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

Commit 0d6eeac4 authored by Morten Rasmussen's avatar Morten Rasmussen Committed by Lingutla Chandrasekhar
Browse files

ANDROID: sched/fair: Avoid unnecessary balancing of asymmetric capacity groups 



On systems with asymmetric cpu capacities, a skewed load distribution
might yield better throughput than balancing load per group capacity.
For example, running compute intensive tasks on high capacity cpus while
leaving low capacity cpus idle. So we let load-balance back off if the
busiest group isn't really overloaded.

cc: Ingo Molnar <mingo@redhat.com>
cc: Peter Zijlstra <peterz@infradead.org>

Signed-off-by: default avatarMorten Rasmussen <morten.rasmussen@arm.com>
Change-Id: I8b08a0fa73f357a9972324bc76cec3912fe293cf
Signed-off-by: default avatarChris Redpath <chris.redpath@arm.com>
Git-commit: c42f9795
Git-repo: https://android.googlesource.com/kernel/common/


Signed-off-by: default avatarSatya Durga Srinivasu Prabhala <satyap@codeaurora.org>
[cslingutla@codeaurora.org: Took partial commit '5494e2ed ("ANDROID:
  sched: Consider misfit tasks when load-balancing")', which defined
  function group_similar_cpu_capacity.]
Signed-off-by: default avatarLingutla Chandrasekhar <clingutla@codeaurora.org>
parent 118a7c02

kernel/sched/fair.c

+22 −0
Original line number Diff line number Diff line
@@ -9146,6 +9146,19 @@ group_smaller_max_cpu_capacity(struct sched_group *sg, struct sched_group *ref) 
						ref->sgc->max_capacity * 1024;

}



/*

 * group_similar_cpu_capacity: Returns true if the minimum capacity of the

 * compared groups differ by less than 12.5%.

 */

static inline bool

group_similar_cpu_capacity(struct sched_group *sg, struct sched_group *ref)

{

	long diff = sg->sgc->min_capacity - ref->sgc->min_capacity;

	long max = max(sg->sgc->min_capacity, ref->sgc->min_capacity);



	return abs(diff) < max >> 3;

}



static inline enum

group_type group_classify(struct sched_group *group,

			  struct sg_lb_stats *sgs)
@@ -9324,6 +9337,15 @@ static bool update_sd_pick_busiest(struct lb_env *env, 
	    group_smaller_min_cpu_capacity(sds->local, sg))

		return false;



	/*

	 * Candidate sg doesn't face any severe imbalance issues so

	 * don't disturb unless the groups are of similar capacity

	 * where balancing is more harmless.

	 */

	if (sgs->group_type == group_other &&

		!group_similar_cpu_capacity(sds->local, sg))

		return false;



	/*

	 * If we have more than one misfit sg go with the biggest misfit.

	 */