[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ZdfQ4jI8yT-bvbV4@google.com>
Date: Thu, 22 Feb 2024 14:55:30 -0800
From: Sean Christopherson <seanjc@...gle.com>
To: Paolo Bonzini <pbonzini@...hat.com>, Marc Zyngier <maz@...nel.org>,
Oliver Upton <oliver.upton@...ux.dev>, Anup Patel <anup@...infault.org>,
Paul Walmsley <paul.walmsley@...ive.com>, Palmer Dabbelt <palmer@...belt.com>,
Albert Ou <aou@...s.berkeley.edu>, Christian Borntraeger <borntraeger@...ux.ibm.com>,
Janosch Frank <frankja@...ux.ibm.com>, Claudio Imbrenda <imbrenda@...ux.ibm.com>, kvm@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org, kvmarm@...ts.linux.dev,
kvm-riscv@...ts.infradead.org, linux-riscv@...ts.infradead.org,
linux-kernel@...r.kernel.org, Vishal Annapurve <vannapurve@...gle.com>,
Ackerley Tng <ackerleytng@...gle.com>, Andrew Jones <andrew.jones@...ux.dev>,
Tom Lendacky <thomas.lendacky@....com>, Michael Roth <michael.roth@....com>,
Peter Gonda <pgonda@...gle.com>
Subject: Re: [PATCH v8 10/10] KVM: selftests: Add a basic SEV smoke test
On Thu, Feb 22, 2024, Sean Christopherson wrote:
> On Tue, Feb 06, 2024, Sean Christopherson wrote:
> > On Fri, Feb 02, 2024, Sean Christopherson wrote:
> > > +int main(int argc, char *argv[])
> > > +{
> > > + TEST_REQUIRE(is_kvm_sev_supported());
> >
> > This also needs
> >
> > TEST_REQUIRE(kvm_cpu_has(X86_FEATURE_SEV));
> >
> > to handle the case where the platform supports SEV, i.e. /dev/sev exists, but
> > KVM doesn't support SEV, e.g. if TDP is disabled, if SEV was explicitly disabled
> > via module param, etc.
>
> Thinking more about this, I think we should simply delete is_kvm_sev_supported().
> (a) it obviously doesn't query _KVM_ support, and (b) if KVM says SEV is supported,
> then it darn well actually be supported.
Ugh, and selftests also need to handle the scenario where SEV is enabled, but all
ASIDs are assigned to SEV-ES. We could try to create a dummy VM and check for an
-EBUSY return, but that is ugly and could result in missing KVM bugs due to tests
being skipped instead of failing.
That's a future problem though, because I think we'll need new KVM functionality
to surface that information.
Powered by blists - more mailing lists