This layer provides Energy Aware Scheduling (EAS) patcheseel_5.1.0eel_5.0.3eel_5.0.2eel/5.1.0eel/5.0.3eel/5.0.25.1.05.0.35.0.2eel

For the moment only for Renesas R-Car Gen3 SoC's.
Can be expanded for other SoC's by setting the machine
feature biglittle and provide the relevant EAS patches.

Bug-AGL: SPEC-813

Change-Id: I2b5e69c515c33e57be19b30466fe208d7b8ac1a5
Signed-off-by: Frode Isaksen <fisaksen@baylibre.com>

1 files changed, 104 insertions, 0 deletions

diff --git a/meta-eas/recipes-kernel/linux/linux-renesas/0060-sched-fair-Energy-aware-wake-up-task-placement.patch b/meta-eas/recipes-kernel/linux/linux-renesas/0060-sched-fair-Energy-aware-wake-up-task-placement.patch
new file mode 100644
index 0000000..d85d6fe
--- /dev/null
+++ b/meta-eas/recipes-kernel/linux/linux-renesas/0060-sched-fair-Energy-aware-wake-up-task-placement.patch
@@ -0,0 +1,104 @@
+From b0ba3376b094bcf111551ee7c5e14183fe15da79 Mon Sep 17 00:00:00 2001
+From: Morten Rasmussen <morten.rasmussen@arm.com>
+Date: Wed, 30 Mar 2016 14:29:48 +0100
+Subject: [PATCH 60/92] sched/fair: Energy-aware wake-up task placement
+
+When the systems is not overutilized, place waking tasks on the most
+energy efficient cpu. Previous attempts reduced the search space by
+matching task utilization to cpu capacity before consulting the energy
+model as this is an expensive operation. The search heuristics didn't
+work very well and lacking any better alternatives this patch takes the
+brute-force route and tries all potential targets.
+
+This approach doesn't scale, but it might be sufficient for many
+embedded applications while work is continuing on a heuristic that can
+minimize the necessary computations. The heuristic must be derrived from
+the platform energy model rather than make additional assumptions, such
+lower capacity implies better energy efficiency. PeterZ mentioned in the
+past that we might be able to derrive some simpler deciding functions
+using mathematical (modal?) analysis.
+
+Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
+(cherry picked from commit 8491fd68555c31e596e1888912796184d42138cc)
+Signed-off-by: Gaku Inami <gaku.inami.xw@bp.renesas.com>
+---
+ kernel/sched/fair.c | 56 +++++++++++++++++++++++++++++++++++++++++++++++++++++
+ 1 file changed, 56 insertions(+)
+
+diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
+index db732bd..58b468c 100644
+--- a/kernel/sched/fair.c
++++ b/kernel/sched/fair.c
+@@ -6148,6 +6148,59 @@ static bool cpu_overutilized(int cpu)
+ 	return (capacity_of(cpu) * 1024) < (cpu_util(cpu) * capacity_margin);
+ }
+ 
++static int select_energy_cpu_brute(struct task_struct *p, int prev_cpu)
++{
++	int i;
++	int min_diff = 0, energy_cpu = prev_cpu, spare_cpu = prev_cpu;
++	unsigned long max_spare = 0;
++	struct sched_domain *sd;
++
++	rcu_read_lock();
++
++	sd = rcu_dereference(per_cpu(sd_ea, prev_cpu));
++
++	if (!sd)
++		return prev_cpu;
++
++	for_each_cpu_and(i, tsk_cpus_allowed(p), sched_domain_span(sd)) {
++		int diff;
++		unsigned long spare;
++
++		struct energy_env eenv = {
++			.util_delta	= task_util(p),
++			.src_cpu	= prev_cpu,
++			.dst_cpu	= i,
++		};
++
++		spare = capacity_spare_wake(i, p);
++
++		if (i == prev_cpu)
++			continue;
++
++		if (spare > max_spare) {
++			max_spare = spare;
++			spare_cpu = i;
++		}
++
++		if (spare * 1024 < capacity_margin * task_util(p))
++			continue;
++
++		diff = energy_diff(&eenv);
++
++		if (diff < min_diff) {
++			min_diff = diff;
++			energy_cpu = i;
++		}
++	}
++
++	rcu_read_unlock();
++
++	if (energy_cpu == prev_cpu && !cpu_overutilized(prev_cpu))
++		return prev_cpu;
++
++	return energy_cpu != prev_cpu ? energy_cpu : spare_cpu;
++}
++
+ /*
+  * select_task_rq_fair: Select target runqueue for the waking task in domains
+  * that have the 'sd_flag' flag set. In practice, this is SD_BALANCE_WAKE,
+@@ -6175,6 +6228,9 @@ static bool cpu_overutilized(int cpu)
+ 			      && cpumask_test_cpu(cpu, tsk_cpus_allowed(p));
+ 	}
+ 
++	if (energy_aware() && !(cpu_rq(prev_cpu)->rd->overutilized))
++		return select_energy_cpu_brute(p, prev_cpu);
++
+ 	rcu_read_lock();
+ 	for_each_domain(cpu, tmp) {
+ 		if (!(tmp->flags & SD_LOAD_BALANCE))
+-- 
+1.9.1
+