TITLE Unmasking an Exception (Exceptions.asm)
; This program shows how to mask (set) and unmask (clear) the divide by zero
; exception flag.
INCLUDE Irvine32.inc
.data
ctrlWord WORD ?
val1 DWORD 1
val2 REAL8 0.0
.code
main PROC
finit ; initialize FPU (divide by zero is masked)
; By unmasking, we enable the divide by zero exception.
fstcw ctrlWord ; get the control word
and ctrlWord,1111111111111011b ; unmask Divide by 0
fldcw ctrlWord ; load it back into FPU
fild val1
fdiv val2 ; divide by zero
fst val2
exit
main ENDP
END main
Hi everyone, i am very new to masm and am taking myself through some online projects i can find and am having trouble with this one, as you can see it unmasks the divide by 0 exception but how can i edit this to mask that same exception? If you can please explain and be as detailed as possible that will help out a ton!
fstcw ctrlWord ; get the control word and ctrlWord,1111111111111011b ; unmask Divide by 0 fldcw ctrlWord ; load it back into FPU
The mask bit (ZM) for the zero-divide exception is bit 2 of the FPU control word. It would improve readability if you didn't write a long masking value like "1111111111111011b".
Since bit 2 is represented by the value 4, a more readable way to clear this bit is:
and ctrlWord, not(4)
Some people will even prefer to write and ctrlWord, not(1<<2) since this one still retains reference to the bit number (2 in this case).
Now when adopting this improvement, masking the zero-divide exception becomes a matter of changing the and into an or and dropping the not operator.
and ctrlWord, not(4) ----> or ctrlWord, 4
As an alternative you can also clear or set a bit using btr and bts instructions:
and ctrlWord, not(4) ----> btr ctrlWord, 2 ;Clears bit 2
or ctrlWord, 4 ----> bts ctrlWord, 2 ;Sets bit 2