Re: [avr-gcc-list] Avr-gcc Produces Incorrect Code with -Os

[Dave N6NZ]

The result of the atan2 does not impact the bit value written to PORTA, 
 so the compiler is free to re-order it.  That's what optimizers do.

-dave

Thomas D. Dean wrote:
> I have a code segment which 
>   1.  sets a bit in PORTA.
>   2.  calls atan2.
>   3.  clears the same bit in PORTA.
>
> The compiler produces code that 
>   1.  sets the bit in PORTA.
>   2.  clears the same bit in PORT.
>   3.  calls atan2.
>
> With -O0, the code is correct.
>
> # uname -a
> FreeBSD dv6000.tddhome 7.0-STABLE FreeBSD 7.0-STABLE #0: Sun May  4 07:58:25 
> PDT 2008     address@hidden:/usr/src/sys/GENERIC  i386
>
> # avr-gcc --version
> avr-gcc (GCC) 4.2.2
> Copyright (C) 2007 Free Software Foundation, Inc.
> This is free software; see the source for copying conditions.  There is NO
> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
>
> The code fragment,
>
>   ...
>   dtostrf(cos_rad,6,3,&line4[14]);
>   line3[20] = ' ';
>   line4[20] = ' ';
>   TRACE_ON(TRACE4);  // portb |= _bv(04)
>   atan_rad = atan2(cos_rad,sin_rad);  
>   TRACE_OFF(TRACE4);  // portb &= ~_bv(04)
>   dtostrf(atan_rad,6,3,&line4[26]); 
>   ...
>
> # gmake ATMEGA16=1 TRACE=1 8_BIT=1 clean all
> rm -f *.o *.elf *.s19 *.b *.a *.map *.hex
> avr-gcc  -mmcu=atmega16  -Wall -Wmissing-prototypes  -Os -DTRACE -DLCD_8_BIT   
> -I.    -c -o main.o main.c
> avr-gcc  -mmcu=atmega16  -Wall -Wmissing-prototypes  -Os -DTRACE -DLCD_8_BIT   
> -I.    -c -o timer.o ../common/timer.c
> avr-gcc  -mmcu=atmega16  -Wall -Wmissing-prototypes  -Os -DTRACE -DLCD_8_BIT   
> -I.    -c -o lcd-8-bit.o ../common/lcd-8-bit.c
> avr-gcc -o main.elf main.o timer.o lcd-8-bit.o -mmcu=atmega16  
> -L../../../asus-avr-1.0/config/stk500 -L../../../asus-avr-1.0/lib  
> -L/usr/local/avr/lib/ -lasus -lc -lm -lprintf_flt -lscanf_flt
> avr-objcopy -j .text -j .data -O ihex main.elf main.hex
>
>
> #  avr-objdump -D main.elf | grep 'call.*atan2' -B12 -A10
>      398:       0e 94 71 03     call    0x6e2   ; 0x6e2 <dtostrf>
>      39c:       80 e2           ldi     r24, 0x20       ; 32
>      39e:       f1 01           movw    r30, r2
>      3a0:       84 8b           std     Z+20, r24       ; 0x14
>      3a2:       f2 01           movw    r30, r4
>      3a4:       84 8b           std     Z+20, r24       ; 0x14
>      3a6:       dc 9a           sbi     0x1b, 4 ; 27
>      3a8:       dc 98           cbi     0x1b, 4 ; 27
>      3aa:       a5 01           movw    r20, r10
>      3ac:       94 01           movw    r18, r8
>      3ae:       c7 01           movw    r24, r14
>      3b0:       b6 01           movw    r22, r12
>      3b2:       0e 94 0f 06     call    0xc1e   ; 0xc1e <atan2>
>      3b6:       cf 54           subi    r28, 0x4F       ; 79
>      3b8:       df 4f           sbci    r29, 0xFF       ; 255
>      3ba:       08 81           ld      r16, Y
>      3bc:       19 81           ldd     r17, Y+1        ; 0x01
>      3be:       c1 5b           subi    r28, 0xB1       ; 177
>      3c0:       d0 40           sbci    r29, 0x00       ; 0
>      3c2:       23 e0           ldi     r18, 0x03       ; 3
>      3c4:       46 e0           ldi     r20, 0x06       ; 6
>      3c6:       0e 94 71 03     call    0x6e2   ; 0x6e2 <dtostrf>
>      3ca:       b1 01           movw    r22, r2
>
>
> _______________________________________________
> AVR-GCC-list mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/avr-gcc-list