Merge tag 'cgroup-for-6.4' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup

Pull cgroup updates from Tejun Heo:

 - cpuset changes including the fix for an incorrect interaction with
   CPU hotplug and an optimization

 - Other doc and cosmetic changes

* tag 'cgroup-for-6.4' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
  docs: cgroup-v1/cpusets: update libcgroup project link
  cgroup/cpuset: Minor updates to test_cpuset_prs.sh
  cgroup/cpuset: Include offline CPUs when tasks' cpumasks in top_cpuset are updated
  cgroup/cpuset: Skip task update if hotplug doesn't affect current cpuset
  cpuset: Clean up cpuset_node_allowed
  cgroup: bpf: use cgroup_lock()/cgroup_unlock() wrappers

10 files changed, 99 insertions, 100 deletions

diff --git a/Documentation/admin-guide/cgroup-v1/cpusets.rst b/Documentation/admin-guide/cgroup-v1/cpusets.rst
index 5d844ed4df69..ae646d621a8a 100644
--- a/Documentation/admin-guide/cgroup-v1/cpusets.rst
+++ b/Documentation/admin-guide/cgroup-v1/cpusets.rst
@@ -719,7 +719,7 @@ There are ways to query or modify cpusets:
    cat, rmdir commands from the shell, or their equivalent from C.
  - via the C library libcpuset.
  - via the C library libcgroup.
-   (http://sourceforge.net/projects/libcg/)
+   (https://github.com/libcgroup/libcgroup/)
  - via the python application cset.
    (http://code.google.com/p/cpuset/)
 
diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h
index d58e0476ee8e..980b76a1237e 100644
--- a/include/linux/cpuset.h
+++ b/include/linux/cpuset.h
@@ -80,18 +80,11 @@ extern nodemask_t cpuset_mems_allowed(struct task_struct *p);
 void cpuset_init_current_mems_allowed(void);
 int cpuset_nodemask_valid_mems_allowed(nodemask_t *nodemask);
 
-extern bool __cpuset_node_allowed(int node, gfp_t gfp_mask);
-
-static inline bool cpuset_node_allowed(int node, gfp_t gfp_mask)
-{
-	if (cpusets_enabled())
-		return __cpuset_node_allowed(node, gfp_mask);
-	return true;
-}
+extern bool cpuset_node_allowed(int node, gfp_t gfp_mask);
 
 static inline bool __cpuset_zone_allowed(struct zone *z, gfp_t gfp_mask)
 {
-	return __cpuset_node_allowed(zone_to_nid(z), gfp_mask);
+	return cpuset_node_allowed(zone_to_nid(z), gfp_mask);
 }
 
 static inline bool cpuset_zone_allowed(struct zone *z, gfp_t gfp_mask)
@@ -223,11 +216,6 @@ static inline int cpuset_nodemask_valid_mems_allowed(nodemask_t *nodemask)
 	return 1;
 }
 
-static inline bool cpuset_node_allowed(int node, gfp_t gfp_mask)
-{
-	return true;
-}
-
 static inline bool __cpuset_zone_allowed(struct zone *z, gfp_t gfp_mask)
 {
 	return true;
diff --git a/kernel/bpf/cgroup.c b/kernel/bpf/cgroup.c
index a06e118a9be5..517b6a5928cc 100644
--- a/kernel/bpf/cgroup.c
+++ b/kernel/bpf/cgroup.c
@@ -173,11 +173,11 @@ void bpf_cgroup_atype_put(int cgroup_atype)
 {
 	int i = cgroup_atype - CGROUP_LSM_START;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	if (--cgroup_lsm_atype[i].refcnt <= 0)
 		cgroup_lsm_atype[i].attach_btf_id = 0;
 	WARN_ON_ONCE(cgroup_lsm_atype[i].refcnt < 0);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 }
 #else
 static enum cgroup_bpf_attach_type
@@ -282,7 +282,7 @@ static void cgroup_bpf_release(struct work_struct *work)
 
 	unsigned int atype;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	for (atype = 0; atype < ARRAY_SIZE(cgrp->bpf.progs); atype++) {
 		struct hlist_head *progs = &cgrp->bpf.progs[atype];
@@ -315,7 +315,7 @@ static void cgroup_bpf_release(struct work_struct *work)
 		bpf_cgroup_storage_free(storage);
 	}
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	for (p = cgroup_parent(cgrp); p; p = cgroup_parent(p))
 		cgroup_bpf_put(p);
@@ -729,9 +729,9 @@ static int cgroup_bpf_attach(struct cgroup *cgrp,
 {
 	int ret;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	ret = __cgroup_bpf_attach(cgrp, prog, replace_prog, link, type, flags);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	return ret;
 }
 
@@ -831,7 +831,7 @@ static int cgroup_bpf_replace(struct bpf_link *link, struct bpf_prog *new_prog,
 
 	cg_link = container_of(link, struct bpf_cgroup_link, link);
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	/* link might have been auto-released by dying cgroup, so fail */
 	if (!cg_link->cgroup) {
 		ret = -ENOLINK;
@@ -843,7 +843,7 @@ static int cgroup_bpf_replace(struct bpf_link *link, struct bpf_prog *new_prog,
 	}
 	ret = __cgroup_bpf_replace(cg_link->cgroup, cg_link, new_prog);
 out_unlock:
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	return ret;
 }
 
@@ -1009,9 +1009,9 @@ static int cgroup_bpf_detach(struct cgroup *cgrp, struct bpf_prog *prog,
 {
 	int ret;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	ret = __cgroup_bpf_detach(cgrp, prog, NULL, type);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	return ret;
 }
 
@@ -1120,9 +1120,9 @@ static int cgroup_bpf_query(struct cgroup *cgrp, const union bpf_attr *attr,
 {
 	int ret;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	ret = __cgroup_bpf_query(cgrp, attr, uattr);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	return ret;
 }
 
@@ -1189,11 +1189,11 @@ static void bpf_cgroup_link_release(struct bpf_link *link)
 	if (!cg_link->cgroup)
 		return;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	/* re-check cgroup under lock again */
 	if (!cg_link->cgroup) {
-		mutex_unlock(&cgroup_mutex);
+		cgroup_unlock();
 		return;
 	}
 
@@ -1205,7 +1205,7 @@ static void bpf_cgroup_link_release(struct bpf_link *link)
 	cg = cg_link->cgroup;
 	cg_link->cgroup = NULL;
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	cgroup_put(cg);
 }
@@ -1232,10 +1232,10 @@ static void bpf_cgroup_link_show_fdinfo(const struct bpf_link *link,
 		container_of(link, struct bpf_cgroup_link, link);
 	u64 cg_id = 0;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	if (cg_link->cgroup)
 		cg_id = cgroup_id(cg_link->cgroup);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	seq_printf(seq,
 		   "cgroup_id:\t%llu\n"
@@ -1251,10 +1251,10 @@ static int bpf_cgroup_link_fill_link_info(const struct bpf_link *link,
 		container_of(link, struct bpf_cgroup_link, link);
 	u64 cg_id = 0;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	if (cg_link->cgroup)
 		cg_id = cgroup_id(cg_link->cgroup);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	info->cgroup.cgroup_id = cg_id;
 	info->cgroup.attach_type = cg_link->type;
diff --git a/kernel/bpf/cgroup_iter.c b/kernel/bpf/cgroup_iter.c
index 06989d278846..810378f04fbc 100644
--- a/kernel/bpf/cgroup_iter.c
+++ b/kernel/bpf/cgroup_iter.c
@@ -58,7 +58,7 @@ static void *cgroup_iter_seq_start(struct seq_file *seq, loff_t *pos)
 {
 	struct cgroup_iter_priv *p = seq->private;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	/* cgroup_iter doesn't support read across multiple sessions. */
 	if (*pos > 0) {
@@ -89,7 +89,7 @@ static void cgroup_iter_seq_stop(struct seq_file *seq, void *v)
 {
 	struct cgroup_iter_priv *p = seq->private;
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	/* pass NULL to the prog for post-processing */
 	if (!v) {
diff --git a/kernel/bpf/local_storage.c b/kernel/bpf/local_storage.c
index 4c7bbec4a9e4..a04f505aefe9 100644
--- a/kernel/bpf/local_storage.c
+++ b/kernel/bpf/local_storage.c
@@ -333,14 +333,14 @@ static void cgroup_storage_map_free(struct bpf_map *_map)
 	struct list_head *storages = &map->list;
 	struct bpf_cgroup_storage *storage, *stmp;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	list_for_each_entry_safe(storage, stmp, storages, list_map) {
 		bpf_cgroup_storage_unlink(storage);
 		bpf_cgroup_storage_free(storage);
 	}
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	WARN_ON(!RB_EMPTY_ROOT(&map->root));
 	WARN_ON(!list_empty(&map->list));
diff --git a/kernel/cgroup/cgroup-v1.c b/kernel/cgroup/cgroup-v1.c
index 52bb5a74a23b..aeef06c465ef 100644
--- a/kernel/cgroup/cgroup-v1.c
+++ b/kernel/cgroup/cgroup-v1.c
@@ -58,7 +58,7 @@ int cgroup_attach_task_all(struct task_struct *from, struct task_struct *tsk)
 	struct cgroup_root *root;
 	int retval = 0;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	cgroup_attach_lock(true);
 	for_each_root(root) {
 		struct cgroup *from_cgrp;
@@ -72,7 +72,7 @@ int cgroup_attach_task_all(struct task_struct *from, struct task_struct *tsk)
 			break;
 	}
 	cgroup_attach_unlock(true);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	return retval;
 }
@@ -106,7 +106,7 @@ int cgroup_transfer_tasks(struct cgroup *to, struct cgroup *from)
 	if (ret)
 		return ret;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	percpu_down_write(&cgroup_threadgroup_rwsem);
 
@@ -145,7 +145,7 @@ int cgroup_transfer_tasks(struct cgroup *to, struct cgroup *from)
 out_err:
 	cgroup_migrate_finish(&mgctx);
 	percpu_up_write(&cgroup_threadgroup_rwsem);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	return ret;
 }
 
@@ -847,13 +847,13 @@ static int cgroup1_rename(struct kernfs_node *kn, struct kernfs_node *new_parent
 	kernfs_break_active_protection(new_parent);
 	kernfs_break_active_protection(kn);
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	ret = kernfs_rename(kn, new_parent, new_name_str);
 	if (!ret)
 		TRACE_CGROUP_PATH(rename, cgrp);
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	kernfs_unbreak_active_protection(kn);
 	kernfs_unbreak_active_protection(new_parent);
@@ -1119,7 +1119,7 @@ int cgroup1_reconfigure(struct fs_context *fc)
 	trace_cgroup_remount(root);
 
  out_unlock:
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	return ret;
 }
 
@@ -1246,7 +1246,7 @@ int cgroup1_get_tree(struct fs_context *fc)
 	if (!ret && !percpu_ref_tryget_live(&ctx->root->cgrp.self.refcnt))
 		ret = 1;	/* restart */
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	if (!ret)
 		ret = cgroup_do_get_tree(fc);
diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index f2a77b1b45a2..625d7483951c 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -1391,7 +1391,7 @@ static void cgroup_destroy_root(struct cgroup_root *root)
 	cgroup_favor_dynmods(root, false);
 	cgroup_exit_root_id(root);
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	cgroup_rstat_exit(cgrp);
 	kernfs_destroy_root(root->kf_root);
@@ -1635,7 +1635,7 @@ void cgroup_kn_unlock(struct kernfs_node *kn)
 	else
 		cgrp = kn->parent->priv;
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	kernfs_unbreak_active_protection(kn);
 	cgroup_put(cgrp);
@@ -1680,7 +1680,7 @@ struct cgroup *cgroup_kn_lock_live(struct kernfs_node *kn, bool drain_offline)
 	if (drain_offline)
 		cgroup_lock_and_drain_offline(cgrp);
 	else
-		mutex_lock(&cgroup_mutex);
+		cgroup_lock();
 
 	if (!cgroup_is_dead(cgrp))
 		return cgrp;
@@ -2177,13 +2177,13 @@ int cgroup_do_get_tree(struct fs_context *fc)
 		struct super_block *sb = fc->root->d_sb;
 		struct cgroup *cgrp;
 
-		mutex_lock(&cgroup_mutex);
+		cgroup_lock();
 		spin_lock_irq(&css_set_lock);
 
 		cgrp = cset_cgroup_from_root(ctx->ns->root_cset, ctx->root);
 
 		spin_unlock_irq(&css_set_lock);
-		mutex_unlock(&cgroup_mutex);
+		cgroup_unlock();
 
 		nsdentry = kernfs_node_dentry(cgrp->kn, sb);
 		dput(fc->root);
@@ -2366,13 +2366,13 @@ int cgroup_path_ns(struct cgroup *cgrp, char *buf, size_t buflen,
 {
 	int ret;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	spin_lock_irq(&css_set_lock);
 
 	ret = cgroup_path_ns_locked(cgrp, buf, buflen, ns);
 
 	spin_unlock_irq(&css_set_lock);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	return ret;
 }
@@ -2398,7 +2398,7 @@ int task_cgroup_path(struct task_struct *task, char *buf, size_t buflen)
 	int hierarchy_id = 1;
 	int ret;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	spin_lock_irq(&css_set_lock);
 
 	root = idr_get_next(&cgroup_hierarchy_idr, &hierarchy_id);
@@ -2412,7 +2412,7 @@ int task_cgroup_path(struct task_struct *task, char *buf, size_t buflen)
 	}
 
 	spin_unlock_irq(&css_set_lock);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	return ret;
 }
 EXPORT_SYMBOL_GPL(task_cgroup_path);
@@ -3121,7 +3121,7 @@ void cgroup_lock_and_drain_offline(struct cgroup *cgrp)
 	int ssid;
 
 restart:
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	cgroup_for_each_live_descendant_post(dsct, d_css, cgrp) {
 		for_each_subsys(ss, ssid) {
@@ -3135,7 +3135,7 @@ restart:
 			prepare_to_wait(&dsct->offline_waitq, &wait,
 					TASK_UNINTERRUPTIBLE);
 
-			mutex_unlock(&cgroup_mutex);
+			cgroup_unlock();
 			schedule();
 			finish_wait(&dsct->offline_waitq, &wait);
 
@@ -4384,9 +4384,9 @@ int cgroup_rm_cftypes(struct cftype *cfts)
 	if (!(cfts[0].flags & __CFTYPE_ADDED))
 		return -ENOENT;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	ret = cgroup_rm_cftypes_locked(cfts);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	return ret;
 }
 
@@ -4418,14 +4418,14 @@ static int cgroup_add_cftypes(struct cgroup_subsys *ss, struct cftype *cfts)
 	if (ret)
 		return ret;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	list_add_tail(&cfts->node, &ss->cfts);
 	ret = cgroup_apply_cftypes(cfts, true);
 	if (ret)
 		cgroup_rm_cftypes_locked(cfts);
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	return ret;
 }
 
@@ -5395,7 +5395,7 @@ static void css_release_work_fn(struct work_struct *work)
 	struct cgroup_subsys *ss = css->ss;
 	struct cgroup *cgrp = css->cgroup;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	css->flags |= CSS_RELEASED;
 	list_del_rcu(&css->sibling);
@@ -5436,7 +5436,7 @@ static void css_release_work_fn(struct work_struct *work)
 					 NULL);
 	}
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	INIT_RCU_WORK(&css->destroy_rwork, css_free_rwork_fn);
 	queue_rcu_work(cgroup_destroy_wq, &css->destroy_rwork);
@@ -5784,7 +5784,7 @@ static void css_killed_work_fn(struct work_struct *work)
 	struct cgroup_subsys_state *css =
 		container_of(work, struct cgroup_subsys_state, destroy_work);
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	do {
 		offline_css(css);
@@ -5793,7 +5793,7 @@ static void css_killed_work_fn(struct work_struct *work)
 		css = css->parent;
 	} while (css && atomic_dec_and_test(&css->online_cnt));
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 }
 
 /* css kill confirmation processing requires process context, bounce */
@@ -5977,7 +5977,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss, bool early)
 
 	pr_debug("Initializing cgroup subsys %s\n", ss->name);
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	idr_init(&ss->css_idr);
 	INIT_LIST_HEAD(&ss->cfts);
@@ -6021,7 +6021,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss, bool early)
 
 	BUG_ON(online_css(css));
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 }
 
 /**
@@ -6081,7 +6081,7 @@ int __init cgroup_init(void)
 
 	get_user_ns(init_cgroup_ns.user_ns);
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 
 	/*
 	 * Add init_css_set to the hash table so that dfl_root can link to
@@ -6092,7 +6092,7 @@ int __init cgroup_init(void)
 
 	BUG_ON(cgroup_setup_root(&cgrp_dfl_root, 0));
 
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 
 	for_each_subsys(ss, ssid) {
 		if (ss->early_init) {
@@ -6144,9 +6144,9 @@ int __init cgroup_init(void)
 		if (ss->bind)
 			ss->bind(init_css_set.subsys[ssid]);
 
-		mutex_lock(&cgroup_mutex);
+		cgroup_lock();
 		css_populate_dir(init_css_set.subsys[ssid]);
-		mutex_unlock(&cgroup_mutex);
+		cgroup_unlock();
 	}
 
 	/* init_css_set.subsys[] has been updated, re-hash */
@@ -6251,7 +6251,7 @@ int proc_cgroup_show(struct seq_file *m, struct pid_namespace *ns,
 	if (!buf)
 		goto out;
 
-	mutex_lock(&cgroup_mutex);
+	cgroup_lock();
 	spin_lock_irq(&css_set_lock);
 
 	for_each_root(root) {
@@ -6306,7 +6306,7 @@ int proc_cgroup_show(struct seq_file *m, struct pid_namespace *ns,
 	retval = 0;
 out_unlock:
 	spin_unlock_irq(&css_set_lock);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	kfree(buf);
 out:
 	return retval;
@@ -6390,7 +6390,7 @@ static int cgroup_css_set_fork(struct kernel_clone_args *kargs)
 	struct file *f;
 
 	if (kargs->flags & CLONE_INTO_CGROUP)
-		mutex_lock(&cgroup_mutex);
+		cgroup_lock();
 
 	cgroup_threadgroup_change_begin(current);
 
@@ -6465,7 +6465,7 @@ static int cgroup_css_set_fork(struct kernel_clone_args *kargs)
 
 err:
 	cgroup_threadgroup_change_end(current);
-	mutex_unlock(&cgroup_mutex);
+	cgroup_unlock();
 	if (f)
 		fput(f);
 	if (dst_cgrp)
@@ -6492,7 +6492,7 @@ static void cgroup_css_set_put_fork(struct kernel_clone_args *kargs)
 		struct cgroup *cgrp = kargs->cgrp;
 		struct css_set *cset = kargs->cset;
 
-		mutex_unlock(&cgroup_mutex);
+		cgroup_unlock();
 
 		if (cset) {
 			put_css_set(cset);
diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index 505d86b16642..e4ca2dd2b764 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -1209,7 +1209,9 @@ void rebuild_sched_domains(void)
  *
  * Iterate through each task of @cs updating its cpus_allowed to the
  * effective cpuset's.  As this function is called with cpuset_rwsem held,
- * cpuset membership stays stable.
+ * cpuset membership stays stable. For top_cpuset, task_cpu_possible_mask()
+ * is used instead of effective_cpus to make sure all offline CPUs are also
+ * included as hotplug code won't update cpumasks for tasks in top_cpuset.
  */
 static void update_tasks_cpumask(struct cpuset *cs, struct cpumask *new_cpus)
 {
@@ -1219,15 +1221,18 @@ static void update_tasks_cpumask(struct cpuset *cs, struct cpumask *new_cpus)
 
 	css_task_iter_start(&cs->css, 0, &it);
 	while ((task = css_task_iter_next(&it))) {
-		/*
-		 * Percpu kthreads in top_cpuset are ignored
-		 */
-		if (top_cs && (task->flags & PF_KTHREAD) &&
-		    kthread_is_per_cpu(task))
-			continue;
+		const struct cpumask *possible_mask = task_cpu_possible_mask(task);
 
-		cpumask_and(new_cpus, cs->effective_cpus,
-			    task_cpu_possible_mask(task));
+		if (top_cs) {
+			/*
+			 * Percpu kthreads in top_cpuset are ignored
+			 */
+			if ((task->flags & PF_KTHREAD) && kthread_is_per_cpu(task))
+				continue;
+			cpumask_andnot(new_cpus, possible_mask, cs->subparts_cpus);
+		} else {
+			cpumask_and(new_cpus, possible_mask, cs->effective_cpus);
+		}
 		set_cpus_allowed_ptr(task, new_cpus);
 	}
 	css_task_iter_end(&it);
@@ -3618,6 +3623,8 @@ retry:
 update_tasks:
 	cpus_updated = !cpumask_equal(&new_cpus, cs->effective_cpus);
 	mems_updated = !nodes_equal(new_mems, cs->effective_mems);
+	if (!cpus_updated && !mems_updated)
+		goto unlock;	/* Hotplug doesn't affect this cpuset */
 
 	if (mems_updated)
 		check_insane_mems_config(&new_mems);
@@ -3629,6 +3636,7 @@ update_tasks:
 		hotplug_update_tasks_legacy(cs, &new_cpus, &new_mems,
 					    cpus_updated, mems_updated);
 
+unlock:
 	percpu_up_write(&cpuset_rwsem);
 }
 
@@ -3941,7 +3949,7 @@ static struct cpuset *nearest_hardwall_ancestor(struct cpuset *cs)
 }
 
 /*
- * __cpuset_node_allowed - Can we allocate on a memory node?
+ * cpuset_node_allowed - Can we allocate on a memory node?
  * @node: is this an allowed node?
  * @gfp_mask: memory allocation flags
  *
@@ -3980,7 +3988,7 @@ static struct cpuset *nearest_hardwall_ancestor(struct cpuset *cs)
  *	GFP_KERNEL   - any node in enclosing hardwalled cpuset ok
  *	GFP_USER     - only nodes in current tasks mems allowed ok.
  */
-bool __cpuset_node_allowed(int node, gfp_t gfp_mask)
+bool cpuset_node_allowed(int node, gfp_t gfp_mask)
 {
 	struct cpuset *cs;		/* current cpuset ancestors */
 	bool allowed;			/* is allocation in zone z allowed? */
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 6da423ec356f..9de2a18519a1 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3385,7 +3385,7 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
 retry:
 	/*
 	 * Scan zonelist, looking for a zone with enough free.
-	 * See also __cpuset_node_allowed() comment in kernel/cgroup/cpuset.c.
+	 * See also cpuset_node_allowed() comment in kernel/cgroup/cpuset.c.
 	 */
 	no_fallback = alloc_flags & ALLOC_NOFRAGMENT;
 	z = ac->preferred_zoneref;
@@ -4059,7 +4059,7 @@ gfp_to_alloc_flags(gfp_t gfp_mask, unsigned int order)
 		/*
 		 * Ignore cpuset mems for non-blocking __GFP_HIGH (probably
 		 * GFP_ATOMIC) rather than fail, see the comment for
-		 * __cpuset_node_allowed().
+		 * cpuset_node_allowed().
 		 */
 		if (alloc_flags & ALLOC_MIN_RESERVE)
 			alloc_flags &= ~ALLOC_CPUSET;
diff --git a/tools/testing/selftests/cgroup/test_cpuset_prs.sh b/tools/testing/selftests/cgroup/test_cpuset_prs.sh
index 75c100de90ff..2b5215cc599f 100755
--- a/tools/testing/selftests/cgroup/test_cpuset_prs.sh
+++ b/tools/testing/selftests/cgroup/test_cpuset_prs.sh
@@ -15,13 +15,6 @@ skip_test() {
 
 [[ $(id -u) -eq 0 ]] || skip_test "Test must be run as root!"
 
-# Set sched verbose flag, if available
-if [[ -d /sys/kernel/debug/sched ]]
-then
-	# Used to restore the original setting during cleanup
-	SCHED_DEBUG=$(cat /sys/kernel/debug/sched/verbose)
-	echo Y > /sys/kernel/debug/sched/verbose
-fi
 
 # Get wait_inotify location
 WAIT_INOTIFY=$(cd $(dirname $0); pwd)/wait_inotify
@@ -37,10 +30,14 @@ CPUS=$(lscpu | grep "^CPU(s):" | sed -e "s/.*:[[:space:]]*//")
 PROG=$1
 VERBOSE=
 DELAY_FACTOR=1
+SCHED_DEBUG=
 while [[ "$1" = -* ]]
 do
 	case "$1" in
 		-v) VERBOSE=1
+		    # Enable sched/verbose can slow thing down
+		    [[ $DELAY_FACTOR -eq 1 ]] &&
+			DELAY_FACTOR=2
 		    break
 		    ;;
 		-d) DELAY_FACTOR=$2
@@ -54,6 +51,14 @@ do
 	shift
 done
 
+# Set sched verbose flag if available when "-v" option is specified
+if [[ -n "$VERBOSE" && -d /sys/kernel/debug/sched ]]
+then
+	# Used to restore the original setting during cleanup
+	SCHED_DEBUG=$(cat /sys/kernel/debug/sched/verbose)
+	echo Y > /sys/kernel/debug/sched/verbose
+fi
+
 cd $CGROUP2
 echo +cpuset > cgroup.subtree_control
 [[ -d test ]] || mkdir test
@@ -65,7 +70,8 @@ cleanup()
 	rmdir A1/A2/A3 A1/A2 A1 B1 > /dev/null 2>&1
 	cd ..
 	rmdir test > /dev/null 2>&1
-	echo "$SCHED_DEBUG" > /sys/kernel/debug/sched/verbose
+	[[ -n "$SCHED_DEBUG" ]] &&
+		echo "$SCHED_DEBUG" > /sys/kernel/debug/sched/verbose
 }
 
 # Pause in ms
@@ -571,7 +577,6 @@ run_state_test()
 			echo "Test $TEST[$I] failed result check!"
 			eval echo \"\${$TEST[$I]}\"
 			dump_states
-			online_cpus
 			exit 1
 		}
 
@@ -582,7 +587,6 @@ run_state_test()
 				eval echo \"\${$TEST[$I]}\"
 				echo
 				dump_states
-				online_cpus
 				exit 1
 			}
 		}
@@ -594,7 +598,6 @@ run_state_test()
 				eval echo \"\${$TEST[$I]}\"
 				echo
 				dump_states
-				online_cpus
 				exit 1
 			}
 		}