trace.c 95.1 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>
Alexey Dobriyan's avatar
Alexey Dobriyan committed
20
#include <linux/smp_lock.h>
21
#include <linux/notifier.h>
22
#include <linux/irqflags.h>
23
#include <linux/debugfs.h>
24
#include <linux/pagemap.h>
25
26
27
#include <linux/hardirq.h>
#include <linux/linkage.h>
#include <linux/uaccess.h>
28
#include <linux/kprobes.h>
29
30
31
#include <linux/ftrace.h>
#include <linux/module.h>
#include <linux/percpu.h>
32
#include <linux/splice.h>
33
#include <linux/kdebug.h>
34
#include <linux/string.h>
35
36
#include <linux/ctype.h>
#include <linux/init.h>
37
#include <linux/poll.h>
38
39
#include <linux/gfp.h>
#include <linux/fs.h>
Ingo Molnar's avatar
Ingo Molnar committed
40

41
#include "trace.h"
42
#include "trace_output.h"
43

44
45
#define TRACE_BUFFER_FLAGS	(RB_FL_OVERWRITE)

46
47
48
49
/*
 * On boot up, the ring buffer is set to the minimum size, so that
 * we do not waste memory on systems that are not using tracing.
 */
50
int ring_buffer_expanded;
51

52
53
/*
 * We need to change this state when a selftest is running.
54
55
 * A selftest will lurk into the ring-buffer to count the
 * entries inserted during the selftest although some concurrent
56
 * insertions into the ring-buffer such as trace_printk could occurred
57
58
 * at the same time, giving false positive or negative results.
 */
59
static bool __read_mostly tracing_selftest_running;
60

61
62
63
/*
 * If a tracer is running, we do not want to run SELFTEST.
 */
64
bool __read_mostly tracing_selftest_disabled;
65

66
67
68
69
70
71
72
73
74
75
76
77
78
79
/* 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;
}
80
81
82
83
84
85
86

/*
 * 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.
 */
87
static int tracing_disabled = 1;
88

89
DEFINE_PER_CPU(local_t, ftrace_cpu_disabled);
90
91
92
93
94
95
96
97
98
99
100
101
102

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();
}

103
static cpumask_var_t __read_mostly	tracing_buffer_mask;
104

105
106
107
/* Define which cpu buffers are currently read in trace_pipe */
static cpumask_var_t			tracing_reader_cpumask;

108
#define for_each_tracing_cpu(cpu)	\
109
	for_each_cpu(cpu, tracing_buffer_mask)
110

111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
/*
 * 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;

126
127
128
129
130
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;
131
132
133

static int __init set_ftrace(char *str)
{
134
135
	strncpy(bootup_tracer_buf, str, BOOTUP_TRACER_SIZE);
	default_bootup_tracer = bootup_tracer_buf;
136
137
	/* We are using ftrace early, expand it */
	ring_buffer_expanded = 1;
138
139
	return 1;
}
140
__setup("ftrace=", set_ftrace);
141

142
143
144
145
146
147
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
148

149
unsigned long long ns2usecs(cycle_t nsec)
150
151
152
153
154
155
{
	nsec += 500;
	do_div(nsec, 1000);
	return nsec;
}

Steven Rostedt's avatar
Steven Rostedt committed
156
157
158
159
160
161
162
163
164
165
166
167
/*
 * 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.
 */
168
169
170
171
static struct trace_array	global_trace;

static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu);

172
173
174
175
176
int filter_current_check_discard(struct ftrace_event_call *call, void *rec,
				 struct ring_buffer_event *event)
{
	return filter_check_discard(call, rec, global_trace.buffer, event);
}
177
EXPORT_SYMBOL_GPL(filter_current_check_discard);
178

179
180
181
182
183
184
185
186
187
188
189
190
191
cycle_t ftrace_now(int cpu)
{
	u64 ts;

	/* Early boot up does not have a buffer yet */
	if (!global_trace.buffer)
		return trace_clock_local();

	ts = ring_buffer_time_stamp(global_trace.buffer, cpu);
	ring_buffer_normalize_time_stamp(global_trace.buffer, cpu, &ts);

	return ts;
}
192

Steven Rostedt's avatar
Steven Rostedt committed
193
194
195
196
197
198
199
200
201
202
/*
 * 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.
 */
203
204
205
206
static struct trace_array	max_tr;

static DEFINE_PER_CPU(struct trace_array_cpu, max_data);

Steven Rostedt's avatar
Steven Rostedt committed
207
/* tracer_enabled is used to toggle activation of a tracer */
208
static int			tracer_enabled = 1;
Steven Rostedt's avatar
Steven Rostedt committed
209

210
211
212
213
214
215
216
217
218
219
220
221
222
/**
 * 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
223
/*
224
225
226
 * 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.
227
228
229
230
231
 *
 * 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
232
 */
233
#define TRACE_BUF_SIZE_DEFAULT	1441792UL /* 16384 * 88 (sizeof(entry)) */
234

235
static unsigned long		trace_buf_size = TRACE_BUF_SIZE_DEFAULT;
236

Steven Rostedt's avatar
Steven Rostedt committed
237
/* trace_types holds a link list of available tracers. */
238
static struct tracer		*trace_types __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
239
240

/* current_trace points to the tracer that is currently active */
241
static struct tracer		*current_trace __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
242
243
244
245
246
247

/*
 * 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.
 */
248
249
static int			max_tracer_type_len;

Steven Rostedt's avatar
Steven Rostedt committed
250
251
252
253
254
255
/*
 * 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.
 */
256
static DEFINE_MUTEX(trace_types_lock);
Steven Rostedt's avatar
Steven Rostedt committed
257
258

/* trace_wait is a waitqueue for tasks blocked on trace_poll */
259
260
static DECLARE_WAIT_QUEUE_HEAD(trace_wait);

261
/* trace_flags holds trace_options default values */
262
unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK |
263
264
	TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO | TRACE_ITER_SLEEP_TIME |
	TRACE_ITER_GRAPH_TIME;
265

Steven Rostedt's avatar
Steven Rostedt committed
266
267
268
269
270
271
/**
 * 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.
 */
272
273
void trace_wake_up(void)
{
Ingo Molnar's avatar
Ingo Molnar committed
274
275
276
277
278
	/*
	 * The runqueue_is_locked() can fail, but this is the best we
	 * have for now:
	 */
	if (!(trace_flags & TRACE_ITER_BLOCK) && !runqueue_is_locked())
279
280
		wake_up(&trace_wait);
}
281

282
static int __init set_buf_size(char *str)
283
{
284
	unsigned long buf_size;
285

286
287
	if (!str)
		return 0;
288
	buf_size = memparse(str, &str);
289
	/* nr_entries can not be zero */
290
	if (buf_size == 0)
291
		return 0;
292
	trace_buf_size = buf_size;
293
294
	return 1;
}
295
__setup("trace_buf_size=", set_buf_size);
296

Steven Rostedt's avatar
Steven Rostedt committed
297
298
299
300
301
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

Steven Rostedt's avatar
Steven Rostedt committed
302
/* These must match the bit postions in trace_iterator_flags */
303
304
305
306
307
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
Ingo Molnar's avatar
Ingo Molnar committed
308
	"raw",
309
	"hex",
Ingo Molnar's avatar
Ingo Molnar committed
310
	"bin",
311
	"block",
Ingo Molnar's avatar
Ingo Molnar committed
312
	"stacktrace",
Ingo Molnar's avatar
Ingo Molnar committed
313
	"sched-tree",
314
	"trace_printk",
315
	"ftrace_preempt",
316
	"branch",
317
	"annotate",
318
	"userstacktrace",
319
	"sym-userobj",
320
	"printk-msg-only",
321
	"context-info",
322
	"latency-format",
323
	"sleep-time",
324
	"graph-time",
325
326
327
	NULL
};

328
329
330
331
332
333
334
335
336
337
static struct {
	u64 (*func)(void);
	const char *name;
} trace_clocks[] = {
	{ trace_clock_local,	"local" },
	{ trace_clock_global,	"global" },
};

int trace_clock_id;

338
339
340
341
342
ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt)
{
	int len;
	int ret;

343
344
345
	if (!cnt)
		return 0;

346
347
348
349
350
351
352
	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);
353
	if (ret == cnt)
354
355
		return -EFAULT;

356
357
	cnt -= ret;

358
	s->readpos += cnt;
359
	return cnt;
360
361
}

362
static ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt)
363
364
365
366
367
368
369
370
371
372
373
374
375
376
{
	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;

377
	s->readpos += cnt;
378
379
380
	return cnt;
}

381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
/*
 * 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.
 *
 * It is also used in other places outside the update_max_tr
 * so it needs to be defined outside of the
 * CONFIG_TRACER_MAX_TRACE.
 */
static raw_spinlock_t ftrace_max_lock =
	(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;

#ifdef CONFIG_TRACER_MAX_TRACE
unsigned long __read_mostly	tracing_max_latency;
unsigned long __read_mostly	tracing_thresh;

/*
 * Copy the new maximum trace into the separate maximum-trace
 * structure. (this way the maximum trace is permanently saved,
 * for later retrieval via /sys/kernel/debug/tracing/latency_trace)
 */
static void
__update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
	struct trace_array_cpu *data = tr->data[cpu];
410
	struct trace_array_cpu *max_data = tr->data[cpu];
411
412
413
414

	max_tr.cpu = cpu;
	max_tr.time_start = data->preempt_timestamp;

415
416
417
418
	max_data = max_tr.data[cpu];
	max_data->saved_latency = tracing_max_latency;
	max_data->critical_start = data->critical_start;
	max_data->critical_end = data->critical_end;
419
420

	memcpy(data->comm, tsk->comm, TASK_COMM_LEN);
421
422
423
424
425
	max_data->pid = tsk->pid;
	max_data->uid = task_uid(tsk);
	max_data->nice = tsk->static_prio - 20 - MAX_RT_PRIO;
	max_data->policy = tsk->policy;
	max_data->rt_priority = tsk->rt_priority;
426
427
428
429
430

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

Steven Rostedt's avatar
Steven Rostedt committed
431
432
433
434
435
436
437
438
439
/**
 * update_max_tr - snapshot all trace buffers from global_trace to max_tr
 * @tr: tracer
 * @tsk: the task with the latency
 * @cpu: The cpu that initiated the trace.
 *
 * Flip the buffers between the @tr and the max_tr and record information
 * about which task was the cause of this latency.
 */
Ingo Molnar's avatar
Ingo Molnar committed
440
void
441
442
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
443
	struct ring_buffer *buf = tr->buffer;
444

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

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

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

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

/**
 * update_max_tr_single - only copy one trace over, and reset the rest
 * @tr - tracer
 * @tsk - task with the latency
 * @cpu - the cpu of the buffer to copy.
Steven Rostedt's avatar
Steven Rostedt committed
464
465
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
466
 */
Ingo Molnar's avatar
Ingo Molnar committed
467
void
468
469
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
470
	int ret;
471

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

475
476
	ftrace_disable_cpu();

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

480
481
	ftrace_enable_cpu();

482
	WARN_ON_ONCE(ret && ret != -EAGAIN);
483
484

	__update_max_tr(tr, tsk, cpu);
485
	__raw_spin_unlock(&ftrace_max_lock);
486
}
487
#endif /* CONFIG_TRACER_MAX_TRACE */
488

Steven Rostedt's avatar
Steven Rostedt committed
489
490
491
492
493
494
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
495
int register_tracer(struct tracer *type)
496
497
__releases(kernel_lock)
__acquires(kernel_lock)
498
499
500
501
502
503
504
505
506
507
{
	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
508
509
510
511
512
513
514
	/*
	 * 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();
515
	mutex_lock(&trace_types_lock);
Ingo Molnar's avatar
Ingo Molnar committed
516

517
518
	tracing_selftest_running = true;

519
520
521
522
523
524
525
526
527
528
	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;
		}
	}

529
530
531
532
533
534
535
	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;
536
537
538
	if (!type->wait_pipe)
		type->wait_pipe = default_wait_pipe;

539

Steven Rostedt's avatar
Steven Rostedt committed
540
#ifdef CONFIG_FTRACE_STARTUP_TEST
541
	if (type->selftest && !tracing_selftest_disabled) {
Steven Rostedt's avatar
Steven Rostedt committed
542
543
544
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
		int i;
545

Steven Rostedt's avatar
Steven Rostedt committed
546
547
548
549
550
551
552
		/*
		 * 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
553
		for_each_tracing_cpu(i)
554
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
555

Steven Rostedt's avatar
Steven Rostedt committed
556
557
558
559
560
561
562
563
564
565
		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
566
		/* Only reset on passing, to avoid touching corrupted buffers */
Ingo Molnar's avatar
Ingo Molnar committed
567
		for_each_tracing_cpu(i)
568
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
569

Steven Rostedt's avatar
Steven Rostedt committed
570
571
572
573
		printk(KERN_CONT "PASSED\n");
	}
#endif

574
575
576
577
578
	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
579

580
 out:
581
	tracing_selftest_running = false;
582
583
	mutex_unlock(&trace_types_lock);

Steven Rostedt's avatar
Steven Rostedt committed
584
585
586
587
588
589
590
591
592
593
594
595
	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;
596
#ifdef CONFIG_FTRACE_STARTUP_TEST
Steven Rostedt's avatar
Steven Rostedt committed
597
598
	printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n",
	       type->name);
599
600
#endif

Steven Rostedt's avatar
Steven Rostedt committed
601
 out_unlock:
602
	lock_kernel();
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
	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;
621
622
623
624
625
626
627
628
629

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

630
631
632
633
634
635
636
637
638
639
640
641
642
	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);
}

643
void tracing_reset(struct trace_array *tr, int cpu)
644
{
645
	ftrace_disable_cpu();
646
	ring_buffer_reset_cpu(tr->buffer, cpu);
647
	ftrace_enable_cpu();
648
649
}

650
651
652
653
654
655
656
657
658
659
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);
}

660
661
662
663
664
665
666
667
668
669
void tracing_reset_current(int cpu)
{
	tracing_reset(&global_trace, cpu);
}

void tracing_reset_current_online_cpus(void)
{
	tracing_reset_online_cpus(&global_trace);
}

670
#define SAVED_CMDLINES 128
671
#define NO_CMDLINE_MAP UINT_MAX
672
673
674
675
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;
676
static raw_spinlock_t trace_cmdline_lock = __RAW_SPIN_LOCK_UNLOCKED;
677
678

/* temporary disable recording */
679
static atomic_t trace_record_cmdline_disabled __read_mostly;
680
681
682

static void trace_init_cmdlines(void)
{
683
684
	memset(&map_pid_to_cmdline, NO_CMDLINE_MAP, sizeof(map_pid_to_cmdline));
	memset(&map_cmdline_to_pid, NO_CMDLINE_MAP, sizeof(map_cmdline_to_pid));
685
686
687
	cmdline_idx = 0;
}

688
689
690
static int trace_stop_count;
static DEFINE_SPINLOCK(tracing_start_lock);

691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
/**
 * 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();
}

706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
/**
 * 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);
721
722
723
724
725
726
	if (--trace_stop_count) {
		if (trace_stop_count < 0) {
			/* Someone screwed up their debugging */
			WARN_ON_ONCE(1);
			trace_stop_count = 0;
		}
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
		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
772
void trace_stop_cmdline_recording(void);
773

Ingo Molnar's avatar
Ingo Molnar committed
774
static void trace_save_cmdline(struct task_struct *tsk)
775
{
776
	unsigned pid, idx;
777
778
779
780
781
782
783
784
785
786

	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.
	 */
787
	if (!__raw_spin_trylock(&trace_cmdline_lock))
788
789
790
		return;

	idx = map_pid_to_cmdline[tsk->pid];
791
	if (idx == NO_CMDLINE_MAP) {
792
793
		idx = (cmdline_idx + 1) % SAVED_CMDLINES;

794
795
796
797
798
799
800
801
802
		/*
		 * Check whether the cmdline buffer at idx has a pid
		 * mapped. We are going to overwrite that entry so we
		 * need to clear the map_pid_to_cmdline. Otherwise we
		 * would read the new comm for the old pid.
		 */
		pid = map_cmdline_to_pid[idx];
		if (pid != NO_CMDLINE_MAP)
			map_pid_to_cmdline[pid] = NO_CMDLINE_MAP;
803

804
		map_cmdline_to_pid[idx] = tsk->pid;
805
806
807
808
809
810
811
		map_pid_to_cmdline[tsk->pid] = idx;

		cmdline_idx = idx;
	}

	memcpy(&saved_cmdlines[idx], tsk->comm, TASK_COMM_LEN);

812
	__raw_spin_unlock(&trace_cmdline_lock);
813
814
}

815
void trace_find_cmdline(int pid, char comm[])
816
817
818
{
	unsigned map;

819
820
821
822
	if (!pid) {
		strcpy(comm, "<idle>");
		return;
	}
823

824
825
826
827
	if (pid > PID_MAX_DEFAULT) {
		strcpy(comm, "<...>");
		return;
	}
828

829
	preempt_disable();
830
	__raw_spin_lock(&trace_cmdline_lock);
831
	map = map_pid_to_cmdline[pid];
832
833
834
835
	if (map != NO_CMDLINE_MAP)
		strcpy(comm, saved_cmdlines[map]);
	else
		strcpy(comm, "<...>");
836

837
	__raw_spin_unlock(&trace_cmdline_lock);
838
	preempt_enable();
839
840
}

Ingo Molnar's avatar
Ingo Molnar committed
841
void tracing_record_cmdline(struct task_struct *tsk)
842
{
843
844
	if (atomic_read(&trace_record_cmdline_disabled) || !tracer_enabled ||
	    !tracing_is_on())
845
846
847
848
849
		return;

	trace_save_cmdline(tsk);
}

850
void
851
852
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
853
854
855
{
	struct task_struct *tsk = current;

856
857
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
858
	entry->tgid			= (tsk) ? tsk->tgid : 0;
859
	entry->flags =
860
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
861
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
862
863
864
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
865
866
867
868
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}
869
EXPORT_SYMBOL_GPL(tracing_generic_entry_update);
870

871
struct ring_buffer_event *trace_buffer_lock_reserve(struct trace_array *tr,
872
						    int type,
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
						    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;
}

889
890
891
892
static inline void __trace_buffer_unlock_commit(struct trace_array *tr,
					struct ring_buffer_event *event,
					unsigned long flags, int pc,
					int wake)
893
894
895
896
897
{
	ring_buffer_unlock_commit(tr->buffer, event);

	ftrace_trace_stack(tr, flags, 6, pc);
	ftrace_trace_userstack(tr, flags, pc);
898
899
900
901
902
903
904
905
906
907

	if (wake)
		trace_wake_up();
}

void trace_buffer_unlock_commit(struct trace_array *tr,
					struct ring_buffer_event *event,
					unsigned long flags, int pc)
{
	__trace_buffer_unlock_commit(tr, event, flags, pc, 1);
908
909
}

910
struct ring_buffer_event *
911
trace_current_buffer_lock_reserve(int type, unsigned long len,
912
913
914
915
916
				  unsigned long flags, int pc)
{
	return trace_buffer_lock_reserve(&global_trace,
					 type, len, flags, pc);
}
917
EXPORT_SYMBOL_GPL(trace_current_buffer_lock_reserve);
918
919
920
921

void trace_current_buffer_unlock_commit(struct ring_buffer_event *event,
					unsigned long flags, int pc)
{
922
	__trace_buffer_unlock_commit(&global_trace, event, flags, pc, 1);
923
}
924
EXPORT_SYMBOL_GPL(trace_current_buffer_unlock_commit);
925
926
927
928

void trace_nowake_buffer_unlock_commit(struct ring_buffer_event *event,
					unsigned long flags, int pc)
{
929
930
	__trace_buffer_unlock_commit(&global_trace, event, flags, pc, 0);
}
931
EXPORT_SYMBOL_GPL(trace_nowake_buffer_unlock_commit);
932
933
934
935

void trace_current_buffer_discard_commit(struct ring_buffer_event *event)
{
	ring_buffer_discard_commit(global_trace.buffer, event);
936
}
937
EXPORT_SYMBOL_GPL(trace_current_buffer_discard_commit);
938

Ingo Molnar's avatar
Ingo Molnar committed
939
void
940
trace_function(struct trace_array *tr,
941
942
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
943
{
944
	struct ftrace_event_call *call = &event_function;
945
	struct ring_buffer_event *event;
946
	struct ftrace_entry *entry;
947

948
949
950
951
	/* If we are reading the ring buffer, don't trace */
	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

952
953
	event = trace_buffer_lock_reserve(tr, TRACE_FN, sizeof(*entry),
					  flags, pc);
954
955
956
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
957
958
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
959

960
961
	if (!filter_check_discard(call, entry, tr->buffer, event))
		ring_buffer_unlock_commit(tr->buffer, event);
962
963
}

Ingo Molnar's avatar
Ingo Molnar committed
964
void
Ingo Molnar's avatar
Ingo Molnar committed
965
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
966
967
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
Ingo Molnar's avatar
Ingo Molnar committed
968
969
{
	if (likely(!atomic_read(&data->disabled)))
970
		trace_function(tr, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
971
972
}

973
#ifdef CONFIG_STACKTRACE
974
975
976
static void __ftrace_trace_stack(struct trace_array *tr,
				 unsigned long flags,
				 int skip, int pc)
Ingo Molnar's avatar
Ingo Molnar committed
977
{
978
	struct ftrace_event_call *call = &event_kernel_stack;
979
	struct ring_buffer_event *event;
980
	struct stack_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
981
982
	struct stack_trace trace;

983
984
	event = trace_buffer_lock_reserve(tr, TRACE_STACK,
					  sizeof(*entry), flags, pc);
985
986
987
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
988
	memset(&entry->caller, 0, sizeof(entry->caller));
Ingo Molnar's avatar
Ingo Molnar committed
989
990
991
992

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
993
	trace.entries		= entry->caller;
Ingo Molnar's avatar
Ingo Molnar committed
994
995

	save_stack_trace(&trace);
996
997
	if (!filter_check_discard(call, entry, tr->buffer, event))
		ring_buffer_unlock_commit(tr->buffer, event);
Ingo Molnar's avatar
Ingo Molnar committed
998
999
}

1000
1001
void ftrace_trace_stack(struct trace_array *tr, unsigned long flags, int skip,
			int pc)
1002
1003
1004
1005
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

1006
	__ftrace_trace_stack(tr, flags, skip, pc);
1007
1008
}

1009
1010
void __trace_stack(struct trace_array *tr, unsigned long flags, int skip,
		   int pc)
1011
{
1012
	__ftrace_trace_stack(tr, flags, skip, pc);
1013
1014
}

1015
void ftrace_trace_userstack(struct trace_array *tr, unsigned long flags, int pc)
1016
{
1017
	struct ftrace_event_call *call = &event_user_stack;
1018
	struct ring_buffer_event *event;
1019
1020
1021
1022
1023
1024
	struct userstack_entry *entry;
	struct stack_trace trace;

	if (!(trace_flags & TRACE_ITER_USERSTACKTRACE))
		return;

1025
1026
	event = trace_buffer_lock_reserve(tr, TRACE_USER_STACK,
					  sizeof(*entry), flags, pc);
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
	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);
1039
1040
	if (!filter_check_discard(call, entry, tr->buffer, event))
		ring_buffer_unlock_commit(tr->buffer, event);
1041
1042
}

1043
1044
#ifdef UNUSED
static void __trace_userstack(struct trace_array *tr, unsigned long flags)
1045
{
1046
	ftrace_trace_userstack(tr, flags, preempt_count());
1047
}
1048
#endif /* UNUSED */
1049

1050
1051
#endif /* CONFIG_STACKTRACE */

1052
static void
1053
ftrace_trace_special(void *__tr,
1054
1055
		     unsigned long arg1, unsigned long arg2, unsigned long arg3,
		     int pc)
Ingo Molnar's avatar
Ingo Molnar committed
1056
{
1057
	struct ring_buffer_event *event;
Ingo Molnar's avatar
Ingo Molnar committed
1058
	struct trace_array *tr = __tr;
1059
	struct special_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
1060

1061
1062
	event = trace_buffer_lock_reserve(tr, TRACE_SPECIAL,
					  sizeof(*entry), 0, pc);
1063
1064
1065
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1066
1067
1068
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
1069
	trace_buffer_unlock_commit(tr, event, 0, pc);
Ingo Molnar's avatar
Ingo Molnar committed
1070
1071
}

1072
1073
1074
1075
void
__trace_special(void *__tr, void *__data,
		unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
1076
	ftrace_trace_special(__tr, arg1, arg2, arg3, preempt_count());
1077
1078
}

1079
1080
1081
1082
1083
void
ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1084
	unsigned long flags;
1085
	int cpu;
1086
	int pc;
1087

1088
	if (tracing_disabled)
1089
1090
		return;

1091
	pc = preempt_count();
1092
	local_irq_save(flags);
1093
1094
1095
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

1096
	if (likely(atomic_inc_return(&data->disabled) == 1))
1097
		ftrace_trace_special(tr, arg1, arg2, arg3, pc);
1098

1099
1100
	atomic_dec(&data->disabled);
	local_irq_restore(flags);
1101
1102
}

1103
/**
1104
 * trace_vbprintk - write binary msg to tracing buffer
1105
1106
 *
 */
1107
int trace_vbprintk(unsigned long ip, const char *fmt, va_list args)
1108
{
1109
1110
	static raw_spinlock_t trace_buf_lock =
		(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
1111
1112
	static u32 trace_buf[TRACE_BUF_SIZE];

1113
	struct ftrace_event_call *call = &event_bprint;
1114
1115
1116
	struct ring_buffer_event *event;
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1117
	struct bprint_entry *entry;
1118
	unsigned long flags;
1119
	int disable;
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
	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];

1134
1135
	disable = atomic_inc_return(&data->disabled);
	if (unlikely(disable != 1))
1136
1137
		goto out;

1138
1139
1140
	/* Lockdep uses trace_printk for lock tracing */
	local_irq_save(flags);
	__raw_spin_lock(&trace_buf_lock);
1141
1142
1143
1144
1145
1146
	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;
1147
	event = trace_buffer_lock_reserve(tr, TRACE_BPRINT, size, flags, pc);
1148
1149
1150
1151
1152
1153
1154
	if (!event)
		goto out_unlock;
	entry = ring_buffer_event_data(event);
	entry->ip			= ip;
	entry->fmt			= fmt;

	memcpy(entry->buf, trace_buf, sizeof(u32) * len);
1155
1156
	if (!filter_check_discard(call, entry, tr->buffer, event))
		ring_buffer_unlock_commit(tr->buffer, event);
1157
1158

out_unlock:
1159
1160
	__raw_spin_unlock(&trace_buf_lock);
	local_irq_restore(flags);
1161
1162

out:
1163
	atomic_dec_return(&data->disabled);
1164
1165
1166
1167
1168
	ftrace_preempt_enable(resched);
	unpause_graph_tracing();

	return len;
}
1169
1170
EXPORT_SYMBOL_GPL(trace_vbprintk);

1171
int trace_vprintk(unsigned long ip, const char *fmt, va_list args)