[PATCH v8 0/2] Add support for Gunyah Watchdog

[PATCH v8 0/2] Add support for Gunyah Watchdog

[Hrishabh Rajput via B4 Relay]

Gunyah is a Type-I hypervisor which was introduced in the patch series
[1]. It is an open source hypervisor. The source repo is available at
[2].

The Gunyah Hypervisor doesn't allow its Virtual Machines to directly
access the MMIO watchdog. It either provides the fully emulated MMIO
based watchdog interface or the SMC-based watchdog interface depending
on the hypervisor configuration.
The SMC-based watchdog follows ARM's SMC Calling Convention (SMCCC)
version 1.1 and uses Vendor Specific Hypervisor Service Calls space.

This patch series adds support for the SMC-based watchdog interface
provided by the Gunyah Hypervisor.

This series is tested on SM8750 platform.

[1]
https://lore.kernel.org/all/20240222-gunyah-v17-0-1e9da6763d38@quicinc.com/

[2]
https://github.com/quic/gunyah-hypervisor

Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
---
Changes in v8:
- Fix error handling in gunyah_wdt_probe() to fail silently with -ENODEV
  if WDT_STATUS call returns -EOPNOTSUPP, indicating support for Gunyah
  watchdog is not present. Fail with logs for other errors.
- Link to v7: https://lore.kernel.org/r/20251114-gunyah_watchdog-v7-0-f5c155b941d5@oss.qualcomm.com

Changes in v7:
- Convert local `const` arrays to `static const` to optimize
  initialization and stack allocation.
- Link to v6: https://lore.kernel.org/r/20251112-gunyah_watchdog-v6-0-38ad01f8dac0@oss.qualcomm.com

Changes in v6:
- Fix build issues reported by the kernel test robot on PowerPC and RISC-V
  architectures by adding IS_ENABLED(CONFIG_HAVE_ARM_SMCCC_DISCOVERY) check
  before calling arm_smccc_hypervisor_has_uuid().
- Link to v5: https://lore.kernel.org/r/20251107-gunyah_watchdog-v5-0-4c6e3fb6eb17@oss.qualcomm.com

Changes in v5:
- Move the gunyah_wdt device registration from the SMEM driver to the
  SCM driver. Add additional logic to check if we're running under the
  Gunyah Hypervisor.
- Implement .remove() for gunyah_wdt driver to make it not persistent.
- Link to v4: https://lore.kernel.org/r/20251031-gunyah_watchdog-v4-0-7abb1ee11315@oss.qualcomm.com

Changes in v4:
- Move the contents of gunyah_wdt_init() to qcom_smem_probe() to make
  sure we're registering the watchdog only on the Qualcomm devices.
- Link to v3: https://lore.kernel.org/r/20251028-gunyah_watchdog-v3-1-e6d1ea438b1d@oss.qualcomm.com

Changes in v3:
- Move back to platform driver model. In module init, determine if we're
  running on a Qualcomm device and there is no supported memory-mapped
  watchdog present. Then proceed to register platform device and driver
  for SMC-based Gunyah watchdog.
- To determine if we're running on a Qualcomm device we're checking the
  presence of "qcom,smem" compatible devicetree node. As an alternative,
  we also tried using socinfo for the same purpose. When both
  gunyah_wdt and socinfo drivers were made built-in, it couldn't be
  ensured that the socinfo driver probed successfully before gunyah_wdt
  init was called. Hence, we resorted to the devicetree node approach.
- Limit the errors listed in gunyah_error to the ones that can be
  produced by the driver.
- Link to v2: https://lore.kernel.org/r/20251006-gunyah_watchdog-v2-1-b99d41d45450@oss.qualcomm.com

Changes in v2:
- Move away from platform driver model since the devicetree overlay does
  not happen by default.
  See https://lore.kernel.org/all/91002189-9d9e-48a2-8424-c42705fed3f8@quicinc.com/
- Only when MMIO-based watchdog device is absent in the devicetree,
  proceed to detect SMC-based watchdog using GUNYAH_WDT_STATUS SMC and
  initialize if SMC returns success.
- Implement pm notifiers as gunyah_wdt is no longer a platform driver so
  dev_pm_ops cannot be used.
- Pretimeout IRQ is no longer supported.
- Remove struct gunyah_wdt since it is not required.
- Move the contents of gunyah_errno.h to gunyah_wdt.c.
- Link to v1: https://lore.kernel.org/r/20250903-gunyah_watchdog-v1-0-3ae690530e4b@oss.qualcomm.com

---
Hrishabh Rajput (2):
      firmware: qcom: scm: Register gunyah watchdog device
      watchdog: Add driver for Gunyah Watchdog

 MAINTAINERS                      |   1 +
 drivers/firmware/qcom/qcom_scm.c |  53 ++++++++
 drivers/watchdog/Kconfig         |  13 ++
 drivers/watchdog/Makefile        |   1 +
 drivers/watchdog/gunyah_wdt.c    | 261 +++++++++++++++++++++++++++++++++++++++
 5 files changed, 329 insertions(+)
---
base-commit: 6a23ae0a96a600d1d12557add110e0bb6e32730c
change-id: 20250903-gunyah_watchdog-2d2649438e29

Best regards,
-- 
Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>

[PATCH v8 1/2] firmware: qcom: scm: Register gunyah watchdog device

[Hrishabh Rajput via B4 Relay]

From: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>

To restrict Gunyah watchdog initialization to Qualcomm platforms running
under the Gunyah Hypervisor, register the watchdog device in the QCOM
SCM driver.

When Gunyah is not present or Gunyah emulates MMIO-based watchdog, we
expect Qualcomm watchdog or ARM SBSA watchdog device to be present in
the devicetree. First, we make sure we're running under the Gunyah
Hypervisor. Then we move to check if any of the above mentioned
watchdog device nodes are present, if not then we proceed to register
the SMC-based Gunyah watchdog device.

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Tested-by: Shivendra Pratap <shivendra.pratap@oss.qualcomm.com>
Tested-by: Neil Armstrong <neil.armstrong@linaro.org>
Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
---
 drivers/firmware/qcom/qcom_scm.c | 53 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 53 insertions(+)

diff --git a/drivers/firmware/qcom/qcom_scm.c b/drivers/firmware/qcom/qcom_scm.c
index e777b7cb9b12..14d0663316e6 100644
--- a/drivers/firmware/qcom/qcom_scm.c
+++ b/drivers/firmware/qcom/qcom_scm.c
@@ -2182,6 +2182,56 @@ int qcom_scm_qtee_callback_response(phys_addr_t buf, size_t buf_size,
 }
 EXPORT_SYMBOL(qcom_scm_qtee_callback_response);
 
+static void qcom_scm_gunyah_wdt_free(void *data)
+{
+	struct platform_device *gunyah_wdt_dev = data;
+
+	platform_device_unregister(gunyah_wdt_dev);
+}
+
+static void qcom_scm_gunyah_wdt_init(struct qcom_scm *scm)
+{
+	struct platform_device *gunyah_wdt_dev;
+	struct device_node *np;
+	bool of_wdt_available;
+	int i;
+	static const uuid_t gunyah_uuid = UUID_INIT(0xc1d58fcd, 0xa453, 0x5fdb,
+						    0x92, 0x65, 0xce, 0x36,
+						    0x67, 0x3d, 0x5f, 0x14);
+	static const char * const of_wdt_compatible[] = {
+		"qcom,kpss-wdt",
+		"arm,sbsa-gwdt",
+	};
+
+	/* Bail out if we are not running under Gunyah */
+	if (!IS_ENABLED(CONFIG_HAVE_ARM_SMCCC_DISCOVERY) ||
+	    !arm_smccc_hypervisor_has_uuid(&gunyah_uuid))
+		return;
+
+	/*
+	 * Gunyah emulates either of Qualcomm watchdog or ARM SBSA watchdog on
+	 * newer platforms. Bail out if we find them in the devicetree.
+	 */
+	for (i = 0; i < ARRAY_SIZE(of_wdt_compatible); i++) {
+		np = of_find_compatible_node(NULL, NULL, of_wdt_compatible[i]);
+		of_wdt_available = of_device_is_available(np);
+		of_node_put(np);
+		if (of_wdt_available)
+			return;
+	}
+
+	gunyah_wdt_dev = platform_device_register_simple("gunyah-wdt", -1,
+							 NULL, 0);
+	if (IS_ERR(gunyah_wdt_dev)) {
+		dev_err(scm->dev, "Failed to register Gunyah watchdog device: %ld\n",
+			PTR_ERR(gunyah_wdt_dev));
+		return;
+	}
+
+	devm_add_action_or_reset(scm->dev, qcom_scm_gunyah_wdt_free,
+				 gunyah_wdt_dev);
+}
+
 static void qcom_scm_qtee_free(void *data)
 {
 	struct platform_device *qtee_dev = data;
@@ -2448,6 +2498,9 @@ static int qcom_scm_probe(struct platform_device *pdev)
 	/* Initialize the QTEE object interface. */
 	qcom_scm_qtee_init(scm);
 
+	/* Initialize the Gunyah watchdog platform device. */
+	qcom_scm_gunyah_wdt_init(scm);
+
 	return 0;
 }
 

-- 
2.43.0

[PATCH v8 2/2] watchdog: Add driver for Gunyah Watchdog

[Hrishabh Rajput via B4 Relay]

From: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>

On Qualcomm SoCs running under the Gunyah hypervisor, access to watchdog
through MMIO is not available on all platforms. Depending on the
hypervisor configuration, the watchdog is either fully emulated or
exposed via ARM's SMC Calling Conventions (SMCCC) through the Vendor
Specific Hypervisor Service Calls space.

Add driver to support the SMC-based watchdog provided by the Gunyah
Hypervisor. Device registration is done in the QCOM SCM driver after
checks to restrict the watchdog initialization to Qualcomm devices
running under Gunyah.

Gunyah watchdog is not a hardware but an SMC-based vendor-specific
hypervisor interface provided by the Gunyah hypervisor. The design
involving QCOM SCM driver for registering the platform device has been
devised to avoid adding non-hardware nodes to devicetree.

Tested-by: Shivendra Pratap <shivendra.pratap@oss.qualcomm.com>
Tested-by: Neil Armstrong <neil.armstrong@linaro.org>
Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
---
 MAINTAINERS                   |   1 +
 drivers/watchdog/Kconfig      |  13 +++
 drivers/watchdog/Makefile     |   1 +
 drivers/watchdog/gunyah_wdt.c | 261 ++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 276 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index e64b94e6b5a9..5d2212c3c917 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -3216,6 +3216,7 @@ F:	arch/arm64/boot/dts/qcom/
 F:	drivers/bus/qcom*
 F:	drivers/firmware/qcom/
 F:	drivers/soc/qcom/
+F:	drivers/watchdog/gunyah_wdt.c
 F:	include/dt-bindings/arm/qcom,ids.h
 F:	include/dt-bindings/firmware/qcom,scm.h
 F:	include/dt-bindings/soc/qcom*
diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
index 05008d937e40..bc6db9a1c116 100644
--- a/drivers/watchdog/Kconfig
+++ b/drivers/watchdog/Kconfig
@@ -2354,4 +2354,17 @@ config KEEMBAY_WATCHDOG
 	  To compile this driver as a module, choose M here: the
 	  module will be called keembay_wdt.
 
+config GUNYAH_WATCHDOG
+	tristate "Qualcomm Gunyah Watchdog"
+	depends on ARCH_QCOM || COMPILE_TEST
+	depends on HAVE_ARM_SMCCC
+	select WATCHDOG_CORE
+	help
+	  Say Y here to include support for watchdog timer provided by the
+	  Gunyah hypervisor. The driver uses ARM SMC Calling Convention (SMCCC)
+	  to interact with Gunyah Watchdog.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called gunyah_wdt.
+
 endif # WATCHDOG
diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
index b680e4d3c1bc..1215efb7816d 100644
--- a/drivers/watchdog/Makefile
+++ b/drivers/watchdog/Makefile
@@ -102,6 +102,7 @@ obj-$(CONFIG_MSC313E_WATCHDOG) += msc313e_wdt.o
 obj-$(CONFIG_APPLE_WATCHDOG) += apple_wdt.o
 obj-$(CONFIG_SUNPLUS_WATCHDOG) += sunplus_wdt.o
 obj-$(CONFIG_MARVELL_GTI_WDT) += marvell_gti_wdt.o
+obj-$(CONFIG_GUNYAH_WATCHDOG) += gunyah_wdt.o
 
 # X86 (i386 + ia64 + x86_64) Architecture
 obj-$(CONFIG_ACQUIRE_WDT) += acquirewdt.o
diff --git a/drivers/watchdog/gunyah_wdt.c b/drivers/watchdog/gunyah_wdt.c
new file mode 100644
index 000000000000..49dfef459e84
--- /dev/null
+++ b/drivers/watchdog/gunyah_wdt.c
@@ -0,0 +1,261 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
+ */
+
+#include <linux/arm-smccc.h>
+#include <linux/delay.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/mod_devicetable.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/watchdog.h>
+
+#define GUNYAH_WDT_SMCCC_CALL_VAL(func_id) \
+	ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL, ARM_SMCCC_SMC_32,\
+			   ARM_SMCCC_OWNER_VENDOR_HYP, func_id)
+
+/* SMCCC function IDs for watchdog operations */
+#define GUNYAH_WDT_CONTROL   GUNYAH_WDT_SMCCC_CALL_VAL(0x0005)
+#define GUNYAH_WDT_STATUS    GUNYAH_WDT_SMCCC_CALL_VAL(0x0006)
+#define GUNYAH_WDT_PING      GUNYAH_WDT_SMCCC_CALL_VAL(0x0007)
+#define GUNYAH_WDT_SET_TIME  GUNYAH_WDT_SMCCC_CALL_VAL(0x0008)
+
+/*
+ * Control values for GUNYAH_WDT_CONTROL.
+ * Bit 0 is used to enable or disable the watchdog. If this bit is set,
+ * then the watchdog is enabled and vice versa.
+ * Bit 1 should always be set to 1 as this bit is reserved in Gunyah and
+ * it's expected to be 1.
+ */
+#define WDT_CTRL_ENABLE  (BIT(1) | BIT(0))
+#define WDT_CTRL_DISABLE BIT(1)
+
+enum gunyah_error {
+	GUNYAH_ERROR_OK				= 0,
+	GUNYAH_ERROR_UNIMPLEMENTED		= -1,
+	GUNYAH_ERROR_ARG_INVAL			= 1,
+};
+
+/**
+ * gunyah_error_remap() - Remap Gunyah hypervisor errors into a Linux error code
+ * @gunyah_error: Gunyah hypercall return value
+ */
+static inline int gunyah_error_remap(enum gunyah_error gunyah_error)
+{
+	switch (gunyah_error) {
+	case GUNYAH_ERROR_OK:
+		return 0;
+	case GUNYAH_ERROR_UNIMPLEMENTED:
+		return -EOPNOTSUPP;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int gunyah_wdt_call(unsigned long func_id, unsigned long arg1,
+			   unsigned long arg2)
+{
+	struct arm_smccc_res res;
+
+	arm_smccc_1_1_smc(func_id, arg1, arg2, &res);
+	return gunyah_error_remap(res.a0);
+}
+
+static int gunyah_wdt_start(struct watchdog_device *wdd)
+{
+	unsigned int timeout_ms;
+	struct device *dev = wdd->parent;
+	int ret;
+
+	ret = gunyah_wdt_call(GUNYAH_WDT_CONTROL, WDT_CTRL_DISABLE, 0);
+	if (ret && watchdog_active(wdd)) {
+		dev_err(dev, "%s: Failed to stop gunyah wdt %d\n", __func__, ret);
+		return ret;
+	}
+
+	timeout_ms = wdd->timeout * 1000;
+	ret = gunyah_wdt_call(GUNYAH_WDT_SET_TIME, timeout_ms, timeout_ms);
+	if (ret) {
+		dev_err(dev, "%s: Failed to set timeout for gunyah wdt %d\n",
+			__func__, ret);
+		return ret;
+	}
+
+	ret = gunyah_wdt_call(GUNYAH_WDT_CONTROL, WDT_CTRL_ENABLE, 0);
+	if (ret)
+		dev_err(dev, "%s: Failed to start gunyah wdt %d\n", __func__, ret);
+
+	return ret;
+}
+
+static int gunyah_wdt_stop(struct watchdog_device *wdd)
+{
+	return gunyah_wdt_call(GUNYAH_WDT_CONTROL, WDT_CTRL_DISABLE, 0);
+}
+
+static int gunyah_wdt_ping(struct watchdog_device *wdd)
+{
+	return gunyah_wdt_call(GUNYAH_WDT_PING, 0, 0);
+}
+
+static int gunyah_wdt_set_timeout(struct watchdog_device *wdd,
+				  unsigned int timeout_sec)
+{
+	wdd->timeout = timeout_sec;
+
+	if (watchdog_active(wdd))
+		return gunyah_wdt_start(wdd);
+
+	return 0;
+}
+
+static int gunyah_wdt_get_time_since_last_ping(void)
+{
+	struct arm_smccc_res res;
+
+	arm_smccc_1_1_smc(GUNYAH_WDT_STATUS, 0, 0, &res);
+	if (res.a0)
+		return gunyah_error_remap(res.a0);
+
+	return res.a2 / 1000;
+}
+
+static unsigned int gunyah_wdt_get_timeleft(struct watchdog_device *wdd)
+{
+	int seconds_since_last_ping;
+
+	seconds_since_last_ping = gunyah_wdt_get_time_since_last_ping();
+	if (seconds_since_last_ping < 0 ||
+	    seconds_since_last_ping > wdd->timeout)
+		return 0;
+
+	return wdd->timeout - seconds_since_last_ping;
+}
+
+static int gunyah_wdt_restart(struct watchdog_device *wdd,
+			      unsigned long action, void *data)
+{
+	/* Set timeout to 1ms and send a ping */
+	gunyah_wdt_call(GUNYAH_WDT_CONTROL, WDT_CTRL_DISABLE, 0);
+	gunyah_wdt_call(GUNYAH_WDT_SET_TIME, 1, 1);
+	gunyah_wdt_call(GUNYAH_WDT_CONTROL, WDT_CTRL_ENABLE, 0);
+	gunyah_wdt_call(GUNYAH_WDT_PING, 0, 0);
+
+	/* Wait to make sure reset occurs */
+	mdelay(100);
+
+	return 0;
+}
+
+static const struct watchdog_info gunyah_wdt_info = {
+	.identity = "Gunyah Watchdog",
+	.options = WDIOF_SETTIMEOUT
+		 | WDIOF_KEEPALIVEPING
+		 | WDIOF_MAGICCLOSE,
+};
+
+static const struct watchdog_ops gunyah_wdt_ops = {
+	.owner = THIS_MODULE,
+	.start = gunyah_wdt_start,
+	.stop = gunyah_wdt_stop,
+	.ping = gunyah_wdt_ping,
+	.set_timeout = gunyah_wdt_set_timeout,
+	.get_timeleft = gunyah_wdt_get_timeleft,
+	.restart = gunyah_wdt_restart
+};
+
+static int gunyah_wdt_probe(struct platform_device *pdev)
+{
+	struct watchdog_device *wdd;
+	struct device *dev = &pdev->dev;
+	int ret;
+
+	ret = gunyah_wdt_call(GUNYAH_WDT_STATUS, 0, 0);
+	if (ret == -EOPNOTSUPP)
+		return -ENODEV;
+
+	if (ret)
+		return dev_err_probe(dev, ret, "status check failed\n");
+
+	wdd = devm_kzalloc(dev, sizeof(*wdd), GFP_KERNEL);
+	if (!wdd)
+		return -ENOMEM;
+
+	wdd->info = &gunyah_wdt_info;
+	wdd->ops = &gunyah_wdt_ops;
+	wdd->parent = dev;
+
+	/*
+	 * Although Gunyah expects 16-bit unsigned int values as timeout values
+	 * in milliseconds, values above 0x8000 are reserved. This limits the
+	 * max timeout value to 32 seconds.
+	 */
+	wdd->max_timeout = 32; /* seconds */
+	wdd->min_timeout = 1; /* seconds */
+	wdd->timeout = wdd->max_timeout;
+
+	gunyah_wdt_stop(wdd);
+	platform_set_drvdata(pdev, wdd);
+	watchdog_set_restart_priority(wdd, 0);
+
+	return devm_watchdog_register_device(dev, wdd);
+}
+
+static void gunyah_wdt_remove(struct platform_device *pdev)
+{
+	struct watchdog_device *wdd = platform_get_drvdata(pdev);
+
+	gunyah_wdt_stop(wdd);
+}
+
+static int gunyah_wdt_suspend(struct device *dev)
+{
+	struct watchdog_device *wdd = dev_get_drvdata(dev);
+
+	if (watchdog_active(wdd))
+		gunyah_wdt_stop(wdd);
+
+	return 0;
+}
+
+static int gunyah_wdt_resume(struct device *dev)
+{
+	struct watchdog_device *wdd = dev_get_drvdata(dev);
+
+	if (watchdog_active(wdd))
+		gunyah_wdt_start(wdd);
+
+	return 0;
+}
+
+static DEFINE_SIMPLE_DEV_PM_OPS(gunyah_wdt_pm_ops, gunyah_wdt_suspend, gunyah_wdt_resume);
+
+/*
+ * Gunyah watchdog is a vendor-specific hypervisor interface provided by the
+ * Gunyah hypervisor. Using QCOM SCM driver to detect Gunyah watchdog SMCCC
+ * hypervisor service and register platform device when the service is available
+ * allows this driver to operate independently of the devicetree and avoids
+ * adding the non-hardware nodes to the devicetree.
+ */
+static const struct platform_device_id gunyah_wdt_id[] = {
+	{ .name = "gunyah-wdt" },
+	{}
+};
+MODULE_DEVICE_TABLE(platform, gunyah_wdt_id);
+
+static struct platform_driver gunyah_wdt_driver = {
+	.driver = {
+		.name = "gunyah-wdt",
+		.pm = pm_sleep_ptr(&gunyah_wdt_pm_ops),
+	},
+	.id_table = gunyah_wdt_id,
+	.probe = gunyah_wdt_probe,
+	.remove = gunyah_wdt_remove,
+};
+
+module_platform_driver(gunyah_wdt_driver);
+
+MODULE_DESCRIPTION("Gunyah Watchdog Driver");
+MODULE_LICENSE("GPL");

-- 
2.43.0

Re: [PATCH v8 1/2] firmware: qcom: scm: Register gunyah watchdog device

[Kathiravan Thirumoorthy]

On 11/18/2025 4:10 PM, Hrishabh Rajput via B4 Relay wrote:
> From: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
>
> To restrict Gunyah watchdog initialization to Qualcomm platforms running
> under the Gunyah Hypervisor, register the watchdog device in the QCOM
> SCM driver.
>
> When Gunyah is not present or Gunyah emulates MMIO-based watchdog, we
> expect Qualcomm watchdog or ARM SBSA watchdog device to be present in
> the devicetree. First, we make sure we're running under the Gunyah
> Hypervisor. Then we move to check if any of the above mentioned
> watchdog device nodes are present, if not then we proceed to register
> the SMC-based Gunyah watchdog device.
>
> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> Tested-by: Shivendra Pratap <shivendra.pratap@oss.qualcomm.com>
> Tested-by: Neil Armstrong <neil.armstrong@linaro.org>
> Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
> ---
>   drivers/firmware/qcom/qcom_scm.c | 53 ++++++++++++++++++++++++++++++++++++++++
>   1 file changed, 53 insertions(+)
>
> diff --git a/drivers/firmware/qcom/qcom_scm.c b/drivers/firmware/qcom/qcom_scm.c
> index e777b7cb9b12..14d0663316e6 100644
> --- a/drivers/firmware/qcom/qcom_scm.c
> +++ b/drivers/firmware/qcom/qcom_scm.c
> @@ -2182,6 +2182,56 @@ int qcom_scm_qtee_callback_response(phys_addr_t buf, size_t buf_size,
>   }
>   EXPORT_SYMBOL(qcom_scm_qtee_callback_response);
>   
> +static void qcom_scm_gunyah_wdt_free(void *data)
> +{
> +	struct platform_device *gunyah_wdt_dev = data;
> +
> +	platform_device_unregister(gunyah_wdt_dev);
> +}
> +
> +static void qcom_scm_gunyah_wdt_init(struct qcom_scm *scm)
> +{
> +	struct platform_device *gunyah_wdt_dev;
> +	struct device_node *np;

nit: Can we use the __cleanup() attribute for device_node like below and 
drop the explicit of_node_put()?

struct device_node *np __free(device_node) = NULL;

FYI - I submitted the patch which takes care of this in the SCM driver[1].

[1] 
https://lore.kernel.org/linux-arm-msm/20251117-scm_cleanup_of_node_put-v1-1-307d36d6b849@oss.qualcomm.com/

> +	bool of_wdt_available;
> +	int i;
> +	static const uuid_t gunyah_uuid = UUID_INIT(0xc1d58fcd, 0xa453, 0x5fdb,
> +						    0x92, 0x65, 0xce, 0x36,
> +						    0x67, 0x3d, 0x5f, 0x14);
> +	static const char * const of_wdt_compatible[] = {
> +		"qcom,kpss-wdt",
> +		"arm,sbsa-gwdt",
> +	};
> +
> +	/* Bail out if we are not running under Gunyah */
> +	if (!IS_ENABLED(CONFIG_HAVE_ARM_SMCCC_DISCOVERY) ||
> +	    !arm_smccc_hypervisor_has_uuid(&gunyah_uuid))
> +		return;
> +
> +	/*
> +	 * Gunyah emulates either of Qualcomm watchdog or ARM SBSA watchdog on
> +	 * newer platforms. Bail out if we find them in the devicetree.
> +	 */
> +	for (i = 0; i < ARRAY_SIZE(of_wdt_compatible); i++) {
> +		np = of_find_compatible_node(NULL, NULL, of_wdt_compatible[i]);
> +		of_wdt_available = of_device_is_available(np);
> +		of_node_put(np);
> +		if (of_wdt_available)
> +			return;
> +	}
> +
> +	gunyah_wdt_dev = platform_device_register_simple("gunyah-wdt", -1,
> +							 NULL, 0);
> +	if (IS_ERR(gunyah_wdt_dev)) {
> +		dev_err(scm->dev, "Failed to register Gunyah watchdog device: %ld\n",
> +			PTR_ERR(gunyah_wdt_dev));
> +		return;
> +	}
> +
> +	devm_add_action_or_reset(scm->dev, qcom_scm_gunyah_wdt_free,
> +				 gunyah_wdt_dev);
> +}
> +
>   static void qcom_scm_qtee_free(void *data)
>   {
>   	struct platform_device *qtee_dev = data;
> @@ -2448,6 +2498,9 @@ static int qcom_scm_probe(struct platform_device *pdev)
>   	/* Initialize the QTEE object interface. */
>   	qcom_scm_qtee_init(scm);
>   
> +	/* Initialize the Gunyah watchdog platform device. */
> +	qcom_scm_gunyah_wdt_init(scm);
> +
>   	return 0;
>   }
>   
>

Re: [PATCH v8 1/2] firmware: qcom: scm: Register gunyah watchdog device

[Krzysztof Kozlowski]

On 18/11/2025 12:18, Kathiravan Thirumoorthy wrote:
> 
> On 11/18/2025 4:10 PM, Hrishabh Rajput via B4 Relay wrote:
>> From: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
>>
>> To restrict Gunyah watchdog initialization to Qualcomm platforms running
>> under the Gunyah Hypervisor, register the watchdog device in the QCOM
>> SCM driver.
>>
>> When Gunyah is not present or Gunyah emulates MMIO-based watchdog, we
>> expect Qualcomm watchdog or ARM SBSA watchdog device to be present in
>> the devicetree. First, we make sure we're running under the Gunyah
>> Hypervisor. Then we move to check if any of the above mentioned
>> watchdog device nodes are present, if not then we proceed to register
>> the SMC-based Gunyah watchdog device.
>>
>> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
>> Tested-by: Shivendra Pratap <shivendra.pratap@oss.qualcomm.com>
>> Tested-by: Neil Armstrong <neil.armstrong@linaro.org>
>> Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
>> ---
>>   drivers/firmware/qcom/qcom_scm.c | 53 ++++++++++++++++++++++++++++++++++++++++
>>   1 file changed, 53 insertions(+)
>>
>> diff --git a/drivers/firmware/qcom/qcom_scm.c b/drivers/firmware/qcom/qcom_scm.c
>> index e777b7cb9b12..14d0663316e6 100644
>> --- a/drivers/firmware/qcom/qcom_scm.c
>> +++ b/drivers/firmware/qcom/qcom_scm.c
>> @@ -2182,6 +2182,56 @@ int qcom_scm_qtee_callback_response(phys_addr_t buf, size_t buf_size,
>>   }
>>   EXPORT_SYMBOL(qcom_scm_qtee_callback_response);
>>   
>> +static void qcom_scm_gunyah_wdt_free(void *data)
>> +{
>> +	struct platform_device *gunyah_wdt_dev = data;
>> +
>> +	platform_device_unregister(gunyah_wdt_dev);
>> +}
>> +
>> +static void qcom_scm_gunyah_wdt_init(struct qcom_scm *scm)
>> +{
>> +	struct platform_device *gunyah_wdt_dev;
>> +	struct device_node *np;
> 
> nit: Can we use the __cleanup() attribute for device_node like below and 
> drop the explicit of_node_put()?
> 
> struct device_node *np __free(device_node) = NULL;

Please don't. It is not desired style, error prone and not helping.

Don't use cleanup.h if you are not accustomed to its specific style.


Best regards,
Krzysztof

Re: [PATCH v8 1/2] firmware: qcom: scm: Register gunyah watchdog device

[Krzysztof Kozlowski]

On 18/11/2025 12:32, Krzysztof Kozlowski wrote:
> On 18/11/2025 12:18, Kathiravan Thirumoorthy wrote:
>>
>> On 11/18/2025 4:10 PM, Hrishabh Rajput via B4 Relay wrote:
>>> From: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
>>>
>>> To restrict Gunyah watchdog initialization to Qualcomm platforms running
>>> under the Gunyah Hypervisor, register the watchdog device in the QCOM
>>> SCM driver.
>>>
>>> When Gunyah is not present or Gunyah emulates MMIO-based watchdog, we
>>> expect Qualcomm watchdog or ARM SBSA watchdog device to be present in
>>> the devicetree. First, we make sure we're running under the Gunyah
>>> Hypervisor. Then we move to check if any of the above mentioned
>>> watchdog device nodes are present, if not then we proceed to register
>>> the SMC-based Gunyah watchdog device.
>>>
>>> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
>>> Tested-by: Shivendra Pratap <shivendra.pratap@oss.qualcomm.com>
>>> Tested-by: Neil Armstrong <neil.armstrong@linaro.org>
>>> Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
>>> ---
>>>   drivers/firmware/qcom/qcom_scm.c | 53 ++++++++++++++++++++++++++++++++++++++++
>>>   1 file changed, 53 insertions(+)
>>>
>>> diff --git a/drivers/firmware/qcom/qcom_scm.c b/drivers/firmware/qcom/qcom_scm.c
>>> index e777b7cb9b12..14d0663316e6 100644
>>> --- a/drivers/firmware/qcom/qcom_scm.c
>>> +++ b/drivers/firmware/qcom/qcom_scm.c
>>> @@ -2182,6 +2182,56 @@ int qcom_scm_qtee_callback_response(phys_addr_t buf, size_t buf_size,
>>>   }
>>>   EXPORT_SYMBOL(qcom_scm_qtee_callback_response);
>>>   
>>> +static void qcom_scm_gunyah_wdt_free(void *data)
>>> +{
>>> +	struct platform_device *gunyah_wdt_dev = data;
>>> +
>>> +	platform_device_unregister(gunyah_wdt_dev);
>>> +}
>>> +
>>> +static void qcom_scm_gunyah_wdt_init(struct qcom_scm *scm)
>>> +{
>>> +	struct platform_device *gunyah_wdt_dev;
>>> +	struct device_node *np;
>>
>> nit: Can we use the __cleanup() attribute for device_node like below and 
>> drop the explicit of_node_put()?
>>
>> struct device_node *np __free(device_node) = NULL;
> 
> Please don't. It is not desired style, error prone and not helping.
> 
> Don't use cleanup.h if you are not accustomed to its specific style.

Good that you linked here your patches, so I can NAK everything. That's
very poor idea. Simple, readable code you transform into buggy,
error-prone discouraged style.

In all your patches the code was left like this ON PURPOSE.


Best regards,
Krzysztof

Re: [PATCH v8 2/2] watchdog: Add driver for Gunyah Watchdog

[Dmitry Baryshkov]

On Tue, Nov 18, 2025 at 10:40:57AM +0000, Hrishabh Rajput via B4 Relay wrote:
> From: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
> 
> On Qualcomm SoCs running under the Gunyah hypervisor, access to watchdog
> through MMIO is not available on all platforms. Depending on the
> hypervisor configuration, the watchdog is either fully emulated or
> exposed via ARM's SMC Calling Conventions (SMCCC) through the Vendor
> Specific Hypervisor Service Calls space.
> 
> Add driver to support the SMC-based watchdog provided by the Gunyah
> Hypervisor. Device registration is done in the QCOM SCM driver after
> checks to restrict the watchdog initialization to Qualcomm devices
> running under Gunyah.
> 
> Gunyah watchdog is not a hardware but an SMC-based vendor-specific
> hypervisor interface provided by the Gunyah hypervisor. The design
> involving QCOM SCM driver for registering the platform device has been
> devised to avoid adding non-hardware nodes to devicetree.
> 
> Tested-by: Shivendra Pratap <shivendra.pratap@oss.qualcomm.com>
> Tested-by: Neil Armstrong <neil.armstrong@linaro.org>
> Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
> ---
>  MAINTAINERS                   |   1 +
>  drivers/watchdog/Kconfig      |  13 +++
>  drivers/watchdog/Makefile     |   1 +
>  drivers/watchdog/gunyah_wdt.c | 261 ++++++++++++++++++++++++++++++++++++++++++
>  4 files changed, 276 insertions(+)

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>


-- 
With best wishes
Dmitry

Re: [PATCH v8 2/2] watchdog: Add driver for Gunyah Watchdog

[Guenter Roeck]

On 11/18/25 02:40, Hrishabh Rajput via B4 Relay wrote:
> From: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
> 
> On Qualcomm SoCs running under the Gunyah hypervisor, access to watchdog
> through MMIO is not available on all platforms. Depending on the
> hypervisor configuration, the watchdog is either fully emulated or
> exposed via ARM's SMC Calling Conventions (SMCCC) through the Vendor
> Specific Hypervisor Service Calls space.
> 
> Add driver to support the SMC-based watchdog provided by the Gunyah
> Hypervisor. Device registration is done in the QCOM SCM driver after
> checks to restrict the watchdog initialization to Qualcomm devices
> running under Gunyah.
> 
> Gunyah watchdog is not a hardware but an SMC-based vendor-specific
> hypervisor interface provided by the Gunyah hypervisor. The design
> involving QCOM SCM driver for registering the platform device has been
> devised to avoid adding non-hardware nodes to devicetree.
> 
> Tested-by: Shivendra Pratap <shivendra.pratap@oss.qualcomm.com>
> Tested-by: Neil Armstrong <neil.armstrong@linaro.org>
> Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
> ---
>   MAINTAINERS                   |   1 +
>   drivers/watchdog/Kconfig      |  13 +++
>   drivers/watchdog/Makefile     |   1 +
>   drivers/watchdog/gunyah_wdt.c | 261 ++++++++++++++++++++++++++++++++++++++++++
>   4 files changed, 276 insertions(+)
> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index e64b94e6b5a9..5d2212c3c917 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -3216,6 +3216,7 @@ F:	arch/arm64/boot/dts/qcom/
>   F:	drivers/bus/qcom*
>   F:	drivers/firmware/qcom/
>   F:	drivers/soc/qcom/
> +F:	drivers/watchdog/gunyah_wdt.c
>   F:	include/dt-bindings/arm/qcom,ids.h
>   F:	include/dt-bindings/firmware/qcom,scm.h
>   F:	include/dt-bindings/soc/qcom*
> diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
> index 05008d937e40..bc6db9a1c116 100644
> --- a/drivers/watchdog/Kconfig
> +++ b/drivers/watchdog/Kconfig
> @@ -2354,4 +2354,17 @@ config KEEMBAY_WATCHDOG
>   	  To compile this driver as a module, choose M here: the
>   	  module will be called keembay_wdt.
>   
> +config GUNYAH_WATCHDOG
> +	tristate "Qualcomm Gunyah Watchdog"
> +	depends on ARCH_QCOM || COMPILE_TEST
> +	depends on HAVE_ARM_SMCCC
> +	select WATCHDOG_CORE
> +	help
> +	  Say Y here to include support for watchdog timer provided by the
> +	  Gunyah hypervisor. The driver uses ARM SMC Calling Convention (SMCCC)
> +	  to interact with Gunyah Watchdog.
> +
> +	  To compile this driver as a module, choose M here: the
> +	  module will be called gunyah_wdt.
> +
>   endif # WATCHDOG
> diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
> index b680e4d3c1bc..1215efb7816d 100644
> --- a/drivers/watchdog/Makefile
> +++ b/drivers/watchdog/Makefile
> @@ -102,6 +102,7 @@ obj-$(CONFIG_MSC313E_WATCHDOG) += msc313e_wdt.o
>   obj-$(CONFIG_APPLE_WATCHDOG) += apple_wdt.o
>   obj-$(CONFIG_SUNPLUS_WATCHDOG) += sunplus_wdt.o
>   obj-$(CONFIG_MARVELL_GTI_WDT) += marvell_gti_wdt.o
> +obj-$(CONFIG_GUNYAH_WATCHDOG) += gunyah_wdt.o
>   
>   # X86 (i386 + ia64 + x86_64) Architecture
>   obj-$(CONFIG_ACQUIRE_WDT) += acquirewdt.o
> diff --git a/drivers/watchdog/gunyah_wdt.c b/drivers/watchdog/gunyah_wdt.c
> new file mode 100644
> index 000000000000..49dfef459e84
> --- /dev/null
> +++ b/drivers/watchdog/gunyah_wdt.c
> @@ -0,0 +1,261 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
> + */
> +
> +#include <linux/arm-smccc.h>
> +#include <linux/delay.h>
> +#include <linux/errno.h>
> +#include <linux/kernel.h>
> +#include <linux/mod_devicetable.h>
> +#include <linux/module.h>
> +#include <linux/platform_device.h>
> +#include <linux/watchdog.h>
> +
> +#define GUNYAH_WDT_SMCCC_CALL_VAL(func_id) \
> +	ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL, ARM_SMCCC_SMC_32,\
> +			   ARM_SMCCC_OWNER_VENDOR_HYP, func_id)
> +
> +/* SMCCC function IDs for watchdog operations */
> +#define GUNYAH_WDT_CONTROL   GUNYAH_WDT_SMCCC_CALL_VAL(0x0005)
> +#define GUNYAH_WDT_STATUS    GUNYAH_WDT_SMCCC_CALL_VAL(0x0006)
> +#define GUNYAH_WDT_PING      GUNYAH_WDT_SMCCC_CALL_VAL(0x0007)
> +#define GUNYAH_WDT_SET_TIME  GUNYAH_WDT_SMCCC_CALL_VAL(0x0008)
> +
> +/*
> + * Control values for GUNYAH_WDT_CONTROL.
> + * Bit 0 is used to enable or disable the watchdog. If this bit is set,
> + * then the watchdog is enabled and vice versa.
> + * Bit 1 should always be set to 1 as this bit is reserved in Gunyah and
> + * it's expected to be 1.
> + */
> +#define WDT_CTRL_ENABLE  (BIT(1) | BIT(0))
> +#define WDT_CTRL_DISABLE BIT(1)
> +
> +enum gunyah_error {
> +	GUNYAH_ERROR_OK				= 0,
> +	GUNYAH_ERROR_UNIMPLEMENTED		= -1,
> +	GUNYAH_ERROR_ARG_INVAL			= 1,
> +};
> +
> +/**
> + * gunyah_error_remap() - Remap Gunyah hypervisor errors into a Linux error code
> + * @gunyah_error: Gunyah hypercall return value
> + */
> +static inline int gunyah_error_remap(enum gunyah_error gunyah_error)
> +{
> +	switch (gunyah_error) {
> +	case GUNYAH_ERROR_OK:
> +		return 0;
> +	case GUNYAH_ERROR_UNIMPLEMENTED:
> +		return -EOPNOTSUPP;
> +	default:
> +		return -EINVAL;
> +	}
> +}
> +
> +static int gunyah_wdt_call(unsigned long func_id, unsigned long arg1,
> +			   unsigned long arg2)
> +{
> +	struct arm_smccc_res res;
> +
> +	arm_smccc_1_1_smc(func_id, arg1, arg2, &res);
> +	return gunyah_error_remap(res.a0);
> +}
> +
> +static int gunyah_wdt_start(struct watchdog_device *wdd)
> +{
> +	unsigned int timeout_ms;
> +	struct device *dev = wdd->parent;
> +	int ret;
> +
> +	ret = gunyah_wdt_call(GUNYAH_WDT_CONTROL, WDT_CTRL_DISABLE, 0);
> +	if (ret && watchdog_active(wdd)) {
> +		dev_err(dev, "%s: Failed to stop gunyah wdt %d\n", __func__, ret);

I am not a friend of such logging noise (it will clog the kernel log if there
is a problem with the firmware), but then I figure that it isn't _my_ system
that gets flooded with noise, so

Reviewed-by: Guenter Roeck <linux@roeck-us.net>

Guenter

Re: [PATCH v8 0/2] Add support for Gunyah Watchdog

[Hrishabh Rajput]

Hi Bjorn, Guenter, and Wim,

Just a gentle ping on this series.

Since the patches have received Reviewed-by tags from Dmitry and 
Guenter, I wanted to confirm the merge strategy.

Bjorn: Are you planning to pick the QCOM SCM changes separately through 
your tree, or would you prefer the whole series go through the Watchdog 
tree?
If the latter, do we need an explicit Acked-by from you for QCOM SCM patch?

Thanks,
Hrishabh

Re: [PATCH v8 0/2] Add support for Gunyah Watchdog

[Krzysztof Kozlowski]

On 02/12/2025 12:23, Hrishabh Rajput wrote:
> Hi Bjorn, Guenter, and Wim,
> 
> Just a gentle ping on this series.

It's merge window. There was no point in pinging just before merge
window and is even worse to ping now. Nothing can happen with this
patchset and such pings is only noise.

> 
> Since the patches have received Reviewed-by tags from Dmitry and 
> Guenter, I wanted to confirm the merge strategy.
> 
> Bjorn: Are you planning to pick the QCOM SCM changes separately through 
> your tree, or would you prefer the whole series go through the Watchdog 
> tree?
> If the latter, do we need an explicit Acked-by from you for QCOM SCM patch?

Where did you document dependencies between patches and any non-obvious
merging? I open cover letter and there is NOTHING. I look at patch
changelog and also NOTHING.

So if you tell us nothing, why would we care to think we need to do
anything special here?

You must explicitly document every dependency, both external and between
patches, in the cover letter. At least cover letter, some people (e.g.
mostly me) don't even read them...

Best regards,
Krzysztof

Re: [PATCH v8 2/2] watchdog: Add driver for Gunyah Watchdog

[Mukesh Ojha]

On Tue, Nov 18, 2025 at 10:40:57AM +0000, Hrishabh Rajput wrote:
> On Qualcomm SoCs running under the Gunyah hypervisor, access to watchdog
> through MMIO is not available on all platforms. Depending on the
> hypervisor configuration, the watchdog is either fully emulated or
> exposed via ARM's SMC Calling Conventions (SMCCC) through the Vendor
> Specific Hypervisor Service Calls space.
> 
> Add driver to support the SMC-based watchdog provided by the Gunyah
> Hypervisor. Device registration is done in the QCOM SCM driver after
> checks to restrict the watchdog initialization to Qualcomm devices
> running under Gunyah.
> 
> Gunyah watchdog is not a hardware but an SMC-based vendor-specific
> hypervisor interface provided by the Gunyah hypervisor. The design
> involving QCOM SCM driver for registering the platform device has been
> devised to avoid adding non-hardware nodes to devicetree.
> 
> Tested-by: Shivendra Pratap <shivendra.pratap@oss.qualcomm.com>
> Tested-by: Neil Armstrong <neil.armstrong@linaro.org>
> Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>

I wanted to test the latest mem-inspect (Minidump) patches, but due to a
firmware bug, my last log after crash messages were not being captured. This
forced me to get another device. I had an SM8750, but it does not
support dump collection because of the missing Gunyah watchdog support
in the upstream. Luckily, this series came to the rescue, and now I am
able to capture the last crash messages.

Feel free to add

Tested-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>

to the series.

-- 
-Mukesh Ojha

Re: [PATCH v8 0/2] Add support for Gunyah Watchdog

[Hrishabh Rajput]

On 12/2/2025 9:29 PM, Krzysztof Kozlowski wrote:
> On 02/12/2025 12:23, Hrishabh Rajput wrote:
>> Hi Bjorn, Guenter, and Wim,
>>
>> Just a gentle ping on this series.
> 
> It's merge window. There was no point in pinging just before merge
> window and is even worse to ping now. Nothing can happen with this
> patchset and such pings is only noise.
> 

Thanks for the guidance and apologies for the noise created during the 
merge window.

>>
>> Since the patches have received Reviewed-by tags from Dmitry and
>> Guenter, I wanted to confirm the merge strategy.
>>
>> Bjorn: Are you planning to pick the QCOM SCM changes separately through
>> your tree, or would you prefer the whole series go through the Watchdog
>> tree?
>> If the latter, do we need an explicit Acked-by from you for QCOM SCM patch?
> 
> Where did you document dependencies between patches and any non-obvious
> merging? I open cover letter and there is NOTHING. I look at patch
> changelog and also NOTHING.
> 
> So if you tell us nothing, why would we care to think we need to do
> anything special here?
> 
> You must explicitly document every dependency, both external and between
> patches, in the cover letter. At least cover letter, some people (e.g.
> mostly me) don't even read them...
> 

This is a miss from my end. The following information should have been 
the part of the cover letter:
```
This series spans 2 subsystems and is split as follows:
- Patch 1: QCOM SCM - Register Gunyah Watchdog Platform device
- Patch 2: Watchdog - Add Gunyah Watchdog driver

Dependency:
There is no build-time dependency between the patches, but Patch 1 is 
required for Patch 2 to function.

Merge strategies:
- Strategy 1: Take both patches via the Watchdog tree.
- Strategy 2: Take Patch 1 via QCM SCM maintainter's tree, Patch 2 via 
Watchdog tree.

Since the patches concern primarily with the Watchdog, I suggest we go 
ahead with Strategy 1. If this is acceptable, I request an Acked-by from 
QCOM SCM maintainer for Patch 1.
```

I understand that this should have been a part of the cover letter. If 
it helps the process, I can add the above information in the cover 
letter and resend as v9. Since there are no other fixes, v9 would only 
contain the cover letter changes.

Thanks,
Hrishabh

Re: [PATCH v8 0/2] Add support for Gunyah Watchdog

[Pavan Kondeti]

Hi Bjorn and Wim,

On Mon, Dec 15, 2025 at 06:30:47PM +0530, Hrishabh Rajput wrote:
> 
> 
> On 12/2/2025 9:29 PM, Krzysztof Kozlowski wrote:
> > On 02/12/2025 12:23, Hrishabh Rajput wrote:
> > > Hi Bjorn, Guenter, and Wim,
> > > 
> > > Just a gentle ping on this series.
> > 
> > It's merge window. There was no point in pinging just before merge
> > window and is even worse to ping now. Nothing can happen with this
> > patchset and such pings is only noise.
> > 
> 
> Thanks for the guidance and apologies for the noise created during the merge
> window.
> 
> > > 
> > > Since the patches have received Reviewed-by tags from Dmitry and
> > > Guenter, I wanted to confirm the merge strategy.
> > > 
> > > Bjorn: Are you planning to pick the QCOM SCM changes separately through
> > > your tree, or would you prefer the whole series go through the Watchdog
> > > tree?
> > > If the latter, do we need an explicit Acked-by from you for QCOM SCM patch?
> > 
> > Where did you document dependencies between patches and any non-obvious
> > merging? I open cover letter and there is NOTHING. I look at patch
> > changelog and also NOTHING.
> > 
> > So if you tell us nothing, why would we care to think we need to do
> > anything special here?
> > 
> > You must explicitly document every dependency, both external and between
> > patches, in the cover letter. At least cover letter, some people (e.g.
> > mostly me) don't even read them...
> > 
> 
> This is a miss from my end. The following information should have been the
> part of the cover letter:
> ```
> This series spans 2 subsystems and is split as follows:
> - Patch 1: QCOM SCM - Register Gunyah Watchdog Platform device
> - Patch 2: Watchdog - Add Gunyah Watchdog driver
> 
> Dependency:
> There is no build-time dependency between the patches, but Patch 1 is
> required for Patch 2 to function.
> 
> Merge strategies:
> - Strategy 1: Take both patches via the Watchdog tree.
> - Strategy 2: Take Patch 1 via QCM SCM maintainter's tree, Patch 2 via
> Watchdog tree.
> 
> Since the patches concern primarily with the Watchdog, I suggest we go ahead
> with Strategy 1. If this is acceptable, I request an Acked-by from QCOM SCM
> maintainer for Patch 1.
> ```
> 

Is it possible to pick it up for v6.20? As mentioned above, both patches
don't have compile time dependency, however the QCOM SCM patch is needed
for probing the watchdog device.

Thanks,
Pavan

Re: [PATCH v8 0/2] Add support for Gunyah Watchdog

[Pavan Kondeti]

Hi Bjorn,

On Tue, Jan 06, 2026 at 11:25:49AM +0530, Pavan Kondeti wrote:
> Hi Bjorn and Wim,
> 
> On Mon, Dec 15, 2025 at 06:30:47PM +0530, Hrishabh Rajput wrote:
> > 
> > 
> > On 12/2/2025 9:29 PM, Krzysztof Kozlowski wrote:
> > > On 02/12/2025 12:23, Hrishabh Rajput wrote:
> > > > Hi Bjorn, Guenter, and Wim,
> > > > 
> > > > Just a gentle ping on this series.
> > > 
> > > It's merge window. There was no point in pinging just before merge
> > > window and is even worse to ping now. Nothing can happen with this
> > > patchset and such pings is only noise.
> > > 
> > 
> > Thanks for the guidance and apologies for the noise created during the merge
> > window.
> > 
> > > > 
> > > > Since the patches have received Reviewed-by tags from Dmitry and
> > > > Guenter, I wanted to confirm the merge strategy.
> > > > 
> > > > Bjorn: Are you planning to pick the QCOM SCM changes separately through
> > > > your tree, or would you prefer the whole series go through the Watchdog
> > > > tree?
> > > > If the latter, do we need an explicit Acked-by from you for QCOM SCM patch?
> > > 
> > > Where did you document dependencies between patches and any non-obvious
> > > merging? I open cover letter and there is NOTHING. I look at patch
> > > changelog and also NOTHING.
> > > 
> > > So if you tell us nothing, why would we care to think we need to do
> > > anything special here?
> > > 
> > > You must explicitly document every dependency, both external and between
> > > patches, in the cover letter. At least cover letter, some people (e.g.
> > > mostly me) don't even read them...
> > > 
> > 
> > This is a miss from my end. The following information should have been the
> > part of the cover letter:
> > ```
> > This series spans 2 subsystems and is split as follows:
> > - Patch 1: QCOM SCM - Register Gunyah Watchdog Platform device
> > - Patch 2: Watchdog - Add Gunyah Watchdog driver
> > 
> > Dependency:
> > There is no build-time dependency between the patches, but Patch 1 is
> > required for Patch 2 to function.
> > 
> > Merge strategies:
> > - Strategy 1: Take both patches via the Watchdog tree.
> > - Strategy 2: Take Patch 1 via QCM SCM maintainter's tree, Patch 2 via
> > Watchdog tree.
> > 
> > Since the patches concern primarily with the Watchdog, I suggest we go ahead
> > with Strategy 1. If this is acceptable, I request an Acked-by from QCOM SCM
> > maintainer for Patch 1.
> > ```
> > 
> 
> Is it possible to pick it up for v6.20? As mentioned above, both patches
> don't have compile time dependency, however the QCOM SCM patch is needed
> for probing the watchdog device.
> 

Please let us know if we need to split the series into two separate
patches? or is it fine to get first patch through qcom-next and 2nd
patch through watchdog tree?

Thanks,
Pavan

Re: [PATCH v8 0/2] Add support for Gunyah Watchdog

[Pavan Kondeti]

Hi Bjorn / Wim,

On Tue, Nov 18, 2025 at 10:40:55AM +0000, Hrishabh Rajput via B4 Relay wrote:
> Gunyah is a Type-I hypervisor which was introduced in the patch series
> [1]. It is an open source hypervisor. The source repo is available at
> [2].
> 
> The Gunyah Hypervisor doesn't allow its Virtual Machines to directly
> access the MMIO watchdog. It either provides the fully emulated MMIO
> based watchdog interface or the SMC-based watchdog interface depending
> on the hypervisor configuration.
> The SMC-based watchdog follows ARM's SMC Calling Convention (SMCCC)
> version 1.1 and uses Vendor Specific Hypervisor Service Calls space.
> 
> This patch series adds support for the SMC-based watchdog interface
> provided by the Gunyah Hypervisor.
> 
> This series is tested on SM8750 platform.
> 
> [1]
> https://lore.kernel.org/all/20240222-gunyah-v17-0-1e9da6763d38@quicinc.com/
> 
> [2]
> https://github.com/quic/gunyah-hypervisor
> 
> Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
> ---
> Changes in v8:
> - Fix error handling in gunyah_wdt_probe() to fail silently with -ENODEV
>   if WDT_STATUS call returns -EOPNOTSUPP, indicating support for Gunyah
>   watchdog is not present. Fail with logs for other errors.
> - Link to v7: https://lore.kernel.org/r/20251114-gunyah_watchdog-v7-0-f5c155b941d5@oss.qualcomm.com
> 

The series applies cleanly on v7.0-rc1. I have tested the watchdog
driver functionality and ramdumps collection on Kodiak RB3 Gen2
platform. Can you please pick this series for v7.1?

As Hrishabh mentioned earlier, both patches don't have any compile time
dependencies.

Thanks,
Pavan

Re: [PATCH v8 0/2] Add support for Gunyah Watchdog

[Pavan Kondeti]

Hi Bjorn / Wim,

On Tue, Feb 24, 2026 at 02:58:05PM +0530, Pavan Kondeti wrote:
> Hi Bjorn / Wim,
> 
> On Tue, Nov 18, 2025 at 10:40:55AM +0000, Hrishabh Rajput via B4 Relay wrote:
> > Gunyah is a Type-I hypervisor which was introduced in the patch series
> > [1]. It is an open source hypervisor. The source repo is available at
> > [2].
> > 
> > The Gunyah Hypervisor doesn't allow its Virtual Machines to directly
> > access the MMIO watchdog. It either provides the fully emulated MMIO
> > based watchdog interface or the SMC-based watchdog interface depending
> > on the hypervisor configuration.
> > The SMC-based watchdog follows ARM's SMC Calling Convention (SMCCC)
> > version 1.1 and uses Vendor Specific Hypervisor Service Calls space.
> > 
> > This patch series adds support for the SMC-based watchdog interface
> > provided by the Gunyah Hypervisor.
> > 
> > This series is tested on SM8750 platform.
> > 
> > [1]
> > https://lore.kernel.org/all/20240222-gunyah-v17-0-1e9da6763d38@quicinc.com/
> > 
> > [2]
> > https://github.com/quic/gunyah-hypervisor
> > 
> > Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
> > ---
> > Changes in v8:
> > - Fix error handling in gunyah_wdt_probe() to fail silently with -ENODEV
> >   if WDT_STATUS call returns -EOPNOTSUPP, indicating support for Gunyah
> >   watchdog is not present. Fail with logs for other errors.
> > - Link to v7: https://lore.kernel.org/r/20251114-gunyah_watchdog-v7-0-f5c155b941d5@oss.qualcomm.com
> > 
> 
> The series applies cleanly on v7.0-rc1. I have tested the watchdog
> driver functionality and ramdumps collection on Kodiak RB3 Gen2
> platform. Can you please pick this series for v7.1?
> 
> As Hrishabh mentioned earlier, both patches don't have any compile time
> dependencies.
> 

Since Hrishabh has left Qualcomm, I am taking over the series. Hence
sending request emails to include this series for v7.0 . I did verify
this on v7.0-rc1 as mentioned above. Since the series applies cleanly, I
have not re-sent the series. However, if you prefer to resend it since a
new person taking over, I can re-send the series with my Signed-off-by.

Please let me know your preference.

Thanks,
Pavan