trace.c 119.3 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11
/*
 * 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
12
 *  Copyright (C) 2004 Nadia Yvette Chambers
13
 */
14
#include <linux/ring_buffer.h>
15
#include <generated/utsrelease.h>
16 17
#include <linux/stacktrace.h>
#include <linux/writeback.h>
18 19
#include <linux/kallsyms.h>
#include <linux/seq_file.h>
20
#include <linux/notifier.h>
21
#include <linux/irqflags.h>
22
#include <linux/irq_work.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
#include <linux/rwsem.h>
36
#include <linux/slab.h>
37 38
#include <linux/ctype.h>
#include <linux/init.h>
39
#include <linux/poll.h>
40
#include <linux/nmi.h>
41
#include <linux/fs.h>
I
Ingo Molnar 已提交
42

43
#include "trace.h"
44
#include "trace_output.h"
45

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 87
/*
 * To prevent the comm cache from being overwritten when no
 * tracing is active, only save the comm when a trace event
 * occurred.
 */
static DEFINE_PER_CPU(bool, trace_cmdline_save);

88 89 90 91 92 93 94 95
/*
 * When a reader is waiting for data, then this variable is
 * set to true.
 */
static bool trace_wakeup_needed;

static struct irq_work trace_work_wakeup;

96 97 98 99 100 101
/*
 * 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.
 */
102
static int tracing_disabled = 1;
103

104
DEFINE_PER_CPU(int, ftrace_cpu_disabled);
105

106
cpumask_var_t __read_mostly	tracing_buffer_mask;
107

108 109 110 111 112 113 114 115 116 117 118
/*
 * 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
119 120 121
 * /proc/sys/kernel/ftrace_dump_on_oops
 * Set 1 if you want to dump buffers of all CPUs
 * Set 2 if you want to dump the buffer of the CPU that triggered oops
122
 */
123 124

enum ftrace_dump_mode ftrace_dump_on_oops;
125

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
static int __init set_ftrace_dump_on_oops(char *str)
{
144 145 146 147 148 149 150 151 152 153 154
	if (*str++ != '=' || !*str) {
		ftrace_dump_on_oops = DUMP_ALL;
		return 1;
	}

	if (!strcmp("orig_cpu", str)) {
		ftrace_dump_on_oops = DUMP_ORIG;
                return 1;
        }

        return 0;
155 156
}
__setup("ftrace_dump_on_oops", set_ftrace_dump_on_oops);
S
Steven Rostedt 已提交
157

158 159 160 161 162 163 164 165 166 167 168 169

static char trace_boot_options_buf[MAX_TRACER_SIZE] __initdata;
static char *trace_boot_options __initdata;

static int __init set_trace_boot_options(char *str)
{
	strncpy(trace_boot_options_buf, str, MAX_TRACER_SIZE);
	trace_boot_options = trace_boot_options_buf;
	return 0;
}
__setup("trace_options=", set_trace_boot_options);

170
unsigned long long ns2usecs(cycle_t nsec)
171 172 173 174 175 176
{
	nsec += 500;
	do_div(nsec, 1000);
	return nsec;
}

S
Steven Rostedt 已提交
177 178 179 180 181 182 183 184 185 186 187 188
/*
 * 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.
 */
189 190 191 192
static struct trace_array	global_trace;

static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu);

193 194
int filter_current_check_discard(struct ring_buffer *buffer,
				 struct ftrace_event_call *call, void *rec,
195 196
				 struct ring_buffer_event *event)
{
197
	return filter_check_discard(call, rec, buffer, event);
198
}
199
EXPORT_SYMBOL_GPL(filter_current_check_discard);
200

201 202 203 204 205 206 207 208 209 210 211 212 213
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;
}
214

S
Steven Rostedt 已提交
215 216 217 218 219 220 221 222 223 224
/*
 * 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.
 */
225 226
static struct trace_array	max_tr;

227
static DEFINE_PER_CPU(struct trace_array_cpu, max_tr_data);
228

229 230
int tracing_is_enabled(void)
{
231
	return tracing_is_on();
232 233
}

S
Steven Rostedt 已提交
234
/*
235 236 237
 * 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.
238 239 240 241 242
 *
 * 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
243
 */
244
#define TRACE_BUF_SIZE_DEFAULT	1441792UL /* 16384 * 88 (sizeof(entry)) */
245

246
static unsigned long		trace_buf_size = TRACE_BUF_SIZE_DEFAULT;
247

S
Steven Rostedt 已提交
248
/* trace_types holds a link list of available tracers. */
249
static struct tracer		*trace_types __read_mostly;
S
Steven Rostedt 已提交
250 251

/* current_trace points to the tracer that is currently active */
252
static struct tracer		*current_trace __read_mostly;
S
Steven Rostedt 已提交
253 254 255 256

/*
 * trace_types_lock is used to protect the trace_types list.
 */
257
static DEFINE_MUTEX(trace_types_lock);
S
Steven Rostedt 已提交
258

259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
/*
 * serialize the access of the ring buffer
 *
 * ring buffer serializes readers, but it is low level protection.
 * The validity of the events (which returns by ring_buffer_peek() ..etc)
 * are not protected by ring buffer.
 *
 * The content of events may become garbage if we allow other process consumes
 * these events concurrently:
 *   A) the page of the consumed events may become a normal page
 *      (not reader page) in ring buffer, and this page will be rewrited
 *      by events producer.
 *   B) The page of the consumed events may become a page for splice_read,
 *      and this page will be returned to system.
 *
 * These primitives allow multi process access to different cpu ring buffer
 * concurrently.
 *
 * These primitives don't distinguish read-only and read-consume access.
 * Multi read-only access are also serialized.
 */

#ifdef CONFIG_SMP
static DECLARE_RWSEM(all_cpu_access_lock);
static DEFINE_PER_CPU(struct mutex, cpu_access_lock);

static inline void trace_access_lock(int cpu)
{
	if (cpu == TRACE_PIPE_ALL_CPU) {
		/* gain it for accessing the whole ring buffer. */
		down_write(&all_cpu_access_lock);
	} else {
		/* gain it for accessing a cpu ring buffer. */

		/* Firstly block other trace_access_lock(TRACE_PIPE_ALL_CPU). */
		down_read(&all_cpu_access_lock);

		/* Secondly block other access to this @cpu ring buffer. */
		mutex_lock(&per_cpu(cpu_access_lock, cpu));
	}
}

static inline void trace_access_unlock(int cpu)
{
	if (cpu == TRACE_PIPE_ALL_CPU) {
		up_write(&all_cpu_access_lock);
	} else {
		mutex_unlock(&per_cpu(cpu_access_lock, cpu));
		up_read(&all_cpu_access_lock);
	}
}

static inline void trace_access_lock_init(void)
{
	int cpu;

	for_each_possible_cpu(cpu)
		mutex_init(&per_cpu(cpu_access_lock, cpu));
}

#else

static DEFINE_MUTEX(access_lock);

static inline void trace_access_lock(int cpu)
{
	(void)cpu;
	mutex_lock(&access_lock);
}

static inline void trace_access_unlock(int cpu)
{
	(void)cpu;
	mutex_unlock(&access_lock);
}

static inline void trace_access_lock_init(void)
{
}

#endif

S
Steven Rostedt 已提交
341
/* trace_wait is a waitqueue for tasks blocked on trace_poll */
342 343
static DECLARE_WAIT_QUEUE_HEAD(trace_wait);

344
/* trace_flags holds trace_options default values */
345
unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK |
346
	TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO | TRACE_ITER_SLEEP_TIME |
347
	TRACE_ITER_GRAPH_TIME | TRACE_ITER_RECORD_CMD | TRACE_ITER_OVERWRITE |
348
	TRACE_ITER_IRQ_INFO | TRACE_ITER_MARKERS;
349

350
static int trace_stop_count;
351
static DEFINE_RAW_SPINLOCK(tracing_start_lock);
352

353 354 355 356 357 358 359 360
/**
 * trace_wake_up - wake up tasks waiting for trace input
 *
 * Schedules a delayed work to wake up any task that is blocked on the
 * trace_wait queue. These is used with trace_poll for tasks polling the
 * trace.
 */
static void trace_wake_up(struct irq_work *work)
361
{
362
	wake_up_all(&trace_wait);
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
/**
 * tracing_on - enable tracing buffers
 *
 * This function enables tracing buffers that may have been
 * disabled with tracing_off.
 */
void tracing_on(void)
{
	if (global_trace.buffer)
		ring_buffer_record_on(global_trace.buffer);
	/*
	 * This flag is only looked at when buffers haven't been
	 * allocated yet. We don't really care about the race
	 * between setting this flag and actually turning
	 * on the buffer.
	 */
	global_trace.buffer_disabled = 0;
}
EXPORT_SYMBOL_GPL(tracing_on);

/**
 * tracing_off - turn off tracing buffers
 *
 * This function stops the tracing buffers from recording data.
 * It does not disable any overhead the tracers themselves may
 * be causing. This function simply causes all recording to
 * the ring buffers to fail.
 */
void tracing_off(void)
{
	if (global_trace.buffer)
397
		ring_buffer_record_off(global_trace.buffer);
398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418
	/*
	 * This flag is only looked at when buffers haven't been
	 * allocated yet. We don't really care about the race
	 * between setting this flag and actually turning
	 * on the buffer.
	 */
	global_trace.buffer_disabled = 1;
}
EXPORT_SYMBOL_GPL(tracing_off);

/**
 * tracing_is_on - show state of ring buffers enabled
 */
int tracing_is_on(void)
{
	if (global_trace.buffer)
		return ring_buffer_record_is_on(global_trace.buffer);
	return !global_trace.buffer_disabled;
}
EXPORT_SYMBOL_GPL(tracing_is_on);

419
static int __init set_buf_size(char *str)
420
{
421
	unsigned long buf_size;
422

423 424
	if (!str)
		return 0;
425
	buf_size = memparse(str, &str);
426
	/* nr_entries can not be zero */
427
	if (buf_size == 0)
428
		return 0;
429
	trace_buf_size = buf_size;
430 431
	return 1;
}
432
__setup("trace_buf_size=", set_buf_size);
433

434 435
static int __init set_tracing_thresh(char *str)
{
436
	unsigned long threshold;
437 438 439 440
	int ret;

	if (!str)
		return 0;
441
	ret = kstrtoul(str, 0, &threshold);
442 443
	if (ret < 0)
		return 0;
444
	tracing_thresh = threshold * 1000;
445 446 447 448
	return 1;
}
__setup("tracing_thresh=", set_tracing_thresh);

S
Steven Rostedt 已提交
449 450 451 452 453
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

S
Steven Rostedt 已提交
454
/* These must match the bit postions in trace_iterator_flags */
455 456 457 458 459
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
I
Ingo Molnar 已提交
460
	"raw",
461
	"hex",
I
Ingo Molnar 已提交
462
	"bin",
463
	"block",
I
Ingo Molnar 已提交
464
	"stacktrace",
465
	"trace_printk",
466
	"ftrace_preempt",
467
	"branch",
468
	"annotate",
469
	"userstacktrace",
470
	"sym-userobj",
471
	"printk-msg-only",
472
	"context-info",
473
	"latency-format",
474
	"sleep-time",
475
	"graph-time",
476
	"record-cmd",
477
	"overwrite",
478
	"disable_on_free",
479
	"irq-info",
480
	"markers",
481 482 483
	NULL
};

484 485 486
static struct {
	u64 (*func)(void);
	const char *name;
487
	int in_ns;		/* is this clock in nanoseconds? */
488
} trace_clocks[] = {
489 490 491
	{ trace_clock_local,	"local",	1 },
	{ trace_clock_global,	"global",	1 },
	{ trace_clock_counter,	"counter",	0 },
D
David Sharp 已提交
492
	ARCH_TRACE_CLOCKS
493 494 495 496
};

int trace_clock_id;

497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573
/*
 * 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)) {
574
		if (parser->idx < parser->size - 1)
575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602
			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;
}

603 604 605 606 607
ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt)
{
	int len;
	int ret;

608 609 610
	if (!cnt)
		return 0;

611 612 613 614 615 616 617
	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);
618
	if (ret == cnt)
619 620
		return -EFAULT;

621 622
	cnt -= ret;

623
	s->readpos += cnt;
624
	return cnt;
S
Steven Rostedt 已提交
625 626
}

627
static ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt)
628 629 630 631 632 633 634 635 636
{
	int len;

	if (s->len <= s->readpos)
		return -EBUSY;

	len = s->len - s->readpos;
	if (cnt > len)
		cnt = len;
637
	memcpy(buf, s->buffer + s->readpos, cnt);
638

639
	s->readpos += cnt;
640 641 642
	return cnt;
}

643 644 645 646 647 648
/*
 * 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.
 *
649
 * This is defined as a arch_spinlock_t in order to help
650 651 652 653 654 655
 * 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.
 */
656
static arch_spinlock_t ftrace_max_lock =
657
	(arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED;
658

659 660
unsigned long __read_mostly	tracing_thresh;

661 662 663 664 665 666 667 668 669 670 671 672
#ifdef CONFIG_TRACER_MAX_TRACE
unsigned long __read_mostly	tracing_max_latency;

/*
 * 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];
673
	struct trace_array_cpu *max_data;
674 675 676 677

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

678 679 680 681
	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;
682

683
	memcpy(max_data->comm, tsk->comm, TASK_COMM_LEN);
684 685 686 687 688
	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;
689 690 691 692 693

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

S
Steven Rostedt 已提交
694 695 696 697 698 699 700 701 702
/**
 * 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
703
void
704 705
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
706
	struct ring_buffer *buf = tr->buffer;
707

708 709 710
	if (trace_stop_count)
		return;

711
	WARN_ON_ONCE(!irqs_disabled());
712 713 714 715
	if (!current_trace->use_max_tr) {
		WARN_ON_ONCE(1);
		return;
	}
716
	arch_spin_lock(&ftrace_max_lock);
717 718 719 720

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

721
	__update_max_tr(tr, tsk, cpu);
722
	arch_spin_unlock(&ftrace_max_lock);
723 724 725 726 727 728 729
}

/**
 * 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
730 731
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
732
 */
I
Ingo Molnar 已提交
733
void
734 735
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
736
	int ret;
737

738 739 740
	if (trace_stop_count)
		return;

741
	WARN_ON_ONCE(!irqs_disabled());
742 743 744 745 746
	if (!current_trace->use_max_tr) {
		WARN_ON_ONCE(1);
		return;
	}

747
	arch_spin_lock(&ftrace_max_lock);
748

749 750
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

751 752 753 754 755 756 757 758 759 760 761 762
	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");
	}

	WARN_ON_ONCE(ret && ret != -EAGAIN && ret != -EBUSY);
763 764

	__update_max_tr(tr, tsk, cpu);
765
	arch_spin_unlock(&ftrace_max_lock);
766
}
767
#endif /* CONFIG_TRACER_MAX_TRACE */
768

769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802
static void default_wait_pipe(struct trace_iterator *iter)
{
	DEFINE_WAIT(wait);

	prepare_to_wait(&trace_wait, &wait, TASK_INTERRUPTIBLE);

	/*
	 * The events can happen in critical sections where
	 * checking a work queue can cause deadlocks.
	 * After adding a task to the queue, this flag is set
	 * only to notify events to try to wake up the queue
	 * using irq_work.
	 *
	 * We don't clear it even if the buffer is no longer
	 * empty. The flag only causes the next event to run
	 * irq_work to do the work queue wake up. The worse
	 * that can happen if we race with !trace_empty() is that
	 * an event will cause an irq_work to try to wake up
	 * an empty queue.
	 *
	 * There's no reason to protect this flag either, as
	 * the work queue and irq_work logic will do the necessary
	 * synchronization for the wake ups. The only thing
	 * that is necessary is that the wake up happens after
	 * a task has been queued. It's OK for spurious wake ups.
	 */
	trace_wakeup_needed = true;

	if (trace_empty(iter))
		schedule();

	finish_wait(&trace_wait, &wait);
}

S
Steven Rostedt 已提交
803 804 805 806 807 808
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
809 810 811 812 813 814 815 816 817 818
int register_tracer(struct tracer *type)
{
	struct tracer *t;
	int ret = 0;

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

819
	if (strlen(type->name) >= MAX_TRACER_SIZE) {
L
Li Zefan 已提交
820 821 822 823
		pr_info("Tracer has a name longer than %d\n", MAX_TRACER_SIZE);
		return -1;
	}

824
	mutex_lock(&trace_types_lock);
I
Ingo Molnar 已提交
825

826 827
	tracing_selftest_running = true;

828 829 830
	for (t = trace_types; t; t = t->next) {
		if (strcmp(type->name, t->name) == 0) {
			/* already found */
L
Li Zefan 已提交
831
			pr_info("Tracer %s already registered\n",
832 833 834 835 836 837
				type->name);
			ret = -1;
			goto out;
		}
	}

838 839 840 841 842 843 844
	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;
845 846 847
	if (!type->wait_pipe)
		type->wait_pipe = default_wait_pipe;

848

S
Steven Rostedt 已提交
849
#ifdef CONFIG_FTRACE_STARTUP_TEST
850
	if (type->selftest && !tracing_selftest_disabled) {
S
Steven Rostedt 已提交
851 852
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
853

S
Steven Rostedt 已提交
854 855 856 857 858 859 860
		/*
		 * 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.
		 */
861
		tracing_reset_online_cpus(tr);
I
Ingo Molnar 已提交
862

S
Steven Rostedt 已提交
863
		current_trace = type;
864 865 866

		/* If we expanded the buffers, make sure the max is expanded too */
		if (ring_buffer_expanded && type->use_max_tr)
867 868
			ring_buffer_resize(max_tr.buffer, trace_buf_size,
						RING_BUFFER_ALL_CPUS);
869

S
Steven Rostedt 已提交
870 871 872 873 874 875 876
		/* 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");
877 878
			/* Add the warning after printing 'FAILED' */
			WARN_ON(1);
S
Steven Rostedt 已提交
879 880
			goto out;
		}
S
Steven Rostedt 已提交
881
		/* Only reset on passing, to avoid touching corrupted buffers */
882
		tracing_reset_online_cpus(tr);
I
Ingo Molnar 已提交
883

884 885
		/* Shrink the max buffer again */
		if (ring_buffer_expanded && type->use_max_tr)
886 887
			ring_buffer_resize(max_tr.buffer, 1,
						RING_BUFFER_ALL_CPUS);
888

S
Steven Rostedt 已提交
889 890 891 892
		printk(KERN_CONT "PASSED\n");
	}
#endif

893 894
	type->next = trace_types;
	trace_types = type;
S
Steven Rostedt 已提交
895

896
 out:
897
	tracing_selftest_running = false;
898 899
	mutex_unlock(&trace_types_lock);

S
Steven Rostedt 已提交
900 901 902
	if (ret || !default_bootup_tracer)
		goto out_unlock;

L
Li Zefan 已提交
903
	if (strncmp(default_bootup_tracer, type->name, MAX_TRACER_SIZE))
S
Steven Rostedt 已提交
904 905 906 907 908 909 910 911
		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;
912
#ifdef CONFIG_FTRACE_STARTUP_TEST
S
Steven Rostedt 已提交
913 914
	printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n",
	       type->name);
915 916
#endif

S
Steven Rostedt 已提交
917
 out_unlock:
918 919 920
	return ret;
}

921 922 923 924 925 926 927 928
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();
929
	ring_buffer_reset_cpu(buffer, cpu);
930 931 932 933

	ring_buffer_record_enable(buffer);
}

934 935
void tracing_reset_online_cpus(struct trace_array *tr)
{
936
	struct ring_buffer *buffer = tr->buffer;
937 938
	int cpu;

939 940 941 942 943
	ring_buffer_record_disable(buffer);

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

944 945 946
	tr->time_start = ftrace_now(tr->cpu);

	for_each_online_cpu(cpu)
947
		ring_buffer_reset_cpu(buffer, cpu);
948 949

	ring_buffer_record_enable(buffer);
950 951
}

952 953 954 955 956 957 958 959 960 961
void tracing_reset_current(int cpu)
{
	tracing_reset(&global_trace, cpu);
}

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

962
#define SAVED_CMDLINES 128
963
#define NO_CMDLINE_MAP UINT_MAX
964 965 966 967
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;
968
static arch_spinlock_t trace_cmdline_lock = __ARCH_SPIN_LOCK_UNLOCKED;
969 970

/* temporary disable recording */
971
static atomic_t trace_record_cmdline_disabled __read_mostly;
972 973 974

static void trace_init_cmdlines(void)
{
975 976
	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));
977 978 979
	cmdline_idx = 0;
}

980 981 982 983 984
int is_tracing_stopped(void)
{
	return trace_stop_count;
}

S
Steven Rostedt 已提交
985 986 987 988 989 990 991 992 993 994 995 996 997 998 999
/**
 * 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();
}

1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013
/**
 * 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;

1014
	raw_spin_lock_irqsave(&tracing_start_lock, flags);
1015 1016 1017 1018 1019 1020
	if (--trace_stop_count) {
		if (trace_stop_count < 0) {
			/* Someone screwed up their debugging */
			WARN_ON_ONCE(1);
			trace_stop_count = 0;
		}
1021 1022 1023
		goto out;
	}

1024 1025
	/* Prevent the buffers from switching */
	arch_spin_lock(&ftrace_max_lock);
1026 1027 1028 1029 1030 1031 1032 1033 1034

	buffer = global_trace.buffer;
	if (buffer)
		ring_buffer_record_enable(buffer);

	buffer = max_tr.buffer;
	if (buffer)
		ring_buffer_record_enable(buffer);

1035 1036
	arch_spin_unlock(&ftrace_max_lock);

1037 1038
	ftrace_start();
 out:
1039
	raw_spin_unlock_irqrestore(&tracing_start_lock, flags);
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053
}

/**
 * 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();
1054
	raw_spin_lock_irqsave(&tracing_start_lock, flags);
1055 1056 1057
	if (trace_stop_count++)
		goto out;

1058 1059 1060
	/* Prevent the buffers from switching */
	arch_spin_lock(&ftrace_max_lock);

1061 1062 1063 1064 1065 1066 1067 1068
	buffer = global_trace.buffer;
	if (buffer)
		ring_buffer_record_disable(buffer);

	buffer = max_tr.buffer;
	if (buffer)
		ring_buffer_record_disable(buffer);

1069 1070
	arch_spin_unlock(&ftrace_max_lock);

1071
 out:
1072
	raw_spin_unlock_irqrestore(&tracing_start_lock, flags);
1073 1074
}

I
Ingo Molnar 已提交
1075
void trace_stop_cmdline_recording(void);
1076

I
Ingo Molnar 已提交
1077
static void trace_save_cmdline(struct task_struct *tsk)
1078
{
1079
	unsigned pid, idx;
1080 1081 1082 1083 1084 1085 1086 1087 1088 1089

	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.
	 */
1090
	if (!arch_spin_trylock(&trace_cmdline_lock))
1091 1092 1093
		return;

	idx = map_pid_to_cmdline[tsk->pid];
1094
	if (idx == NO_CMDLINE_MAP) {
1095 1096
		idx = (cmdline_idx + 1) % SAVED_CMDLINES;

1097 1098 1099 1100 1101 1102 1103 1104 1105
		/*
		 * 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;
1106

1107
		map_cmdline_to_pid[idx] = tsk->pid;
1108 1109 1110 1111 1112 1113 1114
		map_pid_to_cmdline[tsk->pid] = idx;

		cmdline_idx = idx;
	}

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

1115
	arch_spin_unlock(&trace_cmdline_lock);
1116 1117
}

1118
void trace_find_cmdline(int pid, char comm[])
1119 1120 1121
{
	unsigned map;

1122 1123 1124 1125
	if (!pid) {
		strcpy(comm, "<idle>");
		return;
	}
1126

1127 1128 1129 1130 1131
	if (WARN_ON_ONCE(pid < 0)) {
		strcpy(comm, "<XXX>");
		return;
	}

1132 1133 1134 1135
	if (pid > PID_MAX_DEFAULT) {
		strcpy(comm, "<...>");
		return;
	}
1136

1137
	preempt_disable();
1138
	arch_spin_lock(&trace_cmdline_lock);
1139
	map = map_pid_to_cmdline[pid];
1140 1141 1142 1143
	if (map != NO_CMDLINE_MAP)
		strcpy(comm, saved_cmdlines[map]);
	else
		strcpy(comm, "<...>");
1144

1145
	arch_spin_unlock(&trace_cmdline_lock);
1146
	preempt_enable();
1147 1148
}

I
Ingo Molnar 已提交
1149
void tracing_record_cmdline(struct task_struct *tsk)
1150
{
1151
	if (atomic_read(&trace_record_cmdline_disabled) || !tracing_is_on())
1152 1153
		return;

1154 1155 1156 1157 1158
	if (!__this_cpu_read(trace_cmdline_save))
		return;

	__this_cpu_write(trace_cmdline_save, false);

1159 1160 1161
	trace_save_cmdline(tsk);
}

1162
void
1163 1164
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
1165 1166 1167
{
	struct task_struct *tsk = current;

1168 1169
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
1170
	entry->padding			= 0;
1171
	entry->flags =
1172
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
1173
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
1174 1175 1176
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
1177 1178 1179 1180
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}
1181
EXPORT_SYMBOL_GPL(tracing_generic_entry_update);
1182

1183 1184 1185 1186 1187
struct ring_buffer_event *
trace_buffer_lock_reserve(struct ring_buffer *buffer,
			  int type,
			  unsigned long len,
			  unsigned long flags, int pc)
1188 1189 1190
{
	struct ring_buffer_event *event;

1191
	event = ring_buffer_lock_reserve(buffer, len);
1192 1193 1194 1195 1196 1197 1198 1199 1200 1201
	if (event != NULL) {
		struct trace_entry *ent = ring_buffer_event_data(event);

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

	return event;
}

1202 1203 1204 1205
void
__buffer_unlock_commit(struct ring_buffer *buffer, struct ring_buffer_event *event)
{
	__this_cpu_write(trace_cmdline_save, true);
1206 1207 1208 1209 1210
	if (trace_wakeup_needed) {
		trace_wakeup_needed = false;
		/* irq_work_queue() supplies it's own memory barriers */
		irq_work_queue(&trace_work_wakeup);
	}
1211 1212 1213
	ring_buffer_unlock_commit(buffer, event);
}

1214 1215 1216
static inline void
__trace_buffer_unlock_commit(struct ring_buffer *buffer,
			     struct ring_buffer_event *event,
1217
			     unsigned long flags, int pc)
1218
{
1219
	__buffer_unlock_commit(buffer, event);
1220

1221 1222
	ftrace_trace_stack(buffer, flags, 6, pc);
	ftrace_trace_userstack(buffer, flags, pc);
1223 1224
}

1225 1226 1227
void trace_buffer_unlock_commit(struct ring_buffer *buffer,
				struct ring_buffer_event *event,
				unsigned long flags, int pc)
1228
{
1229
	__trace_buffer_unlock_commit(buffer, event, flags, pc);
1230
}
1231
EXPORT_SYMBOL_GPL(trace_buffer_unlock_commit);
1232

1233
struct ring_buffer_event *
1234 1235
trace_current_buffer_lock_reserve(struct ring_buffer **current_rb,
				  int type, unsigned long len,
1236 1237
				  unsigned long flags, int pc)
{
1238 1239
	*current_rb = global_trace.buffer;
	return trace_buffer_lock_reserve(*current_rb,
1240 1241
					 type, len, flags, pc);
}
1242
EXPORT_SYMBOL_GPL(trace_current_buffer_lock_reserve);
1243

1244 1245
void trace_current_buffer_unlock_commit(struct ring_buffer *buffer,
					struct ring_buffer_event *event,
1246 1247
					unsigned long flags, int pc)
{
1248
	__trace_buffer_unlock_commit(buffer, event, flags, pc);
1249
}
1250
EXPORT_SYMBOL_GPL(trace_current_buffer_unlock_commit);
1251

1252 1253 1254 1255
void trace_buffer_unlock_commit_regs(struct ring_buffer *buffer,
				     struct ring_buffer_event *event,
				     unsigned long flags, int pc,
				     struct pt_regs *regs)
1256
{
1257
	__buffer_unlock_commit(buffer, event);
1258 1259 1260 1261

	ftrace_trace_stack_regs(buffer, flags, 0, pc, regs);
	ftrace_trace_userstack(buffer, flags, pc);
}
1262
EXPORT_SYMBOL_GPL(trace_buffer_unlock_commit_regs);
1263

1264 1265
void trace_current_buffer_discard_commit(struct ring_buffer *buffer,
					 struct ring_buffer_event *event)
1266
{
1267
	ring_buffer_discard_commit(buffer, event);
1268
}
1269
EXPORT_SYMBOL_GPL(trace_current_buffer_discard_commit);
1270

I
Ingo Molnar 已提交
1271
void
1272
trace_function(struct trace_array *tr,
1273 1274
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
1275
{
1276
	struct ftrace_event_call *call = &event_function;
1277
	struct ring_buffer *buffer = tr->buffer;
1278
	struct ring_buffer_event *event;
1279
	struct ftrace_entry *entry;
1280

1281
	/* If we are reading the ring buffer, don't trace */
R
Rusty Russell 已提交
1282
	if (unlikely(__this_cpu_read(ftrace_cpu_disabled)))
1283 1284
		return;

1285
	event = trace_buffer_lock_reserve(buffer, TRACE_FN, sizeof(*entry),
1286
					  flags, pc);
1287 1288 1289
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1290 1291
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
1292

1293
	if (!filter_check_discard(call, entry, buffer, event))
1294
		__buffer_unlock_commit(buffer, event);
1295 1296
}

I
Ingo Molnar 已提交
1297
void
I
Ingo Molnar 已提交
1298
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
1299 1300
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
I
Ingo Molnar 已提交
1301 1302
{
	if (likely(!atomic_read(&data->disabled)))
1303
		trace_function(tr, ip, parent_ip, flags, pc);
I
Ingo Molnar 已提交
1304 1305
}

1306
#ifdef CONFIG_STACKTRACE
1307 1308 1309 1310 1311 1312 1313 1314 1315

#define FTRACE_STACK_MAX_ENTRIES (PAGE_SIZE / sizeof(unsigned long))
struct ftrace_stack {
	unsigned long		calls[FTRACE_STACK_MAX_ENTRIES];
};

static DEFINE_PER_CPU(struct ftrace_stack, ftrace_stack);
static DEFINE_PER_CPU(int, ftrace_stack_reserve);

1316
static void __ftrace_trace_stack(struct ring_buffer *buffer,
1317
				 unsigned long flags,
1318
				 int skip, int pc, struct pt_regs *regs)
I
Ingo Molnar 已提交
1319
{
1320
	struct ftrace_event_call *call = &event_kernel_stack;
1321
	struct ring_buffer_event *event;
1322
	struct stack_entry *entry;
I
Ingo Molnar 已提交
1323
	struct stack_trace trace;
1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362
	int use_stack;
	int size = FTRACE_STACK_ENTRIES;

	trace.nr_entries	= 0;
	trace.skip		= skip;

	/*
	 * Since events can happen in NMIs there's no safe way to
	 * use the per cpu ftrace_stacks. We reserve it and if an interrupt
	 * or NMI comes in, it will just have to use the default
	 * FTRACE_STACK_SIZE.
	 */
	preempt_disable_notrace();

	use_stack = ++__get_cpu_var(ftrace_stack_reserve);
	/*
	 * We don't need any atomic variables, just a barrier.
	 * If an interrupt comes in, we don't care, because it would
	 * have exited and put the counter back to what we want.
	 * We just need a barrier to keep gcc from moving things
	 * around.
	 */
	barrier();
	if (use_stack == 1) {
		trace.entries		= &__get_cpu_var(ftrace_stack).calls[0];
		trace.max_entries	= FTRACE_STACK_MAX_ENTRIES;

		if (regs)
			save_stack_trace_regs(regs, &trace);
		else
			save_stack_trace(&trace);

		if (trace.nr_entries > size)
			size = trace.nr_entries;
	} else
		/* From now on, use_stack is a boolean */
		use_stack = 0;

	size *= sizeof(unsigned long);
I
Ingo Molnar 已提交
1363

1364
	event = trace_buffer_lock_reserve(buffer, TRACE_STACK,
1365
					  sizeof(*entry) + size, flags, pc);
1366
	if (!event)
1367 1368
		goto out;
	entry = ring_buffer_event_data(event);
I
Ingo Molnar 已提交
1369

1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384
	memset(&entry->caller, 0, size);

	if (use_stack)
		memcpy(&entry->caller, trace.entries,
		       trace.nr_entries * sizeof(unsigned long));
	else {
		trace.max_entries	= FTRACE_STACK_ENTRIES;
		trace.entries		= entry->caller;
		if (regs)
			save_stack_trace_regs(regs, &trace);
		else
			save_stack_trace(&trace);
	}

	entry->size = trace.nr_entries;
I
Ingo Molnar 已提交
1385

1386
	if (!filter_check_discard(call, entry, buffer, event))
1387
		__buffer_unlock_commit(buffer, event);
1388 1389 1390 1391 1392 1393 1394

 out:
	/* Again, don't let gcc optimize things here */
	barrier();
	__get_cpu_var(ftrace_stack_reserve)--;
	preempt_enable_notrace();

I
Ingo Molnar 已提交
1395 1396
}

1397 1398 1399 1400 1401 1402 1403 1404 1405
void ftrace_trace_stack_regs(struct ring_buffer *buffer, unsigned long flags,
			     int skip, int pc, struct pt_regs *regs)
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

	__ftrace_trace_stack(buffer, flags, skip, pc, regs);
}

1406 1407
void ftrace_trace_stack(struct ring_buffer *buffer, unsigned long flags,
			int skip, int pc)
1408 1409 1410 1411
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

1412
	__ftrace_trace_stack(buffer, flags, skip, pc, NULL);
1413 1414
}

1415 1416
void __trace_stack(struct trace_array *tr, unsigned long flags, int skip,
		   int pc)
1417
{
1418
	__ftrace_trace_stack(tr->buffer, flags, skip, pc, NULL);
1419 1420
}

S
Steven Rostedt 已提交
1421 1422 1423 1424 1425 1426 1427 1428
/**
 * trace_dump_stack - record a stack back trace in the trace buffer
 */
void trace_dump_stack(void)
{
	unsigned long flags;

	if (tracing_disabled || tracing_selftest_running)
1429
		return;
S
Steven Rostedt 已提交
1430 1431 1432 1433

	local_save_flags(flags);

	/* skipping 3 traces, seems to get us at the caller of this function */
1434
	__ftrace_trace_stack(global_trace.buffer, flags, 3, preempt_count(), NULL);
S
Steven Rostedt 已提交
1435 1436
}

1437 1438
static DEFINE_PER_CPU(int, user_stack_count);

1439 1440
void
ftrace_trace_userstack(struct ring_buffer *buffer, unsigned long flags, int pc)
1441
{
1442
	struct ftrace_event_call *call = &event_user_stack;
1443
	struct ring_buffer_event *event;
1444 1445 1446 1447 1448 1449
	struct userstack_entry *entry;
	struct stack_trace trace;

	if (!(trace_flags & TRACE_ITER_USERSTACKTRACE))
		return;

1450 1451 1452 1453 1454 1455
	/*
	 * NMIs can not handle page faults, even with fix ups.
	 * The save user stack can (and often does) fault.
	 */
	if (unlikely(in_nmi()))
		return;
1456

1457 1458 1459 1460 1461 1462 1463 1464 1465 1466
	/*
	 * prevent recursion, since the user stack tracing may
	 * trigger other kernel events.
	 */
	preempt_disable();
	if (__this_cpu_read(user_stack_count))
		goto out;

	__this_cpu_inc(user_stack_count);

1467
	event = trace_buffer_lock_reserve(buffer, TRACE_USER_STACK,
1468
					  sizeof(*entry), flags, pc);
1469
	if (!event)
L
Li Zefan 已提交
1470
		goto out_drop_count;
1471 1472
	entry	= ring_buffer_event_data(event);

1473
	entry->tgid		= current->tgid;
1474 1475 1476 1477 1478 1479 1480 1481
	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);
1482
	if (!filter_check_discard(call, entry, buffer, event))
1483
		__buffer_unlock_commit(buffer, event);
1484

L
Li Zefan 已提交
1485
 out_drop_count:
1486 1487 1488
	__this_cpu_dec(user_stack_count);
 out:
	preempt_enable();
1489 1490
}

1491 1492
#ifdef UNUSED
static void __trace_userstack(struct trace_array *tr, unsigned long flags)
1493
{
1494
	ftrace_trace_userstack(tr, flags, preempt_count());
1495
}
1496
#endif /* UNUSED */
1497

1498 1499
#endif /* CONFIG_STACKTRACE */

1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 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 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583
/* created for use with alloc_percpu */
struct trace_buffer_struct {
	char buffer[TRACE_BUF_SIZE];
};

static struct trace_buffer_struct *trace_percpu_buffer;
static struct trace_buffer_struct *trace_percpu_sirq_buffer;
static struct trace_buffer_struct *trace_percpu_irq_buffer;
static struct trace_buffer_struct *trace_percpu_nmi_buffer;

/*
 * The buffer used is dependent on the context. There is a per cpu
 * buffer for normal context, softirq contex, hard irq context and
 * for NMI context. Thise allows for lockless recording.
 *
 * Note, if the buffers failed to be allocated, then this returns NULL
 */
static char *get_trace_buf(void)
{
	struct trace_buffer_struct *percpu_buffer;
	struct trace_buffer_struct *buffer;

	/*
	 * If we have allocated per cpu buffers, then we do not
	 * need to do any locking.
	 */
	if (in_nmi())
		percpu_buffer = trace_percpu_nmi_buffer;
	else if (in_irq())
		percpu_buffer = trace_percpu_irq_buffer;
	else if (in_softirq())
		percpu_buffer = trace_percpu_sirq_buffer;
	else
		percpu_buffer = trace_percpu_buffer;

	if (!percpu_buffer)
		return NULL;

	buffer = per_cpu_ptr(percpu_buffer, smp_processor_id());

	return buffer->buffer;
}

static int alloc_percpu_trace_buffer(void)
{
	struct trace_buffer_struct *buffers;
	struct trace_buffer_struct *sirq_buffers;
	struct trace_buffer_struct *irq_buffers;
	struct trace_buffer_struct *nmi_buffers;

	buffers = alloc_percpu(struct trace_buffer_struct);
	if (!buffers)
		goto err_warn;

	sirq_buffers = alloc_percpu(struct trace_buffer_struct);
	if (!sirq_buffers)
		goto err_sirq;

	irq_buffers = alloc_percpu(struct trace_buffer_struct);
	if (!irq_buffers)
		goto err_irq;

	nmi_buffers = alloc_percpu(struct trace_buffer_struct);
	if (!nmi_buffers)
		goto err_nmi;

	trace_percpu_buffer = buffers;
	trace_percpu_sirq_buffer = sirq_buffers;
	trace_percpu_irq_buffer = irq_buffers;
	trace_percpu_nmi_buffer = nmi_buffers;

	return 0;

 err_nmi:
	free_percpu(irq_buffers);
 err_irq:
	free_percpu(sirq_buffers);
 err_sirq:
	free_percpu(buffers);
 err_warn:
	WARN(1, "Could not allocate percpu trace_printk buffer");
	return -ENOMEM;
}

1584 1585
static int buffers_allocated;

1586 1587 1588 1589 1590 1591 1592 1593 1594 1595
void trace_printk_init_buffers(void)
{
	if (buffers_allocated)
		return;

	if (alloc_percpu_trace_buffer())
		return;

	pr_info("ftrace: Allocated trace_printk buffers\n");

1596 1597 1598
	/* Expand the buffers to set size */
	tracing_update_buffers();

1599
	buffers_allocated = 1;
1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627

	/*
	 * trace_printk_init_buffers() can be called by modules.
	 * If that happens, then we need to start cmdline recording
	 * directly here. If the global_trace.buffer is already
	 * allocated here, then this was called by module code.
	 */
	if (global_trace.buffer)
		tracing_start_cmdline_record();
}

void trace_printk_start_comm(void)
{
	/* Start tracing comms if trace printk is set */
	if (!buffers_allocated)
		return;
	tracing_start_cmdline_record();
}

static void trace_printk_start_stop_comm(int enabled)
{
	if (!buffers_allocated)
		return;

	if (enabled)
		tracing_start_cmdline_record();
	else
		tracing_stop_cmdline_record();
1628 1629
}

1630
/**
1631
 * trace_vbprintk - write binary msg to tracing buffer
1632 1633
 *
 */
1634
int trace_vbprintk(unsigned long ip, const char *fmt, va_list args)
1635
{
1636
	struct ftrace_event_call *call = &event_bprint;
1637
	struct ring_buffer_event *event;
1638
	struct ring_buffer *buffer;
1639
	struct trace_array *tr = &global_trace;
1640
	struct bprint_entry *entry;
1641
	unsigned long flags;
1642 1643
	char *tbuffer;
	int len = 0, size, pc;
1644 1645 1646 1647 1648 1649 1650 1651

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

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

	pc = preempt_count();
1652
	preempt_disable_notrace();
1653

1654 1655 1656
	tbuffer = get_trace_buf();
	if (!tbuffer) {
		len = 0;
1657
		goto out;
1658
	}
1659

1660
	len = vbin_printf((u32 *)tbuffer, TRACE_BUF_SIZE/sizeof(int), fmt, args);
1661

1662 1663
	if (len > TRACE_BUF_SIZE/sizeof(int) || len < 0)
		goto out;
1664

1665
	local_save_flags(flags);
1666
	size = sizeof(*entry) + sizeof(u32) * len;
1667 1668 1669
	buffer = tr->buffer;
	event = trace_buffer_lock_reserve(buffer, TRACE_BPRINT, size,
					  flags, pc);
1670
	if (!event)
1671
		goto out;
1672 1673 1674 1675
	entry = ring_buffer_event_data(event);
	entry->ip			= ip;
	entry->fmt			= fmt;

1676
	memcpy(entry->buf, tbuffer, sizeof(u32) * len);
1677
	if (!filter_check_discard(call, entry, buffer, event)) {
1678
		__buffer_unlock_commit(buffer, event);
1679 1680
		ftrace_trace_stack(buffer, flags, 6, pc);
	}
1681 1682

out:
1683
	preempt_enable_notrace();
1684 1685 1686 1687
	unpause_graph_tracing();

	return len;
}
1688 1689
EXPORT_SYMBOL_GPL(trace_vbprintk);

1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706
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)
1707
{
1708
	struct ftrace_event_call *call = &event_print;
1709
	struct ring_buffer_event *event;
1710
	struct ring_buffer *buffer;
1711
	int len = 0, size, pc;
1712
	struct print_entry *entry;
1713 1714
	unsigned long flags;
	char *tbuffer;
1715 1716 1717 1718

	if (tracing_disabled || tracing_selftest_running)
		return 0;

1719 1720 1721
	/* Don't pollute graph traces with trace_vprintk internals */
	pause_graph_tracing();

1722 1723 1724
	pc = preempt_count();
	preempt_disable_notrace();

1725 1726 1727 1728

	tbuffer = get_trace_buf();
	if (!tbuffer) {
		len = 0;
1729
		goto out;
1730
	}
1731

1732 1733 1734
	len = vsnprintf(tbuffer, TRACE_BUF_SIZE, fmt, args);
	if (len > TRACE_BUF_SIZE)
		goto out;
1735

1736
	local_save_flags(flags);
1737
	size = sizeof(*entry) + len + 1;
1738 1739
	buffer = tr->buffer;
	event = trace_buffer_lock_reserve(buffer, TRACE_PRINT, size,
1740
					  flags, pc);
1741
	if (!event)
1742
		goto out;
1743
	entry = ring_buffer_event_data(event);
C
Carsten Emde 已提交
1744
	entry->ip = ip;
1745

1746
	memcpy(&entry->buf, tbuffer, len);
C
Carsten Emde 已提交
1747
	entry->buf[len] = '\0';
1748
	if (!filter_check_discard(call, entry, buffer, event)) {
1749
		__buffer_unlock_commit(buffer, event);
1750
		ftrace_trace_stack(buffer, flags, 6, pc);
1751
	}
1752 1753
 out:
	preempt_enable_notrace();
1754
	unpause_graph_tracing();
1755 1756 1757

	return len;
}
1758 1759 1760

int trace_vprintk(unsigned long ip, const char *fmt, va_list args)
{
S
Steven Rostedt 已提交
1761
	return trace_array_vprintk(&global_trace, ip, fmt, args);
1762
}
1763 1764
EXPORT_SYMBOL_GPL(trace_vprintk);

1765
static void trace_iterator_increment(struct trace_iterator *iter)
S
Steven Rostedt 已提交
1766
{
1767 1768
	struct ring_buffer_iter *buf_iter = trace_buffer_iter(iter, iter->cpu);

S
Steven Rostedt 已提交
1769
	iter->idx++;
1770 1771
	if (buf_iter)
		ring_buffer_read(buf_iter, NULL);
S
Steven Rostedt 已提交
1772 1773
}

I
Ingo Molnar 已提交
1774
static struct trace_entry *
1775 1776
peek_next_entry(struct trace_iterator *iter, int cpu, u64 *ts,
		unsigned long *lost_events)
1777
{
1778
	struct ring_buffer_event *event;
1779
	struct ring_buffer_iter *buf_iter = trace_buffer_iter(iter, cpu);
1780

1781 1782 1783
	if (buf_iter)
		event = ring_buffer_iter_peek(buf_iter, ts);
	else
1784 1785
		event = ring_buffer_peek(iter->tr->buffer, cpu, ts,
					 lost_events);
1786

1787 1788 1789 1790 1791 1792
	if (event) {
		iter->ent_size = ring_buffer_event_length(event);
		return ring_buffer_event_data(event);
	}
	iter->ent_size = 0;
	return NULL;
1793
}
1794

1795
static struct trace_entry *
1796 1797
__find_next_entry(struct trace_iterator *iter, int *ent_cpu,
		  unsigned long *missing_events, u64 *ent_ts)
1798
{
1799
	struct ring_buffer *buffer = iter->tr->buffer;
1800
	struct trace_entry *ent, *next = NULL;
1801
	unsigned long lost_events = 0, next_lost = 0;
1802
	int cpu_file = iter->cpu_file;
1803
	u64 next_ts = 0, ts;
1804
	int next_cpu = -1;
1805
	int next_size = 0;
1806 1807
	int cpu;

1808 1809 1810 1811 1812 1813 1814
	/*
	 * 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;
1815
		ent = peek_next_entry(iter, cpu_file, ent_ts, missing_events);
1816 1817 1818 1819 1820 1821
		if (ent_cpu)
			*ent_cpu = cpu_file;

		return ent;
	}

1822
	for_each_tracing_cpu(cpu) {
1823

1824 1825
		if (ring_buffer_empty_cpu(buffer, cpu))
			continue;
1826

1827
		ent = peek_next_entry(iter, cpu, &ts, &lost_events);
1828

I
Ingo Molnar 已提交
1829 1830 1831
		/*
		 * Pick the entry with the smallest timestamp:
		 */
1832
		if (ent && (!next || ts < next_ts)) {
1833 1834
			next = ent;
			next_cpu = cpu;
1835
			next_ts = ts;
1836
			next_lost = lost_events;
1837
			next_size = iter->ent_size;
1838 1839 1840
		}
	}

1841 1842
	iter->ent_size = next_size;

1843 1844 1845
	if (ent_cpu)
		*ent_cpu = next_cpu;

1846 1847 1848
	if (ent_ts)
		*ent_ts = next_ts;

1849 1850 1851
	if (missing_events)
		*missing_events = next_lost;

1852 1853 1854
	return next;
}

1855
/* Find the next real entry, without updating the iterator itself */
1856 1857
struct trace_entry *trace_find_next_entry(struct trace_iterator *iter,
					  int *ent_cpu, u64 *ent_ts)
1858
{
1859
	return __find_next_entry(iter, ent_cpu, NULL, ent_ts);
1860 1861 1862
}

/* Find the next real entry, and increment the iterator to the next entry */
1863
void *trace_find_next_entry_inc(struct trace_iterator *iter)
1864
{
1865 1866
	iter->ent = __find_next_entry(iter, &iter->cpu,
				      &iter->lost_events, &iter->ts);
1867

1868
	if (iter->ent)
1869
		trace_iterator_increment(iter);
1870

1871
	return iter->ent ? iter : NULL;
1872
}
1873

I
Ingo Molnar 已提交
1874
static void trace_consume(struct trace_iterator *iter)
1875
{
1876 1877
	ring_buffer_consume(iter->tr->buffer, iter->cpu, &iter->ts,
			    &iter->lost_events);
1878 1879
}

I
Ingo Molnar 已提交
1880
static void *s_next(struct seq_file *m, void *v, loff_t *pos)
1881 1882 1883
{
	struct trace_iterator *iter = m->private;
	int i = (int)*pos;
I
Ingo Molnar 已提交
1884
	void *ent;
1885

1886 1887
	WARN_ON_ONCE(iter->leftover);

1888 1889 1890 1891 1892 1893 1894
	(*pos)++;

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

	if (iter->idx < 0)
1895
		ent = trace_find_next_entry_inc(iter);
1896 1897 1898 1899
	else
		ent = iter;

	while (ent && iter->idx < i)
1900
		ent = trace_find_next_entry_inc(iter);
1901 1902 1903 1904 1905 1906

	iter->pos = *pos;

	return ent;
}

1907
void tracing_iter_reset(struct trace_iterator *iter, int cpu)
1908 1909 1910 1911 1912 1913 1914 1915 1916
{
	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;

1917 1918
	buf_iter = trace_buffer_iter(iter, cpu);
	if (!buf_iter)
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937
		return;

	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;
}

1938 1939 1940 1941
/*
 * The current tracer is copied to avoid a global locking
 * all around.
 */
1942 1943 1944
static void *s_start(struct seq_file *m, loff_t *pos)
{
	struct trace_iterator *iter = m->private;
1945
	static struct tracer *old_tracer;
1946
	int cpu_file = iter->cpu_file;
1947 1948
	void *p = NULL;
	loff_t l = 0;
1949
	int cpu;
1950

1951
	/* copy the tracer to avoid using a global lock all around */
1952
	mutex_lock(&trace_types_lock);
1953 1954 1955
	if (unlikely(old_tracer != current_trace && current_trace)) {
		old_tracer = current_trace;
		*iter->trace = *current_trace;
1956
	}
1957
	mutex_unlock(&trace_types_lock);
1958 1959 1960 1961 1962 1963 1964 1965

	atomic_inc(&trace_record_cmdline_disabled);

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

1966 1967
		if (cpu_file == TRACE_PIPE_ALL_CPU) {
			for_each_tracing_cpu(cpu)
1968
				tracing_iter_reset(iter, cpu);
1969
		} else
1970
			tracing_iter_reset(iter, cpu_file);
1971

1972
		iter->leftover = 0;
1973 1974 1975 1976
		for (p = iter; p && l < *pos; p = s_next(m, p, &l))
			;

	} else {
1977 1978 1979 1980 1981 1982 1983 1984 1985 1986
		/*
		 * 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);
		}
1987 1988
	}

1989
	trace_event_read_lock();
1990
	trace_access_lock(cpu_file);
1991 1992 1993 1994 1995
	return p;
}

static void s_stop(struct seq_file *m, void *p)
{
1996 1997
	struct trace_iterator *iter = m->private;

1998
	atomic_dec(&trace_record_cmdline_disabled);
1999
	trace_access_unlock(iter->cpu_file);
2000
	trace_event_read_unlock();
2001 2002
}

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
static void
get_total_entries(struct trace_array *tr, unsigned long *total, unsigned long *entries)
{
	unsigned long count;
	int cpu;

	*total = 0;
	*entries = 0;

	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;
	}
}

I
Ingo Molnar 已提交
2030
static void print_lat_help_header(struct seq_file *m)
2031
{
2032 2033 2034 2035 2036
	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");
2037 2038 2039
	seq_puts(m, "#                |||| /     delay             \n");
	seq_puts(m, "#  cmd     pid   ||||| time  |   caller      \n");
	seq_puts(m, "#     \\   /      |||||  \\    |   /           \n");
2040 2041
}

2042
static void print_event_info(struct trace_array *tr, struct seq_file *m)
2043
{
2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055
	unsigned long total;
	unsigned long entries;

	get_total_entries(tr, &total, &entries);
	seq_printf(m, "# entries-in-buffer/entries-written: %lu/%lu   #P:%d\n",
		   entries, total, num_online_cpus());
	seq_puts(m, "#\n");
}

static void print_func_help_header(struct trace_array *tr, struct seq_file *m)
{
	print_event_info(tr, m);
2056
	seq_puts(m, "#           TASK-PID   CPU#      TIMESTAMP  FUNCTION\n");
2057
	seq_puts(m, "#              | |       |          |         |\n");
2058 2059
}

2060
static void print_func_help_header_irq(struct trace_array *tr, struct seq_file *m)
2061
{
2062
	print_event_info(tr, m);
2063 2064 2065 2066 2067 2068 2069 2070
	seq_puts(m, "#                              _-----=> irqs-off\n");
	seq_puts(m, "#                             / _----=> need-resched\n");
	seq_puts(m, "#                            | / _---=> hardirq/softirq\n");
	seq_puts(m, "#                            || / _--=> preempt-depth\n");
	seq_puts(m, "#                            ||| /     delay\n");
	seq_puts(m, "#           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION\n");
	seq_puts(m, "#              | |       |   ||||       |         |\n");
}
2071

2072
void
2073 2074 2075 2076 2077 2078
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;
2079 2080
	unsigned long entries;
	unsigned long total;
2081 2082 2083 2084 2085
	const char *name = "preemption";

	if (type)
		name = type->name;

2086
	get_total_entries(tr, &total, &entries);
2087

2088
	seq_printf(m, "# %s latency trace v1.1.5 on %s\n",
2089
		   name, UTS_RELEASE);
2090
	seq_puts(m, "# -----------------------------------"
2091
		 "---------------------------------\n");
2092
	seq_printf(m, "# latency: %lu us, #%lu/%lu, CPU#%d |"
2093
		   " (M:%s VP:%d, KP:%d, SP:%d HP:%d",
S
Steven Rostedt 已提交
2094
		   nsecs_to_usecs(data->saved_latency),
2095
		   entries,
2096
		   total,
2097 2098 2099 2100 2101
		   tr->cpu,
#if defined(CONFIG_PREEMPT_NONE)
		   "server",
#elif defined(CONFIG_PREEMPT_VOLUNTARY)
		   "desktop",
2102
#elif defined(CONFIG_PREEMPT)
2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113
		   "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
2114 2115
	seq_puts(m, "#    -----------------\n");
	seq_printf(m, "#    | task: %.16s-%d "
2116
		   "(uid:%d nice:%ld policy:%ld rt_prio:%ld)\n",
2117 2118
		   data->comm, data->pid,
		   from_kuid_munged(seq_user_ns(m), data->uid), data->nice,
2119
		   data->policy, data->rt_priority);
2120
	seq_puts(m, "#    -----------------\n");
2121 2122

	if (data->critical_start) {
2123
		seq_puts(m, "#  => started at: ");
S
Steven Rostedt 已提交
2124 2125
		seq_print_ip_sym(&iter->seq, data->critical_start, sym_flags);
		trace_print_seq(m, &iter->seq);
2126
		seq_puts(m, "\n#  => ended at:   ");
S
Steven Rostedt 已提交
2127 2128
		seq_print_ip_sym(&iter->seq, data->critical_end, sym_flags);
		trace_print_seq(m, &iter->seq);
2129
		seq_puts(m, "\n#\n");
2130 2131
	}

2132
	seq_puts(m, "#\n");
2133 2134
}

2135 2136 2137 2138
static void test_cpu_buff_start(struct trace_iterator *iter)
{
	struct trace_seq *s = &iter->seq;

2139 2140 2141 2142 2143 2144
	if (!(trace_flags & TRACE_ITER_ANNOTATE))
		return;

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

2145
	if (cpumask_test_cpu(iter->cpu, iter->started))
2146 2147
		return;

2148 2149 2150
	if (iter->tr->data[iter->cpu]->skipped_entries)
		return;

2151
	cpumask_set_cpu(iter->cpu, iter->started);
2152 2153 2154 2155 2156

	/* 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);
2157 2158
}

2159
static enum print_line_t print_trace_fmt(struct trace_iterator *iter)
2160
{
S
Steven Rostedt 已提交
2161
	struct trace_seq *s = &iter->seq;
2162
	unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK);
I
Ingo Molnar 已提交
2163
	struct trace_entry *entry;
2164
	struct trace_event *event;
2165

I
Ingo Molnar 已提交
2166
	entry = iter->ent;
2167

2168 2169
	test_cpu_buff_start(iter);

2170
	event = ftrace_find_event(entry->type);
2171

2172
	if (trace_flags & TRACE_ITER_CONTEXT_INFO) {
2173 2174 2175 2176 2177 2178 2179
		if (iter->iter_flags & TRACE_FILE_LAT_FMT) {
			if (!trace_print_lat_context(iter))
				goto partial;
		} else {
			if (!trace_print_context(iter))
				goto partial;
		}
2180
	}
2181

2182
	if (event)
2183
		return event->funcs->trace(iter, sym_flags, event);
2184 2185 2186

	if (!trace_seq_printf(s, "Unknown type %d\n", entry->type))
		goto partial;
2187

2188
	return TRACE_TYPE_HANDLED;
2189 2190
partial:
	return TRACE_TYPE_PARTIAL_LINE;
2191 2192
}

2193
static enum print_line_t print_raw_fmt(struct trace_iterator *iter)
I
Ingo Molnar 已提交
2194 2195 2196
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *entry;
2197
	struct trace_event *event;
I
Ingo Molnar 已提交
2198 2199

	entry = iter->ent;
2200

2201
	if (trace_flags & TRACE_ITER_CONTEXT_INFO) {
2202 2203 2204
		if (!trace_seq_printf(s, "%d %d %llu ",
				      entry->pid, iter->cpu, iter->ts))
			goto partial;
2205
	}
I
Ingo Molnar 已提交
2206

2207
	event = ftrace_find_event(entry->type);
2208
	if (event)
2209
		return event->funcs->raw(iter, 0, event);
2210 2211 2212

	if (!trace_seq_printf(s, "%d ?\n", entry->type))
		goto partial;
2213

2214
	return TRACE_TYPE_HANDLED;
2215 2216
partial:
	return TRACE_TYPE_PARTIAL_LINE;
I
Ingo Molnar 已提交
2217 2218
}

2219
static enum print_line_t print_hex_fmt(struct trace_iterator *iter)
2220 2221 2222 2223
{
	struct trace_seq *s = &iter->seq;
	unsigned char newline = '\n';
	struct trace_entry *entry;
2224
	struct trace_event *event;
2225 2226

	entry = iter->ent;
2227

2228 2229 2230 2231 2232
	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);
	}
2233

2234
	event = ftrace_find_event(entry->type);
2235
	if (event) {
2236
		enum print_line_t ret = event->funcs->hex(iter, 0, event);
2237 2238 2239
		if (ret != TRACE_TYPE_HANDLED)
			return ret;
	}
S
Steven Rostedt 已提交
2240

2241 2242
	SEQ_PUT_FIELD_RET(s, newline);

2243
	return TRACE_TYPE_HANDLED;
2244 2245
}

2246
static enum print_line_t print_bin_fmt(struct trace_iterator *iter)
I
Ingo Molnar 已提交
2247 2248 2249
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *entry;
2250
	struct trace_event *event;
I
Ingo Molnar 已提交
2251 2252

	entry = iter->ent;
2253

2254 2255
	if (trace_flags & TRACE_ITER_CONTEXT_INFO) {
		SEQ_PUT_FIELD_RET(s, entry->pid);
2256
		SEQ_PUT_FIELD_RET(s, iter->cpu);
2257 2258
		SEQ_PUT_FIELD_RET(s, iter->ts);
	}
I
Ingo Molnar 已提交
2259

2260
	event = ftrace_find_event(entry->type);
2261 2262
	return event ? event->funcs->binary(iter, 0, event) :
		TRACE_TYPE_HANDLED;
I
Ingo Molnar 已提交
2263 2264
}

2265
int trace_empty(struct trace_iterator *iter)
2266
{
2267
	struct ring_buffer_iter *buf_iter;
2268 2269
	int cpu;

2270 2271 2272
	/* If we are looking at one CPU buffer, only check that one */
	if (iter->cpu_file != TRACE_PIPE_ALL_CPU) {
		cpu = iter->cpu_file;
2273 2274 2275
		buf_iter = trace_buffer_iter(iter, cpu);
		if (buf_iter) {
			if (!ring_buffer_iter_empty(buf_iter))
2276 2277 2278 2279 2280 2281 2282 2283
				return 0;
		} else {
			if (!ring_buffer_empty_cpu(iter->tr->buffer, cpu))
				return 0;
		}
		return 1;
	}

2284
	for_each_tracing_cpu(cpu) {
2285 2286 2287
		buf_iter = trace_buffer_iter(iter, cpu);
		if (buf_iter) {
			if (!ring_buffer_iter_empty(buf_iter))
2288 2289 2290 2291 2292
				return 0;
		} else {
			if (!ring_buffer_empty_cpu(iter->tr->buffer, cpu))
				return 0;
		}
2293
	}
2294

2295
	return 1;
2296 2297
}

2298
/*  Called with trace_event_read_lock() held. */
2299
enum print_line_t print_trace_line(struct trace_iterator *iter)
I
Ingo Molnar 已提交
2300
{
2301 2302
	enum print_line_t ret;

2303 2304 2305 2306
	if (iter->lost_events &&
	    !trace_seq_printf(&iter->seq, "CPU:%d [LOST %lu EVENTS]\n",
				 iter->cpu, iter->lost_events))
		return TRACE_TYPE_PARTIAL_LINE;
2307

2308 2309 2310 2311 2312
	if (iter->trace && iter->trace->print_line) {
		ret = iter->trace->print_line(iter);
		if (ret != TRACE_TYPE_UNHANDLED)
			return ret;
	}
2313

2314 2315 2316
	if (iter->ent->type == TRACE_BPRINT &&
			trace_flags & TRACE_ITER_PRINTK &&
			trace_flags & TRACE_ITER_PRINTK_MSGONLY)
2317
		return trace_print_bprintk_msg_only(iter);
2318

2319 2320 2321
	if (iter->ent->type == TRACE_PRINT &&
			trace_flags & TRACE_ITER_PRINTK &&
			trace_flags & TRACE_ITER_PRINTK_MSGONLY)
2322
		return trace_print_printk_msg_only(iter);
2323

I
Ingo Molnar 已提交
2324 2325 2326
	if (trace_flags & TRACE_ITER_BIN)
		return print_bin_fmt(iter);

2327 2328 2329
	if (trace_flags & TRACE_ITER_HEX)
		return print_hex_fmt(iter);

I
Ingo Molnar 已提交
2330 2331 2332 2333 2334 2335
	if (trace_flags & TRACE_ITER_RAW)
		return print_raw_fmt(iter);

	return print_trace_fmt(iter);
}

2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350
void trace_latency_header(struct seq_file *m)
{
	struct trace_iterator *iter = m->private;

	/* print nothing if the buffers are empty */
	if (trace_empty(iter))
		return;

	if (iter->iter_flags & TRACE_FILE_LAT_FMT)
		print_trace_header(m, iter);

	if (!(trace_flags & TRACE_ITER_VERBOSE))
		print_lat_help_header(m);
}

2351 2352 2353 2354
void trace_default_header(struct seq_file *m)
{
	struct trace_iterator *iter = m->private;

2355 2356 2357
	if (!(trace_flags & TRACE_ITER_CONTEXT_INFO))
		return;

2358 2359 2360 2361 2362 2363 2364 2365
	if (iter->iter_flags & TRACE_FILE_LAT_FMT) {
		/* print nothing if the buffers are empty */
		if (trace_empty(iter))
			return;
		print_trace_header(m, iter);
		if (!(trace_flags & TRACE_ITER_VERBOSE))
			print_lat_help_header(m);
	} else {
2366 2367
		if (!(trace_flags & TRACE_ITER_VERBOSE)) {
			if (trace_flags & TRACE_ITER_IRQ_INFO)
2368
				print_func_help_header_irq(iter->tr, m);
2369
			else
2370
				print_func_help_header(iter->tr, m);
2371
		}
2372 2373 2374
	}
}

2375 2376 2377 2378 2379 2380 2381 2382
static void test_ftrace_alive(struct seq_file *m)
{
	if (!ftrace_is_dead())
		return;
	seq_printf(m, "# WARNING: FUNCTION TRACING IS CORRUPTED\n");
	seq_printf(m, "#          MAY BE MISSING FUNCTION EVENTS\n");
}

2383 2384 2385
static int s_show(struct seq_file *m, void *v)
{
	struct trace_iterator *iter = v;
2386
	int ret;
2387 2388 2389 2390 2391

	if (iter->ent == NULL) {
		if (iter->tr) {
			seq_printf(m, "# tracer: %s\n", iter->trace->name);
			seq_puts(m, "#\n");
2392
			test_ftrace_alive(m);
2393
		}
2394 2395
		if (iter->trace && iter->trace->print_header)
			iter->trace->print_header(m);
2396 2397 2398
		else
			trace_default_header(m);

2399 2400 2401 2402 2403 2404 2405 2406 2407 2408
	} 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;

2409
	} else {
I
Ingo Molnar 已提交
2410
		print_trace_line(iter);
2411 2412 2413 2414 2415 2416 2417 2418 2419
		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;
2420 2421 2422 2423 2424
	}

	return 0;
}

J
James Morris 已提交
2425
static const struct seq_operations tracer_seq_ops = {
I
Ingo Molnar 已提交
2426 2427 2428 2429
	.start		= s_start,
	.next		= s_next,
	.stop		= s_stop,
	.show		= s_show,
2430 2431
};

I
Ingo Molnar 已提交
2432
static struct trace_iterator *
2433
__tracing_open(struct inode *inode, struct file *file)
2434
{
2435
	long cpu_file = (long) inode->i_private;
2436
	struct trace_iterator *iter;
2437
	int cpu;
2438

2439 2440
	if (tracing_disabled)
		return ERR_PTR(-ENODEV);
S
Steven Rostedt 已提交
2441

2442
	iter = __seq_open_private(file, &tracer_seq_ops, sizeof(*iter));
2443 2444
	if (!iter)
		return ERR_PTR(-ENOMEM);
2445

2446 2447
	iter->buffer_iter = kzalloc(sizeof(*iter->buffer_iter) * num_possible_cpus(),
				    GFP_KERNEL);
2448 2449 2450
	if (!iter->buffer_iter)
		goto release;

2451 2452 2453 2454
	/*
	 * We make a copy of the current tracer to avoid concurrent
	 * changes on it while we are reading.
	 */
2455
	mutex_lock(&trace_types_lock);
2456
	iter->trace = kzalloc(sizeof(*iter->trace), GFP_KERNEL);
2457
	if (!iter->trace)
2458
		goto fail;
2459

2460 2461 2462
	if (current_trace)
		*iter->trace = *current_trace;

2463
	if (!zalloc_cpumask_var(&iter->started, GFP_KERNEL))
2464 2465
		goto fail;

2466 2467 2468
	if (current_trace && current_trace->print_max)
		iter->tr = &max_tr;
	else
2469
		iter->tr = &global_trace;
2470
	iter->pos = -1;
2471
	mutex_init(&iter->mutex);
2472
	iter->cpu_file = cpu_file;
2473

2474 2475
	/* Notify the tracer early; before we stop tracing. */
	if (iter->trace && iter->trace->open)
2476
		iter->trace->open(iter);
2477

2478 2479 2480 2481
	/* Annotate start of buffers if we had overruns */
	if (ring_buffer_overruns(iter->tr->buffer))
		iter->iter_flags |= TRACE_FILE_ANNOTATE;

2482 2483 2484 2485
	/* Output in nanoseconds only if we are using a clock in nanoseconds. */
	if (trace_clocks[trace_clock_id].in_ns)
		iter->iter_flags |= TRACE_FILE_TIME_IN_NS;

2486 2487 2488
	/* stop the trace while dumping */
	tracing_stop();

2489 2490 2491
	if (iter->cpu_file == TRACE_PIPE_ALL_CPU) {
		for_each_tracing_cpu(cpu) {
			iter->buffer_iter[cpu] =
2492 2493 2494 2495 2496
				ring_buffer_read_prepare(iter->tr->buffer, cpu);
		}
		ring_buffer_read_prepare_sync();
		for_each_tracing_cpu(cpu) {
			ring_buffer_read_start(iter->buffer_iter[cpu]);
2497
			tracing_iter_reset(iter, cpu);
2498 2499 2500
		}
	} else {
		cpu = iter->cpu_file;
2501
		iter->buffer_iter[cpu] =
2502 2503 2504
			ring_buffer_read_prepare(iter->tr->buffer, cpu);
		ring_buffer_read_prepare_sync();
		ring_buffer_read_start(iter->buffer_iter[cpu]);
2505
		tracing_iter_reset(iter, cpu);
2506 2507
	}

2508 2509 2510
	mutex_unlock(&trace_types_lock);

	return iter;
2511

2512
 fail:
2513
	mutex_unlock(&trace_types_lock);
2514
	kfree(iter->trace);
2515
	kfree(iter->buffer_iter);
2516
release:
2517 2518
	seq_release_private(inode, file);
	return ERR_PTR(-ENOMEM);
2519 2520 2521 2522
}

int tracing_open_generic(struct inode *inode, struct file *filp)
{
S
Steven Rostedt 已提交
2523 2524 2525
	if (tracing_disabled)
		return -ENODEV;

2526 2527 2528 2529
	filp->private_data = inode->i_private;
	return 0;
}

2530
static int tracing_release(struct inode *inode, struct file *file)
2531
{
2532
	struct seq_file *m = file->private_data;
2533
	struct trace_iterator *iter;
2534
	int cpu;
2535

2536 2537 2538 2539 2540
	if (!(file->f_mode & FMODE_READ))
		return 0;

	iter = m->private;

2541
	mutex_lock(&trace_types_lock);
2542 2543 2544 2545 2546
	for_each_tracing_cpu(cpu) {
		if (iter->buffer_iter[cpu])
			ring_buffer_read_finish(iter->buffer_iter[cpu]);
	}

2547 2548 2549 2550
	if (iter->trace && iter->trace->close)
		iter->trace->close(iter);

	/* reenable tracing if it was previously enabled */
2551
	tracing_start();
2552 2553
	mutex_unlock(&trace_types_lock);

2554
	mutex_destroy(&iter->mutex);
2555
	free_cpumask_var(iter->started);
2556
	kfree(iter->trace);
2557
	kfree(iter->buffer_iter);
2558
	seq_release_private(inode, file);
2559 2560 2561 2562 2563
	return 0;
}

static int tracing_open(struct inode *inode, struct file *file)
{
2564 2565
	struct trace_iterator *iter;
	int ret = 0;
2566

2567 2568
	/* If this file was open for write, then erase contents */
	if ((file->f_mode & FMODE_WRITE) &&
2569
	    (file->f_flags & O_TRUNC)) {
2570
		long cpu = (long) inode->i_private;
2571

2572 2573 2574 2575 2576
		if (cpu == TRACE_PIPE_ALL_CPU)
			tracing_reset_online_cpus(&global_trace);
		else
			tracing_reset(&global_trace, cpu);
	}
2577

2578 2579 2580 2581 2582 2583 2584
	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;
	}
2585 2586 2587
	return ret;
}

I
Ingo Molnar 已提交
2588
static void *
2589 2590
t_next(struct seq_file *m, void *v, loff_t *pos)
{
L
Li Zefan 已提交
2591
	struct tracer *t = v;
2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602

	(*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
2603
	struct tracer *t;
2604 2605 2606
	loff_t l = 0;

	mutex_lock(&trace_types_lock);
L
Li Zefan 已提交
2607
	for (t = trace_types; t && l < *pos; t = t_next(m, t, &l))
2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633
		;

	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
2634
static const struct seq_operations show_traces_seq_ops = {
I
Ingo Molnar 已提交
2635 2636 2637 2638
	.start		= t_start,
	.next		= t_next,
	.stop		= t_stop,
	.show		= t_show,
2639 2640 2641 2642
};

static int show_traces_open(struct inode *inode, struct file *file)
{
S
Steven Rostedt 已提交
2643 2644 2645
	if (tracing_disabled)
		return -ENODEV;

L
Li Zefan 已提交
2646
	return seq_open(file, &show_traces_seq_ops);
2647 2648
}

2649 2650 2651 2652 2653 2654 2655
static ssize_t
tracing_write_stub(struct file *filp, const char __user *ubuf,
		   size_t count, loff_t *ppos)
{
	return count;
}

2656 2657 2658 2659 2660 2661 2662 2663
static loff_t tracing_seek(struct file *file, loff_t offset, int origin)
{
	if (file->f_mode & FMODE_READ)
		return seq_lseek(file, offset, origin);
	else
		return 0;
}

2664
static const struct file_operations tracing_fops = {
I
Ingo Molnar 已提交
2665 2666
	.open		= tracing_open,
	.read		= seq_read,
2667
	.write		= tracing_write_stub,
2668
	.llseek		= tracing_seek,
I
Ingo Molnar 已提交
2669
	.release	= tracing_release,
2670 2671
};

2672
static const struct file_operations show_traces_fops = {
I
Ingo Molnar 已提交
2673 2674 2675
	.open		= show_traces_open,
	.read		= seq_read,
	.release	= seq_release,
2676
	.llseek		= seq_lseek,
I
Ingo Molnar 已提交
2677 2678
};

2679 2680 2681
/*
 * Only trace on a CPU if the bitmask is set:
 */
2682
static cpumask_var_t tracing_cpumask;
2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695

/*
 * 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
2696 2697 2698 2699
static ssize_t
tracing_cpumask_read(struct file *filp, char __user *ubuf,
		     size_t count, loff_t *ppos)
{
2700
	int len;
I
Ingo Molnar 已提交
2701 2702

	mutex_lock(&tracing_cpumask_update_lock);
2703

2704
	len = cpumask_scnprintf(mask_str, count, tracing_cpumask);
2705 2706 2707 2708 2709 2710 2711 2712
	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
2713 2714 2715 2716 2717 2718 2719 2720 2721
	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)
{
2722
	int err, cpu;
2723 2724 2725 2726
	cpumask_var_t tracing_cpumask_new;

	if (!alloc_cpumask_var(&tracing_cpumask_new, GFP_KERNEL))
		return -ENOMEM;
I
Ingo Molnar 已提交
2727

2728
	err = cpumask_parse_user(ubuf, count, tracing_cpumask_new);
I
Ingo Molnar 已提交
2729
	if (err)
2730 2731
		goto err_unlock;

2732 2733
	mutex_lock(&tracing_cpumask_update_lock);

2734
	local_irq_disable();
2735
	arch_spin_lock(&ftrace_max_lock);
2736
	for_each_tracing_cpu(cpu) {
2737 2738 2739 2740
		/*
		 * Increase/decrease the disabled counter if we are
		 * about to flip a bit in the cpumask:
		 */
2741 2742
		if (cpumask_test_cpu(cpu, tracing_cpumask) &&
				!cpumask_test_cpu(cpu, tracing_cpumask_new)) {
2743
			atomic_inc(&global_trace.data[cpu]->disabled);
2744
			ring_buffer_record_disable_cpu(global_trace.buffer, cpu);
2745
		}
2746 2747
		if (!cpumask_test_cpu(cpu, tracing_cpumask) &&
				cpumask_test_cpu(cpu, tracing_cpumask_new)) {
2748
			atomic_dec(&global_trace.data[cpu]->disabled);
2749
			ring_buffer_record_enable_cpu(global_trace.buffer, cpu);
2750 2751
		}
	}
2752
	arch_spin_unlock(&ftrace_max_lock);
2753
	local_irq_enable();
2754

2755
	cpumask_copy(tracing_cpumask, tracing_cpumask_new);
2756 2757

	mutex_unlock(&tracing_cpumask_update_lock);
2758
	free_cpumask_var(tracing_cpumask_new);
I
Ingo Molnar 已提交
2759 2760

	return count;
2761 2762

err_unlock:
2763
	free_cpumask_var(tracing_cpumask_new);
2764 2765

	return err;
I
Ingo Molnar 已提交
2766 2767
}

2768
static const struct file_operations tracing_cpumask_fops = {
I
Ingo Molnar 已提交
2769 2770 2771
	.open		= tracing_open_generic,
	.read		= tracing_cpumask_read,
	.write		= tracing_cpumask_write,
2772
	.llseek		= generic_file_llseek,
2773 2774
};

L
Li Zefan 已提交
2775
static int tracing_trace_options_show(struct seq_file *m, void *v)
2776
{
2777 2778 2779
	struct tracer_opt *trace_opts;
	u32 tracer_flags;
	int i;
2780

2781 2782 2783 2784
	mutex_lock(&trace_types_lock);
	tracer_flags = current_trace->flags->val;
	trace_opts = current_trace->flags->opts;

2785 2786
	for (i = 0; trace_options[i]; i++) {
		if (trace_flags & (1 << i))
L
Li Zefan 已提交
2787
			seq_printf(m, "%s\n", trace_options[i]);
2788
		else
L
Li Zefan 已提交
2789
			seq_printf(m, "no%s\n", trace_options[i]);
2790 2791
	}

2792 2793
	for (i = 0; trace_opts[i].name; i++) {
		if (tracer_flags & trace_opts[i].bit)
L
Li Zefan 已提交
2794
			seq_printf(m, "%s\n", trace_opts[i].name);
2795
		else
L
Li Zefan 已提交
2796
			seq_printf(m, "no%s\n", trace_opts[i].name);
2797
	}
2798
	mutex_unlock(&trace_types_lock);
2799

L
Li Zefan 已提交
2800
	return 0;
2801 2802
}

L
Li Zefan 已提交
2803 2804 2805 2806 2807
static int __set_tracer_option(struct tracer *trace,
			       struct tracer_flags *tracer_flags,
			       struct tracer_opt *opts, int neg)
{
	int ret;
2808

L
Li Zefan 已提交
2809 2810 2811 2812 2813 2814 2815 2816 2817
	ret = trace->set_flag(tracer_flags->val, opts->bit, !neg);
	if (ret)
		return ret;

	if (neg)
		tracer_flags->val &= ~opts->bit;
	else
		tracer_flags->val |= opts->bit;
	return 0;
2818 2819
}

2820 2821 2822
/* Try to assign a tracer specific option */
static int set_tracer_option(struct tracer *trace, char *cmp, int neg)
{
2823
	struct tracer_flags *tracer_flags = trace->flags;
2824
	struct tracer_opt *opts = NULL;
L
Li Zefan 已提交
2825
	int i;
2826

2827 2828
	for (i = 0; tracer_flags->opts[i].name; i++) {
		opts = &tracer_flags->opts[i];
2829

L
Li Zefan 已提交
2830 2831 2832
		if (strcmp(cmp, opts->name) == 0)
			return __set_tracer_option(trace, trace->flags,
						   opts, neg);
2833 2834
	}

L
Li Zefan 已提交
2835
	return -EINVAL;
2836 2837
}

2838 2839 2840 2841 2842 2843 2844 2845 2846 2847
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;
2848 2849 2850

	if (mask == TRACE_ITER_RECORD_CMD)
		trace_event_enable_cmd_record(enabled);
2851 2852 2853

	if (mask == TRACE_ITER_OVERWRITE)
		ring_buffer_change_overwrite(global_trace.buffer, enabled);
2854 2855 2856

	if (mask == TRACE_ITER_PRINTK)
		trace_printk_start_stop_comm(enabled);
2857 2858
}

2859
static int trace_set_options(char *option)
2860
{
L
Li Zefan 已提交
2861
	char *cmp;
2862
	int neg = 0;
2863
	int ret = 0;
2864 2865
	int i;

2866
	cmp = strstrip(option);
2867

L
Li Zefan 已提交
2868
	if (strncmp(cmp, "no", 2) == 0) {
2869 2870 2871 2872 2873
		neg = 1;
		cmp += 2;
	}

	for (i = 0; trace_options[i]; i++) {
L
Li Zefan 已提交
2874
		if (strcmp(cmp, trace_options[i]) == 0) {
2875
			set_tracer_flags(1 << i, !neg);
2876 2877 2878
			break;
		}
	}
2879 2880 2881

	/* If no option could be set, test the specific tracer options */
	if (!trace_options[i]) {
2882
		mutex_lock(&trace_types_lock);
2883
		ret = set_tracer_option(current_trace, cmp, neg);
2884
		mutex_unlock(&trace_types_lock);
2885
	}
2886

2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901
	return ret;
}

static ssize_t
tracing_trace_options_write(struct file *filp, const char __user *ubuf,
			size_t cnt, loff_t *ppos)
{
	char buf[64];

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

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

2902 2903
	buf[cnt] = 0;

2904 2905
	trace_set_options(buf);

2906
	*ppos += cnt;
2907 2908 2909 2910

	return cnt;
}

L
Li Zefan 已提交
2911 2912 2913 2914 2915 2916 2917
static int tracing_trace_options_open(struct inode *inode, struct file *file)
{
	if (tracing_disabled)
		return -ENODEV;
	return single_open(file, tracing_trace_options_show, NULL);
}

2918
static const struct file_operations tracing_iter_fops = {
L
Li Zefan 已提交
2919 2920 2921 2922
	.open		= tracing_trace_options_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= single_release,
2923
	.write		= tracing_trace_options_write,
2924 2925
};

I
Ingo Molnar 已提交
2926 2927
static const char readme_msg[] =
	"tracing mini-HOWTO:\n\n"
2928 2929
	"# mount -t debugfs nodev /sys/kernel/debug\n\n"
	"# cat /sys/kernel/debug/tracing/available_tracers\n"
2930
	"wakeup wakeup_rt preemptirqsoff preemptoff irqsoff function nop\n\n"
2931
	"# cat /sys/kernel/debug/tracing/current_tracer\n"
2932
	"nop\n"
2933
	"# echo wakeup > /sys/kernel/debug/tracing/current_tracer\n"
2934
	"# cat /sys/kernel/debug/tracing/current_tracer\n"
2935
	"wakeup\n"
2936
	"# cat /sys/kernel/debug/tracing/trace_options\n"
I
Ingo Molnar 已提交
2937
	"noprint-parent nosym-offset nosym-addr noverbose\n"
2938
	"# echo print-parent > /sys/kernel/debug/tracing/trace_options\n"
2939
	"# echo 1 > /sys/kernel/debug/tracing/tracing_on\n"
2940
	"# cat /sys/kernel/debug/tracing/trace > /tmp/trace.txt\n"
2941
	"# echo 0 > /sys/kernel/debug/tracing/tracing_on\n"
I
Ingo Molnar 已提交
2942 2943 2944 2945 2946 2947 2948 2949 2950 2951
;

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));
}

2952
static const struct file_operations tracing_readme_fops = {
I
Ingo Molnar 已提交
2953 2954
	.open		= tracing_open_generic,
	.read		= tracing_readme_read,
2955
	.llseek		= generic_file_llseek,
I
Ingo Molnar 已提交
2956 2957
};

2958 2959 2960 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 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005
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,
3006
    .llseek	= generic_file_llseek,
3007 3008
};

3009 3010 3011 3012
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
3013
	char buf[MAX_TRACER_SIZE+2];
3014 3015 3016 3017 3018 3019 3020 3021 3022
	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
3023
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
3024 3025
}

3026 3027 3028 3029 3030 3031
int tracer_init(struct tracer *t, struct trace_array *tr)
{
	tracing_reset_online_cpus(tr);
	return t->init(tr);
}

3032 3033 3034 3035 3036 3037 3038
static void set_buffer_entries(struct trace_array *tr, unsigned long val)
{
	int cpu;
	for_each_tracing_cpu(cpu)
		tr->data[cpu]->entries = val;
}

3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063
/* resize @tr's buffer to the size of @size_tr's entries */
static int resize_buffer_duplicate_size(struct trace_array *tr,
					struct trace_array *size_tr, int cpu_id)
{
	int cpu, ret = 0;

	if (cpu_id == RING_BUFFER_ALL_CPUS) {
		for_each_tracing_cpu(cpu) {
			ret = ring_buffer_resize(tr->buffer,
					size_tr->data[cpu]->entries, cpu);
			if (ret < 0)
				break;
			tr->data[cpu]->entries = size_tr->data[cpu]->entries;
		}
	} else {
		ret = ring_buffer_resize(tr->buffer,
					size_tr->data[cpu_id]->entries, cpu_id);
		if (ret == 0)
			tr->data[cpu_id]->entries =
				size_tr->data[cpu_id]->entries;
	}

	return ret;
}

3064
static int __tracing_resize_ring_buffer(unsigned long size, int cpu)
3065 3066 3067 3068 3069
{
	int ret;

	/*
	 * If kernel or user changes the size of the ring buffer
3070 3071
	 * we use the size that was given, and we can forget about
	 * expanding it later.
3072 3073 3074
	 */
	ring_buffer_expanded = 1;

3075 3076 3077 3078
	/* May be called before buffers are initialized */
	if (!global_trace.buffer)
		return 0;

3079
	ret = ring_buffer_resize(global_trace.buffer, size, cpu);
3080 3081 3082
	if (ret < 0)
		return ret;

3083 3084 3085
	if (!current_trace->use_max_tr)
		goto out;

3086
	ret = ring_buffer_resize(max_tr.buffer, size, cpu);
3087
	if (ret < 0) {
3088 3089
		int r = resize_buffer_duplicate_size(&global_trace,
						     &global_trace, cpu);
3090
		if (r < 0) {
3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104
			/*
			 * 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.
			 */
3105 3106 3107 3108 3109 3110
			WARN_ON(1);
			tracing_disabled = 1;
		}
		return ret;
	}

3111 3112 3113 3114 3115
	if (cpu == RING_BUFFER_ALL_CPUS)
		set_buffer_entries(&max_tr, size);
	else
		max_tr.data[cpu]->entries = size;

3116
 out:
3117 3118 3119 3120
	if (cpu == RING_BUFFER_ALL_CPUS)
		set_buffer_entries(&global_trace, size);
	else
		global_trace.data[cpu]->entries = size;
3121 3122 3123 3124

	return ret;
}

3125
static ssize_t tracing_resize_ring_buffer(unsigned long size, int cpu_id)
3126
{
3127
	int ret = size;
3128 3129 3130

	mutex_lock(&trace_types_lock);

3131 3132 3133 3134 3135 3136 3137
	if (cpu_id != RING_BUFFER_ALL_CPUS) {
		/* make sure, this cpu is enabled in the mask */
		if (!cpumask_test_cpu(cpu_id, tracing_buffer_mask)) {
			ret = -EINVAL;
			goto out;
		}
	}
3138

3139
	ret = __tracing_resize_ring_buffer(size, cpu_id);
3140 3141 3142
	if (ret < 0)
		ret = -ENOMEM;

3143
out:
3144 3145 3146 3147 3148
	mutex_unlock(&trace_types_lock);

	return ret;
}

3149

3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163
/**
 * 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;

3164
	mutex_lock(&trace_types_lock);
3165
	if (!ring_buffer_expanded)
3166 3167
		ret = __tracing_resize_ring_buffer(trace_buf_size,
						RING_BUFFER_ALL_CPUS);
3168
	mutex_unlock(&trace_types_lock);
3169 3170 3171 3172

	return ret;
}

3173 3174 3175 3176 3177 3178 3179 3180
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);

3181
static int tracing_set_tracer(const char *buf)
3182
{
3183
	static struct trace_option_dentry *topts;
3184 3185
	struct trace_array *tr = &global_trace;
	struct tracer *t;
3186
	int ret = 0;
3187

3188 3189
	mutex_lock(&trace_types_lock);

3190
	if (!ring_buffer_expanded) {
3191 3192
		ret = __tracing_resize_ring_buffer(trace_buf_size,
						RING_BUFFER_ALL_CPUS);
3193
		if (ret < 0)
3194
			goto out;
3195 3196 3197
		ret = 0;
	}

3198 3199 3200 3201
	for (t = trace_types; t; t = t->next) {
		if (strcmp(t->name, buf) == 0)
			break;
	}
3202 3203 3204 3205 3206
	if (!t) {
		ret = -EINVAL;
		goto out;
	}
	if (t == current_trace)
3207 3208
		goto out;

3209
	trace_branch_disable();
3210 3211
	if (current_trace && current_trace->reset)
		current_trace->reset(tr);
3212 3213 3214 3215 3216 3217
	if (current_trace && current_trace->use_max_tr) {
		/*
		 * We don't free the ring buffer. instead, resize it because
		 * The max_tr ring buffer has some state (e.g. ring->clock) and
		 * we want preserve it.
		 */
3218 3219
		ring_buffer_resize(max_tr.buffer, 1, RING_BUFFER_ALL_CPUS);
		set_buffer_entries(&max_tr, 1);
3220
	}
3221 3222
	destroy_trace_option_files(topts);

3223
	current_trace = &nop_trace;
3224

3225 3226
	topts = create_trace_option_files(t);
	if (t->use_max_tr) {
3227
		/* we need to make per cpu buffer sizes equivalent */
3228 3229 3230 3231
		ret = resize_buffer_duplicate_size(&max_tr, &global_trace,
						   RING_BUFFER_ALL_CPUS);
		if (ret < 0)
			goto out;
3232
	}
3233

3234
	if (t->init) {
3235
		ret = tracer_init(t, tr);
3236 3237 3238
		if (ret)
			goto out;
	}
3239

3240
	current_trace = t;
3241
	trace_branch_enable(tr);
3242 3243 3244
 out:
	mutex_unlock(&trace_types_lock);

3245 3246 3247 3248 3249 3250 3251
	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
3252
	char buf[MAX_TRACER_SIZE+1];
3253 3254
	int i;
	size_t ret;
3255 3256 3257
	int err;

	ret = cnt;
3258

L
Li Zefan 已提交
3259 3260
	if (cnt > MAX_TRACER_SIZE)
		cnt = MAX_TRACER_SIZE;
3261 3262 3263 3264 3265 3266 3267 3268 3269 3270

	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;

3271 3272 3273
	err = tracing_set_tracer(buf);
	if (err)
		return err;
3274

3275
	*ppos += ret;
3276

3277
	return ret;
3278 3279 3280 3281 3282 3283 3284 3285 3286 3287
}

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
3288
	r = snprintf(buf, sizeof(buf), "%ld\n",
3289
		     *ptr == (unsigned long)-1 ? -1 : nsecs_to_usecs(*ptr));
S
Steven Rostedt 已提交
3290 3291
	if (r > sizeof(buf))
		r = sizeof(buf);
I
Ingo Molnar 已提交
3292
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
3293 3294 3295 3296 3297 3298
}

static ssize_t
tracing_max_lat_write(struct file *filp, const char __user *ubuf,
		      size_t cnt, loff_t *ppos)
{
3299 3300
	unsigned long *ptr = filp->private_data;
	unsigned long val;
3301
	int ret;
3302

3303 3304
	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
3305
		return ret;
3306 3307 3308 3309 3310 3311

	*ptr = val * 1000;

	return cnt;
}

3312 3313
static int tracing_open_pipe(struct inode *inode, struct file *filp)
{
3314
	long cpu_file = (long) inode->i_private;
3315
	struct trace_iterator *iter;
3316
	int ret = 0;
3317 3318 3319 3320

	if (tracing_disabled)
		return -ENODEV;

3321 3322
	mutex_lock(&trace_types_lock);

3323 3324
	/* create a buffer to store the information to pass to userspace */
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
3325 3326 3327 3328
	if (!iter) {
		ret = -ENOMEM;
		goto out;
	}
3329

3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341
	/*
	 * 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;

3342
	if (!alloc_cpumask_var(&iter->started, GFP_KERNEL)) {
3343
		ret = -ENOMEM;
3344
		goto fail;
3345 3346
	}

3347
	/* trace pipe does not show start of buffer */
3348
	cpumask_setall(iter->started);
3349

3350 3351 3352
	if (trace_flags & TRACE_ITER_LATENCY_FMT)
		iter->iter_flags |= TRACE_FILE_LAT_FMT;

3353 3354 3355 3356
	/* Output in nanoseconds only if we are using a clock in nanoseconds. */
	if (trace_clocks[trace_clock_id].in_ns)
		iter->iter_flags |= TRACE_FILE_TIME_IN_NS;

3357
	iter->cpu_file = cpu_file;
3358
	iter->tr = &global_trace;
3359
	mutex_init(&iter->mutex);
3360 3361
	filp->private_data = iter;

3362 3363 3364
	if (iter->trace->pipe_open)
		iter->trace->pipe_open(iter);

3365
	nonseekable_open(inode, filp);
3366 3367 3368
out:
	mutex_unlock(&trace_types_lock);
	return ret;
3369 3370 3371 3372 3373 3374

fail:
	kfree(iter->trace);
	kfree(iter);
	mutex_unlock(&trace_types_lock);
	return ret;
3375 3376 3377 3378 3379 3380
}

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

3381 3382
	mutex_lock(&trace_types_lock);

3383
	if (iter->trace->pipe_close)
S
Steven Rostedt 已提交
3384 3385
		iter->trace->pipe_close(iter);

3386 3387
	mutex_unlock(&trace_types_lock);

3388
	free_cpumask_var(iter->started);
3389 3390
	mutex_destroy(&iter->mutex);
	kfree(iter->trace);
3391 3392 3393 3394 3395
	kfree(iter);

	return 0;
}

3396 3397 3398 3399 3400 3401 3402 3403 3404 3405
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
3406
	} else {
3407 3408 3409 3410 3411 3412 3413 3414 3415 3416
		if (!trace_empty(iter))
			return POLLIN | POLLRDNORM;
		poll_wait(filp, &trace_wait, poll_table);
		if (!trace_empty(iter))
			return POLLIN | POLLRDNORM;

		return 0;
	}
}

3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435
/*
 * 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);
}

3436 3437
/* Must be called with trace_types_lock mutex held. */
static int tracing_wait_pipe(struct file *filp)
3438 3439 3440 3441
{
	struct trace_iterator *iter = filp->private_data;

	while (trace_empty(iter)) {
3442

3443
		if ((filp->f_flags & O_NONBLOCK)) {
3444
			return -EAGAIN;
3445
		}
3446

3447
		mutex_unlock(&iter->mutex);
3448

3449
		iter->trace->wait_pipe(iter);
3450

3451
		mutex_lock(&iter->mutex);
3452

3453
		if (signal_pending(current))
3454
			return -EINTR;
3455 3456

		/*
L
Liu Bo 已提交
3457
		 * We block until we read something and tracing is disabled.
3458 3459 3460 3461 3462 3463 3464
		 * 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.
		 */
L
Liu Bo 已提交
3465
		if (!tracing_is_enabled() && iter->pos)
3466 3467 3468
			break;
	}

3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479
	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;
3480
	static struct tracer *old_tracer;
3481 3482 3483 3484 3485 3486 3487
	ssize_t sret;

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

3488
	trace_seq_init(&iter->seq);
3489

3490
	/* copy the tracer to avoid using a global lock all around */
3491
	mutex_lock(&trace_types_lock);
3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503
	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);
3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514
	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;

3515
	/* stop when tracing is finished */
3516 3517
	if (trace_empty(iter)) {
		sret = 0;
3518
		goto out;
3519
	}
3520 3521 3522 3523

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

3524 3525 3526 3527
	/* reset all but tr, trace, and overruns */
	memset(&iter->seq, 0,
	       sizeof(struct trace_iterator) -
	       offsetof(struct trace_iterator, seq));
3528
	iter->pos = -1;
3529

3530
	trace_event_read_lock();
3531
	trace_access_lock(iter->cpu_file);
3532
	while (trace_find_next_entry_inc(iter) != NULL) {
3533
		enum print_line_t ret;
S
Steven Rostedt 已提交
3534 3535
		int len = iter->seq.len;

I
Ingo Molnar 已提交
3536
		ret = print_trace_line(iter);
3537
		if (ret == TRACE_TYPE_PARTIAL_LINE) {
S
Steven Rostedt 已提交
3538 3539
			/* don't print partial lines */
			iter->seq.len = len;
3540
			break;
S
Steven Rostedt 已提交
3541
		}
3542 3543
		if (ret != TRACE_TYPE_NO_CONSUME)
			trace_consume(iter);
3544 3545 3546

		if (iter->seq.len >= cnt)
			break;
3547 3548 3549 3550 3551 3552 3553 3554

		/*
		 * Setting the full flag means we reached the trace_seq buffer
		 * size and we should leave by partial output condition above.
		 * One of the trace_seq_* functions is not used properly.
		 */
		WARN_ONCE(iter->seq.full, "full flag set for trace type %d",
			  iter->ent->type);
3555
	}
3556
	trace_access_unlock(iter->cpu_file);
3557
	trace_event_read_unlock();
3558 3559

	/* Now copy what we have to the user */
3560 3561
	sret = trace_seq_to_user(&iter->seq, ubuf, cnt);
	if (iter->seq.readpos >= iter->seq.len)
3562
		trace_seq_init(&iter->seq);
P
Pekka Paalanen 已提交
3563 3564

	/*
L
Lucas De Marchi 已提交
3565
	 * If there was nothing to send to user, in spite of consuming trace
P
Pekka Paalanen 已提交
3566 3567
	 * entries, go back to wait for more entries.
	 */
3568
	if (sret == -EBUSY)
P
Pekka Paalanen 已提交
3569
		goto waitagain;
3570

3571
out:
3572
	mutex_unlock(&iter->mutex);
3573

3574
	return sret;
3575 3576
}

3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588
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]);
}

3589
static const struct pipe_buf_operations tracing_pipe_buf_ops = {
S
Steven Rostedt 已提交
3590 3591 3592 3593 3594 3595 3596
	.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,
3597 3598
};

S
Steven Rostedt 已提交
3599
static size_t
3600
tracing_fill_pipe_page(size_t rem, struct trace_iterator *iter)
S
Steven Rostedt 已提交
3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619
{
	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;
		}

3620 3621
		if (ret != TRACE_TYPE_NO_CONSUME)
			trace_consume(iter);
S
Steven Rostedt 已提交
3622
		rem -= count;
3623
		if (!trace_find_next_entry_inc(iter))	{
S
Steven Rostedt 已提交
3624 3625 3626 3627 3628 3629 3630 3631 3632
			rem = 0;
			iter->ent = NULL;
			break;
		}
	}

	return rem;
}

3633 3634 3635 3636 3637 3638
static ssize_t tracing_splice_read_pipe(struct file *filp,
					loff_t *ppos,
					struct pipe_inode_info *pipe,
					size_t len,
					unsigned int flags)
{
3639 3640
	struct page *pages_def[PIPE_DEF_BUFFERS];
	struct partial_page partial_def[PIPE_DEF_BUFFERS];
3641 3642
	struct trace_iterator *iter = filp->private_data;
	struct splice_pipe_desc spd = {
3643 3644
		.pages		= pages_def,
		.partial	= partial_def,
S
Steven Rostedt 已提交
3645
		.nr_pages	= 0, /* This gets updated below. */
3646
		.nr_pages_max	= PIPE_DEF_BUFFERS,
S
Steven Rostedt 已提交
3647 3648 3649
		.flags		= flags,
		.ops		= &tracing_pipe_buf_ops,
		.spd_release	= tracing_spd_release_pipe,
3650
	};
3651
	static struct tracer *old_tracer;
3652
	ssize_t ret;
S
Steven Rostedt 已提交
3653
	size_t rem;
3654 3655
	unsigned int i;

3656 3657 3658
	if (splice_grow_spd(pipe, &spd))
		return -ENOMEM;

3659
	/* copy the tracer to avoid using a global lock all around */
3660
	mutex_lock(&trace_types_lock);
3661 3662 3663 3664 3665 3666 3667
	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);
3668 3669 3670 3671 3672

	if (iter->trace->splice_read) {
		ret = iter->trace->splice_read(iter, filp,
					       ppos, pipe, len, flags);
		if (ret)
S
Steven Rostedt 已提交
3673
			goto out_err;
3674 3675 3676 3677
	}

	ret = tracing_wait_pipe(filp);
	if (ret <= 0)
S
Steven Rostedt 已提交
3678
		goto out_err;
3679

3680
	if (!iter->ent && !trace_find_next_entry_inc(iter)) {
3681
		ret = -EFAULT;
S
Steven Rostedt 已提交
3682
		goto out_err;
3683 3684
	}

3685
	trace_event_read_lock();
3686
	trace_access_lock(iter->cpu_file);
3687

3688
	/* Fill as many pages as possible. */
3689 3690 3691
	for (i = 0, rem = len; i < pipe->buffers && rem; i++) {
		spd.pages[i] = alloc_page(GFP_KERNEL);
		if (!spd.pages[i])
S
Steven Rostedt 已提交
3692
			break;
3693

3694
		rem = tracing_fill_pipe_page(rem, iter);
3695 3696 3697

		/* Copy the data into the page, so we can start over. */
		ret = trace_seq_to_buffer(&iter->seq,
3698
					  page_address(spd.pages[i]),
3699 3700
					  iter->seq.len);
		if (ret < 0) {
3701
			__free_page(spd.pages[i]);
3702 3703
			break;
		}
3704 3705
		spd.partial[i].offset = 0;
		spd.partial[i].len = iter->seq.len;
3706

3707
		trace_seq_init(&iter->seq);
3708 3709
	}

3710
	trace_access_unlock(iter->cpu_file);
3711
	trace_event_read_unlock();
3712
	mutex_unlock(&iter->mutex);
3713 3714 3715

	spd.nr_pages = i;

3716 3717
	ret = splice_to_pipe(pipe, &spd);
out:
3718
	splice_shrink_spd(&spd);
3719
	return ret;
3720

S
Steven Rostedt 已提交
3721
out_err:
3722
	mutex_unlock(&iter->mutex);
3723
	goto out;
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
struct ftrace_entries_info {
	struct trace_array	*tr;
	int			cpu;
};

static int tracing_entries_open(struct inode *inode, struct file *filp)
{
	struct ftrace_entries_info *info;

	if (tracing_disabled)
		return -ENODEV;

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

	info->tr = &global_trace;
	info->cpu = (unsigned long)inode->i_private;

	filp->private_data = info;

	return 0;
}

3750 3751 3752 3753
static ssize_t
tracing_entries_read(struct file *filp, char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
3754 3755 3756 3757 3758
	struct ftrace_entries_info *info = filp->private_data;
	struct trace_array *tr = info->tr;
	char buf[64];
	int r = 0;
	ssize_t ret;
3759

3760
	mutex_lock(&trace_types_lock);
3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790

	if (info->cpu == RING_BUFFER_ALL_CPUS) {
		int cpu, buf_size_same;
		unsigned long size;

		size = 0;
		buf_size_same = 1;
		/* check if all cpu sizes are same */
		for_each_tracing_cpu(cpu) {
			/* fill in the size from first enabled cpu */
			if (size == 0)
				size = tr->data[cpu]->entries;
			if (size != tr->data[cpu]->entries) {
				buf_size_same = 0;
				break;
			}
		}

		if (buf_size_same) {
			if (!ring_buffer_expanded)
				r = sprintf(buf, "%lu (expanded: %lu)\n",
					    size >> 10,
					    trace_buf_size >> 10);
			else
				r = sprintf(buf, "%lu\n", size >> 10);
		} else
			r = sprintf(buf, "X\n");
	} else
		r = sprintf(buf, "%lu\n", tr->data[info->cpu]->entries >> 10);

3791 3792
	mutex_unlock(&trace_types_lock);

3793 3794
	ret = simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
	return ret;
3795 3796 3797 3798 3799 3800
}

static ssize_t
tracing_entries_write(struct file *filp, const char __user *ubuf,
		      size_t cnt, loff_t *ppos)
{
3801
	struct ftrace_entries_info *info = filp->private_data;
3802
	unsigned long val;
3803
	int ret;
3804

3805 3806
	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
3807
		return ret;
3808 3809 3810 3811 3812

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

3813 3814 3815
	/* value is in KB */
	val <<= 10;

3816
	ret = tracing_resize_ring_buffer(val, info->cpu);
3817 3818
	if (ret < 0)
		return ret;
3819

3820
	*ppos += cnt;
3821

3822 3823
	return cnt;
}
S
Steven Rostedt 已提交
3824

3825 3826 3827 3828 3829 3830 3831 3832 3833 3834
static int
tracing_entries_release(struct inode *inode, struct file *filp)
{
	struct ftrace_entries_info *info = filp->private_data;

	kfree(info);

	return 0;
}

3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845
static ssize_t
tracing_total_entries_read(struct file *filp, char __user *ubuf,
				size_t cnt, loff_t *ppos)
{
	struct trace_array *tr = filp->private_data;
	char buf[64];
	int r, cpu;
	unsigned long size = 0, expanded_size = 0;

	mutex_lock(&trace_types_lock);
	for_each_tracing_cpu(cpu) {
3846
		size += tr->data[cpu]->entries >> 10;
3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858
		if (!ring_buffer_expanded)
			expanded_size += trace_buf_size >> 10;
	}
	if (ring_buffer_expanded)
		r = sprintf(buf, "%lu\n", size);
	else
		r = sprintf(buf, "%lu (expanded: %lu)\n", size, expanded_size);
	mutex_unlock(&trace_types_lock);

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

3859 3860 3861 3862 3863 3864 3865 3866 3867 3868
static ssize_t
tracing_free_buffer_write(struct file *filp, const char __user *ubuf,
			  size_t cnt, loff_t *ppos)
{
	/*
	 * There is no need to read what the user has written, this function
	 * is just to make sure that there is no error when "echo" is used
	 */

	*ppos += cnt;
3869 3870 3871 3872

	return cnt;
}

3873 3874 3875
static int
tracing_free_buffer_release(struct inode *inode, struct file *filp)
{
3876 3877 3878
	/* disable tracing ? */
	if (trace_flags & TRACE_ITER_STOP_ON_FREE)
		tracing_off();
3879
	/* resize the ring buffer to 0 */
3880
	tracing_resize_ring_buffer(0, RING_BUFFER_ALL_CPUS);
3881 3882 3883 3884

	return 0;
}

3885 3886 3887 3888
static ssize_t
tracing_mark_write(struct file *filp, const char __user *ubuf,
					size_t cnt, loff_t *fpos)
{
3889 3890 3891 3892 3893 3894
	unsigned long addr = (unsigned long)ubuf;
	struct ring_buffer_event *event;
	struct ring_buffer *buffer;
	struct print_entry *entry;
	unsigned long irq_flags;
	struct page *pages[2];
3895
	void *map_page[2];
3896 3897 3898 3899 3900 3901
	int nr_pages = 1;
	ssize_t written;
	int offset;
	int size;
	int len;
	int ret;
3902
	int i;
3903

S
Steven Rostedt 已提交
3904
	if (tracing_disabled)
3905 3906
		return -EINVAL;

3907 3908 3909
	if (!(trace_flags & TRACE_ITER_MARKERS))
		return -EINVAL;

3910 3911 3912
	if (cnt > TRACE_BUF_SIZE)
		cnt = TRACE_BUF_SIZE;

3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927
	/*
	 * Userspace is injecting traces into the kernel trace buffer.
	 * We want to be as non intrusive as possible.
	 * To do so, we do not want to allocate any special buffers
	 * or take any locks, but instead write the userspace data
	 * straight into the ring buffer.
	 *
	 * First we need to pin the userspace buffer into memory,
	 * which, most likely it is, because it just referenced it.
	 * But there's no guarantee that it is. By using get_user_pages_fast()
	 * and kmap_atomic/kunmap_atomic() we can get access to the
	 * pages directly. We then write the data directly into the
	 * ring buffer.
	 */
	BUILD_BUG_ON(TRACE_BUF_SIZE >= PAGE_SIZE);
3928

3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941
	/* check if we cross pages */
	if ((addr & PAGE_MASK) != ((addr + cnt) & PAGE_MASK))
		nr_pages = 2;

	offset = addr & (PAGE_SIZE - 1);
	addr &= PAGE_MASK;

	ret = get_user_pages_fast(addr, nr_pages, 0, pages);
	if (ret < nr_pages) {
		while (--ret >= 0)
			put_page(pages[ret]);
		written = -EFAULT;
		goto out;
3942
	}
3943

3944 3945
	for (i = 0; i < nr_pages; i++)
		map_page[i] = kmap_atomic(pages[i]);
3946 3947 3948 3949 3950 3951 3952 3953 3954 3955

	local_save_flags(irq_flags);
	size = sizeof(*entry) + cnt + 2; /* possible \n added */
	buffer = global_trace.buffer;
	event = trace_buffer_lock_reserve(buffer, TRACE_PRINT, size,
					  irq_flags, preempt_count());
	if (!event) {
		/* Ring buffer disabled, return as if not open for write */
		written = -EBADF;
		goto out_unlock;
3956
	}
3957 3958 3959 3960 3961 3962

	entry = ring_buffer_event_data(event);
	entry->ip = _THIS_IP_;

	if (nr_pages == 2) {
		len = PAGE_SIZE - offset;
3963 3964
		memcpy(&entry->buf, map_page[0] + offset, len);
		memcpy(&entry->buf[len], map_page[1], cnt - len);
C
Carsten Emde 已提交
3965
	} else
3966
		memcpy(&entry->buf, map_page[0] + offset, cnt);
3967

3968 3969 3970 3971 3972 3973
	if (entry->buf[cnt - 1] != '\n') {
		entry->buf[cnt] = '\n';
		entry->buf[cnt + 1] = '\0';
	} else
		entry->buf[cnt] = '\0';

3974
	__buffer_unlock_commit(buffer, event);
3975

3976
	written = cnt;
3977

3978
	*fpos += written;
3979

3980
 out_unlock:
3981 3982 3983 3984
	for (i = 0; i < nr_pages; i++){
		kunmap_atomic(map_page[i]);
		put_page(pages[i]);
	}
3985
 out:
3986
	return written;
3987 3988
}

L
Li Zefan 已提交
3989
static int tracing_clock_show(struct seq_file *m, void *v)
3990 3991 3992 3993
{
	int i;

	for (i = 0; i < ARRAY_SIZE(trace_clocks); i++)
L
Li Zefan 已提交
3994
		seq_printf(m,
3995 3996 3997
			"%s%s%s%s", i ? " " : "",
			i == trace_clock_id ? "[" : "", trace_clocks[i].name,
			i == trace_clock_id ? "]" : "");
L
Li Zefan 已提交
3998
	seq_putc(m, '\n');
3999

L
Li Zefan 已提交
4000
	return 0;
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 4034
}

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);

4035 4036 4037 4038 4039 4040 4041 4042
	/*
	 * New clock may not be consistent with the previous clock.
	 * Reset the buffer so that it doesn't have incomparable timestamps.
	 */
	tracing_reset_online_cpus(&global_trace);
	if (max_tr.buffer)
		tracing_reset_online_cpus(&max_tr);

4043 4044 4045 4046 4047 4048 4049
	mutex_unlock(&trace_types_lock);

	*fpos += cnt;

	return cnt;
}

L
Li Zefan 已提交
4050 4051 4052 4053 4054 4055 4056
static int tracing_clock_open(struct inode *inode, struct file *file)
{
	if (tracing_disabled)
		return -ENODEV;
	return single_open(file, tracing_clock_show, NULL);
}

4057
static const struct file_operations tracing_max_lat_fops = {
I
Ingo Molnar 已提交
4058 4059 4060
	.open		= tracing_open_generic,
	.read		= tracing_max_lat_read,
	.write		= tracing_max_lat_write,
4061
	.llseek		= generic_file_llseek,
4062 4063
};

4064
static const struct file_operations set_tracer_fops = {
I
Ingo Molnar 已提交
4065 4066 4067
	.open		= tracing_open_generic,
	.read		= tracing_set_trace_read,
	.write		= tracing_set_trace_write,
4068
	.llseek		= generic_file_llseek,
4069 4070
};

4071
static const struct file_operations tracing_pipe_fops = {
I
Ingo Molnar 已提交
4072
	.open		= tracing_open_pipe,
4073
	.poll		= tracing_poll_pipe,
I
Ingo Molnar 已提交
4074
	.read		= tracing_read_pipe,
4075
	.splice_read	= tracing_splice_read_pipe,
I
Ingo Molnar 已提交
4076
	.release	= tracing_release_pipe,
4077
	.llseek		= no_llseek,
4078 4079
};

4080
static const struct file_operations tracing_entries_fops = {
4081
	.open		= tracing_entries_open,
4082 4083
	.read		= tracing_entries_read,
	.write		= tracing_entries_write,
4084
	.release	= tracing_entries_release,
4085
	.llseek		= generic_file_llseek,
4086 4087
};

4088 4089 4090 4091 4092 4093
static const struct file_operations tracing_total_entries_fops = {
	.open		= tracing_open_generic,
	.read		= tracing_total_entries_read,
	.llseek		= generic_file_llseek,
};

4094 4095 4096 4097 4098
static const struct file_operations tracing_free_buffer_fops = {
	.write		= tracing_free_buffer_write,
	.release	= tracing_free_buffer_release,
};

4099
static const struct file_operations tracing_mark_fops = {
4100
	.open		= tracing_open_generic,
4101
	.write		= tracing_mark_write,
4102
	.llseek		= generic_file_llseek,
4103 4104
};

4105
static const struct file_operations trace_clock_fops = {
L
Li Zefan 已提交
4106 4107 4108 4109
	.open		= tracing_clock_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= single_release,
4110 4111 4112
	.write		= tracing_clock_write,
};

4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133
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;
4134
	info->spare	= NULL;
4135 4136 4137 4138 4139
	/* Force reading ring buffer for first read */
	info->read	= (unsigned int)-1;

	filp->private_data = info;

4140
	return nonseekable_open(inode, filp);
4141 4142 4143 4144 4145 4146 4147 4148 4149 4150
}

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;
	ssize_t ret;
	size_t size;

4151 4152 4153
	if (!count)
		return 0;

4154
	if (!info->spare)
4155
		info->spare = ring_buffer_alloc_read_page(info->tr->buffer, info->cpu);
4156 4157 4158
	if (!info->spare)
		return -ENOMEM;

4159 4160 4161 4162
	/* Do we have previous read data to read? */
	if (info->read < PAGE_SIZE)
		goto read;

4163
	trace_access_lock(info->cpu);
4164 4165 4166 4167
	ret = ring_buffer_read_page(info->tr->buffer,
				    &info->spare,
				    count,
				    info->cpu, 0);
4168
	trace_access_unlock(info->cpu);
4169 4170 4171
	if (ret < 0)
		return 0;

4172 4173
	info->read = 0;

4174 4175 4176 4177 4178 4179
read:
	size = PAGE_SIZE - info->read;
	if (size > count)
		size = count;

	ret = copy_to_user(ubuf, info->spare + info->read, size);
4180
	if (ret == size)
4181
		return -EFAULT;
4182 4183
	size -= ret;

4184 4185 4186 4187 4188 4189 4190 4191 4192 4193
	*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;

4194 4195
	if (info->spare)
		ring_buffer_free_read_page(info->tr->buffer, info->spare);
4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225 4226 4227 4228
	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 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. */
4229
static const struct pipe_buf_operations buffer_pipe_buf_ops = {
4230 4231 4232 4233 4234
	.can_merge		= 0,
	.map			= generic_pipe_buf_map,
	.unmap			= generic_pipe_buf_unmap,
	.confirm		= generic_pipe_buf_confirm,
	.release		= buffer_pipe_buf_release,
4235
	.steal			= generic_pipe_buf_steal,
4236 4237 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 4253 4254 4255 4256 4257 4258 4259 4260 4261
	.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;
4262 4263
	struct partial_page partial_def[PIPE_DEF_BUFFERS];
	struct page *pages_def[PIPE_DEF_BUFFERS];
4264
	struct splice_pipe_desc spd = {
4265 4266
		.pages		= pages_def,
		.partial	= partial_def,
4267
		.nr_pages_max	= PIPE_DEF_BUFFERS,
4268 4269 4270 4271 4272
		.flags		= flags,
		.ops		= &buffer_pipe_buf_ops,
		.spd_release	= buffer_spd_release,
	};
	struct buffer_ref *ref;
4273
	int entries, size, i;
4274 4275
	size_t ret;

4276 4277 4278
	if (splice_grow_spd(pipe, &spd))
		return -ENOMEM;

4279
	if (*ppos & (PAGE_SIZE - 1)) {
4280 4281
		ret = -EINVAL;
		goto out;
4282 4283 4284
	}

	if (len & (PAGE_SIZE - 1)) {
4285 4286 4287 4288
		if (len < PAGE_SIZE) {
			ret = -EINVAL;
			goto out;
		}
4289 4290 4291
		len &= PAGE_MASK;
	}

4292
	trace_access_lock(info->cpu);
4293 4294
	entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu);

4295
	for (i = 0; i < pipe->buffers && len && entries; i++, len -= PAGE_SIZE) {
4296 4297 4298 4299 4300 4301 4302
		struct page *page;
		int r;

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

4303
		ref->ref = 1;
4304
		ref->buffer = info->tr->buffer;
4305
		ref->page = ring_buffer_alloc_read_page(ref->buffer, info->cpu);
4306 4307 4308 4309 4310 4311
		if (!ref->page) {
			kfree(ref);
			break;
		}

		r = ring_buffer_read_page(ref->buffer, &ref->page,
4312
					  len, info->cpu, 1);
4313
		if (r < 0) {
4314
			ring_buffer_free_read_page(ref->buffer, ref->page);
4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333
			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++;
4334
		*ppos += PAGE_SIZE;
4335 4336

		entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu);
4337 4338
	}

4339
	trace_access_unlock(info->cpu);
4340 4341 4342 4343 4344 4345 4346 4347 4348
	spd.nr_pages = i;

	/* did we read anything? */
	if (!spd.nr_pages) {
		if (flags & SPLICE_F_NONBLOCK)
			ret = -EAGAIN;
		else
			ret = 0;
		/* TODO: block */
4349
		goto out;
4350 4351 4352
	}

	ret = splice_to_pipe(pipe, &spd);
4353
	splice_shrink_spd(&spd);
4354
out:
4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365
	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,
};

4366 4367 4368 4369 4370 4371 4372 4373
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;
4374 4375
	unsigned long long t;
	unsigned long usec_rem;
4376

4377
	s = kmalloc(sizeof(*s), GFP_KERNEL);
4378
	if (!s)
4379
		return -ENOMEM;
4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391

	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);

4392 4393 4394
	cnt = ring_buffer_bytes_cpu(tr->buffer, cpu);
	trace_seq_printf(s, "bytes: %ld\n", cnt);

4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408
	if (trace_clocks[trace_clock_id].in_ns) {
		/* local or global for trace_clock */
		t = ns2usecs(ring_buffer_oldest_event_ts(tr->buffer, cpu));
		usec_rem = do_div(t, USEC_PER_SEC);
		trace_seq_printf(s, "oldest event ts: %5llu.%06lu\n",
								t, usec_rem);

		t = ns2usecs(ring_buffer_time_stamp(tr->buffer, cpu));
		usec_rem = do_div(t, USEC_PER_SEC);
		trace_seq_printf(s, "now ts: %5llu.%06lu\n", t, usec_rem);
	} else {
		/* counter or tsc mode for trace_clock */
		trace_seq_printf(s, "oldest event ts: %llu\n",
				ring_buffer_oldest_event_ts(tr->buffer, cpu));
4409

4410 4411 4412
		trace_seq_printf(s, "now ts: %llu\n",
				ring_buffer_time_stamp(tr->buffer, cpu));
	}
4413

4414 4415 4416
	cnt = ring_buffer_dropped_events_cpu(tr->buffer, cpu);
	trace_seq_printf(s, "dropped events: %ld\n", cnt);

4417 4418 4419 4420 4421 4422 4423 4424 4425 4426
	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,
4427
	.llseek		= generic_file_llseek,
4428 4429
};

4430 4431
#ifdef CONFIG_DYNAMIC_FTRACE

S
Steven Rostedt 已提交
4432 4433 4434 4435 4436
int __weak ftrace_arch_read_dyn_info(char *buf, int size)
{
	return 0;
}

4437
static ssize_t
S
Steven Rostedt 已提交
4438
tracing_read_dyn_info(struct file *filp, char __user *ubuf,
4439 4440
		  size_t cnt, loff_t *ppos)
{
S
Steven Rostedt 已提交
4441 4442
	static char ftrace_dyn_info_buffer[1024];
	static DEFINE_MUTEX(dyn_info_mutex);
4443
	unsigned long *p = filp->private_data;
S
Steven Rostedt 已提交
4444
	char *buf = ftrace_dyn_info_buffer;
S
Steven Rostedt 已提交
4445
	int size = ARRAY_SIZE(ftrace_dyn_info_buffer);
4446 4447
	int r;

S
Steven Rostedt 已提交
4448 4449
	mutex_lock(&dyn_info_mutex);
	r = sprintf(buf, "%ld ", *p);
I
Ingo Molnar 已提交
4450

S
Steven Rostedt 已提交
4451
	r += ftrace_arch_read_dyn_info(buf+r, (size-1)-r);
S
Steven Rostedt 已提交
4452 4453 4454 4455 4456 4457 4458
	buf[r++] = '\n';

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

	mutex_unlock(&dyn_info_mutex);

	return r;
4459 4460
}

4461
static const struct file_operations tracing_dyn_info_fops = {
I
Ingo Molnar 已提交
4462
	.open		= tracing_open_generic,
S
Steven Rostedt 已提交
4463
	.read		= tracing_read_dyn_info,
4464
	.llseek		= generic_file_llseek,
4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476
};
#endif

static struct dentry *d_tracer;

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

	if (d_tracer)
		return d_tracer;

4477 4478 4479
	if (!debugfs_initialized())
		return NULL;

4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490
	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;
}

4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519
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();
4520
	struct dentry *d_cpu;
4521
	char cpu_dir[30]; /* 30 characters should be more than enough */
4522

4523 4524 4525
	if (!d_percpu)
		return;

4526
	snprintf(cpu_dir, 30, "cpu%ld", cpu);
4527 4528 4529 4530 4531
	d_cpu = debugfs_create_dir(cpu_dir, d_percpu);
	if (!d_cpu) {
		pr_warning("Could not create debugfs '%s' entry\n", cpu_dir);
		return;
	}
4532

4533
	/* per cpu trace_pipe */
4534 4535
	trace_create_file("trace_pipe", 0444, d_cpu,
			(void *) cpu, &tracing_pipe_fops);
4536 4537

	/* per cpu trace */
4538 4539
	trace_create_file("trace", 0644, d_cpu,
			(void *) cpu, &tracing_fops);
4540

4541 4542
	trace_create_file("trace_pipe_raw", 0444, d_cpu,
			(void *) cpu, &tracing_buffers_fops);
4543

4544 4545
	trace_create_file("stats", 0444, d_cpu,
			(void *) cpu, &tracing_stats_fops);
4546 4547 4548

	trace_create_file("buffer_size_kb", 0444, d_cpu,
			(void *) cpu, &tracing_entries_fops);
4549 4550
}

S
Steven Rostedt 已提交
4551 4552 4553 4554 4555
#ifdef CONFIG_FTRACE_SELFTEST
/* Let selftest have access to static functions in this file */
#include "trace_selftest.c"
#endif

4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584
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;
	int ret;

4585 4586
	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
4587 4588
		return ret;

L
Li Zefan 已提交
4589 4590
	if (val != 0 && val != 1)
		return -EINVAL;
4591

L
Li Zefan 已提交
4592
	if (!!(topt->flags->val & topt->opt->bit) != val) {
4593
		mutex_lock(&trace_types_lock);
L
Li Zefan 已提交
4594
		ret = __set_tracer_option(current_trace, topt->flags,
4595
					  topt->opt, !val);
4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610
		mutex_unlock(&trace_types_lock);
		if (ret)
			return ret;
	}

	*ppos += cnt;

	return cnt;
}


static const struct file_operations trace_options_fops = {
	.open = tracing_open_generic,
	.read = trace_options_read,
	.write = trace_options_write,
4611
	.llseek	= generic_file_llseek,
4612 4613
};

4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625 4626 4627 4628 4629 4630 4631 4632 4633 4634 4635 4636
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;
	unsigned long val;
	int ret;

4637 4638
	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
4639 4640
		return ret;

4641
	if (val != 0 && val != 1)
4642
		return -EINVAL;
4643
	set_tracer_flags(1 << index, val);
4644 4645 4646 4647 4648 4649 4650 4651 4652 4653

	*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,
4654
	.llseek = generic_file_llseek,
4655 4656
};

4657
struct dentry *trace_create_file(const char *name,
A
Al Viro 已提交
4658
				 umode_t mode,
4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671 4672
				 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;
}


4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692 4693
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;
}

4694 4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707
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;

4708
	topt->entry = trace_create_file(opt->name, 0644, t_options, topt,
4709 4710 4711 4712 4713 4714 4715 4716 4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733
				    &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++)
		;

4734
	topts = kcalloc(cnt + 1, sizeof(*topts), GFP_KERNEL);
4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760
	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);
}

4761 4762 4763 4764 4765 4766 4767 4768 4769
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;

4770
	return trace_create_file(option, 0644, t_options, (void *)index,
4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782
				    &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;

4783 4784
	for (i = 0; trace_options[i]; i++)
		create_trace_option_core_file(trace_options[i], i);
4785 4786
}

4787 4788 4789 4790
static ssize_t
rb_simple_read(struct file *filp, char __user *ubuf,
	       size_t cnt, loff_t *ppos)
{
4791 4792
	struct trace_array *tr = filp->private_data;
	struct ring_buffer *buffer = tr->buffer;
4793 4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809
	char buf[64];
	int r;

	if (buffer)
		r = ring_buffer_record_is_on(buffer);
	else
		r = 0;

	r = sprintf(buf, "%d\n", r);

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

static ssize_t
rb_simple_write(struct file *filp, const char __user *ubuf,
		size_t cnt, loff_t *ppos)
{
4810 4811
	struct trace_array *tr = filp->private_data;
	struct ring_buffer *buffer = tr->buffer;
4812 4813 4814 4815 4816 4817 4818 4819
	unsigned long val;
	int ret;

	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
		return ret;

	if (buffer) {
4820 4821
		mutex_lock(&trace_types_lock);
		if (val) {
4822
			ring_buffer_record_on(buffer);
4823 4824 4825
			if (current_trace->start)
				current_trace->start(tr);
		} else {
4826
			ring_buffer_record_off(buffer);
4827 4828 4829 4830
			if (current_trace->stop)
				current_trace->stop(tr);
		}
		mutex_unlock(&trace_types_lock);
4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844
	}

	(*ppos)++;

	return cnt;
}

static const struct file_operations rb_simple_fops = {
	.open		= tracing_open_generic,
	.read		= rb_simple_read,
	.write		= rb_simple_write,
	.llseek		= default_llseek,
};

4845
static __init int tracer_init_debugfs(void)
4846 4847
{
	struct dentry *d_tracer;
4848
	int cpu;
4849

4850 4851
	trace_access_lock_init();

4852 4853
	d_tracer = tracing_init_dentry();

4854 4855
	trace_create_file("trace_options", 0644, d_tracer,
			NULL, &tracing_iter_fops);
4856

4857 4858 4859 4860 4861
	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);
4862

4863 4864 4865
	trace_create_file("available_tracers", 0444, d_tracer,
			&global_trace, &show_traces_fops);

4866
	trace_create_file("current_tracer", 0644, d_tracer,
4867 4868
			&global_trace, &set_tracer_fops);

4869
#ifdef CONFIG_TRACER_MAX_TRACE
4870 4871
	trace_create_file("tracing_max_latency", 0644, d_tracer,
			&tracing_max_latency, &tracing_max_lat_fops);
4872
#endif
4873 4874 4875

	trace_create_file("tracing_thresh", 0644, d_tracer,
			&tracing_thresh, &tracing_max_lat_fops);
4876

4877
	trace_create_file("README", 0444, d_tracer,
4878 4879 4880
			NULL, &tracing_readme_fops);

	trace_create_file("trace_pipe", 0444, d_tracer,
4881
			(void *) TRACE_PIPE_ALL_CPU, &tracing_pipe_fops);
4882 4883

	trace_create_file("buffer_size_kb", 0644, d_tracer,
4884
			(void *) RING_BUFFER_ALL_CPUS, &tracing_entries_fops);
4885

4886 4887 4888
	trace_create_file("buffer_total_size_kb", 0444, d_tracer,
			&global_trace, &tracing_total_entries_fops);

4889 4890 4891
	trace_create_file("free_buffer", 0644, d_tracer,
			&global_trace, &tracing_free_buffer_fops);

4892 4893
	trace_create_file("trace_marker", 0220, d_tracer,
			NULL, &tracing_mark_fops);
4894

4895 4896
	trace_create_file("saved_cmdlines", 0444, d_tracer,
			NULL, &tracing_saved_cmdlines_fops);
4897

4898 4899 4900
	trace_create_file("trace_clock", 0644, d_tracer, NULL,
			  &trace_clock_fops);

4901
	trace_create_file("tracing_on", 0644, d_tracer,
4902
			    &global_trace, &rb_simple_fops);
4903

4904
#ifdef CONFIG_DYNAMIC_FTRACE
4905 4906
	trace_create_file("dyn_ftrace_total_info", 0444, d_tracer,
			&ftrace_update_tot_cnt, &tracing_dyn_info_fops);
4907
#endif
4908

4909 4910
	create_trace_options_dir();

4911 4912 4913
	for_each_tracing_cpu(cpu)
		tracing_init_debugfs_percpu(cpu);

4914
	return 0;
4915 4916
}

4917 4918 4919
static int trace_panic_handler(struct notifier_block *this,
			       unsigned long event, void *unused)
{
4920
	if (ftrace_dump_on_oops)
4921
		ftrace_dump(ftrace_dump_on_oops);
4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936
	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:
4937
		if (ftrace_dump_on_oops)
4938
			ftrace_dump(ftrace_dump_on_oops);
4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961
		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.
 */
4962
#define KERN_TRACE		KERN_EMERG
4963

4964
void
4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975
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);

4976
	trace_seq_init(s);
4977 4978
}

4979 4980 4981 4982 4983 4984 4985
void trace_init_global_iter(struct trace_iterator *iter)
{
	iter->tr = &global_trace;
	iter->trace = current_trace;
	iter->cpu_file = TRACE_PIPE_ALL_CPU;
}

4986 4987
static void
__ftrace_dump(bool disable_tracing, enum ftrace_dump_mode oops_dump_mode)
4988
{
4989
	static arch_spinlock_t ftrace_dump_lock =
4990
		(arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED;
4991 4992
	/* use static because iter can be a bit big for the stack */
	static struct trace_iterator iter;
4993
	unsigned int old_userobj;
4994
	static int dump_ran;
4995 4996
	unsigned long flags;
	int cnt = 0, cpu;
4997 4998

	/* only one dump */
4999
	local_irq_save(flags);
5000
	arch_spin_lock(&ftrace_dump_lock);
5001 5002 5003 5004 5005
	if (dump_ran)
		goto out;

	dump_ran = 1;

5006
	tracing_off();
5007

5008 5009 5010 5011 5012 5013
	/* Did function tracer already get disabled? */
	if (ftrace_is_dead()) {
		printk("# WARNING: FUNCTION TRACING IS CORRUPTED\n");
		printk("#          MAY BE MISSING FUNCTION EVENTS\n");
	}

5014 5015
	if (disable_tracing)
		ftrace_kill();
5016

5017 5018
	trace_init_global_iter(&iter);

5019
	for_each_tracing_cpu(cpu) {
5020
		atomic_inc(&iter.tr->data[cpu]->disabled);
5021 5022
	}

5023 5024
	old_userobj = trace_flags & TRACE_ITER_SYM_USEROBJ;

5025 5026 5027
	/* don't look at user memory in panic mode */
	trace_flags &= ~TRACE_ITER_SYM_USEROBJ;

5028
	/* Simulate the iterator */
5029 5030
	iter.tr = &global_trace;
	iter.trace = current_trace;
5031 5032 5033 5034 5035 5036 5037 5038 5039 5040 5041 5042 5043 5044 5045 5046

	switch (oops_dump_mode) {
	case DUMP_ALL:
		iter.cpu_file = TRACE_PIPE_ALL_CPU;
		break;
	case DUMP_ORIG:
		iter.cpu_file = raw_smp_processor_id();
		break;
	case DUMP_NONE:
		goto out_enable;
	default:
		printk(KERN_TRACE "Bad dumping mode, switching to all CPUs dump\n");
		iter.cpu_file = TRACE_PIPE_ALL_CPU;
	}

	printk(KERN_TRACE "Dumping ftrace buffer:\n");
5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 5057 5058 5059 5060 5061 5062 5063 5064 5065 5066 5067 5068

	/*
	 * 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;

5069
		if (trace_find_next_entry_inc(&iter) != NULL) {
5070 5071 5072 5073 5074
			int ret;

			ret = print_trace_line(&iter);
			if (ret != TRACE_TYPE_NO_CONSUME)
				trace_consume(&iter);
5075
		}
5076
		touch_nmi_watchdog();
5077 5078 5079 5080 5081 5082 5083 5084 5085

		trace_printk_seq(&iter.seq);
	}

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

5086
 out_enable:
5087 5088 5089 5090 5091
	/* Re-enable tracing if requested */
	if (!disable_tracing) {
		trace_flags |= old_userobj;

		for_each_tracing_cpu(cpu) {
5092
			atomic_dec(&iter.tr->data[cpu]->disabled);
5093 5094 5095 5096
		}
		tracing_on();
	}

5097
 out:
5098
	arch_spin_unlock(&ftrace_dump_lock);
5099
	local_irq_restore(flags);
5100 5101
}

5102
/* By default: disable tracing after the dump */
5103
void ftrace_dump(enum ftrace_dump_mode oops_dump_mode)
5104
{
5105
	__ftrace_dump(true, oops_dump_mode);
5106
}
5107
EXPORT_SYMBOL_GPL(ftrace_dump);
5108

5109
__init static int tracer_alloc_buffers(void)
5110
{
5111
	int ring_buf_size;
5112
	enum ring_buffer_flags rb_flags;
5113
	int i;
5114
	int ret = -ENOMEM;
5115

5116

5117 5118 5119 5120 5121
	if (!alloc_cpumask_var(&tracing_buffer_mask, GFP_KERNEL))
		goto out;

	if (!alloc_cpumask_var(&tracing_cpumask, GFP_KERNEL))
		goto out_free_buffer_mask;
5122

5123 5124
	/* Only allocate trace_printk buffers if a trace_printk exists */
	if (__stop___trace_bprintk_fmt != __start___trace_bprintk_fmt)
5125
		/* Must be called before global_trace.buffer is allocated */
5126 5127
		trace_printk_init_buffers();

5128 5129 5130 5131 5132 5133
	/* 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;

5134 5135
	rb_flags = trace_flags & TRACE_ITER_OVERWRITE ? RB_FL_OVERWRITE : 0;

5136 5137 5138 5139
	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 */
5140
	global_trace.buffer = ring_buffer_alloc(ring_buf_size, rb_flags);
5141 5142 5143
	if (!global_trace.buffer) {
		printk(KERN_ERR "tracer: failed to allocate ring buffer!\n");
		WARN_ON(1);
5144
		goto out_free_cpumask;
5145
	}
5146 5147
	if (global_trace.buffer_disabled)
		tracing_off();
5148

5149

5150
#ifdef CONFIG_TRACER_MAX_TRACE
5151
	max_tr.buffer = ring_buffer_alloc(1, rb_flags);
5152 5153 5154 5155
	if (!max_tr.buffer) {
		printk(KERN_ERR "tracer: failed to allocate max ring buffer!\n");
		WARN_ON(1);
		ring_buffer_free(global_trace.buffer);
5156
		goto out_free_cpumask;
5157
	}
5158
#endif
5159

5160
	/* Allocate the first page for all buffers */
5161
	for_each_tracing_cpu(i) {
5162
		global_trace.data[i] = &per_cpu(global_trace_cpu, i);
5163
		max_tr.data[i] = &per_cpu(max_tr_data, i);
5164
	}
5165

5166 5167
	set_buffer_entries(&global_trace,
			   ring_buffer_size(global_trace.buffer, 0));
5168 5169 5170
#ifdef CONFIG_TRACER_MAX_TRACE
	set_buffer_entries(&max_tr, 1);
#endif
5171 5172

	trace_init_cmdlines();
5173
	init_irq_work(&trace_work_wakeup, trace_wake_up);
5174

5175
	register_tracer(&nop_trace);
5176
	current_trace = &nop_trace;
S
Steven Rostedt 已提交
5177 5178
	/* All seems OK, enable tracing */
	tracing_disabled = 0;
5179

5180 5181 5182 5183
	atomic_notifier_chain_register(&panic_notifier_list,
				       &trace_panic_notifier);

	register_die_notifier(&trace_die_notifier);
5184

5185 5186 5187 5188 5189 5190 5191
	while (trace_boot_options) {
		char *option;

		option = strsep(&trace_boot_options, ",");
		trace_set_options(option);
	}

5192
	return 0;
5193

5194 5195 5196 5197 5198 5199
out_free_cpumask:
	free_cpumask_var(tracing_cpumask);
out_free_buffer_mask:
	free_cpumask_var(tracing_buffer_mask);
out:
	return ret;
5200
}
5201 5202 5203 5204 5205 5206 5207 5208 5209 5210 5211 5212 5213 5214 5215 5216 5217 5218 5219 5220

__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;
}

5221 5222
early_initcall(tracer_alloc_buffers);
fs_initcall(tracer_init_debugfs);
5223
late_initcall(clear_boot_tracer);