProperty.