sched: Nominate a power-efficient ilb in select_nohz_balancer()

The CPU that first goes idle becomes the idle-load-balancer and remains
that until either it picks up a task or till all the CPUs of the system
goes idle.

Optimize this further to allow it to relinquish it's post
once all it's siblings in the power-aware sched_domain go idle, thereby
allowing the whole package-core to go idle. While relinquising the post,
nominate another an idle-load balancer from a semi-idle core/package.

Signed-off-by: Gautham R Shenoy <ego@in.ibm.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <20090414045535.7645.31641.stgit@sofia.in.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

1 files changed, 17 insertions, 1 deletions

diff --git a/kernel/sched.c b/kernel/sched.c
index b0fefa300b40..36d213bca473 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -4414,8 +4414,24 @@ int select_nohz_load_balancer(int stop_tick)
                         /* make me the ilb owner */
                         if (atomic_cmpxchg(&nohz.load_balancer, -1, cpu) == -1)
                                 return 1;
-                } else if (atomic_read(&nohz.load_balancer) == cpu)
+                } else if (atomic_read(&nohz.load_balancer) == cpu) {
+                        int new_ilb;
+
+                        if (!(sched_smt_power_savings ||
+                                                sched_mc_power_savings))
+                                return 1;
+                        /*
+                         * Check to see if there is a more power-efficient
+                         * ilb.
+                         */
+                        new_ilb = find_new_ilb(cpu);
+                        if (new_ilb < nr_cpu_ids && new_ilb != cpu) {
+                                atomic_set(&nohz.load_balancer, -1);
+                                resched_cpu(new_ilb);
+                                return 0;
+                        }
                         return 1;
+                }
         } else {
                 if (!cpumask_test_cpu(cpu, nohz.cpu_mask))
                         return 0;