ftrace.c 99.6 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
/*
 * Infrastructure for profiling code inserted by 'gcc -pg'.
 *
 * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
 * Copyright (C) 2004-2008 Ingo Molnar <mingo@redhat.com>
 *
 * Originally ported from the -rt patch by:
 *   Copyright (C) 2007 Arnaldo Carvalho de Melo <acme@redhat.com>
 *
 * Based on code in the latency_tracer, that is:
 *
 *  Copyright (C) 2004-2006 Ingo Molnar
 *  Copyright (C) 2004 William Lee Irwin III
 */

16 17 18
#include <linux/stop_machine.h>
#include <linux/clocksource.h>
#include <linux/kallsyms.h>
19
#include <linux/seq_file.h>
20
#include <linux/suspend.h>
21
#include <linux/debugfs.h>
22
#include <linux/hardirq.h>
I
Ingo Molnar 已提交
23
#include <linux/kthread.h>
24
#include <linux/uaccess.h>
25
#include <linux/bsearch.h>
26
#include <linux/module.h>
I
Ingo Molnar 已提交
27
#include <linux/ftrace.h>
28
#include <linux/sysctl.h>
29
#include <linux/slab.h>
30
#include <linux/ctype.h>
31
#include <linux/sort.h>
32
#include <linux/list.h>
33
#include <linux/hash.h>
34
#include <linux/rcupdate.h>
35

36
#include <trace/events/sched.h>
37

38
#include <asm/setup.h>
39

40
#include "trace_output.h"
S
Steven Rostedt 已提交
41
#include "trace_stat.h"
42

43
#define FTRACE_WARN_ON(cond)			\
44 45 46
	({					\
		int ___r = cond;		\
		if (WARN_ON(___r))		\
47
			ftrace_kill();		\
48 49
		___r;				\
	})
50 51

#define FTRACE_WARN_ON_ONCE(cond)		\
52 53 54
	({					\
		int ___r = cond;		\
		if (WARN_ON_ONCE(___r))		\
55
			ftrace_kill();		\
56 57
		___r;				\
	})
58

59 60 61
/* hash bits for specific function selection */
#define FTRACE_HASH_BITS 7
#define FTRACE_FUNC_HASHSIZE (1 << FTRACE_HASH_BITS)
62 63
#define FTRACE_HASH_DEFAULT_BITS 10
#define FTRACE_HASH_MAX_BITS 12
64

65 66
#define FL_GLOBAL_CONTROL_MASK (FTRACE_OPS_FL_GLOBAL | FTRACE_OPS_FL_CONTROL)

67 68
/* ftrace_enabled is a method to turn ftrace on or off */
int ftrace_enabled __read_mostly;
69
static int last_ftrace_enabled;
70

71 72 73
/* Quick disabling of function tracer. */
int function_trace_stop;

74 75 76 77 78 79 80
/* List for set_ftrace_pid's pids. */
LIST_HEAD(ftrace_pids);
struct ftrace_pid {
	struct list_head list;
	struct pid *pid;
};

81 82 83 84 85 86
/*
 * ftrace_disabled is set when an anomaly is discovered.
 * ftrace_disabled is much stronger than ftrace_enabled.
 */
static int ftrace_disabled __read_mostly;

87
static DEFINE_MUTEX(ftrace_lock);
88

89
static struct ftrace_ops ftrace_list_end __read_mostly = {
90
	.func		= ftrace_stub,
91 92
};

93
static struct ftrace_ops *ftrace_global_list __read_mostly = &ftrace_list_end;
94
static struct ftrace_ops *ftrace_control_list __read_mostly = &ftrace_list_end;
95
static struct ftrace_ops *ftrace_ops_list __read_mostly = &ftrace_list_end;
96
ftrace_func_t ftrace_trace_function __read_mostly = ftrace_stub;
97
static ftrace_func_t __ftrace_trace_function_delay __read_mostly = ftrace_stub;
98
ftrace_func_t __ftrace_trace_function __read_mostly = ftrace_stub;
99
ftrace_func_t ftrace_pid_function __read_mostly = ftrace_stub;
100
static struct ftrace_ops global_ops;
101
static struct ftrace_ops control_ops;
102

103 104 105
static void
ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip);

106
/*
107
 * Traverse the ftrace_global_list, invoking all entries.  The reason that we
108 109 110
 * can use rcu_dereference_raw() is that elements removed from this list
 * are simply leaked, so there is no need to interact with a grace-period
 * mechanism.  The rcu_dereference_raw() calls are needed to handle
111
 * concurrent insertions into the ftrace_global_list.
112 113 114
 *
 * Silly Alpha and silly pointer-speculation compiler optimizations!
 */
115 116
static void ftrace_global_list_func(unsigned long ip,
				    unsigned long parent_ip)
117
{
118 119 120 121
	struct ftrace_ops *op;

	if (unlikely(trace_recursion_test(TRACE_GLOBAL_BIT)))
		return;
122

123 124
	trace_recursion_set(TRACE_GLOBAL_BIT);
	op = rcu_dereference_raw(ftrace_global_list); /*see above*/
125 126
	while (op != &ftrace_list_end) {
		op->func(ip, parent_ip);
127
		op = rcu_dereference_raw(op->next); /*see above*/
128
	};
129
	trace_recursion_clear(TRACE_GLOBAL_BIT);
130 131
}

132 133
static void ftrace_pid_func(unsigned long ip, unsigned long parent_ip)
{
134
	if (!test_tsk_trace_trace(current))
135 136 137 138 139 140 141 142 143 144 145 146
		return;

	ftrace_pid_function(ip, parent_ip);
}

static void set_ftrace_pid_function(ftrace_func_t func)
{
	/* do not set ftrace_pid_function to itself! */
	if (func != ftrace_pid_func)
		ftrace_pid_function = func;
}

147
/**
148
 * clear_ftrace_function - reset the ftrace function
149
 *
150 151
 * This NULLs the ftrace function and in essence stops
 * tracing.  There may be lag
152
 */
153
void clear_ftrace_function(void)
154
{
155
	ftrace_trace_function = ftrace_stub;
156
	__ftrace_trace_function = ftrace_stub;
157
	__ftrace_trace_function_delay = ftrace_stub;
158
	ftrace_pid_function = ftrace_stub;
159 160
}

161 162 163 164 165 166 167 168 169 170 171 172 173 174
#ifndef CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST
/*
 * For those archs that do not test ftrace_trace_stop in their
 * mcount call site, we need to do it from C.
 */
static void ftrace_test_stop_func(unsigned long ip, unsigned long parent_ip)
{
	if (function_trace_stop)
		return;

	__ftrace_trace_function(ip, parent_ip);
}
#endif

175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200
static void control_ops_disable_all(struct ftrace_ops *ops)
{
	int cpu;

	for_each_possible_cpu(cpu)
		*per_cpu_ptr(ops->disabled, cpu) = 1;
}

static int control_ops_alloc(struct ftrace_ops *ops)
{
	int __percpu *disabled;

	disabled = alloc_percpu(int);
	if (!disabled)
		return -ENOMEM;

	ops->disabled = disabled;
	control_ops_disable_all(ops);
	return 0;
}

static void control_ops_free(struct ftrace_ops *ops)
{
	free_percpu(ops->disabled);
}

201
static void update_global_ops(void)
202 203 204 205 206 207 208 209
{
	ftrace_func_t func;

	/*
	 * If there's only one function registered, then call that
	 * function directly. Otherwise, we need to iterate over the
	 * registered callers.
	 */
210 211 212
	if (ftrace_global_list == &ftrace_list_end ||
	    ftrace_global_list->next == &ftrace_list_end)
		func = ftrace_global_list->func;
213
	else
214
		func = ftrace_global_list_func;
215 216 217 218 219 220

	/* If we filter on pids, update to use the pid function */
	if (!list_empty(&ftrace_pids)) {
		set_ftrace_pid_function(func);
		func = ftrace_pid_func;
	}
221 222 223 224 225 226 227 228 229 230

	global_ops.func = func;
}

static void update_ftrace_function(void)
{
	ftrace_func_t func;

	update_global_ops();

231 232 233 234 235
	/*
	 * If we are at the end of the list and this ops is
	 * not dynamic, then have the mcount trampoline call
	 * the function directly
	 */
236
	if (ftrace_ops_list == &ftrace_list_end ||
237 238
	    (ftrace_ops_list->next == &ftrace_list_end &&
	     !(ftrace_ops_list->flags & FTRACE_OPS_FL_DYNAMIC)))
239 240 241
		func = ftrace_ops_list->func;
	else
		func = ftrace_ops_list_func;
242

243 244
#ifdef CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST
	ftrace_trace_function = func;
245 246 247 248
#else
#ifdef CONFIG_DYNAMIC_FTRACE
	/* do not update till all functions have been modified */
	__ftrace_trace_function_delay = func;
249 250
#else
	__ftrace_trace_function = func;
251
#endif
252 253
	ftrace_trace_function =
		(func == ftrace_stub) ? func : ftrace_test_stop_func;
254 255 256
#endif
}

257
static void add_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
258
{
259
	ops->next = *list;
260
	/*
261
	 * We are entering ops into the list but another
262 263
	 * CPU might be walking that list. We need to make sure
	 * the ops->next pointer is valid before another CPU sees
264
	 * the ops pointer included into the list.
265
	 */
266
	rcu_assign_pointer(*list, ops);
267 268
}

269
static int remove_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
270 271 272 273
{
	struct ftrace_ops **p;

	/*
274 275
	 * If we are removing the last function, then simply point
	 * to the ftrace_stub.
276
	 */
277 278
	if (*list == ops && ops->next == &ftrace_list_end) {
		*list = &ftrace_list_end;
S
Steven Rostedt 已提交
279
		return 0;
280 281
	}

282
	for (p = list; *p != &ftrace_list_end; p = &(*p)->next)
283 284 285
		if (*p == ops)
			break;

S
Steven Rostedt 已提交
286 287
	if (*p != ops)
		return -1;
288 289

	*p = (*p)->next;
290 291
	return 0;
}
292

293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
static void add_ftrace_list_ops(struct ftrace_ops **list,
				struct ftrace_ops *main_ops,
				struct ftrace_ops *ops)
{
	int first = *list == &ftrace_list_end;
	add_ftrace_ops(list, ops);
	if (first)
		add_ftrace_ops(&ftrace_ops_list, main_ops);
}

static int remove_ftrace_list_ops(struct ftrace_ops **list,
				  struct ftrace_ops *main_ops,
				  struct ftrace_ops *ops)
{
	int ret = remove_ftrace_ops(list, ops);
	if (!ret && *list == &ftrace_list_end)
		ret = remove_ftrace_ops(&ftrace_ops_list, main_ops);
	return ret;
}

313 314 315 316 317 318 319 320
static int __register_ftrace_function(struct ftrace_ops *ops)
{
	if (ftrace_disabled)
		return -ENODEV;

	if (FTRACE_WARN_ON(ops == &global_ops))
		return -EINVAL;

321 322 323
	if (WARN_ON(ops->flags & FTRACE_OPS_FL_ENABLED))
		return -EBUSY;

324 325 326 327
	/* We don't support both control and global flags set. */
	if ((ops->flags & FL_GLOBAL_CONTROL_MASK) == FL_GLOBAL_CONTROL_MASK)
		return -EINVAL;

328 329 330
	if (!core_kernel_data((unsigned long)ops))
		ops->flags |= FTRACE_OPS_FL_DYNAMIC;

331
	if (ops->flags & FTRACE_OPS_FL_GLOBAL) {
332
		add_ftrace_list_ops(&ftrace_global_list, &global_ops, ops);
333
		ops->flags |= FTRACE_OPS_FL_ENABLED;
334 335 336 337
	} else if (ops->flags & FTRACE_OPS_FL_CONTROL) {
		if (control_ops_alloc(ops))
			return -ENOMEM;
		add_ftrace_list_ops(&ftrace_control_list, &control_ops, ops);
338 339 340
	} else
		add_ftrace_ops(&ftrace_ops_list, ops);

341 342 343 344 345 346 347 348 349 350 351 352 353
	if (ftrace_enabled)
		update_ftrace_function();

	return 0;
}

static int __unregister_ftrace_function(struct ftrace_ops *ops)
{
	int ret;

	if (ftrace_disabled)
		return -ENODEV;

354 355 356
	if (WARN_ON(!(ops->flags & FTRACE_OPS_FL_ENABLED)))
		return -EBUSY;

357 358 359
	if (FTRACE_WARN_ON(ops == &global_ops))
		return -EINVAL;

360
	if (ops->flags & FTRACE_OPS_FL_GLOBAL) {
361 362
		ret = remove_ftrace_list_ops(&ftrace_global_list,
					     &global_ops, ops);
363 364
		if (!ret)
			ops->flags &= ~FTRACE_OPS_FL_ENABLED;
365 366 367 368 369 370 371 372 373 374 375 376 377
	} else if (ops->flags & FTRACE_OPS_FL_CONTROL) {
		ret = remove_ftrace_list_ops(&ftrace_control_list,
					     &control_ops, ops);
		if (!ret) {
			/*
			 * The ftrace_ops is now removed from the list,
			 * so there'll be no new users. We must ensure
			 * all current users are done before we free
			 * the control data.
			 */
			synchronize_sched();
			control_ops_free(ops);
		}
378 379 380
	} else
		ret = remove_ftrace_ops(&ftrace_ops_list, ops);

381 382
	if (ret < 0)
		return ret;
383

384 385
	if (ftrace_enabled)
		update_ftrace_function();
386

387 388 389 390 391 392 393
	/*
	 * Dynamic ops may be freed, we must make sure that all
	 * callers are done before leaving this function.
	 */
	if (ops->flags & FTRACE_OPS_FL_DYNAMIC)
		synchronize_sched();

S
Steven Rostedt 已提交
394
	return 0;
395 396
}

397 398
static void ftrace_update_pid_func(void)
{
399
	/* Only do something if we are tracing something */
400
	if (ftrace_trace_function == ftrace_stub)
401
		return;
402

403
	update_ftrace_function();
404 405
}

406 407 408 409 410
#ifdef CONFIG_FUNCTION_PROFILER
struct ftrace_profile {
	struct hlist_node		node;
	unsigned long			ip;
	unsigned long			counter;
411 412
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	unsigned long long		time;
413
	unsigned long long		time_squared;
414
#endif
415 416
};

417 418 419 420
struct ftrace_profile_page {
	struct ftrace_profile_page	*next;
	unsigned long			index;
	struct ftrace_profile		records[];
421 422
};

423 424 425 426 427 428 429 430
struct ftrace_profile_stat {
	atomic_t			disabled;
	struct hlist_head		*hash;
	struct ftrace_profile_page	*pages;
	struct ftrace_profile_page	*start;
	struct tracer_stat		stat;
};

431 432
#define PROFILE_RECORDS_SIZE						\
	(PAGE_SIZE - offsetof(struct ftrace_profile_page, records))
433

434 435
#define PROFILES_PER_PAGE					\
	(PROFILE_RECORDS_SIZE / sizeof(struct ftrace_profile))
436

437 438 439 440
static int ftrace_profile_bits __read_mostly;
static int ftrace_profile_enabled __read_mostly;

/* ftrace_profile_lock - synchronize the enable and disable of the profiler */
S
Steven Rostedt 已提交
441 442
static DEFINE_MUTEX(ftrace_profile_lock);

443
static DEFINE_PER_CPU(struct ftrace_profile_stat, ftrace_profile_stats);
444 445 446

#define FTRACE_PROFILE_HASH_SIZE 1024 /* must be power of 2 */

S
Steven Rostedt 已提交
447 448 449
static void *
function_stat_next(void *v, int idx)
{
450 451
	struct ftrace_profile *rec = v;
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
452

453
	pg = (struct ftrace_profile_page *)((unsigned long)rec & PAGE_MASK);
S
Steven Rostedt 已提交
454 455

 again:
L
Li Zefan 已提交
456 457 458
	if (idx != 0)
		rec++;

S
Steven Rostedt 已提交
459 460 461 462 463
	if ((void *)rec >= (void *)&pg->records[pg->index]) {
		pg = pg->next;
		if (!pg)
			return NULL;
		rec = &pg->records[0];
464 465
		if (!rec->counter)
			goto again;
S
Steven Rostedt 已提交
466 467 468 469 470 471 472
	}

	return rec;
}

static void *function_stat_start(struct tracer_stat *trace)
{
473 474 475 476 477 478 479
	struct ftrace_profile_stat *stat =
		container_of(trace, struct ftrace_profile_stat, stat);

	if (!stat || !stat->start)
		return NULL;

	return function_stat_next(&stat->start->records[0], 0);
S
Steven Rostedt 已提交
480 481
}

482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
/* function graph compares on total time */
static int function_stat_cmp(void *p1, void *p2)
{
	struct ftrace_profile *a = p1;
	struct ftrace_profile *b = p2;

	if (a->time < b->time)
		return -1;
	if (a->time > b->time)
		return 1;
	else
		return 0;
}
#else
/* not function graph compares against hits */
S
Steven Rostedt 已提交
498 499
static int function_stat_cmp(void *p1, void *p2)
{
500 501
	struct ftrace_profile *a = p1;
	struct ftrace_profile *b = p2;
S
Steven Rostedt 已提交
502 503 504 505 506 507 508 509

	if (a->counter < b->counter)
		return -1;
	if (a->counter > b->counter)
		return 1;
	else
		return 0;
}
510
#endif
S
Steven Rostedt 已提交
511 512 513

static int function_stat_headers(struct seq_file *m)
{
514
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
515
	seq_printf(m, "  Function                               "
516
		   "Hit    Time            Avg             s^2\n"
517
		      "  --------                               "
518
		   "---    ----            ---             ---\n");
519
#else
S
Steven Rostedt 已提交
520 521
	seq_printf(m, "  Function                               Hit\n"
		      "  --------                               ---\n");
522
#endif
S
Steven Rostedt 已提交
523 524 525 526 527
	return 0;
}

static int function_stat_show(struct seq_file *m, void *v)
{
528
	struct ftrace_profile *rec = v;
S
Steven Rostedt 已提交
529
	char str[KSYM_SYMBOL_LEN];
530
	int ret = 0;
531
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
532 533
	static struct trace_seq s;
	unsigned long long avg;
534
	unsigned long long stddev;
535
#endif
536 537 538 539 540 541 542
	mutex_lock(&ftrace_profile_lock);

	/* we raced with function_profile_reset() */
	if (unlikely(rec->counter == 0)) {
		ret = -EBUSY;
		goto out;
	}
S
Steven Rostedt 已提交
543 544

	kallsyms_lookup(rec->ip, NULL, NULL, NULL, str);
545 546 547 548
	seq_printf(m, "  %-30.30s  %10lu", str, rec->counter);

#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	seq_printf(m, "    ");
549 550 551
	avg = rec->time;
	do_div(avg, rec->counter);

552 553 554 555 556 557 558 559 560 561 562 563
	/* Sample standard deviation (s^2) */
	if (rec->counter <= 1)
		stddev = 0;
	else {
		stddev = rec->time_squared - rec->counter * avg * avg;
		/*
		 * Divide only 1000 for ns^2 -> us^2 conversion.
		 * trace_print_graph_duration will divide 1000 again.
		 */
		do_div(stddev, (rec->counter - 1) * 1000);
	}

564 565 566 567
	trace_seq_init(&s);
	trace_print_graph_duration(rec->time, &s);
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(avg, &s);
568 569
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(stddev, &s);
570 571 572
	trace_print_seq(m, &s);
#endif
	seq_putc(m, '\n');
573 574
out:
	mutex_unlock(&ftrace_profile_lock);
S
Steven Rostedt 已提交
575

576
	return ret;
S
Steven Rostedt 已提交
577 578
}

579
static void ftrace_profile_reset(struct ftrace_profile_stat *stat)
S
Steven Rostedt 已提交
580
{
581
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
582

583
	pg = stat->pages = stat->start;
S
Steven Rostedt 已提交
584

585 586 587 588
	while (pg) {
		memset(pg->records, 0, PROFILE_RECORDS_SIZE);
		pg->index = 0;
		pg = pg->next;
S
Steven Rostedt 已提交
589 590
	}

591
	memset(stat->hash, 0,
592 593
	       FTRACE_PROFILE_HASH_SIZE * sizeof(struct hlist_head));
}
S
Steven Rostedt 已提交
594

595
int ftrace_profile_pages_init(struct ftrace_profile_stat *stat)
596 597
{
	struct ftrace_profile_page *pg;
598 599
	int functions;
	int pages;
600
	int i;
S
Steven Rostedt 已提交
601

602
	/* If we already allocated, do nothing */
603
	if (stat->pages)
604
		return 0;
S
Steven Rostedt 已提交
605

606 607
	stat->pages = (void *)get_zeroed_page(GFP_KERNEL);
	if (!stat->pages)
608
		return -ENOMEM;
S
Steven Rostedt 已提交
609

610 611 612 613 614 615 616 617 618 619 620 621 622
#ifdef CONFIG_DYNAMIC_FTRACE
	functions = ftrace_update_tot_cnt;
#else
	/*
	 * We do not know the number of functions that exist because
	 * dynamic tracing is what counts them. With past experience
	 * we have around 20K functions. That should be more than enough.
	 * It is highly unlikely we will execute every function in
	 * the kernel.
	 */
	functions = 20000;
#endif

623
	pg = stat->start = stat->pages;
S
Steven Rostedt 已提交
624

625 626 627
	pages = DIV_ROUND_UP(functions, PROFILES_PER_PAGE);

	for (i = 0; i < pages; i++) {
628 629
		pg->next = (void *)get_zeroed_page(GFP_KERNEL);
		if (!pg->next)
630
			goto out_free;
631 632 633 634
		pg = pg->next;
	}

	return 0;
635 636 637 638 639 640 641 642 643 644 645 646 647 648 649

 out_free:
	pg = stat->start;
	while (pg) {
		unsigned long tmp = (unsigned long)pg;

		pg = pg->next;
		free_page(tmp);
	}

	free_page((unsigned long)stat->pages);
	stat->pages = NULL;
	stat->start = NULL;

	return -ENOMEM;
S
Steven Rostedt 已提交
650 651
}

652
static int ftrace_profile_init_cpu(int cpu)
S
Steven Rostedt 已提交
653
{
654
	struct ftrace_profile_stat *stat;
655
	int size;
S
Steven Rostedt 已提交
656

657 658 659
	stat = &per_cpu(ftrace_profile_stats, cpu);

	if (stat->hash) {
660
		/* If the profile is already created, simply reset it */
661
		ftrace_profile_reset(stat);
662 663
		return 0;
	}
S
Steven Rostedt 已提交
664

665 666 667 668 669
	/*
	 * We are profiling all functions, but usually only a few thousand
	 * functions are hit. We'll make a hash of 1024 items.
	 */
	size = FTRACE_PROFILE_HASH_SIZE;
S
Steven Rostedt 已提交
670

671
	stat->hash = kzalloc(sizeof(struct hlist_head) * size, GFP_KERNEL);
672

673
	if (!stat->hash)
674 675
		return -ENOMEM;

676 677
	if (!ftrace_profile_bits) {
		size--;
678

679 680 681
		for (; size; size >>= 1)
			ftrace_profile_bits++;
	}
682

683
	/* Preallocate the function profiling pages */
684 685 686
	if (ftrace_profile_pages_init(stat) < 0) {
		kfree(stat->hash);
		stat->hash = NULL;
687 688 689 690
		return -ENOMEM;
	}

	return 0;
S
Steven Rostedt 已提交
691 692
}

693 694 695 696 697 698 699 700 701 702 703 704 705 706
static int ftrace_profile_init(void)
{
	int cpu;
	int ret = 0;

	for_each_online_cpu(cpu) {
		ret = ftrace_profile_init_cpu(cpu);
		if (ret)
			break;
	}

	return ret;
}

707
/* interrupts must be disabled */
708 709
static struct ftrace_profile *
ftrace_find_profiled_func(struct ftrace_profile_stat *stat, unsigned long ip)
S
Steven Rostedt 已提交
710
{
711
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
712 713 714 715 716
	struct hlist_head *hhd;
	struct hlist_node *n;
	unsigned long key;

	key = hash_long(ip, ftrace_profile_bits);
717
	hhd = &stat->hash[key];
S
Steven Rostedt 已提交
718 719 720 721 722 723

	if (hlist_empty(hhd))
		return NULL;

	hlist_for_each_entry_rcu(rec, n, hhd, node) {
		if (rec->ip == ip)
724 725 726 727 728 729
			return rec;
	}

	return NULL;
}

730 731
static void ftrace_add_profile(struct ftrace_profile_stat *stat,
			       struct ftrace_profile *rec)
732 733 734 735
{
	unsigned long key;

	key = hash_long(rec->ip, ftrace_profile_bits);
736
	hlist_add_head_rcu(&rec->node, &stat->hash[key]);
737 738
}

739 740 741
/*
 * The memory is already allocated, this simply finds a new record to use.
 */
742
static struct ftrace_profile *
743
ftrace_profile_alloc(struct ftrace_profile_stat *stat, unsigned long ip)
744 745 746
{
	struct ftrace_profile *rec = NULL;

747
	/* prevent recursion (from NMIs) */
748
	if (atomic_inc_return(&stat->disabled) != 1)
749 750 751
		goto out;

	/*
752 753
	 * Try to find the function again since an NMI
	 * could have added it
754
	 */
755
	rec = ftrace_find_profiled_func(stat, ip);
756
	if (rec)
757
		goto out;
758

759 760 761 762
	if (stat->pages->index == PROFILES_PER_PAGE) {
		if (!stat->pages->next)
			goto out;
		stat->pages = stat->pages->next;
S
Steven Rostedt 已提交
763
	}
764

765
	rec = &stat->pages->records[stat->pages->index++];
766
	rec->ip = ip;
767
	ftrace_add_profile(stat, rec);
768

S
Steven Rostedt 已提交
769
 out:
770
	atomic_dec(&stat->disabled);
S
Steven Rostedt 已提交
771 772 773 774 775 776 777

	return rec;
}

static void
function_profile_call(unsigned long ip, unsigned long parent_ip)
{
778
	struct ftrace_profile_stat *stat;
779
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
780 781 782 783 784 785
	unsigned long flags;

	if (!ftrace_profile_enabled)
		return;

	local_irq_save(flags);
786 787

	stat = &__get_cpu_var(ftrace_profile_stats);
788
	if (!stat->hash || !ftrace_profile_enabled)
789 790 791
		goto out;

	rec = ftrace_find_profiled_func(stat, ip);
792
	if (!rec) {
793
		rec = ftrace_profile_alloc(stat, ip);
794 795 796
		if (!rec)
			goto out;
	}
S
Steven Rostedt 已提交
797 798 799 800 801 802

	rec->counter++;
 out:
	local_irq_restore(flags);
}

803 804 805 806 807 808 809 810 811
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static int profile_graph_entry(struct ftrace_graph_ent *trace)
{
	function_profile_call(trace->func, 0);
	return 1;
}

static void profile_graph_return(struct ftrace_graph_ret *trace)
{
812
	struct ftrace_profile_stat *stat;
813
	unsigned long long calltime;
814
	struct ftrace_profile *rec;
815
	unsigned long flags;
816 817

	local_irq_save(flags);
818
	stat = &__get_cpu_var(ftrace_profile_stats);
819
	if (!stat->hash || !ftrace_profile_enabled)
820 821
		goto out;

822 823 824 825
	/* If the calltime was zero'd ignore it */
	if (!trace->calltime)
		goto out;

826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842
	calltime = trace->rettime - trace->calltime;

	if (!(trace_flags & TRACE_ITER_GRAPH_TIME)) {
		int index;

		index = trace->depth;

		/* Append this call time to the parent time to subtract */
		if (index)
			current->ret_stack[index - 1].subtime += calltime;

		if (current->ret_stack[index].subtime < calltime)
			calltime -= current->ret_stack[index].subtime;
		else
			calltime = 0;
	}

843
	rec = ftrace_find_profiled_func(stat, trace->func);
844
	if (rec) {
845
		rec->time += calltime;
846 847
		rec->time_squared += calltime * calltime;
	}
848

849
 out:
850 851 852 853 854 855 856 857 858 859 860 861 862 863
	local_irq_restore(flags);
}

static int register_ftrace_profiler(void)
{
	return register_ftrace_graph(&profile_graph_return,
				     &profile_graph_entry);
}

static void unregister_ftrace_profiler(void)
{
	unregister_ftrace_graph();
}
#else
864
static struct ftrace_ops ftrace_profile_ops __read_mostly = {
865
	.func		= function_profile_call,
S
Steven Rostedt 已提交
866 867
};

868 869 870 871 872 873 874 875 876 877 878
static int register_ftrace_profiler(void)
{
	return register_ftrace_function(&ftrace_profile_ops);
}

static void unregister_ftrace_profiler(void)
{
	unregister_ftrace_function(&ftrace_profile_ops);
}
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

S
Steven Rostedt 已提交
879 880 881 882 883 884 885
static ssize_t
ftrace_profile_write(struct file *filp, const char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
	unsigned long val;
	int ret;

886 887
	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
S
Steven Rostedt 已提交
888 889 890 891 892 893 894
		return ret;

	val = !!val;

	mutex_lock(&ftrace_profile_lock);
	if (ftrace_profile_enabled ^ val) {
		if (val) {
895 896 897 898 899 900
			ret = ftrace_profile_init();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}

901 902 903 904 905
			ret = register_ftrace_profiler();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}
S
Steven Rostedt 已提交
906 907 908
			ftrace_profile_enabled = 1;
		} else {
			ftrace_profile_enabled = 0;
909 910 911 912
			/*
			 * unregister_ftrace_profiler calls stop_machine
			 * so this acts like an synchronize_sched.
			 */
913
			unregister_ftrace_profiler();
S
Steven Rostedt 已提交
914 915
		}
	}
916
 out:
S
Steven Rostedt 已提交
917 918
	mutex_unlock(&ftrace_profile_lock);

919
	*ppos += cnt;
S
Steven Rostedt 已提交
920 921 922 923

	return cnt;
}

924 925 926 927
static ssize_t
ftrace_profile_read(struct file *filp, char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
928
	char buf[64];		/* big enough to hold a number */
929 930 931 932 933 934
	int r;

	r = sprintf(buf, "%u\n", ftrace_profile_enabled);
	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
}

S
Steven Rostedt 已提交
935 936 937 938
static const struct file_operations ftrace_profile_fops = {
	.open		= tracing_open_generic,
	.read		= ftrace_profile_read,
	.write		= ftrace_profile_write,
939
	.llseek		= default_llseek,
S
Steven Rostedt 已提交
940 941
};

942 943
/* used to initialize the real stat files */
static struct tracer_stat function_stats __initdata = {
944 945 946 947 948 949
	.name		= "functions",
	.stat_start	= function_stat_start,
	.stat_next	= function_stat_next,
	.stat_cmp	= function_stat_cmp,
	.stat_headers	= function_stat_headers,
	.stat_show	= function_stat_show
950 951
};

952
static __init void ftrace_profile_debugfs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
953
{
954
	struct ftrace_profile_stat *stat;
S
Steven Rostedt 已提交
955
	struct dentry *entry;
956
	char *name;
S
Steven Rostedt 已提交
957
	int ret;
958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985
	int cpu;

	for_each_possible_cpu(cpu) {
		stat = &per_cpu(ftrace_profile_stats, cpu);

		/* allocate enough for function name + cpu number */
		name = kmalloc(32, GFP_KERNEL);
		if (!name) {
			/*
			 * The files created are permanent, if something happens
			 * we still do not free memory.
			 */
			WARN(1,
			     "Could not allocate stat file for cpu %d\n",
			     cpu);
			return;
		}
		stat->stat = function_stats;
		snprintf(name, 32, "function%d", cpu);
		stat->stat.name = name;
		ret = register_stat_tracer(&stat->stat);
		if (ret) {
			WARN(1,
			     "Could not register function stat for cpu %d\n",
			     cpu);
			kfree(name);
			return;
		}
S
Steven Rostedt 已提交
986 987 988 989 990 991 992 993 994 995
	}

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

#else /* CONFIG_FUNCTION_PROFILER */
996
static __init void ftrace_profile_debugfs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
997 998 999 1000
{
}
#endif /* CONFIG_FUNCTION_PROFILER */

1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019
static struct pid * const ftrace_swapper_pid = &init_struct_pid;

#ifdef CONFIG_DYNAMIC_FTRACE

#ifndef CONFIG_FTRACE_MCOUNT_RECORD
# error Dynamic ftrace depends on MCOUNT_RECORD
#endif

static struct hlist_head ftrace_func_hash[FTRACE_FUNC_HASHSIZE] __read_mostly;

struct ftrace_func_probe {
	struct hlist_node	node;
	struct ftrace_probe_ops	*ops;
	unsigned long		flags;
	unsigned long		ip;
	void			*data;
	struct rcu_head		rcu;
};

1020 1021 1022 1023 1024 1025 1026 1027 1028
struct ftrace_func_entry {
	struct hlist_node hlist;
	unsigned long ip;
};

struct ftrace_hash {
	unsigned long		size_bits;
	struct hlist_head	*buckets;
	unsigned long		count;
1029
	struct rcu_head		rcu;
1030 1031
};

1032 1033 1034 1035 1036 1037 1038 1039 1040
/*
 * We make these constant because no one should touch them,
 * but they are used as the default "empty hash", to avoid allocating
 * it all the time. These are in a read only section such that if
 * anyone does try to modify it, it will cause an exception.
 */
static const struct hlist_head empty_buckets[1];
static const struct ftrace_hash empty_hash = {
	.buckets = (struct hlist_head *)empty_buckets,
1041
};
1042
#define EMPTY_HASH	((struct ftrace_hash *)&empty_hash)
1043

1044
static struct ftrace_ops global_ops = {
1045
	.func			= ftrace_stub,
1046 1047
	.notrace_hash		= EMPTY_HASH,
	.filter_hash		= EMPTY_HASH,
1048 1049
};

1050 1051 1052 1053
static DEFINE_MUTEX(ftrace_regex_lock);

struct ftrace_page {
	struct ftrace_page	*next;
1054
	struct dyn_ftrace	*records;
1055
	int			index;
1056
	int			size;
1057 1058
};

1059 1060
static struct ftrace_page *ftrace_new_pgs;

1061 1062
#define ENTRY_SIZE sizeof(struct dyn_ftrace)
#define ENTRIES_PER_PAGE (PAGE_SIZE / ENTRY_SIZE)
1063 1064 1065 1066 1067 1068 1069

/* estimate from running different kernels */
#define NR_TO_INIT		10000

static struct ftrace_page	*ftrace_pages_start;
static struct ftrace_page	*ftrace_pages;

1070 1071 1072 1073 1074
static bool ftrace_hash_empty(struct ftrace_hash *hash)
{
	return !hash || !hash->count;
}

1075 1076 1077 1078 1079 1080 1081 1082
static struct ftrace_func_entry *
ftrace_lookup_ip(struct ftrace_hash *hash, unsigned long ip)
{
	unsigned long key;
	struct ftrace_func_entry *entry;
	struct hlist_head *hhd;
	struct hlist_node *n;

1083
	if (ftrace_hash_empty(hash))
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099
		return NULL;

	if (hash->size_bits > 0)
		key = hash_long(ip, hash->size_bits);
	else
		key = 0;

	hhd = &hash->buckets[key];

	hlist_for_each_entry_rcu(entry, n, hhd, hlist) {
		if (entry->ip == ip)
			return entry;
	}
	return NULL;
}

1100 1101
static void __add_hash_entry(struct ftrace_hash *hash,
			     struct ftrace_func_entry *entry)
1102 1103 1104 1105 1106
{
	struct hlist_head *hhd;
	unsigned long key;

	if (hash->size_bits)
1107
		key = hash_long(entry->ip, hash->size_bits);
1108 1109 1110 1111 1112 1113
	else
		key = 0;

	hhd = &hash->buckets[key];
	hlist_add_head(&entry->hlist, hhd);
	hash->count++;
1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
}

static int add_hash_entry(struct ftrace_hash *hash, unsigned long ip)
{
	struct ftrace_func_entry *entry;

	entry = kmalloc(sizeof(*entry), GFP_KERNEL);
	if (!entry)
		return -ENOMEM;

	entry->ip = ip;
	__add_hash_entry(hash, entry);
1126 1127 1128 1129 1130

	return 0;
}

static void
1131
free_hash_entry(struct ftrace_hash *hash,
1132 1133 1134 1135 1136 1137 1138
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	kfree(entry);
	hash->count--;
}

1139 1140 1141 1142 1143 1144 1145 1146
static void
remove_hash_entry(struct ftrace_hash *hash,
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	hash->count--;
}

1147 1148 1149 1150 1151 1152 1153 1154
static void ftrace_hash_clear(struct ftrace_hash *hash)
{
	struct hlist_head *hhd;
	struct hlist_node *tp, *tn;
	struct ftrace_func_entry *entry;
	int size = 1 << hash->size_bits;
	int i;

1155 1156 1157
	if (!hash->count)
		return;

1158 1159 1160
	for (i = 0; i < size; i++) {
		hhd = &hash->buckets[i];
		hlist_for_each_entry_safe(entry, tp, tn, hhd, hlist)
1161
			free_hash_entry(hash, entry);
1162 1163 1164 1165
	}
	FTRACE_WARN_ON(hash->count);
}

1166 1167 1168 1169 1170 1171 1172 1173 1174
static void free_ftrace_hash(struct ftrace_hash *hash)
{
	if (!hash || hash == EMPTY_HASH)
		return;
	ftrace_hash_clear(hash);
	kfree(hash->buckets);
	kfree(hash);
}

1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189
static void __free_ftrace_hash_rcu(struct rcu_head *rcu)
{
	struct ftrace_hash *hash;

	hash = container_of(rcu, struct ftrace_hash, rcu);
	free_ftrace_hash(hash);
}

static void free_ftrace_hash_rcu(struct ftrace_hash *hash)
{
	if (!hash || hash == EMPTY_HASH)
		return;
	call_rcu_sched(&hash->rcu, __free_ftrace_hash_rcu);
}

1190 1191 1192 1193 1194 1195
void ftrace_free_filter(struct ftrace_ops *ops)
{
	free_ftrace_hash(ops->filter_hash);
	free_ftrace_hash(ops->notrace_hash);
}

1196 1197 1198 1199 1200 1201 1202 1203 1204 1205
static struct ftrace_hash *alloc_ftrace_hash(int size_bits)
{
	struct ftrace_hash *hash;
	int size;

	hash = kzalloc(sizeof(*hash), GFP_KERNEL);
	if (!hash)
		return NULL;

	size = 1 << size_bits;
1206
	hash->buckets = kcalloc(size, sizeof(*hash->buckets), GFP_KERNEL);
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232

	if (!hash->buckets) {
		kfree(hash);
		return NULL;
	}

	hash->size_bits = size_bits;

	return hash;
}

static struct ftrace_hash *
alloc_and_copy_ftrace_hash(int size_bits, struct ftrace_hash *hash)
{
	struct ftrace_func_entry *entry;
	struct ftrace_hash *new_hash;
	struct hlist_node *tp;
	int size;
	int ret;
	int i;

	new_hash = alloc_ftrace_hash(size_bits);
	if (!new_hash)
		return NULL;

	/* Empty hash? */
1233
	if (ftrace_hash_empty(hash))
1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253
		return new_hash;

	size = 1 << hash->size_bits;
	for (i = 0; i < size; i++) {
		hlist_for_each_entry(entry, tp, &hash->buckets[i], hlist) {
			ret = add_hash_entry(new_hash, entry->ip);
			if (ret < 0)
				goto free_hash;
		}
	}

	FTRACE_WARN_ON(new_hash->count != hash->count);

	return new_hash;

 free_hash:
	free_ftrace_hash(new_hash);
	return NULL;
}

1254 1255 1256 1257 1258
static void
ftrace_hash_rec_disable(struct ftrace_ops *ops, int filter_hash);
static void
ftrace_hash_rec_enable(struct ftrace_ops *ops, int filter_hash);

1259
static int
1260 1261
ftrace_hash_move(struct ftrace_ops *ops, int enable,
		 struct ftrace_hash **dst, struct ftrace_hash *src)
1262 1263 1264 1265
{
	struct ftrace_func_entry *entry;
	struct hlist_node *tp, *tn;
	struct hlist_head *hhd;
1266 1267
	struct ftrace_hash *old_hash;
	struct ftrace_hash *new_hash;
1268 1269 1270
	unsigned long key;
	int size = src->count;
	int bits = 0;
1271
	int ret;
1272 1273
	int i;

1274 1275 1276 1277 1278 1279
	/*
	 * Remove the current set, update the hash and add
	 * them back.
	 */
	ftrace_hash_rec_disable(ops, enable);

1280 1281 1282 1283 1284
	/*
	 * If the new source is empty, just free dst and assign it
	 * the empty_hash.
	 */
	if (!src->count) {
1285 1286
		free_ftrace_hash_rcu(*dst);
		rcu_assign_pointer(*dst, EMPTY_HASH);
1287 1288 1289
		/* still need to update the function records */
		ret = 0;
		goto out;
1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
	}

	/*
	 * Make the hash size about 1/2 the # found
	 */
	for (size /= 2; size; size >>= 1)
		bits++;

	/* Don't allocate too much */
	if (bits > FTRACE_HASH_MAX_BITS)
		bits = FTRACE_HASH_MAX_BITS;

1302
	ret = -ENOMEM;
1303 1304
	new_hash = alloc_ftrace_hash(bits);
	if (!new_hash)
1305
		goto out;
1306 1307 1308 1309 1310 1311 1312 1313 1314 1315

	size = 1 << src->size_bits;
	for (i = 0; i < size; i++) {
		hhd = &src->buckets[i];
		hlist_for_each_entry_safe(entry, tp, tn, hhd, hlist) {
			if (bits > 0)
				key = hash_long(entry->ip, bits);
			else
				key = 0;
			remove_hash_entry(src, entry);
1316
			__add_hash_entry(new_hash, entry);
1317 1318 1319
		}
	}

1320 1321 1322 1323
	old_hash = *dst;
	rcu_assign_pointer(*dst, new_hash);
	free_ftrace_hash_rcu(old_hash);

1324 1325 1326 1327 1328 1329 1330 1331 1332 1333
	ret = 0;
 out:
	/*
	 * Enable regardless of ret:
	 *  On success, we enable the new hash.
	 *  On failure, we re-enable the original hash.
	 */
	ftrace_hash_rec_enable(ops, enable);

	return ret;
1334 1335
}

1336 1337 1338 1339 1340 1341 1342 1343
/*
 * Test the hashes for this ops to see if we want to call
 * the ops->func or not.
 *
 * It's a match if the ip is in the ops->filter_hash or
 * the filter_hash does not exist or is empty,
 *  AND
 * the ip is not in the ops->notrace_hash.
1344 1345 1346
 *
 * This needs to be called with preemption disabled as
 * the hashes are freed with call_rcu_sched().
1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357
 */
static int
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip)
{
	struct ftrace_hash *filter_hash;
	struct ftrace_hash *notrace_hash;
	int ret;

	filter_hash = rcu_dereference_raw(ops->filter_hash);
	notrace_hash = rcu_dereference_raw(ops->notrace_hash);

1358
	if ((ftrace_hash_empty(filter_hash) ||
1359
	     ftrace_lookup_ip(filter_hash, ip)) &&
1360
	    (ftrace_hash_empty(notrace_hash) ||
1361 1362 1363 1364 1365 1366 1367 1368
	     !ftrace_lookup_ip(notrace_hash, ip)))
		ret = 1;
	else
		ret = 0;

	return ret;
}

1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382
/*
 * This is a double for. Do not use 'break' to break out of the loop,
 * you must use a goto.
 */
#define do_for_each_ftrace_rec(pg, rec)					\
	for (pg = ftrace_pages_start; pg; pg = pg->next) {		\
		int _____i;						\
		for (_____i = 0; _____i < pg->index; _____i++) {	\
			rec = &pg->records[_____i];

#define while_for_each_ftrace_rec()		\
		}				\
	}

1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395

static int ftrace_cmp_recs(const void *a, const void *b)
{
	const struct dyn_ftrace *reca = a;
	const struct dyn_ftrace *recb = b;

	if (reca->ip > recb->ip)
		return 1;
	if (reca->ip < recb->ip)
		return -1;
	return 0;
}

1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408
/**
 * ftrace_location - return true if the ip giving is a traced location
 * @ip: the instruction pointer to check
 *
 * Returns 1 if @ip given is a pointer to a ftrace location.
 * That is, the instruction that is either a NOP or call to
 * the function tracer. It checks the ftrace internal tables to
 * determine if the address belongs or not.
 */
int ftrace_location(unsigned long ip)
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1409
	struct dyn_ftrace key;
1410

1411 1412 1413 1414 1415 1416 1417
	key.ip = ip;

	for (pg = ftrace_pages_start; pg; pg = pg->next) {
		rec = bsearch(&key, pg->records, pg->index,
			      sizeof(struct dyn_ftrace),
			      ftrace_cmp_recs);
		if (rec)
1418
			return 1;
1419
	}
1420 1421 1422 1423

	return 0;
}

1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452
static void __ftrace_hash_rec_update(struct ftrace_ops *ops,
				     int filter_hash,
				     bool inc)
{
	struct ftrace_hash *hash;
	struct ftrace_hash *other_hash;
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
	int count = 0;
	int all = 0;

	/* Only update if the ops has been registered */
	if (!(ops->flags & FTRACE_OPS_FL_ENABLED))
		return;

	/*
	 * In the filter_hash case:
	 *   If the count is zero, we update all records.
	 *   Otherwise we just update the items in the hash.
	 *
	 * In the notrace_hash case:
	 *   We enable the update in the hash.
	 *   As disabling notrace means enabling the tracing,
	 *   and enabling notrace means disabling, the inc variable
	 *   gets inversed.
	 */
	if (filter_hash) {
		hash = ops->filter_hash;
		other_hash = ops->notrace_hash;
1453
		if (ftrace_hash_empty(hash))
1454 1455 1456 1457 1458 1459 1460 1461 1462
			all = 1;
	} else {
		inc = !inc;
		hash = ops->notrace_hash;
		other_hash = ops->filter_hash;
		/*
		 * If the notrace hash has no items,
		 * then there's nothing to do.
		 */
1463
		if (ftrace_hash_empty(hash))
1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476
			return;
	}

	do_for_each_ftrace_rec(pg, rec) {
		int in_other_hash = 0;
		int in_hash = 0;
		int match = 0;

		if (all) {
			/*
			 * Only the filter_hash affects all records.
			 * Update if the record is not in the notrace hash.
			 */
1477
			if (!other_hash || !ftrace_lookup_ip(other_hash, rec->ip))
1478 1479
				match = 1;
		} else {
1480 1481
			in_hash = !!ftrace_lookup_ip(hash, rec->ip);
			in_other_hash = !!ftrace_lookup_ip(other_hash, rec->ip);
1482 1483 1484 1485 1486 1487 1488

			/*
			 *
			 */
			if (filter_hash && in_hash && !in_other_hash)
				match = 1;
			else if (!filter_hash && in_hash &&
1489
				 (in_other_hash || ftrace_hash_empty(other_hash)))
1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522
				match = 1;
		}
		if (!match)
			continue;

		if (inc) {
			rec->flags++;
			if (FTRACE_WARN_ON((rec->flags & ~FTRACE_FL_MASK) == FTRACE_REF_MAX))
				return;
		} else {
			if (FTRACE_WARN_ON((rec->flags & ~FTRACE_FL_MASK) == 0))
				return;
			rec->flags--;
		}
		count++;
		/* Shortcut, if we handled all records, we are done. */
		if (!all && count == hash->count)
			return;
	} while_for_each_ftrace_rec();
}

static void ftrace_hash_rec_disable(struct ftrace_ops *ops,
				    int filter_hash)
{
	__ftrace_hash_rec_update(ops, filter_hash, 0);
}

static void ftrace_hash_rec_enable(struct ftrace_ops *ops,
				   int filter_hash)
{
	__ftrace_hash_rec_update(ops, filter_hash, 1);
}

I
Ingo Molnar 已提交
1523
static struct dyn_ftrace *ftrace_alloc_dyn_node(unsigned long ip)
1524
{
1525 1526 1527 1528
	if (ftrace_pages->index == ftrace_pages->size) {
		/* We should have allocated enough */
		if (WARN_ON(!ftrace_pages->next))
			return NULL;
1529 1530 1531 1532 1533 1534
		ftrace_pages = ftrace_pages->next;
	}

	return &ftrace_pages->records[ftrace_pages->index++];
}

S
Steven Rostedt 已提交
1535
static struct dyn_ftrace *
1536
ftrace_record_ip(unsigned long ip)
1537
{
S
Steven Rostedt 已提交
1538
	struct dyn_ftrace *rec;
1539

1540
	if (ftrace_disabled)
S
Steven Rostedt 已提交
1541
		return NULL;
1542

S
Steven Rostedt 已提交
1543 1544 1545
	rec = ftrace_alloc_dyn_node(ip);
	if (!rec)
		return NULL;
1546

S
Steven Rostedt 已提交
1547
	rec->ip = ip;
1548

S
Steven Rostedt 已提交
1549
	return rec;
1550 1551
}

1552 1553 1554 1555 1556 1557 1558 1559 1560 1561
static void print_ip_ins(const char *fmt, unsigned char *p)
{
	int i;

	printk(KERN_CONT "%s", fmt);

	for (i = 0; i < MCOUNT_INSN_SIZE; i++)
		printk(KERN_CONT "%s%02x", i ? ":" : "", p[i]);
}

1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574
/**
 * ftrace_bug - report and shutdown function tracer
 * @failed: The failed type (EFAULT, EINVAL, EPERM)
 * @ip: The address that failed
 *
 * The arch code that enables or disables the function tracing
 * can call ftrace_bug() when it has detected a problem in
 * modifying the code. @failed should be one of either:
 * EFAULT - if the problem happens on reading the @ip address
 * EINVAL - if what is read at @ip is not what was expected
 * EPERM - if the problem happens on writting to the @ip address
 */
void ftrace_bug(int failed, unsigned long ip)
1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600
{
	switch (failed) {
	case -EFAULT:
		FTRACE_WARN_ON_ONCE(1);
		pr_info("ftrace faulted on modifying ");
		print_ip_sym(ip);
		break;
	case -EINVAL:
		FTRACE_WARN_ON_ONCE(1);
		pr_info("ftrace failed to modify ");
		print_ip_sym(ip);
		print_ip_ins(" actual: ", (unsigned char *)ip);
		printk(KERN_CONT "\n");
		break;
	case -EPERM:
		FTRACE_WARN_ON_ONCE(1);
		pr_info("ftrace faulted on writing ");
		print_ip_sym(ip);
		break;
	default:
		FTRACE_WARN_ON_ONCE(1);
		pr_info("ftrace faulted on unknown error ");
		print_ip_sym(ip);
	}
}

1601

1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615
/* Return 1 if the address range is reserved for ftrace */
int ftrace_text_reserved(void *start, void *end)
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;

	do_for_each_ftrace_rec(pg, rec) {
		if (rec->ip <= (unsigned long)end &&
		    rec->ip + MCOUNT_INSN_SIZE > (unsigned long)start)
			return 1;
	} while_for_each_ftrace_rec();
	return 0;
}

1616
static int ftrace_check_record(struct dyn_ftrace *rec, int enable, int update)
1617
{
1618
	unsigned long flag = 0UL;
1619

S
Steven Rostedt 已提交
1620
	/*
1621
	 * If we are updating calls:
S
Steven Rostedt 已提交
1622
	 *
1623 1624
	 *   If the record has a ref count, then we need to enable it
	 *   because someone is using it.
S
Steven Rostedt 已提交
1625
	 *
1626 1627
	 *   Otherwise we make sure its disabled.
	 *
1628
	 * If we are disabling calls, then disable all records that
1629
	 * are enabled.
S
Steven Rostedt 已提交
1630
	 */
1631
	if (enable && (rec->flags & ~FTRACE_FL_MASK))
1632
		flag = FTRACE_FL_ENABLED;
S
Steven Rostedt 已提交
1633

1634 1635
	/* If the state of this record hasn't changed, then do nothing */
	if ((rec->flags & FTRACE_FL_ENABLED) == flag)
1636
		return FTRACE_UPDATE_IGNORE;
S
Steven Rostedt 已提交
1637

1638
	if (flag) {
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 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691
		if (update)
			rec->flags |= FTRACE_FL_ENABLED;
		return FTRACE_UPDATE_MAKE_CALL;
	}

	if (update)
		rec->flags &= ~FTRACE_FL_ENABLED;

	return FTRACE_UPDATE_MAKE_NOP;
}

/**
 * ftrace_update_record, set a record that now is tracing or not
 * @rec: the record to update
 * @enable: set to 1 if the record is tracing, zero to force disable
 *
 * The records that represent all functions that can be traced need
 * to be updated when tracing has been enabled.
 */
int ftrace_update_record(struct dyn_ftrace *rec, int enable)
{
	return ftrace_check_record(rec, enable, 1);
}

/**
 * ftrace_test_record, check if the record has been enabled or not
 * @rec: the record to test
 * @enable: set to 1 to check if enabled, 0 if it is disabled
 *
 * The arch code may need to test if a record is already set to
 * tracing to determine how to modify the function code that it
 * represents.
 */
int ftrace_test_record(struct dyn_ftrace *rec, int enable)
{
	return ftrace_check_record(rec, enable, 0);
}

static int
__ftrace_replace_code(struct dyn_ftrace *rec, int enable)
{
	unsigned long ftrace_addr;
	int ret;

	ftrace_addr = (unsigned long)FTRACE_ADDR;

	ret = ftrace_update_record(rec, enable);

	switch (ret) {
	case FTRACE_UPDATE_IGNORE:
		return 0;

	case FTRACE_UPDATE_MAKE_CALL:
1692
		return ftrace_make_call(rec, ftrace_addr);
1693 1694 1695

	case FTRACE_UPDATE_MAKE_NOP:
		return ftrace_make_nop(NULL, rec, ftrace_addr);
1696 1697
	}

1698
	return -1; /* unknow ftrace bug */
1699 1700
}

1701
static void ftrace_replace_code(int update)
1702 1703 1704
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
S
Steven Rostedt 已提交
1705
	int failed;
1706

1707 1708 1709
	if (unlikely(ftrace_disabled))
		return;

1710
	do_for_each_ftrace_rec(pg, rec) {
1711
		failed = __ftrace_replace_code(rec, update);
1712
		if (failed) {
1713 1714 1715
			ftrace_bug(failed, rec->ip);
			/* Stop processing */
			return;
1716
		}
1717
	} while_for_each_ftrace_rec();
1718 1719
}

1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
struct ftrace_rec_iter {
	struct ftrace_page	*pg;
	int			index;
};

/**
 * ftrace_rec_iter_start, start up iterating over traced functions
 *
 * Returns an iterator handle that is used to iterate over all
 * the records that represent address locations where functions
 * are traced.
 *
 * May return NULL if no records are available.
 */
struct ftrace_rec_iter *ftrace_rec_iter_start(void)
{
	/*
	 * We only use a single iterator.
	 * Protected by the ftrace_lock mutex.
	 */
	static struct ftrace_rec_iter ftrace_rec_iter;
	struct ftrace_rec_iter *iter = &ftrace_rec_iter;

	iter->pg = ftrace_pages_start;
	iter->index = 0;

	/* Could have empty pages */
	while (iter->pg && !iter->pg->index)
		iter->pg = iter->pg->next;

	if (!iter->pg)
		return NULL;

	return iter;
}

/**
 * ftrace_rec_iter_next, get the next record to process.
 * @iter: The handle to the iterator.
 *
 * Returns the next iterator after the given iterator @iter.
 */
struct ftrace_rec_iter *ftrace_rec_iter_next(struct ftrace_rec_iter *iter)
{
	iter->index++;

	if (iter->index >= iter->pg->index) {
		iter->pg = iter->pg->next;
		iter->index = 0;

		/* Could have empty pages */
		while (iter->pg && !iter->pg->index)
			iter->pg = iter->pg->next;
	}

	if (!iter->pg)
		return NULL;

	return iter;
}

/**
 * ftrace_rec_iter_record, get the record at the iterator location
 * @iter: The current iterator location
 *
 * Returns the record that the current @iter is at.
 */
struct dyn_ftrace *ftrace_rec_iter_record(struct ftrace_rec_iter *iter)
{
	return &iter->pg->records[iter->index];
}

1792
static int
1793
ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec)
1794 1795
{
	unsigned long ip;
1796
	int ret;
1797 1798 1799

	ip = rec->ip;

1800 1801 1802
	if (unlikely(ftrace_disabled))
		return 0;

1803
	ret = ftrace_make_nop(mod, rec, MCOUNT_ADDR);
1804
	if (ret) {
1805
		ftrace_bug(ret, ip);
1806
		return 0;
1807
	}
1808
	return 1;
1809 1810
}

1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828
/*
 * archs can override this function if they must do something
 * before the modifying code is performed.
 */
int __weak ftrace_arch_code_modify_prepare(void)
{
	return 0;
}

/*
 * archs can override this function if they must do something
 * after the modifying code is performed.
 */
int __weak ftrace_arch_code_modify_post_process(void)
{
	return 0;
}

I
Ingo Molnar 已提交
1829
static int __ftrace_modify_code(void *data)
1830
{
1831 1832
	int *command = data;

1833
	if (*command & FTRACE_UPDATE_CALLS)
1834
		ftrace_replace_code(1);
1835
	else if (*command & FTRACE_DISABLE_CALLS)
1836 1837 1838 1839 1840
		ftrace_replace_code(0);

	if (*command & FTRACE_UPDATE_TRACE_FUNC)
		ftrace_update_ftrace_func(ftrace_trace_function);

1841 1842 1843 1844 1845
	if (*command & FTRACE_START_FUNC_RET)
		ftrace_enable_ftrace_graph_caller();
	else if (*command & FTRACE_STOP_FUNC_RET)
		ftrace_disable_ftrace_graph_caller();

1846
	return 0;
1847 1848
}

1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872
/**
 * ftrace_run_stop_machine, go back to the stop machine method
 * @command: The command to tell ftrace what to do
 *
 * If an arch needs to fall back to the stop machine method, the
 * it can call this function.
 */
void ftrace_run_stop_machine(int command)
{
	stop_machine(__ftrace_modify_code, &command, NULL);
}

/**
 * arch_ftrace_update_code, modify the code to trace or not trace
 * @command: The command that needs to be done
 *
 * Archs can override this function if it does not need to
 * run stop_machine() to modify code.
 */
void __weak arch_ftrace_update_code(int command)
{
	ftrace_run_stop_machine(command);
}

I
Ingo Molnar 已提交
1873
static void ftrace_run_update_code(int command)
1874
{
1875 1876 1877 1878 1879 1880
	int ret;

	ret = ftrace_arch_code_modify_prepare();
	FTRACE_WARN_ON(ret);
	if (ret)
		return;
1881 1882 1883 1884 1885
	/*
	 * Do not call function tracer while we update the code.
	 * We are in stop machine.
	 */
	function_trace_stop++;
1886

1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905
	/*
	 * By default we use stop_machine() to modify the code.
	 * But archs can do what ever they want as long as it
	 * is safe. The stop_machine() is the safest, but also
	 * produces the most overhead.
	 */
	arch_ftrace_update_code(command);

#ifndef CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST
	/*
	 * For archs that call ftrace_test_stop_func(), we must
	 * wait till after we update all the function callers
	 * before we update the callback. This keeps different
	 * ops that record different functions from corrupting
	 * each other.
	 */
	__ftrace_trace_function = __ftrace_trace_function_delay;
#endif
	function_trace_stop--;
1906 1907 1908

	ret = ftrace_arch_code_modify_post_process();
	FTRACE_WARN_ON(ret);
1909 1910
}

1911
static ftrace_func_t saved_ftrace_func;
1912
static int ftrace_start_up;
1913
static int global_start_up;
1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926

static void ftrace_startup_enable(int command)
{
	if (saved_ftrace_func != ftrace_trace_function) {
		saved_ftrace_func = ftrace_trace_function;
		command |= FTRACE_UPDATE_TRACE_FUNC;
	}

	if (!command || !ftrace_enabled)
		return;

	ftrace_run_update_code(command);
}
1927

1928
static int ftrace_startup(struct ftrace_ops *ops, int command)
1929
{
1930 1931
	bool hash_enable = true;

1932
	if (unlikely(ftrace_disabled))
1933
		return -ENODEV;
1934

1935
	ftrace_start_up++;
1936
	command |= FTRACE_UPDATE_CALLS;
1937

1938 1939 1940 1941 1942 1943 1944 1945 1946
	/* ops marked global share the filter hashes */
	if (ops->flags & FTRACE_OPS_FL_GLOBAL) {
		ops = &global_ops;
		/* Don't update hash if global is already set */
		if (global_start_up)
			hash_enable = false;
		global_start_up++;
	}

1947
	ops->flags |= FTRACE_OPS_FL_ENABLED;
1948
	if (hash_enable)
1949 1950
		ftrace_hash_rec_enable(ops, 1);

1951
	ftrace_startup_enable(command);
1952 1953

	return 0;
1954 1955
}

1956
static void ftrace_shutdown(struct ftrace_ops *ops, int command)
1957
{
1958 1959
	bool hash_disable = true;

1960 1961 1962
	if (unlikely(ftrace_disabled))
		return;

1963
	ftrace_start_up--;
1964 1965 1966 1967 1968 1969 1970
	/*
	 * Just warn in case of unbalance, no need to kill ftrace, it's not
	 * critical but the ftrace_call callers may be never nopped again after
	 * further ftrace uses.
	 */
	WARN_ON_ONCE(ftrace_start_up < 0);

1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982
	if (ops->flags & FTRACE_OPS_FL_GLOBAL) {
		ops = &global_ops;
		global_start_up--;
		WARN_ON_ONCE(global_start_up < 0);
		/* Don't update hash if global still has users */
		if (global_start_up) {
			WARN_ON_ONCE(!ftrace_start_up);
			hash_disable = false;
		}
	}

	if (hash_disable)
1983 1984
		ftrace_hash_rec_disable(ops, 1);

1985
	if (ops != &global_ops || !global_start_up)
1986
		ops->flags &= ~FTRACE_OPS_FL_ENABLED;
1987

1988
	command |= FTRACE_UPDATE_CALLS;
1989

1990 1991 1992 1993
	if (saved_ftrace_func != ftrace_trace_function) {
		saved_ftrace_func = ftrace_trace_function;
		command |= FTRACE_UPDATE_TRACE_FUNC;
	}
1994

1995
	if (!command || !ftrace_enabled)
S
Steven Rostedt 已提交
1996
		return;
1997 1998

	ftrace_run_update_code(command);
1999 2000
}

I
Ingo Molnar 已提交
2001
static void ftrace_startup_sysctl(void)
2002
{
2003 2004 2005
	if (unlikely(ftrace_disabled))
		return;

2006 2007
	/* Force update next time */
	saved_ftrace_func = NULL;
2008 2009
	/* ftrace_start_up is true if we want ftrace running */
	if (ftrace_start_up)
2010
		ftrace_run_update_code(FTRACE_UPDATE_CALLS);
2011 2012
}

I
Ingo Molnar 已提交
2013
static void ftrace_shutdown_sysctl(void)
2014
{
2015 2016 2017
	if (unlikely(ftrace_disabled))
		return;

2018 2019
	/* ftrace_start_up is true if ftrace is running */
	if (ftrace_start_up)
2020
		ftrace_run_update_code(FTRACE_DISABLE_CALLS);
2021 2022
}

2023 2024 2025 2026
static cycle_t		ftrace_update_time;
static unsigned long	ftrace_update_cnt;
unsigned long		ftrace_update_tot_cnt;

2027 2028 2029 2030 2031
static int ops_traces_mod(struct ftrace_ops *ops)
{
	struct ftrace_hash *hash;

	hash = ops->filter_hash;
2032
	return ftrace_hash_empty(hash);
2033 2034
}

2035
static int ftrace_update_code(struct module *mod)
2036
{
2037
	struct ftrace_page *pg;
2038
	struct dyn_ftrace *p;
2039
	cycle_t start, stop;
2040
	unsigned long ref = 0;
2041
	int i;
2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058

	/*
	 * When adding a module, we need to check if tracers are
	 * currently enabled and if they are set to trace all functions.
	 * If they are, we need to enable the module functions as well
	 * as update the reference counts for those function records.
	 */
	if (mod) {
		struct ftrace_ops *ops;

		for (ops = ftrace_ops_list;
		     ops != &ftrace_list_end; ops = ops->next) {
			if (ops->flags & FTRACE_OPS_FL_ENABLED &&
			    ops_traces_mod(ops))
				ref++;
		}
	}
2059

I
Ingo Molnar 已提交
2060
	start = ftrace_now(raw_smp_processor_id());
2061 2062
	ftrace_update_cnt = 0;

2063
	for (pg = ftrace_new_pgs; pg; pg = pg->next) {
2064

2065 2066 2067 2068
		for (i = 0; i < pg->index; i++) {
			/* If something went wrong, bail without enabling anything */
			if (unlikely(ftrace_disabled))
				return -1;
2069

2070 2071
			p = &pg->records[i];
			p->flags = ref;
2072

2073 2074 2075 2076 2077 2078
			/*
			 * Do the initial record conversion from mcount jump
			 * to the NOP instructions.
			 */
			if (!ftrace_code_disable(mod, p))
				break;
2079

2080
			ftrace_update_cnt++;
2081

2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095
			/*
			 * If the tracing is enabled, go ahead and enable the record.
			 *
			 * The reason not to enable the record immediatelly is the
			 * inherent check of ftrace_make_nop/ftrace_make_call for
			 * correct previous instructions.  Making first the NOP
			 * conversion puts the module to the correct state, thus
			 * passing the ftrace_make_call check.
			 */
			if (ftrace_start_up && ref) {
				int failed = __ftrace_replace_code(p, 1);
				if (failed)
					ftrace_bug(failed, p->ip);
			}
2096
		}
2097 2098
	}

2099 2100
	ftrace_new_pgs = NULL;

I
Ingo Molnar 已提交
2101
	stop = ftrace_now(raw_smp_processor_id());
2102 2103 2104
	ftrace_update_time = stop - start;
	ftrace_update_tot_cnt += ftrace_update_cnt;

2105 2106 2107
	return 0;
}

2108
static int ftrace_allocate_records(struct ftrace_page *pg, int count)
2109
{
2110
	int order;
2111 2112
	int cnt;

2113 2114 2115 2116
	if (WARN_ON(!count))
		return -EINVAL;

	order = get_count_order(DIV_ROUND_UP(count, ENTRIES_PER_PAGE));
2117 2118

	/*
2119 2120
	 * We want to fill as much as possible. No more than a page
	 * may be empty.
2121
	 */
2122 2123
	while ((PAGE_SIZE << order) / ENTRY_SIZE >= count + ENTRIES_PER_PAGE)
		order--;
2124

2125 2126
 again:
	pg->records = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
2127

2128 2129 2130 2131 2132 2133 2134
	if (!pg->records) {
		/* if we can't allocate this size, try something smaller */
		if (!order)
			return -ENOMEM;
		order >>= 1;
		goto again;
	}
2135

2136 2137
	cnt = (PAGE_SIZE << order) / ENTRY_SIZE;
	pg->size = cnt;
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 2168 2169 2170 2171
	if (cnt > count)
		cnt = count;

	return cnt;
}

static struct ftrace_page *
ftrace_allocate_pages(unsigned long num_to_init)
{
	struct ftrace_page *start_pg;
	struct ftrace_page *pg;
	int order;
	int cnt;

	if (!num_to_init)
		return 0;

	start_pg = pg = kzalloc(sizeof(*pg), GFP_KERNEL);
	if (!pg)
		return NULL;

	/*
	 * Try to allocate as much as possible in one continues
	 * location that fills in all of the space. We want to
	 * waste as little space as possible.
	 */
	for (;;) {
		cnt = ftrace_allocate_records(pg, num_to_init);
		if (cnt < 0)
			goto free_pages;

		num_to_init -= cnt;
		if (!num_to_init)
2172 2173
			break;

2174 2175 2176 2177
		pg->next = kzalloc(sizeof(*pg), GFP_KERNEL);
		if (!pg->next)
			goto free_pages;

2178 2179 2180
		pg = pg->next;
	}

2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207
	return start_pg;

 free_pages:
	while (start_pg) {
		order = get_count_order(pg->size / ENTRIES_PER_PAGE);
		free_pages((unsigned long)pg->records, order);
		start_pg = pg->next;
		kfree(pg);
		pg = start_pg;
	}
	pr_info("ftrace: FAILED to allocate memory for functions\n");
	return NULL;
}

static int __init ftrace_dyn_table_alloc(unsigned long num_to_init)
{
	int cnt;

	if (!num_to_init) {
		pr_info("ftrace: No functions to be traced?\n");
		return -1;
	}

	cnt = num_to_init / ENTRIES_PER_PAGE;
	pr_info("ftrace: allocating %ld entries in %d pages\n",
		num_to_init, cnt + 1);

2208 2209 2210
	return 0;
}

2211 2212 2213
#define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */

struct ftrace_iterator {
2214
	loff_t				pos;
2215 2216 2217 2218 2219
	loff_t				func_pos;
	struct ftrace_page		*pg;
	struct dyn_ftrace		*func;
	struct ftrace_func_probe	*probe;
	struct trace_parser		parser;
2220
	struct ftrace_hash		*hash;
2221
	struct ftrace_ops		*ops;
2222 2223 2224
	int				hidx;
	int				idx;
	unsigned			flags;
2225 2226
};

2227
static void *
2228
t_hash_next(struct seq_file *m, loff_t *pos)
2229 2230
{
	struct ftrace_iterator *iter = m->private;
2231
	struct hlist_node *hnd = NULL;
2232 2233 2234
	struct hlist_head *hhd;

	(*pos)++;
2235
	iter->pos = *pos;
2236

2237 2238
	if (iter->probe)
		hnd = &iter->probe->node;
2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260
 retry:
	if (iter->hidx >= FTRACE_FUNC_HASHSIZE)
		return NULL;

	hhd = &ftrace_func_hash[iter->hidx];

	if (hlist_empty(hhd)) {
		iter->hidx++;
		hnd = NULL;
		goto retry;
	}

	if (!hnd)
		hnd = hhd->first;
	else {
		hnd = hnd->next;
		if (!hnd) {
			iter->hidx++;
			goto retry;
		}
	}

2261 2262 2263 2264 2265 2266
	if (WARN_ON_ONCE(!hnd))
		return NULL;

	iter->probe = hlist_entry(hnd, struct ftrace_func_probe, node);

	return iter;
2267 2268 2269 2270 2271 2272
}

static void *t_hash_start(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
	void *p = NULL;
L
Li Zefan 已提交
2273 2274
	loff_t l;

2275 2276 2277
	if (!(iter->flags & FTRACE_ITER_DO_HASH))
		return NULL;

2278 2279
	if (iter->func_pos > *pos)
		return NULL;
2280

L
Li Zefan 已提交
2281
	iter->hidx = 0;
2282
	for (l = 0; l <= (*pos - iter->func_pos); ) {
2283
		p = t_hash_next(m, &l);
L
Li Zefan 已提交
2284 2285 2286
		if (!p)
			break;
	}
2287 2288 2289
	if (!p)
		return NULL;

2290 2291 2292
	/* Only set this if we have an item */
	iter->flags |= FTRACE_ITER_HASH;

2293
	return iter;
2294 2295
}

2296 2297
static int
t_hash_show(struct seq_file *m, struct ftrace_iterator *iter)
2298
{
S
Steven Rostedt 已提交
2299
	struct ftrace_func_probe *rec;
2300

2301 2302 2303
	rec = iter->probe;
	if (WARN_ON_ONCE(!rec))
		return -EIO;
2304

2305 2306 2307
	if (rec->ops->print)
		return rec->ops->print(m, rec->ip, rec->ops, rec->data);

2308
	seq_printf(m, "%ps:%ps", (void *)rec->ip, (void *)rec->ops->func);
2309 2310 2311 2312 2313 2314 2315 2316

	if (rec->data)
		seq_printf(m, ":%p", rec->data);
	seq_putc(m, '\n');

	return 0;
}

I
Ingo Molnar 已提交
2317
static void *
2318 2319 2320
t_next(struct seq_file *m, void *v, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
2321
	struct ftrace_ops *ops = iter->ops;
2322 2323
	struct dyn_ftrace *rec = NULL;

2324 2325 2326
	if (unlikely(ftrace_disabled))
		return NULL;

2327
	if (iter->flags & FTRACE_ITER_HASH)
2328
		return t_hash_next(m, pos);
2329

2330
	(*pos)++;
2331
	iter->pos = iter->func_pos = *pos;
2332

2333
	if (iter->flags & FTRACE_ITER_PRINTALL)
2334
		return t_hash_start(m, pos);
2335

2336 2337 2338 2339 2340 2341 2342 2343 2344
 retry:
	if (iter->idx >= iter->pg->index) {
		if (iter->pg->next) {
			iter->pg = iter->pg->next;
			iter->idx = 0;
			goto retry;
		}
	} else {
		rec = &iter->pg->records[iter->idx++];
2345
		if (((iter->flags & FTRACE_ITER_FILTER) &&
2346
		     !(ftrace_lookup_ip(ops->filter_hash, rec->ip))) ||
S
Steven Rostedt 已提交
2347

2348
		    ((iter->flags & FTRACE_ITER_NOTRACE) &&
2349 2350 2351 2352 2353
		     !ftrace_lookup_ip(ops->notrace_hash, rec->ip)) ||

		    ((iter->flags & FTRACE_ITER_ENABLED) &&
		     !(rec->flags & ~FTRACE_FL_MASK))) {

2354 2355 2356 2357 2358
			rec = NULL;
			goto retry;
		}
	}

2359
	if (!rec)
2360
		return t_hash_start(m, pos);
2361 2362 2363 2364

	iter->func = rec;

	return iter;
2365 2366
}

2367 2368 2369 2370 2371
static void reset_iter_read(struct ftrace_iterator *iter)
{
	iter->pos = 0;
	iter->func_pos = 0;
	iter->flags &= ~(FTRACE_ITER_PRINTALL & FTRACE_ITER_HASH);
2372 2373 2374 2375 2376
}

static void *t_start(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
2377
	struct ftrace_ops *ops = iter->ops;
2378
	void *p = NULL;
2379
	loff_t l;
2380

2381
	mutex_lock(&ftrace_lock);
2382 2383 2384 2385

	if (unlikely(ftrace_disabled))
		return NULL;

2386 2387 2388 2389 2390 2391
	/*
	 * If an lseek was done, then reset and start from beginning.
	 */
	if (*pos < iter->pos)
		reset_iter_read(iter);

2392 2393 2394 2395 2396
	/*
	 * For set_ftrace_filter reading, if we have the filter
	 * off, we can short cut and just print out that all
	 * functions are enabled.
	 */
2397 2398
	if (iter->flags & FTRACE_ITER_FILTER &&
	    ftrace_hash_empty(ops->filter_hash)) {
2399
		if (*pos > 0)
2400
			return t_hash_start(m, pos);
2401
		iter->flags |= FTRACE_ITER_PRINTALL;
2402 2403
		/* reset in case of seek/pread */
		iter->flags &= ~FTRACE_ITER_HASH;
2404 2405 2406
		return iter;
	}

2407 2408 2409
	if (iter->flags & FTRACE_ITER_HASH)
		return t_hash_start(m, pos);

2410 2411 2412 2413 2414
	/*
	 * Unfortunately, we need to restart at ftrace_pages_start
	 * every time we let go of the ftrace_mutex. This is because
	 * those pointers can change without the lock.
	 */
2415 2416 2417 2418 2419 2420
	iter->pg = ftrace_pages_start;
	iter->idx = 0;
	for (l = 0; l <= *pos; ) {
		p = t_next(m, p, &l);
		if (!p)
			break;
2421
	}
2422

2423 2424
	if (!p)
		return t_hash_start(m, pos);
2425 2426

	return iter;
2427 2428 2429 2430
}

static void t_stop(struct seq_file *m, void *p)
{
2431
	mutex_unlock(&ftrace_lock);
2432 2433 2434 2435
}

static int t_show(struct seq_file *m, void *v)
{
2436
	struct ftrace_iterator *iter = m->private;
2437
	struct dyn_ftrace *rec;
2438

2439
	if (iter->flags & FTRACE_ITER_HASH)
2440
		return t_hash_show(m, iter);
2441

2442 2443 2444 2445 2446
	if (iter->flags & FTRACE_ITER_PRINTALL) {
		seq_printf(m, "#### all functions enabled ####\n");
		return 0;
	}

2447 2448
	rec = iter->func;

2449 2450 2451
	if (!rec)
		return 0;

2452 2453 2454 2455 2456
	seq_printf(m, "%ps", (void *)rec->ip);
	if (iter->flags & FTRACE_ITER_ENABLED)
		seq_printf(m, " (%ld)",
			   rec->flags & ~FTRACE_FL_MASK);
	seq_printf(m, "\n");
2457 2458 2459 2460

	return 0;
}

J
James Morris 已提交
2461
static const struct seq_operations show_ftrace_seq_ops = {
2462 2463 2464 2465 2466 2467
	.start = t_start,
	.next = t_next,
	.stop = t_stop,
	.show = t_show,
};

I
Ingo Molnar 已提交
2468
static int
2469 2470 2471 2472
ftrace_avail_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

2473 2474 2475
	if (unlikely(ftrace_disabled))
		return -ENODEV;

2476 2477 2478 2479
	iter = __seq_open_private(file, &show_ftrace_seq_ops, sizeof(*iter));
	if (iter) {
		iter->pg = ftrace_pages_start;
		iter->ops = &global_ops;
I
Ingo Molnar 已提交
2480
	}
2481

2482
	return iter ? 0 : -ENOMEM;
2483 2484
}

2485 2486 2487 2488 2489 2490 2491 2492
static int
ftrace_enabled_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

	if (unlikely(ftrace_disabled))
		return -ENODEV;

2493 2494 2495 2496 2497
	iter = __seq_open_private(file, &show_ftrace_seq_ops, sizeof(*iter));
	if (iter) {
		iter->pg = ftrace_pages_start;
		iter->flags = FTRACE_ITER_ENABLED;
		iter->ops = &global_ops;
2498 2499
	}

2500
	return iter ? 0 : -ENOMEM;
2501 2502
}

2503
static void ftrace_filter_reset(struct ftrace_hash *hash)
2504
{
2505
	mutex_lock(&ftrace_lock);
2506
	ftrace_hash_clear(hash);
2507
	mutex_unlock(&ftrace_lock);
2508 2509
}

2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526
/**
 * ftrace_regex_open - initialize function tracer filter files
 * @ops: The ftrace_ops that hold the hash filters
 * @flag: The type of filter to process
 * @inode: The inode, usually passed in to your open routine
 * @file: The file, usually passed in to your open routine
 *
 * ftrace_regex_open() initializes the filter files for the
 * @ops. Depending on @flag it may process the filter hash or
 * the notrace hash of @ops. With this called from the open
 * routine, you can use ftrace_filter_write() for the write
 * routine if @flag has FTRACE_ITER_FILTER set, or
 * ftrace_notrace_write() if @flag has FTRACE_ITER_NOTRACE set.
 * ftrace_regex_lseek() should be used as the lseek routine, and
 * release must call ftrace_regex_release().
 */
int
2527
ftrace_regex_open(struct ftrace_ops *ops, int flag,
2528
		  struct inode *inode, struct file *file)
2529 2530
{
	struct ftrace_iterator *iter;
2531
	struct ftrace_hash *hash;
2532 2533
	int ret = 0;

2534 2535 2536
	if (unlikely(ftrace_disabled))
		return -ENODEV;

2537 2538 2539 2540
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter)
		return -ENOMEM;

2541 2542 2543 2544 2545
	if (trace_parser_get_init(&iter->parser, FTRACE_BUFF_MAX)) {
		kfree(iter);
		return -ENOMEM;
	}

2546 2547 2548 2549 2550
	if (flag & FTRACE_ITER_NOTRACE)
		hash = ops->notrace_hash;
	else
		hash = ops->filter_hash;

2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564
	iter->ops = ops;
	iter->flags = flag;

	if (file->f_mode & FMODE_WRITE) {
		mutex_lock(&ftrace_lock);
		iter->hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, hash);
		mutex_unlock(&ftrace_lock);

		if (!iter->hash) {
			trace_parser_put(&iter->parser);
			kfree(iter);
			return -ENOMEM;
		}
	}
2565

2566
	mutex_lock(&ftrace_regex_lock);
2567

2568
	if ((file->f_mode & FMODE_WRITE) &&
2569
	    (file->f_flags & O_TRUNC))
2570
		ftrace_filter_reset(iter->hash);
2571 2572 2573 2574 2575 2576 2577 2578

	if (file->f_mode & FMODE_READ) {
		iter->pg = ftrace_pages_start;

		ret = seq_open(file, &show_ftrace_seq_ops);
		if (!ret) {
			struct seq_file *m = file->private_data;
			m->private = iter;
2579
		} else {
2580 2581
			/* Failed */
			free_ftrace_hash(iter->hash);
2582
			trace_parser_put(&iter->parser);
2583
			kfree(iter);
2584
		}
2585 2586
	} else
		file->private_data = iter;
2587
	mutex_unlock(&ftrace_regex_lock);
2588 2589 2590 2591

	return ret;
}

2592 2593 2594
static int
ftrace_filter_open(struct inode *inode, struct file *file)
{
2595 2596 2597
	return ftrace_regex_open(&global_ops,
			FTRACE_ITER_FILTER | FTRACE_ITER_DO_HASH,
			inode, file);
2598 2599 2600 2601 2602
}

static int
ftrace_notrace_open(struct inode *inode, struct file *file)
{
2603
	return ftrace_regex_open(&global_ops, FTRACE_ITER_NOTRACE,
2604
				 inode, file);
2605 2606
}

2607
loff_t
2608
ftrace_regex_lseek(struct file *file, loff_t offset, int origin)
2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619
{
	loff_t ret;

	if (file->f_mode & FMODE_READ)
		ret = seq_lseek(file, offset, origin);
	else
		file->f_pos = ret = 1;

	return ret;
}

2620
static int ftrace_match(char *str, char *regex, int len, int type)
2621 2622
{
	int matched = 0;
2623
	int slen;
2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638

	switch (type) {
	case MATCH_FULL:
		if (strcmp(str, regex) == 0)
			matched = 1;
		break;
	case MATCH_FRONT_ONLY:
		if (strncmp(str, regex, len) == 0)
			matched = 1;
		break;
	case MATCH_MIDDLE_ONLY:
		if (strstr(str, regex))
			matched = 1;
		break;
	case MATCH_END_ONLY:
2639 2640
		slen = strlen(str);
		if (slen >= len && memcmp(str + slen - len, regex, len) == 0)
2641 2642 2643 2644 2645 2646 2647
			matched = 1;
		break;
	}

	return matched;
}

2648
static int
2649
enter_record(struct ftrace_hash *hash, struct dyn_ftrace *rec, int not)
2650
{
2651 2652 2653
	struct ftrace_func_entry *entry;
	int ret = 0;

2654 2655 2656 2657 2658
	entry = ftrace_lookup_ip(hash, rec->ip);
	if (not) {
		/* Do nothing if it doesn't exist */
		if (!entry)
			return 0;
2659

2660
		free_hash_entry(hash, entry);
2661 2662 2663 2664
	} else {
		/* Do nothing if it exists */
		if (entry)
			return 0;
2665

2666
		ret = add_hash_entry(hash, rec->ip);
2667 2668
	}
	return ret;
2669 2670
}

2671
static int
2672 2673
ftrace_match_record(struct dyn_ftrace *rec, char *mod,
		    char *regex, int len, int type)
2674 2675
{
	char str[KSYM_SYMBOL_LEN];
2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688
	char *modname;

	kallsyms_lookup(rec->ip, NULL, NULL, &modname, str);

	if (mod) {
		/* module lookup requires matching the module */
		if (!modname || strcmp(modname, mod))
			return 0;

		/* blank search means to match all funcs in the mod */
		if (!len)
			return 1;
	}
2689 2690 2691 2692

	return ftrace_match(str, regex, len, type);
}

2693 2694 2695
static int
match_records(struct ftrace_hash *hash, char *buff,
	      int len, char *mod, int not)
2696
{
2697
	unsigned search_len = 0;
2698 2699
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
2700 2701
	int type = MATCH_FULL;
	char *search = buff;
2702
	int found = 0;
2703
	int ret;
2704

2705 2706 2707 2708
	if (len) {
		type = filter_parse_regex(buff, len, &search, &not);
		search_len = strlen(search);
	}
2709

2710
	mutex_lock(&ftrace_lock);
2711

2712 2713
	if (unlikely(ftrace_disabled))
		goto out_unlock;
2714

2715
	do_for_each_ftrace_rec(pg, rec) {
2716
		if (ftrace_match_record(rec, mod, search, search_len, type)) {
2717
			ret = enter_record(hash, rec, not);
2718 2719 2720 2721
			if (ret < 0) {
				found = ret;
				goto out_unlock;
			}
2722
			found = 1;
2723 2724
		}
	} while_for_each_ftrace_rec();
2725
 out_unlock:
2726
	mutex_unlock(&ftrace_lock);
2727 2728

	return found;
2729 2730
}

2731
static int
2732
ftrace_match_records(struct ftrace_hash *hash, char *buff, int len)
2733
{
2734
	return match_records(hash, buff, len, NULL, 0);
2735 2736
}

2737 2738
static int
ftrace_match_module_records(struct ftrace_hash *hash, char *buff, char *mod)
2739 2740
{
	int not = 0;
S
Steven Rostedt 已提交
2741

2742 2743 2744 2745 2746 2747 2748 2749 2750 2751
	/* blank or '*' mean the same */
	if (strcmp(buff, "*") == 0)
		buff[0] = 0;

	/* handle the case of 'dont filter this module' */
	if (strcmp(buff, "!") == 0 || strcmp(buff, "!*") == 0) {
		buff[0] = 0;
		not = 1;
	}

2752
	return match_records(hash, buff, strlen(buff), mod, not);
2753 2754
}

2755 2756 2757 2758 2759 2760
/*
 * We register the module command as a template to show others how
 * to register the a command as well.
 */

static int
2761 2762
ftrace_mod_callback(struct ftrace_hash *hash,
		    char *func, char *cmd, char *param, int enable)
2763 2764
{
	char *mod;
2765
	int ret = -EINVAL;
2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776

	/*
	 * cmd == 'mod' because we only registered this func
	 * for the 'mod' ftrace_func_command.
	 * But if you register one func with multiple commands,
	 * you can tell which command was used by the cmd
	 * parameter.
	 */

	/* we must have a module name */
	if (!param)
2777
		return ret;
2778 2779 2780

	mod = strsep(&param, ":");
	if (!strlen(mod))
2781
		return ret;
2782

2783
	ret = ftrace_match_module_records(hash, func, mod);
2784 2785 2786 2787 2788 2789
	if (!ret)
		ret = -EINVAL;
	if (ret < 0)
		return ret;

	return 0;
2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802
}

static struct ftrace_func_command ftrace_mod_cmd = {
	.name			= "mod",
	.func			= ftrace_mod_callback,
};

static int __init ftrace_mod_cmd_init(void)
{
	return register_ftrace_command(&ftrace_mod_cmd);
}
device_initcall(ftrace_mod_cmd_init);

2803
static void
S
Steven Rostedt 已提交
2804
function_trace_probe_call(unsigned long ip, unsigned long parent_ip)
2805
{
S
Steven Rostedt 已提交
2806
	struct ftrace_func_probe *entry;
2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822
	struct hlist_head *hhd;
	struct hlist_node *n;
	unsigned long key;

	key = hash_long(ip, FTRACE_HASH_BITS);

	hhd = &ftrace_func_hash[key];

	if (hlist_empty(hhd))
		return;

	/*
	 * Disable preemption for these calls to prevent a RCU grace
	 * period. This syncs the hash iteration and freeing of items
	 * on the hash. rcu_read_lock is too dangerous here.
	 */
2823
	preempt_disable_notrace();
2824 2825 2826 2827
	hlist_for_each_entry_rcu(entry, n, hhd, node) {
		if (entry->ip == ip)
			entry->ops->func(ip, parent_ip, &entry->data);
	}
2828
	preempt_enable_notrace();
2829 2830
}

S
Steven Rostedt 已提交
2831
static struct ftrace_ops trace_probe_ops __read_mostly =
2832
{
2833
	.func		= function_trace_probe_call,
2834 2835
};

S
Steven Rostedt 已提交
2836
static int ftrace_probe_registered;
2837

S
Steven Rostedt 已提交
2838
static void __enable_ftrace_function_probe(void)
2839
{
2840
	int ret;
2841 2842
	int i;

S
Steven Rostedt 已提交
2843
	if (ftrace_probe_registered)
2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854
		return;

	for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) {
		struct hlist_head *hhd = &ftrace_func_hash[i];
		if (hhd->first)
			break;
	}
	/* Nothing registered? */
	if (i == FTRACE_FUNC_HASHSIZE)
		return;

2855 2856
	ret = __register_ftrace_function(&trace_probe_ops);
	if (!ret)
2857
		ret = ftrace_startup(&trace_probe_ops, 0);
2858

S
Steven Rostedt 已提交
2859
	ftrace_probe_registered = 1;
2860 2861
}

S
Steven Rostedt 已提交
2862
static void __disable_ftrace_function_probe(void)
2863
{
2864
	int ret;
2865 2866
	int i;

S
Steven Rostedt 已提交
2867
	if (!ftrace_probe_registered)
2868 2869 2870 2871 2872 2873 2874 2875 2876
		return;

	for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) {
		struct hlist_head *hhd = &ftrace_func_hash[i];
		if (hhd->first)
			return;
	}

	/* no more funcs left */
2877 2878 2879 2880
	ret = __unregister_ftrace_function(&trace_probe_ops);
	if (!ret)
		ftrace_shutdown(&trace_probe_ops, 0);

S
Steven Rostedt 已提交
2881
	ftrace_probe_registered = 0;
2882 2883 2884 2885 2886
}


static void ftrace_free_entry_rcu(struct rcu_head *rhp)
{
S
Steven Rostedt 已提交
2887 2888
	struct ftrace_func_probe *entry =
		container_of(rhp, struct ftrace_func_probe, rcu);
2889 2890 2891 2892 2893 2894 2895 2896

	if (entry->ops->free)
		entry->ops->free(&entry->data);
	kfree(entry);
}


int
S
Steven Rostedt 已提交
2897
register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
2898 2899
			      void *data)
{
S
Steven Rostedt 已提交
2900
	struct ftrace_func_probe *entry;
2901 2902 2903
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
	int type, len, not;
S
Steven Rostedt 已提交
2904
	unsigned long key;
2905 2906 2907
	int count = 0;
	char *search;

2908
	type = filter_parse_regex(glob, strlen(glob), &search, &not);
2909 2910
	len = strlen(search);

S
Steven Rostedt 已提交
2911
	/* we do not support '!' for function probes */
2912 2913 2914 2915 2916
	if (WARN_ON(not))
		return -EINVAL;

	mutex_lock(&ftrace_lock);

2917 2918
	if (unlikely(ftrace_disabled))
		goto out_unlock;
2919

2920
	do_for_each_ftrace_rec(pg, rec) {
2921

2922
		if (!ftrace_match_record(rec, NULL, search, len, type))
2923 2924 2925 2926
			continue;

		entry = kmalloc(sizeof(*entry), GFP_KERNEL);
		if (!entry) {
S
Steven Rostedt 已提交
2927
			/* If we did not process any, then return error */
2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956
			if (!count)
				count = -ENOMEM;
			goto out_unlock;
		}

		count++;

		entry->data = data;

		/*
		 * The caller might want to do something special
		 * for each function we find. We call the callback
		 * to give the caller an opportunity to do so.
		 */
		if (ops->callback) {
			if (ops->callback(rec->ip, &entry->data) < 0) {
				/* caller does not like this func */
				kfree(entry);
				continue;
			}
		}

		entry->ops = ops;
		entry->ip = rec->ip;

		key = hash_long(entry->ip, FTRACE_HASH_BITS);
		hlist_add_head_rcu(&entry->node, &ftrace_func_hash[key]);

	} while_for_each_ftrace_rec();
S
Steven Rostedt 已提交
2957
	__enable_ftrace_function_probe();
2958 2959 2960 2961 2962 2963 2964 2965

 out_unlock:
	mutex_unlock(&ftrace_lock);

	return count;
}

enum {
S
Steven Rostedt 已提交
2966 2967
	PROBE_TEST_FUNC		= 1,
	PROBE_TEST_DATA		= 2
2968 2969 2970
};

static void
S
Steven Rostedt 已提交
2971
__unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
2972 2973
				  void *data, int flags)
{
S
Steven Rostedt 已提交
2974
	struct ftrace_func_probe *entry;
2975 2976 2977 2978 2979 2980
	struct hlist_node *n, *tmp;
	char str[KSYM_SYMBOL_LEN];
	int type = MATCH_FULL;
	int i, len = 0;
	char *search;

2981
	if (glob && (strcmp(glob, "*") == 0 || !strlen(glob)))
2982
		glob = NULL;
2983
	else if (glob) {
2984 2985
		int not;

2986
		type = filter_parse_regex(glob, strlen(glob), &search, &not);
2987 2988
		len = strlen(search);

S
Steven Rostedt 已提交
2989
		/* we do not support '!' for function probes */
2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000
		if (WARN_ON(not))
			return;
	}

	mutex_lock(&ftrace_lock);
	for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) {
		struct hlist_head *hhd = &ftrace_func_hash[i];

		hlist_for_each_entry_safe(entry, n, tmp, hhd, node) {

			/* break up if statements for readability */
S
Steven Rostedt 已提交
3001
			if ((flags & PROBE_TEST_FUNC) && entry->ops != ops)
3002 3003
				continue;

S
Steven Rostedt 已提交
3004
			if ((flags & PROBE_TEST_DATA) && entry->data != data)
3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018
				continue;

			/* do this last, since it is the most expensive */
			if (glob) {
				kallsyms_lookup(entry->ip, NULL, NULL,
						NULL, str);
				if (!ftrace_match(str, glob, len, type))
					continue;
			}

			hlist_del(&entry->node);
			call_rcu(&entry->rcu, ftrace_free_entry_rcu);
		}
	}
S
Steven Rostedt 已提交
3019
	__disable_ftrace_function_probe();
3020 3021 3022 3023
	mutex_unlock(&ftrace_lock);
}

void
S
Steven Rostedt 已提交
3024
unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3025 3026
				void *data)
{
S
Steven Rostedt 已提交
3027 3028
	__unregister_ftrace_function_probe(glob, ops, data,
					  PROBE_TEST_FUNC | PROBE_TEST_DATA);
3029 3030 3031
}

void
S
Steven Rostedt 已提交
3032
unregister_ftrace_function_probe_func(char *glob, struct ftrace_probe_ops *ops)
3033
{
S
Steven Rostedt 已提交
3034
	__unregister_ftrace_function_probe(glob, ops, NULL, PROBE_TEST_FUNC);
3035 3036
}

S
Steven Rostedt 已提交
3037
void unregister_ftrace_function_probe_all(char *glob)
3038
{
S
Steven Rostedt 已提交
3039
	__unregister_ftrace_function_probe(glob, NULL, NULL, 0);
3040 3041
}

3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082
static LIST_HEAD(ftrace_commands);
static DEFINE_MUTEX(ftrace_cmd_mutex);

int register_ftrace_command(struct ftrace_func_command *cmd)
{
	struct ftrace_func_command *p;
	int ret = 0;

	mutex_lock(&ftrace_cmd_mutex);
	list_for_each_entry(p, &ftrace_commands, list) {
		if (strcmp(cmd->name, p->name) == 0) {
			ret = -EBUSY;
			goto out_unlock;
		}
	}
	list_add(&cmd->list, &ftrace_commands);
 out_unlock:
	mutex_unlock(&ftrace_cmd_mutex);

	return ret;
}

int unregister_ftrace_command(struct ftrace_func_command *cmd)
{
	struct ftrace_func_command *p, *n;
	int ret = -ENODEV;

	mutex_lock(&ftrace_cmd_mutex);
	list_for_each_entry_safe(p, n, &ftrace_commands, list) {
		if (strcmp(cmd->name, p->name) == 0) {
			ret = 0;
			list_del_init(&p->list);
			goto out_unlock;
		}
	}
 out_unlock:
	mutex_unlock(&ftrace_cmd_mutex);

	return ret;
}

3083 3084
static int ftrace_process_regex(struct ftrace_hash *hash,
				char *buff, int len, int enable)
3085
{
3086
	char *func, *command, *next = buff;
S
Steven Rostedt 已提交
3087
	struct ftrace_func_command *p;
3088
	int ret = -EINVAL;
3089 3090 3091 3092

	func = strsep(&next, ":");

	if (!next) {
3093
		ret = ftrace_match_records(hash, func, len);
3094 3095 3096 3097 3098
		if (!ret)
			ret = -EINVAL;
		if (ret < 0)
			return ret;
		return 0;
3099 3100
	}

3101
	/* command found */
3102 3103 3104

	command = strsep(&next, ":");

3105 3106 3107
	mutex_lock(&ftrace_cmd_mutex);
	list_for_each_entry(p, &ftrace_commands, list) {
		if (strcmp(p->name, command) == 0) {
3108
			ret = p->func(hash, func, command, next, enable);
3109 3110
			goto out_unlock;
		}
3111
	}
3112 3113
 out_unlock:
	mutex_unlock(&ftrace_cmd_mutex);
3114

3115
	return ret;
3116 3117
}

I
Ingo Molnar 已提交
3118
static ssize_t
3119 3120
ftrace_regex_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos, int enable)
3121 3122
{
	struct ftrace_iterator *iter;
3123 3124
	struct trace_parser *parser;
	ssize_t ret, read;
3125

3126
	if (!cnt)
3127 3128
		return 0;

3129
	mutex_lock(&ftrace_regex_lock);
3130

3131 3132 3133 3134
	ret = -ENODEV;
	if (unlikely(ftrace_disabled))
		goto out_unlock;

3135 3136 3137 3138 3139 3140
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;
		iter = m->private;
	} else
		iter = file->private_data;

3141 3142
	parser = &iter->parser;
	read = trace_get_user(parser, ubuf, cnt, ppos);
3143

3144
	if (read >= 0 && trace_parser_loaded(parser) &&
3145
	    !trace_parser_cont(parser)) {
3146
		ret = ftrace_process_regex(iter->hash, parser->buffer,
3147
					   parser->idx, enable);
3148
		trace_parser_clear(parser);
3149
		if (ret)
3150
			goto out_unlock;
3151
	}
3152 3153

	ret = read;
3154
out_unlock:
3155
	mutex_unlock(&ftrace_regex_lock);
3156

3157 3158 3159
	return ret;
}

3160
ssize_t
3161 3162 3163 3164 3165 3166
ftrace_filter_write(struct file *file, const char __user *ubuf,
		    size_t cnt, loff_t *ppos)
{
	return ftrace_regex_write(file, ubuf, cnt, ppos, 1);
}

3167
ssize_t
3168 3169 3170 3171 3172 3173
ftrace_notrace_write(struct file *file, const char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
	return ftrace_regex_write(file, ubuf, cnt, ppos, 0);
}

3174
static int
3175 3176
ftrace_set_regex(struct ftrace_ops *ops, unsigned char *buf, int len,
		 int reset, int enable)
3177
{
3178
	struct ftrace_hash **orig_hash;
3179
	struct ftrace_hash *hash;
3180
	int ret;
3181

3182 3183 3184 3185
	/* All global ops uses the global ops filters */
	if (ops->flags & FTRACE_OPS_FL_GLOBAL)
		ops = &global_ops;

3186
	if (unlikely(ftrace_disabled))
3187
		return -ENODEV;
3188

3189
	if (enable)
3190
		orig_hash = &ops->filter_hash;
3191
	else
3192 3193 3194 3195 3196
		orig_hash = &ops->notrace_hash;

	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);
	if (!hash)
		return -ENOMEM;
3197

3198 3199
	mutex_lock(&ftrace_regex_lock);
	if (reset)
3200
		ftrace_filter_reset(hash);
3201 3202 3203 3204
	if (buf && !ftrace_match_records(hash, buf, len)) {
		ret = -EINVAL;
		goto out_regex_unlock;
	}
3205 3206

	mutex_lock(&ftrace_lock);
3207
	ret = ftrace_hash_move(ops, enable, orig_hash, hash);
3208 3209
	if (!ret && ops->flags & FTRACE_OPS_FL_ENABLED
	    && ftrace_enabled)
3210
		ftrace_run_update_code(FTRACE_UPDATE_CALLS);
3211

3212 3213
	mutex_unlock(&ftrace_lock);

3214
 out_regex_unlock:
3215
	mutex_unlock(&ftrace_regex_lock);
3216 3217 3218

	free_ftrace_hash(hash);
	return ret;
3219 3220
}

3221 3222
/**
 * ftrace_set_filter - set a function to filter on in ftrace
3223 3224 3225 3226 3227 3228 3229 3230
 * @ops - the ops to set the filter with
 * @buf - the string that holds the function filter text.
 * @len - the length of the string.
 * @reset - non zero to reset all filters before applying this filter.
 *
 * Filters denote which functions should be enabled when tracing is enabled.
 * If @buf is NULL and reset is set, all functions will be enabled for tracing.
 */
3231
int ftrace_set_filter(struct ftrace_ops *ops, unsigned char *buf,
3232 3233
		       int len, int reset)
{
3234
	return ftrace_set_regex(ops, buf, len, reset, 1);
3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248
}
EXPORT_SYMBOL_GPL(ftrace_set_filter);

/**
 * ftrace_set_notrace - set a function to not trace in ftrace
 * @ops - the ops to set the notrace filter with
 * @buf - the string that holds the function notrace text.
 * @len - the length of the string.
 * @reset - non zero to reset all filters before applying this filter.
 *
 * Notrace Filters denote which functions should not be enabled when tracing
 * is enabled. If @buf is NULL and reset is set, all functions will be enabled
 * for tracing.
 */
3249
int ftrace_set_notrace(struct ftrace_ops *ops, unsigned char *buf,
3250 3251
			int len, int reset)
{
3252
	return ftrace_set_regex(ops, buf, len, reset, 0);
3253 3254 3255 3256 3257
}
EXPORT_SYMBOL_GPL(ftrace_set_notrace);
/**
 * ftrace_set_filter - set a function to filter on in ftrace
 * @ops - the ops to set the filter with
3258 3259 3260 3261 3262 3263 3264
 * @buf - the string that holds the function filter text.
 * @len - the length of the string.
 * @reset - non zero to reset all filters before applying this filter.
 *
 * Filters denote which functions should be enabled when tracing is enabled.
 * If @buf is NULL and reset is set, all functions will be enabled for tracing.
 */
3265
void ftrace_set_global_filter(unsigned char *buf, int len, int reset)
3266
{
3267
	ftrace_set_regex(&global_ops, buf, len, reset, 1);
3268
}
3269
EXPORT_SYMBOL_GPL(ftrace_set_global_filter);
3270

3271 3272
/**
 * ftrace_set_notrace - set a function to not trace in ftrace
3273
 * @ops - the ops to set the notrace filter with
3274 3275 3276 3277 3278 3279 3280 3281
 * @buf - the string that holds the function notrace text.
 * @len - the length of the string.
 * @reset - non zero to reset all filters before applying this filter.
 *
 * Notrace Filters denote which functions should not be enabled when tracing
 * is enabled. If @buf is NULL and reset is set, all functions will be enabled
 * for tracing.
 */
3282
void ftrace_set_global_notrace(unsigned char *buf, int len, int reset)
3283
{
3284
	ftrace_set_regex(&global_ops, buf, len, reset, 0);
3285
}
3286
EXPORT_SYMBOL_GPL(ftrace_set_global_notrace);
3287

3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308
/*
 * command line interface to allow users to set filters on boot up.
 */
#define FTRACE_FILTER_SIZE		COMMAND_LINE_SIZE
static char ftrace_notrace_buf[FTRACE_FILTER_SIZE] __initdata;
static char ftrace_filter_buf[FTRACE_FILTER_SIZE] __initdata;

static int __init set_ftrace_notrace(char *str)
{
	strncpy(ftrace_notrace_buf, str, FTRACE_FILTER_SIZE);
	return 1;
}
__setup("ftrace_notrace=", set_ftrace_notrace);

static int __init set_ftrace_filter(char *str)
{
	strncpy(ftrace_filter_buf, str, FTRACE_FILTER_SIZE);
	return 1;
}
__setup("ftrace_filter=", set_ftrace_filter);

3309
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3310
static char ftrace_graph_buf[FTRACE_FILTER_SIZE] __initdata;
3311 3312
static int ftrace_set_func(unsigned long *array, int *idx, char *buffer);

3313 3314
static int __init set_graph_function(char *str)
{
3315
	strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE);
3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336
	return 1;
}
__setup("ftrace_graph_filter=", set_graph_function);

static void __init set_ftrace_early_graph(char *buf)
{
	int ret;
	char *func;

	while (buf) {
		func = strsep(&buf, ",");
		/* we allow only one expression at a time */
		ret = ftrace_set_func(ftrace_graph_funcs, &ftrace_graph_count,
				      func);
		if (ret)
			printk(KERN_DEBUG "ftrace: function %s not "
					  "traceable\n", func);
	}
}
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

3337 3338
void __init
ftrace_set_early_filter(struct ftrace_ops *ops, char *buf, int enable)
3339 3340 3341 3342 3343
{
	char *func;

	while (buf) {
		func = strsep(&buf, ",");
3344
		ftrace_set_regex(ops, func, strlen(func), 0, enable);
3345 3346 3347 3348 3349 3350
	}
}

static void __init set_ftrace_early_filters(void)
{
	if (ftrace_filter_buf[0])
3351
		ftrace_set_early_filter(&global_ops, ftrace_filter_buf, 1);
3352
	if (ftrace_notrace_buf[0])
3353
		ftrace_set_early_filter(&global_ops, ftrace_notrace_buf, 0);
3354 3355 3356 3357
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	if (ftrace_graph_buf[0])
		set_ftrace_early_graph(ftrace_graph_buf);
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
3358 3359
}

3360
int ftrace_regex_release(struct inode *inode, struct file *file)
3361 3362 3363
{
	struct seq_file *m = (struct seq_file *)file->private_data;
	struct ftrace_iterator *iter;
3364
	struct ftrace_hash **orig_hash;
3365
	struct trace_parser *parser;
3366
	int filter_hash;
3367
	int ret;
3368

3369
	mutex_lock(&ftrace_regex_lock);
3370 3371 3372 3373 3374 3375 3376
	if (file->f_mode & FMODE_READ) {
		iter = m->private;

		seq_release(inode, file);
	} else
		iter = file->private_data;

3377 3378 3379
	parser = &iter->parser;
	if (trace_parser_loaded(parser)) {
		parser->buffer[parser->idx] = 0;
3380
		ftrace_match_records(iter->hash, parser->buffer, parser->idx);
3381 3382
	}

3383 3384
	trace_parser_put(parser);

3385
	if (file->f_mode & FMODE_WRITE) {
3386 3387 3388
		filter_hash = !!(iter->flags & FTRACE_ITER_FILTER);

		if (filter_hash)
3389
			orig_hash = &iter->ops->filter_hash;
3390 3391
		else
			orig_hash = &iter->ops->notrace_hash;
3392

3393
		mutex_lock(&ftrace_lock);
3394 3395 3396 3397
		ret = ftrace_hash_move(iter->ops, filter_hash,
				       orig_hash, iter->hash);
		if (!ret && (iter->ops->flags & FTRACE_OPS_FL_ENABLED)
		    && ftrace_enabled)
3398
			ftrace_run_update_code(FTRACE_UPDATE_CALLS);
3399

3400 3401
		mutex_unlock(&ftrace_lock);
	}
3402 3403
	free_ftrace_hash(iter->hash);
	kfree(iter);
3404

3405
	mutex_unlock(&ftrace_regex_lock);
3406 3407 3408
	return 0;
}

3409
static const struct file_operations ftrace_avail_fops = {
3410 3411 3412
	.open = ftrace_avail_open,
	.read = seq_read,
	.llseek = seq_lseek,
L
Li Zefan 已提交
3413
	.release = seq_release_private,
3414 3415
};

3416 3417 3418 3419 3420 3421 3422
static const struct file_operations ftrace_enabled_fops = {
	.open = ftrace_enabled_open,
	.read = seq_read,
	.llseek = seq_lseek,
	.release = seq_release_private,
};

3423
static const struct file_operations ftrace_filter_fops = {
3424
	.open = ftrace_filter_open,
L
Lai Jiangshan 已提交
3425
	.read = seq_read,
3426
	.write = ftrace_filter_write,
3427
	.llseek = ftrace_regex_lseek,
3428
	.release = ftrace_regex_release,
3429 3430
};

3431
static const struct file_operations ftrace_notrace_fops = {
3432
	.open = ftrace_notrace_open,
L
Lai Jiangshan 已提交
3433
	.read = seq_read,
3434 3435
	.write = ftrace_notrace_write,
	.llseek = ftrace_regex_lseek,
3436
	.release = ftrace_regex_release,
3437 3438
};

3439 3440 3441 3442 3443
#ifdef CONFIG_FUNCTION_GRAPH_TRACER

static DEFINE_MUTEX(graph_lock);

int ftrace_graph_count;
3444
int ftrace_graph_filter_enabled;
3445 3446 3447
unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly;

static void *
3448
__g_next(struct seq_file *m, loff_t *pos)
3449
{
3450
	if (*pos >= ftrace_graph_count)
3451
		return NULL;
3452
	return &ftrace_graph_funcs[*pos];
3453
}
3454

3455 3456 3457 3458 3459
static void *
g_next(struct seq_file *m, void *v, loff_t *pos)
{
	(*pos)++;
	return __g_next(m, pos);
3460 3461 3462 3463 3464 3465
}

static void *g_start(struct seq_file *m, loff_t *pos)
{
	mutex_lock(&graph_lock);

3466
	/* Nothing, tell g_show to print all functions are enabled */
3467
	if (!ftrace_graph_filter_enabled && !*pos)
3468 3469
		return (void *)1;

3470
	return __g_next(m, pos);
3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484
}

static void g_stop(struct seq_file *m, void *p)
{
	mutex_unlock(&graph_lock);
}

static int g_show(struct seq_file *m, void *v)
{
	unsigned long *ptr = v;

	if (!ptr)
		return 0;

3485 3486 3487 3488 3489
	if (ptr == (unsigned long *)1) {
		seq_printf(m, "#### all functions enabled ####\n");
		return 0;
	}

3490
	seq_printf(m, "%ps\n", (void *)*ptr);
3491 3492 3493 3494

	return 0;
}

J
James Morris 已提交
3495
static const struct seq_operations ftrace_graph_seq_ops = {
3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511
	.start = g_start,
	.next = g_next,
	.stop = g_stop,
	.show = g_show,
};

static int
ftrace_graph_open(struct inode *inode, struct file *file)
{
	int ret = 0;

	if (unlikely(ftrace_disabled))
		return -ENODEV;

	mutex_lock(&graph_lock);
	if ((file->f_mode & FMODE_WRITE) &&
3512
	    (file->f_flags & O_TRUNC)) {
3513
		ftrace_graph_filter_enabled = 0;
3514 3515 3516
		ftrace_graph_count = 0;
		memset(ftrace_graph_funcs, 0, sizeof(ftrace_graph_funcs));
	}
3517
	mutex_unlock(&graph_lock);
3518

3519
	if (file->f_mode & FMODE_READ)
3520 3521 3522 3523 3524
		ret = seq_open(file, &ftrace_graph_seq_ops);

	return ret;
}

3525 3526 3527 3528 3529 3530 3531 3532
static int
ftrace_graph_release(struct inode *inode, struct file *file)
{
	if (file->f_mode & FMODE_READ)
		seq_release(inode, file);
	return 0;
}

3533
static int
3534
ftrace_set_func(unsigned long *array, int *idx, char *buffer)
3535 3536 3537
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
3538
	int search_len;
3539
	int fail = 1;
3540 3541 3542 3543
	int type, not;
	char *search;
	bool exists;
	int i;
3544

3545
	/* decode regex */
3546
	type = filter_parse_regex(buffer, strlen(buffer), &search, &not);
3547 3548
	if (!not && *idx >= FTRACE_GRAPH_MAX_FUNCS)
		return -EBUSY;
3549 3550 3551

	search_len = strlen(search);

3552
	mutex_lock(&ftrace_lock);
3553 3554 3555 3556 3557 3558

	if (unlikely(ftrace_disabled)) {
		mutex_unlock(&ftrace_lock);
		return -ENODEV;
	}

3559 3560
	do_for_each_ftrace_rec(pg, rec) {

3561
		if (ftrace_match_record(rec, NULL, search, search_len, type)) {
3562
			/* if it is in the array */
3563
			exists = false;
3564
			for (i = 0; i < *idx; i++) {
3565 3566
				if (array[i] == rec->ip) {
					exists = true;
3567 3568
					break;
				}
3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584
			}

			if (!not) {
				fail = 0;
				if (!exists) {
					array[(*idx)++] = rec->ip;
					if (*idx >= FTRACE_GRAPH_MAX_FUNCS)
						goto out;
				}
			} else {
				if (exists) {
					array[i] = array[--(*idx)];
					array[*idx] = 0;
					fail = 0;
				}
			}
3585
		}
3586
	} while_for_each_ftrace_rec();
3587
out:
3588
	mutex_unlock(&ftrace_lock);
3589

3590 3591 3592 3593 3594
	if (fail)
		return -EINVAL;

	ftrace_graph_filter_enabled = 1;
	return 0;
3595 3596 3597 3598 3599 3600
}

static ssize_t
ftrace_graph_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
3601
	struct trace_parser parser;
3602
	ssize_t read, ret;
3603

3604
	if (!cnt)
3605 3606 3607 3608
		return 0;

	mutex_lock(&graph_lock);

3609 3610
	if (trace_parser_get_init(&parser, FTRACE_BUFF_MAX)) {
		ret = -ENOMEM;
3611
		goto out_unlock;
3612 3613
	}

3614
	read = trace_get_user(&parser, ubuf, cnt, ppos);
3615

3616
	if (read >= 0 && trace_parser_loaded((&parser))) {
3617 3618 3619
		parser.buffer[parser.idx] = 0;

		/* we allow only one expression at a time */
3620
		ret = ftrace_set_func(ftrace_graph_funcs, &ftrace_graph_count,
3621
					parser.buffer);
3622
		if (ret)
3623
			goto out_free;
3624 3625 3626
	}

	ret = read;
3627 3628

out_free:
3629
	trace_parser_put(&parser);
3630
out_unlock:
3631 3632 3633 3634 3635 3636
	mutex_unlock(&graph_lock);

	return ret;
}

static const struct file_operations ftrace_graph_fops = {
3637 3638 3639 3640
	.open		= ftrace_graph_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
	.release	= ftrace_graph_release,
3641
	.llseek		= seq_lseek,
3642 3643 3644
};
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

3645
static __init int ftrace_init_dyn_debugfs(struct dentry *d_tracer)
3646 3647
{

3648 3649
	trace_create_file("available_filter_functions", 0444,
			d_tracer, NULL, &ftrace_avail_fops);
3650

3651 3652 3653
	trace_create_file("enabled_functions", 0444,
			d_tracer, NULL, &ftrace_enabled_fops);

3654 3655
	trace_create_file("set_ftrace_filter", 0644, d_tracer,
			NULL, &ftrace_filter_fops);
3656

3657
	trace_create_file("set_ftrace_notrace", 0644, d_tracer,
3658
				    NULL, &ftrace_notrace_fops);
3659

3660
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3661
	trace_create_file("set_graph_function", 0444, d_tracer,
3662 3663 3664 3665
				    NULL,
				    &ftrace_graph_fops);
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

3666 3667 3668
	return 0;
}

3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679
static void ftrace_swap_recs(void *a, void *b, int size)
{
	struct dyn_ftrace *reca = a;
	struct dyn_ftrace *recb = b;
	struct dyn_ftrace t;

	t = *reca;
	*reca = *recb;
	*recb = t;
}

3680
static int ftrace_process_locs(struct module *mod,
3681
			       unsigned long *start,
3682 3683
			       unsigned long *end)
{
3684 3685
	struct ftrace_page *pg;
	unsigned long count;
3686 3687
	unsigned long *p;
	unsigned long addr;
3688
	unsigned long flags = 0; /* Shut up gcc */
3689 3690 3691 3692 3693 3694 3695 3696 3697 3698
	int ret = -ENOMEM;

	count = end - start;

	if (!count)
		return 0;

	pg = ftrace_allocate_pages(count);
	if (!pg)
		return -ENOMEM;
3699

S
Steven Rostedt 已提交
3700
	mutex_lock(&ftrace_lock);
3701

3702 3703 3704 3705 3706
	/*
	 * Core and each module needs their own pages, as
	 * modules will free them when they are removed.
	 * Force a new page to be allocated for modules.
	 */
3707 3708 3709 3710 3711
	if (!mod) {
		WARN_ON(ftrace_pages || ftrace_pages_start);
		/* First initialization */
		ftrace_pages = ftrace_pages_start = pg;
	} else {
3712
		if (!ftrace_pages)
3713
			goto out;
3714

3715 3716 3717 3718
		if (WARN_ON(ftrace_pages->next)) {
			/* Hmm, we have free pages? */
			while (ftrace_pages->next)
				ftrace_pages = ftrace_pages->next;
3719
		}
3720 3721 3722

		ftrace_pages->next = pg;
		ftrace_pages = pg;
3723 3724
	}

3725 3726 3727
	p = start;
	while (p < end) {
		addr = ftrace_call_adjust(*p++);
3728 3729 3730 3731 3732 3733 3734 3735
		/*
		 * Some architecture linkers will pad between
		 * the different mcount_loc sections of different
		 * object files to satisfy alignments.
		 * Skip any NULL pointers.
		 */
		if (!addr)
			continue;
3736 3737
		if (!ftrace_record_ip(addr))
			break;
3738 3739
	}

3740 3741 3742
	/* These new locations need to be initialized */
	ftrace_new_pgs = pg;

3743 3744 3745 3746 3747
	/* Make each individual set of pages sorted by ips */
	for (; pg; pg = pg->next)
		sort(pg->records, pg->index, sizeof(struct dyn_ftrace),
		     ftrace_cmp_recs, ftrace_swap_recs);

3748
	/*
3749 3750 3751 3752 3753 3754
	 * We only need to disable interrupts on start up
	 * because we are modifying code that an interrupt
	 * may execute, and the modification is not atomic.
	 * But for modules, nothing runs the code we modify
	 * until we are finished with it, and there's no
	 * reason to cause large interrupt latencies while we do it.
3755
	 */
3756 3757
	if (!mod)
		local_irq_save(flags);
3758
	ftrace_update_code(mod);
3759 3760
	if (!mod)
		local_irq_restore(flags);
3761 3762
	ret = 0;
 out:
S
Steven Rostedt 已提交
3763
	mutex_unlock(&ftrace_lock);
3764

3765
	return ret;
3766 3767
}

3768
#ifdef CONFIG_MODULES
3769 3770 3771

#define next_to_ftrace_page(p) container_of(p, struct ftrace_page, next)

3772
void ftrace_release_mod(struct module *mod)
3773 3774
{
	struct dyn_ftrace *rec;
3775
	struct ftrace_page **last_pg;
3776
	struct ftrace_page *pg;
3777
	int order;
3778

3779 3780
	mutex_lock(&ftrace_lock);

3781
	if (ftrace_disabled)
3782
		goto out_unlock;
3783

3784 3785 3786 3787 3788 3789 3790
	/*
	 * Each module has its own ftrace_pages, remove
	 * them from the list.
	 */
	last_pg = &ftrace_pages_start;
	for (pg = ftrace_pages_start; pg; pg = *last_pg) {
		rec = &pg->records[0];
3791
		if (within_module_core(rec->ip, mod)) {
3792
			/*
3793 3794
			 * As core pages are first, the first
			 * page should never be a module page.
3795
			 */
3796 3797 3798 3799 3800 3801 3802 3803
			if (WARN_ON(pg == ftrace_pages_start))
				goto out_unlock;

			/* Check if we are deleting the last page */
			if (pg == ftrace_pages)
				ftrace_pages = next_to_ftrace_page(last_pg);

			*last_pg = pg->next;
3804 3805 3806
			order = get_count_order(pg->size / ENTRIES_PER_PAGE);
			free_pages((unsigned long)pg->records, order);
			kfree(pg);
3807 3808 3809
		} else
			last_pg = &pg->next;
	}
3810
 out_unlock:
3811 3812 3813 3814 3815
	mutex_unlock(&ftrace_lock);
}

static void ftrace_init_module(struct module *mod,
			       unsigned long *start, unsigned long *end)
3816
{
3817
	if (ftrace_disabled || start == end)
3818
		return;
3819
	ftrace_process_locs(mod, start, end);
3820 3821
}

3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833
static int ftrace_module_notify(struct notifier_block *self,
				unsigned long val, void *data)
{
	struct module *mod = data;

	switch (val) {
	case MODULE_STATE_COMING:
		ftrace_init_module(mod, mod->ftrace_callsites,
				   mod->ftrace_callsites +
				   mod->num_ftrace_callsites);
		break;
	case MODULE_STATE_GOING:
3834
		ftrace_release_mod(mod);
3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852
		break;
	}

	return 0;
}
#else
static int ftrace_module_notify(struct notifier_block *self,
				unsigned long val, void *data)
{
	return 0;
}
#endif /* CONFIG_MODULES */

struct notifier_block ftrace_module_nb = {
	.notifier_call = ftrace_module_notify,
	.priority = 0,
};

3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879
extern unsigned long __start_mcount_loc[];
extern unsigned long __stop_mcount_loc[];

void __init ftrace_init(void)
{
	unsigned long count, addr, flags;
	int ret;

	/* Keep the ftrace pointer to the stub */
	addr = (unsigned long)ftrace_stub;

	local_irq_save(flags);
	ftrace_dyn_arch_init(&addr);
	local_irq_restore(flags);

	/* ftrace_dyn_arch_init places the return code in addr */
	if (addr)
		goto failed;

	count = __stop_mcount_loc - __start_mcount_loc;

	ret = ftrace_dyn_table_alloc(count);
	if (ret)
		goto failed;

	last_ftrace_enabled = ftrace_enabled = 1;

3880
	ret = ftrace_process_locs(NULL,
3881
				  __start_mcount_loc,
3882 3883
				  __stop_mcount_loc);

3884
	ret = register_module_notifier(&ftrace_module_nb);
3885
	if (ret)
3886 3887
		pr_warning("Failed to register trace ftrace module notifier\n");

3888 3889
	set_ftrace_early_filters();

3890 3891 3892 3893 3894
	return;
 failed:
	ftrace_disabled = 1;
}

3895
#else
3896

3897
static struct ftrace_ops global_ops = {
3898 3899 3900
	.func			= ftrace_stub,
};

3901 3902 3903 3904 3905 3906 3907
static int __init ftrace_nodyn_init(void)
{
	ftrace_enabled = 1;
	return 0;
}
device_initcall(ftrace_nodyn_init);

3908 3909
static inline int ftrace_init_dyn_debugfs(struct dentry *d_tracer) { return 0; }
static inline void ftrace_startup_enable(int command) { }
3910
/* Keep as macros so we do not need to define the commands */
3911 3912 3913 3914 3915
# define ftrace_startup(ops, command)			\
	({						\
		(ops)->flags |= FTRACE_OPS_FL_ENABLED;	\
		0;					\
	})
3916
# define ftrace_shutdown(ops, command)	do { } while (0)
I
Ingo Molnar 已提交
3917 3918
# define ftrace_startup_sysctl()	do { } while (0)
# define ftrace_shutdown_sysctl()	do { } while (0)
3919 3920 3921 3922 3923 3924 3925

static inline int
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip)
{
	return 1;
}

3926 3927
#endif /* CONFIG_DYNAMIC_FTRACE */

3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957
static void
ftrace_ops_control_func(unsigned long ip, unsigned long parent_ip)
{
	struct ftrace_ops *op;

	if (unlikely(trace_recursion_test(TRACE_CONTROL_BIT)))
		return;

	/*
	 * Some of the ops may be dynamically allocated,
	 * they must be freed after a synchronize_sched().
	 */
	preempt_disable_notrace();
	trace_recursion_set(TRACE_CONTROL_BIT);
	op = rcu_dereference_raw(ftrace_control_list);
	while (op != &ftrace_list_end) {
		if (!ftrace_function_local_disabled(op) &&
		    ftrace_ops_test(op, ip))
			op->func(ip, parent_ip);

		op = rcu_dereference_raw(op->next);
	};
	trace_recursion_clear(TRACE_CONTROL_BIT);
	preempt_enable_notrace();
}

static struct ftrace_ops control_ops = {
	.func = ftrace_ops_control_func,
};

3958 3959 3960
static void
ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip)
{
3961
	struct ftrace_ops *op;
3962

3963 3964 3965 3966
	if (unlikely(trace_recursion_test(TRACE_INTERNAL_BIT)))
		return;

	trace_recursion_set(TRACE_INTERNAL_BIT);
3967 3968 3969 3970 3971 3972
	/*
	 * Some of the ops may be dynamically allocated,
	 * they must be freed after a synchronize_sched().
	 */
	preempt_disable_notrace();
	op = rcu_dereference_raw(ftrace_ops_list);
3973 3974 3975 3976 3977
	while (op != &ftrace_list_end) {
		if (ftrace_ops_test(op, ip))
			op->func(ip, parent_ip);
		op = rcu_dereference_raw(op->next);
	};
3978
	preempt_enable_notrace();
3979
	trace_recursion_clear(TRACE_INTERNAL_BIT);
3980 3981
}

3982
static void clear_ftrace_swapper(void)
S
Steven Rostedt 已提交
3983 3984
{
	struct task_struct *p;
3985
	int cpu;
S
Steven Rostedt 已提交
3986

3987 3988 3989
	get_online_cpus();
	for_each_online_cpu(cpu) {
		p = idle_task(cpu);
S
Steven Rostedt 已提交
3990
		clear_tsk_trace_trace(p);
3991 3992 3993
	}
	put_online_cpus();
}
S
Steven Rostedt 已提交
3994

3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005
static void set_ftrace_swapper(void)
{
	struct task_struct *p;
	int cpu;

	get_online_cpus();
	for_each_online_cpu(cpu) {
		p = idle_task(cpu);
		set_tsk_trace_trace(p);
	}
	put_online_cpus();
S
Steven Rostedt 已提交
4006 4007
}

4008 4009 4010 4011
static void clear_ftrace_pid(struct pid *pid)
{
	struct task_struct *p;

4012
	rcu_read_lock();
4013 4014 4015
	do_each_pid_task(pid, PIDTYPE_PID, p) {
		clear_tsk_trace_trace(p);
	} while_each_pid_task(pid, PIDTYPE_PID, p);
4016 4017
	rcu_read_unlock();

4018 4019 4020 4021
	put_pid(pid);
}

static void set_ftrace_pid(struct pid *pid)
S
Steven Rostedt 已提交
4022 4023 4024
{
	struct task_struct *p;

4025
	rcu_read_lock();
S
Steven Rostedt 已提交
4026 4027 4028
	do_each_pid_task(pid, PIDTYPE_PID, p) {
		set_tsk_trace_trace(p);
	} while_each_pid_task(pid, PIDTYPE_PID, p);
4029
	rcu_read_unlock();
S
Steven Rostedt 已提交
4030 4031
}

4032
static void clear_ftrace_pid_task(struct pid *pid)
4033
{
4034
	if (pid == ftrace_swapper_pid)
4035 4036
		clear_ftrace_swapper();
	else
4037
		clear_ftrace_pid(pid);
4038 4039 4040 4041 4042 4043 4044 4045 4046 4047
}

static void set_ftrace_pid_task(struct pid *pid)
{
	if (pid == ftrace_swapper_pid)
		set_ftrace_swapper();
	else
		set_ftrace_pid(pid);
}

4048
static int ftrace_pid_add(int p)
4049
{
S
Steven Rostedt 已提交
4050
	struct pid *pid;
4051 4052
	struct ftrace_pid *fpid;
	int ret = -EINVAL;
4053

4054
	mutex_lock(&ftrace_lock);
4055

4056 4057 4058 4059
	if (!p)
		pid = ftrace_swapper_pid;
	else
		pid = find_get_pid(p);
4060

4061 4062
	if (!pid)
		goto out;
4063

4064
	ret = 0;
4065

4066 4067 4068
	list_for_each_entry(fpid, &ftrace_pids, list)
		if (fpid->pid == pid)
			goto out_put;
S
Steven Rostedt 已提交
4069

4070
	ret = -ENOMEM;
4071

4072 4073 4074
	fpid = kmalloc(sizeof(*fpid), GFP_KERNEL);
	if (!fpid)
		goto out_put;
4075

4076 4077
	list_add(&fpid->list, &ftrace_pids);
	fpid->pid = pid;
4078

4079
	set_ftrace_pid_task(pid);
S
Steven Rostedt 已提交
4080

4081 4082 4083 4084 4085 4086 4087 4088 4089
	ftrace_update_pid_func();
	ftrace_startup_enable(0);

	mutex_unlock(&ftrace_lock);
	return 0;

out_put:
	if (pid != ftrace_swapper_pid)
		put_pid(pid);
S
Steven Rostedt 已提交
4090

4091 4092 4093 4094 4095 4096 4097 4098
out:
	mutex_unlock(&ftrace_lock);
	return ret;
}

static void ftrace_pid_reset(void)
{
	struct ftrace_pid *fpid, *safe;
S
Steven Rostedt 已提交
4099

4100 4101 4102 4103 4104 4105 4106 4107
	mutex_lock(&ftrace_lock);
	list_for_each_entry_safe(fpid, safe, &ftrace_pids, list) {
		struct pid *pid = fpid->pid;

		clear_ftrace_pid_task(pid);

		list_del(&fpid->list);
		kfree(fpid);
4108 4109 4110 4111 4112
	}

	ftrace_update_pid_func();
	ftrace_startup_enable(0);

S
Steven Rostedt 已提交
4113
	mutex_unlock(&ftrace_lock);
4114
}
4115

4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152 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
static void *fpid_start(struct seq_file *m, loff_t *pos)
{
	mutex_lock(&ftrace_lock);

	if (list_empty(&ftrace_pids) && (!*pos))
		return (void *) 1;

	return seq_list_start(&ftrace_pids, *pos);
}

static void *fpid_next(struct seq_file *m, void *v, loff_t *pos)
{
	if (v == (void *)1)
		return NULL;

	return seq_list_next(v, &ftrace_pids, pos);
}

static void fpid_stop(struct seq_file *m, void *p)
{
	mutex_unlock(&ftrace_lock);
}

static int fpid_show(struct seq_file *m, void *v)
{
	const struct ftrace_pid *fpid = list_entry(v, struct ftrace_pid, list);

	if (v == (void *)1) {
		seq_printf(m, "no pid\n");
		return 0;
	}

	if (fpid->pid == ftrace_swapper_pid)
		seq_printf(m, "swapper tasks\n");
	else
		seq_printf(m, "%u\n", pid_vnr(fpid->pid));

	return 0;
}

static const struct seq_operations ftrace_pid_sops = {
	.start = fpid_start,
	.next = fpid_next,
	.stop = fpid_stop,
	.show = fpid_show,
};

static int
ftrace_pid_open(struct inode *inode, struct file *file)
{
	int ret = 0;

	if ((file->f_mode & FMODE_WRITE) &&
	    (file->f_flags & O_TRUNC))
		ftrace_pid_reset();

	if (file->f_mode & FMODE_READ)
		ret = seq_open(file, &ftrace_pid_sops);

	return ret;
}

4178 4179 4180 4181
static ssize_t
ftrace_pid_write(struct file *filp, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
4182
	char buf[64], *tmp;
4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193
	long val;
	int ret;

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

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

	buf[cnt] = 0;

4194 4195 4196 4197
	/*
	 * Allow "echo > set_ftrace_pid" or "echo -n '' > set_ftrace_pid"
	 * to clean the filter quietly.
	 */
4198 4199
	tmp = strstrip(buf);
	if (strlen(tmp) == 0)
4200 4201
		return 1;

4202
	ret = strict_strtol(tmp, 10, &val);
4203 4204 4205
	if (ret < 0)
		return ret;

4206
	ret = ftrace_pid_add(val);
4207

4208 4209
	return ret ? ret : cnt;
}
4210

4211 4212 4213 4214 4215
static int
ftrace_pid_release(struct inode *inode, struct file *file)
{
	if (file->f_mode & FMODE_READ)
		seq_release(inode, file);
4216

4217
	return 0;
4218 4219
}

4220
static const struct file_operations ftrace_pid_fops = {
4221 4222 4223 4224 4225
	.open		= ftrace_pid_open,
	.write		= ftrace_pid_write,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= ftrace_pid_release,
4226 4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237
};

static __init int ftrace_init_debugfs(void)
{
	struct dentry *d_tracer;

	d_tracer = tracing_init_dentry();
	if (!d_tracer)
		return 0;

	ftrace_init_dyn_debugfs(d_tracer);

4238 4239
	trace_create_file("set_ftrace_pid", 0644, d_tracer,
			    NULL, &ftrace_pid_fops);
4240 4241 4242

	ftrace_profile_debugfs(d_tracer);

4243 4244 4245 4246
	return 0;
}
fs_initcall(ftrace_init_debugfs);

S
Steven Rostedt 已提交
4247
/**
4248
 * ftrace_kill - kill ftrace
S
Steven Rostedt 已提交
4249 4250 4251 4252 4253
 *
 * This function should be used by panic code. It stops ftrace
 * but in a not so nice way. If you need to simply kill ftrace
 * from a non-atomic section, use ftrace_kill.
 */
4254
void ftrace_kill(void)
S
Steven Rostedt 已提交
4255 4256 4257 4258 4259 4260
{
	ftrace_disabled = 1;
	ftrace_enabled = 0;
	clear_ftrace_function();
}

4261 4262 4263 4264 4265 4266 4267 4268
/**
 * Test if ftrace is dead or not.
 */
int ftrace_is_dead(void)
{
	return ftrace_disabled;
}

4269
/**
4270 4271
 * register_ftrace_function - register a function for profiling
 * @ops - ops structure that holds the function for profiling.
4272
 *
4273 4274 4275 4276 4277 4278
 * Register a function to be called by all functions in the
 * kernel.
 *
 * Note: @ops->func and all the functions it calls must be labeled
 *       with "notrace", otherwise it will go into a
 *       recursive loop.
4279
 */
4280
int register_ftrace_function(struct ftrace_ops *ops)
4281
{
4282
	int ret = -1;
4283

S
Steven Rostedt 已提交
4284
	mutex_lock(&ftrace_lock);
4285

4286 4287 4288
	if (unlikely(ftrace_disabled))
		goto out_unlock;

4289
	ret = __register_ftrace_function(ops);
4290
	if (!ret)
4291
		ret = ftrace_startup(ops, 0);
4292

4293

4294
 out_unlock:
S
Steven Rostedt 已提交
4295
	mutex_unlock(&ftrace_lock);
4296
	return ret;
4297
}
4298
EXPORT_SYMBOL_GPL(register_ftrace_function);
4299 4300

/**
4301
 * unregister_ftrace_function - unregister a function for profiling.
4302 4303 4304 4305 4306 4307 4308 4309
 * @ops - ops structure that holds the function to unregister
 *
 * Unregister a function that was added to be called by ftrace profiling.
 */
int unregister_ftrace_function(struct ftrace_ops *ops)
{
	int ret;

S
Steven Rostedt 已提交
4310
	mutex_lock(&ftrace_lock);
4311
	ret = __unregister_ftrace_function(ops);
4312 4313
	if (!ret)
		ftrace_shutdown(ops, 0);
S
Steven Rostedt 已提交
4314
	mutex_unlock(&ftrace_lock);
4315 4316 4317

	return ret;
}
4318
EXPORT_SYMBOL_GPL(unregister_ftrace_function);
4319

I
Ingo Molnar 已提交
4320
int
4321
ftrace_enable_sysctl(struct ctl_table *table, int write,
4322
		     void __user *buffer, size_t *lenp,
4323 4324
		     loff_t *ppos)
{
4325
	int ret = -ENODEV;
4326

S
Steven Rostedt 已提交
4327
	mutex_lock(&ftrace_lock);
4328

4329 4330 4331 4332
	if (unlikely(ftrace_disabled))
		goto out;

	ret = proc_dointvec(table, write, buffer, lenp, ppos);
4333

4334
	if (ret || !write || (last_ftrace_enabled == !!ftrace_enabled))
4335 4336
		goto out;

4337
	last_ftrace_enabled = !!ftrace_enabled;
4338 4339 4340 4341 4342 4343

	if (ftrace_enabled) {

		ftrace_startup_sysctl();

		/* we are starting ftrace again */
4344 4345 4346
		if (ftrace_ops_list != &ftrace_list_end) {
			if (ftrace_ops_list->next == &ftrace_list_end)
				ftrace_trace_function = ftrace_ops_list->func;
4347
			else
4348
				ftrace_trace_function = ftrace_ops_list_func;
4349 4350 4351 4352 4353 4354 4355 4356 4357 4358
		}

	} else {
		/* stopping ftrace calls (just send to ftrace_stub) */
		ftrace_trace_function = ftrace_stub;

		ftrace_shutdown_sysctl();
	}

 out:
S
Steven Rostedt 已提交
4359
	mutex_unlock(&ftrace_lock);
4360
	return ret;
4361
}
4362

4363
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
4364

4365
static int ftrace_graph_active;
4366
static struct notifier_block ftrace_suspend_notifier;
4367

4368 4369 4370 4371 4372
int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace)
{
	return 0;
}

4373 4374 4375
/* The callbacks that hook a function */
trace_func_graph_ret_t ftrace_graph_return =
			(trace_func_graph_ret_t)ftrace_stub;
4376
trace_func_graph_ent_t ftrace_graph_entry = ftrace_graph_entry_stub;
4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406

/* Try to assign a return stack array on FTRACE_RETSTACK_ALLOC_SIZE tasks. */
static int alloc_retstack_tasklist(struct ftrace_ret_stack **ret_stack_list)
{
	int i;
	int ret = 0;
	unsigned long flags;
	int start = 0, end = FTRACE_RETSTACK_ALLOC_SIZE;
	struct task_struct *g, *t;

	for (i = 0; i < FTRACE_RETSTACK_ALLOC_SIZE; i++) {
		ret_stack_list[i] = kmalloc(FTRACE_RETFUNC_DEPTH
					* sizeof(struct ftrace_ret_stack),
					GFP_KERNEL);
		if (!ret_stack_list[i]) {
			start = 0;
			end = i;
			ret = -ENOMEM;
			goto free;
		}
	}

	read_lock_irqsave(&tasklist_lock, flags);
	do_each_thread(g, t) {
		if (start == end) {
			ret = -EAGAIN;
			goto unlock;
		}

		if (t->ret_stack == NULL) {
4407
			atomic_set(&t->tracing_graph_pause, 0);
4408
			atomic_set(&t->trace_overrun, 0);
4409 4410 4411 4412
			t->curr_ret_stack = -1;
			/* Make sure the tasks see the -1 first: */
			smp_wmb();
			t->ret_stack = ret_stack_list[start++];
4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423
		}
	} while_each_thread(g, t);

unlock:
	read_unlock_irqrestore(&tasklist_lock, flags);
free:
	for (i = start; i < end; i++)
		kfree(ret_stack_list[i]);
	return ret;
}

4424
static void
4425 4426
ftrace_graph_probe_sched_switch(void *ignore,
			struct task_struct *prev, struct task_struct *next)
4427 4428 4429 4430
{
	unsigned long long timestamp;
	int index;

4431 4432 4433 4434 4435 4436 4437
	/*
	 * Does the user want to count the time a function was asleep.
	 * If so, do not update the time stamps.
	 */
	if (trace_flags & TRACE_ITER_SLEEP_TIME)
		return;

4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454 4455
	timestamp = trace_clock_local();

	prev->ftrace_timestamp = timestamp;

	/* only process tasks that we timestamped */
	if (!next->ftrace_timestamp)
		return;

	/*
	 * Update all the counters in next to make up for the
	 * time next was sleeping.
	 */
	timestamp -= next->ftrace_timestamp;

	for (index = next->curr_ret_stack; index >= 0; index--)
		next->ret_stack[index].calltime += timestamp;
}

4456
/* Allocate a return stack for each task */
4457
static int start_graph_tracing(void)
4458 4459
{
	struct ftrace_ret_stack **ret_stack_list;
4460
	int ret, cpu;
4461 4462 4463 4464 4465 4466 4467 4468

	ret_stack_list = kmalloc(FTRACE_RETSTACK_ALLOC_SIZE *
				sizeof(struct ftrace_ret_stack *),
				GFP_KERNEL);

	if (!ret_stack_list)
		return -ENOMEM;

4469
	/* The cpu_boot init_task->ret_stack will never be freed */
4470 4471
	for_each_online_cpu(cpu) {
		if (!idle_task(cpu)->ret_stack)
4472
			ftrace_graph_init_idle_task(idle_task(cpu), cpu);
4473
	}
4474

4475 4476 4477 4478
	do {
		ret = alloc_retstack_tasklist(ret_stack_list);
	} while (ret == -EAGAIN);

4479
	if (!ret) {
4480
		ret = register_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
4481 4482 4483 4484 4485
		if (ret)
			pr_info("ftrace_graph: Couldn't activate tracepoint"
				" probe to kernel_sched_switch\n");
	}

4486 4487 4488 4489
	kfree(ret_stack_list);
	return ret;
}

4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510
/*
 * Hibernation protection.
 * The state of the current task is too much unstable during
 * suspend/restore to disk. We want to protect against that.
 */
static int
ftrace_suspend_notifier_call(struct notifier_block *bl, unsigned long state,
							void *unused)
{
	switch (state) {
	case PM_HIBERNATION_PREPARE:
		pause_graph_tracing();
		break;

	case PM_POST_HIBERNATION:
		unpause_graph_tracing();
		break;
	}
	return NOTIFY_DONE;
}

4511 4512
int register_ftrace_graph(trace_func_graph_ret_t retfunc,
			trace_func_graph_ent_t entryfunc)
4513
{
4514 4515
	int ret = 0;

S
Steven Rostedt 已提交
4516
	mutex_lock(&ftrace_lock);
4517

4518
	/* we currently allow only one tracer registered at a time */
4519
	if (ftrace_graph_active) {
4520 4521 4522 4523
		ret = -EBUSY;
		goto out;
	}

4524 4525 4526
	ftrace_suspend_notifier.notifier_call = ftrace_suspend_notifier_call;
	register_pm_notifier(&ftrace_suspend_notifier);

4527
	ftrace_graph_active++;
4528
	ret = start_graph_tracing();
4529
	if (ret) {
4530
		ftrace_graph_active--;
4531 4532
		goto out;
	}
4533

4534 4535
	ftrace_graph_return = retfunc;
	ftrace_graph_entry = entryfunc;
4536

4537
	ret = ftrace_startup(&global_ops, FTRACE_START_FUNC_RET);
4538 4539

out:
S
Steven Rostedt 已提交
4540
	mutex_unlock(&ftrace_lock);
4541
	return ret;
4542 4543
}

4544
void unregister_ftrace_graph(void)
4545
{
S
Steven Rostedt 已提交
4546
	mutex_lock(&ftrace_lock);
4547

4548
	if (unlikely(!ftrace_graph_active))
4549 4550
		goto out;

4551
	ftrace_graph_active--;
4552
	ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub;
4553
	ftrace_graph_entry = ftrace_graph_entry_stub;
4554
	ftrace_shutdown(&global_ops, FTRACE_STOP_FUNC_RET);
4555
	unregister_pm_notifier(&ftrace_suspend_notifier);
4556
	unregister_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
4557

4558
 out:
S
Steven Rostedt 已提交
4559
	mutex_unlock(&ftrace_lock);
4560
}
4561

4562 4563 4564 4565 4566 4567 4568 4569
static DEFINE_PER_CPU(struct ftrace_ret_stack *, idle_ret_stack);

static void
graph_init_task(struct task_struct *t, struct ftrace_ret_stack *ret_stack)
{
	atomic_set(&t->tracing_graph_pause, 0);
	atomic_set(&t->trace_overrun, 0);
	t->ftrace_timestamp = 0;
L
Lucas De Marchi 已提交
4570
	/* make curr_ret_stack visible before we add the ret_stack */
4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604
	smp_wmb();
	t->ret_stack = ret_stack;
}

/*
 * Allocate a return stack for the idle task. May be the first
 * time through, or it may be done by CPU hotplug online.
 */
void ftrace_graph_init_idle_task(struct task_struct *t, int cpu)
{
	t->curr_ret_stack = -1;
	/*
	 * The idle task has no parent, it either has its own
	 * stack or no stack at all.
	 */
	if (t->ret_stack)
		WARN_ON(t->ret_stack != per_cpu(idle_ret_stack, cpu));

	if (ftrace_graph_active) {
		struct ftrace_ret_stack *ret_stack;

		ret_stack = per_cpu(idle_ret_stack, cpu);
		if (!ret_stack) {
			ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH
					    * sizeof(struct ftrace_ret_stack),
					    GFP_KERNEL);
			if (!ret_stack)
				return;
			per_cpu(idle_ret_stack, cpu) = ret_stack;
		}
		graph_init_task(t, ret_stack);
	}
}

4605
/* Allocate a return stack for newly created task */
4606
void ftrace_graph_init_task(struct task_struct *t)
4607
{
4608 4609
	/* Make sure we do not use the parent ret_stack */
	t->ret_stack = NULL;
4610
	t->curr_ret_stack = -1;
4611

4612
	if (ftrace_graph_active) {
4613 4614 4615
		struct ftrace_ret_stack *ret_stack;

		ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH
4616 4617
				* sizeof(struct ftrace_ret_stack),
				GFP_KERNEL);
4618
		if (!ret_stack)
4619
			return;
4620
		graph_init_task(t, ret_stack);
4621
	}
4622 4623
}

4624
void ftrace_graph_exit_task(struct task_struct *t)
4625
{
4626 4627
	struct ftrace_ret_stack	*ret_stack = t->ret_stack;

4628
	t->ret_stack = NULL;
4629 4630 4631 4632
	/* NULL must become visible to IRQs before we free it: */
	barrier();

	kfree(ret_stack);
4633
}
S
Steven Rostedt 已提交
4634 4635 4636 4637 4638

void ftrace_graph_stop(void)
{
	ftrace_stop();
}
4639
#endif