Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Reply-To: Sean Christopherson <seanjc@google.com>
Date:   Tue, 14 Sep 2021 14:09:49 -0700
Message-Id: <20210914210951.2994260-1-seanjc@google.com>
Mime-Version: 1.0
Subject: [PATCH 0/2] KVM: SEV: RECEIVE_UPDATE_DATA bug fixes
From:   Sean Christopherson <seanjc@google.com>
To:     Paolo Bonzini <pbonzini@redhat.com>
Cc:     Sean Christopherson <seanjc@google.com>,
        Vitaly Kuznetsov <vkuznets@redhat.com>,
        Wanpeng Li <wanpengli@tencent.com>,
        Jim Mattson <jmattson@google.com>,
        Joerg Roedel <joro@8bytes.org>, kvm@vger.kernel.org,
        linux-kernel@vger.kernel.org, Peter Gonda <pgonda@google.com>,
        Marc Orr <marcorr@google.com>,
        Tom Lendacky <thomas.lendacky@amd.com>,
        Brijesh Singh <brijesh.singh@amd.com>,
        Masahiro Kozuka <masa.koz@kozuka.jp>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Patch 1 pins the target of RECEIVE_UPDATE_DATA for write instead of read.
The SEV API clearly states that the PSP writes guest memory, which makes
sense given that the guest is receiving data.

Patch 2 adds a CLFLUSH of the guest memory as there is no requirement
that the page already be in an encrypted state before RECEIVE_UPDATE_DATA,
i.e. the cache may have dirty, unencrypted data.

On my end, these patches are compile-tested only as I don't have a
userspace VMM that supports SEV live migration, nor are there tests for
any of this stuff.

Masahiro Kozuka (1):
  KVM: SEV: Flush cache on non-coherent systems before
    RECEIVE_UPDATE_DATA

Sean Christopherson (1):
  KVM: SEV: Pin guest memory for write for RECEIVE_UPDATE_DATA

 arch/x86/kvm/svm/sev.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

-- 
2.33.0.309.g3052b89438-goog