STM32H7 dual core debugging. M4 pc goes rogue.

3 replies
1621 views

Hi Community,

I am getting familiar with the STM32CubeIDE Dual Core Debugger capability for the H7 chip.

I am using code generated from STM32CubeMX of the IDE.

The versions I am using are:

STM32CubeIDE 1.4.1

STM32CubeMX 6.0.1-RC3

NUCLEOH745ZI-Q

This is the code I am using. I added this Variable to help me see evaluate the debugging but it shouldn't affect what I am seeing.

uint16_t thisVariable = 0;
int main(void)
{
 /* USER CODE BEGIN 1 */
 
 /* USER CODE END 1 */
 
/* USER CODE BEGIN Boot_Mode_Sequence_1 */
 /*HW semaphore Clock enable*/
 __HAL_RCC_HSEM_CLK_ENABLE();
// /* Activate HSEM notification for Cortex-M4*/
 HAL_HSEM_ActivateNotification(__HAL_HSEM_SEMID_TO_MASK(HSEM_ID_0));
// /*
// Domain D2 goes to STOP mode (Cortex-M4 in deep-sleep) waiting for Cortex-M7 to
// perform system initialization (system clock config, external memory configuration.. )
// */
 HAL_PWREx_ClearPendingEvent();
 HAL_PWREx_EnterSTOPMode(PWR_MAINREGULATOR_ON, PWR_STOPENTRY_WFE, PWR_D2_DOMAIN);
// /* Clear HSEM flag */
 __HAL_HSEM_CLEAR_FLAG(__HAL_HSEM_SEMID_TO_MASK(HSEM_ID_0));
 
/* USER CODE END Boot_Mode_Sequence_1 */
 /* MCU Configuration--------------------------------------------------------*/
 
 /* Reset of all peripherals, Initializes the Flash interface and the Systick. */
 HAL_Init();
 
 /* USER CODE BEGIN Init */
 
 /* USER CODE END Init */
 
 /* USER CODE BEGIN SysInit */
 
 /* USER CODE END SysInit */
 
 /* Initialize all configured peripherals */
 /* USER CODE BEGIN 2 */
 
 /* USER CODE END 2 */
 
 /* Infinite loop */
 /* USER CODE BEGIN WHILE */
 while (1)
 {
 /* USER CODE END WHILE */
	 thisVariable++;
	 if (thisVariable == 10000){
		 thisVariable = 0;
	 }
 /* USER CODE BEGIN 3 */
 }
 /* USER CODE END 3 */
}

These are the scenarios that I see:

Scenario 1

I start the M7
I start the M4
M4 ends up at some wild address outside program memory

Scenario 2

I start the M7
I step through the M4 code until I reach line 26 Hal_Init()
I then set M4 code to run
I can have either core stop the other as expected

Scenario 3

I start the M7
I comment out line 18 of the M4 code
I start the M4
I can have either core stop the other as expected

The code that affects this is the code generated by the IDE/MX. Has anyone seen something similar and resolved it ? I saw a post about a revision of HW that had a dodgy regulator. I am not sure if that is what I am seeing.

Any help would be appreciated.

This topic has been closed for replies.

TOlli

Senior

Hello,

I've seen this same thing happen on our end as well. I think it is due to the fact that the M4 goes to sleep/stop mode and the debugger is not able to handle that (despite the setting of enabling low power debugging). I think that's the last address you see on the stack, and after that you lose track. By disabling the sleep mode from line 18, you don't enter sleep mode and everything is fine.

Why this doesn't happen with stepping through the line is beyond me, but I've noticed the same thing happening.

Happy to hear some answers on how to overcome this and if it is safe to disable line 18 completely. I would still like to have M4 only continue after M7 has set everything up, so I think a similar functionality is needed.