I have found out something interesting. The attached patch, which I think should fix the bug, is present in kvm-72 Debian sources but not in the 2.6.26-2 kernel source. So maybe rebuilding kvm from source cures the problem; I will test this.
m...@fujiko:/tmp$ diff /home/mud/src/debian-bug-531488/linux-source-2.6.26/arch/x86/kvm/svm.c /tmp/linux-source-2.6.26/arch/x86/kvm/svm.c 1316,1319d1315 < case MSR_K7_PERFCTR0: < case MSR_K7_PERFCTR1: < case MSR_K7_PERFCTR2: < case MSR_K7_PERFCTR3: 1321,1323c1317,1319 < * Just discard all writes to the performance counters; this < * should keep both older linux and windows 64-bit guests < * happy --- > * only support writing 0 to the performance counters for now > * to make Windows happy. Should be replaced by a real > * performance counter emulation later. 1325,1326c1321,1322 < pr_unimpl(vcpu, "unimplemented perfctr wrmsr: 0x%x data 0x%llx\n", ecx, data); < --- > if (data != 0) > goto unhandled; 1328a1325 > unhandled: m...@fujiko:/tmp$
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 5528121..dc91b09 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -1312,16 +1312,19 @@ static int svm_set_msr(struct kvm_vcpu *vcpu, unsigned ecx, u64 data) case MSR_K7_EVNTSEL1: case MSR_K7_EVNTSEL2: case MSR_K7_EVNTSEL3: + case MSR_K7_PERFCTR0: + case MSR_K7_PERFCTR1: + case MSR_K7_PERFCTR2: + case MSR_K7_PERFCTR3: /* - * only support writing 0 to the performance counters for now - * to make Windows happy. Should be replaced by a real - * performance counter emulation later. + * Just discard all writes to the performance counters; this + * should keep both older linux and windows 64-bit guests + * happy */ - if (data != 0) - goto unhandled; + pr_unimpl(vcpu, "unimplemented perfctr wrmsr: 0x%x data 0x%llx\n", ecx, data); + break; default: - unhandled: return kvm_set_msr_common(vcpu, ecx, data); } return 0;
