Re: [RFC PATCH v1 05/14] dt-bindings: clock: thead,th1520: Add support for Video Output subsystem

[Krzysztof Kozlowski <krzk@kernel.org>]

On 04/12/2024 21:21, Stephen Boyd wrote:
>>> No, that's a clear NAK. You claim you have no address space but in the
>>> same time you have address space via regmap.
>>
>> I see your concern. The VOSYS subsystem's address space includes
>> registers for various components, such as clock gates and reset
>> controls, which are scattered throughout the address space as specified
>> in the manual 4.4.1 [2]. Initially, I attempted to use a shared syscon
>> regmap for access, but I realize this might not be the best approach.
>>
>> To address this, I'll specify the 'reg' property in each node to define
>> the address ranges explicitly fragmenting the address space for the VOSYS
>> manually.
>>
>> vosys_clk: clock-controller@ffef528050 {
>>         compatible = "thead,th1520-clk-vo";
>>         reg = <0xff 0xef528050 0x0 0x8>;
>>         #clock-cells = <1>;
>> };
>>
>> pd: power-domain@ffef528000 {
>>         compatible = "thead,th1520-pd";
>>         reg = <0xff 0xef528000 0x0 0x8>;
>>         #power-domain-cells = <1>;
>> };
> 
> You should have one node:
> 
>     clock-controller@ffef528000 {
>       compatible = "thead,th1520-vo";
>       reg = <0xff 0xef528050 0x0 0x1a04>;

Yes, obviously, but probably entire block:

<0xff 0xef528000 0x0 0x1000>;

>       #clock-cells = <1>;
>       #power-domain-cells = <1>;
>     };


Best regards,
Krzysztof