How to use WatchDog in RTOS?

wangerxian

How to use WatchDog in RTOS? [Copy link]

I believe many people have used watchdogs. The main function of the watchdog is that the system cannot be fed when it crashes. When the watchdog timer reaches the specified time, the system will be forced to reset, which can prevent the system from becoming a brick after crashing. How to use the watchdog in RTOS? Generally, a task will be opened to feed the dog. When the system is stuck, the feeding task will not run, and the system will be reset when the watchdog timer reaches the time. It seems that there is no problem. But there is such a problem. I want to monitor the running of other tasks. If the situation is not good, reset the system. If the above method is used, when other tasks are stuck (there are problems with the operation), but the feeding task is still running, the system will not restart. So what method should be used to use the watchdog if you want to monitor whether multiple tasks are stuck ?

freebsder

The watchdog is used to monitor system-level crashes, not task crashes.

Because if the system crashes, for example, it jumps into an abnormal infinite loop, there is no other way to detect and recover/jump out of this state except for the watchdog hardware reset.

If the system is not dead, but a task is dead, you can use a dedicated monitoring task to monitor all other business tasks.

wangerxian

freebsder posted on 2021-12-9 17:53 The watchdog is used to monitor system-level crashes, not task crashes. Because if the system crashes, for example, it jumps to an abnormal crash...

How to achieve this special monitoring task? It seems that it is more complicated to perform the monitoring task.

freebsder

wangerxian posted on 2021-12-10 09:12 How to implement this special monitoring task? It seems that it will be more cumbersome to monitor the task.

If you want stability, don't be afraid of complexity, or improve your testing and programming design capabilities to reduce instability as much as possible.