---

yaml
---
r: 277245
b: refs/heads/master
c: f23f7fa
h: refs/heads/master
i:
  277243: 7c68da5
v: v3

[refs]

@@ -1,2 +1,2 @@
 ---
-refs/heads/master: dff1672d9199fffddb58fa7970ccf59005fc35f3
+refs/heads/master: f23f7fa1c8effca19b52b98fc71016109d21db59

trunk/kernel/rcutree_plugin.h

@@ -1994,8 +1994,40 @@ static void rcu_prepare_for_idle(int cpu)
 
 #else /* #if !defined(CONFIG_RCU_FAST_NO_HZ) */
 
-#define RCU_NEEDS_CPU_FLUSHES 5		/* Allow for callback self-repost. */
+/*
+ * This code is invoked when a CPU goes idle, at which point we want
+ * to have the CPU do everything required for RCU so that it can enter
+ * the energy-efficient dyntick-idle mode.  This is handled by a
+ * state machine implemented by rcu_prepare_for_idle() below.
+ *
+ * The following three proprocessor symbols control this state machine:
+ *
+ * RCU_IDLE_FLUSHES gives the maximum number of times that we will attempt
+ *	to satisfy RCU.  Beyond this point, it is better to incur a periodic
+ *	scheduling-clock interrupt than to loop through the state machine
+ *	at full power.
+ * RCU_IDLE_OPT_FLUSHES gives the number of RCU_IDLE_FLUSHES that are
+ *	optional if RCU does not need anything immediately from this
+ *	CPU, even if this CPU still has RCU callbacks queued.  The first
+ *	times through the state machine are mandatory: we need to give
+ *	the state machine a chance to communicate a quiescent state
+ *	to the RCU core.
+ * RCU_IDLE_GP_DELAY gives the number of jiffies that a CPU is permitted
+ *	to sleep in dyntick-idle mode with RCU callbacks pending.  This
+ *	is sized to be roughly one RCU grace period.  Those energy-efficiency
+ *	benchmarkers who might otherwise be tempted to set this to a large
+ *	number, be warned: Setting RCU_IDLE_GP_DELAY too high can hang your
+ *	system.  And if you are -that- concerned about energy efficiency,
+ *	just power the system down and be done with it!
+ *
+ * The values below work well in practice.  If future workloads require
+ * adjustment, they can be converted into kernel config parameters, though
+ * making the state machine smarter might be a better option.
+ */
+#define RCU_IDLE_FLUSHES 5		/* Number of dyntick-idle tries. */
+#define RCU_IDLE_OPT_FLUSHES 3		/* Optional dyntick-idle tries. */
 #define RCU_IDLE_GP_DELAY 6		/* Roughly one grace period. */
+
 static DEFINE_PER_CPU(int, rcu_dyntick_drain);
 static DEFINE_PER_CPU(unsigned long, rcu_dyntick_holdoff);
 static DEFINE_PER_CPU(struct hrtimer, rcu_idle_gp_timer);
@@ -2110,17 +2142,17 @@ static void rcu_prepare_for_idle(int cpu)
 	/* Check and update the rcu_dyntick_drain sequencing. */
 	if (per_cpu(rcu_dyntick_drain, cpu) <= 0) {
 		/* First time through, initialize the counter. */
-		per_cpu(rcu_dyntick_drain, cpu) = RCU_NEEDS_CPU_FLUSHES;
-	} else if (--per_cpu(rcu_dyntick_drain, cpu) <= 0) {
+		per_cpu(rcu_dyntick_drain, cpu) = RCU_IDLE_FLUSHES;
+	} else if (per_cpu(rcu_dyntick_drain, cpu) <= RCU_IDLE_OPT_FLUSHES &&
+		   !rcu_pending(cpu)) {
 		/* Can we go dyntick-idle despite still having callbacks? */
-		if (!rcu_pending(cpu)) {
-			trace_rcu_prep_idle("Dyntick with callbacks");
-			per_cpu(rcu_dyntick_holdoff, cpu) = jiffies - 1;
-			hrtimer_start(&per_cpu(rcu_idle_gp_timer, cpu),
-				      rcu_idle_gp_wait, HRTIMER_MODE_REL);
-			return; /* Nothing more to do immediately. */
-		}
-
+		trace_rcu_prep_idle("Dyntick with callbacks");
+		per_cpu(rcu_dyntick_drain, cpu) = 0;
+		per_cpu(rcu_dyntick_holdoff, cpu) = jiffies - 1;
+		hrtimer_start(&per_cpu(rcu_idle_gp_timer, cpu),
+			      rcu_idle_gp_wait, HRTIMER_MODE_REL);
+		return; /* Nothing more to do immediately. */
+	} else if (--per_cpu(rcu_dyntick_drain, cpu) <= 0) {
 		/* We have hit the limit, so time to give up. */
 		per_cpu(rcu_dyntick_holdoff, cpu) = jiffies;
 		local_irq_restore(flags);