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;
Date:   Wed, 24 May 2023 13:41:22 -0700
In-Reply-To: <59ef9af0-9528-e220-625a-ff16e6971f23@amd.com>
Mime-Version: 1.0
References: <20230310105346.12302-1-likexu@tencent.com> <20230310105346.12302-6-likexu@tencent.com>
 <ZC99f+AO1tZguu1I@google.com> <509b697f-4e60-94e5-f785-95f7f0a14006@gmail.com>
 <ZDAvDhV/bpPyt3oX@google.com> <34b5dd08-edac-e32f-1884-c8f2b85f7971@gmail.com>
 <59ef9af0-9528-e220-625a-ff16e6971f23@amd.com>
Message-ID: <ZG52cgmjgaqY8jvq@google.com>
Subject: Re: [PATCH 5/5] KVM: x86/pmu: Hide guest counter updates from the
 VMRUN instruction
From:   Sean Christopherson <seanjc@google.com>
To:     Sandipan Das <sandipan.das@amd.com>
Cc:     Like Xu <like.xu.linux@gmail.com>,
        Paolo Bonzini <pbonzini@redhat.com>,
        Ravi Bangoria <ravi.bangoria@amd.com>, kvm@vger.kernel.org,
        linux-kernel@vger.kernel.org,
        Santosh Shukla <santosh.shukla@amd.com>,
        "Tom Lendacky (AMD)" <thomas.lendacky@amd.com>,
        Ananth Narayan <ananth.narayan@amd.com>
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Apr 26, 2023, Sandipan Das wrote:
> Hi Sean, Like,
>=20
> On 4/19/2023 7:11 PM, Like Xu wrote:
> >> Heh, it's very much explicable, it's just not desirable, and you and I=
 would argue
> >> that it's also incorrect.
> >=20
> > This is completely inaccurate from the end guest pmu user's perspective=
.
> >=20
> > I have a toy that looks like virtio-pmu, through which guest users can =
get hypervisor performance data.
> > But the side effect of letting the guest see the VMRUN instruction by d=
efault is unacceptable, isn't it ?
> >=20
> >>
> >> AMD folks, are there plans to document this as an erratum?=EF=BF=BD I =
agree with Like that
> >> counting VMRUN as a taken branch in guest context is a CPU bug, even i=
f the behavior
> >> is known/expected.
> >=20
>=20
> This behaviour is architectural and an erratum will not be issued. Howeve=
r, for clarity, a future
> release of the APM will include additional details like the following:
>=20
>   1) From the perspective of performance monitoring counters, VMRUNs are =
considered as far control
>      transfers and VMEXITs as exceptions.
>=20
>   2) When the performance monitoring counters are set up to count events =
only in certain modes
>      through the "OsUserMode" and "HostGuestOnly" bits, instructions and =
events that change the
>      mode are counted in the target mode. For example, a SYSCALL from CPL=
 3 to CPL 0 with a
>      counter set to count retired instructions with USR=3D1 and OS=3D0 wi=
ll not cause an increment of
>      the counter. However, the SYSRET back from CPL 0 to CPL 3 will cause=
 an increment of the
>      counter and the total count will end up correct. Similarly, when cou=
nting PMCx0C6 (retired
>      far control transfers, including exceptions and interrupts) with Gue=
st=3D1 and Host=3D0, a VMRUN
>      instruction will cause an increment of the counter. However, the sub=
sequent VMEXIT that occurs,
>      since the target is in the host, will not cause an increment of the =
counter and so the total
>      count will end up correct.

The count from the guest's perspective does not "end up correct".  Unlike S=
YSCALL,
where _userspace_ deliberately and synchronously executes a branch instruct=
ion,
VMEXIT and VMRUN are supposed to be transparent to the guest and can be com=
pletely
asynchronous with respect to guest code execution, e.g. if the host is spam=
ming
IRQs, the guest will see a potentially large number of bogus (from it's per=
spective)
branches retired.