51 MCU - Software Watchdog

I had nothing to do today, so I read a book on microcontrollers. Suddenly I realized that I have studied for such a long time but have never used the watchdog thing~~ I feel ashamed, (*^__^*) hehe...

Since this is my first article, let's write something simple but easy to write~~~

001. Software Watchdog

  When the program runs into a temporary infinite loop, if there is no human operation (that is, pressing the reset button ), the software system will be completely paralyzed. Then we need to let the MCU monitor the operation of the system by itself, which is the program operation monitoring system (WatchDog).

  To put it simply, it is like an owner who has a dog. When the owner is working normally, he always remembers to feed the dog every once in a while. The dog will be quiet after eating, so it will not affect the owner's life. But if the owner dozes off and forgets to feed the dog, then when the time comes, the dog will be hungry and find that the owner has not fed it yet, it will bark and wake up the owner. Haha, that's it. . . . . . In foreign countries, this program running monitoring system is called WatchDog. The characteristics of the watchdog are:

   1》It can work independently and basically does not rely on the CPU (it should be explained here that it mainly occupies the CPU for a few microseconds, which is to feed the dog)

   2. The CPU interacts with the system at a fixed time interval (that is, feeding the dog, as mentioned in 1) to indicate that the system is currently normal.

   3. When the CPU falls into an infinite loop, it can be detected immediately and the system can be reset

  The following is the design process of the program~~~

  When the system is stuck in an infinite loop, what kind of program can make it jump out? Only an interrupt program with a higher level than the dead program can take away the dead program's control over the CPU. To this end, we can use a timer as a watchdog, set its overflow interrupt as a high priority interrupt, and then set other interrupts in the system as low priority interrupts (here you must ensure that the interrupt you set for the watchdog must be higher than other interrupts in your program!!!) (There are many ways to set it here, but the principles just mentioned must be met!). For example, use T0 as a watchdog, the timing is about 16ms, if your system does not use INT0, then you do not need to set PTO (I will not explain it here~~you all know).

  When the watchdog is enabled, the system worker must "feed it" frequently, and the interval between each two times must not exceed 16ms (here is still the example above)

  When the program falls into an infinite loop, a T0 interrupt will be triggered within 16ms, generating a high priority interrupt, thus ending the infinite loop. At the same time, it is recommended to directly jump to the error handling program in the T0 interrupt, that is, to place an "LJMP ERROR" in the interrupt vector area (it is easier to explain here in assembly language~~)

  The following is a watchdog program written in assembly language, which includes the simulated main program, the "feed the dog" (DOG) program and the empty bullet return to 0000H (TOP) program.

      ORG       0000H

      AJMP      MAIN

      ORG       000BH

      LGMP      TOP

MAIN: MOV       SP,#06H             ; module hardware reset, these can be increased or decreased according to the system's use of resources

      MOV       PSW,#OOH

      MOV      SCON,#01H

      .

      .

      .

      MOV       IE,#00H

      MOV       IP,#00H

      MOV       TMOD,#01H

      LCALL     DOG                ​​; The time interval for calling DOG should be less than 16ms in the example

      .                           ; Here you don't need to add any programs. You can add them according to your needs. I won't waste my words.

      .

      .

DOG:   MOV       TH0,#0B1H          ; This is the dog feeding program, here is the timer counter 0

      MOV       TL0,#0E0H

      SETB      TR0

      RET

TOP:   POP       ACC                 ; Empty pop breakpoint address, that is, first take out the address you pushed, and then assign a new value

      POP       ACC                 ; Here the value is 0, you know, it means to return to the beginning of the program~~

      CLR       A                   ; Of course you don't have to do this, you can also use a call instruction here and do it yourself!

      PUSH      ACC

      PUSH      ACC

      RETI; (Here is RETI, not RET)

　　The watchdogs here should make it clear~~~

　　This is one of the methods of software anti-interference. Another method is software trap technology, which is also very simple, so I will not go into details~~

Tip: A software trap is a guide instruction that forcibly directs the captured program to a specified address, where there is a special program to handle program errors, used to make the program execute according to the established goal. A real software trap consists of three instructions:

　　NOP

    NOP

    LJMP   ERROR