sched: cpufreq: Allow remote cpufreq callbacks

With Android UI and benchmarks the latency of cpufreq response to
certain scheduling events can become very critical. Currently, callbacks
into cpufreq governors are only made from the scheduler if the target
CPU of the event is the same as the current CPU. This means there are
certain situations where a target CPU may not run the cpufreq governor
for some time.

One testcase to show this behavior is where a task starts running on
CPU0, then a new task is also spawned on CPU0 by a task on CPU1. If the
system is configured such that the new tasks should receive maximum
demand initially, this should result in CPU0 increasing frequency
immediately. But because of the above mentioned limitation though, this
does not occur.

This patch updates the scheduler core to call the cpufreq callbacks for
remote CPUs as well.

The schedutil, ondemand and conservative governors are updated to
process cpufreq utilization update hooks called for remote CPUs where
the remote CPU is managed by the cpufreq policy of the local CPU.

The intel_pstate driver is updated to always reject remote callbacks.

This is tested with couple of usecases (Android: hackbench, recentfling,
galleryfling, vellamo, Ubuntu: hackbench) on ARM hikey board (64 bit
octa-core, single policy). Only galleryfling showed minor improvements,
while others didn't had much deviation.

The reason being that this patch only targets a corner case, where
following are required to be true to improve performance and that
doesn't happen too often with these tests:

- Task is migrated to another CPU.
- The task has high demand, and should take the target CPU to higher
OPPs.
- And the target CPU doesn't call into the cpufreq governor until the
next tick.

Based on initial work from Steve Muckle.

Signed-off-by: Viresh Kumar <[e-mail]>
Acked-by: Saravana Kannan <[e-mail]>
Acked-by: Peter Zijlstra (Intel) <[e-mail]>
Signed-off-by: Rafael J. Wysocki <[e-mail]>

drivers/cpufreq/cpufreq_governor.c

···
 	struct policy_dbs_info *policy_dbs = cdbs->policy_dbs;
 	u64 delta_ns, lst;
 
+	if (!cpufreq_can_do_remote_dvfs(policy_dbs->policy))
+		return;
+
 	/*
 	 * The work may not be allowed to be queued up right now.
 	 * Possible reasons:

drivers/cpufreq/intel_pstate.c

···
 	struct cpudata *cpu = container_of(data, struct cpudata, update_util);
 	u64 delta_ns = time - cpu->sample.time;
 
+	/* Don't allow remote callbacks */
+	if (smp_processor_id() != cpu->cpu)
+		return;
+
 	if ((s64)delta_ns < pid_params.sample_rate_ns)
 		return;
 
···
 {
 	struct cpudata *cpu = container_of(data, struct cpudata, update_util);
 	u64 delta_ns;
+
+	/* Don't allow remote callbacks */
+	if (smp_processor_id() != cpu->cpu)
+		return;
 
 	if (flags & SCHED_CPUFREQ_IOWAIT) {
 		cpu->iowait_boost = int_tofp(1);

include/linux/cpufreq.h

···
 			 size_t count);
 };
 
+static inline bool cpufreq_can_do_remote_dvfs(struct cpufreq_policy *policy)
+{
+	/* Allow remote callbacks only on the CPUs sharing cpufreq policy */
+	if (cpumask_test_cpu(smp_processor_id(), policy->cpus))
+		return true;
+
+	return false;
+}
+
 /*********************************************************************
  *                     FREQUENCY TABLE HELPERS                       *
  *********************************************************************/

kernel/sched/cpufreq_schedutil.c

···
 struct sugov_cpu {
 	struct update_util_data update_util;
 	struct sugov_policy *sg_policy;
+	unsigned int cpu;
 
 	bool iowait_boost_pending;
 	unsigned int iowait_boost;
···
 static bool sugov_should_update_freq(struct sugov_policy *sg_policy, u64 time)
 {
 	s64 delta_ns;
+
+	/*
+	 * Since cpufreq_update_util() is called with rq->lock held for
+	 * the @target_cpu, our per-cpu data is fully serialized.
+	 *
+	 * However, drivers cannot in general deal with cross-cpu
+	 * requests, so while get_next_freq() will work, our
+	 * sugov_update_commit() call may not.
+	 *
+	 * Hence stop here for remote requests if they aren't supported
+	 * by the hardware, as calculating the frequency is pointless if
+	 * we cannot in fact act on it.
+	 */
+	if (!cpufreq_can_do_remote_dvfs(sg_policy->policy))
+		return false;
 
 	if (sg_policy->work_in_progress)
 		return false;
···
 	return cpufreq_driver_resolve_freq(policy, freq);
 }
 
-static void sugov_get_util(unsigned long *util, unsigned long *max)
+static void sugov_get_util(unsigned long *util, unsigned long *max, int cpu)
 {
-	struct rq *rq = this_rq();
+	struct rq *rq = cpu_rq(cpu);
 	unsigned long cfs_max;
 
-	cfs_max = arch_scale_cpu_capacity(NULL, smp_processor_id());
+	cfs_max = arch_scale_cpu_capacity(NULL, cpu);
 
 	*util = min(rq->cfs.avg.util_avg, cfs_max);
 	*max = cfs_max;
···
 	if (flags & SCHED_CPUFREQ_RT_DL) {
 		next_f = policy->cpuinfo.max_freq;
 	} else {
-		sugov_get_util(&util, &max);
+		sugov_get_util(&util, &max, sg_cpu->cpu);
 		sugov_iowait_boost(sg_cpu, &util, &max);
 		next_f = get_next_freq(sg_policy, util, max);
 		/*
···
 	unsigned long util, max;
 	unsigned int next_f;
 
-	sugov_get_util(&util, &max);
+	sugov_get_util(&util, &max, sg_cpu->cpu);
 
 	raw_spin_lock(&sg_policy->update_lock);
 
···
 
 static int __init sugov_register(void)
 {
+	int cpu;
+
+	for_each_possible_cpu(cpu)
+		per_cpu(sugov_cpu, cpu).cpu = cpu;
+
 	return cpufreq_register_governor(&schedutil_gov);
 }
 fs_initcall(sugov_register);

kernel/sched/deadline.c

···
 	}
 
 	/* kick cpufreq (see the comment in kernel/sched/sched.h). */
-	cpufreq_update_this_cpu(rq, SCHED_CPUFREQ_DL);
+	cpufreq_update_util(rq, SCHED_CPUFREQ_DL);
 
 	schedstat_set(curr->se.statistics.exec_max,
 		      max(curr->se.statistics.exec_max, delta_exec));

kernel/sched/fair.c

···
 
 static inline void cfs_rq_util_change(struct cfs_rq *cfs_rq)
 {
-	if (&this_rq()->cfs == cfs_rq) {
+	struct rq *rq = rq_of(cfs_rq);
+
+	if (&rq->cfs == cfs_rq) {
 		/*
 		 * There are a few boundary cases this might miss but it should
 		 * get called often enough that that should (hopefully) not be
···
 		 *
 		 * See cpu_util().
 		 */
-		cpufreq_update_util(rq_of(cfs_rq), 0);
+		cpufreq_update_util(rq, 0);
 	}
 }
 
···
 	 * passed.
 	 */
 	if (p->in_iowait)
-		cpufreq_update_this_cpu(rq, SCHED_CPUFREQ_IOWAIT);
+		cpufreq_update_util(rq, SCHED_CPUFREQ_IOWAIT);
 
 	for_each_sched_entity(se) {
 		if (se->on_rq)

kernel/sched/rt.c

···
 		return;
 
 	/* Kick cpufreq (see the comment in kernel/sched/sched.h). */
-	cpufreq_update_this_cpu(rq, SCHED_CPUFREQ_RT);
+	cpufreq_update_util(rq, SCHED_CPUFREQ_RT);
 
 	schedstat_set(curr->se.statistics.exec_max,
 		      max(curr->se.statistics.exec_max, delta_exec));

kernel/sched/sched.h

···
 {
 	struct update_util_data *data;
 
-	data = rcu_dereference_sched(*this_cpu_ptr(&cpufreq_update_util_data));
+	data = rcu_dereference_sched(*per_cpu_ptr(&cpufreq_update_util_data,
+						  cpu_of(rq)));
 	if (data)
 		data->func(data, rq_clock(rq), flags);
 }
-
-static inline void cpufreq_update_this_cpu(struct rq *rq, unsigned int flags)
-{
-	if (cpu_of(rq) == smp_processor_id())
-		cpufreq_update_util(rq, flags);
-}
 #else
 static inline void cpufreq_update_util(struct rq *rq, unsigned int flags) {}
-static inline void cpufreq_update_this_cpu(struct rq *rq, unsigned int flags) {}
 #endif /* CONFIG_CPU_FREQ */
 
 #ifdef arch_scale_freq_capacity