calling NVIC_SystemReset() in the os task will trigger hard fault

[xyxiao82]

When I calling NVIC_SystemReset() in the os task, the system will enter into hard fault handler error, so I have to call NVIC_SystemReset() again in the HardFault_Handler() to reset the system, so what is the proper way to call NVIC_SystemReset() in the os task?

[TDK]

NVIC_SystemReset() is one correct way to reset the chip.  Calling this from a task as opposed to the main thread should not cause issues.

If a hardfault occurs, I recommend finding the root cause and solving that. STM32CubeIDE has a hard fault analyzer that could be used.

[Andrew Neil]

As @TDK said, you should find out what's causing the Hard Fault:

Debugging Cortex-M Hard Faults

[Pavel A. · Accepted solution]

Do your RTOS tasks run in non-privileged mode? Then of course NVIC_SystemReset() will cause a fault.

>  so what is the proper way to call NVIC_SystemReset() in the os task?

Maybe, do this in SVC handler and make the task execute SVC. Or just leave it as is.

[xyxiao82]

Yes, you are right. The os task is running in the non-privileged mode. As the OS has taken over the SVC_Handler, I configured the OS task to run in Privileged Mode instead.

Thank you for the help.

[xyxiao82]

The root cause of hardfault is "Precise Data Bus Error", and the BFAR register is 0xE000ED0C which is SCB->AIRCR register (the reset register). 

The root cause of my problem is that I run NVIC_SystemReset() under non-privileged mode, so I change the task mode to privileged mode to solve this problem.

Thank you for your support.