Disable LBR virtualization if the CPU doesn't support callstacks, which
were introduced in HSW (see commit e9d7f7cd97c4 ("perf/x86/intel: Add
basic Haswell LBR call stack support"), as KVM unconditionally configures
the perf LBR event with PERF_SAMPLE_BRANCH_CALL_STACK, i.e. LBR
virtualization always fails on pre-HSW CPUs.
Simply disable LBR support on such CPUs, as it has never worked, i.e.
there is no risk of breaking an existing setup, and figuring out a way
to performantly context switch LBRs on old CPUs is not worth the effort.
Fixes: be635e34c284 ("KVM: vmx/pmu: Expose LBR_FMT in the MSR_IA32_PERF_CAPABILITIES")
Cc: Mingwei Zhang <mizhang@xxxxxxxxxx>
Cc: Jim Mattson <jmattson@xxxxxxxxxx>
Signed-off-by: Sean Christopherson <seanjc@xxxxxxxxxx>
---
arch/x86/kvm/vmx/vmx.c | 10 +++++++++-
1 file changed, 9 insertions(+), 1 deletion(-)
diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index 2a7cd66988a5..25a7652bee7c 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -7859,7 +7859,15 @@ static __init u64 vmx_get_perf_capabilities(void)
if (!cpu_feature_enabled(X86_FEATURE_ARCH_LBR)) {
x86_perf_get_lbr(&vmx_lbr_caps);
- if (vmx_lbr_caps.nr)
+
+ /*
+ * KVM requires LBR callstack support, as the overhead due to
+ * context switching LBRs without said support is too high.
+ * See intel_pmu_create_guest_lbr_event() for more info.
+ */
+ if (!vmx_lbr_caps.has_callstack)
+ memset(&vmx_lbr_caps, 0, sizeof(vmx_lbr_caps));
+ else if (vmx_lbr_caps.nr)
perf_cap |= host_perf_cap & PMU_CAP_LBR_FMT;
}
--
2.44.0.278.ge034bb2e1d-goog
