Re: [PATCH v5 4/5] KVM: VMX: Allow direct access to MSR_IA32_SPEC_CTRL

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 02/01/2018 03:19 PM, Konrad Rzeszutek Wilk wrote:
.snip..
+/* Is SPEC_CTRL intercepted for the currently running vCPU? */
+static bool spec_ctrl_intercepted(struct kvm_vcpu *vcpu)
+{
+	unsigned long *msr_bitmap;
+	int f = sizeof(unsigned long);
+
+	if (!cpu_has_vmx_msr_bitmap())
+		return true;
+
+	msr_bitmap = is_guest_mode(vcpu) ?
+			to_vmx(vcpu)->nested.vmcs02.msr_bitmap :
+			to_vmx(vcpu)->vmcs01.msr_bitmap;
+
+	return !!test_bit(MSR_IA32_SPEC_CTRL, msr_bitmap + 0x800 / f);
+}
+
..snip..
@@ -3359,6 +3393,34 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
  	case MSR_IA32_TSC:
  		kvm_write_tsc(vcpu, msr_info);
  		break;
+	case MSR_IA32_SPEC_CTRL:
+		if (!msr_info->host_initiated &&
+		    !guest_cpuid_has(vcpu, X86_FEATURE_IBRS) &&
+		    !guest_cpuid_has(vcpu, X86_FEATURE_SPEC_CTRL))
+			return 1;
+
+		vmx->spec_ctrl_used = true;
+
+		/* The STIBP bit doesn't fault even if it's not advertised */
+		if (data & ~(SPEC_CTRL_IBRS | SPEC_CTRL_STIBP))
+			return 1;
+
+		vmx->spec_ctrl = data;
+
+		/*
+		 * When it's written (to non-zero) for the first time, pass
+		 * it through. This means we don't have to take the perf

.. But only if it is a nested guest (as you have && is_guest_mode).

Do you want to update the comment a bit?

+		 * hit of saving it on vmexit for the common case of guests
+		 * that don't use it.
+		 */
+		if (cpu_has_vmx_msr_bitmap() && data &&
+		    spec_ctrl_intercepted(vcpu) &&
+		    is_guest_mode(vcpu))
                     ^^^^^^^^^^^^^^^^^^ <=== here

Would it be perhaps also good to mention the complexity of how
we ought to be handling L1 and L2 guests in the commit?

We are all stressed and I am sure some of us haven't gotten much
sleep - but it can help in say three months when some unluckly new
soul is trying to understand this and gets utterly confused.

Yup, I will go through the patches and add as much details as possible.

And yes, the is_guest_mode(vcpu) here is inverted :D I blame the late
night :)


+			vmx_disable_intercept_for_msr(
+					vmx->vmcs01.msr_bitmap,
+					MSR_IA32_SPEC_CTRL,
+					MSR_TYPE_RW);
+		break;
  	case MSR_IA32_PRED_CMD:
  		if (!msr_info->host_initiated &&
  		    !guest_cpuid_has(vcpu, X86_FEATURE_IBPB) &&

Amazon Development Center Germany GmbH
Berlin - Dresden - Aachen
main office: Krausenstr. 38, 10117 Berlin
Geschaeftsfuehrer: Dr. Ralf Herbrich, Christian Schlaeger
Ust-ID: DE289237879
Eingetragen am Amtsgericht Charlottenburg HRB 149173 B




[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux