ftrace.c 136.8 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/tracefs.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

252 253 254 255 256 257 258 259 260 261 262 263 264

static ftrace_func_t ftrace_ops_get_list_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;

	return ftrace_ops_get_func(ops);
}

265 266 267 268
static void update_ftrace_function(void)
{
	ftrace_func_t func;

269 270 271 272 273 274 275 276 277 278 279
	/*
	 * 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;

280 281
	/*
	 * If we are at the end of the list and this ops is
282 283
	 * recursion safe and not dynamic and the arch supports passing ops,
	 * then have the mcount trampoline call the function directly.
284
	 */
285
	} else if (ftrace_ops_list->next == &ftrace_list_end) {
286
		func = ftrace_ops_get_list_func(ftrace_ops_list);
287

288 289
	} else {
		/* Just use the default ftrace_ops */
290
		set_function_trace_op = &ftrace_list_end;
291
		func = ftrace_ops_list_func;
292
	}
293

294 295
	update_function_graph_func();

296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338
	/* 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 */

339 340 341
	ftrace_trace_function = func;
}

342 343 344 345 346
int using_ftrace_ops_list_func(void)
{
	return ftrace_trace_function == ftrace_ops_list_func;
}

347
static void add_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
348
{
349
	ops->next = *list;
350
	/*
351
	 * We are entering ops into the list but another
352 353
	 * CPU might be walking that list. We need to make sure
	 * the ops->next pointer is valid before another CPU sees
354
	 * the ops pointer included into the list.
355
	 */
356
	rcu_assign_pointer(*list, ops);
357 358
}

359
static int remove_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
360 361 362 363
{
	struct ftrace_ops **p;

	/*
364 365
	 * If we are removing the last function, then simply point
	 * to the ftrace_stub.
366
	 */
367 368
	if (*list == ops && ops->next == &ftrace_list_end) {
		*list = &ftrace_list_end;
S
Steven Rostedt 已提交
369
		return 0;
370 371
	}

372
	for (p = list; *p != &ftrace_list_end; p = &(*p)->next)
373 374 375
		if (*p == ops)
			break;

S
Steven Rostedt 已提交
376 377
	if (*p != ops)
		return -1;
378 379

	*p = (*p)->next;
380 381
	return 0;
}
382

383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402
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;
}

403 404
static void ftrace_update_trampoline(struct ftrace_ops *ops);

405 406
static int __register_ftrace_function(struct ftrace_ops *ops)
{
407 408 409
	if (ops->flags & FTRACE_OPS_FL_DELETED)
		return -EINVAL;

410 411 412
	if (WARN_ON(ops->flags & FTRACE_OPS_FL_ENABLED))
		return -EBUSY;

413
#ifndef CONFIG_DYNAMIC_FTRACE_WITH_REGS
414 415 416 417 418 419 420 421 422 423 424 425 426
	/*
	 * 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

427 428 429
	if (!core_kernel_data((unsigned long)ops))
		ops->flags |= FTRACE_OPS_FL_DYNAMIC;

430
	if (ops->flags & FTRACE_OPS_FL_CONTROL) {
431 432 433
		if (control_ops_alloc(ops))
			return -ENOMEM;
		add_ftrace_list_ops(&ftrace_control_list, &control_ops, ops);
434 435
		/* The control_ops needs the trampoline update */
		ops = &control_ops;
436 437 438
	} else
		add_ftrace_ops(&ftrace_ops_list, ops);

439 440
	ftrace_update_trampoline(ops);

441 442 443 444 445 446 447 448 449 450
	if (ftrace_enabled)
		update_ftrace_function();

	return 0;
}

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

451 452 453
	if (WARN_ON(!(ops->flags & FTRACE_OPS_FL_ENABLED)))
		return -EBUSY;

454
	if (ops->flags & FTRACE_OPS_FL_CONTROL) {
455 456
		ret = remove_ftrace_list_ops(&ftrace_control_list,
					     &control_ops, ops);
457 458 459
	} else
		ret = remove_ftrace_ops(&ftrace_ops_list, ops);

460 461
	if (ret < 0)
		return ret;
462

463 464
	if (ftrace_enabled)
		update_ftrace_function();
465

S
Steven Rostedt 已提交
466
	return 0;
467 468
}

469 470
static void ftrace_update_pid_func(void)
{
471
	/* Only do something if we are tracing something */
472
	if (ftrace_trace_function == ftrace_stub)
473
		return;
474

475
	update_ftrace_function();
476 477
}

478 479 480 481 482
#ifdef CONFIG_FUNCTION_PROFILER
struct ftrace_profile {
	struct hlist_node		node;
	unsigned long			ip;
	unsigned long			counter;
483 484
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	unsigned long long		time;
485
	unsigned long long		time_squared;
486
#endif
487 488
};

489 490 491 492
struct ftrace_profile_page {
	struct ftrace_profile_page	*next;
	unsigned long			index;
	struct ftrace_profile		records[];
493 494
};

495 496 497 498 499 500 501 502
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;
};

503 504
#define PROFILE_RECORDS_SIZE						\
	(PAGE_SIZE - offsetof(struct ftrace_profile_page, records))
505

506 507
#define PROFILES_PER_PAGE					\
	(PROFILE_RECORDS_SIZE / sizeof(struct ftrace_profile))
508

509 510 511
static int ftrace_profile_enabled __read_mostly;

/* ftrace_profile_lock - synchronize the enable and disable of the profiler */
S
Steven Rostedt 已提交
512 513
static DEFINE_MUTEX(ftrace_profile_lock);

514
static DEFINE_PER_CPU(struct ftrace_profile_stat, ftrace_profile_stats);
515

516 517
#define FTRACE_PROFILE_HASH_BITS 10
#define FTRACE_PROFILE_HASH_SIZE (1 << FTRACE_PROFILE_HASH_BITS)
518

S
Steven Rostedt 已提交
519 520 521
static void *
function_stat_next(void *v, int idx)
{
522 523
	struct ftrace_profile *rec = v;
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
524

525
	pg = (struct ftrace_profile_page *)((unsigned long)rec & PAGE_MASK);
S
Steven Rostedt 已提交
526 527

 again:
L
Li Zefan 已提交
528 529 530
	if (idx != 0)
		rec++;

S
Steven Rostedt 已提交
531 532 533 534 535
	if ((void *)rec >= (void *)&pg->records[pg->index]) {
		pg = pg->next;
		if (!pg)
			return NULL;
		rec = &pg->records[0];
536 537
		if (!rec->counter)
			goto again;
S
Steven Rostedt 已提交
538 539 540 541 542 543 544
	}

	return rec;
}

static void *function_stat_start(struct tracer_stat *trace)
{
545 546 547 548 549 550 551
	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
552 553
}

554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569
#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
570 571
static int function_stat_cmp(void *p1, void *p2)
{
572 573
	struct ftrace_profile *a = p1;
	struct ftrace_profile *b = p2;
S
Steven Rostedt 已提交
574 575 576 577 578 579 580 581

	if (a->counter < b->counter)
		return -1;
	if (a->counter > b->counter)
		return 1;
	else
		return 0;
}
582
#endif
S
Steven Rostedt 已提交
583 584 585

static int function_stat_headers(struct seq_file *m)
{
586
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
587 588 589 590
	seq_puts(m, "  Function                               "
		 "Hit    Time            Avg             s^2\n"
		    "  --------                               "
		 "---    ----            ---             ---\n");
591
#else
592 593
	seq_puts(m, "  Function                               Hit\n"
		    "  --------                               ---\n");
594
#endif
S
Steven Rostedt 已提交
595 596 597 598 599
	return 0;
}

static int function_stat_show(struct seq_file *m, void *v)
{
600
	struct ftrace_profile *rec = v;
S
Steven Rostedt 已提交
601
	char str[KSYM_SYMBOL_LEN];
602
	int ret = 0;
603
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
604 605
	static struct trace_seq s;
	unsigned long long avg;
606
	unsigned long long stddev;
607
#endif
608 609 610 611 612 613 614
	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
615

616 617 618 619 620 621 622
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	avg = rec->time;
	do_div(avg, rec->counter);
	if (tracing_thresh && (avg < tracing_thresh))
		goto out;
#endif

S
Steven Rostedt 已提交
623
	kallsyms_lookup(rec->ip, NULL, NULL, NULL, str);
624 625 626
	seq_printf(m, "  %-30.30s  %10lu", str, rec->counter);

#ifdef CONFIG_FUNCTION_GRAPH_TRACER
627
	seq_puts(m, "    ");
628

629 630 631 632
	/* Sample standard deviation (s^2) */
	if (rec->counter <= 1)
		stddev = 0;
	else {
633 634 635 636 637 638 639
		/*
		 * 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;

640 641 642 643
		/*
		 * Divide only 1000 for ns^2 -> us^2 conversion.
		 * trace_print_graph_duration will divide 1000 again.
		 */
644
		do_div(stddev, rec->counter * (rec->counter - 1) * 1000);
645 646
	}

647 648 649 650
	trace_seq_init(&s);
	trace_print_graph_duration(rec->time, &s);
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(avg, &s);
651 652
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(stddev, &s);
653 654 655
	trace_print_seq(m, &s);
#endif
	seq_putc(m, '\n');
656 657
out:
	mutex_unlock(&ftrace_profile_lock);
S
Steven Rostedt 已提交
658

659
	return ret;
S
Steven Rostedt 已提交
660 661
}

662
static void ftrace_profile_reset(struct ftrace_profile_stat *stat)
S
Steven Rostedt 已提交
663
{
664
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
665

666
	pg = stat->pages = stat->start;
S
Steven Rostedt 已提交
667

668 669 670 671
	while (pg) {
		memset(pg->records, 0, PROFILE_RECORDS_SIZE);
		pg->index = 0;
		pg = pg->next;
S
Steven Rostedt 已提交
672 673
	}

674
	memset(stat->hash, 0,
675 676
	       FTRACE_PROFILE_HASH_SIZE * sizeof(struct hlist_head));
}
S
Steven Rostedt 已提交
677

678
int ftrace_profile_pages_init(struct ftrace_profile_stat *stat)
679 680
{
	struct ftrace_profile_page *pg;
681 682
	int functions;
	int pages;
683
	int i;
S
Steven Rostedt 已提交
684

685
	/* If we already allocated, do nothing */
686
	if (stat->pages)
687
		return 0;
S
Steven Rostedt 已提交
688

689 690
	stat->pages = (void *)get_zeroed_page(GFP_KERNEL);
	if (!stat->pages)
691
		return -ENOMEM;
S
Steven Rostedt 已提交
692

693 694 695 696 697 698 699 700 701 702 703 704 705
#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

706
	pg = stat->start = stat->pages;
S
Steven Rostedt 已提交
707

708 709
	pages = DIV_ROUND_UP(functions, PROFILES_PER_PAGE);

710
	for (i = 1; i < pages; i++) {
711 712
		pg->next = (void *)get_zeroed_page(GFP_KERNEL);
		if (!pg->next)
713
			goto out_free;
714 715 716 717
		pg = pg->next;
	}

	return 0;
718 719 720 721 722 723 724 725 726 727 728 729 730 731

 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
732 733
}

734
static int ftrace_profile_init_cpu(int cpu)
S
Steven Rostedt 已提交
735
{
736
	struct ftrace_profile_stat *stat;
737
	int size;
S
Steven Rostedt 已提交
738

739 740 741
	stat = &per_cpu(ftrace_profile_stats, cpu);

	if (stat->hash) {
742
		/* If the profile is already created, simply reset it */
743
		ftrace_profile_reset(stat);
744 745
		return 0;
	}
S
Steven Rostedt 已提交
746

747 748 749 750 751
	/*
	 * 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
752

753
	stat->hash = kzalloc(sizeof(struct hlist_head) * size, GFP_KERNEL);
754

755
	if (!stat->hash)
756 757
		return -ENOMEM;

758
	/* Preallocate the function profiling pages */
759 760 761
	if (ftrace_profile_pages_init(stat) < 0) {
		kfree(stat->hash);
		stat->hash = NULL;
762 763 764 765
		return -ENOMEM;
	}

	return 0;
S
Steven Rostedt 已提交
766 767
}

768 769 770 771 772
static int ftrace_profile_init(void)
{
	int cpu;
	int ret = 0;

773
	for_each_possible_cpu(cpu) {
774 775 776 777 778 779 780 781
		ret = ftrace_profile_init_cpu(cpu);
		if (ret)
			break;
	}

	return ret;
}

782
/* interrupts must be disabled */
783 784
static struct ftrace_profile *
ftrace_find_profiled_func(struct ftrace_profile_stat *stat, unsigned long ip)
S
Steven Rostedt 已提交
785
{
786
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
787 788 789
	struct hlist_head *hhd;
	unsigned long key;

790
	key = hash_long(ip, FTRACE_PROFILE_HASH_BITS);
791
	hhd = &stat->hash[key];
S
Steven Rostedt 已提交
792 793 794 795

	if (hlist_empty(hhd))
		return NULL;

796
	hlist_for_each_entry_rcu_notrace(rec, hhd, node) {
S
Steven Rostedt 已提交
797
		if (rec->ip == ip)
798 799 800 801 802 803
			return rec;
	}

	return NULL;
}

804 805
static void ftrace_add_profile(struct ftrace_profile_stat *stat,
			       struct ftrace_profile *rec)
806 807 808
{
	unsigned long key;

809
	key = hash_long(rec->ip, FTRACE_PROFILE_HASH_BITS);
810
	hlist_add_head_rcu(&rec->node, &stat->hash[key]);
811 812
}

813 814 815
/*
 * The memory is already allocated, this simply finds a new record to use.
 */
816
static struct ftrace_profile *
817
ftrace_profile_alloc(struct ftrace_profile_stat *stat, unsigned long ip)
818 819 820
{
	struct ftrace_profile *rec = NULL;

821
	/* prevent recursion (from NMIs) */
822
	if (atomic_inc_return(&stat->disabled) != 1)
823 824 825
		goto out;

	/*
826 827
	 * Try to find the function again since an NMI
	 * could have added it
828
	 */
829
	rec = ftrace_find_profiled_func(stat, ip);
830
	if (rec)
831
		goto out;
832

833 834 835 836
	if (stat->pages->index == PROFILES_PER_PAGE) {
		if (!stat->pages->next)
			goto out;
		stat->pages = stat->pages->next;
S
Steven Rostedt 已提交
837
	}
838

839
	rec = &stat->pages->records[stat->pages->index++];
840
	rec->ip = ip;
841
	ftrace_add_profile(stat, rec);
842

S
Steven Rostedt 已提交
843
 out:
844
	atomic_dec(&stat->disabled);
S
Steven Rostedt 已提交
845 846 847 848 849

	return rec;
}

static void
850
function_profile_call(unsigned long ip, unsigned long parent_ip,
851
		      struct ftrace_ops *ops, struct pt_regs *regs)
S
Steven Rostedt 已提交
852
{
853
	struct ftrace_profile_stat *stat;
854
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
855 856 857 858 859 860
	unsigned long flags;

	if (!ftrace_profile_enabled)
		return;

	local_irq_save(flags);
861

862
	stat = this_cpu_ptr(&ftrace_profile_stats);
863
	if (!stat->hash || !ftrace_profile_enabled)
864 865 866
		goto out;

	rec = ftrace_find_profiled_func(stat, ip);
867
	if (!rec) {
868
		rec = ftrace_profile_alloc(stat, ip);
869 870 871
		if (!rec)
			goto out;
	}
S
Steven Rostedt 已提交
872 873 874 875 876 877

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

878 879 880
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static int profile_graph_entry(struct ftrace_graph_ent *trace)
{
881
	function_profile_call(trace->func, 0, NULL, NULL);
882 883 884 885 886
	return 1;
}

static void profile_graph_return(struct ftrace_graph_ret *trace)
{
887
	struct ftrace_profile_stat *stat;
888
	unsigned long long calltime;
889
	struct ftrace_profile *rec;
890
	unsigned long flags;
891 892

	local_irq_save(flags);
893
	stat = this_cpu_ptr(&ftrace_profile_stats);
894
	if (!stat->hash || !ftrace_profile_enabled)
895 896
		goto out;

897 898 899 900
	/* If the calltime was zero'd ignore it */
	if (!trace->calltime)
		goto out;

901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917
	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;
	}

918
	rec = ftrace_find_profiled_func(stat, trace->func);
919
	if (rec) {
920
		rec->time += calltime;
921 922
		rec->time_squared += calltime * calltime;
	}
923

924
 out:
925 926 927 928 929 930 931 932 933 934 935 936 937 938
	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
939
static struct ftrace_ops ftrace_profile_ops __read_mostly = {
940
	.func		= function_profile_call,
941
	.flags		= FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_INITIALIZED,
942
	INIT_OPS_HASH(ftrace_profile_ops)
S
Steven Rostedt 已提交
943 944
};

945 946 947 948 949 950 951 952 953 954 955
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
956 957 958 959 960 961 962
static ssize_t
ftrace_profile_write(struct file *filp, const char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
	unsigned long val;
	int ret;

963 964
	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
	if (ret)
S
Steven Rostedt 已提交
965 966 967 968 969 970 971
		return ret;

	val = !!val;

	mutex_lock(&ftrace_profile_lock);
	if (ftrace_profile_enabled ^ val) {
		if (val) {
972 973 974 975 976 977
			ret = ftrace_profile_init();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}

978 979 980 981 982
			ret = register_ftrace_profiler();
			if (ret < 0) {
				cnt = ret;
				goto out;
			}
S
Steven Rostedt 已提交
983 984 985
			ftrace_profile_enabled = 1;
		} else {
			ftrace_profile_enabled = 0;
986 987 988 989
			/*
			 * unregister_ftrace_profiler calls stop_machine
			 * so this acts like an synchronize_sched.
			 */
990
			unregister_ftrace_profiler();
S
Steven Rostedt 已提交
991 992
		}
	}
993
 out:
S
Steven Rostedt 已提交
994 995
	mutex_unlock(&ftrace_profile_lock);

996
	*ppos += cnt;
S
Steven Rostedt 已提交
997 998 999 1000

	return cnt;
}

1001 1002 1003 1004
static ssize_t
ftrace_profile_read(struct file *filp, char __user *ubuf,
		     size_t cnt, loff_t *ppos)
{
1005
	char buf[64];		/* big enough to hold a number */
1006 1007 1008 1009 1010 1011
	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
1012 1013 1014 1015
static const struct file_operations ftrace_profile_fops = {
	.open		= tracing_open_generic,
	.read		= ftrace_profile_read,
	.write		= ftrace_profile_write,
1016
	.llseek		= default_llseek,
S
Steven Rostedt 已提交
1017 1018
};

1019 1020
/* used to initialize the real stat files */
static struct tracer_stat function_stats __initdata = {
1021 1022 1023 1024 1025 1026
	.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
1027 1028
};

1029
static __init void ftrace_profile_tracefs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
1030
{
1031
	struct ftrace_profile_stat *stat;
S
Steven Rostedt 已提交
1032
	struct dentry *entry;
1033
	char *name;
S
Steven Rostedt 已提交
1034
	int ret;
1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
	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
1063 1064
	}

1065
	entry = tracefs_create_file("function_profile_enabled", 0644,
S
Steven Rostedt 已提交
1066 1067
				    d_tracer, NULL, &ftrace_profile_fops);
	if (!entry)
1068
		pr_warning("Could not create tracefs "
S
Steven Rostedt 已提交
1069 1070 1071 1072
			   "'function_profile_enabled' entry\n");
}

#else /* CONFIG_FUNCTION_PROFILER */
1073
static __init void ftrace_profile_tracefs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
1074 1075 1076 1077
{
}
#endif /* CONFIG_FUNCTION_PROFILER */

1078 1079
static struct pid * const ftrace_swapper_pid = &init_struct_pid;

1080 1081 1082 1083 1084 1085
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static int ftrace_graph_active;
#else
# define ftrace_graph_active 0
#endif

1086 1087
#ifdef CONFIG_DYNAMIC_FTRACE

1088 1089
static struct ftrace_ops *removed_ops;

1090 1091 1092 1093 1094 1095
/*
 * Set when doing a global update, like enabling all recs or disabling them.
 * It is not set when just updating a single ftrace_ops.
 */
static bool update_all_ops;

1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107
#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;
1108
	struct list_head	free_list;
1109 1110
};

1111 1112 1113 1114 1115 1116 1117 1118 1119
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;
1120
	struct rcu_head		rcu;
1121 1122
};

1123 1124 1125 1126 1127 1128 1129 1130 1131
/*
 * 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,
1132
};
1133
#define EMPTY_HASH	((struct ftrace_hash *)&empty_hash)
1134

1135
static struct ftrace_ops global_ops = {
1136 1137 1138 1139 1140 1141
	.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,
1142 1143
};

1144 1145
/*
 * This is used by __kernel_text_address() to return true if the
1146
 * address is on a dynamically allocated trampoline that would
1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180
 * not return true for either core_kernel_text() or
 * is_module_text_address().
 */
bool is_ftrace_trampoline(unsigned long addr)
{
	struct ftrace_ops *op;
	bool ret = false;

	/*
	 * Some of the ops may be dynamically allocated,
	 * they are freed after a synchronize_sched().
	 */
	preempt_disable_notrace();

	do_for_each_ftrace_op(op, ftrace_ops_list) {
		/*
		 * This is to check for dynamically allocated trampolines.
		 * Trampolines that are in kernel text will have
		 * core_kernel_text() return true.
		 */
		if (op->trampoline && op->trampoline_size)
			if (addr >= op->trampoline &&
			    addr < op->trampoline + op->trampoline_size) {
				ret = true;
				goto out;
			}
	} while_for_each_ftrace_op(op);

 out:
	preempt_enable_notrace();

	return ret;
}

1181 1182
struct ftrace_page {
	struct ftrace_page	*next;
1183
	struct dyn_ftrace	*records;
1184
	int			index;
1185
	int			size;
1186 1187
};

1188 1189
#define ENTRY_SIZE sizeof(struct dyn_ftrace)
#define ENTRIES_PER_PAGE (PAGE_SIZE / ENTRY_SIZE)
1190 1191 1192 1193 1194 1195 1196

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

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

1197
static bool __always_inline ftrace_hash_empty(struct ftrace_hash *hash)
1198 1199 1200 1201
{
	return !hash || !hash->count;
}

1202 1203 1204 1205 1206 1207 1208
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;

1209
	if (ftrace_hash_empty(hash))
1210 1211 1212 1213 1214 1215 1216 1217 1218
		return NULL;

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

	hhd = &hash->buckets[key];

1219
	hlist_for_each_entry_rcu_notrace(entry, hhd, hlist) {
1220 1221 1222 1223 1224 1225
		if (entry->ip == ip)
			return entry;
	}
	return NULL;
}

1226 1227
static void __add_hash_entry(struct ftrace_hash *hash,
			     struct ftrace_func_entry *entry)
1228 1229 1230 1231 1232
{
	struct hlist_head *hhd;
	unsigned long key;

	if (hash->size_bits)
1233
		key = hash_long(entry->ip, hash->size_bits);
1234 1235 1236 1237 1238 1239
	else
		key = 0;

	hhd = &hash->buckets[key];
	hlist_add_head(&entry->hlist, hhd);
	hash->count++;
1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251
}

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);
1252 1253 1254 1255 1256

	return 0;
}

static void
1257
free_hash_entry(struct ftrace_hash *hash,
1258 1259 1260 1261 1262 1263 1264
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	kfree(entry);
	hash->count--;
}

1265 1266 1267 1268 1269 1270 1271 1272
static void
remove_hash_entry(struct ftrace_hash *hash,
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	hash->count--;
}

1273 1274 1275
static void ftrace_hash_clear(struct ftrace_hash *hash)
{
	struct hlist_head *hhd;
1276
	struct hlist_node *tn;
1277 1278 1279 1280
	struct ftrace_func_entry *entry;
	int size = 1 << hash->size_bits;
	int i;

1281 1282 1283
	if (!hash->count)
		return;

1284 1285
	for (i = 0; i < size; i++) {
		hhd = &hash->buckets[i];
1286
		hlist_for_each_entry_safe(entry, tn, hhd, hlist)
1287
			free_hash_entry(hash, entry);
1288 1289 1290 1291
	}
	FTRACE_WARN_ON(hash->count);
}

1292 1293 1294 1295 1296 1297 1298 1299 1300
static void free_ftrace_hash(struct ftrace_hash *hash)
{
	if (!hash || hash == EMPTY_HASH)
		return;
	ftrace_hash_clear(hash);
	kfree(hash->buckets);
	kfree(hash);
}

1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
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);
}

1316 1317
void ftrace_free_filter(struct ftrace_ops *ops)
{
1318
	ftrace_ops_init(ops);
1319 1320
	free_ftrace_hash(ops->func_hash->filter_hash);
	free_ftrace_hash(ops->func_hash->notrace_hash);
1321 1322
}

1323 1324 1325 1326 1327 1328 1329 1330 1331 1332
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;
1333
	hash->buckets = kcalloc(size, sizeof(*hash->buckets), GFP_KERNEL);
1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358

	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? */
1359
	if (ftrace_hash_empty(hash))
1360 1361 1362 1363
		return new_hash;

	size = 1 << hash->size_bits;
	for (i = 0; i < size; i++) {
1364
		hlist_for_each_entry(entry, &hash->buckets[i], hlist) {
1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379
			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;
}

1380
static void
1381
ftrace_hash_rec_disable_modify(struct ftrace_ops *ops, int filter_hash);
1382
static void
1383
ftrace_hash_rec_enable_modify(struct ftrace_ops *ops, int filter_hash);
1384

1385 1386 1387
static int ftrace_hash_ipmodify_update(struct ftrace_ops *ops,
				       struct ftrace_hash *new_hash);

1388
static int
1389 1390
ftrace_hash_move(struct ftrace_ops *ops, int enable,
		 struct ftrace_hash **dst, struct ftrace_hash *src)
1391 1392
{
	struct ftrace_func_entry *entry;
1393
	struct hlist_node *tn;
1394
	struct hlist_head *hhd;
1395
	struct ftrace_hash *new_hash;
1396 1397
	int size = src->count;
	int bits = 0;
1398
	int ret;
1399 1400
	int i;

1401 1402 1403 1404
	/* Reject setting notrace hash on IPMODIFY ftrace_ops */
	if (ops->flags & FTRACE_OPS_FL_IPMODIFY && !enable)
		return -EINVAL;

1405 1406 1407 1408 1409
	/*
	 * If the new source is empty, just free dst and assign it
	 * the empty_hash.
	 */
	if (!src->count) {
1410 1411
		new_hash = EMPTY_HASH;
		goto update;
1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
	}

	/*
	 * 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;

1424 1425
	new_hash = alloc_ftrace_hash(bits);
	if (!new_hash)
1426
		return -ENOMEM;
1427 1428 1429 1430

	size = 1 << src->size_bits;
	for (i = 0; i < size; i++) {
		hhd = &src->buckets[i];
1431
		hlist_for_each_entry_safe(entry, tn, hhd, hlist) {
1432
			remove_hash_entry(src, entry);
1433
			__add_hash_entry(new_hash, entry);
1434 1435 1436
		}
	}

1437
update:
1438 1439 1440 1441 1442 1443 1444 1445 1446 1447
	/* Make sure this can be applied if it is IPMODIFY ftrace_ops */
	if (enable) {
		/* IPMODIFY should be updated only when filter_hash updating */
		ret = ftrace_hash_ipmodify_update(ops, new_hash);
		if (ret < 0) {
			free_ftrace_hash(new_hash);
			return ret;
		}
	}

1448 1449 1450 1451
	/*
	 * Remove the current set, update the hash and add
	 * them back.
	 */
1452
	ftrace_hash_rec_disable_modify(ops, enable);
1453

1454 1455
	rcu_assign_pointer(*dst, new_hash);

1456
	ftrace_hash_rec_enable_modify(ops, enable);
1457

1458
	return 0;
1459 1460
}

1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475
static bool hash_contains_ip(unsigned long ip,
			     struct ftrace_ops_hash *hash)
{
	/*
	 * The function record is a match if it exists in the filter
	 * hash and not in the notrace hash. Note, an emty hash is
	 * considered a match for the filter hash, but an empty
	 * notrace hash is considered not in the notrace hash.
	 */
	return (ftrace_hash_empty(hash->filter_hash) ||
		ftrace_lookup_ip(hash->filter_hash, ip)) &&
		(ftrace_hash_empty(hash->notrace_hash) ||
		 !ftrace_lookup_ip(hash->notrace_hash, ip));
}

1476 1477 1478 1479 1480 1481 1482 1483
/*
 * 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.
1484 1485 1486
 *
 * This needs to be called with preemption disabled as
 * the hashes are freed with call_rcu_sched().
1487 1488
 */
static int
1489
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
1490
{
1491
	struct ftrace_ops_hash hash;
1492 1493
	int ret;

1494 1495 1496 1497 1498 1499 1500 1501 1502 1503
#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

1504 1505
	hash.filter_hash = rcu_dereference_raw_notrace(ops->func_hash->filter_hash);
	hash.notrace_hash = rcu_dereference_raw_notrace(ops->func_hash->notrace_hash);
1506

1507
	if (hash_contains_ip(ip, &hash))
1508 1509 1510 1511 1512 1513 1514
		ret = 1;
	else
		ret = 0;

	return ret;
}

1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528
/*
 * 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()		\
		}				\
	}

1529 1530 1531

static int ftrace_cmp_recs(const void *a, const void *b)
{
1532 1533
	const struct dyn_ftrace *key = a;
	const struct dyn_ftrace *rec = b;
1534

1535
	if (key->flags < rec->ip)
1536
		return -1;
1537 1538
	if (key->ip >= rec->ip + MCOUNT_INSN_SIZE)
		return 1;
1539 1540 1541
	return 0;
}

1542
static unsigned long ftrace_location_range(unsigned long start, unsigned long end)
1543 1544 1545
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1546
	struct dyn_ftrace key;
1547

1548 1549
	key.ip = start;
	key.flags = end;	/* overload flags, as it is unsigned long */
1550 1551

	for (pg = ftrace_pages_start; pg; pg = pg->next) {
1552 1553
		if (end < pg->records[0].ip ||
		    start >= (pg->records[pg->index - 1].ip + MCOUNT_INSN_SIZE))
1554
			continue;
1555 1556 1557 1558
		rec = bsearch(&key, pg->records, pg->index,
			      sizeof(struct dyn_ftrace),
			      ftrace_cmp_recs);
		if (rec)
1559
			return rec->ip;
1560
	}
1561 1562 1563 1564

	return 0;
}

1565 1566 1567 1568
/**
 * ftrace_location - return true if the ip giving is a traced location
 * @ip: the instruction pointer to check
 *
1569
 * Returns rec->ip if @ip given is a pointer to a ftrace location.
1570 1571 1572 1573
 * 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.
 */
1574
unsigned long ftrace_location(unsigned long ip)
1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588
{
	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.
 */
1589
int ftrace_text_reserved(const void *start, const void *end)
1590
{
1591 1592 1593 1594 1595 1596
	unsigned long ret;

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

	return (int)!!ret;
1597 1598
}

1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618
/* 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;
}

1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645
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) {
1646 1647
		hash = ops->func_hash->filter_hash;
		other_hash = ops->func_hash->notrace_hash;
1648
		if (ftrace_hash_empty(hash))
1649 1650 1651
			all = 1;
	} else {
		inc = !inc;
1652 1653
		hash = ops->func_hash->notrace_hash;
		other_hash = ops->func_hash->filter_hash;
1654 1655 1656 1657
		/*
		 * If the notrace hash has no items,
		 * then there's nothing to do.
		 */
1658
		if (ftrace_hash_empty(hash))
1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671
			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.
			 */
1672
			if (!other_hash || !ftrace_lookup_ip(other_hash, rec->ip))
1673 1674
				match = 1;
		} else {
1675 1676
			in_hash = !!ftrace_lookup_ip(hash, rec->ip);
			in_other_hash = !!ftrace_lookup_ip(other_hash, rec->ip);
1677 1678

			/*
1679 1680
			 * If filter_hash is set, we want to match all functions
			 * that are in the hash but not in the other hash.
1681
			 *
1682 1683 1684 1685 1686
			 * 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.
1687 1688 1689 1690
			 */
			if (filter_hash && in_hash && !in_other_hash)
				match = 1;
			else if (!filter_hash && in_hash &&
1691
				 (in_other_hash || ftrace_hash_empty(other_hash)))
1692 1693 1694 1695 1696 1697 1698
				match = 1;
		}
		if (!match)
			continue;

		if (inc) {
			rec->flags++;
1699
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == FTRACE_REF_MAX))
1700
				return;
1701 1702 1703 1704 1705 1706

			/*
			 * 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.
			 */
1707
			if (ftrace_rec_count(rec) == 1 && ops->trampoline)
1708
				rec->flags |= FTRACE_FL_TRAMP;
1709
			else
1710 1711 1712
				/*
				 * If we are adding another function callback
				 * to this function, and the previous had a
1713 1714
				 * custom trampoline in use, then we need to go
				 * back to the default trampoline.
1715
				 */
1716
				rec->flags &= ~FTRACE_FL_TRAMP;
1717

1718 1719 1720 1721 1722 1723
			/*
			 * 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;
1724
		} else {
1725
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == 0))
1726 1727
				return;
			rec->flags--;
1728

1729 1730 1731 1732 1733 1734
			/*
			 * 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.
			 */
1735
			if (ftrace_rec_count(rec) > 0 &&
1736 1737 1738 1739 1740
			    rec->flags & FTRACE_FL_REGS &&
			    ops->flags & FTRACE_OPS_FL_SAVE_REGS) {
				if (!test_rec_ops_needs_regs(rec))
					rec->flags &= ~FTRACE_FL_REGS;
			}
1741

1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752
			/*
			 * If the rec had TRAMP enabled, then it needs to
			 * be cleared. As TRAMP can only be enabled iff
			 * there is only a single ops attached to it.
			 * In otherwords, always disable it on decrementing.
			 * In the future, we may set it if rec count is
			 * decremented to one, and the ops that is left
			 * has a trampoline.
			 */
			rec->flags &= ~FTRACE_FL_TRAMP;

1753 1754 1755 1756
			/*
			 * flags will be cleared in ftrace_check_record()
			 * if rec count is zero.
			 */
1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776
		}
		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);
}

1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811
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);
}

1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919
/*
 * Try to update IPMODIFY flag on each ftrace_rec. Return 0 if it is OK
 * or no-needed to update, -EBUSY if it detects a conflict of the flag
 * on a ftrace_rec, and -EINVAL if the new_hash tries to trace all recs.
 * Note that old_hash and new_hash has below meanings
 *  - If the hash is NULL, it hits all recs (if IPMODIFY is set, this is rejected)
 *  - If the hash is EMPTY_HASH, it hits nothing
 *  - Anything else hits the recs which match the hash entries.
 */
static int __ftrace_hash_update_ipmodify(struct ftrace_ops *ops,
					 struct ftrace_hash *old_hash,
					 struct ftrace_hash *new_hash)
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec, *end = NULL;
	int in_old, in_new;

	/* Only update if the ops has been registered */
	if (!(ops->flags & FTRACE_OPS_FL_ENABLED))
		return 0;

	if (!(ops->flags & FTRACE_OPS_FL_IPMODIFY))
		return 0;

	/*
	 * Since the IPMODIFY is a very address sensitive action, we do not
	 * allow ftrace_ops to set all functions to new hash.
	 */
	if (!new_hash || !old_hash)
		return -EINVAL;

	/* Update rec->flags */
	do_for_each_ftrace_rec(pg, rec) {
		/* We need to update only differences of filter_hash */
		in_old = !!ftrace_lookup_ip(old_hash, rec->ip);
		in_new = !!ftrace_lookup_ip(new_hash, rec->ip);
		if (in_old == in_new)
			continue;

		if (in_new) {
			/* New entries must ensure no others are using it */
			if (rec->flags & FTRACE_FL_IPMODIFY)
				goto rollback;
			rec->flags |= FTRACE_FL_IPMODIFY;
		} else /* Removed entry */
			rec->flags &= ~FTRACE_FL_IPMODIFY;
	} while_for_each_ftrace_rec();

	return 0;

rollback:
	end = rec;

	/* Roll back what we did above */
	do_for_each_ftrace_rec(pg, rec) {
		if (rec == end)
			goto err_out;

		in_old = !!ftrace_lookup_ip(old_hash, rec->ip);
		in_new = !!ftrace_lookup_ip(new_hash, rec->ip);
		if (in_old == in_new)
			continue;

		if (in_new)
			rec->flags &= ~FTRACE_FL_IPMODIFY;
		else
			rec->flags |= FTRACE_FL_IPMODIFY;
	} while_for_each_ftrace_rec();

err_out:
	return -EBUSY;
}

static int ftrace_hash_ipmodify_enable(struct ftrace_ops *ops)
{
	struct ftrace_hash *hash = ops->func_hash->filter_hash;

	if (ftrace_hash_empty(hash))
		hash = NULL;

	return __ftrace_hash_update_ipmodify(ops, EMPTY_HASH, hash);
}

/* Disabling always succeeds */
static void ftrace_hash_ipmodify_disable(struct ftrace_ops *ops)
{
	struct ftrace_hash *hash = ops->func_hash->filter_hash;

	if (ftrace_hash_empty(hash))
		hash = NULL;

	__ftrace_hash_update_ipmodify(ops, hash, EMPTY_HASH);
}

static int ftrace_hash_ipmodify_update(struct ftrace_ops *ops,
				       struct ftrace_hash *new_hash)
{
	struct ftrace_hash *old_hash = ops->func_hash->filter_hash;

	if (ftrace_hash_empty(old_hash))
		old_hash = NULL;

	if (ftrace_hash_empty(new_hash))
		new_hash = NULL;

	return __ftrace_hash_update_ipmodify(ops, old_hash, new_hash);
}

1920 1921 1922 1923 1924 1925 1926 1927 1928 1929
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]);
}

1930 1931 1932
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec);

1933 1934 1935
/**
 * ftrace_bug - report and shutdown function tracer
 * @failed: The failed type (EFAULT, EINVAL, EPERM)
1936
 * @rec: The record that failed
1937 1938 1939 1940 1941 1942 1943 1944
 *
 * 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
 */
1945
void ftrace_bug(int failed, struct dyn_ftrace *rec)
1946
{
1947 1948
	unsigned long ip = rec ? rec->ip : 0;

1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959
	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);
1960
		pr_cont("\n");
1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971
		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);
	}
1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989
	if (rec) {
		struct ftrace_ops *ops = NULL;

		pr_info("ftrace record flags: %lx\n", rec->flags);
		pr_cont(" (%ld)%s", ftrace_rec_count(rec),
			rec->flags & FTRACE_FL_REGS ? " R" : "  ");
		if (rec->flags & FTRACE_FL_TRAMP_EN) {
			ops = ftrace_find_tramp_ops_any(rec);
			if (ops)
				pr_cont("\ttramp: %pS",
					(void *)ops->trampoline);
			else
				pr_cont("\ttramp: ERROR!");

		}
		ip = ftrace_get_addr_curr(rec);
		pr_cont(" expected tramp: %lx\n", ip);
	}
1990 1991
}

1992
static int ftrace_check_record(struct dyn_ftrace *rec, int enable, int update)
1993
{
1994
	unsigned long flag = 0UL;
1995

S
Steven Rostedt 已提交
1996
	/*
1997
	 * If we are updating calls:
S
Steven Rostedt 已提交
1998
	 *
1999 2000
	 *   If the record has a ref count, then we need to enable it
	 *   because someone is using it.
S
Steven Rostedt 已提交
2001
	 *
2002 2003
	 *   Otherwise we make sure its disabled.
	 *
2004
	 * If we are disabling calls, then disable all records that
2005
	 * are enabled.
S
Steven Rostedt 已提交
2006
	 */
2007
	if (enable && ftrace_rec_count(rec))
2008
		flag = FTRACE_FL_ENABLED;
S
Steven Rostedt 已提交
2009

2010
	/*
2011 2012 2013
	 * 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.
2014
	 */
2015 2016 2017 2018 2019 2020 2021 2022 2023
	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;
	}
2024

2025 2026
	/* If the state of this record hasn't changed, then do nothing */
	if ((rec->flags & FTRACE_FL_ENABLED) == flag)
2027
		return FTRACE_UPDATE_IGNORE;
S
Steven Rostedt 已提交
2028

2029
	if (flag) {
2030 2031 2032 2033
		/* Save off if rec is being enabled (for return value) */
		flag ^= rec->flags & FTRACE_FL_ENABLED;

		if (update) {
2034
			rec->flags |= FTRACE_FL_ENABLED;
2035 2036 2037 2038 2039 2040
			if (flag & FTRACE_FL_REGS) {
				if (rec->flags & FTRACE_FL_REGS)
					rec->flags |= FTRACE_FL_REGS_EN;
				else
					rec->flags &= ~FTRACE_FL_REGS_EN;
			}
2041 2042 2043 2044 2045 2046
			if (flag & FTRACE_FL_TRAMP) {
				if (rec->flags & FTRACE_FL_TRAMP)
					rec->flags |= FTRACE_FL_TRAMP_EN;
				else
					rec->flags &= ~FTRACE_FL_TRAMP_EN;
			}
2047 2048 2049 2050 2051 2052 2053
		}

		/*
		 * 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
2054
		 *   from the save regs, to a non-save regs function or
2055
		 *   vice versa, or from a trampoline call.
2056 2057 2058
		 */
		if (flag & FTRACE_FL_ENABLED)
			return FTRACE_UPDATE_MAKE_CALL;
2059 2060

		return FTRACE_UPDATE_MODIFY_CALL;
2061 2062
	}

2063 2064
	if (update) {
		/* If there's no more users, clear all flags */
2065
		if (!ftrace_rec_count(rec))
2066 2067
			rec->flags = 0;
		else
2068 2069 2070 2071 2072 2073
			/*
			 * Just disable the record, but keep the ops TRAMP
			 * and REGS states. The _EN flags must be disabled though.
			 */
			rec->flags &= ~(FTRACE_FL_ENABLED | FTRACE_FL_TRAMP_EN |
					FTRACE_FL_REGS_EN);
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

	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);
}

2106 2107 2108 2109
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;
2110
	unsigned long ip = rec->ip;
2111 2112 2113 2114 2115 2116

	do_for_each_ftrace_op(op, ftrace_ops_list) {

		if (!op->trampoline)
			continue;

2117
		if (hash_contains_ip(ip, op->func_hash))
2118 2119 2120 2121 2122 2123
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

2124 2125 2126 2127
static struct ftrace_ops *
ftrace_find_tramp_ops_curr(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;
2128
	unsigned long ip = rec->ip;
2129

2130 2131 2132 2133 2134 2135 2136 2137
	/*
	 * Need to check removed ops first.
	 * If they are being removed, and this rec has a tramp,
	 * and this rec is in the ops list, then it would be the
	 * one with the tramp.
	 */
	if (removed_ops) {
		if (hash_contains_ip(ip, &removed_ops->old_hash))
2138 2139 2140
			return removed_ops;
	}

2141 2142 2143 2144 2145 2146 2147
	/*
	 * Need to find the current trampoline for a rec.
	 * Now, a trampoline is only attached to a rec if there
	 * was a single 'ops' attached to it. But this can be called
	 * when we are adding another op to the rec or removing the
	 * current one. Thus, if the op is being added, we can
	 * ignore it because it hasn't attached itself to the rec
2148 2149 2150 2151 2152 2153 2154 2155 2156 2157
	 * yet.
	 *
	 * If an ops is being modified (hooking to different functions)
	 * then we don't care about the new functions that are being
	 * added, just the old ones (that are probably being removed).
	 *
	 * If we are adding an ops to a function that already is using
	 * a trampoline, it needs to be removed (trampolines are only
	 * for single ops connected), then an ops that is not being
	 * modified also needs to be checked.
2158
	 */
2159
	do_for_each_ftrace_op(op, ftrace_ops_list) {
2160 2161 2162 2163 2164 2165 2166 2167 2168

		if (!op->trampoline)
			continue;

		/*
		 * If the ops is being added, it hasn't gotten to
		 * the point to be removed from this tree yet.
		 */
		if (op->flags & FTRACE_OPS_FL_ADDING)
2169 2170
			continue;

2171

2172
		/*
2173 2174 2175
		 * If the ops is being modified and is in the old
		 * hash, then it is probably being removed from this
		 * function.
2176 2177 2178
		 */
		if ((op->flags & FTRACE_OPS_FL_MODIFYING) &&
		    hash_contains_ip(ip, &op->old_hash))
2179
			return op;
2180 2181 2182 2183 2184 2185 2186 2187
		/*
		 * If the ops is not being added or modified, and it's
		 * in its normal filter hash, then this must be the one
		 * we want!
		 */
		if (!(op->flags & FTRACE_OPS_FL_MODIFYING) &&
		    hash_contains_ip(ip, op->func_hash))
			return op;
2188 2189 2190 2191 2192 2193 2194 2195 2196 2197

	} 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;
2198
	unsigned long ip = rec->ip;
2199 2200 2201

	do_for_each_ftrace_op(op, ftrace_ops_list) {
		/* pass rec in as regs to have non-NULL val */
2202
		if (hash_contains_ip(ip, op->func_hash))
2203 2204 2205 2206 2207 2208
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220
/**
 * 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)
{
2221 2222 2223 2224 2225 2226
	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)) {
2227 2228
			pr_warn("Bad trampoline accounting at: %p (%pS) (%lx)\n",
				(void *)rec->ip, (void *)rec->ip, rec->flags);
2229 2230 2231 2232 2233 2234
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252
	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)
{
2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266
	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;
	}

2267 2268 2269 2270 2271 2272
	if (rec->flags & FTRACE_FL_REGS_EN)
		return (unsigned long)FTRACE_REGS_ADDR;
	else
		return (unsigned long)FTRACE_ADDR;
}

2273 2274 2275
static int
__ftrace_replace_code(struct dyn_ftrace *rec, int enable)
{
2276
	unsigned long ftrace_old_addr;
2277 2278 2279
	unsigned long ftrace_addr;
	int ret;

2280
	ftrace_addr = ftrace_get_addr_new(rec);
2281

2282 2283 2284 2285
	/* 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);
2286

2287 2288 2289 2290 2291
	switch (ret) {
	case FTRACE_UPDATE_IGNORE:
		return 0;

	case FTRACE_UPDATE_MAKE_CALL:
2292
		return ftrace_make_call(rec, ftrace_addr);
2293 2294

	case FTRACE_UPDATE_MAKE_NOP:
2295
		return ftrace_make_nop(NULL, rec, ftrace_old_addr);
2296 2297 2298

	case FTRACE_UPDATE_MODIFY_CALL:
		return ftrace_modify_call(rec, ftrace_old_addr, ftrace_addr);
2299 2300
	}

2301
	return -1; /* unknow ftrace bug */
2302 2303
}

2304
void __weak ftrace_replace_code(int enable)
2305 2306 2307
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
S
Steven Rostedt 已提交
2308
	int failed;
2309

2310 2311 2312
	if (unlikely(ftrace_disabled))
		return;

2313
	do_for_each_ftrace_rec(pg, rec) {
2314
		failed = __ftrace_replace_code(rec, enable);
2315
		if (failed) {
2316
			ftrace_bug(failed, rec);
2317 2318
			/* Stop processing */
			return;
2319
		}
2320
	} while_for_each_ftrace_rec();
2321 2322
}

2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394
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];
}

2395
static int
2396
ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec)
2397
{
2398
	int ret;
2399

2400 2401 2402
	if (unlikely(ftrace_disabled))
		return 0;

2403
	ret = ftrace_make_nop(mod, rec, MCOUNT_ADDR);
2404
	if (ret) {
2405
		ftrace_bug(ret, rec);
2406
		return 0;
2407
	}
2408
	return 1;
2409 2410
}

2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428
/*
 * 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;
}

2429
void ftrace_modify_all_code(int command)
2430
{
2431
	int update = command & FTRACE_UPDATE_TRACE_FUNC;
2432
	int err = 0;
2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443

	/*
	 * 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.
	 */
2444 2445 2446 2447 2448
	if (update) {
		err = ftrace_update_ftrace_func(ftrace_ops_list_func);
		if (FTRACE_WARN_ON(err))
			return;
	}
2449

2450
	if (command & FTRACE_UPDATE_CALLS)
2451
		ftrace_replace_code(1);
2452
	else if (command & FTRACE_DISABLE_CALLS)
2453 2454
		ftrace_replace_code(0);

2455 2456 2457 2458 2459 2460
	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);
2461 2462 2463
		err = ftrace_update_ftrace_func(ftrace_trace_function);
		if (FTRACE_WARN_ON(err))
			return;
2464
	}
2465

2466
	if (command & FTRACE_START_FUNC_RET)
2467
		err = ftrace_enable_ftrace_graph_caller();
2468
	else if (command & FTRACE_STOP_FUNC_RET)
2469 2470
		err = ftrace_disable_ftrace_graph_caller();
	FTRACE_WARN_ON(err);
2471 2472 2473 2474 2475 2476 2477
}

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

	ftrace_modify_all_code(*command);
2478

2479
	return 0;
2480 2481
}

2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505
/**
 * ftrace_run_stop_machine, go back to the stop machine method
 * @command: The command to tell ftrace what to do
 *
 * If an arch needs to fall back to the stop machine method, the
 * it can call this function.
 */
void ftrace_run_stop_machine(int command)
{
	stop_machine(__ftrace_modify_code, &command, NULL);
}

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

I
Ingo Molnar 已提交
2506
static void ftrace_run_update_code(int command)
2507
{
2508 2509 2510 2511 2512 2513 2514
	int ret;

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

2515 2516 2517 2518 2519 2520 2521 2522
	/*
	 * 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);

2523 2524
	ret = ftrace_arch_code_modify_post_process();
	FTRACE_WARN_ON(ret);
2525 2526
}

2527
static void ftrace_run_modify_code(struct ftrace_ops *ops, int command,
2528
				   struct ftrace_ops_hash *old_hash)
2529 2530
{
	ops->flags |= FTRACE_OPS_FL_MODIFYING;
2531 2532
	ops->old_hash.filter_hash = old_hash->filter_hash;
	ops->old_hash.notrace_hash = old_hash->notrace_hash;
2533
	ftrace_run_update_code(command);
2534
	ops->old_hash.filter_hash = NULL;
2535
	ops->old_hash.notrace_hash = NULL;
2536 2537 2538
	ops->flags &= ~FTRACE_OPS_FL_MODIFYING;
}

2539
static ftrace_func_t saved_ftrace_func;
2540
static int ftrace_start_up;
2541

2542 2543 2544 2545
void __weak arch_ftrace_trampoline_free(struct ftrace_ops *ops)
{
}

2546 2547 2548 2549 2550
static void control_ops_free(struct ftrace_ops *ops)
{
	free_percpu(ops->disabled);
}

2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562
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);
}
2563

2564 2565 2566 2567 2568 2569 2570
static void ftrace_startup_all(int command)
{
	update_all_ops = true;
	ftrace_startup_enable(command);
	update_all_ops = false;
}

2571
static int ftrace_startup(struct ftrace_ops *ops, int command)
2572
{
2573
	int ret;
2574

2575
	if (unlikely(ftrace_disabled))
2576
		return -ENODEV;
2577

2578 2579 2580 2581
	ret = __register_ftrace_function(ops);
	if (ret)
		return ret;

2582
	ftrace_start_up++;
2583
	command |= FTRACE_UPDATE_CALLS;
2584

2585 2586 2587 2588 2589 2590 2591 2592 2593
	/*
	 * Note that ftrace probes uses this to start up
	 * and modify functions it will probe. But we still
	 * set the ADDING flag for modification, as probes
	 * do not have trampolines. If they add them in the
	 * future, then the probes will need to distinguish
	 * between adding and updating probes.
	 */
	ops->flags |= FTRACE_OPS_FL_ENABLED | FTRACE_OPS_FL_ADDING;
2594

2595 2596 2597 2598 2599 2600 2601 2602 2603
	ret = ftrace_hash_ipmodify_enable(ops);
	if (ret < 0) {
		/* Rollback registration process */
		__unregister_ftrace_function(ops);
		ftrace_start_up--;
		ops->flags &= ~FTRACE_OPS_FL_ENABLED;
		return ret;
	}

2604
	ftrace_hash_rec_enable(ops, 1);
2605

2606
	ftrace_startup_enable(command);
2607

2608 2609
	ops->flags &= ~FTRACE_OPS_FL_ADDING;

2610
	return 0;
2611 2612
}

2613
static int ftrace_shutdown(struct ftrace_ops *ops, int command)
2614
{
2615
	int ret;
2616

2617
	if (unlikely(ftrace_disabled))
2618 2619 2620 2621 2622
		return -ENODEV;

	ret = __unregister_ftrace_function(ops);
	if (ret)
		return ret;
2623

2624
	ftrace_start_up--;
2625 2626 2627 2628 2629 2630 2631
	/*
	 * 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);

2632 2633
	/* Disabling ipmodify never fails */
	ftrace_hash_ipmodify_disable(ops);
2634
	ftrace_hash_rec_disable(ops, 1);
2635

2636
	ops->flags &= ~FTRACE_OPS_FL_ENABLED;
2637

2638
	command |= FTRACE_UPDATE_CALLS;
2639

2640 2641 2642 2643
	if (saved_ftrace_func != ftrace_trace_function) {
		saved_ftrace_func = ftrace_trace_function;
		command |= FTRACE_UPDATE_TRACE_FUNC;
	}
2644

2645 2646 2647 2648 2649 2650 2651 2652 2653
	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);
2654
		return 0;
2655
	}
2656

2657 2658 2659 2660
	/*
	 * If the ops uses a trampoline, then it needs to be
	 * tested first on update.
	 */
2661
	ops->flags |= FTRACE_OPS_FL_REMOVING;
2662 2663
	removed_ops = ops;

2664 2665 2666 2667
	/* The trampoline logic checks the old hashes */
	ops->old_hash.filter_hash = ops->func_hash->filter_hash;
	ops->old_hash.notrace_hash = ops->func_hash->notrace_hash;

2668
	ftrace_run_update_code(command);
2669

2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684
	/*
	 * If there's no more ops registered with ftrace, run a
	 * sanity check to make sure all rec flags are cleared.
	 */
	if (ftrace_ops_list == &ftrace_list_end) {
		struct ftrace_page *pg;
		struct dyn_ftrace *rec;

		do_for_each_ftrace_rec(pg, rec) {
			if (FTRACE_WARN_ON_ONCE(rec->flags))
				pr_warn("  %pS flags:%lx\n",
					(void *)rec->ip, rec->flags);
		} while_for_each_ftrace_rec();
	}

2685 2686 2687 2688
	ops->old_hash.filter_hash = NULL;
	ops->old_hash.notrace_hash = NULL;

	removed_ops = NULL;
2689
	ops->flags &= ~FTRACE_OPS_FL_REMOVING;
2690

2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707
	/*
	 * 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);

2708 2709
		arch_ftrace_trampoline_free(ops);

2710 2711 2712 2713
		if (ops->flags & FTRACE_OPS_FL_CONTROL)
			control_ops_free(ops);
	}

2714
	return 0;
2715 2716
}

I
Ingo Molnar 已提交
2717
static void ftrace_startup_sysctl(void)
2718
{
2719 2720
	int command;

2721 2722 2723
	if (unlikely(ftrace_disabled))
		return;

2724 2725
	/* Force update next time */
	saved_ftrace_func = NULL;
2726
	/* ftrace_start_up is true if we want ftrace running */
2727 2728 2729 2730
	if (ftrace_start_up) {
		command = FTRACE_UPDATE_CALLS;
		if (ftrace_graph_active)
			command |= FTRACE_START_FUNC_RET;
2731
		ftrace_startup_enable(command);
2732
	}
2733 2734
}

I
Ingo Molnar 已提交
2735
static void ftrace_shutdown_sysctl(void)
2736
{
2737 2738
	int command;

2739 2740 2741
	if (unlikely(ftrace_disabled))
		return;

2742
	/* ftrace_start_up is true if ftrace is running */
2743 2744 2745 2746 2747 2748
	if (ftrace_start_up) {
		command = FTRACE_DISABLE_CALLS;
		if (ftrace_graph_active)
			command |= FTRACE_STOP_FUNC_RET;
		ftrace_run_update_code(command);
	}
2749 2750
}

2751 2752 2753
static cycle_t		ftrace_update_time;
unsigned long		ftrace_update_tot_cnt;

2754
static inline int ops_traces_mod(struct ftrace_ops *ops)
2755
{
2756 2757 2758 2759
	/*
	 * Filter_hash being empty will default to trace module.
	 * But notrace hash requires a test of individual module functions.
	 */
2760 2761
	return ftrace_hash_empty(ops->func_hash->filter_hash) &&
		ftrace_hash_empty(ops->func_hash->notrace_hash);
2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782
}

/*
 * 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 */
2783 2784
	if (!ftrace_hash_empty(ops->func_hash->filter_hash) &&
	    !ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip))
2785
		return 0;
2786

2787
	/* If in notrace hash, we ignore it too */
2788
	if (ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip))
2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804
		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;
2805 2806
}

2807
static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
2808
{
2809
	struct ftrace_page *pg;
2810
	struct dyn_ftrace *p;
2811
	cycle_t start, stop;
2812
	unsigned long update_cnt = 0;
2813
	unsigned long ref = 0;
2814
	bool test = false;
2815
	int i;
2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827

	/*
	 * 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) {
2828 2829 2830 2831 2832 2833
			if (ops->flags & FTRACE_OPS_FL_ENABLED) {
				if (ops_traces_mod(ops))
					ref++;
				else
					test = true;
			}
2834 2835
		}
	}
2836

I
Ingo Molnar 已提交
2837
	start = ftrace_now(raw_smp_processor_id());
2838

2839
	for (pg = new_pgs; pg; pg = pg->next) {
2840

2841
		for (i = 0; i < pg->index; i++) {
2842 2843
			int cnt = ref;

2844 2845 2846
			/* If something went wrong, bail without enabling anything */
			if (unlikely(ftrace_disabled))
				return -1;
2847

2848
			p = &pg->records[i];
2849 2850 2851
			if (test)
				cnt += referenced_filters(p);
			p->flags = cnt;
2852

2853 2854 2855 2856 2857 2858
			/*
			 * Do the initial record conversion from mcount jump
			 * to the NOP instructions.
			 */
			if (!ftrace_code_disable(mod, p))
				break;
2859

2860
			update_cnt++;
2861

2862 2863 2864 2865 2866 2867 2868 2869 2870
			/*
			 * 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.
			 */
2871
			if (ftrace_start_up && cnt) {
2872 2873
				int failed = __ftrace_replace_code(p, 1);
				if (failed)
2874
					ftrace_bug(failed, p);
2875
			}
2876
		}
2877 2878
	}

I
Ingo Molnar 已提交
2879
	stop = ftrace_now(raw_smp_processor_id());
2880
	ftrace_update_time = stop - start;
2881
	ftrace_update_tot_cnt += update_cnt;
2882

2883 2884 2885
	return 0;
}

2886
static int ftrace_allocate_records(struct ftrace_page *pg, int count)
2887
{
2888
	int order;
2889 2890
	int cnt;

2891 2892 2893 2894
	if (WARN_ON(!count))
		return -EINVAL;

	order = get_count_order(DIV_ROUND_UP(count, ENTRIES_PER_PAGE));
2895 2896

	/*
2897 2898
	 * We want to fill as much as possible. No more than a page
	 * may be empty.
2899
	 */
2900 2901
	while ((PAGE_SIZE << order) / ENTRY_SIZE >= count + ENTRIES_PER_PAGE)
		order--;
2902

2903 2904
 again:
	pg->records = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
2905

2906 2907 2908 2909 2910 2911 2912
	if (!pg->records) {
		/* if we can't allocate this size, try something smaller */
		if (!order)
			return -ENOMEM;
		order >>= 1;
		goto again;
	}
2913

2914 2915
	cnt = (PAGE_SIZE << order) / ENTRY_SIZE;
	pg->size = cnt;
2916

2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949
	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)
2950 2951
			break;

2952 2953 2954 2955
		pg->next = kzalloc(sizeof(*pg), GFP_KERNEL);
		if (!pg->next)
			goto free_pages;

2956 2957 2958
		pg = pg->next;
	}

2959 2960 2961
	return start_pg;

 free_pages:
2962 2963
	pg = start_pg;
	while (pg) {
2964 2965 2966 2967 2968 2969 2970 2971 2972 2973
		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;
}

2974 2975 2976
#define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */

struct ftrace_iterator {
2977
	loff_t				pos;
2978 2979 2980 2981 2982
	loff_t				func_pos;
	struct ftrace_page		*pg;
	struct dyn_ftrace		*func;
	struct ftrace_func_probe	*probe;
	struct trace_parser		parser;
2983
	struct ftrace_hash		*hash;
2984
	struct ftrace_ops		*ops;
2985 2986 2987
	int				hidx;
	int				idx;
	unsigned			flags;
2988 2989
};

2990
static void *
2991
t_hash_next(struct seq_file *m, loff_t *pos)
2992 2993
{
	struct ftrace_iterator *iter = m->private;
2994
	struct hlist_node *hnd = NULL;
2995 2996 2997
	struct hlist_head *hhd;

	(*pos)++;
2998
	iter->pos = *pos;
2999

3000 3001
	if (iter->probe)
		hnd = &iter->probe->node;
3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023
 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;
		}
	}

3024 3025 3026 3027 3028 3029
	if (WARN_ON_ONCE(!hnd))
		return NULL;

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

	return iter;
3030 3031 3032 3033 3034 3035
}

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
3036 3037
	loff_t l;

3038 3039 3040
	if (!(iter->flags & FTRACE_ITER_DO_HASH))
		return NULL;

3041 3042
	if (iter->func_pos > *pos)
		return NULL;
3043

L
Li Zefan 已提交
3044
	iter->hidx = 0;
3045
	for (l = 0; l <= (*pos - iter->func_pos); ) {
3046
		p = t_hash_next(m, &l);
L
Li Zefan 已提交
3047 3048 3049
		if (!p)
			break;
	}
3050 3051 3052
	if (!p)
		return NULL;

3053 3054 3055
	/* Only set this if we have an item */
	iter->flags |= FTRACE_ITER_HASH;

3056
	return iter;
3057 3058
}

3059 3060
static int
t_hash_show(struct seq_file *m, struct ftrace_iterator *iter)
3061
{
S
Steven Rostedt 已提交
3062
	struct ftrace_func_probe *rec;
3063

3064 3065 3066
	rec = iter->probe;
	if (WARN_ON_ONCE(!rec))
		return -EIO;
3067

3068 3069 3070
	if (rec->ops->print)
		return rec->ops->print(m, rec->ip, rec->ops, rec->data);

3071
	seq_printf(m, "%ps:%ps", (void *)rec->ip, (void *)rec->ops->func);
3072 3073 3074 3075 3076 3077 3078 3079

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

	return 0;
}

I
Ingo Molnar 已提交
3080
static void *
3081 3082 3083
t_next(struct seq_file *m, void *v, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
3084
	struct ftrace_ops *ops = iter->ops;
3085 3086
	struct dyn_ftrace *rec = NULL;

3087 3088 3089
	if (unlikely(ftrace_disabled))
		return NULL;

3090
	if (iter->flags & FTRACE_ITER_HASH)
3091
		return t_hash_next(m, pos);
3092

3093
	(*pos)++;
3094
	iter->pos = iter->func_pos = *pos;
3095

3096
	if (iter->flags & FTRACE_ITER_PRINTALL)
3097
		return t_hash_start(m, pos);
3098

3099 3100 3101 3102 3103 3104 3105 3106 3107
 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++];
3108
		if (((iter->flags & FTRACE_ITER_FILTER) &&
3109
		     !(ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip))) ||
S
Steven Rostedt 已提交
3110

3111
		    ((iter->flags & FTRACE_ITER_NOTRACE) &&
3112
		     !ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip)) ||
3113 3114

		    ((iter->flags & FTRACE_ITER_ENABLED) &&
3115
		     !(rec->flags & FTRACE_FL_ENABLED))) {
3116

3117 3118 3119 3120 3121
			rec = NULL;
			goto retry;
		}
	}

3122
	if (!rec)
3123
		return t_hash_start(m, pos);
3124 3125 3126 3127

	iter->func = rec;

	return iter;
3128 3129
}

3130 3131 3132 3133
static void reset_iter_read(struct ftrace_iterator *iter)
{
	iter->pos = 0;
	iter->func_pos = 0;
3134
	iter->flags &= ~(FTRACE_ITER_PRINTALL | FTRACE_ITER_HASH);
3135 3136 3137 3138 3139
}

static void *t_start(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
3140
	struct ftrace_ops *ops = iter->ops;
3141
	void *p = NULL;
3142
	loff_t l;
3143

3144
	mutex_lock(&ftrace_lock);
3145 3146 3147 3148

	if (unlikely(ftrace_disabled))
		return NULL;

3149 3150 3151 3152 3153 3154
	/*
	 * If an lseek was done, then reset and start from beginning.
	 */
	if (*pos < iter->pos)
		reset_iter_read(iter);

3155 3156 3157 3158 3159
	/*
	 * For set_ftrace_filter reading, if we have the filter
	 * off, we can short cut and just print out that all
	 * functions are enabled.
	 */
3160
	if ((iter->flags & FTRACE_ITER_FILTER &&
3161
	     ftrace_hash_empty(ops->func_hash->filter_hash)) ||
3162
	    (iter->flags & FTRACE_ITER_NOTRACE &&
3163
	     ftrace_hash_empty(ops->func_hash->notrace_hash))) {
3164
		if (*pos > 0)
3165
			return t_hash_start(m, pos);
3166
		iter->flags |= FTRACE_ITER_PRINTALL;
3167 3168
		/* reset in case of seek/pread */
		iter->flags &= ~FTRACE_ITER_HASH;
3169 3170 3171
		return iter;
	}

3172 3173 3174
	if (iter->flags & FTRACE_ITER_HASH)
		return t_hash_start(m, pos);

3175 3176 3177 3178 3179
	/*
	 * 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.
	 */
3180 3181 3182 3183 3184 3185
	iter->pg = ftrace_pages_start;
	iter->idx = 0;
	for (l = 0; l <= *pos; ) {
		p = t_next(m, p, &l);
		if (!p)
			break;
3186
	}
3187

3188 3189
	if (!p)
		return t_hash_start(m, pos);
3190 3191

	return iter;
3192 3193 3194 3195
}

static void t_stop(struct seq_file *m, void *p)
{
3196
	mutex_unlock(&ftrace_lock);
3197 3198
}

3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214
void * __weak
arch_ftrace_trampoline_func(struct ftrace_ops *ops, struct dyn_ftrace *rec)
{
	return NULL;
}

static void add_trampoline_func(struct seq_file *m, struct ftrace_ops *ops,
				struct dyn_ftrace *rec)
{
	void *ptr;

	ptr = arch_ftrace_trampoline_func(ops, rec);
	if (ptr)
		seq_printf(m, " ->%pS", ptr);
}

3215 3216
static int t_show(struct seq_file *m, void *v)
{
3217
	struct ftrace_iterator *iter = m->private;
3218
	struct dyn_ftrace *rec;
3219

3220
	if (iter->flags & FTRACE_ITER_HASH)
3221
		return t_hash_show(m, iter);
3222

3223
	if (iter->flags & FTRACE_ITER_PRINTALL) {
3224
		if (iter->flags & FTRACE_ITER_NOTRACE)
3225
			seq_puts(m, "#### no functions disabled ####\n");
3226
		else
3227
			seq_puts(m, "#### all functions enabled ####\n");
3228 3229 3230
		return 0;
	}

3231 3232
	rec = iter->func;

3233 3234 3235
	if (!rec)
		return 0;

3236
	seq_printf(m, "%ps", (void *)rec->ip);
3237
	if (iter->flags & FTRACE_ITER_ENABLED) {
3238 3239
		struct ftrace_ops *ops = NULL;

3240
		seq_printf(m, " (%ld)%s%s",
3241
			   ftrace_rec_count(rec),
3242 3243
			   rec->flags & FTRACE_FL_REGS ? " R" : "  ",
			   rec->flags & FTRACE_FL_IPMODIFY ? " I" : "  ");
3244
		if (rec->flags & FTRACE_FL_TRAMP_EN) {
3245
			ops = ftrace_find_tramp_ops_any(rec);
3246
			if (ops)
3247 3248 3249
				seq_printf(m, "\ttramp: %pS",
					   (void *)ops->trampoline);
			else
3250
				seq_puts(m, "\ttramp: ERROR!");
3251

3252
		}
3253
		add_trampoline_func(m, ops, rec);
3254 3255
	}	

3256
	seq_putc(m, '\n');
3257 3258 3259 3260

	return 0;
}

J
James Morris 已提交
3261
static const struct seq_operations show_ftrace_seq_ops = {
3262 3263 3264 3265 3266 3267
	.start = t_start,
	.next = t_next,
	.stop = t_stop,
	.show = t_show,
};

I
Ingo Molnar 已提交
3268
static int
3269 3270 3271 3272
ftrace_avail_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3273 3274 3275
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3276 3277 3278 3279
	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
3280
	}
3281

3282
	return iter ? 0 : -ENOMEM;
3283 3284
}

3285 3286 3287 3288 3289
static int
ftrace_enabled_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3290 3291 3292 3293 3294
	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;
3295 3296
	}

3297
	return iter ? 0 : -ENOMEM;
3298 3299
}

3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312
/**
 * 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.
3313
 * tracing_lseek() should be used as the lseek routine, and
3314 3315 3316
 * release must call ftrace_regex_release().
 */
int
3317
ftrace_regex_open(struct ftrace_ops *ops, int flag,
3318
		  struct inode *inode, struct file *file)
3319 3320
{
	struct ftrace_iterator *iter;
3321
	struct ftrace_hash *hash;
3322 3323
	int ret = 0;

3324 3325
	ftrace_ops_init(ops);

3326 3327 3328
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3329 3330 3331 3332
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter)
		return -ENOMEM;

3333 3334 3335 3336 3337
	if (trace_parser_get_init(&iter->parser, FTRACE_BUFF_MAX)) {
		kfree(iter);
		return -ENOMEM;
	}

3338 3339 3340
	iter->ops = ops;
	iter->flags = flag;

3341
	mutex_lock(&ops->func_hash->regex_lock);
3342

3343
	if (flag & FTRACE_ITER_NOTRACE)
3344
		hash = ops->func_hash->notrace_hash;
3345
	else
3346
		hash = ops->func_hash->filter_hash;
3347

3348
	if (file->f_mode & FMODE_WRITE) {
3349 3350 3351 3352 3353 3354 3355
		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);

3356 3357 3358
		if (!iter->hash) {
			trace_parser_put(&iter->parser);
			kfree(iter);
3359 3360
			ret = -ENOMEM;
			goto out_unlock;
3361 3362
		}
	}
3363

3364 3365 3366 3367 3368 3369 3370
	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;
3371
		} else {
3372 3373
			/* Failed */
			free_ftrace_hash(iter->hash);
3374
			trace_parser_put(&iter->parser);
3375
			kfree(iter);
3376
		}
3377 3378
	} else
		file->private_data = iter;
3379 3380

 out_unlock:
3381
	mutex_unlock(&ops->func_hash->regex_lock);
3382 3383 3384 3385

	return ret;
}

3386 3387 3388
static int
ftrace_filter_open(struct inode *inode, struct file *file)
{
3389 3390 3391
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops,
3392 3393
			FTRACE_ITER_FILTER | FTRACE_ITER_DO_HASH,
			inode, file);
3394 3395 3396 3397 3398
}

static int
ftrace_notrace_open(struct inode *inode, struct file *file)
{
3399 3400 3401
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops, FTRACE_ITER_NOTRACE,
3402
				 inode, file);
3403 3404
}

3405
static int ftrace_match(char *str, char *regex, int len, int type)
3406 3407
{
	int matched = 0;
3408
	int slen;
3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423

	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:
3424 3425
		slen = strlen(str);
		if (slen >= len && memcmp(str + slen - len, regex, len) == 0)
3426 3427 3428 3429 3430 3431 3432
			matched = 1;
		break;
	}

	return matched;
}

3433
static int
3434
enter_record(struct ftrace_hash *hash, struct dyn_ftrace *rec, int not)
3435
{
3436 3437 3438
	struct ftrace_func_entry *entry;
	int ret = 0;

3439 3440 3441 3442 3443
	entry = ftrace_lookup_ip(hash, rec->ip);
	if (not) {
		/* Do nothing if it doesn't exist */
		if (!entry)
			return 0;
3444

3445
		free_hash_entry(hash, entry);
3446 3447 3448 3449
	} else {
		/* Do nothing if it exists */
		if (entry)
			return 0;
3450

3451
		ret = add_hash_entry(hash, rec->ip);
3452 3453
	}
	return ret;
3454 3455
}

3456
static int
3457 3458
ftrace_match_record(struct dyn_ftrace *rec, char *mod,
		    char *regex, int len, int type)
3459 3460
{
	char str[KSYM_SYMBOL_LEN];
3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473
	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;
	}
3474 3475 3476 3477

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

3478 3479 3480
static int
match_records(struct ftrace_hash *hash, char *buff,
	      int len, char *mod, int not)
3481
{
3482
	unsigned search_len = 0;
3483 3484
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
3485 3486
	int type = MATCH_FULL;
	char *search = buff;
3487
	int found = 0;
3488
	int ret;
3489

3490 3491 3492 3493
	if (len) {
		type = filter_parse_regex(buff, len, &search, &not);
		search_len = strlen(search);
	}
3494

3495
	mutex_lock(&ftrace_lock);
3496

3497 3498
	if (unlikely(ftrace_disabled))
		goto out_unlock;
3499

3500
	do_for_each_ftrace_rec(pg, rec) {
3501
		if (ftrace_match_record(rec, mod, search, search_len, type)) {
3502
			ret = enter_record(hash, rec, not);
3503 3504 3505 3506
			if (ret < 0) {
				found = ret;
				goto out_unlock;
			}
3507
			found = 1;
3508 3509
		}
	} while_for_each_ftrace_rec();
3510
 out_unlock:
3511
	mutex_unlock(&ftrace_lock);
3512 3513

	return found;
3514 3515
}

3516
static int
3517
ftrace_match_records(struct ftrace_hash *hash, char *buff, int len)
3518
{
3519
	return match_records(hash, buff, len, NULL, 0);
3520 3521
}

3522 3523
static int
ftrace_match_module_records(struct ftrace_hash *hash, char *buff, char *mod)
3524 3525
{
	int not = 0;
S
Steven Rostedt 已提交
3526

3527 3528 3529 3530 3531 3532 3533 3534 3535 3536
	/* 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;
	}

3537
	return match_records(hash, buff, strlen(buff), mod, not);
3538 3539
}

3540 3541 3542 3543 3544 3545
/*
 * We register the module command as a template to show others how
 * to register the a command as well.
 */

static int
3546 3547
ftrace_mod_callback(struct ftrace_hash *hash,
		    char *func, char *cmd, char *param, int enable)
3548 3549
{
	char *mod;
3550
	int ret = -EINVAL;
3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561

	/*
	 * 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)
3562
		return ret;
3563 3564 3565

	mod = strsep(&param, ":");
	if (!strlen(mod))
3566
		return ret;
3567

3568
	ret = ftrace_match_module_records(hash, func, mod);
3569 3570 3571 3572 3573 3574
	if (!ret)
		ret = -EINVAL;
	if (ret < 0)
		return ret;

	return 0;
3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585
}

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);
}
3586
core_initcall(ftrace_mod_cmd_init);
3587

3588
static void function_trace_probe_call(unsigned long ip, unsigned long parent_ip,
3589
				      struct ftrace_ops *op, struct pt_regs *pt_regs)
3590
{
S
Steven Rostedt 已提交
3591
	struct ftrace_func_probe *entry;
3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606
	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.
	 */
3607
	preempt_disable_notrace();
3608
	hlist_for_each_entry_rcu_notrace(entry, hhd, node) {
3609 3610 3611
		if (entry->ip == ip)
			entry->ops->func(ip, parent_ip, &entry->data);
	}
3612
	preempt_enable_notrace();
3613 3614
}

S
Steven Rostedt 已提交
3615
static struct ftrace_ops trace_probe_ops __read_mostly =
3616
{
3617
	.func		= function_trace_probe_call,
3618
	.flags		= FTRACE_OPS_FL_INITIALIZED,
3619
	INIT_OPS_HASH(trace_probe_ops)
3620 3621
};

S
Steven Rostedt 已提交
3622
static int ftrace_probe_registered;
3623

3624
static void __enable_ftrace_function_probe(struct ftrace_ops_hash *old_hash)
3625
{
3626
	int ret;
3627 3628
	int i;

3629 3630 3631
	if (ftrace_probe_registered) {
		/* still need to update the function call sites */
		if (ftrace_enabled)
3632 3633
			ftrace_run_modify_code(&trace_probe_ops, FTRACE_UPDATE_CALLS,
					       old_hash);
3634
		return;
3635
	}
3636 3637 3638 3639 3640 3641 3642 3643 3644 3645

	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;

3646
	ret = ftrace_startup(&trace_probe_ops, 0);
3647

S
Steven Rostedt 已提交
3648
	ftrace_probe_registered = 1;
3649 3650
}

S
Steven Rostedt 已提交
3651
static void __disable_ftrace_function_probe(void)
3652 3653 3654
{
	int i;

S
Steven Rostedt 已提交
3655
	if (!ftrace_probe_registered)
3656 3657 3658 3659 3660 3661 3662 3663 3664
		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 */
3665
	ftrace_shutdown(&trace_probe_ops, 0);
3666

S
Steven Rostedt 已提交
3667
	ftrace_probe_registered = 0;
3668 3669 3670
}


3671
static void ftrace_free_entry(struct ftrace_func_probe *entry)
3672 3673
{
	if (entry->ops->free)
3674
		entry->ops->free(entry->ops, entry->ip, &entry->data);
3675 3676 3677 3678
	kfree(entry);
}

int
S
Steven Rostedt 已提交
3679
register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3680 3681
			      void *data)
{
3682
	struct ftrace_ops_hash old_hash_ops;
S
Steven Rostedt 已提交
3683
	struct ftrace_func_probe *entry;
3684
	struct ftrace_hash **orig_hash = &trace_probe_ops.func_hash->filter_hash;
3685
	struct ftrace_hash *old_hash = *orig_hash;
3686
	struct ftrace_hash *hash;
3687 3688 3689
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
	int type, len, not;
S
Steven Rostedt 已提交
3690
	unsigned long key;
3691 3692
	int count = 0;
	char *search;
3693
	int ret;
3694

3695
	type = filter_parse_regex(glob, strlen(glob), &search, &not);
3696 3697
	len = strlen(search);

S
Steven Rostedt 已提交
3698
	/* we do not support '!' for function probes */
3699 3700 3701
	if (WARN_ON(not))
		return -EINVAL;

3702
	mutex_lock(&trace_probe_ops.func_hash->regex_lock);
3703

3704 3705 3706 3707
	old_hash_ops.filter_hash = old_hash;
	/* Probes only have filters */
	old_hash_ops.notrace_hash = NULL;

3708
	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, old_hash);
3709 3710
	if (!hash) {
		count = -ENOMEM;
3711
		goto out;
3712 3713 3714 3715
	}

	if (unlikely(ftrace_disabled)) {
		count = -ENODEV;
3716
		goto out;
3717
	}
3718

3719 3720
	mutex_lock(&ftrace_lock);

3721
	do_for_each_ftrace_rec(pg, rec) {
3722

3723
		if (!ftrace_match_record(rec, NULL, search, len, type))
3724 3725 3726 3727
			continue;

		entry = kmalloc(sizeof(*entry), GFP_KERNEL);
		if (!entry) {
S
Steven Rostedt 已提交
3728
			/* If we did not process any, then return error */
3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742
			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.
		 */
3743 3744
		if (ops->init) {
			if (ops->init(ops, rec->ip, &entry->data) < 0) {
3745 3746 3747 3748 3749 3750
				/* caller does not like this func */
				kfree(entry);
				continue;
			}
		}

3751 3752 3753 3754 3755 3756 3757
		ret = enter_record(hash, rec, 0);
		if (ret < 0) {
			kfree(entry);
			count = ret;
			goto out_unlock;
		}

3758 3759 3760 3761 3762 3763 3764
		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();
3765 3766

	ret = ftrace_hash_move(&trace_probe_ops, 1, orig_hash, hash);
3767

3768
	__enable_ftrace_function_probe(&old_hash_ops);
3769

3770 3771 3772
	if (!ret)
		free_ftrace_hash_rcu(old_hash);
	else
3773 3774
		count = ret;

3775
 out_unlock:
3776 3777
	mutex_unlock(&ftrace_lock);
 out:
3778
	mutex_unlock(&trace_probe_ops.func_hash->regex_lock);
3779
	free_ftrace_hash(hash);
3780 3781 3782 3783 3784

	return count;
}

enum {
S
Steven Rostedt 已提交
3785 3786
	PROBE_TEST_FUNC		= 1,
	PROBE_TEST_DATA		= 2
3787 3788 3789
};

static void
S
Steven Rostedt 已提交
3790
__unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3791 3792
				  void *data, int flags)
{
3793
	struct ftrace_func_entry *rec_entry;
S
Steven Rostedt 已提交
3794
	struct ftrace_func_probe *entry;
3795
	struct ftrace_func_probe *p;
3796
	struct ftrace_hash **orig_hash = &trace_probe_ops.func_hash->filter_hash;
3797
	struct ftrace_hash *old_hash = *orig_hash;
3798
	struct list_head free_list;
3799
	struct ftrace_hash *hash;
3800
	struct hlist_node *tmp;
3801 3802 3803 3804
	char str[KSYM_SYMBOL_LEN];
	int type = MATCH_FULL;
	int i, len = 0;
	char *search;
3805
	int ret;
3806

3807
	if (glob && (strcmp(glob, "*") == 0 || !strlen(glob)))
3808
		glob = NULL;
3809
	else if (glob) {
3810 3811
		int not;

3812
		type = filter_parse_regex(glob, strlen(glob), &search, &not);
3813 3814
		len = strlen(search);

S
Steven Rostedt 已提交
3815
		/* we do not support '!' for function probes */
3816 3817 3818 3819
		if (WARN_ON(not))
			return;
	}

3820
	mutex_lock(&trace_probe_ops.func_hash->regex_lock);
3821 3822 3823 3824 3825 3826

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

3827 3828
	INIT_LIST_HEAD(&free_list);

3829 3830 3831
	for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) {
		struct hlist_head *hhd = &ftrace_func_hash[i];

3832
		hlist_for_each_entry_safe(entry, tmp, hhd, node) {
3833 3834

			/* break up if statements for readability */
S
Steven Rostedt 已提交
3835
			if ((flags & PROBE_TEST_FUNC) && entry->ops != ops)
3836 3837
				continue;

S
Steven Rostedt 已提交
3838
			if ((flags & PROBE_TEST_DATA) && entry->data != data)
3839 3840 3841 3842 3843 3844 3845 3846 3847 3848
				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;
			}

3849 3850 3851 3852 3853
			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);

3854
			hlist_del_rcu(&entry->node);
3855
			list_add(&entry->free_list, &free_list);
3856 3857
		}
	}
3858
	mutex_lock(&ftrace_lock);
S
Steven Rostedt 已提交
3859
	__disable_ftrace_function_probe();
3860 3861 3862 3863
	/*
	 * Remove after the disable is called. Otherwise, if the last
	 * probe is removed, a null hash means *all enabled*.
	 */
3864
	ret = ftrace_hash_move(&trace_probe_ops, 1, orig_hash, hash);
3865
	synchronize_sched();
3866 3867 3868
	if (!ret)
		free_ftrace_hash_rcu(old_hash);

3869 3870 3871 3872
	list_for_each_entry_safe(entry, p, &free_list, free_list) {
		list_del(&entry->free_list);
		ftrace_free_entry(entry);
	}
3873
	mutex_unlock(&ftrace_lock);
3874
		
3875
 out_unlock:
3876
	mutex_unlock(&trace_probe_ops.func_hash->regex_lock);
3877
	free_ftrace_hash(hash);
3878 3879 3880
}

void
S
Steven Rostedt 已提交
3881
unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3882 3883
				void *data)
{
S
Steven Rostedt 已提交
3884 3885
	__unregister_ftrace_function_probe(glob, ops, data,
					  PROBE_TEST_FUNC | PROBE_TEST_DATA);
3886 3887 3888
}

void
S
Steven Rostedt 已提交
3889
unregister_ftrace_function_probe_func(char *glob, struct ftrace_probe_ops *ops)
3890
{
S
Steven Rostedt 已提交
3891
	__unregister_ftrace_function_probe(glob, ops, NULL, PROBE_TEST_FUNC);
3892 3893
}

S
Steven Rostedt 已提交
3894
void unregister_ftrace_function_probe_all(char *glob)
3895
{
S
Steven Rostedt 已提交
3896
	__unregister_ftrace_function_probe(glob, NULL, NULL, 0);
3897 3898
}

3899 3900 3901
static LIST_HEAD(ftrace_commands);
static DEFINE_MUTEX(ftrace_cmd_mutex);

3902 3903 3904 3905 3906
/*
 * Currently we only register ftrace commands from __init, so mark this
 * __init too.
 */
__init int register_ftrace_command(struct ftrace_func_command *cmd)
3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924
{
	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;
}

3925 3926 3927 3928 3929
/*
 * Currently we only unregister ftrace commands from __init, so mark
 * this __init too.
 */
__init int unregister_ftrace_command(struct ftrace_func_command *cmd)
3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947
{
	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;
}

3948 3949
static int ftrace_process_regex(struct ftrace_hash *hash,
				char *buff, int len, int enable)
3950
{
3951
	char *func, *command, *next = buff;
S
Steven Rostedt 已提交
3952
	struct ftrace_func_command *p;
3953
	int ret = -EINVAL;
3954 3955 3956 3957

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

	if (!next) {
3958
		ret = ftrace_match_records(hash, func, len);
3959 3960 3961 3962 3963
		if (!ret)
			ret = -EINVAL;
		if (ret < 0)
			return ret;
		return 0;
3964 3965
	}

3966
	/* command found */
3967 3968 3969

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

3970 3971 3972
	mutex_lock(&ftrace_cmd_mutex);
	list_for_each_entry(p, &ftrace_commands, list) {
		if (strcmp(p->name, command) == 0) {
3973
			ret = p->func(hash, func, command, next, enable);
3974 3975
			goto out_unlock;
		}
3976
	}
3977 3978
 out_unlock:
	mutex_unlock(&ftrace_cmd_mutex);
3979

3980
	return ret;
3981 3982
}

I
Ingo Molnar 已提交
3983
static ssize_t
3984 3985
ftrace_regex_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos, int enable)
3986 3987
{
	struct ftrace_iterator *iter;
3988 3989
	struct trace_parser *parser;
	ssize_t ret, read;
3990

3991
	if (!cnt)
3992 3993 3994 3995 3996 3997 3998 3999
		return 0;

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

4000
	if (unlikely(ftrace_disabled))
4001 4002 4003
		return -ENODEV;

	/* iter->hash is a local copy, so we don't need regex_lock */
4004

4005 4006
	parser = &iter->parser;
	read = trace_get_user(parser, ubuf, cnt, ppos);
4007

4008
	if (read >= 0 && trace_parser_loaded(parser) &&
4009
	    !trace_parser_cont(parser)) {
4010
		ret = ftrace_process_regex(iter->hash, parser->buffer,
4011
					   parser->idx, enable);
4012
		trace_parser_clear(parser);
4013
		if (ret < 0)
4014
			goto out;
4015
	}
4016 4017

	ret = read;
4018
 out:
4019 4020 4021
	return ret;
}

4022
ssize_t
4023 4024 4025 4026 4027 4028
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);
}

4029
ssize_t
4030 4031 4032 4033 4034 4035
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);
}

4036
static int
4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054
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);
}

4055
static void ftrace_ops_update_code(struct ftrace_ops *ops,
4056
				   struct ftrace_ops_hash *old_hash)
4057
{
4058 4059 4060 4061 4062 4063
	struct ftrace_ops *op;

	if (!ftrace_enabled)
		return;

	if (ops->flags & FTRACE_OPS_FL_ENABLED) {
4064
		ftrace_run_modify_code(ops, FTRACE_UPDATE_CALLS, old_hash);
4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083
		return;
	}

	/*
	 * If this is the shared global_ops filter, then we need to
	 * check if there is another ops that shares it, is enabled.
	 * If so, we still need to run the modify code.
	 */
	if (ops->func_hash != &global_ops.local_hash)
		return;

	do_for_each_ftrace_op(op, ftrace_ops_list) {
		if (op->func_hash == &global_ops.local_hash &&
		    op->flags & FTRACE_OPS_FL_ENABLED) {
			ftrace_run_modify_code(op, FTRACE_UPDATE_CALLS, old_hash);
			/* Only need to do this once */
			return;
		}
	} while_for_each_ftrace_op(op);
4084 4085
}

4086 4087 4088
static int
ftrace_set_hash(struct ftrace_ops *ops, unsigned char *buf, int len,
		unsigned long ip, int remove, int reset, int enable)
4089
{
4090
	struct ftrace_hash **orig_hash;
4091
	struct ftrace_ops_hash old_hash_ops;
4092
	struct ftrace_hash *old_hash;
4093
	struct ftrace_hash *hash;
4094
	int ret;
4095

4096
	if (unlikely(ftrace_disabled))
4097
		return -ENODEV;
4098

4099
	mutex_lock(&ops->func_hash->regex_lock);
4100

4101
	if (enable)
4102
		orig_hash = &ops->func_hash->filter_hash;
4103
	else
4104
		orig_hash = &ops->func_hash->notrace_hash;
4105

4106 4107 4108 4109 4110
	if (reset)
		hash = alloc_ftrace_hash(FTRACE_HASH_DEFAULT_BITS);
	else
		hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);

4111 4112 4113 4114
	if (!hash) {
		ret = -ENOMEM;
		goto out_regex_unlock;
	}
4115

4116 4117 4118 4119
	if (buf && !ftrace_match_records(hash, buf, len)) {
		ret = -EINVAL;
		goto out_regex_unlock;
	}
4120 4121 4122 4123 4124
	if (ip) {
		ret = ftrace_match_addr(hash, ip, remove);
		if (ret < 0)
			goto out_regex_unlock;
	}
4125 4126

	mutex_lock(&ftrace_lock);
4127
	old_hash = *orig_hash;
4128 4129
	old_hash_ops.filter_hash = ops->func_hash->filter_hash;
	old_hash_ops.notrace_hash = ops->func_hash->notrace_hash;
4130
	ret = ftrace_hash_move(ops, enable, orig_hash, hash);
4131
	if (!ret) {
4132
		ftrace_ops_update_code(ops, &old_hash_ops);
4133 4134
		free_ftrace_hash_rcu(old_hash);
	}
4135 4136
	mutex_unlock(&ftrace_lock);

4137
 out_regex_unlock:
4138
	mutex_unlock(&ops->func_hash->regex_lock);
4139 4140 4141

	free_ftrace_hash(hash);
	return ret;
4142 4143
}

4144 4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163
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)
{
4164
	ftrace_ops_init(ops);
4165 4166 4167 4168 4169 4170 4171 4172 4173 4174 4175
	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);
}

4176 4177
/**
 * ftrace_set_filter - set a function to filter on in ftrace
4178 4179 4180 4181 4182 4183 4184 4185
 * @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.
 */
4186
int ftrace_set_filter(struct ftrace_ops *ops, unsigned char *buf,
4187 4188
		       int len, int reset)
{
4189
	ftrace_ops_init(ops);
4190
	return ftrace_set_regex(ops, buf, len, reset, 1);
4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204
}
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.
 */
4205
int ftrace_set_notrace(struct ftrace_ops *ops, unsigned char *buf,
4206 4207
			int len, int reset)
{
4208
	ftrace_ops_init(ops);
4209
	return ftrace_set_regex(ops, buf, len, reset, 0);
4210 4211 4212
}
EXPORT_SYMBOL_GPL(ftrace_set_notrace);
/**
4213
 * ftrace_set_global_filter - set a function to filter on with global tracers
4214 4215 4216 4217 4218 4219 4220
 * @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.
 */
4221
void ftrace_set_global_filter(unsigned char *buf, int len, int reset)
4222
{
4223
	ftrace_set_regex(&global_ops, buf, len, reset, 1);
4224
}
4225
EXPORT_SYMBOL_GPL(ftrace_set_global_filter);
4226

4227
/**
4228
 * ftrace_set_global_notrace - set a function to not trace with global tracers
4229 4230 4231 4232 4233 4234 4235 4236
 * @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.
 */
4237
void ftrace_set_global_notrace(unsigned char *buf, int len, int reset)
4238
{
4239
	ftrace_set_regex(&global_ops, buf, len, reset, 0);
4240
}
4241
EXPORT_SYMBOL_GPL(ftrace_set_global_notrace);
4242

4243 4244 4245 4246 4247 4248 4249
/*
 * 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;

4250 4251 4252
/* Used by function selftest to not test if filter is set */
bool ftrace_filter_param __initdata;

4253 4254
static int __init set_ftrace_notrace(char *str)
{
4255
	ftrace_filter_param = true;
4256
	strlcpy(ftrace_notrace_buf, str, FTRACE_FILTER_SIZE);
4257 4258 4259 4260 4261 4262
	return 1;
}
__setup("ftrace_notrace=", set_ftrace_notrace);

static int __init set_ftrace_filter(char *str)
{
4263
	ftrace_filter_param = true;
4264
	strlcpy(ftrace_filter_buf, str, FTRACE_FILTER_SIZE);
4265 4266 4267 4268
	return 1;
}
__setup("ftrace_filter=", set_ftrace_filter);

4269
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
4270
static char ftrace_graph_buf[FTRACE_FILTER_SIZE] __initdata;
4271
static char ftrace_graph_notrace_buf[FTRACE_FILTER_SIZE] __initdata;
4272
static int ftrace_set_func(unsigned long *array, int *idx, int size, char *buffer);
4273

4274 4275 4276
static unsigned long save_global_trampoline;
static unsigned long save_global_flags;

4277 4278
static int __init set_graph_function(char *str)
{
4279
	strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE);
4280 4281 4282 4283
	return 1;
}
__setup("ftrace_graph_filter=", set_graph_function);

4284 4285 4286 4287 4288 4289 4290 4291
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)
4292 4293 4294
{
	int ret;
	char *func;
4295 4296 4297 4298 4299 4300 4301
	unsigned long *table = ftrace_graph_funcs;
	int *count = &ftrace_graph_count;

	if (!enable) {
		table = ftrace_graph_notrace_funcs;
		count = &ftrace_graph_notrace_count;
	}
4302 4303 4304 4305

	while (buf) {
		func = strsep(&buf, ",");
		/* we allow only one expression at a time */
4306
		ret = ftrace_set_func(table, count, FTRACE_GRAPH_MAX_FUNCS, func);
4307 4308 4309 4310 4311 4312 4313
		if (ret)
			printk(KERN_DEBUG "ftrace: function %s not "
					  "traceable\n", func);
	}
}
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4314 4315
void __init
ftrace_set_early_filter(struct ftrace_ops *ops, char *buf, int enable)
4316 4317 4318
{
	char *func;

4319 4320
	ftrace_ops_init(ops);

4321 4322
	while (buf) {
		func = strsep(&buf, ",");
4323
		ftrace_set_regex(ops, func, strlen(func), 0, enable);
4324 4325 4326 4327 4328 4329
	}
}

static void __init set_ftrace_early_filters(void)
{
	if (ftrace_filter_buf[0])
4330
		ftrace_set_early_filter(&global_ops, ftrace_filter_buf, 1);
4331
	if (ftrace_notrace_buf[0])
4332
		ftrace_set_early_filter(&global_ops, ftrace_notrace_buf, 0);
4333 4334
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	if (ftrace_graph_buf[0])
4335 4336 4337
		set_ftrace_early_graph(ftrace_graph_buf, 1);
	if (ftrace_graph_notrace_buf[0])
		set_ftrace_early_graph(ftrace_graph_notrace_buf, 0);
4338
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
4339 4340
}

4341
int ftrace_regex_release(struct inode *inode, struct file *file)
4342 4343
{
	struct seq_file *m = (struct seq_file *)file->private_data;
4344
	struct ftrace_ops_hash old_hash_ops;
4345
	struct ftrace_iterator *iter;
4346
	struct ftrace_hash **orig_hash;
4347
	struct ftrace_hash *old_hash;
4348
	struct trace_parser *parser;
4349
	int filter_hash;
4350
	int ret;
4351 4352 4353 4354 4355 4356 4357

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

4358 4359 4360
	parser = &iter->parser;
	if (trace_parser_loaded(parser)) {
		parser->buffer[parser->idx] = 0;
4361
		ftrace_match_records(iter->hash, parser->buffer, parser->idx);
4362 4363
	}

4364 4365
	trace_parser_put(parser);

4366
	mutex_lock(&iter->ops->func_hash->regex_lock);
4367

4368
	if (file->f_mode & FMODE_WRITE) {
4369 4370 4371
		filter_hash = !!(iter->flags & FTRACE_ITER_FILTER);

		if (filter_hash)
4372
			orig_hash = &iter->ops->func_hash->filter_hash;
4373
		else
4374
			orig_hash = &iter->ops->func_hash->notrace_hash;
4375

4376
		mutex_lock(&ftrace_lock);
4377
		old_hash = *orig_hash;
4378 4379
		old_hash_ops.filter_hash = iter->ops->func_hash->filter_hash;
		old_hash_ops.notrace_hash = iter->ops->func_hash->notrace_hash;
4380 4381
		ret = ftrace_hash_move(iter->ops, filter_hash,
				       orig_hash, iter->hash);
4382
		if (!ret) {
4383
			ftrace_ops_update_code(iter->ops, &old_hash_ops);
4384 4385
			free_ftrace_hash_rcu(old_hash);
		}
4386 4387
		mutex_unlock(&ftrace_lock);
	}
4388

4389
	mutex_unlock(&iter->ops->func_hash->regex_lock);
4390 4391
	free_ftrace_hash(iter->hash);
	kfree(iter);
4392

4393 4394 4395
	return 0;
}

4396
static const struct file_operations ftrace_avail_fops = {
4397 4398 4399
	.open = ftrace_avail_open,
	.read = seq_read,
	.llseek = seq_lseek,
L
Li Zefan 已提交
4400
	.release = seq_release_private,
4401 4402
};

4403 4404 4405 4406 4407 4408 4409
static const struct file_operations ftrace_enabled_fops = {
	.open = ftrace_enabled_open,
	.read = seq_read,
	.llseek = seq_lseek,
	.release = seq_release_private,
};

4410
static const struct file_operations ftrace_filter_fops = {
4411
	.open = ftrace_filter_open,
L
Lai Jiangshan 已提交
4412
	.read = seq_read,
4413
	.write = ftrace_filter_write,
4414
	.llseek = tracing_lseek,
4415
	.release = ftrace_regex_release,
4416 4417
};

4418
static const struct file_operations ftrace_notrace_fops = {
4419
	.open = ftrace_notrace_open,
L
Lai Jiangshan 已提交
4420
	.read = seq_read,
4421
	.write = ftrace_notrace_write,
4422
	.llseek = tracing_lseek,
4423
	.release = ftrace_regex_release,
4424 4425
};

4426 4427 4428 4429 4430
#ifdef CONFIG_FUNCTION_GRAPH_TRACER

static DEFINE_MUTEX(graph_lock);

int ftrace_graph_count;
4431
int ftrace_graph_notrace_count;
4432
unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly;
4433
unsigned long ftrace_graph_notrace_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly;
4434

4435 4436 4437 4438 4439 4440 4441
struct ftrace_graph_data {
	unsigned long *table;
	size_t size;
	int *count;
	const struct seq_operations *seq_ops;
};

4442
static void *
4443
__g_next(struct seq_file *m, loff_t *pos)
4444
{
4445 4446 4447
	struct ftrace_graph_data *fgd = m->private;

	if (*pos >= *fgd->count)
4448
		return NULL;
4449
	return &fgd->table[*pos];
4450
}
4451

4452 4453 4454 4455 4456
static void *
g_next(struct seq_file *m, void *v, loff_t *pos)
{
	(*pos)++;
	return __g_next(m, pos);
4457 4458 4459 4460
}

static void *g_start(struct seq_file *m, loff_t *pos)
{
4461 4462
	struct ftrace_graph_data *fgd = m->private;

4463 4464
	mutex_lock(&graph_lock);

4465
	/* Nothing, tell g_show to print all functions are enabled */
4466
	if (!*fgd->count && !*pos)
4467 4468
		return (void *)1;

4469
	return __g_next(m, pos);
4470 4471 4472 4473 4474 4475 4476 4477 4478 4479 4480 4481 4482 4483
}

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;

4484
	if (ptr == (unsigned long *)1) {
4485 4486 4487
		struct ftrace_graph_data *fgd = m->private;

		if (fgd->table == ftrace_graph_funcs)
4488
			seq_puts(m, "#### all functions enabled ####\n");
4489
		else
4490
			seq_puts(m, "#### no functions disabled ####\n");
4491 4492 4493
		return 0;
	}

4494
	seq_printf(m, "%ps\n", (void *)*ptr);
4495 4496 4497 4498

	return 0;
}

J
James Morris 已提交
4499
static const struct seq_operations ftrace_graph_seq_ops = {
4500 4501 4502 4503 4504 4505 4506
	.start = g_start,
	.next = g_next,
	.stop = g_stop,
	.show = g_show,
};

static int
4507 4508
__ftrace_graph_open(struct inode *inode, struct file *file,
		    struct ftrace_graph_data *fgd)
4509 4510 4511 4512 4513
{
	int ret = 0;

	mutex_lock(&graph_lock);
	if ((file->f_mode & FMODE_WRITE) &&
4514
	    (file->f_flags & O_TRUNC)) {
4515 4516
		*fgd->count = 0;
		memset(fgd->table, 0, fgd->size * sizeof(*fgd->table));
4517
	}
4518
	mutex_unlock(&graph_lock);
4519

4520 4521 4522 4523 4524 4525 4526 4527
	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;
4528 4529 4530 4531

	return ret;
}

4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551
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);
}

4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571
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);
}

4572 4573 4574
static int
ftrace_graph_release(struct inode *inode, struct file *file)
{
4575 4576 4577 4578
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;

		kfree(m->private);
4579
		seq_release(inode, file);
4580 4581 4582 4583
	} else {
		kfree(file->private_data);
	}

4584 4585 4586
	return 0;
}

4587
static int
4588
ftrace_set_func(unsigned long *array, int *idx, int size, char *buffer)
4589 4590 4591
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
4592
	int search_len;
4593
	int fail = 1;
4594 4595 4596 4597
	int type, not;
	char *search;
	bool exists;
	int i;
4598

4599
	/* decode regex */
4600
	type = filter_parse_regex(buffer, strlen(buffer), &search, &not);
4601
	if (!not && *idx >= size)
4602
		return -EBUSY;
4603 4604 4605

	search_len = strlen(search);

4606
	mutex_lock(&ftrace_lock);
4607 4608 4609 4610 4611 4612

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

4613 4614
	do_for_each_ftrace_rec(pg, rec) {

4615
		if (ftrace_match_record(rec, NULL, search, search_len, type)) {
4616
			/* if it is in the array */
4617
			exists = false;
4618
			for (i = 0; i < *idx; i++) {
4619 4620
				if (array[i] == rec->ip) {
					exists = true;
4621 4622
					break;
				}
4623 4624 4625 4626 4627 4628
			}

			if (!not) {
				fail = 0;
				if (!exists) {
					array[(*idx)++] = rec->ip;
4629
					if (*idx >= size)
4630 4631 4632 4633 4634 4635 4636 4637 4638
						goto out;
				}
			} else {
				if (exists) {
					array[i] = array[--(*idx)];
					array[*idx] = 0;
					fail = 0;
				}
			}
4639
		}
4640
	} while_for_each_ftrace_rec();
4641
out:
4642
	mutex_unlock(&ftrace_lock);
4643

4644 4645 4646 4647
	if (fail)
		return -EINVAL;

	return 0;
4648 4649 4650 4651 4652 4653
}

static ssize_t
ftrace_graph_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
4654
	struct trace_parser parser;
4655
	ssize_t read, ret = 0;
4656
	struct ftrace_graph_data *fgd = file->private_data;
4657

4658
	if (!cnt)
4659 4660
		return 0;

4661 4662
	if (trace_parser_get_init(&parser, FTRACE_BUFF_MAX))
		return -ENOMEM;
4663

4664
	read = trace_get_user(&parser, ubuf, cnt, ppos);
4665

4666
	if (read >= 0 && trace_parser_loaded((&parser))) {
4667 4668
		parser.buffer[parser.idx] = 0;

4669 4670
		mutex_lock(&graph_lock);

4671
		/* we allow only one expression at a time */
4672 4673
		ret = ftrace_set_func(fgd->table, fgd->count, fgd->size,
				      parser.buffer);
4674 4675

		mutex_unlock(&graph_lock);
4676 4677
	}

4678 4679
	if (!ret)
		ret = read;
4680

4681
	trace_parser_put(&parser);
4682 4683 4684 4685 4686

	return ret;
}

static const struct file_operations ftrace_graph_fops = {
4687 4688 4689
	.open		= ftrace_graph_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
4690
	.llseek		= tracing_lseek,
4691
	.release	= ftrace_graph_release,
4692
};
4693 4694 4695 4696 4697

static const struct file_operations ftrace_graph_notrace_fops = {
	.open		= ftrace_graph_notrace_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
4698
	.llseek		= tracing_lseek,
4699 4700
	.release	= ftrace_graph_release,
};
4701 4702
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4703 4704 4705 4706 4707 4708 4709 4710 4711 4712 4713 4714 4715 4716 4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732
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);
}

4733
static __init int ftrace_init_dyn_tracefs(struct dentry *d_tracer)
4734 4735
{

4736 4737
	trace_create_file("available_filter_functions", 0444,
			d_tracer, NULL, &ftrace_avail_fops);
4738

4739 4740 4741
	trace_create_file("enabled_functions", 0444,
			d_tracer, NULL, &ftrace_enabled_fops);

4742
	ftrace_create_filter_files(&global_ops, d_tracer);
4743

4744
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
4745
	trace_create_file("set_graph_function", 0444, d_tracer,
4746 4747
				    NULL,
				    &ftrace_graph_fops);
4748 4749 4750
	trace_create_file("set_graph_notrace", 0444, d_tracer,
				    NULL,
				    &ftrace_graph_notrace_fops);
4751 4752
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4753 4754 4755
	return 0;
}

4756
static int ftrace_cmp_ips(const void *a, const void *b)
4757
{
4758 4759
	const unsigned long *ipa = a;
	const unsigned long *ipb = b;
4760

4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771 4772 4773 4774 4775 4776
	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;
4777 4778
}

4779
static int ftrace_process_locs(struct module *mod,
4780
			       unsigned long *start,
4781 4782
			       unsigned long *end)
{
4783
	struct ftrace_page *start_pg;
4784
	struct ftrace_page *pg;
4785
	struct dyn_ftrace *rec;
4786
	unsigned long count;
4787 4788
	unsigned long *p;
	unsigned long addr;
4789
	unsigned long flags = 0; /* Shut up gcc */
4790 4791 4792 4793 4794 4795 4796
	int ret = -ENOMEM;

	count = end - start;

	if (!count)
		return 0;

4797 4798 4799
	sort(start, count, sizeof(*start),
	     ftrace_cmp_ips, ftrace_swap_ips);

4800 4801
	start_pg = ftrace_allocate_pages(count);
	if (!start_pg)
4802
		return -ENOMEM;
4803

S
Steven Rostedt 已提交
4804
	mutex_lock(&ftrace_lock);
4805

4806 4807 4808 4809 4810
	/*
	 * 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.
	 */
4811 4812 4813
	if (!mod) {
		WARN_ON(ftrace_pages || ftrace_pages_start);
		/* First initialization */
4814
		ftrace_pages = ftrace_pages_start = start_pg;
4815
	} else {
4816
		if (!ftrace_pages)
4817
			goto out;
4818

4819 4820 4821 4822
		if (WARN_ON(ftrace_pages->next)) {
			/* Hmm, we have free pages? */
			while (ftrace_pages->next)
				ftrace_pages = ftrace_pages->next;
4823
		}
4824

4825
		ftrace_pages->next = start_pg;
4826 4827
	}

4828
	p = start;
4829
	pg = start_pg;
4830 4831
	while (p < end) {
		addr = ftrace_call_adjust(*p++);
4832 4833 4834 4835 4836 4837 4838 4839
		/*
		 * 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;
4840 4841 4842 4843 4844 4845 4846 4847 4848 4849

		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;
4850 4851
	}

4852 4853 4854 4855 4856 4857
	/* We should have used all pages */
	WARN_ON(pg->next);

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

4858
	/*
4859 4860 4861 4862 4863 4864
	 * 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.
4865
	 */
4866 4867
	if (!mod)
		local_irq_save(flags);
4868
	ftrace_update_code(mod, start_pg);
4869 4870
	if (!mod)
		local_irq_restore(flags);
4871 4872
	ret = 0;
 out:
S
Steven Rostedt 已提交
4873
	mutex_unlock(&ftrace_lock);
4874

4875
	return ret;
4876 4877
}

4878
#ifdef CONFIG_MODULES
4879 4880 4881

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

4882
void ftrace_release_mod(struct module *mod)
4883 4884
{
	struct dyn_ftrace *rec;
4885
	struct ftrace_page **last_pg;
4886
	struct ftrace_page *pg;
4887
	int order;
4888

4889 4890
	mutex_lock(&ftrace_lock);

4891
	if (ftrace_disabled)
4892
		goto out_unlock;
4893

4894 4895 4896 4897 4898 4899 4900
	/*
	 * 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];
4901
		if (within_module_core(rec->ip, mod)) {
4902
			/*
4903 4904
			 * As core pages are first, the first
			 * page should never be a module page.
4905
			 */
4906 4907 4908 4909 4910 4911 4912 4913
			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;
4914 4915 4916
			order = get_count_order(pg->size / ENTRIES_PER_PAGE);
			free_pages((unsigned long)pg->records, order);
			kfree(pg);
4917 4918 4919
		} else
			last_pg = &pg->next;
	}
4920
 out_unlock:
4921 4922 4923 4924 4925
	mutex_unlock(&ftrace_lock);
}

static void ftrace_init_module(struct module *mod,
			       unsigned long *start, unsigned long *end)
4926
{
4927
	if (ftrace_disabled || start == end)
4928
		return;
4929
	ftrace_process_locs(mod, start, end);
4930 4931
}

4932
void ftrace_module_init(struct module *mod)
4933
{
4934 4935 4936
	ftrace_init_module(mod, mod->ftrace_callsites,
			   mod->ftrace_callsites +
			   mod->num_ftrace_callsites);
4937 4938 4939 4940 4941 4942 4943 4944
}

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

	if (val == MODULE_STATE_GOING)
4945
		ftrace_release_mod(mod);
4946 4947 4948 4949

	return 0;
}
#else
4950 4951
static int ftrace_module_notify_exit(struct notifier_block *self,
				     unsigned long val, void *data)
4952 4953 4954 4955 4956
{
	return 0;
}
#endif /* CONFIG_MODULES */

4957 4958 4959 4960 4961
struct notifier_block ftrace_module_exit_nb = {
	.notifier_call = ftrace_module_notify_exit,
	.priority = INT_MIN,	/* Run after anything that can remove kprobes */
};

4962 4963
void __init ftrace_init(void)
{
4964 4965
	extern unsigned long __start_mcount_loc[];
	extern unsigned long __stop_mcount_loc[];
4966
	unsigned long count, flags;
4967 4968 4969
	int ret;

	local_irq_save(flags);
4970
	ret = ftrace_dyn_arch_init();
4971
	local_irq_restore(flags);
4972
	if (ret)
4973 4974 4975
		goto failed;

	count = __stop_mcount_loc - __start_mcount_loc;
4976 4977
	if (!count) {
		pr_info("ftrace: No functions to be traced?\n");
4978
		goto failed;
4979 4980 4981 4982
	}

	pr_info("ftrace: allocating %ld entries in %ld pages\n",
		count, count / ENTRIES_PER_PAGE + 1);
4983 4984 4985

	last_ftrace_enabled = ftrace_enabled = 1;

4986
	ret = ftrace_process_locs(NULL,
4987
				  __start_mcount_loc,
4988 4989
				  __stop_mcount_loc);

4990
	ret = register_module_notifier(&ftrace_module_exit_nb);
4991
	if (ret)
4992
		pr_warning("Failed to register trace ftrace module exit notifier\n");
4993

4994 4995
	set_ftrace_early_filters();

4996 4997 4998 4999 5000
	return;
 failed:
	ftrace_disabled = 1;
}

5001 5002 5003 5004 5005 5006 5007
/* Do nothing if arch does not support this */
void __weak arch_ftrace_update_trampoline(struct ftrace_ops *ops)
{
}

static void ftrace_update_trampoline(struct ftrace_ops *ops)
{
5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018

/*
 * Currently there's no safe way to free a trampoline when the kernel
 * is configured with PREEMPT. That is because a task could be preempted
 * when it jumped to the trampoline, it may be preempted for a long time
 * depending on the system load, and currently there's no way to know
 * when it will be off the trampoline. If the trampoline is freed
 * too early, when the task runs again, it will be executing on freed
 * memory and crash.
 */
#ifdef CONFIG_PREEMPT
5019 5020 5021
	/* Currently, only non dynamic ops can have a trampoline */
	if (ops->flags & FTRACE_OPS_FL_DYNAMIC)
		return;
5022
#endif
5023 5024 5025 5026

	arch_ftrace_update_trampoline(ops);
}

5027
#else
5028

5029
static struct ftrace_ops global_ops = {
5030
	.func			= ftrace_stub,
5031
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_INITIALIZED,
5032 5033
};

5034 5035 5036 5037 5038
static int __init ftrace_nodyn_init(void)
{
	ftrace_enabled = 1;
	return 0;
}
5039
core_initcall(ftrace_nodyn_init);
5040

5041
static inline int ftrace_init_dyn_tracefs(struct dentry *d_tracer) { return 0; }
5042
static inline void ftrace_startup_enable(int command) { }
5043
static inline void ftrace_startup_all(int command) { }
5044
/* Keep as macros so we do not need to define the commands */
5045 5046 5047 5048 5049 5050
# define ftrace_startup(ops, command)					\
	({								\
		int ___ret = __register_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags |= FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
5051
	})
5052 5053 5054 5055 5056 5057 5058
# define ftrace_shutdown(ops, command)					\
	({								\
		int ___ret = __unregister_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags &= ~FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
	})
5059

I
Ingo Molnar 已提交
5060 5061
# define ftrace_startup_sysctl()	do { } while (0)
# define ftrace_shutdown_sysctl()	do { } while (0)
5062 5063

static inline int
5064
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
5065 5066 5067 5068
{
	return 1;
}

5069 5070 5071 5072
static void ftrace_update_trampoline(struct ftrace_ops *ops)
{
}

5073 5074
#endif /* CONFIG_DYNAMIC_FTRACE */

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
__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;
}

5103
static void
5104
ftrace_ops_control_func(unsigned long ip, unsigned long parent_ip,
5105
			struct ftrace_ops *op, struct pt_regs *regs)
5106 5107 5108 5109 5110 5111 5112 5113 5114 5115
{
	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);
5116 5117 5118 5119 5120 5121 5122 5123

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

5124
	do_for_each_ftrace_op(op, ftrace_control_list) {
5125 5126
		if (!(op->flags & FTRACE_OPS_FL_STUB) &&
		    !ftrace_function_local_disabled(op) &&
5127
		    ftrace_ops_test(op, ip, regs))
5128
			op->func(ip, parent_ip, op, regs);
5129
	} while_for_each_ftrace_op(op);
5130
 out:
5131 5132 5133 5134 5135
	trace_recursion_clear(TRACE_CONTROL_BIT);
	preempt_enable_notrace();
}

static struct ftrace_ops control_ops = {
5136 5137
	.func	= ftrace_ops_control_func,
	.flags	= FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_INITIALIZED,
5138
	INIT_OPS_HASH(control_ops)
5139 5140
};

5141 5142
static inline void
__ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
5143
		       struct ftrace_ops *ignored, struct pt_regs *regs)
5144
{
5145
	struct ftrace_ops *op;
5146
	int bit;
5147

5148 5149 5150
	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;
5151

5152 5153 5154 5155 5156
	/*
	 * Some of the ops may be dynamically allocated,
	 * they must be freed after a synchronize_sched().
	 */
	preempt_disable_notrace();
5157
	do_for_each_ftrace_op(op, ftrace_ops_list) {
5158
		if (ftrace_ops_test(op, ip, regs)) {
5159 5160
			if (FTRACE_WARN_ON(!op->func)) {
				pr_warn("op=%p %pS\n", op, op);
5161 5162
				goto out;
			}
5163
			op->func(ip, parent_ip, op, regs);
5164
		}
5165
	} while_for_each_ftrace_op(op);
5166
out:
5167
	preempt_enable_notrace();
5168
	trace_clear_recursion(bit);
5169 5170
}

5171 5172 5173 5174 5175
/*
 * 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.
5176 5177 5178
 * 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
5179 5180
 * being NULL, or CONFIG_DYNAMIC_FTRACE_WITH_REGS.
 * Note, CONFIG_DYNAMIC_FTRACE_WITH_REGS expects a full regs to be saved.
5181 5182
 * An architecture can pass partial regs with ftrace_ops and still
 * set the ARCH_SUPPORT_FTARCE_OPS.
5183 5184 5185
 */
#if ARCH_SUPPORTS_FTRACE_OPS
static void ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
5186
				 struct ftrace_ops *op, struct pt_regs *regs)
5187
{
5188
	__ftrace_ops_list_func(ip, parent_ip, NULL, regs);
5189 5190 5191 5192
}
#else
static void ftrace_ops_no_ops(unsigned long ip, unsigned long parent_ip)
{
5193
	__ftrace_ops_list_func(ip, parent_ip, NULL, NULL);
5194 5195 5196
}
#endif

5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 5208 5209 5210 5211 5212 5213 5214 5215
/*
 * 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);
}

5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233 5234 5235 5236 5237 5238 5239
/**
 * 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 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;
}

5240
static void clear_ftrace_swapper(void)
S
Steven Rostedt 已提交
5241 5242
{
	struct task_struct *p;
5243
	int cpu;
S
Steven Rostedt 已提交
5244

5245 5246 5247
	get_online_cpus();
	for_each_online_cpu(cpu) {
		p = idle_task(cpu);
S
Steven Rostedt 已提交
5248
		clear_tsk_trace_trace(p);
5249 5250 5251
	}
	put_online_cpus();
}
S
Steven Rostedt 已提交
5252

5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263
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
5264 5265
}

5266 5267 5268 5269
static void clear_ftrace_pid(struct pid *pid)
{
	struct task_struct *p;

5270
	rcu_read_lock();
5271 5272 5273
	do_each_pid_task(pid, PIDTYPE_PID, p) {
		clear_tsk_trace_trace(p);
	} while_each_pid_task(pid, PIDTYPE_PID, p);
5274 5275
	rcu_read_unlock();

5276 5277 5278 5279
	put_pid(pid);
}

static void set_ftrace_pid(struct pid *pid)
S
Steven Rostedt 已提交
5280 5281 5282
{
	struct task_struct *p;

5283
	rcu_read_lock();
S
Steven Rostedt 已提交
5284 5285 5286
	do_each_pid_task(pid, PIDTYPE_PID, p) {
		set_tsk_trace_trace(p);
	} while_each_pid_task(pid, PIDTYPE_PID, p);
5287
	rcu_read_unlock();
S
Steven Rostedt 已提交
5288 5289
}

5290
static void clear_ftrace_pid_task(struct pid *pid)
5291
{
5292
	if (pid == ftrace_swapper_pid)
5293 5294
		clear_ftrace_swapper();
	else
5295
		clear_ftrace_pid(pid);
5296 5297 5298 5299 5300 5301 5302 5303 5304 5305
}

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

5306
static int ftrace_pid_add(int p)
5307
{
S
Steven Rostedt 已提交
5308
	struct pid *pid;
5309 5310
	struct ftrace_pid *fpid;
	int ret = -EINVAL;
5311

5312
	mutex_lock(&ftrace_lock);
5313

5314 5315 5316 5317
	if (!p)
		pid = ftrace_swapper_pid;
	else
		pid = find_get_pid(p);
5318

5319 5320
	if (!pid)
		goto out;
5321

5322
	ret = 0;
5323

5324 5325 5326
	list_for_each_entry(fpid, &ftrace_pids, list)
		if (fpid->pid == pid)
			goto out_put;
S
Steven Rostedt 已提交
5327

5328
	ret = -ENOMEM;
5329

5330 5331 5332
	fpid = kmalloc(sizeof(*fpid), GFP_KERNEL);
	if (!fpid)
		goto out_put;
5333

5334 5335
	list_add(&fpid->list, &ftrace_pids);
	fpid->pid = pid;
5336

5337
	set_ftrace_pid_task(pid);
S
Steven Rostedt 已提交
5338

5339
	ftrace_update_pid_func();
5340 5341

	ftrace_startup_all(0);
5342 5343 5344 5345 5346 5347 5348

	mutex_unlock(&ftrace_lock);
	return 0;

out_put:
	if (pid != ftrace_swapper_pid)
		put_pid(pid);
S
Steven Rostedt 已提交
5349

5350 5351 5352 5353 5354 5355 5356 5357
out:
	mutex_unlock(&ftrace_lock);
	return ret;
}

static void ftrace_pid_reset(void)
{
	struct ftrace_pid *fpid, *safe;
S
Steven Rostedt 已提交
5358

5359 5360 5361 5362 5363 5364 5365 5366
	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);
5367 5368 5369
	}

	ftrace_update_pid_func();
5370
	ftrace_startup_all(0);
5371

S
Steven Rostedt 已提交
5372
	mutex_unlock(&ftrace_lock);
5373
}
5374

5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401 5402
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) {
5403
		seq_puts(m, "no pid\n");
5404 5405 5406 5407
		return 0;
	}

	if (fpid->pid == ftrace_swapper_pid)
5408
		seq_puts(m, "swapper tasks\n");
5409 5410 5411 5412 5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429 5430 5431 5432 5433 5434 5435 5436
	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;
}

5437 5438 5439 5440
static ssize_t
ftrace_pid_write(struct file *filp, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
5441
	char buf[64], *tmp;
5442 5443 5444 5445 5446 5447 5448 5449 5450 5451 5452
	long val;
	int ret;

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

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

	buf[cnt] = 0;

5453 5454 5455 5456
	/*
	 * Allow "echo > set_ftrace_pid" or "echo -n '' > set_ftrace_pid"
	 * to clean the filter quietly.
	 */
5457 5458
	tmp = strstrip(buf);
	if (strlen(tmp) == 0)
5459 5460
		return 1;

5461
	ret = kstrtol(tmp, 10, &val);
5462 5463 5464
	if (ret < 0)
		return ret;

5465
	ret = ftrace_pid_add(val);
5466

5467 5468
	return ret ? ret : cnt;
}
5469

5470 5471 5472 5473 5474
static int
ftrace_pid_release(struct inode *inode, struct file *file)
{
	if (file->f_mode & FMODE_READ)
		seq_release(inode, file);
5475

5476
	return 0;
5477 5478
}

5479
static const struct file_operations ftrace_pid_fops = {
5480 5481 5482
	.open		= ftrace_pid_open,
	.write		= ftrace_pid_write,
	.read		= seq_read,
5483
	.llseek		= tracing_lseek,
5484
	.release	= ftrace_pid_release,
5485 5486
};

5487
static __init int ftrace_init_tracefs(void)
5488 5489 5490 5491
{
	struct dentry *d_tracer;

	d_tracer = tracing_init_dentry();
5492
	if (IS_ERR(d_tracer))
5493 5494
		return 0;

5495
	ftrace_init_dyn_tracefs(d_tracer);
5496

5497 5498
	trace_create_file("set_ftrace_pid", 0644, d_tracer,
			    NULL, &ftrace_pid_fops);
5499

5500
	ftrace_profile_tracefs(d_tracer);
5501

5502 5503
	return 0;
}
5504
fs_initcall(ftrace_init_tracefs);
5505

S
Steven Rostedt 已提交
5506
/**
5507
 * ftrace_kill - kill ftrace
S
Steven Rostedt 已提交
5508 5509 5510 5511 5512
 *
 * 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.
 */
5513
void ftrace_kill(void)
S
Steven Rostedt 已提交
5514 5515 5516 5517 5518 5519
{
	ftrace_disabled = 1;
	ftrace_enabled = 0;
	clear_ftrace_function();
}

5520 5521 5522 5523 5524 5525 5526 5527
/**
 * Test if ftrace is dead or not.
 */
int ftrace_is_dead(void)
{
	return ftrace_disabled;
}

5528
/**
5529 5530
 * register_ftrace_function - register a function for profiling
 * @ops - ops structure that holds the function for profiling.
5531
 *
5532 5533 5534 5535 5536 5537
 * 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.
5538
 */
5539
int register_ftrace_function(struct ftrace_ops *ops)
5540
{
5541
	int ret = -1;
5542

5543 5544
	ftrace_ops_init(ops);

S
Steven Rostedt 已提交
5545
	mutex_lock(&ftrace_lock);
5546

5547
	ret = ftrace_startup(ops, 0);
5548

S
Steven Rostedt 已提交
5549
	mutex_unlock(&ftrace_lock);
5550

5551
	return ret;
5552
}
5553
EXPORT_SYMBOL_GPL(register_ftrace_function);
5554 5555

/**
5556
 * unregister_ftrace_function - unregister a function for profiling.
5557 5558 5559 5560 5561 5562 5563 5564
 * @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
5565
	mutex_lock(&ftrace_lock);
5566
	ret = ftrace_shutdown(ops, 0);
S
Steven Rostedt 已提交
5567
	mutex_unlock(&ftrace_lock);
5568 5569 5570

	return ret;
}
5571
EXPORT_SYMBOL_GPL(unregister_ftrace_function);
5572

I
Ingo Molnar 已提交
5573
int
5574
ftrace_enable_sysctl(struct ctl_table *table, int write,
5575
		     void __user *buffer, size_t *lenp,
5576 5577
		     loff_t *ppos)
{
5578
	int ret = -ENODEV;
5579

S
Steven Rostedt 已提交
5580
	mutex_lock(&ftrace_lock);
5581

5582 5583 5584 5585
	if (unlikely(ftrace_disabled))
		goto out;

	ret = proc_dointvec(table, write, buffer, lenp, ppos);
5586

5587
	if (ret || !write || (last_ftrace_enabled == !!ftrace_enabled))
5588 5589
		goto out;

5590
	last_ftrace_enabled = !!ftrace_enabled;
5591 5592 5593 5594

	if (ftrace_enabled) {

		/* we are starting ftrace again */
5595 5596
		if (ftrace_ops_list != &ftrace_list_end)
			update_ftrace_function();
5597

5598 5599
		ftrace_startup_sysctl();

5600 5601 5602 5603 5604 5605 5606 5607
	} else {
		/* stopping ftrace calls (just send to ftrace_stub) */
		ftrace_trace_function = ftrace_stub;

		ftrace_shutdown_sysctl();
	}

 out:
S
Steven Rostedt 已提交
5608
	mutex_unlock(&ftrace_lock);
5609
	return ret;
5610
}
5611

5612
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
5613

5614 5615 5616 5617 5618 5619 5620
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,
5621
	/* trampoline_size is only needed for dynamically allocated tramps */
5622 5623 5624 5625
#endif
	ASSIGN_OPS_HASH(graph_ops, &global_ops.local_hash)
};

5626 5627 5628 5629 5630
int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace)
{
	return 0;
}

5631 5632 5633
/* The callbacks that hook a function */
trace_func_graph_ret_t ftrace_graph_return =
			(trace_func_graph_ret_t)ftrace_stub;
5634
trace_func_graph_ent_t ftrace_graph_entry = ftrace_graph_entry_stub;
5635
static trace_func_graph_ent_t __ftrace_graph_entry = ftrace_graph_entry_stub;
5636 5637 5638 5639 5640 5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660 5661 5662 5663 5664 5665

/* 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) {
5666
			atomic_set(&t->tracing_graph_pause, 0);
5667
			atomic_set(&t->trace_overrun, 0);
5668 5669 5670 5671
			t->curr_ret_stack = -1;
			/* Make sure the tasks see the -1 first: */
			smp_wmb();
			t->ret_stack = ret_stack_list[start++];
5672 5673 5674 5675 5676 5677 5678 5679 5680 5681 5682
		}
	} 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;
}

5683
static void
5684 5685
ftrace_graph_probe_sched_switch(void *ignore,
			struct task_struct *prev, struct task_struct *next)
5686 5687 5688 5689
{
	unsigned long long timestamp;
	int index;

5690 5691 5692 5693 5694 5695 5696
	/*
	 * 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;

5697 5698 5699 5700 5701 5702 5703 5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714
	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;
}

5715
/* Allocate a return stack for each task */
5716
static int start_graph_tracing(void)
5717 5718
{
	struct ftrace_ret_stack **ret_stack_list;
5719
	int ret, cpu;
5720 5721 5722 5723 5724 5725 5726 5727

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

	if (!ret_stack_list)
		return -ENOMEM;

5728
	/* The cpu_boot init_task->ret_stack will never be freed */
5729 5730
	for_each_online_cpu(cpu) {
		if (!idle_task(cpu)->ret_stack)
5731
			ftrace_graph_init_idle_task(idle_task(cpu), cpu);
5732
	}
5733

5734 5735 5736 5737
	do {
		ret = alloc_retstack_tasklist(ret_stack_list);
	} while (ret == -EAGAIN);

5738
	if (!ret) {
5739
		ret = register_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
5740 5741 5742 5743 5744
		if (ret)
			pr_info("ftrace_graph: Couldn't activate tracepoint"
				" probe to kernel_sched_switch\n");
	}

5745 5746 5747 5748
	kfree(ret_stack_list);
	return ret;
}

5749 5750 5751 5752 5753 5754 5755 5756 5757 5758 5759 5760 5761 5762 5763 5764 5765 5766 5767 5768 5769
/*
 * 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;
}

5770 5771 5772 5773 5774 5775 5776 5777 5778 5779 5780 5781 5782 5783 5784 5785
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)
{
5786 5787 5788 5789 5790 5791 5792 5793 5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804
	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)
5805
		ftrace_graph_entry = ftrace_graph_entry_test;
5806 5807
	else
		ftrace_graph_entry = __ftrace_graph_entry;
5808 5809
}

5810 5811 5812 5813
static struct notifier_block ftrace_suspend_notifier = {
	.notifier_call = ftrace_suspend_notifier_call,
};

5814 5815
int register_ftrace_graph(trace_func_graph_ret_t retfunc,
			trace_func_graph_ent_t entryfunc)
5816
{
5817 5818
	int ret = 0;

S
Steven Rostedt 已提交
5819
	mutex_lock(&ftrace_lock);
5820

5821
	/* we currently allow only one tracer registered at a time */
5822
	if (ftrace_graph_active) {
5823 5824 5825 5826
		ret = -EBUSY;
		goto out;
	}

5827 5828
	register_pm_notifier(&ftrace_suspend_notifier);

5829
	ftrace_graph_active++;
5830
	ret = start_graph_tracing();
5831
	if (ret) {
5832
		ftrace_graph_active--;
5833 5834
		goto out;
	}
5835

5836
	ftrace_graph_return = retfunc;
5837 5838 5839 5840 5841 5842 5843 5844 5845 5846

	/*
	 * 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();
5847

5848
	ret = ftrace_startup(&graph_ops, FTRACE_START_FUNC_RET);
5849
out:
S
Steven Rostedt 已提交
5850
	mutex_unlock(&ftrace_lock);
5851
	return ret;
5852 5853
}

5854
void unregister_ftrace_graph(void)
5855
{
S
Steven Rostedt 已提交
5856
	mutex_lock(&ftrace_lock);
5857

5858
	if (unlikely(!ftrace_graph_active))
5859 5860
		goto out;

5861
	ftrace_graph_active--;
5862
	ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub;
5863
	ftrace_graph_entry = ftrace_graph_entry_stub;
5864
	__ftrace_graph_entry = ftrace_graph_entry_stub;
5865
	ftrace_shutdown(&graph_ops, FTRACE_STOP_FUNC_RET);
5866
	unregister_pm_notifier(&ftrace_suspend_notifier);
5867
	unregister_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
5868

5869 5870 5871 5872 5873 5874 5875 5876 5877 5878 5879
#ifdef CONFIG_DYNAMIC_FTRACE
	/*
	 * Function graph does not allocate the trampoline, but
	 * other global_ops do. We need to reset the ALLOC_TRAMP flag
	 * if one was used.
	 */
	global_ops.trampoline = save_global_trampoline;
	if (save_global_flags & FTRACE_OPS_FL_ALLOC_TRAMP)
		global_ops.flags |= FTRACE_OPS_FL_ALLOC_TRAMP;
#endif

5880
 out:
S
Steven Rostedt 已提交
5881
	mutex_unlock(&ftrace_lock);
5882
}
5883

5884 5885 5886 5887 5888 5889 5890 5891
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
5892
	/* make curr_ret_stack visible before we add the ret_stack */
5893 5894 5895 5896 5897 5898 5899 5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926
	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);
	}
}

5927
/* Allocate a return stack for newly created task */
5928
void ftrace_graph_init_task(struct task_struct *t)
5929
{
5930 5931
	/* Make sure we do not use the parent ret_stack */
	t->ret_stack = NULL;
5932
	t->curr_ret_stack = -1;
5933

5934
	if (ftrace_graph_active) {
5935 5936 5937
		struct ftrace_ret_stack *ret_stack;

		ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH
5938 5939
				* sizeof(struct ftrace_ret_stack),
				GFP_KERNEL);
5940
		if (!ret_stack)
5941
			return;
5942
		graph_init_task(t, ret_stack);
5943
	}
5944 5945
}

5946
void ftrace_graph_exit_task(struct task_struct *t)
5947
{
5948 5949
	struct ftrace_ret_stack	*ret_stack = t->ret_stack;

5950
	t->ret_stack = NULL;
5951 5952 5953 5954
	/* NULL must become visible to IRQs before we free it: */
	barrier();

	kfree(ret_stack);
5955
}
5956
#endif