The following commit has been merged into the sched/urgent branch of tip:
Commit-ID: 16b0a7a1a0af9db6e008fecd195fe4d6cb366d83
Gitweb: https://git.kernel.org/tip/16b0a7a1a0af9db6e008fecd195fe4d6cb366d83
Author: Vincent Guittot <vincent.guittot@xxxxxxxxxx>
AuthorDate: Mon, 02 Nov 2020 11:24:57 +01:00
Committer: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
CommitterDate: Tue, 10 Nov 2020 18:38:48 +01:00
sched/fair: Ensure tasks spreading in LLC during LB
schbench shows latency increase for 95 percentile above since:
commit 0b0695f2b34a ("sched/fair: Rework load_balance()")
Align the behavior of the load balancer with the wake up path, which tries
to select an idle CPU which belongs to the LLC for a waking task.
calculate_imbalance() will use nr_running instead of the spare
capacity when CPUs share resources (ie cache) at the domain level. This
will ensure a better spread of tasks on idle CPUs.
Running schbench on a hikey (8cores arm64) shows the problem:
tip/sched/core :
schbench -m 2 -t 4 -s 10000 -c 1000000 -r 10
Latency percentiles (usec)
50.0th: 33
75.0th: 45
90.0th: 51
95.0th: 4152
*99.0th: 14288
99.5th: 14288
99.9th: 14288
min=0, max=14276
tip/sched/core + patch :
schbench -m 2 -t 4 -s 10000 -c 1000000 -r 10
Latency percentiles (usec)
50.0th: 34
75.0th: 47
90.0th: 52
95.0th: 78
*99.0th: 94
99.5th: 94
99.9th: 94
min=0, max=94
Fixes: 0b0695f2b34a ("sched/fair: Rework load_balance()")
Reported-by: Chris Mason <clm@xxxxxx>
Suggested-by: Rik van Riel <riel@xxxxxxxxxxx>
Signed-off-by: Vincent Guittot <vincent.guittot@xxxxxxxxxx>
Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Reviewed-by: Rik van Riel <riel@xxxxxxxxxxx>
Tested-by: Rik van Riel <riel@xxxxxxxxxxx>
Link: https://lkml.kernel.org/r/20201102102457.28808-1-vincent.guittot@xxxxxxxxxx
---
kernel/sched/fair.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index aa4c622..210b15f 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -9031,7 +9031,8 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s
* emptying busiest.
*/
if (local->group_type == group_has_spare) {
- if (busiest->group_type > group_fully_busy) {
+ if ((busiest->group_type > group_fully_busy) &&
+ !(env->sd->flags & SD_SHARE_PKG_RESOURCES)) {
/*
* If busiest is overloaded, try to fill spare
* capacity. This might end up creating spare capacity
|