Re: [PATCH V2 RESEND 1/6] dt-bindings: clock: qcom: Add SM8650 video clock controller

[Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>]

Hi Jagadeesh,

On 4/22/24 14:00, Jagadeesh Kona wrote:
> 
> On 4/19/2024 2:31 AM, Vladimir Zapolskiy wrote:
>> Hello Jagadeesh,
>>
>> On 3/25/24 08:07, Jagadeesh Kona wrote:
>>>
>>>
>>> On 3/21/2024 6:42 PM, Dmitry Baryshkov wrote:
>>>> On Thu, 21 Mar 2024 at 11:26, Jagadeesh Kona <quic_jkona@quicinc.com>
>>>> wrote:
>>>>>
>>>>> Extend device tree bindings of SM8450 videocc to add support
>>>>> for SM8650 videocc. While it at, fix the incorrect header
>>>>> include in sm8450 videocc yaml documentation.
>>>>>
>>>>> Signed-off-by: Jagadeesh Kona <quic_jkona@quicinc.com>
>>>>> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
>>>>> ---
>>>>>     .../devicetree/bindings/clock/qcom,sm8450-videocc.yaml    | 4 +++-
>>>>>     include/dt-bindings/clock/qcom,sm8450-videocc.h           | 8
>>>>> +++++++-
>>>>>     2 files changed, 10 insertions(+), 2 deletions(-)
>>>>>
>>>>> diff --git
>>>>> a/Documentation/devicetree/bindings/clock/qcom,sm8450-videocc.yaml
>>>>> b/Documentation/devicetree/bindings/clock/qcom,sm8450-videocc.yaml
>>>>> index bad8f019a8d3..79f55620eb70 100644
>>>>> --- a/Documentation/devicetree/bindings/clock/qcom,sm8450-videocc.yaml
>>>>> +++ b/Documentation/devicetree/bindings/clock/qcom,sm8450-videocc.yaml
>>>>> @@ -8,18 +8,20 @@ title: Qualcomm Video Clock & Reset Controller on
>>>>> SM8450
>>>>>
>>>>>     maintainers:
>>>>>       - Taniya Das <quic_tdas@quicinc.com>
>>>>> +  - Jagadeesh Kona <quic_jkona@quicinc.com>
>>>>>
>>>>>     description: |
>>>>>       Qualcomm video clock control module provides the clocks, resets
>>>>> and power
>>>>>       domains on SM8450.
>>>>>
>>>>> -  See also:: include/dt-bindings/clock/qcom,videocc-sm8450.h
>>>>> +  See also:: include/dt-bindings/clock/qcom,sm8450-videocc.h
>>>>
>>>> This almost pleads to go to a separate patch. Fixes generally should
>>>> be separated from the rest of the changes.
>>>>
>>>
>>> Thanks Dmitry for your review.
>>>
>>> Sure, will separate this into a separate patch in next series.
>>>
>>>>>
>>>>>     properties:
>>>>>       compatible:
>>>>>         enum:
>>>>>           - qcom,sm8450-videocc
>>>>>           - qcom,sm8550-videocc
>>>>> +      - qcom,sm8650-videocc
>>>>>
>>>>>       reg:
>>>>>         maxItems: 1
>>>>> diff --git a/include/dt-bindings/clock/qcom,sm8450-videocc.h
>>>>> b/include/dt-bindings/clock/qcom,sm8450-videocc.h
>>>>> index 9d795adfe4eb..ecfebe52e4bb 100644
>>>>> --- a/include/dt-bindings/clock/qcom,sm8450-videocc.h
>>>>> +++ b/include/dt-bindings/clock/qcom,sm8450-videocc.h
>>>>> @@ -1,6 +1,6 @@
>>>>>     /* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
>>>>>     /*
>>>>> - * Copyright (c) 2023, Qualcomm Innovation Center, Inc. All rights
>>>>> reserved.
>>>>> + * Copyright (c) 2023-2024, Qualcomm Innovation Center, Inc. All
>>>>> rights reserved.
>>>>>      */
>>>>>
>>>>>     #ifndef _DT_BINDINGS_CLK_QCOM_VIDEO_CC_SM8450_H
>>>>> @@ -19,6 +19,11 @@
>>>>>     #define
>>>>> VIDEO_CC_MVS1C_DIV2_DIV_CLK_SRC                                9
>>>>>     #define VIDEO_CC_PLL0                                          10
>>>>>     #define VIDEO_CC_PLL1                                          11
>>>>> +#define
>>>>> VIDEO_CC_MVS0_SHIFT_CLK                                        12
>>>>> +#define VIDEO_CC_MVS0C_SHIFT_CLK                               13
>>>>> +#define
>>>>> VIDEO_CC_MVS1_SHIFT_CLK                                        14
>>>>> +#define VIDEO_CC_MVS1C_SHIFT_CLK                               15
>>>>> +#define VIDEO_CC_XO_CLK_SRC                                    16
>>>>
>>>> Are these values applicable to sm8450?
>>>>
>>>
>>> No, the shift clocks above are part of SM8650 only. To reuse the
>>> existing SM8550 videocc driver for SM8650 and to register these shift
>>> clocks for SM8650, I added them here.
>>>
>>
>> In such case I'd strongly suggest to add a new qcom,sm8650-videocc.h file,
>> and do #include qcom,sm8450-videocc.h in it, thus the new header will be
>> really a short one.
>>
>> This will add pristine clarity.
>>
> 
> Thanks Vladimir for your suggestion. I believe adding a comment for
> these set of clocks should be sufficient to indicate these clocks are
> applicable only for SM8650, I can add the required comment and post the
> next series. Please let me know if this works?

Well, I didn't get any new information to abandon my suggestion, what is
wrong with it or why is it less preferable?

Even if you add a comment in the header file, it means that for SM8450
platforms you'll begin to define inapplicable/unrelated macro for the
platform, which opens a small risk of the misusage, and which can be
easily avoided. I believe that the clarity is better for maintenance.

--
Best wishes,
Vladimir