Merge tag 'kvm-x86-misc-6.6' of https://github.com/kvm-x86/linux into HEAD

KVM x86 changes for 6.6:

 - Misc cleanups

 - Retry APIC optimized recalculation if a vCPU is added/enabled

 - Overhaul emergency reboot code to bring SVM up to par with VMX, tie the
   "emergency disabling" behavior to KVM actually being loaded, and move all of
   the logic within KVM

 - Fix user triggerable WARNs in SVM where KVM incorrectly assumes the TSC
   ratio MSR can diverge from the default iff TSC scaling is enabled, and clean
   up related code

 - Add a framework to allow "caching" feature flags so that KVM can check if
   the guest can use a feature without needing to search guest CPUID

1 files changed, 22 insertions, 24 deletions

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 94fa36ee073c..0b38a046690e 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -237,6 +237,9 @@ EXPORT_SYMBOL_GPL(enable_apicv);
 u64 __read_mostly host_xss;
 EXPORT_SYMBOL_GPL(host_xss);
 
+u64 __read_mostly host_arch_capabilities;
+EXPORT_SYMBOL_GPL(host_arch_capabilities);
+
 const struct _kvm_stats_desc kvm_vm_stats_desc[] = {
 	KVM_GENERIC_VM_STATS(),
 	STATS_DESC_COUNTER(VM, mmu_shadow_zapped),
@@ -1021,7 +1024,7 @@ void kvm_load_guest_xsave_state(struct kvm_vcpu *vcpu)
 		if (vcpu->arch.xcr0 != host_xcr0)
 			xsetbv(XCR_XFEATURE_ENABLED_MASK, vcpu->arch.xcr0);
 
-		if (vcpu->arch.xsaves_enabled &&
+		if (guest_can_use(vcpu, X86_FEATURE_XSAVES) &&
 		    vcpu->arch.ia32_xss != host_xss)
 			wrmsrl(MSR_IA32_XSS, vcpu->arch.ia32_xss);
 	}
@@ -1052,7 +1055,7 @@ void kvm_load_host_xsave_state(struct kvm_vcpu *vcpu)
 		if (vcpu->arch.xcr0 != host_xcr0)
 			xsetbv(XCR_XFEATURE_ENABLED_MASK, host_xcr0);
 
-		if (vcpu->arch.xsaves_enabled &&
+		if (guest_can_use(vcpu, X86_FEATURE_XSAVES) &&
 		    vcpu->arch.ia32_xss != host_xss)
 			wrmsrl(MSR_IA32_XSS, host_xss);
 	}
@@ -1620,12 +1623,7 @@ static bool kvm_is_immutable_feature_msr(u32 msr)
 
 static u64 kvm_get_arch_capabilities(void)
 {
-	u64 data = 0;
-
-	if (boot_cpu_has(X86_FEATURE_ARCH_CAPABILITIES)) {
-		rdmsrl(MSR_IA32_ARCH_CAPABILITIES, data);
-		data &= KVM_SUPPORTED_ARCH_CAP;
-	}
+	u64 data = host_arch_capabilities & KVM_SUPPORTED_ARCH_CAP;
 
 	/*
 	 * If nx_huge_pages is enabled, KVM's shadow paging will ensure that
@@ -2631,7 +2629,7 @@ static void kvm_vcpu_write_tsc_offset(struct kvm_vcpu *vcpu, u64 l1_offset)
 	else
 		vcpu->arch.tsc_offset = l1_offset;
 
-	static_call(kvm_x86_write_tsc_offset)(vcpu, vcpu->arch.tsc_offset);
+	static_call(kvm_x86_write_tsc_offset)(vcpu);
 }
 
 static void kvm_vcpu_write_tsc_multiplier(struct kvm_vcpu *vcpu, u64 l1_multiplier)
@@ -2647,8 +2645,7 @@ static void kvm_vcpu_write_tsc_multiplier(struct kvm_vcpu *vcpu, u64 l1_multipli
 		vcpu->arch.tsc_scaling_ratio = l1_multiplier;
 
 	if (kvm_caps.has_tsc_control)
-		static_call(kvm_x86_write_tsc_multiplier)(
-			vcpu, vcpu->arch.tsc_scaling_ratio);
+		static_call(kvm_x86_write_tsc_multiplier)(vcpu);
 }
 
 static inline bool kvm_check_tsc_unstable(void)
@@ -4665,7 +4662,6 @@ static int kvm_x86_dev_get_attr(struct kvm_device_attr *attr)
 		return 0;
 	default:
 		return -ENXIO;
-		break;
 	}
 }
 
@@ -6532,7 +6528,7 @@ static void kvm_free_msr_filter(struct kvm_x86_msr_filter *msr_filter)
 static int kvm_add_msr_filter(struct kvm_x86_msr_filter *msr_filter,
 			      struct kvm_msr_filter_range *user_range)
 {
-	unsigned long *bitmap = NULL;
+	unsigned long *bitmap;
 	size_t bitmap_size;
 
 	if (!user_range->nmsrs)
@@ -8245,11 +8241,6 @@ static bool emulator_get_cpuid(struct x86_emulate_ctxt *ctxt,
 	return kvm_cpuid(emul_to_vcpu(ctxt), eax, ebx, ecx, edx, exact_only);
 }
 
-static bool emulator_guest_has_long_mode(struct x86_emulate_ctxt *ctxt)
-{
-	return guest_cpuid_has(emul_to_vcpu(ctxt), X86_FEATURE_LM);
-}
-
 static bool emulator_guest_has_movbe(struct x86_emulate_ctxt *ctxt)
 {
 	return guest_cpuid_has(emul_to_vcpu(ctxt), X86_FEATURE_MOVBE);
@@ -8351,7 +8342,6 @@ static const struct x86_emulate_ops emulate_ops = {
 	.fix_hypercall       = emulator_fix_hypercall,
 	.intercept           = emulator_intercept,
 	.get_cpuid           = emulator_get_cpuid,
-	.guest_has_long_mode = emulator_guest_has_long_mode,
 	.guest_has_movbe     = emulator_guest_has_movbe,
 	.guest_has_fxsr      = emulator_guest_has_fxsr,
 	.guest_has_rdpid     = emulator_guest_has_rdpid,
@@ -9172,7 +9162,7 @@ static int kvmclock_cpu_down_prep(unsigned int cpu)
 static void tsc_khz_changed(void *data)
 {
 	struct cpufreq_freqs *freq = data;
-	unsigned long khz = 0;
+	unsigned long khz;
 
 	WARN_ON_ONCE(boot_cpu_has(X86_FEATURE_CONSTANT_TSC));
 
@@ -9512,6 +9502,9 @@ static int __kvm_x86_vendor_init(struct kvm_x86_init_ops *ops)
 
 	kvm_init_pmu_capability(ops->pmu_ops);
 
+	if (boot_cpu_has(X86_FEATURE_ARCH_CAPABILITIES))
+		rdmsrl(MSR_IA32_ARCH_CAPABILITIES, host_arch_capabilities);
+
 	r = ops->hardware_setup();
 	if (r != 0)
 		goto out_mmu_exit;
@@ -11111,12 +11104,17 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu)
 			r = -EINTR;
 			goto out;
 		}
+
 		/*
-		 * It should be impossible for the hypervisor timer to be in
-		 * use before KVM has ever run the vCPU.
+		 * Don't bother switching APIC timer emulation from the
+		 * hypervisor timer to the software timer, the only way for the
+		 * APIC timer to be active is if userspace stuffed vCPU state,
+		 * i.e. put the vCPU into a nonsensical state.  Only an INIT
+		 * will transition the vCPU out of UNINITIALIZED (without more
+		 * state stuffing from userspace), which will reset the local
+		 * APIC and thus cancel the timer or drop the IRQ (if the timer
+		 * already expired).
 		 */
-		WARN_ON_ONCE(kvm_lapic_hv_timer_in_use(vcpu));
-
 		kvm_vcpu_srcu_read_unlock(vcpu);
 		kvm_vcpu_block(vcpu);
 		kvm_vcpu_srcu_read_lock(vcpu);