K 10
svn:author
V 4
neel
K 8
svn:date
V 27
2013-07-04T07:37:29.120528Z
K 7
svn:log
V 492
Revert r252545.

Change r252545 was misguided because of my misinterpretation of the role
of VPIDs when nested page tables are in use. As it turns out they are
indeed useful to allow the TLB to cache guest mappings in a safe manner.

A side-effect of r252545 was turning off the "enable VPID" VM execution
control that in turn caused all guest mappings to be flushed on every
VM entry and exit. This is necessary for correctness but is probably a
drag on performance.

Pointed out by:	grehan

END