On Sun, 10 Feb 2008 13:44:05 +0200 Avi Kivity <[EMAIL PROTECTED]> wrote:
> Anthony Liguori wrote: > > So what we would like to do, is instead of setting up vm86 mode for the > > guest to execute real mode, use x86_emulate() to just emulate the code. > > This means that we wouldn't be using the "vmlaunch" instruction when in > > real mode and instead would be doing an x86_emulate() loop. > > > > As using the emulator is likely to be slower than VT, we can call the > emulator only if we are in a "VT unfriendly" state, so the code might > look like > > if (vmx->rmode.active && big_real_mode(vmx)) > ... > > In addition, there are some protected-mode states that VT can't handle > (cs.rpl != ss.rpl IIRC) so we can emulate those cases as well. > > To improve speed, we may want to emulate >1 instruction per iteration. So the plan is to keep the usage of vm86 mode until we detect a "VT unfriendly" state. I think about a VMentry failure to detect this kind of state? Then, when we are in big_real_mode state, we emulate the code as suggested by Avi with the optimisation that consists to emulate more than one instruction per iteration. Best regards, Guillaume ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ kvm-devel mailing list kvm-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/kvm-devel
