Re: [net-next RFC PATCH v2 3/3] dt-bindings: net: Document support for Aeonsemi PHYs

From: Christian Marangi
Date: Wed Mar 26 2025 - 11:23:44 EST


On Wed, Mar 26, 2025 at 04:08:31PM +0100, Andrew Lunn wrote:
> > + A PHY with not firmware loaded will be exposed on the MDIO bus with ID
> > + 0x7500 0x7500 or 0x7500 0x9410 on C45 registers.
>
> > +select:
> > + properties:
> > + compatible:
> > + contains:
> > + enum:
> > + - ethernet-phy-id7500.9410
> > + - ethernet-phy-id7500.9402
> > + - ethernet-phy-id7500.9412
> > + - ethernet-phy-id7500.9422
> > + - ethernet-phy-id7500.9432
> > + - ethernet-phy-id7500.9442
> > + - ethernet-phy-id7500.9452
> > + - ethernet-phy-id7500.9462
> > + - ethernet-phy-id7500.9472
> > + - ethernet-phy-id7500.9482
> > + - ethernet-phy-id7500.9492
>
> > + ethernet-phy@1f {
> > + compatible = "ethernet-phy-id7500.9410",
> > + "ethernet-phy-ieee802.3-c45";
>
> You need to be careful here. And fully understand what this means. In
> general, you don't list a compatible here, or only
> ethernet-phy-ieee802.3-c45. This is because the bus can be enumerated,
> you can get the ID from the device. What is in the device is more
> likely to be correct than whatever the DT author put here. However,
> you can state a compatible with an ID, and when you do that, it means
> the PHY device ID in the silicon is broken, ignore it, probe based on
> the value here. So if you state ethernet-phy-id7500.9410, it does not
> matter if there is firmware or not in the PHY, what ID the PHY has, it
> will get probed as a ethernet-phy-id7500.9410.
>
> Except, if there is a .match_phy_device in the driver ops. If there is
> a .match_phy_device the driver does whatever it wants to try to
> identify the device and perform a match.
>

Yep I will note this for the PHY driver. I really need to use
match_phy_device for the FW load OPs to prevent any kind of bad
compatible.

In C22 75007500 is reported while in C45 a more correct 75009410 is
reported, this is why the c45 compatible.

Aside from this, the compatible listed here are really just to document
the need for firmware-name and to what PHY it should be needed. It's a
pattern I followed from the aquantia schema.

Example for PHY with ID 7500.9410 in C45, firmware-name is required, for
anything else (example 7500.9422) firmware-name property should not be
used (case with a SPI attached for example).

--
Ansuel