trace.c 71.7 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
61
62
63
64
65
66
67
68
69
/* 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;
}
70
71
72
73
74
75
76
77
78

/*
 * 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;

79
80
81
82
83
84
85
86
87
88
89
90
91
92
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();
}

93
static cpumask_var_t __read_mostly	tracing_buffer_mask;
94
95

#define for_each_tracing_cpu(cpu)	\
96
	for_each_cpu(cpu, tracing_buffer_mask)
97

98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
/*
 * 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;

113
114
115
116
117
118
119
120
121
static int tracing_set_tracer(char *buf);

static int __init set_ftrace(char *str)
{
	tracing_set_tracer(str);
	return 1;
}
__setup("ftrace", set_ftrace);

122
123
124
125
126
127
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
128

129
long
130
131
132
133
134
135
136
ns2usecs(cycle_t nsec)
{
	nsec += 500;
	do_div(nsec, 1000);
	return nsec;
}

Ingo Molnar's avatar
Ingo Molnar committed
137
cycle_t ftrace_now(int cpu)
138
{
139
140
141
	u64 ts = ring_buffer_time_stamp(cpu);
	ring_buffer_normalize_time_stamp(cpu, &ts);
	return ts;
142
143
}

Steven Rostedt's avatar
Steven Rostedt committed
144
145
146
147
148
149
150
151
152
153
154
155
/*
 * 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.
 */
156
157
158
159
static struct trace_array	global_trace;

static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu);

Steven Rostedt's avatar
Steven Rostedt committed
160
161
162
163
164
165
166
167
168
169
/*
 * 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.
 */
170
171
172
173
static struct trace_array	max_tr;

static DEFINE_PER_CPU(struct trace_array_cpu, max_data);

Steven Rostedt's avatar
Steven Rostedt committed
174
/* tracer_enabled is used to toggle activation of a tracer */
175
static int			tracer_enabled = 1;
Steven Rostedt's avatar
Steven Rostedt committed
176

177
178
179
180
181
182
183
184
185
186
187
188
189
/**
 * 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
190
/*
191
192
193
 * 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.
194
195
196
197
198
 *
 * 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
199
 */
200
#define TRACE_BUF_SIZE_DEFAULT	1441792UL /* 16384 * 88 (sizeof(entry)) */
201

202
static unsigned long		trace_buf_size = TRACE_BUF_SIZE_DEFAULT;
203

Steven Rostedt's avatar
Steven Rostedt committed
204
/* trace_types holds a link list of available tracers. */
205
static struct tracer		*trace_types __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
206
207

/* current_trace points to the tracer that is currently active */
208
static struct tracer		*current_trace __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
209
210
211
212
213
214

/*
 * 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.
 */
215
216
static int			max_tracer_type_len;

Steven Rostedt's avatar
Steven Rostedt committed
217
218
219
220
221
222
/*
 * 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.
 */
223
static DEFINE_MUTEX(trace_types_lock);
Steven Rostedt's avatar
Steven Rostedt committed
224
225

/* trace_wait is a waitqueue for tasks blocked on trace_poll */
226
227
static DECLARE_WAIT_QUEUE_HEAD(trace_wait);

228
/* trace_flags holds trace_options default values */
229
230
unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK |
	TRACE_ITER_ANNOTATE;
231

Steven Rostedt's avatar
Steven Rostedt committed
232
233
234
235
236
237
/**
 * 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.
 */
238
239
void trace_wake_up(void)
{
Ingo Molnar's avatar
Ingo Molnar committed
240
241
242
243
244
	/*
	 * The runqueue_is_locked() can fail, but this is the best we
	 * have for now:
	 */
	if (!(trace_flags & TRACE_ITER_BLOCK) && !runqueue_is_locked())
245
246
		wake_up(&trace_wait);
}
247

248
static int __init set_buf_size(char *str)
249
{
250
	unsigned long buf_size;
251
252
	int ret;

253
254
	if (!str)
		return 0;
255
	ret = strict_strtoul(str, 0, &buf_size);
256
	/* nr_entries can not be zero */
257
	if (ret < 0 || buf_size == 0)
258
		return 0;
259
	trace_buf_size = buf_size;
260
261
	return 1;
}
262
__setup("trace_buf_size=", set_buf_size);
263

Steven Rostedt's avatar
Steven Rostedt committed
264
265
266
267
268
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

Steven Rostedt's avatar
Steven Rostedt committed
269
/* These must match the bit postions in trace_iterator_flags */
270
271
272
273
274
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
Ingo Molnar's avatar
Ingo Molnar committed
275
	"raw",
276
	"hex",
Ingo Molnar's avatar
Ingo Molnar committed
277
	"bin",
278
	"block",
Ingo Molnar's avatar
Ingo Molnar committed
279
	"stacktrace",
Ingo Molnar's avatar
Ingo Molnar committed
280
	"sched-tree",
281
	"ftrace_printk",
282
	"ftrace_preempt",
283
	"branch",
284
	"annotate",
285
	"userstacktrace",
286
	"sym-userobj",
287
	"printk-msg-only",
288
289
290
	NULL
};

Steven Rostedt's avatar
Steven Rostedt committed
291
292
293
294
295
296
297
298
299
/*
 * 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.
 */
300
301
static raw_spinlock_t ftrace_max_lock =
	(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
302
303
304
305
306
307

/*
 * 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
308
static void
309
310
311
312
313
314
315
316
317
318
319
320
__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;
321
	data->uid = task_uid(tsk);
322
323
324
325
326
327
328
329
	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
330
static void
331
332
333
trace_seq_reset(struct trace_seq *s)
{
	s->len = 0;
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
	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;
354
355
}

Ingo Molnar's avatar
Ingo Molnar committed
356
static void
357
358
359
360
361
362
363
364
365
366
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
367
368
369
370
371
372
373
374
375
/**
 * 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
376
void
377
378
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
379
	struct ring_buffer *buf = tr->buffer;
380

381
	WARN_ON_ONCE(!irqs_disabled());
382
	__raw_spin_lock(&ftrace_max_lock);
383
384
385
386

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

387
	ftrace_disable_cpu();
388
	ring_buffer_reset(tr->buffer);
389
	ftrace_enable_cpu();
390
391

	__update_max_tr(tr, tsk, cpu);
392
	__raw_spin_unlock(&ftrace_max_lock);
393
394
395
396
397
398
399
}

/**
 * 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
400
401
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
402
 */
Ingo Molnar's avatar
Ingo Molnar committed
403
void
404
405
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
406
	int ret;
407

408
	WARN_ON_ONCE(!irqs_disabled());
409
	__raw_spin_lock(&ftrace_max_lock);
410

411
412
	ftrace_disable_cpu();

413
414
415
	ring_buffer_reset(max_tr.buffer);
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

416
417
	ftrace_enable_cpu();

418
	WARN_ON_ONCE(ret && ret != -EAGAIN);
419
420

	__update_max_tr(tr, tsk, cpu);
421
	__raw_spin_unlock(&ftrace_max_lock);
422
423
}

Steven Rostedt's avatar
Steven Rostedt committed
424
425
426
427
428
429
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
430
431
432
433
434
435
436
437
438
439
440
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
441
442
443
444
445
446
447
	/*
	 * 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();
448
	mutex_lock(&trace_types_lock);
Ingo Molnar's avatar
Ingo Molnar committed
449

450
451
	tracing_selftest_running = true;

452
453
454
455
456
457
458
459
460
461
	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;
		}
	}

462
463
464
465
466
467
468
469
	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
470
471
472
473
474
#ifdef CONFIG_FTRACE_STARTUP_TEST
	if (type->selftest) {
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
		int i;
475

Steven Rostedt's avatar
Steven Rostedt committed
476
477
478
479
480
481
482
		/*
		 * 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
483
		for_each_tracing_cpu(i)
484
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
485

Steven Rostedt's avatar
Steven Rostedt committed
486
487
488
489
490
491
492
493
494
495
		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
496
		/* Only reset on passing, to avoid touching corrupted buffers */
Ingo Molnar's avatar
Ingo Molnar committed
497
		for_each_tracing_cpu(i)
498
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
499

Steven Rostedt's avatar
Steven Rostedt committed
500
501
502
503
		printk(KERN_CONT "PASSED\n");
	}
#endif

504
505
506
507
508
	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
509

510
 out:
511
	tracing_selftest_running = false;
512
	mutex_unlock(&trace_types_lock);
Ingo Molnar's avatar
Ingo Molnar committed
513
	lock_kernel();
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545

	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;
	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);
}

546
void tracing_reset(struct trace_array *tr, int cpu)
547
{
548
	ftrace_disable_cpu();
549
	ring_buffer_reset_cpu(tr->buffer, cpu);
550
	ftrace_enable_cpu();
551
552
}

553
554
555
556
557
558
559
560
561
562
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);
}

563
564
565
566
567
568
#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);
569
570
571

/* temporary disable recording */
atomic_t trace_record_cmdline_disabled __read_mostly;
572
573
574
575
576
577
578
579

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;
}

580
581
582
static int trace_stop_count;
static DEFINE_SPINLOCK(tracing_start_lock);

583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
/**
 * 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();
}

598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
/**
 * 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);
613
614
615
616
617
618
	if (--trace_stop_count) {
		if (trace_stop_count < 0) {
			/* Someone screwed up their debugging */
			WARN_ON_ONCE(1);
			trace_stop_count = 0;
		}
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
		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
664
void trace_stop_cmdline_recording(void);
665

Ingo Molnar's avatar
Ingo Molnar committed
666
static void trace_save_cmdline(struct task_struct *tsk)
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
{
	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);
}

701
char *trace_find_cmdline(int pid)
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
{
	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
722
void tracing_record_cmdline(struct task_struct *tsk)
723
724
725
726
727
728
729
{
	if (atomic_read(&trace_record_cmdline_disabled))
		return;

	trace_save_cmdline(tsk);
}

730
void
731
732
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
733
734
735
{
	struct task_struct *tsk = current;

736
737
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
738
	entry->tgid               	= (tsk) ? tsk->tgid : 0;
739
	entry->flags =
740
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
741
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
742
743
744
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
745
746
747
748
749
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}

Ingo Molnar's avatar
Ingo Molnar committed
750
void
751
trace_function(struct trace_array *tr, struct trace_array_cpu *data,
752
753
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
754
{
755
	struct ring_buffer_event *event;
756
	struct ftrace_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
757
	unsigned long irq_flags;
758

759
760
761
762
	/* If we are reading the ring buffer, don't trace */
	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

763
764
765
766
767
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
768
	tracing_generic_entry_update(&entry->ent, flags, pc);
769
770
771
	entry->ent.type			= TRACE_FN;
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
772
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
773
774
}

775
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
static void __trace_graph_entry(struct trace_array *tr,
				struct trace_array_cpu *data,
				struct ftrace_graph_ent *trace,
				unsigned long flags,
				int pc)
{
	struct ring_buffer_event *event;
	struct ftrace_graph_ent_entry *entry;
	unsigned long irq_flags;

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

	event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
	tracing_generic_entry_update(&entry->ent, flags, pc);
	entry->ent.type			= TRACE_GRAPH_ENT;
	entry->graph_ent			= *trace;
	ring_buffer_unlock_commit(global_trace.buffer, event, irq_flags);
}

static void __trace_graph_return(struct trace_array *tr,
801
				struct trace_array_cpu *data,
802
				struct ftrace_graph_ret *trace,
803
804
805
806
				unsigned long flags,
				int pc)
{
	struct ring_buffer_event *event;
807
	struct ftrace_graph_ret_entry *entry;
808
809
810
811
812
813
814
815
816
817
818
	unsigned long irq_flags;

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

	event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
	tracing_generic_entry_update(&entry->ent, flags, pc);
819
820
	entry->ent.type			= TRACE_GRAPH_RET;
	entry->ret				= *trace;
821
822
823
824
	ring_buffer_unlock_commit(global_trace.buffer, event, irq_flags);
}
#endif

Ingo Molnar's avatar
Ingo Molnar committed
825
void
Ingo Molnar's avatar
Ingo Molnar committed
826
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
827
828
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
Ingo Molnar's avatar
Ingo Molnar committed
829
830
{
	if (likely(!atomic_read(&data->disabled)))
831
		trace_function(tr, data, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
832
833
}

834
835
836
837
static void __ftrace_trace_stack(struct trace_array *tr,
				 struct trace_array_cpu *data,
				 unsigned long flags,
				 int skip, int pc)
Ingo Molnar's avatar
Ingo Molnar committed
838
{
839
#ifdef CONFIG_STACKTRACE
840
	struct ring_buffer_event *event;
841
	struct stack_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
842
	struct stack_trace trace;
843
	unsigned long irq_flags;
Ingo Molnar's avatar
Ingo Molnar committed
844

845
846
847
848
849
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
850
	tracing_generic_entry_update(&entry->ent, flags, pc);
851
	entry->ent.type		= TRACE_STACK;
Ingo Molnar's avatar
Ingo Molnar committed
852

853
	memset(&entry->caller, 0, sizeof(entry->caller));
Ingo Molnar's avatar
Ingo Molnar committed
854
855
856
857

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
858
	trace.entries		= entry->caller;
Ingo Molnar's avatar
Ingo Molnar committed
859
860

	save_stack_trace(&trace);
861
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
862
#endif
Ingo Molnar's avatar
Ingo Molnar committed
863
864
}

865
866
867
868
869
870
871
872
873
874
875
static void ftrace_trace_stack(struct trace_array *tr,
			       struct trace_array_cpu *data,
			       unsigned long flags,
			       int skip, int pc)
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

	__ftrace_trace_stack(tr, data, flags, skip, pc);
}

876
877
878
void __trace_stack(struct trace_array *tr,
		   struct trace_array_cpu *data,
		   unsigned long flags,
879
		   int skip, int pc)
880
{
881
	__ftrace_trace_stack(tr, data, flags, skip, pc);
882
883
}

884
885
886
887
static void ftrace_trace_userstack(struct trace_array *tr,
		   struct trace_array_cpu *data,
		   unsigned long flags, int pc)
{
888
#ifdef CONFIG_STACKTRACE
889
	struct ring_buffer_event *event;
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
	struct userstack_entry *entry;
	struct stack_trace trace;
	unsigned long irq_flags;

	if (!(trace_flags & TRACE_ITER_USERSTACKTRACE))
		return;

	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
	tracing_generic_entry_update(&entry->ent, flags, pc);
	entry->ent.type		= TRACE_USER_STACK;

	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);
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
914
#endif
915
916
917
918
919
920
921
922
923
}

void __trace_userstack(struct trace_array *tr,
		   struct trace_array_cpu *data,
		   unsigned long flags)
{
	ftrace_trace_userstack(tr, data, flags, preempt_count());
}

924
925
926
927
static void
ftrace_trace_special(void *__tr, void *__data,
		     unsigned long arg1, unsigned long arg2, unsigned long arg3,
		     int pc)
Ingo Molnar's avatar
Ingo Molnar committed
928
{
929
	struct ring_buffer_event *event;
Ingo Molnar's avatar
Ingo Molnar committed
930
931
	struct trace_array_cpu *data = __data;
	struct trace_array *tr = __tr;
932
	struct special_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
933
934
	unsigned long irq_flags;

935
936
937
938
939
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
940
	tracing_generic_entry_update(&entry->ent, 0, pc);
941
942
943
944
	entry->ent.type			= TRACE_SPECIAL;
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
945
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
946
	ftrace_trace_stack(tr, data, irq_flags, 4, pc);
947
	ftrace_trace_userstack(tr, data, irq_flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
948
949
950
951

	trace_wake_up();
}

952
953
954
955
956
957
958
void
__trace_special(void *__tr, void *__data,
		unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	ftrace_trace_special(__tr, __data, arg1, arg2, arg3, preempt_count());
}

Ingo Molnar's avatar
Ingo Molnar committed
959
void
960
961
tracing_sched_switch_trace(struct trace_array *tr,
			   struct trace_array_cpu *data,
Ingo Molnar's avatar
Ingo Molnar committed
962
963
			   struct task_struct *prev,
			   struct task_struct *next,
964
			   unsigned long flags, int pc)
965
{
966
	struct ring_buffer_event *event;
967
	struct ctx_switch_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
968
	unsigned long irq_flags;
969

970
971
972
973
974
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					   &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
975
	tracing_generic_entry_update(&entry->ent, flags, pc);
976
977
978
979
980
981
982
983
	entry->ent.type			= TRACE_CTX;
	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);
984
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
985
	ftrace_trace_stack(tr, data, flags, 5, pc);
986
	ftrace_trace_userstack(tr, data, flags, pc);
987
988
}

989
990
991
void
tracing_sched_wakeup_trace(struct trace_array *tr,
			   struct trace_array_cpu *data,
Ingo Molnar's avatar
Ingo Molnar committed
992
993
			   struct task_struct *wakee,
			   struct task_struct *curr,
994
			   unsigned long flags, int pc)
995
{
996
	struct ring_buffer_event *event;
997
	struct ctx_switch_entry *entry;
998
999
	unsigned long irq_flags;

1000
1001
1002
1003
1004
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					   &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1005
	tracing_generic_entry_update(&entry->ent, flags, pc);
1006
1007
1008
1009
1010
1011
1012
1013
	entry->ent.type			= TRACE_WAKE;
	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);
1014
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
1015
	ftrace_trace_stack(tr, data, flags, 6, pc);
1016
	ftrace_trace_userstack(tr, data, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
1017
1018

	trace_wake_up();
1019
1020
}

1021
1022
1023
1024
1025
void
ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1026
	unsigned long flags;
1027
	int cpu;
1028
	int pc;
1029

1030
	if (tracing_disabled)
1031
1032
		return;

1033
	pc = preempt_count();
1034
	local_irq_save(flags);
1035
1036
1037
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

1038
	if (likely(atomic_inc_return(&data->disabled) == 1))
1039
		ftrace_trace_special(tr, data, arg1, arg2, arg3, pc);
1040

1041
1042
	atomic_dec(&data->disabled);
	local_irq_restore(flags);
1043
1044
}

1045
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
1046
int trace_graph_entry(struct ftrace_graph_ent *trace)
1047
1048
1049
1050
1051
1052
1053
1054
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	unsigned long flags;
	long disabled;
	int cpu;
	int pc;

1055
1056
1057
	if (!ftrace_trace_task(current))
		return 0;

1058
1059
1060
	if (!ftrace_graph_addr(trace->func))
		return 0;

1061
	local_irq_save(flags);
1062
1063
1064
1065
1066
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);
	if (likely(disabled == 1)) {
		pc = preempt_count();
1067
1068
		__trace_graph_entry(tr, data, trace, flags, pc);
	}
1069
1070
1071
	/* Only do the atomic if it is not already set */
	if (!test_tsk_trace_graph(current))
		set_tsk_trace_graph(current);
1072
	atomic_dec(&data->disabled);
1073
	local_irq_restore(flags);
1074
1075

	return 1;
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
}

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;

1087
	local_irq_save(flags);
1088
1089
1090
1091
1092
1093
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);
	if (likely(disabled == 1)) {
		pc = preempt_count();
		__trace_graph_return(tr, data, trace, flags, pc);
1094
	}
1095
1096
	if (!trace->depth)
		clear_tsk_trace_graph(current);
1097
	atomic_dec(&data->disabled);
1098
	local_irq_restore(flags);
1099
}
1100
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
1101

1102
1103
enum trace_file_type {
	TRACE_FILE_LAT_FMT	= 1,
1104
	TRACE_FILE_ANNOTATE	= 2,
1105
1106
};

1107
static void trace_iterator_increment(struct trace_iterator *iter)
1108
{
1109
1110
1111
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();

1112
	iter->idx++;
1113
1114
1115
1116
	if (iter->buffer_iter[iter->cpu])
		ring_buffer_read(iter->buffer_iter[iter->cpu], NULL);

	ftrace_enable_cpu();
1117
1118
}

Ingo Molnar's avatar
Ingo Molnar committed
1119
static struct trace_entry *
1120
peek_next_entry(struct trace_iterator *iter, int cpu, u64 *ts)
1121
{
1122
1123
	struct ring_buffer_event *event;
	struct ring_buffer_iter *buf_iter = iter->buffer_iter[cpu];
1124

1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
	/* 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();

1135
	return event ? ring_buffer_event_data(event) : NULL;
1136
}
1137

1138
static struct trace_entry *
1139
__find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts)
1140
{
1141
	struct ring_buffer *buffer = iter->tr->buffer;
1142
	struct trace_entry *ent, *next = NULL;
1143
	u64 next_ts = 0, ts;
1144
1145
1146
	int next_cpu = -1;
	int cpu;

1147
	for_each_tracing_cpu(cpu) {
1148

1149
1150
		if (ring_buffer_empty_cpu(buffer, cpu))
			continue;
1151

1152
		ent = peek_next_entry(iter, cpu, &ts);
1153

Ingo Molnar's avatar
Ingo Molnar committed
1154
1155
1156
		/*
		 * Pick the entry with the smallest timestamp:
		 */
1157
		if (ent && (!next || ts < next_ts)) {
1158
1159
			next = ent;
			next_cpu = cpu;
1160
			next_ts = ts;
1161
1162
1163
1164
1165
1166
		}
	}

	if (ent_cpu)
		*ent_cpu = next_cpu;

1167
1168
1169
	if (ent_ts)
		*ent_ts = next_ts;

1170
1171
1172
	return next;
}

1173
1174
/* Find the next real entry, without updating the iterator itself */
static struct trace_entry *
1175
find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts)
1176
{
1177
	return __find_next_entry(iter, ent_cpu, ent_ts);
1178
1179
1180
1181
1182
}

/* Find the next real entry, and increment the iterator to the next entry */
static void *find_next_entry_inc(struct trace_iterator *iter)
{
1183
	iter->ent = __find_next_entry(iter, &iter->cpu, &iter->ts);
1184

1185
	if (iter->ent)
1186
		trace_iterator_increment(iter);
1187

1188
	return iter->ent ? iter : NULL;
1189
}
1190

Ingo Molnar's avatar
Ingo Molnar committed
1191
static void trace_consume(struct trace_iterator *iter)
1192
{
1193
1194
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();
1195
	ring_buffer_consume(iter->tr->buffer, iter->cpu, &iter->ts);
1196
	ftrace_enable_cpu();
1197
1198
}

Ingo Molnar's avatar
Ingo Molnar committed
1199
static void *s_next(struct seq_file *m, void *v, loff_t *pos)
1200
1201
1202
{
	struct trace_iterator *iter = m->private;
	int i = (int)*pos;
Ingo Molnar's avatar
Ingo Molnar committed
1203
	void *ent;
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228

	(*pos)++;

	/* can't go backwards */
	if (iter->idx > i)
		return NULL;

	if (iter->idx < 0)
		ent = find_next_entry_inc(iter);
	else
		ent = iter;

	while (ent && iter->idx < i)
		ent = find_next_entry_inc(iter);

	iter->pos = *pos;

	return ent;
}

static void *s_start(struct seq_file *m, loff_t *pos)
{
	struct trace_iterator *iter = m->private;
	void *p = NULL;
	loff_t l = 0;
1229
	int cpu;
1230
1231
1232

	mutex_lock(&trace_types_lock);

1233
1234
	if (!current_trace || current_trace != iter->trace) {
		mutex_unlock(&trace_types_lock);
1235
		return NULL;
1236
	}
1237
1238
1239
1240
1241
1242
1243
1244

	atomic_inc(&trace_record_cmdline_disabled);

	if (*pos != iter