Question

I’m working on a project for class, and the professor wants us to use MASM instead of any other assembler. (I know thats is stupid, talk to him not me.) When I run ML.exe to compile my code, it crashes. Not the program I’m writing, but the assembler itself. What could possibly cause that? It’s running for a really long time before it goes, so maybe it’s an infinite loop of some sort.

Here’s my code:

.386
.MODEL FLAT

PUBLIC  compute_b_proc

points          EQU     [ebp + 16]
bs              EQU     [ebp + 12]
n               EQU     [ebp + 10]
m               EQU     [ebp + 8]

.CODE
compute_b_proc  PROC NEAR32

    push    ebpi
    mov     ebp, esp

    mov     ax, n
    cmp     ax, m
    je      base_case

    mov     ax, m
    inc     ax
    pushd   points
    push    n
    push    m
    call    compute_b_proc
    push    eax
    fld     DWORD PTR [esp + 8]
    pop     eax

    mov     ax, n
    dec     ax
    pushd   points
    push    n
    push    m
    call    compute_b_proc
    push    eax
    fld     DWORD PTR [esp + 4]
    pop     eax

    fsubp

    lea     eax, points
    mov     ebx, n
    shr     ebx, 3
    add     eax, ebx
    fild    REAL4 PTR [eax + 4]

    lea     eax, points
    mov     ebx, m
    shr     ebx, 3
    add     ebx, 4
    add     eax, ebx
    fild    REAL4 PTR [eax]

    fsubp
    fdivp
    pushd   0
    fstp    DWORD PTR [eax + 4]
    pop     eax

    mov     esp, ebp
    pop     ebp
    ret     12

    lea     eax, points
    mov     ebx, n
    shl     ebx, 3
    add     eax, ebx
    mov     eax, DWORD PTR[eax]

    mov     esp, ebp
    pop     ebp
    ret     12

compute_b_proc  ENDP
END

MASM compiles other files fine. The error message is just “ML.exe has stopped working,” nothing else. I should note that I’m using MASM version 6.11. It’s what the professor provided.

It seems to be the floating point instructions that make it crash. I reinstalled MASM, but it’s still crashing on everything floating point.

Answer 1

I figured it out… sort of. My floating point instructions were incorrect. fsubp should have been fsub, fmulp should be fmul, etc. Why that crashed the compiler, I have no idea, but changing fixed the problem.