#define CREATE_TRACE_POINTS
#include <trace/events/sched.h>
+#include "walt.h"
DEFINE_MUTEX(sched_domains_mutex);
DEFINE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues);
/* cpus with isolated domains */
cpumask_var_t cpu_isolated_map;
+struct rq *
+lock_rq_of(struct task_struct *p, unsigned long *flags)
+{
+ return task_rq_lock(p, flags);
+}
+
+void
+unlock_rq_of(struct rq *rq, struct task_struct *p, unsigned long *flags)
+{
+ task_rq_unlock(rq, p, flags);
+}
+
/*
* this_rq_lock - lock this runqueue and disable interrupts.
*/
dequeue_task(rq, p, 0);
p->on_rq = TASK_ON_RQ_MIGRATING;
+ double_lock_balance(rq, cpu_rq(new_cpu));
set_task_cpu(p, new_cpu);
+ double_unlock_balance(rq, cpu_rq(new_cpu));
raw_spin_unlock(&rq->lock);
rq = cpu_rq(new_cpu);
p->sched_class->migrate_task_rq(p);
p->se.nr_migrations++;
perf_event_task_migrate(p);
+
+ walt_fixup_busy_time(p, new_cpu);
}
__set_task_cpu(p, new_cpu);
{
unsigned long flags;
int cpu, success = 0;
+#ifdef CONFIG_SMP
+ struct rq *rq;
+ u64 wallclock;
+#endif
/*
* If we are going to wake up a thread waiting for CONDITION we
success = 1; /* we're going to change ->state */
cpu = task_cpu(p);
+ /*
+ * Ensure we load p->on_rq _after_ p->state, otherwise it would
+ * be possible to, falsely, observe p->on_rq == 0 and get stuck
+ * in smp_cond_load_acquire() below.
+ *
+ * sched_ttwu_pending() try_to_wake_up()
+ * [S] p->on_rq = 1; [L] P->state
+ * UNLOCK rq->lock -----.
+ * \
+ * +--- RMB
+ * schedule() /
+ * LOCK rq->lock -----'
+ * UNLOCK rq->lock
+ *
+ * [task p]
+ * [S] p->state = UNINTERRUPTIBLE [L] p->on_rq
+ *
+ * Pairs with the UNLOCK+LOCK on rq->lock from the
+ * last wakeup of our task and the schedule that got our task
+ * current.
+ */
+ smp_rmb();
if (p->on_rq && ttwu_remote(p, wake_flags))
goto stat;
*/
smp_rmb();
+ rq = cpu_rq(task_cpu(p));
+
+ raw_spin_lock(&rq->lock);
+ wallclock = walt_ktime_clock();
+ walt_update_task_ravg(rq->curr, rq, TASK_UPDATE, wallclock, 0);
+ walt_update_task_ravg(p, rq, TASK_WAKE, wallclock, 0);
+ raw_spin_unlock(&rq->lock);
+
p->sched_contributes_to_load = !!task_contributes_to_load(p);
p->state = TASK_WAKING;
p->sched_class->task_waking(p);
cpu = select_task_rq(p, p->wake_cpu, SD_BALANCE_WAKE, wake_flags);
+
if (task_cpu(p) != cpu) {
wake_flags |= WF_MIGRATED;
set_task_cpu(p, cpu);
}
+
#endif /* CONFIG_SMP */
ttwu_queue(p, cpu);
trace_sched_waking(p);
- if (!task_on_rq_queued(p))
+ if (!task_on_rq_queued(p)) {
+ u64 wallclock = walt_ktime_clock();
+
+ walt_update_task_ravg(rq->curr, rq, TASK_UPDATE, wallclock, 0);
+ walt_update_task_ravg(p, rq, TASK_WAKE, wallclock, 0);
ttwu_activate(rq, p, ENQUEUE_WAKEUP);
+ }
ttwu_do_wakeup(rq, p, 0);
ttwu_stat(p, smp_processor_id(), 0);
p->se.nr_migrations = 0;
p->se.vruntime = 0;
INIT_LIST_HEAD(&p->se.group_node);
+ walt_init_new_task_load(p);
#ifdef CONFIG_SCHEDSTATS
memset(&p->se.statistics, 0, sizeof(p->se.statistics));
struct rq *rq;
raw_spin_lock_irqsave(&p->pi_lock, flags);
+
+ walt_init_new_task_load(p);
+
/* Initialize new task's runnable average */
init_entity_runnable_average(&p->se);
#ifdef CONFIG_SMP
#endif
rq = __task_rq_lock(p);
- activate_task(rq, p, 0);
+ walt_mark_task_starting(p);
+ activate_task(rq, p, ENQUEUE_WAKEUP_NEW);
p->on_rq = TASK_ON_RQ_QUEUED;
trace_sched_wakeup_new(p);
check_preempt_curr(rq, p, WF_FORK);
return atomic_read(&this->nr_iowait);
}
+#ifdef CONFIG_CPU_QUIET
+u64 nr_running_integral(unsigned int cpu)
+{
+ unsigned int seqcnt;
+ u64 integral;
+ struct rq *q;
+
+ if (cpu >= nr_cpu_ids)
+ return 0;
+
+ q = cpu_rq(cpu);
+
+ /*
+ * Update average to avoid reading stalled value if there were
+ * no run-queue changes for a long time. On the other hand if
+ * the changes are happening right now, just read current value
+ * directly.
+ */
+
+ seqcnt = read_seqcount_begin(&q->ave_seqcnt);
+ integral = do_nr_running_integral(q);
+ if (read_seqcount_retry(&q->ave_seqcnt, seqcnt)) {
+ read_seqcount_begin(&q->ave_seqcnt);
+ integral = q->nr_running_integral;
+ }
+
+ return integral;
+}
+#endif
+
void get_iowait_load(unsigned long *nr_waiters, unsigned long *load)
{
struct rq *rq = this_rq();
return ns;
}
+#ifdef CONFIG_CPU_FREQ_GOV_SCHED
+
+static inline
+unsigned long add_capacity_margin(unsigned long cpu_capacity)
+{
+ cpu_capacity = cpu_capacity * capacity_margin;
+ cpu_capacity /= SCHED_CAPACITY_SCALE;
+ return cpu_capacity;
+}
+
+static inline
+unsigned long sum_capacity_reqs(unsigned long cfs_cap,
+ struct sched_capacity_reqs *scr)
+{
+ unsigned long total = add_capacity_margin(cfs_cap + scr->rt);
+ return total += scr->dl;
+}
+
+static void sched_freq_tick_pelt(int cpu)
+{
+ unsigned long cpu_utilization = capacity_max;
+ unsigned long capacity_curr = capacity_curr_of(cpu);
+ struct sched_capacity_reqs *scr;
+
+ scr = &per_cpu(cpu_sched_capacity_reqs, cpu);
+ if (sum_capacity_reqs(cpu_utilization, scr) < capacity_curr)
+ return;
+
+ /*
+ * To make free room for a task that is building up its "real"
+ * utilization and to harm its performance the least, request
+ * a jump to a higher OPP as soon as the margin of free capacity
+ * is impacted (specified by capacity_margin).
+ */
+ set_cfs_cpu_capacity(cpu, true, cpu_utilization);
+}
+
+#ifdef CONFIG_SCHED_WALT
+static void sched_freq_tick_walt(int cpu)
+{
+ unsigned long cpu_utilization = cpu_util(cpu);
+ unsigned long capacity_curr = capacity_curr_of(cpu);
+
+ if (walt_disabled || !sysctl_sched_use_walt_cpu_util)
+ return sched_freq_tick_pelt(cpu);
+
+ /*
+ * Add a margin to the WALT utilization.
+ * NOTE: WALT tracks a single CPU signal for all the scheduling
+ * classes, thus this margin is going to be added to the DL class as
+ * well, which is something we do not do in sched_freq_tick_pelt case.
+ */
+ cpu_utilization = add_capacity_margin(cpu_utilization);
+ if (cpu_utilization <= capacity_curr)
+ return;
+
+ /*
+ * It is likely that the load is growing so we
+ * keep the added margin in our request as an
+ * extra boost.
+ */
+ set_cfs_cpu_capacity(cpu, true, cpu_utilization);
+
+}
+#define _sched_freq_tick(cpu) sched_freq_tick_walt(cpu)
+#else
+#define _sched_freq_tick(cpu) sched_freq_tick_pelt(cpu)
+#endif /* CONFIG_SCHED_WALT */
+
+static void sched_freq_tick(int cpu)
+{
+ unsigned long capacity_orig, capacity_curr;
+
+ if (!sched_freq())
+ return;
+
+ capacity_orig = capacity_orig_of(cpu);
+ capacity_curr = capacity_curr_of(cpu);
+ if (capacity_curr == capacity_orig)
+ return;
+
+ _sched_freq_tick(cpu);
+}
+#else
+static inline void sched_freq_tick(int cpu) { }
+#endif /* CONFIG_CPU_FREQ_GOV_SCHED */
+
/*
* This function gets called by the timer code, with HZ frequency.
* We call it with interrupts disabled.
sched_clock_tick();
raw_spin_lock(&rq->lock);
+ walt_set_window_start(rq);
update_rq_clock(rq);
curr->sched_class->task_tick(rq, curr, 0);
update_cpu_load_active(rq);
+ walt_update_task_ravg(rq->curr, rq, TASK_UPDATE,
+ walt_ktime_clock(), 0);
calc_global_load_tick(rq);
+ sched_freq_tick(cpu);
raw_spin_unlock(&rq->lock);
perf_event_task_tick();
unsigned long *switch_count;
struct rq *rq;
int cpu;
+ u64 wallclock;
cpu = smp_processor_id();
rq = cpu_rq(cpu);
update_rq_clock(rq);
next = pick_next_task(rq, prev);
+ wallclock = walt_ktime_clock();
+ walt_update_task_ravg(prev, rq, PUT_PREV_TASK, wallclock, 0);
+ walt_update_task_ravg(next, rq, PICK_NEXT_TASK, wallclock, 0);
clear_tsk_need_resched(prev);
clear_preempt_need_resched();
rq->clock_skip_update = 0;
raw_spin_lock(&rq->lock);
__sched_fork(0, idle);
+
idle->state = TASK_RUNNING;
idle->se.exec_start = sched_clock();
}
}
+static struct ctl_table *
+sd_alloc_ctl_energy_table(struct sched_group_energy *sge)
+{
+ struct ctl_table *table = sd_alloc_ctl_entry(5);
+
+ if (table == NULL)
+ return NULL;
+
+ set_table_entry(&table[0], "nr_idle_states", &sge->nr_idle_states,
+ sizeof(int), 0644, proc_dointvec_minmax, false);
+ set_table_entry(&table[1], "idle_states", &sge->idle_states[0].power,
+ sge->nr_idle_states*sizeof(struct idle_state), 0644,
+ proc_doulongvec_minmax, false);
+ set_table_entry(&table[2], "nr_cap_states", &sge->nr_cap_states,
+ sizeof(int), 0644, proc_dointvec_minmax, false);
+ set_table_entry(&table[3], "cap_states", &sge->cap_states[0].cap,
+ sge->nr_cap_states*sizeof(struct capacity_state), 0644,
+ proc_doulongvec_minmax, false);
+
+ return table;
+}
+
+static struct ctl_table *
+sd_alloc_ctl_group_table(struct sched_group *sg)
+{
+ struct ctl_table *table = sd_alloc_ctl_entry(2);
+
+ if (table == NULL)
+ return NULL;
+
+ table->procname = kstrdup("energy", GFP_KERNEL);
+ table->mode = 0555;
+ table->child = sd_alloc_ctl_energy_table((struct sched_group_energy *)sg->sge);
+
+ return table;
+}
+
static struct ctl_table *
sd_alloc_ctl_domain_table(struct sched_domain *sd)
{
- struct ctl_table *table = sd_alloc_ctl_entry(14);
+ struct ctl_table *table;
+ unsigned int nr_entries = 14;
+
+ int i = 0;
+ struct sched_group *sg = sd->groups;
+
+ if (sg->sge) {
+ int nr_sgs = 0;
+
+ do {} while (nr_sgs++, sg = sg->next, sg != sd->groups);
+
+ nr_entries += nr_sgs;
+ }
+
+ table = sd_alloc_ctl_entry(nr_entries);
if (table == NULL)
return NULL;
sizeof(long), 0644, proc_doulongvec_minmax, false);
set_table_entry(&table[12], "name", sd->name,
CORENAME_MAX_SIZE, 0444, proc_dostring, false);
- /* &table[13] is terminator */
+ sg = sd->groups;
+ if (sg->sge) {
+ char buf[32];
+ struct ctl_table *entry = &table[13];
+
+ do {
+ snprintf(buf, 32, "group%d", i);
+ entry->procname = kstrdup(buf, GFP_KERNEL);
+ entry->mode = 0555;
+ entry->child = sd_alloc_ctl_group_table(sg);
+ } while (entry++, i++, sg = sg->next, sg != sd->groups);
+ }
+ /* &table[nr_entries-1] is terminator */
return table;
}
switch (action & ~CPU_TASKS_FROZEN) {
case CPU_UP_PREPARE:
+ raw_spin_lock_irqsave(&rq->lock, flags);
+ walt_set_window_start(rq);
+ raw_spin_unlock_irqrestore(&rq->lock, flags);
rq->calc_load_update = calc_load_update;
account_reset_rq(rq);
break;
sched_ttwu_pending();
/* Update our root-domain */
raw_spin_lock_irqsave(&rq->lock, flags);
+ walt_migrate_sync_cpu(cpu);
if (rq->rd) {
BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span));
set_rq_offline(rq);
printk(KERN_CONT " %*pbl",
cpumask_pr_args(sched_group_cpus(group)));
if (group->sgc->capacity != SCHED_CAPACITY_SCALE) {
- printk(KERN_CONT " (cpu_capacity = %d)",
+ printk(KERN_CONT " (cpu_capacity = %lu)",
group->sgc->capacity);
}
SD_BALANCE_FORK |
SD_BALANCE_EXEC |
SD_SHARE_CPUCAPACITY |
+ SD_ASYM_CPUCAPACITY |
SD_SHARE_PKG_RESOURCES |
- SD_SHARE_POWERDOMAIN)) {
+ SD_SHARE_POWERDOMAIN |
+ SD_SHARE_CAP_STATES)) {
if (sd->groups != sd->groups->next)
return 0;
}
SD_BALANCE_NEWIDLE |
SD_BALANCE_FORK |
SD_BALANCE_EXEC |
+ SD_ASYM_CPUCAPACITY |
SD_SHARE_CPUCAPACITY |
SD_SHARE_PKG_RESOURCES |
SD_PREFER_SIBLING |
- SD_SHARE_POWERDOMAIN);
+ SD_SHARE_POWERDOMAIN |
+ SD_SHARE_CAP_STATES);
if (nr_node_ids == 1)
pflags &= ~SD_SERIALIZE;
}
if (cpupri_init(&rd->cpupri) != 0)
goto free_rto_mask;
+
+ init_max_cpu_capacity(&rd->max_cpu_capacity);
+
+ rd->max_cap_orig_cpu = rd->min_cap_orig_cpu = -1;
+
return 0;
free_rto_mask:
DEFINE_PER_CPU(struct sched_domain *, sd_numa);
DEFINE_PER_CPU(struct sched_domain *, sd_busy);
DEFINE_PER_CPU(struct sched_domain *, sd_asym);
+DEFINE_PER_CPU(struct sched_domain *, sd_ea);
+DEFINE_PER_CPU(struct sched_domain *, sd_scs);
static void update_top_cache_domain(int cpu)
{
struct sched_domain *sd;
- struct sched_domain *busy_sd = NULL;
+ struct sched_domain *busy_sd = NULL, *ea_sd = NULL;
int id = cpu;
int size = 1;
sd = highest_flag_domain(cpu, SD_ASYM_PACKING);
rcu_assign_pointer(per_cpu(sd_asym, cpu), sd);
+
+ for_each_domain(cpu, sd) {
+ if (sd->groups->sge)
+ ea_sd = sd;
+ else
+ break;
+ }
+ rcu_assign_pointer(per_cpu(sd_ea, cpu), ea_sd);
+
+ sd = highest_flag_domain(cpu, SD_SHARE_CAP_STATES);
+ rcu_assign_pointer(per_cpu(sd_scs, cpu), sd);
}
/*
* die on a /0 trap.
*/
sg->sgc->capacity = SCHED_CAPACITY_SCALE * cpumask_weight(sg_span);
+ sg->sgc->max_capacity = SCHED_CAPACITY_SCALE;
+ sg->sgc->min_capacity = SCHED_CAPACITY_SCALE;
/*
* Make sure the first group of this domain contains the
atomic_set(&sg->sgc->nr_busy_cpus, sg->group_weight);
}
+/*
+ * Check that the per-cpu provided sd energy data is consistent for all cpus
+ * within the mask.
+ */
+static inline void check_sched_energy_data(int cpu, sched_domain_energy_f fn,
+ const struct cpumask *cpumask)
+{
+ const struct sched_group_energy * const sge = fn(cpu);
+ struct cpumask mask;
+ int i;
+
+ if (cpumask_weight(cpumask) <= 1)
+ return;
+
+ cpumask_xor(&mask, cpumask, get_cpu_mask(cpu));
+
+ for_each_cpu(i, &mask) {
+ const struct sched_group_energy * const e = fn(i);
+ int y;
+
+ BUG_ON(e->nr_idle_states != sge->nr_idle_states);
+
+ for (y = 0; y < (e->nr_idle_states); y++) {
+ BUG_ON(e->idle_states[y].power !=
+ sge->idle_states[y].power);
+ }
+
+ BUG_ON(e->nr_cap_states != sge->nr_cap_states);
+
+ for (y = 0; y < (e->nr_cap_states); y++) {
+ BUG_ON(e->cap_states[y].cap != sge->cap_states[y].cap);
+ BUG_ON(e->cap_states[y].power !=
+ sge->cap_states[y].power);
+ }
+ }
+}
+
+static void init_sched_energy(int cpu, struct sched_domain *sd,
+ sched_domain_energy_f fn)
+{
+ if (!(fn && fn(cpu)))
+ return;
+
+ if (cpu != group_balance_cpu(sd->groups))
+ return;
+
+ if (sd->child && !sd->child->groups->sge) {
+ pr_err("BUG: EAS setup broken for CPU%d\n", cpu);
+#ifdef CONFIG_SCHED_DEBUG
+ pr_err(" energy data on %s but not on %s domain\n",
+ sd->name, sd->child->name);
+#endif
+ return;
+ }
+
+ check_sched_energy_data(cpu, fn, sched_group_cpus(sd->groups));
+
+ sd->groups->sge = fn(cpu);
+}
+
/*
* Initializers for schedule domains
* Non-inlined to reduce accumulated stack pressure in build_sched_domains()
/*
* SD_flags allowed in topology descriptions.
*
- * SD_SHARE_CPUCAPACITY - describes SMT topologies
- * SD_SHARE_PKG_RESOURCES - describes shared caches
- * SD_NUMA - describes NUMA topologies
- * SD_SHARE_POWERDOMAIN - describes shared power domain
+ * These flags are purely descriptive of the topology and do not prescribe
+ * behaviour. Behaviour is artificial and mapped in the below sd_init()
+ * function:
+ *
+ * SD_SHARE_CPUCAPACITY - describes SMT topologies
+ * SD_SHARE_PKG_RESOURCES - describes shared caches
+ * SD_NUMA - describes NUMA topologies
+ * SD_SHARE_POWERDOMAIN - describes shared power domain
+ * SD_ASYM_CPUCAPACITY - describes mixed capacity topologies
+ * SD_SHARE_CAP_STATES - describes shared capacity states
+ *
+ * Odd one out, which beside describing the topology has a quirk also
+ * prescribes the desired behaviour that goes along with it:
*
* Odd one out:
* SD_ASYM_PACKING - describes SMT quirks
SD_SHARE_PKG_RESOURCES | \
SD_NUMA | \
SD_ASYM_PACKING | \
- SD_SHARE_POWERDOMAIN)
+ SD_ASYM_CPUCAPACITY | \
+ SD_SHARE_POWERDOMAIN | \
+ SD_SHARE_CAP_STATES)
static struct sched_domain *
-sd_init(struct sched_domain_topology_level *tl, int cpu)
+sd_init(struct sched_domain_topology_level *tl,
+ struct sched_domain *child, int cpu)
{
struct sched_domain *sd = *per_cpu_ptr(tl->data.sd, cpu);
int sd_weight, sd_flags = 0;
.smt_gain = 0,
.max_newidle_lb_cost = 0,
.next_decay_max_lb_cost = jiffies,
+ .child = child,
#ifdef CONFIG_SCHED_DEBUG
.name = tl->name,
#endif
* Convert topological properties into behaviour.
*/
+ if (sd->flags & SD_ASYM_CPUCAPACITY) {
+ struct sched_domain *t = sd;
+
+ for_each_lower_domain(t)
+ t->flags |= SD_BALANCE_WAKE;
+ }
+
if (sd->flags & SD_SHARE_CPUCAPACITY) {
sd->flags |= SD_PREFER_SIBLING;
sd->imbalance_pct = 110;
const struct cpumask *cpu_map, struct sched_domain_attr *attr,
struct sched_domain *child, int cpu)
{
- struct sched_domain *sd = sd_init(tl, cpu);
- if (!sd)
- return child;
+ struct sched_domain *sd = sd_init(tl, child, cpu);
cpumask_and(sched_domain_span(sd), cpu_map, tl->mask(cpu));
if (child) {
sd->level = child->level + 1;
sched_domain_level_max = max(sched_domain_level_max, sd->level);
child->parent = sd;
- sd->child = child;
if (!cpumask_subset(sched_domain_span(child),
sched_domain_span(sd))) {
/* Calculate CPU capacity for physical packages and nodes */
for (i = nr_cpumask_bits-1; i >= 0; i--) {
+ struct sched_domain_topology_level *tl = sched_domain_topology;
+
if (!cpumask_test_cpu(i, cpu_map))
continue;
- for (sd = *per_cpu_ptr(d.sd, i); sd; sd = sd->parent) {
+ for (sd = *per_cpu_ptr(d.sd, i); sd; sd = sd->parent, tl++) {
+ init_sched_energy(i, sd, tl->energy);
claim_allocations(i, sd);
init_sched_groups_capacity(i, sd);
}
/* Attach the domains */
rcu_read_lock();
for_each_cpu(i, cpu_map) {
+ int max_cpu = READ_ONCE(d.rd->max_cap_orig_cpu);
+ int min_cpu = READ_ONCE(d.rd->min_cap_orig_cpu);
+
+ if ((max_cpu < 0) || (cpu_rq(i)->cpu_capacity_orig >
+ cpu_rq(max_cpu)->cpu_capacity_orig))
+ WRITE_ONCE(d.rd->max_cap_orig_cpu, i);
+
+ if ((min_cpu < 0) || (cpu_rq(i)->cpu_capacity_orig <
+ cpu_rq(min_cpu)->cpu_capacity_orig))
+ WRITE_ONCE(d.rd->min_cap_orig_cpu, i);
+
sd = *per_cpu_ptr(d.sd, i);
+
cpu_attach_domain(sd, d.rd, i);
}
rcu_read_unlock();
{
cpumask_var_t non_isolated_cpus;
+ walt_init_cpu_efficiency();
alloc_cpumask_var(&non_isolated_cpus, GFP_KERNEL);
alloc_cpumask_var(&fallback_doms, GFP_KERNEL);
rq->idle_stamp = 0;
rq->avg_idle = 2*sysctl_sched_migration_cost;
rq->max_idle_balance_cost = sysctl_sched_migration_cost;
+#ifdef CONFIG_SCHED_WALT
+ rq->cur_irqload = 0;
+ rq->avg_irqload = 0;
+ rq->irqload_ts = 0;
+#endif
INIT_LIST_HEAD(&rq->cfs_tasks);
return (nested == preempt_offset);
}
+static int __might_sleep_init_called;
+int __init __might_sleep_init(void)
+{
+ __might_sleep_init_called = 1;
+ return 0;
+}
+early_initcall(__might_sleep_init);
+
void __might_sleep(const char *file, int line, int preempt_offset)
{
/*
rcu_sleep_check(); /* WARN_ON_ONCE() by default, no rate limit reqd. */
if ((preempt_count_equals(preempt_offset) && !irqs_disabled() &&
- !is_idle_task(current)) ||
- system_state != SYSTEM_RUNNING || oops_in_progress)
+ !is_idle_task(current)) || oops_in_progress)
+ return;
+ if (system_state != SYSTEM_RUNNING &&
+ (!__might_sleep_init_called || system_state != SYSTEM_BOOTING))
return;
if (time_before(jiffies, prev_jiffy + HZ) && prev_jiffy)
return;