Question

I have a real world program that is similar to this one, which I’ll call test.cpp:

#include <stdlib.h>

extern void f(size_t i);

int sample(size_t x)
{
     size_t a = x;
     size_t i;  

     for (i = a-2; i>=0; i--) {
           f(i);   
     }
}

And my problem is that i is an infinite loop.

If I run the following command:

g++ -S -o test.s test.cpp

I get the following assembly sequence:

        .file   "test.cpp"
        .text
        .globl  _Z6samplem
        .type   _Z6samplem, @function
_Z6samplem:
.LFB0:
        .cfi_startproc
        pushq   %rbp
        .cfi_def_cfa_offset 16
        .cfi_offset 6, -16
        movq    %rsp, %rbp
        .cfi_def_cfa_register 6
        subq    $32, %rsp
        movq    %rdi, -24(%rbp)
        movq    -24(%rbp), %rax
        movq    %rax, -8(%rbp)
        movq    -8(%rbp), %rax
        subq    $2, %rax
        movq    %rax, -16(%rbp)
.L2:
        movq    -16(%rbp), %rax
        movq    %rax, %rdi
        call    _Z1fm
        subq    $1, -16(%rbp)
        jmp     .L2
        .cfi_endproc
.LFE0:
        .size   _Z6samplem, .-_Z6samplem
        .ident  "GCC: (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3"
        .section        .note.GNU-stack,"",@progbits

I’m no expert in assembly language, but I would expect to see code for the comparison i >= 0 and a conditional jump out of the loop. What’s going on here??

GNU C++ 4.6.3 on Ubuntu Linux

Answer 1

size_t is unsigned, so the condition i>=0 is always true. It is impossible for i to be negative.