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

Pull kvm fixes from Paolo Bonzini:
 "ARM:

   - Plug a race in the stage-2 mapping code where the IPA and the PA
     would end up being out of sync

   - Make better use of the bitmap API (bitmap_zero, bitmap_zalloc...)

   - FP/SVE/SME documentation update, in the hope that this field
     becomes clearer...

   - Add workaround for Apple SEIS brokenness to a new SoC

   - Random comment fixes

  x86:

   - add MSR_IA32_TSX_CTRL into msrs_to_save

   - fixes for XCR0 handling in SGX enclaves

  Generic:

   - Fix vcpu_array[0] races

   - Fix race between starting a VM and 'reboot -f'"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: VMX: add MSR_IA32_TSX_CTRL into msrs_to_save
  KVM: x86: Don't adjust guest's CPUID.0x12.1 (allowed SGX enclave XFRM)
  KVM: VMX: Don't rely _only_ on CPUID to enforce XCR0 restrictions for ECREATE
  KVM: Fix vcpu_array[0] races
  KVM: VMX: Fix header file dependency of asm/vmx.h
  KVM: Don't enable hardware after a restart/shutdown is initiated
  KVM: Use syscore_ops instead of reboot_notifier to hook restart/shutdown
  KVM: arm64: vgic: Add Apple M2 PRO/MAX cpus to the list of broken SEIS implementations
  KVM: arm64: Clarify host SME state management
  KVM: arm64: Restructure check for SVE support in FP trap handler
  KVM: arm64: Document check for TIF_FOREIGN_FPSTATE
  KVM: arm64: Fix repeated words in comments
  KVM: arm64: Constify start/end/phys fields of the pgtable walker data
  KVM: arm64: Infer PA offset from VA in hyp map walker
  KVM: arm64: Infer the PA offset from IPA in stage-2 map walker
  KVM: arm64: Use the bitmap API to allocate bitmaps
  KVM: arm64: Slightly optimize flush_context()

13 files changed, 129 insertions, 66 deletions

diff --git a/arch/arm64/include/asm/cputype.h b/arch/arm64/include/asm/cputype.h
index 683ca3af408485..5f6f84837a4903 100644
--- a/arch/arm64/include/asm/cputype.h
+++ b/arch/arm64/include/asm/cputype.h
@@ -126,6 +126,10 @@
 #define APPLE_CPU_PART_M1_FIRESTORM_MAX	0x029
 #define APPLE_CPU_PART_M2_BLIZZARD	0x032
 #define APPLE_CPU_PART_M2_AVALANCHE	0x033
+#define APPLE_CPU_PART_M2_BLIZZARD_PRO	0x034
+#define APPLE_CPU_PART_M2_AVALANCHE_PRO	0x035
+#define APPLE_CPU_PART_M2_BLIZZARD_MAX	0x038
+#define APPLE_CPU_PART_M2_AVALANCHE_MAX	0x039
 
 #define AMPERE_CPU_PART_AMPERE1		0xAC3
 
@@ -181,6 +185,10 @@
 #define MIDR_APPLE_M1_FIRESTORM_MAX MIDR_CPU_MODEL(ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M1_FIRESTORM_MAX)
 #define MIDR_APPLE_M2_BLIZZARD MIDR_CPU_MODEL(ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_BLIZZARD)
 #define MIDR_APPLE_M2_AVALANCHE MIDR_CPU_MODEL(ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_AVALANCHE)
+#define MIDR_APPLE_M2_BLIZZARD_PRO MIDR_CPU_MODEL(ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_BLIZZARD_PRO)
+#define MIDR_APPLE_M2_AVALANCHE_PRO MIDR_CPU_MODEL(ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_AVALANCHE_PRO)
+#define MIDR_APPLE_M2_BLIZZARD_MAX MIDR_CPU_MODEL(ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_BLIZZARD_MAX)
+#define MIDR_APPLE_M2_AVALANCHE_MAX MIDR_CPU_MODEL(ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_AVALANCHE_MAX)
 #define MIDR_AMPERE1 MIDR_CPU_MODEL(ARM_CPU_IMP_AMPERE, AMPERE_CPU_PART_AMPERE1)
 
 /* Fujitsu Erratum 010001 affects A64FX 1.0 and 1.1, (v0r0 and v1r0) */
diff --git a/arch/arm64/include/asm/kvm_pgtable.h b/arch/arm64/include/asm/kvm_pgtable.h
index 4cd6762bda805d..dc3c072e862f1e 100644
--- a/arch/arm64/include/asm/kvm_pgtable.h
+++ b/arch/arm64/include/asm/kvm_pgtable.h
@@ -209,6 +209,7 @@ struct kvm_pgtable_visit_ctx {
 	kvm_pte_t				old;
 	void					*arg;
 	struct kvm_pgtable_mm_ops		*mm_ops;
+	u64					start;
 	u64					addr;
 	u64					end;
 	u32					level;
diff --git a/arch/arm64/kvm/fpsimd.c b/arch/arm64/kvm/fpsimd.c
index 1279949599b5fc..4c9dcd8fc93950 100644
--- a/arch/arm64/kvm/fpsimd.c
+++ b/arch/arm64/kvm/fpsimd.c
@@ -81,26 +81,34 @@ void kvm_arch_vcpu_load_fp(struct kvm_vcpu *vcpu)
 
 	fpsimd_kvm_prepare();
 
+	/*
+	 * We will check TIF_FOREIGN_FPSTATE just before entering the
+	 * guest in kvm_arch_vcpu_ctxflush_fp() and override this to
+	 * FP_STATE_FREE if the flag set.
+	 */
 	vcpu->arch.fp_state = FP_STATE_HOST_OWNED;
 
 	vcpu_clear_flag(vcpu, HOST_SVE_ENABLED);
 	if (read_sysreg(cpacr_el1) & CPACR_EL1_ZEN_EL0EN)
 		vcpu_set_flag(vcpu, HOST_SVE_ENABLED);
 
-	/*
-	 * We don't currently support SME guests but if we leave
-	 * things in streaming mode then when the guest starts running
-	 * FPSIMD or SVE code it may generate SME traps so as a
-	 * special case if we are in streaming mode we force the host
-	 * state to be saved now and exit streaming mode so that we
-	 * don't have to handle any SME traps for valid guest
-	 * operations. Do this for ZA as well for now for simplicity.
-	 */
 	if (system_supports_sme()) {
 		vcpu_clear_flag(vcpu, HOST_SME_ENABLED);
 		if (read_sysreg(cpacr_el1) & CPACR_EL1_SMEN_EL0EN)
 			vcpu_set_flag(vcpu, HOST_SME_ENABLED);
 
+		/*
+		 * If PSTATE.SM is enabled then save any pending FP
+		 * state and disable PSTATE.SM. If we leave PSTATE.SM
+		 * enabled and the guest does not enable SME via
+		 * CPACR_EL1.SMEN then operations that should be valid
+		 * may generate SME traps from EL1 to EL1 which we
+		 * can't intercept and which would confuse the guest.
+		 *
+		 * Do the same for PSTATE.ZA in the case where there
+		 * is state in the registers which has not already
+		 * been saved, this is very unlikely to happen.
+		 */
 		if (read_sysreg_s(SYS_SVCR) & (SVCR_SM_MASK | SVCR_ZA_MASK)) {
 			vcpu->arch.fp_state = FP_STATE_FREE;
 			fpsimd_save_and_flush_cpu_state();
diff --git a/arch/arm64/kvm/hyp/include/hyp/switch.h b/arch/arm64/kvm/hyp/include/hyp/switch.h
index c41166f1a1dd7d..e78a08a72a3c9f 100644
--- a/arch/arm64/kvm/hyp/include/hyp/switch.h
+++ b/arch/arm64/kvm/hyp/include/hyp/switch.h
@@ -177,9 +177,17 @@ static bool kvm_hyp_handle_fpsimd(struct kvm_vcpu *vcpu, u64 *exit_code)
 	sve_guest = vcpu_has_sve(vcpu);
 	esr_ec = kvm_vcpu_trap_get_class(vcpu);
 
-	/* Don't handle SVE traps for non-SVE vcpus here: */
-	if (!sve_guest && esr_ec != ESR_ELx_EC_FP_ASIMD)
+	/* Only handle traps the vCPU can support here: */
+	switch (esr_ec) {
+	case ESR_ELx_EC_FP_ASIMD:
+		break;
+	case ESR_ELx_EC_SVE:
+		if (!sve_guest)
+			return false;
+		break;
+	default:
 		return false;
+	}
 
 	/* Valid trap.  Switch the context: */
 
diff --git a/arch/arm64/kvm/hyp/pgtable.c b/arch/arm64/kvm/hyp/pgtable.c
index 3d61bd3e591d27..5282cb9ca4cff4 100644
--- a/arch/arm64/kvm/hyp/pgtable.c
+++ b/arch/arm64/kvm/hyp/pgtable.c
@@ -58,8 +58,9 @@
 struct kvm_pgtable_walk_data {
 	struct kvm_pgtable_walker	*walker;
 
+	const u64			start;
 	u64				addr;
-	u64				end;
+	const u64			end;
 };
 
 static bool kvm_phys_is_valid(u64 phys)
@@ -201,6 +202,7 @@ static inline int __kvm_pgtable_visit(struct kvm_pgtable_walk_data *data,
 		.old	= READ_ONCE(*ptep),
 		.arg	= data->walker->arg,
 		.mm_ops	= mm_ops,
+		.start	= data->start,
 		.addr	= data->addr,
 		.end	= data->end,
 		.level	= level,
@@ -293,6 +295,7 @@ int kvm_pgtable_walk(struct kvm_pgtable *pgt, u64 addr, u64 size,
 		     struct kvm_pgtable_walker *walker)
 {
 	struct kvm_pgtable_walk_data walk_data = {
+		.start	= ALIGN_DOWN(addr, PAGE_SIZE),
 		.addr	= ALIGN_DOWN(addr, PAGE_SIZE),
 		.end	= PAGE_ALIGN(walk_data.addr + size),
 		.walker	= walker,
@@ -349,7 +352,7 @@ int kvm_pgtable_get_leaf(struct kvm_pgtable *pgt, u64 addr,
 }
 
 struct hyp_map_data {
-	u64				phys;
+	const u64			phys;
 	kvm_pte_t			attr;
 };
 
@@ -407,13 +410,12 @@ enum kvm_pgtable_prot kvm_pgtable_hyp_pte_prot(kvm_pte_t pte)
 static bool hyp_map_walker_try_leaf(const struct kvm_pgtable_visit_ctx *ctx,
 				    struct hyp_map_data *data)
 {
+	u64 phys = data->phys + (ctx->addr - ctx->start);
 	kvm_pte_t new;
-	u64 granule = kvm_granule_size(ctx->level), phys = data->phys;
 
 	if (!kvm_block_mapping_supported(ctx, phys))
 		return false;
 
-	data->phys += granule;
 	new = kvm_init_valid_leaf_pte(phys, data->attr, ctx->level);
 	if (ctx->old == new)
 		return true;
@@ -576,7 +578,7 @@ void kvm_pgtable_hyp_destroy(struct kvm_pgtable *pgt)
 }
 
 struct stage2_map_data {
-	u64				phys;
+	const u64			phys;
 	kvm_pte_t			attr;
 	u8				owner_id;
 
@@ -794,20 +796,43 @@ static bool stage2_pte_executable(kvm_pte_t pte)
 	return !(pte & KVM_PTE_LEAF_ATTR_HI_S2_XN);
 }
 
+static u64 stage2_map_walker_phys_addr(const struct kvm_pgtable_visit_ctx *ctx,
+				       const struct stage2_map_data *data)
+{
+	u64 phys = data->phys;
+
+	/*
+	 * Stage-2 walks to update ownership data are communicated to the map
+	 * walker using an invalid PA. Avoid offsetting an already invalid PA,
+	 * which could overflow and make the address valid again.
+	 */
+	if (!kvm_phys_is_valid(phys))
+		return phys;
+
+	/*
+	 * Otherwise, work out the correct PA based on how far the walk has
+	 * gotten.
+	 */
+	return phys + (ctx->addr - ctx->start);
+}
+
 static bool stage2_leaf_mapping_allowed(const struct kvm_pgtable_visit_ctx *ctx,
 					struct stage2_map_data *data)
 {
+	u64 phys = stage2_map_walker_phys_addr(ctx, data);
+
 	if (data->force_pte && (ctx->level < (KVM_PGTABLE_MAX_LEVELS - 1)))
 		return false;
 
-	return kvm_block_mapping_supported(ctx, data->phys);
+	return kvm_block_mapping_supported(ctx, phys);
 }
 
 static int stage2_map_walker_try_leaf(const struct kvm_pgtable_visit_ctx *ctx,
 				      struct stage2_map_data *data)
 {
 	kvm_pte_t new;
-	u64 granule = kvm_granule_size(ctx->level), phys = data->phys;
+	u64 phys = stage2_map_walker_phys_addr(ctx, data);
+	u64 granule = kvm_granule_size(ctx->level);
 	struct kvm_pgtable *pgt = data->mmu->pgt;
 	struct kvm_pgtable_mm_ops *mm_ops = ctx->mm_ops;
 
@@ -841,8 +866,6 @@ static int stage2_map_walker_try_leaf(const struct kvm_pgtable_visit_ctx *ctx,
 
 	stage2_make_pte(ctx, new);
 
-	if (kvm_phys_is_valid(phys))
-		data->phys += granule;
 	return 0;
 }
 
diff --git a/arch/arm64/kvm/inject_fault.c b/arch/arm64/kvm/inject_fault.c
index 64c3aec0d937c3..0bd93a5f21ce38 100644
--- a/arch/arm64/kvm/inject_fault.c
+++ b/arch/arm64/kvm/inject_fault.c
@@ -204,7 +204,7 @@ void kvm_inject_size_fault(struct kvm_vcpu *vcpu)
 	 * Size Fault at level 0, as if exceeding PARange.
 	 *
 	 * Non-LPAE guests will only get the external abort, as there
-	 * is no way to to describe the ASF.
+	 * is no way to describe the ASF.
 	 */
 	if (vcpu_el1_is_32bit(vcpu) &&
 	    !(vcpu_read_sys_reg(vcpu, TCR_EL1) & TTBCR_EAE))
diff --git a/arch/arm64/kvm/vgic/vgic-v3.c b/arch/arm64/kvm/vgic/vgic-v3.c
index 469d816f356f3f..93a47a515c1371 100644
--- a/arch/arm64/kvm/vgic/vgic-v3.c
+++ b/arch/arm64/kvm/vgic/vgic-v3.c
@@ -616,6 +616,10 @@ static const struct midr_range broken_seis[] = {
 	MIDR_ALL_VERSIONS(MIDR_APPLE_M1_FIRESTORM_MAX),
 	MIDR_ALL_VERSIONS(MIDR_APPLE_M2_BLIZZARD),
 	MIDR_ALL_VERSIONS(MIDR_APPLE_M2_AVALANCHE),
+	MIDR_ALL_VERSIONS(MIDR_APPLE_M2_BLIZZARD_PRO),
+	MIDR_ALL_VERSIONS(MIDR_APPLE_M2_AVALANCHE_PRO),
+	MIDR_ALL_VERSIONS(MIDR_APPLE_M2_BLIZZARD_MAX),
+	MIDR_ALL_VERSIONS(MIDR_APPLE_M2_AVALANCHE_MAX),
 	{},
 };
 
diff --git a/arch/arm64/kvm/vmid.c b/arch/arm64/kvm/vmid.c
index 08978d0672e7e9..7fe8ba1a2851c5 100644
--- a/arch/arm64/kvm/vmid.c
+++ b/arch/arm64/kvm/vmid.c
@@ -47,7 +47,7 @@ static void flush_context(void)
 	int cpu;
 	u64 vmid;
 
-	bitmap_clear(vmid_map, 0, NUM_USER_VMIDS);
+	bitmap_zero(vmid_map, NUM_USER_VMIDS);
 
 	for_each_possible_cpu(cpu) {
 		vmid = atomic64_xchg_relaxed(&per_cpu(active_vmids, cpu), 0);
@@ -182,8 +182,7 @@ int __init kvm_arm_vmid_alloc_init(void)
 	 */
 	WARN_ON(NUM_USER_VMIDS - 1 <= num_possible_cpus());
 	atomic64_set(&vmid_generation, VMID_FIRST_VERSION);
-	vmid_map = kcalloc(BITS_TO_LONGS(NUM_USER_VMIDS),
-			   sizeof(*vmid_map), GFP_KERNEL);
+	vmid_map = bitmap_zalloc(NUM_USER_VMIDS, GFP_KERNEL);
 	if (!vmid_map)
 		return -ENOMEM;
 
@@ -192,5 +191,5 @@ int __init kvm_arm_vmid_alloc_init(void)
 
 void __init kvm_arm_vmid_alloc_free(void)
 {
-	kfree(vmid_map);
+	bitmap_free(vmid_map);
 }
diff --git a/arch/x86/include/asm/vmx.h b/arch/x86/include/asm/vmx.h
index 498dc600bd5c8e..0d02c4aafa6f49 100644
--- a/arch/x86/include/asm/vmx.h
+++ b/arch/x86/include/asm/vmx.h
@@ -13,7 +13,9 @@
 
 
 #include <linux/bitops.h>
+#include <linux/bug.h>
 #include <linux/types.h>
+
 #include <uapi/asm/vmx.h>
 #include <asm/vmxfeatures.h>
 
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index 123bf8b97a4b21..0c9660a07b233f 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -253,7 +253,6 @@ static void __kvm_update_cpuid_runtime(struct kvm_vcpu *vcpu, struct kvm_cpuid_e
 				       int nent)
 {
 	struct kvm_cpuid_entry2 *best;
-	u64 guest_supported_xcr0 = cpuid_get_supported_xcr0(entries, nent);
 
 	best = cpuid_entry2_find(entries, nent, 1, KVM_CPUID_INDEX_NOT_SIGNIFICANT);
 	if (best) {
@@ -292,21 +291,6 @@ static void __kvm_update_cpuid_runtime(struct kvm_vcpu *vcpu, struct kvm_cpuid_e
 					   vcpu->arch.ia32_misc_enable_msr &
 					   MSR_IA32_MISC_ENABLE_MWAIT);
 	}
-
-	/*
-	 * Bits 127:0 of the allowed SECS.ATTRIBUTES (CPUID.0x12.0x1) enumerate
-	 * the supported XSAVE Feature Request Mask (XFRM), i.e. the enclave's
-	 * requested XCR0 value.  The enclave's XFRM must be a subset of XCRO
-	 * at the time of EENTER, thus adjust the allowed XFRM by the guest's
-	 * supported XCR0.  Similar to XCR0 handling, FP and SSE are forced to
-	 * '1' even on CPUs that don't support XSAVE.
-	 */
-	best = cpuid_entry2_find(entries, nent, 0x12, 0x1);
-	if (best) {
-		best->ecx &= guest_supported_xcr0 & 0xffffffff;
-		best->edx &= guest_supported_xcr0 >> 32;
-		best->ecx |= XFEATURE_MASK_FPSSE;
-	}
 }
 
 void kvm_update_cpuid_runtime(struct kvm_vcpu *vcpu)
diff --git a/arch/x86/kvm/vmx/sgx.c b/arch/x86/kvm/vmx/sgx.c
index 0574030b071fbf..2261b684a7d447 100644
--- a/arch/x86/kvm/vmx/sgx.c
+++ b/arch/x86/kvm/vmx/sgx.c
@@ -170,12 +170,19 @@ static int __handle_encls_ecreate(struct kvm_vcpu *vcpu,
 		return 1;
 	}
 
-	/* Enforce CPUID restrictions on MISCSELECT, ATTRIBUTES and XFRM. */
+	/*
+	 * Enforce CPUID restrictions on MISCSELECT, ATTRIBUTES and XFRM.  Note
+	 * that the allowed XFRM (XFeature Request Mask) isn't strictly bound
+	 * by the supported XCR0.  FP+SSE *must* be set in XFRM, even if XSAVE
+	 * is unsupported, i.e. even if XCR0 itself is completely unsupported.
+	 */
 	if ((u32)miscselect & ~sgx_12_0->ebx ||
 	    (u32)attributes & ~sgx_12_1->eax ||
 	    (u32)(attributes >> 32) & ~sgx_12_1->ebx ||
 	    (u32)xfrm & ~sgx_12_1->ecx ||
-	    (u32)(xfrm >> 32) & ~sgx_12_1->edx) {
+	    (u32)(xfrm >> 32) & ~sgx_12_1->edx ||
+	    xfrm & ~(vcpu->arch.guest_supported_xcr0 | XFEATURE_MASK_FPSSE) ||
+	    (xfrm & XFEATURE_MASK_FPSSE) != XFEATURE_MASK_FPSSE) {
 		kvm_inject_gp(vcpu, 0);
 		return 1;
 	}
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index ceb7c5e9cf9e92..c0778ca39650f4 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -1446,7 +1446,7 @@ static const u32 msrs_to_save_base[] = {
 #endif
 	MSR_IA32_TSC, MSR_IA32_CR_PAT, MSR_VM_HSAVE_PA,
 	MSR_IA32_FEAT_CTL, MSR_IA32_BNDCFGS, MSR_TSC_AUX,
-	MSR_IA32_SPEC_CTRL,
+	MSR_IA32_SPEC_CTRL, MSR_IA32_TSX_CTRL,
 	MSR_IA32_RTIT_CTL, MSR_IA32_RTIT_STATUS, MSR_IA32_RTIT_CR3_MATCH,
 	MSR_IA32_RTIT_OUTPUT_BASE, MSR_IA32_RTIT_OUTPUT_MASK,
 	MSR_IA32_RTIT_ADDR0_A, MSR_IA32_RTIT_ADDR0_B,
@@ -7155,6 +7155,10 @@ static void kvm_probe_msr_to_save(u32 msr_index)
 		if (!kvm_cpu_cap_has(X86_FEATURE_XFD))
 			return;
 		break;
+	case MSR_IA32_TSX_CTRL:
+		if (!(kvm_get_arch_capabilities() & ARCH_CAP_TSX_CTRL_MSR))
+			return;
+		break;
 	default:
 		break;
 	}
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index cb5c13eee1936e..479802a892d4f7 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -3962,18 +3962,19 @@ static int kvm_vm_ioctl_create_vcpu(struct kvm *kvm, u32 id)
 	}
 
 	vcpu->vcpu_idx = atomic_read(&kvm->online_vcpus);
-	r = xa_insert(&kvm->vcpu_array, vcpu->vcpu_idx, vcpu, GFP_KERNEL_ACCOUNT);
-	BUG_ON(r == -EBUSY);
+	r = xa_reserve(&kvm->vcpu_array, vcpu->vcpu_idx, GFP_KERNEL_ACCOUNT);
 	if (r)
 		goto unlock_vcpu_destroy;
 
 	/* Now it's all set up, let userspace reach it */
 	kvm_get_kvm(kvm);
 	r = create_vcpu_fd(vcpu);
-	if (r < 0) {
-		xa_erase(&kvm->vcpu_array, vcpu->vcpu_idx);
-		kvm_put_kvm_no_destroy(kvm);
-		goto unlock_vcpu_destroy;
+	if (r < 0)
+		goto kvm_put_xa_release;
+
+	if (KVM_BUG_ON(!!xa_store(&kvm->vcpu_array, vcpu->vcpu_idx, vcpu, 0), kvm)) {
+		r = -EINVAL;
+		goto kvm_put_xa_release;
 	}
 
 	/*
@@ -3988,6 +3989,9 @@ static int kvm_vm_ioctl_create_vcpu(struct kvm *kvm, u32 id)
 	kvm_create_vcpu_debugfs(vcpu);
 	return r;
 
+kvm_put_xa_release:
+	kvm_put_kvm_no_destroy(kvm);
+	xa_release(&kvm->vcpu_array, vcpu->vcpu_idx);
 unlock_vcpu_destroy:
 	mutex_unlock(&kvm->lock);
 	kvm_dirty_ring_free(&vcpu->dirty_ring);
@@ -5184,7 +5188,20 @@ static void hardware_disable_all(void)
 static int hardware_enable_all(void)
 {
 	atomic_t failed = ATOMIC_INIT(0);
-	int r = 0;
+	int r;
+
+	/*
+	 * Do not enable hardware virtualization if the system is going down.
+	 * If userspace initiated a forced reboot, e.g. reboot -f, then it's
+	 * possible for an in-flight KVM_CREATE_VM to trigger hardware enabling
+	 * after kvm_reboot() is called.  Note, this relies on system_state
+	 * being set _before_ kvm_reboot(), which is why KVM uses a syscore ops
+	 * hook instead of registering a dedicated reboot notifier (the latter
+	 * runs before system_state is updated).
+	 */
+	if (system_state == SYSTEM_HALT || system_state == SYSTEM_POWER_OFF ||
+	    system_state == SYSTEM_RESTART)
+		return -EBUSY;
 
 	/*
 	 * When onlining a CPU, cpu_online_mask is set before kvm_online_cpu()
@@ -5197,6 +5214,8 @@ static int hardware_enable_all(void)
 	cpus_read_lock();
 	mutex_lock(&kvm_lock);
 
+	r = 0;
+
 	kvm_usage_count++;
 	if (kvm_usage_count == 1) {
 		on_each_cpu(hardware_enable_nolock, &failed, 1);
@@ -5213,26 +5232,24 @@ static int hardware_enable_all(void)
 	return r;
 }
 
-static int kvm_reboot(struct notifier_block *notifier, unsigned long val,
-		      void *v)
+static void kvm_shutdown(void)
 {
 	/*
-	 * Some (well, at least mine) BIOSes hang on reboot if
-	 * in vmx root mode.
-	 *
-	 * And Intel TXT required VMX off for all cpu when system shutdown.
+	 * Disable hardware virtualization and set kvm_rebooting to indicate
+	 * that KVM has asynchronously disabled hardware virtualization, i.e.
+	 * that relevant errors and exceptions aren't entirely unexpected.
+	 * Some flavors of hardware virtualization need to be disabled before
+	 * transferring control to firmware (to perform shutdown/reboot), e.g.
+	 * on x86, virtualization can block INIT interrupts, which are used by
+	 * firmware to pull APs back under firmware control.  Note, this path
+	 * is used for both shutdown and reboot scenarios, i.e. neither name is
+	 * 100% comprehensive.
 	 */
 	pr_info("kvm: exiting hardware virtualization\n");
 	kvm_rebooting = true;
 	on_each_cpu(hardware_disable_nolock, NULL, 1);
-	return NOTIFY_OK;
 }
 
-static struct notifier_block kvm_reboot_notifier = {
-	.notifier_call = kvm_reboot,
-	.priority = 0,
-};
-
 static int kvm_suspend(void)
 {
 	/*
@@ -5263,6 +5280,7 @@ static void kvm_resume(void)
 static struct syscore_ops kvm_syscore_ops = {
 	.suspend = kvm_suspend,
 	.resume = kvm_resume,
+	.shutdown = kvm_shutdown,
 };
 #else /* CONFIG_KVM_GENERIC_HARDWARE_ENABLING */
 static int hardware_enable_all(void)
@@ -5967,7 +5985,6 @@ int kvm_init(unsigned vcpu_size, unsigned vcpu_align, struct module *module)
 	if (r)
 		return r;
 
-	register_reboot_notifier(&kvm_reboot_notifier);
 	register_syscore_ops(&kvm_syscore_ops);
 #endif
 
@@ -6039,7 +6056,6 @@ err_cpu_kick_mask:
 err_vcpu_cache:
 #ifdef CONFIG_KVM_GENERIC_HARDWARE_ENABLING
 	unregister_syscore_ops(&kvm_syscore_ops);
-	unregister_reboot_notifier(&kvm_reboot_notifier);
 	cpuhp_remove_state_nocalls(CPUHP_AP_KVM_ONLINE);
 #endif
 	return r;
@@ -6065,7 +6081,6 @@ void kvm_exit(void)
 	kvm_async_pf_deinit();
 #ifdef CONFIG_KVM_GENERIC_HARDWARE_ENABLING
 	unregister_syscore_ops(&kvm_syscore_ops);
-	unregister_reboot_notifier(&kvm_reboot_notifier);
 	cpuhp_remove_state_nocalls(CPUHP_AP_KVM_ONLINE);
 #endif
 	kvm_irqfd_exit();