Merge branch 'kvm-nvmx-and-vm-teardown' into HEAD

The immediate issue being fixed here is a nVMX bug where KVM fails to
detect that, after nested VM-Exit, L1 has a pending IRQ (or NMI).
However, checking for a pending interrupt accesses the legacy PIC, and
x86's kvm_arch_destroy_vm() currently frees the PIC before destroying
vCPUs, i.e. checking for IRQs during the forced nested VM-Exit results
in a NULL pointer deref; that's a prerequisite for the nVMX fix.

The remaining patches attempt to bring a bit of sanity to x86's VM
teardown code, which has accumulated a lot of cruft over the years. E.g.
KVM currently unloads each vCPU's MMUs in a separate operation from
destroying vCPUs, all because when guest SMP support was added, KVM had a
kludgy MMU teardown flow that broke when a VM had more than one 1 vCPU.
And that oddity lived on, for 18 years...

Signed-off-by: Paolo Bonzini <[e-mail]>

arch/arm64/include/asm/kvm_host.h

···
 	return cpus_have_final_cap(ARM64_SPECTRE_V3A);
 }
 
-static inline void kvm_arch_sync_events(struct kvm *kvm) {}
-
 void kvm_init_host_debug_data(void);
 void kvm_vcpu_load_debug(struct kvm_vcpu *vcpu);
 void kvm_vcpu_put_debug(struct kvm_vcpu *vcpu);

arch/loongarch/include/asm/kvm_host.h

···
 
 /* Misc */
 static inline void kvm_arch_hardware_unsetup(void) {}
-static inline void kvm_arch_sync_events(struct kvm *kvm) {}
 static inline void kvm_arch_memslots_updated(struct kvm *kvm, u64 gen) {}
 static inline void kvm_arch_vcpu_blocking(struct kvm_vcpu *vcpu) {}
 static inline void kvm_arch_vcpu_unblocking(struct kvm_vcpu *vcpu) {}

arch/mips/include/asm/kvm_host.h

···
 extern int kvm_vcpu_ioctl_interrupt(struct kvm_vcpu *vcpu,
 			     struct kvm_mips_interrupt *irq);
 
-static inline void kvm_arch_sync_events(struct kvm *kvm) {}
 static inline void kvm_arch_free_memslot(struct kvm *kvm,
 					 struct kvm_memory_slot *slot) {}
 static inline void kvm_arch_memslots_updated(struct kvm *kvm, u64 gen) {}

arch/powerpc/include/asm/kvm_host.h

···
 #define __KVM_HAVE_ARCH_WQP
 #define __KVM_HAVE_CREATE_DEVICE
 
-static inline void kvm_arch_sync_events(struct kvm *kvm) {}
 static inline void kvm_arch_memslots_updated(struct kvm *kvm, u64 gen) {}
 static inline void kvm_arch_flush_shadow_all(struct kvm *kvm) {}
 static inline void kvm_arch_vcpu_blocking(struct kvm_vcpu *vcpu) {}

arch/riscv/include/asm/kvm_host.h

···
 	return IS_ENABLED(CONFIG_GUEST_PERF_EVENTS) && !!vcpu;
 }
 
-static inline void kvm_arch_sync_events(struct kvm *kvm) {}
-
 #define KVM_RISCV_GSTAGE_TLB_MIN_ORDER		12
 
 void kvm_riscv_local_hfence_gvma_vmid_gpa(unsigned long vmid,

arch/s390/include/asm/kvm_host.h

···
 extern int kvm_s390_gisc_register(struct kvm *kvm, u32 gisc);
 extern int kvm_s390_gisc_unregister(struct kvm *kvm, u32 gisc);
 
-static inline void kvm_arch_sync_events(struct kvm *kvm) {}
 static inline void kvm_arch_free_memslot(struct kvm *kvm,
 					 struct kvm_memory_slot *slot) {}
 static inline void kvm_arch_memslots_updated(struct kvm *kvm, u64 gen) {}

arch/x86/kvm/x86.c

···
 {
 	int idx;
 
+	kvm_clear_async_pf_completion_queue(vcpu);
+	kvm_mmu_unload(vcpu);
+
 	kvmclock_reset(vcpu);
 
 	kvm_x86_call(vcpu_free)(vcpu);
···
 	return ret;
 }
 
-static void kvm_unload_vcpu_mmu(struct kvm_vcpu *vcpu)
-{
-	vcpu_load(vcpu);
-	kvm_mmu_unload(vcpu);
-	vcpu_put(vcpu);
-}
-
-static void kvm_unload_vcpu_mmus(struct kvm *kvm)
-{
-	unsigned long i;
-	struct kvm_vcpu *vcpu;
-
-	kvm_for_each_vcpu(i, vcpu, kvm) {
-		kvm_clear_async_pf_completion_queue(vcpu);
-		kvm_unload_vcpu_mmu(vcpu);
-	}
-}
-
-void kvm_arch_sync_events(struct kvm *kvm)
-{
-	cancel_delayed_work_sync(&kvm->arch.kvmclock_sync_work);
-	cancel_delayed_work_sync(&kvm->arch.kvmclock_update_work);
-	kvm_free_pit(kvm);
-}
-
 /**
  * __x86_set_memory_region: Setup KVM internal memory slot
  *
···
 
 void kvm_arch_pre_destroy_vm(struct kvm *kvm)
 {
+	/*
+	 * Stop all background workers and kthreads before destroying vCPUs, as
+	 * iterating over vCPUs in a different task while vCPUs are being freed
+	 * is unsafe, i.e. will lead to use-after-free.  The PIT also needs to
+	 * be stopped before IRQ routing is freed.
+	 */
+	cancel_delayed_work_sync(&kvm->arch.kvmclock_sync_work);
+	cancel_delayed_work_sync(&kvm->arch.kvmclock_update_work);
+
+	kvm_free_pit(kvm);
+
 	kvm_mmu_pre_destroy_vm(kvm);
 }
 
···
 		__x86_set_memory_region(kvm, TSS_PRIVATE_MEMSLOT, 0, 0);
 		mutex_unlock(&kvm->slots_lock);
 	}
-	kvm_unload_vcpu_mmus(kvm);
 	kvm_destroy_vcpus(kvm);
 	kvm_x86_call(vm_destroy)(kvm);
 	kvm_free_msr_filter(srcu_dereference_check(kvm->arch.msr_filter, &kvm->srcu, 1));

include/linux/kvm_host.h

···
 
 int kvm_arch_init_vm(struct kvm *kvm, unsigned long type);
 void kvm_arch_destroy_vm(struct kvm *kvm);
-void kvm_arch_sync_events(struct kvm *kvm);
 
 int kvm_cpu_has_pending_timer(struct kvm_vcpu *vcpu);
 

virt/kvm/kvm_main.c

···
 	kvm_for_each_vcpu(i, vcpu, kvm) {
 		kvm_vcpu_destroy(vcpu);
 		xa_erase(&kvm->vcpu_array, i);
+
+		/*
+		 * Assert that the vCPU isn't visible in any way, to ensure KVM
+		 * doesn't trigger a use-after-free if destroying vCPUs results
+		 * in VM-wide request, e.g. to flush remote TLBs when tearing
+		 * down MMUs, or to mark the VM dead if a KVM_BUG_ON() fires.
+		 */
+		WARN_ON_ONCE(xa_load(&kvm->vcpu_array, i) || kvm_get_vcpu(kvm, i));
 	}
 
 	atomic_set(&kvm->online_vcpus, 0);
···
 	kvm_destroy_pm_notifier(kvm);
 	kvm_uevent_notify_change(KVM_EVENT_DESTROY_VM, kvm);
 	kvm_destroy_vm_debugfs(kvm);
-	kvm_arch_sync_events(kvm);
 	mutex_lock(&kvm_lock);
 	list_del(&kvm->vm_list);
 	mutex_unlock(&kvm_lock);