Julian Hall wrote:
>
> >> Other devices may need this as well, so we need the infrastructure
> >> in there.
> >
> >A timing infrastructure is a good thing, but I was
> >thinking of having that more in the terms of guest
> >timing... what point is there in having a host timing
> >infrastructure ?
>
> We probably want to be able to configure the VMs to be able to use devices
> that the emulation doesn't know about, therfore allowing at least one
> virtual machine to run devices that there aren't host native drivers for...
> but in many cases the host would need to be able to respond in real time
> to the device (eg cd writers, tape streamers, etc). Therefore there must
> be some way of configuring a 'real time response' situation (probably
> triggered by IO on ports which are directly mapped through, and lasting up
> to say a second after such IO has finished...) to allow the drivers to get
> realistic timing information.
Uh yeah, but as far as I understood it that's
not the kind of timing Kevin's referring to.
-- Ramon
