From 977dda7c9b540f48b228174346d8b31542c1e99f Mon Sep 17 00:00:00 2001 From: Paul Turner Date: Fri, 14 Jan 2011 17:57:50 -0800 Subject: [PATCH 1/7] sched: Update effective_load() to use global share weights Previously effective_load would approximate the global load weight present on a group taking advantage of: entity_weight = tg->shares ( lw / global_lw ), where entity_weight was provided by tg_shares_up. This worked (approximately) for an 'empty' (at tg level) cpu since we would place boost load representative of what a newly woken task would receive. However, now that load is instantaneously updated this assumption is no longer true and the load calculation is rather incorrect in this case. Fix this (and improve the general case) by re-writing effective_load to take advantage of the new shares distribution code. Signed-off-by: Paul Turner Signed-off-by: Peter Zijlstra LKML-Reference: <20110115015817.069769529@google.com> Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index c62ebae65cf0..414145cf5344 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1362,27 +1362,27 @@ static long effective_load(struct task_group *tg, int cpu, long wl, long wg) return wl; for_each_sched_entity(se) { - long S, rw, s, a, b; + long lw, w; - S = se->my_q->tg->shares; - s = se->load.weight; - rw = se->my_q->load.weight; + tg = se->my_q->tg; + w = se->my_q->load.weight; - a = S*(rw + wl); - b = S*rw + s*wg; + /* use this cpu's instantaneous contribution */ + lw = atomic_read(&tg->load_weight); + lw -= se->my_q->load_contribution; + lw += w + wg; - wl = s*(a-b); + wl += w; - if (likely(b)) - wl /= b; + if (lw > 0 && wl < lw) + wl = (wl * tg->shares) / lw; + else + wl = tg->shares; - /* - * Assume the group is already running and will - * thus already be accounted for in the weight. - * - * That is, moving shares between CPUs, does not - * alter the group weight. - */ + /* zero point is MIN_SHARES */ + if (wl < MIN_SHARES) + wl = MIN_SHARES; + wl -= se->load.weight; wg = 0; } From efe25c2c7b3a5d17b0c70987a758d8fe7af8e3d1 Mon Sep 17 00:00:00 2001 From: Bharata B Rao Date: Tue, 11 Jan 2011 15:41:54 +0530 Subject: [PATCH 2/7] sched: Reinstate group names in /proc/sched_debug Displaying of group names in /proc/sched_debug was dropped in autogroup patches. Add group names while displaying cfs_rq and tasks information. Signed-off-by: Bharata B Rao Signed-off-by: Peter Zijlstra LKML-Reference: <20110111101153.GE4772@in.ibm.com> Signed-off-by: Ingo Molnar --- kernel/sched_debug.c | 35 +++++++++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+) diff --git a/kernel/sched_debug.c b/kernel/sched_debug.c index 1dfae3d014b5..4d36f3726da7 100644 --- a/kernel/sched_debug.c +++ b/kernel/sched_debug.c @@ -16,6 +16,8 @@ #include #include +static DEFINE_SPINLOCK(sched_debug_lock); + /* * This allows printing both to /proc/sched_debug and * to the console @@ -86,6 +88,23 @@ static void print_cfs_group_stats(struct seq_file *m, int cpu, struct task_group } #endif +#ifdef CONFIG_CGROUP_SCHED +static char group_path[PATH_MAX]; + +static char *task_group_path(struct task_group *tg) +{ + /* + * May be NULL if the underlying cgroup isn't fully-created yet + */ + if (!tg->css.cgroup) { + group_path[0] = '\0'; + return group_path; + } + cgroup_path(tg->css.cgroup, group_path, PATH_MAX); + return group_path; +} +#endif + static void print_task(struct seq_file *m, struct rq *rq, struct task_struct *p) { @@ -108,6 +127,9 @@ print_task(struct seq_file *m, struct rq *rq, struct task_struct *p) SEQ_printf(m, "%15Ld %15Ld %15Ld.%06ld %15Ld.%06ld %15Ld.%06ld", 0LL, 0LL, 0LL, 0L, 0LL, 0L, 0LL, 0L); #endif +#ifdef CONFIG_CGROUP_SCHED + SEQ_printf(m, " %s", task_group_path(task_group(p))); +#endif SEQ_printf(m, "\n"); } @@ -144,7 +166,11 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq) struct sched_entity *last; unsigned long flags; +#ifdef CONFIG_FAIR_GROUP_SCHED + SEQ_printf(m, "\ncfs_rq[%d]:%s\n", cpu, task_group_path(cfs_rq->tg)); +#else SEQ_printf(m, "\ncfs_rq[%d]:\n", cpu); +#endif SEQ_printf(m, " .%-30s: %Ld.%06ld\n", "exec_clock", SPLIT_NS(cfs_rq->exec_clock)); @@ -191,7 +217,11 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq) void print_rt_rq(struct seq_file *m, int cpu, struct rt_rq *rt_rq) { +#ifdef CONFIG_RT_GROUP_SCHED + SEQ_printf(m, "\nrt_rq[%d]:%s\n", cpu, task_group_path(rt_rq->tg)); +#else SEQ_printf(m, "\nrt_rq[%d]:\n", cpu); +#endif #define P(x) \ SEQ_printf(m, " .%-30s: %Ld\n", #x, (long long)(rt_rq->x)) @@ -212,6 +242,7 @@ extern __read_mostly int sched_clock_running; static void print_cpu(struct seq_file *m, int cpu) { struct rq *rq = cpu_rq(cpu); + unsigned long flags; #ifdef CONFIG_X86 { @@ -266,10 +297,14 @@ static void print_cpu(struct seq_file *m, int cpu) #undef P #endif + spin_lock_irqsave(&sched_debug_lock, flags); print_cfs_stats(m, cpu); print_rt_stats(m, cpu); + rcu_read_lock(); print_rq(m, rq, cpu); + rcu_read_unlock(); + spin_unlock_irqrestore(&sched_debug_lock, flags); } static const char *sched_tunable_scaling_names[] = { From 8ecedd7a06d27a31dbb36fab88e2ba6e6edd43ca Mon Sep 17 00:00:00 2001 From: Bharata B Rao Date: Tue, 11 Jan 2011 15:42:57 +0530 Subject: [PATCH 3/7] sched: Display autogroup names in /proc/sched_debug Add autogroup name to cfs_rq and tasks information to /proc/sched_debug. Signed-off-by: Bharata B Rao Signed-off-by: Peter Zijlstra LKML-Reference: <20110111101257.GF4772@in.ibm.com> Signed-off-by: Ingo Molnar --- kernel/sched_autogroup.c | 5 +++++ kernel/sched_debug.c | 3 +++ 2 files changed, 8 insertions(+) diff --git a/kernel/sched_autogroup.c b/kernel/sched_autogroup.c index 32a723b8f84c..938d52f80a2d 100644 --- a/kernel/sched_autogroup.c +++ b/kernel/sched_autogroup.c @@ -231,6 +231,11 @@ void proc_sched_autogroup_show_task(struct task_struct *p, struct seq_file *m) #ifdef CONFIG_SCHED_DEBUG static inline int autogroup_path(struct task_group *tg, char *buf, int buflen) { + int enabled = ACCESS_ONCE(sysctl_sched_autogroup_enabled); + + if (!enabled || !tg->autogroup) + return 0; + return snprintf(buf, buflen, "%s-%ld", "/autogroup", tg->autogroup->id); } #endif /* CONFIG_SCHED_DEBUG */ diff --git a/kernel/sched_debug.c b/kernel/sched_debug.c index 4d36f3726da7..e4d37259d490 100644 --- a/kernel/sched_debug.c +++ b/kernel/sched_debug.c @@ -93,6 +93,9 @@ static char group_path[PATH_MAX]; static char *task_group_path(struct task_group *tg) { + if (autogroup_path(tg, group_path, PATH_MAX)) + return group_path; + /* * May be NULL if the underlying cgroup isn't fully-created yet */ From f44937718ce3b8360f72f6c68c9481712517a867 Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Thu, 13 Jan 2011 04:54:50 +0100 Subject: [PATCH 4/7] sched, autogroup: Fix CONFIG_RT_GROUP_SCHED sched_setscheduler() failure If CONFIG_RT_GROUP_SCHED is set, __sched_setscheduler() fails due to autogroup not allocating rt_runtime. Free unused/unusable rt_se and rt_rq, redirect RT tasks to the root task group, and tell __sched_setscheduler() that it's ok. Reported-and-tested-by: Bharata B Rao Signed-off-by: Mike Galbraith Signed-off-by: Peter Zijlstra LKML-Reference: <1294890890.8089.39.camel@marge.simson.net> Signed-off-by: Ingo Molnar --- kernel/sched.c | 3 ++- kernel/sched_autogroup.c | 27 +++++++++++++++++++++++++++ kernel/sched_autogroup.h | 4 ++++ 3 files changed, 33 insertions(+), 1 deletion(-) diff --git a/kernel/sched.c b/kernel/sched.c index a0eb0941fa84..6cbff6bd1a60 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4871,7 +4871,8 @@ recheck: * assigned. */ if (rt_bandwidth_enabled() && rt_policy(policy) && - task_group(p)->rt_bandwidth.rt_runtime == 0) { + task_group(p)->rt_bandwidth.rt_runtime == 0 && + !task_group_is_autogroup(task_group(p))) { __task_rq_unlock(rq); raw_spin_unlock_irqrestore(&p->pi_lock, flags); return -EPERM; diff --git a/kernel/sched_autogroup.c b/kernel/sched_autogroup.c index 938d52f80a2d..9fb656283157 100644 --- a/kernel/sched_autogroup.c +++ b/kernel/sched_autogroup.c @@ -27,6 +27,11 @@ static inline void autogroup_destroy(struct kref *kref) { struct autogroup *ag = container_of(kref, struct autogroup, kref); +#ifdef CONFIG_RT_GROUP_SCHED + /* We've redirected RT tasks to the root task group... */ + ag->tg->rt_se = NULL; + ag->tg->rt_rq = NULL; +#endif sched_destroy_group(ag->tg); } @@ -55,6 +60,10 @@ static inline struct autogroup *autogroup_task_get(struct task_struct *p) return ag; } +#ifdef CONFIG_RT_GROUP_SCHED +static void free_rt_sched_group(struct task_group *tg); +#endif + static inline struct autogroup *autogroup_create(void) { struct autogroup *ag = kzalloc(sizeof(*ag), GFP_KERNEL); @@ -72,6 +81,19 @@ static inline struct autogroup *autogroup_create(void) init_rwsem(&ag->lock); ag->id = atomic_inc_return(&autogroup_seq_nr); ag->tg = tg; +#ifdef CONFIG_RT_GROUP_SCHED + /* + * Autogroup RT tasks are redirected to the root task group + * so we don't have to move tasks around upon policy change, + * or flail around trying to allocate bandwidth on the fly. + * A bandwidth exception in __sched_setscheduler() allows + * the policy change to proceed. Thereafter, task_group() + * returns &root_task_group, so zero bandwidth is required. + */ + free_rt_sched_group(tg); + tg->rt_se = root_task_group.rt_se; + tg->rt_rq = root_task_group.rt_rq; +#endif tg->autogroup = ag; return ag; @@ -106,6 +128,11 @@ task_wants_autogroup(struct task_struct *p, struct task_group *tg) return true; } +static inline bool task_group_is_autogroup(struct task_group *tg) +{ + return tg != &root_task_group && tg->autogroup; +} + static inline struct task_group * autogroup_task_group(struct task_struct *p, struct task_group *tg) { diff --git a/kernel/sched_autogroup.h b/kernel/sched_autogroup.h index 5358e241cb20..7b859ffe5dad 100644 --- a/kernel/sched_autogroup.h +++ b/kernel/sched_autogroup.h @@ -15,6 +15,10 @@ autogroup_task_group(struct task_struct *p, struct task_group *tg); static inline void autogroup_init(struct task_struct *init_task) { } static inline void autogroup_free(struct task_group *tg) { } +static inline bool task_group_is_autogroup(struct task_group *tg) +{ + return 0; +} static inline struct task_group * autogroup_task_group(struct task_struct *p, struct task_group *tg) From fce2097983d914ea8c2338efc6f6e9bb737f6ae4 Mon Sep 17 00:00:00 2001 From: Yong Zhang Date: Fri, 14 Jan 2011 15:57:39 +0800 Subject: [PATCH 5/7] sched: Replace rq->bkl_count with rq->rq_sched_info.bkl_count Now rq->rq_sched_info.bkl_count is not used for rq, scroll rq->bkl_count into it. Thus we can save some space for rq. Signed-off-by: Yong Zhang Signed-off-by: Peter Zijlstra LKML-Reference: <1294991859-13246-1-git-send-email-yong.zhang0@gmail.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 5 +---- kernel/sched_debug.c | 4 +++- 2 files changed, 4 insertions(+), 5 deletions(-) diff --git a/kernel/sched.c b/kernel/sched.c index 6cbff6bd1a60..0a169a85eb3e 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -553,9 +553,6 @@ struct rq { /* try_to_wake_up() stats */ unsigned int ttwu_count; unsigned int ttwu_local; - - /* BKL stats */ - unsigned int bkl_count; #endif }; @@ -3887,7 +3884,7 @@ static inline void schedule_debug(struct task_struct *prev) schedstat_inc(this_rq(), sched_count); #ifdef CONFIG_SCHEDSTATS if (unlikely(prev->lock_depth >= 0)) { - schedstat_inc(this_rq(), bkl_count); + schedstat_inc(this_rq(), rq_sched_info.bkl_count); schedstat_inc(prev, sched_info.bkl_count); } #endif diff --git a/kernel/sched_debug.c b/kernel/sched_debug.c index e4d37259d490..eb6cb8edd075 100644 --- a/kernel/sched_debug.c +++ b/kernel/sched_debug.c @@ -296,9 +296,11 @@ static void print_cpu(struct seq_file *m, int cpu) P(ttwu_count); P(ttwu_local); - P(bkl_count); + SEQ_printf(m, " .%-30s: %d\n", "bkl_count", + rq->rq_sched_info.bkl_count); #undef P +#undef P64 #endif spin_lock_irqsave(&sched_debug_lock, flags); print_cfs_stats(m, cpu); From d7d8294415f0ce4254827d4a2a5ee88b00be52a8 Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Wed, 5 Jan 2011 05:41:17 +0100 Subject: [PATCH 6/7] sched: Fix signed unsigned comparison in check_preempt_tick() Signed unsigned comparison may lead to superfluous resched if leftmost is right of the current task, wasting a few cycles, and inadvertently _lengthening_ the current task's slice. Reported-by: Venkatesh Pallipadi Signed-off-by: Mike Galbraith Signed-off-by: Peter Zijlstra LKML-Reference: <1294202477.9384.5.camel@marge.simson.net> Signed-off-by: Ingo Molnar --- kernel/sched_fair.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 414145cf5344..77e9166d7bbf 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1062,6 +1062,9 @@ check_preempt_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr) struct sched_entity *se = __pick_next_entity(cfs_rq); s64 delta = curr->vruntime - se->vruntime; + if (delta < 0) + return; + if (delta > ideal_runtime) resched_task(rq_of(cfs_rq)->curr); } From 068c5cc5ac7414a8e9eb7856b4bf3cc4d4744267 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 19 Jan 2011 12:26:11 +0100 Subject: [PATCH 7/7] sched, cgroup: Use exit hook to avoid use-after-free crash By not notifying the controller of the on-exit move back to init_css_set, we fail to move the task out of the previous cgroup's cfs_rq. This leads to an opportunity for a cgroup-destroy to come in and free the cgroup (there are no active tasks left in it after all) to which the not-quite dead task is still enqueued. Reported-by: Miklos Vajna Fixed-by: Mike Galbraith Signed-off-by: Peter Zijlstra Cc: Cc: Mike Galbraith Signed-off-by: Ingo Molnar LKML-Reference: <1293206353.29444.205.camel@laptop> --- kernel/sched.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/kernel/sched.c b/kernel/sched.c index 0a169a85eb3e..fa5272a0932c 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -606,6 +606,9 @@ static inline struct task_group *task_group(struct task_struct *p) struct task_group *tg; struct cgroup_subsys_state *css; + if (p->flags & PF_EXITING) + return &root_task_group; + css = task_subsys_state_check(p, cpu_cgroup_subsys_id, lockdep_is_held(&task_rq(p)->lock)); tg = container_of(css, struct task_group, css); @@ -8880,6 +8883,20 @@ cpu_cgroup_attach(struct cgroup_subsys *ss, struct cgroup *cgrp, } } +static void +cpu_cgroup_exit(struct cgroup_subsys *ss, struct task_struct *task) +{ + /* + * cgroup_exit() is called in the copy_process() failure path. + * Ignore this case since the task hasn't ran yet, this avoids + * trying to poke a half freed task state from generic code. + */ + if (!(task->flags & PF_EXITING)) + return; + + sched_move_task(task); +} + #ifdef CONFIG_FAIR_GROUP_SCHED static int cpu_shares_write_u64(struct cgroup *cgrp, struct cftype *cftype, u64 shareval) @@ -8952,6 +8969,7 @@ struct cgroup_subsys cpu_cgroup_subsys = { .destroy = cpu_cgroup_destroy, .can_attach = cpu_cgroup_can_attach, .attach = cpu_cgroup_attach, + .exit = cpu_cgroup_exit, .populate = cpu_cgroup_populate, .subsys_id = cpu_cgroup_subsys_id, .early_init = 1,