Search code examples
assemblyx86-16dosmasm

x86 MS Macro Assembler Hello World program crashes

I have, just for fun, started learning assembler for the x86 architecture, its something i always wanted to know more about. I am using Microsoft Macro Assembler v6.11 running under IBM PC DOS 2000 (VM) everything runs great, the OS, assembler, linker, etc. But when i try to run this program:

.model small
.stack 100h
.data
    hello db 10,13,"Hello World$"
.code
    main proc
    lea dx, hello
    mov ah, 9h
    int 21h
    main endp
    end main

I get a bunch of weird characters and Hello, world at the end, but the process crashes, since its running under DOS, i have no choice but to ctrl-atl-delete and restart. I Would like to know what i am doing wrong? My guess would be that i am out of bounds somehow, but i clearly don't know enough yet, about this.

I do the following to assemble and link the program:

ml hello.asm

This produces no errors or warnings, I have also tried to assemble and link the program manually:

masm hello.asm
link hello.obj

This also gives no errors or warnings, but the program still doesn't work

The program have been tested under IBM PC DOS 2000, and Microsoft Windows 98 SE, the results are the same, except that its only the process running the program that crashes under Windows.

Here is the listing produced, if its of any use:

Microsoft (R) Macro Assembler Version 6.11          12/20/23 15:42:42
hello.asm                            Page 1 - 1


                .model small
                .stack 100h
 0000               .data
 0000 0A 0D 48 65 6C 6C         hello db 10,13,"Hello World$"
       6F 20 57 6F 72 6C
       64 24
 0000               .code
 0000                   main proc
 0000  8D 16 0000 R         lea dx, hello
 0004  B4 09                mov ah, 9h
 0006  CD 21                int 21h
 0008                   main endp
                    end main
Microsoft (R) Macro Assembler Version 6.11          12/20/23 15:42:42
hello.asm                            Symbols 2 - 1




Segments and Groups:

                N a m e                 Size     Length   Align   Combine Class

DGROUP . . . . . . . . . . . . .    GROUP
_DATA  . . . . . . . . . . . . .    16 Bit   000E     Word    Public  'DATA'    
STACK  . . . . . . . . . . . . .    16 Bit   0100     Para    Stack   'STACK'    
_TEXT  . . . . . . . . . . . . .    16 Bit   0008     Word    Public  'CODE'    


Procedures,  parameters and locals:

                N a m e                 Type     Value    Attr

main . . . . . . . . . . . . . .    P Near   0000     _TEXT Length= 0008 Public


Symbols:

                N a m e                 Type     Value    Attr

@CodeSize  . . . . . . . . . . .    Number   0000h   
@DataSize  . . . . . . . . . . .    Number   0000h   
@Interface . . . . . . . . . . .    Number   0000h   
@Model . . . . . . . . . . . . .    Number   0002h   
@code  . . . . . . . . . . . . .    Text     _TEXT
@data  . . . . . . . . . . . . .    Text     DGROUP
@fardata?  . . . . . . . . . . .    Text     FAR_BSS
@fardata . . . . . . . . . . . .    Text     FAR_DATA
@stack . . . . . . . . . . . . .    Text     DGROUP
hello  . . . . . . . . . . . . .    Byte     0000     _DATA 

       0 Warnings
       0 Errors

Update

I found this:

TITLE Hello World

.model small
.stack 100h

.data
message BYTE "Hello World",0dh,0ah,0

.code
main PROC
    mov ax,@data
    mov ds,ax

    mov ah,40h
    mov bx,1
    mov cx,SIZEOF message
    mov dx,OFFSET message
    int 21h

    .exit
main ENDP
END main

This program, works fine, no garbled output, as per the comments, i was under the impression that the "$" was for termination of the string, but the program above seems to append 0dh,0ah,0 to the bytes defined for "message" i have tried to adapt this in my original hello world program, but the output is the same, maybe with a little more garbled output. I will try to compare the two programs.

Solution

  • I get a bunch of weird characters and Hello, world at the end

    The DOS.PrintString function 09h expects a far pointer in the DS:DX register pair. When your .EXE executable starts, the DS segment register points at the PSP (Program Segment Prefix), but in this case you need to make it point at the .data section where your hello message resides:

    mov  ax, @data
mov  ds, ax

    The 'bunch of weird characters' are in fact the textual representation of the PSP and the .code section, finally followed by the (legible) text from the .data section.

    but the process crashes

    Every program needs an exit to its caller (the parent process is most often the OS) and you did not provide one! In the second program (that you found), this is handled by the mention .exit.
    The preferred way to terminate a DOS program is via function 4Ch where you can supply an exitcode in the AL register. Use 0 for a normal termination.

    mov  ax, 4C00h  ; DOS.TerminateWithExitcode
int  21h

    Afterwards, the parent process can inspect this exitcode using function 4Dh.

    mov  ah, 4Dh    ; DOS.GetExitcode
int  21h        ; -> AH exitcode system, AL exitcode child

    i was under the impression that the "$" was for termination of the string, but the program above seems to append 0dh,0ah,0 to the bytes defined for "message" i have tried to adapt this in my original hello world program, but the output is the same

    The $-termination is exclusively used with the DOS.PrintString function 09h. That other program was not using function 09h, but instead uses the DOS.WriteFileOrDevice function 40h with the predefined handle 1 for STDOUT. Because its operation is based on the count of bytes, having message zero-terminated is of no importance. It will display a space character however and that could sometimes mess things up a little.

    To 13,10 or not to 13,10

    Newlining in DOS needs both the carriage return (13) and the linefeed (10). In the real DOS environment the order doesn't matter a bit, and nearly everybody uses (13, 10), but some emulator might not like one or the other. I believe it was emu8086 that isn't particularly fond of (13, 10).