On Sat, Jul 03, 2021 at 05:02:45PM +0100, Jonathan Cameron wrote:

On Sat,  3 Jul 2021 10:42:21 +0200
Andreas Kemnade [off-list ref] wrote:

quoted

Add ADC related nodes and properties. This will allow to wire
up ADC channels to consumers, especially to measure input voltages
by the power subdevice.

Signed-off-by: Andreas Kemnade <andreas@kemnade.info>
---
 .../bindings/mfd/ricoh,rn5t618.yaml           | 53 +++++++++++++++++++
 1 file changed, 53 insertions(+)

diff --git a/Documentation/devicetree/bindings/mfd/ricoh,rn5t618.yaml b/Documentation/devicetree/bindings/mfd/ricoh,rn5t618.yaml
index 032a7fb0b4a7..185f87a14a54 100644
--- a/Documentation/devicetree/bindings/mfd/ricoh,rn5t618.yaml
+++ b/Documentation/devicetree/bindings/mfd/ricoh,rn5t618.yaml

@@ -73,6 +73,48 @@ properties:
     description: |
       See Documentation/devicetree/bindings/power/power-controller.txt
 
+  adc:
+    type: object
+
+    properties:
+      compatible:
+        enum:
+          - ricoh,rn5t618-adc
+          - ricoh,rc5t619-adc
+
+      "#io-channel-cells":
+        const: 1
+
+    additionalProperties: false
+
+    required:
+      - compatible
+      - "#io-channel-cells"

Strictly required?  If not used below (where it is optional)
then why do we require the ADC driver to provided the services?

I don't mind you leave it as it is though if you prefer - it doesn't
do any harm!

The device is either a provider or it is not regardless of whether 
there's a consumer, so I prefer this to be required. Also, if a consumer 
is in an overlay, then it is easier if we can rely on #io-channel-cells 
being present already.

Rob