Re: [Xen-devel] [V0 PATCH 5/6] AMD-PVH: Support TSC_MODE_NEVER_EMULATE for PVH

[Mukesh Rathor <mukesh.rathor@xxxxxxxxxx>]

On Mon, 18 Aug 2014 21:38:35 -0400
Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx> wrote:

> On 08/15/2014 09:53 PM, Mukesh Rathor wrote:
> > On AMD, MSR_AMD64_TSC_RATIO must be set for rdtsc instruction in
> > guest to properly read the cpu tsc. To that end, set tsc_khz in
> > struct domain.
> >
> > Signed-off-by: Mukesh Rathor <mukesh.rathor@xxxxxxxxxx>
> > ---
> >   xen/arch/x86/time.c | 1 +
> >   1 file changed, 1 insertion(+)
> >
> > diff --git a/xen/arch/x86/time.c b/xen/arch/x86/time.c
> > index bd89219..7512aa4 100644
> > --- a/xen/arch/x86/time.c
> > +++ b/xen/arch/x86/time.c
> > @@ -1908,6 +1908,7 @@ void tsc_set_info(struct domain *d,
> >            * but "always_emulate" does not for some reason.  Figure
> > out
> >            * why.
> >            */
> > +        d->arch.tsc_khz = cpu_khz;
> >           switch ( tsc_mode )
> >           {
> >           case TSC_MODE_NEVER_EMULATE:
> 
> I suspect that TSC_MODE DEFAULT may actually work for PVH: since we 
> don't support migration it should be equivalent to NEVER_EMULATE.
> 
> If you replace in tsc_set_info() is_hvm_domain() with 
> is_hvm_container_domain() and allow TSC_MODE_DEFAULT along with 
> TSC_MODE_NEVER_EMULATE under 'if (is_pvh_domain())' then it seems to
> me that things should be fine.

Nope, already tried that. Something is not being set properly in terms
of scaling or offset, and the first timer gets set years in future! Since
you are already going to look into it, I'll let you :)..

thanks,
Mukesh


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel