On Wed, 8 Mar 2000, Kevin P. Lawton wrote:
> We talked about this in the past. It was agreed that since
> this project is so very tied to x86, we weren't going to
> bother with abstracting things that far, and porting to
> other architectures.
>
> There's really nothing portable about virtualization, other
> than the concepts themselves.
Well, you would be the one that would know what you're talking about, and
I would be the one having no idea. I'm mostly just glad that this project
is happening.
It just seems like it should be possible to make it more portable. Well,
I'm certain it's possible, because if necessary you could write a complete
emulator in software. I have a feeling that's where we disagree. I'm
talking about doing full emulation where necessary, and I think you're
talking about sticking to virtual machine monitoring.
Of course, I'm really not sure what I'm talking about. Maybe it would be
better to discribe the layers as:
virtual machine monitor
hardware emulator
hardware interface
So if you were emulating x86 on x86, and you had a chunk of code which
could be execute directly on the hardware w/out virtualization, you could
just skip the emulation layer (it wouldn't need to exist), and go directly
to the hardware interface.
I recognize that these goals may not be reasonable, they just seem ideal
to me. I still dream about being able to run any binary on any platform
(hardware & software). And I still want to be able to upgrade my kernel
without rebooting :)
__________________________________________________________________
PGP fingerprint = 03 5B 9B A0 16 33 91 2F A5 77 BC EE 43 71 98 D4
[EMAIL PROTECTED] / http://www.op.net/~darxus
Chaos reigns.
