diff options
Diffstat (limited to '')
| -rw-r--r-- | kernel/sched/fair.c (renamed from kernel/sched_fair.c) | 1155 |
1 files changed, 890 insertions, 265 deletions
diff --git a/kernel/sched_fair.c b/kernel/sched/fair.c index 5c9e67923b7c..aca16b843b7e 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched/fair.c @@ -23,6 +23,13 @@ #include <linux/latencytop.h> #include <linux/sched.h> #include <linux/cpumask.h> +#include <linux/slab.h> +#include <linux/profile.h> +#include <linux/interrupt.h> + +#include <trace/events/sched.h> + +#include "sched.h" /* * Targeted preemption latency for CPU-bound tasks: @@ -103,7 +110,110 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL; unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL; #endif -static const struct sched_class fair_sched_class; +/* + * Increase the granularity value when there are more CPUs, + * because with more CPUs the 'effective latency' as visible + * to users decreases. But the relationship is not linear, + * so pick a second-best guess by going with the log2 of the + * number of CPUs. + * + * This idea comes from the SD scheduler of Con Kolivas: + */ +static int get_update_sysctl_factor(void) +{ + unsigned int cpus = min_t(int, num_online_cpus(), 8); + unsigned int factor; + + switch (sysctl_sched_tunable_scaling) { + case SCHED_TUNABLESCALING_NONE: + factor = 1; + break; + case SCHED_TUNABLESCALING_LINEAR: + factor = cpus; + break; + case SCHED_TUNABLESCALING_LOG: + default: + factor = 1 + ilog2(cpus); + break; + } + + return factor; +} + +static void update_sysctl(void) +{ + unsigned int factor = get_update_sysctl_factor(); + +#define SET_SYSCTL(name) \ + (sysctl_##name = (factor) * normalized_sysctl_##name) + SET_SYSCTL(sched_min_granularity); + SET_SYSCTL(sched_latency); + SET_SYSCTL(sched_wakeup_granularity); +#undef SET_SYSCTL +} + +void sched_init_granularity(void) +{ + update_sysctl(); +} + +#if BITS_PER_LONG == 32 +# define WMULT_CONST (~0UL) +#else +# define WMULT_CONST (1UL << 32) +#endif + +#define WMULT_SHIFT 32 + +/* + * Shift right and round: + */ +#define SRR(x, y) (((x) + (1UL << ((y) - 1))) >> (y)) + +/* + * delta *= weight / lw + */ +static unsigned long +calc_delta_mine(unsigned long delta_exec, unsigned long weight, + struct load_weight *lw) +{ + u64 tmp; + + /* + * weight can be less than 2^SCHED_LOAD_RESOLUTION for task group sched + * entities since MIN_SHARES = 2. Treat weight as 1 if less than + * 2^SCHED_LOAD_RESOLUTION. + */ + if (likely(weight > (1UL << SCHED_LOAD_RESOLUTION))) + tmp = (u64)delta_exec * scale_load_down(weight); + else + tmp = (u64)delta_exec; + + if (!lw->inv_weight) { + unsigned long w = scale_load_down(lw->weight); + + if (BITS_PER_LONG > 32 && unlikely(w >= WMULT_CONST)) + lw->inv_weight = 1; + else if (unlikely(!w)) + lw->inv_weight = WMULT_CONST; + else + lw->inv_weight = WMULT_CONST / w; + } + + /* + * Check whether we'd overflow the 64-bit multiplication: + */ + if (unlikely(tmp > WMULT_CONST)) + tmp = SRR(SRR(tmp, WMULT_SHIFT/2) * lw->inv_weight, + WMULT_SHIFT/2); + else + tmp = SRR(tmp * lw->inv_weight, WMULT_SHIFT); + + return (unsigned long)min(tmp, (u64)(unsigned long)LONG_MAX); +} + + +const struct sched_class fair_sched_class; /************************************************************** * CFS operations on generic schedulable entities: @@ -413,7 +523,7 @@ static void __dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se) rb_erase(&se->run_node, &cfs_rq->tasks_timeline); } -static struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq) +struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq) { struct rb_node *left = cfs_rq->rb_leftmost; @@ -434,7 +544,7 @@ static struct sched_entity *__pick_next_entity(struct sched_entity *se) } #ifdef CONFIG_SCHED_DEBUG -static struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq) +struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq) { struct rb_node *last = rb_last(&cfs_rq->tasks_timeline); @@ -684,7 +794,7 @@ account_entity_enqueue(struct cfs_rq *cfs_rq, struct sched_entity *se) { update_load_add(&cfs_rq->load, se->load.weight); if (!parent_entity(se)) - inc_cpu_load(rq_of(cfs_rq), se->load.weight); + update_load_add(&rq_of(cfs_rq)->load, se->load.weight); if (entity_is_task(se)) { add_cfs_task_weight(cfs_rq, se->load.weight); list_add(&se->group_node, &cfs_rq->tasks); @@ -697,7 +807,7 @@ account_entity_dequeue(struct cfs_rq *cfs_rq, struct sched_entity *se) { update_load_sub(&cfs_rq->load, se->load.weight); if (!parent_entity(se)) - dec_cpu_load(rq_of(cfs_rq), se->load.weight); + update_load_sub(&rq_of(cfs_rq)->load, se->load.weight); if (entity_is_task(se)) { add_cfs_task_weight(cfs_rq, -se->load.weight); list_del_init(&se->group_node); @@ -772,19 +882,32 @@ static void update_cfs_load(struct cfs_rq *cfs_rq, int global_update) list_del_leaf_cfs_rq(cfs_rq); } +static inline long calc_tg_weight(struct task_group *tg, struct cfs_rq *cfs_rq) +{ + long tg_weight; + + /* + * Use this CPU's actual weight instead of the last load_contribution + * to gain a more accurate current total weight. See + * update_cfs_rq_load_contribution(). + */ + tg_weight = atomic_read(&tg->load_weight); + tg_weight -= cfs_rq->load_contribution; + tg_weight += cfs_rq->load.weight; + + return tg_weight; +} + static long calc_cfs_shares(struct cfs_rq *cfs_rq, struct task_group *tg) { - long load_weight, load, shares; + long tg_weight, load, shares; + tg_weight = calc_tg_weight(tg, cfs_rq); load = cfs_rq->load.weight; - load_weight = atomic_read(&tg->load_weight); - load_weight += load; - load_weight -= cfs_rq->load_contribution; - shares = (tg->shares * load); - if (load_weight) - shares /= load_weight; + if (tg_weight) + shares /= tg_weight; if (shares < MIN_SHARES) shares = MIN_SHARES; @@ -907,6 +1030,8 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se) trace_sched_stat_iowait(tsk, delta); } + trace_sched_stat_blocked(tsk, delta); + /* * Blocking time is in units of nanosecs, so shift by * 20 to get a milliseconds-range estimation of the @@ -1274,6 +1399,32 @@ entity_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr, int queued) */ #ifdef CONFIG_CFS_BANDWIDTH + +#ifdef HAVE_JUMP_LABEL +static struct jump_label_key __cfs_bandwidth_used; + +static inline bool cfs_bandwidth_used(void) +{ + return static_branch(&__cfs_bandwidth_used); +} + +void account_cfs_bandwidth_used(int enabled, int was_enabled) +{ + /* only need to count groups transitioning between enabled/!enabled */ + if (enabled && !was_enabled) + jump_label_inc(&__cfs_bandwidth_used); + else if (!enabled && was_enabled) + jump_label_dec(&__cfs_bandwidth_used); +} +#else /* HAVE_JUMP_LABEL */ +static bool cfs_bandwidth_used(void) +{ + return true; +} + +void account_cfs_bandwidth_used(int enabled, int was_enabled) {} +#endif /* HAVE_JUMP_LABEL */ + /* * default period for cfs group bandwidth. * default: 0.1s, units: nanoseconds @@ -1295,7 +1446,7 @@ static inline u64 sched_cfs_bandwidth_slice(void) * * requires cfs_b->lock */ -static void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b) +void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b) { u64 now; @@ -1307,6 +1458,11 @@ static void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b) cfs_b->runtime_expires = now + ktime_to_ns(cfs_b->period); } +static inline struct cfs_bandwidth *tg_cfs_bandwidth(struct task_group *tg) +{ + return &tg->cfs_bandwidth; +} + /* returns 0 on failure to allocate runtime */ static int assign_cfs_rq_runtime(struct cfs_rq *cfs_rq) { @@ -1408,7 +1564,7 @@ static void __account_cfs_rq_runtime(struct cfs_rq *cfs_rq, static __always_inline void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, unsigned long delta_exec) { - if (!cfs_rq->runtime_enabled) + if (!cfs_bandwidth_used() || !cfs_rq->runtime_enabled) return; __account_cfs_rq_runtime(cfs_rq, delta_exec); @@ -1416,13 +1572,13 @@ static __always_inline void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, static inline int cfs_rq_throttled(struct cfs_rq *cfs_rq) { - return cfs_rq->throttled; + return cfs_bandwidth_used() && cfs_rq->throttled; } /* check whether cfs_rq, or any parent, is throttled */ static inline int throttled_hierarchy(struct cfs_rq *cfs_rq) { - return cfs_rq->throttle_count; + return cfs_bandwidth_used() && cfs_rq->throttle_count; } /* @@ -1517,7 +1673,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq) raw_spin_unlock(&cfs_b->lock); } -static void unthrottle_cfs_rq(struct cfs_rq *cfs_rq) +void unthrottle_cfs_rq(struct cfs_rq *cfs_rq) { struct rq *rq = rq_of(cfs_rq); struct cfs_bandwidth *cfs_b = tg_cfs_bandwidth(cfs_rq->tg); @@ -1743,7 +1899,10 @@ static void __return_cfs_rq_runtime(struct cfs_rq *cfs_rq) static __always_inline void return_cfs_rq_runtime(struct cfs_rq *cfs_rq) { - if (!cfs_rq->runtime_enabled || !cfs_rq->nr_running) + if (!cfs_bandwidth_used()) + return; + + if (!cfs_rq->runtime_enabled || cfs_rq->nr_running) return; __return_cfs_rq_runtime(cfs_rq); @@ -1788,6 +1947,9 @@ static void do_sched_cfs_slack_timer(struct cfs_bandwidth *cfs_b) */ static void check_enqueue_throttle(struct cfs_rq *cfs_rq) { + if (!cfs_bandwidth_used()) + return; + /* an active group must be handled by the update_curr()->put() path */ if (!cfs_rq->runtime_enabled || cfs_rq->curr) return; @@ -1805,6 +1967,9 @@ static void check_enqueue_throttle(struct cfs_rq *cfs_rq) /* conditionally throttle active cfs_rq's from put_prev_entity() */ static void check_cfs_rq_runtime(struct cfs_rq *cfs_rq) { + if (!cfs_bandwidth_used()) + return; + if (likely(!cfs_rq->runtime_enabled || cfs_rq->runtime_remaining > 0)) return; @@ -1817,7 +1982,112 @@ static void check_cfs_rq_runtime(struct cfs_rq *cfs_rq) throttle_cfs_rq(cfs_rq); } -#else + +static inline u64 default_cfs_period(void); +static int do_sched_cfs_period_timer(struct cfs_bandwidth *cfs_b, int overrun); +static void do_sched_cfs_slack_timer(struct cfs_bandwidth *cfs_b); + +static enum hrtimer_restart sched_cfs_slack_timer(struct hrtimer *timer) +{ + struct cfs_bandwidth *cfs_b = + container_of(timer, struct cfs_bandwidth, slack_timer); + do_sched_cfs_slack_timer(cfs_b); + + return HRTIMER_NORESTART; +} + +static enum hrtimer_restart sched_cfs_period_timer(struct hrtimer *timer) +{ + struct cfs_bandwidth *cfs_b = + container_of(timer, struct cfs_bandwidth, period_timer); + ktime_t now; + int overrun; + int idle = 0; + + for (;;) { + now = hrtimer_cb_get_time(timer); + overrun = hrtimer_forward(timer, now, cfs_b->period); + + if (!overrun) + break; + + idle = do_sched_cfs_period_timer(cfs_b, overrun); + } + + return idle ? HRTIMER_NORESTART : HRTIMER_RESTART; +} + +void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b) +{ + raw_spin_lock_init(&cfs_b->lock); + cfs_b->runtime = 0; + cfs_b->quota = RUNTIME_INF; + cfs_b->period = ns_to_ktime(default_cfs_period()); + + INIT_LIST_HEAD(&cfs_b->throttled_cfs_rq); + hrtimer_init(&cfs_b->period_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + cfs_b->period_timer.function = sched_cfs_period_timer; + hrtimer_init(&cfs_b->slack_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + cfs_b->slack_timer.function = sched_cfs_slack_timer; +} + +static void init_cfs_rq_runtime(struct cfs_rq *cfs_rq) +{ + cfs_rq->runtime_enabled = 0; + INIT_LIST_HEAD(&cfs_rq->throttled_list); +} + +/* requires cfs_b->lock, may release to reprogram timer */ +void __start_cfs_bandwidth(struct cfs_bandwidth *cfs_b) +{ + /* + * The timer may be active because we're trying to set a new bandwidth + * period or because we're racing with the tear-down path + * (timer_active==0 becomes visible before the hrtimer call-back + * terminates). In either case we ensure that it's re-programmed + */ + while (unlikely(hrtimer_active(&cfs_b->period_timer))) { + raw_spin_unlock(&cfs_b->lock); + /* ensure cfs_b->lock is available while we wait */ + hrtimer_cancel(&cfs_b->period_timer); + + raw_spin_lock(&cfs_b->lock); + /* if someone else restarted the timer then we're done */ + if (cfs_b->timer_active) + return; + } + + cfs_b->timer_active = 1; + start_bandwidth_timer(&cfs_b->period_timer, cfs_b->period); +} + +static void destroy_cfs_bandwidth(struct cfs_bandwidth *cfs_b) +{ + hrtimer_cancel(&cfs_b->period_timer); + hrtimer_cancel(&cfs_b->slack_timer); +} + +void unthrottle_offline_cfs_rqs(struct rq *rq) +{ + struct cfs_rq *cfs_rq; + + for_each_leaf_cfs_rq(rq, cfs_rq) { + struct cfs_bandwidth *cfs_b = tg_cfs_bandwidth(cfs_rq->tg); + + if (!cfs_rq->runtime_enabled) + continue; + + /* + * clock_task is not advancing so we just need to make sure + * there's some valid quota amount + */ + cfs_rq->runtime_remaining = cfs_b->quota; + if (cfs_rq_throttled(cfs_rq)) + unthrottle_cfs_rq(cfs_rq); + } +} + +#else /* CONFIG_CFS_BANDWIDTH */ static void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, unsigned long delta_exec) {} static void check_cfs_rq_runtime(struct cfs_rq *cfs_rq) {} @@ -1839,8 +2109,22 @@ static inline int throttled_lb_pair(struct task_group *tg, { return 0; } + +void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {} + +#ifdef CONFIG_FAIR_GROUP_SCHED +static void init_cfs_rq_runtime(struct cfs_rq *cfs_rq) {} #endif +static inline struct cfs_bandwidth *tg_cfs_bandwidth(struct task_group *tg) +{ + return NULL; +} +static inline void destroy_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {} +void unthrottle_offline_cfs_rqs(struct rq *rq) {} + +#endif /* CONFIG_CFS_BANDWIDTH */ + /************************************************** * CFS operations on tasks: */ @@ -1853,7 +2137,7 @@ static void hrtick_start_fair(struct rq *rq, struct task_struct *p) WARN_ON(task_rq(p) != rq); - if (hrtick_enabled(rq) && cfs_rq->nr_running > 1) { + if (cfs_rq->nr_running > 1) { u64 slice = sched_slice(cfs_rq, se); u64 ran = se->sum_exec_runtime - se->prev_sum_exec_runtime; s64 delta = slice - ran; @@ -1884,7 +2168,7 @@ static void hrtick_update(struct rq *rq) { struct task_struct *curr = rq->curr; - if (curr->sched_class != &fair_sched_class) + if (!hrtick_enabled(rq) || curr->sched_class != &fair_sched_class) return; if (cfs_rq_of(&curr->se)->nr_running < sched_nr_latency) @@ -2007,6 +2291,61 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags) } #ifdef CONFIG_SMP +/* Used instead of source_load when we know the type == 0 */ +static unsigned long weighted_cpuload(const int cpu) +{ + return cpu_rq(cpu)->load.weight; +} + +/* + * Return a low guess at the load of a migration-source cpu weighted + * according to the scheduling class and "nice" value. + * + * We want to under-estimate the load of migration sources, to + * balance conservatively. + */ +static unsigned long source_load(int cpu, int type) +{ + struct rq *rq = cpu_rq(cpu); + unsigned long total = weighted_cpuload(cpu); + + if (type == 0 || !sched_feat(LB_BIAS)) + return total; + + return min(rq->cpu_load[type-1], total); +} + +/* + * Return a high guess at the load of a migration-target cpu weighted + * according to the scheduling class and "nice" value. + */ +static unsigned long target_load(int cpu, int type) +{ + struct rq *rq = cpu_rq(cpu); + unsigned long total = weighted_cpuload(cpu); + + if (type == 0 || !sched_feat(LB_BIAS)) + return total; + + return max(rq->cpu_load[type-1], total); +} + +static unsigned long power_of(int cpu) +{ + return cpu_rq(cpu)->cpu_power; +} + +static unsigned long cpu_avg_load_per_task(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + unsigned long nr_running = ACCESS_ONCE(rq->nr_running); + + if (nr_running) + return rq->load.weight / nr_running; + + return 0; +} + static void task_waking_fair(struct task_struct *p) { @@ -2036,36 +2375,100 @@ static void task_waking_fair(struct task_struct *p) * Adding load to a group doesn't make a group heavier, but can cause movement * of group shares between cpus. Assuming the shares were perfectly aligned one * can calculate the shift in shares. + * + * Calculate the effective load difference if @wl is added (subtracted) to @tg + * on this @cpu and results in a total addition (subtraction) of @wg to the + * total group weight. + * + * Given a runqueue weight distribution (rw_i) we can compute a shares + * distribution (s_i) using: + * + * s_i = rw_i / \Sum rw_j (1) + * + * Suppose we have 4 CPUs and our @tg is a direct child of the root group and + * has 7 equal weight tasks, distributed as below (rw_i), with the resulting + * shares distribution (s_i): + * + * rw_i = { 2, 4, 1, 0 } + * s_i = { 2/7, 4/7, 1/7, 0 } + * + * As per wake_affine() we're interested in the load of two CPUs (the CPU the + * task used to run on and the CPU the waker is running on), we need to + * compute the effect of waking a task on either CPU and, in case of a sync + * wakeup, compute the effect of the current task going to sleep. + * + * So for a change of @wl to the local @cpu with an overall group weight change + * of @wl we can compute the new shares distribution (s'_i) using: + * + * s'_i = (rw_i + @wl) / (@wg + \Sum rw_j) (2) + * + * Suppose we're interested in CPUs 0 and 1, and want to compute the load + * differences in waking a task to CPU 0. The additional task changes the + * weight and shares distributions like: + * + * rw'_i = { 3, 4, 1, 0 } + * s'_i = { 3/8, 4/8, 1/8, 0 } + * + * We can then compute the difference in effective weight by using: + * + * dw_i = S * (s'_i - s_i) (3) + * + * Where 'S' is the group weight as seen by its parent. + * + * Therefore the effective change in loads on CPU 0 would be 5/56 (3/8 - 2/7) + * times the weight of the group. The effect on CPU 1 would be -4/56 (4/8 - + * 4/7) times the weight of the group. */ static long effective_load(struct task_group *tg, int cpu, long wl, long wg) { struct sched_entity *se = tg->se[cpu]; - if (!tg->parent) + if (!tg->parent) /* the trivial, non-cgroup case */ return wl; for_each_sched_entity(se) { - long lw, w; + long w, W; tg = se->my_q->tg; - w = se->my_q->load.weight; - /* use this cpu's instantaneous contribution */ - lw = atomic_read(&tg->load_weight); - lw -= se->my_q->load_contribution; - lw += w + wg; + /* + * W = @wg + \Sum rw_j + */ + W = wg + calc_tg_weight(tg, se->my_q); - wl += w; + /* + * w = rw_i + @wl + */ + w = se->my_q->load.weight + wl; - if (lw > 0 && wl < lw) - wl = (wl * tg->shares) / lw; + /* + * wl = S * s'_i; see (2) + */ + if (W > 0 && w < W) + wl = (w * tg->shares) / W; else wl = tg->shares; - /* zero point is MIN_SHARES */ + /* + * Per the above, wl is the new se->load.weight value; since + * those are clipped to [MIN_SHARES, ...) do so now. See + * calc_cfs_shares(). + */ if (wl < MIN_SHARES) wl = MIN_SHARES; + + /* + * wl = dw_i = S * (s'_i - s_i); see (3) + */ wl -= se->load.weight; + + /* + * Recursively apply this logic to all parent groups to compute + * the final effective load change on the root group. Since + * only the @tg group gets extra weight, all parent groups can + * only redistribute existing shares. @wl is the shift in shares + * resulting from this level per the above. + */ wg = 0; } @@ -2249,6 +2652,7 @@ static int select_idle_sibling(struct task_struct *p, int target) int cpu = smp_processor_id(); int prev_cpu = task_cpu(p); struct sched_domain *sd; + struct sched_group *sg; int i; /* @@ -2269,25 +2673,28 @@ static int select_idle_sibling(struct task_struct *p, int target) * Otherwise, iterate the domains and find an elegible idle cpu. */ rcu_read_lock(); - for_each_domain(target, sd) { - if (!(sd->flags & SD_SHARE_PKG_RESOURCES)) - break; - for_each_cpu_and(i, sched_domain_span(sd), tsk_cpus_allowed(p)) { - if (idle_cpu(i)) { - target = i; - break; + sd = rcu_dereference(per_cpu(sd_llc, target)); + for_each_lower_domain(sd) { + sg = sd->groups; + do { + if (!cpumask_intersects(sched_group_cpus(sg), + tsk_cpus_allowed(p))) + goto next; + + for_each_cpu(i, sched_group_cpus(sg)) { + if (!idle_cpu(i)) + goto next; } - } - /* - * Lets stop looking for an idle sibling when we reached - * the domain that spans the current cpu and prev_cpu. - */ - if (cpumask_test_cpu(cpu, sched_domain_span(sd)) && - cpumask_test_cpu(prev_cpu, sched_domain_span(sd))) - break; + target = cpumask_first_and(sched_group_cpus(sg), + tsk_cpus_allowed(p)); + goto done; +next: + sg = sg->next; + } while (sg != sd->groups); } +done: rcu_read_unlock(); return target; @@ -2315,6 +2722,9 @@ select_task_rq_fair(struct task_struct *p, int sd_flag, int wake_flags) int want_sd = 1; int sync = wake_flags & WF_SYNC; + if (p->rt.nr_cpus_allowed == 1) + return prev_cpu; + if (sd_flag & SD_BALANCE_WAKE) { if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p))) want_affine = 1; @@ -2599,7 +3009,8 @@ static struct task_struct *pick_next_task_fair(struct rq *rq) } while (cfs_rq); p = task_of(se); - hrtick_start_fair(rq, p); + if (hrtick_enabled(rq)) + hrtick_start_fair(rq, p); return p; } @@ -2643,6 +3054,12 @@ static void yield_task_fair(struct rq *rq) * Update run-time statistics of the 'current'. */ update_curr(cfs_rq); + /* + * Tell update_rq_clock() that we've just updated, + * so we don't do microscopic update in schedule() + * and double the fastpath cost. + */ + rq->skip_clock_update = 1; } set_skip_buddy(se); @@ -2683,12 +3100,50 @@ static void pull_task(struct rq *src_rq, struct task_struct *p, } /* + * Is this task likely cache-hot: + */ +static int +task_hot(struct task_struct *p, u64 now, struct sched_domain *sd) +{ + s64 delta; + + if (p->sched_class != &fair_sched_class) + return 0; + + if (unlikely(p->policy == SCHED_IDLE)) + return 0; + + /* + * Buddy candidates are cache hot: + */ + if (sched_feat(CACHE_HOT_BUDDY) && this_rq()->nr_running && + (&p->se == cfs_rq_of(&p->se)->next || + &p->se == cfs_rq_of(&p->se)->last)) + return 1; + + if (sysctl_sched_migration_cost == -1) + return 1; + if (sysctl_sched_migration_cost == 0) + return 0; + + delta = now - p->se.exec_start; + + return delta < (s64)sysctl_sched_migration_cost; +} + +#define LBF_ALL_PINNED 0x01 +#define LBF_NEED_BREAK 0x02 /* clears into HAD_BREAK */ +#define LBF_HAD_BREAK 0x04 +#define LBF_HAD_BREAKS 0x0C /* count HAD_BREAKs overflows into ABORT */ +#define LBF_ABORT 0x10 + +/* * can_migrate_task - may task p from runqueue rq be migrated to this_cpu? */ static int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned) + int *lb_flags) { int tsk_cache_hot = 0; /* @@ -2701,7 +3156,7 @@ int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, schedstat_inc(p, se.statistics.nr_failed_migrations_affine); return 0; } - *all_pinned = 0; + *lb_flags &= ~LBF_ALL_PINNED; if (task_running(rq, p)) { schedstat_inc(p, se.statistics.nr_failed_migrations_running); @@ -2775,7 +3230,7 @@ move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, static unsigned long balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, + enum cpu_idle_type idle, int *lb_flags, struct cfs_rq *busiest_cfs_rq) { int loops = 0, pulled = 0; @@ -2786,12 +3241,14 @@ balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, goto out; list_for_each_entry_safe(p, n, &busiest_cfs_rq->tasks, se.group_node) { - if (loops++ > sysctl_sched_nr_migrate) + if (loops++ > sysctl_sched_nr_migrate) { + *lb_flags |= LBF_NEED_BREAK; break; + } if ((p->se.load.weight >> 1) > rem_load_move || !can_migrate_task(p, busiest, this_cpu, sd, idle, - all_pinned)) + lb_flags)) continue; pull_task(busiest, p, this_rq, this_cpu); @@ -2804,8 +3261,10 @@ balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, * kernels will stop after the first task is pulled to minimize * the critical section. */ - if (idle == CPU_NEWLY_IDLE) + if (idle == CPU_NEWLY_IDLE) { + *lb_flags |= LBF_ABORT; break; + } #endif /* @@ -2910,7 +3369,7 @@ static unsigned long load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned) + int *lb_flags) { long rem_load_move = max_load_move; struct cfs_rq *busiest_cfs_rq; @@ -2923,6 +3382,9 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long busiest_weight = busiest_cfs_rq->load.weight; u64 rem_load, moved_load; + if (*lb_flags & (LBF_NEED_BREAK|LBF_ABORT)) + break; + /* * empty group or part of a throttled hierarchy */ @@ -2934,7 +3396,7 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, rem_load = div_u64(rem_load, busiest_h_load + 1); moved_load = balance_tasks(this_rq, this_cpu, busiest, - rem_load, sd, idle, all_pinned, + rem_load, sd, idle, lb_flags, busiest_cfs_rq); if (!moved_load) @@ -2960,10 +3422,10 @@ static unsigned long load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned) + int *lb_flags) { return balance_tasks(this_rq, this_cpu, busiest, - max_load_move, sd, idle, all_pinned, + max_load_move, sd, idle, lb_flags, &busiest->cfs); } #endif @@ -2978,29 +3440,30 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned) + int *lb_flags) { unsigned long total_load_moved = 0, load_moved; do { load_moved = load_balance_fair(this_rq, this_cpu, busiest, max_load_move - total_load_moved, - sd, idle, all_pinned); + sd, idle, lb_flags); total_load_moved += load_moved; + if (*lb_flags & (LBF_NEED_BREAK|LBF_ABORT)) + break; + #ifdef CONFIG_PREEMPT /* * NEWIDLE balancing is a source of latency, so preemptible * kernels will stop after the first task is pulled to minimize * the critical section. */ - if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) - break; - - if (raw_spin_is_contended(&this_rq->lock) || - raw_spin_is_contended(&busiest->lock)) + if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) { + *lb_flags |= LBF_ABORT; break; + } #endif } while (load_moved && max_load_move > total_load_moved); @@ -3062,15 +3525,6 @@ struct sg_lb_stats { }; /** - * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. - * @group: The group whose first cpu is to be returned. - */ -static inline unsigned int group_first_cpu(struct sched_group *group) -{ - return cpumask_first(sched_group_cpus(group)); -} - -/** * get_sd_load_idx - Obtain the load index for a given sched domain. * @sd: The sched_domain whose load_idx is to be obtained. * @idle: The Idle status of the CPU for whose sd load_icx is obtained. @@ -3319,7 +3773,7 @@ static void update_cpu_power(struct sched_domain *sd, int cpu) sdg->sgp->power = power; } -static void update_group_power(struct sched_domain *sd, int cpu) +void update_group_power(struct sched_domain *sd, int cpu) { struct sched_domain *child = sd->child; struct sched_group *group, *sdg = sd->groups; @@ -3511,7 +3965,7 @@ static bool update_sd_pick_busiest(struct sched_domain *sd, } /** - * update_sd_lb_stats - Update sched_group's statistics for load balancing. + * update_sd_lb_stats - Update sched_domain's statistics for load balancing. * @sd: sched_domain whose statistics are to be updated. * @this_cpu: Cpu for which load balance is currently performed. * @idle: Idle status of this_cpu @@ -3585,11 +4039,6 @@ static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, } while (sg != sd->groups); } -int __weak arch_sd_sibling_asym_packing(void) -{ - return 0*SD_ASYM_PACKING; -} - /** * check_asym_packing - Check to see if the group is packed into the * sched doman. @@ -3953,7 +4402,7 @@ find_busiest_queue(struct sched_domain *sd, struct sched_group *group, #define MAX_PINNED_INTERVAL 512 /* Working cpumask for load_balance and load_balance_newidle. */ -static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); +DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); static int need_active_balance(struct sched_domain *sd, int idle, int busiest_cpu, int this_cpu) @@ -4004,7 +4453,7 @@ static int load_balance(int this_cpu, struct rq *this_rq, struct sched_domain *sd, enum cpu_idle_type idle, int *balance) { - int ld_moved, all_pinned = 0, active_balance = 0; + int ld_moved, lb_flags = 0, active_balance = 0; struct sched_group *group; unsigned long imbalance; struct rq *busiest; @@ -4045,11 +4494,11 @@ redo: * still unbalanced. ld_moved simply stays zero, so it is * correctly treated as an imbalance. */ - all_pinned = 1; + lb_flags |= LBF_ALL_PINNED; local_irq_save(flags); double_rq_lock(this_rq, busiest); ld_moved = move_tasks(this_rq, this_cpu, busiest, - imbalance, sd, idle, &all_pinned); + imbalance, sd, idle, &lb_flags); double_rq_unlock(this_rq, busiest); local_irq_restore(flags); @@ -4059,8 +4508,18 @@ redo: if (ld_moved && this_cpu != smp_processor_id()) resched_cpu(this_cpu); + if (lb_flags & LBF_ABORT) + goto out_balanced; + + if (lb_flags & LBF_NEED_BREAK) { + lb_flags += LBF_HAD_BREAK - LBF_NEED_BREAK; + if (lb_flags & LBF_ABORT) + goto out_balanced; + goto redo; + } + /* All tasks on this runqueue were pinned by CPU affinity */ - if (unlikely(all_pinned)) { + if (unlikely(lb_flags & LBF_ALL_PINNED)) { cpumask_clear_cpu(cpu_of(busiest), cpus); if (!cpumask_empty(cpus)) goto redo; @@ -4090,7 +4549,7 @@ redo: tsk_cpus_allowed(busiest->curr))) { raw_spin_unlock_irqrestore(&busiest->lock, flags); - all_pinned = 1; + lb_flags |= LBF_ALL_PINNED; goto out_one_pinned; } @@ -4143,7 +4602,8 @@ out_balanced: out_one_pinned: /* tune up the balancing interval */ - if ((all_pinned && sd->balance_interval < MAX_PINNED_INTERVAL) || + if (((lb_flags & LBF_ALL_PINNED) && + sd->balance_interval < MAX_PINNED_INTERVAL) || (sd->balance_interval < sd->max_interval)) sd->balance_interval *= 2; @@ -4156,7 +4616,7 @@ out: * idle_balance is called by schedule() if this_cpu is about to become * idle. Attempts to pull tasks from other CPUs. */ -static void idle_balance(int this_cpu, struct rq *this_rq) +void idle_balance(int this_cpu, struct rq *this_rq) { struct sched_domain *sd; int pulled_task = 0; @@ -4271,28 +4731,16 @@ out_unlock: #ifdef CONFIG_NO_HZ /* * idle load balancing details - * - One of the idle CPUs nominates itself as idle load_balancer, while - * entering idle. - * - This idle load balancer CPU will also go into tickless mode when - * it is idle, just like all other idle CPUs * - When one of the busy CPUs notice that there may be an idle rebalancing * needed, they will kick the idle load balancer, which then does idle * load balancing for all the idle CPUs. */ static struct { - atomic_t load_balancer; - atomic_t first_pick_cpu; - atomic_t second_pick_cpu; cpumask_var_t idle_cpus_mask; - cpumask_var_t grp_idle_mask; + atomic_t nr_cpus; unsigned long next_balance; /* in jiffy units */ } nohz ____cacheline_aligned; -int get_nohz_load_balancer(void) -{ - return atomic_read(&nohz.load_balancer); -} - #if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) /** * lowest_flag_domain - Return lowest sched_domain containing flag. @@ -4329,33 +4777,6 @@ static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) (sd && (sd->flags & flag)); sd = sd->parent) /** - * is_semi_idle_group - Checks if the given sched_group is semi-idle. - * @ilb_group: group to be checked for semi-idleness - * - * Returns: 1 if the group is semi-idle. 0 otherwise. - * - * We define a sched_group to be semi idle if it has atleast one idle-CPU - * and atleast one non-idle CPU. This helper function checks if the given - * sched_group is semi-idle or not. - */ -static inline int is_semi_idle_group(struct sched_group *ilb_group) -{ - cpumask_and(nohz.grp_idle_mask, nohz.idle_cpus_mask, - sched_group_cpus(ilb_group)); - - /* - * A sched_group is semi-idle when it has atleast one busy cpu - * and atleast one idle cpu. - */ - if (cpumask_empty(nohz.grp_idle_mask)) - return 0; - - if (cpumask_equal(nohz.grp_idle_mask, sched_group_cpus(ilb_group))) - return 0; - - return 1; -} -/** * find_new_ilb - Finds the optimum idle load balancer for nomination. * @cpu: The cpu which is nominating a new idle_load_balancer. * @@ -4369,9 +4790,9 @@ static inline int is_semi_idle_group(struct sched_group *ilb_group) */ static int find_new_ilb(int cpu) { + int ilb = cpumask_first(nohz.idle_cpus_mask); + struct sched_group *ilbg; struct sched_domain *sd; - struct sched_group *ilb_group; - int ilb = nr_cpu_ids; /* * Have idle load balancer selection from semi-idle packages only @@ -4389,23 +4810,28 @@ static int find_new_ilb(int cpu) rcu_read_lock(); for_each_flag_domain(cpu, sd, SD_POWERSAVINGS_BALANCE) { - ilb_group = sd->groups; + ilbg = sd->groups; do { - if (is_semi_idle_group(ilb_group)) { - ilb = cpumask_first(nohz.grp_idle_mask); + if (ilbg->group_weight != + atomic_read(&ilbg->sgp->nr_busy_cpus)) { + ilb = cpumask_first_and(nohz.idle_cpus_mask, + sched_group_cpus(ilbg)); goto unlock; } - ilb_group = ilb_group->next; + ilbg = ilbg->next; - } while (ilb_group != sd->groups); + } while (ilbg != sd->groups); } unlock: rcu_read_unlock(); out_done: - return ilb; + if (ilb < nr_cpu_ids && idle_cpu(ilb)) + return ilb; + + return nr_cpu_ids; } #else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ static inline int find_new_ilb(int call_cpu) @@ -4425,102 +4851,98 @@ static void nohz_balancer_kick(int cpu) nohz.next_balance++; - ilb_cpu = get_nohz_load_balancer(); + ilb_cpu = find_new_ilb(cpu); - if (ilb_cpu >= nr_cpu_ids) { - ilb_cpu = cpumask_first(nohz.idle_cpus_mask); - if (ilb_cpu >= nr_cpu_ids) - return; - } + if (ilb_cpu >= nr_cpu_ids) + return; - if (!cpu_rq(ilb_cpu)->nohz_balance_kick) { - cpu_rq(ilb_cpu)->nohz_balance_kick = 1; + if (test_and_set_bit(NOHZ_BALANCE_KICK, nohz_flags(ilb_cpu))) + return; + /* + * Use smp_send_reschedule() instead of resched_cpu(). + * This way we generate a sched IPI on the target cpu which + * is idle. And the softirq performing nohz idle load balance + * will be run before returning from the IPI. + */ + smp_send_reschedule(ilb_cpu); + return; +} - smp_mb(); - /* - * Use smp_send_reschedule() instead of resched_cpu(). - * This way we generate a sched IPI on the target cpu which - * is idle. And the softirq performing nohz idle load balance - * will be run before returning from the IPI. - */ - smp_send_reschedule(ilb_cpu); +static inline void clear_nohz_tick_stopped(int cpu) +{ + if (unlikely(test_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu)))) { + cpumask_clear_cpu(cpu, nohz.idle_cpus_mask); + atomic_dec(&nohz.nr_cpus); + clear_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu)); } - return; } -/* - * This routine will try to nominate the ilb (idle load balancing) - * owner among the cpus whose ticks are stopped. ilb owner will do the idle - * load balancing on behalf of all those cpus. - * - * When the ilb owner becomes busy, we will not have new ilb owner until some - * idle CPU wakes up and goes back to idle or some busy CPU tries to kick - * idle load balancing by kicking one of the idle CPUs. - * - * Ticks are stopped for the ilb owner as well, with busy CPU kicking this - * ilb owner CPU in future (when there is a need for idle load balancing on - * behalf of all idle CPUs). - */ -void select_nohz_load_balancer(int stop_tick) +static inline void set_cpu_sd_state_busy(void) { + struct sched_domain *sd; int cpu = smp_processor_id(); - if (stop_tick) { - if (!cpu_active(cpu)) { - if (atomic_read(&nohz.load_balancer) != cpu) - return; + if (!test_bit(NOHZ_IDLE, nohz_flags(cpu))) + return; + clear_bit(NOHZ_IDLE, nohz_flags(cpu)); - /* - * If we are going offline and still the leader, - * give up! - */ - if (atomic_cmpxchg(&nohz.load_balancer, cpu, - nr_cpu_ids) != cpu) - BUG(); + rcu_read_lock(); + for_each_domain(cpu, sd) + atomic_inc(&sd->groups->sgp->nr_busy_cpus); + rcu_read_unlock(); +} - return; - } +void set_cpu_sd_state_idle(void) +{ + struct sched_domain *sd; + int cpu = smp_processor_id(); - cpumask_set_cpu(cpu, nohz.idle_cpus_mask); + if (test_bit(NOHZ_IDLE, nohz_flags(cpu))) + return; + set_bit(NOHZ_IDLE, nohz_flags(cpu)); - if (atomic_read(&nohz.first_pick_cpu) == cpu) - atomic_cmpxchg(&nohz.first_pick_cpu, cpu, nr_cpu_ids); - if (atomic_read(&nohz.second_pick_cpu) == cpu) - atomic_cmpxchg(&nohz.second_pick_cpu, cpu, nr_cpu_ids); + rcu_read_lock(); + for_each_domain(cpu, sd) + atomic_dec(&sd->groups->sgp->nr_busy_cpus); + rcu_read_unlock(); +} - if (atomic_read(&nohz.load_balancer) >= nr_cpu_ids) { - int new_ilb; +/* + * This routine will record that this cpu is going idle with tick stopped. + * This info will be used in performing idle load balancing in the future. + */ +void select_nohz_load_balancer(int stop_tick) +{ + int cpu = smp_processor_id(); - /* make me the ilb owner */ - if (atomic_cmpxchg(&nohz.load_balancer, nr_cpu_ids, - cpu) != nr_cpu_ids) - return; + /* + * If this cpu is going down, then nothing needs to be done. + */ + if (!cpu_active(cpu)) + return; - /* - * Check to see if there is a more power-efficient - * ilb. - */ - new_ilb = find_new_ilb(cpu); - if (new_ilb < nr_cpu_ids && new_ilb != cpu) { - atomic_set(&nohz.load_balancer, nr_cpu_ids); - resched_cpu(new_ilb); - return; - } - return; - } - } else { - if (!cpumask_test_cpu(cpu, nohz.idle_cpus_mask)) + if (stop_tick) { + if (test_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu))) return; - cpumask_clear_cpu(cpu, nohz.idle_cpus_mask); - - if (atomic_read(&nohz.load_balancer) == cpu) - if (atomic_cmpxchg(&nohz.load_balancer, cpu, - nr_cpu_ids) != cpu) - BUG(); + cpumask_set_cpu(cpu, nohz.idle_cpus_mask); + atomic_inc(&nohz.nr_cpus); + set_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu)); } return; } + +static int __cpuinit sched_ilb_notifier(struct notifier_block *nfb, + unsigned long action, void *hcpu) +{ + switch (action & ~CPU_TASKS_FROZEN) { + case CPU_DYING: + clear_nohz_tick_stopped(smp_processor_id()); + return NOTIFY_OK; + default: + return NOTIFY_DONE; + } +} #endif static DEFINE_SPINLOCK(balancing); @@ -4531,7 +4953,7 @@ static unsigned long __read_mostly max_load_balance_interval = HZ/10; * Scale the max load_balance interval with the number of CPUs in the system. * This trades load-balance latency on larger machines for less cross talk. */ -static void update_max_interval(void) +void update_max_interval(void) { max_load_balance_interval = HZ*num_online_cpus()/10; } @@ -4623,11 +5045,12 @@ static void nohz_idle_balance(int this_cpu, enum cpu_idle_type idle) struct rq *rq; int balance_cpu; - if (idle != CPU_IDLE || !this_rq->nohz_balance_kick) - return; + if (idle != CPU_IDLE || + !test_bit(NOHZ_BALANCE_KICK, nohz_flags(this_cpu))) + goto end; for_each_cpu(balance_cpu, nohz.idle_cpus_mask) { - if (balance_cpu == this_cpu) + if (balance_cpu == this_cpu || !idle_cpu(balance_cpu)) continue; /* @@ -4635,10 +5058,8 @@ static void nohz_idle_balance(int this_cpu, enum cpu_idle_type idle) * work being done for other cpus. Next load * balancing owner will pick it up. */ - if (need_resched()) { - this_rq->nohz_balance_kick = 0; + if (need_resched()) break; - } raw_spin_lock_irq(&this_rq->lock); update_rq_clock(this_rq); @@ -4652,53 +5073,71 @@ static void nohz_idle_balance(int this_cpu, enum cpu_idle_type idle) this_rq->next_balance = rq->next_balance; } nohz.next_balance = this_rq->next_balance; - this_rq->nohz_balance_kick = 0; +end: + clear_bit(NOHZ_BALANCE_KICK, nohz_flags(this_cpu)); } /* - * Current heuristic for kicking the idle load balancer - * - first_pick_cpu is the one of the busy CPUs. It will kick - * idle load balancer when it has more than one process active. This - * eliminates the need for idle load balancing altogether when we have - * only one running process in the system (common case). - * - If there are more than one busy CPU, idle load balancer may have - * to run for active_load_balance to happen (i.e., two busy CPUs are - * SMT or core siblings and can run better if they move to different - * physical CPUs). So, second_pick_cpu is the second of the busy CPUs - * which will kick idle load balancer as soon as it has any load. + * Current heuristic for kicking the idle load balancer in the presence + * of an idle cpu is the system. + * - This rq has more than one task. + * - At any scheduler domain level, this cpu's scheduler group has multiple + * busy cpu's exceeding the group's power. + * - For SD_ASYM_PACKING, if the lower numbered cpu's in the scheduler + * domain span are idle. */ static inline int nohz_kick_needed(struct rq *rq, int cpu) { unsigned long now = jiffies; - int ret; - int first_pick_cpu, second_pick_cpu; + struct sched_domain *sd; - if (time_before(now, nohz.next_balance)) + if (unlikely(idle_cpu(cpu))) return 0; - if (idle_cpu(cpu)) - return 0; + /* + * We may be recently in ticked or tickless idle mode. At the first + * busy tick after returning from idle, we will update the busy stats. + */ + set_cpu_sd_state_busy(); + clear_nohz_tick_stopped(cpu); - first_pick_cpu = atomic_read(&nohz.first_pick_cpu); - second_pick_cpu = atomic_read(&nohz.second_pick_cpu); + /* + * None are in tickless mode and hence no need for NOHZ idle load + * balancing. + */ + if (likely(!atomic_read(&nohz.nr_cpus))) + return 0; - if (first_pick_cpu < nr_cpu_ids && first_pick_cpu != cpu && - second_pick_cpu < nr_cpu_ids && second_pick_cpu != cpu) + if (time_before(now, nohz.next_balance)) return 0; - ret = atomic_cmpxchg(&nohz.first_pick_cpu, nr_cpu_ids, cpu); - if (ret == nr_cpu_ids || ret == cpu) { - atomic_cmpxchg(&nohz.second_pick_cpu, cpu, nr_cpu_ids); - if (rq->nr_running > 1) - return 1; - } else { - ret = atomic_cmpxchg(&nohz.second_pick_cpu, nr_cpu_ids, cpu); - if (ret == nr_cpu_ids || ret == cpu) { - if (rq->nr_running) - return 1; - } + if (rq->nr_running >= 2) + goto need_kick; + + rcu_read_lock(); + for_each_domain(cpu, sd) { + struct sched_group *sg = sd->groups; + struct sched_group_power *sgp = sg->sgp; + int nr_busy = atomic_read(&sgp->nr_busy_cpus); + + if (sd->flags & SD_SHARE_PKG_RESOURCES && nr_busy > 1) + goto need_kick_unlock; + + if (sd->flags & SD_ASYM_PACKING && nr_busy != sg->group_weight + && (cpumask_first_and(nohz.idle_cpus_mask, + sched_domain_span(sd)) < cpu)) + goto need_kick_unlock; + + if (!(sd->flags & (SD_SHARE_PKG_RESOURCES | SD_ASYM_PACKING))) + break; } + rcu_read_unlock(); return 0; + +need_kick_unlock: + rcu_read_unlock(); +need_kick: + return 1; } #else static void nohz_idle_balance(int this_cpu, enum cpu_idle_type idle) { } @@ -4733,14 +5172,14 @@ static inline int on_null_domain(int cpu) /* * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing. */ -static inline void trigger_load_balance(struct rq *rq, int cpu) +void trigger_load_balance(struct rq *rq, int cpu) { /* Don't need to rebalance while attached to NULL domain */ if (time_after_eq(jiffies, rq->next_balance) && likely(!on_null_domain(cpu))) raise_softirq(SCHED_SOFTIRQ); #ifdef CONFIG_NO_HZ - else if (nohz_kick_needed(rq, cpu) && likely(!on_null_domain(cpu))) + if (nohz_kick_needed(rq, cpu) && likely(!on_null_domain(cpu))) nohz_balancer_kick(cpu); #endif } @@ -4755,15 +5194,6 @@ static void rq_offline_fair(struct rq *rq) update_sysctl(); } -#else /* CONFIG_SMP */ - -/* - * on UP we do not need to balance between CPUs: - */ -static inline void idle_balance(int cpu, struct rq *rq) -{ -} - #endif /* CONFIG_SMP */ /* @@ -4787,8 +5217,8 @@ static void task_tick_fair(struct rq *rq, struct task_struct *curr, int queued) */ static void task_fork_fair(struct task_struct *p) { - struct cfs_rq *cfs_rq = task_cfs_rq(current); - struct sched_entity *se = &p->se, *curr = cfs_rq->curr; + struct cfs_rq *cfs_rq; + struct sched_entity *se = &p->se, *curr; int this_cpu = smp_processor_id(); struct rq *rq = this_rq(); unsigned long flags; @@ -4797,6 +5227,9 @@ static void task_fork_fair(struct task_struct *p) update_rq_clock(rq); + cfs_rq = task_cfs_rq(current); + curr = cfs_rq->curr; + if (unlikely(task_cpu(p) != this_cpu)) { rcu_read_lock(); __set_task_cpu(p, this_cpu); @@ -4906,6 +5339,16 @@ static void set_curr_task_fair(struct rq *rq) } } +void init_cfs_rq(struct cfs_rq *cfs_rq) +{ + cfs_rq->tasks_timeline = RB_ROOT; + INIT_LIST_HEAD(&cfs_rq->tasks); + cfs_rq->min_vruntime = (u64)(-(1LL << 20)); +#ifndef CONFIG_64BIT + cfs_rq->min_vruntime_copy = cfs_rq->min_vruntime; +#endif +} + #ifdef CONFIG_FAIR_GROUP_SCHED static void task_move_group_fair(struct task_struct *p, int on_rq) { @@ -4922,13 +5365,182 @@ static void task_move_group_fair(struct task_struct *p, int on_rq) * to another cgroup's rq. This does somewhat interfere with the * fair sleeper stuff for the first placement, but who cares. */ + /* + * When !on_rq, vruntime of the task has usually NOT been normalized. + * But there are some cases where it has already been normalized: + * + * - Moving a forked child which is waiting for being woken up by + * wake_up_new_task(). + * - Moving a task which has been woken up by try_to_wake_up() and + * waiting for actually being woken up by sched_ttwu_pending(). + * + * To prevent boost or penalty in the new cfs_rq caused by delta + * min_vruntime between the two cfs_rqs, we skip vruntime adjustment. + */ + if (!on_rq && (!p->se.sum_exec_runtime || p->state == TASK_WAKING)) + on_rq = 1; + if (!on_rq) p->se.vruntime -= cfs_rq_of(&p->se)->min_vruntime; set_task_rq(p, task_cpu(p)); if (!on_rq) p->se.vruntime += cfs_rq_of(&p->se)->min_vruntime; } + +void free_fair_sched_group(struct task_group *tg) +{ + int i; + + destroy_cfs_bandwidth(tg_cfs_bandwidth(tg)); + + for_each_possible_cpu(i) { + if (tg->cfs_rq) + kfree(tg->cfs_rq[i]); + if (tg->se) + kfree(tg->se[i]); + } + + kfree(tg->cfs_rq); + kfree(tg->se); +} + +int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent) +{ + struct cfs_rq *cfs_rq; + struct sched_entity *se; + int i; + + tg->cfs_rq = kzalloc(sizeof(cfs_rq) * nr_cpu_ids, GFP_KERNEL); + if (!tg->cfs_rq) + goto err; + tg->se = kzalloc(sizeof(se) * nr_cpu_ids, GFP_KERNEL); + if (!tg->se) + goto err; + + tg->shares = NICE_0_LOAD; + + init_cfs_bandwidth(tg_cfs_bandwidth(tg)); + + for_each_possible_cpu(i) { + cfs_rq = kzalloc_node(sizeof(struct cfs_rq), + GFP_KERNEL, cpu_to_node(i)); + if (!cfs_rq) + goto err; + + se = kzalloc_node(sizeof(struct sched_entity), + GFP_KERNEL, cpu_to_node(i)); + if (!se) + goto err_free_rq; + + init_cfs_rq(cfs_rq); + init_tg_cfs_entry(tg, cfs_rq, se, i, parent->se[i]); + } + + return 1; + +err_free_rq: + kfree(cfs_rq); +err: + return 0; +} + +void unregister_fair_sched_group(struct task_group *tg, int cpu) +{ + struct rq *rq = cpu_rq(cpu); + unsigned long flags; + + /* + * Only empty task groups can be destroyed; so we can speculatively + * check on_list without danger of it being re-added. + */ + if (!tg->cfs_rq[cpu]->on_list) + return; + + raw_spin_lock_irqsave(&rq->lock, flags); + list_del_leaf_cfs_rq(tg->cfs_rq[cpu]); + raw_spin_unlock_irqrestore(&rq->lock, flags); +} + +void init_tg_cfs_entry(struct task_group *tg, struct cfs_rq *cfs_rq, + struct sched_entity *se, int cpu, + struct sched_entity *parent) +{ + struct rq *rq = cpu_rq(cpu); + + cfs_rq->tg = tg; + cfs_rq->rq = rq; +#ifdef CONFIG_SMP + /* allow initial update_cfs_load() to truncate */ + cfs_rq->load_stamp = 1; #endif + init_cfs_rq_runtime(cfs_rq); + + tg->cfs_rq[cpu] = cfs_rq; + tg->se[cpu] = se; + + /* se could be NULL for root_task_group */ + if (!se) + return; + + if (!parent) + se->cfs_rq = &rq->cfs; + else + se->cfs_rq = parent->my_q; + + se->my_q = cfs_rq; + update_load_set(&se->load, 0); + se->parent = parent; +} + +static DEFINE_MUTEX(shares_mutex); + +int sched_group_set_shares(struct task_group *tg, unsigned long shares) +{ + int i; + unsigned long flags; + + /* + * We can't change the weight of the root cgroup. + */ + if (!tg->se[0]) + return -EINVAL; + + shares = clamp(shares, scale_load(MIN_SHARES), scale_load(MAX_SHARES)); + + mutex_lock(&shares_mutex); + if (tg->shares == shares) + goto done; + + tg->shares = shares; + for_each_possible_cpu(i) { + struct rq *rq = cpu_rq(i); + struct sched_entity *se; + + se = tg->se[i]; + /* Propagate contribution to hierarchy */ + raw_spin_lock_irqsave(&rq->lock, flags); + for_each_sched_entity(se) + update_cfs_shares(group_cfs_rq(se)); + raw_spin_unlock_irqrestore(&rq->lock, flags); + } + +done: + mutex_unlock(&shares_mutex); + return 0; +} +#else /* CONFIG_FAIR_GROUP_SCHED */ + +void free_fair_sched_group(struct task_group *tg) { } + +int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent) +{ + return 1; +} + +void unregister_fair_sched_group(struct task_group *tg, int cpu) { } + +#endif /* CONFIG_FAIR_GROUP_SCHED */ + static unsigned int get_rr_interval_fair(struct rq *rq, struct task_struct *task) { @@ -4948,7 +5560,7 @@ static unsigned int get_rr_interval_fair(struct rq *rq, struct task_struct *task /* * All the scheduling class methods: */ -static const struct sched_class fair_sched_class = { +const struct sched_class fair_sched_class = { .next = &idle_sched_class, .enqueue_task = enqueue_task_fair, .dequeue_task = dequeue_task_fair, @@ -4985,7 +5597,7 @@ static const struct sched_class fair_sched_class = { }; #ifdef CONFIG_SCHED_DEBUG -static void print_cfs_stats(struct seq_file *m, int cpu) +void print_cfs_stats(struct seq_file *m, int cpu) { struct cfs_rq *cfs_rq; @@ -4995,3 +5607,16 @@ static void print_cfs_stats(struct seq_file *m, int cpu) rcu_read_unlock(); } #endif + +__init void init_sched_fair_class(void) +{ +#ifdef CONFIG_SMP + open_softirq(SCHED_SOFTIRQ, run_rebalance_domains); + +#ifdef CONFIG_NO_HZ + zalloc_cpumask_var(&nohz.idle_cpus_mask, GFP_NOWAIT); + cpu_notifier(sched_ilb_notifier, 0); +#endif +#endif /* SMP */ + +} |