linux/kernel/sched/rt.c

2054 lines
46 KiB
C
Raw Normal View History

/*
* Real-Time Scheduling Class (mapped to the SCHED_FIFO and SCHED_RR
* policies)
*/
#include "sched.h"
#include <linux/slab.h>
static int do_sched_rt_period_timer(struct rt_bandwidth *rt_b, int overrun);
struct rt_bandwidth def_rt_bandwidth;
static enum hrtimer_restart sched_rt_period_timer(struct hrtimer *timer)
{
struct rt_bandwidth *rt_b =
container_of(timer, struct rt_bandwidth, rt_period_timer);
ktime_t now;
int overrun;
int idle = 0;
for (;;) {
now = hrtimer_cb_get_time(timer);
overrun = hrtimer_forward(timer, now, rt_b->rt_period);
if (!overrun)
break;
idle = do_sched_rt_period_timer(rt_b, overrun);
}
return idle ? HRTIMER_NORESTART : HRTIMER_RESTART;
}
void init_rt_bandwidth(struct rt_bandwidth *rt_b, u64 period, u64 runtime)
{
rt_b->rt_period = ns_to_ktime(period);
rt_b->rt_runtime = runtime;
raw_spin_lock_init(&rt_b->rt_runtime_lock);
hrtimer_init(&rt_b->rt_period_timer,
CLOCK_MONOTONIC, HRTIMER_MODE_REL);
rt_b->rt_period_timer.function = sched_rt_period_timer;
}
static void start_rt_bandwidth(struct rt_bandwidth *rt_b)
{
if (!rt_bandwidth_enabled() || rt_b->rt_runtime == RUNTIME_INF)
return;
if (hrtimer_active(&rt_b->rt_period_timer))
return;
raw_spin_lock(&rt_b->rt_runtime_lock);
start_bandwidth_timer(&rt_b->rt_period_timer, rt_b->rt_period);
raw_spin_unlock(&rt_b->rt_runtime_lock);
}
void init_rt_rq(struct rt_rq *rt_rq, struct rq *rq)
{
struct rt_prio_array *array;
int i;
array = &rt_rq->active;
for (i = 0; i < MAX_RT_PRIO; i++) {
INIT_LIST_HEAD(array->queue + i);
__clear_bit(i, array->bitmap);
}
/* delimiter for bitsearch: */
__set_bit(MAX_RT_PRIO, array->bitmap);
#if defined CONFIG_SMP
rt_rq->highest_prio.curr = MAX_RT_PRIO;
rt_rq->highest_prio.next = MAX_RT_PRIO;
rt_rq->rt_nr_migratory = 0;
rt_rq->overloaded = 0;
plist_head_init(&rt_rq->pushable_tasks);
#endif
rt_rq->rt_time = 0;
rt_rq->rt_throttled = 0;
rt_rq->rt_runtime = 0;
raw_spin_lock_init(&rt_rq->rt_runtime_lock);
}
#ifdef CONFIG_RT_GROUP_SCHED
static void destroy_rt_bandwidth(struct rt_bandwidth *rt_b)
{
hrtimer_cancel(&rt_b->rt_period_timer);
}
#define rt_entity_is_task(rt_se) (!(rt_se)->my_q)
static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se)
{
#ifdef CONFIG_SCHED_DEBUG
WARN_ON_ONCE(!rt_entity_is_task(rt_se));
#endif
return container_of(rt_se, struct task_struct, rt);
}
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;
}
void free_rt_sched_group(struct task_group *tg)
{
int i;
if (tg->rt_se)
destroy_rt_bandwidth(&tg->rt_bandwidth);
for_each_possible_cpu(i) {
if (tg->rt_rq)
kfree(tg->rt_rq[i]);
if (tg->rt_se)
kfree(tg->rt_se[i]);
}
kfree(tg->rt_rq);
kfree(tg->rt_se);
}
void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq,
struct sched_rt_entity *rt_se, int cpu,
struct sched_rt_entity *parent)
{
struct rq *rq = cpu_rq(cpu);
rt_rq->highest_prio.curr = MAX_RT_PRIO;
rt_rq->rt_nr_boosted = 0;
rt_rq->rq = rq;
rt_rq->tg = tg;
tg->rt_rq[cpu] = rt_rq;
tg->rt_se[cpu] = rt_se;
if (!rt_se)
return;
if (!parent)
rt_se->rt_rq = &rq->rt;
else
rt_se->rt_rq = parent->my_q;
rt_se->my_q = rt_rq;
rt_se->parent = parent;
INIT_LIST_HEAD(&rt_se->run_list);
}
int alloc_rt_sched_group(struct task_group *tg, struct task_group *parent)
{
struct rt_rq *rt_rq;
struct sched_rt_entity *rt_se;
int i;
tg->rt_rq = kzalloc(sizeof(rt_rq) * nr_cpu_ids, GFP_KERNEL);
if (!tg->rt_rq)
goto err;
tg->rt_se = kzalloc(sizeof(rt_se) * nr_cpu_ids, GFP_KERNEL);
if (!tg->rt_se)
goto err;
init_rt_bandwidth(&tg->rt_bandwidth,
ktime_to_ns(def_rt_bandwidth.rt_period), 0);
for_each_possible_cpu(i) {
rt_rq = kzalloc_node(sizeof(struct rt_rq),
GFP_KERNEL, cpu_to_node(i));
if (!rt_rq)
goto err;
rt_se = kzalloc_node(sizeof(struct sched_rt_entity),
GFP_KERNEL, cpu_to_node(i));
if (!rt_se)
goto err_free_rq;
init_rt_rq(rt_rq, cpu_rq(i));
rt_rq->rt_runtime = tg->rt_bandwidth.rt_runtime;
init_tg_rt_entry(tg, rt_rq, rt_se, i, parent->rt_se[i]);
}
return 1;
err_free_rq:
kfree(rt_rq);
err:
return 0;
}
#else /* CONFIG_RT_GROUP_SCHED */
#define rt_entity_is_task(rt_se) (1)
static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se)
{
return container_of(rt_se, struct task_struct, rt);
}
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;
}
void free_rt_sched_group(struct task_group *tg) { }
int alloc_rt_sched_group(struct task_group *tg, struct task_group *parent)
{
return 1;
}
#endif /* CONFIG_RT_GROUP_SCHED */
#ifdef CONFIG_SMP
static inline int rt_overloaded(struct rq *rq)
{
return atomic_read(&rq->rd->rto_count);
}
static inline void rt_set_overload(struct rq *rq)
{
if (!rq->online)
return;
cpumask_set_cpu(rq->cpu, rq->rd->rto_mask);
/*
* 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();
atomic_inc(&rq->rd->rto_count);
}
static inline void rt_clear_overload(struct rq *rq)
{
if (!rq->online)
return;
/* the order here really doesn't matter */
atomic_dec(&rq->rd->rto_count);
cpumask_clear_cpu(rq->cpu, rq->rd->rto_mask);
}
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
static void update_rt_migration(struct rt_rq *rt_rq)
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
{
if (rt_rq->rt_nr_migratory && rt_rq->rt_nr_total > 1) {
if (!rt_rq->overloaded) {
rt_set_overload(rq_of_rt_rq(rt_rq));
rt_rq->overloaded = 1;
}
} else if (rt_rq->overloaded) {
rt_clear_overload(rq_of_rt_rq(rt_rq));
rt_rq->overloaded = 0;
}
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
}
static void inc_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
{
if (!rt_entity_is_task(rt_se))
return;
rt_rq = &rq_of_rt_rq(rt_rq)->rt;
rt_rq->rt_nr_total++;
if (rt_se->nr_cpus_allowed > 1)
rt_rq->rt_nr_migratory++;
update_rt_migration(rt_rq);
}
static void dec_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
{
if (!rt_entity_is_task(rt_se))
return;
rt_rq = &rq_of_rt_rq(rt_rq)->rt;
rt_rq->rt_nr_total--;
if (rt_se->nr_cpus_allowed > 1)
rt_rq->rt_nr_migratory--;
update_rt_migration(rt_rq);
}
sched: Use pushable_tasks to determine next highest prio Hillf Danton proposed a patch (see link) that cleaned up the sched_rt code that calculates the priority of the next highest priority task to be used in finding run queues to pull from. His patch removed the calculating of the next prio to just use the current prio when deteriming if we should examine a run queue to pull from. The problem with his patch was that it caused more false checks. Because we check a run queue for pushable tasks if the current priority of that run queue is higher in priority than the task about to run on our run queue. But after grabbing the locks and doing the real check, we find that there may not be a task that has a higher prio task to pull. Thus the locks were taken with nothing to do. I added some trace_printks() to record when and how many times the run queue locks were taken to check for pullable tasks, compared to how many times we pulled a task. With the current method, it was: 3806 locks taken vs 2812 pulled tasks With Hillf's patch: 6728 locks taken vs 2804 pulled tasks The number of times locks were taken to pull a task went up almost double with no more success rate. But his patch did get me thinking. When we look at the priority of the highest task to consider taking the locks to do a pull, a failure to pull can be one of the following: (in order of most likely) o RT task was pushed off already between the check and taking the lock o Waiting RT task can not be migrated o RT task's CPU affinity does not include the target run queue's CPU o RT task's priority changed between the check and taking the lock And with Hillf's patch, the thing that caused most of the failures, is the RT task to pull was not at the right priority to pull (not greater than the current RT task priority on the target run queue). Most of the above cases we can't help. But the current method does not check if the next highest prio RT task can be migrated or not, and if it can not, we still grab the locks to do the test (we don't find out about this fact until after we have the locks). I thought about this case, and realized that the pushable task plist that is maintained only holds RT tasks that can migrate. If we move the calculating of the next highest prio task from the inc/dec_rt_task() functions into the queuing of the pushable tasks, then we only measure the priorities of those tasks that we push, and we get this basically for free. Not only does this patch make the code a little more efficient, it cleans it up and makes it a little simpler. Thanks to Hillf Danton for inspiring me on this patch. Signed-off-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Hillf Danton <dhillf@gmail.com> Cc: Gregory Haskins <ghaskins@novell.com> Link: http://lkml.kernel.org/r/BANLkTimQ67180HxCx5vgMqumqw1EkFh3qg@mail.gmail.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-06-17 01:55:23 +00:00
static inline int has_pushable_tasks(struct rq *rq)
{
return !plist_head_empty(&rq->rt.pushable_tasks);
}
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
static void enqueue_pushable_task(struct rq *rq, struct task_struct *p)
{
plist_del(&p->pushable_tasks, &rq->rt.pushable_tasks);
plist_node_init(&p->pushable_tasks, p->prio);
plist_add(&p->pushable_tasks, &rq->rt.pushable_tasks);
sched: Use pushable_tasks to determine next highest prio Hillf Danton proposed a patch (see link) that cleaned up the sched_rt code that calculates the priority of the next highest priority task to be used in finding run queues to pull from. His patch removed the calculating of the next prio to just use the current prio when deteriming if we should examine a run queue to pull from. The problem with his patch was that it caused more false checks. Because we check a run queue for pushable tasks if the current priority of that run queue is higher in priority than the task about to run on our run queue. But after grabbing the locks and doing the real check, we find that there may not be a task that has a higher prio task to pull. Thus the locks were taken with nothing to do. I added some trace_printks() to record when and how many times the run queue locks were taken to check for pullable tasks, compared to how many times we pulled a task. With the current method, it was: 3806 locks taken vs 2812 pulled tasks With Hillf's patch: 6728 locks taken vs 2804 pulled tasks The number of times locks were taken to pull a task went up almost double with no more success rate. But his patch did get me thinking. When we look at the priority of the highest task to consider taking the locks to do a pull, a failure to pull can be one of the following: (in order of most likely) o RT task was pushed off already between the check and taking the lock o Waiting RT task can not be migrated o RT task's CPU affinity does not include the target run queue's CPU o RT task's priority changed between the check and taking the lock And with Hillf's patch, the thing that caused most of the failures, is the RT task to pull was not at the right priority to pull (not greater than the current RT task priority on the target run queue). Most of the above cases we can't help. But the current method does not check if the next highest prio RT task can be migrated or not, and if it can not, we still grab the locks to do the test (we don't find out about this fact until after we have the locks). I thought about this case, and realized that the pushable task plist that is maintained only holds RT tasks that can migrate. If we move the calculating of the next highest prio task from the inc/dec_rt_task() functions into the queuing of the pushable tasks, then we only measure the priorities of those tasks that we push, and we get this basically for free. Not only does this patch make the code a little more efficient, it cleans it up and makes it a little simpler. Thanks to Hillf Danton for inspiring me on this patch. Signed-off-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Hillf Danton <dhillf@gmail.com> Cc: Gregory Haskins <ghaskins@novell.com> Link: http://lkml.kernel.org/r/BANLkTimQ67180HxCx5vgMqumqw1EkFh3qg@mail.gmail.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-06-17 01:55:23 +00:00
/* Update the highest prio pushable task */
if (p->prio < rq->rt.highest_prio.next)
rq->rt.highest_prio.next = p->prio;
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
}
static void dequeue_pushable_task(struct rq *rq, struct task_struct *p)
{
plist_del(&p->pushable_tasks, &rq->rt.pushable_tasks);
sched: Use pushable_tasks to determine next highest prio Hillf Danton proposed a patch (see link) that cleaned up the sched_rt code that calculates the priority of the next highest priority task to be used in finding run queues to pull from. His patch removed the calculating of the next prio to just use the current prio when deteriming if we should examine a run queue to pull from. The problem with his patch was that it caused more false checks. Because we check a run queue for pushable tasks if the current priority of that run queue is higher in priority than the task about to run on our run queue. But after grabbing the locks and doing the real check, we find that there may not be a task that has a higher prio task to pull. Thus the locks were taken with nothing to do. I added some trace_printks() to record when and how many times the run queue locks were taken to check for pullable tasks, compared to how many times we pulled a task. With the current method, it was: 3806 locks taken vs 2812 pulled tasks With Hillf's patch: 6728 locks taken vs 2804 pulled tasks The number of times locks were taken to pull a task went up almost double with no more success rate. But his patch did get me thinking. When we look at the priority of the highest task to consider taking the locks to do a pull, a failure to pull can be one of the following: (in order of most likely) o RT task was pushed off already between the check and taking the lock o Waiting RT task can not be migrated o RT task's CPU affinity does not include the target run queue's CPU o RT task's priority changed between the check and taking the lock And with Hillf's patch, the thing that caused most of the failures, is the RT task to pull was not at the right priority to pull (not greater than the current RT task priority on the target run queue). Most of the above cases we can't help. But the current method does not check if the next highest prio RT task can be migrated or not, and if it can not, we still grab the locks to do the test (we don't find out about this fact until after we have the locks). I thought about this case, and realized that the pushable task plist that is maintained only holds RT tasks that can migrate. If we move the calculating of the next highest prio task from the inc/dec_rt_task() functions into the queuing of the pushable tasks, then we only measure the priorities of those tasks that we push, and we get this basically for free. Not only does this patch make the code a little more efficient, it cleans it up and makes it a little simpler. Thanks to Hillf Danton for inspiring me on this patch. Signed-off-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Hillf Danton <dhillf@gmail.com> Cc: Gregory Haskins <ghaskins@novell.com> Link: http://lkml.kernel.org/r/BANLkTimQ67180HxCx5vgMqumqw1EkFh3qg@mail.gmail.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-06-17 01:55:23 +00:00
/* Update the new highest prio pushable task */
if (has_pushable_tasks(rq)) {
p = plist_first_entry(&rq->rt.pushable_tasks,
struct task_struct, pushable_tasks);
rq->rt.highest_prio.next = p->prio;
} else
rq->rt.highest_prio.next = MAX_RT_PRIO;
}
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
#else
static inline void enqueue_pushable_task(struct rq *rq, struct task_struct *p)
{
}
static inline void dequeue_pushable_task(struct rq *rq, struct task_struct *p)
{
}
static inline
void inc_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
{
}
static inline
void dec_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
{
}
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
#endif /* CONFIG_SMP */
static inline int on_rt_rq(struct sched_rt_entity *rt_se)
{
return !list_empty(&rt_se->run_list);
}
#ifdef CONFIG_RT_GROUP_SCHED
static inline u64 sched_rt_runtime(struct rt_rq *rt_rq)
{
if (!rt_rq->tg)
return RUNTIME_INF;
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);
}
typedef struct task_group *rt_rq_iter_t;
static inline struct task_group *next_task_group(struct task_group *tg)
{
do {
tg = list_entry_rcu(tg->list.next,
typeof(struct task_group), list);
} while (&tg->list != &task_groups && task_group_is_autogroup(tg));
if (&tg->list == &task_groups)
tg = NULL;
return tg;
}
#define for_each_rt_rq(rt_rq, iter, rq) \
for (iter = container_of(&task_groups, typeof(*iter), list); \
(iter = next_task_group(iter)) && \
(rt_rq = iter->rt_rq[cpu_of(rq)]);)
static inline void list_add_leaf_rt_rq(struct rt_rq *rt_rq)
{
list_add_rcu(&rt_rq->leaf_rt_rq_list,
&rq_of_rt_rq(rt_rq)->leaf_rt_rq_list);
}
static inline void list_del_leaf_rt_rq(struct rt_rq *rt_rq)
{
list_del_rcu(&rt_rq->leaf_rt_rq_list);
}
#define for_each_leaf_rt_rq(rt_rq, rq) \
list_for_each_entry_rcu(rt_rq, &rq->leaf_rt_rq_list, leaf_rt_rq_list)
#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, bool head);
static void dequeue_rt_entity(struct sched_rt_entity *rt_se);
static void sched_rt_rq_enqueue(struct rt_rq *rt_rq)
{
sched_rt.c: resch needed in rt_rq_enqueue() for the root rt_rq While working on the new version of the code for SCHED_SPORADIC I noticed something strange in the present throttling mechanism. More specifically in the throttling timer handler in sched_rt.c (do_sched_rt_period_timer()) and in rt_rq_enqueue(). The problem is that, when unthrottling a runqueue, rt_rq_enqueue() only asks for rescheduling if the runqueue has a sched_entity associated to it (i.e., rt_rq->rt_se != NULL). Now, if the runqueue is the root rq (which has a rt_se = NULL) rescheduling does not take place, and it is delayed to some undefined instant in the future. This imply some random bandwidth usage by the RT tasks under throttling. For instance, setting rt_runtime_us/rt_period_us = 950ms/1000ms an RT task will get less than 95%. In our tests we got something varying between 70% to 95%. Using smaller time values, e.g., 95ms/100ms, things are even worse, and I can see values also going down to 20-25%!! The tests we performed are simply running 'yes' as a SCHED_FIFO task, and checking the CPU usage with top, but we can investigate thoroughly if you think it is needed. Things go much better, for us, with the attached patch... Don't know if it is the best approach, but it solved the issue for us. Signed-off-by: Dario Faggioli <raistlin@linux.it> Signed-off-by: Michael Trimarchi <trimarchimichael@yahoo.it> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: <stable@kernel.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-10-03 15:40:46 +00:00
struct task_struct *curr = rq_of_rt_rq(rt_rq)->curr;
struct sched_rt_entity *rt_se;
int cpu = cpu_of(rq_of_rt_rq(rt_rq));
rt_se = rt_rq->tg->rt_se[cpu];
sched_rt.c: resch needed in rt_rq_enqueue() for the root rt_rq While working on the new version of the code for SCHED_SPORADIC I noticed something strange in the present throttling mechanism. More specifically in the throttling timer handler in sched_rt.c (do_sched_rt_period_timer()) and in rt_rq_enqueue(). The problem is that, when unthrottling a runqueue, rt_rq_enqueue() only asks for rescheduling if the runqueue has a sched_entity associated to it (i.e., rt_rq->rt_se != NULL). Now, if the runqueue is the root rq (which has a rt_se = NULL) rescheduling does not take place, and it is delayed to some undefined instant in the future. This imply some random bandwidth usage by the RT tasks under throttling. For instance, setting rt_runtime_us/rt_period_us = 950ms/1000ms an RT task will get less than 95%. In our tests we got something varying between 70% to 95%. Using smaller time values, e.g., 95ms/100ms, things are even worse, and I can see values also going down to 20-25%!! The tests we performed are simply running 'yes' as a SCHED_FIFO task, and checking the CPU usage with top, but we can investigate thoroughly if you think it is needed. Things go much better, for us, with the attached patch... Don't know if it is the best approach, but it solved the issue for us. Signed-off-by: Dario Faggioli <raistlin@linux.it> Signed-off-by: Michael Trimarchi <trimarchimichael@yahoo.it> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: <stable@kernel.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-10-03 15:40:46 +00:00
if (rt_rq->rt_nr_running) {
if (rt_se && !on_rt_rq(rt_se))
enqueue_rt_entity(rt_se, false);
if (rt_rq->highest_prio.curr < curr->prio)
resched_task(curr);
}
}
static void sched_rt_rq_dequeue(struct rt_rq *rt_rq)
{
struct sched_rt_entity *rt_se;
int cpu = cpu_of(rq_of_rt_rq(rt_rq));
rt_se = rt_rq->tg->rt_se[cpu];
if (rt_se && on_rt_rq(rt_se))
dequeue_rt_entity(rt_se);
}
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;
}
#ifdef CONFIG_SMP
static inline const struct cpumask *sched_rt_period_mask(void)
{
return cpu_rq(smp_processor_id())->rd->span;
}
#else
static inline const struct cpumask *sched_rt_period_mask(void)
{
return cpu_online_mask;
}
#endif
static inline
struct rt_rq *sched_rt_period_rt_rq(struct rt_bandwidth *rt_b, int cpu)
{
return container_of(rt_b, struct task_group, rt_bandwidth)->rt_rq[cpu];
}
static inline struct rt_bandwidth *sched_rt_bandwidth(struct rt_rq *rt_rq)
{
return &rt_rq->tg->rt_bandwidth;
}
#else /* !CONFIG_RT_GROUP_SCHED */
static inline u64 sched_rt_runtime(struct rt_rq *rt_rq)
{
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);
}
typedef struct rt_rq *rt_rq_iter_t;
#define for_each_rt_rq(rt_rq, iter, rq) \
for ((void) iter, rt_rq = &rq->rt; rt_rq; rt_rq = NULL)
static inline void list_add_leaf_rt_rq(struct rt_rq *rt_rq)
{
}
static inline void list_del_leaf_rt_rq(struct rt_rq *rt_rq)
{
}
#define for_each_leaf_rt_rq(rt_rq, rq) \
for (rt_rq = &rq->rt; rt_rq; rt_rq = NULL)
#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;
}
static inline void sched_rt_rq_enqueue(struct rt_rq *rt_rq)
{
if (rt_rq->rt_nr_running)
resched_task(rq_of_rt_rq(rt_rq)->curr);
}
static inline void sched_rt_rq_dequeue(struct rt_rq *rt_rq)
{
}
static inline int rt_rq_throttled(struct rt_rq *rt_rq)
{
return rt_rq->rt_throttled;
}
static inline const struct cpumask *sched_rt_period_mask(void)
{
return cpu_online_mask;
}
static inline
struct rt_rq *sched_rt_period_rt_rq(struct rt_bandwidth *rt_b, int cpu)
{
return &cpu_rq(cpu)->rt;
}
static inline struct rt_bandwidth *sched_rt_bandwidth(struct rt_rq *rt_rq)
{
return &def_rt_bandwidth;
}
#endif /* CONFIG_RT_GROUP_SCHED */
#ifdef CONFIG_SMP
/*
* We ran out of runtime, see if we can borrow some from our neighbours.
*/
static int do_balance_runtime(struct rt_rq *rt_rq)
{
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 = cpumask_weight(rd->span);
raw_spin_lock(&rt_b->rt_runtime_lock);
rt_period = ktime_to_ns(rt_b->rt_period);
for_each_cpu(i, rd->span) {
struct rt_rq *iter = sched_rt_period_rt_rq(rt_b, i);
s64 diff;
if (iter == rt_rq)
continue;
raw_spin_lock(&iter->rt_runtime_lock);
/*
* Either all rqs have inf runtime and there's nothing to steal
* or __disable_runtime() below sets a specific rq to inf to
* indicate its been disabled and disalow stealing.
*/
if (iter->rt_runtime == RUNTIME_INF)
goto next;
/*
* From runqueues with spare time, take 1/n part of their
* spare time, but no more than our period.
*/
diff = iter->rt_runtime - iter->rt_time;
if (diff > 0) {
diff = div_u64((u64)diff, weight);
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) {
raw_spin_unlock(&iter->rt_runtime_lock);
break;
}
}
next:
raw_spin_unlock(&iter->rt_runtime_lock);
}
raw_spin_unlock(&rt_b->rt_runtime_lock);
return more;
}
/*
* Ensure this RQ takes back all the runtime it lend to its neighbours.
*/
static void __disable_runtime(struct rq *rq)
{
struct root_domain *rd = rq->rd;
rt_rq_iter_t iter;
struct rt_rq *rt_rq;
if (unlikely(!scheduler_running))
return;
for_each_rt_rq(rt_rq, iter, rq) {
struct rt_bandwidth *rt_b = sched_rt_bandwidth(rt_rq);
s64 want;
int i;
raw_spin_lock(&rt_b->rt_runtime_lock);
raw_spin_lock(&rt_rq->rt_runtime_lock);
/*
* Either we're all inf and nobody needs to borrow, or we're
* already disabled and thus have nothing to do, or we have
* exactly the right amount of runtime to take out.
*/
if (rt_rq->rt_runtime == RUNTIME_INF ||
rt_rq->rt_runtime == rt_b->rt_runtime)
goto balanced;
raw_spin_unlock(&rt_rq->rt_runtime_lock);
/*
* Calculate the difference between what we started out with
* and what we current have, that's the amount of runtime
* we lend and now have to reclaim.
*/
want = rt_b->rt_runtime - rt_rq->rt_runtime;
/*
* Greedy reclaim, take back as much as we can.
*/
for_each_cpu(i, rd->span) {
struct rt_rq *iter = sched_rt_period_rt_rq(rt_b, i);
s64 diff;
/*
* Can't reclaim from ourselves or disabled runqueues.
*/
if (iter == rt_rq || iter->rt_runtime == RUNTIME_INF)
continue;
raw_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;
}
raw_spin_unlock(&iter->rt_runtime_lock);
if (!want)
break;
}
raw_spin_lock(&rt_rq->rt_runtime_lock);
/*
* We cannot be left wanting - that would mean some runtime
* leaked out of the system.
*/
BUG_ON(want);
balanced:
/*
* Disable all the borrow logic by pretending we have inf
* runtime - in which case borrowing doesn't make sense.
*/
rt_rq->rt_runtime = RUNTIME_INF;
raw_spin_unlock(&rt_rq->rt_runtime_lock);
raw_spin_unlock(&rt_b->rt_runtime_lock);
}
}
static void disable_runtime(struct rq *rq)
{
unsigned long flags;
raw_spin_lock_irqsave(&rq->lock, flags);
__disable_runtime(rq);
raw_spin_unlock_irqrestore(&rq->lock, flags);
}
static void __enable_runtime(struct rq *rq)
{
rt_rq_iter_t iter;
struct rt_rq *rt_rq;
if (unlikely(!scheduler_running))
return;
/*
* Reset each runqueue's bandwidth settings
*/
for_each_rt_rq(rt_rq, iter, rq) {
struct rt_bandwidth *rt_b = sched_rt_bandwidth(rt_rq);
raw_spin_lock(&rt_b->rt_runtime_lock);
raw_spin_lock(&rt_rq->rt_runtime_lock);
rt_rq->rt_runtime = rt_b->rt_runtime;
rt_rq->rt_time = 0;
rt_rq->rt_throttled = 0;
raw_spin_unlock(&rt_rq->rt_runtime_lock);
raw_spin_unlock(&rt_b->rt_runtime_lock);
}
}
static void enable_runtime(struct rq *rq)
{
unsigned long flags;
raw_spin_lock_irqsave(&rq->lock, flags);
__enable_runtime(rq);
raw_spin_unlock_irqrestore(&rq->lock, flags);
}
int update_runtime(struct notifier_block *nfb, unsigned long action, void *hcpu)
{
int cpu = (int)(long)hcpu;
switch (action) {
case CPU_DOWN_PREPARE:
case CPU_DOWN_PREPARE_FROZEN:
disable_runtime(cpu_rq(cpu));
return NOTIFY_OK;
case CPU_DOWN_FAILED:
case CPU_DOWN_FAILED_FROZEN:
case CPU_ONLINE:
case CPU_ONLINE_FROZEN:
enable_runtime(cpu_rq(cpu));
return NOTIFY_OK;
default:
return NOTIFY_DONE;
}
}
static int balance_runtime(struct rt_rq *rt_rq)
{
int more = 0;
if (!sched_feat(RT_RUNTIME_SHARE))
return more;
if (rt_rq->rt_time > rt_rq->rt_runtime) {
raw_spin_unlock(&rt_rq->rt_runtime_lock);
more = do_balance_runtime(rt_rq);
raw_spin_lock(&rt_rq->rt_runtime_lock);
}
return more;
}
#else /* !CONFIG_SMP */
static inline int balance_runtime(struct rt_rq *rt_rq)
{
return 0;
}
#endif /* CONFIG_SMP */
static int do_sched_rt_period_timer(struct rt_bandwidth *rt_b, int overrun)
{
int i, idle = 1;
const struct cpumask *span;
if (!rt_bandwidth_enabled() || rt_b->rt_runtime == RUNTIME_INF)
return 1;
span = sched_rt_period_mask();
for_each_cpu(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);
raw_spin_lock(&rq->lock);
if (rt_rq->rt_time) {
u64 runtime;
raw_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;
/*
* Force a clock update if the CPU was idle,
* lest wakeup -> unthrottle time accumulate.
*/
if (rt_rq->rt_nr_running && rq->curr == rq->idle)
rq->skip_clock_update = -1;
}
if (rt_rq->rt_time || rt_rq->rt_nr_running)
idle = 0;
raw_spin_unlock(&rt_rq->rt_runtime_lock);
} else if (rt_rq->rt_nr_running) {
idle = 0;
if (!rt_rq_throttled(rt_rq))
enqueue = 1;
}
if (enqueue)
sched_rt_rq_enqueue(rt_rq);
raw_spin_unlock(&rq->lock);
}
return idle;
}
static inline int rt_se_prio(struct sched_rt_entity *rt_se)
{
#ifdef CONFIG_RT_GROUP_SCHED
struct rt_rq *rt_rq = group_rt_rq(rt_se);
if (rt_rq)
return rt_rq->highest_prio.curr;
#endif
return rt_task_of(rt_se)->prio;
}
static int sched_rt_runtime_exceeded(struct rt_rq *rt_rq)
{
u64 runtime = sched_rt_runtime(rt_rq);
if (rt_rq->rt_throttled)
return rt_rq_throttled(rt_rq);
if (runtime >= sched_rt_period(rt_rq))
return 0;
balance_runtime(rt_rq);
runtime = sched_rt_runtime(rt_rq);
if (runtime == RUNTIME_INF)
return 0;
if (rt_rq->rt_time > runtime) {
rt_rq->rt_throttled = 1;
printk_once(KERN_WARNING "sched: RT throttling activated\n");
if (rt_rq_throttled(rt_rq)) {
sched_rt_rq_dequeue(rt_rq);
return 1;
}
}
return 0;
}
/*
* Update the current task's runtime statistics. Skip current tasks that
* are not in our scheduling class.
*/
static void update_curr_rt(struct rq *rq)
{
struct task_struct *curr = rq->curr;
struct sched_rt_entity *rt_se = &curr->rt;
struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
u64 delta_exec;
if (curr->sched_class != &rt_sched_class)
return;
delta_exec = rq->clock_task - curr->se.exec_start;
if (unlikely((s64)delta_exec < 0))
delta_exec = 0;
schedstat_set(curr->se.statistics.exec_max, max(curr->se.statistics.exec_max, delta_exec));
curr->se.sum_exec_runtime += delta_exec;
timers: fix itimer/many thread hang Overview This patch reworks the handling of POSIX CPU timers, including the ITIMER_PROF, ITIMER_VIRT timers and rlimit handling. It was put together with the help of Roland McGrath, the owner and original writer of this code. The problem we ran into, and the reason for this rework, has to do with using a profiling timer in a process with a large number of threads. It appears that the performance of the old implementation of run_posix_cpu_timers() was at least O(n*3) (where "n" is the number of threads in a process) or worse. Everything is fine with an increasing number of threads until the time taken for that routine to run becomes the same as or greater than the tick time, at which point things degrade rather quickly. This patch fixes bug 9906, "Weird hang with NPTL and SIGPROF." Code Changes This rework corrects the implementation of run_posix_cpu_timers() to make it run in constant time for a particular machine. (Performance may vary between one machine and another depending upon whether the kernel is built as single- or multiprocessor and, in the latter case, depending upon the number of running processors.) To do this, at each tick we now update fields in signal_struct as well as task_struct. The run_posix_cpu_timers() function uses those fields to make its decisions. We define a new structure, "task_cputime," to contain user, system and scheduler times and use these in appropriate places: struct task_cputime { cputime_t utime; cputime_t stime; unsigned long long sum_exec_runtime; }; This is included in the structure "thread_group_cputime," which is a new substructure of signal_struct and which varies for uniprocessor versus multiprocessor kernels. For uniprocessor kernels, it uses "task_cputime" as a simple substructure, while for multiprocessor kernels it is a pointer: struct thread_group_cputime { struct task_cputime totals; }; struct thread_group_cputime { struct task_cputime *totals; }; We also add a new task_cputime substructure directly to signal_struct, to cache the earliest expiration of process-wide timers, and task_cputime also replaces the it_*_expires fields of task_struct (used for earliest expiration of thread timers). The "thread_group_cputime" structure contains process-wide timers that are updated via account_user_time() and friends. In the non-SMP case the structure is a simple aggregator; unfortunately in the SMP case that simplicity was not achievable due to cache-line contention between CPUs (in one measured case performance was actually _worse_ on a 16-cpu system than the same test on a 4-cpu system, due to this contention). For SMP, the thread_group_cputime counters are maintained as a per-cpu structure allocated using alloc_percpu(). The timer functions update only the timer field in the structure corresponding to the running CPU, obtained using per_cpu_ptr(). We define a set of inline functions in sched.h that we use to maintain the thread_group_cputime structure and hide the differences between UP and SMP implementations from the rest of the kernel. The thread_group_cputime_init() function initializes the thread_group_cputime structure for the given task. The thread_group_cputime_alloc() is a no-op for UP; for SMP it calls the out-of-line function thread_group_cputime_alloc_smp() to allocate and fill in the per-cpu structures and fields. The thread_group_cputime_free() function, also a no-op for UP, in SMP frees the per-cpu structures. The thread_group_cputime_clone_thread() function (also a UP no-op) for SMP calls thread_group_cputime_alloc() if the per-cpu structures haven't yet been allocated. The thread_group_cputime() function fills the task_cputime structure it is passed with the contents of the thread_group_cputime fields; in UP it's that simple but in SMP it must also safely check that tsk->signal is non-NULL (if it is it just uses the appropriate fields of task_struct) and, if so, sums the per-cpu values for each online CPU. Finally, the three functions account_group_user_time(), account_group_system_time() and account_group_exec_runtime() are used by timer functions to update the respective fields of the thread_group_cputime structure. Non-SMP operation is trivial and will not be mentioned further. The per-cpu structure is always allocated when a task creates its first new thread, via a call to thread_group_cputime_clone_thread() from copy_signal(). It is freed at process exit via a call to thread_group_cputime_free() from cleanup_signal(). All functions that formerly summed utime/stime/sum_sched_runtime values from from all threads in the thread group now use thread_group_cputime() to snapshot the values in the thread_group_cputime structure or the values in the task structure itself if the per-cpu structure hasn't been allocated. Finally, the code in kernel/posix-cpu-timers.c has changed quite a bit. The run_posix_cpu_timers() function has been split into a fast path and a slow path; the former safely checks whether there are any expired thread timers and, if not, just returns, while the slow path does the heavy lifting. With the dedicated thread group fields, timers are no longer "rebalanced" and the process_timer_rebalance() function and related code has gone away. All summing loops are gone and all code that used them now uses the thread_group_cputime() inline. When process-wide timers are set, the new task_cputime structure in signal_struct is used to cache the earliest expiration; this is checked in the fast path. Performance The fix appears not to add significant overhead to existing operations. It generally performs the same as the current code except in two cases, one in which it performs slightly worse (Case 5 below) and one in which it performs very significantly better (Case 2 below). Overall it's a wash except in those two cases. I've since done somewhat more involved testing on a dual-core Opteron system. Case 1: With no itimer running, for a test with 100,000 threads, the fixed kernel took 1428.5 seconds, 513 seconds more than the unfixed system, all of which was spent in the system. There were twice as many voluntary context switches with the fix as without it. Case 2: With an itimer running at .01 second ticks and 4000 threads (the most an unmodified kernel can handle), the fixed kernel ran the test in eight percent of the time (5.8 seconds as opposed to 70 seconds) and had better tick accuracy (.012 seconds per tick as opposed to .023 seconds per tick). Case 3: A 4000-thread test with an initial timer tick of .01 second and an interval of 10,000 seconds (i.e. a timer that ticks only once) had very nearly the same performance in both cases: 6.3 seconds elapsed for the fixed kernel versus 5.5 seconds for the unfixed kernel. With fewer threads (eight in these tests), the Case 1 test ran in essentially the same time on both the modified and unmodified kernels (5.2 seconds versus 5.8 seconds). The Case 2 test ran in about the same time as well, 5.9 seconds versus 5.4 seconds but again with much better tick accuracy, .013 seconds per tick versus .025 seconds per tick for the unmodified kernel. Since the fix affected the rlimit code, I also tested soft and hard CPU limits. Case 4: With a hard CPU limit of 20 seconds and eight threads (and an itimer running), the modified kernel was very slightly favored in that while it killed the process in 19.997 seconds of CPU time (5.002 seconds of wall time), only .003 seconds of that was system time, the rest was user time. The unmodified kernel killed the process in 20.001 seconds of CPU (5.014 seconds of wall time) of which .016 seconds was system time. Really, though, the results were too close to call. The results were essentially the same with no itimer running. Case 5: With a soft limit of 20 seconds and a hard limit of 2000 seconds (where the hard limit would never be reached) and an itimer running, the modified kernel exhibited worse tick accuracy than the unmodified kernel: .050 seconds/tick versus .028 seconds/tick. Otherwise, performance was almost indistinguishable. With no itimer running this test exhibited virtually identical behavior and times in both cases. In times past I did some limited performance testing. those results are below. On a four-cpu Opteron system without this fix, a sixteen-thread test executed in 3569.991 seconds, of which user was 3568.435s and system was 1.556s. On the same system with the fix, user and elapsed time were about the same, but system time dropped to 0.007 seconds. Performance with eight, four and one thread were comparable. Interestingly, the timer ticks with the fix seemed more accurate: The sixteen-thread test with the fix received 149543 ticks for 0.024 seconds per tick, while the same test without the fix received 58720 for 0.061 seconds per tick. Both cases were configured for an interval of 0.01 seconds. Again, the other tests were comparable. Each thread in this test computed the primes up to 25,000,000. I also did a test with a large number of threads, 100,000 threads, which is impossible without the fix. In this case each thread computed the primes only up to 10,000 (to make the runtime manageable). System time dominated, at 1546.968 seconds out of a total 2176.906 seconds (giving a user time of 629.938s). It received 147651 ticks for 0.015 seconds per tick, still quite accurate. There is obviously no comparable test without the fix. Signed-off-by: Frank Mayhar <fmayhar@google.com> Cc: Roland McGrath <roland@redhat.com> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-09-12 16:54:39 +00:00
account_group_exec_runtime(curr, delta_exec);
curr->se.exec_start = rq->clock_task;
sched: cpu accounting controller (V2) Commit cfb5285660aad4931b2ebbfa902ea48a37dfffa1 removed a useful feature for us, which provided a cpu accounting resource controller. This feature would be useful if someone wants to group tasks only for accounting purpose and doesnt really want to exercise any control over their cpu consumption. The patch below reintroduces the feature. It is based on Paul Menage's original patch (Commit 62d0df64065e7c135d0002f069444fbdfc64768f), with these differences: - Removed load average information. I felt it needs more thought (esp to deal with SMP and virtualized platforms) and can be added for 2.6.25 after more discussions. - Convert group cpu usage to be nanosecond accurate (as rest of the cfs stats are) and invoke cpuacct_charge() from the respective scheduler classes - Make accounting scalable on SMP systems by splitting the usage counter to be per-cpu - Move the code from kernel/cpu_acct.c to kernel/sched.c (since the code is not big enough to warrant a new file and also this rightly needs to live inside the scheduler. Also things like accessing rq->lock while reading cpu usage becomes easier if the code lived in kernel/sched.c) The patch also modifies the cpu controller not to provide the same accounting information. Tested-by: Balbir Singh <balbir@linux.vnet.ibm.com> Tested the patches on top of 2.6.24-rc3. The patches work fine. Ran some simple tests like cpuspin (spin on the cpu), ran several tasks in the same group and timed them. Compared their time stamps with cpuacct.usage. Signed-off-by: Srivatsa Vaddagiri <vatsa@linux.vnet.ibm.com> Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2007-12-02 19:04:49 +00:00
cpuacct_charge(curr, delta_exec);
sched_rt_avg_update(rq, delta_exec);
if (!rt_bandwidth_enabled())
return;
for_each_sched_rt_entity(rt_se) {
rt_rq = rt_rq_of_se(rt_se);
if (sched_rt_runtime(rt_rq) != RUNTIME_INF) {
raw_spin_lock(&rt_rq->rt_runtime_lock);
rt_rq->rt_time += delta_exec;
if (sched_rt_runtime_exceeded(rt_rq))
resched_task(curr);
raw_spin_unlock(&rt_rq->rt_runtime_lock);
}
}
}
#if defined CONFIG_SMP
static void
inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
{
struct rq *rq = rq_of_rt_rq(rt_rq);
sched: Use pushable_tasks to determine next highest prio Hillf Danton proposed a patch (see link) that cleaned up the sched_rt code that calculates the priority of the next highest priority task to be used in finding run queues to pull from. His patch removed the calculating of the next prio to just use the current prio when deteriming if we should examine a run queue to pull from. The problem with his patch was that it caused more false checks. Because we check a run queue for pushable tasks if the current priority of that run queue is higher in priority than the task about to run on our run queue. But after grabbing the locks and doing the real check, we find that there may not be a task that has a higher prio task to pull. Thus the locks were taken with nothing to do. I added some trace_printks() to record when and how many times the run queue locks were taken to check for pullable tasks, compared to how many times we pulled a task. With the current method, it was: 3806 locks taken vs 2812 pulled tasks With Hillf's patch: 6728 locks taken vs 2804 pulled tasks The number of times locks were taken to pull a task went up almost double with no more success rate. But his patch did get me thinking. When we look at the priority of the highest task to consider taking the locks to do a pull, a failure to pull can be one of the following: (in order of most likely) o RT task was pushed off already between the check and taking the lock o Waiting RT task can not be migrated o RT task's CPU affinity does not include the target run queue's CPU o RT task's priority changed between the check and taking the lock And with Hillf's patch, the thing that caused most of the failures, is the RT task to pull was not at the right priority to pull (not greater than the current RT task priority on the target run queue). Most of the above cases we can't help. But the current method does not check if the next highest prio RT task can be migrated or not, and if it can not, we still grab the locks to do the test (we don't find out about this fact until after we have the locks). I thought about this case, and realized that the pushable task plist that is maintained only holds RT tasks that can migrate. If we move the calculating of the next highest prio task from the inc/dec_rt_task() functions into the queuing of the pushable tasks, then we only measure the priorities of those tasks that we push, and we get this basically for free. Not only does this patch make the code a little more efficient, it cleans it up and makes it a little simpler. Thanks to Hillf Danton for inspiring me on this patch. Signed-off-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Hillf Danton <dhillf@gmail.com> Cc: Gregory Haskins <ghaskins@novell.com> Link: http://lkml.kernel.org/r/BANLkTimQ67180HxCx5vgMqumqw1EkFh3qg@mail.gmail.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-06-17 01:55:23 +00:00
if (rq->online && prio < prev_prio)
cpupri_set(&rq->rd->cpupri, rq->cpu, prio);
}
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
static void
dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
{
struct rq *rq = rq_of_rt_rq(rt_rq);
if (rq->online && rt_rq->highest_prio.curr != prev_prio)
cpupri_set(&rq->rd->cpupri, rq->cpu, rt_rq->highest_prio.curr);
}
#else /* CONFIG_SMP */
static inline
void inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) {}
static inline
void dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) {}
#endif /* CONFIG_SMP */
#if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
static void
inc_rt_prio(struct rt_rq *rt_rq, int prio)
{
int prev_prio = rt_rq->highest_prio.curr;
if (prio < prev_prio)
rt_rq->highest_prio.curr = prio;
inc_rt_prio_smp(rt_rq, prio, prev_prio);
}
static void
dec_rt_prio(struct rt_rq *rt_rq, int prio)
{
int prev_prio = rt_rq->highest_prio.curr;
if (rt_rq->rt_nr_running) {
WARN_ON(prio < prev_prio);
/*
* This may have been our highest task, and therefore
* we may have some recomputation to do
*/
if (prio == prev_prio) {
struct rt_prio_array *array = &rt_rq->active;
rt_rq->highest_prio.curr =
sched_find_first_bit(array->bitmap);
}
} else
rt_rq->highest_prio.curr = MAX_RT_PRIO;
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
dec_rt_prio_smp(rt_rq, prio, prev_prio);
}
#else
static inline void inc_rt_prio(struct rt_rq *rt_rq, int prio) {}
static inline void dec_rt_prio(struct rt_rq *rt_rq, int prio) {}
#endif /* CONFIG_SMP || CONFIG_RT_GROUP_SCHED */
#ifdef CONFIG_RT_GROUP_SCHED
static void
inc_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
{
if (rt_se_boosted(rt_se))
rt_rq->rt_nr_boosted++;
if (rt_rq->tg)
start_rt_bandwidth(&rt_rq->tg->rt_bandwidth);
}
static void
dec_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
{
if (rt_se_boosted(rt_se))
rt_rq->rt_nr_boosted--;
WARN_ON(!rt_rq->rt_nr_running && rt_rq->rt_nr_boosted);
}
#else /* CONFIG_RT_GROUP_SCHED */
static void
inc_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
{
start_rt_bandwidth(&def_rt_bandwidth);
}
static inline
void dec_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) {}
#endif /* CONFIG_RT_GROUP_SCHED */
static inline
void inc_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
{
int prio = rt_se_prio(rt_se);
WARN_ON(!rt_prio(prio));
rt_rq->rt_nr_running++;
inc_rt_prio(rt_rq, prio);
inc_rt_migration(rt_se, rt_rq);
inc_rt_group(rt_se, rt_rq);
}
static inline
void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
{
WARN_ON(!rt_prio(rt_se_prio(rt_se)));
WARN_ON(!rt_rq->rt_nr_running);
rt_rq->rt_nr_running--;
dec_rt_prio(rt_rq, rt_se_prio(rt_se));
dec_rt_migration(rt_se, rt_rq);
dec_rt_group(rt_se, rt_rq);
}
static void __enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head)
{
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 22:58:30 +00:00
struct list_head *queue = array->queue + rt_se_prio(rt_se);
/*
* 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))
return;
if (!rt_rq->rt_nr_running)
list_add_leaf_rt_rq(rt_rq);
if (head)
list_add(&rt_se->run_list, queue);
else
list_add_tail(&rt_se->run_list, queue);
__set_bit(rt_se_prio(rt_se), array->bitmap);
inc_rt_tasks(rt_se, rt_rq);
}
static void __dequeue_rt_entity(struct sched_rt_entity *rt_se)
{
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);
if (!rt_rq->rt_nr_running)
list_del_leaf_rt_rq(rt_rq);
}
/*
* Because the prio of an upper entry depends on the lower
* entries, we must remove entries top - down.
*/
static void dequeue_rt_stack(struct sched_rt_entity *rt_se)
{
struct sched_rt_entity *back = NULL;
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))
__dequeue_rt_entity(rt_se);
}
}
static void enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head)
{
dequeue_rt_stack(rt_se);
for_each_sched_rt_entity(rt_se)
__enqueue_rt_entity(rt_se, head);
}
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, false);
}
}
/*
* Adding/removing a task to/from a priority array:
*/
static void
enqueue_task_rt(struct rq *rq, struct task_struct *p, int flags)
{
struct sched_rt_entity *rt_se = &p->rt;
if (flags & ENQUEUE_WAKEUP)
rt_se->timeout = 0;
enqueue_rt_entity(rt_se, flags & ENQUEUE_HEAD);
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
if (!task_current(rq, p) && p->rt.nr_cpus_allowed > 1)
enqueue_pushable_task(rq, p);
inc_nr_running(rq);
}
static void dequeue_task_rt(struct rq *rq, struct task_struct *p, int flags)
{
struct sched_rt_entity *rt_se = &p->rt;
update_curr_rt(rq);
dequeue_rt_entity(rt_se);
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
dequeue_pushable_task(rq, p);
dec_nr_running(rq);
}
/*
* Put task to the head or the end of the run list without the overhead of
* dequeue followed by enqueue.
*/
static void
requeue_rt_entity(struct rt_rq *rt_rq, struct sched_rt_entity *rt_se, int head)
{
if (on_rt_rq(rt_se)) {
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);
}
}
static void requeue_task_rt(struct rq *rq, struct task_struct *p, int head)
{
struct sched_rt_entity *rt_se = &p->rt;
struct rt_rq *rt_rq;
for_each_sched_rt_entity(rt_se) {
rt_rq = rt_rq_of_se(rt_se);
requeue_rt_entity(rt_rq, rt_se, head);
}
}
static void yield_task_rt(struct rq *rq)
{
requeue_task_rt(rq, rq->curr, 0);
}
#ifdef CONFIG_SMP
static int find_lowest_rq(struct task_struct *task);
static int
select_task_rq_rt(struct task_struct *p, int sd_flag, int flags)
{
struct task_struct *curr;
struct rq *rq;
int cpu;
cpu = task_cpu(p);
sched: Use rt.nr_cpus_allowed to recover select_task_rq() cycles rt.nr_cpus_allowed is always available, use it to bail from select_task_rq() when only one cpu can be used, and saves some cycles for pinned tasks. See the line marked with '*' below: # taskset -c 3 pipe-test PerfTop: 997 irqs/sec kernel:89.5% exact: 0.0% [1000Hz cycles], (all, CPU: 3) ------------------------------------------------------------------------------------------------ Virgin Patched samples pcnt function samples pcnt function _______ _____ ___________________________ _______ _____ ___________________________ 2880.00 10.2% __schedule 3136.00 11.3% __schedule 1634.00 5.8% pipe_read 1615.00 5.8% pipe_read 1458.00 5.2% system_call 1534.00 5.5% system_call 1382.00 4.9% _raw_spin_lock_irqsave 1412.00 5.1% _raw_spin_lock_irqsave 1202.00 4.3% pipe_write 1255.00 4.5% copy_user_generic_string 1164.00 4.1% copy_user_generic_string 1241.00 4.5% __switch_to 1097.00 3.9% __switch_to 929.00 3.3% mutex_lock 872.00 3.1% mutex_lock 846.00 3.0% mutex_unlock 687.00 2.4% mutex_unlock 804.00 2.9% pipe_write 682.00 2.4% native_sched_clock 713.00 2.6% native_sched_clock 643.00 2.3% system_call_after_swapgs 653.00 2.3% _raw_spin_unlock_irqrestore 617.00 2.2% sched_clock_local 633.00 2.3% fsnotify 612.00 2.2% fsnotify 605.00 2.2% sched_clock_local 596.00 2.1% _raw_spin_unlock_irqrestore 593.00 2.1% system_call_after_swapgs 542.00 1.9% sysret_check 559.00 2.0% sysret_check 467.00 1.7% fget_light 472.00 1.7% fget_light 462.00 1.6% finish_task_switch 461.00 1.7% finish_task_switch 437.00 1.5% vfs_write 442.00 1.6% vfs_write 431.00 1.5% do_sync_write 428.00 1.5% do_sync_write * 413.00 1.5% select_task_rq_fair 404.00 1.5% _raw_spin_lock_irq 386.00 1.4% update_curr 402.00 1.4% update_curr 385.00 1.4% rw_verify_area 389.00 1.4% do_sync_read 377.00 1.3% _raw_spin_lock_irq 378.00 1.4% vfs_read 369.00 1.3% do_sync_read 340.00 1.2% pipe_iov_copy_from_user 360.00 1.3% vfs_read 316.00 1.1% __wake_up_sync_key 342.00 1.2% hrtick_start_fair 313.00 1.1% __wake_up_common Signed-off-by: Mike Galbraith <efault@gmx.de> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1321971504.6855.15.camel@marge.simson.net Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-11-22 14:18:24 +00:00
if (p->rt.nr_cpus_allowed == 1)
goto out;
/* For anything but wake ups, just return the task_cpu */
if (sd_flag != SD_BALANCE_WAKE && sd_flag != SD_BALANCE_FORK)
goto out;
rq = cpu_rq(cpu);
rcu_read_lock();
curr = ACCESS_ONCE(rq->curr); /* unlocked access */
/*
* If the current task on @p's runqueue 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.
*
sched: Try not to migrate higher priority RT tasks When first working on the RT scheduler design, we concentrated on keeping all CPUs running RT tasks instead of having multiple RT tasks on a single CPU waiting for the migration thread to move them. Instead we take a more proactive stance and push or pull RT tasks from one CPU to another on wakeup or scheduling. When an RT task wakes up on a CPU that is running another RT task, instead of preempting it and killing the cache of the running RT task, we look to see if we can migrate the RT task that is waking up, even if the RT task waking up is of higher priority. This may sound a bit odd, but RT tasks should be limited in migration by the user anyway. But in practice, people do not do this, which causes high prio RT tasks to bounce around the CPUs. This becomes even worse when we have priority inheritance, because a high prio task can block on a lower prio task and boost its priority. When the lower prio task wakes up the high prio task, if it happens to be on the same CPU it will migrate off of it. But in reality, the above does not happen much either, because the wake up of the lower prio task, which has already been boosted, if it was on the same CPU as the higher prio task, it would then migrate off of it. But anyway, we do not want to migrate them either. To examine the scheduling, I created a test program and examined it under kernelshark. The test program created CPU * 2 threads, where each thread had a different priority. The program takes different options. The options used in this change log was to have priority inheritance mutexes or not. All threads did the following loop: static void grab_lock(long id, int iter, int l) { ftrace_write("thread %ld iter %d, taking lock %d\n", id, iter, l); pthread_mutex_lock(&locks[l]); ftrace_write("thread %ld iter %d, took lock %d\n", id, iter, l); busy_loop(nr_tasks - id); ftrace_write("thread %ld iter %d, unlock lock %d\n", id, iter, l); pthread_mutex_unlock(&locks[l]); } void *start_task(void *id) { [...] while (!done) { for (l = 0; l < nr_locks; l++) { grab_lock(id, i, l); ftrace_write("thread %ld iter %d sleeping\n", id, i); ms_sleep(id); } i++; } [...] } The busy_loop(ms) keeps the CPU spinning for ms milliseconds. The ms_sleep(ms) sleeps for ms milliseconds. The ftrace_write() writes to the ftrace buffer to help analyze via ftrace. The higher the id, the higher the prio, the shorter it does the busy loop, but the longer it spins. This is usually the case with RT tasks, the lower priority tasks usually run longer than higher priority tasks. At the end of the test, it records the number of loops each thread took, as well as the number of voluntary preemptions, non-voluntary preemptions, and number of migrations each thread took, taking the information from /proc/$$/sched and /proc/$$/status. Running this on a 4 CPU processor, the results without changes to the kernel looked like this: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 53 3220 1470 98 1: 562 773 724 98 2: 752 933 1375 98 3: 749 39 697 98 4: 758 5 515 98 5: 764 2 679 99 6: 761 2 535 99 7: 757 3 346 99 total: 5156 4977 6341 787 Each thread regardless of priority migrated a few hundred times. The higher priority tasks, were a little better but still took quite an impact. By letting higher priority tasks bump the lower prio task from the CPU, things changed a bit: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 37 2835 1937 98 1: 666 1821 1865 98 2: 654 1003 1385 98 3: 664 635 973 99 4: 698 197 352 99 5: 703 101 159 99 6: 708 1 75 99 7: 713 1 2 99 total: 4843 6594 6748 789 The total # of migrations did not change (several runs showed the difference all within the noise). But we now see a dramatic improvement to the higher priority tasks. (kernelshark showed that the watchdog timer bumped the highest priority task to give it the 2 count. This was actually consistent with every run). Notice that the # of iterations did not change either. The above was with priority inheritance mutexes. That is, when the higher prority task blocked on a lower priority task, the lower priority task would inherit the higher priority task (which shows why task 6 was bumped so many times). When not using priority inheritance mutexes, the current kernel shows this: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 56 3101 1892 95 1: 594 713 937 95 2: 625 188 618 95 3: 628 4 491 96 4: 640 7 468 96 5: 631 2 501 96 6: 641 1 466 96 7: 643 2 497 96 total: 4458 4018 5870 765 Not much changed with or without priority inheritance mutexes. But if we let the high priority task bump lower priority tasks on wakeup we see: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 115 3439 2782 98 1: 633 1354 1583 99 2: 652 919 1218 99 3: 645 713 934 99 4: 690 3 3 99 5: 694 1 4 99 6: 720 3 4 99 7: 747 0 1 100 Which shows a even bigger change. The big difference between task 3 and task 4 is because we have only 4 CPUs on the machine, causing the 4 highest prio tasks to always have preference. Although I did not measure cache misses, and I'm sure there would be little to measure since the test was not data intensive, I could imagine large improvements for higher priority tasks when dealing with lower priority tasks. Thus, I'm satisfied with making the change and agreeing with what Gregory Haskins argued a few years ago when we first had this discussion. One final note. All tasks in the above tests were RT tasks. Any RT task will always preempt a non RT task that is running on the CPU the RT task wants to run on. Signed-off-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Gregory Haskins <ghaskins@novell.com> LKML-Reference: <20100921024138.605460343@goodmis.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-09-21 02:40:03 +00:00
* We want to avoid overloading runqueues. If the woken
* task is a higher priority, then it will stay on this CPU
* and the lower prio task should be moved to another CPU.
* Even though this will probably make the lower prio task
* lose its cache, we do not want to bounce a higher task
* around just because it gave up its CPU, perhaps for a
* lock?
*
* For equal prio tasks, we just let the scheduler sort it out.
*
* Otherwise, just let it ride on the affined RQ and the
* post-schedule router will push the preempted task away
*
* This test is optimistic, if we get it wrong the load-balancer
* will have to sort it out.
*/
if (curr && unlikely(rt_task(curr)) &&
(curr->rt.nr_cpus_allowed < 2 ||
curr->prio <= p->prio) &&
(p->rt.nr_cpus_allowed > 1)) {
int target = find_lowest_rq(p);
if (target != -1)
cpu = target;
}
rcu_read_unlock();
out:
return cpu;
}
static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p)
{
if (rq->curr->rt.nr_cpus_allowed == 1)
return;
if (p->rt.nr_cpus_allowed != 1
&& cpupri_find(&rq->rd->cpupri, p, NULL))
return;
if (!cpupri_find(&rq->rd->cpupri, rq->curr, NULL))
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);
}
#endif /* CONFIG_SMP */
/*
* Preempt the current task with a newly woken task if needed:
*/
static void check_preempt_curr_rt(struct rq *rq, struct task_struct *p, int flags)
{
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 19:20:41 +00:00
if (p->prio < rq->curr->prio) {
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 19:20:41 +00:00
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.
*/
if (p->prio == rq->curr->prio && !test_tsk_need_resched(rq->curr))
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 19:20:41 +00:00
#endif
}
static struct sched_rt_entity *pick_next_rt_entity(struct rq *rq,
struct rt_rq *rt_rq)
{
struct rt_prio_array *array = &rt_rq->active;
struct sched_rt_entity *next = NULL;
struct list_head *queue;
int idx;
idx = sched_find_first_bit(array->bitmap);
BUG_ON(idx >= MAX_RT_PRIO);
queue = array->queue + idx;
next = list_entry(queue->next, struct sched_rt_entity, run_list);
return next;
}
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
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;
rt_rq = &rq->rt;
if (!rt_rq->rt_nr_running)
return NULL;
if (rt_rq_throttled(rt_rq))
return NULL;
do {
rt_se = pick_next_rt_entity(rq, rt_rq);
BUG_ON(!rt_se);
rt_rq = group_rt_rq(rt_se);
} while (rt_rq);
p = rt_task_of(rt_se);
p->se.exec_start = rq->clock_task;
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
return p;
}
static struct task_struct *pick_next_task_rt(struct rq *rq)
{
struct task_struct *p = _pick_next_task_rt(rq);
/* The running task is never eligible for pushing */
if (p)
dequeue_pushable_task(rq, p);
#ifdef CONFIG_SMP
/*
* We detect this state here so that we can avoid taking the RQ
* lock again later if there is no need to push
*/
rq->post_schedule = has_pushable_tasks(rq);
#endif
return p;
}
static void put_prev_task_rt(struct rq *rq, struct task_struct *p)
{
update_curr_rt(rq);
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
/*
* The previous task needs to be made eligible for pushing
* if it is still active
*/
if (on_rt_rq(&p->rt) && p->rt.nr_cpus_allowed > 1)
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
enqueue_pushable_task(rq, p);
}
#ifdef CONFIG_SMP
/* Only try algorithms three times */
#define RT_MAX_TRIES 3
static int pick_rt_task(struct rq *rq, struct task_struct *p, int cpu)
{
if (!task_running(rq, p) &&
(cpu < 0 || cpumask_test_cpu(cpu, tsk_cpus_allowed(p))) &&
(p->rt.nr_cpus_allowed > 1))
return 1;
return 0;
}
/* Return the second highest RT task, NULL otherwise */
static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu)
{
struct task_struct *next = NULL;
struct sched_rt_entity *rt_se;
struct rt_prio_array *array;
struct rt_rq *rt_rq;
int idx;
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;
if (!rt_entity_is_task(rt_se))
continue;
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;
}
}
return next;
}
static DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask);
static int find_lowest_rq(struct task_struct *task)
{
struct sched_domain *sd;
struct cpumask *lowest_mask = __get_cpu_var(local_cpu_mask);
int this_cpu = smp_processor_id();
int cpu = task_cpu(task);
/* Make sure the mask is initialized first */
if (unlikely(!lowest_mask))
return -1;
if (task->rt.nr_cpus_allowed == 1)
return -1; /* No other targets possible */
if (!cpupri_find(&task_rq(task)->rd->cpupri, task, lowest_mask))
return -1; /* No targets found */
/*
* 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 (cpumask_test_cpu(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 (!cpumask_test_cpu(this_cpu, lowest_mask))
this_cpu = -1; /* Skip this_cpu opt if not among lowest */
rcu_read_lock();
for_each_domain(cpu, sd) {
if (sd->flags & SD_WAKE_AFFINE) {
int best_cpu;
/*
* "this_cpu" is cheaper to preempt than a
* remote processor.
*/
if (this_cpu != -1 &&
cpumask_test_cpu(this_cpu, sched_domain_span(sd))) {
rcu_read_unlock();
return this_cpu;
}
best_cpu = cpumask_first_and(lowest_mask,
sched_domain_span(sd));
if (best_cpu < nr_cpu_ids) {
rcu_read_unlock();
return best_cpu;
}
}
}
rcu_read_unlock();
/*
* And finally, if there were no matches within the domains
* just give the caller *something* to work with from the compatible
* locations.
*/
if (this_cpu != -1)
return this_cpu;
cpu = cpumask_any(lowest_mask);
if (cpu < nr_cpu_ids)
return cpu;
return -1;
}
/* Will lock the rq it finds */
static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
{
struct rq *lowest_rq = NULL;
int tries;
int cpu;
for (tries = 0; tries < RT_MAX_TRIES; tries++) {
cpu = find_lowest_rq(task);
if ((cpu == -1) || (cpu == rq->cpu))
break;
lowest_rq = cpu_rq(cpu);
/* if the prio of this runqueue changed, try again */
if (double_lock_balance(rq, lowest_rq)) {
/*
* 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.
*/
if (unlikely(task_rq(task) != rq ||
!cpumask_test_cpu(lowest_rq->cpu,
tsk_cpus_allowed(task)) ||
task_running(rq, task) ||
!task->on_rq)) {
raw_spin_unlock(&lowest_rq->lock);
lowest_rq = NULL;
break;
}
}
/* If this rq is still suitable use it. */
if (lowest_rq->rt.highest_prio.curr > task->prio)
break;
/* try again */
double_unlock_balance(rq, lowest_rq);
lowest_rq = NULL;
}
return lowest_rq;
}
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
static struct task_struct *pick_next_pushable_task(struct rq *rq)
{
struct task_struct *p;
if (!has_pushable_tasks(rq))
return NULL;
p = plist_first_entry(&rq->rt.pushable_tasks,
struct task_struct, pushable_tasks);
BUG_ON(rq->cpu != task_cpu(p));
BUG_ON(task_current(rq, p));
BUG_ON(p->rt.nr_cpus_allowed <= 1);
BUG_ON(!p->on_rq);
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
BUG_ON(!rt_task(p));
return p;
}
/*
* 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.
*/
static int push_rt_task(struct rq *rq)
{
struct task_struct *next_task;
struct rq *lowest_rq;
int ret = 0;
if (!rq->rt.overloaded)
return 0;
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
next_task = pick_next_pushable_task(rq);
if (!next_task)
return 0;
sched/rt: Fix task stack corruption under __ARCH_WANT_INTERRUPTS_ON_CTXSW This issue happens under the following conditions: 1. preemption is off 2. __ARCH_WANT_INTERRUPTS_ON_CTXSW is defined 3. RT scheduling class 4. SMP system Sequence is as follows: 1.suppose current task is A. start schedule() 2.task A is enqueued pushable task at the entry of schedule() __schedule prev = rq->curr; ... put_prev_task put_prev_task_rt enqueue_pushable_task 4.pick the task B as next task. next = pick_next_task(rq); 3.rq->curr set to task B and context_switch is started. rq->curr = next; 4.At the entry of context_swtich, release this cpu's rq->lock. context_switch prepare_task_switch prepare_lock_switch raw_spin_unlock_irq(&rq->lock); 5.Shortly after rq->lock is released, interrupt is occurred and start IRQ context 6.try_to_wake_up() which called by ISR acquires rq->lock try_to_wake_up ttwu_remote rq = __task_rq_lock(p) ttwu_do_wakeup(rq, p, wake_flags); task_woken_rt 7.push_rt_task picks the task A which is enqueued before. task_woken_rt push_rt_tasks(rq) next_task = pick_next_pushable_task(rq) 8.At find_lock_lowest_rq(), If double_lock_balance() returns 0, lowest_rq can be the remote rq. (But,If preemption is on, double_lock_balance always return 1 and it does't happen.) push_rt_task find_lock_lowest_rq if (double_lock_balance(rq, lowest_rq)).. 9.find_lock_lowest_rq return the available rq. task A is migrated to the remote cpu/rq. push_rt_task ... deactivate_task(rq, next_task, 0); set_task_cpu(next_task, lowest_rq->cpu); activate_task(lowest_rq, next_task, 0); 10. But, task A is on irq context at this cpu. So, task A is scheduled by two cpus at the same time until restore from IRQ. Task A's stack is corrupted. To fix it, don't migrate an RT task if it's still running. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Steven Rostedt <rostedt@goodmis.org> Cc: <stable@kernel.org> Link: http://lkml.kernel.org/r/CAOAMb1BHA=5fm7KTewYyke6u-8DP0iUuJMpgQw54vNeXFsGpoQ@mail.gmail.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
2012-01-05 11:00:19 +00:00
#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
if (unlikely(task_running(rq, next_task)))
return 0;
#endif
retry:
if (unlikely(next_task == rq->curr)) {
WARN_ON(1);
return 0;
}
/*
* It's possible that the next_task slipped in of
* higher priority than current. If that's the case
* just reschedule current.
*/
if (unlikely(next_task->prio < rq->curr->prio)) {
resched_task(rq->curr);
return 0;
}
/* We might release rq lock */
get_task_struct(next_task);
/* find_lock_lowest_rq locks the rq if found */
lowest_rq = find_lock_lowest_rq(next_task, rq);
if (!lowest_rq) {
struct task_struct *task;
/*
* find_lock_lowest_rq releases rq->lock
RT: fix push_rt_task() to handle dequeue_pushable properly A panic was discovered by Chirag Jog where a BUG_ON sanity check in the new "pushable_task" logic would trigger a panic under certain circumstances: http://lkml.org/lkml/2008/9/25/189 Gilles Carry discovered that the root cause was attributed to the pushable_tasks list getting corrupted in the push_rt_task logic. This was the result of a dropped rq lock in double_lock_balance allowing a task in the process of being pushed to potentially migrate away, and thus corrupt the pushable_tasks() list. I traced back the problem as introduced by the pushable_tasks patch that went in recently. There is a "retry" path in push_rt_task() that actually had a compound conditional to decide whether to retry or exit. I missed the meaning behind the rationale for the virtual "if(!task) goto out;" portion of the compound statement and thus did not handle it properly. The new pushable_tasks logic actually creates three distinct conditions: 1) an untouched and unpushable task should be dequeued 2) a migrated task where more pushable tasks remain should be retried 3) a migrated task where no more pushable tasks exist should exit The original logic mushed (1) and (3) together, resulting in the system dequeuing a migrated task (against an unlocked foreign run-queue nonetheless). To fix this, we get rid of the notion of "paranoid" and we support the three unique conditions properly. The paranoid feature is no longer relevant with the new pushable logic (since pushable naturally limits the loop) anyway, so lets just remove it. Reported-By: Chirag Jog <chirag@linux.vnet.ibm.com> Found-by: Gilles Carry <gilles.carry@bull.net> Signed-off-by: Gregory Haskins <ghaskins@novell.com>
2008-12-29 14:39:53 +00:00
* so it is possible that next_task has migrated.
*
* We need to make sure that the task is still on the same
* run-queue and is also still the next task eligible for
* pushing.
*/
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
task = pick_next_pushable_task(rq);
RT: fix push_rt_task() to handle dequeue_pushable properly A panic was discovered by Chirag Jog where a BUG_ON sanity check in the new "pushable_task" logic would trigger a panic under certain circumstances: http://lkml.org/lkml/2008/9/25/189 Gilles Carry discovered that the root cause was attributed to the pushable_tasks list getting corrupted in the push_rt_task logic. This was the result of a dropped rq lock in double_lock_balance allowing a task in the process of being pushed to potentially migrate away, and thus corrupt the pushable_tasks() list. I traced back the problem as introduced by the pushable_tasks patch that went in recently. There is a "retry" path in push_rt_task() that actually had a compound conditional to decide whether to retry or exit. I missed the meaning behind the rationale for the virtual "if(!task) goto out;" portion of the compound statement and thus did not handle it properly. The new pushable_tasks logic actually creates three distinct conditions: 1) an untouched and unpushable task should be dequeued 2) a migrated task where more pushable tasks remain should be retried 3) a migrated task where no more pushable tasks exist should exit The original logic mushed (1) and (3) together, resulting in the system dequeuing a migrated task (against an unlocked foreign run-queue nonetheless). To fix this, we get rid of the notion of "paranoid" and we support the three unique conditions properly. The paranoid feature is no longer relevant with the new pushable logic (since pushable naturally limits the loop) anyway, so lets just remove it. Reported-By: Chirag Jog <chirag@linux.vnet.ibm.com> Found-by: Gilles Carry <gilles.carry@bull.net> Signed-off-by: Gregory Haskins <ghaskins@novell.com>
2008-12-29 14:39:53 +00:00
if (task_cpu(next_task) == rq->cpu && task == next_task) {
/*
* The task hasn't migrated, and is still the next
* eligible task, but we failed to find a run-queue
* to push it to. Do not retry in this case, since
* other cpus will pull from us when ready.
RT: fix push_rt_task() to handle dequeue_pushable properly A panic was discovered by Chirag Jog where a BUG_ON sanity check in the new "pushable_task" logic would trigger a panic under certain circumstances: http://lkml.org/lkml/2008/9/25/189 Gilles Carry discovered that the root cause was attributed to the pushable_tasks list getting corrupted in the push_rt_task logic. This was the result of a dropped rq lock in double_lock_balance allowing a task in the process of being pushed to potentially migrate away, and thus corrupt the pushable_tasks() list. I traced back the problem as introduced by the pushable_tasks patch that went in recently. There is a "retry" path in push_rt_task() that actually had a compound conditional to decide whether to retry or exit. I missed the meaning behind the rationale for the virtual "if(!task) goto out;" portion of the compound statement and thus did not handle it properly. The new pushable_tasks logic actually creates three distinct conditions: 1) an untouched and unpushable task should be dequeued 2) a migrated task where more pushable tasks remain should be retried 3) a migrated task where no more pushable tasks exist should exit The original logic mushed (1) and (3) together, resulting in the system dequeuing a migrated task (against an unlocked foreign run-queue nonetheless). To fix this, we get rid of the notion of "paranoid" and we support the three unique conditions properly. The paranoid feature is no longer relevant with the new pushable logic (since pushable naturally limits the loop) anyway, so lets just remove it. Reported-By: Chirag Jog <chirag@linux.vnet.ibm.com> Found-by: Gilles Carry <gilles.carry@bull.net> Signed-off-by: Gregory Haskins <ghaskins@novell.com>
2008-12-29 14:39:53 +00:00
*/
goto out;
}
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
RT: fix push_rt_task() to handle dequeue_pushable properly A panic was discovered by Chirag Jog where a BUG_ON sanity check in the new "pushable_task" logic would trigger a panic under certain circumstances: http://lkml.org/lkml/2008/9/25/189 Gilles Carry discovered that the root cause was attributed to the pushable_tasks list getting corrupted in the push_rt_task logic. This was the result of a dropped rq lock in double_lock_balance allowing a task in the process of being pushed to potentially migrate away, and thus corrupt the pushable_tasks() list. I traced back the problem as introduced by the pushable_tasks patch that went in recently. There is a "retry" path in push_rt_task() that actually had a compound conditional to decide whether to retry or exit. I missed the meaning behind the rationale for the virtual "if(!task) goto out;" portion of the compound statement and thus did not handle it properly. The new pushable_tasks logic actually creates three distinct conditions: 1) an untouched and unpushable task should be dequeued 2) a migrated task where more pushable tasks remain should be retried 3) a migrated task where no more pushable tasks exist should exit The original logic mushed (1) and (3) together, resulting in the system dequeuing a migrated task (against an unlocked foreign run-queue nonetheless). To fix this, we get rid of the notion of "paranoid" and we support the three unique conditions properly. The paranoid feature is no longer relevant with the new pushable logic (since pushable naturally limits the loop) anyway, so lets just remove it. Reported-By: Chirag Jog <chirag@linux.vnet.ibm.com> Found-by: Gilles Carry <gilles.carry@bull.net> Signed-off-by: Gregory Haskins <ghaskins@novell.com>
2008-12-29 14:39:53 +00:00
if (!task)
/* No more tasks, just exit */
goto out;
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
/*
RT: fix push_rt_task() to handle dequeue_pushable properly A panic was discovered by Chirag Jog where a BUG_ON sanity check in the new "pushable_task" logic would trigger a panic under certain circumstances: http://lkml.org/lkml/2008/9/25/189 Gilles Carry discovered that the root cause was attributed to the pushable_tasks list getting corrupted in the push_rt_task logic. This was the result of a dropped rq lock in double_lock_balance allowing a task in the process of being pushed to potentially migrate away, and thus corrupt the pushable_tasks() list. I traced back the problem as introduced by the pushable_tasks patch that went in recently. There is a "retry" path in push_rt_task() that actually had a compound conditional to decide whether to retry or exit. I missed the meaning behind the rationale for the virtual "if(!task) goto out;" portion of the compound statement and thus did not handle it properly. The new pushable_tasks logic actually creates three distinct conditions: 1) an untouched and unpushable task should be dequeued 2) a migrated task where more pushable tasks remain should be retried 3) a migrated task where no more pushable tasks exist should exit The original logic mushed (1) and (3) together, resulting in the system dequeuing a migrated task (against an unlocked foreign run-queue nonetheless). To fix this, we get rid of the notion of "paranoid" and we support the three unique conditions properly. The paranoid feature is no longer relevant with the new pushable logic (since pushable naturally limits the loop) anyway, so lets just remove it. Reported-By: Chirag Jog <chirag@linux.vnet.ibm.com> Found-by: Gilles Carry <gilles.carry@bull.net> Signed-off-by: Gregory Haskins <ghaskins@novell.com>
2008-12-29 14:39:53 +00:00
* Something has shifted, try again.
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
*/
RT: fix push_rt_task() to handle dequeue_pushable properly A panic was discovered by Chirag Jog where a BUG_ON sanity check in the new "pushable_task" logic would trigger a panic under certain circumstances: http://lkml.org/lkml/2008/9/25/189 Gilles Carry discovered that the root cause was attributed to the pushable_tasks list getting corrupted in the push_rt_task logic. This was the result of a dropped rq lock in double_lock_balance allowing a task in the process of being pushed to potentially migrate away, and thus corrupt the pushable_tasks() list. I traced back the problem as introduced by the pushable_tasks patch that went in recently. There is a "retry" path in push_rt_task() that actually had a compound conditional to decide whether to retry or exit. I missed the meaning behind the rationale for the virtual "if(!task) goto out;" portion of the compound statement and thus did not handle it properly. The new pushable_tasks logic actually creates three distinct conditions: 1) an untouched and unpushable task should be dequeued 2) a migrated task where more pushable tasks remain should be retried 3) a migrated task where no more pushable tasks exist should exit The original logic mushed (1) and (3) together, resulting in the system dequeuing a migrated task (against an unlocked foreign run-queue nonetheless). To fix this, we get rid of the notion of "paranoid" and we support the three unique conditions properly. The paranoid feature is no longer relevant with the new pushable logic (since pushable naturally limits the loop) anyway, so lets just remove it. Reported-By: Chirag Jog <chirag@linux.vnet.ibm.com> Found-by: Gilles Carry <gilles.carry@bull.net> Signed-off-by: Gregory Haskins <ghaskins@novell.com>
2008-12-29 14:39:53 +00:00
put_task_struct(next_task);
next_task = task;
goto retry;
}
deactivate_task(rq, next_task, 0);
set_task_cpu(next_task, lowest_rq->cpu);
activate_task(lowest_rq, next_task, 0);
ret = 1;
resched_task(lowest_rq->curr);
double_unlock_balance(rq, lowest_rq);
out:
put_task_struct(next_task);
return ret;
}
static void push_rt_tasks(struct rq *rq)
{
/* push_rt_task will return true if it moved an RT */
while (push_rt_task(rq))
;
}
static int pull_rt_task(struct rq *this_rq)
{
int this_cpu = this_rq->cpu, ret = 0, cpu;
struct task_struct *p;
struct rq *src_rq;
if (likely(!rt_overloaded(this_rq)))
return 0;
for_each_cpu(cpu, this_rq->rd->rto_mask) {
if (this_cpu == cpu)
continue;
src_rq = cpu_rq(cpu);
/*
* Don't bother taking the src_rq->lock if the next highest
* task is known to be lower-priority than our current task.
* This may look racy, but if this value is about to go
* logically higher, the src_rq will push this task away.
* And if its going logically lower, we do not care
*/
if (src_rq->rt.highest_prio.next >=
this_rq->rt.highest_prio.curr)
continue;
/*
* We can potentially drop this_rq's lock in
* double_lock_balance, and another CPU could
* alter this_rq
*/
double_lock_balance(this_rq, src_rq);
/*
* Are there still pullable RT tasks?
*/
if (src_rq->rt.rt_nr_running <= 1)
goto skip;
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 && (p->prio < this_rq->rt.highest_prio.curr)) {
WARN_ON(p == src_rq->curr);
WARN_ON(!p->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
*/
if (p->prio < src_rq->curr->prio)
goto skip;
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 likelihood
* but possible)
*/
}
skip:
double_unlock_balance(this_rq, src_rq);
}
return ret;
}
static void pre_schedule_rt(struct rq *rq, struct task_struct *prev)
{
/* Try to pull RT tasks here if we lower this rq's prio */
if (rq->rt.highest_prio.curr > prev->prio)
pull_rt_task(rq);
}
static void post_schedule_rt(struct rq *rq)
{
push_rt_tasks(rq);
}
/*
* If we are not running and we are not going to reschedule soon, we should
* try to push tasks away now
*/
static void task_woken_rt(struct rq *rq, struct task_struct *p)
{
if (!task_running(rq, p) &&
!test_tsk_need_resched(rq->curr) &&
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
has_pushable_tasks(rq) &&
p->rt.nr_cpus_allowed > 1 &&
sched: Try not to migrate higher priority RT tasks When first working on the RT scheduler design, we concentrated on keeping all CPUs running RT tasks instead of having multiple RT tasks on a single CPU waiting for the migration thread to move them. Instead we take a more proactive stance and push or pull RT tasks from one CPU to another on wakeup or scheduling. When an RT task wakes up on a CPU that is running another RT task, instead of preempting it and killing the cache of the running RT task, we look to see if we can migrate the RT task that is waking up, even if the RT task waking up is of higher priority. This may sound a bit odd, but RT tasks should be limited in migration by the user anyway. But in practice, people do not do this, which causes high prio RT tasks to bounce around the CPUs. This becomes even worse when we have priority inheritance, because a high prio task can block on a lower prio task and boost its priority. When the lower prio task wakes up the high prio task, if it happens to be on the same CPU it will migrate off of it. But in reality, the above does not happen much either, because the wake up of the lower prio task, which has already been boosted, if it was on the same CPU as the higher prio task, it would then migrate off of it. But anyway, we do not want to migrate them either. To examine the scheduling, I created a test program and examined it under kernelshark. The test program created CPU * 2 threads, where each thread had a different priority. The program takes different options. The options used in this change log was to have priority inheritance mutexes or not. All threads did the following loop: static void grab_lock(long id, int iter, int l) { ftrace_write("thread %ld iter %d, taking lock %d\n", id, iter, l); pthread_mutex_lock(&locks[l]); ftrace_write("thread %ld iter %d, took lock %d\n", id, iter, l); busy_loop(nr_tasks - id); ftrace_write("thread %ld iter %d, unlock lock %d\n", id, iter, l); pthread_mutex_unlock(&locks[l]); } void *start_task(void *id) { [...] while (!done) { for (l = 0; l < nr_locks; l++) { grab_lock(id, i, l); ftrace_write("thread %ld iter %d sleeping\n", id, i); ms_sleep(id); } i++; } [...] } The busy_loop(ms) keeps the CPU spinning for ms milliseconds. The ms_sleep(ms) sleeps for ms milliseconds. The ftrace_write() writes to the ftrace buffer to help analyze via ftrace. The higher the id, the higher the prio, the shorter it does the busy loop, but the longer it spins. This is usually the case with RT tasks, the lower priority tasks usually run longer than higher priority tasks. At the end of the test, it records the number of loops each thread took, as well as the number of voluntary preemptions, non-voluntary preemptions, and number of migrations each thread took, taking the information from /proc/$$/sched and /proc/$$/status. Running this on a 4 CPU processor, the results without changes to the kernel looked like this: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 53 3220 1470 98 1: 562 773 724 98 2: 752 933 1375 98 3: 749 39 697 98 4: 758 5 515 98 5: 764 2 679 99 6: 761 2 535 99 7: 757 3 346 99 total: 5156 4977 6341 787 Each thread regardless of priority migrated a few hundred times. The higher priority tasks, were a little better but still took quite an impact. By letting higher priority tasks bump the lower prio task from the CPU, things changed a bit: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 37 2835 1937 98 1: 666 1821 1865 98 2: 654 1003 1385 98 3: 664 635 973 99 4: 698 197 352 99 5: 703 101 159 99 6: 708 1 75 99 7: 713 1 2 99 total: 4843 6594 6748 789 The total # of migrations did not change (several runs showed the difference all within the noise). But we now see a dramatic improvement to the higher priority tasks. (kernelshark showed that the watchdog timer bumped the highest priority task to give it the 2 count. This was actually consistent with every run). Notice that the # of iterations did not change either. The above was with priority inheritance mutexes. That is, when the higher prority task blocked on a lower priority task, the lower priority task would inherit the higher priority task (which shows why task 6 was bumped so many times). When not using priority inheritance mutexes, the current kernel shows this: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 56 3101 1892 95 1: 594 713 937 95 2: 625 188 618 95 3: 628 4 491 96 4: 640 7 468 96 5: 631 2 501 96 6: 641 1 466 96 7: 643 2 497 96 total: 4458 4018 5870 765 Not much changed with or without priority inheritance mutexes. But if we let the high priority task bump lower priority tasks on wakeup we see: Task vol nonvol migrated iterations ---- --- ------ -------- ---------- 0: 115 3439 2782 98 1: 633 1354 1583 99 2: 652 919 1218 99 3: 645 713 934 99 4: 690 3 3 99 5: 694 1 4 99 6: 720 3 4 99 7: 747 0 1 100 Which shows a even bigger change. The big difference between task 3 and task 4 is because we have only 4 CPUs on the machine, causing the 4 highest prio tasks to always have preference. Although I did not measure cache misses, and I'm sure there would be little to measure since the test was not data intensive, I could imagine large improvements for higher priority tasks when dealing with lower priority tasks. Thus, I'm satisfied with making the change and agreeing with what Gregory Haskins argued a few years ago when we first had this discussion. One final note. All tasks in the above tests were RT tasks. Any RT task will always preempt a non RT task that is running on the CPU the RT task wants to run on. Signed-off-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Gregory Haskins <ghaskins@novell.com> LKML-Reference: <20100921024138.605460343@goodmis.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-09-21 02:40:03 +00:00
rt_task(rq->curr) &&
(rq->curr->rt.nr_cpus_allowed < 2 ||
rq->curr->prio <= p->prio))
push_rt_tasks(rq);
}
static void set_cpus_allowed_rt(struct task_struct *p,
const struct cpumask *new_mask)
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
{
int weight = cpumask_weight(new_mask);
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
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.
*/
if (p->on_rq && (weight != p->rt.nr_cpus_allowed)) {
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
struct rq *rq = task_rq(p);
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
if (!task_current(rq, p)) {
/*
* Make sure we dequeue this task from the pushable list
* before going further. It will either remain off of
* the list because we are no longer pushable, or it
* will be requeued.
*/
if (p->rt.nr_cpus_allowed > 1)
dequeue_pushable_task(rq, p);
/*
* Requeue if our weight is changing and still > 1
*/
if (weight > 1)
enqueue_pushable_task(rq, p);
}
if ((p->rt.nr_cpus_allowed <= 1) && (weight > 1)) {
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
rq->rt.rt_nr_migratory++;
} else if ((p->rt.nr_cpus_allowed > 1) && (weight <= 1)) {
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
BUG_ON(!rq->rt.rt_nr_migratory);
rq->rt.rt_nr_migratory--;
}
update_rt_migration(&rq->rt);
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
}
}
/* Assumes rq->lock is held */
static void rq_online_rt(struct rq *rq)
{
if (rq->rt.overloaded)
rt_set_overload(rq);
__enable_runtime(rq);
cpupri_set(&rq->rd->cpupri, rq->cpu, rq->rt.highest_prio.curr);
}
/* Assumes rq->lock is held */
static void rq_offline_rt(struct rq *rq)
{
if (rq->rt.overloaded)
rt_clear_overload(rq);
__disable_runtime(rq);
cpupri_set(&rq->rd->cpupri, rq->cpu, CPUPRI_INVALID);
}
/*
* 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)
{
/*
* 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 (p->on_rq && !rq->rt.rt_nr_running)
pull_rt_task(rq);
}
void init_sched_rt_class(void)
{
unsigned int i;
for_each_possible_cpu(i) {
zalloc_cpumask_var_node(&per_cpu(local_cpu_mask, i),
GFP_KERNEL, cpu_to_node(i));
}
}
#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 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 (p->on_rq && rq->curr != p) {
#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)
{
if (!p->on_rq)
return;
if (rq->curr == p) {
#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
* 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.
*/
if (p->prio > rq->rt.highest_prio.curr && rq->curr == p)
resched_task(p);
#else
/* For UP simply resched on drop of prio */
if (oldprio < p->prio)
resched_task(p);
#endif /* CONFIG_SMP */
} 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);
}
}
static void watchdog(struct rq *rq, struct task_struct *p)
{
unsigned long soft, hard;
/* max may change after cur was read, this will be fixed next tick */
soft = task_rlimit(p, RLIMIT_RTTIME);
hard = task_rlimit_max(p, RLIMIT_RTTIME);
if (soft != RLIM_INFINITY) {
unsigned long next;
p->rt.timeout++;
next = DIV_ROUND_UP(min(soft, hard), USEC_PER_SEC/HZ);
if (p->rt.timeout > next)
timers: fix itimer/many thread hang Overview This patch reworks the handling of POSIX CPU timers, including the ITIMER_PROF, ITIMER_VIRT timers and rlimit handling. It was put together with the help of Roland McGrath, the owner and original writer of this code. The problem we ran into, and the reason for this rework, has to do with using a profiling timer in a process with a large number of threads. It appears that the performance of the old implementation of run_posix_cpu_timers() was at least O(n*3) (where "n" is the number of threads in a process) or worse. Everything is fine with an increasing number of threads until the time taken for that routine to run becomes the same as or greater than the tick time, at which point things degrade rather quickly. This patch fixes bug 9906, "Weird hang with NPTL and SIGPROF." Code Changes This rework corrects the implementation of run_posix_cpu_timers() to make it run in constant time for a particular machine. (Performance may vary between one machine and another depending upon whether the kernel is built as single- or multiprocessor and, in the latter case, depending upon the number of running processors.) To do this, at each tick we now update fields in signal_struct as well as task_struct. The run_posix_cpu_timers() function uses those fields to make its decisions. We define a new structure, "task_cputime," to contain user, system and scheduler times and use these in appropriate places: struct task_cputime { cputime_t utime; cputime_t stime; unsigned long long sum_exec_runtime; }; This is included in the structure "thread_group_cputime," which is a new substructure of signal_struct and which varies for uniprocessor versus multiprocessor kernels. For uniprocessor kernels, it uses "task_cputime" as a simple substructure, while for multiprocessor kernels it is a pointer: struct thread_group_cputime { struct task_cputime totals; }; struct thread_group_cputime { struct task_cputime *totals; }; We also add a new task_cputime substructure directly to signal_struct, to cache the earliest expiration of process-wide timers, and task_cputime also replaces the it_*_expires fields of task_struct (used for earliest expiration of thread timers). The "thread_group_cputime" structure contains process-wide timers that are updated via account_user_time() and friends. In the non-SMP case the structure is a simple aggregator; unfortunately in the SMP case that simplicity was not achievable due to cache-line contention between CPUs (in one measured case performance was actually _worse_ on a 16-cpu system than the same test on a 4-cpu system, due to this contention). For SMP, the thread_group_cputime counters are maintained as a per-cpu structure allocated using alloc_percpu(). The timer functions update only the timer field in the structure corresponding to the running CPU, obtained using per_cpu_ptr(). We define a set of inline functions in sched.h that we use to maintain the thread_group_cputime structure and hide the differences between UP and SMP implementations from the rest of the kernel. The thread_group_cputime_init() function initializes the thread_group_cputime structure for the given task. The thread_group_cputime_alloc() is a no-op for UP; for SMP it calls the out-of-line function thread_group_cputime_alloc_smp() to allocate and fill in the per-cpu structures and fields. The thread_group_cputime_free() function, also a no-op for UP, in SMP frees the per-cpu structures. The thread_group_cputime_clone_thread() function (also a UP no-op) for SMP calls thread_group_cputime_alloc() if the per-cpu structures haven't yet been allocated. The thread_group_cputime() function fills the task_cputime structure it is passed with the contents of the thread_group_cputime fields; in UP it's that simple but in SMP it must also safely check that tsk->signal is non-NULL (if it is it just uses the appropriate fields of task_struct) and, if so, sums the per-cpu values for each online CPU. Finally, the three functions account_group_user_time(), account_group_system_time() and account_group_exec_runtime() are used by timer functions to update the respective fields of the thread_group_cputime structure. Non-SMP operation is trivial and will not be mentioned further. The per-cpu structure is always allocated when a task creates its first new thread, via a call to thread_group_cputime_clone_thread() from copy_signal(). It is freed at process exit via a call to thread_group_cputime_free() from cleanup_signal(). All functions that formerly summed utime/stime/sum_sched_runtime values from from all threads in the thread group now use thread_group_cputime() to snapshot the values in the thread_group_cputime structure or the values in the task structure itself if the per-cpu structure hasn't been allocated. Finally, the code in kernel/posix-cpu-timers.c has changed quite a bit. The run_posix_cpu_timers() function has been split into a fast path and a slow path; the former safely checks whether there are any expired thread timers and, if not, just returns, while the slow path does the heavy lifting. With the dedicated thread group fields, timers are no longer "rebalanced" and the process_timer_rebalance() function and related code has gone away. All summing loops are gone and all code that used them now uses the thread_group_cputime() inline. When process-wide timers are set, the new task_cputime structure in signal_struct is used to cache the earliest expiration; this is checked in the fast path. Performance The fix appears not to add significant overhead to existing operations. It generally performs the same as the current code except in two cases, one in which it performs slightly worse (Case 5 below) and one in which it performs very significantly better (Case 2 below). Overall it's a wash except in those two cases. I've since done somewhat more involved testing on a dual-core Opteron system. Case 1: With no itimer running, for a test with 100,000 threads, the fixed kernel took 1428.5 seconds, 513 seconds more than the unfixed system, all of which was spent in the system. There were twice as many voluntary context switches with the fix as without it. Case 2: With an itimer running at .01 second ticks and 4000 threads (the most an unmodified kernel can handle), the fixed kernel ran the test in eight percent of the time (5.8 seconds as opposed to 70 seconds) and had better tick accuracy (.012 seconds per tick as opposed to .023 seconds per tick). Case 3: A 4000-thread test with an initial timer tick of .01 second and an interval of 10,000 seconds (i.e. a timer that ticks only once) had very nearly the same performance in both cases: 6.3 seconds elapsed for the fixed kernel versus 5.5 seconds for the unfixed kernel. With fewer threads (eight in these tests), the Case 1 test ran in essentially the same time on both the modified and unmodified kernels (5.2 seconds versus 5.8 seconds). The Case 2 test ran in about the same time as well, 5.9 seconds versus 5.4 seconds but again with much better tick accuracy, .013 seconds per tick versus .025 seconds per tick for the unmodified kernel. Since the fix affected the rlimit code, I also tested soft and hard CPU limits. Case 4: With a hard CPU limit of 20 seconds and eight threads (and an itimer running), the modified kernel was very slightly favored in that while it killed the process in 19.997 seconds of CPU time (5.002 seconds of wall time), only .003 seconds of that was system time, the rest was user time. The unmodified kernel killed the process in 20.001 seconds of CPU (5.014 seconds of wall time) of which .016 seconds was system time. Really, though, the results were too close to call. The results were essentially the same with no itimer running. Case 5: With a soft limit of 20 seconds and a hard limit of 2000 seconds (where the hard limit would never be reached) and an itimer running, the modified kernel exhibited worse tick accuracy than the unmodified kernel: .050 seconds/tick versus .028 seconds/tick. Otherwise, performance was almost indistinguishable. With no itimer running this test exhibited virtually identical behavior and times in both cases. In times past I did some limited performance testing. those results are below. On a four-cpu Opteron system without this fix, a sixteen-thread test executed in 3569.991 seconds, of which user was 3568.435s and system was 1.556s. On the same system with the fix, user and elapsed time were about the same, but system time dropped to 0.007 seconds. Performance with eight, four and one thread were comparable. Interestingly, the timer ticks with the fix seemed more accurate: The sixteen-thread test with the fix received 149543 ticks for 0.024 seconds per tick, while the same test without the fix received 58720 for 0.061 seconds per tick. Both cases were configured for an interval of 0.01 seconds. Again, the other tests were comparable. Each thread in this test computed the primes up to 25,000,000. I also did a test with a large number of threads, 100,000 threads, which is impossible without the fix. In this case each thread computed the primes only up to 10,000 (to make the runtime manageable). System time dominated, at 1546.968 seconds out of a total 2176.906 seconds (giving a user time of 629.938s). It received 147651 ticks for 0.015 seconds per tick, still quite accurate. There is obviously no comparable test without the fix. Signed-off-by: Frank Mayhar <fmayhar@google.com> Cc: Roland McGrath <roland@redhat.com> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-09-12 16:54:39 +00:00
p->cputime_expires.sched_exp = p->se.sum_exec_runtime;
}
}
static void task_tick_rt(struct rq *rq, struct task_struct *p, int queued)
{
update_curr_rt(rq);
watchdog(rq, p);
/*
* RR tasks need a special form of timeslice management.
* FIFO tasks have no timeslices.
*/
if (p->policy != SCHED_RR)
return;
if (--p->rt.time_slice)
return;
p->rt.time_slice = DEF_TIMESLICE;
/*
* Requeue to the end of queue if we are not the only element
* on the queue:
*/
if (p->rt.run_list.prev != p->rt.run_list.next) {
requeue_task_rt(rq, p, 0);
set_tsk_need_resched(p);
}
}
static void set_curr_task_rt(struct rq *rq)
{
struct task_struct *p = rq->curr;
p->se.exec_start = rq->clock_task;
sched: create "pushable_tasks" list to limit pushing to one attempt The RT scheduler employs a "push/pull" design to actively balance tasks within the system (on a per disjoint cpuset basis). When a task is awoken, it is immediately determined if there are any lower priority cpus which should be preempted. This is opposed to the way normal SCHED_OTHER tasks behave, which will wait for a periodic rebalancing operation to occur before spreading out load. When a particular RQ has more than 1 active RT task, it is said to be in an "overloaded" state. Once this occurs, the system enters the active balancing mode, where it will try to push the task away, or persuade a different cpu to pull it over. The system will stay in this state until the system falls back below the <= 1 queued RT task per RQ. However, the current implementation suffers from a limitation in the push logic. Once overloaded, all tasks (other than current) on the RQ are analyzed on every push operation, even if it was previously unpushable (due to affinity, etc). Whats more, the operation stops at the first task that is unpushable and will not look at items lower in the queue. This causes two problems: 1) We can have the same tasks analyzed over and over again during each push, which extends out the fast path in the scheduler for no gain. Consider a RQ that has dozens of tasks that are bound to a core. Each one of those tasks will be encountered and skipped for each push operation while they are queued. 2) There may be lower-priority tasks under the unpushable task that could have been successfully pushed, but will never be considered until either the unpushable task is cleared, or a pull operation succeeds. The net result is a potential latency source for mid priority tasks. This patch aims to rectify these two conditions by introducing a new priority sorted list: "pushable_tasks". A task is added to the list each time a task is activated or preempted. It is removed from the list any time it is deactivated, made current, or fails to push. This works because a task only needs to be attempted to push once. After an initial failure to push, the other cpus will eventually try to pull the task when the conditions are proper. This also solves the problem that we don't completely analyze all tasks due to encountering an unpushable tasks. Now every task will have a push attempted (when appropriate). This reduces latency both by shorting the critical section of the rq->lock for certain workloads, and by making sure the algorithm considers all eligible tasks in the system. [ rostedt: added a couple more BUG_ONs ] Signed-off-by: Gregory Haskins <ghaskins@novell.com> Acked-by: Steven Rostedt <srostedt@redhat.com>
2008-12-29 14:39:53 +00:00
/* The running task is never eligible for pushing */
dequeue_pushable_task(rq, p);
}
static unsigned int get_rr_interval_rt(struct rq *rq, struct task_struct *task)
{
/*
* Time slice is 0 for SCHED_FIFO tasks
*/
if (task->policy == SCHED_RR)
return DEF_TIMESLICE;
else
return 0;
}
const struct sched_class rt_sched_class = {
.next = &fair_sched_class,
.enqueue_task = enqueue_task_rt,
.dequeue_task = dequeue_task_rt,
.yield_task = yield_task_rt,
.check_preempt_curr = check_preempt_curr_rt,
.pick_next_task = pick_next_task_rt,
.put_prev_task = put_prev_task_rt,
#ifdef CONFIG_SMP
.select_task_rq = select_task_rq_rt,
sched: add RT-balance cpu-weight Some RT tasks (particularly kthreads) are bound to one specific CPU. It is fairly common for two or more bound tasks to get queued up at the same time. Consider, for instance, softirq_timer and softirq_sched. A timer goes off in an ISR which schedules softirq_thread to run at RT50. Then the timer handler determines that it's time to smp-rebalance the system so it schedules softirq_sched to run. So we are in a situation where we have two RT50 tasks queued, and the system will go into rt-overload condition to request other CPUs for help. This causes two problems in the current code: 1) If a high-priority bound task and a low-priority unbounded task queue up behind the running task, we will fail to ever relocate the unbounded task because we terminate the search on the first unmovable task. 2) We spend precious futile cycles in the fast-path trying to pull overloaded tasks over. It is therefore optimial to strive to avoid the overhead all together if we can cheaply detect the condition before overload even occurs. This patch tries to achieve this optimization by utilizing the hamming weight of the task->cpus_allowed mask. A weight of 1 indicates that the task cannot be migrated. We will then utilize this information to skip non-migratable tasks and to eliminate uncessary rebalance attempts. We introduce a per-rq variable to count the number of migratable tasks that are currently running. We only go into overload if we have more than one rt task, AND at least one of them is migratable. In addition, we introduce a per-task variable to cache the cpus_allowed weight, since the hamming calculation is probably relatively expensive. We only update the cached value when the mask is updated which should be relatively infrequent, especially compared to scheduling frequency in the fast path. Signed-off-by: Gregory Haskins <ghaskins@novell.com> Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-01-25 20:08:07 +00:00
.set_cpus_allowed = set_cpus_allowed_rt,
.rq_online = rq_online_rt,
.rq_offline = rq_offline_rt,
.pre_schedule = pre_schedule_rt,
.post_schedule = post_schedule_rt,
.task_woken = task_woken_rt,
.switched_from = switched_from_rt,
#endif
.set_curr_task = set_curr_task_rt,
.task_tick = task_tick_rt,
.get_rr_interval = get_rr_interval_rt,
.prio_changed = prio_changed_rt,
.switched_to = switched_to_rt,
};
#ifdef CONFIG_SCHED_DEBUG
extern void print_rt_rq(struct seq_file *m, int cpu, struct rt_rq *rt_rq);
void print_rt_stats(struct seq_file *m, int cpu)
{
rt_rq_iter_t iter;
struct rt_rq *rt_rq;
rcu_read_lock();
for_each_rt_rq(rt_rq, iter, cpu_rq(cpu))
print_rt_rq(m, cpu, rt_rq);
rcu_read_unlock();
}
#endif /* CONFIG_SCHED_DEBUG */