Re: [PATCH 3/3] arm64: dts: amlogic: a4: add power domain controller node

[neil.armstrong@linaro.org]

On 28/05/2024 11:00, Xianwei Zhao wrote:
> Hi Neil,
>     Thanks for your quickly reply.
> 
> On 2024/5/28 16:46, Neil Armstrong wrote:
>> [ EXTERNAL EMAIL ]
>>
>> On 28/05/2024 10:39, Xianwei Zhao via B4 Relay wrote:
>>> From: Xianwei Zhao <xianwei.zhao@amlogic.com>
>>>
>>> Add power domain controller node for Amlogic A4 SoC
>>>
>>> Signed-off-by: Xianwei Zhao <xianwei.zhao@amlogic.com>
>>> ---
>>>   arch/arm64/boot/dts/amlogic/amlogic-a4-common.dtsi | 4 ++++
>>>   arch/arm64/boot/dts/amlogic/amlogic-a4.dtsi        | 5 +++++
>>>   2 files changed, 9 insertions(+)
>>>
>>> diff --git a/arch/arm64/boot/dts/amlogic/amlogic-a4-common.dtsi b/arch/arm64/boot/dts/amlogic/amlogic-a4-common.dtsi
>>> index b6106ad4a072..eebde77ae5b4 100644
>>> --- a/arch/arm64/boot/dts/amlogic/amlogic-a4-common.dtsi
>>> +++ b/arch/arm64/boot/dts/amlogic/amlogic-a4-common.dtsi
>>> @@ -27,6 +27,10 @@ xtal: xtal-clk {
>>>               #clock-cells = <0>;
>>>       };
>>>
>>> +     sm: secure-monitor {
>>> +             compatible = "amlogic,meson-gxbb-sm";
>>> +     };
>>> +
>>>       soc {
>>>               compatible = "simple-bus";
>>>               #address-cells = <2>;
>>> diff --git a/arch/arm64/boot/dts/amlogic/amlogic-a4.dtsi b/arch/arm64/boot/dts/amlogic/amlogic-a4.dtsi
>>> index 73ca1d7eed81..917c05219b9c 100644
>>> --- a/arch/arm64/boot/dts/amlogic/amlogic-a4.dtsi
>>> +++ b/arch/arm64/boot/dts/amlogic/amlogic-a4.dtsi
>>> @@ -37,4 +37,9 @@ cpu3: cpu@3 {
>>>                       enable-method = "psci";
>>>               };
>>>       };
>>> +
>>> +     pwrc: power-controller {
>>> +             compatible = "amlogic,a4-pwrc";
>>> +             #power-domain-cells = <1>;
>>> +     };
>>
>> pwrc is supposed to be a child of secure-monitor.
>>
> Considered writing it like this when I wrote this.
> 
> Here are two approaches: one is to include secure-monitor in the comm dtsi and fill power-controller by aliases in dtsi of each chip, while the other is to directly include secure-monitor in the dtsi of each chip. Which one do you suggest?

The bindings mandates it to be a child of the secure monitor.

Neil

> 
>> Neil
>>
>>>   };
>>>
>>