Drew Northup wrote:
> nature (Kev, how is that virtualizer coming, anyway...... and nice to hear
> you back from a safe vacation).
Things are coming along well. I had to put some time into
figuring out the best way to emulate instructions, so we could
model some subtleties in the way a Pentium behaves. Also
planned out some architectural details for allowing code
to run without SBE on (for running ring3 code as-is with
no virtualization intervention). Have to be careful coding
here since we don't maintain the segment register caches
(need to fetch them before using etc). And needed to
figure out cases where SBE had to be on, for instance
when there was descriptor cache vs descriptor table
incoherencies (stuff we talked about on this list).
And planned for allowing guest physical pages to be
dynamically unlocked and added to the host swap eligible
list. Also recoded the host<-->monitor/guest interaction
for much more coding flexibility. Other stuff, like
coding as though we're supporting paging etc.
A lot of this stuff is codependent, so it was worth the
time to pause and figure it out. I'm happy with things,
so now I'm back rehashing the emulation to get back
to the functionality we had before I hacked it up. :^)
Though the SBE code is more integrated, and we'll be
able to do a lot more without hitting hurdles.
Will release a tarball as soon as I get stuff back
and running using new code.
-Kevin
