MSP430F5335: Problem with nested interrupts and LPM

John Bolton74

Part Number: MSP430F5335

I am observing an issue with nested interrupts. I know I should avoid using nested interrupts but believe me, in this case there was no other choice.

This is the scenario:

- MSP is in Low Power Mode 3 (LPM3).

- RTC interrupt occurs.

- Inside the RTC's ISR, I set GIE to enable nesting.

- TA0 (or others) interrupts before returning from RTC's ISR (nesting).

- Inside the TA0 I change the Status Register (SR) bits to exit LPM (using __bic_SR_register_on_exit).

- The MSP does not respond anymore.

Note: If I don't try exiting LPM from the nested interrupt, the MSP behaves as expected.

In the following thread, Jens-Michael Gross, mentioned that "the nested interrupt cannot exit LPM".

https://e2e.ti.com/support/microcontrollers/msp430/f/166/t/349661?Nested-interrupts-and-jump-table

I understand the SR and PC are saved in the stack twice. Then, at the end of the nested interrupt ISR, the latest SR saved in the stack is modified (to exit LPM) and retrieved, then the PC will retrieved an instruction from the RTC's ISR and will continue executing the RTC's ISR. Then, at the end of the RTC's ISR, the SR will be retrieved again form the stack with the values saved before entering the RTC's ISR. For that reason, the MSP will not go out form LPM in that moment.

But, what I don't understand is why the MSP doesn't wake up at all, later when another interrupt occurs and there is no nesting enabled.

I would really appreciate any inputs on this matter.

Thanks in advance.

over 6 years ago

0 Bruce McKenney over 6 years ago

Guru 100735 points

I don't have an F5335. When I run this program on an F5529 (Launchpad), it functions as expected.

I'm not sure what you and I are doing differently.

///
//  nest-5529
//  If nesting works, LED1 blinks 2Hz, LED2 0.5Hz
//
#include <msp430.h> 
#define HZ  1000000UL           // 1MHz from reset
int
main(void)
{
	WDTCTL = WDTPW | WDTHOLD;	// stop watchdog timer
	P4OUT &= ~BIT7;             // Launchpad LED2
	P4DIR |=  BIT7;
	P1OUT &= ~BIT0;             // Launchpad LED1
	P1DIR |=  BIT0;
	//  Blink at 2Hz from 4 interrupts/sec
	TA0CCR0 = HZ/4/4;           // HZ/4 with ID=2
	TA0CCTL0 = CCIE;
	TA0CTL = TASSEL_2 | ID_2 | MC_1 | TACLR; // SMCLK/4, Up [,Clear]
	//  RTC in counter mode, interrupt via PS1 1/sec
	RTCPS0CTL = RT0PSDIV_7|(0*RT0PSHOLD);   // ACLK/256
	RTCPS1CTL = RT1SSEL_3|RT1PSDIV_6|(0*RT1PSHOLD)|RT1IP_6|RT1PSIE; // PS0/128, count+interrupt
    RTCCTL0 = 0;
    RTCCTL1 = (0*RTCHOLD); // Run
	__enable_interrupt();
	while (1)
	{
	    LPM3;
	}
	/*NOTREACHED*/
	return 0;
}
#pragma vector=TIMER0_A0_VECTOR
__interrupt void
TA0_0_ISR(void)
{
    P1OUT ^= BIT0;                  // LED1
    __bic_SR_register_on_exit(LPM3_bits);  // Wake up half the time
    return;
}
#pragma vector=RTC_VECTOR
__interrupt void
RTC_ISR(void)
{
    (void)RTCIV;                    // We know it's PS1
    P4OUT ^= BIT7;                  // LED2
    __bis_SR_register(GIE);         // Allow nesting
    __delay_cycles(3*HZ/4);         // 3/4-second visibly blocks TA0 without GIE
    return;
}

0 Dietmar Walther over 6 years ago in reply to Bruce McKenney

TI__Genius 14165 points

Hi John,

Can you please post the part of the code where you go into LPM?

To clear the SR inside the TimerA ISR has only an effect if the interrupt of your TimerA appeared while you were in LPM. If it would be executed as a nested interrupt your device will anyway return into AcitveMode because it came from RTC right?

So either you cause some kind of stack overflow or you might return into AM within your main loop and do not catch the CPU in a loop. These are my initial ideas if you say it does not respond anymore.

**Attention** This is a public forum

MSP low-power microcontrollers

MSP low-power microcontroller forum

MSP430F5335: Problem with nested interrupts and LPM