trace.c 100.1 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>
A
Alexey Dobriyan 已提交
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>
I
Ingo Molnar 已提交
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(int, ftrace_cpu_disabled);
90 91 92 93

static inline void ftrace_disable_cpu(void)
{
	preempt_disable();
94
	__this_cpu_inc(per_cpu_var(ftrace_cpu_disabled));
95 96 97 98
}

static inline void ftrace_enable_cpu(void)
{
99
	__this_cpu_dec(per_cpu_var(ftrace_cpu_disabled));
100 101 102
	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
static int tracing_set_tracer(const char *buf);

L
Li Zefan 已提交
128 129
#define MAX_TRACER_SIZE		100
static char bootup_tracer_buf[MAX_TRACER_SIZE] __initdata;
130
static char *default_bootup_tracer;
131

132
static int __init set_cmdline_ftrace(char *str)
133
{
L
Li Zefan 已提交
134
	strncpy(bootup_tracer_buf, str, MAX_TRACER_SIZE);
135
	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_cmdline_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);
S
Steven Rostedt 已提交
148

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

S
Steven Rostedt 已提交
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
int filter_current_check_discard(struct ring_buffer *buffer,
				 struct ftrace_event_call *call, void *rec,
174 175
				 struct ring_buffer_event *event)
{
176
	return filter_check_discard(call, rec, buffer, event);
177
}
178
EXPORT_SYMBOL_GPL(filter_current_check_discard);
179

180 181 182 183 184 185 186 187 188 189 190 191 192
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;
}
193

S
Steven Rostedt 已提交
194 195 196 197 198 199 200 201 202 203
/*
 * 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.
 */
204 205
static struct trace_array	max_tr;

206
static DEFINE_PER_CPU(struct trace_array_cpu, max_tr_data);
207

S
Steven Rostedt 已提交
208
/* tracer_enabled is used to toggle activation of a tracer */
209
static int			tracer_enabled = 1;
S
Steven Rostedt 已提交
210

211 212 213 214 215 216 217 218 219 220 221 222 223
/**
 * 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;
}

S
Steven Rostedt 已提交
224
/*
225 226 227
 * 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.
228 229 230 231 232
 *
 * 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.
S
Steven Rostedt 已提交
233
 */
234
#define TRACE_BUF_SIZE_DEFAULT	1441792UL /* 16384 * 88 (sizeof(entry)) */
235

236
static unsigned long		trace_buf_size = TRACE_BUF_SIZE_DEFAULT;
237

S
Steven Rostedt 已提交
238
/* trace_types holds a link list of available tracers. */
239
static struct tracer		*trace_types __read_mostly;
S
Steven Rostedt 已提交
240 241

/* current_trace points to the tracer that is currently active */
242
static struct tracer		*current_trace __read_mostly;
S
Steven Rostedt 已提交
243 244 245 246 247 248 249

/*
 * 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.
 */
250
static DEFINE_MUTEX(trace_types_lock);
S
Steven Rostedt 已提交
251 252

/* trace_wait is a waitqueue for tasks blocked on trace_poll */
253 254
static DECLARE_WAIT_QUEUE_HEAD(trace_wait);

255
/* trace_flags holds trace_options default values */
256
unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK |
257 258
	TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO | TRACE_ITER_SLEEP_TIME |
	TRACE_ITER_GRAPH_TIME;
259

260 261 262
static int trace_stop_count;
static DEFINE_SPINLOCK(tracing_start_lock);

S
Steven Rostedt 已提交
263 264 265 266 267 268
/**
 * 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.
 */
269 270
void trace_wake_up(void)
{
271 272 273 274
	int cpu;

	if (trace_flags & TRACE_ITER_BLOCK)
		return;
I
Ingo Molnar 已提交
275 276 277 278
	/*
	 * The runqueue_is_locked() can fail, but this is the best we
	 * have for now:
	 */
279 280
	cpu = get_cpu();
	if (!runqueue_is_locked(cpu))
281
		wake_up(&trace_wait);
282
	put_cpu();
283
}
284

285
static int __init set_buf_size(char *str)
286
{
287
	unsigned long buf_size;
288

289 290
	if (!str)
		return 0;
291
	buf_size = memparse(str, &str);
292
	/* nr_entries can not be zero */
293
	if (buf_size == 0)
294
		return 0;
295
	trace_buf_size = buf_size;
296 297
	return 1;
}
298
__setup("trace_buf_size=", set_buf_size);
299

S
Steven Rostedt 已提交
300 301 302 303 304
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

S
Steven Rostedt 已提交
305
/* These must match the bit postions in trace_iterator_flags */
306 307 308 309 310
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
I
Ingo Molnar 已提交
311
	"raw",
312
	"hex",
I
Ingo Molnar 已提交
313
	"bin",
314
	"block",
I
Ingo Molnar 已提交
315
	"stacktrace",
I
Ingo Molnar 已提交
316
	"sched-tree",
317
	"trace_printk",
318
	"ftrace_preempt",
319
	"branch",
320
	"annotate",
321
	"userstacktrace",
322
	"sym-userobj",
323
	"printk-msg-only",
324
	"context-info",
325
	"latency-format",
326
	"sleep-time",
327
	"graph-time",
328 329 330
	NULL
};

331 332 333 334 335 336 337 338 339 340
static struct {
	u64 (*func)(void);
	const char *name;
} trace_clocks[] = {
	{ trace_clock_local,	"local" },
	{ trace_clock_global,	"global" },
};

int trace_clock_id;

341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417
/*
 * trace_parser_get_init - gets the buffer for trace parser
 */
int trace_parser_get_init(struct trace_parser *parser, int size)
{
	memset(parser, 0, sizeof(*parser));

	parser->buffer = kmalloc(size, GFP_KERNEL);
	if (!parser->buffer)
		return 1;

	parser->size = size;
	return 0;
}

/*
 * trace_parser_put - frees the buffer for trace parser
 */
void trace_parser_put(struct trace_parser *parser)
{
	kfree(parser->buffer);
}

/*
 * trace_get_user - reads the user input string separated by  space
 * (matched by isspace(ch))
 *
 * For each string found the 'struct trace_parser' is updated,
 * and the function returns.
 *
 * Returns number of bytes read.
 *
 * See kernel/trace/trace.h for 'struct trace_parser' details.
 */
int trace_get_user(struct trace_parser *parser, const char __user *ubuf,
	size_t cnt, loff_t *ppos)
{
	char ch;
	size_t read = 0;
	ssize_t ret;

	if (!*ppos)
		trace_parser_clear(parser);

	ret = get_user(ch, ubuf++);
	if (ret)
		goto out;

	read++;
	cnt--;

	/*
	 * The parser is not finished with the last write,
	 * continue reading the user input without skipping spaces.
	 */
	if (!parser->cont) {
		/* skip white space */
		while (cnt && isspace(ch)) {
			ret = get_user(ch, ubuf++);
			if (ret)
				goto out;
			read++;
			cnt--;
		}

		/* only spaces were written */
		if (isspace(ch)) {
			*ppos += read;
			ret = read;
			goto out;
		}

		parser->idx = 0;
	}

	/* read the non-space input */
	while (cnt && !isspace(ch)) {
418
		if (parser->idx < parser->size - 1)
419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446
			parser->buffer[parser->idx++] = ch;
		else {
			ret = -EINVAL;
			goto out;
		}
		ret = get_user(ch, ubuf++);
		if (ret)
			goto out;
		read++;
		cnt--;
	}

	/* We either got finished input or we have to wait for another call. */
	if (isspace(ch)) {
		parser->buffer[parser->idx] = 0;
		parser->cont = false;
	} else {
		parser->cont = true;
		parser->buffer[parser->idx++] = ch;
	}

	*ppos += read;
	ret = read;

out:
	return ret;
}

447 448 449 450 451
ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt)
{
	int len;
	int ret;

452 453 454
	if (!cnt)
		return 0;

455 456 457 458 459 460 461
	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);
462
	if (ret == cnt)
463 464
		return -EFAULT;

465 466
	cnt -= ret;

467
	s->readpos += cnt;
468
	return cnt;
S
Steven Rostedt 已提交
469 470
}

471
static ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt)
472 473 474 475 476 477 478 479 480 481 482 483 484 485
{
	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;

486
	s->readpos += cnt;
487 488 489
	return cnt;
}

490 491 492 493 494 495
/*
 * 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.
 *
496
 * This is defined as a arch_spinlock_t in order to help
497 498 499 500 501 502
 * 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.
 */
503
static arch_spinlock_t ftrace_max_lock =
504
	(arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED;
505 506 507 508 509 510 511 512 513 514 515 516 517 518

#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];
519
	struct trace_array_cpu *max_data = tr->data[cpu];
520 521 522 523

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

524 525 526 527
	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;
528 529

	memcpy(data->comm, tsk->comm, TASK_COMM_LEN);
530 531 532 533 534
	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;
535 536 537 538 539

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

S
Steven Rostedt 已提交
540 541 542 543 544 545 546 547 548
/**
 * 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.
 */
I
Ingo Molnar 已提交
549
void
550 551
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
552
	struct ring_buffer *buf = tr->buffer;
553

554 555 556
	if (trace_stop_count)
		return;

557
	WARN_ON_ONCE(!irqs_disabled());
558
	arch_spin_lock(&ftrace_max_lock);
559 560 561 562

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

563
	__update_max_tr(tr, tsk, cpu);
564
	arch_spin_unlock(&ftrace_max_lock);
565 566 567 568 569 570 571
}

/**
 * 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.
S
Steven Rostedt 已提交
572 573
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
574
 */
I
Ingo Molnar 已提交
575
void
576 577
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
578
	int ret;
579

580 581 582
	if (trace_stop_count)
		return;

583
	WARN_ON_ONCE(!irqs_disabled());
584
	arch_spin_lock(&ftrace_max_lock);
585

586 587
	ftrace_disable_cpu();

588 589
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

590 591 592 593 594 595 596 597 598 599 600
	if (ret == -EBUSY) {
		/*
		 * We failed to swap the buffer due to a commit taking
		 * place on this CPU. We fail to record, but we reset
		 * the max trace buffer (no one writes directly to it)
		 * and flag that it failed.
		 */
		trace_array_printk(&max_tr, _THIS_IP_,
			"Failed to swap buffers due to commit in progress\n");
	}

601 602
	ftrace_enable_cpu();

603
	WARN_ON_ONCE(ret && ret != -EAGAIN && ret != -EBUSY);
604 605

	__update_max_tr(tr, tsk, cpu);
606
	arch_spin_unlock(&ftrace_max_lock);
607
}
608
#endif /* CONFIG_TRACER_MAX_TRACE */
609

S
Steven Rostedt 已提交
610 611 612 613 614 615
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
616
int register_tracer(struct tracer *type)
617 618
__releases(kernel_lock)
__acquires(kernel_lock)
619 620 621 622 623 624 625 626 627
{
	struct tracer *t;
	int ret = 0;

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

L
Li Zefan 已提交
628 629 630 631 632
	if (strlen(type->name) > MAX_TRACER_SIZE) {
		pr_info("Tracer has a name longer than %d\n", MAX_TRACER_SIZE);
		return -1;
	}

I
Ingo Molnar 已提交
633 634 635 636 637 638 639
	/*
	 * 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();
640
	mutex_lock(&trace_types_lock);
I
Ingo Molnar 已提交
641

642 643
	tracing_selftest_running = true;

644 645 646
	for (t = trace_types; t; t = t->next) {
		if (strcmp(type->name, t->name) == 0) {
			/* already found */
L
Li Zefan 已提交
647
			pr_info("Tracer %s already registered\n",
648 649 650 651 652 653
				type->name);
			ret = -1;
			goto out;
		}
	}

654 655 656 657 658 659 660
	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;
661 662 663
	if (!type->wait_pipe)
		type->wait_pipe = default_wait_pipe;

664

S
Steven Rostedt 已提交
665
#ifdef CONFIG_FTRACE_STARTUP_TEST
666
	if (type->selftest && !tracing_selftest_disabled) {
S
Steven Rostedt 已提交
667 668
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
669

S
Steven Rostedt 已提交
670 671 672 673 674 675 676
		/*
		 * 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.
		 */
677
		tracing_reset_online_cpus(tr);
I
Ingo Molnar 已提交
678

S
Steven Rostedt 已提交
679 680 681 682 683 684 685 686 687 688
		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;
		}
S
Steven Rostedt 已提交
689
		/* Only reset on passing, to avoid touching corrupted buffers */
690
		tracing_reset_online_cpus(tr);
I
Ingo Molnar 已提交
691

S
Steven Rostedt 已提交
692 693 694 695
		printk(KERN_CONT "PASSED\n");
	}
#endif

696 697
	type->next = trace_types;
	trace_types = type;
S
Steven Rostedt 已提交
698

699
 out:
700
	tracing_selftest_running = false;
701 702
	mutex_unlock(&trace_types_lock);

S
Steven Rostedt 已提交
703 704 705
	if (ret || !default_bootup_tracer)
		goto out_unlock;

L
Li Zefan 已提交
706
	if (strncmp(default_bootup_tracer, type->name, MAX_TRACER_SIZE))
S
Steven Rostedt 已提交
707 708 709 710 711 712 713 714
		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;
715
#ifdef CONFIG_FTRACE_STARTUP_TEST
S
Steven Rostedt 已提交
716 717
	printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n",
	       type->name);
718 719
#endif

S
Steven Rostedt 已提交
720
 out_unlock:
721
	lock_kernel();
722 723 724 725 726 727 728 729 730 731 732 733
	return ret;
}

void unregister_tracer(struct tracer *type)
{
	struct tracer **t;

	mutex_lock(&trace_types_lock);
	for (t = &trace_types; *t; t = &(*t)->next) {
		if (*t == type)
			goto found;
	}
L
Li Zefan 已提交
734
	pr_info("Tracer %s not registered\n", type->name);
735 736 737 738
	goto out;

 found:
	*t = (*t)->next;
739 740 741 742 743 744 745 746

	if (type == current_trace && tracer_enabled) {
		tracer_enabled = 0;
		tracing_stop();
		if (current_trace->stop)
			current_trace->stop(&global_trace);
		current_trace = &nop_trace;
	}
L
Li Zefan 已提交
747
out:
748 749 750
	mutex_unlock(&trace_types_lock);
}

751
static void __tracing_reset(struct trace_array *tr, int cpu)
752
{
753
	ftrace_disable_cpu();
754
	ring_buffer_reset_cpu(tr->buffer, cpu);
755
	ftrace_enable_cpu();
756 757
}

758 759 760 761 762 763 764 765 766 767 768 769 770
void tracing_reset(struct trace_array *tr, int cpu)
{
	struct ring_buffer *buffer = tr->buffer;

	ring_buffer_record_disable(buffer);

	/* Make sure all commits have finished */
	synchronize_sched();
	__tracing_reset(tr, cpu);

	ring_buffer_record_enable(buffer);
}

771 772
void tracing_reset_online_cpus(struct trace_array *tr)
{
773
	struct ring_buffer *buffer = tr->buffer;
774 775
	int cpu;

776 777 778 779 780
	ring_buffer_record_disable(buffer);

	/* Make sure all commits have finished */
	synchronize_sched();

781 782 783
	tr->time_start = ftrace_now(tr->cpu);

	for_each_online_cpu(cpu)
784
		__tracing_reset(tr, cpu);
785 786

	ring_buffer_record_enable(buffer);
787 788
}

789 790 791 792 793 794 795 796 797 798
void tracing_reset_current(int cpu)
{
	tracing_reset(&global_trace, cpu);
}

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

799
#define SAVED_CMDLINES 128
800
#define NO_CMDLINE_MAP UINT_MAX
801 802 803 804
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;
805
static arch_spinlock_t trace_cmdline_lock = __ARCH_SPIN_LOCK_UNLOCKED;
806 807

/* temporary disable recording */
808
static atomic_t trace_record_cmdline_disabled __read_mostly;
809 810 811

static void trace_init_cmdlines(void)
{
812 813
	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));
814 815 816
	cmdline_idx = 0;
}

817 818 819 820 821
int is_tracing_stopped(void)
{
	return trace_stop_count;
}

S
Steven Rostedt 已提交
822 823 824 825 826 827 828 829 830 831 832 833 834 835 836
/**
 * 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();
}

837 838 839 840 841 842 843 844 845 846 847 848 849 850 851
/**
 * 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);
852 853 854 855 856 857
	if (--trace_stop_count) {
		if (trace_stop_count < 0) {
			/* Someone screwed up their debugging */
			WARN_ON_ONCE(1);
			trace_stop_count = 0;
		}
858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902
		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);
}

I
Ingo Molnar 已提交
903
void trace_stop_cmdline_recording(void);
904

I
Ingo Molnar 已提交
905
static void trace_save_cmdline(struct task_struct *tsk)
906
{
907
	unsigned pid, idx;
908 909 910 911 912 913 914 915 916 917

	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.
	 */
918
	if (!arch_spin_trylock(&trace_cmdline_lock))
919 920 921
		return;

	idx = map_pid_to_cmdline[tsk->pid];
922
	if (idx == NO_CMDLINE_MAP) {
923 924
		idx = (cmdline_idx + 1) % SAVED_CMDLINES;

925 926 927 928 929 930 931 932 933
		/*
		 * 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;
934

935
		map_cmdline_to_pid[idx] = tsk->pid;
936 937 938 939 940 941 942
		map_pid_to_cmdline[tsk->pid] = idx;

		cmdline_idx = idx;
	}

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

943
	arch_spin_unlock(&trace_cmdline_lock);
944 945
}

946
void trace_find_cmdline(int pid, char comm[])
947 948 949
{
	unsigned map;

950 951 952 953
	if (!pid) {
		strcpy(comm, "<idle>");
		return;
	}
954

955 956 957 958
	if (pid > PID_MAX_DEFAULT) {
		strcpy(comm, "<...>");
		return;
	}
959

960
	preempt_disable();
961
	arch_spin_lock(&trace_cmdline_lock);
962
	map = map_pid_to_cmdline[pid];
963 964 965 966
	if (map != NO_CMDLINE_MAP)
		strcpy(comm, saved_cmdlines[map]);
	else
		strcpy(comm, "<...>");
967

968
	arch_spin_unlock(&trace_cmdline_lock);
969
	preempt_enable();
970 971
}

I
Ingo Molnar 已提交
972
void tracing_record_cmdline(struct task_struct *tsk)
973
{
974 975
	if (atomic_read(&trace_record_cmdline_disabled) || !tracer_enabled ||
	    !tracing_is_on())
976 977 978 979 980
		return;

	trace_save_cmdline(tsk);
}

981
void
982 983
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
984 985 986
{
	struct task_struct *tsk = current;

987 988
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
989
	entry->lock_depth		= (tsk) ? tsk->lock_depth : 0;
990
	entry->flags =
991
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
992
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
993 994 995
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
996 997 998 999
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}
1000
EXPORT_SYMBOL_GPL(tracing_generic_entry_update);
1001

1002 1003 1004 1005 1006
struct ring_buffer_event *
trace_buffer_lock_reserve(struct ring_buffer *buffer,
			  int type,
			  unsigned long len,
			  unsigned long flags, int pc)
1007 1008 1009
{
	struct ring_buffer_event *event;

1010
	event = ring_buffer_lock_reserve(buffer, len);
1011 1012 1013 1014 1015 1016 1017 1018 1019 1020
	if (event != NULL) {
		struct trace_entry *ent = ring_buffer_event_data(event);

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

	return event;
}

1021 1022 1023 1024 1025
static inline void
__trace_buffer_unlock_commit(struct ring_buffer *buffer,
			     struct ring_buffer_event *event,
			     unsigned long flags, int pc,
			     int wake)
1026
{
1027
	ring_buffer_unlock_commit(buffer, event);
1028

1029 1030
	ftrace_trace_stack(buffer, flags, 6, pc);
	ftrace_trace_userstack(buffer, flags, pc);
1031 1032 1033 1034 1035

	if (wake)
		trace_wake_up();
}

1036 1037 1038
void trace_buffer_unlock_commit(struct ring_buffer *buffer,
				struct ring_buffer_event *event,
				unsigned long flags, int pc)
1039
{
1040
	__trace_buffer_unlock_commit(buffer, event, flags, pc, 1);
1041 1042
}

1043
struct ring_buffer_event *
1044 1045
trace_current_buffer_lock_reserve(struct ring_buffer **current_rb,
				  int type, unsigned long len,
1046 1047
				  unsigned long flags, int pc)
{
1048 1049
	*current_rb = global_trace.buffer;
	return trace_buffer_lock_reserve(*current_rb,
1050 1051
					 type, len, flags, pc);
}
1052
EXPORT_SYMBOL_GPL(trace_current_buffer_lock_reserve);
1053

1054 1055
void trace_current_buffer_unlock_commit(struct ring_buffer *buffer,
					struct ring_buffer_event *event,
1056 1057
					unsigned long flags, int pc)
{
1058
	__trace_buffer_unlock_commit(buffer, event, flags, pc, 1);
1059
}
1060
EXPORT_SYMBOL_GPL(trace_current_buffer_unlock_commit);
1061

1062 1063 1064
void trace_nowake_buffer_unlock_commit(struct ring_buffer *buffer,
				       struct ring_buffer_event *event,
				       unsigned long flags, int pc)
1065
{
1066
	__trace_buffer_unlock_commit(buffer, event, flags, pc, 0);
1067
}
1068
EXPORT_SYMBOL_GPL(trace_nowake_buffer_unlock_commit);
1069

1070 1071
void trace_current_buffer_discard_commit(struct ring_buffer *buffer,
					 struct ring_buffer_event *event)
1072
{
1073
	ring_buffer_discard_commit(buffer, event);
1074
}
1075
EXPORT_SYMBOL_GPL(trace_current_buffer_discard_commit);
1076

I
Ingo Molnar 已提交
1077
void
1078
trace_function(struct trace_array *tr,
1079 1080
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
1081
{
1082
	struct ftrace_event_call *call = &event_function;
1083
	struct ring_buffer *buffer = tr->buffer;
1084
	struct ring_buffer_event *event;
1085
	struct ftrace_entry *entry;
1086

1087
	/* If we are reading the ring buffer, don't trace */
1088
	if (unlikely(__this_cpu_read(per_cpu_var(ftrace_cpu_disabled))))
1089 1090
		return;

1091
	event = trace_buffer_lock_reserve(buffer, TRACE_FN, sizeof(*entry),
1092
					  flags, pc);
1093 1094 1095
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1096 1097
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
1098

1099 1100
	if (!filter_check_discard(call, entry, buffer, event))
		ring_buffer_unlock_commit(buffer, event);
1101 1102
}

I
Ingo Molnar 已提交
1103
void
I
Ingo Molnar 已提交
1104
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
1105 1106
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
I
Ingo Molnar 已提交
1107 1108
{
	if (likely(!atomic_read(&data->disabled)))
1109
		trace_function(tr, ip, parent_ip, flags, pc);
I
Ingo Molnar 已提交
1110 1111
}

1112
#ifdef CONFIG_STACKTRACE
1113
static void __ftrace_trace_stack(struct ring_buffer *buffer,
1114 1115
				 unsigned long flags,
				 int skip, int pc)
I
Ingo Molnar 已提交
1116
{
1117
	struct ftrace_event_call *call = &event_kernel_stack;
1118
	struct ring_buffer_event *event;
1119
	struct stack_entry *entry;
I
Ingo Molnar 已提交
1120 1121
	struct stack_trace trace;

1122
	event = trace_buffer_lock_reserve(buffer, TRACE_STACK,
1123
					  sizeof(*entry), flags, pc);
1124 1125 1126
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1127
	memset(&entry->caller, 0, sizeof(entry->caller));
I
Ingo Molnar 已提交
1128 1129 1130 1131

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
1132
	trace.entries		= entry->caller;
I
Ingo Molnar 已提交
1133 1134

	save_stack_trace(&trace);
1135 1136
	if (!filter_check_discard(call, entry, buffer, event))
		ring_buffer_unlock_commit(buffer, event);
I
Ingo Molnar 已提交
1137 1138
}

1139 1140
void ftrace_trace_stack(struct ring_buffer *buffer, unsigned long flags,
			int skip, int pc)
1141 1142 1143 1144
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

1145
	__ftrace_trace_stack(buffer, flags, skip, pc);
1146 1147
}

1148 1149
void __trace_stack(struct trace_array *tr, unsigned long flags, int skip,
		   int pc)
1150
{
1151
	__ftrace_trace_stack(tr->buffer, flags, skip, pc);
1152 1153
}

1154 1155
void
ftrace_trace_userstack(struct ring_buffer *buffer, unsigned long flags, int pc)
1156
{
1157
	struct ftrace_event_call *call = &event_user_stack;
1158
	struct ring_buffer_event *event;
1159 1160 1161 1162 1163 1164
	struct userstack_entry *entry;
	struct stack_trace trace;

	if (!(trace_flags & TRACE_ITER_USERSTACKTRACE))
		return;

1165
	event = trace_buffer_lock_reserve(buffer, TRACE_USER_STACK,
1166
					  sizeof(*entry), flags, pc);
1167 1168 1169 1170
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);

1171
	entry->tgid		= current->tgid;
1172 1173 1174 1175 1176 1177 1178 1179
	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);
1180 1181
	if (!filter_check_discard(call, entry, buffer, event))
		ring_buffer_unlock_commit(buffer, event);
1182 1183
}

1184 1185
#ifdef UNUSED
static void __trace_userstack(struct trace_array *tr, unsigned long flags)
1186
{
1187
	ftrace_trace_userstack(tr, flags, preempt_count());
1188
}
1189
#endif /* UNUSED */
1190

1191 1192
#endif /* CONFIG_STACKTRACE */

1193
static void
1194
ftrace_trace_special(void *__tr,
1195 1196
		     unsigned long arg1, unsigned long arg2, unsigned long arg3,
		     int pc)
I
Ingo Molnar 已提交
1197
{
1198
	struct ftrace_event_call *call = &event_special;
1199
	struct ring_buffer_event *event;
I
Ingo Molnar 已提交
1200
	struct trace_array *tr = __tr;
1201
	struct ring_buffer *buffer = tr->buffer;
1202
	struct special_entry *entry;
I
Ingo Molnar 已提交
1203

1204
	event = trace_buffer_lock_reserve(buffer, TRACE_SPECIAL,
1205
					  sizeof(*entry), 0, pc);
1206 1207 1208
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1209 1210 1211
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
1212 1213 1214

	if (!filter_check_discard(call, entry, buffer, event))
		trace_buffer_unlock_commit(buffer, event, 0, pc);
I
Ingo Molnar 已提交
1215 1216
}

1217 1218 1219 1220
void
__trace_special(void *__tr, void *__data,
		unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
1221
	ftrace_trace_special(__tr, arg1, arg2, arg3, preempt_count());
1222 1223
}

1224 1225 1226 1227 1228
void
ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1229
	unsigned long flags;
1230
	int cpu;
1231
	int pc;
1232

S
Steven Rostedt 已提交
1233
	if (tracing_disabled)
1234 1235
		return;

1236
	pc = preempt_count();
1237
	local_irq_save(flags);
1238 1239 1240
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

1241
	if (likely(atomic_inc_return(&data->disabled) == 1))
1242
		ftrace_trace_special(tr, arg1, arg2, arg3, pc);
1243

1244 1245
	atomic_dec(&data->disabled);
	local_irq_restore(flags);
1246 1247
}

1248
/**
1249
 * trace_vbprintk - write binary msg to tracing buffer
1250 1251
 *
 */
1252
int trace_vbprintk(unsigned long ip, const char *fmt, va_list args)
1253
{
1254
	static arch_spinlock_t trace_buf_lock =
1255
		(arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED;
1256 1257
	static u32 trace_buf[TRACE_BUF_SIZE];

1258
	struct ftrace_event_call *call = &event_bprint;
1259
	struct ring_buffer_event *event;
1260
	struct ring_buffer *buffer;
1261 1262
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1263
	struct bprint_entry *entry;
1264
	unsigned long flags;
1265
	int disable;
1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279
	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];

1280 1281
	disable = atomic_inc_return(&data->disabled);
	if (unlikely(disable != 1))
1282 1283
		goto out;

1284 1285
	/* Lockdep uses trace_printk for lock tracing */
	local_irq_save(flags);
1286
	arch_spin_lock(&trace_buf_lock);
1287 1288 1289 1290 1291 1292
	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;
1293 1294 1295
	buffer = tr->buffer;
	event = trace_buffer_lock_reserve(buffer, TRACE_BPRINT, size,
					  flags, pc);
1296 1297 1298 1299 1300 1301 1302
	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);
1303 1304
	if (!filter_check_discard(call, entry, buffer, event))
		ring_buffer_unlock_commit(buffer, event);
1305 1306

out_unlock:
1307
	arch_spin_unlock(&trace_buf_lock);
1308
	local_irq_restore(flags);
1309 1310

out:
1311
	atomic_dec_return(&data->disabled);
1312 1313 1314 1315 1316
	ftrace_preempt_enable(resched);
	unpause_graph_tracing();

	return len;
}
1317 1318
EXPORT_SYMBOL_GPL(trace_vbprintk);

1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335
int trace_array_printk(struct trace_array *tr,
		       unsigned long ip, const char *fmt, ...)
{
	int ret;
	va_list ap;

	if (!(trace_flags & TRACE_ITER_PRINTK))
		return 0;

	va_start(ap, fmt);
	ret = trace_array_vprintk(tr, ip, fmt, ap);
	va_end(ap);
	return ret;
}

int trace_array_vprintk(struct trace_array *tr,
			unsigned long ip, const char *fmt, va_list args)
1336
{
1337
	static arch_spinlock_t trace_buf_lock = __ARCH_SPIN_LOCK_UNLOCKED;
1338 1339
	static char trace_buf[TRACE_BUF_SIZE];

1340
	struct ftrace_event_call *call = &event_print;
1341
	struct ring_buffer_event *event;
1342
	struct ring_buffer *buffer;
1343 1344 1345 1346
	struct trace_array_cpu *data;
	int cpu, len = 0, size, pc;
	struct print_entry *entry;
	unsigned long irq_flags;
1347
	int disable;
1348 1349 1350 1351 1352 1353 1354 1355 1356

	if (tracing_disabled || tracing_selftest_running)
		return 0;

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

1357 1358
	disable = atomic_inc_return(&data->disabled);
	if (unlikely(disable != 1))
1359 1360 1361 1362
		goto out;

	pause_graph_tracing();
	raw_local_irq_save(irq_flags);
1363
	arch_spin_lock(&trace_buf_lock);
1364
	len = vsnprintf(trace_buf, TRACE_BUF_SIZE, fmt, args);
1365 1366

	size = sizeof(*entry) + len + 1;
1367 1368 1369
	buffer = tr->buffer;
	event = trace_buffer_lock_reserve(buffer, TRACE_PRINT, size,
					  irq_flags, pc);
1370 1371 1372
	if (!event)
		goto out_unlock;
	entry = ring_buffer_event_data(event);
C
Carsten Emde 已提交
1373
	entry->ip = ip;
1374 1375

	memcpy(&entry->buf, trace_buf, len);
C
Carsten Emde 已提交
1376
	entry->buf[len] = '\0';
1377 1378
	if (!filter_check_discard(call, entry, buffer, event))
		ring_buffer_unlock_commit(buffer, event);
1379 1380

 out_unlock:
1381
	arch_spin_unlock(&trace_buf_lock);
1382 1383 1384
	raw_local_irq_restore(irq_flags);
	unpause_graph_tracing();
 out:
1385
	atomic_dec_return(&data->disabled);
1386 1387 1388 1389
	preempt_enable_notrace();

	return len;
}
1390 1391 1392

int trace_vprintk(unsigned long ip, const char *fmt, va_list args)
{
S
Steven Rostedt 已提交
1393
	return trace_array_vprintk(&global_trace, ip, fmt, args);
1394
}
1395 1396
EXPORT_SYMBOL_GPL(trace_vprintk);

1397 1398
enum trace_file_type {
	TRACE_FILE_LAT_FMT	= 1,
1399
	TRACE_FILE_ANNOTATE	= 2,
1400 1401
};

1402
static void trace_iterator_increment(struct trace_iterator *iter)
S
Steven Rostedt 已提交
1403
{
1404 1405 1406
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();

S
Steven Rostedt 已提交
1407
	iter->idx++;
1408 1409 1410 1411
	if (iter->buffer_iter[iter->cpu])
		ring_buffer_read(iter->buffer_iter[iter->cpu], NULL);

	ftrace_enable_cpu();
S
Steven Rostedt 已提交
1412 1413
}

I
Ingo Molnar 已提交
1414
static struct trace_entry *
1415
peek_next_entry(struct trace_iterator *iter, int cpu, u64 *ts)
1416
{
1417 1418
	struct ring_buffer_event *event;
	struct ring_buffer_iter *buf_iter = iter->buffer_iter[cpu];
1419

1420 1421 1422 1423 1424 1425 1426 1427 1428 1429
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();

	if (buf_iter)
		event = ring_buffer_iter_peek(buf_iter, ts);
	else
		event = ring_buffer_peek(iter->tr->buffer, cpu, ts);

	ftrace_enable_cpu();

1430
	return event ? ring_buffer_event_data(event) : NULL;
1431
}
1432

1433
static struct trace_entry *
1434
__find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts)
1435
{
1436
	struct ring_buffer *buffer = iter->tr->buffer;
1437
	struct trace_entry *ent, *next = NULL;
1438
	int cpu_file = iter->cpu_file;
1439
	u64 next_ts = 0, ts;
1440 1441 1442
	int next_cpu = -1;
	int cpu;

1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456
	/*
	 * If we are in a per_cpu trace file, don't bother by iterating over
	 * all cpu and peek directly.
	 */
	if (cpu_file > TRACE_PIPE_ALL_CPU) {
		if (ring_buffer_empty_cpu(buffer, cpu_file))
			return NULL;
		ent = peek_next_entry(iter, cpu_file, ent_ts);
		if (ent_cpu)
			*ent_cpu = cpu_file;

		return ent;
	}

1457
	for_each_tracing_cpu(cpu) {
1458

1459 1460
		if (ring_buffer_empty_cpu(buffer, cpu))
			continue;
1461

1462
		ent = peek_next_entry(iter, cpu, &ts);
1463

I
Ingo Molnar 已提交
1464 1465 1466
		/*
		 * Pick the entry with the smallest timestamp:
		 */
1467
		if (ent && (!next || ts < next_ts)) {
1468 1469
			next = ent;
			next_cpu = cpu;
1470
			next_ts = ts;
1471 1472 1473 1474 1475 1476
		}
	}

	if (ent_cpu)
		*ent_cpu = next_cpu;

1477 1478 1479
	if (ent_ts)
		*ent_ts = next_ts;

1480 1481 1482
	return next;
}

1483
/* Find the next real entry, without updating the iterator itself */
1484 1485
struct trace_entry *trace_find_next_entry(struct trace_iterator *iter,
					  int *ent_cpu, u64 *ent_ts)
1486
{
1487
	return __find_next_entry(iter, ent_cpu, ent_ts);
1488 1489 1490 1491 1492
}

/* Find the next real entry, and increment the iterator to the next entry */
static void *find_next_entry_inc(struct trace_iterator *iter)
{
1493
	iter->ent = __find_next_entry(iter, &iter->cpu, &iter->ts);
1494

1495
	if (iter->ent)
1496
		trace_iterator_increment(iter);
1497

1498
	return iter->ent ? iter : NULL;
1499
}
1500

I
Ingo Molnar 已提交
1501
static void trace_consume(struct trace_iterator *iter)
1502
{
1503 1504
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();
1505
	ring_buffer_consume(iter->tr->buffer, iter->cpu, &iter->ts);
1506
	ftrace_enable_cpu();
1507 1508
}

I
Ingo Molnar 已提交
1509
static void *s_next(struct seq_file *m, void *v, loff_t *pos)
1510 1511 1512
{
	struct trace_iterator *iter = m->private;
	int i = (int)*pos;
I
Ingo Molnar 已提交
1513
	void *ent;
1514

1515 1516
	WARN_ON_ONCE(iter->leftover);

1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535
	(*pos)++;

	/* can't go backwards */
	if (iter->idx > i)
		return NULL;

	if (iter->idx < 0)
		ent = find_next_entry_inc(iter);
	else
		ent = iter;

	while (ent && iter->idx < i)
		ent = find_next_entry_inc(iter);

	iter->pos = *pos;

	return ent;
}

1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566
static void tracing_iter_reset(struct trace_iterator *iter, int cpu)
{
	struct trace_array *tr = iter->tr;
	struct ring_buffer_event *event;
	struct ring_buffer_iter *buf_iter;
	unsigned long entries = 0;
	u64 ts;

	tr->data[cpu]->skipped_entries = 0;

	if (!iter->buffer_iter[cpu])
		return;

	buf_iter = iter->buffer_iter[cpu];
	ring_buffer_iter_reset(buf_iter);

	/*
	 * We could have the case with the max latency tracers
	 * that a reset never took place on a cpu. This is evident
	 * by the timestamp being before the start of the buffer.
	 */
	while ((event = ring_buffer_iter_peek(buf_iter, &ts))) {
		if (ts >= iter->tr->time_start)
			break;
		entries++;
		ring_buffer_read(buf_iter, NULL);
	}

	tr->data[cpu]->skipped_entries = entries;
}

1567 1568 1569 1570 1571 1572 1573 1574 1575 1576
/*
 * No necessary locking here. The worst thing which can
 * happen is loosing events consumed at the same time
 * by a trace_pipe reader.
 * Other than that, we don't risk to crash the ring buffer
 * because it serializes the readers.
 *
 * The current tracer is copied to avoid a global locking
 * all around.
 */
1577 1578 1579
static void *s_start(struct seq_file *m, loff_t *pos)
{
	struct trace_iterator *iter = m->private;
1580
	static struct tracer *old_tracer;
1581
	int cpu_file = iter->cpu_file;
1582 1583
	void *p = NULL;
	loff_t l = 0;
1584
	int cpu;
1585

1586
	/* copy the tracer to avoid using a global lock all around */
1587
	mutex_lock(&trace_types_lock);
1588 1589 1590
	if (unlikely(old_tracer != current_trace && current_trace)) {
		old_tracer = current_trace;
		*iter->trace = *current_trace;
1591
	}
1592
	mutex_unlock(&trace_types_lock);
1593 1594 1595 1596 1597 1598 1599 1600

	atomic_inc(&trace_record_cmdline_disabled);

	if (*pos != iter->pos) {
		iter->ent = NULL;
		iter->cpu = 0;
		iter->idx = -1;

1601 1602
		ftrace_disable_cpu();

1603 1604
		if (cpu_file == TRACE_PIPE_ALL_CPU) {
			for_each_tracing_cpu(cpu)
1605
				tracing_iter_reset(iter, cpu);
1606
		} else
1607
			tracing_iter_reset(iter, cpu_file);
1608

1609 1610
		ftrace_enable_cpu();

1611 1612 1613 1614
		for (p = iter; p && l < *pos; p = s_next(m, p, &l))
			;

	} else {
1615 1616 1617 1618 1619 1620 1621 1622 1623 1624
		/*
		 * If we overflowed the seq_file before, then we want
		 * to just reuse the trace_seq buffer again.
		 */
		if (iter->leftover)
			p = iter;
		else {
			l = *pos - 1;
			p = s_next(m, p, &l);
		}
1625 1626
	}

1627
	trace_event_read_lock();
1628 1629 1630 1631 1632 1633
	return p;
}

static void s_stop(struct seq_file *m, void *p)
{
	atomic_dec(&trace_record_cmdline_disabled);
1634
	trace_event_read_unlock();
1635 1636
}

I
Ingo Molnar 已提交
1637
static void print_lat_help_header(struct seq_file *m)
1638
{
1639 1640 1641 1642 1643
	seq_puts(m, "#                  _------=> CPU#            \n");
	seq_puts(m, "#                 / _-----=> irqs-off        \n");
	seq_puts(m, "#                | / _----=> need-resched    \n");
	seq_puts(m, "#                || / _---=> hardirq/softirq \n");
	seq_puts(m, "#                ||| / _--=> preempt-depth   \n");
1644 1645 1646 1647
	seq_puts(m, "#                |||| /_--=> lock-depth       \n");
	seq_puts(m, "#                |||||/     delay             \n");
	seq_puts(m, "#  cmd     pid   |||||| time  |   caller      \n");
	seq_puts(m, "#     \\   /      ||||||   \\   |   /           \n");
1648 1649
}

I
Ingo Molnar 已提交
1650
static void print_func_help_header(struct seq_file *m)
1651
{
1652 1653
	seq_puts(m, "#           TASK-PID    CPU#    TIMESTAMP  FUNCTION\n");
	seq_puts(m, "#              | |       |          |         |\n");
1654 1655 1656
}


I
Ingo Molnar 已提交
1657
static void
1658 1659 1660 1661 1662 1663
print_trace_header(struct seq_file *m, struct trace_iterator *iter)
{
	unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK);
	struct trace_array *tr = iter->tr;
	struct trace_array_cpu *data = tr->data[tr->cpu];
	struct tracer *type = current_trace;
1664 1665 1666
	unsigned long entries = 0;
	unsigned long total = 0;
	unsigned long count;
1667
	const char *name = "preemption";
1668
	int cpu;
1669 1670 1671 1672

	if (type)
		name = type->name;

1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689

	for_each_tracing_cpu(cpu) {
		count = ring_buffer_entries_cpu(tr->buffer, cpu);
		/*
		 * If this buffer has skipped entries, then we hold all
		 * entries for the trace and we need to ignore the
		 * ones before the time stamp.
		 */
		if (tr->data[cpu]->skipped_entries) {
			count -= tr->data[cpu]->skipped_entries;
			/* total is the same as the entries */
			total += count;
		} else
			total += count +
				ring_buffer_overrun_cpu(tr->buffer, cpu);
		entries += count;
	}
1690

1691
	seq_printf(m, "# %s latency trace v1.1.5 on %s\n",
1692
		   name, UTS_RELEASE);
1693
	seq_puts(m, "# -----------------------------------"
1694
		 "---------------------------------\n");
1695
	seq_printf(m, "# latency: %lu us, #%lu/%lu, CPU#%d |"
1696
		   " (M:%s VP:%d, KP:%d, SP:%d HP:%d",
S
Steven Rostedt 已提交
1697
		   nsecs_to_usecs(data->saved_latency),
1698
		   entries,
1699
		   total,
1700 1701 1702 1703 1704
		   tr->cpu,
#if defined(CONFIG_PREEMPT_NONE)
		   "server",
#elif defined(CONFIG_PREEMPT_VOLUNTARY)
		   "desktop",
1705
#elif defined(CONFIG_PREEMPT)
1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716
		   "preempt",
#else
		   "unknown",
#endif
		   /* These are reserved for later use */
		   0, 0, 0, 0);
#ifdef CONFIG_SMP
	seq_printf(m, " #P:%d)\n", num_online_cpus());
#else
	seq_puts(m, ")\n");
#endif
1717 1718
	seq_puts(m, "#    -----------------\n");
	seq_printf(m, "#    | task: %.16s-%d "
1719 1720 1721
		   "(uid:%d nice:%ld policy:%ld rt_prio:%ld)\n",
		   data->comm, data->pid, data->uid, data->nice,
		   data->policy, data->rt_priority);
1722
	seq_puts(m, "#    -----------------\n");
1723 1724

	if (data->critical_start) {
1725
		seq_puts(m, "#  => started at: ");
S
Steven Rostedt 已提交
1726 1727
		seq_print_ip_sym(&iter->seq, data->critical_start, sym_flags);
		trace_print_seq(m, &iter->seq);
1728
		seq_puts(m, "\n#  => ended at:   ");
S
Steven Rostedt 已提交
1729 1730
		seq_print_ip_sym(&iter->seq, data->critical_end, sym_flags);
		trace_print_seq(m, &iter->seq);
1731
		seq_puts(m, "\n#\n");
1732 1733
	}

1734
	seq_puts(m, "#\n");
1735 1736
}

1737 1738 1739 1740
static void test_cpu_buff_start(struct trace_iterator *iter)
{
	struct trace_seq *s = &iter->seq;

1741 1742 1743 1744 1745 1746
	if (!(trace_flags & TRACE_ITER_ANNOTATE))
		return;

	if (!(iter->iter_flags & TRACE_FILE_ANNOTATE))
		return;

1747
	if (cpumask_test_cpu(iter->cpu, iter->started))
1748 1749
		return;

1750 1751 1752
	if (iter->tr->data[iter->cpu]->skipped_entries)
		return;

1753
	cpumask_set_cpu(iter->cpu, iter->started);
1754 1755 1756 1757 1758

	/* Don't print started cpu buffer for the first entry of the trace */
	if (iter->idx > 1)
		trace_seq_printf(s, "##### CPU %u buffer started ####\n",
				iter->cpu);
1759 1760
}

1761
static enum print_line_t print_trace_fmt(struct trace_iterator *iter)
1762
{
S
Steven Rostedt 已提交
1763
	struct trace_seq *s = &iter->seq;
1764
	unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK);
I
Ingo Molnar 已提交
1765
	struct trace_entry *entry;
1766
	struct trace_event *event;
1767

I
Ingo Molnar 已提交
1768
	entry = iter->ent;
1769

1770 1771
	test_cpu_buff_start(iter);

1772
	event = ftrace_find_event(entry->type);
1773

1774
	if (trace_flags & TRACE_ITER_CONTEXT_INFO) {
1775 1776 1777 1778 1779 1780 1781
		if (iter->iter_flags & TRACE_FILE_LAT_FMT) {
			if (!trace_print_lat_context(iter))
				goto partial;
		} else {
			if (!trace_print_context(iter))
				goto partial;
		}
1782
	}
1783

1784
	if (event)
1785 1786 1787 1788
		return event->trace(iter, sym_flags);

	if (!trace_seq_printf(s, "Unknown type %d\n", entry->type))
		goto partial;
1789

1790
	return TRACE_TYPE_HANDLED;
1791 1792
partial:
	return TRACE_TYPE_PARTIAL_LINE;
1793 1794
}

1795
static enum print_line_t print_raw_fmt(struct trace_iterator *iter)
I
Ingo Molnar 已提交
1796 1797 1798
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *entry;
1799
	struct trace_event *event;
I
Ingo Molnar 已提交
1800 1801

	entry = iter->ent;
1802

1803
	if (trace_flags & TRACE_ITER_CONTEXT_INFO) {
1804 1805 1806
		if (!trace_seq_printf(s, "%d %d %llu ",
				      entry->pid, iter->cpu, iter->ts))
			goto partial;
1807
	}
I
Ingo Molnar 已提交
1808

1809
	event = ftrace_find_event(entry->type);
1810
	if (event)
1811 1812 1813 1814
		return event->raw(iter, 0);

	if (!trace_seq_printf(s, "%d ?\n", entry->type))
		goto partial;
1815

1816
	return TRACE_TYPE_HANDLED;
1817 1818
partial:
	return TRACE_TYPE_PARTIAL_LINE;
I
Ingo Molnar 已提交
1819 1820
}

1821
static enum print_line_t print_hex_fmt(struct trace_iterator *iter)
1822 1823 1824 1825
{
	struct trace_seq *s = &iter->seq;
	unsigned char newline = '\n';
	struct trace_entry *entry;
1826
	struct trace_event *event;
1827 1828

	entry = iter->ent;
1829

1830 1831 1832 1833 1834
	if (trace_flags & TRACE_ITER_CONTEXT_INFO) {
		SEQ_PUT_HEX_FIELD_RET(s, entry->pid);
		SEQ_PUT_HEX_FIELD_RET(s, iter->cpu);
		SEQ_PUT_HEX_FIELD_RET(s, iter->ts);
	}
1835

1836
	event = ftrace_find_event(entry->type);
1837
	if (event) {
1838
		enum print_line_t ret = event->hex(iter, 0);
1839 1840 1841
		if (ret != TRACE_TYPE_HANDLED)
			return ret;
	}
S
Steven Rostedt 已提交
1842

1843 1844
	SEQ_PUT_FIELD_RET(s, newline);

1845
	return TRACE_TYPE_HANDLED;
1846 1847
}

1848
static enum print_line_t print_bin_fmt(struct trace_iterator *iter)
I
Ingo Molnar 已提交
1849 1850 1851
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *entry;
1852
	struct trace_event *event;
I
Ingo Molnar 已提交
1853 1854

	entry = iter->ent;
1855

1856 1857
	if (trace_flags & TRACE_ITER_CONTEXT_INFO) {
		SEQ_PUT_FIELD_RET(s, entry->pid);
1858
		SEQ_PUT_FIELD_RET(s, iter->cpu);
1859 1860
		SEQ_PUT_FIELD_RET(s, iter->ts);
	}
I
Ingo Molnar 已提交
1861

1862
	event = ftrace_find_event(entry->type);
1863
	return event ? event->binary(iter, 0) : TRACE_TYPE_HANDLED;
I
Ingo Molnar 已提交
1864 1865
}

1866 1867 1868 1869
static int trace_empty(struct trace_iterator *iter)
{
	int cpu;

1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882
	/* If we are looking at one CPU buffer, only check that one */
	if (iter->cpu_file != TRACE_PIPE_ALL_CPU) {
		cpu = iter->cpu_file;
		if (iter->buffer_iter[cpu]) {
			if (!ring_buffer_iter_empty(iter->buffer_iter[cpu]))
				return 0;
		} else {
			if (!ring_buffer_empty_cpu(iter->tr->buffer, cpu))
				return 0;
		}
		return 1;
	}

1883
	for_each_tracing_cpu(cpu) {
1884 1885 1886 1887 1888 1889 1890
		if (iter->buffer_iter[cpu]) {
			if (!ring_buffer_iter_empty(iter->buffer_iter[cpu]))
				return 0;
		} else {
			if (!ring_buffer_empty_cpu(iter->tr->buffer, cpu))
				return 0;
		}
1891
	}
1892

1893
	return 1;
1894 1895
}

1896
/*  Called with trace_event_read_lock() held. */
1897
static enum print_line_t print_trace_line(struct trace_iterator *iter)
I
Ingo Molnar 已提交
1898
{
1899 1900 1901 1902 1903 1904 1905
	enum print_line_t ret;

	if (iter->trace && iter->trace->print_line) {
		ret = iter->trace->print_line(iter);
		if (ret != TRACE_TYPE_UNHANDLED)
			return ret;
	}
1906

1907 1908 1909
	if (iter->ent->type == TRACE_BPRINT &&
			trace_flags & TRACE_ITER_PRINTK &&
			trace_flags & TRACE_ITER_PRINTK_MSGONLY)
1910
		return trace_print_bprintk_msg_only(iter);
1911

1912 1913 1914
	if (iter->ent->type == TRACE_PRINT &&
			trace_flags & TRACE_ITER_PRINTK &&
			trace_flags & TRACE_ITER_PRINTK_MSGONLY)
1915
		return trace_print_printk_msg_only(iter);
1916

I
Ingo Molnar 已提交
1917 1918 1919
	if (trace_flags & TRACE_ITER_BIN)
		return print_bin_fmt(iter);

1920 1921 1922
	if (trace_flags & TRACE_ITER_HEX)
		return print_hex_fmt(iter);

I
Ingo Molnar 已提交
1923 1924 1925 1926 1927 1928
	if (trace_flags & TRACE_ITER_RAW)
		return print_raw_fmt(iter);

	return print_trace_fmt(iter);
}

1929 1930 1931
static int s_show(struct seq_file *m, void *v)
{
	struct trace_iterator *iter = v;
1932
	int ret;
1933 1934 1935 1936 1937 1938

	if (iter->ent == NULL) {
		if (iter->tr) {
			seq_printf(m, "# tracer: %s\n", iter->trace->name);
			seq_puts(m, "#\n");
		}
1939 1940 1941
		if (iter->trace && iter->trace->print_header)
			iter->trace->print_header(m);
		else if (iter->iter_flags & TRACE_FILE_LAT_FMT) {
1942 1943 1944 1945 1946 1947 1948 1949 1950 1951
			/* print nothing if the buffers are empty */
			if (trace_empty(iter))
				return 0;
			print_trace_header(m, iter);
			if (!(trace_flags & TRACE_ITER_VERBOSE))
				print_lat_help_header(m);
		} else {
			if (!(trace_flags & TRACE_ITER_VERBOSE))
				print_func_help_header(m);
		}
1952 1953 1954 1955 1956 1957 1958 1959 1960 1961
	} else if (iter->leftover) {
		/*
		 * If we filled the seq_file buffer earlier, we
		 * want to just show it now.
		 */
		ret = trace_print_seq(m, &iter->seq);

		/* ret should this time be zero, but you never know */
		iter->leftover = ret;

1962
	} else {
I
Ingo Molnar 已提交
1963
		print_trace_line(iter);
1964 1965 1966 1967 1968 1969 1970 1971 1972
		ret = trace_print_seq(m, &iter->seq);
		/*
		 * If we overflow the seq_file buffer, then it will
		 * ask us for this data again at start up.
		 * Use that instead.
		 *  ret is 0 if seq_file write succeeded.
		 *        -1 otherwise.
		 */
		iter->leftover = ret;
1973 1974 1975 1976 1977
	}

	return 0;
}

J
James Morris 已提交
1978
static const struct seq_operations tracer_seq_ops = {
I
Ingo Molnar 已提交
1979 1980 1981 1982
	.start		= s_start,
	.next		= s_next,
	.stop		= s_stop,
	.show		= s_show,
1983 1984
};

I
Ingo Molnar 已提交
1985
static struct trace_iterator *
1986
__tracing_open(struct inode *inode, struct file *file)
1987
{
1988
	long cpu_file = (long) inode->i_private;
1989
	void *fail_ret = ERR_PTR(-ENOMEM);
1990
	struct trace_iterator *iter;
1991
	struct seq_file *m;
1992
	int cpu, ret;
1993

1994 1995
	if (tracing_disabled)
		return ERR_PTR(-ENODEV);
S
Steven Rostedt 已提交
1996

1997
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
1998 1999
	if (!iter)
		return ERR_PTR(-ENOMEM);
2000

2001 2002 2003 2004
	/*
	 * We make a copy of the current tracer to avoid concurrent
	 * changes on it while we are reading.
	 */
2005
	mutex_lock(&trace_types_lock);
2006
	iter->trace = kzalloc(sizeof(*iter->trace), GFP_KERNEL);
2007
	if (!iter->trace)
2008
		goto fail;
2009

2010 2011 2012
	if (current_trace)
		*iter->trace = *current_trace;

2013
	if (!zalloc_cpumask_var(&iter->started, GFP_KERNEL))
2014 2015
		goto fail;

2016 2017 2018
	if (current_trace && current_trace->print_max)
		iter->tr = &max_tr;
	else
2019
		iter->tr = &global_trace;
2020
	iter->pos = -1;
2021
	mutex_init(&iter->mutex);
2022
	iter->cpu_file = cpu_file;
2023

2024 2025
	/* Notify the tracer early; before we stop tracing. */
	if (iter->trace && iter->trace->open)
2026
		iter->trace->open(iter);
2027

2028 2029 2030 2031
	/* Annotate start of buffers if we had overruns */
	if (ring_buffer_overruns(iter->tr->buffer))
		iter->iter_flags |= TRACE_FILE_ANNOTATE;

2032 2033 2034
	/* stop the trace while dumping */
	tracing_stop();

2035 2036
	if (iter->cpu_file == TRACE_PIPE_ALL_CPU) {
		for_each_tracing_cpu(cpu) {
2037

2038 2039
			iter->buffer_iter[cpu] =
				ring_buffer_read_start(iter->tr->buffer, cpu);
2040
			tracing_iter_reset(iter, cpu);
2041 2042 2043
		}
	} else {
		cpu = iter->cpu_file;
2044
		iter->buffer_iter[cpu] =
2045
				ring_buffer_read_start(iter->tr->buffer, cpu);
2046
		tracing_iter_reset(iter, cpu);
2047 2048
	}

2049 2050 2051
	ret = seq_open(file, &tracer_seq_ops);
	if (ret < 0) {
		fail_ret = ERR_PTR(ret);
2052
		goto fail_buffer;
2053
	}
2054

2055 2056
	m = file->private_data;
	m->private = iter;
2057 2058 2059 2060

	mutex_unlock(&trace_types_lock);

	return iter;
2061 2062 2063 2064 2065 2066

 fail_buffer:
	for_each_tracing_cpu(cpu) {
		if (iter->buffer_iter[cpu])
			ring_buffer_read_finish(iter->buffer_iter[cpu]);
	}
2067
	free_cpumask_var(iter->started);
2068
	tracing_start();
2069
 fail:
2070
	mutex_unlock(&trace_types_lock);
2071
	kfree(iter->trace);
2072
	kfree(iter);
2073

2074
	return fail_ret;
2075 2076 2077 2078
}

int tracing_open_generic(struct inode *inode, struct file *filp)
{
S
Steven Rostedt 已提交
2079 2080 2081
	if (tracing_disabled)
		return -ENODEV;

2082 2083 2084 2085
	filp->private_data = inode->i_private;
	return 0;
}

2086
static int tracing_release(struct inode *inode, struct file *file)
2087 2088
{
	struct seq_file *m = (struct seq_file *)file->private_data;
2089
	struct trace_iterator *iter;
2090
	int cpu;
2091

2092 2093 2094 2095 2096
	if (!(file->f_mode & FMODE_READ))
		return 0;

	iter = m->private;

2097
	mutex_lock(&trace_types_lock);
2098 2099 2100 2101 2102
	for_each_tracing_cpu(cpu) {
		if (iter->buffer_iter[cpu])
			ring_buffer_read_finish(iter->buffer_iter[cpu]);
	}

2103 2104 2105 2106
	if (iter->trace && iter->trace->close)
		iter->trace->close(iter);

	/* reenable tracing if it was previously enabled */
2107
	tracing_start();
2108 2109 2110
	mutex_unlock(&trace_types_lock);

	seq_release(inode, file);
2111
	mutex_destroy(&iter->mutex);
2112
	free_cpumask_var(iter->started);
2113
	kfree(iter->trace);
2114 2115 2116 2117 2118 2119
	kfree(iter);
	return 0;
}

static int tracing_open(struct inode *inode, struct file *file)
{
2120 2121
	struct trace_iterator *iter;
	int ret = 0;
2122

2123 2124
	/* If this file was open for write, then erase contents */
	if ((file->f_mode & FMODE_WRITE) &&
2125
	    (file->f_flags & O_TRUNC)) {
2126
		long cpu = (long) inode->i_private;
2127

2128 2129 2130 2131 2132
		if (cpu == TRACE_PIPE_ALL_CPU)
			tracing_reset_online_cpus(&global_trace);
		else
			tracing_reset(&global_trace, cpu);
	}
2133

2134 2135 2136 2137 2138 2139 2140
	if (file->f_mode & FMODE_READ) {
		iter = __tracing_open(inode, file);
		if (IS_ERR(iter))
			ret = PTR_ERR(iter);
		else if (trace_flags & TRACE_ITER_LATENCY_FMT)
			iter->iter_flags |= TRACE_FILE_LAT_FMT;
	}
2141 2142 2143
	return ret;
}

I
Ingo Molnar 已提交
2144
static void *
2145 2146
t_next(struct seq_file *m, void *v, loff_t *pos)
{
L
Li Zefan 已提交
2147
	struct tracer *t = v;
2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158

	(*pos)++;

	if (t)
		t = t->next;

	return t;
}

static void *t_start(struct seq_file *m, loff_t *pos)
{
L
Li Zefan 已提交
2159
	struct tracer *t;
2160 2161 2162
	loff_t l = 0;

	mutex_lock(&trace_types_lock);
L
Li Zefan 已提交
2163
	for (t = trace_types; t && l < *pos; t = t_next(m, t, &l))
2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189
		;

	return t;
}

static void t_stop(struct seq_file *m, void *p)
{
	mutex_unlock(&trace_types_lock);
}

static int t_show(struct seq_file *m, void *v)
{
	struct tracer *t = v;

	if (!t)
		return 0;

	seq_printf(m, "%s", t->name);
	if (t->next)
		seq_putc(m, ' ');
	else
		seq_putc(m, '\n');

	return 0;
}

J
James Morris 已提交
2190
static const struct seq_operations show_traces_seq_ops = {
I
Ingo Molnar 已提交
2191 2192 2193 2194
	.start		= t_start,
	.next		= t_next,
	.stop		= t_stop,
	.show		= t_show,
2195 2196 2197 2198
};

static int show_traces_open(struct inode *inode, struct file *file)
{
S
Steven Rostedt 已提交
2199 2200 2201
	if (tracing_disabled)
		return -ENODEV;

L
Li Zefan 已提交
2202
	return seq_open(file, &show_traces_seq_ops);
2203 2204
}

2205 2206 2207 2208 2209 2210 2211
static ssize_t
tracing_write_stub(struct file *filp, const char __user *ubuf,
		   size_t count, loff_t *ppos)
{
	return count;
}

2212
static const struct file_operations tracing_fops = {
I
Ingo Molnar 已提交
2213 2214
	.open		= tracing_open,
	.read		= seq_read,
2215
	.write		= tracing_write_stub,
I
Ingo Molnar 已提交
2216 2217
	.llseek		= seq_lseek,
	.release	= tracing_release,
2218 2219
};

2220
static const struct file_operations show_traces_fops = {
I
Ingo Molnar 已提交
2221 2222 2223 2224 2225
	.open		= show_traces_open,
	.read		= seq_read,
	.release	= seq_release,
};

2226 2227 2228
/*
 * Only trace on a CPU if the bitmask is set:
 */
2229
static cpumask_var_t tracing_cpumask;
2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242

/*
 * The tracer itself will not take this lock, but still we want
 * to provide a consistent cpumask to user-space:
 */
static DEFINE_MUTEX(tracing_cpumask_update_lock);

/*
 * Temporary storage for the character representation of the
 * CPU bitmask (and one more byte for the newline):
 */
static char mask_str[NR_CPUS + 1];

I
Ingo Molnar 已提交
2243 2244 2245 2246
static ssize_t
tracing_cpumask_read(struct file *filp, char __user *ubuf,
		     size_t count, loff_t *ppos)
{
2247
	int len;
I
Ingo Molnar 已提交
2248 2249

	mutex_lock(&tracing_cpumask_update_lock);
2250

2251
	len = cpumask_scnprintf(mask_str, count, tracing_cpumask);
2252 2253 2254 2255 2256 2257 2258 2259
	if (count - len < 2) {
		count = -EINVAL;
		goto out_err;
	}
	len += sprintf(mask_str + len, "\n");
	count = simple_read_from_buffer(ubuf, count, ppos, mask_str, NR_CPUS+1);

out_err:
I
Ingo Molnar 已提交
2260 2261 2262 2263 2264 2265 2266 2267 2268
	mutex_unlock(&tracing_cpumask_update_lock);

	return count;
}

static ssize_t
tracing_cpumask_write(struct file *filp, const char __user *ubuf,
		      size_t count, loff_t *ppos)
{
2269
	int err, cpu;
2270 2271 2272 2273
	cpumask_var_t tracing_cpumask_new;

	if (!alloc_cpumask_var(&tracing_cpumask_new, GFP_KERNEL))
		return -ENOMEM;
I
Ingo Molnar 已提交
2274

2275
	err = cpumask_parse_user(ubuf, count, tracing_cpumask_new);
I
Ingo Molnar 已提交
2276
	if (err)
2277 2278
		goto err_unlock;

2279 2280
	mutex_lock(&tracing_cpumask_update_lock);

2281
	local_irq_disable();
2282
	arch_spin_lock(&ftrace_max_lock);
2283
	for_each_tracing_cpu(cpu) {
2284 2285 2286 2287
		/*
		 * Increase/decrease the disabled counter if we are
		 * about to flip a bit in the cpumask:
		 */
2288 2289
		if (cpumask_test_cpu(cpu, tracing_cpumask) &&
				!cpumask_test_cpu(cpu, tracing_cpumask_new)) {
2290 2291
			atomic_inc(&global_trace.data[cpu]->disabled);
		}
2292 2293
		if (!cpumask_test_cpu(cpu, tracing_cpumask) &&
				cpumask_test_cpu(cpu, tracing_cpumask_new)) {
2294 2295 2296
			atomic_dec(&global_trace.data[cpu]->disabled);
		}
	}
2297
	arch_spin_unlock(&ftrace_max_lock);
2298
	local_irq_enable();
2299

2300
	cpumask_copy(tracing_cpumask, tracing_cpumask_new);
2301 2302

	mutex_unlock(&tracing_cpumask_update_lock);
2303
	free_cpumask_var(tracing_cpumask_new);
I
Ingo Molnar 已提交
2304 2305

	return count;
2306 2307

err_unlock:
2308
	free_cpumask_var(tracing_cpumask_new);
2309 2310

	return err;
I
Ingo Molnar 已提交
2311 2312
}

2313
static const struct file_operations tracing_cpumask_fops = {
I
Ingo Molnar 已提交
2314 2315 2316
	.open		= tracing_open_generic,
	.read		= tracing_cpumask_read,
	.write		= tracing_cpumask_write,
2317 2318 2319
};

static ssize_t
2320
tracing_trace_options_read(struct file *filp, char __user *ubuf,
2321 2322
		       size_t cnt, loff_t *ppos)
{
2323 2324 2325
	struct tracer_opt *trace_opts;
	u32 tracer_flags;
	int len = 0;
2326 2327
	char *buf;
	int r = 0;
2328
	int i;
2329

2330

W
Wenji Huang 已提交
2331
	/* calculate max size */
2332 2333
	for (i = 0; trace_options[i]; i++) {
		len += strlen(trace_options[i]);
2334
		len += 3; /* "no" and newline */
2335 2336
	}

2337 2338 2339 2340
	mutex_lock(&trace_types_lock);
	tracer_flags = current_trace->flags->val;
	trace_opts = current_trace->flags->opts;

2341 2342 2343 2344 2345 2346
	/*
	 * Increase the size with names of options specific
	 * of the current tracer.
	 */
	for (i = 0; trace_opts[i].name; i++) {
		len += strlen(trace_opts[i].name);
2347
		len += 3; /* "no" and newline */
2348 2349
	}

2350 2351
	/* +1 for \0 */
	buf = kmalloc(len + 1, GFP_KERNEL);
2352 2353
	if (!buf) {
		mutex_unlock(&trace_types_lock);
2354
		return -ENOMEM;
2355
	}
2356 2357 2358

	for (i = 0; trace_options[i]; i++) {
		if (trace_flags & (1 << i))
2359
			r += sprintf(buf + r, "%s\n", trace_options[i]);
2360
		else
2361
			r += sprintf(buf + r, "no%s\n", trace_options[i]);
2362 2363
	}

2364 2365
	for (i = 0; trace_opts[i].name; i++) {
		if (tracer_flags & trace_opts[i].bit)
2366
			r += sprintf(buf + r, "%s\n",
2367 2368
				trace_opts[i].name);
		else
2369
			r += sprintf(buf + r, "no%s\n",
2370 2371
				trace_opts[i].name);
	}
2372
	mutex_unlock(&trace_types_lock);
2373

2374
	WARN_ON(r >= len + 1);
2375

2376
	r = simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
2377 2378 2379 2380 2381

	kfree(buf);
	return r;
}

2382 2383 2384
/* Try to assign a tracer specific option */
static int set_tracer_option(struct tracer *trace, char *cmp, int neg)
{
2385
	struct tracer_flags *tracer_flags = trace->flags;
2386 2387 2388 2389
	struct tracer_opt *opts = NULL;
	int ret = 0, i = 0;
	int len;

2390 2391
	for (i = 0; tracer_flags->opts[i].name; i++) {
		opts = &tracer_flags->opts[i];
2392 2393 2394
		len = strlen(opts->name);

		if (strncmp(cmp, opts->name, len) == 0) {
2395
			ret = trace->set_flag(tracer_flags->val,
2396 2397 2398 2399 2400
				opts->bit, !neg);
			break;
		}
	}
	/* Not found */
2401
	if (!tracer_flags->opts[i].name)
2402 2403 2404 2405 2406 2407 2408
		return -EINVAL;

	/* Refused to handle */
	if (ret)
		return ret;

	if (neg)
2409
		tracer_flags->val &= ~opts->bit;
2410
	else
2411
		tracer_flags->val |= opts->bit;
2412 2413 2414 2415

	return 0;
}

2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427
static void set_tracer_flags(unsigned int mask, int enabled)
{
	/* do nothing if flag is already set */
	if (!!(trace_flags & mask) == !!enabled)
		return;

	if (enabled)
		trace_flags |= mask;
	else
		trace_flags &= ~mask;
}

2428
static ssize_t
2429
tracing_trace_options_write(struct file *filp, const char __user *ubuf,
2430 2431 2432 2433 2434
			size_t cnt, loff_t *ppos)
{
	char buf[64];
	char *cmp = buf;
	int neg = 0;
2435
	int ret;
2436 2437
	int i;

S
Steven Rostedt 已提交
2438 2439
	if (cnt >= sizeof(buf))
		return -EINVAL;
2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454

	if (copy_from_user(&buf, ubuf, cnt))
		return -EFAULT;

	buf[cnt] = 0;

	if (strncmp(buf, "no", 2) == 0) {
		neg = 1;
		cmp += 2;
	}

	for (i = 0; trace_options[i]; i++) {
		int len = strlen(trace_options[i]);

		if (strncmp(cmp, trace_options[i], len) == 0) {
2455
			set_tracer_flags(1 << i, !neg);
2456 2457 2458
			break;
		}
	}
2459 2460 2461

	/* If no option could be set, test the specific tracer options */
	if (!trace_options[i]) {
2462
		mutex_lock(&trace_types_lock);
2463
		ret = set_tracer_option(current_trace, cmp, neg);
2464
		mutex_unlock(&trace_types_lock);
2465 2466 2467
		if (ret)
			return ret;
	}
2468

2469
	*ppos += cnt;
2470 2471 2472 2473

	return cnt;
}

2474
static const struct file_operations tracing_iter_fops = {
I
Ingo Molnar 已提交
2475
	.open		= tracing_open_generic,
2476 2477
	.read		= tracing_trace_options_read,
	.write		= tracing_trace_options_write,
2478 2479
};

I
Ingo Molnar 已提交
2480 2481
static const char readme_msg[] =
	"tracing mini-HOWTO:\n\n"
2482 2483
	"# mount -t debugfs nodev /sys/kernel/debug\n\n"
	"# cat /sys/kernel/debug/tracing/available_tracers\n"
2484
	"wakeup preemptirqsoff preemptoff irqsoff function sched_switch nop\n\n"
2485
	"# cat /sys/kernel/debug/tracing/current_tracer\n"
2486
	"nop\n"
2487 2488
	"# echo sched_switch > /sys/kernel/debug/tracing/current_tracer\n"
	"# cat /sys/kernel/debug/tracing/current_tracer\n"
I
Ingo Molnar 已提交
2489
	"sched_switch\n"
2490
	"# cat /sys/kernel/debug/tracing/trace_options\n"
I
Ingo Molnar 已提交
2491
	"noprint-parent nosym-offset nosym-addr noverbose\n"
2492 2493 2494 2495
	"# echo print-parent > /sys/kernel/debug/tracing/trace_options\n"
	"# echo 1 > /sys/kernel/debug/tracing/tracing_enabled\n"
	"# cat /sys/kernel/debug/tracing/trace > /tmp/trace.txt\n"
	"# echo 0 > /sys/kernel/debug/tracing/tracing_enabled\n"
I
Ingo Molnar 已提交
2496 2497 2498 2499 2500 2501 2502 2503 2504 2505
;

static ssize_t
tracing_readme_read(struct file *filp, char __user *ubuf,
		       size_t cnt, loff_t *ppos)
{
	return simple_read_from_buffer(ubuf, cnt, ppos,
					readme_msg, strlen(readme_msg));
}

2506
static const struct file_operations tracing_readme_fops = {
I
Ingo Molnar 已提交
2507 2508
	.open		= tracing_open_generic,
	.read		= tracing_readme_read,
I
Ingo Molnar 已提交
2509 2510
};

2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560
static ssize_t
tracing_saved_cmdlines_read(struct file *file, char __user *ubuf,
				size_t cnt, loff_t *ppos)
{
	char *buf_comm;
	char *file_buf;
	char *buf;
	int len = 0;
	int pid;
	int i;

	file_buf = kmalloc(SAVED_CMDLINES*(16+TASK_COMM_LEN), GFP_KERNEL);
	if (!file_buf)
		return -ENOMEM;

	buf_comm = kmalloc(TASK_COMM_LEN, GFP_KERNEL);
	if (!buf_comm) {
		kfree(file_buf);
		return -ENOMEM;
	}

	buf = file_buf;

	for (i = 0; i < SAVED_CMDLINES; i++) {
		int r;

		pid = map_cmdline_to_pid[i];
		if (pid == -1 || pid == NO_CMDLINE_MAP)
			continue;

		trace_find_cmdline(pid, buf_comm);
		r = sprintf(buf, "%d %s\n", pid, buf_comm);
		buf += r;
		len += r;
	}

	len = simple_read_from_buffer(ubuf, cnt, ppos,
				      file_buf, len);

	kfree(file_buf);
	kfree(buf_comm);

	return len;
}

static const struct file_operations tracing_saved_cmdlines_fops = {
    .open       = tracing_open_generic,
    .read       = tracing_saved_cmdlines_read,
};

2561 2562 2563 2564 2565 2566 2567
static ssize_t
tracing_ctrl_read(struct file *filp, char __user *ubuf,
		  size_t cnt, loff_t *ppos)
{
	char buf[64];
	int r;

2568
	r = sprintf(buf, "%u\n", tracer_enabled);
I
Ingo Molnar 已提交
2569
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
2570 2571 2572 2573 2574 2575 2576 2577
}

static ssize_t
tracing_ctrl_write(struct file *filp, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
	struct trace_array *tr = filp->private_data;
	char buf[64];
2578
	unsigned long val;
2579
	int ret;
2580

S
Steven Rostedt 已提交
2581 2582
	if (cnt >= sizeof(buf))
		return -EINVAL;
2583 2584 2585 2586 2587 2588

	if (copy_from_user(&buf, ubuf, cnt))
		return -EFAULT;

	buf[cnt] = 0;

2589 2590 2591
	ret = strict_strtoul(buf, 10, &val);
	if (ret < 0)
		return ret;
2592 2593 2594 2595

	val = !!val;

	mutex_lock(&trace_types_lock);
2596 2597
	if (tracer_enabled ^ val) {
		if (val) {
2598
			tracer_enabled = 1;
2599 2600 2601 2602
			if (current_trace->start)
				current_trace->start(tr);
			tracing_start();
		} else {
2603
			tracer_enabled = 0;
2604 2605 2606 2607
			tracing_stop();
			if (current_trace->stop)
				current_trace->stop(tr);
		}
2608 2609 2610
	}
	mutex_unlock(&trace_types_lock);

2611
	*ppos += cnt;
2612 2613 2614 2615 2616 2617 2618 2619

	return cnt;
}

static ssize_t
tracing_set_trace_read(struct file *filp, char __user *ubuf,
		       size_t cnt, loff_t *ppos)
{
L
Li Zefan 已提交
2620
	char buf[MAX_TRACER_SIZE+2];
2621 2622 2623 2624 2625 2626 2627 2628 2629
	int r;

	mutex_lock(&trace_types_lock);
	if (current_trace)
		r = sprintf(buf, "%s\n", current_trace->name);
	else
		r = sprintf(buf, "\n");
	mutex_unlock(&trace_types_lock);

I
Ingo Molnar 已提交
2630
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
2631 2632
}

2633 2634 2635 2636 2637 2638
int tracer_init(struct tracer *t, struct trace_array *tr)
{
	tracing_reset_online_cpus(tr);
	return t->init(tr);
}

2639 2640 2641 2642 2643 2644
static int tracing_resize_ring_buffer(unsigned long size)
{
	int ret;

	/*
	 * If kernel or user changes the size of the ring buffer
2645 2646
	 * we use the size that was given, and we can forget about
	 * expanding it later.
2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660
	 */
	ring_buffer_expanded = 1;

	ret = ring_buffer_resize(global_trace.buffer, size);
	if (ret < 0)
		return ret;

	ret = ring_buffer_resize(max_tr.buffer, size);
	if (ret < 0) {
		int r;

		r = ring_buffer_resize(global_trace.buffer,
				       global_trace.entries);
		if (r < 0) {
2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674
			/*
			 * AARGH! We are left with different
			 * size max buffer!!!!
			 * The max buffer is our "snapshot" buffer.
			 * When a tracer needs a snapshot (one of the
			 * latency tracers), it swaps the max buffer
			 * with the saved snap shot. We succeeded to
			 * update the size of the main buffer, but failed to
			 * update the size of the max buffer. But when we tried
			 * to reset the main buffer to the original size, we
			 * failed there too. This is very unlikely to
			 * happen, but if it does, warn and kill all
			 * tracing.
			 */
2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685
			WARN_ON(1);
			tracing_disabled = 1;
		}
		return ret;
	}

	global_trace.entries = size;

	return ret;
}

2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699
/**
 * tracing_update_buffers - used by tracing facility to expand ring buffers
 *
 * To save on memory when the tracing is never used on a system with it
 * configured in. The ring buffers are set to a minimum size. But once
 * a user starts to use the tracing facility, then they need to grow
 * to their default size.
 *
 * This function is to be called when a tracer is about to be used.
 */
int tracing_update_buffers(void)
{
	int ret = 0;

2700
	mutex_lock(&trace_types_lock);
2701 2702
	if (!ring_buffer_expanded)
		ret = tracing_resize_ring_buffer(trace_buf_size);
2703
	mutex_unlock(&trace_types_lock);
2704 2705 2706 2707

	return ret;
}

2708 2709 2710 2711 2712 2713 2714 2715
struct trace_option_dentry;

static struct trace_option_dentry *
create_trace_option_files(struct tracer *tracer);

static void
destroy_trace_option_files(struct trace_option_dentry *topts);

2716
static int tracing_set_tracer(const char *buf)
2717
{
2718
	static struct trace_option_dentry *topts;
2719 2720
	struct trace_array *tr = &global_trace;
	struct tracer *t;
2721
	int ret = 0;
2722

2723 2724
	mutex_lock(&trace_types_lock);

2725 2726 2727
	if (!ring_buffer_expanded) {
		ret = tracing_resize_ring_buffer(trace_buf_size);
		if (ret < 0)
2728
			goto out;
2729 2730 2731
		ret = 0;
	}

2732 2733 2734 2735
	for (t = trace_types; t; t = t->next) {
		if (strcmp(t->name, buf) == 0)
			break;
	}
2736 2737 2738 2739 2740
	if (!t) {
		ret = -EINVAL;
		goto out;
	}
	if (t == current_trace)
2741 2742
		goto out;

2743
	trace_branch_disable();
2744 2745 2746
	if (current_trace && current_trace->reset)
		current_trace->reset(tr);

2747 2748
	destroy_trace_option_files(topts);

2749
	current_trace = t;
2750 2751 2752

	topts = create_trace_option_files(current_trace);

2753
	if (t->init) {
2754
		ret = tracer_init(t, tr);
2755 2756 2757
		if (ret)
			goto out;
	}
2758

2759
	trace_branch_enable(tr);
2760 2761 2762
 out:
	mutex_unlock(&trace_types_lock);

2763 2764 2765 2766 2767 2768 2769
	return ret;
}

static ssize_t
tracing_set_trace_write(struct file *filp, const char __user *ubuf,
			size_t cnt, loff_t *ppos)
{
L
Li Zefan 已提交
2770
	char buf[MAX_TRACER_SIZE+1];
2771 2772
	int i;
	size_t ret;
2773 2774 2775
	int err;

	ret = cnt;
2776

L
Li Zefan 已提交
2777 2778
	if (cnt > MAX_TRACER_SIZE)
		cnt = MAX_TRACER_SIZE;
2779 2780 2781 2782 2783 2784 2785 2786 2787 2788

	if (copy_from_user(&buf, ubuf, cnt))
		return -EFAULT;

	buf[cnt] = 0;

	/* strip ending whitespace. */
	for (i = cnt - 1; i > 0 && isspace(buf[i]); i--)
		buf[i] = 0;

2789 2790 2791
	err = tracing_set_tracer(buf);
	if (err)
		return err;
2792

2793
	*ppos += ret;
2794

2795
	return ret;
2796 2797 2798 2799 2800 2801 2802 2803 2804 2805
}

static ssize_t
tracing_max_lat_read(struct file *filp, char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
	unsigned long *ptr = filp->private_data;
	char buf[64];
	int r;

S
Steven Rostedt 已提交
2806
	r = snprintf(buf, sizeof(buf), "%ld\n",
2807
		     *ptr == (unsigned long)-1 ? -1 : nsecs_to_usecs(*ptr));
S
Steven Rostedt 已提交
2808 2809
	if (r > sizeof(buf))
		r = sizeof(buf);
I
Ingo Molnar 已提交
2810
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
2811 2812 2813 2814 2815 2816
}

static ssize_t
tracing_max_lat_write(struct file *filp, const char __user *ubuf,
		      size_t cnt, loff_t *ppos)
{
2817
	unsigned long *ptr = filp->private_data;
2818
	char buf[64];
2819
	unsigned long val;
2820
	int ret;
2821

S
Steven Rostedt 已提交
2822 2823
	if (cnt >= sizeof(buf))
		return -EINVAL;
2824 2825 2826 2827 2828 2829

	if (copy_from_user(&buf, ubuf, cnt))
		return -EFAULT;

	buf[cnt] = 0;

2830 2831 2832
	ret = strict_strtoul(buf, 10, &val);
	if (ret < 0)
		return ret;
2833 2834 2835 2836 2837 2838

	*ptr = val * 1000;

	return cnt;
}

2839 2840
static int tracing_open_pipe(struct inode *inode, struct file *filp)
{
2841
	long cpu_file = (long) inode->i_private;
2842
	struct trace_iterator *iter;
2843
	int ret = 0;
2844 2845 2846 2847

	if (tracing_disabled)
		return -ENODEV;

2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863
	mutex_lock(&trace_types_lock);

	/* We only allow one reader per cpu */
	if (cpu_file == TRACE_PIPE_ALL_CPU) {
		if (!cpumask_empty(tracing_reader_cpumask)) {
			ret = -EBUSY;
			goto out;
		}
		cpumask_setall(tracing_reader_cpumask);
	} else {
		if (!cpumask_test_cpu(cpu_file, tracing_reader_cpumask))
			cpumask_set_cpu(cpu_file, tracing_reader_cpumask);
		else {
			ret = -EBUSY;
			goto out;
		}
2864 2865 2866 2867
	}

	/* create a buffer to store the information to pass to userspace */
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
2868 2869 2870 2871
	if (!iter) {
		ret = -ENOMEM;
		goto out;
	}
2872

2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884
	/*
	 * We make a copy of the current tracer to avoid concurrent
	 * changes on it while we are reading.
	 */
	iter->trace = kmalloc(sizeof(*iter->trace), GFP_KERNEL);
	if (!iter->trace) {
		ret = -ENOMEM;
		goto fail;
	}
	if (current_trace)
		*iter->trace = *current_trace;

2885
	if (!alloc_cpumask_var(&iter->started, GFP_KERNEL)) {
2886
		ret = -ENOMEM;
2887
		goto fail;
2888 2889
	}

2890
	/* trace pipe does not show start of buffer */
2891
	cpumask_setall(iter->started);
2892

2893 2894 2895
	if (trace_flags & TRACE_ITER_LATENCY_FMT)
		iter->iter_flags |= TRACE_FILE_LAT_FMT;

2896
	iter->cpu_file = cpu_file;
2897
	iter->tr = &global_trace;
2898
	mutex_init(&iter->mutex);
2899 2900
	filp->private_data = iter;

2901 2902 2903
	if (iter->trace->pipe_open)
		iter->trace->pipe_open(iter);

2904 2905 2906
out:
	mutex_unlock(&trace_types_lock);
	return ret;
2907 2908 2909 2910 2911 2912

fail:
	kfree(iter->trace);
	kfree(iter);
	mutex_unlock(&trace_types_lock);
	return ret;
2913 2914 2915 2916 2917 2918
}

static int tracing_release_pipe(struct inode *inode, struct file *file)
{
	struct trace_iterator *iter = file->private_data;

2919 2920 2921 2922 2923 2924 2925
	mutex_lock(&trace_types_lock);

	if (iter->cpu_file == TRACE_PIPE_ALL_CPU)
		cpumask_clear(tracing_reader_cpumask);
	else
		cpumask_clear_cpu(iter->cpu_file, tracing_reader_cpumask);

S
Steven Rostedt 已提交
2926

2927
	if (iter->trace->pipe_close)
S
Steven Rostedt 已提交
2928 2929
		iter->trace->pipe_close(iter);

2930 2931
	mutex_unlock(&trace_types_lock);

2932
	free_cpumask_var(iter->started);
2933 2934
	mutex_destroy(&iter->mutex);
	kfree(iter->trace);
2935 2936 2937 2938 2939
	kfree(iter);

	return 0;
}

2940 2941 2942 2943 2944 2945 2946 2947 2948 2949
static unsigned int
tracing_poll_pipe(struct file *filp, poll_table *poll_table)
{
	struct trace_iterator *iter = filp->private_data;

	if (trace_flags & TRACE_ITER_BLOCK) {
		/*
		 * Always select as readable when in blocking mode
		 */
		return POLLIN | POLLRDNORM;
I
Ingo Molnar 已提交
2950
	} else {
2951 2952 2953 2954 2955 2956 2957 2958 2959 2960
		if (!trace_empty(iter))
			return POLLIN | POLLRDNORM;
		poll_wait(filp, &trace_wait, poll_table);
		if (!trace_empty(iter))
			return POLLIN | POLLRDNORM;

		return 0;
	}
}

2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992

void default_wait_pipe(struct trace_iterator *iter)
{
	DEFINE_WAIT(wait);

	prepare_to_wait(&trace_wait, &wait, TASK_INTERRUPTIBLE);

	if (trace_empty(iter))
		schedule();

	finish_wait(&trace_wait, &wait);
}

/*
 * This is a make-shift waitqueue.
 * A tracer might use this callback on some rare cases:
 *
 *  1) the current tracer might hold the runqueue lock when it wakes up
 *     a reader, hence a deadlock (sched, function, and function graph tracers)
 *  2) the function tracers, trace all functions, we don't want
 *     the overhead of calling wake_up and friends
 *     (and tracing them too)
 *
 *     Anyway, this is really very primitive wakeup.
 */
void poll_wait_pipe(struct trace_iterator *iter)
{
	set_current_state(TASK_INTERRUPTIBLE);
	/* sleep for 100 msecs, and try again. */
	schedule_timeout(HZ / 10);
}

2993 2994
/* Must be called with trace_types_lock mutex held. */
static int tracing_wait_pipe(struct file *filp)
2995 2996 2997 2998
{
	struct trace_iterator *iter = filp->private_data;

	while (trace_empty(iter)) {
2999

3000
		if ((filp->f_flags & O_NONBLOCK)) {
3001
			return -EAGAIN;
3002
		}
3003

3004
		mutex_unlock(&iter->mutex);
3005

3006
		iter->trace->wait_pipe(iter);
3007

3008
		mutex_lock(&iter->mutex);
3009

3010
		if (signal_pending(current))
3011
			return -EINTR;
3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025

		/*
		 * We block until we read something and tracing is disabled.
		 * We still block if tracing is disabled, but we have never
		 * read anything. This allows a user to cat this file, and
		 * then enable tracing. But after we have read something,
		 * we give an EOF when tracing is again disabled.
		 *
		 * iter->pos will be 0 if we haven't read anything.
		 */
		if (!tracer_enabled && iter->pos)
			break;
	}

3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036
	return 1;
}

/*
 * Consumer reader.
 */
static ssize_t
tracing_read_pipe(struct file *filp, char __user *ubuf,
		  size_t cnt, loff_t *ppos)
{
	struct trace_iterator *iter = filp->private_data;
3037
	static struct tracer *old_tracer;
3038 3039 3040 3041 3042 3043 3044
	ssize_t sret;

	/* return any leftover data */
	sret = trace_seq_to_user(&iter->seq, ubuf, cnt);
	if (sret != -EBUSY)
		return sret;

3045
	trace_seq_init(&iter->seq);
3046

3047
	/* copy the tracer to avoid using a global lock all around */
3048
	mutex_lock(&trace_types_lock);
3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060
	if (unlikely(old_tracer != current_trace && current_trace)) {
		old_tracer = current_trace;
		*iter->trace = *current_trace;
	}
	mutex_unlock(&trace_types_lock);

	/*
	 * Avoid more than one consumer on a single file descriptor
	 * This is just a matter of traces coherency, the ring buffer itself
	 * is protected.
	 */
	mutex_lock(&iter->mutex);
3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071
	if (iter->trace->read) {
		sret = iter->trace->read(iter, filp, ubuf, cnt, ppos);
		if (sret)
			goto out;
	}

waitagain:
	sret = tracing_wait_pipe(filp);
	if (sret <= 0)
		goto out;

3072
	/* stop when tracing is finished */
3073 3074
	if (trace_empty(iter)) {
		sret = 0;
3075
		goto out;
3076
	}
3077 3078 3079 3080

	if (cnt >= PAGE_SIZE)
		cnt = PAGE_SIZE - 1;

3081 3082 3083 3084
	/* reset all but tr, trace, and overruns */
	memset(&iter->seq, 0,
	       sizeof(struct trace_iterator) -
	       offsetof(struct trace_iterator, seq));
3085
	iter->pos = -1;
3086

3087
	trace_event_read_lock();
S
Steven Rostedt 已提交
3088
	while (find_next_entry_inc(iter) != NULL) {
3089
		enum print_line_t ret;
S
Steven Rostedt 已提交
3090 3091
		int len = iter->seq.len;

I
Ingo Molnar 已提交
3092
		ret = print_trace_line(iter);
3093
		if (ret == TRACE_TYPE_PARTIAL_LINE) {
S
Steven Rostedt 已提交
3094 3095
			/* don't print partial lines */
			iter->seq.len = len;
3096
			break;
S
Steven Rostedt 已提交
3097
		}
3098 3099
		if (ret != TRACE_TYPE_NO_CONSUME)
			trace_consume(iter);
3100 3101 3102 3103

		if (iter->seq.len >= cnt)
			break;
	}
3104
	trace_event_read_unlock();
3105 3106

	/* Now copy what we have to the user */
3107 3108
	sret = trace_seq_to_user(&iter->seq, ubuf, cnt);
	if (iter->seq.readpos >= iter->seq.len)
3109
		trace_seq_init(&iter->seq);
P
Pekka Paalanen 已提交
3110 3111 3112 3113 3114

	/*
	 * If there was nothing to send to user, inspite of consuming trace
	 * entries, go back to wait for more entries.
	 */
3115
	if (sret == -EBUSY)
P
Pekka Paalanen 已提交
3116
		goto waitagain;
3117

3118
out:
3119
	mutex_unlock(&iter->mutex);
3120

3121
	return sret;
3122 3123
}

3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136
static void tracing_pipe_buf_release(struct pipe_inode_info *pipe,
				     struct pipe_buffer *buf)
{
	__free_page(buf->page);
}

static void tracing_spd_release_pipe(struct splice_pipe_desc *spd,
				     unsigned int idx)
{
	__free_page(spd->pages[idx]);
}

static struct pipe_buf_operations tracing_pipe_buf_ops = {
S
Steven Rostedt 已提交
3137 3138 3139 3140 3141 3142 3143
	.can_merge		= 0,
	.map			= generic_pipe_buf_map,
	.unmap			= generic_pipe_buf_unmap,
	.confirm		= generic_pipe_buf_confirm,
	.release		= tracing_pipe_buf_release,
	.steal			= generic_pipe_buf_steal,
	.get			= generic_pipe_buf_get,
3144 3145
};

S
Steven Rostedt 已提交
3146
static size_t
3147
tracing_fill_pipe_page(size_t rem, struct trace_iterator *iter)
S
Steven Rostedt 已提交
3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166
{
	size_t count;
	int ret;

	/* Seq buffer is page-sized, exactly what we need. */
	for (;;) {
		count = iter->seq.len;
		ret = print_trace_line(iter);
		count = iter->seq.len - count;
		if (rem < count) {
			rem = 0;
			iter->seq.len -= count;
			break;
		}
		if (ret == TRACE_TYPE_PARTIAL_LINE) {
			iter->seq.len -= count;
			break;
		}

3167 3168
		if (ret != TRACE_TYPE_NO_CONSUME)
			trace_consume(iter);
S
Steven Rostedt 已提交
3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179
		rem -= count;
		if (!find_next_entry_inc(iter))	{
			rem = 0;
			iter->ent = NULL;
			break;
		}
	}

	return rem;
}

3180 3181 3182 3183 3184 3185 3186 3187 3188 3189
static ssize_t tracing_splice_read_pipe(struct file *filp,
					loff_t *ppos,
					struct pipe_inode_info *pipe,
					size_t len,
					unsigned int flags)
{
	struct page *pages[PIPE_BUFFERS];
	struct partial_page partial[PIPE_BUFFERS];
	struct trace_iterator *iter = filp->private_data;
	struct splice_pipe_desc spd = {
S
Steven Rostedt 已提交
3190 3191 3192 3193 3194 3195
		.pages		= pages,
		.partial	= partial,
		.nr_pages	= 0, /* This gets updated below. */
		.flags		= flags,
		.ops		= &tracing_pipe_buf_ops,
		.spd_release	= tracing_spd_release_pipe,
3196
	};
3197
	static struct tracer *old_tracer;
3198
	ssize_t ret;
S
Steven Rostedt 已提交
3199
	size_t rem;
3200 3201
	unsigned int i;

3202
	/* copy the tracer to avoid using a global lock all around */
3203
	mutex_lock(&trace_types_lock);
3204 3205 3206 3207 3208 3209 3210
	if (unlikely(old_tracer != current_trace && current_trace)) {
		old_tracer = current_trace;
		*iter->trace = *current_trace;
	}
	mutex_unlock(&trace_types_lock);

	mutex_lock(&iter->mutex);
3211 3212 3213 3214 3215

	if (iter->trace->splice_read) {
		ret = iter->trace->splice_read(iter, filp,
					       ppos, pipe, len, flags);
		if (ret)
S
Steven Rostedt 已提交
3216
			goto out_err;
3217 3218 3219 3220
	}

	ret = tracing_wait_pipe(filp);
	if (ret <= 0)
S
Steven Rostedt 已提交
3221
		goto out_err;
3222 3223 3224

	if (!iter->ent && !find_next_entry_inc(iter)) {
		ret = -EFAULT;
S
Steven Rostedt 已提交
3225
		goto out_err;
3226 3227
	}

3228 3229
	trace_event_read_lock();

3230 3231 3232
	/* Fill as many pages as possible. */
	for (i = 0, rem = len; i < PIPE_BUFFERS && rem; i++) {
		pages[i] = alloc_page(GFP_KERNEL);
S
Steven Rostedt 已提交
3233 3234
		if (!pages[i])
			break;
3235

3236
		rem = tracing_fill_pipe_page(rem, iter);
3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248

		/* Copy the data into the page, so we can start over. */
		ret = trace_seq_to_buffer(&iter->seq,
					  page_address(pages[i]),
					  iter->seq.len);
		if (ret < 0) {
			__free_page(pages[i]);
			break;
		}
		partial[i].offset = 0;
		partial[i].len = iter->seq.len;

3249
		trace_seq_init(&iter->seq);
3250 3251
	}

3252
	trace_event_read_unlock();
3253
	mutex_unlock(&iter->mutex);
3254 3255 3256 3257 3258

	spd.nr_pages = i;

	return splice_to_pipe(pipe, &spd);

S
Steven Rostedt 已提交
3259
out_err:
3260
	mutex_unlock(&iter->mutex);
3261 3262 3263 3264

	return ret;
}

3265 3266 3267 3268 3269
static ssize_t
tracing_entries_read(struct file *filp, char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
	struct trace_array *tr = filp->private_data;
3270
	char buf[96];
3271 3272
	int r;

3273 3274 3275 3276 3277 3278 3279 3280 3281
	mutex_lock(&trace_types_lock);
	if (!ring_buffer_expanded)
		r = sprintf(buf, "%lu (expanded: %lu)\n",
			    tr->entries >> 10,
			    trace_buf_size >> 10);
	else
		r = sprintf(buf, "%lu\n", tr->entries >> 10);
	mutex_unlock(&trace_types_lock);

3282 3283 3284 3285 3286 3287 3288 3289 3290
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
}

static ssize_t
tracing_entries_write(struct file *filp, const char __user *ubuf,
		      size_t cnt, loff_t *ppos)
{
	unsigned long val;
	char buf[64];
S
Steven Rostedt 已提交
3291
	int ret, cpu;
3292

S
Steven Rostedt 已提交
3293 3294
	if (cnt >= sizeof(buf))
		return -EINVAL;
3295 3296 3297 3298 3299 3300

	if (copy_from_user(&buf, ubuf, cnt))
		return -EFAULT;

	buf[cnt] = 0;

3301 3302 3303
	ret = strict_strtoul(buf, 10, &val);
	if (ret < 0)
		return ret;
3304 3305 3306 3307 3308 3309 3310

	/* must have at least 1 entry */
	if (!val)
		return -EINVAL;

	mutex_lock(&trace_types_lock);

S
Steven Rostedt 已提交
3311
	tracing_stop();
3312

S
Steven Rostedt 已提交
3313 3314 3315 3316 3317 3318 3319 3320
	/* disable all cpu buffers */
	for_each_tracing_cpu(cpu) {
		if (global_trace.data[cpu])
			atomic_inc(&global_trace.data[cpu]->disabled);
		if (max_tr.data[cpu])
			atomic_inc(&max_tr.data[cpu]->disabled);
	}

3321 3322 3323
	/* value is in KB */
	val <<= 10;

3324
	if (val != global_trace.entries) {
3325
		ret = tracing_resize_ring_buffer(val);
3326 3327
		if (ret < 0) {
			cnt = ret;
S
Steven Rostedt 已提交
3328 3329
			goto out;
		}
3330 3331
	}

3332
	*ppos += cnt;
3333

3334 3335 3336
	/* If check pages failed, return ENOMEM */
	if (tracing_disabled)
		cnt = -ENOMEM;
3337
 out:
S
Steven Rostedt 已提交
3338 3339 3340 3341 3342 3343 3344
	for_each_tracing_cpu(cpu) {
		if (global_trace.data[cpu])
			atomic_dec(&global_trace.data[cpu]->disabled);
		if (max_tr.data[cpu])
			atomic_dec(&max_tr.data[cpu]->disabled);
	}

S
Steven Rostedt 已提交
3345
	tracing_start();
3346 3347 3348 3349 3350 3351
	max_tr.entries = global_trace.entries;
	mutex_unlock(&trace_types_lock);

	return cnt;
}

3352 3353 3354 3355 3356 3357 3358 3359 3360 3361
static int mark_printk(const char *fmt, ...)
{
	int ret;
	va_list args;
	va_start(args, fmt);
	ret = trace_vprintk(0, fmt, args);
	va_end(args);
	return ret;
}

3362 3363 3364 3365 3366 3367
static ssize_t
tracing_mark_write(struct file *filp, const char __user *ubuf,
					size_t cnt, loff_t *fpos)
{
	char *buf;

S
Steven Rostedt 已提交
3368
	if (tracing_disabled)
3369 3370 3371 3372 3373
		return -EINVAL;

	if (cnt > TRACE_BUF_SIZE)
		cnt = TRACE_BUF_SIZE;

C
Carsten Emde 已提交
3374
	buf = kmalloc(cnt + 2, GFP_KERNEL);
3375 3376 3377 3378 3379 3380 3381
	if (buf == NULL)
		return -ENOMEM;

	if (copy_from_user(buf, ubuf, cnt)) {
		kfree(buf);
		return -EFAULT;
	}
C
Carsten Emde 已提交
3382 3383 3384 3385 3386
	if (buf[cnt-1] != '\n') {
		buf[cnt] = '\n';
		buf[cnt+1] = '\0';
	} else
		buf[cnt] = '\0';
3387

3388
	cnt = mark_printk("%s", buf);
3389 3390 3391 3392 3393 3394
	kfree(buf);
	*fpos += cnt;

	return cnt;
}

3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450
static ssize_t tracing_clock_read(struct file *filp, char __user *ubuf,
				  size_t cnt, loff_t *ppos)
{
	char buf[64];
	int bufiter = 0;
	int i;

	for (i = 0; i < ARRAY_SIZE(trace_clocks); i++)
		bufiter += snprintf(buf + bufiter, sizeof(buf) - bufiter,
			"%s%s%s%s", i ? " " : "",
			i == trace_clock_id ? "[" : "", trace_clocks[i].name,
			i == trace_clock_id ? "]" : "");
	bufiter += snprintf(buf + bufiter, sizeof(buf) - bufiter, "\n");

	return simple_read_from_buffer(ubuf, cnt, ppos, buf, bufiter);
}

static ssize_t tracing_clock_write(struct file *filp, const char __user *ubuf,
				   size_t cnt, loff_t *fpos)
{
	char buf[64];
	const char *clockstr;
	int i;

	if (cnt >= sizeof(buf))
		return -EINVAL;

	if (copy_from_user(&buf, ubuf, cnt))
		return -EFAULT;

	buf[cnt] = 0;

	clockstr = strstrip(buf);

	for (i = 0; i < ARRAY_SIZE(trace_clocks); i++) {
		if (strcmp(trace_clocks[i].name, clockstr) == 0)
			break;
	}
	if (i == ARRAY_SIZE(trace_clocks))
		return -EINVAL;

	trace_clock_id = i;

	mutex_lock(&trace_types_lock);

	ring_buffer_set_clock(global_trace.buffer, trace_clocks[i].func);
	if (max_tr.buffer)
		ring_buffer_set_clock(max_tr.buffer, trace_clocks[i].func);

	mutex_unlock(&trace_types_lock);

	*fpos += cnt;

	return cnt;
}

3451
static const struct file_operations tracing_max_lat_fops = {
I
Ingo Molnar 已提交
3452 3453 3454
	.open		= tracing_open_generic,
	.read		= tracing_max_lat_read,
	.write		= tracing_max_lat_write,
3455 3456
};

3457
static const struct file_operations tracing_ctrl_fops = {
I
Ingo Molnar 已提交
3458 3459 3460
	.open		= tracing_open_generic,
	.read		= tracing_ctrl_read,
	.write		= tracing_ctrl_write,
3461 3462
};

3463
static const struct file_operations set_tracer_fops = {
I
Ingo Molnar 已提交
3464 3465 3466
	.open		= tracing_open_generic,
	.read		= tracing_set_trace_read,
	.write		= tracing_set_trace_write,
3467 3468
};

3469
static const struct file_operations tracing_pipe_fops = {
I
Ingo Molnar 已提交
3470
	.open		= tracing_open_pipe,
3471
	.poll		= tracing_poll_pipe,
I
Ingo Molnar 已提交
3472
	.read		= tracing_read_pipe,
3473
	.splice_read	= tracing_splice_read_pipe,
I
Ingo Molnar 已提交
3474
	.release	= tracing_release_pipe,
3475 3476
};

3477
static const struct file_operations tracing_entries_fops = {
3478 3479 3480 3481 3482
	.open		= tracing_open_generic,
	.read		= tracing_entries_read,
	.write		= tracing_entries_write,
};

3483
static const struct file_operations tracing_mark_fops = {
3484
	.open		= tracing_open_generic,
3485 3486 3487
	.write		= tracing_mark_write,
};

3488 3489 3490 3491 3492 3493
static const struct file_operations trace_clock_fops = {
	.open		= tracing_open_generic,
	.read		= tracing_clock_read,
	.write		= tracing_clock_write,
};

3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514
struct ftrace_buffer_info {
	struct trace_array	*tr;
	void			*spare;
	int			cpu;
	unsigned int		read;
};

static int tracing_buffers_open(struct inode *inode, struct file *filp)
{
	int cpu = (int)(long)inode->i_private;
	struct ftrace_buffer_info *info;

	if (tracing_disabled)
		return -ENODEV;

	info = kzalloc(sizeof(*info), GFP_KERNEL);
	if (!info)
		return -ENOMEM;

	info->tr	= &global_trace;
	info->cpu	= cpu;
3515
	info->spare	= NULL;
3516 3517 3518 3519 3520
	/* Force reading ring buffer for first read */
	info->read	= (unsigned int)-1;

	filp->private_data = info;

3521
	return nonseekable_open(inode, filp);
3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532
}

static ssize_t
tracing_buffers_read(struct file *filp, char __user *ubuf,
		     size_t count, loff_t *ppos)
{
	struct ftrace_buffer_info *info = filp->private_data;
	unsigned int pos;
	ssize_t ret;
	size_t size;

3533 3534 3535
	if (!count)
		return 0;

3536 3537 3538 3539 3540
	if (!info->spare)
		info->spare = ring_buffer_alloc_read_page(info->tr->buffer);
	if (!info->spare)
		return -ENOMEM;

3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564
	/* Do we have previous read data to read? */
	if (info->read < PAGE_SIZE)
		goto read;

	info->read = 0;

	ret = ring_buffer_read_page(info->tr->buffer,
				    &info->spare,
				    count,
				    info->cpu, 0);
	if (ret < 0)
		return 0;

	pos = ring_buffer_page_len(info->spare);

	if (pos < PAGE_SIZE)
		memset(info->spare + pos, 0, PAGE_SIZE - pos);

read:
	size = PAGE_SIZE - info->read;
	if (size > count)
		size = count;

	ret = copy_to_user(ubuf, info->spare + info->read, size);
3565
	if (ret == size)
3566
		return -EFAULT;
3567 3568
	size -= ret;

3569 3570 3571 3572 3573 3574 3575 3576 3577 3578
	*ppos += size;
	info->read += size;

	return size;
}

static int tracing_buffers_release(struct inode *inode, struct file *file)
{
	struct ftrace_buffer_info *info = file->private_data;

3579 3580
	if (info->spare)
		ring_buffer_free_read_page(info->tr->buffer, info->spare);
3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662
	kfree(info);

	return 0;
}

struct buffer_ref {
	struct ring_buffer	*buffer;
	void			*page;
	int			ref;
};

static void buffer_pipe_buf_release(struct pipe_inode_info *pipe,
				    struct pipe_buffer *buf)
{
	struct buffer_ref *ref = (struct buffer_ref *)buf->private;

	if (--ref->ref)
		return;

	ring_buffer_free_read_page(ref->buffer, ref->page);
	kfree(ref);
	buf->private = 0;
}

static int buffer_pipe_buf_steal(struct pipe_inode_info *pipe,
				 struct pipe_buffer *buf)
{
	return 1;
}

static void buffer_pipe_buf_get(struct pipe_inode_info *pipe,
				struct pipe_buffer *buf)
{
	struct buffer_ref *ref = (struct buffer_ref *)buf->private;

	ref->ref++;
}

/* Pipe buffer operations for a buffer. */
static struct pipe_buf_operations buffer_pipe_buf_ops = {
	.can_merge		= 0,
	.map			= generic_pipe_buf_map,
	.unmap			= generic_pipe_buf_unmap,
	.confirm		= generic_pipe_buf_confirm,
	.release		= buffer_pipe_buf_release,
	.steal			= buffer_pipe_buf_steal,
	.get			= buffer_pipe_buf_get,
};

/*
 * Callback from splice_to_pipe(), if we need to release some pages
 * at the end of the spd in case we error'ed out in filling the pipe.
 */
static void buffer_spd_release(struct splice_pipe_desc *spd, unsigned int i)
{
	struct buffer_ref *ref =
		(struct buffer_ref *)spd->partial[i].private;

	if (--ref->ref)
		return;

	ring_buffer_free_read_page(ref->buffer, ref->page);
	kfree(ref);
	spd->partial[i].private = 0;
}

static ssize_t
tracing_buffers_splice_read(struct file *file, loff_t *ppos,
			    struct pipe_inode_info *pipe, size_t len,
			    unsigned int flags)
{
	struct ftrace_buffer_info *info = file->private_data;
	struct partial_page partial[PIPE_BUFFERS];
	struct page *pages[PIPE_BUFFERS];
	struct splice_pipe_desc spd = {
		.pages		= pages,
		.partial	= partial,
		.flags		= flags,
		.ops		= &buffer_pipe_buf_ops,
		.spd_release	= buffer_spd_release,
	};
	struct buffer_ref *ref;
3663
	int entries, size, i;
3664 3665
	size_t ret;

3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677
	if (*ppos & (PAGE_SIZE - 1)) {
		WARN_ONCE(1, "Ftrace: previous read must page-align\n");
		return -EINVAL;
	}

	if (len & (PAGE_SIZE - 1)) {
		WARN_ONCE(1, "Ftrace: splice_read should page-align\n");
		if (len < PAGE_SIZE)
			return -EINVAL;
		len &= PAGE_MASK;
	}

3678 3679 3680
	entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu);

	for (i = 0; i < PIPE_BUFFERS && len && entries; i++, len -= PAGE_SIZE) {
3681 3682 3683 3684 3685 3686 3687
		struct page *page;
		int r;

		ref = kzalloc(sizeof(*ref), GFP_KERNEL);
		if (!ref)
			break;

3688
		ref->ref = 1;
3689 3690 3691 3692 3693 3694 3695 3696
		ref->buffer = info->tr->buffer;
		ref->page = ring_buffer_alloc_read_page(ref->buffer);
		if (!ref->page) {
			kfree(ref);
			break;
		}

		r = ring_buffer_read_page(ref->buffer, &ref->page,
3697
					  len, info->cpu, 1);
3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719
		if (r < 0) {
			ring_buffer_free_read_page(ref->buffer,
						   ref->page);
			kfree(ref);
			break;
		}

		/*
		 * zero out any left over data, this is going to
		 * user land.
		 */
		size = ring_buffer_page_len(ref->page);
		if (size < PAGE_SIZE)
			memset(ref->page + size, 0, PAGE_SIZE - size);

		page = virt_to_page(ref->page);

		spd.pages[i] = page;
		spd.partial[i].len = PAGE_SIZE;
		spd.partial[i].offset = 0;
		spd.partial[i].private = (unsigned long)ref;
		spd.nr_pages++;
3720
		*ppos += PAGE_SIZE;
3721 3722

		entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu);
3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749
	}

	spd.nr_pages = i;

	/* did we read anything? */
	if (!spd.nr_pages) {
		if (flags & SPLICE_F_NONBLOCK)
			ret = -EAGAIN;
		else
			ret = 0;
		/* TODO: block */
		return ret;
	}

	ret = splice_to_pipe(pipe, &spd);

	return ret;
}

static const struct file_operations tracing_buffers_fops = {
	.open		= tracing_buffers_open,
	.read		= tracing_buffers_read,
	.release	= tracing_buffers_release,
	.splice_read	= tracing_buffers_splice_read,
	.llseek		= no_llseek,
};

3750 3751 3752 3753 3754 3755 3756 3757 3758
static ssize_t
tracing_stats_read(struct file *filp, char __user *ubuf,
		   size_t count, loff_t *ppos)
{
	unsigned long cpu = (unsigned long)filp->private_data;
	struct trace_array *tr = &global_trace;
	struct trace_seq *s;
	unsigned long cnt;

3759
	s = kmalloc(sizeof(*s), GFP_KERNEL);
3760
	if (!s)
3761
		return -ENOMEM;
3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785

	trace_seq_init(s);

	cnt = ring_buffer_entries_cpu(tr->buffer, cpu);
	trace_seq_printf(s, "entries: %ld\n", cnt);

	cnt = ring_buffer_overrun_cpu(tr->buffer, cpu);
	trace_seq_printf(s, "overrun: %ld\n", cnt);

	cnt = ring_buffer_commit_overrun_cpu(tr->buffer, cpu);
	trace_seq_printf(s, "commit overrun: %ld\n", cnt);

	count = simple_read_from_buffer(ubuf, count, ppos, s->buffer, s->len);

	kfree(s);

	return count;
}

static const struct file_operations tracing_stats_fops = {
	.open		= tracing_open_generic,
	.read		= tracing_stats_read,
};

3786 3787
#ifdef CONFIG_DYNAMIC_FTRACE

S
Steven Rostedt 已提交
3788 3789 3790 3791 3792
int __weak ftrace_arch_read_dyn_info(char *buf, int size)
{
	return 0;
}

3793
static ssize_t
S
Steven Rostedt 已提交
3794
tracing_read_dyn_info(struct file *filp, char __user *ubuf,
3795 3796
		  size_t cnt, loff_t *ppos)
{
S
Steven Rostedt 已提交
3797 3798
	static char ftrace_dyn_info_buffer[1024];
	static DEFINE_MUTEX(dyn_info_mutex);
3799
	unsigned long *p = filp->private_data;
S
Steven Rostedt 已提交
3800
	char *buf = ftrace_dyn_info_buffer;
S
Steven Rostedt 已提交
3801
	int size = ARRAY_SIZE(ftrace_dyn_info_buffer);
3802 3803
	int r;

S
Steven Rostedt 已提交
3804 3805
	mutex_lock(&dyn_info_mutex);
	r = sprintf(buf, "%ld ", *p);
I
Ingo Molnar 已提交
3806

S
Steven Rostedt 已提交
3807
	r += ftrace_arch_read_dyn_info(buf+r, (size-1)-r);
S
Steven Rostedt 已提交
3808 3809 3810 3811 3812 3813 3814
	buf[r++] = '\n';

	r = simple_read_from_buffer(ubuf, cnt, ppos, buf, r);

	mutex_unlock(&dyn_info_mutex);

	return r;
3815 3816
}

3817
static const struct file_operations tracing_dyn_info_fops = {
I
Ingo Molnar 已提交
3818
	.open		= tracing_open_generic,
S
Steven Rostedt 已提交
3819
	.read		= tracing_read_dyn_info,
3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831
};
#endif

static struct dentry *d_tracer;

struct dentry *tracing_init_dentry(void)
{
	static int once;

	if (d_tracer)
		return d_tracer;

3832 3833 3834
	if (!debugfs_initialized())
		return NULL;

3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845
	d_tracer = debugfs_create_dir("tracing", NULL);

	if (!d_tracer && !once) {
		once = 1;
		pr_warning("Could not create debugfs directory 'tracing'\n");
		return NULL;
	}

	return d_tracer;
}

3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874
static struct dentry *d_percpu;

struct dentry *tracing_dentry_percpu(void)
{
	static int once;
	struct dentry *d_tracer;

	if (d_percpu)
		return d_percpu;

	d_tracer = tracing_init_dentry();

	if (!d_tracer)
		return NULL;

	d_percpu = debugfs_create_dir("per_cpu", d_tracer);

	if (!d_percpu && !once) {
		once = 1;
		pr_warning("Could not create debugfs directory 'per_cpu'\n");
		return NULL;
	}

	return d_percpu;
}

static void tracing_init_debugfs_percpu(long cpu)
{
	struct dentry *d_percpu = tracing_dentry_percpu();
3875
	struct dentry *d_cpu;
3876 3877
	/* strlen(cpu) + MAX(log10(cpu)) + '\0' */
	char cpu_dir[7];
3878 3879 3880 3881

	if (cpu > 999 || cpu < 0)
		return;

3882 3883 3884 3885 3886 3887
	sprintf(cpu_dir, "cpu%ld", cpu);
	d_cpu = debugfs_create_dir(cpu_dir, d_percpu);
	if (!d_cpu) {
		pr_warning("Could not create debugfs '%s' entry\n", cpu_dir);
		return;
	}
3888

3889
	/* per cpu trace_pipe */
3890 3891
	trace_create_file("trace_pipe", 0444, d_cpu,
			(void *) cpu, &tracing_pipe_fops);
3892 3893

	/* per cpu trace */
3894 3895
	trace_create_file("trace", 0644, d_cpu,
			(void *) cpu, &tracing_fops);
3896

3897 3898
	trace_create_file("trace_pipe_raw", 0444, d_cpu,
			(void *) cpu, &tracing_buffers_fops);
3899

3900 3901
	trace_create_file("stats", 0444, d_cpu,
			(void *) cpu, &tracing_stats_fops);
3902 3903
}

S
Steven Rostedt 已提交
3904 3905 3906 3907 3908
#ifdef CONFIG_FTRACE_SELFTEST
/* Let selftest have access to static functions in this file */
#include "trace_selftest.c"
#endif

3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997
struct trace_option_dentry {
	struct tracer_opt		*opt;
	struct tracer_flags		*flags;
	struct dentry			*entry;
};

static ssize_t
trace_options_read(struct file *filp, char __user *ubuf, size_t cnt,
			loff_t *ppos)
{
	struct trace_option_dentry *topt = filp->private_data;
	char *buf;

	if (topt->flags->val & topt->opt->bit)
		buf = "1\n";
	else
		buf = "0\n";

	return simple_read_from_buffer(ubuf, cnt, ppos, buf, 2);
}

static ssize_t
trace_options_write(struct file *filp, const char __user *ubuf, size_t cnt,
			 loff_t *ppos)
{
	struct trace_option_dentry *topt = filp->private_data;
	unsigned long val;
	char buf[64];
	int ret;

	if (cnt >= sizeof(buf))
		return -EINVAL;

	if (copy_from_user(&buf, ubuf, cnt))
		return -EFAULT;

	buf[cnt] = 0;

	ret = strict_strtoul(buf, 10, &val);
	if (ret < 0)
		return ret;

	ret = 0;
	switch (val) {
	case 0:
		/* do nothing if already cleared */
		if (!(topt->flags->val & topt->opt->bit))
			break;

		mutex_lock(&trace_types_lock);
		if (current_trace->set_flag)
			ret = current_trace->set_flag(topt->flags->val,
						      topt->opt->bit, 0);
		mutex_unlock(&trace_types_lock);
		if (ret)
			return ret;
		topt->flags->val &= ~topt->opt->bit;
		break;
	case 1:
		/* do nothing if already set */
		if (topt->flags->val & topt->opt->bit)
			break;

		mutex_lock(&trace_types_lock);
		if (current_trace->set_flag)
			ret = current_trace->set_flag(topt->flags->val,
						      topt->opt->bit, 1);
		mutex_unlock(&trace_types_lock);
		if (ret)
			return ret;
		topt->flags->val |= topt->opt->bit;
		break;

	default:
		return -EINVAL;
	}

	*ppos += cnt;

	return cnt;
}


static const struct file_operations trace_options_fops = {
	.open = tracing_open_generic,
	.read = trace_options_read,
	.write = trace_options_write,
};

3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033
static ssize_t
trace_options_core_read(struct file *filp, char __user *ubuf, size_t cnt,
			loff_t *ppos)
{
	long index = (long)filp->private_data;
	char *buf;

	if (trace_flags & (1 << index))
		buf = "1\n";
	else
		buf = "0\n";

	return simple_read_from_buffer(ubuf, cnt, ppos, buf, 2);
}

static ssize_t
trace_options_core_write(struct file *filp, const char __user *ubuf, size_t cnt,
			 loff_t *ppos)
{
	long index = (long)filp->private_data;
	char buf[64];
	unsigned long val;
	int ret;

	if (cnt >= sizeof(buf))
		return -EINVAL;

	if (copy_from_user(&buf, ubuf, cnt))
		return -EFAULT;

	buf[cnt] = 0;

	ret = strict_strtoul(buf, 10, &val);
	if (ret < 0)
		return ret;

4034
	if (val != 0 && val != 1)
4035
		return -EINVAL;
4036
	set_tracer_flags(1 << index, val);
4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048

	*ppos += cnt;

	return cnt;
}

static const struct file_operations trace_options_core_fops = {
	.open = tracing_open_generic,
	.read = trace_options_core_read,
	.write = trace_options_core_write,
};

4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064
struct dentry *trace_create_file(const char *name,
				 mode_t mode,
				 struct dentry *parent,
				 void *data,
				 const struct file_operations *fops)
{
	struct dentry *ret;

	ret = debugfs_create_file(name, mode, parent, data, fops);
	if (!ret)
		pr_warning("Could not create debugfs '%s' entry\n", name);

	return ret;
}


4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085
static struct dentry *trace_options_init_dentry(void)
{
	struct dentry *d_tracer;
	static struct dentry *t_options;

	if (t_options)
		return t_options;

	d_tracer = tracing_init_dentry();
	if (!d_tracer)
		return NULL;

	t_options = debugfs_create_dir("options", d_tracer);
	if (!t_options) {
		pr_warning("Could not create debugfs directory 'options'\n");
		return NULL;
	}

	return t_options;
}

4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099
static void
create_trace_option_file(struct trace_option_dentry *topt,
			 struct tracer_flags *flags,
			 struct tracer_opt *opt)
{
	struct dentry *t_options;

	t_options = trace_options_init_dentry();
	if (!t_options)
		return;

	topt->flags = flags;
	topt->opt = opt;

4100
	topt->entry = trace_create_file(opt->name, 0644, t_options, topt,
4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125
				    &trace_options_fops);

}

static struct trace_option_dentry *
create_trace_option_files(struct tracer *tracer)
{
	struct trace_option_dentry *topts;
	struct tracer_flags *flags;
	struct tracer_opt *opts;
	int cnt;

	if (!tracer)
		return NULL;

	flags = tracer->flags;

	if (!flags || !flags->opts)
		return NULL;

	opts = flags->opts;

	for (cnt = 0; opts[cnt].name; cnt++)
		;

4126
	topts = kcalloc(cnt + 1, sizeof(*topts), GFP_KERNEL);
4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152
	if (!topts)
		return NULL;

	for (cnt = 0; opts[cnt].name; cnt++)
		create_trace_option_file(&topts[cnt], flags,
					 &opts[cnt]);

	return topts;
}

static void
destroy_trace_option_files(struct trace_option_dentry *topts)
{
	int cnt;

	if (!topts)
		return;

	for (cnt = 0; topts[cnt].opt; cnt++) {
		if (topts[cnt].entry)
			debugfs_remove(topts[cnt].entry);
	}

	kfree(topts);
}

4153 4154 4155 4156 4157 4158 4159 4160 4161
static struct dentry *
create_trace_option_core_file(const char *option, long index)
{
	struct dentry *t_options;

	t_options = trace_options_init_dentry();
	if (!t_options)
		return NULL;

4162
	return trace_create_file(option, 0644, t_options, (void *)index,
4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173 4174
				    &trace_options_core_fops);
}

static __init void create_trace_options_dir(void)
{
	struct dentry *t_options;
	int i;

	t_options = trace_options_init_dentry();
	if (!t_options)
		return;

4175 4176
	for (i = 0; trace_options[i]; i++)
		create_trace_option_core_file(trace_options[i], i);
4177 4178
}

4179
static __init int tracer_init_debugfs(void)
4180 4181
{
	struct dentry *d_tracer;
4182
	int cpu;
4183 4184 4185

	d_tracer = tracing_init_dentry();

4186 4187
	trace_create_file("tracing_enabled", 0644, d_tracer,
			&global_trace, &tracing_ctrl_fops);
4188

4189 4190
	trace_create_file("trace_options", 0644, d_tracer,
			NULL, &tracing_iter_fops);
4191

4192 4193 4194 4195 4196
	trace_create_file("tracing_cpumask", 0644, d_tracer,
			NULL, &tracing_cpumask_fops);

	trace_create_file("trace", 0644, d_tracer,
			(void *) TRACE_PIPE_ALL_CPU, &tracing_fops);
4197

4198 4199 4200
	trace_create_file("available_tracers", 0444, d_tracer,
			&global_trace, &show_traces_fops);

4201
	trace_create_file("current_tracer", 0644, d_tracer,
4202 4203
			&global_trace, &set_tracer_fops);

4204
#ifdef CONFIG_TRACER_MAX_TRACE
4205 4206 4207 4208 4209
	trace_create_file("tracing_max_latency", 0644, d_tracer,
			&tracing_max_latency, &tracing_max_lat_fops);

	trace_create_file("tracing_thresh", 0644, d_tracer,
			&tracing_thresh, &tracing_max_lat_fops);
4210
#endif
4211

4212
	trace_create_file("README", 0444, d_tracer,
4213 4214 4215
			NULL, &tracing_readme_fops);

	trace_create_file("trace_pipe", 0444, d_tracer,
4216
			(void *) TRACE_PIPE_ALL_CPU, &tracing_pipe_fops);
4217 4218 4219 4220 4221 4222

	trace_create_file("buffer_size_kb", 0644, d_tracer,
			&global_trace, &tracing_entries_fops);

	trace_create_file("trace_marker", 0220, d_tracer,
			NULL, &tracing_mark_fops);
4223

4224 4225
	trace_create_file("saved_cmdlines", 0444, d_tracer,
			NULL, &tracing_saved_cmdlines_fops);
4226

4227 4228 4229
	trace_create_file("trace_clock", 0644, d_tracer, NULL,
			  &trace_clock_fops);

4230
#ifdef CONFIG_DYNAMIC_FTRACE
4231 4232
	trace_create_file("dyn_ftrace_total_info", 0444, d_tracer,
			&ftrace_update_tot_cnt, &tracing_dyn_info_fops);
4233
#endif
I
Ingo Molnar 已提交
4234 4235 4236
#ifdef CONFIG_SYSPROF_TRACER
	init_tracer_sysprof_debugfs(d_tracer);
#endif
4237

4238 4239
	create_trace_options_dir();

4240 4241 4242
	for_each_tracing_cpu(cpu)
		tracing_init_debugfs_percpu(cpu);

4243
	return 0;
4244 4245
}

4246 4247 4248
static int trace_panic_handler(struct notifier_block *this,
			       unsigned long event, void *unused)
{
4249 4250
	if (ftrace_dump_on_oops)
		ftrace_dump();
4251 4252 4253 4254 4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265
	return NOTIFY_OK;
}

static struct notifier_block trace_panic_notifier = {
	.notifier_call  = trace_panic_handler,
	.next           = NULL,
	.priority       = 150   /* priority: INT_MAX >= x >= 0 */
};

static int trace_die_handler(struct notifier_block *self,
			     unsigned long val,
			     void *data)
{
	switch (val) {
	case DIE_OOPS:
4266 4267
		if (ftrace_dump_on_oops)
			ftrace_dump();
4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290
		break;
	default:
		break;
	}
	return NOTIFY_OK;
}

static struct notifier_block trace_die_notifier = {
	.notifier_call = trace_die_handler,
	.priority = 200
};

/*
 * printk is set to max of 1024, we really don't need it that big.
 * Nothing should be printing 1000 characters anyway.
 */
#define TRACE_MAX_PRINT		1000

/*
 * Define here KERN_TRACE so that we have one place to modify
 * it if we decide to change what log level the ftrace dump
 * should be at.
 */
4291
#define KERN_TRACE		KERN_EMERG
4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304

static void
trace_printk_seq(struct trace_seq *s)
{
	/* Probably should print a warning here. */
	if (s->len >= 1000)
		s->len = 1000;

	/* should be zero ended, but we are paranoid. */
	s->buffer[s->len] = 0;

	printk(KERN_TRACE "%s", s->buffer);

4305
	trace_seq_init(s);
4306 4307
}

4308
static void __ftrace_dump(bool disable_tracing)
4309
{
4310
	static arch_spinlock_t ftrace_dump_lock =
4311
		(arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED;
4312 4313
	/* use static because iter can be a bit big for the stack */
	static struct trace_iterator iter;
4314
	unsigned int old_userobj;
4315
	static int dump_ran;
4316 4317
	unsigned long flags;
	int cnt = 0, cpu;
4318 4319

	/* only one dump */
4320
	local_irq_save(flags);
4321
	arch_spin_lock(&ftrace_dump_lock);
4322 4323 4324 4325 4326
	if (dump_ran)
		goto out;

	dump_ran = 1;

4327
	tracing_off();
4328 4329 4330

	if (disable_tracing)
		ftrace_kill();
4331

4332 4333 4334 4335
	for_each_tracing_cpu(cpu) {
		atomic_inc(&global_trace.data[cpu]->disabled);
	}

4336 4337
	old_userobj = trace_flags & TRACE_ITER_SYM_USEROBJ;

4338 4339 4340
	/* don't look at user memory in panic mode */
	trace_flags &= ~TRACE_ITER_SYM_USEROBJ;

4341 4342
	printk(KERN_TRACE "Dumping ftrace buffer:\n");

4343
	/* Simulate the iterator */
4344 4345
	iter.tr = &global_trace;
	iter.trace = current_trace;
4346
	iter.cpu_file = TRACE_PIPE_ALL_CPU;
4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369

	/*
	 * We need to stop all tracing on all CPUS to read the
	 * the next buffer. This is a bit expensive, but is
	 * not done often. We fill all what we can read,
	 * and then release the locks again.
	 */

	while (!trace_empty(&iter)) {

		if (!cnt)
			printk(KERN_TRACE "---------------------------------\n");

		cnt++;

		/* reset all but tr, trace, and overruns */
		memset(&iter.seq, 0,
		       sizeof(struct trace_iterator) -
		       offsetof(struct trace_iterator, seq));
		iter.iter_flags |= TRACE_FILE_LAT_FMT;
		iter.pos = -1;

		if (find_next_entry_inc(&iter) != NULL) {
4370 4371 4372 4373 4374
			int ret;

			ret = print_trace_line(&iter);
			if (ret != TRACE_TYPE_NO_CONSUME)
				trace_consume(&iter);
4375 4376 4377 4378 4379 4380 4381 4382 4383 4384
		}

		trace_printk_seq(&iter.seq);
	}

	if (!cnt)
		printk(KERN_TRACE "   (ftrace buffer empty)\n");
	else
		printk(KERN_TRACE "---------------------------------\n");

4385 4386 4387 4388 4389 4390 4391 4392 4393 4394
	/* Re-enable tracing if requested */
	if (!disable_tracing) {
		trace_flags |= old_userobj;

		for_each_tracing_cpu(cpu) {
			atomic_dec(&global_trace.data[cpu]->disabled);
		}
		tracing_on();
	}

4395
 out:
4396
	arch_spin_unlock(&ftrace_dump_lock);
4397
	local_irq_restore(flags);
4398 4399
}

4400 4401 4402 4403 4404 4405
/* By default: disable tracing after the dump */
void ftrace_dump(void)
{
	__ftrace_dump(true);
}

4406
__init static int tracer_alloc_buffers(void)
4407
{
4408
	int ring_buf_size;
4409
	int i;
4410
	int ret = -ENOMEM;
4411

4412 4413 4414 4415 4416
	if (!alloc_cpumask_var(&tracing_buffer_mask, GFP_KERNEL))
		goto out;

	if (!alloc_cpumask_var(&tracing_cpumask, GFP_KERNEL))
		goto out_free_buffer_mask;
4417

4418
	if (!zalloc_cpumask_var(&tracing_reader_cpumask, GFP_KERNEL))
4419 4420
		goto out_free_tracing_cpumask;

4421 4422 4423 4424 4425 4426
	/* To save memory, keep the ring buffer size to its minimum */
	if (ring_buffer_expanded)
		ring_buf_size = trace_buf_size;
	else
		ring_buf_size = 1;

4427 4428 4429 4430
	cpumask_copy(tracing_buffer_mask, cpu_possible_mask);
	cpumask_copy(tracing_cpumask, cpu_all_mask);

	/* TODO: make the number of buffers hot pluggable with CPUS */
4431
	global_trace.buffer = ring_buffer_alloc(ring_buf_size,
4432 4433 4434 4435
						   TRACE_BUFFER_FLAGS);
	if (!global_trace.buffer) {
		printk(KERN_ERR "tracer: failed to allocate ring buffer!\n");
		WARN_ON(1);
4436
		goto out_free_cpumask;
4437
	}
4438
	global_trace.entries = ring_buffer_size(global_trace.buffer);
4439

4440

4441
#ifdef CONFIG_TRACER_MAX_TRACE
4442
	max_tr.buffer = ring_buffer_alloc(ring_buf_size,
4443 4444 4445 4446 4447
					     TRACE_BUFFER_FLAGS);
	if (!max_tr.buffer) {
		printk(KERN_ERR "tracer: failed to allocate max ring buffer!\n");
		WARN_ON(1);
		ring_buffer_free(global_trace.buffer);
4448
		goto out_free_cpumask;
4449
	}
4450 4451
	max_tr.entries = ring_buffer_size(max_tr.buffer);
	WARN_ON(max_tr.entries != global_trace.entries);
4452
#endif
4453

4454
	/* Allocate the first page for all buffers */
4455
	for_each_tracing_cpu(i) {
4456
		global_trace.data[i] = &per_cpu(global_trace_cpu, i);
4457
		max_tr.data[i] = &per_cpu(max_tr_data, i);
4458
	}
4459 4460 4461

	trace_init_cmdlines();

4462
	register_tracer(&nop_trace);
4463
	current_trace = &nop_trace;
4464 4465 4466
#ifdef CONFIG_BOOT_TRACER
	register_tracer(&boot_tracer);
#endif
S
Steven Rostedt 已提交
4467 4468
	/* All seems OK, enable tracing */
	tracing_disabled = 0;
4469

4470 4471 4472 4473
	atomic_notifier_chain_register(&panic_notifier_list,
				       &trace_panic_notifier);

	register_die_notifier(&trace_die_notifier);
4474 4475

	return 0;
4476

4477
out_free_cpumask:
4478 4479
	free_cpumask_var(tracing_reader_cpumask);
out_free_tracing_cpumask:
4480 4481 4482 4483 4484
	free_cpumask_var(tracing_cpumask);
out_free_buffer_mask:
	free_cpumask_var(tracing_buffer_mask);
out:
	return ret;
4485
}
4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505

__init static int clear_boot_tracer(void)
{
	/*
	 * The default tracer at boot buffer is an init section.
	 * This function is called in lateinit. If we did not
	 * find the boot tracer, then clear it out, to prevent
	 * later registration from accessing the buffer that is
	 * about to be freed.
	 */
	if (!default_bootup_tracer)
		return 0;

	printk(KERN_INFO "ftrace bootup tracer '%s' not registered.\n",
	       default_bootup_tracer);
	default_bootup_tracer = NULL;

	return 0;
}

4506 4507
early_initcall(tracer_alloc_buffers);
fs_initcall(tracer_init_debugfs);
4508
late_initcall(clear_boot_tracer);