On Mon, Aug 14, 2023 at 01:06:19PM +0300, Arınç ÜNAL wrote:
> On 13.08.2023 22:02, Vladimir Oltean wrote:
> > On Sun, Aug 13, 2023 at 05:58:57PM +0300, Arınç ÜNAL wrote:
> > > On top of this, I'd argue to document the internal MDIO bus on the
> > > ethernet-switch.yaml schema instead.
> >
> > Why?
>
> Because a generic switch can have an internal MDIO bus, it's not specific
> to a DSA controlled switch.
>
> Arınç
I'm not sure it's that simple. Check out arch/mips/boot/dts/mscc/ocelot.dtsi.
Its switch IP ("mscc,vsc7514-switch") is on the same hierarchical level
with the "mscc,ocelot-miim" nodes (so, not a child), because the MDIO controller
isn't part of the address space of the switching IP. Actually that could equally
be considered true for DSA. Placing the "mdio" node as a child of the switch is
one of the possible options, but it has its limitations and is certainly
not the only one.
