From eb95308ee2a69403909e111837b9068c64cfc349 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 17 Apr 2012 13:38:40 +0200 Subject: sched: Fix more load-balancing fallout Commits 367456c756a6 ("sched: Ditch per cgroup task lists for load-balancing") and 5d6523ebd ("sched: Fix load-balance wreckage") left some more wreckage. By setting loop_max unconditionally to ->nr_running load-balancing could take a lot of time on very long runqueues (hackbench!). So keep the sysctl as max limit of the amount of tasks we'll iterate. Furthermore, the min load filter for migration completely fails with cgroups since inequality in per-cpu state can easily lead to such small loads :/ Furthermore the change to add new tasks to the tail of the queue instead of the head seems to have some effect.. not quite sure I understand why. Combined these fixes solve the huge hackbench regression reported by Tim when hackbench is ran in a cgroup. Reported-by: Tim Chen Acked-by: Tim Chen Signed-off-by: Peter Zijlstra Cc: Linus Torvalds Cc: Andrew Morton Link: http://lkml.kernel.org/r/1335365763.28150.267.camel@twins [ got rid of the CONFIG_PREEMPT tuning and made small readability edits ] Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 18 ++++++++++-------- kernel/sched/features.h | 1 + 2 files changed, 11 insertions(+), 8 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 0d97ebdc58f..e9553640c1c 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -784,7 +784,7 @@ account_entity_enqueue(struct cfs_rq *cfs_rq, struct sched_entity *se) update_load_add(&rq_of(cfs_rq)->load, se->load.weight); #ifdef CONFIG_SMP if (entity_is_task(se)) - list_add_tail(&se->group_node, &rq_of(cfs_rq)->cfs_tasks); + list_add(&se->group_node, &rq_of(cfs_rq)->cfs_tasks); #endif cfs_rq->nr_running++; } @@ -3215,6 +3215,8 @@ static int move_one_task(struct lb_env *env) static unsigned long task_h_load(struct task_struct *p); +static const unsigned int sched_nr_migrate_break = 32; + /* * move_tasks tries to move up to load_move weighted load from busiest to * this_rq, as part of a balancing operation within domain "sd". @@ -3242,7 +3244,7 @@ static int move_tasks(struct lb_env *env) /* take a breather every nr_migrate tasks */ if (env->loop > env->loop_break) { - env->loop_break += sysctl_sched_nr_migrate; + env->loop_break += sched_nr_migrate_break; env->flags |= LBF_NEED_BREAK; break; } @@ -3252,7 +3254,7 @@ static int move_tasks(struct lb_env *env) load = task_h_load(p); - if (load < 16 && !env->sd->nr_balance_failed) + if (sched_feat(LB_MIN) && load < 16 && !env->sd->nr_balance_failed) goto next; if ((load / 2) > env->load_move) @@ -4407,7 +4409,7 @@ static int load_balance(int this_cpu, struct rq *this_rq, .dst_cpu = this_cpu, .dst_rq = this_rq, .idle = idle, - .loop_break = sysctl_sched_nr_migrate, + .loop_break = sched_nr_migrate_break, }; cpumask_copy(cpus, cpu_active_mask); @@ -4445,10 +4447,10 @@ redo: * correctly treated as an imbalance. */ env.flags |= LBF_ALL_PINNED; - env.load_move = imbalance; - env.src_cpu = busiest->cpu; - env.src_rq = busiest; - env.loop_max = busiest->nr_running; + env.load_move = imbalance; + env.src_cpu = busiest->cpu; + env.src_rq = busiest; + env.loop_max = min_t(unsigned long, sysctl_sched_nr_migrate, busiest->nr_running); more_balance: local_irq_save(flags); diff --git a/kernel/sched/features.h b/kernel/sched/features.h index e61fd73913d..de00a486c5c 100644 --- a/kernel/sched/features.h +++ b/kernel/sched/features.h @@ -68,3 +68,4 @@ SCHED_FEAT(TTWU_QUEUE, true) SCHED_FEAT(FORCE_SD_OVERLAP, false) SCHED_FEAT(RT_RUNTIME_SHARE, true) +SCHED_FEAT(LB_MIN, false) -- cgit v1.2.3-18-g5258 From fb2cf2c660971bea0ad86a9a5c19ad39eab61344 Mon Sep 17 00:00:00 2001 From: "he, bo" Date: Wed, 25 Apr 2012 19:59:21 +0800 Subject: sched: Fix OOPS when build_sched_domains() percpu allocation fails Under extreme memory used up situations, percpu allocation might fail. We hit it when system goes to suspend-to-ram, causing a kworker panic: EIP: [] build_sched_domains+0x23a/0xad0 Kernel panic - not syncing: Fatal exception Pid: 3026, comm: kworker/u:3 3.0.8-137473-gf42fbef #1 Call Trace: [] panic+0x66/0x16c [...] [] partition_sched_domains+0x287/0x4b0 [] cpuset_update_active_cpus+0x1fe/0x210 [] cpuset_cpu_inactive+0x1d/0x30 [...] With this fix applied build_sched_domains() will return -ENOMEM and the suspend attempt fails. Signed-off-by: he, bo Reviewed-by: Zhang, Yanmin Reviewed-by: Srivatsa S. Bhat Signed-off-by: Peter Zijlstra Cc: Linus Torvalds Cc: Andrew Morton Cc: Link: http://lkml.kernel.org/r/1335355161.5892.17.camel@hebo [ So, we fail to deallocate a CPU because we cannot allocate RAM :-/ I don't like that kind of sad behavior but nevertheless it should not crash under high memory load. ] Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 22 ++++++++++++++++------ 1 file changed, 16 insertions(+), 6 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 4603b9d8f30..0533a688ce2 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -6405,16 +6405,26 @@ static void __sdt_free(const struct cpumask *cpu_map) struct sd_data *sdd = &tl->data; for_each_cpu(j, cpu_map) { - struct sched_domain *sd = *per_cpu_ptr(sdd->sd, j); - if (sd && (sd->flags & SD_OVERLAP)) - free_sched_groups(sd->groups, 0); - kfree(*per_cpu_ptr(sdd->sd, j)); - kfree(*per_cpu_ptr(sdd->sg, j)); - kfree(*per_cpu_ptr(sdd->sgp, j)); + struct sched_domain *sd; + + if (sdd->sd) { + sd = *per_cpu_ptr(sdd->sd, j); + if (sd && (sd->flags & SD_OVERLAP)) + free_sched_groups(sd->groups, 0); + kfree(*per_cpu_ptr(sdd->sd, j)); + } + + if (sdd->sg) + kfree(*per_cpu_ptr(sdd->sg, j)); + if (sdd->sgp) + kfree(*per_cpu_ptr(sdd->sgp, j)); } free_percpu(sdd->sd); + sdd->sd = NULL; free_percpu(sdd->sg); + sdd->sg = NULL; free_percpu(sdd->sgp); + sdd->sgp = NULL; } } -- cgit v1.2.3-18-g5258