ftrace.c 142.2 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
#ifdef CONFIG_DYNAMIC_FTRACE
66 67 68
#define INIT_OPS_HASH(opsname)	\
	.func_hash		= &opsname.local_hash,			\
	.local_hash.regex_lock	= __MUTEX_INITIALIZER(opsname.local_hash.regex_lock),
69 70 71
#define ASSIGN_OPS_HASH(opsname, val) \
	.func_hash		= val, \
	.local_hash.regex_lock	= __MUTEX_INITIALIZER(opsname.local_hash.regex_lock),
72
#else
73
#define INIT_OPS_HASH(opsname)
74
#define ASSIGN_OPS_HASH(opsname, val)
75 76
#endif

77 78
static struct ftrace_ops ftrace_list_end __read_mostly = {
	.func		= ftrace_stub,
79
	.flags		= FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_STUB,
80
	INIT_OPS_HASH(ftrace_list_end)
81 82
};

83 84
/* ftrace_enabled is a method to turn ftrace on or off */
int ftrace_enabled __read_mostly;
85
static int last_ftrace_enabled;
86

87 88
/* Current function tracing op */
struct ftrace_ops *function_trace_op __read_mostly = &ftrace_list_end;
89 90
/* What to set function_trace_op to */
static struct ftrace_ops *set_function_trace_op;
91

92
static bool ftrace_pids_enabled(struct ftrace_ops *ops)
93
{
94 95 96 97 98 99 100 101
	struct trace_array *tr;

	if (!(ops->flags & FTRACE_OPS_FL_PID) || !ops->private)
		return false;

	tr = ops->private;

	return tr->function_pids != NULL;
102 103 104 105
}

static void ftrace_update_trampoline(struct ftrace_ops *ops);

106 107 108 109 110 111
/*
 * ftrace_disabled is set when an anomaly is discovered.
 * ftrace_disabled is much stronger than ftrace_enabled.
 */
static int ftrace_disabled __read_mostly;

112
static DEFINE_MUTEX(ftrace_lock);
113

114
static struct ftrace_ops *ftrace_ops_list __read_mostly = &ftrace_list_end;
115
ftrace_func_t ftrace_trace_function __read_mostly = ftrace_stub;
116
static struct ftrace_ops global_ops;
117

118 119
#if ARCH_SUPPORTS_FTRACE_OPS
static void ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
120
				 struct ftrace_ops *op, struct pt_regs *regs);
121 122 123 124 125
#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
126

127 128
/*
 * Traverse the ftrace_global_list, invoking all entries.  The reason that we
129
 * can use rcu_dereference_raw_notrace() is that elements removed from this list
130
 * are simply leaked, so there is no need to interact with a grace-period
131
 * mechanism.  The rcu_dereference_raw_notrace() calls are needed to handle
132 133 134 135 136
 * concurrent insertions into the ftrace_global_list.
 *
 * Silly Alpha and silly pointer-speculation compiler optimizations!
 */
#define do_for_each_ftrace_op(op, list)			\
137
	op = rcu_dereference_raw_notrace(list);			\
138 139 140 141 142 143
	do

/*
 * Optimized for just a single item in the list (as that is the normal case).
 */
#define while_for_each_ftrace_op(op)				\
144
	while (likely(op = rcu_dereference_raw_notrace((op)->next)) &&	\
145 146
	       unlikely((op) != &ftrace_list_end))

147 148 149 150
static inline void ftrace_ops_init(struct ftrace_ops *ops)
{
#ifdef CONFIG_DYNAMIC_FTRACE
	if (!(ops->flags & FTRACE_OPS_FL_INITIALIZED)) {
151 152
		mutex_init(&ops->local_hash.regex_lock);
		ops->func_hash = &ops->local_hash;
153 154 155 156 157
		ops->flags |= FTRACE_OPS_FL_INITIALIZED;
	}
#endif
}

158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178
/**
 * 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;
}

179
static void ftrace_pid_func(unsigned long ip, unsigned long parent_ip,
180
			    struct ftrace_ops *op, struct pt_regs *regs)
181
{
182 183 184
	struct trace_array *tr = op->private;

	if (tr && this_cpu_read(tr->trace_buffer.data->ftrace_ignore_pid))
185 186
		return;

187
	op->saved_func(ip, parent_ip, op, regs);
188 189
}

190
/**
191
 * clear_ftrace_function - reset the ftrace function
192
 *
193 194
 * This NULLs the ftrace function and in essence stops
 * tracing.  There may be lag
195
 */
196
void clear_ftrace_function(void)
197
{
198 199 200
	ftrace_trace_function = ftrace_stub;
}

201
static void per_cpu_ops_disable_all(struct ftrace_ops *ops)
202 203 204 205 206 207 208
{
	int cpu;

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

209
static int per_cpu_ops_alloc(struct ftrace_ops *ops)
210 211 212
{
	int __percpu *disabled;

213 214 215
	if (WARN_ON_ONCE(!(ops->flags & FTRACE_OPS_FL_PER_CPU)))
		return -EINVAL;

216 217 218 219 220
	disabled = alloc_percpu(int);
	if (!disabled)
		return -ENOMEM;

	ops->disabled = disabled;
221
	per_cpu_ops_disable_all(ops);
222 223 224
	return 0;
}

225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241
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();
}

242 243
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static void update_function_graph_func(void);
244 245 246 247 248

/* Both enabled by default (can be cleared by function_graph tracer flags */
static bool fgraph_sleep_time = true;
static bool fgraph_graph_time = true;

249 250 251 252
#else
static inline void update_function_graph_func(void) { }
#endif

253 254 255 256

static ftrace_func_t ftrace_ops_get_list_func(struct ftrace_ops *ops)
{
	/*
257
	 * If this is a dynamic, RCU, or per CPU ops, or we force list func,
258 259
	 * then it needs to call the list anyway.
	 */
260 261
	if (ops->flags & (FTRACE_OPS_FL_DYNAMIC | FTRACE_OPS_FL_PER_CPU |
			  FTRACE_OPS_FL_RCU) || FTRACE_FORCE_LIST_FUNC)
262 263 264 265 266
		return ftrace_ops_list_func;

	return ftrace_ops_get_func(ops);
}

267 268 269 270
static void update_ftrace_function(void)
{
	ftrace_func_t func;

271 272 273 274 275 276 277 278 279 280 281
	/*
	 * 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;

282 283
	/*
	 * If we are at the end of the list and this ops is
284 285
	 * recursion safe and not dynamic and the arch supports passing ops,
	 * then have the mcount trampoline call the function directly.
286
	 */
287
	} else if (ftrace_ops_list->next == &ftrace_list_end) {
288
		func = ftrace_ops_get_list_func(ftrace_ops_list);
289

290 291
	} else {
		/* Just use the default ftrace_ops */
292
		set_function_trace_op = &ftrace_list_end;
293
		func = ftrace_ops_list_func;
294
	}
295

296 297
	update_function_graph_func();

298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
	/* 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 */

341 342 343
	ftrace_trace_function = func;
}

344 345 346 347 348
int using_ftrace_ops_list_func(void)
{
	return ftrace_trace_function == ftrace_ops_list_func;
}

349
static void add_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
350
{
351
	ops->next = *list;
352
	/*
353
	 * We are entering ops into the list but another
354 355
	 * CPU might be walking that list. We need to make sure
	 * the ops->next pointer is valid before another CPU sees
356
	 * the ops pointer included into the list.
357
	 */
358
	rcu_assign_pointer(*list, ops);
359 360
}

361
static int remove_ftrace_ops(struct ftrace_ops **list, struct ftrace_ops *ops)
362 363 364 365
{
	struct ftrace_ops **p;

	/*
366 367
	 * If we are removing the last function, then simply point
	 * to the ftrace_stub.
368
	 */
369 370
	if (*list == ops && ops->next == &ftrace_list_end) {
		*list = &ftrace_list_end;
S
Steven Rostedt 已提交
371
		return 0;
372 373
	}

374
	for (p = list; *p != &ftrace_list_end; p = &(*p)->next)
375 376 377
		if (*p == ops)
			break;

S
Steven Rostedt 已提交
378 379
	if (*p != ops)
		return -1;
380 381

	*p = (*p)->next;
382 383
	return 0;
}
384

385 386
static void ftrace_update_trampoline(struct ftrace_ops *ops);

387 388
static int __register_ftrace_function(struct ftrace_ops *ops)
{
389 390 391
	if (ops->flags & FTRACE_OPS_FL_DELETED)
		return -EINVAL;

392 393 394
	if (WARN_ON(ops->flags & FTRACE_OPS_FL_ENABLED))
		return -EBUSY;

395
#ifndef CONFIG_DYNAMIC_FTRACE_WITH_REGS
396 397 398 399 400 401 402 403 404 405 406 407 408
	/*
	 * 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

409 410 411
	if (!core_kernel_data((unsigned long)ops))
		ops->flags |= FTRACE_OPS_FL_DYNAMIC;

412 413
	if (ops->flags & FTRACE_OPS_FL_PER_CPU) {
		if (per_cpu_ops_alloc(ops))
414
			return -ENOMEM;
415 416 417
	}

	add_ftrace_ops(&ftrace_ops_list, ops);
418

419 420 421
	/* Always save the function, and reset at unregistering */
	ops->saved_func = ops->func;

422
	if (ftrace_pids_enabled(ops))
423 424
		ops->func = ftrace_pid_func;

425 426
	ftrace_update_trampoline(ops);

427 428 429 430 431 432 433 434 435 436
	if (ftrace_enabled)
		update_ftrace_function();

	return 0;
}

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

437 438 439
	if (WARN_ON(!(ops->flags & FTRACE_OPS_FL_ENABLED)))
		return -EBUSY;

440
	ret = remove_ftrace_ops(&ftrace_ops_list, ops);
441

442 443
	if (ret < 0)
		return ret;
444

445 446
	if (ftrace_enabled)
		update_ftrace_function();
447

448 449
	ops->func = ops->saved_func;

S
Steven Rostedt 已提交
450
	return 0;
451 452
}

453 454
static void ftrace_update_pid_func(void)
{
455 456
	struct ftrace_ops *op;

457
	/* Only do something if we are tracing something */
458
	if (ftrace_trace_function == ftrace_stub)
459
		return;
460

461 462
	do_for_each_ftrace_op(op, ftrace_ops_list) {
		if (op->flags & FTRACE_OPS_FL_PID) {
463 464
			op->func = ftrace_pids_enabled(op) ?
				ftrace_pid_func : op->saved_func;
465 466 467 468
			ftrace_update_trampoline(op);
		}
	} while_for_each_ftrace_op(op);

469
	update_ftrace_function();
470 471
}

472 473 474 475 476
#ifdef CONFIG_FUNCTION_PROFILER
struct ftrace_profile {
	struct hlist_node		node;
	unsigned long			ip;
	unsigned long			counter;
477 478
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	unsigned long long		time;
479
	unsigned long long		time_squared;
480
#endif
481 482
};

483 484 485 486
struct ftrace_profile_page {
	struct ftrace_profile_page	*next;
	unsigned long			index;
	struct ftrace_profile		records[];
487 488
};

489 490 491 492 493 494 495 496
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;
};

497 498
#define PROFILE_RECORDS_SIZE						\
	(PAGE_SIZE - offsetof(struct ftrace_profile_page, records))
499

500 501
#define PROFILES_PER_PAGE					\
	(PROFILE_RECORDS_SIZE / sizeof(struct ftrace_profile))
502

503 504 505
static int ftrace_profile_enabled __read_mostly;

/* ftrace_profile_lock - synchronize the enable and disable of the profiler */
S
Steven Rostedt 已提交
506 507
static DEFINE_MUTEX(ftrace_profile_lock);

508
static DEFINE_PER_CPU(struct ftrace_profile_stat, ftrace_profile_stats);
509

510 511
#define FTRACE_PROFILE_HASH_BITS 10
#define FTRACE_PROFILE_HASH_SIZE (1 << FTRACE_PROFILE_HASH_BITS)
512

S
Steven Rostedt 已提交
513 514 515
static void *
function_stat_next(void *v, int idx)
{
516 517
	struct ftrace_profile *rec = v;
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
518

519
	pg = (struct ftrace_profile_page *)((unsigned long)rec & PAGE_MASK);
S
Steven Rostedt 已提交
520 521

 again:
L
Li Zefan 已提交
522 523 524
	if (idx != 0)
		rec++;

S
Steven Rostedt 已提交
525 526 527 528 529
	if ((void *)rec >= (void *)&pg->records[pg->index]) {
		pg = pg->next;
		if (!pg)
			return NULL;
		rec = &pg->records[0];
530 531
		if (!rec->counter)
			goto again;
S
Steven Rostedt 已提交
532 533 534 535 536 537 538
	}

	return rec;
}

static void *function_stat_start(struct tracer_stat *trace)
{
539 540 541 542 543 544 545
	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
546 547
}

548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563
#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
564 565
static int function_stat_cmp(void *p1, void *p2)
{
566 567
	struct ftrace_profile *a = p1;
	struct ftrace_profile *b = p2;
S
Steven Rostedt 已提交
568 569 570 571 572 573 574 575

	if (a->counter < b->counter)
		return -1;
	if (a->counter > b->counter)
		return 1;
	else
		return 0;
}
576
#endif
S
Steven Rostedt 已提交
577 578 579

static int function_stat_headers(struct seq_file *m)
{
580
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
581 582 583 584
	seq_puts(m, "  Function                               "
		 "Hit    Time            Avg             s^2\n"
		    "  --------                               "
		 "---    ----            ---             ---\n");
585
#else
586 587
	seq_puts(m, "  Function                               Hit\n"
		    "  --------                               ---\n");
588
#endif
S
Steven Rostedt 已提交
589 590 591 592 593
	return 0;
}

static int function_stat_show(struct seq_file *m, void *v)
{
594
	struct ftrace_profile *rec = v;
S
Steven Rostedt 已提交
595
	char str[KSYM_SYMBOL_LEN];
596
	int ret = 0;
597
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
598 599
	static struct trace_seq s;
	unsigned long long avg;
600
	unsigned long long stddev;
601
#endif
602 603 604 605 606 607 608
	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
609

610 611 612 613 614 615 616
#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
617
	kallsyms_lookup(rec->ip, NULL, NULL, NULL, str);
618 619 620
	seq_printf(m, "  %-30.30s  %10lu", str, rec->counter);

#ifdef CONFIG_FUNCTION_GRAPH_TRACER
621
	seq_puts(m, "    ");
622

623 624 625 626
	/* Sample standard deviation (s^2) */
	if (rec->counter <= 1)
		stddev = 0;
	else {
627 628 629 630 631 632 633
		/*
		 * 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;

634 635 636 637
		/*
		 * Divide only 1000 for ns^2 -> us^2 conversion.
		 * trace_print_graph_duration will divide 1000 again.
		 */
638
		do_div(stddev, rec->counter * (rec->counter - 1) * 1000);
639 640
	}

641 642 643 644
	trace_seq_init(&s);
	trace_print_graph_duration(rec->time, &s);
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(avg, &s);
645 646
	trace_seq_puts(&s, "    ");
	trace_print_graph_duration(stddev, &s);
647 648 649
	trace_print_seq(m, &s);
#endif
	seq_putc(m, '\n');
650 651
out:
	mutex_unlock(&ftrace_profile_lock);
S
Steven Rostedt 已提交
652

653
	return ret;
S
Steven Rostedt 已提交
654 655
}

656
static void ftrace_profile_reset(struct ftrace_profile_stat *stat)
S
Steven Rostedt 已提交
657
{
658
	struct ftrace_profile_page *pg;
S
Steven Rostedt 已提交
659

660
	pg = stat->pages = stat->start;
S
Steven Rostedt 已提交
661

662 663 664 665
	while (pg) {
		memset(pg->records, 0, PROFILE_RECORDS_SIZE);
		pg->index = 0;
		pg = pg->next;
S
Steven Rostedt 已提交
666 667
	}

668
	memset(stat->hash, 0,
669 670
	       FTRACE_PROFILE_HASH_SIZE * sizeof(struct hlist_head));
}
S
Steven Rostedt 已提交
671

672
int ftrace_profile_pages_init(struct ftrace_profile_stat *stat)
673 674
{
	struct ftrace_profile_page *pg;
675 676
	int functions;
	int pages;
677
	int i;
S
Steven Rostedt 已提交
678

679
	/* If we already allocated, do nothing */
680
	if (stat->pages)
681
		return 0;
S
Steven Rostedt 已提交
682

683 684
	stat->pages = (void *)get_zeroed_page(GFP_KERNEL);
	if (!stat->pages)
685
		return -ENOMEM;
S
Steven Rostedt 已提交
686

687 688 689 690 691 692 693 694 695 696 697 698 699
#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

700
	pg = stat->start = stat->pages;
S
Steven Rostedt 已提交
701

702 703
	pages = DIV_ROUND_UP(functions, PROFILES_PER_PAGE);

704
	for (i = 1; i < pages; i++) {
705 706
		pg->next = (void *)get_zeroed_page(GFP_KERNEL);
		if (!pg->next)
707
			goto out_free;
708 709 710 711
		pg = pg->next;
	}

	return 0;
712 713 714 715 716 717 718 719 720 721 722 723 724 725

 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
726 727
}

728
static int ftrace_profile_init_cpu(int cpu)
S
Steven Rostedt 已提交
729
{
730
	struct ftrace_profile_stat *stat;
731
	int size;
S
Steven Rostedt 已提交
732

733 734 735
	stat = &per_cpu(ftrace_profile_stats, cpu);

	if (stat->hash) {
736
		/* If the profile is already created, simply reset it */
737
		ftrace_profile_reset(stat);
738 739
		return 0;
	}
S
Steven Rostedt 已提交
740

741 742 743 744 745
	/*
	 * 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
746

747
	stat->hash = kzalloc(sizeof(struct hlist_head) * size, GFP_KERNEL);
748

749
	if (!stat->hash)
750 751
		return -ENOMEM;

752
	/* Preallocate the function profiling pages */
753 754 755
	if (ftrace_profile_pages_init(stat) < 0) {
		kfree(stat->hash);
		stat->hash = NULL;
756 757 758 759
		return -ENOMEM;
	}

	return 0;
S
Steven Rostedt 已提交
760 761
}

762 763 764 765 766
static int ftrace_profile_init(void)
{
	int cpu;
	int ret = 0;

767
	for_each_possible_cpu(cpu) {
768 769 770 771 772 773 774 775
		ret = ftrace_profile_init_cpu(cpu);
		if (ret)
			break;
	}

	return ret;
}

776
/* interrupts must be disabled */
777 778
static struct ftrace_profile *
ftrace_find_profiled_func(struct ftrace_profile_stat *stat, unsigned long ip)
S
Steven Rostedt 已提交
779
{
780
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
781 782 783
	struct hlist_head *hhd;
	unsigned long key;

784
	key = hash_long(ip, FTRACE_PROFILE_HASH_BITS);
785
	hhd = &stat->hash[key];
S
Steven Rostedt 已提交
786 787 788 789

	if (hlist_empty(hhd))
		return NULL;

790
	hlist_for_each_entry_rcu_notrace(rec, hhd, node) {
S
Steven Rostedt 已提交
791
		if (rec->ip == ip)
792 793 794 795 796 797
			return rec;
	}

	return NULL;
}

798 799
static void ftrace_add_profile(struct ftrace_profile_stat *stat,
			       struct ftrace_profile *rec)
800 801 802
{
	unsigned long key;

803
	key = hash_long(rec->ip, FTRACE_PROFILE_HASH_BITS);
804
	hlist_add_head_rcu(&rec->node, &stat->hash[key]);
805 806
}

807 808 809
/*
 * The memory is already allocated, this simply finds a new record to use.
 */
810
static struct ftrace_profile *
811
ftrace_profile_alloc(struct ftrace_profile_stat *stat, unsigned long ip)
812 813 814
{
	struct ftrace_profile *rec = NULL;

815
	/* prevent recursion (from NMIs) */
816
	if (atomic_inc_return(&stat->disabled) != 1)
817 818 819
		goto out;

	/*
820 821
	 * Try to find the function again since an NMI
	 * could have added it
822
	 */
823
	rec = ftrace_find_profiled_func(stat, ip);
824
	if (rec)
825
		goto out;
826

827 828 829 830
	if (stat->pages->index == PROFILES_PER_PAGE) {
		if (!stat->pages->next)
			goto out;
		stat->pages = stat->pages->next;
S
Steven Rostedt 已提交
831
	}
832

833
	rec = &stat->pages->records[stat->pages->index++];
834
	rec->ip = ip;
835
	ftrace_add_profile(stat, rec);
836

S
Steven Rostedt 已提交
837
 out:
838
	atomic_dec(&stat->disabled);
S
Steven Rostedt 已提交
839 840 841 842 843

	return rec;
}

static void
844
function_profile_call(unsigned long ip, unsigned long parent_ip,
845
		      struct ftrace_ops *ops, struct pt_regs *regs)
S
Steven Rostedt 已提交
846
{
847
	struct ftrace_profile_stat *stat;
848
	struct ftrace_profile *rec;
S
Steven Rostedt 已提交
849 850 851 852 853 854
	unsigned long flags;

	if (!ftrace_profile_enabled)
		return;

	local_irq_save(flags);
855

856
	stat = this_cpu_ptr(&ftrace_profile_stats);
857
	if (!stat->hash || !ftrace_profile_enabled)
858 859 860
		goto out;

	rec = ftrace_find_profiled_func(stat, ip);
861
	if (!rec) {
862
		rec = ftrace_profile_alloc(stat, ip);
863 864 865
		if (!rec)
			goto out;
	}
S
Steven Rostedt 已提交
866 867 868 869 870 871

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

872 873 874
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static int profile_graph_entry(struct ftrace_graph_ent *trace)
{
875 876
	int index = trace->depth;

877
	function_profile_call(trace->func, 0, NULL, NULL);
878 879 880 881

	if (index >= 0 && index < FTRACE_RETFUNC_DEPTH)
		current->ret_stack[index].subtime = 0;

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
	calltime = trace->rettime - trace->calltime;

903
	if (!fgraph_graph_time) {
904 905 906 907 908 909 910 911 912 913 914 915 916 917
		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
	int cpu;

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

1040
		name = kasprintf(GFP_KERNEL, "function%d", cpu);
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060
		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;
		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
1061 1062
	}

1063
	entry = tracefs_create_file("function_profile_enabled", 0644,
S
Steven Rostedt 已提交
1064 1065
				    d_tracer, NULL, &ftrace_profile_fops);
	if (!entry)
1066
		pr_warn("Could not create tracefs 'function_profile_enabled' entry\n");
S
Steven Rostedt 已提交
1067 1068 1069
}

#else /* CONFIG_FUNCTION_PROFILER */
1070
static __init void ftrace_profile_tracefs(struct dentry *d_tracer)
S
Steven Rostedt 已提交
1071 1072 1073 1074
{
}
#endif /* CONFIG_FUNCTION_PROFILER */

1075 1076
static struct pid * const ftrace_swapper_pid = &init_struct_pid;

1077 1078 1079 1080 1081 1082
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
static int ftrace_graph_active;
#else
# define ftrace_graph_active 0
#endif

1083 1084
#ifdef CONFIG_DYNAMIC_FTRACE

1085 1086
static struct ftrace_ops *removed_ops;

1087 1088 1089 1090 1091 1092
/*
 * 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;

1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
#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;
1105
	struct list_head	free_list;
1106 1107
};

1108 1109 1110 1111 1112
struct ftrace_func_entry {
	struct hlist_node hlist;
	unsigned long ip;
};

1113 1114 1115 1116 1117 1118 1119 1120 1121
/*
 * 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,
1122
};
1123
#define EMPTY_HASH	((struct ftrace_hash *)&empty_hash)
1124

1125
static struct ftrace_ops global_ops = {
1126 1127 1128 1129 1130
	.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 |
1131 1132
					  FTRACE_OPS_FL_INITIALIZED |
					  FTRACE_OPS_FL_PID,
1133 1134
};

1135 1136
/*
 * This is used by __kernel_text_address() to return true if the
1137
 * address is on a dynamically allocated trampoline that would
1138 1139 1140 1141 1142 1143 1144 1145 1146 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
 * 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;
}

1172 1173
struct ftrace_page {
	struct ftrace_page	*next;
1174
	struct dyn_ftrace	*records;
1175
	int			index;
1176
	int			size;
1177 1178
};

1179 1180
#define ENTRY_SIZE sizeof(struct dyn_ftrace)
#define ENTRIES_PER_PAGE (PAGE_SIZE / ENTRY_SIZE)
1181 1182 1183 1184 1185 1186 1187

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

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

1188
struct ftrace_func_entry *
1189 1190 1191 1192 1193 1194
ftrace_lookup_ip(struct ftrace_hash *hash, unsigned long ip)
{
	unsigned long key;
	struct ftrace_func_entry *entry;
	struct hlist_head *hhd;

1195
	if (ftrace_hash_empty(hash))
1196 1197 1198 1199 1200 1201 1202 1203 1204
		return NULL;

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

	hhd = &hash->buckets[key];

1205
	hlist_for_each_entry_rcu_notrace(entry, hhd, hlist) {
1206 1207 1208 1209 1210 1211
		if (entry->ip == ip)
			return entry;
	}
	return NULL;
}

1212 1213
static void __add_hash_entry(struct ftrace_hash *hash,
			     struct ftrace_func_entry *entry)
1214 1215 1216 1217 1218
{
	struct hlist_head *hhd;
	unsigned long key;

	if (hash->size_bits)
1219
		key = hash_long(entry->ip, hash->size_bits);
1220 1221 1222 1223 1224 1225
	else
		key = 0;

	hhd = &hash->buckets[key];
	hlist_add_head(&entry->hlist, hhd);
	hash->count++;
1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237
}

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);
1238 1239 1240 1241 1242

	return 0;
}

static void
1243
free_hash_entry(struct ftrace_hash *hash,
1244 1245 1246 1247 1248 1249 1250
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	kfree(entry);
	hash->count--;
}

1251 1252 1253 1254 1255 1256 1257 1258
static void
remove_hash_entry(struct ftrace_hash *hash,
		  struct ftrace_func_entry *entry)
{
	hlist_del(&entry->hlist);
	hash->count--;
}

1259 1260 1261
static void ftrace_hash_clear(struct ftrace_hash *hash)
{
	struct hlist_head *hhd;
1262
	struct hlist_node *tn;
1263 1264 1265 1266
	struct ftrace_func_entry *entry;
	int size = 1 << hash->size_bits;
	int i;

1267 1268 1269
	if (!hash->count)
		return;

1270 1271
	for (i = 0; i < size; i++) {
		hhd = &hash->buckets[i];
1272
		hlist_for_each_entry_safe(entry, tn, hhd, hlist)
1273
			free_hash_entry(hash, entry);
1274 1275 1276 1277
	}
	FTRACE_WARN_ON(hash->count);
}

1278 1279 1280 1281 1282 1283 1284 1285 1286
static void free_ftrace_hash(struct ftrace_hash *hash)
{
	if (!hash || hash == EMPTY_HASH)
		return;
	ftrace_hash_clear(hash);
	kfree(hash->buckets);
	kfree(hash);
}

1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
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);
}

1302 1303
void ftrace_free_filter(struct ftrace_ops *ops)
{
1304
	ftrace_ops_init(ops);
1305 1306
	free_ftrace_hash(ops->func_hash->filter_hash);
	free_ftrace_hash(ops->func_hash->notrace_hash);
1307 1308
}

1309 1310 1311 1312 1313 1314 1315 1316 1317 1318
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;
1319
	hash->buckets = kcalloc(size, sizeof(*hash->buckets), GFP_KERNEL);
1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344

	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? */
1345
	if (ftrace_hash_empty(hash))
1346 1347 1348 1349
		return new_hash;

	size = 1 << hash->size_bits;
	for (i = 0; i < size; i++) {
1350
		hlist_for_each_entry(entry, &hash->buckets[i], hlist) {
1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365
			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;
}

1366
static void
1367
ftrace_hash_rec_disable_modify(struct ftrace_ops *ops, int filter_hash);
1368
static void
1369
ftrace_hash_rec_enable_modify(struct ftrace_ops *ops, int filter_hash);
1370

1371 1372 1373
static int ftrace_hash_ipmodify_update(struct ftrace_ops *ops,
				       struct ftrace_hash *new_hash);

1374 1375
static struct ftrace_hash *
__ftrace_hash_move(struct ftrace_hash *src)
1376 1377
{
	struct ftrace_func_entry *entry;
1378
	struct hlist_node *tn;
1379
	struct hlist_head *hhd;
1380
	struct ftrace_hash *new_hash;
1381 1382 1383 1384 1385
	int size = src->count;
	int bits = 0;
	int i;

	/*
1386
	 * If the new source is empty, just return the empty_hash.
1387
	 */
1388 1389
	if (!src->count)
		return EMPTY_HASH;
1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400

	/*
	 * 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;

1401 1402
	new_hash = alloc_ftrace_hash(bits);
	if (!new_hash)
1403
		return NULL;
1404 1405 1406 1407

	size = 1 << src->size_bits;
	for (i = 0; i < size; i++) {
		hhd = &src->buckets[i];
1408
		hlist_for_each_entry_safe(entry, tn, hhd, hlist) {
1409
			remove_hash_entry(src, entry);
1410
			__add_hash_entry(new_hash, entry);
1411 1412 1413
		}
	}

1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431
	return new_hash;
}

static int
ftrace_hash_move(struct ftrace_ops *ops, int enable,
		 struct ftrace_hash **dst, struct ftrace_hash *src)
{
	struct ftrace_hash *new_hash;
	int ret;

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

	new_hash = __ftrace_hash_move(src);
	if (!new_hash)
		return -ENOMEM;

1432 1433 1434 1435 1436 1437 1438 1439 1440 1441
	/* 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;
		}
	}

1442 1443 1444 1445
	/*
	 * Remove the current set, update the hash and add
	 * them back.
	 */
1446
	ftrace_hash_rec_disable_modify(ops, enable);
1447

1448 1449
	rcu_assign_pointer(*dst, new_hash);

1450
	ftrace_hash_rec_enable_modify(ops, enable);
1451

1452
	return 0;
1453 1454
}

1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469
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));
}

1470 1471 1472 1473 1474 1475 1476 1477
/*
 * 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.
1478 1479 1480
 *
 * This needs to be called with preemption disabled as
 * the hashes are freed with call_rcu_sched().
1481 1482
 */
static int
1483
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
1484
{
1485
	struct ftrace_ops_hash hash;
1486 1487
	int ret;

1488 1489 1490 1491 1492 1493 1494 1495 1496 1497
#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

1498 1499
	hash.filter_hash = rcu_dereference_raw_notrace(ops->func_hash->filter_hash);
	hash.notrace_hash = rcu_dereference_raw_notrace(ops->func_hash->notrace_hash);
1500

1501
	if (hash_contains_ip(ip, &hash))
1502 1503 1504 1505 1506 1507 1508
		ret = 1;
	else
		ret = 0;

	return ret;
}

1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522
/*
 * 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()		\
		}				\
	}

1523 1524 1525

static int ftrace_cmp_recs(const void *a, const void *b)
{
1526 1527
	const struct dyn_ftrace *key = a;
	const struct dyn_ftrace *rec = b;
1528

1529
	if (key->flags < rec->ip)
1530
		return -1;
1531 1532
	if (key->ip >= rec->ip + MCOUNT_INSN_SIZE)
		return 1;
1533 1534 1535
	return 0;
}

1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548
/**
 * ftrace_location_range - return the first address of a traced location
 *	if it touches the given ip range
 * @start: start of range to search.
 * @end: end of range to search (inclusive). @end points to the last byte
 *	to check.
 *
 * Returns rec->ip if the related ftrace location is a least partly within
 * the given address range. That is, the first address of 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.
 */
unsigned long ftrace_location_range(unsigned long start, unsigned long end)
1549 1550 1551
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1552
	struct dyn_ftrace key;
1553

1554 1555
	key.ip = start;
	key.flags = end;	/* overload flags, as it is unsigned long */
1556 1557

	for (pg = ftrace_pages_start; pg; pg = pg->next) {
1558 1559
		if (end < pg->records[0].ip ||
		    start >= (pg->records[pg->index - 1].ip + MCOUNT_INSN_SIZE))
1560
			continue;
1561 1562 1563 1564
		rec = bsearch(&key, pg->records, pg->index,
			      sizeof(struct dyn_ftrace),
			      ftrace_cmp_recs);
		if (rec)
1565
			return rec->ip;
1566
	}
1567 1568 1569 1570

	return 0;
}

1571 1572 1573 1574
/**
 * ftrace_location - return true if the ip giving is a traced location
 * @ip: the instruction pointer to check
 *
1575
 * Returns rec->ip if @ip given is a pointer to a ftrace location.
1576 1577 1578 1579
 * 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.
 */
1580
unsigned long ftrace_location(unsigned long ip)
1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594
{
	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.
 */
1595
int ftrace_text_reserved(const void *start, const void *end)
1596
{
1597 1598 1599 1600 1601 1602
	unsigned long ret;

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

	return (int)!!ret;
1603 1604
}

1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624
/* 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;
}

1625
static bool __ftrace_hash_rec_update(struct ftrace_ops *ops,
1626 1627 1628 1629 1630 1631 1632
				     int filter_hash,
				     bool inc)
{
	struct ftrace_hash *hash;
	struct ftrace_hash *other_hash;
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
1633
	bool update = false;
1634 1635 1636 1637 1638
	int count = 0;
	int all = 0;

	/* Only update if the ops has been registered */
	if (!(ops->flags & FTRACE_OPS_FL_ENABLED))
1639
		return false;
1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652

	/*
	 * 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) {
1653 1654
		hash = ops->func_hash->filter_hash;
		other_hash = ops->func_hash->notrace_hash;
1655
		if (ftrace_hash_empty(hash))
1656 1657 1658
			all = 1;
	} else {
		inc = !inc;
1659 1660
		hash = ops->func_hash->notrace_hash;
		other_hash = ops->func_hash->filter_hash;
1661 1662 1663 1664
		/*
		 * If the notrace hash has no items,
		 * then there's nothing to do.
		 */
1665
		if (ftrace_hash_empty(hash))
1666
			return false;
1667 1668 1669 1670 1671 1672 1673
	}

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

1674 1675 1676
		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

1677 1678 1679 1680 1681
		if (all) {
			/*
			 * Only the filter_hash affects all records.
			 * Update if the record is not in the notrace hash.
			 */
1682
			if (!other_hash || !ftrace_lookup_ip(other_hash, rec->ip))
1683 1684
				match = 1;
		} else {
1685 1686
			in_hash = !!ftrace_lookup_ip(hash, rec->ip);
			in_other_hash = !!ftrace_lookup_ip(other_hash, rec->ip);
1687 1688

			/*
1689 1690
			 * If filter_hash is set, we want to match all functions
			 * that are in the hash but not in the other hash.
1691
			 *
1692 1693 1694 1695 1696
			 * 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.
1697 1698 1699 1700
			 */
			if (filter_hash && in_hash && !in_other_hash)
				match = 1;
			else if (!filter_hash && in_hash &&
1701
				 (in_other_hash || ftrace_hash_empty(other_hash)))
1702 1703 1704 1705 1706 1707 1708
				match = 1;
		}
		if (!match)
			continue;

		if (inc) {
			rec->flags++;
1709
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == FTRACE_REF_MAX))
1710
				return false;
1711 1712 1713 1714 1715 1716

			/*
			 * 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.
			 */
1717
			if (ftrace_rec_count(rec) == 1 && ops->trampoline)
1718
				rec->flags |= FTRACE_FL_TRAMP;
1719
			else
1720 1721 1722
				/*
				 * If we are adding another function callback
				 * to this function, and the previous had a
1723 1724
				 * custom trampoline in use, then we need to go
				 * back to the default trampoline.
1725
				 */
1726
				rec->flags &= ~FTRACE_FL_TRAMP;
1727

1728 1729 1730 1731 1732 1733
			/*
			 * 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;
1734
		} else {
1735
			if (FTRACE_WARN_ON(ftrace_rec_count(rec) == 0))
1736
				return false;
1737
			rec->flags--;
1738

1739 1740 1741 1742 1743 1744
			/*
			 * 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.
			 */
1745
			if (ftrace_rec_count(rec) > 0 &&
1746 1747 1748 1749 1750
			    rec->flags & FTRACE_FL_REGS &&
			    ops->flags & FTRACE_OPS_FL_SAVE_REGS) {
				if (!test_rec_ops_needs_regs(rec))
					rec->flags &= ~FTRACE_FL_REGS;
			}
1751

1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762
			/*
			 * 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;

1763 1764 1765 1766
			/*
			 * flags will be cleared in ftrace_check_record()
			 * if rec count is zero.
			 */
1767 1768
		}
		count++;
1769 1770 1771 1772

		/* Must match FTRACE_UPDATE_CALLS in ftrace_modify_all_code() */
		update |= ftrace_test_record(rec, 1) != FTRACE_UPDATE_IGNORE;

1773 1774
		/* Shortcut, if we handled all records, we are done. */
		if (!all && count == hash->count)
1775
			return update;
1776
	} while_for_each_ftrace_rec();
1777 1778

	return update;
1779 1780
}

1781
static bool ftrace_hash_rec_disable(struct ftrace_ops *ops,
1782 1783
				    int filter_hash)
{
1784
	return __ftrace_hash_rec_update(ops, filter_hash, 0);
1785 1786
}

1787
static bool ftrace_hash_rec_enable(struct ftrace_ops *ops,
1788 1789
				   int filter_hash)
{
1790
	return __ftrace_hash_rec_update(ops, filter_hash, 1);
1791 1792
}

1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827
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);
}

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
/*
 * 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) {
1861 1862 1863 1864

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886
		/* 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) {
1887 1888 1889 1890

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

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 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943
		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);
}

1944
static void print_ip_ins(const char *fmt, const unsigned char *p)
1945 1946 1947 1948 1949 1950 1951 1952 1953
{
	int i;

	printk(KERN_CONT "%s", fmt);

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

1954 1955
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec);
1956 1957
static struct ftrace_ops *
ftrace_find_tramp_ops_next(struct dyn_ftrace *rec, struct ftrace_ops *ops);
1958

1959
enum ftrace_bug_type ftrace_bug_type;
1960
const void *ftrace_expected;
1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981

static void print_bug_type(void)
{
	switch (ftrace_bug_type) {
	case FTRACE_BUG_UNKNOWN:
		break;
	case FTRACE_BUG_INIT:
		pr_info("Initializing ftrace call sites\n");
		break;
	case FTRACE_BUG_NOP:
		pr_info("Setting ftrace call site to NOP\n");
		break;
	case FTRACE_BUG_CALL:
		pr_info("Setting ftrace call site to call ftrace function\n");
		break;
	case FTRACE_BUG_UPDATE:
		pr_info("Updating ftrace call site to call a different ftrace function\n");
		break;
	}
}

1982 1983 1984
/**
 * ftrace_bug - report and shutdown function tracer
 * @failed: The failed type (EFAULT, EINVAL, EPERM)
1985
 * @rec: The record that failed
1986 1987 1988 1989 1990 1991 1992 1993
 *
 * 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
 */
1994
void ftrace_bug(int failed, struct dyn_ftrace *rec)
1995
{
1996 1997
	unsigned long ip = rec ? rec->ip : 0;

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
	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);
2008
		print_ip_ins(" actual:   ", (unsigned char *)ip);
2009
		pr_cont("\n");
2010 2011 2012 2013
		if (ftrace_expected) {
			print_ip_ins(" expected: ", ftrace_expected);
			pr_cont("\n");
		}
2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024
		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);
	}
2025
	print_bug_type();
2026 2027 2028 2029 2030 2031 2032 2033
	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);
2034 2035 2036 2037 2038 2039 2040 2041
			if (ops) {
				do {
					pr_cont("\ttramp: %pS (%pS)",
						(void *)ops->trampoline,
						(void *)ops->func);
					ops = ftrace_find_tramp_ops_next(rec, ops);
				} while (ops);
			} else
2042 2043 2044 2045
				pr_cont("\ttramp: ERROR!");

		}
		ip = ftrace_get_addr_curr(rec);
2046
		pr_cont("\n expected tramp: %lx\n", ip);
2047
	}
2048 2049
}

2050
static int ftrace_check_record(struct dyn_ftrace *rec, int enable, int update)
2051
{
2052
	unsigned long flag = 0UL;
2053

2054 2055
	ftrace_bug_type = FTRACE_BUG_UNKNOWN;

2056 2057 2058
	if (rec->flags & FTRACE_FL_DISABLED)
		return FTRACE_UPDATE_IGNORE;

S
Steven Rostedt 已提交
2059
	/*
2060
	 * If we are updating calls:
S
Steven Rostedt 已提交
2061
	 *
2062 2063
	 *   If the record has a ref count, then we need to enable it
	 *   because someone is using it.
S
Steven Rostedt 已提交
2064
	 *
2065 2066
	 *   Otherwise we make sure its disabled.
	 *
2067
	 * If we are disabling calls, then disable all records that
2068
	 * are enabled.
S
Steven Rostedt 已提交
2069
	 */
2070
	if (enable && ftrace_rec_count(rec))
2071
		flag = FTRACE_FL_ENABLED;
S
Steven Rostedt 已提交
2072

2073
	/*
2074 2075 2076
	 * 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.
2077
	 */
2078 2079 2080 2081 2082 2083 2084 2085 2086
	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;
	}
2087

2088 2089
	/* If the state of this record hasn't changed, then do nothing */
	if ((rec->flags & FTRACE_FL_ENABLED) == flag)
2090
		return FTRACE_UPDATE_IGNORE;
S
Steven Rostedt 已提交
2091

2092
	if (flag) {
2093 2094 2095 2096
		/* Save off if rec is being enabled (for return value) */
		flag ^= rec->flags & FTRACE_FL_ENABLED;

		if (update) {
2097
			rec->flags |= FTRACE_FL_ENABLED;
2098 2099 2100 2101 2102 2103
			if (flag & FTRACE_FL_REGS) {
				if (rec->flags & FTRACE_FL_REGS)
					rec->flags |= FTRACE_FL_REGS_EN;
				else
					rec->flags &= ~FTRACE_FL_REGS_EN;
			}
2104 2105 2106 2107 2108 2109
			if (flag & FTRACE_FL_TRAMP) {
				if (rec->flags & FTRACE_FL_TRAMP)
					rec->flags |= FTRACE_FL_TRAMP_EN;
				else
					rec->flags &= ~FTRACE_FL_TRAMP_EN;
			}
2110 2111 2112 2113 2114 2115 2116
		}

		/*
		 * 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
2117
		 *   from the save regs, to a non-save regs function or
2118
		 *   vice versa, or from a trampoline call.
2119
		 */
2120 2121
		if (flag & FTRACE_FL_ENABLED) {
			ftrace_bug_type = FTRACE_BUG_CALL;
2122
			return FTRACE_UPDATE_MAKE_CALL;
2123
		}
2124

2125
		ftrace_bug_type = FTRACE_BUG_UPDATE;
2126
		return FTRACE_UPDATE_MODIFY_CALL;
2127 2128
	}

2129 2130
	if (update) {
		/* If there's no more users, clear all flags */
2131
		if (!ftrace_rec_count(rec))
2132 2133
			rec->flags = 0;
		else
2134 2135 2136 2137 2138 2139
			/*
			 * 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);
2140
	}
2141

2142
	ftrace_bug_type = FTRACE_BUG_NOP;
2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172
	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);
}

2173 2174 2175 2176
static struct ftrace_ops *
ftrace_find_tramp_ops_any(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;
2177
	unsigned long ip = rec->ip;
2178 2179 2180 2181 2182 2183

	do_for_each_ftrace_op(op, ftrace_ops_list) {

		if (!op->trampoline)
			continue;

2184
		if (hash_contains_ip(ip, op->func_hash))
2185 2186 2187 2188 2189 2190
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208
static struct ftrace_ops *
ftrace_find_tramp_ops_next(struct dyn_ftrace *rec,
			   struct ftrace_ops *op)
{
	unsigned long ip = rec->ip;

	while_for_each_ftrace_op(op) {

		if (!op->trampoline)
			continue;

		if (hash_contains_ip(ip, op->func_hash))
			return op;
	} 

	return NULL;
}

2209 2210 2211 2212
static struct ftrace_ops *
ftrace_find_tramp_ops_curr(struct dyn_ftrace *rec)
{
	struct ftrace_ops *op;
2213
	unsigned long ip = rec->ip;
2214

2215 2216 2217 2218 2219 2220 2221 2222
	/*
	 * 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))
2223 2224 2225
			return removed_ops;
	}

2226 2227 2228 2229 2230 2231 2232
	/*
	 * 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
2233 2234 2235 2236 2237 2238 2239 2240 2241 2242
	 * 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.
2243
	 */
2244
	do_for_each_ftrace_op(op, ftrace_ops_list) {
2245 2246 2247 2248 2249 2250 2251 2252 2253

		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)
2254 2255
			continue;

2256

2257
		/*
2258 2259 2260
		 * If the ops is being modified and is in the old
		 * hash, then it is probably being removed from this
		 * function.
2261 2262 2263
		 */
		if ((op->flags & FTRACE_OPS_FL_MODIFYING) &&
		    hash_contains_ip(ip, &op->old_hash))
2264
			return op;
2265 2266 2267 2268 2269 2270 2271 2272
		/*
		 * 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;
2273 2274 2275 2276 2277 2278 2279 2280 2281 2282

	} 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;
2283
	unsigned long ip = rec->ip;
2284 2285 2286

	do_for_each_ftrace_op(op, ftrace_ops_list) {
		/* pass rec in as regs to have non-NULL val */
2287
		if (hash_contains_ip(ip, op->func_hash))
2288 2289 2290 2291 2292 2293
			return op;
	} while_for_each_ftrace_op(op);

	return NULL;
}

2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305
/**
 * 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)
{
2306 2307 2308 2309 2310 2311
	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)) {
2312 2313
			pr_warn("Bad trampoline accounting at: %p (%pS) (%lx)\n",
				(void *)rec->ip, (void *)rec->ip, rec->flags);
2314 2315 2316 2317 2318 2319
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337
	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)
{
2338 2339 2340 2341 2342 2343
	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)) {
2344 2345
			pr_warn("Bad trampoline accounting at: %p (%pS)\n",
				(void *)rec->ip, (void *)rec->ip);
2346 2347 2348 2349 2350 2351
			/* Ftrace is shutting down, return anything */
			return (unsigned long)FTRACE_ADDR;
		}
		return ops->trampoline;
	}

2352 2353 2354 2355 2356 2357
	if (rec->flags & FTRACE_FL_REGS_EN)
		return (unsigned long)FTRACE_REGS_ADDR;
	else
		return (unsigned long)FTRACE_ADDR;
}

2358 2359 2360
static int
__ftrace_replace_code(struct dyn_ftrace *rec, int enable)
{
2361
	unsigned long ftrace_old_addr;
2362 2363 2364
	unsigned long ftrace_addr;
	int ret;

2365
	ftrace_addr = ftrace_get_addr_new(rec);
2366

2367 2368 2369 2370
	/* 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);
2371

2372 2373
	ftrace_bug_type = FTRACE_BUG_UNKNOWN;

2374 2375 2376 2377 2378
	switch (ret) {
	case FTRACE_UPDATE_IGNORE:
		return 0;

	case FTRACE_UPDATE_MAKE_CALL:
2379
		ftrace_bug_type = FTRACE_BUG_CALL;
2380
		return ftrace_make_call(rec, ftrace_addr);
2381 2382

	case FTRACE_UPDATE_MAKE_NOP:
2383
		ftrace_bug_type = FTRACE_BUG_NOP;
2384
		return ftrace_make_nop(NULL, rec, ftrace_old_addr);
2385 2386

	case FTRACE_UPDATE_MODIFY_CALL:
2387
		ftrace_bug_type = FTRACE_BUG_UPDATE;
2388
		return ftrace_modify_call(rec, ftrace_old_addr, ftrace_addr);
2389 2390
	}

2391
	return -1; /* unknow ftrace bug */
2392 2393
}

2394
void __weak ftrace_replace_code(int enable)
2395 2396 2397
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
S
Steven Rostedt 已提交
2398
	int failed;
2399

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

2403
	do_for_each_ftrace_rec(pg, rec) {
2404 2405 2406 2407

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

2408
		failed = __ftrace_replace_code(rec, enable);
2409
		if (failed) {
2410
			ftrace_bug(failed, rec);
2411 2412
			/* Stop processing */
			return;
2413
		}
2414
	} while_for_each_ftrace_rec();
2415 2416
}

2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488
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];
}

2489
static int
2490
ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec)
2491
{
2492
	int ret;
2493

2494 2495 2496
	if (unlikely(ftrace_disabled))
		return 0;

2497
	ret = ftrace_make_nop(mod, rec, MCOUNT_ADDR);
2498
	if (ret) {
2499
		ftrace_bug_type = FTRACE_BUG_INIT;
2500
		ftrace_bug(ret, rec);
2501
		return 0;
2502
	}
2503
	return 1;
2504 2505
}

2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523
/*
 * 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;
}

2524
void ftrace_modify_all_code(int command)
2525
{
2526
	int update = command & FTRACE_UPDATE_TRACE_FUNC;
2527
	int err = 0;
2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538

	/*
	 * 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.
	 */
2539 2540 2541 2542 2543
	if (update) {
		err = ftrace_update_ftrace_func(ftrace_ops_list_func);
		if (FTRACE_WARN_ON(err))
			return;
	}
2544

2545
	if (command & FTRACE_UPDATE_CALLS)
2546
		ftrace_replace_code(1);
2547
	else if (command & FTRACE_DISABLE_CALLS)
2548 2549
		ftrace_replace_code(0);

2550 2551 2552 2553 2554 2555
	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);
2556 2557 2558
		err = ftrace_update_ftrace_func(ftrace_trace_function);
		if (FTRACE_WARN_ON(err))
			return;
2559
	}
2560

2561
	if (command & FTRACE_START_FUNC_RET)
2562
		err = ftrace_enable_ftrace_graph_caller();
2563
	else if (command & FTRACE_STOP_FUNC_RET)
2564 2565
		err = ftrace_disable_ftrace_graph_caller();
	FTRACE_WARN_ON(err);
2566 2567 2568 2569 2570 2571 2572
}

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

	ftrace_modify_all_code(*command);
2573

2574
	return 0;
2575 2576
}

2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600
/**
 * 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
2601
static void ftrace_run_update_code(int command)
2602
{
2603 2604 2605 2606 2607 2608 2609
	int ret;

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

2610 2611 2612 2613 2614 2615 2616 2617
	/*
	 * 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);

2618 2619
	ret = ftrace_arch_code_modify_post_process();
	FTRACE_WARN_ON(ret);
2620 2621
}

2622
static void ftrace_run_modify_code(struct ftrace_ops *ops, int command,
2623
				   struct ftrace_ops_hash *old_hash)
2624 2625
{
	ops->flags |= FTRACE_OPS_FL_MODIFYING;
2626 2627
	ops->old_hash.filter_hash = old_hash->filter_hash;
	ops->old_hash.notrace_hash = old_hash->notrace_hash;
2628
	ftrace_run_update_code(command);
2629
	ops->old_hash.filter_hash = NULL;
2630
	ops->old_hash.notrace_hash = NULL;
2631 2632 2633
	ops->flags &= ~FTRACE_OPS_FL_MODIFYING;
}

2634
static ftrace_func_t saved_ftrace_func;
2635
static int ftrace_start_up;
2636

2637 2638 2639 2640
void __weak arch_ftrace_trampoline_free(struct ftrace_ops *ops)
{
}

2641
static void per_cpu_ops_free(struct ftrace_ops *ops)
2642 2643 2644 2645
{
	free_percpu(ops->disabled);
}

2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657
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);
}
2658

2659 2660 2661 2662 2663 2664 2665
static void ftrace_startup_all(int command)
{
	update_all_ops = true;
	ftrace_startup_enable(command);
	update_all_ops = false;
}

2666
static int ftrace_startup(struct ftrace_ops *ops, int command)
2667
{
2668
	int ret;
2669

2670
	if (unlikely(ftrace_disabled))
2671
		return -ENODEV;
2672

2673 2674 2675 2676
	ret = __register_ftrace_function(ops);
	if (ret)
		return ret;

2677
	ftrace_start_up++;
2678

2679 2680 2681 2682 2683 2684 2685 2686 2687
	/*
	 * 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;
2688

2689 2690 2691 2692 2693 2694 2695 2696 2697
	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;
	}

2698 2699
	if (ftrace_hash_rec_enable(ops, 1))
		command |= FTRACE_UPDATE_CALLS;
2700

2701
	ftrace_startup_enable(command);
2702

2703 2704
	ops->flags &= ~FTRACE_OPS_FL_ADDING;

2705
	return 0;
2706 2707
}

2708
static int ftrace_shutdown(struct ftrace_ops *ops, int command)
2709
{
2710
	int ret;
2711

2712
	if (unlikely(ftrace_disabled))
2713 2714 2715 2716 2717
		return -ENODEV;

	ret = __unregister_ftrace_function(ops);
	if (ret)
		return ret;
2718

2719
	ftrace_start_up--;
2720 2721 2722 2723 2724 2725 2726
	/*
	 * 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);

2727 2728
	/* Disabling ipmodify never fails */
	ftrace_hash_ipmodify_disable(ops);
2729

2730 2731
	if (ftrace_hash_rec_disable(ops, 1))
		command |= FTRACE_UPDATE_CALLS;
2732

2733
	ops->flags &= ~FTRACE_OPS_FL_ENABLED;
2734

2735 2736 2737 2738
	if (saved_ftrace_func != ftrace_trace_function) {
		saved_ftrace_func = ftrace_trace_function;
		command |= FTRACE_UPDATE_TRACE_FUNC;
	}
2739

2740 2741
	if (!command || !ftrace_enabled) {
		/*
2742
		 * If these are per_cpu ops, they still need their
2743 2744 2745 2746
		 * per_cpu field freed. Since, function tracing is
		 * not currently active, we can just free them
		 * without synchronizing all CPUs.
		 */
2747 2748
		if (ops->flags & FTRACE_OPS_FL_PER_CPU)
			per_cpu_ops_free(ops);
2749
		return 0;
2750
	}
2751

2752 2753 2754 2755
	/*
	 * If the ops uses a trampoline, then it needs to be
	 * tested first on update.
	 */
2756
	ops->flags |= FTRACE_OPS_FL_REMOVING;
2757 2758
	removed_ops = ops;

2759 2760 2761 2762
	/* 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;

2763
	ftrace_run_update_code(command);
2764

2765 2766 2767 2768 2769 2770 2771 2772 2773
	/*
	 * 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) {
2774
			if (FTRACE_WARN_ON_ONCE(rec->flags & ~FTRACE_FL_DISABLED))
2775 2776 2777 2778 2779
				pr_warn("  %pS flags:%lx\n",
					(void *)rec->ip, rec->flags);
		} while_for_each_ftrace_rec();
	}

2780 2781 2782 2783
	ops->old_hash.filter_hash = NULL;
	ops->old_hash.notrace_hash = NULL;

	removed_ops = NULL;
2784
	ops->flags &= ~FTRACE_OPS_FL_REMOVING;
2785

2786 2787 2788
	/*
	 * Dynamic ops may be freed, we must make sure that all
	 * callers are done before leaving this function.
2789
	 * The same goes for freeing the per_cpu data of the per_cpu
2790 2791 2792 2793 2794 2795 2796 2797 2798 2799
	 * 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.
	 */
2800
	if (ops->flags & (FTRACE_OPS_FL_DYNAMIC | FTRACE_OPS_FL_PER_CPU)) {
2801 2802
		schedule_on_each_cpu(ftrace_sync);

2803 2804
		arch_ftrace_trampoline_free(ops);

2805 2806
		if (ops->flags & FTRACE_OPS_FL_PER_CPU)
			per_cpu_ops_free(ops);
2807 2808
	}

2809
	return 0;
2810 2811
}

I
Ingo Molnar 已提交
2812
static void ftrace_startup_sysctl(void)
2813
{
2814 2815
	int command;

2816 2817 2818
	if (unlikely(ftrace_disabled))
		return;

2819 2820
	/* Force update next time */
	saved_ftrace_func = NULL;
2821
	/* ftrace_start_up is true if we want ftrace running */
2822 2823 2824 2825
	if (ftrace_start_up) {
		command = FTRACE_UPDATE_CALLS;
		if (ftrace_graph_active)
			command |= FTRACE_START_FUNC_RET;
2826
		ftrace_startup_enable(command);
2827
	}
2828 2829
}

I
Ingo Molnar 已提交
2830
static void ftrace_shutdown_sysctl(void)
2831
{
2832 2833
	int command;

2834 2835 2836
	if (unlikely(ftrace_disabled))
		return;

2837
	/* ftrace_start_up is true if ftrace is running */
2838 2839 2840 2841 2842 2843
	if (ftrace_start_up) {
		command = FTRACE_DISABLE_CALLS;
		if (ftrace_graph_active)
			command |= FTRACE_STOP_FUNC_RET;
		ftrace_run_update_code(command);
	}
2844 2845
}

2846
static u64		ftrace_update_time;
2847 2848
unsigned long		ftrace_update_tot_cnt;

2849
static inline int ops_traces_mod(struct ftrace_ops *ops)
2850
{
2851 2852 2853 2854
	/*
	 * Filter_hash being empty will default to trace module.
	 * But notrace hash requires a test of individual module functions.
	 */
2855 2856
	return ftrace_hash_empty(ops->func_hash->filter_hash) &&
		ftrace_hash_empty(ops->func_hash->notrace_hash);
2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872
}

/*
 * 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;

2873
	/* If ops traces all then it includes this function */
2874
	if (ops_traces_mod(ops))
2875
		return 1;
2876 2877

	/* The function must be in the filter */
2878 2879
	if (!ftrace_hash_empty(ops->func_hash->filter_hash) &&
	    !ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip))
2880
		return 0;
2881

2882
	/* If in notrace hash, we ignore it too */
2883
	if (ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip))
2884 2885 2886 2887 2888
		return 0;

	return 1;
}

2889
static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
2890
{
2891
	struct ftrace_page *pg;
2892
	struct dyn_ftrace *p;
2893
	u64 start, stop;
2894
	unsigned long update_cnt = 0;
2895
	unsigned long rec_flags = 0;
2896
	int i;
2897

2898 2899
	start = ftrace_now(raw_smp_processor_id());

2900
	/*
2901 2902 2903 2904 2905 2906 2907 2908 2909
	 * When a module is loaded, this function is called to convert
	 * the calls to mcount in its text to nops, and also to create
	 * an entry in the ftrace data. Now, if ftrace is activated
	 * after this call, but before the module sets its text to
	 * read-only, the modification of enabling ftrace can fail if
	 * the read-only is done while ftrace is converting the calls.
	 * To prevent this, the module's records are set as disabled
	 * and will be enabled after the call to set the module's text
	 * to read-only.
2910
	 */
2911 2912
	if (mod)
		rec_flags |= FTRACE_FL_DISABLED;
2913

2914
	for (pg = new_pgs; pg; pg = pg->next) {
2915

2916
		for (i = 0; i < pg->index; i++) {
2917

2918 2919 2920
			/* If something went wrong, bail without enabling anything */
			if (unlikely(ftrace_disabled))
				return -1;
2921

2922
			p = &pg->records[i];
2923
			p->flags = rec_flags;
2924

2925 2926 2927 2928 2929 2930
			/*
			 * Do the initial record conversion from mcount jump
			 * to the NOP instructions.
			 */
			if (!ftrace_code_disable(mod, p))
				break;
2931

2932
			update_cnt++;
2933
		}
2934 2935
	}

I
Ingo Molnar 已提交
2936
	stop = ftrace_now(raw_smp_processor_id());
2937
	ftrace_update_time = stop - start;
2938
	ftrace_update_tot_cnt += update_cnt;
2939

2940 2941 2942
	return 0;
}

2943
static int ftrace_allocate_records(struct ftrace_page *pg, int count)
2944
{
2945
	int order;
2946 2947
	int cnt;

2948 2949 2950 2951
	if (WARN_ON(!count))
		return -EINVAL;

	order = get_count_order(DIV_ROUND_UP(count, ENTRIES_PER_PAGE));
2952 2953

	/*
2954 2955
	 * We want to fill as much as possible. No more than a page
	 * may be empty.
2956
	 */
2957 2958
	while ((PAGE_SIZE << order) / ENTRY_SIZE >= count + ENTRIES_PER_PAGE)
		order--;
2959

2960 2961
 again:
	pg->records = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
2962

2963 2964 2965 2966 2967 2968 2969
	if (!pg->records) {
		/* if we can't allocate this size, try something smaller */
		if (!order)
			return -ENOMEM;
		order >>= 1;
		goto again;
	}
2970

2971 2972
	cnt = (PAGE_SIZE << order) / ENTRY_SIZE;
	pg->size = cnt;
2973

2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006
	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)
3007 3008
			break;

3009 3010 3011 3012
		pg->next = kzalloc(sizeof(*pg), GFP_KERNEL);
		if (!pg->next)
			goto free_pages;

3013 3014 3015
		pg = pg->next;
	}

3016 3017 3018
	return start_pg;

 free_pages:
3019 3020
	pg = start_pg;
	while (pg) {
3021 3022 3023 3024 3025 3026 3027 3028 3029 3030
		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;
}

3031 3032 3033
#define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */

struct ftrace_iterator {
3034
	loff_t				pos;
3035 3036 3037 3038 3039
	loff_t				func_pos;
	struct ftrace_page		*pg;
	struct dyn_ftrace		*func;
	struct ftrace_func_probe	*probe;
	struct trace_parser		parser;
3040
	struct ftrace_hash		*hash;
3041
	struct ftrace_ops		*ops;
3042 3043 3044
	int				hidx;
	int				idx;
	unsigned			flags;
3045 3046
};

3047
static void *
3048
t_hash_next(struct seq_file *m, loff_t *pos)
3049 3050
{
	struct ftrace_iterator *iter = m->private;
3051
	struct hlist_node *hnd = NULL;
3052 3053 3054
	struct hlist_head *hhd;

	(*pos)++;
3055
	iter->pos = *pos;
3056

3057 3058
	if (iter->probe)
		hnd = &iter->probe->node;
3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080
 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;
		}
	}

3081 3082 3083 3084 3085 3086
	if (WARN_ON_ONCE(!hnd))
		return NULL;

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

	return iter;
3087 3088 3089 3090 3091 3092
}

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
3093 3094
	loff_t l;

3095 3096 3097
	if (!(iter->flags & FTRACE_ITER_DO_HASH))
		return NULL;

3098 3099
	if (iter->func_pos > *pos)
		return NULL;
3100

L
Li Zefan 已提交
3101
	iter->hidx = 0;
3102
	for (l = 0; l <= (*pos - iter->func_pos); ) {
3103
		p = t_hash_next(m, &l);
L
Li Zefan 已提交
3104 3105 3106
		if (!p)
			break;
	}
3107 3108 3109
	if (!p)
		return NULL;

3110 3111 3112
	/* Only set this if we have an item */
	iter->flags |= FTRACE_ITER_HASH;

3113
	return iter;
3114 3115
}

3116 3117
static int
t_hash_show(struct seq_file *m, struct ftrace_iterator *iter)
3118
{
S
Steven Rostedt 已提交
3119
	struct ftrace_func_probe *rec;
3120

3121 3122 3123
	rec = iter->probe;
	if (WARN_ON_ONCE(!rec))
		return -EIO;
3124

3125 3126 3127
	if (rec->ops->print)
		return rec->ops->print(m, rec->ip, rec->ops, rec->data);

3128
	seq_printf(m, "%ps:%ps", (void *)rec->ip, (void *)rec->ops->func);
3129 3130 3131 3132 3133 3134 3135 3136

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

	return 0;
}

I
Ingo Molnar 已提交
3137
static void *
3138 3139 3140
t_next(struct seq_file *m, void *v, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
3141
	struct ftrace_ops *ops = iter->ops;
3142 3143
	struct dyn_ftrace *rec = NULL;

3144 3145 3146
	if (unlikely(ftrace_disabled))
		return NULL;

3147
	if (iter->flags & FTRACE_ITER_HASH)
3148
		return t_hash_next(m, pos);
3149

3150
	(*pos)++;
3151
	iter->pos = iter->func_pos = *pos;
3152

3153
	if (iter->flags & FTRACE_ITER_PRINTALL)
3154
		return t_hash_start(m, pos);
3155

3156 3157 3158 3159 3160 3161 3162 3163 3164
 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++];
3165
		if (((iter->flags & FTRACE_ITER_FILTER) &&
3166
		     !(ftrace_lookup_ip(ops->func_hash->filter_hash, rec->ip))) ||
S
Steven Rostedt 已提交
3167

3168
		    ((iter->flags & FTRACE_ITER_NOTRACE) &&
3169
		     !ftrace_lookup_ip(ops->func_hash->notrace_hash, rec->ip)) ||
3170 3171

		    ((iter->flags & FTRACE_ITER_ENABLED) &&
3172
		     !(rec->flags & FTRACE_FL_ENABLED))) {
3173

3174 3175 3176 3177 3178
			rec = NULL;
			goto retry;
		}
	}

3179
	if (!rec)
3180
		return t_hash_start(m, pos);
3181 3182 3183 3184

	iter->func = rec;

	return iter;
3185 3186
}

3187 3188 3189 3190
static void reset_iter_read(struct ftrace_iterator *iter)
{
	iter->pos = 0;
	iter->func_pos = 0;
3191
	iter->flags &= ~(FTRACE_ITER_PRINTALL | FTRACE_ITER_HASH);
3192 3193 3194 3195 3196
}

static void *t_start(struct seq_file *m, loff_t *pos)
{
	struct ftrace_iterator *iter = m->private;
3197
	struct ftrace_ops *ops = iter->ops;
3198
	void *p = NULL;
3199
	loff_t l;
3200

3201
	mutex_lock(&ftrace_lock);
3202 3203 3204 3205

	if (unlikely(ftrace_disabled))
		return NULL;

3206 3207 3208 3209 3210 3211
	/*
	 * If an lseek was done, then reset and start from beginning.
	 */
	if (*pos < iter->pos)
		reset_iter_read(iter);

3212 3213 3214 3215 3216
	/*
	 * For set_ftrace_filter reading, if we have the filter
	 * off, we can short cut and just print out that all
	 * functions are enabled.
	 */
3217
	if ((iter->flags & FTRACE_ITER_FILTER &&
3218
	     ftrace_hash_empty(ops->func_hash->filter_hash)) ||
3219
	    (iter->flags & FTRACE_ITER_NOTRACE &&
3220
	     ftrace_hash_empty(ops->func_hash->notrace_hash))) {
3221
		if (*pos > 0)
3222
			return t_hash_start(m, pos);
3223
		iter->flags |= FTRACE_ITER_PRINTALL;
3224 3225
		/* reset in case of seek/pread */
		iter->flags &= ~FTRACE_ITER_HASH;
3226 3227 3228
		return iter;
	}

3229 3230 3231
	if (iter->flags & FTRACE_ITER_HASH)
		return t_hash_start(m, pos);

3232 3233 3234 3235 3236
	/*
	 * 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.
	 */
3237 3238 3239 3240 3241 3242
	iter->pg = ftrace_pages_start;
	iter->idx = 0;
	for (l = 0; l <= *pos; ) {
		p = t_next(m, p, &l);
		if (!p)
			break;
3243
	}
3244

3245 3246
	if (!p)
		return t_hash_start(m, pos);
3247 3248

	return iter;
3249 3250 3251 3252
}

static void t_stop(struct seq_file *m, void *p)
{
3253
	mutex_unlock(&ftrace_lock);
3254 3255
}

3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271
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);
}

3272 3273
static int t_show(struct seq_file *m, void *v)
{
3274
	struct ftrace_iterator *iter = m->private;
3275
	struct dyn_ftrace *rec;
3276

3277
	if (iter->flags & FTRACE_ITER_HASH)
3278
		return t_hash_show(m, iter);
3279

3280
	if (iter->flags & FTRACE_ITER_PRINTALL) {
3281
		if (iter->flags & FTRACE_ITER_NOTRACE)
3282
			seq_puts(m, "#### no functions disabled ####\n");
3283
		else
3284
			seq_puts(m, "#### all functions enabled ####\n");
3285 3286 3287
		return 0;
	}

3288 3289
	rec = iter->func;

3290 3291 3292
	if (!rec)
		return 0;

3293
	seq_printf(m, "%ps", (void *)rec->ip);
3294
	if (iter->flags & FTRACE_ITER_ENABLED) {
3295
		struct ftrace_ops *ops;
3296

3297
		seq_printf(m, " (%ld)%s%s",
3298
			   ftrace_rec_count(rec),
3299 3300
			   rec->flags & FTRACE_FL_REGS ? " R" : "  ",
			   rec->flags & FTRACE_FL_IPMODIFY ? " I" : "  ");
3301
		if (rec->flags & FTRACE_FL_TRAMP_EN) {
3302
			ops = ftrace_find_tramp_ops_any(rec);
3303 3304 3305 3306 3307
			if (ops) {
				do {
					seq_printf(m, "\ttramp: %pS (%pS)",
						   (void *)ops->trampoline,
						   (void *)ops->func);
3308
					add_trampoline_func(m, ops, rec);
3309 3310 3311
					ops = ftrace_find_tramp_ops_next(rec, ops);
				} while (ops);
			} else
3312
				seq_puts(m, "\ttramp: ERROR!");
3313 3314
		} else {
			add_trampoline_func(m, NULL, rec);
3315 3316 3317
		}
	}	

3318
	seq_putc(m, '\n');
3319 3320 3321 3322

	return 0;
}

J
James Morris 已提交
3323
static const struct seq_operations show_ftrace_seq_ops = {
3324 3325 3326 3327 3328 3329
	.start = t_start,
	.next = t_next,
	.stop = t_stop,
	.show = t_show,
};

I
Ingo Molnar 已提交
3330
static int
3331 3332 3333 3334
ftrace_avail_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3335 3336 3337
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3338 3339 3340 3341
	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
3342
	}
3343

3344
	return iter ? 0 : -ENOMEM;
3345 3346
}

3347 3348 3349 3350 3351
static int
ftrace_enabled_open(struct inode *inode, struct file *file)
{
	struct ftrace_iterator *iter;

3352 3353 3354 3355 3356
	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;
3357 3358
	}

3359
	return iter ? 0 : -ENOMEM;
3360 3361
}

3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374
/**
 * 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.
3375
 * tracing_lseek() should be used as the lseek routine, and
3376 3377 3378
 * release must call ftrace_regex_release().
 */
int
3379
ftrace_regex_open(struct ftrace_ops *ops, int flag,
3380
		  struct inode *inode, struct file *file)
3381 3382
{
	struct ftrace_iterator *iter;
3383
	struct ftrace_hash *hash;
3384 3385
	int ret = 0;

3386 3387
	ftrace_ops_init(ops);

3388 3389 3390
	if (unlikely(ftrace_disabled))
		return -ENODEV;

3391 3392 3393 3394
	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
	if (!iter)
		return -ENOMEM;

3395 3396 3397 3398 3399
	if (trace_parser_get_init(&iter->parser, FTRACE_BUFF_MAX)) {
		kfree(iter);
		return -ENOMEM;
	}

3400 3401 3402
	iter->ops = ops;
	iter->flags = flag;

3403
	mutex_lock(&ops->func_hash->regex_lock);
3404

3405
	if (flag & FTRACE_ITER_NOTRACE)
3406
		hash = ops->func_hash->notrace_hash;
3407
	else
3408
		hash = ops->func_hash->filter_hash;
3409

3410
	if (file->f_mode & FMODE_WRITE) {
3411 3412 3413 3414 3415 3416 3417
		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);

3418 3419 3420
		if (!iter->hash) {
			trace_parser_put(&iter->parser);
			kfree(iter);
3421 3422
			ret = -ENOMEM;
			goto out_unlock;
3423 3424
		}
	}
3425

3426 3427 3428 3429 3430 3431 3432
	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;
3433
		} else {
3434 3435
			/* Failed */
			free_ftrace_hash(iter->hash);
3436
			trace_parser_put(&iter->parser);
3437
			kfree(iter);
3438
		}
3439 3440
	} else
		file->private_data = iter;
3441 3442

 out_unlock:
3443
	mutex_unlock(&ops->func_hash->regex_lock);
3444 3445 3446 3447

	return ret;
}

3448 3449 3450
static int
ftrace_filter_open(struct inode *inode, struct file *file)
{
3451 3452 3453
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops,
3454 3455
			FTRACE_ITER_FILTER | FTRACE_ITER_DO_HASH,
			inode, file);
3456 3457 3458 3459 3460
}

static int
ftrace_notrace_open(struct inode *inode, struct file *file)
{
3461 3462 3463
	struct ftrace_ops *ops = inode->i_private;

	return ftrace_regex_open(ops, FTRACE_ITER_NOTRACE,
3464
				 inode, file);
3465 3466
}

3467 3468 3469 3470 3471 3472 3473
/* Type for quick search ftrace basic regexes (globs) from filter_parse_regex */
struct ftrace_glob {
	char *search;
	unsigned len;
	int type;
};

3474 3475 3476 3477 3478 3479 3480 3481 3482 3483
/*
 * If symbols in an architecture don't correspond exactly to the user-visible
 * name of what they represent, it is possible to define this function to
 * perform the necessary adjustments.
*/
char * __weak arch_ftrace_match_adjust(char *str, const char *search)
{
	return str;
}

3484
static int ftrace_match(char *str, struct ftrace_glob *g)
3485 3486
{
	int matched = 0;
3487
	int slen;
3488

3489 3490
	str = arch_ftrace_match_adjust(str, g->search);

3491
	switch (g->type) {
3492
	case MATCH_FULL:
3493
		if (strcmp(str, g->search) == 0)
3494 3495 3496
			matched = 1;
		break;
	case MATCH_FRONT_ONLY:
3497
		if (strncmp(str, g->search, g->len) == 0)
3498 3499 3500
			matched = 1;
		break;
	case MATCH_MIDDLE_ONLY:
3501
		if (strstr(str, g->search))
3502 3503 3504
			matched = 1;
		break;
	case MATCH_END_ONLY:
3505
		slen = strlen(str);
3506 3507
		if (slen >= g->len &&
		    memcmp(str + slen - g->len, g->search, g->len) == 0)
3508 3509
			matched = 1;
		break;
3510 3511 3512 3513
	case MATCH_GLOB:
		if (glob_match(g->search, str))
			matched = 1;
		break;
3514 3515 3516 3517 3518
	}

	return matched;
}

3519
static int
3520
enter_record(struct ftrace_hash *hash, struct dyn_ftrace *rec, int clear_filter)
3521
{
3522 3523 3524
	struct ftrace_func_entry *entry;
	int ret = 0;

3525
	entry = ftrace_lookup_ip(hash, rec->ip);
3526
	if (clear_filter) {
3527 3528 3529
		/* Do nothing if it doesn't exist */
		if (!entry)
			return 0;
3530

3531
		free_hash_entry(hash, entry);
3532 3533 3534 3535
	} else {
		/* Do nothing if it exists */
		if (entry)
			return 0;
3536

3537
		ret = add_hash_entry(hash, rec->ip);
3538 3539
	}
	return ret;
3540 3541
}

3542
static int
D
Dmitry Safonov 已提交
3543 3544
ftrace_match_record(struct dyn_ftrace *rec, struct ftrace_glob *func_g,
		struct ftrace_glob *mod_g, int exclude_mod)
3545 3546
{
	char str[KSYM_SYMBOL_LEN];
3547 3548 3549 3550
	char *modname;

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

D
Dmitry Safonov 已提交
3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570
	if (mod_g) {
		int mod_matches = (modname) ? ftrace_match(modname, mod_g) : 0;

		/* blank module name to match all modules */
		if (!mod_g->len) {
			/* blank module globbing: modname xor exclude_mod */
			if ((!exclude_mod) != (!modname))
				goto func_match;
			return 0;
		}

		/* not matching the module */
		if (!modname || !mod_matches) {
			if (exclude_mod)
				goto func_match;
			else
				return 0;
		}

		if (mod_matches && exclude_mod)
3571 3572
			return 0;

D
Dmitry Safonov 已提交
3573
func_match:
3574
		/* blank search means to match all funcs in the mod */
3575
		if (!func_g->len)
3576 3577
			return 1;
	}
3578

3579
	return ftrace_match(str, func_g);
3580 3581
}

3582
static int
3583
match_records(struct ftrace_hash *hash, char *func, int len, char *mod)
3584 3585 3586
{
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
3587
	struct ftrace_glob func_g = { .type = MATCH_FULL };
D
Dmitry Safonov 已提交
3588 3589 3590
	struct ftrace_glob mod_g = { .type = MATCH_FULL };
	struct ftrace_glob *mod_match = (mod) ? &mod_g : NULL;
	int exclude_mod = 0;
3591
	int found = 0;
3592
	int ret;
3593
	int clear_filter;
3594

D
Dmitry Safonov 已提交
3595
	if (func) {
3596 3597 3598
		func_g.type = filter_parse_regex(func, len, &func_g.search,
						 &clear_filter);
		func_g.len = strlen(func_g.search);
3599
	}
3600

D
Dmitry Safonov 已提交
3601 3602 3603 3604
	if (mod) {
		mod_g.type = filter_parse_regex(mod, strlen(mod),
				&mod_g.search, &exclude_mod);
		mod_g.len = strlen(mod_g.search);
3605
	}
3606

3607
	mutex_lock(&ftrace_lock);
3608

3609 3610
	if (unlikely(ftrace_disabled))
		goto out_unlock;
3611

3612
	do_for_each_ftrace_rec(pg, rec) {
3613 3614 3615 3616

		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

D
Dmitry Safonov 已提交
3617
		if (ftrace_match_record(rec, &func_g, mod_match, exclude_mod)) {
3618
			ret = enter_record(hash, rec, clear_filter);
3619 3620 3621 3622
			if (ret < 0) {
				found = ret;
				goto out_unlock;
			}
3623
			found = 1;
3624 3625
		}
	} while_for_each_ftrace_rec();
3626
 out_unlock:
3627
	mutex_unlock(&ftrace_lock);
3628 3629

	return found;
3630 3631
}

3632
static int
3633
ftrace_match_records(struct ftrace_hash *hash, char *buff, int len)
3634
{
3635
	return match_records(hash, buff, len, NULL);
3636 3637 3638
}


3639 3640 3641 3642 3643 3644
/*
 * We register the module command as a template to show others how
 * to register the a command as well.
 */

static int
3645
ftrace_mod_callback(struct ftrace_hash *hash,
3646
		    char *func, char *cmd, char *module, int enable)
3647
{
3648
	int ret;
3649 3650 3651 3652 3653 3654 3655 3656

	/*
	 * 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.
	 */
3657
	ret = match_records(hash, func, strlen(func), module);
3658
	if (!ret)
3659
		return -EINVAL;
3660 3661 3662
	if (ret < 0)
		return ret;
	return 0;
3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673
}

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);
}
3674
core_initcall(ftrace_mod_cmd_init);
3675

3676
static void function_trace_probe_call(unsigned long ip, unsigned long parent_ip,
3677
				      struct ftrace_ops *op, struct pt_regs *pt_regs)
3678
{
S
Steven Rostedt 已提交
3679
	struct ftrace_func_probe *entry;
3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694
	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.
	 */
3695
	preempt_disable_notrace();
3696
	hlist_for_each_entry_rcu_notrace(entry, hhd, node) {
3697 3698 3699
		if (entry->ip == ip)
			entry->ops->func(ip, parent_ip, &entry->data);
	}
3700
	preempt_enable_notrace();
3701 3702
}

S
Steven Rostedt 已提交
3703
static struct ftrace_ops trace_probe_ops __read_mostly =
3704
{
3705
	.func		= function_trace_probe_call,
3706
	.flags		= FTRACE_OPS_FL_INITIALIZED,
3707
	INIT_OPS_HASH(trace_probe_ops)
3708 3709
};

S
Steven Rostedt 已提交
3710
static int ftrace_probe_registered;
3711

3712
static void __enable_ftrace_function_probe(struct ftrace_ops_hash *old_hash)
3713
{
3714
	int ret;
3715 3716
	int i;

3717 3718 3719
	if (ftrace_probe_registered) {
		/* still need to update the function call sites */
		if (ftrace_enabled)
3720 3721
			ftrace_run_modify_code(&trace_probe_ops, FTRACE_UPDATE_CALLS,
					       old_hash);
3722
		return;
3723
	}
3724 3725 3726 3727 3728 3729 3730 3731 3732 3733

	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;

3734
	ret = ftrace_startup(&trace_probe_ops, 0);
3735

S
Steven Rostedt 已提交
3736
	ftrace_probe_registered = 1;
3737 3738
}

S
Steven Rostedt 已提交
3739
static void __disable_ftrace_function_probe(void)
3740 3741 3742
{
	int i;

S
Steven Rostedt 已提交
3743
	if (!ftrace_probe_registered)
3744 3745 3746 3747 3748 3749 3750 3751 3752
		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 */
3753
	ftrace_shutdown(&trace_probe_ops, 0);
3754

S
Steven Rostedt 已提交
3755
	ftrace_probe_registered = 0;
3756 3757 3758
}


3759
static void ftrace_free_entry(struct ftrace_func_probe *entry)
3760 3761
{
	if (entry->ops->free)
3762
		entry->ops->free(entry->ops, entry->ip, &entry->data);
3763 3764 3765 3766
	kfree(entry);
}

int
S
Steven Rostedt 已提交
3767
register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3768 3769
			      void *data)
{
3770
	struct ftrace_ops_hash old_hash_ops;
S
Steven Rostedt 已提交
3771
	struct ftrace_func_probe *entry;
3772
	struct ftrace_glob func_g;
3773
	struct ftrace_hash **orig_hash = &trace_probe_ops.func_hash->filter_hash;
3774
	struct ftrace_hash *old_hash = *orig_hash;
3775
	struct ftrace_hash *hash;
3776 3777
	struct ftrace_page *pg;
	struct dyn_ftrace *rec;
3778
	int not;
S
Steven Rostedt 已提交
3779
	unsigned long key;
3780
	int count = 0;
3781
	int ret;
3782

3783 3784 3785
	func_g.type = filter_parse_regex(glob, strlen(glob),
			&func_g.search, &not);
	func_g.len = strlen(func_g.search);
3786

S
Steven Rostedt 已提交
3787
	/* we do not support '!' for function probes */
3788 3789 3790
	if (WARN_ON(not))
		return -EINVAL;

3791
	mutex_lock(&trace_probe_ops.func_hash->regex_lock);
3792

3793 3794 3795 3796
	old_hash_ops.filter_hash = old_hash;
	/* Probes only have filters */
	old_hash_ops.notrace_hash = NULL;

3797
	hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, old_hash);
3798 3799
	if (!hash) {
		count = -ENOMEM;
3800
		goto out;
3801 3802 3803 3804
	}

	if (unlikely(ftrace_disabled)) {
		count = -ENODEV;
3805
		goto out;
3806
	}
3807

3808 3809
	mutex_lock(&ftrace_lock);

3810
	do_for_each_ftrace_rec(pg, rec) {
3811

3812 3813 3814
		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

D
Dmitry Safonov 已提交
3815
		if (!ftrace_match_record(rec, &func_g, NULL, 0))
3816 3817 3818 3819
			continue;

		entry = kmalloc(sizeof(*entry), GFP_KERNEL);
		if (!entry) {
S
Steven Rostedt 已提交
3820
			/* If we did not process any, then return error */
3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834
			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.
		 */
3835 3836
		if (ops->init) {
			if (ops->init(ops, rec->ip, &entry->data) < 0) {
3837 3838 3839 3840 3841 3842
				/* caller does not like this func */
				kfree(entry);
				continue;
			}
		}

3843 3844 3845 3846 3847 3848 3849
		ret = enter_record(hash, rec, 0);
		if (ret < 0) {
			kfree(entry);
			count = ret;
			goto out_unlock;
		}

3850 3851 3852 3853 3854 3855 3856
		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();
3857 3858

	ret = ftrace_hash_move(&trace_probe_ops, 1, orig_hash, hash);
3859

3860
	__enable_ftrace_function_probe(&old_hash_ops);
3861

3862 3863 3864
	if (!ret)
		free_ftrace_hash_rcu(old_hash);
	else
3865 3866
		count = ret;

3867
 out_unlock:
3868 3869
	mutex_unlock(&ftrace_lock);
 out:
3870
	mutex_unlock(&trace_probe_ops.func_hash->regex_lock);
3871
	free_ftrace_hash(hash);
3872 3873 3874 3875 3876

	return count;
}

enum {
S
Steven Rostedt 已提交
3877 3878
	PROBE_TEST_FUNC		= 1,
	PROBE_TEST_DATA		= 2
3879 3880 3881
};

static void
S
Steven Rostedt 已提交
3882
__unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3883 3884
				  void *data, int flags)
{
3885
	struct ftrace_func_entry *rec_entry;
S
Steven Rostedt 已提交
3886
	struct ftrace_func_probe *entry;
3887
	struct ftrace_func_probe *p;
3888
	struct ftrace_glob func_g;
3889
	struct ftrace_hash **orig_hash = &trace_probe_ops.func_hash->filter_hash;
3890
	struct ftrace_hash *old_hash = *orig_hash;
3891
	struct list_head free_list;
3892
	struct ftrace_hash *hash;
3893
	struct hlist_node *tmp;
3894
	char str[KSYM_SYMBOL_LEN];
3895
	int i, ret;
3896

3897
	if (glob && (strcmp(glob, "*") == 0 || !strlen(glob)))
3898
		func_g.search = NULL;
3899
	else if (glob) {
3900 3901
		int not;

3902 3903 3904 3905
		func_g.type = filter_parse_regex(glob, strlen(glob),
						 &func_g.search, &not);
		func_g.len = strlen(func_g.search);
		func_g.search = glob;
3906

S
Steven Rostedt 已提交
3907
		/* we do not support '!' for function probes */
3908 3909 3910 3911
		if (WARN_ON(not))
			return;
	}

3912
	mutex_lock(&trace_probe_ops.func_hash->regex_lock);
3913 3914 3915 3916 3917 3918

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

3919 3920
	INIT_LIST_HEAD(&free_list);

3921 3922 3923
	for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) {
		struct hlist_head *hhd = &ftrace_func_hash[i];

3924
		hlist_for_each_entry_safe(entry, tmp, hhd, node) {
3925 3926

			/* break up if statements for readability */
S
Steven Rostedt 已提交
3927
			if ((flags & PROBE_TEST_FUNC) && entry->ops != ops)
3928 3929
				continue;

S
Steven Rostedt 已提交
3930
			if ((flags & PROBE_TEST_DATA) && entry->data != data)
3931 3932 3933
				continue;

			/* do this last, since it is the most expensive */
3934
			if (func_g.search) {
3935 3936
				kallsyms_lookup(entry->ip, NULL, NULL,
						NULL, str);
3937
				if (!ftrace_match(str, &func_g))
3938 3939 3940
					continue;
			}

3941 3942 3943 3944 3945
			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);

3946
			hlist_del_rcu(&entry->node);
3947
			list_add(&entry->free_list, &free_list);
3948 3949
		}
	}
3950
	mutex_lock(&ftrace_lock);
S
Steven Rostedt 已提交
3951
	__disable_ftrace_function_probe();
3952 3953 3954 3955
	/*
	 * Remove after the disable is called. Otherwise, if the last
	 * probe is removed, a null hash means *all enabled*.
	 */
3956
	ret = ftrace_hash_move(&trace_probe_ops, 1, orig_hash, hash);
3957
	synchronize_sched();
3958 3959 3960
	if (!ret)
		free_ftrace_hash_rcu(old_hash);

3961 3962 3963 3964
	list_for_each_entry_safe(entry, p, &free_list, free_list) {
		list_del(&entry->free_list);
		ftrace_free_entry(entry);
	}
3965
	mutex_unlock(&ftrace_lock);
3966

3967
 out_unlock:
3968
	mutex_unlock(&trace_probe_ops.func_hash->regex_lock);
3969
	free_ftrace_hash(hash);
3970 3971 3972
}

void
S
Steven Rostedt 已提交
3973
unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops,
3974 3975
				void *data)
{
S
Steven Rostedt 已提交
3976 3977
	__unregister_ftrace_function_probe(glob, ops, data,
					  PROBE_TEST_FUNC | PROBE_TEST_DATA);
3978 3979 3980
}

void
S
Steven Rostedt 已提交
3981
unregister_ftrace_function_probe_func(char *glob, struct ftrace_probe_ops *ops)
3982
{
S
Steven Rostedt 已提交
3983
	__unregister_ftrace_function_probe(glob, ops, NULL, PROBE_TEST_FUNC);
3984 3985
}

S
Steven Rostedt 已提交
3986
void unregister_ftrace_function_probe_all(char *glob)
3987
{
S
Steven Rostedt 已提交
3988
	__unregister_ftrace_function_probe(glob, NULL, NULL, 0);
3989 3990
}

3991 3992 3993
static LIST_HEAD(ftrace_commands);
static DEFINE_MUTEX(ftrace_cmd_mutex);

3994 3995 3996 3997 3998
/*
 * Currently we only register ftrace commands from __init, so mark this
 * __init too.
 */
__init int register_ftrace_command(struct ftrace_func_command *cmd)
3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016
{
	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;
}

4017 4018 4019 4020 4021
/*
 * Currently we only unregister ftrace commands from __init, so mark
 * this __init too.
 */
__init int unregister_ftrace_command(struct ftrace_func_command *cmd)
4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039
{
	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;
}

4040 4041
static int ftrace_process_regex(struct ftrace_hash *hash,
				char *buff, int len, int enable)
4042
{
4043
	char *func, *command, *next = buff;
S
Steven Rostedt 已提交
4044
	struct ftrace_func_command *p;
4045
	int ret = -EINVAL;
4046 4047 4048 4049

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

	if (!next) {
4050
		ret = ftrace_match_records(hash, func, len);
4051 4052 4053 4054 4055
		if (!ret)
			ret = -EINVAL;
		if (ret < 0)
			return ret;
		return 0;
4056 4057
	}

4058
	/* command found */
4059 4060 4061

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

4062 4063 4064
	mutex_lock(&ftrace_cmd_mutex);
	list_for_each_entry(p, &ftrace_commands, list) {
		if (strcmp(p->name, command) == 0) {
4065
			ret = p->func(hash, func, command, next, enable);
4066 4067
			goto out_unlock;
		}
4068
	}
4069 4070
 out_unlock:
	mutex_unlock(&ftrace_cmd_mutex);
4071

4072
	return ret;
4073 4074
}

I
Ingo Molnar 已提交
4075
static ssize_t
4076 4077
ftrace_regex_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos, int enable)
4078 4079
{
	struct ftrace_iterator *iter;
4080 4081
	struct trace_parser *parser;
	ssize_t ret, read;
4082

4083
	if (!cnt)
4084 4085 4086 4087 4088 4089 4090 4091
		return 0;

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

4092
	if (unlikely(ftrace_disabled))
4093 4094 4095
		return -ENODEV;

	/* iter->hash is a local copy, so we don't need regex_lock */
4096

4097 4098
	parser = &iter->parser;
	read = trace_get_user(parser, ubuf, cnt, ppos);
4099

4100
	if (read >= 0 && trace_parser_loaded(parser) &&
4101
	    !trace_parser_cont(parser)) {
4102
		ret = ftrace_process_regex(iter->hash, parser->buffer,
4103
					   parser->idx, enable);
4104
		trace_parser_clear(parser);
4105
		if (ret < 0)
4106
			goto out;
4107
	}
4108 4109

	ret = read;
4110
 out:
4111 4112 4113
	return ret;
}

4114
ssize_t
4115 4116 4117 4118 4119 4120
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);
}

4121
ssize_t
4122 4123 4124 4125 4126 4127
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);
}

4128
static int
4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146
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);
}

4147
static void ftrace_ops_update_code(struct ftrace_ops *ops,
4148
				   struct ftrace_ops_hash *old_hash)
4149
{
4150 4151 4152 4153 4154 4155
	struct ftrace_ops *op;

	if (!ftrace_enabled)
		return;

	if (ops->flags & FTRACE_OPS_FL_ENABLED) {
4156
		ftrace_run_modify_code(ops, FTRACE_UPDATE_CALLS, old_hash);
4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173 4174 4175
		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);
4176 4177
}

4178 4179 4180
static int
ftrace_set_hash(struct ftrace_ops *ops, unsigned char *buf, int len,
		unsigned long ip, int remove, int reset, int enable)
4181
{
4182
	struct ftrace_hash **orig_hash;
4183
	struct ftrace_ops_hash old_hash_ops;
4184
	struct ftrace_hash *old_hash;
4185
	struct ftrace_hash *hash;
4186
	int ret;
4187

4188
	if (unlikely(ftrace_disabled))
4189
		return -ENODEV;
4190

4191
	mutex_lock(&ops->func_hash->regex_lock);
4192

4193
	if (enable)
4194
		orig_hash = &ops->func_hash->filter_hash;
4195
	else
4196
		orig_hash = &ops->func_hash->notrace_hash;
4197

4198 4199 4200 4201 4202
	if (reset)
		hash = alloc_ftrace_hash(FTRACE_HASH_DEFAULT_BITS);
	else
		hash = alloc_and_copy_ftrace_hash(FTRACE_HASH_DEFAULT_BITS, *orig_hash);

4203 4204 4205 4206
	if (!hash) {
		ret = -ENOMEM;
		goto out_regex_unlock;
	}
4207

4208 4209 4210 4211
	if (buf && !ftrace_match_records(hash, buf, len)) {
		ret = -EINVAL;
		goto out_regex_unlock;
	}
4212 4213 4214 4215 4216
	if (ip) {
		ret = ftrace_match_addr(hash, ip, remove);
		if (ret < 0)
			goto out_regex_unlock;
	}
4217 4218

	mutex_lock(&ftrace_lock);
4219
	old_hash = *orig_hash;
4220 4221
	old_hash_ops.filter_hash = ops->func_hash->filter_hash;
	old_hash_ops.notrace_hash = ops->func_hash->notrace_hash;
4222
	ret = ftrace_hash_move(ops, enable, orig_hash, hash);
4223
	if (!ret) {
4224
		ftrace_ops_update_code(ops, &old_hash_ops);
4225 4226
		free_ftrace_hash_rcu(old_hash);
	}
4227 4228
	mutex_unlock(&ftrace_lock);

4229
 out_regex_unlock:
4230
	mutex_unlock(&ops->func_hash->regex_lock);
4231 4232 4233

	free_ftrace_hash(hash);
	return ret;
4234 4235
}

4236 4237 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 4253 4254 4255
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)
{
4256
	ftrace_ops_init(ops);
4257 4258 4259 4260
	return ftrace_set_addr(ops, ip, remove, reset, 1);
}
EXPORT_SYMBOL_GPL(ftrace_set_filter_ip);

4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277
/**
 * ftrace_ops_set_global_filter - setup ops to use global filters
 * @ops - the ops which will use the global filters
 *
 * ftrace users who need global function trace filtering should call this.
 * It can set the global filter only if ops were not initialized before.
 */
void ftrace_ops_set_global_filter(struct ftrace_ops *ops)
{
	if (ops->flags & FTRACE_OPS_FL_INITIALIZED)
		return;

	ftrace_ops_init(ops);
	ops->func_hash = &global_ops.local_hash;
}
EXPORT_SYMBOL_GPL(ftrace_ops_set_global_filter);

4278 4279 4280 4281 4282 4283 4284
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);
}

4285 4286
/**
 * ftrace_set_filter - set a function to filter on in ftrace
4287 4288 4289 4290 4291 4292 4293 4294
 * @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.
 */
4295
int ftrace_set_filter(struct ftrace_ops *ops, unsigned char *buf,
4296 4297
		       int len, int reset)
{
4298
	ftrace_ops_init(ops);
4299
	return ftrace_set_regex(ops, buf, len, reset, 1);
4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313
}
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.
 */
4314
int ftrace_set_notrace(struct ftrace_ops *ops, unsigned char *buf,
4315 4316
			int len, int reset)
{
4317
	ftrace_ops_init(ops);
4318
	return ftrace_set_regex(ops, buf, len, reset, 0);
4319 4320 4321
}
EXPORT_SYMBOL_GPL(ftrace_set_notrace);
/**
4322
 * ftrace_set_global_filter - set a function to filter on with global tracers
4323 4324 4325 4326 4327 4328 4329
 * @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.
 */
4330
void ftrace_set_global_filter(unsigned char *buf, int len, int reset)
4331
{
4332
	ftrace_set_regex(&global_ops, buf, len, reset, 1);
4333
}
4334
EXPORT_SYMBOL_GPL(ftrace_set_global_filter);
4335

4336
/**
4337
 * ftrace_set_global_notrace - set a function to not trace with global tracers
4338 4339 4340 4341 4342 4343 4344 4345
 * @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.
 */
4346
void ftrace_set_global_notrace(unsigned char *buf, int len, int reset)
4347
{
4348
	ftrace_set_regex(&global_ops, buf, len, reset, 0);
4349
}
4350
EXPORT_SYMBOL_GPL(ftrace_set_global_notrace);
4351

4352 4353 4354 4355 4356 4357 4358
/*
 * 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;

4359 4360 4361
/* Used by function selftest to not test if filter is set */
bool ftrace_filter_param __initdata;

4362 4363
static int __init set_ftrace_notrace(char *str)
{
4364
	ftrace_filter_param = true;
4365
	strlcpy(ftrace_notrace_buf, str, FTRACE_FILTER_SIZE);
4366 4367 4368 4369 4370 4371
	return 1;
}
__setup("ftrace_notrace=", set_ftrace_notrace);

static int __init set_ftrace_filter(char *str)
{
4372
	ftrace_filter_param = true;
4373
	strlcpy(ftrace_filter_buf, str, FTRACE_FILTER_SIZE);
4374 4375 4376 4377
	return 1;
}
__setup("ftrace_filter=", set_ftrace_filter);

4378
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
4379
static char ftrace_graph_buf[FTRACE_FILTER_SIZE] __initdata;
4380
static char ftrace_graph_notrace_buf[FTRACE_FILTER_SIZE] __initdata;
4381
static int ftrace_graph_set_hash(struct ftrace_hash *hash, char *buffer);
4382

4383 4384 4385
static unsigned long save_global_trampoline;
static unsigned long save_global_flags;

4386 4387
static int __init set_graph_function(char *str)
{
4388
	strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE);
4389 4390 4391 4392
	return 1;
}
__setup("ftrace_graph_filter=", set_graph_function);

4393 4394 4395 4396 4397 4398 4399 4400
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)
4401 4402 4403
{
	int ret;
	char *func;
4404
	struct ftrace_hash *hash;
4405

4406 4407 4408 4409
	if (enable)
		hash = ftrace_graph_hash;
	else
		hash = ftrace_graph_notrace_hash;
4410 4411 4412 4413

	while (buf) {
		func = strsep(&buf, ",");
		/* we allow only one expression at a time */
4414
		ret = ftrace_graph_set_hash(hash, func);
4415 4416 4417 4418 4419 4420 4421
		if (ret)
			printk(KERN_DEBUG "ftrace: function %s not "
					  "traceable\n", func);
	}
}
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4422 4423
void __init
ftrace_set_early_filter(struct ftrace_ops *ops, char *buf, int enable)
4424 4425 4426
{
	char *func;

4427 4428
	ftrace_ops_init(ops);

4429 4430
	while (buf) {
		func = strsep(&buf, ",");
4431
		ftrace_set_regex(ops, func, strlen(func), 0, enable);
4432 4433 4434 4435 4436 4437
	}
}

static void __init set_ftrace_early_filters(void)
{
	if (ftrace_filter_buf[0])
4438
		ftrace_set_early_filter(&global_ops, ftrace_filter_buf, 1);
4439
	if (ftrace_notrace_buf[0])
4440
		ftrace_set_early_filter(&global_ops, ftrace_notrace_buf, 0);
4441 4442
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
	if (ftrace_graph_buf[0])
4443 4444 4445
		set_ftrace_early_graph(ftrace_graph_buf, 1);
	if (ftrace_graph_notrace_buf[0])
		set_ftrace_early_graph(ftrace_graph_notrace_buf, 0);
4446
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
4447 4448
}

4449
int ftrace_regex_release(struct inode *inode, struct file *file)
4450 4451
{
	struct seq_file *m = (struct seq_file *)file->private_data;
4452
	struct ftrace_ops_hash old_hash_ops;
4453
	struct ftrace_iterator *iter;
4454
	struct ftrace_hash **orig_hash;
4455
	struct ftrace_hash *old_hash;
4456
	struct trace_parser *parser;
4457
	int filter_hash;
4458
	int ret;
4459 4460 4461 4462 4463 4464 4465

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

4466 4467 4468
	parser = &iter->parser;
	if (trace_parser_loaded(parser)) {
		parser->buffer[parser->idx] = 0;
4469
		ftrace_match_records(iter->hash, parser->buffer, parser->idx);
4470 4471
	}

4472 4473
	trace_parser_put(parser);

4474
	mutex_lock(&iter->ops->func_hash->regex_lock);
4475

4476
	if (file->f_mode & FMODE_WRITE) {
4477 4478 4479
		filter_hash = !!(iter->flags & FTRACE_ITER_FILTER);

		if (filter_hash)
4480
			orig_hash = &iter->ops->func_hash->filter_hash;
4481
		else
4482
			orig_hash = &iter->ops->func_hash->notrace_hash;
4483

4484
		mutex_lock(&ftrace_lock);
4485
		old_hash = *orig_hash;
4486 4487
		old_hash_ops.filter_hash = iter->ops->func_hash->filter_hash;
		old_hash_ops.notrace_hash = iter->ops->func_hash->notrace_hash;
4488 4489
		ret = ftrace_hash_move(iter->ops, filter_hash,
				       orig_hash, iter->hash);
4490
		if (!ret) {
4491
			ftrace_ops_update_code(iter->ops, &old_hash_ops);
4492 4493
			free_ftrace_hash_rcu(old_hash);
		}
4494 4495
		mutex_unlock(&ftrace_lock);
	}
4496

4497
	mutex_unlock(&iter->ops->func_hash->regex_lock);
4498 4499
	free_ftrace_hash(iter->hash);
	kfree(iter);
4500

4501 4502 4503
	return 0;
}

4504
static const struct file_operations ftrace_avail_fops = {
4505 4506 4507
	.open = ftrace_avail_open,
	.read = seq_read,
	.llseek = seq_lseek,
L
Li Zefan 已提交
4508
	.release = seq_release_private,
4509 4510
};

4511 4512 4513 4514 4515 4516 4517
static const struct file_operations ftrace_enabled_fops = {
	.open = ftrace_enabled_open,
	.read = seq_read,
	.llseek = seq_lseek,
	.release = seq_release_private,
};

4518
static const struct file_operations ftrace_filter_fops = {
4519
	.open = ftrace_filter_open,
L
Lai Jiangshan 已提交
4520
	.read = seq_read,
4521
	.write = ftrace_filter_write,
4522
	.llseek = tracing_lseek,
4523
	.release = ftrace_regex_release,
4524 4525
};

4526
static const struct file_operations ftrace_notrace_fops = {
4527
	.open = ftrace_notrace_open,
L
Lai Jiangshan 已提交
4528
	.read = seq_read,
4529
	.write = ftrace_notrace_write,
4530
	.llseek = tracing_lseek,
4531
	.release = ftrace_regex_release,
4532 4533
};

4534 4535 4536 4537
#ifdef CONFIG_FUNCTION_GRAPH_TRACER

static DEFINE_MUTEX(graph_lock);

4538 4539 4540 4541 4542 4543 4544
struct ftrace_hash *ftrace_graph_hash = EMPTY_HASH;
struct ftrace_hash *ftrace_graph_notrace_hash = EMPTY_HASH;

enum graph_filter_type {
	GRAPH_FILTER_NOTRACE	= 0,
	GRAPH_FILTER_FUNCTION,
};
4545

4546
struct ftrace_graph_data {
4547 4548 4549 4550 4551
	struct ftrace_hash *hash;
	struct ftrace_func_entry *entry;
	int idx;   /* for hash table iteration */
	enum graph_filter_type type;
	struct ftrace_hash *new_hash;
4552 4553 4554
	const struct seq_operations *seq_ops;
};

4555
static void *
4556
__g_next(struct seq_file *m, loff_t *pos)
4557
{
4558
	struct ftrace_graph_data *fgd = m->private;
4559 4560 4561
	struct ftrace_func_entry *entry = fgd->entry;
	struct hlist_head *head;
	int i, idx = fgd->idx;
4562

4563
	if (*pos >= fgd->hash->count)
4564
		return NULL;
4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583

	if (entry) {
		hlist_for_each_entry_continue(entry, hlist) {
			fgd->entry = entry;
			return entry;
		}

		idx++;
	}

	for (i = idx; i < 1 << fgd->hash->size_bits; i++) {
		head = &fgd->hash->buckets[i];
		hlist_for_each_entry(entry, head, hlist) {
			fgd->entry = entry;
			fgd->idx = i;
			return entry;
		}
	}
	return NULL;
4584
}
4585

4586 4587 4588 4589 4590
static void *
g_next(struct seq_file *m, void *v, loff_t *pos)
{
	(*pos)++;
	return __g_next(m, pos);
4591 4592 4593 4594
}

static void *g_start(struct seq_file *m, loff_t *pos)
{
4595 4596
	struct ftrace_graph_data *fgd = m->private;

4597 4598
	mutex_lock(&graph_lock);

4599
	/* Nothing, tell g_show to print all functions are enabled */
4600
	if (ftrace_hash_empty(fgd->hash) && !*pos)
4601 4602
		return (void *)1;

4603 4604
	fgd->idx = 0;
	fgd->entry = NULL;
4605
	return __g_next(m, pos);
4606 4607 4608 4609 4610 4611 4612 4613 4614
}

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

static int g_show(struct seq_file *m, void *v)
{
4615
	struct ftrace_func_entry *entry = v;
4616

4617
	if (!entry)
4618 4619
		return 0;

4620
	if (entry == (void *)1) {
4621 4622
		struct ftrace_graph_data *fgd = m->private;

4623
		if (fgd->type == GRAPH_FILTER_FUNCTION)
4624
			seq_puts(m, "#### all functions enabled ####\n");
4625
		else
4626
			seq_puts(m, "#### no functions disabled ####\n");
4627 4628 4629
		return 0;
	}

4630
	seq_printf(m, "%ps\n", (void *)entry->ip);
4631 4632 4633 4634

	return 0;
}

J
James Morris 已提交
4635
static const struct seq_operations ftrace_graph_seq_ops = {
4636 4637 4638 4639 4640 4641 4642
	.start = g_start,
	.next = g_next,
	.stop = g_stop,
	.show = g_show,
};

static int
4643 4644
__ftrace_graph_open(struct inode *inode, struct file *file,
		    struct ftrace_graph_data *fgd)
4645 4646
{
	int ret = 0;
4647
	struct ftrace_hash *new_hash = NULL;
4648

4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660
	if (file->f_mode & FMODE_WRITE) {
		const int size_bits = FTRACE_HASH_DEFAULT_BITS;

		if (file->f_flags & O_TRUNC)
			new_hash = alloc_ftrace_hash(size_bits);
		else
			new_hash = alloc_and_copy_ftrace_hash(size_bits,
							      fgd->hash);
		if (!new_hash) {
			ret = -ENOMEM;
			goto out;
		}
4661 4662
	}

4663
	if (file->f_mode & FMODE_READ) {
4664
		ret = seq_open(file, &ftrace_graph_seq_ops);
4665 4666 4667
		if (!ret) {
			struct seq_file *m = file->private_data;
			m->private = fgd;
4668 4669 4670 4671
		} else {
			/* Failed */
			free_ftrace_hash(new_hash);
			new_hash = NULL;
4672 4673 4674
		}
	} else
		file->private_data = fgd;
4675

4676 4677
out:
	fgd->new_hash = new_hash;
4678 4679 4680
	return ret;
}

4681 4682 4683 4684
static int
ftrace_graph_open(struct inode *inode, struct file *file)
{
	struct ftrace_graph_data *fgd;
4685
	int ret;
4686 4687 4688 4689 4690 4691 4692 4693

	if (unlikely(ftrace_disabled))
		return -ENODEV;

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

4694 4695 4696 4697
	mutex_lock(&graph_lock);

	fgd->hash = ftrace_graph_hash;
	fgd->type = GRAPH_FILTER_FUNCTION;
4698 4699
	fgd->seq_ops = &ftrace_graph_seq_ops;

4700 4701 4702 4703 4704 4705
	ret = __ftrace_graph_open(inode, file, fgd);
	if (ret < 0)
		kfree(fgd);

	mutex_unlock(&graph_lock);
	return ret;
4706 4707
}

4708 4709 4710 4711
static int
ftrace_graph_notrace_open(struct inode *inode, struct file *file)
{
	struct ftrace_graph_data *fgd;
4712
	int ret;
4713 4714 4715 4716 4717 4718 4719 4720

	if (unlikely(ftrace_disabled))
		return -ENODEV;

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

4721 4722 4723 4724
	mutex_lock(&graph_lock);

	fgd->hash = ftrace_graph_notrace_hash;
	fgd->type = GRAPH_FILTER_NOTRACE;
4725 4726
	fgd->seq_ops = &ftrace_graph_seq_ops;

4727 4728 4729 4730 4731 4732
	ret = __ftrace_graph_open(inode, file, fgd);
	if (ret < 0)
		kfree(fgd);

	mutex_unlock(&graph_lock);
	return ret;
4733 4734
}

4735 4736 4737
static int
ftrace_graph_release(struct inode *inode, struct file *file)
{
4738 4739
	struct ftrace_graph_data *fgd;

4740 4741 4742
	if (file->f_mode & FMODE_READ) {
		struct seq_file *m = file->private_data;

4743
		fgd = m->private;
4744
		seq_release(inode, file);
4745
	} else {
4746
		fgd = file->private_data;
4747 4748
	}

4749 4750 4751
	kfree(fgd->new_hash);
	kfree(fgd);

4752 4753 4754
	return 0;
}

4755
static int
4756
ftrace_graph_set_hash(struct ftrace_hash *hash, char *buffer)
4757
{
4758
	struct ftrace_glob func_g;
4759 4760
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;
4761
	struct ftrace_func_entry *entry;
4762
	int fail = 1;
4763
	int not;
4764

4765
	/* decode regex */
4766 4767
	func_g.type = filter_parse_regex(buffer, strlen(buffer),
					 &func_g.search, &not);
4768

4769
	func_g.len = strlen(func_g.search);
4770

4771
	mutex_lock(&ftrace_lock);
4772 4773 4774 4775 4776 4777

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

4778 4779
	do_for_each_ftrace_rec(pg, rec) {

4780 4781 4782
		if (rec->flags & FTRACE_FL_DISABLED)
			continue;

D
Dmitry Safonov 已提交
4783
		if (ftrace_match_record(rec, &func_g, NULL, 0)) {
4784
			entry = ftrace_lookup_ip(hash, rec->ip);
4785 4786 4787

			if (!not) {
				fail = 0;
4788 4789 4790 4791 4792

				if (entry)
					continue;
				if (add_hash_entry(hash, rec->ip) < 0)
					goto out;
4793
			} else {
4794 4795
				if (entry) {
					free_hash_entry(hash, entry);
4796 4797 4798
					fail = 0;
				}
			}
4799
		}
4800
	} while_for_each_ftrace_rec();
4801
out:
4802
	mutex_unlock(&ftrace_lock);
4803

4804 4805 4806 4807
	if (fail)
		return -EINVAL;

	return 0;
4808 4809 4810 4811 4812 4813
}

static ssize_t
ftrace_graph_write(struct file *file, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
4814
	struct trace_parser parser;
4815
	ssize_t read, ret = 0;
4816
	struct ftrace_graph_data *fgd = file->private_data;
4817
	struct ftrace_hash *old_hash, *new_hash;
4818

4819
	if (!cnt)
4820 4821
		return 0;

4822 4823
	if (trace_parser_get_init(&parser, FTRACE_BUFF_MAX))
		return -ENOMEM;
4824

4825
	read = trace_get_user(&parser, ubuf, cnt, ppos);
4826

4827
	if (read >= 0 && trace_parser_loaded((&parser))) {
4828 4829
		parser.buffer[parser.idx] = 0;

4830 4831
		mutex_lock(&graph_lock);

4832
		/* we allow only one expression at a time */
4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844
		ret = ftrace_graph_set_hash(fgd->new_hash,
					    parser.buffer);

		old_hash = fgd->hash;
		new_hash = __ftrace_hash_move(fgd->new_hash);
		if (!new_hash)
			ret = -ENOMEM;

		if (fgd->type == GRAPH_FILTER_FUNCTION)
			rcu_assign_pointer(ftrace_graph_hash, new_hash);
		else
			rcu_assign_pointer(ftrace_graph_notrace_hash, new_hash);
4845 4846

		mutex_unlock(&graph_lock);
4847 4848 4849 4850 4851

		/* Wait till all users are no longer using the old hash */
		synchronize_sched();

		free_ftrace_hash(old_hash);
4852 4853
	}

4854 4855
	if (!ret)
		ret = read;
4856

4857
	trace_parser_put(&parser);
4858 4859 4860 4861 4862

	return ret;
}

static const struct file_operations ftrace_graph_fops = {
4863 4864 4865
	.open		= ftrace_graph_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
4866
	.llseek		= tracing_lseek,
4867
	.release	= ftrace_graph_release,
4868
};
4869 4870 4871 4872 4873

static const struct file_operations ftrace_graph_notrace_fops = {
	.open		= ftrace_graph_notrace_open,
	.read		= seq_read,
	.write		= ftrace_graph_write,
4874
	.llseek		= tracing_lseek,
4875 4876
	.release	= ftrace_graph_release,
};
4877 4878
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908
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);
}

4909
static __init int ftrace_init_dyn_tracefs(struct dentry *d_tracer)
4910 4911
{

4912 4913
	trace_create_file("available_filter_functions", 0444,
			d_tracer, NULL, &ftrace_avail_fops);
4914

4915 4916 4917
	trace_create_file("enabled_functions", 0444,
			d_tracer, NULL, &ftrace_enabled_fops);

4918
	ftrace_create_filter_files(&global_ops, d_tracer);
4919

4920
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
4921
	trace_create_file("set_graph_function", 0444, d_tracer,
4922 4923
				    NULL,
				    &ftrace_graph_fops);
4924 4925 4926
	trace_create_file("set_graph_notrace", 0444, d_tracer,
				    NULL,
				    &ftrace_graph_notrace_fops);
4927 4928
#endif /* CONFIG_FUNCTION_GRAPH_TRACER */

4929 4930 4931
	return 0;
}

4932
static int ftrace_cmp_ips(const void *a, const void *b)
4933
{
4934 4935
	const unsigned long *ipa = a;
	const unsigned long *ipb = b;
4936

4937 4938 4939 4940 4941 4942 4943
	if (*ipa > *ipb)
		return 1;
	if (*ipa < *ipb)
		return -1;
	return 0;
}

4944
static int ftrace_process_locs(struct module *mod,
4945
			       unsigned long *start,
4946 4947
			       unsigned long *end)
{
4948
	struct ftrace_page *start_pg;
4949
	struct ftrace_page *pg;
4950
	struct dyn_ftrace *rec;
4951
	unsigned long count;
4952 4953
	unsigned long *p;
	unsigned long addr;
4954
	unsigned long flags = 0; /* Shut up gcc */
4955 4956 4957 4958 4959 4960 4961
	int ret = -ENOMEM;

	count = end - start;

	if (!count)
		return 0;

4962
	sort(start, count, sizeof(*start),
4963
	     ftrace_cmp_ips, NULL);
4964

4965 4966
	start_pg = ftrace_allocate_pages(count);
	if (!start_pg)
4967
		return -ENOMEM;
4968

S
Steven Rostedt 已提交
4969
	mutex_lock(&ftrace_lock);
4970

4971 4972 4973 4974 4975
	/*
	 * 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.
	 */
4976 4977 4978
	if (!mod) {
		WARN_ON(ftrace_pages || ftrace_pages_start);
		/* First initialization */
4979
		ftrace_pages = ftrace_pages_start = start_pg;
4980
	} else {
4981
		if (!ftrace_pages)
4982
			goto out;
4983

4984 4985 4986 4987
		if (WARN_ON(ftrace_pages->next)) {
			/* Hmm, we have free pages? */
			while (ftrace_pages->next)
				ftrace_pages = ftrace_pages->next;
4988
		}
4989

4990
		ftrace_pages->next = start_pg;
4991 4992
	}

4993
	p = start;
4994
	pg = start_pg;
4995 4996
	while (p < end) {
		addr = ftrace_call_adjust(*p++);
4997 4998 4999 5000 5001 5002 5003 5004
		/*
		 * 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;
5005 5006 5007 5008 5009 5010 5011 5012 5013 5014

		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;
5015 5016
	}

5017 5018 5019 5020 5021 5022
	/* We should have used all pages */
	WARN_ON(pg->next);

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

5023
	/*
5024 5025 5026 5027 5028 5029
	 * 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.
5030
	 */
5031 5032
	if (!mod)
		local_irq_save(flags);
5033
	ftrace_update_code(mod, start_pg);
5034 5035
	if (!mod)
		local_irq_restore(flags);
5036 5037
	ret = 0;
 out:
S
Steven Rostedt 已提交
5038
	mutex_unlock(&ftrace_lock);
5039

5040
	return ret;
5041 5042
}

5043
#ifdef CONFIG_MODULES
5044 5045 5046

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

5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 5057 5058 5059
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;
}

5060
void ftrace_release_mod(struct module *mod)
5061 5062
{
	struct dyn_ftrace *rec;
5063
	struct ftrace_page **last_pg;
5064
	struct ftrace_page *pg;
5065
	int order;
5066

5067 5068
	mutex_lock(&ftrace_lock);

5069
	if (ftrace_disabled)
5070
		goto out_unlock;
5071

5072 5073 5074 5075 5076 5077 5078
	/*
	 * 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];
5079
		if (within_module_core(rec->ip, mod)) {
5080
			/*
5081 5082
			 * As core pages are first, the first
			 * page should never be a module page.
5083
			 */
5084 5085 5086 5087 5088 5089 5090 5091
			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;
5092 5093 5094
			order = get_count_order(pg->size / ENTRIES_PER_PAGE);
			free_pages((unsigned long)pg->records, order);
			kfree(pg);
5095 5096 5097
		} else
			last_pg = &pg->next;
	}
5098
 out_unlock:
5099 5100 5101
	mutex_unlock(&ftrace_lock);
}

5102
void ftrace_module_enable(struct module *mod)
5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169 5170
{
	struct dyn_ftrace *rec;
	struct ftrace_page *pg;

	mutex_lock(&ftrace_lock);

	if (ftrace_disabled)
		goto out_unlock;

	/*
	 * 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.
	 *
	 * We also delay this to after the module code already set the
	 * text to read-only, as we now need to set it back to read-write
	 * so that we can modify the text.
	 */
	if (ftrace_start_up)
		ftrace_arch_code_modify_prepare();

	do_for_each_ftrace_rec(pg, rec) {
		int cnt;
		/*
		 * do_for_each_ftrace_rec() is a double loop.
		 * module text shares the pg. If a record is
		 * not part of this module, then skip this pg,
		 * which the "break" will do.
		 */
		if (!within_module_core(rec->ip, mod))
			break;

		cnt = 0;

		/*
		 * When adding a module, we need to check if tracers are
		 * currently enabled and if they are, and can trace this record,
		 * we need to enable the module functions as well as update the
		 * reference counts for those function records.
		 */
		if (ftrace_start_up)
			cnt += referenced_filters(rec);

		/* This clears FTRACE_FL_DISABLED */
		rec->flags = cnt;

		if (ftrace_start_up && cnt) {
			int failed = __ftrace_replace_code(rec, 1);
			if (failed) {
				ftrace_bug(failed, rec);
				goto out_loop;
			}
		}

	} while_for_each_ftrace_rec();

 out_loop:
	if (ftrace_start_up)
		ftrace_arch_code_modify_post_process();

 out_unlock:
	mutex_unlock(&ftrace_lock);
}

5171
void ftrace_module_init(struct module *mod)
5172
{
5173
	if (ftrace_disabled || !mod->num_ftrace_callsites)
5174
		return;
5175

5176 5177
	ftrace_process_locs(mod, mod->ftrace_callsites,
			    mod->ftrace_callsites + mod->num_ftrace_callsites);
5178
}
5179 5180
#endif /* CONFIG_MODULES */

5181 5182
void __init ftrace_init(void)
{
5183 5184
	extern unsigned long __start_mcount_loc[];
	extern unsigned long __stop_mcount_loc[];
5185
	unsigned long count, flags;
5186 5187 5188
	int ret;

	local_irq_save(flags);
5189
	ret = ftrace_dyn_arch_init();
5190
	local_irq_restore(flags);
5191
	if (ret)
5192 5193 5194
		goto failed;

	count = __stop_mcount_loc - __start_mcount_loc;
5195 5196
	if (!count) {
		pr_info("ftrace: No functions to be traced?\n");
5197
		goto failed;
5198 5199 5200 5201
	}

	pr_info("ftrace: allocating %ld entries in %ld pages\n",
		count, count / ENTRIES_PER_PAGE + 1);
5202 5203 5204

	last_ftrace_enabled = ftrace_enabled = 1;

5205
	ret = ftrace_process_locs(NULL,
5206
				  __start_mcount_loc,
5207 5208
				  __stop_mcount_loc);

5209 5210
	set_ftrace_early_filters();

5211 5212 5213 5214 5215
	return;
 failed:
	ftrace_disabled = 1;
}

5216 5217 5218 5219 5220 5221 5222
/* 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)
{
5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233

/*
 * 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
5234 5235 5236
	/* Currently, only non dynamic ops can have a trampoline */
	if (ops->flags & FTRACE_OPS_FL_DYNAMIC)
		return;
5237
#endif
5238 5239 5240 5241

	arch_ftrace_update_trampoline(ops);
}

5242
#else
5243

5244
static struct ftrace_ops global_ops = {
5245
	.func			= ftrace_stub,
5246 5247 5248
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE |
				  FTRACE_OPS_FL_INITIALIZED |
				  FTRACE_OPS_FL_PID,
5249 5250
};

5251 5252 5253 5254 5255
static int __init ftrace_nodyn_init(void)
{
	ftrace_enabled = 1;
	return 0;
}
5256
core_initcall(ftrace_nodyn_init);
5257

5258
static inline int ftrace_init_dyn_tracefs(struct dentry *d_tracer) { return 0; }
5259
static inline void ftrace_startup_enable(int command) { }
5260
static inline void ftrace_startup_all(int command) { }
5261
/* Keep as macros so we do not need to define the commands */
5262 5263 5264 5265 5266 5267
# define ftrace_startup(ops, command)					\
	({								\
		int ___ret = __register_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags |= FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
5268
	})
5269 5270 5271 5272 5273 5274 5275
# define ftrace_shutdown(ops, command)					\
	({								\
		int ___ret = __unregister_ftrace_function(ops);		\
		if (!___ret)						\
			(ops)->flags &= ~FTRACE_OPS_FL_ENABLED;		\
		___ret;							\
	})
5276

I
Ingo Molnar 已提交
5277 5278
# define ftrace_startup_sysctl()	do { } while (0)
# define ftrace_shutdown_sysctl()	do { } while (0)
5279 5280

static inline int
5281
ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip, void *regs)
5282 5283 5284 5285
{
	return 1;
}

5286 5287 5288 5289
static void ftrace_update_trampoline(struct ftrace_ops *ops)
{
}

5290 5291
#endif /* CONFIG_DYNAMIC_FTRACE */

5292 5293 5294 5295 5296 5297 5298 5299 5300 5301 5302 5303 5304 5305 5306 5307 5308 5309 5310 5311 5312 5313 5314
__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);
	}
	tr->ops->func = func;
	tr->ops->private = tr;
}

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

5315 5316
static inline void
__ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
5317
		       struct ftrace_ops *ignored, struct pt_regs *regs)
5318
{
5319
	struct ftrace_ops *op;
5320
	int bit;
5321

5322 5323 5324
	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;
5325

5326 5327 5328 5329 5330
	/*
	 * Some of the ops may be dynamically allocated,
	 * they must be freed after a synchronize_sched().
	 */
	preempt_disable_notrace();
5331

5332
	do_for_each_ftrace_op(op, ftrace_ops_list) {
5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344 5345 5346
		/*
		 * Check the following for each ops before calling their func:
		 *  if RCU flag is set, then rcu_is_watching() must be true
		 *  if PER_CPU is set, then ftrace_function_local_disable()
		 *                          must be false
		 *  Otherwise test if the ip matches the ops filter
		 *
		 * If any of the above fails then the op->func() is not executed.
		 */
		if ((!(op->flags & FTRACE_OPS_FL_RCU) || rcu_is_watching()) &&
		    (!(op->flags & FTRACE_OPS_FL_PER_CPU) ||
		     !ftrace_function_local_disabled(op)) &&
		    ftrace_ops_test(op, ip, regs)) {
		    
5347 5348
			if (FTRACE_WARN_ON(!op->func)) {
				pr_warn("op=%p %pS\n", op, op);
5349 5350
				goto out;
			}
5351
			op->func(ip, parent_ip, op, regs);
5352
		}
5353
	} while_for_each_ftrace_op(op);
5354
out:
5355
	preempt_enable_notrace();
5356
	trace_clear_recursion(bit);
5357 5358
}

5359 5360 5361 5362 5363
/*
 * 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.
5364 5365 5366
 * 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
5367 5368
 * being NULL, or CONFIG_DYNAMIC_FTRACE_WITH_REGS.
 * Note, CONFIG_DYNAMIC_FTRACE_WITH_REGS expects a full regs to be saved.
5369
 * An architecture can pass partial regs with ftrace_ops and still
L
Li Bin 已提交
5370
 * set the ARCH_SUPPORTS_FTRACE_OPS.
5371 5372 5373
 */
#if ARCH_SUPPORTS_FTRACE_OPS
static void ftrace_ops_list_func(unsigned long ip, unsigned long parent_ip,
5374
				 struct ftrace_ops *op, struct pt_regs *regs)
5375
{
5376
	__ftrace_ops_list_func(ip, parent_ip, NULL, regs);
5377 5378 5379 5380
}
#else
static void ftrace_ops_no_ops(unsigned long ip, unsigned long parent_ip)
{
5381
	__ftrace_ops_list_func(ip, parent_ip, NULL, NULL);
5382 5383 5384
}
#endif

5385 5386
/*
 * If there's only one function registered but it does not support
5387 5388
 * recursion, needs RCU protection and/or requires per cpu handling, then
 * this function will be called by the mcount trampoline.
5389
 */
5390
static void ftrace_ops_assist_func(unsigned long ip, unsigned long parent_ip,
5391 5392 5393 5394
				   struct ftrace_ops *op, struct pt_regs *regs)
{
	int bit;

5395 5396 5397
	if ((op->flags & FTRACE_OPS_FL_RCU) && !rcu_is_watching())
		return;

5398 5399 5400 5401
	bit = trace_test_and_set_recursion(TRACE_LIST_START, TRACE_LIST_MAX);
	if (bit < 0)
		return;

5402
	preempt_disable_notrace();
5403

5404 5405 5406 5407 5408 5409
	if (!(op->flags & FTRACE_OPS_FL_PER_CPU) ||
	    !ftrace_function_local_disabled(op)) {
		op->func(ip, parent_ip, op, regs);
	}

	preempt_enable_notrace();
5410 5411 5412
	trace_clear_recursion(bit);
}

5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426
/**
 * 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)
{
	/*
5427 5428
	 * If the function does not handle recursion, needs to be RCU safe,
	 * or does per cpu logic, then we need to call the assist handler.
5429
	 */
5430 5431 5432
	if (!(ops->flags & FTRACE_OPS_FL_RECURSION_SAFE) ||
	    ops->flags & (FTRACE_OPS_FL_RCU | FTRACE_OPS_FL_PER_CPU))
		return ftrace_ops_assist_func;
5433 5434 5435 5436

	return ops->func;
}

5437 5438 5439
static void
ftrace_filter_pid_sched_switch_probe(void *data, bool preempt,
		    struct task_struct *prev, struct task_struct *next)
S
Steven Rostedt 已提交
5440
{
5441 5442
	struct trace_array *tr = data;
	struct trace_pid_list *pid_list;
S
Steven Rostedt 已提交
5443

5444
	pid_list = rcu_dereference_sched(tr->function_pids);
5445

5446 5447
	this_cpu_write(tr->trace_buffer.data->ftrace_ignore_pid,
		       trace_ignore_this_task(pid_list, next));
S
Steven Rostedt 已提交
5448 5449
}

5450
static void clear_ftrace_pids(struct trace_array *tr)
5451
{
5452 5453
	struct trace_pid_list *pid_list;
	int cpu;
5454

5455 5456 5457 5458
	pid_list = rcu_dereference_protected(tr->function_pids,
					     lockdep_is_held(&ftrace_lock));
	if (!pid_list)
		return;
5459

5460
	unregister_trace_sched_switch(ftrace_filter_pid_sched_switch_probe, tr);
5461

5462 5463
	for_each_possible_cpu(cpu)
		per_cpu_ptr(tr->trace_buffer.data, cpu)->ftrace_ignore_pid = false;
S
Steven Rostedt 已提交
5464

5465
	rcu_assign_pointer(tr->function_pids, NULL);
S
Steven Rostedt 已提交
5466

5467 5468
	/* Wait till all users are no longer using pid filtering */
	synchronize_sched();
5469

5470
	trace_free_pid_list(pid_list);
5471 5472
}

5473
static void ftrace_pid_reset(struct trace_array *tr)
5474
{
5475
	mutex_lock(&ftrace_lock);
5476
	clear_ftrace_pids(tr);
S
Steven Rostedt 已提交
5477

5478
	ftrace_update_pid_func();
5479
	ftrace_startup_all(0);
5480 5481 5482 5483

	mutex_unlock(&ftrace_lock);
}

5484 5485
/* Greater than any max PID */
#define FTRACE_NO_PIDS		(void *)(PID_MAX_LIMIT + 1)
5486

5487
static void *fpid_start(struct seq_file *m, loff_t *pos)
5488
	__acquires(RCU)
5489
{
5490 5491 5492
	struct trace_pid_list *pid_list;
	struct trace_array *tr = m->private;

5493
	mutex_lock(&ftrace_lock);
5494 5495 5496
	rcu_read_lock_sched();

	pid_list = rcu_dereference_sched(tr->function_pids);
5497

5498 5499
	if (!pid_list)
		return !(*pos) ? FTRACE_NO_PIDS : NULL;
5500

5501
	return trace_pid_start(pid_list, pos);
5502 5503 5504 5505
}

static void *fpid_next(struct seq_file *m, void *v, loff_t *pos)
{
5506 5507 5508 5509
	struct trace_array *tr = m->private;
	struct trace_pid_list *pid_list = rcu_dereference_sched(tr->function_pids);

	if (v == FTRACE_NO_PIDS)
5510 5511
		return NULL;

5512
	return trace_pid_next(pid_list, v, pos);
5513 5514 5515
}

static void fpid_stop(struct seq_file *m, void *p)
5516
	__releases(RCU)
5517
{
5518
	rcu_read_unlock_sched();
5519 5520 5521 5522 5523
	mutex_unlock(&ftrace_lock);
}

static int fpid_show(struct seq_file *m, void *v)
{
5524
	if (v == FTRACE_NO_PIDS) {
5525
		seq_puts(m, "no pid\n");
5526 5527 5528
		return 0;
	}

5529
	return trace_pid_show(m, v);
5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540 5541
}

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)
{
5542 5543
	struct trace_array *tr = inode->i_private;
	struct seq_file *m;
5544 5545
	int ret = 0;

5546 5547 5548
	if (trace_array_get(tr) < 0)
		return -ENODEV;

5549 5550
	if ((file->f_mode & FMODE_WRITE) &&
	    (file->f_flags & O_TRUNC))
5551
		ftrace_pid_reset(tr);
5552

5553 5554 5555 5556 5557 5558 5559 5560
	ret = seq_open(file, &ftrace_pid_sops);
	if (ret < 0) {
		trace_array_put(tr);
	} else {
		m = file->private_data;
		/* copy tr over to seq ops */
		m->private = tr;
	}
5561 5562 5563 5564

	return ret;
}

5565 5566 5567 5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580
static void ignore_task_cpu(void *data)
{
	struct trace_array *tr = data;
	struct trace_pid_list *pid_list;

	/*
	 * This function is called by on_each_cpu() while the
	 * event_mutex is held.
	 */
	pid_list = rcu_dereference_protected(tr->function_pids,
					     mutex_is_locked(&ftrace_lock));

	this_cpu_write(tr->trace_buffer.data->ftrace_ignore_pid,
		       trace_ignore_this_task(pid_list, current));
}

5581 5582 5583 5584
static ssize_t
ftrace_pid_write(struct file *filp, const char __user *ubuf,
		   size_t cnt, loff_t *ppos)
{
5585 5586 5587 5588 5589
	struct seq_file *m = filp->private_data;
	struct trace_array *tr = m->private;
	struct trace_pid_list *filtered_pids = NULL;
	struct trace_pid_list *pid_list;
	ssize_t ret;
5590

5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601
	if (!cnt)
		return 0;

	mutex_lock(&ftrace_lock);

	filtered_pids = rcu_dereference_protected(tr->function_pids,
					     lockdep_is_held(&ftrace_lock));

	ret = trace_pid_write(filtered_pids, &pid_list, ubuf, cnt);
	if (ret < 0)
		goto out;
5602

5603
	rcu_assign_pointer(tr->function_pids, pid_list);
5604

5605 5606 5607 5608 5609 5610 5611
	if (filtered_pids) {
		synchronize_sched();
		trace_free_pid_list(filtered_pids);
	} else if (pid_list) {
		/* Register a probe to set whether to ignore the tracing of a task */
		register_trace_sched_switch(ftrace_filter_pid_sched_switch_probe, tr);
	}
5612

5613
	/*
5614 5615 5616
	 * Ignoring of pids is done at task switch. But we have to
	 * check for those tasks that are currently running.
	 * Always do this in case a pid was appended or removed.
5617
	 */
5618
	on_each_cpu(ignore_task_cpu, tr, 1);
5619

5620 5621 5622 5623
	ftrace_update_pid_func();
	ftrace_startup_all(0);
 out:
	mutex_unlock(&ftrace_lock);
5624

5625 5626
	if (ret > 0)
		*ppos += ret;
5627

5628
	return ret;
5629
}
5630

5631 5632 5633
static int
ftrace_pid_release(struct inode *inode, struct file *file)
{
5634
	struct trace_array *tr = inode->i_private;
5635

5636 5637 5638
	trace_array_put(tr);

	return seq_release(inode, file);
5639 5640
}

5641
static const struct file_operations ftrace_pid_fops = {
5642 5643 5644
	.open		= ftrace_pid_open,
	.write		= ftrace_pid_write,
	.read		= seq_read,
5645
	.llseek		= tracing_lseek,
5646
	.release	= ftrace_pid_release,
5647 5648
};

5649
void ftrace_init_tracefs(struct trace_array *tr, struct dentry *d_tracer)
5650
{
5651
	trace_create_file("set_ftrace_pid", 0644, d_tracer,
5652
			    tr, &ftrace_pid_fops);
5653 5654
}

5655 5656 5657 5658 5659 5660 5661 5662 5663 5664
void __init ftrace_init_tracefs_toplevel(struct trace_array *tr,
					 struct dentry *d_tracer)
{
	/* Only the top level directory has the dyn_tracefs and profile */
	WARN_ON(!(tr->flags & TRACE_ARRAY_FL_GLOBAL));

	ftrace_init_dyn_tracefs(d_tracer);
	ftrace_profile_tracefs(d_tracer);
}

S
Steven Rostedt 已提交
5665
/**
5666
 * ftrace_kill - kill ftrace
S
Steven Rostedt 已提交
5667 5668 5669 5670 5671
 *
 * 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.
 */
5672
void ftrace_kill(void)
S
Steven Rostedt 已提交
5673 5674 5675 5676 5677 5678
{
	ftrace_disabled = 1;
	ftrace_enabled = 0;
	clear_ftrace_function();
}

5679 5680 5681 5682 5683 5684 5685 5686
/**
 * Test if ftrace is dead or not.
 */
int ftrace_is_dead(void)
{
	return ftrace_disabled;
}

5687
/**
5688 5689
 * register_ftrace_function - register a function for profiling
 * @ops - ops structure that holds the function for profiling.
5690
 *
5691 5692 5693 5694 5695 5696
 * 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.
5697
 */
5698
int register_ftrace_function(struct ftrace_ops *ops)
5699
{
5700
	int ret = -1;
5701

5702 5703
	ftrace_ops_init(ops);

S
Steven Rostedt 已提交
5704
	mutex_lock(&ftrace_lock);
5705

5706
	ret = ftrace_startup(ops, 0);
5707

S
Steven Rostedt 已提交
5708
	mutex_unlock(&ftrace_lock);
5709

5710
	return ret;
5711
}
5712
EXPORT_SYMBOL_GPL(register_ftrace_function);
5713 5714

/**
5715
 * unregister_ftrace_function - unregister a function for profiling.
5716 5717 5718 5719 5720 5721 5722 5723
 * @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
5724
	mutex_lock(&ftrace_lock);
5725
	ret = ftrace_shutdown(ops, 0);
S
Steven Rostedt 已提交
5726
	mutex_unlock(&ftrace_lock);
5727 5728 5729

	return ret;
}
5730
EXPORT_SYMBOL_GPL(unregister_ftrace_function);
5731

I
Ingo Molnar 已提交
5732
int
5733
ftrace_enable_sysctl(struct ctl_table *table, int write,
5734
		     void __user *buffer, size_t *lenp,
5735 5736
		     loff_t *ppos)
{
5737
	int ret = -ENODEV;
5738

S
Steven Rostedt 已提交
5739
	mutex_lock(&ftrace_lock);
5740

5741 5742 5743 5744
	if (unlikely(ftrace_disabled))
		goto out;

	ret = proc_dointvec(table, write, buffer, lenp, ppos);
5745

5746
	if (ret || !write || (last_ftrace_enabled == !!ftrace_enabled))
5747 5748
		goto out;

5749
	last_ftrace_enabled = !!ftrace_enabled;
5750 5751 5752 5753

	if (ftrace_enabled) {

		/* we are starting ftrace again */
5754 5755
		if (ftrace_ops_list != &ftrace_list_end)
			update_ftrace_function();
5756

5757 5758
		ftrace_startup_sysctl();

5759 5760 5761 5762 5763 5764 5765 5766
	} else {
		/* stopping ftrace calls (just send to ftrace_stub) */
		ftrace_trace_function = ftrace_stub;

		ftrace_shutdown_sysctl();
	}

 out:
S
Steven Rostedt 已提交
5767
	mutex_unlock(&ftrace_lock);
5768
	return ret;
5769
}
5770

5771
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
5772

5773 5774 5775 5776
static struct ftrace_ops graph_ops = {
	.func			= ftrace_stub,
	.flags			= FTRACE_OPS_FL_RECURSION_SAFE |
				   FTRACE_OPS_FL_INITIALIZED |
5777
				   FTRACE_OPS_FL_PID |
5778 5779 5780
				   FTRACE_OPS_FL_STUB,
#ifdef FTRACE_GRAPH_TRAMP_ADDR
	.trampoline		= FTRACE_GRAPH_TRAMP_ADDR,
5781
	/* trampoline_size is only needed for dynamically allocated tramps */
5782 5783 5784 5785
#endif
	ASSIGN_OPS_HASH(graph_ops, &global_ops.local_hash)
};

5786 5787 5788 5789 5790 5791 5792 5793 5794 5795
void ftrace_graph_sleep_time_control(bool enable)
{
	fgraph_sleep_time = enable;
}

void ftrace_graph_graph_time_control(bool enable)
{
	fgraph_graph_time = enable;
}

5796 5797 5798 5799 5800
int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace)
{
	return 0;
}

5801 5802 5803
/* The callbacks that hook a function */
trace_func_graph_ret_t ftrace_graph_return =
			(trace_func_graph_ret_t)ftrace_stub;
5804
trace_func_graph_ent_t ftrace_graph_entry = ftrace_graph_entry_stub;
5805
static trace_func_graph_ent_t __ftrace_graph_entry = ftrace_graph_entry_stub;
5806 5807 5808 5809 5810 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 5825 5826

/* 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;
	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;
		}
	}

5827
	read_lock(&tasklist_lock);
5828 5829 5830 5831 5832 5833 5834
	do_each_thread(g, t) {
		if (start == end) {
			ret = -EAGAIN;
			goto unlock;
		}

		if (t->ret_stack == NULL) {
5835
			atomic_set(&t->tracing_graph_pause, 0);
5836
			atomic_set(&t->trace_overrun, 0);
5837 5838 5839 5840
			t->curr_ret_stack = -1;
			/* Make sure the tasks see the -1 first: */
			smp_wmb();
			t->ret_stack = ret_stack_list[start++];
5841 5842 5843 5844
		}
	} while_each_thread(g, t);

unlock:
5845
	read_unlock(&tasklist_lock);
5846 5847 5848 5849 5850 5851
free:
	for (i = start; i < end; i++)
		kfree(ret_stack_list[i]);
	return ret;
}

5852
static void
5853
ftrace_graph_probe_sched_switch(void *ignore, bool preempt,
5854
			struct task_struct *prev, struct task_struct *next)
5855 5856 5857 5858
{
	unsigned long long timestamp;
	int index;

5859 5860 5861 5862
	/*
	 * Does the user want to count the time a function was asleep.
	 * If so, do not update the time stamps.
	 */
5863
	if (fgraph_sleep_time)
5864 5865
		return;

5866 5867 5868 5869 5870 5871 5872 5873 5874 5875 5876 5877 5878 5879 5880 5881 5882 5883
	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;
}

5884
/* Allocate a return stack for each task */
5885
static int start_graph_tracing(void)
5886 5887
{
	struct ftrace_ret_stack **ret_stack_list;
5888
	int ret, cpu;
5889 5890 5891 5892 5893 5894 5895 5896

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

	if (!ret_stack_list)
		return -ENOMEM;

5897
	/* The cpu_boot init_task->ret_stack will never be freed */
5898 5899
	for_each_online_cpu(cpu) {
		if (!idle_task(cpu)->ret_stack)
5900
			ftrace_graph_init_idle_task(idle_task(cpu), cpu);
5901
	}
5902

5903 5904 5905 5906
	do {
		ret = alloc_retstack_tasklist(ret_stack_list);
	} while (ret == -EAGAIN);

5907
	if (!ret) {
5908
		ret = register_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
5909 5910 5911 5912 5913
		if (ret)
			pr_info("ftrace_graph: Couldn't activate tracepoint"
				" probe to kernel_sched_switch\n");
	}

5914 5915 5916 5917
	kfree(ret_stack_list);
	return ret;
}

5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 5932 5933 5934 5935 5936 5937 5938
/*
 * 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;
}

5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954
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)
{
5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973
	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)
5974
		ftrace_graph_entry = ftrace_graph_entry_test;
5975 5976
	else
		ftrace_graph_entry = __ftrace_graph_entry;
5977 5978
}

5979 5980 5981 5982
static struct notifier_block ftrace_suspend_notifier = {
	.notifier_call = ftrace_suspend_notifier_call,
};

5983 5984
int register_ftrace_graph(trace_func_graph_ret_t retfunc,
			trace_func_graph_ent_t entryfunc)
5985
{
5986 5987
	int ret = 0;

S
Steven Rostedt 已提交
5988
	mutex_lock(&ftrace_lock);
5989

5990
	/* we currently allow only one tracer registered at a time */
5991
	if (ftrace_graph_active) {
5992 5993 5994 5995
		ret = -EBUSY;
		goto out;
	}

5996 5997
	register_pm_notifier(&ftrace_suspend_notifier);

5998
	ftrace_graph_active++;
5999
	ret = start_graph_tracing();
6000
	if (ret) {
6001
		ftrace_graph_active--;
6002 6003
		goto out;
	}
6004

6005
	ftrace_graph_return = retfunc;
6006 6007 6008 6009 6010 6011 6012 6013 6014 6015

	/*
	 * 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();
6016

6017
	ret = ftrace_startup(&graph_ops, FTRACE_START_FUNC_RET);
6018
out:
S
Steven Rostedt 已提交
6019
	mutex_unlock(&ftrace_lock);
6020
	return ret;
6021 6022
}

6023
void unregister_ftrace_graph(void)
6024
{
S
Steven Rostedt 已提交
6025
	mutex_lock(&ftrace_lock);
6026

6027
	if (unlikely(!ftrace_graph_active))
6028 6029
		goto out;

6030
	ftrace_graph_active--;
6031
	ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub;
6032
	ftrace_graph_entry = ftrace_graph_entry_stub;
6033
	__ftrace_graph_entry = ftrace_graph_entry_stub;
6034
	ftrace_shutdown(&graph_ops, FTRACE_STOP_FUNC_RET);
6035
	unregister_pm_notifier(&ftrace_suspend_notifier);
6036
	unregister_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
6037

6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048
#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

6049
 out:
S
Steven Rostedt 已提交
6050
	mutex_unlock(&ftrace_lock);
6051
}
6052

6053 6054 6055 6056 6057 6058 6059 6060
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
6061
	/* make curr_ret_stack visible before we add the ret_stack */
6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088 6089 6090 6091 6092 6093 6094 6095
	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);
	}
}

6096
/* Allocate a return stack for newly created task */
6097
void ftrace_graph_init_task(struct task_struct *t)
6098
{
6099 6100
	/* Make sure we do not use the parent ret_stack */
	t->ret_stack = NULL;
6101
	t->curr_ret_stack = -1;
6102

6103
	if (ftrace_graph_active) {
6104 6105 6106
		struct ftrace_ret_stack *ret_stack;

		ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH
6107 6108
				* sizeof(struct ftrace_ret_stack),
				GFP_KERNEL);
6109
		if (!ret_stack)
6110
			return;
6111
		graph_init_task(t, ret_stack);
6112
	}
6113 6114
}

6115
void ftrace_graph_exit_task(struct task_struct *t)
6116
{
6117 6118
	struct ftrace_ret_stack	*ret_stack = t->ret_stack;

6119
	t->ret_stack = NULL;
6120 6121 6122 6123
	/* NULL must become visible to IRQs before we free it: */
	barrier();

	kfree(ret_stack);
6124
}
6125
#endif