Skip to contentSkip to conversation

watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature

PATCH New Janusz Uzycki 2 participants 15 messages Sep 22, 2014 – Sep 30, 2014

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/6 watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature

Some applications require to start watchdog before userspace software. This patch enables such feature. Only the flag is necessary to enable it.
Moreover kernel's ping is re-enabled when userspace software closed watchdog using the magic character. The features improves kernel's reliability if hardware watchdog is available.

---
 drivers/watchdog/watchdog_dev.c | 58 ++++++++++++++++++++++++++-
 include/linux/watchdog.h        |  5 +++
 2 files changed, 61 insertions(+), 2 deletions(-)
drivers/watchdog/watchdog_dev.c +56 -2
index 6aaefba..51a65f6 100644
@@ -41,6 +41,7 @@
 #include <linux/miscdevice.h>	/* For handling misc devices */
 #include <linux/init.h>		/* For __init/__exit/... */
 #include <linux/uaccess.h>	/* For copy_to_user/put_user/... */
+#include <linux/jiffies.h>	/* for ping timer */
 
 #include "watchdog_core.h"
 
@@ -277,6 +278,27 @@ out_ioctl:
 	return err;
 }
 
+/* 'keep on' feature */
+static void watchdog_ping_timer_cb(unsigned long data)
+{
+	struct watchdog_device *wdd = (struct watchdog_device *)data;
+	watchdog_ping(wdd);
+	/* call next ping half the timeout value */
+	mod_timer(&wdd->ping_timer,
+			jiffies + msecs_to_jiffies(wdd->timeout * 500));
+}
+
+static void watchdog_keepon_start(struct watchdog_device *wdd)
+{
+	watchdog_start(wdd);
+	watchdog_ping_timer_cb((unsigned long)wdd);
+}
+
+static void watchdog_keepon_stop(struct watchdog_device *wdd)
+{
+	del_timer_sync(&wdd->ping_timer);
+}
+
 /*
  *	watchdog_write: writes to the watchdog.
  *	@file: file from VFS
@@ -430,6 +452,9 @@ static int watchdog_open(struct inode *inode, struct file *file)
 	if (!try_module_get(wdd->ops->owner))
 		goto out;
 
+	if (test_bit(WDOG_KEEP_ON, &wdd->status))
+		watchdog_keepon_stop(wdd);
+
 	err = watchdog_start(wdd);
 	if (err < 0)
 		goto out_mod;
@@ -472,8 +497,13 @@ static int watchdog_release(struct inode *inode, struct file *file)
 	if (!test_bit(WDOG_ACTIVE, &wdd->status))
 		err = 0;
 	else if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
-		 !(wdd->info->options & WDIOF_MAGICCLOSE))
-		err = watchdog_stop(wdd);
+		 !(wdd->info->options & WDIOF_MAGICCLOSE)) {
+		if (test_bit(WDOG_KEEP_ON, &wdd->status)) {
+			watchdog_keepon_start(wdd);
+			err = 0;
+		} else
+			err = watchdog_stop(wdd);
+	}
 
 	/* If the watchdog was not stopped, send a keepalive ping */
 	if (err < 0) {
@@ -524,6 +554,14 @@ int watchdog_dev_register(struct watchdog_device *watchdog)
 {
 	int err, devno;
 
+	if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+		if (!try_module_get(watchdog->ops->owner))
+			return -ENODEV;
+		setup_timer(&watchdog->ping_timer, watchdog_ping_timer_cb,
+				(unsigned long)watchdog);
+		watchdog_keepon_start(watchdog);
+	}
+
 	if (watchdog->id == 0) {
 		old_wdd = watchdog;
 		watchdog_miscdev.parent = watchdog->parent;
@@ -535,6 +573,11 @@ int watchdog_dev_register(struct watchdog_device *watchdog)
 				pr_err("%s: a legacy watchdog module is probably present.\n",
 					watchdog->info->identity);
 			old_wdd = NULL;
+			if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+				watchdog_keepon_stop(watchdog);
+				watchdog_stop(watchdog);
+				module_put(watchdog->ops->owner);
+			}
 			return err;
 		}
 	}
@@ -553,6 +596,11 @@ int watchdog_dev_register(struct watchdog_device *watchdog)
 			misc_deregister(&watchdog_miscdev);
 			old_wdd = NULL;
 		}
+		if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+			watchdog_keepon_stop(watchdog);
+			watchdog_stop(watchdog);
+			module_put(watchdog->ops->owner);
+		}
 	}
 	return err;
 }
@@ -575,6 +623,12 @@ int watchdog_dev_unregister(struct watchdog_device *watchdog)
 		misc_deregister(&watchdog_miscdev);
 		old_wdd = NULL;
 	}
+
+	if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+		watchdog_keepon_stop(watchdog);
+		watchdog_stop(watchdog);
+		module_put(watchdog->ops->owner);
+	}
 	return 0;
 }
 
include/linux/watchdog.h +5 -0
index 2a3038e..650e0d5 100644
@@ -12,6 +12,7 @@
 #include <linux/bitops.h>
 #include <linux/device.h>
 #include <linux/cdev.h>
+#include <linux/timer.h>		/* for ping timer */
 #include <uapi/linux/watchdog.h>
 
 struct watchdog_ops;
@@ -95,6 +96,8 @@ struct watchdog_device {
 #define WDOG_ALLOW_RELEASE	2	/* Did we receive the magic char ? */
 #define WDOG_NO_WAY_OUT		3	/* Is 'nowayout' feature set ? */
 #define WDOG_UNREGISTERED	4	/* Has the device been unregistered */
+#define WDOG_KEEP_ON		5	/* Is 'keep on' feature set? */
+	struct timer_list ping_timer;	/* timer to keep on hardware ping */
 };
 
 #ifdef CONFIG_WATCHDOG_NOWAYOUT
@@ -104,6 +107,8 @@ struct watchdog_device {
 #define WATCHDOG_NOWAYOUT		0
 #define WATCHDOG_NOWAYOUT_INIT_STATUS	0
 #endif
+/* other proposal: WATCHDOG_ALWAYS_ACTIVE */
+#define WATCHDOG_KEEP_ON		(1 << WDOG_KEEP_ON)
 
 /* Use the following function to check whether or not the watchdog is active */
 static inline bool watchdog_active(struct watchdog_device *wdd)
Signature · 1 line
1.7.11.3
#1 lore raw

PATCH 2/6 watchdog: boottime protection feature (requires 'keep on')

After the boot time active watchdog(s) have to be maintained by userland.
The protection works only if boot time is non-zero and keep on feature is enabled.

---
 drivers/watchdog/watchdog_core.c |  4 ++++
 drivers/watchdog/watchdog_dev.c  | 11 +++++++++++
 2 files changed, 15 insertions(+)
drivers/watchdog/watchdog_core.c +4 -0
index cec9b55..13f3d3a 100644
@@ -43,6 +43,10 @@
 static DEFINE_IDA(watchdog_ida);
 static struct class *watchdog_class;
 
+unsigned int boottime = 0;
+module_param(boottime, uint, 0);
+MODULE_PARM_DESC(boottime, "After the boot time active watchdog(s) have to be maintained by userland. 0 disables the protection.");
+
 static void watchdog_check_min_max_timeout(struct watchdog_device *wdd)
 {
 	/*
drivers/watchdog/watchdog_dev.c +11 -0
index 51a65f6..d289eab 100644
@@ -42,6 +42,7 @@
 #include <linux/init.h>		/* For __init/__exit/... */
 #include <linux/uaccess.h>	/* For copy_to_user/put_user/... */
 #include <linux/jiffies.h>	/* for ping timer */
+#include <linux/time.h>		/* for uptime */
 
 #include "watchdog_core.h"
 
@@ -282,6 +283,16 @@ out_ioctl:
 static void watchdog_ping_timer_cb(unsigned long data)
 {
 	struct watchdog_device *wdd = (struct watchdog_device *)data;
+	extern unsigned int boottime;
+	if (boottime) {
+		struct timespec uptime;
+		get_monotonic_boottime(&uptime);
+		if (uptime.tv_sec > boottime) {
+			dev_info(wdd->dev,
+				 "Boottime for userland exceeded. Reboot expected!\n");
+			return;
+		}
+	}
 	watchdog_ping(wdd);
 	/* call next ping half the timeout value */
 	mod_timer(&wdd->ping_timer,
Signature · 1 line
1.7.11.3
#2 lore raw

PATCH 3/6 stmp3xxx_rtc_wdt: Add suspend/resume PM support

In reply to Janusz Uzycki

There is no conflict with rtc/rtc-stmp3xxx.c parent because modified registers in PM functions of stmp3xxx_rtc_wdt are different.

---
 drivers/watchdog/stmp3xxx_rtc_wdt.c | 24 +++++++++++++++++++++++
 1 file changed, 24 insertions(+)
drivers/watchdog/stmp3xxx_rtc_wdt.c +24 -0
index b4d6b34..77936b6 100644
@@ -95,9 +95,33 @@ static int stmp3xxx_wdt_remove(struct platform_device *pdev)
 	return 0;
 }
 
+static int __maybe_unused stmp3xxx_wdt_suspend(struct device *dev)
+{
+	struct watchdog_device *wdd = &stmp3xxx_wdd;
+
+	if (watchdog_active(wdd))
+		return wdt_stop(wdd);
+
+	return 0;
+}
+
+static int __maybe_unused stmp3xxx_wdt_resume(struct device *dev)
+{
+	struct watchdog_device *wdd = &stmp3xxx_wdd;
+
+	if (watchdog_active(wdd))
+		return wdt_start(wdd);
+
+	return 0;
+}
+
+static SIMPLE_DEV_PM_OPS(stmp3xxx_wdt_pm_ops,
+			 stmp3xxx_wdt_suspend, stmp3xxx_wdt_resume);
+
 static struct platform_driver stmp3xxx_wdt_driver = {
 	.driver = {
 		.name = "stmp3xxx_rtc_wdt",
+		.pm = &stmp3xxx_wdt_pm_ops,
 	},
 	.probe = stmp3xxx_wdt_probe,
 	.remove = stmp3xxx_wdt_remove,
Signature · 1 line
1.7.11.3
maintainer
#3 lore raw

PATCH 3/6 stmp3xxx_rtc_wdt: Add suspend/resume PM support

Quoted 5 lines · Janusz U?ycki

There is no conflict with rtc/rtc-stmp3xxx.c parent because modified registers in PM functions of stmp3xxx_rtc_wdt are different.

Quoted 42 lines
---
 drivers/watchdog/stmp3xxx_rtc_wdt.c | 24 +++++++++++++++++++++++
 1 file changed, 24 insertions(+)
drivers/watchdog/stmp3xxx_rtc_wdt.c +24 -0
index b4d6b34..77936b6 100644
@@ -95,9 +95,33 @@ static int stmp3xxx_wdt_remove(struct platform_device *pdev)
 	return 0;
 }
 
+static int __maybe_unused stmp3xxx_wdt_suspend(struct device *dev)
+{
+	struct watchdog_device *wdd = &stmp3xxx_wdd;
+
+	if (watchdog_active(wdd))
+		return wdt_stop(wdd);
+
+	return 0;
+}
+
+static int __maybe_unused stmp3xxx_wdt_resume(struct device *dev)
+{
+	struct watchdog_device *wdd = &stmp3xxx_wdd;
+
+	if (watchdog_active(wdd))
+		return wdt_start(wdd);
+
+	return 0;
+}
+
+static SIMPLE_DEV_PM_OPS(stmp3xxx_wdt_pm_ops,
+			 stmp3xxx_wdt_suspend, stmp3xxx_wdt_resume);
+
 static struct platform_driver stmp3xxx_wdt_driver = {
 	.driver = {
 		.name = "stmp3xxx_rtc_wdt",
+		.pm = &stmp3xxx_wdt_pm_ops,
 	},
 	.probe = stmp3xxx_wdt_probe,
 	.remove = stmp3xxx_wdt_remove,
#4 lore raw

PATCH 4/6 stmp3xxx_rtc_wdt: WATCHDOG_KEEP_ON enabled

In reply to Janusz Uzycki
---
 drivers/watchdog/stmp3xxx_rtc_wdt.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/watchdog/stmp3xxx_rtc_wdt.c +1 -1
index 77936b6..7e753b9 100644
@@ -67,7 +67,7 @@ static struct watchdog_device stmp3xxx_wdd = {
 	.ops = &stmp3xxx_wdt_ops,
 	.min_timeout = 1,
 	.max_timeout = STMP3XXX_MAX_TIMEOUT,
-	.status = WATCHDOG_NOWAYOUT_INIT_STATUS,
+	.status = WATCHDOG_NOWAYOUT_INIT_STATUS | WATCHDOG_KEEP_ON,
 };
 
 static int stmp3xxx_wdt_probe(struct platform_device *pdev)
Signature · 1 line
1.7.11.3
#5 lore raw

PATCH 5/6 watchdog: suspend/resume PM helper

In reply to Janusz Uzycki

The helper replaces watchdog_active() condition usually used in device-specific drivers for PM support. It also solves race between ping timer and driver-specific suspend function.

---
 drivers/watchdog/watchdog_dev.c | 38 +++++++++++++++++++++++++++
 include/linux/watchdog.h        |  4 +++
 2 files changed, 42 insertions(+)
drivers/watchdog/watchdog_dev.c +38 -0
index d289eab..2df88b7 100644
@@ -311,6 +311,44 @@ static void watchdog_keepon_stop(struct watchdog_device *wdd)
 }
 
 /*
+ *	watchdog_suspend: deactivate ping timer if enabled and
+ *			  stop the watchdog if active
+ *	@wdd: the watchdog device to do the suspend on
+ */
+void watchdog_suspend(struct watchdog_device *wdd)
+{
+	if (test_bit(WDOG_UNREGISTERED, &wdd->status))
+		return;
+
+	if (test_bit(WDOG_KEEP_ON, &wdd->status) &&
+	   !test_bit(WDOG_DEV_OPEN, &wdd->status))
+		del_timer_sync(&wdd->ping_timer);
+
+	if (watchdog_active(wdd))
+		wdd->ops->stop(wdd);
+}
+EXPORT_SYMBOL_GPL(watchdog_suspend);
+
+/*
+ *	watchdog_resume: start the watchdog if it was active
+ *			 and activate ping timer if it was enabled
+ *	@wdd: the watchdog device to do the resume on
+ */
+void watchdog_resume(struct watchdog_device *wdd)
+{
+	if (test_bit(WDOG_UNREGISTERED, &wdd->status))
+		return;
+
+	if (watchdog_active(wdd))
+		wdd->ops->start(wdd);
+
+	if (test_bit(WDOG_KEEP_ON, &wdd->status) &&
+	   !test_bit(WDOG_DEV_OPEN, &wdd->status))
+		watchdog_ping_timer_cb((unsigned long)wdd);
+}
+EXPORT_SYMBOL_GPL(watchdog_resume);
+
+/*
  *	watchdog_write: writes to the watchdog.
  *	@file: file from VFS
  *	@data: user address of data
include/linux/watchdog.h +4 -0
index 650e0d5..eee1b65 100644
@@ -147,4 +147,8 @@ extern int watchdog_init_timeout(struct watchdog_device *wdd,
 extern int watchdog_register_device(struct watchdog_device *);
 extern void watchdog_unregister_device(struct watchdog_device *);
 
+/* drivers/watchdog/watchdog_dev.c */
+extern void watchdog_suspend(struct watchdog_device *);
+extern void watchdog_resume(struct watchdog_device *);
+
 #endif  /* ifndef _LINUX_WATCHDOG_H */
Signature · 1 line
1.7.11.3
#6 lore raw

PATCH 6/6 stmp3xxx_rtc_wdt: use watchdog's PM helper

In reply to Janusz Uzycki

It also fixes PM race for WATCHDOG_KEEP_ON enabled.

---
 drivers/watchdog/stmp3xxx_rtc_wdt.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
drivers/watchdog/stmp3xxx_rtc_wdt.c +2 -4
index 7e753b9..0ba9edc 100644
@@ -99,8 +99,7 @@ static int __maybe_unused stmp3xxx_wdt_suspend(struct device *dev)
 {
 	struct watchdog_device *wdd = &stmp3xxx_wdd;
 
-	if (watchdog_active(wdd))
-		return wdt_stop(wdd);
+	watchdog_suspend(wdd);
 
 	return 0;
 }
@@ -109,8 +108,7 @@ static int __maybe_unused stmp3xxx_wdt_resume(struct device *dev)
 {
 	struct watchdog_device *wdd = &stmp3xxx_wdd;
 
-	if (watchdog_active(wdd))
-		return wdt_start(wdd);
+	watchdog_resume(wdd);
 
 	return 0;
 }
Signature · 1 line
1.7.11.3
maintainer
#7 lore raw

PATCH 1/6 watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature

In reply to Janusz Uzycki
Quoted 8 lines · Janusz Uzycki

Some applications require to start watchdog before userspace software. This patch enables such feature. Only the flag is necessary to enable it.
Moreover kernel's ping is re-enabled when userspace software closed watchdog using the magic character. The features improves kernel's reliability if hardware watchdog is available.

Hi Janusz,

This patch set is trying to solve four problems at once:

1) Auto-start watchdog when its driver registers
2) Keep watchdog running when its driver registers until userspace opens it
3) Handle watchdogs which can not be stopped after being started
4) Keep watchdog running with kernel timer after it has been closed, even if it can be stopped.

The next time adds 'boot time protection', which is really another term for an initial timeout, and case 5).

That is a bit too much for a single patch and, even more so, a single flag. Let's look at one case after another.

Auto-start watchdog when its driver registers - this makes sense as a feature just by itself. A good name for its flag might be something like WDT_AUTOSTART. A matching module parameter might also make sense.

autostart:
Set to 0 to disable, -1 to enable with unlimited timeout, or <n> for an initial timeout of <n> seconds.

This could be accompanied by a variable in watchdog_device:
int init_timeout; /* initial timeout in seconds */

An API function such as watchdog_set_autostart() with the initial timeout as parameter would also be helpful. This function could then be used to implement 2).

	if (autostart || (keep_running && this_watchdog_is_running())
		watchdog_set_autostart(&wdd, autostart ? : keep_running);

keep_running could then be a another module parameter with the same meaning as autostart.

Together this would also solve problem 5) while at the same time keeping the use cases separate.

For 3) we really need another flag. Actually, it might be sufficient to have watchdog drivers with this condition simply not provide a 'stop' function. If we use a flag, something like WDOG_HW_NO_WAY_OUT with matching WATCHDOG_HW_NOWAYOUT might make sense. Its functionality is slightly different to the other conditions: It would not auto-start a watchdog, but keep it running with the internal timer when the watchdog file is closed.

As for 4), I don't really know if it makes sense to have this functionality.

Does this all make sense ?

Some more comments below.

Thanks,
Guenter

Quoted 22 lines
---
  drivers/watchdog/watchdog_dev.c | 58 ++++++++++++++++++++++++++-
  include/linux/watchdog.h        |  5 +++
  2 files changed, 61 insertions(+), 2 deletions(-)
drivers/watchdog/watchdog_dev.c +2 -0
index 6aaefba..51a65f6 100644
@@ -41,6 +41,7 @@
  #include <linux/miscdevice.h>	/* For handling misc devices */
  #include <linux/init.h>		/* For __init/__exit/... */
  #include <linux/uaccess.h>	/* For copy_to_user/put_user/... */
+#include <linux/jiffies.h>	/* for ping timer */
 
  #include "watchdog_core.h"
 
@@ -277,6 +278,27 @@ out_ioctl:
  	return err;
  }
 
+/* 'keep on' feature */

drivers/watchdog/watchdog_dev.c out_ioctl: · line 291

Quoted 10 lines
drivers/watchdog/watchdog_dev.c +10 -0
+static void watchdog_ping_timer_cb(unsigned long data)
+{
+	struct watchdog_device *wdd = (struct watchdog_device *)data;
+	watchdog_ping(wdd);
+	/* call next ping half the timeout value */
+	mod_timer(&wdd->ping_timer,
+			jiffies + msecs_to_jiffies(wdd->timeout * 500));
+}
+
+static void watchdog_keepon_start(struct watchdog_device *wdd)

This name reflects the intended use case, but not the functionality. Something like "watchdog_timer_start" might make more sense here.

drivers/watchdog/watchdog_dev.c out_ioctl: · line 293

Quoted 2 lines
drivers/watchdog/watchdog_dev.c +2 -0
+{
+	watchdog_start(wdd);

This should probably be handled by the caller, to let us separate cases where the watchdog is already running (for example on close).

drivers/watchdog/watchdog_dev.c out_ioctl: · line 297

Quoted 4 lines
drivers/watchdog/watchdog_dev.c +4 -0
+	watchdog_ping_timer_cb((unsigned long)wdd);
+}
+
+static void watchdog_keepon_stop(struct watchdog_device *wdd)

Same name comment as above. keepon -> timer.

Quoted 20 lines
drivers/watchdog/watchdog_dev.c +7 -0
+{
+	del_timer_sync(&wdd->ping_timer);
+}
+
  /*
   *	watchdog_write: writes to the watchdog.
   *	@file: file from VFS
@@ -430,6 +452,9 @@ static int watchdog_open(struct inode *inode, struct file *file)
  	if (!try_module_get(wdd->ops->owner))
  		goto out;
 
+	if (test_bit(WDOG_KEEP_ON, &wdd->status))
+		watchdog_keepon_stop(wdd);
+
  	err = watchdog_start(wdd);
  	if (err < 0)
  		goto out_mod;
@@ -472,8 +497,13 @@ static int watchdog_release(struct inode *inode, struct file *file)
  	if (!test_bit(WDOG_ACTIVE, &wdd->status))
  		err = 0;

drivers/watchdog/watchdog_dev.c static int watchdog_release(struct inode *inode, struct file *file) · line 506

Quoted 10 lines
drivers/watchdog/watchdog_dev.c +7 -2
  	else if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
-		 !(wdd->info->options & WDIOF_MAGICCLOSE))
-		err = watchdog_stop(wdd);
+		 !(wdd->info->options & WDIOF_MAGICCLOSE)) {
+		if (test_bit(WDOG_KEEP_ON, &wdd->status)) {
+			watchdog_keepon_start(wdd);
+			err = 0;
+		} else
+			err = watchdog_stop(wdd);
+	}

I think this all should probably go into some helper function. The conditions here are getting a bit complicated.

Quoted 5 lines
drivers/watchdog/watchdog_dev.c +0 -0
  	/* If the watchdog was not stopped, send a keepalive ping */
  	if (err < 0) {
@@ -524,6 +554,14 @@ int watchdog_dev_register(struct watchdog_device *watchdog)
  {

drivers/watchdog/watchdog_dev.c int watchdog_dev_register(struct watchdog_device *watchdog) · line 564

Quoted 10 lines
drivers/watchdog/watchdog_dev.c +8 -0
  	int err, devno;
 
+	if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+		if (!try_module_get(watchdog->ops->owner))
+			return -ENODEV;
+		setup_timer(&watchdog->ping_timer, watchdog_ping_timer_cb,
+				(unsigned long)watchdog);
+		watchdog_keepon_start(watchdog);
+	}
+

This should be the last action in the probe function. Reason is that we might have a watchdog which can not be stopped after it was started once. If one of the error cases below happens, we would be stuck with a running watchdog and no means to stop it or to keep it alive.

drivers/watchdog/watchdog_dev.c int watchdog_dev_register(struct watchdog_device *watchdog) · line 580

Quoted 12 lines
drivers/watchdog/watchdog_dev.c +5 -0
  	if (watchdog->id == 0) {
  		old_wdd = watchdog;
  		watchdog_miscdev.parent = watchdog->parent;
@@ -535,6 +573,11 @@ int watchdog_dev_register(struct watchdog_device *watchdog)
  				pr_err("%s: a legacy watchdog module is probably present.\n",
  					watchdog->info->identity);
  			old_wdd = NULL;
+			if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+				watchdog_keepon_stop(watchdog);
+				watchdog_stop(watchdog);
+				module_put(watchdog->ops->owner);
+			}

Won't apply after moving above functions, but in general complex error cleanup like this should be handled with a goto to an error handler at the end of the function.

drivers/watchdog/watchdog_dev.c int watchdog_dev_register(struct watchdog_device *watchdog) · line 603

Quoted 12 lines
drivers/watchdog/watchdog_dev.c +5 -0
  			return err;
  		}
  	}
@@ -553,6 +596,11 @@ int watchdog_dev_register(struct watchdog_device *watchdog)
  			misc_deregister(&watchdog_miscdev);
  			old_wdd = NULL;
  		}
+		if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+			watchdog_keepon_stop(watchdog);
+			watchdog_stop(watchdog);
+			module_put(watchdog->ops->owner);
+		}

... to avoid situations with duplicated error handling code like this, but also to make code easier to read. See CodingStyle.

Quoted 47 lines
drivers/watchdog/watchdog_dev.c +6 -0
  	}
  	return err;
  }
@@ -575,6 +623,12 @@ int watchdog_dev_unregister(struct watchdog_device *watchdog)
  		misc_deregister(&watchdog_miscdev);
  		old_wdd = NULL;
  	}
+
+	if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+		watchdog_keepon_stop(watchdog);
+		watchdog_stop(watchdog);
+		module_put(watchdog->ops->owner);
+	}
  	return 0;
  }
 
include/linux/watchdog.h +5 -0
index 2a3038e..650e0d5 100644
@@ -12,6 +12,7 @@
  #include <linux/bitops.h>
  #include <linux/device.h>
  #include <linux/cdev.h>
+#include <linux/timer.h>		/* for ping timer */
  #include <uapi/linux/watchdog.h>
 
  struct watchdog_ops;
@@ -95,6 +96,8 @@ struct watchdog_device {
  #define WDOG_ALLOW_RELEASE	2	/* Did we receive the magic char ? */
  #define WDOG_NO_WAY_OUT		3	/* Is 'nowayout' feature set ? */
  #define WDOG_UNREGISTERED	4	/* Has the device been unregistered */
+#define WDOG_KEEP_ON		5	/* Is 'keep on' feature set? */
+	struct timer_list ping_timer;	/* timer to keep on hardware ping */
  };
 
  #ifdef CONFIG_WATCHDOG_NOWAYOUT
@@ -104,6 +107,8 @@ struct watchdog_device {
  #define WATCHDOG_NOWAYOUT		0
  #define WATCHDOG_NOWAYOUT_INIT_STATUS	0
  #endif
+/* other proposal: WATCHDOG_ALWAYS_ACTIVE */
+#define WATCHDOG_KEEP_ON		(1 << WDOG_KEEP_ON)
 
  /* Use the following function to check whether or not the watchdog is active */
  static inline bool watchdog_active(struct watchdog_device *wdd)
#8 lore raw

PATCH 1/6 watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature

W dniu 2014-09-26 06:01, Guenter Roeck pisze:

Quoted 10 lines · Janusz Uzycki
Quoted 8 lines

Some applications require to start watchdog before userspace software. This patch enables such feature. Only the flag is necessary to enable it.
Moreover kernel's ping is re-enabled when userspace software closed watchdog using the magic character. The features improves kernel's reliability if hardware watchdog is available.

Hi Guenter,

Quoted 10 lines

This patch set is trying to solve four problems at once:

1) Auto-start watchdog when its driver registers
2) Keep watchdog running when its driver registers until userspace opens it
3) Handle watchdogs which can not be stopped after being started
4) Keep watchdog running with kernel timer after it has been closed, even if it can be stopped.

Quoted 5 lines

The next time adds 'boot time protection', which is really another term for an initial timeout, and case 5).

That is a bit too much for a single patch and, even more so, a single flag.

OK, but I think [PATCH 3/6] could be applied.
Do you agree? Should I resent it separately?
I omited in the comment
"The patch adds suspend/resume PM support to stmp3xxx_rtc_wdt watchdog driver" because the subject is almost the same.

Quoted 9 lines

Let's look at one case after another.

Auto-start watchdog when its driver registers - this makes sense as a feature just by itself. A good name for its flag might be something like WDT_AUTOSTART. A matching module parameter might also make sense.

autostart:
Set to 0 to disable, -1 to enable with unlimited timeout, or <n> for an initial timeout of <n> seconds.

Current start(1) + keep-on(2,3,4) + boottime(5) combined. It looks OK.

Quoted 3 lines

This could be accompanied by a variable in watchdog_device:
int init_timeout; /* initial timeout in seconds */

As the module parameter, instead of "boottime" in watchdog_core?

Quoted 8 lines

An API function such as watchdog_set_autostart() with the initial timeout as parameter would also be helpful. This function could then be used to implement 2).

    if (autostart || (keep_running && this_watchdog_is_running())
        watchdog_set_autostart(&wdd, autostart ? : keep_running);

I don't understand the difference exactly and why to check the watchdog is running? This means watchdog is active or something new?

Quoted 3 lines

keep_running could then be a another module parameter with the same meaning
as autostart.

But autostart and keep_running aren't in conflict. So I don't understand also "autostart ? : keep_running".

Quoted 3 lines

Together this would also solve problem 5) while at the same time keeping the use cases separate.

It is solved by current code.

Quoted 5 lines

For 3) we really need another flag. Actually, it might be sufficient to have
watchdog drivers with this condition simply not provide a 'stop' function.

or use "NOT SUPPORTED" error code in stop, Stop could be called on register and new flag is set.

Quoted 2 lines

If we use a flag, something like WDOG_HW_NO_WAY_OUT with matching WATCHDOG_HW_NOWAYOUT might make sense. Its functionality is slightly

What is difference between WDOG_HW_NO_WAY_OUT
and WATCHDOG_HW_NOWAYOUT?

Quoted 6 lines

different to the other conditions: It would not auto-start a watchdog, but keep it running with the internal timer when the watchdog file is closed.

As for 4), I don't really know if it makes sense to have this functionality.

Yes, it is rootfs specific need. Script based code runs watchdog before critical function and after exit the watchdog using magic char. Critical section has timeout equal watchdog timeout value. The feature allow to avoid userland application for watchdog and does not cost much in the kernel.

Quoted 2 lines

Does this all make sense ?

I need more details because not all is clear for me.

Quoted 2 lines

Some more comments below.

thanks

Quoted 26 lines

Thanks,
Guenter

Quoted 22 lines
---
  drivers/watchdog/watchdog_dev.c | 58 ++++++++++++++++++++++++++-
  include/linux/watchdog.h        |  5 +++
  2 files changed, 61 insertions(+), 2 deletions(-)
drivers/watchdog/watchdog_dev.c +1 -0
index 6aaefba..51a65f6 100644
@@ -41,6 +41,7 @@
  #include <linux/miscdevice.h>    /* For handling misc devices */
  #include <linux/init.h>        /* For __init/__exit/... */
  #include <linux/uaccess.h>    /* For copy_to_user/put_user/... */
+#include <linux/jiffies.h>    /* for ping timer */
 
  #include "watchdog_core.h"
 
@@ -277,6 +278,27 @@ out_ioctl:
      return err;
  }
 
Quoted 14 lines

drivers/watchdog/watchdog_dev.c out_ioctl: · line 291

Quoted 11 lines
drivers/watchdog/watchdog_dev.c +11 -0
+/* 'keep on' feature */
+static void watchdog_ping_timer_cb(unsigned long data)
+{
+    struct watchdog_device *wdd = (struct watchdog_device *)data;
+    watchdog_ping(wdd);
+    /* call next ping half the timeout value */
+    mod_timer(&wdd->ping_timer,
+            jiffies + msecs_to_jiffies(wdd->timeout * 500));
+}
+
+static void watchdog_keepon_start(struct watchdog_device *wdd)

This name reflects the intended use case, but not the functionality. Something like "watchdog_timer_start" might make more sense here.

I see

Quoted 6 lines

drivers/watchdog/watchdog_dev.c out_ioctl: · line 293

Quoted 3 lines
drivers/watchdog/watchdog_dev.c +2 -0
+{
+    watchdog_start(wdd);

This should probably be handled by the caller, to let us separate cases where the watchdog is already running (for example on close).

The goal is to have watchdog always enabled.
This option could be disabled only if watchdog can't be stopped because then suspend is also impossible.

Quoted 7 lines

drivers/watchdog/watchdog_dev.c out_ioctl: · line 297

Quoted 5 lines
drivers/watchdog/watchdog_dev.c +4 -0
+    watchdog_ping_timer_cb((unsigned long)wdd);
+}
+
+static void watchdog_keepon_stop(struct watchdog_device *wdd)

Same name comment as above. keepon -> timer.

sure

Quoted 28 lines
Quoted 28 lines
drivers/watchdog/watchdog_dev.c +7 -0
+{
+    del_timer_sync(&wdd->ping_timer);
+}
+
  /*
   *    watchdog_write: writes to the watchdog.
   *    @file: file from VFS
@@ -430,6 +452,9 @@ static int watchdog_open(struct inode *inode, 
 struct file *file)
      if (!try_module_get(wdd->ops->owner))
          goto out;
 
+    if (test_bit(WDOG_KEEP_ON, &wdd->status))
+        watchdog_keepon_stop(wdd);
+
      err = watchdog_start(wdd);
      if (err < 0)
          goto out_mod;
drivers/watchdog/watchdog_dev.c +2 -2
@@ -472,8 +497,13 @@ static int watchdog_release(struct inode *inode, 
 struct file *file)
      if (!test_bit(WDOG_ACTIVE, &wdd->status))
          err = 0;
      else if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
-         !(wdd->info->options & WDIOF_MAGICCLOSE))
-        err = watchdog_stop(wdd);
+         !(wdd->info->options & WDIOF_MAGICCLOSE)) {
+        if (test_bit(WDOG_KEEP_ON, &wdd->status)) {
Quoted 8 lines

drivers/watchdog/watchdog_dev.c static int watchdog_release(struct inode *inode, struct file *file) · line 506

Quoted 5 lines
drivers/watchdog/watchdog_dev.c +5 -0
+            watchdog_keepon_start(wdd);
+            err = 0;
+        } else
+            err = watchdog_stop(wdd);
+    }

I think this all should probably go into some helper function. The conditions here are getting a bit complicated.

I will think about.

Quoted 9 lines
Quoted 9 lines
drivers/watchdog/watchdog_dev.c +0 -0
      /* If the watchdog was not stopped, send a keepalive ping */
      if (err < 0) {
@@ -524,6 +554,14 @@ int watchdog_dev_register(struct watchdog_device 
 *watchdog)
  {
      int err, devno;
 
Quoted 13 lines

drivers/watchdog/watchdog_dev.c int watchdog_dev_register(struct watchdog_device *watchdog) · line 564

Quoted 8 lines
drivers/watchdog/watchdog_dev.c +8 -0
+    if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+        if (!try_module_get(watchdog->ops->owner))
+            return -ENODEV;
+        setup_timer(&watchdog->ping_timer, watchdog_ping_timer_cb,
+                (unsigned long)watchdog);
+        watchdog_keepon_start(watchdog);
+    }
+

This should be the last action in the probe function. Reason is that we might have a watchdog which can not be stopped after it was started once. If one of the error cases below happens, we would be stuck with a running watchdog and no means to stop it or to keep it alive.

Right. But why not the last in the register function? Isn't the register function the last in the probe?

Quoted 11 lines
Quoted 11 lines
drivers/watchdog/watchdog_dev.c +0 -0
      if (watchdog->id == 0) {
          old_wdd = watchdog;
          watchdog_miscdev.parent = watchdog->parent;
drivers/watchdog/watchdog_dev.c +1 -0
@@ -535,6 +573,11 @@ int watchdog_dev_register(struct watchdog_device 
 *watchdog)
                  pr_err("%s: a legacy watchdog module is probably 
 present.\n",
                      watchdog->info->identity);
              old_wdd = NULL;
+            if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
Quoted 8 lines

drivers/watchdog/watchdog_dev.c int watchdog_dev_register(struct watchdog_device *watchdog) · line 580

Quoted 4 lines
drivers/watchdog/watchdog_dev.c +4 -0
+                watchdog_keepon_stop(watchdog);
+                watchdog_stop(watchdog);
+                module_put(watchdog->ops->owner);
+            }

Won't apply after moving above functions, but in general complex error cleanup like this should be handled with a goto to an error handler at the end of the function.

OK

Quoted 17 lines
Quoted 5 lines
drivers/watchdog/watchdog_dev.c +0 -0
              return err;
          }
      }

drivers/watchdog/watchdog_dev.c int watchdog_dev_register(struct watchdog_device · line 604

Quoted 9 lines
drivers/watchdog/watchdog_dev.c +5 -0
@@ -553,6 +596,11 @@ int watchdog_dev_register(struct watchdog_device 
 *watchdog)
              misc_deregister(&watchdog_miscdev);
              old_wdd = NULL;
          }
+        if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+            watchdog_keepon_stop(watchdog);
+            watchdog_stop(watchdog);
+            module_put(watchdog->ops->owner);
+        }

... to avoid situations with duplicated error handling code like this, but also to make code easier to read. See CodingStyle.

of course

Quoted 54 lines
Quoted 53 lines
drivers/watchdog/watchdog_dev.c +0 -0
      }
      return err;
  }
drivers/watchdog/watchdog_dev.c +6 -0
@@ -575,6 +623,12 @@ int watchdog_dev_unregister(struct 
 watchdog_device *watchdog)
          misc_deregister(&watchdog_miscdev);
          old_wdd = NULL;
      }
+
+    if (test_bit(WDOG_KEEP_ON, &watchdog->status)) {
+        watchdog_keepon_stop(watchdog);
+        watchdog_stop(watchdog);
+        module_put(watchdog->ops->owner);
+    }
      return 0;
  }
include/linux/watchdog.h +1 -0
index 2a3038e..650e0d5 100644
@@ -12,6 +12,7 @@
  #include <linux/bitops.h>
  #include <linux/device.h>
  #include <linux/cdev.h>
+#include <linux/timer.h>        /* for ping timer */
  #include <uapi/linux/watchdog.h>
 
  struct watchdog_ops;
@@ -95,6 +96,8 @@ struct watchdog_device {
  #define WDOG_ALLOW_RELEASE    2    /* Did we receive the magic char 

? */
#define WDOG_NO_WAY_OUT 3 /* Is 'nowayout' feature set ? */
#define WDOG_UNREGISTERED 4 /* Has the device been
unregistered */

include/linux/watchdog.h +2 -0
+#define WDOG_KEEP_ON        5    /* Is 'keep on' feature set? */
+    struct timer_list ping_timer;    /* timer to keep on hardware 

ping */
};

  #ifdef CONFIG_WATCHDOG_NOWAYOUT
include/linux/watchdog.h +2 -0
@@ -104,6 +107,8 @@ struct watchdog_device {
  #define WATCHDOG_NOWAYOUT        0
  #define WATCHDOG_NOWAYOUT_INIT_STATUS    0
  #endif
+/* other proposal: WATCHDOG_ALWAYS_ACTIVE */
+#define WATCHDOG_KEEP_ON        (1 << WDOG_KEEP_ON)
 
  /* Use the following function to check whether or not the watchdog 

is active */
static inline bool watchdog_active(struct watchdog_device *wdd)

maintainer
#9 lore raw

PATCH 1/6 watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature

Quoted 24 lines · Janusz U?ycki

W dniu 2014-09-26 06:01, Guenter Roeck pisze:

Quoted 10 lines · Janusz Uzycki
Quoted 8 lines

Some applications require to start watchdog before userspace software. This patch enables such feature. Only the flag is necessary to enable it.
Moreover kernel's ping is re-enabled when userspace software closed watchdog using the magic character. The features improves kernel's reliability if hardware watchdog is available.

Hi Guenter,

Quoted 9 lines

This patch set is trying to solve four problems at once:

1) Auto-start watchdog when its driver registers
2) Keep watchdog running when its driver registers until userspace opens it
3) Handle watchdogs which can not be stopped after being started
4) Keep watchdog running with kernel timer after it has been closed, even if it can be stopped.

Quoted 7 lines
Quoted 4 lines

The next time adds 'boot time protection', which is really another term for an initial timeout, and case 5).

That is a bit too much for a single patch and, even more so, a single flag.

OK, but I think [PATCH 3/6] could be applied.
Do you agree? Should I resent it separately?

Yes, it looks ok.

Quoted 10 lines

I omited in the comment
"The patch adds suspend/resume PM support to stmp3xxx_rtc_wdt watchdog driver" because the subject is almost the same.

Quoted 6 lines

Let's look at one case after another.

Auto-start watchdog when its driver registers - this makes sense as a feature just by itself. A good name for its flag might be something like WDT_AUTOSTART. A matching module parameter might also make sense.

Quoted 6 lines
Quoted 3 lines

autostart:
Set to 0 to disable, -1 to enable with unlimited timeout, or <n> for an initial timeout of <n> seconds.

Current start(1) + keep-on(2,3,4) + boottime(5) combined. It looks OK.

Maybe for you. For me they are different cases.

Quoted 14 lines
Quoted 3 lines

This could be accompanied by a variable in watchdog_device:
int init_timeout; /* initial timeout in seconds */

As the module parameter, instead of "boottime" in watchdog_core?

Quoted 8 lines

An API function such as watchdog_set_autostart() with the initial timeout as parameter would also be helpful. This function could then be used to implement 2).

    if (autostart || (keep_running && this_watchdog_is_running())
        watchdog_set_autostart(&wdd, autostart ? : keep_running);
Quoted 8 lines

I don't understand the difference exactly and why to check the watchdog is running? This means watchdog is active or something new?

Quoted 2 lines

keep_running could then be a another module parameter with the same meaning as autostart.

But autostart and keep_running aren't in conflict. So I don't understand also "autostart ? : keep_running".

The functionality is distinctively different.

autostart: start watchdog on module load
keep_running: If the watchdog is already running, keep it running. Otherwise do nothing.

Both have different use cases which should not be combined.

Quoted 4 lines
Quoted 3 lines

Together this would also solve problem 5) while at the same time keeping the use cases separate.

Quoted 8 lines

It is solved by current code.

Quoted 3 lines

For 3) we really need another flag. Actually, it might be sufficient to have watchdog drivers with this condition simply not provide a 'stop' function.

or use "NOT SUPPORTED" error code in stop, Stop could be called on register and new flag is set.

Seems to add complexity for no real benefit. Please explain why you think it is a good idea to have multiple drivers implement the same function just to return an error and do nothing else,

Quoted 6 lines
Quoted 2 lines

If we use a flag, something like WDOG_HW_NO_WAY_OUT with matching WATCHDOG_HW_NOWAYOUT might make sense. Its functionality is slightly

What is difference between WDOG_HW_NO_WAY_OUT
and WATCHDOG_HW_NOWAYOUT?

Similar to other flags

#define WDOG_HW_NO_WAY_OUT 5
#define WATCHDOG_HW_NOWAYOUT (1 << WDOG_HW_NO_WAY_OUT)

Quoted 11 lines
Quoted 4 lines

different to the other conditions: It would not auto-start a watchdog, but keep it running with the internal timer when the watchdog file is closed.

As for 4), I don't really know if it makes sense to have this functionality.

Yes, it is rootfs specific need. Script based code runs watchdog before critical function and after exit the watchdog using magic char. Critical section has timeout equal watchdog timeout value. The feature allow to avoid userland application for watchdog and does not cost much in the kernel.

Sorry, I can not follow your logic here. A basic userland implementation doesn't cost much either, is much safer, and even init systems such as systemd implement it nowadays.

Guenter

#10 lore raw

PATCH 1/6 watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature

W dniu 2014-09-30 06:37, Guenter Roeck pisze:

Quoted 14 lines · Janusz U?ycki
Quoted 13 lines
Quoted 13 lines

This patch set is trying to solve four problems at once:

1) Auto-start watchdog when its driver registers
2) Keep watchdog running when its driver registers until userspace opens it
3) Handle watchdogs which can not be stopped after being started
4) Keep watchdog running with kernel timer after it has been closed, even if it can be stopped.

The next time adds 'boot time protection', which is really another term for an initial timeout, and case 5).

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

That is a bit too much for a single patch and, even more so, a single flag.

OK, but I think [PATCH 3/6] could be applied.
Do you agree? Should I resent it separately?

Yes, it looks ok.

Can you apply it?

Quoted 12 lines
Quoted 12 lines

I omited in the comment
"The patch adds suspend/resume PM support to stmp3xxx_rtc_wdt watchdog driver" because the subject is almost the same.

Quoted 7 lines

Let's look at one case after another.

Auto-start watchdog when its driver registers - this makes sense as a feature just by itself. A good name for its flag might be something like
WDT_AUTOSTART. A matching module parameter might also make sense.

Quoted 7 lines
Quoted 6 lines
Quoted 3 lines

autostart:
Set to 0 to disable, -1 to enable with unlimited timeout, or <n> for an initial timeout of <n> seconds.

Current start(1) + keep-on(2,3,4) + boottime(5) combined. It looks OK.

Maybe for you. For me they are different cases.

They are different. I missed some words in the first sentence :) However autostart automatically combine them together again :) The problem is common timer so the patches are dependent. There is no reason to use more timers.

Quoted 26 lines
Quoted 16 lines
Quoted 4 lines

This could be accompanied by a variable in watchdog_device:
int init_timeout; /* initial timeout in seconds */

As the module parameter, instead of "boottime" in watchdog_core?

Quoted 9 lines

An API function such as watchdog_set_autostart() with the initial timeout
as parameter would also be helpful. This function could then be used to implement 2).

    if (autostart || (keep_running && this_watchdog_is_running())
        watchdog_set_autostart(&wdd, autostart ? : keep_running);
Quoted 9 lines

I don't understand the difference exactly and why to check the watchdog is running? This means watchdog is active or something new?

Quoted 3 lines

keep_running could then be a another module parameter with the same meaning
as autostart.

But autostart and keep_running aren't in conflict. So I don't understand also "autostart ? : keep_running".

Quoted 7 lines

The functionality is distinctively different.

autostart: start watchdog on module load
keep_running: If the watchdog is already running, keep it running. Otherwise do nothing.

Both have different use cases which should not be combined.

according to autostart value:
-1: current "keep-on" feature

Quoted 1 line

0: current "boottime" feature

Does watchdog_set_autostart() start/activate watchog?
Does keep_running differ from autostart=-1 "this_watchdog_is_running()" only?
How/where is this_watchdog_is_running() implemented?

What type is keep_running?
* bool
* 0/1
* -1/0 for watchdog_set_autostart()?

Quoted 13 lines
Quoted 13 lines
Quoted 5 lines

Together this would also solve problem 5) while at the same time keeping
the use cases separate.

It is solved by current code.

Quoted 5 lines

For 3) we really need another flag. Actually, it might be sufficient to have
watchdog drivers with this condition simply not provide a 'stop' function.

Quoted 7 lines
Quoted 3 lines

or use "NOT SUPPORTED" error code in stop, Stop could be called on register and new flag is set.

Seems to add complexity for no real benefit. Please explain why you think it is a good idea to have multiple drivers implement the same function just to return an error and do nothing else,

Specific driver knows if the watchdog is stoppable. wddev->ops->stop() is called from watchdog_dev and if the stop() returns an error
watchdog_dev prints warning. It assumes wddev->ops->stop() is always implemented
even if not supported, ie. there is no condition. So either ENOSUP can be used
or the condition for optional (not mandatory) wddev->ops->stop(). Today a lot of drivers returns 0 only instead of ENOSUP.
So stop() could be called before autostart() to set the flag (code complexity)
or as you wrote just use the flag directly. The last one is indeed better.

Quoted 11 lines
Quoted 7 lines
Quoted 3 lines

If we use a flag, something like WDOG_HW_NO_WAY_OUT with matching WATCHDOG_HW_NOWAYOUT might make sense. Its functionality is slightly

What is difference between WDOG_HW_NO_WAY_OUT
and WATCHDOG_HW_NOWAYOUT?

Similar to other flags

#define WDOG_HW_NO_WAY_OUT 5
#define WATCHDOG_HW_NOWAYOUT (1 << WDOG_HW_NO_WAY_OUT)

oh, right

Quoted 14 lines
Quoted 5 lines
Quoted 5 lines

different to the other conditions: It would not auto-start a watchdog, but keep it running with the internal timer when the watchdog file is closed.

Quoted 9 lines
Quoted 2 lines

As for 4), I don't really know if it makes sense to have this functionality.

Yes, it is rootfs specific need. Script based code runs watchdog before critical function and after exit the watchdog using magic char. Critical section has timeout equal watchdog timeout value. The feature allow to avoid userland application for watchdog and does not cost much in the kernel.

Quoted 3 lines

Sorry, I can not follow your logic here. A basic userland implementation doesn't cost much either, is much safer, and even init systems such as systemd implement it nowadays.

True but not always for embedded systems where systemd is overweight today. Let's notice that 4) works only if magic char is sent on exit. Thanks to 4) magic char works also for non-stoppable watchdogs.

Janusz

maintainer
#11 lore raw

PATCH 1/6 watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature

Quoted 15 lines · Janusz U?ycki

W dniu 2014-09-30 06:37, Guenter Roeck pisze:

Quoted 13 lines · Janusz U?ycki
Quoted 12 lines
Quoted 12 lines

This patch set is trying to solve four problems at once:

1) Auto-start watchdog when its driver registers
2) Keep watchdog running when its driver registers until userspace opens it
3) Handle watchdogs which can not be stopped after being started
4) Keep watchdog running with kernel timer after it has been closed, even if it can be stopped.

The next time adds 'boot time protection', which is really another term for an initial timeout, and case 5).

Quoted 9 lines
Quoted 6 lines
Quoted 4 lines
Quoted 1 line

That is a bit too much for a single patch and, even more so, a single flag.

OK, but I think [PATCH 3/6] could be applied.
Do you agree? Should I resent it separately?

Yes, it looks ok.

Can you apply it?

I don't apply watchdog patches; I only provide review feedback.

Quoted 15 lines
Quoted 15 lines
Quoted 15 lines

I omited in the comment
"The patch adds suspend/resume PM support to stmp3xxx_rtc_wdt watchdog driver" because the subject is almost the same.

Quoted 9 lines

Let's look at one case after another.

Auto-start watchdog when its driver registers - this makes sense as a feature just by itself. A good name for its flag might be something like WDT_AUTOSTART. A matching module parameter might also make sense.

autostart:
Set to 0 to disable, -1 to enable with unlimited timeout, or <n> for an initial timeout of <n> seconds.

Quoted 9 lines
Quoted 3 lines
Quoted 2 lines

Current start(1) + keep-on(2,3,4) + boottime(5) combined. It looks OK.

Maybe for you. For me they are different cases.

They are different. I missed some words in the first sentence :) However autostart automatically combine them together again :) The problem is common timer so the patches are dependent. There is no reason to use more timers.

I did not say use multiple timers. I said use multiple flags.

Guenter

#12 lore raw

PATCH 1/6 watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature

In reply to Guenter Roeck

TODO: change WDOG_KEEP_ON to autostart module parameter

changelog:
[PATCH 1/6] watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature
* clean up old comments
* watchdog_keepon_start() renamed to watchdog_timer_start()
* watchdog_keepon_stop() renamed to watchdog_timer_stop()
* watchdog_timer_register() added
* watchdog_timer_unregister() added
* watchdog_timer_register() is the last action
in watchdog_dev_register().
FIXME: Really should be in probe()?
* TODO: should watchdog_timer_start() call watchdog_start()?
* watchdog_release() uses watchdog_timer_restart() helper function

#13 lore raw

PATCH 1/6 watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature

Some applications require to start watchdog before userspace software. This patch enables such feature. Only the flag is necessary to enable it.
Moreover kernel's ping is re-enabled when userspace software closed watchdog using the magic character. The features improves kernel's reliability if hardware watchdog is available.

---
 drivers/watchdog/watchdog_dev.c | 65 ++++++++++++++++++++++++++-
 include/linux/watchdog.h        |  4 ++
 2 files changed, 68 insertions(+), 1 deletion(-)
drivers/watchdog/watchdog_dev.c +64 -1
index 6aaefba..afc14f7 100644
@@ -41,6 +41,7 @@
 #include <linux/miscdevice.h>	/* For handling misc devices */
 #include <linux/init.h>		/* For __init/__exit/... */
 #include <linux/uaccess.h>	/* For copy_to_user/put_user/... */
+#include <linux/jiffies.h>	/* for ping timer */
 
 #include "watchdog_core.h"
 
@@ -277,6 +278,60 @@ out_ioctl:
 	return err;
 }
 
+static void watchdog_ping_timer_cb(unsigned long data)
+{
+	struct watchdog_device *wdd = (struct watchdog_device *)data;
+	watchdog_ping(wdd);
+	/* call next ping half the timeout value */
+	mod_timer(&wdd->ping_timer,
+			jiffies + msecs_to_jiffies(wdd->timeout * 500));
+}
+
+static void watchdog_timer_start(struct watchdog_device *wdd)
+{
+	/* TODO: watchdog_start():
+	 * it should probably be handled by the caller, to let us separate
+	 * cases where the watchdog is already running (for example on close). */
+	watchdog_start(wdd);
+	watchdog_ping_timer_cb((unsigned long)wdd);
+}
+
+static void watchdog_timer_stop(struct watchdog_device *wdd)
+{
+	del_timer_sync(&wdd->ping_timer);
+}
+
+static int watchdog_timer_restart(struct watchdog_device *wdd)
+{
+	if (test_bit(WDOG_KEEP_ON, &wdd->status)) {
+		watchdog_timer_start(wdd);
+		return 0;
+	}
+	return watchdog_stop(wdd);
+}
+
+static int watchdog_timer_register(struct watchdog_device *wdd)
+{
+	if (test_bit(WDOG_KEEP_ON, &wdd->status)) {
+		if (!try_module_get(wdd->ops->owner))
+			return -ENODEV;
+		setup_timer(&wdd->ping_timer, watchdog_ping_timer_cb,
+				(unsigned long)wdd);
+		watchdog_timer_start(wdd);
+	}
+	return 0;
+}
+
+static int watchdog_timer_unregister(struct watchdog_device *wdd)
+{
+	if (test_bit(WDOG_KEEP_ON, &wdd->status)) {
+		watchdog_timer_stop(wdd);
+		watchdog_stop(wdd);
+		module_put(wdd->ops->owner);
+	}
+	return 0;
+}
+
 /*
  *	watchdog_write: writes to the watchdog.
  *	@file: file from VFS
@@ -430,6 +485,9 @@ static int watchdog_open(struct inode *inode, struct file *file)
 	if (!try_module_get(wdd->ops->owner))
 		goto out;
 
+	if (test_bit(WDOG_KEEP_ON, &wdd->status))
+		watchdog_timer_stop(wdd);
+
 	err = watchdog_start(wdd);
 	if (err < 0)
 		goto out_mod;
@@ -473,7 +531,7 @@ static int watchdog_release(struct inode *inode, struct file *file)
 		err = 0;
 	else if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
 		 !(wdd->info->options & WDIOF_MAGICCLOSE))
-		err = watchdog_stop(wdd);
+		err = watchdog_timer_restart(wdd);
 
 	/* If the watchdog was not stopped, send a keepalive ping */
 	if (err < 0) {
@@ -553,7 +611,10 @@ int watchdog_dev_register(struct watchdog_device *watchdog)
 			misc_deregister(&watchdog_miscdev);
 			old_wdd = NULL;
 		}
+		return err;
 	}
+
+	err = watchdog_timer_register(watchdog);
 	return err;
 }
 
@@ -575,6 +636,8 @@ int watchdog_dev_unregister(struct watchdog_device *watchdog)
 		misc_deregister(&watchdog_miscdev);
 		old_wdd = NULL;
 	}
+
+	watchdog_timer_unregister(watchdog);
 	return 0;
 }
 
include/linux/watchdog.h +4 -0
index 2a3038e..0afeabd 100644
@@ -12,6 +12,7 @@
 #include <linux/bitops.h>
 #include <linux/device.h>
 #include <linux/cdev.h>
+#include <linux/timer.h>		/* for ping timer */
 #include <uapi/linux/watchdog.h>
 
 struct watchdog_ops;
@@ -95,6 +96,8 @@ struct watchdog_device {
 #define WDOG_ALLOW_RELEASE	2	/* Did we receive the magic char ? */
 #define WDOG_NO_WAY_OUT		3	/* Is 'nowayout' feature set ? */
 #define WDOG_UNREGISTERED	4	/* Has the device been unregistered */
+#define WDOG_KEEP_ON		5	/* Is 'keep on' feature set? */
+	struct timer_list ping_timer;	/* timer to keep on hardware ping */
 };
 
 #ifdef CONFIG_WATCHDOG_NOWAYOUT
@@ -104,6 +107,7 @@ struct watchdog_device {
 #define WATCHDOG_NOWAYOUT		0
 #define WATCHDOG_NOWAYOUT_INIT_STATUS	0
 #endif
+#define WATCHDOG_KEEP_ON		(1 << WDOG_KEEP_ON)
 
 /* Use the following function to check whether or not the watchdog is active */
 static inline bool watchdog_active(struct watchdog_device *wdd)
Signature · 1 line
1.7.11.3
#14 lore raw

PATCH 1/6 watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature

I sent also the changelog:
[PATCH 1/6] watchdog: watchdog_dev: WATCHDOG_KEEP_ON feature
* clean up old comments
* watchdog_keepon_start() renamed to watchdog_timer_start()
* watchdog_keepon_stop() renamed to watchdog_timer_stop()
* watchdog_timer_register() added
* watchdog_timer_unregister() added
* watchdog_timer_register() is the last action
in watchdog_dev_register().
FIXME: Really should be in probe()?
* TODO: should watchdog_timer_start() call watchdog_start()?
* watchdog_release() uses watchdog_timer_restart() helper function

* TODO: change WDOG_KEEP_ON to autostart module parameter Here I am still not sure about module parameters.

Quoted 3 lines

autostart:
Set to 0 to disable, -1 to enable with unlimited timeout, or <n> for an initial timeout of <n> seconds.

[1/6]: support of 0 and -1, where -1 is instead of WDOG_KEEP_ON? [2/6]: support of <n> instead of boottime

what about:
* int init_timeout; /* initial timeout in seconds */
It looks as <n> so why duplicate the autostart parameter?
* keep_running
It looks like -1 without watchdog_start()
* this_watchdog_is_running()
Where defined?
* watchdog_set_autostart()
How to split new patches?

Janusz

W dniu 2014-09-30 14:46, Janusz Uzycki pisze:

Quoted 154 lines

Some applications require to start watchdog before userspace software. This patch enables such feature. Only the flag is necessary to enable it.
Moreover kernel's ping is re-enabled when userspace software closed watchdog using the magic character. The features improves kernel's reliability if hardware watchdog is available.

---
  drivers/watchdog/watchdog_dev.c | 65 ++++++++++++++++++++++++++-
  include/linux/watchdog.h        |  4 ++
  2 files changed, 68 insertions(+), 1 deletion(-)
drivers/watchdog/watchdog_dev.c +64 -1
index 6aaefba..afc14f7 100644
@@ -41,6 +41,7 @@
  #include <linux/miscdevice.h>	/* For handling misc devices */
  #include <linux/init.h>		/* For __init/__exit/... */
  #include <linux/uaccess.h>	/* For copy_to_user/put_user/... */
+#include <linux/jiffies.h>	/* for ping timer */
 
  #include "watchdog_core.h"
 
@@ -277,6 +278,60 @@ out_ioctl:
  	return err;
  }
 
+static void watchdog_ping_timer_cb(unsigned long data)
+{
+	struct watchdog_device *wdd = (struct watchdog_device *)data;
+	watchdog_ping(wdd);
+	/* call next ping half the timeout value */
+	mod_timer(&wdd->ping_timer,
+			jiffies + msecs_to_jiffies(wdd->timeout * 500));
+}
+
+static void watchdog_timer_start(struct watchdog_device *wdd)
+{
+	/* TODO: watchdog_start():
+	 * it should probably be handled by the caller, to let us separate
+	 * cases where the watchdog is already running (for example on close). */
+	watchdog_start(wdd);
+	watchdog_ping_timer_cb((unsigned long)wdd);
+}
+
+static void watchdog_timer_stop(struct watchdog_device *wdd)
+{
+	del_timer_sync(&wdd->ping_timer);
+}
+
+static int watchdog_timer_restart(struct watchdog_device *wdd)
+{
+	if (test_bit(WDOG_KEEP_ON, &wdd->status)) {
+		watchdog_timer_start(wdd);
+		return 0;
+	}
+	return watchdog_stop(wdd);
+}
+
+static int watchdog_timer_register(struct watchdog_device *wdd)
+{
+	if (test_bit(WDOG_KEEP_ON, &wdd->status)) {
+		if (!try_module_get(wdd->ops->owner))
+			return -ENODEV;
+		setup_timer(&wdd->ping_timer, watchdog_ping_timer_cb,
+				(unsigned long)wdd);
+		watchdog_timer_start(wdd);
+	}
+	return 0;
+}
+
+static int watchdog_timer_unregister(struct watchdog_device *wdd)
+{
+	if (test_bit(WDOG_KEEP_ON, &wdd->status)) {
+		watchdog_timer_stop(wdd);
+		watchdog_stop(wdd);
+		module_put(wdd->ops->owner);
+	}
+	return 0;
+}
+
  /*
   *	watchdog_write: writes to the watchdog.
   *	@file: file from VFS
@@ -430,6 +485,9 @@ static int watchdog_open(struct inode *inode, struct file *file)
  	if (!try_module_get(wdd->ops->owner))
  		goto out;
 
+	if (test_bit(WDOG_KEEP_ON, &wdd->status))
+		watchdog_timer_stop(wdd);
+
  	err = watchdog_start(wdd);
  	if (err < 0)
  		goto out_mod;
@@ -473,7 +531,7 @@ static int watchdog_release(struct inode *inode, struct file *file)
  		err = 0;
  	else if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
  		 !(wdd->info->options & WDIOF_MAGICCLOSE))
-		err = watchdog_stop(wdd);
+		err = watchdog_timer_restart(wdd);
 
  	/* If the watchdog was not stopped, send a keepalive ping */
  	if (err < 0) {
@@ -553,7 +611,10 @@ int watchdog_dev_register(struct watchdog_device *watchdog)
  			misc_deregister(&watchdog_miscdev);
  			old_wdd = NULL;
  		}
+		return err;
  	}
+
+	err = watchdog_timer_register(watchdog);
  	return err;
  }
 
@@ -575,6 +636,8 @@ int watchdog_dev_unregister(struct watchdog_device *watchdog)
  		misc_deregister(&watchdog_miscdev);
  		old_wdd = NULL;
  	}
+
+	watchdog_timer_unregister(watchdog);
  	return 0;
  }
 
include/linux/watchdog.h +4 -0
index 2a3038e..0afeabd 100644
@@ -12,6 +12,7 @@
  #include <linux/bitops.h>
  #include <linux/device.h>
  #include <linux/cdev.h>
+#include <linux/timer.h>		/* for ping timer */
  #include <uapi/linux/watchdog.h>
 
  struct watchdog_ops;
@@ -95,6 +96,8 @@ struct watchdog_device {
  #define WDOG_ALLOW_RELEASE	2	/* Did we receive the magic char ? */
  #define WDOG_NO_WAY_OUT		3	/* Is 'nowayout' feature set ? */
  #define WDOG_UNREGISTERED	4	/* Has the device been unregistered */
+#define WDOG_KEEP_ON		5	/* Is 'keep on' feature set? */
+	struct timer_list ping_timer;	/* timer to keep on hardware ping */
  };
 
  #ifdef CONFIG_WATCHDOG_NOWAYOUT
@@ -104,6 +107,7 @@ struct watchdog_device {
  #define WATCHDOG_NOWAYOUT		0
  #define WATCHDOG_NOWAYOUT_INIT_STATUS	0
  #endif
+#define WATCHDOG_KEEP_ON		(1 << WDOG_KEEP_ON)
 
  /* Use the following function to check whether or not the watchdog is active */
  static inline bool watchdog_active(struct watchdog_device *wdd)