On Sat, 2019-05-11 at 18:59 +0200, Martin Blumenstingl wrote:
> Hi Jerome,
>
> On Fri, May 10, 2019 at 6:49 PM Jerome Brunet <jbrunet@xxxxxxxxxxxx> wrote:
> > Add the g12a mdio multiplexer which allows to connect to either
> > an external phy through the SoC pins or the internal 10/100 phy
> >
> > Signed-off-by: Jerome Brunet <jbrunet@xxxxxxxxxxxx>
> > ---
> > arch/arm64/boot/dts/amlogic/meson-g12a.dtsi | 32 +++++++++++++++++++++
> > 1 file changed, 32 insertions(+)
> >
> > diff --git a/arch/arm64/boot/dts/amlogic/meson-g12a.dtsi b/arch/arm64/boot/dts/amlogic/meson-g12a.dtsi
> > index fe0f73730525..6e9587aafb5d 100644
> > --- a/arch/arm64/boot/dts/amlogic/meson-g12a.dtsi
> > +++ b/arch/arm64/boot/dts/amlogic/meson-g12a.dtsi
> > @@ -460,6 +460,38 @@
> > assigned-clock-rates = <100000000>;
> > #phy-cells = <1>;
> > };
> > +
> > + eth_phy: mdio-multiplexer@4c000 {
> > + compatible = "amlogic,g12a-mdio-mux";
> > + reg = <0x0 0x4c000 0x0 0xa4>;
> > + clocks = <&clkc CLKID_ETH_PHY>,
> > + <&xtal>,
> > + <&clkc CLKID_MPLL_5OM>;
> I haven't noticed that before but there's a typo in the MPLL_5OM clock
> definition:
> the O (capital o) should be a 0 (zero).
> can you fix this typo in an additional clock patch for v5.2 - then we
> don't have to do it in v5.3 where this .dtsi might already use it
>
> > + clock-names = "pclk", "clkin0", "clkin1";
> > + mdio-parent-bus = <&mdio0>;
> > + #address-cells = <1>;
> > + #size-cells = <0>;
> > +
> > + ext_mdio: mdio@0 {
> > + reg = <0>;
> > + #address-cells = <1>;
> > + #size-cells = <0>;
> > + };
> > +
> > + int_mdio: mdio@1 {
> > + reg = <1>;
> > + #address-cells = <1>;
> > + #size-cells = <0>;
> > +
> > + internal_ephy: ethernet_phy@8 {
> > + compatible = "ethernet-phy-id0180.3301",
> > + "ethernet-phy-ieee802.3-c22";
> please drop the compatible string and replace it with a comment (if
> you feel that it's needed).
> quote from Documentation/devicetree/bindings/net/phy.txt:
> > If the PHY reports an incorrect ID (or none at all) then the
> > "compatible" list may contain an entry with the correct PHY ID in the
> > form: "ethernet-phy-idAAAA.BBBB"
I would keep this for the internal phy. The id completely made up by the MDIO mux, so it
is a lot weaker than the usual PHY.
The fact is that I made a mistake (which I'm going to correct) in the g12a mdio mux and
the id is incorrect. Thanks to this, we know the PHY the correct internal phy driver
is picked up.
>
> I am going to send a patch for other Amlogic boards to remove any
> ethernet-phy-id comaptible string
Make sense for the external phys but I think it is wiser to keep it for the internal ones.
For external phys, A manufacturer could replace the device w/o any notice, so we should
rely on the id to pick it up correctly. For a the PHY embedded in the SoC, we which one it
is and it can't change (w/o notice at least)
>
>
> Regards
> Martin
