[PATCH 2/3] dt: tegra20: Add ehci host controller nodes.

[Stephen Warren]

Stephen Warren wrote at Tuesday, July 19, 2011 5:00 PM:
> achew at nvidia.com wrote at Tuesday, July 19, 2011 4:47 PM:
> > From: Andrew Chew <achew at nvidia.com>
> >
> > These values were derived from various headers in arch/arm/mach-tegra.
...
> > diff --git a/arch/arm/boot/dts/tegra20.dtsi b/arch/arm/boot/dts/tegra20.dtsi
> > index 83fedf3..a0353bb 100644
> > --- a/arch/arm/boot/dts/tegra20.dtsi
> > +++ b/arch/arm/boot/dts/tegra20.dtsi
> > @@ -160,5 +160,57 @@
> >  		interrupts = < 63 >;
> >  		status = "disabled";
> >  	};
> > +
> > +	ehci at c5000000 {
> > +		#address-cells = <1>;
> > +		#size-cells = <0>;
> > +		compatible = "nvidia,tegra20-ehci";
> > +		reg = <0xc5000000 0x4000>;
> > +		interrupts = < 52 >;
> > +		mode = < 1 >;
> > +		power_down_on_bus_suspend;
> > +		type = "utmi";
> > +		hssync_start_delay = < 0 >;
> > +		idle_wait_delay = < 17 >;
> > +		elastic_limit = < 16 >;
> > +		term_range_adj = < 6 >;
> > +		xcvr_setup = < 15 >;
> > +		xcvr_lsfslew = < 2 >;
> > +		xcvr_lsrslew = < 2 >;
> > +		status = "disabled";
> 
> In the latest devicetree/test, tegra20.dtsi doesn't specify status for
> any nodes, but leaves everything enabled. Instead, board-specific files
> disable any unused devices.

Andrew pointed out that isn't actually true right now. However, Grant
did state that he was *going* to make Tegra work that way, so the
comment basically still stands.

-- 
nvpublic