[deliverable/linux.git] / kernel / sched / debug.c

/*
 * kernel/sched/debug.c
 *
 * Print the CFS rbtree
 *
 * Copyright(C) 2007, Red Hat, Inc., Ingo Molnar
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */

#include <linux/proc_fs.h>
#include <linux/sched.h>
#include <linux/seq_file.h>
#include <linux/kallsyms.h>
#include <linux/utsname.h>
#include <linux/mempolicy.h>

#include "sched.h"

static DEFINE_SPINLOCK(sched_debug_lock);

/*
 * This allows printing both to /proc/sched_debug and
 * to the console
 */
#define SEQ_printf(m, x...)			\
 do {						\
	if (m)					\
		seq_printf(m, x);		\
	else					\
		printk(x);			\
 } while (0)

/*
 * Ease the printing of nsec fields:
 */
static long long nsec_high(unsigned long long nsec)
{
	if ((long long)nsec < 0) {
		nsec = -nsec;
		do_div(nsec, 1000000);
		return -nsec;
	}
	do_div(nsec, 1000000);

	return nsec;
}

static unsigned long nsec_low(unsigned long long nsec)
{
	if ((long long)nsec < 0)
		nsec = -nsec;

	return do_div(nsec, 1000000);
}

#define SPLIT_NS(x) nsec_high(x), nsec_low(x)

#ifdef CONFIG_FAIR_GROUP_SCHED
static void print_cfs_group_stats(struct seq_file *m, int cpu, struct task_group *tg)
{
	struct sched_entity *se = tg->se[cpu];

#define P(F) \
	SEQ_printf(m, "  .%-30s: %lld\n", #F, (long long)F)
#define PN(F) \
	SEQ_printf(m, "  .%-30s: %lld.%06ld\n", #F, SPLIT_NS((long long)F))

	if (!se) {
		struct sched_avg *avg = &cpu_rq(cpu)->avg;
		P(avg->runnable_avg_sum);
		P(avg->avg_period);
		return;
	}


	PN(se->exec_start);
	PN(se->vruntime);
	PN(se->sum_exec_runtime);
#ifdef CONFIG_SCHEDSTATS
	PN(se->statistics.wait_start);
	PN(se->statistics.sleep_start);
	PN(se->statistics.block_start);
	PN(se->statistics.sleep_max);
	PN(se->statistics.block_max);
	PN(se->statistics.exec_max);
	PN(se->statistics.slice_max);
	PN(se->statistics.wait_max);
	PN(se->statistics.wait_sum);
	P(se->statistics.wait_count);
#endif
	P(se->load.weight);
#ifdef CONFIG_SMP
	P(se->avg.runnable_avg_sum);
	P(se->avg.running_avg_sum);
	P(se->avg.avg_period);
	P(se->avg.load_avg_contrib);
	P(se->avg.utilization_avg_contrib);
	P(se->avg.decay_count);
#endif
#undef PN
#undef P
}
#endif

#ifdef CONFIG_CGROUP_SCHED
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;

	return cgroup_path(tg->css.cgroup, group_path, PATH_MAX);
}
#endif

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, " ");

	SEQ_printf(m, "%15s %5d %9Ld.%06ld %9Ld %5d ",
		p->comm, task_pid_nr(p),
		SPLIT_NS(p->se.vruntime),
		(long long)(p->nvcsw + p->nivcsw),
		p->prio);
#ifdef CONFIG_SCHEDSTATS
	SEQ_printf(m, "%9Ld.%06ld %9Ld.%06ld %9Ld.%06ld",
		SPLIT_NS(p->se.statistics.wait_sum),
		SPLIT_NS(p->se.sum_exec_runtime),
		SPLIT_NS(p->se.statistics.sum_sleep_runtime));
#else
	SEQ_printf(m, "%9Ld.%06ld %9Ld.%06ld %9Ld.%06ld",
		0LL, 0L,
		SPLIT_NS(p->se.sum_exec_runtime),
		0LL, 0L);
#endif
#ifdef CONFIG_NUMA_BALANCING
	SEQ_printf(m, " %d %d", task_node(p), task_numa_group_id(p));
#endif
#ifdef CONFIG_CGROUP_SCHED
	SEQ_printf(m, " %s", task_group_path(task_group(p)));
#endif

	SEQ_printf(m, "\n");
}

static void print_rq(struct seq_file *m, struct rq *rq, int rq_cpu)
{
	struct task_struct *g, *p;

	SEQ_printf(m,
	"\nrunnable tasks:\n"
	"            task   PID         tree-key  switches  prio"
	"     wait-time             sum-exec        sum-sleep\n"
	"------------------------------------------------------"
	"----------------------------------------------------\n");

	rcu_read_lock();
	for_each_process_thread(g, p) {
		if (task_cpu(p) != rq_cpu)
			continue;

		print_task(m, rq, p);
	}
	rcu_read_unlock();
}

void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
{
	s64 MIN_vruntime = -1, min_vruntime, max_vruntime = -1,
		spread, rq0_min_vruntime, spread0;
	struct rq *rq = cpu_rq(cpu);
	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));

	raw_spin_lock_irqsave(&rq->lock, flags);
	if (cfs_rq->rb_leftmost)
		MIN_vruntime = (__pick_first_entity(cfs_rq))->vruntime;
	last = __pick_last_entity(cfs_rq);
	if (last)
		max_vruntime = last->vruntime;
	min_vruntime = cfs_rq->min_vruntime;
	rq0_min_vruntime = cpu_rq(0)->cfs.min_vruntime;
	raw_spin_unlock_irqrestore(&rq->lock, flags);
	SEQ_printf(m, "  .%-30s: %Ld.%06ld\n", "MIN_vruntime",
			SPLIT_NS(MIN_vruntime));
	SEQ_printf(m, "  .%-30s: %Ld.%06ld\n", "min_vruntime",
			SPLIT_NS(min_vruntime));
	SEQ_printf(m, "  .%-30s: %Ld.%06ld\n", "max_vruntime",
			SPLIT_NS(max_vruntime));
	spread = max_vruntime - MIN_vruntime;
	SEQ_printf(m, "  .%-30s: %Ld.%06ld\n", "spread",
			SPLIT_NS(spread));
	spread0 = min_vruntime - rq0_min_vruntime;
	SEQ_printf(m, "  .%-30s: %Ld.%06ld\n", "spread0",
			SPLIT_NS(spread0));
	SEQ_printf(m, "  .%-30s: %d\n", "nr_spread_over",
			cfs_rq->nr_spread_over);
	SEQ_printf(m, "  .%-30s: %d\n", "nr_running", cfs_rq->nr_running);
	SEQ_printf(m, "  .%-30s: %ld\n", "load", cfs_rq->load.weight);
#ifdef CONFIG_SMP
	SEQ_printf(m, "  .%-30s: %ld\n", "runnable_load_avg",
			cfs_rq->runnable_load_avg);
	SEQ_printf(m, "  .%-30s: %ld\n", "blocked_load_avg",
			cfs_rq->blocked_load_avg);
	SEQ_printf(m, "  .%-30s: %ld\n", "utilization_load_avg",
			cfs_rq->utilization_load_avg);
#ifdef CONFIG_FAIR_GROUP_SCHED
	SEQ_printf(m, "  .%-30s: %ld\n", "tg_load_contrib",
			cfs_rq->tg_load_contrib);
	SEQ_printf(m, "  .%-30s: %d\n", "tg_runnable_contrib",
			cfs_rq->tg_runnable_contrib);
	SEQ_printf(m, "  .%-30s: %ld\n", "tg_load_avg",
			atomic_long_read(&cfs_rq->tg->load_avg));
	SEQ_printf(m, "  .%-30s: %d\n", "tg->runnable_avg",
			atomic_read(&cfs_rq->tg->runnable_avg));
#endif
#endif
#ifdef CONFIG_CFS_BANDWIDTH
	SEQ_printf(m, "  .%-30s: %d\n", "throttled",
			cfs_rq->throttled);
	SEQ_printf(m, "  .%-30s: %d\n", "throttle_count",
			cfs_rq->throttle_count);
#endif

#ifdef CONFIG_FAIR_GROUP_SCHED
	print_cfs_group_stats(m, cpu, cfs_rq->tg);
#endif
}

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))
#define PN(x) \
	SEQ_printf(m, "  .%-30s: %Ld.%06ld\n", #x, SPLIT_NS(rt_rq->x))

	P(rt_nr_running);
	P(rt_throttled);
	PN(rt_time);
	PN(rt_runtime);

#undef PN
#undef P
}

void print_dl_rq(struct seq_file *m, int cpu, struct dl_rq *dl_rq)
{
	SEQ_printf(m, "\ndl_rq[%d]:\n", cpu);
	SEQ_printf(m, "  .%-30s: %ld\n", "dl_nr_running", dl_rq->dl_nr_running);
}

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
	{
		unsigned int freq = cpu_khz ? : 1;

		SEQ_printf(m, "cpu#%d, %u.%03u MHz\n",
			   cpu, freq / 1000, (freq % 1000));
	}
#else
	SEQ_printf(m, "cpu#%d\n", cpu);
#endif

#define P(x)								\
do {									\
	if (sizeof(rq->x) == 4)						\
		SEQ_printf(m, "  .%-30s: %ld\n", #x, (long)(rq->x));	\
	else								\
		SEQ_printf(m, "  .%-30s: %Ld\n", #x, (long long)(rq->x));\
} while (0)

#define PN(x) \
	SEQ_printf(m, "  .%-30s: %Ld.%06ld\n", #x, SPLIT_NS(rq->x))

	P(nr_running);
	SEQ_printf(m, "  .%-30s: %lu\n", "load",
		   rq->load.weight);
	P(nr_switches);
	P(nr_load_updates);
	P(nr_uninterruptible);
	PN(next_balance);
	SEQ_printf(m, "  .%-30s: %ld\n", "curr->pid", (long)(task_pid_nr(rq->curr)));
	PN(clock);
	PN(clock_task);
	P(cpu_load[0]);
	P(cpu_load[1]);
	P(cpu_load[2]);
	P(cpu_load[3]);
	P(cpu_load[4]);
#undef P
#undef PN

#ifdef CONFIG_SCHEDSTATS
#define P(n) SEQ_printf(m, "  .%-30s: %d\n", #n, rq->n);
#define P64(n) SEQ_printf(m, "  .%-30s: %Ld\n", #n, rq->n);

	P(yld_count);

	P(sched_count);
	P(sched_goidle);
#ifdef CONFIG_SMP
	P64(avg_idle);
	P64(max_idle_balance_cost);
#endif

	P(ttwu_count);
	P(ttwu_local);

#undef P
#undef P64
#endif
	spin_lock_irqsave(&sched_debug_lock, flags);
	print_cfs_stats(m, cpu);
	print_rt_stats(m, cpu);
	print_dl_stats(m, cpu);

	print_rq(m, rq, cpu);
	spin_unlock_irqrestore(&sched_debug_lock, flags);
	SEQ_printf(m, "\n");
}

static const char *sched_tunable_scaling_names[] = {
	"none",
	"logaritmic",
	"linear"
};

static void sched_debug_header(struct seq_file *m)
{
	u64 ktime, sched_clk, cpu_clk;
	unsigned long flags;

	local_irq_save(flags);
	ktime = ktime_to_ns(ktime_get());
	sched_clk = sched_clock();
	cpu_clk = local_clock();
	local_irq_restore(flags);

	SEQ_printf(m, "Sched Debug Version: v0.11, %s %.*s\n",
		init_utsname()->release,
		(int)strcspn(init_utsname()->version, " "),
		init_utsname()->version);

#define P(x) \
	SEQ_printf(m, "%-40s: %Ld\n", #x, (long long)(x))
#define PN(x) \
	SEQ_printf(m, "%-40s: %Ld.%06ld\n", #x, SPLIT_NS(x))
	PN(ktime);
	PN(sched_clk);
	PN(cpu_clk);
	P(jiffies);
#ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
	P(sched_clock_stable());
#endif
#undef PN
#undef P

	SEQ_printf(m, "\n");
	SEQ_printf(m, "sysctl_sched\n");

#define P(x) \
	SEQ_printf(m, "  .%-40s: %Ld\n", #x, (long long)(x))
#define PN(x) \
	SEQ_printf(m, "  .%-40s: %Ld.%06ld\n", #x, SPLIT_NS(x))
	PN(sysctl_sched_latency);
	PN(sysctl_sched_min_granularity);
	PN(sysctl_sched_wakeup_granularity);
	P(sysctl_sched_child_runs_first);
	P(sysctl_sched_features);
#undef PN
#undef P

	SEQ_printf(m, "  .%-40s: %d (%s)\n",
		"sysctl_sched_tunable_scaling",
		sysctl_sched_tunable_scaling,
		sched_tunable_scaling_names[sysctl_sched_tunable_scaling]);
	SEQ_printf(m, "\n");
}

static int sched_debug_show(struct seq_file *m, void *v)
{
	int cpu = (unsigned long)(v - 2);

	if (cpu != -1)
		print_cpu(m, cpu);
	else
		sched_debug_header(m);

	return 0;
}

void sysrq_sched_debug_show(void)
{
	int cpu;

	sched_debug_header(NULL);
	for_each_online_cpu(cpu)
		print_cpu(NULL, cpu);

}

/*
 * This itererator needs some explanation.
 * It returns 1 for the header position.
 * This means 2 is cpu 0.
 * In a hotplugged system some cpus, including cpu 0, may be missing so we have
 * to use cpumask_* to iterate over the cpus.
 */
static void *sched_debug_start(struct seq_file *file, loff_t *offset)
{
	unsigned long n = *offset;

	if (n == 0)
		return (void *) 1;

	n--;

	if (n > 0)
		n = cpumask_next(n - 1, cpu_online_mask);
	else
		n = cpumask_first(cpu_online_mask);

	*offset = n + 1;

	if (n < nr_cpu_ids)
		return (void *)(unsigned long)(n + 2);
	return NULL;
}

static void *sched_debug_next(struct seq_file *file, void *data, loff_t *offset)
{
	(*offset)++;
	return sched_debug_start(file, offset);
}

static void sched_debug_stop(struct seq_file *file, void *data)
{
}

static const struct seq_operations sched_debug_sops = {
	.start = sched_debug_start,
	.next = sched_debug_next,
	.stop = sched_debug_stop,
	.show = sched_debug_show,
};

static int sched_debug_release(struct inode *inode, struct file *file)
{
	seq_release(inode, file);

	return 0;
}

static int sched_debug_open(struct inode *inode, struct file *filp)
{
	int ret = 0;

	ret = seq_open(filp, &sched_debug_sops);

	return ret;
}

static const struct file_operations sched_debug_fops = {
	.open		= sched_debug_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= sched_debug_release,
};

static int __init init_sched_debug_procfs(void)
{
	struct proc_dir_entry *pe;

	pe = proc_create("sched_debug", 0444, NULL, &sched_debug_fops);
	if (!pe)
		return -ENOMEM;
	return 0;
}

__initcall(init_sched_debug_procfs);

#define __P(F) \
	SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)F)
#define P(F) \
	SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)p->F)
#define __PN(F) \
	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)F))
#define PN(F) \
	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F))


#ifdef CONFIG_NUMA_BALANCING
void print_numa_stats(struct seq_file *m, int node, unsigned long tsf,
		unsigned long tpf, unsigned long gsf, unsigned long gpf)
{
	SEQ_printf(m, "numa_faults node=%d ", node);
	SEQ_printf(m, "task_private=%lu task_shared=%lu ", tsf, tpf);
	SEQ_printf(m, "group_private=%lu group_shared=%lu\n", gsf, gpf);
}
#endif


static void sched_show_numa(struct task_struct *p, struct seq_file *m)
{
#ifdef CONFIG_NUMA_BALANCING
	struct mempolicy *pol;

	if (p->mm)
		P(mm->numa_scan_seq);

	task_lock(p);
	pol = p->mempolicy;
	if (pol && !(pol->flags & MPOL_F_MORON))
		pol = NULL;
	mpol_get(pol);
	task_unlock(p);

	P(numa_pages_migrated);
	P(numa_preferred_nid);
	P(total_numa_faults);
	SEQ_printf(m, "current_node=%d, numa_group_id=%d\n",
			task_node(p), task_numa_group_id(p));
	show_numa_stats(p, m);
	mpol_put(pol);
#endif
}

void proc_sched_show_task(struct task_struct *p, struct seq_file *m)
{
	unsigned long nr_switches;

	SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr(p),
						get_nr_threads(p));
	SEQ_printf(m,
		"---------------------------------------------------------"
		"----------\n");
#define __P(F) \
	SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)F)
#define P(F) \
	SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)p->F)
#define __PN(F) \
	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)F))
#define PN(F) \
	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F))

	PN(se.exec_start);
	PN(se.vruntime);
	PN(se.sum_exec_runtime);

	nr_switches = p->nvcsw + p->nivcsw;

#ifdef CONFIG_SCHEDSTATS
	PN(se.statistics.sum_sleep_runtime);
	PN(se.statistics.wait_start);
	PN(se.statistics.sleep_start);
	PN(se.statistics.block_start);
	PN(se.statistics.sleep_max);
	PN(se.statistics.block_max);
	PN(se.statistics.exec_max);
	PN(se.statistics.slice_max);
	PN(se.statistics.wait_max);
	PN(se.statistics.wait_sum);
	P(se.statistics.wait_count);
	PN(se.statistics.iowait_sum);
	P(se.statistics.iowait_count);
	P(se.nr_migrations);
	P(se.statistics.nr_migrations_cold);
	P(se.statistics.nr_failed_migrations_affine);
	P(se.statistics.nr_failed_migrations_running);
	P(se.statistics.nr_failed_migrations_hot);
	P(se.statistics.nr_forced_migrations);
	P(se.statistics.nr_wakeups);
	P(se.statistics.nr_wakeups_sync);
	P(se.statistics.nr_wakeups_migrate);
	P(se.statistics.nr_wakeups_local);
	P(se.statistics.nr_wakeups_remote);
	P(se.statistics.nr_wakeups_affine);
	P(se.statistics.nr_wakeups_affine_attempts);
	P(se.statistics.nr_wakeups_passive);
	P(se.statistics.nr_wakeups_idle);

	{
		u64 avg_atom, avg_per_cpu;

		avg_atom = p->se.sum_exec_runtime;
		if (nr_switches)
			avg_atom = div64_ul(avg_atom, nr_switches);
		else
			avg_atom = -1LL;

		avg_per_cpu = p->se.sum_exec_runtime;
		if (p->se.nr_migrations) {
			avg_per_cpu = div64_u64(avg_per_cpu,
						p->se.nr_migrations);
		} else {
			avg_per_cpu = -1LL;
		}

		__PN(avg_atom);
		__PN(avg_per_cpu);
	}
#endif
	__P(nr_switches);
	SEQ_printf(m, "%-45s:%21Ld\n",
		   "nr_voluntary_switches", (long long)p->nvcsw);
	SEQ_printf(m, "%-45s:%21Ld\n",
		   "nr_involuntary_switches", (long long)p->nivcsw);

	P(se.load.weight);
#ifdef CONFIG_SMP
	P(se.avg.runnable_avg_sum);
	P(se.avg.running_avg_sum);
	P(se.avg.avg_period);
	P(se.avg.load_avg_contrib);
	P(se.avg.utilization_avg_contrib);
	P(se.avg.decay_count);
#endif
	P(policy);
	P(prio);
#undef PN
#undef __PN
#undef P
#undef __P

	{
		unsigned int this_cpu = raw_smp_processor_id();
		u64 t0, t1;

		t0 = cpu_clock(this_cpu);
		t1 = cpu_clock(this_cpu);
		SEQ_printf(m, "%-45s:%21Ld\n",
			   "clock-delta", (long long)(t1-t0));
	}

	sched_show_numa(p, m);
}

void proc_sched_set_task(struct task_struct *p)
{
#ifdef CONFIG_SCHEDSTATS
	memset(&p->se.statistics, 0, sizeof(p->se.statistics));
#endif
}
Commit	Line	Data
	1	/*
	2	* kernel/sched/debug.c
	3	*
	4	* Print the CFS rbtree
	5	*
	6	* Copyright(C) 2007, Red Hat, Inc., Ingo Molnar
	7	*
	8	* This program is free software; you can redistribute it and/or modify
	9	* it under the terms of the GNU General Public License version 2 as
	10	* published by the Free Software Foundation.
	11	*/
	12
	13	#include <linux/proc_fs.h>
	14	#include <linux/sched.h>
	15	#include <linux/seq_file.h>
	16	#include <linux/kallsyms.h>
	17	#include <linux/utsname.h>
	18	#include <linux/mempolicy.h>
	19
	20	#include "sched.h"
	21
	22	static DEFINE_SPINLOCK(sched_debug_lock);
	23
	24	/*
	25	* This allows printing both to /proc/sched_debug and
	26	* to the console
	27	*/
	28	#define SEQ_printf(m, x...) \
	29	do { \
	30	if (m) \
	31	seq_printf(m, x); \
	32	else \
	33	printk(x); \
	34	} while (0)
	35
	36	/*
	37	* Ease the printing of nsec fields:
	38	*/
	39	static long long nsec_high(unsigned long long nsec)
	40	{
	41	if ((long long)nsec < 0) {
	42	nsec = -nsec;
	43	do_div(nsec, 1000000);
	44	return -nsec;
	45	}
	46	do_div(nsec, 1000000);
	47
	48	return nsec;
	49	}
	50
	51	static unsigned long nsec_low(unsigned long long nsec)
	52	{
	53	if ((long long)nsec < 0)
	54	nsec = -nsec;
	55
	56	return do_div(nsec, 1000000);
	57	}
	58
	59	#define SPLIT_NS(x) nsec_high(x), nsec_low(x)
	60
	61	#ifdef CONFIG_FAIR_GROUP_SCHED
	62	static void print_cfs_group_stats(struct seq_file m, int cpu, struct task_group tg)
	63	{
	64	struct sched_entity *se = tg->se[cpu];
	65
	66	#define P(F) \
	67	SEQ_printf(m, " .%-30s: %lld\n", #F, (long long)F)
	68	#define PN(F) \
	69	SEQ_printf(m, " .%-30s: %lld.%06ld\n", #F, SPLIT_NS((long long)F))
	70
	71	if (!se) {
	72	struct sched_avg *avg = &cpu_rq(cpu)->avg;
	73	P(avg->runnable_avg_sum);
	74	P(avg->avg_period);
	75	return;
	76	}
	77
	78
	79	PN(se->exec_start);
	80	PN(se->vruntime);
	81	PN(se->sum_exec_runtime);
	82	#ifdef CONFIG_SCHEDSTATS
	83	PN(se->statistics.wait_start);
	84	PN(se->statistics.sleep_start);
	85	PN(se->statistics.block_start);
	86	PN(se->statistics.sleep_max);
	87	PN(se->statistics.block_max);
	88	PN(se->statistics.exec_max);
	89	PN(se->statistics.slice_max);
	90	PN(se->statistics.wait_max);
	91	PN(se->statistics.wait_sum);
	92	P(se->statistics.wait_count);
	93	#endif
	94	P(se->load.weight);
	95	#ifdef CONFIG_SMP
	96	P(se->avg.runnable_avg_sum);
	97	P(se->avg.running_avg_sum);
	98	P(se->avg.avg_period);
	99	P(se->avg.load_avg_contrib);
	100	P(se->avg.utilization_avg_contrib);
	101	P(se->avg.decay_count);
	102	#endif
	103	#undef PN
	104	#undef P
	105	}
	106	#endif
	107
	108	#ifdef CONFIG_CGROUP_SCHED
	109	static char group_path[PATH_MAX];
	110
	111	static char task_group_path(struct task_group tg)
	112	{
	113	if (autogroup_path(tg, group_path, PATH_MAX))
	114	return group_path;
	115
	116	return cgroup_path(tg->css.cgroup, group_path, PATH_MAX);
	117	}
	118	#endif
	119
	120	static void
	121	print_task(struct seq_file m, struct rq rq, struct task_struct *p)
	122	{
	123	if (rq->curr == p)
	124	SEQ_printf(m, "R");
	125	else
	126	SEQ_printf(m, " ");
	127
	128	SEQ_printf(m, "%15s %5d %9Ld.%06ld %9Ld %5d ",
	129	p->comm, task_pid_nr(p),
	130	SPLIT_NS(p->se.vruntime),
	131	(long long)(p->nvcsw + p->nivcsw),
	132	p->prio);
	133	#ifdef CONFIG_SCHEDSTATS
	134	SEQ_printf(m, "%9Ld.%06ld %9Ld.%06ld %9Ld.%06ld",
	135	SPLIT_NS(p->se.statistics.wait_sum),
	136	SPLIT_NS(p->se.sum_exec_runtime),
	137	SPLIT_NS(p->se.statistics.sum_sleep_runtime));
	138	#else
	139	SEQ_printf(m, "%9Ld.%06ld %9Ld.%06ld %9Ld.%06ld",
	140	0LL, 0L,
	141	SPLIT_NS(p->se.sum_exec_runtime),
	142	0LL, 0L);
	143	#endif
	144	#ifdef CONFIG_NUMA_BALANCING
	145	SEQ_printf(m, " %d %d", task_node(p), task_numa_group_id(p));
	146	#endif
	147	#ifdef CONFIG_CGROUP_SCHED
	148	SEQ_printf(m, " %s", task_group_path(task_group(p)));
	149	#endif
	150
	151	SEQ_printf(m, "\n");
	152	}
	153
	154	static void print_rq(struct seq_file m, struct rq rq, int rq_cpu)
	155	{
	156	struct task_struct g, p;
	157
	158	SEQ_printf(m,
	159	"\nrunnable tasks:\n"
	160	" task PID tree-key switches prio"
	161	" wait-time sum-exec sum-sleep\n"
	162	"------------------------------------------------------"
	163	"----------------------------------------------------\n");
	164
	165	rcu_read_lock();
	166	for_each_process_thread(g, p) {
	167	if (task_cpu(p) != rq_cpu)
	168	continue;
	169
	170	print_task(m, rq, p);
	171	}
	172	rcu_read_unlock();
	173	}
	174
	175	void print_cfs_rq(struct seq_file m, int cpu, struct cfs_rq cfs_rq)
	176	{
	177	s64 MIN_vruntime = -1, min_vruntime, max_vruntime = -1,
	178	spread, rq0_min_vruntime, spread0;
	179	struct rq *rq = cpu_rq(cpu);
	180	struct sched_entity *last;
	181	unsigned long flags;
	182
	183	#ifdef CONFIG_FAIR_GROUP_SCHED
	184	SEQ_printf(m, "\ncfs_rq[%d]:%s\n", cpu, task_group_path(cfs_rq->tg));
	185	#else
	186	SEQ_printf(m, "\ncfs_rq[%d]:\n", cpu);
	187	#endif
	188	SEQ_printf(m, " .%-30s: %Ld.%06ld\n", "exec_clock",
	189	SPLIT_NS(cfs_rq->exec_clock));
	190
	191	raw_spin_lock_irqsave(&rq->lock, flags);
	192	if (cfs_rq->rb_leftmost)
	193	MIN_vruntime = (__pick_first_entity(cfs_rq))->vruntime;
	194	last = __pick_last_entity(cfs_rq);
	195	if (last)
	196	max_vruntime = last->vruntime;
	197	min_vruntime = cfs_rq->min_vruntime;
	198	rq0_min_vruntime = cpu_rq(0)->cfs.min_vruntime;
	199	raw_spin_unlock_irqrestore(&rq->lock, flags);
	200	SEQ_printf(m, " .%-30s: %Ld.%06ld\n", "MIN_vruntime",
	201	SPLIT_NS(MIN_vruntime));
	202	SEQ_printf(m, " .%-30s: %Ld.%06ld\n", "min_vruntime",
	203	SPLIT_NS(min_vruntime));
	204	SEQ_printf(m, " .%-30s: %Ld.%06ld\n", "max_vruntime",
	205	SPLIT_NS(max_vruntime));
	206	spread = max_vruntime - MIN_vruntime;
	207	SEQ_printf(m, " .%-30s: %Ld.%06ld\n", "spread",
	208	SPLIT_NS(spread));
	209	spread0 = min_vruntime - rq0_min_vruntime;
	210	SEQ_printf(m, " .%-30s: %Ld.%06ld\n", "spread0",
	211	SPLIT_NS(spread0));
	212	SEQ_printf(m, " .%-30s: %d\n", "nr_spread_over",
	213	cfs_rq->nr_spread_over);
	214	SEQ_printf(m, " .%-30s: %d\n", "nr_running", cfs_rq->nr_running);
	215	SEQ_printf(m, " .%-30s: %ld\n", "load", cfs_rq->load.weight);
	216	#ifdef CONFIG_SMP
	217	SEQ_printf(m, " .%-30s: %ld\n", "runnable_load_avg",
	218	cfs_rq->runnable_load_avg);
	219	SEQ_printf(m, " .%-30s: %ld\n", "blocked_load_avg",
	220	cfs_rq->blocked_load_avg);
	221	SEQ_printf(m, " .%-30s: %ld\n", "utilization_load_avg",
	222	cfs_rq->utilization_load_avg);
	223	#ifdef CONFIG_FAIR_GROUP_SCHED
	224	SEQ_printf(m, " .%-30s: %ld\n", "tg_load_contrib",
	225	cfs_rq->tg_load_contrib);
	226	SEQ_printf(m, " .%-30s: %d\n", "tg_runnable_contrib",
	227	cfs_rq->tg_runnable_contrib);
	228	SEQ_printf(m, " .%-30s: %ld\n", "tg_load_avg",
	229	atomic_long_read(&cfs_rq->tg->load_avg));
	230	SEQ_printf(m, " .%-30s: %d\n", "tg->runnable_avg",
	231	atomic_read(&cfs_rq->tg->runnable_avg));
	232	#endif
	233	#endif
	234	#ifdef CONFIG_CFS_BANDWIDTH
	235	SEQ_printf(m, " .%-30s: %d\n", "throttled",
	236	cfs_rq->throttled);
	237	SEQ_printf(m, " .%-30s: %d\n", "throttle_count",
	238	cfs_rq->throttle_count);
	239	#endif
	240
	241	#ifdef CONFIG_FAIR_GROUP_SCHED
	242	print_cfs_group_stats(m, cpu, cfs_rq->tg);
	243	#endif
	244	}
	245
	246	void print_rt_rq(struct seq_file m, int cpu, struct rt_rq rt_rq)
	247	{
	248	#ifdef CONFIG_RT_GROUP_SCHED
	249	SEQ_printf(m, "\nrt_rq[%d]:%s\n", cpu, task_group_path(rt_rq->tg));
	250	#else
	251	SEQ_printf(m, "\nrt_rq[%d]:\n", cpu);
	252	#endif
	253
	254	#define P(x) \
	255	SEQ_printf(m, " .%-30s: %Ld\n", #x, (long long)(rt_rq->x))
	256	#define PN(x) \
	257	SEQ_printf(m, " .%-30s: %Ld.%06ld\n", #x, SPLIT_NS(rt_rq->x))
	258
	259	P(rt_nr_running);
	260	P(rt_throttled);
	261	PN(rt_time);
	262	PN(rt_runtime);
	263
	264	#undef PN
	265	#undef P
	266	}
	267
	268	void print_dl_rq(struct seq_file m, int cpu, struct dl_rq dl_rq)
	269	{
	270	SEQ_printf(m, "\ndl_rq[%d]:\n", cpu);
	271	SEQ_printf(m, " .%-30s: %ld\n", "dl_nr_running", dl_rq->dl_nr_running);
	272	}
	273
	274	extern __read_mostly int sched_clock_running;
	275
	276	static void print_cpu(struct seq_file *m, int cpu)
	277	{
	278	struct rq *rq = cpu_rq(cpu);
	279	unsigned long flags;
	280
	281	#ifdef CONFIG_X86
	282	{
	283	unsigned int freq = cpu_khz ? : 1;
	284
	285	SEQ_printf(m, "cpu#%d, %u.%03u MHz\n",
	286	cpu, freq / 1000, (freq % 1000));
	287	}
	288	#else
	289	SEQ_printf(m, "cpu#%d\n", cpu);
	290	#endif
	291
	292	#define P(x) \
	293	do { \
	294	if (sizeof(rq->x) == 4) \
	295	SEQ_printf(m, " .%-30s: %ld\n", #x, (long)(rq->x)); \
	296	else \
	297	SEQ_printf(m, " .%-30s: %Ld\n", #x, (long long)(rq->x));\
	298	} while (0)
	299
	300	#define PN(x) \
	301	SEQ_printf(m, " .%-30s: %Ld.%06ld\n", #x, SPLIT_NS(rq->x))
	302
	303	P(nr_running);
	304	SEQ_printf(m, " .%-30s: %lu\n", "load",
	305	rq->load.weight);
	306	P(nr_switches);
	307	P(nr_load_updates);
	308	P(nr_uninterruptible);
	309	PN(next_balance);
	310	SEQ_printf(m, " .%-30s: %ld\n", "curr->pid", (long)(task_pid_nr(rq->curr)));
	311	PN(clock);
	312	PN(clock_task);
	313	P(cpu_load[0]);
	314	P(cpu_load[1]);
	315	P(cpu_load[2]);
	316	P(cpu_load[3]);
	317	P(cpu_load[4]);
	318	#undef P
	319	#undef PN
	320
	321	#ifdef CONFIG_SCHEDSTATS
	322	#define P(n) SEQ_printf(m, " .%-30s: %d\n", #n, rq->n);
	323	#define P64(n) SEQ_printf(m, " .%-30s: %Ld\n", #n, rq->n);
	324
	325	P(yld_count);
	326
	327	P(sched_count);
	328	P(sched_goidle);
	329	#ifdef CONFIG_SMP
	330	P64(avg_idle);
	331	P64(max_idle_balance_cost);
	332	#endif
	333
	334	P(ttwu_count);
	335	P(ttwu_local);
	336
	337	#undef P
	338	#undef P64
	339	#endif
	340	spin_lock_irqsave(&sched_debug_lock, flags);
	341	print_cfs_stats(m, cpu);
	342	print_rt_stats(m, cpu);
	343	print_dl_stats(m, cpu);
	344
	345	print_rq(m, rq, cpu);
	346	spin_unlock_irqrestore(&sched_debug_lock, flags);
	347	SEQ_printf(m, "\n");
	348	}
	349
	350	static const char *sched_tunable_scaling_names[] = {
	351	"none",
	352	"logaritmic",
	353	"linear"
	354	};
	355
	356	static void sched_debug_header(struct seq_file *m)
	357	{
	358	u64 ktime, sched_clk, cpu_clk;
	359	unsigned long flags;
	360
	361	local_irq_save(flags);
	362	ktime = ktime_to_ns(ktime_get());
	363	sched_clk = sched_clock();
	364	cpu_clk = local_clock();
	365	local_irq_restore(flags);
	366
	367	SEQ_printf(m, "Sched Debug Version: v0.11, %s %.*s\n",
	368	init_utsname()->release,
	369	(int)strcspn(init_utsname()->version, " "),
	370	init_utsname()->version);
	371
	372	#define P(x) \
	373	SEQ_printf(m, "%-40s: %Ld\n", #x, (long long)(x))
	374	#define PN(x) \
	375	SEQ_printf(m, "%-40s: %Ld.%06ld\n", #x, SPLIT_NS(x))
	376	PN(ktime);
	377	PN(sched_clk);
	378	PN(cpu_clk);
	379	P(jiffies);
	380	#ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
	381	P(sched_clock_stable());
	382	#endif
	383	#undef PN
	384	#undef P
	385
	386	SEQ_printf(m, "\n");
	387	SEQ_printf(m, "sysctl_sched\n");
	388
	389	#define P(x) \
	390	SEQ_printf(m, " .%-40s: %Ld\n", #x, (long long)(x))
	391	#define PN(x) \
	392	SEQ_printf(m, " .%-40s: %Ld.%06ld\n", #x, SPLIT_NS(x))
	393	PN(sysctl_sched_latency);
	394	PN(sysctl_sched_min_granularity);
	395	PN(sysctl_sched_wakeup_granularity);
	396	P(sysctl_sched_child_runs_first);
	397	P(sysctl_sched_features);
	398	#undef PN
	399	#undef P
	400
	401	SEQ_printf(m, " .%-40s: %d (%s)\n",
	402	"sysctl_sched_tunable_scaling",
	403	sysctl_sched_tunable_scaling,
	404	sched_tunable_scaling_names[sysctl_sched_tunable_scaling]);
	405	SEQ_printf(m, "\n");
	406	}
	407
	408	static int sched_debug_show(struct seq_file m, void v)
	409	{
	410	int cpu = (unsigned long)(v - 2);
	411
	412	if (cpu != -1)
	413	print_cpu(m, cpu);
	414	else
	415	sched_debug_header(m);
	416
	417	return 0;
	418	}
	419
	420	void sysrq_sched_debug_show(void)
	421	{
	422	int cpu;
	423
	424	sched_debug_header(NULL);
	425	for_each_online_cpu(cpu)
	426	print_cpu(NULL, cpu);
	427
	428	}
	429
	430	/*
	431	* This itererator needs some explanation.
	432	* It returns 1 for the header position.
	433	* This means 2 is cpu 0.
	434	* In a hotplugged system some cpus, including cpu 0, may be missing so we have
	435	* to use cpumask_* to iterate over the cpus.
	436	*/
	437	static void sched_debug_start(struct seq_file file, loff_t *offset)
	438	{
	439	unsigned long n = *offset;
	440
	441	if (n == 0)
	442	return (void *) 1;
	443
	444	n--;
	445
	446	if (n > 0)
	447	n = cpumask_next(n - 1, cpu_online_mask);
	448	else
	449	n = cpumask_first(cpu_online_mask);
	450
	451	*offset = n + 1;
	452
	453	if (n < nr_cpu_ids)
	454	return (void *)(unsigned long)(n + 2);
	455	return NULL;
	456	}
	457
	458	static void sched_debug_next(struct seq_file file, void data, loff_t offset)
	459	{
	460	(*offset)++;
	461	return sched_debug_start(file, offset);
	462	}
	463
	464	static void sched_debug_stop(struct seq_file file, void data)
	465	{
	466	}
	467
	468	static const struct seq_operations sched_debug_sops = {
	469	.start = sched_debug_start,
	470	.next = sched_debug_next,
	471	.stop = sched_debug_stop,
	472	.show = sched_debug_show,
	473	};
	474
	475	static int sched_debug_release(struct inode inode, struct file file)
	476	{
	477	seq_release(inode, file);
	478
	479	return 0;
	480	}
	481
	482	static int sched_debug_open(struct inode inode, struct file filp)
	483	{
	484	int ret = 0;
	485
	486	ret = seq_open(filp, &sched_debug_sops);
	487
	488	return ret;
	489	}
	490
	491	static const struct file_operations sched_debug_fops = {
	492	.open = sched_debug_open,
	493	.read = seq_read,
	494	.llseek = seq_lseek,
	495	.release = sched_debug_release,
	496	};
	497
	498	static int __init init_sched_debug_procfs(void)
	499	{
	500	struct proc_dir_entry *pe;
	501
	502	pe = proc_create("sched_debug", 0444, NULL, &sched_debug_fops);
	503	if (!pe)
	504	return -ENOMEM;
	505	return 0;
	506	}
	507
	508	__initcall(init_sched_debug_procfs);
	509
	510	#define __P(F) \
	511	SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)F)
	512	#define P(F) \
	513	SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)p->F)
	514	#define __PN(F) \
	515	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)F))
	516	#define PN(F) \
	517	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F))
	518
	519
	520	#ifdef CONFIG_NUMA_BALANCING
	521	void print_numa_stats(struct seq_file *m, int node, unsigned long tsf,
	522	unsigned long tpf, unsigned long gsf, unsigned long gpf)
	523	{
	524	SEQ_printf(m, "numa_faults node=%d ", node);
	525	SEQ_printf(m, "task_private=%lu task_shared=%lu ", tsf, tpf);
	526	SEQ_printf(m, "group_private=%lu group_shared=%lu\n", gsf, gpf);
	527	}
	528	#endif
	529
	530
	531	static void sched_show_numa(struct task_struct p, struct seq_file m)
	532	{
	533	#ifdef CONFIG_NUMA_BALANCING
	534	struct mempolicy *pol;
	535
	536	if (p->mm)
	537	P(mm->numa_scan_seq);
	538
	539	task_lock(p);
	540	pol = p->mempolicy;
	541	if (pol && !(pol->flags & MPOL_F_MORON))
	542	pol = NULL;
	543	mpol_get(pol);
	544	task_unlock(p);
	545
	546	P(numa_pages_migrated);
	547	P(numa_preferred_nid);
	548	P(total_numa_faults);
	549	SEQ_printf(m, "current_node=%d, numa_group_id=%d\n",
	550	task_node(p), task_numa_group_id(p));
	551	show_numa_stats(p, m);
	552	mpol_put(pol);
	553	#endif
	554	}
	555
	556	void proc_sched_show_task(struct task_struct p, struct seq_file m)
	557	{
	558	unsigned long nr_switches;
	559
	560	SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr(p),
	561	get_nr_threads(p));
	562	SEQ_printf(m,
	563	"---------------------------------------------------------"
	564	"----------\n");
	565	#define __P(F) \
	566	SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)F)
	567	#define P(F) \
	568	SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)p->F)
	569	#define __PN(F) \
	570	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)F))
	571	#define PN(F) \
	572	SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F))
	573
	574	PN(se.exec_start);
	575	PN(se.vruntime);
	576	PN(se.sum_exec_runtime);
	577
	578	nr_switches = p->nvcsw + p->nivcsw;
	579
	580	#ifdef CONFIG_SCHEDSTATS
	581	PN(se.statistics.sum_sleep_runtime);
	582	PN(se.statistics.wait_start);
	583	PN(se.statistics.sleep_start);
	584	PN(se.statistics.block_start);
	585	PN(se.statistics.sleep_max);
	586	PN(se.statistics.block_max);
	587	PN(se.statistics.exec_max);
	588	PN(se.statistics.slice_max);
	589	PN(se.statistics.wait_max);
	590	PN(se.statistics.wait_sum);
	591	P(se.statistics.wait_count);
	592	PN(se.statistics.iowait_sum);
	593	P(se.statistics.iowait_count);
	594	P(se.nr_migrations);
	595	P(se.statistics.nr_migrations_cold);
	596	P(se.statistics.nr_failed_migrations_affine);
	597	P(se.statistics.nr_failed_migrations_running);
	598	P(se.statistics.nr_failed_migrations_hot);
	599	P(se.statistics.nr_forced_migrations);
	600	P(se.statistics.nr_wakeups);
	601	P(se.statistics.nr_wakeups_sync);
	602	P(se.statistics.nr_wakeups_migrate);
	603	P(se.statistics.nr_wakeups_local);
	604	P(se.statistics.nr_wakeups_remote);
	605	P(se.statistics.nr_wakeups_affine);
	606	P(se.statistics.nr_wakeups_affine_attempts);
	607	P(se.statistics.nr_wakeups_passive);
	608	P(se.statistics.nr_wakeups_idle);
	609
	610	{
	611	u64 avg_atom, avg_per_cpu;
	612
	613	avg_atom = p->se.sum_exec_runtime;
	614	if (nr_switches)
	615	avg_atom = div64_ul(avg_atom, nr_switches);
	616	else
	617	avg_atom = -1LL;
	618
	619	avg_per_cpu = p->se.sum_exec_runtime;
	620	if (p->se.nr_migrations) {
	621	avg_per_cpu = div64_u64(avg_per_cpu,
	622	p->se.nr_migrations);
	623	} else {
	624	avg_per_cpu = -1LL;
	625	}
	626
	627	__PN(avg_atom);
	628	__PN(avg_per_cpu);
	629	}
	630	#endif
	631	__P(nr_switches);
	632	SEQ_printf(m, "%-45s:%21Ld\n",
	633	"nr_voluntary_switches", (long long)p->nvcsw);
	634	SEQ_printf(m, "%-45s:%21Ld\n",
	635	"nr_involuntary_switches", (long long)p->nivcsw);
	636
	637	P(se.load.weight);
	638	#ifdef CONFIG_SMP
	639	P(se.avg.runnable_avg_sum);
	640	P(se.avg.running_avg_sum);
	641	P(se.avg.avg_period);
	642	P(se.avg.load_avg_contrib);
	643	P(se.avg.utilization_avg_contrib);
	644	P(se.avg.decay_count);
	645	#endif
	646	P(policy);
	647	P(prio);
	648	#undef PN
	649	#undef __PN
	650	#undef P
	651	#undef __P
	652
	653	{
	654	unsigned int this_cpu = raw_smp_processor_id();
	655	u64 t0, t1;
	656
	657	t0 = cpu_clock(this_cpu);
	658	t1 = cpu_clock(this_cpu);
	659	SEQ_printf(m, "%-45s:%21Ld\n",
	660	"clock-delta", (long long)(t1-t0));
	661	}
	662
	663	sched_show_numa(p, m);
	664	}
	665
	666	void proc_sched_set_task(struct task_struct *p)
	667	{
	668	#ifdef CONFIG_SCHEDSTATS
	669	memset(&p->se.statistics, 0, sizeof(p->se.statistics));
	670	#endif
	671	}