Ulrich Weigand wrote:

> Now, how to virtualize real/v86 mode is a whole other question ;-)
> I really think we should get protected mode working first ...
> B.t.w., in v86 mode, every software interrupt drops back into
> protected mode as well (unless you use this semi-documented
> 'extended v86 mode' ...)


Speaking of virtual-8086 mode extensions, there is an
even less documented protected mode virtual interrupts
feature as well (PVI).  I only found one article on PVI
on the 'net.  The x86.org site features an article

  http://www.rcollins.org/articles/pvi1/pvi1.html


If anyone has other info on PVI, please let me know.
The article above didn't cover everything but was
very useful.

Granted PVI was implemented properly, there may
be some use of it for virtualization.  From the
article, it sounds like Intel implemented it in
a broken way.

When the time comes, I may revisit the PVI article,
and do some real tests on the hardware to fill in
the gaps left by it.


-Kevin

Reply via email to