[PATCH] sched: use softirq for load balancing

Call rebalance_tick (renamed to run_rebalance_domains) from a newly introduced
softirq.

We calculate the earliest time for each layer of sched domains to be rescanned
(this is the rescan time for idle) and use the earliest of those to schedule
the softirq via a new field "next_balance" added to struct rq.

Signed-off-by: Christoph Lameter <[email protected]>
Cc: Peter Williams <[email protected]>
Cc: Nick Piggin <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: "Siddha, Suresh B" <[email protected]>
Cc: "Chen, Kenneth W" <[email protected]>
Acked-by: Ingo Molnar <[email protected]>
Cc: KAMEZAWA Hiroyuki <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

Author: Christoph Lameter

include/linux/interrupt.h

@@ -231,7 +231,8 @@ enum
 	NET_TX_SOFTIRQ,
 	NET_RX_SOFTIRQ,
 	BLOCK_SOFTIRQ,
-	TASKLET_SOFTIRQ
+	TASKLET_SOFTIRQ,
+	SCHED_SOFTIRQ,
 };
 
 /* softirq mask and active fields moved to irq_cpustat_t in

kernel/sched.c

@@ -227,6 +227,7 @@ struct rq {
 	unsigned long expired_timestamp;
 	unsigned long long timestamp_last_tick;
 	struct task_struct *curr, *idle;
+	unsigned long next_balance;
 	struct mm_struct *prev_mm;
 	struct prio_array *active, *expired, arrays[2];
 	int best_expired_prio;
@@ -2858,17 +2859,18 @@ static void update_load(struct rq *this_rq)
 }
 
 /*
- * rebalance_tick will get called every timer tick, on every CPU.
+ * run_rebalance_domains is triggered when needed from the scheduler tick.
  *
  * It checks each scheduling domain to see if it is due to be balanced,
  * and initiates a balancing operation if so.
  *
  * Balancing parameters are set up in arch_init_sched_domains.
  */
 
-static void
-rebalance_tick(int this_cpu, struct rq *this_rq)
+static void run_rebalance_domains(struct softirq_action *h)
 {
+	int this_cpu = smp_processor_id();
+	struct rq *this_rq = cpu_rq(this_cpu);
 	unsigned long interval;
 	struct sched_domain *sd;
 	/*
@@ -2877,6 +2879,8 @@ rebalance_tick(int this_cpu, struct rq *this_rq)
 	 */
 	enum idle_type idle = !this_rq->nr_running ?
 				SCHED_IDLE : NOT_IDLE;
+	/* Earliest time when we have to call run_rebalance_domains again */
+	unsigned long next_balance = jiffies + 60*HZ;
 
 	for_each_domain(this_cpu, sd) {
 		if (!(sd->flags & SD_LOAD_BALANCE))
@@ -2891,7 +2895,7 @@ rebalance_tick(int this_cpu, struct rq *this_rq)
 		if (unlikely(!interval))
 			interval = 1;
 
-		if (jiffies - sd->last_balance >= interval) {
+		if (time_after_eq(jiffies, sd->last_balance + interval)) {
 			if (load_balance(this_cpu, this_rq, sd, idle)) {
 				/*
 				 * We've pulled tasks over so either we're no
@@ -2902,7 +2906,10 @@ rebalance_tick(int this_cpu, struct rq *this_rq)
 			}
 			sd->last_balance += interval;
 		}
+		if (time_after(next_balance, sd->last_balance + interval))
+			next_balance = sd->last_balance + interval;
 	}
+	this_rq->next_balance = next_balance;
 }
 #else
 /*
@@ -3155,7 +3162,8 @@ void scheduler_tick(void)
 		task_running_tick(rq, p);
 #ifdef CONFIG_SMP
 	update_load(rq);
-	rebalance_tick(cpu, rq);
+	if (time_after_eq(jiffies, rq->next_balance))
+		raise_softirq(SCHED_SOFTIRQ);
 #endif
 }
 
@@ -6859,6 +6867,10 @@ void __init sched_init(void)
 
 	set_load_weight(&init_task);
 
+#ifdef CONFIG_SMP
+	open_softirq(SCHED_SOFTIRQ, run_rebalance_domains, NULL);
+#endif
+
 #ifdef CONFIG_RT_MUTEXES
 	plist_head_init(&init_task.pi_waiters, &init_task.pi_lock);
 #endif