Introduce nohz_flags in the struct rq, which will track two flags for now.

NOHZ_TICK_STOPPED will convey the stop tick status that gets set when
the tick is stopped and cleared during the first busy tick after the tick
is restarted.

NOHZ_BALANCE_KICK will track the need for nohz idle load balance
on this rq.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com> 
---
 kernel/sched/core.c  |    5 +++--
 kernel/sched/fair.c  |   40 +++++++++++++++++++++-------------------
 kernel/sched/sched.h |   10 +++++++++-
 3 files changed, 33 insertions(+), 22 deletions(-)

Index: tip/kernel/sched/core.c
===================================================================
--- tip.orig/kernel/sched/core.c
+++ tip/kernel/sched/core.c
@@ -575,7 +575,8 @@ void wake_up_idle_cpu(int cpu)
 
 static inline bool got_nohz_idle_kick(void)
 {
-	return idle_cpu(smp_processor_id()) && this_rq()->nohz_balance_kick;
+	int cpu = smp_processor_id();
+	return idle_cpu(cpu) && test_bit(NOHZ_BALANCE_KICK, nohz_flags(cpu));
 }
 
 #else /* CONFIG_NO_HZ */
@@ -6833,7 +6834,7 @@ void __init sched_init(void)
 		rq->avg_idle = 2*sysctl_sched_migration_cost;
 		rq_attach_root(rq, &def_root_domain);
 #ifdef CONFIG_NO_HZ
-		rq->nohz_balance_kick = 0;
+		rq->nohz_flags = 0;
 #endif
 #endif
 		init_rq_hrtick(rq);
Index: tip/kernel/sched/fair.c
===================================================================
--- tip.orig/kernel/sched/fair.c
+++ tip/kernel/sched/fair.c
@@ -4866,18 +4866,15 @@ static void nohz_balancer_kick(int cpu)
 			return;
 	}
 
-	if (!cpu_rq(ilb_cpu)->nohz_balance_kick) {
-		cpu_rq(ilb_cpu)->nohz_balance_kick = 1;
-
-		smp_mb();
-		/*
-		 * Use smp_send_reschedule() instead of resched_cpu().
-		 * This way we generate a sched IPI on the target cpu which
-		 * is idle. And the softirq performing nohz idle load balance
-		 * will be run before returning from the IPI.
-		 */
-		smp_send_reschedule(ilb_cpu);
-	}
+	if (test_and_set_bit(NOHZ_BALANCE_KICK, nohz_flags(cpu)))
+		return;
+	/*
+	 * Use smp_send_reschedule() instead of resched_cpu().
+	 * This way we generate a sched IPI on the target cpu which
+	 * is idle. And the softirq performing nohz idle load balance
+	 * will be run before returning from the IPI.
+	 */
+	smp_send_reschedule(ilb_cpu);
 	return;
 }
 
@@ -4941,6 +4938,8 @@ void select_nohz_load_balancer(int stop_
 			}
 			return;
 		}
+
+		set_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
 	} else {
 		if (!cpumask_test_cpu(cpu, nohz.idle_cpus_mask))
 			return;
@@ -5056,8 +5055,9 @@ static void nohz_idle_balance(int this_c
 	struct rq *rq;
 	int balance_cpu;
 
-	if (idle != CPU_IDLE || !this_rq->nohz_balance_kick)
-		return;
+	if (idle != CPU_IDLE ||
+	    !test_bit(NOHZ_BALANCE_KICK, nohz_flags(this_cpu)))
+		goto end;
 
 	for_each_cpu(balance_cpu, nohz.idle_cpus_mask) {
 		if (balance_cpu == this_cpu)
@@ -5068,10 +5068,8 @@ static void nohz_idle_balance(int this_c
 		 * work being done for other cpus. Next load
 		 * balancing owner will pick it up.
 		 */
-		if (need_resched()) {
-			this_rq->nohz_balance_kick = 0;
+		if (need_resched())
 			break;
-		}
 
 		raw_spin_lock_irq(&this_rq->lock);
 		update_rq_clock(this_rq);
@@ -5085,7 +5083,8 @@ static void nohz_idle_balance(int this_c
 			this_rq->next_balance = rq->next_balance;
 	}
 	nohz.next_balance = this_rq->next_balance;
-	this_rq->nohz_balance_kick = 0;
+end:
+	clear_bit(NOHZ_BALANCE_KICK, nohz_flags(this_cpu));
 }
 
 /*
@@ -5106,6 +5105,9 @@ static inline int nohz_kick_needed(struc
 	int ret;
 	int first_pick_cpu, second_pick_cpu;
 
+	if (unlikely(test_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu))))
+		clear_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
+
 	if (time_before(now, nohz.next_balance))
 		return 0;
 
@@ -5173,7 +5175,7 @@ void trigger_load_balance(struct rq *rq,
 	    likely(!on_null_domain(cpu)))
 		raise_softirq(SCHED_SOFTIRQ);
 #ifdef CONFIG_NO_HZ
-	else if (nohz_kick_needed(rq, cpu) && likely(!on_null_domain(cpu)))
+	if (nohz_kick_needed(rq, cpu) && likely(!on_null_domain(cpu)))
 		nohz_balancer_kick(cpu);
 #endif
 }
Index: tip/kernel/sched/sched.h
===================================================================
--- tip.orig/kernel/sched/sched.h
+++ tip/kernel/sched/sched.h
@@ -371,7 +371,7 @@ struct rq {
 	unsigned long last_load_update_tick;
 #ifdef CONFIG_NO_HZ
 	u64 nohz_stamp;
-	unsigned char nohz_balance_kick;
+	unsigned long nohz_flags;
 #endif
 	int skip_clock_update;
 
@@ -1062,3 +1062,11 @@ extern void init_rt_rq(struct rt_rq *rt_
 extern void unthrottle_offline_cfs_rqs(struct rq *rq);
 
 extern void account_cfs_bandwidth_used(int enabled, int was_enabled);
+
+#ifdef CONFIG_NO_HZ
+enum rq_nohz_flag_bits {
+	NOHZ_TICK_STOPPED,
+	NOHZ_BALANCE_KICK,
+};
+#define nohz_flags(cpu)		(&cpu_rq(cpu)->nohz_flags)
+#endif /* CONFIG_NO_HZ */


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/