diff --git a/drivers/target/target_core_tpg.c b/drivers/target/target_core_tpg.c
index c036595b17cfc9cfc75f9bd57d146bfb4179d67b..fddfae61222f2e3bf3a063d6ab11d6438e4ff897 100644
--- a/drivers/target/target_core_tpg.c
+++ b/drivers/target/target_core_tpg.c
@@ -825,7 +825,7 @@ int core_tpg_add_lun(
 
 	ret = core_dev_export(dev, tpg, lun);
 	if (ret < 0) {
-		percpu_ref_cancel_init(&lun->lun_ref);
+		percpu_ref_exit(&lun->lun_ref);
 		return ret;
 	}
 
@@ -880,5 +880,7 @@ int core_tpg_post_dellun(
 	lun->lun_status = TRANSPORT_LUN_STATUS_FREE;
 	spin_unlock(&tpg->tpg_lun_lock);
 
+	percpu_ref_exit(&lun->lun_ref);
+
 	return 0;
 }
diff --git a/fs/aio.c b/fs/aio.c
index 5e0d7f9cb693480f5e214c567585aaa5bdf32792..ea1bc2e8f4f30c2c05d1fd597e0c32cb80bd0c41 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -506,6 +506,8 @@ static void free_ioctx(struct work_struct *work)
 
 	aio_free_ring(ctx);
 	free_percpu(ctx->cpu);
+	percpu_ref_exit(&ctx->reqs);
+	percpu_ref_exit(&ctx->users);
 	kmem_cache_free(kioctx_cachep, ctx);
 }
 
@@ -715,8 +717,8 @@ static struct kioctx *ioctx_alloc(unsigned nr_events)
 err:
 	mutex_unlock(&ctx->ring_lock);
 	free_percpu(ctx->cpu);
-	percpu_ref_cancel_init(&ctx->reqs);
-	percpu_ref_cancel_init(&ctx->users);
+	percpu_ref_exit(&ctx->reqs);
+	percpu_ref_exit(&ctx->users);
 	kmem_cache_free(kioctx_cachep, ctx);
 	pr_debug("error allocating ioctx %d\n", err);
 	return ERR_PTR(err);
diff --git a/include/linux/percpu-refcount.h b/include/linux/percpu-refcount.h
index 6f8cd4c0546cb63ca18447962cd6cbf787adb8f6..0ddd2839ca846a63ba7f835a05d8d4bfa60e6c54 100644
--- a/include/linux/percpu-refcount.h
+++ b/include/linux/percpu-refcount.h
@@ -57,9 +57,7 @@ struct percpu_ref {
 	atomic_t		count;
 	/*
 	 * The low bit of the pointer indicates whether the ref is in percpu
-	 * mode; if set, then get/put will manipulate the atomic_t (this is a
-	 * hack because we need to keep the pointer around for
-	 * percpu_ref_kill_rcu())
+	 * mode; if set, then get/put will manipulate the atomic_t.
 	 */
 	unsigned long		pcpu_count_ptr;
 	percpu_ref_func_t	*release;
@@ -69,7 +67,7 @@ struct percpu_ref {
 
 int __must_check percpu_ref_init(struct percpu_ref *ref,
 				 percpu_ref_func_t *release);
-void percpu_ref_cancel_init(struct percpu_ref *ref);
+void percpu_ref_exit(struct percpu_ref *ref);
 void percpu_ref_kill_and_confirm(struct percpu_ref *ref,
 				 percpu_ref_func_t *confirm_kill);
 
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 7868fc3c0bc59b3223490114ca8e27f83b4c864c..c06aa5e257a89ab3e1e9bbf777713636bf222661 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -1638,7 +1638,7 @@ static int cgroup_setup_root(struct cgroup_root *root, unsigned int ss_mask)
 exit_root_id:
 	cgroup_exit_root_id(root);
 cancel_ref:
-	percpu_ref_cancel_init(&root_cgrp->self.refcnt);
+	percpu_ref_exit(&root_cgrp->self.refcnt);
 out:
 	free_cgrp_cset_links(&tmp_links);
 	return ret;
@@ -4133,6 +4133,8 @@ static void css_free_work_fn(struct work_struct *work)
 		container_of(work, struct cgroup_subsys_state, destroy_work);
 	struct cgroup *cgrp = css->cgroup;
 
+	percpu_ref_exit(&css->refcnt);
+
 	if (css->ss) {
 		/* css free path */
 		if (css->parent)
@@ -4330,7 +4332,7 @@ static int create_css(struct cgroup *cgrp, struct cgroup_subsys *ss)
 err_free_id:
 	cgroup_idr_remove(&ss->css_idr, css->id);
 err_free_percpu_ref:
-	percpu_ref_cancel_init(&css->refcnt);
+	percpu_ref_exit(&css->refcnt);
 err_free_css:
 	call_rcu(&css->rcu_head, css_free_rcu_fn);
 	return err;
@@ -4441,7 +4443,7 @@ static int cgroup_mkdir(struct kernfs_node *parent_kn, const char *name,
 out_free_id:
 	cgroup_idr_remove(&root->cgroup_idr, cgrp->id);
 out_cancel_ref:
-	percpu_ref_cancel_init(&cgrp->self.refcnt);
+	percpu_ref_exit(&cgrp->self.refcnt);
 out_free_cgrp:
 	kfree(cgrp);
 out_unlock:
diff --git a/lib/percpu-refcount.c b/lib/percpu-refcount.c
index 94e5b624de640612a639a12c5fdfe092f77b1270..ac4299120087e3ff25c9cc707f4c175dccfb2da5 100644
--- a/lib/percpu-refcount.c
+++ b/lib/percpu-refcount.c
@@ -61,36 +61,25 @@ int percpu_ref_init(struct percpu_ref *ref, percpu_ref_func_t *release)
 EXPORT_SYMBOL_GPL(percpu_ref_init);
 
 /**
- * percpu_ref_cancel_init - cancel percpu_ref_init()
- * @ref: percpu_ref to cancel init for
+ * percpu_ref_exit - undo percpu_ref_init()
+ * @ref: percpu_ref to exit
  *
- * Once a percpu_ref is initialized, its destruction is initiated by
- * percpu_ref_kill() and completes asynchronously, which can be painful to
- * do when destroying a half-constructed object in init failure path.
- *
- * This function destroys @ref without invoking @ref->release and the
- * memory area containing it can be freed immediately on return.  To
- * prevent accidental misuse, it's required that @ref has finished
- * percpu_ref_init(), whether successful or not, but never used.
- *
- * The weird name and usage restriction are to prevent people from using
- * this function by mistake for normal shutdown instead of
- * percpu_ref_kill().
+ * This function exits @ref.  The caller is responsible for ensuring that
+ * @ref is no longer in active use.  The usual places to invoke this
+ * function from are the @ref->release() callback or in init failure path
+ * where percpu_ref_init() succeeded but other parts of the initialization
+ * of the embedding object failed.
  */
-void percpu_ref_cancel_init(struct percpu_ref *ref)
+void percpu_ref_exit(struct percpu_ref *ref)
 {
 	unsigned __percpu *pcpu_count = pcpu_count_ptr(ref);
-	int cpu;
-
-	WARN_ON_ONCE(atomic_read(&ref->count) != 1 + PCPU_COUNT_BIAS);
 
 	if (pcpu_count) {
-		for_each_possible_cpu(cpu)
-			WARN_ON_ONCE(*per_cpu_ptr(pcpu_count, cpu));
 		free_percpu(pcpu_count);
+		ref->pcpu_count_ptr = PCPU_REF_DEAD;
 	}
 }
-EXPORT_SYMBOL_GPL(percpu_ref_cancel_init);
+EXPORT_SYMBOL_GPL(percpu_ref_exit);
 
 static void percpu_ref_kill_rcu(struct rcu_head *rcu)
 {
@@ -102,8 +91,6 @@ static void percpu_ref_kill_rcu(struct rcu_head *rcu)
 	for_each_possible_cpu(cpu)
 		count += *per_cpu_ptr(pcpu_count, cpu);
 
-	free_percpu(pcpu_count);
-
 	pr_debug("global %i pcpu %i", atomic_read(&ref->count), (int) count);
 
 	/*