The Linux Kernel Mailing List
 help / color / mirror / Atom feed
From: Zide Chen <zide.chen@intel.com>
To: Sean Christopherson <seanjc@google.com>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Peter Zijlstra <peterz@infradead.org>
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
	Jim Mattson <jmattson@google.com>,
	Mingwei Zhang <mizhang@google.com>,
	Zide Chen <zide.chen@intel.com>,
	Das Sandipan <Sandipan.Das@amd.com>,
	Shukla Manali <Manali.Shukla@amd.com>,
	Dapeng Mi <dapeng1.mi@linux.intel.com>,
	Falcon Thomas <thomas.falcon@intel.com>,
	Xudong Hao <xudong.hao@intel.com>
Subject: [PATCH 11/15] KVM: x86/pmu: Emulate streamlined Freeze-LBR-on-PMI
Date: Tue,  7 Jul 2026 11:34:01 -0700	[thread overview]
Message-ID: <20260707183405.15571-12-zide.chen@intel.com> (raw)
In-Reply-To: <20260707183405.15571-1-zide.chen@intel.com>

PerfMon v4 streamlined the Freeze-LBR-on-PMI mechanism. When
DEBUGCTLMSR_FREEZE_LBRS_ON_PMI is set and a PMI fires, hardware sets
IA32_PERF_GLOBAL_STATUS.LBR_Frz instead of clearing DEBUGCTLMSR_LBR.

Guest PerfMon v4+ is supported under mediated vPMU only. When KVM
relays a guest-induced PMI, the mediated vPMU framework already
propagates IA32_PERF_GLOBAL_STATUS.LBR_Frz, and no additional handling
is required.

For PMIs generated by KVM-emulated PMU events, however, KVM must
emulate IA32_PERF_GLOBAL_STATUS.LBR_Frz so that guest LBR recording
is frozen as required by the architectural behavior.

Signed-off-by: Zide Chen <zide.chen@intel.com>
---
 arch/x86/kvm/vmx/pmu_intel.c | 47 ++++++++++++++++++++++--------------
 1 file changed, 29 insertions(+), 18 deletions(-)

diff --git a/arch/x86/kvm/vmx/pmu_intel.c b/arch/x86/kvm/vmx/pmu_intel.c
index cb6f9c272e03..556c119d5e91 100644
--- a/arch/x86/kvm/vmx/pmu_intel.c
+++ b/arch/x86/kvm/vmx/pmu_intel.c
@@ -660,6 +660,8 @@ static void intel_pmu_refresh(struct kvm_vcpu *vcpu)
 
 	pmu->global_status_rsvd = pmu->global_ctrl_rsvd
 			& ~(GLOBAL_STATUS_BUFFER_OVF | GLOBAL_STATUS_COND_CHG);
+	if (pmu->version > 3)
+		pmu->global_status_rsvd &= ~GLOBAL_STATUS_LBRS_FROZEN;
 	if (vmx_pt_mode_is_host_guest())
 		pmu->global_status_rsvd &= ~GLOBAL_STATUS_TRACE_TOPAPMI;
 
@@ -718,32 +720,41 @@ static void intel_pmu_reset(struct kvm_vcpu *vcpu)
 }
 
 /*
- * Emulate LBR_On_PMI behavior for 1 < pmu.version < 4.
- *
- * If Freeze_LBR_On_PMI = 1, the LBR is frozen on PMI and
- * the KVM emulates to clear the LBR bit (bit 0) in IA32_DEBUGCTL.
- *
- * Guest needs to re-enable LBR to resume branches recording.
+ * Emulate legacy and streamlined Freeze_LBR_On_PMI behavior.
+ * In either case, guest needs to re-enable LBR to resume branches recording.
  */
-static void intel_pmu_legacy_freezing_lbrs_on_pmi(struct kvm_vcpu *vcpu)
+static void intel_pmu_freeze_lbr_on_pmi(struct kvm_vcpu *vcpu)
 {
-	u64 data = vmx_guest_debugctl_read();
+	u8 version;
+	u64 data;
 
-	if (data & DEBUGCTLMSR_FREEZE_LBRS_ON_PMI) {
-		data &= ~DEBUGCTLMSR_LBR;
-		vmx_guest_debugctl_write(vcpu, data);
+	if (!intel_pmu_lbr_is_enabled(vcpu))
+		return;
+
+	data = vmx_guest_debugctl_read();
+	version = vcpu_to_pmu(vcpu)->version;
+
+	if (!(data & DEBUGCTLMSR_FREEZE_LBRS_ON_PMI))
+		return;
+
+	if (version > 1 && version < 4) {
+		if (data & DEBUGCTLMSR_LBR) {
+			data &= ~DEBUGCTLMSR_LBR;
+			vmx_guest_debugctl_write(vcpu, data);
+		}
+	} else if (vcpu_to_lbr_desc(vcpu)->msr_passthrough &&
+		   kvm_vcpu_has_mediated_pmu(vcpu)) {
+		/*
+		 * This will be restored to guest before VM-Entry, setting
+		 * LBR_Frz to freeze LBR recording until the guest clears it.
+		 */
+		vcpu_to_pmu(vcpu)->global_status |= GLOBAL_STATUS_LBRS_FROZEN;
 	}
 }
 
 static void intel_pmu_deliver_pmi(struct kvm_vcpu *vcpu)
 {
-	u8 version = vcpu_to_pmu(vcpu)->version;
-
-	if (!intel_pmu_lbr_is_enabled(vcpu))
-		return;
-
-	if (version > 1 && version < 4)
-		intel_pmu_legacy_freezing_lbrs_on_pmi(vcpu);
+	intel_pmu_freeze_lbr_on_pmi(vcpu);
 }
 
 static void vmx_update_intercept_for_lbr_msrs(struct kvm_vcpu *vcpu, bool set)
-- 
2.54.0


  parent reply	other threads:[~2026-07-07 18:44 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-07-07 18:33 [PATCH 00/15] KVM: x86/pmu: Add mediated vPMU PerfMon v5 support Zide Chen
2026-07-07 18:33 ` [PATCH 01/15] KVM: x86/pmu: Remove redundant Perf Global Status MSR bit definitions Zide Chen
2026-07-07 18:33 ` [PATCH 02/15] KVM: x86/pmu: Rename all_valid_pmc_idx to all_valid_pmc_mask Zide Chen
2026-07-07 18:33 ` [PATCH 03/15] KVM: x86/pmu: Rename reserved_bits to eventsel_rsvd in kvm_pmu Zide Chen
2026-07-07 18:33 ` [PATCH 04/15] KVM: x86/pmu: Add PMC bitmap accessor helpers Zide Chen
2026-07-07 18:33 ` [PATCH 05/15] KVM: x86/pmu: Drop nr_arch_{gp,fixed}_counters from kvm_pmu Zide Chen
2026-07-07 18:33 ` [PATCH 06/15] KVM: x86/pmu: Expose kvm_host_pmu to vendor modules Zide Chen
2026-07-07 18:33 ` [PATCH 07/15] perf/x86: Plumb counter bitmap from x86_pmu to x86_pmu_cap Zide Chen
2026-07-07 18:33 ` [PATCH 08/15] KVM: x86/pmu: Switch to bitmask-based KVM PMU capabilities Zide Chen
2026-07-07 18:33 ` [PATCH 09/15] perf/x86: Remove num_counters_{gp,fixed} from x86_pmu_capability Zide Chen
2026-07-07 18:34 ` [PATCH 10/15] KVM: x86/pmu: Emulate the GLOBAL_STATUS_SET and GLOBAL_INUSE MSRs Zide Chen
2026-07-07 18:34 ` Zide Chen [this message]
2026-07-07 18:34 ` [PATCH 12/15] KVM: x86/pmu: Populate CPUID.0AH:ECX fixed-counter bitmap Zide Chen
2026-07-07 18:34 ` [PATCH 13/15] KVM: x86/pmu: Ignore AnyThread bit if CPUID.0AH:EDX[15] is not set Zide Chen
2026-07-07 18:34 ` [PATCH 14/15] KVM: x86/pmu: Advertise PerfMon version 5 on Intel hosts Zide Chen
2026-07-07 18:34 ` [PATCH 15/15] KVM: selftests: Support fixed counters bitmap in pmu_counters_test Zide Chen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260707183405.15571-12-zide.chen@intel.com \
    --to=zide.chen@intel.com \
    --cc=Manali.Shukla@amd.com \
    --cc=Sandipan.Das@amd.com \
    --cc=dapeng1.mi@linux.intel.com \
    --cc=jmattson@google.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mizhang@google.com \
    --cc=pbonzini@redhat.com \
    --cc=peterz@infradead.org \
    --cc=seanjc@google.com \
    --cc=thomas.falcon@intel.com \
    --cc=xudong.hao@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox