2023-03-28 02:20:40

by Jacky Huang

[permalink] [raw]
Subject: [PATCH v6 00/12] Introduce Nuvoton ma35d1 SoC

From: Jacky Huang <[email protected]>

This patchset adds initial support for the Nuvoton ma35d1 SoC, including
initial device tree, clock driver, reset driver, and serial driver.

This patchset cover letter is based from the initial support for Nuvoton
ma35d1 to keep tracking the version history.

This patchset had been applied to Linux kernel 6.3-rc3 and tested on the
Nuvoton ma35d1 SOM evaluation board.

(ma35d1 information: https://www.nuvoton.com/products/microprocessors/arm-cortex-a35-mpus/)
MA35D1 porting on linux-5.10.y can be found at: https://github.com/OpenNuvoton/MPU-Family

v6:
- Combine nuvoton,ma35d1-clk.yaml and nuvoton,ma35d1-clk.h into one patch
- Combine nuvoton,ma35d1-reset.yaml and nuvoton,ma35d1-reset.h into one patch
- rename Documentation/devicetree/bindings/arm/npcm directory as nuvoton
- Remove patch for adding include/linux/mfd/ma35d1-sys.h as it's not required
- Update dtsi & dts files and move board-specific nodes to dts
- Modify reset driver
- Modify serial driver, fix coding style issues
- Modify clock driver, rewrite the PLL calculation functions

v5:
- Add ARCH_NUVOTON to arm64 Kconfig
- Add ARCH_NUVOTON to defconfig
- Add the clock driver
- Add the reset driver
- Add the serial driver
- Add us to the maintainer

v4:
- patch 4/5 is a resend
- Fixed dt_binding_check errors of nuvoton,ma35d1-clk.yaml
- Modify ma35d1.dtsi
1. Add a node hxt_24m
2. Fixed the base address of gic node
3. Add clocks and clock-names to clock node
- Fixed borad binding mistakes of nuvoton.yaml

v3:
- added patch 4/5 and 5/5
- introduce CONFIG_ARCH_NUVOTON option
- add initial bindings for Nuvoton Platform boards
- fixed coding style problem of nuvoton,ma35d1-clk.h
- added CAPLL to clock-controller node
- modify the chosen node of ma35d1-evb.dts
- modify clock yaml "clk-pll-mode" to "nuvoton,clk-pll-mode"

v2:
- fixed dt_binding_check failed of nuvoton,ma35d1-clk.yaml

Jacky Huang (12):
arm64: Kconfig.platforms: Add config for Nuvoton MA35 platform
arm64: defconfig: Add support for Nuvoton MA35 family SoCs
dt-bindings: clock: nuvoton: add binding for ma35d1 clock controller
dt-bindings: reset: nuvoton: add binding for ma35d1 IP reset control
dt-bindings: mfd: syscon: Add nuvoton,ma35d1-sys compatible
dt-bindings: arm: Add initial bindings for Nuvoton platform
dt-bindings: serial: Document ma35d1 uart controller
arm64: dts: nuvoton: Add initial ma35d1 device tree
clk: nuvoton: Add clock driver for ma35d1 clock controller
reset: Add Nuvoton ma35d1 reset driver support
tty: serial: Add Nuvoton ma35d1 serial driver support
MAINTAINERS: Add entry for NUVOTON MA35

.../bindings/arm/nuvoton/nuvoton,ma35d1.yaml | 30 +
.../nuvoton,npcm-gcr.yaml} | 2 +-
.../npcm.yaml => nuvoton/nuvoton,npcm.yaml} | 2 +-
.../bindings/clock/nuvoton,ma35d1-clk.yaml | 72 ++
.../devicetree/bindings/mfd/syscon.yaml | 1 +
.../bindings/reset/nuvoton,ma35d1-reset.yaml | 44 +
.../serial/nuvoton,ma35d1-serial.yaml | 48 +
MAINTAINERS | 9 +
arch/arm64/Kconfig.platforms | 9 +
arch/arm64/boot/dts/nuvoton/Makefile | 2 +
.../boot/dts/nuvoton/ma35d1-iot-512m.dts | 56 +
.../boot/dts/nuvoton/ma35d1-som-256m.dts | 56 +
arch/arm64/boot/dts/nuvoton/ma35d1.dtsi | 231 +++++
arch/arm64/configs/defconfig | 1 +
drivers/clk/Makefile | 1 +
drivers/clk/nuvoton/Kconfig | 19 +
drivers/clk/nuvoton/Makefile | 4 +
drivers/clk/nuvoton/clk-ma35d1-divider.c | 140 +++
drivers/clk/nuvoton/clk-ma35d1-pll.c | 350 +++++++
drivers/clk/nuvoton/clk-ma35d1.c | 963 ++++++++++++++++++
drivers/clk/nuvoton/clk-ma35d1.h | 123 +++
drivers/reset/Kconfig | 6 +
drivers/reset/Makefile | 1 +
drivers/reset/reset-ma35d1.c | 152 +++
drivers/tty/serial/Kconfig | 18 +
drivers/tty/serial/Makefile | 1 +
drivers/tty/serial/ma35d1_serial.c | 802 +++++++++++++++
.../dt-bindings/clock/nuvoton,ma35d1-clk.h | 253 +++++
.../dt-bindings/reset/nuvoton,ma35d1-reset.h | 108 ++
29 files changed, 3502 insertions(+), 2 deletions(-)
create mode 100644 Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml
rename Documentation/devicetree/bindings/arm/{npcm/nuvoton,gcr.yaml => nuvoton/nuvoton,npcm-gcr.yaml} (93%)
rename Documentation/devicetree/bindings/arm/{npcm/npcm.yaml => nuvoton/nuvoton,npcm.yaml} (93%)
create mode 100644 Documentation/devicetree/bindings/clock/nuvoton,ma35d1-clk.yaml
create mode 100644 Documentation/devicetree/bindings/reset/nuvoton,ma35d1-reset.yaml
create mode 100644 Documentation/devicetree/bindings/serial/nuvoton,ma35d1-serial.yaml
create mode 100644 arch/arm64/boot/dts/nuvoton/ma35d1-iot-512m.dts
create mode 100644 arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts
create mode 100644 arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
create mode 100644 drivers/clk/nuvoton/Kconfig
create mode 100644 drivers/clk/nuvoton/Makefile
create mode 100644 drivers/clk/nuvoton/clk-ma35d1-divider.c
create mode 100644 drivers/clk/nuvoton/clk-ma35d1-pll.c
create mode 100644 drivers/clk/nuvoton/clk-ma35d1.c
create mode 100644 drivers/clk/nuvoton/clk-ma35d1.h
create mode 100644 drivers/reset/reset-ma35d1.c
create mode 100644 drivers/tty/serial/ma35d1_serial.c
create mode 100644 include/dt-bindings/clock/nuvoton,ma35d1-clk.h
create mode 100644 include/dt-bindings/reset/nuvoton,ma35d1-reset.h

--
2.34.1


2023-03-28 02:20:56

by Jacky Huang

[permalink] [raw]
Subject: [PATCH v6 07/12] dt-bindings: serial: Document ma35d1 uart controller

From: Jacky Huang <[email protected]>

Add documentation to describe nuvoton ma35d1 uart driver bindings.

Signed-off-by: Jacky Huang <[email protected]>
---
.../serial/nuvoton,ma35d1-serial.yaml | 48 +++++++++++++++++++
1 file changed, 48 insertions(+)
create mode 100644 Documentation/devicetree/bindings/serial/nuvoton,ma35d1-serial.yaml

diff --git a/Documentation/devicetree/bindings/serial/nuvoton,ma35d1-serial.yaml b/Documentation/devicetree/bindings/serial/nuvoton,ma35d1-serial.yaml
new file mode 100644
index 000000000000..a76af0f6009b
--- /dev/null
+++ b/Documentation/devicetree/bindings/serial/nuvoton,ma35d1-serial.yaml
@@ -0,0 +1,48 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/serial/nuvoton,ma35d1-serial.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Nuvoton MA35D1 Universal Asynchronous Receiver/Transmitter (UART)
+
+maintainers:
+ - Min-Jen Chen <[email protected]>
+ - Jacky Huang <[email protected]>
+
+allOf:
+ - $ref: serial.yaml
+
+properties:
+ compatible:
+ const: nuvoton,ma35d1-uart
+
+ reg:
+ maxItems: 1
+
+ interrupts:
+ maxItems: 1
+
+ clocks:
+ maxItems: 1
+
+required:
+ - compatible
+ - reg
+ - interrupts
+ - clocks
+
+unevaluatedProperties: false
+
+examples:
+ - |
+ #include <dt-bindings/interrupt-controller/arm-gic.h>
+ #include <dt-bindings/clock/nuvoton,ma35d1-clk.h>
+
+ serial@40700000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x40700000 0x100>;
+ interrupts = <GIC_SPI 59 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART0_GATE>;
+ };
+...
--
2.34.1

2023-03-28 02:21:02

by Jacky Huang

[permalink] [raw]
Subject: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform

From: Jacky Huang <[email protected]>

Rename the bindings/arm/npcm directory as nuvoton.
Add binding for ARMv8 based Nuvotn SoCs and platform boards.
Add initial bindings for ma35d1 series development boards.

Signed-off-by: Jacky Huang <[email protected]>
---
.../bindings/arm/nuvoton/nuvoton,ma35d1.yaml | 30 +++++++++++++++++++
.../nuvoton,npcm-gcr.yaml} | 2 +-
.../npcm.yaml => nuvoton/nuvoton,npcm.yaml} | 2 +-
3 files changed, 32 insertions(+), 2 deletions(-)
create mode 100644 Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml
rename Documentation/devicetree/bindings/arm/{npcm/nuvoton,gcr.yaml => nuvoton/nuvoton,npcm-gcr.yaml} (93%)
rename Documentation/devicetree/bindings/arm/{npcm/npcm.yaml => nuvoton/nuvoton,npcm.yaml} (93%)

diff --git a/Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml b/Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml
new file mode 100644
index 000000000000..fb190db61525
--- /dev/null
+++ b/Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml
@@ -0,0 +1,30 @@
+# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/arm/nuvoton/nuvoton,ma35d1.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Nuvoton MA35 series SoC based platforms
+
+maintainers:
+ - Jacky Huang <[email protected]>
+
+description: |
+ Boards with an ARMv8 based Nuvoton MA35 series SoC shall have
+ the following properties.
+
+properties:
+ $nodename:
+ const: '/'
+ compatible:
+ oneOf:
+
+ - description: MA35D1 based boards
+ items:
+ - enum:
+ - nuvoton,ma35d1-iot
+ - nuvoton,ma35d1-som
+ - const: nuvoton,ma35d1
+
+additionalProperties: true
+...
diff --git a/Documentation/devicetree/bindings/arm/npcm/nuvoton,gcr.yaml b/Documentation/devicetree/bindings/arm/nuvoton/nuvoton,npcm-gcr.yaml
similarity index 93%
rename from Documentation/devicetree/bindings/arm/npcm/nuvoton,gcr.yaml
rename to Documentation/devicetree/bindings/arm/nuvoton/nuvoton,npcm-gcr.yaml
index 94e72f25b331..1f354aa900ea 100644
--- a/Documentation/devicetree/bindings/arm/npcm/nuvoton,gcr.yaml
+++ b/Documentation/devicetree/bindings/arm/nuvoton/nuvoton,npcm-gcr.yaml
@@ -1,7 +1,7 @@
# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause
%YAML 1.2
---
-$id: http://devicetree.org/schemas/arm/npcm/nuvoton,gcr.yaml#
+$id: http://devicetree.org/schemas/arm/nuvoton/nuvoton,npcm-gcr.yaml#
$schema: http://devicetree.org/meta-schemas/core.yaml#

title: Global Control Registers block in Nuvoton SoCs
diff --git a/Documentation/devicetree/bindings/arm/npcm/npcm.yaml b/Documentation/devicetree/bindings/arm/nuvoton/nuvoton,npcm.yaml
similarity index 93%
rename from Documentation/devicetree/bindings/arm/npcm/npcm.yaml
rename to Documentation/devicetree/bindings/arm/nuvoton/nuvoton,npcm.yaml
index 6871483947c5..d386744c8815 100644
--- a/Documentation/devicetree/bindings/arm/npcm/npcm.yaml
+++ b/Documentation/devicetree/bindings/arm/nuvoton/nuvoton,npcm.yaml
@@ -1,7 +1,7 @@
# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause
%YAML 1.2
---
-$id: http://devicetree.org/schemas/arm/npcm/npcm.yaml#
+$id: http://devicetree.org/schemas/arm/nuvoton/nuvoton,npcm.yaml#
$schema: http://devicetree.org/meta-schemas/core.yaml#

title: NPCM Platforms
--
2.34.1

2023-03-28 02:21:05

by Jacky Huang

[permalink] [raw]
Subject: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

From: Jacky Huang <[email protected]>

Add initial device tree support for Nuvoton ma35d1 SoC, including
cpu, clock, reset, and serial controllers.
Add reference boards som-256m and iot-512m.

Signed-off-by: Jacky Huang <[email protected]>
---
arch/arm64/boot/dts/nuvoton/Makefile | 2 +
.../boot/dts/nuvoton/ma35d1-iot-512m.dts | 56 +++++
.../boot/dts/nuvoton/ma35d1-som-256m.dts | 56 +++++
arch/arm64/boot/dts/nuvoton/ma35d1.dtsi | 231 ++++++++++++++++++
4 files changed, 345 insertions(+)
create mode 100644 arch/arm64/boot/dts/nuvoton/ma35d1-iot-512m.dts
create mode 100644 arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts
create mode 100644 arch/arm64/boot/dts/nuvoton/ma35d1.dtsi

diff --git a/arch/arm64/boot/dts/nuvoton/Makefile b/arch/arm64/boot/dts/nuvoton/Makefile
index a99dab90472a..c11ab4eac9c7 100644
--- a/arch/arm64/boot/dts/nuvoton/Makefile
+++ b/arch/arm64/boot/dts/nuvoton/Makefile
@@ -1,2 +1,4 @@
# SPDX-License-Identifier: GPL-2.0
dtb-$(CONFIG_ARCH_NPCM) += nuvoton-npcm845-evb.dtb
+dtb-$(CONFIG_ARCH_NUVOTON) += ma35d1-iot-512m.dtb
+dtb-$(CONFIG_ARCH_NUVOTON) += ma35d1-som-256m.dtb
diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1-iot-512m.dts b/arch/arm64/boot/dts/nuvoton/ma35d1-iot-512m.dts
new file mode 100644
index 000000000000..b89e2be6abae
--- /dev/null
+++ b/arch/arm64/boot/dts/nuvoton/ma35d1-iot-512m.dts
@@ -0,0 +1,56 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2023 Nuvoton Technology Corp.
+ * Author: Shan-Chun Hung <[email protected]>
+ * Jacky huang <[email protected]>
+ */
+
+/dts-v1/;
+#include "ma35d1.dtsi"
+
+/ {
+ model = "Nuvoton MA35D1-IoT";
+ compatible = "nuvoton,ma35d1-iot", "nuvoton,ma35d1";
+
+ aliases {
+ serial0 = &uart0;
+ };
+
+ chosen {
+ stdout-path = "serial0:115200n8";
+ };
+
+ mem: memory@80000000 {
+ device_type = "memory";
+ reg = <0x00000000 0x80000000 0 0x20000000>; /* 512M DRAM */
+ };
+
+ clk_hxt: clock-hxt {
+ compatible = "fixed-clock";
+ #clock-cells = <0>;
+ clock-frequency = <24000000>;
+ clock-output-names = "clk_hxt";
+ };
+};
+
+&uart0 {
+ status = "okay";
+};
+
+&clk {
+ assigned-clocks = <&clk CAPLL>,
+ <&clk DDRPLL>,
+ <&clk APLL>,
+ <&clk EPLL>,
+ <&clk VPLL>;
+ assigned-clock-rates = <800000000>,
+ <266000000>,
+ <180000000>,
+ <500000000>,
+ <102000000>;
+ nuvoton,pll-mode = "integer",
+ "fractional",
+ "integer",
+ "integer",
+ "integer";
+};
diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts b/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts
new file mode 100644
index 000000000000..a1ebddecb7f8
--- /dev/null
+++ b/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts
@@ -0,0 +1,56 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2023 Nuvoton Technology Corp.
+ * Author: Shan-Chun Hung <[email protected]>
+ * Jacky huang <[email protected]>
+ */
+
+/dts-v1/;
+#include "ma35d1.dtsi"
+
+/ {
+ model = "Nuvoton MA35D1-SOM";
+ compatible = "nuvoton,ma35d1-som", "nuvoton,ma35d1";
+
+ aliases {
+ serial0 = &uart0;
+ };
+
+ chosen {
+ stdout-path = "serial0:115200n8";
+ };
+
+ mem: memory@80000000 {
+ device_type = "memory";
+ reg = <0x00000000 0x80000000 0 0x10000000>; /* 256M DRAM */
+ };
+
+ clk_hxt: clock-hxt {
+ compatible = "fixed-clock";
+ #clock-cells = <0>;
+ clock-frequency = <24000000>;
+ clock-output-names = "clk_hxt";
+ };
+};
+
+&uart0 {
+ status = "okay";
+};
+
+&clk {
+ assigned-clocks = <&clk CAPLL>,
+ <&clk DDRPLL>,
+ <&clk APLL>,
+ <&clk EPLL>,
+ <&clk VPLL>;
+ assigned-clock-rates = <800000000>,
+ <266000000>,
+ <180000000>,
+ <500000000>,
+ <102000000>;
+ nuvoton,pll-mode = "integer",
+ "fractional",
+ "integer",
+ "integer",
+ "integer";
+};
diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
new file mode 100644
index 000000000000..0740b0b218a7
--- /dev/null
+++ b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
@@ -0,0 +1,231 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2023 Nuvoton Technology Corp.
+ * Author: Shan-Chun Hung <[email protected]>
+ * Jacky huang <[email protected]>
+ */
+
+#include <dt-bindings/interrupt-controller/arm-gic.h>
+#include <dt-bindings/input/input.h>
+#include <dt-bindings/gpio/gpio.h>
+#include <dt-bindings/clock/nuvoton,ma35d1-clk.h>
+#include <dt-bindings/reset/nuvoton,ma35d1-reset.h>
+
+/ {
+ compatible = "nuvoton,ma35d1";
+ interrupt-parent = <&gic>;
+ #address-cells = <2>;
+ #size-cells = <2>;
+
+ cpus {
+ #address-cells = <2>;
+ #size-cells = <0>;
+
+ cpu0: cpu@0 {
+ device_type = "cpu";
+ compatible = "arm,cortex-a35";
+ reg = <0x0 0x0>;
+ enable-method = "psci";
+ next-level-cache = <&L2_0>;
+ };
+
+ cpu1: cpu@1 {
+ device_type = "cpu";
+ compatible = "arm,cortex-a35";
+ reg = <0x0 0x1>;
+ enable-method = "psci";
+ next-level-cache = <&L2_0>;
+ };
+
+ L2_0: l2-cache0 {
+ compatible = "cache";
+ cache-level = <2>;
+ };
+ };
+
+ psci {
+ compatible = "arm,psci-0.2";
+ method = "smc";
+ };
+
+ timer {
+ compatible = "arm,armv8-timer";
+ interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(4) |
+ IRQ_TYPE_LEVEL_LOW)>, /* Physical Secure */
+ <GIC_PPI 14 (GIC_CPU_MASK_SIMPLE(4) |
+ IRQ_TYPE_LEVEL_LOW)>, /* Physical Non-Secure */
+ <GIC_PPI 11 (GIC_CPU_MASK_SIMPLE(4) |
+ IRQ_TYPE_LEVEL_LOW)>, /* Virtual */
+ <GIC_PPI 10 (GIC_CPU_MASK_SIMPLE(4) |
+ IRQ_TYPE_LEVEL_LOW)>; /* Hypervisor */
+ clock-frequency = <12000000>;
+ interrupt-parent = <&gic>;
+ };
+
+ sys: system-management@40460000 {
+ compatible = "nuvoton,ma35d1-sys", "syscon", "simple-mfd";
+ reg = <0x0 0x40460000 0x0 0x200>;
+
+ reset: reset-controller {
+ compatible = "nuvoton,ma35d1-reset";
+ #reset-cells = <1>;
+ };
+ };
+
+ clk: clock-controller@40460200 {
+ compatible = "nuvoton,ma35d1-clk", "syscon";
+ reg = <0x00000000 0x40460200 0x0 0x100>;
+ #clock-cells = <1>;
+ clocks = <&clk_hxt>;
+ nuvoton,sys = <&sys>;
+ };
+
+ gic: interrupt-controller@50801000 {
+ compatible = "arm,gic-400";
+ reg = <0x0 0x50801000 0 0x1000>, /* GICD */
+ <0x0 0x50802000 0 0x2000>, /* GICC */
+ <0x0 0x50804000 0 0x2000>, /* GICH */
+ <0x0 0x50806000 0 0x2000>; /* GICV */
+ #interrupt-cells = <3>;
+ interrupt-parent = <&gic>;
+ interrupt-controller;
+ interrupts = <GIC_PPI 9 (GIC_CPU_MASK_RAW(0x13) |
+ IRQ_TYPE_LEVEL_HIGH)>;
+ };
+
+ uart0:serial@40700000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40700000 0x0 0x100>;
+ interrupts = <GIC_SPI 59 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART0_GATE>;
+ status = "disabled";
+ };
+
+ uart1:serial@40710000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40710000 0x0 0x100>;
+ interrupts = <GIC_SPI 60 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART1_GATE>;
+ status = "disabled";
+ };
+
+ uart2:serial@40720000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40720000 0x0 0x100>;
+ interrupts = <GIC_SPI 61 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART2_GATE>;
+ status = "disabled";
+ };
+
+ uart3:serial@40730000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40730000 0x0 0x100>;
+ interrupts = <GIC_SPI 62 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART3_GATE>;
+ status = "disabled";
+ };
+
+ uart4:serial@40740000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40740000 0x0 0x100>;
+ interrupts = <GIC_SPI 63 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART4_GATE>;
+ status = "disabled";
+ };
+
+ uart5:serial@40750000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40750000 0x0 0x100>;
+ interrupts = <GIC_SPI 64 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART5_GATE>;
+ status = "disabled";
+ };
+
+ uart6:serial@40760000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40760000 0x0 0x100>;
+ interrupts = <GIC_SPI 91 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART6_GATE>;
+ status = "disabled";
+ };
+
+ uart7:serial@40770000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40770000 0x0 0x100>;
+ interrupts = <GIC_SPI 92 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART7_GATE>;
+ status = "disabled";
+ };
+
+ uart8:serial@40780000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40780000 0x0 0x100>;
+ interrupts = <GIC_SPI 93 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART8_GATE>;
+ status = "disabled";
+ };
+
+ uart9:serial@40790000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40790000 0x0 0x100>;
+ interrupts = <GIC_SPI 94 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART9_GATE>;
+ status = "disabled";
+ };
+
+ uart10:serial@407a0000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x407a0000 0x0 0x100>;
+ interrupts = <GIC_SPI 95 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART10_GATE>;
+ status = "disabled";
+ };
+
+ uart11:serial@407b0000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x407b0000 0x0 0x100>;
+ interrupts = <GIC_SPI 96 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART11_GATE>;
+ status = "disabled";
+ };
+
+ uart12:serial@407c0000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x407c0000 0x0 0x100>;
+ interrupts = <GIC_SPI 118 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART12_GATE>;
+ status = "disabled";
+ };
+
+ uart13:serial@407d0000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x407d0000 0x0 0x100>;
+ interrupts = <GIC_SPI 119 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART13_GATE>;
+ status = "disabled";
+ };
+
+ uart14:serial@407e0000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x407e0000 0x0 0x100>;
+ interrupts = <GIC_SPI 120 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART14_GATE>;
+ status = "disabled";
+ };
+
+ uart15:serial@407f0000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x407f0000 0x0 0x100>;
+ interrupts = <GIC_SPI 121 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART15_GATE>;
+ status = "disabled";
+ };
+
+ uart16:serial@40880000 {
+ compatible = "nuvoton,ma35d1-uart";
+ reg = <0x0 0x40880000 0x0 0x100>;
+ interrupts = <GIC_SPI 122 IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&clk UART16_GATE>;
+ status = "disabled";
+ };
+};
--
2.34.1

2023-03-28 02:21:17

by Jacky Huang

[permalink] [raw]
Subject: [PATCH v6 09/12] clk: nuvoton: Add clock driver for ma35d1 clock controller

From: Jacky Huang <[email protected]>

The clock controller generates clocks for the whole chip, including
system clocks and all peripheral clocks. This driver support ma35d1
clock gating, divider, and individual PLL configuration.

There are 6 PLLs in ma35d1 SoC:
- CA-PLL for the two Cortex-A35 CPU clock
- SYS-PLL for system bus, which comes from the companion MCU
and cannot be programmed by clock controller.
- DDR-PLL for DDR
- EPLL for GMAC and GFX, Display, and VDEC IPs.
- VPLL for video output pixel clock
- APLL for SDHC, I2S audio, and other IPs.
CA-PLL has only one operation mode.
DDR-PLL, EPLL, VPLL, and APLL are advanced PLLs which have 3
operation modes: integer mode, fraction mode, and spread specturm mode.

Signed-off-by: Jacky Huang <[email protected]>
---
drivers/clk/Makefile | 1 +
drivers/clk/nuvoton/Kconfig | 19 +
drivers/clk/nuvoton/Makefile | 4 +
drivers/clk/nuvoton/clk-ma35d1-divider.c | 140 ++++
drivers/clk/nuvoton/clk-ma35d1-pll.c | 350 ++++++++
drivers/clk/nuvoton/clk-ma35d1.c | 963 +++++++++++++++++++++++
drivers/clk/nuvoton/clk-ma35d1.h | 123 +++
7 files changed, 1600 insertions(+)
create mode 100644 drivers/clk/nuvoton/Kconfig
create mode 100644 drivers/clk/nuvoton/Makefile
create mode 100644 drivers/clk/nuvoton/clk-ma35d1-divider.c
create mode 100644 drivers/clk/nuvoton/clk-ma35d1-pll.c
create mode 100644 drivers/clk/nuvoton/clk-ma35d1.c
create mode 100644 drivers/clk/nuvoton/clk-ma35d1.h

diff --git a/drivers/clk/Makefile b/drivers/clk/Makefile
index e3ca0d058a25..2e7916d269e1 100644
--- a/drivers/clk/Makefile
+++ b/drivers/clk/Makefile
@@ -103,6 +103,7 @@ endif
obj-y += mstar/
obj-y += mvebu/
obj-$(CONFIG_ARCH_MXS) += mxs/
+obj-$(CONFIG_ARCH_NUVOTON) += nuvoton/
obj-$(CONFIG_COMMON_CLK_NXP) += nxp/
obj-$(CONFIG_COMMON_CLK_PISTACHIO) += pistachio/
obj-$(CONFIG_COMMON_CLK_PXA) += pxa/
diff --git a/drivers/clk/nuvoton/Kconfig b/drivers/clk/nuvoton/Kconfig
new file mode 100644
index 000000000000..c1324efedcb9
--- /dev/null
+++ b/drivers/clk/nuvoton/Kconfig
@@ -0,0 +1,19 @@
+# SPDX-License-Identifier: GPL-2.0
+# common clock support for Nuvoton SoC family.
+
+config COMMON_CLK_NUVOTON
+ bool "Nuvoton clock controller common support"
+ depends on ARCH_NUVOTON
+ default ARCH_NUVOTON
+ help
+ Say y here to enable common clock controller for Nuvoton platforms.
+
+if COMMON_CLK_NUVOTON
+config CLK_MA35D1
+ bool "Nuvoton MA35D1 clock controller support"
+ depends on ARM64 || COMPILE_TEST
+ default y
+ help
+ Build the driver for MA35D1 Clock Driver.
+
+endif
diff --git a/drivers/clk/nuvoton/Makefile b/drivers/clk/nuvoton/Makefile
new file mode 100644
index 000000000000..d2c092541b8d
--- /dev/null
+++ b/drivers/clk/nuvoton/Makefile
@@ -0,0 +1,4 @@
+# SPDX-License-Identifier: GPL-2.0-only
+obj-$(CONFIG_ARCH_NUVOTON) += clk-ma35d1.o
+obj-$(CONFIG_ARCH_NUVOTON) += clk-ma35d1-divider.o
+obj-$(CONFIG_ARCH_NUVOTON) += clk-ma35d1-pll.o
diff --git a/drivers/clk/nuvoton/clk-ma35d1-divider.c b/drivers/clk/nuvoton/clk-ma35d1-divider.c
new file mode 100644
index 000000000000..340a889a1417
--- /dev/null
+++ b/drivers/clk/nuvoton/clk-ma35d1-divider.c
@@ -0,0 +1,140 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2023 Nuvoton Technology Corp.
+ * Author: Chi-Fang Li <[email protected]>
+ */
+
+#include <linux/clk-provider.h>
+#include <linux/device.h>
+#include <linux/regmap.h>
+#include <linux/spinlock.h>
+
+#include "clk-ma35d1.h"
+
+#define clk_div_mask(width) ((1 << (width)) - 1)
+
+struct ma35d1_adc_clk_div {
+ struct clk_hw hw;
+ void __iomem *reg;
+ u8 shift;
+ u8 width;
+ u32 mask;
+ const struct clk_div_table *table;
+ /* protects concurrent access to clock divider registers */
+ spinlock_t *lock;
+};
+
+static inline struct ma35d1_adc_clk_div *to_ma35d1_adc_clk_div(struct clk_hw *_hw)
+{
+ return container_of(_hw, struct ma35d1_adc_clk_div, hw);
+}
+
+static inline unsigned long ma35d1_clkdiv_recalc_rate(struct clk_hw *hw,
+ unsigned long parent_rate)
+{
+ unsigned int val;
+ struct ma35d1_adc_clk_div *dclk = to_ma35d1_adc_clk_div(hw);
+
+ val = readl_relaxed(dclk->reg) >> dclk->shift;
+ val &= clk_div_mask(dclk->width);
+ val += 1;
+ return divider_recalc_rate(hw, parent_rate, val, dclk->table,
+ CLK_DIVIDER_ROUND_CLOSEST, dclk->width);
+}
+
+static inline long ma35d1_clkdiv_round_rate(struct clk_hw *hw,
+ unsigned long rate,
+ unsigned long *prate)
+{
+ struct ma35d1_adc_clk_div *dclk = to_ma35d1_adc_clk_div(hw);
+
+ return divider_round_rate(hw, rate, prate, dclk->table,
+ dclk->width, CLK_DIVIDER_ROUND_CLOSEST);
+}
+
+static inline int ma35d1_clkdiv_set_rate(struct clk_hw *hw,
+ unsigned long rate,
+ unsigned long parent_rate)
+{
+ int value;
+ unsigned long flags = 0;
+ u32 data;
+ struct ma35d1_adc_clk_div *dclk = to_ma35d1_adc_clk_div(hw);
+
+ value = divider_get_val(rate, parent_rate, dclk->table,
+ dclk->width, CLK_DIVIDER_ROUND_CLOSEST);
+
+ if (dclk->lock)
+ spin_lock_irqsave(dclk->lock, flags);
+
+ data = readl_relaxed(dclk->reg);
+ data &= ~(clk_div_mask(dclk->width) << dclk->shift);
+ data |= (value - 1) << dclk->shift;
+ data |= dclk->mask;
+
+ writel_relaxed(data, dclk->reg);
+
+ if (dclk->lock)
+ spin_unlock_irqrestore(dclk->lock, flags);
+
+ return 0;
+}
+
+static const struct clk_ops ma35d1_adc_clkdiv_ops = {
+ .recalc_rate = ma35d1_clkdiv_recalc_rate,
+ .round_rate = ma35d1_clkdiv_round_rate,
+ .set_rate = ma35d1_clkdiv_set_rate,
+};
+
+struct clk_hw *ma35d1_reg_adc_clkdiv(struct device *dev, const char *name,
+ const char *parent_name,
+ spinlock_t *lock,
+ unsigned long flags, void __iomem *reg,
+ u8 shift, u8 width, u32 mask_bit)
+{
+ struct ma35d1_adc_clk_div *div;
+ struct clk_init_data init;
+ struct clk_div_table *table;
+ u32 max_div, min_div;
+ struct clk_hw *hw;
+ int ret;
+ int i;
+
+ div = devm_kzalloc(dev, sizeof(*div), GFP_KERNEL);
+ if (!div)
+ return ERR_PTR(-ENOMEM);
+
+ max_div = clk_div_mask(width) + 1;
+ min_div = 1;
+
+ table = devm_kcalloc(dev, max_div + 1, sizeof(*table), GFP_KERNEL);
+ if (!table)
+ return ERR_PTR(-ENOMEM);
+
+ for (i = 0; i < max_div; i++) {
+ table[i].val = (min_div + i);
+ table[i].div = 2 * table[i].val;
+ }
+ table[max_div].val = 0;
+ table[max_div].div = 0;
+
+ init.name = name;
+ init.ops = &ma35d1_adc_clkdiv_ops;
+ init.flags |= flags;
+ init.parent_names = parent_name ? &parent_name : NULL;
+ init.num_parents = parent_name ? 1 : 0;
+
+ div->reg = reg;
+ div->shift = shift;
+ div->width = width;
+ div->mask = mask_bit ? BIT(mask_bit) : 0;
+ div->lock = lock;
+ div->hw.init = &init;
+ div->table = table;
+
+ hw = &div->hw;
+ ret = devm_clk_hw_register(dev, hw);
+ if (ret < 0)
+ return ERR_PTR(ret);
+ return hw;
+}
diff --git a/drivers/clk/nuvoton/clk-ma35d1-pll.c b/drivers/clk/nuvoton/clk-ma35d1-pll.c
new file mode 100644
index 000000000000..b36fbda4fa0a
--- /dev/null
+++ b/drivers/clk/nuvoton/clk-ma35d1-pll.c
@@ -0,0 +1,350 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2023 Nuvoton Technology Corp.
+ * Author: Chi-Fang Li <[email protected]>
+ */
+
+#include <linux/bitfield.h>
+#include <linux/clk-provider.h>
+#include <linux/slab.h>
+#include <linux/regmap.h>
+
+#include "clk-ma35d1.h"
+
+#define REG_SYS_RLKTZNS 0x1a4 /* Register Lock Control */
+
+struct ma35d1_clk_pll {
+ struct clk_hw hw;
+ u8 type;
+ u8 mode;
+ unsigned long rate;
+ void __iomem *ctl0_base;
+ void __iomem *ctl1_base;
+ void __iomem *ctl2_base;
+ struct regmap *regmap;
+};
+
+static inline struct ma35d1_clk_pll *to_ma35d1_clk_pll(struct clk_hw *_hw)
+{
+ return container_of(_hw, struct ma35d1_clk_pll, hw);
+}
+
+static void ma35d1_unlock_regs(struct ma35d1_clk_pll *pll)
+{
+ int ret;
+
+ do {
+ regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x59);
+ regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x16);
+ regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x88);
+ regmap_read(pll->regmap, REG_SYS_RLKTZNS, &ret);
+ } while (ret == 0);
+}
+
+static void ma35d1_lock_regs(struct ma35d1_clk_pll *pll)
+{
+ regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x0);
+}
+
+static unsigned long ma35d1_calc_smic_pll_freq(u32 pll0_ctl0,
+ unsigned long parent_rate)
+{
+ u32 m, n, p, outdiv;
+ u64 pll_freq;
+ u32 clk_div_table[] = { 1, 2, 4, 8 };
+
+ if (pll0_ctl0 & SPLL0_CTL0_BP)
+ return parent_rate;
+
+ n = FIELD_GET(SPLL0_CTL0_FBDIV, pll0_ctl0);
+ m = FIELD_GET(SPLL0_CTL0_INDIV, pll0_ctl0);
+ p = FIELD_GET(SPLL0_CTL0_OUTDIV, pll0_ctl0);
+ outdiv = clk_div_table[p];
+ pll_freq = (u64)parent_rate * n;
+ do_div(pll_freq, m * outdiv);
+ return (unsigned long)pll_freq;
+}
+
+static unsigned long ma35d1_calc_pll_freq(u8 mode, u32 *reg_ctl,
+ unsigned long parent_rate)
+{
+ u32 m, n, p;
+ u64 pll_freq, x;
+
+ if (reg_ctl[1] & PLL_CTL1_BP)
+ return parent_rate;
+
+ if (mode == PLL_MODE_INT) {
+ n = FIELD_GET(PLL_CTL0_FBDIV, reg_ctl[0]);
+ m = FIELD_GET(PLL_CTL0_INDIV, reg_ctl[0]);
+ p = FIELD_GET(PLL_CTL1_OUTDIV, reg_ctl[1]);
+ pll_freq = (u64)parent_rate * n;
+ do_div(pll_freq, m * p);
+ } else {
+ n = FIELD_GET(PLL_CTL0_FBDIV, reg_ctl[0]);
+ m = FIELD_GET(PLL_CTL0_INDIV, reg_ctl[0]);
+ p = FIELD_GET(PLL_CTL1_OUTDIV, reg_ctl[1]);
+ x = FIELD_GET(PLL_CTL1_FRAC, reg_ctl[1]);
+ /* 2 decimal places floating to integer (ex. 1.23 to 123) */
+ n = n * 100 + ((x * 100) / FIELD_MAX(PLL_CTL1_FRAC));
+ pll_freq = ((u64)parent_rate * n) / 100 / m / p;
+ }
+ return (unsigned long)pll_freq;
+}
+
+static int ma35d1_pll_find_closest(struct ma35d1_clk_pll *pll,
+ unsigned long rate,
+ unsigned long parent_rate,
+ u32 *reg_ctl, unsigned long *freq)
+{
+ int p, m, n;
+ int fbdiv_min, fbdiv_max;
+ unsigned long diff = 0xffffffff;
+
+ *freq = 0;
+ if ((rate < PLL_FCLKO_MIN_FREQ) || (rate > PLL_FCLKO_MAX_FREQ))
+ return -EINVAL;
+
+ if (pll->mode == PLL_MODE_INT) {
+ fbdiv_min = FBDIV_MIN;
+ fbdiv_max = FBDIV_MAX;
+ } else {
+ fbdiv_min = FBDIV_FRAC_MIN;
+ fbdiv_max = FBDIV_FRAC_MAX;
+ }
+
+ for (m = INDIV_MIN; m <= INDIV_MAX; m++) {
+ for (n = fbdiv_min; n <= fbdiv_max; n++) {
+ for (p = OUTDIV_MIN; p <= OUTDIV_MAX; p++) {
+ unsigned long tmp, fout;
+ u64 fclk;
+
+ tmp = parent_rate / m;
+ if (tmp < PLL_FREF_M_MIN_FREQ ||
+ tmp > PLL_FREF_M_MAX_FREQ)
+ continue; /* constrain */
+
+ fclk = (u64)parent_rate * n / m;
+ /* for 2 decimal places */
+ if (pll->mode != PLL_MODE_INT)
+ fclk /= 100;
+
+ if (fclk < PLL_FCLK_MIN_FREQ ||
+ fclk > PLL_FCLK_MAX_FREQ)
+ continue; /* constrain */
+
+ fout = (unsigned long)(fclk / p);
+ if (fout < PLL_FCLKO_MIN_FREQ ||
+ fout > PLL_FCLKO_MAX_FREQ)
+ continue; /* constrain */
+
+ if (abs(rate - fout) < diff) {
+ reg_ctl[0] = FIELD_PREP(PLL_CTL0_INDIV, m) |
+ FIELD_PREP(PLL_CTL0_FBDIV, n);
+ reg_ctl[1] = FIELD_PREP(PLL_CTL1_OUTDIV, p);
+ *freq = fout;
+ diff = abs(rate - fout);
+ if (diff == 0)
+ break;
+ }
+ }
+ }
+ }
+ if (*freq == 0)
+ return -EINVAL; /* cannot find even one valid setting */
+ return 0;
+}
+
+static int ma35d1_clk_pll_set_rate(struct clk_hw *hw, unsigned long rate,
+ unsigned long parent_rate)
+{
+ struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
+ u32 reg_ctl[3] = { 0 };
+ unsigned long pll_freq;
+ int ret;
+
+ if ((parent_rate < PLL_FREF_MIN_FREQ) ||
+ (parent_rate > PLL_FREF_MAX_FREQ))
+ return -EINVAL;
+
+ if ((pll->type == MA35D1_CAPLL) || (pll->type == MA35D1_DDRPLL)) {
+ pr_warn("Nuvoton MA35D1 CAPLL/DDRPLL is read only.\n");
+ return -EACCES;
+ }
+
+ ret = ma35d1_pll_find_closest(pll, rate, parent_rate, reg_ctl, &pll_freq);
+ if (ret != 0)
+ return ret;
+ pll->rate = pll_freq;
+
+ switch (pll->mode) {
+ case PLL_MODE_INT:
+ reg_ctl[0] |= FIELD_PREP(PLL_CTL0_MODE, PLL_MODE_INT);
+ break;
+ case PLL_MODE_FRAC:
+ reg_ctl[0] |= FIELD_PREP(PLL_CTL0_MODE, PLL_MODE_FRAC);
+ break;
+ case PLL_MODE_SS:
+ reg_ctl[0] |= FIELD_PREP(PLL_CTL0_MODE, PLL_MODE_SS) |
+ FIELD_PREP(PLL_CTL0_SSRATE, PLL_SS_RATE);
+ reg_ctl[2] = FIELD_PREP(PLL_CTL2_SLOPE, PLL_SLOPE);
+ break;
+ }
+ reg_ctl[1] |= PLL_CTL1_PD;
+
+ ma35d1_unlock_regs(pll);
+ writel_relaxed(reg_ctl[0], pll->ctl0_base);
+ writel_relaxed(reg_ctl[1], pll->ctl1_base);
+ writel_relaxed(reg_ctl[2], pll->ctl2_base);
+ ma35d1_lock_regs(pll);
+ return 0;
+}
+
+static unsigned long ma35d1_clk_pll_recalc_rate(struct clk_hw *hw,
+ unsigned long parent_rate)
+{
+ struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
+ u32 reg_ctl[3];
+ unsigned long pll_freq;
+
+ if ((parent_rate < PLL_FREF_MIN_FREQ) || (parent_rate > PLL_FREF_MAX_FREQ))
+ return 0;
+
+ switch (pll->type) {
+ case MA35D1_CAPLL:
+ reg_ctl[0] = readl_relaxed(pll->ctl0_base);
+ pll_freq = ma35d1_calc_smic_pll_freq(reg_ctl[0], parent_rate);
+ break;
+ case MA35D1_DDRPLL:
+ case MA35D1_APLL:
+ case MA35D1_EPLL:
+ case MA35D1_VPLL:
+ reg_ctl[0] = readl_relaxed(pll->ctl0_base);
+ reg_ctl[1] = readl_relaxed(pll->ctl1_base);
+ pll_freq = ma35d1_calc_pll_freq(pll->mode, reg_ctl, parent_rate);
+ break;
+ }
+ return pll_freq;
+}
+
+static long ma35d1_clk_pll_round_rate(struct clk_hw *hw, unsigned long rate,
+ unsigned long *parent_rate)
+{
+ struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
+ u32 reg_ctl[3] = { 0 };
+ unsigned long pll_freq;
+ long ret;
+
+ if (*parent_rate < PLL_FREF_MIN_FREQ || *parent_rate > PLL_FREF_MAX_FREQ)
+ return -EINVAL;
+
+ ret = ma35d1_pll_find_closest(pll, rate, *parent_rate, reg_ctl, &pll_freq);
+ if (ret != 0)
+ return ret;
+
+ switch (pll->type) {
+ case MA35D1_CAPLL:
+ reg_ctl[0] = readl_relaxed(pll->ctl0_base);
+ pll_freq = ma35d1_calc_smic_pll_freq(reg_ctl[0], *parent_rate);
+ break;
+ case MA35D1_DDRPLL:
+ case MA35D1_APLL:
+ case MA35D1_EPLL:
+ case MA35D1_VPLL:
+ reg_ctl[0] = readl_relaxed(pll->ctl0_base);
+ reg_ctl[1] = readl_relaxed(pll->ctl1_base);
+ pll_freq = ma35d1_calc_pll_freq(pll->mode, reg_ctl, *parent_rate);
+ break;
+ }
+ return pll_freq;
+}
+
+static int ma35d1_clk_pll_is_prepared(struct clk_hw *hw)
+{
+ struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
+ u32 val = readl_relaxed(pll->ctl1_base);
+
+ return val & PLL_CTL1_PD ? 0 : 1;
+}
+
+static int ma35d1_clk_pll_prepare(struct clk_hw *hw)
+{
+ struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
+ u32 val;
+
+ if ((pll->type == MA35D1_CAPLL) || (pll->type == MA35D1_DDRPLL)) {
+ pr_warn("Nuvoton MA35D1 CAPLL/DDRPLL is read only.\n");
+ return -EACCES;
+ }
+
+ ma35d1_unlock_regs(pll);
+ val = readl_relaxed(pll->ctl1_base);
+ val &= ~PLL_CTL1_PD;
+ writel_relaxed(val, pll->ctl1_base);
+ ma35d1_lock_regs(pll);
+ return 0;
+}
+
+static void ma35d1_clk_pll_unprepare(struct clk_hw *hw)
+{
+ struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
+ u32 val;
+
+ if ((pll->type == MA35D1_CAPLL) || (pll->type == MA35D1_DDRPLL)) {
+ pr_warn("Nuvoton MA35D1 CAPLL/DDRPLL is read only.\n");
+ } else {
+ val = readl_relaxed(pll->ctl1_base);
+ val |= PLL_CTL1_PD;
+ writel_relaxed(val, pll->ctl1_base);
+ }
+}
+
+static const struct clk_ops ma35d1_clk_pll_ops = {
+ .is_prepared = ma35d1_clk_pll_is_prepared,
+ .prepare = ma35d1_clk_pll_prepare,
+ .unprepare = ma35d1_clk_pll_unprepare,
+ .set_rate = ma35d1_clk_pll_set_rate,
+ .recalc_rate = ma35d1_clk_pll_recalc_rate,
+ .round_rate = ma35d1_clk_pll_round_rate,
+};
+
+struct clk_hw *ma35d1_reg_clk_pll(enum ma35d1_pll_type type,
+ u8 u8mode, const char *name,
+ const char *parent,
+ unsigned long targetFreq,
+ void __iomem *base,
+ struct regmap *regmap)
+{
+ struct ma35d1_clk_pll *pll;
+ struct clk_hw *hw;
+ struct clk_init_data init;
+ int ret;
+
+ pll = kmalloc(sizeof(*pll), GFP_KERNEL);
+ if (!pll)
+ return ERR_PTR(-ENOMEM);
+
+ pll->type = type;
+ pll->mode = u8mode;
+ pll->rate = targetFreq;
+ pll->ctl0_base = base + REG_PLL_CTL0_OFFSET;
+ pll->ctl1_base = base + REG_PLL_CTL1_OFFSET;
+ pll->ctl2_base = base + REG_PLL_CTL2_OFFSET;
+ pll->regmap = regmap;
+
+ init.name = name;
+ init.flags = 0;
+ init.parent_names = &parent;
+ init.num_parents = 1;
+ init.ops = &ma35d1_clk_pll_ops;
+ pll->hw.init = &init;
+ hw = &pll->hw;
+
+ ret = clk_hw_register(NULL, hw);
+ if (ret) {
+ pr_err("failed to register vsi-pll clock!!!\n");
+ kfree(pll);
+ return ERR_PTR(ret);
+ }
+ return hw;
+}
diff --git a/drivers/clk/nuvoton/clk-ma35d1.c b/drivers/clk/nuvoton/clk-ma35d1.c
new file mode 100644
index 000000000000..e4d3ced396a3
--- /dev/null
+++ b/drivers/clk/nuvoton/clk-ma35d1.c
@@ -0,0 +1,963 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2023 Nuvoton Technology Corp.
+ * Author: Chi-Fang Li <[email protected]>
+ */
+
+#include <linux/clk-provider.h>
+#include <linux/mfd/syscon.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/platform_device.h>
+#include <linux/spinlock.h>
+#include <dt-bindings/clock/nuvoton,ma35d1-clk.h>
+
+#include "clk-ma35d1.h"
+
+static DEFINE_SPINLOCK(ma35d1_lock);
+
+static const char *const ca35clk_sel_clks[] = {
+ "hxt", "capll", "ddrpll", "dummy"
+};
+
+static const char *const sysclk0_sel_clks[] = {
+ "epll_div2", "syspll"
+};
+
+static const char *const sysclk1_sel_clks[] = {
+ "hxt", "syspll"
+};
+
+static const char *const axiclk_sel_clks[] = {
+ "capll_div2", "capll_div4"
+};
+
+static const char *const ccap_sel_clks[] = {
+ "hxt", "vpll", "apll", "syspll"
+};
+
+static const char *const sdh_sel_clks[] = {
+ "syspll", "apll", "dummy", "dummy"
+};
+
+static const char *const dcu_sel_clks[] = {
+ "epll_div2", "syspll"
+};
+
+static const char *const gfx_sel_clks[] = {
+ "epll", "syspll"
+};
+
+static const char *const dbg_sel_clks[] = {
+ "hirc", "syspll"
+};
+
+static const char *const timer0_sel_clks[] = {
+ "hxt", "lxt", "pclk0", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer1_sel_clks[] = {
+ "hxt", "lxt", "pclk0", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer2_sel_clks[] = {
+ "hxt", "lxt", "pclk1", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer3_sel_clks[] = {
+ "hxt", "lxt", "pclk1", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer4_sel_clks[] = {
+ "hxt", "lxt", "pclk2", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer5_sel_clks[] = {
+ "hxt", "lxt", "pclk2", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer6_sel_clks[] = {
+ "hxt", "lxt", "pclk0", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer7_sel_clks[] = {
+ "hxt", "lxt", "pclk0", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer8_sel_clks[] = {
+ "hxt", "lxt", "pclk1", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer9_sel_clks[] = {
+ "hxt", "lxt", "pclk1", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer10_sel_clks[] = {
+ "hxt", "lxt", "pclk2", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const timer11_sel_clks[] = {
+ "hxt", "lxt", "pclk2", "dummy", "dummy", "lirc", "dummy", "hirc"
+};
+
+static const char *const uart_sel_clks[] = {
+ "hxt", "sysclk1_div2", "dummy", "dummy"
+};
+
+static const char *const wdt0_sel_clks[] = {
+ "dummy", "lxt", "pclk3_div4096", "lirc"
+};
+
+static const char *const wdt1_sel_clks[] = {
+ "dummy", "lxt", "pclk3_div4096", "lirc"
+};
+
+static const char *const wdt2_sel_clks[] = {
+ "dummy", "lxt", "pclk4_div4096", "lirc"
+};
+
+static const char *const wwdt0_sel_clks[] = {
+ "dummy", "dummy", "pclk3_div4096", "lirc"
+};
+
+static const char *const wwdt1_sel_clks[] = {
+ "dummy", "dummy", "pclk3_div4096", "lirc"
+};
+
+static const char *const wwdt2_sel_clks[] = {
+ "dummy", "dummy", "pclk4_div4096", "lirc"
+};
+
+static const char *const spi0_sel_clks[] = {
+ "pclk1", "apll", "dummy", "dummy"
+};
+
+static const char *const spi1_sel_clks[] = {
+ "pclk2", "apll", "dummy", "dummy"
+};
+
+static const char *const spi2_sel_clks[] = {
+ "pclk1", "apll", "dummy", "dummy"
+};
+
+static const char *const spi3_sel_clks[] = {
+ "pclk2", "apll", "dummy", "dummy"
+};
+
+static const char *const qspi0_sel_clks[] = {
+ "pclk0", "apll", "dummy", "dummy"
+};
+
+static const char *const qspi1_sel_clks[] = {
+ "pclk0", "apll", "dummy", "dummy"
+};
+
+static const char *const i2s0_sel_clks[] = {
+ "apll", "sysclk1_div2", "dummy", "dummy"
+};
+
+static const char *const i2s1_sel_clks[] = {
+ "apll", "sysclk1_div2", "dummy", "dummy"
+};
+
+static const char *const can_sel_clks[] = {
+ "apll", "vpll"
+};
+
+static const char *const cko_sel_clks[] = {
+ "hxt", "lxt", "hirc", "lirc", "capll_div4", "syspll",
+ "ddrpll", "epll_div2", "apll", "vpll", "dummy", "dummy",
+ "dummy", "dummy", "dummy", "dummy"
+};
+
+static const char *const smc_sel_clks[] = {
+ "hxt", "pclk4"
+};
+
+static const char *const kpi_sel_clks[] = {
+ "hxt", "lxt"
+};
+
+static const struct clk_div_table ip_div_table[] = {
+ {0, 2}, {1, 4}, {2, 6}, {3, 8}, {4, 10},
+ {5, 12}, {6, 14}, {7, 16}, {0, 0},
+};
+
+static const struct clk_div_table eadc_div_table[] = {
+ {0, 2}, {1, 4}, {2, 6}, {3, 8}, {4, 10},
+ {5, 12}, {6, 14}, {7, 16}, {8, 18},
+ {9, 20}, {10, 22}, {11, 24}, {12, 26},
+ {13, 28}, {14, 30}, {15, 32}, {0, 0},
+};
+
+static inline struct clk_hw *ma35d1_clk_fixed(const char *name, int rate)
+{
+ return clk_hw_register_fixed_rate(NULL, name, NULL, 0, rate);
+}
+
+static inline struct clk_hw *ma35d1_clk_mux(const char *name,
+ void __iomem *reg,
+ u8 shift, u8 width,
+ const char *const *parents,
+ int num_parents)
+{
+ return clk_hw_register_mux(NULL, name, parents, num_parents,
+ CLK_SET_RATE_NO_REPARENT, reg, shift,
+ width, 0, &ma35d1_lock);
+}
+
+static inline struct clk_hw *ma35d1_clk_divider(const char *name,
+ const char *parent,
+ void __iomem *reg, u8 shift,
+ u8 width)
+{
+ return clk_hw_register_divider(NULL, name, parent, CLK_SET_RATE_PARENT,
+ reg, shift, width, 0, &ma35d1_lock);
+}
+
+static inline struct clk_hw *ma35d1_clk_divider_pow2(const char *name,
+ const char *parent,
+ void __iomem *reg,
+ u8 shift, u8 width)
+{
+ return clk_hw_register_divider(NULL, name, parent,
+ CLK_DIVIDER_POWER_OF_TWO, reg, shift,
+ width, 0, &ma35d1_lock);
+}
+
+static inline struct clk_hw *ma35d1_clk_divider_table(const char *name,
+ const char *parent,
+ void __iomem *reg,
+ u8 shift, u8 width,
+ const struct clk_div_table *table)
+{
+ return clk_hw_register_divider_table(NULL, name, parent, 0,
+ reg, shift, width, 0, table,
+ &ma35d1_lock);
+}
+
+static inline struct clk_hw *ma35d1_clk_fixed_factor(const char *name,
+ const char *parent,
+ unsigned int mult,
+ unsigned int div)
+{
+ return clk_hw_register_fixed_factor(NULL, name, parent,
+ CLK_SET_RATE_PARENT, mult, div);
+}
+
+static inline struct clk_hw *ma35d1_clk_gate(const char *name,
+ const char *parent,
+ void __iomem *reg, u8 shift)
+{
+ return clk_hw_register_gate(NULL, name, parent, CLK_SET_RATE_PARENT,
+ reg, shift, 0, &ma35d1_lock);
+}
+
+static int ma35d1_get_pll_setting(struct device_node *clk_node,
+ u32 *pllmode, u32 *pllfreq)
+{
+ const char *of_str;
+ int i;
+
+ for (i = 0; i < PLL_MAX_NUM; i++) {
+ if (of_property_read_string_index(clk_node, "nuvoton,pll-mode", i, &of_str))
+ return -EINVAL;
+ if (!strcmp(of_str, "integer"))
+ pllmode[i] = PLL_MODE_INT;
+ else if (!strcmp(of_str, "fractional"))
+ pllmode[i] = PLL_MODE_FRAC;
+ else if (!strcmp(of_str, "spread-spectrum"))
+ pllmode[i] = PLL_MODE_SS;
+ else
+ return -EINVAL;
+ }
+ return of_property_read_u32_array(clk_node, "assigned-clock-rates",
+ pllfreq, PLL_MAX_NUM);
+}
+
+static int ma35d1_clocks_probe(struct platform_device *pdev)
+{
+ struct device_node *clk_node = pdev->dev.of_node;
+ struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ void __iomem *clk_base;
+ struct regmap *regmap;
+ static struct clk_hw **hws;
+ static struct clk_hw_onecell_data *ma35d1_hw_data;
+ u32 pllmode[PLL_MAX_NUM];
+ u32 pllfreq[PLL_MAX_NUM];
+ int ret;
+
+ ma35d1_hw_data = devm_kzalloc(&pdev->dev, struct_size(ma35d1_hw_data,
+ hws, CLK_MAX_IDX), GFP_KERNEL);
+ if (WARN_ON(!ma35d1_hw_data))
+ return -ENOMEM;
+
+ ma35d1_hw_data->num = CLK_MAX_IDX;
+ hws = ma35d1_hw_data->hws;
+
+ clk_base = devm_ioremap_resource(&pdev->dev, res);
+ if (IS_ERR(clk_base))
+ return PTR_ERR(clk_base);
+
+ regmap = syscon_regmap_lookup_by_phandle(pdev->dev.of_node, "nuvoton,sys");
+ if (IS_ERR(regmap))
+ return PTR_ERR(regmap);
+
+ ret = ma35d1_get_pll_setting(clk_node, pllmode, pllfreq);
+ if (ret < 0) {
+ dev_err(&pdev->dev, "Invalid PLL setting!\n");
+ return -EINVAL;
+ }
+
+ hws[HXT] = ma35d1_clk_fixed("hxt", 24000000);
+ hws[HXT_GATE] = ma35d1_clk_gate("hxt_gate", "hxt",
+ clk_base + REG_CLK_PWRCTL, 0);
+ hws[LXT] = ma35d1_clk_fixed("lxt", 32768);
+ hws[LXT_GATE] = ma35d1_clk_gate("lxt_gate", "lxt",
+ clk_base + REG_CLK_PWRCTL, 1);
+ hws[HIRC] = ma35d1_clk_fixed("hirc", 12000000);
+ hws[HIRC_GATE] = ma35d1_clk_gate("hirc_gate", "hirc",
+ clk_base + REG_CLK_PWRCTL, 2);
+ hws[LIRC] = ma35d1_clk_fixed("lirc", 32000);
+ hws[LIRC_GATE] = ma35d1_clk_gate("lirc_gate", "lirc",
+ clk_base + REG_CLK_PWRCTL, 3);
+
+ hws[CAPLL] = ma35d1_reg_clk_pll(MA35D1_CAPLL, pllmode[0], "capll",
+ "hxt", pllfreq[0],
+ clk_base + REG_CLK_PLL0CTL0, regmap);
+ hws[SYSPLL] = ma35d1_clk_fixed("syspll", 180000000);
+
+ hws[DDRPLL] = ma35d1_reg_clk_pll(MA35D1_DDRPLL, pllmode[1],
+ "ddrpll", "hxt", pllfreq[1],
+ clk_base + REG_CLK_PLL2CTL0, regmap);
+ hws[APLL] = ma35d1_reg_clk_pll(MA35D1_APLL, pllmode[2], "apll", "hxt",
+ pllfreq[2], clk_base + REG_CLK_PLL3CTL0, regmap);
+ hws[EPLL] = ma35d1_reg_clk_pll(MA35D1_EPLL, pllmode[3], "epll", "hxt",
+ pllfreq[3], clk_base + REG_CLK_PLL4CTL0, regmap);
+ hws[VPLL] = ma35d1_reg_clk_pll(MA35D1_VPLL, pllmode[4], "vpll", "hxt",
+ pllfreq[4], clk_base + REG_CLK_PLL5CTL0, regmap);
+ hws[EPLL_DIV2] = ma35d1_clk_fixed_factor("epll_div2", "epll", 1, 2);
+ hws[EPLL_DIV4] = ma35d1_clk_fixed_factor("epll_div4", "epll", 1, 4);
+ hws[EPLL_DIV8] = ma35d1_clk_fixed_factor("epll_div8", "epll", 1, 8);
+
+ hws[CA35CLK_MUX] = ma35d1_clk_mux("ca35clk_mux", clk_base + REG_CLK_CLKSEL0, 0, 2,
+ ca35clk_sel_clks, ARRAY_SIZE(ca35clk_sel_clks));
+ hws[AXICLK_DIV2] = ma35d1_clk_fixed_factor("capll_div2", "ca35clk_mux", 1, 2);
+ hws[AXICLK_DIV4] = ma35d1_clk_fixed_factor("capll_div4", "ca35clk_mux", 1, 4);
+ hws[AXICLK_MUX] = ma35d1_clk_mux("axiclk_mux", clk_base + REG_CLK_CLKDIV0, 26, 1,
+ axiclk_sel_clks, ARRAY_SIZE(axiclk_sel_clks));
+
+ hws[SYSCLK0_MUX] = ma35d1_clk_mux("sysclk0_mux", clk_base + REG_CLK_CLKSEL0, 2, 1,
+ sysclk0_sel_clks, ARRAY_SIZE(sysclk0_sel_clks));
+ hws[SYSCLK1_MUX] = ma35d1_clk_mux("sysclk1_mux", clk_base + REG_CLK_CLKSEL0, 4, 1,
+ sysclk1_sel_clks, ARRAY_SIZE(sysclk1_sel_clks));
+ hws[SYSCLK1_DIV2] = ma35d1_clk_fixed_factor("sysclk1_div2", "sysclk1_mux", 1, 2);
+
+ /* HCLK0~3 & PCLK0~4 */
+ hws[HCLK0] = ma35d1_clk_fixed_factor("hclk0", "sysclk1_mux", 1, 1);
+ hws[HCLK1] = ma35d1_clk_fixed_factor("hclk1", "sysclk1_mux", 1, 1);
+ hws[HCLK2] = ma35d1_clk_fixed_factor("hclk2", "sysclk1_mux", 1, 1);
+ hws[PCLK0] = ma35d1_clk_fixed_factor("pclk0", "sysclk1_mux", 1, 1);
+ hws[PCLK1] = ma35d1_clk_fixed_factor("pclk1", "sysclk1_mux", 1, 1);
+ hws[PCLK2] = ma35d1_clk_fixed_factor("pclk2", "sysclk1_mux", 1, 1);
+
+ hws[HCLK3] = ma35d1_clk_fixed_factor("hclk3", "sysclk1_mux", 1, 2);
+ hws[PCLK3] = ma35d1_clk_fixed_factor("pclk3", "sysclk1_mux", 1, 2);
+ hws[PCLK4] = ma35d1_clk_fixed_factor("pclk4", "sysclk1_mux", 1, 2);
+
+ hws[USBPHY0] = ma35d1_clk_fixed("usbphy0", 480000000);
+ hws[USBPHY1] = ma35d1_clk_fixed("usbphy1", 480000000);
+
+ /* DDR */
+ hws[DDR0_GATE] = ma35d1_clk_gate("ddr0_gate", "ddrpll",
+ clk_base + REG_CLK_SYSCLK0, 4);
+ hws[DDR6_GATE] = ma35d1_clk_gate("ddr6_gate", "ddrpll",
+ clk_base + REG_CLK_SYSCLK0, 5);
+
+ hws[CAN0_MUX] = ma35d1_clk_mux("can0_mux", clk_base + REG_CLK_CLKSEL4,
+ 16, 1,
+ can_sel_clks, ARRAY_SIZE(can_sel_clks));
+ hws[CAN0_DIV] = ma35d1_clk_divider_table("can0_div", "can0_mux",
+ clk_base + REG_CLK_CLKDIV0,
+ 0, 3, ip_div_table);
+ hws[CAN0_GATE] = ma35d1_clk_gate("can0_gate", "can0_div",
+ clk_base + REG_CLK_SYSCLK0, 8);
+ hws[CAN1_MUX] = ma35d1_clk_mux("can1_mux", clk_base + REG_CLK_CLKSEL4,
+ 17, 1, can_sel_clks,
+ ARRAY_SIZE(can_sel_clks));
+ hws[CAN1_DIV] = ma35d1_clk_divider_table("can1_div", "can1_mux",
+ clk_base + REG_CLK_CLKDIV0,
+ 4, 3, ip_div_table);
+ hws[CAN1_GATE] = ma35d1_clk_gate("can1_gate", "can1_div",
+ clk_base + REG_CLK_SYSCLK0, 9);
+ hws[CAN2_MUX] = ma35d1_clk_mux("can2_mux", clk_base + REG_CLK_CLKSEL4,
+ 18, 1, can_sel_clks,
+ ARRAY_SIZE(can_sel_clks));
+ hws[CAN2_DIV] = ma35d1_clk_divider_table("can2_div", "can2_mux",
+ clk_base + REG_CLK_CLKDIV0,
+ 8, 3, ip_div_table);
+ hws[CAN2_GATE] = ma35d1_clk_gate("can2_gate", "can2_div",
+ clk_base + REG_CLK_SYSCLK0, 10);
+ hws[CAN3_MUX] = ma35d1_clk_mux("can3_mux", clk_base + REG_CLK_CLKSEL4,
+ 19, 1, can_sel_clks,
+ ARRAY_SIZE(can_sel_clks));
+ hws[CAN3_DIV] = ma35d1_clk_divider_table("can3_div", "can3_mux",
+ clk_base + REG_CLK_CLKDIV0,
+ 12, 3, ip_div_table);
+ hws[CAN3_GATE] = ma35d1_clk_gate("can3_gate", "can3_div",
+ clk_base + REG_CLK_SYSCLK0, 11);
+
+ hws[SDH0_MUX] = ma35d1_clk_mux("sdh0_mux", clk_base + REG_CLK_CLKSEL0,
+ 16, 2, sdh_sel_clks,
+ ARRAY_SIZE(sdh_sel_clks));
+ hws[SDH0_GATE] = ma35d1_clk_gate("sdh0_gate", "sdh0_mux",
+ clk_base + REG_CLK_SYSCLK0, 16);
+ hws[SDH1_MUX] = ma35d1_clk_mux("sdh1_mux", clk_base + REG_CLK_CLKSEL0,
+ 18, 2, sdh_sel_clks,
+ ARRAY_SIZE(sdh_sel_clks));
+ hws[SDH1_GATE] = ma35d1_clk_gate("sdh1_gate", "sdh1_mux",
+ clk_base + REG_CLK_SYSCLK0, 17);
+
+ hws[NAND_GATE] = ma35d1_clk_gate("nand_gate", "hclk1",
+ clk_base + REG_CLK_SYSCLK0, 18);
+
+ hws[USBD_GATE] = ma35d1_clk_gate("usbd_gate", "usbphy0",
+ clk_base + REG_CLK_SYSCLK0, 19);
+ hws[USBH_GATE] = ma35d1_clk_gate("usbh_gate", "usbphy0",
+ clk_base + REG_CLK_SYSCLK0, 20);
+ hws[HUSBH0_GATE] = ma35d1_clk_gate("husbh0_gate", "usbphy0",
+ clk_base + REG_CLK_SYSCLK0, 21);
+ hws[HUSBH1_GATE] = ma35d1_clk_gate("husbh1_gate", "usbphy0",
+ clk_base + REG_CLK_SYSCLK0, 22);
+
+ hws[GFX_MUX] = ma35d1_clk_mux("gfx_mux", clk_base + REG_CLK_CLKSEL0,
+ 26, 1, gfx_sel_clks,
+ ARRAY_SIZE(gfx_sel_clks));
+ hws[GFX_GATE] = ma35d1_clk_gate("gfx_gate", "gfx_mux",
+ clk_base + REG_CLK_SYSCLK0, 24);
+ hws[VC8K_GATE] = ma35d1_clk_gate("vc8k_gate", "sysclk0_mux",
+ clk_base + REG_CLK_SYSCLK0, 25);
+ hws[DCU_MUX] = ma35d1_clk_mux("dcu_mux", clk_base + REG_CLK_CLKSEL0,
+ 24, 1, dcu_sel_clks,
+ ARRAY_SIZE(dcu_sel_clks));
+ hws[DCU_GATE] = ma35d1_clk_gate("dcu_gate", "dcu_mux",
+ clk_base + REG_CLK_SYSCLK0, 26);
+ hws[DCUP_DIV] = ma35d1_clk_divider_table("dcup_div", "vpll",
+ clk_base + REG_CLK_CLKDIV0,
+ 16, 3, ip_div_table);
+
+ hws[EMAC0_GATE] = ma35d1_clk_gate("emac0_gate", "epll_div2",
+ clk_base + REG_CLK_SYSCLK0, 27);
+ hws[EMAC1_GATE] = ma35d1_clk_gate("emac1_gate", "epll_div2",
+ clk_base + REG_CLK_SYSCLK0, 28);
+
+ hws[CCAP0_MUX] = ma35d1_clk_mux("ccap0_mux",
+ clk_base + REG_CLK_CLKSEL0,
+ 12, 1, ccap_sel_clks,
+ ARRAY_SIZE(ccap_sel_clks));
+ hws[CCAP0_DIV] = ma35d1_clk_divider("ccap0_div", "ccap0_mux",
+ clk_base + REG_CLK_CLKDIV1, 8, 4);
+ hws[CCAP0_GATE] = ma35d1_clk_gate("ccap0_gate", "ccap0_div",
+ clk_base + REG_CLK_SYSCLK0, 29);
+ hws[CCAP1_MUX] = ma35d1_clk_mux("ccap1_mux",
+ clk_base + REG_CLK_CLKSEL0,
+ 14, 1, ccap_sel_clks,
+ ARRAY_SIZE(ccap_sel_clks));
+ hws[CCAP1_DIV] = ma35d1_clk_divider("ccap1_div", "ccap1_mux",
+ clk_base + REG_CLK_CLKDIV1,
+ 12, 4);
+ hws[CCAP1_GATE] = ma35d1_clk_gate("ccap1_gate", "ccap1_div",
+ clk_base + REG_CLK_SYSCLK0, 30);
+
+ hws[PDMA0_GATE] = ma35d1_clk_gate("pdma0_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 0);
+ hws[PDMA1_GATE] = ma35d1_clk_gate("pdma1_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 1);
+ hws[PDMA2_GATE] = ma35d1_clk_gate("pdma2_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 2);
+ hws[PDMA3_GATE] = ma35d1_clk_gate("pdma3_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 3);
+
+ hws[WH0_GATE] = ma35d1_clk_gate("wh0_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 4);
+ hws[WH1_GATE] = ma35d1_clk_gate("wh1_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 5);
+
+ hws[HWS_GATE] = ma35d1_clk_gate("hws_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 6);
+
+ hws[EBI_GATE] = ma35d1_clk_gate("ebi_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 7);
+
+ hws[SRAM0_GATE] = ma35d1_clk_gate("sram0_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 8);
+ hws[SRAM1_GATE] = ma35d1_clk_gate("sram1_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 9);
+
+ hws[ROM_GATE] = ma35d1_clk_gate("rom_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 10);
+
+ hws[TRA_GATE] = ma35d1_clk_gate("tra_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 11);
+
+ hws[DBG_MUX] = ma35d1_clk_mux("dbg_mux", clk_base + REG_CLK_CLKSEL0,
+ 27, 1, dbg_sel_clks,
+ ARRAY_SIZE(dbg_sel_clks));
+ hws[DBG_GATE] = ma35d1_clk_gate("dbg_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 12);
+
+ hws[CKO_MUX] = ma35d1_clk_mux("cko_mux", clk_base + REG_CLK_CLKSEL4,
+ 24, 4, cko_sel_clks,
+ ARRAY_SIZE(cko_sel_clks));
+ hws[CKO_DIV] = ma35d1_clk_divider_pow2("cko_div", "cko_mux",
+ clk_base + REG_CLK_CLKOCTL, 0, 4);
+ hws[CKO_GATE] = ma35d1_clk_gate("cko_gate", "cko_div",
+ clk_base + REG_CLK_SYSCLK1, 13);
+
+ hws[GTMR_GATE] = ma35d1_clk_gate("gtmr_gate", "hirc",
+ clk_base + REG_CLK_SYSCLK1, 14);
+
+ hws[GPA_GATE] = ma35d1_clk_gate("gpa_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 16);
+ hws[GPB_GATE] = ma35d1_clk_gate("gpb_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 17);
+ hws[GPC_GATE] = ma35d1_clk_gate("gpc_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 18);
+ hws[GPD_GATE] = ma35d1_clk_gate("gpd_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 19);
+ hws[GPE_GATE] = ma35d1_clk_gate("gpe_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 20);
+ hws[GPF_GATE] = ma35d1_clk_gate("gpf_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 21);
+ hws[GPG_GATE] = ma35d1_clk_gate("gpg_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 22);
+ hws[GPH_GATE] = ma35d1_clk_gate("gph_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 23);
+ hws[GPI_GATE] = ma35d1_clk_gate("gpi_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 24);
+ hws[GPJ_GATE] = ma35d1_clk_gate("gpj_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 25);
+ hws[GPK_GATE] = ma35d1_clk_gate("gpk_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 26);
+ hws[GPL_GATE] = ma35d1_clk_gate("gpl_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 27);
+ hws[GPM_GATE] = ma35d1_clk_gate("gpm_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 28);
+ hws[GPN_GATE] = ma35d1_clk_gate("gpn_gate", "hclk0",
+ clk_base + REG_CLK_SYSCLK1, 29);
+
+ hws[TMR0_MUX] = ma35d1_clk_mux("tmr0_mux", clk_base + REG_CLK_CLKSEL1,
+ 0, 3, timer0_sel_clks,
+ ARRAY_SIZE(timer0_sel_clks));
+ hws[TMR0_GATE] = ma35d1_clk_gate("tmr0_gate", "tmr0_mux",
+ clk_base + REG_CLK_APBCLK0, 0);
+ hws[TMR1_MUX] = ma35d1_clk_mux("tmr1_mux", clk_base + REG_CLK_CLKSEL1,
+ 4, 3, timer1_sel_clks,
+ ARRAY_SIZE(timer1_sel_clks));
+ hws[TMR1_GATE] = ma35d1_clk_gate("tmr1_gate", "tmr1_mux",
+ clk_base + REG_CLK_APBCLK0, 1);
+ hws[TMR2_MUX] = ma35d1_clk_mux("tmr2_mux", clk_base + REG_CLK_CLKSEL1,
+ 8, 3, timer2_sel_clks,
+ ARRAY_SIZE(timer2_sel_clks));
+ hws[TMR2_GATE] = ma35d1_clk_gate("tmr2_gate", "tmr2_mux",
+ clk_base + REG_CLK_APBCLK0, 2);
+ hws[TMR3_MUX] = ma35d1_clk_mux("tmr3_mux", clk_base + REG_CLK_CLKSEL1,
+ 12, 3, timer3_sel_clks,
+ ARRAY_SIZE(timer3_sel_clks));
+ hws[TMR3_GATE] = ma35d1_clk_gate("tmr3_gate", "tmr3_mux",
+ clk_base + REG_CLK_APBCLK0, 3);
+ hws[TMR4_MUX] = ma35d1_clk_mux("tmr4_mux", clk_base + REG_CLK_CLKSEL1,
+ 16, 3, timer4_sel_clks,
+ ARRAY_SIZE(timer4_sel_clks));
+ hws[TMR4_GATE] = ma35d1_clk_gate("tmr4_gate", "tmr4_mux",
+ clk_base + REG_CLK_APBCLK0, 4);
+ hws[TMR5_MUX] = ma35d1_clk_mux("tmr5_mux", clk_base + REG_CLK_CLKSEL1,
+ 20, 3, timer5_sel_clks,
+ ARRAY_SIZE(timer5_sel_clks));
+ hws[TMR5_GATE] = ma35d1_clk_gate("tmr5_gate", "tmr5_mux",
+ clk_base + REG_CLK_APBCLK0, 5);
+ hws[TMR6_MUX] = ma35d1_clk_mux("tmr6_mux", clk_base + REG_CLK_CLKSEL1,
+ 24, 3, timer6_sel_clks,
+ ARRAY_SIZE(timer6_sel_clks));
+ hws[TMR6_GATE] = ma35d1_clk_gate("tmr6_gate", "tmr6_mux",
+ clk_base + REG_CLK_APBCLK0, 6);
+ hws[TMR7_MUX] = ma35d1_clk_mux("tmr7_mux", clk_base + REG_CLK_CLKSEL1,
+ 28, 3, timer7_sel_clks,
+ ARRAY_SIZE(timer7_sel_clks));
+ hws[TMR7_GATE] = ma35d1_clk_gate("tmr7_gate", "tmr7_mux",
+ clk_base + REG_CLK_APBCLK0, 7);
+ hws[TMR8_MUX] = ma35d1_clk_mux("tmr8_mux", clk_base + REG_CLK_CLKSEL2,
+ 0, 3, timer8_sel_clks,
+ ARRAY_SIZE(timer8_sel_clks));
+ hws[TMR8_GATE] = ma35d1_clk_gate("tmr8_gate", "tmr8_mux",
+ clk_base + REG_CLK_APBCLK0, 8);
+ hws[TMR9_MUX] = ma35d1_clk_mux("tmr9_mux", clk_base + REG_CLK_CLKSEL2,
+ 4, 3, timer9_sel_clks,
+ ARRAY_SIZE(timer9_sel_clks));
+ hws[TMR9_GATE] = ma35d1_clk_gate("tmr9_gate", "tmr9_mux",
+ clk_base + REG_CLK_APBCLK0, 9);
+ hws[TMR10_MUX] = ma35d1_clk_mux("tmr10_mux",
+ clk_base + REG_CLK_CLKSEL2,
+ 8, 3, timer10_sel_clks,
+ ARRAY_SIZE(timer10_sel_clks));
+ hws[TMR10_GATE] = ma35d1_clk_gate("tmr10_gate", "tmr10_mux",
+ clk_base + REG_CLK_APBCLK0, 10);
+ hws[TMR11_MUX] = ma35d1_clk_mux("tmr11_mux",
+ clk_base + REG_CLK_CLKSEL2,
+ 12, 3, timer11_sel_clks,
+ ARRAY_SIZE(timer11_sel_clks));
+ hws[TMR11_GATE] = ma35d1_clk_gate("tmr11_gate", "tmr11_mux",
+ clk_base + REG_CLK_APBCLK0, 11);
+
+ hws[UART0_MUX] = ma35d1_clk_mux("uart0_mux",
+ clk_base + REG_CLK_CLKSEL2,
+ 16, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART0_DIV] = ma35d1_clk_divider("uart0_div", "uart0_mux",
+ clk_base + REG_CLK_CLKDIV1,
+ 16, 4);
+ hws[UART0_GATE] = ma35d1_clk_gate("uart0_gate", "uart0_div",
+ clk_base + REG_CLK_APBCLK0, 12);
+ hws[UART1_MUX] = ma35d1_clk_mux("uart1_mux",
+ clk_base + REG_CLK_CLKSEL2,
+ 18, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART1_DIV] = ma35d1_clk_divider("uart1_div", "uart1_mux",
+ clk_base + REG_CLK_CLKDIV1,
+ 20, 4);
+ hws[UART1_GATE] = ma35d1_clk_gate("uart1_gate", "uart1_div",
+ clk_base + REG_CLK_APBCLK0, 13);
+ hws[UART2_MUX] = ma35d1_clk_mux("uart2_mux",
+ clk_base + REG_CLK_CLKSEL2,
+ 20, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART2_DIV] = ma35d1_clk_divider("uart2_div", "uart2_mux",
+ clk_base + REG_CLK_CLKDIV1,
+ 24, 4);
+ hws[UART2_GATE] = ma35d1_clk_gate("uart2_gate", "uart2_div",
+ clk_base + REG_CLK_APBCLK0, 14);
+ hws[UART3_MUX] = ma35d1_clk_mux("uart3_mux",
+ clk_base + REG_CLK_CLKSEL2,
+ 22, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART3_DIV] = ma35d1_clk_divider("uart3_div", "uart3_mux",
+ clk_base + REG_CLK_CLKDIV1,
+ 28, 4);
+ hws[UART3_GATE] = ma35d1_clk_gate("uart3_gate", "uart3_div",
+ clk_base + REG_CLK_APBCLK0, 15);
+ hws[UART4_MUX] = ma35d1_clk_mux("uart4_mux",
+ clk_base + REG_CLK_CLKSEL2,
+ 24, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART4_DIV] = ma35d1_clk_divider("uart4_div", "uart4_mux",
+ clk_base + REG_CLK_CLKDIV2,
+ 0, 4);
+ hws[UART4_GATE] = ma35d1_clk_gate("uart4_gate", "uart4_div",
+ clk_base + REG_CLK_APBCLK0, 16);
+ hws[UART5_MUX] = ma35d1_clk_mux("uart5_mux",
+ clk_base + REG_CLK_CLKSEL2,
+ 26, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART5_DIV] = ma35d1_clk_divider("uart5_div", "uart5_mux",
+ clk_base + REG_CLK_CLKDIV2,
+ 4, 4);
+ hws[UART5_GATE] = ma35d1_clk_gate("uart5_gate", "uart5_div",
+ clk_base + REG_CLK_APBCLK0, 17);
+ hws[UART6_MUX] = ma35d1_clk_mux("uart6_mux",
+ clk_base + REG_CLK_CLKSEL2,
+ 28, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART6_DIV] = ma35d1_clk_divider("uart6_div", "uart6_mux",
+ clk_base + REG_CLK_CLKDIV2,
+ 8, 4);
+ hws[UART6_GATE] = ma35d1_clk_gate("uart6_gate", "uart6_div",
+ clk_base + REG_CLK_APBCLK0, 18);
+ hws[UART7_MUX] = ma35d1_clk_mux("uart7_mux",
+ clk_base + REG_CLK_CLKSEL2,
+ 30, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART7_DIV] = ma35d1_clk_divider("uart7_div", "uart7_mux",
+ clk_base + REG_CLK_CLKDIV2,
+ 12, 4);
+ hws[UART7_GATE] = ma35d1_clk_gate("uart7_gate", "uart7_div",
+ clk_base + REG_CLK_APBCLK0, 19);
+ hws[UART8_MUX] = ma35d1_clk_mux("uart8_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 0, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART8_DIV] = ma35d1_clk_divider("uart8_div", "uart8_mux",
+ clk_base + REG_CLK_CLKDIV2,
+ 16, 4);
+ hws[UART8_GATE] = ma35d1_clk_gate("uart8_gate", "uart8_div",
+ clk_base + REG_CLK_APBCLK0, 20);
+ hws[UART9_MUX] = ma35d1_clk_mux("uart9_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 2, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART9_DIV] = ma35d1_clk_divider("uart9_div", "uart9_mux",
+ clk_base + REG_CLK_CLKDIV2,
+ 20, 4);
+ hws[UART9_GATE] = ma35d1_clk_gate("uart9_gate", "uart9_div",
+ clk_base + REG_CLK_APBCLK0, 21);
+ hws[UART10_MUX] = ma35d1_clk_mux("uart10_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 4, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART10_DIV] = ma35d1_clk_divider("uart10_div", "uart10_mux",
+ clk_base + REG_CLK_CLKDIV2,
+ 24, 4);
+ hws[UART10_GATE] = ma35d1_clk_gate("uart10_gate", "uart10_div",
+ clk_base + REG_CLK_APBCLK0, 22);
+ hws[UART11_MUX] = ma35d1_clk_mux("uart11_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 6, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART11_DIV] = ma35d1_clk_divider("uart11_div", "uart11_mux",
+ clk_base + REG_CLK_CLKDIV2,
+ 28, 4);
+ hws[UART11_GATE] = ma35d1_clk_gate("uart11_gate", "uart11_div",
+ clk_base + REG_CLK_APBCLK0, 23);
+ hws[UART12_MUX] = ma35d1_clk_mux("uart12_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 8, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART12_DIV] = ma35d1_clk_divider("uart12_div", "uart12_mux",
+ clk_base + REG_CLK_CLKDIV3,
+ 0, 4);
+ hws[UART12_GATE] = ma35d1_clk_gate("uart12_gate", "uart12_div",
+ clk_base + REG_CLK_APBCLK0, 24);
+ hws[UART13_MUX] = ma35d1_clk_mux("uart13_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 10, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART13_DIV] = ma35d1_clk_divider("uart13_div", "uart13_mux",
+ clk_base + REG_CLK_CLKDIV3,
+ 4, 4);
+ hws[UART13_GATE] = ma35d1_clk_gate("uart13_gate", "uart13_div",
+ clk_base + REG_CLK_APBCLK0, 25);
+ hws[UART14_MUX] = ma35d1_clk_mux("uart14_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 12, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART14_DIV] = ma35d1_clk_divider("uart14_div", "uart14_mux",
+ clk_base + REG_CLK_CLKDIV3,
+ 8, 4);
+ hws[UART14_GATE] = ma35d1_clk_gate("uart14_gate", "uart14_div",
+ clk_base + REG_CLK_APBCLK0, 26);
+ hws[UART15_MUX] = ma35d1_clk_mux("uart15_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 14, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART15_DIV] = ma35d1_clk_divider("uart15_div", "uart15_mux",
+ clk_base + REG_CLK_CLKDIV3,
+ 12, 4);
+ hws[UART15_GATE] = ma35d1_clk_gate("uart15_gate", "uart15_div",
+ clk_base + REG_CLK_APBCLK0, 27);
+ hws[UART16_MUX] = ma35d1_clk_mux("uart16_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 16, 2, uart_sel_clks,
+ ARRAY_SIZE(uart_sel_clks));
+ hws[UART16_DIV] = ma35d1_clk_divider("uart16_div", "uart16_mux",
+ clk_base + REG_CLK_CLKDIV3,
+ 16, 4);
+ hws[UART16_GATE] = ma35d1_clk_gate("uart16_gate", "uart16_div",
+ clk_base + REG_CLK_APBCLK0, 28);
+
+ hws[RTC_GATE] = ma35d1_clk_gate("rtc_gate", "lxt",
+ clk_base + REG_CLK_APBCLK0, 29);
+ hws[DDR_GATE] = ma35d1_clk_gate("ddr_gate", "ddrpll",
+ clk_base + REG_CLK_APBCLK0, 30);
+
+ hws[KPI_MUX] = ma35d1_clk_mux("kpi_mux", clk_base + REG_CLK_CLKSEL4,
+ 30, 1, kpi_sel_clks,
+ ARRAY_SIZE(kpi_sel_clks));
+ hws[KPI_DIV] = ma35d1_clk_divider("kpi_div", "kpi_mux",
+ clk_base + REG_CLK_CLKDIV4,
+ 24, 8);
+ hws[KPI_GATE] = ma35d1_clk_gate("kpi_gate", "kpi_div",
+ clk_base + REG_CLK_APBCLK0, 31);
+
+ hws[I2C0_GATE] = ma35d1_clk_gate("i2c0_gate", "pclk0",
+ clk_base + REG_CLK_APBCLK1, 0);
+ hws[I2C1_GATE] = ma35d1_clk_gate("i2c1_gate", "pclk1",
+ clk_base + REG_CLK_APBCLK1, 1);
+ hws[I2C2_GATE] = ma35d1_clk_gate("i2c2_gate", "pclk2",
+ clk_base + REG_CLK_APBCLK1, 2);
+ hws[I2C3_GATE] = ma35d1_clk_gate("i2c3_gate", "pclk0",
+ clk_base + REG_CLK_APBCLK1, 3);
+ hws[I2C4_GATE] = ma35d1_clk_gate("i2c4_gate", "pclk1",
+ clk_base + REG_CLK_APBCLK1, 4);
+ hws[I2C5_GATE] = ma35d1_clk_gate("i2c5_gate", "pclk2",
+ clk_base + REG_CLK_APBCLK1, 5);
+
+ hws[QSPI0_MUX] = ma35d1_clk_mux("qspi0_mux",
+ clk_base + REG_CLK_CLKSEL4,
+ 8, 2, qspi0_sel_clks,
+ ARRAY_SIZE(qspi0_sel_clks));
+ hws[QSPI0_GATE] = ma35d1_clk_gate("qspi0_gate", "qspi0_mux",
+ clk_base + REG_CLK_APBCLK1, 6);
+ hws[QSPI1_MUX] = ma35d1_clk_mux("qspi1_mux",
+ clk_base + REG_CLK_CLKSEL4,
+ 10, 2, qspi1_sel_clks,
+ ARRAY_SIZE(qspi1_sel_clks));
+ hws[QSPI1_GATE] = ma35d1_clk_gate("qspi1_gate", "qspi1_mux",
+ clk_base + REG_CLK_APBCLK1, 7);
+
+ hws[SMC0_MUX] = ma35d1_clk_mux("smc0_mux",
+ clk_base + REG_CLK_CLKSEL4,
+ 28, 1, smc_sel_clks,
+ ARRAY_SIZE(smc_sel_clks));
+ hws[SMC0_DIV] = ma35d1_clk_divider("smc0_div", "smc0_mux",
+ clk_base + REG_CLK_CLKDIV1,
+ 0, 4);
+ hws[SMC0_GATE] = ma35d1_clk_gate("smc0_gate", "smc0_div",
+ clk_base + REG_CLK_APBCLK1, 12);
+ hws[SMC1_MUX] = ma35d1_clk_mux("smc1_mux",
+ clk_base + REG_CLK_CLKSEL4,
+ 29, 1, smc_sel_clks,
+ ARRAY_SIZE(smc_sel_clks));
+ hws[SMC1_DIV] = ma35d1_clk_divider("smc1_div", "smc1_mux",
+ clk_base + REG_CLK_CLKDIV1,
+ 4, 4);
+ hws[SMC1_GATE] = ma35d1_clk_gate("smc1_gate", "smc1_div",
+ clk_base + REG_CLK_APBCLK1, 13);
+
+ hws[WDT0_MUX] = ma35d1_clk_mux("wdt0_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 20, 2, wdt0_sel_clks,
+ ARRAY_SIZE(wdt0_sel_clks));
+ hws[WDT0_GATE] = ma35d1_clk_gate("wdt0_gate", "wdt0_mux",
+ clk_base + REG_CLK_APBCLK1, 16);
+ hws[WDT1_MUX] = ma35d1_clk_mux("wdt1_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 24, 2, wdt1_sel_clks,
+ ARRAY_SIZE(wdt1_sel_clks));
+ hws[WDT1_GATE] = ma35d1_clk_gate("wdt1_gate", "wdt1_mux",
+ clk_base + REG_CLK_APBCLK1, 17);
+ hws[WDT2_MUX] = ma35d1_clk_mux("wdt2_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 28, 2, wdt2_sel_clks,
+ ARRAY_SIZE(wdt2_sel_clks));
+ hws[WDT2_GATE] = ma35d1_clk_gate("wdt2_gate", "wdt2_mux",
+ clk_base + REG_CLK_APBCLK1, 18);
+
+ hws[WWDT0_MUX] = ma35d1_clk_mux("wwdt0_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 22, 2, wwdt0_sel_clks,
+ ARRAY_SIZE(wwdt0_sel_clks));
+ hws[WWDT1_MUX] = ma35d1_clk_mux("wwdt1_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 26, 2, wwdt1_sel_clks,
+ ARRAY_SIZE(wwdt1_sel_clks));
+ hws[WWDT2_MUX] = ma35d1_clk_mux("wwdt2_mux",
+ clk_base + REG_CLK_CLKSEL3,
+ 30, 2, wwdt2_sel_clks,
+ ARRAY_SIZE(wwdt2_sel_clks));
+
+ hws[EPWM0_GATE] = ma35d1_clk_gate("epwm0_gate", "pclk1",
+ clk_base + REG_CLK_APBCLK1, 24);
+ hws[EPWM1_GATE] = ma35d1_clk_gate("epwm1_gate", "pclk2",
+ clk_base + REG_CLK_APBCLK1, 25);
+ hws[EPWM2_GATE] = ma35d1_clk_gate("epwm2_gate", "pclk1",
+ clk_base + REG_CLK_APBCLK1, 26);
+
+ hws[I2S0_MUX] = ma35d1_clk_mux("i2s0_mux",
+ clk_base + REG_CLK_CLKSEL4,
+ 12, 2, i2s0_sel_clks,
+ ARRAY_SIZE(i2s0_sel_clks));
+ hws[I2S0_GATE] = ma35d1_clk_gate("i2s0_gate", "i2s0_mux",
+ clk_base + REG_CLK_APBCLK2, 0);
+ hws[I2S1_MUX] = ma35d1_clk_mux("i2s1_mux",
+ clk_base + REG_CLK_CLKSEL4,
+ 14, 2, i2s1_sel_clks,
+ ARRAY_SIZE(i2s1_sel_clks));
+ hws[I2S1_GATE] = ma35d1_clk_gate("i2s1_gate", "i2s1_mux",
+ clk_base + REG_CLK_APBCLK2, 1);
+
+ hws[SSMCC_GATE] = ma35d1_clk_gate("ssmcc_gate", "pclk3",
+ clk_base + REG_CLK_APBCLK2, 2);
+ hws[SSPCC_GATE] = ma35d1_clk_gate("sspcc_gate", "pclk3",
+ clk_base + REG_CLK_APBCLK2, 3);
+
+ hws[SPI0_MUX] = ma35d1_clk_mux("spi0_mux",
+ clk_base + REG_CLK_CLKSEL4,
+ 0, 2, spi0_sel_clks,
+ ARRAY_SIZE(spi0_sel_clks));
+ hws[SPI0_GATE] = ma35d1_clk_gate("spi0_gate", "spi0_mux",
+ clk_base + REG_CLK_APBCLK2, 4);
+ hws[SPI1_MUX] = ma35d1_clk_mux("spi1_mux",
+ clk_base + REG_CLK_CLKSEL4,
+ 2, 2, spi1_sel_clks,
+ ARRAY_SIZE(spi1_sel_clks));
+ hws[SPI1_GATE] = ma35d1_clk_gate("spi1_gate", "spi1_mux",
+ clk_base + REG_CLK_APBCLK2, 5);
+ hws[SPI2_MUX] = ma35d1_clk_mux("spi2_mux",
+ clk_base + REG_CLK_CLKSEL4,
+ 4, 2, spi2_sel_clks,
+ ARRAY_SIZE(spi2_sel_clks));
+ hws[SPI2_GATE] = ma35d1_clk_gate("spi2_gate", "spi2_mux",
+ clk_base + REG_CLK_APBCLK2, 6);
+ hws[SPI3_MUX] = ma35d1_clk_mux("spi3_mux",
+ clk_base + REG_CLK_CLKSEL4,
+ 6, 2, spi3_sel_clks,
+ ARRAY_SIZE(spi3_sel_clks));
+ hws[SPI3_GATE] = ma35d1_clk_gate("spi3_gate", "spi3_mux",
+ clk_base + REG_CLK_APBCLK2, 7);
+
+ hws[ECAP0_GATE] = ma35d1_clk_gate("ecap0_gate", "pclk1",
+ clk_base + REG_CLK_APBCLK2, 8);
+ hws[ECAP1_GATE] = ma35d1_clk_gate("ecap1_gate", "pclk2",
+ clk_base + REG_CLK_APBCLK2, 9);
+ hws[ECAP2_GATE] = ma35d1_clk_gate("ecap2_gate", "pclk1",
+ clk_base + REG_CLK_APBCLK2, 10);
+
+ hws[QEI0_GATE] = ma35d1_clk_gate("qei0_gate", "pclk1",
+ clk_base + REG_CLK_APBCLK2, 12);
+ hws[QEI1_GATE] = ma35d1_clk_gate("qei1_gate", "pclk2",
+ clk_base + REG_CLK_APBCLK2, 13);
+ hws[QEI2_GATE] = ma35d1_clk_gate("qei2_gate", "pclk1",
+ clk_base + REG_CLK_APBCLK2, 14);
+
+ hws[ADC_DIV] = ma35d1_reg_adc_clkdiv(&pdev->dev, "adc_div", "pclk0",
+ &ma35d1_lock, 0,
+ clk_base + REG_CLK_CLKDIV4,
+ 4, 17, 0x1ffff);
+ hws[ADC_GATE] = ma35d1_clk_gate("adc_gate", "adc_div",
+ clk_base + REG_CLK_APBCLK2, 24);
+
+ hws[EADC_DIV] = ma35d1_clk_divider_table("eadc_div", "pclk2",
+ clk_base + REG_CLK_CLKDIV4,
+ 0, 4, eadc_div_table);
+ hws[EADC_GATE] = ma35d1_clk_gate("eadc_gate", "eadc_div",
+ clk_base + REG_CLK_APBCLK2, 25);
+
+ return devm_of_clk_add_hw_provider(&pdev->dev,
+ of_clk_hw_onecell_get,
+ ma35d1_hw_data);
+}
+
+static const struct of_device_id ma35d1_clk_of_match[] = {
+ { .compatible = "nuvoton,ma35d1-clk" },
+ { }
+};
+MODULE_DEVICE_TABLE(of, ma35d1_clk_of_match);
+
+static struct platform_driver ma35d1_clk_driver = {
+ .probe = ma35d1_clocks_probe,
+ .driver = {
+ .name = "ma35d1-clk",
+ .of_match_table = ma35d1_clk_of_match,
+ },
+};
+
+static int __init ma35d1_clocks_init(void)
+{
+ return platform_driver_register(&ma35d1_clk_driver);
+}
+
+postcore_initcall(ma35d1_clocks_init);
+
+MODULE_AUTHOR("Chi-Fang Li <[email protected]>");
+MODULE_DESCRIPTION("NUVOTON MA35D1 Clock Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/clk/nuvoton/clk-ma35d1.h b/drivers/clk/nuvoton/clk-ma35d1.h
new file mode 100644
index 000000000000..2c47667bfd5d
--- /dev/null
+++ b/drivers/clk/nuvoton/clk-ma35d1.h
@@ -0,0 +1,123 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2023 Nuvoton Technology Corp.
+ * Author: Chi-Fang Li <[email protected]>
+ */
+
+#ifndef __DRV_CLK_NUVOTON_MA35D1_H
+#define __DRV_CLK_NUVOTON_MA35D1_H
+
+enum ma35d1_pll_type {
+ MA35D1_CAPLL,
+ MA35D1_DDRPLL,
+ MA35D1_APLL,
+ MA35D1_EPLL,
+ MA35D1_VPLL,
+};
+
+enum ma35d1_pll_mode {
+ PLL_MODE_INT,
+ PLL_MODE_FRAC,
+ PLL_MODE_SS,
+};
+
+#define PLL_MAX_NUM 5
+
+/* PLL frequency limits */
+#define PLL_FREF_MAX_FREQ 200000000UL
+#define PLL_FREF_MIN_FREQ 1000000UL
+#define PLL_FREF_M_MAX_FREQ 40000000UL
+#define PLL_FREF_M_MIN_FREQ 10000000UL
+#define PLL_FCLK_MAX_FREQ 2400000000UL
+#define PLL_FCLK_MIN_FREQ 600000000UL
+#define PLL_FCLKO_MAX_FREQ 2400000000UL
+#define PLL_FCLKO_MIN_FREQ 85700000UL
+#define PLL_SS_RATE 0x77
+#define PLL_SLOPE 0x58CFA
+
+/* Clock Control Registers Offset */
+#define REG_CLK_PWRCTL 0x00
+#define REG_CLK_SYSCLK0 0x04
+#define REG_CLK_SYSCLK1 0x08
+#define REG_CLK_APBCLK0 0x0C
+#define REG_CLK_APBCLK1 0x10
+#define REG_CLK_APBCLK2 0x14
+#define REG_CLK_CLKSEL0 0x18
+#define REG_CLK_CLKSEL1 0x1C
+#define REG_CLK_CLKSEL2 0x20
+#define REG_CLK_CLKSEL3 0x24
+#define REG_CLK_CLKSEL4 0x28
+#define REG_CLK_CLKDIV0 0x2C
+#define REG_CLK_CLKDIV1 0x30
+#define REG_CLK_CLKDIV2 0x34
+#define REG_CLK_CLKDIV3 0x38
+#define REG_CLK_CLKDIV4 0x3C
+#define REG_CLK_CLKOCTL 0x40
+#define REG_CLK_STATUS 0x50
+#define REG_CLK_PLL0CTL0 0x60
+#define REG_CLK_PLL2CTL0 0x80
+#define REG_CLK_PLL2CTL1 0x84
+#define REG_CLK_PLL2CTL2 0x88
+#define REG_CLK_PLL3CTL0 0x90
+#define REG_CLK_PLL3CTL1 0x94
+#define REG_CLK_PLL3CTL2 0x98
+#define REG_CLK_PLL4CTL0 0xA0
+#define REG_CLK_PLL4CTL1 0xA4
+#define REG_CLK_PLL4CTL2 0xA8
+#define REG_CLK_PLL5CTL0 0xB0
+#define REG_CLK_PLL5CTL1 0xB4
+#define REG_CLK_PLL5CTL2 0xB8
+#define REG_CLK_CLKDCTL 0xC0
+#define REG_CLK_CLKDSTS 0xC4
+#define REG_CLK_CDUPB 0xC8
+#define REG_CLK_CDLOWB 0xCC
+#define REG_CLK_CKFLTRCTL 0xD0
+#define REG_CLK_TESTCLK 0xF0
+#define REG_CLK_PLLCTL 0x40
+
+#define REG_PLL_CTL0_OFFSET 0x0
+#define REG_PLL_CTL1_OFFSET 0x4
+#define REG_PLL_CTL2_OFFSET 0x8
+
+/* bit fields for REG_CLK_PLL0CTL0, which is SMIC PLL design */
+#define SPLL0_CTL0_FBDIV GENMASK(7, 0)
+#define SPLL0_CTL0_INDIV GENMASK(11, 8)
+#define SPLL0_CTL0_OUTDIV GENMASK(13, 12)
+#define SPLL0_CTL0_PD BIT(16)
+#define SPLL0_CTL0_BP BIT(17)
+
+/* bit fields for REG_CLK_PLLxCTL0 ~ REG_CLK_PLLxCTL2, where x = 2 ~ 5 */
+#define PLL_CTL0_FBDIV GENMASK(10, 0)
+#define PLL_CTL0_INDIV GENMASK(17, 12)
+#define PLL_CTL0_MODE GENMASK(19, 18)
+#define PLL_CTL0_SSRATE GENMASK(30, 20)
+#define PLL_CTL1_PD BIT(0)
+#define PLL_CTL1_BP BIT(1)
+#define PLL_CTL1_OUTDIV GENMASK(6, 4)
+#define PLL_CTL1_FRAC GENMASK(31, 24)
+#define PLL_CTL2_SLOPE GENMASK(23, 0)
+
+#define INDIV_MIN 1
+#define INDIV_MAX 63
+#define FBDIV_MIN 16
+#define FBDIV_MAX 2047
+#define FBDIV_FRAC_MIN 1600
+#define FBDIV_FRAC_MAX 204700
+#define OUTDIV_MIN 1
+#define OUTDIV_MAX 7
+
+struct clk_hw *ma35d1_reg_clk_pll(enum ma35d1_pll_type type, u8 u8mode,
+ const char *name, const char *parent,
+ unsigned long targetFreq,
+ void __iomem *base,
+ struct regmap *regmap);
+
+struct clk_hw *ma35d1_reg_adc_clkdiv(struct device *dev,
+ const char *name,
+ const char *parent_name,
+ spinlock_t *lock,
+ unsigned long flags,
+ void __iomem *reg, u8 shift,
+ u8 width, u32 mask_bit);
+
+#endif /* __DRV_CLK_NUVOTON_MA35D1_H */
--
2.34.1

2023-03-28 02:21:22

by Jacky Huang

[permalink] [raw]
Subject: [PATCH v6 10/12] reset: Add Nuvoton ma35d1 reset driver support

From: Jacky Huang <[email protected]>

This driver supports individual IP reset for ma35d1. The reset
control registers is a subset of system control registers.

Signed-off-by: Jacky Huang <[email protected]>
---
drivers/reset/Kconfig | 6 ++
drivers/reset/Makefile | 1 +
drivers/reset/reset-ma35d1.c | 152 +++++++++++++++++++++++++++++++++++
3 files changed, 159 insertions(+)
create mode 100644 drivers/reset/reset-ma35d1.c

diff --git a/drivers/reset/Kconfig b/drivers/reset/Kconfig
index 2a52c990d4fe..58477c6ca9b8 100644
--- a/drivers/reset/Kconfig
+++ b/drivers/reset/Kconfig
@@ -143,6 +143,12 @@ config RESET_NPCM
This enables the reset controller driver for Nuvoton NPCM
BMC SoCs.

+config RESET_NUVOTON_MA35D1
+ bool "Nuvton MA35D1 Reset Driver"
+ default ARCH_NUVOTON || COMPILE_TEST
+ help
+ This enables the reset controller driver for Nuvoton MA35D1 SoC.
+
config RESET_OXNAS
bool

diff --git a/drivers/reset/Makefile b/drivers/reset/Makefile
index 3e7e5fd633a8..fd52dcf66a99 100644
--- a/drivers/reset/Makefile
+++ b/drivers/reset/Makefile
@@ -20,6 +20,7 @@ obj-$(CONFIG_RESET_MCHP_SPARX5) += reset-microchip-sparx5.o
obj-$(CONFIG_RESET_MESON) += reset-meson.o
obj-$(CONFIG_RESET_MESON_AUDIO_ARB) += reset-meson-audio-arb.o
obj-$(CONFIG_RESET_NPCM) += reset-npcm.o
+obj-$(CONFIG_RESET_NUVOTON_MA35D1) += reset-ma35d1.o
obj-$(CONFIG_RESET_OXNAS) += reset-oxnas.o
obj-$(CONFIG_RESET_PISTACHIO) += reset-pistachio.o
obj-$(CONFIG_RESET_POLARFIRE_SOC) += reset-mpfs.o
diff --git a/drivers/reset/reset-ma35d1.c b/drivers/reset/reset-ma35d1.c
new file mode 100644
index 000000000000..221299e7b873
--- /dev/null
+++ b/drivers/reset/reset-ma35d1.c
@@ -0,0 +1,152 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2023 Nuvoton Technology Corp.
+ * Author: Chi-Fang Li <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/mfd/syscon.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/reboot.h>
+#include <linux/regmap.h>
+#include <linux/reset-controller.h>
+#include <dt-bindings/reset/nuvoton,ma35d1-reset.h>
+
+#define REG_SYS_IPRST0 0x020
+#define REG_SYS_IPRST1 0x024
+#define REG_SYS_IPRST2 0x028
+#define REG_SYS_IPRST3 0x02C
+
+#define RST_PRE_REG 32
+
+struct ma35d1_reset_data {
+ struct reset_controller_dev rcdev;
+ struct regmap *regmap;
+};
+
+struct ma35d1_reboot_data {
+ struct notifier_block restart_handler;
+ struct regmap *regmap;
+};
+
+static int ma35d1_restart_handler(struct notifier_block *this,
+ unsigned long mode, void *cmd)
+{
+ struct ma35d1_reboot_data *data =
+ container_of(this, struct ma35d1_reboot_data,
+ restart_handler);
+ return regmap_write(data->regmap, REG_SYS_IPRST0, 1 << MA35D1_RESET_CHIP);
+}
+
+static int ma35d1_reset_update(struct reset_controller_dev *rcdev,
+ unsigned long id, bool assert)
+{
+ unsigned int reg;
+ int ret;
+ int offset = (id / RST_PRE_REG) * 4;
+ struct ma35d1_reset_data *data =
+ container_of(rcdev, struct ma35d1_reset_data, rcdev);
+
+ ret = regmap_read(data->regmap, REG_SYS_IPRST0 + offset, &reg);
+ if (ret < 0)
+ return ret;
+ if (assert)
+ reg |= 1 << (id % RST_PRE_REG);
+ else
+ reg &= ~(1 << (id % RST_PRE_REG));
+
+ return regmap_write(data->regmap, REG_SYS_IPRST0 + offset, reg);
+}
+
+static int ma35d1_reset_assert(struct reset_controller_dev *rcdev,
+ unsigned long id)
+{
+ return ma35d1_reset_update(rcdev, id, true);
+}
+
+static int ma35d1_reset_deassert(struct reset_controller_dev *rcdev,
+ unsigned long id)
+{
+ return ma35d1_reset_update(rcdev, id, false);
+}
+
+static int ma35d1_reset_status(struct reset_controller_dev *rcdev,
+ unsigned long id)
+{
+ int reg, ret;
+ int offset = id / RST_PRE_REG;
+ struct ma35d1_reset_data *data =
+ container_of(rcdev, struct ma35d1_reset_data, rcdev);
+
+ ret = regmap_read(data->regmap, REG_SYS_IPRST0 + offset, &reg);
+ if (ret < 0)
+ return ret;
+ return !!(reg & BIT(id % RST_PRE_REG));
+}
+
+static const struct reset_control_ops ma35d1_reset_ops = {
+ .assert = ma35d1_reset_assert,
+ .deassert = ma35d1_reset_deassert,
+ .status = ma35d1_reset_status,
+};
+
+static const struct of_device_id ma35d1_reset_dt_ids[] = {
+ { .compatible = "nuvoton,ma35d1-reset" },
+ { },
+};
+
+static int ma35d1_reset_probe(struct platform_device *pdev)
+{
+ int err;
+ struct device *dev = &pdev->dev;
+ struct device_node *parent;
+ struct ma35d1_reset_data *reset_data;
+ struct ma35d1_reboot_data *reboot_data;
+
+ if (!pdev->dev.of_node) {
+ dev_err(&pdev->dev, "Device tree node not found\n");
+ return -EINVAL;
+ }
+
+ reset_data = devm_kzalloc(dev, sizeof(*reset_data), GFP_KERNEL);
+ if (!reset_data)
+ return -ENOMEM;
+
+ reboot_data = devm_kzalloc(dev, sizeof(*reboot_data), GFP_KERNEL);
+ if (!reboot_data)
+ return -ENOMEM;
+
+ parent = of_get_parent(dev->of_node); /* parent should be syscon node */
+ reset_data->regmap = syscon_node_to_regmap(parent);
+ of_node_put(parent);
+ if (IS_ERR(reset_data->regmap)) {
+ dev_err(&pdev->dev, "Failed to get SYS register base\n");
+ return PTR_ERR(reset_data->regmap);
+ }
+
+ reset_data->rcdev.owner = THIS_MODULE;
+ reset_data->rcdev.nr_resets = MA35D1_RESET_COUNT;
+ reset_data->rcdev.ops = &ma35d1_reset_ops;
+ reset_data->rcdev.of_node = dev->of_node;
+
+ reboot_data->regmap = reset_data->regmap;
+ reboot_data->restart_handler.notifier_call = ma35d1_restart_handler;
+ reboot_data->restart_handler.priority = 192;
+
+ err = register_restart_handler(&reboot_data->restart_handler);
+ if (err)
+ dev_warn(&pdev->dev, "failed to register restart handler\n");
+
+ return devm_reset_controller_register(dev, &reset_data->rcdev);
+}
+
+static struct platform_driver ma35d1_reset_driver = {
+ .probe = ma35d1_reset_probe,
+ .driver = {
+ .name = "ma35d1-reset",
+ .of_match_table = ma35d1_reset_dt_ids,
+ },
+};
+
+builtin_platform_driver(ma35d1_reset_driver);
--
2.34.1

2023-03-28 02:21:33

by Jacky Huang

[permalink] [raw]
Subject: [PATCH v6 03/12] dt-bindings: clock: nuvoton: add binding for ma35d1 clock controller

From: Jacky Huang <[email protected]>

Add the dt-bindings header for Nuvoton ma35d1, that gets shared
between the clock controller and clock references in the dts.
Add documentation to describe nuvoton ma35d1 clock driver.

Signed-off-by: Jacky Huang <[email protected]>
---
.../bindings/clock/nuvoton,ma35d1-clk.yaml | 72 +++++
.../dt-bindings/clock/nuvoton,ma35d1-clk.h | 253 ++++++++++++++++++
2 files changed, 325 insertions(+)
create mode 100644 Documentation/devicetree/bindings/clock/nuvoton,ma35d1-clk.yaml
create mode 100644 include/dt-bindings/clock/nuvoton,ma35d1-clk.h

diff --git a/Documentation/devicetree/bindings/clock/nuvoton,ma35d1-clk.yaml b/Documentation/devicetree/bindings/clock/nuvoton,ma35d1-clk.yaml
new file mode 100644
index 000000000000..0a9c2a115e6b
--- /dev/null
+++ b/Documentation/devicetree/bindings/clock/nuvoton,ma35d1-clk.yaml
@@ -0,0 +1,72 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/clock/nuvoton,ma35d1-clk.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Nuvoton MA35D1 Clock Controller Module
+
+maintainers:
+ - Chi-Fang Li <[email protected]>
+ - Jacky Huang <[email protected]>
+
+description: |
+ The MA35D1 clock controller generates clocks for the whole chip,
+ including system clocks and all peripheral clocks.
+
+ See also:
+ include/dt-bindings/clock/ma35d1-clk.h
+
+properties:
+ compatible:
+ items:
+ - const: nuvoton,ma35d1-clk
+ - const: syscon
+
+ reg:
+ maxItems: 1
+
+ "#clock-cells":
+ const: 1
+
+ clocks:
+ maxItems: 1
+
+ nuvoton,pll-mode:
+ description:
+ A list of PLL operation mode corresponding to CAPLL, DDRPLL, APLL,
+ EPLL, and VPLL in sequential.
+ maxItems: 5
+ items:
+ enum:
+ - integer
+ - fractional
+ - spread-spectrum
+ $ref: /schemas/types.yaml#/definitions/non-unique-string-array
+
+ nuvoton,sys:
+ description:
+ Use to unlock and lock some clock controller registers. The lock
+ control register is in system controller.
+ $ref: /schemas/types.yaml#/definitions/phandle
+
+required:
+ - compatible
+ - reg
+ - "#clock-cells"
+ - clocks
+ - nuvoton,sys
+
+additionalProperties: false
+
+examples:
+ - |
+
+ clock-controller@40460200 {
+ compatible = "nuvoton,ma35d1-clk", "syscon";
+ reg = <0x40460200 0x100>;
+ #clock-cells = <1>;
+ clocks = <&clk_hxt>;
+ nuvoton,sys = <&sys>;
+ };
+...
diff --git a/include/dt-bindings/clock/nuvoton,ma35d1-clk.h b/include/dt-bindings/clock/nuvoton,ma35d1-clk.h
new file mode 100644
index 000000000000..ba2d70f776a6
--- /dev/null
+++ b/include/dt-bindings/clock/nuvoton,ma35d1-clk.h
@@ -0,0 +1,253 @@
+/* SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause) */
+/*
+ * Copyright (C) 2023 Nuvoton Technologies.
+ */
+
+#ifndef __DT_BINDINGS_CLOCK_NUVOTON_MA35D1_CLK_H
+#define __DT_BINDINGS_CLOCK_NUVOTON_MA35D1_CLK_H
+
+/* external and internal oscillator clocks */
+#define HXT 0
+#define HXT_GATE 1
+#define LXT 2
+#define LXT_GATE 3
+#define HIRC 4
+#define HIRC_GATE 5
+#define LIRC 6
+#define LIRC_GATE 7
+/* PLLs */
+#define CAPLL 8
+#define SYSPLL 9
+#define DDRPLL 10
+#define APLL 11
+#define EPLL 12
+#define VPLL 13
+/* EPLL divider */
+#define EPLL_DIV2 14
+#define EPLL_DIV4 15
+#define EPLL_DIV8 16
+/* CPU clock, system clock, AXI, HCLK and PCLK */
+#define CA35CLK_MUX 17
+#define AXICLK_DIV2 18
+#define AXICLK_DIV4 19
+#define AXICLK_MUX 20
+#define SYSCLK0_MUX 21
+#define SYSCLK1_MUX 22
+#define SYSCLK1_DIV2 23
+#define HCLK0 24
+#define HCLK1 25
+#define HCLK2 26
+#define PCLK0 27
+#define PCLK1 28
+#define PCLK2 29
+#define HCLK3 30
+#define PCLK3 31
+#define PCLK4 32
+/* AXI and AHB peripheral clocks */
+#define USBPHY0 33
+#define USBPHY1 34
+#define DDR0_GATE 35
+#define DDR6_GATE 36
+#define CAN0_MUX 37
+#define CAN0_DIV 38
+#define CAN0_GATE 39
+#define CAN1_MUX 40
+#define CAN1_DIV 41
+#define CAN1_GATE 42
+#define CAN2_MUX 43
+#define CAN2_DIV 44
+#define CAN2_GATE 45
+#define CAN3_MUX 46
+#define CAN3_DIV 47
+#define CAN3_GATE 48
+#define SDH0_MUX 49
+#define SDH0_GATE 50
+#define SDH1_MUX 51
+#define SDH1_GATE 52
+#define NAND_GATE 53
+#define USBD_GATE 54
+#define USBH_GATE 55
+#define HUSBH0_GATE 56
+#define HUSBH1_GATE 57
+#define GFX_MUX 58
+#define GFX_GATE 59
+#define VC8K_GATE 60
+#define DCU_MUX 61
+#define DCU_GATE 62
+#define DCUP_DIV 63
+#define EMAC0_GATE 64
+#define EMAC1_GATE 65
+#define CCAP0_MUX 66
+#define CCAP0_DIV 67
+#define CCAP0_GATE 68
+#define CCAP1_MUX 69
+#define CCAP1_DIV 70
+#define CCAP1_GATE 71
+#define PDMA0_GATE 72
+#define PDMA1_GATE 73
+#define PDMA2_GATE 74
+#define PDMA3_GATE 75
+#define WH0_GATE 76
+#define WH1_GATE 77
+#define HWS_GATE 78
+#define EBI_GATE 79
+#define SRAM0_GATE 80
+#define SRAM1_GATE 81
+#define ROM_GATE 82
+#define TRA_GATE 83
+#define DBG_MUX 84
+#define DBG_GATE 85
+#define CKO_MUX 86
+#define CKO_DIV 87
+#define CKO_GATE 88
+#define GTMR_GATE 89
+#define GPA_GATE 90
+#define GPB_GATE 91
+#define GPC_GATE 92
+#define GPD_GATE 93
+#define GPE_GATE 94
+#define GPF_GATE 95
+#define GPG_GATE 96
+#define GPH_GATE 97
+#define GPI_GATE 98
+#define GPJ_GATE 99
+#define GPK_GATE 100
+#define GPL_GATE 101
+#define GPM_GATE 102
+#define GPN_GATE 103
+/* APB peripheral clocks */
+#define TMR0_MUX 104
+#define TMR0_GATE 105
+#define TMR1_MUX 106
+#define TMR1_GATE 107
+#define TMR2_MUX 108
+#define TMR2_GATE 109
+#define TMR3_MUX 110
+#define TMR3_GATE 111
+#define TMR4_MUX 112
+#define TMR4_GATE 113
+#define TMR5_MUX 114
+#define TMR5_GATE 115
+#define TMR6_MUX 116
+#define TMR6_GATE 117
+#define TMR7_MUX 118
+#define TMR7_GATE 119
+#define TMR8_MUX 120
+#define TMR8_GATE 121
+#define TMR9_MUX 122
+#define TMR9_GATE 123
+#define TMR10_MUX 124
+#define TMR10_GATE 125
+#define TMR11_MUX 126
+#define TMR11_GATE 127
+#define UART0_MUX 128
+#define UART0_DIV 129
+#define UART0_GATE 130
+#define UART1_MUX 131
+#define UART1_DIV 132
+#define UART1_GATE 133
+#define UART2_MUX 134
+#define UART2_DIV 135
+#define UART2_GATE 136
+#define UART3_MUX 137
+#define UART3_DIV 138
+#define UART3_GATE 139
+#define UART4_MUX 140
+#define UART4_DIV 141
+#define UART4_GATE 142
+#define UART5_MUX 143
+#define UART5_DIV 144
+#define UART5_GATE 145
+#define UART6_MUX 146
+#define UART6_DIV 147
+#define UART6_GATE 148
+#define UART7_MUX 149
+#define UART7_DIV 150
+#define UART7_GATE 151
+#define UART8_MUX 152
+#define UART8_DIV 153
+#define UART8_GATE 154
+#define UART9_MUX 155
+#define UART9_DIV 156
+#define UART9_GATE 157
+#define UART10_MUX 158
+#define UART10_DIV 159
+#define UART10_GATE 160
+#define UART11_MUX 161
+#define UART11_DIV 162
+#define UART11_GATE 163
+#define UART12_MUX 164
+#define UART12_DIV 165
+#define UART12_GATE 166
+#define UART13_MUX 167
+#define UART13_DIV 168
+#define UART13_GATE 169
+#define UART14_MUX 170
+#define UART14_DIV 171
+#define UART14_GATE 172
+#define UART15_MUX 173
+#define UART15_DIV 174
+#define UART15_GATE 175
+#define UART16_MUX 176
+#define UART16_DIV 177
+#define UART16_GATE 178
+#define RTC_GATE 179
+#define DDR_GATE 180
+#define KPI_MUX 181
+#define KPI_DIV 182
+#define KPI_GATE 183
+#define I2C0_GATE 184
+#define I2C1_GATE 185
+#define I2C2_GATE 186
+#define I2C3_GATE 187
+#define I2C4_GATE 188
+#define I2C5_GATE 189
+#define QSPI0_MUX 190
+#define QSPI0_GATE 191
+#define QSPI1_MUX 192
+#define QSPI1_GATE 193
+#define SMC0_MUX 194
+#define SMC0_DIV 195
+#define SMC0_GATE 196
+#define SMC1_MUX 197
+#define SMC1_DIV 198
+#define SMC1_GATE 199
+#define WDT0_MUX 200
+#define WDT0_GATE 201
+#define WDT1_MUX 202
+#define WDT1_GATE 203
+#define WDT2_MUX 204
+#define WDT2_GATE 205
+#define WWDT0_MUX 206
+#define WWDT1_MUX 207
+#define WWDT2_MUX 208
+#define EPWM0_GATE 209
+#define EPWM1_GATE 210
+#define EPWM2_GATE 211
+#define I2S0_MUX 212
+#define I2S0_GATE 213
+#define I2S1_MUX 214
+#define I2S1_GATE 215
+#define SSMCC_GATE 216
+#define SSPCC_GATE 217
+#define SPI0_MUX 218
+#define SPI0_GATE 219
+#define SPI1_MUX 220
+#define SPI1_GATE 221
+#define SPI2_MUX 222
+#define SPI2_GATE 223
+#define SPI3_MUX 224
+#define SPI3_GATE 225
+#define ECAP0_GATE 226
+#define ECAP1_GATE 227
+#define ECAP2_GATE 228
+#define QEI0_GATE 229
+#define QEI1_GATE 230
+#define QEI2_GATE 231
+#define ADC_DIV 232
+#define ADC_GATE 233
+#define EADC_DIV 234
+#define EADC_GATE 235
+#define CLK_MAX_IDX 236
+
+#endif /* __DT_BINDINGS_CLOCK_NUVOTON_MA35D1_CLK_H */
--
2.34.1

2023-03-28 02:21:55

by Jacky Huang

[permalink] [raw]
Subject: [PATCH v6 12/12] MAINTAINERS: Add entry for NUVOTON MA35

From: Jacky Huang <[email protected]>

Add entry for Nuvton ma35d1 maintainer and files.

Signed-off-by: Jacky Huang <[email protected]>
---
MAINTAINERS | 9 +++++++++
1 file changed, 9 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index 1dc8bd26b6cf..a3912717f1da 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2501,6 +2501,15 @@ F: drivers/rtc/rtc-ab8500.c
F: drivers/rtc/rtc-pl031.c
F: drivers/soc/ux500/

+ARM/NUVOTON MA35 ARCHITECTURE
+M: Jacky Huang <[email protected]>
+M: Shan-Chun Hung <[email protected]>
+L: [email protected] (moderated for non-subscribers)
+S: Supported
+F: Documentation/devicetree/bindings/*/*nuvoton*
+F: arch/arm64/boot/dts/nuvoton/*ma35*
+K: ma35d1
+
ARM/NUVOTON NPCM ARCHITECTURE
M: Avi Fishman <[email protected]>
M: Tomer Maimon <[email protected]>
--
2.34.1

2023-03-28 02:22:19

by Jacky Huang

[permalink] [raw]
Subject: [PATCH v6 11/12] tty: serial: Add Nuvoton ma35d1 serial driver support

From: Jacky Huang <[email protected]>

This adds UART and console driver for Nuvoton ma35d1 Soc.
It supports full-duplex communication, FIFO control, and
hardware flow control.

Signed-off-by: Jacky Huang <[email protected]>
---
drivers/tty/serial/Kconfig | 18 +
drivers/tty/serial/Makefile | 1 +
drivers/tty/serial/ma35d1_serial.c | 802 +++++++++++++++++++++++++++++
3 files changed, 821 insertions(+)
create mode 100644 drivers/tty/serial/ma35d1_serial.c

diff --git a/drivers/tty/serial/Kconfig b/drivers/tty/serial/Kconfig
index 0072892ca7fc..daa4ae6bc08e 100644
--- a/drivers/tty/serial/Kconfig
+++ b/drivers/tty/serial/Kconfig
@@ -1562,6 +1562,24 @@ config SERIAL_SUNPLUS_CONSOLE
you can alter that using a kernel command line option such as
"console=ttySUPx".

+config SERIAL_NUVOTON_MA35D1
+ tristate "Nuvoton MA35D1 family UART support"
+ depends on ARCH_NUVOTON || COMPILE_TEST
+ select SERIAL_CORE
+ help
+ This driver supports Nuvoton MA35D1 family UART ports. If you would
+ like to use them, you must answer Y or M to this option. Note that
+ for use as console, it must be included in kernel and not as a
+ module
+
+config SERIAL_NUVOTON_MA35D1_CONSOLE
+ bool "Console on a Nuvotn MA35D1 family UART port"
+ depends on SERIAL_NUVOTON_MA35D1=y
+ select SERIAL_CORE_CONSOLE
+ help
+ Select this options if you'd like to use the UART port0 of the
+ Nuvoton MA35D1 family as a console.
+
endmenu

config SERIAL_MCTRL_GPIO
diff --git a/drivers/tty/serial/Makefile b/drivers/tty/serial/Makefile
index cd9afd9e3018..71ebeba06ff2 100644
--- a/drivers/tty/serial/Makefile
+++ b/drivers/tty/serial/Makefile
@@ -93,3 +93,4 @@ obj-$(CONFIG_SERIAL_MCTRL_GPIO) += serial_mctrl_gpio.o

obj-$(CONFIG_SERIAL_KGDB_NMI) += kgdb_nmi.o
obj-$(CONFIG_KGDB_SERIAL_CONSOLE) += kgdboc.o
+obj-$(CONFIG_SERIAL_NUVOTON_MA35D1) += ma35d1_serial.o
diff --git a/drivers/tty/serial/ma35d1_serial.c b/drivers/tty/serial/ma35d1_serial.c
new file mode 100644
index 000000000000..75618ab0b2ca
--- /dev/null
+++ b/drivers/tty/serial/ma35d1_serial.c
@@ -0,0 +1,802 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * MA35D1 serial driver
+ * Copyright (C) 2023 Nuvoton Technology Corp.
+ */
+
+#include <linux/bitfield.h>
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/of.h>
+#include <linux/of_platform.h>
+#include <linux/serial_core.h>
+#include <linux/slab.h>
+#include <linux/tty_flip.h>
+
+#define UART_NR 17
+
+#define UART_REG_RBR 0x00
+#define UART_REG_THR 0x00
+#define UART_REG_IER 0x04
+#define UART_REG_FCR 0x08
+#define UART_REG_LCR 0x0C
+#define UART_REG_MCR 0x10
+#define UART_REG_MSR 0x14
+#define UART_REG_FSR 0x18
+#define UART_REG_ISR 0x1C
+#define UART_REG_TOR 0x20
+#define UART_REG_BAUD 0x24
+#define UART_REG_ALTCTL 0x2C
+#define UART_FUN_SEL 0x30
+#define UART_REG_WKCTL 0x40
+#define UART_REG_WKSTS 0x44
+
+/* UART_REG_IER - Interrupt Enable Register */
+#define IER_RDA_IEN BIT(0) /* RBR Available Interrupt Enable */
+#define IER_THRE_IEN BIT(1) /* THR Empty Interrupt Enable */
+#define IER_RLS_IEN BIT(2) /* RX Line Status Interrupt Enable */
+#define IER_RTO_IEN BIT(4) /* RX Time-out Interrupt Enable */
+#define IER_BUFERR_IEN BIT(5) /* Buffer Error Interrupt Enable */
+#define IER_TIME_OUT_EN BIT(11) /* RX Buffer Time-out Counter Enable */
+#define IER_AUTO_RTS BIT(12) /* nRTS Auto-flow Control Enable */
+#define IER_AUTO_CTS BIT(13) /* nCTS Auto-flow Control Enable */
+
+/* UART_REG_FCR - FIFO Control Register */
+#define FCR_RFR BIT(1) /* RX Field Software Reset */
+#define FCR_TFR BIT(2) /* TX Field Software Reset */
+#define FCR_RFITL_MASK GENMASK(7, 4) /* RX FIFO Interrupt Trigger Level */
+#define FCR_RFITL_1BYTE FIELD_PREP(FCR_RFITL_MASK, 0)
+#define FCR_RFITL_4BYTES FIELD_PREP(FCR_RFITL_MASK, 1)
+#define FCR_RFITL_8BYTES FIELD_PREP(FCR_RFITL_MASK, 2)
+#define FCR_RFITL_14BYTES FIELD_PREP(FCR_RFITL_MASK, 3)
+#define FCR_RFITL_30BYTES FIELD_PREP(FCR_RFITL_MASK, 4)
+#define FCR_RTSTRGLV_MASK GENMASK(19, 16) /* nRTS Trigger Level */
+#define FCR_RTSTRGLV_1BYTE FIELD_PREP(FCR_RTSTRGLV_MASK, 0)
+#define FCR_RTSTRGLV_4BYTES FIELD_PREP(FCR_RTSTRGLV_MASK, 1)
+#define FCR_RTSTRGLV_8BYTES FIELD_PREP(FCR_RTSTRGLV_MASK, 2)
+#define FCR_RTSTRGLV_14BYTES FIELD_PREP(FCR_RTSTRGLV_MASK, 3)
+#define FCR_RTSTRGLVL_30BYTES FIELD_PREP(FCR_RTSTRGLV_MASK, 4)
+
+/* UART_REG_LCR - Line Control Register */
+#define LCR_NSB BIT(2) /* Number of “STOP Bit” */
+#define LCR_PBE BIT(3) /* Parity Bit Enable */
+#define LCR_EPE BIT(4) /* Even Parity Enable */
+#define LCR_SPE BIT(5) /* Stick Parity Enable */
+#define LCR_BREAK BIT(6) /* Break Control */
+#define LCR_WLS_MASK GENMASK(1, 0) /* Word Length Selection */
+#define LCR_WLS_5BITS FIELD_PREP(LCR_WLS_MASK, 0)
+#define LCR_WLS_6BITS FIELD_PREP(LCR_WLS_MASK, 1)
+#define LCR_WLS_7BITS FIELD_PREP(LCR_WLS_MASK, 2)
+#define LCR_WLS_8BITS FIELD_PREP(LCR_WLS_MASK, 3)
+
+/* UART_REG_MCR - Modem Control Register */
+#define MCR_RTS_CTRL BIT(1) /* nRTS Signal Control */
+#define MCR_RTSACTLV BIT(9) /* nRTS Pin Active Level */
+#define MCR_RTSSTS BIT(13) /* nRTS Pin Status (Read Only) */
+
+/* UART_REG_MSR - Modem Status Register */
+#define MSR_CTSDETF BIT(0) /* Detect nCTS State Change Flag */
+#define MSR_CTSSTS BIT(4) /* nCTS Pin Status (Read Only) */
+#define MSR_CTSACTLV BIT(8) /* nCTS Pin Active Level */
+
+/* UART_REG_FSR - FIFO Status Register */
+#define FSR_RX_OVER_IF BIT(0) /* RX Overflow Error Interrupt Flag */
+#define FSR_PEF BIT(4) /* Parity Error Flag*/
+#define FSR_FEF BIT(5) /* Framing Error Flag */
+#define FSR_BIF BIT(6) /* Break Interrupt Flag */
+#define FSR_RX_EMPTY BIT(14) /* Receiver FIFO Empty (Read Only) */
+#define FSR_RX_FULL BIT(15) /* Receiver FIFO Full (Read Only) */
+#define FSR_TX_EMPTY BIT(22) /* Transmitter FIFO Empty (Read Only) */
+#define FSR_TX_FULL BIT(23) /* Transmitter FIFO Full (Read Only) */
+#define FSR_TX_OVER_IF BIT(24) /* TX Overflow Error Interrupt Flag */
+#define FSR_TE_FLAG BIT(28) /* Transmitter Empty Flag (Read Only) */
+#define FSR_RXPTR_MSK GENMASK(13, 8) /* TX FIFO Pointer mask */
+#define FSR_TXPTR_MSK GENMASK(21, 16) /* RX FIFO Pointer mask */
+
+/* UART_REG_ISR - Interrupt Status Register */
+#define ISR_RDA_IF BIT(0) /* RBR Available Interrupt Flag */
+#define ISR_THRE_IF BIT(1) /* THR Empty Interrupt Flag */
+#define ISR_RLSIF BIT(2) /* Receive Line Interrupt Flag */
+#define ISR_MODEMIF BIT(3) /* MODEM Interrupt Flag */
+#define ISR_RXTO_IF BIT(4) /* RX Time-out Interrupt Flag */
+#define ISR_BUFEIF BIT(5) /* Buffer Error Interrupt Flag */
+#define UART_ISR_WK_IF BIT(6) /* UART Wake-up Interrupt Flag */
+#define UART_ISR_RDAINT BIT(8) /* RBR Available Interrupt Indicator */
+#define ISR_THRE_INT BIT(9) /* THR Empty Interrupt Indicator */
+#define ISR_ALL 0xFFFFFFFF
+
+/* UART_REG_BAUD - Baud Rate Divider Register */
+#define BAUD_MODE_MASK GENMASK(29, 28)
+#define BAUD_MODE0 FIELD_PREP(BAUD_MODE_MASK, 0)
+#define BAUD_MODE1 FIELD_PREP(BAUD_MODE_MASK, 2)
+#define BAUD_MODE2 FIELD_PREP(BAUD_MODE_MASK, 3)
+
+/* UART_REG_ALTCTL - Alternate Control/Status Register */
+#define ALTCTL_RS485AUD BIT(10) /* RS-485 Auto Direction Function */
+
+/* UART_FUN_SEL - Function Select Register */
+#define FUN_SEL_MASK GENMASK(2, 0)
+#define FUN_SEL_UART FIELD_PREP(FUN_SEL_MASK, 0)
+#define FUN_SEL_RS485 FIELD_PREP(FUN_SEL_MASK, 3)
+
+/* UART FIFO depth */
+#define UART_FIFO_DEPTH 32
+/* UART console clock */
+#define UART_CONSOLE_CLK 24000000
+/* UART register ioremap size */
+#define UART_REG_SIZE 0x100
+/* Rx Timeout */
+#define UART_RX_TOUT 0x40
+
+#define UART_ISR_IF_CHECK (ISR_RDA_IF | ISR_RXTO_IF | ISR_THRE_INT | ISR_BUFEIF)
+
+#define LOOP_TIMEOUT 1000
+
+static struct uart_driver ma35d1serial_reg;
+struct clk *clk;
+
+struct uart_ma35d1_port {
+ struct uart_port port;
+ u16 capabilities; /* port capabilities */
+ u8 ier;
+ u8 lcr;
+ u8 mcr;
+ u32 baud_rate;
+ u32 console_baud_rate;
+ u32 console_line;
+ u32 console_int;
+};
+
+static struct device_node *ma35d1serial_uart_nodes[UART_NR];
+static struct uart_ma35d1_port ma35d1serial_ports[UART_NR] = { 0 };
+
+static struct uart_ma35d1_port *to_ma35d1_uart_port(struct uart_port *uart)
+{
+ return container_of(uart, struct uart_ma35d1_port, port);
+}
+
+static u32 serial_in(struct uart_ma35d1_port *p, u32 offset)
+{
+ return readl_relaxed(p->port.membase + offset);
+}
+
+static void serial_out(struct uart_ma35d1_port *p, u32 offset, u32 value)
+{
+ writel_relaxed(value, p->port.membase + offset);
+}
+
+static void __stop_tx(struct uart_ma35d1_port *p)
+{
+ u32 ier;
+
+ ier = serial_in(p, UART_REG_IER);
+ if (ier & IER_THRE_IEN)
+ serial_out(p, UART_REG_IER, ier & ~IER_THRE_IEN);
+}
+
+static void ma35d1serial_stop_tx(struct uart_port *port)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+
+ __stop_tx(up);
+}
+
+static void transmit_chars(struct uart_ma35d1_port *up)
+{
+ struct circ_buf *xmit = &up->port.state->xmit;
+ int count;
+ u8 ch;
+
+ if (uart_tx_stopped(&up->port)) {
+ ma35d1serial_stop_tx(&up->port);
+ return;
+ }
+ if (uart_circ_empty(xmit)) {
+ __stop_tx(up);
+ return;
+ }
+
+ count = UART_FIFO_DEPTH - ((serial_in(up, UART_REG_FSR) & FSR_TXPTR_MSK) >> 16);
+
+ uart_port_tx_limited(&up->port, ch, count,
+ !(serial_in(up, UART_REG_FSR) & FSR_TX_FULL),
+ serial_out(up, UART_REG_THR, ch),
+ ({}));
+
+ if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
+ uart_write_wakeup(&up->port);
+
+ if (uart_circ_empty(xmit))
+ __stop_tx(up);
+}
+
+static void ma35d1serial_start_tx(struct uart_port *port)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+ u32 ier;
+
+ ier = serial_in(up, UART_REG_IER);
+ serial_out(up, UART_REG_IER, ier & ~IER_THRE_IEN);
+ transmit_chars(up);
+ serial_out(up, UART_REG_IER, ier | IER_THRE_IEN);
+}
+
+static void ma35d1serial_stop_rx(struct uart_port *port)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+
+ serial_out(up, UART_REG_IER, serial_in(up, UART_REG_IER) & ~IER_RDA_IEN);
+}
+
+static void receive_chars(struct uart_ma35d1_port *up)
+{
+ u8 flag;
+ u32 fsr;
+ unsigned int ch;
+ int max_count = 256;
+
+ fsr = serial_in(up, UART_REG_FSR);
+ do {
+ flag = TTY_NORMAL;
+ up->port.icount.rx++;
+
+ if (unlikely(fsr & (FSR_BIF | FSR_FEF | FSR_PEF | FSR_RX_OVER_IF))) {
+ if (fsr & FSR_BIF) {
+ up->port.icount.brk++;
+ if (uart_handle_break(&up->port))
+ continue;
+ }
+ if (fsr & FSR_FEF)
+ up->port.icount.frame++;
+ if (fsr & FSR_PEF)
+ up->port.icount.parity++;
+ if (fsr & FSR_RX_OVER_IF)
+ up->port.icount.overrun++;
+
+ serial_out(up, UART_REG_FSR, fsr &
+ (FSR_BIF | FSR_FEF | FSR_PEF | FSR_RX_OVER_IF));
+
+ if (fsr & FSR_BIF)
+ flag = TTY_BREAK;
+ else if (fsr & FSR_PEF)
+ flag = TTY_PARITY;
+ else if (fsr & FSR_FEF)
+ flag = TTY_FRAME;
+ }
+
+ ch = serial_in(up, UART_REG_RBR);
+ if (uart_handle_sysrq_char(&up->port, ch))
+ continue;
+
+ spin_lock(&up->port.lock);
+ uart_insert_char(&up->port, fsr, FSR_RX_OVER_IF, ch, flag);
+ spin_unlock(&up->port.lock);
+
+ fsr = serial_in(up, UART_REG_FSR);
+ } while (!(fsr & FSR_RX_EMPTY) && (max_count-- > 0));
+
+ spin_lock(&up->port.lock);
+ tty_flip_buffer_push(&up->port.state->port);
+ spin_unlock(&up->port.lock);
+}
+
+static irqreturn_t ma35d1serial_interrupt(int irq, void *dev_id)
+{
+ struct uart_port *port = dev_id;
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+ u32 isr, fsr;
+
+ isr = serial_in(up, UART_REG_ISR);
+ fsr = serial_in(up, UART_REG_FSR);
+
+ if (!(isr & UART_ISR_IF_CHECK))
+ return IRQ_NONE;
+
+ if (isr & (ISR_RDA_IF | ISR_RXTO_IF))
+ receive_chars(up);
+ if (isr & ISR_THRE_INT)
+ transmit_chars(up);
+ if (fsr & FSR_TX_OVER_IF)
+ serial_out(up, UART_REG_FSR, FSR_TX_OVER_IF);
+
+ return IRQ_HANDLED;
+}
+
+static u32 ma35d1serial_tx_empty(struct uart_port *port)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+ u32 fsr;
+
+ fsr = serial_in(up, UART_REG_FSR);
+ return (fsr & (FSR_TE_FLAG | FSR_TX_EMPTY)) ==
+ (FSR_TE_FLAG | FSR_TX_EMPTY) ? TIOCSER_TEMT : 0;
+}
+
+static u32 ma35d1serial_get_mctrl(struct uart_port *port)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+ u32 status;
+ u32 ret = 0;
+
+ status = serial_in(up, UART_REG_MSR);
+ if (!(status & MSR_CTSSTS))
+ ret |= TIOCM_CTS;
+ return ret;
+}
+
+static void ma35d1serial_set_mctrl(struct uart_port *port, u32 mctrl)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+ u32 mcr = 0;
+ u32 ier = 0;
+
+ if (mctrl & TIOCM_RTS) {
+ mcr = serial_in(up, UART_REG_MCR);
+ mcr |= MCR_RTSACTLV;
+ mcr &= ~MCR_RTS_CTRL;
+ }
+ if (up->mcr & UART_MCR_AFE) {
+ mcr = serial_in(up, UART_REG_MCR);
+ mcr |= MCR_RTSACTLV;
+ mcr &= ~MCR_RTS_CTRL;
+
+ serial_out(up, UART_REG_IER,
+ (serial_in(up, UART_REG_IER) | IER_AUTO_RTS | IER_AUTO_CTS));
+
+ up->port.flags |= UPF_HARD_FLOW;
+ } else {
+ ier = serial_in(up, UART_REG_IER);
+ ier &= ~(IER_AUTO_RTS | IER_AUTO_CTS);
+ serial_out(up, UART_REG_IER, ier);
+
+ up->port.flags &= ~UPF_HARD_FLOW;
+ }
+ serial_out(up, UART_REG_MSR, (serial_in(up, UART_REG_MSR) | MSR_CTSACTLV));
+ serial_out(up, UART_REG_MCR, mcr);
+}
+
+static void ma35d1serial_break_ctl(struct uart_port *port, int break_state)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+ unsigned long flags;
+ u32 lcr;
+
+ spin_lock_irqsave(&up->port.lock, flags);
+ lcr = serial_in(up, UART_REG_LCR);
+ if (break_state != 0)
+ lcr |= LCR_BREAK;
+ else
+ lcr &= ~LCR_BREAK;
+ serial_out(up, UART_REG_LCR, lcr);
+ spin_unlock_irqrestore(&up->port.lock, flags);
+}
+
+static int ma35d1serial_startup(struct uart_port *port)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+ int retval;
+
+ /* Reset FIFO */
+ serial_out(up, UART_REG_FCR, FCR_TFR | FCR_RFR);
+
+ /* Clear pending interrupts */
+ serial_out(up, UART_REG_ISR, ISR_ALL);
+
+ retval = request_irq(port->irq, ma35d1serial_interrupt, 0,
+ dev_name(port->dev), port);
+ if (retval) {
+ dev_err(up->port.dev, "request irq failed.\n");
+ return retval;
+ }
+
+ serial_out(up, UART_REG_FCR, serial_in(up, UART_REG_FCR) |
+ FCR_RFITL_4BYTES | FCR_RTSTRGLV_8BYTES);
+ serial_out(up, UART_REG_LCR, LCR_WLS_8BITS);
+ serial_out(up, UART_REG_TOR, UART_RX_TOUT);
+ serial_out(up, UART_REG_IER, IER_RTO_IEN | IER_RDA_IEN |
+ IER_TIME_OUT_EN | IER_BUFERR_IEN);
+ return 0;
+}
+
+static void ma35d1serial_shutdown(struct uart_port *port)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+
+ serial_out(up, UART_REG_IER, 0);
+ free_irq(port->irq, port);
+}
+
+static u32 ma35d1serial_get_divisor(struct uart_port *port, u32 baud)
+{
+ return (port->uartclk / baud) - 2;
+}
+
+static void ma35d1serial_set_termios(struct uart_port *port,
+ struct ktermios *termios,
+ const struct ktermios *old)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+ u32 lcr = 0;
+ unsigned long flags;
+ u32 baud, quot;
+
+ lcr = UART_LCR_WLEN(tty_get_char_size(termios->c_cflag));
+
+ if (termios->c_cflag & CSTOPB)
+ lcr |= LCR_NSB;
+ if (termios->c_cflag & PARENB)
+ lcr |= LCR_PBE;
+ if (!(termios->c_cflag & PARODD))
+ lcr |= LCR_EPE;
+ if (termios->c_cflag & CMSPAR)
+ lcr |= LCR_SPE;
+
+ baud = uart_get_baud_rate(port, termios, old, port->uartclk / 0xffff,
+ port->uartclk / 11);
+
+ quot = ma35d1serial_get_divisor(port, baud);
+
+ /*
+ * Ok, we're now changing the port state. Do it with
+ * interrupts disabled.
+ */
+ spin_lock_irqsave(&up->port.lock, flags);
+
+ up->port.read_status_mask = FSR_RX_OVER_IF;
+ if (termios->c_iflag & INPCK)
+ up->port.read_status_mask |= FSR_FEF | FSR_PEF;
+ if (termios->c_iflag & (BRKINT | PARMRK))
+ up->port.read_status_mask |= FSR_BIF;
+
+ /*
+ * Characteres to ignore
+ */
+ up->port.ignore_status_mask = 0;
+ if (termios->c_iflag & IGNPAR)
+ up->port.ignore_status_mask |= FSR_FEF | FSR_PEF;
+ if (termios->c_iflag & IGNBRK) {
+ up->port.ignore_status_mask |= FSR_BIF;
+ /*
+ * If we're ignoring parity and break indicators,
+ * ignore overruns too (for real raw support).
+ */
+ if (termios->c_iflag & IGNPAR)
+ up->port.ignore_status_mask |= FSR_RX_OVER_IF;
+ }
+ if (termios->c_cflag & CRTSCTS)
+ up->mcr |= UART_MCR_AFE;
+ else
+ up->mcr &= ~UART_MCR_AFE;
+
+ uart_update_timeout(port, termios->c_cflag, baud);
+ ma35d1serial_set_mctrl(&up->port, up->port.mctrl);
+ serial_out(up, UART_REG_BAUD, quot | BAUD_MODE2);
+ serial_out(up, UART_REG_LCR, lcr);
+ spin_unlock_irqrestore(&up->port.lock, flags);
+}
+
+static const char *ma35d1serial_type(struct uart_port *port)
+{
+ return "ma35d1-uart";
+}
+
+static void ma35d1serial_config_port(struct uart_port *port, int flags)
+{
+ /*
+ * Driver core for serial ports forces a non-zero value for port type.
+ * Write an arbitrary value here to accommodate the serial core driver,
+ * as ID part of UAPI is redundant.
+ */
+ port->type = 1;
+}
+
+static int ma35d1serial_verify_port(struct uart_port *port,
+ struct serial_struct *ser)
+{
+ if (port->type != PORT_UNKNOWN && ser->type != 1)
+ return -EINVAL;
+
+ return 0;
+}
+
+static const struct uart_ops ma35d1serial_ops = {
+ .tx_empty = ma35d1serial_tx_empty,
+ .set_mctrl = ma35d1serial_set_mctrl,
+ .get_mctrl = ma35d1serial_get_mctrl,
+ .stop_tx = ma35d1serial_stop_tx,
+ .start_tx = ma35d1serial_start_tx,
+ .stop_rx = ma35d1serial_stop_rx,
+ .break_ctl = ma35d1serial_break_ctl,
+ .startup = ma35d1serial_startup,
+ .shutdown = ma35d1serial_shutdown,
+ .set_termios = ma35d1serial_set_termios,
+ .type = ma35d1serial_type,
+ .config_port = ma35d1serial_config_port,
+ .verify_port = ma35d1serial_verify_port,
+};
+
+static const struct of_device_id ma35d1_serial_of_match[] = {
+ { .compatible = "nuvoton,ma35d1-uart" },
+ {},
+};
+MODULE_DEVICE_TABLE(of, ma35d1_serial_of_match);
+
+#ifdef CONFIG_SERIAL_NUVOTON_MA35D1_CONSOLE
+
+static void wait_for_xmitr(struct uart_ma35d1_port *up)
+{
+ unsigned int tmout;
+
+ /* Wait up to 10ms for the character(s) to be sent. */
+ tmout = 10000;
+ while (--tmout) {
+ if (serial_in(up, UART_REG_FSR) & FSR_TX_EMPTY)
+ break;
+ udelay(1);
+ }
+}
+
+static void ma35d1serial_console_putchar(struct uart_port *port, unsigned char ch)
+{
+ struct uart_ma35d1_port *up = to_ma35d1_uart_port(port);
+
+ wait_for_xmitr(up);
+ serial_out(up, UART_REG_THR, ch);
+}
+
+/*
+ * Print a string to the serial port trying not to disturb
+ * any possible real use of the port...
+ *
+ * The console_lock must be held when we get here.
+ */
+static void ma35d1serial_console_write(struct console *co,
+ const char *s, u32 count)
+{
+ struct uart_ma35d1_port *up = &ma35d1serial_ports[co->index];
+ unsigned long flags;
+ u32 ier;
+
+ local_irq_save(flags);
+
+ /*
+ * First save the IER then disable the interrupts
+ */
+ ier = serial_in(up, UART_REG_IER);
+ serial_out(up, UART_REG_IER, 0);
+
+ uart_console_write(&up->port, s, count, ma35d1serial_console_putchar);
+
+ wait_for_xmitr(up);
+
+ serial_out(up, UART_REG_IER, ier);
+ local_irq_restore(flags);
+}
+
+static int __init ma35d1serial_console_setup(struct console *co,
+ char *options)
+{
+ struct device_node *np = ma35d1serial_uart_nodes[co->index];
+ struct uart_ma35d1_port *p = &ma35d1serial_ports[co->index];
+ u32 val32[4];
+ struct uart_port *port;
+ int baud = 115200;
+ int bits = 8;
+ int parity = 'n';
+ int flow = 'n';
+
+ /*
+ * Check whether an invalid uart number has been specified, and
+ * if so, search for the first available port that does have
+ * console support.
+ */
+ if ((co->index < 0) || (co->index >= UART_NR)) {
+ pr_debug("Console Port%x out of range\n", co->index);
+ return -EINVAL;
+ }
+
+ if (of_property_read_u32_array(np, "reg", val32, 4) != 0)
+ return -EINVAL;
+ p->port.iobase = val32[1];
+ p->port.membase = ioremap(p->port.iobase, UART_REG_SIZE);
+ p->port.ops = &ma35d1serial_ops;
+ p->port.line = 0;
+ p->port.uartclk = UART_CONSOLE_CLK;
+
+ port = &ma35d1serial_ports[co->index].port;
+ return uart_set_options(port, co, baud, parity, bits, flow);
+}
+
+static struct console ma35d1serial_console = {
+ .name = "ttyS",
+ .write = ma35d1serial_console_write,
+ .device = uart_console_device,
+ .setup = ma35d1serial_console_setup,
+ .flags = CON_PRINTBUFFER | CON_ENABLED,
+ .index = -1,
+ .data = &ma35d1serial_reg,
+};
+
+static void ma35d1serial_console_init_port(void)
+{
+ int i = 0;
+ struct device_node *np;
+
+ for_each_matching_node(np, ma35d1_serial_of_match) {
+ if (ma35d1serial_uart_nodes[i] == NULL) {
+ of_node_get(np);
+ ma35d1serial_uart_nodes[i] = np;
+ i++;
+ if (i == UART_NR)
+ break;
+ }
+ }
+}
+
+static int __init ma35d1serial_console_init(void)
+{
+ ma35d1serial_console_init_port();
+ register_console(&ma35d1serial_console);
+ return 0;
+}
+console_initcall(ma35d1serial_console_init);
+
+#define MA35D1SERIAL_CONSOLE (&ma35d1serial_console)
+#else
+#define MA35D1SERIAL_CONSOLE NULL
+#endif
+
+static struct uart_driver ma35d1serial_reg = {
+ .owner = THIS_MODULE,
+ .driver_name = "serial",
+ .dev_name = "ttyS",
+ .major = TTY_MAJOR,
+ .minor = 64,
+ .cons = MA35D1SERIAL_CONSOLE,
+ .nr = UART_NR,
+};
+
+/*
+ * Register a set of serial devices attached to a platform device.
+ * The list is terminated with a zero flags entry, which means we expect
+ * all entries to have at least UPF_BOOT_AUTOCONF set.
+ */
+static int ma35d1serial_probe(struct platform_device *pdev)
+{
+ struct resource *res_mem;
+ struct uart_ma35d1_port *up;
+ int ret;
+ struct clk *clk;
+ int err;
+
+ if (pdev->dev.of_node) {
+ ret = of_alias_get_id(pdev->dev.of_node, "serial");
+ if (ret < 0) {
+ dev_err(&pdev->dev, "failed to get alias/pdev id, errno %d\n", ret);
+ return ret;
+ }
+ }
+ up = &ma35d1serial_ports[ret];
+ up->port.line = ret;
+ res_mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ if (!res_mem)
+ return -ENODEV;
+
+ up->port.iobase = res_mem->start;
+ up->port.membase = ioremap(up->port.iobase, UART_REG_SIZE);
+ up->port.ops = &ma35d1serial_ops;
+
+ spin_lock_init(&up->port.lock);
+
+ clk = of_clk_get(pdev->dev.of_node, 0);
+ if (IS_ERR(clk)) {
+ err = PTR_ERR(clk);
+ dev_err(&pdev->dev, "failed to get core clk: %d\n", err);
+ return -ENOENT;
+ }
+ err = clk_prepare_enable(clk);
+ if (err)
+ return -ENOENT;
+
+ if (up->port.line != 0)
+ up->port.uartclk = clk_get_rate(clk);
+ up->port.irq = platform_get_irq(pdev, 0);
+ up->port.dev = &pdev->dev;
+ up->port.flags = UPF_BOOT_AUTOCONF;
+ ret = uart_add_one_port(&ma35d1serial_reg, &up->port);
+ platform_set_drvdata(pdev, up);
+ return 0;
+}
+
+/*
+ * Remove serial ports registered against a platform device.
+ */
+static int ma35d1serial_remove(struct platform_device *dev)
+{
+ struct uart_port *port = platform_get_drvdata(dev);
+
+ if (port) {
+ uart_remove_one_port(&ma35d1serial_reg, port);
+ free_irq(port->irq, port);
+ }
+ return 0;
+}
+
+static int ma35d1serial_suspend(struct platform_device *dev, pm_message_t state)
+{
+ int i;
+ struct uart_ma35d1_port *up;
+
+ if (dev->dev.of_node)
+ i = of_alias_get_id(dev->dev.of_node, "serial");
+ if (i < 0) {
+ dev_err(&dev->dev, "failed to get alias/pdev id, errno %d\n", i);
+ return i;
+ }
+ up = &ma35d1serial_ports[i];
+ if (i == 0) {
+ up->console_baud_rate = serial_in(up, UART_REG_BAUD);
+ up->console_line = serial_in(up, UART_REG_LCR);
+ up->console_int = serial_in(up, UART_REG_IER);
+ }
+ return 0;
+}
+
+static int ma35d1serial_resume(struct platform_device *dev)
+{
+ int i;
+ struct uart_ma35d1_port *up;
+
+ if (dev->dev.of_node)
+ i = of_alias_get_id(dev->dev.of_node, "serial");
+ if (i < 0) {
+ dev_err(&dev->dev, "failed to get alias/pdev id, errno %d\n", i);
+ return i;
+ }
+ up = &ma35d1serial_ports[i];
+ if (i == 0) {
+ serial_out(up, UART_REG_BAUD, up->console_baud_rate);
+ serial_out(up, UART_REG_LCR, up->console_line);
+ serial_out(up, UART_REG_IER, up->console_int);
+ }
+ return 0;
+}
+
+static struct platform_driver ma35d1serial_driver = {
+ .probe = ma35d1serial_probe,
+ .remove = ma35d1serial_remove,
+ .suspend = ma35d1serial_suspend,
+ .resume = ma35d1serial_resume,
+ .driver = {
+ .name = "ma35d1-uart",
+ .owner = THIS_MODULE,
+ .of_match_table = of_match_ptr(ma35d1_serial_of_match),
+ },
+};
+
+static int __init ma35d1serial_init(void)
+{
+ int ret;
+
+ ret = uart_register_driver(&ma35d1serial_reg);
+ if (ret)
+ return ret;
+ ret = platform_driver_register(&ma35d1serial_driver);
+ if (ret)
+ uart_unregister_driver(&ma35d1serial_reg);
+ return ret;
+}
+
+static void __exit ma35d1serial_exit(void)
+{
+ platform_driver_unregister(&ma35d1serial_driver);
+ uart_unregister_driver(&ma35d1serial_reg);
+}
+
+module_init(ma35d1serial_init);
+module_exit(ma35d1serial_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("MA35D1 serial driver");
+MODULE_ALIAS_CHARDEV_MAJOR(TTY_MAJOR);
+
--
2.34.1

2023-03-28 09:26:15

by Jiri Slaby

[permalink] [raw]
Subject: Re: [PATCH v6 11/12] tty: serial: Add Nuvoton ma35d1 serial driver support

On 28. 03. 23, 4:19, Jacky Huang wrote:
> +static void transmit_chars(struct uart_ma35d1_port *up)
> +{
> + struct circ_buf *xmit = &up->port.state->xmit;
> + int count;
> + u8 ch;
> +
> + if (uart_tx_stopped(&up->port)) {
> + ma35d1serial_stop_tx(&up->port);
> + return;
> + }
> + if (uart_circ_empty(xmit)) {
> + __stop_tx(up);
> + return;
> + }

Why is this necessary?

> + count = UART_FIFO_DEPTH - ((serial_in(up, UART_REG_FSR) & FSR_TXPTR_MSK) >> 16);
> +
> + uart_port_tx_limited(&up->port, ch, count,
> + !(serial_in(up, UART_REG_FSR) & FSR_TX_FULL),
> + serial_out(up, UART_REG_THR, ch),
> + ({}));
> +
> + if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
> + uart_write_wakeup(&up->port);
> +
> + if (uart_circ_empty(xmit))
> + __stop_tx(up);

uart_port_tx_limited() should take care about the above and this too, right?

> +}
...
> +static void receive_chars(struct uart_ma35d1_port *up)
> +{
> + u8 flag;
> + u32 fsr;
> + unsigned int ch;

Shouldn't ch be u8 too?

> + int max_count = 256;
> +
> + fsr = serial_in(up, UART_REG_FSR);
> + do {
> + flag = TTY_NORMAL;
> + up->port.icount.rx++;
> +
> + if (unlikely(fsr & (FSR_BIF | FSR_FEF | FSR_PEF | FSR_RX_OVER_IF))) {
> + if (fsr & FSR_BIF) {
> + up->port.icount.brk++;
> + if (uart_handle_break(&up->port))
> + continue;
> + }
> + if (fsr & FSR_FEF)
> + up->port.icount.frame++;
> + if (fsr & FSR_PEF)
> + up->port.icount.parity++;
> + if (fsr & FSR_RX_OVER_IF)
> + up->port.icount.overrun++;
> +
> + serial_out(up, UART_REG_FSR, fsr &
> + (FSR_BIF | FSR_FEF | FSR_PEF | FSR_RX_OVER_IF));
> +
> + if (fsr & FSR_BIF)
> + flag = TTY_BREAK;
> + else if (fsr & FSR_PEF)
> + flag = TTY_PARITY;
> + else if (fsr & FSR_FEF)
> + flag = TTY_FRAME;
> + }
> +
> + ch = serial_in(up, UART_REG_RBR);
> + if (uart_handle_sysrq_char(&up->port, ch))
> + continue;
> +
> + spin_lock(&up->port.lock);
> + uart_insert_char(&up->port, fsr, FSR_RX_OVER_IF, ch, flag);
> + spin_unlock(&up->port.lock);
> +
> + fsr = serial_in(up, UART_REG_FSR);
> + } while (!(fsr & FSR_RX_EMPTY) && (max_count-- > 0));
> +
> + spin_lock(&up->port.lock);
> + tty_flip_buffer_push(&up->port.state->port);
> + spin_unlock(&up->port.lock);
> +}
...
> +static int ma35d1serial_remove(struct platform_device *dev)
> +{
> + struct uart_port *port = platform_get_drvdata(dev);
> +
> + if (port) {

Can this ever be NULL?

> + uart_remove_one_port(&ma35d1serial_reg, port);
> + free_irq(port->irq, port);
> + }
> + return 0;
> +}
> +
> +static int ma35d1serial_suspend(struct platform_device *dev, pm_message_t state)
> +{
> + int i;
> + struct uart_ma35d1_port *up;
> +
> + if (dev->dev.of_node)
> + i = of_alias_get_id(dev->dev.of_node, "serial");
> + if (i < 0) {
> + dev_err(&dev->dev, "failed to get alias/pdev id, errno %d\n", i);
> + return i;
> + }
> + up = &ma35d1serial_ports[i];

platform_get_drvdata(dev) ?

> + if (i == 0) {
> + up->console_baud_rate = serial_in(up, UART_REG_BAUD);
> + up->console_line = serial_in(up, UART_REG_LCR);
> + up->console_int = serial_in(up, UART_REG_IER);
> + }
> + return 0;
> +}
> +
> +static int ma35d1serial_resume(struct platform_device *dev)
> +{
> + int i;
> + struct uart_ma35d1_port *up;
> +
> + if (dev->dev.of_node)
> + i = of_alias_get_id(dev->dev.of_node, "serial");
> + if (i < 0) {
> + dev_err(&dev->dev, "failed to get alias/pdev id, errno %d\n", i);
> + return i;
> + }
> + up = &ma35d1serial_ports[i];
> + if (i == 0) {
> + serial_out(up, UART_REG_BAUD, up->console_baud_rate);
> + serial_out(up, UART_REG_LCR, up->console_line);
> + serial_out(up, UART_REG_IER, up->console_int);
> + }
> + return 0;
> +}

No uart_suspend_port()/uart_resume_port()? You don't wait for
transmitter to be empty in suspend. You don't stop tx etc.

> +static struct platform_driver ma35d1serial_driver = {
> + .probe = ma35d1serial_probe,
> + .remove = ma35d1serial_remove,
> + .suspend = ma35d1serial_suspend,
> + .resume = ma35d1serial_resume,
> + .driver = {
> + .name = "ma35d1-uart",
> + .owner = THIS_MODULE,
> + .of_match_table = of_match_ptr(ma35d1_serial_of_match),
> + },
> +};
> +
> +static int __init ma35d1serial_init(void)
> +{
> + int ret;
> +
> + ret = uart_register_driver(&ma35d1serial_reg);
> + if (ret)
> + return ret;
> + ret = platform_driver_register(&ma35d1serial_driver);
> + if (ret)
> + uart_unregister_driver(&ma35d1serial_reg);
> + return ret;
> +}
> +
> +static void __exit ma35d1serial_exit(void)
> +{
> + platform_driver_unregister(&ma35d1serial_driver);
> + uart_unregister_driver(&ma35d1serial_reg);
> +}
> +
> +module_init(ma35d1serial_init);
> +module_exit(ma35d1serial_exit);
> +
> +MODULE_LICENSE("GPL");
> +MODULE_DESCRIPTION("MA35D1 serial driver");


> +MODULE_ALIAS_CHARDEV_MAJOR(TTY_MAJOR);

Why is this needed? How are other platform drivers autoloaded?

thanks,
--
js
suse labs

2023-03-28 15:47:04

by kernel test robot

[permalink] [raw]
Subject: Re: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform

Hi Jacky,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on robh/for-next]
[also build test WARNING on clk/clk-next linus/master pza/reset/next v6.3-rc4 next-20230328]
[cannot apply to pza/imx-drm/next]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url: https://github.com/intel-lab-lkp/linux/commits/Jacky-Huang/arm64-Kconfig-platforms-Add-config-for-Nuvoton-MA35-platform/20230328-102245
base: https://git.kernel.org/pub/scm/linux/kernel/git/robh/linux.git for-next
patch link: https://lore.kernel.org/r/20230328021912.177301-7-ychuang570808%40gmail.com
patch subject: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform
reproduce:
# https://github.com/intel-lab-lkp/linux/commit/d023e63cdc203398c5db3c81ca76d34d22e23fe5
git remote add linux-review https://github.com/intel-lab-lkp/linux
git fetch --no-tags linux-review Jacky-Huang/arm64-Kconfig-platforms-Add-config-for-Nuvoton-MA35-platform/20230328-102245
git checkout d023e63cdc203398c5db3c81ca76d34d22e23fe5
make menuconfig
# enable CONFIG_COMPILE_TEST, CONFIG_WARN_MISSING_DOCUMENTS, CONFIG_WARN_ABI_ERRORS
make htmldocs

If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <[email protected]>
| Link: https://lore.kernel.org/oe-kbuild-all/[email protected]/

All warnings (new ones prefixed by >>):

>> Warning: MAINTAINERS references a file that doesn't exist: Documentation/devicetree/bindings/arm/npcm/*

--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests

2023-03-28 18:10:44

by Stephen Boyd

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Quoting Jacky Huang (2023-03-27 19:19:08)
> diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
> new file mode 100644
> index 000000000000..0740b0b218a7
> --- /dev/null
> +++ b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
> @@ -0,0 +1,231 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Copyright (C) 2023 Nuvoton Technology Corp.
> + * Author: Shan-Chun Hung <[email protected]>
> + * Jacky huang <[email protected]>
> + */
> +
> +#include <dt-bindings/interrupt-controller/arm-gic.h>
> +#include <dt-bindings/input/input.h>
> +#include <dt-bindings/gpio/gpio.h>
> +#include <dt-bindings/clock/nuvoton,ma35d1-clk.h>
> +#include <dt-bindings/reset/nuvoton,ma35d1-reset.h>
> +
> +/ {
> + compatible = "nuvoton,ma35d1";
> + interrupt-parent = <&gic>;
> + #address-cells = <2>;
> + #size-cells = <2>;
> +
> + cpus {
> + #address-cells = <2>;
> + #size-cells = <0>;
> +
> + cpu0: cpu@0 {
> + device_type = "cpu";
> + compatible = "arm,cortex-a35";
> + reg = <0x0 0x0>;
> + enable-method = "psci";
> + next-level-cache = <&L2_0>;
> + };
> +
> + cpu1: cpu@1 {
> + device_type = "cpu";
> + compatible = "arm,cortex-a35";
> + reg = <0x0 0x1>;
> + enable-method = "psci";
> + next-level-cache = <&L2_0>;
> + };
> +
> + L2_0: l2-cache0 {

Just l2-cache for the node name. Doesn't it go under the cpu0 node as
well?

> + compatible = "cache";
> + cache-level = <2>;
> + };
> + };
> +
> + psci {
> + compatible = "arm,psci-0.2";
> + method = "smc";
> + };
> +
> + timer {
> + compatible = "arm,armv8-timer";
> + interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(4) |
> + IRQ_TYPE_LEVEL_LOW)>, /* Physical Secure */
> + <GIC_PPI 14 (GIC_CPU_MASK_SIMPLE(4) |
> + IRQ_TYPE_LEVEL_LOW)>, /* Physical Non-Secure */
> + <GIC_PPI 11 (GIC_CPU_MASK_SIMPLE(4) |
> + IRQ_TYPE_LEVEL_LOW)>, /* Virtual */
> + <GIC_PPI 10 (GIC_CPU_MASK_SIMPLE(4) |
> + IRQ_TYPE_LEVEL_LOW)>; /* Hypervisor */
> + clock-frequency = <12000000>;

Remove this property. The frequency should be read by the driver.

> + interrupt-parent = <&gic>;
> + };

Please create an 'soc' node for the SoC to hold all the nodes that have
a reg property.

> +
> + sys: system-management@40460000 {
> + compatible = "nuvoton,ma35d1-sys", "syscon", "simple-mfd";
> + reg = <0x0 0x40460000 0x0 0x200>;
> +
> + reset: reset-controller {
> + compatible = "nuvoton,ma35d1-reset";
> + #reset-cells = <1>;
> + };
> + };
> +
> + clk: clock-controller@40460200 {
> + compatible = "nuvoton,ma35d1-clk", "syscon";
> + reg = <0x00000000 0x40460200 0x0 0x100>;
> + #clock-cells = <1>;
> + clocks = <&clk_hxt>;
> + nuvoton,sys = <&sys>;
> + };

It looks like the device at 40460000 is a reset and clock controller.
Just make it one node and register the clk or reset device as an
auxiliary device.

> +
> + gic: interrupt-controller@50801000 {
> + compatible = "arm,gic-400";
> + reg = <0x0 0x50801000 0 0x1000>, /* GICD */
> + <0x0 0x50802000 0 0x2000>, /* GICC */
> + <0x0 0x50804000 0 0x2000>, /* GICH */
> + <0x0 0x50806000 0 0x2000>; /* GICV */
> + #interrupt-cells = <3>;
> + interrupt-parent = <&gic>;
> + interrupt-controller;
> + interrupts = <GIC_PPI 9 (GIC_CPU_MASK_RAW(0x13) |
> + IRQ_TYPE_LEVEL_HIGH)>;
> + };
> +
> + uart0:serial@40700000 {

Add a space after :

> + compatible = "nuvoton,ma35d1-uart";
> + reg = <0x0 0x40700000 0x0 0x100>;
> + interrupts = <GIC_SPI 59 IRQ_TYPE_LEVEL_HIGH>;
> + clocks = <&clk UART0_GATE>;
> + status = "disabled";
> + };

2023-03-28 18:38:49

by Stephen Boyd

[permalink] [raw]
Subject: Re: [PATCH v6 09/12] clk: nuvoton: Add clock driver for ma35d1 clock controller

Quoting Jacky Huang (2023-03-27 19:19:09)
> diff --git a/drivers/clk/nuvoton/Kconfig b/drivers/clk/nuvoton/Kconfig
> new file mode 100644
> index 000000000000..c1324efedcb9
> --- /dev/null
> +++ b/drivers/clk/nuvoton/Kconfig
> @@ -0,0 +1,19 @@
> +# SPDX-License-Identifier: GPL-2.0
> +# common clock support for Nuvoton SoC family.
> +
> +config COMMON_CLK_NUVOTON
> + bool "Nuvoton clock controller common support"
> + depends on ARCH_NUVOTON

Add || COMPILE_TEST

> + default ARCH_NUVOTON
> + help
> + Say y here to enable common clock controller for Nuvoton platforms.
> +
> +if COMMON_CLK_NUVOTON

Add a newline here.

> +config CLK_MA35D1
> + bool "Nuvoton MA35D1 clock controller support"
> + depends on ARM64 || COMPILE_TEST

Drop this

> + default y

And this.

> + help
> + Build the driver for MA35D1 Clock Driver.

This says driver twice. "Build the clk driver for MA35D1 SoCs"?

> +
> +endif
> diff --git a/drivers/clk/nuvoton/clk-ma35d1-divider.c b/drivers/clk/nuvoton/clk-ma35d1-divider.c
> new file mode 100644
> index 000000000000..340a889a1417
> --- /dev/null
> +++ b/drivers/clk/nuvoton/clk-ma35d1-divider.c
> @@ -0,0 +1,140 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Copyright (C) 2023 Nuvoton Technology Corp.
> + * Author: Chi-Fang Li <[email protected]>
> + */
> +
> +#include <linux/clk-provider.h>
> +#include <linux/device.h>
> +#include <linux/regmap.h>
> +#include <linux/spinlock.h>
> +
> +#include "clk-ma35d1.h"
> +
> +#define clk_div_mask(width) ((1 << (width)) - 1)

clk-provider.h defines this already, use that?

> +
> +struct ma35d1_adc_clk_div {
> + struct clk_hw hw;
> + void __iomem *reg;
> + u8 shift;
> + u8 width;
> + u32 mask;
> + const struct clk_div_table *table;
> + /* protects concurrent access to clock divider registers */
> + spinlock_t *lock;
> +};
> +
> +static inline struct ma35d1_adc_clk_div *to_ma35d1_adc_clk_div(struct clk_hw *_hw)
> +{
> + return container_of(_hw, struct ma35d1_adc_clk_div, hw);
> +}
> +
> +static inline unsigned long ma35d1_clkdiv_recalc_rate(struct clk_hw *hw,
> + unsigned long parent_rate)
> +{
> + unsigned int val;
> + struct ma35d1_adc_clk_div *dclk = to_ma35d1_adc_clk_div(hw);
> +
> + val = readl_relaxed(dclk->reg) >> dclk->shift;
> + val &= clk_div_mask(dclk->width);
> + val += 1;
> + return divider_recalc_rate(hw, parent_rate, val, dclk->table,
> + CLK_DIVIDER_ROUND_CLOSEST, dclk->width);
> +}
> +
> +static inline long ma35d1_clkdiv_round_rate(struct clk_hw *hw,
> + unsigned long rate,
> + unsigned long *prate)
> +{
> + struct ma35d1_adc_clk_div *dclk = to_ma35d1_adc_clk_div(hw);
> +
> + return divider_round_rate(hw, rate, prate, dclk->table,
> + dclk->width, CLK_DIVIDER_ROUND_CLOSEST);
> +}
> +
> +static inline int ma35d1_clkdiv_set_rate(struct clk_hw *hw,
> + unsigned long rate,
> + unsigned long parent_rate)
> +{
> + int value;
> + unsigned long flags = 0;
> + u32 data;
> + struct ma35d1_adc_clk_div *dclk = to_ma35d1_adc_clk_div(hw);
> +
> + value = divider_get_val(rate, parent_rate, dclk->table,
> + dclk->width, CLK_DIVIDER_ROUND_CLOSEST);
> +
> + if (dclk->lock)
> + spin_lock_irqsave(dclk->lock, flags);

Can you just always lock? That makes the conditional locking go away.

> +
> + data = readl_relaxed(dclk->reg);
> + data &= ~(clk_div_mask(dclk->width) << dclk->shift);
> + data |= (value - 1) << dclk->shift;
> + data |= dclk->mask;
> +
> + writel_relaxed(data, dclk->reg);
> +
> + if (dclk->lock)
> + spin_unlock_irqrestore(dclk->lock, flags);
> +
> + return 0;
> +}
> +
> +static const struct clk_ops ma35d1_adc_clkdiv_ops = {
> + .recalc_rate = ma35d1_clkdiv_recalc_rate,
> + .round_rate = ma35d1_clkdiv_round_rate,
> + .set_rate = ma35d1_clkdiv_set_rate,
> +};
> +
> +struct clk_hw *ma35d1_reg_adc_clkdiv(struct device *dev, const char *name,
> + const char *parent_name,
> + spinlock_t *lock,
> + unsigned long flags, void __iomem *reg,
> + u8 shift, u8 width, u32 mask_bit)
> +{
> + struct ma35d1_adc_clk_div *div;
> + struct clk_init_data init;
> + struct clk_div_table *table;
> + u32 max_div, min_div;
> + struct clk_hw *hw;
> + int ret;
> + int i;
> +
> + div = devm_kzalloc(dev, sizeof(*div), GFP_KERNEL);
> + if (!div)
> + return ERR_PTR(-ENOMEM);
> +
> + max_div = clk_div_mask(width) + 1;
> + min_div = 1;
> +
> + table = devm_kcalloc(dev, max_div + 1, sizeof(*table), GFP_KERNEL);
> + if (!table)
> + return ERR_PTR(-ENOMEM);
> +
> + for (i = 0; i < max_div; i++) {
> + table[i].val = (min_div + i);

Drop useless parenthesis.

> + table[i].div = 2 * table[i].val;
> + }
> + table[max_div].val = 0;
> + table[max_div].div = 0;
> +
> + init.name = name;
> + init.ops = &ma35d1_adc_clkdiv_ops;
> + init.flags |= flags;
> + init.parent_names = parent_name ? &parent_name : NULL;
> + init.num_parents = parent_name ? 1 : 0;
> +
> + div->reg = reg;
> + div->shift = shift;
> + div->width = width;
> + div->mask = mask_bit ? BIT(mask_bit) : 0;
> + div->lock = lock;
> + div->hw.init = &init;
> + div->table = table;
> +
> + hw = &div->hw;
> + ret = devm_clk_hw_register(dev, hw);
> + if (ret < 0)

Use if (ret) instead.

> + return ERR_PTR(ret);
> + return hw;
> +}
> diff --git a/drivers/clk/nuvoton/clk-ma35d1-pll.c b/drivers/clk/nuvoton/clk-ma35d1-pll.c
> new file mode 100644
> index 000000000000..b36fbda4fa0a
> --- /dev/null
> +++ b/drivers/clk/nuvoton/clk-ma35d1-pll.c
> @@ -0,0 +1,350 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Copyright (C) 2023 Nuvoton Technology Corp.
> + * Author: Chi-Fang Li <[email protected]>
> + */
> +
> +#include <linux/bitfield.h>
> +#include <linux/clk-provider.h>
> +#include <linux/slab.h>
> +#include <linux/regmap.h>
> +
> +#include "clk-ma35d1.h"
> +
> +#define REG_SYS_RLKTZNS 0x1a4 /* Register Lock Control */
> +
> +struct ma35d1_clk_pll {
> + struct clk_hw hw;
> + u8 type;
> + u8 mode;
> + unsigned long rate;
> + void __iomem *ctl0_base;

This needs an include for __iomem define. Just include kernel.h for now.

> + void __iomem *ctl1_base;
> + void __iomem *ctl2_base;
> + struct regmap *regmap;
> +};
> +
> +static inline struct ma35d1_clk_pll *to_ma35d1_clk_pll(struct clk_hw *_hw)
> +{
> + return container_of(_hw, struct ma35d1_clk_pll, hw);

Missing include as well.

> +}
> +
> +static void ma35d1_unlock_regs(struct ma35d1_clk_pll *pll)
> +{
> + int ret;
> +
> + do {
> + regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x59);
> + regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x16);
> + regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x88);
> + regmap_read(pll->regmap, REG_SYS_RLKTZNS, &ret);
> + } while (ret == 0);

Any sort of timeout?

> +}
> +
> +static void ma35d1_lock_regs(struct ma35d1_clk_pll *pll)
> +{
> + regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x0);
> +}
> +
> +static unsigned long ma35d1_calc_smic_pll_freq(u32 pll0_ctl0,
> + unsigned long parent_rate)
> +{
> + u32 m, n, p, outdiv;
> + u64 pll_freq;
> + u32 clk_div_table[] = { 1, 2, 4, 8 };

const

> +
> + if (pll0_ctl0 & SPLL0_CTL0_BP)
> + return parent_rate;
> +
> + n = FIELD_GET(SPLL0_CTL0_FBDIV, pll0_ctl0);
> + m = FIELD_GET(SPLL0_CTL0_INDIV, pll0_ctl0);
> + p = FIELD_GET(SPLL0_CTL0_OUTDIV, pll0_ctl0);
> + outdiv = clk_div_table[p];
> + pll_freq = (u64)parent_rate * n;
> + do_div(pll_freq, m * outdiv);
> + return (unsigned long)pll_freq;

Remove useless cast.

> +}
> +
> +static unsigned long ma35d1_calc_pll_freq(u8 mode, u32 *reg_ctl,
> + unsigned long parent_rate)
> +{
> + u32 m, n, p;
> + u64 pll_freq, x;
> +
> + if (reg_ctl[1] & PLL_CTL1_BP)
> + return parent_rate;
> +
> + if (mode == PLL_MODE_INT) {
> + n = FIELD_GET(PLL_CTL0_FBDIV, reg_ctl[0]);
> + m = FIELD_GET(PLL_CTL0_INDIV, reg_ctl[0]);
> + p = FIELD_GET(PLL_CTL1_OUTDIV, reg_ctl[1]);
> + pll_freq = (u64)parent_rate * n;
> + do_div(pll_freq, m * p);
> + } else {
> + n = FIELD_GET(PLL_CTL0_FBDIV, reg_ctl[0]);
> + m = FIELD_GET(PLL_CTL0_INDIV, reg_ctl[0]);
> + p = FIELD_GET(PLL_CTL1_OUTDIV, reg_ctl[1]);

The n, m, p are the same, so pull them out of the if-else to deduplicate.

> + x = FIELD_GET(PLL_CTL1_FRAC, reg_ctl[1]);
> + /* 2 decimal places floating to integer (ex. 1.23 to 123) */
> + n = n * 100 + ((x * 100) / FIELD_MAX(PLL_CTL1_FRAC));

Is this mult_frac()?

> + pll_freq = ((u64)parent_rate * n) / 100 / m / p;
> + }

Add a newline

> + return (unsigned long)pll_freq;

Remove useless cast.

> +}
> +
> +static int ma35d1_pll_find_closest(struct ma35d1_clk_pll *pll,
> + unsigned long rate,
> + unsigned long parent_rate,
> + u32 *reg_ctl, unsigned long *freq)
> +{
> + int p, m, n;
> + int fbdiv_min, fbdiv_max;
> + unsigned long diff = 0xffffffff;
> +
> + *freq = 0;
> + if ((rate < PLL_FCLKO_MIN_FREQ) || (rate > PLL_FCLKO_MAX_FREQ))
> + return -EINVAL;
> +
> + if (pll->mode == PLL_MODE_INT) {
> + fbdiv_min = FBDIV_MIN;
> + fbdiv_max = FBDIV_MAX;
> + } else {
> + fbdiv_min = FBDIV_FRAC_MIN;
> + fbdiv_max = FBDIV_FRAC_MAX;
> + }
> +
> + for (m = INDIV_MIN; m <= INDIV_MAX; m++) {
> + for (n = fbdiv_min; n <= fbdiv_max; n++) {
> + for (p = OUTDIV_MIN; p <= OUTDIV_MAX; p++) {
> + unsigned long tmp, fout;
> + u64 fclk;
> +
> + tmp = parent_rate / m;
> + if (tmp < PLL_FREF_M_MIN_FREQ ||
> + tmp > PLL_FREF_M_MAX_FREQ)
> + continue; /* constrain */
> +
> + fclk = (u64)parent_rate * n / m;
> + /* for 2 decimal places */
> + if (pll->mode != PLL_MODE_INT)
> + fclk /= 100;
> +
> + if (fclk < PLL_FCLK_MIN_FREQ ||
> + fclk > PLL_FCLK_MAX_FREQ)
> + continue; /* constrain */
> +
> + fout = (unsigned long)(fclk / p);
> + if (fout < PLL_FCLKO_MIN_FREQ ||
> + fout > PLL_FCLKO_MAX_FREQ)
> + continue; /* constrain */
> +
> + if (abs(rate - fout) < diff) {
> + reg_ctl[0] = FIELD_PREP(PLL_CTL0_INDIV, m) |
> + FIELD_PREP(PLL_CTL0_FBDIV, n);
> + reg_ctl[1] = FIELD_PREP(PLL_CTL1_OUTDIV, p);
> + *freq = fout;
> + diff = abs(rate - fout);
> + if (diff == 0)
> + break;
> + }
> + }
> + }
> + }
> + if (*freq == 0)
> + return -EINVAL; /* cannot find even one valid setting */
> + return 0;
> +}
> +
> +static int ma35d1_clk_pll_set_rate(struct clk_hw *hw, unsigned long rate,
> + unsigned long parent_rate)
> +{
> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
> + u32 reg_ctl[3] = { 0 };
> + unsigned long pll_freq;
> + int ret;
> +
> + if ((parent_rate < PLL_FREF_MIN_FREQ) ||
> + (parent_rate > PLL_FREF_MAX_FREQ))

Remove useless parenthesis.

> + return -EINVAL;
> +
> + if ((pll->type == MA35D1_CAPLL) || (pll->type == MA35D1_DDRPLL)) {

Remove useless parenthesis.

> + pr_warn("Nuvoton MA35D1 CAPLL/DDRPLL is read only.\n");
> + return -EACCES;
> + }
> +
> + ret = ma35d1_pll_find_closest(pll, rate, parent_rate, reg_ctl, &pll_freq);
> + if (ret != 0)
> + return ret;
> + pll->rate = pll_freq;
> +
> + switch (pll->mode) {
> + case PLL_MODE_INT:
> + reg_ctl[0] |= FIELD_PREP(PLL_CTL0_MODE, PLL_MODE_INT);
> + break;
> + case PLL_MODE_FRAC:
> + reg_ctl[0] |= FIELD_PREP(PLL_CTL0_MODE, PLL_MODE_FRAC);
> + break;
> + case PLL_MODE_SS:
> + reg_ctl[0] |= FIELD_PREP(PLL_CTL0_MODE, PLL_MODE_SS) |
> + FIELD_PREP(PLL_CTL0_SSRATE, PLL_SS_RATE);
> + reg_ctl[2] = FIELD_PREP(PLL_CTL2_SLOPE, PLL_SLOPE);
> + break;
> + }
> + reg_ctl[1] |= PLL_CTL1_PD;
> +
> + ma35d1_unlock_regs(pll);
> + writel_relaxed(reg_ctl[0], pll->ctl0_base);
> + writel_relaxed(reg_ctl[1], pll->ctl1_base);
> + writel_relaxed(reg_ctl[2], pll->ctl2_base);
> + ma35d1_lock_regs(pll);
> + return 0;
> +}
> +
> +static unsigned long ma35d1_clk_pll_recalc_rate(struct clk_hw *hw,
> + unsigned long parent_rate)
> +{
> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
> + u32 reg_ctl[3];
> + unsigned long pll_freq;
> +
> + if ((parent_rate < PLL_FREF_MIN_FREQ) || (parent_rate > PLL_FREF_MAX_FREQ))

Remove useless parenthesis.

> + return 0;
> +
> + switch (pll->type) {
> + case MA35D1_CAPLL:
> + reg_ctl[0] = readl_relaxed(pll->ctl0_base);
> + pll_freq = ma35d1_calc_smic_pll_freq(reg_ctl[0], parent_rate);

return?

> + break;
> + case MA35D1_DDRPLL:
> + case MA35D1_APLL:
> + case MA35D1_EPLL:
> + case MA35D1_VPLL:
> + reg_ctl[0] = readl_relaxed(pll->ctl0_base);
> + reg_ctl[1] = readl_relaxed(pll->ctl1_base);
> + pll_freq = ma35d1_calc_pll_freq(pll->mode, reg_ctl, parent_rate);

return?

> + break;
> + }
> + return pll_freq;

return 0?

> +}
> +
> +static long ma35d1_clk_pll_round_rate(struct clk_hw *hw, unsigned long rate,
> + unsigned long *parent_rate)
> +{
> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
> + u32 reg_ctl[3] = { 0 };
> + unsigned long pll_freq;
> + long ret;
> +
> + if (*parent_rate < PLL_FREF_MIN_FREQ || *parent_rate > PLL_FREF_MAX_FREQ)
> + return -EINVAL;
> +
> + ret = ma35d1_pll_find_closest(pll, rate, *parent_rate, reg_ctl, &pll_freq);
> + if (ret != 0)
> + return ret;
> +
> + switch (pll->type) {
> + case MA35D1_CAPLL:
> + reg_ctl[0] = readl_relaxed(pll->ctl0_base);
> + pll_freq = ma35d1_calc_smic_pll_freq(reg_ctl[0], *parent_rate);

return?

> + break;
> + case MA35D1_DDRPLL:
> + case MA35D1_APLL:
> + case MA35D1_EPLL:
> + case MA35D1_VPLL:
> + reg_ctl[0] = readl_relaxed(pll->ctl0_base);
> + reg_ctl[1] = readl_relaxed(pll->ctl1_base);
> + pll_freq = ma35d1_calc_pll_freq(pll->mode, reg_ctl, *parent_rate);

return?

> + break;
> + }
> + return pll_freq;

return 0?

> +}
> +
> +static int ma35d1_clk_pll_is_prepared(struct clk_hw *hw)
> +{
> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
> + u32 val = readl_relaxed(pll->ctl1_base);
> +
> + return val & PLL_CTL1_PD ? 0 : 1;
> +}
> +
> +static int ma35d1_clk_pll_prepare(struct clk_hw *hw)
> +{
> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
> + u32 val;
> +
> + if ((pll->type == MA35D1_CAPLL) || (pll->type == MA35D1_DDRPLL)) {

Drop useless parenthesis.

> + pr_warn("Nuvoton MA35D1 CAPLL/DDRPLL is read only.\n");
> + return -EACCES;
> + }
> +
> + ma35d1_unlock_regs(pll);
> + val = readl_relaxed(pll->ctl1_base);
> + val &= ~PLL_CTL1_PD;
> + writel_relaxed(val, pll->ctl1_base);
> + ma35d1_lock_regs(pll);
> + return 0;
> +}
> +
> +static void ma35d1_clk_pll_unprepare(struct clk_hw *hw)
> +{
> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
> + u32 val;
> +
> + if ((pll->type == MA35D1_CAPLL) || (pll->type == MA35D1_DDRPLL)) {

Drop useless parenthesis.

> + pr_warn("Nuvoton MA35D1 CAPLL/DDRPLL is read only.\n");

These read-only checks should be removed and different clk_ops should be
assigned for the read-only type of clks.

> + } else {
> + val = readl_relaxed(pll->ctl1_base);
> + val |= PLL_CTL1_PD;
> + writel_relaxed(val, pll->ctl1_base);
> + }
> +}
> +
> +static const struct clk_ops ma35d1_clk_pll_ops = {
> + .is_prepared = ma35d1_clk_pll_is_prepared,
> + .prepare = ma35d1_clk_pll_prepare,
> + .unprepare = ma35d1_clk_pll_unprepare,
> + .set_rate = ma35d1_clk_pll_set_rate,
> + .recalc_rate = ma35d1_clk_pll_recalc_rate,
> + .round_rate = ma35d1_clk_pll_round_rate,
> +};
> +
> +struct clk_hw *ma35d1_reg_clk_pll(enum ma35d1_pll_type type,

Pass a device here.

> + u8 u8mode, const char *name,
> + const char *parent,
> + unsigned long targetFreq,
> + void __iomem *base,
> + struct regmap *regmap)
> +{
> + struct ma35d1_clk_pll *pll;
> + struct clk_hw *hw;
> + struct clk_init_data init;

init = {};

> + int ret;
> +
> + pll = kmalloc(sizeof(*pll), GFP_KERNEL);

Just use kzalloc(). This isn't a hot path.

> + if (!pll)
> + return ERR_PTR(-ENOMEM);
> +
> + pll->type = type;
> + pll->mode = u8mode;
> + pll->rate = targetFreq;
> + pll->ctl0_base = base + REG_PLL_CTL0_OFFSET;
> + pll->ctl1_base = base + REG_PLL_CTL1_OFFSET;
> + pll->ctl2_base = base + REG_PLL_CTL2_OFFSET;
> + pll->regmap = regmap;
> +
> + init.name = name;
> + init.flags = 0;
> + init.parent_names = &parent;
> + init.num_parents = 1;
> + init.ops = &ma35d1_clk_pll_ops;
> + pll->hw.init = &init;
> + hw = &pll->hw;
> +
> + ret = clk_hw_register(NULL, hw);

Use devm_

> + if (ret) {
> + pr_err("failed to register vsi-pll clock!!!\n");

I'm going to put a dev_err_probe() into clk registration code. Remove
this printk.

> + kfree(pll);
> + return ERR_PTR(ret);
> + }
> + return hw;
> +}
> diff --git a/drivers/clk/nuvoton/clk-ma35d1.c b/drivers/clk/nuvoton/clk-ma35d1.c
> new file mode 100644
> index 000000000000..e4d3ced396a3
> --- /dev/null
> +++ b/drivers/clk/nuvoton/clk-ma35d1.c
> @@ -0,0 +1,963 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Copyright (C) 2023 Nuvoton Technology Corp.
> + * Author: Chi-Fang Li <[email protected]>
> + */
> +
> +#include <linux/clk-provider.h>
> +#include <linux/mfd/syscon.h>
> +#include <linux/module.h>
> +#include <linux/of.h>
> +#include <linux/of_address.h>
> +#include <linux/platform_device.h>
> +#include <linux/spinlock.h>
> +#include <dt-bindings/clock/nuvoton,ma35d1-clk.h>
> +
> +#include "clk-ma35d1.h"
> +
> +static DEFINE_SPINLOCK(ma35d1_lock);
> +
> +static const char *const ca35clk_sel_clks[] = {

Use clk_parent_data instead.

> + "hxt", "capll", "ddrpll", "dummy"
> +};
> +
[...]
> +
> +static inline struct clk_hw *ma35d1_clk_fixed(const char *name, int rate)
> +{
> + return clk_hw_register_fixed_rate(NULL, name, NULL, 0, rate);
> +}
> +
> +static inline struct clk_hw *ma35d1_clk_mux(const char *name,
> + void __iomem *reg,
> + u8 shift, u8 width,
> + const char *const *parents,
> + int num_parents)
> +{
> + return clk_hw_register_mux(NULL, name, parents, num_parents,
> + CLK_SET_RATE_NO_REPARENT, reg, shift,
> + width, 0, &ma35d1_lock);
> +}
> +
> +static inline struct clk_hw *ma35d1_clk_divider(const char *name,
> + const char *parent,
> + void __iomem *reg, u8 shift,
> + u8 width)
> +{
> + return clk_hw_register_divider(NULL, name, parent, CLK_SET_RATE_PARENT,
> + reg, shift, width, 0, &ma35d1_lock);
> +}
> +
> +static inline struct clk_hw *ma35d1_clk_divider_pow2(const char *name,
> + const char *parent,
> + void __iomem *reg,
> + u8 shift, u8 width)
> +{
> + return clk_hw_register_divider(NULL, name, parent,
> + CLK_DIVIDER_POWER_OF_TWO, reg, shift,
> + width, 0, &ma35d1_lock);
> +}
> +
> +static inline struct clk_hw *ma35d1_clk_divider_table(const char *name,
> + const char *parent,
> + void __iomem *reg,
> + u8 shift, u8 width,
> + const struct clk_div_table *table)
> +{
> + return clk_hw_register_divider_table(NULL, name, parent, 0,
> + reg, shift, width, 0, table,
> + &ma35d1_lock);
> +}
> +
> +static inline struct clk_hw *ma35d1_clk_fixed_factor(const char *name,
> + const char *parent,
> + unsigned int mult,
> + unsigned int div)
> +{
> + return clk_hw_register_fixed_factor(NULL, name, parent,
> + CLK_SET_RATE_PARENT, mult, div);
> +}
> +
> +static inline struct clk_hw *ma35d1_clk_gate(const char *name,
> + const char *parent,
> + void __iomem *reg, u8 shift)
> +{
> + return clk_hw_register_gate(NULL, name, parent, CLK_SET_RATE_PARENT,
> + reg, shift, 0, &ma35d1_lock);
> +}

These need to use devm_ to unregister on failure.

> +
> +static int ma35d1_get_pll_setting(struct device_node *clk_node,
> + u32 *pllmode, u32 *pllfreq)
> +{
> + const char *of_str;
> + int i;
> +
> + for (i = 0; i < PLL_MAX_NUM; i++) {
> + if (of_property_read_string_index(clk_node, "nuvoton,pll-mode", i, &of_str))
> + return -EINVAL;
> + if (!strcmp(of_str, "integer"))
> + pllmode[i] = PLL_MODE_INT;
> + else if (!strcmp(of_str, "fractional"))
> + pllmode[i] = PLL_MODE_FRAC;
> + else if (!strcmp(of_str, "spread-spectrum"))
> + pllmode[i] = PLL_MODE_SS;
> + else
> + return -EINVAL;
> + }
> + return of_property_read_u32_array(clk_node, "assigned-clock-rates",

The clk framework looks at this property. Why is the driver looking at
it too?

> + pllfreq, PLL_MAX_NUM);
> +}
> +

2023-03-29 02:09:19

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Dear Stephen,


Thanks for your advice.


On 2023/3/29 上午 01:57, Stephen Boyd wrote:
> Quoting Jacky Huang (2023-03-27 19:19:08)
>> diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
>> new file mode 100644
>> index 000000000000..0740b0b218a7
>> --- /dev/null
>> +++ b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
>> @@ -0,0 +1,231 @@
>> +// SPDX-License-Identifier: GPL-2.0
>> +/*
>> + * Copyright (C) 2023 Nuvoton Technology Corp.
>> + * Author: Shan-Chun Hung <[email protected]>
>> + * Jacky huang <[email protected]>
>> + */
>> +
>> +#include <dt-bindings/interrupt-controller/arm-gic.h>
>> +#include <dt-bindings/input/input.h>
>> +#include <dt-bindings/gpio/gpio.h>
>> +#include <dt-bindings/clock/nuvoton,ma35d1-clk.h>
>> +#include <dt-bindings/reset/nuvoton,ma35d1-reset.h>
>> +
>> +/ {
>> + compatible = "nuvoton,ma35d1";
>> + interrupt-parent = <&gic>;
>> + #address-cells = <2>;
>> + #size-cells = <2>;
>> +
>> + cpus {
>> + #address-cells = <2>;
>> + #size-cells = <0>;
>> +
>> + cpu0: cpu@0 {
>> + device_type = "cpu";
>> + compatible = "arm,cortex-a35";
>> + reg = <0x0 0x0>;
>> + enable-method = "psci";
>> + next-level-cache = <&L2_0>;
>> + };
>> +
>> + cpu1: cpu@1 {
>> + device_type = "cpu";
>> + compatible = "arm,cortex-a35";
>> + reg = <0x0 0x1>;
>> + enable-method = "psci";
>> + next-level-cache = <&L2_0>;
>> + };
>> +
>> + L2_0: l2-cache0 {
> Just l2-cache for the node name. Doesn't it go under the cpu0 node as
> well?

This describes the level-2 cache which is external to and shared by cpu0
& cpu1.
And only level-1 cache is inside of CPU core.
L2_0 is must, because both cpu0 and cpu1 has a next-level-cache =
<&L2_0> property.

Many identical example of l2-cache node can be found in arm64 dts, such
as k3-arm642.dtsi,
rk3328.dtsi, mt8195.dtsi, etc. Here is just a copy of similar arm64
multi-core SoCs.

So we would like to keep this unchanged. Is it OK for you? Thanks.


>> + compatible = "cache";
>> + cache-level = <2>;
>> + };
>> + };
>> +
>> + psci {
>> + compatible = "arm,psci-0.2";
>> + method = "smc";
>> + };
>> +
>> + timer {
>> + compatible = "arm,armv8-timer";
>> + interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(4) |
>> + IRQ_TYPE_LEVEL_LOW)>, /* Physical Secure */
>> + <GIC_PPI 14 (GIC_CPU_MASK_SIMPLE(4) |
>> + IRQ_TYPE_LEVEL_LOW)>, /* Physical Non-Secure */
>> + <GIC_PPI 11 (GIC_CPU_MASK_SIMPLE(4) |
>> + IRQ_TYPE_LEVEL_LOW)>, /* Virtual */
>> + <GIC_PPI 10 (GIC_CPU_MASK_SIMPLE(4) |
>> + IRQ_TYPE_LEVEL_LOW)>; /* Hypervisor */
>> + clock-frequency = <12000000>;
> Remove this property. The frequency should be read by the driver.

I will remove it.

>> + interrupt-parent = <&gic>;
>> + };
> Please create an 'soc' node for the SoC to hold all the nodes that have
> a reg property.

OK, we will use soc node in the next version.

>> +
>> + sys: system-management@40460000 {
>> + compatible = "nuvoton,ma35d1-sys", "syscon", "simple-mfd";
>> + reg = <0x0 0x40460000 0x0 0x200>;
>> +
>> + reset: reset-controller {
>> + compatible = "nuvoton,ma35d1-reset";
>> + #reset-cells = <1>;
>> + };
>> + };
>> +
>> + clk: clock-controller@40460200 {
>> + compatible = "nuvoton,ma35d1-clk", "syscon";
>> + reg = <0x00000000 0x40460200 0x0 0x100>;
>> + #clock-cells = <1>;
>> + clocks = <&clk_hxt>;
>> + nuvoton,sys = <&sys>;
>> + };
> It looks like the device at 40460000 is a reset and clock controller.
> Just make it one node and register the clk or reset device as an
> auxiliary device.

40460000 is for system control registers, including power contrl,
multifunction pin control,
usb phy control, IP reset control, power-on setting information, and
many other miscellaneous controls.
The registers of reset controller is only a subset of system control
registers.

40460200 is for clock controller which is independent of the system
control integration
The register base of clock controller is very close to system
controller, but in fact the two are independent.


>> +
>> + gic: interrupt-controller@50801000 {
>> + compatible = "arm,gic-400";
>> + reg = <0x0 0x50801000 0 0x1000>, /* GICD */
>> + <0x0 0x50802000 0 0x2000>, /* GICC */
>> + <0x0 0x50804000 0 0x2000>, /* GICH */
>> + <0x0 0x50806000 0 0x2000>; /* GICV */
>> + #interrupt-cells = <3>;
>> + interrupt-parent = <&gic>;
>> + interrupt-controller;
>> + interrupts = <GIC_PPI 9 (GIC_CPU_MASK_RAW(0x13) |
>> + IRQ_TYPE_LEVEL_HIGH)>;
>> + };
>> +
>> + uart0:serial@40700000 {
> Add a space after :

I will fix it. Thank you.

>> + compatible = "nuvoton,ma35d1-uart";
>> + reg = <0x0 0x40700000 0x0 0x100>;
>> + interrupts = <GIC_SPI 59 IRQ_TYPE_LEVEL_HIGH>;
>> + clocks = <&clk UART0_GATE>;
>> + status = "disabled";
>> + };


Best regards,
Jacky Huang


2023-03-29 02:25:13

by Stephen Boyd

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Quoting Jacky Huang (2023-03-28 19:03:24)
> On 2023/3/29 上午 01:57, Stephen Boyd wrote:
> > Quoting Jacky Huang (2023-03-27 19:19:08)
> >> diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
> >> new file mode 100644
> >> index 000000000000..0740b0b218a7
> >> --- /dev/null
> >> +++ b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
> >> @@ -0,0 +1,231 @@
[...]
> >> +
> >> + L2_0: l2-cache0 {
> > Just l2-cache for the node name. Doesn't it go under the cpu0 node as
> > well?
>
> This describes the level-2 cache which is external to and shared by cpu0
> & cpu1.
> And only level-1 cache is inside of CPU core.
> L2_0 is must, because both cpu0 and cpu1 has a next-level-cache =
> <&L2_0> property.

Ok. The name should just be l2-cache then, not l2-cache0.

>
> Many identical example of l2-cache node can be found in arm64 dts, such
> as k3-arm642.dtsi,
> rk3328.dtsi, mt8195.dtsi, etc. Here is just a copy of similar arm64
> multi-core SoCs.
>
> So we would like to keep this unchanged. Is it OK for you? Thanks.
>

Mostly ok, yes.

>
> >> +
> >> + sys: system-management@40460000 {
> >> + compatible = "nuvoton,ma35d1-sys", "syscon", "simple-mfd";
> >> + reg = <0x0 0x40460000 0x0 0x200>;
> >> +
> >> + reset: reset-controller {
> >> + compatible = "nuvoton,ma35d1-reset";
> >> + #reset-cells = <1>;
> >> + };
> >> + };
> >> +
> >> + clk: clock-controller@40460200 {
> >> + compatible = "nuvoton,ma35d1-clk", "syscon";
> >> + reg = <0x00000000 0x40460200 0x0 0x100>;
> >> + #clock-cells = <1>;
> >> + clocks = <&clk_hxt>;
> >> + nuvoton,sys = <&sys>;
> >> + };
> > It looks like the device at 40460000 is a reset and clock controller.
> > Just make it one node and register the clk or reset device as an
> > auxiliary device.
>
> 40460000 is for system control registers, including power contrl,
> multifunction pin control,
> usb phy control, IP reset control, power-on setting information, and
> many other miscellaneous controls.
> The registers of reset controller is only a subset of system control
> registers.
>
> 40460200 is for clock controller which is independent of the system
> control integration
> The register base of clock controller is very close to system
> controller, but in fact the two are independent.

What do you use the syscon for then? The clock driver must want to use
the syscon for something, implying that they are the same device.

2023-03-29 02:40:44

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Dear Stephen,


On 2023/3/29 上午 10:19, Stephen Boyd wrote:
> Quoting Jacky Huang (2023-03-28 19:03:24)
>> On 2023/3/29 上午 01:57, Stephen Boyd wrote:
>>> Quoting Jacky Huang (2023-03-27 19:19:08)
>>>> diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
>>>> new file mode 100644
>>>> index 000000000000..0740b0b218a7
>>>> --- /dev/null
>>>> +++ b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
>>>> @@ -0,0 +1,231 @@
> [...]
>>>> +
>>>> + L2_0: l2-cache0 {
>>> Just l2-cache for the node name. Doesn't it go under the cpu0 node as
>>> well?
>> This describes the level-2 cache which is external to and shared by cpu0
>> & cpu1.
>> And only level-1 cache is inside of CPU core.
>> L2_0 is must, because both cpu0 and cpu1 has a next-level-cache =
>> <&L2_0> property.
> Ok. The name should just be l2-cache then, not l2-cache0.

OK, I will fix it.

>> Many identical example of l2-cache node can be found in arm64 dts, such
>> as k3-arm642.dtsi,
>> rk3328.dtsi, mt8195.dtsi, etc. Here is just a copy of similar arm64
>> multi-core SoCs.
>>
>> So we would like to keep this unchanged. Is it OK for you? Thanks.
>>
> Mostly ok, yes.

Thank you for the agreement.

>
>>>> +
>>>> + sys: system-management@40460000 {
>>>> + compatible = "nuvoton,ma35d1-sys", "syscon", "simple-mfd";
>>>> + reg = <0x0 0x40460000 0x0 0x200>;
>>>> +
>>>> + reset: reset-controller {
>>>> + compatible = "nuvoton,ma35d1-reset";
>>>> + #reset-cells = <1>;
>>>> + };
>>>> + };
>>>> +
>>>> + clk: clock-controller@40460200 {
>>>> + compatible = "nuvoton,ma35d1-clk", "syscon";
>>>> + reg = <0x00000000 0x40460200 0x0 0x100>;
>>>> + #clock-cells = <1>;
>>>> + clocks = <&clk_hxt>;
>>>> + nuvoton,sys = <&sys>;
>>>> + };
>>> It looks like the device at 40460000 is a reset and clock controller.
>>> Just make it one node and register the clk or reset device as an
>>> auxiliary device.
>> 40460000 is for system control registers, including power contrl,
>> multifunction pin control,
>> usb phy control, IP reset control, power-on setting information, and
>> many other miscellaneous controls.
>> The registers of reset controller is only a subset of system control
>> registers.
>>
>> 40460200 is for clock controller which is independent of the system
>> control integration
>> The register base of clock controller is very close to system
>> controller, but in fact the two are independent.
> What do you use the syscon for then? The clock driver must want to use
> the syscon for something, implying that they are the same device.

The register lock mechanism is applied to protect many critical
registers from false written.
The register lock control register is one register in system controller.
Some registers of the clock controller are lock protected. Not only
clock controller, but other
IP such as RTC, PWM, ADC, etc, also have lock protected registers. All
these IP requires
syscon to access the lock/unlock control register in the system controller.
That's why we add a <&sys> to the clock controller.

Should we implement a ma35d1-sysctl driver to protect register_lock()
and register_unlock()
and export to those drivers?  If yes, we can remove the <&sys> from
clock controller.


Best regards,
Jacky Huang








2023-03-29 03:01:32

by Stephen Boyd

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Quoting Jacky Huang (2023-03-28 19:39:36)
> On 2023/3/29 上午 10:19, Stephen Boyd wrote:
> > What do you use the syscon for then? The clock driver must want to use
> > the syscon for something, implying that they are the same device.
>
> The register lock mechanism is applied to protect many critical
> registers from false written.
> The register lock control register is one register in system controller.
> Some registers of the clock controller are lock protected. Not only
> clock controller, but other
> IP such as RTC, PWM, ADC, etc, also have lock protected registers. All
> these IP requires
> syscon to access the lock/unlock control register in the system controller.
> That's why we add a <&sys> to the clock controller.
>
> Should we implement a ma35d1-sysctl driver to protect register_lock()
> and register_unlock()
> and export to those drivers?  If yes, we can remove the <&sys> from
> clock controller.
>

You can implement the lock and unlock in the hwspinlock framework. See
drivers/hwspinlock.

2023-03-29 03:30:56

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Dear Stephen,


On 2023/3/29 上午 10:46, Stephen Boyd wrote:
> Quoting Jacky Huang (2023-03-28 19:39:36)
>> On 2023/3/29 上午 10:19, Stephen Boyd wrote:
>>> What do you use the syscon for then? The clock driver must want to use
>>> the syscon for something, implying that they are the same device.
>> The register lock mechanism is applied to protect many critical
>> registers from false written.
>> The register lock control register is one register in system controller.
>> Some registers of the clock controller are lock protected. Not only
>> clock controller, but other
>> IP such as RTC, PWM, ADC, etc, also have lock protected registers. All
>> these IP requires
>> syscon to access the lock/unlock control register in the system controller.
>> That's why we add a <&sys> to the clock controller.
>>
>> Should we implement a ma35d1-sysctl driver to protect register_lock()
>> and register_unlock()
>> and export to those drivers?  If yes, we can remove the <&sys> from
>> clock controller.
>>
> You can implement the lock and unlock in the hwspinlock framework. See
> drivers/hwspinlock.

I may not explain clearly enough. The lock/unlock register of system
controller is more like
a kind of write protection for specific registers, rather than
preventing hetero-core CPU access.
In many different IP of ma35d1 contain write protected registers.
In fact, ma35d1 has a "hardware semaphore" IP, and we have implemented
the driver in drivers/hwspinlock.
Even the control register of "hardware semaphore" is also write protected.

So, should we implement a system controller driver to provide
register_unlock() function?
Is it OK to have such a driver in drivers/mfd?
Or, just use syscon in device tree for those devices that have write
protect registers.


Best regards,
Jacky Huang


2023-03-29 03:35:24

by Stephen Boyd

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Quoting Jacky Huang (2023-03-28 20:13:11)
> Dear Stephen,
>
>
> On 2023/3/29 上午 10:46, Stephen Boyd wrote:
> > Quoting Jacky Huang (2023-03-28 19:39:36)
> >> On 2023/3/29 上午 10:19, Stephen Boyd wrote:
> >>> What do you use the syscon for then? The clock driver must want to use
> >>> the syscon for something, implying that they are the same device.
> >> The register lock mechanism is applied to protect many critical
> >> registers from false written.
> >> The register lock control register is one register in system controller.
> >> Some registers of the clock controller are lock protected. Not only
> >> clock controller, but other
> >> IP such as RTC, PWM, ADC, etc, also have lock protected registers. All
> >> these IP requires
> >> syscon to access the lock/unlock control register in the system controller.
> >> That's why we add a <&sys> to the clock controller.
> >>
> >> Should we implement a ma35d1-sysctl driver to protect register_lock()
> >> and register_unlock()
> >> and export to those drivers?  If yes, we can remove the <&sys> from
> >> clock controller.
> >>
> > You can implement the lock and unlock in the hwspinlock framework. See
> > drivers/hwspinlock.
>
> I may not explain clearly enough. The lock/unlock register of system
> controller is more like
> a kind of write protection for specific registers, rather than
> preventing hetero-core CPU access.
> In many different IP of ma35d1 contain write protected registers.
> In fact, ma35d1 has a "hardware semaphore" IP, and we have implemented
> the driver in drivers/hwspinlock.
> Even the control register of "hardware semaphore" is also write protected.

What's the need to lock and unlock the registers? Is some other
processor also writing to the registers that we need to synchronize
against? Or is Linux the only entity reading and writing the registers?
I'm wondering if we should simply unlock the registers and never lock
them.

>
> So, should we implement a system controller driver to provide
> register_unlock() function?
> Is it OK to have such a driver in drivers/mfd?
> Or, just use syscon in device tree for those devices that have write
> protect registers.
>

The hwspinlock framework doesn't require there to be another entity
accessing some resource. It's there to implement hardware locks. I don't
see why it can't be used here.

2023-03-29 03:50:57

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Dear Stephen,


On 2023/3/29 上午 11:25, Stephen Boyd wrote:
> Quoting Jacky Huang (2023-03-28 20:13:11)
>> Dear Stephen,
>>
>>
>> On 2023/3/29 上午 10:46, Stephen Boyd wrote:
>>> Quoting Jacky Huang (2023-03-28 19:39:36)
>>>> On 2023/3/29 上午 10:19, Stephen Boyd wrote:
>>>>> What do you use the syscon for then? The clock driver must want to use
>>>>> the syscon for something, implying that they are the same device.
>>>> The register lock mechanism is applied to protect many critical
>>>> registers from false written.
>>>> The register lock control register is one register in system controller.
>>>> Some registers of the clock controller are lock protected. Not only
>>>> clock controller, but other
>>>> IP such as RTC, PWM, ADC, etc, also have lock protected registers. All
>>>> these IP requires
>>>> syscon to access the lock/unlock control register in the system controller.
>>>> That's why we add a <&sys> to the clock controller.
>>>>
>>>> Should we implement a ma35d1-sysctl driver to protect register_lock()
>>>> and register_unlock()
>>>> and export to those drivers?  If yes, we can remove the <&sys> from
>>>> clock controller.
>>>>
>>> You can implement the lock and unlock in the hwspinlock framework. See
>>> drivers/hwspinlock.
>> I may not explain clearly enough. The lock/unlock register of system
>> controller is more like
>> a kind of write protection for specific registers, rather than
>> preventing hetero-core CPU access.
>> In many different IP of ma35d1 contain write protected registers.
>> In fact, ma35d1 has a "hardware semaphore" IP, and we have implemented
>> the driver in drivers/hwspinlock.
>> Even the control register of "hardware semaphore" is also write protected.
> What's the need to lock and unlock the registers? Is some other
> processor also writing to the registers that we need to synchronize
> against? Or is Linux the only entity reading and writing the registers?
> I'm wondering if we should simply unlock the registers and never lock
> them.
>
>> So, should we implement a system controller driver to provide
>> register_unlock() function?
>> Is it OK to have such a driver in drivers/mfd?
>> Or, just use syscon in device tree for those devices that have write
>> protect registers.
>>
> The hwspinlock framework doesn't require there to be another entity
> accessing some resource. It's there to implement hardware locks. I don't
> see why it can't be used here.

The current usage of register lock/unlock protect is as the following code:

static void ma35d1_unlock_regs(struct ma35d1_clk_pll *pll)
{
    int ret;

    do {
        regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x59);
        regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x16);
        regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x88);
        regmap_read(pll->regmap, REG_SYS_RLKTZNS, &ret);
    } while (ret == 0);
}

static void ma35d1_lock_regs(struct ma35d1_clk_pll *pll)
{
    regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x0);
}

And the following code is to unlock registers for write and then lock again.

    ma35d1_unlock_regs(pll);
    writel_relaxed(reg_ctl[0], pll->ctl0_base);
    writel_relaxed(reg_ctl[1], pll->ctl1_base);
    writel_relaxed(reg_ctl[2], pll->ctl2_base);
    ma35d1_lock_regs(pll);

The above code is from the clk-ma35d1-pll.c from this patchset.

We just employ regmap mechansim for the access to REG_SYS_RLKTZNS register.
Is this implementation OK for you?  Thank you.


Best regards,
Jacky Huang





2023-03-29 03:56:16

by Stephen Boyd

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Quoting Jacky Huang (2023-03-28 20:43:23)
> On 2023/3/29 上午 11:25, Stephen Boyd wrote:
> > Quoting Jacky Huang (2023-03-28 20:13:11)
> >> I may not explain clearly enough. The lock/unlock register of system
> >> controller is more like
> >> a kind of write protection for specific registers, rather than
> >> preventing hetero-core CPU access.
> >> In many different IP of ma35d1 contain write protected registers.
> >> In fact, ma35d1 has a "hardware semaphore" IP, and we have implemented
> >> the driver in drivers/hwspinlock.
> >> Even the control register of "hardware semaphore" is also write protected.
> > What's the need to lock and unlock the registers? Is some other
> > processor also writing to the registers that we need to synchronize
> > against? Or is Linux the only entity reading and writing the registers?
> > I'm wondering if we should simply unlock the registers and never lock
> > them.

Can you answer this question?

> >
> >> So, should we implement a system controller driver to provide
> >> register_unlock() function?
> >> Is it OK to have such a driver in drivers/mfd?
> >> Or, just use syscon in device tree for those devices that have write
> >> protect registers.
> >>
> > The hwspinlock framework doesn't require there to be another entity
> > accessing some resource. It's there to implement hardware locks. I don't
> > see why it can't be used here.
>
> The current usage of register lock/unlock protect is as the following code:
>
> static void ma35d1_unlock_regs(struct ma35d1_clk_pll *pll)
> {
>     int ret;
>
>     do {
>         regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x59);
>         regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x16);
>         regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x88);
>         regmap_read(pll->regmap, REG_SYS_RLKTZNS, &ret);
>     } while (ret == 0);
> }
>
> static void ma35d1_lock_regs(struct ma35d1_clk_pll *pll)
> {
>     regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x0);
> }
>
> And the following code is to unlock registers for write and then lock again.
>
>     ma35d1_unlock_regs(pll);
>     writel_relaxed(reg_ctl[0], pll->ctl0_base);
>     writel_relaxed(reg_ctl[1], pll->ctl1_base);
>     writel_relaxed(reg_ctl[2], pll->ctl2_base);
>     ma35d1_lock_regs(pll);
>
> The above code is from the clk-ma35d1-pll.c from this patchset.

Yeah I understand that you write some registers in the syscon to lock
the registers.

>
> We just employ regmap mechansim for the access to REG_SYS_RLKTZNS register.
> Is this implementation OK for you?  Thank you.
>

No. Why can't that be a hwspinlock? Or why can't it be unlocked all the
time and rely on software spinlocks in the kernel to prevent concurrent
access to the registers accessed by a driver, like a lock for the clk
registers and a lock for the reset registers, etc. Or if no two clks or
resets exist within one 32-bit word then no lock is necessary.

2023-03-29 06:04:53

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Dear Stephen,



On 2023/3/29 上午 11:54, Stephen Boyd wrote:
> Quoting Jacky Huang (2023-03-28 20:43:23)
>> On 2023/3/29 上午 11:25, Stephen Boyd wrote:
>>> Quoting Jacky Huang (2023-03-28 20:13:11)
>>>> I may not explain clearly enough. The lock/unlock register of system
>>>> controller is more like
>>>> a kind of write protection for specific registers, rather than
>>>> preventing hetero-core CPU access.
>>>> In many different IP of ma35d1 contain write protected registers.
>>>> In fact, ma35d1 has a "hardware semaphore" IP, and we have implemented
>>>> the driver in drivers/hwspinlock.
>>>> Even the control register of "hardware semaphore" is also write protected.
>>> What's the need to lock and unlock the registers? Is some other
>>> processor also writing to the registers that we need to synchronize
>>> against? Or is Linux the only entity reading and writing the registers?
>>> I'm wondering if we should simply unlock the registers and never lock
>>> them.
> Can you answer this question?

Sorry, I miss this.
The lock and unlock register mechanism is a hardware design of ma35d1 SoC.
The purpose is to prevent from unexpected write to some registers.
However, I think this is a redundant design if s/w is done properly.
Even though I think it's a redundant design, it's out there and we have
to deal with it.
And of course we have unlock and lock pair, I just lost to write in the
above.

>>>> So, should we implement a system controller driver to provide
>>>> register_unlock() function?
>>>> Is it OK to have such a driver in drivers/mfd?
>>>> Or, just use syscon in device tree for those devices that have write
>>>> protect registers.
>>>>
>>> The hwspinlock framework doesn't require there to be another entity
>>> accessing some resource. It's there to implement hardware locks. I don't
>>> see why it can't be used here.
>> The current usage of register lock/unlock protect is as the following code:
>>
>> static void ma35d1_unlock_regs(struct ma35d1_clk_pll *pll)
>> {
>>     int ret;
>>
>>     do {
>>         regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x59);
>>         regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x16);
>>         regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x88);
>>         regmap_read(pll->regmap, REG_SYS_RLKTZNS, &ret);
>>     } while (ret == 0);
>> }
>>
>> static void ma35d1_lock_regs(struct ma35d1_clk_pll *pll)
>> {
>>     regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x0);
>> }
>>
>> And the following code is to unlock registers for write and then lock again.
>>
>>     ma35d1_unlock_regs(pll);
>>     writel_relaxed(reg_ctl[0], pll->ctl0_base);
>>     writel_relaxed(reg_ctl[1], pll->ctl1_base);
>>     writel_relaxed(reg_ctl[2], pll->ctl2_base);
>>     ma35d1_lock_regs(pll);
>>
>> The above code is from the clk-ma35d1-pll.c from this patchset.
> Yeah I understand that you write some registers in the syscon to lock
> the registers.
>
>> We just employ regmap mechansim for the access to REG_SYS_RLKTZNS register.
>> Is this implementation OK for you?  Thank you.
>>
> No. Why can't that be a hwspinlock? Or why can't it be unlocked all the
> time and rely on software spinlocks in the kernel to prevent concurrent
> access to the registers accessed by a driver, like a lock for the clk
> registers and a lock for the reset registers, etc. Or if no two clks or
> resets exist within one 32-bit word then no lock is necessary.

You gave a good suggestion here. Yes, of course we can let it be
unlocked all the time.
We can unlock the "register lock" before entering Linux.
As a result, the unlonk and lock register related code is not required.
Thus, we can remove syscon from the clock controller node.

If you agree with this, we will modify it in the next version.


Best regards,
Jacky Huang




2023-03-29 08:17:32

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 11/12] tty: serial: Add Nuvoton ma35d1 serial driver support

Dear Jiri,


Thanks for your advice.

On 2023/3/28 下午 05:23, Jiri Slaby wrote:
> On 28. 03. 23, 4:19, Jacky Huang wrote:
>> +static void transmit_chars(struct uart_ma35d1_port *up)
>> +{
>> +    struct circ_buf *xmit = &up->port.state->xmit;
>> +    int count;
>> +    u8 ch;
>> +
>> +    if (uart_tx_stopped(&up->port)) {
>> +        ma35d1serial_stop_tx(&up->port);
>> +        return;
>> +    }
>> +    if (uart_circ_empty(xmit)) {
>> +        __stop_tx(up);
>> +        return;
>> +    }
>
> Why is this necessary?
>

I will remove this reduntant uart_circ_empty() check.

>> +    count = UART_FIFO_DEPTH - ((serial_in(up, UART_REG_FSR) &
>> FSR_TXPTR_MSK) >> 16);
>> +
>> +    uart_port_tx_limited(&up->port, ch, count,
>> +                 !(serial_in(up, UART_REG_FSR) & FSR_TX_FULL),
>> +                 serial_out(up, UART_REG_THR, ch),
>> +                 ({}));
>> +
>> +    if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
>> +        uart_write_wakeup(&up->port);
>> +
>> +    if (uart_circ_empty(xmit))
>> +        __stop_tx(up);
>
> uart_port_tx_limited() should take care about the above and this too,
> right?

Yes, you are right. These two are also redundant. I will remove them.
Thank you.

>
>> +}
> ...
>> +static void receive_chars(struct uart_ma35d1_port *up)
>> +{
>> +    u8 flag;
>> +    u32 fsr;
>> +    unsigned int ch;
>
> Shouldn't ch be u8 too?

OK, I will fix it as u8.

>
>> +    int max_count = 256;
>> +
>> +    fsr = serial_in(up, UART_REG_FSR);
>> +    do {
>> +        flag = TTY_NORMAL;
>> +        up->port.icount.rx++;
>> +
>> +        if (unlikely(fsr & (FSR_BIF | FSR_FEF | FSR_PEF |
>> FSR_RX_OVER_IF))) {
>> +            if (fsr & FSR_BIF) {
>> +                up->port.icount.brk++;
>> +                if (uart_handle_break(&up->port))
>> +                    continue;
>> +            }
>> +            if (fsr & FSR_FEF)
>> +                up->port.icount.frame++;
>> +            if (fsr & FSR_PEF)
>> +                up->port.icount.parity++;
>> +            if (fsr & FSR_RX_OVER_IF)
>> +                up->port.icount.overrun++;
>> +
>> +            serial_out(up, UART_REG_FSR, fsr &
>> +                   (FSR_BIF | FSR_FEF | FSR_PEF | FSR_RX_OVER_IF));
>> +
>> +            if (fsr & FSR_BIF)
>> +                flag = TTY_BREAK;
>> +            else if (fsr & FSR_PEF)
>> +                flag = TTY_PARITY;
>> +            else if (fsr & FSR_FEF)
>> +                flag = TTY_FRAME;
>> +        }
>> +
>> +        ch = serial_in(up, UART_REG_RBR);
>> +        if (uart_handle_sysrq_char(&up->port, ch))
>> +            continue;
>> +
>> +        spin_lock(&up->port.lock);
>> +        uart_insert_char(&up->port, fsr, FSR_RX_OVER_IF, ch, flag);
>> +        spin_unlock(&up->port.lock);
>> +
>> +        fsr = serial_in(up, UART_REG_FSR);
>> +    } while (!(fsr & FSR_RX_EMPTY) && (max_count-- > 0));
>> +
>> +    spin_lock(&up->port.lock);
>> +    tty_flip_buffer_push(&up->port.state->port);
>> +    spin_unlock(&up->port.lock);
>> +}
> ...
>> +static int ma35d1serial_remove(struct platform_device *dev)
>> +{
>> +    struct uart_port *port = platform_get_drvdata(dev);
>> +
>> +    if (port) {
>
> Can this ever be NULL?

OK, I will remove this NULL check.

>
>> + uart_remove_one_port(&ma35d1serial_reg, port);
>> +        free_irq(port->irq, port);
>> +    }
>> +    return 0;
>> +}
>> +
>> +static int ma35d1serial_suspend(struct platform_device *dev,
>> pm_message_t state)
>> +{
>> +    int i;
>> +    struct uart_ma35d1_port *up;
>> +
>> +    if (dev->dev.of_node)
>> +        i = of_alias_get_id(dev->dev.of_node, "serial");
>> +    if (i < 0) {
>> +        dev_err(&dev->dev, "failed to get alias/pdev id, errno
>> %d\n", i);
>> +        return i;
>> +    }
>> +    up = &ma35d1serial_ports[i];
>
> platform_get_drvdata(dev) ?

We will fix it.

>
>> +    if (i == 0) {
>> +        up->console_baud_rate = serial_in(up, UART_REG_BAUD);
>> +        up->console_line = serial_in(up, UART_REG_LCR);
>> +        up->console_int = serial_in(up, UART_REG_IER);
>> +    }
>> +    return 0;
>> +}
>> +
>> +static int ma35d1serial_resume(struct platform_device *dev)
>> +{
>> +    int i;
>> +    struct uart_ma35d1_port *up;
>> +
>> +    if (dev->dev.of_node)
>> +        i = of_alias_get_id(dev->dev.of_node, "serial");
>> +    if (i < 0) {
>> +        dev_err(&dev->dev, "failed to get alias/pdev id, errno
>> %d\n", i);
>> +        return i;
>> +    }
>> +    up = &ma35d1serial_ports[i];
>> +    if (i == 0) {
>> +        serial_out(up, UART_REG_BAUD, up->console_baud_rate);
>> +        serial_out(up, UART_REG_LCR, up->console_line);
>> +        serial_out(up, UART_REG_IER, up->console_int);
>> +    }
>> +    return 0;
>> +}
>
> No uart_suspend_port()/uart_resume_port()? You don't wait for
> transmitter to be empty in suspend. You don't stop tx etc.
>

We will fix it in the next version.

>> +static struct platform_driver ma35d1serial_driver = {
>> +    .probe      = ma35d1serial_probe,
>> +    .remove     = ma35d1serial_remove,
>> +    .suspend    = ma35d1serial_suspend,
>> +    .resume     = ma35d1serial_resume,
>> +    .driver     = {
>> +        .name   = "ma35d1-uart",
>> +        .owner  = THIS_MODULE,
>> +        .of_match_table = of_match_ptr(ma35d1_serial_of_match),
>> +    },
>> +};
>> +
>> +static int __init ma35d1serial_init(void)
>> +{
>> +    int ret;
>> +
>> +    ret = uart_register_driver(&ma35d1serial_reg);
>> +    if (ret)
>> +        return ret;
>> +    ret = platform_driver_register(&ma35d1serial_driver);
>> +    if (ret)
>> +        uart_unregister_driver(&ma35d1serial_reg);
>> +    return ret;
>> +}
>> +
>> +static void __exit ma35d1serial_exit(void)
>> +{
>> +    platform_driver_unregister(&ma35d1serial_driver);
>> +    uart_unregister_driver(&ma35d1serial_reg);
>> +}
>> +
>> +module_init(ma35d1serial_init);
>> +module_exit(ma35d1serial_exit);
>> +
>> +MODULE_LICENSE("GPL");
>> +MODULE_DESCRIPTION("MA35D1 serial driver");
>
>
>> +MODULE_ALIAS_CHARDEV_MAJOR(TTY_MAJOR);
>
> Why is this needed? How are other platform drivers autoloaded?
>
> thanks,

Yes, it's not needed. I will remove it.


Best regards,
Jacky Huang



2023-03-29 08:18:40

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH v6 03/12] dt-bindings: clock: nuvoton: add binding for ma35d1 clock controller

On 28/03/2023 04:19, Jacky Huang wrote:
> From: Jacky Huang <[email protected]>
>
> Add the dt-bindings header for Nuvoton ma35d1, that gets shared
> between the clock controller and clock references in the dts.
> Add documentation to describe nuvoton ma35d1 clock driver.
>
> Signed-off-by: Jacky Huang <[email protected]>
> ---

Reviewed-by: Krzysztof Kozlowski <[email protected]>

Best regards,
Krzysztof

2023-03-29 08:34:42

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform

On 28/03/2023 04:19, Jacky Huang wrote:
> From: Jacky Huang <[email protected]>
>
> Rename the bindings/arm/npcm directory as nuvoton.
> Add binding for ARMv8 based Nuvotn SoCs and platform boards.

Typo: Nucotn?

> Add initial bindings for ma35d1 series development boards.
>
> Signed-off-by: Jacky Huang <[email protected]>
> ---
> .../bindings/arm/nuvoton/nuvoton,ma35d1.yaml | 30 +++++++++++++++++++

Anyway robot pointed to issues here...


Best regards,
Krzysztof

2023-03-29 08:39:00

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH v6 07/12] dt-bindings: serial: Document ma35d1 uart controller

On 28/03/2023 04:19, Jacky Huang wrote:
> From: Jacky Huang <[email protected]>
>
> Add documentation to describe nuvoton ma35d1 uart driver bindings.
>

This depends on clock patches, so it must be taken together.

Reviewed-by: Krzysztof Kozlowski <[email protected]>

Best regards,
Krzysztof

2023-03-29 08:39:54

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

On 28/03/2023 04:19, Jacky Huang wrote:
> From: Jacky Huang <[email protected]>
>
> Add initial device tree support for Nuvoton ma35d1 SoC, including
> cpu, clock, reset, and serial controllers.
> Add reference boards som-256m and iot-512m.
>
> Signed-off-by: Jacky Huang <[email protected]>



> + gic: interrupt-controller@50801000 {
> + compatible = "arm,gic-400";
> + reg = <0x0 0x50801000 0 0x1000>, /* GICD */
> + <0x0 0x50802000 0 0x2000>, /* GICC */
> + <0x0 0x50804000 0 0x2000>, /* GICH */
> + <0x0 0x50806000 0 0x2000>; /* GICV */
> + #interrupt-cells = <3>;
> + interrupt-parent = <&gic>;
> + interrupt-controller;
> + interrupts = <GIC_PPI 9 (GIC_CPU_MASK_RAW(0x13) |
> + IRQ_TYPE_LEVEL_HIGH)>;
> + };
> +
> + uart0:serial@40700000 {

There is always space after label:.


Best regards,
Krzysztof

2023-03-29 08:41:06

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform


Dear Krzysztof,


On 2023/3/29 下午 04:19, Krzysztof Kozlowski wrote:
> On 28/03/2023 04:19, Jacky Huang wrote:
>> From: Jacky Huang <[email protected]>
>>
>> Rename the bindings/arm/npcm directory as nuvoton.
>> Add binding for ARMv8 based Nuvotn SoCs and platform boards.
> Typo: Nucotn?

I will fix it. Thank you.

>> Add initial bindings for ma35d1 series development boards.
>>
>> Signed-off-by: Jacky Huang <[email protected]>
>> ---
>> .../bindings/arm/nuvoton/nuvoton,ma35d1.yaml | 30 +++++++++++++++++++
> Anyway robot pointed to issues here...
>
>
> Best regards,
> Krzysztof
>

The robot report path error in MAINTAINER. I should modify it when I
rename the npcm directory.


Best regards,
Jacky Huang


2023-03-29 08:42:06

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Dear Krzysztof,


On 2023/3/29 下午 04:21, Krzysztof Kozlowski wrote:
> On 28/03/2023 04:19, Jacky Huang wrote:
>> From: Jacky Huang <[email protected]>
>>
>> Add initial device tree support for Nuvoton ma35d1 SoC, including
>> cpu, clock, reset, and serial controllers.
>> Add reference boards som-256m and iot-512m.
>>
>> Signed-off-by: Jacky Huang <[email protected]>
>
>
>> + gic: interrupt-controller@50801000 {
>> + compatible = "arm,gic-400";
>> + reg = <0x0 0x50801000 0 0x1000>, /* GICD */
>> + <0x0 0x50802000 0 0x2000>, /* GICC */
>> + <0x0 0x50804000 0 0x2000>, /* GICH */
>> + <0x0 0x50806000 0 0x2000>; /* GICV */
>> + #interrupt-cells = <3>;
>> + interrupt-parent = <&gic>;
>> + interrupt-controller;
>> + interrupts = <GIC_PPI 9 (GIC_CPU_MASK_RAW(0x13) |
>> + IRQ_TYPE_LEVEL_HIGH)>;
>> + };
>> +
>> + uart0:serial@40700000 {
> There is always space after label:.
>
>
> Best regards,
> Krzysztof
>

I will fix them all.


Best regards,
Jacky Huang

2023-03-29 08:46:35

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 07/12] dt-bindings: serial: Document ma35d1 uart controller

Dear Krzysztof,


Thanks for your review.


On 2023/3/29 下午 04:20, Krzysztof Kozlowski wrote:
> On 28/03/2023 04:19, Jacky Huang wrote:
>> From: Jacky Huang <[email protected]>
>>
>> Add documentation to describe nuvoton ma35d1 uart driver bindings.
>>
> This depends on clock patches, so it must be taken together.
>
> Reviewed-by: Krzysztof Kozlowski <[email protected]>
>
> Best regards,
> Krzysztof
>

OK, I will combine this patch with the clock patch.


Best regards,
Jacky Huang

2023-03-29 13:10:13

by Rob Herring

[permalink] [raw]
Subject: Re: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform

On Mon, Mar 27, 2023 at 9:19 PM Jacky Huang <[email protected]> wrote:
>
> From: Jacky Huang <[email protected]>
>
> Rename the bindings/arm/npcm directory as nuvoton.
> Add binding for ARMv8 based Nuvotn SoCs and platform boards.
> Add initial bindings for ma35d1 series development boards.
>
> Signed-off-by: Jacky Huang <[email protected]>
> ---
> .../bindings/arm/nuvoton/nuvoton,ma35d1.yaml | 30 +++++++++++++++++++
> .../nuvoton,npcm-gcr.yaml} | 2 +-
> .../npcm.yaml => nuvoton/nuvoton,npcm.yaml} | 2 +-
> 3 files changed, 32 insertions(+), 2 deletions(-)
> create mode 100644 Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml
> rename Documentation/devicetree/bindings/arm/{npcm/nuvoton,gcr.yaml => nuvoton/nuvoton,npcm-gcr.yaml} (93%)

Since you are moving it, this one should be moved to bindings/soc/nuvoton/

> rename Documentation/devicetree/bindings/arm/{npcm/npcm.yaml => nuvoton/nuvoton,npcm.yaml} (93%)

2023-03-29 18:02:12

by Stephen Boyd

[permalink] [raw]
Subject: Re: [PATCH v6 08/12] arm64: dts: nuvoton: Add initial ma35d1 device tree

Quoting Jacky Huang (2023-03-28 23:02:31)
>
> On 2023/3/29 上午 11:54, Stephen Boyd wrote:
> > Quoting Jacky Huang (2023-03-28 20:43:23)
> >
> >> We just employ regmap mechansim for the access to REG_SYS_RLKTZNS register.
> >> Is this implementation OK for you?  Thank you.
> >>
> > No. Why can't that be a hwspinlock? Or why can't it be unlocked all the
> > time and rely on software spinlocks in the kernel to prevent concurrent
> > access to the registers accessed by a driver, like a lock for the clk
> > registers and a lock for the reset registers, etc. Or if no two clks or
> > resets exist within one 32-bit word then no lock is necessary.
>
> You gave a good suggestion here. Yes, of course we can let it be
> unlocked all the time.
> We can unlock the "register lock" before entering Linux.
> As a result, the unlonk and lock register related code is not required.
> Thus, we can remove syscon from the clock controller node.
>
> If you agree with this, we will modify it in the next version.
>

Sure, that works.

2023-03-30 07:41:00

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH v6 07/12] dt-bindings: serial: Document ma35d1 uart controller

On 29/03/2023 10:44, Jacky Huang wrote:
> Dear Krzysztof,
>
>
> Thanks for your review.
>
>
> On 2023/3/29 下午 04:20, Krzysztof Kozlowski wrote:
>> On 28/03/2023 04:19, Jacky Huang wrote:
>>> From: Jacky Huang <[email protected]>
>>>
>>> Add documentation to describe nuvoton ma35d1 uart driver bindings.
>>>
>> This depends on clock patches, so it must be taken together.
>>
>> Reviewed-by: Krzysztof Kozlowski <[email protected]>
>>
>> Best regards,
>> Krzysztof
>>
>
> OK, I will combine this patch with the clock patch.

I don't know what you mean by "combine", but I don't think this is what
we want. You must clearly state all dependencies and patch merging
strategy/way.

Best regards,
Krzysztof

2023-03-30 10:37:20

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 09/12] clk: nuvoton: Add clock driver for ma35d1 clock controller

Dear Stephen,


Thanks for your advice.

On 2023/3/29 上午 02:18, Stephen Boyd wrote:
> Quoting Jacky Huang (2023-03-27 19:19:09)
>> diff --git a/drivers/clk/nuvoton/Kconfig b/drivers/clk/nuvoton/Kconfig
>> new file mode 100644
>> index 000000000000..c1324efedcb9
>> --- /dev/null
>> +++ b/drivers/clk/nuvoton/Kconfig
>> @@ -0,0 +1,19 @@
>> +# SPDX-License-Identifier: GPL-2.0
>> +# common clock support for Nuvoton SoC family.
>> +
>> +config COMMON_CLK_NUVOTON
>> + bool "Nuvoton clock controller common support"
>> + depends on ARCH_NUVOTON
> Add || COMPILE_TEST
>
>> + default ARCH_NUVOTON
>> + help
>> + Say y here to enable common clock controller for Nuvoton platforms.
>> +
>> +if COMMON_CLK_NUVOTON
> Add a newline here.
>
>> +config CLK_MA35D1
>> + bool "Nuvoton MA35D1 clock controller support"
>> + depends on ARM64 || COMPILE_TEST
> Drop this
>
>> + default y
> And this.
>
>> + help
>> + Build the driver for MA35D1 Clock Driver.
> This says driver twice. "Build the clk driver for MA35D1 SoCs"?

I will fix all above of Kconfig file. Thank you.

>> +
>> +endif
>> diff --git a/drivers/clk/nuvoton/clk-ma35d1-divider.c b/drivers/clk/nuvoton/clk-ma35d1-divider.c
>> new file mode 100644
>> index 000000000000..340a889a1417
>> --- /dev/null
>> +++ b/drivers/clk/nuvoton/clk-ma35d1-divider.c
>> @@ -0,0 +1,140 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * Copyright (C) 2023 Nuvoton Technology Corp.
>> + * Author: Chi-Fang Li <[email protected]>
>> + */
>> +
>> +#include <linux/clk-provider.h>
>> +#include <linux/device.h>
>> +#include <linux/regmap.h>
>> +#include <linux/spinlock.h>
>> +
>> +#include "clk-ma35d1.h"
>> +
>> +#define clk_div_mask(width) ((1 << (width)) - 1)
> clk-provider.h defines this already, use that?

Yes, I will remove this.

>> +
>> +struct ma35d1_adc_clk_div {
>> + struct clk_hw hw;
>> + void __iomem *reg;
>> + u8 shift;
>> + u8 width;
>> + u32 mask;
>> + const struct clk_div_table *table;
>> + /* protects concurrent access to clock divider registers */
>> + spinlock_t *lock;
>> +};
>> +
>> +static inline struct ma35d1_adc_clk_div *to_ma35d1_adc_clk_div(struct clk_hw *_hw)
>> +{
>> + return container_of(_hw, struct ma35d1_adc_clk_div, hw);
>> +}
>> +
>> +static inline unsigned long ma35d1_clkdiv_recalc_rate(struct clk_hw *hw,
>> + unsigned long parent_rate)
>> +{
>> + unsigned int val;
>> + struct ma35d1_adc_clk_div *dclk = to_ma35d1_adc_clk_div(hw);
>> +
>> + val = readl_relaxed(dclk->reg) >> dclk->shift;
>> + val &= clk_div_mask(dclk->width);
>> + val += 1;
>> + return divider_recalc_rate(hw, parent_rate, val, dclk->table,
>> + CLK_DIVIDER_ROUND_CLOSEST, dclk->width);
>> +}
>> +
>> +static inline long ma35d1_clkdiv_round_rate(struct clk_hw *hw,
>> + unsigned long rate,
>> + unsigned long *prate)
>> +{
>> + struct ma35d1_adc_clk_div *dclk = to_ma35d1_adc_clk_div(hw);
>> +
>> + return divider_round_rate(hw, rate, prate, dclk->table,
>> + dclk->width, CLK_DIVIDER_ROUND_CLOSEST);
>> +}
>> +
>> +static inline int ma35d1_clkdiv_set_rate(struct clk_hw *hw,
>> + unsigned long rate,
>> + unsigned long parent_rate)
>> +{
>> + int value;
>> + unsigned long flags = 0;
>> + u32 data;
>> + struct ma35d1_adc_clk_div *dclk = to_ma35d1_adc_clk_div(hw);
>> +
>> + value = divider_get_val(rate, parent_rate, dclk->table,
>> + dclk->width, CLK_DIVIDER_ROUND_CLOSEST);
>> +
>> + if (dclk->lock)
>> + spin_lock_irqsave(dclk->lock, flags);
> Can you just always lock? That makes the conditional locking go away.

OK, I will remove the "if (dclk->lock)" check.

>
>> +
>> + data = readl_relaxed(dclk->reg);
>> + data &= ~(clk_div_mask(dclk->width) << dclk->shift);
>> + data |= (value - 1) << dclk->shift;
>> + data |= dclk->mask;
>> +
>> + writel_relaxed(data, dclk->reg);
>> +
>> + if (dclk->lock)
>> + spin_unlock_irqrestore(dclk->lock, flags);
>> +
>> + return 0;
>> +}
>> +
>> +static const struct clk_ops ma35d1_adc_clkdiv_ops = {
>> + .recalc_rate = ma35d1_clkdiv_recalc_rate,
>> + .round_rate = ma35d1_clkdiv_round_rate,
>> + .set_rate = ma35d1_clkdiv_set_rate,
>> +};
>> +
>> +struct clk_hw *ma35d1_reg_adc_clkdiv(struct device *dev, const char *name,
>> + const char *parent_name,
>> + spinlock_t *lock,
>> + unsigned long flags, void __iomem *reg,
>> + u8 shift, u8 width, u32 mask_bit)
>> +{
>> + struct ma35d1_adc_clk_div *div;
>> + struct clk_init_data init;
>> + struct clk_div_table *table;
>> + u32 max_div, min_div;
>> + struct clk_hw *hw;
>> + int ret;
>> + int i;
>> +
>> + div = devm_kzalloc(dev, sizeof(*div), GFP_KERNEL);
>> + if (!div)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + max_div = clk_div_mask(width) + 1;
>> + min_div = 1;
>> +
>> + table = devm_kcalloc(dev, max_div + 1, sizeof(*table), GFP_KERNEL);
>> + if (!table)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + for (i = 0; i < max_div; i++) {
>> + table[i].val = (min_div + i);
> Drop useless parenthesis.

I will fix it.

>
>> + table[i].div = 2 * table[i].val;
>> + }
>> + table[max_div].val = 0;
>> + table[max_div].div = 0;
>> +
>> + init.name = name;
>> + init.ops = &ma35d1_adc_clkdiv_ops;
>> + init.flags |= flags;
>> + init.parent_names = parent_name ? &parent_name : NULL;
>> + init.num_parents = parent_name ? 1 : 0;
>> +
>> + div->reg = reg;
>> + div->shift = shift;
>> + div->width = width;
>> + div->mask = mask_bit ? BIT(mask_bit) : 0;
>> + div->lock = lock;
>> + div->hw.init = &init;
>> + div->table = table;
>> +
>> + hw = &div->hw;
>> + ret = devm_clk_hw_register(dev, hw);
>> + if (ret < 0)
> Use if (ret) instead.

I will fix it.

>
>> + return ERR_PTR(ret);
>> + return hw;
>> +}
>> diff --git a/drivers/clk/nuvoton/clk-ma35d1-pll.c b/drivers/clk/nuvoton/clk-ma35d1-pll.c
>> new file mode 100644
>> index 000000000000..b36fbda4fa0a
>> --- /dev/null
>> +++ b/drivers/clk/nuvoton/clk-ma35d1-pll.c
>> @@ -0,0 +1,350 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * Copyright (C) 2023 Nuvoton Technology Corp.
>> + * Author: Chi-Fang Li <[email protected]>
>> + */
>> +
>> +#include <linux/bitfield.h>
>> +#include <linux/clk-provider.h>
>> +#include <linux/slab.h>
>> +#include <linux/regmap.h>
>> +
>> +#include "clk-ma35d1.h"
>> +
>> +#define REG_SYS_RLKTZNS 0x1a4 /* Register Lock Control */
>> +
>> +struct ma35d1_clk_pll {
>> + struct clk_hw hw;
>> + u8 type;
>> + u8 mode;
>> + unsigned long rate;
>> + void __iomem *ctl0_base;
> This needs an include for __iomem define. Just include kernel.h for now.

OK, I will add it.

>> + void __iomem *ctl1_base;
>> + void __iomem *ctl2_base;
>> + struct regmap *regmap;
>> +};
>> +
>> +static inline struct ma35d1_clk_pll *to_ma35d1_clk_pll(struct clk_hw *_hw)
>> +{
>> + return container_of(_hw, struct ma35d1_clk_pll, hw);
> Missing include as well.

I will add the
#include <linux/container_of.h>

>
>> +}
>> +
>> +static void ma35d1_unlock_regs(struct ma35d1_clk_pll *pll)
>> +{
>> + int ret;
>> +
>> + do {
>> + regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x59);
>> + regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x16);
>> + regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x88);
>> + regmap_read(pll->regmap, REG_SYS_RLKTZNS, &ret);
>> + } while (ret == 0);
> Any sort of timeout?

As we have discussed in [PATCH v6 08/12] of this patchset, we will
remove the register lock.
So, this function will be removed in the next version.

>> +}
>> +
>> +static void ma35d1_lock_regs(struct ma35d1_clk_pll *pll)
>> +{
>> + regmap_write(pll->regmap, REG_SYS_RLKTZNS, 0x0);
>> +}
>> +
>> +static unsigned long ma35d1_calc_smic_pll_freq(u32 pll0_ctl0,
>> + unsigned long parent_rate)
>> +{
>> + u32 m, n, p, outdiv;
>> + u64 pll_freq;
>> + u32 clk_div_table[] = { 1, 2, 4, 8 };
> const

OK, I will add const to the clk_div_table[].
>
>> +
>> + if (pll0_ctl0 & SPLL0_CTL0_BP)
>> + return parent_rate;
>> +
>> + n = FIELD_GET(SPLL0_CTL0_FBDIV, pll0_ctl0);
>> + m = FIELD_GET(SPLL0_CTL0_INDIV, pll0_ctl0);
>> + p = FIELD_GET(SPLL0_CTL0_OUTDIV, pll0_ctl0);
>> + outdiv = clk_div_table[p];
>> + pll_freq = (u64)parent_rate * n;
>> + do_div(pll_freq, m * outdiv);
>> + return (unsigned long)pll_freq;
> Remove useless cast.

I will fix it.

>> +}
>> +
>> +static unsigned long ma35d1_calc_pll_freq(u8 mode, u32 *reg_ctl,
>> + unsigned long parent_rate)
>> +{
>> + u32 m, n, p;
>> + u64 pll_freq, x;
>> +
>> + if (reg_ctl[1] & PLL_CTL1_BP)
>> + return parent_rate;
>> +
>> + if (mode == PLL_MODE_INT) {
>> + n = FIELD_GET(PLL_CTL0_FBDIV, reg_ctl[0]);
>> + m = FIELD_GET(PLL_CTL0_INDIV, reg_ctl[0]);
>> + p = FIELD_GET(PLL_CTL1_OUTDIV, reg_ctl[1]);
>> + pll_freq = (u64)parent_rate * n;
>> + do_div(pll_freq, m * p);
>> + } else {
>> + n = FIELD_GET(PLL_CTL0_FBDIV, reg_ctl[0]);
>> + m = FIELD_GET(PLL_CTL0_INDIV, reg_ctl[0]);
>> + p = FIELD_GET(PLL_CTL1_OUTDIV, reg_ctl[1]);
> The n, m, p are the same, so pull them out of the if-else to deduplicate.

Ok, I will modify it.

>
>> + x = FIELD_GET(PLL_CTL1_FRAC, reg_ctl[1]);
>> + /* 2 decimal places floating to integer (ex. 1.23 to 123) */
>> + n = n * 100 + ((x * 100) / FIELD_MAX(PLL_CTL1_FRAC));
> Is this mult_frac()?

Not exactly. The number is represented as n.x.
Suppose the number is 123.45, then n is 123 and x is 45.
We have it be multiplied with 100 to be integer 12345, which is n.

In the following
pll_freq = ((u64)parent_rate * n) / 100 / m / p;

We can keep precision of n.

It seems not applicable to mult_frac().

>
>> + pll_freq = ((u64)parent_rate * n) / 100 / m / p;
>> + }
> Add a newline
>
>> + return (unsigned long)pll_freq;
> Remove useless cast.

I will fix them.

>> +}
>> +
>> +static int ma35d1_pll_find_closest(struct ma35d1_clk_pll *pll,
>> + unsigned long rate,
>> + unsigned long parent_rate,
>> + u32 *reg_ctl, unsigned long *freq)
>> +{
>> + int p, m, n;
>> + int fbdiv_min, fbdiv_max;
>> + unsigned long diff = 0xffffffff;
>> +
>> + *freq = 0;
>> + if ((rate < PLL_FCLKO_MIN_FREQ) || (rate > PLL_FCLKO_MAX_FREQ))
>> + return -EINVAL;
>> +
>> + if (pll->mode == PLL_MODE_INT) {
>> + fbdiv_min = FBDIV_MIN;
>> + fbdiv_max = FBDIV_MAX;
>> + } else {
>> + fbdiv_min = FBDIV_FRAC_MIN;
>> + fbdiv_max = FBDIV_FRAC_MAX;
>> + }
>> +
>> + for (m = INDIV_MIN; m <= INDIV_MAX; m++) {
>> + for (n = fbdiv_min; n <= fbdiv_max; n++) {
>> + for (p = OUTDIV_MIN; p <= OUTDIV_MAX; p++) {
>> + unsigned long tmp, fout;
>> + u64 fclk;
>> +
>> + tmp = parent_rate / m;
>> + if (tmp < PLL_FREF_M_MIN_FREQ ||
>> + tmp > PLL_FREF_M_MAX_FREQ)
>> + continue; /* constrain */
>> +
>> + fclk = (u64)parent_rate * n / m;
>> + /* for 2 decimal places */
>> + if (pll->mode != PLL_MODE_INT)
>> + fclk /= 100;
>> +
>> + if (fclk < PLL_FCLK_MIN_FREQ ||
>> + fclk > PLL_FCLK_MAX_FREQ)
>> + continue; /* constrain */
>> +
>> + fout = (unsigned long)(fclk / p);
>> + if (fout < PLL_FCLKO_MIN_FREQ ||
>> + fout > PLL_FCLKO_MAX_FREQ)
>> + continue; /* constrain */
>> +
>> + if (abs(rate - fout) < diff) {
>> + reg_ctl[0] = FIELD_PREP(PLL_CTL0_INDIV, m) |
>> + FIELD_PREP(PLL_CTL0_FBDIV, n);
>> + reg_ctl[1] = FIELD_PREP(PLL_CTL1_OUTDIV, p);
>> + *freq = fout;
>> + diff = abs(rate - fout);
>> + if (diff == 0)
>> + break;
>> + }
>> + }
>> + }
>> + }
>> + if (*freq == 0)
>> + return -EINVAL; /* cannot find even one valid setting */
>> + return 0;
>> +}
>> +
>> +static int ma35d1_clk_pll_set_rate(struct clk_hw *hw, unsigned long rate,
>> + unsigned long parent_rate)
>> +{
>> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
>> + u32 reg_ctl[3] = { 0 };
>> + unsigned long pll_freq;
>> + int ret;
>> +
>> + if ((parent_rate < PLL_FREF_MIN_FREQ) ||
>> + (parent_rate > PLL_FREF_MAX_FREQ))
> Remove useless parenthesis.
>
>> + return -EINVAL;
>> +
>> + if ((pll->type == MA35D1_CAPLL) || (pll->type == MA35D1_DDRPLL)) {
> Remove useless parenthesis.

I will fix them.

>> + pr_warn("Nuvoton MA35D1 CAPLL/DDRPLL is read only.\n");
>> + return -EACCES;
>> + }
>> +
>> + ret = ma35d1_pll_find_closest(pll, rate, parent_rate, reg_ctl, &pll_freq);
>> + if (ret != 0)
>> + return ret;
>> + pll->rate = pll_freq;
>> +
>> + switch (pll->mode) {
>> + case PLL_MODE_INT:
>> + reg_ctl[0] |= FIELD_PREP(PLL_CTL0_MODE, PLL_MODE_INT);
>> + break;
>> + case PLL_MODE_FRAC:
>> + reg_ctl[0] |= FIELD_PREP(PLL_CTL0_MODE, PLL_MODE_FRAC);
>> + break;
>> + case PLL_MODE_SS:
>> + reg_ctl[0] |= FIELD_PREP(PLL_CTL0_MODE, PLL_MODE_SS) |
>> + FIELD_PREP(PLL_CTL0_SSRATE, PLL_SS_RATE);
>> + reg_ctl[2] = FIELD_PREP(PLL_CTL2_SLOPE, PLL_SLOPE);
>> + break;
>> + }
>> + reg_ctl[1] |= PLL_CTL1_PD;
>> +
>> + ma35d1_unlock_regs(pll);
>> + writel_relaxed(reg_ctl[0], pll->ctl0_base);
>> + writel_relaxed(reg_ctl[1], pll->ctl1_base);
>> + writel_relaxed(reg_ctl[2], pll->ctl2_base);
>> + ma35d1_lock_regs(pll);
>> + return 0;
>> +}
>> +
>> +static unsigned long ma35d1_clk_pll_recalc_rate(struct clk_hw *hw,
>> + unsigned long parent_rate)
>> +{
>> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
>> + u32 reg_ctl[3];
>> + unsigned long pll_freq;
>> +
>> + if ((parent_rate < PLL_FREF_MIN_FREQ) || (parent_rate > PLL_FREF_MAX_FREQ))
> Remove useless parenthesis.

I will fix it.

>
>> + return 0;
>> +
>> + switch (pll->type) {
>> + case MA35D1_CAPLL:
>> + reg_ctl[0] = readl_relaxed(pll->ctl0_base);
>> + pll_freq = ma35d1_calc_smic_pll_freq(reg_ctl[0], parent_rate);
> return?

Yes, I will have it return here.

>> + break;
>> + case MA35D1_DDRPLL:
>> + case MA35D1_APLL:
>> + case MA35D1_EPLL:
>> + case MA35D1_VPLL:
>> + reg_ctl[0] = readl_relaxed(pll->ctl0_base);
>> + reg_ctl[1] = readl_relaxed(pll->ctl1_base);
>> + pll_freq = ma35d1_calc_pll_freq(pll->mode, reg_ctl, parent_rate);
> return?
>
>> + break;
>> + }
>> + return pll_freq;
> return 0?

Yes, I will fix them.

>
>> +}
>> +
>> +static long ma35d1_clk_pll_round_rate(struct clk_hw *hw, unsigned long rate,
>> + unsigned long *parent_rate)
>> +{
>> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
>> + u32 reg_ctl[3] = { 0 };
>> + unsigned long pll_freq;
>> + long ret;
>> +
>> + if (*parent_rate < PLL_FREF_MIN_FREQ || *parent_rate > PLL_FREF_MAX_FREQ)
>> + return -EINVAL;
>> +
>> + ret = ma35d1_pll_find_closest(pll, rate, *parent_rate, reg_ctl, &pll_freq);
>> + if (ret != 0)
>> + return ret;
>> +
>> + switch (pll->type) {
>> + case MA35D1_CAPLL:
>> + reg_ctl[0] = readl_relaxed(pll->ctl0_base);
>> + pll_freq = ma35d1_calc_smic_pll_freq(reg_ctl[0], *parent_rate);
> return?
>
>> + break;
>> + case MA35D1_DDRPLL:
>> + case MA35D1_APLL:
>> + case MA35D1_EPLL:
>> + case MA35D1_VPLL:
>> + reg_ctl[0] = readl_relaxed(pll->ctl0_base);
>> + reg_ctl[1] = readl_relaxed(pll->ctl1_base);
>> + pll_freq = ma35d1_calc_pll_freq(pll->mode, reg_ctl, *parent_rate);
> return?
>
>> + break;
>> + }
>> + return pll_freq;
> return 0?

Yes, I will fix them.

>> +}
>> +
>> +static int ma35d1_clk_pll_is_prepared(struct clk_hw *hw)
>> +{
>> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
>> + u32 val = readl_relaxed(pll->ctl1_base);
>> +
>> + return val & PLL_CTL1_PD ? 0 : 1;
>> +}
>> +
>> +static int ma35d1_clk_pll_prepare(struct clk_hw *hw)
>> +{
>> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
>> + u32 val;
>> +
>> + if ((pll->type == MA35D1_CAPLL) || (pll->type == MA35D1_DDRPLL)) {
> Drop useless parenthesis.

I will fix it.

>
>> + pr_warn("Nuvoton MA35D1 CAPLL/DDRPLL is read only.\n");
>> + return -EACCES;
>> + }
>> +
>> + ma35d1_unlock_regs(pll);
>> + val = readl_relaxed(pll->ctl1_base);
>> + val &= ~PLL_CTL1_PD;
>> + writel_relaxed(val, pll->ctl1_base);
>> + ma35d1_lock_regs(pll);
>> + return 0;
>> +}
>> +
>> +static void ma35d1_clk_pll_unprepare(struct clk_hw *hw)
>> +{
>> + struct ma35d1_clk_pll *pll = to_ma35d1_clk_pll(hw);
>> + u32 val;
>> +
>> + if ((pll->type == MA35D1_CAPLL) || (pll->type == MA35D1_DDRPLL)) {
> Drop useless parenthesis.

I will fix it.

>
>> + pr_warn("Nuvoton MA35D1 CAPLL/DDRPLL is read only.\n");
> These read-only checks should be removed and different clk_ops should be
> assigned for the read-only type of clks.

OK, I will fix it.

>> + } else {
>> + val = readl_relaxed(pll->ctl1_base);
>> + val |= PLL_CTL1_PD;
>> + writel_relaxed(val, pll->ctl1_base);
>> + }
>> +}
>> +
>> +static const struct clk_ops ma35d1_clk_pll_ops = {
>> + .is_prepared = ma35d1_clk_pll_is_prepared,
>> + .prepare = ma35d1_clk_pll_prepare,
>> + .unprepare = ma35d1_clk_pll_unprepare,
>> + .set_rate = ma35d1_clk_pll_set_rate,
>> + .recalc_rate = ma35d1_clk_pll_recalc_rate,
>> + .round_rate = ma35d1_clk_pll_round_rate,
>> +};
>> +
>> +struct clk_hw *ma35d1_reg_clk_pll(enum ma35d1_pll_type type,
> Pass a device here.

OK, I will add it.

>> + u8 u8mode, const char *name,
>> + const char *parent,
>> + unsigned long targetFreq,
>> + void __iomem *base,
>> + struct regmap *regmap)
>> +{
>> + struct ma35d1_clk_pll *pll;
>> + struct clk_hw *hw;
>> + struct clk_init_data init;
> init = {};
>
>> + int ret;
>> +
>> + pll = kmalloc(sizeof(*pll), GFP_KERNEL);
> Just use kzalloc(). This isn't a hot path.

I will use devm_kzalloc().

>> + if (!pll)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + pll->type = type;
>> + pll->mode = u8mode;
>> + pll->rate = targetFreq;
>> + pll->ctl0_base = base + REG_PLL_CTL0_OFFSET;
>> + pll->ctl1_base = base + REG_PLL_CTL1_OFFSET;
>> + pll->ctl2_base = base + REG_PLL_CTL2_OFFSET;
>> + pll->regmap = regmap;
>> +
>> + init.name = name;
>> + init.flags = 0;
>> + init.parent_names = &parent;
>> + init.num_parents = 1;
>> + init.ops = &ma35d1_clk_pll_ops;
>> + pll->hw.init = &init;
>> + hw = &pll->hw;
>> +
>> + ret = clk_hw_register(NULL, hw);
> Use devm_

OK, I will fix it.

>
>> + if (ret) {
>> + pr_err("failed to register vsi-pll clock!!!\n");
> I'm going to put a dev_err_probe() into clk registration code. Remove
> this printk.

OK, I will remove it.

>> + kfree(pll);
>> + return ERR_PTR(ret);
>> + }
>> + return hw;
>> +}
>> diff --git a/drivers/clk/nuvoton/clk-ma35d1.c b/drivers/clk/nuvoton/clk-ma35d1.c
>> new file mode 100644
>> index 000000000000..e4d3ced396a3
>> --- /dev/null
>> +++ b/drivers/clk/nuvoton/clk-ma35d1.c
>> @@ -0,0 +1,963 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * Copyright (C) 2023 Nuvoton Technology Corp.
>> + * Author: Chi-Fang Li <[email protected]>
>> + */
>> +
>> +#include <linux/clk-provider.h>
>> +#include <linux/mfd/syscon.h>
>> +#include <linux/module.h>
>> +#include <linux/of.h>
>> +#include <linux/of_address.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/spinlock.h>
>> +#include <dt-bindings/clock/nuvoton,ma35d1-clk.h>
>> +
>> +#include "clk-ma35d1.h"
>> +
>> +static DEFINE_SPINLOCK(ma35d1_lock);
>> +
>> +static const char *const ca35clk_sel_clks[] = {
> Use clk_parent_data instead.

I will fix it.

>
>> + "hxt", "capll", "ddrpll", "dummy"
>> +};
>> +
> [...]
>> +
>> +static inline struct clk_hw *ma35d1_clk_fixed(const char *name, int rate)
>> +{
>> + return clk_hw_register_fixed_rate(NULL, name, NULL, 0, rate);
>> +}
>> +
>> +static inline struct clk_hw *ma35d1_clk_mux(const char *name,
>> + void __iomem *reg,
>> + u8 shift, u8 width,
>> + const char *const *parents,
>> + int num_parents)
>> +{
>> + return clk_hw_register_mux(NULL, name, parents, num_parents,
>> + CLK_SET_RATE_NO_REPARENT, reg, shift,
>> + width, 0, &ma35d1_lock);
>> +}
>> +
>> +static inline struct clk_hw *ma35d1_clk_divider(const char *name,
>> + const char *parent,
>> + void __iomem *reg, u8 shift,
>> + u8 width)
>> +{
>> + return clk_hw_register_divider(NULL, name, parent, CLK_SET_RATE_PARENT,
>> + reg, shift, width, 0, &ma35d1_lock);
>> +}
>> +
>> +static inline struct clk_hw *ma35d1_clk_divider_pow2(const char *name,
>> + const char *parent,
>> + void __iomem *reg,
>> + u8 shift, u8 width)
>> +{
>> + return clk_hw_register_divider(NULL, name, parent,
>> + CLK_DIVIDER_POWER_OF_TWO, reg, shift,
>> + width, 0, &ma35d1_lock);
>> +}
>> +
>> +static inline struct clk_hw *ma35d1_clk_divider_table(const char *name,
>> + const char *parent,
>> + void __iomem *reg,
>> + u8 shift, u8 width,
>> + const struct clk_div_table *table)
>> +{
>> + return clk_hw_register_divider_table(NULL, name, parent, 0,
>> + reg, shift, width, 0, table,
>> + &ma35d1_lock);
>> +}
>> +
>> +static inline struct clk_hw *ma35d1_clk_fixed_factor(const char *name,
>> + const char *parent,
>> + unsigned int mult,
>> + unsigned int div)
>> +{
>> + return clk_hw_register_fixed_factor(NULL, name, parent,
>> + CLK_SET_RATE_PARENT, mult, div);
>> +}
>> +
>> +static inline struct clk_hw *ma35d1_clk_gate(const char *name,
>> + const char *parent,
>> + void __iomem *reg, u8 shift)
>> +{
>> + return clk_hw_register_gate(NULL, name, parent, CLK_SET_RATE_PARENT,
>> + reg, shift, 0, &ma35d1_lock);
>> +}
> These need to use devm_ to unregister on failure.

OK, I will fix them all.

>> +
>> +static int ma35d1_get_pll_setting(struct device_node *clk_node,
>> + u32 *pllmode, u32 *pllfreq)
>> +{
>> + const char *of_str;
>> + int i;
>> +
>> + for (i = 0; i < PLL_MAX_NUM; i++) {
>> + if (of_property_read_string_index(clk_node, "nuvoton,pll-mode", i, &of_str))
>> + return -EINVAL;
>> + if (!strcmp(of_str, "integer"))
>> + pllmode[i] = PLL_MODE_INT;
>> + else if (!strcmp(of_str, "fractional"))
>> + pllmode[i] = PLL_MODE_FRAC;
>> + else if (!strcmp(of_str, "spread-spectrum"))
>> + pllmode[i] = PLL_MODE_SS;
>> + else
>> + return -EINVAL;
>> + }
>> + return of_property_read_u32_array(clk_node, "assigned-clock-rates",
> The clk framework looks at this property. Why is the driver looking at
> it too?

We should not use "assigned-clock-rates", instead we should create a
vendor-specific property to describe the pll frequency array.

I will fix it in the next version.


>> + pllfreq, PLL_MAX_NUM);
>> +}
>> +


Best regards,
Jacky Huang

2023-03-30 10:45:22

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform

Dear Rob,


On 2023/3/29 下午 09:07, Rob Herring wrote:
> On Mon, Mar 27, 2023 at 9:19 PM Jacky Huang <[email protected]> wrote:
>> From: Jacky Huang <[email protected]>
>>
>> Rename the bindings/arm/npcm directory as nuvoton.
>> Add binding for ARMv8 based Nuvotn SoCs and platform boards.
>> Add initial bindings for ma35d1 series development boards.
>>
>> Signed-off-by: Jacky Huang <[email protected]>
>> ---
>> .../bindings/arm/nuvoton/nuvoton,ma35d1.yaml | 30 +++++++++++++++++++
>> .../nuvoton,npcm-gcr.yaml} | 2 +-
>> .../npcm.yaml => nuvoton/nuvoton,npcm.yaml} | 2 +-
>> 3 files changed, 32 insertions(+), 2 deletions(-)
>> create mode 100644 Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml
>> rename Documentation/devicetree/bindings/arm/{npcm/nuvoton,gcr.yaml => nuvoton/nuvoton,npcm-gcr.yaml} (93%)
> Since you are moving it, this one should be moved to bindings/soc/nuvoton/
>
>> rename Documentation/devicetree/bindings/arm/{npcm/npcm.yaml => nuvoton/nuvoton,npcm.yaml} (93%)

OK, I will move "Documentation/devicetree/bindings/arm/npcm" to
"Documentation/devicetree/bindings/soc/nuvoton".


Best regards,
Jacky Huang


2023-03-30 10:54:10

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 07/12] dt-bindings: serial: Document ma35d1 uart controller



On 2023/3/30 下午 03:33, Krzysztof Kozlowski wrote:
> On 29/03/2023 10:44, Jacky Huang wrote:
>> Dear Krzysztof,
>>
>>
>> Thanks for your review.
>>
>>
>> On 2023/3/29 下午 04:20, Krzysztof Kozlowski wrote:
>>> On 28/03/2023 04:19, Jacky Huang wrote:
>>>> From: Jacky Huang <[email protected]>
>>>>
>>>> Add documentation to describe nuvoton ma35d1 uart driver bindings.
>>>>
>>> This depends on clock patches, so it must be taken together.
>>>
>>> Reviewed-by: Krzysztof Kozlowski <[email protected]>
>>>
>>> Best regards,
>>> Krzysztof
>>>
>> OK, I will combine this patch with the clock patch.
> I don't know what you mean by "combine", but I don't think this is what
> we want. You must clearly state all dependencies and patch merging
> strategy/way.
>
> Best regards,
> Krzysztof
>

Dear Krzysztof,


I may be misunderstanding your request.
I thought it was moving the "nuvoton,ma35d1-serial.yaml" to the patch of
nuvoton,ma35d1-clk.h.
Would you please provide guidance what should I do next on this [PATCH
v6 07/12]. Thank you.


Best regards,
Jacky Huang







2023-03-30 13:18:51

by Rob Herring

[permalink] [raw]
Subject: Re: [PATCH v6 07/12] dt-bindings: serial: Document ma35d1 uart controller

On Thu, Mar 30, 2023 at 06:52:11PM +0800, Jacky Huang wrote:
>
>
> On 2023/3/30 下午 03:33, Krzysztof Kozlowski wrote:
> > On 29/03/2023 10:44, Jacky Huang wrote:
> > > Dear Krzysztof,
> > >
> > >
> > > Thanks for your review.
> > >
> > >
> > > On 2023/3/29 下午 04:20, Krzysztof Kozlowski wrote:
> > > > On 28/03/2023 04:19, Jacky Huang wrote:
> > > > > From: Jacky Huang <[email protected]>
> > > > >
> > > > > Add documentation to describe nuvoton ma35d1 uart driver bindings.
> > > > >
> > > > This depends on clock patches, so it must be taken together.
> > > >
> > > > Reviewed-by: Krzysztof Kozlowski <[email protected]>
> > > >
> > > > Best regards,
> > > > Krzysztof
> > > >
> > > OK, I will combine this patch with the clock patch.
> > I don't know what you mean by "combine", but I don't think this is what
> > we want. You must clearly state all dependencies and patch merging
> > strategy/way.
> >
> > Best regards,
> > Krzysztof
> >
>
> Dear Krzysztof,
>
>
> I may be misunderstanding your request.
> I thought it was moving the "nuvoton,ma35d1-serial.yaml" to the patch of
> nuvoton,ma35d1-clk.h.
> Would you please provide guidance what should I do next on this [PATCH v6
> 07/12]. Thank you.

The only combining you should do is putting the 2 patches in the same
series. Combining to 1 patch would be wrong.

Your other option is just drop the header in the example and use the
raw numbers. Then there is no dependency to manage and each subsystem
maintainer can take the relevant patches.

Rob

2023-03-30 13:28:38

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform

On 30/03/2023 12:41, Jacky Huang wrote:
> Dear Rob,
>
>
> On 2023/3/29 下午 09:07, Rob Herring wrote:
>> On Mon, Mar 27, 2023 at 9:19 PM Jacky Huang <[email protected]> wrote:
>>> From: Jacky Huang <[email protected]>
>>>
>>> Rename the bindings/arm/npcm directory as nuvoton.
>>> Add binding for ARMv8 based Nuvotn SoCs and platform boards.
>>> Add initial bindings for ma35d1 series development boards.
>>>
>>> Signed-off-by: Jacky Huang <[email protected]>
>>> ---
>>> .../bindings/arm/nuvoton/nuvoton,ma35d1.yaml | 30 +++++++++++++++++++
>>> .../nuvoton,npcm-gcr.yaml} | 2 +-
>>> .../npcm.yaml => nuvoton/nuvoton,npcm.yaml} | 2 +-
>>> 3 files changed, 32 insertions(+), 2 deletions(-)
>>> create mode 100644 Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml
>>> rename Documentation/devicetree/bindings/arm/{npcm/nuvoton,gcr.yaml => nuvoton/nuvoton,npcm-gcr.yaml} (93%)
>> Since you are moving it, this one should be moved to bindings/soc/nuvoton/
>>
>>> rename Documentation/devicetree/bindings/arm/{npcm/npcm.yaml => nuvoton/nuvoton,npcm.yaml} (93%)
>
> OK, I will move "Documentation/devicetree/bindings/arm/npcm" to
> "Documentation/devicetree/bindings/soc/nuvoton".

To clarify - Rob asked *this one*. Not all.



Best regards,
Krzysztof

2023-03-31 00:32:18

by kernel test robot

[permalink] [raw]
Subject: Re: [PATCH v6 11/12] tty: serial: Add Nuvoton ma35d1 serial driver support

Hi Jacky,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on robh/for-next]
[also build test WARNING on clk/clk-next linus/master]
[cannot apply to pza/reset/next pza/imx-drm/next]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url: https://github.com/intel-lab-lkp/linux/commits/Jacky-Huang/arm64-Kconfig-platforms-Add-config-for-Nuvoton-MA35-platform/20230328-102245
base: https://git.kernel.org/pub/scm/linux/kernel/git/robh/linux.git for-next
patch link: https://lore.kernel.org/r/20230328021912.177301-12-ychuang570808%40gmail.com
patch subject: [PATCH v6 11/12] tty: serial: Add Nuvoton ma35d1 serial driver support
config: powerpc-allyesconfig (https://download.01.org/0day-ci/archive/20230331/[email protected]/config)
compiler: clang version 17.0.0 (https://github.com/llvm/llvm-project 67409911353323ca5edf2049ef0df54132fa1ca7)
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# install powerpc cross compiling tool for clang build
# apt-get install binutils-powerpc-linux-gnu
# https://github.com/intel-lab-lkp/linux/commit/380d83a62e873855024ca4c660865c654a62748a
git remote add linux-review https://github.com/intel-lab-lkp/linux
git fetch --no-tags linux-review Jacky-Huang/arm64-Kconfig-platforms-Add-config-for-Nuvoton-MA35-platform/20230328-102245
git checkout 380d83a62e873855024ca4c660865c654a62748a
# save the config file
mkdir build_dir && cp config build_dir/.config
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=powerpc olddefconfig
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=powerpc SHELL=/bin/bash drivers/tty/serial/

If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <[email protected]>
| Link: https://lore.kernel.org/oe-kbuild-all/[email protected]/

All warnings (new ones prefixed by >>):

>> drivers/tty/serial/ma35d1_serial.c:672:6: warning: variable 'ret' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized]
if (pdev->dev.of_node) {
^~~~~~~~~~~~~~~~~
drivers/tty/serial/ma35d1_serial.c:679:27: note: uninitialized use occurs here
up = &ma35d1serial_ports[ret];
^~~
drivers/tty/serial/ma35d1_serial.c:672:2: note: remove the 'if' if its condition is always true
if (pdev->dev.of_node) {
^~~~~~~~~~~~~~~~~~~~~~~
drivers/tty/serial/ma35d1_serial.c:668:9: note: initialize the variable 'ret' to silence this warning
int ret;
^
= 0
>> drivers/tty/serial/ma35d1_serial.c:730:6: warning: variable 'i' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized]
if (dev->dev.of_node)
^~~~~~~~~~~~~~~~
drivers/tty/serial/ma35d1_serial.c:732:6: note: uninitialized use occurs here
if (i < 0) {
^
drivers/tty/serial/ma35d1_serial.c:730:2: note: remove the 'if' if its condition is always true
if (dev->dev.of_node)
^~~~~~~~~~~~~~~~~~~~~
drivers/tty/serial/ma35d1_serial.c:727:7: note: initialize the variable 'i' to silence this warning
int i;
^
= 0
drivers/tty/serial/ma35d1_serial.c:750:6: warning: variable 'i' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized]
if (dev->dev.of_node)
^~~~~~~~~~~~~~~~
drivers/tty/serial/ma35d1_serial.c:752:6: note: uninitialized use occurs here
if (i < 0) {
^
drivers/tty/serial/ma35d1_serial.c:750:2: note: remove the 'if' if its condition is always true
if (dev->dev.of_node)
^~~~~~~~~~~~~~~~~~~~~
drivers/tty/serial/ma35d1_serial.c:747:7: note: initialize the variable 'i' to silence this warning
int i;
^
= 0
3 warnings generated.


vim +672 drivers/tty/serial/ma35d1_serial.c

658
659 /*
660 * Register a set of serial devices attached to a platform device.
661 * The list is terminated with a zero flags entry, which means we expect
662 * all entries to have at least UPF_BOOT_AUTOCONF set.
663 */
664 static int ma35d1serial_probe(struct platform_device *pdev)
665 {
666 struct resource *res_mem;
667 struct uart_ma35d1_port *up;
668 int ret;
669 struct clk *clk;
670 int err;
671
> 672 if (pdev->dev.of_node) {
673 ret = of_alias_get_id(pdev->dev.of_node, "serial");
674 if (ret < 0) {
675 dev_err(&pdev->dev, "failed to get alias/pdev id, errno %d\n", ret);
676 return ret;
677 }
678 }
679 up = &ma35d1serial_ports[ret];
680 up->port.line = ret;
681 res_mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
682 if (!res_mem)
683 return -ENODEV;
684
685 up->port.iobase = res_mem->start;
686 up->port.membase = ioremap(up->port.iobase, UART_REG_SIZE);
687 up->port.ops = &ma35d1serial_ops;
688
689 spin_lock_init(&up->port.lock);
690
691 clk = of_clk_get(pdev->dev.of_node, 0);
692 if (IS_ERR(clk)) {
693 err = PTR_ERR(clk);
694 dev_err(&pdev->dev, "failed to get core clk: %d\n", err);
695 return -ENOENT;
696 }
697 err = clk_prepare_enable(clk);
698 if (err)
699 return -ENOENT;
700
701 if (up->port.line != 0)
702 up->port.uartclk = clk_get_rate(clk);
703 up->port.irq = platform_get_irq(pdev, 0);
704 up->port.dev = &pdev->dev;
705 up->port.flags = UPF_BOOT_AUTOCONF;
706 ret = uart_add_one_port(&ma35d1serial_reg, &up->port);
707 platform_set_drvdata(pdev, up);
708 return 0;
709 }
710
711 /*
712 * Remove serial ports registered against a platform device.
713 */
714 static int ma35d1serial_remove(struct platform_device *dev)
715 {
716 struct uart_port *port = platform_get_drvdata(dev);
717
718 if (port) {
719 uart_remove_one_port(&ma35d1serial_reg, port);
720 free_irq(port->irq, port);
721 }
722 return 0;
723 }
724
725 static int ma35d1serial_suspend(struct platform_device *dev, pm_message_t state)
726 {
727 int i;
728 struct uart_ma35d1_port *up;
729
> 730 if (dev->dev.of_node)
731 i = of_alias_get_id(dev->dev.of_node, "serial");
732 if (i < 0) {
733 dev_err(&dev->dev, "failed to get alias/pdev id, errno %d\n", i);
734 return i;
735 }
736 up = &ma35d1serial_ports[i];
737 if (i == 0) {
738 up->console_baud_rate = serial_in(up, UART_REG_BAUD);
739 up->console_line = serial_in(up, UART_REG_LCR);
740 up->console_int = serial_in(up, UART_REG_IER);
741 }
742 return 0;
743 }
744

--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests

2023-03-31 02:21:50

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 07/12] dt-bindings: serial: Document ma35d1 uart controller



On 2023/3/30 下午 09:12, Rob Herring wrote:
> On Thu, Mar 30, 2023 at 06:52:11PM +0800, Jacky Huang wrote:
>>
>> On 2023/3/30 下午 03:33, Krzysztof Kozlowski wrote:
>>> On 29/03/2023 10:44, Jacky Huang wrote:
>>>> Dear Krzysztof,
>>>>
>>>>
>>>> Thanks for your review.
>>>>
>>>>
>>>> On 2023/3/29 下午 04:20, Krzysztof Kozlowski wrote:
>>>>> On 28/03/2023 04:19, Jacky Huang wrote:
>>>>>> From: Jacky Huang <[email protected]>
>>>>>>
>>>>>> Add documentation to describe nuvoton ma35d1 uart driver bindings.
>>>>>>
>>>>> This depends on clock patches, so it must be taken together.
>>>>>
>>>>> Reviewed-by: Krzysztof Kozlowski <[email protected]>
>>>>>
>>>>> Best regards,
>>>>> Krzysztof
>>>>>
>>>> OK, I will combine this patch with the clock patch.
>>> I don't know what you mean by "combine", but I don't think this is what
>>> we want. You must clearly state all dependencies and patch merging
>>> strategy/way.
>>>
>>> Best regards,
>>> Krzysztof
>>>
>> Dear Krzysztof,
>>
>>
>> I may be misunderstanding your request.
>> I thought it was moving the "nuvoton,ma35d1-serial.yaml" to the patch of
>> nuvoton,ma35d1-clk.h.
>> Would you please provide guidance what should I do next on this [PATCH v6
>> 07/12]. Thank you.
> The only combining you should do is putting the 2 patches in the same
> series. Combining to 1 patch would be wrong.
>
> Your other option is just drop the header in the example and use the
> raw numbers. Then there is no dependency to manage and each subsystem
> maintainer can take the relevant patches.
>
> Rob

Dear Rob,

Thanks for your detailed explanation. I got it now.


Best regards,
Jacky Huang




2023-03-31 02:24:32

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform



On 2023/3/30 下午 09:25, Krzysztof Kozlowski wrote:
> On 30/03/2023 12:41, Jacky Huang wrote:
>> Dear Rob,
>>
>>
>> On 2023/3/29 下午 09:07, Rob Herring wrote:
>>> On Mon, Mar 27, 2023 at 9:19 PM Jacky Huang <[email protected]> wrote:
>>>> From: Jacky Huang <[email protected]>
>>>>
>>>> Rename the bindings/arm/npcm directory as nuvoton.
>>>> Add binding for ARMv8 based Nuvotn SoCs and platform boards.
>>>> Add initial bindings for ma35d1 series development boards.
>>>>
>>>> Signed-off-by: Jacky Huang <[email protected]>
>>>> ---
>>>> .../bindings/arm/nuvoton/nuvoton,ma35d1.yaml | 30 +++++++++++++++++++
>>>> .../nuvoton,npcm-gcr.yaml} | 2 +-
>>>> .../npcm.yaml => nuvoton/nuvoton,npcm.yaml} | 2 +-
>>>> 3 files changed, 32 insertions(+), 2 deletions(-)
>>>> create mode 100644 Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml
>>>> rename Documentation/devicetree/bindings/arm/{npcm/nuvoton,gcr.yaml => nuvoton/nuvoton,npcm-gcr.yaml} (93%)
>>> Since you are moving it, this one should be moved to bindings/soc/nuvoton/
>>>
>>>> rename Documentation/devicetree/bindings/arm/{npcm/npcm.yaml => nuvoton/nuvoton,npcm.yaml} (93%)
>> OK, I will move "Documentation/devicetree/bindings/arm/npcm" to
>> "Documentation/devicetree/bindings/soc/nuvoton".
> To clarify - Rob asked *this one*. Not all.
>
>
>
> Best regards,
> Krzysztof
>

Dear Krzysztof,

I am sorry that still cannot get it exactly.

1. Just only submit  the nuvoton,ma35d1.yaml to
Documentation/devicetree/bindings/soc/nuvoton/nuvoton,ma35d1.yaml
    and keep Documentation/devicetree/bindings/arm/npcm unchanged

2. Move the directory "Documentation/devicetree/bindings/arm/npcm" to
"Documentation/devicetree/bindings/soc/nuvoton"
   and submit  the nuvoton,ma35d1.yaml to
"Documentation/devicetree/bindings/soc/nuvoton"

I should do 1 or 2?  Thank you very much.


Best regards,
Jacky Huang


2023-04-03 20:40:46

by Rob Herring

[permalink] [raw]
Subject: Re: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform

On Fri, Mar 31, 2023 at 10:15:16AM +0800, Jacky Huang wrote:
>
>
> On 2023/3/30 下午 09:25, Krzysztof Kozlowski wrote:
> > On 30/03/2023 12:41, Jacky Huang wrote:
> > > Dear Rob,
> > >
> > >
> > > On 2023/3/29 下午 09:07, Rob Herring wrote:
> > > > On Mon, Mar 27, 2023 at 9:19 PM Jacky Huang <[email protected]> wrote:
> > > > > From: Jacky Huang <[email protected]>
> > > > >
> > > > > Rename the bindings/arm/npcm directory as nuvoton.
> > > > > Add binding for ARMv8 based Nuvotn SoCs and platform boards.
> > > > > Add initial bindings for ma35d1 series development boards.
> > > > >
> > > > > Signed-off-by: Jacky Huang <[email protected]>
> > > > > ---
> > > > > .../bindings/arm/nuvoton/nuvoton,ma35d1.yaml | 30 +++++++++++++++++++
> > > > > .../nuvoton,npcm-gcr.yaml} | 2 +-
> > > > > .../npcm.yaml => nuvoton/nuvoton,npcm.yaml} | 2 +-
> > > > > 3 files changed, 32 insertions(+), 2 deletions(-)
> > > > > create mode 100644 Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml
> > > > > rename Documentation/devicetree/bindings/arm/{npcm/nuvoton,gcr.yaml => nuvoton/nuvoton,npcm-gcr.yaml} (93%)
> > > > Since you are moving it, this one should be moved to bindings/soc/nuvoton/
> > > >
> > > > > rename Documentation/devicetree/bindings/arm/{npcm/npcm.yaml => nuvoton/nuvoton,npcm.yaml} (93%)
> > > OK, I will move "Documentation/devicetree/bindings/arm/npcm" to
> > > "Documentation/devicetree/bindings/soc/nuvoton".
> > To clarify - Rob asked *this one*. Not all.

nuvoton,gcr.yaml is what should be moved to bindings/soc/... The rest
are the correct locations.

> >
> >
> >
> > Best regards,
> > Krzysztof
> >
>
> Dear Krzysztof,
>
> I am sorry that still cannot get it exactly.
>
> 1. Just only submit  the nuvoton,ma35d1.yaml to
> Documentation/devicetree/bindings/soc/nuvoton/nuvoton,ma35d1.yaml
>     and keep Documentation/devicetree/bindings/arm/npcm unchanged
>
> 2. Move the directory "Documentation/devicetree/bindings/arm/npcm" to
> "Documentation/devicetree/bindings/soc/nuvoton"
>    and submit  the nuvoton,ma35d1.yaml to
> "Documentation/devicetree/bindings/soc/nuvoton"
>
> I should do 1 or 2?  Thank you very much.

Neither.

Arm board and soc root node bindings go in bindings/arm/...

Miscellaneous system registers that don't fit into any other bindings/*
subsystem directory go in bindings/soc/

Rob

2023-04-06 02:19:34

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 06/12] dt-bindings: arm: Add initial bindings for Nuvoton platform



On 2023/4/4 上午 04:33, Rob Herring wrote:
> On Fri, Mar 31, 2023 at 10:15:16AM +0800, Jacky Huang wrote:
>>
>> On 2023/3/30 下午 09:25, Krzysztof Kozlowski wrote:
>>> On 30/03/2023 12:41, Jacky Huang wrote:
>>>> Dear Rob,
>>>>
>>>>
>>>> On 2023/3/29 下午 09:07, Rob Herring wrote:
>>>>> On Mon, Mar 27, 2023 at 9:19 PM Jacky Huang <[email protected]> wrote:
>>>>>> From: Jacky Huang <[email protected]>
>>>>>>
>>>>>> Rename the bindings/arm/npcm directory as nuvoton.
>>>>>> Add binding for ARMv8 based Nuvotn SoCs and platform boards.
>>>>>> Add initial bindings for ma35d1 series development boards.
>>>>>>
>>>>>> Signed-off-by: Jacky Huang <[email protected]>
>>>>>> ---
>>>>>> .../bindings/arm/nuvoton/nuvoton,ma35d1.yaml | 30 +++++++++++++++++++
>>>>>> .../nuvoton,npcm-gcr.yaml} | 2 +-
>>>>>> .../npcm.yaml => nuvoton/nuvoton,npcm.yaml} | 2 +-
>>>>>> 3 files changed, 32 insertions(+), 2 deletions(-)
>>>>>> create mode 100644 Documentation/devicetree/bindings/arm/nuvoton/nuvoton,ma35d1.yaml
>>>>>> rename Documentation/devicetree/bindings/arm/{npcm/nuvoton,gcr.yaml => nuvoton/nuvoton,npcm-gcr.yaml} (93%)
>>>>> Since you are moving it, this one should be moved to bindings/soc/nuvoton/
>>>>>
>>>>>> rename Documentation/devicetree/bindings/arm/{npcm/npcm.yaml => nuvoton/nuvoton,npcm.yaml} (93%)
>>>> OK, I will move "Documentation/devicetree/bindings/arm/npcm" to
>>>> "Documentation/devicetree/bindings/soc/nuvoton".
>>> To clarify - Rob asked *this one*. Not all.
> nuvoton,gcr.yaml is what should be moved to bindings/soc/... The rest
> are the correct locations.
>
>>>
>>>
>>> Best regards,
>>> Krzysztof
>>>
>> Dear Krzysztof,
>>
>> I am sorry that still cannot get it exactly.
>>
>> 1. Just only submit  the nuvoton,ma35d1.yaml to
>> Documentation/devicetree/bindings/soc/nuvoton/nuvoton,ma35d1.yaml
>>     and keep Documentation/devicetree/bindings/arm/npcm unchanged
>>
>> 2. Move the directory "Documentation/devicetree/bindings/arm/npcm" to
>> "Documentation/devicetree/bindings/soc/nuvoton"
>>    and submit  the nuvoton,ma35d1.yaml to
>> "Documentation/devicetree/bindings/soc/nuvoton"
>>
>> I should do 1 or 2?  Thank you very much.
> Neither.
>
> Arm board and soc root node bindings go in bindings/arm/...
>
> Miscellaneous system registers that don't fit into any other bindings/*
> subsystem directory go in bindings/soc/
>
> Rob

Dear Rob,

Thank you very much. I got it now.


Best regards,
Jacky Huang



2023-04-24 20:07:06

by Philipp Zabel

[permalink] [raw]
Subject: Re: [PATCH v6 10/12] reset: Add Nuvoton ma35d1 reset driver support

Hi Jacky,

On Tue, Mar 28, 2023 at 02:19:10AM +0000, Jacky Huang wrote:
> From: Jacky Huang <[email protected]>
>
> This driver supports individual IP reset for ma35d1. The reset
> control registers is a subset of system control registers.
>
> Signed-off-by: Jacky Huang <[email protected]>
> ---
> drivers/reset/Kconfig | 6 ++
> drivers/reset/Makefile | 1 +
> drivers/reset/reset-ma35d1.c | 152 +++++++++++++++++++++++++++++++++++
> 3 files changed, 159 insertions(+)
> create mode 100644 drivers/reset/reset-ma35d1.c
>
[...]
> diff --git a/drivers/reset/reset-ma35d1.c b/drivers/reset/reset-ma35d1.c
> new file mode 100644
> index 000000000000..221299e7b873
> --- /dev/null
> +++ b/drivers/reset/reset-ma35d1.c
> @@ -0,0 +1,152 @@
[...]
> +static int ma35d1_reset_update(struct reset_controller_dev *rcdev,
> + unsigned long id, bool assert)
> +{
> + unsigned int reg;
> + int ret;
> + int offset = (id / RST_PRE_REG) * 4;
> + struct ma35d1_reset_data *data =
> + container_of(rcdev, struct ma35d1_reset_data, rcdev);
> +
> + ret = regmap_read(data->regmap, REG_SYS_IPRST0 + offset, &reg);
> + if (ret < 0)
> + return ret;
> + if (assert)
> + reg |= 1 << (id % RST_PRE_REG);
> + else
> + reg &= ~(1 << (id % RST_PRE_REG));
> +
> + return regmap_write(data->regmap, REG_SYS_IPRST0 + offset, reg);

This should use regmap_update_bits().

[...]
> +static int ma35d1_reset_status(struct reset_controller_dev *rcdev,
> + unsigned long id)
> +{
> + int reg, ret;
> + int offset = id / RST_PRE_REG;

Should this be

int offset = (id / RST_PRE_REG) * 4;

?

> +static int ma35d1_reset_probe(struct platform_device *pdev)
> +{
> + int err;
> + struct device *dev = &pdev->dev;
> + struct device_node *parent;
> + struct ma35d1_reset_data *reset_data;
> + struct ma35d1_reboot_data *reboot_data;
> +
> + if (!pdev->dev.of_node) {
> + dev_err(&pdev->dev, "Device tree node not found\n");
> + return -EINVAL;
> + }
> +
> + reset_data = devm_kzalloc(dev, sizeof(*reset_data), GFP_KERNEL);
> + if (!reset_data)
> + return -ENOMEM;
> +
> + reboot_data = devm_kzalloc(dev, sizeof(*reboot_data), GFP_KERNEL);
> + if (!reboot_data)
> + return -ENOMEM;

These structures could be combined into one.

regards
Philipp

2023-04-25 01:28:19

by Jacky Huang

[permalink] [raw]
Subject: Re: [PATCH v6 10/12] reset: Add Nuvoton ma35d1 reset driver support

Dear Philipp,



On 2023/4/25 上午 04:02, Philipp Zabel wrote:
> Hi Jacky,
>
> On Tue, Mar 28, 2023 at 02:19:10AM +0000, Jacky Huang wrote:
>> From: Jacky Huang <[email protected]>
>>
>> This driver supports individual IP reset for ma35d1. The reset
>> control registers is a subset of system control registers.
>>
>> Signed-off-by: Jacky Huang <[email protected]>
>> ---
>> drivers/reset/Kconfig | 6 ++
>> drivers/reset/Makefile | 1 +
>> drivers/reset/reset-ma35d1.c | 152 +++++++++++++++++++++++++++++++++++
>> 3 files changed, 159 insertions(+)
>> create mode 100644 drivers/reset/reset-ma35d1.c
>>
> [...]
>> diff --git a/drivers/reset/reset-ma35d1.c b/drivers/reset/reset-ma35d1.c
>> new file mode 100644
>> index 000000000000..221299e7b873
>> --- /dev/null
>> +++ b/drivers/reset/reset-ma35d1.c
>> @@ -0,0 +1,152 @@
> [...]
>> +static int ma35d1_reset_update(struct reset_controller_dev *rcdev,
>> + unsigned long id, bool assert)
>> +{
>> + unsigned int reg;
>> + int ret;
>> + int offset = (id / RST_PRE_REG) * 4;
>> + struct ma35d1_reset_data *data =
>> + container_of(rcdev, struct ma35d1_reset_data, rcdev);
>> +
>> + ret = regmap_read(data->regmap, REG_SYS_IPRST0 + offset, &reg);
>> + if (ret < 0)
>> + return ret;
>> + if (assert)
>> + reg |= 1 << (id % RST_PRE_REG);
>> + else
>> + reg &= ~(1 << (id % RST_PRE_REG));
>> +
>> + return regmap_write(data->regmap, REG_SYS_IPRST0 + offset, reg);
> This should use regmap_update_bits().
>
> [...]
>> +static int ma35d1_reset_status(struct reset_controller_dev *rcdev,
>> + unsigned long id)
>> +{
>> + int reg, ret;
>> + int offset = id / RST_PRE_REG;
> Should this be
>
> int offset = (id / RST_PRE_REG) * 4;
>
> ?

Yes, here is a coding mistake.
As the register offset was modified to be indexed by lookup table in v7,
this
code was obsoleted.
Thank you for pointing out this.

>> +static int ma35d1_reset_probe(struct platform_device *pdev)
>> +{
>> + int err;
>> + struct device *dev = &pdev->dev;
>> + struct device_node *parent;
>> + struct ma35d1_reset_data *reset_data;
>> + struct ma35d1_reboot_data *reboot_data;
>> +
>> + if (!pdev->dev.of_node) {
>> + dev_err(&pdev->dev, "Device tree node not found\n");
>> + return -EINVAL;
>> + }
>> +
>> + reset_data = devm_kzalloc(dev, sizeof(*reset_data), GFP_KERNEL);
>> + if (!reset_data)
>> + return -ENOMEM;
>> +
>> + reboot_data = devm_kzalloc(dev, sizeof(*reboot_data), GFP_KERNEL);
>> + if (!reboot_data)
>> + return -ENOMEM;
> These structures could be combined into one.

OK, we will combine them into one.

> regards
> Philipp

Best regards,
Jacky Huang