Re: [PATCH 7/8] dt-bindings: i2c: realtek,rtl9301-i2c: extend for RTL9607C support

[Krzysztof Kozlowski <krzk@kernel.org>]

On 15/03/2026 11:10, Rustam Adilov wrote:
> Hello,
> 
> On 2026-03-15 08:56, Krzysztof Kozlowski wrote:
>> On Sat, Mar 14, 2026 at 01:26:27PM +0500, Rustam Adilov wrote:
>>> Add the "realtek,rtl9607-i2c" compatible for i2c controller on the
>>> RTL9607C SoC series.
>>>
>>> Add a clocks property to the properties since RTL9607C requires it
>>> along with the realtek,scl.
>>>
>>> Signed-off-by: Rustam Adilov <adilov@disroot.org>
>>> ---
>>>  .../bindings/i2c/realtek,rtl9301-i2c.yaml         | 15 +++++++++++++++
>>>  1 file changed, 15 insertions(+)
>>>
>>> diff --git a/Documentation/devicetree/bindings/i2c/realtek,rtl9301-i2c.yaml b/Documentation/devicetree/bindings/i2c/realtek,rtl9301-i2c.yaml
>>> index f9a449fee2b0..5873cfdc5b3e 100644
>>> --- a/Documentation/devicetree/bindings/i2c/realtek,rtl9301-i2c.yaml
>>> +++ b/Documentation/devicetree/bindings/i2c/realtek,rtl9301-i2c.yaml
>>> @@ -15,6 +15,8 @@ description:
>>>    assigned to either I2C controller.
>>>    RTL9310 SoCs have equal capabilities but support 12 common SDA lines which
>>>    can be assigned to either I2C controller.
>>> +  RTL9607C SoCs have equal capabilities but each controller only supports 1
>>> +  SCL/SDA line.
>>>  
>>>  properties:
>>>    compatible:
>>> @@ -34,6 +36,7 @@ properties:
>>>        - enum:
>>>            - realtek,rtl9301-i2c
>>>            - realtek,rtl9310-i2c
>>> +          - realtek,rtl9607-i2c
>>>  
>>>    reg:
>>>      items:
>>> @@ -51,6 +54,9 @@ properties:
>>>        The SCL line number of this I2C controller.
>>>      enum: [ 0, 1 ]
>>>  
>>> +  clocks:
>>> +    maxItems: 1
>>> +
>>>  patternProperties:
>>>    '^i2c@[0-9ab]$':
>>>      $ref: /schemas/i2c/i2c-controller.yaml
>>> @@ -81,6 +87,15 @@ allOf:
>>>      then:
>>>        patternProperties:
>>>          '^i2c@[89ab]$': false
>>
>> Other devices do not have any clock input? Hard to believe, but if that
>> was the case then previous "if:then:" should disallow this. But if you
>> assume they have clock, which is expected, then document this in the
>> commit msg. Now you just silently add clock to each of existing variant
>> without any explanation WHY.
> 
> As far as i can tell, that does seem to be the case that they don't need the
> clocks input. In RTL9300 and RL9310 there is a simple SCL_FREQ bit field that
> sets the frequency from preselected option (100khz, 400kHz, 2.5MHz and 50kHz) and
> it is gotten straight from the clock-frequency of the i2c child nodes.

And from where does this clock gets its frequency? Really SoC IP block
not having clock input would mean it is not part of the SoC, because it
would have completely independent clock domain.

> But in RTL9607C there is CLK_DIV bit field instead and it is set to the calculated
> value of "clk_get_rate(i2c->clk) / i2c->bus_freq - 1" and that is why it requires
> clocks.
> 
> With this, what should be changed in this i2c bindings file? I was under assumption

If devices *do not have* clock, you set it as false (see example schema
and even line above!). Clue here is what I wrote "devices" and "do not
have".

Best regards,
Krzysztof