Received-SPF: pass (google.com: domain of linux-kernel+bounces-141425-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Date: Thu, 11 Apr 2024 10:46:38 -0700
In-Reply-To: <CALMp9eRwsyBUHRtjKZDyU6i13hr5tif3ty7tpNjfs=Zq3RA8RA@mail.gmail.com>
Precedence: bulk
Mime-Version: 1.0
References: <20240126085444.324918-1-xiong.y.zhang@linux.intel.com>
 <20240126085444.324918-2-xiong.y.zhang@linux.intel.com> <ZhgYD4B1szpbvlHq@google.com>
 <56a98cae-36c5-40f8-8554-77f9d9c9a1b0@linux.intel.com> <CALMp9eRwsyBUHRtjKZDyU6i13hr5tif3ty7tpNjfs=Zq3RA8RA@mail.gmail.com>
Message-ID: <Zhgh_vQYx2MCzma6@google.com>
Subject: Re: [RFC PATCH 01/41] perf: x86/intel: Support PERF_PMU_CAP_VPMU_PASSTHROUGH
From: Sean Christopherson <seanjc@google.com>
To: Jim Mattson <jmattson@google.com>
Cc: Kan Liang <kan.liang@linux.intel.com>, Xiong Zhang <xiong.y.zhang@linux.intel.com>, 
	pbonzini@redhat.com, peterz@infradead.org, mizhang@google.com, 
	kan.liang@intel.com, zhenyuw@linux.intel.com, dapeng1.mi@linux.intel.com, 
	kvm@vger.kernel.org, linux-perf-users@vger.kernel.org, 
	linux-kernel@vger.kernel.org, zhiyuan.lv@intel.com, eranian@google.com, 
	irogers@google.com, samantha.alt@intel.com, like.xu.linux@gmail.com, 
	chao.gao@intel.com
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable

On Thu, Apr 11, 2024, Jim Mattson wrote:
> On Thu, Apr 11, 2024 at 10:21=E2=80=AFAM Liang, Kan <kan.liang@linux.inte=
l.com> wrote:
> > On 2024-04-11 1:04 p.m., Sean Christopherson wrote:
> > > On Fri, Jan 26, 2024, Xiong Zhang wrote:
> > >> From: Kan Liang <kan.liang@linux.intel.com>
> > >>
> > >> Define and apply the PERF_PMU_CAP_VPMU_PASSTHROUGH flag for the vers=
ion 4
> > >> and later PMUs
> > >
> > > Why?  I get that is an RFC, but it's not at all obvious to me why thi=
s needs to
> > > take a dependency on v4+.
> >
> > The IA32_PERF_GLOBAL_STATUS_RESET/SET MSRs are introduced in v4. They
> > are used in the save/restore of PMU state. Please see PATCH 23/41.
> > So it's limited to v4+ for now.
>=20
> Prior to version 4, semi-passthrough is possible, but IA32_PERF_GLOBAL_ST=
ATUS
> has to be intercepted and emulated, since it is non-trivial to set bits i=
n
> this MSR.

Ah, then this _perf_ capability should be PERF_PMU_CAP_WRITABLE_GLOBAL_STAT=
US or
so, especially since it's introduced in advance of the KVM side of things. =
 Then
whether or not to support a mediated PMU becomes a KVM decision, e.g. inter=
cepting
accesses to IA32_PERF_GLOBAL_STATUS doesn't seem like a complete deal break=
er
(or maybe it is, I now see the comment about it being used to do the contex=
t switch).

And peeking ahead, IIUC perf effectively _forces_ a passthrough model when
has_vpmu_passthrough_cap() is true, which is wrong.  There needs to be a us=
er/admin
opt-in (or opt-out) to that behavior, at a kernel/perf level, not just at a=
 KVM
level.  Hmm, or is perf relying on KVM to do that right thing?  I.e. relyin=
g on
KVM to do perf_guest_{enter,exit}() if and only if the PMU can support the
passthrough model.

If that's the case, most of the has_vpmu_passthrough_cap() checks are grati=
utous
and confusing, e.g. just WARN if KVM (or some other module) tries to trigge=
r a
PMU context switch when it's not supported by perf.