Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Reply-To: Sean Christopherson <seanjc@google.com>
Date:   Fri, 21 Jul 2023 15:43:35 -0700
Mime-Version: 1.0
Message-ID: <20230721224337.2335137-1-seanjc@google.com>
Subject: [PATCH 0/2] KVM: x86: Acquire SRCU in fastpath handler
From:   Sean Christopherson <seanjc@google.com>
To:     Sean Christopherson <seanjc@google.com>,
        Paolo Bonzini <pbonzini@redhat.com>
Cc:     kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
        Greg Thelen <gthelen@google.com>,
        Aaron Lewis <aaronlewis@google.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Acquire SRCU for read when handling fastpath MSR writes so that side
effects like

Note, the PMU case could (and should) also be fixed by making the PMU
filter code smarter, e.g. by snapshotting which PMC events need to be
emulated, thus avoiding the filter lookup entirely.  But acquiring SRCU
is relatively cheap, and this isn't the first bug of this nature.

Which is a perfect segue into patch 2, which reverts a hack-a-fix to
fudge around SVM needing to do the front half of emulation when skipping
the WRMSR.

Note #2, the fastpath also doesn't honor the MSR filter for TSC_DEADLINE.
That's a problem for another day.

Sean Christopherson (2):
  KVM: x86: Acquire SRCU read lock when handling fastpath MSR writes
  Revert "KVM: SVM: Skip WRMSR fastpath on VM-Exit if next RIP isn't
    valid"

 arch/x86/kvm/svm/svm.c | 10 ++--------
 arch/x86/kvm/x86.c     |  4 ++++
 2 files changed, 6 insertions(+), 8 deletions(-)


base-commit: fdf0eaf11452d72945af31804e2a1048ee1b574c
-- 
2.41.0.487.g6d72f3e995-goog