>Понедельник, 21 января 2019, 19:25 +03:00 от Rob Herring <robh@xxxxxxxxxx>:
...
>> >On Sat, Dec 22, 2018 at 06:55:20PM +0300, Alexander Shiyan wrote:
>> >> Since the driver switched to using the syscon_regmap_lookup_by_phandle()
>> >> function, we need to specify the phandle for the driver. This patch
>> >> updates the driver binding documentation for these changes.
>> ...
>> >> diff --git a/Documentation/devicetree/bindings/input/clps711x-keypad.txt b/Documentation/devicetree/bindings/input/clps711x-keypad.txt
>> >> index 3eed881..aebedc3 100644
>> >> --- a/Documentation/devicetree/bindings/input/clps711x-keypad.txt
>> >> +++ b/Documentation/devicetree/bindings/input/clps711x-keypad.txt
>> >> @@ -6,6 +6,7 @@ Required Properties:
>> >> - poll-interval: Poll interval time in milliseconds.
>> >> - linux,keymap: The definition can be found at
>> >> bindings/input/matrix-keymap.txt.
>> >> +- syscon: Phandle that points to the CPU syscon1 node.
>> >
>> >What does this provide? If the keypad registers are a sub-block of the
>> >syscon1, then you should make this node a child node.
>>
>> Syscon does not populate child devices.
>
>I don't have any context of what syscon1 looks like. Maybe it needs
>simple-mfd or a driver for the syscon.
The SYSCON1 on this processor is already being processed by the
mfd/syscon driver, so it is represents in DTS, as shown below:
...
syscon1: syscon@80000100 {
compatible = "syscon";
reg = <0x80000100 0x80>;
};
...
keypad: keypad {
compatible = "cirrus,ep7209-keypad";
syscon = <&syscon1>;
status = "disabled";
};
Maybe I need to expand the example section in the DT-bindings so that we can see it?
---
