Re: [avr-gcc-list] Compiled code; output Bug or Not?

[Joerg Wunsch]

Panther Electronics <address@hidden> wrote:

> avr-gcc (GCC) 3.3 20021209 (experimental)

OK.

> avr-gcc -c -g -Os -funsigned-char -funsigned-bitfields -fpack-struct 
> -fshort-enums -Wall -Wstrict-prototypes -Wa,-ahlms=MXTPort.lst 
> -mmcu=at90s2313 -I. MXTPort.c -o MXTPort.o

Since you sent me the C file in a private mail, i could try
reproducing this.

Hmm, when i replace the -c by a -S there and omit the -o MXTPort.o,
the generated MXTPort.s contains code for check_pa() that has exactly
one RET instruction.  So sorry, i still don't see where there would be
a problem.

The generated MXTPort.lst seems to agree with this.

The disassembled ELF file (avr-objcopy -d) agrees, too.  However, the
addresses shown don't match the snippet you've been posting (from
AVRstudio?) in any way.  My compiler version is slightly newer
(20030108), but this should not cause such drastical differences
anyway.

> MXTPort.c: In function `check_pa':
> MXTPort.c:269: warning: assignment makes pointer from integer without a cast

Btw., are you absolutely sure that you want to read an SRAM pointer
directly from an IO port?  This looks a bit suspicious to me.

-- 
J"org Wunsch                                           Unix support engineer
address@hidden        http://www.interface-systems.de/~j/
avr-gcc-list at http://avr1.org