trace.c 94 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
49
50
51
52
/*
 * 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.
 */
static int ring_buffer_expanded;

53
54
/*
 * We need to change this state when a selftest is running.
55
56
 * A selftest will lurk into the ring-buffer to count the
 * entries inserted during the selftest although some concurrent
57
 * insertions into the ring-buffer such as trace_printk could occurred
58
59
 * at the same time, giving false positive or negative results.
 */
60
static bool __read_mostly tracing_selftest_running;
61

62
63
64
65
66
/*
 * If a tracer is running, we do not want to run SELFTEST.
 */
static bool __read_mostly tracing_selftest_disabled;

67
68
69
70
71
72
73
74
75
76
77
78
79
80
/* 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;
}
81
82
83
84
85
86
87

/*
 * 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.
 */
88
static int tracing_disabled = 1;
89

90
91
92
93
94
95
96
97
98
99
100
101
102
103
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();
}

104
static cpumask_var_t __read_mostly	tracing_buffer_mask;
105

106
107
108
/* Define which cpu buffers are currently read in trace_pipe */
static cpumask_var_t			tracing_reader_cpumask;

109
#define for_each_tracing_cpu(cpu)	\
110
	for_each_cpu(cpu, tracing_buffer_mask)
111

112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
/*
 * 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;

127
128
129
130
131
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;
132
133
134

static int __init set_ftrace(char *str)
{
135
136
	strncpy(bootup_tracer_buf, str, BOOTUP_TRACER_SIZE);
	default_bootup_tracer = bootup_tracer_buf;
137
138
	/* We are using ftrace early, expand it */
	ring_buffer_expanded = 1;
139
140
	return 1;
}
141
__setup("ftrace=", set_ftrace);
142

143
144
145
146
147
148
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
149

150
long
151
152
153
154
155
156
157
ns2usecs(cycle_t nsec)
{
	nsec += 500;
	do_div(nsec, 1000);
	return nsec;
}

Steven Rostedt's avatar
Steven Rostedt committed
158
159
160
161
162
163
164
165
166
167
168
169
/*
 * 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.
 */
170
171
172
173
static struct trace_array	global_trace;

static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu);

174
175
176
177
178
179
180
181
182
183
184
185
186
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;
}
187

Steven Rostedt's avatar
Steven Rostedt committed
188
189
190
191
192
193
194
195
196
197
/*
 * 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.
 */
198
199
200
201
static struct trace_array	max_tr;

static DEFINE_PER_CPU(struct trace_array_cpu, max_data);

Steven Rostedt's avatar
Steven Rostedt committed
202
/* tracer_enabled is used to toggle activation of a tracer */
203
static int			tracer_enabled = 1;
Steven Rostedt's avatar
Steven Rostedt committed
204

205
206
207
208
209
210
211
212
213
214
215
216
217
/**
 * 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
218
/*
219
220
221
 * 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.
222
223
224
225
226
 *
 * 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
227
 */
228
#define TRACE_BUF_SIZE_DEFAULT	1441792UL /* 16384 * 88 (sizeof(entry)) */
229

230
static unsigned long		trace_buf_size = TRACE_BUF_SIZE_DEFAULT;
231

Steven Rostedt's avatar
Steven Rostedt committed
232
/* trace_types holds a link list of available tracers. */
233
static struct tracer		*trace_types __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
234
235

/* current_trace points to the tracer that is currently active */
236
static struct tracer		*current_trace __read_mostly;
Steven Rostedt's avatar
Steven Rostedt committed
237
238
239
240
241
242

/*
 * 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.
 */
243
244
static int			max_tracer_type_len;

Steven Rostedt's avatar
Steven Rostedt committed
245
246
247
248
249
250
/*
 * 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.
 */
251
static DEFINE_MUTEX(trace_types_lock);
Steven Rostedt's avatar
Steven Rostedt committed
252
253

/* trace_wait is a waitqueue for tasks blocked on trace_poll */
254
255
static DECLARE_WAIT_QUEUE_HEAD(trace_wait);

256
/* trace_flags holds trace_options default values */
257
unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK |
258
	TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO;
259

Steven Rostedt's avatar
Steven Rostedt committed
260
261
262
263
264
265
/**
 * 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.
 */
266
267
void trace_wake_up(void)
{
Ingo Molnar's avatar
Ingo Molnar committed
268
269
270
271
272
	/*
	 * The runqueue_is_locked() can fail, but this is the best we
	 * have for now:
	 */
	if (!(trace_flags & TRACE_ITER_BLOCK) && !runqueue_is_locked())
273
274
		wake_up(&trace_wait);
}
275

276
static int __init set_buf_size(char *str)
277
{
278
	unsigned long buf_size;
279
280
	int ret;

281
282
	if (!str)
		return 0;
283
	ret = strict_strtoul(str, 0, &buf_size);
284
	/* nr_entries can not be zero */
285
	if (ret < 0 || buf_size == 0)
286
		return 0;
287
	trace_buf_size = buf_size;
288
289
	return 1;
}
290
__setup("trace_buf_size=", set_buf_size);
291

Steven Rostedt's avatar
Steven Rostedt committed
292
293
294
295
296
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

Steven Rostedt's avatar
Steven Rostedt committed
297
/* These must match the bit postions in trace_iterator_flags */
298
299
300
301
302
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
Ingo Molnar's avatar
Ingo Molnar committed
303
	"raw",
304
	"hex",
Ingo Molnar's avatar
Ingo Molnar committed
305
	"bin",
306
	"block",
Ingo Molnar's avatar
Ingo Molnar committed
307
	"stacktrace",
Ingo Molnar's avatar
Ingo Molnar committed
308
	"sched-tree",
309
	"trace_printk",
310
	"ftrace_preempt",
311
	"branch",
312
	"annotate",
313
	"userstacktrace",
314
	"sym-userobj",
315
	"printk-msg-only",
316
	"context-info",
317
	"latency-format",
318
	"global-clock",
319
320
321
	NULL
};

Steven Rostedt's avatar
Steven Rostedt committed
322
323
324
325
326
327
328
329
330
/*
 * 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.
 */
331
332
static raw_spinlock_t ftrace_max_lock =
	(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
333
334
335
336
337
338

/*
 * 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
339
static void
340
341
342
343
344
345
346
347
348
349
350
351
__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;
352
	data->uid = task_uid(tsk);
353
354
355
356
357
	data->nice = tsk->static_prio - 20 - MAX_RT_PRIO;
	data->policy = tsk->policy;
	data->rt_priority = tsk->rt_priority;

	/* record this tasks comm */
358
	tracing_record_cmdline(tsk);
359
360
}

361
362
363
364
365
ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt)
{
	int len;
	int ret;

366
367
368
	if (!cnt)
		return 0;

369
370
371
372
373
374
375
	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);
376
	if (ret == cnt)
377
378
		return -EFAULT;

379
380
	cnt -= ret;

381
	s->readpos += cnt;
382
	return cnt;
383
384
}

385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
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;

400
	s->readpos += cnt;
401
402
403
	return cnt;
}

Ingo Molnar's avatar
Ingo Molnar committed
404
static void
405
406
407
408
409
410
411
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);

412
	trace_seq_init(s);
413
414
}

Steven Rostedt's avatar
Steven Rostedt committed
415
416
417
418
419
420
421
422
423
/**
 * 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
424
void
425
426
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
427
	struct ring_buffer *buf = tr->buffer;
428

429
	WARN_ON_ONCE(!irqs_disabled());
430
	__raw_spin_lock(&ftrace_max_lock);
431
432
433
434

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

435
	ftrace_disable_cpu();
436
	ring_buffer_reset(tr->buffer);
437
	ftrace_enable_cpu();
438
439

	__update_max_tr(tr, tsk, cpu);
440
	__raw_spin_unlock(&ftrace_max_lock);
441
442
443
444
445
446
447
}

/**
 * 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
448
449
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
450
 */
Ingo Molnar's avatar
Ingo Molnar committed
451
void
452
453
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
454
	int ret;
455

456
	WARN_ON_ONCE(!irqs_disabled());
457
	__raw_spin_lock(&ftrace_max_lock);
458

459
460
	ftrace_disable_cpu();

461
462
463
	ring_buffer_reset(max_tr.buffer);
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

464
465
	ftrace_enable_cpu();

466
	WARN_ON_ONCE(ret && ret != -EAGAIN);
467
468

	__update_max_tr(tr, tsk, cpu);
469
	__raw_spin_unlock(&ftrace_max_lock);
470
471
}

Steven Rostedt's avatar
Steven Rostedt committed
472
473
474
475
476
477
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
478
int register_tracer(struct tracer *type)
479
480
__releases(kernel_lock)
__acquires(kernel_lock)
481
482
483
484
485
486
487
488
489
490
{
	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
491
492
493
494
495
496
497
	/*
	 * 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();
498
	mutex_lock(&trace_types_lock);
Ingo Molnar's avatar
Ingo Molnar committed
499

500
501
	tracing_selftest_running = true;

502
503
504
505
506
507
508
509
510
511
	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;
		}
	}

512
513
514
515
516
517
518
	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;
519
520
521
	if (!type->wait_pipe)
		type->wait_pipe = default_wait_pipe;

522

Steven Rostedt's avatar
Steven Rostedt committed
523
#ifdef CONFIG_FTRACE_STARTUP_TEST
524
	if (type->selftest && !tracing_selftest_disabled) {
Steven Rostedt's avatar
Steven Rostedt committed
525
526
527
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
		int i;
528

Steven Rostedt's avatar
Steven Rostedt committed
529
530
531
532
533
534
535
		/*
		 * 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
536
		for_each_tracing_cpu(i)
537
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
538

Steven Rostedt's avatar
Steven Rostedt committed
539
540
541
542
543
544
545
546
547
548
		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
549
		/* Only reset on passing, to avoid touching corrupted buffers */
Ingo Molnar's avatar
Ingo Molnar committed
550
		for_each_tracing_cpu(i)
551
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
552

Steven Rostedt's avatar
Steven Rostedt committed
553
554
555
556
		printk(KERN_CONT "PASSED\n");
	}
#endif

557
558
559
560
561
	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
562

563
 out:
564
	tracing_selftest_running = false;
565
566
	mutex_unlock(&trace_types_lock);

Steven Rostedt's avatar
Steven Rostedt committed
567
568
569
570
571
572
573
574
575
576
577
578
	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;
579
#ifdef CONFIG_FTRACE_STARTUP_TEST
Steven Rostedt's avatar
Steven Rostedt committed
580
581
	printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n",
	       type->name);
582
583
#endif

Steven Rostedt's avatar
Steven Rostedt committed
584
 out_unlock:
585
	lock_kernel();
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
	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;
604
605
606
607
608
609
610
611
612

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

613
614
615
616
617
618
619
620
621
622
623
624
625
	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);
}

626
void tracing_reset(struct trace_array *tr, int cpu)
627
{
628
	ftrace_disable_cpu();
629
	ring_buffer_reset_cpu(tr->buffer, cpu);
630
	ftrace_enable_cpu();
631
632
}

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

643
#define SAVED_CMDLINES 128
644
#define NO_CMDLINE_MAP UINT_MAX
645
646
647
648
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;
649
static raw_spinlock_t trace_cmdline_lock = __RAW_SPIN_LOCK_UNLOCKED;
650
651

/* temporary disable recording */
652
static atomic_t trace_record_cmdline_disabled __read_mostly;
653
654
655

static void trace_init_cmdlines(void)
{
656
657
	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));
658
659
660
	cmdline_idx = 0;
}

661
662
663
static int trace_stop_count;
static DEFINE_SPINLOCK(tracing_start_lock);

664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
/**
 * 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();
}

679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
/**
 * 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);
694
695
696
697
698
699
	if (--trace_stop_count) {
		if (trace_stop_count < 0) {
			/* Someone screwed up their debugging */
			WARN_ON_ONCE(1);
			trace_stop_count = 0;
		}
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
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
		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
745
void trace_stop_cmdline_recording(void);
746

Ingo Molnar's avatar
Ingo Molnar committed
747
static void trace_save_cmdline(struct task_struct *tsk)
748
{
749
	unsigned pid, idx;
750
751
752
753
754
755
756
757
758
759

	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.
	 */
760
	if (!__raw_spin_trylock(&trace_cmdline_lock))
761
762
763
		return;

	idx = map_pid_to_cmdline[tsk->pid];
764
	if (idx == NO_CMDLINE_MAP) {
765
766
		idx = (cmdline_idx + 1) % SAVED_CMDLINES;

767
768
769
770
771
772
773
774
775
		/*
		 * 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;
776

777
		map_cmdline_to_pid[idx] = tsk->pid;
778
779
780
781
782
783
784
		map_pid_to_cmdline[tsk->pid] = idx;

		cmdline_idx = idx;
	}

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

785
	__raw_spin_unlock(&trace_cmdline_lock);
786
787
}

788
void trace_find_cmdline(int pid, char comm[])
789
790
791
{
	unsigned map;

792
793
794
795
	if (!pid) {
		strcpy(comm, "<idle>");
		return;
	}
796

797
798
799
800
	if (pid > PID_MAX_DEFAULT) {
		strcpy(comm, "<...>");
		return;
	}
801

802
	__raw_spin_lock(&trace_cmdline_lock);
803
	map = map_pid_to_cmdline[pid];
804
805
806
807
	if (map != NO_CMDLINE_MAP)
		strcpy(comm, saved_cmdlines[map]);
	else
		strcpy(comm, "<...>");
808

809
	__raw_spin_unlock(&trace_cmdline_lock);
810
811
}

Ingo Molnar's avatar
Ingo Molnar committed
812
void tracing_record_cmdline(struct task_struct *tsk)
813
{
814
815
	if (atomic_read(&trace_record_cmdline_disabled) || !tracer_enabled ||
	    !tracing_is_on())
816
817
818
819
820
		return;

	trace_save_cmdline(tsk);
}

821
void
822
823
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
824
825
826
{
	struct task_struct *tsk = current;

827
828
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
829
	entry->tgid			= (tsk) ? tsk->tgid : 0;
830
	entry->flags =
831
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
832
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
833
834
835
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
836
837
838
839
840
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}

841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
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();
}

874
875
876
877
878
879
880
881
882
883
884
885
886
887
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
888
void
889
trace_function(struct trace_array *tr,
890
891
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
892
{
893
	struct ring_buffer_event *event;
894
	struct ftrace_entry *entry;
895

896
897
898
899
	/* If we are reading the ring buffer, don't trace */
	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

900
901
	event = trace_buffer_lock_reserve(tr, TRACE_FN, sizeof(*entry),
					  flags, pc);
902
903
904
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
905
906
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
907
	ring_buffer_unlock_commit(tr->buffer, event);
908
909
}

910
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
911
912
913
914
915
916
917
918
919
920
921
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;

922
923
	event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_ENT,
					  sizeof(*entry), flags, pc);
924
925
926
927
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
	entry->graph_ent			= *trace;
928
	ring_buffer_unlock_commit(global_trace.buffer, event);
929
930
931
}

static void __trace_graph_return(struct trace_array *tr,
932
				struct ftrace_graph_ret *trace,
933
934
935
936
				unsigned long flags,
				int pc)
{
	struct ring_buffer_event *event;
937
	struct ftrace_graph_ret_entry *entry;
938
939
940
941

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

942
943
	event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_RET,
					  sizeof(*entry), flags, pc);
944
945
946
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
947
	entry->ret				= *trace;
948
	ring_buffer_unlock_commit(global_trace.buffer, event);
949
950
951
}
#endif

Ingo Molnar's avatar
Ingo Molnar committed
952
void
Ingo Molnar's avatar
Ingo Molnar committed
953
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
954
955
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
Ingo Molnar's avatar
Ingo Molnar committed
956
957
{
	if (likely(!atomic_read(&data->disabled)))
958
		trace_function(tr, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
959
960
}

961
962
963
static void __ftrace_trace_stack(struct trace_array *tr,
				 unsigned long flags,
				 int skip, int pc)
Ingo Molnar's avatar
Ingo Molnar committed
964
{
965
#ifdef CONFIG_STACKTRACE
966
	struct ring_buffer_event *event;
967
	struct stack_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
968
969
	struct stack_trace trace;

970
971
	event = trace_buffer_lock_reserve(tr, TRACE_STACK,
					  sizeof(*entry), flags, pc);
972
973
974
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
975
	memset(&entry->caller, 0, sizeof(entry->caller));
Ingo Molnar's avatar
Ingo Molnar committed
976
977
978
979

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
980
	trace.entries		= entry->caller;
Ingo Molnar's avatar
Ingo Molnar committed
981
982

	save_stack_trace(&trace);
983
	ring_buffer_unlock_commit(tr->buffer, event);
984
#endif
Ingo Molnar's avatar
Ingo Molnar committed
985
986
}

987
988
989
990
991
992
993
static void ftrace_trace_stack(struct trace_array *tr,
			       unsigned long flags,
			       int skip, int pc)
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

994
	__ftrace_trace_stack(tr, flags, skip, pc);
995
996
}

997
998
void __trace_stack(struct trace_array *tr,
		   unsigned long flags,
999
		   int skip, int pc)
1000
{
1001
	__ftrace_trace_stack(tr, flags, skip, pc);
1002
1003
}

1004
static void ftrace_trace_userstack(struct trace_array *tr,
1005
				   unsigned long flags, int pc)
1006
{
1007
#ifdef CONFIG_STACKTRACE
1008
	struct ring_buffer_event *event;
1009
1010
1011
1012
1013
1014
	struct userstack_entry *entry;
	struct stack_trace trace;

	if (!(trace_flags & TRACE_ITER_USERSTACKTRACE))
		return;

1015
1016
	event = trace_buffer_lock_reserve(tr, TRACE_USER_STACK,
					  sizeof(*entry), flags, pc);
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
	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);
1029
	ring_buffer_unlock_commit(tr->buffer, event);
1030
#endif
1031
1032
}

1033
1034
#ifdef UNUSED
static void __trace_userstack(struct trace_array *tr, unsigned long flags)
1035
{
1036
	ftrace_trace_userstack(tr, flags, preempt_count());
1037
}
1038
#endif /* UNUSED */
1039

1040
static void
1041
ftrace_trace_special(void *__tr,
1042
1043
		     unsigned long arg1, unsigned long arg2, unsigned long arg3,
		     int pc)
Ingo Molnar's avatar
Ingo Molnar committed
1044
{
1045
	struct ring_buffer_event *event;
Ingo Molnar's avatar
Ingo Molnar committed
1046
	struct trace_array *tr = __tr;
1047
	struct special_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
1048

1049
1050
	event = trace_buffer_lock_reserve(tr, TRACE_SPECIAL,
					  sizeof(*entry), 0, pc);
1051
1052
1053
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1054
1055
1056
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
1057
	trace_buffer_unlock_commit(tr, event, 0, pc);
Ingo Molnar's avatar
Ingo Molnar committed
1058
1059
}

1060
1061
1062
1063
void
__trace_special(void *__tr, void *__data,
		unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
1064
	ftrace_trace_special(__tr, arg1, arg2, arg3, preempt_count());
1065
1066
}

Ingo Molnar's avatar
Ingo Molnar committed
1067
void
1068
tracing_sched_switch_trace(struct trace_array *tr,
Ingo Molnar's avatar
Ingo Molnar committed
1069
1070
			   struct task_struct *prev,
			   struct task_struct *next,
1071
			   unsigned long flags, int pc)
1072
{
1073
	struct ring_buffer_event *event;
1074
	struct ctx_switch_entry *entry;
1075

1076
1077
	event = trace_buffer_lock_reserve(tr, TRACE_CTX,
					  sizeof(*entry), flags, pc);
1078
1079
1080
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1081
1082
1083
1084
1085
1086
1087
	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);
1088
	trace_buffer_unlock_commit(tr, event, flags, pc);
1089
1090
}

1091
1092
void
tracing_sched_wakeup_trace(struct trace_array *tr,
Ingo Molnar's avatar
Ingo Molnar committed
1093
1094
			   struct task_struct *wakee,
			   struct task_struct *curr,
1095
			   unsigned long flags, int pc)
1096
{
1097
	struct ring_buffer_event *event;
1098
	struct ctx_switch_entry *entry;
1099

1100
1101
	event = trace_buffer_lock_reserve(tr, TRACE_WAKE,
					  sizeof(*entry), flags, pc);
1102
1103
1104
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1105
1106
1107
1108
1109
1110
1111
	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);
1112
1113
1114
1115

	ring_buffer_unlock_commit(tr->buffer, event);
	ftrace_trace_stack(tr, flags, 6, pc);
	ftrace_trace_userstack(tr, flags, pc);
1116
1117
}

1118
1119
1120
1121
1122
void
ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1123
	unsigned long flags;
1124
	int cpu;
1125
	int pc;
1126

1127
	if (tracing_disabled)
1128
1129
		return;

1130
	pc = preempt_count();
1131
	local_irq_save(flags);
1132
1133
1134
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

1135
	if (likely(atomic_inc_return(&data->disabled) == 1))
1136
		ftrace_trace_special(tr, arg1, arg2, arg3, pc);
1137

1138
1139
	atomic_dec(&data->disabled);
	local_irq_restore(flags);
1140
1141
}

1142
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
1143
int trace_graph_entry(struct ftrace_graph_ent *trace)
1144
1145
1146
1147
1148
1149
1150
1151
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	unsigned long flags;
	long disabled;
	int cpu;
	int pc;

1152
1153
1154
	if (!ftrace_trace_task(current))
		return 0;

1155
1156
1157
	if (!ftrace_graph_addr(trace->func))
		return 0;

1158
	local_irq_save(flags);
1159
1160
1161
1162
1163
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);
	if (likely(disabled == 1)) {
		pc = preempt_count();
1164
		__trace_graph_entry(tr, trace, flags, pc);
1165
	}
1166
1167
1168
	/* Only do the atomic if it is not already set */
	if (!test_tsk_trace_graph(current))
		set_tsk_trace_graph(current);
1169
	atomic_dec(&data->disabled);
1170
	local_irq_restore(flags);
1171
1172

	return 1;
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
}

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;

1184
	local_irq_save(flags);
1185
1186
1187
1188
1189
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);
	if (likely(disabled == 1)) {
		pc = preempt_count();