Re: [net-next 2/2] net: dsa: realtek: load switch variants on demand

[Vladimir Oltean <olteanv@gmail.com>]

On Thu, Dec 07, 2023 at 04:50:12PM -0300, Luiz Angelo Daros de Luca wrote:
> HI Vladmir,
> 
> We discussed something about that in the past:
> 
> https://lkml.kernel.org/netdev/20220630200423.tieprdu5fpabflj7@bang-olufsen.dk/T/#m04e6cf7d0c1f18b02c2bf40266ada915b1f02f3d
> 
> The code is able to handle only a single node and binding docs say it
> should be named "mdio". The compatible string wasn't a requirement
> since the beginning and I don't think it is worth it to rename the
> compatible string. I suggest we simply switch to
> of_get_child_by_name() and look for a node named "mdio". If that node
> is not found, we can still look for the old compatible string
> (backwards compatibility) and probably warn the "user" (targeting not
> the end-user but the one creating the DT for a new device).
> 
> I don't know how to handle the binding docs as the compatible string
> is still a requirement for older kernel versions. Is it ok to update
> the device-tree bindings docs in such a way it would break old
> drivers? Or should we keep it there until the last LTS kernel
> requiring it reaches EOL? As device-tree bindings docs should not
> consider how the driver was implemented, I think it would be strange
> to have a note like "required by kernel up to 6.x".
> 
> Regards,
> 
> Luiz

And did you ever answer this question?

"And why do you even need to remove the compatible string from the MDIO
node, can't you just ignore it, does it bother you in any way?"

I'm very confused as to what you're after.