Conflict in interrupt handling in TM4C123

Hi Hemant,

Which part are you exactly using? Which one from TM4C123.

Without providing the code we can only guess the problem, could you provide it?

Hi Luis,

Thanks for the response.
I am using TM4C123GH6PZ. It may be difficult to upload the code fully. If you can give pointers I can supply the snippets as appropriate.

Regards,

Hemant

Hello Hemant,

When a system hangs, the approach should be to be connected via the debugger to see where it hangs before the reset is done. Did you check where the code hangs?

Regards
Amit

Hi Luis,

The keypad is connected as:
Port K (0 to 4): Output rows
Port F (0 to 6): Input columns
Port A (0): Input column

function KeyboardInit:
{
// PortK Keyboard ROW
SysCtlPeripheralEnable(SYSCTL_PERIPH_GPIOK);
GPIOPinTypeGPIOOutput(GPIO_PORTK_BASE, GPIO_PIN_0|GPIO_PIN_1|GPIO_PIN_2|GPIO_PIN_3);

// Output default Keyboard Row mask (0x00) to PortK
GPIOPinWrite(GPIO_PORTK_BASE, GPIO_PIN_0|GPIO_PIN_1|GPIO_PIN_2|GPIO_PIN_3, 0);

// PortF.D6 to .D0 and PortA.D7 Keyboard COL
SysCtlPeripheralEnable(SYSCTL_PERIPH_GPIOF);

// Unlock PF0
HWREG(GPIO_PORTF_BASE + GPIO_O_LOCK) = GPIO_LOCK_KEY;
HWREG(GPIO_PORTF_BASE + GPIO_O_CR) |= 0x01;
HWREG(GPIO_PORTF_BASE + GPIO_O_LOCK) = 0;

GPIOPinTypeGPIOInput(GPIO_PORTF_BASE, GPIO_PIN_6|GPIO_PIN_5|GPIO_PIN_4|GPIO_PIN_3|GPIO_PIN_2|GPIO_PIN_1|GPIO_PIN_0);
GPIOPadConfigSet(GPIO_PORTF_BASE, 0x0000007F,GPIO_STRENGTH_2MA,GPIO_PIN_TYPE_STD_WPU);
// Configure PortF and PortA for falling edge interrupt
GPIOIntTypeSet(GPIO_PORTF_BASE, GPIO_PIN_6|GPIO_PIN_5|GPIO_PIN_4|GPIO_PIN_3
|GPIO_PIN_2|GPIO_PIN_1|GPIO_PIN_0,GPIO_BOTH_EDGES);
// GPIOIntTypeSet(GPIO_PORTF_BASE, GPIO_PIN_6,GPIO_BOTH_EDGES);
GPIOIntEnable(GPIO_PORTF_BASE,GPIO_INT_PIN_6|GPIO_INT_PIN_5|GPIO_INT_PIN_4
|GPIO_INT_PIN_3|GPIO_INT_PIN_2|GPIO_INT_PIN_1|GPIO_INT_PIN_0);

SysCtlPeripheralEnable(SYSCTL_PERIPH_GPIOA);
GPIOPinTypeGPIOInput(GPIO_PORTA_BASE, GPIO_PIN_7);
GPIOPadConfigSet(GPIO_PORTA_BASE, GPIO_PIN_7, GPIO_STRENGTH_2MA,GPIO_PIN_TYPE_STD_WPU);
GPIOIntTypeSet(GPIO_PORTA_BASE, GPIO_PIN_7,GPIO_BOTH_EDGES);
GPIOIntEnable(GPIO_PORTA_BASE, GPIO_INT_PIN_7);
}
function CommunicationInit:
{
SysCtlPeripheralEnable(SYSCTL_PERIPH_UART0);
// SysCtlPeripheralEnable(SYSCTL_PERIPH_GPIOA);

// Set GPIO A0 and A1 as UART pins.
GPIOPinConfigure(GPIO_PA0_U0RX);
GPIOPinConfigure(GPIO_PA1_U0TX);
GPIOPinTypeUART(GPIO_PORTA_BASE, GPIO_PIN_0 | GPIO_PIN_1);

// Configure the UART for 115,200, 8-N-1 operation.
UARTConfigSetExpClk(UART0_BASE, SysCtlClockGet(), 9600,(UART_CONFIG_WLEN_8 | UART_CONFIG_STOP_ONE |UART_CONFIG_PAR_NONE));
UARTEnable(UART0_BASE);
UARTFIFODisable(UART0_BASE);

UARTIntEnable(UART0_BASE, UART_INT_RX);
}

in main()
{
// initialisation
// Initialise keyboard interface
KeyboardInit();
ui32GPIO_PORTAStatus = GPIOIntStatus(GPIO_PORTA_BASE, 1);
ui32GPIO_PORTFStatus = GPIOIntStatus(GPIO_PORTF_BASE, 1);
GPIOIntClear(GPIO_PORTF_BASE, ui32GPIO_PORTFStatus);
GPIOIntClear(GPIO_PORTA_BASE, ui32GPIO_PORTAStatus);
IntEnable(INT_GPIOF);
IntEnable(INT_GPIOA);

// Initialise UART0 for communication with EDM Controller
CommunicationInit();
IntEnable(INT_UART0);
IntMasterEnable();
// Rest of the code in main()

Sorry Port K 0 to 3 pins are used

Hi Amit,

Yes we had the debugger on and it went to FaultISR. We then checked the NVIC_FAULT_STAT and NVIC_FAULT_ADDR but we have no clue how to proceed beyond this.

Regards,

Hemant

Hi Hemant,

After you enable the clocks for the peripherals you need at least 3 clock cycles before you access their registers.
Try adding SysCtlDelay(3) after all clocks enabled.

Hi Luis,

Sorry for the lapse.

My colleague in the team has done that in the latest code. What I sent you is a wee bit old.

Regards,

Hemant

Hello Hemant,

What is the value of the two registers? Did you look it up the Data sheet?

Regards
Amit

Hi Amar,

The registers under various conditions show different faults. The register we checked is NVIC_DBG_DATA.

Latest exercise in debugging indicates that a Time used in the application to refresh the HMI periodicallt may be causing the hanging.

We have used Time1 Sub timer A as a periodic Timer with a period of 50msec (tried different values too). The code is shown below.
********************************************************************************************************/
void HMIRefreshInit(void)
{

SysCtlPeripheralEnable(SYSCTL_PERIPH_TIMER1);
SysCtlDelay(3);

TimerConfigure(TIMER1_BASE, TIMER_CFG_A_PERIODIC);
TimerLoadSet(TIMER1_BASE, TIMER_A, SysCtlClockGet()/20);

TimerIntEnable(TIMER1_BASE, TIMER_TIMA_TIMEOUT);
IntEnable(INT_TIMER1A);
}

In the interrupt handler for this timer, we have a switch case block that sends query to the other system over UART. This is shown below.

********************************************************************************************************/
void HMIRefreshIntHandler(void) // Uses Timer1A
{
static uint8_t ui8ParameterRequest=0;
switch (ui8ParameterRequest)
{
case 0:
//Query GapV
UARTSend("g");
ui8ParameterRequest++;
break;
case 1:
//Query DROZ
UARTSend("z");
ui8ParameterRequest++;
break;
case 2:
//Query DROX
UARTSend("x");
ui8ParameterRequest++;
break;
case 3:
//Query DROY
UARTSend("y");
ui8ParameterRequest++;
break;
case 4:
//Query Alarms
UARTSend("f");
ui8ParameterRequest=0;
break;
}
TimerIntClear(TIMER1_BASE, TIMER_TIMA_TIMEOUT);
}
/********************************************************************************************************

We start and stop the timer as required in varous places in the code. We use the HMIRefreshStart() and HMIRefreshPause() functions to enable and disable the timer. This is shown below.

********************************************************************************************************/
void HMIRefreshStart(void)
{
// Enable HourCountTimerInt
TimerEnable(TIMER1_BASE, TIMER_A);
// SysTickEnable();
}
********************************************************************************************************/
void HMIRefreshPause(void)
{
// Disable HourCountTimerInt
TimerDisable(TIMER1_BASE, TIMER_A);
// SysTickDisable();
SysCtlDelay(80000);
}
/**************************************************************************************************

The HMIRefreshStart() is called for the first time after HMIRefreshInit in the main(). later we toggle the timer as required. For debugging, we have kept HMIRefreshpause() and HMIRefreshStart() in only one place for simplicty.

I noticed that if I disable HMIRefreshStart() and HMIRefreshPause() in ALL places, the application does not hang up. (Though HMIRefreshInit() still is executed in main().

Amar, if you are in India time zone, can we talk over telephone/skype as the resolution of this issue is SUPER URGENT :(

Thanks for your time.

Regards.

Hemant

Hi,

The app note AN01286–May 2012 Section 3.3 Bus Fault Scenario 2 suggests debugging price and imprecise bus faults with an example on TimeLoadSet() function. However, no solution is suggested in the Summary of the section. Could this be an issue?

Regards,

Hemant

Hello Hemant,

Precise faults are easy. Imprecise faults are the tricky one and hence the App Note mentions the procedure to do so with an example. I think the problem is arising from the manner in which the Timer is being enabled and being disabled.. If for example you do not disable the timer ever (comment out HMIRefreshPause) then does the application still hang.

Also you are using it in Periodic Mode, while the software operation is being done as a One Shot. Is this an optimal solution?

Regards
Amit

Hi Amit,

Yes, the applications hangs if the HMIRefreshPause is commented out.

On debugging further per the document mentioned above we reached a point where we believe the application enters FaultISR during execution of usprintf(). This function is called inside the UART0 Interrupt handler. the TI wiki for C I/O says C I/O functions should not be called inside Interrupt handlers as they are not re-entrant. Does this apply to usprintf() also. Will taking that part of the code out of the interrupt handler help?

Regards,

Hemant

Hello Hemant

That may be the issue. From the same wiki

Here are some examples of functions which access the C I/O buffer, or access other global state in the C I/O subsystem:

the entire printf family except sprintf

Regards
Amit

I am curious,

The functions being non re-entrant does it really make them not possible to call inside the interrupt?
Are you using in your main code any of those printf functions? Only then that would be a problem. I know it's non-optimal at all to use them inside the interrupt even if not used anywhere else but still, it's relevant to the problem.