[PATCH 3/5] arm64: dts: imx: add imx8qxp support

Mark Rutland mark.rutland at arm.com
Sun Apr 29 22:54:39 PDT 2018


Hi,

On Sat, Apr 28, 2018 at 03:06:15AM +0800, Dong Aisheng wrote:
> +/{
> +	cpus {
> +		#address-cells = <2>;
> +		#size-cells = <0>;
> +
> +		/* We have 1 clusters having 4 Cortex-A35 cores */

> +	};
> +
> +	pmu {
> +		compatible = "arm,armv8-pmuv3";
> +		interrupts = <GIC_PPI 7
> +			(GIC_CPU_MASK_SIMPLE(6) | IRQ_TYPE_LEVEL_HIGH)>;
> +		interrupt-affinity = <&A35_0>, <&A35_1>, <&A35_2>, <&A35_3>;
> +	};

If using a PPI and all cores are identical, interrupt-affinity isn't necessary.

I note you're using GICv3, whihch doesn't encode a cpumask in its PPI
specifier. All instances of GIC_CPU_MASK_SIMPLE should go from dts{i,} files
for this platform.

> +
> +	psci {
> +		compatible = "arm,psci-1.0";
> +		method = "smc";
> +		cpu_suspend   = <0xc4000001>;
> +		cpu_off	      = <0xc4000002>;
> +		cpu_on	      = <0xc4000003>;

You can drop the cpu_* function IDs if you only have "arm,psci-1.0" in the
compatible list. These IDs (and more) are implied by PSCI 1.0+.

> +	};
> +};
> diff --git a/arch/arm64/boot/dts/freescale/fsl-imx8qxp-mek.dts b/arch/arm64/boot/dts/freescale/fsl-imx8qxp-mek.dts
> new file mode 100644
> index 0000000..920018e
> --- /dev/null
> +++ b/arch/arm64/boot/dts/freescale/fsl-imx8qxp-mek.dts
> @@ -0,0 +1,138 @@
> +// SPDX-License-Identifier: GPL-2.0+
> +/*
> + * Copyright 2017~2018 NXP
> + */
> +
> +/dts-v1/;
> +
> +#include "fsl-imx8qxp.dtsi"
> +
> +/ {
> +	model = "Freescale i.MX8QXP MEK";
> +	compatible = "fsl,imx8qxp-mek", "fsl,imx8qxp";
> +
> +	chosen {
> +		stdout-path = &dma_lpuart0;
> +	};

No baud-rate configuration?

[...]

> diff --git a/arch/arm64/boot/dts/freescale/fsl-imx8qxp.dtsi b/arch/arm64/boot/dts/freescale/fsl-imx8qxp.dtsi
> new file mode 100644
> index 0000000..d7cc836
> --- /dev/null
> +++ b/arch/arm64/boot/dts/freescale/fsl-imx8qxp.dtsi
> @@ -0,0 +1,794 @@
> +// SPDX-License-Identifier: GPL-2.0+
> +/*
> + * Copyright (C) 2016 Freescale Semiconductor, Inc.
> + * Copyright 2017~2018 NXP
> + *	Dong Aisheng <aisheng.dong at nxp.com>
> + */
> +
> +#include <dt-bindings/clock/imx8qxp-clock.h>
> +#include <dt-bindings/gpio/gpio.h>
> +#include <dt-bindings/pinctrl/pads-imx8qxp.h>
> +#include <dt-bindings/power/imx-rsrc.h>
> +
> +#include "fsl-imx8-ca35.dtsi"
> +
> +/ {
> +	compatible = "fsl,imx8qxp";
> +	interrupt-parent = <&gic>;
> +	#address-cells = <2>;
> +	#size-cells = <2>;
> +
> +	aliases {
> +		serial0 = &dma_lpuart0;
> +		mmc0 = &usdhc1;
> +		mmc1 = &usdhc2;
> +		mmc2 = &usdhc3;
> +	};
> +
> +	memory at 80000000 {
> +		device_type = "memory";
> +		reg = <0x00000000 0x80000000 0 0x40000000>;
> +	};
> +
> +	gic: interrupt-controller at 51a00000 {
> +		compatible = "arm,gic-v3";
> +		reg = <0x0 0x51a00000 0 0x10000>, /* GIC Dist */
> +		      <0x0 0x51b00000 0 0xC0000>; /* GICR (RD_base + SGI_base) */
> +		#interrupt-cells = <3>;
> +		interrupt-controller;
> +		interrupts = <GIC_PPI 9
> +			(GIC_CPU_MASK_SIMPLE(6) | IRQ_TYPE_LEVEL_HIGH)>;

As above, GIC_CPU_MASK_SIMPLE does not apply to GICv3, and should go.

> +	};
> +
> +	timer {
> +		compatible = "arm,armv8-timer";
> +		interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(6) | IRQ_TYPE_LEVEL_LOW)>, /* Physical Secure */
> +			     <GIC_PPI 14 (GIC_CPU_MASK_SIMPLE(6) | IRQ_TYPE_LEVEL_LOW)>, /* Physical Non-Secure */
> +			     <GIC_PPI 11 (GIC_CPU_MASK_SIMPLE(6) | IRQ_TYPE_LEVEL_LOW)>, /* Virtual */
> +			     <GIC_PPI 10 (GIC_CPU_MASK_SIMPLE(6) | IRQ_TYPE_LEVEL_LOW)>; /* Hypervisor */

... likewise.

> +		clock-frequency = <8000000>;

Your firmware should be configuring CNTFRQ, so this clock-frequency property
should go.

Thanks,
Mark.



More information about the linux-arm-kernel mailing list