Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit fa3b6ddc authored by Siddha, Suresh B's avatar Siddha, Suresh B Committed by Linus Torvalds
Browse files

[PATCH] sched: don't kick ALB in the presence of pinned task 



Jack Steiner brought this issue at my OLS talk.

Take a scenario where two tasks are pinned to two HT threads in a physical
package.  Idle packages in the system will keep kicking migration_thread on
the busy package with out any success.

We will run into similar scenarios in the presence of CMP/NUMA.

Signed-off-by: default avatarSuresh Siddha <suresh.b.siddha@intel.com>
Acked-by: default avatarIngo Molnar <mingo@elte.hu>
Signed-off-by: default avatarAndrew Morton <akpm@osdl.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@osdl.org>
parent 5927ad78

kernel/sched.c

+14 −1
Original line number Original line Diff line number Diff line
@@ -2125,6 +2125,16 @@ static int load_balance(int this_cpu, runqueue_t *this_rq, 
		if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) {

		if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) {





			spin_lock(&busiest->lock);

			spin_lock(&busiest->lock);



			/* don't kick the migration_thread, if the curr

			 * task on busiest cpu can't be moved to this_cpu

			 */

			if (!cpu_isset(this_cpu, busiest->curr->cpus_allowed)) {

				spin_unlock(&busiest->lock);

				all_pinned = 1;

				goto out_one_pinned;

			}



			if (!busiest->active_balance) {

			if (!busiest->active_balance) {

				busiest->active_balance = 1;

				busiest->active_balance = 1;

				busiest->push_cpu = this_cpu;

				busiest->push_cpu = this_cpu;
@@ -2165,6 +2175,8 @@ static int load_balance(int this_cpu, runqueue_t *this_rq, 
	schedstat_inc(sd, lb_balanced[idle]);

	schedstat_inc(sd, lb_balanced[idle]);





	sd->nr_balance_failed = 0;

	sd->nr_balance_failed = 0;



out_one_pinned:

	/* tune up the balancing interval */

	/* tune up the balancing interval */

	if ((all_pinned && sd->balance_interval < MAX_PINNED_INTERVAL) ||

	if ((all_pinned && sd->balance_interval < MAX_PINNED_INTERVAL) ||

			(sd->balance_interval < sd->max_interval))

			(sd->balance_interval < sd->max_interval))
@@ -2357,7 +2369,8 @@ static void rebalance_tick(int this_cpu, runqueue_t *this_rq, 




		if (j - sd->last_balance >= interval) {

		if (j - sd->last_balance >= interval) {

			if (load_balance(this_cpu, this_rq, sd, idle)) {

			if (load_balance(this_cpu, this_rq, sd, idle)) {

				/* We've pulled tasks over so either we're no

				/*

				 * We've pulled tasks over so either we're no

				 * longer idle, or one of our SMT siblings is

				 * longer idle, or one of our SMT siblings is

				 * not idle.

				 * not idle.

				 */

				 */