From 74dc3384fc7983b78cc46ebb1824968a3db85eb1 Mon Sep 17 00:00:00 2001 From: Aleksa Sarai Date: Sun, 6 Aug 2017 14:41:41 +1000 Subject: sched/debug: Use task_pid_nr_ns in /proc/$pid/sched It appears as though the addition of the PID namespace did not update the output code for /proc/*/sched, which resulted in it providing PIDs that were not self-consistent with the /proc mount. This additionally made it trivial to detect whether a process was inside &init_pid_ns from userspace, making container detection trivial: https://github.com/jessfraz/amicontained This leads to situations such as: % unshare -pmf % mount -t proc proc /proc % head -n1 /proc/1/sched head (10047, #threads: 1) Fix this by just using task_pid_nr_ns for the output of /proc/*/sched. All of the other uses of task_pid_nr in kernel/sched/debug.c are from a sysctl context and thus don't need to be namespaced. Signed-off-by: Aleksa Sarai Signed-off-by: Peter Zijlstra (Intel) Acked-by: Eric W. Biederman Cc: Jess Frazelle Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: cyphar@cyphar.com Link: http://lkml.kernel.org/r/20170806044141.5093-1-asarai@suse.com Signed-off-by: Ingo Molnar --- kernel/sched/debug.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel/sched/debug.c') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index 4fa66de52bd6..ac345115877b 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -872,11 +872,12 @@ static void sched_show_numa(struct task_struct *p, struct seq_file *m) #endif } -void proc_sched_show_task(struct task_struct *p, struct seq_file *m) +void proc_sched_show_task(struct task_struct *p, struct pid_namespace *ns, + struct seq_file *m) { unsigned long nr_switches; - SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr(p), + SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr_ns(p, ns), get_nr_threads(p)); SEQ_printf(m, "---------------------------------------------------------" -- cgit v1.2.3 From e8c164954b926f06f109a42fb8595ed01275b141 Mon Sep 17 00:00:00 2001 From: Xie XiuQi Date: Mon, 7 Aug 2017 16:44:22 +0800 Subject: sched/debug: Show task state in /proc/sched_debug Currently we print the runnable task in /proc/sched_debug, but there is no task state information. We don't know which task is in the runqueue and which task is sleeping. Add task state in the runnable task list, like this: runnable tasks: S task PID tree-key switches prio wait-time sum-exec sum-sleep ----------------------------------------------------------------------------------------------------------- S watchdog/239 1452 -11.917445 2811 0 0.000000 8.949306 0.000000 7 0 / S migration/239 1453 20686.367740 8 0 0.000000 16215.720897 0.000000 7 0 / S ksoftirqd/239 1454 115383.841071 12 120 0.000000 0.200683 0.000000 7 0 / >R test 21287 4872.190970 407 120 0.000000 4874.911790 0.000000 7 0 /autogroup-150 R test 21288 4868.385454 401 120 0.000000 3672.341489 0.000000 7 0 /autogroup-150 R test 21289 4868.326776 384 120 0.000000 3424.934159 0.000000 7 0 /autogroup-150 Signed-off-by: Xie XiuQi Signed-off-by: Peter Zijlstra (Intel) Cc: Cc: Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/1502095463-160172-2-git-send-email-xiexiuqi@huawei.com Signed-off-by: Ingo Molnar --- kernel/sched/debug.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) (limited to 'kernel/sched/debug.c') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index ac345115877b..d8d2ea215b85 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -421,13 +421,19 @@ static char *task_group_path(struct task_group *tg) } #endif +static const char stat_nam[] = TASK_STATE_TO_CHAR_STR; + static void print_task(struct seq_file *m, struct rq *rq, struct task_struct *p) { - if (rq->curr == p) - SEQ_printf(m, "R"); - else - SEQ_printf(m, " "); + unsigned long state; + + if (rq->curr == p) { + SEQ_printf(m, ">R"); + } else { + state = p->state ? __ffs(p->state) + 1 : 0; + SEQ_printf(m, " %c", state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?'); + } SEQ_printf(m, "%15s %5d %9Ld.%06ld %9Ld %5d ", p->comm, task_pid_nr(p), @@ -456,9 +462,9 @@ static void print_rq(struct seq_file *m, struct rq *rq, int rq_cpu) SEQ_printf(m, "\nrunnable tasks:\n" - " task PID tree-key switches prio" + " S task PID tree-key switches prio" " wait-time sum-exec sum-sleep\n" - "------------------------------------------------------" + "-------------------------------------------------------" "----------------------------------------------------\n"); rcu_read_lock(); -- cgit v1.2.3 From 20435d84e5f2041c64c792399ab6f2948a2c2252 Mon Sep 17 00:00:00 2001 From: Xie XiuQi Date: Mon, 7 Aug 2017 16:44:23 +0800 Subject: sched/debug: Intruduce task_state_to_char() helper function Now that we have more than one place to get the task state, intruduce the task_state_to_char() helper function to save some code. No functionality changed. Signed-off-by: Xie XiuQi Signed-off-by: Peter Zijlstra (Intel) Cc: Cc: Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/1502095463-160172-3-git-send-email-xiexiuqi@huawei.com Signed-off-by: Ingo Molnar --- include/linux/sched.h | 13 +++++++++++++ kernel/sched/core.c | 15 ++++----------- kernel/sched/debug.c | 10 +++------- 3 files changed, 20 insertions(+), 18 deletions(-) (limited to 'kernel/sched/debug.c') diff --git a/include/linux/sched.h b/include/linux/sched.h index 8337e2db0bb2..c28b182c9833 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1229,6 +1229,19 @@ static inline pid_t task_pgrp_nr(struct task_struct *tsk) return task_pgrp_nr_ns(tsk, &init_pid_ns); } +static inline char task_state_to_char(struct task_struct *task) +{ + const char stat_nam[] = TASK_STATE_TO_CHAR_STR; + unsigned long state = task->state; + + state = state ? __ffs(state) + 1 : 0; + + /* Make sure the string lines up properly with the number of task states: */ + BUILD_BUG_ON(sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1); + + return state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?'; +} + /** * is_global_init - check if a task structure is init. Since init * is free to have sub-threads we need to check tgid. diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 6d91c10b1814..f9f9948e2470 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5103,24 +5103,17 @@ out_unlock: return retval; } -static const char stat_nam[] = TASK_STATE_TO_CHAR_STR; - void sched_show_task(struct task_struct *p) { unsigned long free = 0; int ppid; - unsigned long state = p->state; - - /* Make sure the string lines up properly with the number of task states: */ - BUILD_BUG_ON(sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1); if (!try_get_task_stack(p)) return; - if (state) - state = __ffs(state) + 1; - printk(KERN_INFO "%-15.15s %c", p->comm, - state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?'); - if (state == TASK_RUNNING) + + printk(KERN_INFO "%-15.15s %c", p->comm, task_state_to_char(p)); + + if (p->state == TASK_RUNNING) printk(KERN_CONT " running task "); #ifdef CONFIG_DEBUG_STACK_USAGE free = stack_not_used(p); diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index d8d2ea215b85..cfd84f79e075 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -426,14 +426,10 @@ static const char stat_nam[] = TASK_STATE_TO_CHAR_STR; static void print_task(struct seq_file *m, struct rq *rq, struct task_struct *p) { - unsigned long state; - - if (rq->curr == p) { + if (rq->curr == p) SEQ_printf(m, ">R"); - } else { - state = p->state ? __ffs(p->state) + 1 : 0; - SEQ_printf(m, " %c", state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?'); - } + else + SEQ_printf(m, " %c", task_state_to_char(p)); SEQ_printf(m, "%15s %5d %9Ld.%06ld %9Ld %5d ", p->comm, task_pid_nr(p), -- cgit v1.2.3 From bbdacdfed2f5fa50a2cc9f500a36e05990a0837d Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 10 Aug 2017 17:10:26 +0200 Subject: sched/debug: Optimize sched_domain sysctl generation Currently we unconditionally destroy all sysctl bits and regenerate them after we've rebuild the domains (even if that rebuild is a no-op). And since we unconditionally (re)build the sysctl for all possible CPUs, onlining all CPUs gets us O(n^2) time. Instead change this to only rebuild the bits for CPUs we've actually installed new domains on. Reported-by: Ofer Levi(SW) Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Signed-off-by: Ingo Molnar --- kernel/sched/debug.c | 68 +++++++++++++++++++++++++++++++++++++++---------- kernel/sched/sched.h | 4 +++ kernel/sched/topology.c | 1 + 3 files changed, 59 insertions(+), 14 deletions(-) (limited to 'kernel/sched/debug.c') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index cfd84f79e075..4a23bbc3111b 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -327,38 +327,78 @@ static struct ctl_table *sd_alloc_ctl_cpu_table(int cpu) return table; } +static cpumask_var_t sd_sysctl_cpus; static struct ctl_table_header *sd_sysctl_header; + void register_sched_domain_sysctl(void) { - int i, cpu_num = num_possible_cpus(); - struct ctl_table *entry = sd_alloc_ctl_entry(cpu_num + 1); + static struct ctl_table *cpu_entries; + static struct ctl_table **cpu_idx; char buf[32]; + int i; - WARN_ON(sd_ctl_dir[0].child); - sd_ctl_dir[0].child = entry; + if (!cpu_entries) { + cpu_entries = sd_alloc_ctl_entry(num_possible_cpus() + 1); + if (!cpu_entries) + return; - if (entry == NULL) - return; + WARN_ON(sd_ctl_dir[0].child); + sd_ctl_dir[0].child = cpu_entries; + } - for_each_possible_cpu(i) { - snprintf(buf, 32, "cpu%d", i); - entry->procname = kstrdup(buf, GFP_KERNEL); - entry->mode = 0555; - entry->child = sd_alloc_ctl_cpu_table(i); - entry++; + if (!cpu_idx) { + struct ctl_table *e = cpu_entries; + + cpu_idx = kcalloc(nr_cpu_ids, sizeof(struct ctl_table*), GFP_KERNEL); + if (!cpu_idx) + return; + + /* deal with sparse possible map */ + for_each_possible_cpu(i) { + cpu_idx[i] = e; + e++; + } + } + + if (!cpumask_available(sd_sysctl_cpus)) { + if (!alloc_cpumask_var(&sd_sysctl_cpus, GFP_KERNEL)) + return; + + /* init to possible to not have holes in @cpu_entries */ + cpumask_copy(sd_sysctl_cpus, cpu_possible_mask); + } + + for_each_cpu(i, sd_sysctl_cpus) { + struct ctl_table *e = cpu_idx[i]; + + if (e->child) + sd_free_ctl_entry(&e->child); + + if (!e->procname) { + snprintf(buf, 32, "cpu%d", i); + e->procname = kstrdup(buf, GFP_KERNEL); + } + e->mode = 0555; + e->child = sd_alloc_ctl_cpu_table(i); + + __cpumask_clear_cpu(i, sd_sysctl_cpus); } WARN_ON(sd_sysctl_header); sd_sysctl_header = register_sysctl_table(sd_ctl_root); } +void dirty_sched_domain_sysctl(int cpu) +{ + if (cpumask_available(sd_sysctl_cpus)) + __cpumask_set_cpu(cpu, sd_sysctl_cpus); +} + /* may be called multiple times per register */ void unregister_sched_domain_sysctl(void) { unregister_sysctl_table(sd_sysctl_header); sd_sysctl_header = NULL; - if (sd_ctl_dir[0].child) - sd_free_ctl_entry(&sd_ctl_dir[0].child); } #endif /* CONFIG_SYSCTL */ #endif /* CONFIG_SMP */ diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index eeef1a3086d1..25e5cb1107f3 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -1120,11 +1120,15 @@ extern int group_balance_cpu(struct sched_group *sg); #if defined(CONFIG_SCHED_DEBUG) && defined(CONFIG_SYSCTL) void register_sched_domain_sysctl(void); +void dirty_sched_domain_sysctl(int cpu); void unregister_sched_domain_sysctl(void); #else static inline void register_sched_domain_sysctl(void) { } +static inline void dirty_sched_domain_sysctl(int cpu) +{ +} static inline void unregister_sched_domain_sysctl(void) { } diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c index 727daa2a0abe..6f7b43982f73 100644 --- a/kernel/sched/topology.c +++ b/kernel/sched/topology.c @@ -459,6 +459,7 @@ cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu) rq_attach_root(rq, rd); tmp = rq->sd; rcu_assign_pointer(rq->sd, sd); + dirty_sched_domain_sysctl(cpu); destroy_sched_domains(tmp); update_top_cache_domain(cpu); -- cgit v1.2.3 From bfb068892d30dcf0a32b89302fe293347adeaaaa Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Fri, 8 Sep 2017 16:14:55 -0700 Subject: sched/fair: replace cfs_rq->rb_leftmost ... with the generic rbtree flavor instead. No changes in semantics whatsoever. Link: http://lkml.kernel.org/r/20170719014603.19029-8-dave@stgolabs.net Signed-off-by: Davidlohr Bueso Acked-by: Peter Zijlstra (Intel) Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/sched/debug.c | 2 +- kernel/sched/fair.c | 39 +++++++++++++-------------------------- kernel/sched/sched.h | 3 +-- 3 files changed, 15 insertions(+), 29 deletions(-) (limited to 'kernel/sched/debug.c') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index 4a23bbc3111b..8e536d963652 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -530,7 +530,7 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq) SPLIT_NS(cfs_rq->exec_clock)); raw_spin_lock_irqsave(&rq->lock, flags); - if (cfs_rq->rb_leftmost) + if (rb_first_cached(&cfs_rq->tasks_timeline)) MIN_vruntime = (__pick_first_entity(cfs_rq))->vruntime; last = __pick_last_entity(cfs_rq); if (last) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 8bc0a883d190..a5d83ed8dd82 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -513,6 +513,7 @@ static inline int entity_before(struct sched_entity *a, static void update_min_vruntime(struct cfs_rq *cfs_rq) { struct sched_entity *curr = cfs_rq->curr; + struct rb_node *leftmost = rb_first_cached(&cfs_rq->tasks_timeline); u64 vruntime = cfs_rq->min_vruntime; @@ -523,10 +524,9 @@ static void update_min_vruntime(struct cfs_rq *cfs_rq) curr = NULL; } - if (cfs_rq->rb_leftmost) { - struct sched_entity *se = rb_entry(cfs_rq->rb_leftmost, - struct sched_entity, - run_node); + if (leftmost) { /* non-empty tree */ + struct sched_entity *se; + se = rb_entry(leftmost, struct sched_entity, run_node); if (!curr) vruntime = se->vruntime; @@ -547,10 +547,10 @@ static void update_min_vruntime(struct cfs_rq *cfs_rq) */ static void __enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se) { - struct rb_node **link = &cfs_rq->tasks_timeline.rb_node; + struct rb_node **link = &cfs_rq->tasks_timeline.rb_root.rb_node; struct rb_node *parent = NULL; struct sched_entity *entry; - int leftmost = 1; + bool leftmost = true; /* * Find the right place in the rbtree: @@ -566,36 +566,23 @@ static void __enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se) link = &parent->rb_left; } else { link = &parent->rb_right; - leftmost = 0; + leftmost = false; } } - /* - * Maintain a cache of leftmost tree entries (it is frequently - * used): - */ - if (leftmost) - cfs_rq->rb_leftmost = &se->run_node; - rb_link_node(&se->run_node, parent, link); - rb_insert_color(&se->run_node, &cfs_rq->tasks_timeline); + rb_insert_color_cached(&se->run_node, + &cfs_rq->tasks_timeline, leftmost); } static void __dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se) { - if (cfs_rq->rb_leftmost == &se->run_node) { - struct rb_node *next_node; - - next_node = rb_next(&se->run_node); - cfs_rq->rb_leftmost = next_node; - } - - rb_erase(&se->run_node, &cfs_rq->tasks_timeline); + rb_erase_cached(&se->run_node, &cfs_rq->tasks_timeline); } struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq) { - struct rb_node *left = cfs_rq->rb_leftmost; + struct rb_node *left = rb_first_cached(&cfs_rq->tasks_timeline); if (!left) return NULL; @@ -616,7 +603,7 @@ static struct sched_entity *__pick_next_entity(struct sched_entity *se) #ifdef CONFIG_SCHED_DEBUG struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq) { - struct rb_node *last = rb_last(&cfs_rq->tasks_timeline); + struct rb_node *last = rb_last(&cfs_rq->tasks_timeline.rb_root); if (!last) return NULL; @@ -9312,7 +9299,7 @@ static void set_curr_task_fair(struct rq *rq) void init_cfs_rq(struct cfs_rq *cfs_rq) { - cfs_rq->tasks_timeline = RB_ROOT; + cfs_rq->tasks_timeline = RB_ROOT_CACHED; cfs_rq->min_vruntime = (u64)(-(1LL << 20)); #ifndef CONFIG_64BIT cfs_rq->min_vruntime_copy = cfs_rq->min_vruntime; diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 6ed7962dc896..c30c57563dbc 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -426,8 +426,7 @@ struct cfs_rq { u64 min_vruntime_copy; #endif - struct rb_root tasks_timeline; - struct rb_node *rb_leftmost; + struct rb_root_cached tasks_timeline; /* * 'curr' points to currently running entity on this cfs_rq. -- cgit v1.2.3 From 9469eb01db891b55367ee7539f1b9f7f6fd2819d Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 7 Sep 2017 17:03:53 +0200 Subject: sched/debug: Add debugfs knob for "sched_debug" I'm forever late for editing my kernel cmdline, add a runtime knob to disable the "sched_debug" thing. Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/20170907150614.142924283@infradead.org Signed-off-by: Ingo Molnar --- kernel/sched/debug.c | 5 +++++ kernel/sched/sched.h | 2 ++ kernel/sched/topology.c | 4 +--- 3 files changed, 8 insertions(+), 3 deletions(-) (limited to 'kernel/sched/debug.c') diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index 4a23bbc3111b..b19d06ea6e10 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -181,11 +181,16 @@ static const struct file_operations sched_feat_fops = { .release = single_release, }; +__read_mostly bool sched_debug_enabled; + static __init int sched_init_debug(void) { debugfs_create_file("sched_features", 0644, NULL, NULL, &sched_feat_fops); + debugfs_create_bool("sched_debug", 0644, NULL, + &sched_debug_enabled); + return 0; } late_initcall(sched_init_debug); diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index ab1c7f5409a0..7ea2a0339771 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -1954,6 +1954,8 @@ extern struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq); extern struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq); #ifdef CONFIG_SCHED_DEBUG +extern bool sched_debug_enabled; + extern void print_cfs_stats(struct seq_file *m, int cpu); extern void print_rt_stats(struct seq_file *m, int cpu); extern void print_dl_stats(struct seq_file *m, int cpu); diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c index 6f7b43982f73..2ab2aa68c796 100644 --- a/kernel/sched/topology.c +++ b/kernel/sched/topology.c @@ -14,11 +14,9 @@ cpumask_var_t sched_domains_tmpmask2; #ifdef CONFIG_SCHED_DEBUG -static __read_mostly int sched_debug_enabled; - static int __init sched_debug_setup(char *str) { - sched_debug_enabled = 1; + sched_debug_enabled = true; return 0; } -- cgit v1.2.3