[PATCH 1/4] dt-bindings: reset: thead,th1520-reset: Add controllers for more subsys
Krzysztof Kozlowski
krzk at kernel.org
Tue Sep 2 08:43:43 PDT 2025
On 02/09/2025 15:57, Philipp Zabel wrote:
> On Di, 2025-09-02 at 15:44 +0200, Krzysztof Kozlowski wrote:
>> On 02/09/2025 11:04, Yao Zi wrote:
>>> On Tue, Sep 02, 2025 at 10:27:53AM +0200, Krzysztof Kozlowski wrote:
>>>> On Mon, Sep 01, 2025 at 04:23:17AM +0000, Yao Zi wrote:
>>>>> +/* VO Subsystem */
>>>>> #define TH1520_RESET_ID_GPU 0
>>>>> #define TH1520_RESET_ID_GPU_CLKGEN 1
>>>>> -#define TH1520_RESET_ID_NPU 2
>>>>> -#define TH1520_RESET_ID_WDT0 3
>>>>> -#define TH1520_RESET_ID_WDT1 4
>>>>
>>>> This is ABI break and deserves explanation and its own patchset.
>>>
>>> The registers in control of TH1520_RESET_ID_{NPU,WDT0,WDT1} don't belong
>>> to the VO reset controller (documented as "thead,th1520-reset"), and
>>> thus cannot be implemented by it. They're in fact AP subsystem resets,
>>> which gets supported in Linux with this series.
>>>
>>> Is it okay for you to separate a patch to delete these wrong IDs and add
>>> them back for the AP reset controller latter? Anyway, I should have
>>> provided more information about these three resets. Thanks for catching
>>> this.
>>
>> So feels like separate patch dropping these resets with above explanation.
>
> They happen to be reintroduced with exactly the same values, just for
> the AP subsystem reset controller:
Yes, I noticed, but that's different reset controller, so previous ABI
for that controller stops working. Anyway my comment still stays.
Best regards,
Krzysztof
More information about the linux-riscv
mailing list