问题
I am trying to write an array-reversing function that works correctly whether the static array holds BYTEs or WORDs.
This is what I have so far where I assumed the array was of size WORD
.data
myArray WORD 1231h, 2342h, 3353h, 4564h, 5675h, 7, 9
.code
main proc
mov eax, 0 ;index of the left side of the array
mov esi, SIZEOF myArray
sub esi, TYPE myArray ;index of the right side of the array
Switch:
movsx edx, [myArray + eax] ;puts the left element of the array in eax
xchg dx, [myArray + esi] ;exchange edx with the right element of the array
mov [myArray + eax], dx ;puts the right element into the left part of the array
add eax, TYPE myArray ;eax is currently pointing to leftPtr so we add the appropriate amount to
;it depending on the size of each element in myArray. This way it will point
;to the next element in the array
sub esi, TYPE myArray ;same concept as above except we are subtracting the amount from the right pointer
cmp esi, eax ;compare the right pointer to the left pointer
jnle Switch ;Jump if the right pointer is !(<=) the left pointer
main endp
end main
I can use the movzx/movsx instruction to move the smaller sized value from the array into a 32-bit register.
The problem is writing something that assembles to an 8bit store or 16bit store depending on TYPE.
回答1:
You can define a symbol conditionally on the type of myArray to be either the text dl, dx or edx and use that in place of the register. Something like:
IF TYPE myArray EQ TYPE BYTE
@DX TEXTEQU <dl>
ELSEIF TYPE myArray EQ TYPE WORD
@DX TEXTEQU <dx>
ELSEIF TYPE myArray EQ TYPE DWORD
@DX TEXTEQU <edx>
ENDIF
mov @DX, [myArray + eax] ;puts the left element of the array in DX
xchg @DX, [myArray + esi] ;exchange DX with the right element of the array
mov [myArray + eax], @DX ;puts the right element into the left part of the array
来源:https://stackoverflow.com/questions/37770559/generic-masm-code-storing-8-or-16-bits-of-a-register-depending-on-type-of-a-sym