CC3235SF: Reset cause is not cleared

Hi Jan,
I always get the resetCause reason as:
Reset Cause Register: 0x00000007 which is PRCM_HIB_EXIT. Even the MCU has undergone a watchdog reset.

UART_PRINT("\n\n=== BOOT DEBUG INFO ===\n");
UART_PRINT("Reset Cause Register: 0x%08X\n", resetCause);
UART_PRINT("\nIndividual Flags:\n");
UART_PRINT(" PRCM_POWER_ON: %s\n", (resetCause == PRCM_POWER_ON) ? "SET" : "CLEAR");
UART_PRINT(" PRCM_LPDS_EXIT: %s\n", (resetCause == PRCM_LPDS_EXIT) ? "SET" : "CLEAR");
UART_PRINT(" PRCM_HIB_EXIT: %s\n", (resetCause == PRCM_HIB_EXIT) ? "SET" : "CLEAR");
UART_PRINT(" PRCM_WDT_RESET: %s\n", (resetCause == PRCM_WDT_RESET) ? "SET" : "CLEAR");
UART_PRINT(" PRCM_MCU_RESET: %s\n", (resetCause == PRCM_MCU_RESET) ? "SET" : "CLEAR");
UART_PRINT(" PRCM_SOC_RESET: %s\n", (resetCause == PRCM_SOC_RESET) ? "SET" : "CLEAR");

If the controller get reset by watcdog reset, I get;

=== BOOT DEBUG INFO ===

Reset Cause Register: 0x00000007

Individual Flags:

PRCM_POWER_ON: CLEAR

PRCM_LPDS_EXIT: CLEAR

PRCM_HIB_EXIT: SET

PRCM_WDT_RESET: CLEAR

PRCM_MCU_RESET: CLEAR

PRCM_SOC_RESET: CLEAR

Hi,

I haven't never seen such issue. Are you sure that device is restarted by WDT? What status do you see when you trigger reset by the reset pin? Can you test at CC3235 LaunchPad?

Jan

Hi Jan,

I'm intentionally performing the Watchdog reset by not refreshing the watchdog. At that time also I'm getting PRCM_HIB_EXIT only.

void initWatchdog(void)
{
Watchdog_Params params;

Watchdog_init();
Watchdog_Params_init(&params);
params.resetMode = Watchdog_RESET_ON;
params.debugStallMode = Watchdog_DEBUG_STALL_ON;

watchDogHandle = Watchdog_open(Board_WATCHDOG_0, &params);

if (watchDogHandle == NULL) {
while (1); // Error
}

reload = Watchdog_convertMsToTicks(watchDogHandle, 15000); // 10s timeout
Watchdog_setReload(watchDogHandle, reload);
}

void* watchDogTask(void *args)
{
for(;;)
{
vTaskDelay(3000);
if ((taskFlags & ALLTASKSETBIT) == ALLTASKSETBIT)
{
Watchdog_clear(watchDogHandle);
Watchdog_setReload(watchDogHandle, reload);
taskFlags = 0;
}
else
{
UART_PRINT("Watchdog not cleared! Flags: 0x%02X\n", taskFlags);
}

}
return NULL;
}

Hi,

What you see when you try other reset sources?

Jan

Hi,

I'm performing power shut down by calling this function:

void powershutdown (uint32_t shutdownTime)
{
(void)Power_registerNotify(&hibSignal, PowerCC32XX_ENTERING_SHUTDOWN,
                         &preHibConfig, (uintptr_t) NULL);
    (void)Power_shutdown(0, shutdownTime);
}
I'm getting the same reset cause reason:

Reset Cause Register: 0x00000007

Individual Flags:

PRCM_POWER_ON: CLEAR

PRCM_LPDS_EXIT: CLEAR

PRCM_HIB_EXIT: SET

PRCM_WDT_RESET: CLEAR

PRCM_MCU_RESET: CLEAR

PRCM_SOC_RESET: CLEAR

Hi,

I think this is expected, because for full SoC reset is used hibernation API. Try reset by pin of power cycle.

You can check implementation of Power_shutdown() inside \source\ti\drivers\power\PowerCC32XX.c.

Jan

Hi Jan,
I tried this Power_shutdown() shutdow,
void powershutdown (uint32_t shutdownTime)
{
    (void)Power_shutdown(0, shutdownTime);
}
for this also I'm getting the same reset Cause reason:

Reset Cause Register: 0x00000007

Individual Flags:

PRCM_POWER_ON: CLEAR

PRCM_LPDS_EXIT: CLEAR

PRCM_HIB_EXIT: SET

PRCM_WDT_RESET: CLEAR

PRCM_MCU_RESET: CLEAR

PRCM_SOC_RESET: CLEAR

And the another point is, I have checked the 
uint32_t wakeCause = PRCMHibernateWakeupCauseGet();
while checking the HIbernateWakeup casue reason,
I get PRCM_HIB_WAKEUP_CAUSE_GPIO (0x00000004) when the MCU has undergone a power cycle reset and PRCM_HIB_WAKEUP_CAUSE_SLOW_CLOCK (0x00000002) when the MCU Has undergone a watchdog reset.

Hi,

As I said if you are using Power_shutdown() return code 0x7 from PRCMSysResetCauseGet() is expected. This is normal. What do you see when you try other reset sources like POR or reset pin (nRESET)?

Jan

Hi,
I tried the same power_shutdown with the CC3235 LaunchPad and checked the PRCMSysResetCauseGet().
I got the resetCause reason PRCM_WDT_RESET when the MCU has undergone the Watchdog and PRCM_POWER_ON when the MCU has undergone the power shut down reset.

Hi,

That means same code at CC3235 LaunchPad works without any issue? Right? It is correct to say that we narrowed down issue to your hardware?

Jan

Hi,
I always get reset cause reason Hibernate exit (PRCM_HIB_EXIT) if I flash the code using uniflash in my hardware and CC3235sf Launchpad for watchdog reset and power cycle reset. I get the resetCause reason PRCM_WDT_RESET when the MCU has undergone the Watchdog and PRCM_POWER_ON when the MCU has undergone the power shut down reset when I perform debug using CCS. I'm using the XDS110 DEBUG PROBE.

Hi,

I am having trouble understanding what you are saying. 

You're saying that when you flash code to conduct watchdog and power cycle reset functionality, your reset reason is hibernation exit? You also get a reset reason of watchdog timer reset during debug mode?

Yes

Hi Jan,

What configuration do I need to change to get the exact reset cause? For example, if the MCU has undergone a watchdog reset, I need to identify it as a watchdog reset, and similarly for a power cycle reset.

Hi,

I never seen such issue as yours. I am not sure what is going on at your case. Unfortunately I am not able test this at my end. Please wait for answer from TI.

Jan

Hi KS,

Can you send me a dumb-down version of your project so that I can see if I can reproduce your issue on my end, or at least whatever code you are currently running to trigger non-shutdown events that supposedly result in an incorrect reset cause?

Hi,

I'm using simplelink SDK 6.10 version, and CCS version 12.0.0 in CC3235SF. I'm using FreeRTOS in my project. I have configured watchdog to check whether all the task are reporting, if any errors occurs the watchdog will reset the MCU. I keep on storing my data in network processor file system(NVS), if the watchdog reset happens I'll read the data from file system continuing the procedure from the left over state. To identify the reset reason I'm using the function PRCMSysResetCauseGet() after the main. But, the function returns the reset reason as PRCM_HIB_EXIT even the reset happens because of watchdog. I'm sure that the reset happens because of watchdog because I'm intentionally performing that.
I want to know the configuration that I need to perform to get the proper reset cause reason.

Hi,

Hold on. This is not related to your case. How often do you save values into NWP filesystem (sFlash). If you do this often, this is very bad idea because you will soon damage SPI flash (it have guaranteed 100k write cycles / sector only).

Jan

Ok, I won't do the filesystem stuff but will replicate the rest of your experiment and let you know what I see.