STM8 cannot jump out of TIM2 update interrupt

I just learned STM8 and recorded how to solve the problems that occurred during debugging.

The chip used is STM8S103F3P6. I am going to make a card swiping module, in which the delay function is implemented with the tim2 timer. In the routines downloaded from the Internet, the delay function unit is 1ms. I think the interval time is a bit long, so I am going to change it to 10us, use the HSE external crystal oscillator 8MHz, 8 division, count 10 numbers to generate update interrupts. Then make a ms delay function, and implement 100 10us delay functions.

Problem phenomenon:

During the running process, it was found that after timer2 was initialized and the interrupt function enableinterrupts was turned on, the program could not jump out.

Problem analysis steps:

1. I debugged it first and found that the program could not jump out of the interrupt service function of tim2. At first, I thought that the code for clearing the interrupt flag was not executed, or that the code for clearing the interrupt flag should be executed immediately after entering the interrupt service function. I moved the code to the first sentence in the interrupt service function and ran it again, but it still did not work.

2. Open the register display window, refer to the Chinese manual of stm8s, find the register related content of tim2, debug in single step, and find that the clear interrupt bit of TIM2->SR1 is set to 0, but it will be set to 1 again after a few steps. I first analyzed that it might be a problem in a certain code, but after restarting four or five times, I found that the timing of setting to 1 is not fixed. Only then did I notice the CNTRL register. It seems that SR1 is set to 1 every time the value of this register becomes 0x09. Only then did I react that it might be that during the processing of the interrupt service function of tim2, another update interrupt came in and reset the update interrupt bit of SR1 to 1, which means that the timing of my tim2 is too short.

problem solved:

After extending the update time of the timer (100us), the problem was solved.

Summarize:

During the debugging process, I discovered the power of checking registers, which is a way to solve problems faster. When I used stm32 before, due to the rich online resources, basically any kind of problem could be searched. At the beginning, I also learned the library function version, so I didn’t use register debugging much, and mostly used the watch window to check the variable value. However, there are fewer online resources for stm8, and more of it needs to be done by yourself, which reflects the advantage of checking register values. It took about three hours from discovering the problem to solving it.