sched/fair: Add eas (& cas) specific rq, sd and task stats

The statistic counter are placed in the eas (& cas) wakeup path. Each
of them has one representation for the runqueue (rq), the sched_domain
(sd) and the task.
A task counter is always incremented. A rq counter is always
incremented for the rq the scheduler is currently running on. A sd
counter is only incremented if a relation to a sd exists.

The counters are exposed:

(1) In /proc/schedstat for rq's and sd's:

$ cat /proc/schedstat
...
cpu0 71422 0 2321254 ...
eas  44144 0 0 19446 0 24698 568435 51621 156932 133 222011 17459 120279 516814 83 0 156962 359235 176439 139981
  <- runqueue for cpu0
...
domain0 3 42430 42331 ...
eas 0 0 0 14200 0 0 0 0 0 0 0 0 0 0 0 0 0 0 66355 0  <- MC sched domain for cpu0
...

The per-cpu eas vector has the following elements:

sis_attempts  sis_idle   sis_cache_affine sis_suff_cap    sis_idle_cpu    sis_count               ||
secb_attempts secb_sync  secb_idle_bt     secb_insuff_cap secb_no_nrg_sav secb_nrg_sav secb_count ||
fbt_attempts  fbt_no_cpu fbt_no_sd        fbt_pref_idle   fbt_count                               ||
cas_attempts  cas_count

The following relations exist between these counters (from cpu0 eas
vector above):

sis_attempts = sis_idle + sis_cache_affine + sis_suff_cap + sis_idle_cpu + sis_count

44144        = 0        + 0                + 19446        + 0            + 24698

secb_attempts = secb_sync + secb_idle_bt + secb_insuff_cap + secb_no_nrg_sav + secb_nrg_sav + secb_count

568435        = 51621     + 156932       + 133             + 222011          + 17459        + 120279

fbt_attempts = fbt_no_cpu + fbt_no_sd + fbt_pref_idle + fbt_count + (return -1)

516814       = 83         + 0         + 156962        + 359235    + (534)

cas_attempts = cas_count + (return -1 or smp_processor_id())

176439       = 139981    + (36458)

(2) In /proc/$PROCESS_PID/task/$TASK_PID/sched for a task.

example: main thread of system_server

$ cat /proc/1083/task/1083/sched

...
se.statistics.nr_wakeups_sis_attempts        :                  945
se.statistics.nr_wakeups_sis_idle            :                    0
se.statistics.nr_wakeups_sis_cache_affine    :                    0
se.statistics.nr_wakeups_sis_suff_cap        :                  219
se.statistics.nr_wakeups_sis_idle_cpu        :                    0
se.statistics.nr_wakeups_sis_count           :                  726
se.statistics.nr_wakeups_secb_attempts       :                10376
se.statistics.nr_wakeups_secb_sync           :                 1462
se.statistics.nr_wakeups_secb_idle_bt        :                 6984
se.statistics.nr_wakeups_secb_insuff_cap     :                    3
se.statistics.nr_wakeups_secb_no_nrg_sav     :                  927
se.statistics.nr_wakeups_secb_nrg_sav        :                  206
se.statistics.nr_wakeups_secb_count          :                  794
se.statistics.nr_wakeups_fbt_attempts        :                 8914
se.statistics.nr_wakeups_fbt_no_cpu          :                    0
se.statistics.nr_wakeups_fbt_no_sd           :                    0
se.statistics.nr_wakeups_fbt_pref_idle       :                 6987
se.statistics.nr_wakeups_fbt_count           :                 1554
se.statistics.nr_wakeups_cas_attempts        :                 3107
se.statistics.nr_wakeups_cas_count           :                 1195
...

The same relation between the counters as in the per-cpu case apply.

Change-Id: Ie7d01267c78a3f41f60a3ef52917d5a5d463f195
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Chris Redpath <chris.redpath@arm.com>

4 files changed, 165 insertions, 36 deletions

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 641511771ae6..7f7116622631 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -597,6 +597,32 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m)
 	P(se.statistics.nr_wakeups_affine_attempts);
 	P(se.statistics.nr_wakeups_passive);
 	P(se.statistics.nr_wakeups_idle);
+	/* eas */
+	/* select_idle_sibling() */
+	P(se.statistics.nr_wakeups_sis_attempts);
+	P(se.statistics.nr_wakeups_sis_idle);
+	P(se.statistics.nr_wakeups_sis_cache_affine);
+	P(se.statistics.nr_wakeups_sis_suff_cap);
+	P(se.statistics.nr_wakeups_sis_idle_cpu);
+	P(se.statistics.nr_wakeups_sis_count);
+	/* select_energy_cpu_brute() */
+	P(se.statistics.nr_wakeups_secb_attempts);
+	P(se.statistics.nr_wakeups_secb_sync);
+	P(se.statistics.nr_wakeups_secb_idle_bt);
+	P(se.statistics.nr_wakeups_secb_insuff_cap);
+	P(se.statistics.nr_wakeups_secb_no_nrg_sav);
+	P(se.statistics.nr_wakeups_secb_nrg_sav);
+	P(se.statistics.nr_wakeups_secb_count);
+	/* find_best_target() */
+	P(se.statistics.nr_wakeups_fbt_attempts);
+	P(se.statistics.nr_wakeups_fbt_no_cpu);
+	P(se.statistics.nr_wakeups_fbt_no_sd);
+	P(se.statistics.nr_wakeups_fbt_pref_idle);
+	P(se.statistics.nr_wakeups_fbt_count);
+	/* cas */
+	/* select_task_rq_fair() */
+	P(se.statistics.nr_wakeups_cas_attempts);
+	P(se.statistics.nr_wakeups_cas_count);
 
 	{
 		u64 avg_atom, avg_per_cpu;
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index c84e15e3cd56..dfd150b02e63 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -6059,15 +6059,24 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target)
 	int best_idle_cstate = INT_MAX;
 	unsigned long best_idle_capacity = ULONG_MAX;
 
+	schedstat_inc(p, se.statistics.nr_wakeups_sis_attempts);
+	schedstat_inc(this_rq(), eas_stats.sis_attempts);
+
 	if (!sysctl_sched_cstate_aware) {
-		if (idle_cpu(target))
+		if (idle_cpu(target)) {
+			schedstat_inc(p, se.statistics.nr_wakeups_sis_idle);
+			schedstat_inc(this_rq(), eas_stats.sis_idle);
 			return target;
+		}
 
 		/*
 		 * If the prevous cpu is cache affine and idle, don't be stupid.
 		 */
-		if (prev != target && cpus_share_cache(prev, target) && idle_cpu(prev))
+		if (prev != target && cpus_share_cache(prev, target) && idle_cpu(prev)) {
+			schedstat_inc(p, se.statistics.nr_wakeups_sis_cache_affine);
+			schedstat_inc(this_rq(), eas_stats.sis_cache_affine);
 			return prev;
+		}
 	}
 
 	/*
@@ -6091,8 +6100,12 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target)
 					if (new_usage > capacity_orig || !idle_cpu(i))
 						goto next;
 
-					if (i == target && new_usage <= capacity_curr_of(target))
+					if (i == target && new_usage <= capacity_curr_of(target)) {
+						schedstat_inc(p, se.statistics.nr_wakeups_sis_suff_cap);
+						schedstat_inc(this_rq(), eas_stats.sis_suff_cap);
+						schedstat_inc(sd, eas_stats.sis_suff_cap);
 						return target;
+					}
 
 					if (idle_idx < best_idle_cstate &&
 					    capacity_orig <= best_idle_capacity) {
@@ -6109,6 +6122,9 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target)
 
 				target = cpumask_first_and(sched_group_cpus(sg),
 					tsk_cpus_allowed(p));
+				schedstat_inc(p, se.statistics.nr_wakeups_sis_idle_cpu);
+				schedstat_inc(this_rq(), eas_stats.sis_idle_cpu);
+				schedstat_inc(sd, eas_stats.sis_idle_cpu);
 				goto done;
 			}
 next:
@@ -6120,6 +6136,9 @@ next:
 		target = best_idle_cpu;
 
 done:
+	schedstat_inc(p, se.statistics.nr_wakeups_sis_count);
+	schedstat_inc(this_rq(), eas_stats.sis_count);
+
 	return target;
 }
 
@@ -6146,13 +6165,22 @@ static inline int find_best_target(struct task_struct *p, bool boosted, bool pre
 	struct sched_group *sg;
 	int cpu = start_cpu(boosted);
 
-	if (cpu < 0)
+	schedstat_inc(p, se.statistics.nr_wakeups_fbt_attempts);
+	schedstat_inc(this_rq(), eas_stats.fbt_attempts);
+
+	if (cpu < 0) {
+		schedstat_inc(p, se.statistics.nr_wakeups_fbt_no_cpu);
+		schedstat_inc(this_rq(), eas_stats.fbt_no_cpu);
 		return target_cpu;
+	}
 
 	sd = rcu_dereference(per_cpu(sd_ea, cpu));
 
-	if (!sd)
+	if (!sd) {
+		schedstat_inc(p, se.statistics.nr_wakeups_fbt_no_sd);
+		schedstat_inc(this_rq(), eas_stats.fbt_no_sd);
 		return target_cpu;
+	}
 
 	sg = sd->groups;
 
@@ -6191,8 +6219,11 @@ static inline int find_best_target(struct task_struct *p, bool boosted, bool pre
 			 * Unconditionally favoring tasks that prefer idle cpus to
 			 * improve latency.
 			 */
-			if (idle_cpu(i) && prefer_idle)
+			if (idle_cpu(i) && prefer_idle) {
+				schedstat_inc(p, se.statistics.nr_wakeups_fbt_pref_idle);
+				schedstat_inc(this_rq(), eas_stats.fbt_pref_idle);
 				return i;
+			}
 
 			cur_capacity = capacity_curr_of(i);
 
@@ -6228,6 +6259,11 @@ static inline int find_best_target(struct task_struct *p, bool boosted, bool pre
 	if (target_cpu < 0)
 		target_cpu = best_idle_cpu >= 0 ? best_idle_cpu : backup_cpu;
 
+	if (target_cpu >= 0) {
+		schedstat_inc(p, se.statistics.nr_wakeups_fbt_count);
+		schedstat_inc(this_rq(), eas_stats.fbt_count);
+	}
+
 	return target_cpu;
 }
 
@@ -6279,11 +6315,17 @@ static int select_energy_cpu_brute(struct task_struct *p, int prev_cpu, int sync
 	int target_cpu = prev_cpu, tmp_target;
 	bool boosted, prefer_idle;
 
+	schedstat_inc(p, se.statistics.nr_wakeups_secb_attempts);
+	schedstat_inc(this_rq(), eas_stats.secb_attempts);
+
 	if (sysctl_sched_sync_hint_enable && sync) {
 		int cpu = smp_processor_id();
 
-		if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p)))
+		if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p))) {
+			schedstat_inc(p, se.statistics.nr_wakeups_secb_sync);
+			schedstat_inc(this_rq(), eas_stats.secb_sync);
 			return cpu;
+		}
 	}
 
 	rcu_read_lock();
@@ -6303,8 +6345,11 @@ static int select_energy_cpu_brute(struct task_struct *p, int prev_cpu, int sync
 		goto unlock;
 	if (tmp_target >= 0) {
 		target_cpu = tmp_target;
-		if ((boosted || prefer_idle) && idle_cpu(target_cpu))
+		if ((boosted || prefer_idle) && idle_cpu(target_cpu)) {
+			schedstat_inc(p, se.statistics.nr_wakeups_secb_idle_bt);
+			schedstat_inc(this_rq(), eas_stats.secb_idle_bt);
 			goto unlock;
+		}
 	}
 
 	if (target_cpu != prev_cpu) {
@@ -6316,15 +6361,30 @@ static int select_energy_cpu_brute(struct task_struct *p, int prev_cpu, int sync
 		};
 
 		/* Not enough spare capacity on previous cpu */
-		if (cpu_overutilized(prev_cpu))
+		if (cpu_overutilized(prev_cpu)) {
+			schedstat_inc(p, se.statistics.nr_wakeups_secb_insuff_cap);
+			schedstat_inc(this_rq(), eas_stats.secb_insuff_cap);
 			goto unlock;
+		}
 
-		if (energy_diff(&eenv) >= 0)
+		if (energy_diff(&eenv) >= 0) {
+			schedstat_inc(p, se.statistics.nr_wakeups_secb_no_nrg_sav);
+			schedstat_inc(this_rq(), eas_stats.secb_no_nrg_sav);
 			target_cpu = prev_cpu;
+			goto unlock;
+		}
+
+		schedstat_inc(p, se.statistics.nr_wakeups_secb_nrg_sav);
+		schedstat_inc(this_rq(), eas_stats.secb_nrg_sav);
+		goto unlock;
 	}
 
+	schedstat_inc(p, se.statistics.nr_wakeups_secb_count);
+	schedstat_inc(this_rq(), eas_stats.secb_count);
+
 unlock:
 	rcu_read_unlock();
+
 	return target_cpu;
 }
 
@@ -6387,39 +6447,57 @@ select_task_rq_fair(struct task_struct *p, int prev_cpu, int sd_flag, int wake_f
 		if (sd_flag & SD_BALANCE_WAKE) /* XXX always ? */
 			new_cpu = select_idle_sibling(p, prev_cpu, new_cpu);
 
-	} else while (sd) {
-		struct sched_group *group;
-		int weight;
+	} else {
+		int wu = sd_flag & SD_BALANCE_WAKE;
+		int cas_cpu = -1;
 
-		if (!(sd->flags & sd_flag)) {
-			sd = sd->child;
-			continue;
+		if (wu) {
+			schedstat_inc(p, se.statistics.nr_wakeups_cas_attempts);
+			schedstat_inc(this_rq(), eas_stats.cas_attempts);
 		}
 
-		group = find_idlest_group(sd, p, cpu, sd_flag);
-		if (!group) {
-			sd = sd->child;
-			continue;
-		}
+		while (sd) {
+			struct sched_group *group;
+			int weight;
 
-		new_cpu = find_idlest_cpu(group, p, cpu);
-		if (new_cpu == -1 || new_cpu == cpu) {
-			/* Now try balancing at a lower domain level of cpu */
-			sd = sd->child;
-			continue;
+			if (wu)
+				schedstat_inc(sd, eas_stats.cas_attempts);
+
+			if (!(sd->flags & sd_flag)) {
+				sd = sd->child;
+				continue;
+			}
+
+			group = find_idlest_group(sd, p, cpu, sd_flag);
+			if (!group) {
+				sd = sd->child;
+				continue;
+			}
+
+			new_cpu = find_idlest_cpu(group, p, cpu);
+			if (new_cpu == -1 || new_cpu == cpu) {
+				/* Now try balancing at a lower domain level of cpu */
+				sd = sd->child;
+				continue;
+			}
+
+			/* Now try balancing at a lower domain level of new_cpu */
+			cpu = cas_cpu = new_cpu;
+			weight = sd->span_weight;
+			sd = NULL;
+			for_each_domain(cpu, tmp) {
+				if (weight <= tmp->span_weight)
+					break;
+				if (tmp->flags & sd_flag)
+					sd = tmp;
+			}
+			/* while loop will break here if sd == NULL */
 		}
 
-		/* Now try balancing at a lower domain level of new_cpu */
-		cpu = new_cpu;
-		weight = sd->span_weight;
-		sd = NULL;
-		for_each_domain(cpu, tmp) {
-			if (weight <= tmp->span_weight)
-				break;
-			if (tmp->flags & sd_flag)
-				sd = tmp;
+		if (wu && (cas_cpu >= 0)) {
+			schedstat_inc(p, se.statistics.nr_wakeups_cas_count);
+			schedstat_inc(this_rq(), eas_stats.cas_count);
 		}
-		/* while loop will break here if sd == NULL */
 	}
 	rcu_read_unlock();
 
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index d0971e5c9269..6880fbc39760 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -734,6 +734,8 @@ struct rq {
 	/* try_to_wake_up() stats */
 	unsigned int ttwu_count;
 	unsigned int ttwu_local;
+
+	struct eas_stats eas_stats;
 #endif
 
 #ifdef CONFIG_SMP
diff --git a/kernel/sched/stats.c b/kernel/sched/stats.c
index 87e2c9f0c33e..b63879918cd6 100644
--- a/kernel/sched/stats.c
+++ b/kernel/sched/stats.c
@@ -12,6 +12,26 @@
  */
 #define SCHEDSTAT_VERSION 15
 
+static inline void show_easstat(struct seq_file *seq, struct eas_stats *stats)
+{
+	/* eas-specific runqueue stats */
+	seq_printf(seq, "eas %llu %llu %llu %llu %llu %llu ",
+	    stats->sis_attempts, stats->sis_idle, stats->sis_cache_affine,
+	    stats->sis_suff_cap, stats->sis_idle_cpu, stats->sis_count);
+
+	seq_printf(seq, "%llu %llu %llu %llu %llu %llu %llu ",
+	    stats->secb_attempts, stats->secb_sync, stats->secb_idle_bt,
+	    stats->secb_insuff_cap, stats->secb_no_nrg_sav,
+	    stats->secb_nrg_sav, stats->secb_count);
+
+	seq_printf(seq, "%llu %llu %llu %llu %llu ",
+	    stats->fbt_attempts, stats->fbt_no_cpu, stats->fbt_no_sd,
+	    stats->fbt_pref_idle, stats->fbt_count);
+
+	seq_printf(seq, "%llu %llu\n",
+	    stats->cas_attempts, stats->cas_count);
+}
+
 static int show_schedstat(struct seq_file *seq, void *v)
 {
 	int cpu;
@@ -39,6 +59,7 @@ static int show_schedstat(struct seq_file *seq, void *v)
 
 		seq_printf(seq, "\n");
 
+		show_easstat(seq, &rq->eas_stats);
 #ifdef CONFIG_SMP
 		/* domain-specific stats */
 		rcu_read_lock();
@@ -66,6 +87,8 @@ static int show_schedstat(struct seq_file *seq, void *v)
 			    sd->sbf_count, sd->sbf_balanced, sd->sbf_pushed,
 			    sd->ttwu_wake_remote, sd->ttwu_move_affine,
 			    sd->ttwu_move_balance);
+
+			show_easstat(seq, &sd->eas_stats);
 		}
 		rcu_read_unlock();
 #endif