Received-SPF: pass (google.com: domain of linux-kernel+bounces-94817-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Reply-To: Sean Christopherson <seanjc@google.com>
Date: Wed,  6 Mar 2024 17:13:41 -0800
Precedence: bulk
Mime-Version: 1.0
Message-ID: <20240307011344.835640-1-seanjc@google.com>
Subject: [PATCH 0/3] KVM: VMX: Disable LBRs if CPU doesn't have callstacks
From: Sean Christopherson <seanjc@google.com>
To: Peter Zijlstra <peterz@infradead.org>, Ingo Molnar <mingo@redhat.com>, 
	Arnaldo Carvalho de Melo <acme@kernel.org>, Namhyung Kim <namhyung@kernel.org>, 
	Sean Christopherson <seanjc@google.com>, Paolo Bonzini <pbonzini@redhat.com>
Cc: linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org, 
	kvm@vger.kernel.org, Mingwei Zhang <mizhang@google.com>, Jim Mattson <jmattson@google.com>
Content-Type: text/plain; charset="UTF-8"

Disable LBR virtualization if the CPU (or I guess perf) doesn't support
LBR callstacks, as KVM unconditionally creates the associated perf LBR
event with PERF_SAMPLE_BRANCH_CALL_STACK.  That results in perf rejecting
the event, and cause LBR virtualization to silently fail.

This was detected by running vmx_pmu_caps_test on older hardware.  I didn't
tag it for stable because I can't imagine anyone is trying to use KVM's LBR
virtualization on pre-HSW.

Sean Christopherson (3):
  KVM: VMX: Snapshot LBR capabilities during module initialization
  perf/x86/intel: Expose existence of callback support to KVM
  KVM: VMX: Disable LBR virtualization if the CPU doesn't support LBR
    callstacks

 arch/x86/events/intel/lbr.c       |  1 +
 arch/x86/include/asm/perf_event.h |  1 +
 arch/x86/kvm/vmx/pmu_intel.c      |  2 +-
 arch/x86/kvm/vmx/vmx.c            | 17 +++++++++++++----
 arch/x86/kvm/vmx/vmx.h            |  2 ++
 5 files changed, 18 insertions(+), 5 deletions(-)


base-commit: 0c64952fec3ea01cb5b09f00134200f3e7ab40d5
-- 
2.44.0.278.ge034bb2e1d-goog