avr-gcc-list
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[avr-gcc-list] Mega32 wierdness


From: Jonny Dyer
Subject: [avr-gcc-list] Mega32 wierdness
Date: Sun, 23 Jan 2005 01:38:41 -0800

I am trying to get started working with GCC for AVR and am having a terrible time with a Mega32. Note that I am experience with AVR assembly and the aTinys, but this is my first C AVR project.

As a simple test program, I have Timer0 interrupt enabled and am running Timer0 at clk/1024. On overflow, I am toggling all the pins on Port D. If I do just this and nothing else, the program works fine. However, I wanted to add a counter to the overflow function so that I can divide the timer overflows further and only toggle the LED, say, every 5 overflows. If I define my counter variable as volatile (as I have read is necessary when working with interrupts several other places), increment it in the interrupt function and leave the PORTD toggling in that function, it still operates the same (as it should):

SIGNAL(SIG_OVERFLOW0)
{
        timer_step++;
        PORTD ^= 0xff;
}

However, if I do anything to timer_step in the rest of the code (namely in main()), it's as if the timer overflow stops being called and the pins on PORTD no longer flash. Note that I am doing nothing at all to my interrupt routine:

volatile uint8_t timer_step;

int main (void)
{
        uint8_t save;
        timer_step = 0;
        DDRD = 0xFF;
        PORTD = 0x00;
        
        TIFR  = BV(TOIE0);
        TCCR0  = BV(CS02) | BV(CS00); /* CTC, prescale = 128 */
        TCNT0  = 0;
        TIMSK = BV(TOIE0);    /* enable Timer0 overflow interrupt*/
        sei();
        
        for(;;)
        {
                if(timer_step > 10)
                {
                        timer_step = 0;
                }
        }
}

SIGNAL(SIG_OVERFLOW0)
{
        timer_step++;
        PORTD ^= 0xff;
}

I have disassembled the code and I can see nothing wrong with the code GCC is generating either. It is almost as if somehow messing with timer_step (which gcc is saving at r0) outside of the interrupt function corrupts TIMSK or TIFR or the interrupt vector. Any help is greatly appreciated.

Thanks
Jonny

reply via email to

[Prev in Thread] Current Thread [Next in Thread]