trace.c 73.6 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
40
#include "trace.h"

41
42
#define TRACE_BUFFER_FLAGS	(RB_FL_OVERWRITE)

43
44
45
unsigned long __read_mostly	tracing_max_latency = (cycle_t)ULONG_MAX;
unsigned long __read_mostly	tracing_thresh;

46
47
48
49
50
51
52
53
54
55
56
57
58
59
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();
}

60
61
62
63
64
static cpumask_t __read_mostly		tracing_buffer_mask;

#define for_each_tracing_cpu(cpu)	\
	for_each_cpu_mask(cpu, tracing_buffer_mask)

Steven Rostedt's avatar
Steven Rostedt committed
65
66
static int tracing_disabled = 1;

67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
/*
 * 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;

82
83
84
85
86
87
88
89
90
static int tracing_set_tracer(char *buf);

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

91
92
93
94
95
96
97
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);

98
long
99
100
101
102
103
104
105
ns2usecs(cycle_t nsec)
{
	nsec += 500;
	do_div(nsec, 1000);
	return nsec;
}

Ingo Molnar's avatar
Ingo Molnar committed
106
cycle_t ftrace_now(int cpu)
107
{
108
109
110
	u64 ts = ring_buffer_time_stamp(cpu);
	ring_buffer_normalize_time_stamp(cpu, &ts);
	return ts;
111
112
}

Steven Rostedt's avatar
Steven Rostedt committed
113
114
115
116
117
118
119
120
121
122
123
124
/*
 * 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.
 */
125
126
127
128
static struct trace_array	global_trace;

static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu);

Steven Rostedt's avatar
Steven Rostedt committed
129
130
131
132
133
134
135
136
137
138
/*
 * 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.
 */
139
140
141
142
static struct trace_array	max_tr;

static DEFINE_PER_CPU(struct trace_array_cpu, max_data);

Steven Rostedt's avatar
Steven Rostedt committed
143
/* tracer_enabled is used to toggle activation of a tracer */
144
static int			tracer_enabled = 1;
Steven Rostedt's avatar
Steven Rostedt committed
145

146
147
148
/* function tracing enabled */
int				ftrace_function_enabled;

Steven Rostedt's avatar
Steven Rostedt committed
149
/*
150
151
152
 * 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.
153
154
155
156
157
 *
 * 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
158
 */
159
#define TRACE_BUF_SIZE_DEFAULT	1441792UL /* 16384 * 88 (sizeof(entry)) */
160

161
static unsigned long		trace_buf_size = TRACE_BUF_SIZE_DEFAULT;
162

Steven Rostedt's avatar
Steven Rostedt committed
163
/* trace_types holds a link list of available tracers. */
164
static struct tracer		*trace_types __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
165
166

/* current_trace points to the tracer that is currently active */
167
static struct tracer		*current_trace __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
168
169
170
171
172
173

/*
 * 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.
 */
174
175
static int			max_tracer_type_len;

Steven Rostedt's avatar
Steven Rostedt committed
176
177
178
179
180
181
/*
 * 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.
 */
182
static DEFINE_MUTEX(trace_types_lock);
Steven Rostedt's avatar
Steven Rostedt committed
183
184

/* trace_wait is a waitqueue for tasks blocked on trace_poll */
185
186
static DECLARE_WAIT_QUEUE_HEAD(trace_wait);

Steven Rostedt's avatar
Steven Rostedt committed
187
/* trace_flags holds iter_ctrl options */
188
189
unsigned long trace_flags = TRACE_ITER_PRINT_PARENT;

Steven Rostedt's avatar
Steven Rostedt committed
190
191
192
193
194
195
/**
 * 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.
 */
196
197
void trace_wake_up(void)
{
Ingo Molnar's avatar
Ingo Molnar committed
198
199
200
201
202
	/*
	 * The runqueue_is_locked() can fail, but this is the best we
	 * have for now:
	 */
	if (!(trace_flags & TRACE_ITER_BLOCK) && !runqueue_is_locked())
203
204
		wake_up(&trace_wait);
}
205

206
static int __init set_buf_size(char *str)
207
{
208
	unsigned long buf_size;
209
210
	int ret;

211
212
	if (!str)
		return 0;
213
	ret = strict_strtoul(str, 0, &buf_size);
214
	/* nr_entries can not be zero */
215
	if (ret < 0 || buf_size == 0)
216
		return 0;
217
	trace_buf_size = buf_size;
218
219
	return 1;
}
220
__setup("trace_buf_size=", set_buf_size);
221

Steven Rostedt's avatar
Steven Rostedt committed
222
223
224
225
226
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

Steven Rostedt's avatar
Steven Rostedt committed
227
228
229
230
/*
 * TRACE_ITER_SYM_MASK masks the options in trace_flags that
 * control the output of kernel symbols.
 */
231
232
233
#define TRACE_ITER_SYM_MASK \
	(TRACE_ITER_PRINT_PARENT|TRACE_ITER_SYM_OFFSET|TRACE_ITER_SYM_ADDR)

Steven Rostedt's avatar
Steven Rostedt committed
234
/* These must match the bit postions in trace_iterator_flags */
235
236
237
238
239
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
Ingo Molnar's avatar
Ingo Molnar committed
240
	"raw",
241
	"hex",
Ingo Molnar's avatar
Ingo Molnar committed
242
	"bin",
243
	"block",
Ingo Molnar's avatar
Ingo Molnar committed
244
	"stacktrace",
Ingo Molnar's avatar
Ingo Molnar committed
245
	"sched-tree",
246
	"ftrace_printk",
247
	"ftrace_preempt",
248
249
250
	NULL
};

Steven Rostedt's avatar
Steven Rostedt committed
251
252
253
254
255
256
257
258
259
/*
 * 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.
 */
260
261
static raw_spinlock_t ftrace_max_lock =
	(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
262
263
264
265
266
267

/*
 * 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
268
static void
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
__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;
	data->uid = tsk->uid;
	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);
}

Steven Rostedt's avatar
Steven Rostedt committed
290
291
292
293
294
295
296
297
298
299
300
/**
 * trace_seq_printf - sequence printing of trace information
 * @s: trace sequence descriptor
 * @fmt: printf format string
 *
 * The tracer may use either sequence operations or its own
 * copy to user routines. To simplify formating of a trace
 * trace_seq_printf is used to store strings into a special
 * buffer (@s). Then the output may be either used by
 * the sequencer or pulled into another buffer.
 */
301
int
302
303
304
305
trace_seq_printf(struct trace_seq *s, const char *fmt, ...)
{
	int len = (PAGE_SIZE - 1) - s->len;
	va_list ap;
306
	int ret;
307
308
309
310
311

	if (!len)
		return 0;

	va_start(ap, fmt);
312
	ret = vsnprintf(s->buffer + s->len, len, fmt, ap);
313
314
	va_end(ap);

315
	/* If we can't write it all, don't bother writing anything */
316
	if (ret >= len)
317
318
319
		return 0;

	s->len += ret;
320
321
322
323

	return len;
}

Steven Rostedt's avatar
Steven Rostedt committed
324
325
326
327
328
329
330
331
332
333
/**
 * trace_seq_puts - trace sequence printing of simple string
 * @s: trace sequence descriptor
 * @str: simple string to record
 *
 * The tracer may use either the sequence operations or its own
 * copy to user routines. This function records a simple string
 * into a special buffer (@s) for later retrieval by a sequencer
 * or other mechanism.
 */
Ingo Molnar's avatar
Ingo Molnar committed
334
static int
335
336
337
338
339
trace_seq_puts(struct trace_seq *s, const char *str)
{
	int len = strlen(str);

	if (len > ((PAGE_SIZE - 1) - s->len))
340
		return 0;
341
342
343
344
345
346
347

	memcpy(s->buffer + s->len, str, len);
	s->len += len;

	return len;
}

Ingo Molnar's avatar
Ingo Molnar committed
348
static int
349
350
351
352
353
354
355
356
357
358
trace_seq_putc(struct trace_seq *s, unsigned char c)
{
	if (s->len >= (PAGE_SIZE - 1))
		return 0;

	s->buffer[s->len++] = c;

	return 1;
}

Ingo Molnar's avatar
Ingo Molnar committed
359
static int
Ingo Molnar's avatar
Ingo Molnar committed
360
361
362
363
364
365
366
367
368
369
370
trace_seq_putmem(struct trace_seq *s, void *mem, size_t len)
{
	if (len > ((PAGE_SIZE - 1) - s->len))
		return 0;

	memcpy(s->buffer + s->len, mem, len);
	s->len += len;

	return len;
}

371
372
#define MAX_MEMHEX_BYTES	8
#define HEX_CHARS		(MAX_MEMHEX_BYTES*2 + 1)
373

Ingo Molnar's avatar
Ingo Molnar committed
374
static int
375
376
377
trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len)
{
	unsigned char hex[HEX_CHARS];
Thomas Gleixner's avatar
Thomas Gleixner committed
378
	unsigned char *data = mem;
379
380
381
382
383
384
385
	int i, j;

#ifdef __BIG_ENDIAN
	for (i = 0, j = 0; i < len; i++) {
#else
	for (i = len-1, j = 0; i >= 0; i--) {
#endif
386
387
		hex[j++] = hex_asc_hi(data[i]);
		hex[j++] = hex_asc_lo(data[i]);
388
	}
Thomas Gleixner's avatar
Thomas Gleixner committed
389
	hex[j++] = ' ';
390
391
392
393

	return trace_seq_putmem(s, hex, j);
}

Ingo Molnar's avatar
Ingo Molnar committed
394
static void
395
396
397
trace_seq_reset(struct trace_seq *s)
{
	s->len = 0;
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
	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;
418
419
}

Ingo Molnar's avatar
Ingo Molnar committed
420
static void
421
422
423
424
425
426
427
428
429
430
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
431
432
433
434
435
436
437
438
439
/**
 * 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
440
void
441
442
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
443
	struct ring_buffer *buf = tr->buffer;
444

445
	WARN_ON_ONCE(!irqs_disabled());
446
	__raw_spin_lock(&ftrace_max_lock);
447
448
449
450

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

451
	ftrace_disable_cpu();
452
	ring_buffer_reset(tr->buffer);
453
	ftrace_enable_cpu();
454
455

	__update_max_tr(tr, tsk, cpu);
456
	__raw_spin_unlock(&ftrace_max_lock);
457
458
459
460
461
462
463
}

/**
 * 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
464
465
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
466
 */
Ingo Molnar's avatar
Ingo Molnar committed
467
void
468
469
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
470
	int ret;
471

472
	WARN_ON_ONCE(!irqs_disabled());
473
	__raw_spin_lock(&ftrace_max_lock);
474

475
476
	ftrace_disable_cpu();

477
478
479
	ring_buffer_reset(max_tr.buffer);
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

480
481
	ftrace_enable_cpu();

482
	WARN_ON_ONCE(ret);
483
484

	__update_max_tr(tr, tsk, cpu);
485
	__raw_spin_unlock(&ftrace_max_lock);
486
487
}

Steven Rostedt's avatar
Steven Rostedt committed
488
489
490
491
492
493
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
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;
	}

	mutex_lock(&trace_types_lock);
	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;
		}
	}

Steven Rostedt's avatar
Steven Rostedt committed
516
517
518
519
520
521
522
523
524
525
526
527
528
#ifdef CONFIG_FTRACE_STARTUP_TEST
	if (type->selftest) {
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
		int saved_ctrl = tr->ctrl;
		int i;
		/*
		 * 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.
		 */
529
		for_each_tracing_cpu(i) {
530
			tracing_reset(tr, i);
Steven Rostedt's avatar
Steven Rostedt committed
531
532
533
534
535
536
537
538
539
540
541
542
543
		}
		current_trace = type;
		tr->ctrl = 0;
		/* 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;
		tr->ctrl = saved_ctrl;
		if (ret) {
			printk(KERN_CONT "FAILED!\n");
			goto out;
		}
Steven Rostedt's avatar
Steven Rostedt committed
544
		/* Only reset on passing, to avoid touching corrupted buffers */
545
		for_each_tracing_cpu(i) {
546
			tracing_reset(tr, i);
Steven Rostedt's avatar
Steven Rostedt committed
547
		}
Steven Rostedt's avatar
Steven Rostedt committed
548
549
550
551
		printk(KERN_CONT "PASSED\n");
	}
#endif

552
553
554
555
556
	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
557

558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
 out:
	mutex_unlock(&trace_types_lock);

	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);
}

592
void tracing_reset(struct trace_array *tr, int cpu)
593
{
594
	ftrace_disable_cpu();
595
	ring_buffer_reset_cpu(tr->buffer, cpu);
596
	ftrace_enable_cpu();
597
598
599
600
601
602
603
604
}

#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);
605
606
607

/* temporary disable recording */
atomic_t trace_record_cmdline_disabled __read_mostly;
608
609
610
611
612
613
614
615

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;
}

Ingo Molnar's avatar
Ingo Molnar committed
616
void trace_stop_cmdline_recording(void);
617

Ingo Molnar's avatar
Ingo Molnar committed
618
static void trace_save_cmdline(struct task_struct *tsk)
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
{
	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);
}

Ingo Molnar's avatar
Ingo Molnar committed
653
static char *trace_find_cmdline(int pid)
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
{
	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
674
void tracing_record_cmdline(struct task_struct *tsk)
675
676
677
678
679
680
681
{
	if (atomic_read(&trace_record_cmdline_disabled))
		return;

	trace_save_cmdline(tsk);
}

682
void
683
684
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
685
686
687
{
	struct task_struct *tsk = current;

688
689
690
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
	entry->flags =
691
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
692
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
693
694
695
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
696
697
698
699
700
		((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
701
void
702
trace_function(struct trace_array *tr, struct trace_array_cpu *data,
703
704
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
705
{
706
	struct ring_buffer_event *event;
707
	struct ftrace_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
708
	unsigned long irq_flags;
709

710
711
712
713
	/* If we are reading the ring buffer, don't trace */
	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

714
715
716
717
718
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
719
	tracing_generic_entry_update(&entry->ent, flags, pc);
720
721
722
	entry->ent.type			= TRACE_FN;
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
723
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
724
725
}

Ingo Molnar's avatar
Ingo Molnar committed
726
void
Ingo Molnar's avatar
Ingo Molnar committed
727
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
728
729
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
Ingo Molnar's avatar
Ingo Molnar committed
730
731
{
	if (likely(!atomic_read(&data->disabled)))
732
		trace_function(tr, data, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
733
734
}

735
736
737
738
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
739
{
740
#ifdef CONFIG_STACKTRACE
741
	struct ring_buffer_event *event;
742
	struct stack_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
743
	struct stack_trace trace;
744
	unsigned long irq_flags;
Ingo Molnar's avatar
Ingo Molnar committed
745
746
747
748

	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

749
750
751
752
753
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
754
	tracing_generic_entry_update(&entry->ent, flags, pc);
755
	entry->ent.type		= TRACE_STACK;
Ingo Molnar's avatar
Ingo Molnar committed
756

757
	memset(&entry->caller, 0, sizeof(entry->caller));
Ingo Molnar's avatar
Ingo Molnar committed
758
759
760
761

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
762
	trace.entries		= entry->caller;
Ingo Molnar's avatar
Ingo Molnar committed
763
764

	save_stack_trace(&trace);
765
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
766
#endif
Ingo Molnar's avatar
Ingo Molnar committed
767
768
}

769
770
771
772
773
774
775
776
777
778
779
780
void __trace_stack(struct trace_array *tr,
		   struct trace_array_cpu *data,
		   unsigned long flags,
		   int skip)
{
	ftrace_trace_stack(tr, data, flags, skip, preempt_count());
}

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
781
{
782
	struct ring_buffer_event *event;
Ingo Molnar's avatar
Ingo Molnar committed
783
784
	struct trace_array_cpu *data = __data;
	struct trace_array *tr = __tr;
785
	struct special_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
786
787
	unsigned long irq_flags;

788
789
790
791
792
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
793
	tracing_generic_entry_update(&entry->ent, 0, pc);
794
795
796
797
	entry->ent.type			= TRACE_SPECIAL;
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
798
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
799
	ftrace_trace_stack(tr, data, irq_flags, 4, pc);
Ingo Molnar's avatar
Ingo Molnar committed
800
801
802
803

	trace_wake_up();
}

804
805
806
807
808
809
810
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
811
void
812
813
tracing_sched_switch_trace(struct trace_array *tr,
			   struct trace_array_cpu *data,
Ingo Molnar's avatar
Ingo Molnar committed
814
815
			   struct task_struct *prev,
			   struct task_struct *next,
816
			   unsigned long flags, int pc)
817
{
818
	struct ring_buffer_event *event;
819
	struct ctx_switch_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
820
	unsigned long irq_flags;
821

822
823
824
825
826
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					   &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
827
	tracing_generic_entry_update(&entry->ent, flags, pc);
828
829
830
831
832
833
834
835
	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);
836
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
837
	ftrace_trace_stack(tr, data, flags, 5, pc);
838
839
}

840
841
842
void
tracing_sched_wakeup_trace(struct trace_array *tr,
			   struct trace_array_cpu *data,
Ingo Molnar's avatar
Ingo Molnar committed
843
844
			   struct task_struct *wakee,
			   struct task_struct *curr,
845
			   unsigned long flags, int pc)
846
{
847
	struct ring_buffer_event *event;
848
	struct ctx_switch_entry *entry;
849
850
	unsigned long irq_flags;

851
852
853
854
855
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					   &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
856
	tracing_generic_entry_update(&entry->ent, flags, pc);
857
858
859
860
861
862
863
864
	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);
865
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
866
	ftrace_trace_stack(tr, data, flags, 6, pc);
Ingo Molnar's avatar
Ingo Molnar committed
867
868

	trace_wake_up();
869
870
}

871
872
873
874
875
876
void
ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	int cpu;
877
	int pc;
878

879
	if (tracing_disabled || !tr->ctrl)
880
881
		return;

882
883
	pc = preempt_count();
	preempt_disable_notrace();
884
885
886
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

887
	if (likely(!atomic_read(&data->disabled)))
888
		ftrace_trace_special(tr, data, arg1, arg2, arg3, pc);
889

890
	preempt_enable_notrace();
891
892
}

893
#ifdef CONFIG_FUNCTION_TRACER
Ingo Molnar's avatar
Ingo Molnar committed
894
static void
895
function_trace_call_preempt_only(unsigned long ip, unsigned long parent_ip)
Ingo Molnar's avatar
Ingo Molnar committed
896
897
898
899
900
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	unsigned long flags;
	long disabled;
901
902
	int cpu, resched;
	int pc;
Ingo Molnar's avatar
Ingo Molnar committed
903

904
	if (unlikely(!ftrace_function_enabled))
Ingo Molnar's avatar
Ingo Molnar committed
905
906
		return;

907
	pc = preempt_count();
908
	resched = ftrace_preempt_disable();
909
	local_save_flags(flags);
Ingo Molnar's avatar
Ingo Molnar committed
910
911
912
913
914
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);

	if (likely(disabled == 1))
915
		trace_function(tr, data, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
916
917

	atomic_dec(&data->disabled);
918
	ftrace_preempt_enable(resched);
Ingo Molnar's avatar
Ingo Molnar committed
919
920
}

921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
static void
function_trace_call(unsigned long ip, unsigned long parent_ip)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	unsigned long flags;
	long disabled;
	int cpu;
	int pc;

	if (unlikely(!ftrace_function_enabled))
		return;

	/*
	 * Need to use raw, since this must be called before the
	 * recursive protection is performed.
	 */
	raw_local_irq_save(flags);
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);

	if (likely(disabled == 1)) {
		pc = preempt_count();
		trace_function(tr, data, ip, parent_ip, flags, pc);
	}

	atomic_dec(&data->disabled);
	raw_local_irq_restore(flags);
}

Ingo Molnar's avatar
Ingo Molnar committed
952
953
954
955
956
static struct ftrace_ops trace_ops __read_mostly =
{
	.func = function_trace_call,
};

Ingo Molnar's avatar
Ingo Molnar committed
957
void tracing_start_function_trace(void)
Ingo Molnar's avatar
Ingo Molnar committed
958
{
959
	ftrace_function_enabled = 0;
960
961
962
963
964
965

	if (trace_flags & TRACE_ITER_PREEMPTONLY)
		trace_ops.func = function_trace_call_preempt_only;
	else
		trace_ops.func = function_trace_call;

Ingo Molnar's avatar
Ingo Molnar committed
966
	register_ftrace_function(&trace_ops);
967
968
	if (tracer_enabled)
		ftrace_function_enabled = 1;
Ingo Molnar's avatar
Ingo Molnar committed
969
970
}

Ingo Molnar's avatar
Ingo Molnar committed
971
void tracing_stop_function_trace(void)
Ingo Molnar's avatar
Ingo Molnar committed
972
{
973
	ftrace_function_enabled = 0;
Ingo Molnar's avatar
Ingo Molnar committed
974
975
976
977
	unregister_ftrace_function(&trace_ops);
}
#endif

978
979
980
981
enum trace_file_type {
	TRACE_FILE_LAT_FMT	= 1,
};

982
983
static void trace_iterator_increment(struct trace_iterator *iter, int cpu)
{
984
985
986
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();

987
	iter->idx++;
988
989
990
991
	if (iter->buffer_iter[iter->cpu])
		ring_buffer_read(iter->buffer_iter[iter->cpu], NULL);

	ftrace_enable_cpu();
992
993
}

Ingo Molnar's avatar
Ingo Molnar committed
994
static struct trace_entry *
995
peek_next_entry(struct trace_iterator *iter, int cpu, u64 *ts)
996
{
997
998
	struct ring_buffer_event *event;
	struct ring_buffer_iter *buf_iter = iter->buffer_iter[cpu];
999

1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
	/* 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();

1010
	return event ? ring_buffer_event_data(event) : NULL;
1011
}
1012

1013
static struct trace_entry *
1014
__find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts)
1015
{
1016
	struct ring_buffer *buffer = iter->tr->buffer;
1017
	struct trace_entry *ent, *next = NULL;
1018
	u64 next_ts = 0, ts;
1019
1020
1021
	int next_cpu = -1;
	int cpu;

1022
	for_each_tracing_cpu(cpu) {
1023

1024
1025
		if (ring_buffer_empty_cpu(buffer, cpu))
			continue;
1026

1027
		ent = peek_next_entry(iter, cpu, &ts);
1028

Ingo Molnar's avatar
Ingo Molnar committed
1029
1030
1031
		/*
		 * Pick the entry with the smallest timestamp:
		 */
1032
		if (ent && (!next || ts < next_ts)) {
1033
1034
			next = ent;
			next_cpu = cpu;
1035
			next_ts = ts;
1036
1037
1038
1039
1040
1041
		}
	}

	if (ent_cpu)
		*ent_cpu = next_cpu;

1042
1043
1044
	if (ent_ts)
		*ent_ts = next_ts;

1045
1046
1047
	return next;
}

1048
1049
/* Find the next real entry, without updating the iterator itself */
static struct trace_entry *
1050
find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts)
1051
{
1052
	return __find_next_entry(iter, ent_cpu, ent_ts);
1053
1054
1055
1056
1057
}

/* Find the next real entry, and increment the iterator to the next entry */
static void *find_next_entry_inc(struct trace_iterator *iter)
{
1058
	iter->ent = __find_next_entry(iter, &iter->cpu, &iter->ts);
1059

1060
	if (iter->ent)
1061
1062
		trace_iterator_increment(iter, iter->cpu);

1063
	return iter->ent ? iter : NULL;
1064
}
1065

Ingo Molnar's avatar
Ingo Molnar committed
1066
static void trace_consume(struct trace_iterator *iter)
1067
{
1068
1069
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();
1070
	ring_buffer_consume(iter->tr->buffer, iter->cpu, &iter->ts);
1071
	ftrace_enable_cpu();
1072
1073
}

Ingo Molnar's avatar
Ingo Molnar committed
1074
static void *s_next(struct seq_file *m, void *v, loff_t *pos)
1075
1076
1077
{
	struct trace_iterator *iter = m->private;
	int i = (int)*pos;
Ingo Molnar's avatar
Ingo Molnar committed
1078
	void *ent;
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103

	(*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;
1104
	int cpu;
1105
1106
1107

	mutex_lock(&trace_types_lock);

1108
1109
	if (!current_trace || current_trace != iter->trace) {
		mutex_unlock(&trace_types_lock);
1110
		return NULL;
1111
	}
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123

	atomic_inc(&trace_record_cmdline_disabled);

	/* let the tracer grab locks here if needed */
	if (current_trace->start)
		current_trace->start(iter);

	if (*pos != iter->pos) {
		iter->ent = NULL;
		iter->cpu = 0;
		iter->idx = -1;

1124
1125
		ftrace_disable_cpu();

1126
1127
		for_each_tracing_cpu(cpu) {
			ring_buffer_iter_reset(iter->buffer_iter[cpu]);
1128
		}
1129

1130
1131
		ftrace_enable_cpu();

1132
1133
1134
1135
		for (p = iter; p && l < *pos; p = s_next(m, p, &l))
			;

	} else {
1136
		l = *pos - 1;
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
		p = s_next(m, p, &l);
	}

	return p;
}

static void s_stop(struct seq_file *m, void *p)
{
	struct trace_iterator *iter = m->private;

	atomic_dec(&trace_record_cmdline_disabled);

	/* let the tracer release locks here if needed */
	if (current_trace && current_trace == iter->trace && iter->trace->stop)
		iter->trace->stop(iter);

	mutex_unlock(&trace_types_lock);
}

1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
#define KRETPROBE_MSG "[unknown/kretprobe'd]"

#ifdef CONFIG_KRETPROBES
static inline int kretprobed(unsigned long addr)
{
	return addr == (unsigned long)kretprobe_trampoline;
}
#else
static inline int kretprobed(unsigned long addr)
{
	return 0;
}
#endif /* CONFIG_KRETPROBES */

1170
static int
1171
seq_print_sym_short(struct trace_seq *s, const char *fmt, unsigned long address)
1172
1173
1174
1175
1176
1177
{
#ifdef CONFIG_KALLSYMS
	char str[KSYM_SYMBOL_LEN];

	kallsyms_lookup(address, NULL, NULL, NULL, str);

1178
	return trace_seq_printf(s, fmt, str);
1179
#endif
1180
	return 1;
1181
1182
}

1183
static int
1184
1185
seq_print_sym_offset(struct trace_seq *s, const char *fmt,
		     unsigned long address)
1186
1187
1188
1189
1190
{
#ifdef CONFIG_KALLSYMS
	char str[KSYM_SYMBOL_LEN];

	sprint_symbol(str, address);
1191
	return trace_seq_printf(s, fmt, str);
1192
#endif
1193
	return 1;
1194
1195
1196
1197
1198
1199
1200
1201
}

#ifndef CONFIG_64BIT
# define IP_FMT "%08lx"
#else
# define IP_FMT "%016lx"
#endif

Ingo Molnar's avatar
Ingo Molnar committed
1202
static int
1203
seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags)
1204
{
1205
1206
1207
1208
	int ret;

	if (!ip)
		return trace_seq_printf(s, "0");
1209
1210

	if (sym_flags & TRACE_ITER_SYM_OFFSET)
1211
		ret = seq_print_sym_offset(s, "%s", ip);
1212
	else
1213
1214
1215
1216
		ret = seq_print_sym_short(s, "%s", ip);

	if (!ret)
		return 0;
1217
1218

	if (sym_flags & TRACE_ITER_SYM_ADDR)
1219
1220
		ret = trace_seq_printf(s, " <" IP_FMT ">", ip);
	return ret;
1221
1222
}

Ingo Molnar's avatar
Ingo Molnar committed
1223
static void print_lat_help_header(struct seq_file *m)
1224
{
1225
1226
1227
1228
1229
1230
1231
1232
1233
	seq_puts(m, "#                  _------=> CPU#            \n");
	seq_puts(m, "#                 / _-----=> irqs-off        \n");
	seq_puts(m, "#                | / _----=> need-resched    \n");
	seq_puts(m, "#                || / _---=> hardirq/softirq \n");
	seq_puts(m, "#                ||| / _--=> preempt-depth   \n");
	seq_puts(m, "#                |||| /                      \n");
	seq_puts(m, "#                |||||     delay             \n");
	seq_puts(m, "#  cmd     pid   ||||| time  |   caller      \n");
	seq_puts(m, "#     \\   /      |||||   \\   |   /           \n");
1234
1235
}