Re: [PATCH v4 1/2] ASoC: dt-bindings: cs35l45: Adds SYNC properties

[Krzysztof Kozlowski <krzk@kernel.org>]

On 27/08/2026 22:07, Ricardo Rivera-Matos wrote:
> 
> 
> On 8/27/26 7:31 AM, Krzysztof Kozlowski wrote:
>> On 22/08/2026 00:41, Ricardo Rivera-Matos wrote:
>>> Introduces support for the multidevice synchronization (SYNC)
>>> bus. The properties introduced allow for the configuration
>>> of the device's transmitter IDs.
>>>
>>> Signed-off-by: Ricardo Rivera-Matos <rriveram@opensource.cirrus.com>
>>> ---
>>>   .../devicetree/bindings/sound/cirrus,cs35l45.yaml  | 27 ++++++++++++++++++++++
>>>   1 file changed, 27 insertions(+)
>>>
>>> diff --git a/Documentation/devicetree/bindings/sound/cirrus,cs35l45.yaml b/Documentation/devicetree/bindings/sound/cirrus,cs35l45.yaml
>>> index 70f6c62aedca..20d758dc442f 100644
>>> --- a/Documentation/devicetree/bindings/sound/cirrus,cs35l45.yaml
>>> +++ b/Documentation/devicetree/bindings/sound/cirrus,cs35l45.yaml
>>> @@ -53,6 +53,29 @@ properties:
>>>       maximum: 3
>>>       default: 2
>>>   
>>> +  sync:
>>
>> Same comment. You did not respond to Sashiko report because "it
>> contradicts itself" is pointless. It did not... unless you meant
>> irrelevant part about example. But that is just irrelevant to the binding.
> 
> I will explain the contradiction again. Pay close attention to the word bus.
> 
> Sashiko says "This commit introduces support for the multidevice 
> synchronization (SYNC) bus".
> 
> Sashiko and I agree, it is a bus.
> 
> Sashiko then complains "Is it necessary to use a child node here just to 
> group configuration properties? In Device Tree design, child nodes 
> usually represent distinct hardware components, sub-devices, buses, or 
> standardized structural elements rather than serving as logical folders 
> to group related configuration properties."
> 
> I am confused, do buses merit a child node or not? This is a 
> contradiction without additional context. Do certain kinds of buses get 
> child nodes and others do not?
> 
> Sashiko is an automated review tool that is being ambiguous. I need a 
> human maintainer decision made here. Does a bus merit a child node, yes 
> or no?
> 

I understand now your judgment.

Skipping then whatever Sashiko invented, I do not see there a bus
represented by that "sync" node. Bus in DT comes with children, thus
sync would have children. Or your DEVICE is a bus, then sync would be a
child of that bus.

But your description clearly says it is not a bus. It is a "bus
configuration". Configuration is not a bus and configuration does not
receive separate device nodes.

Best regards,
Krzysztof