trace.c 70.3 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>
Ingo Molnar's avatar
Ingo Molnar committed
37

38
39
#include "trace.h"

40
41
#define TRACE_BUFFER_FLAGS	(RB_FL_OVERWRITE)

42
43
44
unsigned long __read_mostly	tracing_max_latency = (cycle_t)ULONG_MAX;
unsigned long __read_mostly	tracing_thresh;

45
46
47
48
49
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
50
51
static int tracing_disabled = 1;

52
long
53
54
55
56
57
58
59
ns2usecs(cycle_t nsec)
{
	nsec += 500;
	do_div(nsec, 1000);
	return nsec;
}

Ingo Molnar's avatar
Ingo Molnar committed
60
cycle_t ftrace_now(int cpu)
61
{
62
63
64
	u64 ts = ring_buffer_time_stamp(cpu);
	ring_buffer_normalize_time_stamp(cpu, &ts);
	return ts;
65
66
}

Steven Rostedt's avatar
Steven Rostedt committed
67
68
69
70
71
72
73
74
75
76
77
78
/*
 * 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.
 */
79
80
81
82
static struct trace_array	global_trace;

static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu);

Steven Rostedt's avatar
Steven Rostedt committed
83
84
85
86
87
88
89
90
91
92
/*
 * 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.
 */
93
94
95
96
static struct trace_array	max_tr;

static DEFINE_PER_CPU(struct trace_array_cpu, max_data);

Steven Rostedt's avatar
Steven Rostedt committed
97
/* tracer_enabled is used to toggle activation of a tracer */
98
static int			tracer_enabled = 1;
Steven Rostedt's avatar
Steven Rostedt committed
99

100
101
102
/* function tracing enabled */
int				ftrace_function_enabled;

Steven Rostedt's avatar
Steven Rostedt committed
103
/*
104
105
106
 * 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.
107
108
109
110
111
 *
 * 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
112
 */
113
#define TRACE_BUF_SIZE_DEFAULT	1441792UL /* 16384 * 88 (sizeof(entry)) */
114

115
static unsigned long		trace_buf_size = TRACE_BUF_SIZE_DEFAULT;
116

Steven Rostedt's avatar
Steven Rostedt committed
117
/* trace_types holds a link list of available tracers. */
118
static struct tracer		*trace_types __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
119
120

/* current_trace points to the tracer that is currently active */
121
static struct tracer		*current_trace __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
122
123
124
125
126
127

/*
 * 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.
 */
128
129
static int			max_tracer_type_len;

Steven Rostedt's avatar
Steven Rostedt committed
130
131
132
133
134
135
/*
 * 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.
 */
136
static DEFINE_MUTEX(trace_types_lock);
Steven Rostedt's avatar
Steven Rostedt committed
137
138

/* trace_wait is a waitqueue for tasks blocked on trace_poll */
139
140
static DECLARE_WAIT_QUEUE_HEAD(trace_wait);

Steven Rostedt's avatar
Steven Rostedt committed
141
/* trace_flags holds iter_ctrl options */
142
143
unsigned long trace_flags = TRACE_ITER_PRINT_PARENT;

Steven Rostedt's avatar
Steven Rostedt committed
144
145
146
147
148
149
/**
 * 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.
 */
150
151
void trace_wake_up(void)
{
Ingo Molnar's avatar
Ingo Molnar committed
152
153
154
155
156
	/*
	 * The runqueue_is_locked() can fail, but this is the best we
	 * have for now:
	 */
	if (!(trace_flags & TRACE_ITER_BLOCK) && !runqueue_is_locked())
157
158
		wake_up(&trace_wait);
}
159

160
static int __init set_buf_size(char *str)
161
{
162
	unsigned long buf_size;
163
164
	int ret;

165
166
	if (!str)
		return 0;
167
	ret = strict_strtoul(str, 0, &buf_size);
168
	/* nr_entries can not be zero */
169
	if (ret < 0 || buf_size == 0)
170
		return 0;
171
	trace_buf_size = buf_size;
172
173
	return 1;
}
174
__setup("trace_buf_size=", set_buf_size);
175

Steven Rostedt's avatar
Steven Rostedt committed
176
177
178
179
180
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

Steven Rostedt's avatar
Steven Rostedt committed
181
182
183
184
/*
 * TRACE_ITER_SYM_MASK masks the options in trace_flags that
 * control the output of kernel symbols.
 */
185
186
187
#define TRACE_ITER_SYM_MASK \
	(TRACE_ITER_PRINT_PARENT|TRACE_ITER_SYM_OFFSET|TRACE_ITER_SYM_ADDR)

Steven Rostedt's avatar
Steven Rostedt committed
188
/* These must match the bit postions in trace_iterator_flags */
189
190
191
192
193
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
Ingo Molnar's avatar
Ingo Molnar committed
194
	"raw",
195
	"hex",
Ingo Molnar's avatar
Ingo Molnar committed
196
	"bin",
197
	"block",
Ingo Molnar's avatar
Ingo Molnar committed
198
	"stacktrace",
Ingo Molnar's avatar
Ingo Molnar committed
199
	"sched-tree",
200
	"ftrace_printk",
201
202
203
	NULL
};

Steven Rostedt's avatar
Steven Rostedt committed
204
205
206
207
208
209
210
211
212
/*
 * 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.
 */
213
214
static raw_spinlock_t ftrace_max_lock =
	(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
215
216
217
218
219
220

/*
 * 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
221
static void
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
__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
243
244
245
246
247
248
249
250
251
252
253
/**
 * 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.
 */
254
int
255
256
257
258
trace_seq_printf(struct trace_seq *s, const char *fmt, ...)
{
	int len = (PAGE_SIZE - 1) - s->len;
	va_list ap;
259
	int ret;
260
261
262
263
264

	if (!len)
		return 0;

	va_start(ap, fmt);
265
	ret = vsnprintf(s->buffer + s->len, len, fmt, ap);
266
267
	va_end(ap);

268
	/* If we can't write it all, don't bother writing anything */
269
	if (ret >= len)
270
271
272
		return 0;

	s->len += ret;
273
274
275
276

	return len;
}

Steven Rostedt's avatar
Steven Rostedt committed
277
278
279
280
281
282
283
284
285
286
/**
 * 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
287
static int
288
289
290
291
292
trace_seq_puts(struct trace_seq *s, const char *str)
{
	int len = strlen(str);

	if (len > ((PAGE_SIZE - 1) - s->len))
293
		return 0;
294
295
296
297
298
299
300

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

	return len;
}

Ingo Molnar's avatar
Ingo Molnar committed
301
static int
302
303
304
305
306
307
308
309
310
311
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
312
static int
Ingo Molnar's avatar
Ingo Molnar committed
313
314
315
316
317
318
319
320
321
322
323
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;
}

324
#define HEX_CHARS 17
Thomas Gleixner's avatar
Thomas Gleixner committed
325
static const char hex2asc[] = "0123456789abcdef";
326

Ingo Molnar's avatar
Ingo Molnar committed
327
static int
328
329
330
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
331
	unsigned char *data = mem;
332
333
334
335
336
337
338
339
340
341
342
343
	unsigned char byte;
	int i, j;

	BUG_ON(len >= HEX_CHARS);

#ifdef __BIG_ENDIAN
	for (i = 0, j = 0; i < len; i++) {
#else
	for (i = len-1, j = 0; i >= 0; i--) {
#endif
		byte = data[i];

Thomas Gleixner's avatar
Thomas Gleixner committed
344
345
		hex[j++] = hex2asc[byte & 0x0f];
		hex[j++] = hex2asc[byte >> 4];
346
	}
Thomas Gleixner's avatar
Thomas Gleixner committed
347
	hex[j++] = ' ';
348
349
350
351

	return trace_seq_putmem(s, hex, j);
}

Ingo Molnar's avatar
Ingo Molnar committed
352
static void
353
354
355
trace_seq_reset(struct trace_seq *s)
{
	s->len = 0;
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
	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;
376
377
}

Ingo Molnar's avatar
Ingo Molnar committed
378
static void
379
380
381
382
383
384
385
386
387
388
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
389
390
391
392
393
394
395
396
397
/**
 * 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
398
void
399
400
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
401
	struct ring_buffer *buf = tr->buffer;
402

403
	WARN_ON_ONCE(!irqs_disabled());
404
	__raw_spin_lock(&ftrace_max_lock);
405
406
407
408
409

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

	ring_buffer_reset(tr->buffer);
410
411

	__update_max_tr(tr, tsk, cpu);
412
	__raw_spin_unlock(&ftrace_max_lock);
413
414
415
416
417
418
419
}

/**
 * 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
420
421
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
422
 */
Ingo Molnar's avatar
Ingo Molnar committed
423
void
424
425
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
426
	int ret;
427

428
	WARN_ON_ONCE(!irqs_disabled());
429
	__raw_spin_lock(&ftrace_max_lock);
430

431
432
433
434
	ring_buffer_reset(max_tr.buffer);
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

	WARN_ON_ONCE(ret);
435
436

	__update_max_tr(tr, tsk, cpu);
437
	__raw_spin_unlock(&ftrace_max_lock);
438
439
}

Steven Rostedt's avatar
Steven Rostedt committed
440
441
442
443
444
445
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
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
468
469
470
471
472
473
474
475
476
477
478
479
480
#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.
		 */
481
		for_each_tracing_cpu(i) {
482
			tracing_reset(tr, i);
Steven Rostedt's avatar
Steven Rostedt committed
483
484
485
486
487
488
489
490
491
492
493
494
495
		}
		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
496
		/* Only reset on passing, to avoid touching corrupted buffers */
497
		for_each_tracing_cpu(i) {
498
			tracing_reset(tr, i);
Steven Rostedt's avatar
Steven Rostedt 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
511
512
513
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
 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);
}

544
void tracing_reset(struct trace_array *tr, int cpu)
545
{
546
	ring_buffer_reset_cpu(tr->buffer, cpu);
547
548
549
550
551
552
553
554
}

#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);
555
556
557

/* temporary disable recording */
atomic_t trace_record_cmdline_disabled __read_mostly;
558
559
560
561
562
563
564
565

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
566
void trace_stop_cmdline_recording(void);
567

Ingo Molnar's avatar
Ingo Molnar committed
568
static void trace_save_cmdline(struct task_struct *tsk)
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
{
	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
603
static char *trace_find_cmdline(int pid)
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
{
	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
624
void tracing_record_cmdline(struct task_struct *tsk)
625
626
627
628
629
630
631
{
	if (atomic_read(&trace_record_cmdline_disabled))
		return;

	trace_save_cmdline(tsk);
}

632
void
Ingo Molnar's avatar
Ingo Molnar committed
633
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags)
634
635
636
637
638
639
{
	struct task_struct *tsk = current;
	unsigned long pc;

	pc = preempt_count();

640
641
642
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
	entry->flags =
643
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
644
645
646
647
648
		((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
649
void
650
651
trace_function(struct trace_array *tr, struct trace_array_cpu *data,
	       unsigned long ip, unsigned long parent_ip, unsigned long flags)
652
{
653
	struct ring_buffer_event *event;
654
	struct ftrace_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
655
	unsigned long irq_flags;
656

657
658
659
660
661
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
662
663
664
665
	tracing_generic_entry_update(&entry->ent, flags);
	entry->ent.type			= TRACE_FN;
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
666
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
667
668
}

Ingo Molnar's avatar
Ingo Molnar committed
669
void
Ingo Molnar's avatar
Ingo Molnar committed
670
671
672
673
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
       unsigned long ip, unsigned long parent_ip, unsigned long flags)
{
	if (likely(!atomic_read(&data->disabled)))
674
		trace_function(tr, data, ip, parent_ip, flags);
Ingo Molnar's avatar
Ingo Molnar committed
675
676
}

Ingo Molnar's avatar
Ingo Molnar committed
677
678
679
680
681
void __trace_stack(struct trace_array *tr,
		   struct trace_array_cpu *data,
		   unsigned long flags,
		   int skip)
{
682
	struct ring_buffer_event *event;
683
	struct stack_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
684
	struct stack_trace trace;
685
	unsigned long irq_flags;
Ingo Molnar's avatar
Ingo Molnar committed
686
687
688
689

	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

690
691
692
693
694
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
695
696
	tracing_generic_entry_update(&entry->ent, flags);
	entry->ent.type		= TRACE_STACK;
Ingo Molnar's avatar
Ingo Molnar committed
697

698
	memset(&entry->caller, 0, sizeof(entry->caller));
Ingo Molnar's avatar
Ingo Molnar committed
699
700
701
702

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
703
	trace.entries		= entry->caller;
Ingo Molnar's avatar
Ingo Molnar committed
704
705

	save_stack_trace(&trace);
706
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
Ingo Molnar's avatar
Ingo Molnar committed
707
708
}

Ingo Molnar's avatar
Ingo Molnar committed
709
710
711
712
void
__trace_special(void *__tr, void *__data,
		unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
713
	struct ring_buffer_event *event;
Ingo Molnar's avatar
Ingo Molnar committed
714
715
	struct trace_array_cpu *data = __data;
	struct trace_array *tr = __tr;
716
	struct special_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
717
718
	unsigned long irq_flags;

719
720
721
722
723
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					 &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
724
725
726
727
728
	tracing_generic_entry_update(&entry->ent, 0);
	entry->ent.type			= TRACE_SPECIAL;
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
729
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
Ingo Molnar's avatar
Ingo Molnar committed
730
731
732
733
734
	__trace_stack(tr, data, irq_flags, 4);

	trace_wake_up();
}

Ingo Molnar's avatar
Ingo Molnar committed
735
void
736
737
tracing_sched_switch_trace(struct trace_array *tr,
			   struct trace_array_cpu *data,
Ingo Molnar's avatar
Ingo Molnar committed
738
739
			   struct task_struct *prev,
			   struct task_struct *next,
740
741
			   unsigned long flags)
{
742
	struct ring_buffer_event *event;
743
	struct ctx_switch_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
744
	unsigned long irq_flags;
745

746
747
748
749
750
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					   &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
751
752
753
754
755
756
757
758
759
	tracing_generic_entry_update(&entry->ent, flags);
	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);
760
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
Ingo Molnar's avatar
Ingo Molnar committed
761
	__trace_stack(tr, data, flags, 5);
762
763
}

764
765
766
void
tracing_sched_wakeup_trace(struct trace_array *tr,
			   struct trace_array_cpu *data,
Ingo Molnar's avatar
Ingo Molnar committed
767
768
			   struct task_struct *wakee,
			   struct task_struct *curr,
769
770
			   unsigned long flags)
{
771
	struct ring_buffer_event *event;
772
	struct ctx_switch_entry *entry;
773
774
	unsigned long irq_flags;

775
776
777
778
779
	event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
					   &irq_flags);
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
780
781
782
783
784
785
786
787
788
	tracing_generic_entry_update(&entry->ent, flags);
	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);
789
	ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
Ingo Molnar's avatar
Ingo Molnar committed
790
	__trace_stack(tr, data, flags, 6);
Ingo Molnar's avatar
Ingo Molnar committed
791
792

	trace_wake_up();
793
794
}

795
796
797
798
799
800
801
802
803
void
ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	unsigned long flags;
	long disabled;
	int cpu;

804
	if (tracing_disabled || !tr->ctrl)
805
806
807
808
809
810
811
812
813
814
815
816
817
818
		return;

	local_irq_save(flags);
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);

	if (likely(disabled == 1))
		__trace_special(tr, data, arg1, arg2, arg3);

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

Ingo Molnar's avatar
Ingo Molnar committed
819
#ifdef CONFIG_FTRACE
Ingo Molnar's avatar
Ingo Molnar committed
820
static void
Ingo Molnar's avatar
Ingo Molnar committed
821
822
823
824
825
826
827
828
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;

829
	if (unlikely(!ftrace_function_enabled))
Ingo Molnar's avatar
Ingo Molnar committed
830
831
		return;

832
833
834
	if (skip_trace(ip))
		return;

Ingo Molnar's avatar
Ingo Molnar committed
835
836
837
838
839
840
	local_irq_save(flags);
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);

	if (likely(disabled == 1))
841
		trace_function(tr, data, ip, parent_ip, flags);
Ingo Molnar's avatar
Ingo Molnar committed
842
843
844
845
846
847
848
849
850
851

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

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

Ingo Molnar's avatar
Ingo Molnar committed
852
void tracing_start_function_trace(void)
Ingo Molnar's avatar
Ingo Molnar committed
853
{
854
	ftrace_function_enabled = 0;
Ingo Molnar's avatar
Ingo Molnar committed
855
	register_ftrace_function(&trace_ops);
856
857
	if (tracer_enabled)
		ftrace_function_enabled = 1;
Ingo Molnar's avatar
Ingo Molnar committed
858
859
}

Ingo Molnar's avatar
Ingo Molnar committed
860
void tracing_stop_function_trace(void)
Ingo Molnar's avatar
Ingo Molnar committed
861
{
862
	ftrace_function_enabled = 0;
Ingo Molnar's avatar
Ingo Molnar committed
863
864
865
866
	unregister_ftrace_function(&trace_ops);
}
#endif

867
868
869
870
enum trace_file_type {
	TRACE_FILE_LAT_FMT	= 1,
};

871
872
873
static void trace_iterator_increment(struct trace_iterator *iter, int cpu)
{
	iter->idx++;
874
	ring_buffer_read(iter->buffer_iter[iter->cpu], NULL);
875
876
}

Ingo Molnar's avatar
Ingo Molnar committed
877
static struct trace_entry *
878
peek_next_entry(struct trace_iterator *iter, int cpu, u64 *ts)
879
{
880
881
	struct ring_buffer_event *event;
	struct ring_buffer_iter *buf_iter = iter->buffer_iter[cpu];
882

883
884
	event = ring_buffer_iter_peek(buf_iter, ts);
	return event ? ring_buffer_event_data(event) : NULL;
885
886
}
static struct trace_entry *
887
__find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts)
888
{
889
	struct ring_buffer *buffer = iter->tr->buffer;
890
	struct trace_entry *ent, *next = NULL;
891
	u64 next_ts = 0, ts;
892
893
894
	int next_cpu = -1;
	int cpu;

895
	for_each_tracing_cpu(cpu) {
896

897
898
		if (ring_buffer_empty_cpu(buffer, cpu))
			continue;
899

900
		ent = peek_next_entry(iter, cpu, &ts);
901

Ingo Molnar's avatar
Ingo Molnar committed
902
903
904
		/*
		 * Pick the entry with the smallest timestamp:
		 */
905
		if (ent && (!next || ts < next_ts)) {
906
907
			next = ent;
			next_cpu = cpu;
908
			next_ts = ts;
909
910
911
912
913
914
		}
	}

	if (ent_cpu)
		*ent_cpu = next_cpu;

915
916
917
	if (ent_ts)
		*ent_ts = next_ts;

918
919
920
	return next;
}

921
922
/* Find the next real entry, without updating the iterator itself */
static struct trace_entry *
923
find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts)
924
{
925
	return __find_next_entry(iter, ent_cpu, ent_ts);
926
927
928
929
930
}

/* Find the next real entry, and increment the iterator to the next entry */
static void *find_next_entry_inc(struct trace_iterator *iter)
{
931
	iter->ent = __find_next_entry(iter, &iter->cpu, &iter->ts);
932

933
	if (iter->ent)
934
935
		trace_iterator_increment(iter, iter->cpu);

936
	return iter->ent ? iter : NULL;
937
}
938

Ingo Molnar's avatar
Ingo Molnar committed
939
static void trace_consume(struct trace_iterator *iter)
940
{
941
	ring_buffer_consume(iter->tr->buffer, iter->cpu, &iter->ts);
942
943
}

Ingo Molnar's avatar
Ingo Molnar committed
944
static void *s_next(struct seq_file *m, void *v, loff_t *pos)
945
946
947
{
	struct trace_iterator *iter = m->private;
	int i = (int)*pos;
Ingo Molnar's avatar
Ingo Molnar committed
948
	void *ent;
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973

	(*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;
974
	int cpu;
975
976
977

	mutex_lock(&trace_types_lock);

978
979
	if (!current_trace || current_trace != iter->trace) {
		mutex_unlock(&trace_types_lock);
980
		return NULL;
981
	}
982
983
984
985
986
987
988
989
990
991
992
993

	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;

994
995
		for_each_tracing_cpu(cpu) {
			ring_buffer_iter_reset(iter->buffer_iter[cpu]);
996
		}
997
998
999
1000
1001

		for (p = iter; p && l < *pos; p = s_next(m, p, &l))
			;

	} else {
1002
		l = *pos - 1;
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
		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);
}

1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
#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 */

1036
static int
1037
seq_print_sym_short(struct trace_seq *s, const char *fmt, unsigned long address)
1038
1039
1040
1041
1042
1043
{
#ifdef CONFIG_KALLSYMS
	char str[KSYM_SYMBOL_LEN];

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

1044
	return trace_seq_printf(s, fmt, str);
1045
#endif
1046
	return 1;
1047
1048
}

1049
static int
1050
1051
seq_print_sym_offset(struct trace_seq *s, const char *fmt,
		     unsigned long address)
1052
1053
1054
1055
1056
{
#ifdef CONFIG_KALLSYMS
	char str[KSYM_SYMBOL_LEN];

	sprint_symbol(str, address);
1057
	return trace_seq_printf(s, fmt, str);
1058
#endif
1059
	return 1;
1060
1061
1062
1063
1064
1065
1066
1067
}

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

Ingo Molnar's avatar
Ingo Molnar committed
1068
static int
1069
seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags)
1070
{
1071
1072
1073
1074
	int ret;

	if (!ip)
		return trace_seq_printf(s, "0");
1075
1076

	if (sym_flags & TRACE_ITER_SYM_OFFSET)
1077
		ret = seq_print_sym_offset(s, "%s", ip);
1078
	else
1079
1080
1081
1082
		ret = seq_print_sym_short(s, "%s", ip);

	if (!ret)
		return 0;
1083
1084

	if (sym_flags & TRACE_ITER_SYM_ADDR)
1085
1086
		ret = trace_seq_printf(s, " <" IP_FMT ">", ip);
	return ret;
1087
1088
}

Ingo Molnar's avatar
Ingo Molnar committed
1089
static void print_lat_help_header(struct seq_file *m)
1090
{
1091
1092
1093
1094
1095
1096
1097
1098
1099
	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");
1100
1101
}

Ingo Molnar's avatar
Ingo Molnar committed
1102
static void print_func_help_header(struct seq_file *m)
1103
{
1104
1105
	seq_puts(m, "#           TASK-PID    CPU#    TIMESTAMP  FUNCTION\n");
	seq_puts(m, "#              | |       |          |         |\n");
1106
1107
1108
}


Ingo Molnar's avatar
Ingo Molnar committed
1109
static void
1110
1111
1112
1113
1114
1115
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;
1116
1117
	unsigned long total;
	unsigned long entries;
1118
1119
1120
1121
1122
	const char *name = "preemption";

	if (type)
		name = type->name;

1123
1124
1125
	entries = ring_buffer_entries(iter->tr->buffer);
	total = entries +
		ring_buffer_overruns(iter->tr->buffer);
1126
1127
1128
1129
1130
1131
1132

	seq_printf(m, "%s latency trace v1.1.5 on %s\n",
		   name, UTS_RELEASE);
	seq_puts(m, "-----------------------------------"
		 "---------------------------------\n");
	seq_printf(m, " latency: %lu us, #%lu/%lu, CPU#%d |"
		   " (M:%s VP:%d, KP:%d, SP:%d HP:%d",
Steven Rostedt's avatar
Steven Rostedt committed
1133
		   nsecs_to_usecs(data->saved_latency),
1134
		   entries,
1135
		   total,
1136
1137
1138
1139
1140
		   tr->cpu,
#if defined(CONFIG_PREEMPT_NONE)
		   "server",
#elif defined(CONFIG_PREEMPT_VOLUNTARY)
		   "desktop",
1141
#elif defined(CONFIG_PREEMPT)
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
		   "preempt",
#else
		   "unknown",
#endif
		   /* These are reserved for later use */
		   0, 0, 0, 0);
#ifdef CONFIG_SMP
	seq_printf(m, " #P:%d)\n", num_online_cpus());
#else
	seq_puts(m, ")\n");
#endif
	seq_puts(m, "    -----------------\n");
	seq_printf(m, "    | task: %.16s-%d "
		   "(uid:%d nice:%ld policy:%ld rt_prio:%ld)\n",
		   data->comm, data->pid, data->uid, data->nice,
		   data->policy, data->rt_priority);
	seq_puts(m, "    -----------------\n");

	if (data->critical_start) {
		seq_puts(m, " => started at: ");
1162
1163
		seq_print_ip_sym(&iter->seq, data->critical_start, sym_flags);
		trace_print_seq(m, &iter->seq);
1164
		seq_puts(m, "\n => ended at:   ");
1165
1166
		seq_print_ip_sym(&iter->seq, data->critical_end, sym_flags);
		trace_print_seq(m, &iter->seq);
1167
1168
1169
1170
1171
1172
		seq_puts(m, "\n");
	}

	seq_puts(m, "\n");
}

Ingo Molnar's avatar
Ingo Molnar committed
1173
static void
1174
lat_print_generic(struct trace_seq *s, struct trace_entry *entry, int cpu)
1175
1176
1177
1178
{
	int hardirq, softirq;
	char *comm;

1179
	comm = trace_find_cmdline(entry->pid);
1180

1181
	trace_seq_printf(s, "%8.8s-%-5d ", comm, entry->pid);
1182
	trace_seq_printf(s, "%3d", cpu);
1183
	trace_seq_printf(s, "%c%c",
1184
1185
			(entry->flags & TRACE_FLAG_IRQS_OFF) ? 'd' : '.',
			((entry->flags & TRACE_FLAG_NEED_RESCHED) ? 'N' : '.'));
1186

1187
1188
	hardirq = entry->flags & TRACE_FLAG_HARDIRQ;
	softirq = entry->flags & TRACE_FLAG_SOFTIRQ;
Ingo Molnar's avatar
Ingo Molnar committed
1189
	if (hardirq && softirq) {
1190
		trace_seq_putc(s, 'H');
Ingo Molnar's avatar
Ingo Molnar committed
1191
1192
	} else {
		if (hardirq) {
1193
			trace_seq_putc(s, 'h');
Ingo Molnar's avatar
Ingo Molnar committed
1194
		} else {
1195
			if (softirq)
1196
				trace_seq_putc(s, 's');
1197
			else
1198
				trace_seq_putc(s, '.');
1199
1200
1201
		}
	}

1202
1203
	if (entry->preempt_count)
		trace_seq_printf(s, "%x", entry->preempt_count);
1204
	else
1205
		trace_seq_puts(s, ".");
1206
1207
1208
1209
}

unsigned long preempt_mark_thresh = 100;

Ingo Molnar's avatar
Ingo Molnar committed
1210
static void
1211
lat_print_timestamp(struct trace_seq *s, u64 abs_usecs,
1212
1213
		    unsigned long rel_usecs)
{
1214
	trace_seq_printf(s, " %4lldus", abs_usecs);
1215
	if (rel_usecs > preempt_mark_thresh)
1216
		trace_seq_puts(s, "!: ");
1217
	else if (rel_usecs > 1)
1218
		trace_seq_puts(s, "+: ");
1219
	else
1220
		trace_seq_puts(s, " : ");
1221
1222
1223
1224
}

static const char state_to_char[] = TASK_STATE_TO_CHAR_STR;

1225
1226
1227
1228
1229
/*
 * The message is supposed to contain an ending newline.
 * If the printing stops prematurely, try to add a newline of our own.
 */
void trace_seq_print_cont(struct trace_seq *s, struct trace_iterator *iter)
1230
1231
{
	struct trace_entry *ent;