2007-07-09 12:51:58 -04:00
|
|
|
/*
|
|
|
|
* Real-Time Scheduling Class (mapped to the SCHED_FIFO and SCHED_RR
|
|
|
|
* policies)
|
|
|
|
*/
|
|
|
|
|
2008-01-25 15:08:06 -05:00
|
|
|
#ifdef CONFIG_SMP
|
2008-01-25 15:08:15 -05:00
|
|
|
|
2008-01-25 15:08:18 -05:00
|
|
|
static inline int rt_overloaded(struct rq *rq)
|
2008-01-25 15:08:06 -05:00
|
|
|
{
|
2008-01-25 15:08:18 -05:00
|
|
|
return atomic_read(&rq->rd->rto_count);
|
2008-01-25 15:08:06 -05:00
|
|
|
}
|
2008-01-25 15:08:15 -05:00
|
|
|
|
2008-01-25 15:08:06 -05:00
|
|
|
static inline void rt_set_overload(struct rq *rq)
|
|
|
|
{
|
2008-06-04 15:04:05 -04:00
|
|
|
if (!rq->online)
|
|
|
|
return;
|
|
|
|
|
2008-01-25 15:08:18 -05:00
|
|
|
cpu_set(rq->cpu, rq->rd->rto_mask);
|
2008-01-25 15:08:06 -05:00
|
|
|
/*
|
|
|
|
* Make sure the mask is visible before we set
|
|
|
|
* the overload count. That is checked to determine
|
|
|
|
* if we should look at the mask. It would be a shame
|
|
|
|
* if we looked at the mask, but the mask was not
|
|
|
|
* updated yet.
|
|
|
|
*/
|
|
|
|
wmb();
|
2008-01-25 15:08:18 -05:00
|
|
|
atomic_inc(&rq->rd->rto_count);
|
2008-01-25 15:08:06 -05:00
|
|
|
}
|
2008-01-25 15:08:15 -05:00
|
|
|
|
2008-01-25 15:08:06 -05:00
|
|
|
static inline void rt_clear_overload(struct rq *rq)
|
|
|
|
{
|
2008-06-04 15:04:05 -04:00
|
|
|
if (!rq->online)
|
|
|
|
return;
|
|
|
|
|
2008-01-25 15:08:06 -05:00
|
|
|
/* the order here really doesn't matter */
|
2008-01-25 15:08:18 -05:00
|
|
|
atomic_dec(&rq->rd->rto_count);
|
|
|
|
cpu_clear(rq->cpu, rq->rd->rto_mask);
|
2008-01-25 15:08:06 -05:00
|
|
|
}
|
2008-01-25 15:08:07 -05:00
|
|
|
|
|
|
|
static void update_rt_migration(struct rq *rq)
|
|
|
|
{
|
2008-01-25 15:08:18 -05:00
|
|
|
if (rq->rt.rt_nr_migratory && (rq->rt.rt_nr_running > 1)) {
|
2008-01-25 15:08:23 -05:00
|
|
|
if (!rq->rt.overloaded) {
|
|
|
|
rt_set_overload(rq);
|
|
|
|
rq->rt.overloaded = 1;
|
|
|
|
}
|
|
|
|
} else if (rq->rt.overloaded) {
|
2008-01-25 15:08:07 -05:00
|
|
|
rt_clear_overload(rq);
|
2008-01-25 15:08:18 -05:00
|
|
|
rq->rt.overloaded = 0;
|
|
|
|
}
|
2008-01-25 15:08:07 -05:00
|
|
|
}
|
2008-01-25 15:08:06 -05:00
|
|
|
#endif /* CONFIG_SMP */
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se)
|
2008-01-25 15:08:29 -05:00
|
|
|
{
|
2008-01-25 15:08:30 -05:00
|
|
|
return container_of(rt_se, struct task_struct, rt);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline int on_rt_rq(struct sched_rt_entity *rt_se)
|
|
|
|
{
|
|
|
|
return !list_empty(&rt_se->run_list);
|
|
|
|
}
|
|
|
|
|
2008-02-13 09:45:40 -05:00
|
|
|
#ifdef CONFIG_RT_GROUP_SCHED
|
2008-01-25 15:08:30 -05:00
|
|
|
|
2008-02-13 09:45:39 -05:00
|
|
|
static inline u64 sched_rt_runtime(struct rt_rq *rt_rq)
|
2008-01-25 15:08:30 -05:00
|
|
|
{
|
|
|
|
if (!rt_rq->tg)
|
2008-02-13 09:45:39 -05:00
|
|
|
return RUNTIME_INF;
|
2008-01-25 15:08:30 -05:00
|
|
|
|
2008-04-19 13:44:58 -04:00
|
|
|
return rt_rq->rt_runtime;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline u64 sched_rt_period(struct rt_rq *rt_rq)
|
|
|
|
{
|
|
|
|
return ktime_to_ns(rt_rq->tg->rt_bandwidth.rt_period);
|
2008-01-25 15:08:30 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
#define for_each_leaf_rt_rq(rt_rq, rq) \
|
|
|
|
list_for_each_entry(rt_rq, &rq->leaf_rt_rq_list, leaf_rt_rq_list)
|
|
|
|
|
|
|
|
static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
|
|
|
|
{
|
|
|
|
return rt_rq->rq;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
|
|
|
|
{
|
|
|
|
return rt_se->rt_rq;
|
|
|
|
}
|
|
|
|
|
|
|
|
#define for_each_sched_rt_entity(rt_se) \
|
|
|
|
for (; rt_se; rt_se = rt_se->parent)
|
|
|
|
|
|
|
|
static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se)
|
|
|
|
{
|
|
|
|
return rt_se->my_q;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void enqueue_rt_entity(struct sched_rt_entity *rt_se);
|
|
|
|
static void dequeue_rt_entity(struct sched_rt_entity *rt_se);
|
|
|
|
|
2008-02-13 09:45:39 -05:00
|
|
|
static void sched_rt_rq_enqueue(struct rt_rq *rt_rq)
|
2008-01-25 15:08:30 -05:00
|
|
|
{
|
|
|
|
struct sched_rt_entity *rt_se = rt_rq->rt_se;
|
|
|
|
|
|
|
|
if (rt_se && !on_rt_rq(rt_se) && rt_rq->rt_nr_running) {
|
2008-01-25 15:08:32 -05:00
|
|
|
struct task_struct *curr = rq_of_rt_rq(rt_rq)->curr;
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
enqueue_rt_entity(rt_se);
|
2008-01-25 15:08:32 -05:00
|
|
|
if (rt_rq->highest_prio < curr->prio)
|
|
|
|
resched_task(curr);
|
2008-01-25 15:08:30 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-02-13 09:45:39 -05:00
|
|
|
static void sched_rt_rq_dequeue(struct rt_rq *rt_rq)
|
2008-01-25 15:08:30 -05:00
|
|
|
{
|
|
|
|
struct sched_rt_entity *rt_se = rt_rq->rt_se;
|
|
|
|
|
|
|
|
if (rt_se && on_rt_rq(rt_se))
|
|
|
|
dequeue_rt_entity(rt_se);
|
|
|
|
}
|
|
|
|
|
2008-02-13 09:45:39 -05:00
|
|
|
static inline int rt_rq_throttled(struct rt_rq *rt_rq)
|
|
|
|
{
|
|
|
|
return rt_rq->rt_throttled && !rt_rq->rt_nr_boosted;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int rt_se_boosted(struct sched_rt_entity *rt_se)
|
|
|
|
{
|
|
|
|
struct rt_rq *rt_rq = group_rt_rq(rt_se);
|
|
|
|
struct task_struct *p;
|
|
|
|
|
|
|
|
if (rt_rq)
|
|
|
|
return !!rt_rq->rt_nr_boosted;
|
|
|
|
|
|
|
|
p = rt_task_of(rt_se);
|
|
|
|
return p->prio != p->normal_prio;
|
|
|
|
}
|
|
|
|
|
2008-04-19 13:44:57 -04:00
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
static inline cpumask_t sched_rt_period_mask(void)
|
|
|
|
{
|
|
|
|
return cpu_rq(smp_processor_id())->rd->span;
|
|
|
|
}
|
2008-01-25 15:08:30 -05:00
|
|
|
#else
|
2008-04-19 13:44:57 -04:00
|
|
|
static inline cpumask_t sched_rt_period_mask(void)
|
|
|
|
{
|
|
|
|
return cpu_online_map;
|
|
|
|
}
|
|
|
|
#endif
|
2008-01-25 15:08:30 -05:00
|
|
|
|
2008-04-19 13:44:57 -04:00
|
|
|
static inline
|
|
|
|
struct rt_rq *sched_rt_period_rt_rq(struct rt_bandwidth *rt_b, int cpu)
|
2008-01-25 15:08:30 -05:00
|
|
|
{
|
2008-04-19 13:44:57 -04:00
|
|
|
return container_of(rt_b, struct task_group, rt_bandwidth)->rt_rq[cpu];
|
|
|
|
}
|
2008-02-13 09:45:39 -05:00
|
|
|
|
2008-04-19 13:44:58 -04:00
|
|
|
static inline struct rt_bandwidth *sched_rt_bandwidth(struct rt_rq *rt_rq)
|
|
|
|
{
|
|
|
|
return &rt_rq->tg->rt_bandwidth;
|
|
|
|
}
|
|
|
|
|
2008-06-24 14:09:43 -04:00
|
|
|
#else /* !CONFIG_RT_GROUP_SCHED */
|
2008-04-19 13:44:57 -04:00
|
|
|
|
|
|
|
static inline u64 sched_rt_runtime(struct rt_rq *rt_rq)
|
|
|
|
{
|
2008-04-19 13:44:58 -04:00
|
|
|
return rt_rq->rt_runtime;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline u64 sched_rt_period(struct rt_rq *rt_rq)
|
|
|
|
{
|
|
|
|
return ktime_to_ns(def_rt_bandwidth.rt_period);
|
2008-01-25 15:08:30 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
#define for_each_leaf_rt_rq(rt_rq, rq) \
|
|
|
|
for (rt_rq = &rq->rt; rt_rq; rt_rq = NULL)
|
|
|
|
|
|
|
|
static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
|
|
|
|
{
|
|
|
|
return container_of(rt_rq, struct rq, rt);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
|
|
|
|
{
|
|
|
|
struct task_struct *p = rt_task_of(rt_se);
|
|
|
|
struct rq *rq = task_rq(p);
|
|
|
|
|
|
|
|
return &rq->rt;
|
|
|
|
}
|
|
|
|
|
|
|
|
#define for_each_sched_rt_entity(rt_se) \
|
|
|
|
for (; rt_se; rt_se = NULL)
|
|
|
|
|
|
|
|
static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se)
|
|
|
|
{
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2008-02-13 09:45:39 -05:00
|
|
|
static inline void sched_rt_rq_enqueue(struct rt_rq *rt_rq)
|
2008-01-25 15:08:30 -05:00
|
|
|
{
|
2008-08-26 15:09:43 -04:00
|
|
|
if (rt_rq->rt_nr_running)
|
|
|
|
resched_task(rq_of_rt_rq(rt_rq)->curr);
|
2008-01-25 15:08:30 -05:00
|
|
|
}
|
|
|
|
|
2008-02-13 09:45:39 -05:00
|
|
|
static inline void sched_rt_rq_dequeue(struct rt_rq *rt_rq)
|
2008-01-25 15:08:30 -05:00
|
|
|
{
|
|
|
|
}
|
|
|
|
|
2008-02-13 09:45:39 -05:00
|
|
|
static inline int rt_rq_throttled(struct rt_rq *rt_rq)
|
|
|
|
{
|
|
|
|
return rt_rq->rt_throttled;
|
|
|
|
}
|
2008-04-19 13:44:57 -04:00
|
|
|
|
|
|
|
static inline cpumask_t sched_rt_period_mask(void)
|
|
|
|
{
|
|
|
|
return cpu_online_map;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline
|
|
|
|
struct rt_rq *sched_rt_period_rt_rq(struct rt_bandwidth *rt_b, int cpu)
|
|
|
|
{
|
|
|
|
return &cpu_rq(cpu)->rt;
|
|
|
|
}
|
|
|
|
|
2008-04-19 13:44:58 -04:00
|
|
|
static inline struct rt_bandwidth *sched_rt_bandwidth(struct rt_rq *rt_rq)
|
|
|
|
{
|
|
|
|
return &def_rt_bandwidth;
|
|
|
|
}
|
|
|
|
|
2008-06-24 14:09:43 -04:00
|
|
|
#endif /* CONFIG_RT_GROUP_SCHED */
|
2008-04-19 13:44:57 -04:00
|
|
|
|
2008-04-19 13:44:58 -04:00
|
|
|
#ifdef CONFIG_SMP
|
2008-06-19 08:22:25 -04:00
|
|
|
static int do_balance_runtime(struct rt_rq *rt_rq)
|
2008-04-19 13:44:58 -04:00
|
|
|
{
|
|
|
|
struct rt_bandwidth *rt_b = sched_rt_bandwidth(rt_rq);
|
|
|
|
struct root_domain *rd = cpu_rq(smp_processor_id())->rd;
|
|
|
|
int i, weight, more = 0;
|
|
|
|
u64 rt_period;
|
|
|
|
|
|
|
|
weight = cpus_weight(rd->span);
|
|
|
|
|
|
|
|
spin_lock(&rt_b->rt_runtime_lock);
|
|
|
|
rt_period = ktime_to_ns(rt_b->rt_period);
|
2008-05-12 15:21:13 -04:00
|
|
|
for_each_cpu_mask_nr(i, rd->span) {
|
2008-04-19 13:44:58 -04:00
|
|
|
struct rt_rq *iter = sched_rt_period_rt_rq(rt_b, i);
|
|
|
|
s64 diff;
|
|
|
|
|
|
|
|
if (iter == rt_rq)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
spin_lock(&iter->rt_runtime_lock);
|
2008-06-05 08:49:58 -04:00
|
|
|
if (iter->rt_runtime == RUNTIME_INF)
|
|
|
|
goto next;
|
|
|
|
|
2008-04-19 13:44:58 -04:00
|
|
|
diff = iter->rt_runtime - iter->rt_time;
|
|
|
|
if (diff > 0) {
|
2008-07-24 06:43:13 -04:00
|
|
|
diff = div_u64((u64)diff, weight);
|
2008-04-19 13:44:58 -04:00
|
|
|
if (rt_rq->rt_runtime + diff > rt_period)
|
|
|
|
diff = rt_period - rt_rq->rt_runtime;
|
|
|
|
iter->rt_runtime -= diff;
|
|
|
|
rt_rq->rt_runtime += diff;
|
|
|
|
more = 1;
|
|
|
|
if (rt_rq->rt_runtime == rt_period) {
|
|
|
|
spin_unlock(&iter->rt_runtime_lock);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2008-06-05 08:49:58 -04:00
|
|
|
next:
|
2008-04-19 13:44:58 -04:00
|
|
|
spin_unlock(&iter->rt_runtime_lock);
|
|
|
|
}
|
|
|
|
spin_unlock(&rt_b->rt_runtime_lock);
|
|
|
|
|
|
|
|
return more;
|
|
|
|
}
|
2008-06-05 08:49:58 -04:00
|
|
|
|
|
|
|
static void __disable_runtime(struct rq *rq)
|
|
|
|
{
|
|
|
|
struct root_domain *rd = rq->rd;
|
|
|
|
struct rt_rq *rt_rq;
|
|
|
|
|
|
|
|
if (unlikely(!scheduler_running))
|
|
|
|
return;
|
|
|
|
|
|
|
|
for_each_leaf_rt_rq(rt_rq, rq) {
|
|
|
|
struct rt_bandwidth *rt_b = sched_rt_bandwidth(rt_rq);
|
|
|
|
s64 want;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
spin_lock(&rt_b->rt_runtime_lock);
|
|
|
|
spin_lock(&rt_rq->rt_runtime_lock);
|
|
|
|
if (rt_rq->rt_runtime == RUNTIME_INF ||
|
|
|
|
rt_rq->rt_runtime == rt_b->rt_runtime)
|
|
|
|
goto balanced;
|
|
|
|
spin_unlock(&rt_rq->rt_runtime_lock);
|
|
|
|
|
|
|
|
want = rt_b->rt_runtime - rt_rq->rt_runtime;
|
|
|
|
|
|
|
|
for_each_cpu_mask(i, rd->span) {
|
|
|
|
struct rt_rq *iter = sched_rt_period_rt_rq(rt_b, i);
|
|
|
|
s64 diff;
|
|
|
|
|
2008-08-14 09:49:00 -04:00
|
|
|
if (iter == rt_rq || iter->rt_runtime == RUNTIME_INF)
|
2008-06-05 08:49:58 -04:00
|
|
|
continue;
|
|
|
|
|
|
|
|
spin_lock(&iter->rt_runtime_lock);
|
|
|
|
if (want > 0) {
|
|
|
|
diff = min_t(s64, iter->rt_runtime, want);
|
|
|
|
iter->rt_runtime -= diff;
|
|
|
|
want -= diff;
|
|
|
|
} else {
|
|
|
|
iter->rt_runtime -= want;
|
|
|
|
want -= want;
|
|
|
|
}
|
|
|
|
spin_unlock(&iter->rt_runtime_lock);
|
|
|
|
|
|
|
|
if (!want)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
spin_lock(&rt_rq->rt_runtime_lock);
|
|
|
|
BUG_ON(want);
|
|
|
|
balanced:
|
|
|
|
rt_rq->rt_runtime = RUNTIME_INF;
|
|
|
|
spin_unlock(&rt_rq->rt_runtime_lock);
|
|
|
|
spin_unlock(&rt_b->rt_runtime_lock);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void disable_runtime(struct rq *rq)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&rq->lock, flags);
|
|
|
|
__disable_runtime(rq);
|
|
|
|
spin_unlock_irqrestore(&rq->lock, flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __enable_runtime(struct rq *rq)
|
|
|
|
{
|
|
|
|
struct rt_rq *rt_rq;
|
|
|
|
|
|
|
|
if (unlikely(!scheduler_running))
|
|
|
|
return;
|
|
|
|
|
|
|
|
for_each_leaf_rt_rq(rt_rq, rq) {
|
|
|
|
struct rt_bandwidth *rt_b = sched_rt_bandwidth(rt_rq);
|
|
|
|
|
|
|
|
spin_lock(&rt_b->rt_runtime_lock);
|
|
|
|
spin_lock(&rt_rq->rt_runtime_lock);
|
|
|
|
rt_rq->rt_runtime = rt_b->rt_runtime;
|
|
|
|
rt_rq->rt_time = 0;
|
|
|
|
spin_unlock(&rt_rq->rt_runtime_lock);
|
|
|
|
spin_unlock(&rt_b->rt_runtime_lock);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void enable_runtime(struct rq *rq)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&rq->lock, flags);
|
|
|
|
__enable_runtime(rq);
|
|
|
|
spin_unlock_irqrestore(&rq->lock, flags);
|
|
|
|
}
|
|
|
|
|
2008-06-19 08:22:26 -04:00
|
|
|
static int balance_runtime(struct rt_rq *rt_rq)
|
|
|
|
{
|
|
|
|
int more = 0;
|
|
|
|
|
|
|
|
if (rt_rq->rt_time > rt_rq->rt_runtime) {
|
|
|
|
spin_unlock(&rt_rq->rt_runtime_lock);
|
|
|
|
more = do_balance_runtime(rt_rq);
|
|
|
|
spin_lock(&rt_rq->rt_runtime_lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
return more;
|
|
|
|
}
|
2008-06-24 14:09:43 -04:00
|
|
|
#else /* !CONFIG_SMP */
|
2008-06-19 08:22:26 -04:00
|
|
|
static inline int balance_runtime(struct rt_rq *rt_rq)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
2008-06-24 14:09:43 -04:00
|
|
|
#endif /* CONFIG_SMP */
|
2008-04-19 13:44:58 -04:00
|
|
|
|
2008-06-19 08:22:26 -04:00
|
|
|
static int do_sched_rt_period_timer(struct rt_bandwidth *rt_b, int overrun)
|
|
|
|
{
|
|
|
|
int i, idle = 1;
|
|
|
|
cpumask_t span;
|
|
|
|
|
|
|
|
if (rt_b->rt_runtime == RUNTIME_INF)
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
span = sched_rt_period_mask();
|
|
|
|
for_each_cpu_mask(i, span) {
|
|
|
|
int enqueue = 0;
|
|
|
|
struct rt_rq *rt_rq = sched_rt_period_rt_rq(rt_b, i);
|
|
|
|
struct rq *rq = rq_of_rt_rq(rt_rq);
|
|
|
|
|
|
|
|
spin_lock(&rq->lock);
|
|
|
|
if (rt_rq->rt_time) {
|
|
|
|
u64 runtime;
|
|
|
|
|
|
|
|
spin_lock(&rt_rq->rt_runtime_lock);
|
|
|
|
if (rt_rq->rt_throttled)
|
|
|
|
balance_runtime(rt_rq);
|
|
|
|
runtime = rt_rq->rt_runtime;
|
|
|
|
rt_rq->rt_time -= min(rt_rq->rt_time, overrun*runtime);
|
|
|
|
if (rt_rq->rt_throttled && rt_rq->rt_time < runtime) {
|
|
|
|
rt_rq->rt_throttled = 0;
|
|
|
|
enqueue = 1;
|
|
|
|
}
|
|
|
|
if (rt_rq->rt_time || rt_rq->rt_nr_running)
|
|
|
|
idle = 0;
|
|
|
|
spin_unlock(&rt_rq->rt_runtime_lock);
|
2008-06-19 08:22:28 -04:00
|
|
|
} else if (rt_rq->rt_nr_running)
|
|
|
|
idle = 0;
|
2008-06-19 08:22:26 -04:00
|
|
|
|
|
|
|
if (enqueue)
|
|
|
|
sched_rt_rq_enqueue(rt_rq);
|
|
|
|
spin_unlock(&rq->lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
return idle;
|
|
|
|
}
|
2008-04-19 13:44:58 -04:00
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
static inline int rt_se_prio(struct sched_rt_entity *rt_se)
|
|
|
|
{
|
2008-02-13 09:45:40 -05:00
|
|
|
#ifdef CONFIG_RT_GROUP_SCHED
|
2008-01-25 15:08:30 -05:00
|
|
|
struct rt_rq *rt_rq = group_rt_rq(rt_se);
|
|
|
|
|
|
|
|
if (rt_rq)
|
|
|
|
return rt_rq->highest_prio;
|
|
|
|
#endif
|
|
|
|
|
|
|
|
return rt_task_of(rt_se)->prio;
|
|
|
|
}
|
|
|
|
|
2008-02-13 09:45:39 -05:00
|
|
|
static int sched_rt_runtime_exceeded(struct rt_rq *rt_rq)
|
2008-01-25 15:08:30 -05:00
|
|
|
{
|
2008-02-13 09:45:39 -05:00
|
|
|
u64 runtime = sched_rt_runtime(rt_rq);
|
2008-01-25 15:08:29 -05:00
|
|
|
|
|
|
|
if (rt_rq->rt_throttled)
|
2008-02-13 09:45:39 -05:00
|
|
|
return rt_rq_throttled(rt_rq);
|
2008-01-25 15:08:29 -05:00
|
|
|
|
2008-04-19 13:44:58 -04:00
|
|
|
if (sched_rt_runtime(rt_rq) >= sched_rt_period(rt_rq))
|
|
|
|
return 0;
|
|
|
|
|
2008-06-19 08:22:25 -04:00
|
|
|
balance_runtime(rt_rq);
|
|
|
|
runtime = sched_rt_runtime(rt_rq);
|
|
|
|
if (runtime == RUNTIME_INF)
|
|
|
|
return 0;
|
2008-04-19 13:44:58 -04:00
|
|
|
|
2008-02-13 09:45:39 -05:00
|
|
|
if (rt_rq->rt_time > runtime) {
|
2008-01-25 15:08:30 -05:00
|
|
|
rt_rq->rt_throttled = 1;
|
2008-02-13 09:45:39 -05:00
|
|
|
if (rt_rq_throttled(rt_rq)) {
|
2008-02-13 09:45:39 -05:00
|
|
|
sched_rt_rq_dequeue(rt_rq);
|
2008-02-13 09:45:39 -05:00
|
|
|
return 1;
|
|
|
|
}
|
2008-01-25 15:08:29 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-07-09 12:51:58 -04:00
|
|
|
/*
|
|
|
|
* Update the current task's runtime statistics. Skip current tasks that
|
|
|
|
* are not in our scheduling class.
|
|
|
|
*/
|
2007-10-15 11:00:13 -04:00
|
|
|
static void update_curr_rt(struct rq *rq)
|
2007-07-09 12:51:58 -04:00
|
|
|
{
|
|
|
|
struct task_struct *curr = rq->curr;
|
2008-01-25 15:08:30 -05:00
|
|
|
struct sched_rt_entity *rt_se = &curr->rt;
|
|
|
|
struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
|
2007-07-09 12:51:58 -04:00
|
|
|
u64 delta_exec;
|
|
|
|
|
|
|
|
if (!task_has_rt_policy(curr))
|
|
|
|
return;
|
|
|
|
|
2007-08-09 05:16:47 -04:00
|
|
|
delta_exec = rq->clock - curr->se.exec_start;
|
2007-07-09 12:51:58 -04:00
|
|
|
if (unlikely((s64)delta_exec < 0))
|
|
|
|
delta_exec = 0;
|
2007-08-02 11:41:40 -04:00
|
|
|
|
|
|
|
schedstat_set(curr->se.exec_max, max(curr->se.exec_max, delta_exec));
|
2007-07-09 12:51:58 -04:00
|
|
|
|
|
|
|
curr->se.sum_exec_runtime += delta_exec;
|
2007-08-09 05:16:47 -04:00
|
|
|
curr->se.exec_start = rq->clock;
|
2007-12-02 14:04:49 -05:00
|
|
|
cpuacct_charge(curr, delta_exec);
|
2008-01-25 15:08:29 -05:00
|
|
|
|
2008-04-19 13:44:59 -04:00
|
|
|
for_each_sched_rt_entity(rt_se) {
|
|
|
|
rt_rq = rt_rq_of_se(rt_se);
|
|
|
|
|
|
|
|
spin_lock(&rt_rq->rt_runtime_lock);
|
2008-08-19 06:33:03 -04:00
|
|
|
if (sched_rt_runtime(rt_rq) != RUNTIME_INF) {
|
|
|
|
rt_rq->rt_time += delta_exec;
|
|
|
|
if (sched_rt_runtime_exceeded(rt_rq))
|
|
|
|
resched_task(curr);
|
|
|
|
}
|
2008-04-19 13:44:59 -04:00
|
|
|
spin_unlock(&rt_rq->rt_runtime_lock);
|
|
|
|
}
|
2007-07-09 12:51:58 -04:00
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
static inline
|
|
|
|
void inc_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
|
2008-01-25 15:08:03 -05:00
|
|
|
{
|
2008-01-25 15:08:30 -05:00
|
|
|
WARN_ON(!rt_prio(rt_se_prio(rt_se)));
|
|
|
|
rt_rq->rt_nr_running++;
|
2008-02-13 09:45:40 -05:00
|
|
|
#if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
|
2008-05-12 15:21:01 -04:00
|
|
|
if (rt_se_prio(rt_se) < rt_rq->highest_prio) {
|
2008-07-11 08:34:54 -04:00
|
|
|
#ifdef CONFIG_SMP
|
2008-05-12 15:21:01 -04:00
|
|
|
struct rq *rq = rq_of_rt_rq(rt_rq);
|
2008-07-11 08:34:54 -04:00
|
|
|
#endif
|
2008-06-04 15:04:05 -04:00
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
rt_rq->highest_prio = rt_se_prio(rt_se);
|
2008-06-05 06:25:37 -04:00
|
|
|
#ifdef CONFIG_SMP
|
2008-06-04 15:04:05 -04:00
|
|
|
if (rq->online)
|
|
|
|
cpupri_set(&rq->rd->cpupri, rq->cpu,
|
|
|
|
rt_se_prio(rt_se));
|
2008-06-05 06:25:37 -04:00
|
|
|
#endif
|
2008-05-12 15:21:01 -04:00
|
|
|
}
|
2008-01-25 15:08:30 -05:00
|
|
|
#endif
|
2008-01-25 15:08:04 -05:00
|
|
|
#ifdef CONFIG_SMP
|
2008-01-25 15:08:30 -05:00
|
|
|
if (rt_se->nr_cpus_allowed > 1) {
|
|
|
|
struct rq *rq = rq_of_rt_rq(rt_rq);
|
2008-06-05 06:25:37 -04:00
|
|
|
|
2008-01-25 15:08:07 -05:00
|
|
|
rq->rt.rt_nr_migratory++;
|
2008-01-25 15:08:30 -05:00
|
|
|
}
|
2008-01-25 15:08:07 -05:00
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
update_rt_migration(rq_of_rt_rq(rt_rq));
|
|
|
|
#endif
|
2008-02-13 09:45:40 -05:00
|
|
|
#ifdef CONFIG_RT_GROUP_SCHED
|
2008-02-13 09:45:39 -05:00
|
|
|
if (rt_se_boosted(rt_se))
|
|
|
|
rt_rq->rt_nr_boosted++;
|
2008-04-19 13:44:57 -04:00
|
|
|
|
|
|
|
if (rt_rq->tg)
|
|
|
|
start_rt_bandwidth(&rt_rq->tg->rt_bandwidth);
|
|
|
|
#else
|
|
|
|
start_rt_bandwidth(&def_rt_bandwidth);
|
2008-02-13 09:45:39 -05:00
|
|
|
#endif
|
2008-01-25 15:08:03 -05:00
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
static inline
|
|
|
|
void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
|
2008-01-25 15:08:03 -05:00
|
|
|
{
|
2008-05-12 15:21:01 -04:00
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
int highest_prio = rt_rq->highest_prio;
|
|
|
|
#endif
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
WARN_ON(!rt_prio(rt_se_prio(rt_se)));
|
|
|
|
WARN_ON(!rt_rq->rt_nr_running);
|
|
|
|
rt_rq->rt_nr_running--;
|
2008-02-13 09:45:40 -05:00
|
|
|
#if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
|
2008-01-25 15:08:30 -05:00
|
|
|
if (rt_rq->rt_nr_running) {
|
2008-01-25 15:08:04 -05:00
|
|
|
struct rt_prio_array *array;
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
WARN_ON(rt_se_prio(rt_se) < rt_rq->highest_prio);
|
|
|
|
if (rt_se_prio(rt_se) == rt_rq->highest_prio) {
|
2008-01-25 15:08:04 -05:00
|
|
|
/* recalculate */
|
2008-01-25 15:08:30 -05:00
|
|
|
array = &rt_rq->active;
|
|
|
|
rt_rq->highest_prio =
|
2008-01-25 15:08:04 -05:00
|
|
|
sched_find_first_bit(array->bitmap);
|
|
|
|
} /* otherwise leave rq->highest prio alone */
|
|
|
|
} else
|
2008-01-25 15:08:30 -05:00
|
|
|
rt_rq->highest_prio = MAX_RT_PRIO;
|
|
|
|
#endif
|
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
if (rt_se->nr_cpus_allowed > 1) {
|
|
|
|
struct rq *rq = rq_of_rt_rq(rt_rq);
|
2008-01-25 15:08:07 -05:00
|
|
|
rq->rt.rt_nr_migratory--;
|
2008-01-25 15:08:30 -05:00
|
|
|
}
|
2008-01-25 15:08:07 -05:00
|
|
|
|
2008-05-12 15:21:01 -04:00
|
|
|
if (rt_rq->highest_prio != highest_prio) {
|
|
|
|
struct rq *rq = rq_of_rt_rq(rt_rq);
|
2008-06-04 15:04:05 -04:00
|
|
|
|
|
|
|
if (rq->online)
|
|
|
|
cpupri_set(&rq->rd->cpupri, rq->cpu,
|
|
|
|
rt_rq->highest_prio);
|
2008-05-12 15:21:01 -04:00
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
update_rt_migration(rq_of_rt_rq(rt_rq));
|
2008-01-25 15:08:04 -05:00
|
|
|
#endif /* CONFIG_SMP */
|
2008-02-13 09:45:40 -05:00
|
|
|
#ifdef CONFIG_RT_GROUP_SCHED
|
2008-02-13 09:45:39 -05:00
|
|
|
if (rt_se_boosted(rt_se))
|
|
|
|
rt_rq->rt_nr_boosted--;
|
|
|
|
|
|
|
|
WARN_ON(!rt_rq->rt_nr_running && rt_rq->rt_nr_boosted);
|
|
|
|
#endif
|
2008-01-25 15:08:03 -05:00
|
|
|
}
|
|
|
|
|
2008-06-19 03:06:57 -04:00
|
|
|
static void __enqueue_rt_entity(struct sched_rt_entity *rt_se)
|
2007-07-09 12:51:58 -04:00
|
|
|
{
|
2008-01-25 15:08:30 -05:00
|
|
|
struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
|
|
|
|
struct rt_prio_array *array = &rt_rq->active;
|
|
|
|
struct rt_rq *group_rq = group_rt_rq(rt_se);
|
sched: rework of "prioritize non-migratable tasks over migratable ones"
regarding this commit: 45c01e824991b2dd0a332e19efc4901acb31209f
I think we can do it simpler. Please take a look at the patch below.
Instead of having 2 separate arrays (which is + ~800 bytes on x86_32 and
twice so on x86_64), let's add "exclusive" (the ones that are bound to
this CPU) tasks to the head of the queue and "shared" ones -- to the
end.
In case of a few newly woken up "exclusive" tasks, they are 'stacked'
(not queued as now), meaning that a task {i+1} is being placed in front
of the previously woken up task {i}. But I don't think that this
behavior may cause any realistic problems.
There are a couple of changes on top of this one.
(1) in check_preempt_curr_rt()
I don't think there is a need for the "pick_next_rt_entity(rq, &rq->rt)
!= &rq->curr->rt" check.
enqueue_task_rt(p) and check_preempt_curr_rt() are always called one
after another with rq->lock being held so the following check
"p->rt.nr_cpus_allowed == 1 && rq->curr->rt.nr_cpus_allowed != 1" should
be enough (well, just its left part) to guarantee that 'p' has been
queued in front of the 'curr'.
(2) in set_cpus_allowed_rt()
I don't thinks there is a need for requeue_task_rt() here.
Perhaps, the only case when 'requeue' (+ reschedule) might be useful is
as follows:
i) weight == 1 && cpu_isset(task_cpu(p), *new_mask)
i.e. a task is being bound to this CPU);
ii) 'p' != rq->curr
but here, 'p' has already been on this CPU for a while and was not
migrated. i.e. it's possible that 'rq->curr' would not have high chances
to be migrated right at this particular moment (although, has chance in
a bit longer term), should we allow it to be preempted.
Anyway, I think we should not perhaps make it more complex trying to
address some rare corner cases. For instance, that's why a single queue
approach would be preferable. Unless I'm missing something obvious, this
approach gives us similar functionality at lower cost.
Verified only compilation-wise.
(Almost)-Signed-off-by: Dmitry Adamushko <dmitry.adamushko@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-06-10 18:58:30 -04:00
|
|
|
struct list_head *queue = array->queue + rt_se_prio(rt_se);
|
2007-07-09 12:51:58 -04:00
|
|
|
|
2008-06-19 03:06:57 -04:00
|
|
|
/*
|
|
|
|
* Don't enqueue the group if its throttled, or when empty.
|
|
|
|
* The latter is a consequence of the former when a child group
|
|
|
|
* get throttled and the current group doesn't have any other
|
|
|
|
* active members.
|
|
|
|
*/
|
|
|
|
if (group_rq && (rt_rq_throttled(group_rq) || !group_rq->rt_nr_running))
|
2008-01-25 15:08:30 -05:00
|
|
|
return;
|
2008-01-25 15:08:03 -05:00
|
|
|
|
2008-07-01 17:32:15 -04:00
|
|
|
list_add_tail(&rt_se->run_list, queue);
|
2008-01-25 15:08:30 -05:00
|
|
|
__set_bit(rt_se_prio(rt_se), array->bitmap);
|
2008-01-25 15:08:27 -05:00
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
inc_rt_tasks(rt_se, rt_rq);
|
|
|
|
}
|
|
|
|
|
2008-06-19 03:06:57 -04:00
|
|
|
static void __dequeue_rt_entity(struct sched_rt_entity *rt_se)
|
2008-01-25 15:08:30 -05:00
|
|
|
{
|
|
|
|
struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
|
|
|
|
struct rt_prio_array *array = &rt_rq->active;
|
|
|
|
|
|
|
|
list_del_init(&rt_se->run_list);
|
|
|
|
if (list_empty(array->queue + rt_se_prio(rt_se)))
|
|
|
|
__clear_bit(rt_se_prio(rt_se), array->bitmap);
|
|
|
|
|
|
|
|
dec_rt_tasks(rt_se, rt_rq);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Because the prio of an upper entry depends on the lower
|
|
|
|
* entries, we must remove entries top - down.
|
|
|
|
*/
|
2008-06-19 03:06:57 -04:00
|
|
|
static void dequeue_rt_stack(struct sched_rt_entity *rt_se)
|
2008-01-25 15:08:30 -05:00
|
|
|
{
|
2008-06-19 03:06:57 -04:00
|
|
|
struct sched_rt_entity *back = NULL;
|
2008-01-25 15:08:30 -05:00
|
|
|
|
2008-04-19 13:45:00 -04:00
|
|
|
for_each_sched_rt_entity(rt_se) {
|
|
|
|
rt_se->back = back;
|
|
|
|
back = rt_se;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (rt_se = back; rt_se; rt_se = rt_se->back) {
|
|
|
|
if (on_rt_rq(rt_se))
|
2008-06-19 03:06:57 -04:00
|
|
|
__dequeue_rt_entity(rt_se);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void enqueue_rt_entity(struct sched_rt_entity *rt_se)
|
|
|
|
{
|
|
|
|
dequeue_rt_stack(rt_se);
|
|
|
|
for_each_sched_rt_entity(rt_se)
|
|
|
|
__enqueue_rt_entity(rt_se);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void dequeue_rt_entity(struct sched_rt_entity *rt_se)
|
|
|
|
{
|
|
|
|
dequeue_rt_stack(rt_se);
|
|
|
|
|
|
|
|
for_each_sched_rt_entity(rt_se) {
|
|
|
|
struct rt_rq *rt_rq = group_rt_rq(rt_se);
|
|
|
|
|
|
|
|
if (rt_rq && rt_rq->rt_nr_running)
|
|
|
|
__enqueue_rt_entity(rt_se);
|
2008-04-19 13:45:00 -04:00
|
|
|
}
|
2007-07-09 12:51:58 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Adding/removing a task to/from a priority array:
|
|
|
|
*/
|
2008-01-25 15:08:30 -05:00
|
|
|
static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup)
|
|
|
|
{
|
|
|
|
struct sched_rt_entity *rt_se = &p->rt;
|
|
|
|
|
|
|
|
if (wakeup)
|
|
|
|
rt_se->timeout = 0;
|
|
|
|
|
2008-06-19 03:06:57 -04:00
|
|
|
enqueue_rt_entity(rt_se);
|
2008-06-27 07:41:14 -04:00
|
|
|
|
|
|
|
inc_cpu_load(rq, p->se.load.weight);
|
2008-01-25 15:08:30 -05:00
|
|
|
}
|
|
|
|
|
2007-08-09 05:16:48 -04:00
|
|
|
static void dequeue_task_rt(struct rq *rq, struct task_struct *p, int sleep)
|
2007-07-09 12:51:58 -04:00
|
|
|
{
|
2008-01-25 15:08:30 -05:00
|
|
|
struct sched_rt_entity *rt_se = &p->rt;
|
2007-07-09 12:51:58 -04:00
|
|
|
|
2007-08-09 05:16:48 -04:00
|
|
|
update_curr_rt(rq);
|
2008-06-19 03:06:57 -04:00
|
|
|
dequeue_rt_entity(rt_se);
|
2008-06-27 07:41:14 -04:00
|
|
|
|
|
|
|
dec_cpu_load(rq, p->se.load.weight);
|
2007-07-09 12:51:58 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Put task to the end of the run list without the overhead of dequeue
|
|
|
|
* followed by enqueue.
|
|
|
|
*/
|
2008-07-01 17:32:15 -04:00
|
|
|
static void
|
|
|
|
requeue_rt_entity(struct rt_rq *rt_rq, struct sched_rt_entity *rt_se, int head)
|
2008-01-25 15:08:30 -05:00
|
|
|
{
|
2008-06-19 03:09:15 -04:00
|
|
|
if (on_rt_rq(rt_se)) {
|
2008-07-01 17:32:15 -04:00
|
|
|
struct rt_prio_array *array = &rt_rq->active;
|
|
|
|
struct list_head *queue = array->queue + rt_se_prio(rt_se);
|
|
|
|
|
|
|
|
if (head)
|
|
|
|
list_move(&rt_se->run_list, queue);
|
|
|
|
else
|
|
|
|
list_move_tail(&rt_se->run_list, queue);
|
2008-06-19 03:09:15 -04:00
|
|
|
}
|
2008-01-25 15:08:30 -05:00
|
|
|
}
|
|
|
|
|
2008-07-01 17:32:15 -04:00
|
|
|
static void requeue_task_rt(struct rq *rq, struct task_struct *p, int head)
|
2007-07-09 12:51:58 -04:00
|
|
|
{
|
2008-01-25 15:08:30 -05:00
|
|
|
struct sched_rt_entity *rt_se = &p->rt;
|
|
|
|
struct rt_rq *rt_rq;
|
2007-07-09 12:51:58 -04:00
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
for_each_sched_rt_entity(rt_se) {
|
|
|
|
rt_rq = rt_rq_of_se(rt_se);
|
2008-07-01 17:32:15 -04:00
|
|
|
requeue_rt_entity(rt_rq, rt_se, head);
|
2008-01-25 15:08:30 -05:00
|
|
|
}
|
2007-07-09 12:51:58 -04:00
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
static void yield_task_rt(struct rq *rq)
|
2007-07-09 12:51:58 -04:00
|
|
|
{
|
2008-07-01 17:32:15 -04:00
|
|
|
requeue_task_rt(rq, rq->curr, 0);
|
2007-07-09 12:51:58 -04:00
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:09 -05:00
|
|
|
#ifdef CONFIG_SMP
|
2008-01-25 15:08:10 -05:00
|
|
|
static int find_lowest_rq(struct task_struct *task);
|
|
|
|
|
2008-01-25 15:08:09 -05:00
|
|
|
static int select_task_rq_rt(struct task_struct *p, int sync)
|
|
|
|
{
|
2008-01-25 15:08:10 -05:00
|
|
|
struct rq *rq = task_rq(p);
|
|
|
|
|
|
|
|
/*
|
2008-01-25 15:08:12 -05:00
|
|
|
* If the current task is an RT task, then
|
|
|
|
* try to see if we can wake this RT task up on another
|
|
|
|
* runqueue. Otherwise simply start this RT task
|
|
|
|
* on its current runqueue.
|
|
|
|
*
|
|
|
|
* We want to avoid overloading runqueues. Even if
|
|
|
|
* the RT task is of higher priority than the current RT task.
|
|
|
|
* RT tasks behave differently than other tasks. If
|
|
|
|
* one gets preempted, we try to push it off to another queue.
|
|
|
|
* So trying to keep a preempting RT task on the same
|
|
|
|
* cache hot CPU will force the running RT task to
|
|
|
|
* a cold CPU. So we waste all the cache for the lower
|
|
|
|
* RT task in hopes of saving some of a RT task
|
|
|
|
* that is just being woken and probably will have
|
|
|
|
* cold cache anyway.
|
2008-01-25 15:08:10 -05:00
|
|
|
*/
|
2008-01-25 15:08:13 -05:00
|
|
|
if (unlikely(rt_task(rq->curr)) &&
|
2008-01-25 15:08:30 -05:00
|
|
|
(p->rt.nr_cpus_allowed > 1)) {
|
2008-01-25 15:08:10 -05:00
|
|
|
int cpu = find_lowest_rq(p);
|
|
|
|
|
|
|
|
return (cpu == -1) ? task_cpu(p) : cpu;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Otherwise, just let it ride on the affined RQ and the
|
|
|
|
* post-schedule router will push the preempted task away
|
|
|
|
*/
|
2008-01-25 15:08:09 -05:00
|
|
|
return task_cpu(p);
|
|
|
|
}
|
2008-07-01 17:32:15 -04:00
|
|
|
|
|
|
|
static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p)
|
|
|
|
{
|
|
|
|
cpumask_t mask;
|
|
|
|
|
|
|
|
if (rq->curr->rt.nr_cpus_allowed == 1)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (p->rt.nr_cpus_allowed != 1
|
|
|
|
&& cpupri_find(&rq->rd->cpupri, p, &mask))
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (!cpupri_find(&rq->rd->cpupri, rq->curr, &mask))
|
|
|
|
return;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* There appears to be other cpus that can accept
|
|
|
|
* current and none to run 'p', so lets reschedule
|
|
|
|
* to try and push current away:
|
|
|
|
*/
|
|
|
|
requeue_task_rt(rq, p, 1);
|
|
|
|
resched_task(rq->curr);
|
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:09 -05:00
|
|
|
#endif /* CONFIG_SMP */
|
|
|
|
|
2007-07-09 12:51:58 -04:00
|
|
|
/*
|
|
|
|
* Preempt the current task with a newly woken task if needed:
|
|
|
|
*/
|
|
|
|
static void check_preempt_curr_rt(struct rq *rq, struct task_struct *p)
|
|
|
|
{
|
sched: prioritize non-migratable tasks over migratable ones
Dmitry Adamushko pointed out a known flaw in the rt-balancing algorithm
that could allow suboptimal balancing if a non-migratable task gets
queued behind a running migratable one. It is discussed in this thread:
http://lkml.org/lkml/2008/4/22/296
This issue has been further exacerbated by a recent checkin to
sched-devel (git-id 5eee63a5ebc19a870ac40055c0be49457f3a89a3).
>From a pure priority standpoint, the run-queue is doing the "right"
thing. Using Dmitry's nomenclature, if T0 is on cpu1 first, and T1
wakes up at equal or lower priority (affined only to cpu1) later, it
*should* wait for T0 to finish. However, in reality that is likely
suboptimal from a system perspective if there are other cores that
could allow T0 and T1 to run concurrently. Since T1 can not migrate,
the only choice for higher concurrency is to try to move T0. This is
not something we addessed in the recent rt-balancing re-work.
This patch tries to enhance the balancing algorithm by accomodating this
scenario. It accomplishes this by incorporating the migratability of a
task into its priority calculation. Within a numerical tsk->prio, a
non-migratable task is logically higher than a migratable one. We
maintain this by introducing a new per-priority queue (xqueue, or
exclusive-queue) for holding non-migratable tasks. The scheduler will
draw from the xqueue over the standard shared-queue (squeue) when
available.
There are several details for utilizing this properly.
1) During task-wake-up, we not only need to check if the priority
preempts the current task, but we also need to check for this
non-migratable condition. Therefore, if a non-migratable task wakes
up and sees an equal priority migratable task already running, it
will attempt to preempt it *if* there is a likelyhood that the
current task will find an immediate home.
2) Tasks only get this non-migratable "priority boost" on wake-up. Any
requeuing will result in the non-migratable task being queued to the
end of the shared queue. This is an attempt to prevent the system
from being completely unfair to migratable tasks during things like
SCHED_RR timeslicing.
I am sure this patch introduces potentially "odd" behavior if you
concoct a scenario where a bunch of non-migratable threads could starve
migratable ones given the right pattern. I am not yet convinced that
this is a problem since we are talking about tasks of equal RT priority
anyway, and there never is much in the way of guarantees against
starvation under that scenario anyway. (e.g. you could come up with a
similar scenario with a specific timing environment verses an affinity
environment). I can be convinced otherwise, but for now I think this is
"ok".
Signed-off-by: Gregory Haskins <ghaskins@novell.com>
CC: Dmitry Adamushko <dmitry.adamushko@gmail.com>
CC: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 15:20:41 -04:00
|
|
|
if (p->prio < rq->curr->prio) {
|
2007-07-09 12:51:58 -04:00
|
|
|
resched_task(rq->curr);
|
sched: prioritize non-migratable tasks over migratable ones
Dmitry Adamushko pointed out a known flaw in the rt-balancing algorithm
that could allow suboptimal balancing if a non-migratable task gets
queued behind a running migratable one. It is discussed in this thread:
http://lkml.org/lkml/2008/4/22/296
This issue has been further exacerbated by a recent checkin to
sched-devel (git-id 5eee63a5ebc19a870ac40055c0be49457f3a89a3).
>From a pure priority standpoint, the run-queue is doing the "right"
thing. Using Dmitry's nomenclature, if T0 is on cpu1 first, and T1
wakes up at equal or lower priority (affined only to cpu1) later, it
*should* wait for T0 to finish. However, in reality that is likely
suboptimal from a system perspective if there are other cores that
could allow T0 and T1 to run concurrently. Since T1 can not migrate,
the only choice for higher concurrency is to try to move T0. This is
not something we addessed in the recent rt-balancing re-work.
This patch tries to enhance the balancing algorithm by accomodating this
scenario. It accomplishes this by incorporating the migratability of a
task into its priority calculation. Within a numerical tsk->prio, a
non-migratable task is logically higher than a migratable one. We
maintain this by introducing a new per-priority queue (xqueue, or
exclusive-queue) for holding non-migratable tasks. The scheduler will
draw from the xqueue over the standard shared-queue (squeue) when
available.
There are several details for utilizing this properly.
1) During task-wake-up, we not only need to check if the priority
preempts the current task, but we also need to check for this
non-migratable condition. Therefore, if a non-migratable task wakes
up and sees an equal priority migratable task already running, it
will attempt to preempt it *if* there is a likelyhood that the
current task will find an immediate home.
2) Tasks only get this non-migratable "priority boost" on wake-up. Any
requeuing will result in the non-migratable task being queued to the
end of the shared queue. This is an attempt to prevent the system
from being completely unfair to migratable tasks during things like
SCHED_RR timeslicing.
I am sure this patch introduces potentially "odd" behavior if you
concoct a scenario where a bunch of non-migratable threads could starve
migratable ones given the right pattern. I am not yet convinced that
this is a problem since we are talking about tasks of equal RT priority
anyway, and there never is much in the way of guarantees against
starvation under that scenario anyway. (e.g. you could come up with a
similar scenario with a specific timing environment verses an affinity
environment). I can be convinced otherwise, but for now I think this is
"ok".
Signed-off-by: Gregory Haskins <ghaskins@novell.com>
CC: Dmitry Adamushko <dmitry.adamushko@gmail.com>
CC: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 15:20:41 -04:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
/*
|
|
|
|
* If:
|
|
|
|
*
|
|
|
|
* - the newly woken task is of equal priority to the current task
|
|
|
|
* - the newly woken task is non-migratable while current is migratable
|
|
|
|
* - current will be preempted on the next reschedule
|
|
|
|
*
|
|
|
|
* we should check to see if current can readily move to a different
|
|
|
|
* cpu. If so, we will reschedule to allow the push logic to try
|
|
|
|
* to move current somewhere else, making room for our non-migratable
|
|
|
|
* task.
|
|
|
|
*/
|
2008-07-01 17:32:15 -04:00
|
|
|
if (p->prio == rq->curr->prio && !need_resched())
|
|
|
|
check_preempt_equal_prio(rq, p);
|
sched: prioritize non-migratable tasks over migratable ones
Dmitry Adamushko pointed out a known flaw in the rt-balancing algorithm
that could allow suboptimal balancing if a non-migratable task gets
queued behind a running migratable one. It is discussed in this thread:
http://lkml.org/lkml/2008/4/22/296
This issue has been further exacerbated by a recent checkin to
sched-devel (git-id 5eee63a5ebc19a870ac40055c0be49457f3a89a3).
>From a pure priority standpoint, the run-queue is doing the "right"
thing. Using Dmitry's nomenclature, if T0 is on cpu1 first, and T1
wakes up at equal or lower priority (affined only to cpu1) later, it
*should* wait for T0 to finish. However, in reality that is likely
suboptimal from a system perspective if there are other cores that
could allow T0 and T1 to run concurrently. Since T1 can not migrate,
the only choice for higher concurrency is to try to move T0. This is
not something we addessed in the recent rt-balancing re-work.
This patch tries to enhance the balancing algorithm by accomodating this
scenario. It accomplishes this by incorporating the migratability of a
task into its priority calculation. Within a numerical tsk->prio, a
non-migratable task is logically higher than a migratable one. We
maintain this by introducing a new per-priority queue (xqueue, or
exclusive-queue) for holding non-migratable tasks. The scheduler will
draw from the xqueue over the standard shared-queue (squeue) when
available.
There are several details for utilizing this properly.
1) During task-wake-up, we not only need to check if the priority
preempts the current task, but we also need to check for this
non-migratable condition. Therefore, if a non-migratable task wakes
up and sees an equal priority migratable task already running, it
will attempt to preempt it *if* there is a likelyhood that the
current task will find an immediate home.
2) Tasks only get this non-migratable "priority boost" on wake-up. Any
requeuing will result in the non-migratable task being queued to the
end of the shared queue. This is an attempt to prevent the system
from being completely unfair to migratable tasks during things like
SCHED_RR timeslicing.
I am sure this patch introduces potentially "odd" behavior if you
concoct a scenario where a bunch of non-migratable threads could starve
migratable ones given the right pattern. I am not yet convinced that
this is a problem since we are talking about tasks of equal RT priority
anyway, and there never is much in the way of guarantees against
starvation under that scenario anyway. (e.g. you could come up with a
similar scenario with a specific timing environment verses an affinity
environment). I can be convinced otherwise, but for now I think this is
"ok".
Signed-off-by: Gregory Haskins <ghaskins@novell.com>
CC: Dmitry Adamushko <dmitry.adamushko@gmail.com>
CC: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 15:20:41 -04:00
|
|
|
#endif
|
2007-07-09 12:51:58 -04:00
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
static struct sched_rt_entity *pick_next_rt_entity(struct rq *rq,
|
|
|
|
struct rt_rq *rt_rq)
|
2007-07-09 12:51:58 -04:00
|
|
|
{
|
2008-01-25 15:08:30 -05:00
|
|
|
struct rt_prio_array *array = &rt_rq->active;
|
|
|
|
struct sched_rt_entity *next = NULL;
|
2007-07-09 12:51:58 -04:00
|
|
|
struct list_head *queue;
|
|
|
|
int idx;
|
|
|
|
|
|
|
|
idx = sched_find_first_bit(array->bitmap);
|
2008-01-25 15:08:30 -05:00
|
|
|
BUG_ON(idx >= MAX_RT_PRIO);
|
2007-07-09 12:51:58 -04:00
|
|
|
|
|
|
|
queue = array->queue + idx;
|
2008-01-25 15:08:30 -05:00
|
|
|
next = list_entry(queue->next, struct sched_rt_entity, run_list);
|
2008-01-25 15:08:34 -05:00
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
return next;
|
|
|
|
}
|
2007-07-09 12:51:58 -04:00
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
static struct task_struct *pick_next_task_rt(struct rq *rq)
|
|
|
|
{
|
|
|
|
struct sched_rt_entity *rt_se;
|
|
|
|
struct task_struct *p;
|
|
|
|
struct rt_rq *rt_rq;
|
2007-07-09 12:51:58 -04:00
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
rt_rq = &rq->rt;
|
|
|
|
|
|
|
|
if (unlikely(!rt_rq->rt_nr_running))
|
|
|
|
return NULL;
|
|
|
|
|
2008-02-13 09:45:39 -05:00
|
|
|
if (rt_rq_throttled(rt_rq))
|
2008-01-25 15:08:30 -05:00
|
|
|
return NULL;
|
|
|
|
|
|
|
|
do {
|
|
|
|
rt_se = pick_next_rt_entity(rq, rt_rq);
|
2008-01-25 15:08:34 -05:00
|
|
|
BUG_ON(!rt_se);
|
2008-01-25 15:08:30 -05:00
|
|
|
rt_rq = group_rt_rq(rt_se);
|
|
|
|
} while (rt_rq);
|
|
|
|
|
|
|
|
p = rt_task_of(rt_se);
|
|
|
|
p->se.exec_start = rq->clock;
|
|
|
|
return p;
|
2007-07-09 12:51:58 -04:00
|
|
|
}
|
|
|
|
|
2007-08-09 05:16:49 -04:00
|
|
|
static void put_prev_task_rt(struct rq *rq, struct task_struct *p)
|
2007-07-09 12:51:58 -04:00
|
|
|
{
|
2007-08-09 05:16:48 -04:00
|
|
|
update_curr_rt(rq);
|
2007-07-09 12:51:58 -04:00
|
|
|
p->se.exec_start = 0;
|
|
|
|
}
|
|
|
|
|
2007-10-24 12:23:51 -04:00
|
|
|
#ifdef CONFIG_SMP
|
2008-01-25 15:08:30 -05:00
|
|
|
|
2008-01-25 15:08:05 -05:00
|
|
|
/* Only try algorithms three times */
|
|
|
|
#define RT_MAX_TRIES 3
|
|
|
|
|
|
|
|
static int double_lock_balance(struct rq *this_rq, struct rq *busiest);
|
2008-08-11 03:30:22 -04:00
|
|
|
static void double_unlock_balance(struct rq *this_rq, struct rq *busiest);
|
|
|
|
|
2008-01-25 15:08:05 -05:00
|
|
|
static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep);
|
|
|
|
|
2008-01-25 15:08:07 -05:00
|
|
|
static int pick_rt_task(struct rq *rq, struct task_struct *p, int cpu)
|
|
|
|
{
|
|
|
|
if (!task_running(rq, p) &&
|
2008-01-25 15:08:07 -05:00
|
|
|
(cpu < 0 || cpu_isset(cpu, p->cpus_allowed)) &&
|
2008-01-25 15:08:30 -05:00
|
|
|
(p->rt.nr_cpus_allowed > 1))
|
2008-01-25 15:08:07 -05:00
|
|
|
return 1;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:05 -05:00
|
|
|
/* Return the second highest RT task, NULL otherwise */
|
2008-01-25 15:08:14 -05:00
|
|
|
static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu)
|
2008-01-25 15:08:05 -05:00
|
|
|
{
|
2008-01-25 15:08:30 -05:00
|
|
|
struct task_struct *next = NULL;
|
|
|
|
struct sched_rt_entity *rt_se;
|
|
|
|
struct rt_prio_array *array;
|
|
|
|
struct rt_rq *rt_rq;
|
2008-01-25 15:08:05 -05:00
|
|
|
int idx;
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
for_each_leaf_rt_rq(rt_rq, rq) {
|
|
|
|
array = &rt_rq->active;
|
|
|
|
idx = sched_find_first_bit(array->bitmap);
|
|
|
|
next_idx:
|
|
|
|
if (idx >= MAX_RT_PRIO)
|
|
|
|
continue;
|
|
|
|
if (next && next->prio < idx)
|
|
|
|
continue;
|
|
|
|
list_for_each_entry(rt_se, array->queue + idx, run_list) {
|
|
|
|
struct task_struct *p = rt_task_of(rt_se);
|
|
|
|
if (pick_rt_task(rq, p, cpu)) {
|
|
|
|
next = p;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (!next) {
|
|
|
|
idx = find_next_bit(array->bitmap, MAX_RT_PRIO, idx+1);
|
|
|
|
goto next_idx;
|
|
|
|
}
|
2008-01-25 15:08:07 -05:00
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:05 -05:00
|
|
|
return next;
|
|
|
|
}
|
|
|
|
|
|
|
|
static DEFINE_PER_CPU(cpumask_t, local_cpu_mask);
|
|
|
|
|
2008-01-25 15:08:11 -05:00
|
|
|
static inline int pick_optimal_cpu(int this_cpu, cpumask_t *mask)
|
|
|
|
{
|
|
|
|
int first;
|
|
|
|
|
|
|
|
/* "this_cpu" is cheaper to preempt than a remote processor */
|
|
|
|
if ((this_cpu != -1) && cpu_isset(this_cpu, *mask))
|
|
|
|
return this_cpu;
|
|
|
|
|
|
|
|
first = first_cpu(*mask);
|
|
|
|
if (first != NR_CPUS)
|
|
|
|
return first;
|
|
|
|
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int find_lowest_rq(struct task_struct *task)
|
|
|
|
{
|
|
|
|
struct sched_domain *sd;
|
|
|
|
cpumask_t *lowest_mask = &__get_cpu_var(local_cpu_mask);
|
|
|
|
int this_cpu = smp_processor_id();
|
|
|
|
int cpu = task_cpu(task);
|
2008-01-25 15:08:13 -05:00
|
|
|
|
2008-05-12 15:21:01 -04:00
|
|
|
if (task->rt.nr_cpus_allowed == 1)
|
|
|
|
return -1; /* No other targets possible */
|
2008-01-25 15:08:11 -05:00
|
|
|
|
2008-05-12 15:21:01 -04:00
|
|
|
if (!cpupri_find(&task_rq(task)->rd->cpupri, task, lowest_mask))
|
|
|
|
return -1; /* No targets found */
|
2008-01-25 15:08:11 -05:00
|
|
|
|
2008-07-15 07:43:49 -04:00
|
|
|
/*
|
|
|
|
* Only consider CPUs that are usable for migration.
|
|
|
|
* I guess we might want to change cpupri_find() to ignore those
|
|
|
|
* in the first place.
|
|
|
|
*/
|
|
|
|
cpus_and(*lowest_mask, *lowest_mask, cpu_active_map);
|
|
|
|
|
2008-01-25 15:08:11 -05:00
|
|
|
/*
|
|
|
|
* At this point we have built a mask of cpus representing the
|
|
|
|
* lowest priority tasks in the system. Now we want to elect
|
|
|
|
* the best one based on our affinity and topology.
|
|
|
|
*
|
|
|
|
* We prioritize the last cpu that the task executed on since
|
|
|
|
* it is most likely cache-hot in that location.
|
|
|
|
*/
|
|
|
|
if (cpu_isset(cpu, *lowest_mask))
|
|
|
|
return cpu;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Otherwise, we consult the sched_domains span maps to figure
|
|
|
|
* out which cpu is logically closest to our hot cache data.
|
|
|
|
*/
|
|
|
|
if (this_cpu == cpu)
|
|
|
|
this_cpu = -1; /* Skip this_cpu opt if the same */
|
|
|
|
|
|
|
|
for_each_domain(cpu, sd) {
|
|
|
|
if (sd->flags & SD_WAKE_AFFINE) {
|
|
|
|
cpumask_t domain_mask;
|
|
|
|
int best_cpu;
|
|
|
|
|
|
|
|
cpus_and(domain_mask, sd->span, *lowest_mask);
|
|
|
|
|
|
|
|
best_cpu = pick_optimal_cpu(this_cpu,
|
|
|
|
&domain_mask);
|
|
|
|
if (best_cpu != -1)
|
|
|
|
return best_cpu;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* And finally, if there were no matches within the domains
|
|
|
|
* just give the caller *something* to work with from the compatible
|
|
|
|
* locations.
|
|
|
|
*/
|
|
|
|
return pick_optimal_cpu(this_cpu, lowest_mask);
|
2008-01-25 15:08:10 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Will lock the rq it finds */
|
2008-01-25 15:08:15 -05:00
|
|
|
static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
|
2008-01-25 15:08:10 -05:00
|
|
|
{
|
|
|
|
struct rq *lowest_rq = NULL;
|
|
|
|
int tries;
|
2008-01-25 15:08:15 -05:00
|
|
|
int cpu;
|
2008-01-25 15:08:05 -05:00
|
|
|
|
2008-01-25 15:08:10 -05:00
|
|
|
for (tries = 0; tries < RT_MAX_TRIES; tries++) {
|
|
|
|
cpu = find_lowest_rq(task);
|
|
|
|
|
2008-01-25 15:08:10 -05:00
|
|
|
if ((cpu == -1) || (cpu == rq->cpu))
|
2008-01-25 15:08:05 -05:00
|
|
|
break;
|
|
|
|
|
2008-01-25 15:08:10 -05:00
|
|
|
lowest_rq = cpu_rq(cpu);
|
|
|
|
|
2008-01-25 15:08:05 -05:00
|
|
|
/* if the prio of this runqueue changed, try again */
|
2008-01-25 15:08:10 -05:00
|
|
|
if (double_lock_balance(rq, lowest_rq)) {
|
2008-01-25 15:08:05 -05:00
|
|
|
/*
|
|
|
|
* We had to unlock the run queue. In
|
|
|
|
* the mean time, task could have
|
|
|
|
* migrated already or had its affinity changed.
|
|
|
|
* Also make sure that it wasn't scheduled on its rq.
|
|
|
|
*/
|
2008-01-25 15:08:10 -05:00
|
|
|
if (unlikely(task_rq(task) != rq ||
|
2008-01-25 15:08:15 -05:00
|
|
|
!cpu_isset(lowest_rq->cpu,
|
|
|
|
task->cpus_allowed) ||
|
2008-01-25 15:08:10 -05:00
|
|
|
task_running(rq, task) ||
|
2008-01-25 15:08:05 -05:00
|
|
|
!task->se.on_rq)) {
|
2008-01-25 15:08:15 -05:00
|
|
|
|
2008-01-25 15:08:05 -05:00
|
|
|
spin_unlock(&lowest_rq->lock);
|
|
|
|
lowest_rq = NULL;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* If this rq is still suitable use it. */
|
|
|
|
if (lowest_rq->rt.highest_prio > task->prio)
|
|
|
|
break;
|
|
|
|
|
|
|
|
/* try again */
|
2008-08-11 03:30:22 -04:00
|
|
|
double_unlock_balance(rq, lowest_rq);
|
2008-01-25 15:08:05 -05:00
|
|
|
lowest_rq = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return lowest_rq;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If the current CPU has more than one RT task, see if the non
|
|
|
|
* running task can migrate over to a CPU that is running a task
|
|
|
|
* of lesser priority.
|
|
|
|
*/
|
2008-01-25 15:08:09 -05:00
|
|
|
static int push_rt_task(struct rq *rq)
|
2008-01-25 15:08:05 -05:00
|
|
|
{
|
|
|
|
struct task_struct *next_task;
|
|
|
|
struct rq *lowest_rq;
|
|
|
|
int ret = 0;
|
|
|
|
int paranoid = RT_MAX_TRIES;
|
|
|
|
|
2008-01-25 15:08:12 -05:00
|
|
|
if (!rq->rt.overloaded)
|
|
|
|
return 0;
|
|
|
|
|
2008-01-25 15:08:09 -05:00
|
|
|
next_task = pick_next_highest_task_rt(rq, -1);
|
2008-01-25 15:08:05 -05:00
|
|
|
if (!next_task)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
retry:
|
2008-01-25 15:08:09 -05:00
|
|
|
if (unlikely(next_task == rq->curr)) {
|
2008-01-25 15:08:07 -05:00
|
|
|
WARN_ON(1);
|
2008-01-25 15:08:05 -05:00
|
|
|
return 0;
|
2008-01-25 15:08:07 -05:00
|
|
|
}
|
2008-01-25 15:08:05 -05:00
|
|
|
|
|
|
|
/*
|
|
|
|
* It's possible that the next_task slipped in of
|
|
|
|
* higher priority than current. If that's the case
|
|
|
|
* just reschedule current.
|
|
|
|
*/
|
2008-01-25 15:08:09 -05:00
|
|
|
if (unlikely(next_task->prio < rq->curr->prio)) {
|
|
|
|
resched_task(rq->curr);
|
2008-01-25 15:08:05 -05:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:09 -05:00
|
|
|
/* We might release rq lock */
|
2008-01-25 15:08:05 -05:00
|
|
|
get_task_struct(next_task);
|
|
|
|
|
|
|
|
/* find_lock_lowest_rq locks the rq if found */
|
2008-01-25 15:08:09 -05:00
|
|
|
lowest_rq = find_lock_lowest_rq(next_task, rq);
|
2008-01-25 15:08:05 -05:00
|
|
|
if (!lowest_rq) {
|
|
|
|
struct task_struct *task;
|
|
|
|
/*
|
2008-01-25 15:08:09 -05:00
|
|
|
* find lock_lowest_rq releases rq->lock
|
2008-01-25 15:08:05 -05:00
|
|
|
* so it is possible that next_task has changed.
|
|
|
|
* If it has, then try again.
|
|
|
|
*/
|
2008-01-25 15:08:09 -05:00
|
|
|
task = pick_next_highest_task_rt(rq, -1);
|
2008-01-25 15:08:05 -05:00
|
|
|
if (unlikely(task != next_task) && task && paranoid--) {
|
|
|
|
put_task_struct(next_task);
|
|
|
|
next_task = task;
|
|
|
|
goto retry;
|
|
|
|
}
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:09 -05:00
|
|
|
deactivate_task(rq, next_task, 0);
|
2008-01-25 15:08:05 -05:00
|
|
|
set_task_cpu(next_task, lowest_rq->cpu);
|
|
|
|
activate_task(lowest_rq, next_task, 0);
|
|
|
|
|
|
|
|
resched_task(lowest_rq->curr);
|
|
|
|
|
2008-08-11 03:30:22 -04:00
|
|
|
double_unlock_balance(rq, lowest_rq);
|
2008-01-25 15:08:05 -05:00
|
|
|
|
|
|
|
ret = 1;
|
|
|
|
out:
|
|
|
|
put_task_struct(next_task);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* TODO: Currently we just use the second highest prio task on
|
|
|
|
* the queue, and stop when it can't migrate (or there's
|
|
|
|
* no more RT tasks). There may be a case where a lower
|
|
|
|
* priority RT task has a different affinity than the
|
|
|
|
* higher RT task. In this case the lower RT task could
|
|
|
|
* possibly be able to migrate where as the higher priority
|
|
|
|
* RT task could not. We currently ignore this issue.
|
|
|
|
* Enhancements are welcome!
|
|
|
|
*/
|
|
|
|
static void push_rt_tasks(struct rq *rq)
|
|
|
|
{
|
|
|
|
/* push_rt_task will return true if it moved an RT */
|
|
|
|
while (push_rt_task(rq))
|
|
|
|
;
|
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:07 -05:00
|
|
|
static int pull_rt_task(struct rq *this_rq)
|
|
|
|
{
|
2008-01-25 15:08:17 -05:00
|
|
|
int this_cpu = this_rq->cpu, ret = 0, cpu;
|
|
|
|
struct task_struct *p, *next;
|
2008-01-25 15:08:07 -05:00
|
|
|
struct rq *src_rq;
|
|
|
|
|
2008-01-25 15:08:18 -05:00
|
|
|
if (likely(!rt_overloaded(this_rq)))
|
2008-01-25 15:08:07 -05:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
next = pick_next_task_rt(this_rq);
|
|
|
|
|
2008-05-12 15:21:13 -04:00
|
|
|
for_each_cpu_mask_nr(cpu, this_rq->rd->rto_mask) {
|
2008-01-25 15:08:07 -05:00
|
|
|
if (this_cpu == cpu)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
src_rq = cpu_rq(cpu);
|
|
|
|
/*
|
|
|
|
* We can potentially drop this_rq's lock in
|
|
|
|
* double_lock_balance, and another CPU could
|
|
|
|
* steal our next task - hence we must cause
|
|
|
|
* the caller to recalculate the next task
|
|
|
|
* in that case:
|
|
|
|
*/
|
|
|
|
if (double_lock_balance(this_rq, src_rq)) {
|
|
|
|
struct task_struct *old_next = next;
|
2008-01-25 15:08:17 -05:00
|
|
|
|
2008-01-25 15:08:07 -05:00
|
|
|
next = pick_next_task_rt(this_rq);
|
|
|
|
if (next != old_next)
|
|
|
|
ret = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Are there still pullable RT tasks?
|
|
|
|
*/
|
2008-01-25 15:08:30 -05:00
|
|
|
if (src_rq->rt.rt_nr_running <= 1)
|
|
|
|
goto skip;
|
2008-01-25 15:08:07 -05:00
|
|
|
|
|
|
|
p = pick_next_highest_task_rt(src_rq, this_cpu);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Do we have an RT task that preempts
|
|
|
|
* the to-be-scheduled task?
|
|
|
|
*/
|
|
|
|
if (p && (!next || (p->prio < next->prio))) {
|
|
|
|
WARN_ON(p == src_rq->curr);
|
|
|
|
WARN_ON(!p->se.on_rq);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* There's a chance that p is higher in priority
|
|
|
|
* than what's currently running on its cpu.
|
|
|
|
* This is just that p is wakeing up and hasn't
|
|
|
|
* had a chance to schedule. We only pull
|
|
|
|
* p if it is lower in priority than the
|
|
|
|
* current task on the run queue or
|
|
|
|
* this_rq next task is lower in prio than
|
|
|
|
* the current task on that rq.
|
|
|
|
*/
|
|
|
|
if (p->prio < src_rq->curr->prio ||
|
|
|
|
(next && next->prio < src_rq->curr->prio))
|
2008-01-25 15:08:30 -05:00
|
|
|
goto skip;
|
2008-01-25 15:08:07 -05:00
|
|
|
|
|
|
|
ret = 1;
|
|
|
|
|
|
|
|
deactivate_task(src_rq, p, 0);
|
|
|
|
set_task_cpu(p, this_cpu);
|
|
|
|
activate_task(this_rq, p, 0);
|
|
|
|
/*
|
|
|
|
* We continue with the search, just in
|
|
|
|
* case there's an even higher prio task
|
|
|
|
* in another runqueue. (low likelyhood
|
|
|
|
* but possible)
|
2008-01-25 15:08:17 -05:00
|
|
|
*
|
2008-01-25 15:08:07 -05:00
|
|
|
* Update next so that we won't pick a task
|
|
|
|
* on another cpu with a priority lower (or equal)
|
|
|
|
* than the one we just picked.
|
|
|
|
*/
|
|
|
|
next = p;
|
|
|
|
|
|
|
|
}
|
2008-01-25 15:08:30 -05:00
|
|
|
skip:
|
2008-08-11 03:30:22 -04:00
|
|
|
double_unlock_balance(this_rq, src_rq);
|
2008-01-25 15:08:07 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:22 -05:00
|
|
|
static void pre_schedule_rt(struct rq *rq, struct task_struct *prev)
|
2008-01-25 15:08:07 -05:00
|
|
|
{
|
|
|
|
/* Try to pull RT tasks here if we lower this rq's prio */
|
2008-01-25 15:08:17 -05:00
|
|
|
if (unlikely(rt_task(prev)) && rq->rt.highest_prio > prev->prio)
|
2008-01-25 15:08:07 -05:00
|
|
|
pull_rt_task(rq);
|
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:22 -05:00
|
|
|
static void post_schedule_rt(struct rq *rq)
|
2008-01-25 15:08:05 -05:00
|
|
|
{
|
|
|
|
/*
|
|
|
|
* If we have more than one rt_task queued, then
|
|
|
|
* see if we can push the other rt_tasks off to other CPUS.
|
|
|
|
* Note we may release the rq lock, and since
|
|
|
|
* the lock was owned by prev, we need to release it
|
|
|
|
* first via finish_lock_switch and then reaquire it here.
|
|
|
|
*/
|
2008-01-25 15:08:12 -05:00
|
|
|
if (unlikely(rq->rt.overloaded)) {
|
2008-01-25 15:08:05 -05:00
|
|
|
spin_lock_irq(&rq->lock);
|
|
|
|
push_rt_tasks(rq);
|
|
|
|
spin_unlock_irq(&rq->lock);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-04-23 07:13:29 -04:00
|
|
|
/*
|
|
|
|
* If we are not running and we are not going to reschedule soon, we should
|
|
|
|
* try to push tasks away now
|
|
|
|
*/
|
2008-01-25 15:08:22 -05:00
|
|
|
static void task_wake_up_rt(struct rq *rq, struct task_struct *p)
|
2008-01-25 15:08:07 -05:00
|
|
|
{
|
2008-01-25 15:08:22 -05:00
|
|
|
if (!task_running(rq, p) &&
|
2008-04-23 07:13:29 -04:00
|
|
|
!test_tsk_need_resched(rq->curr) &&
|
2008-01-25 15:08:12 -05:00
|
|
|
rq->rt.overloaded)
|
2008-01-25 15:08:07 -05:00
|
|
|
push_rt_tasks(rq);
|
|
|
|
}
|
|
|
|
|
sched: simplify move_tasks()
The move_tasks() function is currently multiplexed with two distinct
capabilities:
1. attempt to move a specified amount of weighted load from one run
queue to another; and
2. attempt to move a specified number of tasks from one run queue to
another.
The first of these capabilities is used in two places, load_balance()
and load_balance_idle(), and in both of these cases the return value of
move_tasks() is used purely to decide if tasks/load were moved and no
notice of the actual number of tasks moved is taken.
The second capability is used in exactly one place,
active_load_balance(), to attempt to move exactly one task and, as
before, the return value is only used as an indicator of success or failure.
This multiplexing of sched_task() was introduced, by me, as part of the
smpnice patches and was motivated by the fact that the alternative, one
function to move specified load and one to move a single task, would
have led to two functions of roughly the same complexity as the old
move_tasks() (or the new balance_tasks()). However, the new modular
design of the new CFS scheduler allows a simpler solution to be adopted
and this patch addresses that solution by:
1. adding a new function, move_one_task(), to be used by
active_load_balance(); and
2. making move_tasks() a single purpose function that tries to move a
specified weighted load and returns 1 for success and 0 for failure.
One of the consequences of these changes is that neither move_one_task()
or the new move_tasks() care how many tasks sched_class.load_balance()
moves and this enables its interface to be simplified by returning the
amount of load moved as its result and removing the load_moved pointer
from the argument list. This helps simplify the new move_tasks() and
slightly reduces the amount of work done in each of
sched_class.load_balance()'s implementations.
Further simplification, e.g. changes to balance_tasks(), are possible
but (slightly) complicated by the special needs of load_balance_fair()
so I've left them to a later patch (if this one gets accepted).
NB Since move_tasks() gets called with two run queue locks held even
small reductions in overhead are worthwhile.
[ mingo@elte.hu ]
this change also reduces code size nicely:
text data bss dec hex filename
39216 3618 24 42858 a76a sched.o.before
39173 3618 24 42815 a73f sched.o.after
Signed-off-by: Peter Williams <pwil3058@bigpond.net.au>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2007-08-09 05:16:46 -04:00
|
|
|
static unsigned long
|
2007-07-09 12:51:58 -04:00
|
|
|
load_balance_rt(struct rq *this_rq, int this_cpu, struct rq *busiest,
|
2007-10-24 12:23:51 -04:00
|
|
|
unsigned long max_load_move,
|
|
|
|
struct sched_domain *sd, enum cpu_idle_type idle,
|
|
|
|
int *all_pinned, int *this_best_prio)
|
2007-07-09 12:51:58 -04:00
|
|
|
{
|
2008-01-25 15:08:07 -05:00
|
|
|
/* don't touch RT tasks */
|
|
|
|
return 0;
|
2007-10-24 12:23:51 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
move_one_task_rt(struct rq *this_rq, int this_cpu, struct rq *busiest,
|
|
|
|
struct sched_domain *sd, enum cpu_idle_type idle)
|
|
|
|
{
|
2008-01-25 15:08:07 -05:00
|
|
|
/* don't touch RT tasks */
|
|
|
|
return 0;
|
2007-07-09 12:51:58 -04:00
|
|
|
}
|
2008-01-25 15:08:15 -05:00
|
|
|
|
2008-03-26 17:23:49 -04:00
|
|
|
static void set_cpus_allowed_rt(struct task_struct *p,
|
|
|
|
const cpumask_t *new_mask)
|
2008-01-25 15:08:07 -05:00
|
|
|
{
|
|
|
|
int weight = cpus_weight(*new_mask);
|
|
|
|
|
|
|
|
BUG_ON(!rt_task(p));
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Update the migration status of the RQ if we have an RT task
|
|
|
|
* which is running AND changing its weight value.
|
|
|
|
*/
|
2008-01-25 15:08:30 -05:00
|
|
|
if (p->se.on_rq && (weight != p->rt.nr_cpus_allowed)) {
|
2008-01-25 15:08:07 -05:00
|
|
|
struct rq *rq = task_rq(p);
|
|
|
|
|
2008-01-25 15:08:30 -05:00
|
|
|
if ((p->rt.nr_cpus_allowed <= 1) && (weight > 1)) {
|
2008-01-25 15:08:07 -05:00
|
|
|
rq->rt.rt_nr_migratory++;
|
2008-01-25 15:08:30 -05:00
|
|
|
} else if ((p->rt.nr_cpus_allowed > 1) && (weight <= 1)) {
|
2008-01-25 15:08:07 -05:00
|
|
|
BUG_ON(!rq->rt.rt_nr_migratory);
|
|
|
|
rq->rt.rt_nr_migratory--;
|
|
|
|
}
|
|
|
|
|
|
|
|
update_rt_migration(rq);
|
|
|
|
}
|
|
|
|
|
|
|
|
p->cpus_allowed = *new_mask;
|
2008-01-25 15:08:30 -05:00
|
|
|
p->rt.nr_cpus_allowed = weight;
|
2008-01-25 15:08:07 -05:00
|
|
|
}
|
2008-01-25 15:08:15 -05:00
|
|
|
|
2008-01-25 15:08:18 -05:00
|
|
|
/* Assumes rq->lock is held */
|
2008-06-04 15:04:05 -04:00
|
|
|
static void rq_online_rt(struct rq *rq)
|
2008-01-25 15:08:18 -05:00
|
|
|
{
|
|
|
|
if (rq->rt.overloaded)
|
|
|
|
rt_set_overload(rq);
|
2008-05-12 15:21:01 -04:00
|
|
|
|
2008-06-05 08:49:58 -04:00
|
|
|
__enable_runtime(rq);
|
|
|
|
|
2008-05-12 15:21:01 -04:00
|
|
|
cpupri_set(&rq->rd->cpupri, rq->cpu, rq->rt.highest_prio);
|
2008-01-25 15:08:18 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Assumes rq->lock is held */
|
2008-06-04 15:04:05 -04:00
|
|
|
static void rq_offline_rt(struct rq *rq)
|
2008-01-25 15:08:18 -05:00
|
|
|
{
|
|
|
|
if (rq->rt.overloaded)
|
|
|
|
rt_clear_overload(rq);
|
2008-05-12 15:21:01 -04:00
|
|
|
|
2008-06-05 08:49:58 -04:00
|
|
|
__disable_runtime(rq);
|
|
|
|
|
2008-05-12 15:21:01 -04:00
|
|
|
cpupri_set(&rq->rd->cpupri, rq->cpu, CPUPRI_INVALID);
|
2008-01-25 15:08:18 -05:00
|
|
|
}
|
2008-01-25 15:08:22 -05:00
|
|
|
|
|
|
|
/*
|
|
|
|
* When switch from the rt queue, we bring ourselves to a position
|
|
|
|
* that we might want to pull RT tasks from other runqueues.
|
|
|
|
*/
|
|
|
|
static void switched_from_rt(struct rq *rq, struct task_struct *p,
|
|
|
|
int running)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* If there are other RT tasks then we will reschedule
|
|
|
|
* and the scheduling of the other RT tasks will handle
|
|
|
|
* the balancing. But if we are the last RT task
|
|
|
|
* we may need to handle the pulling of RT tasks
|
|
|
|
* now.
|
|
|
|
*/
|
|
|
|
if (!rq->rt.rt_nr_running)
|
|
|
|
pull_rt_task(rq);
|
|
|
|
}
|
|
|
|
#endif /* CONFIG_SMP */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* When switching a task to RT, we may overload the runqueue
|
|
|
|
* with RT tasks. In this case we try to push them off to
|
|
|
|
* other runqueues.
|
|
|
|
*/
|
|
|
|
static void switched_to_rt(struct rq *rq, struct task_struct *p,
|
|
|
|
int running)
|
|
|
|
{
|
|
|
|
int check_resched = 1;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If we are already running, then there's nothing
|
|
|
|
* that needs to be done. But if we are not running
|
|
|
|
* we may need to preempt the current running task.
|
|
|
|
* If that current running task is also an RT task
|
|
|
|
* then see if we can move to another run queue.
|
|
|
|
*/
|
|
|
|
if (!running) {
|
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
if (rq->rt.overloaded && push_rt_task(rq) &&
|
|
|
|
/* Don't resched if we changed runqueues */
|
|
|
|
rq != task_rq(p))
|
|
|
|
check_resched = 0;
|
|
|
|
#endif /* CONFIG_SMP */
|
|
|
|
if (check_resched && p->prio < rq->curr->prio)
|
|
|
|
resched_task(rq->curr);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Priority of the task has changed. This may cause
|
|
|
|
* us to initiate a push or pull.
|
|
|
|
*/
|
|
|
|
static void prio_changed_rt(struct rq *rq, struct task_struct *p,
|
|
|
|
int oldprio, int running)
|
|
|
|
{
|
|
|
|
if (running) {
|
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
/*
|
|
|
|
* If our priority decreases while running, we
|
|
|
|
* may need to pull tasks to this runqueue.
|
|
|
|
*/
|
|
|
|
if (oldprio < p->prio)
|
|
|
|
pull_rt_task(rq);
|
|
|
|
/*
|
|
|
|
* If there's a higher priority task waiting to run
|
2008-03-05 10:00:12 -05:00
|
|
|
* then reschedule. Note, the above pull_rt_task
|
|
|
|
* can release the rq lock and p could migrate.
|
|
|
|
* Only reschedule if p is still on the same runqueue.
|
2008-01-25 15:08:22 -05:00
|
|
|
*/
|
2008-03-05 10:00:12 -05:00
|
|
|
if (p->prio > rq->rt.highest_prio && rq->curr == p)
|
2008-01-25 15:08:22 -05:00
|
|
|
resched_task(p);
|
|
|
|
#else
|
|
|
|
/* For UP simply resched on drop of prio */
|
|
|
|
if (oldprio < p->prio)
|
|
|
|
resched_task(p);
|
2008-01-25 15:08:05 -05:00
|
|
|
#endif /* CONFIG_SMP */
|
2008-01-25 15:08:22 -05:00
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* This task is not running, but if it is
|
|
|
|
* greater than the current running task
|
|
|
|
* then reschedule.
|
|
|
|
*/
|
|
|
|
if (p->prio < rq->curr->prio)
|
|
|
|
resched_task(rq->curr);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-01-25 15:08:27 -05:00
|
|
|
static void watchdog(struct rq *rq, struct task_struct *p)
|
|
|
|
{
|
|
|
|
unsigned long soft, hard;
|
|
|
|
|
|
|
|
if (!p->signal)
|
|
|
|
return;
|
|
|
|
|
|
|
|
soft = p->signal->rlim[RLIMIT_RTTIME].rlim_cur;
|
|
|
|
hard = p->signal->rlim[RLIMIT_RTTIME].rlim_max;
|
|
|
|
|
|
|
|
if (soft != RLIM_INFINITY) {
|
|
|
|
unsigned long next;
|
|
|
|
|
|
|
|
p->rt.timeout++;
|
|
|
|
next = DIV_ROUND_UP(min(soft, hard), USEC_PER_SEC/HZ);
|
2008-01-25 15:08:32 -05:00
|
|
|
if (p->rt.timeout > next)
|
2008-01-25 15:08:27 -05:00
|
|
|
p->it_sched_expires = p->se.sum_exec_runtime;
|
|
|
|
}
|
|
|
|
}
|
2007-07-09 12:51:58 -04:00
|
|
|
|
2008-01-25 15:08:29 -05:00
|
|
|
static void task_tick_rt(struct rq *rq, struct task_struct *p, int queued)
|
2007-07-09 12:51:58 -04:00
|
|
|
{
|
2007-12-20 09:01:17 -05:00
|
|
|
update_curr_rt(rq);
|
|
|
|
|
2008-01-25 15:08:27 -05:00
|
|
|
watchdog(rq, p);
|
|
|
|
|
2007-07-09 12:51:58 -04:00
|
|
|
/*
|
|
|
|
* RR tasks need a special form of timeslice management.
|
|
|
|
* FIFO tasks have no timeslices.
|
|
|
|
*/
|
|
|
|
if (p->policy != SCHED_RR)
|
|
|
|
return;
|
|
|
|
|
2008-01-25 15:08:27 -05:00
|
|
|
if (--p->rt.time_slice)
|
2007-07-09 12:51:58 -04:00
|
|
|
return;
|
|
|
|
|
2008-01-25 15:08:27 -05:00
|
|
|
p->rt.time_slice = DEF_TIMESLICE;
|
2007-07-09 12:51:58 -04:00
|
|
|
|
2007-08-24 14:39:10 -04:00
|
|
|
/*
|
|
|
|
* Requeue to the end of queue if we are not the only element
|
|
|
|
* on the queue:
|
|
|
|
*/
|
2008-01-25 15:08:27 -05:00
|
|
|
if (p->rt.run_list.prev != p->rt.run_list.next) {
|
2008-07-01 17:32:15 -04:00
|
|
|
requeue_task_rt(rq, p, 0);
|
2007-08-24 14:39:10 -04:00
|
|
|
set_tsk_need_resched(p);
|
|
|
|
}
|
2007-07-09 12:51:58 -04:00
|
|
|
}
|
|
|
|
|
2007-10-15 11:00:08 -04:00
|
|
|
static void set_curr_task_rt(struct rq *rq)
|
|
|
|
{
|
|
|
|
struct task_struct *p = rq->curr;
|
|
|
|
|
|
|
|
p->se.exec_start = rq->clock;
|
|
|
|
}
|
|
|
|
|
2008-04-25 13:53:13 -04:00
|
|
|
static const struct sched_class rt_sched_class = {
|
2007-10-15 11:00:12 -04:00
|
|
|
.next = &fair_sched_class,
|
2007-07-09 12:51:58 -04:00
|
|
|
.enqueue_task = enqueue_task_rt,
|
|
|
|
.dequeue_task = dequeue_task_rt,
|
|
|
|
.yield_task = yield_task_rt,
|
2008-01-25 15:08:09 -05:00
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
.select_task_rq = select_task_rq_rt,
|
|
|
|
#endif /* CONFIG_SMP */
|
2007-07-09 12:51:58 -04:00
|
|
|
|
|
|
|
.check_preempt_curr = check_preempt_curr_rt,
|
|
|
|
|
|
|
|
.pick_next_task = pick_next_task_rt,
|
|
|
|
.put_prev_task = put_prev_task_rt,
|
|
|
|
|
2007-10-24 12:23:51 -04:00
|
|
|
#ifdef CONFIG_SMP
|
2007-07-09 12:51:58 -04:00
|
|
|
.load_balance = load_balance_rt,
|
2007-10-24 12:23:51 -04:00
|
|
|
.move_one_task = move_one_task_rt,
|
2008-01-25 15:08:07 -05:00
|
|
|
.set_cpus_allowed = set_cpus_allowed_rt,
|
2008-06-04 15:04:05 -04:00
|
|
|
.rq_online = rq_online_rt,
|
|
|
|
.rq_offline = rq_offline_rt,
|
2008-01-25 15:08:22 -05:00
|
|
|
.pre_schedule = pre_schedule_rt,
|
|
|
|
.post_schedule = post_schedule_rt,
|
|
|
|
.task_wake_up = task_wake_up_rt,
|
2008-01-25 15:08:22 -05:00
|
|
|
.switched_from = switched_from_rt,
|
2007-10-24 12:23:51 -04:00
|
|
|
#endif
|
2007-07-09 12:51:58 -04:00
|
|
|
|
2007-10-15 11:00:08 -04:00
|
|
|
.set_curr_task = set_curr_task_rt,
|
2007-07-09 12:51:58 -04:00
|
|
|
.task_tick = task_tick_rt,
|
2008-01-25 15:08:22 -05:00
|
|
|
|
|
|
|
.prio_changed = prio_changed_rt,
|
|
|
|
.switched_to = switched_to_rt,
|
2007-07-09 12:51:58 -04:00
|
|
|
};
|
2008-06-19 08:22:24 -04:00
|
|
|
|
|
|
|
#ifdef CONFIG_SCHED_DEBUG
|
|
|
|
extern void print_rt_rq(struct seq_file *m, int cpu, struct rt_rq *rt_rq);
|
|
|
|
|
|
|
|
static void print_rt_stats(struct seq_file *m, int cpu)
|
|
|
|
{
|
|
|
|
struct rt_rq *rt_rq;
|
|
|
|
|
|
|
|
rcu_read_lock();
|
|
|
|
for_each_leaf_rt_rq(rt_rq, cpu_rq(cpu))
|
|
|
|
print_rt_rq(m, cpu, rt_rq);
|
|
|
|
rcu_read_unlock();
|
|
|
|
}
|
2008-06-24 14:09:43 -04:00
|
|
|
#endif /* CONFIG_SCHED_DEBUG */
|