trace.c 103.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13
/*
 * ring buffer based function tracer
 *
 * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
 * Copyright (C) 2008 Ingo Molnar <mingo@redhat.com>
 *
 * Originally taken from the RT patch by:
 *    Arnaldo Carvalho de Melo <acme@redhat.com>
 *
 * Based on code from the latency_tracer, that is:
 *  Copyright (C) 2004-2006 Ingo Molnar
 *  Copyright (C) 2004 William Lee Irwin III
 */
14
#include <linux/ring_buffer.h>
15
#include <generated/utsrelease.h>
16 17
#include <linux/stacktrace.h>
#include <linux/writeback.h>
18 19
#include <linux/kallsyms.h>
#include <linux/seq_file.h>
A
Alexey Dobriyan 已提交
20
#include <linux/smp_lock.h>
21
#include <linux/notifier.h>
22
#include <linux/irqflags.h>
23
#include <linux/debugfs.h>
24
#include <linux/pagemap.h>
25 26 27
#include <linux/hardirq.h>
#include <linux/linkage.h>
#include <linux/uaccess.h>
28
#include <linux/kprobes.h>
29 30 31
#include <linux/ftrace.h>
#include <linux/module.h>
#include <linux/percpu.h>
32
#include <linux/splice.h>
33
#include <linux/kdebug.h>
34
#include <linux/string.h>
35
#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/fs.h>
I
Ingo Molnar 已提交
41

42
#include "trace.h"
43
#include "trace_output.h"
44

45 46
#define TRACE_BUFFER_FLAGS	(RB_FL_OVERWRITE)

47 48 49 50
/*
 * 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.
 */
51
int ring_buffer_expanded;
52

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

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

67 68 69 70 71 72 73 74 75 76 77 78 79 80
/* For tracers that don't implement custom flags */
static struct tracer_opt dummy_tracer_opt[] = {
	{ }
};

static struct tracer_flags dummy_tracer_flags = {
	.val = 0,
	.opts = dummy_tracer_opt
};

static int dummy_set_flag(u32 old_flags, u32 bit, int set)
{
	return 0;
}
81 82 83 84 85 86 87

/*
 * Kill all tracing for good (never come back).
 * It is initialized to 1 but will turn to zero if the initialization
 * of the tracer is successful. But that is the only place that sets
 * this back to zero.
 */
88
static int tracing_disabled = 1;
89

90
DEFINE_PER_CPU(int, ftrace_cpu_disabled);
91 92 93 94

static inline void ftrace_disable_cpu(void)
{
	preempt_disable();
R
Rusty Russell 已提交
95
	__this_cpu_inc(ftrace_cpu_disabled);
96 97 98 99
}

static inline void ftrace_enable_cpu(void)
{
R
Rusty Russell 已提交
100
	__this_cpu_dec(ftrace_cpu_disabled);
101 102 103
	preempt_enable();
}

104
static cpumask_var_t __read_mostly	tracing_buffer_mask;
105 106

#define for_each_tracing_cpu(cpu)	\
107
	for_each_cpu(cpu, tracing_buffer_mask)
108

109 110 111 112 113 114 115 116 117 118 119
/*
 * 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
120 121 122
 * /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
123
 */
124 125

enum ftrace_dump_mode ftrace_dump_on_oops;
126

127 128
static int tracing_set_tracer(const char *buf);

L
Li Zefan 已提交
129 130
#define MAX_TRACER_SIZE		100
static char bootup_tracer_buf[MAX_TRACER_SIZE] __initdata;
131
static char *default_bootup_tracer;
132

133
static int __init set_cmdline_ftrace(char *str)
134
{
L
Li Zefan 已提交
135
	strncpy(bootup_tracer_buf, str, MAX_TRACER_SIZE);
136
	default_bootup_tracer = bootup_tracer_buf;
137 138
	/* We are using ftrace early, expand it */
	ring_buffer_expanded = 1;
139 140
	return 1;
}
141
__setup("ftrace=", set_cmdline_ftrace);
142

143 144
static int __init set_ftrace_dump_on_oops(char *str)
{
145 146 147 148 149 150 151 152 153 154 155
	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;
156 157
}
__setup("ftrace_dump_on_oops", set_ftrace_dump_on_oops);
S
Steven Rostedt 已提交
158

159
unsigned long long ns2usecs(cycle_t nsec)
160 161 162 163 164 165
{
	nsec += 500;
	do_div(nsec, 1000);
	return nsec;
}

S
Steven Rostedt 已提交
166 167 168 169 170 171 172 173 174 175 176 177
/*
 * 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.
 */
178 179 180 181
static struct trace_array	global_trace;

static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu);

182 183
int filter_current_check_discard(struct ring_buffer *buffer,
				 struct ftrace_event_call *call, void *rec,
184 185
				 struct ring_buffer_event *event)
{
186
	return filter_check_discard(call, rec, buffer, event);
187
}
188
EXPORT_SYMBOL_GPL(filter_current_check_discard);
189

190 191 192 193 194 195 196 197 198 199 200 201 202
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;
}
203

S
Steven Rostedt 已提交
204 205 206 207 208 209 210 211 212 213
/*
 * 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.
 */
214 215
static struct trace_array	max_tr;

216
static DEFINE_PER_CPU(struct trace_array_cpu, max_tr_data);
217

S
Steven Rostedt 已提交
218
/* tracer_enabled is used to toggle activation of a tracer */
219
static int			tracer_enabled = 1;
S
Steven Rostedt 已提交
220

221 222 223 224 225 226 227 228 229 230 231 232 233
/**
 * tracing_is_enabled - return tracer_enabled status
 *
 * This function is used by other tracers to know the status
 * of the tracer_enabled flag.  Tracers may use this function
 * to know if it should enable their features when starting
 * up. See irqsoff tracer for an example (start_irqsoff_tracer).
 */
int tracing_is_enabled(void)
{
	return tracer_enabled;
}

S
Steven Rostedt 已提交
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;
348

349 350 351
static int trace_stop_count;
static DEFINE_SPINLOCK(tracing_start_lock);

S
Steven Rostedt 已提交
352 353 354 355 356 357
/**
 * trace_wake_up - wake up tasks waiting for trace input
 *
 * Simply wakes up any task that is blocked on the trace_wait
 * queue. These is used with trace_poll for tasks polling the trace.
 */
358 359
void trace_wake_up(void)
{
360 361 362 363
	int cpu;

	if (trace_flags & TRACE_ITER_BLOCK)
		return;
I
Ingo Molnar 已提交
364 365 366 367
	/*
	 * The runqueue_is_locked() can fail, but this is the best we
	 * have for now:
	 */
368 369
	cpu = get_cpu();
	if (!runqueue_is_locked(cpu))
370
		wake_up(&trace_wait);
371
	put_cpu();
372
}
373

374
static int __init set_buf_size(char *str)
375
{
376
	unsigned long buf_size;
377

378 379
	if (!str)
		return 0;
380
	buf_size = memparse(str, &str);
381
	/* nr_entries can not be zero */
382
	if (buf_size == 0)
383
		return 0;
384
	trace_buf_size = buf_size;
385 386
	return 1;
}
387
__setup("trace_buf_size=", set_buf_size);
388

389 390 391 392 393 394 395 396 397 398 399 400 401 402 403
static int __init set_tracing_thresh(char *str)
{
	unsigned long threshhold;
	int ret;

	if (!str)
		return 0;
	ret = strict_strtoul(str, 0, &threshhold);
	if (ret < 0)
		return 0;
	tracing_thresh = threshhold * 1000;
	return 1;
}
__setup("tracing_thresh=", set_tracing_thresh);

S
Steven Rostedt 已提交
404 405 406 407 408
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

S
Steven Rostedt 已提交
409
/* These must match the bit postions in trace_iterator_flags */
410 411 412 413 414
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
I
Ingo Molnar 已提交
415
	"raw",
416
	"hex",
I
Ingo Molnar 已提交
417
	"bin",
418
	"block",
I
Ingo Molnar 已提交
419
	"stacktrace",
420
	"trace_printk",
421
	"ftrace_preempt",
422
	"branch",
423
	"annotate",
424
	"userstacktrace",
425
	"sym-userobj",
426
	"printk-msg-only",
427
	"context-info",
428
	"latency-format",
429
	"sleep-time",
430
	"graph-time",
431
	"record-cmd",
432 433 434
	NULL
};

435 436 437 438 439 440 441 442 443 444
static struct {
	u64 (*func)(void);
	const char *name;
} trace_clocks[] = {
	{ trace_clock_local,	"local" },
	{ trace_clock_global,	"global" },
};

int trace_clock_id;

445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521
/*
 * 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)) {
522
		if (parser->idx < parser->size - 1)
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
			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;
}

551 552 553 554 555
ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt)
{
	int len;
	int ret;

556 557 558
	if (!cnt)
		return 0;

559 560 561 562 563 564 565
	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);
566
	if (ret == cnt)
567 568
		return -EFAULT;

569 570
	cnt -= ret;

571
	s->readpos += cnt;
572
	return cnt;
S
Steven Rostedt 已提交
573 574
}

575
static ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt)
576 577 578 579 580 581 582 583 584 585 586 587 588 589
{
	int len;
	void *ret;

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

	len = s->len - s->readpos;
	if (cnt > len)
		cnt = len;
	ret = memcpy(buf, s->buffer + s->readpos, cnt);
	if (!ret)
		return -EFAULT;

590
	s->readpos += cnt;
591 592 593
	return cnt;
}

594 595 596 597 598 599
/*
 * 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.
 *
600
 * This is defined as a arch_spinlock_t in order to help
601 602 603 604 605 606
 * 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.
 */
607
static arch_spinlock_t ftrace_max_lock =
608
	(arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED;
609

610 611
unsigned long __read_mostly	tracing_thresh;

612 613 614 615 616 617 618 619 620 621 622 623
#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];
624
	struct trace_array_cpu *max_data;
625 626 627 628

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

629 630 631 632
	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;
633

634
	memcpy(max_data->comm, tsk->comm, TASK_COMM_LEN);
635 636 637 638 639
	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;
640 641 642 643 644

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

S
Steven Rostedt 已提交
645 646 647 648 649 650 651 652 653
/**
 * 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
654
void
655 656
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
657
	struct ring_buffer *buf = tr->buffer;
658

659 660 661
	if (trace_stop_count)
		return;

662
	WARN_ON_ONCE(!irqs_disabled());
663 664 665 666
	if (!current_trace->use_max_tr) {
		WARN_ON_ONCE(1);
		return;
	}
667
	arch_spin_lock(&ftrace_max_lock);
668 669 670 671

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

672
	__update_max_tr(tr, tsk, cpu);
673
	arch_spin_unlock(&ftrace_max_lock);
674 675 676 677 678 679 680
}

/**
 * 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
681 682
 *
 * Flip the trace of a single CPU buffer between the @tr and the max_tr.
683
 */
I
Ingo Molnar 已提交
684
void
685 686
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
687
	int ret;
688

689 690 691
	if (trace_stop_count)
		return;

692
	WARN_ON_ONCE(!irqs_disabled());
693 694 695 696 697
	if (!current_trace->use_max_tr) {
		WARN_ON_ONCE(1);
		return;
	}

698
	arch_spin_lock(&ftrace_max_lock);
699

700 701
	ftrace_disable_cpu();

702 703
	ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu);

704 705 706 707 708 709 710 711 712 713 714
	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");
	}

715 716
	ftrace_enable_cpu();

717
	WARN_ON_ONCE(ret && ret != -EAGAIN && ret != -EBUSY);
718 719

	__update_max_tr(tr, tsk, cpu);
720
	arch_spin_unlock(&ftrace_max_lock);
721
}
722
#endif /* CONFIG_TRACER_MAX_TRACE */
723

S
Steven Rostedt 已提交
724 725 726 727 728 729
/**
 * register_tracer - register a tracer with the ftrace system.
 * @type - the plugin for the tracer
 *
 * Register a new plugin tracer.
 */
730
int register_tracer(struct tracer *type)
731 732
__releases(kernel_lock)
__acquires(kernel_lock)
733 734 735 736 737 738 739 740 741
{
	struct tracer *t;
	int ret = 0;

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

742
	if (strlen(type->name) >= MAX_TRACER_SIZE) {
L
Li Zefan 已提交
743 744 745 746
		pr_info("Tracer has a name longer than %d\n", MAX_TRACER_SIZE);
		return -1;
	}

I
Ingo Molnar 已提交
747 748 749 750 751 752 753
	/*
	 * When this gets called we hold the BKL which means that
	 * preemption is disabled. Various trace selftests however
	 * need to disable and enable preemption for successful tests.
	 * So we drop the BKL here and grab it after the tests again.
	 */
	unlock_kernel();
754
	mutex_lock(&trace_types_lock);
I
Ingo Molnar 已提交
755

756 757
	tracing_selftest_running = true;

758 759 760
	for (t = trace_types; t; t = t->next) {
		if (strcmp(type->name, t->name) == 0) {
			/* already found */
L
Li Zefan 已提交
761
			pr_info("Tracer %s already registered\n",
762 763 764 765 766 767
				type->name);
			ret = -1;
			goto out;
		}
	}

768 769 770 771 772 773 774
	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;
775 776 777
	if (!type->wait_pipe)
		type->wait_pipe = default_wait_pipe;

778

S
Steven Rostedt 已提交
779
#ifdef CONFIG_FTRACE_STARTUP_TEST
780
	if (type->selftest && !tracing_selftest_disabled) {
S
Steven Rostedt 已提交
781 782
		struct tracer *saved_tracer = current_trace;
		struct trace_array *tr = &global_trace;
783

S
Steven Rostedt 已提交
784 785 786 787 788 789 790
		/*
		 * 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.
		 */
791
		tracing_reset_online_cpus(tr);
I
Ingo Molnar 已提交
792

S
Steven Rostedt 已提交
793 794 795 796 797 798 799 800 801 802
		current_trace = type;
		/* the test is responsible for initializing and enabling */
		pr_info("Testing tracer %s: ", type->name);
		ret = type->selftest(type, tr);
		/* the test is responsible for resetting too */
		current_trace = saved_tracer;
		if (ret) {
			printk(KERN_CONT "FAILED!\n");
			goto out;
		}
S
Steven Rostedt 已提交
803
		/* Only reset on passing, to avoid touching corrupted buffers */
804
		tracing_reset_online_cpus(tr);
I
Ingo Molnar 已提交
805

S
Steven Rostedt 已提交
806 807 808 809
		printk(KERN_CONT "PASSED\n");
	}
#endif

810 811
	type->next = trace_types;
	trace_types = type;
S
Steven Rostedt 已提交
812

813
 out:
814
	tracing_selftest_running = false;
815 816
	mutex_unlock(&trace_types_lock);

S
Steven Rostedt 已提交
817 818 819
	if (ret || !default_bootup_tracer)
		goto out_unlock;

L
Li Zefan 已提交
820
	if (strncmp(default_bootup_tracer, type->name, MAX_TRACER_SIZE))
S
Steven Rostedt 已提交
821 822 823 824 825 826 827 828
		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;
829
#ifdef CONFIG_FTRACE_STARTUP_TEST
S
Steven Rostedt 已提交
830 831
	printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n",
	       type->name);
832 833
#endif

S
Steven Rostedt 已提交
834
 out_unlock:
835
	lock_kernel();
836 837 838 839 840 841 842 843 844 845 846 847
	return ret;
}

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

	mutex_lock(&trace_types_lock);
	for (t = &trace_types; *t; t = &(*t)->next) {
		if (*t == type)
			goto found;
	}
L
Li Zefan 已提交
848
	pr_info("Tracer %s not registered\n", type->name);
849 850 851 852
	goto out;

 found:
	*t = (*t)->next;
853 854 855 856 857 858 859 860

	if (type == current_trace && tracer_enabled) {
		tracer_enabled = 0;
		tracing_stop();
		if (current_trace->stop)
			current_trace->stop(&global_trace);
		current_trace = &nop_trace;
	}
L
Li Zefan 已提交
861
out:
862 863 864
	mutex_unlock(&trace_types_lock);
}

865
static void __tracing_reset(struct ring_buffer *buffer, int cpu)
866
{
867
	ftrace_disable_cpu();
868
	ring_buffer_reset_cpu(buffer, cpu);
869
	ftrace_enable_cpu();
870 871
}

872 873 874 875 876 877 878 879
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();
880
	__tracing_reset(buffer, cpu);
881 882 883 884

	ring_buffer_record_enable(buffer);
}

885 886
void tracing_reset_online_cpus(struct trace_array *tr)
{
887
	struct ring_buffer *buffer = tr->buffer;
888 889
	int cpu;

890 891 892 893 894
	ring_buffer_record_disable(buffer);

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

895 896 897
	tr->time_start = ftrace_now(tr->cpu);

	for_each_online_cpu(cpu)
898
		__tracing_reset(buffer, cpu);
899 900

	ring_buffer_record_enable(buffer);
901 902
}

903 904 905 906 907 908 909 910 911 912
void tracing_reset_current(int cpu)
{
	tracing_reset(&global_trace, cpu);
}

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

913
#define SAVED_CMDLINES 128
914
#define NO_CMDLINE_MAP UINT_MAX
915 916 917 918
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;
919
static arch_spinlock_t trace_cmdline_lock = __ARCH_SPIN_LOCK_UNLOCKED;
920 921

/* temporary disable recording */
922
static atomic_t trace_record_cmdline_disabled __read_mostly;
923 924 925

static void trace_init_cmdlines(void)
{
926 927
	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));
928 929 930
	cmdline_idx = 0;
}

931 932 933 934 935
int is_tracing_stopped(void)
{
	return trace_stop_count;
}

S
Steven Rostedt 已提交
936 937 938 939 940 941 942 943 944 945 946 947 948 949 950
/**
 * 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();
}

951 952 953 954 955 956 957 958 959 960 961 962 963 964 965
/**
 * tracing_start - quick start of the tracer
 *
 * If tracing is enabled but was stopped by tracing_stop,
 * this will start the tracer back up.
 */
void tracing_start(void)
{
	struct ring_buffer *buffer;
	unsigned long flags;

	if (tracing_disabled)
		return;

	spin_lock_irqsave(&tracing_start_lock, flags);
966 967 968 969 970 971
	if (--trace_stop_count) {
		if (trace_stop_count < 0) {
			/* Someone screwed up their debugging */
			WARN_ON_ONCE(1);
			trace_stop_count = 0;
		}
972 973 974
		goto out;
	}

975 976
	/* Prevent the buffers from switching */
	arch_spin_lock(&ftrace_max_lock);
977 978 979 980 981 982 983 984 985

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

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

986 987
	arch_spin_unlock(&ftrace_max_lock);

988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008
	ftrace_start();
 out:
	spin_unlock_irqrestore(&tracing_start_lock, flags);
}

/**
 * tracing_stop - quick stop of the tracer
 *
 * Light weight way to stop tracing. Use in conjunction with
 * tracing_start.
 */
void tracing_stop(void)
{
	struct ring_buffer *buffer;
	unsigned long flags;

	ftrace_stop();
	spin_lock_irqsave(&tracing_start_lock, flags);
	if (trace_stop_count++)
		goto out;

1009 1010 1011
	/* Prevent the buffers from switching */
	arch_spin_lock(&ftrace_max_lock);

1012 1013 1014 1015 1016 1017 1018 1019
	buffer = global_trace.buffer;
	if (buffer)
		ring_buffer_record_disable(buffer);

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

1020 1021
	arch_spin_unlock(&ftrace_max_lock);

1022 1023 1024 1025
 out:
	spin_unlock_irqrestore(&tracing_start_lock, flags);
}

I
Ingo Molnar 已提交
1026
void trace_stop_cmdline_recording(void);
1027

I
Ingo Molnar 已提交
1028
static void trace_save_cmdline(struct task_struct *tsk)
1029
{
1030
	unsigned pid, idx;
1031 1032 1033 1034 1035 1036 1037 1038 1039 1040

	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.
	 */
1041
	if (!arch_spin_trylock(&trace_cmdline_lock))
1042 1043 1044
		return;

	idx = map_pid_to_cmdline[tsk->pid];
1045
	if (idx == NO_CMDLINE_MAP) {
1046 1047
		idx = (cmdline_idx + 1) % SAVED_CMDLINES;

1048 1049 1050 1051 1052 1053 1054 1055 1056
		/*
		 * 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;
1057

1058
		map_cmdline_to_pid[idx] = tsk->pid;
1059 1060 1061 1062 1063 1064 1065
		map_pid_to_cmdline[tsk->pid] = idx;

		cmdline_idx = idx;
	}

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

1066
	arch_spin_unlock(&trace_cmdline_lock);
1067 1068
}

1069
void trace_find_cmdline(int pid, char comm[])
1070 1071 1072
{
	unsigned map;

1073 1074 1075 1076
	if (!pid) {
		strcpy(comm, "<idle>");
		return;
	}
1077

1078 1079 1080 1081 1082
	if (WARN_ON_ONCE(pid < 0)) {
		strcpy(comm, "<XXX>");
		return;
	}

1083 1084 1085 1086
	if (pid > PID_MAX_DEFAULT) {
		strcpy(comm, "<...>");
		return;
	}
1087

1088
	preempt_disable();
1089
	arch_spin_lock(&trace_cmdline_lock);
1090
	map = map_pid_to_cmdline[pid];
1091 1092 1093 1094
	if (map != NO_CMDLINE_MAP)
		strcpy(comm, saved_cmdlines[map]);
	else
		strcpy(comm, "<...>");
1095

1096
	arch_spin_unlock(&trace_cmdline_lock);
1097
	preempt_enable();
1098 1099
}

I
Ingo Molnar 已提交
1100
void tracing_record_cmdline(struct task_struct *tsk)
1101
{
1102 1103
	if (atomic_read(&trace_record_cmdline_disabled) || !tracer_enabled ||
	    !tracing_is_on())
1104 1105 1106 1107 1108
		return;

	trace_save_cmdline(tsk);
}

1109
void
1110 1111
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags,
			     int pc)
1112 1113 1114
{
	struct task_struct *tsk = current;

1115 1116
	entry->preempt_count		= pc & 0xff;
	entry->pid			= (tsk) ? tsk->pid : 0;
1117
	entry->lock_depth		= (tsk) ? tsk->lock_depth : 0;
1118
	entry->flags =
1119
#ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
1120
		(irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
1121 1122 1123
#else
		TRACE_FLAG_IRQS_NOSUPPORT |
#endif
1124 1125 1126 1127
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}
1128
EXPORT_SYMBOL_GPL(tracing_generic_entry_update);
1129

1130 1131 1132 1133 1134
struct ring_buffer_event *
trace_buffer_lock_reserve(struct ring_buffer *buffer,
			  int type,
			  unsigned long len,
			  unsigned long flags, int pc)
1135 1136 1137
{
	struct ring_buffer_event *event;

1138
	event = ring_buffer_lock_reserve(buffer, len);
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148
	if (event != NULL) {
		struct trace_entry *ent = ring_buffer_event_data(event);

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

	return event;
}

1149 1150 1151 1152 1153
static inline void
__trace_buffer_unlock_commit(struct ring_buffer *buffer,
			     struct ring_buffer_event *event,
			     unsigned long flags, int pc,
			     int wake)
1154
{
1155
	ring_buffer_unlock_commit(buffer, event);
1156

1157 1158
	ftrace_trace_stack(buffer, flags, 6, pc);
	ftrace_trace_userstack(buffer, flags, pc);
1159 1160 1161 1162 1163

	if (wake)
		trace_wake_up();
}

1164 1165 1166
void trace_buffer_unlock_commit(struct ring_buffer *buffer,
				struct ring_buffer_event *event,
				unsigned long flags, int pc)
1167
{
1168
	__trace_buffer_unlock_commit(buffer, event, flags, pc, 1);
1169 1170
}

1171
struct ring_buffer_event *
1172 1173
trace_current_buffer_lock_reserve(struct ring_buffer **current_rb,
				  int type, unsigned long len,
1174 1175
				  unsigned long flags, int pc)
{
1176 1177
	*current_rb = global_trace.buffer;
	return trace_buffer_lock_reserve(*current_rb,
1178 1179
					 type, len, flags, pc);
}
1180
EXPORT_SYMBOL_GPL(trace_current_buffer_lock_reserve);
1181

1182 1183
void trace_current_buffer_unlock_commit(struct ring_buffer *buffer,
					struct ring_buffer_event *event,
1184 1185
					unsigned long flags, int pc)
{
1186
	__trace_buffer_unlock_commit(buffer, event, flags, pc, 1);
1187
}
1188
EXPORT_SYMBOL_GPL(trace_current_buffer_unlock_commit);
1189

1190 1191 1192
void trace_nowake_buffer_unlock_commit(struct ring_buffer *buffer,
				       struct ring_buffer_event *event,
				       unsigned long flags, int pc)
1193
{
1194
	__trace_buffer_unlock_commit(buffer, event, flags, pc, 0);
1195
}
1196
EXPORT_SYMBOL_GPL(trace_nowake_buffer_unlock_commit);
1197

1198 1199
void trace_current_buffer_discard_commit(struct ring_buffer *buffer,
					 struct ring_buffer_event *event)
1200
{
1201
	ring_buffer_discard_commit(buffer, event);
1202
}
1203
EXPORT_SYMBOL_GPL(trace_current_buffer_discard_commit);
1204

I
Ingo Molnar 已提交
1205
void
1206
trace_function(struct trace_array *tr,
1207 1208
	       unsigned long ip, unsigned long parent_ip, unsigned long flags,
	       int pc)
1209
{
1210
	struct ftrace_event_call *call = &event_function;
1211
	struct ring_buffer *buffer = tr->buffer;
1212
	struct ring_buffer_event *event;
1213
	struct ftrace_entry *entry;
1214

1215
	/* If we are reading the ring buffer, don't trace */
R
Rusty Russell 已提交
1216
	if (unlikely(__this_cpu_read(ftrace_cpu_disabled)))
1217 1218
		return;

1219
	event = trace_buffer_lock_reserve(buffer, TRACE_FN, sizeof(*entry),
1220
					  flags, pc);
1221 1222 1223
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1224 1225
	entry->ip			= ip;
	entry->parent_ip		= parent_ip;
1226

1227 1228
	if (!filter_check_discard(call, entry, buffer, event))
		ring_buffer_unlock_commit(buffer, event);
1229 1230
}

I
Ingo Molnar 已提交
1231
void
I
Ingo Molnar 已提交
1232
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
1233 1234
       unsigned long ip, unsigned long parent_ip, unsigned long flags,
       int pc)
I
Ingo Molnar 已提交
1235 1236
{
	if (likely(!atomic_read(&data->disabled)))
1237
		trace_function(tr, ip, parent_ip, flags, pc);
I
Ingo Molnar 已提交
1238 1239
}

1240
#ifdef CONFIG_STACKTRACE
1241
static void __ftrace_trace_stack(struct ring_buffer *buffer,
1242 1243
				 unsigned long flags,
				 int skip, int pc)
I
Ingo Molnar 已提交
1244
{
1245
	struct ftrace_event_call *call = &event_kernel_stack;
1246
	struct ring_buffer_event *event;
1247
	struct stack_entry *entry;
I
Ingo Molnar 已提交
1248 1249
	struct stack_trace trace;

1250
	event = trace_buffer_lock_reserve(buffer, TRACE_STACK,
1251
					  sizeof(*entry), flags, pc);
1252 1253 1254
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1255
	memset(&entry->caller, 0, sizeof(entry->caller));
I
Ingo Molnar 已提交
1256 1257 1258 1259

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
1260
	trace.entries		= entry->caller;
I
Ingo Molnar 已提交
1261 1262

	save_stack_trace(&trace);
1263 1264
	if (!filter_check_discard(call, entry, buffer, event))
		ring_buffer_unlock_commit(buffer, event);
I
Ingo Molnar 已提交
1265 1266
}

1267 1268
void ftrace_trace_stack(struct ring_buffer *buffer, unsigned long flags,
			int skip, int pc)
1269 1270 1271 1272
{
	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

1273
	__ftrace_trace_stack(buffer, flags, skip, pc);
1274 1275
}

1276 1277
void __trace_stack(struct trace_array *tr, unsigned long flags, int skip,
		   int pc)
1278
{
1279
	__ftrace_trace_stack(tr->buffer, flags, skip, pc);
1280 1281
}

S
Steven Rostedt 已提交
1282 1283 1284 1285 1286 1287 1288 1289
/**
 * 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)
1290
		return;
S
Steven Rostedt 已提交
1291 1292 1293 1294 1295 1296 1297

	local_save_flags(flags);

	/* skipping 3 traces, seems to get us at the caller of this function */
	__ftrace_trace_stack(global_trace.buffer, flags, 3, preempt_count());
}

1298 1299
void
ftrace_trace_userstack(struct ring_buffer *buffer, unsigned long flags, int pc)
1300
{
1301
	struct ftrace_event_call *call = &event_user_stack;
1302
	struct ring_buffer_event *event;
1303 1304 1305 1306 1307 1308
	struct userstack_entry *entry;
	struct stack_trace trace;

	if (!(trace_flags & TRACE_ITER_USERSTACKTRACE))
		return;

1309 1310 1311 1312 1313 1314
	/*
	 * NMIs can not handle page faults, even with fix ups.
	 * The save user stack can (and often does) fault.
	 */
	if (unlikely(in_nmi()))
		return;
1315

1316
	event = trace_buffer_lock_reserve(buffer, TRACE_USER_STACK,
1317
					  sizeof(*entry), flags, pc);
1318 1319 1320 1321
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);

1322
	entry->tgid		= current->tgid;
1323 1324 1325 1326 1327 1328 1329 1330
	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);
1331 1332
	if (!filter_check_discard(call, entry, buffer, event))
		ring_buffer_unlock_commit(buffer, event);
1333 1334
}

1335 1336
#ifdef UNUSED
static void __trace_userstack(struct trace_array *tr, unsigned long flags)
1337
{
1338
	ftrace_trace_userstack(tr, flags, preempt_count());
1339
}
1340
#endif /* UNUSED */
1341

1342 1343
#endif /* CONFIG_STACKTRACE */

1344
static void
1345
ftrace_trace_special(void *__tr,
1346 1347
		     unsigned long arg1, unsigned long arg2, unsigned long arg3,
		     int pc)
I
Ingo Molnar 已提交
1348
{
1349
	struct ftrace_event_call *call = &event_special;
1350
	struct ring_buffer_event *event;
I
Ingo Molnar 已提交
1351
	struct trace_array *tr = __tr;
1352
	struct ring_buffer *buffer = tr->buffer;
1353
	struct special_entry *entry;
I
Ingo Molnar 已提交
1354

1355
	event = trace_buffer_lock_reserve(buffer, TRACE_SPECIAL,
1356
					  sizeof(*entry), 0, pc);
1357 1358 1359
	if (!event)
		return;
	entry	= ring_buffer_event_data(event);
1360 1361 1362
	entry->arg1			= arg1;
	entry->arg2			= arg2;
	entry->arg3			= arg3;
1363 1364 1365

	if (!filter_check_discard(call, entry, buffer, event))
		trace_buffer_unlock_commit(buffer, event, 0, pc);
I
Ingo Molnar 已提交
1366 1367
}

1368 1369 1370 1371
void
__trace_special(void *__tr, void *__data,
		unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
1372
	ftrace_trace_special(__tr, arg1, arg2, arg3, preempt_count());
1373 1374
}

1375 1376 1377 1378 1379
void
ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1380
	unsigned long flags;
1381
	int cpu;
1382
	int pc;
1383

S
Steven Rostedt 已提交
1384
	if (tracing_disabled)
1385 1386
		return;

1387
	pc = preempt_count();
1388
	local_irq_save(flags);
1389 1390 1391
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

1392
	if (likely(atomic_inc_return(&data->disabled) == 1))
1393
		ftrace_trace_special(tr, arg1, arg2, arg3, pc);
1394

1395 1396
	atomic_dec(&data->disabled);
	local_irq_restore(flags);
1397 1398
}

1399
/**
1400
 * trace_vbprintk - write binary msg to tracing buffer
1401 1402
 *
 */
1403
int trace_vbprintk(unsigned long ip, const char *fmt, va_list args)
1404
{
1405
	static arch_spinlock_t trace_buf_lock =
1406
		(arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED;
1407 1408
	static u32 trace_buf[TRACE_BUF_SIZE];

1409
	struct ftrace_event_call *call = &event_bprint;
1410
	struct ring_buffer_event *event;
1411
	struct ring_buffer *buffer;
1412 1413
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
1414
	struct bprint_entry *entry;
1415
	unsigned long flags;
1416
	int disable;
1417 1418 1419 1420 1421 1422 1423 1424 1425
	int cpu, len = 0, size, pc;

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

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

	pc = preempt_count();
1426
	preempt_disable_notrace();
1427 1428 1429
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];

1430 1431
	disable = atomic_inc_return(&data->disabled);
	if (unlikely(disable != 1))
1432 1433
		goto out;

1434 1435
	/* Lockdep uses trace_printk for lock tracing */
	local_irq_save(flags);
1436
	arch_spin_lock(&trace_buf_lock);
1437 1438 1439 1440 1441 1442
	len = vbin_printf(trace_buf, TRACE_BUF_SIZE, fmt, args);

	if (len > TRACE_BUF_SIZE || len < 0)
		goto out_unlock;

	size = sizeof(*entry) + sizeof(u32) * len;
1443 1444 1445
	buffer = tr->buffer;
	event = trace_buffer_lock_reserve(buffer, TRACE_BPRINT, size,
					  flags, pc);
1446 1447 1448 1449 1450 1451 1452
	if (!event)
		goto out_unlock;
	entry = ring_buffer_event_data(event);
	entry->ip			= ip;
	entry->fmt			= fmt;

	memcpy(entry->buf, trace_buf, sizeof(u32) * len);
1453
	if (!filter_check_discard(call, entry, buffer, event)) {
1454
		ring_buffer_unlock_commit(buffer, event);
1455 1456
		ftrace_trace_stack(buffer, flags, 6, pc);
	}
1457 1458

out_unlock:
1459
	arch_spin_unlock(&trace_buf_lock);
1460
	local_irq_restore(flags);
1461 1462

out:
1463
	atomic_dec_return(&data->disabled);
1464
	preempt_enable_notrace();
1465 1466 1467 1468
	unpause_graph_tracing();

	return len;
}
1469 1470
EXPORT_SYMBOL_GPL(trace_vbprintk);

1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487
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)
1488
{
1489
	static arch_spinlock_t trace_buf_lock = __ARCH_SPIN_LOCK_UNLOCKED;
1490 1491
	static char trace_buf[TRACE_BUF_SIZE];

1492
	struct ftrace_event_call *call = &event_print;
1493
	struct ring_buffer_event *event;
1494
	struct ring_buffer *buffer;
1495 1496 1497 1498
	struct trace_array_cpu *data;
	int cpu, len = 0, size, pc;
	struct print_entry *entry;
	unsigned long irq_flags;
1499
	int disable;
1500 1501 1502 1503 1504 1505 1506 1507 1508

	if (tracing_disabled || tracing_selftest_running)
		return 0;

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

1509 1510
	disable = atomic_inc_return(&data->disabled);
	if (unlikely(disable != 1))
1511 1512 1513 1514
		goto out;

	pause_graph_tracing();
	raw_local_irq_save(irq_flags);
1515
	arch_spin_lock(&trace_buf_lock);
1516
	len = vsnprintf(trace_buf, TRACE_BUF_SIZE, fmt, args);
1517 1518

	size = sizeof(*entry) + len + 1;
1519 1520 1521
	buffer = tr->buffer;
	event = trace_buffer_lock_reserve(buffer, TRACE_PRINT, size,
					  irq_flags, pc);
1522 1523 1524
	if (!event)
		goto out_unlock;
	entry = ring_buffer_event_data(event);
C
Carsten Emde 已提交
1525
	entry->ip = ip;
1526 1527

	memcpy(&entry->buf, trace_buf, len);
C
Carsten Emde 已提交
1528
	entry->buf[len] = '\0';
1529
	if (!filter_check_discard(call, entry, buffer, event)) {
1530
		ring_buffer_unlock_commit(buffer, event);
1531 1532
		ftrace_trace_stack(buffer, irq_flags, 6, pc);
	}
1533 1534

 out_unlock:
1535
	arch_spin_unlock(&trace_buf_lock);
1536 1537 1538
	raw_local_irq_restore(irq_flags);
	unpause_graph_tracing();
 out:
1539
	atomic_dec_return(&data->disabled);
1540 1541 1542 1543
	preempt_enable_notrace();

	return len;
}
1544 1545 1546

int trace_vprintk(unsigned long ip, const char *fmt, va_list args)
{
S
Steven Rostedt 已提交
1547
	return trace_array_vprintk(&global_trace, ip, fmt, args);
1548
}
1549 1550
EXPORT_SYMBOL_GPL(trace_vprintk);

1551 1552
enum trace_file_type {
	TRACE_FILE_LAT_FMT	= 1,
1553
	TRACE_FILE_ANNOTATE	= 2,
1554 1555
};

1556
static void trace_iterator_increment(struct trace_iterator *iter)
S
Steven Rostedt 已提交
1557
{
1558 1559 1560
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();

S
Steven Rostedt 已提交
1561
	iter->idx++;
1562 1563 1564 1565
	if (iter->buffer_iter[iter->cpu])
		ring_buffer_read(iter->buffer_iter[iter->cpu], NULL);

	ftrace_enable_cpu();
S
Steven Rostedt 已提交
1566 1567
}

I
Ingo Molnar 已提交
1568
static struct trace_entry *
1569 1570
peek_next_entry(struct trace_iterator *iter, int cpu, u64 *ts,
		unsigned long *lost_events)
1571
{
1572 1573
	struct ring_buffer_event *event;
	struct ring_buffer_iter *buf_iter = iter->buffer_iter[cpu];
1574

1575 1576 1577 1578 1579 1580
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();

	if (buf_iter)
		event = ring_buffer_iter_peek(buf_iter, ts);
	else
1581 1582
		event = ring_buffer_peek(iter->tr->buffer, cpu, ts,
					 lost_events);
1583 1584 1585

	ftrace_enable_cpu();

1586
	return event ? ring_buffer_event_data(event) : NULL;
1587
}
1588

1589
static struct trace_entry *
1590 1591
__find_next_entry(struct trace_iterator *iter, int *ent_cpu,
		  unsigned long *missing_events, u64 *ent_ts)
1592
{
1593
	struct ring_buffer *buffer = iter->tr->buffer;
1594
	struct trace_entry *ent, *next = NULL;
1595
	unsigned long lost_events = 0, next_lost = 0;
1596
	int cpu_file = iter->cpu_file;
1597
	u64 next_ts = 0, ts;
1598 1599 1600
	int next_cpu = -1;
	int cpu;

1601 1602 1603 1604 1605 1606 1607
	/*
	 * 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;
1608
		ent = peek_next_entry(iter, cpu_file, ent_ts, missing_events);
1609 1610 1611 1612 1613 1614
		if (ent_cpu)
			*ent_cpu = cpu_file;

		return ent;
	}

1615
	for_each_tracing_cpu(cpu) {
1616

1617 1618
		if (ring_buffer_empty_cpu(buffer, cpu))
			continue;
1619

1620
		ent = peek_next_entry(iter, cpu, &ts, &lost_events);
1621

I
Ingo Molnar 已提交
1622 1623 1624
		/*
		 * Pick the entry with the smallest timestamp:
		 */
1625
		if (ent && (!next || ts < next_ts)) {
1626 1627
			next = ent;
			next_cpu = cpu;
1628
			next_ts = ts;
1629
			next_lost = lost_events;
1630 1631 1632 1633 1634 1635
		}
	}

	if (ent_cpu)
		*ent_cpu = next_cpu;

1636 1637 1638
	if (ent_ts)
		*ent_ts = next_ts;

1639 1640 1641
	if (missing_events)
		*missing_events = next_lost;

1642 1643 1644
	return next;
}

1645
/* Find the next real entry, without updating the iterator itself */
1646 1647
struct trace_entry *trace_find_next_entry(struct trace_iterator *iter,
					  int *ent_cpu, u64 *ent_ts)
1648
{
1649
	return __find_next_entry(iter, ent_cpu, NULL, ent_ts);
1650 1651 1652 1653 1654
}

/* Find the next real entry, and increment the iterator to the next entry */
static void *find_next_entry_inc(struct trace_iterator *iter)
{
1655 1656
	iter->ent = __find_next_entry(iter, &iter->cpu,
				      &iter->lost_events, &iter->ts);
1657

1658
	if (iter->ent)
1659
		trace_iterator_increment(iter);
1660

1661
	return iter->ent ? iter : NULL;
1662
}
1663

I
Ingo Molnar 已提交
1664
static void trace_consume(struct trace_iterator *iter)
1665
{
1666 1667
	/* Don't allow ftrace to trace into the ring buffers */
	ftrace_disable_cpu();
1668 1669
	ring_buffer_consume(iter->tr->buffer, iter->cpu, &iter->ts,
			    &iter->lost_events);
1670
	ftrace_enable_cpu();
1671 1672
}

I
Ingo Molnar 已提交
1673
static void *s_next(struct seq_file *m, void *v, loff_t *pos)
1674 1675 1676
{
	struct trace_iterator *iter = m->private;
	int i = (int)*pos;
I
Ingo Molnar 已提交
1677
	void *ent;
1678

1679 1680
	WARN_ON_ONCE(iter->leftover);

1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699
	(*pos)++;

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

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

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

	iter->pos = *pos;

	return ent;
}

1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730
static void tracing_iter_reset(struct trace_iterator *iter, int cpu)
{
	struct trace_array *tr = iter->tr;
	struct ring_buffer_event *event;
	struct ring_buffer_iter *buf_iter;
	unsigned long entries = 0;
	u64 ts;

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

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

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

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

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

1731 1732 1733 1734
/*
 * The current tracer is copied to avoid a global locking
 * all around.
 */
1735 1736 1737
static void *s_start(struct seq_file *m, loff_t *pos)
{
	struct trace_iterator *iter = m->private;
1738
	static struct tracer *old_tracer;
1739
	int cpu_file = iter->cpu_file;
1740 1741
	void *p = NULL;
	loff_t l = 0;
1742
	int cpu;
1743

1744
	/* copy the tracer to avoid using a global lock all around */
1745
	mutex_lock(&trace_types_lock);
1746 1747 1748
	if (unlikely(old_tracer != current_trace && current_trace)) {
		old_tracer = current_trace;
		*iter->trace = *current_trace;
1749
	}
1750
	mutex_unlock(&trace_types_lock);
1751 1752 1753 1754 1755 1756 1757 1758

	atomic_inc(&trace_record_cmdline_disabled);

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

1759 1760
		ftrace_disable_cpu();

1761 1762
		if (cpu_file == TRACE_PIPE_ALL_CPU) {
			for_each_tracing_cpu(cpu)
1763
				tracing_iter_reset(iter, cpu);
1764
		} else
1765
			tracing_iter_reset(iter, cpu_file);
1766

1767 1768
		ftrace_enable_cpu();

1769
		iter->leftover = 0;
1770 1771 1772 1773
		for (p = iter; p && l < *pos; p = s_next(m, p, &l))
			;

	} else {
1774 1775 1776 1777 1778 1779 1780 1781 1782 1783
		/*
		 * 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);
		}
1784 1785
	}

1786
	trace_event_read_lock();
1787
	trace_access_lock(cpu_file);
1788 1789 1790 1791 1792
	return p;
}

static void s_stop(struct seq_file *m, void *p)
{
1793 1794
	struct trace_iterator *iter = m->private;

1795
	atomic_dec(&trace_record_cmdline_disabled);
1796
	trace_access_unlock(iter->cpu_file);
1797
	trace_event_read_unlock();
1798 1799
}

I
Ingo Molnar 已提交
1800
static void print_lat_help_header(struct seq_file *m)
1801
{
1802 1803 1804 1805 1806
	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");
1807 1808 1809 1810
	seq_puts(m, "#                |||| /_--=> lock-depth       \n");
	seq_puts(m, "#                |||||/     delay             \n");
	seq_puts(m, "#  cmd     pid   |||||| time  |   caller      \n");
	seq_puts(m, "#     \\   /      ||||||   \\   |   /           \n");
1811 1812
}

I
Ingo Molnar 已提交
1813
static void print_func_help_header(struct seq_file *m)
1814
{
1815 1816
	seq_puts(m, "#           TASK-PID    CPU#    TIMESTAMP  FUNCTION\n");
	seq_puts(m, "#              | |       |          |         |\n");
1817 1818 1819
}


1820
void
1821 1822 1823 1824 1825 1826
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;
1827 1828 1829
	unsigned long entries = 0;
	unsigned long total = 0;
	unsigned long count;
1830
	const char *name = "preemption";
1831
	int cpu;
1832 1833 1834 1835

	if (type)
		name = type->name;

1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852

	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;
	}
1853

1854
	seq_printf(m, "# %s latency trace v1.1.5 on %s\n",
1855
		   name, UTS_RELEASE);
1856
	seq_puts(m, "# -----------------------------------"
1857
		 "---------------------------------\n");
1858
	seq_printf(m, "# latency: %lu us, #%lu/%lu, CPU#%d |"
1859
		   " (M:%s VP:%d, KP:%d, SP:%d HP:%d",
S
Steven Rostedt 已提交
1860
		   nsecs_to_usecs(data->saved_latency),
1861
		   entries,
1862
		   total,
1863 1864 1865 1866 1867
		   tr->cpu,
#if defined(CONFIG_PREEMPT_NONE)
		   "server",
#elif defined(CONFIG_PREEMPT_VOLUNTARY)
		   "desktop",
1868
#elif defined(CONFIG_PREEMPT)
1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879
		   "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
1880 1881
	seq_puts(m, "#    -----------------\n");
	seq_printf(m, "#    | task: %.16s-%d "
1882 1883 1884
		   "(uid:%d nice:%ld policy:%ld rt_prio:%ld)\n",
		   data->comm, data->pid, data->uid, data->nice,
		   data->policy, data->rt_priority);
1885
	seq_puts(m, "#    -----------------\n");
1886 1887

	if (data->critical_start) {
1888
		seq_puts(m, "#  => started at: ");
S
Steven Rostedt 已提交
1889 1890
		seq_print_ip_sym(&iter->seq, data->critical_start, sym_flags);
		trace_print_seq(m, &iter->seq);
1891
		seq_puts(m, "\n#  => ended at:   ");
S
Steven Rostedt 已提交
1892 1893
		seq_print_ip_sym(&iter->seq, data->critical_end, sym_flags);
		trace_print_seq(m, &iter->seq);
1894
		seq_puts(m, "\n#\n");
1895 1896
	}

1897
	seq_puts(m, "#\n");
1898 1899
}

1900 1901 1902 1903
static void test_cpu_buff_start(struct trace_iterator *iter)
{
	struct trace_seq *s = &iter->seq;

1904 1905 1906 1907 1908 1909
	if (!(trace_flags & TRACE_ITER_ANNOTATE))
		return;

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

1910
	if (cpumask_test_cpu(iter->cpu, iter->started))
1911 1912
		return;

1913 1914 1915
	if (iter->tr->data[iter->cpu]->skipped_entries)
		return;

1916
	cpumask_set_cpu(iter->cpu, iter->started);
1917 1918 1919 1920 1921

	/* 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);
1922 1923
}

1924
static enum print_line_t print_trace_fmt(struct trace_iterator *iter)
1925
{
S
Steven Rostedt 已提交
1926
	struct trace_seq *s = &iter->seq;
1927
	unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK);
I
Ingo Molnar 已提交
1928
	struct trace_entry *entry;
1929
	struct trace_event *event;
1930

I
Ingo Molnar 已提交
1931
	entry = iter->ent;
1932

1933 1934
	test_cpu_buff_start(iter);

1935
	event = ftrace_find_event(entry->type);
1936

1937
	if (trace_flags & TRACE_ITER_CONTEXT_INFO) {
1938 1939 1940 1941 1942 1943 1944
		if (iter->iter_flags & TRACE_FILE_LAT_FMT) {
			if (!trace_print_lat_context(iter))
				goto partial;
		} else {
			if (!trace_print_context(iter))
				goto partial;
		}
1945
	}
1946

1947
	if (event)
1948
		return event->funcs->trace(iter, sym_flags, event);
1949 1950 1951

	if (!trace_seq_printf(s, "Unknown type %d\n", entry->type))
		goto partial;
1952

1953
	return TRACE_TYPE_HANDLED;
1954 1955
partial:
	return TRACE_TYPE_PARTIAL_LINE;
1956 1957
}

1958
static enum print_line_t print_raw_fmt(struct trace_iterator *iter)
I
Ingo Molnar 已提交
1959 1960 1961
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *entry;
1962
	struct trace_event *event;
I
Ingo Molnar 已提交
1963 1964

	entry = iter->ent;
1965

1966
	if (trace_flags & TRACE_ITER_CONTEXT_INFO) {
1967 1968 1969
		if (!trace_seq_printf(s, "%d %d %llu ",
				      entry->pid, iter->cpu, iter->ts))
			goto partial;
1970
	}
I
Ingo Molnar 已提交
1971

1972
	event = ftrace_find_event(entry->type);
1973
	if (event)
1974
		return event->funcs->raw(iter, 0, event);
1975 1976 1977

	if (!trace_seq_printf(s, "%d ?\n", entry->type))
		goto partial;
1978

1979
	return TRACE_TYPE_HANDLED;
1980 1981
partial:
	return TRACE_TYPE_PARTIAL_LINE;
I
Ingo Molnar 已提交
1982 1983
}

1984
static enum print_line_t print_hex_fmt(struct trace_iterator *iter)
1985 1986 1987 1988
{
	struct trace_seq *s = &iter->seq;
	unsigned char newline = '\n';
	struct trace_entry *entry;
1989
	struct trace_event *event;
1990 1991

	entry = iter->ent;
1992

1993 1994 1995 1996 1997
	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);
	}
1998

1999
	event = ftrace_find_event(entry->type);
2000
	if (event) {
2001
		enum print_line_t ret = event->funcs->hex(iter, 0, event);
2002 2003 2004
		if (ret != TRACE_TYPE_HANDLED)
			return ret;
	}
S
Steven Rostedt 已提交
2005

2006 2007
	SEQ_PUT_FIELD_RET(s, newline);

2008
	return TRACE_TYPE_HANDLED;
2009 2010
}

2011
static enum print_line_t print_bin_fmt(struct trace_iterator *iter)
I
Ingo Molnar 已提交
2012 2013 2014
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *entry;
2015
	struct trace_event *event;
I
Ingo Molnar 已提交
2016 2017

	entry = iter->ent;
2018

2019 2020
	if (trace_flags & TRACE_ITER_CONTEXT_INFO) {
		SEQ_PUT_FIELD_RET(s, entry->pid);
2021
		SEQ_PUT_FIELD_RET(s, iter->cpu);
2022 2023
		SEQ_PUT_FIELD_RET(s, iter->ts);
	}
I
Ingo Molnar 已提交
2024

2025
	event = ftrace_find_event(entry->type);
2026 2027
	return event ? event->funcs->binary(iter, 0, event) :
		TRACE_TYPE_HANDLED;
I
Ingo Molnar 已提交
2028 2029
}

2030
int trace_empty(struct trace_iterator *iter)
2031 2032 2033
{
	int cpu;

2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046
	/* If we are looking at one CPU buffer, only check that one */
	if (iter->cpu_file != TRACE_PIPE_ALL_CPU) {
		cpu = iter->cpu_file;
		if (iter->buffer_iter[cpu]) {
			if (!ring_buffer_iter_empty(iter->buffer_iter[cpu]))
				return 0;
		} else {
			if (!ring_buffer_empty_cpu(iter->tr->buffer, cpu))
				return 0;
		}
		return 1;
	}

2047
	for_each_tracing_cpu(cpu) {
2048 2049 2050 2051 2052 2053 2054
		if (iter->buffer_iter[cpu]) {
			if (!ring_buffer_iter_empty(iter->buffer_iter[cpu]))
				return 0;
		} else {
			if (!ring_buffer_empty_cpu(iter->tr->buffer, cpu))
				return 0;
		}
2055
	}
2056

2057
	return 1;
2058 2059
}

2060
/*  Called with trace_event_read_lock() held. */
2061
static enum print_line_t print_trace_line(struct trace_iterator *iter)
I
Ingo Molnar 已提交
2062
{
2063 2064
	enum print_line_t ret;

2065 2066 2067 2068
	if (iter->lost_events)
		trace_seq_printf(&iter->seq, "CPU:%d [LOST %lu EVENTS]\n",
				 iter->cpu, iter->lost_events);

2069 2070 2071 2072 2073
	if (iter->trace && iter->trace->print_line) {
		ret = iter->trace->print_line(iter);
		if (ret != TRACE_TYPE_UNHANDLED)
			return ret;
	}
2074

2075 2076 2077
	if (iter->ent->type == TRACE_BPRINT &&
			trace_flags & TRACE_ITER_PRINTK &&
			trace_flags & TRACE_ITER_PRINTK_MSGONLY)
2078
		return trace_print_bprintk_msg_only(iter);
2079

2080 2081 2082
	if (iter->ent->type == TRACE_PRINT &&
			trace_flags & TRACE_ITER_PRINTK &&
			trace_flags & TRACE_ITER_PRINTK_MSGONLY)
2083
		return trace_print_printk_msg_only(iter);
2084

I
Ingo Molnar 已提交
2085 2086 2087
	if (trace_flags & TRACE_ITER_BIN)
		return print_bin_fmt(iter);

2088 2089 2090
	if (trace_flags & TRACE_ITER_HEX)
		return print_hex_fmt(iter);

I
Ingo Molnar 已提交
2091 2092 2093 2094 2095 2096
	if (trace_flags & TRACE_ITER_RAW)
		return print_raw_fmt(iter);

	return print_trace_fmt(iter);
}

2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113
void trace_default_header(struct seq_file *m)
{
	struct trace_iterator *iter = m->private;

	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 {
		if (!(trace_flags & TRACE_ITER_VERBOSE))
			print_func_help_header(m);
	}
}

2114 2115 2116
static int s_show(struct seq_file *m, void *v)
{
	struct trace_iterator *iter = v;
2117
	int ret;
2118 2119 2120 2121 2122 2123

	if (iter->ent == NULL) {
		if (iter->tr) {
			seq_printf(m, "# tracer: %s\n", iter->trace->name);
			seq_puts(m, "#\n");
		}
2124 2125
		if (iter->trace && iter->trace->print_header)
			iter->trace->print_header(m);
2126 2127 2128
		else
			trace_default_header(m);

2129 2130 2131 2132 2133 2134 2135 2136 2137 2138
	} 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;

2139
	} else {
I
Ingo Molnar 已提交
2140
		print_trace_line(iter);
2141 2142 2143 2144 2145 2146 2147 2148 2149
		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;
2150 2151 2152 2153 2154
	}

	return 0;
}

J
James Morris 已提交
2155
static const struct seq_operations tracer_seq_ops = {
I
Ingo Molnar 已提交
2156 2157 2158 2159
	.start		= s_start,
	.next		= s_next,
	.stop		= s_stop,
	.show		= s_show,
2160 2161
};

I
Ingo Molnar 已提交
2162
static struct trace_iterator *
2163
__tracing_open(struct inode *inode, struct file *file)
2164
{
2165
	long cpu_file = (long) inode->i_private;
2166
	void *fail_ret = ERR_PTR(-ENOMEM);
2167
	struct trace_iterator *iter;
2168
	struct seq_file *m;
2169
	int cpu, ret;
2170

2171 2172
	if (tracing_disabled)
		return ERR_PTR(-ENODEV);
S
Steven Rostedt 已提交
2173

2174
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
2175 2176
	if (!iter)
		return ERR_PTR(-ENOMEM);
2177

2178 2179 2180 2181
	/*
	 * We make a copy of the current tracer to avoid concurrent
	 * changes on it while we are reading.
	 */
2182
	mutex_lock(&trace_types_lock);
2183
	iter->trace = kzalloc(sizeof(*iter->trace), GFP_KERNEL);
2184
	if (!iter->trace)
2185
		goto fail;
2186

2187 2188 2189
	if (current_trace)
		*iter->trace = *current_trace;

2190
	if (!zalloc_cpumask_var(&iter->started, GFP_KERNEL))
2191 2192
		goto fail;

2193 2194 2195
	if (current_trace && current_trace->print_max)
		iter->tr = &max_tr;
	else
2196
		iter->tr = &global_trace;
2197
	iter->pos = -1;
2198
	mutex_init(&iter->mutex);
2199
	iter->cpu_file = cpu_file;
2200

2201 2202
	/* Notify the tracer early; before we stop tracing. */
	if (iter->trace && iter->trace->open)
2203
		iter->trace->open(iter);
2204

2205 2206 2207 2208
	/* Annotate start of buffers if we had overruns */
	if (ring_buffer_overruns(iter->tr->buffer))
		iter->iter_flags |= TRACE_FILE_ANNOTATE;

2209 2210 2211
	/* stop the trace while dumping */
	tracing_stop();

2212 2213 2214
	if (iter->cpu_file == TRACE_PIPE_ALL_CPU) {
		for_each_tracing_cpu(cpu) {
			iter->buffer_iter[cpu] =
2215 2216 2217 2218 2219
				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]);
2220
			tracing_iter_reset(iter, cpu);
2221 2222 2223
		}
	} else {
		cpu = iter->cpu_file;
2224
		iter->buffer_iter[cpu] =
2225 2226 2227
			ring_buffer_read_prepare(iter->tr->buffer, cpu);
		ring_buffer_read_prepare_sync();
		ring_buffer_read_start(iter->buffer_iter[cpu]);
2228
		tracing_iter_reset(iter, cpu);
2229 2230
	}

2231 2232 2233
	ret = seq_open(file, &tracer_seq_ops);
	if (ret < 0) {
		fail_ret = ERR_PTR(ret);
2234
		goto fail_buffer;
2235
	}
2236

2237 2238
	m = file->private_data;
	m->private = iter;
2239 2240 2241 2242

	mutex_unlock(&trace_types_lock);

	return iter;
2243 2244 2245 2246 2247 2248

 fail_buffer:
	for_each_tracing_cpu(cpu) {
		if (iter->buffer_iter[cpu])
			ring_buffer_read_finish(iter->buffer_iter[cpu]);
	}
2249
	free_cpumask_var(iter->started);
2250
	tracing_start();
2251
 fail:
2252
	mutex_unlock(&trace_types_lock);
2253
	kfree(iter->trace);
2254
	kfree(iter);
2255

2256
	return fail_ret;
2257 2258 2259 2260
}

int tracing_open_generic(struct inode *inode, struct file *filp)
{
S
Steven Rostedt 已提交
2261 2262 2263
	if (tracing_disabled)
		return -ENODEV;

2264 2265 2266 2267
	filp->private_data = inode->i_private;
	return 0;
}

2268
static int tracing_release(struct inode *inode, struct file *file)
2269 2270
{
	struct seq_file *m = (struct seq_file *)file->private_data;
2271
	struct trace_iterator *iter;
2272
	int cpu;
2273

2274 2275 2276 2277 2278
	if (!(file->f_mode & FMODE_READ))
		return 0;

	iter = m->private;

2279
	mutex_lock(&trace_types_lock);
2280 2281 2282 2283 2284
	for_each_tracing_cpu(cpu) {
		if (iter->buffer_iter[cpu])
			ring_buffer_read_finish(iter->buffer_iter[cpu]);
	}

2285 2286 2287 2288
	if (iter->trace && iter->trace->close)
		iter->trace->close(iter);

	/* reenable tracing if it was previously enabled */
2289
	tracing_start();
2290 2291 2292
	mutex_unlock(&trace_types_lock);

	seq_release(inode, file);
2293
	mutex_destroy(&iter->mutex);
2294
	free_cpumask_var(iter->started);
2295
	kfree(iter->trace);
2296 2297 2298 2299 2300 2301
	kfree(iter);
	return 0;
}

static int tracing_open(struct inode *inode, struct file *file)
{
2302 2303
	struct trace_iterator *iter;
	int ret = 0;
2304

2305 2306
	/* If this file was open for write, then erase contents */
	if ((file->f_mode & FMODE_WRITE) &&
2307
	    (file->f_flags & O_TRUNC)) {
2308
		long cpu = (long) inode->i_private;
2309

2310 2311 2312 2313 2314
		if (cpu == TRACE_PIPE_ALL_CPU)
			tracing_reset_online_cpus(&global_trace);
		else
			tracing_reset(&global_trace, cpu);
	}
2315

2316 2317 2318 2319 2320 2321 2322
	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;
	}
2323 2324 2325
	return ret;
}

I
Ingo Molnar 已提交
2326
static void *
2327 2328
t_next(struct seq_file *m, void *v, loff_t *pos)
{
L
Li Zefan 已提交
2329
	struct tracer *t = v;
2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340

	(*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
2341
	struct tracer *t;
2342 2343 2344
	loff_t l = 0;

	mutex_lock(&trace_types_lock);
L
Li Zefan 已提交
2345
	for (t = trace_types; t && l < *pos; t = t_next(m, t, &l))
2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371
		;

	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
2372
static const struct seq_operations show_traces_seq_ops = {
I
Ingo Molnar 已提交
2373 2374 2375 2376
	.start		= t_start,
	.next		= t_next,
	.stop		= t_stop,
	.show		= t_show,
2377 2378 2379 2380
};

static int show_traces_open(struct inode *inode, struct file *file)
{
S
Steven Rostedt 已提交
2381 2382 2383
	if (tracing_disabled)
		return -ENODEV;

L
Li Zefan 已提交
2384
	return seq_open(file, &show_traces_seq_ops);
2385 2386
}

2387 2388 2389 2390 2391 2392 2393
static ssize_t
tracing_write_stub(struct file *filp, const char __user *ubuf,
		   size_t count, loff_t *ppos)
{
	return count;
}

2394
static const struct file_operations tracing_fops = {
I
Ingo Molnar 已提交
2395 2396
	.open		= tracing_open,
	.read		= seq_read,
2397
	.write		= tracing_write_stub,
I
Ingo Molnar 已提交
2398 2399
	.llseek		= seq_lseek,
	.release	= tracing_release,
2400 2401
};

2402
static const struct file_operations show_traces_fops = {
I
Ingo Molnar 已提交
2403 2404 2405 2406 2407
	.open		= show_traces_open,
	.read		= seq_read,
	.release	= seq_release,
};

2408 2409 2410
/*
 * Only trace on a CPU if the bitmask is set:
 */
2411
static cpumask_var_t tracing_cpumask;
2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424

/*
 * 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
2425 2426 2427 2428
static ssize_t
tracing_cpumask_read(struct file *filp, char __user *ubuf,
		     size_t count, loff_t *ppos)
{
2429
	int len;
I
Ingo Molnar 已提交
2430 2431

	mutex_lock(&tracing_cpumask_update_lock);
2432

2433
	len = cpumask_scnprintf(mask_str, count, tracing_cpumask);
2434 2435 2436 2437 2438 2439 2440 2441
	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
2442 2443 2444 2445 2446 2447 2448 2449 2450
	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)
{
2451
	int err, cpu;
2452 2453 2454 2455
	cpumask_var_t tracing_cpumask_new;

	if (!alloc_cpumask_var(&tracing_cpumask_new, GFP_KERNEL))
		return -ENOMEM;
I
Ingo Molnar 已提交
2456

2457
	err = cpumask_parse_user(ubuf, count, tracing_cpumask_new);
I
Ingo Molnar 已提交
2458
	if (err)
2459 2460
		goto err_unlock;

2461 2462
	mutex_lock(&tracing_cpumask_update_lock);

2463
	local_irq_disable();
2464
	arch_spin_lock(&ftrace_max_lock);
2465
	for_each_tracing_cpu(cpu) {
2466 2467 2468 2469
		/*
		 * Increase/decrease the disabled counter if we are
		 * about to flip a bit in the cpumask:
		 */
2470 2471
		if (cpumask_test_cpu(cpu, tracing_cpumask) &&
				!cpumask_test_cpu(cpu, tracing_cpumask_new)) {
2472 2473
			atomic_inc(&global_trace.data[cpu]->disabled);
		}
2474 2475
		if (!cpumask_test_cpu(cpu, tracing_cpumask) &&
				cpumask_test_cpu(cpu, tracing_cpumask_new)) {
2476 2477 2478
			atomic_dec(&global_trace.data[cpu]->disabled);
		}
	}
2479
	arch_spin_unlock(&ftrace_max_lock);
2480
	local_irq_enable();
2481

2482
	cpumask_copy(tracing_cpumask, tracing_cpumask_new);
2483 2484

	mutex_unlock(&tracing_cpumask_update_lock);
2485
	free_cpumask_var(tracing_cpumask_new);
I
Ingo Molnar 已提交
2486 2487

	return count;
2488 2489

err_unlock:
2490
	free_cpumask_var(tracing_cpumask_new);
2491 2492

	return err;
I
Ingo Molnar 已提交
2493 2494
}

2495
static const struct file_operations tracing_cpumask_fops = {
I
Ingo Molnar 已提交
2496 2497 2498
	.open		= tracing_open_generic,
	.read		= tracing_cpumask_read,
	.write		= tracing_cpumask_write,
2499 2500
};

L
Li Zefan 已提交
2501
static int tracing_trace_options_show(struct seq_file *m, void *v)
2502
{
2503 2504 2505
	struct tracer_opt *trace_opts;
	u32 tracer_flags;
	int i;
2506

2507 2508 2509 2510
	mutex_lock(&trace_types_lock);
	tracer_flags = current_trace->flags->val;
	trace_opts = current_trace->flags->opts;

2511 2512
	for (i = 0; trace_options[i]; i++) {
		if (trace_flags & (1 << i))
L
Li Zefan 已提交
2513
			seq_printf(m, "%s\n", trace_options[i]);
2514
		else
L
Li Zefan 已提交
2515
			seq_printf(m, "no%s\n", trace_options[i]);
2516 2517
	}

2518 2519
	for (i = 0; trace_opts[i].name; i++) {
		if (tracer_flags & trace_opts[i].bit)
L
Li Zefan 已提交
2520
			seq_printf(m, "%s\n", trace_opts[i].name);
2521
		else
L
Li Zefan 已提交
2522
			seq_printf(m, "no%s\n", trace_opts[i].name);
2523
	}
2524
	mutex_unlock(&trace_types_lock);
2525

L
Li Zefan 已提交
2526
	return 0;
2527 2528
}

L
Li Zefan 已提交
2529 2530 2531 2532 2533
static int __set_tracer_option(struct tracer *trace,
			       struct tracer_flags *tracer_flags,
			       struct tracer_opt *opts, int neg)
{
	int ret;
2534

L
Li Zefan 已提交
2535 2536 2537 2538 2539 2540 2541 2542 2543
	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;
2544 2545
}

2546 2547 2548
/* Try to assign a tracer specific option */
static int set_tracer_option(struct tracer *trace, char *cmp, int neg)
{
2549
	struct tracer_flags *tracer_flags = trace->flags;
2550
	struct tracer_opt *opts = NULL;
L
Li Zefan 已提交
2551
	int i;
2552

2553 2554
	for (i = 0; tracer_flags->opts[i].name; i++) {
		opts = &tracer_flags->opts[i];
2555

L
Li Zefan 已提交
2556 2557 2558
		if (strcmp(cmp, opts->name) == 0)
			return __set_tracer_option(trace, trace->flags,
						   opts, neg);
2559 2560
	}

L
Li Zefan 已提交
2561
	return -EINVAL;
2562 2563
}

2564 2565 2566 2567 2568 2569 2570 2571 2572 2573
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;
2574 2575 2576

	if (mask == TRACE_ITER_RECORD_CMD)
		trace_event_enable_cmd_record(enabled);
2577 2578
}

2579
static ssize_t
2580
tracing_trace_options_write(struct file *filp, const char __user *ubuf,
2581 2582 2583
			size_t cnt, loff_t *ppos)
{
	char buf[64];
L
Li Zefan 已提交
2584
	char *cmp;
2585
	int neg = 0;
2586
	int ret;
2587 2588
	int i;

S
Steven Rostedt 已提交
2589 2590
	if (cnt >= sizeof(buf))
		return -EINVAL;
2591 2592 2593 2594 2595

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

	buf[cnt] = 0;
L
Li Zefan 已提交
2596
	cmp = strstrip(buf);
2597

L
Li Zefan 已提交
2598
	if (strncmp(cmp, "no", 2) == 0) {
2599 2600 2601 2602 2603
		neg = 1;
		cmp += 2;
	}

	for (i = 0; trace_options[i]; i++) {
L
Li Zefan 已提交
2604
		if (strcmp(cmp, trace_options[i]) == 0) {
2605
			set_tracer_flags(1 << i, !neg);
2606 2607 2608
			break;
		}
	}
2609 2610 2611

	/* If no option could be set, test the specific tracer options */
	if (!trace_options[i]) {
2612
		mutex_lock(&trace_types_lock);
2613
		ret = set_tracer_option(current_trace, cmp, neg);
2614
		mutex_unlock(&trace_types_lock);
2615 2616 2617
		if (ret)
			return ret;
	}
2618

2619
	*ppos += cnt;
2620 2621 2622 2623

	return cnt;
}

L
Li Zefan 已提交
2624 2625 2626 2627 2628 2629 2630
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);
}

2631
static const struct file_operations tracing_iter_fops = {
L
Li Zefan 已提交
2632 2633 2634 2635
	.open		= tracing_trace_options_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= single_release,
2636
	.write		= tracing_trace_options_write,
2637 2638
};

I
Ingo Molnar 已提交
2639 2640
static const char readme_msg[] =
	"tracing mini-HOWTO:\n\n"
2641 2642
	"# mount -t debugfs nodev /sys/kernel/debug\n\n"
	"# cat /sys/kernel/debug/tracing/available_tracers\n"
2643
	"wakeup preemptirqsoff preemptoff irqsoff function sched_switch nop\n\n"
2644
	"# cat /sys/kernel/debug/tracing/current_tracer\n"
2645
	"nop\n"
2646 2647
	"# echo sched_switch > /sys/kernel/debug/tracing/current_tracer\n"
	"# cat /sys/kernel/debug/tracing/current_tracer\n"
I
Ingo Molnar 已提交
2648
	"sched_switch\n"
2649
	"# cat /sys/kernel/debug/tracing/trace_options\n"
I
Ingo Molnar 已提交
2650
	"noprint-parent nosym-offset nosym-addr noverbose\n"
2651 2652 2653 2654
	"# echo print-parent > /sys/kernel/debug/tracing/trace_options\n"
	"# echo 1 > /sys/kernel/debug/tracing/tracing_enabled\n"
	"# cat /sys/kernel/debug/tracing/trace > /tmp/trace.txt\n"
	"# echo 0 > /sys/kernel/debug/tracing/tracing_enabled\n"
I
Ingo Molnar 已提交
2655 2656 2657 2658 2659 2660 2661 2662 2663 2664
;

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));
}

2665
static const struct file_operations tracing_readme_fops = {
I
Ingo Molnar 已提交
2666 2667
	.open		= tracing_open_generic,
	.read		= tracing_readme_read,
I
Ingo Molnar 已提交
2668 2669
};

2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719
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,
};

2720 2721 2722 2723 2724 2725 2726
static ssize_t
tracing_ctrl_read(struct file *filp, char __user *ubuf,
		  size_t cnt, loff_t *ppos)
{
	char buf[64];
	int r;

2727
	r = sprintf(buf, "%u\n", tracer_enabled);
I
Ingo Molnar 已提交
2728
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
2729 2730 2731 2732 2733 2734 2735 2736
}

static ssize_t
tracing_ctrl_write(struct file *filp, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
	struct trace_array *tr = filp->private_data;
	char buf[64];
2737
	unsigned long val;
2738
	int ret;
2739

S
Steven Rostedt 已提交
2740 2741
	if (cnt >= sizeof(buf))
		return -EINVAL;
2742 2743 2744 2745 2746 2747

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

	buf[cnt] = 0;

2748 2749 2750
	ret = strict_strtoul(buf, 10, &val);
	if (ret < 0)
		return ret;
2751 2752 2753 2754

	val = !!val;

	mutex_lock(&trace_types_lock);
2755 2756
	if (tracer_enabled ^ val) {
		if (val) {
2757
			tracer_enabled = 1;
2758 2759 2760 2761
			if (current_trace->start)
				current_trace->start(tr);
			tracing_start();
		} else {
2762
			tracer_enabled = 0;
2763 2764 2765 2766
			tracing_stop();
			if (current_trace->stop)
				current_trace->stop(tr);
		}
2767 2768 2769
	}
	mutex_unlock(&trace_types_lock);

2770
	*ppos += cnt;
2771 2772 2773 2774 2775 2776 2777 2778

	return cnt;
}

static ssize_t
tracing_set_trace_read(struct file *filp, char __user *ubuf,
		       size_t cnt, loff_t *ppos)
{
L
Li Zefan 已提交
2779
	char buf[MAX_TRACER_SIZE+2];
2780 2781 2782 2783 2784 2785 2786 2787 2788
	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
2789
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
2790 2791
}

2792 2793 2794 2795 2796 2797
int tracer_init(struct tracer *t, struct trace_array *tr)
{
	tracing_reset_online_cpus(tr);
	return t->init(tr);
}

2798 2799 2800 2801 2802 2803
static int tracing_resize_ring_buffer(unsigned long size)
{
	int ret;

	/*
	 * If kernel or user changes the size of the ring buffer
2804 2805
	 * we use the size that was given, and we can forget about
	 * expanding it later.
2806 2807 2808 2809 2810 2811 2812
	 */
	ring_buffer_expanded = 1;

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

2813 2814 2815
	if (!current_trace->use_max_tr)
		goto out;

2816 2817 2818 2819 2820 2821 2822
	ret = ring_buffer_resize(max_tr.buffer, size);
	if (ret < 0) {
		int r;

		r = ring_buffer_resize(global_trace.buffer,
				       global_trace.entries);
		if (r < 0) {
2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836
			/*
			 * 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.
			 */
2837 2838 2839 2840 2841 2842
			WARN_ON(1);
			tracing_disabled = 1;
		}
		return ret;
	}

2843 2844
	max_tr.entries = size;
 out:
2845 2846 2847 2848 2849
	global_trace.entries = size;

	return ret;
}

2850

2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864
/**
 * 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;

2865
	mutex_lock(&trace_types_lock);
2866 2867
	if (!ring_buffer_expanded)
		ret = tracing_resize_ring_buffer(trace_buf_size);
2868
	mutex_unlock(&trace_types_lock);
2869 2870 2871 2872

	return ret;
}

2873 2874 2875 2876 2877 2878 2879 2880
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);

2881
static int tracing_set_tracer(const char *buf)
2882
{
2883
	static struct trace_option_dentry *topts;
2884 2885
	struct trace_array *tr = &global_trace;
	struct tracer *t;
2886
	int ret = 0;
2887

2888 2889
	mutex_lock(&trace_types_lock);

2890 2891 2892
	if (!ring_buffer_expanded) {
		ret = tracing_resize_ring_buffer(trace_buf_size);
		if (ret < 0)
2893
			goto out;
2894 2895 2896
		ret = 0;
	}

2897 2898 2899 2900
	for (t = trace_types; t; t = t->next) {
		if (strcmp(t->name, buf) == 0)
			break;
	}
2901 2902 2903 2904 2905
	if (!t) {
		ret = -EINVAL;
		goto out;
	}
	if (t == current_trace)
2906 2907
		goto out;

2908
	trace_branch_disable();
2909 2910
	if (current_trace && current_trace->reset)
		current_trace->reset(tr);
2911 2912 2913 2914 2915 2916 2917 2918 2919
	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.
		 */
		ring_buffer_resize(max_tr.buffer, 1);
		max_tr.entries = 1;
	}
2920 2921
	destroy_trace_option_files(topts);

2922
	current_trace = t;
2923 2924

	topts = create_trace_option_files(current_trace);
2925 2926 2927 2928 2929 2930
	if (current_trace->use_max_tr) {
		ret = ring_buffer_resize(max_tr.buffer, global_trace.entries);
		if (ret < 0)
			goto out;
		max_tr.entries = global_trace.entries;
	}
2931

2932
	if (t->init) {
2933
		ret = tracer_init(t, tr);
2934 2935 2936
		if (ret)
			goto out;
	}
2937

2938
	trace_branch_enable(tr);
2939 2940 2941
 out:
	mutex_unlock(&trace_types_lock);

2942 2943 2944 2945 2946 2947 2948
	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
2949
	char buf[MAX_TRACER_SIZE+1];
2950 2951
	int i;
	size_t ret;
2952 2953 2954
	int err;

	ret = cnt;
2955

L
Li Zefan 已提交
2956 2957
	if (cnt > MAX_TRACER_SIZE)
		cnt = MAX_TRACER_SIZE;
2958 2959 2960 2961 2962 2963 2964 2965 2966 2967

	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;

2968 2969 2970
	err = tracing_set_tracer(buf);
	if (err)
		return err;
2971

2972
	*ppos += ret;
2973

2974
	return ret;
2975 2976 2977 2978 2979 2980 2981 2982 2983 2984
}

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
2985
	r = snprintf(buf, sizeof(buf), "%ld\n",
2986
		     *ptr == (unsigned long)-1 ? -1 : nsecs_to_usecs(*ptr));
S
Steven Rostedt 已提交
2987 2988
	if (r > sizeof(buf))
		r = sizeof(buf);
I
Ingo Molnar 已提交
2989
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
2990 2991 2992 2993 2994 2995
}

static ssize_t
tracing_max_lat_write(struct file *filp, const char __user *ubuf,
		      size_t cnt, loff_t *ppos)
{
2996
	unsigned long *ptr = filp->private_data;
2997
	char buf[64];
2998
	unsigned long val;
2999
	int ret;
3000

S
Steven Rostedt 已提交
3001 3002
	if (cnt >= sizeof(buf))
		return -EINVAL;
3003 3004 3005 3006 3007 3008

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

	buf[cnt] = 0;

3009 3010 3011
	ret = strict_strtoul(buf, 10, &val);
	if (ret < 0)
		return ret;
3012 3013 3014 3015 3016 3017

	*ptr = val * 1000;

	return cnt;
}

3018 3019
static int tracing_open_pipe(struct inode *inode, struct file *filp)
{
3020
	long cpu_file = (long) inode->i_private;
3021
	struct trace_iterator *iter;
3022
	int ret = 0;
3023 3024 3025 3026

	if (tracing_disabled)
		return -ENODEV;

3027 3028
	mutex_lock(&trace_types_lock);

3029 3030
	/* create a buffer to store the information to pass to userspace */
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
3031 3032 3033 3034
	if (!iter) {
		ret = -ENOMEM;
		goto out;
	}
3035

3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047
	/*
	 * 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;

3048
	if (!alloc_cpumask_var(&iter->started, GFP_KERNEL)) {
3049
		ret = -ENOMEM;
3050
		goto fail;
3051 3052
	}

3053
	/* trace pipe does not show start of buffer */
3054
	cpumask_setall(iter->started);
3055

3056 3057 3058
	if (trace_flags & TRACE_ITER_LATENCY_FMT)
		iter->iter_flags |= TRACE_FILE_LAT_FMT;

3059
	iter->cpu_file = cpu_file;
3060
	iter->tr = &global_trace;
3061
	mutex_init(&iter->mutex);
3062 3063
	filp->private_data = iter;

3064 3065 3066
	if (iter->trace->pipe_open)
		iter->trace->pipe_open(iter);

3067 3068 3069
out:
	mutex_unlock(&trace_types_lock);
	return ret;
3070 3071 3072 3073 3074 3075

fail:
	kfree(iter->trace);
	kfree(iter);
	mutex_unlock(&trace_types_lock);
	return ret;
3076 3077 3078 3079 3080 3081
}

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

3082 3083
	mutex_lock(&trace_types_lock);

3084
	if (iter->trace->pipe_close)
S
Steven Rostedt 已提交
3085 3086
		iter->trace->pipe_close(iter);

3087 3088
	mutex_unlock(&trace_types_lock);

3089
	free_cpumask_var(iter->started);
3090 3091
	mutex_destroy(&iter->mutex);
	kfree(iter->trace);
3092 3093 3094 3095 3096
	kfree(iter);

	return 0;
}

3097 3098 3099 3100 3101 3102 3103 3104 3105 3106
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
3107
	} else {
3108 3109 3110 3111 3112 3113 3114 3115 3116 3117
		if (!trace_empty(iter))
			return POLLIN | POLLRDNORM;
		poll_wait(filp, &trace_wait, poll_table);
		if (!trace_empty(iter))
			return POLLIN | POLLRDNORM;

		return 0;
	}
}

3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149

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

	prepare_to_wait(&trace_wait, &wait, TASK_INTERRUPTIBLE);

	if (trace_empty(iter))
		schedule();

	finish_wait(&trace_wait, &wait);
}

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

3150 3151
/* Must be called with trace_types_lock mutex held. */
static int tracing_wait_pipe(struct file *filp)
3152 3153 3154 3155
{
	struct trace_iterator *iter = filp->private_data;

	while (trace_empty(iter)) {
3156

3157
		if ((filp->f_flags & O_NONBLOCK)) {
3158
			return -EAGAIN;
3159
		}
3160

3161
		mutex_unlock(&iter->mutex);
3162

3163
		iter->trace->wait_pipe(iter);
3164

3165
		mutex_lock(&iter->mutex);
3166

3167
		if (signal_pending(current))
3168
			return -EINTR;
3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182

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

3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193
	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;
3194
	static struct tracer *old_tracer;
3195 3196 3197 3198 3199 3200 3201
	ssize_t sret;

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

3202
	trace_seq_init(&iter->seq);
3203

3204
	/* copy the tracer to avoid using a global lock all around */
3205
	mutex_lock(&trace_types_lock);
3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217
	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);
3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228
	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;

3229
	/* stop when tracing is finished */
3230 3231
	if (trace_empty(iter)) {
		sret = 0;
3232
		goto out;
3233
	}
3234 3235 3236 3237

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

3238 3239 3240 3241
	/* reset all but tr, trace, and overruns */
	memset(&iter->seq, 0,
	       sizeof(struct trace_iterator) -
	       offsetof(struct trace_iterator, seq));
3242
	iter->pos = -1;
3243

3244
	trace_event_read_lock();
3245
	trace_access_lock(iter->cpu_file);
S
Steven Rostedt 已提交
3246
	while (find_next_entry_inc(iter) != NULL) {
3247
		enum print_line_t ret;
S
Steven Rostedt 已提交
3248 3249
		int len = iter->seq.len;

I
Ingo Molnar 已提交
3250
		ret = print_trace_line(iter);
3251
		if (ret == TRACE_TYPE_PARTIAL_LINE) {
S
Steven Rostedt 已提交
3252 3253
			/* don't print partial lines */
			iter->seq.len = len;
3254
			break;
S
Steven Rostedt 已提交
3255
		}
3256 3257
		if (ret != TRACE_TYPE_NO_CONSUME)
			trace_consume(iter);
3258 3259 3260 3261

		if (iter->seq.len >= cnt)
			break;
	}
3262
	trace_access_unlock(iter->cpu_file);
3263
	trace_event_read_unlock();
3264 3265

	/* Now copy what we have to the user */
3266 3267
	sret = trace_seq_to_user(&iter->seq, ubuf, cnt);
	if (iter->seq.readpos >= iter->seq.len)
3268
		trace_seq_init(&iter->seq);
P
Pekka Paalanen 已提交
3269 3270 3271 3272 3273

	/*
	 * If there was nothing to send to user, inspite of consuming trace
	 * entries, go back to wait for more entries.
	 */
3274
	if (sret == -EBUSY)
P
Pekka Paalanen 已提交
3275
		goto waitagain;
3276

3277
out:
3278
	mutex_unlock(&iter->mutex);
3279

3280
	return sret;
3281 3282
}

3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294
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]);
}

3295
static const struct pipe_buf_operations tracing_pipe_buf_ops = {
S
Steven Rostedt 已提交
3296 3297 3298 3299 3300 3301 3302
	.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,
3303 3304
};

S
Steven Rostedt 已提交
3305
static size_t
3306
tracing_fill_pipe_page(size_t rem, struct trace_iterator *iter)
S
Steven Rostedt 已提交
3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325
{
	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;
		}

3326 3327
		if (ret != TRACE_TYPE_NO_CONSUME)
			trace_consume(iter);
S
Steven Rostedt 已提交
3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338
		rem -= count;
		if (!find_next_entry_inc(iter))	{
			rem = 0;
			iter->ent = NULL;
			break;
		}
	}

	return rem;
}

3339 3340 3341 3342 3343 3344
static ssize_t tracing_splice_read_pipe(struct file *filp,
					loff_t *ppos,
					struct pipe_inode_info *pipe,
					size_t len,
					unsigned int flags)
{
3345 3346
	struct page *pages_def[PIPE_DEF_BUFFERS];
	struct partial_page partial_def[PIPE_DEF_BUFFERS];
3347 3348
	struct trace_iterator *iter = filp->private_data;
	struct splice_pipe_desc spd = {
3349 3350
		.pages		= pages_def,
		.partial	= partial_def,
S
Steven Rostedt 已提交
3351 3352 3353 3354
		.nr_pages	= 0, /* This gets updated below. */
		.flags		= flags,
		.ops		= &tracing_pipe_buf_ops,
		.spd_release	= tracing_spd_release_pipe,
3355
	};
3356
	static struct tracer *old_tracer;
3357
	ssize_t ret;
S
Steven Rostedt 已提交
3358
	size_t rem;
3359 3360
	unsigned int i;

3361 3362 3363
	if (splice_grow_spd(pipe, &spd))
		return -ENOMEM;

3364
	/* copy the tracer to avoid using a global lock all around */
3365
	mutex_lock(&trace_types_lock);
3366 3367 3368 3369 3370 3371 3372
	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);
3373 3374 3375 3376 3377

	if (iter->trace->splice_read) {
		ret = iter->trace->splice_read(iter, filp,
					       ppos, pipe, len, flags);
		if (ret)
S
Steven Rostedt 已提交
3378
			goto out_err;
3379 3380 3381 3382
	}

	ret = tracing_wait_pipe(filp);
	if (ret <= 0)
S
Steven Rostedt 已提交
3383
		goto out_err;
3384 3385 3386

	if (!iter->ent && !find_next_entry_inc(iter)) {
		ret = -EFAULT;
S
Steven Rostedt 已提交
3387
		goto out_err;
3388 3389
	}

3390
	trace_event_read_lock();
3391
	trace_access_lock(iter->cpu_file);
3392

3393
	/* Fill as many pages as possible. */
3394 3395 3396
	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
3397
			break;
3398

3399
		rem = tracing_fill_pipe_page(rem, iter);
3400 3401 3402

		/* Copy the data into the page, so we can start over. */
		ret = trace_seq_to_buffer(&iter->seq,
3403
					  page_address(spd.pages[i]),
3404 3405
					  iter->seq.len);
		if (ret < 0) {
3406
			__free_page(spd.pages[i]);
3407 3408
			break;
		}
3409 3410
		spd.partial[i].offset = 0;
		spd.partial[i].len = iter->seq.len;
3411

3412
		trace_seq_init(&iter->seq);
3413 3414
	}

3415
	trace_access_unlock(iter->cpu_file);
3416
	trace_event_read_unlock();
3417
	mutex_unlock(&iter->mutex);
3418 3419 3420

	spd.nr_pages = i;

3421 3422 3423 3424
	ret = splice_to_pipe(pipe, &spd);
out:
	splice_shrink_spd(pipe, &spd);
	return ret;
3425

S
Steven Rostedt 已提交
3426
out_err:
3427
	mutex_unlock(&iter->mutex);
3428
	goto out;
3429 3430
}

3431 3432 3433 3434 3435
static ssize_t
tracing_entries_read(struct file *filp, char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
	struct trace_array *tr = filp->private_data;
3436
	char buf[96];
3437 3438
	int r;

3439 3440 3441 3442 3443 3444 3445 3446 3447
	mutex_lock(&trace_types_lock);
	if (!ring_buffer_expanded)
		r = sprintf(buf, "%lu (expanded: %lu)\n",
			    tr->entries >> 10,
			    trace_buf_size >> 10);
	else
		r = sprintf(buf, "%lu\n", tr->entries >> 10);
	mutex_unlock(&trace_types_lock);

3448 3449 3450 3451 3452 3453 3454 3455 3456
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
}

static ssize_t
tracing_entries_write(struct file *filp, const char __user *ubuf,
		      size_t cnt, loff_t *ppos)
{
	unsigned long val;
	char buf[64];
S
Steven Rostedt 已提交
3457
	int ret, cpu;
3458

S
Steven Rostedt 已提交
3459 3460
	if (cnt >= sizeof(buf))
		return -EINVAL;
3461 3462 3463 3464 3465 3466

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

	buf[cnt] = 0;

3467 3468 3469
	ret = strict_strtoul(buf, 10, &val);
	if (ret < 0)
		return ret;
3470 3471 3472 3473 3474 3475 3476

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

	mutex_lock(&trace_types_lock);

S
Steven Rostedt 已提交
3477
	tracing_stop();
3478

S
Steven Rostedt 已提交
3479 3480 3481 3482 3483 3484 3485 3486
	/* disable all cpu buffers */
	for_each_tracing_cpu(cpu) {
		if (global_trace.data[cpu])
			atomic_inc(&global_trace.data[cpu]->disabled);
		if (max_tr.data[cpu])
			atomic_inc(&max_tr.data[cpu]->disabled);
	}

3487 3488 3489
	/* value is in KB */
	val <<= 10;

3490
	if (val != global_trace.entries) {
3491
		ret = tracing_resize_ring_buffer(val);
3492 3493
		if (ret < 0) {
			cnt = ret;
S
Steven Rostedt 已提交
3494 3495
			goto out;
		}
3496 3497
	}

3498
	*ppos += cnt;
3499

3500 3501 3502
	/* If check pages failed, return ENOMEM */
	if (tracing_disabled)
		cnt = -ENOMEM;
3503
 out:
S
Steven Rostedt 已提交
3504 3505 3506 3507 3508 3509 3510
	for_each_tracing_cpu(cpu) {
		if (global_trace.data[cpu])
			atomic_dec(&global_trace.data[cpu]->disabled);
		if (max_tr.data[cpu])
			atomic_dec(&max_tr.data[cpu]->disabled);
	}

S
Steven Rostedt 已提交
3511
	tracing_start();
3512 3513 3514 3515 3516
	mutex_unlock(&trace_types_lock);

	return cnt;
}

3517 3518 3519 3520 3521 3522 3523 3524 3525 3526
static int mark_printk(const char *fmt, ...)
{
	int ret;
	va_list args;
	va_start(args, fmt);
	ret = trace_vprintk(0, fmt, args);
	va_end(args);
	return ret;
}

3527 3528 3529 3530 3531 3532
static ssize_t
tracing_mark_write(struct file *filp, const char __user *ubuf,
					size_t cnt, loff_t *fpos)
{
	char *buf;

S
Steven Rostedt 已提交
3533
	if (tracing_disabled)
3534 3535 3536 3537 3538
		return -EINVAL;

	if (cnt > TRACE_BUF_SIZE)
		cnt = TRACE_BUF_SIZE;

C
Carsten Emde 已提交
3539
	buf = kmalloc(cnt + 2, GFP_KERNEL);
3540 3541 3542 3543 3544 3545 3546
	if (buf == NULL)
		return -ENOMEM;

	if (copy_from_user(buf, ubuf, cnt)) {
		kfree(buf);
		return -EFAULT;
	}
C
Carsten Emde 已提交
3547 3548 3549 3550 3551
	if (buf[cnt-1] != '\n') {
		buf[cnt] = '\n';
		buf[cnt+1] = '\0';
	} else
		buf[cnt] = '\0';
3552

3553
	cnt = mark_printk("%s", buf);
3554 3555 3556 3557 3558 3559
	kfree(buf);
	*fpos += cnt;

	return cnt;
}

L
Li Zefan 已提交
3560
static int tracing_clock_show(struct seq_file *m, void *v)
3561 3562 3563 3564
{
	int i;

	for (i = 0; i < ARRAY_SIZE(trace_clocks); i++)
L
Li Zefan 已提交
3565
		seq_printf(m,
3566 3567 3568
			"%s%s%s%s", i ? " " : "",
			i == trace_clock_id ? "[" : "", trace_clocks[i].name,
			i == trace_clock_id ? "]" : "");
L
Li Zefan 已提交
3569
	seq_putc(m, '\n');
3570

L
Li Zefan 已提交
3571
	return 0;
3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612
}

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

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

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

	buf[cnt] = 0;

	clockstr = strstrip(buf);

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

	trace_clock_id = i;

	mutex_lock(&trace_types_lock);

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

	mutex_unlock(&trace_types_lock);

	*fpos += cnt;

	return cnt;
}

L
Li Zefan 已提交
3613 3614 3615 3616 3617 3618 3619
static int tracing_clock_open(struct inode *inode, struct file *file)
{
	if (tracing_disabled)
		return -ENODEV;
	return single_open(file, tracing_clock_show, NULL);
}

3620
static const struct file_operations tracing_max_lat_fops = {
I
Ingo Molnar 已提交
3621 3622 3623
	.open		= tracing_open_generic,
	.read		= tracing_max_lat_read,
	.write		= tracing_max_lat_write,
3624 3625
};

3626
static const struct file_operations tracing_ctrl_fops = {
I
Ingo Molnar 已提交
3627 3628 3629
	.open		= tracing_open_generic,
	.read		= tracing_ctrl_read,
	.write		= tracing_ctrl_write,
3630 3631
};

3632
static const struct file_operations set_tracer_fops = {
I
Ingo Molnar 已提交
3633 3634 3635
	.open		= tracing_open_generic,
	.read		= tracing_set_trace_read,
	.write		= tracing_set_trace_write,
3636 3637
};

3638
static const struct file_operations tracing_pipe_fops = {
I
Ingo Molnar 已提交
3639
	.open		= tracing_open_pipe,
3640
	.poll		= tracing_poll_pipe,
I
Ingo Molnar 已提交
3641
	.read		= tracing_read_pipe,
3642
	.splice_read	= tracing_splice_read_pipe,
I
Ingo Molnar 已提交
3643
	.release	= tracing_release_pipe,
3644 3645
};

3646
static const struct file_operations tracing_entries_fops = {
3647 3648 3649 3650 3651
	.open		= tracing_open_generic,
	.read		= tracing_entries_read,
	.write		= tracing_entries_write,
};

3652
static const struct file_operations tracing_mark_fops = {
3653
	.open		= tracing_open_generic,
3654 3655 3656
	.write		= tracing_mark_write,
};

3657
static const struct file_operations trace_clock_fops = {
L
Li Zefan 已提交
3658 3659 3660 3661
	.open		= tracing_clock_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= single_release,
3662 3663 3664
	.write		= tracing_clock_write,
};

3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685
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;
3686
	info->spare	= NULL;
3687 3688 3689 3690 3691
	/* Force reading ring buffer for first read */
	info->read	= (unsigned int)-1;

	filp->private_data = info;

3692
	return nonseekable_open(inode, filp);
3693 3694 3695 3696 3697 3698 3699 3700 3701 3702
}

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;

3703 3704 3705
	if (!count)
		return 0;

3706 3707 3708 3709 3710
	if (!info->spare)
		info->spare = ring_buffer_alloc_read_page(info->tr->buffer);
	if (!info->spare)
		return -ENOMEM;

3711 3712 3713 3714 3715 3716
	/* Do we have previous read data to read? */
	if (info->read < PAGE_SIZE)
		goto read;

	info->read = 0;

3717
	trace_access_lock(info->cpu);
3718 3719 3720 3721
	ret = ring_buffer_read_page(info->tr->buffer,
				    &info->spare,
				    count,
				    info->cpu, 0);
3722
	trace_access_unlock(info->cpu);
3723 3724 3725 3726 3727 3728 3729 3730 3731
	if (ret < 0)
		return 0;

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

	ret = copy_to_user(ubuf, info->spare + info->read, size);
3732
	if (ret == size)
3733
		return -EFAULT;
3734 3735
	size -= ret;

3736 3737 3738 3739 3740 3741 3742 3743 3744 3745
	*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;

3746 3747
	if (info->spare)
		ring_buffer_free_read_page(info->tr->buffer, info->spare);
3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 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
	kfree(info);

	return 0;
}

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

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

	if (--ref->ref)
		return;

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

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

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

	ref->ref++;
}

/* Pipe buffer operations for a buffer. */
3787
static const struct pipe_buf_operations buffer_pipe_buf_ops = {
3788 3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819
	.can_merge		= 0,
	.map			= generic_pipe_buf_map,
	.unmap			= generic_pipe_buf_unmap,
	.confirm		= generic_pipe_buf_confirm,
	.release		= buffer_pipe_buf_release,
	.steal			= buffer_pipe_buf_steal,
	.get			= buffer_pipe_buf_get,
};

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

	if (--ref->ref)
		return;

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

static ssize_t
tracing_buffers_splice_read(struct file *file, loff_t *ppos,
			    struct pipe_inode_info *pipe, size_t len,
			    unsigned int flags)
{
	struct ftrace_buffer_info *info = file->private_data;
3820 3821
	struct partial_page partial_def[PIPE_DEF_BUFFERS];
	struct page *pages_def[PIPE_DEF_BUFFERS];
3822
	struct splice_pipe_desc spd = {
3823 3824
		.pages		= pages_def,
		.partial	= partial_def,
3825 3826 3827 3828 3829
		.flags		= flags,
		.ops		= &buffer_pipe_buf_ops,
		.spd_release	= buffer_spd_release,
	};
	struct buffer_ref *ref;
3830
	int entries, size, i;
3831 3832
	size_t ret;

3833 3834 3835
	if (splice_grow_spd(pipe, &spd))
		return -ENOMEM;

3836 3837
	if (*ppos & (PAGE_SIZE - 1)) {
		WARN_ONCE(1, "Ftrace: previous read must page-align\n");
3838 3839
		ret = -EINVAL;
		goto out;
3840 3841 3842 3843
	}

	if (len & (PAGE_SIZE - 1)) {
		WARN_ONCE(1, "Ftrace: splice_read should page-align\n");
3844 3845 3846 3847
		if (len < PAGE_SIZE) {
			ret = -EINVAL;
			goto out;
		}
3848 3849 3850
		len &= PAGE_MASK;
	}

3851
	trace_access_lock(info->cpu);
3852 3853
	entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu);

3854
	for (i = 0; i < pipe->buffers && len && entries; i++, len -= PAGE_SIZE) {
3855 3856 3857 3858 3859 3860 3861
		struct page *page;
		int r;

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

3862
		ref->ref = 1;
3863 3864 3865 3866 3867 3868 3869 3870
		ref->buffer = info->tr->buffer;
		ref->page = ring_buffer_alloc_read_page(ref->buffer);
		if (!ref->page) {
			kfree(ref);
			break;
		}

		r = ring_buffer_read_page(ref->buffer, &ref->page,
3871
					  len, info->cpu, 1);
3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893
		if (r < 0) {
			ring_buffer_free_read_page(ref->buffer,
						   ref->page);
			kfree(ref);
			break;
		}

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

		page = virt_to_page(ref->page);

		spd.pages[i] = page;
		spd.partial[i].len = PAGE_SIZE;
		spd.partial[i].offset = 0;
		spd.partial[i].private = (unsigned long)ref;
		spd.nr_pages++;
3894
		*ppos += PAGE_SIZE;
3895 3896

		entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu);
3897 3898
	}

3899
	trace_access_unlock(info->cpu);
3900 3901 3902 3903 3904 3905 3906 3907 3908
	spd.nr_pages = i;

	/* did we read anything? */
	if (!spd.nr_pages) {
		if (flags & SPLICE_F_NONBLOCK)
			ret = -EAGAIN;
		else
			ret = 0;
		/* TODO: block */
3909
		goto out;
3910 3911 3912
	}

	ret = splice_to_pipe(pipe, &spd);
3913 3914
	splice_shrink_spd(pipe, &spd);
out:
3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925
	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,
};

3926 3927 3928 3929 3930 3931 3932 3933 3934
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;

3935
	s = kmalloc(sizeof(*s), GFP_KERNEL);
3936
	if (!s)
3937
		return -ENOMEM;
3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961

	trace_seq_init(s);

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

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

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

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

	kfree(s);

	return count;
}

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

3962 3963
#ifdef CONFIG_DYNAMIC_FTRACE

S
Steven Rostedt 已提交
3964 3965 3966 3967 3968
int __weak ftrace_arch_read_dyn_info(char *buf, int size)
{
	return 0;
}

3969
static ssize_t
S
Steven Rostedt 已提交
3970
tracing_read_dyn_info(struct file *filp, char __user *ubuf,
3971 3972
		  size_t cnt, loff_t *ppos)
{
S
Steven Rostedt 已提交
3973 3974
	static char ftrace_dyn_info_buffer[1024];
	static DEFINE_MUTEX(dyn_info_mutex);
3975
	unsigned long *p = filp->private_data;
S
Steven Rostedt 已提交
3976
	char *buf = ftrace_dyn_info_buffer;
S
Steven Rostedt 已提交
3977
	int size = ARRAY_SIZE(ftrace_dyn_info_buffer);
3978 3979
	int r;

S
Steven Rostedt 已提交
3980 3981
	mutex_lock(&dyn_info_mutex);
	r = sprintf(buf, "%ld ", *p);
I
Ingo Molnar 已提交
3982

S
Steven Rostedt 已提交
3983
	r += ftrace_arch_read_dyn_info(buf+r, (size-1)-r);
S
Steven Rostedt 已提交
3984 3985 3986 3987 3988 3989 3990
	buf[r++] = '\n';

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

	mutex_unlock(&dyn_info_mutex);

	return r;
3991 3992
}

3993
static const struct file_operations tracing_dyn_info_fops = {
I
Ingo Molnar 已提交
3994
	.open		= tracing_open_generic,
S
Steven Rostedt 已提交
3995
	.read		= tracing_read_dyn_info,
3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007
};
#endif

static struct dentry *d_tracer;

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

	if (d_tracer)
		return d_tracer;

4008 4009 4010
	if (!debugfs_initialized())
		return NULL;

4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021
	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;
}

4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050
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();
4051
	struct dentry *d_cpu;
4052 4053
	/* strlen(cpu) + MAX(log10(cpu)) + '\0' */
	char cpu_dir[7];
4054 4055 4056 4057

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

4058 4059 4060 4061 4062 4063
	sprintf(cpu_dir, "cpu%ld", cpu);
	d_cpu = debugfs_create_dir(cpu_dir, d_percpu);
	if (!d_cpu) {
		pr_warning("Could not create debugfs '%s' entry\n", cpu_dir);
		return;
	}
4064

4065
	/* per cpu trace_pipe */
4066 4067
	trace_create_file("trace_pipe", 0444, d_cpu,
			(void *) cpu, &tracing_pipe_fops);
4068 4069

	/* per cpu trace */
4070 4071
	trace_create_file("trace", 0644, d_cpu,
			(void *) cpu, &tracing_fops);
4072

4073 4074
	trace_create_file("trace_pipe_raw", 0444, d_cpu,
			(void *) cpu, &tracing_buffers_fops);
4075

4076 4077
	trace_create_file("stats", 0444, d_cpu,
			(void *) cpu, &tracing_stats_fops);
4078 4079
}

S
Steven Rostedt 已提交
4080 4081 4082 4083 4084
#ifdef CONFIG_FTRACE_SELFTEST
/* Let selftest have access to static functions in this file */
#include "trace_selftest.c"
#endif

4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126
struct trace_option_dentry {
	struct tracer_opt		*opt;
	struct tracer_flags		*flags;
	struct dentry			*entry;
};

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

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

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

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

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

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

	buf[cnt] = 0;

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

L
Li Zefan 已提交
4127 4128
	if (val != 0 && val != 1)
		return -EINVAL;
4129

L
Li Zefan 已提交
4130
	if (!!(topt->flags->val & topt->opt->bit) != val) {
4131
		mutex_lock(&trace_types_lock);
L
Li Zefan 已提交
4132
		ret = __set_tracer_option(current_trace, topt->flags,
4133
					  topt->opt, !val);
4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150
		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,
};

4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186
static ssize_t
trace_options_core_read(struct file *filp, char __user *ubuf, size_t cnt,
			loff_t *ppos)
{
	long index = (long)filp->private_data;
	char *buf;

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

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

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

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

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

	buf[cnt] = 0;

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

4187
	if (val != 0 && val != 1)
4188
		return -EINVAL;
4189
	set_tracer_flags(1 << index, val);
4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201

	*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,
};

4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217
struct dentry *trace_create_file(const char *name,
				 mode_t mode,
				 struct dentry *parent,
				 void *data,
				 const struct file_operations *fops)
{
	struct dentry *ret;

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

	return ret;
}


4218 4219 4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237 4238
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;
}

4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252
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;

4253
	topt->entry = trace_create_file(opt->name, 0644, t_options, topt,
4254 4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278
				    &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++)
		;

4279
	topts = kcalloc(cnt + 1, sizeof(*topts), GFP_KERNEL);
4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305
	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);
}

4306 4307 4308 4309 4310 4311 4312 4313 4314
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;

4315
	return trace_create_file(option, 0644, t_options, (void *)index,
4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327
				    &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;

4328 4329
	for (i = 0; trace_options[i]; i++)
		create_trace_option_core_file(trace_options[i], i);
4330 4331
}

4332
static __init int tracer_init_debugfs(void)
4333 4334
{
	struct dentry *d_tracer;
4335
	int cpu;
4336

4337 4338
	trace_access_lock_init();

4339 4340
	d_tracer = tracing_init_dentry();

4341 4342
	trace_create_file("tracing_enabled", 0644, d_tracer,
			&global_trace, &tracing_ctrl_fops);
4343

4344 4345
	trace_create_file("trace_options", 0644, d_tracer,
			NULL, &tracing_iter_fops);
4346

4347 4348 4349 4350 4351
	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);
4352

4353 4354 4355
	trace_create_file("available_tracers", 0444, d_tracer,
			&global_trace, &show_traces_fops);

4356
	trace_create_file("current_tracer", 0644, d_tracer,
4357 4358
			&global_trace, &set_tracer_fops);

4359
#ifdef CONFIG_TRACER_MAX_TRACE
4360 4361
	trace_create_file("tracing_max_latency", 0644, d_tracer,
			&tracing_max_latency, &tracing_max_lat_fops);
4362
#endif
4363 4364 4365

	trace_create_file("tracing_thresh", 0644, d_tracer,
			&tracing_thresh, &tracing_max_lat_fops);
4366

4367
	trace_create_file("README", 0444, d_tracer,
4368 4369 4370
			NULL, &tracing_readme_fops);

	trace_create_file("trace_pipe", 0444, d_tracer,
4371
			(void *) TRACE_PIPE_ALL_CPU, &tracing_pipe_fops);
4372 4373 4374 4375 4376 4377

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

	trace_create_file("trace_marker", 0220, d_tracer,
			NULL, &tracing_mark_fops);
4378

4379 4380
	trace_create_file("saved_cmdlines", 0444, d_tracer,
			NULL, &tracing_saved_cmdlines_fops);
4381

4382 4383 4384
	trace_create_file("trace_clock", 0644, d_tracer, NULL,
			  &trace_clock_fops);

4385
#ifdef CONFIG_DYNAMIC_FTRACE
4386 4387
	trace_create_file("dyn_ftrace_total_info", 0444, d_tracer,
			&ftrace_update_tot_cnt, &tracing_dyn_info_fops);
4388
#endif
I
Ingo Molnar 已提交
4389 4390 4391
#ifdef CONFIG_SYSPROF_TRACER
	init_tracer_sysprof_debugfs(d_tracer);
#endif
4392

4393 4394
	create_trace_options_dir();

4395 4396 4397
	for_each_tracing_cpu(cpu)
		tracing_init_debugfs_percpu(cpu);

4398
	return 0;
4399 4400
}

4401 4402 4403
static int trace_panic_handler(struct notifier_block *this,
			       unsigned long event, void *unused)
{
4404
	if (ftrace_dump_on_oops)
4405
		ftrace_dump(ftrace_dump_on_oops);
4406 4407 4408 4409 4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420
	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:
4421
		if (ftrace_dump_on_oops)
4422
			ftrace_dump(ftrace_dump_on_oops);
4423 4424 4425 4426 4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445
		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.
 */
4446
#define KERN_TRACE		KERN_EMERG
4447 4448 4449 4450 4451 4452 4453 4454 4455 4456 4457 4458 4459

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

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

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

4460
	trace_seq_init(s);
4461 4462
}

4463 4464
static void
__ftrace_dump(bool disable_tracing, enum ftrace_dump_mode oops_dump_mode)
4465
{
4466
	static arch_spinlock_t ftrace_dump_lock =
4467
		(arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED;
4468 4469
	/* use static because iter can be a bit big for the stack */
	static struct trace_iterator iter;
4470
	unsigned int old_userobj;
4471
	static int dump_ran;
4472 4473
	unsigned long flags;
	int cnt = 0, cpu;
4474 4475

	/* only one dump */
4476
	local_irq_save(flags);
4477
	arch_spin_lock(&ftrace_dump_lock);
4478 4479 4480 4481 4482
	if (dump_ran)
		goto out;

	dump_ran = 1;

4483
	tracing_off();
4484 4485 4486

	if (disable_tracing)
		ftrace_kill();
4487

4488 4489 4490 4491
	for_each_tracing_cpu(cpu) {
		atomic_inc(&global_trace.data[cpu]->disabled);
	}

4492 4493
	old_userobj = trace_flags & TRACE_ITER_SYM_USEROBJ;

4494 4495 4496
	/* don't look at user memory in panic mode */
	trace_flags &= ~TRACE_ITER_SYM_USEROBJ;

4497
	/* Simulate the iterator */
4498 4499
	iter.tr = &global_trace;
	iter.trace = current_trace;
4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515

	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");
4516 4517 4518 4519 4520 4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538

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

	while (!trace_empty(&iter)) {

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

		cnt++;

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

		if (find_next_entry_inc(&iter) != NULL) {
4539 4540 4541 4542 4543
			int ret;

			ret = print_trace_line(&iter);
			if (ret != TRACE_TYPE_NO_CONSUME)
				trace_consume(&iter);
4544 4545 4546 4547 4548 4549 4550 4551 4552 4553
		}

		trace_printk_seq(&iter.seq);
	}

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

4554
 out_enable:
4555 4556 4557 4558 4559 4560 4561 4562 4563 4564
	/* Re-enable tracing if requested */
	if (!disable_tracing) {
		trace_flags |= old_userobj;

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

4565
 out:
4566
	arch_spin_unlock(&ftrace_dump_lock);
4567
	local_irq_restore(flags);
4568 4569
}

4570
/* By default: disable tracing after the dump */
4571
void ftrace_dump(enum ftrace_dump_mode oops_dump_mode)
4572
{
4573
	__ftrace_dump(true, oops_dump_mode);
4574 4575
}

4576
__init static int tracer_alloc_buffers(void)
4577
{
4578
	int ring_buf_size;
4579
	int i;
4580
	int ret = -ENOMEM;
4581

4582 4583 4584 4585 4586
	if (!alloc_cpumask_var(&tracing_buffer_mask, GFP_KERNEL))
		goto out;

	if (!alloc_cpumask_var(&tracing_cpumask, GFP_KERNEL))
		goto out_free_buffer_mask;
4587

4588 4589 4590 4591 4592 4593
	/* 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;

4594 4595 4596 4597
	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 */
4598
	global_trace.buffer = ring_buffer_alloc(ring_buf_size,
4599 4600 4601 4602
						   TRACE_BUFFER_FLAGS);
	if (!global_trace.buffer) {
		printk(KERN_ERR "tracer: failed to allocate ring buffer!\n");
		WARN_ON(1);
4603
		goto out_free_cpumask;
4604
	}
4605
	global_trace.entries = ring_buffer_size(global_trace.buffer);
4606

4607

4608
#ifdef CONFIG_TRACER_MAX_TRACE
4609
	max_tr.buffer = ring_buffer_alloc(1, TRACE_BUFFER_FLAGS);
4610 4611 4612 4613
	if (!max_tr.buffer) {
		printk(KERN_ERR "tracer: failed to allocate max ring buffer!\n");
		WARN_ON(1);
		ring_buffer_free(global_trace.buffer);
4614
		goto out_free_cpumask;
4615
	}
4616
	max_tr.entries = 1;
4617
#endif
4618

4619
	/* Allocate the first page for all buffers */
4620
	for_each_tracing_cpu(i) {
4621
		global_trace.data[i] = &per_cpu(global_trace_cpu, i);
4622
		max_tr.data[i] = &per_cpu(max_tr_data, i);
4623
	}
4624 4625 4626

	trace_init_cmdlines();

4627
	register_tracer(&nop_trace);
4628
	current_trace = &nop_trace;
S
Steven Rostedt 已提交
4629 4630
	/* All seems OK, enable tracing */
	tracing_disabled = 0;
4631

4632 4633 4634 4635
	atomic_notifier_chain_register(&panic_notifier_list,
				       &trace_panic_notifier);

	register_die_notifier(&trace_die_notifier);
4636 4637

	return 0;
4638

4639 4640 4641 4642 4643 4644
out_free_cpumask:
	free_cpumask_var(tracing_cpumask);
out_free_buffer_mask:
	free_cpumask_var(tracing_buffer_mask);
out:
	return ret;
4645
}
4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662 4663 4664 4665

__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;
}

4666 4667
early_initcall(tracer_alloc_buffers);
fs_initcall(tracer_init_debugfs);
4668
late_initcall(clear_boot_tracer);