Re: [Xen-devel] [PATCH 15/16] vmx: nest: capability reporting MSRs
On Wednesday 15 September 2010 16:28:29 Dong, Eddie wrote:
> Christoph Egger wrote:
> > On Wednesday 08 September 2010 17:22:23 Qing He wrote:
> >> handles VMX capability reporting MSRs.
> >> Some features are masked so L1 would see a rather
> >> simple configuration
> >> Signed-off-by: Qing He <qing.he@xxxxxxxxx>
> >> Signed-off-by: Eddie Dong <eddie.dong@xxxxxxxxx>
> > Are there any vmx capability features that are read out via cpuid ?
> > If yes, then that code belongs into the tools patch.
> That is in 16.txt, you can include in your code. Once you fixed the MAX
> LEAF issue, I can ack that patch.
Andre tried you to explain why the MAXLEAF change is not a problem.
Is there an other problem ?
> > In SVM the nestedhvm_vcpu_features hook is empty and for MSRs
> > there are already two msr hooks namely msr_read_intercept and
> > msr_write_intercept. I assume the functions below all called from
> > there directly or indirectly.
> No. this interception is for L1 guest.
Yes, that is how I understood.
What I mean is that you can call the functions from
> > That renders the nestedhvm_vcpu_features hook useless and will remove
> > it.
> A step toward my wish of light weight wrapper, glad to see!
I would have removed it earlier, if you were telling me what I said above
which is what I want to know.
> The more you removed, the more I can ack, untill the point only those
> necessary APIs were left, such as nested EPT/NPT, heavily revisited
> interrupt injection API after removing the new namespace.
When you tell me which adjustments you exactly need I will do the changes
w/o breaking SVM, of course.
---to satisfy European Law for business letters:
Advanced Micro Devices GmbH
Einsteinring 24, 85609 Dornach b. Muenchen
Geschaeftsfuehrer: Alberto Bozzo, Andrew Bowd
Sitz: Dornach, Gemeinde Aschheim, Landkreis Muenchen
Registergericht Muenchen, HRB Nr. 43632
Xen-devel mailing list