Hi Rob,

On Mon, May 11, 2020 at 04:47:27PM -0500, Rob Herring wrote:

On Fri, Apr 24, 2020 at 05:33:44PM +0200, Maxime Ripard wrote:

quoted

The firmware running on the RPi VideoCore can be used to discover and
change the various clocks running in the BCM2711. Since devices will
need to use them through the DT, let's add a pretty simple binding.

Cc: Michael Turquette <mturquette@baylibre.com>
Cc: Stephen Boyd <sboyd@kernel.org>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: linux-clk@vger.kernel.org
Cc: devicetree@vger.kernel.org
Signed-off-by: Maxime Ripard <redacted>
---
 Documentation/devicetree/bindings/arm/bcm/raspberrypi,bcm2835-firmware.yaml | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

diff --git a/Documentation/devicetree/bindings/arm/bcm/raspberrypi,bcm2835-firmware.yaml b/Documentation/devicetree/bindings/arm/bcm/raspberrypi,bcm2835-firmware.yaml
index cec540c052b6..b48ed875eb8e 100644
--- a/Documentation/devicetree/bindings/arm/bcm/raspberrypi,bcm2835-firmware.yaml
+++ b/Documentation/devicetree/bindings/arm/bcm/raspberrypi,bcm2835-firmware.yaml

@@ -22,6 +22,25 @@ properties:
       Phandle to the firmware device's Mailbox.
       (See: ../mailbox/mailbox.txt for more information)
 
+  clocks:
+    type: object
+
+    properties:
+      compatible:
+        const: raspberrypi,firmware-clocks
+
+      "#clock-cells":
+        const: 1
+        description: >
+          The argument is the ID of the clocks contained by the
+          firmware messages.

Not really any reason to make this a child node. The parent can be a 
clock provider.

This was meant to provide some consistency for that driver. It already has a nod
for the GPIOs exposed through the firmware in the RPi where you could make the
exact same argument, so I guess that, while we shouldn't make the same choices
if we had to design it from scratch, it's more important to keep the same design
choices for a given binding?

Maxime