[PATCH v3 0/2] Add drivers for Intel Keem Bay SoC watchdog

[PATCH v3 0/2] Add drivers for Intel Keem Bay SoC watchdog

[vijayakannan.ayyathurai]

From: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>

Hi,

This patch set adds the watchdog timer driver support for Intel Keem Bay Soc.

Patch 1 holds the driver and Patch 2 holds the Device Tree
binding documentation.

This driver was tested on the Keem Bay evaluation module board.

Thank you,
Vijay

Changes since v2:
 - Remove reduntant write to TIM_WATCHDOG as per Guenter.
 - Rebase the driver with 5.10-rc5 tag.

Changes since v1:
 - Fix indentation error in the dt-bindings file.
 - Use true/false in the second arg of keembay_wdt_set_timeout_reg().
 - Fix the watchdog start sequence.
 - Avoid reduntant timeout register setting.
 - Remove min usage to find actual time at keembay_wdt_set_timeout().
 - Remove timeout configuration boundary check at
   keembay_wdt_set_pretimeout().
 - Use devm_watchdog_register_device() for wdt registration, which
   eventually supports driver unload functionality as well.

Vijayakannan Ayyathurai (2):
  watchdog: Add watchdog driver for Intel Keembay Soc
  dt-bindings: watchdog: Add bindings for Intel Keem Bay SoC

 .../bindings/watchdog/intel,keembay-wdt.yaml  |  57 ++++
 drivers/watchdog/Kconfig                      |  13 +
 drivers/watchdog/Makefile                     |   1 +
 drivers/watchdog/keembay_wdt.c                | 284 ++++++++++++++++++
 4 files changed, 355 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/watchdog/intel,keembay-wdt.yaml
 create mode 100644 drivers/watchdog/keembay_wdt.c


base-commit: 418baf2c28f3473039f2f7377760bd8f6897ae18
prerequisite-patch-id: 822987dcf4c969ef6ac70359b088af06ba39042b
prerequisite-patch-id: 0a348762b660d0d817b8e70cc71647e83173c78c
prerequisite-patch-id: 54c661a006c7362053cb7602448d6c77419d5cf9
prerequisite-patch-id: d140d8534fb828778e0652fe5fcf6282e027f985
-- 
2.17.1

[PATCH v3 1/2] watchdog: Add watchdog driver for Intel Keembay Soc

[vijayakannan.ayyathurai]

From: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>

Intel Keembay Soc requires watchdog timer support.
Add watchdog driver to enable this.

Signed-off-by: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
Acked-by: Mark Gross <mgross@linux.intel.com>
Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 drivers/watchdog/Kconfig       |  13 ++
 drivers/watchdog/Makefile      |   1 +
 drivers/watchdog/keembay_wdt.c | 284 +++++++++++++++++++++++++++++++++
 3 files changed, 298 insertions(+)
 create mode 100644 drivers/watchdog/keembay_wdt.c

diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
index fd7968635e6d..f412cf2d0f1a 100644
--- a/drivers/watchdog/Kconfig
+++ b/drivers/watchdog/Kconfig
@@ -2163,4 +2163,17 @@ config USBPCWATCHDOG
 
 	  Most people will say N.
 
+config KEEMBAY_WATCHDOG
+	tristate "Intel Keem Bay SoC non-secure watchdog"
+	depends on ARCH_KEEMBAY || (ARM64 && COMPILE_TEST)
+	select WATCHDOG_CORE
+	help
+	 This option enable support for an In-secure watchdog timer driver for
+	 Intel Keem Bay SoC. This WDT has a 32 bit timer and decrements in every
+	 count unit. An interrupt will be triggered, when the count crosses
+	 the thershold configured in the register.
+
+	 To compile this driver as a module, choose M here: the
+	 module will be called keembay_wdt.
+
 endif # WATCHDOG
diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
index 071a2e50be98..f6f9f434f407 100644
--- a/drivers/watchdog/Makefile
+++ b/drivers/watchdog/Makefile
@@ -146,6 +146,7 @@ obj-$(CONFIG_INTEL_MEI_WDT) += mei_wdt.o
 obj-$(CONFIG_NI903X_WDT) += ni903x_wdt.o
 obj-$(CONFIG_NIC7018_WDT) += nic7018_wdt.o
 obj-$(CONFIG_MLX_WDT) += mlx_wdt.o
+obj-$(CONFIG_KEEMBAY_WATCHDOG) += keembay_wdt.o
 
 # M68K Architecture
 obj-$(CONFIG_M54xx_WATCHDOG) += m54xx_wdt.o
diff --git a/drivers/watchdog/keembay_wdt.c b/drivers/watchdog/keembay_wdt.c
new file mode 100644
index 000000000000..1e1a1447c729
--- /dev/null
+++ b/drivers/watchdog/keembay_wdt.c
@@ -0,0 +1,284 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Watchdog driver for Intel Keem Bay non-secure watchdog.
+ *
+ * Copyright (C) 2020 Intel Corporation
+ */
+
+#include <linux/arm-smccc.h>
+#include <linux/bits.h>
+#include <linux/clk.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/limits.h>
+#include <linux/module.h>
+#include <linux/mod_devicetable.h>
+#include <linux/platform_device.h>
+#include <linux/reboot.h>
+#include <linux/watchdog.h>
+
+/* Non-secure watchdog register offsets */
+#define TIM_WATCHDOG		0x0
+#define TIM_WATCHDOG_INT_THRES	0x4
+#define TIM_WDOG_EN		0x8
+#define TIM_SAFE		0xc
+
+#define WDT_ISR_MASK		GENMASK(9, 8)
+#define WDT_ISR_CLEAR		0x8200ff18
+#define WDT_UNLOCK		0xf1d0dead
+#define WDT_LOAD_MAX		U32_MAX
+#define WDT_LOAD_MIN		1
+#define WDT_TIMEOUT		5
+
+static unsigned int timeout = WDT_TIMEOUT;
+module_param(timeout, int, 0);
+MODULE_PARM_DESC(timeout, "Watchdog timeout period in seconds (default = "
+		 __MODULE_STRING(WDT_TIMEOUT) ")");
+
+static bool nowayout = WATCHDOG_NOWAYOUT;
+module_param(nowayout, bool, 0);
+MODULE_PARM_DESC(nowayout, "Watchdog cannot be stopped once started (default = "
+		 __MODULE_STRING(WATCHDOG_NOWAYOUT) ")");
+
+struct keembay_wdt {
+	struct watchdog_device	wdd;
+	struct clk		*clk;
+	unsigned int		rate;
+	int			to_irq;
+	int			th_irq;
+	void __iomem		*base;
+};
+
+static inline u32 keembay_wdt_readl(struct keembay_wdt *wdt, u32 offset)
+{
+	return readl(wdt->base + offset);
+}
+
+static inline void keembay_wdt_writel(struct keembay_wdt *wdt,
+				      u32 offset, u32 val)
+{
+	writel(WDT_UNLOCK, wdt->base + TIM_SAFE);
+	writel(val, wdt->base + offset);
+}
+
+static void keembay_wdt_set_timeout_reg(struct watchdog_device *wdog, bool ping)
+{
+	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
+	u32 th_val = 0;
+
+	if (!ping && wdog->pretimeout) {
+		th_val = wdog->timeout - wdog->pretimeout;
+		keembay_wdt_writel(wdt, TIM_WATCHDOG_INT_THRES, th_val * wdt->rate);
+	}
+
+	keembay_wdt_writel(wdt, TIM_WATCHDOG, wdog->timeout * wdt->rate);
+}
+
+static int keembay_wdt_start(struct watchdog_device *wdog)
+{
+	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
+
+	keembay_wdt_set_timeout_reg(wdog, false);
+	keembay_wdt_writel(wdt, TIM_WDOG_EN, 1);
+
+	return 0;
+}
+
+static int keembay_wdt_stop(struct watchdog_device *wdog)
+{
+	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
+
+	keembay_wdt_writel(wdt, TIM_WDOG_EN, 0);
+
+	return 0;
+}
+
+static int keembay_wdt_ping(struct watchdog_device *wdog)
+{
+	keembay_wdt_set_timeout_reg(wdog, true);
+
+	return 0;
+}
+
+static int keembay_wdt_set_timeout(struct watchdog_device *wdog, u32 t)
+{
+	wdog->timeout = t;
+	keembay_wdt_set_timeout_reg(wdog, false);
+
+	return 0;
+}
+
+static int keembay_wdt_set_pretimeout(struct watchdog_device *wdog, u32 t)
+{
+	wdog->pretimeout = t;
+	keembay_wdt_set_timeout_reg(wdog, false);
+
+	return 0;
+}
+
+static unsigned int keembay_wdt_get_timeleft(struct watchdog_device *wdog)
+{
+	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
+
+	return keembay_wdt_readl(wdt, TIM_WATCHDOG) / wdt->rate;
+}
+
+/*
+ * SMC call is used to clear the interrupt bits, because the TIM_GEN_CONFIG
+ * register is in the secure bank.
+ */
+static irqreturn_t keembay_wdt_to_isr(int irq, void *dev_id)
+{
+	struct keembay_wdt *wdt = dev_id;
+	struct arm_smccc_res res;
+
+	keembay_wdt_writel(wdt, TIM_WATCHDOG, 1);
+	arm_smccc_smc(WDT_ISR_CLEAR, WDT_ISR_MASK, 0, 0, 0, 0, 0, 0, &res);
+	dev_crit(wdt->wdd.parent, "Intel Keem Bay non-sec wdt timeout.\n");
+	emergency_restart();
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t keembay_wdt_th_isr(int irq, void *dev_id)
+{
+	struct keembay_wdt *wdt = dev_id;
+	struct arm_smccc_res res;
+	u32 th_val = 0;
+
+	if (wdt->wdd.pretimeout)
+		th_val = wdt->wdd.timeout - wdt->wdd.pretimeout;
+
+	keembay_wdt_writel(wdt, TIM_WATCHDOG, th_val * wdt->rate + 1);
+	arm_smccc_smc(WDT_ISR_CLEAR, WDT_ISR_MASK, 0, 0, 0, 0, 0, 0, &res);
+	dev_crit(wdt->wdd.parent, "Intel Keem Bay non-sec wdt pre-timeout.\n");
+	watchdog_notify_pretimeout(&wdt->wdd);
+
+	return IRQ_HANDLED;
+}
+
+static const struct watchdog_info keembay_wdt_info = {
+	.identity	= "Intel Keem Bay Watchdog Timer",
+	.options	= WDIOF_SETTIMEOUT |
+			  WDIOF_PRETIMEOUT |
+			  WDIOF_MAGICCLOSE |
+			  WDIOF_KEEPALIVEPING,
+};
+
+static const struct watchdog_ops keembay_wdt_ops = {
+	.owner		= THIS_MODULE,
+	.start		= keembay_wdt_start,
+	.stop		= keembay_wdt_stop,
+	.ping		= keembay_wdt_ping,
+	.set_timeout	= keembay_wdt_set_timeout,
+	.set_pretimeout	= keembay_wdt_set_pretimeout,
+	.get_timeleft	= keembay_wdt_get_timeleft,
+};
+
+static int keembay_wdt_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct keembay_wdt *wdt;
+	int ret;
+
+	wdt = devm_kzalloc(dev, sizeof(*wdt), GFP_KERNEL);
+	if (!wdt)
+		return -ENOMEM;
+
+	wdt->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(wdt->base))
+		return PTR_ERR(wdt->base);
+
+	/* we do not need to enable the clock as it is enabled by default */
+	wdt->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(wdt->clk))
+		return dev_err_probe(dev, PTR_ERR(wdt->clk), "Failed to get clock\n");
+
+	wdt->rate = clk_get_rate(wdt->clk);
+	if (!wdt->rate)
+		return dev_err_probe(dev, -EINVAL, "Failed to get clock rate\n");
+
+	wdt->th_irq = platform_get_irq_byname(pdev, "threshold");
+	if (wdt->th_irq < 0)
+		return dev_err_probe(dev, wdt->th_irq, "Failed to get IRQ for threshold\n");
+
+	ret = devm_request_irq(dev, wdt->th_irq, keembay_wdt_th_isr, 0,
+			       "keembay-wdt", wdt);
+	if (ret)
+		return dev_err_probe(dev, ret, "Failed to request IRQ for threshold\n");
+
+	wdt->to_irq = platform_get_irq_byname(pdev, "timeout");
+	if (wdt->to_irq < 0)
+		return dev_err_probe(dev, wdt->to_irq, "Failed to get IRQ for timeout\n");
+
+	ret = devm_request_irq(dev, wdt->to_irq, keembay_wdt_to_isr, 0,
+			       "keembay-wdt", wdt);
+	if (ret)
+		return dev_err_probe(dev, ret, "Failed to request IRQ for timeout\n");
+
+	wdt->wdd.parent		= dev;
+	wdt->wdd.info		= &keembay_wdt_info;
+	wdt->wdd.ops		= &keembay_wdt_ops;
+	wdt->wdd.min_timeout	= WDT_LOAD_MIN;
+	wdt->wdd.max_timeout	= WDT_LOAD_MAX / wdt->rate;
+	wdt->wdd.timeout	= WDT_TIMEOUT;
+
+	watchdog_set_drvdata(&wdt->wdd, wdt);
+	watchdog_set_nowayout(&wdt->wdd, nowayout);
+	watchdog_init_timeout(&wdt->wdd, timeout, dev);
+	keembay_wdt_set_timeout(&wdt->wdd, wdt->wdd.timeout);
+
+	ret = devm_watchdog_register_device(dev, &wdt->wdd);
+	if (ret)
+		return dev_err_probe(dev, ret, "Failed to register watchdog device.\n");
+
+	platform_set_drvdata(pdev, wdt);
+	dev_info(dev, "Initial timeout %d sec%s.\n",
+		 wdt->wdd.timeout, nowayout ? ", nowayout" : "");
+
+	return 0;
+}
+
+static int __maybe_unused keembay_wdt_suspend(struct device *dev)
+{
+	struct keembay_wdt *wdt = dev_get_drvdata(dev);
+
+	if (watchdog_active(&wdt->wdd))
+		return keembay_wdt_stop(&wdt->wdd);
+
+	return 0;
+}
+
+static int __maybe_unused keembay_wdt_resume(struct device *dev)
+{
+	struct keembay_wdt *wdt = dev_get_drvdata(dev);
+
+	if (watchdog_active(&wdt->wdd))
+		return keembay_wdt_start(&wdt->wdd);
+
+	return 0;
+}
+
+static SIMPLE_DEV_PM_OPS(keembay_wdt_pm_ops, keembay_wdt_suspend,
+			 keembay_wdt_resume);
+
+static const struct of_device_id keembay_wdt_match[] = {
+	{ .compatible = "intel,keembay-wdt" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, keembay_wdt_match);
+
+static struct platform_driver keembay_wdt_driver = {
+	.probe		= keembay_wdt_probe,
+	.driver		= {
+		.name		= "keembay_wdt",
+		.of_match_table	= keembay_wdt_match,
+		.pm		= &keembay_wdt_pm_ops,
+	},
+};
+
+module_platform_driver(keembay_wdt_driver);
+
+MODULE_DESCRIPTION("Intel Keem Bay SoC watchdog driver");
+MODULE_AUTHOR("Wan Ahmad Zainie <wan.ahmad.zainie.wan.mohamad@intel.com");
+MODULE_LICENSE("GPL v2");
-- 
2.17.1

[PATCH v3 2/2] dt-bindings: watchdog: Add bindings for Intel Keem Bay SoC

[vijayakannan.ayyathurai]

From: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>

Add Device Tree binding document for Watchdog IP in the Intel Keem Bay SoC.

Signed-off-by: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
Acked-by: Mark Gross <mgross@linux.intel.com>
Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 .../bindings/watchdog/intel,keembay-wdt.yaml  | 57 +++++++++++++++++++
 1 file changed, 57 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/watchdog/intel,keembay-wdt.yaml

diff --git a/Documentation/devicetree/bindings/watchdog/intel,keembay-wdt.yaml b/Documentation/devicetree/bindings/watchdog/intel,keembay-wdt.yaml
new file mode 100644
index 000000000000..1437ff8a122f
--- /dev/null
+++ b/Documentation/devicetree/bindings/watchdog/intel,keembay-wdt.yaml
@@ -0,0 +1,57 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/watchdog/intel,keembay-wdt.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Intel Keem Bay SoC non-secure Watchdog Timer
+
+maintainers:
+  - Wan Ahmad Zainie <wan.ahmad.zainie.wan.mohamad@intel.com>
+
+properties:
+  compatible:
+    enum:
+      - intel,keembay-wdt
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    maxItems: 1
+
+  interrupts:
+    items:
+      - description: interrupt specifier for threshold interrupt line
+      - description: interrupt specifier for timeout interrupt line
+
+  interrupt-names:
+    items:
+      - const: threshold
+      - const: timeout
+
+required:
+  - compatible
+  - reg
+  - interrupts
+  - interrupt-names
+  - clocks
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/interrupt-controller/arm-gic.h>
+    #include <dt-bindings/interrupt-controller/irq.h>
+    #define KEEM_BAY_A53_TIM
+
+    watchdog: watchdog@2033009c {
+        compatible = "intel,keembay-wdt";
+        reg = <0x2033009c 0x10>;
+        interrupts = <GIC_SPI 1 IRQ_TYPE_LEVEL_HIGH>,
+                     <GIC_SPI 2 IRQ_TYPE_LEVEL_HIGH>;
+        interrupt-names = "threshold", "timeout";
+        clocks = <&scmi_clk KEEM_BAY_A53_TIM>;
+    };
+
+...
-- 
2.17.1

RE: [PATCH v3 2/2] dt-bindings: watchdog: Add bindings for Intel Keem Bay SoC

[Ayyathurai, Vijayakannan]

Hi,

> From: Ayyathurai, Vijayakannan <vijayakannan.ayyathurai@intel.com>
> 
> Add Device Tree binding document for Watchdog IP in the Intel Keem Bay SoC.
> 
> Signed-off-by: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
> Acked-by: Mark Gross <mgross@linux.intel.com>
> Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Sorry, by mistake I missed adding Rob's Reviewed-by: tag.

> ---
>  .../bindings/watchdog/intel,keembay-wdt.yaml  | 57 +++++++++++++++++++
>  1 file changed, 57 insertions(+)
>  create mode 100644
> Documentation/devicetree/bindings/watchdog/intel,keembay-wdt.yaml
>
Thanks,
Vijay

Re: [PATCH v3 1/2] watchdog: Add watchdog driver for Intel Keembay Soc

[Guenter Roeck]

On Tue, Dec 01, 2020 at 11:10:33PM +0800, vijayakannan.ayyathurai@intel.com wrote:
> From: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
> 
> Intel Keembay Soc requires watchdog timer support.
> Add watchdog driver to enable this.
> 
> Signed-off-by: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
> Acked-by: Mark Gross <mgross@linux.intel.com>
> Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---
>  drivers/watchdog/Kconfig       |  13 ++
>  drivers/watchdog/Makefile      |   1 +
>  drivers/watchdog/keembay_wdt.c | 284 +++++++++++++++++++++++++++++++++
>  3 files changed, 298 insertions(+)
>  create mode 100644 drivers/watchdog/keembay_wdt.c
> 
> diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
> index fd7968635e6d..f412cf2d0f1a 100644
> --- a/drivers/watchdog/Kconfig
> +++ b/drivers/watchdog/Kconfig
> @@ -2163,4 +2163,17 @@ config USBPCWATCHDOG
>  
>  	  Most people will say N.
>  
> +config KEEMBAY_WATCHDOG
> +	tristate "Intel Keem Bay SoC non-secure watchdog"
> +	depends on ARCH_KEEMBAY || (ARM64 && COMPILE_TEST)
> +	select WATCHDOG_CORE
> +	help
> +	 This option enable support for an In-secure watchdog timer driver for
> +	 Intel Keem Bay SoC. This WDT has a 32 bit timer and decrements in every
> +	 count unit. An interrupt will be triggered, when the count crosses
> +	 the thershold configured in the register.
> +
> +	 To compile this driver as a module, choose M here: the
> +	 module will be called keembay_wdt.
> +
>  endif # WATCHDOG
> diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
> index 071a2e50be98..f6f9f434f407 100644
> --- a/drivers/watchdog/Makefile
> +++ b/drivers/watchdog/Makefile
> @@ -146,6 +146,7 @@ obj-$(CONFIG_INTEL_MEI_WDT) += mei_wdt.o
>  obj-$(CONFIG_NI903X_WDT) += ni903x_wdt.o
>  obj-$(CONFIG_NIC7018_WDT) += nic7018_wdt.o
>  obj-$(CONFIG_MLX_WDT) += mlx_wdt.o
> +obj-$(CONFIG_KEEMBAY_WATCHDOG) += keembay_wdt.o
>  
>  # M68K Architecture
>  obj-$(CONFIG_M54xx_WATCHDOG) += m54xx_wdt.o
> diff --git a/drivers/watchdog/keembay_wdt.c b/drivers/watchdog/keembay_wdt.c
> new file mode 100644
> index 000000000000..1e1a1447c729
> --- /dev/null
> +++ b/drivers/watchdog/keembay_wdt.c
> @@ -0,0 +1,284 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Watchdog driver for Intel Keem Bay non-secure watchdog.
> + *
> + * Copyright (C) 2020 Intel Corporation
> + */
> +
> +#include <linux/arm-smccc.h>
> +#include <linux/bits.h>
> +#include <linux/clk.h>
> +#include <linux/interrupt.h>
> +#include <linux/io.h>
> +#include <linux/limits.h>
> +#include <linux/module.h>
> +#include <linux/mod_devicetable.h>
> +#include <linux/platform_device.h>
> +#include <linux/reboot.h>
> +#include <linux/watchdog.h>
> +
> +/* Non-secure watchdog register offsets */
> +#define TIM_WATCHDOG		0x0
> +#define TIM_WATCHDOG_INT_THRES	0x4
> +#define TIM_WDOG_EN		0x8
> +#define TIM_SAFE		0xc
> +
> +#define WDT_ISR_MASK		GENMASK(9, 8)
> +#define WDT_ISR_CLEAR		0x8200ff18
> +#define WDT_UNLOCK		0xf1d0dead
> +#define WDT_LOAD_MAX		U32_MAX
> +#define WDT_LOAD_MIN		1
> +#define WDT_TIMEOUT		5
> +
> +static unsigned int timeout = WDT_TIMEOUT;
> +module_param(timeout, int, 0);
> +MODULE_PARM_DESC(timeout, "Watchdog timeout period in seconds (default = "
> +		 __MODULE_STRING(WDT_TIMEOUT) ")");
> +
> +static bool nowayout = WATCHDOG_NOWAYOUT;
> +module_param(nowayout, bool, 0);
> +MODULE_PARM_DESC(nowayout, "Watchdog cannot be stopped once started (default = "
> +		 __MODULE_STRING(WATCHDOG_NOWAYOUT) ")");
> +
> +struct keembay_wdt {
> +	struct watchdog_device	wdd;
> +	struct clk		*clk;
> +	unsigned int		rate;
> +	int			to_irq;
> +	int			th_irq;
> +	void __iomem		*base;
> +};
> +
> +static inline u32 keembay_wdt_readl(struct keembay_wdt *wdt, u32 offset)
> +{
> +	return readl(wdt->base + offset);
> +}
> +
> +static inline void keembay_wdt_writel(struct keembay_wdt *wdt,
> +				      u32 offset, u32 val)
> +{
> +	writel(WDT_UNLOCK, wdt->base + TIM_SAFE);
> +	writel(val, wdt->base + offset);
> +}
> +
> +static void keembay_wdt_set_timeout_reg(struct watchdog_device *wdog, bool ping)
> +{
> +	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
> +	u32 th_val = 0;
> +
> +	if (!ping && wdog->pretimeout) {
> +		th_val = wdog->timeout - wdog->pretimeout;
> +		keembay_wdt_writel(wdt, TIM_WATCHDOG_INT_THRES, th_val * wdt->rate);

Sorry for annoying you now, but I may have found another potential problem.

What happens if the user sets a pretimeout, then removes it ?
What should TIM_WATCHDOG_INT_THRES be set to in that case ?
Right now TIM_WATCHDOG_INT_THRES won't be updated anymore
in that case, which seems wrong. This might get worse with
the following sequence.

- set pretimeout
- clear pretimeout
- set timeout to some other value

Thanks,
Guenter

> +	}
> +
> +	keembay_wdt_writel(wdt, TIM_WATCHDOG, wdog->timeout * wdt->rate);
> +}
> +
> +static int keembay_wdt_start(struct watchdog_device *wdog)
> +{
> +	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
> +
> +	keembay_wdt_set_timeout_reg(wdog, false);
> +	keembay_wdt_writel(wdt, TIM_WDOG_EN, 1);
> +
> +	return 0;
> +}
> +
> +static int keembay_wdt_stop(struct watchdog_device *wdog)
> +{
> +	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
> +
> +	keembay_wdt_writel(wdt, TIM_WDOG_EN, 0);
> +
> +	return 0;
> +}
> +
> +static int keembay_wdt_ping(struct watchdog_device *wdog)
> +{
> +	keembay_wdt_set_timeout_reg(wdog, true);
> +
> +	return 0;
> +}
> +
> +static int keembay_wdt_set_timeout(struct watchdog_device *wdog, u32 t)
> +{
> +	wdog->timeout = t;
> +	keembay_wdt_set_timeout_reg(wdog, false);
> +
> +	return 0;
> +}
> +
> +static int keembay_wdt_set_pretimeout(struct watchdog_device *wdog, u32 t)
> +{
> +	wdog->pretimeout = t;
> +	keembay_wdt_set_timeout_reg(wdog, false);
> +
> +	return 0;
> +}
> +
> +static unsigned int keembay_wdt_get_timeleft(struct watchdog_device *wdog)
> +{
> +	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
> +
> +	return keembay_wdt_readl(wdt, TIM_WATCHDOG) / wdt->rate;
> +}
> +
> +/*
> + * SMC call is used to clear the interrupt bits, because the TIM_GEN_CONFIG
> + * register is in the secure bank.
> + */
> +static irqreturn_t keembay_wdt_to_isr(int irq, void *dev_id)
> +{
> +	struct keembay_wdt *wdt = dev_id;
> +	struct arm_smccc_res res;
> +
> +	keembay_wdt_writel(wdt, TIM_WATCHDOG, 1);
> +	arm_smccc_smc(WDT_ISR_CLEAR, WDT_ISR_MASK, 0, 0, 0, 0, 0, 0, &res);
> +	dev_crit(wdt->wdd.parent, "Intel Keem Bay non-sec wdt timeout.\n");
> +	emergency_restart();
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static irqreturn_t keembay_wdt_th_isr(int irq, void *dev_id)
> +{
> +	struct keembay_wdt *wdt = dev_id;
> +	struct arm_smccc_res res;
> +	u32 th_val = 0;
> +
> +	if (wdt->wdd.pretimeout)
> +		th_val = wdt->wdd.timeout - wdt->wdd.pretimeout;
> +
> +	keembay_wdt_writel(wdt, TIM_WATCHDOG, th_val * wdt->rate + 1);
> +	arm_smccc_smc(WDT_ISR_CLEAR, WDT_ISR_MASK, 0, 0, 0, 0, 0, 0, &res);
> +	dev_crit(wdt->wdd.parent, "Intel Keem Bay non-sec wdt pre-timeout.\n");
> +	watchdog_notify_pretimeout(&wdt->wdd);
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static const struct watchdog_info keembay_wdt_info = {
> +	.identity	= "Intel Keem Bay Watchdog Timer",
> +	.options	= WDIOF_SETTIMEOUT |
> +			  WDIOF_PRETIMEOUT |
> +			  WDIOF_MAGICCLOSE |
> +			  WDIOF_KEEPALIVEPING,
> +};
> +
> +static const struct watchdog_ops keembay_wdt_ops = {
> +	.owner		= THIS_MODULE,
> +	.start		= keembay_wdt_start,
> +	.stop		= keembay_wdt_stop,
> +	.ping		= keembay_wdt_ping,
> +	.set_timeout	= keembay_wdt_set_timeout,
> +	.set_pretimeout	= keembay_wdt_set_pretimeout,
> +	.get_timeleft	= keembay_wdt_get_timeleft,
> +};
> +
> +static int keembay_wdt_probe(struct platform_device *pdev)
> +{
> +	struct device *dev = &pdev->dev;
> +	struct keembay_wdt *wdt;
> +	int ret;
> +
> +	wdt = devm_kzalloc(dev, sizeof(*wdt), GFP_KERNEL);
> +	if (!wdt)
> +		return -ENOMEM;
> +
> +	wdt->base = devm_platform_ioremap_resource(pdev, 0);
> +	if (IS_ERR(wdt->base))
> +		return PTR_ERR(wdt->base);
> +
> +	/* we do not need to enable the clock as it is enabled by default */
> +	wdt->clk = devm_clk_get(dev, NULL);
> +	if (IS_ERR(wdt->clk))
> +		return dev_err_probe(dev, PTR_ERR(wdt->clk), "Failed to get clock\n");
> +
> +	wdt->rate = clk_get_rate(wdt->clk);
> +	if (!wdt->rate)
> +		return dev_err_probe(dev, -EINVAL, "Failed to get clock rate\n");
> +
> +	wdt->th_irq = platform_get_irq_byname(pdev, "threshold");
> +	if (wdt->th_irq < 0)
> +		return dev_err_probe(dev, wdt->th_irq, "Failed to get IRQ for threshold\n");
> +
> +	ret = devm_request_irq(dev, wdt->th_irq, keembay_wdt_th_isr, 0,
> +			       "keembay-wdt", wdt);
> +	if (ret)
> +		return dev_err_probe(dev, ret, "Failed to request IRQ for threshold\n");
> +
> +	wdt->to_irq = platform_get_irq_byname(pdev, "timeout");
> +	if (wdt->to_irq < 0)
> +		return dev_err_probe(dev, wdt->to_irq, "Failed to get IRQ for timeout\n");
> +
> +	ret = devm_request_irq(dev, wdt->to_irq, keembay_wdt_to_isr, 0,
> +			       "keembay-wdt", wdt);
> +	if (ret)
> +		return dev_err_probe(dev, ret, "Failed to request IRQ for timeout\n");
> +
> +	wdt->wdd.parent		= dev;
> +	wdt->wdd.info		= &keembay_wdt_info;
> +	wdt->wdd.ops		= &keembay_wdt_ops;
> +	wdt->wdd.min_timeout	= WDT_LOAD_MIN;
> +	wdt->wdd.max_timeout	= WDT_LOAD_MAX / wdt->rate;
> +	wdt->wdd.timeout	= WDT_TIMEOUT;
> +
> +	watchdog_set_drvdata(&wdt->wdd, wdt);
> +	watchdog_set_nowayout(&wdt->wdd, nowayout);
> +	watchdog_init_timeout(&wdt->wdd, timeout, dev);
> +	keembay_wdt_set_timeout(&wdt->wdd, wdt->wdd.timeout);
> +
> +	ret = devm_watchdog_register_device(dev, &wdt->wdd);
> +	if (ret)
> +		return dev_err_probe(dev, ret, "Failed to register watchdog device.\n");
> +
> +	platform_set_drvdata(pdev, wdt);
> +	dev_info(dev, "Initial timeout %d sec%s.\n",
> +		 wdt->wdd.timeout, nowayout ? ", nowayout" : "");
> +
> +	return 0;
> +}
> +
> +static int __maybe_unused keembay_wdt_suspend(struct device *dev)
> +{
> +	struct keembay_wdt *wdt = dev_get_drvdata(dev);
> +
> +	if (watchdog_active(&wdt->wdd))
> +		return keembay_wdt_stop(&wdt->wdd);
> +
> +	return 0;
> +}
> +
> +static int __maybe_unused keembay_wdt_resume(struct device *dev)
> +{
> +	struct keembay_wdt *wdt = dev_get_drvdata(dev);
> +
> +	if (watchdog_active(&wdt->wdd))
> +		return keembay_wdt_start(&wdt->wdd);
> +
> +	return 0;
> +}
> +
> +static SIMPLE_DEV_PM_OPS(keembay_wdt_pm_ops, keembay_wdt_suspend,
> +			 keembay_wdt_resume);
> +
> +static const struct of_device_id keembay_wdt_match[] = {
> +	{ .compatible = "intel,keembay-wdt" },
> +	{ }
> +};
> +MODULE_DEVICE_TABLE(of, keembay_wdt_match);
> +
> +static struct platform_driver keembay_wdt_driver = {
> +	.probe		= keembay_wdt_probe,
> +	.driver		= {
> +		.name		= "keembay_wdt",
> +		.of_match_table	= keembay_wdt_match,
> +		.pm		= &keembay_wdt_pm_ops,
> +	},
> +};
> +
> +module_platform_driver(keembay_wdt_driver);
> +
> +MODULE_DESCRIPTION("Intel Keem Bay SoC watchdog driver");
> +MODULE_AUTHOR("Wan Ahmad Zainie <wan.ahmad.zainie.wan.mohamad@intel.com");
> +MODULE_LICENSE("GPL v2");
> -- 
> 2.17.1
> 

RE: [PATCH v3 1/2] watchdog: Add watchdog driver for Intel Keembay Soc

[Ayyathurai, Vijayakannan]

Hi Guenter,

> From: Guenter Roeck <linux@roeck-us.net>
> Sent: Wednesday, 2 December, 2020 12:31 AM
> 
> On Tue, Dec 01, 2020 at 11:10:33PM +0800,
> vijayakannan.ayyathurai@intel.com wrote:
> > From: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
> >
> > Intel Keembay Soc requires watchdog timer support.
> > Add watchdog driver to enable this.
> >
> > +static void keembay_wdt_set_timeout_reg(struct watchdog_device *wdog,
> bool ping)
> > +{
> > +	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
> > +	u32 th_val = 0;
> > +
> > +	if (!ping && wdog->pretimeout) {
> > +		th_val = wdog->timeout - wdog->pretimeout;
> > +		keembay_wdt_writel(wdt, TIM_WATCHDOG_INT_THRES, th_val
> * wdt->rate);
> 
> Sorry for annoying you now, but I may have found another potential problem.
> 
> What happens if the user sets a pretimeout, then removes it ?
> What should TIM_WATCHDOG_INT_THRES be set to in that case ?
> Right now TIM_WATCHDOG_INT_THRES won't be updated anymore

It is a good catch. Indeed, I don't have coverage like this.

> in that case, which seems wrong. This might get worse with
> the following sequence.
> 
> - set pretimeout
> - clear pretimeout
> - set timeout to some other value
> 

Can the below method resolve this issue?


static int keembay_wdt_set_pretimeout(struct watchdog_device *wdog, u32 t)
{
        struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);

        if(!t)
                keembay_wdt_writel(wdt, TIM_WATCHDOG_INT_THRES, 0);

        wdog->pretimeout = t;
        keembay_wdt_set_timeout_reg(wdog, false);

        return 0;
}

> Thanks,
> Guenter
> 

Thanks,
Vijay

Re: [PATCH v3 1/2] watchdog: Add watchdog driver for Intel Keembay Soc

[Guenter Roeck]

On Wed, Dec 02, 2020 at 02:55:32PM +0000, Ayyathurai, Vijayakannan wrote:
> Hi Guenter,
> 
> > From: Guenter Roeck <linux@roeck-us.net>
> > Sent: Wednesday, 2 December, 2020 12:31 AM
> > 
> > On Tue, Dec 01, 2020 at 11:10:33PM +0800,
> > vijayakannan.ayyathurai@intel.com wrote:
> > > From: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
> > >
> > > Intel Keembay Soc requires watchdog timer support.
> > > Add watchdog driver to enable this.
> > >
> > > +static void keembay_wdt_set_timeout_reg(struct watchdog_device *wdog,
> > bool ping)
> > > +{
> > > +	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
> > > +	u32 th_val = 0;
> > > +
> > > +	if (!ping && wdog->pretimeout) {
> > > +		th_val = wdog->timeout - wdog->pretimeout;
> > > +		keembay_wdt_writel(wdt, TIM_WATCHDOG_INT_THRES, th_val
> > * wdt->rate);
> > 
> > Sorry for annoying you now, but I may have found another potential problem.
> > 
> > What happens if the user sets a pretimeout, then removes it ?
> > What should TIM_WATCHDOG_INT_THRES be set to in that case ?
> > Right now TIM_WATCHDOG_INT_THRES won't be updated anymore
> 
> It is a good catch. Indeed, I don't have coverage like this.
> 
> > in that case, which seems wrong. This might get worse with
> > the following sequence.
> > 
> > - set pretimeout
> > - clear pretimeout
> > - set timeout to some other value
> > 
> 
> Can the below method resolve this issue?
> 
> 
> static int keembay_wdt_set_pretimeout(struct watchdog_device *wdog, u32 t)
> {
>         struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
> 
>         if(!t)
>                 keembay_wdt_writel(wdt, TIM_WATCHDOG_INT_THRES, 0);
> 

Partially, only it makes for awkward code. After all, it is never really
necessary to set the timeout register after updating the pretimeout.
The "ping" parameter makes less and less sense with this in mind.
It might be better to split the set_timeout_reg function into
set_timeout_reg and set_pretimeout_reg and call those functions as needed
(and handle the if() above in the set_pretimeout_reg function).

Thanks,
Guenter

>         wdog->pretimeout = t;
>         keembay_wdt_set_timeout_reg(wdog, false);
> 
>         return 0;
> }
> 
> > Thanks,
> > Guenter
> > 
> 
> Thanks,
> Vijay

RE: [PATCH v3 1/2] watchdog: Add watchdog driver for Intel Keembay Soc

[Ayyathurai, Vijayakannan]

Hi Guenter,

> From: Guenter Roeck <linux@roeck-us.net>
> 
> On Wed, Dec 02, 2020 at 02:55:32PM +0000, Ayyathurai, Vijayakannan wrote:
> >
> > > From: Guenter Roeck <linux@roeck-us.net>
> > > Sent: Wednesday, 2 December, 2020 12:31 AM
> > >
> > > On Tue, Dec 01, 2020 at 11:10:33PM +0800,
> > > > From: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
> > > >
> > > >
> > > > +static void keembay_wdt_set_timeout_reg(struct watchdog_device
> *wdog,
> > > bool ping)
> > > > +{
> > > > +	struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
> > > > +	u32 th_val = 0;
> > > > +
> > > > +	if (!ping && wdog->pretimeout) {
> > > > +		th_val = wdog->timeout - wdog->pretimeout;
> > > > +		keembay_wdt_writel(wdt, TIM_WATCHDOG_INT_THRES, th_val
> > > * wdt->rate);
> > >
> > > Sorry for annoying you now, but I may have found another potential
> problem.
> > >
> > > What happens if the user sets a pretimeout, then removes it ?
> > > What should TIM_WATCHDOG_INT_THRES be set to in that case ?
> > > Right now TIM_WATCHDOG_INT_THRES won't be updated anymore
> >
> > It is a good catch. Indeed, I don't have coverage like this.
> >
> > > in that case, which seems wrong. This might get worse with
> > > the following sequence.
> > >
> > > - set pretimeout
> > > - clear pretimeout
> > > - set timeout to some other value
> > >
> >
> > Can the below method resolve this issue?
> >
> >
> > static int keembay_wdt_set_pretimeout(struct watchdog_device *wdog, u32
> t)
> > {
> >         struct keembay_wdt *wdt = watchdog_get_drvdata(wdog);
> >
> >         if(!t)
> >                 keembay_wdt_writel(wdt, TIM_WATCHDOG_INT_THRES, 0);
> >
> 
> Partially, only it makes for awkward code. After all, it is never really
> necessary to set the timeout register after updating the pretimeout.
> The "ping" parameter makes less and less sense with this in mind.

Yes.

> It might be better to split the set_timeout_reg function into
> set_timeout_reg and set_pretimeout_reg and call those functions as needed
> (and handle the if() above in the set_pretimeout_reg function).
> 

Ok. Thanks for your suggestion.
Let me incorporate necessary changes in the next version.

> Thanks,
> Guenter
>

Thanks,
Vijay

Re: [PATCH v3 2/2] dt-bindings: watchdog: Add bindings for Intel Keem Bay SoC

[Rob Herring]

On Tue, 01 Dec 2020 23:10:34 +0800, vijayakannan.ayyathurai@intel.com wrote:
> From: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
> 
> Add Device Tree binding document for Watchdog IP in the Intel Keem Bay SoC.
> 
> Signed-off-by: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
> Acked-by: Mark Gross <mgross@linux.intel.com>
> Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---
>  .../bindings/watchdog/intel,keembay-wdt.yaml  | 57 +++++++++++++++++++
>  1 file changed, 57 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/watchdog/intel,keembay-wdt.yaml
> 


Please add Acked-by/Reviewed-by tags when posting new versions. However,
there's no need to repost patches *only* to add the tags. The upstream
maintainer will do that for acks received on the version they apply.

If a tag was not added on purpose, please state why and what changed.

RE: [PATCH v3 2/2] dt-bindings: watchdog: Add bindings for Intel Keem Bay SoC

[Ayyathurai, Vijayakannan]

Hi Rob,

> From: Rob Herring <robh@kernel.org>
> 
> > From: Vijayakannan Ayyathurai <vijayakannan.ayyathurai@intel.com>
> >
> > Add Device Tree binding document for Watchdog IP in the Intel Keem Bay
> SoC.
> >
> > Signed-off-by: Vijayakannan Ayyathurai
> <vijayakannan.ayyathurai@intel.com>
> > Acked-by: Mark Gross <mgross@linux.intel.com>
> > Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> > ---
> >  .../bindings/watchdog/intel,keembay-wdt.yaml  | 57 +++++++++++++++++++
> >  1 file changed, 57 insertions(+)
> >  create mode 100644
> Documentation/devicetree/bindings/watchdog/intel,keembay-wdt.yaml
> >
> 
> 
> Please add Acked-by/Reviewed-by tags when posting new versions. However,
> there's no need to repost patches *only* to add the tags. The upstream
> maintainer will do that for acks received on the version they apply.
> 
> If a tag was not added on purpose, please state why and what changed.

Sure. It is my mistake. I will add your tag, when posting the next version.

Thanks,
Vijay