John Rogers wrote:
>
> interesting : the values of the segment registers DS,ES and the value of CX
> (low 16-bits of ECX) seem to have no resemblance to the code that should
> have loaded them with 0x7C00,0x9000,0x0100 respectively...
I had a realmode program that supposedly crashed
on an incl %ecx instruction ;)
I think I can clear up the puzzlement:
Our decoder etc. in the user program do not
work with real-mode. Addresses are probably
being decoded in a pretty wrong way when
the processor is in realmode. This needs
to be fixed; nobody expected Kevin to come
up with RM support this fast ;)
-- Ramon
