Re: [PATCH 1/2] dt-bindings: net: Add rfkill-gpio binding

On Mi, 2022-12-21 at 08:45 -0600, Rob Herring wrote:

On Wed, Dec 21, 2022 at 11:48:02AM +0100, Philipp Zabel wrote:

quoted

Add a device tree binding document for GPIO controlled rfkill switches.
The name, type, shutdown-gpios and reset-gpios properties are the same
as defined for ACPI.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
---
 .../devicetree/bindings/net/rfkill-gpio.yaml  | 60 +++++++++++++++++++
 1 file changed, 60 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/rfkill-gpio.yaml

diff --git a/Documentation/devicetree/bindings/net/rfkill-gpio.yaml b/Documentation/devicetree/bindings/net/rfkill-gpio.yaml
new file mode 100644
index 000000000000..6e62e6c96456
--- /dev/null
+++ b/Documentation/devicetree/bindings/net/rfkill-gpio.yaml

@@ -0,0 +1,60 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: "http://devicetree.org/schemas/net/rfkill-gpio.yaml#"
+$schema: "http://devicetree.org/meta-schemas/core.yaml#"
+
+title: GPIO controlled rfkill switch
+
+maintainers:
+  - Johannes Berg <johannes@sipsolutions.net>
+  - Philipp Zabel <p.zabel@pengutronix.de>
+
+properties:
+  compatible:
+    const: rfkill-gpio
+
+  name:

Did you test this? Something should complain, but maybe not. The problem 
is 'name' is already a property in the unflattened DT (and old FDT 
formats).

Thank you. Maybe this was hidden by the fact that I set the name
property to the same string as the node's name.

'label' would be appropriate perhaps, but why do we care what the name 
is?

This is meant to be the identifier of the rfkill API object. It is the
content of /sys/class/rfkill/rfkill0/name, and the 'ID' in the rfkill
command line tool, that can be used to select a switch, in case a
device has multiple radios of the same type.

quoted

+    $ref: /schemas/types.yaml#/definitions/string
+    description: rfkill switch name, defaults to node name
+
+  type:

Too generic. Property names should ideally have 1 type globally. I think 
'type' is already in use. 'radio-type' instead?

These values correspond to the 'enum rfkill_type' in Linux UAPI, but I
think in this context 'radio-type' would be better than 'rfkill-type'.

quoted

+    description: rfkill radio type
+    enum:
+      - wlan
+      - bluetooth
+      - ultrawideband
+      - wimax
+      - wwan
+      - gps
+      - fm
+      - nfc
+
+  shutdown-gpios:
+    maxItems: 1
+
+  reset-gpios:
+    maxItems: 1

I'm lost as to why there are 2 GPIOs.

I don't know either.  My assumption is that this is for devices that
are radio silenced by just asserting their reset pin (for example GPS
chips). The driver handles them the same.

I could remove reset-gpios and make shutdown-gpios required.

quoted

+
+required:
+  - compatible
+  - type
+
+oneOf:
+  - required:
+      - shutdown-gpios
+  - required:
+      - reset-gpios

But only 1 can be present? So just define 1 GPIO name.

The intent was that only one of them would be required.

quoted

+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+
+    rfkill-pcie-wlan {

Node names should be generic.

What could be a generic name for this - is "rfkill" acceptable even
though it is a Linux subsystem name? Or would "rf-kill-switch" be
better?

How should they be called if there are multiple of them?

quoted

+        compatible = "rfkill-gpio";
+        name = "rfkill-pcie-wlan";
+        type = "wlan";
+        shutdown-gpios = <&gpio2 25 GPIO_ACTIVE_HIGH>;
+    };
-- 
2.30.2

regards
Philipp