Merge branch 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull scheduler updates from Ingo Molnar:
 "The main changes in this cycle were:

   - fix affine wakeups (Peter Zijlstra)

   - improve CPU onlining (and general bootup) scalability on systems
     with ridiculous number (thousands) of CPUs (Peter Zijlstra)

   - sched/numa updates (Rik van Riel)

   - sched/deadline updates (Byungchul Park)

   - sched/cpufreq enhancements and related cleanups (Viresh Kumar)

   - sched/debug enhancements (Xie XiuQi)

   - various fixes"

* 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (27 commits)
  sched/debug: Optimize sched_domain sysctl generation
  sched/topology: Avoid pointless rebuild
  sched/topology, cpuset: Avoid spurious/wrong domain rebuilds
  sched/topology: Improve comments
  sched/topology: Fix memory leak in __sdt_alloc()
  sched/completion: Document that reinit_completion() must be called after complete_all()
  sched/autogroup: Fix error reporting printk text in autogroup_create()
  sched/fair: Fix wake_affine() for !NUMA_BALANCING
  sched/debug: Intruduce task_state_to_char() helper function
  sched/debug: Show task state in /proc/sched_debug
  sched/debug: Use task_pid_nr_ns in /proc/$pid/sched
  sched/core: Remove unnecessary initialization init_idle_bootup_task()
  sched/deadline: Change return value of cpudl_find()
  sched/deadline: Make find_later_rq() choose a closer CPU in topology
  sched/numa: Scale scan period with tasks in group and shared/private
  sched/numa: Slow down scan rate if shared faults dominate
  sched/pelt: Fix false running accounting
  sched: Mark pick_next_task_dl() and build_sched_domain() as static
  sched/cpupri: Don't re-initialize 'struct cpupri'
  sched/deadline: Don't re-initialize 'struct cpudl'
  ...

arch/x86/include/asm/topology.h

@@ -75,12 +75,6 @@ static inline const struct cpumask *cpumask_of_node(int node)
 
 extern void setup_node_to_cpumask_map(void);
 
-/*
- * Returns the number of the node containing Node 'node'. This
- * architecture is flat, so it is a pretty simple function!
- */
-#define parent_node(node) (node)
-
 #define pcibus_to_node(bus) __pcibus_to_node(bus)
 
 extern int __node_distance(int, int);

fs/proc/base.c

@@ -1408,12 +1408,13 @@ static const struct file_operations proc_fail_nth_operations = {
 static int sched_show(struct seq_file *m, void *v)
 {
 	struct inode *inode = m->private;
+	struct pid_namespace *ns = inode->i_sb->s_fs_info;
 	struct task_struct *p;
 
 	p = get_proc_task(inode);
 	if (!p)
 		return -ESRCH;
-	proc_sched_show_task(p, m);
+	proc_sched_show_task(p, ns, m);
 
 	put_task_struct(p);
 

include/linux/sched.h

@@ -1233,6 +1233,19 @@ static inline pid_t task_pgrp_nr(struct task_struct *tsk)
 	return task_pgrp_nr_ns(tsk, &init_pid_ns);
 }
 
+static inline char task_state_to_char(struct task_struct *task)
+{
+	const char stat_nam[] = TASK_STATE_TO_CHAR_STR;
+	unsigned long state = task->state;
+
+	state = state ? __ffs(state) + 1 : 0;
+
+	/* Make sure the string lines up properly with the number of task states: */
+	BUILD_BUG_ON(sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1);
+
+	return state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?';
+}
+
 /**
  * is_global_init - check if a task structure is init. Since init
  * is free to have sub-threads we need to check tgid.

include/linux/sched/debug.h

@@ -6,6 +6,7 @@
  */
 
 struct task_struct;
+struct pid_namespace;
 
 extern void dump_cpu_task(int cpu);
 
@@ -34,7 +35,8 @@ extern void sched_show_task(struct task_struct *p);
 
 #ifdef CONFIG_SCHED_DEBUG
 struct seq_file;
-extern void proc_sched_show_task(struct task_struct *p, struct seq_file *m);
+extern void proc_sched_show_task(struct task_struct *p,
+				 struct pid_namespace *ns, struct seq_file *m);
 extern void proc_sched_set_task(struct task_struct *p);
 #endif
 

include/linux/sched/task.h

@@ -30,7 +30,6 @@ extern int lockdep_tasklist_lock_is_held(void);
 
 extern asmlinkage void schedule_tail(struct task_struct *prev);
 extern void init_idle(struct task_struct *idle, int cpu);
-extern void init_idle_bootup_task(struct task_struct *idle);
 
 extern int sched_fork(unsigned long clone_flags, struct task_struct *p);
 extern void sched_dead(struct task_struct *p);

include/linux/sched/topology.h

@@ -71,6 +71,14 @@ struct sched_domain_shared {
 	atomic_t	ref;
 	atomic_t	nr_busy_cpus;
 	int		has_idle_cores;
+
+	/*
+	 * Some variables from the most recent sd_lb_stats for this domain,
+	 * used by wake_affine().
+	 */
+	unsigned long	nr_running;
+	unsigned long	load;
+	unsigned long	capacity;
 };
 
 struct sched_domain {

init/main.c

@@ -430,7 +430,6 @@ static noinline void __ref rest_init(void)
 	 * The boot idle thread must execute schedule()
 	 * at least once to get things moving:
 	 */
-	init_idle_bootup_task(current);
 	schedule_preempt_disabled();
 	/* Call into cpu_idle with preempt disabled */
 	cpu_startup_entry(CPUHP_ONLINE);

kernel/cgroup/cpuset.c

@@ -2344,13 +2344,7 @@ void cpuset_update_active_cpus(void)
 	 * We're inside cpu hotplug critical region which usually nests
 	 * inside cgroup synchronization.  Bounce actual hotplug processing
 	 * to a work item to avoid reverse locking order.
-	 *
-	 * We still need to do partition_sched_domains() synchronously;
-	 * otherwise, the scheduler will get confused and put tasks to the
-	 * dead CPU.  Fall back to the default single domain.
-	 * cpuset_hotplug_workfn() will rebuild it as necessary.
 	 */
-	partition_sched_domains(1, NULL, NULL);
 	schedule_work(&cpuset_hotplug_work);
 }
 

kernel/sched/autogroup.c

@@ -71,7 +71,6 @@ static inline struct autogroup *autogroup_create(void)
 		goto out_fail;
 
 	tg = sched_create_group(&root_task_group);
-
 	if (IS_ERR(tg))
 		goto out_free;
 
@@ -101,7 +100,7 @@ static inline struct autogroup *autogroup_create(void)
 out_fail:
 	if (printk_ratelimit()) {
 		printk(KERN_WARNING "autogroup_create: %s failure.\n",
-			ag ? "sched_create_group()" : "kmalloc()");
+			ag ? "sched_create_group()" : "kzalloc()");
 	}
 
 	return autogroup_kref_get(&autogroup_default);

kernel/sched/completion.c

@@ -47,6 +47,13 @@ EXPORT_SYMBOL(complete);
  *
  * It may be assumed that this function implies a write memory barrier before
  * changing the task state if and only if any tasks are woken up.
+ *
+ * Since complete_all() sets the completion of @x permanently to done
+ * to allow multiple waiters to finish, a call to reinit_completion()
+ * must be used on @x if @x is to be used again. The code must make
+ * sure that all waiters have woken and finished before reinitializing
+ * @x. Also note that the function completion_done() can not be used
+ * to know if there are still waiters after complete_all() has been called.
  */
 void complete_all(struct completion *x)
 {
@@ -297,6 +304,7 @@ EXPORT_SYMBOL(try_wait_for_completion);
  *	Return: 0 if there are waiters (wait_for_completion() in progress)
  *		 1 if there are no waiters.
  *
+ *	Note, this will always return true if complete_all() was called on @X.
  */
 bool completion_done(struct completion *x)
 {

kernel/sched/core.c

@@ -5133,24 +5133,17 @@ SYSCALL_DEFINE2(sched_rr_get_interval, pid_t, pid,
 	return retval;
 }
 
-static const char stat_nam[] = TASK_STATE_TO_CHAR_STR;
-
 void sched_show_task(struct task_struct *p)
 {
 	unsigned long free = 0;
 	int ppid;
-	unsigned long state = p->state;
-
-	/* Make sure the string lines up properly with the number of task states: */
-	BUILD_BUG_ON(sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1);
 
 	if (!try_get_task_stack(p))
 		return;
-	if (state)
-		state = __ffs(state) + 1;
-	printk(KERN_INFO "%-15.15s %c", p->comm,
-		state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?');
-	if (state == TASK_RUNNING)
+
+	printk(KERN_INFO "%-15.15s %c", p->comm, task_state_to_char(p));
+
+	if (p->state == TASK_RUNNING)
 		printk(KERN_CONT "  running task    ");
 #ifdef CONFIG_DEBUG_STACK_USAGE
 	free = stack_not_used(p);
@@ -5207,11 +5200,6 @@ void show_state_filter(unsigned long state_filter)
 		debug_show_all_locks();
 }
 
-void init_idle_bootup_task(struct task_struct *idle)
-{
-	idle->sched_class = &idle_sched_class;
-}
-
 /**
  * init_idle - set up an idle thread for a given CPU
  * @idle: task in question
@@ -5468,7 +5456,7 @@ static void migrate_tasks(struct rq *dead_rq, struct rq_flags *rf)
 		 */
 		next = pick_next_task(rq, &fake_task, rf);
 		BUG_ON(!next);
-		next->sched_class->put_prev_task(rq, next);
+		put_prev_task(rq, next);
 
 		/*
 		 * Rules for changing task_struct::cpus_allowed are holding

kernel/sched/cpudeadline.c

@@ -119,29 +119,29 @@ static inline int cpudl_maximum(struct cpudl *cp)
  * @p: the task
  * @later_mask: a mask to fill in with the selected CPUs (or NULL)
  *
- * Returns: int - best CPU (heap maximum if suitable)
+ * Returns: int - CPUs were found
  */
 int cpudl_find(struct cpudl *cp, struct task_struct *p,
 	       struct cpumask *later_mask)
 {
-	int best_cpu = -1;
 	const struct sched_dl_entity *dl_se = &p->dl;
 
 	if (later_mask &&
 	    cpumask_and(later_mask, cp->free_cpus, &p->cpus_allowed)) {
-		best_cpu = cpumask_any(later_mask);
-		goto out;
-	} else if (cpumask_test_cpu(cpudl_maximum(cp), &p->cpus_allowed) &&
-			dl_time_before(dl_se->deadline, cp->elements[0].dl)) {
-		best_cpu = cpudl_maximum(cp);
-		if (later_mask)
-			cpumask_set_cpu(best_cpu, later_mask);
-	}
+		return 1;
+	} else {
+		int best_cpu = cpudl_maximum(cp);
+		WARN_ON(best_cpu != -1 && !cpu_present(best_cpu));
 
-out:
-	WARN_ON(best_cpu != -1 && !cpu_present(best_cpu));
+		if (cpumask_test_cpu(best_cpu, &p->cpus_allowed) &&
+		    dl_time_before(dl_se->deadline, cp->elements[0].dl)) {
+			if (later_mask)
+				cpumask_set_cpu(best_cpu, later_mask);
 
-	return best_cpu;
+			return 1;
+		}
+	}
+	return 0;
 }
 
 /*
@@ -246,7 +246,6 @@ int cpudl_init(struct cpudl *cp)
 {
 	int i;
 
-	memset(cp, 0, sizeof(*cp));
 	raw_spin_lock_init(&cp->lock);
 	cp->size = 0;
 

kernel/sched/cpupri.c

@@ -209,8 +209,6 @@ int cpupri_init(struct cpupri *cp)
 {
 	int i;
 
-	memset(cp, 0, sizeof(*cp));
-
 	for (i = 0; i < CPUPRI_NR_PRIORITIES; i++) {
 		struct cpupri_vec *vec = &cp->pri_to_cpu[i];
 

kernel/sched/deadline.c

@@ -1594,7 +1594,7 @@ static void check_preempt_equal_dl(struct rq *rq, struct task_struct *p)
 	 * let's hope p can move out.
 	 */
 	if (rq->curr->nr_cpus_allowed == 1 ||
-	    cpudl_find(&rq->rd->cpudl, rq->curr, NULL) == -1)
+	    !cpudl_find(&rq->rd->cpudl, rq->curr, NULL))
 		return;
 
 	/*
@@ -1602,7 +1602,7 @@ static void check_preempt_equal_dl(struct rq *rq, struct task_struct *p)
 	 * see if it is pushed or pulled somewhere else.
 	 */
 	if (p->nr_cpus_allowed != 1 &&
-	    cpudl_find(&rq->rd->cpudl, p, NULL) != -1)
+	    cpudl_find(&rq->rd->cpudl, p, NULL))
 		return;
 
 	resched_curr(rq);
@@ -1655,7 +1655,7 @@ static struct sched_dl_entity *pick_next_dl_entity(struct rq *rq,
 	return rb_entry(left, struct sched_dl_entity, rb_node);
 }
 
-struct task_struct *
+static struct task_struct *
 pick_next_task_dl(struct rq *rq, struct task_struct *prev, struct rq_flags *rf)
 {
 	struct sched_dl_entity *dl_se;
@@ -1798,7 +1798,7 @@ static int find_later_rq(struct task_struct *task)
 	struct sched_domain *sd;
 	struct cpumask *later_mask = this_cpu_cpumask_var_ptr(local_cpu_mask_dl);
 	int this_cpu = smp_processor_id();
-	int best_cpu, cpu = task_cpu(task);
+	int cpu = task_cpu(task);
 
 	/* Make sure the mask is initialized first */
 	if (unlikely(!later_mask))
@@ -1811,17 +1811,14 @@ static int find_later_rq(struct task_struct *task)
 	 * We have to consider system topology and task affinity
 	 * first, then we can look for a suitable cpu.
 	 */
-	best_cpu = cpudl_find(&task_rq(task)->rd->cpudl,
-			task, later_mask);
-	if (best_cpu == -1)
+	if (!cpudl_find(&task_rq(task)->rd->cpudl, task, later_mask))
 		return -1;
 
 	/*
-	 * If we are here, some target has been found,
-	 * the most suitable of which is cached in best_cpu.
-	 * This is, among the runqueues where the current tasks
-	 * have later deadlines than the task's one, the rq
-	 * with the latest possible one.
+	 * If we are here, some targets have been found, including
+	 * the most suitable which is, among the runqueues where the
+	 * current tasks have later deadlines than the task's one, the
+	 * rq with the latest possible one.
 	 *
 	 * Now we check how well this matches with task's
 	 * affinity and system topology.
@@ -1841,6 +1838,7 @@ static int find_later_rq(struct task_struct *task)
 	rcu_read_lock();
 	for_each_domain(cpu, sd) {
 		if (sd->flags & SD_WAKE_AFFINE) {
+			int best_cpu;
 
 			/*
 			 * If possible, preempting this_cpu is
@@ -1852,12 +1850,15 @@ static int find_later_rq(struct task_struct *task)
 				return this_cpu;
 			}
 
+			best_cpu = cpumask_first_and(later_mask,
+							sched_domain_span(sd));
 			/*
-			 * Last chance: if best_cpu is valid and is
-			 * in the mask, that becomes our choice.
+			 * Last chance: if a cpu being in both later_mask
+			 * and current sd span is valid, that becomes our
+			 * choice. Of course, the latest possible cpu is
+			 * already under consideration through later_mask.
 			 */
-			if (best_cpu < nr_cpu_ids &&
-			    cpumask_test_cpu(best_cpu, sched_domain_span(sd))) {
+			if (best_cpu < nr_cpu_ids) {
 				rcu_read_unlock();
 				return best_cpu;
 			}

kernel/sched/debug.c

@@ -327,38 +327,78 @@ static struct ctl_table *sd_alloc_ctl_cpu_table(int cpu)
 	return table;
 }
 
+static cpumask_var_t sd_sysctl_cpus;
 static struct ctl_table_header *sd_sysctl_header;
+
 void register_sched_domain_sysctl(void)
 {
-	int i, cpu_num = num_possible_cpus();
-	struct ctl_table *entry = sd_alloc_ctl_entry(cpu_num + 1);
+	static struct ctl_table *cpu_entries;
+	static struct ctl_table **cpu_idx;
 	char buf[32];
+	int i;
 
-	WARN_ON(sd_ctl_dir[0].child);
-	sd_ctl_dir[0].child = entry;
+	if (!cpu_entries) {
+		cpu_entries = sd_alloc_ctl_entry(num_possible_cpus() + 1);
+		if (!cpu_entries)
+			return;
 
-	if (entry == NULL)
-		return;
+		WARN_ON(sd_ctl_dir[0].child);
+		sd_ctl_dir[0].child = cpu_entries;
+	}
 
-	for_each_possible_cpu(i) {
-		snprintf(buf, 32, "cpu%d", i);
-		entry->procname = kstrdup(buf, GFP_KERNEL);
-		entry->mode = 0555;
-		entry->child = sd_alloc_ctl_cpu_table(i);
-		entry++;
+	if (!cpu_idx) {
+		struct ctl_table *e = cpu_entries;
+
+		cpu_idx = kcalloc(nr_cpu_ids, sizeof(struct ctl_table*), GFP_KERNEL);
+		if (!cpu_idx)
+			return;
+
+		/* deal with sparse possible map */
+		for_each_possible_cpu(i) {
+			cpu_idx[i] = e;
+			e++;
+		}
+	}
+
+	if (!cpumask_available(sd_sysctl_cpus)) {
+		if (!alloc_cpumask_var(&sd_sysctl_cpus, GFP_KERNEL))
+			return;
+
+		/* init to possible to not have holes in @cpu_entries */
+		cpumask_copy(sd_sysctl_cpus, cpu_possible_mask);
+	}
+
+	for_each_cpu(i, sd_sysctl_cpus) {
+		struct ctl_table *e = cpu_idx[i];
+
+		if (e->child)
+			sd_free_ctl_entry(&e->child);
+
+		if (!e->procname) {
+			snprintf(buf, 32, "cpu%d", i);
+			e->procname = kstrdup(buf, GFP_KERNEL);
+		}
+		e->mode = 0555;
+		e->child = sd_alloc_ctl_cpu_table(i);
+
+		__cpumask_clear_cpu(i, sd_sysctl_cpus);
 	}
 
 	WARN_ON(sd_sysctl_header);
 	sd_sysctl_header = register_sysctl_table(sd_ctl_root);
 }
 
+void dirty_sched_domain_sysctl(int cpu)
+{
+	if (cpumask_available(sd_sysctl_cpus))
+		__cpumask_set_cpu(cpu, sd_sysctl_cpus);
+}
+
 /* may be called multiple times per register */
 void unregister_sched_domain_sysctl(void)
 {
 	unregister_sysctl_table(sd_sysctl_header);
 	sd_sysctl_header = NULL;
-	if (sd_ctl_dir[0].child)
-		sd_free_ctl_entry(&sd_ctl_dir[0].child);
 }
 #endif /* CONFIG_SYSCTL */
 #endif /* CONFIG_SMP */
@@ -421,13 +461,15 @@ static char *task_group_path(struct task_group *tg)
 }
 #endif
 
+static const char stat_nam[] = TASK_STATE_TO_CHAR_STR;
+
 static void
 print_task(struct seq_file *m, struct rq *rq, struct task_struct *p)
 {
 	if (rq->curr == p)
-		SEQ_printf(m, "R");
+		SEQ_printf(m, ">R");
 	else
-		SEQ_printf(m, " ");
+		SEQ_printf(m, " %c", task_state_to_char(p));
 
 	SEQ_printf(m, "%15s %5d %9Ld.%06ld %9Ld %5d ",
 		p->comm, task_pid_nr(p),
@@ -456,9 +498,9 @@ static void print_rq(struct seq_file *m, struct rq *rq, int rq_cpu)
 
 	SEQ_printf(m,
 	"\nrunnable tasks:\n"
-	"            task   PID         tree-key  switches  prio"
+	" S           task   PID         tree-key  switches  prio"
 	"     wait-time             sum-exec        sum-sleep\n"
-	"------------------------------------------------------"
+	"-------------------------------------------------------"
 	"----------------------------------------------------\n");
 
 	rcu_read_lock();
@@ -872,11 +914,12 @@ static void sched_show_numa(struct task_struct *p, struct seq_file *m)
 #endif
 }
 
-void proc_sched_show_task(struct task_struct *p, struct seq_file *m)
+void proc_sched_show_task(struct task_struct *p, struct pid_namespace *ns,
+						  struct seq_file *m)
 {
 	unsigned long nr_switches;
 
-	SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr(p),
+	SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr_ns(p, ns),
 						get_nr_threads(p));
 	SEQ_printf(m,
 		"---------------------------------------------------------"

kernel/sched/fair.c

@@ -806,7 +806,7 @@ void post_init_entity_util_avg(struct sched_entity *se)
 			/*
 			 * For !fair tasks do:
 			 *
-			update_cfs_rq_load_avg(now, cfs_rq, false);
+			update_cfs_rq_load_avg(now, cfs_rq);
 			attach_entity_load_avg(cfs_rq, se);
 			switched_from_fair(rq, p);
 			 *
@@ -1071,6 +1071,29 @@ unsigned int sysctl_numa_balancing_scan_size = 256;
 /* Scan @scan_size MB every @scan_period after an initial @scan_delay in ms */
 unsigned int sysctl_numa_balancing_scan_delay = 1000;
 
+struct numa_group {
+	atomic_t refcount;
+
+	spinlock_t lock; /* nr_tasks, tasks */
+	int nr_tasks;
+	pid_t gid;
+	int active_nodes;
+
+	struct rcu_head rcu;
+	unsigned long total_faults;
+	unsigned long max_faults_cpu;
+	/*
+	 * Faults_cpu is used to decide whether memory should move
+	 * towards the CPU. As a consequence, these stats are weighted
+	 * more by CPU use than by memory faults.
+	 */
+	unsigned long *faults_cpu;
+	unsigned long faults[0];
+};
+
+static inline unsigned long group_faults_priv(struct numa_group *ng);
+static inline unsigned long group_faults_shared(struct numa_group *ng);
+
 static unsigned int task_nr_scan_windows(struct task_struct *p)
 {
 	unsigned long rss = 0;
@@ -1107,13 +1130,47 @@ static unsigned int task_scan_min(struct task_struct *p)
 	return max_t(unsigned int, floor, scan);
 }
 
+static unsigned int task_scan_start(struct task_struct *p)
+{
+	unsigned long smin = task_scan_min(p);
+	unsigned long period = smin;
+
+	/* Scale the maximum scan period with the amount of shared memory. */
+	if (p->numa_group) {
+		struct numa_group *ng = p->numa_group;
+		unsigned long shared = group_faults_shared(ng);
+		unsigned long private = group_faults_priv(ng);
+
+		period *= atomic_read(&ng->refcount);
+		period *= shared + 1;
+		period /= private + shared + 1;
+	}
+
+	return max(smin, period);
+}
+
 static unsigned int task_scan_max(struct task_struct *p)
 {
-	unsigned int smin = task_scan_min(p);
-	unsigned int smax;
+	unsigned long smin = task_scan_min(p);
+	unsigned long smax;
 
 	/* Watch for min being lower than max due to floor calculations */
 	smax = sysctl_numa_balancing_scan_period_max / task_nr_scan_windows(p);
+
+	/* Scale the maximum scan period with the amount of shared memory. */
+	if (p->numa_group) {
+		struct numa_group *ng = p->numa_group;
+		unsigned long shared = group_faults_shared(ng);
+		unsigned long private = group_faults_priv(ng);
+		unsigned long period = smax;
+
+		period *= atomic_read(&ng->refcount);
+		period *= shared + 1;
+		period /= private + shared + 1;
+
+		smax = max(smax, period);
+	}
+
 	return max(smin, smax);
 }
 
@@ -1129,26 +1186,6 @@ static void account_numa_dequeue(struct rq *rq, struct task_struct *p)
 	rq->nr_preferred_running -= (p->numa_preferred_nid == task_node(p));
 }
 
-struct numa_group {
-	atomic_t refcount;
-
-	spinlock_t lock; /* nr_tasks, tasks */
-	int nr_tasks;
-	pid_t gid;
-	int active_nodes;
-
-	struct rcu_head rcu;
-	unsigned long total_faults;
-	unsigned long max_faults_cpu;
-	/*
-	 * Faults_cpu is used to decide whether memory should move
-	 * towards the CPU. As a consequence, these stats are weighted
-	 * more by CPU use than by memory faults.
-	 */
-	unsigned long *faults_cpu;
-	unsigned long faults[0];
-};
-
 /* Shared or private faults. */
 #define NR_NUMA_HINT_FAULT_TYPES 2
 
@@ -1198,6 +1235,30 @@ static inline unsigned long group_faults_cpu(struct numa_group *group, int nid)
 		group->faults_cpu[task_faults_idx(NUMA_MEM, nid, 1)];
 }
 
+static inline unsigned long group_faults_priv(struct numa_group *ng)
+{
+	unsigned long faults = 0;
+	int node;
+
+	for_each_online_node(node) {
+		faults += ng->faults[task_faults_idx(NUMA_MEM, node, 1)];
+	}
+
+	return faults;
+}
+
+static inline unsigned long group_faults_shared(struct numa_group *ng)
+{
+	unsigned long faults = 0;
+	int node;
+
+	for_each_online_node(node) {
+		faults += ng->faults[task_faults_idx(NUMA_MEM, node, 0)];
+	}
+
+	return faults;
+}
+
 /*
  * A node triggering more than 1/3 as many NUMA faults as the maximum is
  * considered part of a numa group's pseudo-interleaving set. Migrations
@@ -1378,7 +1439,7 @@ bool should_numa_migrate_memory(struct task_struct *p, struct page * page,
 	       group_faults_cpu(ng, src_nid) * group_faults(p, dst_nid) * 4;
 }
 
-static unsigned long weighted_cpuload(const int cpu);
+static unsigned long weighted_cpuload(struct rq *rq);
 static unsigned long source_load(int cpu, int type);
 static unsigned long target_load(int cpu, int type);
 static unsigned long capacity_of(int cpu);
@@ -1409,7 +1470,7 @@ static void update_numa_stats(struct numa_stats *ns, int nid)
 		struct rq *rq = cpu_rq(cpu);
 
 		ns->nr_running += rq->nr_running;
-		ns->load += weighted_cpuload(cpu);
+		ns->load += weighted_cpuload(rq);
 		ns->compute_capacity += capacity_of(cpu);
 
 		cpus++;
@@ -1808,7 +1869,7 @@ static int task_numa_migrate(struct task_struct *p)
 	 * Reset the scan period if the task is being rescheduled on an
 	 * alternative node to recheck if the tasks is now properly placed.
 	 */
-	p->numa_scan_period = task_scan_min(p);
+	p->numa_scan_period = task_scan_start(p);
 
 	if (env.best_task == NULL) {
 		ret = migrate_task_to(p, env.best_cpu);
@@ -1892,7 +1953,7 @@ static void update_task_scan_period(struct task_struct *p,
 			unsigned long shared, unsigned long private)
 {
 	unsigned int period_slot;
-	int ratio;
+	int lr_ratio, ps_ratio;
 	int diff;
 
 	unsigned long remote = p->numa_faults_locality[0];
@@ -1922,25 +1983,36 @@ static void update_task_scan_period(struct task_struct *p,
 	 *	 >= NUMA_PERIOD_THRESHOLD scan period increases (scan slower)
 	 */
 	period_slot = DIV_ROUND_UP(p->numa_scan_period, NUMA_PERIOD_SLOTS);
-	ratio = (local * NUMA_PERIOD_SLOTS) / (local + remote);
-	if (ratio >= NUMA_PERIOD_THRESHOLD) {
-		int slot = ratio - NUMA_PERIOD_THRESHOLD;
+	lr_ratio = (local * NUMA_PERIOD_SLOTS) / (local + remote);
+	ps_ratio = (private * NUMA_PERIOD_SLOTS) / (private + shared);
+
+	if (ps_ratio >= NUMA_PERIOD_THRESHOLD) {
+		/*
+		 * Most memory accesses are local. There is no need to
+		 * do fast NUMA scanning, since memory is already local.
+		 */
+		int slot = ps_ratio - NUMA_PERIOD_THRESHOLD;
+		if (!slot)
+			slot = 1;
+		diff = slot * period_slot;
+	} else if (lr_ratio >= NUMA_PERIOD_THRESHOLD) {
+		/*
+		 * Most memory accesses are shared with other tasks.
+		 * There is no point in continuing fast NUMA scanning,
+		 * since other tasks may just move the memory elsewhere.
+		 */
+		int slot = lr_ratio - NUMA_PERIOD_THRESHOLD;
 		if (!slot)
 			slot = 1;
 		diff = slot * period_slot;
 	} else {
-		diff = -(NUMA_PERIOD_THRESHOLD - ratio) * period_slot;
-
 		/*
-		 * Scale scan rate increases based on sharing. There is an
-		 * inverse relationship between the degree of sharing and
-		 * the adjustment made to the scanning period. Broadly
-		 * speaking the intent is that there is little point
-		 * scanning faster if shared accesses dominate as it may
-		 * simply bounce migrations uselessly
+		 * Private memory faults exceed (SLOTS-THRESHOLD)/SLOTS,
+		 * yet they are not on the local NUMA node. Speed up
+		 * NUMA scanning to get the memory moved over.
 		 */
-		ratio = DIV_ROUND_UP(private * NUMA_PERIOD_SLOTS, (private + shared + 1));
-		diff = (diff * ratio) / NUMA_PERIOD_SLOTS;
+		int ratio = max(lr_ratio, ps_ratio);
+		diff = -(NUMA_PERIOD_THRESHOLD - ratio) * period_slot;
 	}
 
 	p->numa_scan_period = clamp(p->numa_scan_period + diff,
@@ -2448,7 +2520,7 @@ void task_numa_work(struct callback_head *work)
 
 	if (p->numa_scan_period == 0) {
 		p->numa_scan_period_max = task_scan_max(p);
-		p->numa_scan_period = task_scan_min(p);
+		p->numa_scan_period = task_scan_start(p);
 	}
 
 	next_scan = now + msecs_to_jiffies(p->numa_scan_period);
@@ -2576,7 +2648,7 @@ void task_tick_numa(struct rq *rq, struct task_struct *curr)
 
 	if (now > curr->node_stamp + period) {
 		if (!curr->node_stamp)
-			curr->numa_scan_period = task_scan_min(curr);
+			curr->numa_scan_period = task_scan_start(curr);
 		curr->node_stamp += period;
 
 		if (!time_before(jiffies, curr->mm->numa_next_scan)) {
@@ -2586,59 +2658,6 @@ void task_tick_numa(struct rq *rq, struct task_struct *curr)
 	}
 }
 
-/*
- * Can a task be moved from prev_cpu to this_cpu without causing a load
- * imbalance that would trigger the load balancer?
- */
-static inline bool numa_wake_affine(struct sched_domain *sd,
-				    struct task_struct *p, int this_cpu,
-				    int prev_cpu, int sync)
-{
-	struct numa_stats prev_load, this_load;
-	s64 this_eff_load, prev_eff_load;
-
-	update_numa_stats(&prev_load, cpu_to_node(prev_cpu));
-	update_numa_stats(&this_load, cpu_to_node(this_cpu));
-
-	/*
-	 * If sync wakeup then subtract the (maximum possible)
-	 * effect of the currently running task from the load
-	 * of the current CPU:
-	 */
-	if (sync) {
-		unsigned long current_load = task_h_load(current);
-
-		if (this_load.load > current_load)
-			this_load.load -= current_load;
-		else
-			this_load.load = 0;
-	}
-
-	/*
-	 * In low-load situations, where this_cpu's node is idle due to the
-	 * sync cause above having dropped this_load.load to 0, move the task.
-	 * Moving to an idle socket will not create a bad imbalance.
-	 *
-	 * Otherwise check if the nodes are near enough in load to allow this
-	 * task to be woken on this_cpu's node.
-	 */
-	if (this_load.load > 0) {
-		unsigned long task_load = task_h_load(p);
-
-		this_eff_load = 100;
-		this_eff_load *= prev_load.compute_capacity;
-
-		prev_eff_load = 100 + (sd->imbalance_pct - 100) / 2;
-		prev_eff_load *= this_load.compute_capacity;
-
-		this_eff_load *= this_load.load + task_load;
-		prev_eff_load *= prev_load.load - task_load;
-
-		return this_eff_load <= prev_eff_load;
-	}
-
-	return true;
-}
 #else
 static void task_tick_numa(struct rq *rq, struct task_struct *curr)
 {
@@ -2652,14 +2671,6 @@ static inline void account_numa_dequeue(struct rq *rq, struct task_struct *p)
 {
 }
 
-#ifdef CONFIG_SMP
-static inline bool numa_wake_affine(struct sched_domain *sd,
-				    struct task_struct *p, int this_cpu,
-				    int prev_cpu, int sync)
-{
-	return true;
-}
-#endif /* !SMP */
 #endif /* CONFIG_NUMA_BALANCING */
 
 static void
@@ -2790,6 +2801,29 @@ static inline void update_cfs_shares(struct sched_entity *se)
 }
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 
+static inline void cfs_rq_util_change(struct cfs_rq *cfs_rq)
+{
+	if (&this_rq()->cfs == cfs_rq) {
+		/*
+		 * There are a few boundary cases this might miss but it should
+		 * get called often enough that that should (hopefully) not be
+		 * a real problem -- added to that it only calls on the local
+		 * CPU, so if we enqueue remotely we'll miss an update, but
+		 * the next tick/schedule should update.
+		 *
+		 * It will not get called when we go idle, because the idle
+		 * thread is a different class (!fair), nor will the utilization
+		 * number include things like RT tasks.
+		 *
+		 * As is, the util number is not freq-invariant (we'd have to
+		 * implement arch_scale_freq_capacity() for that).
+		 *
+		 * See cpu_util().
+		 */
+		cpufreq_update_util(rq_of(cfs_rq), 0);
+	}
+}
+
 #ifdef CONFIG_SMP
 /*
  * Approximate:
@@ -2967,6 +3001,18 @@ ___update_load_avg(u64 now, int cpu, struct sched_avg *sa,
 
 	sa->last_update_time += delta << 10;
 
+	/*
+	 * running is a subset of runnable (weight) so running can't be set if
+	 * runnable is clear. But there are some corner cases where the current
+	 * se has been already dequeued but cfs_rq->curr still points to it.
+	 * This means that weight will be 0 but not running for a sched_entity
+	 * but also for a cfs_rq if the latter becomes idle. As an example,
+	 * this happens during idle_balance() which calls
+	 * update_blocked_averages()
+	 */
+	if (!weight)
+		running = 0;
+
 	/*
 	 * Now we know we crossed measurement unit boundaries. The *_avg
 	 * accrues by two steps:
@@ -3276,29 +3322,6 @@ static inline void set_tg_cfs_propagate(struct cfs_rq *cfs_rq) {}
 
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 
-static inline void cfs_rq_util_change(struct cfs_rq *cfs_rq)
-{
-	if (&this_rq()->cfs == cfs_rq) {
-		/*
-		 * There are a few boundary cases this might miss but it should
-		 * get called often enough that that should (hopefully) not be
-		 * a real problem -- added to that it only calls on the local
-		 * CPU, so if we enqueue remotely we'll miss an update, but
-		 * the next tick/schedule should update.
-		 *
-		 * It will not get called when we go idle, because the idle
-		 * thread is a different class (!fair), nor will the utilization
-		 * number include things like RT tasks.
-		 *
-		 * As is, the util number is not freq-invariant (we'd have to
-		 * implement arch_scale_freq_capacity() for that).
-		 *
-		 * See cpu_util().
-		 */
-		cpufreq_update_util(rq_of(cfs_rq), 0);
-	}
-}
-
 /*
  * Unsigned subtract and clamp on underflow.
  *
@@ -3320,7 +3343,6 @@ static inline void cfs_rq_util_change(struct cfs_rq *cfs_rq)
  * update_cfs_rq_load_avg - update the cfs_rq's load/util averages
  * @now: current time, as per cfs_rq_clock_task()
  * @cfs_rq: cfs_rq to update
- * @update_freq: should we call cfs_rq_util_change() or will the call do so
  *
  * The cfs_rq avg is the direct sum of all its entities (blocked and runnable)
  * avg. The immediate corollary is that all (fair) tasks must be attached, see
@@ -3334,7 +3356,7 @@ static inline void cfs_rq_util_change(struct cfs_rq *cfs_rq)
  * call update_tg_load_avg() when this function returns true.
  */
 static inline int
-update_cfs_rq_load_avg(u64 now, struct cfs_rq *cfs_rq, bool update_freq)
+update_cfs_rq_load_avg(u64 now, struct cfs_rq *cfs_rq)
 {
 	struct sched_avg *sa = &cfs_rq->avg;
 	int decayed, removed_load = 0, removed_util = 0;
@@ -3362,7 +3384,7 @@ update_cfs_rq_load_avg(u64 now, struct cfs_rq *cfs_rq, bool update_freq)
 	cfs_rq->load_last_update_time_copy = sa->last_update_time;
 #endif
 
-	if (update_freq && (decayed || removed_util))
+	if (decayed || removed_util)
 		cfs_rq_util_change(cfs_rq);
 
 	return decayed || removed_load;
@@ -3390,7 +3412,7 @@ static inline void update_load_avg(struct sched_entity *se, int flags)
 	if (se->avg.last_update_time && !(flags & SKIP_AGE_LOAD))
 		__update_load_avg_se(now, cpu, cfs_rq, se);
 
-	decayed  = update_cfs_rq_load_avg(now, cfs_rq, true);
+	decayed  = update_cfs_rq_load_avg(now, cfs_rq);
 	decayed |= propagate_entity_load_avg(se);
 
 	if (decayed && (flags & UPDATE_TG))
@@ -3534,7 +3556,7 @@ static int idle_balance(struct rq *this_rq, struct rq_flags *rf);
 #else /* CONFIG_SMP */
 
 static inline int
-update_cfs_rq_load_avg(u64 now, struct cfs_rq *cfs_rq, bool update_freq)
+update_cfs_rq_load_avg(u64 now, struct cfs_rq *cfs_rq)
 {
 	return 0;
 }
@@ -3544,7 +3566,7 @@ update_cfs_rq_load_avg(u64 now, struct cfs_rq *cfs_rq, bool update_freq)
 
 static inline void update_load_avg(struct sched_entity *se, int not_used1)
 {
-	cpufreq_update_util(rq_of(cfs_rq_of(se)), 0);
+	cfs_rq_util_change(cfs_rq_of(se));
 }
 
 static inline void
@@ -5125,9 +5147,9 @@ static void cpu_load_update(struct rq *this_rq, unsigned long this_load,
 }
 
 /* Used instead of source_load when we know the type == 0 */
-static unsigned long weighted_cpuload(const int cpu)
+static unsigned long weighted_cpuload(struct rq *rq)
 {
-	return cfs_rq_runnable_load_avg(&cpu_rq(cpu)->cfs);
+	return cfs_rq_runnable_load_avg(&rq->cfs);
 }
 
 #ifdef CONFIG_NO_HZ_COMMON
@@ -5172,7 +5194,7 @@ static void cpu_load_update_idle(struct rq *this_rq)
 	/*
 	 * bail if there's load or we're actually up-to-date.
 	 */
-	if (weighted_cpuload(cpu_of(this_rq)))
+	if (weighted_cpuload(this_rq))
 		return;
 
 	cpu_load_update_nohz(this_rq, READ_ONCE(jiffies), 0);
@@ -5193,7 +5215,7 @@ void cpu_load_update_nohz_start(void)
 	 * concurrently we'll exit nohz. And cpu_load write can race with
 	 * cpu_load_update_idle() but both updater would be writing the same.
 	 */
-	this_rq->cpu_load[0] = weighted_cpuload(cpu_of(this_rq));
+	this_rq->cpu_load[0] = weighted_cpuload(this_rq);
 }
 
 /*
@@ -5209,7 +5231,7 @@ void cpu_load_update_nohz_stop(void)
 	if (curr_jiffies == this_rq->last_load_update_tick)
 		return;
 
-	load = weighted_cpuload(cpu_of(this_rq));
+	load = weighted_cpuload(this_rq);
 	rq_lock(this_rq, &rf);
 	update_rq_clock(this_rq);
 	cpu_load_update_nohz(this_rq, curr_jiffies, load);
@@ -5235,7 +5257,7 @@ static void cpu_load_update_periodic(struct rq *this_rq, unsigned long load)
  */
 void cpu_load_update_active(struct rq *this_rq)
 {
-	unsigned long load = weighted_cpuload(cpu_of(this_rq));
+	unsigned long load = weighted_cpuload(this_rq);
 
 	if (tick_nohz_tick_stopped())
 		cpu_load_update_nohz(this_rq, READ_ONCE(jiffies), load);
@@ -5253,7 +5275,7 @@ void cpu_load_update_active(struct rq *this_rq)
 static unsigned long source_load(int cpu, int type)
 {
 	struct rq *rq = cpu_rq(cpu);
-	unsigned long total = weighted_cpuload(cpu);
+	unsigned long total = weighted_cpuload(rq);
 
 	if (type == 0 || !sched_feat(LB_BIAS))
 		return total;
@@ -5268,7 +5290,7 @@ static unsigned long source_load(int cpu, int type)
 static unsigned long target_load(int cpu, int type)
 {
 	struct rq *rq = cpu_rq(cpu);
-	unsigned long total = weighted_cpuload(cpu);
+	unsigned long total = weighted_cpuload(rq);
 
 	if (type == 0 || !sched_feat(LB_BIAS))
 		return total;
@@ -5290,7 +5312,7 @@ static unsigned long cpu_avg_load_per_task(int cpu)
 {
 	struct rq *rq = cpu_rq(cpu);
 	unsigned long nr_running = READ_ONCE(rq->cfs.h_nr_running);
-	unsigned long load_avg = weighted_cpuload(cpu);
+	unsigned long load_avg = weighted_cpuload(rq);
 
 	if (nr_running)
 		return load_avg / nr_running;
@@ -5345,20 +5367,115 @@ static int wake_wide(struct task_struct *p)
 	return 1;
 }
 
+struct llc_stats {
+	unsigned long	nr_running;
+	unsigned long	load;
+	unsigned long	capacity;
+	int		has_capacity;
+};
+
+static bool get_llc_stats(struct llc_stats *stats, int cpu)
+{
+	struct sched_domain_shared *sds = rcu_dereference(per_cpu(sd_llc_shared, cpu));
+
+	if (!sds)
+		return false;
+
+	stats->nr_running	= READ_ONCE(sds->nr_running);
+	stats->load		= READ_ONCE(sds->load);
+	stats->capacity		= READ_ONCE(sds->capacity);
+	stats->has_capacity	= stats->nr_running < per_cpu(sd_llc_size, cpu);
+
+	return true;
+}
+
+/*
+ * Can a task be moved from prev_cpu to this_cpu without causing a load
+ * imbalance that would trigger the load balancer?
+ *
+ * Since we're running on 'stale' values, we might in fact create an imbalance
+ * but recomputing these values is expensive, as that'd mean iteration 2 cache
+ * domains worth of CPUs.
+ */
+static bool
+wake_affine_llc(struct sched_domain *sd, struct task_struct *p,
+		int this_cpu, int prev_cpu, int sync)
+{
+	struct llc_stats prev_stats, this_stats;
+	s64 this_eff_load, prev_eff_load;
+	unsigned long task_load;
+
+	if (!get_llc_stats(&prev_stats, prev_cpu) ||
+	    !get_llc_stats(&this_stats, this_cpu))
+		return false;
+
+	/*
+	 * If sync wakeup then subtract the (maximum possible)
+	 * effect of the currently running task from the load
+	 * of the current LLC.
+	 */
+	if (sync) {
+		unsigned long current_load = task_h_load(current);
+
+		/* in this case load hits 0 and this LLC is considered 'idle' */
+		if (current_load > this_stats.load)
+			return true;
+
+		this_stats.load -= current_load;
+	}
+
+	/*
+	 * The has_capacity stuff is not SMT aware, but by trying to balance
+	 * the nr_running on both ends we try and fill the domain at equal
+	 * rates, thereby first consuming cores before siblings.
+	 */
+
+	/* if the old cache has capacity, stay there */
+	if (prev_stats.has_capacity && prev_stats.nr_running < this_stats.nr_running+1)
+		return false;
+
+	/* if this cache has capacity, come here */
+	if (this_stats.has_capacity && this_stats.nr_running < prev_stats.nr_running+1)
+		return true;
+
+	/*
+	 * Check to see if we can move the load without causing too much
+	 * imbalance.
+	 */
+	task_load = task_h_load(p);
+
+	this_eff_load = 100;
+	this_eff_load *= prev_stats.capacity;
+
+	prev_eff_load = 100 + (sd->imbalance_pct - 100) / 2;
+	prev_eff_load *= this_stats.capacity;
+
+	this_eff_load *= this_stats.load + task_load;
+	prev_eff_load *= prev_stats.load - task_load;
+
+	return this_eff_load <= prev_eff_load;
+}
+
 static int wake_affine(struct sched_domain *sd, struct task_struct *p,
 		       int prev_cpu, int sync)
 {
 	int this_cpu = smp_processor_id();
-	bool affine = false;
+	bool affine;
 
 	/*
-	 * Common case: CPUs are in the same socket, and select_idle_sibling()
-	 * will do its thing regardless of what we return:
+	 * Default to no affine wakeups; wake_affine() should not effect a task
+	 * placement the load-balancer feels inclined to undo. The conservative
+	 * option is therefore to not move tasks when they wake up.
 	 */
-	if (cpus_share_cache(prev_cpu, this_cpu))
-		affine = true;
-	else
-		affine = numa_wake_affine(sd, p, this_cpu, prev_cpu, sync);
+	affine = false;
+
+	/*
+	 * If the wakeup is across cache domains, try to evaluate if movement
+	 * makes sense, otherwise rely on select_idle_siblings() to do
+	 * placement inside the cache domain.
+	 */
+	if (!cpus_share_cache(prev_cpu, this_cpu))
+		affine = wake_affine_llc(sd, p, this_cpu, prev_cpu, sync);
 
 	schedstat_inc(p->se.statistics.nr_wakeups_affine_attempts);
 	if (affine) {
@@ -5550,7 +5667,7 @@ find_idlest_cpu(struct sched_group *group, struct task_struct *p, int this_cpu)
 				shallowest_idle_cpu = i;
 			}
 		} else if (shallowest_idle_cpu == -1) {
-			load = weighted_cpuload(i);
+			load = weighted_cpuload(cpu_rq(i));
 			if (load < min_load || (load == min_load && i == this_cpu)) {
 				min_load = load;
 				least_loaded_cpu = i;
@@ -6187,10 +6304,10 @@ pick_next_task_fair(struct rq *rq, struct task_struct *prev, struct rq_flags *rf
 	int new_tasks;
 
 again:
-#ifdef CONFIG_FAIR_GROUP_SCHED
 	if (!cfs_rq->nr_running)
 		goto idle;
 
+#ifdef CONFIG_FAIR_GROUP_SCHED
 	if (prev->sched_class != &fair_sched_class)
 		goto simple;
 
@@ -6220,11 +6337,17 @@ pick_next_task_fair(struct rq *rq, struct task_struct *prev, struct rq_flags *rf
 			/*
 			 * This call to check_cfs_rq_runtime() will do the
 			 * throttle and dequeue its entity in the parent(s).
-			 * Therefore the 'simple' nr_running test will indeed
+			 * Therefore the nr_running test will indeed
 			 * be correct.
 			 */
-			if (unlikely(check_cfs_rq_runtime(cfs_rq)))
+			if (unlikely(check_cfs_rq_runtime(cfs_rq))) {
+				cfs_rq = &rq->cfs;
+
+				if (!cfs_rq->nr_running)
+					goto idle;
+
 				goto simple;
+			}
 		}
 
 		se = pick_next_entity(cfs_rq, curr);
@@ -6264,12 +6387,8 @@ pick_next_task_fair(struct rq *rq, struct task_struct *prev, struct rq_flags *rf
 
 	return p;
 simple:
-	cfs_rq = &rq->cfs;
 #endif
 
-	if (!cfs_rq->nr_running)
-		goto idle;
-
 	put_prev_task(rq, prev);
 
 	do {
@@ -6917,7 +7036,7 @@ static void update_blocked_averages(int cpu)
 		if (throttled_hierarchy(cfs_rq))
 			continue;
 
-		if (update_cfs_rq_load_avg(cfs_rq_clock_task(cfs_rq), cfs_rq, true))
+		if (update_cfs_rq_load_avg(cfs_rq_clock_task(cfs_rq), cfs_rq))
 			update_tg_load_avg(cfs_rq, 0);
 
 		/* Propagate pending load changes to the parent, if any: */
@@ -6990,7 +7109,7 @@ static inline void update_blocked_averages(int cpu)
 
 	rq_lock_irqsave(rq, &rf);
 	update_rq_clock(rq);
-	update_cfs_rq_load_avg(cfs_rq_clock_task(cfs_rq), cfs_rq, true);
+	update_cfs_rq_load_avg(cfs_rq_clock_task(cfs_rq), cfs_rq);
 	rq_unlock_irqrestore(rq, &rf);
 }
 
@@ -7036,6 +7155,7 @@ struct sg_lb_stats {
 struct sd_lb_stats {
 	struct sched_group *busiest;	/* Busiest group in this sd */
 	struct sched_group *local;	/* Local group in this sd */
+	unsigned long total_running;
 	unsigned long total_load;	/* Total load of all groups in sd */
 	unsigned long total_capacity;	/* Total capacity of all groups in sd */
 	unsigned long avg_load;	/* Average load across all groups in sd */
@@ -7055,6 +7175,7 @@ static inline void init_sd_lb_stats(struct sd_lb_stats *sds)
 	*sds = (struct sd_lb_stats){
 		.busiest = NULL,
 		.local = NULL,
+		.total_running = 0UL,
 		.total_load = 0UL,
 		.total_capacity = 0UL,
 		.busiest_stat = {
@@ -7363,7 +7484,7 @@ static inline void update_sg_lb_stats(struct lb_env *env,
 		sgs->nr_numa_running += rq->nr_numa_running;
 		sgs->nr_preferred_running += rq->nr_preferred_running;
 #endif
-		sgs->sum_weighted_load += weighted_cpuload(i);
+		sgs->sum_weighted_load += weighted_cpuload(rq);
 		/*
 		 * No need to call idle_cpu() if nr_running is not 0
 		 */
@@ -7490,6 +7611,7 @@ static inline enum fbq_type fbq_classify_rq(struct rq *rq)
  */
 static inline void update_sd_lb_stats(struct lb_env *env, struct sd_lb_stats *sds)
 {
+	struct sched_domain_shared *shared = env->sd->shared;
 	struct sched_domain *child = env->sd->child;
 	struct sched_group *sg = env->sd->groups;
 	struct sg_lb_stats *local = &sds->local_stat;
@@ -7546,6 +7668,7 @@ static inline void update_sd_lb_stats(struct lb_env *env, struct sd_lb_stats *sd
 
 next_group:
 		/* Now, start updating sd_lb_stats */
+		sds->total_running += sgs->sum_nr_running;
 		sds->total_load += sgs->group_load;
 		sds->total_capacity += sgs->group_capacity;
 
@@ -7561,6 +7684,21 @@ static inline void update_sd_lb_stats(struct lb_env *env, struct sd_lb_stats *sd
 			env->dst_rq->rd->overload = overload;
 	}
 
+	if (!shared)
+		return;
+
+	/*
+	 * Since these are sums over groups they can contain some CPUs
+	 * multiple times for the NUMA domains.
+	 *
+	 * Currently only wake_affine_llc() and find_busiest_group()
+	 * uses these numbers, only the last is affected by this problem.
+	 *
+	 * XXX fix that.
+	 */
+	WRITE_ONCE(shared->nr_running,	sds->total_running);
+	WRITE_ONCE(shared->load,	sds->total_load);
+	WRITE_ONCE(shared->capacity,	sds->total_capacity);
 }
 
 /**
@@ -7790,6 +7928,7 @@ static struct sched_group *find_busiest_group(struct lb_env *env)
 	if (!sds.busiest || busiest->sum_nr_running == 0)
 		goto out_balanced;
 
+	/* XXX broken for overlapping NUMA groups */
 	sds.avg_load = (SCHED_CAPACITY_SCALE * sds.total_load)
 						/ sds.total_capacity;
 
@@ -7892,7 +8031,7 @@ static struct rq *find_busiest_queue(struct lb_env *env,
 
 		capacity = capacity_of(i);
 
-		wl = weighted_cpuload(i);
+		wl = weighted_cpuload(rq);
 
 		/*
 		 * When comparing with imbalance, use weighted_cpuload()

kernel/sched/sched.h

@@ -1120,11 +1120,15 @@ extern int group_balance_cpu(struct sched_group *sg);
 
 #if defined(CONFIG_SCHED_DEBUG) && defined(CONFIG_SYSCTL)
 void register_sched_domain_sysctl(void);
+void dirty_sched_domain_sysctl(int cpu);
 void unregister_sched_domain_sysctl(void);
 #else
 static inline void register_sched_domain_sysctl(void)
 {
 }
+static inline void dirty_sched_domain_sysctl(int cpu)
+{
+}
 static inline void unregister_sched_domain_sysctl(void)
 {
 }

kernel/sched/topology.c

@@ -261,8 +261,6 @@ void rq_attach_root(struct rq *rq, struct root_domain *rd)
 
 static int init_rootdomain(struct root_domain *rd)
 {
-	memset(rd, 0, sizeof(*rd));
-
 	if (!zalloc_cpumask_var(&rd->span, GFP_KERNEL))
 		goto out;
 	if (!zalloc_cpumask_var(&rd->online, GFP_KERNEL))
@@ -311,7 +309,7 @@ static struct root_domain *alloc_rootdomain(void)
 {
 	struct root_domain *rd;
 
-	rd = kmalloc(sizeof(*rd), GFP_KERNEL);
+	rd = kzalloc(sizeof(*rd), GFP_KERNEL);
 	if (!rd)
 		return NULL;
 
@@ -337,7 +335,8 @@ static void free_sched_groups(struct sched_group *sg, int free_sgc)
 		if (free_sgc && atomic_dec_and_test(&sg->sgc->ref))
 			kfree(sg->sgc);
 
-		kfree(sg);
+		if (atomic_dec_and_test(&sg->ref))
+			kfree(sg);
 		sg = tmp;
 	} while (sg != first);
 }
@@ -345,15 +344,12 @@ static void free_sched_groups(struct sched_group *sg, int free_sgc)
 static void destroy_sched_domain(struct sched_domain *sd)
 {
 	/*
-	 * If its an overlapping domain it has private groups, iterate and
-	 * nuke them all.
+	 * A normal sched domain may have multiple group references, an
+	 * overlapping domain, having private groups, only one.  Iterate,
+	 * dropping group/capacity references, freeing where none remain.
 	 */
-	if (sd->flags & SD_OVERLAP) {
-		free_sched_groups(sd->groups, 1);
-	} else if (atomic_dec_and_test(&sd->groups->ref)) {
-		kfree(sd->groups->sgc);
-		kfree(sd->groups);
-	}
+	free_sched_groups(sd->groups, 1);
+
 	if (sd->shared && atomic_dec_and_test(&sd->shared->ref))
 		kfree(sd->shared);
 	kfree(sd);
@@ -463,6 +459,7 @@ cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu)
 	rq_attach_root(rq, rd);
 	tmp = rq->sd;
 	rcu_assign_pointer(rq->sd, sd);
+	dirty_sched_domain_sysctl(cpu);
 	destroy_sched_domains(tmp);
 
 	update_top_cache_domain(cpu);
@@ -670,6 +667,7 @@ build_group_from_child_sched_domain(struct sched_domain *sd, int cpu)
 	else
 		cpumask_copy(sg_span, sched_domain_span(sd));
 
+	atomic_inc(&sg->ref);
 	return sg;
 }
 
@@ -1595,7 +1593,7 @@ static void __sdt_free(const struct cpumask *cpu_map)
 	}
 }
 
-struct sched_domain *build_sched_domain(struct sched_domain_topology_level *tl,
+static struct sched_domain *build_sched_domain(struct sched_domain_topology_level *tl,
 		const struct cpumask *cpu_map, struct sched_domain_attr *attr,
 		struct sched_domain *child, int cpu)
 {
@@ -1854,7 +1852,17 @@ void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
 	/* Let the architecture update CPU core mappings: */
 	new_topology = arch_update_cpu_topology();
 
-	n = doms_new ? ndoms_new : 0;
+	if (!doms_new) {
+		WARN_ON_ONCE(dattr_new);
+		n = 0;
+		doms_new = alloc_sched_domains(1);
+		if (doms_new) {
+			n = 1;
+			cpumask_andnot(doms_new[0], cpu_active_mask, cpu_isolated_map);
+		}
+	} else {
+		n = ndoms_new;
+	}
 
 	/* Destroy deleted domains: */
 	for (i = 0; i < ndoms_cur; i++) {
@@ -1870,11 +1878,10 @@ void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
 	}
 
 	n = ndoms_cur;
-	if (doms_new == NULL) {
+	if (!doms_new) {
 		n = 0;
 		doms_new = &fallback_doms;
 		cpumask_andnot(doms_new[0], cpu_active_mask, cpu_isolated_map);
-		WARN_ON_ONCE(dattr_new);
 	}
 
 	/* Build new domains: */