diff --git a/include/linux/sched/topology.h b/include/linux/sched/topology.h index 9680886277c97ce4e8e4fafc752b61d3a3995a86..62a1e22b057f4f1d64dce55cc0c187c7b217b4c6 100644 --- a/include/linux/sched/topology.h +++ b/include/linux/sched/topology.h @@ -153,8 +153,12 @@ struct sched_domain { struct rcu_head rcu; /* used during destruction */ }; struct sched_domain_shared *shared; - +#ifndef __GENKSYMS__ + unsigned int imb_numa_nr; /* Nr running tasks that allows a NUMA imbalance */ + KABI_FILL_HOLE(unsigned int kabi_hole) +#else KABI_RESERVE(1) +#endif KABI_RESERVE(2) unsigned int span_weight; diff --git a/kernel/sched/core.c b/kernel/sched/core.c index d06ac41204c63a461c1c7c4957ff8ba9c93f23c5..ee6e459611f7a0692e6bf7937db048880ffffdf5 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2933,13 +2933,6 @@ void sched_ttwu_pending(void *arg) if (!llist) return; - /* - * rq::ttwu_pending racy indication of out-standing wakeups. - * Races such that false-negatives are possible, since they - * are shorter lived that false-positives would be. - */ - WRITE_ONCE(rq->ttwu_pending, 0); - rq_lock_irqsave(rq, &rf); update_rq_clock(rq); @@ -2953,6 +2946,17 @@ void sched_ttwu_pending(void *arg) ttwu_do_activate(rq, p, p->sched_remote_wakeup ? WF_MIGRATED : 0, &rf); } + /* + * Must be after enqueueing at least once task such that + * idle_cpu() does not observe a false-negative -- if it does, + * it is possible for select_idle_siblings() to stack a number + * of tasks on this CPU during that window. + * + * It is ok to clear ttwu_pending when another task pending. + * We will receive IPI after local irq enabled and then enqueue it. + * Since now nr_running > 0, idle_cpu() will always get correct result. + */ + WRITE_ONCE(rq->ttwu_pending, 0); rq_unlock_irqrestore(rq, &rf); } @@ -3026,7 +3030,7 @@ bool cpus_share_lowest_cache(int this_cpu, int that_cpu) return per_cpu(sd_lowest_cache_id, this_cpu) == per_cpu(sd_lowest_cache_id, that_cpu); } -static inline bool ttwu_queue_cond(int cpu, int wake_flags) +static inline bool ttwu_queue_cond(int cpu) { /* * If the CPU does not share cache, then queue the task on the @@ -3035,17 +3039,21 @@ static inline bool ttwu_queue_cond(int cpu, int wake_flags) if (!cpus_share_cache(smp_processor_id(), cpu)) return true; + if (cpu == smp_processor_id()) + return false; + /* - * If the task is descheduling and the only running task on the - * CPU then use the wakelist to offload the task activation to - * the soon-to-be-idle CPU as the current CPU is likely busy. - * nr_running is checked to avoid unnecessary task stacking. + * If the wakee cpu is idle, or the task is descheduling and the + * only running task on the CPU, then use the wakelist to offload + * the task activation to the idle (or soon-to-be-idle) CPU as + * the current CPU is likely busy. nr_running is checked to + * avoid unnecessary task stacking. * * Note that we can only get here with (wakee) p->on_rq=0, * p->on_cpu can be whatever, we've done the dequeue, so * the wakee has been accounted out of ->nr_running. */ - if ((wake_flags & WF_ON_CPU) && !cpu_rq(cpu)->nr_running) + if (!cpu_rq(cpu)->nr_running) return true; return false; @@ -3053,10 +3061,7 @@ static inline bool ttwu_queue_cond(int cpu, int wake_flags) static bool ttwu_queue_wakelist(struct task_struct *p, int cpu, int wake_flags) { - if (sched_feat(TTWU_QUEUE) && ttwu_queue_cond(cpu, wake_flags)) { - if (WARN_ON_ONCE(cpu == smp_processor_id())) - return false; - + if (sched_feat(TTWU_QUEUE) && ttwu_queue_cond(cpu)) { sched_clock_cpu(cpu); /* Sync clocks across CPUs */ __ttwu_queue_wakelist(p, cpu, wake_flags); return true; @@ -3333,7 +3338,7 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags) * scheduling. */ if (smp_load_acquire(&p->on_cpu) && - ttwu_queue_wakelist(p, task_cpu(p), wake_flags | WF_ON_CPU)) + ttwu_queue_wakelist(p, task_cpu(p), wake_flags)) goto unlock; /* @@ -3895,7 +3900,8 @@ static inline void prepare_task(struct task_struct *next) * Claim the task as running, we do this before switching to it * such that any running task will have this set. * - * See the ttwu() WF_ON_CPU case and its ordering comment. + * See the smp_load_acquire(&p->on_cpu) case in ttwu() and + * its ordering comment. */ WRITE_ONCE(next->on_cpu, 1); #endif diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 32ef2dab67080f9cba417574454fafa81ca92ad1..0b168eacd8bf6f165653a483bd2e2b39dd7e8113 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -1524,6 +1524,7 @@ struct task_numa_env { int src_cpu, src_nid; int dst_cpu, dst_nid; + int imb_numa_nr; struct numa_stats src_stats, dst_stats; @@ -1539,7 +1540,7 @@ static unsigned long cpu_load(struct rq *rq); static unsigned long cpu_runnable(struct rq *rq); static unsigned long cpu_util(int cpu); static inline long adjust_numa_imbalance(int imbalance, - int dst_running, int dst_weight); + int dst_running, int imb_numa_nr); static inline enum numa_type numa_classify(unsigned int imbalance_pct, @@ -1920,7 +1921,7 @@ static void task_numa_find_cpu(struct task_numa_env *env, dst_running = env->dst_stats.nr_running + 1; imbalance = max(0, dst_running - src_running); imbalance = adjust_numa_imbalance(imbalance, dst_running, - env->dst_stats.weight); + env->imb_numa_nr); /* Use idle CPU if there is no imbalance */ if (!imbalance) { @@ -1985,8 +1986,10 @@ static int task_numa_migrate(struct task_struct *p) */ rcu_read_lock(); sd = rcu_dereference(per_cpu(sd_numa, env.src_cpu)); - if (sd) + if (sd) { env.imbalance_pct = 100 + (sd->imbalance_pct - 100) / 2; + env.imb_numa_nr = sd->imb_numa_nr; + } rcu_read_unlock(); /* @@ -3084,6 +3087,9 @@ dequeue_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se) { sub_positive(&cfs_rq->avg.load_avg, se->avg.load_avg); sub_positive(&cfs_rq->avg.load_sum, se_weight(se) * se->avg.load_sum); + /* See update_cfs_rq_load_avg() */ + cfs_rq->avg.load_sum = max_t(u32, cfs_rq->avg.load_sum, + cfs_rq->avg.load_avg * PELT_MIN_DIVIDER); } #else static inline void @@ -3449,11 +3455,11 @@ void set_task_rq_fair(struct sched_entity *se, static inline void update_tg_cfs_util(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq *gcfs_rq) { - long delta = gcfs_rq->avg.util_avg - se->avg.util_avg; - u32 divider; + long delta_sum, delta_avg = gcfs_rq->avg.util_avg - se->avg.util_avg; + u32 new_sum, divider; /* Nothing to update */ - if (!delta) + if (!delta_avg) return; /* @@ -3462,23 +3468,30 @@ update_tg_cfs_util(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq */ divider = get_pelt_divider(&cfs_rq->avg); + /* Set new sched_entity's utilization */ se->avg.util_avg = gcfs_rq->avg.util_avg; - se->avg.util_sum = se->avg.util_avg * divider; + new_sum = se->avg.util_avg * divider; + delta_sum = (long)new_sum - (long)se->avg.util_sum; + se->avg.util_sum = new_sum; /* Update parent cfs_rq utilization */ - add_positive(&cfs_rq->avg.util_avg, delta); - cfs_rq->avg.util_sum = cfs_rq->avg.util_avg * divider; + add_positive(&cfs_rq->avg.util_avg, delta_avg); + add_positive(&cfs_rq->avg.util_sum, delta_sum); + + /* See update_cfs_rq_load_avg() */ + cfs_rq->avg.util_sum = max_t(u32, cfs_rq->avg.util_sum, + cfs_rq->avg.util_avg * PELT_MIN_DIVIDER); } static inline void update_tg_cfs_runnable(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq *gcfs_rq) { - long delta = gcfs_rq->avg.runnable_avg - se->avg.runnable_avg; - u32 divider; + long delta_sum, delta_avg = gcfs_rq->avg.runnable_avg - se->avg.runnable_avg; + u32 new_sum, divider; /* Nothing to update */ - if (!delta) + if (!delta_avg) return; /* @@ -3489,19 +3502,25 @@ update_tg_cfs_runnable(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cf /* Set new sched_entity's runnable */ se->avg.runnable_avg = gcfs_rq->avg.runnable_avg; - se->avg.runnable_sum = se->avg.runnable_avg * divider; + new_sum = se->avg.runnable_avg * divider; + delta_sum = (long)new_sum - (long)se->avg.runnable_sum; + se->avg.runnable_sum = new_sum; /* Update parent cfs_rq runnable */ - add_positive(&cfs_rq->avg.runnable_avg, delta); - cfs_rq->avg.runnable_sum = cfs_rq->avg.runnable_avg * divider; + add_positive(&cfs_rq->avg.runnable_avg, delta_avg); + add_positive(&cfs_rq->avg.runnable_sum, delta_sum); + /* See update_cfs_rq_load_avg() */ + cfs_rq->avg.runnable_sum = max_t(u32, cfs_rq->avg.runnable_sum, + cfs_rq->avg.runnable_avg * PELT_MIN_DIVIDER); } static inline void update_tg_cfs_load(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq *gcfs_rq) { - long delta, running_sum, runnable_sum = gcfs_rq->prop_runnable_sum; + long delta_avg, running_sum, runnable_sum = gcfs_rq->prop_runnable_sum; unsigned long load_avg; u64 load_sum = 0; + s64 delta_sum; u32 divider; if (!runnable_sum) @@ -3528,7 +3547,7 @@ update_tg_cfs_load(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq * assuming all tasks are equally runnable. */ if (scale_load_down(gcfs_rq->load.weight)) { - load_sum = div_s64(gcfs_rq->avg.load_sum, + load_sum = div_u64(gcfs_rq->avg.load_sum, scale_load_down(gcfs_rq->load.weight)); } @@ -3545,16 +3564,22 @@ update_tg_cfs_load(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq running_sum = se->avg.util_sum >> SCHED_CAPACITY_SHIFT; runnable_sum = max(runnable_sum, running_sum); - load_sum = (s64)se_weight(se) * runnable_sum; - load_avg = div_s64(load_sum, divider); + load_sum = se_weight(se) * runnable_sum; + load_avg = div_u64(load_sum, divider); + + delta_avg = load_avg - se->avg.load_avg; + if (!delta_avg) + return; - delta = load_avg - se->avg.load_avg; + delta_sum = load_sum - (s64)se_weight(se) * se->avg.load_sum; se->avg.load_sum = runnable_sum; se->avg.load_avg = load_avg; - - add_positive(&cfs_rq->avg.load_avg, delta); - cfs_rq->avg.load_sum = cfs_rq->avg.load_avg * divider; + add_positive(&cfs_rq->avg.load_avg, delta_avg); + add_positive(&cfs_rq->avg.load_sum, delta_sum); + /* See update_cfs_rq_load_avg() */ + cfs_rq->avg.load_sum = max_t(u32, cfs_rq->avg.load_sum, + cfs_rq->avg.load_avg * PELT_MIN_DIVIDER); } static inline void add_tg_cfs_propagate(struct cfs_rq *cfs_rq, long runnable_sum) @@ -3670,7 +3695,9 @@ update_cfs_rq_load_avg(u64 now, struct cfs_rq *cfs_rq) r = removed_load; sub_positive(&sa->load_avg, r); - sa->load_sum = sa->load_avg * divider; + sub_positive(&sa->load_sum, r * divider); + /* See sa->util_sum below */ + sa->load_sum = max_t(u32, sa->load_sum, sa->load_avg * PELT_MIN_DIVIDER); r = removed_util; sub_positive(&sa->util_avg, r); @@ -3690,7 +3717,10 @@ update_cfs_rq_load_avg(u64 now, struct cfs_rq *cfs_rq) r = removed_runnable; sub_positive(&sa->runnable_avg, r); - sa->runnable_sum = sa->runnable_avg * divider; + sub_positive(&sa->runnable_sum, r * divider); + /* See sa->util_sum above */ + sa->runnable_sum = max_t(u32, sa->runnable_sum, + sa->runnable_avg * PELT_MIN_DIVIDER); /* * removed_runnable is the unweighted version of removed_load so we @@ -3777,17 +3807,18 @@ static void attach_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *s */ static void detach_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se) { - /* - * cfs_rq->avg.period_contrib can be used for both cfs_rq and se. - * See ___update_load_avg() for details. - */ - u32 divider = get_pelt_divider(&cfs_rq->avg); - dequeue_load_avg(cfs_rq, se); sub_positive(&cfs_rq->avg.util_avg, se->avg.util_avg); - cfs_rq->avg.util_sum = cfs_rq->avg.util_avg * divider; + sub_positive(&cfs_rq->avg.util_sum, se->avg.util_sum); + /* See update_cfs_rq_load_avg() */ + cfs_rq->avg.util_sum = max_t(u32, cfs_rq->avg.util_sum, + cfs_rq->avg.util_avg * PELT_MIN_DIVIDER); + sub_positive(&cfs_rq->avg.runnable_avg, se->avg.runnable_avg); - cfs_rq->avg.runnable_sum = cfs_rq->avg.runnable_avg * divider; + sub_positive(&cfs_rq->avg.runnable_sum, se->avg.runnable_sum); + /* See update_cfs_rq_load_avg() */ + cfs_rq->avg.runnable_sum = max_t(u32, cfs_rq->avg.runnable_sum, + cfs_rq->avg.runnable_avg * PELT_MIN_DIVIDER); add_tg_cfs_propagate(cfs_rq, -se->avg.load_sum); @@ -9965,9 +9996,9 @@ static bool update_pick_idlest(struct sched_group *idlest, * This is an approximation as the number of running tasks may not be * related to the number of busy CPUs due to sched_setaffinity. */ -static inline bool allow_numa_imbalance(int dst_running, int dst_weight) +static inline bool allow_numa_imbalance(int running, int imb_numa_nr) { - return (dst_running < (dst_weight >> 2)); + return running <= imb_numa_nr; } /* @@ -10106,12 +10137,13 @@ find_idlest_group(struct sched_domain *sd, struct task_struct *p, int this_cpu) return idlest; #endif /* - * Otherwise, keep the task on this node to stay close - * its wakeup source and improve locality. If there is - * a real need of migration, periodic load balance will - * take care of it. + * Otherwise, keep the task close to the wakeup source + * and improve locality if the number of running tasks + * would remain below threshold where an imbalance is + * allowed. If there is a real need of migration, + * periodic load balance will take care of it. */ - if (allow_numa_imbalance(local_sgs.sum_nr_running, sd->span_weight)) + if (allow_numa_imbalance(local_sgs.sum_nr_running + 1, sd->imb_numa_nr)) return NULL; } @@ -10291,9 +10323,9 @@ static inline void update_sd_lb_stats(struct lb_env *env, struct sd_lb_stats *sd #define NUMA_IMBALANCE_MIN 2 static inline long adjust_numa_imbalance(int imbalance, - int dst_running, int dst_weight) + int dst_running, int imb_numa_nr) { - if (!allow_numa_imbalance(dst_running, dst_weight)) + if (!allow_numa_imbalance(dst_running, imb_numa_nr)) return imbalance; /* @@ -10405,7 +10437,7 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s /* Consider allowing a small imbalance between NUMA groups */ if (env->sd->flags & SD_NUMA) { env->imbalance = adjust_numa_imbalance(env->imbalance, - busiest->sum_nr_running, busiest->group_weight); + local->sum_nr_running + 1, env->sd->imb_numa_nr); } return; diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 3bf1e7e3f9958861d4a022720e6fbc4392f07b18..10b0d7e52a9893c2277b79d02808da3339a39333 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -2085,7 +2085,6 @@ static inline int task_on_rq_migrating(struct task_struct *p) #define WF_SYNC 0x01 /* Waker goes to sleep after wakeup */ #define WF_FORK 0x02 /* Child wakeup after fork */ #define WF_MIGRATED 0x04 /* Internal use, task got migrated */ -#define WF_ON_CPU 0x08 /* Wakee is on_cpu */ /* * To aid in avoiding the subversion of "niceness" due to uneven distribution diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c index 3208293d68fac66d1b2d158349b5dadab3c2d506..abdb60ca1c80e273b5ed776917913625855df58d 100644 --- a/kernel/sched/topology.c +++ b/kernel/sched/topology.c @@ -2343,6 +2343,59 @@ build_sched_domains(const struct cpumask *cpu_map, struct sched_domain_attr *att } } + /* + * Calculate an allowed NUMA imbalance such that LLCs do not get + * imbalanced. + */ + for_each_cpu(i, cpu_map) { + unsigned int imb = 0; + unsigned int imb_span = 1; + + for (sd = *per_cpu_ptr(d.sd, i); sd; sd = sd->parent) { + struct sched_domain *child = sd->child; + + if (!(sd->flags & SD_SHARE_PKG_RESOURCES) && child && + (child->flags & SD_SHARE_PKG_RESOURCES)) { + struct sched_domain *top, *top_p; + unsigned int nr_llcs; + + /* + * For a single LLC per node, allow an + * imbalance up to 25% of the node. This is an + * arbitrary cutoff based on SMT-2 to balance + * between memory bandwidth and avoiding + * premature sharing of HT resources and SMT-4 + * or SMT-8 *may* benefit from a different + * cutoff. + * + * For multiple LLCs, allow an imbalance + * until multiple tasks would share an LLC + * on one node while LLCs on another node + * remain idle. + */ + nr_llcs = sd->span_weight / child->span_weight; + if (nr_llcs == 1) + imb = sd->span_weight >> 2; + else + imb = nr_llcs; + sd->imb_numa_nr = imb; + + /* Set span based on the first NUMA domain. */ + top = sd; + top_p = top->parent; + while (top_p && !(top_p->flags & SD_NUMA)) { + top = top->parent; + top_p = top->parent; + } + imb_span = top_p ? top_p->span_weight : sd->span_weight; + } else { + int factor = max(1U, (sd->span_weight / imb_span)); + + sd->imb_numa_nr = imb * factor; + } + } + } + /* Calculate CPU capacity for physical packages and nodes */ for (i = nr_cpumask_bits-1; i >= 0; i--) { if (!cpumask_test_cpu(i, cpu_map))