Ulrich Weigand wrote:
>
> Hello,
>
> I've added a Intel instruction decoder (which I'd written for another
> project) to the 'user' app; this could be useful for debugging
> purposes at least (it's probably too general, i.e. bulky and slow,
> to be used in the monitor).
Sounds good, having a separate instruction decode should be handy in
the user program. I agree the monitor decode has to
be custom built. Will have to code this as part of the
scan-before-execute technique.
First, I'm looking into changing the guest<->host transition
code and other components so they are completely self contained
and relocatable. The transition code, interrupt handler,
a kernel stack page, and the real versions of the IDT, the GDT, and TSS
all need to be mapped into the linear address space of any given guest
(because when an interrupt hits, it needs all of these things), so we need to
make them be able to "float" so they can be placed in unused/unaccessible
areas of the guest's address space.
When the unused/unaccessible areas of linear space change to
in-use, we can easily then migrate these structures to other
unused/unaccessible areas. Just need to update any segment
base/offsets where pertinent.
-Kevin
