[PATCH] sched: decrease number of load balances
Currently at a particular domain, each cpu in the sched group will do a
load balance at the frequency of balance_interval. More the cores and
threads, more the cpus will be in each sched group at SMP and NUMA domain.
And we endup spending quite a bit of time doing load balancing in those
domains.
Fix this by making only one cpu(first idle cpu or first cpu in the group if
all the cpus are busy) in the sched group do the load balance at that
particular sched domain and this load will slowly percolate down to the
other cpus with in that group(when they do load balancing at lower
domains).
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/kernel/sched.c b/kernel/sched.c
index 15ce772..4e45343 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -428,7 +428,7 @@
* bump this up when changing the output format or the meaning of an existing
* format, so that tools can adapt (or abort)
*/
-#define SCHEDSTAT_VERSION 12
+#define SCHEDSTAT_VERSION 13
static int show_schedstat(struct seq_file *seq, void *v)
{
@@ -466,7 +466,7 @@
seq_printf(seq, "domain%d %s", dcnt++, mask_str);
for (itype = SCHED_IDLE; itype < MAX_IDLE_TYPES;
itype++) {
- seq_printf(seq, " %lu %lu %lu %lu %lu %lu %lu %lu",
+ seq_printf(seq, " %lu %lu %lu %lu %lu %lu %lu %lu %lu",
sd->lb_cnt[itype],
sd->lb_balanced[itype],
sd->lb_failed[itype],
@@ -474,7 +474,8 @@
sd->lb_gained[itype],
sd->lb_hot_gained[itype],
sd->lb_nobusyq[itype],
- sd->lb_nobusyg[itype]);
+ sd->lb_nobusyg[itype],
+ sd->lb_stopbalance[itype]);
}
seq_printf(seq, " %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu\n",
sd->alb_cnt, sd->alb_failed, sd->alb_pushed,
@@ -2249,7 +2250,7 @@
static struct sched_group *
find_busiest_group(struct sched_domain *sd, int this_cpu,
unsigned long *imbalance, enum idle_type idle, int *sd_idle,
- cpumask_t *cpus)
+ cpumask_t *cpus, int *balance)
{
struct sched_group *busiest = NULL, *this = NULL, *group = sd->groups;
unsigned long max_load, avg_load, total_load, this_load, total_pwr;
@@ -2278,10 +2279,14 @@
unsigned long load, group_capacity;
int local_group;
int i;
+ unsigned int balance_cpu = -1, first_idle_cpu = 0;
unsigned long sum_nr_running, sum_weighted_load;
local_group = cpu_isset(this_cpu, group->cpumask);
+ if (local_group)
+ balance_cpu = first_cpu(group->cpumask);
+
/* Tally up the load of all CPUs in the group */
sum_weighted_load = sum_nr_running = avg_load = 0;
@@ -2297,9 +2302,14 @@
*sd_idle = 0;
/* Bias balancing toward cpus of our domain */
- if (local_group)
+ if (local_group) {
+ if (idle_cpu(i) && !first_idle_cpu) {
+ first_idle_cpu = 1;
+ balance_cpu = i;
+ }
+
load = target_load(i, load_idx);
- else
+ } else
load = source_load(i, load_idx);
avg_load += load;
@@ -2307,6 +2317,16 @@
sum_weighted_load += rq->raw_weighted_load;
}
+ /*
+ * First idle cpu or the first cpu(busiest) in this sched group
+ * is eligible for doing load balancing at this and above
+ * domains.
+ */
+ if (local_group && balance_cpu != this_cpu && balance) {
+ *balance = 0;
+ goto ret;
+ }
+
total_load += avg_load;
total_pwr += group->cpu_power;
@@ -2498,8 +2518,8 @@
*imbalance = min_load_per_task;
return group_min;
}
-ret:
#endif
+ret:
*imbalance = 0;
return NULL;
}
@@ -2550,7 +2570,8 @@
* tasks if there is an imbalance.
*/
static int load_balance(int this_cpu, struct rq *this_rq,
- struct sched_domain *sd, enum idle_type idle)
+ struct sched_domain *sd, enum idle_type idle,
+ int *balance)
{
int nr_moved, all_pinned = 0, active_balance = 0, sd_idle = 0;
struct sched_group *group;
@@ -2573,7 +2594,13 @@
redo:
group = find_busiest_group(sd, this_cpu, &imbalance, idle, &sd_idle,
- &cpus);
+ &cpus, balance);
+
+ if (*balance == 0) {
+ schedstat_inc(sd, lb_stopbalance[idle]);
+ goto out_balanced;
+ }
+
if (!group) {
schedstat_inc(sd, lb_nobusyg[idle]);
goto out_balanced;
@@ -2715,7 +2742,7 @@
schedstat_inc(sd, lb_cnt[NEWLY_IDLE]);
redo:
group = find_busiest_group(sd, this_cpu, &imbalance, NEWLY_IDLE,
- &sd_idle, &cpus);
+ &sd_idle, &cpus, NULL);
if (!group) {
schedstat_inc(sd, lb_nobusyg[NEWLY_IDLE]);
goto out_balanced;
@@ -2885,7 +2912,7 @@
static void run_rebalance_domains(struct softirq_action *h)
{
- int this_cpu = smp_processor_id();
+ int this_cpu = smp_processor_id(), balance = 1;
struct rq *this_rq = cpu_rq(this_cpu);
unsigned long interval;
struct sched_domain *sd;
@@ -2917,7 +2944,7 @@
}
if (time_after_eq(jiffies, sd->last_balance + interval)) {
- if (load_balance(this_cpu, this_rq, sd, idle)) {
+ if (load_balance(this_cpu, this_rq, sd, idle, &balance)) {
/*
* We've pulled tasks over so either we're no
* longer idle, or one of our SMT siblings is
@@ -2932,6 +2959,14 @@
out:
if (time_after(next_balance, sd->last_balance + interval))
next_balance = sd->last_balance + interval;
+
+ /*
+ * Stop the load balance at this level. There is another
+ * CPU in our sched group which is doing load balancing more
+ * actively.
+ */
+ if (!balance)
+ break;
}
this_rq->next_balance = next_balance;
}
