[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <175079269575.517785.12142408340158295471.b4-ty@google.com>
Date: Tue, 24 Jun 2025 12:38:20 -0700
From: Sean Christopherson <seanjc@...gle.com>
To: Sean Christopherson <seanjc@...gle.com>, Paolo Bonzini <pbonzini@...hat.com>
Cc: kvm@...r.kernel.org, linux-kernel@...r.kernel.org,
Alexander Potapenko <glider@...gle.com>, James Houghton <jthoughton@...gle.com>,
Peter Gonda <pgonda@...gle.com>, Tom Lendacky <thomas.lendacky@....com>
Subject: Re: [PATCH 0/2] KVM: SVM: Fix a NULL VMSA deref with MOVE_ENC_CONTEXT
On Mon, 02 Jun 2025 15:44:57 -0700, Sean Christopherson wrote:
> Fix a NULL VMSA deref bug (which is probably the tip of the iceberg with
> respect to what all can go wrong) due to a race between KVM_CREATE_VCPU and
> KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM, where a non-SEV-ES vCPU can be created in
> an SEV-ES VM.
>
> Found by running syzkaller on a bare metal SEV-ES host. C repro below.
>
> [...]
Applied to kvm-x86 fixes. Paolo's "Queued, thanks!" seems to have been
premature (though Paolo's mail saved me; I completely forgot about swapping
EINVAL to EBUSY).
[1/2] KVM: SVM: Reject SEV{-ES} intra host migration if vCPU creation is in-flight
https://github.com/kvm-x86/linux/commit/ecf371f8b02d
[2/2] KVM: SVM: Initialize vmsa_pa in VMCB to INVALID_PAGE if VMSA page is NULL
https://github.com/kvm-x86/linux/commit/48f15f62418
--
https://github.com/kvm-x86/kvm-unit-tests/tree/next
Powered by blists - more mailing lists