trace.c 72.7 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
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
248
249
	NULL
};

Steven Rostedt's avatar
Steven Rostedt committed
250
251
252
253
254
255
256
257
258
/*
 * 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.
 */
259
260
static raw_spinlock_t ftrace_max_lock =
	(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
261
262
263
264
265
266

/*
 * 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
267
static void
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
__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
289
290
291
292
293
294
295
296
297
298
299
/**
 * 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.
 */
300
int
301
302
303
304
trace_seq_printf(struct trace_seq *s, const char *fmt, ...)
{
	int len = (PAGE_SIZE - 1) - s->len;
	va_list ap;
305
	int ret;
306
307
308
309
310

	if (!len)
		return 0;

	va_start(ap, fmt);
311
	ret = vsnprintf(s->buffer + s->len, len, fmt, ap);
312
313
	va_end(ap);

314
	/* If we can't write it all, don't bother writing anything */
315
	if (ret >= len)
316
317
318
		return 0;

	s->len += ret;
319
320
321
322

	return len;
}

Steven Rostedt's avatar
Steven Rostedt committed
323
324
325
326
327
328
329
330
331
332
/**
 * 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
333
static int
334
335
336
337
338
trace_seq_puts(struct trace_seq *s, const char *str)
{
	int len = strlen(str);

	if (len > ((PAGE_SIZE - 1) - s->len))
339
		return 0;
340
341
342
343
344
345
346

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

	return len;
}

Ingo Molnar's avatar
Ingo Molnar committed
347
static int
348
349
350
351
352
353
354
355
356
357
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
358
static int
Ingo Molnar's avatar
Ingo Molnar committed
359
360
361
362
363
364
365
366
367
368
369
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;
}

370
371
#define MAX_MEMHEX_BYTES	8
#define HEX_CHARS		(MAX_MEMHEX_BYTES*2 + 1)
372

Ingo Molnar's avatar
Ingo Molnar committed
373
static int
374
375
376
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
377
	unsigned char *data = mem;
378
379
380
381
382
383
384
	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
385
386
		hex[j++] = hex_asc_hi(data[i]);
		hex[j++] = hex_asc_lo(data[i]);
387
	}
Thomas Gleixner's avatar
Thomas Gleixner committed
388
	hex[j++] = ' ';
389
390
391
392

	return trace_seq_putmem(s, hex, j);
}

Ingo Molnar's avatar
Ingo Molnar committed
393
static void
394
395
396
trace_seq_reset(struct trace_seq *s)
{
	s->len = 0;
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
	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;
417
418
}

Ingo Molnar's avatar
Ingo Molnar committed
419
static void
420
421
422
423
424
425
426
427
428
429
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
430
431
432
433
434
435
436
437
438
/**
 * 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
439
void
440
441
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
442
	struct ring_buffer *buf = tr->buffer;
443

444
	WARN_ON_ONCE(!irqs_disabled());
445
	__raw_spin_lock(&ftrace_max_lock);
446
447
448
449

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

450
	ftrace_disable_cpu();
451
	ring_buffer_reset(tr->buffer);
452
	ftrace_enable_cpu();
453
454

	__update_max_tr(tr, tsk, cpu);
455
	__raw_spin_unlock(&ftrace_max_lock);
456
457
458
459
460
461
462
}

/**
 * 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
463
464
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
465
 */
Ingo Molnar's avatar
Ingo Molnar committed
466
void
467
468
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
469
	int ret;
470

471
	WARN_ON_ONCE(!irqs_disabled());
472
	__raw_spin_lock(&ftrace_max_lock);
473

474
475
	ftrace_disable_cpu();

476
477
478
	ring_buffer_reset(max_tr.buffer);
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

479
480
	ftrace_enable_cpu();

481
	WARN_ON_ONCE(ret);
482
483

	__update_max_tr(tr, tsk, cpu);
484
	__raw_spin_unlock(&ftrace_max_lock);
485
486
}

Steven Rostedt's avatar
Steven Rostedt committed
487
488
489
490
491
492
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
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
515
516
517
518
519
520
521
522
523
524
525
526
527
#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.
		 */
528
		for_each_tracing_cpu(i) {
529
			tracing_reset(tr, i);
Steven Rostedt's avatar
Steven Rostedt committed
530
531
532
533
534
535
536
537
538
539
540
541
542
		}
		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
543
		/* Only reset on passing, to avoid touching corrupted buffers */
544
		for_each_tracing_cpu(i) {
545
			tracing_reset(tr, i);
Steven Rostedt's avatar
Steven Rostedt committed
546
		}
Steven Rostedt's avatar
Steven Rostedt committed
547
548
549
550
		printk(KERN_CONT "PASSED\n");
	}
#endif

551
552
553
554
555
	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
556

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
 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);
}

591
void tracing_reset(struct trace_array *tr, int cpu)
592
{
593
	ftrace_disable_cpu();
594
	ring_buffer_reset_cpu(tr->buffer, cpu);
595
	ftrace_enable_cpu();
596
597
598
599
600
601
602
603
}

#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);
604
605
606

/* temporary disable recording */
atomic_t trace_record_cmdline_disabled __read_mostly;
607
608
609
610
611
612
613
614

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
615
void trace_stop_cmdline_recording(void);
616

Ingo Molnar's avatar
Ingo Molnar committed
617
static void trace_save_cmdline(struct task_struct *tsk)
618
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
{
	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
652
static char *trace_find_cmdline(int pid)
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
{
	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
673
void tracing_record_cmdline(struct task_struct *tsk)
674
675
676
677
678
679
680
{
	if (atomic_read(&trace_record_cmdline_disabled))
		return;

	trace_save_cmdline(tsk);
}

681
void
682
683
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
684
685
686
{
	struct task_struct *tsk = current;

687
688
689
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
	entry->flags =
690
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
691
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
692
693
694
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
695
696
697
698
699
		((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
700
void
701
trace_function(struct trace_array *tr, struct trace_array_cpu *data,
702
703
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
704
{
705
	struct ring_buffer_event *event;
706
	struct ftrace_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
707
	unsigned long irq_flags;
708

709
710
711
712
	/* If we are reading the ring buffer, don't trace */
	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

713
714
715
716
717
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
718
	tracing_generic_entry_update(&entry->ent, flags, pc);
719
720
721
	entry->ent.type			= TRACE_FN;
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
722
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
723
724
}

Ingo Molnar's avatar
Ingo Molnar committed
725
void
Ingo Molnar's avatar
Ingo Molnar committed
726
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
727
728
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
Ingo Molnar's avatar
Ingo Molnar committed
729
730
{
	if (likely(!atomic_read(&data->disabled)))
731
		trace_function(tr, data, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
732
733
}

734
735
736
737
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
738
{
739
#ifdef CONFIG_STACKTRACE
740
	struct ring_buffer_event *event;
741
	struct stack_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
742
	struct stack_trace trace;
743
	unsigned long irq_flags;
Ingo Molnar's avatar
Ingo Molnar committed
744
745
746
747

	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

748
749
750
751
752
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
753
	tracing_generic_entry_update(&entry->ent, flags, pc);
754
	entry->ent.type		= TRACE_STACK;
Ingo Molnar's avatar
Ingo Molnar committed
755

756
	memset(&entry->caller, 0, sizeof(entry->caller));
Ingo Molnar's avatar
Ingo Molnar committed
757
758
759
760

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
761
	trace.entries		= entry->caller;
Ingo Molnar's avatar
Ingo Molnar committed
762
763

	save_stack_trace(&trace);
764
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
765
#endif
Ingo Molnar's avatar
Ingo Molnar committed
766
767
}

768
769
770
771
772
773
774
775
776
777
778
779
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
780
{
781
	struct ring_buffer_event *event;
Ingo Molnar's avatar
Ingo Molnar committed
782
783
	struct trace_array_cpu *data = __data;
	struct trace_array *tr = __tr;
784
	struct special_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
785
786
	unsigned long irq_flags;

787
788
789
790
791
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
792
	tracing_generic_entry_update(&entry->ent, 0, pc);
793
794
795
796
	entry->ent.type			= TRACE_SPECIAL;
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
797
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
798
	ftrace_trace_stack(tr, data, irq_flags, 4, pc);
Ingo Molnar's avatar
Ingo Molnar committed
799
800
801
802

	trace_wake_up();
}

803
804
805
806
807
808
809
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
810
void
811
812
tracing_sched_switch_trace(struct trace_array *tr,
			   struct trace_array_cpu *data,
Ingo Molnar's avatar
Ingo Molnar committed
813
814
			   struct task_struct *prev,
			   struct task_struct *next,
815
			   unsigned long flags, int pc)
816
{
817
	struct ring_buffer_event *event;
818
	struct ctx_switch_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
819
	unsigned long irq_flags;
820

821
822
823
824
825
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					   &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
826
	tracing_generic_entry_update(&entry->ent, flags, pc);
827
828
829
830
831
832
833
834
	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);
835
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
836
	ftrace_trace_stack(tr, data, flags, 5, pc);
837
838
}

839
840
841
void
tracing_sched_wakeup_trace(struct trace_array *tr,
			   struct trace_array_cpu *data,
Ingo Molnar's avatar
Ingo Molnar committed
842
843
			   struct task_struct *wakee,
			   struct task_struct *curr,
844
			   unsigned long flags, int pc)
845
{
846
	struct ring_buffer_event *event;
847
	struct ctx_switch_entry *entry;
848
849
	unsigned long irq_flags;

850
851
852
853
854
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					   &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
855
	tracing_generic_entry_update(&entry->ent, flags, pc);
856
857
858
859
860
861
862
863
	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);
864
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
865
	ftrace_trace_stack(tr, data, flags, 6, pc);
Ingo Molnar's avatar
Ingo Molnar committed
866
867

	trace_wake_up();
868
869
}

870
871
872
873
874
875
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;
876
	int pc;
877

878
	if (tracing_disabled || !tr->ctrl)
879
880
		return;

881
882
	pc = preempt_count();
	preempt_disable_notrace();
883
884
885
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

886
	if (likely(!atomic_read(&data->disabled)))
887
		ftrace_trace_special(tr, data, arg1, arg2, arg3, pc);
888

889
	preempt_enable_notrace();
890
891
}

892
#ifdef CONFIG_FUNCTION_TRACER
Ingo Molnar's avatar
Ingo Molnar committed
893
static void
Ingo Molnar's avatar
Ingo Molnar committed
894
895
896
897
898
899
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;
900
901
	int cpu, resched;
	int pc;
Ingo Molnar's avatar
Ingo Molnar committed
902

903
	if (unlikely(!ftrace_function_enabled))
Ingo Molnar's avatar
Ingo Molnar committed
904
905
		return;

906
	pc = preempt_count();
907
	resched = ftrace_preempt_disable();
908
	local_save_flags(flags);
Ingo Molnar's avatar
Ingo Molnar committed
909
910
911
912
913
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);

	if (likely(disabled == 1))
914
		trace_function(tr, data, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
915
916

	atomic_dec(&data->disabled);
917
	ftrace_preempt_enable(resched);
Ingo Molnar's avatar
Ingo Molnar committed
918
919
920
921
922
923
924
}

static struct ftrace_ops trace_ops __read_mostly =
{
	.func = function_trace_call,
};

Ingo Molnar's avatar
Ingo Molnar committed
925
void tracing_start_function_trace(void)
Ingo Molnar's avatar
Ingo Molnar committed
926
{
927
	ftrace_function_enabled = 0;
Ingo Molnar's avatar
Ingo Molnar committed
928
	register_ftrace_function(&trace_ops);
929
930
	if (tracer_enabled)
		ftrace_function_enabled = 1;
Ingo Molnar's avatar
Ingo Molnar committed
931
932
}

Ingo Molnar's avatar
Ingo Molnar committed
933
void tracing_stop_function_trace(void)
Ingo Molnar's avatar
Ingo Molnar committed
934
{
935
	ftrace_function_enabled = 0;
Ingo Molnar's avatar
Ingo Molnar committed
936
937
938
939
	unregister_ftrace_function(&trace_ops);
}
#endif

940
941
942
943
enum trace_file_type {
	TRACE_FILE_LAT_FMT	= 1,
};

944
945
static void trace_iterator_increment(struct trace_iterator *iter, int cpu)
{
946
947
948
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();

949
	iter->idx++;
950
951
952
953
	if (iter->buffer_iter[iter->cpu])
		ring_buffer_read(iter->buffer_iter[iter->cpu], NULL);

	ftrace_enable_cpu();
954
955
}

Ingo Molnar's avatar
Ingo Molnar committed
956
static struct trace_entry *
957
peek_next_entry(struct trace_iterator *iter, int cpu, u64 *ts)
958
{
959
960
	struct ring_buffer_event *event;
	struct ring_buffer_iter *buf_iter = iter->buffer_iter[cpu];
961

962
963
964
965
966
967
968
969
970
971
	/* 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();

972
	return event ? ring_buffer_event_data(event) : NULL;
973
}
974

975
static struct trace_entry *
976
__find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts)
977
{
978
	struct ring_buffer *buffer = iter->tr->buffer;
979
	struct trace_entry *ent, *next = NULL;
980
	u64 next_ts = 0, ts;
981
982
983
	int next_cpu = -1;
	int cpu;

984
	for_each_tracing_cpu(cpu) {
985

986
987
		if (ring_buffer_empty_cpu(buffer, cpu))
			continue;
988

989
		ent = peek_next_entry(iter, cpu, &ts);
990

Ingo Molnar's avatar
Ingo Molnar committed
991
992
993
		/*
		 * Pick the entry with the smallest timestamp:
		 */
994
		if (ent && (!next || ts < next_ts)) {
995
996
			next = ent;
			next_cpu = cpu;
997
			next_ts = ts;
998
999
1000
1001
1002
1003
		}
	}

	if (ent_cpu)
		*ent_cpu = next_cpu;

1004
1005
1006
	if (ent_ts)
		*ent_ts = next_ts;

1007
1008
1009
	return next;
}

1010
1011
/* Find the next real entry, without updating the iterator itself */
static struct trace_entry *
1012
find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts)
1013
{
1014
	return __find_next_entry(iter, ent_cpu, ent_ts);
1015
1016
1017
1018
1019
}

/* Find the next real entry, and increment the iterator to the next entry */
static void *find_next_entry_inc(struct trace_iterator *iter)
{
1020
	iter->ent = __find_next_entry(iter, &iter->cpu, &iter->ts);
1021

1022
	if (iter->ent)
1023
1024
		trace_iterator_increment(iter, iter->cpu);

1025
	return iter->ent ? iter : NULL;
1026
}
1027

Ingo Molnar's avatar
Ingo Molnar committed
1028
static void trace_consume(struct trace_iterator *iter)
1029
{
1030
1031
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();
1032
	ring_buffer_consume(iter->tr->buffer, iter->cpu, &iter->ts);
1033
	ftrace_enable_cpu();
1034
1035
}

Ingo Molnar's avatar
Ingo Molnar committed
1036
static void *s_next(struct seq_file *m, void *v, loff_t *pos)
1037
1038
1039
{
	struct trace_iterator *iter = m->private;
	int i = (int)*pos;
Ingo Molnar's avatar
Ingo Molnar committed
1040
	void *ent;
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065

	(*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;
1066
	int cpu;
1067
1068
1069

	mutex_lock(&trace_types_lock);

1070
1071
	if (!current_trace || current_trace != iter->trace) {
		mutex_unlock(&trace_types_lock);
1072
		return NULL;
1073
	}
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085

	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;

1086
1087
		ftrace_disable_cpu();

1088
1089
		for_each_tracing_cpu(cpu) {
			ring_buffer_iter_reset(iter->buffer_iter[cpu]);
1090
		}
1091

1092
1093
		ftrace_enable_cpu();

1094
1095
1096
1097
		for (p = iter; p && l < *pos; p = s_next(m, p, &l))
			;

	} else {
1098
		l = *pos - 1;
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
		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);
}

1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
#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 */

1132
static int
1133
seq_print_sym_short(struct trace_seq *s, const char *fmt, unsigned long address)
1134
1135
1136
1137
1138
1139
{
#ifdef CONFIG_KALLSYMS
	char str[KSYM_SYMBOL_LEN];

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

1140
	return trace_seq_printf(s, fmt, str);
1141
#endif
1142
	return 1;
1143
1144
}

1145
static int
1146
1147
seq_print_sym_offset(struct trace_seq *s, const char *fmt,
		     unsigned long address)
1148
1149
1150
1151
1152
{
#ifdef CONFIG_KALLSYMS
	char str[KSYM_SYMBOL_LEN];

	sprint_symbol(str, address);
1153
	return trace_seq_printf(s, fmt, str);
1154
#endif
1155
	return 1;
1156
1157
1158
1159
1160
1161
1162
1163
}

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

Ingo Molnar's avatar
Ingo Molnar committed
1164
static int
1165
seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags)
1166
{
1167
1168
1169
1170
	int ret;

	if (!ip)
		return trace_seq_printf(s, "0");
1171
1172

	if (sym_flags & TRACE_ITER_SYM_OFFSET)
1173
		ret = seq_print_sym_offset(s, "%s", ip);
1174
	else
1175
1176
1177
1178
		ret = seq_print_sym_short(s, "%s", ip);

	if (!ret)
		return 0;
1179
1180

	if (sym_flags & TRACE_ITER_SYM_ADDR)
1181
1182
		ret = trace_seq_printf(s, " <" IP_FMT ">", ip);
	return ret;
1183
1184
}

Ingo Molnar's avatar
Ingo Molnar committed
1185
static void print_lat_help_header(struct seq_file *m)
1186
{
1187
1188
1189
1190
1191
1192
1193
1194
1195
	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");
1196
1197
}

Ingo Molnar's avatar
Ingo Molnar committed
1198
static void print_func_help_header(struct seq_file *m)
1199
{
1200
1201
	seq_puts(m, "#           TASK-PID    CPU#    TIMESTAMP  FUNCTION\n");
	seq_puts(m, "#              | |       |          |         |\n");
1202
1203
1204
}


Ingo Molnar's avatar
Ingo Molnar committed
1205
static void
1206
1207
1208
1209
1210
1211
print_trace_header(struct seq_file *m, struct trace_iterator *iter)
{
	unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK);
	struct trace_array *tr = iter->tr;
	struct trace_array_cpu *data = tr->data[tr->cpu];
	struct tracer *type = current_trace;
1212
1213
	unsigned long total;
	unsigned long entries;
1214
1215
1216
1217
1218
	const char *name = "preemption";

	if (type)
		name = type->name;