Skip to contentSkip to conversation

Add timer driver for StarFive JH7110 RISC-V SoC

PATCH Superseded Xingyu Wu 3 participants 15 messages Oct 12, 2023 – Oct 16, 2023

Subsystems (4)

v6 of 8 Superseded Latest is v8 →

lore
Share & tools
lore

Xingyu Wu

author
#0 lore raw

v6 cover Add timer driver for StarFive JH7110 RISC-V SoC

This patch serises are to add timer driver for the StarFive JH7110 RISC-V SoC. The first patch adds documentation to describe device tree bindings. The subsequent patch adds timer driver and support JH7110 SoC. The last patch adds device node about timer in JH7110 dts.

This timer has four free-running 32 bit counters and runs in 24MHz clock on StarFive JH7110 SoC. And each channel(counter) triggers an interrupt when timeout. They support one-shot mode and continuous-run mode.

Changes since v5:
- Rebased on 6.6-rc5.
- Chnaged the number about characters of name.
- Made the clkevt->periodic to a local variable.
- Dropped the variables of device and base.
- Used clkevt->evt.irq directly and dropped the extra copy of irq.

V5: https://lore.kernel.org/all/[email protected]/

Changes since v4:
- Rebased on 6.5.
- Dropped the useless enum and used value directly when writing registers.
- Modified the description in Kconfig.
- Add the reviewed tag in patch 3.

v4: https://lore.kernel.org/all/[email protected]/

Changes since v3:
- Rebased on 6.5-rc6
- Dropped the useless enum names like 'JH7110_TIMER_CH_0'.
- Dropped the platform data about JH7110 and used the register offsets directly.
- Drroped the useless functions of clk_disable_unprepare().

v3: https://lore.kernel.org/all/[email protected]/

Changes since v2:
- Rebased on 6.4-rc7.
- Merged the header file into the c file.
- Renamed the functions from 'starfive_' to 'jh7110_'
- Used function 'clocksource_register_hz' instead of 'clocksource_mmio_init'.

v2: https://lore.kernel.org/all/[email protected]/

Changes since v1:
- Added description about timer and modified properties' description in dt-bindings.
- Dropped the 'interrupt-names' and 'clock-frequency' in dt-bindings.
- Renamed the functions and added 'starfive_'
- Modified that the driver probe by platform bus.

v1: https://lore.kernel.org/all/[email protected]/

Xingyu Wu (3):
dt-bindings: timer: Add timer for StarFive JH7110 SoC
clocksource: Add JH7110 timer driver
riscv: dts: jh7110: starfive: Add timer node

 .../bindings/timer/starfive,jh7110-timer.yaml |  96 +++++
 MAINTAINERS                                   |   7 +
 arch/riscv/boot/dts/starfive/jh7110.dtsi      |  20 +
 drivers/clocksource/Kconfig                   |  11 +
 drivers/clocksource/Makefile                  |   1 +
 drivers/clocksource/timer-jh7110.c            | 387 ++++++++++++++++++
 6 files changed, 522 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/timer/starfive,jh7110-timer.yaml
 create mode 100644 drivers/clocksource/timer-jh7110.c
Signature · 1 line
2.25.1
author
#1 lore raw

PATCH v6 1/3 dt-bindings: timer: Add timer for StarFive JH7110 SoC

Add bindings for the timer on the JH7110 RISC-V SoC by StarFive Technology Ltd.

---
 .../bindings/timer/starfive,jh7110-timer.yaml | 96 +++++++++++++++++++
 1 file changed, 96 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/timer/starfive,jh7110-timer.yaml
Documentation/devicetree/bindings/timer/starfive,jh7110-timer.yaml new file +96 -0
new file mode 100644
index 000000000000..9a2dac11eb06
@@ -0,0 +1,96 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/timer/starfive,jh7110-timer.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: StarFive JH7110 Timer
+
+maintainers:
+  - Xingyu Wu <[email protected]>
+  - Samin Guo <[email protected]>
+
+description:
+  This timer has four free-running 32 bit counters in StarFive JH7110 SoC.
+  And each channel(counter) triggers an interrupt when timeout. They support
+  one-shot mode and continuous-run mode.
+
+properties:
+  compatible:
+    const: starfive,jh7110-timer
+
+  reg:
+    maxItems: 1
+
+  interrupts:
+    items:
+      - description: channel 0
+      - description: channel 1
+      - description: channel 2
+      - description: channel 3
+
+  clocks:
+    items:
+      - description: timer APB
+      - description: channel 0
+      - description: channel 1
+      - description: channel 2
+      - description: channel 3
+
+  clock-names:
+    items:
+      - const: apb
+      - const: ch0
+      - const: ch1
+      - const: ch2
+      - const: ch3
+
+  resets:
+    items:
+      - description: timer APB
+      - description: channel 0
+      - description: channel 1
+      - description: channel 2
+      - description: channel 3
+
+  reset-names:
+    items:
+      - const: apb
+      - const: ch0
+      - const: ch1
+      - const: ch2
+      - const: ch3
+
+required:
+  - compatible
+  - reg
+  - interrupts
+  - clocks
+  - clock-names
+  - resets
+  - reset-names
+
+additionalProperties: false
+
+examples:
+  - |
+    timer@13050000 {
+        compatible = "starfive,jh7110-timer";
+        reg = <0x13050000 0x10000>;
+        interrupts = <69>, <70>, <71> ,<72>;
+        clocks = <&clk 124>,
+                 <&clk 125>,
+                 <&clk 126>,
+                 <&clk 127>,
+                 <&clk 128>;
+        clock-names = "apb", "ch0", "ch1",
+                      "ch2", "ch3";
+        resets = <&rst 117>,
+                 <&rst 118>,
+                 <&rst 119>,
+                 <&rst 120>,
+                 <&rst 121>;
+        reset-names = "apb", "ch0", "ch1",
+                      "ch2", "ch3";
+    };
+
Signature · 1 line
2.25.1
author
#2 lore raw

PATCH v6 2/3 clocksource: Add JH7110 timer driver

In reply to Xingyu Wu

Add timer driver for the StarFive JH7110 SoC.

---
 MAINTAINERS                        |   7 +
 drivers/clocksource/Kconfig        |  11 +
 drivers/clocksource/Makefile       |   1 +
 drivers/clocksource/timer-jh7110.c | 387 +++++++++++++++++++++++++++++
 4 files changed, 406 insertions(+)
 create mode 100644 drivers/clocksource/timer-jh7110.c
MAINTAINERS +7 -0
index 6c4cce45a09d..1525f031d4a2 100644
@@ -20472,6 +20472,13 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/sound/starfive,jh7110-tdm.yaml
 F:	sound/soc/starfive/jh7110_tdm.c
 
+STARFIVE JH7110 TIMER DRIVER
+M:	Samin Guo <[email protected]>
+M:	Xingyu Wu <[email protected]>
+S:	Supported
+F:	Documentation/devicetree/bindings/timer/starfive,jh7110-timer.yaml
+F:	drivers/clocksource/timer-jh7110.c
+
 STARFIVE JH71X0 CLOCK DRIVERS
 M:	Emil Renner Berthing <[email protected]>
 M:	Hal Feng <[email protected]>
drivers/clocksource/Kconfig +11 -0
index 0ba0dc4ecf06..821abcc1e517 100644
@@ -641,6 +641,17 @@ config RISCV_TIMER
 	  is accessed via both the SBI and the rdcycle instruction.  This is
 	  required for all RISC-V systems.
 
+config STARFIVE_JH7110_TIMER
+	bool "Timer for the STARFIVE JH7110 SoC"
+	depends on ARCH_STARFIVE || COMPILE_TEST
+	select TIMER_OF
+	select CLKSRC_MMIO
+	default ARCH_STARFIVE
+	help
+	  This enables the timer for StarFive JH7110 SoC. On RISC-V platform,
+	  the system has started RISCV_TIMER, but you can also use this timer
+	  which can provide four channels to do a lot more things on JH7110 SoC.
+
 config CLINT_TIMER
 	bool "CLINT Timer for the RISC-V platform" if COMPILE_TEST
 	depends on GENERIC_SCHED_CLOCK && RISCV
drivers/clocksource/Makefile +1 -0
index 368c3461dab8..b66ac05ec086 100644
@@ -80,6 +80,7 @@ obj-$(CONFIG_INGENIC_TIMER)		+= ingenic-timer.o
 obj-$(CONFIG_CLKSRC_ST_LPC)		+= clksrc_st_lpc.o
 obj-$(CONFIG_X86_NUMACHIP)		+= numachip.o
 obj-$(CONFIG_RISCV_TIMER)		+= timer-riscv.o
+obj-$(CONFIG_STARFIVE_JH7110_TIMER)	+= timer-jh7110.o
 obj-$(CONFIG_CLINT_TIMER)		+= timer-clint.o
 obj-$(CONFIG_CSKY_MP_TIMER)		+= timer-mp-csky.o
 obj-$(CONFIG_GX6605S_TIMER)		+= timer-gx6605s.o
drivers/clocksource/timer-jh7110.c new file +387 -0

387 lines not on this page · raw message

Signature · 1 line
2.25.1
#3 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

Le 12/10/2023 à 10:10, Xingyu Wu a écrit :

Quoted 3 lines

Add timer driver for the StarFive JH7110 SoC.

...

Quoted 17 lines
drivers/clocksource/timer-jh7110.c new file +17 -0
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+	struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+	char name[4];
+	struct clk *pclk;
+	struct reset_control *rst;
+	int ch;
+	int ret;
+	void __iomem *base;
+
+	base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(base))
+		return dev_err_probe(&pdev->dev, PTR_ERR(base),
+				     "failed to map registers\n");
+
+	rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+	if (IS_ERR(rst))

drivers/clocksource/timer-jh7110.c line 324

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+		return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+	pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+	if (IS_ERR(pclk))
+		return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+				     "failed to get & enable apb clock\n");
+
+	ret = reset_control_deassert(rst);
+	if (ret)
+		return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");

Hi,

I'm not very familiar with the reset_control_[de]assert() functions, but shouldn't this be undone by a reset_control_assert() call if an error occurs later?

Quoted 13 lines
drivers/clocksource/timer-jh7110.c new file +13 -0
+
+	for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+		clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+		if (!clkevt[ch])
+			return -ENOMEM;
+
+		snprintf(name, sizeof(name), "ch%d", ch);
+
+		clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+		/* Ensure timer is disabled */
+		jh7110_timer_disable(clkevt[ch]);
+
+		rst = devm_reset_control_get_exclusive(&pdev->dev, name);

drivers/clocksource/timer-jh7110.c line 347

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+		if (IS_ERR(rst))
+			return PTR_ERR(rst);
+
+		clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+		if (IS_ERR(clkevt[ch]->clk))
+			return PTR_ERR(clkevt[ch]->clk);
+
+		ret = reset_control_deassert(rst);
+		if (ret)
+			return ret;

Same here.

Quoted 5 lines
drivers/clocksource/timer-jh7110.c new file +5 -0
+
+		clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+		if (clkevt[ch]->evt.irq < 0)
+			return clkevt[ch]->evt.irq;
+

drivers/clocksource/timer-jh7110.c line 362

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+		snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+		jh7110_clockevents_register(clkevt[ch]);
+
+		ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+				       IRQF_TIMER | IRQF_IRQPOLL,
+				       clkevt[ch]->name, &clkevt[ch]->evt);
+		if (ret)
+			return ret;
+
+		ret = jh7110_clocksource_init(clkevt[ch]);

Does something should be done if this fails?

CJ

Quoted 25 lines
drivers/clocksource/timer-jh7110.c new file +25 -0
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+	{ .compatible = "starfive,jh7110-timer", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+	.probe = jh7110_timer_probe,
+	.driver = {
+		.name = "jh7110-timer",
+		.of_match_table = jh7110_timer_match,
+	},
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");
author
#4 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

Quoted 6 lines · Christophe JAILLET

Le 12/10/2023 à 10:10, Xingyu Wu a écrit :

Quoted 3 lines

Add timer driver for the StarFive JH7110 SoC.

...

It looks normal in my email and the web. Is this due to some settings?

Quoted 31 lines
Quoted 18 lines
drivers/clocksource/timer-jh7110.c new file +17 -0
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+    struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+    char name[4];
+    struct clk *pclk;
+    struct reset_control *rst;
+    int ch;
+    int ret;
+    void __iomem *base;
+
+    base = devm_platform_ioremap_resource(pdev, 0);
+    if (IS_ERR(base))
+        return dev_err_probe(&pdev->dev, PTR_ERR(base),
+                     "failed to map registers\n");
+
+    rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+    if (IS_ERR(rst))

drivers/clocksource/timer-jh7110.c line 324

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+    pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+    if (IS_ERR(pclk))
+        return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+                     "failed to get & enable apb clock\n");
+
+    ret = reset_control_deassert(rst);
+    if (ret)
+        return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");

Hi,

Quoted 1 line

I'm not very familiar with the reset_control_[de]assert() functions, but shouldn't this be undone by a reset_control_assert() call if an error occurs later?

In this case, the reset controller is set from 'assert' state to 'deassert' state. If it is failed and still 'assert' state, I don't think it need to call reset_control_assert().

Quoted 45 lines
Quoted 14 lines
drivers/clocksource/timer-jh7110.c new file +13 -0
+
+    for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+        clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+        if (!clkevt[ch])
+            return -ENOMEM;
+
+        snprintf(name, sizeof(name), "ch%d", ch);
+
+        clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+        /* Ensure timer is disabled */
+        jh7110_timer_disable(clkevt[ch]);
+
+        rst = devm_reset_control_get_exclusive(&pdev->dev, name);

drivers/clocksource/timer-jh7110.c line 347

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        if (IS_ERR(rst))
+            return PTR_ERR(rst);
+
+        clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+        if (IS_ERR(clkevt[ch]->clk))
+            return PTR_ERR(clkevt[ch]->clk);
+
+        ret = reset_control_deassert(rst);
+        if (ret)
+            return ret;

Same here.

Quoted 5 lines
drivers/clocksource/timer-jh7110.c new file +5 -0
+
+        clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+        if (clkevt[ch]->evt.irq < 0)
+            return clkevt[ch]->evt.irq;
+

drivers/clocksource/timer-jh7110.c line 362

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+        jh7110_clockevents_register(clkevt[ch]);
+
+        ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+                       IRQF_TIMER | IRQF_IRQPOLL,
+                       clkevt[ch]->name, &clkevt[ch]->evt);
+        if (ret)
+            return ret;
+
+        ret = jh7110_clocksource_init(clkevt[ch]);

Does something should be done if this fails?

Quoted 1 line

CJ

Yes, it should be call reset_control_assert() here and I will add it in next version.

Quoted 18 lines
Quoted 18 lines

+        if (ret)
+            return ret;
+    }

drivers/clocksource/timer-jh7110.c new file +14 -0
+
+    return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+    { .compatible = "starfive,jh7110-timer", },
+    { /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+    .probe = jh7110_timer_probe,
+    .driver = {
+        .name = "jh7110-timer",
Quoted 9 lines

drivers/clocksource/timer-jh7110.c line 387

Quoted 8 lines

+        .of_match_table = jh7110_timer_match,
+    },

drivers/clocksource/timer-jh7110.c new file +6 -0
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");

Thanks,
Xingyu Wu

#5 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

Quoted 41 lines · Xingyu Wu
Quoted 6 lines · Christophe JAILLET

Le 12/10/2023 à 10:10, Xingyu Wu a écrit :

Quoted 3 lines

Add timer driver for the StarFive JH7110 SoC.

...

It looks normal in my email and the web. Is this due to some settings?

Quoted 31 lines
Quoted 18 lines
drivers/clocksource/timer-jh7110.c new file +17 -0
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+    struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+    char name[4];
+    struct clk *pclk;
+    struct reset_control *rst;
+    int ch;
+    int ret;
+    void __iomem *base;
+
+    base = devm_platform_ioremap_resource(pdev, 0);
+    if (IS_ERR(base))
+        return dev_err_probe(&pdev->dev, PTR_ERR(base),
+                     "failed to map registers\n");
+
+    rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+    if (IS_ERR(rst))

drivers/clocksource/timer-jh7110.c line 324

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+    pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+    if (IS_ERR(pclk))
+        return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+                     "failed to get & enable apb clock\n");
+
+    ret = reset_control_deassert(rst);
+    if (ret)
+        return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");

Hi,

Quoted 3 lines
Quoted 1 line

I'm not very familiar with the reset_control_[de]assert() functions, but shouldn't this be undone by a reset_control_assert() call if an error occurs later?

In this case, the reset controller is set from 'assert' state to 'deassert' state. If it is failed and still 'assert' state, I don't think it need to call reset_control_assert().

The problem is if the assert succeeds, but the probe function fails later. Then there is nothing to undo the deassert. A common pattern I see is something like this:

static void jh7110_timer_reset_control_assert(void *data)

{
	reset_control_assert(data);
}

...

	ret = reset_control_deassert(rst);
	if (ret)
		return dev_err_probe(...);

	ret = devm_add_action_or_reset(&pdev->dev,
jh7110_timer_reset_control_assert, rst);
	if (ret)
		return ret;

/Emil

Quoted 85 lines
Quoted 45 lines
Quoted 14 lines
drivers/clocksource/timer-jh7110.c new file +13 -0
+
+    for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+        clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+        if (!clkevt[ch])
+            return -ENOMEM;
+
+        snprintf(name, sizeof(name), "ch%d", ch);
+
+        clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+        /* Ensure timer is disabled */
+        jh7110_timer_disable(clkevt[ch]);
+
+        rst = devm_reset_control_get_exclusive(&pdev->dev, name);

drivers/clocksource/timer-jh7110.c line 347

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        if (IS_ERR(rst))
+            return PTR_ERR(rst);
+
+        clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+        if (IS_ERR(clkevt[ch]->clk))
+            return PTR_ERR(clkevt[ch]->clk);
+
+        ret = reset_control_deassert(rst);
+        if (ret)
+            return ret;

Same here.

Quoted 5 lines
drivers/clocksource/timer-jh7110.c new file +5 -0
+
+        clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+        if (clkevt[ch]->evt.irq < 0)
+            return clkevt[ch]->evt.irq;
+

drivers/clocksource/timer-jh7110.c line 362

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+        jh7110_clockevents_register(clkevt[ch]);
+
+        ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+                       IRQF_TIMER | IRQF_IRQPOLL,
+                       clkevt[ch]->name, &clkevt[ch]->evt);
+        if (ret)
+            return ret;
+
+        ret = jh7110_clocksource_init(clkevt[ch]);

Does something should be done if this fails?

Quoted 1 line

CJ

Yes, it should be call reset_control_assert() here and I will add it in next version.

Quoted 18 lines
Quoted 18 lines

+        if (ret)
+            return ret;
+    }

drivers/clocksource/timer-jh7110.c new file +14 -0
+
+    return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+    { .compatible = "starfive,jh7110-timer", },
+    { /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+    .probe = jh7110_timer_probe,
+    .driver = {
+        .name = "jh7110-timer",
Quoted 9 lines

drivers/clocksource/timer-jh7110.c line 387

Quoted 8 lines

+        .of_match_table = jh7110_timer_match,
+    },

drivers/clocksource/timer-jh7110.c new file +6 -0
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");

Thanks,
Xingyu Wu

_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv

author
#6 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

Quoted 66 lines · Emil Renner Berthing
Quoted 41 lines · Xingyu Wu
Quoted 6 lines · Christophe JAILLET

Le 12/10/2023 à 10:10, Xingyu Wu a écrit :

Quoted 3 lines

Add timer driver for the StarFive JH7110 SoC.

...

It looks normal in my email and the web. Is this due to some settings?

Quoted 31 lines
Quoted 18 lines
drivers/clocksource/timer-jh7110.c new file +17 -0
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+    struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+    char name[4];
+    struct clk *pclk;
+    struct reset_control *rst;
+    int ch;
+    int ret;
+    void __iomem *base;
+
+    base = devm_platform_ioremap_resource(pdev, 0);
+    if (IS_ERR(base))
+        return dev_err_probe(&pdev->dev, PTR_ERR(base),
+                     "failed to map registers\n");
+
+    rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+    if (IS_ERR(rst))

drivers/clocksource/timer-jh7110.c line 324

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+    pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+    if (IS_ERR(pclk))
+        return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+                     "failed to get & enable apb clock\n");
+
+    ret = reset_control_deassert(rst);
+    if (ret)
+        return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");

Hi,

Quoted 3 lines
Quoted 1 line

I'm not very familiar with the reset_control_[de]assert() functions, but shouldn't this be undone by a reset_control_assert() call if an error occurs later?

In this case, the reset controller is set from 'assert' state to 'deassert' state. If it is failed and still 'assert' state, I don't think it need to call reset_control_assert().

The problem is if the assert succeeds, but the probe function fails later. Then there is nothing to undo the deassert. A common pattern I see is something like this:

static void jh7110_timer_reset_control_assert(void *data)

{
	reset_control_assert(data);
}

...

	ret = reset_control_deassert(rst);
	if (ret)
		return dev_err_probe(...);

	ret = devm_add_action_or_reset(&pdev->dev,
jh7110_timer_reset_control_assert, rst);
	if (ret)
		return ret;
Quoted 1 line

/Emil

It is a good idea! Let me try it.

Thanks,
Xingyu Wu

Quoted 86 lines
Quoted 86 lines
Quoted 46 lines
Quoted 15 lines
drivers/clocksource/timer-jh7110.c new file +13 -0
+
+    for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+        clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+        if (!clkevt[ch])
+            return -ENOMEM;
+
+        snprintf(name, sizeof(name), "ch%d", ch);
+
+        clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+        /* Ensure timer is disabled */
+        jh7110_timer_disable(clkevt[ch]);
+
+        rst = devm_reset_control_get_exclusive(&pdev->dev, name);

drivers/clocksource/timer-jh7110.c line 347

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        if (IS_ERR(rst))
+            return PTR_ERR(rst);
+
+        clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+        if (IS_ERR(clkevt[ch]->clk))
+            return PTR_ERR(clkevt[ch]->clk);
+
+        ret = reset_control_deassert(rst);
+        if (ret)
+            return ret;

Same here.

Quoted 5 lines
drivers/clocksource/timer-jh7110.c new file +5 -0
+
+        clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+        if (clkevt[ch]->evt.irq < 0)
+            return clkevt[ch]->evt.irq;
+

drivers/clocksource/timer-jh7110.c line 362

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+        jh7110_clockevents_register(clkevt[ch]);
+
+        ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+                       IRQF_TIMER | IRQF_IRQPOLL,
+                       clkevt[ch]->name, &clkevt[ch]->evt);
+        if (ret)
+            return ret;
+
+        ret = jh7110_clocksource_init(clkevt[ch]);

Does something should be done if this fails?

Quoted 1 line

CJ

Yes, it should be call reset_control_assert() here and I will add it in next version.

Quoted 18 lines
Quoted 18 lines

+        if (ret)
+            return ret;
+    }

drivers/clocksource/timer-jh7110.c new file +14 -0
+
+    return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+    { .compatible = "starfive,jh7110-timer", },
+    { /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+    .probe = jh7110_timer_probe,
+    .driver = {
+        .name = "jh7110-timer",
Quoted 9 lines

drivers/clocksource/timer-jh7110.c line 387

Quoted 8 lines

+        .of_match_table = jh7110_timer_match,
+    },

drivers/clocksource/timer-jh7110.c new file +6 -0
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");

Thanks,
Xingyu Wu

_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv

#7 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

In reply to Xingyu Wu

Le 13/10/2023 à 11:34, Xingyu Wu a écrit :

Quoted 9 lines · Christophe JAILLET
Quoted 6 lines

Le 12/10/2023 à 10:10, Xingyu Wu a écrit :

Quoted 3 lines

Add timer driver for the StarFive JH7110 SoC.

...

It looks normal in my email and the web. Is this due to some settings?

Hi,

I use gmane.org and a news reader (Thunderbird). Gmane sometimes (not always!) obfuscate e-mail addresses.

Do not pay attantion to these strange rewritten addresses.

Quoted 32 lines
Quoted 32 lines
Quoted 19 lines
drivers/clocksource/timer-jh7110.c new file +17 -0
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+    struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+    char name[4];
+    struct clk *pclk;
+    struct reset_control *rst;
+    int ch;
+    int ret;
+    void __iomem *base;
+
+    base = devm_platform_ioremap_resource(pdev, 0);
+    if (IS_ERR(base))
+        return dev_err_probe(&pdev->dev, PTR_ERR(base),
+                     "failed to map registers\n");
+
+    rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+    if (IS_ERR(rst))

drivers/clocksource/timer-jh7110.c line 324

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+    pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+    if (IS_ERR(pclk))
+        return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+                     "failed to get & enable apb clock\n");
+
+    ret = reset_control_deassert(rst);
+    if (ret)
+        return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");

Hi,

Quoted 3 lines
Quoted 1 line

I'm not very familiar with the reset_control_[de]assert() functions, but shouldn't this be undone by a reset_control_assert() call if an error occurs later?

In this case, the reset controller is set from 'assert' state to 'deassert' state. If it is failed and still 'assert' state, I don't think it need to call reset_control_assert().

Emil already explained what I meaned (sorry for not being clear enough). I do agree with his proposed approach.

Quoted 46 lines
Quoted 46 lines
Quoted 15 lines
drivers/clocksource/timer-jh7110.c new file +13 -0
+
+    for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+        clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+        if (!clkevt[ch])
+            return -ENOMEM;
+
+        snprintf(name, sizeof(name), "ch%d", ch);
+
+        clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+        /* Ensure timer is disabled */
+        jh7110_timer_disable(clkevt[ch]);
+
+        rst = devm_reset_control_get_exclusive(&pdev->dev, name);

drivers/clocksource/timer-jh7110.c line 347

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        if (IS_ERR(rst))
+            return PTR_ERR(rst);
+
+        clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+        if (IS_ERR(clkevt[ch]->clk))
+            return PTR_ERR(clkevt[ch]->clk);
+
+        ret = reset_control_deassert(rst);
+        if (ret)
+            return ret;

Same here.

Quoted 5 lines
drivers/clocksource/timer-jh7110.c new file +5 -0
+
+        clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+        if (clkevt[ch]->evt.irq < 0)
+            return clkevt[ch]->evt.irq;
+

drivers/clocksource/timer-jh7110.c line 362

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+        jh7110_clockevents_register(clkevt[ch]);
+
+        ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+                       IRQF_TIMER | IRQF_IRQPOLL,
+                       clkevt[ch]->name, &clkevt[ch]->evt);
+        if (ret)
+            return ret;
+
+        ret = jh7110_clocksource_init(clkevt[ch]);

Does something should be done if this fails?

Quoted 3 lines
Quoted 1 line

CJ

Yes, it should be call reset_control_assert() here and I will add it in next version.

My point was for the above reset_control_assert() but also for the resources allocated within this for loop.

I have not checked all paths, but in case of error in the probe:
- There is another reset_control_deassert()

- jh7110_clocksource_init() --> jh7110_timer_int_init_enable() -->
jh7110_timer_enable()
Should jh7110_timer_disable() be called?

- jh7110_clocksource_init() --> clocksource_register_hz(). Should clocksource_unregister() be called?

If I'm correct and depending on how you update the code, a .remove function may be needed as well.

CJ

Quoted 34 lines
Quoted 19 lines
Quoted 19 lines

+        if (ret)
+            return ret;
+    }

drivers/clocksource/timer-jh7110.c new file +14 -0
+
+    return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+    { .compatible = "starfive,jh7110-timer", },
+    { /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+    .probe = jh7110_timer_probe,
+    .driver = {
+        .name = "jh7110-timer",
Quoted 9 lines

drivers/clocksource/timer-jh7110.c line 387

Quoted 8 lines

+        .of_match_table = jh7110_timer_match,
+    },

drivers/clocksource/timer-jh7110.c new file +6 -0
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <xingyu.wu-bONrM45KWFOXmMXjJBpWqg-XMD5yJDbdMRS5n6/[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");

Thanks,
Xingyu Wu

author
#8 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

Quoted 115 lines · Christophe JAILLET

Le 13/10/2023 à 11:34, Xingyu Wu a écrit :

Quoted 9 lines · Christophe JAILLET
Quoted 6 lines

Le 12/10/2023 à 10:10, Xingyu Wu a écrit :

Quoted 3 lines

Add timer driver for the StarFive JH7110 SoC.

...

It looks normal in my email and the web. Is this due to some settings?

Hi,

I use gmane.org and a news reader (Thunderbird). Gmane sometimes (not always!) obfuscate e-mail addresses.

Do not pay attantion to these strange rewritten addresses.

Quoted 32 lines
Quoted 32 lines
Quoted 19 lines
drivers/clocksource/timer-jh7110.c new file +17 -0
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+    struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+    char name[4];
+    struct clk *pclk;
+    struct reset_control *rst;
+    int ch;
+    int ret;
+    void __iomem *base;
+
+    base = devm_platform_ioremap_resource(pdev, 0);
+    if (IS_ERR(base))
+        return dev_err_probe(&pdev->dev, PTR_ERR(base),
+                     "failed to map registers\n");
+
+    rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+    if (IS_ERR(rst))

drivers/clocksource/timer-jh7110.c line 324

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+    pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+    if (IS_ERR(pclk))
+        return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+                     "failed to get & enable apb clock\n");
+
+    ret = reset_control_deassert(rst);
+    if (ret)
+        return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");

Hi,

Quoted 3 lines
Quoted 1 line

I'm not very familiar with the reset_control_[de]assert() functions, but shouldn't this be undone by a reset_control_assert() call if an error occurs later?

In this case, the reset controller is set from 'assert' state to 'deassert' state. If it is failed and still 'assert' state, I don't think it need to call reset_control_assert().

Emil already explained what I meaned (sorry for not being clear enough). I do agree with his proposed approach.

Quoted 46 lines
Quoted 46 lines
Quoted 15 lines
drivers/clocksource/timer-jh7110.c new file +13 -0
+
+    for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+        clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+        if (!clkevt[ch])
+            return -ENOMEM;
+
+        snprintf(name, sizeof(name), "ch%d", ch);
+
+        clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+        /* Ensure timer is disabled */
+        jh7110_timer_disable(clkevt[ch]);
+
+        rst = devm_reset_control_get_exclusive(&pdev->dev, name);

drivers/clocksource/timer-jh7110.c line 347

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        if (IS_ERR(rst))
+            return PTR_ERR(rst);
+
+        clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+        if (IS_ERR(clkevt[ch]->clk))
+            return PTR_ERR(clkevt[ch]->clk);
+
+        ret = reset_control_deassert(rst);
+        if (ret)
+            return ret;

Same here.

Quoted 5 lines
drivers/clocksource/timer-jh7110.c new file +5 -0
+
+        clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+        if (clkevt[ch]->evt.irq < 0)
+            return clkevt[ch]->evt.irq;
+

drivers/clocksource/timer-jh7110.c line 362

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+        snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+        jh7110_clockevents_register(clkevt[ch]);
+
+        ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+                       IRQF_TIMER | IRQF_IRQPOLL,
+                       clkevt[ch]->name, &clkevt[ch]->evt);
+        if (ret)
+            return ret;
+
+        ret = jh7110_clocksource_init(clkevt[ch]);

Does something should be done if this fails?

Quoted 3 lines
Quoted 1 line

CJ

Yes, it should be call reset_control_assert() here and I will add it in next version.

My point was for the above reset_control_assert() but also for the resources allocated within this for loop.

I have not checked all paths, but in case of error in the probe:
  - There is another reset_control_deassert()

  - jh7110_clocksource_init() --> jh7110_timer_int_init_enable() --> jh7110_timer_enable() Should jh7110_timer_disable() be called?

Quoted 7 lines

  - jh7110_clocksource_init() --> clocksource_register_hz(). Should clocksource_unregister() be called?

If I'm correct and depending on how you update the code, a .remove function may be needed as well.

CJ

Great. I'm going to add a action which Emil had said to do this job when remove or make the error in the loop.

Thanks,
Xingyu Wu

Quoted 35 lines
Quoted 34 lines
Quoted 19 lines
Quoted 19 lines

+        if (ret)
+            return ret;
+    }

drivers/clocksource/timer-jh7110.c new file +14 -0
+
+    return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+    { .compatible = "starfive,jh7110-timer", },
+    { /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+    .probe = jh7110_timer_probe,
+    .driver = {
+        .name = "jh7110-timer",
Quoted 9 lines

drivers/clocksource/timer-jh7110.c line 387

Quoted 8 lines

+        .of_match_table = jh7110_timer_match,
+    },

drivers/clocksource/timer-jh7110.c new file +6 -0
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <xingyu.wu-bONrM45KWFOXmMXjJBpWqg-XMD5yJDbdMRS5n6/[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");

Thanks,
Xingyu Wu

#9 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

In reply to Xingyu Wu
Quoted 115 lines · Xingyu Wu

Add timer driver for the StarFive JH7110 SoC.

---
 MAINTAINERS                        |   7 +
 drivers/clocksource/Kconfig        |  11 +
 drivers/clocksource/Makefile       |   1 +
 drivers/clocksource/timer-jh7110.c | 387 +++++++++++++++++++++++++++++
 4 files changed, 406 insertions(+)
 create mode 100644 drivers/clocksource/timer-jh7110.c
MAINTAINERS +7 -0
index 6c4cce45a09d..1525f031d4a2 100644
@@ -20472,6 +20472,13 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/sound/starfive,jh7110-tdm.yaml
 F:	sound/soc/starfive/jh7110_tdm.c
 
+STARFIVE JH7110 TIMER DRIVER
+M:	Samin Guo <[email protected]>
+M:	Xingyu Wu <[email protected]>
+S:	Supported
+F:	Documentation/devicetree/bindings/timer/starfive,jh7110-timer.yaml
+F:	drivers/clocksource/timer-jh7110.c
+
 STARFIVE JH71X0 CLOCK DRIVERS
 M:	Emil Renner Berthing <[email protected]>
 M:	Hal Feng <[email protected]>
drivers/clocksource/Kconfig +11 -0
index 0ba0dc4ecf06..821abcc1e517 100644
@@ -641,6 +641,17 @@ config RISCV_TIMER
 	  is accessed via both the SBI and the rdcycle instruction.  This is
 	  required for all RISC-V systems.
 
+config STARFIVE_JH7110_TIMER
+	bool "Timer for the STARFIVE JH7110 SoC"
+	depends on ARCH_STARFIVE || COMPILE_TEST
+	select TIMER_OF
+	select CLKSRC_MMIO
+	default ARCH_STARFIVE
+	help
+	  This enables the timer for StarFive JH7110 SoC. On RISC-V platform,
+	  the system has started RISCV_TIMER, but you can also use this timer
+	  which can provide four channels to do a lot more things on JH7110 SoC.
+
 config CLINT_TIMER
 	bool "CLINT Timer for the RISC-V platform" if COMPILE_TEST
 	depends on GENERIC_SCHED_CLOCK && RISCV
drivers/clocksource/Makefile +1 -0
index 368c3461dab8..b66ac05ec086 100644
@@ -80,6 +80,7 @@ obj-$(CONFIG_INGENIC_TIMER)		+= ingenic-timer.o
 obj-$(CONFIG_CLKSRC_ST_LPC)		+= clksrc_st_lpc.o
 obj-$(CONFIG_X86_NUMACHIP)		+= numachip.o
 obj-$(CONFIG_RISCV_TIMER)		+= timer-riscv.o
+obj-$(CONFIG_STARFIVE_JH7110_TIMER)	+= timer-jh7110.o
 obj-$(CONFIG_CLINT_TIMER)		+= timer-clint.o
 obj-$(CONFIG_CSKY_MP_TIMER)		+= timer-mp-csky.o
 obj-$(CONFIG_GX6605S_TIMER)		+= timer-gx6605s.o
drivers/clocksource/timer-jh7110.c new file +46 -0
new file mode 100644
index 000000000000..914424368290
@@ -0,0 +1,387 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Starfive JH7110 Timer driver
+ *
+ * Copyright (C) 2022-2023 StarFive Technology Co., Ltd.
+ *
+ * Author:
+ * Xingyu Wu <[email protected]>
+ * Samin Guo <[email protected]>
+ */
+
+#include <linux/clk.h>
+#include <linux/clockchips.h>
+#include <linux/clocksource.h>
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/irq.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+#include <linux/sched_clock.h>
+
+/* Bias: Ch0-0x0, Ch1-0x40, Ch2-0x80, and so on. */
+#define JH7110_TIMER_CH_LEN		0x40
+#define JH7110_TIMER_CH_BASE(x)		((x) * JH7110_TIMER_CH_LEN)
+#define JH7110_TIMER_CH_MAX		4
+
+#define JH7110_CLOCK_SOURCE_RATING	200
+#define JH7110_VALID_BITS		32
+#define JH7110_DELAY_US			0
+#define JH7110_TIMEOUT_US		10000
+#define JH7110_CLOCKEVENT_RATING	300
+#define JH7110_TIMER_MAX_TICKS		0xffffffff
+#define JH7110_TIMER_MIN_TICKS		0xf
+#define JH7110_TIMER_NAME_NUM		19
+
+#define JH7110_TIMER_INT_STATUS		0x00 /* RO[0:4]: Interrupt Status for channel0~4 */
+#define JH7110_TIMER_CTL		0x04 /* RW[0]: 0-continuous run, 1-single run */
+#define JH7110_TIMER_LOAD		0x08 /* RW: load value to counter */
+#define JH7110_TIMER_ENABLE		0x10 /* RW[0]: timer enable register */
+#define JH7110_TIMER_RELOAD		0x14 /* RW: write 1 or 0 both reload counter */

drivers/clocksource/timer-jh7110.c line 56

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+#define JH7110_TIMER_VALUE		0x18 /* RO: timer value register */
+#define JH7110_TIMER_INT_CLR		0x20 /* RW: timer interrupt clear register */
+#define JH7110_TIMER_INT_MASK		0x24 /* RW[0]: timer interrupt mask register */
+#define JH7110_TIMER_INT_CLR_AVA_MASK	BIT(1)
+
+struct jh7110_clkevt {
+	struct clock_event_device evt;
+	struct clocksource cs;
+	struct clk *clk;
+	char name[JH7110_TIMER_NAME_NUM];

Instead of this define maybe just use something like

	char name[sizeof("jh7110-timer.chX")];

drivers/clocksource/timer-jh7110.c line 57

Quoted 1 line
drivers/clocksource/timer-jh7110.c new file +1 -0
+	int irq;

You wrote that you "dropped the extra copy of irq", but here you didn't.

Quoted 21 lines
drivers/clocksource/timer-jh7110.c new file +21 -0
+	u32 rate;
+	u32 reload_val;
+	void __iomem *base;
+};
+
+static inline struct jh7110_clkevt *to_jh7110_clkevt(struct clock_event_device *evt)
+{
+	return container_of(evt, struct jh7110_clkevt, evt);
+}
+
+/* 0:continuous-run mode, 1:single-run mode */
+static inline void jh7110_timer_set_continuous_mod(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_CTL);
+}
+
+static inline void jh7110_timer_set_single_mod(struct jh7110_clkevt *clkevt)
+{
+	writel(1, clkevt->base + JH7110_TIMER_CTL);
+}
+

drivers/clocksource/timer-jh7110.c line 88

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+/* Interrupt Mask Register, 0:Unmask, 1:Mask */
+static inline void jh7110_timer_int_enable(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_INT_MASK);
+}
+
+static inline void jh7110_timer_int_disable(struct jh7110_clkevt *clkevt)
+{
+	writel(1, clkevt->base + JH7110_TIMER_INT_MASK);
+}

I really don't think all these wrappers make the code any clearer. Please just inline the writel() calls. What you can do to make it a bit clearer is something like

enum {
	JH7110_TIMER_MODE_CONTINUOUS,
	JH7110_TIMER_MODE_SINGLE,
};

so you can write

	writel(JH7110_TIMER_MODE_CONTINUOUS, clkevt->base + JH7110_TIMER_CTL);
Quoted 44 lines
drivers/clocksource/timer-jh7110.c new file +44 -0
+
+/*
+ * BIT(0): Read value represent channel intr status.
+ * Write 1 to this bit to clear interrupt. Write 0 has no effects.
+ * BIT(1): "1" means that it is clearing interrupt. BIT(0) can not be written.
+ */
+static inline int jh7110_timer_int_clear(struct jh7110_clkevt *clkevt)
+{
+	u32 value;
+	int ret;
+
+	/* waiting interrupt can be to clearing */
+	ret = readl_poll_timeout_atomic(clkevt->base + JH7110_TIMER_INT_CLR, value,
+					!(value & JH7110_TIMER_INT_CLR_AVA_MASK),
+					JH7110_DELAY_US, JH7110_TIMEOUT_US);
+	if (!ret)
+		writel(0x1, clkevt->base + JH7110_TIMER_INT_CLR);
+
+	return ret;
+}
+
+/*
+ * The initial value to be loaded into the
+ * counter and is also used as the reload value.
+ * val = clock rate --> 1s
+ */
+static inline void jh7110_timer_set_load(struct jh7110_clkevt *clkevt, u32 val)
+{
+	writel(val, clkevt->base + JH7110_TIMER_LOAD);
+}
+
+static inline u32 jh7110_timer_get_val(struct jh7110_clkevt *clkevt)
+{
+	return readl(clkevt->base + JH7110_TIMER_VALUE);
+}
+
+/*
+ * Write RELOAD register to reload preset value to counter.
+ * Write 0 and write 1 are both ok.
+ */
+static inline void jh7110_timer_set_reload(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_RELOAD);
+}

drivers/clocksource/timer-jh7110.c line 142

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+
+static inline void jh7110_timer_enable(struct jh7110_clkevt *clkevt)
+{
+	writel(1, clkevt->base + JH7110_TIMER_ENABLE);
+}
+
+static inline void jh7110_timer_disable(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_ENABLE);
+}

The same thing goes for all these wrappers. If you do insist on having these wrappers then at least group them all together so you can quickly scroll past them to get to the "real" code.

Quoted 253 lines
drivers/clocksource/timer-jh7110.c new file +245 -1
+
+static int jh7110_timer_int_init_enable(struct jh7110_clkevt *clkevt)
+{
+	int ret;
+
+	jh7110_timer_int_disable(clkevt);
+	ret = jh7110_timer_int_clear(clkevt);
+	if (ret)
+		return ret;
+
+	jh7110_timer_int_enable(clkevt);
+	jh7110_timer_enable(clkevt);
+
+	return 0;
+}
+
+static int jh7110_timer_shutdown(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_disable(clkevt);
+	return jh7110_timer_int_clear(clkevt);
+}
+
+static void jh7110_timer_suspend(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	clkevt->reload_val = jh7110_timer_get_val(clkevt);
+	jh7110_timer_shutdown(evt);
+}
+
+static void jh7110_timer_resume(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_set_load(clkevt, clkevt->reload_val);
+	jh7110_timer_set_reload(clkevt);
+	jh7110_timer_int_enable(clkevt);
+	jh7110_timer_enable(clkevt);
+}
+
+static int jh7110_timer_tick_resume(struct clock_event_device *evt)
+{
+	jh7110_timer_resume(evt);
+
+	return 0;
+}
+
+/* IRQ handler for the timer */
+static irqreturn_t jh7110_timer_interrupt(int irq, void *priv)
+{
+	struct clock_event_device *evt = (struct clock_event_device *)priv;
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	if (jh7110_timer_int_clear(clkevt))
+		return IRQ_NONE;
+
+	if (evt->event_handler)
+		evt->event_handler(evt);
+
+	return IRQ_HANDLED;
+}
+
+static int jh7110_timer_set_periodic(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+	u32 periodic = DIV_ROUND_CLOSEST(clkevt->rate, HZ);
+
+	jh7110_timer_disable(clkevt);
+	jh7110_timer_set_continuous_mod(clkevt);
+	jh7110_timer_set_load(clkevt, periodic);
+
+	return jh7110_timer_int_init_enable(clkevt);
+}
+
+static int jh7110_timer_set_oneshot(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_disable(clkevt);
+	jh7110_timer_set_single_mod(clkevt);
+	jh7110_timer_set_load(clkevt, JH7110_TIMER_MAX_TICKS);
+
+	return jh7110_timer_int_init_enable(clkevt);
+}
+
+static int jh7110_timer_set_next_event(unsigned long next,
+				       struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_disable(clkevt);
+	jh7110_timer_set_single_mod(clkevt);
+	jh7110_timer_set_load(clkevt, next);
+	jh7110_timer_enable(clkevt);
+
+	return 0;
+}
+
+static void jh7110_set_clockevent(struct clock_event_device *evt)
+{
+	evt->features = CLOCK_EVT_FEAT_PERIODIC |
+			CLOCK_EVT_FEAT_ONESHOT |
+			CLOCK_EVT_FEAT_DYNIRQ;
+	evt->set_state_shutdown = jh7110_timer_shutdown;
+	evt->set_state_periodic = jh7110_timer_set_periodic;
+	evt->set_state_oneshot = jh7110_timer_set_oneshot;
+	evt->set_state_oneshot_stopped = jh7110_timer_shutdown;
+	evt->tick_resume = jh7110_timer_tick_resume;
+	evt->set_next_event = jh7110_timer_set_next_event;
+	evt->suspend = jh7110_timer_suspend;
+	evt->resume = jh7110_timer_resume;
+	evt->rating = JH7110_CLOCKEVENT_RATING;
+}
+
+static u64 jh7110_timer_clocksource_read(struct clocksource *cs)
+{
+	struct jh7110_clkevt *clkevt = container_of(cs, struct jh7110_clkevt, cs);
+
+	return (u64)jh7110_timer_get_val(clkevt);
+}
+
+static int jh7110_clocksource_init(struct jh7110_clkevt *clkevt)
+{
+	int ret;
+
+	jh7110_timer_set_continuous_mod(clkevt);
+	jh7110_timer_set_load(clkevt, JH7110_TIMER_MAX_TICKS);
+
+	ret = jh7110_timer_int_init_enable(clkevt);
+	if (ret)
+		return ret;
+
+	clkevt->cs.name = clkevt->name;
+	clkevt->cs.rating = JH7110_CLOCK_SOURCE_RATING;
+	clkevt->cs.read = jh7110_timer_clocksource_read;
+	clkevt->cs.mask = CLOCKSOURCE_MASK(JH7110_VALID_BITS);
+	clkevt->cs.flags = CLOCK_SOURCE_IS_CONTINUOUS;
+
+	return clocksource_register_hz(&clkevt->cs, clkevt->rate);
+}
+
+static void jh7110_clockevents_register(struct jh7110_clkevt *clkevt)
+{
+	clkevt->rate = clk_get_rate(clkevt->clk);
+
+	jh7110_set_clockevent(&clkevt->evt);
+	clkevt->evt.name = clkevt->name;
+	clkevt->evt.cpumask = cpu_possible_mask;
+
+	clockevents_config_and_register(&clkevt->evt, clkevt->rate,
+					JH7110_TIMER_MIN_TICKS, JH7110_TIMER_MAX_TICKS);
+}
+
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+	struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+	char name[4];
+	struct clk *pclk;
+	struct reset_control *rst;
+	int ch;
+	int ret;
+	void __iomem *base;
+
+	base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(base))
+		return dev_err_probe(&pdev->dev, PTR_ERR(base),
+				     "failed to map registers\n");
+
+	rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+	if (IS_ERR(rst))
+		return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+	pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+	if (IS_ERR(pclk))
+		return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+				     "failed to get & enable apb clock\n");
+
+	ret = reset_control_deassert(rst);
+	if (ret)
+		return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");
+
+	for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+		clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+		if (!clkevt[ch])
+			return -ENOMEM;
+
+		snprintf(name, sizeof(name), "ch%d", ch);
+
+		clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+		/* Ensure timer is disabled */
+		jh7110_timer_disable(clkevt[ch]);
+
+		rst = devm_reset_control_get_exclusive(&pdev->dev, name);
+		if (IS_ERR(rst))
+			return PTR_ERR(rst);
+
+		clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+		if (IS_ERR(clkevt[ch]->clk))
+			return PTR_ERR(clkevt[ch]->clk);
+
+		ret = reset_control_deassert(rst);
+		if (ret)
+			return ret;
+
+		clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+		if (clkevt[ch]->evt.irq < 0)
+			return clkevt[ch]->evt.irq;
+
+		snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+		jh7110_clockevents_register(clkevt[ch]);
+
+		ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+				       IRQF_TIMER | IRQF_IRQPOLL,
+				       clkevt[ch]->name, &clkevt[ch]->evt);
+		if (ret)
+			return ret;
+
+		ret = jh7110_clocksource_init(clkevt[ch]);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+	{ .compatible = "starfive,jh7110-timer", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+	.probe = jh7110_timer_probe,
+	.driver = {
+		.name = "jh7110-timer",
+		.of_match_table = jh7110_timer_match,
+	},
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");
--

2.25.1

_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv

author
#10 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

Quoted 129 lines · Emil Renner Berthing
Quoted 115 lines · Xingyu Wu

Add timer driver for the StarFive JH7110 SoC.

---
 MAINTAINERS                        |   7 +
 drivers/clocksource/Kconfig        |  11 +
 drivers/clocksource/Makefile       |   1 +
 drivers/clocksource/timer-jh7110.c | 387 +++++++++++++++++++++++++++++
 4 files changed, 406 insertions(+)
 create mode 100644 drivers/clocksource/timer-jh7110.c
MAINTAINERS +7 -0
index 6c4cce45a09d..1525f031d4a2 100644
@@ -20472,6 +20472,13 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/sound/starfive,jh7110-tdm.yaml
 F:	sound/soc/starfive/jh7110_tdm.c
 
+STARFIVE JH7110 TIMER DRIVER
+M:	Samin Guo <[email protected]>
+M:	Xingyu Wu <[email protected]>
+S:	Supported
+F:	Documentation/devicetree/bindings/timer/starfive,jh7110-timer.yaml
+F:	drivers/clocksource/timer-jh7110.c
+
 STARFIVE JH71X0 CLOCK DRIVERS
 M:	Emil Renner Berthing <[email protected]>
 M:	Hal Feng <[email protected]>
drivers/clocksource/Kconfig +11 -0
index 0ba0dc4ecf06..821abcc1e517 100644
@@ -641,6 +641,17 @@ config RISCV_TIMER
 	  is accessed via both the SBI and the rdcycle instruction.  This is
 	  required for all RISC-V systems.
 
+config STARFIVE_JH7110_TIMER
+	bool "Timer for the STARFIVE JH7110 SoC"
+	depends on ARCH_STARFIVE || COMPILE_TEST
+	select TIMER_OF
+	select CLKSRC_MMIO
+	default ARCH_STARFIVE
+	help
+	  This enables the timer for StarFive JH7110 SoC. On RISC-V platform,
+	  the system has started RISCV_TIMER, but you can also use this timer
+	  which can provide four channels to do a lot more things on JH7110 SoC.
+
 config CLINT_TIMER
 	bool "CLINT Timer for the RISC-V platform" if COMPILE_TEST
 	depends on GENERIC_SCHED_CLOCK && RISCV
drivers/clocksource/Makefile +1 -0
index 368c3461dab8..b66ac05ec086 100644
@@ -80,6 +80,7 @@ obj-$(CONFIG_INGENIC_TIMER)		+= ingenic-timer.o
 obj-$(CONFIG_CLKSRC_ST_LPC)		+= clksrc_st_lpc.o
 obj-$(CONFIG_X86_NUMACHIP)		+= numachip.o
 obj-$(CONFIG_RISCV_TIMER)		+= timer-riscv.o
+obj-$(CONFIG_STARFIVE_JH7110_TIMER)	+= timer-jh7110.o
 obj-$(CONFIG_CLINT_TIMER)		+= timer-clint.o
 obj-$(CONFIG_CSKY_MP_TIMER)		+= timer-mp-csky.o
 obj-$(CONFIG_GX6605S_TIMER)		+= timer-gx6605s.o
drivers/clocksource/timer-jh7110.c new file +46 -0
new file mode 100644
index 000000000000..914424368290
@@ -0,0 +1,387 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Starfive JH7110 Timer driver
+ *
+ * Copyright (C) 2022-2023 StarFive Technology Co., Ltd.
+ *
+ * Author:
+ * Xingyu Wu <[email protected]>
+ * Samin Guo <[email protected]>
+ */
+
+#include <linux/clk.h>
+#include <linux/clockchips.h>
+#include <linux/clocksource.h>
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/irq.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+#include <linux/sched_clock.h>
+
+/* Bias: Ch0-0x0, Ch1-0x40, Ch2-0x80, and so on. */
+#define JH7110_TIMER_CH_LEN		0x40
+#define JH7110_TIMER_CH_BASE(x)		((x) * JH7110_TIMER_CH_LEN)
+#define JH7110_TIMER_CH_MAX		4
+
+#define JH7110_CLOCK_SOURCE_RATING	200
+#define JH7110_VALID_BITS		32
+#define JH7110_DELAY_US			0
+#define JH7110_TIMEOUT_US		10000
+#define JH7110_CLOCKEVENT_RATING	300
+#define JH7110_TIMER_MAX_TICKS		0xffffffff
+#define JH7110_TIMER_MIN_TICKS		0xf
+#define JH7110_TIMER_NAME_NUM		19
+
+#define JH7110_TIMER_INT_STATUS		0x00 /* RO[0:4]: Interrupt Status for channel0~4 */
+#define JH7110_TIMER_CTL		0x04 /* RW[0]: 0-continuous run, 1-single run */
+#define JH7110_TIMER_LOAD		0x08 /* RW: load value to counter */
+#define JH7110_TIMER_ENABLE		0x10 /* RW[0]: timer enable register */
+#define JH7110_TIMER_RELOAD		0x14 /* RW: write 1 or 0 both reload counter */

drivers/clocksource/timer-jh7110.c line 56

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+#define JH7110_TIMER_VALUE		0x18 /* RO: timer value register */
+#define JH7110_TIMER_INT_CLR		0x20 /* RW: timer interrupt clear register */
+#define JH7110_TIMER_INT_MASK		0x24 /* RW[0]: timer interrupt mask register */
+#define JH7110_TIMER_INT_CLR_AVA_MASK	BIT(1)
+
+struct jh7110_clkevt {
+	struct clock_event_device evt;
+	struct clocksource cs;
+	struct clk *clk;
+	char name[JH7110_TIMER_NAME_NUM];

Instead of this define maybe just use something like

Quoted 2 lines
	char name[sizeof("jh7110-timer.chX")];

Will fix.

Quoted 3 lines

drivers/clocksource/timer-jh7110.c line 57

Quoted 1 line
drivers/clocksource/timer-jh7110.c new file +1 -0
+	int irq;

You wrote that you "dropped the extra copy of irq", but here you didn't.

What I mean is to drop the local variable of irq in the probe. Do you mean to use a local variable instead this in the probe?

Quoted 37 lines
Quoted 22 lines
drivers/clocksource/timer-jh7110.c new file +21 -0
+	u32 rate;
+	u32 reload_val;
+	void __iomem *base;
+};
+
+static inline struct jh7110_clkevt *to_jh7110_clkevt(struct clock_event_device *evt)
+{
+	return container_of(evt, struct jh7110_clkevt, evt);
+}
+
+/* 0:continuous-run mode, 1:single-run mode */
+static inline void jh7110_timer_set_continuous_mod(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_CTL);
+}
+
+static inline void jh7110_timer_set_single_mod(struct jh7110_clkevt *clkevt)
+{
+	writel(1, clkevt->base + JH7110_TIMER_CTL);
+}
+

drivers/clocksource/timer-jh7110.c line 88

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+/* Interrupt Mask Register, 0:Unmask, 1:Mask */
+static inline void jh7110_timer_int_enable(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_INT_MASK);
+}
+
+static inline void jh7110_timer_int_disable(struct jh7110_clkevt *clkevt)
+{
+	writel(1, clkevt->base + JH7110_TIMER_INT_MASK);
+}

I really don't think all these wrappers make the code any clearer. Please just inline the writel() calls. What you can do to make it a bit clearer is something like

Quoted 9 lines
enum {
	JH7110_TIMER_MODE_CONTINUOUS,
	JH7110_TIMER_MODE_SINGLE,
};

so you can write

	writel(JH7110_TIMER_MODE_CONTINUOUS, clkevt->base + JH7110_TIMER_CTL);

Will fix.

Quoted 50 lines
Quoted 50 lines
drivers/clocksource/timer-jh7110.c new file +50 -0
+
+/*
+ * BIT(0): Read value represent channel intr status.
+ * Write 1 to this bit to clear interrupt. Write 0 has no effects.
+ * BIT(1): "1" means that it is clearing interrupt. BIT(0) can not be written.
+ */
+static inline int jh7110_timer_int_clear(struct jh7110_clkevt *clkevt)
+{
+	u32 value;
+	int ret;
+
+	/* waiting interrupt can be to clearing */
+	ret = readl_poll_timeout_atomic(clkevt->base + JH7110_TIMER_INT_CLR, value,
+					!(value & JH7110_TIMER_INT_CLR_AVA_MASK),
+					JH7110_DELAY_US, JH7110_TIMEOUT_US);
+	if (!ret)
+		writel(0x1, clkevt->base + JH7110_TIMER_INT_CLR);
+
+	return ret;
+}
+
+/*
+ * The initial value to be loaded into the
+ * counter and is also used as the reload value.
+ * val = clock rate --> 1s
+ */
+static inline void jh7110_timer_set_load(struct jh7110_clkevt *clkevt, u32 val)
+{
+	writel(val, clkevt->base + JH7110_TIMER_LOAD);
+}
+
+static inline u32 jh7110_timer_get_val(struct jh7110_clkevt *clkevt)
+{
+	return readl(clkevt->base + JH7110_TIMER_VALUE);
+}
+
+/*
+ * Write RELOAD register to reload preset value to counter.
+ * Write 0 and write 1 are both ok.
+ */
+static inline void jh7110_timer_set_reload(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_RELOAD);
+}
+
+static inline void jh7110_timer_enable(struct jh7110_clkevt *clkevt)
+{
+	writel(1, clkevt->base + JH7110_TIMER_ENABLE);
+}
+
Quoted 8 lines

drivers/clocksource/timer-jh7110.c line 142

Quoted 4 lines
drivers/clocksource/timer-jh7110.c new file +4 -0
+static inline void jh7110_timer_disable(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_ENABLE);
+}

The same thing goes for all these wrappers. If you do insist on having these wrappers then at least group them all together so you can quickly scroll past them to get to the "real" code.

Will fix.

Quoted 250 lines
Quoted 250 lines
drivers/clocksource/timer-jh7110.c new file +245 -1
+
+static int jh7110_timer_int_init_enable(struct jh7110_clkevt *clkevt)
+{
+	int ret;
+
+	jh7110_timer_int_disable(clkevt);
+	ret = jh7110_timer_int_clear(clkevt);
+	if (ret)
+		return ret;
+
+	jh7110_timer_int_enable(clkevt);
+	jh7110_timer_enable(clkevt);
+
+	return 0;
+}
+
+static int jh7110_timer_shutdown(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_disable(clkevt);
+	return jh7110_timer_int_clear(clkevt);
+}
+
+static void jh7110_timer_suspend(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	clkevt->reload_val = jh7110_timer_get_val(clkevt);
+	jh7110_timer_shutdown(evt);
+}
+
+static void jh7110_timer_resume(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_set_load(clkevt, clkevt->reload_val);
+	jh7110_timer_set_reload(clkevt);
+	jh7110_timer_int_enable(clkevt);
+	jh7110_timer_enable(clkevt);
+}
+
+static int jh7110_timer_tick_resume(struct clock_event_device *evt)
+{
+	jh7110_timer_resume(evt);
+
+	return 0;
+}
+
+/* IRQ handler for the timer */
+static irqreturn_t jh7110_timer_interrupt(int irq, void *priv)
+{
+	struct clock_event_device *evt = (struct clock_event_device *)priv;
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	if (jh7110_timer_int_clear(clkevt))
+		return IRQ_NONE;
+
+	if (evt->event_handler)
+		evt->event_handler(evt);
+
+	return IRQ_HANDLED;
+}
+
+static int jh7110_timer_set_periodic(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+	u32 periodic = DIV_ROUND_CLOSEST(clkevt->rate, HZ);
+
+	jh7110_timer_disable(clkevt);
+	jh7110_timer_set_continuous_mod(clkevt);
+	jh7110_timer_set_load(clkevt, periodic);
+
+	return jh7110_timer_int_init_enable(clkevt);
+}
+
+static int jh7110_timer_set_oneshot(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_disable(clkevt);
+	jh7110_timer_set_single_mod(clkevt);
+	jh7110_timer_set_load(clkevt, JH7110_TIMER_MAX_TICKS);
+
+	return jh7110_timer_int_init_enable(clkevt);
+}
+
+static int jh7110_timer_set_next_event(unsigned long next,
+				       struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_disable(clkevt);
+	jh7110_timer_set_single_mod(clkevt);
+	jh7110_timer_set_load(clkevt, next);
+	jh7110_timer_enable(clkevt);
+
+	return 0;
+}
+
+static void jh7110_set_clockevent(struct clock_event_device *evt)
+{
+	evt->features = CLOCK_EVT_FEAT_PERIODIC |
+			CLOCK_EVT_FEAT_ONESHOT |
+			CLOCK_EVT_FEAT_DYNIRQ;
+	evt->set_state_shutdown = jh7110_timer_shutdown;
+	evt->set_state_periodic = jh7110_timer_set_periodic;
+	evt->set_state_oneshot = jh7110_timer_set_oneshot;
+	evt->set_state_oneshot_stopped = jh7110_timer_shutdown;
+	evt->tick_resume = jh7110_timer_tick_resume;
+	evt->set_next_event = jh7110_timer_set_next_event;
+	evt->suspend = jh7110_timer_suspend;
+	evt->resume = jh7110_timer_resume;
+	evt->rating = JH7110_CLOCKEVENT_RATING;
+}
+
+static u64 jh7110_timer_clocksource_read(struct clocksource *cs)
+{
+	struct jh7110_clkevt *clkevt = container_of(cs, struct jh7110_clkevt, cs);
+
+	return (u64)jh7110_timer_get_val(clkevt);
+}
+
+static int jh7110_clocksource_init(struct jh7110_clkevt *clkevt)
+{
+	int ret;
+
+	jh7110_timer_set_continuous_mod(clkevt);
+	jh7110_timer_set_load(clkevt, JH7110_TIMER_MAX_TICKS);
+
+	ret = jh7110_timer_int_init_enable(clkevt);
+	if (ret)
+		return ret;
+
+	clkevt->cs.name = clkevt->name;
+	clkevt->cs.rating = JH7110_CLOCK_SOURCE_RATING;
+	clkevt->cs.read = jh7110_timer_clocksource_read;
+	clkevt->cs.mask = CLOCKSOURCE_MASK(JH7110_VALID_BITS);
+	clkevt->cs.flags = CLOCK_SOURCE_IS_CONTINUOUS;
+
+	return clocksource_register_hz(&clkevt->cs, clkevt->rate);
+}
+
+static void jh7110_clockevents_register(struct jh7110_clkevt *clkevt)
+{
+	clkevt->rate = clk_get_rate(clkevt->clk);
+
+	jh7110_set_clockevent(&clkevt->evt);
+	clkevt->evt.name = clkevt->name;
+	clkevt->evt.cpumask = cpu_possible_mask;
+
+	clockevents_config_and_register(&clkevt->evt, clkevt->rate,
+					JH7110_TIMER_MIN_TICKS, JH7110_TIMER_MAX_TICKS);
+}
+
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+	struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+	char name[4];
+	struct clk *pclk;
+	struct reset_control *rst;
+	int ch;
+	int ret;
+	void __iomem *base;
+
+	base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(base))
+		return dev_err_probe(&pdev->dev, PTR_ERR(base),
+				     "failed to map registers\n");
+
+	rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+	if (IS_ERR(rst))
+		return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+	pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+	if (IS_ERR(pclk))
+		return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+				     "failed to get & enable apb clock\n");
+
+	ret = reset_control_deassert(rst);
+	if (ret)
+		return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");
+
+	for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+		clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+		if (!clkevt[ch])
+			return -ENOMEM;
+
+		snprintf(name, sizeof(name), "ch%d", ch);
+
+		clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+		/* Ensure timer is disabled */
+		jh7110_timer_disable(clkevt[ch]);
+
+		rst = devm_reset_control_get_exclusive(&pdev->dev, name);
+		if (IS_ERR(rst))
+			return PTR_ERR(rst);
+
+		clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+		if (IS_ERR(clkevt[ch]->clk))
+			return PTR_ERR(clkevt[ch]->clk);
+
+		ret = reset_control_deassert(rst);
+		if (ret)
+			return ret;
+
+		clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+		if (clkevt[ch]->evt.irq < 0)
+			return clkevt[ch]->evt.irq;
+
+		snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+		jh7110_clockevents_register(clkevt[ch]);
+
+		ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+				       IRQF_TIMER | IRQF_IRQPOLL,
+				       clkevt[ch]->name, &clkevt[ch]->evt);
+		if (ret)
+			return ret;
+
+		ret = jh7110_clocksource_init(clkevt[ch]);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+	{ .compatible = "starfive,jh7110-timer", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+	.probe = jh7110_timer_probe,
+	.driver = {
+		.name = "jh7110-timer",
+		.of_match_table = jh7110_timer_match,
+	},
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");
--

2.25.1

Quoted 4 lines
Quoted 4 lines

_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv

Best regards,
Xingyu Wu

#11 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

Quoted 133 lines · Xingyu Wu
Quoted 129 lines · Emil Renner Berthing
Quoted 115 lines · Xingyu Wu

Add timer driver for the StarFive JH7110 SoC.

---
 MAINTAINERS                        |   7 +
 drivers/clocksource/Kconfig        |  11 +
 drivers/clocksource/Makefile       |   1 +
 drivers/clocksource/timer-jh7110.c | 387 +++++++++++++++++++++++++++++
 4 files changed, 406 insertions(+)
 create mode 100644 drivers/clocksource/timer-jh7110.c
MAINTAINERS +7 -0
index 6c4cce45a09d..1525f031d4a2 100644
@@ -20472,6 +20472,13 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/sound/starfive,jh7110-tdm.yaml
 F:	sound/soc/starfive/jh7110_tdm.c
 
+STARFIVE JH7110 TIMER DRIVER
+M:	Samin Guo <[email protected]>
+M:	Xingyu Wu <[email protected]>
+S:	Supported
+F:	Documentation/devicetree/bindings/timer/starfive,jh7110-timer.yaml
+F:	drivers/clocksource/timer-jh7110.c
+
 STARFIVE JH71X0 CLOCK DRIVERS
 M:	Emil Renner Berthing <[email protected]>
 M:	Hal Feng <[email protected]>
drivers/clocksource/Kconfig +11 -0
index 0ba0dc4ecf06..821abcc1e517 100644
@@ -641,6 +641,17 @@ config RISCV_TIMER
 	  is accessed via both the SBI and the rdcycle instruction.  This is
 	  required for all RISC-V systems.
 
+config STARFIVE_JH7110_TIMER
+	bool "Timer for the STARFIVE JH7110 SoC"
+	depends on ARCH_STARFIVE || COMPILE_TEST
+	select TIMER_OF
+	select CLKSRC_MMIO
+	default ARCH_STARFIVE
+	help
+	  This enables the timer for StarFive JH7110 SoC. On RISC-V platform,
+	  the system has started RISCV_TIMER, but you can also use this timer
+	  which can provide four channels to do a lot more things on JH7110 SoC.
+
 config CLINT_TIMER
 	bool "CLINT Timer for the RISC-V platform" if COMPILE_TEST
 	depends on GENERIC_SCHED_CLOCK && RISCV
drivers/clocksource/Makefile +1 -0
index 368c3461dab8..b66ac05ec086 100644
@@ -80,6 +80,7 @@ obj-$(CONFIG_INGENIC_TIMER)		+= ingenic-timer.o
 obj-$(CONFIG_CLKSRC_ST_LPC)		+= clksrc_st_lpc.o
 obj-$(CONFIG_X86_NUMACHIP)		+= numachip.o
 obj-$(CONFIG_RISCV_TIMER)		+= timer-riscv.o
+obj-$(CONFIG_STARFIVE_JH7110_TIMER)	+= timer-jh7110.o
 obj-$(CONFIG_CLINT_TIMER)		+= timer-clint.o
 obj-$(CONFIG_CSKY_MP_TIMER)		+= timer-mp-csky.o
 obj-$(CONFIG_GX6605S_TIMER)		+= timer-gx6605s.o
drivers/clocksource/timer-jh7110.c new file +46 -0
new file mode 100644
index 000000000000..914424368290
@@ -0,0 +1,387 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Starfive JH7110 Timer driver
+ *
+ * Copyright (C) 2022-2023 StarFive Technology Co., Ltd.
+ *
+ * Author:
+ * Xingyu Wu <[email protected]>
+ * Samin Guo <[email protected]>
+ */
+
+#include <linux/clk.h>
+#include <linux/clockchips.h>
+#include <linux/clocksource.h>
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/irq.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+#include <linux/sched_clock.h>
+
+/* Bias: Ch0-0x0, Ch1-0x40, Ch2-0x80, and so on. */
+#define JH7110_TIMER_CH_LEN		0x40
+#define JH7110_TIMER_CH_BASE(x)		((x) * JH7110_TIMER_CH_LEN)
+#define JH7110_TIMER_CH_MAX		4
+
+#define JH7110_CLOCK_SOURCE_RATING	200
+#define JH7110_VALID_BITS		32
+#define JH7110_DELAY_US			0
+#define JH7110_TIMEOUT_US		10000
+#define JH7110_CLOCKEVENT_RATING	300
+#define JH7110_TIMER_MAX_TICKS		0xffffffff
+#define JH7110_TIMER_MIN_TICKS		0xf
+#define JH7110_TIMER_NAME_NUM		19
+
+#define JH7110_TIMER_INT_STATUS		0x00 /* RO[0:4]: Interrupt Status for channel0~4 */
+#define JH7110_TIMER_CTL		0x04 /* RW[0]: 0-continuous run, 1-single run */
+#define JH7110_TIMER_LOAD		0x08 /* RW: load value to counter */
+#define JH7110_TIMER_ENABLE		0x10 /* RW[0]: timer enable register */
+#define JH7110_TIMER_RELOAD		0x14 /* RW: write 1 or 0 both reload counter */

drivers/clocksource/timer-jh7110.c line 56

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+#define JH7110_TIMER_VALUE		0x18 /* RO: timer value register */
+#define JH7110_TIMER_INT_CLR		0x20 /* RW: timer interrupt clear register */
+#define JH7110_TIMER_INT_MASK		0x24 /* RW[0]: timer interrupt mask register */
+#define JH7110_TIMER_INT_CLR_AVA_MASK	BIT(1)
+
+struct jh7110_clkevt {
+	struct clock_event_device evt;
+	struct clocksource cs;
+	struct clk *clk;
+	char name[JH7110_TIMER_NAME_NUM];

Instead of this define maybe just use something like

Quoted 2 lines
	char name[sizeof("jh7110-timer.chX")];
Quoted 7 lines

Will fix.

Quoted 3 lines

drivers/clocksource/timer-jh7110.c line 57

Quoted 1 line
drivers/clocksource/timer-jh7110.c new file +1 -0
+	int irq;

You wrote that you "dropped the extra copy of irq", but here you didn't.

What I mean is to drop the local variable of irq in the probe. Do you mean to use a local variable instead this in the probe?

No, what I mean is that this field is no longer used, so you can delete it.

Quoted 359 lines
Quoted 38 lines
Quoted 23 lines
drivers/clocksource/timer-jh7110.c new file +21 -0
+	u32 rate;
+	u32 reload_val;
+	void __iomem *base;
+};
+
+static inline struct jh7110_clkevt *to_jh7110_clkevt(struct clock_event_device *evt)
+{
+	return container_of(evt, struct jh7110_clkevt, evt);
+}
+
+/* 0:continuous-run mode, 1:single-run mode */
+static inline void jh7110_timer_set_continuous_mod(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_CTL);
+}
+
+static inline void jh7110_timer_set_single_mod(struct jh7110_clkevt *clkevt)
+{
+	writel(1, clkevt->base + JH7110_TIMER_CTL);
+}
+

drivers/clocksource/timer-jh7110.c line 88

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+/* Interrupt Mask Register, 0:Unmask, 1:Mask */
+static inline void jh7110_timer_int_enable(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_INT_MASK);
+}
+
+static inline void jh7110_timer_int_disable(struct jh7110_clkevt *clkevt)
+{
+	writel(1, clkevt->base + JH7110_TIMER_INT_MASK);
+}

I really don't think all these wrappers make the code any clearer. Please just inline the writel() calls. What you can do to make it a bit clearer is something like

Quoted 9 lines
enum {
	JH7110_TIMER_MODE_CONTINUOUS,
	JH7110_TIMER_MODE_SINGLE,
};

so you can write

	writel(JH7110_TIMER_MODE_CONTINUOUS, clkevt->base + JH7110_TIMER_CTL);

Will fix.

Quoted 50 lines
Quoted 50 lines
drivers/clocksource/timer-jh7110.c new file +50 -0
+
+/*
+ * BIT(0): Read value represent channel intr status.
+ * Write 1 to this bit to clear interrupt. Write 0 has no effects.
+ * BIT(1): "1" means that it is clearing interrupt. BIT(0) can not be written.
+ */
+static inline int jh7110_timer_int_clear(struct jh7110_clkevt *clkevt)
+{
+	u32 value;
+	int ret;
+
+	/* waiting interrupt can be to clearing */
+	ret = readl_poll_timeout_atomic(clkevt->base + JH7110_TIMER_INT_CLR, value,
+					!(value & JH7110_TIMER_INT_CLR_AVA_MASK),
+					JH7110_DELAY_US, JH7110_TIMEOUT_US);
+	if (!ret)
+		writel(0x1, clkevt->base + JH7110_TIMER_INT_CLR);
+
+	return ret;
+}
+
+/*
+ * The initial value to be loaded into the
+ * counter and is also used as the reload value.
+ * val = clock rate --> 1s
+ */
+static inline void jh7110_timer_set_load(struct jh7110_clkevt *clkevt, u32 val)
+{
+	writel(val, clkevt->base + JH7110_TIMER_LOAD);
+}
+
+static inline u32 jh7110_timer_get_val(struct jh7110_clkevt *clkevt)
+{
+	return readl(clkevt->base + JH7110_TIMER_VALUE);
+}
+
+/*
+ * Write RELOAD register to reload preset value to counter.
+ * Write 0 and write 1 are both ok.
+ */
+static inline void jh7110_timer_set_reload(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_RELOAD);
+}
+
+static inline void jh7110_timer_enable(struct jh7110_clkevt *clkevt)
+{
+	writel(1, clkevt->base + JH7110_TIMER_ENABLE);
+}
+
Quoted 8 lines

drivers/clocksource/timer-jh7110.c line 142

Quoted 4 lines
drivers/clocksource/timer-jh7110.c new file +4 -0
+static inline void jh7110_timer_disable(struct jh7110_clkevt *clkevt)
+{
+	writel(0, clkevt->base + JH7110_TIMER_ENABLE);
+}

The same thing goes for all these wrappers. If you do insist on having these wrappers then at least group them all together so you can quickly scroll past them to get to the "real" code.

Will fix.

Quoted 248 lines
Quoted 248 lines
drivers/clocksource/timer-jh7110.c new file +245 -1
+
+static int jh7110_timer_int_init_enable(struct jh7110_clkevt *clkevt)
+{
+	int ret;
+
+	jh7110_timer_int_disable(clkevt);
+	ret = jh7110_timer_int_clear(clkevt);
+	if (ret)
+		return ret;
+
+	jh7110_timer_int_enable(clkevt);
+	jh7110_timer_enable(clkevt);
+
+	return 0;
+}
+
+static int jh7110_timer_shutdown(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_disable(clkevt);
+	return jh7110_timer_int_clear(clkevt);
+}
+
+static void jh7110_timer_suspend(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	clkevt->reload_val = jh7110_timer_get_val(clkevt);
+	jh7110_timer_shutdown(evt);
+}
+
+static void jh7110_timer_resume(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_set_load(clkevt, clkevt->reload_val);
+	jh7110_timer_set_reload(clkevt);
+	jh7110_timer_int_enable(clkevt);
+	jh7110_timer_enable(clkevt);
+}
+
+static int jh7110_timer_tick_resume(struct clock_event_device *evt)
+{
+	jh7110_timer_resume(evt);
+
+	return 0;
+}
+
+/* IRQ handler for the timer */
+static irqreturn_t jh7110_timer_interrupt(int irq, void *priv)
+{
+	struct clock_event_device *evt = (struct clock_event_device *)priv;
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	if (jh7110_timer_int_clear(clkevt))
+		return IRQ_NONE;
+
+	if (evt->event_handler)
+		evt->event_handler(evt);
+
+	return IRQ_HANDLED;
+}
+
+static int jh7110_timer_set_periodic(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+	u32 periodic = DIV_ROUND_CLOSEST(clkevt->rate, HZ);
+
+	jh7110_timer_disable(clkevt);
+	jh7110_timer_set_continuous_mod(clkevt);
+	jh7110_timer_set_load(clkevt, periodic);
+
+	return jh7110_timer_int_init_enable(clkevt);
+}
+
+static int jh7110_timer_set_oneshot(struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_disable(clkevt);
+	jh7110_timer_set_single_mod(clkevt);
+	jh7110_timer_set_load(clkevt, JH7110_TIMER_MAX_TICKS);
+
+	return jh7110_timer_int_init_enable(clkevt);
+}
+
+static int jh7110_timer_set_next_event(unsigned long next,
+				       struct clock_event_device *evt)
+{
+	struct jh7110_clkevt *clkevt = to_jh7110_clkevt(evt);
+
+	jh7110_timer_disable(clkevt);
+	jh7110_timer_set_single_mod(clkevt);
+	jh7110_timer_set_load(clkevt, next);
+	jh7110_timer_enable(clkevt);
+
+	return 0;
+}
+
+static void jh7110_set_clockevent(struct clock_event_device *evt)
+{
+	evt->features = CLOCK_EVT_FEAT_PERIODIC |
+			CLOCK_EVT_FEAT_ONESHOT |
+			CLOCK_EVT_FEAT_DYNIRQ;
+	evt->set_state_shutdown = jh7110_timer_shutdown;
+	evt->set_state_periodic = jh7110_timer_set_periodic;
+	evt->set_state_oneshot = jh7110_timer_set_oneshot;
+	evt->set_state_oneshot_stopped = jh7110_timer_shutdown;
+	evt->tick_resume = jh7110_timer_tick_resume;
+	evt->set_next_event = jh7110_timer_set_next_event;
+	evt->suspend = jh7110_timer_suspend;
+	evt->resume = jh7110_timer_resume;
+	evt->rating = JH7110_CLOCKEVENT_RATING;
+}
+
+static u64 jh7110_timer_clocksource_read(struct clocksource *cs)
+{
+	struct jh7110_clkevt *clkevt = container_of(cs, struct jh7110_clkevt, cs);
+
+	return (u64)jh7110_timer_get_val(clkevt);
+}
+
+static int jh7110_clocksource_init(struct jh7110_clkevt *clkevt)
+{
+	int ret;
+
+	jh7110_timer_set_continuous_mod(clkevt);
+	jh7110_timer_set_load(clkevt, JH7110_TIMER_MAX_TICKS);
+
+	ret = jh7110_timer_int_init_enable(clkevt);
+	if (ret)
+		return ret;
+
+	clkevt->cs.name = clkevt->name;
+	clkevt->cs.rating = JH7110_CLOCK_SOURCE_RATING;
+	clkevt->cs.read = jh7110_timer_clocksource_read;
+	clkevt->cs.mask = CLOCKSOURCE_MASK(JH7110_VALID_BITS);
+	clkevt->cs.flags = CLOCK_SOURCE_IS_CONTINUOUS;
+
+	return clocksource_register_hz(&clkevt->cs, clkevt->rate);
+}
+
+static void jh7110_clockevents_register(struct jh7110_clkevt *clkevt)
+{
+	clkevt->rate = clk_get_rate(clkevt->clk);
+
+	jh7110_set_clockevent(&clkevt->evt);
+	clkevt->evt.name = clkevt->name;
+	clkevt->evt.cpumask = cpu_possible_mask;
+
+	clockevents_config_and_register(&clkevt->evt, clkevt->rate,
+					JH7110_TIMER_MIN_TICKS, JH7110_TIMER_MAX_TICKS);
+}
+
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+	struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+	char name[4];
+	struct clk *pclk;
+	struct reset_control *rst;
+	int ch;
+	int ret;
+	void __iomem *base;
+
+	base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(base))
+		return dev_err_probe(&pdev->dev, PTR_ERR(base),
+				     "failed to map registers\n");
+
+	rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+	if (IS_ERR(rst))
+		return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+	pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+	if (IS_ERR(pclk))
+		return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+				     "failed to get & enable apb clock\n");
+
+	ret = reset_control_deassert(rst);
+	if (ret)
+		return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");
+
+	for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+		clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+		if (!clkevt[ch])
+			return -ENOMEM;
+
+		snprintf(name, sizeof(name), "ch%d", ch);
+
+		clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+		/* Ensure timer is disabled */
+		jh7110_timer_disable(clkevt[ch]);
+
+		rst = devm_reset_control_get_exclusive(&pdev->dev, name);
+		if (IS_ERR(rst))
+			return PTR_ERR(rst);
+
+		clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+		if (IS_ERR(clkevt[ch]->clk))
+			return PTR_ERR(clkevt[ch]->clk);
+
+		ret = reset_control_deassert(rst);
+		if (ret)
+			return ret;
+
+		clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+		if (clkevt[ch]->evt.irq < 0)
+			return clkevt[ch]->evt.irq;
+
+		snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+		jh7110_clockevents_register(clkevt[ch]);
+
+		ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+				       IRQF_TIMER | IRQF_IRQPOLL,
+				       clkevt[ch]->name, &clkevt[ch]->evt);
+		if (ret)
+			return ret;
+
+		ret = jh7110_clocksource_init(clkevt[ch]);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+	{ .compatible = "starfive,jh7110-timer", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+	.probe = jh7110_timer_probe,
+	.driver = {
+		.name = "jh7110-timer",
+		.of_match_table = jh7110_timer_match,
+	},
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");
--

2.25.1

author
#12 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

Quoted 136 lines · Emil Renner Berthing
Quoted 135 lines · Xingyu Wu
Quoted 129 lines · Emil Renner Berthing
Quoted 115 lines · Xingyu Wu

Add timer driver for the StarFive JH7110 SoC.

---
 MAINTAINERS                        |   7 +
 drivers/clocksource/Kconfig        |  11 +
 drivers/clocksource/Makefile       |   1 +
 drivers/clocksource/timer-jh7110.c | 387 +++++++++++++++++++++++++++++
 4 files changed, 406 insertions(+)
 create mode 100644 drivers/clocksource/timer-jh7110.c
MAINTAINERS +7 -0
index 6c4cce45a09d..1525f031d4a2 100644
@@ -20472,6 +20472,13 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/sound/starfive,jh7110-tdm.yaml
 F:	sound/soc/starfive/jh7110_tdm.c
 
+STARFIVE JH7110 TIMER DRIVER
+M:	Samin Guo <[email protected]>
+M:	Xingyu Wu <[email protected]>
+S:	Supported
+F:	Documentation/devicetree/bindings/timer/starfive,jh7110-timer.yaml
+F:	drivers/clocksource/timer-jh7110.c
+
 STARFIVE JH71X0 CLOCK DRIVERS
 M:	Emil Renner Berthing <[email protected]>
 M:	Hal Feng <[email protected]>
drivers/clocksource/Kconfig +11 -0
index 0ba0dc4ecf06..821abcc1e517 100644
@@ -641,6 +641,17 @@ config RISCV_TIMER
 	  is accessed via both the SBI and the rdcycle instruction.  This is
 	  required for all RISC-V systems.
 
+config STARFIVE_JH7110_TIMER
+	bool "Timer for the STARFIVE JH7110 SoC"
+	depends on ARCH_STARFIVE || COMPILE_TEST
+	select TIMER_OF
+	select CLKSRC_MMIO
+	default ARCH_STARFIVE
+	help
+	  This enables the timer for StarFive JH7110 SoC. On RISC-V platform,
+	  the system has started RISCV_TIMER, but you can also use this timer
+	  which can provide four channels to do a lot more things on JH7110 SoC.
+
 config CLINT_TIMER
 	bool "CLINT Timer for the RISC-V platform" if COMPILE_TEST
 	depends on GENERIC_SCHED_CLOCK && RISCV
drivers/clocksource/Makefile +1 -0
index 368c3461dab8..b66ac05ec086 100644
@@ -80,6 +80,7 @@ obj-$(CONFIG_INGENIC_TIMER)		+= ingenic-timer.o
 obj-$(CONFIG_CLKSRC_ST_LPC)		+= clksrc_st_lpc.o
 obj-$(CONFIG_X86_NUMACHIP)		+= numachip.o
 obj-$(CONFIG_RISCV_TIMER)		+= timer-riscv.o
+obj-$(CONFIG_STARFIVE_JH7110_TIMER)	+= timer-jh7110.o
 obj-$(CONFIG_CLINT_TIMER)		+= timer-clint.o
 obj-$(CONFIG_CSKY_MP_TIMER)		+= timer-mp-csky.o
 obj-$(CONFIG_GX6605S_TIMER)		+= timer-gx6605s.o
drivers/clocksource/timer-jh7110.c new file +46 -0
new file mode 100644
index 000000000000..914424368290
@@ -0,0 +1,387 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Starfive JH7110 Timer driver
+ *
+ * Copyright (C) 2022-2023 StarFive Technology Co., Ltd.
+ *
+ * Author:
+ * Xingyu Wu <[email protected]>
+ * Samin Guo <[email protected]>
+ */
+
+#include <linux/clk.h>
+#include <linux/clockchips.h>
+#include <linux/clocksource.h>
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/irq.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+#include <linux/sched_clock.h>
+
+/* Bias: Ch0-0x0, Ch1-0x40, Ch2-0x80, and so on. */
+#define JH7110_TIMER_CH_LEN		0x40
+#define JH7110_TIMER_CH_BASE(x)		((x) * JH7110_TIMER_CH_LEN)
+#define JH7110_TIMER_CH_MAX		4
+
+#define JH7110_CLOCK_SOURCE_RATING	200
+#define JH7110_VALID_BITS		32
+#define JH7110_DELAY_US			0
+#define JH7110_TIMEOUT_US		10000
+#define JH7110_CLOCKEVENT_RATING	300
+#define JH7110_TIMER_MAX_TICKS		0xffffffff
+#define JH7110_TIMER_MIN_TICKS		0xf
+#define JH7110_TIMER_NAME_NUM		19
+
+#define JH7110_TIMER_INT_STATUS		0x00 /* RO[0:4]: Interrupt Status for channel0~4 */
+#define JH7110_TIMER_CTL		0x04 /* RW[0]: 0-continuous run, 1-single run */
+#define JH7110_TIMER_LOAD		0x08 /* RW: load value to counter */
+#define JH7110_TIMER_ENABLE		0x10 /* RW[0]: timer enable register */
+#define JH7110_TIMER_RELOAD		0x14 /* RW: write 1 or 0 both reload counter */

drivers/clocksource/timer-jh7110.c line 56

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+#define JH7110_TIMER_VALUE		0x18 /* RO: timer value register */
+#define JH7110_TIMER_INT_CLR		0x20 /* RW: timer interrupt clear register */
+#define JH7110_TIMER_INT_MASK		0x24 /* RW[0]: timer interrupt mask register */
+#define JH7110_TIMER_INT_CLR_AVA_MASK	BIT(1)
+
+struct jh7110_clkevt {
+	struct clock_event_device evt;
+	struct clocksource cs;
+	struct clk *clk;
+	char name[JH7110_TIMER_NAME_NUM];

Instead of this define maybe just use something like

Quoted 2 lines
	char name[sizeof("jh7110-timer.chX")];

Will fix.

Quoted 8 lines
Quoted 5 lines
Quoted 3 lines

drivers/clocksource/timer-jh7110.c line 57

Quoted 1 line
drivers/clocksource/timer-jh7110.c new file +1 -0
+	int irq;

You wrote that you "dropped the extra copy of irq", but here you didn't.

What I mean is to drop the local variable of irq in the probe. Do you mean to use a local variable instead this in the probe?

No, what I mean is that this field is no longer used, so you can delete it.

OK, drop it here.

Quoted 48 lines
Quoted 48 lines
Quoted 48 lines
Quoted 48 lines
drivers/clocksource/timer-jh7110.c new file +47 -0
+
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+	struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+	char name[4];
+	struct clk *pclk;
+	struct reset_control *rst;
+	int ch;
+	int ret;
+	void __iomem *base
+
+	base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(base))
+		return dev_err_probe(&pdev->dev, PTR_ERR(base),
+				     "failed to map registers\n");
+
+	rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+	if (IS_ERR(rst))
+		return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+	pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+	if (IS_ERR(pclk))
+		return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+				     "failed to get & enable apb clock\n");
+
+	ret = reset_control_deassert(rst);
+	if (ret)
+		return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");
+
+	for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+		clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+		if (!clkevt[ch])
+			return -ENOMEM;
+
+		snprintf(name, sizeof(name), "ch%d", ch);
+
+		clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+		/* Ensure timer is disabled */
+		jh7110_timer_disable(clkevt[ch]);
+
+		rst = devm_reset_control_get_exclusive(&pdev->dev, name);
+		if (IS_ERR(rst))
+			return PTR_ERR(rst);
+
+		clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+		if (IS_ERR(clkevt[ch]->clk))
+			return PTR_ERR(clkevt[ch]->clk);
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

drivers/clocksource/timer-jh7110.c line 351

Quoted 8 lines
drivers/clocksource/timer-jh7110.c new file +8 -0
+
+		ret = reset_control_deassert(rst);
+		if (ret)
+			return ret;
+
+		clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+		if (clkevt[ch]->evt.irq < 0)
+			return clkevt[ch]->evt.irq;

I should use a local variable of irq instead here. Right?

Quoted 30 lines
Quoted 30 lines
Quoted 30 lines
Quoted 30 lines
drivers/clocksource/timer-jh7110.c new file +30 -0
+
+		snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+		jh7110_clockevents_register(clkevt[ch]);
+
+		ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+				       IRQF_TIMER | IRQF_IRQPOLL,
+				       clkevt[ch]->name, &clkevt[ch]->evt);
+		if (ret)
+			return ret;
+
+		ret = jh7110_clocksource_init(clkevt[ch]);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+	{ .compatible = "starfive,jh7110-timer", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+	.probe = jh7110_timer_probe,
+	.driver = {
+		.name = "jh7110-timer",
+		.of_match_table = jh7110_timer_match,
+	},
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
drivers/clocksource/timer-jh7110.c new file +6 -1
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");
--

2.25.1

Thanks,
Xingyu Wu

#13 lore raw

Re: [PATCH v6 2/3] clocksource: Add JH7110 timer driver

Quoted 197 lines · Xingyu Wu
Quoted 136 lines · Emil Renner Berthing
Quoted 135 lines · Xingyu Wu
Quoted 129 lines · Emil Renner Berthing
Quoted 115 lines · Xingyu Wu

Add timer driver for the StarFive JH7110 SoC.

---
 MAINTAINERS                        |   7 +
 drivers/clocksource/Kconfig        |  11 +
 drivers/clocksource/Makefile       |   1 +
 drivers/clocksource/timer-jh7110.c | 387 +++++++++++++++++++++++++++++
 4 files changed, 406 insertions(+)
 create mode 100644 drivers/clocksource/timer-jh7110.c
MAINTAINERS +7 -0
index 6c4cce45a09d..1525f031d4a2 100644
@@ -20472,6 +20472,13 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/sound/starfive,jh7110-tdm.yaml
 F:	sound/soc/starfive/jh7110_tdm.c
 
+STARFIVE JH7110 TIMER DRIVER
+M:	Samin Guo <[email protected]>
+M:	Xingyu Wu <[email protected]>
+S:	Supported
+F:	Documentation/devicetree/bindings/timer/starfive,jh7110-timer.yaml
+F:	drivers/clocksource/timer-jh7110.c
+
 STARFIVE JH71X0 CLOCK DRIVERS
 M:	Emil Renner Berthing <[email protected]>
 M:	Hal Feng <[email protected]>
drivers/clocksource/Kconfig +11 -0
index 0ba0dc4ecf06..821abcc1e517 100644
@@ -641,6 +641,17 @@ config RISCV_TIMER
 	  is accessed via both the SBI and the rdcycle instruction.  This is
 	  required for all RISC-V systems.
 
+config STARFIVE_JH7110_TIMER
+	bool "Timer for the STARFIVE JH7110 SoC"
+	depends on ARCH_STARFIVE || COMPILE_TEST
+	select TIMER_OF
+	select CLKSRC_MMIO
+	default ARCH_STARFIVE
+	help
+	  This enables the timer for StarFive JH7110 SoC. On RISC-V platform,
+	  the system has started RISCV_TIMER, but you can also use this timer
+	  which can provide four channels to do a lot more things on JH7110 SoC.
+
 config CLINT_TIMER
 	bool "CLINT Timer for the RISC-V platform" if COMPILE_TEST
 	depends on GENERIC_SCHED_CLOCK && RISCV
drivers/clocksource/Makefile +1 -0
index 368c3461dab8..b66ac05ec086 100644
@@ -80,6 +80,7 @@ obj-$(CONFIG_INGENIC_TIMER)		+= ingenic-timer.o
 obj-$(CONFIG_CLKSRC_ST_LPC)		+= clksrc_st_lpc.o
 obj-$(CONFIG_X86_NUMACHIP)		+= numachip.o
 obj-$(CONFIG_RISCV_TIMER)		+= timer-riscv.o
+obj-$(CONFIG_STARFIVE_JH7110_TIMER)	+= timer-jh7110.o
 obj-$(CONFIG_CLINT_TIMER)		+= timer-clint.o
 obj-$(CONFIG_CSKY_MP_TIMER)		+= timer-mp-csky.o
 obj-$(CONFIG_GX6605S_TIMER)		+= timer-gx6605s.o
drivers/clocksource/timer-jh7110.c new file +46 -0
new file mode 100644
index 000000000000..914424368290
@@ -0,0 +1,387 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Starfive JH7110 Timer driver
+ *
+ * Copyright (C) 2022-2023 StarFive Technology Co., Ltd.
+ *
+ * Author:
+ * Xingyu Wu <[email protected]>
+ * Samin Guo <[email protected]>
+ */
+
+#include <linux/clk.h>
+#include <linux/clockchips.h>
+#include <linux/clocksource.h>
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/irq.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+#include <linux/sched_clock.h>
+
+/* Bias: Ch0-0x0, Ch1-0x40, Ch2-0x80, and so on. */
+#define JH7110_TIMER_CH_LEN		0x40
+#define JH7110_TIMER_CH_BASE(x)		((x) * JH7110_TIMER_CH_LEN)
+#define JH7110_TIMER_CH_MAX		4
+
+#define JH7110_CLOCK_SOURCE_RATING	200
+#define JH7110_VALID_BITS		32
+#define JH7110_DELAY_US			0
+#define JH7110_TIMEOUT_US		10000
+#define JH7110_CLOCKEVENT_RATING	300
+#define JH7110_TIMER_MAX_TICKS		0xffffffff
+#define JH7110_TIMER_MIN_TICKS		0xf
+#define JH7110_TIMER_NAME_NUM		19
+
+#define JH7110_TIMER_INT_STATUS		0x00 /* RO[0:4]: Interrupt Status for channel0~4 */
+#define JH7110_TIMER_CTL		0x04 /* RW[0]: 0-continuous run, 1-single run */
+#define JH7110_TIMER_LOAD		0x08 /* RW: load value to counter */
+#define JH7110_TIMER_ENABLE		0x10 /* RW[0]: timer enable register */
+#define JH7110_TIMER_RELOAD		0x14 /* RW: write 1 or 0 both reload counter */

drivers/clocksource/timer-jh7110.c line 56

Quoted 10 lines
drivers/clocksource/timer-jh7110.c new file +10 -0
+#define JH7110_TIMER_VALUE		0x18 /* RO: timer value register */
+#define JH7110_TIMER_INT_CLR		0x20 /* RW: timer interrupt clear register */
+#define JH7110_TIMER_INT_MASK		0x24 /* RW[0]: timer interrupt mask register */
+#define JH7110_TIMER_INT_CLR_AVA_MASK	BIT(1)
+
+struct jh7110_clkevt {
+	struct clock_event_device evt;
+	struct clocksource cs;
+	struct clk *clk;
+	char name[JH7110_TIMER_NAME_NUM];

Instead of this define maybe just use something like

Quoted 2 lines
	char name[sizeof("jh7110-timer.chX")];

Will fix.

Quoted 8 lines
Quoted 5 lines
Quoted 3 lines

drivers/clocksource/timer-jh7110.c line 57

Quoted 1 line
drivers/clocksource/timer-jh7110.c new file +1 -0
+	int irq;

You wrote that you "dropped the extra copy of irq", but here you didn't.

What I mean is to drop the local variable of irq in the probe. Do you mean to use a local variable instead this in the probe?

No, what I mean is that this field is no longer used, so you can delete it.

OK, drop it here.

Quoted 49 lines
Quoted 49 lines
Quoted 49 lines
Quoted 49 lines
drivers/clocksource/timer-jh7110.c new file +48 -0
+
+static int jh7110_timer_probe(struct platform_device *pdev)
+{
+	struct jh7110_clkevt *clkevt[JH7110_TIMER_CH_MAX];
+	char name[4];
+	struct clk *pclk;
+	struct reset_control *rst;
+	int ch;
+	int ret;
+	void __iomem *base
+
+	base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(base))
+		return dev_err_probe(&pdev->dev, PTR_ERR(base),
+				     "failed to map registers\n");
+
+	rst = devm_reset_control_get_exclusive(&pdev->dev, "apb");
+	if (IS_ERR(rst))
+		return dev_err_probe(&pdev->dev, PTR_ERR(rst), "failed to get apb reset\n");
+
+	pclk = devm_clk_get_enabled(&pdev->dev, "apb");
+	if (IS_ERR(pclk))
+		return dev_err_probe(&pdev->dev, PTR_ERR(pclk),
+				     "failed to get & enable apb clock\n");
+
+	ret = reset_control_deassert(rst);
+	if (ret)
+		return dev_err_probe(&pdev->dev, ret, "failed to deassert apb reset\n");
+
+	for (ch = 0; ch < JH7110_TIMER_CH_MAX; ch++) {
+		clkevt[ch] = devm_kzalloc(&pdev->dev, sizeof(*clkevt[ch]), GFP_KERNEL);
+		if (!clkevt[ch])
+			return -ENOMEM;
+
+		snprintf(name, sizeof(name), "ch%d", ch);
+
+		clkevt[ch]->base = base + JH7110_TIMER_CH_BASE(ch);
+		/* Ensure timer is disabled */
+		jh7110_timer_disable(clkevt[ch]);
+
+		rst = devm_reset_control_get_exclusive(&pdev->dev, name);
+		if (IS_ERR(rst))
+			return PTR_ERR(rst);
+
+		clkevt[ch]->clk = devm_clk_get_enabled(&pdev->dev, name);
+		if (IS_ERR(clkevt[ch]->clk))
+			return PTR_ERR(clkevt[ch]->clk);
+
Quoted 8 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

drivers/clocksource/timer-jh7110.c line 351

Quoted 7 lines
drivers/clocksource/timer-jh7110.c new file +7 -0
+		ret = reset_control_deassert(rst);
+		if (ret)
+			return ret;
+
+		clkevt[ch]->evt.irq = platform_get_irq(pdev, ch);
+		if (clkevt[ch]->evt.irq < 0)
+			return clkevt[ch]->evt.irq;

I should use a local variable of irq instead here. Right?

No, I don't think so.

Quoted 38 lines
Quoted 38 lines
Quoted 38 lines
Quoted 38 lines
Quoted 38 lines
drivers/clocksource/timer-jh7110.c new file +36 -1
+
+		snprintf(clkevt[ch]->name, sizeof(clkevt[ch]->name), "%s.ch%d", pdev->name, ch);
+		jh7110_clockevents_register(clkevt[ch]);
+
+		ret = devm_request_irq(&pdev->dev, clkevt[ch]->evt.irq, jh7110_timer_interrupt,
+				       IRQF_TIMER | IRQF_IRQPOLL,
+				       clkevt[ch]->name, &clkevt[ch]->evt);
+		if (ret)
+			return ret;
+
+		ret = jh7110_clocksource_init(clkevt[ch]);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static const struct of_device_id jh7110_timer_match[] = {
+	{ .compatible = "starfive,jh7110-timer", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, jh7110_timer_match);
+
+static struct platform_driver jh7110_timer_driver = {
+	.probe = jh7110_timer_probe,
+	.driver = {
+		.name = "jh7110-timer",
+		.of_match_table = jh7110_timer_match,
+	},
+};
+module_platform_driver(jh7110_timer_driver);
+
+MODULE_AUTHOR("Xingyu Wu <[email protected]>");
+MODULE_DESCRIPTION("StarFive JH7110 timer driver");
+MODULE_LICENSE("GPL");
--

2.25.1

author
#14 lore raw

PATCH v6 3/3 riscv: dts: jh7110: starfive: Add timer node

In reply to Xingyu Wu

Add the timer node for the Starfive JH7110 SoC.

---
 arch/riscv/boot/dts/starfive/jh7110.dtsi | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)
arch/riscv/boot/dts/starfive/jh7110.dtsi +20 -0
index e85464c328d0..d17c8005aaf6 100644
@@ -811,6 +811,26 @@ sysgpio: pinctrl@13040000 {
 			#gpio-cells = <2>;
 		};
 
+		timer@13050000 {
+			compatible = "starfive,jh7110-timer";
+			reg = <0x0 0x13050000 0x0 0x10000>;
+			interrupts = <69>, <70>, <71> ,<72>;
+			clocks = <&syscrg JH7110_SYSCLK_TIMER_APB>,
+				 <&syscrg JH7110_SYSCLK_TIMER0>,
+				 <&syscrg JH7110_SYSCLK_TIMER1>,
+				 <&syscrg JH7110_SYSCLK_TIMER2>,
+				 <&syscrg JH7110_SYSCLK_TIMER3>;
+			clock-names = "apb", "ch0", "ch1",
+				      "ch2", "ch3";
+			resets = <&syscrg JH7110_SYSRST_TIMER_APB>,
+				 <&syscrg JH7110_SYSRST_TIMER0>,
+				 <&syscrg JH7110_SYSRST_TIMER1>,
+				 <&syscrg JH7110_SYSRST_TIMER2>,
+				 <&syscrg JH7110_SYSRST_TIMER3>;
+			reset-names = "apb", "ch0", "ch1",
+				      "ch2", "ch3";
+		};
+
 		watchdog@13070000 {
 			compatible = "starfive,jh7110-wdt";
 			reg = <0x0 0x13070000 0x0 0x10000>;
Signature · 1 line
2.25.1