sched: fair-group: SMP-nice for group scheduling

Implement SMP nice support for the full group hierarchy.

On each load-balance action, compile a sched_domain wide view of the full
task_group tree. We compute the domain wide view when walking down the
hierarchy, and readjust the weights when walking back up.

After collecting and readjusting the domain wide view, we try to balance the
tasks within the task_groups. The current approach is a naively balance each
task group until we've moved the targeted amount of load.

Inspired by Srivatsa Vaddsgiri's previous code and Abhishek Chandra's H-SMP
paper.

XXX: there will be some numerical issues due to the limited nature of
     SCHED_LOAD_SCALE wrt to representing a task_groups influence on the
     total weight. When the tree is deep enough, or the task weight small
     enough, we'll run out of bits.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
CC: Abhishek Chandra <chandra@cs.umn.edu>
CC: Srivatsa Vaddagiri <vatsa@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

include/linux/sched.h

@@ -758,6 +758,7 @@ struct sched_domain {
 	struct sched_domain *child;	/* bottom domain must be null terminated */
 	struct sched_group *groups;	/* the balancing groups of the domain */
 	cpumask_t span;			/* span of all CPUs in this domain */
+	int first_cpu;			/* cache of the first cpu in this domain */
 	unsigned long min_interval;	/* Minimum balance interval ms */
 	unsigned long max_interval;	/* Maximum balance interval ms */
 	unsigned int busy_factor;	/* less balancing by factor if busy */

kernel/sched_fair.c

@@ -492,10 +492,27 @@ update_stats_curr_start(struct cfs_rq *cfs_rq, struct sched_entity *se)
  * Scheduling class queueing methods:
  */
 
+#if defined CONFIG_SMP && defined CONFIG_FAIR_GROUP_SCHED
+static void
+add_cfs_task_weight(struct cfs_rq *cfs_rq, unsigned long weight)
+{
+	cfs_rq->task_weight += weight;
+}
+#else
+static inline void
+add_cfs_task_weight(struct cfs_rq *cfs_rq, unsigned long weight)
+{
+}
+#endif
+
 static void
 account_entity_enqueue(struct cfs_rq *cfs_rq, struct sched_entity *se)
 {
 	update_load_add(&cfs_rq->load, se->load.weight);
+	if (!parent_entity(se))
+		inc_cpu_load(rq_of(cfs_rq), se->load.weight);
+	if (entity_is_task(se))
+		add_cfs_task_weight(cfs_rq, se->load.weight);
 	cfs_rq->nr_running++;
 	se->on_rq = 1;
 }
@@ -504,6 +521,10 @@ static void
 account_entity_dequeue(struct cfs_rq *cfs_rq, struct sched_entity *se)
 {
 	update_load_sub(&cfs_rq->load, se->load.weight);
+	if (!parent_entity(se))
+		dec_cpu_load(rq_of(cfs_rq), se->load.weight);
+	if (entity_is_task(se))
+		add_cfs_task_weight(cfs_rq, -se->load.weight);
 	cfs_rq->nr_running--;
 	se->on_rq = 0;
 }
@@ -1286,75 +1307,90 @@ static struct task_struct *load_balance_next_fair(void *arg)
 	return __load_balance_iterator(cfs_rq, cfs_rq->rb_load_balance_curr);
 }
 
-#ifdef CONFIG_FAIR_GROUP_SCHED
-static int cfs_rq_best_prio(struct cfs_rq *cfs_rq)
+static unsigned long
+__load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest,
+		unsigned long max_load_move, struct sched_domain *sd,
+		enum cpu_idle_type idle, int *all_pinned, int *this_best_prio,
+		struct cfs_rq *cfs_rq)
 {
-	struct sched_entity *curr;
-	struct task_struct *p;
-
-	if (!cfs_rq->nr_running || !first_fair(cfs_rq))
-		return MAX_PRIO;
-
-	curr = cfs_rq->curr;
-	if (!curr)
-		curr = __pick_next_entity(cfs_rq);
+	struct rq_iterator cfs_rq_iterator;
 
-	p = task_of(curr);
+	cfs_rq_iterator.start = load_balance_start_fair;
+	cfs_rq_iterator.next = load_balance_next_fair;
+	cfs_rq_iterator.arg = cfs_rq;
 
-	return p->prio;
+	return balance_tasks(this_rq, this_cpu, busiest,
+			max_load_move, sd, idle, all_pinned,
+			this_best_prio, &cfs_rq_iterator);
 }
-#endif
 
+#ifdef CONFIG_FAIR_GROUP_SCHED
 static unsigned long
 load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest,
 		  unsigned long max_load_move,
 		  struct sched_domain *sd, enum cpu_idle_type idle,
 		  int *all_pinned, int *this_best_prio)
 {
-	struct cfs_rq *busy_cfs_rq;
 	long rem_load_move = max_load_move;
-	struct rq_iterator cfs_rq_iterator;
+	int busiest_cpu = cpu_of(busiest);
+	struct task_group *tg;
 
-	cfs_rq_iterator.start = load_balance_start_fair;
-	cfs_rq_iterator.next = load_balance_next_fair;
-
-	for_each_leaf_cfs_rq(busiest, busy_cfs_rq) {
-#ifdef CONFIG_FAIR_GROUP_SCHED
-		struct cfs_rq *this_cfs_rq;
+	rcu_read_lock();
+	list_for_each_entry(tg, &task_groups, list) {
 		long imbalance;
-		unsigned long maxload;
+		unsigned long this_weight, busiest_weight;
+		long rem_load, max_load, moved_load;
+
+		/*
+		 * empty group
+		 */
+		if (!aggregate(tg, sd)->task_weight)
+			continue;
+
+		rem_load = rem_load_move * aggregate(tg, sd)->rq_weight;
+		rem_load /= aggregate(tg, sd)->load + 1;
 
-		this_cfs_rq = cpu_cfs_rq(busy_cfs_rq, this_cpu);
+		this_weight = tg->cfs_rq[this_cpu]->task_weight;
+		busiest_weight = tg->cfs_rq[busiest_cpu]->task_weight;
 
-		imbalance = busy_cfs_rq->load.weight - this_cfs_rq->load.weight;
-		/* Don't pull if this_cfs_rq has more load than busy_cfs_rq */
-		if (imbalance <= 0)
+		imbalance = (busiest_weight - this_weight) / 2;
+
+		if (imbalance < 0)
+			imbalance = busiest_weight;
+
+		max_load = max(rem_load, imbalance);
+		moved_load = __load_balance_fair(this_rq, this_cpu, busiest,
+				max_load, sd, idle, all_pinned, this_best_prio,
+				tg->cfs_rq[busiest_cpu]);
+
+		if (!moved_load)
 			continue;
 
-		/* Don't pull more than imbalance/2 */
-		imbalance /= 2;
-		maxload = min(rem_load_move, imbalance);
+		move_group_shares(tg, sd, busiest_cpu, this_cpu);
 
-		*this_best_prio = cfs_rq_best_prio(this_cfs_rq);
-#else
-# define maxload rem_load_move
-#endif
-		/*
-		 * pass busy_cfs_rq argument into
-		 * load_balance_[start|next]_fair iterators
-		 */
-		cfs_rq_iterator.arg = busy_cfs_rq;
-		rem_load_move -= balance_tasks(this_rq, this_cpu, busiest,
-					       maxload, sd, idle, all_pinned,
-					       this_best_prio,
-					       &cfs_rq_iterator);
+		moved_load *= aggregate(tg, sd)->load;
+		moved_load /= aggregate(tg, sd)->rq_weight + 1;
 
-		if (rem_load_move <= 0)
+		rem_load_move -= moved_load;
+		if (rem_load_move < 0)
 			break;
 	}
+	rcu_read_unlock();
 
 	return max_load_move - rem_load_move;
 }
+#else
+static unsigned long
+load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest,
+		  unsigned long max_load_move,
+		  struct sched_domain *sd, enum cpu_idle_type idle,
+		  int *all_pinned, int *this_best_prio)
+{
+	return __load_balance_fair(this_rq, this_cpu, busiest,
+			max_load_move, sd, idle, all_pinned,
+			this_best_prio, &busiest->cfs);
+}
+#endif
 
 static int
 move_one_task_fair(struct rq *this_rq, int this_cpu, struct rq *busiest,

kernel/sched_rt.c

@@ -518,6 +518,8 @@ static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup)
 	 */
 	for_each_sched_rt_entity(rt_se)
 		enqueue_rt_entity(rt_se);
+
+	inc_cpu_load(rq, p->se.load.weight);
 }
 
 static void dequeue_task_rt(struct rq *rq, struct task_struct *p, int sleep)
@@ -537,6 +539,8 @@ static void dequeue_task_rt(struct rq *rq, struct task_struct *p, int sleep)
 		if (rt_rq && rt_rq->rt_nr_running)
 			enqueue_rt_entity(rt_se);
 	}
+
+	dec_cpu_load(rq, p->se.load.weight);
 }
 
 /*