Re: [PATCH 8/8] drm/bridge: it66121: Allow link this driver as a lib

[Sui Jingfeng]

Hi,


On 2023/11/23 16:08, Dmitry Baryshkov wrote:
> > I'm agree with the idea that drm bridges drivers involved toward to a direction
> > that support more complex design, but I think we should also leave a way for the
> > most frequent use case. Make it straight-forward as a canonical design.
> Not having anything connector-related in the drm_bridge driver is a
> canonical design.

What you said is just for the more complex uses case. I can't agree, sorry.

By choosing the word "canonical design", I means that the most frequently used
cases in practice are the canonical design, 95+% motherboards I have seen has
only one *onboard* display bridges chip. For my driver, I abstract the internal
(inside of the chip) encoder as drm_encoder and abstract the external TX chip as
drm_bridge, this design still works very well.


Originally, I means that this is a concept of the hardware design.
You are wrong even through in the software design context, the
transparent simple drm bridge drivers(simple-bridge.c) also *allow*
to create drm connector manually. I don't think I need to emulate
more example, please read the code by youself.

Canonical or not canonical is not a question to argue, if other programmers
are allowed to do such kind of abstraction, I should also allowed too. Thanks.