KVM: SVM: Add support for AMD's OSVW feature in guests

In some cases guests should not provide workarounds for errata even when the
physical processor is affected. For example, because of erratum 400 on family
10h processors a Linux guest will read an MSR (resulting in VMEXIT) before
going to idle in order to avoid getting stuck in a non-C0 state. This is not
necessary: HLT and IO instructions are intercepted and therefore there is no
reason for erratum 400 workaround in the guest.

This patch allows us to present a guest with certain errata as fixed,
regardless of the state of actual hardware.

Signed-off-by: Boris Ostrovsky <boris.ostrovsky@amd.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>

arch/x86/include/asm/kvm_host.h

@@ -478,6 +478,12 @@ struct kvm_vcpu_arch {
 		u32 id;
 		bool send_user_only;
 	} apf;
+
+	/* OSVW MSRs (AMD only) */
+	struct {
+		u64 length;
+		u64 status;
+	} osvw;
 };
 
 struct kvm_arch {

arch/x86/kvm/cpuid.c

@@ -236,7 +236,7 @@ static int do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 	const u32 kvm_supported_word6_x86_features =
 		F(LAHF_LM) | F(CMP_LEGACY) | 0 /*SVM*/ | 0 /* ExtApicSpace */ |
 		F(CR8_LEGACY) | F(ABM) | F(SSE4A) | F(MISALIGNSSE) |
-		F(3DNOWPREFETCH) | 0 /* OSVW */ | 0 /* IBS */ | F(XOP) |
+		F(3DNOWPREFETCH) | F(OSVW) | 0 /* IBS */ | F(XOP) |
 		0 /* SKINIT, WDT, LWP */ | F(FMA4) | F(TBM);
 
 	/* cpuid 0xC0000001.edx */

arch/x86/kvm/cpuid.h

@@ -43,4 +43,12 @@ static inline bool guest_cpuid_has_fsgsbase(struct kvm_vcpu *vcpu)
 	return best && (best->ebx & bit(X86_FEATURE_FSGSBASE));
 }
 
+static inline bool guest_cpuid_has_osvw(struct kvm_vcpu *vcpu)
+{
+	struct kvm_cpuid_entry2 *best;
+
+	best = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
+	return best && (best->ecx & bit(X86_FEATURE_OSVW));
+}
+
 #endif

arch/x86/kvm/svm.c

@@ -110,6 +110,12 @@ struct nested_state {
 #define MSRPM_OFFSETS	16
 static u32 msrpm_offsets[MSRPM_OFFSETS] __read_mostly;
 
+/*
+ * Set osvw_len to higher value when updated Revision Guides
+ * are published and we know what the new status bits are
+ */
+static uint64_t osvw_len = 4, osvw_status;
+
 struct vcpu_svm {
 	struct kvm_vcpu vcpu;
 	struct vmcb *vmcb;
@@ -556,6 +562,27 @@ static void svm_init_erratum_383(void)
 	erratum_383_found = true;
 }
 
+static void svm_init_osvw(struct kvm_vcpu *vcpu)
+{
+	/*
+	 * Guests should see errata 400 and 415 as fixed (assuming that
+	 * HLT and IO instructions are intercepted).
+	 */
+	vcpu->arch.osvw.length = (osvw_len >= 3) ? (osvw_len) : 3;
+	vcpu->arch.osvw.status = osvw_status & ~(6ULL);
+
+	/*
+	 * By increasing VCPU's osvw.length to 3 we are telling the guest that
+	 * all osvw.status bits inside that length, including bit 0 (which is
+	 * reserved for erratum 298), are valid. However, if host processor's
+	 * osvw_len is 0 then osvw_status[0] carries no information. We need to
+	 * be conservative here and therefore we tell the guest that erratum 298
+	 * is present (because we really don't know).
+	 */
+	if (osvw_len == 0 && boot_cpu_data.x86 == 0x10)
+		vcpu->arch.osvw.status |= 1;
+}
+
 static int has_svm(void)
 {
 	const char *msg;
@@ -620,6 +647,36 @@ static int svm_hardware_enable(void *garbage)
 		__get_cpu_var(current_tsc_ratio) = TSC_RATIO_DEFAULT;
 	}
 
+
+	/*
+	 * Get OSVW bits.
+	 *
+	 * Note that it is possible to have a system with mixed processor
+	 * revisions and therefore different OSVW bits. If bits are not the same
+	 * on different processors then choose the worst case (i.e. if erratum
+	 * is present on one processor and not on another then assume that the
+	 * erratum is present everywhere).
+	 */
+	if (cpu_has(&boot_cpu_data, X86_FEATURE_OSVW)) {
+		uint64_t len, status = 0;
+		int err;
+
+		len = native_read_msr_safe(MSR_AMD64_OSVW_ID_LENGTH, &err);
+		if (!err)
+			status = native_read_msr_safe(MSR_AMD64_OSVW_STATUS,
+						      &err);
+
+		if (err)
+			osvw_status = osvw_len = 0;
+		else {
+			if (len < osvw_len)
+				osvw_len = len;
+			osvw_status |= status;
+			osvw_status &= (1ULL << osvw_len) - 1;
+		}
+	} else
+		osvw_status = osvw_len = 0;
+
 	svm_init_erratum_383();
 
 	return 0;
@@ -1186,6 +1243,8 @@ static struct kvm_vcpu *svm_create_vcpu(struct kvm *kvm, unsigned int id)
 	if (kvm_vcpu_is_bsp(&svm->vcpu))
 		svm->vcpu.arch.apic_base |= MSR_IA32_APICBASE_BSP;
 
+	svm_init_osvw(&svm->vcpu);
+
 	return &svm->vcpu;
 
 free_page4:

arch/x86/kvm/x86.c

@@ -1675,6 +1675,16 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, u32 msr, u64 data)
 		 */
 		pr_unimpl(vcpu, "ignored wrmsr: 0x%x data %llx\n", msr, data);
 		break;
+	case MSR_AMD64_OSVW_ID_LENGTH:
+		if (!guest_cpuid_has_osvw(vcpu))
+			return 1;
+		vcpu->arch.osvw.length = data;
+		break;
+	case MSR_AMD64_OSVW_STATUS:
+		if (!guest_cpuid_has_osvw(vcpu))
+			return 1;
+		vcpu->arch.osvw.status = data;
+		break;
 	default:
 		if (msr && (msr == vcpu->kvm->arch.xen_hvm_config.msr))
 			return xen_hvm_config(vcpu, data);
@@ -1959,6 +1969,16 @@ int kvm_get_msr_common(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata)
 		 */
 		data = 0xbe702111;
 		break;
+	case MSR_AMD64_OSVW_ID_LENGTH:
+		if (!guest_cpuid_has_osvw(vcpu))
+			return 1;
+		data = vcpu->arch.osvw.length;
+		break;
+	case MSR_AMD64_OSVW_STATUS:
+		if (!guest_cpuid_has_osvw(vcpu))
+			return 1;
+		data = vcpu->arch.osvw.status;
+		break;
 	default:
 		if (kvm_pmu_msr(vcpu, msr))
 			return kvm_pmu_get_msr(vcpu, msr, pdata);