trace.c 58.6 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
/*
 * 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
 */
#include <linux/utsrelease.h>
#include <linux/kallsyms.h>
#include <linux/seq_file.h>
#include <linux/debugfs.h>
18
#include <linux/pagemap.h>
19 20 21 22 23 24 25 26
#include <linux/hardirq.h>
#include <linux/linkage.h>
#include <linux/uaccess.h>
#include <linux/ftrace.h>
#include <linux/module.h>
#include <linux/percpu.h>
#include <linux/ctype.h>
#include <linux/init.h>
27
#include <linux/poll.h>
28 29 30
#include <linux/gfp.h>
#include <linux/fs.h>

I
Ingo Molnar 已提交
31 32
#include <linux/stacktrace.h>

33 34 35 36 37
#include "trace.h"

unsigned long __read_mostly	tracing_max_latency = (cycle_t)ULONG_MAX;
unsigned long __read_mostly	tracing_thresh;

S
Steven Rostedt 已提交
38 39
static int tracing_disabled = 1;

I
Ingo Molnar 已提交
40
static long
41 42 43 44 45 46 47
ns2usecs(cycle_t nsec)
{
	nsec += 500;
	do_div(nsec, 1000);
	return nsec;
}

I
Ingo Molnar 已提交
48
cycle_t ftrace_now(int cpu)
I
Ingo Molnar 已提交
49
{
I
Ingo Molnar 已提交
50
	return cpu_clock(cpu);
I
Ingo Molnar 已提交
51 52
}

53 54 55 56 57 58 59 60
static struct trace_array	global_trace;

static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu);

static struct trace_array	max_tr;

static DEFINE_PER_CPU(struct trace_array_cpu, max_data);

61
static int			tracer_enabled = 1;
62
static unsigned long		trace_nr_entries = 65536UL;
63 64 65 66 67 68

static struct tracer		*trace_types __read_mostly;
static struct tracer		*current_trace __read_mostly;
static int			max_tracer_type_len;

static DEFINE_MUTEX(trace_types_lock);
69 70 71 72 73 74
static DECLARE_WAIT_QUEUE_HEAD(trace_wait);

unsigned long trace_flags = TRACE_ITER_PRINT_PARENT;

void trace_wake_up(void)
{
I
Ingo Molnar 已提交
75 76 77 78 79
	/*
	 * The runqueue_is_locked() can fail, but this is the best we
	 * have for now:
	 */
	if (!(trace_flags & TRACE_ITER_BLOCK) && !runqueue_is_locked())
80 81
		wake_up(&trace_wait);
}
82

83 84
#define ENTRIES_PER_PAGE (PAGE_SIZE / sizeof(struct trace_entry))

85 86 87 88 89 90 91 92 93
static int __init set_nr_entries(char *str)
{
	if (!str)
		return 0;
	trace_nr_entries = simple_strtoul(str, &str, 0);
	return 1;
}
__setup("trace_entries=", set_nr_entries);

S
Steven Rostedt 已提交
94 95 96 97 98
unsigned long nsecs_to_usecs(unsigned long nsecs)
{
	return nsecs / 1000;
}

99 100 101 102 103
enum trace_type {
	__TRACE_FIRST_TYPE = 0,

	TRACE_FN,
	TRACE_CTX,
104
	TRACE_WAKE,
I
Ingo Molnar 已提交
105
	TRACE_STACK,
I
Ingo Molnar 已提交
106
	TRACE_SPECIAL,
107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126

	__TRACE_LAST_TYPE
};

enum trace_flag_type {
	TRACE_FLAG_IRQS_OFF		= 0x01,
	TRACE_FLAG_NEED_RESCHED		= 0x02,
	TRACE_FLAG_HARDIRQ		= 0x04,
	TRACE_FLAG_SOFTIRQ		= 0x08,
};

#define TRACE_ITER_SYM_MASK \
	(TRACE_ITER_PRINT_PARENT|TRACE_ITER_SYM_OFFSET|TRACE_ITER_SYM_ADDR)

/* These must match the bit postions above */
static const char *trace_options[] = {
	"print-parent",
	"sym-offset",
	"sym-addr",
	"verbose",
I
Ingo Molnar 已提交
127
	"raw",
128
	"hex",
I
Ingo Molnar 已提交
129
	"bin",
130
	"block",
I
Ingo Molnar 已提交
131
	"stacktrace",
I
Ingo Molnar 已提交
132
	"sched-tree",
133 134 135
	NULL
};

136
static DEFINE_SPINLOCK(ftrace_max_lock);
137 138 139 140 141 142

/*
 * Copy the new maximum trace into the separate maximum-trace
 * structure. (this way the maximum trace is permanently saved,
 * for later retrieval via /debugfs/tracing/latency_trace)
 */
I
Ingo Molnar 已提交
143
static void
144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164
__update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
	struct trace_array_cpu *data = tr->data[cpu];

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

	data = max_tr.data[cpu];
	data->saved_latency = tracing_max_latency;

	memcpy(data->comm, tsk->comm, TASK_COMM_LEN);
	data->pid = tsk->pid;
	data->uid = tsk->uid;
	data->nice = tsk->static_prio - 20 - MAX_RT_PRIO;
	data->policy = tsk->policy;
	data->rt_priority = tsk->rt_priority;

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

I
Ingo Molnar 已提交
165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
void check_pages(struct trace_array_cpu *data)
{
	struct page *page, *tmp;

	BUG_ON(data->trace_pages.next->prev != &data->trace_pages);
	BUG_ON(data->trace_pages.prev->next != &data->trace_pages);

	list_for_each_entry_safe(page, tmp, &data->trace_pages, lru) {
		BUG_ON(page->lru.next->prev != &page->lru);
		BUG_ON(page->lru.prev->next != &page->lru);
	}
}

void *head_page(struct trace_array_cpu *data)
{
	struct page *page;

	check_pages(data);
	if (list_empty(&data->trace_pages))
		return NULL;

	page = list_entry(data->trace_pages.next, struct page, lru);
	BUG_ON(&page->lru == &data->trace_pages);

	return page_address(page);
}

I
Ingo Molnar 已提交
192
static int
S
Steven Rostedt 已提交
193 194 195 196
trace_seq_printf(struct trace_seq *s, const char *fmt, ...)
{
	int len = (PAGE_SIZE - 1) - s->len;
	va_list ap;
197
	int ret;
S
Steven Rostedt 已提交
198 199 200 201 202

	if (!len)
		return 0;

	va_start(ap, fmt);
203
	ret = vsnprintf(s->buffer + s->len, len, fmt, ap);
S
Steven Rostedt 已提交
204 205
	va_end(ap);

206 207 208 209 210
	/* If we can't write it all, don't bother writing anything */
	if (ret > len)
		return 0;

	s->len += ret;
S
Steven Rostedt 已提交
211 212 213 214

	return len;
}

I
Ingo Molnar 已提交
215
static int
S
Steven Rostedt 已提交
216 217 218 219 220
trace_seq_puts(struct trace_seq *s, const char *str)
{
	int len = strlen(str);

	if (len > ((PAGE_SIZE - 1) - s->len))
221
		return 0;
S
Steven Rostedt 已提交
222 223 224 225 226 227 228

	memcpy(s->buffer + s->len, str, len);
	s->len += len;

	return len;
}

I
Ingo Molnar 已提交
229
static int
S
Steven Rostedt 已提交
230 231 232 233 234 235 236 237 238 239
trace_seq_putc(struct trace_seq *s, unsigned char c)
{
	if (s->len >= (PAGE_SIZE - 1))
		return 0;

	s->buffer[s->len++] = c;

	return 1;
}

I
Ingo Molnar 已提交
240
static int
I
Ingo Molnar 已提交
241 242 243 244 245 246 247 248 249 250 251
trace_seq_putmem(struct trace_seq *s, void *mem, size_t len)
{
	if (len > ((PAGE_SIZE - 1) - s->len))
		return 0;

	memcpy(s->buffer + s->len, mem, len);
	s->len += len;

	return len;
}

252 253
#define HEX_CHARS 17

I
Ingo Molnar 已提交
254
static int
255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292
trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len)
{
	unsigned char hex[HEX_CHARS];
	unsigned char *data;
	unsigned char byte;
	int i, j;

	BUG_ON(len >= HEX_CHARS);

	data = mem;

#ifdef __BIG_ENDIAN
	for (i = 0, j = 0; i < len; i++) {
#else
	for (i = len-1, j = 0; i >= 0; i--) {
#endif
		byte = data[i];

		hex[j]   = byte & 0x0f;
		if (hex[j] >= 10)
			hex[j] += 'a' - 10;
		else
			hex[j] += '0';
		j++;

		hex[j] = byte >> 4;
		if (hex[j] >= 10)
			hex[j] += 'a' - 10;
		else
			hex[j] += '0';
		j++;
	}
	hex[j] = ' ';
	j++;

	return trace_seq_putmem(s, hex, j);
}

I
Ingo Molnar 已提交
293
static void
S
Steven Rostedt 已提交
294 295 296 297 298
trace_seq_reset(struct trace_seq *s)
{
	s->len = 0;
}

I
Ingo Molnar 已提交
299
static void
S
Steven Rostedt 已提交
300 301 302 303 304 305 306 307 308 309
trace_print_seq(struct seq_file *m, struct trace_seq *s)
{
	int len = s->len >= PAGE_SIZE ? PAGE_SIZE - 1 : s->len;

	s->buffer[len] = 0;
	seq_puts(m, s->buffer);

	trace_seq_reset(s);
}

I
Ingo Molnar 已提交
310
static void
I
Ingo Molnar 已提交
311 312 313 314 315 316
flip_trace(struct trace_array_cpu *tr1, struct trace_array_cpu *tr2)
{
	struct list_head flip_pages;

	INIT_LIST_HEAD(&flip_pages);

317
	memcpy(&tr1->trace_head_idx, &tr2->trace_head_idx,
I
Ingo Molnar 已提交
318
		sizeof(struct trace_array_cpu) -
319
		offsetof(struct trace_array_cpu, trace_head_idx));
I
Ingo Molnar 已提交
320 321 322 323 324 325 326 327 328 329 330

	check_pages(tr1);
	check_pages(tr2);
	list_splice_init(&tr1->trace_pages, &flip_pages);
	list_splice_init(&tr2->trace_pages, &tr1->trace_pages);
	list_splice_init(&flip_pages, &tr2->trace_pages);
	BUG_ON(!list_empty(&flip_pages));
	check_pages(tr1);
	check_pages(tr2);
}

I
Ingo Molnar 已提交
331
void
332 333 334 335 336
update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
	struct trace_array_cpu *data;
	int i;

337 338
	WARN_ON_ONCE(!irqs_disabled());
	spin_lock(&ftrace_max_lock);
339 340 341
	/* clear out all the previous traces */
	for_each_possible_cpu(i) {
		data = tr->data[i];
I
Ingo Molnar 已提交
342
		flip_trace(max_tr.data[i], data);
343
		tracing_reset(data);
344 345 346
	}

	__update_max_tr(tr, tsk, cpu);
347
	spin_unlock(&ftrace_max_lock);
348 349 350 351 352 353 354 355
}

/**
 * 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.
 */
I
Ingo Molnar 已提交
356
void
357 358 359 360 361
update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu)
{
	struct trace_array_cpu *data = tr->data[cpu];
	int i;

362 363
	WARN_ON_ONCE(!irqs_disabled());
	spin_lock(&ftrace_max_lock);
364 365 366
	for_each_possible_cpu(i)
		tracing_reset(max_tr.data[i]);

I
Ingo Molnar 已提交
367
	flip_trace(max_tr.data[cpu], data);
368
	tracing_reset(data);
369 370

	__update_max_tr(tr, tsk, cpu);
371
	spin_unlock(&ftrace_max_lock);
372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395
}

int register_tracer(struct tracer *type)
{
	struct tracer *t;
	int len;
	int ret = 0;

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

	mutex_lock(&trace_types_lock);
	for (t = trace_types; t; t = t->next) {
		if (strcmp(type->name, t->name) == 0) {
			/* already found */
			pr_info("Trace %s already registered\n",
				type->name);
			ret = -1;
			goto out;
		}
	}

S
Steven Rostedt 已提交
396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411
#ifdef CONFIG_FTRACE_STARTUP_TEST
	if (type->selftest) {
		struct tracer *saved_tracer = current_trace;
		struct trace_array_cpu *data;
		struct trace_array *tr = &global_trace;
		int saved_ctrl = tr->ctrl;
		int i;
		/*
		 * 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.
		 */
		for_each_possible_cpu(i) {
			data = tr->data[i];
I
Ingo Molnar 已提交
412 413
			if (!head_page(data))
				continue;
S
Steven Rostedt 已提交
414 415 416 417 418 419 420 421 422 423 424 425 426 427
			tracing_reset(data);
		}
		current_trace = type;
		tr->ctrl = 0;
		/* 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;
		tr->ctrl = saved_ctrl;
		if (ret) {
			printk(KERN_CONT "FAILED!\n");
			goto out;
		}
S
Steven Rostedt 已提交
428 429 430 431 432 433 434
		/* Only reset on passing, to avoid touching corrupted buffers */
		for_each_possible_cpu(i) {
			data = tr->data[i];
			if (!head_page(data))
				continue;
			tracing_reset(data);
		}
S
Steven Rostedt 已提交
435 436 437 438
		printk(KERN_CONT "PASSED\n");
	}
#endif

439 440 441 442 443
	type->next = trace_types;
	trace_types = type;
	len = strlen(type->name);
	if (len > max_tracer_type_len)
		max_tracer_type_len = len;
S
Steven Rostedt 已提交
444

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
 out:
	mutex_unlock(&trace_types_lock);

	return ret;
}

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

	mutex_lock(&trace_types_lock);
	for (t = &trace_types; *t; t = &(*t)->next) {
		if (*t == type)
			goto found;
	}
	pr_info("Trace %s not registered\n", type->name);
	goto out;

 found:
	*t = (*t)->next;
	if (strlen(type->name) != max_tracer_type_len)
		goto out;

	max_tracer_type_len = 0;
	for (t = &trace_types; *t; t = &(*t)->next) {
		len = strlen((*t)->name);
		if (len > max_tracer_type_len)
			max_tracer_type_len = len;
	}
 out:
	mutex_unlock(&trace_types_lock);
}

I
Ingo Molnar 已提交
479
void tracing_reset(struct trace_array_cpu *data)
480 481
{
	data->trace_idx = 0;
482 483 484
	data->trace_head = data->trace_tail = head_page(data);
	data->trace_head_idx = 0;
	data->trace_tail_idx = 0;
485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501
}

#define SAVED_CMDLINES 128
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;
static DEFINE_SPINLOCK(trace_cmdline_lock);
atomic_t trace_record_cmdline_disabled;

static void trace_init_cmdlines(void)
{
	memset(&map_pid_to_cmdline, -1, sizeof(map_pid_to_cmdline));
	memset(&map_cmdline_to_pid, -1, sizeof(map_cmdline_to_pid));
	cmdline_idx = 0;
}

I
Ingo Molnar 已提交
502
void trace_stop_cmdline_recording(void);
503

I
Ingo Molnar 已提交
504
static void trace_save_cmdline(struct task_struct *tsk)
505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538
{
	unsigned map;
	unsigned idx;

	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.
	 */
	if (!spin_trylock(&trace_cmdline_lock))
		return;

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

		map = map_cmdline_to_pid[idx];
		if (map <= PID_MAX_DEFAULT)
			map_pid_to_cmdline[map] = (unsigned)-1;

		map_pid_to_cmdline[tsk->pid] = idx;

		cmdline_idx = idx;
	}

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

	spin_unlock(&trace_cmdline_lock);
}

I
Ingo Molnar 已提交
539
static char *trace_find_cmdline(int pid)
540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559
{
	char *cmdline = "<...>";
	unsigned map;

	if (!pid)
		return "<idle>";

	if (pid > PID_MAX_DEFAULT)
		goto out;

	map = map_pid_to_cmdline[pid];
	if (map >= SAVED_CMDLINES)
		goto out;

	cmdline = saved_cmdlines[map];

 out:
	return cmdline;
}

I
Ingo Molnar 已提交
560
void tracing_record_cmdline(struct task_struct *tsk)
561 562 563 564 565 566 567
{
	if (atomic_read(&trace_record_cmdline_disabled))
		return;

	trace_save_cmdline(tsk);
}

I
Ingo Molnar 已提交
568
static inline struct list_head *
569 570 571 572 573 574 575 576 577 578 579 580 581
trace_next_list(struct trace_array_cpu *data, struct list_head *next)
{
	/*
	 * Roundrobin - but skip the head (which is not a real page):
	 */
	next = next->next;
	if (unlikely(next == &data->trace_pages))
		next = next->next;
	BUG_ON(next == &data->trace_pages);

	return next;
}

I
Ingo Molnar 已提交
582
static inline void *
583 584 585 586 587 588 589 590 591 592 593 594 595
trace_next_page(struct trace_array_cpu *data, void *addr)
{
	struct list_head *next;
	struct page *page;

	page = virt_to_page(addr);

	next = trace_next_list(data, &page->lru);
	page = list_entry(next, struct page, lru);

	return page_address(page);
}

I
Ingo Molnar 已提交
596
static inline struct trace_entry *
I
Ingo Molnar 已提交
597
tracing_get_trace_entry(struct trace_array *tr, struct trace_array_cpu *data)
598 599 600 601
{
	unsigned long idx, idx_next;
	struct trace_entry *entry;

602
	data->trace_idx++;
603
	idx = data->trace_head_idx;
604 605
	idx_next = idx + 1;

I
Ingo Molnar 已提交
606 607
	BUG_ON(idx * TRACE_ENTRY_SIZE >= PAGE_SIZE);

608
	entry = data->trace_head + idx * TRACE_ENTRY_SIZE;
609 610

	if (unlikely(idx_next >= ENTRIES_PER_PAGE)) {
611
		data->trace_head = trace_next_page(data, data->trace_head);
612 613 614
		idx_next = 0;
	}

615 616 617 618 619 620 621 622 623 624 625 626
	if (data->trace_head == data->trace_tail &&
	    idx_next == data->trace_tail_idx) {
		/* overrun */
		data->trace_tail_idx++;
		if (data->trace_tail_idx >= ENTRIES_PER_PAGE) {
			data->trace_tail =
				trace_next_page(data, data->trace_tail);
			data->trace_tail_idx = 0;
		}
	}

	data->trace_head_idx = idx_next;
627 628 629 630

	return entry;
}

I
Ingo Molnar 已提交
631
static inline void
I
Ingo Molnar 已提交
632
tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags)
633 634 635 636 637 638
{
	struct task_struct *tsk = current;
	unsigned long pc;

	pc = preempt_count();

I
Ingo Molnar 已提交
639 640
	entry->preempt_count	= pc & 0xff;
	entry->pid		= tsk->pid;
I
Ingo Molnar 已提交
641
	entry->t		= ftrace_now(raw_smp_processor_id());
642 643 644 645 646 647
	entry->flags = (irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) |
		((pc & HARDIRQ_MASK) ? TRACE_FLAG_HARDIRQ : 0) |
		((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) |
		(need_resched() ? TRACE_FLAG_NEED_RESCHED : 0);
}

I
Ingo Molnar 已提交
648
void
649 650
trace_function(struct trace_array *tr, struct trace_array_cpu *data,
	       unsigned long ip, unsigned long parent_ip, unsigned long flags)
651 652
{
	struct trace_entry *entry;
I
Ingo Molnar 已提交
653
	unsigned long irq_flags;
654

I
Ingo Molnar 已提交
655
	spin_lock_irqsave(&data->lock, irq_flags);
I
Ingo Molnar 已提交
656
	entry			= tracing_get_trace_entry(tr, data);
657
	tracing_generic_entry_update(entry, flags);
I
Ingo Molnar 已提交
658 659 660
	entry->type		= TRACE_FN;
	entry->fn.ip		= ip;
	entry->fn.parent_ip	= parent_ip;
I
Ingo Molnar 已提交
661
	spin_unlock_irqrestore(&data->lock, irq_flags);
662 663
}

I
Ingo Molnar 已提交
664
void
I
Ingo Molnar 已提交
665 666 667 668
ftrace(struct trace_array *tr, struct trace_array_cpu *data,
       unsigned long ip, unsigned long parent_ip, unsigned long flags)
{
	if (likely(!atomic_read(&data->disabled)))
669
		trace_function(tr, data, ip, parent_ip, flags);
I
Ingo Molnar 已提交
670 671
}

I
Ingo Molnar 已提交
672
void
673 674
__trace_special(void *__tr, void *__data,
		unsigned long arg1, unsigned long arg2, unsigned long arg3)
I
Ingo Molnar 已提交
675
{
676 677
	struct trace_array_cpu *data = __data;
	struct trace_array *tr = __tr;
I
Ingo Molnar 已提交
678
	struct trace_entry *entry;
I
Ingo Molnar 已提交
679
	unsigned long irq_flags;
I
Ingo Molnar 已提交
680

I
Ingo Molnar 已提交
681
	spin_lock_irqsave(&data->lock, irq_flags);
I
Ingo Molnar 已提交
682 683 684 685 686 687
	entry			= tracing_get_trace_entry(tr, data);
	tracing_generic_entry_update(entry, 0);
	entry->type		= TRACE_SPECIAL;
	entry->special.arg1	= arg1;
	entry->special.arg2	= arg2;
	entry->special.arg3	= arg3;
I
Ingo Molnar 已提交
688
	spin_unlock_irqrestore(&data->lock, irq_flags);
I
Ingo Molnar 已提交
689 690

	trace_wake_up();
I
Ingo Molnar 已提交
691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715
}

void __trace_stack(struct trace_array *tr,
		   struct trace_array_cpu *data,
		   unsigned long flags,
		   int skip)
{
	struct trace_entry *entry;
	struct stack_trace trace;

	if (!(trace_flags & TRACE_ITER_STACKTRACE))
		return;

	entry			= tracing_get_trace_entry(tr, data);
	tracing_generic_entry_update(entry, flags);
	entry->type		= TRACE_STACK;

	memset(&entry->stack, 0, sizeof(entry->stack));

	trace.nr_entries	= 0;
	trace.max_entries	= FTRACE_STACK_ENTRIES;
	trace.skip		= skip;
	trace.entries		= entry->stack.caller;

	save_stack_trace(&trace);
I
Ingo Molnar 已提交
716 717
}

I
Ingo Molnar 已提交
718
void
719 720
tracing_sched_switch_trace(struct trace_array *tr,
			   struct trace_array_cpu *data,
I
Ingo Molnar 已提交
721 722
			   struct task_struct *prev,
			   struct task_struct *next,
723 724 725
			   unsigned long flags)
{
	struct trace_entry *entry;
I
Ingo Molnar 已提交
726
	unsigned long irq_flags;
727

I
Ingo Molnar 已提交
728
	spin_lock_irqsave(&data->lock, irq_flags);
I
Ingo Molnar 已提交
729
	entry			= tracing_get_trace_entry(tr, data);
730 731 732 733 734 735 736
	tracing_generic_entry_update(entry, flags);
	entry->type		= TRACE_CTX;
	entry->ctx.prev_pid	= prev->pid;
	entry->ctx.prev_prio	= prev->prio;
	entry->ctx.prev_state	= prev->state;
	entry->ctx.next_pid	= next->pid;
	entry->ctx.next_prio	= next->prio;
P
Peter Zijlstra 已提交
737
	entry->ctx.next_state	= next->state;
I
Ingo Molnar 已提交
738
	__trace_stack(tr, data, flags, 4);
I
Ingo Molnar 已提交
739
	spin_unlock_irqrestore(&data->lock, irq_flags);
740 741
}

742 743 744
void
tracing_sched_wakeup_trace(struct trace_array *tr,
			   struct trace_array_cpu *data,
I
Ingo Molnar 已提交
745 746
			   struct task_struct *wakee,
			   struct task_struct *curr,
747 748 749 750 751 752 753 754 755 756 757 758 759 760
			   unsigned long flags)
{
	struct trace_entry *entry;
	unsigned long irq_flags;

	spin_lock_irqsave(&data->lock, irq_flags);
	entry			= tracing_get_trace_entry(tr, data);
	tracing_generic_entry_update(entry, flags);
	entry->type		= TRACE_WAKE;
	entry->ctx.prev_pid	= curr->pid;
	entry->ctx.prev_prio	= curr->prio;
	entry->ctx.prev_state	= curr->state;
	entry->ctx.next_pid	= wakee->pid;
	entry->ctx.next_prio	= wakee->prio;
P
Peter Zijlstra 已提交
761
	entry->ctx.next_state	= wakee->state;
I
Ingo Molnar 已提交
762
	__trace_stack(tr, data, flags, 5);
763
	spin_unlock_irqrestore(&data->lock, irq_flags);
I
Ingo Molnar 已提交
764 765

	trace_wake_up();
766 767
}

I
Ingo Molnar 已提交
768
#ifdef CONFIG_FTRACE
I
Ingo Molnar 已提交
769
static void
I
Ingo Molnar 已提交
770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786
function_trace_call(unsigned long ip, unsigned long parent_ip)
{
	struct trace_array *tr = &global_trace;
	struct trace_array_cpu *data;
	unsigned long flags;
	long disabled;
	int cpu;

	if (unlikely(!tracer_enabled))
		return;

	local_irq_save(flags);
	cpu = raw_smp_processor_id();
	data = tr->data[cpu];
	disabled = atomic_inc_return(&data->disabled);

	if (likely(disabled == 1))
787
		trace_function(tr, data, ip, parent_ip, flags);
I
Ingo Molnar 已提交
788 789 790 791 792 793 794 795 796 797

	atomic_dec(&data->disabled);
	local_irq_restore(flags);
}

static struct ftrace_ops trace_ops __read_mostly =
{
	.func = function_trace_call,
};

I
Ingo Molnar 已提交
798
void tracing_start_function_trace(void)
I
Ingo Molnar 已提交
799 800 801 802
{
	register_ftrace_function(&trace_ops);
}

I
Ingo Molnar 已提交
803
void tracing_stop_function_trace(void)
I
Ingo Molnar 已提交
804 805 806 807 808
{
	unregister_ftrace_function(&trace_ops);
}
#endif

809 810 811 812 813
enum trace_file_type {
	TRACE_FILE_LAT_FMT	= 1,
};

static struct trace_entry *
814 815
trace_entry_idx(struct trace_array *tr, struct trace_array_cpu *data,
		struct trace_iterator *iter, int cpu)
816
{
817 818
	struct page *page;
	struct trace_entry *array;
819

820
	if (iter->next_idx[cpu] >= tr->entries ||
821 822 823
	    iter->next_idx[cpu] >= data->trace_idx ||
	    (data->trace_head == data->trace_tail &&
	     data->trace_head_idx == data->trace_tail_idx))
824 825
		return NULL;

826
	if (!iter->next_page[cpu]) {
827 828 829 830 831
		/* Initialize the iterator for this cpu trace buffer */
		WARN_ON(!data->trace_tail);
		page = virt_to_page(data->trace_tail);
		iter->next_page[cpu] = &page->lru;
		iter->next_page_idx[cpu] = data->trace_tail_idx;
832
	}
833

834
	page = list_entry(iter->next_page[cpu], struct page, lru);
I
Ingo Molnar 已提交
835 836
	BUG_ON(&data->trace_pages == &page->lru);

837 838
	array = page_address(page);

839
	WARN_ON(iter->next_page_idx[cpu] >= ENTRIES_PER_PAGE);
840
	return &array[iter->next_page_idx[cpu]];
841 842
}

I
Ingo Molnar 已提交
843
static struct trace_entry *
844 845 846 847 848 849 850 851
find_next_entry(struct trace_iterator *iter, int *ent_cpu)
{
	struct trace_array *tr = iter->tr;
	struct trace_entry *ent, *next = NULL;
	int next_cpu = -1;
	int cpu;

	for_each_possible_cpu(cpu) {
I
Ingo Molnar 已提交
852
		if (!head_page(tr->data[cpu]))
853
			continue;
854
		ent = trace_entry_idx(tr, tr->data[cpu], iter, cpu);
I
Ingo Molnar 已提交
855 856 857 858
		/*
		 * Pick the entry with the smallest timestamp:
		 */
		if (ent && (!next || ent->t < next->t)) {
859 860 861 862 863 864 865 866 867 868 869
			next = ent;
			next_cpu = cpu;
		}
	}

	if (ent_cpu)
		*ent_cpu = next_cpu;

	return next;
}

I
Ingo Molnar 已提交
870
static void trace_iterator_increment(struct trace_iterator *iter)
871
{
872 873 874
	iter->idx++;
	iter->next_idx[iter->cpu]++;
	iter->next_page_idx[iter->cpu]++;
I
Ingo Molnar 已提交
875

876 877
	if (iter->next_page_idx[iter->cpu] >= ENTRIES_PER_PAGE) {
		struct trace_array_cpu *data = iter->tr->data[iter->cpu];
878

879 880 881 882 883
		iter->next_page_idx[iter->cpu] = 0;
		iter->next_page[iter->cpu] =
			trace_next_list(data, iter->next_page[iter->cpu]);
	}
}
884

I
Ingo Molnar 已提交
885
static void trace_consume(struct trace_iterator *iter)
886 887 888 889 890 891 892 893
{
	struct trace_array_cpu *data = iter->tr->data[iter->cpu];

	data->trace_tail_idx++;
	if (data->trace_tail_idx >= ENTRIES_PER_PAGE) {
		data->trace_tail = trace_next_page(data, data->trace_tail);
		data->trace_tail_idx = 0;
	}
I
Ingo Molnar 已提交
894

895 896 897 898 899 900
	/* Check if we empty it, then reset the index */
	if (data->trace_head == data->trace_tail &&
	    data->trace_head_idx == data->trace_tail_idx)
		data->trace_idx = 0;
}

I
Ingo Molnar 已提交
901
static void *find_next_entry_inc(struct trace_iterator *iter)
902 903 904 905 906
{
	struct trace_entry *next;
	int next_cpu = -1;

	next = find_next_entry(iter, &next_cpu);
907

I
Ingo Molnar 已提交
908 909 910
	iter->prev_ent = iter->ent;
	iter->prev_cpu = iter->cpu;

911 912 913
	iter->ent = next;
	iter->cpu = next_cpu;

914 915 916
	if (next)
		trace_iterator_increment(iter);

917 918 919
	return next ? iter : NULL;
}

I
Ingo Molnar 已提交
920
static void *s_next(struct seq_file *m, void *v, loff_t *pos)
921 922 923 924
{
	struct trace_iterator *iter = m->private;
	void *last_ent = iter->ent;
	int i = (int)*pos;
I
Ingo Molnar 已提交
925
	void *ent;
926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970

	(*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;

	if (last_ent && !ent)
		seq_puts(m, "\n\nvim:ft=help\n");

	return ent;
}

static void *s_start(struct seq_file *m, loff_t *pos)
{
	struct trace_iterator *iter = m->private;
	void *p = NULL;
	loff_t l = 0;
	int i;

	mutex_lock(&trace_types_lock);

	if (!current_trace || current_trace != iter->trace)
		return NULL;

	atomic_inc(&trace_record_cmdline_disabled);

	/* let the tracer grab locks here if needed */
	if (current_trace->start)
		current_trace->start(iter);

	if (*pos != iter->pos) {
		iter->ent = NULL;
		iter->cpu = 0;
		iter->idx = -1;
I
Ingo Molnar 已提交
971 972
		iter->prev_ent = NULL;
		iter->prev_cpu = -1;
973

974
		for_each_possible_cpu(i) {
975
			iter->next_idx[i] = 0;
976 977
			iter->next_page[i] = NULL;
		}
978 979 980 981 982

		for (p = iter; p && l < *pos; p = s_next(m, p, &l))
			;

	} else {
983
		l = *pos - 1;
984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002
		p = s_next(m, p, &l);
	}

	return p;
}

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

	atomic_dec(&trace_record_cmdline_disabled);

	/* let the tracer release locks here if needed */
	if (current_trace && current_trace == iter->trace && iter->trace->stop)
		iter->trace->stop(iter);

	mutex_unlock(&trace_types_lock);
}

1003
static int
S
Steven Rostedt 已提交
1004
seq_print_sym_short(struct trace_seq *s, const char *fmt, unsigned long address)
1005 1006 1007 1008 1009 1010
{
#ifdef CONFIG_KALLSYMS
	char str[KSYM_SYMBOL_LEN];

	kallsyms_lookup(address, NULL, NULL, NULL, str);

1011
	return trace_seq_printf(s, fmt, str);
1012
#endif
1013
	return 1;
1014 1015
}

1016
static int
S
Steven Rostedt 已提交
1017 1018
seq_print_sym_offset(struct trace_seq *s, const char *fmt,
		     unsigned long address)
1019 1020 1021 1022 1023
{
#ifdef CONFIG_KALLSYMS
	char str[KSYM_SYMBOL_LEN];

	sprint_symbol(str, address);
1024
	return trace_seq_printf(s, fmt, str);
1025
#endif
1026
	return 1;
1027 1028 1029 1030 1031 1032 1033 1034
}

#ifndef CONFIG_64BIT
# define IP_FMT "%08lx"
#else
# define IP_FMT "%016lx"
#endif

I
Ingo Molnar 已提交
1035
static int
S
Steven Rostedt 已提交
1036
seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags)
1037
{
1038 1039 1040 1041
	int ret;

	if (!ip)
		return trace_seq_printf(s, "0");
1042 1043

	if (sym_flags & TRACE_ITER_SYM_OFFSET)
1044
		ret = seq_print_sym_offset(s, "%s", ip);
1045
	else
1046 1047 1048 1049
		ret = seq_print_sym_short(s, "%s", ip);

	if (!ret)
		return 0;
1050 1051

	if (sym_flags & TRACE_ITER_SYM_ADDR)
1052 1053
		ret = trace_seq_printf(s, " <" IP_FMT ">", ip);
	return ret;
1054 1055
}

I
Ingo Molnar 已提交
1056
static void print_lat_help_header(struct seq_file *m)
1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068
{
	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");
	seq_puts(m, "#              |||| /                      \n");
	seq_puts(m, "#              |||||     delay             \n");
	seq_puts(m, "#  cmd     pid ||||| time  |   caller      \n");
	seq_puts(m, "#     \\   /    |||||   \\   |   /           \n");
}

I
Ingo Molnar 已提交
1069
static void print_func_help_header(struct seq_file *m)
1070 1071 1072 1073 1074 1075
{
	seq_puts(m, "#           TASK-PID   CPU#    TIMESTAMP  FUNCTION\n");
	seq_puts(m, "#              | |      |          |         |\n");
}


I
Ingo Molnar 已提交
1076
static void
1077 1078 1079 1080 1081 1082
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;
1083 1084
	unsigned long total   = 0;
	unsigned long entries = 0;
1085 1086 1087 1088 1089 1090 1091
	int cpu;
	const char *name = "preemption";

	if (type)
		name = type->name;

	for_each_possible_cpu(cpu) {
I
Ingo Molnar 已提交
1092
		if (head_page(tr->data[cpu])) {
1093 1094
			total += tr->data[cpu]->trace_idx;
			if (tr->data[cpu]->trace_idx > tr->entries)
1095
				entries += tr->entries;
1096
			else
1097 1098 1099 1100 1101 1102 1103 1104 1105 1106
				entries += tr->data[cpu]->trace_idx;
		}
	}

	seq_printf(m, "%s latency trace v1.1.5 on %s\n",
		   name, UTS_RELEASE);
	seq_puts(m, "-----------------------------------"
		 "---------------------------------\n");
	seq_printf(m, " latency: %lu us, #%lu/%lu, CPU#%d |"
		   " (M:%s VP:%d, KP:%d, SP:%d HP:%d",
S
Steven Rostedt 已提交
1107
		   nsecs_to_usecs(data->saved_latency),
1108
		   entries,
1109
		   total,
1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135
		   tr->cpu,
#if defined(CONFIG_PREEMPT_NONE)
		   "server",
#elif defined(CONFIG_PREEMPT_VOLUNTARY)
		   "desktop",
#elif defined(CONFIG_PREEMPT_DESKTOP)
		   "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
	seq_puts(m, "    -----------------\n");
	seq_printf(m, "    | task: %.16s-%d "
		   "(uid:%d nice:%ld policy:%ld rt_prio:%ld)\n",
		   data->comm, data->pid, data->uid, data->nice,
		   data->policy, data->rt_priority);
	seq_puts(m, "    -----------------\n");

	if (data->critical_start) {
		seq_puts(m, " => started at: ");
S
Steven Rostedt 已提交
1136 1137
		seq_print_ip_sym(&iter->seq, data->critical_start, sym_flags);
		trace_print_seq(m, &iter->seq);
1138
		seq_puts(m, "\n => ended at:   ");
S
Steven Rostedt 已提交
1139 1140
		seq_print_ip_sym(&iter->seq, data->critical_end, sym_flags);
		trace_print_seq(m, &iter->seq);
1141 1142 1143 1144 1145 1146
		seq_puts(m, "\n");
	}

	seq_puts(m, "\n");
}

I
Ingo Molnar 已提交
1147
static void
S
Steven Rostedt 已提交
1148
lat_print_generic(struct trace_seq *s, struct trace_entry *entry, int cpu)
1149 1150 1151 1152 1153 1154
{
	int hardirq, softirq;
	char *comm;

	comm = trace_find_cmdline(entry->pid);

S
Steven Rostedt 已提交
1155 1156 1157 1158 1159
	trace_seq_printf(s, "%8.8s-%-5d ", comm, entry->pid);
	trace_seq_printf(s, "%d", cpu);
	trace_seq_printf(s, "%c%c",
			(entry->flags & TRACE_FLAG_IRQS_OFF) ? 'd' : '.',
			((entry->flags & TRACE_FLAG_NEED_RESCHED) ? 'N' : '.'));
1160 1161 1162 1163

	hardirq = entry->flags & TRACE_FLAG_HARDIRQ;
	softirq = entry->flags & TRACE_FLAG_SOFTIRQ;
	if (hardirq && softirq)
S
Steven Rostedt 已提交
1164
		trace_seq_putc(s, 'H');
1165 1166
	else {
		if (hardirq)
S
Steven Rostedt 已提交
1167
			trace_seq_putc(s, 'h');
1168 1169
		else {
			if (softirq)
S
Steven Rostedt 已提交
1170
				trace_seq_putc(s, 's');
1171
			else
S
Steven Rostedt 已提交
1172
				trace_seq_putc(s, '.');
1173 1174 1175 1176
		}
	}

	if (entry->preempt_count)
S
Steven Rostedt 已提交
1177
		trace_seq_printf(s, "%x", entry->preempt_count);
1178
	else
S
Steven Rostedt 已提交
1179
		trace_seq_puts(s, ".");
1180 1181 1182 1183
}

unsigned long preempt_mark_thresh = 100;

I
Ingo Molnar 已提交
1184
static void
S
Steven Rostedt 已提交
1185
lat_print_timestamp(struct trace_seq *s, unsigned long long abs_usecs,
1186 1187
		    unsigned long rel_usecs)
{
S
Steven Rostedt 已提交
1188
	trace_seq_printf(s, " %4lldus", abs_usecs);
1189
	if (rel_usecs > preempt_mark_thresh)
S
Steven Rostedt 已提交
1190
		trace_seq_puts(s, "!: ");
1191
	else if (rel_usecs > 1)
S
Steven Rostedt 已提交
1192
		trace_seq_puts(s, "+: ");
1193
	else
S
Steven Rostedt 已提交
1194
		trace_seq_puts(s, " : ");
1195 1196 1197 1198
}

static const char state_to_char[] = TASK_STATE_TO_CHAR_STR;

I
Ingo Molnar 已提交
1199
static int
S
Steven Rostedt 已提交
1200
print_lat_fmt(struct trace_iterator *iter, unsigned int trace_idx, int cpu)
1201
{
S
Steven Rostedt 已提交
1202
	struct trace_seq *s = &iter->seq;
1203 1204 1205 1206 1207 1208 1209
	unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK);
	struct trace_entry *next_entry = find_next_entry(iter, NULL);
	unsigned long verbose = (trace_flags & TRACE_ITER_VERBOSE);
	struct trace_entry *entry = iter->ent;
	unsigned long abs_usecs;
	unsigned long rel_usecs;
	char *comm;
P
Peter Zijlstra 已提交
1210
	int S, T;
I
Ingo Molnar 已提交
1211
	int i;
1212 1213 1214 1215 1216 1217 1218 1219

	if (!next_entry)
		next_entry = entry;
	rel_usecs = ns2usecs(next_entry->t - entry->t);
	abs_usecs = ns2usecs(entry->t - iter->tr->time_start);

	if (verbose) {
		comm = trace_find_cmdline(entry->pid);
S
Steven Rostedt 已提交
1220 1221 1222 1223 1224 1225 1226 1227 1228
		trace_seq_printf(s, "%16s %5d %d %d %08x %08x [%08lx]"
				 " %ld.%03ldms (+%ld.%03ldms): ",
				 comm,
				 entry->pid, cpu, entry->flags,
				 entry->preempt_count, trace_idx,
				 ns2usecs(entry->t),
				 abs_usecs/1000,
				 abs_usecs % 1000, rel_usecs/1000,
				 rel_usecs % 1000);
1229
	} else {
I
Ingo Molnar 已提交
1230 1231
		lat_print_generic(s, entry, cpu);
		lat_print_timestamp(s, abs_usecs, rel_usecs);
1232 1233 1234
	}
	switch (entry->type) {
	case TRACE_FN:
S
Steven Rostedt 已提交
1235 1236 1237 1238
		seq_print_ip_sym(s, entry->fn.ip, sym_flags);
		trace_seq_puts(s, " (");
		seq_print_ip_sym(s, entry->fn.parent_ip, sym_flags);
		trace_seq_puts(s, ")\n");
1239 1240
		break;
	case TRACE_CTX:
1241
	case TRACE_WAKE:
1242 1243
		S = entry->ctx.prev_state < sizeof(state_to_char) ?
			state_to_char[entry->ctx.prev_state] : 'X';
P
Peter Zijlstra 已提交
1244 1245 1246
		T = entry->ctx.next_state < sizeof(state_to_char) ?
			state_to_char[entry->ctx.next_state] : 'X';

1247
		comm = trace_find_cmdline(entry->ctx.next_pid);
P
Peter Zijlstra 已提交
1248
		trace_seq_printf(s, " %5d:%3d:%c %s %5d:%3d:%c %s\n",
S
Steven Rostedt 已提交
1249 1250
				 entry->ctx.prev_pid,
				 entry->ctx.prev_prio,
1251
				 S, entry->type == TRACE_CTX ? "==>" : "  +",
S
Steven Rostedt 已提交
1252 1253
				 entry->ctx.next_pid,
				 entry->ctx.next_prio,
P
Peter Zijlstra 已提交
1254
				 T, comm);
1255
		break;
I
Ingo Molnar 已提交
1256
	case TRACE_SPECIAL:
I
Ingo Molnar 已提交
1257
		trace_seq_printf(s, "# %ld %ld %ld\n",
I
Ingo Molnar 已提交
1258 1259 1260 1261
				 entry->special.arg1,
				 entry->special.arg2,
				 entry->special.arg3);
		break;
I
Ingo Molnar 已提交
1262 1263 1264 1265 1266 1267 1268 1269
	case TRACE_STACK:
		for (i = 0; i < FTRACE_STACK_ENTRIES; i++) {
			if (i)
				trace_seq_puts(s, " <= ");
			seq_print_ip_sym(s, entry->stack.caller[i], sym_flags);
		}
		trace_seq_puts(s, "\n");
		break;
1270
	default:
S
Steven Rostedt 已提交
1271
		trace_seq_printf(s, "Unknown type %d\n", entry->type);
1272
	}
I
Ingo Molnar 已提交
1273
	return 1;
1274 1275
}

I
Ingo Molnar 已提交
1276
static int print_trace_fmt(struct trace_iterator *iter)
1277
{
S
Steven Rostedt 已提交
1278
	struct trace_seq *s = &iter->seq;
1279
	unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK);
I
Ingo Molnar 已提交
1280
	struct trace_entry *entry;
1281 1282 1283 1284
	unsigned long usec_rem;
	unsigned long long t;
	unsigned long secs;
	char *comm;
1285
	int ret;
P
Peter Zijlstra 已提交
1286
	int S, T;
I
Ingo Molnar 已提交
1287
	int i;
1288

I
Ingo Molnar 已提交
1289 1290
	entry = iter->ent;

1291 1292
	comm = trace_find_cmdline(iter->ent->pid);

I
Ingo Molnar 已提交
1293
	t = ns2usecs(entry->t);
1294 1295 1296
	usec_rem = do_div(t, 1000000ULL);
	secs = (unsigned long)t;

I
Ingo Molnar 已提交
1297 1298 1299 1300 1301 1302 1303 1304 1305
	ret = trace_seq_printf(s, "%16s-%-5d ", comm, entry->pid);
	if (!ret)
		return 0;
	ret = trace_seq_printf(s, "[%02d] ", iter->cpu);
	if (!ret)
		return 0;
	ret = trace_seq_printf(s, "%5lu.%06lu: ", secs, usec_rem);
	if (!ret)
		return 0;
1306 1307 1308

	switch (entry->type) {
	case TRACE_FN:
1309 1310 1311
		ret = seq_print_ip_sym(s, entry->fn.ip, sym_flags);
		if (!ret)
			return 0;
1312 1313
		if ((sym_flags & TRACE_ITER_PRINT_PARENT) &&
						entry->fn.parent_ip) {
1314 1315 1316 1317 1318 1319 1320
			ret = trace_seq_printf(s, " <-");
			if (!ret)
				return 0;
			ret = seq_print_ip_sym(s, entry->fn.parent_ip,
					       sym_flags);
			if (!ret)
				return 0;
1321
		}
1322 1323 1324
		ret = trace_seq_printf(s, "\n");
		if (!ret)
			return 0;
1325 1326
		break;
	case TRACE_CTX:
1327
	case TRACE_WAKE:
1328 1329
		S = entry->ctx.prev_state < sizeof(state_to_char) ?
			state_to_char[entry->ctx.prev_state] : 'X';
P
Peter Zijlstra 已提交
1330 1331 1332
		T = entry->ctx.next_state < sizeof(state_to_char) ?
			state_to_char[entry->ctx.next_state] : 'X';
		ret = trace_seq_printf(s, " %5d:%3d:%c %s %5d:%3d:%c\n",
1333 1334 1335
				       entry->ctx.prev_pid,
				       entry->ctx.prev_prio,
				       S,
1336
				       entry->type == TRACE_CTX ? "==>" : "  +",
1337
				       entry->ctx.next_pid,
P
Peter Zijlstra 已提交
1338 1339
				       entry->ctx.next_prio,
				       T);
1340 1341
		if (!ret)
			return 0;
1342
		break;
I
Ingo Molnar 已提交
1343
	case TRACE_SPECIAL:
I
Ingo Molnar 已提交
1344
		ret = trace_seq_printf(s, "# %ld %ld %ld\n",
I
Ingo Molnar 已提交
1345 1346 1347 1348 1349 1350
				 entry->special.arg1,
				 entry->special.arg2,
				 entry->special.arg3);
		if (!ret)
			return 0;
		break;
I
Ingo Molnar 已提交
1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
	case TRACE_STACK:
		for (i = 0; i < FTRACE_STACK_ENTRIES; i++) {
			if (i) {
				ret = trace_seq_puts(s, " <= ");
				if (!ret)
					return 0;
			}
			ret = seq_print_ip_sym(s, entry->stack.caller[i],
					       sym_flags);
			if (!ret)
				return 0;
		}
		ret = trace_seq_puts(s, "\n");
		if (!ret)
			return 0;
		break;
1367
	}
1368
	return 1;
1369 1370
}

I
Ingo Molnar 已提交
1371
static int print_raw_fmt(struct trace_iterator *iter)
I
Ingo Molnar 已提交
1372 1373 1374 1375
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *entry;
	int ret;
P
Peter Zijlstra 已提交
1376
	int S, T;
I
Ingo Molnar 已提交
1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392

	entry = iter->ent;

	ret = trace_seq_printf(s, "%d %d %llu ",
		entry->pid, iter->cpu, entry->t);
	if (!ret)
		return 0;

	switch (entry->type) {
	case TRACE_FN:
		ret = trace_seq_printf(s, "%x %x\n",
					entry->fn.ip, entry->fn.parent_ip);
		if (!ret)
			return 0;
		break;
	case TRACE_CTX:
1393
	case TRACE_WAKE:
I
Ingo Molnar 已提交
1394 1395
		S = entry->ctx.prev_state < sizeof(state_to_char) ?
			state_to_char[entry->ctx.prev_state] : 'X';
P
Peter Zijlstra 已提交
1396 1397
		T = entry->ctx.next_state < sizeof(state_to_char) ?
			state_to_char[entry->ctx.next_state] : 'X';
1398 1399
		if (entry->type == TRACE_WAKE)
			S = '+';
P
Peter Zijlstra 已提交
1400
		ret = trace_seq_printf(s, "%d %d %c %d %d %c\n",
I
Ingo Molnar 已提交
1401 1402 1403 1404
				       entry->ctx.prev_pid,
				       entry->ctx.prev_prio,
				       S,
				       entry->ctx.next_pid,
P
Peter Zijlstra 已提交
1405 1406
				       entry->ctx.next_prio,
				       T);
I
Ingo Molnar 已提交
1407 1408 1409
		if (!ret)
			return 0;
		break;
I
Ingo Molnar 已提交
1410
	case TRACE_SPECIAL:
I
Ingo Molnar 已提交
1411
	case TRACE_STACK:
I
Ingo Molnar 已提交
1412
		ret = trace_seq_printf(s, "# %ld %ld %ld\n",
I
Ingo Molnar 已提交
1413 1414 1415 1416 1417 1418
				 entry->special.arg1,
				 entry->special.arg2,
				 entry->special.arg3);
		if (!ret)
			return 0;
		break;
I
Ingo Molnar 已提交
1419 1420 1421 1422
	}
	return 1;
}

I
Ingo Molnar 已提交
1423 1424 1425 1426 1427 1428
#define SEQ_PUT_FIELD_RET(s, x)				\
do {							\
	if (!trace_seq_putmem(s, &(x), sizeof(x)))	\
		return 0;				\
} while (0)

1429 1430 1431 1432 1433 1434
#define SEQ_PUT_HEX_FIELD_RET(s, x)			\
do {							\
	if (!trace_seq_putmem_hex(s, &(x), sizeof(x)))	\
		return 0;				\
} while (0)

I
Ingo Molnar 已提交
1435
static int print_hex_fmt(struct trace_iterator *iter)
1436 1437 1438 1439
{
	struct trace_seq *s = &iter->seq;
	unsigned char newline = '\n';
	struct trace_entry *entry;
P
Peter Zijlstra 已提交
1440
	int S, T;
1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453

	entry = iter->ent;

	SEQ_PUT_HEX_FIELD_RET(s, entry->pid);
	SEQ_PUT_HEX_FIELD_RET(s, iter->cpu);
	SEQ_PUT_HEX_FIELD_RET(s, entry->t);

	switch (entry->type) {
	case TRACE_FN:
		SEQ_PUT_HEX_FIELD_RET(s, entry->fn.ip);
		SEQ_PUT_HEX_FIELD_RET(s, entry->fn.parent_ip);
		break;
	case TRACE_CTX:
1454
	case TRACE_WAKE:
1455 1456
		S = entry->ctx.prev_state < sizeof(state_to_char) ?
			state_to_char[entry->ctx.prev_state] : 'X';
P
Peter Zijlstra 已提交
1457 1458
		T = entry->ctx.next_state < sizeof(state_to_char) ?
			state_to_char[entry->ctx.next_state] : 'X';
1459 1460
		if (entry->type == TRACE_WAKE)
			S = '+';
1461 1462 1463 1464 1465 1466
		SEQ_PUT_HEX_FIELD_RET(s, entry->ctx.prev_pid);
		SEQ_PUT_HEX_FIELD_RET(s, entry->ctx.prev_prio);
		SEQ_PUT_HEX_FIELD_RET(s, S);
		SEQ_PUT_HEX_FIELD_RET(s, entry->ctx.next_pid);
		SEQ_PUT_HEX_FIELD_RET(s, entry->ctx.next_prio);
		SEQ_PUT_HEX_FIELD_RET(s, entry->fn.parent_ip);
P
Peter Zijlstra 已提交
1467
		SEQ_PUT_HEX_FIELD_RET(s, T);
1468 1469
		break;
	case TRACE_SPECIAL:
I
Ingo Molnar 已提交
1470
	case TRACE_STACK:
1471 1472 1473 1474 1475 1476 1477 1478 1479 1480
		SEQ_PUT_HEX_FIELD_RET(s, entry->special.arg1);
		SEQ_PUT_HEX_FIELD_RET(s, entry->special.arg2);
		SEQ_PUT_HEX_FIELD_RET(s, entry->special.arg3);
		break;
	}
	SEQ_PUT_FIELD_RET(s, newline);

	return 1;
}

I
Ingo Molnar 已提交
1481
static int print_bin_fmt(struct trace_iterator *iter)
I
Ingo Molnar 已提交
1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *entry;

	entry = iter->ent;

	SEQ_PUT_FIELD_RET(s, entry->pid);
	SEQ_PUT_FIELD_RET(s, entry->cpu);
	SEQ_PUT_FIELD_RET(s, entry->t);

	switch (entry->type) {
	case TRACE_FN:
		SEQ_PUT_FIELD_RET(s, entry->fn.ip);
		SEQ_PUT_FIELD_RET(s, entry->fn.parent_ip);
		break;
	case TRACE_CTX:
		SEQ_PUT_FIELD_RET(s, entry->ctx.prev_pid);
		SEQ_PUT_FIELD_RET(s, entry->ctx.prev_prio);
		SEQ_PUT_FIELD_RET(s, entry->ctx.prev_state);
		SEQ_PUT_FIELD_RET(s, entry->ctx.next_pid);
		SEQ_PUT_FIELD_RET(s, entry->ctx.next_prio);
P
Peter Zijlstra 已提交
1503
		SEQ_PUT_FIELD_RET(s, entry->ctx.next_state);
I
Ingo Molnar 已提交
1504
		break;
I
Ingo Molnar 已提交
1505
	case TRACE_SPECIAL:
I
Ingo Molnar 已提交
1506
	case TRACE_STACK:
I
Ingo Molnar 已提交
1507 1508 1509 1510
		SEQ_PUT_FIELD_RET(s, entry->special.arg1);
		SEQ_PUT_FIELD_RET(s, entry->special.arg2);
		SEQ_PUT_FIELD_RET(s, entry->special.arg3);
		break;
I
Ingo Molnar 已提交
1511 1512 1513 1514
	}
	return 1;
}

1515 1516 1517 1518 1519 1520 1521 1522
static int trace_empty(struct trace_iterator *iter)
{
	struct trace_array_cpu *data;
	int cpu;

	for_each_possible_cpu(cpu) {
		data = iter->tr->data[cpu];

1523 1524 1525
		if (head_page(data) && data->trace_idx &&
		    (data->trace_tail != data->trace_head ||
		     data->trace_tail_idx != data->trace_head_idx))
1526 1527 1528 1529 1530
			return 0;
	}
	return 1;
}

I
Ingo Molnar 已提交
1531 1532
static int print_trace_line(struct trace_iterator *iter)
{
I
Ingo Molnar 已提交
1533 1534 1535
	if (trace_flags & TRACE_ITER_BIN)
		return print_bin_fmt(iter);

1536 1537 1538
	if (trace_flags & TRACE_ITER_HEX)
		return print_hex_fmt(iter);

I
Ingo Molnar 已提交
1539 1540 1541 1542 1543 1544 1545 1546 1547
	if (trace_flags & TRACE_ITER_RAW)
		return print_raw_fmt(iter);

	if (iter->iter_flags & TRACE_FILE_LAT_FMT)
		return print_lat_fmt(iter, iter->idx, iter->cpu);

	return print_trace_fmt(iter);
}

1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568
static int s_show(struct seq_file *m, void *v)
{
	struct trace_iterator *iter = v;

	if (iter->ent == NULL) {
		if (iter->tr) {
			seq_printf(m, "# tracer: %s\n", iter->trace->name);
			seq_puts(m, "#\n");
		}
		if (iter->iter_flags & TRACE_FILE_LAT_FMT) {
			/* print nothing if the buffers are empty */
			if (trace_empty(iter))
				return 0;
			print_trace_header(m, iter);
			if (!(trace_flags & TRACE_ITER_VERBOSE))
				print_lat_help_header(m);
		} else {
			if (!(trace_flags & TRACE_ITER_VERBOSE))
				print_func_help_header(m);
		}
	} else {
I
Ingo Molnar 已提交
1569
		print_trace_line(iter);
S
Steven Rostedt 已提交
1570
		trace_print_seq(m, &iter->seq);
1571 1572 1573 1574 1575 1576
	}

	return 0;
}

static struct seq_operations tracer_seq_ops = {
I
Ingo Molnar 已提交
1577 1578 1579 1580
	.start		= s_start,
	.next		= s_next,
	.stop		= s_stop,
	.show		= s_show,
1581 1582
};

I
Ingo Molnar 已提交
1583
static struct trace_iterator *
1584 1585 1586 1587
__tracing_open(struct inode *inode, struct file *file, int *ret)
{
	struct trace_iterator *iter;

S
Steven Rostedt 已提交
1588 1589 1590 1591 1592
	if (tracing_disabled) {
		*ret = -ENODEV;
		return NULL;
	}

1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter) {
		*ret = -ENOMEM;
		goto out;
	}

	mutex_lock(&trace_types_lock);
	if (current_trace && current_trace->print_max)
		iter->tr = &max_tr;
	else
		iter->tr = inode->i_private;
	iter->trace = current_trace;
	iter->pos = -1;

	/* TODO stop tracer */
	*ret = seq_open(file, &tracer_seq_ops);
	if (!*ret) {
		struct seq_file *m = file->private_data;
		m->private = iter;

		/* stop the trace while dumping */
		if (iter->tr->ctrl)
			tracer_enabled = 0;

		if (iter->trace && iter->trace->open)
			iter->trace->open(iter);
	} else {
		kfree(iter);
		iter = NULL;
	}
	mutex_unlock(&trace_types_lock);

 out:
	return iter;
}

int tracing_open_generic(struct inode *inode, struct file *filp)
{
S
Steven Rostedt 已提交
1631 1632 1633
	if (tracing_disabled)
		return -ENODEV;

1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679
	filp->private_data = inode->i_private;
	return 0;
}

int tracing_release(struct inode *inode, struct file *file)
{
	struct seq_file *m = (struct seq_file *)file->private_data;
	struct trace_iterator *iter = m->private;

	mutex_lock(&trace_types_lock);
	if (iter->trace && iter->trace->close)
		iter->trace->close(iter);

	/* reenable tracing if it was previously enabled */
	if (iter->tr->ctrl)
		tracer_enabled = 1;
	mutex_unlock(&trace_types_lock);

	seq_release(inode, file);
	kfree(iter);
	return 0;
}

static int tracing_open(struct inode *inode, struct file *file)
{
	int ret;

	__tracing_open(inode, file, &ret);

	return ret;
}

static int tracing_lt_open(struct inode *inode, struct file *file)
{
	struct trace_iterator *iter;
	int ret;

	iter = __tracing_open(inode, file, &ret);

	if (!ret)
		iter->iter_flags |= TRACE_FILE_LAT_FMT;

	return ret;
}


I
Ingo Molnar 已提交
1680
static void *
1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 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
t_next(struct seq_file *m, void *v, loff_t *pos)
{
	struct tracer *t = m->private;

	(*pos)++;

	if (t)
		t = t->next;

	m->private = t;

	return t;
}

static void *t_start(struct seq_file *m, loff_t *pos)
{
	struct tracer *t = m->private;
	loff_t l = 0;

	mutex_lock(&trace_types_lock);
	for (; t && l < *pos; t = t_next(m, t, &l))
		;

	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;
}

static struct seq_operations show_traces_seq_ops = {
I
Ingo Molnar 已提交
1729 1730 1731 1732
	.start		= t_start,
	.next		= t_next,
	.stop		= t_stop,
	.show		= t_show,
1733 1734 1735 1736 1737 1738
};

static int show_traces_open(struct inode *inode, struct file *file)
{
	int ret;

S
Steven Rostedt 已提交
1739 1740 1741
	if (tracing_disabled)
		return -ENODEV;

1742 1743 1744 1745 1746 1747 1748 1749 1750 1751
	ret = seq_open(file, &show_traces_seq_ops);
	if (!ret) {
		struct seq_file *m = file->private_data;
		m->private = trace_types;
	}

	return ret;
}

static struct file_operations tracing_fops = {
I
Ingo Molnar 已提交
1752 1753 1754 1755
	.open		= tracing_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= tracing_release,
1756 1757 1758
};

static struct file_operations tracing_lt_fops = {
I
Ingo Molnar 已提交
1759 1760 1761 1762
	.open		= tracing_lt_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= tracing_release,
1763 1764 1765
};

static struct file_operations show_traces_fops = {
I
Ingo Molnar 已提交
1766 1767 1768 1769 1770
	.open		= show_traces_open,
	.read		= seq_read,
	.release	= seq_release,
};

1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793
/*
 * Only trace on a CPU if the bitmask is set:
 */
static cpumask_t tracing_cpumask = CPU_MASK_ALL;

/*
 * When tracing/tracing_cpu_mask is modified then this holds
 * the new bitmask we are about to install:
 */
static cpumask_t tracing_cpumask_new;

/*
 * 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
1794 1795 1796 1797
static ssize_t
tracing_cpumask_read(struct file *filp, char __user *ubuf,
		     size_t count, loff_t *ppos)
{
1798
	int len;
I
Ingo Molnar 已提交
1799 1800

	mutex_lock(&tracing_cpumask_update_lock);
1801 1802 1803 1804 1805 1806 1807 1808 1809 1810

	len = cpumask_scnprintf(mask_str, count, tracing_cpumask);
	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
1811 1812 1813 1814 1815 1816 1817 1818 1819
	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)
{
1820
	int err, cpu;
I
Ingo Molnar 已提交
1821 1822

	mutex_lock(&tracing_cpumask_update_lock);
1823
	err = cpumask_parse_user(ubuf, count, tracing_cpumask_new);
I
Ingo Molnar 已提交
1824
	if (err)
1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846
		goto err_unlock;

	spin_lock_irq(&ftrace_max_lock);
	for_each_possible_cpu(cpu) {
		/*
		 * Increase/decrease the disabled counter if we are
		 * about to flip a bit in the cpumask:
		 */
		if (cpu_isset(cpu, tracing_cpumask) &&
				!cpu_isset(cpu, tracing_cpumask_new)) {
			atomic_inc(&global_trace.data[cpu]->disabled);
		}
		if (!cpu_isset(cpu, tracing_cpumask) &&
				cpu_isset(cpu, tracing_cpumask_new)) {
			atomic_dec(&global_trace.data[cpu]->disabled);
		}
	}
	spin_unlock_irq(&ftrace_max_lock);

	tracing_cpumask = tracing_cpumask_new;

	mutex_unlock(&tracing_cpumask_update_lock);
I
Ingo Molnar 已提交
1847 1848

	return count;
1849 1850 1851 1852 1853

err_unlock:
	mutex_unlock(&tracing_cpumask_update_lock);

	return err;
I
Ingo Molnar 已提交
1854 1855 1856 1857 1858 1859
}

static struct file_operations tracing_cpumask_fops = {
	.open		= tracing_open_generic,
	.read		= tracing_cpumask_read,
	.write		= tracing_cpumask_write,
1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891
};

static ssize_t
tracing_iter_ctrl_read(struct file *filp, char __user *ubuf,
		       size_t cnt, loff_t *ppos)
{
	char *buf;
	int r = 0;
	int len = 0;
	int i;

	/* calulate max size */
	for (i = 0; trace_options[i]; i++) {
		len += strlen(trace_options[i]);
		len += 3; /* "no" and space */
	}

	/* +2 for \n and \0 */
	buf = kmalloc(len + 2, GFP_KERNEL);
	if (!buf)
		return -ENOMEM;

	for (i = 0; trace_options[i]; i++) {
		if (trace_flags & (1 << i))
			r += sprintf(buf + r, "%s ", trace_options[i]);
		else
			r += sprintf(buf + r, "no%s ", trace_options[i]);
	}

	r += sprintf(buf + r, "\n");
	WARN_ON(r >= len + 2);

1892
	r = simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931

	kfree(buf);

	return r;
}

static ssize_t
tracing_iter_ctrl_write(struct file *filp, const char __user *ubuf,
			size_t cnt, loff_t *ppos)
{
	char buf[64];
	char *cmp = buf;
	int neg = 0;
	int i;

	if (cnt > 63)
		cnt = 63;

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

	buf[cnt] = 0;

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

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

		if (strncmp(cmp, trace_options[i], len) == 0) {
			if (neg)
				trace_flags &= ~(1 << i);
			else
				trace_flags |= (1 << i);
			break;
		}
	}
I
Ingo Molnar 已提交
1932 1933 1934 1935 1936
	/*
	 * If no option could be set, return an error:
	 */
	if (!trace_options[i])
		return -EINVAL;
1937 1938 1939 1940 1941 1942 1943

	filp->f_pos += cnt;

	return cnt;
}

static struct file_operations tracing_iter_fops = {
I
Ingo Molnar 已提交
1944 1945 1946
	.open		= tracing_open_generic,
	.read		= tracing_iter_ctrl_read,
	.write		= tracing_iter_ctrl_write,
1947 1948
};

I
Ingo Molnar 已提交
1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976
static const char readme_msg[] =
	"tracing mini-HOWTO:\n\n"
	"# mkdir /debug\n"
	"# mount -t debugfs nodev /debug\n\n"
	"# cat /debug/tracing/available_tracers\n"
	"wakeup preemptirqsoff preemptoff irqsoff ftrace sched_switch none\n\n"
	"# cat /debug/tracing/current_tracer\n"
	"none\n"
	"# echo sched_switch > /debug/tracing/current_tracer\n"
	"# cat /debug/tracing/current_tracer\n"
	"sched_switch\n"
	"# cat /debug/tracing/iter_ctrl\n"
	"noprint-parent nosym-offset nosym-addr noverbose\n"
	"# echo print-parent > /debug/tracing/iter_ctrl\n"
	"# echo 1 > /debug/tracing/tracing_enabled\n"
	"# cat /debug/tracing/trace > /tmp/trace.txt\n"
	"echo 0 > /debug/tracing/tracing_enabled\n"
;

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));
}

static struct file_operations tracing_readme_fops = {
I
Ingo Molnar 已提交
1977 1978
	.open		= tracing_open_generic,
	.read		= tracing_readme_read,
I
Ingo Molnar 已提交
1979 1980
};

1981 1982 1983 1984 1985 1986 1987 1988 1989
static ssize_t
tracing_ctrl_read(struct file *filp, char __user *ubuf,
		  size_t cnt, loff_t *ppos)
{
	struct trace_array *tr = filp->private_data;
	char buf[64];
	int r;

	r = sprintf(buf, "%ld\n", tr->ctrl);
I
Ingo Molnar 已提交
1990
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045
}

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;
	long val;
	char buf[64];

	if (cnt > 63)
		cnt = 63;

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

	buf[cnt] = 0;

	val = simple_strtoul(buf, NULL, 10);

	val = !!val;

	mutex_lock(&trace_types_lock);
	if (tr->ctrl ^ val) {
		if (val)
			tracer_enabled = 1;
		else
			tracer_enabled = 0;

		tr->ctrl = val;

		if (current_trace && current_trace->ctrl_update)
			current_trace->ctrl_update(tr);
	}
	mutex_unlock(&trace_types_lock);

	filp->f_pos += cnt;

	return cnt;
}

static ssize_t
tracing_set_trace_read(struct file *filp, char __user *ubuf,
		       size_t cnt, loff_t *ppos)
{
	char buf[max_tracer_type_len+2];
	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
2046
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104
}

static ssize_t
tracing_set_trace_write(struct file *filp, const char __user *ubuf,
			size_t cnt, loff_t *ppos)
{
	struct trace_array *tr = &global_trace;
	struct tracer *t;
	char buf[max_tracer_type_len+1];
	int i;

	if (cnt > max_tracer_type_len)
		cnt = max_tracer_type_len;

	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;

	mutex_lock(&trace_types_lock);
	for (t = trace_types; t; t = t->next) {
		if (strcmp(t->name, buf) == 0)
			break;
	}
	if (!t || t == current_trace)
		goto out;

	if (current_trace && current_trace->reset)
		current_trace->reset(tr);

	current_trace = t;
	if (t->init)
		t->init(tr);

 out:
	mutex_unlock(&trace_types_lock);

	filp->f_pos += cnt;

	return cnt;
}

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;

	r = snprintf(buf, 64, "%ld\n",
		     *ptr == (unsigned long)-1 ? -1 : nsecs_to_usecs(*ptr));
	if (r > 64)
		r = 64;
I
Ingo Molnar 已提交
2105
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130
}

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

	if (cnt > 63)
		cnt = 63;

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

	buf[cnt] = 0;

	val = simple_strtoul(buf, NULL, 10);

	*ptr = val * 1000;

	return cnt;
}

2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167
static atomic_t tracing_reader;

static int tracing_open_pipe(struct inode *inode, struct file *filp)
{
	struct trace_iterator *iter;

	if (tracing_disabled)
		return -ENODEV;

	/* We only allow for reader of the pipe */
	if (atomic_inc_return(&tracing_reader) != 1) {
		atomic_dec(&tracing_reader);
		return -EBUSY;
	}

	/* create a buffer to store the information to pass to userspace */
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter)
		return -ENOMEM;

	iter->tr = &global_trace;

	filp->private_data = iter;

	return 0;
}

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

	kfree(iter);
	atomic_dec(&tracing_reader);

	return 0;
}

2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189
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;
	}
	else {
		if (!trace_empty(iter))
			return POLLIN | POLLRDNORM;
		poll_wait(filp, &trace_wait, poll_table);
		if (!trace_empty(iter))
			return POLLIN | POLLRDNORM;

		return 0;
	}
}

2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201
/*
 * 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;
	struct trace_array_cpu *data;
	static cpumask_t mask;
	static int start;
	unsigned long flags;
I
Ingo Molnar 已提交
2202
#ifdef CONFIG_FTRACE
I
Ingo Molnar 已提交
2203
	int ftrace_save;
I
Ingo Molnar 已提交
2204
#endif
2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227
	int read = 0;
	int cpu;
	int len;
	int ret;

	/* return any leftover data */
	if (iter->seq.len > start) {
		len = iter->seq.len - start;
		if (cnt > len)
			cnt = len;
		ret = copy_to_user(ubuf, iter->seq.buffer + start, cnt);
		if (ret)
			cnt = -EFAULT;

		start += len;

		return cnt;
	}

	trace_seq_reset(&iter->seq);
	start = 0;

	while (trace_empty(iter)) {
2228 2229
		if (!(trace_flags & TRACE_ITER_BLOCK))
			return -EWOULDBLOCK;
2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284
		/*
		 * This is a make-shift waitqueue. The reason we don't use
		 * an actual wait queue is because:
		 *  1) we only ever have one waiter
		 *  2) the tracing, traces 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.
		 */
		set_current_state(TASK_INTERRUPTIBLE);
		iter->tr->waiter = current;

		/* sleep for one second, and try again. */
		schedule_timeout(HZ);

		iter->tr->waiter = NULL;

		if (signal_pending(current))
			return -EINTR;

		/*
		 * 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;

		continue;
	}

	/* stop when tracing is finished */
	if (trace_empty(iter))
		return 0;

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

	memset(iter, 0, sizeof(*iter));
	iter->tr = &global_trace;
	iter->pos = -1;

	/*
	 * 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.
	 */

	cpus_clear(mask);
	local_irq_save(flags);
I
Ingo Molnar 已提交
2285
#ifdef CONFIG_FTRACE
I
Ingo Molnar 已提交
2286 2287
	ftrace_save = ftrace_enabled;
	ftrace_enabled = 0;
I
Ingo Molnar 已提交
2288
#endif
I
Ingo Molnar 已提交
2289
	smp_wmb();
2290 2291 2292 2293 2294 2295 2296 2297 2298 2299
	for_each_possible_cpu(cpu) {
		data = iter->tr->data[cpu];

		if (!head_page(data) || !data->trace_idx)
			continue;

		atomic_inc(&data->disabled);
		cpu_set(cpu, mask);
	}

I
Ingo Molnar 已提交
2300 2301 2302 2303 2304
	for_each_cpu_mask(cpu, mask) {
		data = iter->tr->data[cpu];
		spin_lock(&data->lock);
	}

S
Steven Rostedt 已提交
2305 2306 2307
	while (find_next_entry_inc(iter) != NULL) {
		int len = iter->seq.len;

I
Ingo Molnar 已提交
2308
		ret = print_trace_line(iter);
S
Steven Rostedt 已提交
2309 2310 2311
		if (!ret) {
			/* don't print partial lines */
			iter->seq.len = len;
2312
			break;
S
Steven Rostedt 已提交
2313
		}
2314 2315 2316 2317 2318 2319 2320

		trace_consume(iter);

		if (iter->seq.len >= cnt)
			break;
	}

I
Ingo Molnar 已提交
2321
	for_each_cpu_mask(cpu, mask) {
2322 2323
		data = iter->tr->data[cpu];
		spin_unlock(&data->lock);
I
Ingo Molnar 已提交
2324 2325 2326 2327
	}

	for_each_cpu_mask(cpu, mask) {
		data = iter->tr->data[cpu];
2328 2329
		atomic_dec(&data->disabled);
	}
I
Ingo Molnar 已提交
2330
#ifdef CONFIG_FTRACE
I
Ingo Molnar 已提交
2331
	ftrace_enabled = ftrace_save;
I
Ingo Molnar 已提交
2332
#endif
2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352
	local_irq_restore(flags);

	/* Now copy what we have to the user */
	read = iter->seq.len;
	if (read > cnt)
		read = cnt;

	ret = copy_to_user(ubuf, iter->seq.buffer, read);

	if (read < iter->seq.len)
		start = read;
	else
		trace_seq_reset(&iter->seq);

	if (ret)
		read = -EFAULT;

	return read;
}

2353
static struct file_operations tracing_max_lat_fops = {
I
Ingo Molnar 已提交
2354 2355 2356
	.open		= tracing_open_generic,
	.read		= tracing_max_lat_read,
	.write		= tracing_max_lat_write,
2357 2358 2359
};

static struct file_operations tracing_ctrl_fops = {
I
Ingo Molnar 已提交
2360 2361 2362
	.open		= tracing_open_generic,
	.read		= tracing_ctrl_read,
	.write		= tracing_ctrl_write,
2363 2364 2365
};

static struct file_operations set_tracer_fops = {
I
Ingo Molnar 已提交
2366 2367 2368
	.open		= tracing_open_generic,
	.read		= tracing_set_trace_read,
	.write		= tracing_set_trace_write,
2369 2370
};

2371
static struct file_operations tracing_pipe_fops = {
I
Ingo Molnar 已提交
2372
	.open		= tracing_open_pipe,
2373
	.poll		= tracing_poll_pipe,
I
Ingo Molnar 已提交
2374 2375
	.read		= tracing_read_pipe,
	.release	= tracing_release_pipe,
2376 2377
};

2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388
#ifdef CONFIG_DYNAMIC_FTRACE

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

	r = sprintf(buf, "%ld\n", *p);
I
Ingo Molnar 已提交
2389 2390

	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
2391 2392 2393
}

static struct file_operations tracing_read_long_fops = {
I
Ingo Molnar 已提交
2394 2395
	.open		= tracing_open_generic,
	.read		= tracing_read_long,
2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418
};
#endif

static struct dentry *d_tracer;

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

	if (d_tracer)
		return d_tracer;

	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;
}

S
Steven Rostedt 已提交
2419 2420 2421 2422 2423
#ifdef CONFIG_FTRACE_SELFTEST
/* Let selftest have access to static functions in this file */
#include "trace_selftest.c"
#endif

2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440
static __init void tracer_init_debugfs(void)
{
	struct dentry *d_tracer;
	struct dentry *entry;

	d_tracer = tracing_init_dentry();

	entry = debugfs_create_file("tracing_enabled", 0644, d_tracer,
				    &global_trace, &tracing_ctrl_fops);
	if (!entry)
		pr_warning("Could not create debugfs 'tracing_enabled' entry\n");

	entry = debugfs_create_file("iter_ctrl", 0644, d_tracer,
				    NULL, &tracing_iter_fops);
	if (!entry)
		pr_warning("Could not create debugfs 'iter_ctrl' entry\n");

I
Ingo Molnar 已提交
2441 2442 2443 2444 2445
	entry = debugfs_create_file("tracing_cpumask", 0644, d_tracer,
				    NULL, &tracing_cpumask_fops);
	if (!entry)
		pr_warning("Could not create debugfs 'tracing_cpumask' entry\n");

2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477
	entry = debugfs_create_file("latency_trace", 0444, d_tracer,
				    &global_trace, &tracing_lt_fops);
	if (!entry)
		pr_warning("Could not create debugfs 'latency_trace' entry\n");

	entry = debugfs_create_file("trace", 0444, d_tracer,
				    &global_trace, &tracing_fops);
	if (!entry)
		pr_warning("Could not create debugfs 'trace' entry\n");

	entry = debugfs_create_file("available_tracers", 0444, d_tracer,
				    &global_trace, &show_traces_fops);
	if (!entry)
		pr_warning("Could not create debugfs 'trace' entry\n");

	entry = debugfs_create_file("current_tracer", 0444, d_tracer,
				    &global_trace, &set_tracer_fops);
	if (!entry)
		pr_warning("Could not create debugfs 'trace' entry\n");

	entry = debugfs_create_file("tracing_max_latency", 0644, d_tracer,
				    &tracing_max_latency,
				    &tracing_max_lat_fops);
	if (!entry)
		pr_warning("Could not create debugfs "
			   "'tracing_max_latency' entry\n");

	entry = debugfs_create_file("tracing_thresh", 0644, d_tracer,
				    &tracing_thresh, &tracing_max_lat_fops);
	if (!entry)
		pr_warning("Could not create debugfs "
			   "'tracing_threash' entry\n");
I
Ingo Molnar 已提交
2478 2479 2480 2481 2482
	entry = debugfs_create_file("README", 0644, d_tracer,
				    NULL, &tracing_readme_fops);
	if (!entry)
		pr_warning("Could not create debugfs 'README' entry\n");

2483 2484 2485 2486 2487
	entry = debugfs_create_file("trace_pipe", 0644, d_tracer,
				    NULL, &tracing_pipe_fops);
	if (!entry)
		pr_warning("Could not create debugfs "
			   "'tracing_threash' entry\n");
2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501

#ifdef CONFIG_DYNAMIC_FTRACE
	entry = debugfs_create_file("dyn_ftrace_total_info", 0444, d_tracer,
				    &ftrace_update_tot_cnt,
				    &tracing_read_long_fops);
	if (!entry)
		pr_warning("Could not create debugfs "
			   "'dyn_ftrace_total_info' entry\n");
#endif
}

/* dummy trace to disable tracing */
static struct tracer no_tracer __read_mostly =
{
I
Ingo Molnar 已提交
2502
	.name		= "none",
2503 2504
};

2505
static int trace_alloc_page(void)
2506
{
2507 2508 2509
	struct trace_array_cpu *data;
	struct page *page, *tmp;
	LIST_HEAD(pages);
I
Ingo Molnar 已提交
2510
	void *array;
2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540
	int i;

	/* first allocate a page for each CPU */
	for_each_possible_cpu(i) {
		array = (void *)__get_free_page(GFP_KERNEL);
		if (array == NULL) {
			printk(KERN_ERR "tracer: failed to allocate page"
			       "for trace buffer!\n");
			goto free_pages;
		}

		page = virt_to_page(array);
		list_add(&page->lru, &pages);

/* Only allocate if we are actually using the max trace */
#ifdef CONFIG_TRACER_MAX_TRACE
		array = (void *)__get_free_page(GFP_KERNEL);
		if (array == NULL) {
			printk(KERN_ERR "tracer: failed to allocate page"
			       "for trace buffer!\n");
			goto free_pages;
		}
		page = virt_to_page(array);
		list_add(&page->lru, &pages);
#endif
	}

	/* Now that we successfully allocate a page per CPU, add them */
	for_each_possible_cpu(i) {
		data = global_trace.data[i];
2541
		spin_lock_init(&data->lock);
I
Ingo Molnar 已提交
2542
		lockdep_set_class(&data->lock, &data->lock_key);
2543
		page = list_entry(pages.next, struct page, lru);
I
Ingo Molnar 已提交
2544
		list_del_init(&page->lru);
2545 2546 2547 2548 2549
		list_add_tail(&page->lru, &data->trace_pages);
		ClearPageLRU(page);

#ifdef CONFIG_TRACER_MAX_TRACE
		data = max_tr.data[i];
2550
		spin_lock_init(&data->lock);
I
Ingo Molnar 已提交
2551
		lockdep_set_class(&data->lock, &data->lock_key);
2552
		page = list_entry(pages.next, struct page, lru);
I
Ingo Molnar 已提交
2553
		list_del_init(&page->lru);
2554 2555 2556 2557 2558 2559 2560 2561 2562 2563
		list_add_tail(&page->lru, &data->trace_pages);
		SetPageLRU(page);
#endif
	}
	global_trace.entries += ENTRIES_PER_PAGE;

	return 0;

 free_pages:
	list_for_each_entry_safe(page, tmp, &pages, lru) {
I
Ingo Molnar 已提交
2564
		list_del_init(&page->lru);
2565 2566 2567
		__free_page(page);
	}
	return -ENOMEM;
2568 2569 2570 2571
}

__init static int tracer_alloc_buffers(void)
{
2572 2573 2574 2575
	struct trace_array_cpu *data;
	void *array;
	struct page *page;
	int pages = 0;
S
Steven Rostedt 已提交
2576
	int ret = -ENOMEM;
2577 2578
	int i;

2579 2580
	global_trace.ctrl = tracer_enabled;

2581
	/* Allocate the first page for all buffers */
2582
	for_each_possible_cpu(i) {
2583
		data = global_trace.data[i] = &per_cpu(global_trace_cpu, i);
2584 2585
		max_tr.data[i] = &per_cpu(max_data, i);

2586
		array = (void *)__get_free_page(GFP_KERNEL);
2587
		if (array == NULL) {
2588 2589
			printk(KERN_ERR "tracer: failed to allocate page"
			       "for trace buffer!\n");
2590 2591
			goto free_buffers;
		}
2592 2593 2594 2595 2596 2597 2598

		/* set the array to the list */
		INIT_LIST_HEAD(&data->trace_pages);
		page = virt_to_page(array);
		list_add(&page->lru, &data->trace_pages);
		/* use the LRU flag to differentiate the two buffers */
		ClearPageLRU(page);
2599 2600 2601

/* Only allocate if we are actually using the max trace */
#ifdef CONFIG_TRACER_MAX_TRACE
2602
		array = (void *)__get_free_page(GFP_KERNEL);
2603
		if (array == NULL) {
2604 2605
			printk(KERN_ERR "tracer: failed to allocate page"
			       "for trace buffer!\n");
2606 2607
			goto free_buffers;
		}
2608 2609 2610 2611 2612

		INIT_LIST_HEAD(&max_tr.data[i]->trace_pages);
		page = virt_to_page(array);
		list_add(&page->lru, &max_tr.data[i]->trace_pages);
		SetPageLRU(page);
2613 2614 2615 2616 2617 2618 2619
#endif
	}

	/*
	 * Since we allocate by orders of pages, we may be able to
	 * round up a bit.
	 */
2620 2621 2622 2623 2624 2625 2626 2627
	global_trace.entries = ENTRIES_PER_PAGE;
	pages++;

	while (global_trace.entries < trace_nr_entries) {
		if (trace_alloc_page())
			break;
		pages++;
	}
2628
	max_tr.entries = global_trace.entries;
2629

2630 2631
	pr_info("tracer: %d pages allocated for %ld",
		pages, trace_nr_entries);
2632 2633 2634 2635 2636 2637 2638 2639 2640 2641
	pr_info(" entries of %ld bytes\n", (long)TRACE_ENTRY_SIZE);
	pr_info("   actual entries %ld\n", global_trace.entries);

	tracer_init_debugfs();

	trace_init_cmdlines();

	register_tracer(&no_tracer);
	current_trace = &no_tracer;

S
Steven Rostedt 已提交
2642 2643 2644
	/* All seems OK, enable tracing */
	tracing_disabled = 0;

2645 2646 2647 2648
	return 0;

 free_buffers:
	for (i-- ; i >= 0; i--) {
2649
		struct page *page, *tmp;
2650 2651
		struct trace_array_cpu *data = global_trace.data[i];

I
Ingo Molnar 已提交
2652
		if (data) {
2653 2654
			list_for_each_entry_safe(page, tmp,
						 &data->trace_pages, lru) {
I
Ingo Molnar 已提交
2655
				list_del_init(&page->lru);
2656 2657
				__free_page(page);
			}
2658 2659 2660 2661
		}

#ifdef CONFIG_TRACER_MAX_TRACE
		data = max_tr.data[i];
I
Ingo Molnar 已提交
2662
		if (data) {
2663 2664
			list_for_each_entry_safe(page, tmp,
						 &data->trace_pages, lru) {
I
Ingo Molnar 已提交
2665
				list_del_init(&page->lru);
2666 2667
				__free_page(page);
			}
2668 2669 2670
		}
#endif
	}
S
Steven Rostedt 已提交
2671
	return ret;
2672
}
S
Steven Rostedt 已提交
2673
fs_initcall(tracer_alloc_buffers);