On Mon, 30 Jan 2012 16:26:06 +0100
Kevin Wolf <[email protected]> wrote:
> > The other case is what we're looking at, task switch. To actually
> > update cpl, set_segment() needs to look at cr0.pe and eflags, but these
> > might not have been committed yet. It's all solvable but the solution
> > involves knowing exactly what kvm and the emulator depend on, which
> > makes it very fragile. I think giving the emulator less complicated
> > primitives will make it more readable.
>
> I think the main problem here is that you have two sets of registers,
> one in the vcpu struct and one in the emulator context.
>
I think we can, partly?, eliminate the usage of the latter by moving the
register initialization to the decode/emulation stage, as once talked on
kvm ml, and changing the writeback code to use callbacks.
But still some refactoring is needed before that.
Takuya
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
