Skip to contentSkip to conversation

watchdog: core: don't try to stop device if not running

RFC PATCH Superseded Hector Palacios 2 participants 5 messages Apr 5, 2013 – Apr 8, 2013 Watchdog device drivers

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

RFC PATCH 1/1 watchdog: core: don't try to stop device if not running

A watchdog device may be stopped from userspace using WDIOC_SETOPTIONS ioctl and flag WDIOS_DISABLECARD. If the device is closed after this operation, watchdog_release() is called and status bits checked for stopping it. Besides, if the device has not been unregistered a critical message "watchdog did not stop!" is printed, although the ioctl may have successfully stopped it already.

Without the patch a user application sample code like this will successfully stop the watchdog, but the kernel will output the message "watchdog did not stop!":

	wd_fd = open("/dev/watchdog", O_RDWR);

	flags = WDIOS_DISABLECARD;
	ioctl(wd_fd, WDIOC_SETOPTIONS, &flags);

	close(wd_fd);
---
 drivers/watchdog/watchdog_dev.c |   38 +++++++++++++++++++++-----------------
 1 file changed, 21 insertions(+), 17 deletions(-)
drivers/watchdog/watchdog_dev.c +21 -17
index ef8edec..adadbdd 100644
@@ -463,24 +463,28 @@ out:
 static int watchdog_release(struct inode *inode, struct file *file)
 {
 	struct watchdog_device *wdd = file->private_data;
-	int err = -EBUSY;
 
-	/*
-	 * We only stop the watchdog if we received the magic character
-	 * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then
-	 * watchdog_stop will fail.
-	 */
-	if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
-	    !(wdd->info->options & WDIOF_MAGICCLOSE))
-		err = watchdog_stop(wdd);
-
-	/* If the watchdog was not stopped, send a keepalive ping */
-	if (err < 0) {
-		mutex_lock(&wdd->lock);
-		if (!test_bit(WDOG_UNREGISTERED, &wdd->status))
-			dev_crit(wdd->dev, "watchdog did not stop!\n");
-		mutex_unlock(&wdd->lock);
-		watchdog_ping(wdd);
+	if (test_bit(WDOG_ACTIVE, &wdd->status)) {
+		int err = -EBUSY;
+
+		/*
+		 * We only stop the watchdog if we received the magic character
+		 * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then
+		 * watchdog_stop will fail.
+		 */
+		if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
+		    !(wdd->info->options & WDIOF_MAGICCLOSE)) {
+			err = watchdog_stop(wdd);
+		}
+
+		/* If the watchdog was not stopped, send a keepalive ping */
+		if (err < 0) {
+			mutex_lock(&wdd->lock);
+			if (!test_bit(WDOG_UNREGISTERED, &wdd->status))
+				dev_crit(wdd->dev, "watchdog did not stop!\n");
+			mutex_unlock(&wdd->lock);
+			watchdog_ping(wdd);
+		}
 	}
 
 	/* Allow the owner module to be unloaded again */
Signature · 1 line
1.7.9.5
maintainer
#1 lore raw
Quoted 11 lines · Hector Palacios

A watchdog device may be stopped from userspace using WDIOC_SETOPTIONS ioctl and flag WDIOS_DISABLECARD. If the device is closed after this operation, watchdog_release() is called and status bits checked for stopping it. Besides, if the device has not been unregistered a critical message "watchdog did not stop!" is printed, although the ioctl may have successfully stopped it already.

Without the patch a user application sample code like this will successfully stop the watchdog, but the kernel will output the message "watchdog did not stop!":

Quoted 8 lines
	wd_fd = open("/dev/watchdog", O_RDWR);

	flags = WDIOS_DISABLECARD;
	ioctl(wd_fd, WDIOC_SETOPTIONS, &flags);

	close(wd_fd);

How about the following patch instead ?

drivers/watchdog/watchdog_dev.c +3 -1
index 08b48bb..9775e8d 100644
@@ -469,7 +469,9 @@ static int watchdog_release(struct inode *inode, struct file *file)
 	 * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then
 	 * watchdog_stop will fail.
 	 */
-	if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
+	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);
 

Much less invasive and the result is the same.

Thanks,
Guenter

Quoted 61 lines
---
 drivers/watchdog/watchdog_dev.c |   38 +++++++++++++++++++++-----------------
 1 file changed, 21 insertions(+), 17 deletions(-)
drivers/watchdog/watchdog_dev.c +21 -17
index ef8edec..adadbdd 100644
@@ -463,24 +463,28 @@ out:
 static int watchdog_release(struct inode *inode, struct file *file)
 {
 	struct watchdog_device *wdd = file->private_data;
-	int err = -EBUSY;
 
-	/*
-	 * We only stop the watchdog if we received the magic character
-	 * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then
-	 * watchdog_stop will fail.
-	 */
-	if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
-	    !(wdd->info->options & WDIOF_MAGICCLOSE))
-		err = watchdog_stop(wdd);
-
-	/* If the watchdog was not stopped, send a keepalive ping */
-	if (err < 0) {
-		mutex_lock(&wdd->lock);
-		if (!test_bit(WDOG_UNREGISTERED, &wdd->status))
-			dev_crit(wdd->dev, "watchdog did not stop!\n");
-		mutex_unlock(&wdd->lock);
-		watchdog_ping(wdd);
+	if (test_bit(WDOG_ACTIVE, &wdd->status)) {
+		int err = -EBUSY;
+
+		/*
+		 * We only stop the watchdog if we received the magic character
+		 * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then
+		 * watchdog_stop will fail.
+		 */
+		if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
+		    !(wdd->info->options & WDIOF_MAGICCLOSE)) {
+			err = watchdog_stop(wdd);
+		}
+
+		/* If the watchdog was not stopped, send a keepalive ping */
+		if (err < 0) {
+			mutex_lock(&wdd->lock);
+			if (!test_bit(WDOG_UNREGISTERED, &wdd->status))
+				dev_crit(wdd->dev, "watchdog did not stop!\n");
+			mutex_unlock(&wdd->lock);
+			watchdog_ping(wdd);
+		}
 	}
 
 	/* Allow the owner module to be unloaded again */
Signature · 5 lines
1.7.9.5
--
To unsubscribe from this list: send the line "unsubscribe linux-watchdog" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#2 lore raw
Quoted 38 lines · Guenter Roeck
Quoted 11 lines · Hector Palacios

A watchdog device may be stopped from userspace using WDIOC_SETOPTIONS ioctl and flag WDIOS_DISABLECARD. If the device is closed after this operation, watchdog_release() is called and status bits checked for stopping it. Besides, if the device has not been unregistered a critical message "watchdog did not stop!" is printed, although the ioctl may have successfully stopped it already.

Without the patch a user application sample code like this will successfully stop the watchdog, but the kernel will output the message "watchdog did not stop!":

Quoted 8 lines
	wd_fd = open("/dev/watchdog", O_RDWR);

	flags = WDIOS_DISABLECARD;
	ioctl(wd_fd, WDIOC_SETOPTIONS, &flags);

	close(wd_fd);

How about the following patch instead ?

drivers/watchdog/watchdog_dev.c +3 -1
index 08b48bb..9775e8d 100644
@@ -469,7 +469,9 @@ static int watchdog_release(struct inode *inode, struct file *file)
  	 * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then
  	 * watchdog_stop will fail.
  	 */
-	if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
+	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);
 
Quoted 1 line

Much less invasive and the result is the same.

I like the simplicity but it is kind of inverted logic to initially define err = -EBUSY only to turn it to zero later, so I'm rebuilding your approach like this:

drivers/watchdog/watchdog_dev.c +6 -3
index ef8edec..a4163cd 100644
@@ -463,16 +463,19 @@ out:
 static int watchdog_release(struct inode *inode, struct file *file)
 {
        struct watchdog_device *wdd = file->private_data;
-       int err = -EBUSY;
+       int err = 0;
 
        /*
         * We only stop the watchdog if we received the magic character
         * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then
         * watchdog_stop will fail.
         */
-       if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
-           !(wdd->info->options & WDIOF_MAGICCLOSE))
+       if (test_bit(WDOG_ACTIVE, &wdd->status))
+               err = -EBUSY;
+       else if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
+           !(wdd->info->options & WDIOF_MAGICCLOSE)) {
                err = watchdog_stop(wdd);
+       }
 
        /* If the watchdog was not stopped, send a keepalive ping */
 

which makes the code more readable.
Thanks,

Signature · 5 lines
Héctor Palacios
--
To unsubscribe from this list: send the line "unsubscribe linux-watchdog" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
maintainer
#3 lore raw
Quoted 59 lines · Hector Palacios
Quoted 38 lines · Guenter Roeck
Quoted 11 lines · Hector Palacios

A watchdog device may be stopped from userspace using WDIOC_SETOPTIONS ioctl and flag WDIOS_DISABLECARD. If the device is closed after this operation, watchdog_release() is called and status bits checked for stopping it. Besides, if the device has not been unregistered a critical message "watchdog did not stop!" is printed, although the ioctl may have successfully stopped it already.

Without the patch a user application sample code like this will successfully stop the watchdog, but the kernel will output the message "watchdog did not stop!":

Quoted 8 lines
	wd_fd = open("/dev/watchdog", O_RDWR);

	flags = WDIOS_DISABLECARD;
	ioctl(wd_fd, WDIOC_SETOPTIONS, &flags);

	close(wd_fd);

How about the following patch instead ?

drivers/watchdog/watchdog_dev.c +3 -1
index 08b48bb..9775e8d 100644
@@ -469,7 +469,9 @@ static int watchdog_release(struct inode *inode, struct file *file)
 	 * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then
 	 * watchdog_stop will fail.
 	 */
-	if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
+	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);
 
Quoted 1 line

Much less invasive and the result is the same.

I like the simplicity but it is kind of inverted logic to initially define err = -EBUSY only to turn it to zero later, so I'm rebuilding your approach like this:

drivers/watchdog/watchdog_dev.c +1 -1
index ef8edec..a4163cd 100644
@@ -463,16 +463,19 @@ out:
 static int watchdog_release(struct inode *inode, struct file *file)
 {
        struct watchdog_device *wdd = file->private_data;
-       int err = -EBUSY;
+       int err = 0;
 
        /*
         * We only stop the watchdog if we received the magic character
         * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then

drivers/watchdog/watchdog_dev.c out: · line 478

Quoted 10 lines
drivers/watchdog/watchdog_dev.c +5 -2
         * watchdog_stop will fail.
         */
-       if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
-           !(wdd->info->options & WDIOF_MAGICCLOSE))
+       if (test_bit(WDOG_ACTIVE, &wdd->status))
+               err = -EBUSY;
+       else if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
+           !(wdd->info->options & WDIOF_MAGICCLOSE)) {
                err = watchdog_stop(wdd);
+       }

Ok, but the added { } are unnecessary and violate coding style rules.

Guenter

#4 lore raw
Quoted 57 lines · Guenter Roeck
Quoted 56 lines · Hector Palacios
Quoted 38 lines · Guenter Roeck
Quoted 11 lines · Hector Palacios

A watchdog device may be stopped from userspace using WDIOC_SETOPTIONS ioctl and flag WDIOS_DISABLECARD. If the device is closed after this operation, watchdog_release() is called and status bits checked for stopping it. Besides, if the device has not been unregistered a critical message "watchdog did not stop!" is printed, although the ioctl may have successfully stopped it already.

Without the patch a user application sample code like this will successfully stop the watchdog, but the kernel will output the message "watchdog did not stop!":

Quoted 8 lines
	wd_fd = open("/dev/watchdog", O_RDWR);

	flags = WDIOS_DISABLECARD;
	ioctl(wd_fd, WDIOC_SETOPTIONS, &flags);

	close(wd_fd);

How about the following patch instead ?

drivers/watchdog/watchdog_dev.c +3 -1
index 08b48bb..9775e8d 100644
@@ -469,7 +469,9 @@ static int watchdog_release(struct inode *inode, struct file *file)
  	 * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then
  	 * watchdog_stop will fail.
  	 */
-	if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
+	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);
 
Quoted 1 line

Much less invasive and the result is the same.

I like the simplicity but it is kind of inverted logic to initially define err = -EBUSY only to turn it to zero later, so I'm rebuilding your approach like this:

drivers/watchdog/watchdog_dev.c +1 -1
index ef8edec..a4163cd 100644
@@ -463,16 +463,19 @@ out:
  static int watchdog_release(struct inode *inode, struct file *file)
  {
         struct watchdog_device *wdd = file->private_data;
-       int err = -EBUSY;
+       int err = 0;
 
Quoted 15 lines

drivers/watchdog/watchdog_dev.c out: · line 478

Quoted 13 lines
drivers/watchdog/watchdog_dev.c +5 -2
         /*
          * We only stop the watchdog if we received the magic character
          * or if WDIOF_MAGICCLOSE is not set. If nowayout was set then
          * watchdog_stop will fail.
          */
-       if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
-           !(wdd->info->options & WDIOF_MAGICCLOSE))
+       if (test_bit(WDOG_ACTIVE, &wdd->status))
+               err = -EBUSY;
+       else if (test_and_clear_bit(WDOG_ALLOW_RELEASE, &wdd->status) ||
+           !(wdd->info->options & WDIOF_MAGICCLOSE)) {
                 err = watchdog_stop(wdd);
+       }

Ok, but the added { } are unnecessary and violate coding style rules.

Oops. Remainders of a debug message. Thanks for pointing out.

Signature · 5 lines
Héctor Palacios
--
To unsubscribe from this list: send the line "unsubscribe linux-watchdog" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html