trace.c 88.8 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
 * A selftest will lurk into the ring-buffer to count the
 * entries inserted during the selftest although some concurrent
51
 * insertions into the ring-buffer such as trace_printk could occurred
52
53
 * 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
	"trace_printk",
295
	"ftrace_preempt",
296
	"branch",
297
	"annotate",
298
	"userstacktrace",
299
	"sym-userobj",
300
	"printk-msg-only",
301
	"context-info",
302
	"latency-format",
303
304
305
	NULL
};

Steven Rostedt's avatar
Steven Rostedt committed
306
307
308
309
310
311
312
313
314
/*
 * 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.
 */
315
316
static raw_spinlock_t ftrace_max_lock =
	(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
317
318
319
320
321
322

/*
 * 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
323
static void
324
325
326
327
328
329
330
331
332
333
334
335
__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;
336
	data->uid = task_uid(tsk);
337
338
339
340
341
	data->nice = tsk->static_prio - 20 - MAX_RT_PRIO;
	data->policy = tsk->policy;
	data->rt_priority = tsk->rt_priority;

	/* record this tasks comm */
342
	tracing_record_cmdline(tsk);
343
344
}

345
346
347
348
349
ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt)
{
	int len;
	int ret;

350
351
352
	if (!cnt)
		return 0;

353
354
355
356
357
358
359
	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);
360
	if (ret == cnt)
361
362
		return -EFAULT;

363
364
	cnt -= ret;

365
	s->readpos += cnt;
366
	return cnt;
367
368
}

369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
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;

384
	s->readpos += cnt;
385
386
387
	return cnt;
}

Ingo Molnar's avatar
Ingo Molnar committed
388
static void
389
390
391
392
393
394
395
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);

396
	trace_seq_init(s);
397
398
}

Steven Rostedt's avatar
Steven Rostedt committed
399
400
401
402
403
404
405
406
407
/**
 * 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
408
void
409
410
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
411
	struct ring_buffer *buf = tr->buffer;
412

413
	WARN_ON_ONCE(!irqs_disabled());
414
	__raw_spin_lock(&ftrace_max_lock);
415
416
417
418

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

419
	ftrace_disable_cpu();
420
	ring_buffer_reset(tr->buffer);
421
	ftrace_enable_cpu();
422
423

	__update_max_tr(tr, tsk, cpu);
424
	__raw_spin_unlock(&ftrace_max_lock);
425
426
427
428
429
430
431
}

/**
 * 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
432
433
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
434
 */
Ingo Molnar's avatar
Ingo Molnar committed
435
void
436
437
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
438
	int ret;
439

440
	WARN_ON_ONCE(!irqs_disabled());
441
	__raw_spin_lock(&ftrace_max_lock);
442

443
444
	ftrace_disable_cpu();

445
446
447
	ring_buffer_reset(max_tr.buffer);
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

448
449
	ftrace_enable_cpu();

450
	WARN_ON_ONCE(ret && ret != -EAGAIN);
451
452

	__update_max_tr(tr, tsk, cpu);
453
	__raw_spin_unlock(&ftrace_max_lock);
454
455
}

Steven Rostedt's avatar
Steven Rostedt committed
456
457
458
459
460
461
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
462
int register_tracer(struct tracer *type)
463
464
__releases(kernel_lock)
__acquires(kernel_lock)
465
466
467
468
469
470
471
472
473
474
{
	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
475
476
477
478
479
480
481
	/*
	 * 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();
482
	mutex_lock(&trace_types_lock);
Ingo Molnar's avatar
Ingo Molnar committed
483

484
485
	tracing_selftest_running = true;

486
487
488
489
490
491
492
493
494
495
	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;
		}
	}

496
497
498
499
500
501
502
	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;
503
504
505
	if (!type->wait_pipe)
		type->wait_pipe = default_wait_pipe;

506

Steven Rostedt's avatar
Steven Rostedt committed
507
#ifdef CONFIG_FTRACE_STARTUP_TEST
508
	if (type->selftest && !tracing_selftest_disabled) {
Steven Rostedt's avatar
Steven Rostedt committed
509
510
511
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
		int i;
512

Steven Rostedt's avatar
Steven Rostedt committed
513
514
515
516
517
518
519
		/*
		 * 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
520
		for_each_tracing_cpu(i)
521
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
522

Steven Rostedt's avatar
Steven Rostedt committed
523
524
525
526
527
528
529
530
531
532
		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
533
		/* Only reset on passing, to avoid touching corrupted buffers */
Ingo Molnar's avatar
Ingo Molnar committed
534
		for_each_tracing_cpu(i)
535
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
536

Steven Rostedt's avatar
Steven Rostedt committed
537
538
539
540
		printk(KERN_CONT "PASSED\n");
	}
#endif

541
542
543
544
545
	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
546

547
 out:
548
	tracing_selftest_running = false;
549
550
	mutex_unlock(&trace_types_lock);

Steven Rostedt's avatar
Steven Rostedt committed
551
552
553
554
555
556
557
558
559
560
561
562
	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;
563
#ifdef CONFIG_FTRACE_STARTUP_TEST
Steven Rostedt's avatar
Steven Rostedt committed
564
565
	printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n",
	       type->name);
566
567
#endif

Steven Rostedt's avatar
Steven Rostedt committed
568
 out_unlock:
569
	lock_kernel();
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
	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;
588
589
590
591
592
593
594
595
596

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

597
598
599
600
601
602
603
604
605
606
607
608
609
	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);
}

610
void tracing_reset(struct trace_array *tr, int cpu)
611
{
612
	ftrace_disable_cpu();
613
	ring_buffer_reset_cpu(tr->buffer, cpu);
614
	ftrace_enable_cpu();
615
616
}

617
618
619
620
621
622
623
624
625
626
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);
}

627
628
629
630
631
#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;
632
static raw_spinlock_t trace_cmdline_lock = __RAW_SPIN_LOCK_UNLOCKED;
633
634

/* temporary disable recording */
635
static atomic_t trace_record_cmdline_disabled __read_mostly;
636
637
638
639
640
641
642
643

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;
}

644
645
646
static int trace_stop_count;
static DEFINE_SPINLOCK(tracing_start_lock);

647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
/**
 * 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();
}

662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
/**
 * 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);
677
678
679
680
681
682
	if (--trace_stop_count) {
		if (trace_stop_count < 0) {
			/* Someone screwed up their debugging */
			WARN_ON_ONCE(1);
			trace_stop_count = 0;
		}
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
727
		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
728
void trace_stop_cmdline_recording(void);
729

Ingo Molnar's avatar
Ingo Molnar committed
730
static void trace_save_cmdline(struct task_struct *tsk)
731
732
733
734
735
736
737
738
739
740
741
742
743
{
	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.
	 */
744
	if (!__raw_spin_trylock(&trace_cmdline_lock))
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
		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);

762
	__raw_spin_unlock(&trace_cmdline_lock);
763
764
}

765
char *trace_find_cmdline(int pid)
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
{
	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
786
void tracing_record_cmdline(struct task_struct *tsk)
787
788
789
790
791
792
793
{
	if (atomic_read(&trace_record_cmdline_disabled))
		return;

	trace_save_cmdline(tsk);
}

794
void
795
796
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
797
798
799
{
	struct task_struct *tsk = current;

800
801
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
802
	entry->tgid			= (tsk) ? tsk->tgid : 0;
803
	entry->flags =
804
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
805
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
806
807
808
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
809
810
811
812
813
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}

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
846
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();
}

847
848
849
850
851
852
853
854
855
856
857
858
859
860
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
861
void
862
trace_function(struct trace_array *tr,
863
864
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
865
{
866
	struct ring_buffer_event *event;
867
	struct ftrace_entry *entry;
868

869
870
871
872
	/* If we are reading the ring buffer, don't trace */
	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

873
874
	event = trace_buffer_lock_reserve(tr, TRACE_FN, sizeof(*entry),
					  flags, pc);
875
876
877
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
878
879
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
880
	ring_buffer_unlock_commit(tr->buffer, event);
881
882
}

883
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
884
885
886
887
888
889
890
891
892
893
894
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;

895
896
	event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_ENT,
					  sizeof(*entry), flags, pc);
897
898
899
900
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
	entry->graph_ent			= *trace;
901
	ring_buffer_unlock_commit(global_trace.buffer, event);
902
903
904
}

static void __trace_graph_return(struct trace_array *tr,
905
				struct ftrace_graph_ret *trace,
906
907
908
909
				unsigned long flags,
				int pc)
{
	struct ring_buffer_event *event;
910
	struct ftrace_graph_ret_entry *entry;
911
912
913
914

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

915
916
	event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_RET,
					  sizeof(*entry), flags, pc);
917
918
919
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
920
	entry->ret				= *trace;
921
	ring_buffer_unlock_commit(global_trace.buffer, event);
922
923
924
}
#endif

Ingo Molnar's avatar
Ingo Molnar committed
925
void
Ingo Molnar's avatar
Ingo Molnar committed
926
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
927
928
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
Ingo Molnar's avatar
Ingo Molnar committed
929
930
{
	if (likely(!atomic_read(&data->disabled)))
931
		trace_function(tr, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
932
933
}

934
935
936
static void __ftrace_trace_stack(struct trace_array *tr,
				 unsigned long flags,
				 int skip, int pc)
Ingo Molnar's avatar
Ingo Molnar committed
937
{
938
#ifdef CONFIG_STACKTRACE
939
	struct ring_buffer_event *event;
940
	struct stack_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
941
942
	struct stack_trace trace;

943
944
	event = trace_buffer_lock_reserve(tr, TRACE_STACK,
					  sizeof(*entry), flags, pc);
945
946
947
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
948
	memset(&entry->caller, 0, sizeof(entry->caller));
Ingo Molnar's avatar
Ingo Molnar committed
949
950
951
952

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
953
	trace.entries		= entry->caller;
Ingo Molnar's avatar
Ingo Molnar committed
954
955

	save_stack_trace(&trace);
956
	ring_buffer_unlock_commit(tr->buffer, event);
957
#endif
Ingo Molnar's avatar
Ingo Molnar committed
958
959
}

960
961
962
963
964
965
966
static void ftrace_trace_stack(struct trace_array *tr,
			       unsigned long flags,
			       int skip, int pc)
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

967
	__ftrace_trace_stack(tr, flags, skip, pc);
968
969
}

970
971
void __trace_stack(struct trace_array *tr,
		   unsigned long flags,
972
		   int skip, int pc)
973
{
974
	__ftrace_trace_stack(tr, flags, skip, pc);
975
976
}

977
static void ftrace_trace_userstack(struct trace_array *tr,
978
				   unsigned long flags, int pc)
979
{
980
#ifdef CONFIG_STACKTRACE
981
	struct ring_buffer_event *event;
982
983
984
985
986
987
	struct userstack_entry *entry;
	struct stack_trace trace;

	if (!(trace_flags & TRACE_ITER_USERSTACKTRACE))
		return;

988
989
	event = trace_buffer_lock_reserve(tr, TRACE_USER_STACK,
					  sizeof(*entry), flags, pc);
990
991
992
993
994
995
996
997
998
999
1000
1001
	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);
1002
	ring_buffer_unlock_commit(tr->buffer, event);
1003
#endif
1004
1005
}

1006
1007
#ifdef UNUSED
static void __trace_userstack(struct trace_array *tr, unsigned long flags)
1008
{
1009
	ftrace_trace_userstack(tr, flags, preempt_count());
1010
}
1011
#endif /* UNUSED */
1012

1013
static void
1014
ftrace_trace_special(void *__tr,
1015
1016
		     unsigned long arg1, unsigned long arg2, unsigned long arg3,
		     int pc)
Ingo Molnar's avatar
Ingo Molnar committed
1017
{
1018
	struct ring_buffer_event *event;
Ingo Molnar's avatar
Ingo Molnar committed
1019
	struct trace_array *tr = __tr;
1020
	struct special_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
1021

1022
1023
	event = trace_buffer_lock_reserve(tr, TRACE_SPECIAL,
					  sizeof(*entry), 0, pc);
1024
1025
1026
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1027
1028
1029
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
1030
	trace_buffer_unlock_commit(tr, event, 0, pc);
Ingo Molnar's avatar
Ingo Molnar committed
1031
1032
}

1033
1034
1035
1036
void
__trace_special(void *__tr, void *__data,
		unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
1037
	ftrace_trace_special(__tr, arg1, arg2, arg3, preempt_count());
1038
1039
}

Ingo Molnar's avatar
Ingo Molnar committed
1040
void
1041
tracing_sched_switch_trace(struct trace_array *tr,
Ingo Molnar's avatar
Ingo Molnar committed
1042
1043
			   struct task_struct *prev,
			   struct task_struct *next,
1044
			   unsigned long flags, int pc)
1045
{
1046
	struct ring_buffer_event *event;
1047
	struct ctx_switch_entry *entry;
1048

1049
1050
	event = trace_buffer_lock_reserve(tr, TRACE_CTX,
					  sizeof(*entry), flags, pc);
1051
1052
1053
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1054
1055
1056
1057
1058
1059
1060
	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);
1061
	trace_buffer_unlock_commit(tr, event, flags, pc);
1062
1063
}

1064
1065
void
tracing_sched_wakeup_trace(struct trace_array *tr,
Ingo Molnar's avatar
Ingo Molnar committed
1066
1067
			   struct task_struct *wakee,
			   struct task_struct *curr,
1068
			   unsigned long flags, int pc)
1069
{
1070
	struct ring_buffer_event *event;
1071
	struct ctx_switch_entry *entry;
1072

1073
1074
	event = trace_buffer_lock_reserve(tr, TRACE_WAKE,
					  sizeof(*entry), flags, pc);
1075
1076
1077
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1078
1079
1080
1081
1082
1083
1084
	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);
1085
1086
1087
1088

	ring_buffer_unlock_commit(tr->buffer, event);
	ftrace_trace_stack(tr, flags, 6, pc);
	ftrace_trace_userstack(tr, flags, pc);
1089
1090
}

1091
1092
1093
1094
1095
void
ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1096
	unsigned long flags;
1097
	int cpu;
1098
	int pc;
1099

1100
	if (tracing_disabled)
1101
1102
		return;

1103
	pc = preempt_count();
1104
	local_irq_save(flags);
1105
1106
1107
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

1108
	if (likely(atomic_inc_return(&data->disabled) == 1))
1109
		ftrace_trace_special(tr, arg1, arg2, arg3, pc);
1110

1111
1112
	atomic_dec(&data->disabled);
	local_irq_restore(flags);
1113
1114
}

1115
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
1116
int trace_graph_entry(struct ftrace_graph_ent *trace)
1117
1118
1119
1120
1121
1122
1123
1124
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	unsigned long flags;
	long disabled;
	int cpu;
	int pc;

1125
1126
1127
	if (!ftrace_trace_task(current))
		return 0;

1128
1129
1130
	if (!ftrace_graph_addr(trace->func))
		return 0;

1131
	local_irq_save(flags);
1132
1133
1134
1135
1136
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);
	if (likely(disabled == 1)) {
		pc = preempt_count();
1137
		__trace_graph_entry(tr, trace, flags, pc);
1138
	}
1139
1140
1141
	/* Only do the atomic if it is not already set */
	if (!test_tsk_trace_graph(current))
		set_tsk_trace_graph(current);
1142
	atomic_dec(&data->disabled);
1143
	local_irq_restore(flags);
1144
1145

	return 1;
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
}

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;

1157
	local_irq_save(flags);
1158
1159
1160
1161
1162
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);
	if (likely(disabled == 1)) {
		pc = preempt_count();
1163
		__trace_graph_return(tr, trace, flags, pc);
1164
	}
1165
1166
	if (!trace->depth)
		clear_tsk_trace_graph(current);
1167
	atomic_dec(&data->disabled);
1168
	local_irq_restore(flags);
1169
}
1170
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
1171

1172
1173
1174
1175
1176
1177
1178

/**
 * trace_vprintk - write binary msg to tracing buffer
 *
 */
int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args)
{
1179
1180
	static raw_spinlock_t trace_buf_lock =
		(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
	static u32 trace_buf[TRACE_BUF_SIZE];

	struct ring_buffer_event *event;
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	struct print_entry *entry;
	unsigned long flags;
	int resched;
	int cpu, len = 0, size, pc;

	if (unlikely(tracing_selftest_running || tracing_disabled))
		return 0;

	/* Don't pollute graph traces with trace_vprintk internals */
	pause_graph_tracing();

	pc = preempt_count();
	resched = ftrace_preempt_disable();
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

	if (unlikely(atomic_read(&data->disabled)))
		goto out;

1205
1206
1207
	/* Lockdep uses trace_printk for lock tracing */
	local_irq_save(flags);
	__raw_spin_lock(&trace_buf_lock);
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
	len = vbin_printf(trace_buf, TRACE_BUF_SIZE, fmt, args);

	if (len > TRACE_BUF_SIZE || len < 0)
		goto out_unlock;

	size = sizeof(*entry) + sizeof(u32) * len;
	event = trace_buffer_lock_reserve(tr, TRACE_PRINT, size, flags, pc);
	if (!event)
		goto out_unlock;
	entry = ring_buffer_event_data(event);
	entry->ip			= ip;
	entry->depth			= depth;
	entry->fmt			= fmt;

	memcpy(entry->buf, trace_buf, sizeof(u32) * len);
	ring_buffer_unlock_commit(tr->buffer, event);

out_unlock:
1226
1227
	__raw_spin_unlock(&trace_buf_lock);
	local_irq_restore(flags);
1228
1229
1230
1231