trace.c 89.9 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
/*
 * 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
 */
14
#include <linux/ring_buffer.h>
15
#include <linux/utsrelease.h>
16
17
#include <linux/stacktrace.h>
#include <linux/writeback.h>
18
19
#include <linux/kallsyms.h>
#include <linux/seq_file.h>
20
#include <linux/notifier.h>
21
#include <linux/irqflags.h>
22
#include <linux/debugfs.h>
23
#include <linux/pagemap.h>
24
25
26
#include <linux/hardirq.h>
#include <linux/linkage.h>
#include <linux/uaccess.h>
27
#include <linux/kprobes.h>
28
29
30
#include <linux/ftrace.h>
#include <linux/module.h>
#include <linux/percpu.h>
31
#include <linux/splice.h>
32
#include <linux/kdebug.h>
33
34
#include <linux/ctype.h>
#include <linux/init.h>
35
#include <linux/poll.h>
36
37
#include <linux/gfp.h>
#include <linux/fs.h>
Ingo Molnar's avatar
Ingo Molnar committed
38

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

42
43
#define TRACE_BUFFER_FLAGS	(RB_FL_OVERWRITE)

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

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

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

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

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

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

/*
 * Kill all tracing for good (never come back).
 * It is initialized to 1 but will turn to zero if the initialization
 * of the tracer is successful. But that is the only place that sets
 * this back to zero.
 */
82
static int tracing_disabled = 1;
83

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

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

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

98
static cpumask_var_t __read_mostly	tracing_buffer_mask;
99

100
101
102
/* Define which cpu buffers are currently read in trace_pipe */
static cpumask_var_t			tracing_reader_cpumask;

103
#define for_each_tracing_cpu(cpu)	\
104
	for_each_cpu(cpu, tracing_buffer_mask)
105

106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
/*
 * 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;

121
122
123
124
125
static int tracing_set_tracer(const char *buf);

#define BOOTUP_TRACER_SIZE		100
static char bootup_tracer_buf[BOOTUP_TRACER_SIZE] __initdata;
static char *default_bootup_tracer;
126
127
128

static int __init set_ftrace(char *str)
{
129
130
	strncpy(bootup_tracer_buf, str, BOOTUP_TRACER_SIZE);
	default_bootup_tracer = bootup_tracer_buf;
131
132
	return 1;
}
133
__setup("ftrace=", set_ftrace);
134

135
136
137
138
139
140
static int __init set_ftrace_dump_on_oops(char *str)
{
	ftrace_dump_on_oops = 1;
	return 1;
}
__setup("ftrace_dump_on_oops", set_ftrace_dump_on_oops);
Steven Rostedt's avatar
Steven Rostedt committed
141

142
long
143
144
145
146
147
148
149
ns2usecs(cycle_t nsec)
{
	nsec += 500;
	do_div(nsec, 1000);
	return nsec;
}

Ingo Molnar's avatar
Ingo Molnar committed
150
cycle_t ftrace_now(int cpu)
151
{
152
153
154
	u64 ts = ring_buffer_time_stamp(cpu);
	ring_buffer_normalize_time_stamp(cpu, &ts);
	return ts;
155
156
}

Steven Rostedt's avatar
Steven Rostedt committed
157
158
159
160
161
162
163
164
165
166
167
168
/*
 * 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.
 */
169
170
171
172
static struct trace_array	global_trace;

static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu);

Steven Rostedt's avatar
Steven Rostedt committed
173
174
175
176
177
178
179
180
181
182
/*
 * 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.
 */
183
184
185
186
static struct trace_array	max_tr;

static DEFINE_PER_CPU(struct trace_array_cpu, max_data);

Steven Rostedt's avatar
Steven Rostedt committed
187
/* tracer_enabled is used to toggle activation of a tracer */
188
static int			tracer_enabled = 1;
Steven Rostedt's avatar
Steven Rostedt committed
189

190
191
192
193
194
195
196
197
198
199
200
201
202
/**
 * tracing_is_enabled - return tracer_enabled status
 *
 * This function is used by other tracers to know the status
 * of the tracer_enabled flag.  Tracers may use this function
 * to know if it should enable their features when starting
 * up. See irqsoff tracer for an example (start_irqsoff_tracer).
 */
int tracing_is_enabled(void)
{
	return tracer_enabled;
}

Steven Rostedt's avatar
Steven Rostedt committed
203
/*
204
205
206
 * 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.
207
208
209
210
211
 *
 * 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
212
 */
213
#define TRACE_BUF_SIZE_DEFAULT	1441792UL /* 16384 * 88 (sizeof(entry)) */
214

215
static unsigned long		trace_buf_size = TRACE_BUF_SIZE_DEFAULT;
216

Steven Rostedt's avatar
Steven Rostedt committed
217
/* trace_types holds a link list of available tracers. */
218
static struct tracer		*trace_types __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
219
220

/* current_trace points to the tracer that is currently active */
221
static struct tracer		*current_trace __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
222
223
224
225
226
227

/*
 * 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.
 */
228
229
static int			max_tracer_type_len;

Steven Rostedt's avatar
Steven Rostedt committed
230
231
232
233
234
235
/*
 * 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.
 */
236
static DEFINE_MUTEX(trace_types_lock);
Steven Rostedt's avatar
Steven Rostedt committed
237
238

/* trace_wait is a waitqueue for tasks blocked on trace_poll */
239
240
static DECLARE_WAIT_QUEUE_HEAD(trace_wait);

241
/* trace_flags holds trace_options default values */
242
unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK |
243
	TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO;
244

Steven Rostedt's avatar
Steven Rostedt committed
245
246
247
248
249
250
/**
 * 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.
 */
251
252
void trace_wake_up(void)
{
Ingo Molnar's avatar
Ingo Molnar committed
253
254
255
256
257
	/*
	 * The runqueue_is_locked() can fail, but this is the best we
	 * have for now:
	 */
	if (!(trace_flags & TRACE_ITER_BLOCK) && !runqueue_is_locked())
258
259
		wake_up(&trace_wait);
}
260

261
static int __init set_buf_size(char *str)
262
{
263
	unsigned long buf_size;
264
265
	int ret;

266
267
	if (!str)
		return 0;
268
	ret = strict_strtoul(str, 0, &buf_size);
269
	/* nr_entries can not be zero */
270
	if (ret < 0 || buf_size == 0)
271
		return 0;
272
	trace_buf_size = buf_size;
273
274
	return 1;
}
275
__setup("trace_buf_size=", set_buf_size);
276

Steven Rostedt's avatar
Steven Rostedt committed
277
278
279
280
281
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

Steven Rostedt's avatar
Steven Rostedt committed
282
/* These must match the bit postions in trace_iterator_flags */
283
284
285
286
287
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
Ingo Molnar's avatar
Ingo Molnar committed
288
	"raw",
289
	"hex",
Ingo Molnar's avatar
Ingo Molnar committed
290
	"bin",
291
	"block",
Ingo Molnar's avatar
Ingo Molnar committed
292
	"stacktrace",
Ingo Molnar's avatar
Ingo Molnar committed
293
	"sched-tree",
294
	"ftrace_printk",
295
	"ftrace_preempt",
296
	"branch",
297
	"annotate",
298
	"userstacktrace",
299
	"sym-userobj",
300
	"printk-msg-only",
301
	"context-info",
302
303
304
	NULL
};

Steven Rostedt's avatar
Steven Rostedt committed
305
306
307
308
309
310
311
312
313
/*
 * 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.
 */
314
315
static raw_spinlock_t ftrace_max_lock =
	(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
316
317
318
319
320
321

/*
 * 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
322
static void
323
324
325
326
327
328
329
330
331
332
333
334
__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;
335
	data->uid = task_uid(tsk);
336
337
338
339
340
	data->nice = tsk->static_prio - 20 - MAX_RT_PRIO;
	data->policy = tsk->policy;
	data->rt_priority = tsk->rt_priority;

	/* record this tasks comm */
341
	tracing_record_cmdline(tsk);
342
343
}

344
345
346
347
348
ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt)
{
	int len;
	int ret;

349
350
351
	if (!cnt)
		return 0;

352
353
354
355
356
357
358
	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);
359
	if (ret == cnt)
360
361
		return -EFAULT;

362
363
	cnt -= ret;

364
365
	s->readpos += len;
	return cnt;
366
367
}

368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt)
{
	int len;
	void *ret;

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

	len = s->len - s->readpos;
	if (cnt > len)
		cnt = len;
	ret = memcpy(buf, s->buffer + s->readpos, cnt);
	if (!ret)
		return -EFAULT;

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

Ingo Molnar's avatar
Ingo Molnar committed
387
static void
388
389
390
391
392
393
394
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);

395
	trace_seq_init(s);
396
397
}

Steven Rostedt's avatar
Steven Rostedt committed
398
399
400
401
402
403
404
405
406
/**
 * 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
407
void
408
409
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
410
	struct ring_buffer *buf = tr->buffer;
411

412
	WARN_ON_ONCE(!irqs_disabled());
413
	__raw_spin_lock(&ftrace_max_lock);
414
415
416
417

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

418
	ftrace_disable_cpu();
419
	ring_buffer_reset(tr->buffer);
420
	ftrace_enable_cpu();
421
422

	__update_max_tr(tr, tsk, cpu);
423
	__raw_spin_unlock(&ftrace_max_lock);
424
425
426
427
428
429
430
}

/**
 * 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
431
432
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
433
 */
Ingo Molnar's avatar
Ingo Molnar committed
434
void
435
436
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
437
	int ret;
438

439
	WARN_ON_ONCE(!irqs_disabled());
440
	__raw_spin_lock(&ftrace_max_lock);
441

442
443
	ftrace_disable_cpu();

444
445
446
	ring_buffer_reset(max_tr.buffer);
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

447
448
	ftrace_enable_cpu();

449
	WARN_ON_ONCE(ret && ret != -EAGAIN);
450
451

	__update_max_tr(tr, tsk, cpu);
452
	__raw_spin_unlock(&ftrace_max_lock);
453
454
}

Steven Rostedt's avatar
Steven Rostedt committed
455
456
457
458
459
460
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
461
int register_tracer(struct tracer *type)
462
463
__releases(kernel_lock)
__acquires(kernel_lock)
464
465
466
467
468
469
470
471
472
473
{
	struct tracer *t;
	int len;
	int ret = 0;

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

Ingo Molnar's avatar
Ingo Molnar committed
474
475
476
477
478
479
480
	/*
	 * When this gets called we hold the BKL which means that
	 * preemption is disabled. Various trace selftests however
	 * need to disable and enable preemption for successful tests.
	 * So we drop the BKL here and grab it after the tests again.
	 */
	unlock_kernel();
481
	mutex_lock(&trace_types_lock);
Ingo Molnar's avatar
Ingo Molnar committed
482

483
484
	tracing_selftest_running = true;

485
486
487
488
489
490
491
492
493
494
	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;
		}
	}

495
496
497
498
499
500
501
	if (!type->set_flag)
		type->set_flag = &dummy_set_flag;
	if (!type->flags)
		type->flags = &dummy_tracer_flags;
	else
		if (!type->flags->opts)
			type->flags->opts = dummy_tracer_opt;
502
503
504
	if (!type->wait_pipe)
		type->wait_pipe = default_wait_pipe;

505

Steven Rostedt's avatar
Steven Rostedt committed
506
#ifdef CONFIG_FTRACE_STARTUP_TEST
507
	if (type->selftest && !tracing_selftest_disabled) {
Steven Rostedt's avatar
Steven Rostedt committed
508
509
510
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
		int i;
511

Steven Rostedt's avatar
Steven Rostedt committed
512
513
514
515
516
517
518
		/*
		 * Run a selftest on this tracer.
		 * Here we reset the trace buffer, and set the current
		 * tracer to be this tracer. The tracer can then run some
		 * internal tracing to verify that everything is in order.
		 * If we fail, we do not register this tracer.
		 */
Ingo Molnar's avatar
Ingo Molnar committed
519
		for_each_tracing_cpu(i)
520
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
521

Steven Rostedt's avatar
Steven Rostedt committed
522
523
524
525
526
527
528
529
530
531
		current_trace = type;
		/* the test is responsible for initializing and enabling */
		pr_info("Testing tracer %s: ", type->name);
		ret = type->selftest(type, tr);
		/* the test is responsible for resetting too */
		current_trace = saved_tracer;
		if (ret) {
			printk(KERN_CONT "FAILED!\n");
			goto out;
		}
Steven Rostedt's avatar
Steven Rostedt committed
532
		/* Only reset on passing, to avoid touching corrupted buffers */
Ingo Molnar's avatar
Ingo Molnar committed
533
		for_each_tracing_cpu(i)
534
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
535

Steven Rostedt's avatar
Steven Rostedt committed
536
537
538
539
		printk(KERN_CONT "PASSED\n");
	}
#endif

540
541
542
543
544
	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
545

546
 out:
547
	tracing_selftest_running = false;
548
549
	mutex_unlock(&trace_types_lock);

Steven Rostedt's avatar
Steven Rostedt committed
550
551
552
553
554
555
556
557
558
559
560
561
	if (ret || !default_bootup_tracer)
		goto out_unlock;

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

	printk(KERN_INFO "Starting tracer '%s'\n", type->name);
	/* Do we want this tracer to start on bootup? */
	tracing_set_tracer(type->name);
	default_bootup_tracer = NULL;
	/* disable other selftests, since this will break it. */
	tracing_selftest_disabled = 1;
562
#ifdef CONFIG_FTRACE_STARTUP_TEST
Steven Rostedt's avatar
Steven Rostedt committed
563
564
	printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n",
	       type->name);
565
566
#endif

Steven Rostedt's avatar
Steven Rostedt committed
567
 out_unlock:
568
	lock_kernel();
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
	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;
587
588
589
590
591
592
593
594
595

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

596
597
598
599
600
601
602
603
604
605
606
607
608
	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);
}

609
void tracing_reset(struct trace_array *tr, int cpu)
610
{
611
	ftrace_disable_cpu();
612
	ring_buffer_reset_cpu(tr->buffer, cpu);
613
	ftrace_enable_cpu();
614
615
}

616
617
618
619
620
621
622
623
624
625
void tracing_reset_online_cpus(struct trace_array *tr)
{
	int cpu;

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

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

626
627
628
629
630
#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;
631
static raw_spinlock_t trace_cmdline_lock = __RAW_SPIN_LOCK_UNLOCKED;
632
633

/* temporary disable recording */
634
static atomic_t trace_record_cmdline_disabled __read_mostly;
635
636
637
638
639
640
641
642

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;
}

643
644
645
static int trace_stop_count;
static DEFINE_SPINLOCK(tracing_start_lock);

646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
/**
 * ftrace_off_permanent - disable all ftrace code permanently
 *
 * This should only be called when a serious anomally has
 * been detected.  This will turn off the function tracing,
 * ring buffers, and other tracing utilites. It takes no
 * locks and can be called from any context.
 */
void ftrace_off_permanent(void)
{
	tracing_disabled = 1;
	ftrace_stop();
	tracing_off_permanent();
}

661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
/**
 * tracing_start - quick start of the tracer
 *
 * If tracing is enabled but was stopped by tracing_stop,
 * this will start the tracer back up.
 */
void tracing_start(void)
{
	struct ring_buffer *buffer;
	unsigned long flags;

	if (tracing_disabled)
		return;

	spin_lock_irqsave(&tracing_start_lock, flags);
676
677
678
679
680
681
	if (--trace_stop_count) {
		if (trace_stop_count < 0) {
			/* Someone screwed up their debugging */
			WARN_ON_ONCE(1);
			trace_stop_count = 0;
		}
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
		goto out;
	}


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

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

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

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

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

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

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

 out:
	spin_unlock_irqrestore(&tracing_start_lock, flags);
}

Ingo Molnar's avatar
Ingo Molnar committed
727
void trace_stop_cmdline_recording(void);
728

Ingo Molnar's avatar
Ingo Molnar committed
729
static void trace_save_cmdline(struct task_struct *tsk)
730
731
732
733
734
735
736
737
738
739
740
741
742
{
	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.
	 */
743
	if (!__raw_spin_trylock(&trace_cmdline_lock))
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
		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);

761
	__raw_spin_unlock(&trace_cmdline_lock);
762
763
}

764
char *trace_find_cmdline(int pid)
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
{
	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
785
void tracing_record_cmdline(struct task_struct *tsk)
786
787
788
789
790
791
792
{
	if (atomic_read(&trace_record_cmdline_disabled))
		return;

	trace_save_cmdline(tsk);
}

793
void
794
795
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
796
797
798
{
	struct task_struct *tsk = current;

799
800
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
801
	entry->tgid               	= (tsk) ? tsk->tgid : 0;
802
	entry->flags =
803
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
804
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
805
806
807
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
808
809
810
811
812
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}

813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
struct ring_buffer_event *trace_buffer_lock_reserve(struct trace_array *tr,
						    unsigned char type,
						    unsigned long len,
						    unsigned long flags, int pc)
{
	struct ring_buffer_event *event;

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

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

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

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

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

846
847
848
849
850
851
852
853
854
855
856
857
858
859
struct ring_buffer_event *
trace_current_buffer_lock_reserve(unsigned char type, unsigned long len,
				  unsigned long flags, int pc)
{
	return trace_buffer_lock_reserve(&global_trace,
					 type, len, flags, pc);
}

void trace_current_buffer_unlock_commit(struct ring_buffer_event *event,
					unsigned long flags, int pc)
{
	return trace_buffer_unlock_commit(&global_trace, event, flags, pc);
}

Ingo Molnar's avatar
Ingo Molnar committed
860
void
861
trace_function(struct trace_array *tr,
862
863
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
864
{
865
	struct ring_buffer_event *event;
866
	struct ftrace_entry *entry;
867

868
869
870
871
	/* If we are reading the ring buffer, don't trace */
	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

872
873
	event = trace_buffer_lock_reserve(tr, TRACE_FN, sizeof(*entry),
					  flags, pc);
874
875
876
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
877
878
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
879
	ring_buffer_unlock_commit(tr->buffer, event);
880
881
}

882
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
883
884
885
886
887
888
889
890
891
892
893
static void __trace_graph_entry(struct trace_array *tr,
				struct ftrace_graph_ent *trace,
				unsigned long flags,
				int pc)
{
	struct ring_buffer_event *event;
	struct ftrace_graph_ent_entry *entry;

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

894
895
	event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_ENT,
					  sizeof(*entry), flags, pc);
896
897
898
899
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
	entry->graph_ent			= *trace;
900
	ring_buffer_unlock_commit(global_trace.buffer, event);
901
902
903
}

static void __trace_graph_return(struct trace_array *tr,
904
				struct ftrace_graph_ret *trace,
905
906
907
908
				unsigned long flags,
				int pc)
{
	struct ring_buffer_event *event;
909
	struct ftrace_graph_ret_entry *entry;
910
911
912
913

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

914
915
	event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_RET,
					  sizeof(*entry), flags, pc);
916
917
918
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
919
	entry->ret				= *trace;
920
	ring_buffer_unlock_commit(global_trace.buffer, event);
921
922
923
}
#endif

Ingo Molnar's avatar
Ingo Molnar committed
924
void
Ingo Molnar's avatar
Ingo Molnar committed
925
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
926
927
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
Ingo Molnar's avatar
Ingo Molnar committed
928
929
{
	if (likely(!atomic_read(&data->disabled)))
930
		trace_function(tr, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
931
932
}

933
934
935
static void __ftrace_trace_stack(struct trace_array *tr,
				 unsigned long flags,
				 int skip, int pc)
Ingo Molnar's avatar
Ingo Molnar committed
936
{
937
#ifdef CONFIG_STACKTRACE
938
	struct ring_buffer_event *event;
939
	struct stack_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
940
941
	struct stack_trace trace;

942
943
	event = trace_buffer_lock_reserve(tr, TRACE_STACK,
					  sizeof(*entry), flags, pc);
944
945
946
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
947
	memset(&entry->caller, 0, sizeof(entry->caller));
Ingo Molnar's avatar
Ingo Molnar committed
948
949
950
951

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
952
	trace.entries		= entry->caller;
Ingo Molnar's avatar
Ingo Molnar committed
953
954

	save_stack_trace(&trace);
955
	ring_buffer_unlock_commit(tr->buffer, event);
956
#endif
Ingo Molnar's avatar
Ingo Molnar committed
957
958
}

959
960
961
962
963
964
965
static void ftrace_trace_stack(struct trace_array *tr,
			       unsigned long flags,
			       int skip, int pc)
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

966
	__ftrace_trace_stack(tr, flags, skip, pc);
967
968
}

969
970
void __trace_stack(struct trace_array *tr,
		   unsigned long flags,
971
		   int skip, int pc)
972
{
973
	__ftrace_trace_stack(tr, flags, skip, pc);
974
975
}

976
static void ftrace_trace_userstack(struct trace_array *tr,
977
				   unsigned long flags, int pc)
978
{
979
#ifdef CONFIG_STACKTRACE
980
	struct ring_buffer_event *event;
981
982
983
984
985
986
	struct userstack_entry *entry;
	struct stack_trace trace;

	if (!(trace_flags & TRACE_ITER_USERSTACKTRACE))
		return;

987
988
	event = trace_buffer_lock_reserve(tr, TRACE_USER_STACK,
					  sizeof(*entry), flags, pc);
989
990
991
992
993
994
995
996
997
998
999
1000
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);

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

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

	save_stack_trace_user(&trace);
1001
	ring_buffer_unlock_commit(tr->buffer, event);
1002
#endif
1003
1004
}

1005
1006
#ifdef UNUSED
static void __trace_userstack(struct trace_array *tr, unsigned long flags)
1007
{
1008
	ftrace_trace_userstack(tr, flags, preempt_count());
1009
}
1010
#endif /* UNUSED */
1011

1012
static void
1013
ftrace_trace_special(void *__tr,
1014
1015
		     unsigned long arg1, unsigned long arg2, unsigned long arg3,
		     int pc)
Ingo Molnar's avatar
Ingo Molnar committed
1016
{
1017
	struct ring_buffer_event *event;
Ingo Molnar's avatar
Ingo Molnar committed
1018
	struct trace_array *tr = __tr;
1019
	struct special_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
1020

1021
1022
	event = trace_buffer_lock_reserve(tr, TRACE_SPECIAL,
					  sizeof(*entry), 0, pc);
1023
1024
1025
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1026
1027
1028
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
1029
	trace_buffer_unlock_commit(tr, event, 0, pc);
Ingo Molnar's avatar
Ingo Molnar committed
1030
1031
}

1032
1033
1034
1035
void
__trace_special(void *__tr, void *__data,
		unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
1036
	ftrace_trace_special(__tr, arg1, arg2, arg3, preempt_count());
1037
1038
}

Ingo Molnar's avatar
Ingo Molnar committed
1039
void
1040
tracing_sched_switch_trace(struct trace_array *tr,
Ingo Molnar's avatar
Ingo Molnar committed
1041
1042
			   struct task_struct *prev,
			   struct task_struct *next,
1043
			   unsigned long flags, int pc)
1044
{
1045
	struct ring_buffer_event *event;
1046
	struct ctx_switch_entry *entry;
1047

1048
1049
	event = trace_buffer_lock_reserve(tr, TRACE_CTX,
					  sizeof(*entry), flags, pc);
1050
1051
1052
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1053
1054
1055
1056
1057
1058
1059
	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);
1060
	trace_buffer_unlock_commit(tr, event, flags, pc);
1061
1062
}

1063
1064
void
tracing_sched_wakeup_trace(struct trace_array *tr,
Ingo Molnar's avatar
Ingo Molnar committed
1065
1066
			   struct task_struct *wakee,
			   struct task_struct *curr,
1067
			   unsigned long flags, int pc)
1068
{
1069
	struct ring_buffer_event *event;
1070
	struct ctx_switch_entry *entry;
1071

1072
1073
	event = trace_buffer_lock_reserve(tr, TRACE_WAKE,
					  sizeof(*entry), flags, pc);
1074
1075
1076
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1077
1078
1079
1080
1081
1082
1083
	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);
1084
1085
1086
1087

	ring_buffer_unlock_commit(tr->buffer, event);
	ftrace_trace_stack(tr, flags, 6, pc);
	ftrace_trace_userstack(tr, flags, pc);
1088
1089
}

1090
1091
1092
1093
1094
void
ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1095
	unsigned long flags;
1096
	int cpu;
1097
	int pc;
1098

1099
	if (tracing_disabled)
1100
1101
		return;

1102
	pc = preempt_count();
1103
	local_irq_save(flags);
1104
1105
1106
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

1107
	if (likely(atomic_inc_return(&data->disabled) == 1))
1108
		ftrace_trace_special(tr, arg1, arg2, arg3, pc);
1109

1110
1111
	atomic_dec(&data->disabled);
	local_irq_restore(flags);
1112
1113
}

1114
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
1115
int trace_graph_entry(struct ftrace_graph_ent *trace)
1116
1117
1118
1119
1120
1121
1122
1123
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	unsigned long flags;
	long disabled;
	int cpu;
	int pc;

1124
1125
1126
	if (!ftrace_trace_task(current))
		return 0;

1127
1128
1129
	if (!ftrace_graph_addr(trace->func))
		return 0;

1130
	local_irq_save(flags);
1131
1132
1133
1134
1135
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);
	if (likely(disabled == 1)) {
		pc = preempt_count();
1136
		__trace_graph_entry(tr, trace, flags, pc);
1137
	}
1138
1139
1140
	/* Only do the atomic if it is not already set */
	if (!test_tsk_trace_graph(current))
		set_tsk_trace_graph(current);
1141
	atomic_dec(&data->disabled);
1142
	local_irq_restore(flags);
1143
1144

	return 1;
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
}

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

1156
	local_irq_save(flags);
1157
1158
1159
1160
1161
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);
	if (likely(disabled == 1)) {
		pc = preempt_count();
1162
		__trace_graph_return(tr, trace, flags, pc);
1163
	}
1164
1165
	if (!trace->depth)
		clear_tsk_trace_graph(current);
1166
	atomic_dec(&data->disabled);
1167
	local_irq_restore(flags);
1168
}
1169
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
1170

1171
1172
enum trace_file_type {
	TRACE_FILE_LAT_FMT	= 1,
1173
	TRACE_FILE_ANNOTATE	= 2,
1174
1175
};

1176
static void trace_iterator_increment(struct trace_iterator *iter)
1177
{
1178
1179
1180
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();

1181
	iter->idx++;
1182
1183
1184
1185
	if (iter->buffer_iter[iter->cpu])
		ring_buffer_read(iter->buffer_iter[iter->cpu], NULL);

	ftrace_enable_cpu();
1186
1187
}

Ingo Molnar's avatar
Ingo Molnar committed
1188
static struct trace_entry *
1189
peek_next_entry(struct trace_iterator *iter, int cpu, u64 *ts)
1190
{
1191
1192
	struct ring_buffer_event *event;
	struct ring_buffer_iter *buf_iter = iter->buffer_iter[cpu];
1193

1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
	/* 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();

1204
	return event ? ring_buffer_event_data(event) : NULL;
1205
}
1206

1207
static struct trace_entry *