Skip to contentSkip to conversation

arm64: Introduce CIX P1 (SKY1) SoC

PATCH Superseded Peter Chen 3 participants 15 messages Feb 27, 2025 – Mar 4, 2025

Subsystems (4)

v3 of 11 Superseded Latest is v11 →

lore
Share & tools
lore

Peter Chen

author
#0 lore raw

v3 cover arm64: Introduce CIX P1 (SKY1) SoC

Cixtech P1 (internal name sky1) is high performance generic Armv9 SoC. Orion O6 is the world's first open source Arm V9 Motherboard built by Radxa. You could find brief introduction for SoC and related boards at:
https://radxa.com/products/orion/o6#overview

In this series, we add initial SoC and board support for Kernel building. Patch 1-2: Add dt-binding doc for CIX and its sky1 SoC
Patch 3: add related maintainter entry
Patch 4-5: add Arm64 build support
Patch 6: add initial dts support for SoC and Orion O6 board

To run upstream kernel at Orion O6 board, you need to use BIOS released by Radxa:
https://docs.radxa.com/en/orion/o6/bios/install-bios

Changes for v3:
- Patch 1: Add Krzysztof Kozlowski's Acked-by Tag
- Patch 2: Add Krzysztof Kozlowski's Reviewed-by Tag
- Patch 6: Fix two dts coding sytle issues

Changes for v2:
- Pass dts build check with below commands:
make O=$OUTKNL dt_binding_check DT_SCHEMA_FILES=vendor-prefixes.yaml make O=$OUTKNL dt_binding_check DT_SCHEMA_FILES=arm/cix.yaml make O=$OUTKNL CHECK_DTBS=y W=1 cix/sky1-orion-o6.dtb
- Re-order the patch set, and move vendor-perfixes to the 1st patch.
- Patch 4: Ordered Kconfig config entry by alpha-numerically
- Patch 5: Corrects the Ack tag's name
- Patch 6: see below.
1) Corrects the SoF tag's name
2) Fix several coding sytle issues
3) move linux,cma node to dts file
4) delete memory node, memory size is passed by firmware
5) delete uart2 node which will be added in future patches
6) Improve for pmu and cpu node to stands for more specific cpu model
7) Improve the timer node and add hypervisor virtual timer irq

Fugang Duan (1):
arm64: Kconfig: add ARCH_CIX for cix silicons

Peter Chen (5):
dt-bindings: vendor-prefixes: Add CIX Technology Group Co., Ltd.
dt-bindings: arm: add CIX P1 (SKY1) SoC
MAINTAINERS: Add CIX SoC maintainer entry
arm64: defconfig: Enable CIX SoC
arm64: dts: cix: add initial CIX P1(SKY1) dts support

 .../devicetree/bindings/arm/cix.yaml          |  26 +++
 .../devicetree/bindings/vendor-prefixes.yaml  |   2 +
 MAINTAINERS                                   |   8 +
 arch/arm64/Kconfig.platforms                  |   6 +
 arch/arm64/boot/dts/Makefile                  |   1 +
 arch/arm64/boot/dts/cix/Makefile              |   2 +
 arch/arm64/boot/dts/cix/sky1-orion-o6.dts     |  26 +++
 arch/arm64/boot/dts/cix/sky1.dtsi             | 216 ++++++++++++++++++
 arch/arm64/configs/defconfig                  |   1 +
 9 files changed, 288 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/arm/cix.yaml
 create mode 100644 arch/arm64/boot/dts/cix/Makefile
 create mode 100644 arch/arm64/boot/dts/cix/sky1-orion-o6.dts
 create mode 100644 arch/arm64/boot/dts/cix/sky1.dtsi
Signature · 1 line
2.25.1
author
#1 lore raw

PATCH v3 1/6 dt-bindings: vendor-prefixes: Add CIX Technology Group Co., Ltd.

CIX Technology Group Co., Ltd. is a high performance Arm SoC design company. Link: https://www.cixtech.com/.

---
Changes for v3:
- Add Krzysztof Kozlowski's Acked-by Tag

 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index 5079ca6ce1d1..5e76223e4570 100644
@@ -302,6 +302,8 @@ patternProperties:
     description: Cirrus Logic, Inc.
   "^cisco,.*":
     description: Cisco Systems, Inc.
+  "^cix,.*":
+    description: CIX Technology Group Co., Ltd.
   "^clockwork,.*":
     description: Clockwork Tech LLC
   "^cloos,.*":
Signature · 1 line
2.25.1
author
#2 lore raw

PATCH v3 2/6 dt-bindings: arm: add CIX P1 (SKY1) SoC

In reply to Peter Chen

Add device tree bindings for CIX P1 (Internal name sky1) Arm SoC, it consists several SoC models like CP8180, CD8180, etc.

---
Changes for v3:
- Add Krzysztof Kozlowski's Reviewed-by Tag

 .../devicetree/bindings/arm/cix.yaml          | 26 +++++++++++++++++++
 1 file changed, 26 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/arm/cix.yaml
Documentation/devicetree/bindings/arm/cix.yaml new file +26 -0
new file mode 100644
index 000000000000..114dab4bc4d2
@@ -0,0 +1,26 @@
+# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/arm/cix.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: CIX platforms
+
+maintainers:
+  - Peter Chen <[email protected]>
+  - Fugang Duan <[email protected]>
+
+properties:
+  $nodename:
+    const: '/'
+  compatible:
+    oneOf:
+
+      - description: Radxa Orion O6
+        items:
+          - const: radxa,orion-o6
+          - const: cix,sky1
+
+additionalProperties: true
+
+...
Signature · 1 line
2.25.1
author
#3 lore raw

PATCH v3 3/6 MAINTAINERS: Add CIX SoC maintainer entry

In reply to Peter Chen

Using this entry as the maintainers information for CIX SKY series SoCs.

---
 MAINTAINERS | 8 ++++++++
 1 file changed, 8 insertions(+)
MAINTAINERS +8 -0
index efee40ea589f..200d2529c8e5 100644
@@ -2354,6 +2354,14 @@ F:	arch/arm/boot/compressed/misc-ep93xx.h
 F:	arch/arm/mach-ep93xx/
 F:	drivers/iio/adc/ep93xx_adc.c
 
+ARM/CIX SKY ARM ARCHITECTURE
+M:	Peter Chen <[email protected]>
+M:	Fugang Duan <[email protected]>
+R:	CIX Linux Kernel Upstream Group <[email protected]>
+L:	[email protected] (moderated for non-subscribers)
+S:	Maintained
+F:	Documentation/devicetree/bindings/arm/cix.yaml
+
 ARM/CLKDEV SUPPORT
 M:	Russell King <[email protected]>
 L:	[email protected] (moderated for non-subscribers)
Signature · 1 line
2.25.1
#4 lore raw

Re: [PATCH v3 3/6] MAINTAINERS: Add CIX SoC maintainer entry

MAINTAINERS F: arch/arm/boot/compressed/misc-ep93xx.h · line 2363

Quoted 7 lines · Peter Chen
MAINTAINERS +7 -0
+ARM/CIX SKY ARM ARCHITECTURE
+M:	Peter Chen <[email protected]>
+M:	Fugang Duan <[email protected]>
+R:	CIX Linux Kernel Upstream Group <[email protected]>
+L:	[email protected] (moderated for non-subscribers)
+S:	Maintained
+F:	Documentation/devicetree/bindings/arm/cix.yaml

Move the patch to the end of patchset and add here dts.

Best regards,
Krzysztof

author
#5 lore raw

PATCH v3 4/6 arm64: Kconfig: add ARCH_CIX for cix silicons

In reply to Peter Chen

From: Fugang Duan <[email protected]>

Add ARCH_CIX for CIX SoC series support.

---
 arch/arm64/Kconfig.platforms | 6 ++++++
 1 file changed, 6 insertions(+)
arch/arm64/Kconfig.platforms +6 -0
index 02f9248f7c84..abe41db9b9b3 100644
@@ -106,6 +106,12 @@ config ARCH_BLAIZE
 	help
 	  This enables support for the Blaize SoC family
 
+config ARCH_CIX
+	bool "Cixtech SoC family"
+	help
+	  This enables support for the Cixtech SoC family,
+	  like P1(sky1).
+
 config ARCH_EXYNOS
 	bool "Samsung Exynos SoC family"
 	select COMMON_CLK_SAMSUNG
Signature · 1 line
2.25.1
author
#7 lore raw

PATCH v3 5/6 arm64: defconfig: Enable CIX SoC

In reply to Peter Chen

Enable CIX SoC support at ARM64 defconfig

---
 arch/arm64/configs/defconfig | 1 +
 1 file changed, 1 insertion(+)
arch/arm64/configs/defconfig +1 -0
index cb7da4415599..1dd46d200401 100644
@@ -45,6 +45,7 @@ CONFIG_ARCH_BCMBCA=y
 CONFIG_ARCH_BRCMSTB=y
 CONFIG_ARCH_BERLIN=y
 CONFIG_ARCH_BLAIZE=y
+CONFIG_ARCH_CIX=y
 CONFIG_ARCH_EXYNOS=y
 CONFIG_ARCH_SPARX5=y
 CONFIG_ARCH_K3=y
Signature · 1 line
2.25.1
author
#9 lore raw

PATCH v3 6/6 arm64: dts: cix: add initial CIX P1(SKY1) dts support

In reply to Peter Chen

CIX SKY1 SoC is high performance Armv9 SoC designed by Cixtech, and Orion O6 is open source motherboard launched by Radxa. See below for detail:
https://docs.radxa.com/en/orion/o6/getting-started/introduction

In this commit, it only adds limited components for running initramfs at Orion O6.

---
Changes for v3:
- Fix two dts coding sytle issues

 arch/arm64/boot/dts/Makefile              |   1 +
 arch/arm64/boot/dts/cix/Makefile          |   2 +
 arch/arm64/boot/dts/cix/sky1-orion-o6.dts |  26 +++
 arch/arm64/boot/dts/cix/sky1.dtsi         | 216 ++++++++++++++++++++++
 4 files changed, 245 insertions(+)
 create mode 100644 arch/arm64/boot/dts/cix/Makefile
 create mode 100644 arch/arm64/boot/dts/cix/sky1-orion-o6.dts
 create mode 100644 arch/arm64/boot/dts/cix/sky1.dtsi
arch/arm64/boot/dts/Makefile +1 -0
index 79b73a21ddc2..8e7ccd0027bd 100644
@@ -13,6 +13,7 @@ subdir-y += bitmain
 subdir-y += blaize
 subdir-y += broadcom
 subdir-y += cavium
+subdir-y += cix
 subdir-y += exynos
 subdir-y += freescale
 subdir-y += hisilicon
arch/arm64/boot/dts/cix/Makefile new file +2 -0
new file mode 100644
index 000000000000..ed3713982012
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0
+dtb-$(CONFIG_ARCH_CIX) += sky1-orion-o6.dtb
arch/arm64/boot/dts/cix/sky1-orion-o6.dts new file +26 -0
new file mode 100644
index 000000000000..78f4fcd87216
@@ -0,0 +1,26 @@
+// SPDX-License-Identifier: BSD-3-Clause
+/*
+ * Copyright 2025 Cix Technology Group Co., Ltd.
+ *
+ */
+
+/dts-v1/;
+
+#include "sky1.dtsi"
+/ {
+	model = "Radxa Orion O6";
+	compatible = "radxa,orion-o6", "cix,sky1";
+
+	reserved-memory {
+		#address-cells = <2>;
+		#size-cells = <2>;
+		ranges;
+
+		linux,cma {
+			compatible = "shared-dma-pool";
+			reusable;
+			size = <0x0 0x28000000>;
+			linux,cma-default;
+		};
+	};
+};
arch/arm64/boot/dts/cix/sky1.dtsi new file +216 -0

216 lines not on this page · raw message

Signature · 1 line
2.25.1
maintainer Reviewed-by on v3 6/6
#10 lore raw

Re: [PATCH v3 6/6] arm64: dts: cix: add initial CIX P1(SKY1) dts support

Quoted 5 lines · Peter Chen

CIX SKY1 SoC is high performance Armv9 SoC designed by Cixtech, and Orion O6 is open source motherboard launched by Radxa. See below for detail:
https://docs.radxa.com/en/orion/o6/getting-started/introduction

Quoted 8 lines

In this commit, it only adds limited components for running initramfs at Orion O6.

---
Changes for v3:
- Fix two dts coding sytle issues

Best regards,
Krzysztof

#11 lore raw

Re: [PATCH v3 6/6] arm64: dts: cix: add initial CIX P1(SKY1) dts support

In reply to Peter Chen
Quoted 184 lines · Peter Chen

CIX SKY1 SoC is high performance Armv9 SoC designed by Cixtech, and Orion O6 is open source motherboard launched by Radxa. See below for detail:
https://docs.radxa.com/en/orion/o6/getting-started/introduction

In this commit, it only adds limited components for running initramfs at Orion O6.

---
Changes for v3:
- Fix two dts coding sytle issues

 arch/arm64/boot/dts/Makefile              |   1 +
 arch/arm64/boot/dts/cix/Makefile          |   2 +
 arch/arm64/boot/dts/cix/sky1-orion-o6.dts |  26 +++
 arch/arm64/boot/dts/cix/sky1.dtsi         | 216 ++++++++++++++++++++++
 4 files changed, 245 insertions(+)
 create mode 100644 arch/arm64/boot/dts/cix/Makefile
 create mode 100644 arch/arm64/boot/dts/cix/sky1-orion-o6.dts
 create mode 100644 arch/arm64/boot/dts/cix/sky1.dtsi
arch/arm64/boot/dts/Makefile +1 -0
index 79b73a21ddc2..8e7ccd0027bd 100644
@@ -13,6 +13,7 @@ subdir-y += bitmain
 subdir-y += blaize
 subdir-y += broadcom
 subdir-y += cavium
+subdir-y += cix
 subdir-y += exynos
 subdir-y += freescale
 subdir-y += hisilicon
arch/arm64/boot/dts/cix/Makefile new file +2 -0
new file mode 100644
index 000000000000..ed3713982012
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0
+dtb-$(CONFIG_ARCH_CIX) += sky1-orion-o6.dtb
arch/arm64/boot/dts/cix/sky1-orion-o6.dts new file +26 -0
new file mode 100644
index 000000000000..78f4fcd87216
@@ -0,0 +1,26 @@
+// SPDX-License-Identifier: BSD-3-Clause
+/*
+ * Copyright 2025 Cix Technology Group Co., Ltd.
+ *
+ */
+
+/dts-v1/;
+
+#include "sky1.dtsi"
+/ {
+	model = "Radxa Orion O6";
+	compatible = "radxa,orion-o6", "cix,sky1";
+
+	reserved-memory {
+		#address-cells = <2>;
+		#size-cells = <2>;
+		ranges;
+
+		linux,cma {
+			compatible = "shared-dma-pool";
+			reusable;
+			size = <0x0 0x28000000>;
+			linux,cma-default;
+		};
+	};
+};
arch/arm64/boot/dts/cix/sky1.dtsi new file +102 -0
new file mode 100644
index 000000000000..c6d7a48e9893
@@ -0,0 +1,216 @@
+// SPDX-License-Identifier: BSD-3-Clause
+/*
+ * Copyright 2025 Cix Technology Group Co., Ltd.
+ *
+ */
+
+#include <dt-bindings/interrupt-controller/arm-gic.h>
+
+/ {
+	interrupt-parent = <&gic>;
+	#address-cells = <2>;
+	#size-cells = <2>;
+
+	cpus {
+		#address-cells = <2>;
+		#size-cells = <0>;
+
+		cpu0: cpu@0 {
+			compatible = "arm,cortex-a520";
+			enable-method = "psci";
+			reg = <0x0 0x0>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <403>;
+		};
+
+		cpu1: cpu@100 {
+			compatible = "arm,cortex-a520";
+			enable-method = "psci";
+			reg = <0x0 0x100>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <403>;
+		};
+
+		cpu2: cpu@200 {
+			compatible = "arm,cortex-a520";
+			enable-method = "psci";
+			reg = <0x0 0x200>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <403>;
+		};
+
+		cpu3: cpu@300 {
+			compatible = "arm,cortex-a520";
+			enable-method = "psci";
+			reg = <0x0 0x300>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <403>;
+		};
+
+		cpu4: cpu@400 {
+			compatible = "arm,cortex-a720";
+			enable-method = "psci";
+			reg = <0x0 0x400>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <1024>;
+		};
+
+		cpu5: cpu@500 {
+			compatible = "arm,cortex-a720";
+			enable-method = "psci";
+			reg = <0x0 0x500>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <1024>;
+		};
+
+		cpu6: cpu@600 {
+			compatible = "arm,cortex-a720";
+			enable-method = "psci";
+			reg = <0x0 0x600>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <1024>;
+		};
+
+		cpu7: cpu@700 {
+			compatible = "arm,cortex-a720";
+			enable-method = "psci";
+			reg = <0x0 0x700>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <1024>;
+		};
+
+		cpu8: cpu@800 {
+			compatible = "arm,cortex-a720";
+			enable-method = "psci";
+			reg = <0x0 0x800>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <1024>;
+		};
+
+		cpu9: cpu@900 {
+			compatible = "arm,cortex-a720";
+			enable-method = "psci";
+			reg = <0x0 0x900>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <1024>;
+		};
+
+		cpu10: cpu@a00 {
+			compatible = "arm,cortex-a720";
+			enable-method = "psci";
+			reg = <0x0 0xa00>;
+			device_type = "cpu";

arch/arm64/boot/dts/cix/sky1.dtsi line 112

Quoted 10 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +10 -0
+			capacity-dmips-mhz = <1024>;
+		};
+
+		cpu11: cpu@b00 {
+			compatible = "arm,cortex-a720";
+			enable-method = "psci";
+			reg = <0x0 0xb00>;
+			device_type = "cpu";
+			capacity-dmips-mhz = <1024>;
+		};

Given that half the A720s are advertised with lower clock speed, how comes they all have the same capacity?

Quoted 42 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +42 -0
+
+		cpu-map {
+			cluster0 {
+				core0 {
+					cpu = <&cpu0>;
+				};
+				core1 {
+					cpu = <&cpu1>;
+				};
+				core2 {
+					cpu = <&cpu2>;
+				};
+				core3 {
+					cpu = <&cpu3>;
+				};
+				core4 {
+					cpu = <&cpu4>;
+				};
+				core5 {
+					cpu = <&cpu5>;
+				};
+				core6 {
+					cpu = <&cpu6>;
+				};
+				core7 {
+					cpu = <&cpu7>;
+				};
+				core8 {
+					cpu = <&cpu8>;
+				};
+				core9 {
+					cpu = <&cpu9>;
+				};
+				core10 {
+					cpu = <&cpu10>;
+				};
+				core11 {
+					cpu = <&cpu11>;
+				};
+			};
+		};
+	};

arch/arm64/boot/dts/cix/sky1.dtsi line 164

Quoted 10 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +10 -0
+
+	pmu-a520 {
+		compatible = "arm,cortex-a520-pmu";
+		interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW>;
+	};
+
+	pmu-a720 {
+		compatible = "arm,cortex-a720-pmu";
+		interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW>;
+	};

This is wrong. The default configuration for PPIs is to expose the *same* device on all CPUs. You must use PPI affinities for your PMUs. Please see the GICv3 binding for the details.

Quoted 14 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +14 -0
+
+	pmu-spe {
+		compatible = "arm,statistical-profiling-extension-v1";
+		interrupts = <GIC_PPI 5 IRQ_TYPE_LEVEL_LOW>;
+	};
+
+	psci {
+		compatible = "arm,psci-1.0";
+		method = "smc";
+	};
+
+	soc@0 {
+		compatible = "simple-bus";
+		ranges = <0 0 0 0 0x20 0>;

arch/arm64/boot/dts/cix/sky1.dtsi line 188

Quoted 10 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +10 -0
+		dma-ranges;
+		#address-cells = <2>;
+		#size-cells = <2>;
+
+		gic: interrupt-controller@e010000 {
+			compatible = "arm,gic-v3";
+			reg = <0x0 0x0e010000 0 0x10000>,	/* GICD */
+			      <0x0 0x0e090000 0 0x300000>;       /* GICR * 12 */
+			interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_LOW>;
+			#interrupt-cells = <3>;

This will need to be bumped up to 4, and all the interrupt specifiers adjusted.

arch/arm64/boot/dts/cix/sky1.dtsi line 190

Quoted 2 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +2 -0
+			interrupt-controller;
+			#redistributor-regions = <1>;

Drop this, this is useless. It is pretty obvious that there is a single RD region, and 1 is the default.

arch/arm64/boot/dts/cix/sky1.dtsi line 191

Quoted 1 line
arch/arm64/boot/dts/cix/sky1.dtsi new file +1 -0
+			redistributor-stride = <0 0x40000>;

Drop this. This is a standard GIC700 that doesn't need any help computing the stride as it obeys the architecture.

arch/arm64/boot/dts/cix/sky1.dtsi line 193

Quoted 2 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +2 -0
+			#address-cells = <2>;
+			#size-cells = <2>;

I don't understand why you repeat this on every sub-nodes.

Quoted 10 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +10 -0
+			ranges;
+
+			gic_its: msi-controller@e050000 {
+				compatible = "arm,gic-v3-its";
+				reg = <0x0 0x0e050000 0x0 0x30000>;
+				msi-controller;
+				#msi-cells = <1>;
+			};
+		};
+	};

arch/arm64/boot/dts/cix/sky1.dtsi line 213

Quoted 10 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +10 -0
+
+	timer {
+		compatible = "arm,armv8-timer";
+		interrupt-names = "sec-phys", "phys", "virt", "hyp-phys", "hyp-virt";
+		interrupts = <GIC_PPI 13 IRQ_TYPE_LEVEL_LOW>,
+			     <GIC_PPI 14 IRQ_TYPE_LEVEL_LOW>,
+			     <GIC_PPI 11 IRQ_TYPE_LEVEL_LOW>,
+			     <GIC_PPI 10 IRQ_TYPE_LEVEL_LOW>,
+			     <GIC_PPI 12 IRQ_TYPE_LEVEL_LOW>;
+		clock-frequency = <1000000000>;

Drop this. The firmware already sets CNTFRQ_EL0 to the correct value, it seems. And if it doesn't, please fix the firmware.

arch/arm64/boot/dts/cix/sky1.dtsi line 214

Quoted 1 line
arch/arm64/boot/dts/cix/sky1.dtsi new file +1 -0
+		arm,no-tick-in-suspend;

Why do you need this? Is the HW so broken that you have implemented the global counter in a power domain that isn't always on?

As it stands, this DT is completely broken and needs major fixing.

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.
author
#12 lore raw

Re: [PATCH v3 6/6] arm64: dts: cix: add initial CIX P1(SKY1) dts support

On 25-02-28 15:10:24, Marc Zyngier wrote:

Hi Marc,

Thanks for your detail review.

Quoted 11 lines
Quoted 11 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +11 -0
+
+             cpu10: cpu@a00 {
+                     compatible = "arm,cortex-a720";
+                     enable-method = "psci";
+                     reg = <0x0 0xa00>;
+                     device_type = "cpu";
+                     capacity-dmips-mhz = <1024>;
+             };
+
+             cpu11: cpu@b00 {
+                     compatible = "arm,cortex-a720";
Quoted 8 lines

arch/arm64/boot/dts/cix/sky1.dtsi line 112

Quoted 5 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +5 -0
+                     enable-method = "psci";
+                     reg = <0x0 0xb00>;
+                     device_type = "cpu";
+                     capacity-dmips-mhz = <1024>;
+             };

Given that half the A720s are advertised with lower clock speed, how comes they all have the same capacity?

According to Documentation/devicetree/bindings/cpu/cpu-capacity.txt "capacity-dmips-mhz" is u32 value representing CPU capacity expressed in normalized DMIPS/MHz, it means CPU capability per MHz. For sky1 SoC, both middle and big cores are A720, so their capability per MHz are the same.

Quoted 6 lines
Quoted 6 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +6 -0
+
+     pmu-a520 {
+             compatible = "arm,cortex-a520-pmu";
+             interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW>;
+     };
+
Quoted 8 lines

arch/arm64/boot/dts/cix/sky1.dtsi line 164

Quoted 4 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +4 -0
+     pmu-a720 {
+             compatible = "arm,cortex-a720-pmu";
+             interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW>;
+     };

This is wrong. The default configuration for PPIs is to expose the *same* device on all CPUs. You must use PPI affinities for your PMUs. Please see the GICv3 binding for the details.

We have discussed internally, we have not seen the benefits routing different PPI interrupt to dedicated CPUs. Any use cases?

I prefer changing pmu nodes as one generic Armv8 PMU node. Is it accepted? Or must I keep both pmu for A520 and A720, and add PPI affinities to describe hardware well?

Quoted 19 lines
Quoted 19 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +18 -0
+
+     pmu-spe {
+             compatible = "arm,statistical-profiling-extension-v1";
+             interrupts = <GIC_PPI 5 IRQ_TYPE_LEVEL_LOW>;
+     };
+
+     psci {
+             compatible = "arm,psci-1.0";
+             method = "smc";
+     };
+
+     soc@0 {
+             compatible = "simple-bus";
+             ranges = <0 0 0 0 0x20 0>;
+             dma-ranges;
+             #address-cells = <2>;
+             #size-cells = <2>;
+
Quoted 8 lines

arch/arm64/boot/dts/cix/sky1.dtsi line 188

Quoted 6 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +6 -0
+             gic: interrupt-controller@e010000 {
+                     compatible = "arm,gic-v3";
+                     reg = <0x0 0x0e010000 0 0x10000>,       /* GICD */
+                           <0x0 0x0e090000 0 0x300000>;       /* GICR * 12 */
+                     interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_LOW>;
+                     #interrupt-cells = <3>;

This will need to be bumped up to 4, and all the interrupt specifiers adjusted.

Depends on if PPI affinities is must.

Quoted 11 lines

arch/arm64/boot/dts/cix/sky1.dtsi line 190

Quoted 3 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +2 -0
+                     interrupt-controller;
+                     #redistributor-regions = <1>;

Drop this, this is useless. It is pretty obvious that there is a single RD region, and 1 is the default.

Quoted 1 line

+ redistributor-stride = <0 0x40000>;

Drop this. This is a standard GIC700 that doesn't need any help computing the stride as it obeys the architecture.

Will drop above two properties.

Quoted 5 lines

Review

Quoted 3 lines
(quoted diff) +2 -0
+                     #address-cells = <2>;
+                     #size-cells = <2>;

I don't understand why you repeat this on every sub-nodes.

Because there is a child node for gic_its below

Quoted 16 lines
Quoted 16 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +15 -0
+                     ranges;
+
+                     gic_its: msi-controller@e050000 {
+                             compatible = "arm,gic-v3-its";
+                             reg = <0x0 0x0e050000 0x0 0x30000>;
+                             msi-controller;
+                             #msi-cells = <1>;
+                     };
+             };
+     };
+
+     timer {
+             compatible = "arm,armv8-timer";
+             interrupt-names = "sec-phys", "phys", "virt", "hyp-phys", "hyp-virt";
+             interrupts = <GIC_PPI 13 IRQ_TYPE_LEVEL_LOW>,
Quoted 8 lines

arch/arm64/boot/dts/cix/sky1.dtsi line 213

Quoted 5 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +5 -0
+                          <GIC_PPI 14 IRQ_TYPE_LEVEL_LOW>,
+                          <GIC_PPI 11 IRQ_TYPE_LEVEL_LOW>,
+                          <GIC_PPI 10 IRQ_TYPE_LEVEL_LOW>,
+                          <GIC_PPI 12 IRQ_TYPE_LEVEL_LOW>;
+             clock-frequency = <1000000000>;

Drop this. The firmware already sets CNTFRQ_EL0 to the correct value, it seems. And if it doesn't, please fix the firmware.

Yes, you are right, firmware configures it, I will delete it at next version.

Quoted 6 lines
Quoted 2 lines

+ arm,no-tick-in-suspend;

Why do you need this? Is the HW so broken that you have implemented the global counter in a power domain that isn't always on?

Not hardware broken, just arch timer will be powered off at cpu idle and system suspend due to power consumption reason.

Signature · 2 lines
Best regards,
Peter
#13 lore raw

Re: [PATCH v3 6/6] arm64: dts: cix: add initial CIX P1(SKY1) dts support

Quoted 27 lines · Peter Chen

On 25-02-28 15:10:24, Marc Zyngier wrote:

Hi Marc,

Thanks for your detail review.

Quoted 11 lines
Quoted 11 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +11 -0
+
+             cpu10: cpu@a00 {
+                     compatible = "arm,cortex-a720";
+                     enable-method = "psci";
+                     reg = <0x0 0xa00>;
+                     device_type = "cpu";
+                     capacity-dmips-mhz = <1024>;
+             };
+
+             cpu11: cpu@b00 {
+                     compatible = "arm,cortex-a720";
Quoted 8 lines

arch/arm64/boot/dts/cix/sky1.dtsi line 112

Quoted 5 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +5 -0
+                     enable-method = "psci";
+                     reg = <0x0 0xb00>;
+                     device_type = "cpu";
+                     capacity-dmips-mhz = <1024>;
+             };

Given that half the A720s are advertised with lower clock speed, how comes they all have the same capacity?

Quoted 5 lines

According to Documentation/devicetree/bindings/cpu/cpu-capacity.txt "capacity-dmips-mhz" is u32 value representing CPU capacity expressed in normalized DMIPS/MHz, it means CPU capability per MHz. For sky1 SoC, both middle and big cores are A720, so their capability per MHz are the same.

Ah, fair enough. I missed that detail.

Quoted 16 lines
Quoted 15 lines

arch/arm64/boot/dts/cix/sky1.dtsi line 164

Quoted 11 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +10 -0
+
+     pmu-a520 {
+             compatible = "arm,cortex-a520-pmu";
+             interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW>;
+     };
+
+     pmu-a720 {
+             compatible = "arm,cortex-a720-pmu";
+             interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW>;
+     };

This is wrong. The default configuration for PPIs is to expose the *same* device on all CPUs. You must use PPI affinities for your PMUs. Please see the GICv3 binding for the details.

Quoted 2 lines

We have discussed internally, we have not seen the benefits routing different PPI interrupt to dedicated CPUs. Any use cases?

This isn't about changing the PPI. It is about matching CPUs with their PMU. Here, you are saying "both PMU types are connected to all the CPUs using PPI7".

That's obviously not the case.

Quoted 1 line

I prefer changing pmu nodes as one generic Armv8 PMU node. Is it accepted?

No, that's not acceptable.

Quoted 2 lines

Or must I keep both pmu for A520 and A720, and add PPI affinities to describe hardware well?

This is an established practice on all big-little systems: each PMU node has an affinity that indicates which CPUs they are connected to. For GICv3+, this is carried by the interrupt specifier.

Please look at existing SoCs supported, such as rk3399, for example.

Quoted 29 lines
Quoted 20 lines
Quoted 20 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +18 -0
+
+     pmu-spe {
+             compatible = "arm,statistical-profiling-extension-v1";
+             interrupts = <GIC_PPI 5 IRQ_TYPE_LEVEL_LOW>;
+     };
+
+     psci {
+             compatible = "arm,psci-1.0";
+             method = "smc";
+     };
+
+     soc@0 {
+             compatible = "simple-bus";
+             ranges = <0 0 0 0 0x20 0>;
+             dma-ranges;
+             #address-cells = <2>;
+             #size-cells = <2>;
+
Quoted 8 lines

arch/arm64/boot/dts/cix/sky1.dtsi line 188

Quoted 6 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +6 -0
+             gic: interrupt-controller@e010000 {
+                     compatible = "arm,gic-v3";
+                     reg = <0x0 0x0e010000 0 0x10000>,       /* GICD */
+                           <0x0 0x0e090000 0 0x300000>;       /* GICR * 12 */
+                     interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_LOW>;
+                     #interrupt-cells = <3>;

This will need to be bumped up to 4, and all the interrupt specifiers adjusted.

Quoted 1 line

Depends on if PPI affinities is must.

Definitely a must, unless you want to completely remove all traces of the PMU, which is of course silly, but a valid alternative.

[...]

Quoted 8 lines
Quoted 5 lines
Quoted 1 line

+ arm,no-tick-in-suspend;

Why do you need this? Is the HW so broken that you have implemented the global counter in a power domain that isn't always on?

Not hardware broken, just arch timer will be powered off at cpu idle and system suspend due to power consumption reason.

This is not about the timer. This is about the global counter. If your counter stops ticking when you're in idle or suspended, your system is broken and you need this property. If the timer (or more precisely the comparator) is turned off because the CPU is off, then that's the expected behaviour and you don't need this property.

Signature · 1 line
Without deviation from the norm, progress is not possible.
author
#14 lore raw

Re: [PATCH v3 6/6] arm64: dts: cix: add initial CIX P1(SKY1) dts support

Quoted 28 lines · Marc Zyngier
Quoted 15 lines
Quoted 6 lines
Quoted 6 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +6 -0
+
+     pmu-a520 {
+             compatible = "arm,cortex-a520-pmu";
+             interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW>;
+     };
+
Quoted 8 lines

arch/arm64/boot/dts/cix/sky1.dtsi line 164

Quoted 4 lines
arch/arm64/boot/dts/cix/sky1.dtsi new file +4 -0
+     pmu-a720 {
+             compatible = "arm,cortex-a720-pmu";
+             interrupts = <GIC_PPI 7 IRQ_TYPE_LEVEL_LOW>;
+     };

This is wrong. The default configuration for PPIs is to expose the *same* device on all CPUs. You must use PPI affinities for your PMUs. Please see the GICv3 binding for the details.

Quoted 2 lines

We have discussed internally, we have not seen the benefits routing different PPI interrupt to dedicated CPUs. Any use cases?

This isn't about changing the PPI. It is about matching CPUs with their PMU. Here, you are saying "both PMU types are connected to all the CPUs using PPI7".

That's obviously not the case.

Quoted 1 line

I prefer changing pmu nodes as one generic Armv8 PMU node. Is it accepted?

No, that's not acceptable.

Quoted 8 lines
Quoted 2 lines

Or must I keep both pmu for A520 and A720, and add PPI affinities to describe hardware well?

This is an established practice on all big-little systems: each PMU node has an affinity that indicates which CPUs they are connected to. For GICv3+, this is carried by the interrupt specifier.

Please look at existing SoCs supported, such as rk3399, for example.

I see. I will add ppi-partitions for gic-v3 node.

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

This will need to be bumped up to 4, and all the interrupt specifiers adjusted.

Depends on if PPI affinities is must.

Definitely a must, unless you want to completely remove all traces of the PMU, which is of course silly, but a valid alternative.

I will change #interrupt-cells to 4, and applies to all interrupt specifiers.

Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
Quoted 2 lines

+ arm,no-tick-in-suspend;

Why do you need this? Is the HW so broken that you have implemented the global counter in a power domain that isn't always on?

Quoted 9 lines
Quoted 2 lines

Not hardware broken, just arch timer will be powered off at cpu idle and system suspend due to power consumption reason.

This is not about the timer. This is about the global counter. If your counter stops ticking when you're in idle or suspended, your system is broken and you need this property. If the timer (or more precisely the comparator) is turned off because the CPU is off, then that's the expected behaviour and you don't need this property.

I will delete this property.

Signature · 2 lines
Best regards,
Peter