On Tue, May 02, 2023 at 02:22:22PM +0200, Johan Hovold wrote:
> On Mon, Apr 24, 2023 at 08:40:09PM -0700, Bjorn Andersson wrote:
> > diff --git a/arch/arm64/boot/dts/qcom/sc8280xp-crd.dts b/arch/arm64/boot/dts/qcom/sc8280xp-crd.dts
[..]
> > &mdss0_dp0_out {
> > data-lanes = <0 1>;
> > - remote-endpoint = <&pmic_glink_con0_ss>;
> > + remote-endpoint = <&usb_0_qmpphy_dp_in>;
> > };
>
> It's a bit hard to follow what going on when using place holder nodes
> from the dtsi like this (instead of describing all the ports directly in
> the board dts). IIRC we went a bit back and forth over this earlier and
> we already use this scheme for the display port controllers, so I guess
> this is the price we pay for being consistent.
>
I agree, this is why I argued in favour of keeping the of_graphs
together in a single node. But as long as we label things appropriately
it's pretty ok - and the alternative would be yet another undocumented
"rule".
So let's stick with this...
> > diff --git a/arch/arm64/boot/dts/qcom/sc8280xp.dtsi b/arch/arm64/boot/dts/qcom/sc8280xp.dtsi
[..]
> > + ports {
> > + #address-cells = <1>;
> > + #size-cells = <0>;
> > +
> > + port@0 {
> > + reg = <0>;
> > +
> > + usb_0_qmpphy_out: endpoint {};
> > + };
> > +
> > + port@1 {
> > + reg = <1>;
> > +
> > + usb_0_qmpphy_dp_in: endpoint {};
> > + };
> > + };
> > };
>
> The binding describes three ports, where dp-in is port 2.
>
> Perhaps you don't need to describe ss-in yet, but shouldn't the port
> numbers match? Should some of these be described as required in the
> binding?
>
This should certainly be port@2, thanks for spotting.
Regards,
Bjorn
