ftrace.c 99.7 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
252 253 254 255
	ftrace_trace_function = ftrace_test_stop_func;
#endif
}

256
static void add_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
257
{
258
	ops->next = *list;
259
	/*
260
	 * We are entering ops into the list but another
261 262
	 * CPU might be walking that list. We need to make sure
	 * the ops->next pointer is valid before another CPU sees
263
	 * the ops pointer included into the list.
264
	 */
265
	rcu_assign_pointer(*list, ops);
266 267
}

268
static int remove_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
269 270 271 272
{
	struct ftrace_ops **p;

	/*
273 274
	 * If we are removing the last function, then simply point
	 * to the ftrace_stub.
275
	 */
276 277
	if (*list == ops && ops->next == &ftrace_list_end) {
		*list = &ftrace_list_end;
S
Steven Rostedt 已提交
278
		return 0;
279 280
	}

281
	for (p = list; *p != &ftrace_list_end; p = &(*p)->next)
282 283 284
		if (*p == ops)
			break;

S
Steven Rostedt 已提交
285 286
	if (*p != ops)
		return -1;
287 288

	*p = (*p)->next;
289 290
	return 0;
}
291

292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311
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;
}

312 313 314 315 316 317 318 319
static int __register_ftrace_function(struct ftrace_ops *ops)
{
	if (ftrace_disabled)
		return -ENODEV;

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

320 321 322
	if (WARN_ON(ops->flags & FTRACE_OPS_FL_ENABLED))
		return -EBUSY;

323 324 325 326
	/* We don't support both control and global flags set. */
	if ((ops->flags & FL_GLOBAL_CONTROL_MASK) == FL_GLOBAL_CONTROL_MASK)
		return -EINVAL;

327 328 329
	if (!core_kernel_data((unsigned long)ops))
		ops->flags |= FTRACE_OPS_FL_DYNAMIC;

330
	if (ops->flags & FTRACE_OPS_FL_GLOBAL) {
331
		add_ftrace_list_ops(&ftrace_global_list, &global_ops, ops);
332
		ops->flags |= FTRACE_OPS_FL_ENABLED;
333 334 335 336
	} 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);
337 338 339
	} else
		add_ftrace_ops(&ftrace_ops_list, ops);

340 341 342 343 344 345 346 347 348 349 350 351 352
	if (ftrace_enabled)
		update_ftrace_function();

	return 0;
}

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

	if (ftrace_disabled)
		return -ENODEV;

353 354 355
	if (WARN_ON(!(ops->flags & FTRACE_OPS_FL_ENABLED)))
		return -EBUSY;

356 357 358
	if (FTRACE_WARN_ON(ops == &global_ops))
		return -EINVAL;

359
	if (ops->flags & FTRACE_OPS_FL_GLOBAL) {
360 361
		ret = remove_ftrace_list_ops(&ftrace_global_list,
					     &global_ops, ops);
362 363
		if (!ret)
			ops->flags &= ~FTRACE_OPS_FL_ENABLED;
364 365 366 367 368 369 370 371 372 373 374 375 376
	} 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);
		}
377 378 379
	} else
		ret = remove_ftrace_ops(&ftrace_ops_list, ops);

380 381
	if (ret < 0)
		return ret;
382

383 384
	if (ftrace_enabled)
		update_ftrace_function();
385

386 387 388 389 390 391 392
	/*
	 * 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
393
	return 0;
394 395
}

396 397
static void ftrace_update_pid_func(void)
{
398
	/* Only do something if we are tracing something */
399
	if (ftrace_trace_function == ftrace_stub)
400
		return;
401

402
	update_ftrace_function();
403 404
}

405 406 407 408 409
#ifdef CONFIG_FUNCTION_PROFILER
struct ftrace_profile {
	struct hlist_node		node;
	unsigned long			ip;
	unsigned long			counter;
410 411
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	unsigned long long		time;
412
	unsigned long long		time_squared;
413
#endif
414 415
};

416 417 418 419
struct ftrace_profile_page {
	struct ftrace_profile_page	*next;
	unsigned long			index;
	struct ftrace_profile		records[];
420 421
};

422 423 424 425 426 427 428 429
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;
};

430 431
#define PROFILE_RECORDS_SIZE						\
	(PAGE_SIZE - offsetof(struct ftrace_profile_page, records))
432

433 434
#define PROFILES_PER_PAGE					\
	(PROFILE_RECORDS_SIZE / sizeof(struct ftrace_profile))
435

436 437 438 439
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
440 441
static DEFINE_MUTEX(ftrace_profile_lock);

442
static DEFINE_PER_CPU(struct ftrace_profile_stat, ftrace_profile_stats);
443 444 445

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

S
Steven Rostedt 已提交
446 447 448
static void *
function_stat_next(void *v, int idx)
{
449 450
	struct ftrace_profile *rec = v;
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
451

452
	pg = (struct ftrace_profile_page *)((unsigned long)rec & PAGE_MASK);
S
Steven Rostedt 已提交
453 454

 again:
L
Li Zefan 已提交
455 456 457
	if (idx != 0)
		rec++;

S
Steven Rostedt 已提交
458 459 460 461 462
	if ((void *)rec >= (void *)&pg->records[pg->index]) {
		pg = pg->next;
		if (!pg)
			return NULL;
		rec = &pg->records[0];
463 464
		if (!rec->counter)
			goto again;
S
Steven Rostedt 已提交
465 466 467 468 469 470 471
	}

	return rec;
}

static void *function_stat_start(struct tracer_stat *trace)
{
472 473 474 475 476 477 478
	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
479 480
}

481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496
#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
497 498
static int function_stat_cmp(void *p1, void *p2)
{
499 500
	struct ftrace_profile *a = p1;
	struct ftrace_profile *b = p2;
S
Steven Rostedt 已提交
501 502 503 504 505 506 507 508

	if (a->counter < b->counter)
		return -1;
	if (a->counter > b->counter)
		return 1;
	else
		return 0;
}
509
#endif
S
Steven Rostedt 已提交
510 511 512

static int function_stat_headers(struct seq_file *m)
{
513
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
514
	seq_printf(m, "  Function                               "
515
		   "Hit    Time            Avg             s^2\n"
516
		      "  --------                               "
517
		   "---    ----            ---             ---\n");
518
#else
S
Steven Rostedt 已提交
519 520
	seq_printf(m, "  Function                               Hit\n"
		      "  --------                               ---\n");
521
#endif
S
Steven Rostedt 已提交
522 523 524 525 526
	return 0;
}

static int function_stat_show(struct seq_file *m, void *v)
{
527
	struct ftrace_profile *rec = v;
S
Steven Rostedt 已提交
528
	char str[KSYM_SYMBOL_LEN];
529
	int ret = 0;
530
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
531 532
	static struct trace_seq s;
	unsigned long long avg;
533
	unsigned long long stddev;
534
#endif
535 536 537 538 539 540 541
	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
542 543

	kallsyms_lookup(rec->ip, NULL, NULL, NULL, str);
544 545 546 547
	seq_printf(m, "  %-30.30s  %10lu", str, rec->counter);

#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	seq_printf(m, "    ");
548 549 550
	avg = rec->time;
	do_div(avg, rec->counter);

551 552 553 554 555 556 557 558 559 560 561 562
	/* 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);
	}

563 564 565 566
	trace_seq_init(&s);
	trace_print_graph_duration(rec->time, &s);
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(avg, &s);
567 568
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(stddev, &s);
569 570 571
	trace_print_seq(m, &s);
#endif
	seq_putc(m, '\n');
572 573
out:
	mutex_unlock(&ftrace_profile_lock);
S
Steven Rostedt 已提交
574

575
	return ret;
S
Steven Rostedt 已提交
576 577
}

578
static void ftrace_profile_reset(struct ftrace_profile_stat *stat)
S
Steven Rostedt 已提交
579
{
580
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
581

582
	pg = stat->pages = stat->start;
S
Steven Rostedt 已提交
583

584 585 586 587
	while (pg) {
		memset(pg->records, 0, PROFILE_RECORDS_SIZE);
		pg->index = 0;
		pg = pg->next;
S
Steven Rostedt 已提交
588 589
	}

590
	memset(stat->hash, 0,
591 592
	       FTRACE_PROFILE_HASH_SIZE * sizeof(struct hlist_head));
}
S
Steven Rostedt 已提交
593

594
int ftrace_profile_pages_init(struct ftrace_profile_stat *stat)
595 596
{
	struct ftrace_profile_page *pg;
597 598
	int functions;
	int pages;
599
	int i;
S
Steven Rostedt 已提交
600

601
	/* If we already allocated, do nothing */
602
	if (stat->pages)
603
		return 0;
S
Steven Rostedt 已提交
604

605 606
	stat->pages = (void *)get_zeroed_page(GFP_KERNEL);
	if (!stat->pages)
607
		return -ENOMEM;
S
Steven Rostedt 已提交
608

609 610 611 612 613 614 615 616 617 618 619 620 621
#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

622
	pg = stat->start = stat->pages;
S
Steven Rostedt 已提交
623

624 625 626
	pages = DIV_ROUND_UP(functions, PROFILES_PER_PAGE);

	for (i = 0; i < pages; i++) {
627 628
		pg->next = (void *)get_zeroed_page(GFP_KERNEL);
		if (!pg->next)
629
			goto out_free;
630 631 632 633
		pg = pg->next;
	}

	return 0;
634 635 636 637 638 639 640 641 642 643 644 645 646 647 648

 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
649 650
}

651
static int ftrace_profile_init_cpu(int cpu)
S
Steven Rostedt 已提交
652
{
653
	struct ftrace_profile_stat *stat;
654
	int size;
S
Steven Rostedt 已提交
655

656 657 658
	stat = &per_cpu(ftrace_profile_stats, cpu);

	if (stat->hash) {
659
		/* If the profile is already created, simply reset it */
660
		ftrace_profile_reset(stat);
661 662
		return 0;
	}
S
Steven Rostedt 已提交
663

664 665 666 667 668
	/*
	 * 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
669

670
	stat->hash = kzalloc(sizeof(struct hlist_head) * size, GFP_KERNEL);
671

672
	if (!stat->hash)
673 674
		return -ENOMEM;

675 676
	if (!ftrace_profile_bits) {
		size--;
677

678 679 680
		for (; size; size >>= 1)
			ftrace_profile_bits++;
	}
681

682
	/* Preallocate the function profiling pages */
683 684 685
	if (ftrace_profile_pages_init(stat) < 0) {
		kfree(stat->hash);
		stat->hash = NULL;
686 687 688 689
		return -ENOMEM;
	}

	return 0;
S
Steven Rostedt 已提交
690 691
}

692 693 694 695 696 697 698 699 700 701 702 703 704 705
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;
}

706
/* interrupts must be disabled */
707 708
static struct ftrace_profile *
ftrace_find_profiled_func(struct ftrace_profile_stat *stat, unsigned long ip)
S
Steven Rostedt 已提交
709
{
710
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
711 712 713 714 715
	struct hlist_head *hhd;
	struct hlist_node *n;
	unsigned long key;

	key = hash_long(ip, ftrace_profile_bits);
716
	hhd = &stat->hash[key];
S
Steven Rostedt 已提交
717 718 719 720 721 722

	if (hlist_empty(hhd))
		return NULL;

	hlist_for_each_entry_rcu(rec, n, hhd, node) {
		if (rec->ip == ip)
723 724 725 726 727 728
			return rec;
	}

	return NULL;
}

729 730
static void ftrace_add_profile(struct ftrace_profile_stat *stat,
			       struct ftrace_profile *rec)
731 732 733 734
{
	unsigned long key;

	key = hash_long(rec->ip, ftrace_profile_bits);
735
	hlist_add_head_rcu(&rec->node, &stat->hash[key]);
736 737
}

738 739 740
/*
 * The memory is already allocated, this simply finds a new record to use.
 */
741
static struct ftrace_profile *
742
ftrace_profile_alloc(struct ftrace_profile_stat *stat, unsigned long ip)
743 744 745
{
	struct ftrace_profile *rec = NULL;

746
	/* prevent recursion (from NMIs) */
747
	if (atomic_inc_return(&stat->disabled) != 1)
748 749 750
		goto out;

	/*
751 752
	 * Try to find the function again since an NMI
	 * could have added it
753
	 */
754
	rec = ftrace_find_profiled_func(stat, ip);
755
	if (rec)
756
		goto out;
757

758 759 760 761
	if (stat->pages->index == PROFILES_PER_PAGE) {
		if (!stat->pages->next)
			goto out;
		stat->pages = stat->pages->next;
S
Steven Rostedt 已提交
762
	}
763

764
	rec = &stat->pages->records[stat->pages->index++];
765
	rec->ip = ip;
766
	ftrace_add_profile(stat, rec);
767

S
Steven Rostedt 已提交
768
 out:
769
	atomic_dec(&stat->disabled);
S
Steven Rostedt 已提交
770 771 772 773 774 775 776

	return rec;
}

static void
function_profile_call(unsigned long ip, unsigned long parent_ip)
{
777
	struct ftrace_profile_stat *stat;
778
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
779 780 781 782 783 784
	unsigned long flags;

	if (!ftrace_profile_enabled)
		return;

	local_irq_save(flags);
785 786

	stat = &__get_cpu_var(ftrace_profile_stats);
787
	if (!stat->hash || !ftrace_profile_enabled)
788 789 790
		goto out;

	rec = ftrace_find_profiled_func(stat, ip);
791
	if (!rec) {
792
		rec = ftrace_profile_alloc(stat, ip);
793 794 795
		if (!rec)
			goto out;
	}
S
Steven Rostedt 已提交
796 797 798 799 800 801

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

802 803 804 805 806 807 808 809 810
#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)
{
811
	struct ftrace_profile_stat *stat;
812
	unsigned long long calltime;
813
	struct ftrace_profile *rec;
814
	unsigned long flags;
815 816

	local_irq_save(flags);
817
	stat = &__get_cpu_var(ftrace_profile_stats);
818
	if (!stat->hash || !ftrace_profile_enabled)
819 820
		goto out;

821 822 823 824
	/* If the calltime was zero'd ignore it */
	if (!trace->calltime)
		goto out;

825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841
	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;
	}

842
	rec = ftrace_find_profiled_func(stat, trace->func);
843
	if (rec) {
844
		rec->time += calltime;
845 846
		rec->time_squared += calltime * calltime;
	}
847

848
 out:
849 850 851 852 853 854 855 856 857 858 859 860 861 862
	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
863
static struct ftrace_ops ftrace_profile_ops __read_mostly = {
864
	.func		= function_profile_call,
S
Steven Rostedt 已提交
865 866
};

867 868 869 870 871 872 873 874 875 876 877
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
878 879 880 881 882 883 884
static ssize_t
ftrace_profile_write(struct file *filp, const char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
	unsigned long val;
	int ret;

885 886
	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
S
Steven Rostedt 已提交
887 888 889 890 891 892 893
		return ret;

	val = !!val;

	mutex_lock(&ftrace_profile_lock);
	if (ftrace_profile_enabled ^ val) {
		if (val) {
894 895 896 897 898 899
			ret = ftrace_profile_init();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}

900 901 902 903 904
			ret = register_ftrace_profiler();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}
S
Steven Rostedt 已提交
905 906 907
			ftrace_profile_enabled = 1;
		} else {
			ftrace_profile_enabled = 0;
908 909 910 911
			/*
			 * unregister_ftrace_profiler calls stop_machine
			 * so this acts like an synchronize_sched.
			 */
912
			unregister_ftrace_profiler();
S
Steven Rostedt 已提交
913 914
		}
	}
915
 out:
S
Steven Rostedt 已提交
916 917
	mutex_unlock(&ftrace_profile_lock);

918
	*ppos += cnt;
S
Steven Rostedt 已提交
919 920 921 922

	return cnt;
}

923 924 925 926
static ssize_t
ftrace_profile_read(struct file *filp, char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
927
	char buf[64];		/* big enough to hold a number */
928 929 930 931 932 933
	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
934 935 936 937
static const struct file_operations ftrace_profile_fops = {
	.open		= tracing_open_generic,
	.read		= ftrace_profile_read,
	.write		= ftrace_profile_write,
938
	.llseek		= default_llseek,
S
Steven Rostedt 已提交
939 940
};

941 942
/* used to initialize the real stat files */
static struct tracer_stat function_stats __initdata = {
943 944 945 946 947 948
	.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
949 950
};

951
static __init void ftrace_profile_debugfs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
952
{
953
	struct ftrace_profile_stat *stat;
S
Steven Rostedt 已提交
954
	struct dentry *entry;
955
	char *name;
S
Steven Rostedt 已提交
956
	int ret;
957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984
	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
985 986 987 988 989 990 991 992 993 994
	}

	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 */
995
static __init void ftrace_profile_debugfs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
996 997 998 999
{
}
#endif /* CONFIG_FUNCTION_PROFILER */

1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018
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;
};

1019 1020 1021 1022 1023 1024 1025 1026 1027
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;
1028
	struct rcu_head		rcu;
1029 1030
};

1031 1032 1033 1034 1035 1036 1037 1038 1039
/*
 * 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,
1040
};
1041
#define EMPTY_HASH	((struct ftrace_hash *)&empty_hash)
1042

1043
static struct ftrace_ops global_ops = {
1044
	.func			= ftrace_stub,
1045 1046
	.notrace_hash		= EMPTY_HASH,
	.filter_hash		= EMPTY_HASH,
1047 1048
};

1049 1050 1051 1052
static DEFINE_MUTEX(ftrace_regex_lock);

struct ftrace_page {
	struct ftrace_page	*next;
1053
	struct dyn_ftrace	*records;
1054
	int			index;
1055
	int			size;
1056 1057
};

1058 1059
static struct ftrace_page *ftrace_new_pgs;

1060 1061
#define ENTRY_SIZE sizeof(struct dyn_ftrace)
#define ENTRIES_PER_PAGE (PAGE_SIZE / ENTRY_SIZE)
1062 1063 1064 1065 1066 1067 1068

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

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

1069 1070 1071 1072 1073
static bool ftrace_hash_empty(struct ftrace_hash *hash)
{
	return !hash || !hash->count;
}

1074 1075 1076 1077 1078 1079 1080 1081
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;

1082
	if (ftrace_hash_empty(hash))
1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098
		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;
}

1099 1100
static void __add_hash_entry(struct ftrace_hash *hash,
			     struct ftrace_func_entry *entry)
1101 1102 1103 1104 1105
{
	struct hlist_head *hhd;
	unsigned long key;

	if (hash->size_bits)
1106
		key = hash_long(entry->ip, hash->size_bits);
1107 1108 1109 1110 1111 1112
	else
		key = 0;

	hhd = &hash->buckets[key];
	hlist_add_head(&entry->hlist, hhd);
	hash->count++;
1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124
}

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);
1125 1126 1127 1128 1129

	return 0;
}

static void
1130
free_hash_entry(struct ftrace_hash *hash,
1131 1132 1133 1134 1135 1136 1137
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	kfree(entry);
	hash->count--;
}

1138 1139 1140 1141 1142 1143 1144 1145
static void
remove_hash_entry(struct ftrace_hash *hash,
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	hash->count--;
}

1146 1147 1148 1149 1150 1151 1152 1153
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;

1154 1155 1156
	if (!hash->count)
		return;

1157 1158 1159
	for (i = 0; i < size; i++) {
		hhd = &hash->buckets[i];
		hlist_for_each_entry_safe(entry, tp, tn, hhd, hlist)
1160
			free_hash_entry(hash, entry);
1161 1162 1163 1164
	}
	FTRACE_WARN_ON(hash->count);
}

1165 1166 1167 1168 1169 1170 1171 1172 1173
static void free_ftrace_hash(struct ftrace_hash *hash)
{
	if (!hash || hash == EMPTY_HASH)
		return;
	ftrace_hash_clear(hash);
	kfree(hash->buckets);
	kfree(hash);
}

1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188
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);
}

1189 1190 1191 1192 1193 1194 1195 1196 1197 1198
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;
1199
	hash->buckets = kcalloc(size, sizeof(*hash->buckets), GFP_KERNEL);
1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225

	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? */
1226
	if (ftrace_hash_empty(hash))
1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246
		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;
}

1247 1248 1249 1250 1251
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);

1252
static int
1253 1254
ftrace_hash_move(struct ftrace_ops *ops, int enable,
		 struct ftrace_hash **dst, struct ftrace_hash *src)
1255 1256 1257 1258
{
	struct ftrace_func_entry *entry;
	struct hlist_node *tp, *tn;
	struct hlist_head *hhd;
1259 1260
	struct ftrace_hash *old_hash;
	struct ftrace_hash *new_hash;
1261 1262 1263
	unsigned long key;
	int size = src->count;
	int bits = 0;
1264
	int ret;
1265 1266
	int i;

1267 1268 1269 1270 1271 1272
	/*
	 * Remove the current set, update the hash and add
	 * them back.
	 */
	ftrace_hash_rec_disable(ops, enable);

1273 1274 1275 1276 1277
	/*
	 * If the new source is empty, just free dst and assign it
	 * the empty_hash.
	 */
	if (!src->count) {
1278 1279
		free_ftrace_hash_rcu(*dst);
		rcu_assign_pointer(*dst, EMPTY_HASH);
1280 1281 1282
		/* still need to update the function records */
		ret = 0;
		goto out;
1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294
	}

	/*
	 * 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;

1295
	ret = -ENOMEM;
1296 1297
	new_hash = alloc_ftrace_hash(bits);
	if (!new_hash)
1298
		goto out;
1299 1300 1301 1302 1303 1304 1305 1306 1307 1308

	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);
1309
			__add_hash_entry(new_hash, entry);
1310 1311 1312
		}
	}

1313 1314 1315 1316
	old_hash = *dst;
	rcu_assign_pointer(*dst, new_hash);
	free_ftrace_hash_rcu(old_hash);

1317 1318 1319 1320 1321 1322 1323 1324 1325 1326
	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;
1327 1328
}

1329 1330 1331 1332 1333 1334 1335 1336
/*
 * 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.
1337 1338 1339
 *
 * This needs to be called with preemption disabled as
 * the hashes are freed with call_rcu_sched().
1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350
 */
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);

1351
	if ((ftrace_hash_empty(filter_hash) ||
1352
	     ftrace_lookup_ip(filter_hash, ip)) &&
1353
	    (ftrace_hash_empty(notrace_hash) ||
1354 1355 1356 1357 1358 1359 1360 1361
	     !ftrace_lookup_ip(notrace_hash, ip)))
		ret = 1;
	else
		ret = 0;

	return ret;
}

1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375
/*
 * 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()		\
		}				\
	}

1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388

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;
}

1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401
/**
 * 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;
1402
	struct dyn_ftrace key;
1403

1404 1405 1406 1407 1408 1409 1410
	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)
1411
			return 1;
1412
	}
1413 1414 1415 1416

	return 0;
}

1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445
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;
1446
		if (ftrace_hash_empty(hash))
1447 1448 1449 1450 1451 1452 1453 1454 1455
			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.
		 */
1456
		if (ftrace_hash_empty(hash))
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469
			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.
			 */
1470
			if (!other_hash || !ftrace_lookup_ip(other_hash, rec->ip))
1471 1472
				match = 1;
		} else {
1473 1474
			in_hash = !!ftrace_lookup_ip(hash, rec->ip);
			in_other_hash = !!ftrace_lookup_ip(other_hash, rec->ip);
1475 1476 1477 1478 1479 1480 1481

			/*
			 *
			 */
			if (filter_hash && in_hash && !in_other_hash)
				match = 1;
			else if (!filter_hash && in_hash &&
1482
				 (in_other_hash || ftrace_hash_empty(other_hash)))
1483 1484 1485 1486 1487 1488 1489 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
				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
1516
static struct dyn_ftrace *ftrace_alloc_dyn_node(unsigned long ip)
1517
{
1518 1519 1520 1521
	if (ftrace_pages->index == ftrace_pages->size) {
		/* We should have allocated enough */
		if (WARN_ON(!ftrace_pages->next))
			return NULL;
1522 1523 1524 1525 1526 1527
		ftrace_pages = ftrace_pages->next;
	}

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

S
Steven Rostedt 已提交
1528
static struct dyn_ftrace *
1529
ftrace_record_ip(unsigned long ip)
1530
{
S
Steven Rostedt 已提交
1531
	struct dyn_ftrace *rec;
1532

1533
	if (ftrace_disabled)
S
Steven Rostedt 已提交
1534
		return NULL;
1535

S
Steven Rostedt 已提交
1536 1537 1538
	rec = ftrace_alloc_dyn_node(ip);
	if (!rec)
		return NULL;
1539

S
Steven Rostedt 已提交
1540
	rec->ip = ip;
1541

S
Steven Rostedt 已提交
1542
	return rec;
1543 1544
}

1545 1546 1547 1548 1549 1550 1551 1552 1553 1554
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]);
}

1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567
/**
 * 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)
1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593
{
	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);
	}
}

1594

1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608
/* 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;
}

1609
static int ftrace_check_record(struct dyn_ftrace *rec, int enable, int update)
1610
{
1611
	unsigned long flag = 0UL;
1612

S
Steven Rostedt 已提交
1613
	/*
1614
	 * If we are updating calls:
S
Steven Rostedt 已提交
1615
	 *
1616 1617
	 *   If the record has a ref count, then we need to enable it
	 *   because someone is using it.
S
Steven Rostedt 已提交
1618
	 *
1619 1620
	 *   Otherwise we make sure its disabled.
	 *
1621
	 * If we are disabling calls, then disable all records that
1622
	 * are enabled.
S
Steven Rostedt 已提交
1623
	 */
1624
	if (enable && (rec->flags & ~FTRACE_FL_MASK))
1625
		flag = FTRACE_FL_ENABLED;
S
Steven Rostedt 已提交
1626

1627 1628
	/* If the state of this record hasn't changed, then do nothing */
	if ((rec->flags & FTRACE_FL_ENABLED) == flag)
1629
		return FTRACE_UPDATE_IGNORE;
S
Steven Rostedt 已提交
1630

1631
	if (flag) {
1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684
		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:
1685
		return ftrace_make_call(rec, ftrace_addr);
1686 1687 1688

	case FTRACE_UPDATE_MAKE_NOP:
		return ftrace_make_nop(NULL, rec, ftrace_addr);
1689 1690
	}

1691
	return -1; /* unknow ftrace bug */
1692 1693
}

1694
static void ftrace_replace_code(int update)
1695 1696 1697
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
S
Steven Rostedt 已提交
1698
	int failed;
1699

1700 1701 1702
	if (unlikely(ftrace_disabled))
		return;

1703
	do_for_each_ftrace_rec(pg, rec) {
1704
		failed = __ftrace_replace_code(rec, update);
1705
		if (failed) {
1706 1707 1708
			ftrace_bug(failed, rec->ip);
			/* Stop processing */
			return;
1709
		}
1710
	} while_for_each_ftrace_rec();
1711 1712
}

1713 1714 1715 1716 1717 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
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];
}

1785
static int
1786
ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec)
1787 1788
{
	unsigned long ip;
1789
	int ret;
1790 1791 1792

	ip = rec->ip;

1793 1794 1795
	if (unlikely(ftrace_disabled))
		return 0;

1796
	ret = ftrace_make_nop(mod, rec, MCOUNT_ADDR);
1797
	if (ret) {
1798
		ftrace_bug(ret, ip);
1799
		return 0;
1800
	}
1801
	return 1;
1802 1803
}

1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821
/*
 * 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
1822
static int __ftrace_modify_code(void *data)
1823
{
1824 1825
	int *command = data;

1826
	if (*command & FTRACE_UPDATE_CALLS)
1827
		ftrace_replace_code(1);
1828
	else if (*command & FTRACE_DISABLE_CALLS)
1829 1830 1831 1832 1833
		ftrace_replace_code(0);

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

1834 1835 1836 1837 1838
	if (*command & FTRACE_START_FUNC_RET)
		ftrace_enable_ftrace_graph_caller();
	else if (*command & FTRACE_STOP_FUNC_RET)
		ftrace_disable_ftrace_graph_caller();

1839
	return 0;
1840 1841
}

1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865
/**
 * 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
1866
static void ftrace_run_update_code(int command)
1867
{
1868 1869 1870 1871 1872 1873
	int ret;

	ret = ftrace_arch_code_modify_prepare();
	FTRACE_WARN_ON(ret);
	if (ret)
		return;
1874 1875 1876 1877 1878
	/*
	 * Do not call function tracer while we update the code.
	 * We are in stop machine.
	 */
	function_trace_stop++;
1879

1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898
	/*
	 * 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--;
1899 1900 1901

	ret = ftrace_arch_code_modify_post_process();
	FTRACE_WARN_ON(ret);
1902 1903
}

1904
static ftrace_func_t saved_ftrace_func;
1905
static int ftrace_start_up;
1906
static int global_start_up;
1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919

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);
}
1920

1921
static int ftrace_startup(struct ftrace_ops *ops, int command)
1922
{
1923 1924
	bool hash_enable = true;

1925
	if (unlikely(ftrace_disabled))
1926
		return -ENODEV;
1927

1928
	ftrace_start_up++;
1929
	command |= FTRACE_UPDATE_CALLS;
1930

1931 1932 1933 1934 1935 1936 1937 1938 1939
	/* 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++;
	}

1940
	ops->flags |= FTRACE_OPS_FL_ENABLED;
1941
	if (hash_enable)
1942 1943
		ftrace_hash_rec_enable(ops, 1);

1944
	ftrace_startup_enable(command);
1945 1946

	return 0;
1947 1948
}

1949
static void ftrace_shutdown(struct ftrace_ops *ops, int command)
1950
{
1951 1952
	bool hash_disable = true;

1953 1954 1955
	if (unlikely(ftrace_disabled))
		return;

1956
	ftrace_start_up--;
1957 1958 1959 1960 1961 1962 1963
	/*
	 * 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);

1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
	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)
1976 1977
		ftrace_hash_rec_disable(ops, 1);

1978
	if (ops != &global_ops || !global_start_up)
1979
		ops->flags &= ~FTRACE_OPS_FL_ENABLED;
1980

1981
	command |= FTRACE_UPDATE_CALLS;
1982

1983 1984 1985 1986
	if (saved_ftrace_func != ftrace_trace_function) {
		saved_ftrace_func = ftrace_trace_function;
		command |= FTRACE_UPDATE_TRACE_FUNC;
	}
1987

1988
	if (!command || !ftrace_enabled)
S
Steven Rostedt 已提交
1989
		return;
1990 1991

	ftrace_run_update_code(command);
1992 1993
}

I
Ingo Molnar 已提交
1994
static void ftrace_startup_sysctl(void)
1995
{
1996 1997 1998
	if (unlikely(ftrace_disabled))
		return;

1999 2000
	/* Force update next time */
	saved_ftrace_func = NULL;
2001 2002
	/* ftrace_start_up is true if we want ftrace running */
	if (ftrace_start_up)
2003
		ftrace_run_update_code(FTRACE_UPDATE_CALLS);
2004 2005
}

I
Ingo Molnar 已提交
2006
static void ftrace_shutdown_sysctl(void)
2007
{
2008 2009 2010
	if (unlikely(ftrace_disabled))
		return;

2011 2012
	/* ftrace_start_up is true if ftrace is running */
	if (ftrace_start_up)
2013
		ftrace_run_update_code(FTRACE_DISABLE_CALLS);
2014 2015
}

2016 2017 2018 2019
static cycle_t		ftrace_update_time;
static unsigned long	ftrace_update_cnt;
unsigned long		ftrace_update_tot_cnt;

2020 2021 2022 2023 2024
static int ops_traces_mod(struct ftrace_ops *ops)
{
	struct ftrace_hash *hash;

	hash = ops->filter_hash;
2025
	return ftrace_hash_empty(hash);
2026 2027
}

2028
static int ftrace_update_code(struct module *mod)
2029
{
2030
	struct ftrace_page *pg;
2031
	struct dyn_ftrace *p;
2032
	cycle_t start, stop;
2033
	unsigned long ref = 0;
2034
	int i;
2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051

	/*
	 * 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++;
		}
	}
2052

I
Ingo Molnar 已提交
2053
	start = ftrace_now(raw_smp_processor_id());
2054 2055
	ftrace_update_cnt = 0;

2056
	for (pg = ftrace_new_pgs; pg; pg = pg->next) {
2057

2058 2059 2060 2061
		for (i = 0; i < pg->index; i++) {
			/* If something went wrong, bail without enabling anything */
			if (unlikely(ftrace_disabled))
				return -1;
2062

2063 2064
			p = &pg->records[i];
			p->flags = ref;
2065

2066 2067 2068 2069 2070 2071
			/*
			 * Do the initial record conversion from mcount jump
			 * to the NOP instructions.
			 */
			if (!ftrace_code_disable(mod, p))
				break;
2072

2073
			ftrace_update_cnt++;
2074

2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088
			/*
			 * 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);
			}
2089
		}
2090 2091
	}

2092 2093
	ftrace_new_pgs = NULL;

I
Ingo Molnar 已提交
2094
	stop = ftrace_now(raw_smp_processor_id());
2095 2096 2097
	ftrace_update_time = stop - start;
	ftrace_update_tot_cnt += ftrace_update_cnt;

2098 2099 2100
	return 0;
}

2101
static int ftrace_allocate_records(struct ftrace_page *pg, int count)
2102
{
2103
	int order;
2104 2105
	int cnt;

2106 2107 2108 2109
	if (WARN_ON(!count))
		return -EINVAL;

	order = get_count_order(DIV_ROUND_UP(count, ENTRIES_PER_PAGE));
2110 2111

	/*
2112 2113
	 * We want to fill as much as possible. No more than a page
	 * may be empty.
2114
	 */
2115 2116
	while ((PAGE_SIZE << order) / ENTRY_SIZE >= count + ENTRIES_PER_PAGE)
		order--;
2117

2118 2119
 again:
	pg->records = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
2120

2121 2122 2123 2124 2125 2126 2127
	if (!pg->records) {
		/* if we can't allocate this size, try something smaller */
		if (!order)
			return -ENOMEM;
		order >>= 1;
		goto again;
	}
2128

2129 2130
	cnt = (PAGE_SIZE << order) / ENTRY_SIZE;
	pg->size = cnt;
2131

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

2167 2168 2169 2170
		pg->next = kzalloc(sizeof(*pg), GFP_KERNEL);
		if (!pg->next)
			goto free_pages;

2171 2172 2173
		pg = pg->next;
	}

2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200
	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);

2201 2202 2203
	return 0;
}

2204 2205 2206
#define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */

struct ftrace_iterator {
2207
	loff_t				pos;
2208 2209 2210 2211 2212
	loff_t				func_pos;
	struct ftrace_page		*pg;
	struct dyn_ftrace		*func;
	struct ftrace_func_probe	*probe;
	struct trace_parser		parser;
2213
	struct ftrace_hash		*hash;
2214
	struct ftrace_ops		*ops;
2215 2216 2217
	int				hidx;
	int				idx;
	unsigned			flags;
2218 2219
};

2220
static void *
2221
t_hash_next(struct seq_file *m, loff_t *pos)
2222 2223
{
	struct ftrace_iterator *iter = m->private;
2224
	struct hlist_node *hnd = NULL;
2225 2226 2227
	struct hlist_head *hhd;

	(*pos)++;
2228
	iter->pos = *pos;
2229

2230 2231
	if (iter->probe)
		hnd = &iter->probe->node;
2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253
 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;
		}
	}

2254 2255 2256 2257 2258 2259
	if (WARN_ON_ONCE(!hnd))
		return NULL;

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

	return iter;
2260 2261 2262 2263 2264 2265
}

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
2266 2267
	loff_t l;

2268 2269 2270
	if (!(iter->flags & FTRACE_ITER_DO_HASH))
		return NULL;

2271 2272
	if (iter->func_pos > *pos)
		return NULL;
2273

L
Li Zefan 已提交
2274
	iter->hidx = 0;
2275
	for (l = 0; l <= (*pos - iter->func_pos); ) {
2276
		p = t_hash_next(m, &l);
L
Li Zefan 已提交
2277 2278 2279
		if (!p)
			break;
	}
2280 2281 2282
	if (!p)
		return NULL;

2283 2284 2285
	/* Only set this if we have an item */
	iter->flags |= FTRACE_ITER_HASH;

2286
	return iter;
2287 2288
}

2289 2290
static int
t_hash_show(struct seq_file *m, struct ftrace_iterator *iter)
2291
{
S
Steven Rostedt 已提交
2292
	struct ftrace_func_probe *rec;
2293

2294 2295 2296
	rec = iter->probe;
	if (WARN_ON_ONCE(!rec))
		return -EIO;
2297

2298 2299 2300
	if (rec->ops->print)
		return rec->ops->print(m, rec->ip, rec->ops, rec->data);

2301
	seq_printf(m, "%ps:%ps", (void *)rec->ip, (void *)rec->ops->func);
2302 2303 2304 2305 2306 2307 2308 2309

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

	return 0;
}

I
Ingo Molnar 已提交
2310
static void *
2311 2312 2313
t_next(struct seq_file *m, void *v, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
2314
	struct ftrace_ops *ops = iter->ops;
2315 2316
	struct dyn_ftrace *rec = NULL;

2317 2318 2319
	if (unlikely(ftrace_disabled))
		return NULL;

2320
	if (iter->flags & FTRACE_ITER_HASH)
2321
		return t_hash_next(m, pos);
2322

2323
	(*pos)++;
2324
	iter->pos = iter->func_pos = *pos;
2325

2326
	if (iter->flags & FTRACE_ITER_PRINTALL)
2327
		return t_hash_start(m, pos);
2328

2329 2330 2331 2332 2333 2334 2335 2336 2337
 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++];
2338
		if (((iter->flags & FTRACE_ITER_FILTER) &&
2339
		     !(ftrace_lookup_ip(ops->filter_hash, rec->ip))) ||
S
Steven Rostedt 已提交
2340

2341
		    ((iter->flags & FTRACE_ITER_NOTRACE) &&
2342 2343 2344 2345 2346
		     !ftrace_lookup_ip(ops->notrace_hash, rec->ip)) ||

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

2347 2348 2349 2350 2351
			rec = NULL;
			goto retry;
		}
	}

2352
	if (!rec)
2353
		return t_hash_start(m, pos);
2354 2355 2356 2357

	iter->func = rec;

	return iter;
2358 2359
}

2360 2361 2362 2363 2364
static void reset_iter_read(struct ftrace_iterator *iter)
{
	iter->pos = 0;
	iter->func_pos = 0;
	iter->flags &= ~(FTRACE_ITER_PRINTALL & FTRACE_ITER_HASH);
2365 2366 2367 2368 2369
}

static void *t_start(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
2370
	struct ftrace_ops *ops = iter->ops;
2371
	void *p = NULL;
2372
	loff_t l;
2373

2374
	mutex_lock(&ftrace_lock);
2375 2376 2377 2378

	if (unlikely(ftrace_disabled))
		return NULL;

2379 2380 2381 2382 2383 2384
	/*
	 * If an lseek was done, then reset and start from beginning.
	 */
	if (*pos < iter->pos)
		reset_iter_read(iter);

2385 2386 2387 2388 2389
	/*
	 * For set_ftrace_filter reading, if we have the filter
	 * off, we can short cut and just print out that all
	 * functions are enabled.
	 */
2390 2391
	if (iter->flags & FTRACE_ITER_FILTER &&
	    ftrace_hash_empty(ops->filter_hash)) {
2392
		if (*pos > 0)
2393
			return t_hash_start(m, pos);
2394
		iter->flags |= FTRACE_ITER_PRINTALL;
2395 2396
		/* reset in case of seek/pread */
		iter->flags &= ~FTRACE_ITER_HASH;
2397 2398 2399
		return iter;
	}

2400 2401 2402
	if (iter->flags & FTRACE_ITER_HASH)
		return t_hash_start(m, pos);

2403 2404 2405 2406 2407
	/*
	 * 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.
	 */
2408 2409 2410 2411 2412 2413
	iter->pg = ftrace_pages_start;
	iter->idx = 0;
	for (l = 0; l <= *pos; ) {
		p = t_next(m, p, &l);
		if (!p)
			break;
2414
	}
2415

2416 2417
	if (!p)
		return t_hash_start(m, pos);
2418 2419

	return iter;
2420 2421 2422 2423
}

static void t_stop(struct seq_file *m, void *p)
{
2424
	mutex_unlock(&ftrace_lock);
2425 2426 2427 2428
}

static int t_show(struct seq_file *m, void *v)
{
2429
	struct ftrace_iterator *iter = m->private;
2430
	struct dyn_ftrace *rec;
2431

2432
	if (iter->flags & FTRACE_ITER_HASH)
2433
		return t_hash_show(m, iter);
2434

2435 2436 2437 2438 2439
	if (iter->flags & FTRACE_ITER_PRINTALL) {
		seq_printf(m, "#### all functions enabled ####\n");
		return 0;
	}

2440 2441
	rec = iter->func;

2442 2443 2444
	if (!rec)
		return 0;

2445 2446 2447 2448 2449
	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");
2450 2451 2452 2453

	return 0;
}

J
James Morris 已提交
2454
static const struct seq_operations show_ftrace_seq_ops = {
2455 2456 2457 2458 2459 2460
	.start = t_start,
	.next = t_next,
	.stop = t_stop,
	.show = t_show,
};

I
Ingo Molnar 已提交
2461
static int
2462 2463 2464 2465 2466
ftrace_avail_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;
	int ret;

2467 2468 2469
	if (unlikely(ftrace_disabled))
		return -ENODEV;

2470 2471 2472 2473 2474
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter)
		return -ENOMEM;

	iter->pg = ftrace_pages_start;
2475
	iter->ops = &global_ops;
2476 2477 2478 2479

	ret = seq_open(file, &show_ftrace_seq_ops);
	if (!ret) {
		struct seq_file *m = file->private_data;
I
Ingo Molnar 已提交
2480

2481
		m->private = iter;
I
Ingo Molnar 已提交
2482
	} else {
2483
		kfree(iter);
I
Ingo Molnar 已提交
2484
	}
2485 2486 2487 2488

	return ret;
}

2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503
static int
ftrace_enabled_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;
	int ret;

	if (unlikely(ftrace_disabled))
		return -ENODEV;

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

	iter->pg = ftrace_pages_start;
	iter->flags = FTRACE_ITER_ENABLED;
2504
	iter->ops = &global_ops;
2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517

	ret = seq_open(file, &show_ftrace_seq_ops);
	if (!ret) {
		struct seq_file *m = file->private_data;

		m->private = iter;
	} else {
		kfree(iter);
	}

	return ret;
}

2518
static void ftrace_filter_reset(struct ftrace_hash *hash)
2519
{
2520
	mutex_lock(&ftrace_lock);
2521
	ftrace_hash_clear(hash);
2522
	mutex_unlock(&ftrace_lock);
2523 2524
}

2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541
/**
 * 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
2542
ftrace_regex_open(struct ftrace_ops *ops, int flag,
2543
		  struct inode *inode, struct file *file)
2544 2545
{
	struct ftrace_iterator *iter;
2546
	struct ftrace_hash *hash;
2547 2548
	int ret = 0;

2549 2550 2551
	if (unlikely(ftrace_disabled))
		return -ENODEV;

2552 2553 2554 2555
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter)
		return -ENOMEM;

2556 2557 2558 2559 2560
	if (trace_parser_get_init(&iter->parser, FTRACE_BUFF_MAX)) {
		kfree(iter);
		return -ENOMEM;
	}

2561 2562 2563 2564 2565
	if (flag & FTRACE_ITER_NOTRACE)
		hash = ops->notrace_hash;
	else
		hash = ops->filter_hash;

2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579
	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;
		}
	}
2580

2581
	mutex_lock(&ftrace_regex_lock);
2582

2583
	if ((file->f_mode & FMODE_WRITE) &&
2584
	    (file->f_flags & O_TRUNC))
2585
		ftrace_filter_reset(iter->hash);
2586 2587 2588 2589 2590 2591 2592 2593

	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;
2594
		} else {
2595 2596
			/* Failed */
			free_ftrace_hash(iter->hash);
2597
			trace_parser_put(&iter->parser);
2598
			kfree(iter);
2599
		}
2600 2601
	} else
		file->private_data = iter;
2602
	mutex_unlock(&ftrace_regex_lock);
2603 2604 2605 2606

	return ret;
}

2607 2608 2609
static int
ftrace_filter_open(struct inode *inode, struct file *file)
{
2610 2611 2612
	return ftrace_regex_open(&global_ops,
			FTRACE_ITER_FILTER | FTRACE_ITER_DO_HASH,
			inode, file);
2613 2614 2615 2616 2617
}

static int
ftrace_notrace_open(struct inode *inode, struct file *file)
{
2618
	return ftrace_regex_open(&global_ops, FTRACE_ITER_NOTRACE,
2619
				 inode, file);
2620 2621
}

2622
loff_t
2623
ftrace_regex_lseek(struct file *file, loff_t offset, int origin)
2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634
{
	loff_t ret;

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

	return ret;
}

2635
static int ftrace_match(char *str, char *regex, int len, int type)
2636 2637
{
	int matched = 0;
2638
	int slen;
2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653

	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:
2654 2655
		slen = strlen(str);
		if (slen >= len && memcmp(str + slen - len, regex, len) == 0)
2656 2657 2658 2659 2660 2661 2662
			matched = 1;
		break;
	}

	return matched;
}

2663
static int
2664
enter_record(struct ftrace_hash *hash, struct dyn_ftrace *rec, int not)
2665
{
2666 2667 2668
	struct ftrace_func_entry *entry;
	int ret = 0;

2669 2670 2671 2672 2673
	entry = ftrace_lookup_ip(hash, rec->ip);
	if (not) {
		/* Do nothing if it doesn't exist */
		if (!entry)
			return 0;
2674

2675
		free_hash_entry(hash, entry);
2676 2677 2678 2679
	} else {
		/* Do nothing if it exists */
		if (entry)
			return 0;
2680

2681
		ret = add_hash_entry(hash, rec->ip);
2682 2683
	}
	return ret;
2684 2685
}

2686
static int
2687 2688
ftrace_match_record(struct dyn_ftrace *rec, char *mod,
		    char *regex, int len, int type)
2689 2690
{
	char str[KSYM_SYMBOL_LEN];
2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703
	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;
	}
2704 2705 2706 2707

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

2708 2709 2710
static int
match_records(struct ftrace_hash *hash, char *buff,
	      int len, char *mod, int not)
2711
{
2712
	unsigned search_len = 0;
2713 2714
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
2715 2716
	int type = MATCH_FULL;
	char *search = buff;
2717
	int found = 0;
2718
	int ret;
2719

2720 2721 2722 2723
	if (len) {
		type = filter_parse_regex(buff, len, &search, &not);
		search_len = strlen(search);
	}
2724

2725
	mutex_lock(&ftrace_lock);
2726

2727 2728
	if (unlikely(ftrace_disabled))
		goto out_unlock;
2729

2730
	do_for_each_ftrace_rec(pg, rec) {
2731
		if (ftrace_match_record(rec, mod, search, search_len, type)) {
2732
			ret = enter_record(hash, rec, not);
2733 2734 2735 2736
			if (ret < 0) {
				found = ret;
				goto out_unlock;
			}
2737
			found = 1;
2738 2739
		}
	} while_for_each_ftrace_rec();
2740
 out_unlock:
2741
	mutex_unlock(&ftrace_lock);
2742 2743

	return found;
2744 2745
}

2746
static int
2747
ftrace_match_records(struct ftrace_hash *hash, char *buff, int len)
2748
{
2749
	return match_records(hash, buff, len, NULL, 0);
2750 2751
}

2752 2753
static int
ftrace_match_module_records(struct ftrace_hash *hash, char *buff, char *mod)
2754 2755
{
	int not = 0;
S
Steven Rostedt 已提交
2756

2757 2758 2759 2760 2761 2762 2763 2764 2765 2766
	/* 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;
	}

2767
	return match_records(hash, buff, strlen(buff), mod, not);
2768 2769
}

2770 2771 2772 2773 2774 2775
/*
 * We register the module command as a template to show others how
 * to register the a command as well.
 */

static int
2776 2777
ftrace_mod_callback(struct ftrace_hash *hash,
		    char *func, char *cmd, char *param, int enable)
2778 2779
{
	char *mod;
2780
	int ret = -EINVAL;
2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791

	/*
	 * 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)
2792
		return ret;
2793 2794 2795

	mod = strsep(&param, ":");
	if (!strlen(mod))
2796
		return ret;
2797

2798
	ret = ftrace_match_module_records(hash, func, mod);
2799 2800 2801 2802 2803 2804
	if (!ret)
		ret = -EINVAL;
	if (ret < 0)
		return ret;

	return 0;
2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817
}

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);

2818
static void
S
Steven Rostedt 已提交
2819
function_trace_probe_call(unsigned long ip, unsigned long parent_ip)
2820
{
S
Steven Rostedt 已提交
2821
	struct ftrace_func_probe *entry;
2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837
	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.
	 */
2838
	preempt_disable_notrace();
2839 2840 2841 2842
	hlist_for_each_entry_rcu(entry, n, hhd, node) {
		if (entry->ip == ip)
			entry->ops->func(ip, parent_ip, &entry->data);
	}
2843
	preempt_enable_notrace();
2844 2845
}

S
Steven Rostedt 已提交
2846
static struct ftrace_ops trace_probe_ops __read_mostly =
2847
{
2848
	.func		= function_trace_probe_call,
2849 2850
};

S
Steven Rostedt 已提交
2851
static int ftrace_probe_registered;
2852

S
Steven Rostedt 已提交
2853
static void __enable_ftrace_function_probe(void)
2854
{
2855
	int ret;
2856 2857
	int i;

S
Steven Rostedt 已提交
2858
	if (ftrace_probe_registered)
2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869
		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;

2870 2871
	ret = __register_ftrace_function(&trace_probe_ops);
	if (!ret)
2872
		ret = ftrace_startup(&trace_probe_ops, 0);
2873

S
Steven Rostedt 已提交
2874
	ftrace_probe_registered = 1;
2875 2876
}

S
Steven Rostedt 已提交
2877
static void __disable_ftrace_function_probe(void)
2878
{
2879
	int ret;
2880 2881
	int i;

S
Steven Rostedt 已提交
2882
	if (!ftrace_probe_registered)
2883 2884 2885 2886 2887 2888 2889 2890 2891
		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 */
2892 2893 2894 2895
	ret = __unregister_ftrace_function(&trace_probe_ops);
	if (!ret)
		ftrace_shutdown(&trace_probe_ops, 0);

S
Steven Rostedt 已提交
2896
	ftrace_probe_registered = 0;
2897 2898 2899 2900 2901
}


static void ftrace_free_entry_rcu(struct rcu_head *rhp)
{
S
Steven Rostedt 已提交
2902 2903
	struct ftrace_func_probe *entry =
		container_of(rhp, struct ftrace_func_probe, rcu);
2904 2905 2906 2907 2908 2909 2910 2911

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


int
S
Steven Rostedt 已提交
2912
register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
2913 2914
			      void *data)
{
S
Steven Rostedt 已提交
2915
	struct ftrace_func_probe *entry;
2916 2917 2918
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
	int type, len, not;
S
Steven Rostedt 已提交
2919
	unsigned long key;
2920 2921 2922
	int count = 0;
	char *search;

2923
	type = filter_parse_regex(glob, strlen(glob), &search, &not);
2924 2925
	len = strlen(search);

S
Steven Rostedt 已提交
2926
	/* we do not support '!' for function probes */
2927 2928 2929 2930 2931
	if (WARN_ON(not))
		return -EINVAL;

	mutex_lock(&ftrace_lock);

2932 2933
	if (unlikely(ftrace_disabled))
		goto out_unlock;
2934

2935
	do_for_each_ftrace_rec(pg, rec) {
2936

2937
		if (!ftrace_match_record(rec, NULL, search, len, type))
2938 2939 2940 2941
			continue;

		entry = kmalloc(sizeof(*entry), GFP_KERNEL);
		if (!entry) {
S
Steven Rostedt 已提交
2942
			/* If we did not process any, then return error */
2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971
			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
2972
	__enable_ftrace_function_probe();
2973 2974 2975 2976 2977 2978 2979 2980

 out_unlock:
	mutex_unlock(&ftrace_lock);

	return count;
}

enum {
S
Steven Rostedt 已提交
2981 2982
	PROBE_TEST_FUNC		= 1,
	PROBE_TEST_DATA		= 2
2983 2984 2985
};

static void
S
Steven Rostedt 已提交
2986
__unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
2987 2988
				  void *data, int flags)
{
S
Steven Rostedt 已提交
2989
	struct ftrace_func_probe *entry;
2990 2991 2992 2993 2994 2995
	struct hlist_node *n, *tmp;
	char str[KSYM_SYMBOL_LEN];
	int type = MATCH_FULL;
	int i, len = 0;
	char *search;

2996
	if (glob && (strcmp(glob, "*") == 0 || !strlen(glob)))
2997
		glob = NULL;
2998
	else if (glob) {
2999 3000
		int not;

3001
		type = filter_parse_regex(glob, strlen(glob), &search, &not);
3002 3003
		len = strlen(search);

S
Steven Rostedt 已提交
3004
		/* we do not support '!' for function probes */
3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015
		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
3016
			if ((flags & PROBE_TEST_FUNC) && entry->ops != ops)
3017 3018
				continue;

S
Steven Rostedt 已提交
3019
			if ((flags & PROBE_TEST_DATA) && entry->data != data)
3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033
				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
3034
	__disable_ftrace_function_probe();
3035 3036 3037 3038
	mutex_unlock(&ftrace_lock);
}

void
S
Steven Rostedt 已提交
3039
unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3040 3041
				void *data)
{
S
Steven Rostedt 已提交
3042 3043
	__unregister_ftrace_function_probe(glob, ops, data,
					  PROBE_TEST_FUNC | PROBE_TEST_DATA);
3044 3045 3046
}

void
S
Steven Rostedt 已提交
3047
unregister_ftrace_function_probe_func(char *glob, struct ftrace_probe_ops *ops)
3048
{
S
Steven Rostedt 已提交
3049
	__unregister_ftrace_function_probe(glob, ops, NULL, PROBE_TEST_FUNC);
3050 3051
}

S
Steven Rostedt 已提交
3052
void unregister_ftrace_function_probe_all(char *glob)
3053
{
S
Steven Rostedt 已提交
3054
	__unregister_ftrace_function_probe(glob, NULL, NULL, 0);
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 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097
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;
}

3098 3099
static int ftrace_process_regex(struct ftrace_hash *hash,
				char *buff, int len, int enable)
3100
{
3101
	char *func, *command, *next = buff;
S
Steven Rostedt 已提交
3102
	struct ftrace_func_command *p;
3103
	int ret = -EINVAL;
3104 3105 3106 3107

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

	if (!next) {
3108
		ret = ftrace_match_records(hash, func, len);
3109 3110 3111 3112 3113
		if (!ret)
			ret = -EINVAL;
		if (ret < 0)
			return ret;
		return 0;
3114 3115
	}

3116
	/* command found */
3117 3118 3119

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

3120 3121 3122
	mutex_lock(&ftrace_cmd_mutex);
	list_for_each_entry(p, &ftrace_commands, list) {
		if (strcmp(p->name, command) == 0) {
3123
			ret = p->func(hash, func, command, next, enable);
3124 3125
			goto out_unlock;
		}
3126
	}
3127 3128
 out_unlock:
	mutex_unlock(&ftrace_cmd_mutex);
3129

3130
	return ret;
3131 3132
}

I
Ingo Molnar 已提交
3133
static ssize_t
3134 3135
ftrace_regex_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos, int enable)
3136 3137
{
	struct ftrace_iterator *iter;
3138 3139
	struct trace_parser *parser;
	ssize_t ret, read;
3140

3141
	if (!cnt)
3142 3143
		return 0;

3144
	mutex_lock(&ftrace_regex_lock);
3145

3146 3147 3148 3149
	ret = -ENODEV;
	if (unlikely(ftrace_disabled))
		goto out_unlock;

3150 3151 3152 3153 3154 3155
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;
		iter = m->private;
	} else
		iter = file->private_data;

3156 3157
	parser = &iter->parser;
	read = trace_get_user(parser, ubuf, cnt, ppos);
3158

3159
	if (read >= 0 && trace_parser_loaded(parser) &&
3160
	    !trace_parser_cont(parser)) {
3161
		ret = ftrace_process_regex(iter->hash, parser->buffer,
3162
					   parser->idx, enable);
3163
		trace_parser_clear(parser);
3164
		if (ret)
3165
			goto out_unlock;
3166
	}
3167 3168

	ret = read;
3169
out_unlock:
3170
	mutex_unlock(&ftrace_regex_lock);
3171

3172 3173 3174
	return ret;
}

3175
ssize_t
3176 3177 3178 3179 3180 3181
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);
}

3182
ssize_t
3183 3184 3185 3186 3187 3188
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);
}

3189
static int
3190 3191
ftrace_set_regex(struct ftrace_ops *ops, unsigned char *buf, int len,
		 int reset, int enable)
3192
{
3193
	struct ftrace_hash **orig_hash;
3194
	struct ftrace_hash *hash;
3195
	int ret;
3196

3197 3198 3199 3200
	/* All global ops uses the global ops filters */
	if (ops->flags & FTRACE_OPS_FL_GLOBAL)
		ops = &global_ops;

3201
	if (unlikely(ftrace_disabled))
3202
		return -ENODEV;
3203

3204
	if (enable)
3205
		orig_hash = &ops->filter_hash;
3206
	else
3207 3208 3209 3210 3211
		orig_hash = &ops->notrace_hash;

	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);
	if (!hash)
		return -ENOMEM;
3212

3213 3214
	mutex_lock(&ftrace_regex_lock);
	if (reset)
3215
		ftrace_filter_reset(hash);
3216 3217 3218 3219
	if (buf && !ftrace_match_records(hash, buf, len)) {
		ret = -EINVAL;
		goto out_regex_unlock;
	}
3220 3221

	mutex_lock(&ftrace_lock);
3222
	ret = ftrace_hash_move(ops, enable, orig_hash, hash);
3223 3224
	if (!ret && ops->flags & FTRACE_OPS_FL_ENABLED
	    && ftrace_enabled)
3225
		ftrace_run_update_code(FTRACE_UPDATE_CALLS);
3226

3227 3228
	mutex_unlock(&ftrace_lock);

3229
 out_regex_unlock:
3230
	mutex_unlock(&ftrace_regex_lock);
3231 3232 3233

	free_ftrace_hash(hash);
	return ret;
3234 3235
}

3236 3237
/**
 * ftrace_set_filter - set a function to filter on in ftrace
3238 3239 3240 3241 3242 3243 3244 3245
 * @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.
 */
3246
int ftrace_set_filter(struct ftrace_ops *ops, unsigned char *buf,
3247 3248
		       int len, int reset)
{
3249
	return ftrace_set_regex(ops, buf, len, reset, 1);
3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263
}
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.
 */
3264
int ftrace_set_notrace(struct ftrace_ops *ops, unsigned char *buf,
3265 3266
			int len, int reset)
{
3267
	return ftrace_set_regex(ops, buf, len, reset, 0);
3268 3269 3270 3271 3272
}
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
3273 3274 3275 3276 3277 3278 3279
 * @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.
 */
3280
void ftrace_set_global_filter(unsigned char *buf, int len, int reset)
3281
{
3282
	ftrace_set_regex(&global_ops, buf, len, reset, 1);
3283
}
3284
EXPORT_SYMBOL_GPL(ftrace_set_global_filter);
3285

3286 3287
/**
 * ftrace_set_notrace - set a function to not trace in ftrace
3288
 * @ops - the ops to set the notrace filter with
3289 3290 3291 3292 3293 3294 3295 3296
 * @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.
 */
3297
void ftrace_set_global_notrace(unsigned char *buf, int len, int reset)
3298
{
3299
	ftrace_set_regex(&global_ops, buf, len, reset, 0);
3300
}
3301
EXPORT_SYMBOL_GPL(ftrace_set_global_notrace);
3302

3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323
/*
 * 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);

3324
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3325
static char ftrace_graph_buf[FTRACE_FILTER_SIZE] __initdata;
3326 3327
static int ftrace_set_func(unsigned long *array, int *idx, char *buffer);

3328 3329
static int __init set_graph_function(char *str)
{
3330
	strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE);
3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351
	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 */

3352 3353
void __init
ftrace_set_early_filter(struct ftrace_ops *ops, char *buf, int enable)
3354 3355 3356 3357 3358
{
	char *func;

	while (buf) {
		func = strsep(&buf, ",");
3359
		ftrace_set_regex(ops, func, strlen(func), 0, enable);
3360 3361 3362 3363 3364 3365
	}
}

static void __init set_ftrace_early_filters(void)
{
	if (ftrace_filter_buf[0])
3366
		ftrace_set_early_filter(&global_ops, ftrace_filter_buf, 1);
3367
	if (ftrace_notrace_buf[0])
3368
		ftrace_set_early_filter(&global_ops, ftrace_notrace_buf, 0);
3369 3370 3371 3372
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	if (ftrace_graph_buf[0])
		set_ftrace_early_graph(ftrace_graph_buf);
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
3373 3374
}

3375
int ftrace_regex_release(struct inode *inode, struct file *file)
3376 3377 3378
{
	struct seq_file *m = (struct seq_file *)file->private_data;
	struct ftrace_iterator *iter;
3379
	struct ftrace_hash **orig_hash;
3380
	struct trace_parser *parser;
3381
	int filter_hash;
3382
	int ret;
3383

3384
	mutex_lock(&ftrace_regex_lock);
3385 3386 3387 3388 3389 3390 3391
	if (file->f_mode & FMODE_READ) {
		iter = m->private;

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

3392 3393 3394
	parser = &iter->parser;
	if (trace_parser_loaded(parser)) {
		parser->buffer[parser->idx] = 0;
3395
		ftrace_match_records(iter->hash, parser->buffer, parser->idx);
3396 3397
	}

3398 3399
	trace_parser_put(parser);

3400
	if (file->f_mode & FMODE_WRITE) {
3401 3402 3403
		filter_hash = !!(iter->flags & FTRACE_ITER_FILTER);

		if (filter_hash)
3404
			orig_hash = &iter->ops->filter_hash;
3405 3406
		else
			orig_hash = &iter->ops->notrace_hash;
3407

3408
		mutex_lock(&ftrace_lock);
3409 3410 3411 3412
		ret = ftrace_hash_move(iter->ops, filter_hash,
				       orig_hash, iter->hash);
		if (!ret && (iter->ops->flags & FTRACE_OPS_FL_ENABLED)
		    && ftrace_enabled)
3413
			ftrace_run_update_code(FTRACE_UPDATE_CALLS);
3414

3415 3416
		mutex_unlock(&ftrace_lock);
	}
3417 3418
	free_ftrace_hash(iter->hash);
	kfree(iter);
3419

3420
	mutex_unlock(&ftrace_regex_lock);
3421 3422 3423
	return 0;
}

3424
static const struct file_operations ftrace_avail_fops = {
3425 3426 3427
	.open = ftrace_avail_open,
	.read = seq_read,
	.llseek = seq_lseek,
L
Li Zefan 已提交
3428
	.release = seq_release_private,
3429 3430
};

3431 3432 3433 3434 3435 3436 3437
static const struct file_operations ftrace_enabled_fops = {
	.open = ftrace_enabled_open,
	.read = seq_read,
	.llseek = seq_lseek,
	.release = seq_release_private,
};

3438
static const struct file_operations ftrace_filter_fops = {
3439
	.open = ftrace_filter_open,
L
Lai Jiangshan 已提交
3440
	.read = seq_read,
3441
	.write = ftrace_filter_write,
3442
	.llseek = ftrace_regex_lseek,
3443
	.release = ftrace_regex_release,
3444 3445
};

3446
static const struct file_operations ftrace_notrace_fops = {
3447
	.open = ftrace_notrace_open,
L
Lai Jiangshan 已提交
3448
	.read = seq_read,
3449 3450
	.write = ftrace_notrace_write,
	.llseek = ftrace_regex_lseek,
3451
	.release = ftrace_regex_release,
3452 3453
};

3454 3455 3456 3457 3458
#ifdef CONFIG_FUNCTION_GRAPH_TRACER

static DEFINE_MUTEX(graph_lock);

int ftrace_graph_count;
3459
int ftrace_graph_filter_enabled;
3460 3461 3462
unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly;

static void *
3463
__g_next(struct seq_file *m, loff_t *pos)
3464
{
3465
	if (*pos >= ftrace_graph_count)
3466
		return NULL;
3467
	return &ftrace_graph_funcs[*pos];
3468
}
3469

3470 3471 3472 3473 3474
static void *
g_next(struct seq_file *m, void *v, loff_t *pos)
{
	(*pos)++;
	return __g_next(m, pos);
3475 3476 3477 3478 3479 3480
}

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

3481
	/* Nothing, tell g_show to print all functions are enabled */
3482
	if (!ftrace_graph_filter_enabled && !*pos)
3483 3484
		return (void *)1;

3485
	return __g_next(m, pos);
3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499
}

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;

3500 3501 3502 3503 3504
	if (ptr == (unsigned long *)1) {
		seq_printf(m, "#### all functions enabled ####\n");
		return 0;
	}

3505
	seq_printf(m, "%ps\n", (void *)*ptr);
3506 3507 3508 3509

	return 0;
}

J
James Morris 已提交
3510
static const struct seq_operations ftrace_graph_seq_ops = {
3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526
	.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) &&
3527
	    (file->f_flags & O_TRUNC)) {
3528
		ftrace_graph_filter_enabled = 0;
3529 3530 3531
		ftrace_graph_count = 0;
		memset(ftrace_graph_funcs, 0, sizeof(ftrace_graph_funcs));
	}
3532
	mutex_unlock(&graph_lock);
3533

3534
	if (file->f_mode & FMODE_READ)
3535 3536 3537 3538 3539
		ret = seq_open(file, &ftrace_graph_seq_ops);

	return ret;
}

3540 3541 3542 3543 3544 3545 3546 3547
static int
ftrace_graph_release(struct inode *inode, struct file *file)
{
	if (file->f_mode & FMODE_READ)
		seq_release(inode, file);
	return 0;
}

3548
static int
3549
ftrace_set_func(unsigned long *array, int *idx, char *buffer)
3550 3551 3552
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
3553
	int search_len;
3554
	int fail = 1;
3555 3556 3557 3558
	int type, not;
	char *search;
	bool exists;
	int i;
3559

3560
	/* decode regex */
3561
	type = filter_parse_regex(buffer, strlen(buffer), &search, &not);
3562 3563
	if (!not && *idx >= FTRACE_GRAPH_MAX_FUNCS)
		return -EBUSY;
3564 3565 3566

	search_len = strlen(search);

3567
	mutex_lock(&ftrace_lock);
3568 3569 3570 3571 3572 3573

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

3574 3575
	do_for_each_ftrace_rec(pg, rec) {

3576
		if (ftrace_match_record(rec, NULL, search, search_len, type)) {
3577
			/* if it is in the array */
3578
			exists = false;
3579
			for (i = 0; i < *idx; i++) {
3580 3581
				if (array[i] == rec->ip) {
					exists = true;
3582 3583
					break;
				}
3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599
			}

			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;
				}
			}
3600
		}
3601
	} while_for_each_ftrace_rec();
3602
out:
3603
	mutex_unlock(&ftrace_lock);
3604

3605 3606 3607 3608 3609
	if (fail)
		return -EINVAL;

	ftrace_graph_filter_enabled = 1;
	return 0;
3610 3611 3612 3613 3614 3615
}

static ssize_t
ftrace_graph_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
3616
	struct trace_parser parser;
3617
	ssize_t read, ret;
3618

3619
	if (!cnt)
3620 3621 3622 3623
		return 0;

	mutex_lock(&graph_lock);

3624 3625
	if (trace_parser_get_init(&parser, FTRACE_BUFF_MAX)) {
		ret = -ENOMEM;
3626
		goto out_unlock;
3627 3628
	}

3629
	read = trace_get_user(&parser, ubuf, cnt, ppos);
3630

3631
	if (read >= 0 && trace_parser_loaded((&parser))) {
3632 3633 3634
		parser.buffer[parser.idx] = 0;

		/* we allow only one expression at a time */
3635
		ret = ftrace_set_func(ftrace_graph_funcs, &ftrace_graph_count,
3636
					parser.buffer);
3637
		if (ret)
3638
			goto out_free;
3639 3640 3641
	}

	ret = read;
3642 3643

out_free:
3644
	trace_parser_put(&parser);
3645
out_unlock:
3646 3647 3648 3649 3650 3651
	mutex_unlock(&graph_lock);

	return ret;
}

static const struct file_operations ftrace_graph_fops = {
3652 3653 3654 3655
	.open		= ftrace_graph_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
	.release	= ftrace_graph_release,
3656
	.llseek		= seq_lseek,
3657 3658 3659
};
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

3660
static __init int ftrace_init_dyn_debugfs(struct dentry *d_tracer)
3661 3662
{

3663 3664
	trace_create_file("available_filter_functions", 0444,
			d_tracer, NULL, &ftrace_avail_fops);
3665

3666 3667 3668
	trace_create_file("enabled_functions", 0444,
			d_tracer, NULL, &ftrace_enabled_fops);

3669 3670
	trace_create_file("set_ftrace_filter", 0644, d_tracer,
			NULL, &ftrace_filter_fops);
3671

3672
	trace_create_file("set_ftrace_notrace", 0644, d_tracer,
3673
				    NULL, &ftrace_notrace_fops);
3674

3675
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3676
	trace_create_file("set_graph_function", 0444, d_tracer,
3677 3678 3679 3680
				    NULL,
				    &ftrace_graph_fops);
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

3681 3682 3683
	return 0;
}

3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694
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;
}

3695
static int ftrace_process_locs(struct module *mod,
3696
			       unsigned long *start,
3697 3698
			       unsigned long *end)
{
3699 3700
	struct ftrace_page *pg;
	unsigned long count;
3701 3702
	unsigned long *p;
	unsigned long addr;
3703
	unsigned long flags = 0; /* Shut up gcc */
3704 3705 3706 3707 3708 3709 3710 3711 3712 3713
	int ret = -ENOMEM;

	count = end - start;

	if (!count)
		return 0;

	pg = ftrace_allocate_pages(count);
	if (!pg)
		return -ENOMEM;
3714

S
Steven Rostedt 已提交
3715
	mutex_lock(&ftrace_lock);
3716

3717 3718 3719 3720 3721
	/*
	 * 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.
	 */
3722 3723 3724 3725 3726
	if (!mod) {
		WARN_ON(ftrace_pages || ftrace_pages_start);
		/* First initialization */
		ftrace_pages = ftrace_pages_start = pg;
	} else {
3727
		if (!ftrace_pages)
3728
			goto out;
3729

3730 3731 3732 3733
		if (WARN_ON(ftrace_pages->next)) {
			/* Hmm, we have free pages? */
			while (ftrace_pages->next)
				ftrace_pages = ftrace_pages->next;
3734
		}
3735 3736 3737

		ftrace_pages->next = pg;
		ftrace_pages = pg;
3738 3739
	}

3740 3741 3742
	p = start;
	while (p < end) {
		addr = ftrace_call_adjust(*p++);
3743 3744 3745 3746 3747 3748 3749 3750
		/*
		 * 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;
3751 3752
		if (!ftrace_record_ip(addr))
			break;
3753 3754
	}

3755 3756 3757
	/* These new locations need to be initialized */
	ftrace_new_pgs = pg;

3758 3759 3760 3761 3762
	/* 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);

3763
	/*
3764 3765 3766 3767 3768 3769
	 * 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.
3770
	 */
3771 3772
	if (!mod)
		local_irq_save(flags);
3773
	ftrace_update_code(mod);
3774 3775
	if (!mod)
		local_irq_restore(flags);
3776 3777
	ret = 0;
 out:
S
Steven Rostedt 已提交
3778
	mutex_unlock(&ftrace_lock);
3779

3780
	return ret;
3781 3782
}

3783
#ifdef CONFIG_MODULES
3784 3785 3786

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

3787
void ftrace_release_mod(struct module *mod)
3788 3789
{
	struct dyn_ftrace *rec;
3790
	struct ftrace_page **last_pg;
3791
	struct ftrace_page *pg;
3792
	int order;
3793

3794 3795
	mutex_lock(&ftrace_lock);

3796
	if (ftrace_disabled)
3797
		goto out_unlock;
3798

3799 3800 3801 3802 3803 3804 3805
	/*
	 * 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];
3806
		if (within_module_core(rec->ip, mod)) {
3807
			/*
3808 3809
			 * As core pages are first, the first
			 * page should never be a module page.
3810
			 */
3811 3812 3813 3814 3815 3816 3817 3818
			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;
3819 3820 3821
			order = get_count_order(pg->size / ENTRIES_PER_PAGE);
			free_pages((unsigned long)pg->records, order);
			kfree(pg);
3822 3823 3824
		} else
			last_pg = &pg->next;
	}
3825
 out_unlock:
3826 3827 3828 3829 3830
	mutex_unlock(&ftrace_lock);
}

static void ftrace_init_module(struct module *mod,
			       unsigned long *start, unsigned long *end)
3831
{
3832
	if (ftrace_disabled || start == end)
3833
		return;
3834
	ftrace_process_locs(mod, start, end);
3835 3836
}

3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848
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:
3849
		ftrace_release_mod(mod);
3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867
		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,
};

3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894
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;

3895
	ret = ftrace_process_locs(NULL,
3896
				  __start_mcount_loc,
3897 3898
				  __stop_mcount_loc);

3899
	ret = register_module_notifier(&ftrace_module_nb);
3900
	if (ret)
3901 3902
		pr_warning("Failed to register trace ftrace module notifier\n");

3903 3904
	set_ftrace_early_filters();

3905 3906 3907 3908 3909
	return;
 failed:
	ftrace_disabled = 1;
}

3910
#else
3911

3912
static struct ftrace_ops global_ops = {
3913 3914 3915
	.func			= ftrace_stub,
};

3916 3917 3918 3919 3920 3921 3922
static int __init ftrace_nodyn_init(void)
{
	ftrace_enabled = 1;
	return 0;
}
device_initcall(ftrace_nodyn_init);

3923 3924
static inline int ftrace_init_dyn_debugfs(struct dentry *d_tracer) { return 0; }
static inline void ftrace_startup_enable(int command) { }
3925
/* Keep as macros so we do not need to define the commands */
3926 3927 3928 3929 3930
# define ftrace_startup(ops, command)			\
	({						\
		(ops)->flags |= FTRACE_OPS_FL_ENABLED;	\
		0;					\
	})
3931
# define ftrace_shutdown(ops, command)	do { } while (0)
I
Ingo Molnar 已提交
3932 3933
# define ftrace_startup_sysctl()	do { } while (0)
# define ftrace_shutdown_sysctl()	do { } while (0)
3934 3935 3936 3937 3938 3939 3940

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

3941 3942
#endif /* CONFIG_DYNAMIC_FTRACE */

3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972
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,
};

3973 3974 3975
static void
ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip)
{
3976
	struct ftrace_ops *op;
3977

3978 3979 3980 3981
	if (unlikely(trace_recursion_test(TRACE_INTERNAL_BIT)))
		return;

	trace_recursion_set(TRACE_INTERNAL_BIT);
3982 3983 3984 3985 3986 3987
	/*
	 * 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);
3988 3989 3990 3991 3992
	while (op != &ftrace_list_end) {
		if (ftrace_ops_test(op, ip))
			op->func(ip, parent_ip);
		op = rcu_dereference_raw(op->next);
	};
3993
	preempt_enable_notrace();
3994
	trace_recursion_clear(TRACE_INTERNAL_BIT);
3995 3996
}

3997
static void clear_ftrace_swapper(void)
S
Steven Rostedt 已提交
3998 3999
{
	struct task_struct *p;
4000
	int cpu;
S
Steven Rostedt 已提交
4001

4002 4003 4004
	get_online_cpus();
	for_each_online_cpu(cpu) {
		p = idle_task(cpu);
S
Steven Rostedt 已提交
4005
		clear_tsk_trace_trace(p);
4006 4007 4008
	}
	put_online_cpus();
}
S
Steven Rostedt 已提交
4009

4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020
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
4021 4022
}

4023 4024 4025 4026
static void clear_ftrace_pid(struct pid *pid)
{
	struct task_struct *p;

4027
	rcu_read_lock();
4028 4029 4030
	do_each_pid_task(pid, PIDTYPE_PID, p) {
		clear_tsk_trace_trace(p);
	} while_each_pid_task(pid, PIDTYPE_PID, p);
4031 4032
	rcu_read_unlock();

4033 4034 4035 4036
	put_pid(pid);
}

static void set_ftrace_pid(struct pid *pid)
S
Steven Rostedt 已提交
4037 4038 4039
{
	struct task_struct *p;

4040
	rcu_read_lock();
S
Steven Rostedt 已提交
4041 4042 4043
	do_each_pid_task(pid, PIDTYPE_PID, p) {
		set_tsk_trace_trace(p);
	} while_each_pid_task(pid, PIDTYPE_PID, p);
4044
	rcu_read_unlock();
S
Steven Rostedt 已提交
4045 4046
}

4047
static void clear_ftrace_pid_task(struct pid *pid)
4048
{
4049
	if (pid == ftrace_swapper_pid)
4050 4051
		clear_ftrace_swapper();
	else
4052
		clear_ftrace_pid(pid);
4053 4054 4055 4056 4057 4058 4059 4060 4061 4062
}

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

4063
static int ftrace_pid_add(int p)
4064
{
S
Steven Rostedt 已提交
4065
	struct pid *pid;
4066 4067
	struct ftrace_pid *fpid;
	int ret = -EINVAL;
4068

4069
	mutex_lock(&ftrace_lock);
4070

4071 4072 4073 4074
	if (!p)
		pid = ftrace_swapper_pid;
	else
		pid = find_get_pid(p);
4075

4076 4077
	if (!pid)
		goto out;
4078

4079
	ret = 0;
4080

4081 4082 4083
	list_for_each_entry(fpid, &ftrace_pids, list)
		if (fpid->pid == pid)
			goto out_put;
S
Steven Rostedt 已提交
4084

4085
	ret = -ENOMEM;
4086

4087 4088 4089
	fpid = kmalloc(sizeof(*fpid), GFP_KERNEL);
	if (!fpid)
		goto out_put;
4090

4091 4092
	list_add(&fpid->list, &ftrace_pids);
	fpid->pid = pid;
4093

4094
	set_ftrace_pid_task(pid);
S
Steven Rostedt 已提交
4095

4096 4097 4098 4099 4100 4101 4102 4103 4104
	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
4105

4106 4107 4108 4109 4110 4111 4112 4113
out:
	mutex_unlock(&ftrace_lock);
	return ret;
}

static void ftrace_pid_reset(void)
{
	struct ftrace_pid *fpid, *safe;
S
Steven Rostedt 已提交
4114

4115 4116 4117 4118 4119 4120 4121 4122
	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);
4123 4124 4125 4126 4127
	}

	ftrace_update_pid_func();
	ftrace_startup_enable(0);

S
Steven Rostedt 已提交
4128
	mutex_unlock(&ftrace_lock);
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 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192
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;
}

4193 4194 4195 4196
static ssize_t
ftrace_pid_write(struct file *filp, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
4197
	char buf[64], *tmp;
4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208
	long val;
	int ret;

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

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

	buf[cnt] = 0;

4209 4210 4211 4212
	/*
	 * Allow "echo > set_ftrace_pid" or "echo -n '' > set_ftrace_pid"
	 * to clean the filter quietly.
	 */
4213 4214
	tmp = strstrip(buf);
	if (strlen(tmp) == 0)
4215 4216
		return 1;

4217
	ret = strict_strtol(tmp, 10, &val);
4218 4219 4220
	if (ret < 0)
		return ret;

4221
	ret = ftrace_pid_add(val);
4222

4223 4224
	return ret ? ret : cnt;
}
4225

4226 4227 4228 4229 4230
static int
ftrace_pid_release(struct inode *inode, struct file *file)
{
	if (file->f_mode & FMODE_READ)
		seq_release(inode, file);
4231

4232
	return 0;
4233 4234
}

4235
static const struct file_operations ftrace_pid_fops = {
4236 4237 4238 4239 4240
	.open		= ftrace_pid_open,
	.write		= ftrace_pid_write,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= ftrace_pid_release,
4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252
};

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);

4253 4254
	trace_create_file("set_ftrace_pid", 0644, d_tracer,
			    NULL, &ftrace_pid_fops);
4255 4256 4257

	ftrace_profile_debugfs(d_tracer);

4258 4259 4260 4261
	return 0;
}
fs_initcall(ftrace_init_debugfs);

S
Steven Rostedt 已提交
4262
/**
4263
 * ftrace_kill - kill ftrace
S
Steven Rostedt 已提交
4264 4265 4266 4267 4268
 *
 * 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.
 */
4269
void ftrace_kill(void)
S
Steven Rostedt 已提交
4270 4271 4272 4273 4274 4275
{
	ftrace_disabled = 1;
	ftrace_enabled = 0;
	clear_ftrace_function();
}

4276 4277 4278 4279 4280 4281 4282 4283
/**
 * Test if ftrace is dead or not.
 */
int ftrace_is_dead(void)
{
	return ftrace_disabled;
}

4284
/**
4285 4286
 * register_ftrace_function - register a function for profiling
 * @ops - ops structure that holds the function for profiling.
4287
 *
4288 4289 4290 4291 4292 4293
 * 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.
4294
 */
4295
int register_ftrace_function(struct ftrace_ops *ops)
4296
{
4297
	int ret = -1;
4298

S
Steven Rostedt 已提交
4299
	mutex_lock(&ftrace_lock);
4300

4301 4302 4303
	if (unlikely(ftrace_disabled))
		goto out_unlock;

4304
	ret = __register_ftrace_function(ops);
4305
	if (!ret)
4306
		ret = ftrace_startup(ops, 0);
4307

4308

4309
 out_unlock:
S
Steven Rostedt 已提交
4310
	mutex_unlock(&ftrace_lock);
4311
	return ret;
4312
}
4313
EXPORT_SYMBOL_GPL(register_ftrace_function);
4314 4315

/**
4316
 * unregister_ftrace_function - unregister a function for profiling.
4317 4318 4319 4320 4321 4322 4323 4324
 * @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
4325
	mutex_lock(&ftrace_lock);
4326
	ret = __unregister_ftrace_function(ops);
4327 4328
	if (!ret)
		ftrace_shutdown(ops, 0);
S
Steven Rostedt 已提交
4329
	mutex_unlock(&ftrace_lock);
4330 4331 4332

	return ret;
}
4333
EXPORT_SYMBOL_GPL(unregister_ftrace_function);
4334

I
Ingo Molnar 已提交
4335
int
4336
ftrace_enable_sysctl(struct ctl_table *table, int write,
4337
		     void __user *buffer, size_t *lenp,
4338 4339
		     loff_t *ppos)
{
4340
	int ret = -ENODEV;
4341

S
Steven Rostedt 已提交
4342
	mutex_lock(&ftrace_lock);
4343

4344 4345 4346 4347
	if (unlikely(ftrace_disabled))
		goto out;

	ret = proc_dointvec(table, write, buffer, lenp, ppos);
4348

4349
	if (ret || !write || (last_ftrace_enabled == !!ftrace_enabled))
4350 4351
		goto out;

4352
	last_ftrace_enabled = !!ftrace_enabled;
4353 4354 4355 4356 4357 4358

	if (ftrace_enabled) {

		ftrace_startup_sysctl();

		/* we are starting ftrace again */
4359 4360 4361
		if (ftrace_ops_list != &ftrace_list_end) {
			if (ftrace_ops_list->next == &ftrace_list_end)
				ftrace_trace_function = ftrace_ops_list->func;
4362
			else
4363
				ftrace_trace_function = ftrace_ops_list_func;
4364 4365 4366 4367 4368 4369 4370 4371 4372 4373
		}

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

		ftrace_shutdown_sysctl();
	}

 out:
S
Steven Rostedt 已提交
4374
	mutex_unlock(&ftrace_lock);
4375
	return ret;
4376
}
4377

4378
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
4379

4380
static int ftrace_graph_active;
4381
static struct notifier_block ftrace_suspend_notifier;
4382

4383 4384 4385 4386 4387
int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace)
{
	return 0;
}

4388 4389 4390
/* The callbacks that hook a function */
trace_func_graph_ret_t ftrace_graph_return =
			(trace_func_graph_ret_t)ftrace_stub;
4391
trace_func_graph_ent_t ftrace_graph_entry = ftrace_graph_entry_stub;
4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421

/* 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) {
4422
			atomic_set(&t->tracing_graph_pause, 0);
4423
			atomic_set(&t->trace_overrun, 0);
4424 4425 4426 4427
			t->curr_ret_stack = -1;
			/* Make sure the tasks see the -1 first: */
			smp_wmb();
			t->ret_stack = ret_stack_list[start++];
4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438
		}
	} 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;
}

4439
static void
4440 4441
ftrace_graph_probe_sched_switch(void *ignore,
			struct task_struct *prev, struct task_struct *next)
4442 4443 4444 4445
{
	unsigned long long timestamp;
	int index;

4446 4447 4448 4449 4450 4451 4452
	/*
	 * 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;

4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470
	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;
}

4471
/* Allocate a return stack for each task */
4472
static int start_graph_tracing(void)
4473 4474
{
	struct ftrace_ret_stack **ret_stack_list;
4475
	int ret, cpu;
4476 4477 4478 4479 4480 4481 4482 4483

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

	if (!ret_stack_list)
		return -ENOMEM;

4484
	/* The cpu_boot init_task->ret_stack will never be freed */
4485 4486
	for_each_online_cpu(cpu) {
		if (!idle_task(cpu)->ret_stack)
4487
			ftrace_graph_init_idle_task(idle_task(cpu), cpu);
4488
	}
4489

4490 4491 4492 4493
	do {
		ret = alloc_retstack_tasklist(ret_stack_list);
	} while (ret == -EAGAIN);

4494
	if (!ret) {
4495
		ret = register_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
4496 4497 4498 4499 4500
		if (ret)
			pr_info("ftrace_graph: Couldn't activate tracepoint"
				" probe to kernel_sched_switch\n");
	}

4501 4502 4503 4504
	kfree(ret_stack_list);
	return ret;
}

4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522 4523 4524 4525
/*
 * 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;
}

4526 4527
int register_ftrace_graph(trace_func_graph_ret_t retfunc,
			trace_func_graph_ent_t entryfunc)
4528
{
4529 4530
	int ret = 0;

S
Steven Rostedt 已提交
4531
	mutex_lock(&ftrace_lock);
4532

4533
	/* we currently allow only one tracer registered at a time */
4534
	if (ftrace_graph_active) {
4535 4536 4537 4538
		ret = -EBUSY;
		goto out;
	}

4539 4540 4541
	ftrace_suspend_notifier.notifier_call = ftrace_suspend_notifier_call;
	register_pm_notifier(&ftrace_suspend_notifier);

4542
	ftrace_graph_active++;
4543
	ret = start_graph_tracing();
4544
	if (ret) {
4545
		ftrace_graph_active--;
4546 4547
		goto out;
	}
4548

4549 4550
	ftrace_graph_return = retfunc;
	ftrace_graph_entry = entryfunc;
4551

4552
	ret = ftrace_startup(&global_ops, FTRACE_START_FUNC_RET);
4553 4554

out:
S
Steven Rostedt 已提交
4555
	mutex_unlock(&ftrace_lock);
4556
	return ret;
4557 4558
}

4559
void unregister_ftrace_graph(void)
4560
{
S
Steven Rostedt 已提交
4561
	mutex_lock(&ftrace_lock);
4562

4563
	if (unlikely(!ftrace_graph_active))
4564 4565
		goto out;

4566
	ftrace_graph_active--;
4567
	ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub;
4568
	ftrace_graph_entry = ftrace_graph_entry_stub;
4569
	ftrace_shutdown(&global_ops, FTRACE_STOP_FUNC_RET);
4570
	unregister_pm_notifier(&ftrace_suspend_notifier);
4571
	unregister_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
4572

4573
 out:
S
Steven Rostedt 已提交
4574
	mutex_unlock(&ftrace_lock);
4575
}
4576

4577 4578 4579 4580 4581 4582 4583 4584
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
4585
	/* make curr_ret_stack visible before we add the ret_stack */
4586 4587 4588 4589 4590 4591 4592 4593 4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619
	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);
	}
}

4620
/* Allocate a return stack for newly created task */
4621
void ftrace_graph_init_task(struct task_struct *t)
4622
{
4623 4624
	/* Make sure we do not use the parent ret_stack */
	t->ret_stack = NULL;
4625
	t->curr_ret_stack = -1;
4626

4627
	if (ftrace_graph_active) {
4628 4629 4630
		struct ftrace_ret_stack *ret_stack;

		ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH
4631 4632
				* sizeof(struct ftrace_ret_stack),
				GFP_KERNEL);
4633
		if (!ret_stack)
4634
			return;
4635
		graph_init_task(t, ret_stack);
4636
	}
4637 4638
}

4639
void ftrace_graph_exit_task(struct task_struct *t)
4640
{
4641 4642
	struct ftrace_ret_stack	*ret_stack = t->ret_stack;

4643
	t->ret_stack = NULL;
4644 4645 4646 4647
	/* NULL must become visible to IRQs before we free it: */
	barrier();

	kfree(ret_stack);
4648
}
S
Steven Rostedt 已提交
4649 4650 4651 4652 4653

void ftrace_graph_stop(void)
{
	ftrace_stop();
}
4654
#endif