SystemClock_Config() generated by STM32CubeMX fails after jumping from the built-in bootloader (system memory)

[ChristophR]

Hi,

We are using a STM32G071GB uC, and have created an application using STM32CubeMX 5.4.0 (project file is attached).

When using the go command to jump from the built-in bootloader to the user application, HAL_RCC_OscConfig() called by SystemClock_Config() fails:

void SystemClock_Config(void)
{
  RCC_OscInitTypeDef RCC_OscInitStruct = {0};
  RCC_ClkInitTypeDef RCC_ClkInitStruct = {0};
  RCC_PeriphCLKInitTypeDef PeriphClkInit = {0};
 
  /** Configure the main internal regulator output voltage 
  */
  HAL_PWREx_ControlVoltageScaling(PWR_REGULATOR_VOLTAGE_SCALE1);
  /** Initializes the CPU, AHB and APB busses clocks 
  */
  RCC_OscInitStruct.OscillatorType = RCC_OSCILLATORTYPE_HSI;
  RCC_OscInitStruct.HSIState = RCC_HSI_ON;
  RCC_OscInitStruct.HSIDiv = RCC_HSI_DIV1;
  RCC_OscInitStruct.HSICalibrationValue = RCC_HSICALIBRATION_DEFAULT;
  RCC_OscInitStruct.PLL.PLLState = RCC_PLL_ON;
  RCC_OscInitStruct.PLL.PLLSource = RCC_PLLSOURCE_HSI;
  RCC_OscInitStruct.PLL.PLLM = RCC_PLLM_DIV1;
  RCC_OscInitStruct.PLL.PLLN = 8;
  RCC_OscInitStruct.PLL.PLLP = RCC_PLLP_DIV2;
  RCC_OscInitStruct.PLL.PLLQ = RCC_PLLQ_DIV2;
  RCC_OscInitStruct.PLL.PLLR = RCC_PLLR_DIV2;
  if (HAL_RCC_OscConfig(&RCC_OscInitStruct) != HAL_OK)
  {
    Error_Handler(); // fails here
  }

HAL_StatusTypeDef HAL_RCC_OscConfig(RCC_OscInitTypeDef  *RCC_OscInitStruct)
{
...
    else
    {
      /* Check if there is a request to disable the PLL used as System clock source */
      if((RCC_OscInitStruct->PLL.PLLState) == RCC_PLL_OFF)
      {
        return HAL_ERROR;
      }
      else
      {   
        /* Do not return HAL_ERROR if request repeats the current configuration */
        temp_pllckcfg = RCC->PLLCFGR;
        if((READ_BIT(temp_pllckcfg, RCC_PLLCFGR_PLLSRC) != RCC_OscInitStruct->PLL.PLLSource) ||
           (READ_BIT(temp_pllckcfg, RCC_PLLCFGR_PLLM) != RCC_OscInitStruct->PLL.PLLM) ||
           (READ_BIT(temp_pllckcfg, RCC_PLLCFGR_PLLN) != (RCC_OscInitStruct->PLL.PLLN << RCC_PLLCFGR_PLLN_Pos)) ||
           (READ_BIT(temp_pllckcfg, RCC_PLLCFGR_PLLP) != RCC_OscInitStruct->PLL.PLLP) ||
#if defined (RCC_PLLQ_SUPPORT)
           (READ_BIT(temp_pllckcfg, RCC_PLLCFGR_PLLQ) != RCC_OscInitStruct->PLL.PLLQ) ||
#endif
           (READ_BIT(temp_pllckcfg, RCC_PLLCFGR_PLLR) != RCC_OscInitStruct->PLL.PLLR))
        {
          return HAL_ERROR; // fails here
        }

This happens only when jumping from the bootloader, not when directly booting the user application.

Did I make a mistake, or should this be fixed in the HAL library?

Thanks!

BR,

Christoph

[Antoine Odonne]

Dear Christoph,

It is very possible that the System memory, prior executing the jump to your application using go Command don't reset the RCC setting. Maybe check if DeInit those configuration help.

Monitoring the memory map in the SYSCFG register and set it to appropriate value can also be a good point to check for your investigation.

Regards,

Antoine

[RaJa]

HAL_RCC_DeInit() before SystemClock_Config() helps properly config in CubeMX 6.2.1. In previous versions of CubeMX this wasn't needed.

[Tesla DeLorean]

Probably assumptions about currently running clocks, etc.

Could also a) transition to HSI source or b) make a determination about the current run-time conditions, and then not teardown HSE, PLL, etc when not necessary.