TMS570LC4357: Example of configuration watchdog with FreeRTOS

Hi Pablo,

The HALCoGen doesn't generate RTI driver and DWD APs when freeRTOS is selected.

The RTI Watchdog KEY register should be programmed in privileged mode. The MCU can execute code in either Privileged mode or User (unprivileged) mode. The code executing in User mode cannot write protected registers and cannot access RAM outside its own stack and the MPU regions that are configured for it.

In xTaskCreate(), uxPriority is used to set the priority at which the task is initially created. In xTaskCreateRestricted(), it is also used to set the task to either Privileged mode or User mode. To create a Privileged mode task, bitwise OR the required task priority with portPRIVILEGE_BIT.

Hi QJ Wan,

    // Create FreeRTOS watchdogTask
    watchdogTaskHandle = xTaskCreateStatic(watchdogTask, "watchdogTask", EC_WATCHDOG_TASK_STACK_SIZE, NULL, (3 | portPRIVILEGE_BIT), watchdogStackBuffer, &watchdogTCB);
    appAssert(watchdogTaskHandle != NULL, "### ERROR ### when trying to create watchdog task");

int main(void)
{
    dwdInit(rtiREG1,4095);

    epic_common_Init();
    epic_protocol_Init();
    PCU_init();

    vTaskStartScheduler();

    return 0;
}

static void watchdogTask(void *arg)
{

    dwdCounterEnable(rtiREG1);

    for(;;)
    {
          vTaskDelay( (TickType_t) (10/portTICK_RATE_MS));

          if(wtd == 0)
              dwdReset(rtiREG1);

//          if( !(EC_WATCHDOG_CHECK_PATTERN ^ watchdogFlags) ) dwdReset(rtiREG1);
//            watchdogFlags = 0;
    }
}

Thanks

Pablo Llull said:

but for some reason after MCU reset, this causes the abort.

What kind of abort? There are several kind of abort: data abort, prefetch abort, and undefine abort.

Please refer to this FAQ:

https://e2e.ti.com/support/microcontrollers/arm-based-microcontrollers-group/arm-based-microcontrollers/f/arm-based-microcontrollers-forum/1029488/faq-tms570lc4357-troubleshooting-the-abort-exceptions-on-hercules-devices?tisearch=e2e-sitesearch&keymatch=FAQ%252520AND%252520abort#

Hi,

Thanks for the FAQ. The kind of abort is the Data Abort exception. On the first watchdog reset when it enters the dwdInit() call and exits, a few instructions later it catches the exception.

Hi,

Please let me know the value of the Data Fault Status Register (DFSR) and R14_abt register.

The below is the status description:

Hi!

 I'm sorry for the delay, I attach two images with the requested registers. 

 Data Fault Status Register (DFSR):

and R14_abt:

From DFSR=0x1008, we know that the data abort is caused by the ECC error in memory interface or in the cache, and it is a asynchronous abort which means that it is difficult to trace the abort location using the address in DFAR register. From the value of ADFSR (0x00500000), we can confirm that access was made via AXI Peripheral Port,

r14_abt = 0xD0CC-0x8,  r14_abt - 0x8 = 0xD0C4 which is the address near the instruction that caused the data abort. Please check if there is str instructions around 0xD0C4.

correction:

r14_abt = 0xD0CC