[avr-gcc-list] Re: Insight & simulavr ->korrektion :-)

[Theodore A. Roth]

Hi Klaus,

I'm not seeing anything wrong happen in simulavr. I ran the attached 
program in the sim with avr-gdb (no insight). Run avr-gdb as such:

  $ avr-gdb -nx -x gdb.in

Here's the disassembly of the resulting code:

000000ca <main>:

volatile int foo;

int main (void)
{
  ca:   cf ef           ldi r28, 0xFF   ; 255
  cc:   d0 e1           ldi r29, 0x10   ; 16
  ce:   de bf           out 0x3e, r29   ; 62
  d0:   cd bf           out 0x3d, r28   ; 61
    for (;;)
    {
        foo++;
  d2:   80 91 00 01     lds r24, 0x0100
  d6:   90 91 01 01     lds r25, 0x0101
  da:   01 96           adiw    r24, 0x01   ; 1
  dc:   90 93 01 01     sts 0x0101, r25
  e0:   80 93 00 01     sts 0x0100, r24
  e4:   f6 cf           rjmp    .-20        ; 0xd2

which looks simular to what you posted.

I don't have insight on my computer at home, but will retry this on my 
computer at work later today.

Ted Roth

On Fri, 27 Dec 2002, Klaus Rudolph wrote:

:)>Hi all, hi Tod,
:)>
:)>
:)>I saw a funny but absolutly wrong result in testing insight
:)>today.
:)>
:)>I compiled a simple program and got the following window 
:)>in insight:
:)>     5       int main() {
:)>-    0x56    <main>:         ldi     r28, 0x5F       ; 95
:)>-    0x58    <main+2>:               ldi     r29, 0x02       ; 2
:)>-    0x5a    <main+4>:               out     0x3e, r29       ; 62
:)>-    0x5c    <main+6>:               out     0x3d, r28       ; 61
:)>-    0x5e    <main+8>:               nop
:)>     6               for (;;) {
:)>-    0x72    <main+28>:              rjmp    .-20            ; 0x60
:)>     7                       a++;
:)>     0x60    <main+10>:              lds     r24, 0x0060
:)>-    0x64    <main+14>:              lds     r25, 0x0061
:)>-    0x68    <main+18>:              adiw    r24, 0x01       ; 1
:)>-    0x6a    <main+20>:              sts     0x0061, r25
:)>-    0x6e    <main+24>:              sts     0x0060, r24
:)>-    0x74    <__stop_progIi__>:              rjmp    .-2             ; 0x74
:)>
:)>In address 0x68 the r24/25 register should be added by one, but simulavr
:)>add not
:)>one, it add any possible value. Looking into memory give a total
:)>confused output. Every assembler step the complete memory content 
:)>change :-(
:)
:)Sorry, the memory content is ok, I forget to ad 0x800000 and the dump
:)was
:)relativ to $PC. Sorry, "change user error".
:)
:)But simulavr add the wrong value, this behavior did not change :-(
:)
:) 
:)>I use:
:)>simulavr-0.1.1
:)>insight-5.3
:)
:)>Maybe some compatibility problems with the downloaded versions??
:)
:)>Thanks
:)>     Klaus
:)

test.tar.gz
Description: GNU Zip compressed data