trace.c 70.1 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
/*
 * ring buffer based function tracer
 *
 * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
 * Copyright (C) 2008 Ingo Molnar <mingo@redhat.com>
 *
 * Originally taken from the RT patch by:
 *    Arnaldo Carvalho de Melo <acme@redhat.com>
 *
 * Based on code from the latency_tracer, that is:
 *  Copyright (C) 2004-2006 Ingo Molnar
 *  Copyright (C) 2004 William Lee Irwin III
 */
#include <linux/utsrelease.h>
#include <linux/kallsyms.h>
#include <linux/seq_file.h>
17
#include <linux/notifier.h>
18
#include <linux/debugfs.h>
19
#include <linux/pagemap.h>
20
21
22
23
24
25
#include <linux/hardirq.h>
#include <linux/linkage.h>
#include <linux/uaccess.h>
#include <linux/ftrace.h>
#include <linux/module.h>
#include <linux/percpu.h>
26
#include <linux/kdebug.h>
27
28
#include <linux/ctype.h>
#include <linux/init.h>
29
#include <linux/poll.h>
30
31
#include <linux/gfp.h>
#include <linux/fs.h>
32
#include <linux/kprobes.h>
Steven Rostedt's avatar
Steven Rostedt committed
33
#include <linux/writeback.h>
34

Ingo Molnar's avatar
Ingo Molnar committed
35
#include <linux/stacktrace.h>
36
#include <linux/ring_buffer.h>
37
#include <linux/irqflags.h>
Ingo Molnar's avatar
Ingo Molnar committed
38

39
#include "trace.h"
40
#include "trace_output.h"
41

42
43
#define TRACE_BUFFER_FLAGS	(RB_FL_OVERWRITE)

44
unsigned long __read_mostly	tracing_max_latency;
45
46
unsigned long __read_mostly	tracing_thresh;

47
48
/*
 * We need to change this state when a selftest is running.
49
50
51
52
53
 * A selftest will lurk into the ring-buffer to count the
 * entries inserted during the selftest although some concurrent
 * insertions into the ring-buffer such as ftrace_printk could occurred
 * at the same time, giving false positive or negative results.
 */
54
static bool __read_mostly tracing_selftest_running;
55

56
57
58
59
60
/*
 * If a tracer is running, we do not want to run SELFTEST.
 */
static bool __read_mostly tracing_selftest_disabled;

61
62
63
64
65
66
67
68
69
70
71
72
73
74
/* For tracers that don't implement custom flags */
static struct tracer_opt dummy_tracer_opt[] = {
	{ }
};

static struct tracer_flags dummy_tracer_flags = {
	.val = 0,
	.opts = dummy_tracer_opt
};

static int dummy_set_flag(u32 old_flags, u32 bit, int set)
{
	return 0;
}
75
76
77
78
79
80
81
82
83

/*
 * Kill all tracing for good (never come back).
 * It is initialized to 1 but will turn to zero if the initialization
 * of the tracer is successful. But that is the only place that sets
 * this back to zero.
 */
int tracing_disabled = 1;

84
85
86
87
88
89
90
91
92
93
94
95
96
97
static DEFINE_PER_CPU(local_t, ftrace_cpu_disabled);

static inline void ftrace_disable_cpu(void)
{
	preempt_disable();
	local_inc(&__get_cpu_var(ftrace_cpu_disabled));
}

static inline void ftrace_enable_cpu(void)
{
	local_dec(&__get_cpu_var(ftrace_cpu_disabled));
	preempt_enable();
}

98
static cpumask_var_t __read_mostly	tracing_buffer_mask;
99
100

#define for_each_tracing_cpu(cpu)	\
101
	for_each_cpu(cpu, tracing_buffer_mask)
102

103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
/*
 * ftrace_dump_on_oops - variable to dump ftrace buffer on oops
 *
 * If there is an oops (or kernel panic) and the ftrace_dump_on_oops
 * is set, then ftrace_dump is called. This will output the contents
 * of the ftrace buffers to the console.  This is very useful for
 * capturing traces that lead to crashes and outputing it to a
 * serial console.
 *
 * It is default off, but you can enable it with either specifying
 * "ftrace_dump_on_oops" in the kernel command line, or setting
 * /proc/sys/kernel/ftrace_dump_on_oops to true.
 */
int ftrace_dump_on_oops;

118
119
120
121
122
static int tracing_set_tracer(const char *buf);

#define BOOTUP_TRACER_SIZE		100
static char bootup_tracer_buf[BOOTUP_TRACER_SIZE] __initdata;
static char *default_bootup_tracer;
123
124
125

static int __init set_ftrace(char *str)
{
126
127
	strncpy(bootup_tracer_buf, str, BOOTUP_TRACER_SIZE);
	default_bootup_tracer = bootup_tracer_buf;
128
129
	return 1;
}
130
__setup("ftrace=", set_ftrace);
131

132
133
134
135
136
137
static int __init set_ftrace_dump_on_oops(char *str)
{
	ftrace_dump_on_oops = 1;
	return 1;
}
__setup("ftrace_dump_on_oops", set_ftrace_dump_on_oops);
Steven Rostedt's avatar
Steven Rostedt committed
138

139
long
140
141
142
143
144
145
146
ns2usecs(cycle_t nsec)
{
	nsec += 500;
	do_div(nsec, 1000);
	return nsec;
}

Ingo Molnar's avatar
Ingo Molnar committed
147
cycle_t ftrace_now(int cpu)
148
{
149
150
151
	u64 ts = ring_buffer_time_stamp(cpu);
	ring_buffer_normalize_time_stamp(cpu, &ts);
	return ts;
152
153
}

Steven Rostedt's avatar
Steven Rostedt committed
154
155
156
157
158
159
160
161
162
163
164
165
/*
 * The global_trace is the descriptor that holds the tracing
 * buffers for the live tracing. For each CPU, it contains
 * a link list of pages that will store trace entries. The
 * page descriptor of the pages in the memory is used to hold
 * the link list by linking the lru item in the page descriptor
 * to each of the pages in the buffer per CPU.
 *
 * For each active CPU there is a data field that holds the
 * pages for the buffer for that CPU. Each CPU has the same number
 * of pages allocated for its buffer.
 */
166
167
168
169
static struct trace_array	global_trace;

static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu);

Steven Rostedt's avatar
Steven Rostedt committed
170
171
172
173
174
175
176
177
178
179
/*
 * The max_tr is used to snapshot the global_trace when a maximum
 * latency is reached. Some tracers will use this to store a maximum
 * trace while it continues examining live traces.
 *
 * The buffers for the max_tr are set up the same as the global_trace.
 * When a snapshot is taken, the link list of the max_tr is swapped
 * with the link list of the global_trace and the buffers are reset for
 * the global_trace so the tracing can continue.
 */
180
181
182
183
static struct trace_array	max_tr;

static DEFINE_PER_CPU(struct trace_array_cpu, max_data);

Steven Rostedt's avatar
Steven Rostedt committed
184
/* tracer_enabled is used to toggle activation of a tracer */
185
static int			tracer_enabled = 1;
Steven Rostedt's avatar
Steven Rostedt committed
186

187
188
189
190
191
192
193
194
195
196
197
198
199
/**
 * tracing_is_enabled - return tracer_enabled status
 *
 * This function is used by other tracers to know the status
 * of the tracer_enabled flag.  Tracers may use this function
 * to know if it should enable their features when starting
 * up. See irqsoff tracer for an example (start_irqsoff_tracer).
 */
int tracing_is_enabled(void)
{
	return tracer_enabled;
}

Steven Rostedt's avatar
Steven Rostedt committed
200
/*
201
202
203
 * trace_buf_size is the size in bytes that is allocated
 * for a buffer. Note, the number of bytes is always rounded
 * to page size.
204
205
206
207
208
 *
 * This number is purposely set to a low number of 16384.
 * If the dump on oops happens, it will be much appreciated
 * to not have to wait for all that output. Anyway this can be
 * boot time and run time configurable.
Steven Rostedt's avatar
Steven Rostedt committed
209
 */
210
#define TRACE_BUF_SIZE_DEFAULT	1441792UL /* 16384 * 88 (sizeof(entry)) */
211

212
static unsigned long		trace_buf_size = TRACE_BUF_SIZE_DEFAULT;
213

Steven Rostedt's avatar
Steven Rostedt committed
214
/* trace_types holds a link list of available tracers. */
215
static struct tracer		*trace_types __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
216
217

/* current_trace points to the tracer that is currently active */
218
static struct tracer		*current_trace __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
219
220
221
222
223
224

/*
 * max_tracer_type_len is used to simplify the allocating of
 * buffers to read userspace tracer names. We keep track of
 * the longest tracer name registered.
 */
225
226
static int			max_tracer_type_len;

Steven Rostedt's avatar
Steven Rostedt committed
227
228
229
230
231
232
/*
 * trace_types_lock is used to protect the trace_types list.
 * This lock is also used to keep user access serialized.
 * Accesses from userspace will grab this lock while userspace
 * activities happen inside the kernel.
 */
233
static DEFINE_MUTEX(trace_types_lock);
Steven Rostedt's avatar
Steven Rostedt committed
234
235

/* trace_wait is a waitqueue for tasks blocked on trace_poll */
236
237
static DECLARE_WAIT_QUEUE_HEAD(trace_wait);

238
/* trace_flags holds trace_options default values */
239
unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK |
240
	TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO;
241

Steven Rostedt's avatar
Steven Rostedt committed
242
243
244
245
246
247
/**
 * trace_wake_up - wake up tasks waiting for trace input
 *
 * Simply wakes up any task that is blocked on the trace_wait
 * queue. These is used with trace_poll for tasks polling the trace.
 */
248
249
void trace_wake_up(void)
{
Ingo Molnar's avatar
Ingo Molnar committed
250
251
252
253
254
	/*
	 * The runqueue_is_locked() can fail, but this is the best we
	 * have for now:
	 */
	if (!(trace_flags & TRACE_ITER_BLOCK) && !runqueue_is_locked())
255
256
		wake_up(&trace_wait);
}
257

258
static int __init set_buf_size(char *str)
259
{
260
	unsigned long buf_size;
261
262
	int ret;

263
264
	if (!str)
		return 0;
265
	ret = strict_strtoul(str, 0, &buf_size);
266
	/* nr_entries can not be zero */
267
	if (ret < 0 || buf_size == 0)
268
		return 0;
269
	trace_buf_size = buf_size;
270
271
	return 1;
}
272
__setup("trace_buf_size=", set_buf_size);
273

Steven Rostedt's avatar
Steven Rostedt committed
274
275
276
277
278
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

Steven Rostedt's avatar
Steven Rostedt committed
279
/* These must match the bit postions in trace_iterator_flags */
280
281
282
283
284
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
Ingo Molnar's avatar
Ingo Molnar committed
285
	"raw",
286
	"hex",
Ingo Molnar's avatar
Ingo Molnar committed
287
	"bin",
288
	"block",
Ingo Molnar's avatar
Ingo Molnar committed
289
	"stacktrace",
Ingo Molnar's avatar
Ingo Molnar committed
290
	"sched-tree",
291
	"ftrace_printk",
292
	"ftrace_preempt",
293
	"branch",
294
	"annotate",
295
	"userstacktrace",
296
	"sym-userobj",
297
	"printk-msg-only",
298
	"context-info",
299
300
301
	NULL
};

Steven Rostedt's avatar
Steven Rostedt committed
302
303
304
305
306
307
308
309
310
/*
 * ftrace_max_lock is used to protect the swapping of buffers
 * when taking a max snapshot. The buffers themselves are
 * protected by per_cpu spinlocks. But the action of the swap
 * needs its own lock.
 *
 * This is defined as a raw_spinlock_t in order to help
 * with performance when lockdep debugging is enabled.
 */
311
312
static raw_spinlock_t ftrace_max_lock =
	(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
313
314
315
316
317
318

/*
 * Copy the new maximum trace into the separate maximum-trace
 * structure. (this way the maximum trace is permanently saved,
 * for later retrieval via /debugfs/tracing/latency_trace)
 */
Ingo Molnar's avatar
Ingo Molnar committed
319
static void
320
321
322
323
324
325
326
327
328
329
330
331
__update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
	struct trace_array_cpu *data = tr->data[cpu];

	max_tr.cpu = cpu;
	max_tr.time_start = data->preempt_timestamp;

	data = max_tr.data[cpu];
	data->saved_latency = tracing_max_latency;

	memcpy(data->comm, tsk->comm, TASK_COMM_LEN);
	data->pid = tsk->pid;
332
	data->uid = task_uid(tsk);
333
334
335
336
337
338
339
340
	data->nice = tsk->static_prio - 20 - MAX_RT_PRIO;
	data->policy = tsk->policy;
	data->rt_priority = tsk->rt_priority;

	/* record this tasks comm */
	tracing_record_cmdline(current);
}

Ingo Molnar's avatar
Ingo Molnar committed
341
static void
342
343
344
trace_seq_reset(struct trace_seq *s)
{
	s->len = 0;
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
	s->readpos = 0;
}

ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt)
{
	int len;
	int ret;

	if (s->len <= s->readpos)
		return -EBUSY;

	len = s->len - s->readpos;
	if (cnt > len)
		cnt = len;
	ret = copy_to_user(ubuf, s->buffer + s->readpos, cnt);
	if (ret)
		return -EFAULT;

	s->readpos += len;
	return cnt;
365
366
}

Ingo Molnar's avatar
Ingo Molnar committed
367
static void
368
369
370
371
372
373
374
375
376
377
trace_print_seq(struct seq_file *m, struct trace_seq *s)
{
	int len = s->len >= PAGE_SIZE ? PAGE_SIZE - 1 : s->len;

	s->buffer[len] = 0;
	seq_puts(m, s->buffer);

	trace_seq_reset(s);
}

Steven Rostedt's avatar
Steven Rostedt committed
378
379
380
381
382
383
384
385
386
/**
 * update_max_tr - snapshot all trace buffers from global_trace to max_tr
 * @tr: tracer
 * @tsk: the task with the latency
 * @cpu: The cpu that initiated the trace.
 *
 * Flip the buffers between the @tr and the max_tr and record information
 * about which task was the cause of this latency.
 */
Ingo Molnar's avatar
Ingo Molnar committed
387
void
388
389
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
390
	struct ring_buffer *buf = tr->buffer;
391

392
	WARN_ON_ONCE(!irqs_disabled());
393
	__raw_spin_lock(&ftrace_max_lock);
394
395
396
397

	tr->buffer = max_tr.buffer;
	max_tr.buffer = buf;

398
	ftrace_disable_cpu();
399
	ring_buffer_reset(tr->buffer);
400
	ftrace_enable_cpu();
401
402

	__update_max_tr(tr, tsk, cpu);
403
	__raw_spin_unlock(&ftrace_max_lock);
404
405
406
407
408
409
410
}

/**
 * update_max_tr_single - only copy one trace over, and reset the rest
 * @tr - tracer
 * @tsk - task with the latency
 * @cpu - the cpu of the buffer to copy.
Steven Rostedt's avatar
Steven Rostedt committed
411
412
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
413
 */
Ingo Molnar's avatar
Ingo Molnar committed
414
void
415
416
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
417
	int ret;
418

419
	WARN_ON_ONCE(!irqs_disabled());
420
	__raw_spin_lock(&ftrace_max_lock);
421

422
423
	ftrace_disable_cpu();

424
425
426
	ring_buffer_reset(max_tr.buffer);
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

427
428
	ftrace_enable_cpu();

429
	WARN_ON_ONCE(ret && ret != -EAGAIN);
430
431

	__update_max_tr(tr, tsk, cpu);
432
	__raw_spin_unlock(&ftrace_max_lock);
433
434
}

Steven Rostedt's avatar
Steven Rostedt committed
435
436
437
438
439
440
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
441
442
443
444
445
446
447
448
449
450
451
int register_tracer(struct tracer *type)
{
	struct tracer *t;
	int len;
	int ret = 0;

	if (!type->name) {
		pr_info("Tracer must have a name\n");
		return -1;
	}

Ingo Molnar's avatar
Ingo Molnar committed
452
453
454
455
456
457
458
	/*
	 * When this gets called we hold the BKL which means that
	 * preemption is disabled. Various trace selftests however
	 * need to disable and enable preemption for successful tests.
	 * So we drop the BKL here and grab it after the tests again.
	 */
	unlock_kernel();
459
	mutex_lock(&trace_types_lock);
Ingo Molnar's avatar
Ingo Molnar committed
460

461
462
	tracing_selftest_running = true;

463
464
465
466
467
468
469
470
471
472
	for (t = trace_types; t; t = t->next) {
		if (strcmp(type->name, t->name) == 0) {
			/* already found */
			pr_info("Trace %s already registered\n",
				type->name);
			ret = -1;
			goto out;
		}
	}

473
474
475
476
477
478
479
480
	if (!type->set_flag)
		type->set_flag = &dummy_set_flag;
	if (!type->flags)
		type->flags = &dummy_tracer_flags;
	else
		if (!type->flags->opts)
			type->flags->opts = dummy_tracer_opt;

Steven Rostedt's avatar
Steven Rostedt committed
481
#ifdef CONFIG_FTRACE_STARTUP_TEST
482
	if (type->selftest && !tracing_selftest_disabled) {
Steven Rostedt's avatar
Steven Rostedt committed
483
484
485
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
		int i;
486

Steven Rostedt's avatar
Steven Rostedt committed
487
488
489
490
491
492
493
		/*
		 * Run a selftest on this tracer.
		 * Here we reset the trace buffer, and set the current
		 * tracer to be this tracer. The tracer can then run some
		 * internal tracing to verify that everything is in order.
		 * If we fail, we do not register this tracer.
		 */
Ingo Molnar's avatar
Ingo Molnar committed
494
		for_each_tracing_cpu(i)
495
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
496

Steven Rostedt's avatar
Steven Rostedt committed
497
498
499
500
501
502
503
504
505
506
		current_trace = type;
		/* the test is responsible for initializing and enabling */
		pr_info("Testing tracer %s: ", type->name);
		ret = type->selftest(type, tr);
		/* the test is responsible for resetting too */
		current_trace = saved_tracer;
		if (ret) {
			printk(KERN_CONT "FAILED!\n");
			goto out;
		}
Steven Rostedt's avatar
Steven Rostedt committed
507
		/* Only reset on passing, to avoid touching corrupted buffers */
Ingo Molnar's avatar
Ingo Molnar committed
508
		for_each_tracing_cpu(i)
509
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
510

Steven Rostedt's avatar
Steven Rostedt committed
511
512
513
514
		printk(KERN_CONT "PASSED\n");
	}
#endif

515
516
517
518
519
	type->next = trace_types;
	trace_types = type;
	len = strlen(type->name);
	if (len > max_tracer_type_len)
		max_tracer_type_len = len;
Steven Rostedt's avatar
Steven Rostedt committed
520

521
 out:
522
	tracing_selftest_running = false;
523
524
	mutex_unlock(&trace_types_lock);

Steven Rostedt's avatar
Steven Rostedt committed
525
526
527
528
529
530
531
532
533
534
535
536
	if (ret || !default_bootup_tracer)
		goto out_unlock;

	if (strncmp(default_bootup_tracer, type->name, BOOTUP_TRACER_SIZE))
		goto out_unlock;

	printk(KERN_INFO "Starting tracer '%s'\n", type->name);
	/* Do we want this tracer to start on bootup? */
	tracing_set_tracer(type->name);
	default_bootup_tracer = NULL;
	/* disable other selftests, since this will break it. */
	tracing_selftest_disabled = 1;
537
#ifdef CONFIG_FTRACE_STARTUP_TEST
Steven Rostedt's avatar
Steven Rostedt committed
538
539
	printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n",
	       type->name);
540
541
#endif

Steven Rostedt's avatar
Steven Rostedt committed
542
 out_unlock:
543
	lock_kernel();
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
	return ret;
}

void unregister_tracer(struct tracer *type)
{
	struct tracer **t;
	int len;

	mutex_lock(&trace_types_lock);
	for (t = &trace_types; *t; t = &(*t)->next) {
		if (*t == type)
			goto found;
	}
	pr_info("Trace %s not registered\n", type->name);
	goto out;

 found:
	*t = (*t)->next;
562
563
564
565
566
567
568
569
570

	if (type == current_trace && tracer_enabled) {
		tracer_enabled = 0;
		tracing_stop();
		if (current_trace->stop)
			current_trace->stop(&global_trace);
		current_trace = &nop_trace;
	}

571
572
573
574
575
576
577
578
579
580
581
582
583
	if (strlen(type->name) != max_tracer_type_len)
		goto out;

	max_tracer_type_len = 0;
	for (t = &trace_types; *t; t = &(*t)->next) {
		len = strlen((*t)->name);
		if (len > max_tracer_type_len)
			max_tracer_type_len = len;
	}
 out:
	mutex_unlock(&trace_types_lock);
}

584
void tracing_reset(struct trace_array *tr, int cpu)
585
{
586
	ftrace_disable_cpu();
587
	ring_buffer_reset_cpu(tr->buffer, cpu);
588
	ftrace_enable_cpu();
589
590
}

591
592
593
594
595
596
597
598
599
600
void tracing_reset_online_cpus(struct trace_array *tr)
{
	int cpu;

	tr->time_start = ftrace_now(tr->cpu);

	for_each_online_cpu(cpu)
		tracing_reset(tr, cpu);
}

601
602
603
604
605
606
#define SAVED_CMDLINES 128
static unsigned map_pid_to_cmdline[PID_MAX_DEFAULT+1];
static unsigned map_cmdline_to_pid[SAVED_CMDLINES];
static char saved_cmdlines[SAVED_CMDLINES][TASK_COMM_LEN];
static int cmdline_idx;
static DEFINE_SPINLOCK(trace_cmdline_lock);
607
608
609

/* temporary disable recording */
atomic_t trace_record_cmdline_disabled __read_mostly;
610
611
612
613
614
615
616
617

static void trace_init_cmdlines(void)
{
	memset(&map_pid_to_cmdline, -1, sizeof(map_pid_to_cmdline));
	memset(&map_cmdline_to_pid, -1, sizeof(map_cmdline_to_pid));
	cmdline_idx = 0;
}

618
619
620
static int trace_stop_count;
static DEFINE_SPINLOCK(tracing_start_lock);

621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
/**
 * ftrace_off_permanent - disable all ftrace code permanently
 *
 * This should only be called when a serious anomally has
 * been detected.  This will turn off the function tracing,
 * ring buffers, and other tracing utilites. It takes no
 * locks and can be called from any context.
 */
void ftrace_off_permanent(void)
{
	tracing_disabled = 1;
	ftrace_stop();
	tracing_off_permanent();
}

636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
/**
 * tracing_start - quick start of the tracer
 *
 * If tracing is enabled but was stopped by tracing_stop,
 * this will start the tracer back up.
 */
void tracing_start(void)
{
	struct ring_buffer *buffer;
	unsigned long flags;

	if (tracing_disabled)
		return;

	spin_lock_irqsave(&tracing_start_lock, flags);
651
652
653
654
655
656
	if (--trace_stop_count) {
		if (trace_stop_count < 0) {
			/* Someone screwed up their debugging */
			WARN_ON_ONCE(1);
			trace_stop_count = 0;
		}
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
		goto out;
	}


	buffer = global_trace.buffer;
	if (buffer)
		ring_buffer_record_enable(buffer);

	buffer = max_tr.buffer;
	if (buffer)
		ring_buffer_record_enable(buffer);

	ftrace_start();
 out:
	spin_unlock_irqrestore(&tracing_start_lock, flags);
}

/**
 * tracing_stop - quick stop of the tracer
 *
 * Light weight way to stop tracing. Use in conjunction with
 * tracing_start.
 */
void tracing_stop(void)
{
	struct ring_buffer *buffer;
	unsigned long flags;

	ftrace_stop();
	spin_lock_irqsave(&tracing_start_lock, flags);
	if (trace_stop_count++)
		goto out;

	buffer = global_trace.buffer;
	if (buffer)
		ring_buffer_record_disable(buffer);

	buffer = max_tr.buffer;
	if (buffer)
		ring_buffer_record_disable(buffer);

 out:
	spin_unlock_irqrestore(&tracing_start_lock, flags);
}

Ingo Molnar's avatar
Ingo Molnar committed
702
void trace_stop_cmdline_recording(void);
703

Ingo Molnar's avatar
Ingo Molnar committed
704
static void trace_save_cmdline(struct task_struct *tsk)
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
{
	unsigned map;
	unsigned idx;

	if (!tsk->pid || unlikely(tsk->pid > PID_MAX_DEFAULT))
		return;

	/*
	 * It's not the end of the world if we don't get
	 * the lock, but we also don't want to spin
	 * nor do we want to disable interrupts,
	 * so if we miss here, then better luck next time.
	 */
	if (!spin_trylock(&trace_cmdline_lock))
		return;

	idx = map_pid_to_cmdline[tsk->pid];
	if (idx >= SAVED_CMDLINES) {
		idx = (cmdline_idx + 1) % SAVED_CMDLINES;

		map = map_cmdline_to_pid[idx];
		if (map <= PID_MAX_DEFAULT)
			map_pid_to_cmdline[map] = (unsigned)-1;

		map_pid_to_cmdline[tsk->pid] = idx;

		cmdline_idx = idx;
	}

	memcpy(&saved_cmdlines[idx], tsk->comm, TASK_COMM_LEN);

	spin_unlock(&trace_cmdline_lock);
}

739
char *trace_find_cmdline(int pid)
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
{
	char *cmdline = "<...>";
	unsigned map;

	if (!pid)
		return "<idle>";

	if (pid > PID_MAX_DEFAULT)
		goto out;

	map = map_pid_to_cmdline[pid];
	if (map >= SAVED_CMDLINES)
		goto out;

	cmdline = saved_cmdlines[map];

 out:
	return cmdline;
}

Ingo Molnar's avatar
Ingo Molnar committed
760
void tracing_record_cmdline(struct task_struct *tsk)
761
762
763
764
765
766
767
{
	if (atomic_read(&trace_record_cmdline_disabled))
		return;

	trace_save_cmdline(tsk);
}

768
void
769
770
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
771
772
773
{
	struct task_struct *tsk = current;

774
775
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
776
	entry->tgid               	= (tsk) ? tsk->tgid : 0;
777
	entry->flags =
778
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
779
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
780
781
782
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
783
784
785
786
787
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}

788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
struct ring_buffer_event *trace_buffer_lock_reserve(struct trace_array *tr,
						    unsigned char type,
						    unsigned long len,
						    unsigned long flags, int pc)
{
	struct ring_buffer_event *event;

	event = ring_buffer_lock_reserve(tr->buffer, len);
	if (event != NULL) {
		struct trace_entry *ent = ring_buffer_event_data(event);

		tracing_generic_entry_update(ent, flags, pc);
		ent->type = type;
	}

	return event;
}
static void ftrace_trace_stack(struct trace_array *tr,
			       unsigned long flags, int skip, int pc);
static void ftrace_trace_userstack(struct trace_array *tr,
				   unsigned long flags, int pc);

void trace_buffer_unlock_commit(struct trace_array *tr,
				struct ring_buffer_event *event,
				unsigned long flags, int pc)
{
	ring_buffer_unlock_commit(tr->buffer, event);

	ftrace_trace_stack(tr, flags, 6, pc);
	ftrace_trace_userstack(tr, flags, pc);
	trace_wake_up();
}

Ingo Molnar's avatar
Ingo Molnar committed
821
void
822
trace_function(struct trace_array *tr,
823
824
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
825
{
826
	struct ring_buffer_event *event;
827
	struct ftrace_entry *entry;
828

829
830
831
832
	/* If we are reading the ring buffer, don't trace */
	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

833
834
	event = trace_buffer_lock_reserve(tr, TRACE_FN, sizeof(*entry),
					  flags, pc);
835
836
837
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
838
839
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
840
	ring_buffer_unlock_commit(tr->buffer, event);
841
842
}

843
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
844
845
846
847
848
849
850
851
852
853
854
static void __trace_graph_entry(struct trace_array *tr,
				struct ftrace_graph_ent *trace,
				unsigned long flags,
				int pc)
{
	struct ring_buffer_event *event;
	struct ftrace_graph_ent_entry *entry;

	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

855
856
	event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_ENT,
					  sizeof(*entry), flags, pc);
857
858
859
860
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
	entry->graph_ent			= *trace;
861
	ring_buffer_unlock_commit(global_trace.buffer, event);
862
863
864
}

static void __trace_graph_return(struct trace_array *tr,
865
				struct ftrace_graph_ret *trace,
866
867
868
869
				unsigned long flags,
				int pc)
{
	struct ring_buffer_event *event;
870
	struct ftrace_graph_ret_entry *entry;
871
872
873
874

	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

875
876
	event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_RET,
					  sizeof(*entry), flags, pc);
877
878
879
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
880
	entry->ret				= *trace;
881
	ring_buffer_unlock_commit(global_trace.buffer, event);
882
883
884
}
#endif

Ingo Molnar's avatar
Ingo Molnar committed
885
void
Ingo Molnar's avatar
Ingo Molnar committed
886
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
887
888
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
Ingo Molnar's avatar
Ingo Molnar committed
889
890
{
	if (likely(!atomic_read(&data->disabled)))
891
		trace_function(tr, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
892
893
}

894
895
896
static void __ftrace_trace_stack(struct trace_array *tr,
				 unsigned long flags,
				 int skip, int pc)
Ingo Molnar's avatar
Ingo Molnar committed
897
{
898
#ifdef CONFIG_STACKTRACE
899
	struct ring_buffer_event *event;
900
	struct stack_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
901
902
	struct stack_trace trace;

903
904
	event = trace_buffer_lock_reserve(tr, TRACE_STACK,
					  sizeof(*entry), flags, pc);
905
906
907
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
908
	memset(&entry->caller, 0, sizeof(entry->caller));
Ingo Molnar's avatar
Ingo Molnar committed
909
910
911
912

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
913
	trace.entries		= entry->caller;
Ingo Molnar's avatar
Ingo Molnar committed
914
915

	save_stack_trace(&trace);
916
	ring_buffer_unlock_commit(tr->buffer, event);
917
#endif
Ingo Molnar's avatar
Ingo Molnar committed
918
919
}

920
921
922
923
924
925
926
static void ftrace_trace_stack(struct trace_array *tr,
			       unsigned long flags,
			       int skip, int pc)
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

927
	__ftrace_trace_stack(tr, flags, skip, pc);
928
929
}

930
931
void __trace_stack(struct trace_array *tr,
		   unsigned long flags,
932
		   int skip, int pc)
933
{
934
	__ftrace_trace_stack(tr, flags, skip, pc);
935
936
}

937
static void ftrace_trace_userstack(struct trace_array *tr,
938
				   unsigned long flags, int pc)
939
{
940
#ifdef CONFIG_STACKTRACE
941
	struct ring_buffer_event *event;
942
943
944
945
946
947
	struct userstack_entry *entry;
	struct stack_trace trace;

	if (!(trace_flags & TRACE_ITER_USERSTACKTRACE))
		return;

948
949
	event = trace_buffer_lock_reserve(tr, TRACE_USER_STACK,
					  sizeof(*entry), flags, pc);
950
951
952
953
954
955
956
957
958
959
960
961
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);

	memset(&entry->caller, 0, sizeof(entry->caller));

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= 0;
	trace.entries		= entry->caller;

	save_stack_trace_user(&trace);
962
	ring_buffer_unlock_commit(tr->buffer, event);
963
#endif
964
965
}

966
void __trace_userstack(struct trace_array *tr, unsigned long flags)
967
{
968
	ftrace_trace_userstack(tr, flags, preempt_count());
969
970
}

971
static void
972
ftrace_trace_special(void *__tr,
973
974
		     unsigned long arg1, unsigned long arg2, unsigned long arg3,
		     int pc)
Ingo Molnar's avatar
Ingo Molnar committed
975
{
976
	struct ring_buffer_event *event;
Ingo Molnar's avatar
Ingo Molnar committed
977
	struct trace_array *tr = __tr;
978
	struct special_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
979

980
981
	event = trace_buffer_lock_reserve(tr, TRACE_SPECIAL,
					  sizeof(*entry), 0, pc);
982
983
984
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
985
986
987
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
988
	trace_buffer_unlock_commit(tr, event, 0, pc);
Ingo Molnar's avatar
Ingo Molnar committed
989
990
}

991
992
993
994
void
__trace_special(void *__tr, void *__data,
		unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
995
	ftrace_trace_special(__tr, arg1, arg2, arg3, preempt_count());
996
997
}

Ingo Molnar's avatar
Ingo Molnar committed
998
void
999
tracing_sched_switch_trace(struct trace_array *tr,
Ingo Molnar's avatar
Ingo Molnar committed
1000
1001
			   struct task_struct *prev,
			   struct task_struct *next,
1002
			   unsigned long flags, int pc)
1003
{
1004
	struct ring_buffer_event *event;
1005
	struct ctx_switch_entry *entry;
1006

1007
1008
	event = trace_buffer_lock_reserve(tr, TRACE_CTX,
					  sizeof(*entry), flags, pc);
1009
1010
1011
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1012
1013
1014
1015
1016
1017
1018
	entry->prev_pid			= prev->pid;
	entry->prev_prio		= prev->prio;
	entry->prev_state		= prev->state;
	entry->next_pid			= next->pid;
	entry->next_prio		= next->prio;
	entry->next_state		= next->state;
	entry->next_cpu	= task_cpu(next);
1019
	trace_buffer_unlock_commit(tr, event, flags, pc);
1020
1021
}

1022
1023
void
tracing_sched_wakeup_trace(struct trace_array *tr,
Ingo Molnar's avatar
Ingo Molnar committed
1024
1025
			   struct task_struct *wakee,
			   struct task_struct *curr,
1026
			   unsigned long flags, int pc)
1027
{
1028
	struct ring_buffer_event *event;
1029
	struct ctx_switch_entry *entry;
1030

1031
1032
	event = trace_buffer_lock_reserve(tr, TRACE_WAKE,
					  sizeof(*entry), flags, pc);
1033
1034
1035
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1036
1037
1038
1039
1040
1041
1042
	entry->prev_pid			= curr->pid;
	entry->prev_prio		= curr->prio;
	entry->prev_state		= curr->state;
	entry->next_pid			= wakee->pid;
	entry->next_prio		= wakee->prio;
	entry->next_state		= wakee->state;
	entry->next_cpu			= task_cpu(wakee);
1043
	trace_buffer_unlock_commit(tr, event, flags, pc);
1044
1045
}

1046
1047
1048
1049
1050
void
ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1051
	unsigned long flags;
1052
	int cpu;
1053
	int pc;
1054

1055
	if (tracing_disabled)
1056
1057
		return;

1058
	pc = preempt_count();
1059
	local_irq_save(flags);
1060
1061
1062
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

1063
	if (likely(atomic_inc_return(&data->disabled) == 1))
1064
		ftrace_trace_special(tr, arg1, arg2, arg3, pc);
1065

1066
1067
	atomic_dec(&data->disabled);
	local_irq_restore(flags);
1068
1069
}

1070
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
1071
int trace_graph_entry(struct ftrace_graph_ent *trace)
1072
1073
1074
1075
1076
1077
1078
1079
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	unsigned long flags;
	long disabled;
	int cpu;
	int pc;

1080
1081
1082
	if (!ftrace_trace_task(current))
		return 0;

1083
1084
1085
	if (!ftrace_graph_addr(trace->func))
		return 0;

1086
	local_irq_save(flags);
1087
1088
1089
1090
1091
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);
	if (likely(disabled == 1)) {
		pc = preempt_count();
1092
		__trace_graph_entry(tr, trace, flags, pc);
1093
	}
1094
1095
1096
	/* Only do the atomic if it is not already set */
	if (!test_tsk_trace_graph(current))
		set_tsk_trace_graph(current);
1097
	atomic_dec(&data->disabled);
1098
	local_irq_restore(flags);
1099
1100

	return 1;
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
}

void trace_graph_return(struct ftrace_graph_ret *trace)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	unsigned long flags;
	long disabled;
	int cpu;
	int pc;

1112
	local_irq_save(flags);
1113
1114
1115
1116
1117
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);
	if (likely(disabled == 1)) {
		pc = preempt_count();
1118
		__trace_graph_return(tr, trace, flags, pc);
1119
	}
1120
1121
	if (!trace->depth)
		clear_tsk_trace_graph(current);
1122
	atomic_dec(&data->disabled);
1123
	local_irq_restore(flags);
1124
}
1125
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
1126

1127
1128
enum trace_file_type {
	TRACE_FILE_LAT_FMT	= 1,
1129
	TRACE_FILE_ANNOTATE	= 2,
1130
1131
};

1132
static void trace_iterator_increment(struct trace_iterator *iter)
1133
{
1134
1135
1136
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();

1137
	iter->idx++;
1138
1139
1140
1141
	if (iter->buffer_iter[iter->cpu])
		ring_buffer_read(iter->buffer_iter[iter->cpu], NULL);

	ftrace_enable_cpu();
1142
1143
}

Ingo Molnar's avatar
Ingo Molnar committed
1144
static struct trace_entry *
1145
peek_next_entry(struct trace_iterator *iter, int cpu, u64 *ts)
1146
{
1147
1148
	struct ring_buffer_event *event;
	struct ring_buffer_iter *buf_iter = iter->buffer_iter[cpu];
1149

1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();

	if (buf_iter)
		event = ring_buffer_iter_peek(buf_iter, ts);
	else
		event = ring_buffer_peek(iter->tr->buffer, cpu, ts);

	ftrace_enable_cpu();

1160
	return event ? ring_buffer_event_data(event) : NULL;
1161
}
1162

1163
static struct trace_entry *
1164
__find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts)
1165
{
1166
	struct ring_buffer *buffer = iter->tr->buffer;
1167
	struct trace_entry *ent, *next = NULL;
1168
	u64 next_ts = 0, ts;
1169
1170
1171
	int next_cpu = -1;
	int cpu;

1172
	for_each_tracing_cpu(cpu) {
1173

1174
1175
		if (ring_buffer_empty_cpu(buffer, cpu))
			continue;
1176

1177
		ent = peek_next_entry(iter, cpu, &ts);
1178

Ingo Molnar's avatar
Ingo Molnar committed
1179
1180
1181
		/*
		 * Pick the entry with the smallest timestamp:
		 */
1182
		if (ent && (!next || ts < next_ts)) {
1183
1184
			next = ent;
			next_cpu = cpu;
1185
			next_ts = ts;
1186
1187
1188
1189
1190
1191
		}
	}

	if (ent_cpu)
		*ent_cpu = next_cpu;

1192
1193
1194
	if (ent_ts)
		*ent_ts = next_ts;

1195
1196
1197
	return next;
}

1198
/* Find the next real entry, without updating the iterator itself */
1199
1200
struct trace_entry *trace_find_next_entry(struct trace_iterator *iter,
					  int *ent_cpu, u64 *ent_ts)
1201
{
1202
	return __find_next_entry(iter, ent_cpu, ent_ts);
1203
1204
1205
1206
1207
}

/* Find the next real entry, and increment the iterator to the next entry */
static void *find_next_entry_inc(struct trace_iterator *iter)
{
1208
	iter->ent = __find_next_entry(iter, &iter->cpu, &iter->ts);
1209

1210
	if (iter->ent)
1211
		trace_iterator_increment(iter);