Re: [PATCH v6 1/2] dt-bindings: i3c: Add adi-i3c-master

[Krzysztof Kozlowski <krzk@kernel.org>]

On Sun, Jul 20, 2025 at 06:27:26PM -0500, Rob Herring wrote:
> > +description: |
> > +  FPGA-based I3C controller designed to interface with I3C and I2C peripherals,
> > +  implementing a subset of the I3C-basic specification. The IP core is tested
> > +  on arm, microblaze, and arm64 architectures.
> > +
> > +  https://analogdevicesinc.github.io/hdl/library/i3c_controller
> > +
> > +maintainers:
> > +  - Jorge Marques <jorge.marques@analog.com>
> > +
> > +properties:
> > +  compatible:
> > +    const: adi,i3c-master-v1
> 
> If you want to use version numbers, they need to correlate to something 
> and you need to document what that is. I don't see anything in the above 
> link about a version 1. Kind of feels like you just made it up.

I asked already at v4 to document the naming/versioning, which was a
result of one of previous discussions, in the binding description. :/

> 
> > +
> > +  reg:
> > +    maxItems: 1
> > +
> > +  clocks:
> > +    minItems: 1
> > +    items:
> > +      - description: The AXI interconnect clock, drives the register map.
> > +      - description: The I3C controller clock. AXI clock drives all logic if not provided.
> 
> Is that a description of how the h/w works? The controller clock input 
> can literally be left disconnected? If 1 clock source drives both 
> inputs, then the binding should reflect that.

This was explained in reply, but never made as proper explanation to the binding.

Jorge,
When you answer to a review about uncertain pieces like that, usually
outcome of the discussion must end up also in new patch - either in
commit msg or better in the binding itself. I also asked about this -
documenting the outcode - in v4.

Best regards,
Krzysztof