BM> The ModRegR/m byte does not specify whether the segment is Use16 or 
Use32..
BM> The segment declarations determine this, and the assembler would probably 
m
BM> use of the address-size and/or operand-size overrides when calling one 
type
BM> segment from the other type.  I must admit I haven't done much with this, 
b
BM> the microprocessor must have some way to tell whether it's running a 
Use16 
BM> a Use32 segment.
    The D bit tells us if we assume 32 bit (=1) or 16 bit (=0). For
    real mode and virtual 8086, d is 0 and not changeable (as far as I know).
    In win32 programs, you can have both 32 and 16 bit segments requesting
    a D bit change (dunno how to do it).
    Since my target is DOS/DOS window, I can assume 16bit coding and use
    op size overrides for 32 bit encoding.
--- GEcho 1.00
---------------