On 25.02.2021 14:03, Jan Beulich wrote: > When none of the physical address bits in PTEs are reserved, we can't > create any 4k (leaf) PTEs which would trigger reserved bit faults. Hence > the present SHOPT_FAST_FAULT_PATH machinery needs to be suppressed in > this case, which is most easily achieved by never creating any magic > entries. > > To compensate a little, eliminate sh_write_p2m_entry_post()'s impact on > such hardware.
As to 4.15: Without this shadow mode simply won't work on such (new) hardware. Hence something needs to be done anyway. An alternative would be to limit the change to just the guest-no-present entries (to at least allow PV guests to be migrated), and refuse to enable shadow mode for HVM guests on such hardware. (In this case we'd probably better take care of ... > While at it, also avoid using an MMIO magic entry when that would > truncate the incoming GFN. ... this long-standing issue, perhaps as outlined in a post-commit- message remark.) The main risk here is (in particular for the MMIO part of the change I suppose) execution suddenly going a different path, which has been unused / untested (for this specific case) for years. Jan
