Merge tag 'kvm-4.13-2' of git://git.kernel.org/pub/scm/virt/kvm/kvm

Pull more KVM updates from Radim Krčmář:
 "Second batch of KVM updates for v4.13

  Common:
   - add uevents for VM creation/destruction
   - annotate and properly access RCU-protected objects

  s390:
   - rename IOCTL added in the first v4.13 merge

  x86:
   - emulate VMLOAD VMSAVE feature in SVM
   - support paravirtual asynchronous page fault while nested
   - add Hyper-V userspace interfaces for better migration
   - improve master clock corner cases
   - extend internal error reporting after EPT misconfig
   - correct single-stepping of emulated instructions in SVM
   - handle MCE during VM entry
   - fix nVMX VM entry checks and nVMX VMCS shadowing"

* tag 'kvm-4.13-2' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (28 commits)
  kvm: x86: hyperv: make VP_INDEX managed by userspace
  KVM: async_pf: Let guest support delivery of async_pf from guest mode
  KVM: async_pf: Force a nested vmexit if the injected #PF is async_pf
  KVM: async_pf: Add L1 guest async_pf #PF vmexit handler
  KVM: x86: Simplify kvm_x86_ops->queue_exception parameter list
  kvm: x86: hyperv: add KVM_CAP_HYPERV_SYNIC2
  KVM: x86: make backwards_tsc_observed a per-VM variable
  KVM: trigger uevents when creating or destroying a VM
  KVM: SVM: Enable Virtual VMLOAD VMSAVE feature
  KVM: SVM: Add Virtual VMLOAD VMSAVE feature definition
  KVM: SVM: Rename lbr_ctl field in the vmcb control area
  KVM: SVM: Prepare for new bit definition in lbr_ctl
  KVM: SVM: handle singlestep exception when skipping emulated instructions
  KVM: x86: take slots_lock in kvm_free_pit
  KVM: s390: Fix KVM_S390_GET_CMMA_BITS ioctl definition
  kvm: vmx: Properly handle machine check during VM-entry
  KVM: x86: update master clock before computing kvmclock_offset
  kvm: nVMX: Shadow "high" parts of shadowed 64-bit VMCS fields
  kvm: nVMX: Fix nested_vmx_check_msr_bitmap_controls
  kvm: nVMX: Validate the I/O bitmaps on nested VM-entry
  ...

3 files changed, 115 insertions, 26 deletions

diff --git a/virt/kvm/eventfd.c b/virt/kvm/eventfd.c
index 9120edf3c94b..f2ac53ab8243 100644
--- a/virt/kvm/eventfd.c
+++ b/virt/kvm/eventfd.c
@@ -825,7 +825,7 @@ static int kvm_assign_ioeventfd_idx(struct kvm *kvm,
 	if (ret < 0)
 		goto unlock_fail;
 
-	kvm->buses[bus_idx]->ioeventfd_count++;
+	kvm_get_bus(kvm, bus_idx)->ioeventfd_count++;
 	list_add_tail(&p->list, &kvm->ioeventfds);
 
 	mutex_unlock(&kvm->slots_lock);
@@ -848,6 +848,7 @@ kvm_deassign_ioeventfd_idx(struct kvm *kvm, enum kvm_bus bus_idx,
 {
 	struct _ioeventfd        *p, *tmp;
 	struct eventfd_ctx       *eventfd;
+	struct kvm_io_bus	 *bus;
 	int                       ret = -ENOENT;
 
 	eventfd = eventfd_ctx_fdget(args->fd);
@@ -870,8 +871,9 @@ kvm_deassign_ioeventfd_idx(struct kvm *kvm, enum kvm_bus bus_idx,
 			continue;
 
 		kvm_io_bus_unregister_dev(kvm, bus_idx, &p->dev);
-		if (kvm->buses[bus_idx])
-			kvm->buses[bus_idx]->ioeventfd_count--;
+		bus = kvm_get_bus(kvm, bus_idx);
+		if (bus)
+			bus->ioeventfd_count--;
 		ioeventfd_release(p);
 		ret = 0;
 		break;
diff --git a/virt/kvm/irqchip.c b/virt/kvm/irqchip.c
index 31e40c9e81df..b1286c4e0712 100644
--- a/virt/kvm/irqchip.c
+++ b/virt/kvm/irqchip.c
@@ -230,7 +230,7 @@ int kvm_set_irq_routing(struct kvm *kvm,
 	}
 
 	mutex_lock(&kvm->irq_lock);
-	old = kvm->irq_routing;
+	old = rcu_dereference_protected(kvm->irq_routing, 1);
 	rcu_assign_pointer(kvm->irq_routing, new);
 	kvm_irq_routing_update(kvm);
 	kvm_arch_irq_routing_update(kvm);
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 19f0ecb9b93e..82987d457b8b 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -130,6 +130,12 @@ EXPORT_SYMBOL_GPL(kvm_rebooting);
 
 static bool largepages_enabled = true;
 
+#define KVM_EVENT_CREATE_VM 0
+#define KVM_EVENT_DESTROY_VM 1
+static void kvm_uevent_notify_change(unsigned int type, struct kvm *kvm);
+static unsigned long long kvm_createvm_count;
+static unsigned long long kvm_active_vms;
+
 bool kvm_is_reserved_pfn(kvm_pfn_t pfn)
 {
 	if (pfn_valid(pfn))
@@ -187,12 +193,23 @@ static void ack_flush(void *_completed)
 {
 }
 
+static inline bool kvm_kick_many_cpus(const struct cpumask *cpus, bool wait)
+{
+	if (unlikely(!cpus))
+		cpus = cpu_online_mask;
+
+	if (cpumask_empty(cpus))
+		return false;
+
+	smp_call_function_many(cpus, ack_flush, NULL, wait);
+	return true;
+}
+
 bool kvm_make_all_cpus_request(struct kvm *kvm, unsigned int req)
 {
 	int i, cpu, me;
 	cpumask_var_t cpus;
-	bool called = true;
-	bool wait = req & KVM_REQUEST_WAIT;
+	bool called;
 	struct kvm_vcpu *vcpu;
 
 	zalloc_cpumask_var(&cpus, GFP_ATOMIC);
@@ -207,14 +224,9 @@ bool kvm_make_all_cpus_request(struct kvm *kvm, unsigned int req)
 
 		if (cpus != NULL && cpu != -1 && cpu != me &&
 		    kvm_request_needs_ipi(vcpu, req))
-			cpumask_set_cpu(cpu, cpus);
+			__cpumask_set_cpu(cpu, cpus);
 	}
-	if (unlikely(cpus == NULL))
-		smp_call_function_many(cpu_online_mask, ack_flush, NULL, wait);
-	else if (!cpumask_empty(cpus))
-		smp_call_function_many(cpus, ack_flush, NULL, wait);
-	else
-		called = false;
+	called = kvm_kick_many_cpus(cpus, !!(req & KVM_REQUEST_WAIT));
 	put_cpu();
 	free_cpumask_var(cpus);
 	return called;
@@ -293,7 +305,12 @@ EXPORT_SYMBOL_GPL(kvm_vcpu_init);
 
 void kvm_vcpu_uninit(struct kvm_vcpu *vcpu)
 {
-	put_pid(vcpu->pid);
+	/*
+	 * no need for rcu_read_lock as VCPU_RUN is the only place that
+	 * will change the vcpu->pid pointer and on uninit all file
+	 * descriptors are already gone.
+	 */
+	put_pid(rcu_dereference_protected(vcpu->pid, 1));
 	kvm_arch_vcpu_uninit(vcpu);
 	free_page((unsigned long)vcpu->run);
 }
@@ -674,8 +691,8 @@ static struct kvm *kvm_create_vm(unsigned long type)
 	if (init_srcu_struct(&kvm->irq_srcu))
 		goto out_err_no_irq_srcu;
 	for (i = 0; i < KVM_NR_BUSES; i++) {
-		kvm->buses[i] = kzalloc(sizeof(struct kvm_io_bus),
-					GFP_KERNEL);
+		rcu_assign_pointer(kvm->buses[i],
+			kzalloc(sizeof(struct kvm_io_bus), GFP_KERNEL));
 		if (!kvm->buses[i])
 			goto out_err;
 	}
@@ -700,9 +717,10 @@ out_err_no_srcu:
 	hardware_disable_all();
 out_err_no_disable:
 	for (i = 0; i < KVM_NR_BUSES; i++)
-		kfree(kvm->buses[i]);
+		kfree(rcu_access_pointer(kvm->buses[i]));
 	for (i = 0; i < KVM_ADDRESS_SPACE_NUM; i++)
-		kvm_free_memslots(kvm, kvm->memslots[i]);
+		kvm_free_memslots(kvm,
+			rcu_dereference_protected(kvm->memslots[i], 1));
 	kvm_arch_free_vm(kvm);
 	mmdrop(current->mm);
 	return ERR_PTR(r);
@@ -728,6 +746,7 @@ static void kvm_destroy_vm(struct kvm *kvm)
 	int i;
 	struct mm_struct *mm = kvm->mm;
 
+	kvm_uevent_notify_change(KVM_EVENT_DESTROY_VM, kvm);
 	kvm_destroy_vm_debugfs(kvm);
 	kvm_arch_sync_events(kvm);
 	spin_lock(&kvm_lock);
@@ -735,8 +754,11 @@ static void kvm_destroy_vm(struct kvm *kvm)
 	spin_unlock(&kvm_lock);
 	kvm_free_irq_routing(kvm);
 	for (i = 0; i < KVM_NR_BUSES; i++) {
-		if (kvm->buses[i])
-			kvm_io_bus_destroy(kvm->buses[i]);
+		struct kvm_io_bus *bus;
+
+		bus = rcu_dereference_protected(kvm->buses[i], 1);
+		if (bus)
+			kvm_io_bus_destroy(bus);
 		kvm->buses[i] = NULL;
 	}
 	kvm_coalesced_mmio_free(kvm);
@@ -748,7 +770,8 @@ static void kvm_destroy_vm(struct kvm *kvm)
 	kvm_arch_destroy_vm(kvm);
 	kvm_destroy_devices(kvm);
 	for (i = 0; i < KVM_ADDRESS_SPACE_NUM; i++)
-		kvm_free_memslots(kvm, kvm->memslots[i]);
+		kvm_free_memslots(kvm,
+			rcu_dereference_protected(kvm->memslots[i], 1));
 	cleanup_srcu_struct(&kvm->irq_srcu);
 	cleanup_srcu_struct(&kvm->srcu);
 	kvm_arch_free_vm(kvm);
@@ -2551,13 +2574,14 @@ static long kvm_vcpu_ioctl(struct file *filp,
 	if (r)
 		return r;
 	switch (ioctl) {
-	case KVM_RUN:
+	case KVM_RUN: {
+		struct pid *oldpid;
 		r = -EINVAL;
 		if (arg)
 			goto out;
-		if (unlikely(vcpu->pid != current->pids[PIDTYPE_PID].pid)) {
+		oldpid = rcu_access_pointer(vcpu->pid);
+		if (unlikely(oldpid != current->pids[PIDTYPE_PID].pid)) {
 			/* The thread running this VCPU changed. */
-			struct pid *oldpid = vcpu->pid;
 			struct pid *newpid = get_task_pid(current, PIDTYPE_PID);
 
 			rcu_assign_pointer(vcpu->pid, newpid);
@@ -2568,6 +2592,7 @@ static long kvm_vcpu_ioctl(struct file *filp,
 		r = kvm_arch_vcpu_ioctl_run(vcpu, vcpu->run);
 		trace_kvm_userspace_exit(vcpu->run->exit_reason, r);
 		break;
+	}
 	case KVM_GET_REGS: {
 		struct kvm_regs *kvm_regs;
 
@@ -3202,6 +3227,7 @@ static int kvm_dev_ioctl_create_vm(unsigned long type)
 		fput(file);
 		return -ENOMEM;
 	}
+	kvm_uevent_notify_change(KVM_EVENT_CREATE_VM, kvm);
 
 	fd_install(r, file);
 	return r;
@@ -3563,7 +3589,7 @@ int kvm_io_bus_register_dev(struct kvm *kvm, enum kvm_bus bus_idx, gpa_t addr,
 {
 	struct kvm_io_bus *new_bus, *bus;
 
-	bus = kvm->buses[bus_idx];
+	bus = kvm_get_bus(kvm, bus_idx);
 	if (!bus)
 		return -ENOMEM;
 
@@ -3592,7 +3618,7 @@ void kvm_io_bus_unregister_dev(struct kvm *kvm, enum kvm_bus bus_idx,
 	int i;
 	struct kvm_io_bus *new_bus, *bus;
 
-	bus = kvm->buses[bus_idx];
+	bus = kvm_get_bus(kvm, bus_idx);
 	if (!bus)
 		return;
 
@@ -3854,6 +3880,67 @@ static const struct file_operations *stat_fops[] = {
 	[KVM_STAT_VM]   = &vm_stat_fops,
 };
 
+static void kvm_uevent_notify_change(unsigned int type, struct kvm *kvm)
+{
+	struct kobj_uevent_env *env;
+	char *tmp, *pathbuf = NULL;
+	unsigned long long created, active;
+
+	if (!kvm_dev.this_device || !kvm)
+		return;
+
+	spin_lock(&kvm_lock);
+	if (type == KVM_EVENT_CREATE_VM) {
+		kvm_createvm_count++;
+		kvm_active_vms++;
+	} else if (type == KVM_EVENT_DESTROY_VM) {
+		kvm_active_vms--;
+	}
+	created = kvm_createvm_count;
+	active = kvm_active_vms;
+	spin_unlock(&kvm_lock);
+
+	env = kzalloc(sizeof(*env), GFP_KERNEL);
+	if (!env)
+		return;
+
+	add_uevent_var(env, "CREATED=%llu", created);
+	add_uevent_var(env, "COUNT=%llu", active);
+
+	if (type == KVM_EVENT_CREATE_VM)
+		add_uevent_var(env, "EVENT=create");
+	else if (type == KVM_EVENT_DESTROY_VM)
+		add_uevent_var(env, "EVENT=destroy");
+
+	if (kvm->debugfs_dentry) {
+		char p[ITOA_MAX_LEN];
+
+		snprintf(p, sizeof(p), "%s", kvm->debugfs_dentry->d_name.name);
+		tmp = strchrnul(p + 1, '-');
+		*tmp = '\0';
+		add_uevent_var(env, "PID=%s", p);
+		pathbuf = kmalloc(PATH_MAX, GFP_KERNEL);
+		if (pathbuf) {
+			/* sizeof counts the final '\0' */
+			int len = sizeof("STATS_PATH=") - 1;
+			const char *pvar = "STATS_PATH=";
+
+			tmp = dentry_path_raw(kvm->debugfs_dentry,
+					      pathbuf + len,
+					      PATH_MAX - len);
+			if (!IS_ERR(tmp)) {
+				memcpy(tmp - len, pvar, len);
+				env->envp[env->envp_idx++] = tmp - len;
+			}
+		}
+	}
+	/* no need for checks, since we are adding at most only 5 keys */
+	env->envp[env->envp_idx++] = NULL;
+	kobject_uevent_env(&kvm_dev.this_device->kobj, KOBJ_CHANGE, env->envp);
+	kfree(env);
+	kfree(pathbuf);
+}
+
 static int kvm_init_debug(void)
 {
 	int r = -EEXIST;