RE: [avr-gcc-list] possible 4.1.2 bug

[John Regehr]

Hi Eric-

I went and looked at the original code and it does use a 32-bit int, see 
below.  I screwed this up while creating a small example to send to the 
list -- the specific case I was looking at casts a 16-bit int into 32 and 
I propagated the smaller variable too far forwards.

Not clear by how much this beats divide-by-4096 using avr-gcc intrinsics, 
but I imagine by a good bit (assuming it's correct).

This is from the "Megasquirt AVR" firmware that I've been messing with.  
Open-source engine control -- fun stuff.

John Regehr


static inline uint16_t div4096(uint32_t a) {
        uint16_t result;

        asm(
                "\n"
                "swap %B1 \n\t"
                "swap %C1 \n\t"
                "swap %D1 \n\t"
                "ldi %A1, 0xf0 \n\t"

                "eor %D1, %C1 \n\t"
                "and %D1, %A1 \n\t"
                "eor %D1, %C1 \n\t"
                                
                "eor %C1, %B1 \n\t"
                "and %C1, %A1 \n\t"
                "eor %C1, %B1 \n\t"
                "movw %A0, %C1 \n\t"

                : "=&r" (result)
                : "r" (a)
        );
        return result;
}