OSDN Git Service

sched: Update capacity and load scale factor for all clusters at boot
authorSyed Rameez Mustafa <rameezmustafa@codeaurora.org>
Sat, 21 Jan 2017 00:23:42 +0000 (16:23 -0800)
committerSyed Rameez Mustafa <rameezmustafa@codeaurora.org>
Sat, 21 Jan 2017 00:58:11 +0000 (16:58 -0800)
Cluster capacities should reflect differences in efficiency of
different clusters even in the absence of cpufreq. Currently
capacity is updated only when cpufreq policy notifier is received.
Therefore placement is suboptimal when cpufreq is turned off. Fix
this by updating capacities and load scaling factors during cluster
detection.

Change-Id: I47f63c1e374bbfd247a4302525afb37d55334bad
Signed-off-by: Syed Rameez Mustafa <rameezmustafa@codeaurora.org>
kernel/sched/hmp.c

index 40df4f8..95125c5 100644 (file)
@@ -590,6 +590,7 @@ void update_cluster_topology(void)
         * cluster_head visible.
         */
        move_list(&cluster_head, &new_head, false);
+       update_all_clusters_stats();
 }
 
 void init_clusters(void)