trace.c 93.9 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
/*
 * ring buffer based function tracer
 *
 * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
 * Copyright (C) 2008 Ingo Molnar <mingo@redhat.com>
 *
 * Originally taken from the RT patch by:
 *    Arnaldo Carvalho de Melo <acme@redhat.com>
 *
 * Based on code from the latency_tracer, that is:
 *  Copyright (C) 2004-2006 Ingo Molnar
 *  Copyright (C) 2004 William Lee Irwin III
 */
14
#include <linux/ring_buffer.h>
15
#include <linux/utsrelease.h>
16
17
#include <linux/stacktrace.h>
#include <linux/writeback.h>
18
19
#include <linux/kallsyms.h>
#include <linux/seq_file.h>
20
#include <linux/notifier.h>
21
#include <linux/irqflags.h>
22
#include <linux/debugfs.h>
23
#include <linux/pagemap.h>
24
25
26
#include <linux/hardirq.h>
#include <linux/linkage.h>
#include <linux/uaccess.h>
27
#include <linux/kprobes.h>
28
29
30
#include <linux/ftrace.h>
#include <linux/module.h>
#include <linux/percpu.h>
31
#include <linux/splice.h>
32
#include <linux/kdebug.h>
33
34
#include <linux/ctype.h>
#include <linux/init.h>
35
#include <linux/poll.h>
36
37
#include <linux/gfp.h>
#include <linux/fs.h>
Ingo Molnar's avatar
Ingo Molnar committed
38

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

42
43
#define TRACE_BUFFER_FLAGS	(RB_FL_OVERWRITE)

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

47
48
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 | TRACE_ITER_SLEEP_TIME;
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
	"sleep-time",
320
321
322
	NULL
};

Steven Rostedt's avatar
Steven Rostedt committed
323
324
325
326
327
328
329
330
331
/*
 * 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.
 */
332
333
static raw_spinlock_t ftrace_max_lock =
	(raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
334
335
336
337
338
339

/*
 * 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
340
static void
341
342
343
344
345
346
347
348
349
350
351
352
__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;
353
	data->uid = task_uid(tsk);
354
355
356
357
358
	data->nice = tsk->static_prio - 20 - MAX_RT_PRIO;
	data->policy = tsk->policy;
	data->rt_priority = tsk->rt_priority;

	/* record this tasks comm */
359
	tracing_record_cmdline(tsk);
360
361
}

362
363
364
365
366
ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt)
{
	int len;
	int ret;

367
368
369
	if (!cnt)
		return 0;

370
371
372
373
374
375
376
	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);
377
	if (ret == cnt)
378
379
		return -EFAULT;

380
381
	cnt -= ret;

382
	s->readpos += cnt;
383
	return cnt;
384
385
}

386
static ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt)
387
388
389
390
391
392
393
394
395
396
397
398
399
400
{
	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;

401
	s->readpos += cnt;
402
403
404
	return cnt;
}

Steven Rostedt's avatar
Steven Rostedt committed
405
406
407
408
409
410
411
412
413
/**
 * 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
414
void
415
416
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
417
	struct ring_buffer *buf = tr->buffer;
418

419
	WARN_ON_ONCE(!irqs_disabled());
420
	__raw_spin_lock(&ftrace_max_lock);
421
422
423
424

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

425
	ftrace_disable_cpu();
426
	ring_buffer_reset(tr->buffer);
427
	ftrace_enable_cpu();
428
429

	__update_max_tr(tr, tsk, cpu);
430
	__raw_spin_unlock(&ftrace_max_lock);
431
432
433
434
435
436
437
}

/**
 * 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
438
439
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
440
 */
Ingo Molnar's avatar
Ingo Molnar committed
441
void
442
443
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
444
	int ret;
445

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

449
450
	ftrace_disable_cpu();

451
452
453
	ring_buffer_reset(max_tr.buffer);
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

454
455
	ftrace_enable_cpu();

456
	WARN_ON_ONCE(ret && ret != -EAGAIN);
457
458

	__update_max_tr(tr, tsk, cpu);
459
	__raw_spin_unlock(&ftrace_max_lock);
460
461
}

Steven Rostedt's avatar
Steven Rostedt committed
462
463
464
465
466
467
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
468
int register_tracer(struct tracer *type)
469
470
__releases(kernel_lock)
__acquires(kernel_lock)
471
472
473
474
475
476
477
478
479
480
{
	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
481
482
483
484
485
486
487
	/*
	 * 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();
488
	mutex_lock(&trace_types_lock);
Ingo Molnar's avatar
Ingo Molnar committed
489

490
491
	tracing_selftest_running = true;

492
493
494
495
496
497
498
499
500
501
	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;
		}
	}

502
503
504
505
506
507
508
	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;
509
510
511
	if (!type->wait_pipe)
		type->wait_pipe = default_wait_pipe;

512

Steven Rostedt's avatar
Steven Rostedt committed
513
#ifdef CONFIG_FTRACE_STARTUP_TEST
514
	if (type->selftest && !tracing_selftest_disabled) {
Steven Rostedt's avatar
Steven Rostedt committed
515
516
517
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
		int i;
518

Steven Rostedt's avatar
Steven Rostedt committed
519
520
521
522
523
524
525
		/*
		 * 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
526
		for_each_tracing_cpu(i)
527
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
528

Steven Rostedt's avatar
Steven Rostedt committed
529
530
531
532
533
534
535
536
537
538
		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
539
		/* Only reset on passing, to avoid touching corrupted buffers */
Ingo Molnar's avatar
Ingo Molnar committed
540
		for_each_tracing_cpu(i)
541
			tracing_reset(tr, i);
Ingo Molnar's avatar
Ingo Molnar committed
542

Steven Rostedt's avatar
Steven Rostedt committed
543
544
545
546
		printk(KERN_CONT "PASSED\n");
	}
#endif

547
548
549
550
551
	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
552

553
 out:
554
	tracing_selftest_running = false;
555
556
	mutex_unlock(&trace_types_lock);

Steven Rostedt's avatar
Steven Rostedt committed
557
558
559
560
561
562
563
564
565
566
567
568
	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;
569
#ifdef CONFIG_FTRACE_STARTUP_TEST
Steven Rostedt's avatar
Steven Rostedt committed
570
571
	printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n",
	       type->name);
572
573
#endif

Steven Rostedt's avatar
Steven Rostedt committed
574
 out_unlock:
575
	lock_kernel();
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
	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;
594
595
596
597
598
599
600
601
602

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

603
604
605
606
607
608
609
610
611
612
613
614
615
	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);
}

616
void tracing_reset(struct trace_array *tr, int cpu)
617
{
618
	ftrace_disable_cpu();
619
	ring_buffer_reset_cpu(tr->buffer, cpu);
620
	ftrace_enable_cpu();
621
622
}

623
624
625
626
627
628
629
630
631
632
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);
}

633
#define SAVED_CMDLINES 128
634
#define NO_CMDLINE_MAP UINT_MAX
635
636
637
638
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;
639
static raw_spinlock_t trace_cmdline_lock = __RAW_SPIN_LOCK_UNLOCKED;
640
641

/* temporary disable recording */
642
static atomic_t trace_record_cmdline_disabled __read_mostly;
643
644
645

static void trace_init_cmdlines(void)
{
646
647
	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));
648
649
650
	cmdline_idx = 0;
}

651
652
653
static int trace_stop_count;
static DEFINE_SPINLOCK(tracing_start_lock);

654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
/**
 * 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();
}

669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
/**
 * 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);
684
685
686
687
688
689
	if (--trace_stop_count) {
		if (trace_stop_count < 0) {
			/* Someone screwed up their debugging */
			WARN_ON_ONCE(1);
			trace_stop_count = 0;
		}
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
728
729
730
731
732
733
734
		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
735
void trace_stop_cmdline_recording(void);
736

Ingo Molnar's avatar
Ingo Molnar committed
737
static void trace_save_cmdline(struct task_struct *tsk)
738
{
739
	unsigned pid, idx;
740
741
742
743
744
745
746
747
748
749

	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.
	 */
750
	if (!__raw_spin_trylock(&trace_cmdline_lock))
751
752
753
		return;

	idx = map_pid_to_cmdline[tsk->pid];
754
	if (idx == NO_CMDLINE_MAP) {
755
756
		idx = (cmdline_idx + 1) % SAVED_CMDLINES;

757
758
759
760
761
762
763
764
765
		/*
		 * 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;
766

767
		map_cmdline_to_pid[idx] = tsk->pid;
768
769
770
771
772
773
774
		map_pid_to_cmdline[tsk->pid] = idx;

		cmdline_idx = idx;
	}

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

775
	__raw_spin_unlock(&trace_cmdline_lock);
776
777
}

778
void trace_find_cmdline(int pid, char comm[])
779
780
781
{
	unsigned map;

782
783
784
785
	if (!pid) {
		strcpy(comm, "<idle>");
		return;
	}
786

787
788
789
790
	if (pid > PID_MAX_DEFAULT) {
		strcpy(comm, "<...>");
		return;
	}
791

792
	__raw_spin_lock(&trace_cmdline_lock);
793
	map = map_pid_to_cmdline[pid];
794
795
796
797
	if (map != NO_CMDLINE_MAP)
		strcpy(comm, saved_cmdlines[map]);
	else
		strcpy(comm, "<...>");
798

799
	__raw_spin_unlock(&trace_cmdline_lock);
800
801
}

Ingo Molnar's avatar
Ingo Molnar committed
802
void tracing_record_cmdline(struct task_struct *tsk)
803
{
804
805
	if (atomic_read(&trace_record_cmdline_disabled) || !tracer_enabled ||
	    !tracing_is_on())
806
807
808
809
810
		return;

	trace_save_cmdline(tsk);
}

811
void
812
813
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
814
815
816
{
	struct task_struct *tsk = current;

817
818
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
819
	entry->tgid			= (tsk) ? tsk->tgid : 0;
820
	entry->flags =
821
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
822
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
823
824
825
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
826
827
828
829
830
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}

831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
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();
}

864
865
866
867
868
869
870
871
872
873
874
875
876
877
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
878
void
879
trace_function(struct trace_array *tr,
880
881
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
882
{
883
	struct ring_buffer_event *event;
884
	struct ftrace_entry *entry;
885

886
887
888
889
	/* If we are reading the ring buffer, don't trace */
	if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled))))
		return;

890
891
	event = trace_buffer_lock_reserve(tr, TRACE_FN, sizeof(*entry),
					  flags, pc);
892
893
894
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
895
896
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
897
	ring_buffer_unlock_commit(tr->buffer, event);
898
899
}

900
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
901
902
903
904
905
906
907
908
909
910
911
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;

912
913
	event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_ENT,
					  sizeof(*entry), flags, pc);
914
915
916
917
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
	entry->graph_ent			= *trace;
918
	ring_buffer_unlock_commit(global_trace.buffer, event);
919
920
921
}

static void __trace_graph_return(struct trace_array *tr,
922
				struct ftrace_graph_ret *trace,
923
924
925
926
				unsigned long flags,
				int pc)
{
	struct ring_buffer_event *event;
927
	struct ftrace_graph_ret_entry *entry;
928
929
930
931

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

932
933
	event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_RET,
					  sizeof(*entry), flags, pc);
934
935
936
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
937
	entry->ret				= *trace;
938
	ring_buffer_unlock_commit(global_trace.buffer, event);
939
940
941
}
#endif

Ingo Molnar's avatar
Ingo Molnar committed
942
void
Ingo Molnar's avatar
Ingo Molnar committed
943
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
944
945
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
Ingo Molnar's avatar
Ingo Molnar committed
946
947
{
	if (likely(!atomic_read(&data->disabled)))
948
		trace_function(tr, ip, parent_ip, flags, pc);
Ingo Molnar's avatar
Ingo Molnar committed
949
950
}

951
952
953
static void __ftrace_trace_stack(struct trace_array *tr,
				 unsigned long flags,
				 int skip, int pc)
Ingo Molnar's avatar
Ingo Molnar committed
954
{
955
#ifdef CONFIG_STACKTRACE
956
	struct ring_buffer_event *event;
957
	struct stack_entry *entry;
Ingo Molnar's avatar
Ingo Molnar committed
958
959
	struct stack_trace trace;

960
961
	event = trace_buffer_lock_reserve(tr, TRACE_STACK,
					  sizeof(*entry), flags, pc);
962
963
964
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
965
	memset(&entry->caller, 0, sizeof(entry->caller));
Ingo Molnar's avatar
Ingo Molnar committed
966
967
968
969

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
970
	trace.entries		= entry->caller;
Ingo Molnar's avatar
Ingo Molnar committed
971
972

	save_stack_trace(&trace);
973
	ring_buffer_unlock_commit(tr->buffer, event);
974
#endif
Ingo Molnar's avatar
Ingo Molnar committed
975
976
}

977
978
979
980
981
982
983
static void ftrace_trace_stack(struct trace_array *tr,
			       unsigned long flags,
			       int skip, int pc)
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

984
	__ftrace_trace_stack(tr, flags, skip, pc);
985
986
}

987
988
void __trace_stack(struct trace_array *tr,
		   unsigned long flags,
989
		   int skip, int pc)
990
{
991
	__ftrace_trace_stack(tr, flags, skip, pc);
992
993
}

994
static void ftrace_trace_userstack(struct trace_array *tr,
995
				   unsigned long flags, int pc)
996
{
997
#ifdef CONFIG_STACKTRACE
998
	struct ring_buffer_event *event;
999
1000
	struct userstack_entry *entry;
	struct stack_trace trace;
For faster browsing, not all history is shown. View entire blame