NASM trying to multiply integer by float, printf writes garbage

Question

I tried creating a very simple example: a conversion program, where one number is an int and the other is a double.

global _main
extern _printf

section .data
    km_h    dd  70
    format  db '%fmph=%dkm/h',10,13,0
    km_mi   dq 1.609

section .code
_main:
    push ebp
    mov ebp, esp
    
    sub esp, 16 ;for the parameters
    
    fld qword [km_mi]
    fimul dword [km_h]
    
    mov dword [esp], format
    mov eax, dword [km_h]
    mov [esp+12], eax
    fstp qword [esp+4]
    push format
    call _printf
    
    mov esp, ebp
    pop ebp
    ret

I tried cc -S on a C source and went as far as writing it with mov [esp+X] instead of pushing, but whatever I try, it outputs garbage.

Why do you have both `mov dword [esp], format` and `push format` ? — Michael, May 16 '23 at 20:00
Because I'm an idiot and didn't notice that I left a push there after all... it's very likely the cause, I'll remove it... — legacywolf, May 16 '23 at 20:09
I reversed the variable names I guess when translating source to English to make it understandable. — legacywolf, May 16 '23 at 21:08

legacywolf · Answer 1 · 2023-05-17T18:47:00.123

0

Okay, I've removed the push and it works now. The conversion is still backwards, but that's easy to fix, and of course, that didn't really matter, it was just a homework to practice the opcodes.

edited May 17 '23 at 18:47

answered May 17 '23 at 18:44

legacywolf

1
3

NASM trying to multiply integer by float, printf writes garbage

1 Answers1