cgroup: Changes for v6.13

- cpu.stat now also shows niced CPU time.
 
 - Freezer and cpuset optimizations.
 
 - Other misc changes.
 -----BEGIN PGP SIGNATURE-----
 
 iIQEABYKACwWIQTfIjM1kS57o3GsC/uxYfJx3gVYGQUCZztlgg4cdGpAa2VybmVs
 Lm9yZwAKCRCxYfJx3gVYGbohAQDE/enqpAX9vSOpQPne4ZzgcPlGTrCwBcka3Z5z
 4aOF0AD/SmdjcJ/EULisD/2O27ovsGAtqDjngrrZwNUTbCNkTQQ=
 =pKyo
 -----END PGP SIGNATURE-----

Merge tag 'cgroup-for-6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup

Pull cgroup updates from Tejun Heo:

 - cpu.stat now also shows niced CPU time

 - Freezer and cpuset optimizations

 - Other misc changes

* tag 'cgroup-for-6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
  cgroup/cpuset: Disable cpuset_cpumask_can_shrink() test if not load balancing
  cgroup/cpuset: Further optimize code if CONFIG_CPUSETS_V1 not set
  cgroup/cpuset: Enforce at most one rebuild_sched_domains_locked() call per operation
  cgroup/cpuset: Revert "Allow suppression of sched domain rebuild in update_cpumasks_hier()"
  MAINTAINERS: remove Zefan Li
  cgroup/freezer: Add cgroup CGRP_FROZEN flag update helper
  cgroup/freezer: Reduce redundant traversal for cgroup_freeze
  cgroup/bpf: only cgroup v2 can be attached by bpf programs
  Revert "cgroup: Fix memory leak caused by missing cgroup_bpf_offline"
  selftests/cgroup: Fix compile error in test_cpu.c
  cgroup/rstat: Selftests for niced CPU statistics
  cgroup/rstat: Tracking cgroup-level niced CPU time
  cgroup/cpuset: Fix spelling errors in file kernel/cgroup/cpuset.c

CREDITS

@@ -579,6 +579,9 @@ N: Zach Brown
 E: zab@zabbo.net
 D: maestro pci sound
 
+N: Zefan Li
+D: Contribution to control group stuff
+
 N: David Brownell
 D: Kernel engineer, mentor, and friend.  Maintained USB EHCI and
 D: gadget layers, SPI subsystem, GPIO subsystem, and more than a few

MAINTAINERS

@@ -5756,7 +5756,6 @@ F:	kernel/context_tracking.c
 
 CONTROL GROUP (CGROUP)
 M:	Tejun Heo <tj@kernel.org>
-M:	Zefan Li <lizefan.x@bytedance.com>
 M:	Johannes Weiner <hannes@cmpxchg.org>
 M:	Michal Koutný <mkoutny@suse.com>
 L:	cgroups@vger.kernel.org
@@ -5785,7 +5784,6 @@ F:	include/linux/blk-cgroup.h
 
 CONTROL GROUP - CPUSET
 M:	Waiman Long <longman@redhat.com>
-M:	Zefan Li <lizefan.x@bytedance.com>
 L:	cgroups@vger.kernel.org
 S:	Maintained
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git

include/linux/cgroup-defs.h

@@ -327,6 +327,7 @@ struct cgroup_base_stat {
 #ifdef CONFIG_SCHED_CORE
 	u64 forceidle_sum;
 #endif
+	u64 ntime;
 };
 
 /*
@@ -397,7 +398,7 @@ struct cgroup_freezer_state {
 	bool freeze;
 
 	/* Should the cgroup actually be frozen? */
-	int e_freeze;
+	bool e_freeze;
 
 	/* Fields below are protected by css_set_lock */
 

kernel/cgroup/cgroup.c

@@ -2140,8 +2140,10 @@ int cgroup_setup_root(struct cgroup_root *root, u16 ss_mask)
 	if (ret)
 		goto exit_stats;
 
-	ret = cgroup_bpf_inherit(root_cgrp);
+	if (root == &cgrp_dfl_root) {
-	WARN_ON_ONCE(ret);
+		ret = cgroup_bpf_inherit(root_cgrp);
+		WARN_ON_ONCE(ret);
+	}
 
 	trace_cgroup_setup_root(root);
 
@@ -2314,10 +2316,8 @@ static void cgroup_kill_sb(struct super_block *sb)
 	 * And don't kill the default root.
 	 */
 	if (list_empty(&root->cgrp.self.children) && root != &cgrp_dfl_root &&
-	    !percpu_ref_is_dying(&root->cgrp.self.refcnt)) {
+	    !percpu_ref_is_dying(&root->cgrp.self.refcnt))
-		cgroup_bpf_offline(&root->cgrp);
 		percpu_ref_kill(&root->cgrp.self.refcnt);
-	}
 	cgroup_put(&root->cgrp);
 	kernfs_kill_sb(sb);
 }
@@ -5710,9 +5710,11 @@ static struct cgroup *cgroup_create(struct cgroup *parent, const char *name,
 	if (ret)
 		goto out_kernfs_remove;
 
-	ret = cgroup_bpf_inherit(cgrp);
+	if (cgrp->root == &cgrp_dfl_root) {
-	if (ret)
+		ret = cgroup_bpf_inherit(cgrp);
-		goto out_psi_free;
+		if (ret)
+			goto out_psi_free;
+	}
 
 	/*
 	 * New cgroup inherits effective freeze counter, and
@@ -6026,7 +6028,8 @@ static int cgroup_destroy_locked(struct cgroup *cgrp)
 
 	cgroup1_check_for_release(parent);
 
-	cgroup_bpf_offline(cgrp);
+	if (cgrp->root == &cgrp_dfl_root)
+		cgroup_bpf_offline(cgrp);
 
 	/* put the base reference */
 	percpu_ref_kill(&cgrp->self.refcnt);

kernel/cgroup/cpuset.c

@@ -84,9 +84,19 @@ static bool		have_boot_isolcpus;
 static struct list_head remote_children;
 
 /*
- * A flag to force sched domain rebuild at the end of an operation while
+ * A flag to force sched domain rebuild at the end of an operation.
- * inhibiting it in the intermediate stages when set. Currently it is only
+ * It can be set in
- * set in hotplug code.
+ *  - update_partition_sd_lb()
+ *  - remote_partition_check()
+ *  - update_cpumasks_hier()
+ *  - cpuset_update_flag()
+ *  - cpuset_hotplug_update_tasks()
+ *  - cpuset_handle_hotplug()
+ *
+ * Protected by cpuset_mutex (with cpus_read_lock held) or cpus_write_lock.
+ *
+ * Note that update_relax_domain_level() in cpuset-v1.c can still call
+ * rebuild_sched_domains_locked() directly without using this flag.
  */
 static bool force_sd_rebuild;
 
@@ -283,6 +293,12 @@ static inline void dec_attach_in_progress(struct cpuset *cs)
 	mutex_unlock(&cpuset_mutex);
 }
 
+static inline bool cpuset_v2(void)
+{
+	return !IS_ENABLED(CONFIG_CPUSETS_V1) ||
+		cgroup_subsys_on_dfl(cpuset_cgrp_subsys);
+}
+
 /*
  * Cgroup v2 behavior is used on the "cpus" and "mems" control files when
  * on default hierarchy or when the cpuset_v2_mode flag is set by mounting
@@ -293,7 +309,7 @@ static inline void dec_attach_in_progress(struct cpuset *cs)
  */
 static inline bool is_in_v2_mode(void)
 {
-	return cgroup_subsys_on_dfl(cpuset_cgrp_subsys) ||
+	return cpuset_v2() ||
 	      (cpuset_cgrp_subsys.root->flags & CGRP_ROOT_CPUSET_V2_MODE);
 }
 
@@ -565,12 +581,24 @@ static int validate_change(struct cpuset *cur, struct cpuset *trial)
 
 	/*
 	 * We can't shrink if we won't have enough room for SCHED_DEADLINE
-	 * tasks.
+	 * tasks. This check is not done when scheduling is disabled as the
+	 * users should know what they are doing.
+	 *
+	 * For v1, effective_cpus == cpus_allowed & user_xcpus() returns
+	 * cpus_allowed.
+	 *
+	 * For v2, is_cpu_exclusive() & is_sched_load_balance() are true only
+	 * for non-isolated partition root. At this point, the target
+	 * effective_cpus isn't computed yet. user_xcpus() is the best
+	 * approximation.
+	 *
+	 * TBD: May need to precompute the real effective_cpus here in case
+	 * incorrect scheduling of SCHED_DEADLINE tasks in a partition
+	 * becomes an issue.
 	 */
 	ret = -EBUSY;
-	if (is_cpu_exclusive(cur) &&
+	if (is_cpu_exclusive(cur) && is_sched_load_balance(cur) &&
-	    !cpuset_cpumask_can_shrink(cur->cpus_allowed,
+	    !cpuset_cpumask_can_shrink(cur->effective_cpus, user_xcpus(trial)))
-				       trial->cpus_allowed))
 		goto out;
 
 	/*
@@ -728,7 +756,7 @@ static int generate_sched_domains(cpumask_var_t **domains,
 	int nslot;		/* next empty doms[] struct cpumask slot */
 	struct cgroup_subsys_state *pos_css;
 	bool root_load_balance = is_sched_load_balance(&top_cpuset);
-	bool cgrpv2 = cgroup_subsys_on_dfl(cpuset_cgrp_subsys);
+	bool cgrpv2 = cpuset_v2();
 	int nslot_update;
 
 	doms = NULL;
@@ -990,6 +1018,7 @@ void rebuild_sched_domains_locked(void)
 
 	lockdep_assert_cpus_held();
 	lockdep_assert_held(&cpuset_mutex);
+	force_sd_rebuild = false;
 
 	/*
 	 * If we have raced with CPU hotplug, return early to avoid
@@ -1164,8 +1193,8 @@ static void update_partition_sd_lb(struct cpuset *cs, int old_prs)
 			clear_bit(CS_SCHED_LOAD_BALANCE, &cs->flags);
 	}
 
-	if (rebuild_domains && !force_sd_rebuild)
+	if (rebuild_domains)
-		rebuild_sched_domains_locked();
+		cpuset_force_rebuild();
 }
 
 /*
@@ -1187,7 +1216,7 @@ static void reset_partition_data(struct cpuset *cs)
 {
 	struct cpuset *parent = parent_cs(cs);
 
-	if (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys))
+	if (!cpuset_v2())
 		return;
 
 	lockdep_assert_held(&callback_lock);
@@ -1339,7 +1368,7 @@ static inline bool is_local_partition(struct cpuset *cs)
  * remote_partition_enable - Enable current cpuset as a remote partition root
  * @cs: the cpuset to update
  * @new_prs: new partition_root_state
- * @tmp: temparary masks
+ * @tmp: temporary masks
  * Return: 0 if successful, errcode if error
  *
  * Enable the current cpuset to become a remote partition root taking CPUs
@@ -1377,7 +1406,7 @@ static int remote_partition_enable(struct cpuset *cs, int new_prs,
 	update_unbound_workqueue_cpumask(isolcpus_updated);
 
 	/*
-	 * Proprogate changes in top_cpuset's effective_cpus down the hierarchy.
+	 * Propagate changes in top_cpuset's effective_cpus down the hierarchy.
 	 */
 	cpuset_update_tasks_cpumask(&top_cpuset, tmp->new_cpus);
 	update_sibling_cpumasks(&top_cpuset, NULL, tmp);
@@ -1387,7 +1416,7 @@ static int remote_partition_enable(struct cpuset *cs, int new_prs,
 /*
  * remote_partition_disable - Remove current cpuset from remote partition list
  * @cs: the cpuset to update
- * @tmp: temparary masks
+ * @tmp: temporary masks
  *
  * The effective_cpus is also updated.
  *
@@ -1413,7 +1442,7 @@ static void remote_partition_disable(struct cpuset *cs, struct tmpmasks *tmp)
 	update_unbound_workqueue_cpumask(isolcpus_updated);
 
 	/*
-	 * Proprogate changes in top_cpuset's effective_cpus down the hierarchy.
+	 * Propagate changes in top_cpuset's effective_cpus down the hierarchy.
 	 */
 	cpuset_update_tasks_cpumask(&top_cpuset, tmp->new_cpus);
 	update_sibling_cpumasks(&top_cpuset, NULL, tmp);
@@ -1423,7 +1452,7 @@ static void remote_partition_disable(struct cpuset *cs, struct tmpmasks *tmp)
  * remote_cpus_update - cpus_exclusive change of remote partition
  * @cs: the cpuset to be updated
  * @newmask: the new effective_xcpus mask
- * @tmp: temparary masks
+ * @tmp: temporary masks
  *
  * top_cpuset and subpartitions_cpus will be updated or partition can be
  * invalidated.
@@ -1465,7 +1494,7 @@ static void remote_cpus_update(struct cpuset *cs, struct cpumask *newmask,
 	update_unbound_workqueue_cpumask(isolcpus_updated);
 
 	/*
-	 * Proprogate changes in top_cpuset's effective_cpus down the hierarchy.
+	 * Propagate changes in top_cpuset's effective_cpus down the hierarchy.
 	 */
 	cpuset_update_tasks_cpumask(&top_cpuset, tmp->new_cpus);
 	update_sibling_cpumasks(&top_cpuset, NULL, tmp);
@@ -1480,7 +1509,7 @@ static void remote_cpus_update(struct cpuset *cs, struct cpumask *newmask,
  * @cs: the cpuset to be updated
  * @newmask: the new effective_xcpus mask
  * @delmask: temporary mask for deletion (not in tmp)
- * @tmp: temparary masks
+ * @tmp: temporary masks
  *
  * This should be called before the given cs has updated its cpus_allowed
  * and/or effective_xcpus.
@@ -1512,8 +1541,8 @@ static void remote_partition_check(struct cpuset *cs, struct cpumask *newmask,
 			remote_partition_disable(child, tmp);
 			disable_cnt++;
 		}
-	if (disable_cnt && !force_sd_rebuild)
+	if (disable_cnt)
-		rebuild_sched_domains_locked();
+		cpuset_force_rebuild();
 }
 
 /*
@@ -1922,12 +1951,6 @@ static void compute_partition_effective_cpumask(struct cpuset *cs,
 	rcu_read_unlock();
 }
 
-/*
- * update_cpumasks_hier() flags
- */
-#define HIER_CHECKALL		0x01	/* Check all cpusets with no skipping */
-#define HIER_NO_SD_REBUILD	0x02	/* Don't rebuild sched domains */
-
 /*
  * update_cpumasks_hier - Update effective cpumasks and tasks in the subtree
  * @cs:  the cpuset to consider
@@ -1942,7 +1965,7 @@ static void compute_partition_effective_cpumask(struct cpuset *cs,
  * Called with cpuset_mutex held
  */
 static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp,
-				 int flags)
+				 bool force)
 {
 	struct cpuset *cp;
 	struct cgroup_subsys_state *pos_css;
@@ -2007,12 +2030,12 @@ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp,
 		 * Skip the whole subtree if
 		 * 1) the cpumask remains the same,
 		 * 2) has no partition root state,
-		 * 3) HIER_CHECKALL flag not set, and
+		 * 3) force flag not set, and
 		 * 4) for v2 load balance state same as its parent.
 		 */
-		if (!cp->partition_root_state && !(flags & HIER_CHECKALL) &&
+		if (!cp->partition_root_state && !force &&
 		    cpumask_equal(tmp->new_cpus, cp->effective_cpus) &&
-		    (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys) ||
+		    (!cpuset_v2() ||
 		    (is_sched_load_balance(parent) == is_sched_load_balance(cp)))) {
 			pos_css = css_rightmost_descendant(pos_css);
 			continue;
@@ -2086,8 +2109,7 @@ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp,
 		 * from parent if current cpuset isn't a valid partition root
 		 * and their load balance states differ.
 		 */
-		if (cgroup_subsys_on_dfl(cpuset_cgrp_subsys) &&
+		if (cpuset_v2() && !is_partition_valid(cp) &&
-		    !is_partition_valid(cp) &&
 		    (is_sched_load_balance(parent) != is_sched_load_balance(cp))) {
 			if (is_sched_load_balance(parent))
 				set_bit(CS_SCHED_LOAD_BALANCE, &cp->flags);
@@ -2103,8 +2125,7 @@ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp,
 		 */
 		if (!cpumask_empty(cp->cpus_allowed) &&
 		    is_sched_load_balance(cp) &&
-		   (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys) ||
+		   (!cpuset_v2() || is_partition_valid(cp)))
-		    is_partition_valid(cp)))
 			need_rebuild_sched_domains = true;
 
 		rcu_read_lock();
@@ -2112,9 +2133,8 @@ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp,
 	}
 	rcu_read_unlock();
 
-	if (need_rebuild_sched_domains && !(flags & HIER_NO_SD_REBUILD) &&
+	if (need_rebuild_sched_domains)
-	    !force_sd_rebuild)
+		cpuset_force_rebuild();
-		rebuild_sched_domains_locked();
 }
 
 /**
@@ -2141,9 +2161,7 @@ static void update_sibling_cpumasks(struct cpuset *parent, struct cpuset *cs,
 	 * directly.
 	 *
 	 * The update_cpumasks_hier() function may sleep. So we have to
-	 * release the RCU read lock before calling it. HIER_NO_SD_REBUILD
+	 * release the RCU read lock before calling it.
-	 * flag is used to suppress rebuild of sched domains as the callers
-	 * will take care of that.
 	 */
 	rcu_read_lock();
 	cpuset_for_each_child(sibling, pos_css, parent) {
@@ -2159,7 +2177,7 @@ static void update_sibling_cpumasks(struct cpuset *parent, struct cpuset *cs,
 			continue;
 
 		rcu_read_unlock();
-		update_cpumasks_hier(sibling, tmp, HIER_NO_SD_REBUILD);
+		update_cpumasks_hier(sibling, tmp, false);
 		rcu_read_lock();
 		css_put(&sibling->css);
 	}
@@ -2179,7 +2197,7 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs,
 	struct tmpmasks tmp;
 	struct cpuset *parent = parent_cs(cs);
 	bool invalidate = false;
-	int hier_flags = 0;
+	bool force = false;
 	int old_prs = cs->partition_root_state;
 
 	/* top_cpuset.cpus_allowed tracks cpu_online_mask; it's read-only */
@@ -2206,7 +2224,7 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs,
 			return -EINVAL;
 
 		/*
-		 * When exclusive_cpus isn't explicitly set, it is constrainted
+		 * When exclusive_cpus isn't explicitly set, it is constrained
 		 * by cpus_allowed and parent's effective_xcpus. Otherwise,
 		 * trialcs->effective_xcpus is used as a temporary cpumask
 		 * for checking validity of the partition root.
@@ -2240,12 +2258,11 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs,
 	 * Check all the descendants in update_cpumasks_hier() if
 	 * effective_xcpus is to be changed.
 	 */
-	if (!cpumask_equal(cs->effective_xcpus, trialcs->effective_xcpus))
+	force = !cpumask_equal(cs->effective_xcpus, trialcs->effective_xcpus);
-		hier_flags = HIER_CHECKALL;
 
 	retval = validate_change(cs, trialcs);
 
-	if ((retval == -EINVAL) && cgroup_subsys_on_dfl(cpuset_cgrp_subsys)) {
+	if ((retval == -EINVAL) && cpuset_v2()) {
 		struct cgroup_subsys_state *css;
 		struct cpuset *cp;
 
@@ -2309,7 +2326,7 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs,
 	spin_unlock_irq(&callback_lock);
 
 	/* effective_cpus/effective_xcpus will be updated here */
-	update_cpumasks_hier(cs, &tmp, hier_flags);
+	update_cpumasks_hier(cs, &tmp, force);
 
 	/* Update CS_SCHED_LOAD_BALANCE and/or sched_domains, if necessary */
 	if (cs->partition_root_state)
@@ -2334,7 +2351,7 @@ static int update_exclusive_cpumask(struct cpuset *cs, struct cpuset *trialcs,
 	struct tmpmasks tmp;
 	struct cpuset *parent = parent_cs(cs);
 	bool invalidate = false;
-	int hier_flags = 0;
+	bool force = false;
 	int old_prs = cs->partition_root_state;
 
 	if (!*buf) {
@@ -2357,8 +2374,7 @@ static int update_exclusive_cpumask(struct cpuset *cs, struct cpuset *trialcs,
 	 * Check all the descendants in update_cpumasks_hier() if
 	 * effective_xcpus is to be changed.
 	 */
-	if (!cpumask_equal(cs->effective_xcpus, trialcs->effective_xcpus))
+	force = !cpumask_equal(cs->effective_xcpus, trialcs->effective_xcpus);
-		hier_flags = HIER_CHECKALL;
 
 	retval = validate_change(cs, trialcs);
 	if (retval)
@@ -2411,8 +2427,8 @@ static int update_exclusive_cpumask(struct cpuset *cs, struct cpuset *trialcs,
 	 * of the subtree when it is a valid partition root or effective_xcpus
 	 * is updated.
 	 */
-	if (is_partition_valid(cs) || hier_flags)
+	if (is_partition_valid(cs) || force)
-		update_cpumasks_hier(cs, &tmp, hier_flags);
+		update_cpumasks_hier(cs, &tmp, force);
 
 	/* Update CS_SCHED_LOAD_BALANCE and/or sched_domains, if necessary */
 	if (cs->partition_root_state)
@@ -2737,9 +2753,12 @@ int cpuset_update_flag(cpuset_flagbits_t bit, struct cpuset *cs,
 	cs->flags = trialcs->flags;
 	spin_unlock_irq(&callback_lock);
 
-	if (!cpumask_empty(trialcs->cpus_allowed) && balance_flag_changed &&
+	if (!cpumask_empty(trialcs->cpus_allowed) && balance_flag_changed) {
-	    !force_sd_rebuild)
+		if (cpuset_v2())
-		rebuild_sched_domains_locked();
+			cpuset_force_rebuild();
+		else
+			rebuild_sched_domains_locked();
+	}
 
 	if (spread_flag_changed)
 		cpuset1_update_tasks_flags(cs);
@@ -2853,12 +2872,14 @@ static int update_prstate(struct cpuset *cs, int new_prs)
 	update_unbound_workqueue_cpumask(new_xcpus_state);
 
 	/* Force update if switching back to member */
-	update_cpumasks_hier(cs, &tmpmask, !new_prs ? HIER_CHECKALL : 0);
+	update_cpumasks_hier(cs, &tmpmask, !new_prs);
 
 	/* Update sched domains and load balance flag */
 	update_partition_sd_lb(cs, old_prs);
 
 	notify_partition_change(cs, old_prs);
+	if (force_sd_rebuild)
+		rebuild_sched_domains_locked();
 	free_cpumasks(NULL, &tmpmask);
 	return 0;
 }
@@ -2919,8 +2940,7 @@ static int cpuset_can_attach(struct cgroup_taskset *tset)
 		 * migration permission derives from hierarchy ownership in
 		 * cgroup_procs_write_permission()).
 		 */
-		if (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys) ||
+		if (!cpuset_v2() || (cpus_updated || mems_updated)) {
-		    (cpus_updated || mems_updated)) {
 			ret = security_task_setscheduler(task);
 			if (ret)
 				goto out_unlock;
@@ -3034,8 +3054,7 @@ static void cpuset_attach(struct cgroup_taskset *tset)
 	 * in effective cpus and mems. In that case, we can optimize out
 	 * by skipping the task iteration and update.
 	 */
-	if (cgroup_subsys_on_dfl(cpuset_cgrp_subsys) &&
+	if (cpuset_v2() && !cpus_updated && !mems_updated) {
-	    !cpus_updated && !mems_updated) {
 		cpuset_attach_nodemask_to = cs->effective_mems;
 		goto out;
 	}
@@ -3152,6 +3171,8 @@ ssize_t cpuset_write_resmask(struct kernfs_open_file *of,
 	}
 
 	free_cpuset(trialcs);
+	if (force_sd_rebuild)
+		rebuild_sched_domains_locked();
 out_unlock:
 	mutex_unlock(&cpuset_mutex);
 	cpus_read_unlock();
@@ -3383,7 +3404,7 @@ cpuset_css_alloc(struct cgroup_subsys_state *parent_css)
 	INIT_LIST_HEAD(&cs->remote_sibling);
 
 	/* Set CS_MEMORY_MIGRATE for default hierarchy */
-	if (cgroup_subsys_on_dfl(cpuset_cgrp_subsys))
+	if (cpuset_v2())
 		__set_bit(CS_MEMORY_MIGRATE, &cs->flags);
 
 	return &cs->css;
@@ -3410,8 +3431,7 @@ static int cpuset_css_online(struct cgroup_subsys_state *css)
 	/*
 	 * For v2, clear CS_SCHED_LOAD_BALANCE if parent is isolated
 	 */
-	if (cgroup_subsys_on_dfl(cpuset_cgrp_subsys) &&
+	if (cpuset_v2() && !is_sched_load_balance(parent))
-	    !is_sched_load_balance(parent))
 		clear_bit(CS_SCHED_LOAD_BALANCE, &cs->flags);
 
 	cpuset_inc();
@@ -3481,8 +3501,7 @@ static void cpuset_css_offline(struct cgroup_subsys_state *css)
 	if (is_partition_valid(cs))
 		update_prstate(cs, 0);
 
-	if (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys) &&
+	if (!cpuset_v2() && is_sched_load_balance(cs))
-	    is_sched_load_balance(cs))
 		cpuset_update_flag(CS_SCHED_LOAD_BALANCE, cs, 0);
 
 	cpuset_dec();
@@ -3896,11 +3915,9 @@ static void cpuset_handle_hotplug(void)
 		rcu_read_unlock();
 	}
 
-	/* rebuild sched domains if cpus_allowed has changed */
+	/* rebuild sched domains if necessary */
-	if (force_sd_rebuild) {
+	if (force_sd_rebuild)
-		force_sd_rebuild = false;
 		rebuild_sched_domains_cpuslocked();
-	}
 
 	free_cpumasks(NULL, ptmp);
 }

kernel/cgroup/freezer.c

@@ -8,6 +8,28 @@
 
 #include <trace/events/cgroup.h>
 
+/*
+ * Update CGRP_FROZEN of cgroup.flag
+ * Return true if flags is updated; false if flags has no change
+ */
+static bool cgroup_update_frozen_flag(struct cgroup *cgrp, bool frozen)
+{
+	lockdep_assert_held(&css_set_lock);
+
+	/* Already there? */
+	if (test_bit(CGRP_FROZEN, &cgrp->flags) == frozen)
+		return false;
+
+	if (frozen)
+		set_bit(CGRP_FROZEN, &cgrp->flags);
+	else
+		clear_bit(CGRP_FROZEN, &cgrp->flags);
+
+	cgroup_file_notify(&cgrp->events_file);
+	TRACE_CGROUP_PATH(notify_frozen, cgrp, frozen);
+	return true;
+}
+
 /*
  * Propagate the cgroup frozen state upwards by the cgroup tree.
  */
@@ -24,24 +46,16 @@ static void cgroup_propagate_frozen(struct cgroup *cgrp, bool frozen)
 	while ((cgrp = cgroup_parent(cgrp))) {
 		if (frozen) {
 			cgrp->freezer.nr_frozen_descendants += desc;
-			if (!test_bit(CGRP_FROZEN, &cgrp->flags) &&
+			if (!test_bit(CGRP_FREEZE, &cgrp->flags) ||
-			    test_bit(CGRP_FREEZE, &cgrp->flags) &&
+			    (cgrp->freezer.nr_frozen_descendants !=
-			    cgrp->freezer.nr_frozen_descendants ==
+			    cgrp->nr_descendants))
-			    cgrp->nr_descendants) {
+				continue;
-				set_bit(CGRP_FROZEN, &cgrp->flags);
-				cgroup_file_notify(&cgrp->events_file);
-				TRACE_CGROUP_PATH(notify_frozen, cgrp, 1);
-				desc++;
-			}
 		} else {
 			cgrp->freezer.nr_frozen_descendants -= desc;
-			if (test_bit(CGRP_FROZEN, &cgrp->flags)) {
-				clear_bit(CGRP_FROZEN, &cgrp->flags);
-				cgroup_file_notify(&cgrp->events_file);
-				TRACE_CGROUP_PATH(notify_frozen, cgrp, 0);
-				desc++;
-			}
 		}
+
+		if (cgroup_update_frozen_flag(cgrp, frozen))
+			desc++;
 	}
 }
 
@@ -53,8 +67,6 @@ void cgroup_update_frozen(struct cgroup *cgrp)
 {
 	bool frozen;
 
-	lockdep_assert_held(&css_set_lock);
-
 	/*
 	 * If the cgroup has to be frozen (CGRP_FREEZE bit set),
 	 * and all tasks are frozen and/or stopped, let's consider
@@ -63,24 +75,9 @@ void cgroup_update_frozen(struct cgroup *cgrp)
 	frozen = test_bit(CGRP_FREEZE, &cgrp->flags) &&
 		cgrp->freezer.nr_frozen_tasks == __cgroup_task_count(cgrp);
 
-	if (frozen) {
+	/* If flags is updated, update the state of ancestor cgroups. */
-		/* Already there? */
+	if (cgroup_update_frozen_flag(cgrp, frozen))
-		if (test_bit(CGRP_FROZEN, &cgrp->flags))
+		cgroup_propagate_frozen(cgrp, frozen);
-			return;
-
-		set_bit(CGRP_FROZEN, &cgrp->flags);
-	} else {
-		/* Already there? */
-		if (!test_bit(CGRP_FROZEN, &cgrp->flags))
-			return;
-
-		clear_bit(CGRP_FROZEN, &cgrp->flags);
-	}
-	cgroup_file_notify(&cgrp->events_file);
-	TRACE_CGROUP_PATH(notify_frozen, cgrp, frozen);
-
-	/* Update the state of ancestor cgroups. */
-	cgroup_propagate_frozen(cgrp, frozen);
 }
 
 /*
@@ -260,8 +257,10 @@ void cgroup_freezer_migrate_task(struct task_struct *task,
 void cgroup_freeze(struct cgroup *cgrp, bool freeze)
 {
 	struct cgroup_subsys_state *css;
+	struct cgroup *parent;
 	struct cgroup *dsct;
 	bool applied = false;
+	bool old_e;
 
 	lockdep_assert_held(&cgroup_mutex);
 
@@ -282,22 +281,18 @@ void cgroup_freeze(struct cgroup *cgrp, bool freeze)
 		if (cgroup_is_dead(dsct))
 			continue;
 
-		if (freeze) {
+		/*
-			dsct->freezer.e_freeze++;
+		 * e_freeze is affected by parent's e_freeze and dst's freeze.
-			/*
+		 * If old e_freeze eq new e_freeze, no change, its children
-			 * Already frozen because of ancestor's settings?
+		 * will not be affected. So do nothing and skip the subtree
-			 */
+		 */
-			if (dsct->freezer.e_freeze > 1)
+		old_e = dsct->freezer.e_freeze;
-				continue;
+		parent = cgroup_parent(dsct);
-		} else {
+		dsct->freezer.e_freeze = (dsct->freezer.freeze ||
-			dsct->freezer.e_freeze--;
+					  parent->freezer.e_freeze);
-			/*
+		if (dsct->freezer.e_freeze == old_e) {
-			 * Still frozen because of ancestor's settings?
+			css = css_rightmost_descendant(css);
-			 */
+			continue;
-			if (dsct->freezer.e_freeze > 0)
-				continue;
-
-			WARN_ON_ONCE(dsct->freezer.e_freeze < 0);
 		}
 
 		/*

kernel/cgroup/rstat.c

@@ -444,6 +444,7 @@ static void cgroup_base_stat_add(struct cgroup_base_stat *dst_bstat,
 #ifdef CONFIG_SCHED_CORE
 	dst_bstat->forceidle_sum += src_bstat->forceidle_sum;
 #endif
+	dst_bstat->ntime += src_bstat->ntime;
 }
 
 static void cgroup_base_stat_sub(struct cgroup_base_stat *dst_bstat,
@@ -455,6 +456,7 @@ static void cgroup_base_stat_sub(struct cgroup_base_stat *dst_bstat,
 #ifdef CONFIG_SCHED_CORE
 	dst_bstat->forceidle_sum -= src_bstat->forceidle_sum;
 #endif
+	dst_bstat->ntime -= src_bstat->ntime;
 }
 
 static void cgroup_base_stat_flush(struct cgroup *cgrp, int cpu)
@@ -534,8 +536,10 @@ void __cgroup_account_cputime_field(struct cgroup *cgrp,
 	rstatc = cgroup_base_stat_cputime_account_begin(cgrp, &flags);
 
 	switch (index) {
-	case CPUTIME_USER:
 	case CPUTIME_NICE:
+		rstatc->bstat.ntime += delta_exec;
+		fallthrough;
+	case CPUTIME_USER:
 		rstatc->bstat.cputime.utime += delta_exec;
 		break;
 	case CPUTIME_SYSTEM:
@@ -591,6 +595,7 @@ static void root_cgroup_cputime(struct cgroup_base_stat *bstat)
 #ifdef CONFIG_SCHED_CORE
 		bstat->forceidle_sum += cpustat[CPUTIME_FORCEIDLE];
 #endif
+		bstat->ntime += cpustat[CPUTIME_NICE];
 	}
 }
 
@@ -608,13 +613,14 @@ static void cgroup_force_idle_show(struct seq_file *seq, struct cgroup_base_stat
 void cgroup_base_stat_cputime_show(struct seq_file *seq)
 {
 	struct cgroup *cgrp = seq_css(seq)->cgroup;
-	u64 usage, utime, stime;
+	u64 usage, utime, stime, ntime;
 
 	if (cgroup_parent(cgrp)) {
 		cgroup_rstat_flush_hold(cgrp);
 		usage = cgrp->bstat.cputime.sum_exec_runtime;
 		cputime_adjust(&cgrp->bstat.cputime, &cgrp->prev_cputime,
 			       &utime, &stime);
+		ntime = cgrp->bstat.ntime;
 		cgroup_rstat_flush_release(cgrp);
 	} else {
 		/* cgrp->bstat of root is not actually used, reuse it */
@@ -622,16 +628,19 @@ void cgroup_base_stat_cputime_show(struct seq_file *seq)
 		usage = cgrp->bstat.cputime.sum_exec_runtime;
 		utime = cgrp->bstat.cputime.utime;
 		stime = cgrp->bstat.cputime.stime;
+		ntime = cgrp->bstat.ntime;
 	}
 
 	do_div(usage, NSEC_PER_USEC);
 	do_div(utime, NSEC_PER_USEC);
 	do_div(stime, NSEC_PER_USEC);
+	do_div(ntime, NSEC_PER_USEC);
 
 	seq_printf(seq, "usage_usec %llu\n"
-		   "user_usec %llu\n"
+			"user_usec %llu\n"
-		   "system_usec %llu\n",
+			"system_usec %llu\n"
-		   usage, utime, stime);
+			"nice_usec %llu\n",
+			usage, utime, stime, ntime);
 
 	cgroup_force_idle_show(seq, &cgrp->bstat);
 }

tools/testing/selftests/cgroup/test_cpu.c

@@ -8,6 +8,7 @@
 #include <pthread.h>
 #include <stdio.h>
 #include <time.h>
+#include <unistd.h>
 
 #include "../kselftest.h"
 #include "cgroup_util.h"
@@ -229,6 +230,79 @@ static int test_cpucg_stats(const char *root)
 	return ret;
 }
 
+/*
+ * Creates a nice process that consumes CPU and checks that the elapsed
+ * usertime in the cgroup is close to the expected time.
+ */
+static int test_cpucg_nice(const char *root)
+{
+	int ret = KSFT_FAIL;
+	int status;
+	long user_usec, nice_usec;
+	long usage_seconds = 2;
+	long expected_nice_usec = usage_seconds * USEC_PER_SEC;
+	char *cpucg;
+	pid_t pid;
+
+	cpucg = cg_name(root, "cpucg_test");
+	if (!cpucg)
+		goto cleanup;
+
+	if (cg_create(cpucg))
+		goto cleanup;
+
+	user_usec = cg_read_key_long(cpucg, "cpu.stat", "user_usec");
+	nice_usec = cg_read_key_long(cpucg, "cpu.stat", "nice_usec");
+	if (nice_usec == -1)
+		ret = KSFT_SKIP;
+	if (user_usec != 0 || nice_usec != 0)
+		goto cleanup;
+
+	/*
+	 * We fork here to create a new process that can be niced without
+	 * polluting the nice value of other selftests
+	 */
+	pid = fork();
+	if (pid < 0) {
+		goto cleanup;
+	} else if (pid == 0) {
+		struct cpu_hog_func_param param = {
+			.nprocs = 1,
+			.ts = {
+				.tv_sec = usage_seconds,
+				.tv_nsec = 0,
+			},
+			.clock_type = CPU_HOG_CLOCK_PROCESS,
+		};
+		char buf[64];
+		snprintf(buf, sizeof(buf), "%d", getpid());
+		if (cg_write(cpucg, "cgroup.procs", buf))
+			goto cleanup;
+
+		/* Try to keep niced CPU usage as constrained to hog_cpu as possible */
+		nice(1);
+		hog_cpus_timed(cpucg, &param);
+		exit(0);
+	} else {
+		waitpid(pid, &status, 0);
+		if (!WIFEXITED(status))
+			goto cleanup;
+
+		user_usec = cg_read_key_long(cpucg, "cpu.stat", "user_usec");
+		nice_usec = cg_read_key_long(cpucg, "cpu.stat", "nice_usec");
+		if (!values_close(nice_usec, expected_nice_usec, 1))
+			goto cleanup;
+
+		ret = KSFT_PASS;
+	}
+
+cleanup:
+	cg_destroy(cpucg);
+	free(cpucg);
+
+	return ret;
+}
+
 static int
 run_cpucg_weight_test(
 		const char *root,
@@ -686,6 +760,7 @@ struct cpucg_test {
 } tests[] = {
 	T(test_cpucg_subtree_control),
 	T(test_cpucg_stats),
+	T(test_cpucg_nice),
 	T(test_cpucg_weight_overprovisioned),
 	T(test_cpucg_weight_underprovisioned),
 	T(test_cpucg_nested_weight_overprovisioned),