From 45c01e824991b2dd0a332e19efc4901acb31209f Mon Sep 17 00:00:00 2001 From: Gregory Haskins Date: Mon, 12 May 2008 21:20:41 +0200 Subject: 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 CC: Dmitry Adamushko CC: Steven Rostedt Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- kernel/sched.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index bfb8ad8ed17..7178b8c2351 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -151,7 +151,8 @@ static inline int task_has_rt_policy(struct task_struct *p) */ struct rt_prio_array { DECLARE_BITMAP(bitmap, MAX_RT_PRIO+1); /* include 1 bit for delimiter */ - struct list_head queue[MAX_RT_PRIO]; + struct list_head xqueue[MAX_RT_PRIO]; /* exclusive queue */ + struct list_head squeue[MAX_RT_PRIO]; /* shared queue */ }; struct rt_bandwidth { @@ -7542,7 +7543,8 @@ static void init_rt_rq(struct rt_rq *rt_rq, struct rq *rq) array = &rt_rq->active; for (i = 0; i < MAX_RT_PRIO; i++) { - INIT_LIST_HEAD(array->queue + i); + INIT_LIST_HEAD(array->xqueue + i); + INIT_LIST_HEAD(array->squeue + i); __clear_bit(i, array->bitmap); } /* delimiter for bitsearch: */ -- cgit v1.2.3 From f333fdc9098b71e2687e4e9b6349fcb352960d66 Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Mon, 12 May 2008 21:20:55 +0200 Subject: sched: make !hrtick faster it is safe to ignore timers and flags when the feature is disabled. Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- kernel/sched.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 7178b8c2351..aa960b84b88 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4134,7 +4134,7 @@ asmlinkage void __sched schedule(void) struct task_struct *prev, *next; unsigned long *switch_count; struct rq *rq; - int cpu; + int cpu, hrtick = sched_feat(HRTICK); need_resched: preempt_disable(); @@ -4149,7 +4149,8 @@ need_resched_nonpreemptible: schedule_debug(prev); - hrtick_clear(rq); + if (hrtick) + hrtick_clear(rq); /* * Do the rq-clock update outside the rq lock: @@ -4197,7 +4198,8 @@ need_resched_nonpreemptible: } else spin_unlock_irq(&rq->lock); - hrtick_set(rq); + if (hrtick) + hrtick_set(rq); if (unlikely(reacquire_kernel_lock(current) < 0)) goto need_resched_nonpreemptible; -- cgit v1.2.3 From 6e0534f278199f1e3dd1049b9bc19a7a5b87ada1 Mon Sep 17 00:00:00 2001 From: Gregory Haskins Date: Mon, 12 May 2008 21:21:01 +0200 Subject: sched: use a 2-d bitmap for searching lowest-pri CPU The current code use a linear algorithm which causes scaling issues on larger SMP machines. This patch replaces that algorithm with a 2-dimensional bitmap to reduce latencies in the wake-up path. Signed-off-by: Gregory Haskins Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- kernel/sched.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index aa960b84b88..8a1257b6556 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -74,6 +74,8 @@ #include #include +#include "sched_cpupri.h" + /* * Convert user-nice values [ -20 ... 0 ... 19 ] * to static priority [ MAX_RT_PRIO..MAX_PRIO-1 ], @@ -450,6 +452,9 @@ struct root_domain { */ cpumask_t rto_mask; atomic_t rto_count; +#ifdef CONFIG_SMP + struct cpupri cpupri; +#endif }; /* @@ -6392,6 +6397,8 @@ static void init_rootdomain(struct root_domain *rd) cpus_clear(rd->span); cpus_clear(rd->online); + + cpupri_init(&rd->cpupri); } static void init_defrootdomain(void) -- cgit v1.2.3 From d07355f5def74d060333563b36ab51b89fd44cdd Mon Sep 17 00:00:00 2001 From: Dmitry Adamushko Date: Mon, 12 May 2008 21:21:15 +0200 Subject: sched: check for SD_SERIALIZE atomically in rebalance_domains() Nothing really serious here, mainly just a matter of nit-picking :-/ From: Dmitry Adamushko For CONFIG_SCHED_DEBUG && CONFIG_SYSCT configs, sd->flags can be altered while being manipulated in rebalance_domains(). Let's do an atomic check. We rely here on the atomicity of read/write accesses for aligned words. Signed-off-by: Dmitry Adamushko Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- kernel/sched.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 8a1257b6556..90329f1f894 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -3668,6 +3668,7 @@ static void rebalance_domains(int cpu, enum cpu_idle_type idle) /* Earliest time when we have to do rebalance again */ unsigned long next_balance = jiffies + 60*HZ; int update_next_balance = 0; + int need_serialize; cpumask_t tmp; for_each_domain(cpu, sd) { @@ -3685,8 +3686,9 @@ static void rebalance_domains(int cpu, enum cpu_idle_type idle) if (interval > HZ*NR_CPUS/10) interval = HZ*NR_CPUS/10; + need_serialize = sd->flags & SD_SERIALIZE; - if (sd->flags & SD_SERIALIZE) { + if (need_serialize) { if (!spin_trylock(&balancing)) goto out; } @@ -3702,7 +3704,7 @@ static void rebalance_domains(int cpu, enum cpu_idle_type idle) } sd->last_balance = jiffies; } - if (sd->flags & SD_SERIALIZE) + if (need_serialize) spin_unlock(&balancing); out: if (time_after(next_balance, sd->last_balance + interval)) { -- cgit v1.2.3 From f7dcd80bbc8e7032443e6539ea1b830364f82200 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sat, 24 May 2008 23:20:38 +0200 Subject: namespacecheck: fixes in kernel/sched.c Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 90329f1f894..02a5eeedcb9 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1633,7 +1633,7 @@ inline int task_curr(const struct task_struct *p) } /* Used instead of source_load when we know the type == 0 */ -unsigned long weighted_cpuload(const int cpu) +static unsigned long weighted_cpuload(const int cpu) { return cpu_rq(cpu)->load.weight; } -- cgit v1.2.3 From 81d41d7ece23a1c3b4bcd1604026d3a06cc4dc79 Mon Sep 17 00:00:00 2001 From: Rabin Vincent Date: Sun, 11 May 2008 05:55:33 +0530 Subject: sched: fix defined-but-unused warning Fix this warning, which appears with !CONFIG_SMP: kernel/sched.c:1216: warning: `init_hrtick' defined but not used Signed-off-by: Rabin Vincent Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- kernel/sched.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 02a5eeedcb9..f3faec52c5a 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1130,6 +1130,7 @@ static enum hrtimer_restart hrtick(struct hrtimer *timer) return HRTIMER_NORESTART; } +#ifdef CONFIG_SMP static void hotplug_hrtick_disable(int cpu) { struct rq *rq = cpu_rq(cpu); @@ -1185,6 +1186,7 @@ static void init_hrtick(void) { hotcpu_notifier(hotplug_hrtick, 0); } +#endif /* CONFIG_SMP */ static void init_rq_hrtick(struct rq *rq) { -- cgit v1.2.3 From e21f5b153b9b4a6775d7d41964e372e13a9178ab Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Fri, 23 May 2008 09:05:58 -0700 Subject: sched: print module list in the "scheduling while atomic" warning For the normal WARN_ON() etc we added a print-the-modules-list already, which is very useful to figure out candidates for certain types of bugs. This patch adds the same print to the "scheduling while atomic" BUG warning, for the same reason: when we get here it's very useful to see which modules are loaded, to narrow down the candidate code list. Signed-off-by: Arjan van de Ven Cc: mingo@elte.hu Signed-off-by: Thomas Gleixner --- kernel/sched.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index f3faec52c5a..84a360670b9 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4070,6 +4070,7 @@ static noinline void __schedule_bug(struct task_struct *prev) prev->comm, prev->pid, preempt_count()); debug_show_held_locks(prev); + print_modules(); if (irqs_disabled()) print_irqtrace_events(prev); -- cgit v1.2.3 From 6d6bc0ad867c46896d0994bb039e7550ecb9b51d Mon Sep 17 00:00:00 2001 From: Dhaval Giani Date: Fri, 30 May 2008 14:23:45 +0200 Subject: sched: add comments for ifdefs in sched.c make sched.c easier to read. Signed-off-by: Dhaval Giani Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- kernel/sched.c | 76 +++++++++++++++++++++++++++++----------------------------- 1 file changed, 38 insertions(+), 38 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 84a360670b9..ef4e25604bb 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -292,15 +292,15 @@ struct task_group root_task_group; static DEFINE_PER_CPU(struct sched_entity, init_sched_entity); /* Default task group's cfs_rq on each cpu */ static DEFINE_PER_CPU(struct cfs_rq, init_cfs_rq) ____cacheline_aligned_in_smp; -#endif +#endif /* CONFIG_FAIR_GROUP_SCHED */ #ifdef CONFIG_RT_GROUP_SCHED static DEFINE_PER_CPU(struct sched_rt_entity, init_sched_rt_entity); static DEFINE_PER_CPU(struct rt_rq, init_rt_rq) ____cacheline_aligned_in_smp; -#endif -#else +#endif /* CONFIG_RT_GROUP_SCHED */ +#else /* !CONFIG_FAIR_GROUP_SCHED */ #define root_task_group init_task_group -#endif +#endif /* CONFIG_FAIR_GROUP_SCHED */ /* task_group_lock serializes add/remove of task groups and also changes to * a task group's cpu shares. @@ -310,9 +310,9 @@ static DEFINE_SPINLOCK(task_group_lock); #ifdef CONFIG_FAIR_GROUP_SCHED #ifdef CONFIG_USER_SCHED # define INIT_TASK_GROUP_LOAD (2*NICE_0_LOAD) -#else +#else /* !CONFIG_USER_SCHED */ # define INIT_TASK_GROUP_LOAD NICE_0_LOAD -#endif +#endif /* CONFIG_USER_SCHED */ /* * A weight of 0, 1 or ULONG_MAX can cause arithmetics problems. @@ -1316,15 +1316,15 @@ void wake_up_idle_cpu(int cpu) if (!tsk_is_polling(rq->idle)) smp_send_reschedule(cpu); } -#endif +#endif /* CONFIG_NO_HZ */ -#else +#else /* !CONFIG_SMP */ static void __resched_task(struct task_struct *p, int tif_bit) { assert_spin_locked(&task_rq(p)->lock); set_tsk_thread_flag(p, tif_bit); } -#endif +#endif /* CONFIG_SMP */ #if BITS_PER_LONG == 32 # define WMULT_CONST (~0UL) @@ -2129,7 +2129,7 @@ static int try_to_wake_up(struct task_struct *p, unsigned int state, int sync) } } } -#endif +#endif /* CONFIG_SCHEDSTATS */ out_activate: #endif /* CONFIG_SMP */ @@ -2329,7 +2329,7 @@ fire_sched_out_preempt_notifiers(struct task_struct *curr, notifier->ops->sched_out(notifier, next); } -#else +#else /* !CONFIG_PREEMPT_NOTIFIERS */ static void fire_sched_in_preempt_notifiers(struct task_struct *curr) { @@ -2341,7 +2341,7 @@ fire_sched_out_preempt_notifiers(struct task_struct *curr, { } -#endif +#endif /* CONFIG_PREEMPT_NOTIFIERS */ /** * prepare_task_switch - prepare to switch tasks @@ -6300,9 +6300,9 @@ static void sched_domain_debug(struct sched_domain *sd, int cpu) } kfree(groupmask); } -#else +#else /* !CONFIG_SCHED_DEBUG */ # define sched_domain_debug(sd, cpu) do { } while (0) -#endif +#endif /* CONFIG_SCHED_DEBUG */ static int sd_degenerate(struct sched_domain *sd) { @@ -6598,7 +6598,7 @@ static void sched_domain_node_span(int node, cpumask_t *span) cpus_or(*span, *span, *nodemask); } } -#endif +#endif /* CONFIG_NUMA */ int sched_smt_power_savings = 0, sched_mc_power_savings = 0; @@ -6617,7 +6617,7 @@ cpu_to_cpu_group(int cpu, const cpumask_t *cpu_map, struct sched_group **sg, *sg = &per_cpu(sched_group_cpus, cpu); return cpu; } -#endif +#endif /* CONFIG_SCHED_SMT */ /* * multi-core sched-domains: @@ -6625,7 +6625,7 @@ cpu_to_cpu_group(int cpu, const cpumask_t *cpu_map, struct sched_group **sg, #ifdef CONFIG_SCHED_MC static DEFINE_PER_CPU(struct sched_domain, core_domains); static DEFINE_PER_CPU(struct sched_group, sched_group_core); -#endif +#endif /* CONFIG_SCHED_MC */ #if defined(CONFIG_SCHED_MC) && defined(CONFIG_SCHED_SMT) static int @@ -6727,7 +6727,7 @@ static void init_numa_sched_groups_power(struct sched_group *group_head) sg = sg->next; } while (sg != group_head); } -#endif +#endif /* CONFIG_NUMA */ #ifdef CONFIG_NUMA /* Free memory allocated for various sched_group structures */ @@ -6764,11 +6764,11 @@ next_sg: sched_group_nodes_bycpu[cpu] = NULL; } } -#else +#else /* !CONFIG_NUMA */ static void free_sched_groups(const cpumask_t *cpu_map, cpumask_t *nodemask) { } -#endif +#endif /* CONFIG_NUMA */ /* * Initialize sched groups cpu_power. @@ -7459,7 +7459,7 @@ int sched_create_sysfs_power_savings_entries(struct sysdev_class *cls) #endif return err; } -#endif +#endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ /* * Force a reinitialization of the sched domains hierarchy. The domains @@ -7677,8 +7677,8 @@ void __init sched_init(void) root_task_group.cfs_rq = (struct cfs_rq **)ptr; ptr += nr_cpu_ids * sizeof(void **); -#endif -#endif +#endif /* CONFIG_USER_SCHED */ +#endif /* CONFIG_FAIR_GROUP_SCHED */ #ifdef CONFIG_RT_GROUP_SCHED init_task_group.rt_se = (struct sched_rt_entity **)ptr; ptr += nr_cpu_ids * sizeof(void **); @@ -7692,8 +7692,8 @@ void __init sched_init(void) root_task_group.rt_rq = (struct rt_rq **)ptr; ptr += nr_cpu_ids * sizeof(void **); -#endif -#endif +#endif /* CONFIG_USER_SCHED */ +#endif /* CONFIG_RT_GROUP_SCHED */ } #ifdef CONFIG_SMP @@ -7709,8 +7709,8 @@ void __init sched_init(void) #ifdef CONFIG_USER_SCHED init_rt_bandwidth(&root_task_group.rt_bandwidth, global_rt_period(), RUNTIME_INF); -#endif -#endif +#endif /* CONFIG_USER_SCHED */ +#endif /* CONFIG_RT_GROUP_SCHED */ #ifdef CONFIG_GROUP_SCHED list_add(&init_task_group.list, &task_groups); @@ -7720,8 +7720,8 @@ void __init sched_init(void) INIT_LIST_HEAD(&root_task_group.children); init_task_group.parent = &root_task_group; list_add(&init_task_group.siblings, &root_task_group.children); -#endif -#endif +#endif /* CONFIG_USER_SCHED */ +#endif /* CONFIG_GROUP_SCHED */ for_each_possible_cpu(i) { struct rq *rq; @@ -8040,7 +8040,7 @@ static inline void unregister_fair_sched_group(struct task_group *tg, int cpu) { list_del_rcu(&tg->cfs_rq[cpu]->leaf_cfs_rq_list); } -#else +#else /* !CONFG_FAIR_GROUP_SCHED */ static inline void free_fair_sched_group(struct task_group *tg) { } @@ -8058,7 +8058,7 @@ static inline void register_fair_sched_group(struct task_group *tg, int cpu) static inline void unregister_fair_sched_group(struct task_group *tg, int cpu) { } -#endif +#endif /* CONFIG_FAIR_GROUP_SCHED */ #ifdef CONFIG_RT_GROUP_SCHED static void free_rt_sched_group(struct task_group *tg) @@ -8129,7 +8129,7 @@ static inline void unregister_rt_sched_group(struct task_group *tg, int cpu) { list_del_rcu(&tg->rt_rq[cpu]->leaf_rt_rq_list); } -#else +#else /* !CONFIG_RT_GROUP_SCHED */ static inline void free_rt_sched_group(struct task_group *tg) { } @@ -8147,7 +8147,7 @@ static inline void register_rt_sched_group(struct task_group *tg, int cpu) static inline void unregister_rt_sched_group(struct task_group *tg, int cpu) { } -#endif +#endif /* CONFIG_RT_GROUP_SCHED */ #ifdef CONFIG_GROUP_SCHED static void free_sched_group(struct task_group *tg) @@ -8258,7 +8258,7 @@ void sched_move_task(struct task_struct *tsk) task_rq_unlock(rq, &flags); } -#endif +#endif /* CONFIG_GROUP_SCHED */ #ifdef CONFIG_FAIR_GROUP_SCHED static void set_se_shares(struct sched_entity *se, unsigned long shares) @@ -8508,7 +8508,7 @@ static int sched_rt_global_constraints(void) return ret; } -#else +#else /* !CONFIG_RT_GROUP_SCHED */ static int sched_rt_global_constraints(void) { unsigned long flags; @@ -8526,7 +8526,7 @@ static int sched_rt_global_constraints(void) return 0; } -#endif +#endif /* CONFIG_RT_GROUP_SCHED */ int sched_rt_handler(struct ctl_table *table, int write, struct file *filp, void __user *buffer, size_t *lenp, @@ -8634,7 +8634,7 @@ static u64 cpu_shares_read_u64(struct cgroup *cgrp, struct cftype *cft) return (u64) tg->shares; } -#endif +#endif /* CONFIG_FAIR_GROUP_SCHED */ #ifdef CONFIG_RT_GROUP_SCHED static int cpu_rt_runtime_write(struct cgroup *cgrp, struct cftype *cft, @@ -8658,7 +8658,7 @@ static u64 cpu_rt_period_read_uint(struct cgroup *cgrp, struct cftype *cft) { return sched_group_rt_period(cgroup_tg(cgrp)); } -#endif +#endif /* CONFIG_RT_GROUP_SCHED */ static struct cftype cpu_files[] = { #ifdef CONFIG_FAIR_GROUP_SCHED -- cgit v1.2.3 From 099f98c8a1f13501a98afbfff4756395a610581c Mon Sep 17 00:00:00 2001 From: Gautham R Shenoy Date: Thu, 29 May 2008 20:56:32 +0530 Subject: sched: print the sd->level in sched_domain_debug code While printing out the visual representation of the sched-domains, print the level (MC, SMT, CPU, NODE, ... ) of each of the sched_domains. Credit: Peter Zijlstra Signed-off-by: Gautham R Shenoy Signed-off-by: Ingo Molnar --- kernel/sched.c | 25 ++++++++++++++++++++++++- 1 file changed, 24 insertions(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index ef4e25604bb..dc0be113f41 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -6197,6 +6197,28 @@ void __init migration_init(void) #ifdef CONFIG_SCHED_DEBUG +static inline const char *sd_level_to_string(enum sched_domain_level lvl) +{ + switch (lvl) { + case SD_LV_NONE: + return "NONE"; + case SD_LV_SIBLING: + return "SIBLING"; + case SD_LV_MC: + return "MC"; + case SD_LV_CPU: + return "CPU"; + case SD_LV_NODE: + return "NODE"; + case SD_LV_ALLNODES: + return "ALLNODES"; + case SD_LV_MAX: + return "MAX"; + + } + return "MAX"; +} + static int sched_domain_debug_one(struct sched_domain *sd, int cpu, int level, cpumask_t *groupmask) { @@ -6216,7 +6238,8 @@ static int sched_domain_debug_one(struct sched_domain *sd, int cpu, int level, return -1; } - printk(KERN_CONT "span %s\n", str); + printk(KERN_CONT "span %s level %s\n", + str, sd_level_to_string(sd->level)); if (!cpu_isset(cpu, sd->span)) { printk(KERN_ERR "ERROR: domain->span does not contain " -- cgit v1.2.3 From 1f11eb6a8bc92536d9e93ead48fa3ffbd1478571 Mon Sep 17 00:00:00 2001 From: Gregory Haskins Date: Wed, 4 Jun 2008 15:04:05 -0400 Subject: sched: fix cpupri hotplug support The RT folks over at RedHat found an issue w.r.t. hotplug support which was traced to problems with the cpupri infrastructure in the scheduler: https://bugzilla.redhat.com/show_bug.cgi?id=449676 This bug affects 23-rt12+, 24-rtX, 25-rtX, and sched-devel. This patch applies to 25.4-rt4, though it should trivially apply to most cpupri enabled kernels mentioned above. It turned out that the issue was that offline cpus could get inadvertently registered with cpupri so that they were erroneously selected during migration decisions. The end result would be an OOPS as the offline cpu had tasks routed to it. This patch generalizes the old join/leave domain interface into an online/offline interface, and adjusts the root-domain/hotplug code to utilize it. I was able to easily reproduce the issue prior to this patch, and am no longer able to reproduce it after this patch. I can offline cpus indefinately and everything seems to be in working order. Thanks to Arnaldo (acme), Thomas, and Peter for doing the legwork to point me in the right direction. Also thank you to Peter for reviewing the early iterations of this patch. Signed-off-by: Gregory Haskins Cc: Peter Zijlstra Cc: Steven Rostedt Cc: Arnaldo Carvalho de Melo Cc: Ingo Molnar Signed-off-by: Thomas Gleixner --- kernel/sched.c | 54 ++++++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 40 insertions(+), 14 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index dc0be113f41..f0ed81b7128 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -529,6 +529,7 @@ struct rq { int push_cpu; /* cpu of this runqueue: */ int cpu; + int online; struct task_struct *migration_thread; struct list_head migration_queue; @@ -1498,6 +1499,8 @@ static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) #endif #define sched_class_highest (&rt_sched_class) +#define for_each_class(class) \ + for (class = sched_class_highest; class; class = class->next) static inline void inc_load(struct rq *rq, const struct task_struct *p) { @@ -6065,6 +6068,36 @@ static void unregister_sched_domain_sysctl(void) } #endif +static void set_rq_online(struct rq *rq) +{ + if (!rq->online) { + const struct sched_class *class; + + cpu_set(rq->cpu, rq->rd->online); + rq->online = 1; + + for_each_class(class) { + if (class->rq_online) + class->rq_online(rq); + } + } +} + +static void set_rq_offline(struct rq *rq) +{ + if (rq->online) { + const struct sched_class *class; + + for_each_class(class) { + if (class->rq_offline) + class->rq_offline(rq); + } + + cpu_clear(rq->cpu, rq->rd->online); + rq->online = 0; + } +} + /* * migration_call - callback that gets triggered when a CPU is added. * Here we can start up the necessary migration thread for the new CPU. @@ -6102,7 +6135,8 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) spin_lock_irqsave(&rq->lock, flags); if (rq->rd) { BUG_ON(!cpu_isset(cpu, rq->rd->span)); - cpu_set(cpu, rq->rd->online); + + set_rq_online(rq); } spin_unlock_irqrestore(&rq->lock, flags); break; @@ -6163,7 +6197,7 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) spin_lock_irqsave(&rq->lock, flags); if (rq->rd) { BUG_ON(!cpu_isset(cpu, rq->rd->span)); - cpu_clear(cpu, rq->rd->online); + set_rq_offline(rq); } spin_unlock_irqrestore(&rq->lock, flags); break; @@ -6385,20 +6419,16 @@ sd_parent_degenerate(struct sched_domain *sd, struct sched_domain *parent) static void rq_attach_root(struct rq *rq, struct root_domain *rd) { unsigned long flags; - const struct sched_class *class; spin_lock_irqsave(&rq->lock, flags); if (rq->rd) { struct root_domain *old_rd = rq->rd; - for (class = sched_class_highest; class; class = class->next) { - if (class->leave_domain) - class->leave_domain(rq); - } + if (cpu_isset(rq->cpu, old_rd->online)) + set_rq_offline(rq); cpu_clear(rq->cpu, old_rd->span); - cpu_clear(rq->cpu, old_rd->online); if (atomic_dec_and_test(&old_rd->refcount)) kfree(old_rd); @@ -6409,12 +6439,7 @@ static void rq_attach_root(struct rq *rq, struct root_domain *rd) cpu_set(rq->cpu, rd->span); if (cpu_isset(rq->cpu, cpu_online_map)) - cpu_set(rq->cpu, rd->online); - - for (class = sched_class_highest; class; class = class->next) { - if (class->join_domain) - class->join_domain(rq); - } + set_rq_online(rq); spin_unlock_irqrestore(&rq->lock, flags); } @@ -7824,6 +7849,7 @@ void __init sched_init(void) rq->next_balance = jiffies; rq->push_cpu = 0; rq->cpu = i; + rq->online = 0; rq->migration_thread = NULL; INIT_LIST_HEAD(&rq->migration_queue); rq_attach_root(rq, &def_root_domain); -- cgit v1.2.3 From 5c8e1ed1d204a6770ca2854cd3b3597070fe7e5a Mon Sep 17 00:00:00 2001 From: Max Krasnyansky Date: Thu, 29 May 2008 11:17:01 -0700 Subject: sched: CPU hotplug events must not destroy scheduler domains created by the cpusets First issue is not related to the cpusets. We're simply leaking doms_cur. It's allocated in arch_init_sched_domains() which is called for every hotplug event. So we just keep reallocation doms_cur without freeing it. I introduced free_sched_domains() function that cleans things up. Second issue is that sched domains created by the cpusets are completely destroyed by the CPU hotplug events. For all CPU hotplug events scheduler attaches all CPUs to the NULL domain and then puts them all into the single domain thereby destroying domains created by the cpusets (partition_sched_domains). The solution is simple, when cpusets are enabled scheduler should not create default domain and instead let cpusets do that. Which is exactly what the patch does. Signed-off-by: Max Krasnyansky Cc: pj@sgi.com Cc: menage@google.com Cc: rostedt@goodmis.org Cc: mingo@elte.hu Acked-by: Peter Zijlstra Signed-off-by: Thomas Gleixner --- kernel/sched.c | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index f0ed81b7128..1ddb0a8c797 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -7292,6 +7292,18 @@ void __attribute__((weak)) arch_update_cpu_topology(void) { } +/* + * Free current domain masks. + * Called after all cpus are attached to NULL domain. + */ +static void free_sched_domains(void) +{ + ndoms_cur = 0; + if (doms_cur != &fallback_doms) + kfree(doms_cur); + doms_cur = &fallback_doms; +} + /* * Set up scheduler domains and groups. Callers must hold the hotplug lock. * For now this just excludes isolated cpus, but could be used to @@ -7439,6 +7451,7 @@ int arch_reinit_sched_domains(void) get_online_cpus(); mutex_lock(&sched_domains_mutex); detach_destroy_domains(&cpu_online_map); + free_sched_domains(); err = arch_init_sched_domains(&cpu_online_map); mutex_unlock(&sched_domains_mutex); put_online_cpus(); @@ -7524,6 +7537,7 @@ static int update_sched_domains(struct notifier_block *nfb, case CPU_DOWN_PREPARE: case CPU_DOWN_PREPARE_FROZEN: detach_destroy_domains(&cpu_online_map); + free_sched_domains(); return NOTIFY_OK; case CPU_UP_CANCELED: @@ -7542,8 +7556,16 @@ static int update_sched_domains(struct notifier_block *nfb, return NOTIFY_DONE; } +#ifndef CONFIG_CPUSETS + /* + * Create default domain partitioning if cpusets are disabled. + * Otherwise we let cpusets rebuild the domains based on the + * current setup. + */ + /* The hotplug lock is already held by cpu_up/cpu_down */ arch_init_sched_domains(&cpu_online_map); +#endif return NOTIFY_OK; } -- cgit v1.2.3 From 68f4f1ec08e3d95730a2693b99df8260aa0d06ae Mon Sep 17 00:00:00 2001 From: Max Krasnyansky Date: Thu, 29 May 2008 11:17:02 -0700 Subject: sched: Move cpu masks from kernel/sched.c into kernel/cpu.c kernel/cpu.c seems a more logical place for those maps since they do not really have much to do with the scheduler these days. kernel/cpu.c is now built for the UP kernel too, but it does not affect the size the kernel sections. $ size vmlinux before text data bss dec hex filename 3313797 307060 310352 3931209 3bfc49 vmlinux after text data bss dec hex filename 3313797 307060 310352 3931209 3bfc49 vmlinux Signed-off-by: Max Krasnyansky Cc: pj@sgi.com Cc: menage@google.com Cc: rostedt@goodmis.org Cc: mingo@elte.hu Acked-by: Peter Zijlstra Signed-off-by: Thomas Gleixner --- kernel/sched.c | 18 ------------------ 1 file changed, 18 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 1ddb0a8c797..f36f549e574 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -5080,24 +5080,6 @@ asmlinkage long sys_sched_setaffinity(pid_t pid, unsigned int len, return sched_setaffinity(pid, &new_mask); } -/* - * Represents all cpu's present in the system - * In systems capable of hotplug, this map could dynamically grow - * as new cpu's are detected in the system via any platform specific - * method, such as ACPI for e.g. - */ - -cpumask_t cpu_present_map __read_mostly; -EXPORT_SYMBOL(cpu_present_map); - -#ifndef CONFIG_SMP -cpumask_t cpu_online_map __read_mostly = CPU_MASK_ALL; -EXPORT_SYMBOL(cpu_online_map); - -cpumask_t cpu_possible_map __read_mostly = CPU_MASK_ALL; -EXPORT_SYMBOL(cpu_possible_map); -#endif - long sched_getaffinity(pid_t pid, cpumask_t *mask) { struct task_struct *p; -- cgit v1.2.3 From e958b3600484533ff801920290468adc8135f89d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 4 Jun 2008 23:22:32 +0200 Subject: sched: move weighted_cpuload into #ifdef CONFIG_SMP section weighted_cpuload is only used on SMP. move it into the CONFIG_SMP section. Signed-off-by: Thomas Gleixner --- kernel/sched.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index f36f549e574..727bdef7616 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1637,12 +1637,6 @@ inline int task_curr(const struct task_struct *p) return cpu_curr(task_cpu(p)) == p; } -/* 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; -} - static inline void __set_task_cpu(struct task_struct *p, unsigned int cpu) { set_task_rq(p, cpu); @@ -1671,6 +1665,12 @@ static inline void check_class_changed(struct rq *rq, struct task_struct *p, #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; +} + /* * Is this task likely cache-hot: */ -- cgit v1.2.3 From 7def2be1dc679984f4c4fb3ef19a8a081b2454ec Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 5 Jun 2008 14:49:58 +0200 Subject: sched: fix hotplug cpus on ia64 Cliff Wickman wrote: > I built an ia64 kernel from Andrew's tree (2.6.26-rc2-mm1) > and get a very predictable hotplug cpu problem. > billberry1:/tmp/cpw # ./dis > disabled cpu 17 > enabled cpu 17 > billberry1:/tmp/cpw # ./dis > disabled cpu 17 > enabled cpu 17 > billberry1:/tmp/cpw # ./dis > > The script that disables the cpu always hangs (unkillable) > on the 3rd attempt. > > And a bit further: > The kstopmachine thread always sits on the run queue (real time) for about > 30 minutes before running. this fix solves some (but not all) issues between CPU hotplug and RT bandwidth throttling. Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 727bdef7616..e9c24a12865 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -7513,21 +7513,28 @@ int sched_create_sysfs_power_savings_entries(struct sysdev_class *cls) static int update_sched_domains(struct notifier_block *nfb, unsigned long action, void *hcpu) { + int cpu = (int)(long)hcpu; + switch (action) { - case CPU_UP_PREPARE: - case CPU_UP_PREPARE_FROZEN: case CPU_DOWN_PREPARE: case CPU_DOWN_PREPARE_FROZEN: + disable_runtime(cpu_rq(cpu)); + /* fall-through */ + case CPU_UP_PREPARE: + case CPU_UP_PREPARE_FROZEN: detach_destroy_domains(&cpu_online_map); free_sched_domains(); return NOTIFY_OK; - case CPU_UP_CANCELED: - case CPU_UP_CANCELED_FROZEN: + case CPU_DOWN_FAILED: case CPU_DOWN_FAILED_FROZEN: case CPU_ONLINE: case CPU_ONLINE_FROZEN: + enable_runtime(cpu_rq(cpu)); + /* fall-through */ + case CPU_UP_CANCELED: + case CPU_UP_CANCELED_FROZEN: case CPU_DEAD: case CPU_DEAD_FROZEN: /* -- cgit v1.2.3 From 9985b0bab332289f14837eff3c6e0bcc658b58f7 Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Thu, 5 Jun 2008 12:57:11 -0700 Subject: sched: prevent bound kthreads from changing cpus_allowed Kthreads that have called kthread_bind() are bound to specific cpus, so other tasks should not be able to change their cpus_allowed from under them. Otherwise, it is possible to move kthreads, such as the migration or software watchdog threads, so they are not allowed access to the cpu they work on. Cc: Peter Zijlstra Cc: Paul Menage Cc: Paul Jackson Signed-off-by: David Rientjes Signed-off-by: Ingo Molnar --- kernel/sched.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index e9c24a12865..164fe7fe0d8 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -5563,6 +5563,12 @@ int set_cpus_allowed_ptr(struct task_struct *p, const cpumask_t *new_mask) goto out; } + if (unlikely((p->flags & PF_THREAD_BOUND) && p != current && + !cpus_equal(p->cpus_allowed, *new_mask))) { + ret = -EINVAL; + goto out; + } + if (p->sched_class->set_cpus_allowed) p->sched_class->set_cpus_allowed(p, new_mask); else { -- cgit v1.2.3 From e9886ca3a93d7d041d3de8e5acebe213da777d59 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 9 Jun 2008 17:12:24 +0900 Subject: sched: kill off dead cfs_rq_set_shares() Building with CONFIG_FAIR_GROUP_SCHED=y on UP results in an unused cfs_rq_set_shares() reference. As nothing is using this dummy function in the first place, just kill it off. Signed-off-by: Paul Mundt Signed-off-by: Ingo Molnar --- kernel/sched.c | 8 -------- 1 file changed, 8 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 164fe7fe0d8..07d5472dee9 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1480,16 +1480,8 @@ static unsigned long source_load(int cpu, int type); static unsigned long target_load(int cpu, int type); static unsigned long cpu_avg_load_per_task(int cpu); static int task_hot(struct task_struct *p, u64 now, struct sched_domain *sd); -#else /* CONFIG_SMP */ - -#ifdef CONFIG_FAIR_GROUP_SCHED -static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) -{ -} #endif -#endif /* CONFIG_SMP */ - #include "sched_stats.h" #include "sched_idletask.c" #include "sched_fair.c" -- cgit v1.2.3 From 20b6331bfed1f07ba1e5006889a5d64adc53615e Mon Sep 17 00:00:00 2001 From: Dmitry Adamushko Date: Wed, 11 Jun 2008 00:58:30 +0200 Subject: 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 Signed-off-by: Ingo Molnar --- kernel/sched.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 554de400980..cc1d558406f 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -153,8 +153,7 @@ static inline int task_has_rt_policy(struct task_struct *p) */ struct rt_prio_array { DECLARE_BITMAP(bitmap, MAX_RT_PRIO+1); /* include 1 bit for delimiter */ - struct list_head xqueue[MAX_RT_PRIO]; /* exclusive queue */ - struct list_head squeue[MAX_RT_PRIO]; /* shared queue */ + struct list_head queue[MAX_RT_PRIO]; }; struct rt_bandwidth { @@ -7620,8 +7619,7 @@ static void init_rt_rq(struct rt_rq *rt_rq, struct rq *rq) array = &rt_rq->active; for (i = 0; i < MAX_RT_PRIO; i++) { - INIT_LIST_HEAD(array->xqueue + i); - INIT_LIST_HEAD(array->squeue + i); + INIT_LIST_HEAD(array->queue + i); __clear_bit(i, array->bitmap); } /* delimiter for bitsearch: */ -- cgit v1.2.3 From 10b612f440a22a294e87ec7e8f03f9eea3338628 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 19 Jun 2008 14:22:27 +0200 Subject: sched: rt: fix the bandwidth contraint computations Signed-off-by: Peter Zijlstra Cc: "Daniel K." Signed-off-by: Ingo Molnar --- kernel/sched.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 5b307da827e..1f711a58a2b 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -8419,7 +8419,7 @@ static unsigned long to_ratio(u64 period, u64 runtime) #ifdef CONFIG_CGROUP_SCHED static int __rt_schedulable(struct task_group *tg, u64 period, u64 runtime) { - struct task_group *tgi, *parent = tg ? tg->parent : NULL; + struct task_group *tgi, *parent = tg->parent; unsigned long total = 0; if (!parent) { @@ -8443,7 +8443,7 @@ static int __rt_schedulable(struct task_group *tg, u64 period, u64 runtime) } rcu_read_unlock(); - return total + to_ratio(period, runtime) < + return total + to_ratio(period, runtime) <= to_ratio(ktime_to_ns(parent->rt_bandwidth.rt_period), parent->rt_bandwidth.rt_runtime); } @@ -8560,10 +8560,15 @@ long sched_group_rt_period(struct task_group *tg) static int sched_rt_global_constraints(void) { + struct task_group *tg = &root_task_group; + u64 rt_runtime, rt_period; int ret = 0; + rt_period = ktime_to_ns(tg->rt_bandwidth.rt_period); + rt_runtime = tg->rt_bandwidth.rt_runtime; + mutex_lock(&rt_constraints_mutex); - if (!__rt_schedulable(NULL, 1, 0)) + if (!__rt_schedulable(tg, rt_period, rt_runtime)) ret = -EINVAL; mutex_unlock(&rt_constraints_mutex); -- cgit v1.2.3 From a7be37ac8e1565e00880531f4e2aff421a21c803 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:11 +0200 Subject: sched: revert the revert of: weight calculations Try again.. initial commit: 8f1bc385cfbab474db6c27b5af1e439614f3025c revert: f9305d4a0968201b2818dbed0dc8cb0d4ee7aeb3 Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index c51d9fae8cd..f653af684fb 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1342,6 +1342,9 @@ static void __resched_task(struct task_struct *p, int tif_bit) */ #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) @@ -1369,12 +1372,6 @@ calc_delta_mine(unsigned long delta_exec, unsigned long weight, return (unsigned long)min(tmp, (u64)(unsigned long)LONG_MAX); } -static inline unsigned long -calc_delta_fair(unsigned long delta_exec, struct load_weight *lw) -{ - return calc_delta_mine(delta_exec, NICE_0_LOAD, lw); -} - static inline void update_load_add(struct load_weight *lw, unsigned long inc) { lw->weight += inc; -- cgit v1.2.3 From c09595f63bb1909c5dc4dca288f4fe818561b5f3 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:14 +0200 Subject: sched: revert revert of: fair-group: SMP-nice for group scheduling Try again.. Initial commit: 18d95a2832c1392a2d63227a7a6d433cb9f2037e Revert: 6363ca57c76b7b83639ca8c83fc285fa26a7880e Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 430 ++++++++++++++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 399 insertions(+), 31 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index f653af684fb..874b6da1543 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -403,6 +403,43 @@ struct cfs_rq { */ struct list_head leaf_cfs_rq_list; struct task_group *tg; /* group that "owns" this runqueue */ + +#ifdef CONFIG_SMP + unsigned long task_weight; + unsigned long shares; + /* + * We need space to build a sched_domain wide view of the full task + * group tree, in order to avoid depending on dynamic memory allocation + * during the load balancing we place this in the per cpu task group + * hierarchy. This limits the load balancing to one instance per cpu, + * but more should not be needed anyway. + */ + struct aggregate_struct { + /* + * load = weight(cpus) * f(tg) + * + * Where f(tg) is the recursive weight fraction assigned to + * this group. + */ + unsigned long load; + + /* + * part of the group weight distributed to this span. + */ + unsigned long shares; + + /* + * The sum of all runqueue weights within this span. + */ + unsigned long rq_weight; + + /* + * Weight contributed by tasks; this is the part we can + * influence by moving tasks around. + */ + unsigned long task_weight; + } aggregate; +#endif #endif }; @@ -1484,6 +1521,326 @@ static unsigned long source_load(int cpu, int type); static unsigned long target_load(int cpu, int type); static unsigned long cpu_avg_load_per_task(int cpu); static int task_hot(struct task_struct *p, u64 now, struct sched_domain *sd); + +#ifdef CONFIG_FAIR_GROUP_SCHED + +/* + * Group load balancing. + * + * We calculate a few balance domain wide aggregate numbers; load and weight. + * Given the pictures below, and assuming each item has equal weight: + * + * root 1 - thread + * / | \ A - group + * A 1 B + * /|\ / \ + * C 2 D 3 4 + * | | + * 5 6 + * + * load: + * A and B get 1/3-rd of the total load. C and D get 1/3-rd of A's 1/3-rd, + * which equals 1/9-th of the total load. + * + * shares: + * The weight of this group on the selected cpus. + * + * rq_weight: + * Direct sum of all the cpu's their rq weight, e.g. A would get 3 while + * B would get 2. + * + * task_weight: + * Part of the rq_weight contributed by tasks; all groups except B would + * get 1, B gets 2. + */ + +static inline struct aggregate_struct * +aggregate(struct task_group *tg, struct sched_domain *sd) +{ + return &tg->cfs_rq[sd->first_cpu]->aggregate; +} + +typedef void (*aggregate_func)(struct task_group *, struct sched_domain *); + +/* + * Iterate the full tree, calling @down when first entering a node and @up when + * leaving it for the final time. + */ +static +void aggregate_walk_tree(aggregate_func down, aggregate_func up, + struct sched_domain *sd) +{ + struct task_group *parent, *child; + + rcu_read_lock(); + parent = &root_task_group; +down: + (*down)(parent, sd); + list_for_each_entry_rcu(child, &parent->children, siblings) { + parent = child; + goto down; + +up: + continue; + } + (*up)(parent, sd); + + child = parent; + parent = parent->parent; + if (parent) + goto up; + rcu_read_unlock(); +} + +/* + * Calculate the aggregate runqueue weight. + */ +static +void aggregate_group_weight(struct task_group *tg, struct sched_domain *sd) +{ + unsigned long rq_weight = 0; + unsigned long task_weight = 0; + int i; + + for_each_cpu_mask(i, sd->span) { + rq_weight += tg->cfs_rq[i]->load.weight; + task_weight += tg->cfs_rq[i]->task_weight; + } + + aggregate(tg, sd)->rq_weight = rq_weight; + aggregate(tg, sd)->task_weight = task_weight; +} + +/* + * Compute the weight of this group on the given cpus. + */ +static +void aggregate_group_shares(struct task_group *tg, struct sched_domain *sd) +{ + unsigned long shares = 0; + int i; + + for_each_cpu_mask(i, sd->span) + shares += tg->cfs_rq[i]->shares; + + if ((!shares && aggregate(tg, sd)->rq_weight) || shares > tg->shares) + shares = tg->shares; + + aggregate(tg, sd)->shares = shares; +} + +/* + * Compute the load fraction assigned to this group, relies on the aggregate + * weight and this group's parent's load, i.e. top-down. + */ +static +void aggregate_group_load(struct task_group *tg, struct sched_domain *sd) +{ + unsigned long load; + + if (!tg->parent) { + int i; + + load = 0; + for_each_cpu_mask(i, sd->span) + load += cpu_rq(i)->load.weight; + + } else { + load = aggregate(tg->parent, sd)->load; + + /* + * shares is our weight in the parent's rq so + * shares/parent->rq_weight gives our fraction of the load + */ + load *= aggregate(tg, sd)->shares; + load /= aggregate(tg->parent, sd)->rq_weight + 1; + } + + aggregate(tg, sd)->load = load; +} + +static void __set_se_shares(struct sched_entity *se, unsigned long shares); + +/* + * Calculate and set the cpu's group shares. + */ +static void +__update_group_shares_cpu(struct task_group *tg, struct sched_domain *sd, + int tcpu) +{ + int boost = 0; + unsigned long shares; + unsigned long rq_weight; + + if (!tg->se[tcpu]) + return; + + rq_weight = tg->cfs_rq[tcpu]->load.weight; + + /* + * If there are currently no tasks on the cpu pretend there is one of + * average load so that when a new task gets to run here it will not + * get delayed by group starvation. + */ + if (!rq_weight) { + boost = 1; + rq_weight = NICE_0_LOAD; + } + + /* + * \Sum shares * rq_weight + * shares = ----------------------- + * \Sum rq_weight + * + */ + shares = aggregate(tg, sd)->shares * rq_weight; + shares /= aggregate(tg, sd)->rq_weight + 1; + + /* + * record the actual number of shares, not the boosted amount. + */ + tg->cfs_rq[tcpu]->shares = boost ? 0 : shares; + + if (shares < MIN_SHARES) + shares = MIN_SHARES; + else if (shares > MAX_SHARES) + shares = MAX_SHARES; + + __set_se_shares(tg->se[tcpu], shares); +} + +/* + * Re-adjust the weights on the cpu the task came from and on the cpu the + * task went to. + */ +static void +__move_group_shares(struct task_group *tg, struct sched_domain *sd, + int scpu, int dcpu) +{ + unsigned long shares; + + shares = tg->cfs_rq[scpu]->shares + tg->cfs_rq[dcpu]->shares; + + __update_group_shares_cpu(tg, sd, scpu); + __update_group_shares_cpu(tg, sd, dcpu); + + /* + * ensure we never loose shares due to rounding errors in the + * above redistribution. + */ + shares -= tg->cfs_rq[scpu]->shares + tg->cfs_rq[dcpu]->shares; + if (shares) + tg->cfs_rq[dcpu]->shares += shares; +} + +/* + * Because changing a group's shares changes the weight of the super-group + * we need to walk up the tree and change all shares until we hit the root. + */ +static void +move_group_shares(struct task_group *tg, struct sched_domain *sd, + int scpu, int dcpu) +{ + while (tg) { + __move_group_shares(tg, sd, scpu, dcpu); + tg = tg->parent; + } +} + +static +void aggregate_group_set_shares(struct task_group *tg, struct sched_domain *sd) +{ + unsigned long shares = aggregate(tg, sd)->shares; + int i; + + for_each_cpu_mask(i, sd->span) { + struct rq *rq = cpu_rq(i); + unsigned long flags; + + spin_lock_irqsave(&rq->lock, flags); + __update_group_shares_cpu(tg, sd, i); + spin_unlock_irqrestore(&rq->lock, flags); + } + + aggregate_group_shares(tg, sd); + + /* + * ensure we never loose shares due to rounding errors in the + * above redistribution. + */ + shares -= aggregate(tg, sd)->shares; + if (shares) { + tg->cfs_rq[sd->first_cpu]->shares += shares; + aggregate(tg, sd)->shares += shares; + } +} + +/* + * Calculate the accumulative weight and recursive load of each task group + * while walking down the tree. + */ +static +void aggregate_get_down(struct task_group *tg, struct sched_domain *sd) +{ + aggregate_group_weight(tg, sd); + aggregate_group_shares(tg, sd); + aggregate_group_load(tg, sd); +} + +/* + * Rebalance the cpu shares while walking back up the tree. + */ +static +void aggregate_get_up(struct task_group *tg, struct sched_domain *sd) +{ + aggregate_group_set_shares(tg, sd); +} + +static DEFINE_PER_CPU(spinlock_t, aggregate_lock); + +static void __init init_aggregate(void) +{ + int i; + + for_each_possible_cpu(i) + spin_lock_init(&per_cpu(aggregate_lock, i)); +} + +static int get_aggregate(struct sched_domain *sd) +{ + if (!spin_trylock(&per_cpu(aggregate_lock, sd->first_cpu))) + return 0; + + aggregate_walk_tree(aggregate_get_down, aggregate_get_up, sd); + return 1; +} + +static void put_aggregate(struct sched_domain *sd) +{ + spin_unlock(&per_cpu(aggregate_lock, sd->first_cpu)); +} + +static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) +{ + cfs_rq->shares = shares; +} + +#else + +static inline void init_aggregate(void) +{ +} + +static inline int get_aggregate(struct sched_domain *sd) +{ + return 0; +} + +static inline void put_aggregate(struct sched_domain *sd) +{ +} +#endif + #endif #include "sched_stats.h" @@ -1498,26 +1855,14 @@ static int task_hot(struct task_struct *p, u64 now, struct sched_domain *sd); #define for_each_class(class) \ for (class = sched_class_highest; class; class = class->next) -static inline void inc_load(struct rq *rq, const struct task_struct *p) -{ - update_load_add(&rq->load, p->se.load.weight); -} - -static inline void dec_load(struct rq *rq, const struct task_struct *p) -{ - update_load_sub(&rq->load, p->se.load.weight); -} - -static void inc_nr_running(struct task_struct *p, struct rq *rq) +static void inc_nr_running(struct rq *rq) { rq->nr_running++; - inc_load(rq, p); } -static void dec_nr_running(struct task_struct *p, struct rq *rq) +static void dec_nr_running(struct rq *rq) { rq->nr_running--; - dec_load(rq, p); } static void set_load_weight(struct task_struct *p) @@ -1609,7 +1954,7 @@ static void activate_task(struct rq *rq, struct task_struct *p, int wakeup) rq->nr_uninterruptible--; enqueue_task(rq, p, wakeup); - inc_nr_running(p, rq); + inc_nr_running(rq); } /* @@ -1621,7 +1966,7 @@ static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep) rq->nr_uninterruptible++; dequeue_task(rq, p, sleep); - dec_nr_running(p, rq); + dec_nr_running(rq); } /** @@ -2274,7 +2619,7 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) * management (if any): */ p->sched_class->task_new(rq, p); - inc_nr_running(p, rq); + inc_nr_running(rq); } check_preempt_curr(rq, p); #ifdef CONFIG_SMP @@ -3265,9 +3610,12 @@ static int load_balance(int this_cpu, struct rq *this_rq, unsigned long imbalance; struct rq *busiest; unsigned long flags; + int unlock_aggregate; cpus_setall(*cpus); + unlock_aggregate = get_aggregate(sd); + /* * When power savings policy is enabled for the parent domain, idle * sibling can pick up load irrespective of busy siblings. In this case, @@ -3383,8 +3731,9 @@ redo: if (!ld_moved && !sd_idle && sd->flags & SD_SHARE_CPUPOWER && !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - return ld_moved; + ld_moved = -1; + + goto out; out_balanced: schedstat_inc(sd, lb_balanced[idle]); @@ -3399,8 +3748,13 @@ out_one_pinned: if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - return 0; + ld_moved = -1; + else + ld_moved = 0; +out: + if (unlock_aggregate) + put_aggregate(sd); + return ld_moved; } /* @@ -4588,10 +4942,8 @@ void set_user_nice(struct task_struct *p, long nice) goto out_unlock; } on_rq = p->se.on_rq; - if (on_rq) { + if (on_rq) dequeue_task(rq, p, 0); - dec_load(rq, p); - } p->static_prio = NICE_TO_PRIO(nice); set_load_weight(p); @@ -4601,7 +4953,6 @@ void set_user_nice(struct task_struct *p, long nice) if (on_rq) { enqueue_task(rq, p, 0); - inc_load(rq, p); /* * If the task increased its priority or is running and * lowered its priority, then reschedule its CPU: @@ -7016,6 +7367,7 @@ static int __build_sched_domains(const cpumask_t *cpu_map, SD_INIT(sd, ALLNODES); set_domain_attribute(sd, attr); sd->span = *cpu_map; + sd->first_cpu = first_cpu(sd->span); cpu_to_allnodes_group(i, cpu_map, &sd->groups, tmpmask); p = sd; sd_allnodes = 1; @@ -7026,6 +7378,7 @@ static int __build_sched_domains(const cpumask_t *cpu_map, SD_INIT(sd, NODE); set_domain_attribute(sd, attr); sched_domain_node_span(cpu_to_node(i), &sd->span); + sd->first_cpu = first_cpu(sd->span); sd->parent = p; if (p) p->child = sd; @@ -7037,6 +7390,7 @@ static int __build_sched_domains(const cpumask_t *cpu_map, SD_INIT(sd, CPU); set_domain_attribute(sd, attr); sd->span = *nodemask; + sd->first_cpu = first_cpu(sd->span); sd->parent = p; if (p) p->child = sd; @@ -7048,6 +7402,7 @@ static int __build_sched_domains(const cpumask_t *cpu_map, SD_INIT(sd, MC); set_domain_attribute(sd, attr); sd->span = cpu_coregroup_map(i); + sd->first_cpu = first_cpu(sd->span); cpus_and(sd->span, sd->span, *cpu_map); sd->parent = p; p->child = sd; @@ -7060,6 +7415,7 @@ static int __build_sched_domains(const cpumask_t *cpu_map, SD_INIT(sd, SIBLING); set_domain_attribute(sd, attr); sd->span = per_cpu(cpu_sibling_map, i); + sd->first_cpu = first_cpu(sd->span); cpus_and(sd->span, sd->span, *cpu_map); sd->parent = p; p->child = sd; @@ -7757,6 +8113,7 @@ void __init sched_init(void) } #ifdef CONFIG_SMP + init_aggregate(); init_defrootdomain(); #endif @@ -8322,14 +8679,11 @@ void sched_move_task(struct task_struct *tsk) #endif /* CONFIG_GROUP_SCHED */ #ifdef CONFIG_FAIR_GROUP_SCHED -static void set_se_shares(struct sched_entity *se, unsigned long shares) +static void __set_se_shares(struct sched_entity *se, unsigned long shares) { struct cfs_rq *cfs_rq = se->cfs_rq; - struct rq *rq = cfs_rq->rq; int on_rq; - spin_lock_irq(&rq->lock); - on_rq = se->on_rq; if (on_rq) dequeue_entity(cfs_rq, se, 0); @@ -8339,8 +8693,17 @@ static void set_se_shares(struct sched_entity *se, unsigned long shares) if (on_rq) enqueue_entity(cfs_rq, se, 0); +} - spin_unlock_irq(&rq->lock); +static void set_se_shares(struct sched_entity *se, unsigned long shares) +{ + struct cfs_rq *cfs_rq = se->cfs_rq; + struct rq *rq = cfs_rq->rq; + unsigned long flags; + + spin_lock_irqsave(&rq->lock, flags); + __set_se_shares(se, shares); + spin_unlock_irqrestore(&rq->lock, flags); } static DEFINE_MUTEX(shares_mutex); @@ -8379,8 +8742,13 @@ int sched_group_set_shares(struct task_group *tg, unsigned long shares) * w/o tripping rebalance_share or load_balance_fair. */ tg->shares = shares; - for_each_possible_cpu(i) + for_each_possible_cpu(i) { + /* + * force a rebalance + */ + cfs_rq_set_shares(tg->cfs_rq[i], 0); set_se_shares(tg->se[i], shares); + } /* * Enable load balance activity on this group, by inserting it back on -- cgit v1.2.3 From 76a2a6ee8a0660a29127f05989ac59ae1ce865fa Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:15 +0200 Subject: sched: sched_clock_cpu() based cpu_clock() with sched_clock_cpu() being reasonably in sync between cpus (max 1 jiffy difference) use this to provide cpu_clock(). Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 76 ---------------------------------------------------------- 1 file changed, 76 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 874b6da1543..eb3454c410f 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -818,82 +818,6 @@ static inline u64 global_rt_runtime(void) return (u64)sysctl_sched_rt_runtime * NSEC_PER_USEC; } -unsigned long long time_sync_thresh = 100000; - -static DEFINE_PER_CPU(unsigned long long, time_offset); -static DEFINE_PER_CPU(unsigned long long, prev_cpu_time); - -/* - * Global lock which we take every now and then to synchronize - * the CPUs time. This method is not warp-safe, but it's good - * enough to synchronize slowly diverging time sources and thus - * it's good enough for tracing: - */ -static DEFINE_SPINLOCK(time_sync_lock); -static unsigned long long prev_global_time; - -static unsigned long long __sync_cpu_clock(unsigned long long time, int cpu) -{ - /* - * We want this inlined, to not get tracer function calls - * in this critical section: - */ - spin_acquire(&time_sync_lock.dep_map, 0, 0, _THIS_IP_); - __raw_spin_lock(&time_sync_lock.raw_lock); - - if (time < prev_global_time) { - per_cpu(time_offset, cpu) += prev_global_time - time; - time = prev_global_time; - } else { - prev_global_time = time; - } - - __raw_spin_unlock(&time_sync_lock.raw_lock); - spin_release(&time_sync_lock.dep_map, 1, _THIS_IP_); - - return time; -} - -static unsigned long long __cpu_clock(int cpu) -{ - unsigned long long now; - - /* - * Only call sched_clock() if the scheduler has already been - * initialized (some code might call cpu_clock() very early): - */ - if (unlikely(!scheduler_running)) - return 0; - - now = sched_clock_cpu(cpu); - - return now; -} - -/* - * For kernel-internal use: high-speed (but slightly incorrect) per-cpu - * clock constructed from sched_clock(): - */ -unsigned long long cpu_clock(int cpu) -{ - unsigned long long prev_cpu_time, time, delta_time; - unsigned long flags; - - local_irq_save(flags); - prev_cpu_time = per_cpu(prev_cpu_time, cpu); - time = __cpu_clock(cpu) + per_cpu(time_offset, cpu); - delta_time = time-prev_cpu_time; - - if (unlikely(delta_time > time_sync_thresh)) { - time = __sync_cpu_clock(time, cpu); - per_cpu(prev_cpu_time, cpu) = time; - } - local_irq_restore(flags); - - return time; -} -EXPORT_SYMBOL_GPL(cpu_clock); - #ifndef prepare_arch_switch # define prepare_arch_switch(next) do { } while (0) #endif -- cgit v1.2.3 From 103638d95ba5b0c53c8d9c0cb581156ccc8513ee Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:16 +0200 Subject: sched: fix wakeup granularity and buddy granularity Uncouple buddy selection from wakeup granularity. The initial idea was that buddies could run ahead as far as a normal task can - do this by measuring a pair 'slice' just as we do for a normal task. This means we can drop the wakeup_granularity back to 5ms. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index eb3454c410f..7d282c52bd4 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -375,6 +375,7 @@ struct cfs_rq { u64 exec_clock; u64 min_vruntime; + u64 pair_start; struct rb_root tasks_timeline; struct rb_node *rb_leftmost; -- cgit v1.2.3 From b6a86c746f5b708012809958462234d19e9c8177 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:18 +0200 Subject: sched: fix sched_domain aggregation Keeping the aggregate on the first cpu of the sched domain has two problems: - it could collide between different sched domains on different cpus - it could slow things down because of the remote accesses Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 113 +++++++++++++++++++++++++++------------------------------ 1 file changed, 54 insertions(+), 59 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 7d282c52bd4..160d3c209b8 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1480,12 +1480,12 @@ static int task_hot(struct task_struct *p, u64 now, struct sched_domain *sd); */ static inline struct aggregate_struct * -aggregate(struct task_group *tg, struct sched_domain *sd) +aggregate(struct task_group *tg, int cpu) { - return &tg->cfs_rq[sd->first_cpu]->aggregate; + return &tg->cfs_rq[cpu]->aggregate; } -typedef void (*aggregate_func)(struct task_group *, struct sched_domain *); +typedef void (*aggregate_func)(struct task_group *, int, struct sched_domain *); /* * Iterate the full tree, calling @down when first entering a node and @up when @@ -1493,14 +1493,14 @@ typedef void (*aggregate_func)(struct task_group *, struct sched_domain *); */ static void aggregate_walk_tree(aggregate_func down, aggregate_func up, - struct sched_domain *sd) + int cpu, struct sched_domain *sd) { struct task_group *parent, *child; rcu_read_lock(); parent = &root_task_group; down: - (*down)(parent, sd); + (*down)(parent, cpu, sd); list_for_each_entry_rcu(child, &parent->children, siblings) { parent = child; goto down; @@ -1508,7 +1508,7 @@ down: up: continue; } - (*up)(parent, sd); + (*up)(parent, cpu, sd); child = parent; parent = parent->parent; @@ -1520,8 +1520,8 @@ up: /* * Calculate the aggregate runqueue weight. */ -static -void aggregate_group_weight(struct task_group *tg, struct sched_domain *sd) +static void +aggregate_group_weight(struct task_group *tg, int cpu, struct sched_domain *sd) { unsigned long rq_weight = 0; unsigned long task_weight = 0; @@ -1532,15 +1532,15 @@ void aggregate_group_weight(struct task_group *tg, struct sched_domain *sd) task_weight += tg->cfs_rq[i]->task_weight; } - aggregate(tg, sd)->rq_weight = rq_weight; - aggregate(tg, sd)->task_weight = task_weight; + aggregate(tg, cpu)->rq_weight = rq_weight; + aggregate(tg, cpu)->task_weight = task_weight; } /* * Compute the weight of this group on the given cpus. */ -static -void aggregate_group_shares(struct task_group *tg, struct sched_domain *sd) +static void +aggregate_group_shares(struct task_group *tg, int cpu, struct sched_domain *sd) { unsigned long shares = 0; int i; @@ -1548,18 +1548,18 @@ void aggregate_group_shares(struct task_group *tg, struct sched_domain *sd) for_each_cpu_mask(i, sd->span) shares += tg->cfs_rq[i]->shares; - if ((!shares && aggregate(tg, sd)->rq_weight) || shares > tg->shares) + if ((!shares && aggregate(tg, cpu)->rq_weight) || shares > tg->shares) shares = tg->shares; - aggregate(tg, sd)->shares = shares; + aggregate(tg, cpu)->shares = shares; } /* * Compute the load fraction assigned to this group, relies on the aggregate * weight and this group's parent's load, i.e. top-down. */ -static -void aggregate_group_load(struct task_group *tg, struct sched_domain *sd) +static void +aggregate_group_load(struct task_group *tg, int cpu, struct sched_domain *sd) { unsigned long load; @@ -1571,17 +1571,17 @@ void aggregate_group_load(struct task_group *tg, struct sched_domain *sd) load += cpu_rq(i)->load.weight; } else { - load = aggregate(tg->parent, sd)->load; + load = aggregate(tg->parent, cpu)->load; /* * shares is our weight in the parent's rq so * shares/parent->rq_weight gives our fraction of the load */ - load *= aggregate(tg, sd)->shares; - load /= aggregate(tg->parent, sd)->rq_weight + 1; + load *= aggregate(tg, cpu)->shares; + load /= aggregate(tg->parent, cpu)->rq_weight + 1; } - aggregate(tg, sd)->load = load; + aggregate(tg, cpu)->load = load; } static void __set_se_shares(struct sched_entity *se, unsigned long shares); @@ -1590,8 +1590,8 @@ static void __set_se_shares(struct sched_entity *se, unsigned long shares); * Calculate and set the cpu's group shares. */ static void -__update_group_shares_cpu(struct task_group *tg, struct sched_domain *sd, - int tcpu) +__update_group_shares_cpu(struct task_group *tg, int cpu, + struct sched_domain *sd, int tcpu) { int boost = 0; unsigned long shares; @@ -1618,8 +1618,8 @@ __update_group_shares_cpu(struct task_group *tg, struct sched_domain *sd, * \Sum rq_weight * */ - shares = aggregate(tg, sd)->shares * rq_weight; - shares /= aggregate(tg, sd)->rq_weight + 1; + shares = aggregate(tg, cpu)->shares * rq_weight; + shares /= aggregate(tg, cpu)->rq_weight + 1; /* * record the actual number of shares, not the boosted amount. @@ -1639,15 +1639,15 @@ __update_group_shares_cpu(struct task_group *tg, struct sched_domain *sd, * task went to. */ static void -__move_group_shares(struct task_group *tg, struct sched_domain *sd, +__move_group_shares(struct task_group *tg, int cpu, struct sched_domain *sd, int scpu, int dcpu) { unsigned long shares; shares = tg->cfs_rq[scpu]->shares + tg->cfs_rq[dcpu]->shares; - __update_group_shares_cpu(tg, sd, scpu); - __update_group_shares_cpu(tg, sd, dcpu); + __update_group_shares_cpu(tg, cpu, sd, scpu); + __update_group_shares_cpu(tg, cpu, sd, dcpu); /* * ensure we never loose shares due to rounding errors in the @@ -1663,19 +1663,19 @@ __move_group_shares(struct task_group *tg, struct sched_domain *sd, * we need to walk up the tree and change all shares until we hit the root. */ static void -move_group_shares(struct task_group *tg, struct sched_domain *sd, +move_group_shares(struct task_group *tg, int cpu, struct sched_domain *sd, int scpu, int dcpu) { while (tg) { - __move_group_shares(tg, sd, scpu, dcpu); + __move_group_shares(tg, cpu, sd, scpu, dcpu); tg = tg->parent; } } -static -void aggregate_group_set_shares(struct task_group *tg, struct sched_domain *sd) +static void +aggregate_group_set_shares(struct task_group *tg, int cpu, struct sched_domain *sd) { - unsigned long shares = aggregate(tg, sd)->shares; + unsigned long shares = aggregate(tg, cpu)->shares; int i; for_each_cpu_mask(i, sd->span) { @@ -1683,20 +1683,20 @@ void aggregate_group_set_shares(struct task_group *tg, struct sched_domain *sd) unsigned long flags; spin_lock_irqsave(&rq->lock, flags); - __update_group_shares_cpu(tg, sd, i); + __update_group_shares_cpu(tg, cpu, sd, i); spin_unlock_irqrestore(&rq->lock, flags); } - aggregate_group_shares(tg, sd); + aggregate_group_shares(tg, cpu, sd); /* * ensure we never loose shares due to rounding errors in the * above redistribution. */ - shares -= aggregate(tg, sd)->shares; + shares -= aggregate(tg, cpu)->shares; if (shares) { - tg->cfs_rq[sd->first_cpu]->shares += shares; - aggregate(tg, sd)->shares += shares; + tg->cfs_rq[cpu]->shares += shares; + aggregate(tg, cpu)->shares += shares; } } @@ -1704,21 +1704,21 @@ void aggregate_group_set_shares(struct task_group *tg, struct sched_domain *sd) * Calculate the accumulative weight and recursive load of each task group * while walking down the tree. */ -static -void aggregate_get_down(struct task_group *tg, struct sched_domain *sd) +static void +aggregate_get_down(struct task_group *tg, int cpu, struct sched_domain *sd) { - aggregate_group_weight(tg, sd); - aggregate_group_shares(tg, sd); - aggregate_group_load(tg, sd); + aggregate_group_weight(tg, cpu, sd); + aggregate_group_shares(tg, cpu, sd); + aggregate_group_load(tg, cpu, sd); } /* * Rebalance the cpu shares while walking back up the tree. */ -static -void aggregate_get_up(struct task_group *tg, struct sched_domain *sd) +static void +aggregate_get_up(struct task_group *tg, int cpu, struct sched_domain *sd) { - aggregate_group_set_shares(tg, sd); + aggregate_group_set_shares(tg, cpu, sd); } static DEFINE_PER_CPU(spinlock_t, aggregate_lock); @@ -1731,18 +1731,18 @@ static void __init init_aggregate(void) spin_lock_init(&per_cpu(aggregate_lock, i)); } -static int get_aggregate(struct sched_domain *sd) +static int get_aggregate(int cpu, struct sched_domain *sd) { - if (!spin_trylock(&per_cpu(aggregate_lock, sd->first_cpu))) + if (!spin_trylock(&per_cpu(aggregate_lock, cpu))) return 0; - aggregate_walk_tree(aggregate_get_down, aggregate_get_up, sd); + aggregate_walk_tree(aggregate_get_down, aggregate_get_up, cpu, sd); return 1; } -static void put_aggregate(struct sched_domain *sd) +static void put_aggregate(int cpu, struct sched_domain *sd) { - spin_unlock(&per_cpu(aggregate_lock, sd->first_cpu)); + spin_unlock(&per_cpu(aggregate_lock, cpu)); } static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) @@ -1756,12 +1756,12 @@ static inline void init_aggregate(void) { } -static inline int get_aggregate(struct sched_domain *sd) +static inline int get_aggregate(int cpu, struct sched_domain *sd) { return 0; } -static inline void put_aggregate(struct sched_domain *sd) +static inline void put_aggregate(int cpu, struct sched_domain *sd) { } #endif @@ -3539,7 +3539,7 @@ static int load_balance(int this_cpu, struct rq *this_rq, cpus_setall(*cpus); - unlock_aggregate = get_aggregate(sd); + unlock_aggregate = get_aggregate(this_cpu, sd); /* * When power savings policy is enabled for the parent domain, idle @@ -3678,7 +3678,7 @@ out_one_pinned: ld_moved = 0; out: if (unlock_aggregate) - put_aggregate(sd); + put_aggregate(this_cpu, sd); return ld_moved; } @@ -7292,7 +7292,6 @@ static int __build_sched_domains(const cpumask_t *cpu_map, SD_INIT(sd, ALLNODES); set_domain_attribute(sd, attr); sd->span = *cpu_map; - sd->first_cpu = first_cpu(sd->span); cpu_to_allnodes_group(i, cpu_map, &sd->groups, tmpmask); p = sd; sd_allnodes = 1; @@ -7303,7 +7302,6 @@ static int __build_sched_domains(const cpumask_t *cpu_map, SD_INIT(sd, NODE); set_domain_attribute(sd, attr); sched_domain_node_span(cpu_to_node(i), &sd->span); - sd->first_cpu = first_cpu(sd->span); sd->parent = p; if (p) p->child = sd; @@ -7315,7 +7313,6 @@ static int __build_sched_domains(const cpumask_t *cpu_map, SD_INIT(sd, CPU); set_domain_attribute(sd, attr); sd->span = *nodemask; - sd->first_cpu = first_cpu(sd->span); sd->parent = p; if (p) p->child = sd; @@ -7327,7 +7324,6 @@ static int __build_sched_domains(const cpumask_t *cpu_map, SD_INIT(sd, MC); set_domain_attribute(sd, attr); sd->span = cpu_coregroup_map(i); - sd->first_cpu = first_cpu(sd->span); cpus_and(sd->span, sd->span, *cpu_map); sd->parent = p; p->child = sd; @@ -7340,7 +7336,6 @@ static int __build_sched_domains(const cpumask_t *cpu_map, SD_INIT(sd, SIBLING); set_domain_attribute(sd, attr); sd->span = per_cpu(cpu_sibling_map, i); - sd->first_cpu = first_cpu(sd->span); cpus_and(sd->span, sd->span, *cpu_map); sd->parent = p; p->child = sd; -- cgit v1.2.3 From 4d8d595dfa69e1c807bf928f364668a7f30da5dc Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:19 +0200 Subject: sched: update aggregate when holding the RQs It was observed that in __update_group_shares_cpu() rq_weight > aggregate()->rq_weight This is caused by forks/wakeups in between the initial aggregate pass and locking of the RQs for load balance. To avoid this situation partially re-do the aggregation once we have the RQs locked (which avoids new tasks from appearing). Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 160d3c209b8..dae20199dc9 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1721,6 +1721,11 @@ aggregate_get_up(struct task_group *tg, int cpu, struct sched_domain *sd) aggregate_group_set_shares(tg, cpu, sd); } +static void +aggregate_get_nop(struct task_group *tg, int cpu, struct sched_domain *sd) +{ +} + static DEFINE_PER_CPU(spinlock_t, aggregate_lock); static void __init init_aggregate(void) @@ -1740,6 +1745,11 @@ static int get_aggregate(int cpu, struct sched_domain *sd) return 1; } +static void update_aggregate(int cpu, struct sched_domain *sd) +{ + aggregate_walk_tree(aggregate_get_down, aggregate_get_nop, cpu, sd); +} + static void put_aggregate(int cpu, struct sched_domain *sd) { spin_unlock(&per_cpu(aggregate_lock, cpu)); @@ -1761,6 +1771,10 @@ static inline int get_aggregate(int cpu, struct sched_domain *sd) return 0; } +static inline void update_aggregate(int cpu, struct sched_domain *sd) +{ +} + static inline void put_aggregate(int cpu, struct sched_domain *sd) { } @@ -2192,6 +2206,12 @@ find_idlest_group(struct sched_domain *sd, struct task_struct *p, int this_cpu) int load_idx = sd->forkexec_idx; int imbalance = 100 + (sd->imbalance_pct-100)/2; + /* + * now that we have both rqs locked the rq weight won't change + * anymore - so update the stats. + */ + update_aggregate(this_cpu, sd); + do { unsigned long load, avg_load; int local_group; -- cgit v1.2.3 From d3f40dbab954d83383b6a516582d5c09cc216dcc Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:21 +0200 Subject: sched: dont micro manage share losses We used to try and contain the loss of 'shares' by playing arithmetic games. Replace that by noticing that at the top sched_domain we'll always have the full weight in shares to distribute. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 26 +++----------------------- 1 file changed, 3 insertions(+), 23 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index dae20199dc9..28229c5d498 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1551,6 +1551,9 @@ aggregate_group_shares(struct task_group *tg, int cpu, struct sched_domain *sd) if ((!shares && aggregate(tg, cpu)->rq_weight) || shares > tg->shares) shares = tg->shares; + if (!sd->parent || !(sd->parent->flags & SD_LOAD_BALANCE)) + shares = tg->shares; + aggregate(tg, cpu)->shares = shares; } @@ -1642,20 +1645,8 @@ static void __move_group_shares(struct task_group *tg, int cpu, struct sched_domain *sd, int scpu, int dcpu) { - unsigned long shares; - - shares = tg->cfs_rq[scpu]->shares + tg->cfs_rq[dcpu]->shares; - __update_group_shares_cpu(tg, cpu, sd, scpu); __update_group_shares_cpu(tg, cpu, sd, dcpu); - - /* - * ensure we never loose shares due to rounding errors in the - * above redistribution. - */ - shares -= tg->cfs_rq[scpu]->shares + tg->cfs_rq[dcpu]->shares; - if (shares) - tg->cfs_rq[dcpu]->shares += shares; } /* @@ -1675,7 +1666,6 @@ move_group_shares(struct task_group *tg, int cpu, struct sched_domain *sd, static void aggregate_group_set_shares(struct task_group *tg, int cpu, struct sched_domain *sd) { - unsigned long shares = aggregate(tg, cpu)->shares; int i; for_each_cpu_mask(i, sd->span) { @@ -1688,16 +1678,6 @@ aggregate_group_set_shares(struct task_group *tg, int cpu, struct sched_domain * } aggregate_group_shares(tg, cpu, sd); - - /* - * ensure we never loose shares due to rounding errors in the - * above redistribution. - */ - shares -= aggregate(tg, cpu)->shares; - if (shares) { - tg->cfs_rq[cpu]->shares += shares; - aggregate(tg, cpu)->shares += shares; - } } /* -- cgit v1.2.3 From a25b5aca8740ea99d5e18dfc71235a52b685dcf7 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:22 +0200 Subject: sched: no need to aggregate task_weight We only need to know the task_weight of the busiest rq - nothing to do if there are no tasks there. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 16 +--------------- 1 file changed, 1 insertion(+), 15 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 28229c5d498..716cfc8e099 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -433,12 +433,6 @@ struct cfs_rq { * The sum of all runqueue weights within this span. */ unsigned long rq_weight; - - /* - * Weight contributed by tasks; this is the part we can - * influence by moving tasks around. - */ - unsigned long task_weight; } aggregate; #endif #endif @@ -1473,10 +1467,6 @@ static int task_hot(struct task_struct *p, u64 now, struct sched_domain *sd); * rq_weight: * Direct sum of all the cpu's their rq weight, e.g. A would get 3 while * B would get 2. - * - * task_weight: - * Part of the rq_weight contributed by tasks; all groups except B would - * get 1, B gets 2. */ static inline struct aggregate_struct * @@ -1524,16 +1514,12 @@ static void aggregate_group_weight(struct task_group *tg, int cpu, struct sched_domain *sd) { unsigned long rq_weight = 0; - unsigned long task_weight = 0; int i; - for_each_cpu_mask(i, sd->span) { + for_each_cpu_mask(i, sd->span) rq_weight += tg->cfs_rq[i]->load.weight; - task_weight += tg->cfs_rq[i]->task_weight; - } aggregate(tg, cpu)->rq_weight = rq_weight; - aggregate(tg, cpu)->task_weight = task_weight; } /* -- cgit v1.2.3 From c8cba857b4997d5b00451d01474638f6a153f713 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:23 +0200 Subject: sched: simplify the group load balancer While thinking about the previous patch - I realized that using per domain aggregate load values in load_balance_fair() is wrong. We should use the load value for that CPU. By not needing per domain hierarchical load values we don't need to store per domain aggregate shares, which greatly simplifies all the math. It basically falls apart in two separate computations: - per domain update of the shares - per CPU update of the hierarchical load Also get rid of the move_group_shares() stuff - just re-compute the shares again after a successful load balance. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 286 +++++++++++++-------------------------------------------- 1 file changed, 64 insertions(+), 222 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 716cfc8e099..f864b751fd1 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -406,34 +406,23 @@ struct cfs_rq { struct task_group *tg; /* group that "owns" this runqueue */ #ifdef CONFIG_SMP - unsigned long task_weight; - unsigned long shares; /* - * We need space to build a sched_domain wide view of the full task - * group tree, in order to avoid depending on dynamic memory allocation - * during the load balancing we place this in the per cpu task group - * hierarchy. This limits the load balancing to one instance per cpu, - * but more should not be needed anyway. + * the part of load.weight contributed by tasks */ - struct aggregate_struct { - /* - * load = weight(cpus) * f(tg) - * - * Where f(tg) is the recursive weight fraction assigned to - * this group. - */ - unsigned long load; + unsigned long task_weight; - /* - * part of the group weight distributed to this span. - */ - unsigned long shares; + /* + * h_load = weight * f(tg) + * + * Where f(tg) is the recursive weight fraction assigned to + * this group. + */ + unsigned long h_load; - /* - * The sum of all runqueue weights within this span. - */ - unsigned long rq_weight; - } aggregate; + /* + * this cpu's part of tg->shares + */ + unsigned long shares; #endif #endif }; @@ -1443,47 +1432,14 @@ static int task_hot(struct task_struct *p, u64 now, struct sched_domain *sd); #ifdef CONFIG_FAIR_GROUP_SCHED -/* - * Group load balancing. - * - * We calculate a few balance domain wide aggregate numbers; load and weight. - * Given the pictures below, and assuming each item has equal weight: - * - * root 1 - thread - * / | \ A - group - * A 1 B - * /|\ / \ - * C 2 D 3 4 - * | | - * 5 6 - * - * load: - * A and B get 1/3-rd of the total load. C and D get 1/3-rd of A's 1/3-rd, - * which equals 1/9-th of the total load. - * - * shares: - * The weight of this group on the selected cpus. - * - * rq_weight: - * Direct sum of all the cpu's their rq weight, e.g. A would get 3 while - * B would get 2. - */ - -static inline struct aggregate_struct * -aggregate(struct task_group *tg, int cpu) -{ - return &tg->cfs_rq[cpu]->aggregate; -} - -typedef void (*aggregate_func)(struct task_group *, int, struct sched_domain *); +typedef void (*tg_visitor)(struct task_group *, int, struct sched_domain *); /* * Iterate the full tree, calling @down when first entering a node and @up when * leaving it for the final time. */ -static -void aggregate_walk_tree(aggregate_func down, aggregate_func up, - int cpu, struct sched_domain *sd) +static void +walk_tg_tree(tg_visitor down, tg_visitor up, int cpu, struct sched_domain *sd) { struct task_group *parent, *child; @@ -1507,72 +1463,6 @@ up: rcu_read_unlock(); } -/* - * Calculate the aggregate runqueue weight. - */ -static void -aggregate_group_weight(struct task_group *tg, int cpu, struct sched_domain *sd) -{ - unsigned long rq_weight = 0; - int i; - - for_each_cpu_mask(i, sd->span) - rq_weight += tg->cfs_rq[i]->load.weight; - - aggregate(tg, cpu)->rq_weight = rq_weight; -} - -/* - * Compute the weight of this group on the given cpus. - */ -static void -aggregate_group_shares(struct task_group *tg, int cpu, struct sched_domain *sd) -{ - unsigned long shares = 0; - int i; - - for_each_cpu_mask(i, sd->span) - shares += tg->cfs_rq[i]->shares; - - if ((!shares && aggregate(tg, cpu)->rq_weight) || shares > tg->shares) - shares = tg->shares; - - if (!sd->parent || !(sd->parent->flags & SD_LOAD_BALANCE)) - shares = tg->shares; - - aggregate(tg, cpu)->shares = shares; -} - -/* - * Compute the load fraction assigned to this group, relies on the aggregate - * weight and this group's parent's load, i.e. top-down. - */ -static void -aggregate_group_load(struct task_group *tg, int cpu, struct sched_domain *sd) -{ - unsigned long load; - - if (!tg->parent) { - int i; - - load = 0; - for_each_cpu_mask(i, sd->span) - load += cpu_rq(i)->load.weight; - - } else { - load = aggregate(tg->parent, cpu)->load; - - /* - * shares is our weight in the parent's rq so - * shares/parent->rq_weight gives our fraction of the load - */ - load *= aggregate(tg, cpu)->shares; - load /= aggregate(tg->parent, cpu)->rq_weight + 1; - } - - aggregate(tg, cpu)->load = load; -} - static void __set_se_shares(struct sched_entity *se, unsigned long shares); /* @@ -1580,16 +1470,16 @@ static void __set_se_shares(struct sched_entity *se, unsigned long shares); */ static void __update_group_shares_cpu(struct task_group *tg, int cpu, - struct sched_domain *sd, int tcpu) + unsigned long sd_shares, unsigned long sd_rq_weight) { int boost = 0; unsigned long shares; unsigned long rq_weight; - if (!tg->se[tcpu]) + if (!tg->se[cpu]) return; - rq_weight = tg->cfs_rq[tcpu]->load.weight; + rq_weight = tg->cfs_rq[cpu]->load.weight; /* * If there are currently no tasks on the cpu pretend there is one of @@ -1601,124 +1491,97 @@ __update_group_shares_cpu(struct task_group *tg, int cpu, rq_weight = NICE_0_LOAD; } + if (unlikely(rq_weight > sd_rq_weight)) + rq_weight = sd_rq_weight; + /* * \Sum shares * rq_weight * shares = ----------------------- * \Sum rq_weight * */ - shares = aggregate(tg, cpu)->shares * rq_weight; - shares /= aggregate(tg, cpu)->rq_weight + 1; + shares = (sd_shares * rq_weight) / (sd_rq_weight + 1); /* * record the actual number of shares, not the boosted amount. */ - tg->cfs_rq[tcpu]->shares = boost ? 0 : shares; + tg->cfs_rq[cpu]->shares = boost ? 0 : shares; if (shares < MIN_SHARES) shares = MIN_SHARES; else if (shares > MAX_SHARES) shares = MAX_SHARES; - __set_se_shares(tg->se[tcpu], shares); + __set_se_shares(tg->se[cpu], shares); } /* - * Re-adjust the weights on the cpu the task came from and on the cpu the - * task went to. + * Re-compute the task group their per cpu shares over the given domain. + * This needs to be done in a bottom-up fashion because the rq weight of a + * parent group depends on the shares of its child groups. */ static void -__move_group_shares(struct task_group *tg, int cpu, struct sched_domain *sd, - int scpu, int dcpu) +tg_shares_up(struct task_group *tg, int cpu, struct sched_domain *sd) { - __update_group_shares_cpu(tg, cpu, sd, scpu); - __update_group_shares_cpu(tg, cpu, sd, dcpu); -} + unsigned long rq_weight = 0; + unsigned long shares = 0; + int i; -/* - * Because changing a group's shares changes the weight of the super-group - * we need to walk up the tree and change all shares until we hit the root. - */ -static void -move_group_shares(struct task_group *tg, int cpu, struct sched_domain *sd, - int scpu, int dcpu) -{ - while (tg) { - __move_group_shares(tg, cpu, sd, scpu, dcpu); - tg = tg->parent; + for_each_cpu_mask(i, sd->span) { + rq_weight += tg->cfs_rq[i]->load.weight; + shares += tg->cfs_rq[i]->shares; } -} -static void -aggregate_group_set_shares(struct task_group *tg, int cpu, struct sched_domain *sd) -{ - int i; + if ((!shares && rq_weight) || shares > tg->shares) + shares = tg->shares; + + if (!sd->parent || !(sd->parent->flags & SD_LOAD_BALANCE)) + shares = tg->shares; for_each_cpu_mask(i, sd->span) { struct rq *rq = cpu_rq(i); unsigned long flags; spin_lock_irqsave(&rq->lock, flags); - __update_group_shares_cpu(tg, cpu, sd, i); + __update_group_shares_cpu(tg, i, shares, rq_weight); spin_unlock_irqrestore(&rq->lock, flags); } - - aggregate_group_shares(tg, cpu, sd); } /* - * Calculate the accumulative weight and recursive load of each task group - * while walking down the tree. + * Compute the cpu's hierarchical load factor for each task group. + * This needs to be done in a top-down fashion because the load of a child + * group is a fraction of its parents load. */ static void -aggregate_get_down(struct task_group *tg, int cpu, struct sched_domain *sd) +tg_load_down(struct task_group *tg, int cpu, struct sched_domain *sd) { - aggregate_group_weight(tg, cpu, sd); - aggregate_group_shares(tg, cpu, sd); - aggregate_group_load(tg, cpu, sd); -} - -/* - * Rebalance the cpu shares while walking back up the tree. - */ -static void -aggregate_get_up(struct task_group *tg, int cpu, struct sched_domain *sd) -{ - aggregate_group_set_shares(tg, cpu, sd); -} - -static void -aggregate_get_nop(struct task_group *tg, int cpu, struct sched_domain *sd) -{ -} - -static DEFINE_PER_CPU(spinlock_t, aggregate_lock); + unsigned long load; -static void __init init_aggregate(void) -{ - int i; + if (!tg->parent) { + load = cpu_rq(cpu)->load.weight; + } else { + load = tg->parent->cfs_rq[cpu]->h_load; + load *= tg->cfs_rq[cpu]->shares; + load /= tg->parent->cfs_rq[cpu]->load.weight + 1; + } - for_each_possible_cpu(i) - spin_lock_init(&per_cpu(aggregate_lock, i)); + tg->cfs_rq[cpu]->h_load = load; } -static int get_aggregate(int cpu, struct sched_domain *sd) +static void +tg_nop(struct task_group *tg, int cpu, struct sched_domain *sd) { - if (!spin_trylock(&per_cpu(aggregate_lock, cpu))) - return 0; - - aggregate_walk_tree(aggregate_get_down, aggregate_get_up, cpu, sd); - return 1; } -static void update_aggregate(int cpu, struct sched_domain *sd) +static void update_shares(struct sched_domain *sd) { - aggregate_walk_tree(aggregate_get_down, aggregate_get_nop, cpu, sd); + walk_tg_tree(tg_nop, tg_shares_up, 0, sd); } -static void put_aggregate(int cpu, struct sched_domain *sd) +static void update_h_load(int cpu) { - spin_unlock(&per_cpu(aggregate_lock, cpu)); + walk_tg_tree(tg_load_down, tg_nop, cpu, NULL); } static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) @@ -1728,22 +1591,10 @@ static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) #else -static inline void init_aggregate(void) -{ -} - -static inline int get_aggregate(int cpu, struct sched_domain *sd) -{ - return 0; -} - -static inline void update_aggregate(int cpu, struct sched_domain *sd) +static inline void update_shares(struct sched_domain *sd) { } -static inline void put_aggregate(int cpu, struct sched_domain *sd) -{ -} #endif #endif @@ -2172,12 +2023,6 @@ find_idlest_group(struct sched_domain *sd, struct task_struct *p, int this_cpu) int load_idx = sd->forkexec_idx; int imbalance = 100 + (sd->imbalance_pct-100)/2; - /* - * now that we have both rqs locked the rq weight won't change - * anymore - so update the stats. - */ - update_aggregate(this_cpu, sd); - do { unsigned long load, avg_load; int local_group; @@ -3521,12 +3366,9 @@ static int load_balance(int this_cpu, struct rq *this_rq, unsigned long imbalance; struct rq *busiest; unsigned long flags; - int unlock_aggregate; cpus_setall(*cpus); - unlock_aggregate = get_aggregate(this_cpu, sd); - /* * When power savings policy is enabled for the parent domain, idle * sibling can pick up load irrespective of busy siblings. In this case, @@ -3540,6 +3382,7 @@ static int load_balance(int this_cpu, struct rq *this_rq, schedstat_inc(sd, lb_count[idle]); redo: + update_shares(sd); group = find_busiest_group(sd, this_cpu, &imbalance, idle, &sd_idle, cpus, balance); @@ -3663,8 +3506,8 @@ out_one_pinned: else ld_moved = 0; out: - if (unlock_aggregate) - put_aggregate(this_cpu, sd); + if (ld_moved) + update_shares(sd); return ld_moved; } @@ -8019,7 +7862,6 @@ void __init sched_init(void) } #ifdef CONFIG_SMP - init_aggregate(); init_defrootdomain(); #endif -- cgit v1.2.3 From 3e5459b4bea3ca2618cc02d56d12639f2cba531d Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:24 +0200 Subject: sched: fix newidle smp group balancing Re-compute the shares on newidle - so we can make a decision based on recent data. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index f864b751fd1..cdd09462fc9 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1579,6 +1579,13 @@ static void update_shares(struct sched_domain *sd) walk_tg_tree(tg_nop, tg_shares_up, 0, sd); } +static void update_shares_locked(struct rq *rq, struct sched_domain *sd) +{ + spin_unlock(&rq->lock); + update_shares(sd); + spin_lock(&rq->lock); +} + static void update_h_load(int cpu) { walk_tg_tree(tg_load_down, tg_nop, cpu, NULL); @@ -1595,6 +1602,10 @@ static inline void update_shares(struct sched_domain *sd) { } +static inline void update_shares_locked(struct rq *rq, struct sched_domain *sd) +{ +} + #endif #endif @@ -3543,6 +3554,7 @@ load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd, schedstat_inc(sd, lb_count[CPU_NEWLY_IDLE]); redo: + update_shares_locked(this_rq, sd); group = find_busiest_group(sd, this_cpu, &imbalance, CPU_NEWLY_IDLE, &sd_idle, cpus, NULL); if (!group) { @@ -3586,6 +3598,7 @@ redo: } else sd->nr_balance_failed = 0; + update_shares_locked(this_rq, sd); return ld_moved; out_balanced: -- cgit v1.2.3 From 039a1c41b3a489e34593ea1e1687f6fdad6b13ab Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:25 +0200 Subject: sched: fix sched_balance_self() smp group balancing Finding the least idle cpu is more accurate when done with updated shares. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index cdd09462fc9..39d5495540d 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2128,6 +2128,9 @@ static int sched_balance_self(int cpu, int flag) sd = tmp; } + if (sd) + update_shares(sd); + while (sd) { cpumask_t span, tmpmask; struct sched_group *group; -- cgit v1.2.3 From a8a51d5e59561aa5b4d66e19eca819b537783e8f Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:26 +0200 Subject: sched: persistent average load per task Remove the fall-back to SCHED_LOAD_SCALE by remembering the previous value of cpu_avg_load_per_task() - this is useful because of the hierarchical group model in which task weight can be much smaller. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 25 ++++++++++++------------- 1 file changed, 12 insertions(+), 13 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 39d5495540d..6a6b0139eb3 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -554,6 +554,8 @@ struct rq { int cpu; int online; + unsigned long avg_load_per_task; + struct task_struct *migration_thread; struct list_head migration_queue; #endif @@ -1427,9 +1429,18 @@ static inline void dec_cpu_load(struct rq *rq, unsigned long load) #ifdef CONFIG_SMP static unsigned long source_load(int cpu, int type); static unsigned long target_load(int cpu, int type); -static unsigned long cpu_avg_load_per_task(int cpu); static int task_hot(struct task_struct *p, u64 now, struct sched_domain *sd); +static unsigned long cpu_avg_load_per_task(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + + if (rq->nr_running) + rq->avg_load_per_task = rq->load.weight / rq->nr_running; + + return rq->avg_load_per_task; +} + #ifdef CONFIG_FAIR_GROUP_SCHED typedef void (*tg_visitor)(struct task_group *, int, struct sched_domain *); @@ -2010,18 +2021,6 @@ static unsigned long target_load(int cpu, int type) return max(rq->cpu_load[type-1], total); } -/* - * Return the average load per task on the cpu's run queue - */ -static unsigned long cpu_avg_load_per_task(int cpu) -{ - struct rq *rq = cpu_rq(cpu); - unsigned long total = weighted_cpuload(cpu); - unsigned long n = rq->nr_running; - - return n ? total / n : SCHED_LOAD_SCALE; -} - /* * find_idlest_group finds and returns the least busy CPU group within the * domain. -- cgit v1.2.3 From 408ed066b11cf9ee4536573b4269ee3613bd735e Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:28 +0200 Subject: sched: hierarchical load vs find_busiest_group find_busiest_group() has some assumptions about task weight being in the NICE_0_LOAD range. Hierarchical task groups break this assumption - fix this by replacing it with the average task weight, which will adapt the situation. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 26 +++++++++++++++++++++++--- 1 file changed, 23 insertions(+), 3 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 6a6b0139eb3..5e2aa394a81 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -3050,6 +3050,7 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, max_load = this_load = total_load = total_pwr = 0; busiest_load_per_task = busiest_nr_running = 0; this_load_per_task = this_nr_running = 0; + if (idle == CPU_NOT_IDLE) load_idx = sd->busy_idx; else if (idle == CPU_NEWLY_IDLE) @@ -3064,6 +3065,8 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, int __group_imb = 0; unsigned int balance_cpu = -1, first_idle_cpu = 0; unsigned long sum_nr_running, sum_weighted_load; + unsigned long sum_avg_load_per_task; + unsigned long avg_load_per_task; local_group = cpu_isset(this_cpu, group->cpumask); @@ -3072,6 +3075,8 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, /* Tally up the load of all CPUs in the group */ sum_weighted_load = sum_nr_running = avg_load = 0; + sum_avg_load_per_task = avg_load_per_task = 0; + max_cpu_load = 0; min_cpu_load = ~0UL; @@ -3105,6 +3110,8 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, avg_load += load; sum_nr_running += rq->nr_running; sum_weighted_load += weighted_cpuload(i); + + sum_avg_load_per_task += cpu_avg_load_per_task(i); } /* @@ -3126,7 +3133,20 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, avg_load = sg_div_cpu_power(group, avg_load * SCHED_LOAD_SCALE); - if ((max_cpu_load - min_cpu_load) > SCHED_LOAD_SCALE) + + /* + * Consider the group unbalanced when the imbalance is larger + * than the average weight of two tasks. + * + * APZ: with cgroup the avg task weight can vary wildly and + * might not be a suitable number - should we keep a + * normalized nr_running number somewhere that negates + * the hierarchy? + */ + avg_load_per_task = sg_div_cpu_power(group, + sum_avg_load_per_task * SCHED_LOAD_SCALE); + + if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) __group_imb = 1; group_capacity = group->__cpu_power / SCHED_LOAD_SCALE; @@ -3267,9 +3287,9 @@ small_imbalance: if (busiest_load_per_task > this_load_per_task) imbn = 1; } else - this_load_per_task = SCHED_LOAD_SCALE; + this_load_per_task = cpu_avg_load_per_task(this_cpu); - if (max_load - this_load + SCHED_LOAD_SCALE_FUZZ >= + if (max_load - this_load + 2*busiest_load_per_task >= busiest_load_per_task * imbn) { *imbalance = busiest_load_per_task; return busiest; -- cgit v1.2.3 From 051c67640e771fd6ad1b676fc0c16c379b3c6f80 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:31 +0200 Subject: sched: remove prio preference from balance decisions Priority looses much of its meaning in a hierarchical context. So don't use it in balance decisions. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 12 +++--------- 1 file changed, 3 insertions(+), 9 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 5e2aa394a81..10d43f5bf0f 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2896,7 +2896,7 @@ balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, enum cpu_idle_type idle, int *all_pinned, int *this_best_prio, struct rq_iterator *iterator) { - int loops = 0, pulled = 0, pinned = 0, skip_for_load; + int loops = 0, pulled = 0, pinned = 0; struct task_struct *p; long rem_load_move = max_load_move; @@ -2912,14 +2912,8 @@ balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, next: if (!p || loops++ > sysctl_sched_nr_migrate) goto out; - /* - * To help distribute high priority tasks across CPUs we don't - * skip a task if it will be the highest priority task (i.e. smallest - * prio value) on its new queue regardless of its load weight - */ - skip_for_load = (p->se.load.weight >> 1) > rem_load_move + - SCHED_LOAD_SCALE_FUZZ; - if ((skip_for_load && p->prio >= *this_best_prio) || + + if ((p->se.load.weight >> 1) > rem_load_move || !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { p = iterator->next(iterator->arg); goto next; -- cgit v1.2.3 From 93b75217df39e6d75889cc6f8050343286aff4a5 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:33 +0200 Subject: sched: disable source/target_load bias The bias given by source/target_load functions can be very large, disable it by default to get faster convergence. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 10d43f5bf0f..6c5eb3bc37e 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2000,7 +2000,7 @@ static unsigned long source_load(int cpu, int type) struct rq *rq = cpu_rq(cpu); unsigned long total = weighted_cpuload(cpu); - if (type == 0) + if (type == 0 || !sched_feat(LB_BIAS)) return total; return min(rq->cpu_load[type-1], total); @@ -2015,7 +2015,7 @@ static unsigned long target_load(int cpu, int type) struct rq *rq = cpu_rq(cpu); unsigned long total = weighted_cpuload(cpu); - if (type == 0) + if (type == 0 || !sched_feat(LB_BIAS)) return total; return max(rq->cpu_load[type-1], total); -- cgit v1.2.3 From cd80917e4ff465ea77106f8e4fb631eedc4cf426 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:34 +0200 Subject: sched: fix shares boost logic In case the domain is empty, pretend there is a single task on each cpu, so that together with the boost logic we end up giving 1/n shares to each cpu. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 6c5eb3bc37e..1cff969f664 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1549,6 +1549,9 @@ tg_shares_up(struct task_group *tg, int cpu, struct sched_domain *sd) if (!sd->parent || !(sd->parent->flags & SD_LOAD_BALANCE)) shares = tg->shares; + if (!rq_weight) + rq_weight = cpus_weight(sd->span) * NICE_0_LOAD; + for_each_cpu_mask(i, sd->span) { struct rq *rq = cpu_rq(i); unsigned long flags; -- cgit v1.2.3 From 2398f2c6d34b43025f274fc42eaca34d23ec2320 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:35 +0200 Subject: sched: update shares on wakeup We found that the affine wakeup code needs rather accurate load figures to be effective. The trouble is that updating the load figures is fairly expensive with group scheduling. Therefore ratelimit the updating. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 30 +++++++++++++++++++++++++++++- 1 file changed, 29 insertions(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 1cff969f664..62db0891025 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -777,6 +777,12 @@ late_initcall(sched_init_debug); */ const_debug unsigned int sysctl_sched_nr_migrate = 32; +/* + * ratelimit for updating the group shares. + * default: 0.5ms + */ +const_debug unsigned int sysctl_sched_shares_ratelimit = 500000; + /* * period over which we measure -rt task cpu usage in us. * default: 1s @@ -1590,7 +1596,13 @@ tg_nop(struct task_group *tg, int cpu, struct sched_domain *sd) static void update_shares(struct sched_domain *sd) { - walk_tg_tree(tg_nop, tg_shares_up, 0, sd); + u64 now = cpu_clock(raw_smp_processor_id()); + s64 elapsed = now - sd->last_update; + + if (elapsed >= (s64)(u64)sysctl_sched_shares_ratelimit) { + sd->last_update = now; + walk_tg_tree(tg_nop, tg_shares_up, 0, sd); + } } static void update_shares_locked(struct rq *rq, struct sched_domain *sd) @@ -2199,6 +2211,22 @@ static int try_to_wake_up(struct task_struct *p, unsigned int state, int sync) if (!sched_feat(SYNC_WAKEUPS)) sync = 0; +#ifdef CONFIG_SMP + if (sched_feat(LB_WAKEUP_UPDATE)) { + struct sched_domain *sd; + + this_cpu = raw_smp_processor_id(); + cpu = task_cpu(p); + + for_each_domain(this_cpu, sd) { + if (cpu_isset(cpu, sd->span)) { + update_shares(sd); + break; + } + } + } +#endif + smp_wmb(); rq = task_rq_lock(p, &flags); old_state = p->state; -- cgit v1.2.3 From 83378269a5fad98f562ebc0f09c349575e6cbfe1 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:37 +0200 Subject: sched: correct wakeup weight calculations rw_i = {2, 4, 1, 0} s_i = {2/7, 4/7, 1/7, 0} wakeup on cpu0, weight=1 rw'_i = {3, 4, 1, 0} s'_i = {3/8, 4/8, 1/8, 0} s_0 = S * rw_0 / \Sum rw_j -> \Sum rw_j = S*rw_0/s_0 = 1*2*7/2 = 7 (correct) s'_0 = S * (rw_0 + 1) / (\Sum rw_j + 1) = 1 * (2+1) / (7+1) = 3/8 (correct so we find that adding 1 to cpu0 gains 5/56 in weight if say the other cpu were, cpu1, we'd also have to calculate its 4/56 loss Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 62db0891025..01d3e51b711 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -365,6 +365,10 @@ static inline void set_task_rq(struct task_struct *p, unsigned int cpu) #else static inline void set_task_rq(struct task_struct *p, unsigned int cpu) { } +static inline struct task_group *task_group(struct task_struct *p) +{ + return NULL; +} #endif /* CONFIG_GROUP_SCHED */ -- cgit v1.2.3 From f1d239f73200a5803a89e5929fb3abc1596b7589 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 Jun 2008 13:41:38 +0200 Subject: sched: incremental effective_load() Increase the accuracy of the effective_load values. Not only consider the current increment (as per the attempted wakeup), but also consider the delta between when we last adjusted the shares and the current situation. Signed-off-by: Peter Zijlstra Cc: Srivatsa Vaddagiri Cc: Mike Galbraith Signed-off-by: Ingo Molnar --- kernel/sched.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 01d3e51b711..7613f69f097 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -427,6 +427,11 @@ struct cfs_rq { * this cpu's part of tg->shares */ unsigned long shares; + + /* + * load.weight at the time we set shares + */ + unsigned long rq_weight; #endif #endif }; @@ -1527,6 +1532,7 @@ __update_group_shares_cpu(struct task_group *tg, int cpu, * record the actual number of shares, not the boosted amount. */ tg->cfs_rq[cpu]->shares = boost ? 0 : shares; + tg->cfs_rq[cpu]->rq_weight = rq_weight; if (shares < MIN_SHARES) shares = MIN_SHARES; -- cgit v1.2.3 From 34e83e850f5e5ee2a18cd77a5d70d31972a632e6 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 27 Jun 2008 15:42:36 +0200 Subject: sched: build fix MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit fix: kernel/sched.c: In function ‘sched_group_set_shares': kernel/sched.c:8635: error: implicit declaration of function ‘cfs_rq_set_shares' Signed-off-by: Ingo Molnar --- kernel/sched.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 7613f69f097..058250a63b6 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1627,11 +1627,6 @@ static void update_h_load(int cpu) walk_tg_tree(tg_load_down, tg_nop, cpu, NULL); } -static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) -{ - cfs_rq->shares = shares; -} - #else static inline void update_shares(struct sched_domain *sd) @@ -1646,6 +1641,13 @@ static inline void update_shares_locked(struct rq *rq, struct sched_domain *sd) #endif +static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) +{ +#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED) + cfs_rq->shares = shares; +#endif +} + #include "sched_stats.h" #include "sched_idletask.c" #include "sched_fair.c" -- cgit v1.2.3 From 30432094a7f506ad24997a3ba6aed913ab61c01d Mon Sep 17 00:00:00 2001 From: Vegard Nossum Date: Fri, 27 Jun 2008 21:35:50 +0200 Subject: sched: fix warning This patch fixes the following warning: kernel/sched.c:1667: warning: 'cfs_rq_set_shares' defined but not used This seems the correct way to fix this; cfs_rq_set_shares() is only used in a single place, which is also inside #ifdef CONFIG_FAIR_GROUP_SCHED. Signed-off-by: Vegard Nossum Cc: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 058250a63b6..677c80b9a6b 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1641,12 +1641,14 @@ static inline void update_shares_locked(struct rq *rq, struct sched_domain *sd) #endif +#ifdef CONFIG_FAIR_GROUP_SCHED static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) { -#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED) +#ifdef CONFIG_SMP cfs_rq->shares = shares; #endif } +#endif #include "sched_stats.h" #include "sched_idletask.c" -- cgit v1.2.3 From c4acb2c0669c5c5c9b28e9d02a34b5c67edf7092 Mon Sep 17 00:00:00 2001 From: Gregory Haskins Date: Fri, 27 Jun 2008 14:29:55 -0600 Subject: sched: terminate newidle balancing once at least one task has moved over Inspired by Peter Zijlstra. Signed-off-by: Gregory Haskins Cc: npiggin@suse.de Cc: rostedt@goodmis.org Acked-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 677c80b9a6b..d99aeabeb72 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -3013,6 +3013,10 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, max_load_move - total_load_moved, sd, idle, all_pinned, &this_best_prio); class = class->next; + + if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) + break; + } while (class && max_load_move > total_load_moved); return total_load_moved > 0; -- cgit v1.2.3 From 2087a1ad822cd3a68b73338457047fcc54da726b Mon Sep 17 00:00:00 2001 From: Gregory Haskins Date: Fri, 27 Jun 2008 14:30:00 -0600 Subject: sched: add avg-overlap support to RT tasks We have the notion of tracking process-coupling (a.k.a. buddy-wake) via the p->se.last_wake / p->se.avg_overlap facilities, but it is only used for cfs to cfs interactions. There is no reason why an rt to cfs interaction cannot share in establishing a relationhip in a similar manner. Because PREEMPT_RT runs many kernel threads as FIFO priority, we often times have heavy interaction between RT threads waking CFS applications. This patch offers a substantial boost (50-60%+) in perfomance under those circumstances. Signed-off-by: Gregory Haskins Cc: npiggin@suse.de Cc: rostedt@goodmis.org Acked-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index d99aeabeb72..bbc40c3a065 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1693,6 +1693,12 @@ static void set_load_weight(struct task_struct *p) p->se.load.inv_weight = prio_to_wmult[p->static_prio - MAX_RT_PRIO]; } +static void update_avg(u64 *avg, u64 sample) +{ + s64 diff = sample - *avg; + *avg += diff >> 3; +} + static void enqueue_task(struct rq *rq, struct task_struct *p, int wakeup) { sched_info_queued(p); @@ -1702,6 +1708,12 @@ static void enqueue_task(struct rq *rq, struct task_struct *p, int wakeup) static void dequeue_task(struct rq *rq, struct task_struct *p, int sleep) { + if (sleep && p->se.last_wakeup) { + update_avg(&p->se.avg_overlap, + p->se.sum_exec_runtime - p->se.last_wakeup); + p->se.last_wakeup = 0; + } + p->sched_class->dequeue_task(rq, p, sleep); p->se.on_rq = 0; } @@ -2313,6 +2325,8 @@ out_running: p->sched_class->task_wake_up(rq, p); #endif out: + current->se.last_wakeup = current->se.sum_exec_runtime; + task_rq_unlock(rq, &flags); return success; -- cgit v1.2.3 From 46ac22bab42cc868b9c1d0e915ddbc8e8065a44d Mon Sep 17 00:00:00 2001 From: Ankita Garg Date: Tue, 1 Jul 2008 14:30:06 +0530 Subject: sched: fix accounting in task delay accounting & migration On Thu, Jun 19, 2008 at 12:27:14PM +0200, Peter Zijlstra wrote: > On Thu, 2008-06-05 at 10:50 +0530, Ankita Garg wrote: > > > Thanks Peter for the explanation... > > > > I agree with the above and that is the reason why I did not see weird > > values with cpu_time. But, run_delay still would suffer skews as the end > > points for delta could be taken on different cpus due to migration (more > > so on RT kernel due to the push-pull operations). With the below patch, > > I could not reproduce the issue I had seen earlier. After every dequeue, > > we take the delta and start wait measurements from zero when moved to a > > different rq. > > OK, so task delay delay accounting is broken because it doesn't take > migration into account. > > What you've done is make it symmetric wrt enqueue, and account it like > > cpu0 cpu1 > > enqueue > > dequeue > enqueue > > run > > Where you add both d1 and d2 to the run_delay,.. right? > Thanks for reviewing the patch. The above is exactly what I have done. > This seems like a good fix, however it looks like the patch will break > compilation in !CONFIG_SCHEDSTATS && !CONFIG_TASK_DELAY_ACCT, of it > failing to provide a stub for sched_info_dequeue() in that case. Fixed. Pl. find the new patch below. Signed-off-by: Ankita Garg Acked-by: Peter Zijlstra Cc: Gregory Haskins Cc: rostedt@goodmis.org Cc: suresh.b.siddha@intel.com Cc: aneesh.kumar@linux.vnet.ibm.com Cc: dhaval@linux.vnet.ibm.com Cc: vatsa@linux.vnet.ibm.com Cc: David Bahi Signed-off-by: Ingo Molnar --- kernel/sched.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index bbc40c3a065..996bc15196a 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1714,6 +1714,7 @@ static void dequeue_task(struct rq *rq, struct task_struct *p, int sleep) p->se.last_wakeup = 0; } + sched_info_dequeued(p); p->sched_class->dequeue_task(rq, p, sleep); p->se.on_rq = 0; } -- cgit v1.2.3