(Not so) large (anymore) parts of this series are to further isolate pieces which are needed for HVM only, and hence would better not be built with HVM=n. But there are also a few other items which I've noticed along the road, including the new in v2 4th patch.
1: mm: no-one passes a NULL domain to init_xen_l4_slots() 2: shadow: sh_update_linear_entries() is a no-op for PV 3: mm: monitor table is HVM-only 4: adjustments to guest handle treatment Jan