[linux-block.git] / kernel / sched_debug.c

/*
 * kernel/time/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>

/*
 * 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)

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 %15Ld %13Ld %13Ld %9Ld %5d ",
		p->comm, p->pid,
		(long long)p->se.fair_key,
		(long long)(p->se.fair_key - rq->cfs.fair_clock),
		(long long)p->se.wait_runtime,
		(long long)(p->nvcsw + p->nivcsw),
		p->prio);
#ifdef CONFIG_SCHEDSTATS
	SEQ_printf(m, "%15Ld %15Ld %15Ld %15Ld %15Ld %15Ld\n",
		(long long)p->se.vruntime,
		(long long)p->se.sum_exec_runtime,
		(long long)p->se.sum_wait_runtime,
		(long long)p->se.sum_sleep_runtime,
		(long long)p->se.wait_runtime_overruns,
		(long long)p->se.wait_runtime_underruns);
#else
	SEQ_printf(m, "%15Ld %15Ld %15Ld %15Ld %15Ld\n",
		0LL, 0LL, 0LL, 0LL, 0LL);
#endif
}

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         delta       waiting"
	"  switches  prio"
	"    exec-runtime        sum-exec        sum-wait       sum-sleep"
	"    wait-overrun   wait-underrun\n"
	"------------------------------------------------------------------"
	"--------------------------------"
	"------------------------------------------------"
	"--------------------------------\n");

	read_lock_irq(&tasklist_lock);

	do_each_thread(g, p) {
		if (!p->se.on_rq || task_cpu(p) != rq_cpu)
			continue;

		print_task(m, rq, p);
	} while_each_thread(g, p);

	read_unlock_irq(&tasklist_lock);
}

static void
print_cfs_rq_runtime_sum(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
{
	s64 wait_runtime_rq_sum = 0;
	struct task_struct *p;
	struct rb_node *curr;
	unsigned long flags;
	struct rq *rq = &per_cpu(runqueues, cpu);

	spin_lock_irqsave(&rq->lock, flags);
	curr = first_fair(cfs_rq);
	while (curr) {
		p = rb_entry(curr, struct task_struct, se.run_node);
		wait_runtime_rq_sum += p->se.wait_runtime;

		curr = rb_next(curr);
	}
	spin_unlock_irqrestore(&rq->lock, flags);

	SEQ_printf(m, "  .%-30s: %Ld\n", "wait_runtime_rq_sum",
		(long long)wait_runtime_rq_sum);
}

void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
{
	s64 MIN_vruntime = -1, max_vruntime = -1, spread;
	struct rq *rq = &per_cpu(runqueues, cpu);
	struct sched_entity *last;
	unsigned long flags;

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

#define P(x) \
	SEQ_printf(m, "  .%-30s: %Ld\n", #x, (long long)(cfs_rq->x))

	P(fair_clock);
	P(exec_clock);
	P(min_vruntime);

	spin_lock_irqsave(&rq->lock, flags);
	if (cfs_rq->rb_leftmost)
		MIN_vruntime = (__pick_next_entity(cfs_rq))->vruntime;
	last = __pick_last_entity(cfs_rq);
	if (last)
		max_vruntime = last->vruntime;
	spin_unlock_irqrestore(&rq->lock, flags);
	SEQ_printf(m, "  .%-30s: %Ld\n", "MIN_vruntime",
			(long long)MIN_vruntime);
	SEQ_printf(m, "  .%-30s: %Ld\n", "max_vruntime",
			(long long)max_vruntime);
	spread = max_vruntime - MIN_vruntime;
	SEQ_printf(m, "  .%-30s: %Ld\n", "spread",
			(long long)spread);

	P(wait_runtime);
	P(wait_runtime_overruns);
	P(wait_runtime_underruns);
	P(sleeper_bonus);
#undef P

	print_cfs_rq_runtime_sum(m, cpu, cfs_rq);
}

static void print_cpu(struct seq_file *m, int cpu)
{
	struct rq *rq = &per_cpu(runqueues, cpu);

#ifdef CONFIG_X86
	{
		unsigned int freq = cpu_khz ? : 1;

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

#define P(x) \
	SEQ_printf(m, "  .%-30s: %Ld\n", #x, (long long)(rq->x))

	P(nr_running);
	SEQ_printf(m, "  .%-30s: %lu\n", "load",
		   rq->ls.load.weight);
	P(nr_switches);
	P(nr_load_updates);
	P(nr_uninterruptible);
	SEQ_printf(m, "  .%-30s: %lu\n", "jiffies", jiffies);
	P(next_balance);
	P(curr->pid);
	P(clock);
	P(idle_clock);
	P(prev_clock_raw);
	P(clock_warps);
	P(clock_overflows);
	P(clock_deep_idle_events);
	P(clock_max_delta);
	P(cpu_load[0]);
	P(cpu_load[1]);
	P(cpu_load[2]);
	P(cpu_load[3]);
	P(cpu_load[4]);
#undef P

	print_cfs_stats(m, cpu);

	print_rq(m, rq, cpu);
}

static int sched_debug_show(struct seq_file *m, void *v)
{
	u64 now = ktime_to_ns(ktime_get());
	int cpu;

	SEQ_printf(m, "Sched Debug Version: v0.05-v20, %s %.*s\n",
		init_utsname()->release,
		(int)strcspn(init_utsname()->version, " "),
		init_utsname()->version);

	SEQ_printf(m, "now at %Lu nsecs\n", (unsigned long long)now);

	for_each_online_cpu(cpu)
		print_cpu(m, cpu);

	SEQ_printf(m, "\n");

	return 0;
}

static void sysrq_sched_debug_show(void)
{
	sched_debug_show(NULL, NULL);
}

static int sched_debug_open(struct inode *inode, struct file *filp)
{
	return single_open(filp, sched_debug_show, NULL);
}

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

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

	pe = create_proc_entry("sched_debug", 0644, NULL);
	if (!pe)
		return -ENOMEM;

	pe->proc_fops = &sched_debug_fops;

	return 0;
}

__initcall(init_sched_debug_procfs);

void proc_sched_show_task(struct task_struct *p, struct seq_file *m)
{
	unsigned long flags;
	int num_threads = 1;

	rcu_read_lock();
	if (lock_task_sighand(p, &flags)) {
		num_threads = atomic_read(&p->signal->count);
		unlock_task_sighand(p, &flags);
	}
	rcu_read_unlock();

	SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, p->pid, num_threads);
	SEQ_printf(m, "----------------------------------------------\n");
#define P(F) \
	SEQ_printf(m, "%-25s:%20Ld\n", #F, (long long)p->F)

	P(se.wait_runtime);
	P(se.wait_start_fair);
	P(se.exec_start);
	P(se.sleep_start_fair);
	P(se.vruntime);
	P(se.sum_exec_runtime);

#ifdef CONFIG_SCHEDSTATS
	P(se.wait_start);
	P(se.sleep_start);
	P(se.block_start);
	P(se.sleep_max);
	P(se.block_max);
	P(se.exec_max);
	P(se.slice_max);
	P(se.wait_max);
	P(se.wait_runtime_overruns);
	P(se.wait_runtime_underruns);
	P(se.sum_wait_runtime);
#endif
	SEQ_printf(m, "%-25s:%20Ld\n",
		   "nr_switches", (long long)(p->nvcsw + p->nivcsw));
	P(se.load.weight);
	P(policy);
	P(prio);
#undef P

	{
		u64 t0, t1;

		t0 = sched_clock();
		t1 = sched_clock();
		SEQ_printf(m, "%-25s:%20Ld\n",
			   "clock-delta", (long long)(t1-t0));
	}
}

void proc_sched_set_task(struct task_struct *p)
{
#ifdef CONFIG_SCHEDSTATS
	p->se.sleep_max			= 0;
	p->se.block_max			= 0;
	p->se.exec_max			= 0;
	p->se.slice_max			= 0;
	p->se.wait_max			= 0;
	p->se.wait_runtime_overruns	= 0;
	p->se.wait_runtime_underruns	= 0;
#endif
	p->se.sum_exec_runtime		= 0;
	p->se.prev_sum_exec_runtime	= 0;
}
Commit	Line	Data
43ae34cb IM	1	/*
	2	* kernel/time/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
	19	/*
	20	* This allows printing both to /proc/sched_debug and
	21	* to the console
	22	*/
	23	#define SEQ_printf(m, x...) \
	24	do { \
	25	if (m) \
	26	seq_printf(m, x); \
	27	else \
	28	printk(x); \
	29	} while (0)
	30
	31	static void
a48da48b	32	print_task(struct seq_file m, struct rq rq, struct task_struct *p)
43ae34cb IM	33	{
	34	if (rq->curr == p)
	35	SEQ_printf(m, "R");
	36	else
	37	SEQ_printf(m, " ");
	38
6f605d83	39	SEQ_printf(m, "%15s %5d %15Ld %13Ld %13Ld %9Ld %5d ",
43ae34cb IM	40	p->comm, p->pid,
	41	(long long)p->se.fair_key,
	42	(long long)(p->se.fair_key - rq->cfs.fair_clock),
	43	(long long)p->se.wait_runtime,
	44	(long long)(p->nvcsw + p->nivcsw),
6f605d83	45	p->prio);
6cfb0d5d	46	#ifdef CONFIG_SCHEDSTATS
67e12eac IM	47	SEQ_printf(m, "%15Ld %15Ld %15Ld %15Ld %15Ld %15Ld\n",
67e12eac IM	48	(long long)p->se.vruntime,
43ae34cb IM	49	(long long)p->se.sum_exec_runtime,
	50	(long long)p->se.sum_wait_runtime,
	51	(long long)p->se.sum_sleep_runtime,
	52	(long long)p->se.wait_runtime_overruns,
6f605d83	53	(long long)p->se.wait_runtime_underruns);
6cfb0d5d	54	#else
6f605d83 AV	55	SEQ_printf(m, "%15Ld %15Ld %15Ld %15Ld %15Ld\n",
6f605d83 AV	56	0LL, 0LL, 0LL, 0LL, 0LL);
6cfb0d5d	57	#endif
43ae34cb IM	58	}
43ae34cb IM	59
a48da48b	60	static void print_rq(struct seq_file m, struct rq rq, int rq_cpu)
43ae34cb IM	61	{
	62	struct task_struct g, p;
	63
	64	SEQ_printf(m,
	65	"\nrunnable tasks:\n"
	66	" task PID tree-key delta waiting"
	67	" switches prio"
67e12eac	68	" exec-runtime sum-exec sum-wait sum-sleep"
43ae34cb IM	69	" wait-overrun wait-underrun\n"
43ae34cb IM	70	"------------------------------------------------------------------"
67e12eac	71	"--------------------------------"
43ae34cb IM	72	"------------------------------------------------"
	73	"--------------------------------\n");
	74
	75	read_lock_irq(&tasklist_lock);
	76
	77	do_each_thread(g, p) {
	78	if (!p->se.on_rq \|\| task_cpu(p) != rq_cpu)
	79	continue;
	80
a48da48b	81	print_task(m, rq, p);
43ae34cb IM	82	} while_each_thread(g, p);
	83
	84	read_unlock_irq(&tasklist_lock);
	85	}
	86
	87	static void
	88	print_cfs_rq_runtime_sum(struct seq_file m, int cpu, struct cfs_rq cfs_rq)
	89	{
	90	s64 wait_runtime_rq_sum = 0;
	91	struct task_struct *p;
	92	struct rb_node *curr;
	93	unsigned long flags;
	94	struct rq *rq = &per_cpu(runqueues, cpu);
	95
	96	spin_lock_irqsave(&rq->lock, flags);
	97	curr = first_fair(cfs_rq);
	98	while (curr) {
	99	p = rb_entry(curr, struct task_struct, se.run_node);
	100	wait_runtime_rq_sum += p->se.wait_runtime;
	101
	102	curr = rb_next(curr);
	103	}
	104	spin_unlock_irqrestore(&rq->lock, flags);
	105
	106	SEQ_printf(m, " .%-30s: %Ld\n", "wait_runtime_rq_sum",
	107	(long long)wait_runtime_rq_sum);
	108	}
	109
5cef9eca	110	void print_cfs_rq(struct seq_file m, int cpu, struct cfs_rq cfs_rq)
43ae34cb	111	{
67e12eac IM	112	s64 MIN_vruntime = -1, max_vruntime = -1, spread;
	113	struct rq *rq = &per_cpu(runqueues, cpu);
	114	struct sched_entity *last;
	115	unsigned long flags;
	116
5167e75f	117	SEQ_printf(m, "\ncfs_rq\n");
43ae34cb IM	118
	119	#define P(x) \
	120	SEQ_printf(m, " .%-30s: %Ld\n", #x, (long long)(cfs_rq->x))
	121
	122	P(fair_clock);
	123	P(exec_clock);
67e12eac IM	124	P(min_vruntime);
	125
	126	spin_lock_irqsave(&rq->lock, flags);
	127	if (cfs_rq->rb_leftmost)
	128	MIN_vruntime = (__pick_next_entity(cfs_rq))->vruntime;
	129	last = __pick_last_entity(cfs_rq);
	130	if (last)
	131	max_vruntime = last->vruntime;
	132	spin_unlock_irqrestore(&rq->lock, flags);
	133	SEQ_printf(m, " .%-30s: %Ld\n", "MIN_vruntime",
	134	(long long)MIN_vruntime);
	135	SEQ_printf(m, " .%-30s: %Ld\n", "max_vruntime",
	136	(long long)max_vruntime);
	137	spread = max_vruntime - MIN_vruntime;
	138	SEQ_printf(m, " .%-30s: %Ld\n", "spread",
	139	(long long)spread);
	140
43ae34cb IM	141	P(wait_runtime);
	142	P(wait_runtime_overruns);
	143	P(wait_runtime_underruns);
	144	P(sleeper_bonus);
	145	#undef P
	146
	147	print_cfs_rq_runtime_sum(m, cpu, cfs_rq);
	148	}
	149
a48da48b	150	static void print_cpu(struct seq_file *m, int cpu)
43ae34cb IM	151	{
	152	struct rq *rq = &per_cpu(runqueues, cpu);
	153
	154	#ifdef CONFIG_X86
	155	{
	156	unsigned int freq = cpu_khz ? : 1;
	157
	158	SEQ_printf(m, "\ncpu#%d, %u.%03u MHz\n",
	159	cpu, freq / 1000, (freq % 1000));
	160	}
	161	#else
	162	SEQ_printf(m, "\ncpu#%d\n", cpu);
	163	#endif
	164
	165	#define P(x) \
	166	SEQ_printf(m, " .%-30s: %Ld\n", #x, (long long)(rq->x))
	167
	168	P(nr_running);
	169	SEQ_printf(m, " .%-30s: %lu\n", "load",
	170	rq->ls.load.weight);
43ae34cb IM	171	P(nr_switches);
	172	P(nr_load_updates);
	173	P(nr_uninterruptible);
	174	SEQ_printf(m, " .%-30s: %lu\n", "jiffies", jiffies);
	175	P(next_balance);
	176	P(curr->pid);
	177	P(clock);
2aa44d05	178	P(idle_clock);
43ae34cb IM	179	P(prev_clock_raw);
	180	P(clock_warps);
	181	P(clock_overflows);
2aa44d05	182	P(clock_deep_idle_events);
43ae34cb IM	183	P(clock_max_delta);
	184	P(cpu_load[0]);
	185	P(cpu_load[1]);
	186	P(cpu_load[2]);
	187	P(cpu_load[3]);
	188	P(cpu_load[4]);
	189	#undef P
	190
5cef9eca	191	print_cfs_stats(m, cpu);
43ae34cb	192
a48da48b	193	print_rq(m, rq, cpu);
43ae34cb IM	194	}
	195
	196	static int sched_debug_show(struct seq_file m, void v)
	197	{
	198	u64 now = ktime_to_ns(ktime_get());
	199	int cpu;
	200
6cfb0d5d	201	SEQ_printf(m, "Sched Debug Version: v0.05-v20, %s %.*s\n",
43ae34cb IM	202	init_utsname()->release,
	203	(int)strcspn(init_utsname()->version, " "),
	204	init_utsname()->version);
	205
	206	SEQ_printf(m, "now at %Lu nsecs\n", (unsigned long long)now);
	207
	208	for_each_online_cpu(cpu)
a48da48b	209	print_cpu(m, cpu);
43ae34cb IM	210
	211	SEQ_printf(m, "\n");
	212
	213	return 0;
	214	}
	215
f3373461	216	static void sysrq_sched_debug_show(void)
43ae34cb IM	217	{
	218	sched_debug_show(NULL, NULL);
	219	}
	220
	221	static int sched_debug_open(struct inode inode, struct file filp)
	222	{
	223	return single_open(filp, sched_debug_show, NULL);
	224	}
	225
	226	static struct file_operations sched_debug_fops = {
	227	.open = sched_debug_open,
	228	.read = seq_read,
	229	.llseek = seq_lseek,
5ea473a1	230	.release = single_release,
43ae34cb IM	231	};
	232
	233	static int __init init_sched_debug_procfs(void)
	234	{
	235	struct proc_dir_entry *pe;
	236
	237	pe = create_proc_entry("sched_debug", 0644, NULL);
	238	if (!pe)
	239	return -ENOMEM;
	240
	241	pe->proc_fops = &sched_debug_fops;
	242
	243	return 0;
	244	}
	245
	246	__initcall(init_sched_debug_procfs);
	247
	248	void proc_sched_show_task(struct task_struct p, struct seq_file m)
	249	{
	250	unsigned long flags;
	251	int num_threads = 1;
	252
	253	rcu_read_lock();
	254	if (lock_task_sighand(p, &flags)) {
	255	num_threads = atomic_read(&p->signal->count);
	256	unlock_task_sighand(p, &flags);
	257	}
	258	rcu_read_unlock();
	259
	260	SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, p->pid, num_threads);
	261	SEQ_printf(m, "----------------------------------------------\n");
	262	#define P(F) \
	263	SEQ_printf(m, "%-25s:%20Ld\n", #F, (long long)p->F)
	264
6cfb0d5d	265	P(se.wait_runtime);
43ae34cb IM	266	P(se.wait_start_fair);
43ae34cb IM	267	P(se.exec_start);
43ae34cb	268	P(se.sleep_start_fair);
67e12eac	269	P(se.vruntime);
6cfb0d5d IM	270	P(se.sum_exec_runtime);
	271
	272	#ifdef CONFIG_SCHEDSTATS
	273	P(se.wait_start);
	274	P(se.sleep_start);
43ae34cb IM	275	P(se.block_start);
	276	P(se.sleep_max);
	277	P(se.block_max);
	278	P(se.exec_max);
eba1ed4b	279	P(se.slice_max);
43ae34cb	280	P(se.wait_max);
43ae34cb IM	281	P(se.wait_runtime_overruns);
	282	P(se.wait_runtime_underruns);
	283	P(se.sum_wait_runtime);
6cfb0d5d	284	#endif
43ae34cb IM	285	SEQ_printf(m, "%-25s:%20Ld\n",
	286	"nr_switches", (long long)(p->nvcsw + p->nivcsw));
	287	P(se.load.weight);
	288	P(policy);
	289	P(prio);
	290	#undef P
	291
	292	{
	293	u64 t0, t1;
	294
	295	t0 = sched_clock();
	296	t1 = sched_clock();
	297	SEQ_printf(m, "%-25s:%20Ld\n",
	298	"clock-delta", (long long)(t1-t0));
	299	}
	300	}
	301
	302	void proc_sched_set_task(struct task_struct *p)
	303	{
6cfb0d5d	304	#ifdef CONFIG_SCHEDSTATS
a4b29ba2 IM	305	p->se.sleep_max = 0;
	306	p->se.block_max = 0;
	307	p->se.exec_max = 0;
eba1ed4b	308	p->se.slice_max = 0;
a4b29ba2 IM	309	p->se.wait_max = 0;
	310	p->se.wait_runtime_overruns = 0;
	311	p->se.wait_runtime_underruns = 0;
6cfb0d5d	312	#endif
a4b29ba2	313	p->se.sum_exec_runtime = 0;
2491b2b8	314	p->se.prev_sum_exec_runtime = 0;
43ae34cb	315	}