CODEWITHSUNDEEP
linuxassemblyx86nasmstrlen
Michel
Michel

Reputation: 113

strlen in assembly

I made my own implementation of strlen in assembly, but it doesn't return the correct value. It returns the string length + 4. Consequently. I don't see why.. and I hope any of you do...

Assembly source:

section .text
    [GLOBAL stringlen:] ; C function

stringlen:  
    push ebp
    mov ebp, esp        ; setup the stack frame

    mov ecx, [ebp+8]

    xor eax, eax        ; loop counter


startLoop:
    xor edx, edx
    mov edx, [ecx+eax]
    inc eax

    cmp edx, 0x0 ; null byte    
    jne startLoop
end:    
    pop ebp

    ret

And the main routine:

#include <stdio.h>

extern int stringlen(char *);

int main(void)
{
  printf("%d", stringlen("h"));

  return 0;
}

Thanks

Upvotes: 6

Views: 29759

Answers (6)

sharow
sharow

Reputation: 19

More easy way here(ASCII zero terminated string only):

REPE SCAS m8

http://pdos.csail.mit.edu/6.828/2006/readings/i386/REP.htm

Upvotes: 0

Michel
Michel

Reputation: 113

Thanks for your answers. Under here working code for anyone who has the same problem as me.

section .text
    [GLOBAL stringlen:]

stringlen:  
    push ebp
    mov ebp, esp

    mov edx, [ebp+8]    ; the string
    xor eax, eax        ; loop counter

    jmp if

then:
    inc eax

if:
    mov cl, [edx+eax]
    cmp cl, 0x0
    jne then

end:
    pop ebp
    ret

Upvotes: 5

Zimbabao
Zimbabao

Reputation: 8240

Change the line

mov edx, [ecx+eax]

to 

mov dl, byte [ecx+eax]

and 

  cmp edx, 0x0 ; null byte

to 

  cmp dl, 0x0 ; null byte

Because you have to compare only byte at a time. Following is the code. Your original code got off-by-one error. For "h" it will return two h + null character.

section .text
    [GLOBAL stringlen:] ; C function

stringlen:
    push ebp
    mov ebp, esp        ; setup the stack frame

    mov ecx, [ebp+8]

    xor eax, eax        ; loop counter


startLoop:
    xor dx, dx
    mov dl, byte [ecx+eax]
    inc eax

    cmp dl, 0x0 ; null byte
    jne startLoop
end:
    pop ebp

    ret

Upvotes: 1

Jester
Jester

Reputation: 58772

You are not accessing bytes (characters), but doublewords. So your code is not looking for a single terminating zero, it is looking for 4 consecutive zeroes. Note that won't always return correct value +4, it depends on what the memory after your string contains.

To fix, you should use byte accesses, for example by changing edx to dl.

Upvotes: 6

Satya
Satya

Reputation: 4478

I think your inc should be after the jne. I'm not familiar with this assembly, so I don't really know.

Upvotes: -2

unwind
unwind

Reputation: 399891

Not sure about the four, but it seems obvious it will always return the proper length + 1, since eax is always increased, even if the first byte read from the string is zero.

Upvotes: 1

Related Questions