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

Commit de366870 authored by Satya Durga Srinivasu Prabhala's avatar Satya Durga Srinivasu Prabhala
Browse files

sched: sched_avg: Fix sched_get_nr_running_avg for tri-cluster systems 



sched_get_nr_running_avg() sets values for the various arguments
passed to it and one of them is max number of tasks running
on the max capacity CPUs which breaks tri-cluster systems.
Fix it by updating condition by checking for CPUs other than min
capacity CPUs.

Change-Id: I3f519a7fd1ee89796818e03967abe77f9a72cbb8
Signed-off-by: default avatarSatya Durga Srinivasu Prabhala <satyap@codeaurora.org>
parent 07fb2060

kernel/sched/sched_avg.c

+2 −2
Original line number Diff line number Diff line 
/* Copyright (c) 2012, 2015-2017, The Linux Foundation. All rights reserved.

/* Copyright (c) 2012, 2015-2018, The Linux Foundation. All rights reserved.

 *

 * This program is free software; you can redistribute it and/or modify

 * it under the terms of the GNU General Public License version 2 and
@@ -89,7 +89,7 @@ void sched_get_nr_running_avg(int *avg, int *iowait_avg, int *big_avg, 
		if (*max_nr < per_cpu(nr_max, cpu))

			*max_nr = per_cpu(nr_max, cpu);



		if (is_max_capacity_cpu(cpu)) {

		if (!is_min_capacity_cpu(cpu)) {

			if (*big_max_nr < per_cpu(nr_max, cpu))

				*big_max_nr = per_cpu(nr_max, cpu);

		}