[PATCH v7 0/3] arm64: mvebu: Support for Marvell 98DX2530 (and variants)

[PATCH v7 0/3] arm64: mvebu: Support for Marvell 98DX2530 (and variants)

[Chris Packham]

This series adds support for the Marvell 98DX2530 SoC which is the Control and
Management CPU integrated into the AlleyCat5/AlleyCat5X series of Marvell
switches.

The CPU core is an ARM Cortex-A55 with neon, simd and crypto extensions.

This is fairly similar to the Armada-3700 SoC so most of the required
peripherals are already supported. This series adds a devicetree and pinctrl
driver for the SoC and the RD-AC5X-32G16HVG6HLG reference board.

The pinctrl changes from v4 have been picked up and are in linux-next so I
haven't included them in this round. That leaves just the dts files and a minor
Kconfig update for arm64.

Chris Packham (3):
  dt-bindings: marvell: Document the AC5/AC5X compatibles
  arm64: dts: marvell: Add Armada 98DX2530 SoC and RD-AC5X board
  arm64: marvell: enable the 98DX2530 pinctrl driver

 .../bindings/arm/marvell/armada-98dx25xx.yaml |  32 ++
 arch/arm64/Kconfig.platforms                  |   2 +
 arch/arm64/boot/dts/marvell/Makefile          |   1 +
 .../boot/dts/marvell/armada-98dx25xx.dtsi     | 295 ++++++++++++++++++
 .../boot/dts/marvell/armada-98dx35xx-rd.dts   | 101 ++++++
 .../boot/dts/marvell/armada-98dx35xx.dtsi     |  13 +
 6 files changed, 444 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/arm/marvell/armada-98dx25xx.yaml
 create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
 create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx-rd.dts
 create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx.dtsi

-- 
2.36.0

[PATCH v7 1/3] dt-bindings: marvell: Document the AC5/AC5X compatibles

[Chris Packham]

Describe the compatible properties for the Marvell Alleycat5/5X switches
with integrated CPUs.

Alleycat5:
* 98DX2538: 24x1G + 2x10G + 2x10G Stack
* 98DX2535: 24x1G + 4x1G Stack
* 98DX2532: 8x1G + 2x10G + 2x1G Stack
* 98DX2531: 8x1G + 4x1G Stack
* 98DX2528: 24x1G + 2x10G + 2x10G Stack
* 98DX2525: 24x1G + 4x1G Stack
* 98DX2522: 8x1G + 2x10G + 2x1G Stack
* 98DX2521: 8x1G + 4x1G Stack
* 98DX2518: 24x1G + 2x10G + 2x10G Stack
* 98DX2515: 24x1G + 4x1G Stack
* 98DX2512: 8x1G + 2x10G + 2x1G Stack
* 98DX2511: 8x1G + 4x1G Stack

Alleycat5X:
* 98DX3500: 24x1G + 6x25G
* 98DX3501: 16x1G + 6x10G
* 98DX3510: 48x1G + 6x25G
* 98DX3520: 24x2.5G + 6x25G
* 98DX3530: 48x2.5G + 6x25G
* 98DX3540: 12x5G/6x10G + 6x25G
* 98DX3550: 24x5G/12x10G + 6x25G

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---

Notes:
    Changes in v7:
    - Add rd-ac5 and rd-ac5x boards to binding.
    - Rename to armada-98dx25xx.yaml
    Changes in v6:
    - New

 .../bindings/arm/marvell/armada-98dx25xx.yaml | 32 +++++++++++++++++++
 1 file changed, 32 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/arm/marvell/armada-98dx25xx.yaml

diff --git a/Documentation/devicetree/bindings/arm/marvell/armada-98dx25xx.yaml b/Documentation/devicetree/bindings/arm/marvell/armada-98dx25xx.yaml
new file mode 100644
index 000000000000..b99cfb096277
--- /dev/null
+++ b/Documentation/devicetree/bindings/arm/marvell/armada-98dx25xx.yaml
@@ -0,0 +1,32 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/arm/marvell/armada-98dx25xx.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Marvell Alleycat5/5X Platforms
+
+maintainers:
+  - Chris Packham <chris.packham@alliedtelesis.co.nz>
+
+properties:
+  $nodename:
+    const: '/'
+  compatible:
+    oneOf:
+      - description: Alleycat5 (98DX25xx) Reference Design
+        items:
+          - enum:
+              - marvell,rd-ac5
+          - const: marvell,ac5
+
+      - description: Alleycat5X (98DX35xx) Reference Design
+        items:
+          - enum:
+              - marvell,rd-ac5x
+          - const: marvell,ac5x
+          - const: marvell,ac5
+
+additionalProperties: true
+
+...
-- 
2.36.0

[PATCH v7 2/3] arm64: dts: marvell: Add Armada 98DX2530 SoC and RD-AC5X board

[Chris Packham]

The 98DX2530 SoC is the Control and Management CPU integrated into
the Marvell 98DX25xx and 98DX35xx series of switch chip (internally
referred to as AlleyCat5 and AlleyCat5X).

These files have been taken from the Marvell SDK and lightly cleaned
up with the License and copyright retained.

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---

Notes:
    The Marvell SDK has a number of new compatible strings. I've brought
    through some of the drivers or where possible used an in-tree
    alternative (e.g. there is SDK code for a ac5-gpio but two instances of
    the existing marvell,orion-gpio seems to cover what is needed if you use
    an appropriate binding). I expect that there will a new series of
    patches when I get some different hardware (or additions to this series
    depending on if/when it lands).
    
    Changes in v7:
    - Add missing compatible on usb1
    - Add "rd-ac5x" compatible for board
    - Move aliases to board dts
    - Move board specific usb info to board dts
    - Consolidate usb1 board settings and remove unnecessary compatible
    - Add Allied Telesis copyright
    - Rename files after mailng-list discussion
    Changes in v6:
    - Move CPU nodes above the SoC (Krzysztof)
    - Minor formatting clean ups (Krzysztof)
    - Run through `make dtbs_check`
    - Move gic nodes inside SoC
    - Group clocks under a clock node
    Changes in v5:
    - add #{address,size}-cells property to i2c nodes
    - make i2c nodes disabled in the SoC and enable them in the board
    - add interrupt controller attributes to gpio nodes
    - Move fixed-clock nodes up a level and remove unnecessary @0
    Changes in v4:
    - use 'phy-handle' instead of 'phy'
    - move status="okay" on usb nodes to board dts
    - Add review from Andrew
    Changes in v3:
    - Move memory node to board
    - Use single digit reg value for phy address
    - Remove MMC node (driver needs work)
    - Remove syscon & simple-mfd for pinctrl
    Changes in v2:
    - Make pinctrl a child node of a syscon node
    - Use marvell,armada-8k-gpio instead of orion-gpio
    - Remove nand peripheral. The Marvell SDK does have some changes for the
      ac5-nand-controller but I currently lack hardware with NAND fitted so
      I can't test it right now. I've therefore chosen to omit the node and
      not attempted to bring in the driver or binding.
    - Remove pcie peripheral. Again there are changes in the SDK and I have
      no way of testing them.
    - Remove prestera node.
    - Remove "marvell,ac5-ehci" compatible from USB node as
      "marvell,orion-ehci" is sufficient
    - Remove watchdog node. There is a buggy driver for the ac5 watchdog in
      the SDK but it needs some work so I've dropped the node for now.

 arch/arm64/boot/dts/marvell/Makefile          |   1 +
 .../boot/dts/marvell/armada-98dx25xx.dtsi     | 295 ++++++++++++++++++
 .../boot/dts/marvell/armada-98dx35xx-rd.dts   | 101 ++++++
 .../boot/dts/marvell/armada-98dx35xx.dtsi     |  13 +
 4 files changed, 410 insertions(+)
 create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
 create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx-rd.dts
 create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx.dtsi

diff --git a/arch/arm64/boot/dts/marvell/Makefile b/arch/arm64/boot/dts/marvell/Makefile
index 1c794cdcb8e6..b7a4c715afbb 100644
--- a/arch/arm64/boot/dts/marvell/Makefile
+++ b/arch/arm64/boot/dts/marvell/Makefile
@@ -24,3 +24,4 @@ dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db.dtb
 dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db-B.dtb
 dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-A.dtb
 dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-B.dtb
+dtb-$(CONFIG_ARCH_MVEBU) += armada-98dx35xx-rd.dtb
diff --git a/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
new file mode 100644
index 000000000000..55ab4cd843a9
--- /dev/null
+++ b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
@@ -0,0 +1,295 @@
+// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
+/*
+ * Device Tree For AC5.
+ *
+ * Copyright (C) 2021 Marvell
+ * Copyright (C) 2022 Allied Telesis Labs
+ */
+
+#include <dt-bindings/gpio/gpio.h>
+#include <dt-bindings/interrupt-controller/arm-gic.h>
+
+/ {
+	model = "Marvell AC5 SoC";
+	compatible = "marvell,ac5";
+	interrupt-parent = <&gic>;
+	#address-cells = <2>;
+	#size-cells = <2>;
+
+	cpus {
+		#address-cells = <2>;
+		#size-cells = <0>;
+
+		cpu-map {
+			cluster0 {
+				core0 {
+					cpu = <&cpu0>;
+				};
+				core1 {
+					cpu = <&cpu1>;
+				};
+			};
+		};
+
+		cpu0: cpu@0 {
+			device_type = "cpu";
+			compatible = "arm,armv8";
+			reg = <0x0 0x0>;
+			enable-method = "psci";
+			next-level-cache = <&l2>;
+		};
+
+		cpu1: cpu@1 {
+			device_type = "cpu";
+			compatible = "arm,armv8";
+			reg = <0x0 0x100>;
+			enable-method = "psci";
+			next-level-cache = <&l2>;
+		};
+
+		l2: l2-cache {
+			compatible = "cache";
+		};
+	};
+
+
+	psci {
+		compatible = "arm,psci-0.2";
+		method = "smc";
+	};
+
+	timer {
+		compatible = "arm,armv8-timer";
+		interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>,
+				 <GIC_PPI 8 IRQ_TYPE_LEVEL_HIGH>,
+				 <GIC_PPI 10 IRQ_TYPE_LEVEL_HIGH>,
+				 <GIC_PPI 7 IRQ_TYPE_LEVEL_HIGH>;
+		clock-frequency = <25000000>;
+	};
+
+	pmu {
+		compatible = "arm,armv8-pmuv3";
+		interrupts = <GIC_PPI 12 IRQ_TYPE_LEVEL_HIGH>;
+	};
+
+	soc {
+		compatible = "simple-bus";
+		#address-cells = <2>;
+		#size-cells = <2>;
+		ranges;
+		dma-ranges;
+
+		internal-regs@7f000000 {
+			#address-cells = <1>;
+			#size-cells = <1>;
+			compatible = "simple-bus";
+			/* 16M internal register @ 0x7f00_0000 */
+			ranges = <0x0 0x0 0x7f000000 0x1000000>;
+			dma-coherent;
+
+			uart0: serial@12000 {
+				compatible = "snps,dw-apb-uart";
+				reg = <0x12000 0x100>;
+				reg-shift = <2>;
+				interrupts = <GIC_SPI 83 IRQ_TYPE_LEVEL_HIGH>;
+				reg-io-width = <1>;
+				clock-frequency = <328000000>;
+				status = "okay";
+			};
+
+			mdio: mdio@22004 {
+				#address-cells = <1>;
+				#size-cells = <0>;
+				compatible = "marvell,orion-mdio";
+				reg = <0x22004 0x4>;
+				clocks = <&core_clock>;
+			};
+
+			i2c0: i2c@11000{
+				compatible = "marvell,mv78230-i2c";
+				reg = <0x11000 0x20>;
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				clocks = <&core_clock>;
+				clock-names = "core";
+				interrupts = <GIC_SPI 87 IRQ_TYPE_LEVEL_HIGH>;
+				clock-frequency=<100000>;
+
+				pinctrl-names = "default", "gpio";
+				pinctrl-0 = <&i2c0_pins>;
+				pinctrl-1 = <&i2c0_gpio>;
+				scl_gpio = <&gpio0 26 GPIO_ACTIVE_HIGH>;
+				sda_gpio = <&gpio0 27 GPIO_ACTIVE_HIGH>;
+				status = "disabled";
+			};
+
+			i2c1: i2c@11100{
+				compatible = "marvell,mv78230-i2c";
+				reg = <0x11100 0x20>;
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				clocks = <&core_clock>;
+				clock-names = "core";
+				interrupts = <GIC_SPI 88 IRQ_TYPE_LEVEL_HIGH>;
+				clock-frequency=<100000>;
+
+				pinctrl-names = "default", "gpio";
+				pinctrl-0 = <&i2c1_pins>;
+				pinctrl-1 = <&i2c1_gpio>;
+				scl_gpio = <&gpio0 20 GPIO_ACTIVE_HIGH>;
+				sda_gpio = <&gpio0 21 GPIO_ACTIVE_HIGH>;
+				status = "disabled";
+			};
+
+			gpio0: gpio@18100 {
+				compatible = "marvell,orion-gpio";
+				reg = <0x18100 0x40>;
+				ngpios = <32>;
+				gpio-controller;
+				#gpio-cells = <2>;
+				gpio-ranges = <&pinctrl0 0 0 32>;
+				marvell,pwm-offset = <0x1f0>;
+				interrupt-controller;
+				#interrupt-cells = <2>;
+			};
+
+			gpio1: gpio@18140 {
+				reg = <0x18140 0x40>;
+				compatible = "marvell,orion-gpio";
+				ngpios = <14>;
+				gpio-controller;
+				#gpio-cells = <2>;
+				gpio-ranges = <&pinctrl0 0 32 14>;
+				marvell,pwm-offset = <0x1f0>;
+				interrupt-controller;
+				#interrupt-cells = <2>;
+			};
+		};
+
+		/*
+		 * Dedicated section for devices behind 32bit controllers so we
+		 * can configure specific DMA mapping for them
+		 */
+		behind-32bit-controller@7f000000 {
+			compatible = "simple-bus";
+			#address-cells = <0x2>;
+			#size-cells = <0x2>;
+			ranges = <0x0 0x0 0x0 0x7f000000 0x0 0x1000000>;
+			/* Host phy ram starts at 0x200M */
+			dma-ranges = <0x0 0x0 0x2 0x0 0x1 0x0>;
+			dma-coherent;
+
+			eth0: ethernet@20000 {
+				compatible = "marvell,armada-ac5-neta";
+				reg = <0x0 0x20000 0x0 0x4000>;
+				interrupts = <GIC_SPI 45 IRQ_TYPE_LEVEL_HIGH>;
+				clocks = <&core_clock>;
+				phy-mode = "sgmii";
+				status = "disabled";
+			};
+
+			eth1: ethernet@24000 {
+				compatible = "marvell,armada-ac5-neta";
+				reg = <0x0 0x24000 0x0 0x4000>;
+				interrupts = <GIC_SPI 55 IRQ_TYPE_LEVEL_HIGH>;
+				clocks = <&core_clock>;
+				phy-mode = "sgmii";
+				status = "disabled";
+			};
+
+			usb0: usb@80000 {
+				compatible = "marvell,orion-ehci";
+				reg = <0x0 0x80000 0x0 0x500>;
+				interrupts = <GIC_SPI 67 IRQ_TYPE_LEVEL_HIGH>;
+				status = "disabled";
+			};
+
+			usb1: usb@a0000 {
+				compatible = "marvell,orion-ehci";
+				reg = <0x0 0xa0000 0x0 0x500>;
+				interrupts = <GIC_SPI 69 IRQ_TYPE_LEVEL_HIGH>;
+				status = "disabled";
+			};
+		};
+
+		pinctrl0: pinctrl@80020100 {
+			compatible = "marvell,ac5-pinctrl";
+			reg = <0 0x80020100 0 0x20>;
+
+			i2c0_pins: i2c0-pins {
+				marvell,pins = "mpp26", "mpp27";
+				marvell,function = "i2c0";
+			};
+
+			i2c0_gpio: i2c0-gpio-pins {
+				marvell,pins = "mpp26", "mpp27";
+				marvell,function = "gpio";
+			};
+
+			i2c1_pins: i2c1-pins {
+				marvell,pins = "mpp20", "mpp21";
+				marvell,function = "i2c1";
+			};
+
+			i2c1_gpio: i2c1-gpio-pins {
+				marvell,pins = "mpp20", "mpp21";
+				marvell,function = "i2c1";
+			};
+		};
+
+		spi0: spi@805a0000 {
+			compatible = "marvell,armada-3700-spi";
+			reg = <0x0 0x805a0000 0x0 0x50>;
+			#address-cells = <0x1>;
+			#size-cells = <0x0>;
+			clocks = <&spi_clock>;
+			interrupts = <GIC_SPI 90 IRQ_TYPE_LEVEL_HIGH>;
+			num-cs = <1>;
+			status = "disabled";
+		};
+
+		spi1: spi@805a8000 {
+			compatible = "marvell,armada-3700-spi";
+			reg = <0x0 0x805a8000 0x0 0x50>;
+			#address-cells = <0x1>;
+			#size-cells = <0x0>;
+			clocks = <&spi_clock>;
+			interrupts = <GIC_SPI 91 IRQ_TYPE_LEVEL_HIGH>;
+			num-cs = <1>;
+			status = "disabled";
+		};
+
+		gic: interrupt-controller@80600000 {
+			compatible = "arm,gic-v3";
+			#interrupt-cells = <3>;
+			interrupt-controller;
+			/*#redistributor-regions = <1>;*/
+			redistributor-stride = <0x0 0x20000>;	// 128kB stride
+			reg = <0x0 0x80600000 0x0 0x10000>, /* GICD */
+			      <0x0 0x80660000 0x0 0x40000>; /* GICR */
+			interrupts = <GIC_PPI 6 IRQ_TYPE_LEVEL_HIGH>;
+		};
+	};
+
+	clocks {
+		core_clock: core-clock {
+			compatible = "fixed-clock";
+			#clock-cells = <0>;
+			clock-frequency = <400000000>;
+		};
+
+		axi_clock: axi-clock {
+			compatible = "fixed-clock";
+			#clock-cells = <0>;
+			clock-frequency = <325000000>;
+		};
+
+		spi_clock: spi-clock {
+			compatible = "fixed-clock";
+			#clock-cells = <0>;
+			clock-frequency = <200000000>;
+		};
+	};
+};
diff --git a/arch/arm64/boot/dts/marvell/armada-98dx35xx-rd.dts b/arch/arm64/boot/dts/marvell/armada-98dx35xx-rd.dts
new file mode 100644
index 000000000000..c20e032410fa
--- /dev/null
+++ b/arch/arm64/boot/dts/marvell/armada-98dx35xx-rd.dts
@@ -0,0 +1,101 @@
+// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
+/*
+ * Device Tree For RD-AC5X.
+ *
+ * Copyright (C) 2021 Marvell
+ * Copyright (C) 2022 Allied Telesis Labs
+ */
+/*
+ * Device Tree file for Marvell Alleycat 5X development board
+ * This board file supports the B configuration of the board
+ */
+
+/dts-v1/;
+
+#include "armada-98dx35xx.dtsi"
+
+/ {
+	model = "Marvell RD-AC5X Board";
+	compatible = "marvell,rd-ac5x", "marvell,ac5x", "marvell,ac5";
+
+	aliases {
+		serial0 = &uart0;
+		spiflash0 = &spiflash0;
+		gpio0 = &gpio0;
+		gpio1 = &gpio1;
+		ethernet0 = &eth0;
+		ethernet1 = &eth1;
+	};
+
+	memory@0 {
+		device_type = "memory";
+		reg = <0x2 0x00000000 0x0 0x40000000>;
+	};
+
+	usb1phy: usb-phy {
+		compatible = "usb-nop-xceiv";
+		#phy-cells = <0>;
+	};
+};
+
+&mdio {
+	phy0: ethernet-phy@0 {
+		reg = <0>;
+	};
+};
+
+&i2c0 {
+	status = "okay";
+};
+
+&i2c1 {
+	status = "okay";
+};
+
+&eth0 {
+	status = "okay";
+	phy-handle = <&phy0>;
+};
+
+/* USB0 is a host USB */
+&usb0 {
+	status = "okay";
+};
+
+/* USB1 is a peripheral USB */
+&usb1 {
+	status = "okay";
+	phys = <&usb1phy>;
+	phy-names = "usb-phy";
+	dr_mode = "peripheral";
+};
+
+&spi0 {
+	status = "okay";
+
+	spiflash0: flash@0 {
+		compatible = "jedec,spi-nor";
+		spi-max-frequency = <50000000>;
+		spi-tx-bus-width = <1>; /* 1-single, 2-dual, 4-quad */
+		spi-rx-bus-width = <1>; /* 1-single, 2-dual, 4-quad */
+		reg = <0>;
+
+		#address-cells = <1>;
+		#size-cells = <1>;
+
+		partition@0 {
+			label = "spi_flash_part0";
+			reg = <0x0 0x800000>;
+		};
+
+		parition@1 {
+			label = "spi_flash_part1";
+			reg = <0x800000 0x700000>;
+		};
+
+		parition@2 {
+			label = "spi_flash_part2";
+			reg = <0xF00000 0x100000>;
+		};
+	};
+};
diff --git a/arch/arm64/boot/dts/marvell/armada-98dx35xx.dtsi b/arch/arm64/boot/dts/marvell/armada-98dx35xx.dtsi
new file mode 100644
index 000000000000..62d2650513aa
--- /dev/null
+++ b/arch/arm64/boot/dts/marvell/armada-98dx35xx.dtsi
@@ -0,0 +1,13 @@
+// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
+/*
+ * Device Tree For AC5X.
+ *
+ * Copyright (C) 2022 Allied Telesis Labs
+ */
+
+#include "armada-98dx25xx.dtsi"
+
+/ {
+	model = "Marvell AC5X SoC";
+	compatible = "marvell,ac5x", "marvell,ac5";
+};
-- 
2.36.0

[PATCH v7 3/3] arm64: marvell: enable the 98DX2530 pinctrl driver

[Chris Packham]

This commit makes sure the drivers for the 98DX2530 pin controller is
enabled.

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---

Notes:
    Changes in v7:
    - None
    Changes in v6:
    - None
    Changes in v5:
    - None
    Changes in v4:
    - None
    Changes in v3:
    - Add review from Andrew
    Changes in v2:
    - None

 arch/arm64/Kconfig.platforms | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/arm64/Kconfig.platforms b/arch/arm64/Kconfig.platforms
index 30b123cde02c..229571d57496 100644
--- a/arch/arm64/Kconfig.platforms
+++ b/arch/arm64/Kconfig.platforms
@@ -183,11 +183,13 @@ config ARCH_MVEBU
 	select PINCTRL_ARMADA_37XX
 	select PINCTRL_ARMADA_AP806
 	select PINCTRL_ARMADA_CP110
+	select PINCTRL_AC5
 	help
 	  This enables support for Marvell EBU familly, including:
 	   - Armada 3700 SoC Family
 	   - Armada 7K SoC Family
 	   - Armada 8K SoC Family
+	   - 98DX2530 SoC Family
 
 config ARCH_MXC
 	bool "ARMv8 based NXP i.MX SoC family"
-- 
2.36.0

Re: [PATCH v7 2/3] arm64: dts: marvell: Add Armada 98DX2530 SoC and RD-AC5X board

[Marc Zyngier]

On Thu, 12 May 2022 05:25:00 +0100,
Chris Packham <chris.packham@alliedtelesis.co.nz> wrote:
> 
> The 98DX2530 SoC is the Control and Management CPU integrated into
> the Marvell 98DX25xx and 98DX35xx series of switch chip (internally
> referred to as AlleyCat5 and AlleyCat5X).
> 
> These files have been taken from the Marvell SDK and lightly cleaned
> up with the License and copyright retained.
> 
> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>
> ---
> 
> Notes:
>     The Marvell SDK has a number of new compatible strings. I've brought
>     through some of the drivers or where possible used an in-tree
>     alternative (e.g. there is SDK code for a ac5-gpio but two instances of
>     the existing marvell,orion-gpio seems to cover what is needed if you use
>     an appropriate binding). I expect that there will a new series of
>     patches when I get some different hardware (or additions to this series
>     depending on if/when it lands).
>     
>     Changes in v7:
>     - Add missing compatible on usb1
>     - Add "rd-ac5x" compatible for board
>     - Move aliases to board dts
>     - Move board specific usb info to board dts
>     - Consolidate usb1 board settings and remove unnecessary compatible
>     - Add Allied Telesis copyright
>     - Rename files after mailng-list discussion
>     Changes in v6:
>     - Move CPU nodes above the SoC (Krzysztof)
>     - Minor formatting clean ups (Krzysztof)
>     - Run through `make dtbs_check`
>     - Move gic nodes inside SoC
>     - Group clocks under a clock node
>     Changes in v5:
>     - add #{address,size}-cells property to i2c nodes
>     - make i2c nodes disabled in the SoC and enable them in the board
>     - add interrupt controller attributes to gpio nodes
>     - Move fixed-clock nodes up a level and remove unnecessary @0
>     Changes in v4:
>     - use 'phy-handle' instead of 'phy'
>     - move status="okay" on usb nodes to board dts
>     - Add review from Andrew
>     Changes in v3:
>     - Move memory node to board
>     - Use single digit reg value for phy address
>     - Remove MMC node (driver needs work)
>     - Remove syscon & simple-mfd for pinctrl
>     Changes in v2:
>     - Make pinctrl a child node of a syscon node
>     - Use marvell,armada-8k-gpio instead of orion-gpio
>     - Remove nand peripheral. The Marvell SDK does have some changes for the
>       ac5-nand-controller but I currently lack hardware with NAND fitted so
>       I can't test it right now. I've therefore chosen to omit the node and
>       not attempted to bring in the driver or binding.
>     - Remove pcie peripheral. Again there are changes in the SDK and I have
>       no way of testing them.
>     - Remove prestera node.
>     - Remove "marvell,ac5-ehci" compatible from USB node as
>       "marvell,orion-ehci" is sufficient
>     - Remove watchdog node. There is a buggy driver for the ac5 watchdog in
>       the SDK but it needs some work so I've dropped the node for now.
> 
>  arch/arm64/boot/dts/marvell/Makefile          |   1 +
>  .../boot/dts/marvell/armada-98dx25xx.dtsi     | 295 ++++++++++++++++++
>  .../boot/dts/marvell/armada-98dx35xx-rd.dts   | 101 ++++++
>  .../boot/dts/marvell/armada-98dx35xx.dtsi     |  13 +
>  4 files changed, 410 insertions(+)
>  create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>  create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx-rd.dts
>  create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx.dtsi
> 
> diff --git a/arch/arm64/boot/dts/marvell/Makefile b/arch/arm64/boot/dts/marvell/Makefile
> index 1c794cdcb8e6..b7a4c715afbb 100644
> --- a/arch/arm64/boot/dts/marvell/Makefile
> +++ b/arch/arm64/boot/dts/marvell/Makefile
> @@ -24,3 +24,4 @@ dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db.dtb
>  dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db-B.dtb
>  dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-A.dtb
>  dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-B.dtb
> +dtb-$(CONFIG_ARCH_MVEBU) += armada-98dx35xx-rd.dtb
> diff --git a/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
> new file mode 100644
> index 000000000000..55ab4cd843a9
> --- /dev/null
> +++ b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
> @@ -0,0 +1,295 @@
> +// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
> +/*
> + * Device Tree For AC5.
> + *
> + * Copyright (C) 2021 Marvell
> + * Copyright (C) 2022 Allied Telesis Labs
> + */
> +
> +#include <dt-bindings/gpio/gpio.h>
> +#include <dt-bindings/interrupt-controller/arm-gic.h>
> +
> +/ {
> +	model = "Marvell AC5 SoC";
> +	compatible = "marvell,ac5";
> +	interrupt-parent = <&gic>;
> +	#address-cells = <2>;
> +	#size-cells = <2>;
> +
> +	cpus {
> +		#address-cells = <2>;
> +		#size-cells = <0>;
> +
> +		cpu-map {
> +			cluster0 {
> +				core0 {
> +					cpu = <&cpu0>;
> +				};
> +				core1 {
> +					cpu = <&cpu1>;
> +				};
> +			};
> +		};
> +
> +		cpu0: cpu@0 {
> +			device_type = "cpu";
> +			compatible = "arm,armv8";
> +			reg = <0x0 0x0>;
> +			enable-method = "psci";
> +			next-level-cache = <&l2>;
> +		};
> +
> +		cpu1: cpu@1 {
> +			device_type = "cpu";
> +			compatible = "arm,armv8";
> +			reg = <0x0 0x100>;
> +			enable-method = "psci";
> +			next-level-cache = <&l2>;
> +		};
> +
> +		l2: l2-cache {
> +			compatible = "cache";
> +		};
> +	};
> +
> +
> +	psci {
> +		compatible = "arm,psci-0.2";
> +		method = "smc";
> +	};
> +
> +	timer {
> +		compatible = "arm,armv8-timer";
> +		interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>,
> +				 <GIC_PPI 8 IRQ_TYPE_LEVEL_HIGH>,
> +				 <GIC_PPI 10 IRQ_TYPE_LEVEL_HIGH>,
> +				 <GIC_PPI 7 IRQ_TYPE_LEVEL_HIGH>;
> +		clock-frequency = <25000000>;

I said no to this hack in a past version of this patch, and I'm going
to say it *again*. Please fix your firmware to program CNTFRQ_EL0, and
remove this useless property.

You are also missing a PPI for the EL2 virtual timer which is present
on any ARMv8.1+ CPU (and since this system is using A55, it definitely
has it).

[...]

> +
> +		gic: interrupt-controller@80600000 {
> +			compatible = "arm,gic-v3";
> +			#interrupt-cells = <3>;
> +			interrupt-controller;
> +			/*#redistributor-regions = <1>;*/
> +			redistributor-stride = <0x0 0x20000>;	// 128kB stride

You don't need this at all. This is the architected value for GICv3.

> +			reg = <0x0 0x80600000 0x0 0x10000>, /* GICD */
> +			      <0x0 0x80660000 0x0 0x40000>; /* GICR */
> +			interrupts = <GIC_PPI 6 IRQ_TYPE_LEVEL_HIGH>;
> +		};
> +	};

Thanks,

	M.

-- 
Without deviation from the norm, progress is not possible.

Re: [PATCH v7 2/3] arm64: dts: marvell: Add Armada 98DX2530 SoC and RD-AC5X board

[Chris Packham]

Hi Marc,

On 12/05/22 19:38, Marc Zyngier wrote:
> On Thu, 12 May 2022 05:25:00 +0100,
> Chris Packham <chris.packham@alliedtelesis.co.nz> wrote:
>> The 98DX2530 SoC is the Control and Management CPU integrated into
>> the Marvell 98DX25xx and 98DX35xx series of switch chip (internally
>> referred to as AlleyCat5 and AlleyCat5X).
>>
>> These files have been taken from the Marvell SDK and lightly cleaned
>> up with the License and copyright retained.
>>
>> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
>> Reviewed-by: Andrew Lunn <andrew@lunn.ch>
>> ---
>>
>> Notes:
>>      The Marvell SDK has a number of new compatible strings. I've brought
>>      through some of the drivers or where possible used an in-tree
>>      alternative (e.g. there is SDK code for a ac5-gpio but two instances of
>>      the existing marvell,orion-gpio seems to cover what is needed if you use
>>      an appropriate binding). I expect that there will a new series of
>>      patches when I get some different hardware (or additions to this series
>>      depending on if/when it lands).
>>      
>>      Changes in v7:
>>      - Add missing compatible on usb1
>>      - Add "rd-ac5x" compatible for board
>>      - Move aliases to board dts
>>      - Move board specific usb info to board dts
>>      - Consolidate usb1 board settings and remove unnecessary compatible
>>      - Add Allied Telesis copyright
>>      - Rename files after mailng-list discussion
>>      Changes in v6:
>>      - Move CPU nodes above the SoC (Krzysztof)
>>      - Minor formatting clean ups (Krzysztof)
>>      - Run through `make dtbs_check`
>>      - Move gic nodes inside SoC
>>      - Group clocks under a clock node
>>      Changes in v5:
>>      - add #{address,size}-cells property to i2c nodes
>>      - make i2c nodes disabled in the SoC and enable them in the board
>>      - add interrupt controller attributes to gpio nodes
>>      - Move fixed-clock nodes up a level and remove unnecessary @0
>>      Changes in v4:
>>      - use 'phy-handle' instead of 'phy'
>>      - move status="okay" on usb nodes to board dts
>>      - Add review from Andrew
>>      Changes in v3:
>>      - Move memory node to board
>>      - Use single digit reg value for phy address
>>      - Remove MMC node (driver needs work)
>>      - Remove syscon & simple-mfd for pinctrl
>>      Changes in v2:
>>      - Make pinctrl a child node of a syscon node
>>      - Use marvell,armada-8k-gpio instead of orion-gpio
>>      - Remove nand peripheral. The Marvell SDK does have some changes for the
>>        ac5-nand-controller but I currently lack hardware with NAND fitted so
>>        I can't test it right now. I've therefore chosen to omit the node and
>>        not attempted to bring in the driver or binding.
>>      - Remove pcie peripheral. Again there are changes in the SDK and I have
>>        no way of testing them.
>>      - Remove prestera node.
>>      - Remove "marvell,ac5-ehci" compatible from USB node as
>>        "marvell,orion-ehci" is sufficient
>>      - Remove watchdog node. There is a buggy driver for the ac5 watchdog in
>>        the SDK but it needs some work so I've dropped the node for now.
>>
>>   arch/arm64/boot/dts/marvell/Makefile          |   1 +
>>   .../boot/dts/marvell/armada-98dx25xx.dtsi     | 295 ++++++++++++++++++
>>   .../boot/dts/marvell/armada-98dx35xx-rd.dts   | 101 ++++++
>>   .../boot/dts/marvell/armada-98dx35xx.dtsi     |  13 +
>>   4 files changed, 410 insertions(+)
>>   create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>>   create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx-rd.dts
>>   create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx.dtsi
>>
>> diff --git a/arch/arm64/boot/dts/marvell/Makefile b/arch/arm64/boot/dts/marvell/Makefile
>> index 1c794cdcb8e6..b7a4c715afbb 100644
>> --- a/arch/arm64/boot/dts/marvell/Makefile
>> +++ b/arch/arm64/boot/dts/marvell/Makefile
>> @@ -24,3 +24,4 @@ dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db.dtb
>>   dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db-B.dtb
>>   dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-A.dtb
>>   dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-B.dtb
>> +dtb-$(CONFIG_ARCH_MVEBU) += armada-98dx35xx-rd.dtb
>> diff --git a/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>> new file mode 100644
>> index 000000000000..55ab4cd843a9
>> --- /dev/null
>> +++ b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>> @@ -0,0 +1,295 @@
>> +// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
>> +/*
>> + * Device Tree For AC5.
>> + *
>> + * Copyright (C) 2021 Marvell
>> + * Copyright (C) 2022 Allied Telesis Labs
>> + */
>> +
>> +#include <dt-bindings/gpio/gpio.h>
>> +#include <dt-bindings/interrupt-controller/arm-gic.h>
>> +
>> +/ {
>> +	model = "Marvell AC5 SoC";
>> +	compatible = "marvell,ac5";
>> +	interrupt-parent = <&gic>;
>> +	#address-cells = <2>;
>> +	#size-cells = <2>;
>> +
>> +	cpus {
>> +		#address-cells = <2>;
>> +		#size-cells = <0>;
>> +
>> +		cpu-map {
>> +			cluster0 {
>> +				core0 {
>> +					cpu = <&cpu0>;
>> +				};
>> +				core1 {
>> +					cpu = <&cpu1>;
>> +				};
>> +			};
>> +		};
>> +
>> +		cpu0: cpu@0 {
>> +			device_type = "cpu";
>> +			compatible = "arm,armv8";
>> +			reg = <0x0 0x0>;
>> +			enable-method = "psci";
>> +			next-level-cache = <&l2>;
>> +		};
>> +
>> +		cpu1: cpu@1 {
>> +			device_type = "cpu";
>> +			compatible = "arm,armv8";
>> +			reg = <0x0 0x100>;
>> +			enable-method = "psci";
>> +			next-level-cache = <&l2>;
>> +		};
>> +
>> +		l2: l2-cache {
>> +			compatible = "cache";
>> +		};
>> +	};
>> +
>> +
>> +	psci {
>> +		compatible = "arm,psci-0.2";
>> +		method = "smc";
>> +	};
>> +
>> +	timer {
>> +		compatible = "arm,armv8-timer";
>> +		interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>,
>> +				 <GIC_PPI 8 IRQ_TYPE_LEVEL_HIGH>,
>> +				 <GIC_PPI 10 IRQ_TYPE_LEVEL_HIGH>,
>> +				 <GIC_PPI 7 IRQ_TYPE_LEVEL_HIGH>;
>> +		clock-frequency = <25000000>;
> I said no to this hack in a past version of this patch, and I'm going
> to say it *again*.
Sorry I must have missed it.
> Please fix your firmware to program CNTFRQ_EL0, and
> remove this useless property.
I'm kind of at the mercy of what Marvell have provided for ATF. I am 
working on the bootloader portion in parallel and am getting things 
ready for submitting the u-boot support upstream. I was hoping to leave 
ATF alone I can at least see if they haven't fixed this already (the 
original dtsi I started with was fairly old) and if they haven't I'll 
raise it via their support system.
> You are also missing a PPI for the EL2 virtual timer which is present
> on any ARMv8.1+ CPU (and since this system is using A55, it definitely
> has it).
>
> [...]
Will add.
>> +
>> +		gic: interrupt-controller@80600000 {
>> +			compatible = "arm,gic-v3";
>> +			#interrupt-cells = <3>;
>> +			interrupt-controller;
>> +			/*#redistributor-regions = <1>;*/
>> +			redistributor-stride = <0x0 0x20000>;	// 128kB stride
> You don't need this at all. This is the architected value for GICv3.
Will remove.
>
>> +			reg = <0x0 0x80600000 0x0 0x10000>, /* GICD */
>> +			      <0x0 0x80660000 0x0 0x40000>; /* GICR */
>> +			interrupts = <GIC_PPI 6 IRQ_TYPE_LEVEL_HIGH>;
>> +		};
>> +	};
> Thanks,
>
> 	M.
>

Re: [PATCH v7 2/3] arm64: dts: marvell: Add Armada 98DX2530 SoC and RD-AC5X board

[Chris Packham]

Hi Marc,

On 13/05/22 10:10, Chris Packham wrote:
> Hi Marc,
>
> On 12/05/22 19:38, Marc Zyngier wrote:
>> On Thu, 12 May 2022 05:25:00 +0100,
>> Chris Packham <chris.packham@alliedtelesis.co.nz> wrote:
>>> The 98DX2530 SoC is the Control and Management CPU integrated into
>>> the Marvell 98DX25xx and 98DX35xx series of switch chip (internally
>>> referred to as AlleyCat5 and AlleyCat5X).
>>>
>>> These files have been taken from the Marvell SDK and lightly cleaned
>>> up with the License and copyright retained.
>>>
>>> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
>>> Reviewed-by: Andrew Lunn <andrew@lunn.ch>
>>> ---
>>>
>>> Notes:
>>>      The Marvell SDK has a number of new compatible strings. I've 
>>> brought
>>>      through some of the drivers or where possible used an in-tree
>>>      alternative (e.g. there is SDK code for a ac5-gpio but two 
>>> instances of
>>>      the existing marvell,orion-gpio seems to cover what is needed 
>>> if you use
>>>      an appropriate binding). I expect that there will a new series of
>>>      patches when I get some different hardware (or additions to 
>>> this series
>>>      depending on if/when it lands).
>>>           Changes in v7:
>>>      - Add missing compatible on usb1
>>>      - Add "rd-ac5x" compatible for board
>>>      - Move aliases to board dts
>>>      - Move board specific usb info to board dts
>>>      - Consolidate usb1 board settings and remove unnecessary 
>>> compatible
>>>      - Add Allied Telesis copyright
>>>      - Rename files after mailng-list discussion
>>>      Changes in v6:
>>>      - Move CPU nodes above the SoC (Krzysztof)
>>>      - Minor formatting clean ups (Krzysztof)
>>>      - Run through `make dtbs_check`
>>>      - Move gic nodes inside SoC
>>>      - Group clocks under a clock node
>>>      Changes in v5:
>>>      - add #{address,size}-cells property to i2c nodes
>>>      - make i2c nodes disabled in the SoC and enable them in the board
>>>      - add interrupt controller attributes to gpio nodes
>>>      - Move fixed-clock nodes up a level and remove unnecessary @0
>>>      Changes in v4:
>>>      - use 'phy-handle' instead of 'phy'
>>>      - move status="okay" on usb nodes to board dts
>>>      - Add review from Andrew
>>>      Changes in v3:
>>>      - Move memory node to board
>>>      - Use single digit reg value for phy address
>>>      - Remove MMC node (driver needs work)
>>>      - Remove syscon & simple-mfd for pinctrl
>>>      Changes in v2:
>>>      - Make pinctrl a child node of a syscon node
>>>      - Use marvell,armada-8k-gpio instead of orion-gpio
>>>      - Remove nand peripheral. The Marvell SDK does have some 
>>> changes for the
>>>        ac5-nand-controller but I currently lack hardware with NAND 
>>> fitted so
>>>        I can't test it right now. I've therefore chosen to omit the 
>>> node and
>>>        not attempted to bring in the driver or binding.
>>>      - Remove pcie peripheral. Again there are changes in the SDK 
>>> and I have
>>>        no way of testing them.
>>>      - Remove prestera node.
>>>      - Remove "marvell,ac5-ehci" compatible from USB node as
>>>        "marvell,orion-ehci" is sufficient
>>>      - Remove watchdog node. There is a buggy driver for the ac5 
>>> watchdog in
>>>        the SDK but it needs some work so I've dropped the node for now.
>>>
>>>   arch/arm64/boot/dts/marvell/Makefile          |   1 +
>>>   .../boot/dts/marvell/armada-98dx25xx.dtsi     | 295 
>>> ++++++++++++++++++
>>>   .../boot/dts/marvell/armada-98dx35xx-rd.dts   | 101 ++++++
>>>   .../boot/dts/marvell/armada-98dx35xx.dtsi     |  13 +
>>>   4 files changed, 410 insertions(+)
>>>   create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>>>   create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx-rd.dts
>>>   create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx.dtsi
>>>
>>> diff --git a/arch/arm64/boot/dts/marvell/Makefile 
>>> b/arch/arm64/boot/dts/marvell/Makefile
>>> index 1c794cdcb8e6..b7a4c715afbb 100644
>>> --- a/arch/arm64/boot/dts/marvell/Makefile
>>> +++ b/arch/arm64/boot/dts/marvell/Makefile
>>> @@ -24,3 +24,4 @@ dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db.dtb
>>>   dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db-B.dtb
>>>   dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-A.dtb
>>>   dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-B.dtb
>>> +dtb-$(CONFIG_ARCH_MVEBU) += armada-98dx35xx-rd.dtb
>>> diff --git a/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi 
>>> b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>>> new file mode 100644
>>> index 000000000000..55ab4cd843a9
>>> --- /dev/null
>>> +++ b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>>> @@ -0,0 +1,295 @@
>>> +// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
>>> +/*
>>> + * Device Tree For AC5.
>>> + *
>>> + * Copyright (C) 2021 Marvell
>>> + * Copyright (C) 2022 Allied Telesis Labs
>>> + */
>>> +
>>> +#include <dt-bindings/gpio/gpio.h>
>>> +#include <dt-bindings/interrupt-controller/arm-gic.h>
>>> +
>>> +/ {
>>> +    model = "Marvell AC5 SoC";
>>> +    compatible = "marvell,ac5";
>>> +    interrupt-parent = <&gic>;
>>> +    #address-cells = <2>;
>>> +    #size-cells = <2>;
>>> +
>>> +    cpus {
>>> +        #address-cells = <2>;
>>> +        #size-cells = <0>;
>>> +
>>> +        cpu-map {
>>> +            cluster0 {
>>> +                core0 {
>>> +                    cpu = <&cpu0>;
>>> +                };
>>> +                core1 {
>>> +                    cpu = <&cpu1>;
>>> +                };
>>> +            };
>>> +        };
>>> +
>>> +        cpu0: cpu@0 {
>>> +            device_type = "cpu";
>>> +            compatible = "arm,armv8";
>>> +            reg = <0x0 0x0>;
>>> +            enable-method = "psci";
>>> +            next-level-cache = <&l2>;
>>> +        };
>>> +
>>> +        cpu1: cpu@1 {
>>> +            device_type = "cpu";
>>> +            compatible = "arm,armv8";
>>> +            reg = <0x0 0x100>;
>>> +            enable-method = "psci";
>>> +            next-level-cache = <&l2>;
>>> +        };
>>> +
>>> +        l2: l2-cache {
>>> +            compatible = "cache";
>>> +        };
>>> +    };
>>> +
>>> +
>>> +    psci {
>>> +        compatible = "arm,psci-0.2";
>>> +        method = "smc";
>>> +    };
>>> +
>>> +    timer {
>>> +        compatible = "arm,armv8-timer";
>>> +        interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>,
>>> +                 <GIC_PPI 8 IRQ_TYPE_LEVEL_HIGH>,
>>> +                 <GIC_PPI 10 IRQ_TYPE_LEVEL_HIGH>,
>>> +                 <GIC_PPI 7 IRQ_TYPE_LEVEL_HIGH>;
>>> +        clock-frequency = <25000000>;
>> I said no to this hack in a past version of this patch, and I'm going
>> to say it *again*.
> Sorry I must have missed it.
>> Please fix your firmware to program CNTFRQ_EL0, and
>> remove this useless property.
> I'm kind of at the mercy of what Marvell have provided for ATF. I am 
> working on the bootloader portion in parallel and am getting things 
> ready for submitting the u-boot support upstream. I was hoping to 
> leave ATF alone I can at least see if they haven't fixed this already 
> (the original dtsi I started with was fairly old) and if they haven't 
> I'll raise it via their support system.
Seems to work fine without the clock so I'll drop it.
>> You are also missing a PPI for the EL2 virtual timer which is present
>> on any ARMv8.1+ CPU (and since this system is using A55, it definitely
>> has it).
>>
>> [...]
> Will add.
I assume you're talking about the 5th PPI per the 
timer/arm,arch_timer.yaml ("hypervisor virtual timer irq"). Helpfully 
Marvell don't include the PPI interrupt numbers in their datasheet. But 
then I also notice that none of the other boards that have a 
"arm,armv8-timer" provide a 5th interrupt either, have I misunderstood 
something?
>>> +
>>> +        gic: interrupt-controller@80600000 {
>>> +            compatible = "arm,gic-v3";
>>> +            #interrupt-cells = <3>;
>>> +            interrupt-controller;
>>> +            /*#redistributor-regions = <1>;*/
>>> +            redistributor-stride = <0x0 0x20000>;    // 128kB stride
>> You don't need this at all. This is the architected value for GICv3.
> Will remove.
>>
>>> +            reg = <0x0 0x80600000 0x0 0x10000>, /* GICD */
>>> +                  <0x0 0x80660000 0x0 0x40000>; /* GICR */
>>> +            interrupts = <GIC_PPI 6 IRQ_TYPE_LEVEL_HIGH>;
>>> +        };
>>> +    };
>> Thanks,
>>
>>     M.
>>

Re: [PATCH v7 1/3] dt-bindings: marvell: Document the AC5/AC5X compatibles

[Krzysztof Kozlowski]

On 12/05/2022 06:24, Chris Packham wrote:
> Describe the compatible properties for the Marvell Alleycat5/5X switches
> with integrated CPUs.
> 


Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>


Best regards,
Krzysztof

Re: [PATCH v7 2/3] arm64: dts: marvell: Add Armada 98DX2530 SoC and RD-AC5X board

[Marc Zyngier]

On Fri, 13 May 2022 02:26:21 +0100,
Chris Packham <Chris.Packham@alliedtelesis.co.nz> wrote:
> 
> Hi Marc,
> 
> On 13/05/22 10:10, Chris Packham wrote:
> > Hi Marc,
> >
> > On 12/05/22 19:38, Marc Zyngier wrote:
> >> On Thu, 12 May 2022 05:25:00 +0100,
> >> Chris Packham <chris.packham@alliedtelesis.co.nz> wrote:
> >>> The 98DX2530 SoC is the Control and Management CPU integrated into
> >>> the Marvell 98DX25xx and 98DX35xx series of switch chip (internally
> >>> referred to as AlleyCat5 and AlleyCat5X).
> >>>
> >>> These files have been taken from the Marvell SDK and lightly cleaned
> >>> up with the License and copyright retained.
> >>>
> >>> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
> >>> Reviewed-by: Andrew Lunn <andrew@lunn.ch>
> >>> ---
> >>>
> >>> Notes:
> >>>      The Marvell SDK has a number of new compatible strings. I've 
> >>> brought
> >>>      through some of the drivers or where possible used an in-tree
> >>>      alternative (e.g. there is SDK code for a ac5-gpio but two 
> >>> instances of
> >>>      the existing marvell,orion-gpio seems to cover what is needed 
> >>> if you use
> >>>      an appropriate binding). I expect that there will a new series of
> >>>      patches when I get some different hardware (or additions to 
> >>> this series
> >>>      depending on if/when it lands).
> >>>           Changes in v7:
> >>>      - Add missing compatible on usb1
> >>>      - Add "rd-ac5x" compatible for board
> >>>      - Move aliases to board dts
> >>>      - Move board specific usb info to board dts
> >>>      - Consolidate usb1 board settings and remove unnecessary 
> >>> compatible
> >>>      - Add Allied Telesis copyright
> >>>      - Rename files after mailng-list discussion
> >>>      Changes in v6:
> >>>      - Move CPU nodes above the SoC (Krzysztof)
> >>>      - Minor formatting clean ups (Krzysztof)
> >>>      - Run through `make dtbs_check`
> >>>      - Move gic nodes inside SoC
> >>>      - Group clocks under a clock node
> >>>      Changes in v5:
> >>>      - add #{address,size}-cells property to i2c nodes
> >>>      - make i2c nodes disabled in the SoC and enable them in the board
> >>>      - add interrupt controller attributes to gpio nodes
> >>>      - Move fixed-clock nodes up a level and remove unnecessary @0
> >>>      Changes in v4:
> >>>      - use 'phy-handle' instead of 'phy'
> >>>      - move status="okay" on usb nodes to board dts
> >>>      - Add review from Andrew
> >>>      Changes in v3:
> >>>      - Move memory node to board
> >>>      - Use single digit reg value for phy address
> >>>      - Remove MMC node (driver needs work)
> >>>      - Remove syscon & simple-mfd for pinctrl
> >>>      Changes in v2:
> >>>      - Make pinctrl a child node of a syscon node
> >>>      - Use marvell,armada-8k-gpio instead of orion-gpio
> >>>      - Remove nand peripheral. The Marvell SDK does have some 
> >>> changes for the
> >>>        ac5-nand-controller but I currently lack hardware with NAND 
> >>> fitted so
> >>>        I can't test it right now. I've therefore chosen to omit the 
> >>> node and
> >>>        not attempted to bring in the driver or binding.
> >>>      - Remove pcie peripheral. Again there are changes in the SDK 
> >>> and I have
> >>>        no way of testing them.
> >>>      - Remove prestera node.
> >>>      - Remove "marvell,ac5-ehci" compatible from USB node as
> >>>        "marvell,orion-ehci" is sufficient
> >>>      - Remove watchdog node. There is a buggy driver for the ac5 
> >>> watchdog in
> >>>        the SDK but it needs some work so I've dropped the node for now.
> >>>
> >>>   arch/arm64/boot/dts/marvell/Makefile          |   1 +
> >>>   .../boot/dts/marvell/armada-98dx25xx.dtsi     | 295 
> >>> ++++++++++++++++++
> >>>   .../boot/dts/marvell/armada-98dx35xx-rd.dts   | 101 ++++++
> >>>   .../boot/dts/marvell/armada-98dx35xx.dtsi     |  13 +
> >>>   4 files changed, 410 insertions(+)
> >>>   create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
> >>>   create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx-rd.dts
> >>>   create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx.dtsi
> >>>
> >>> diff --git a/arch/arm64/boot/dts/marvell/Makefile 
> >>> b/arch/arm64/boot/dts/marvell/Makefile
> >>> index 1c794cdcb8e6..b7a4c715afbb 100644
> >>> --- a/arch/arm64/boot/dts/marvell/Makefile
> >>> +++ b/arch/arm64/boot/dts/marvell/Makefile
> >>> @@ -24,3 +24,4 @@ dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db.dtb
> >>>   dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db-B.dtb
> >>>   dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-A.dtb
> >>>   dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-B.dtb
> >>> +dtb-$(CONFIG_ARCH_MVEBU) += armada-98dx35xx-rd.dtb
> >>> diff --git a/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi 
> >>> b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
> >>> new file mode 100644
> >>> index 000000000000..55ab4cd843a9
> >>> --- /dev/null
> >>> +++ b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
> >>> @@ -0,0 +1,295 @@
> >>> +// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
> >>> +/*
> >>> + * Device Tree For AC5.
> >>> + *
> >>> + * Copyright (C) 2021 Marvell
> >>> + * Copyright (C) 2022 Allied Telesis Labs
> >>> + */
> >>> +
> >>> +#include <dt-bindings/gpio/gpio.h>
> >>> +#include <dt-bindings/interrupt-controller/arm-gic.h>
> >>> +
> >>> +/ {
> >>> +    model = "Marvell AC5 SoC";
> >>> +    compatible = "marvell,ac5";
> >>> +    interrupt-parent = <&gic>;
> >>> +    #address-cells = <2>;
> >>> +    #size-cells = <2>;
> >>> +
> >>> +    cpus {
> >>> +        #address-cells = <2>;
> >>> +        #size-cells = <0>;
> >>> +
> >>> +        cpu-map {
> >>> +            cluster0 {
> >>> +                core0 {
> >>> +                    cpu = <&cpu0>;
> >>> +                };
> >>> +                core1 {
> >>> +                    cpu = <&cpu1>;
> >>> +                };
> >>> +            };
> >>> +        };
> >>> +
> >>> +        cpu0: cpu@0 {
> >>> +            device_type = "cpu";
> >>> +            compatible = "arm,armv8";
> >>> +            reg = <0x0 0x0>;
> >>> +            enable-method = "psci";
> >>> +            next-level-cache = <&l2>;
> >>> +        };
> >>> +
> >>> +        cpu1: cpu@1 {
> >>> +            device_type = "cpu";
> >>> +            compatible = "arm,armv8";
> >>> +            reg = <0x0 0x100>;
> >>> +            enable-method = "psci";
> >>> +            next-level-cache = <&l2>;
> >>> +        };
> >>> +
> >>> +        l2: l2-cache {
> >>> +            compatible = "cache";
> >>> +        };
> >>> +    };
> >>> +
> >>> +
> >>> +    psci {
> >>> +        compatible = "arm,psci-0.2";
> >>> +        method = "smc";
> >>> +    };
> >>> +
> >>> +    timer {
> >>> +        compatible = "arm,armv8-timer";
> >>> +        interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>,
> >>> +                 <GIC_PPI 8 IRQ_TYPE_LEVEL_HIGH>,
> >>> +                 <GIC_PPI 10 IRQ_TYPE_LEVEL_HIGH>,
> >>> +                 <GIC_PPI 7 IRQ_TYPE_LEVEL_HIGH>;
> >>> +        clock-frequency = <25000000>;
> >> I said no to this hack in a past version of this patch, and I'm going
> >> to say it *again*.
> > Sorry I must have missed it.
> >> Please fix your firmware to program CNTFRQ_EL0, and
> >> remove this useless property.
> > I'm kind of at the mercy of what Marvell have provided for ATF. I am 
> > working on the bootloader portion in parallel and am getting things 
> > ready for submitting the u-boot support upstream. I was hoping to 
> > leave ATF alone I can at least see if they haven't fixed this already 
> > (the original dtsi I started with was fairly old) and if they haven't 
> > I'll raise it via their support system.
> Seems to work fine without the clock so I'll drop it.

Thanks. If you can, please verify that this is set on both CPUs (I
have seen plenty of firmware only setting it on CPU0 in the past).

> >> You are also missing a PPI for the EL2 virtual timer which is present
> >> on any ARMv8.1+ CPU (and since this system is using A55, it definitely
> >> has it).
> >>
> >> [...]
> > Will add.
> I assume you're talking about the 5th PPI per the 
> timer/arm,arch_timer.yaml ("hypervisor virtual timer irq").

Indeed.

> Helpfully
> Marvell don't include the PPI interrupt numbers in their datasheet. But 
> then I also notice that none of the other boards that have a 
> "arm,armv8-timer" provide a 5th interrupt either, have I misunderstood 
> something?

This was only recently added to the DT binding, but the interrupt
definitely exist at the CPU level for anything that implements ARMv8.1
and up. AFAIK, the M1 is the only machine to expose this interrupt in
DT, but this doesn't mean the interrupt doesn't exist on all the other
systems that have the same architecture revision.

If you have contacts in Marvell, maybe try and find out whether they
have simply decided not to wire the interrupt (I wouldn't be
surprised). In this case, please add a comment.

Thanks,

	M.

-- 
Without deviation from the norm, progress is not possible.

Re: [PATCH v7 2/3] arm64: dts: marvell: Add Armada 98DX2530 SoC and RD-AC5X board

[Chris Packham]

On 16/05/22 21:48, Marc Zyngier wrote:
> On Fri, 13 May 2022 02:26:21 +0100,
> Chris Packham <Chris.Packham@alliedtelesis.co.nz> wrote:
>> Hi Marc,
>>
>> On 13/05/22 10:10, Chris Packham wrote:
>>> Hi Marc,
>>>
>>> On 12/05/22 19:38, Marc Zyngier wrote:
>>>> On Thu, 12 May 2022 05:25:00 +0100,
>>>> Chris Packham <chris.packham@alliedtelesis.co.nz> wrote:
>>>>> The 98DX2530 SoC is the Control and Management CPU integrated into
>>>>> the Marvell 98DX25xx and 98DX35xx series of switch chip (internally
>>>>> referred to as AlleyCat5 and AlleyCat5X).
>>>>>
>>>>> These files have been taken from the Marvell SDK and lightly cleaned
>>>>> up with the License and copyright retained.
>>>>>
>>>>> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
>>>>> Reviewed-by: Andrew Lunn <andrew@lunn.ch>
>>>>> ---
>>>>>
>>>>> Notes:
>>>>>       The Marvell SDK has a number of new compatible strings. I've
>>>>> brought
>>>>>       through some of the drivers or where possible used an in-tree
>>>>>       alternative (e.g. there is SDK code for a ac5-gpio but two
>>>>> instances of
>>>>>       the existing marvell,orion-gpio seems to cover what is needed
>>>>> if you use
>>>>>       an appropriate binding). I expect that there will a new series of
>>>>>       patches when I get some different hardware (or additions to
>>>>> this series
>>>>>       depending on if/when it lands).
>>>>>            Changes in v7:
>>>>>       - Add missing compatible on usb1
>>>>>       - Add "rd-ac5x" compatible for board
>>>>>       - Move aliases to board dts
>>>>>       - Move board specific usb info to board dts
>>>>>       - Consolidate usb1 board settings and remove unnecessary
>>>>> compatible
>>>>>       - Add Allied Telesis copyright
>>>>>       - Rename files after mailng-list discussion
>>>>>       Changes in v6:
>>>>>       - Move CPU nodes above the SoC (Krzysztof)
>>>>>       - Minor formatting clean ups (Krzysztof)
>>>>>       - Run through `make dtbs_check`
>>>>>       - Move gic nodes inside SoC
>>>>>       - Group clocks under a clock node
>>>>>       Changes in v5:
>>>>>       - add #{address,size}-cells property to i2c nodes
>>>>>       - make i2c nodes disabled in the SoC and enable them in the board
>>>>>       - add interrupt controller attributes to gpio nodes
>>>>>       - Move fixed-clock nodes up a level and remove unnecessary @0
>>>>>       Changes in v4:
>>>>>       - use 'phy-handle' instead of 'phy'
>>>>>       - move status="okay" on usb nodes to board dts
>>>>>       - Add review from Andrew
>>>>>       Changes in v3:
>>>>>       - Move memory node to board
>>>>>       - Use single digit reg value for phy address
>>>>>       - Remove MMC node (driver needs work)
>>>>>       - Remove syscon & simple-mfd for pinctrl
>>>>>       Changes in v2:
>>>>>       - Make pinctrl a child node of a syscon node
>>>>>       - Use marvell,armada-8k-gpio instead of orion-gpio
>>>>>       - Remove nand peripheral. The Marvell SDK does have some
>>>>> changes for the
>>>>>         ac5-nand-controller but I currently lack hardware with NAND
>>>>> fitted so
>>>>>         I can't test it right now. I've therefore chosen to omit the
>>>>> node and
>>>>>         not attempted to bring in the driver or binding.
>>>>>       - Remove pcie peripheral. Again there are changes in the SDK
>>>>> and I have
>>>>>         no way of testing them.
>>>>>       - Remove prestera node.
>>>>>       - Remove "marvell,ac5-ehci" compatible from USB node as
>>>>>         "marvell,orion-ehci" is sufficient
>>>>>       - Remove watchdog node. There is a buggy driver for the ac5
>>>>> watchdog in
>>>>>         the SDK but it needs some work so I've dropped the node for now.
>>>>>
>>>>>    arch/arm64/boot/dts/marvell/Makefile          |   1 +
>>>>>    .../boot/dts/marvell/armada-98dx25xx.dtsi     | 295
>>>>> ++++++++++++++++++
>>>>>    .../boot/dts/marvell/armada-98dx35xx-rd.dts   | 101 ++++++
>>>>>    .../boot/dts/marvell/armada-98dx35xx.dtsi     |  13 +
>>>>>    4 files changed, 410 insertions(+)
>>>>>    create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>>>>>    create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx-rd.dts
>>>>>    create mode 100644 arch/arm64/boot/dts/marvell/armada-98dx35xx.dtsi
>>>>>
>>>>> diff --git a/arch/arm64/boot/dts/marvell/Makefile
>>>>> b/arch/arm64/boot/dts/marvell/Makefile
>>>>> index 1c794cdcb8e6..b7a4c715afbb 100644
>>>>> --- a/arch/arm64/boot/dts/marvell/Makefile
>>>>> +++ b/arch/arm64/boot/dts/marvell/Makefile
>>>>> @@ -24,3 +24,4 @@ dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db.dtb
>>>>>    dtb-$(CONFIG_ARCH_MVEBU) += cn9132-db-B.dtb
>>>>>    dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-A.dtb
>>>>>    dtb-$(CONFIG_ARCH_MVEBU) += cn9130-crb-B.dtb
>>>>> +dtb-$(CONFIG_ARCH_MVEBU) += armada-98dx35xx-rd.dtb
>>>>> diff --git a/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>>>>> b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>>>>> new file mode 100644
>>>>> index 000000000000..55ab4cd843a9
>>>>> --- /dev/null
>>>>> +++ b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
>>>>> @@ -0,0 +1,295 @@
>>>>> +// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
>>>>> +/*
>>>>> + * Device Tree For AC5.
>>>>> + *
>>>>> + * Copyright (C) 2021 Marvell
>>>>> + * Copyright (C) 2022 Allied Telesis Labs
>>>>> + */
>>>>> +
>>>>> +#include <dt-bindings/gpio/gpio.h>
>>>>> +#include <dt-bindings/interrupt-controller/arm-gic.h>
>>>>> +
>>>>> +/ {
>>>>> +    model = "Marvell AC5 SoC";
>>>>> +    compatible = "marvell,ac5";
>>>>> +    interrupt-parent = <&gic>;
>>>>> +    #address-cells = <2>;
>>>>> +    #size-cells = <2>;
>>>>> +
>>>>> +    cpus {
>>>>> +        #address-cells = <2>;
>>>>> +        #size-cells = <0>;
>>>>> +
>>>>> +        cpu-map {
>>>>> +            cluster0 {
>>>>> +                core0 {
>>>>> +                    cpu = <&cpu0>;
>>>>> +                };
>>>>> +                core1 {
>>>>> +                    cpu = <&cpu1>;
>>>>> +                };
>>>>> +            };
>>>>> +        };
>>>>> +
>>>>> +        cpu0: cpu@0 {
>>>>> +            device_type = "cpu";
>>>>> +            compatible = "arm,armv8";
>>>>> +            reg = <0x0 0x0>;
>>>>> +            enable-method = "psci";
>>>>> +            next-level-cache = <&l2>;
>>>>> +        };
>>>>> +
>>>>> +        cpu1: cpu@1 {
>>>>> +            device_type = "cpu";
>>>>> +            compatible = "arm,armv8";
>>>>> +            reg = <0x0 0x100>;
>>>>> +            enable-method = "psci";
>>>>> +            next-level-cache = <&l2>;
>>>>> +        };
>>>>> +
>>>>> +        l2: l2-cache {
>>>>> +            compatible = "cache";
>>>>> +        };
>>>>> +    };
>>>>> +
>>>>> +
>>>>> +    psci {
>>>>> +        compatible = "arm,psci-0.2";
>>>>> +        method = "smc";
>>>>> +    };
>>>>> +
>>>>> +    timer {
>>>>> +        compatible = "arm,armv8-timer";
>>>>> +        interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>,
>>>>> +                 <GIC_PPI 8 IRQ_TYPE_LEVEL_HIGH>,
>>>>> +                 <GIC_PPI 10 IRQ_TYPE_LEVEL_HIGH>,
>>>>> +                 <GIC_PPI 7 IRQ_TYPE_LEVEL_HIGH>;
>>>>> +        clock-frequency = <25000000>;
>>>> I said no to this hack in a past version of this patch, and I'm going
>>>> to say it *again*.
>>> Sorry I must have missed it.
>>>> Please fix your firmware to program CNTFRQ_EL0, and
>>>> remove this useless property.
>>> I'm kind of at the mercy of what Marvell have provided for ATF. I am
>>> working on the bootloader portion in parallel and am getting things
>>> ready for submitting the u-boot support upstream. I was hoping to
>>> leave ATF alone I can at least see if they haven't fixed this already
>>> (the original dtsi I started with was fairly old) and if they haven't
>>> I'll raise it via their support system.
>> Seems to work fine without the clock so I'll drop it.
> Thanks. If you can, please verify that this is set on both CPUs (I
> have seen plenty of firmware only setting it on CPU0 in the past).
The arch_timer interrupts are counting up on both CPUs and things 
generally seem to be getting scheduled (I don't have much of a userland 
at the moment so it's not exactly a stress test). Do you think that is 
sufficient to say the clock property is unnecessary and whatever 
firmware I have is working as expected.
>>>> You are also missing a PPI for the EL2 virtual timer which is present
>>>> on any ARMv8.1+ CPU (and since this system is using A55, it definitely
>>>> has it).
>>>>
>>>> [...]
>>> Will add.
>> I assume you're talking about the 5th PPI per the
>> timer/arm,arch_timer.yaml ("hypervisor virtual timer irq").
> Indeed.
>
>> Helpfully
>> Marvell don't include the PPI interrupt numbers in their datasheet. But
>> then I also notice that none of the other boards that have a
>> "arm,armv8-timer" provide a 5th interrupt either, have I misunderstood
>> something?
> This was only recently added to the DT binding, but the interrupt
> definitely exist at the CPU level for anything that implements ARMv8.1
> and up. AFAIK, the M1 is the only machine to expose this interrupt in
> DT, but this doesn't mean the interrupt doesn't exist on all the other
> systems that have the same architecture revision.
>
> If you have contacts in Marvell, maybe try and find out whether they
> have simply decided not to wire the interrupt (I wouldn't be
> surprised). In this case, please add a comment.

I've reached out via their customer support portal. So far they just 
want to know why I'm refusing to use their out of date SDK (maybe I 
should direct them at some of Jon Corbet's presentations :P).

These integrated chips are sometimes a bit problematic because the 
support goes via the Switching group but these questions are really 
about IP blocks that have been taken from the SoC group. It may take a 
while before I get a response from someone that actually knows the 
internals.

>
> Thanks,
>
> 	M.
>

Re: [PATCH v7 2/3] arm64: dts: marvell: Add Armada 98DX2530 SoC and RD-AC5X board

[Marc Zyngier]

On Mon, 16 May 2022 22:56:44 +0100,
Chris Packham <Chris.Packham@alliedtelesis.co.nz> wrote:
> 
> >>>> Please fix your firmware to program CNTFRQ_EL0, and
> >>>> remove this useless property.
> >>> I'm kind of at the mercy of what Marvell have provided for ATF. I am
> >>> working on the bootloader portion in parallel and am getting things
> >>> ready for submitting the u-boot support upstream. I was hoping to
> >>> leave ATF alone I can at least see if they haven't fixed this already
> >>> (the original dtsi I started with was fairly old) and if they haven't
> >>> I'll raise it via their support system.
> >> Seems to work fine without the clock so I'll drop it.
> > Thanks. If you can, please verify that this is set on both CPUs (I
> > have seen plenty of firmware only setting it on CPU0 in the past).
> The arch_timer interrupts are counting up on both CPUs and things 
> generally seem to be getting scheduled (I don't have much of a userland 
> at the moment so it's not exactly a stress test). Do you think that is 
> sufficient to say the clock property is unnecessary and whatever 
> firmware I have is working as expected.

No, the counter always count, and CNTFRQ_EL0 is only an indication of
the frequency for SW to find out. You can directly read CNTFRQ_EL0
from userspace on each CPU and find whether they have the same value.

> >>>> You are also missing a PPI for the EL2 virtual timer which is present
> >>>> on any ARMv8.1+ CPU (and since this system is using A55, it definitely
> >>>> has it).
> >>>>
> >>>> [...]
> >>> Will add.
> >> I assume you're talking about the 5th PPI per the
> >> timer/arm,arch_timer.yaml ("hypervisor virtual timer irq").
> > Indeed.
> >
> >> Helpfully
> >> Marvell don't include the PPI interrupt numbers in their datasheet. But
> >> then I also notice that none of the other boards that have a
> >> "arm,armv8-timer" provide a 5th interrupt either, have I misunderstood
> >> something?
> > This was only recently added to the DT binding, but the interrupt
> > definitely exist at the CPU level for anything that implements ARMv8.1
> > and up. AFAIK, the M1 is the only machine to expose this interrupt in
> > DT, but this doesn't mean the interrupt doesn't exist on all the other
> > systems that have the same architecture revision.
> >
> > If you have contacts in Marvell, maybe try and find out whether they
> > have simply decided not to wire the interrupt (I wouldn't be
> > surprised). In this case, please add a comment.
> 
> I've reached out via their customer support portal. So far they just 
> want to know why I'm refusing to use their out of date SDK (maybe I 
> should direct them at some of Jon Corbet's presentations :P).

The fact that they are asking is already saying everything there is to
know, sadly...

> These integrated chips are sometimes a bit problematic because the 
> support goes via the Switching group but these questions are really 
> about IP blocks that have been taken from the SoC group. It may take a 
> while before I get a response from someone that actually knows the 
> internals.

Fair enough. Until then, please drop a comment in the DT indicating
that the fate of this PPI is unknown. If you eventually find out, just
add it to the DT (it is easy to add things, much harder to remove
them).

Thanks,

	M.

-- 
Without deviation from the norm, progress is not possible.

Re: [PATCH v7 2/3] arm64: dts: marvell: Add Armada 98DX2530 SoC and RD-AC5X board

[Chris Packham]

On 17/05/22 18:42, Marc Zyngier wrote:
> On Mon, 16 May 2022 22:56:44 +0100,
> Chris Packham <Chris.Packham@alliedtelesis.co.nz> wrote:
>>>>>> Please fix your firmware to program CNTFRQ_EL0, and
>>>>>> remove this useless property.
>>>>> I'm kind of at the mercy of what Marvell have provided for ATF. I am
>>>>> working on the bootloader portion in parallel and am getting things
>>>>> ready for submitting the u-boot support upstream. I was hoping to
>>>>> leave ATF alone I can at least see if they haven't fixed this already
>>>>> (the original dtsi I started with was fairly old) and if they haven't
>>>>> I'll raise it via their support system.
>>>> Seems to work fine without the clock so I'll drop it.
>>> Thanks. If you can, please verify that this is set on both CPUs (I
>>> have seen plenty of firmware only setting it on CPU0 in the past).
>> The arch_timer interrupts are counting up on both CPUs and things
>> generally seem to be getting scheduled (I don't have much of a userland
>> at the moment so it's not exactly a stress test). Do you think that is
>> sufficient to say the clock property is unnecessary and whatever
>> firmware I have is working as expected.
> No, the counter always count, and CNTFRQ_EL0 is only an indication of
> the frequency for SW to find out. You can directly read CNTFRQ_EL0
> from userspace on each CPU and find whether they have the same value.

Here's my test program

$ cat CNTFRQ_EL0.c
#include <stdio.h>
#include <stdint.h>
#include <inttypes.h>

int main(int argc, char *argv[])
{
         uint64_t val;

         asm volatile("mrs %0, CNTFRQ_EL0" : "=r" (val));
         printf("CNTFRQ_EL0 = %" PRIu64 "\n", val);

         return 0;
}

And running on the RD-AC5X board

[root@linuxbox tmp]# taskset 0x1 ./CNTFRQ_EL0
CNTFRQ_EL0 = 25000000
[root@linuxbox tmp]# taskset 0x2 ./CNTFRQ_EL0
CNTFRQ_EL0 = 25000000


>
>>>>>> You are also missing a PPI for the EL2 virtual timer which is present
>>>>>> on any ARMv8.1+ CPU (and since this system is using A55, it definitely
>>>>>> has it).
>>>>>>
>>>>>> [...]
>>>>> Will add.
>>>> I assume you're talking about the 5th PPI per the
>>>> timer/arm,arch_timer.yaml ("hypervisor virtual timer irq").
>>> Indeed.
>>>
>>>> Helpfully
>>>> Marvell don't include the PPI interrupt numbers in their datasheet. But
>>>> then I also notice that none of the other boards that have a
>>>> "arm,armv8-timer" provide a 5th interrupt either, have I misunderstood
>>>> something?
>>> This was only recently added to the DT binding, but the interrupt
>>> definitely exist at the CPU level for anything that implements ARMv8.1
>>> and up. AFAIK, the M1 is the only machine to expose this interrupt in
>>> DT, but this doesn't mean the interrupt doesn't exist on all the other
>>> systems that have the same architecture revision.
>>>
>>> If you have contacts in Marvell, maybe try and find out whether they
>>> have simply decided not to wire the interrupt (I wouldn't be
>>> surprised). In this case, please add a comment.
>> I've reached out via their customer support portal. So far they just
>> want to know why I'm refusing to use their out of date SDK (maybe I
>> should direct them at some of Jon Corbet's presentations :P).
> The fact that they are asking is already saying everything there is to
> know, sadly...
>
>> These integrated chips are sometimes a bit problematic because the
>> support goes via the Switching group but these questions are really
>> about IP blocks that have been taken from the SoC group. It may take a
>> while before I get a response from someone that actually knows the
>> internals.
> Fair enough. Until then, please drop a comment in the DT indicating
> that the fate of this PPI is unknown. If you eventually find out, just
> add it to the DT (it is easy to add things, much harder to remove
> them).

I'll include the following in the next round

diff --git a/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi 
b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
index 88edc741c008..7a3693a2ad04 100644
--- a/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
+++ b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
@@ -63,6 +63,7 @@ timer {
                              <GIC_PPI 8 IRQ_TYPE_LEVEL_HIGH>,
                              <GIC_PPI 10 IRQ_TYPE_LEVEL_HIGH>,
                              <GIC_PPI 7 IRQ_TYPE_LEVEL_HIGH>;
+                            /* PPI for EL2 virtual timer is undocumented */
         };

         pmu {

> Thanks,
>
> 	M.
>

Re: [PATCH v7 2/3] arm64: dts: marvell: Add Armada 98DX2530 SoC and RD-AC5X board

[Marc Zyngier]

On 2022-05-17 23:56, Chris Packham wrote:
> On 17/05/22 18:42, Marc Zyngier wrote:
>> On Mon, 16 May 2022 22:56:44 +0100,
>> Chris Packham <Chris.Packham@alliedtelesis.co.nz> wrote:
>>>>>>> Please fix your firmware to program CNTFRQ_EL0, and
>>>>>>> remove this useless property.
>>>>>> I'm kind of at the mercy of what Marvell have provided for ATF. I 
>>>>>> am
>>>>>> working on the bootloader portion in parallel and am getting 
>>>>>> things
>>>>>> ready for submitting the u-boot support upstream. I was hoping to
>>>>>> leave ATF alone I can at least see if they haven't fixed this 
>>>>>> already
>>>>>> (the original dtsi I started with was fairly old) and if they 
>>>>>> haven't
>>>>>> I'll raise it via their support system.
>>>>> Seems to work fine without the clock so I'll drop it.
>>>> Thanks. If you can, please verify that this is set on both CPUs (I
>>>> have seen plenty of firmware only setting it on CPU0 in the past).
>>> The arch_timer interrupts are counting up on both CPUs and things
>>> generally seem to be getting scheduled (I don't have much of a 
>>> userland
>>> at the moment so it's not exactly a stress test). Do you think that 
>>> is
>>> sufficient to say the clock property is unnecessary and whatever
>>> firmware I have is working as expected.
>> No, the counter always count, and CNTFRQ_EL0 is only an indication of
>> the frequency for SW to find out. You can directly read CNTFRQ_EL0
>> from userspace on each CPU and find whether they have the same value.
> 
> Here's my test program
> 
> $ cat CNTFRQ_EL0.c
> #include <stdio.h>
> #include <stdint.h>
> #include <inttypes.h>
> 
> int main(int argc, char *argv[])
> {
>          uint64_t val;
> 
>          asm volatile("mrs %0, CNTFRQ_EL0" : "=r" (val));
>          printf("CNTFRQ_EL0 = %" PRIu64 "\n", val);
> 
>          return 0;
> }
> 
> And running on the RD-AC5X board
> 
> [root@linuxbox tmp]# taskset 0x1 ./CNTFRQ_EL0
> CNTFRQ_EL0 = 25000000
> [root@linuxbox tmp]# taskset 0x2 ./CNTFRQ_EL0
> CNTFRQ_EL0 = 25000000

Great. So the DT attribute was only cargo-culted by MRVL,
and they don't realise what that's for...

Thanks for going the extra mile and checking this!

> 
>> 
>>>>>>> You are also missing a PPI for the EL2 virtual timer which is 
>>>>>>> present
>>>>>>> on any ARMv8.1+ CPU (and since this system is using A55, it 
>>>>>>> definitely
>>>>>>> has it).
>>>>>>> 
>>>>>>> [...]
>>>>>> Will add.
>>>>> I assume you're talking about the 5th PPI per the
>>>>> timer/arm,arch_timer.yaml ("hypervisor virtual timer irq").
>>>> Indeed.
>>>> 
>>>>> Helpfully
>>>>> Marvell don't include the PPI interrupt numbers in their datasheet. 
>>>>> But
>>>>> then I also notice that none of the other boards that have a
>>>>> "arm,armv8-timer" provide a 5th interrupt either, have I 
>>>>> misunderstood
>>>>> something?
>>>> This was only recently added to the DT binding, but the interrupt
>>>> definitely exist at the CPU level for anything that implements 
>>>> ARMv8.1
>>>> and up. AFAIK, the M1 is the only machine to expose this interrupt 
>>>> in
>>>> DT, but this doesn't mean the interrupt doesn't exist on all the 
>>>> other
>>>> systems that have the same architecture revision.
>>>> 
>>>> If you have contacts in Marvell, maybe try and find out whether they
>>>> have simply decided not to wire the interrupt (I wouldn't be
>>>> surprised). In this case, please add a comment.
>>> I've reached out via their customer support portal. So far they just
>>> want to know why I'm refusing to use their out of date SDK (maybe I
>>> should direct them at some of Jon Corbet's presentations :P).
>> The fact that they are asking is already saying everything there is to
>> know, sadly...
>> 
>>> These integrated chips are sometimes a bit problematic because the
>>> support goes via the Switching group but these questions are really
>>> about IP blocks that have been taken from the SoC group. It may take 
>>> a
>>> while before I get a response from someone that actually knows the
>>> internals.
>> Fair enough. Until then, please drop a comment in the DT indicating
>> that the fate of this PPI is unknown. If you eventually find out, just
>> add it to the DT (it is easy to add things, much harder to remove
>> them).
> 
> I'll include the following in the next round
> 
> diff --git a/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
> b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
> index 88edc741c008..7a3693a2ad04 100644
> --- a/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
> +++ b/arch/arm64/boot/dts/marvell/armada-98dx25xx.dtsi
> @@ -63,6 +63,7 @@ timer {
>                               <GIC_PPI 8 IRQ_TYPE_LEVEL_HIGH>,
>                               <GIC_PPI 10 IRQ_TYPE_LEVEL_HIGH>,
>                               <GIC_PPI 7 IRQ_TYPE_LEVEL_HIGH>;
> +                            /* PPI for EL2 virtual timer is 
> undocumented */
>          };
> 
>          pmu {

Looks good, thank you.

         M.
-- 
Jazz is not dead. It just smells funny...