Re: [PATCH v1 1/3] dt-bindings: timer: Add timer for StarFive JH7110 SoC

[Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>]

On 10/01/2023 03:14, Xingyu Wu wrote:
> On 2022/12/23 18:25, Krzysztof Kozlowski wrote:
>> On 23/12/2022 10:47, Xingyu Wu wrote:
>>> Add bindings for the timer on the JH7110
>>> RISC-V SoC by StarFive Technology Ltd.
>>
>> Please wrap commit message according to Linux coding style / submission
>> process (neither too early nor over the limit):
>> https://elixir.bootlin.com/linux/v5.18-rc4/source/Documentation/process/submitting-patches.rst#L586
>>
>>
>>>
>>> Signed-off-by: Xingyu Wu <xingyu.wu@starfivetech.com>
>>> ---
>>>  .../timer/starfive,jh7110-timers.yaml         | 105 ++++++++++++++++++
>>>  1 file changed, 105 insertions(+)
>>>  create mode 100644 Documentation/devicetree/bindings/timer/starfive,jh7110-timers.yaml
>>>
>>> diff --git a/Documentation/devicetree/bindings/timer/starfive,jh7110-timers.yaml b/Documentation/devicetree/bindings/timer/starfive,jh7110-timers.yaml
>>> new file mode 100644
>>> index 000000000000..fe58dc056313
>>> --- /dev/null
>>> +++ b/Documentation/devicetree/bindings/timer/starfive,jh7110-timers.yaml
>>> @@ -0,0 +1,105 @@
>>> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
>>> +%YAML 1.2
>>> +---
>>> +$id: http://devicetree.org/schemas/timer/starfive,jh7110-timers.yaml#
>>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>>> +
>>> +title: StarFive Timers
>>
>>
>> Not enough, really not enough. Describe the hardware.
> 
> Will add. Thanks.
> 
>>
>>> +
>>> +maintainers:
>>> +  - Samin Guo <samin.guo@starfivetech.com>
>>> +  - Xingyu Wu <xingyu.wu@starfivetech.com>
>>> +
>>> +properties:
>>> +  compatible:
>>> +    const: starfive,jh7110-timers
>>
>> Why plural "timers", not "timer"? The module is usually called timer -
>> see other hardware that type.
>>
> 
> Will fix. Thanks.
> 
>>> +
>>> +  reg:
>>> +    maxItems: 1
>>> +
>>> +  interrupts:
>>> +    items:
>>> +      - description: timer channel 0 interrupt
>>> +      - description: timer channel 1 interrupt
>>> +      - description: timer channel 2 interrupt
>>> +      - description: timer channel 3 interrupt
>>> +
>>> +  interrupt-names:
>>> +    items:
>>> +      - const: timer0
>>> +      - const: timer1
>>> +      - const: timer2
>>> +      - const: timer3
>>
>> I would just drop the names, not really useful. Unless you plan to add
>> here some generic interrupt (like you did for clock-names)?
> 
> Will drop. Thanks.
> 
>>
>>> +
>>> +  clocks:
>>> +    items:
>>> +      - description: timer channel 0 clock
>>> +      - description: timer channel 1 clock
>>> +      - description: timer channel 2 clock
>>> +      - description: timer channel 3 clock
>>> +      - description: APB clock
>>> +
>>> +  clock-names:
>>> +    items:
>>> +      - const: timer0
>>> +      - const: timer1
>>> +      - const: timer2
>>> +      - const: timer3
>>> +      - const: apb
>>> +
>>> +  resets:
>>> +    items:
>>> +      - description: timer channel 0 reset
>>> +      - description: timer channel 1 reset
>>> +      - description: timer channel 2 reset
>>> +      - description: timer channel 3 reset
>>> +      - description: APB reset
>>> +
>>> +  reset-names:
>>> +    items:
>>> +      - const: timer0
>>> +      - const: timer1
>>> +      - const: timer2
>>> +      - const: timer3
>>> +      - const: apb
>>> +
>>> +  clock-frequency:
>>> +    description: The frequency of the clock that drives the counter, in Hz.
>>
>> Why do you need it? Use common clk framework to get that frequency.
> 
> Because normally this timer driver is loaded earlier than the clock tree driver, it won't get
> that frequency by clk framework and this 'clock-frequency' node is used instead.

I don't think that clk framework or fixed clocks are not available at
this time... of_clk_init is before timer.

> 
>>
>> Also, sort the nodes somehow, e.g.
>> compatible/reg/clocks/clock-frequency/interrupts/resets.
> 
> Will reorder. Thanks.
> 
>>
>>
>>> +
>>> +required:
>>> +  - compatible
>>> +  - reg
>>> +  - interrupts
>>> +  - interrupt-names
>>> +  - clocks
>>> +  - clock-names
>>> +  - resets
>>> +  - reset-names
>>> +  - clock-frequency
>>> +
>>> +unevaluatedProperties: false
>>
>> Did you test the binding?
> 
> Yes, I had tested by 'dt_binding_check'. Do you mean the 'unevaluatedProperties' is wrong
> and use 'additionalProperties'?

Yes, previously it was generating a warning but I do not see Rob's bot
answer so maybe something changed.

Best regards,
Krzysztof