Re: [linux-sunxi] Re: [PATCH 1/3] ARM: dts: sun8i: Add basic dtsi file for Allwinner R40

[icenowy]

å 2017-08-24 14:07ïMaxime Ripard åéï
> On Wed, Aug 23, 2017 at 11:13:04PM +0800, icenowy@xxxxxxx wrote:
> > å 2017-08-23 22:35ïMaxime Ripard åéï
> > > On Wed, Aug 23, 2017 at 07:56:29PM +0800, icenowy@xxxxxxx wrote:
> > > > > > +			reg = <0x01c0f000 0x1000>;
> > > > > > +			clocks = <&ccu CLK_BUS_MMC0>, <&ccu CLK_MMC0>;
> > > > > > +			clock-names = "ahb", "mmc";
> > > > > > +			resets = <&ccu RST_BUS_MMC0>;
> > > > > > +			reset-names = "ahb";
> > > > > > +			pinctrl-0 = <&mmc0_pins>;
> > > > > > +			pinctrl-names = "default";
> > > > > > +			interrupts = <GIC_SPI 32 IRQ_TYPE_LEVEL_HIGH>;
> > > > > > +			max-frequency = <150000000>;
> > > > >
> > > > > have you tested that frequency?
> > > >
> > > > I think the frequency should be kept here, although my cards cannot
> > > > reach this frequency.
> > > >
> > > > The numbers are same as the corresponding controllers in A64.
> > > >
> > > > Maybe I should add a comment saying it's educated guess?
> > >
> > > I'd rather have it tested by someone, and then add the proper
> > > frequencies. It took quite a while to figure out how these modes were
> > > supposed to be working on the A64, so it's not obvious that they're
> > > just going to work.
> >
> > Should I add my results here?
> >
> > MMC0: 25MHz
> > MMC1: 50MHz
> > MMC2: 52MHz
> > MMC3: not wired :-(
> >
> > I think it's conservative enough and works well ;-)
>
> And that's my point. You didn't test HS200, SDR50 or SDR104 for
> example. So you have no idea whether the frequencies from 50MHz to
> 150MHz are actually working or not.

OK.

I will leave a TODO here, although I cannot measure the frequencies
with Banana Pis -- it will requires custom boards.

> > >
> > > > > > +		gic: interrupt-controller@1c81000 {
> > > > > > +			compatible = "arm,gic-400";
> > > > > > +			reg = <0x01c81000 0x1000>,
> > > > > > +			      <0x01c82000 0x1000>,
> > > > > > +			      <0x01c84000 0x2000>,
> > > > > > +			      <0x01c86000 0x2000>;
> > > > > > +			interrupt-controller;
> > > > > > +			#interrupt-cells = <3>;
> > > > > > +			interrupts = <GIC_PPI 9 (GIC_CPU_MASK_SIMPLE(4) |
> > > > > > IRQ_TYPE_LEVEL_HIGH)>;
> > > > > > +		};
> > > > > > +	};
> > > > > > +
> > > > > > +	timer {
> > > > > > +		compatible = "arm,armv7-timer";
> > > > > > +		interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(4) |
> > > > > > IRQ_TYPE_LEVEL_LOW)>,
> > > > > > +			     <GIC_PPI 14 (GIC_CPU_MASK_SIMPLE(4) | IRQ_TYPE_LEVEL_LOW)>,
> > > > > > +			     <GIC_PPI 11 (GIC_CPU_MASK_SIMPLE(4) | IRQ_TYPE_LEVEL_LOW)>,
> > > > > > +			     <GIC_PPI 10 (GIC_CPU_MASK_SIMPLE(4) | IRQ_TYPE_LEVEL_LOW)>;
> > > > >
> > > > > Those masks are wrong.
> > > >
> > > > I compared it with other sun8i SoCs' device tree.
> > > >
> > > > Where's wrong?
> > >
> > > It's supposed to be a mask of the CPUs in your system. Since you just
> > > have one of them, it shouldn't be 4.
> >
> > R40 has 4 cores...
> >
> > Or I didn't understand this?
>
> Gah, sorry, I mistook this for the V3s for some reason...

Thanks.

I think I should send a fix for V3s.

> Maxime
>
> --
> Maxime Ripard, Free Electrons
> Embedded Linux and Kernel engineering
> http://free-electrons.com