Search code examples
assemblydivide

"Division by zero" error when dividing by a non-zero


Previously when I had "division by zero" error in Turbo Debugger, I solved it by putting a zero into the DX register before DIV-iding by a two-byte value.

The problem returned when I added the di-register-containing lines (marked in the code with the "recently added" comments).

Looking at turbo debugger, the Division By Zero error occurs after the first return to the "Lab:" label.

Besides in this case, what other things cause the Division By Zero error?

.MODEL SMALL

Print   EQU 2
Exit    EQU 4Ch

.DATA ;------------------------------------------------------

a   DW  59
ten DW  10

.CODE ;------------------------------------------------------

Start   PROC

    mov ax, SEG DGROUP
    mov ds, ax

    mov     ax, a
    mov di, 1       ; recently added
Lab:
    mov dx, 0
    div     ten     ; the "ten" variable is always non-zero (...right?)

    mov [di], dx    ; recently added
    inc di          ; recently added

    mov bx, ax

    add dx, '0' 
    mov ah, Print      
    int 21h        

    mov ax, bx

    cmp ax, 0
    jne Lab

    mov ah, Exit
    int 21h

; -----------------------------------------------------------

Start   ENDP
    .STACK 512

    END Start

Solution

  • Here is the explanation of your problem :

    .MODEL SMALL
    
    Print   EQU 2
    Exit    EQU 4Ch
    
    .DATA ;------------------------------------------------------
    
    a   DW  59
    ten DW  10
    
    .CODE ;------------------------------------------------------
    
    Start   PROC
    
        mov ax, SEG DGROUP
        mov ds, ax
    
        mov     ax, a
        mov di, 1       ;DI POINTS TO SECOND BYTE (1) IN DATA SEGMENT.
                        ;THIS SECOND BYTE BELONGS TO "A". SO DI IS
                        ;POINTING TO THE SECOND BYTE OF "A".
    Lab:
        mov dx, 0
        div     ten     
    
        mov [di], dx    ;DI IS NOT THE SAME THAN [DI]. MOVING DX (REMAINDER 
                        ;OF DIVISION) TO [DI], AND DI POINTS TO SECOND BYTE 
                        ;OF "A", SO THE VALUE OF DX IS OVERWRITING THE SECOND 
                        ;BYTE OF "A" AND THE FIRST BYTE OF "TEN", BECAUSE
                        ;DX SIZE IS 2 BYTES. SO "TEN" IS NO LONGER 10, IT'S 0.
        inc di
    
        mov bx, ax
    
        add dx, '0' 
        mov ah, Print      
        int 21h        
    
        mov ax, bx
    
        cmp ax, 0
        jne Lab
    
        mov ah, Exit
        int 21h
    
    ; -----------------------------------------------------------
    
    Start   ENDP
        .STACK 512
    
        END Start
    

    The problem is that [di] is affecting two variables, like this :

    enter image description here