ftrace.c 127.4 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12
/*
 * 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
13
 *  Copyright (C) 2004 Nadia Yvette Chambers
14 15
 */

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
#define FL_GLOBAL_CONTROL_MASK (FTRACE_OPS_FL_CONTROL)
66

67
#ifdef CONFIG_DYNAMIC_FTRACE
68 69 70
#define INIT_OPS_HASH(opsname)	\
	.func_hash		= &opsname.local_hash,			\
	.local_hash.regex_lock	= __MUTEX_INITIALIZER(opsname.local_hash.regex_lock),
71 72 73
#define ASSIGN_OPS_HASH(opsname, val) \
	.func_hash		= val, \
	.local_hash.regex_lock	= __MUTEX_INITIALIZER(opsname.local_hash.regex_lock),
74
#else
75
#define INIT_OPS_HASH(opsname)
76
#define ASSIGN_OPS_HASH(opsname, val)
77 78
#endif

79 80
static struct ftrace_ops ftrace_list_end __read_mostly = {
	.func		= ftrace_stub,
81
	.flags		= FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_STUB,
82
	INIT_OPS_HASH(ftrace_list_end)
83 84
};

85 86
/* ftrace_enabled is a method to turn ftrace on or off */
int ftrace_enabled __read_mostly;
87
static int last_ftrace_enabled;
88

89 90
/* Current function tracing op */
struct ftrace_ops *function_trace_op __read_mostly = &ftrace_list_end;
91 92
/* What to set function_trace_op to */
static struct ftrace_ops *set_function_trace_op;
93

94 95 96 97 98 99 100
/* List for set_ftrace_pid's pids. */
LIST_HEAD(ftrace_pids);
struct ftrace_pid {
	struct list_head list;
	struct pid *pid;
};

101 102 103 104 105 106
/*
 * ftrace_disabled is set when an anomaly is discovered.
 * ftrace_disabled is much stronger than ftrace_enabled.
 */
static int ftrace_disabled __read_mostly;

107
static DEFINE_MUTEX(ftrace_lock);
108

109
static struct ftrace_ops *ftrace_control_list __read_mostly = &ftrace_list_end;
110
static struct ftrace_ops *ftrace_ops_list __read_mostly = &ftrace_list_end;
111
ftrace_func_t ftrace_trace_function __read_mostly = ftrace_stub;
112
ftrace_func_t ftrace_pid_function __read_mostly = ftrace_stub;
113
static struct ftrace_ops global_ops;
114
static struct ftrace_ops control_ops;
115

116 117 118
static void ftrace_ops_recurs_func(unsigned long ip, unsigned long parent_ip,
				   struct ftrace_ops *op, struct pt_regs *regs);

119 120
#if ARCH_SUPPORTS_FTRACE_OPS
static void ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
121
				 struct ftrace_ops *op, struct pt_regs *regs);
122 123 124 125 126
#else
/* See comment below, where ftrace_ops_list_func is defined */
static void ftrace_ops_no_ops(unsigned long ip, unsigned long parent_ip);
#define ftrace_ops_list_func ((ftrace_func_t)ftrace_ops_no_ops)
#endif
127

128 129
/*
 * Traverse the ftrace_global_list, invoking all entries.  The reason that we
130
 * can use rcu_dereference_raw_notrace() is that elements removed from this list
131
 * are simply leaked, so there is no need to interact with a grace-period
132
 * mechanism.  The rcu_dereference_raw_notrace() calls are needed to handle
133 134 135 136 137
 * concurrent insertions into the ftrace_global_list.
 *
 * Silly Alpha and silly pointer-speculation compiler optimizations!
 */
#define do_for_each_ftrace_op(op, list)			\
138
	op = rcu_dereference_raw_notrace(list);			\
139 140 141 142 143 144
	do

/*
 * Optimized for just a single item in the list (as that is the normal case).
 */
#define while_for_each_ftrace_op(op)				\
145
	while (likely(op = rcu_dereference_raw_notrace((op)->next)) &&	\
146 147
	       unlikely((op) != &ftrace_list_end))

148 149 150 151
static inline void ftrace_ops_init(struct ftrace_ops *ops)
{
#ifdef CONFIG_DYNAMIC_FTRACE
	if (!(ops->flags & FTRACE_OPS_FL_INITIALIZED)) {
152 153
		mutex_init(&ops->local_hash.regex_lock);
		ops->func_hash = &ops->local_hash;
154 155 156 157 158
		ops->flags |= FTRACE_OPS_FL_INITIALIZED;
	}
#endif
}

159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179
/**
 * ftrace_nr_registered_ops - return number of ops registered
 *
 * Returns the number of ftrace_ops registered and tracing functions
 */
int ftrace_nr_registered_ops(void)
{
	struct ftrace_ops *ops;
	int cnt = 0;

	mutex_lock(&ftrace_lock);

	for (ops = ftrace_ops_list;
	     ops != &ftrace_list_end; ops = ops->next)
		cnt++;

	mutex_unlock(&ftrace_lock);

	return cnt;
}

180
static void ftrace_pid_func(unsigned long ip, unsigned long parent_ip,
181
			    struct ftrace_ops *op, struct pt_regs *regs)
182
{
183
	if (!test_tsk_trace_trace(current))
184 185
		return;

186
	ftrace_pid_function(ip, parent_ip, op, regs);
187 188 189 190 191 192 193 194 195
}

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;
}

196
/**
197
 * clear_ftrace_function - reset the ftrace function
198
 *
199 200
 * This NULLs the ftrace function and in essence stops
 * tracing.  There may be lag
201
 */
202
void clear_ftrace_function(void)
203
{
204
	ftrace_trace_function = ftrace_stub;
205
	ftrace_pid_function = ftrace_stub;
206 207
}

208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228
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;
}

229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245
static void ftrace_sync(struct work_struct *work)
{
	/*
	 * This function is just a stub to implement a hard force
	 * of synchronize_sched(). This requires synchronizing
	 * tasks even in userspace and idle.
	 *
	 * Yes, function tracing is rude.
	 */
}

static void ftrace_sync_ipi(void *data)
{
	/* Probably not needed, but do it anyway */
	smp_rmb();
}

246 247 248 249 250 251
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static void update_function_graph_func(void);
#else
static inline void update_function_graph_func(void) { }
#endif

252 253 254 255
static void update_ftrace_function(void)
{
	ftrace_func_t func;

256 257 258 259 260 261 262 263 264 265 266
	/*
	 * Prepare the ftrace_ops that the arch callback will use.
	 * If there's only one ftrace_ops registered, the ftrace_ops_list
	 * will point to the ops we want.
	 */
	set_function_trace_op = ftrace_ops_list;

	/* If there's no ftrace_ops registered, just call the stub function */
	if (ftrace_ops_list == &ftrace_list_end) {
		func = ftrace_stub;

267 268
	/*
	 * If we are at the end of the list and this ops is
269 270
	 * recursion safe and not dynamic and the arch supports passing ops,
	 * then have the mcount trampoline call the function directly.
271
	 */
272
	} else if (ftrace_ops_list->next == &ftrace_list_end) {
273
		func = ftrace_ops_get_func(ftrace_ops_list);
274

275 276
	} else {
		/* Just use the default ftrace_ops */
277
		set_function_trace_op = &ftrace_list_end;
278
		func = ftrace_ops_list_func;
279
	}
280

281 282
	update_function_graph_func();

283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325
	/* If there's no change, then do nothing more here */
	if (ftrace_trace_function == func)
		return;

	/*
	 * If we are using the list function, it doesn't care
	 * about the function_trace_ops.
	 */
	if (func == ftrace_ops_list_func) {
		ftrace_trace_function = func;
		/*
		 * Don't even bother setting function_trace_ops,
		 * it would be racy to do so anyway.
		 */
		return;
	}

#ifndef CONFIG_DYNAMIC_FTRACE
	/*
	 * For static tracing, we need to be a bit more careful.
	 * The function change takes affect immediately. Thus,
	 * we need to coorditate the setting of the function_trace_ops
	 * with the setting of the ftrace_trace_function.
	 *
	 * Set the function to the list ops, which will call the
	 * function we want, albeit indirectly, but it handles the
	 * ftrace_ops and doesn't depend on function_trace_op.
	 */
	ftrace_trace_function = ftrace_ops_list_func;
	/*
	 * Make sure all CPUs see this. Yes this is slow, but static
	 * tracing is slow and nasty to have enabled.
	 */
	schedule_on_each_cpu(ftrace_sync);
	/* Now all cpus are using the list ops. */
	function_trace_op = set_function_trace_op;
	/* Make sure the function_trace_op is visible on all CPUs */
	smp_wmb();
	/* Nasty way to force a rmb on all cpus */
	smp_call_function(ftrace_sync_ipi, NULL, 1);
	/* OK, we are all set to update the ftrace_trace_function now! */
#endif /* !CONFIG_DYNAMIC_FTRACE */

326 327 328
	ftrace_trace_function = func;
}

329 330 331 332 333
int using_ftrace_ops_list_func(void)
{
	return ftrace_trace_function == ftrace_ops_list_func;
}

334
static void add_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
335
{
336
	ops->next = *list;
337
	/*
338
	 * We are entering ops into the list but another
339 340
	 * CPU might be walking that list. We need to make sure
	 * the ops->next pointer is valid before another CPU sees
341
	 * the ops pointer included into the list.
342
	 */
343
	rcu_assign_pointer(*list, ops);
344 345
}

346
static int remove_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
347 348 349 350
{
	struct ftrace_ops **p;

	/*
351 352
	 * If we are removing the last function, then simply point
	 * to the ftrace_stub.
353
	 */
354 355
	if (*list == ops && ops->next == &ftrace_list_end) {
		*list = &ftrace_list_end;
S
Steven Rostedt 已提交
356
		return 0;
357 358
	}

359
	for (p = list; *p != &ftrace_list_end; p = &(*p)->next)
360 361 362
		if (*p == ops)
			break;

S
Steven Rostedt 已提交
363 364
	if (*p != ops)
		return -1;
365 366

	*p = (*p)->next;
367 368
	return 0;
}
369

370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389
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;
}

390 391
static int __register_ftrace_function(struct ftrace_ops *ops)
{
392 393 394
	if (ops->flags & FTRACE_OPS_FL_DELETED)
		return -EINVAL;

395 396 397
	if (WARN_ON(ops->flags & FTRACE_OPS_FL_ENABLED))
		return -EBUSY;

398
#ifndef CONFIG_DYNAMIC_FTRACE_WITH_REGS
399 400 401 402 403 404 405 406 407 408 409 410 411
	/*
	 * If the ftrace_ops specifies SAVE_REGS, then it only can be used
	 * if the arch supports it, or SAVE_REGS_IF_SUPPORTED is also set.
	 * Setting SAVE_REGS_IF_SUPPORTED makes SAVE_REGS irrelevant.
	 */
	if (ops->flags & FTRACE_OPS_FL_SAVE_REGS &&
	    !(ops->flags & FTRACE_OPS_FL_SAVE_REGS_IF_SUPPORTED))
		return -EINVAL;

	if (ops->flags & FTRACE_OPS_FL_SAVE_REGS_IF_SUPPORTED)
		ops->flags |= FTRACE_OPS_FL_SAVE_REGS;
#endif

412 413 414
	if (!core_kernel_data((unsigned long)ops))
		ops->flags |= FTRACE_OPS_FL_DYNAMIC;

415
	if (ops->flags & FTRACE_OPS_FL_CONTROL) {
416 417 418
		if (control_ops_alloc(ops))
			return -ENOMEM;
		add_ftrace_list_ops(&ftrace_control_list, &control_ops, ops);
419 420 421
	} else
		add_ftrace_ops(&ftrace_ops_list, ops);

422 423 424 425 426 427 428 429 430 431
	if (ftrace_enabled)
		update_ftrace_function();

	return 0;
}

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

432 433 434
	if (WARN_ON(!(ops->flags & FTRACE_OPS_FL_ENABLED)))
		return -EBUSY;

435
	if (ops->flags & FTRACE_OPS_FL_CONTROL) {
436 437
		ret = remove_ftrace_list_ops(&ftrace_control_list,
					     &control_ops, ops);
438 439 440
	} else
		ret = remove_ftrace_ops(&ftrace_ops_list, ops);

441 442
	if (ret < 0)
		return ret;
443

444 445
	if (ftrace_enabled)
		update_ftrace_function();
446

S
Steven Rostedt 已提交
447
	return 0;
448 449
}

450 451
static void ftrace_update_pid_func(void)
{
452
	/* Only do something if we are tracing something */
453
	if (ftrace_trace_function == ftrace_stub)
454
		return;
455

456
	update_ftrace_function();
457 458
}

459 460 461 462 463
#ifdef CONFIG_FUNCTION_PROFILER
struct ftrace_profile {
	struct hlist_node		node;
	unsigned long			ip;
	unsigned long			counter;
464 465
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	unsigned long long		time;
466
	unsigned long long		time_squared;
467
#endif
468 469
};

470 471 472 473
struct ftrace_profile_page {
	struct ftrace_profile_page	*next;
	unsigned long			index;
	struct ftrace_profile		records[];
474 475
};

476 477 478 479 480 481 482 483
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;
};

484 485
#define PROFILE_RECORDS_SIZE						\
	(PAGE_SIZE - offsetof(struct ftrace_profile_page, records))
486

487 488
#define PROFILES_PER_PAGE					\
	(PROFILE_RECORDS_SIZE / sizeof(struct ftrace_profile))
489

490 491 492
static int ftrace_profile_enabled __read_mostly;

/* ftrace_profile_lock - synchronize the enable and disable of the profiler */
S
Steven Rostedt 已提交
493 494
static DEFINE_MUTEX(ftrace_profile_lock);

495
static DEFINE_PER_CPU(struct ftrace_profile_stat, ftrace_profile_stats);
496

497 498
#define FTRACE_PROFILE_HASH_BITS 10
#define FTRACE_PROFILE_HASH_SIZE (1 << FTRACE_PROFILE_HASH_BITS)
499

S
Steven Rostedt 已提交
500 501 502
static void *
function_stat_next(void *v, int idx)
{
503 504
	struct ftrace_profile *rec = v;
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
505

506
	pg = (struct ftrace_profile_page *)((unsigned long)rec & PAGE_MASK);
S
Steven Rostedt 已提交
507 508

 again:
L
Li Zefan 已提交
509 510 511
	if (idx != 0)
		rec++;

S
Steven Rostedt 已提交
512 513 514 515 516
	if ((void *)rec >= (void *)&pg->records[pg->index]) {
		pg = pg->next;
		if (!pg)
			return NULL;
		rec = &pg->records[0];
517 518
		if (!rec->counter)
			goto again;
S
Steven Rostedt 已提交
519 520 521 522 523 524 525
	}

	return rec;
}

static void *function_stat_start(struct tracer_stat *trace)
{
526 527 528 529 530 531 532
	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
533 534
}

535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550
#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
551 552
static int function_stat_cmp(void *p1, void *p2)
{
553 554
	struct ftrace_profile *a = p1;
	struct ftrace_profile *b = p2;
S
Steven Rostedt 已提交
555 556 557 558 559 560 561 562

	if (a->counter < b->counter)
		return -1;
	if (a->counter > b->counter)
		return 1;
	else
		return 0;
}
563
#endif
S
Steven Rostedt 已提交
564 565 566

static int function_stat_headers(struct seq_file *m)
{
567
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
568
	seq_printf(m, "  Function                               "
569
		   "Hit    Time            Avg             s^2\n"
570
		      "  --------                               "
571
		   "---    ----            ---             ---\n");
572
#else
S
Steven Rostedt 已提交
573 574
	seq_printf(m, "  Function                               Hit\n"
		      "  --------                               ---\n");
575
#endif
S
Steven Rostedt 已提交
576 577 578 579 580
	return 0;
}

static int function_stat_show(struct seq_file *m, void *v)
{
581
	struct ftrace_profile *rec = v;
S
Steven Rostedt 已提交
582
	char str[KSYM_SYMBOL_LEN];
583
	int ret = 0;
584
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
585 586
	static struct trace_seq s;
	unsigned long long avg;
587
	unsigned long long stddev;
588
#endif
589 590 591 592 593 594 595
	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
596 597

	kallsyms_lookup(rec->ip, NULL, NULL, NULL, str);
598 599 600 601
	seq_printf(m, "  %-30.30s  %10lu", str, rec->counter);

#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	seq_printf(m, "    ");
602 603 604
	avg = rec->time;
	do_div(avg, rec->counter);

605 606 607 608
	/* Sample standard deviation (s^2) */
	if (rec->counter <= 1)
		stddev = 0;
	else {
609 610 611 612 613 614 615
		/*
		 * Apply Welford's method:
		 * s^2 = 1 / (n * (n-1)) * (n * \Sum (x_i)^2 - (\Sum x_i)^2)
		 */
		stddev = rec->counter * rec->time_squared -
			 rec->time * rec->time;

616 617 618 619
		/*
		 * Divide only 1000 for ns^2 -> us^2 conversion.
		 * trace_print_graph_duration will divide 1000 again.
		 */
620
		do_div(stddev, rec->counter * (rec->counter - 1) * 1000);
621 622
	}

623 624 625 626
	trace_seq_init(&s);
	trace_print_graph_duration(rec->time, &s);
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(avg, &s);
627 628
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(stddev, &s);
629 630 631
	trace_print_seq(m, &s);
#endif
	seq_putc(m, '\n');
632 633
out:
	mutex_unlock(&ftrace_profile_lock);
S
Steven Rostedt 已提交
634

635
	return ret;
S
Steven Rostedt 已提交
636 637
}

638
static void ftrace_profile_reset(struct ftrace_profile_stat *stat)
S
Steven Rostedt 已提交
639
{
640
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
641

642
	pg = stat->pages = stat->start;
S
Steven Rostedt 已提交
643

644 645 646 647
	while (pg) {
		memset(pg->records, 0, PROFILE_RECORDS_SIZE);
		pg->index = 0;
		pg = pg->next;
S
Steven Rostedt 已提交
648 649
	}

650
	memset(stat->hash, 0,
651 652
	       FTRACE_PROFILE_HASH_SIZE * sizeof(struct hlist_head));
}
S
Steven Rostedt 已提交
653

654
int ftrace_profile_pages_init(struct ftrace_profile_stat *stat)
655 656
{
	struct ftrace_profile_page *pg;
657 658
	int functions;
	int pages;
659
	int i;
S
Steven Rostedt 已提交
660

661
	/* If we already allocated, do nothing */
662
	if (stat->pages)
663
		return 0;
S
Steven Rostedt 已提交
664

665 666
	stat->pages = (void *)get_zeroed_page(GFP_KERNEL);
	if (!stat->pages)
667
		return -ENOMEM;
S
Steven Rostedt 已提交
668

669 670 671 672 673 674 675 676 677 678 679 680 681
#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

682
	pg = stat->start = stat->pages;
S
Steven Rostedt 已提交
683

684 685
	pages = DIV_ROUND_UP(functions, PROFILES_PER_PAGE);

686
	for (i = 1; i < pages; i++) {
687 688
		pg->next = (void *)get_zeroed_page(GFP_KERNEL);
		if (!pg->next)
689
			goto out_free;
690 691 692 693
		pg = pg->next;
	}

	return 0;
694 695 696 697 698 699 700 701 702 703 704 705 706 707

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

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

	stat->pages = NULL;
	stat->start = NULL;

	return -ENOMEM;
S
Steven Rostedt 已提交
708 709
}

710
static int ftrace_profile_init_cpu(int cpu)
S
Steven Rostedt 已提交
711
{
712
	struct ftrace_profile_stat *stat;
713
	int size;
S
Steven Rostedt 已提交
714

715 716 717
	stat = &per_cpu(ftrace_profile_stats, cpu);

	if (stat->hash) {
718
		/* If the profile is already created, simply reset it */
719
		ftrace_profile_reset(stat);
720 721
		return 0;
	}
S
Steven Rostedt 已提交
722

723 724 725 726 727
	/*
	 * 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
728

729
	stat->hash = kzalloc(sizeof(struct hlist_head) * size, GFP_KERNEL);
730

731
	if (!stat->hash)
732 733
		return -ENOMEM;

734
	/* Preallocate the function profiling pages */
735 736 737
	if (ftrace_profile_pages_init(stat) < 0) {
		kfree(stat->hash);
		stat->hash = NULL;
738 739 740 741
		return -ENOMEM;
	}

	return 0;
S
Steven Rostedt 已提交
742 743
}

744 745 746 747 748
static int ftrace_profile_init(void)
{
	int cpu;
	int ret = 0;

749
	for_each_possible_cpu(cpu) {
750 751 752 753 754 755 756 757
		ret = ftrace_profile_init_cpu(cpu);
		if (ret)
			break;
	}

	return ret;
}

758
/* interrupts must be disabled */
759 760
static struct ftrace_profile *
ftrace_find_profiled_func(struct ftrace_profile_stat *stat, unsigned long ip)
S
Steven Rostedt 已提交
761
{
762
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
763 764 765
	struct hlist_head *hhd;
	unsigned long key;

766
	key = hash_long(ip, FTRACE_PROFILE_HASH_BITS);
767
	hhd = &stat->hash[key];
S
Steven Rostedt 已提交
768 769 770 771

	if (hlist_empty(hhd))
		return NULL;

772
	hlist_for_each_entry_rcu_notrace(rec, hhd, node) {
S
Steven Rostedt 已提交
773
		if (rec->ip == ip)
774 775 776 777 778 779
			return rec;
	}

	return NULL;
}

780 781
static void ftrace_add_profile(struct ftrace_profile_stat *stat,
			       struct ftrace_profile *rec)
782 783 784
{
	unsigned long key;

785
	key = hash_long(rec->ip, FTRACE_PROFILE_HASH_BITS);
786
	hlist_add_head_rcu(&rec->node, &stat->hash[key]);
787 788
}

789 790 791
/*
 * The memory is already allocated, this simply finds a new record to use.
 */
792
static struct ftrace_profile *
793
ftrace_profile_alloc(struct ftrace_profile_stat *stat, unsigned long ip)
794 795 796
{
	struct ftrace_profile *rec = NULL;

797
	/* prevent recursion (from NMIs) */
798
	if (atomic_inc_return(&stat->disabled) != 1)
799 800 801
		goto out;

	/*
802 803
	 * Try to find the function again since an NMI
	 * could have added it
804
	 */
805
	rec = ftrace_find_profiled_func(stat, ip);
806
	if (rec)
807
		goto out;
808

809 810 811 812
	if (stat->pages->index == PROFILES_PER_PAGE) {
		if (!stat->pages->next)
			goto out;
		stat->pages = stat->pages->next;
S
Steven Rostedt 已提交
813
	}
814

815
	rec = &stat->pages->records[stat->pages->index++];
816
	rec->ip = ip;
817
	ftrace_add_profile(stat, rec);
818

S
Steven Rostedt 已提交
819
 out:
820
	atomic_dec(&stat->disabled);
S
Steven Rostedt 已提交
821 822 823 824 825

	return rec;
}

static void
826
function_profile_call(unsigned long ip, unsigned long parent_ip,
827
		      struct ftrace_ops *ops, struct pt_regs *regs)
S
Steven Rostedt 已提交
828
{
829
	struct ftrace_profile_stat *stat;
830
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
831 832 833 834 835 836
	unsigned long flags;

	if (!ftrace_profile_enabled)
		return;

	local_irq_save(flags);
837

838
	stat = this_cpu_ptr(&ftrace_profile_stats);
839
	if (!stat->hash || !ftrace_profile_enabled)
840 841 842
		goto out;

	rec = ftrace_find_profiled_func(stat, ip);
843
	if (!rec) {
844
		rec = ftrace_profile_alloc(stat, ip);
845 846 847
		if (!rec)
			goto out;
	}
S
Steven Rostedt 已提交
848 849 850 851 852 853

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

854 855 856
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static int profile_graph_entry(struct ftrace_graph_ent *trace)
{
857
	function_profile_call(trace->func, 0, NULL, NULL);
858 859 860 861 862
	return 1;
}

static void profile_graph_return(struct ftrace_graph_ret *trace)
{
863
	struct ftrace_profile_stat *stat;
864
	unsigned long long calltime;
865
	struct ftrace_profile *rec;
866
	unsigned long flags;
867 868

	local_irq_save(flags);
869
	stat = this_cpu_ptr(&ftrace_profile_stats);
870
	if (!stat->hash || !ftrace_profile_enabled)
871 872
		goto out;

873 874 875 876
	/* If the calltime was zero'd ignore it */
	if (!trace->calltime)
		goto out;

877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893
	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;
	}

894
	rec = ftrace_find_profiled_func(stat, trace->func);
895
	if (rec) {
896
		rec->time += calltime;
897 898
		rec->time_squared += calltime * calltime;
	}
899

900
 out:
901 902 903 904 905 906 907 908 909 910 911 912 913 914
	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
915
static struct ftrace_ops ftrace_profile_ops __read_mostly = {
916
	.func		= function_profile_call,
917
	.flags		= FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_INITIALIZED,
918
	INIT_OPS_HASH(ftrace_profile_ops)
S
Steven Rostedt 已提交
919 920
};

921 922 923 924 925 926 927 928 929 930 931
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
932 933 934 935 936 937 938
static ssize_t
ftrace_profile_write(struct file *filp, const char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
	unsigned long val;
	int ret;

939 940
	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
S
Steven Rostedt 已提交
941 942 943 944 945 946 947
		return ret;

	val = !!val;

	mutex_lock(&ftrace_profile_lock);
	if (ftrace_profile_enabled ^ val) {
		if (val) {
948 949 950 951 952 953
			ret = ftrace_profile_init();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}

954 955 956 957 958
			ret = register_ftrace_profiler();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}
S
Steven Rostedt 已提交
959 960 961
			ftrace_profile_enabled = 1;
		} else {
			ftrace_profile_enabled = 0;
962 963 964 965
			/*
			 * unregister_ftrace_profiler calls stop_machine
			 * so this acts like an synchronize_sched.
			 */
966
			unregister_ftrace_profiler();
S
Steven Rostedt 已提交
967 968
		}
	}
969
 out:
S
Steven Rostedt 已提交
970 971
	mutex_unlock(&ftrace_profile_lock);

972
	*ppos += cnt;
S
Steven Rostedt 已提交
973 974 975 976

	return cnt;
}

977 978 979 980
static ssize_t
ftrace_profile_read(struct file *filp, char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
981
	char buf[64];		/* big enough to hold a number */
982 983 984 985 986 987
	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
988 989 990 991
static const struct file_operations ftrace_profile_fops = {
	.open		= tracing_open_generic,
	.read		= ftrace_profile_read,
	.write		= ftrace_profile_write,
992
	.llseek		= default_llseek,
S
Steven Rostedt 已提交
993 994
};

995 996
/* used to initialize the real stat files */
static struct tracer_stat function_stats __initdata = {
997 998 999 1000 1001 1002
	.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
1003 1004
};

1005
static __init void ftrace_profile_debugfs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
1006
{
1007
	struct ftrace_profile_stat *stat;
S
Steven Rostedt 已提交
1008
	struct dentry *entry;
1009
	char *name;
S
Steven Rostedt 已提交
1010
	int ret;
1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038
	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
1039 1040 1041 1042 1043 1044 1045 1046 1047 1048
	}

	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 */
1049
static __init void ftrace_profile_debugfs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
1050 1051 1052 1053
{
}
#endif /* CONFIG_FUNCTION_PROFILER */

1054 1055 1056 1057
static struct pid * const ftrace_swapper_pid = &init_struct_pid;

#ifdef CONFIG_DYNAMIC_FTRACE

1058 1059
static struct ftrace_ops *removed_ops;

1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071
#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;
1072
	struct list_head	free_list;
1073 1074
};

1075 1076 1077 1078 1079 1080 1081 1082 1083
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;
1084
	struct rcu_head		rcu;
1085 1086
};

1087 1088 1089 1090 1091 1092 1093 1094 1095
/*
 * 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,
1096
};
1097
#define EMPTY_HASH	((struct ftrace_hash *)&empty_hash)
1098

1099
static struct ftrace_ops global_ops = {
1100 1101 1102 1103 1104 1105
	.func				= ftrace_stub,
	.local_hash.notrace_hash	= EMPTY_HASH,
	.local_hash.filter_hash		= EMPTY_HASH,
	INIT_OPS_HASH(global_ops)
	.flags				= FTRACE_OPS_FL_RECURSION_SAFE |
					  FTRACE_OPS_FL_INITIALIZED,
1106 1107
};

1108 1109
struct ftrace_page {
	struct ftrace_page	*next;
1110
	struct dyn_ftrace	*records;
1111
	int			index;
1112
	int			size;
1113 1114
};

1115 1116
#define ENTRY_SIZE sizeof(struct dyn_ftrace)
#define ENTRIES_PER_PAGE (PAGE_SIZE / ENTRY_SIZE)
1117 1118 1119 1120 1121 1122 1123

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

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

1124
static bool __always_inline ftrace_hash_empty(struct ftrace_hash *hash)
1125 1126 1127 1128
{
	return !hash || !hash->count;
}

1129 1130 1131 1132 1133 1134 1135
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;

1136
	if (ftrace_hash_empty(hash))
1137 1138 1139 1140 1141 1142 1143 1144 1145
		return NULL;

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

	hhd = &hash->buckets[key];

1146
	hlist_for_each_entry_rcu_notrace(entry, hhd, hlist) {
1147 1148 1149 1150 1151 1152
		if (entry->ip == ip)
			return entry;
	}
	return NULL;
}

1153 1154
static void __add_hash_entry(struct ftrace_hash *hash,
			     struct ftrace_func_entry *entry)
1155 1156 1157 1158 1159
{
	struct hlist_head *hhd;
	unsigned long key;

	if (hash->size_bits)
1160
		key = hash_long(entry->ip, hash->size_bits);
1161 1162 1163 1164 1165 1166
	else
		key = 0;

	hhd = &hash->buckets[key];
	hlist_add_head(&entry->hlist, hhd);
	hash->count++;
1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178
}

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);
1179 1180 1181 1182 1183

	return 0;
}

static void
1184
free_hash_entry(struct ftrace_hash *hash,
1185 1186 1187 1188 1189 1190 1191
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	kfree(entry);
	hash->count--;
}

1192 1193 1194 1195 1196 1197 1198 1199
static void
remove_hash_entry(struct ftrace_hash *hash,
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	hash->count--;
}

1200 1201 1202
static void ftrace_hash_clear(struct ftrace_hash *hash)
{
	struct hlist_head *hhd;
1203
	struct hlist_node *tn;
1204 1205 1206 1207
	struct ftrace_func_entry *entry;
	int size = 1 << hash->size_bits;
	int i;

1208 1209 1210
	if (!hash->count)
		return;

1211 1212
	for (i = 0; i < size; i++) {
		hhd = &hash->buckets[i];
1213
		hlist_for_each_entry_safe(entry, tn, hhd, hlist)
1214
			free_hash_entry(hash, entry);
1215 1216 1217 1218
	}
	FTRACE_WARN_ON(hash->count);
}

1219 1220 1221 1222 1223 1224 1225 1226 1227
static void free_ftrace_hash(struct ftrace_hash *hash)
{
	if (!hash || hash == EMPTY_HASH)
		return;
	ftrace_hash_clear(hash);
	kfree(hash->buckets);
	kfree(hash);
}

1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242
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);
}

1243 1244
void ftrace_free_filter(struct ftrace_ops *ops)
{
1245
	ftrace_ops_init(ops);
1246 1247
	free_ftrace_hash(ops->func_hash->filter_hash);
	free_ftrace_hash(ops->func_hash->notrace_hash);
1248 1249
}

1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
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;
1260
	hash->buckets = kcalloc(size, sizeof(*hash->buckets), GFP_KERNEL);
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285

	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;
	int size;
	int ret;
	int i;

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

	/* Empty hash? */
1286
	if (ftrace_hash_empty(hash))
1287 1288 1289 1290
		return new_hash;

	size = 1 << hash->size_bits;
	for (i = 0; i < size; i++) {
1291
		hlist_for_each_entry(entry, &hash->buckets[i], hlist) {
1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306
			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;
}

1307
static void
1308
ftrace_hash_rec_disable_modify(struct ftrace_ops *ops, int filter_hash);
1309
static void
1310
ftrace_hash_rec_enable_modify(struct ftrace_ops *ops, int filter_hash);
1311

1312
static int
1313 1314
ftrace_hash_move(struct ftrace_ops *ops, int enable,
		 struct ftrace_hash **dst, struct ftrace_hash *src)
1315 1316
{
	struct ftrace_func_entry *entry;
1317
	struct hlist_node *tn;
1318
	struct hlist_head *hhd;
1319
	struct ftrace_hash *new_hash;
1320 1321 1322 1323 1324 1325 1326 1327 1328
	int size = src->count;
	int bits = 0;
	int i;

	/*
	 * If the new source is empty, just free dst and assign it
	 * the empty_hash.
	 */
	if (!src->count) {
1329 1330
		new_hash = EMPTY_HASH;
		goto update;
1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342
	}

	/*
	 * 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;

1343 1344
	new_hash = alloc_ftrace_hash(bits);
	if (!new_hash)
1345
		return -ENOMEM;
1346 1347 1348 1349

	size = 1 << src->size_bits;
	for (i = 0; i < size; i++) {
		hhd = &src->buckets[i];
1350
		hlist_for_each_entry_safe(entry, tn, hhd, hlist) {
1351
			remove_hash_entry(src, entry);
1352
			__add_hash_entry(new_hash, entry);
1353 1354 1355
		}
	}

1356 1357 1358 1359 1360
update:
	/*
	 * Remove the current set, update the hash and add
	 * them back.
	 */
1361
	ftrace_hash_rec_disable_modify(ops, enable);
1362

1363 1364
	rcu_assign_pointer(*dst, new_hash);

1365
	ftrace_hash_rec_enable_modify(ops, enable);
1366

1367
	return 0;
1368 1369
}

1370 1371 1372 1373 1374 1375 1376 1377
/*
 * 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.
1378 1379 1380
 *
 * This needs to be called with preemption disabled as
 * the hashes are freed with call_rcu_sched().
1381 1382
 */
static int
1383
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
1384 1385 1386 1387 1388
{
	struct ftrace_hash *filter_hash;
	struct ftrace_hash *notrace_hash;
	int ret;

1389 1390 1391 1392 1393 1394 1395 1396 1397 1398
#ifdef CONFIG_DYNAMIC_FTRACE_WITH_REGS
	/*
	 * There's a small race when adding ops that the ftrace handler
	 * that wants regs, may be called without them. We can not
	 * allow that handler to be called if regs is NULL.
	 */
	if (regs == NULL && (ops->flags & FTRACE_OPS_FL_SAVE_REGS))
		return 0;
#endif

1399 1400
	filter_hash = rcu_dereference_raw_notrace(ops->func_hash->filter_hash);
	notrace_hash = rcu_dereference_raw_notrace(ops->func_hash->notrace_hash);
1401

1402
	if ((ftrace_hash_empty(filter_hash) ||
1403
	     ftrace_lookup_ip(filter_hash, ip)) &&
1404
	    (ftrace_hash_empty(notrace_hash) ||
1405 1406 1407 1408 1409 1410 1411 1412
	     !ftrace_lookup_ip(notrace_hash, ip)))
		ret = 1;
	else
		ret = 0;

	return ret;
}

1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426
/*
 * 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()		\
		}				\
	}

1427 1428 1429

static int ftrace_cmp_recs(const void *a, const void *b)
{
1430 1431
	const struct dyn_ftrace *key = a;
	const struct dyn_ftrace *rec = b;
1432

1433
	if (key->flags < rec->ip)
1434
		return -1;
1435 1436
	if (key->ip >= rec->ip + MCOUNT_INSN_SIZE)
		return 1;
1437 1438 1439
	return 0;
}

1440
static unsigned long ftrace_location_range(unsigned long start, unsigned long end)
1441 1442 1443
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1444
	struct dyn_ftrace key;
1445

1446 1447
	key.ip = start;
	key.flags = end;	/* overload flags, as it is unsigned long */
1448 1449

	for (pg = ftrace_pages_start; pg; pg = pg->next) {
1450 1451
		if (end < pg->records[0].ip ||
		    start >= (pg->records[pg->index - 1].ip + MCOUNT_INSN_SIZE))
1452
			continue;
1453 1454 1455 1456
		rec = bsearch(&key, pg->records, pg->index,
			      sizeof(struct dyn_ftrace),
			      ftrace_cmp_recs);
		if (rec)
1457
			return rec->ip;
1458
	}
1459 1460 1461 1462

	return 0;
}

1463 1464 1465 1466
/**
 * ftrace_location - return true if the ip giving is a traced location
 * @ip: the instruction pointer to check
 *
1467
 * Returns rec->ip if @ip given is a pointer to a ftrace location.
1468 1469 1470 1471
 * 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.
 */
1472
unsigned long ftrace_location(unsigned long ip)
1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486
{
	return ftrace_location_range(ip, ip);
}

/**
 * ftrace_text_reserved - return true if range contains an ftrace location
 * @start: start of range to search
 * @end: end of range to search (inclusive). @end points to the last byte to check.
 *
 * Returns 1 if @start and @end contains 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.
 */
1487
int ftrace_text_reserved(const void *start, const void *end)
1488
{
1489 1490 1491 1492 1493 1494
	unsigned long ret;

	ret = ftrace_location_range((unsigned long)start,
				    (unsigned long)end);

	return (int)!!ret;
1495 1496
}

1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516
/* Test if ops registered to this rec needs regs */
static bool test_rec_ops_needs_regs(struct dyn_ftrace *rec)
{
	struct ftrace_ops *ops;
	bool keep_regs = false;

	for (ops = ftrace_ops_list;
	     ops != &ftrace_list_end; ops = ops->next) {
		/* pass rec in as regs to have non-NULL val */
		if (ftrace_ops_test(ops, rec->ip, rec)) {
			if (ops->flags & FTRACE_OPS_FL_SAVE_REGS) {
				keep_regs = true;
				break;
			}
		}
	}

	return  keep_regs;
}

1517 1518 1519
static void ftrace_remove_tramp(struct ftrace_ops *ops,
				struct dyn_ftrace *rec)
{
1520 1521
	/* If TRAMP is not set, no ops should have a trampoline for this */
	if (!(rec->flags & FTRACE_FL_TRAMP))
1522 1523
		return;

1524 1525 1526 1527 1528 1529
	rec->flags &= ~FTRACE_FL_TRAMP;

	if ((!ftrace_hash_empty(ops->func_hash->filter_hash) &&
	     !ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip)) ||
	    ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip))
		return;
1530 1531 1532 1533
	/*
	 * The tramp_hash entry will be removed at time
	 * of update.
	 */
1534
	ops->nr_trampolines--;
1535 1536
}

1537
static void ftrace_clear_tramps(struct dyn_ftrace *rec, struct ftrace_ops *ops)
1538 1539 1540
{
	struct ftrace_ops *op;

1541 1542 1543 1544
	/* If TRAMP is not set, no ops should have a trampoline for this */
	if (!(rec->flags & FTRACE_FL_TRAMP))
		return;

1545
	do_for_each_ftrace_op(op, ftrace_ops_list) {
1546 1547 1548 1549 1550 1551
		/*
		 * This function is called to clear other tramps
		 * not the one that is being updated.
		 */
		if (op == ops)
			continue;
1552
		if (op->nr_trampolines)
1553 1554 1555 1556
			ftrace_remove_tramp(op, rec);
	} while_for_each_ftrace_op(op);
}

1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583
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) {
1584 1585
		hash = ops->func_hash->filter_hash;
		other_hash = ops->func_hash->notrace_hash;
1586
		if (ftrace_hash_empty(hash))
1587 1588 1589
			all = 1;
	} else {
		inc = !inc;
1590 1591
		hash = ops->func_hash->notrace_hash;
		other_hash = ops->func_hash->filter_hash;
1592 1593 1594 1595
		/*
		 * If the notrace hash has no items,
		 * then there's nothing to do.
		 */
1596
		if (ftrace_hash_empty(hash))
1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609
			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.
			 */
1610
			if (!other_hash || !ftrace_lookup_ip(other_hash, rec->ip))
1611 1612
				match = 1;
		} else {
1613 1614
			in_hash = !!ftrace_lookup_ip(hash, rec->ip);
			in_other_hash = !!ftrace_lookup_ip(other_hash, rec->ip);
1615 1616

			/*
1617 1618
			 * If filter_hash is set, we want to match all functions
			 * that are in the hash but not in the other hash.
1619
			 *
1620 1621 1622 1623 1624
			 * If filter_hash is not set, then we are decrementing.
			 * That means we match anything that is in the hash
			 * and also in the other_hash. That is, we need to turn
			 * off functions in the other hash because they are disabled
			 * by this hash.
1625 1626 1627 1628
			 */
			if (filter_hash && in_hash && !in_other_hash)
				match = 1;
			else if (!filter_hash && in_hash &&
1629
				 (in_other_hash || ftrace_hash_empty(other_hash)))
1630 1631 1632 1633 1634 1635 1636
				match = 1;
		}
		if (!match)
			continue;

		if (inc) {
			rec->flags++;
1637
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == FTRACE_REF_MAX))
1638
				return;
1639 1640 1641 1642 1643 1644 1645 1646

			/*
			 * If there's only a single callback registered to a
			 * function, and the ops has a trampoline registered
			 * for it, then we can call it directly.
			 */
			if (ftrace_rec_count(rec) == 1 && ops->trampoline) {
				rec->flags |= FTRACE_FL_TRAMP;
1647
				ops->nr_trampolines++;
1648 1649 1650 1651
			} else {
				/*
				 * If we are adding another function callback
				 * to this function, and the previous had a
1652 1653
				 * custom trampoline in use, then we need to go
				 * back to the default trampoline.
1654
				 */
1655
				ftrace_clear_tramps(rec, ops);
1656 1657
			}

1658 1659 1660 1661 1662 1663
			/*
			 * If any ops wants regs saved for this function
			 * then all ops will get saved regs.
			 */
			if (ops->flags & FTRACE_OPS_FL_SAVE_REGS)
				rec->flags |= FTRACE_FL_REGS;
1664
		} else {
1665
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == 0))
1666 1667
				return;
			rec->flags--;
1668 1669 1670 1671

			if (ops->trampoline && !ftrace_rec_count(rec))
				ftrace_remove_tramp(ops, rec);

1672 1673 1674 1675 1676 1677
			/*
			 * If the rec had REGS enabled and the ops that is
			 * being removed had REGS set, then see if there is
			 * still any ops for this record that wants regs.
			 * If not, we can stop recording them.
			 */
1678
			if (ftrace_rec_count(rec) > 0 &&
1679 1680 1681 1682 1683
			    rec->flags & FTRACE_FL_REGS &&
			    ops->flags & FTRACE_OPS_FL_SAVE_REGS) {
				if (!test_rec_ops_needs_regs(rec))
					rec->flags &= ~FTRACE_FL_REGS;
			}
1684 1685 1686 1687 1688

			/*
			 * flags will be cleared in ftrace_check_record()
			 * if rec count is zero.
			 */
1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708
		}
		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);
}

1709 1710 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
static void ftrace_hash_rec_update_modify(struct ftrace_ops *ops,
					  int filter_hash, int inc)
{
	struct ftrace_ops *op;

	__ftrace_hash_rec_update(ops, filter_hash, inc);

	if (ops->func_hash != &global_ops.local_hash)
		return;

	/*
	 * If the ops shares the global_ops hash, then we need to update
	 * all ops that are enabled and use this hash.
	 */
	do_for_each_ftrace_op(op, ftrace_ops_list) {
		/* Already done */
		if (op == ops)
			continue;
		if (op->func_hash == &global_ops.local_hash)
			__ftrace_hash_rec_update(op, filter_hash, inc);
	} while_for_each_ftrace_op(op);
}

static void ftrace_hash_rec_disable_modify(struct ftrace_ops *ops,
					   int filter_hash)
{
	ftrace_hash_rec_update_modify(ops, filter_hash, 0);
}

static void ftrace_hash_rec_enable_modify(struct ftrace_ops *ops,
					  int filter_hash)
{
	ftrace_hash_rec_update_modify(ops, filter_hash, 1);
}

1744 1745 1746 1747 1748 1749 1750 1751 1752 1753
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]);
}

1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766
/**
 * 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)
1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792
{
	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);
	}
}

1793
static int ftrace_check_record(struct dyn_ftrace *rec, int enable, int update)
1794
{
1795
	unsigned long flag = 0UL;
1796

S
Steven Rostedt 已提交
1797
	/*
1798
	 * If we are updating calls:
S
Steven Rostedt 已提交
1799
	 *
1800 1801
	 *   If the record has a ref count, then we need to enable it
	 *   because someone is using it.
S
Steven Rostedt 已提交
1802
	 *
1803 1804
	 *   Otherwise we make sure its disabled.
	 *
1805
	 * If we are disabling calls, then disable all records that
1806
	 * are enabled.
S
Steven Rostedt 已提交
1807
	 */
1808
	if (enable && ftrace_rec_count(rec))
1809
		flag = FTRACE_FL_ENABLED;
S
Steven Rostedt 已提交
1810

1811
	/*
1812 1813 1814
	 * If enabling and the REGS flag does not match the REGS_EN, or
	 * the TRAMP flag doesn't match the TRAMP_EN, then do not ignore
	 * this record. Set flags to fail the compare against ENABLED.
1815
	 */
1816 1817 1818 1819 1820 1821 1822 1823 1824
	if (flag) {
		if (!(rec->flags & FTRACE_FL_REGS) != 
		    !(rec->flags & FTRACE_FL_REGS_EN))
			flag |= FTRACE_FL_REGS;

		if (!(rec->flags & FTRACE_FL_TRAMP) != 
		    !(rec->flags & FTRACE_FL_TRAMP_EN))
			flag |= FTRACE_FL_TRAMP;
	}
1825

1826 1827
	/* If the state of this record hasn't changed, then do nothing */
	if ((rec->flags & FTRACE_FL_ENABLED) == flag)
1828
		return FTRACE_UPDATE_IGNORE;
S
Steven Rostedt 已提交
1829

1830
	if (flag) {
1831 1832 1833 1834
		/* Save off if rec is being enabled (for return value) */
		flag ^= rec->flags & FTRACE_FL_ENABLED;

		if (update) {
1835
			rec->flags |= FTRACE_FL_ENABLED;
1836 1837 1838 1839 1840 1841
			if (flag & FTRACE_FL_REGS) {
				if (rec->flags & FTRACE_FL_REGS)
					rec->flags |= FTRACE_FL_REGS_EN;
				else
					rec->flags &= ~FTRACE_FL_REGS_EN;
			}
1842 1843 1844 1845 1846 1847
			if (flag & FTRACE_FL_TRAMP) {
				if (rec->flags & FTRACE_FL_TRAMP)
					rec->flags |= FTRACE_FL_TRAMP_EN;
				else
					rec->flags &= ~FTRACE_FL_TRAMP_EN;
			}
1848 1849 1850 1851 1852 1853 1854
		}

		/*
		 * If this record is being updated from a nop, then
		 *   return UPDATE_MAKE_CALL.
		 * Otherwise,
		 *   return UPDATE_MODIFY_CALL to tell the caller to convert
1855
		 *   from the save regs, to a non-save regs function or
1856
		 *   vice versa, or from a trampoline call.
1857 1858 1859
		 */
		if (flag & FTRACE_FL_ENABLED)
			return FTRACE_UPDATE_MAKE_CALL;
1860 1861

		return FTRACE_UPDATE_MODIFY_CALL;
1862 1863
	}

1864 1865
	if (update) {
		/* If there's no more users, clear all flags */
1866
		if (!ftrace_rec_count(rec))
1867 1868 1869 1870 1871
			rec->flags = 0;
		else
			/* Just disable the record (keep REGS state) */
			rec->flags &= ~FTRACE_FL_ENABLED;
	}
1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902

	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);
}

1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;

	do_for_each_ftrace_op(op, ftrace_ops_list) {

		if (!op->trampoline)
			continue;

		if (ftrace_lookup_ip(op->func_hash->filter_hash, rec->ip) &&
		    (ftrace_hash_empty(op->func_hash->notrace_hash) ||
		     !ftrace_lookup_ip(op->func_hash->notrace_hash, rec->ip)))
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958
static struct ftrace_ops *
ftrace_find_tramp_ops_curr(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;

	/* Removed ops need to be tested first */
	if (removed_ops && removed_ops->tramp_hash) {
		if (ftrace_lookup_ip(removed_ops->tramp_hash, rec->ip))
			return removed_ops;
	}

	do_for_each_ftrace_op(op, ftrace_ops_list) {
		if (!op->tramp_hash)
			continue;

		if (ftrace_lookup_ip(op->tramp_hash, rec->ip))
			return op;

	} while_for_each_ftrace_op(op);

	return NULL;
}

static struct ftrace_ops *
ftrace_find_tramp_ops_new(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;

	do_for_each_ftrace_op(op, ftrace_ops_list) {
		/* pass rec in as regs to have non-NULL val */
		if (ftrace_ops_test(op, rec->ip, rec))
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970
/**
 * ftrace_get_addr_new - Get the call address to set to
 * @rec:  The ftrace record descriptor
 *
 * If the record has the FTRACE_FL_REGS set, that means that it
 * wants to convert to a callback that saves all regs. If FTRACE_FL_REGS
 * is not not set, then it wants to convert to the normal callback.
 *
 * Returns the address of the trampoline to set to
 */
unsigned long ftrace_get_addr_new(struct dyn_ftrace *rec)
{
1971 1972 1973 1974 1975 1976
	struct ftrace_ops *ops;

	/* Trampolines take precedence over regs */
	if (rec->flags & FTRACE_FL_TRAMP) {
		ops = ftrace_find_tramp_ops_new(rec);
		if (FTRACE_WARN_ON(!ops || !ops->trampoline)) {
1977 1978
			pr_warn("Bad trampoline accounting at: %p (%pS) (%lx)\n",
				(void *)rec->ip, (void *)rec->ip, rec->flags);
1979 1980 1981 1982 1983 1984
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
	if (rec->flags & FTRACE_FL_REGS)
		return (unsigned long)FTRACE_REGS_ADDR;
	else
		return (unsigned long)FTRACE_ADDR;
}

/**
 * ftrace_get_addr_curr - Get the call address that is already there
 * @rec:  The ftrace record descriptor
 *
 * The FTRACE_FL_REGS_EN is set when the record already points to
 * a function that saves all the regs. Basically the '_EN' version
 * represents the current state of the function.
 *
 * Returns the address of the trampoline that is currently being called
 */
unsigned long ftrace_get_addr_curr(struct dyn_ftrace *rec)
{
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
	struct ftrace_ops *ops;

	/* Trampolines take precedence over regs */
	if (rec->flags & FTRACE_FL_TRAMP_EN) {
		ops = ftrace_find_tramp_ops_curr(rec);
		if (FTRACE_WARN_ON(!ops)) {
			pr_warning("Bad trampoline accounting at: %p (%pS)\n",
				    (void *)rec->ip, (void *)rec->ip);
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

2017 2018 2019 2020 2021 2022
	if (rec->flags & FTRACE_FL_REGS_EN)
		return (unsigned long)FTRACE_REGS_ADDR;
	else
		return (unsigned long)FTRACE_ADDR;
}

2023 2024 2025
static int
__ftrace_replace_code(struct dyn_ftrace *rec, int enable)
{
2026
	unsigned long ftrace_old_addr;
2027 2028 2029
	unsigned long ftrace_addr;
	int ret;

2030
	ftrace_addr = ftrace_get_addr_new(rec);
2031

2032 2033 2034 2035
	/* This needs to be done before we call ftrace_update_record */
	ftrace_old_addr = ftrace_get_addr_curr(rec);

	ret = ftrace_update_record(rec, enable);
2036

2037 2038 2039 2040 2041
	switch (ret) {
	case FTRACE_UPDATE_IGNORE:
		return 0;

	case FTRACE_UPDATE_MAKE_CALL:
2042
		return ftrace_make_call(rec, ftrace_addr);
2043 2044

	case FTRACE_UPDATE_MAKE_NOP:
2045
		return ftrace_make_nop(NULL, rec, ftrace_old_addr);
2046 2047 2048

	case FTRACE_UPDATE_MODIFY_CALL:
		return ftrace_modify_call(rec, ftrace_old_addr, ftrace_addr);
2049 2050
	}

2051
	return -1; /* unknow ftrace bug */
2052 2053
}

2054
void __weak ftrace_replace_code(int enable)
2055 2056 2057
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
S
Steven Rostedt 已提交
2058
	int failed;
2059

2060 2061 2062
	if (unlikely(ftrace_disabled))
		return;

2063
	do_for_each_ftrace_rec(pg, rec) {
2064
		failed = __ftrace_replace_code(rec, enable);
2065
		if (failed) {
2066 2067 2068
			ftrace_bug(failed, rec->ip);
			/* Stop processing */
			return;
2069
		}
2070
	} while_for_each_ftrace_rec();
2071 2072
}

2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144
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];
}

2145
static int
2146
ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec)
2147 2148
{
	unsigned long ip;
2149
	int ret;
2150 2151 2152

	ip = rec->ip;

2153 2154 2155
	if (unlikely(ftrace_disabled))
		return 0;

2156
	ret = ftrace_make_nop(mod, rec, MCOUNT_ADDR);
2157
	if (ret) {
2158
		ftrace_bug(ret, ip);
2159
		return 0;
2160
	}
2161
	return 1;
2162 2163
}

2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181
/*
 * 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;
}

2182
void ftrace_modify_all_code(int command)
2183
{
2184
	int update = command & FTRACE_UPDATE_TRACE_FUNC;
2185
	int err = 0;
2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196

	/*
	 * If the ftrace_caller calls a ftrace_ops func directly,
	 * we need to make sure that it only traces functions it
	 * expects to trace. When doing the switch of functions,
	 * we need to update to the ftrace_ops_list_func first
	 * before the transition between old and new calls are set,
	 * as the ftrace_ops_list_func will check the ops hashes
	 * to make sure the ops are having the right functions
	 * traced.
	 */
2197 2198 2199 2200 2201
	if (update) {
		err = ftrace_update_ftrace_func(ftrace_ops_list_func);
		if (FTRACE_WARN_ON(err))
			return;
	}
2202

2203
	if (command & FTRACE_UPDATE_CALLS)
2204
		ftrace_replace_code(1);
2205
	else if (command & FTRACE_DISABLE_CALLS)
2206 2207
		ftrace_replace_code(0);

2208 2209 2210 2211 2212 2213
	if (update && ftrace_trace_function != ftrace_ops_list_func) {
		function_trace_op = set_function_trace_op;
		smp_wmb();
		/* If irqs are disabled, we are in stop machine */
		if (!irqs_disabled())
			smp_call_function(ftrace_sync_ipi, NULL, 1);
2214 2215 2216
		err = ftrace_update_ftrace_func(ftrace_trace_function);
		if (FTRACE_WARN_ON(err))
			return;
2217
	}
2218

2219
	if (command & FTRACE_START_FUNC_RET)
2220
		err = ftrace_enable_ftrace_graph_caller();
2221
	else if (command & FTRACE_STOP_FUNC_RET)
2222 2223
		err = ftrace_disable_ftrace_graph_caller();
	FTRACE_WARN_ON(err);
2224 2225 2226 2227 2228 2229 2230
}

static int __ftrace_modify_code(void *data)
{
	int *command = data;

	ftrace_modify_all_code(*command);
2231

2232
	return 0;
2233 2234
}

2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258
/**
 * 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);
}

2259 2260 2261 2262 2263 2264 2265
static int ftrace_save_ops_tramp_hash(struct ftrace_ops *ops)
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
	int size, bits;
	int ret;

2266
	size = ops->nr_trampolines;
2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288
	bits = 0;
	/*
	 * Make the hash size about 1/2 the # found
	 */
	for (size /= 2; size; size >>= 1)
		bits++;

	ops->tramp_hash = alloc_ftrace_hash(bits);
	/*
	 * TODO: a failed allocation is going to screw up
	 * the accounting of what needs to be modified
	 * and not. For now, we kill ftrace if we fail
	 * to allocate here. But there are ways around this,
	 * but that will take a little more work.
	 */
	if (!ops->tramp_hash)
		return -ENOMEM;

	do_for_each_ftrace_rec(pg, rec) {
		if (ftrace_rec_count(rec) == 1 &&
		    ftrace_ops_test(ops, rec->ip, rec)) {

2289 2290 2291 2292 2293 2294 2295 2296
			/*
			 * If another ops adds to a rec, the rec will
			 * lose its trampoline and never get it back
			 * until all ops are off of it.
			 */
			if (!(rec->flags & FTRACE_FL_TRAMP))
				continue;

2297 2298 2299 2300 2301 2302 2303 2304 2305 2306
			/* This record had better have a trampoline */
			if (FTRACE_WARN_ON(!(rec->flags & FTRACE_FL_TRAMP_EN)))
				return -1;

			ret = add_hash_entry(ops->tramp_hash, rec->ip);
			if (ret < 0)
				return ret;
		}
	} while_for_each_ftrace_rec();

2307
	/* The number of recs in the hash must match nr_trampolines */
2308 2309 2310 2311
	if (FTRACE_WARN_ON(ops->tramp_hash->count != ops->nr_trampolines))
		pr_warn("count=%ld trampolines=%d\n",
			ops->tramp_hash->count,
			ops->nr_trampolines);
2312

2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333
	return 0;
}

static int ftrace_save_tramp_hashes(void)
{
	struct ftrace_ops *op;
	int ret;

	/*
	 * Now that any trampoline is being used, we need to save the
	 * hashes for the ops that have them. This allows the mapping
	 * back from the record to the ops that has the trampoline to
	 * know what code is being replaced. Modifying code must always
	 * verify what it is changing.
	 */
	do_for_each_ftrace_op(op, ftrace_ops_list) {

		/* The tramp_hash is recreated each time. */
		free_ftrace_hash(op->tramp_hash);
		op->tramp_hash = NULL;

2334
		if (op->nr_trampolines) {
2335 2336 2337 2338 2339 2340 2341 2342 2343 2344
			ret = ftrace_save_ops_tramp_hash(op);
			if (ret)
				return ret;
		}

	} while_for_each_ftrace_op(op);

	return 0;
}

I
Ingo Molnar 已提交
2345
static void ftrace_run_update_code(int command)
2346
{
2347 2348 2349 2350 2351 2352 2353
	int ret;

	ret = ftrace_arch_code_modify_prepare();
	FTRACE_WARN_ON(ret);
	if (ret)
		return;

2354 2355 2356 2357 2358 2359 2360 2361
	/*
	 * 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);

2362 2363
	ret = ftrace_arch_code_modify_post_process();
	FTRACE_WARN_ON(ret);
2364 2365 2366

	ret = ftrace_save_tramp_hashes();
	FTRACE_WARN_ON(ret);
2367 2368
}

2369
static ftrace_func_t saved_ftrace_func;
2370
static int ftrace_start_up;
2371

2372 2373 2374 2375 2376
static void control_ops_free(struct ftrace_ops *ops)
{
	free_percpu(ops->disabled);
}

2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388
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);
}
2389

2390
static int ftrace_startup(struct ftrace_ops *ops, int command)
2391
{
2392
	int ret;
2393

2394
	if (unlikely(ftrace_disabled))
2395
		return -ENODEV;
2396

2397 2398 2399 2400
	ret = __register_ftrace_function(ops);
	if (ret)
		return ret;

2401
	ftrace_start_up++;
2402
	command |= FTRACE_UPDATE_CALLS;
2403

2404
	ops->flags |= FTRACE_OPS_FL_ENABLED;
2405 2406

	ftrace_hash_rec_enable(ops, 1);
2407

2408
	ftrace_startup_enable(command);
2409 2410

	return 0;
2411 2412
}

2413
static int ftrace_shutdown(struct ftrace_ops *ops, int command)
2414
{
2415
	int ret;
2416

2417
	if (unlikely(ftrace_disabled))
2418 2419 2420 2421 2422
		return -ENODEV;

	ret = __unregister_ftrace_function(ops);
	if (ret)
		return ret;
2423

2424
	ftrace_start_up--;
2425 2426 2427 2428 2429 2430 2431
	/*
	 * 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);

2432
	ftrace_hash_rec_disable(ops, 1);
2433

2434
	ops->flags &= ~FTRACE_OPS_FL_ENABLED;
2435

2436
	command |= FTRACE_UPDATE_CALLS;
2437

2438 2439 2440 2441
	if (saved_ftrace_func != ftrace_trace_function) {
		saved_ftrace_func = ftrace_trace_function;
		command |= FTRACE_UPDATE_TRACE_FUNC;
	}
2442

2443 2444 2445 2446 2447 2448 2449 2450 2451
	if (!command || !ftrace_enabled) {
		/*
		 * If these are control ops, they still need their
		 * per_cpu field freed. Since, function tracing is
		 * not currently active, we can just free them
		 * without synchronizing all CPUs.
		 */
		if (ops->flags & FTRACE_OPS_FL_CONTROL)
			control_ops_free(ops);
2452
		return 0;
2453
	}
2454

2455 2456 2457 2458 2459 2460
	/*
	 * If the ops uses a trampoline, then it needs to be
	 * tested first on update.
	 */
	removed_ops = ops;

2461
	ftrace_run_update_code(command);
2462

2463 2464
	removed_ops = NULL;

2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485
	/*
	 * Dynamic ops may be freed, we must make sure that all
	 * callers are done before leaving this function.
	 * The same goes for freeing the per_cpu data of the control
	 * ops.
	 *
	 * Again, normal synchronize_sched() is not good enough.
	 * We need to do a hard force of sched synchronization.
	 * This is because we use preempt_disable() to do RCU, but
	 * the function tracers can be called where RCU is not watching
	 * (like before user_exit()). We can not rely on the RCU
	 * infrastructure to do the synchronization, thus we must do it
	 * ourselves.
	 */
	if (ops->flags & (FTRACE_OPS_FL_DYNAMIC | FTRACE_OPS_FL_CONTROL)) {
		schedule_on_each_cpu(ftrace_sync);

		if (ops->flags & FTRACE_OPS_FL_CONTROL)
			control_ops_free(ops);
	}

2486
	return 0;
2487 2488
}

I
Ingo Molnar 已提交
2489
static void ftrace_startup_sysctl(void)
2490
{
2491 2492 2493
	if (unlikely(ftrace_disabled))
		return;

2494 2495
	/* Force update next time */
	saved_ftrace_func = NULL;
2496 2497
	/* ftrace_start_up is true if we want ftrace running */
	if (ftrace_start_up)
2498
		ftrace_run_update_code(FTRACE_UPDATE_CALLS);
2499 2500
}

I
Ingo Molnar 已提交
2501
static void ftrace_shutdown_sysctl(void)
2502
{
2503 2504 2505
	if (unlikely(ftrace_disabled))
		return;

2506 2507
	/* ftrace_start_up is true if ftrace is running */
	if (ftrace_start_up)
2508
		ftrace_run_update_code(FTRACE_DISABLE_CALLS);
2509 2510
}

2511 2512 2513
static cycle_t		ftrace_update_time;
unsigned long		ftrace_update_tot_cnt;

2514
static inline int ops_traces_mod(struct ftrace_ops *ops)
2515
{
2516 2517 2518 2519
	/*
	 * Filter_hash being empty will default to trace module.
	 * But notrace hash requires a test of individual module functions.
	 */
2520 2521
	return ftrace_hash_empty(ops->func_hash->filter_hash) &&
		ftrace_hash_empty(ops->func_hash->notrace_hash);
2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542
}

/*
 * Check if the current ops references the record.
 *
 * If the ops traces all functions, then it was already accounted for.
 * If the ops does not trace the current record function, skip it.
 * If the ops ignores the function via notrace filter, skip it.
 */
static inline bool
ops_references_rec(struct ftrace_ops *ops, struct dyn_ftrace *rec)
{
	/* If ops isn't enabled, ignore it */
	if (!(ops->flags & FTRACE_OPS_FL_ENABLED))
		return 0;

	/* If ops traces all mods, we already accounted for it */
	if (ops_traces_mod(ops))
		return 0;

	/* The function must be in the filter */
2543 2544
	if (!ftrace_hash_empty(ops->func_hash->filter_hash) &&
	    !ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip))
2545
		return 0;
2546

2547
	/* If in notrace hash, we ignore it too */
2548
	if (ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip))
2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564
		return 0;

	return 1;
}

static int referenced_filters(struct dyn_ftrace *rec)
{
	struct ftrace_ops *ops;
	int cnt = 0;

	for (ops = ftrace_ops_list; ops != &ftrace_list_end; ops = ops->next) {
		if (ops_references_rec(ops, rec))
		    cnt++;
	}

	return cnt;
2565 2566
}

2567
static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
2568
{
2569
	struct ftrace_page *pg;
2570
	struct dyn_ftrace *p;
2571
	cycle_t start, stop;
2572
	unsigned long update_cnt = 0;
2573
	unsigned long ref = 0;
2574
	bool test = false;
2575
	int i;
2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587

	/*
	 * 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) {
2588 2589 2590 2591 2592 2593
			if (ops->flags & FTRACE_OPS_FL_ENABLED) {
				if (ops_traces_mod(ops))
					ref++;
				else
					test = true;
			}
2594 2595
		}
	}
2596

I
Ingo Molnar 已提交
2597
	start = ftrace_now(raw_smp_processor_id());
2598

2599
	for (pg = new_pgs; pg; pg = pg->next) {
2600

2601
		for (i = 0; i < pg->index; i++) {
2602 2603
			int cnt = ref;

2604 2605 2606
			/* If something went wrong, bail without enabling anything */
			if (unlikely(ftrace_disabled))
				return -1;
2607

2608
			p = &pg->records[i];
2609 2610 2611
			if (test)
				cnt += referenced_filters(p);
			p->flags = cnt;
2612

2613 2614 2615 2616 2617 2618
			/*
			 * Do the initial record conversion from mcount jump
			 * to the NOP instructions.
			 */
			if (!ftrace_code_disable(mod, p))
				break;
2619

2620
			update_cnt++;
2621

2622 2623 2624 2625 2626 2627 2628 2629 2630
			/*
			 * 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.
			 */
2631
			if (ftrace_start_up && cnt) {
2632 2633 2634 2635
				int failed = __ftrace_replace_code(p, 1);
				if (failed)
					ftrace_bug(failed, p->ip);
			}
2636
		}
2637 2638
	}

I
Ingo Molnar 已提交
2639
	stop = ftrace_now(raw_smp_processor_id());
2640
	ftrace_update_time = stop - start;
2641
	ftrace_update_tot_cnt += update_cnt;
2642

2643 2644 2645
	return 0;
}

2646
static int ftrace_allocate_records(struct ftrace_page *pg, int count)
2647
{
2648
	int order;
2649 2650
	int cnt;

2651 2652 2653 2654
	if (WARN_ON(!count))
		return -EINVAL;

	order = get_count_order(DIV_ROUND_UP(count, ENTRIES_PER_PAGE));
2655 2656

	/*
2657 2658
	 * We want to fill as much as possible. No more than a page
	 * may be empty.
2659
	 */
2660 2661
	while ((PAGE_SIZE << order) / ENTRY_SIZE >= count + ENTRIES_PER_PAGE)
		order--;
2662

2663 2664
 again:
	pg->records = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
2665

2666 2667 2668 2669 2670 2671 2672
	if (!pg->records) {
		/* if we can't allocate this size, try something smaller */
		if (!order)
			return -ENOMEM;
		order >>= 1;
		goto again;
	}
2673

2674 2675
	cnt = (PAGE_SIZE << order) / ENTRY_SIZE;
	pg->size = cnt;
2676

2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709
	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)
2710 2711
			break;

2712 2713 2714 2715
		pg->next = kzalloc(sizeof(*pg), GFP_KERNEL);
		if (!pg->next)
			goto free_pages;

2716 2717 2718
		pg = pg->next;
	}

2719 2720 2721
	return start_pg;

 free_pages:
2722 2723
	pg = start_pg;
	while (pg) {
2724 2725 2726 2727 2728 2729 2730 2731 2732 2733
		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;
}

2734 2735 2736
#define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */

struct ftrace_iterator {
2737
	loff_t				pos;
2738 2739 2740 2741 2742
	loff_t				func_pos;
	struct ftrace_page		*pg;
	struct dyn_ftrace		*func;
	struct ftrace_func_probe	*probe;
	struct trace_parser		parser;
2743
	struct ftrace_hash		*hash;
2744
	struct ftrace_ops		*ops;
2745 2746 2747
	int				hidx;
	int				idx;
	unsigned			flags;
2748 2749
};

2750
static void *
2751
t_hash_next(struct seq_file *m, loff_t *pos)
2752 2753
{
	struct ftrace_iterator *iter = m->private;
2754
	struct hlist_node *hnd = NULL;
2755 2756 2757
	struct hlist_head *hhd;

	(*pos)++;
2758
	iter->pos = *pos;
2759

2760 2761
	if (iter->probe)
		hnd = &iter->probe->node;
2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783
 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;
		}
	}

2784 2785 2786 2787 2788 2789
	if (WARN_ON_ONCE(!hnd))
		return NULL;

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

	return iter;
2790 2791 2792 2793 2794 2795
}

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
2796 2797
	loff_t l;

2798 2799 2800
	if (!(iter->flags & FTRACE_ITER_DO_HASH))
		return NULL;

2801 2802
	if (iter->func_pos > *pos)
		return NULL;
2803

L
Li Zefan 已提交
2804
	iter->hidx = 0;
2805
	for (l = 0; l <= (*pos - iter->func_pos); ) {
2806
		p = t_hash_next(m, &l);
L
Li Zefan 已提交
2807 2808 2809
		if (!p)
			break;
	}
2810 2811 2812
	if (!p)
		return NULL;

2813 2814 2815
	/* Only set this if we have an item */
	iter->flags |= FTRACE_ITER_HASH;

2816
	return iter;
2817 2818
}

2819 2820
static int
t_hash_show(struct seq_file *m, struct ftrace_iterator *iter)
2821
{
S
Steven Rostedt 已提交
2822
	struct ftrace_func_probe *rec;
2823

2824 2825 2826
	rec = iter->probe;
	if (WARN_ON_ONCE(!rec))
		return -EIO;
2827

2828 2829 2830
	if (rec->ops->print)
		return rec->ops->print(m, rec->ip, rec->ops, rec->data);

2831
	seq_printf(m, "%ps:%ps", (void *)rec->ip, (void *)rec->ops->func);
2832 2833 2834 2835 2836 2837 2838 2839

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

	return 0;
}

I
Ingo Molnar 已提交
2840
static void *
2841 2842 2843
t_next(struct seq_file *m, void *v, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
2844
	struct ftrace_ops *ops = iter->ops;
2845 2846
	struct dyn_ftrace *rec = NULL;

2847 2848 2849
	if (unlikely(ftrace_disabled))
		return NULL;

2850
	if (iter->flags & FTRACE_ITER_HASH)
2851
		return t_hash_next(m, pos);
2852

2853
	(*pos)++;
2854
	iter->pos = iter->func_pos = *pos;
2855

2856
	if (iter->flags & FTRACE_ITER_PRINTALL)
2857
		return t_hash_start(m, pos);
2858

2859 2860 2861 2862 2863 2864 2865 2866 2867
 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++];
2868
		if (((iter->flags & FTRACE_ITER_FILTER) &&
2869
		     !(ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip))) ||
S
Steven Rostedt 已提交
2870

2871
		    ((iter->flags & FTRACE_ITER_NOTRACE) &&
2872
		     !ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip)) ||
2873 2874

		    ((iter->flags & FTRACE_ITER_ENABLED) &&
2875
		     !(rec->flags & FTRACE_FL_ENABLED))) {
2876

2877 2878 2879 2880 2881
			rec = NULL;
			goto retry;
		}
	}

2882
	if (!rec)
2883
		return t_hash_start(m, pos);
2884 2885 2886 2887

	iter->func = rec;

	return iter;
2888 2889
}

2890 2891 2892 2893
static void reset_iter_read(struct ftrace_iterator *iter)
{
	iter->pos = 0;
	iter->func_pos = 0;
2894
	iter->flags &= ~(FTRACE_ITER_PRINTALL | FTRACE_ITER_HASH);
2895 2896 2897 2898 2899
}

static void *t_start(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
2900
	struct ftrace_ops *ops = iter->ops;
2901
	void *p = NULL;
2902
	loff_t l;
2903

2904
	mutex_lock(&ftrace_lock);
2905 2906 2907 2908

	if (unlikely(ftrace_disabled))
		return NULL;

2909 2910 2911 2912 2913 2914
	/*
	 * If an lseek was done, then reset and start from beginning.
	 */
	if (*pos < iter->pos)
		reset_iter_read(iter);

2915 2916 2917 2918 2919
	/*
	 * For set_ftrace_filter reading, if we have the filter
	 * off, we can short cut and just print out that all
	 * functions are enabled.
	 */
2920
	if ((iter->flags & FTRACE_ITER_FILTER &&
2921
	     ftrace_hash_empty(ops->func_hash->filter_hash)) ||
2922
	    (iter->flags & FTRACE_ITER_NOTRACE &&
2923
	     ftrace_hash_empty(ops->func_hash->notrace_hash))) {
2924
		if (*pos > 0)
2925
			return t_hash_start(m, pos);
2926
		iter->flags |= FTRACE_ITER_PRINTALL;
2927 2928
		/* reset in case of seek/pread */
		iter->flags &= ~FTRACE_ITER_HASH;
2929 2930 2931
		return iter;
	}

2932 2933 2934
	if (iter->flags & FTRACE_ITER_HASH)
		return t_hash_start(m, pos);

2935 2936 2937 2938 2939
	/*
	 * 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.
	 */
2940 2941 2942 2943 2944 2945
	iter->pg = ftrace_pages_start;
	iter->idx = 0;
	for (l = 0; l <= *pos; ) {
		p = t_next(m, p, &l);
		if (!p)
			break;
2946
	}
2947

2948 2949
	if (!p)
		return t_hash_start(m, pos);
2950 2951

	return iter;
2952 2953 2954 2955
}

static void t_stop(struct seq_file *m, void *p)
{
2956
	mutex_unlock(&ftrace_lock);
2957 2958 2959 2960
}

static int t_show(struct seq_file *m, void *v)
{
2961
	struct ftrace_iterator *iter = m->private;
2962
	struct dyn_ftrace *rec;
2963

2964
	if (iter->flags & FTRACE_ITER_HASH)
2965
		return t_hash_show(m, iter);
2966

2967
	if (iter->flags & FTRACE_ITER_PRINTALL) {
2968 2969 2970 2971
		if (iter->flags & FTRACE_ITER_NOTRACE)
			seq_printf(m, "#### no functions disabled ####\n");
		else
			seq_printf(m, "#### all functions enabled ####\n");
2972 2973 2974
		return 0;
	}

2975 2976
	rec = iter->func;

2977 2978 2979
	if (!rec)
		return 0;

2980
	seq_printf(m, "%ps", (void *)rec->ip);
2981
	if (iter->flags & FTRACE_ITER_ENABLED) {
2982
		seq_printf(m, " (%ld)%s",
2983
			   ftrace_rec_count(rec),
2984 2985 2986 2987
			   rec->flags & FTRACE_FL_REGS ? " R" : "  ");
		if (rec->flags & FTRACE_FL_TRAMP_EN) {
			struct ftrace_ops *ops;

2988
			ops = ftrace_find_tramp_ops_any(rec);
2989 2990 2991 2992 2993 2994 2995 2996
			if (ops && ops->trampoline)
				seq_printf(m, "\ttramp: %pS",
					   (void *)ops->trampoline);
			else
				seq_printf(m, "\ttramp: ERROR!");
		}
	}	

2997
	seq_printf(m, "\n");
2998 2999 3000 3001

	return 0;
}

J
James Morris 已提交
3002
static const struct seq_operations show_ftrace_seq_ops = {
3003 3004 3005 3006 3007 3008
	.start = t_start,
	.next = t_next,
	.stop = t_stop,
	.show = t_show,
};

I
Ingo Molnar 已提交
3009
static int
3010 3011 3012 3013
ftrace_avail_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3014 3015 3016
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3017 3018 3019 3020
	iter = __seq_open_private(file, &show_ftrace_seq_ops, sizeof(*iter));
	if (iter) {
		iter->pg = ftrace_pages_start;
		iter->ops = &global_ops;
I
Ingo Molnar 已提交
3021
	}
3022

3023
	return iter ? 0 : -ENOMEM;
3024 3025
}

3026 3027 3028 3029 3030 3031 3032 3033
static int
ftrace_enabled_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

	if (unlikely(ftrace_disabled))
		return -ENODEV;

3034 3035 3036 3037 3038
	iter = __seq_open_private(file, &show_ftrace_seq_ops, sizeof(*iter));
	if (iter) {
		iter->pg = ftrace_pages_start;
		iter->flags = FTRACE_ITER_ENABLED;
		iter->ops = &global_ops;
3039 3040
	}

3041
	return iter ? 0 : -ENOMEM;
3042 3043
}

3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056
/**
 * 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.
3057
 * tracing_lseek() should be used as the lseek routine, and
3058 3059 3060
 * release must call ftrace_regex_release().
 */
int
3061
ftrace_regex_open(struct ftrace_ops *ops, int flag,
3062
		  struct inode *inode, struct file *file)
3063 3064
{
	struct ftrace_iterator *iter;
3065
	struct ftrace_hash *hash;
3066 3067
	int ret = 0;

3068 3069
	ftrace_ops_init(ops);

3070 3071 3072
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3073 3074 3075 3076
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter)
		return -ENOMEM;

3077 3078 3079 3080 3081
	if (trace_parser_get_init(&iter->parser, FTRACE_BUFF_MAX)) {
		kfree(iter);
		return -ENOMEM;
	}

3082 3083 3084
	iter->ops = ops;
	iter->flags = flag;

3085
	mutex_lock(&ops->func_hash->regex_lock);
3086

3087
	if (flag & FTRACE_ITER_NOTRACE)
3088
		hash = ops->func_hash->notrace_hash;
3089
	else
3090
		hash = ops->func_hash->filter_hash;
3091

3092
	if (file->f_mode & FMODE_WRITE) {
3093 3094 3095 3096 3097 3098 3099
		const int size_bits = FTRACE_HASH_DEFAULT_BITS;

		if (file->f_flags & O_TRUNC)
			iter->hash = alloc_ftrace_hash(size_bits);
		else
			iter->hash = alloc_and_copy_ftrace_hash(size_bits, hash);

3100 3101 3102
		if (!iter->hash) {
			trace_parser_put(&iter->parser);
			kfree(iter);
3103 3104
			ret = -ENOMEM;
			goto out_unlock;
3105 3106
		}
	}
3107

3108 3109 3110 3111 3112 3113 3114
	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;
3115
		} else {
3116 3117
			/* Failed */
			free_ftrace_hash(iter->hash);
3118
			trace_parser_put(&iter->parser);
3119
			kfree(iter);
3120
		}
3121 3122
	} else
		file->private_data = iter;
3123 3124

 out_unlock:
3125
	mutex_unlock(&ops->func_hash->regex_lock);
3126 3127 3128 3129

	return ret;
}

3130 3131 3132
static int
ftrace_filter_open(struct inode *inode, struct file *file)
{
3133 3134 3135
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops,
3136 3137
			FTRACE_ITER_FILTER | FTRACE_ITER_DO_HASH,
			inode, file);
3138 3139 3140 3141 3142
}

static int
ftrace_notrace_open(struct inode *inode, struct file *file)
{
3143 3144 3145
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops, FTRACE_ITER_NOTRACE,
3146
				 inode, file);
3147 3148
}

3149
static int ftrace_match(char *str, char *regex, int len, int type)
3150 3151
{
	int matched = 0;
3152
	int slen;
3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167

	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:
3168 3169
		slen = strlen(str);
		if (slen >= len && memcmp(str + slen - len, regex, len) == 0)
3170 3171 3172 3173 3174 3175 3176
			matched = 1;
		break;
	}

	return matched;
}

3177
static int
3178
enter_record(struct ftrace_hash *hash, struct dyn_ftrace *rec, int not)
3179
{
3180 3181 3182
	struct ftrace_func_entry *entry;
	int ret = 0;

3183 3184 3185 3186 3187
	entry = ftrace_lookup_ip(hash, rec->ip);
	if (not) {
		/* Do nothing if it doesn't exist */
		if (!entry)
			return 0;
3188

3189
		free_hash_entry(hash, entry);
3190 3191 3192 3193
	} else {
		/* Do nothing if it exists */
		if (entry)
			return 0;
3194

3195
		ret = add_hash_entry(hash, rec->ip);
3196 3197
	}
	return ret;
3198 3199
}

3200
static int
3201 3202
ftrace_match_record(struct dyn_ftrace *rec, char *mod,
		    char *regex, int len, int type)
3203 3204
{
	char str[KSYM_SYMBOL_LEN];
3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217
	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;
	}
3218 3219 3220 3221

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

3222 3223 3224
static int
match_records(struct ftrace_hash *hash, char *buff,
	      int len, char *mod, int not)
3225
{
3226
	unsigned search_len = 0;
3227 3228
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
3229 3230
	int type = MATCH_FULL;
	char *search = buff;
3231
	int found = 0;
3232
	int ret;
3233

3234 3235 3236 3237
	if (len) {
		type = filter_parse_regex(buff, len, &search, &not);
		search_len = strlen(search);
	}
3238

3239
	mutex_lock(&ftrace_lock);
3240

3241 3242
	if (unlikely(ftrace_disabled))
		goto out_unlock;
3243

3244
	do_for_each_ftrace_rec(pg, rec) {
3245
		if (ftrace_match_record(rec, mod, search, search_len, type)) {
3246
			ret = enter_record(hash, rec, not);
3247 3248 3249 3250
			if (ret < 0) {
				found = ret;
				goto out_unlock;
			}
3251
			found = 1;
3252 3253
		}
	} while_for_each_ftrace_rec();
3254
 out_unlock:
3255
	mutex_unlock(&ftrace_lock);
3256 3257

	return found;
3258 3259
}

3260
static int
3261
ftrace_match_records(struct ftrace_hash *hash, char *buff, int len)
3262
{
3263
	return match_records(hash, buff, len, NULL, 0);
3264 3265
}

3266 3267
static int
ftrace_match_module_records(struct ftrace_hash *hash, char *buff, char *mod)
3268 3269
{
	int not = 0;
S
Steven Rostedt 已提交
3270

3271 3272 3273 3274 3275 3276 3277 3278 3279 3280
	/* 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;
	}

3281
	return match_records(hash, buff, strlen(buff), mod, not);
3282 3283
}

3284 3285 3286 3287 3288 3289
/*
 * We register the module command as a template to show others how
 * to register the a command as well.
 */

static int
3290 3291
ftrace_mod_callback(struct ftrace_hash *hash,
		    char *func, char *cmd, char *param, int enable)
3292 3293
{
	char *mod;
3294
	int ret = -EINVAL;
3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305

	/*
	 * 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)
3306
		return ret;
3307 3308 3309

	mod = strsep(&param, ":");
	if (!strlen(mod))
3310
		return ret;
3311

3312
	ret = ftrace_match_module_records(hash, func, mod);
3313 3314 3315 3316 3317 3318
	if (!ret)
		ret = -EINVAL;
	if (ret < 0)
		return ret;

	return 0;
3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329
}

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);
}
3330
core_initcall(ftrace_mod_cmd_init);
3331

3332
static void function_trace_probe_call(unsigned long ip, unsigned long parent_ip,
3333
				      struct ftrace_ops *op, struct pt_regs *pt_regs)
3334
{
S
Steven Rostedt 已提交
3335
	struct ftrace_func_probe *entry;
3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350
	struct hlist_head *hhd;
	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.
	 */
3351
	preempt_disable_notrace();
3352
	hlist_for_each_entry_rcu_notrace(entry, hhd, node) {
3353 3354 3355
		if (entry->ip == ip)
			entry->ops->func(ip, parent_ip, &entry->data);
	}
3356
	preempt_enable_notrace();
3357 3358
}

S
Steven Rostedt 已提交
3359
static struct ftrace_ops trace_probe_ops __read_mostly =
3360
{
3361
	.func		= function_trace_probe_call,
3362
	.flags		= FTRACE_OPS_FL_INITIALIZED,
3363
	INIT_OPS_HASH(trace_probe_ops)
3364 3365
};

S
Steven Rostedt 已提交
3366
static int ftrace_probe_registered;
3367

S
Steven Rostedt 已提交
3368
static void __enable_ftrace_function_probe(void)
3369
{
3370
	int ret;
3371 3372
	int i;

3373 3374 3375 3376
	if (ftrace_probe_registered) {
		/* still need to update the function call sites */
		if (ftrace_enabled)
			ftrace_run_update_code(FTRACE_UPDATE_CALLS);
3377
		return;
3378
	}
3379 3380 3381 3382 3383 3384 3385 3386 3387 3388

	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;

3389
	ret = ftrace_startup(&trace_probe_ops, 0);
3390

S
Steven Rostedt 已提交
3391
	ftrace_probe_registered = 1;
3392 3393
}

S
Steven Rostedt 已提交
3394
static void __disable_ftrace_function_probe(void)
3395 3396 3397
{
	int i;

S
Steven Rostedt 已提交
3398
	if (!ftrace_probe_registered)
3399 3400 3401 3402 3403 3404 3405 3406 3407
		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 */
3408
	ftrace_shutdown(&trace_probe_ops, 0);
3409

S
Steven Rostedt 已提交
3410
	ftrace_probe_registered = 0;
3411 3412 3413
}


3414
static void ftrace_free_entry(struct ftrace_func_probe *entry)
3415 3416
{
	if (entry->ops->free)
3417
		entry->ops->free(entry->ops, entry->ip, &entry->data);
3418 3419 3420 3421
	kfree(entry);
}

int
S
Steven Rostedt 已提交
3422
register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3423 3424
			      void *data)
{
S
Steven Rostedt 已提交
3425
	struct ftrace_func_probe *entry;
3426
	struct ftrace_hash **orig_hash = &trace_probe_ops.func_hash->filter_hash;
3427
	struct ftrace_hash *old_hash = *orig_hash;
3428
	struct ftrace_hash *hash;
3429 3430 3431
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
	int type, len, not;
S
Steven Rostedt 已提交
3432
	unsigned long key;
3433 3434
	int count = 0;
	char *search;
3435
	int ret;
3436

3437
	type = filter_parse_regex(glob, strlen(glob), &search, &not);
3438 3439
	len = strlen(search);

S
Steven Rostedt 已提交
3440
	/* we do not support '!' for function probes */
3441 3442 3443
	if (WARN_ON(not))
		return -EINVAL;

3444
	mutex_lock(&trace_probe_ops.func_hash->regex_lock);
3445

3446
	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, old_hash);
3447 3448
	if (!hash) {
		count = -ENOMEM;
3449
		goto out;
3450 3451 3452 3453
	}

	if (unlikely(ftrace_disabled)) {
		count = -ENODEV;
3454
		goto out;
3455
	}
3456

3457 3458
	mutex_lock(&ftrace_lock);

3459
	do_for_each_ftrace_rec(pg, rec) {
3460

3461
		if (!ftrace_match_record(rec, NULL, search, len, type))
3462 3463 3464 3465
			continue;

		entry = kmalloc(sizeof(*entry), GFP_KERNEL);
		if (!entry) {
S
Steven Rostedt 已提交
3466
			/* If we did not process any, then return error */
3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480
			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.
		 */
3481 3482
		if (ops->init) {
			if (ops->init(ops, rec->ip, &entry->data) < 0) {
3483 3484 3485 3486 3487 3488
				/* caller does not like this func */
				kfree(entry);
				continue;
			}
		}

3489 3490 3491 3492 3493 3494 3495
		ret = enter_record(hash, rec, 0);
		if (ret < 0) {
			kfree(entry);
			count = ret;
			goto out_unlock;
		}

3496 3497 3498 3499 3500 3501 3502
		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();
3503 3504

	ret = ftrace_hash_move(&trace_probe_ops, 1, orig_hash, hash);
3505 3506 3507
	if (!ret)
		free_ftrace_hash_rcu(old_hash);
	else
3508 3509
		count = ret;

S
Steven Rostedt 已提交
3510
	__enable_ftrace_function_probe();
3511 3512

 out_unlock:
3513 3514
	mutex_unlock(&ftrace_lock);
 out:
3515
	mutex_unlock(&trace_probe_ops.func_hash->regex_lock);
3516
	free_ftrace_hash(hash);
3517 3518 3519 3520 3521

	return count;
}

enum {
S
Steven Rostedt 已提交
3522 3523
	PROBE_TEST_FUNC		= 1,
	PROBE_TEST_DATA		= 2
3524 3525 3526
};

static void
S
Steven Rostedt 已提交
3527
__unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3528 3529
				  void *data, int flags)
{
3530
	struct ftrace_func_entry *rec_entry;
S
Steven Rostedt 已提交
3531
	struct ftrace_func_probe *entry;
3532
	struct ftrace_func_probe *p;
3533
	struct ftrace_hash **orig_hash = &trace_probe_ops.func_hash->filter_hash;
3534
	struct ftrace_hash *old_hash = *orig_hash;
3535
	struct list_head free_list;
3536
	struct ftrace_hash *hash;
3537
	struct hlist_node *tmp;
3538 3539 3540 3541
	char str[KSYM_SYMBOL_LEN];
	int type = MATCH_FULL;
	int i, len = 0;
	char *search;
3542
	int ret;
3543

3544
	if (glob && (strcmp(glob, "*") == 0 || !strlen(glob)))
3545
		glob = NULL;
3546
	else if (glob) {
3547 3548
		int not;

3549
		type = filter_parse_regex(glob, strlen(glob), &search, &not);
3550 3551
		len = strlen(search);

S
Steven Rostedt 已提交
3552
		/* we do not support '!' for function probes */
3553 3554 3555 3556
		if (WARN_ON(not))
			return;
	}

3557
	mutex_lock(&trace_probe_ops.func_hash->regex_lock);
3558 3559 3560 3561 3562 3563

	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);
	if (!hash)
		/* Hmm, should report this somehow */
		goto out_unlock;

3564 3565
	INIT_LIST_HEAD(&free_list);

3566 3567 3568
	for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) {
		struct hlist_head *hhd = &ftrace_func_hash[i];

3569
		hlist_for_each_entry_safe(entry, tmp, hhd, node) {
3570 3571

			/* break up if statements for readability */
S
Steven Rostedt 已提交
3572
			if ((flags & PROBE_TEST_FUNC) && entry->ops != ops)
3573 3574
				continue;

S
Steven Rostedt 已提交
3575
			if ((flags & PROBE_TEST_DATA) && entry->data != data)
3576 3577 3578 3579 3580 3581 3582 3583 3584 3585
				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;
			}

3586 3587 3588 3589 3590
			rec_entry = ftrace_lookup_ip(hash, entry->ip);
			/* It is possible more than one entry had this ip */
			if (rec_entry)
				free_hash_entry(hash, rec_entry);

3591
			hlist_del_rcu(&entry->node);
3592
			list_add(&entry->free_list, &free_list);
3593 3594
		}
	}
3595
	mutex_lock(&ftrace_lock);
S
Steven Rostedt 已提交
3596
	__disable_ftrace_function_probe();
3597 3598 3599 3600
	/*
	 * Remove after the disable is called. Otherwise, if the last
	 * probe is removed, a null hash means *all enabled*.
	 */
3601
	ret = ftrace_hash_move(&trace_probe_ops, 1, orig_hash, hash);
3602
	synchronize_sched();
3603 3604 3605
	if (!ret)
		free_ftrace_hash_rcu(old_hash);

3606 3607 3608 3609
	list_for_each_entry_safe(entry, p, &free_list, free_list) {
		list_del(&entry->free_list);
		ftrace_free_entry(entry);
	}
3610
	mutex_unlock(&ftrace_lock);
3611
		
3612
 out_unlock:
3613
	mutex_unlock(&trace_probe_ops.func_hash->regex_lock);
3614
	free_ftrace_hash(hash);
3615 3616 3617
}

void
S
Steven Rostedt 已提交
3618
unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3619 3620
				void *data)
{
S
Steven Rostedt 已提交
3621 3622
	__unregister_ftrace_function_probe(glob, ops, data,
					  PROBE_TEST_FUNC | PROBE_TEST_DATA);
3623 3624 3625
}

void
S
Steven Rostedt 已提交
3626
unregister_ftrace_function_probe_func(char *glob, struct ftrace_probe_ops *ops)
3627
{
S
Steven Rostedt 已提交
3628
	__unregister_ftrace_function_probe(glob, ops, NULL, PROBE_TEST_FUNC);
3629 3630
}

S
Steven Rostedt 已提交
3631
void unregister_ftrace_function_probe_all(char *glob)
3632
{
S
Steven Rostedt 已提交
3633
	__unregister_ftrace_function_probe(glob, NULL, NULL, 0);
3634 3635
}

3636 3637 3638
static LIST_HEAD(ftrace_commands);
static DEFINE_MUTEX(ftrace_cmd_mutex);

3639 3640 3641 3642 3643
/*
 * Currently we only register ftrace commands from __init, so mark this
 * __init too.
 */
__init int register_ftrace_command(struct ftrace_func_command *cmd)
3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661
{
	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;
}

3662 3663 3664 3665 3666
/*
 * Currently we only unregister ftrace commands from __init, so mark
 * this __init too.
 */
__init int unregister_ftrace_command(struct ftrace_func_command *cmd)
3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684
{
	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;
}

3685 3686
static int ftrace_process_regex(struct ftrace_hash *hash,
				char *buff, int len, int enable)
3687
{
3688
	char *func, *command, *next = buff;
S
Steven Rostedt 已提交
3689
	struct ftrace_func_command *p;
3690
	int ret = -EINVAL;
3691 3692 3693 3694

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

	if (!next) {
3695
		ret = ftrace_match_records(hash, func, len);
3696 3697 3698 3699 3700
		if (!ret)
			ret = -EINVAL;
		if (ret < 0)
			return ret;
		return 0;
3701 3702
	}

3703
	/* command found */
3704 3705 3706

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

3707 3708 3709
	mutex_lock(&ftrace_cmd_mutex);
	list_for_each_entry(p, &ftrace_commands, list) {
		if (strcmp(p->name, command) == 0) {
3710
			ret = p->func(hash, func, command, next, enable);
3711 3712
			goto out_unlock;
		}
3713
	}
3714 3715
 out_unlock:
	mutex_unlock(&ftrace_cmd_mutex);
3716

3717
	return ret;
3718 3719
}

I
Ingo Molnar 已提交
3720
static ssize_t
3721 3722
ftrace_regex_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos, int enable)
3723 3724
{
	struct ftrace_iterator *iter;
3725 3726
	struct trace_parser *parser;
	ssize_t ret, read;
3727

3728
	if (!cnt)
3729 3730 3731 3732 3733 3734 3735 3736
		return 0;

	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;
		iter = m->private;
	} else
		iter = file->private_data;

3737
	if (unlikely(ftrace_disabled))
3738 3739 3740
		return -ENODEV;

	/* iter->hash is a local copy, so we don't need regex_lock */
3741

3742 3743
	parser = &iter->parser;
	read = trace_get_user(parser, ubuf, cnt, ppos);
3744

3745
	if (read >= 0 && trace_parser_loaded(parser) &&
3746
	    !trace_parser_cont(parser)) {
3747
		ret = ftrace_process_regex(iter->hash, parser->buffer,
3748
					   parser->idx, enable);
3749
		trace_parser_clear(parser);
3750
		if (ret < 0)
3751
			goto out;
3752
	}
3753 3754

	ret = read;
3755
 out:
3756 3757 3758
	return ret;
}

3759
ssize_t
3760 3761 3762 3763 3764 3765
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);
}

3766
ssize_t
3767 3768 3769 3770 3771 3772
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);
}

3773
static int
3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791
ftrace_match_addr(struct ftrace_hash *hash, unsigned long ip, int remove)
{
	struct ftrace_func_entry *entry;

	if (!ftrace_location(ip))
		return -EINVAL;

	if (remove) {
		entry = ftrace_lookup_ip(hash, ip);
		if (!entry)
			return -ENOENT;
		free_hash_entry(hash, entry);
		return 0;
	}

	return add_hash_entry(hash, ip);
}

3792 3793 3794 3795 3796 3797
static void ftrace_ops_update_code(struct ftrace_ops *ops)
{
	if (ops->flags & FTRACE_OPS_FL_ENABLED && ftrace_enabled)
		ftrace_run_update_code(FTRACE_UPDATE_CALLS);
}

3798 3799 3800
static int
ftrace_set_hash(struct ftrace_ops *ops, unsigned char *buf, int len,
		unsigned long ip, int remove, int reset, int enable)
3801
{
3802
	struct ftrace_hash **orig_hash;
3803
	struct ftrace_hash *old_hash;
3804
	struct ftrace_hash *hash;
3805
	int ret;
3806

3807
	if (unlikely(ftrace_disabled))
3808
		return -ENODEV;
3809

3810
	mutex_lock(&ops->func_hash->regex_lock);
3811

3812
	if (enable)
3813
		orig_hash = &ops->func_hash->filter_hash;
3814
	else
3815
		orig_hash = &ops->func_hash->notrace_hash;
3816

3817 3818 3819 3820 3821
	if (reset)
		hash = alloc_ftrace_hash(FTRACE_HASH_DEFAULT_BITS);
	else
		hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);

3822 3823 3824 3825
	if (!hash) {
		ret = -ENOMEM;
		goto out_regex_unlock;
	}
3826

3827 3828 3829 3830
	if (buf && !ftrace_match_records(hash, buf, len)) {
		ret = -EINVAL;
		goto out_regex_unlock;
	}
3831 3832 3833 3834 3835
	if (ip) {
		ret = ftrace_match_addr(hash, ip, remove);
		if (ret < 0)
			goto out_regex_unlock;
	}
3836 3837

	mutex_lock(&ftrace_lock);
3838
	old_hash = *orig_hash;
3839
	ret = ftrace_hash_move(ops, enable, orig_hash, hash);
3840
	if (!ret) {
3841
		ftrace_ops_update_code(ops);
3842 3843
		free_ftrace_hash_rcu(old_hash);
	}
3844 3845
	mutex_unlock(&ftrace_lock);

3846
 out_regex_unlock:
3847
	mutex_unlock(&ops->func_hash->regex_lock);
3848 3849 3850

	free_ftrace_hash(hash);
	return ret;
3851 3852
}

3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872
static int
ftrace_set_addr(struct ftrace_ops *ops, unsigned long ip, int remove,
		int reset, int enable)
{
	return ftrace_set_hash(ops, 0, 0, ip, remove, reset, enable);
}

/**
 * ftrace_set_filter_ip - set a function to filter on in ftrace by address
 * @ops - the ops to set the filter with
 * @ip - the address to add to or remove from the filter.
 * @remove - non zero to remove the ip from the filter
 * @reset - non zero to reset all filters before applying this filter.
 *
 * Filters denote which functions should be enabled when tracing is enabled
 * If @ip is NULL, it failes to update filter.
 */
int ftrace_set_filter_ip(struct ftrace_ops *ops, unsigned long ip,
			 int remove, int reset)
{
3873
	ftrace_ops_init(ops);
3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884
	return ftrace_set_addr(ops, ip, remove, reset, 1);
}
EXPORT_SYMBOL_GPL(ftrace_set_filter_ip);

static int
ftrace_set_regex(struct ftrace_ops *ops, unsigned char *buf, int len,
		 int reset, int enable)
{
	return ftrace_set_hash(ops, buf, len, 0, 0, reset, enable);
}

3885 3886
/**
 * ftrace_set_filter - set a function to filter on in ftrace
3887 3888 3889 3890 3891 3892 3893 3894
 * @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.
 */
3895
int ftrace_set_filter(struct ftrace_ops *ops, unsigned char *buf,
3896 3897
		       int len, int reset)
{
3898
	ftrace_ops_init(ops);
3899
	return ftrace_set_regex(ops, buf, len, reset, 1);
3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913
}
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.
 */
3914
int ftrace_set_notrace(struct ftrace_ops *ops, unsigned char *buf,
3915 3916
			int len, int reset)
{
3917
	ftrace_ops_init(ops);
3918
	return ftrace_set_regex(ops, buf, len, reset, 0);
3919 3920 3921
}
EXPORT_SYMBOL_GPL(ftrace_set_notrace);
/**
3922
 * ftrace_set_global_filter - set a function to filter on with global tracers
3923 3924 3925 3926 3927 3928 3929
 * @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.
 */
3930
void ftrace_set_global_filter(unsigned char *buf, int len, int reset)
3931
{
3932
	ftrace_set_regex(&global_ops, buf, len, reset, 1);
3933
}
3934
EXPORT_SYMBOL_GPL(ftrace_set_global_filter);
3935

3936
/**
3937
 * ftrace_set_global_notrace - set a function to not trace with global tracers
3938 3939 3940 3941 3942 3943 3944 3945
 * @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.
 */
3946
void ftrace_set_global_notrace(unsigned char *buf, int len, int reset)
3947
{
3948
	ftrace_set_regex(&global_ops, buf, len, reset, 0);
3949
}
3950
EXPORT_SYMBOL_GPL(ftrace_set_global_notrace);
3951

3952 3953 3954 3955 3956 3957 3958
/*
 * 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;

3959 3960 3961
/* Used by function selftest to not test if filter is set */
bool ftrace_filter_param __initdata;

3962 3963
static int __init set_ftrace_notrace(char *str)
{
3964
	ftrace_filter_param = true;
3965
	strlcpy(ftrace_notrace_buf, str, FTRACE_FILTER_SIZE);
3966 3967 3968 3969 3970 3971
	return 1;
}
__setup("ftrace_notrace=", set_ftrace_notrace);

static int __init set_ftrace_filter(char *str)
{
3972
	ftrace_filter_param = true;
3973
	strlcpy(ftrace_filter_buf, str, FTRACE_FILTER_SIZE);
3974 3975 3976 3977
	return 1;
}
__setup("ftrace_filter=", set_ftrace_filter);

3978
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
3979
static char ftrace_graph_buf[FTRACE_FILTER_SIZE] __initdata;
3980
static char ftrace_graph_notrace_buf[FTRACE_FILTER_SIZE] __initdata;
3981
static int ftrace_set_func(unsigned long *array, int *idx, int size, char *buffer);
3982

3983 3984
static int __init set_graph_function(char *str)
{
3985
	strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE);
3986 3987 3988 3989
	return 1;
}
__setup("ftrace_graph_filter=", set_graph_function);

3990 3991 3992 3993 3994 3995 3996 3997
static int __init set_graph_notrace_function(char *str)
{
	strlcpy(ftrace_graph_notrace_buf, str, FTRACE_FILTER_SIZE);
	return 1;
}
__setup("ftrace_graph_notrace=", set_graph_notrace_function);

static void __init set_ftrace_early_graph(char *buf, int enable)
3998 3999 4000
{
	int ret;
	char *func;
4001 4002 4003 4004 4005 4006 4007
	unsigned long *table = ftrace_graph_funcs;
	int *count = &ftrace_graph_count;

	if (!enable) {
		table = ftrace_graph_notrace_funcs;
		count = &ftrace_graph_notrace_count;
	}
4008 4009 4010 4011

	while (buf) {
		func = strsep(&buf, ",");
		/* we allow only one expression at a time */
4012
		ret = ftrace_set_func(table, count, FTRACE_GRAPH_MAX_FUNCS, func);
4013 4014 4015 4016 4017 4018 4019
		if (ret)
			printk(KERN_DEBUG "ftrace: function %s not "
					  "traceable\n", func);
	}
}
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4020 4021
void __init
ftrace_set_early_filter(struct ftrace_ops *ops, char *buf, int enable)
4022 4023 4024
{
	char *func;

4025 4026
	ftrace_ops_init(ops);

4027 4028
	while (buf) {
		func = strsep(&buf, ",");
4029
		ftrace_set_regex(ops, func, strlen(func), 0, enable);
4030 4031 4032 4033 4034 4035
	}
}

static void __init set_ftrace_early_filters(void)
{
	if (ftrace_filter_buf[0])
4036
		ftrace_set_early_filter(&global_ops, ftrace_filter_buf, 1);
4037
	if (ftrace_notrace_buf[0])
4038
		ftrace_set_early_filter(&global_ops, ftrace_notrace_buf, 0);
4039 4040
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	if (ftrace_graph_buf[0])
4041 4042 4043
		set_ftrace_early_graph(ftrace_graph_buf, 1);
	if (ftrace_graph_notrace_buf[0])
		set_ftrace_early_graph(ftrace_graph_notrace_buf, 0);
4044
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
4045 4046
}

4047
int ftrace_regex_release(struct inode *inode, struct file *file)
4048 4049 4050
{
	struct seq_file *m = (struct seq_file *)file->private_data;
	struct ftrace_iterator *iter;
4051
	struct ftrace_hash **orig_hash;
4052
	struct ftrace_hash *old_hash;
4053
	struct trace_parser *parser;
4054
	int filter_hash;
4055
	int ret;
4056 4057 4058 4059 4060 4061 4062

	if (file->f_mode & FMODE_READ) {
		iter = m->private;
		seq_release(inode, file);
	} else
		iter = file->private_data;

4063 4064 4065
	parser = &iter->parser;
	if (trace_parser_loaded(parser)) {
		parser->buffer[parser->idx] = 0;
4066
		ftrace_match_records(iter->hash, parser->buffer, parser->idx);
4067 4068
	}

4069 4070
	trace_parser_put(parser);

4071
	mutex_lock(&iter->ops->func_hash->regex_lock);
4072

4073
	if (file->f_mode & FMODE_WRITE) {
4074 4075 4076
		filter_hash = !!(iter->flags & FTRACE_ITER_FILTER);

		if (filter_hash)
4077
			orig_hash = &iter->ops->func_hash->filter_hash;
4078
		else
4079
			orig_hash = &iter->ops->func_hash->notrace_hash;
4080

4081
		mutex_lock(&ftrace_lock);
4082
		old_hash = *orig_hash;
4083 4084
		ret = ftrace_hash_move(iter->ops, filter_hash,
				       orig_hash, iter->hash);
4085
		if (!ret) {
4086
			ftrace_ops_update_code(iter->ops);
4087 4088
			free_ftrace_hash_rcu(old_hash);
		}
4089 4090
		mutex_unlock(&ftrace_lock);
	}
4091

4092
	mutex_unlock(&iter->ops->func_hash->regex_lock);
4093 4094
	free_ftrace_hash(iter->hash);
	kfree(iter);
4095

4096 4097 4098
	return 0;
}

4099
static const struct file_operations ftrace_avail_fops = {
4100 4101 4102
	.open = ftrace_avail_open,
	.read = seq_read,
	.llseek = seq_lseek,
L
Li Zefan 已提交
4103
	.release = seq_release_private,
4104 4105
};

4106 4107 4108 4109 4110 4111 4112
static const struct file_operations ftrace_enabled_fops = {
	.open = ftrace_enabled_open,
	.read = seq_read,
	.llseek = seq_lseek,
	.release = seq_release_private,
};

4113
static const struct file_operations ftrace_filter_fops = {
4114
	.open = ftrace_filter_open,
L
Lai Jiangshan 已提交
4115
	.read = seq_read,
4116
	.write = ftrace_filter_write,
4117
	.llseek = tracing_lseek,
4118
	.release = ftrace_regex_release,
4119 4120
};

4121
static const struct file_operations ftrace_notrace_fops = {
4122
	.open = ftrace_notrace_open,
L
Lai Jiangshan 已提交
4123
	.read = seq_read,
4124
	.write = ftrace_notrace_write,
4125
	.llseek = tracing_lseek,
4126
	.release = ftrace_regex_release,
4127 4128
};

4129 4130 4131 4132 4133
#ifdef CONFIG_FUNCTION_GRAPH_TRACER

static DEFINE_MUTEX(graph_lock);

int ftrace_graph_count;
4134
int ftrace_graph_notrace_count;
4135
unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly;
4136
unsigned long ftrace_graph_notrace_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly;
4137

4138 4139 4140 4141 4142 4143 4144
struct ftrace_graph_data {
	unsigned long *table;
	size_t size;
	int *count;
	const struct seq_operations *seq_ops;
};

4145
static void *
4146
__g_next(struct seq_file *m, loff_t *pos)
4147
{
4148 4149 4150
	struct ftrace_graph_data *fgd = m->private;

	if (*pos >= *fgd->count)
4151
		return NULL;
4152
	return &fgd->table[*pos];
4153
}
4154

4155 4156 4157 4158 4159
static void *
g_next(struct seq_file *m, void *v, loff_t *pos)
{
	(*pos)++;
	return __g_next(m, pos);
4160 4161 4162 4163
}

static void *g_start(struct seq_file *m, loff_t *pos)
{
4164 4165
	struct ftrace_graph_data *fgd = m->private;

4166 4167
	mutex_lock(&graph_lock);

4168
	/* Nothing, tell g_show to print all functions are enabled */
4169
	if (!*fgd->count && !*pos)
4170 4171
		return (void *)1;

4172
	return __g_next(m, pos);
4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186
}

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;

4187
	if (ptr == (unsigned long *)1) {
4188 4189 4190 4191 4192 4193
		struct ftrace_graph_data *fgd = m->private;

		if (fgd->table == ftrace_graph_funcs)
			seq_printf(m, "#### all functions enabled ####\n");
		else
			seq_printf(m, "#### no functions disabled ####\n");
4194 4195 4196
		return 0;
	}

4197
	seq_printf(m, "%ps\n", (void *)*ptr);
4198 4199 4200 4201

	return 0;
}

J
James Morris 已提交
4202
static const struct seq_operations ftrace_graph_seq_ops = {
4203 4204 4205 4206 4207 4208 4209
	.start = g_start,
	.next = g_next,
	.stop = g_stop,
	.show = g_show,
};

static int
4210 4211
__ftrace_graph_open(struct inode *inode, struct file *file,
		    struct ftrace_graph_data *fgd)
4212 4213 4214 4215 4216
{
	int ret = 0;

	mutex_lock(&graph_lock);
	if ((file->f_mode & FMODE_WRITE) &&
4217
	    (file->f_flags & O_TRUNC)) {
4218 4219
		*fgd->count = 0;
		memset(fgd->table, 0, fgd->size * sizeof(*fgd->table));
4220
	}
4221
	mutex_unlock(&graph_lock);
4222

4223 4224 4225 4226 4227 4228 4229 4230
	if (file->f_mode & FMODE_READ) {
		ret = seq_open(file, fgd->seq_ops);
		if (!ret) {
			struct seq_file *m = file->private_data;
			m->private = fgd;
		}
	} else
		file->private_data = fgd;
4231 4232 4233 4234

	return ret;
}

4235 4236 4237 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 4253 4254
static int
ftrace_graph_open(struct inode *inode, struct file *file)
{
	struct ftrace_graph_data *fgd;

	if (unlikely(ftrace_disabled))
		return -ENODEV;

	fgd = kmalloc(sizeof(*fgd), GFP_KERNEL);
	if (fgd == NULL)
		return -ENOMEM;

	fgd->table = ftrace_graph_funcs;
	fgd->size = FTRACE_GRAPH_MAX_FUNCS;
	fgd->count = &ftrace_graph_count;
	fgd->seq_ops = &ftrace_graph_seq_ops;

	return __ftrace_graph_open(inode, file, fgd);
}

4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274
static int
ftrace_graph_notrace_open(struct inode *inode, struct file *file)
{
	struct ftrace_graph_data *fgd;

	if (unlikely(ftrace_disabled))
		return -ENODEV;

	fgd = kmalloc(sizeof(*fgd), GFP_KERNEL);
	if (fgd == NULL)
		return -ENOMEM;

	fgd->table = ftrace_graph_notrace_funcs;
	fgd->size = FTRACE_GRAPH_MAX_FUNCS;
	fgd->count = &ftrace_graph_notrace_count;
	fgd->seq_ops = &ftrace_graph_seq_ops;

	return __ftrace_graph_open(inode, file, fgd);
}

4275 4276 4277
static int
ftrace_graph_release(struct inode *inode, struct file *file)
{
4278 4279 4280 4281
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;

		kfree(m->private);
4282
		seq_release(inode, file);
4283 4284 4285 4286
	} else {
		kfree(file->private_data);
	}

4287 4288 4289
	return 0;
}

4290
static int
4291
ftrace_set_func(unsigned long *array, int *idx, int size, char *buffer)
4292 4293 4294
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
4295
	int search_len;
4296
	int fail = 1;
4297 4298 4299 4300
	int type, not;
	char *search;
	bool exists;
	int i;
4301

4302
	/* decode regex */
4303
	type = filter_parse_regex(buffer, strlen(buffer), &search, &not);
4304
	if (!not && *idx >= size)
4305
		return -EBUSY;
4306 4307 4308

	search_len = strlen(search);

4309
	mutex_lock(&ftrace_lock);
4310 4311 4312 4313 4314 4315

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

4316 4317
	do_for_each_ftrace_rec(pg, rec) {

4318
		if (ftrace_match_record(rec, NULL, search, search_len, type)) {
4319
			/* if it is in the array */
4320
			exists = false;
4321
			for (i = 0; i < *idx; i++) {
4322 4323
				if (array[i] == rec->ip) {
					exists = true;
4324 4325
					break;
				}
4326 4327 4328 4329 4330 4331
			}

			if (!not) {
				fail = 0;
				if (!exists) {
					array[(*idx)++] = rec->ip;
4332
					if (*idx >= size)
4333 4334 4335 4336 4337 4338 4339 4340 4341
						goto out;
				}
			} else {
				if (exists) {
					array[i] = array[--(*idx)];
					array[*idx] = 0;
					fail = 0;
				}
			}
4342
		}
4343
	} while_for_each_ftrace_rec();
4344
out:
4345
	mutex_unlock(&ftrace_lock);
4346

4347 4348 4349 4350
	if (fail)
		return -EINVAL;

	return 0;
4351 4352 4353 4354 4355 4356
}

static ssize_t
ftrace_graph_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
4357
	struct trace_parser parser;
4358
	ssize_t read, ret = 0;
4359
	struct ftrace_graph_data *fgd = file->private_data;
4360

4361
	if (!cnt)
4362 4363
		return 0;

4364 4365
	if (trace_parser_get_init(&parser, FTRACE_BUFF_MAX))
		return -ENOMEM;
4366

4367
	read = trace_get_user(&parser, ubuf, cnt, ppos);
4368

4369
	if (read >= 0 && trace_parser_loaded((&parser))) {
4370 4371
		parser.buffer[parser.idx] = 0;

4372 4373
		mutex_lock(&graph_lock);

4374
		/* we allow only one expression at a time */
4375 4376
		ret = ftrace_set_func(fgd->table, fgd->count, fgd->size,
				      parser.buffer);
4377 4378

		mutex_unlock(&graph_lock);
4379 4380
	}

4381 4382
	if (!ret)
		ret = read;
4383

4384
	trace_parser_put(&parser);
4385 4386 4387 4388 4389

	return ret;
}

static const struct file_operations ftrace_graph_fops = {
4390 4391 4392
	.open		= ftrace_graph_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
4393
	.llseek		= tracing_lseek,
4394
	.release	= ftrace_graph_release,
4395
};
4396 4397 4398 4399 4400

static const struct file_operations ftrace_graph_notrace_fops = {
	.open		= ftrace_graph_notrace_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
4401
	.llseek		= tracing_lseek,
4402 4403
	.release	= ftrace_graph_release,
};
4404 4405
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4406 4407 4408 4409 4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429 4430 4431 4432 4433 4434 4435
void ftrace_create_filter_files(struct ftrace_ops *ops,
				struct dentry *parent)
{

	trace_create_file("set_ftrace_filter", 0644, parent,
			  ops, &ftrace_filter_fops);

	trace_create_file("set_ftrace_notrace", 0644, parent,
			  ops, &ftrace_notrace_fops);
}

/*
 * The name "destroy_filter_files" is really a misnomer. Although
 * in the future, it may actualy delete the files, but this is
 * really intended to make sure the ops passed in are disabled
 * and that when this function returns, the caller is free to
 * free the ops.
 *
 * The "destroy" name is only to match the "create" name that this
 * should be paired with.
 */
void ftrace_destroy_filter_files(struct ftrace_ops *ops)
{
	mutex_lock(&ftrace_lock);
	if (ops->flags & FTRACE_OPS_FL_ENABLED)
		ftrace_shutdown(ops, 0);
	ops->flags |= FTRACE_OPS_FL_DELETED;
	mutex_unlock(&ftrace_lock);
}

4436
static __init int ftrace_init_dyn_debugfs(struct dentry *d_tracer)
4437 4438
{

4439 4440
	trace_create_file("available_filter_functions", 0444,
			d_tracer, NULL, &ftrace_avail_fops);
4441

4442 4443 4444
	trace_create_file("enabled_functions", 0444,
			d_tracer, NULL, &ftrace_enabled_fops);

4445
	ftrace_create_filter_files(&global_ops, d_tracer);
4446

4447
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
4448
	trace_create_file("set_graph_function", 0444, d_tracer,
4449 4450
				    NULL,
				    &ftrace_graph_fops);
4451 4452 4453
	trace_create_file("set_graph_notrace", 0444, d_tracer,
				    NULL,
				    &ftrace_graph_notrace_fops);
4454 4455
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4456 4457 4458
	return 0;
}

4459
static int ftrace_cmp_ips(const void *a, const void *b)
4460
{
4461 4462
	const unsigned long *ipa = a;
	const unsigned long *ipb = b;
4463

4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476 4477 4478 4479
	if (*ipa > *ipb)
		return 1;
	if (*ipa < *ipb)
		return -1;
	return 0;
}

static void ftrace_swap_ips(void *a, void *b, int size)
{
	unsigned long *ipa = a;
	unsigned long *ipb = b;
	unsigned long t;

	t = *ipa;
	*ipa = *ipb;
	*ipb = t;
4480 4481
}

4482
static int ftrace_process_locs(struct module *mod,
4483
			       unsigned long *start,
4484 4485
			       unsigned long *end)
{
4486
	struct ftrace_page *start_pg;
4487
	struct ftrace_page *pg;
4488
	struct dyn_ftrace *rec;
4489
	unsigned long count;
4490 4491
	unsigned long *p;
	unsigned long addr;
4492
	unsigned long flags = 0; /* Shut up gcc */
4493 4494 4495 4496 4497 4498 4499
	int ret = -ENOMEM;

	count = end - start;

	if (!count)
		return 0;

4500 4501 4502
	sort(start, count, sizeof(*start),
	     ftrace_cmp_ips, ftrace_swap_ips);

4503 4504
	start_pg = ftrace_allocate_pages(count);
	if (!start_pg)
4505
		return -ENOMEM;
4506

S
Steven Rostedt 已提交
4507
	mutex_lock(&ftrace_lock);
4508

4509 4510 4511 4512 4513
	/*
	 * 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.
	 */
4514 4515 4516
	if (!mod) {
		WARN_ON(ftrace_pages || ftrace_pages_start);
		/* First initialization */
4517
		ftrace_pages = ftrace_pages_start = start_pg;
4518
	} else {
4519
		if (!ftrace_pages)
4520
			goto out;
4521

4522 4523 4524 4525
		if (WARN_ON(ftrace_pages->next)) {
			/* Hmm, we have free pages? */
			while (ftrace_pages->next)
				ftrace_pages = ftrace_pages->next;
4526
		}
4527

4528
		ftrace_pages->next = start_pg;
4529 4530
	}

4531
	p = start;
4532
	pg = start_pg;
4533 4534
	while (p < end) {
		addr = ftrace_call_adjust(*p++);
4535 4536 4537 4538 4539 4540 4541 4542
		/*
		 * 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;
4543 4544 4545 4546 4547 4548 4549 4550 4551 4552

		if (pg->index == pg->size) {
			/* We should have allocated enough */
			if (WARN_ON(!pg->next))
				break;
			pg = pg->next;
		}

		rec = &pg->records[pg->index++];
		rec->ip = addr;
4553 4554
	}

4555 4556 4557 4558 4559 4560
	/* We should have used all pages */
	WARN_ON(pg->next);

	/* Assign the last page to ftrace_pages */
	ftrace_pages = pg;

4561
	/*
4562 4563 4564 4565 4566 4567
	 * 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.
4568
	 */
4569 4570
	if (!mod)
		local_irq_save(flags);
4571
	ftrace_update_code(mod, start_pg);
4572 4573
	if (!mod)
		local_irq_restore(flags);
4574 4575
	ret = 0;
 out:
S
Steven Rostedt 已提交
4576
	mutex_unlock(&ftrace_lock);
4577

4578
	return ret;
4579 4580
}

4581
#ifdef CONFIG_MODULES
4582 4583 4584

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

4585
void ftrace_release_mod(struct module *mod)
4586 4587
{
	struct dyn_ftrace *rec;
4588
	struct ftrace_page **last_pg;
4589
	struct ftrace_page *pg;
4590
	int order;
4591

4592 4593
	mutex_lock(&ftrace_lock);

4594
	if (ftrace_disabled)
4595
		goto out_unlock;
4596

4597 4598 4599 4600 4601 4602 4603
	/*
	 * 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];
4604
		if (within_module_core(rec->ip, mod)) {
4605
			/*
4606 4607
			 * As core pages are first, the first
			 * page should never be a module page.
4608
			 */
4609 4610 4611 4612 4613 4614 4615 4616
			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;
4617 4618 4619
			order = get_count_order(pg->size / ENTRIES_PER_PAGE);
			free_pages((unsigned long)pg->records, order);
			kfree(pg);
4620 4621 4622
		} else
			last_pg = &pg->next;
	}
4623
 out_unlock:
4624 4625 4626 4627 4628
	mutex_unlock(&ftrace_lock);
}

static void ftrace_init_module(struct module *mod,
			       unsigned long *start, unsigned long *end)
4629
{
4630
	if (ftrace_disabled || start == end)
4631
		return;
4632
	ftrace_process_locs(mod, start, end);
4633 4634
}

4635
void ftrace_module_init(struct module *mod)
4636
{
4637 4638 4639
	ftrace_init_module(mod, mod->ftrace_callsites,
			   mod->ftrace_callsites +
			   mod->num_ftrace_callsites);
4640 4641 4642 4643 4644 4645 4646 4647
}

static int ftrace_module_notify_exit(struct notifier_block *self,
				     unsigned long val, void *data)
{
	struct module *mod = data;

	if (val == MODULE_STATE_GOING)
4648
		ftrace_release_mod(mod);
4649 4650 4651 4652

	return 0;
}
#else
4653 4654
static int ftrace_module_notify_exit(struct notifier_block *self,
				     unsigned long val, void *data)
4655 4656 4657 4658 4659
{
	return 0;
}
#endif /* CONFIG_MODULES */

4660 4661 4662 4663 4664
struct notifier_block ftrace_module_exit_nb = {
	.notifier_call = ftrace_module_notify_exit,
	.priority = INT_MIN,	/* Run after anything that can remove kprobes */
};

4665 4666
void __init ftrace_init(void)
{
4667 4668
	extern unsigned long __start_mcount_loc[];
	extern unsigned long __stop_mcount_loc[];
4669
	unsigned long count, flags;
4670 4671 4672
	int ret;

	local_irq_save(flags);
4673
	ret = ftrace_dyn_arch_init();
4674
	local_irq_restore(flags);
4675
	if (ret)
4676 4677 4678
		goto failed;

	count = __stop_mcount_loc - __start_mcount_loc;
4679 4680
	if (!count) {
		pr_info("ftrace: No functions to be traced?\n");
4681
		goto failed;
4682 4683 4684 4685
	}

	pr_info("ftrace: allocating %ld entries in %ld pages\n",
		count, count / ENTRIES_PER_PAGE + 1);
4686 4687 4688

	last_ftrace_enabled = ftrace_enabled = 1;

4689
	ret = ftrace_process_locs(NULL,
4690
				  __start_mcount_loc,
4691 4692
				  __stop_mcount_loc);

4693
	ret = register_module_notifier(&ftrace_module_exit_nb);
4694
	if (ret)
4695
		pr_warning("Failed to register trace ftrace module exit notifier\n");
4696

4697 4698
	set_ftrace_early_filters();

4699 4700 4701 4702 4703
	return;
 failed:
	ftrace_disabled = 1;
}

4704
#else
4705

4706
static struct ftrace_ops global_ops = {
4707
	.func			= ftrace_stub,
4708
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_INITIALIZED,
4709 4710
};

4711 4712 4713 4714 4715
static int __init ftrace_nodyn_init(void)
{
	ftrace_enabled = 1;
	return 0;
}
4716
core_initcall(ftrace_nodyn_init);
4717

4718 4719
static inline int ftrace_init_dyn_debugfs(struct dentry *d_tracer) { return 0; }
static inline void ftrace_startup_enable(int command) { }
4720
/* Keep as macros so we do not need to define the commands */
4721 4722 4723 4724 4725 4726
# define ftrace_startup(ops, command)					\
	({								\
		int ___ret = __register_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags |= FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
4727
	})
4728 4729 4730 4731 4732 4733 4734
# define ftrace_shutdown(ops, command)					\
	({								\
		int ___ret = __unregister_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags &= ~FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
	})
4735

I
Ingo Molnar 已提交
4736 4737
# define ftrace_startup_sysctl()	do { } while (0)
# define ftrace_shutdown_sysctl()	do { } while (0)
4738 4739

static inline int
4740
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
4741 4742 4743 4744
{
	return 1;
}

4745 4746
#endif /* CONFIG_DYNAMIC_FTRACE */

4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771 4772 4773 4774
__init void ftrace_init_global_array_ops(struct trace_array *tr)
{
	tr->ops = &global_ops;
	tr->ops->private = tr;
}

void ftrace_init_array_ops(struct trace_array *tr, ftrace_func_t func)
{
	/* If we filter on pids, update to use the pid function */
	if (tr->flags & TRACE_ARRAY_FL_GLOBAL) {
		if (WARN_ON(tr->ops->func != ftrace_stub))
			printk("ftrace ops had %pS for function\n",
			       tr->ops->func);
		/* Only the top level instance does pid tracing */
		if (!list_empty(&ftrace_pids)) {
			set_ftrace_pid_function(func);
			func = ftrace_pid_func;
		}
	}
	tr->ops->func = func;
	tr->ops->private = tr;
}

void ftrace_reset_array_ops(struct trace_array *tr)
{
	tr->ops->func = ftrace_stub;
}

4775
static void
4776
ftrace_ops_control_func(unsigned long ip, unsigned long parent_ip,
4777
			struct ftrace_ops *op, struct pt_regs *regs)
4778 4779 4780 4781 4782 4783 4784 4785 4786 4787
{
	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);
4788 4789 4790 4791 4792 4793 4794 4795

	/*
	 * Control funcs (perf) uses RCU. Only trace if
	 * RCU is currently active.
	 */
	if (!rcu_is_watching())
		goto out;

4796
	do_for_each_ftrace_op(op, ftrace_control_list) {
4797 4798
		if (!(op->flags & FTRACE_OPS_FL_STUB) &&
		    !ftrace_function_local_disabled(op) &&
4799
		    ftrace_ops_test(op, ip, regs))
4800
			op->func(ip, parent_ip, op, regs);
4801
	} while_for_each_ftrace_op(op);
4802
 out:
4803 4804 4805 4806 4807
	trace_recursion_clear(TRACE_CONTROL_BIT);
	preempt_enable_notrace();
}

static struct ftrace_ops control_ops = {
4808 4809
	.func	= ftrace_ops_control_func,
	.flags	= FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_INITIALIZED,
4810
	INIT_OPS_HASH(control_ops)
4811 4812
};

4813 4814
static inline void
__ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
4815
		       struct ftrace_ops *ignored, struct pt_regs *regs)
4816
{
4817
	struct ftrace_ops *op;
4818
	int bit;
4819

4820 4821 4822
	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;
4823

4824 4825 4826 4827 4828
	/*
	 * Some of the ops may be dynamically allocated,
	 * they must be freed after a synchronize_sched().
	 */
	preempt_disable_notrace();
4829
	do_for_each_ftrace_op(op, ftrace_ops_list) {
4830
		if (ftrace_ops_test(op, ip, regs)) {
4831 4832
			if (FTRACE_WARN_ON(!op->func)) {
				pr_warn("op=%p %pS\n", op, op);
4833 4834
				goto out;
			}
4835
			op->func(ip, parent_ip, op, regs);
4836
		}
4837
	} while_for_each_ftrace_op(op);
4838
out:
4839
	preempt_enable_notrace();
4840
	trace_clear_recursion(bit);
4841 4842
}

4843 4844 4845 4846 4847
/*
 * Some archs only support passing ip and parent_ip. Even though
 * the list function ignores the op parameter, we do not want any
 * C side effects, where a function is called without the caller
 * sending a third parameter.
4848 4849 4850
 * Archs are to support both the regs and ftrace_ops at the same time.
 * If they support ftrace_ops, it is assumed they support regs.
 * If call backs want to use regs, they must either check for regs
4851 4852
 * being NULL, or CONFIG_DYNAMIC_FTRACE_WITH_REGS.
 * Note, CONFIG_DYNAMIC_FTRACE_WITH_REGS expects a full regs to be saved.
4853 4854
 * An architecture can pass partial regs with ftrace_ops and still
 * set the ARCH_SUPPORT_FTARCE_OPS.
4855 4856 4857
 */
#if ARCH_SUPPORTS_FTRACE_OPS
static void ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
4858
				 struct ftrace_ops *op, struct pt_regs *regs)
4859
{
4860
	__ftrace_ops_list_func(ip, parent_ip, NULL, regs);
4861 4862 4863 4864
}
#else
static void ftrace_ops_no_ops(unsigned long ip, unsigned long parent_ip)
{
4865
	__ftrace_ops_list_func(ip, parent_ip, NULL, NULL);
4866 4867 4868
}
#endif

4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887
/*
 * If there's only one function registered but it does not support
 * recursion, this function will be called by the mcount trampoline.
 * This function will handle recursion protection.
 */
static void ftrace_ops_recurs_func(unsigned long ip, unsigned long parent_ip,
				   struct ftrace_ops *op, struct pt_regs *regs)
{
	int bit;

	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;

	op->func(ip, parent_ip, op, regs);

	trace_clear_recursion(bit);
}

4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918
/**
 * ftrace_ops_get_func - get the function a trampoline should call
 * @ops: the ops to get the function for
 *
 * Normally the mcount trampoline will call the ops->func, but there
 * are times that it should not. For example, if the ops does not
 * have its own recursion protection, then it should call the
 * ftrace_ops_recurs_func() instead.
 *
 * Returns the function that the trampoline should call for @ops.
 */
ftrace_func_t ftrace_ops_get_func(struct ftrace_ops *ops)
{
	/*
	 * If this is a dynamic ops or we force list func,
	 * then it needs to call the list anyway.
	 */
	if (ops->flags & FTRACE_OPS_FL_DYNAMIC || FTRACE_FORCE_LIST_FUNC)
		return ftrace_ops_list_func;

	/*
	 * If the func handles its own recursion, call it directly.
	 * Otherwise call the recursion protected function that
	 * will call the ftrace ops function.
	 */
	if (!(ops->flags & FTRACE_OPS_FL_RECURSION_SAFE))
		return ftrace_ops_recurs_func;

	return ops->func;
}

4919
static void clear_ftrace_swapper(void)
S
Steven Rostedt 已提交
4920 4921
{
	struct task_struct *p;
4922
	int cpu;
S
Steven Rostedt 已提交
4923

4924 4925 4926
	get_online_cpus();
	for_each_online_cpu(cpu) {
		p = idle_task(cpu);
S
Steven Rostedt 已提交
4927
		clear_tsk_trace_trace(p);
4928 4929 4930
	}
	put_online_cpus();
}
S
Steven Rostedt 已提交
4931

4932 4933 4934 4935 4936 4937 4938 4939 4940 4941 4942
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
4943 4944
}

4945 4946 4947 4948
static void clear_ftrace_pid(struct pid *pid)
{
	struct task_struct *p;

4949
	rcu_read_lock();
4950 4951 4952
	do_each_pid_task(pid, PIDTYPE_PID, p) {
		clear_tsk_trace_trace(p);
	} while_each_pid_task(pid, PIDTYPE_PID, p);
4953 4954
	rcu_read_unlock();

4955 4956 4957 4958
	put_pid(pid);
}

static void set_ftrace_pid(struct pid *pid)
S
Steven Rostedt 已提交
4959 4960 4961
{
	struct task_struct *p;

4962
	rcu_read_lock();
S
Steven Rostedt 已提交
4963 4964 4965
	do_each_pid_task(pid, PIDTYPE_PID, p) {
		set_tsk_trace_trace(p);
	} while_each_pid_task(pid, PIDTYPE_PID, p);
4966
	rcu_read_unlock();
S
Steven Rostedt 已提交
4967 4968
}

4969
static void clear_ftrace_pid_task(struct pid *pid)
4970
{
4971
	if (pid == ftrace_swapper_pid)
4972 4973
		clear_ftrace_swapper();
	else
4974
		clear_ftrace_pid(pid);
4975 4976 4977 4978 4979 4980 4981 4982 4983 4984
}

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

4985
static int ftrace_pid_add(int p)
4986
{
S
Steven Rostedt 已提交
4987
	struct pid *pid;
4988 4989
	struct ftrace_pid *fpid;
	int ret = -EINVAL;
4990

4991
	mutex_lock(&ftrace_lock);
4992

4993 4994 4995 4996
	if (!p)
		pid = ftrace_swapper_pid;
	else
		pid = find_get_pid(p);
4997

4998 4999
	if (!pid)
		goto out;
5000

5001
	ret = 0;
5002

5003 5004 5005
	list_for_each_entry(fpid, &ftrace_pids, list)
		if (fpid->pid == pid)
			goto out_put;
S
Steven Rostedt 已提交
5006

5007
	ret = -ENOMEM;
5008

5009 5010 5011
	fpid = kmalloc(sizeof(*fpid), GFP_KERNEL);
	if (!fpid)
		goto out_put;
5012

5013 5014
	list_add(&fpid->list, &ftrace_pids);
	fpid->pid = pid;
5015

5016
	set_ftrace_pid_task(pid);
S
Steven Rostedt 已提交
5017

5018 5019 5020 5021 5022 5023 5024 5025 5026
	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
5027

5028 5029 5030 5031 5032 5033 5034 5035
out:
	mutex_unlock(&ftrace_lock);
	return ret;
}

static void ftrace_pid_reset(void)
{
	struct ftrace_pid *fpid, *safe;
S
Steven Rostedt 已提交
5036

5037 5038 5039 5040 5041 5042 5043 5044
	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);
5045 5046 5047 5048 5049
	}

	ftrace_update_pid_func();
	ftrace_startup_enable(0);

S
Steven Rostedt 已提交
5050
	mutex_unlock(&ftrace_lock);
5051
}
5052

5053 5054 5055 5056 5057 5058 5059 5060 5061 5062 5063 5064 5065 5066 5067 5068 5069 5070 5071 5072 5073 5074 5075 5076 5077 5078 5079 5080 5081 5082 5083 5084 5085 5086 5087 5088 5089 5090 5091 5092 5093 5094 5095 5096 5097 5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114
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;
}

5115 5116 5117 5118
static ssize_t
ftrace_pid_write(struct file *filp, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
5119
	char buf[64], *tmp;
5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130
	long val;
	int ret;

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

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

	buf[cnt] = 0;

5131 5132 5133 5134
	/*
	 * Allow "echo > set_ftrace_pid" or "echo -n '' > set_ftrace_pid"
	 * to clean the filter quietly.
	 */
5135 5136
	tmp = strstrip(buf);
	if (strlen(tmp) == 0)
5137 5138
		return 1;

5139
	ret = kstrtol(tmp, 10, &val);
5140 5141 5142
	if (ret < 0)
		return ret;

5143
	ret = ftrace_pid_add(val);
5144

5145 5146
	return ret ? ret : cnt;
}
5147

5148 5149 5150 5151 5152
static int
ftrace_pid_release(struct inode *inode, struct file *file)
{
	if (file->f_mode & FMODE_READ)
		seq_release(inode, file);
5153

5154
	return 0;
5155 5156
}

5157
static const struct file_operations ftrace_pid_fops = {
5158 5159 5160
	.open		= ftrace_pid_open,
	.write		= ftrace_pid_write,
	.read		= seq_read,
5161
	.llseek		= tracing_lseek,
5162
	.release	= ftrace_pid_release,
5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173 5174
};

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);

5175 5176
	trace_create_file("set_ftrace_pid", 0644, d_tracer,
			    NULL, &ftrace_pid_fops);
5177 5178 5179

	ftrace_profile_debugfs(d_tracer);

5180 5181 5182 5183
	return 0;
}
fs_initcall(ftrace_init_debugfs);

S
Steven Rostedt 已提交
5184
/**
5185
 * ftrace_kill - kill ftrace
S
Steven Rostedt 已提交
5186 5187 5188 5189 5190
 *
 * 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.
 */
5191
void ftrace_kill(void)
S
Steven Rostedt 已提交
5192 5193 5194 5195 5196 5197
{
	ftrace_disabled = 1;
	ftrace_enabled = 0;
	clear_ftrace_function();
}

5198 5199 5200 5201 5202 5203 5204 5205
/**
 * Test if ftrace is dead or not.
 */
int ftrace_is_dead(void)
{
	return ftrace_disabled;
}

5206
/**
5207 5208
 * register_ftrace_function - register a function for profiling
 * @ops - ops structure that holds the function for profiling.
5209
 *
5210 5211 5212 5213 5214 5215
 * 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.
5216
 */
5217
int register_ftrace_function(struct ftrace_ops *ops)
5218
{
5219
	int ret = -1;
5220

5221 5222
	ftrace_ops_init(ops);

S
Steven Rostedt 已提交
5223
	mutex_lock(&ftrace_lock);
5224

5225
	ret = ftrace_startup(ops, 0);
5226

S
Steven Rostedt 已提交
5227
	mutex_unlock(&ftrace_lock);
5228

5229
	return ret;
5230
}
5231
EXPORT_SYMBOL_GPL(register_ftrace_function);
5232 5233

/**
5234
 * unregister_ftrace_function - unregister a function for profiling.
5235 5236 5237 5238 5239 5240 5241 5242
 * @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
5243
	mutex_lock(&ftrace_lock);
5244
	ret = ftrace_shutdown(ops, 0);
S
Steven Rostedt 已提交
5245
	mutex_unlock(&ftrace_lock);
5246 5247 5248

	return ret;
}
5249
EXPORT_SYMBOL_GPL(unregister_ftrace_function);
5250

I
Ingo Molnar 已提交
5251
int
5252
ftrace_enable_sysctl(struct ctl_table *table, int write,
5253
		     void __user *buffer, size_t *lenp,
5254 5255
		     loff_t *ppos)
{
5256
	int ret = -ENODEV;
5257

S
Steven Rostedt 已提交
5258
	mutex_lock(&ftrace_lock);
5259

5260 5261 5262 5263
	if (unlikely(ftrace_disabled))
		goto out;

	ret = proc_dointvec(table, write, buffer, lenp, ppos);
5264

5265
	if (ret || !write || (last_ftrace_enabled == !!ftrace_enabled))
5266 5267
		goto out;

5268
	last_ftrace_enabled = !!ftrace_enabled;
5269 5270 5271 5272 5273 5274

	if (ftrace_enabled) {

		ftrace_startup_sysctl();

		/* we are starting ftrace again */
5275 5276
		if (ftrace_ops_list != &ftrace_list_end)
			update_ftrace_function();
5277 5278 5279 5280 5281 5282 5283 5284 5285

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

		ftrace_shutdown_sysctl();
	}

 out:
S
Steven Rostedt 已提交
5286
	mutex_unlock(&ftrace_lock);
5287
	return ret;
5288
}
5289

5290
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
5291

5292 5293 5294 5295 5296 5297 5298 5299 5300 5301 5302
static struct ftrace_ops graph_ops = {
	.func			= ftrace_stub,
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE |
				   FTRACE_OPS_FL_INITIALIZED |
				   FTRACE_OPS_FL_STUB,
#ifdef FTRACE_GRAPH_TRAMP_ADDR
	.trampoline		= FTRACE_GRAPH_TRAMP_ADDR,
#endif
	ASSIGN_OPS_HASH(graph_ops, &global_ops.local_hash)
};

5303
static int ftrace_graph_active;
5304

5305 5306 5307 5308 5309
int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace)
{
	return 0;
}

5310 5311 5312
/* The callbacks that hook a function */
trace_func_graph_ret_t ftrace_graph_return =
			(trace_func_graph_ret_t)ftrace_stub;
5313
trace_func_graph_ent_t ftrace_graph_entry = ftrace_graph_entry_stub;
5314
static trace_func_graph_ent_t __ftrace_graph_entry = ftrace_graph_entry_stub;
5315 5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344

/* 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) {
5345
			atomic_set(&t->tracing_graph_pause, 0);
5346
			atomic_set(&t->trace_overrun, 0);
5347 5348 5349 5350
			t->curr_ret_stack = -1;
			/* Make sure the tasks see the -1 first: */
			smp_wmb();
			t->ret_stack = ret_stack_list[start++];
5351 5352 5353 5354 5355 5356 5357 5358 5359 5360 5361
		}
	} 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;
}

5362
static void
5363 5364
ftrace_graph_probe_sched_switch(void *ignore,
			struct task_struct *prev, struct task_struct *next)
5365 5366 5367 5368
{
	unsigned long long timestamp;
	int index;

5369 5370 5371 5372 5373 5374 5375
	/*
	 * 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;

5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393
	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;
}

5394
/* Allocate a return stack for each task */
5395
static int start_graph_tracing(void)
5396 5397
{
	struct ftrace_ret_stack **ret_stack_list;
5398
	int ret, cpu;
5399 5400 5401 5402 5403 5404 5405 5406

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

	if (!ret_stack_list)
		return -ENOMEM;

5407
	/* The cpu_boot init_task->ret_stack will never be freed */
5408 5409
	for_each_online_cpu(cpu) {
		if (!idle_task(cpu)->ret_stack)
5410
			ftrace_graph_init_idle_task(idle_task(cpu), cpu);
5411
	}
5412

5413 5414 5415 5416
	do {
		ret = alloc_retstack_tasklist(ret_stack_list);
	} while (ret == -EAGAIN);

5417
	if (!ret) {
5418
		ret = register_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
5419 5420 5421 5422 5423
		if (ret)
			pr_info("ftrace_graph: Couldn't activate tracepoint"
				" probe to kernel_sched_switch\n");
	}

5424 5425 5426 5427
	kfree(ret_stack_list);
	return ret;
}

5428 5429 5430 5431 5432 5433 5434 5435 5436 5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448
/*
 * 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;
}

5449 5450 5451 5452 5453 5454 5455 5456 5457 5458 5459 5460 5461 5462 5463 5464
static int ftrace_graph_entry_test(struct ftrace_graph_ent *trace)
{
	if (!ftrace_ops_test(&global_ops, trace->func, NULL))
		return 0;
	return __ftrace_graph_entry(trace);
}

/*
 * The function graph tracer should only trace the functions defined
 * by set_ftrace_filter and set_ftrace_notrace. If another function
 * tracer ops is registered, the graph tracer requires testing the
 * function against the global ops, and not just trace any function
 * that any ftrace_ops registered.
 */
static void update_function_graph_func(void)
{
5465 5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483
	struct ftrace_ops *op;
	bool do_test = false;

	/*
	 * The graph and global ops share the same set of functions
	 * to test. If any other ops is on the list, then
	 * the graph tracing needs to test if its the function
	 * it should call.
	 */
	do_for_each_ftrace_op(op, ftrace_ops_list) {
		if (op != &global_ops && op != &graph_ops &&
		    op != &ftrace_list_end) {
			do_test = true;
			/* in double loop, break out with goto */
			goto out;
		}
	} while_for_each_ftrace_op(op);
 out:
	if (do_test)
5484
		ftrace_graph_entry = ftrace_graph_entry_test;
5485 5486
	else
		ftrace_graph_entry = __ftrace_graph_entry;
5487 5488
}

5489 5490 5491 5492
static struct notifier_block ftrace_suspend_notifier = {
	.notifier_call = ftrace_suspend_notifier_call,
};

5493 5494
int register_ftrace_graph(trace_func_graph_ret_t retfunc,
			trace_func_graph_ent_t entryfunc)
5495
{
5496 5497
	int ret = 0;

S
Steven Rostedt 已提交
5498
	mutex_lock(&ftrace_lock);
5499

5500
	/* we currently allow only one tracer registered at a time */
5501
	if (ftrace_graph_active) {
5502 5503 5504 5505
		ret = -EBUSY;
		goto out;
	}

5506 5507
	register_pm_notifier(&ftrace_suspend_notifier);

5508
	ftrace_graph_active++;
5509
	ret = start_graph_tracing();
5510
	if (ret) {
5511
		ftrace_graph_active--;
5512 5513
		goto out;
	}
5514

5515
	ftrace_graph_return = retfunc;
5516 5517 5518 5519 5520 5521 5522 5523 5524 5525

	/*
	 * Update the indirect function to the entryfunc, and the
	 * function that gets called to the entry_test first. Then
	 * call the update fgraph entry function to determine if
	 * the entryfunc should be called directly or not.
	 */
	__ftrace_graph_entry = entryfunc;
	ftrace_graph_entry = ftrace_graph_entry_test;
	update_function_graph_func();
5526

5527
	ret = ftrace_startup(&graph_ops, FTRACE_START_FUNC_RET);
5528 5529

out:
S
Steven Rostedt 已提交
5530
	mutex_unlock(&ftrace_lock);
5531
	return ret;
5532 5533
}

5534
void unregister_ftrace_graph(void)
5535
{
S
Steven Rostedt 已提交
5536
	mutex_lock(&ftrace_lock);
5537

5538
	if (unlikely(!ftrace_graph_active))
5539 5540
		goto out;

5541
	ftrace_graph_active--;
5542
	ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub;
5543
	ftrace_graph_entry = ftrace_graph_entry_stub;
5544
	__ftrace_graph_entry = ftrace_graph_entry_stub;
5545
	ftrace_shutdown(&graph_ops, FTRACE_STOP_FUNC_RET);
5546
	unregister_pm_notifier(&ftrace_suspend_notifier);
5547
	unregister_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
5548

5549
 out:
S
Steven Rostedt 已提交
5550
	mutex_unlock(&ftrace_lock);
5551
}
5552

5553 5554 5555 5556 5557 5558 5559 5560
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
5561
	/* make curr_ret_stack visible before we add the ret_stack */
5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592 5593 5594 5595
	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);
	}
}

5596
/* Allocate a return stack for newly created task */
5597
void ftrace_graph_init_task(struct task_struct *t)
5598
{
5599 5600
	/* Make sure we do not use the parent ret_stack */
	t->ret_stack = NULL;
5601
	t->curr_ret_stack = -1;
5602

5603
	if (ftrace_graph_active) {
5604 5605 5606
		struct ftrace_ret_stack *ret_stack;

		ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH
5607 5608
				* sizeof(struct ftrace_ret_stack),
				GFP_KERNEL);
5609
		if (!ret_stack)
5610
			return;
5611
		graph_init_task(t, ret_stack);
5612
	}
5613 5614
}

5615
void ftrace_graph_exit_task(struct task_struct *t)
5616
{
5617 5618
	struct ftrace_ret_stack	*ret_stack = t->ret_stack;

5619
	t->ret_stack = NULL;
5620 5621 5622 5623
	/* NULL must become visible to IRQs before we free it: */
	barrier();

	kfree(ret_stack);
5624
}
5625
#endif