Since rq->nr_running would include both migration and rt task, it is not
reasonable to seek to move nr_running number of task in the load_balance
function, since it only apply to cfs type.
Change it to cfs's h_nr_running, which could well present the task
number in current cfs queue.
Signed-off-by: Lei Wen <lei...@marvell.com>
---
kernel/sched/fair.c | 8 +++++---
1 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index f918635..d6153c8 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -5096,17 +5096,19 @@ redo:
schedstat_add(sd, lb_imbalance[idle], env.imbalance);
ld_moved = 0;
- if (busiest->nr_running > 1) {
+ /* load balance only apply to CFS task, we use h_nr_running here */
+ if (busiest->cfs.h_nr_running > 1) {
/*
* Attempt to move tasks. If find_busiest_group has found
- * an imbalance but busiest->nr_running <= 1, the group is
+ * an imbalance but busiest->cfs.h_nr_running <= 1, the group is
* still unbalanced. ld_moved simply stays zero, so it is
* correctly treated as an imbalance.
*/
env.flags |= LBF_ALL_PINNED;
env.src_cpu = busiest->cpu;
env.src_rq = busiest;
- env.loop_max = min(sysctl_sched_nr_migrate,
busiest->nr_running);
+ env.loop_max = min(sysctl_sched_nr_migrate,
+ busiest->cfs.h_nr_running);
update_h_load(env.src_cpu);
more_balance:
--
1.7.5.4
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
