Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm

Pull more kvm fixes from Paolo Bonzini:
 "Five small fixes.

  The nested migration bug will be fixed with a better API in 5.10 or
  5.11, for now this is a fix that works with existing userspace but
  keeps the current ugly API"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: SVM: Add a dedicated INVD intercept routine
  KVM: x86: Reset MMU context if guest toggles CR4.SMAP or CR4.PKE
  KVM: x86: fix MSR_IA32_TSC read for nested migration
  selftests: kvm: Fix assert failure in single-step test
  KVM: x86: VMX: Make smaller physical guest address space support user-configurable

Author: torvalds

arch/x86/kvm/svm/svm.c

@@ -2183,6 +2183,12 @@ static int iret_interception(struct vcpu_svm *svm)
 	return 1;
 }
 
+static int invd_interception(struct vcpu_svm *svm)
+{
+	/* Treat an INVD instruction as a NOP and just skip it. */
+	return kvm_skip_emulated_instruction(&svm->vcpu);
+}
+
 static int invlpg_interception(struct vcpu_svm *svm)
 {
 	if (!static_cpu_has(X86_FEATURE_DECODEASSISTS))
@@ -2774,7 +2780,7 @@ static int (*const svm_exit_handlers[])(struct vcpu_svm *svm) = {
 	[SVM_EXIT_RDPMC]			= rdpmc_interception,
 	[SVM_EXIT_CPUID]			= cpuid_interception,
 	[SVM_EXIT_IRET]                         = iret_interception,
-	[SVM_EXIT_INVD]                         = emulate_on_interception,
+	[SVM_EXIT_INVD]                         = invd_interception,
 	[SVM_EXIT_PAUSE]			= pause_interception,
 	[SVM_EXIT_HLT]				= halt_interception,
 	[SVM_EXIT_INVLPG]			= invlpg_interception,

arch/x86/kvm/vmx/vmx.c

@@ -129,6 +129,9 @@ static bool __read_mostly enable_preemption_timer = 1;
 module_param_named(preemption_timer, enable_preemption_timer, bool, S_IRUGO);
 #endif
 
+extern bool __read_mostly allow_smaller_maxphyaddr;
+module_param(allow_smaller_maxphyaddr, bool, S_IRUGO);
+
 #define KVM_VM_CR0_ALWAYS_OFF (X86_CR0_NW | X86_CR0_CD)
 #define KVM_VM_CR0_ALWAYS_ON_UNRESTRICTED_GUEST X86_CR0_NE
 #define KVM_VM_CR0_ALWAYS_ON				\
@@ -4803,6 +4806,7 @@ static int handle_exception_nmi(struct kvm_vcpu *vcpu)
 			 * EPT will cause page fault only if we need to
 			 * detect illegal GPAs.
 			 */
+			WARN_ON_ONCE(!allow_smaller_maxphyaddr);
 			kvm_fixup_and_inject_pf_error(vcpu, cr2, error_code);
 			return 1;
 		} else
@@ -5331,7 +5335,7 @@ static int handle_ept_violation(struct kvm_vcpu *vcpu)
 	 * would also use advanced VM-exit information for EPT violations to
 	 * reconstruct the page fault error code.
 	 */
-	if (unlikely(kvm_mmu_is_illegal_gpa(vcpu, gpa)))
+	if (unlikely(allow_smaller_maxphyaddr && kvm_mmu_is_illegal_gpa(vcpu, gpa)))
 		return kvm_emulate_instruction(vcpu, 0);
 
 	return kvm_mmu_page_fault(vcpu, gpa, error_code, NULL, 0);
@@ -8305,11 +8309,12 @@ static int __init vmx_init(void)
 	vmx_check_vmcs12_offsets();
 
 	/*
-	 * Intel processors don't have problems with
-	 * GUEST_MAXPHYADDR < HOST_MAXPHYADDR so enable
-	 * it for VMX by default
+	 * Shadow paging doesn't have a (further) performance penalty
+	 * from GUEST_MAXPHYADDR < HOST_MAXPHYADDR so enable it
+	 * by default
 	 */
-	allow_smaller_maxphyaddr = true;
+	if (!enable_ept)
+		allow_smaller_maxphyaddr = true;
 
 	return 0;
 }

arch/x86/kvm/vmx/vmx.h

@@ -552,7 +552,10 @@ static inline bool vmx_has_waitpkg(struct vcpu_vmx *vmx)
 
 static inline bool vmx_need_pf_intercept(struct kvm_vcpu *vcpu)
 {
-	return !enable_ept || cpuid_maxphyaddr(vcpu) < boot_cpu_data.x86_phys_bits;
+	if (!enable_ept)
+		return true;
+
+	return allow_smaller_maxphyaddr && cpuid_maxphyaddr(vcpu) < boot_cpu_data.x86_phys_bits;
 }
 
 void dump_vmcs(void);

arch/x86/kvm/x86.c

@@ -188,7 +188,7 @@ static struct kvm_shared_msrs __percpu *shared_msrs;
 u64 __read_mostly host_efer;
 EXPORT_SYMBOL_GPL(host_efer);
 
-bool __read_mostly allow_smaller_maxphyaddr;
+bool __read_mostly allow_smaller_maxphyaddr = 0;
 EXPORT_SYMBOL_GPL(allow_smaller_maxphyaddr);
 
 static u64 __read_mostly host_xss;
@@ -976,6 +976,7 @@ int kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
 	unsigned long old_cr4 = kvm_read_cr4(vcpu);
 	unsigned long pdptr_bits = X86_CR4_PGE | X86_CR4_PSE | X86_CR4_PAE |
 				   X86_CR4_SMEP;
+	unsigned long mmu_role_bits = pdptr_bits | X86_CR4_SMAP | X86_CR4_PKE;
 
 	if (kvm_valid_cr4(vcpu, cr4))
 		return 1;
@@ -1003,7 +1004,7 @@ int kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
 	if (kvm_x86_ops.set_cr4(vcpu, cr4))
 		return 1;
 
-	if (((cr4 ^ old_cr4) & pdptr_bits) ||
+	if (((cr4 ^ old_cr4) & mmu_role_bits) ||
 	    (!(cr4 & X86_CR4_PCIDE) && (old_cr4 & X86_CR4_PCIDE)))
 		kvm_mmu_reset_context(vcpu);
 
@@ -3221,9 +3222,22 @@ int kvm_get_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 	case MSR_IA32_POWER_CTL:
 		msr_info->data = vcpu->arch.msr_ia32_power_ctl;
 		break;
-	case MSR_IA32_TSC:
-		msr_info->data = kvm_scale_tsc(vcpu, rdtsc()) + vcpu->arch.tsc_offset;
+	case MSR_IA32_TSC: {
+		/*
+		 * Intel SDM states that MSR_IA32_TSC read adds the TSC offset
+		 * even when not intercepted. AMD manual doesn't explicitly
+		 * state this but appears to behave the same.
+		 *
+		 * On userspace reads and writes, however, we unconditionally
+		 * operate L1's TSC value to ensure backwards-compatible
+		 * behavior for migration.
+		 */
+		u64 tsc_offset = msr_info->host_initiated ? vcpu->arch.l1_tsc_offset :
+							    vcpu->arch.tsc_offset;
+
+		msr_info->data = kvm_scale_tsc(vcpu, rdtsc()) + tsc_offset;
 		break;
+	}
 	case MSR_MTRRcap:
 	case 0x200 ... 0x2ff:
 		return kvm_mtrr_get_msr(vcpu, msr_info->index, &msr_info->data);

tools/testing/selftests/kvm/x86_64/debug_regs.c

@@ -73,7 +73,7 @@ int main(void)
 	int i;
 	/* Instruction lengths starting at ss_start */
 	int ss_size[4] = {
-		3,		/* xor */
+		2,		/* xor */
 		2,		/* cpuid */
 		5,		/* mov */
 		2,		/* rdmsr */