On Wed, Apr 17, 2013 at 05:41:19PM +0300, Abel Gordon wrote:
>
>
> Gleb Natapov <gleb@xxxxxxxxxx> wrote on 17/04/2013 05:10:28 PM:
>
>
> > On Wed, Apr 17, 2013 at 02:53:10PM +0300, Abel Gordon wrote:
> > > Allocate a shadow vmcs used by the processor to shadow part of the
> fields
> > > stored in the software defined VMCS12 (let L1 access fields without
> causing
> > > exits). Note we keep a shadow vmcs only for the current vmcs12.
> > Once a vmcs12
> > > becomes non-current, its shadow vmcs is released.
> > >
> > >
> > > Signed-off-by: Abel Gordon <abelg@xxxxxxxxxx>
> > > ---
> > > arch/x86/kvm/vmx.c | 15 +++++++++++++++
> > > 1 file changed, 15 insertions(+)
> > >
> > > --- .before/arch/x86/kvm/vmx.c 2013-04-17 14:20:50.000000000 +0300
> > > +++ .after/arch/x86/kvm/vmx.c 2013-04-17 14:20:50.000000000 +0300
> > > @@ -355,6 +355,7 @@ struct nested_vmx {
> > > /* The host-usable pointer to the above */
> > > struct page *current_vmcs12_page;
> > > struct vmcs12 *current_vmcs12;
> > > + struct vmcs *current_shadow_vmcs;
> > >
> > > /* vmcs02_list cache of VMCSs recently used to run L2 guests */
> > > struct list_head vmcs02_pool;
> > > @@ -5980,6 +5981,7 @@ static int handle_vmptrld(struct kvm_vcp
> > > gva_t gva;
> > > gpa_t vmptr;
> > > struct x86_exception e;
> > > + struct vmcs *shadow_vmcs;
> > >
> > > if (!nested_vmx_check_permission(vcpu))
> > > return 1;
> > > @@ -6026,6 +6028,19 @@ static int handle_vmptrld(struct kvm_vcp
> > > vmx->nested.current_vmptr = vmptr;
> > > vmx->nested.current_vmcs12 = new_vmcs12;
> > > vmx->nested.current_vmcs12_page = page;
> > > + if (enable_shadow_vmcs) {
> > > + shadow_vmcs = alloc_vmcs();
> > Next patch frees vmx->nested.current_shadow_vmcs couple of lines above.
> > What about reusing previous page instead of allocation new one each
> > time?
>
> Yes, we could have a single shadow vmcs per L1 vcpu that is used to shadow
> multiple L2 vcpus. I preferred not to do that because I didn't want to
> share the same page (physical vmcs) for different vmcs12s. However, this is
> not an issues because we overwrite the shadowed fields every time we sync
> the content.
> It's your call. If you prefer to re-use, I'll send a new version that
> do that. Please confirm.
Yes, I prefer to reuse explicitly. There is a big chance that the same
page will be reused even if you go through freeing and allocation.
--
Gleb.
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
