KVM: x86/mmu: Avoid collision with !PRESENT SPTEs in TDP MMU lpage stats
authorSean Christopherson <seanjc@google.com>
Tue, 3 Aug 2021 04:46:06 +0000 (21:46 -0700)
committerPaolo Bonzini <pbonzini@redhat.com>
Fri, 20 Aug 2021 20:06:34 +0000 (16:06 -0400)
commit088acd23526647844aec1c39db4ad02552c86c7b
tree3613091262a70aa0414d2b692a3dd03106b6dbd8
parent4293ddb788c1a98bdfa6479bcfd63ad5ce0a5ce6
KVM: x86/mmu: Avoid collision with !PRESENT SPTEs in TDP MMU lpage stats

Factor in whether or not the old/new SPTEs are shadow-present when
adjusting the large page stats in the TDP MMU.  A modified MMIO SPTE can
toggle the page size bit, as bit 7 is used to store the MMIO generation,
i.e. is_large_pte() can get a false positive when called on a MMIO SPTE.
Ditto for nuking SPTEs with REMOVED_SPTE, which sets bit 7 in its magic
value.

Opportunistically move the logic below the check to verify at least one
of the old/new SPTEs is shadow present.

Use is/was_leaf even though is/was_present would suffice.  The code
generation is roughly equivalent since all flags need to be computed
prior to the code in question, and using the *_leaf flags will minimize
the diff in a future enhancement to account all pages, i.e. will change
the check to "is_leaf != was_leaf".

Reviewed-by: David Matlack <dmatlack@google.com>
Reviewed-by: Ben Gardon <bgardon@google.com>
Fixes: 1699f65c8b65 ("kvm/x86: Fix 'lpages' kvm stat for TDM MMU")
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Mingwei Zhang <mizhang@google.com>
Message-Id: <20210803044607.599629-3-mizhang@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
arch/x86/kvm/mmu/tdp_mmu.c